[From the U.S. Government Printing Office, www.gpo.gov]
"The Ocean Challenge" Coastal Zone ,. Information Center ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~ jni, ' 1wg X~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~V ~u~~~~~ceanso~A VIW "The Ocean Challenge" Fourth Annual Combined The Marine Technology Society and Conference Sponsored by: The Institute of Electrical and Electronics Engineers *ep~ Council on Oceanic Engineering a Washington Section September 6-8, 1978 Participating Organizations Sheraton-Park Hotel 0 American Association for the Advance- Washington, D.C. ment of Science e American Geophysical Union A!-'~ * American Institute of Aeronautics and Astronautics * American Oceanic Organization * American Society of Civil Engineers * American Society of Mechanical ....... -'" .....Engineers ... L * American Society of Naval Engineers U P. $. [4EFARTME.ITI F, C( [,A 1Q~ * National Association of Corrosion COASTAL SF,\'R` ES CENTER Engineers 2234 S(DUli Hl E'SON AVENUE 0 National Ocean Industries Association CNHARLESTO N,0 S$SC 2�9405-241 l Sea Grant Association * Society of Naval Architects and Marine Engineers Chairman's Message iii Special Activities v OCEANS '78 Conference Committee vi Plenary and Policy Sessions vii Table of Contents ix Authors, Chairmen & Panelists xix Exhibitors xxi Sessions/Papers Start on page 1 78CH1356-6 OEC :Poperty of Csc Library Published by: The Marine Technology Society, 1730 M Street, N.W., Washington, D.C. 20036 and The Institute of Electrical and Electronics Engineers, Inc., 345 East 47th Street, New York, New York 10017 Responsibility of the contents rests upon the authors, and not upon the MTS, the IEEE, or their members. Copies of this Record may be purchased from MTS or IEEE. Reference publication 78 CH 0134-7 OEC Order from: MTS: 1730 M Street, N.W., Washington, D.C. 20036 IEEE: 445 Hoes Lane, Piscataway, NJ, 08854 All rights, including translations, are reserved by MTS and IEEE. Abstracting is permitted with mention of source. Request for republication should be addressed to The Marine Technology Society, 1730 M Street, N.W., Washington, D.C. 20036 or the IEEE Editorial Department, 345 East 47th Street, New York, N.Y. 10017. Copyright � 1978 by The Marine Technology Society. Library of Congress Card Number 74-76790 IEEE Publication Number 78 CH 0134-7 0EC Printed in the USA ii '~~cc no loM'Q*1 Chairman's Message The "challenge" chosen as the theme for the Oceans '78 Conference isto be daring, imaginative, and perseverant-to tap the vast resources of teoceans and to maintain the ocean's ecology for the benefit of man- kid. I am convinced that scientists and engineers and managers can over- come the problems inherent in understanding the behavior of the sea and - perfecting the technologies required for its preservation and development The technical papers presented at the Conference are solid evidence of progress. The displayed instruments and systems are further manifestatio that real strides have been made in the last decade in terms of equipment performance, reliability and economy. Although we have demonstrated our ability to meet distant and dif- ficult goals once we move forward, our capacity to focus on the correct objectives and to commit the necessary resources in the face of comn- peting needs is less clear. This aspect of the challenge is central to the selection of the series of policy discussions presented as part of this year's program. The opportunities for debate and questioning provided bythe Conference policy panels will play an important role in advancing the public dialogue on these issues. The Oceans '78 Conference is the fourth and final in a series sponsored jointly by the Marine Technology Society and the Institute of Electrical and Electronic Engineers. The members of the Conference Committee, assembled from these two technical societies, deserve special appreciation for their dedicated and effective efforts. We also particularly appreciate the help of the firms that joined us to exhibit the latest developments in marine technology. Finally, on behalf of the Conference Committee and the sponsoring organizations, I want to thank those attending the Conference for their active participation. Richard A. Frank Administrator National Oceanic and Atmospheric Administration Cecil Andrus Juanita Kreps Secretary of the Interior Secretary of Commerce Lloyd Maudlin James A. Rickard President President, Marine Technology Society Council on Oceanic Engineering, IEEE Richard A. Frank William M. Nicholson Administrator, National Oceanic and Office of Marine Technology, National Atmospheric Administration Oceanic and Atmospheric Administration iv Special Activities Plenary Session The conference opened with a keynote address by Secretary of the Interior Cecil D. Andrus who discussed the benefits and challenges associated with Government- Industry cooperation in ocean development. The more spe- cific issue of Government regulation was then discussed by a panel chaired by Richard A. Frank, Administrator, National Oceanic and Atmospheric Administration. Panel members presented their views of the effectiveness of governmental regulation of ocean industries, whether Government regulates too much or too little, and what the proper role of Government should be so that our Nation's goals for the oceans are best realized. Policy Program The Policy Program, an innovation in OCEANS '78, consisted of five separate 21/2 to 3 hour discussions and debates on critical oceans issues. Actual and potential policy of the Federal Government was the focus of the discussions for speakers and panels that ran concurrently with the Technical Program. The format consisted of speakers who presented differing points of view followed by a panel discussion and questions from the audience. The five subject areas were: - Outer Continental Shelf Oil and Gas: The desirable pace of OCS oil and gas development and the desirable degree of government regulation of oil and gas extraction from the Outer Continental Shelf. - Coastal Zone Management: This panel addressed the problems of consistency between Coastal Zone Management and Federally allowed activities in the oceans, with particular emphasis on the extent and degree of State control beyond the territorial seas. The goals and methods for management of the territorial seas were also discussed. - Ocean Mining: The developing ocean mining industry may need extensive governmental assistance. Is spe- cial assistance appropriate and, if so, what types of assistance, to whom, how much and when are some of the questions which were asked. This session also addressed the question of whether the U.S. should go ahead now with ocean mining or wait for an international treaty. -Fisheries Policy: This session focused first on issues related to limitations on domestic fishing activity for the purpose of maintaining optimal yield, and second on issues related to joint ventures, import limitations and other aspects of foreign trade. - Ocean Research & Development: The final Policy Session considered the complex questions involved in how our limited federal resources for scientific programs supporting ocean activities should be allocated. Senator Lowell P. Weicker (R-Conn.) provided a congressional viewpoint. Early Bird Reception- Richmond Room (Cash Bar); 4:00 p.m. to 8:00 p.m., Tuesday, September 5. Chairman's Luncheon-No Host Reception 11:45 a.m. to 12:30 p.m., Wednesday, September 6. Richard A. Frank, Administrator of the National Oceanic and Atmospheric Administration and Chairman of OCEANS 78, will address luncheon attendees. Participants of the Plenary Session are cordially invited to join conference attendees at this luncheon. Exhibit Hall Cocktail Party-5:00 p.m. to 7:00 p.m., Wednesday, September 6. Conference participants and exhibitors are invited, at this time, to a free cocktail party. Ocean Challenge Banquet-No Host Reception 6:15 p.m. to 7:00 p.m., Thursday, September 6. The banquet will be highlighted by a speech given by the Honorable Juanita Kreps, Secretary of Commerce. Acting as Master of Ceremonies will be Dr. John Craven, University of Hawaii. Presidents' Award Luncheon- 12:30 p.m., Friday, September 8. At this time, MTS President Dr. James A. Rickard will present four awards on behalf of the Society and IEEE Vice President Mr. Lester Hogan will present three IEEE awards. Oceans '78 Conference Committee Chairman RICHARD A. FRANK Arrangements DANIEL J. WEI LER Administrator Naval Ship Engineering Center National Oceanic and A tmospheric Administration DONALD CEBULSKI Naval Ship Engineering Center Vice-Chairman WILLIAM M. NICHOLSON Office of Marine Technology Conference Coordinator JAN BENNETT National Ocean Survey JB Inc. Na tional Oceanic and A tmos- pheric Administration Steering Committee JOSEPH R. VADUS Office of Ocean Engineering Publications PETER B. MYERS National Oceanic and Atmo- Magnavox spheric Administration Exhibits RICHARD M. SHAMP EDWARD M. MacCUTCHEON Engineering Service Associates LEONARD T. LYNCH, JR. Technical Program ANTHONY I. ELLER B-K Dynamics, Inc. Naval Research Laboratory ROBERT H. CASSIS Public Policy Program HEATHER L. ROSS U.S. Coast Guard Deputy Asst. Secretary- Policy, Budget & Administration J. BARRY OAKES Dept of Interior Johns Hopkins Applied Physics Laboratory RAY KARAM Dept of Interior ALVIN REINER Federal Communications PAUL STANG Commission Dept of Interior Finance RICHARD F. CROSS R. F. Cross Associates, Ltd. Publicity KATHI JENSEN College of Marine Sciences University of Delaware ROLAND D. PAINE Office of Public Affairs National Oceanic and A tmo- spheric Administration vi Plenary and Policy Sessions Session (0) Plenary Session This opening session of the Conference addressed the broad theme of its public policy program - how can government and industry best play out their respective roles so as to jointly realize for society the greatest possible benefits from the vast ocean resources which surround it. The Keynote Speaker set the stage with a presentation of guiding principles for public resources management policy. What is the government trying to achieve in its management of ocean resources, and how does it hope to achieve it? The panel followed with discussion of a major management tool which has recently been growing in public use and controversy - government regulation. Is government regulation tuned to realizing the greatest possi- ble social benefits from the oceans? Or has it gone too far? Or not far enough? Widely divergent views were heard from experts from the private and public sectors, with opportunity for questions from the audience. Session (6) Public Policy Outer Continental Shelf Oil and Gas The Nation's energy needs require the development of its OCS resources. But how fast should this be done? How can the need for domestic supplies of oil and gas be balanced against the potential costs of development? These costs include possible environmental degradation, conflicts with other ocean activities, and adverse onshore impacts. What pace of leasing is implied by a truly balanced approach? The speakers and panel first explored the pace of leasing questions, and then the statute authorizing regulation of oil and gas extraction. The OCS Lands Act amendments will specify new working relationships between industry, Federal and State Governments and increase the degree of control, reporting and consultation requirements on all who are associated with the offshore oil and gas program. Do the provisions in the law strike a proper balance so that oil and gas development can proceed while affected shorelines are adequately protected, competing users accounted for, and the environment preserved? Session (12) Public Policy Coastal Zone Management In this session the controversial effects of the Federal consistency provisions of the coastal zone management program were debated by representatives of industry and State and Federal governments. In the second part of the session two presentations and panel comments focused on the issues related to increasing State and Federal con- trols on ocean uses nearshore and seaward. Session (18) Public Policy Deep Ocean Mining The Ocean mining issues cut across economic, environmental, international and national security concerns. The focus of this session was the economic and international aspects. What are the arguments for and against spe- cial public support of ocean mining ventures? If support is warranted, in what form should it be provided? On the international front, are we better off going ahead now with ocean mining activity, or waiting for an international treaty? At what point, in terms of belated timing or adverse content of a treaty, does it make better sense for the U.S. to get out in front alone? The speakers and panel addressed these issues and answered questions from the floor. Session (24) Public Policy Debate: Fisheries The session focused on some of the key policy questions related to commercial fisheries. The first half of the session was devoted to a debate and discussion of the critical issues of limitations on domestic fishing activity for the purpose of maintaining optimum yield. The second half was broader, covering obstacles to growth, market structure, effects of foreign trade barriers and subsidies, joint ventures, appropriate rolls of Federal agencies, incen- tives needed for development, and other aspects of commercial fisheries management. Session (30) Public Policy - Ocean Research and Development One of the great challenges that face this nation in the next decade is how we handle the oceans in terms of its utilization and protection. Increased use of the oceans for such things as food, energy, recreation, and transpor- tation will require larger programs in marine research and development than now exist. Congress must, therefore, shift its attention and push for a National Oceans Policy and a proper organizational structure to administer it. The tremendous job ahead will be expensive however, and require new Federal funding patterns; but perhaps more im- portantly, new program priorities will be necessary so that Congress can meet its fiscal responsibilities. Vill Table of Contents Session 0. Plenary Session Session 2. Acoustic Systems Keynote Speaker: Honorable Cecil D. Andrus, Secretary of the Interior- Chairman: James F. Bartram, Raytheon Co. "Realizing Our Ocean Riches" Co-Chairman: Panel Chairman: Co-Chairman: Salvatore D. Morgera, Raytheon Co. Richard A. Frank, Administrator, NOAA 2A A Digital Survey Fathometer Providing a Chart Dis- Panelists: play and a Digital Output, J. Russell Hogan, James N. Barnes, Center for Law and Social Policy play and a Digital Output. J. Russell Hogan, Raytheon Co ............................ 42 Charles DiBona, Executive Vice President, American Petroleum 2B A Deterministic and Random Error Model for Institute a Multibeam Hydrographic Sonar, Dick Angelari, National Oceanic andAtmospheric Administration . 48 William D. Nordhaus, Member of the President's Council of Economic Advisers 2C Portable Acoustic Tracking Systems (PATS), Robert W. Gill, University of Washington; and William C. Gardner, Naval Facilities Engineering Command ............................. 54 Session 1. Buoy Technology 2D Swept Carrier Acoustic Underwater Communica- Chairman: tions, A. Zielinski and L. Barbour, Memorial Uni- Kenneth M. Ferer, Naval Ocean R&D Activity versity of Newfoundland .................. . 60 Co-Chairman: 2E Small Submersible Acoustic Communications Sys- Gene Edgerton, Western Instrument Corp. tem Design, Salvatore D. Morgera, Raytheon Co. .... 66 1A Evolution of Buoy Electronics and Telemetry, Jerry C. 2F Frequency Shift Arrival Time Determination of McCall and Edmund G. Kerut, NOAA Data Buoy Acoustic Signals, Edward W. Early, John E. Office; and George Haas and R. Mueller, Sperry Ehrenberg and Roundney K. Liem, University of Support Services ......................... Washington ............................ 72 1 B New Mooring Design for a Telemetering Offshore Oceanographic Buoy, Paul D. Higley and Arthur B. Joyal, Raytheon Co ...................... 10 Session 3. Ocean Sciences 1 C A Successful Meteorological Drifting Buoy Chairman: F. A. Reid, Hermes Electronics Ltd............. 18 Walter L. Brundage Jr., Naval Research Laboratory 1 D An Analog Wave Spectrum Analyzer System for Co-Chairman: Operational Buoys, Kenneth E. Steele, NOAA Data Patrick J. Hannan, Naval Research Laboratory Buoy Office; and A. S. Hananel, Sperry Support Services .............................. 22 3A Shipboard Evaluation of Drag Coefficients for the Marine Regime, C. Fairall, G. Schacker, T. Houlihan 1E A GOES Reporting Waverider Buoy, Louis C. Adamo, and K. Davidson, Naval Postgraduate School ...... 76 Louis C. Adamo, Inc., and Kenneth E. Steel, NOAA Data Buoy Office; and Ernest L. Burdette, Com- 3B The Mechanical Properties of Chitosan Membranes, puter Sciences Corp ....................... 31 B. L. Averbach and R. B. Clark, Massachusetts Institute of Technology ................... . 82 1 F Wave Direction Measurement by a Single Wave Follower Buoy, Wayne I. Sternberger, L. R. LeBlanc 3C Sediment as a cause of MACROCYSTIS gametophyte and F. H. Middleton, University of Rhode Island . 37 Non-Survival, Louis A. Volse, U.S. Coast Guard Academy ............................. 87 3D Navy Research on Marine Borers and the Laboratory Culturing of Limnorians, Kathleen Parrish and John D. Bultman, Naval Research Laboratory ..... 92 ix 3E The Motility of the Cells and the Extracellular 5D Great Lakes Shoreline Erosion-Western Lake Matrix of Labyrinthula-A Colonial Organism Michigan, J. M. Pezzetta, Ontario, Canada, and Found on Zastera Marina, Eugene Bell, Dennis J. R. Moore, University of Alaska ......... .... 143 Solomon, and Stephanie Sher, Massachusetts Institute of Technology .................... 99 5E Environmental Impact Assessment of Shoreline Bulkheading: A Historical Approach, Michael E. Penney, University of Rhode Island ............ 151 Session 4. Underwater Work Systems & Procedures Chairman: Session 6. Public Policy-Outer Continental Shelf Oil and Gas W. F. Searle, Jr., Searle Consultants Chairman: Co-Chairman: CTom Blockwick, MAR, Inc. Heather L. Ross, Deputy Assistant Secretary, Department of the Tom Blockwick, MAR, Inc. Interior 4A A Non-Acoustic Long Distance Underwater Com- munications System, R. G. Williscroft, NOAA Ship Speakers: Surveyor; and N. C. McLeod, TDC Underwater Frances Beinecke, Natural Defense Council, Lloyd Ottman, Systems .............................. 105 Shell Oil Company and Robert Bybee, Exxon Company, USA 4B Lift Line Dynamics: A Technique Tailored for the Panel: User, William S. Pope and Carl S. Albro, Battelle- Lowell D. Evian, Director, Georgia State Office of Planning Columbus Laboratories .................. 109 and Budget, Hope M. Babcock, Deputy Assistant Secretary Dept. of the Interior and Paul Kelly, VP. Zapata Corp. 4C External Work of Breathing of Demand-Air Under- water Breathing Apparatus, J. J. Grodski, Defence & Civil Institute of Environmen tal Medicine, Ontario . .............................. 115 Session 7. Instrumentation I 4D Underwater Repair of Nuclear Reactors, John D. Chairman: Peters, System Development Corporation; and Thomas M. Dauphinee, National Research Council, Canada Leon Manry, The Ocean Corp. ............... 122 Co-Chairman: 4E Underwater Stud Welding, Koichi Masubuchi, J. Barry Oakes, Applied Physics Laboratory, Johns Hopkins Uni- Hironori Ozaki, and Jun-ichi Chiba, Massachusetts versity Institute of Technology .................... 125 7A Drifting Buoys for Ocean Data Collection, 4F Development of a Design Baseline for Remotely John Garrett, Institute of Ocean Sciences ........ 155 Controlled Underwater Work Systems, Robert Lee Wernli, Naval Ocean Systems Center ............ 130 7B A Stable Platform for Micro-Scale Studies of Velocity, Temperature and Conductivity, Bill P. Johnson, Scripps Institution of Oceanography ..... 160 Session 5. Economics and Management of Coastal Regions 7C Marine Instrumentation: An Assessement of Tech- nology Versus Needs, S. H. Koeppen, P. C. Etter, Chairman: T. H. Hesselbecher and A. L. Inderbitzen, MAR, Milton Steinmueller, Dept. of Resource Development, Inc .................................. 164 Michigan State University 7D A Salinity Chain for Estuarine Use, E. L. Lewis 5A Coastal Recreation Values and Environmental Pertur- R. B. Sudar and Ron Perkin, Institute of Ocean bations, Elizabeth A. Wilman, Resources for the ............................. Future, Washington, D.C ................... 7E Development of a Small In Situ Conductivity 5B Coastal Demographic Changes in the U.S. 1950-1974, Instrument, Arthur M. Pederson and Michael C. Neils West, University of Rhode Island .......... 137 Gregg,APL, University of Washington ........... 173 5C Coastal Zone Management-A Decision Process, 7F Evaluation of Metal Film Temperature and Velocity Joel M. Goodman, Frederic R. Harris, Inc......... Sensors and the Stability of SPURV for Making Measurements of Ocean Turbulence, James D. Irish and Wayne E. Nodland, APL, University of Washington .................... Session 8. Fisheries Session 9. Unmanned Underwater Vehicles Chairman: Chairman: Fred L. Olson, National Oceanic and Atmospheric Norman Doelling, Massachusetts Institute of Technology Administration 9A Undersea Vehicle System Used for Photographic Co-Chairman: Reconnaissance of the Wreck of the USS Monitor, Francis M. Schuler, Jr., National Oceanic and Atmospheric Chris Tietze, Harbor Branch Foundation, Inc..... Administration 9C Trafficability and Performance Analyses for a Deep 8A Implications of the Japanese Experience in Aqua- Ocean Running, Gear Module for an Undersea culture Development for Thirty-Three Food-Short Vehicle, Herbert G. Herrmann, Naval Construction Countries, Yoshiaki Matsuda, Woods Hole Oceano- Battalion Center ......................... 723 graphic Institution ....................... 188 9D Underwater Profiling of Icebergs Using Submersibles, 8B Optimal Management of Global Tuna Resources, Boris Sukhov, NORDCO Ltd ................ 225 Man-Loong Chan, John Buchanan, Wendy Garling, and John Sasser, Tetra Tech .................. 199 9E Monitoring System for a New Ocean Cable Plow, Kazuomi Yamamura, Yoshinao Iwamoto, and 8C Acoustic Modeling of the Dolphin-Tuna Associative Yuichi Shirasaki, KDD Laboratories ............ 231 Behavior, James W. Fitzgerald and Richard D. Doolittle, WoodmontAssociates, Ltd.; and Phyllis H. 9F Display Aids for Remote Control of Untethered Cahn, Long Island University ............... Undersea Vehicles, William L. Verplank, Massachusetts Institute of Technology .................... 238 8D The Detection of Tuna by Sonar, Richard D. Doolittle and James W. Fitzgerald, Woodmont Associates, Ltd.; and Phyllis H. Cahn, Long Island University ................... .......... Session 10. Law and Policy 8E A Microcomputer-Based Echo Integration System Chairman: for Fish Population Assessment, Raynard Y. John Norton Moore, Center for Oceans Law & Policy, Kanemori and John E. Ehrenberg, University of University of Virginia Washington ............................ 204 10A Major Marine Legislative Actions in the 95th Congress, 8F Prospects for the Development of a SWATH Fishing Alfred W. Anderson, National Oceanic and Atmo- Vessel, Mark S. Rice and J. B. Harmon, SEACO, Inc. 208 spheric Administration ..................... 242 8G Developments in the Law - The Fishery Conservation 10B Ocean Area Management-An Emerging Concept, LeBoeuf, Lamb, Leiby & MacRae ............. 214 10C The Economics & Technology of an Integrated 8H Input-Output Analysis for Salt Marsh Bioproductivity, Ocean Surveillance System for Enforcement, Thomas Leschine and Leah J. Smith, Woods Hole Norman Louis Stone, Sanders Associates, Inc. ..... 248 Oceanographic Institution .................. 219 10D The Marine Sanctuary Program: Implications for Ocean Management, Carl S. Pavetto, National Oceanic and Atmospheric Administration .............. 10E Cultural Resource Management in the Coastal Zone, Daniel Koski-Karell, The Karell Institute ......... 255 10F The Impact of the 1978 Geneva Session of UNCLOS III on Ocean Development, Lewis M. Alexander, Francis X. Cameron, and Dennis Nixon, University of Rhode Island ............... 259 10G International Practice in Implementing 200 Mile Zones, Myron Nordquist, Nossaman, Krueger & Marsh ................................ xi Session 11. Electromechanical Cables/Connectors and Panelists: Their Components Robert W. Knecht, Michael E. Shapiro, National Oceanic and Atmospheric Administration, and Deb Stirling, Staff, Senate Chairman: Committee on Commerce, Sciences and Transportation, G. G. Rodriguez, David Taylor Naval Ship R&D Center William Frick, Van Ness, Feldman & Sutcliffe Co-Chairman: Part II Robert Forbes, David Taylor Naval Ship R&D Center Comprehensive Coastal Water Management and ocean uses 11A Forced Motions of a Cable Suspended from a Speakers: Floating Structure, Raymond I. Bisplinghoff, Tyce John Armstrong, University of Michigan, and Charles Williams, Laboratories; and David O. Libby, Simplex Wire & C. W. Williams, Inc. Cable Co. ............................. 263 Panelists: 11B Effects of Long Term Tension on Kevlar Ropes: Robert W. Knecht, Deb Stirling, G. William Frick and Paul Stang, Some Preliminary Results, Thomas P. Bourgault, Dept. of Interior Naval Underwater Systems Center ............. 268 11C Performance/Failure Analysis of Acoustic Array Connectors and Cables After 6-10 Years of Service, Session 13. Instrumentation II Colin J. Sandwith, APL, University of Washington. .. 273 Chairman: 11iD Flow-induced Transverse Motions of a Flexible J. Barry Oakes, Applied Physics Laboratory, Johns Hopkins Cable Aligned with the Flow Direction, R. J. Hansen and C. C. Ni, Naval Research Laboratory ........ . 287 Co-Chairman: 11E The State of Technical Data on the Hydrodynamic Thomas M. Dauphinee, National Research Council, Canada Characteristics of Moored Array Components, J. H. Pattison and P. P. Rispin ................... 291 13A The Development and Testing of Current Meters for Long Term Development on the Continental Shelf, 11 F Mooring Component Performance Kevlar Mooring Kenneth I. Heldebrandt, Texas Instruments; and Lines, G. A. Fowler and R. Reiniger, Bedford Eduardo D. Michelena, NOAA Data Buoy Office; Institute of Oceanography and Leslie S. Trest, Sperry Support Services ....... 308 11G Specifying and Using Contra-Helically Armored 13B Charting of Outer Continental Shelf Surface Cur- Cables for Maximum Life and Reliability, Albert G. rents by Aerial Tracking of Tracers, Timothy L. Berian, The Rochester Corp .................. 302 Flynn, Aero-Marine Surveys, and David O. Cook, Raytheon Co. .......................... 315 13C An Automated Tide Gauging Network, Donald C. Session 12. Public Policy-Coastal Zone Management and Knudsen, Ocean & Aquatic Sciences ............ 728 Ocean Use 13D An Underwater Towed Electromagnetic Source for Chairman: Geophysical Exploration, Barry E. Tossman, W. A. Robert W. Knecht, Acting Assistant Administrator for Ocean Swartz, and D. L. Thayer, Applied Physics Labora- Coastal Zone Management, National Oceanic and Atmospheric tory, Johns Hopkins University ............... 321 Administration 13E Undersea Stereoscopic Television, R. B. Fugitt and Part I R. S. Acks, Remote Ocean Systems, Inc .......... 322 Federal Consistency Provisions of the CZM Act 13F The Comparison of Discus Buoy Wave Spectra Pro- Introduction: duced by Buoy-Fixed and Vertically Stabilized Michael E. Shapiro, National Oceanic and Atmospheric Admini- Accelerometers, Ernest L. Burdette, Computer stration Sciences Corp ........................... 327 Speakers: James Ross, State of Oregon Phillip A. Clark, American Petroleum Institute xii Session 14. Acoustic Sources and Sonars Session 16. Remote Sensing from Satellites and Aircraft Chairman: Chairman: Charles W. Votaw, Naval Research Laboratory David E. Weissman, Hofstra University 14A A Small, Wide-band, Low-frequency, High-power Co-Chairman: Sound Source Utilizing the Flextensional Transducer Donald M. Bolle, Brown University Concept, J. Pagliarini and R. White, Sanders As- sociates, Inc. ........................... 333 16A Mission Horizons for Oceanic Environmental Monitor- ing, S. Walter McCandless Jr., National Aeronautics 14B An Analysis of the Distribution of Measured Bubble and Space Administration Pulse Periods of Explosive Sources, Richard J. Hecht and Louis A. Mole, Underwater Systems, Inc. ..... 339 16B Oceanographic & Geodetic Surveillance-The Expanding Role of Satellite Radar Altimetry, 14C Statistical Measures of Beam Output Noise Due to Tom Godbey, General Electric Co............ Shipping, M. Moll, R. M. Zeskind, and W. Scott, Bolt, Beranek and Newman, Inc ............... 345 16C SAR Imaging of Waves in Ice, Byron R. Dawe and Surendra K. Parashar, Memorial University of 14D A Wideband Sonar Equation, L. W. Brooks, Binary Newfoundland .......................... 379 Systems, Inc ............................ 350 16D The Comparison Between Synthetic Aperture 14E Synthetic Aperture Imaging Using Acoustical Radar Imageries and Surface Truth of Ocean Waves, Holography, Woon Siong Gan, Acoustical Services S. Vincent Hsiao, Jet Propulsion Laboratory ...... 385 Pte, Ltd., Singapore ...................... 355 16E Laser Bathymetry for Near Shore Charting Applica- tion (Preliminary Field Text Results), Gary C. Guenther, Lowell R. Goodman, and David B. Enabnit, Session 15. Large-Scale Development of Ocean Energy NOAA/National Ocean Survey; and Robert N. Swift Resources and Robert W. Thomas, EG&G ............. .. 390 Chairman: 16F Regional Integration, Validation and Application of Owen M. Griffin, Naval Research Laboratory SEASAT-A Satellite Scatterometer Data, Victor E. Delnore, Rutgers University ................. 397 15A Environmental Considerations for Siting an Ocean Thermal Energy Conversion Early Ocean Testing Platform at Four Proposed Areas, M. Dale Sands, Session 17. Institutional Aspects of Ocean Development Interstate Electronics Corp. and Pat Wilde, Lawrence Berkeley Laboratories ..................... 358 Chairman: Henry H. Esterly, New York City Community College 15B Large Scale Wave Energy-A New Approach, F. R. Haselton, Imagineering ...................C. o-Chairman: Alexander Bigler, Environmental Interface 15C Technology Considerations in the Design of a Com- mercial Offshore Energy Conversion (OTEC) Discussants: Plant, William W. Rogalski, Jr. and Robert J. Scott, Kirill Chekotillo, United Nations Gibbs and Cox, Inc.; and Julio G. Giannotti, Arvid Pardo, University of Southern California Giannotti and Buck Associates .........36.. 363 17A Regional Aspects of Ocean Management, Lewis M. 15D Philosophy of Monitoring, Inspection and Repair for Alexander, University of Rhode Island . 398 Fixed Offshore Platforms, Peter W. Marshall, Shell Oil Company ........................... 369 17B The Proposed International Seabed Authority: A Unique Institution, Samuel R. Levering, U.S. 15E A Proposal Concerning Total Energy Self-Supply Plan Committee fns, S auel R. Levering , D.C. and From Ocean Surfaces in Japan, Yoshiro Kasahara, Lee Kimball, United Methodist LOS Project ....... 401 Kinki University, Osaka ................... 377 17C Hydrocarbons on the Continental Margins-Some of the Issues Addressed at the Current Law of the Sea Treaty Negotiations, John N. Garrett, Gulf Energy & Minerals Co........................... 17D The Impact of UNCLOS ill on Marine Scientific Research, William L. Sullivan, Jr., U.S. Dept. of State, Office of Marine Science & Technology Affairs ............................... 403 xiii 17E CAESAR, SOSUS and Submarines. Economic and Session 20. Marine Pollution Analysis & Monitoring Institutional Implications of ASW Technologies, Harvey B. Silverstein, University of South Carolina . 406 Chairman: Stanley G. Chamberlain, Raytheon Co. 17F The Economic Potential of the Oceans and African States, Penelope Ferreira, Duquesne University .... 411 20A Advancement in Oil Spill Trajectory Modelling, David R. Garver, Dow Chemical Co., and Glen N. Session 18. PUBLIC POLICY-Deep Ocean Mining Williams, Texas A&M University .............. 441 Chairman: 20B An Objective Procedure for Evaluating Pollutant Charles P. Eddy, Deputy Assistant Secretary, Department Dispersion and Flushing in Tidal Canals and of the Interior Optimizing Canal System Design, Frederick W. Morris, IV, University of Florida ................... Speakers: Leigh S. Ratiner, Dickstein, Shapiro and Morin 20C Marine Pollution in Kuwait, B. L. Oostdam, Millers- ville State College ........................ 445 Helen B. Junz, Deputy Assistant Secretary of the Treasury 20D Stratified Water Circulation Forecasting for Water Samuel R. Levering, U.S. Committee for the Oceans Quality Management Planning in the Coastal Zone, Stanley G. Chamberlain, Prekimi V. Tawari, and Panel: William J. Comery, Raytheon Co. ........... .. 451 Alan Berlind, Director, Law of the Sea Office, Department of State 20E Ocean Disposal in the Mid-Atlantic Bight, William C. Muir, Gilbert Horwitz, and George D. Pence, Jr., Thomas R. Kitsos, Staff House Committee on Merchant Environmental Protection Agency ............. 456 Marine and Fisheries U.S. House of Representatives 20F Application of Sonar Techniques to the Surveillance H. William Menard, Director, U.S. Geological Survey and Enforcement of Dredged Material Disposal Operations, Gilbert L. Chase, U.S. Army Corps of Engineers . Session 19. Satellite Radiometric & Visible Sensing 20G Geological Impacts on Sludge Dumping on the Continental Shelf: Estimation by Math Modelling, Chairman: David O. Cook, James W. Demenkow, and Priya J. James J. Gallagher, Naval Underwater Systems Center Wickramaratne, Raytheon Co 19A Satellite Imagery and the Geomorphology of Bermuda, Stanley A. Ashmore, Defense Mapping Agency ............................ ... 417 Session 21. Navigation 19B Utility of Polarized Visible Radiation in Oceanog- Chairman: raphy, H. V. Soule, OAO Corporation .......... 420 Karl R. Schroeder, U.S. Coast Guard 19C The European Space Agency-its Ancestry, Ob- Co-Chairman: jectives and Programs, W. J. Mellors, European Space B. Harl Romine, U.S. Coast Guard Agency ............................... 21A Submersible Acoustic Navigation for Precise Under- 19D Satellite Ocean-Related Imagery Applications water Surveys, R. C. Abbott, AMF Sea-Link Sys- Program, Edward Wedler and Richard D. Worsfold, terns ................................. 462 Memorial University of Newfoundland .......... 427 21B Optimal Estimation of Undersea Acoustic Trans- 19E Some Automatic Digital Techniques for Enhancing ponder Locations, David G. Carta, Jet Propulsion and Extracting Oceanic Features from Satellite Laboratory ............................ 466 Infra-red Pictures, D. G. Nichol, Naval Underwater System Center .......................... 433 21 C Accurate and Continuous Geophysical Navigation Control with Newly Developed Radio System, 19F Satellite Interferometry for Ocean Surveillance, L. E. Ott, NA V-CON, Inc. .................. 472 Richard S. Orr, Stanford Communications, Inc .... 437 21D How to Add Satellite Navigational Accuracy and Stability to Existing Offshore NAV Systems, Phineas J. Icenbice, JMR Instruments ........... 478 xiv 21 E Performance of a Real World Minimum Variance Session 23. Information & Data Systems Integrated Navigation System, Ronald F. McPherson, Sperry Marine Systems ......... ... 485 Chairman: Salvatore D. Morgera, Raytheon Co. 21 F A Satellite-Based System for Redundant Maritime Distress Signalling, Walter C. Scales, The MITRE Co-Chairman: Corporation ............................ 490 James F. Bartram, Raytheon Co. 23A Microprocessor Acoustic Data Buoy, Robert C. Spindell, Kenneth R. Peal, and Donald E. Koelsch, Session 22. Sea Floor Engineering Woods Hole Oceanographic Institution .......... 527 Chairman: 23B A Low Power Microprocessor System for Deep-Sea Anton L. Inderbitzen, MAR, Inc. Free Vehicle Applications, Richard E. Ackermann, C. M. Butler and R. F. Weiss, Scripps Institution Co-Chairman: of Oceanography ........................ 532 Terrence McGuinness, National Oceans and A tmospheric TerAdministr ness, National Oceans andAtmospheric 23C Real-Time Oceanographic Data from Georges Bank, Robert N. Lobecker, Earl R. Lind, Gary Milano, and 22A Adhesive Resistance of Underconsolidated Sediments, Ronald Franklin, Raytheon, Co. .............. 536 T. R. Chari, S. N. Guha, and K. Muthukrishnaiah, Memorial University of Newfoundland .......... 23D Use of Industry Standards for Shipboard Data Systems, Kenneth R. Peal and Albert M. Bradley, 22B Submarine Landslides-East Coast Continental Slope Woods Hole Oceanographic Institution .......... 547 and Upper Rise, Alexander Malahoff and Richard B. Perry, National Ocean Survey/NOAA; and Robert W. Perry, National Ocean Survey/NOAA; and Robert W. 23E A Microcomputer Based Ocean Observation System, Embley, Lamont-Doherty Geological Observatory .. 503 Domingo Munoz, San Jose State University ....... 552 22C Design of an Ocean Bottom Seismometer with 23F Ships Thruster Allocation Logic With Rudder 22C Design of an Ocean Bottom Seismometer with Response from 25 Hz to 100 Seconds, Robert G. Interaction, Daniel O. Molnar and M. R. Cramer, Bookbinder, Arthur C. Hubbard, William G. Honeywell ............................. 556 McDonald, and Paul W. Pomeroy, Lamont-Doherty Geological Observatory .................... 510 22D Preliminary Results on Sub-Bottom Surveys at Session 24. PUBLIC POLICY-Fisheries Golfo De Campeche Area for Engineering Purposes, Jose Angle Aguilar-Castro and Pablo E. Torres- Salmeron, Compania Mexicana Aerofoto, S.A ...... Chairman: Terry L. Leitzell, National Marine Fisheries Service, National 22E A Unique Ocean Test Range, J. Blaine Davidson, Oceanic and Atmospheric Administration Florida Atlantic University .................. 516 I. Limited Entry 22F Broadband Seismic Data used for Seafloor Sediment Classification, A. D. Dunsiger, Memorial University Speaker: of Newfoundland; and R. R. Macisaac, Acres Virgil J. Norton, University of Maryland. ConsultingServices, Ltd ................ 521 Access Limitations as a Fisheries Management Alternative. Speaker: Jacob J. Dykstra, Point Judith Fishermen's Cooperative Limited entry is a regulatory nightmare which places unfair con- straints on U.S. fishermen. Panel: Terry Leitzell, National Marine Fisheries Service, National Oceanic and Atmospheric Administration, Christopher Koch, Senate Committee on Commerce, Science & Transportation, Norman C. Venzke, R. Adm., USCG, Lee Weddig, National Fisheries Institute. xv II. U.S. Commercial Fishing Industry Session 26. Novel Devices for Extracting Energy from the Ocean Speaker: Julio Pontecorvo, Columbia University Chairman: Current Problems in Fisheries Management Michael E. McCormick, U.S. Naval Academy Speaker: 26A Dynamic Dam for Harnessing Ocean and River Cur- Richard N. Sharoud, Nat'l Federation of Fisheries rents and Tidal Power, James T. Yen, Grumman Constraints on Grovath in the U.S. Fishing Industry in the Post Aerospace Corp., and John Isaacs, Scripps Institution 200 Mile Era. of Oceanography ......................... 582 Panel: 26B A Wave Activated Electric Generator, Thore Omholt, Terry L. Leitzell, National Marine Fisheries Service, National SUNY Maritime College .................... 585 Oceanic and Atmospheric Administration; Christopher Koch, Senate Committee on Science, Commerce & Transportation, 26C Energy from Sea Waves: System Optimization by Lee Weddig, National Fisheries Institute; John D. Negroponte, Diffraction Theory, Gaetano Sebastiani, Marco Berta, Deputy Asst. Secy State for Oceans & Fisheries Affairs and A. Blandino, Technomare, SpA ............. 590 26D A Hydrostatically Actuated Piston and Hydraulic System for Deep-Sea Free Vehicle Applications, Oskar H. Kirsten and R. F. Weiss, Scripps Institution Session 25. Education & Training for Ocean Involvement Oskar H . Kirsten and R. F.Weiss, Scripps Institution of Oceanography ........................ 596 Chairman: hMatirene S. Berryman, University of the District of 26E The Use of Ocean Energy In A Hydrostatic Motor, Matilene S. Berryman, University of the District of Columbia Stephen Selwyn and F. W. McCoy, Lamont-Doherty Geological Observatory .................... 599 25A The Marine Technologist, Roger M. Lloyd, Jr., Florida Junior College ................... .. 561 25B Education-The Bridge to Ocean Exploration, Session 27. Deep Sea Mining Matilene S. Berryman, University of the District of Columbia .............................564 Session sponsored by Marine Minerals Committee (MTS) and MS-5 Panel of SNAME 25C Undergraduate Preparation for a Career in Physical Oceanography, Victor E. Delnore, Rutgers Chairman: University .............................568 Raymond Kaufman, Deepsea Ventures, Inc. 25D Submarine Radiometry and Physical Environment 27A Activities & Economics of Existing Manganese Nodule Monitoring, Joel Vandenberghe, Institut Superior Mining Consortia, C. Richard Tinsley, Continental D'Electronique du Nord, France; and Robert Kling, Bank ................................ 602 Universite des Sciences and Techniques, France .... 571 27B A Cost Model for a First Generation Deep Ocean 25E Simplified Guidance for Engineers Associates with Mining Venture, G. D. Nyhart, MIT ............ 739 Design Construction and Certification of Hyperbaric Research Facilities, Terry Hayes, Naval Facilities 27C Technological Gaps in Deep Ocean Mining, John B. Engineering Command . . ................... 575 Herbich and John E. Flipse, Texas A&M University.. 606 27D Utilization of Available Technology to Support Ocean Mining, Sam Rothman and Robert J. Lynch, Jr., George Washington University ............. 611 27E Coastal Zone Implications of Deep Seabed Mining, M. Karl Jugel, NOAA Marine Minerals Division; and Daniel M. Bragg, TexasA&M University ......... 734 xvi Session 28. Wave Direction Measurement Technology Session 29. Financing Ocean Development Chairman: Chairman: Edward C. Brainard II, Environmental Devices Corp. James J. White, USNR (Retired) Co-Chairman: Co-Chairman: Marshall D. Earle, National Oceanic and Atmospheric George English, National Oceanic and Atmospheric Administration Administration 28A A System for the Routine Measurement of Directional 29A International Programs, Financing the International Wave Spectra from Large Discus Buoys, Kenneth E. Seabed Authority, Hartsell Cash, International Bank Steele, NOAA Data Buoy Office; Ernest L. for Reconstruction and Development, Washington, Burdette, Computer Sciences Corp.; and Anthony D.C ................................. Trampus, Federal Energy Regulatory Commission, Dept. of Energy ................... ...... 614 29B Offshore Petroleum Development Financing, Arthur Eschenlaur, Morgan Guaranty Trust Co .... 649 28B Directional Wave Spectral Data from a Large Discus Buoy, Ernest L. Burdette, Computer Science Corp.; 29C Financial Requirements for Ocean Thermal Energy Kenneth E. Steele, NOAA Data Buoy Office; Conversion, Frederick E. Naef and R. J. Pont, and Anthony Trampus, Federal Energy Regulatory Lockheed Corp .......................... 654 Commission, Dept. of Energy ............. ... 622 29D Cost Standards for Ocean Research Vessels, 28C Directional Wave Spectra Monitoring Horizontal Spar W. L. Martino, U.S. General Accounting Office .... 655 Buoy, Andrew Van Baird, Massachusetts Institute of Technology ............................ 629 29E Deep Sea Mining: Who Should Pay? Myron Nordquist, Nossaman, Krueger, and Marsh .... ........... 28D Measurement of Directional Ocean Wave Spectra from Fixed Platforms, Leon E. Borgman and Evangelos Yfantis, University of Wyoming ............. .. 634 Session 30. Public Policy Ocean Research & Development 28E Ocean Wave Detection and Direction Measurements with Microwave Radars, Paul G. Teleki, U.S. Geologi- Chairman: cal Survey; Wi liam McLeish and Duncan R oss, NOAA Dr. John B. Slaughter, National Science Foundation, Assistant Atlantic Oceanographic & Meteorological Labora- Director for Astronomical, Atmospheric, Earth, and Ocean tories; Robert A. Shuchman, Environmental Research Sciences Institute of Michigan; Walter E. Brown, Jr., Jet Pro- pulsion Laboratory; and Michael Mattie, U.S. Army Speaker: Corps of Engineers Hon. Lowell P. Weicker, Senator from Connecticut Panelists: Dr. G. Ross Heath, Dean of the School of Oceanography, Oregon State University, Corvallis, Oregon Mr. John R. Botzum, Editor, Nautilus Press Inc. Mr. Marvin Pitkin, Assistant Administrator for Commercial Development, Maritime Administration Mr. Otto Klima, Vice President and General Manager, Reentry and Environmental Systems Division, General Electric Company xvii Poster Sessions 1. A DYNAMIC HYDROTHERMAL EXPERIMENTAL STATION 657 Andrew Gustajtis, Memorial University of Newfoundland 2. DEPLOYMENT OF LONG-TERM SEAFLOOR MINERAL EXPOSURE EXPERIMENTS TO MEASURE CHANGES OF MINERALOGY AND COMPOSITION OF MANGANESE NODULES 662 R. G. Burns, V. M. Burns, H. W. Stockman and C. T. Stockman, MIT, and M. D. Benson, Scripps Institution of Oceanography 3. MOBILE INITIAL CONDITION INDICATOR, A DYNAMIC APPROACH TO MARINE EXPLORATION 668 Lawrence H. Befferman, Em-Log Enterprises 4. COMPREHENSIVE OCEANOGRAPHIC MONITORING PROGRAM IN THE ARABIAN GULF 674 Chung-Sung Ahn, Ebasco Services Inc. 5. A SYSTEMS APPROACH TO UNDERWATER CAMERA DATA CORRELATION AND INTERPRETATION 678 G. G. Hayward and W. D. McElroy, Jr., Benthos, Inc., and D. S. Hosom and W. M. Marquet, Woods Hole Oceanographic Institution 6. ICEBERG ALARM SYSTEM 682 R. Boetes, P. B. Sukhov, R. W. Crocker and W. E. Russell, Nordco, Ltd. 7. USE OF FREE FALL PENETROMETER IN SEAFLOOR ENGINEERING 686 T. R. Chari, W. G. Smith and A. Zielinski, Memorial Universiy of Newfoundland 8. THE DEVELOPMENT OF A SEAFLOOR EARTHQUAKE MEASUREMENT SYSTEM 692 David Ryerson and Eric Reece, Sandia Laboratories 9. MICROWAVE DETECTION OF ADSORPTION AND SPREADING FILMS AT THE AIR-WATER INTERFACE 696 R. Cini, Universita di Firenze and P. P. Lombardini, Universita di Torino 10. ELECTRODEPOSITION OF MINERALS IN SEAWATER 699 Wolf Hilbertz, University of Texas 11. ARCTIC DATA ACQUISITION AND PROCESSING SYSTEM E. O. Lewis, Canadian Department of Environment 12. JANUS IV CONSTRUCTION DIVING AT 460 M (1509 FT) NEW WORLD RECORD AT 501 M (1644 FT) 707 H. G. DeLauze and J. L. Chasserot, Comex Industries 13. SUBSURFACE LAYERED MEDIA PARAMETER ESTIMATION USING A LINEARIZED MULTIPLE REFLECTION MODEL 712 F. M. El-Hawary and W. J. Vetter, Memorial University of Newfoundland 14. R & D ACTIVITIES OF THE AMERICAN UNDERWATER BAND 718 Richard Bailey xviii Authors Index Page Abbott, R. C. 462 Flynn, T. L. 315 Maclsaac, R. R. 521 Ackerman, R. E. 532 Fowler, G. A. 297 Malahoff, A. 503 Acks, R. S. 322 Franklin, R. 536 Manry, L. 122 Adamo, L. C. 31 Fugitt, R. B. 322 Marquet. WV. M. 678 Ahn, C. S. 674 Gan, WV. S. 355 Marshall, P. W. 369 Albro, C. S. 109 Gardner, W. C. 54 Martino, WN. L. 655 Alexander, L. M. 259, 398 Garling, W. 199 Masubuchi, K. 125 Anderson, A.W. 242 Garrett, J. 155 Matsuda, Y. 188 Angelari, R. D. 48 Garver, D. R. 441 Mattie, M. 639 Ashmore, S. A. 417 G iannotti, J. G. 363 McCall, J. C. 1 Averbach, B. L. 82 Gill, R. WN. 54 McCoy, F. W. 599 Bailey, R. 718 Goodman, L. R. 390 McDonald, WN. G. 510 Barbour, L. 60 Gregg, M. C. 173 McElroy, WN. D., Jr. 678 Befferman, L. H. 668 Grodski, J. J. 115 Mc Leish, WN. 639 Bell, E. 99 Guenther, G. C. 390 McLeod, N. C. 105 Benson, M. D. 662 Guha, S. N. 497 McPherson, R. F. 485 Berian, A. G. 302 Gustajtis, A. 657 Michelena, E. D. 308 Berryman, M. S. 564 Haas, G. 1 Middleton, F. H. 37 Berta, M. 590 Hananel, A. S. 22 Milano, G. 536 Bisplinghoff, R. I. 263 Hansen, R. J. 287 Mole, L. A. 339 Blandino, A. 590 Harmon, J. B. 208 Moll, M. 345 Boetes, R. 682 Hayes, T. 575 Molnar, D. 0. 556 Bookbinder, R. G. 510 Hayward, G. G. 678 Moore, J. R. 143 Borgman, L. E. 634 Hecht, R. J. 339 Morgera, S. D. 66 Bourgault, T. P. 268 Heldebrandt, K. 1. 308 Mueller, R. 1 Bradley, A. M. 547 Herbich, J. B. 606 Muir, WN. C. 456 Bragg, D. M. 734 Herrmann, H. G. 723 Munoz, D. 552 Brooks, L. WN. 350 Hesselbecher, T. H. 164 Muthukrishnaiah, K. 497 Brown, W. E., Jr. 639 Higley, P. D. 10 Naef, F. E. 654 Buchanan,J.E. 199 H il bertz, W. H. 699 Ni, C. C. 287 Bultman, J. D. 92 Hogan, J. R. 42 Nichol,D1. G. 433 Burdette, E. L. 31,327,614,622 Horwitz, G. 456 Nixon, D. 259 Burns, R. G. 662 H osom, D. S. 678 Nodland, W. E. 180 Burns, V. M. 662 Houlihan, T. 76 Nyhart, G. D. 739 Butler, C. M. 532 Hsiao, S. V. 385 Omholt, T. 585 Cameron, F. X. 259 Hubbard, A. C. 510 Oostdam, S. L. 445 Carta, D. G. 466 Icenbice, P. J. 478 Orr, R. S. 437 Chamberlain, S. G. 451 Inderbitzen, A. L. 164 Ott, L. E. 472 Chan, M. L. 199 Irish, J. D. 180 Ozaki, H. 125 Chari, T. R. 497, 686 Isaacs, J. 582 Pagliarini, J. 333 Chasserot, J. L. 707 Iwamoto, K. 231 Parashar, S. K. 379 Chiba, J. I. 125 Johnson, B. P. 160 Parrish, K. M. 92 Cini, R. 696 Joyal, A. B. 10 Pattison, J. H. 291 Clark, R. B. 82 Jugel, M. K. 734 Peal, K. R. 527, 547 Comery, WN. J. 451 Kanemori, R. Y. 204 Pederson, A. M. 173 Cook, D. 0. 315 Kasahara, Y. 377 Pence, G. D., Jr. 456 Cramner, M. R. 556 Kerut, E. G. I Penney, M. E. 151 Davidson, J. B. 516 Kirsten, 0. H. 596 Peters, J. D. 122 Davidson, K. L. 76 Kling, R. 571 Pezzetta, J. M. 143 Dawve, B. R. 379 Knudsen, D. C. 728 Pomeroy, P. WN. 510 DeLauze, H. G. 707 Koelsch, D. E. 527 Pont, R. J. 654 Delnore, V. E. 397, 568 Koeppen, S. H. 164 Pope, W. S. 109 Dunsiger, A. D. 521 Koski-Karell, D. 255 Reece, E. WN. 692 Early, E. WN. 72 LeBlanc, L. R. 37 Reid, F. A. 18 Ehrenberg, J. E. 72, 204 Leschine, T. 219 Reiniger, R. 297 EI-Hawary, F. M. 712 Levering, S. R. 401 Rice, M. S. 208 Embley, R. WN. 503 Lewis, E. L. 169 Rispin, P. P. 291 Enabanit, D. B. 390 Libby, D. 0. 263 Rogalski, W. WN., Jr. 363 Eschenlauer, A. 649 Liem, R. K. B. 72 Ross, D. 639 Etter, P. C. 164 Lind, E. R. 536 Rothman, S. 611 Fairall, C. 76 Liloyd, R. M., Jr. 561 R ussel1, W. E. 682 Ferreira, P. 411 Lobecker, R. N. 536 Ryerson, D. E. 692 Fidell, E. R. 214 Lombardini, P. P. 696 Sands, M. D. 358 Flipse, J. E. 606 Lynch, R. J., Jr. 611 Sandwith, C. J. 273 Xxx Sasser, J. 199 Stockman, C. T. 662 Verplank, W. L. 238 Scales, W. C. 490 Stockman, H. W. 662 Vetter, W. J. 712 Schacker, G. 76 Stone, N. L. 248 Volse, L. A. 87 Scott, R. J. 363 Sudlar, R. B. 169 Wedler, E. 427 Scott, W. L. 345 Sukhov, B. 225, 682 Weiss, R. F. 532, 596 Sebastiani, G. 590 Sullivan, W. L., Jr. 403 Wernli, R. L. 130 Selwyn, S. 599 Swartz, W. A. 321 West, N. 137 Sher, S. 99 Swift, R. N. 390 White, R. P. 333 Shirasaki, Y. 231 Tawari, P. V. 451 Wilde, P. 358 Shuchman, R. A. 639 Teleki, P. G. 639 Williams, G. N. .441 Silverstein, H. B. 406 Thayer, D. L. 321 Williscroft, R. G. 105 Smith, L. J. 219 Thomas, R. W. 390 Worsfold, R. D. 427 Smith, W. G. 686 Tinsley, C. R. 602 Yamamura, K. 231 Solomon, D. 99 Tossman, B. E. 321 Yen, J. T. 582 Soule, H. V. 420 Trampus, A. 614,622 Yfantis, F. 634 Spindell, R. C. 527 Trest, L. S. 308 Zeskind, R. M. 345 Steele, K. E. 22, 31.614. 622 VanBaird, A. 629 Zielinski, A. 60,686 Sternberger, W. 1. 37 Vandenberghe, J. 571 Chairmen, Speakers and Panelists Session Andrus, C. D. PLENARY Evjen, L. D. 6 Norton, V. 24 Armstrong, J. 1 2 Ferer, K. M. 1 Oakes, J. B. 7,13 Babcock, H. M. 6 Forbes, R. I11 Olson, F. L. 8 Barnes, J. N. PLENARY Frank, R. A. PLENARY Otteman, L. 6 Bartram, J. F. 2,23 Frick, G. W. 1 2 Pontecorvo, J. 24 Beinecke, F. 6 Gallagher, J. J. 19 Ratiner, L. S. 18 Berl ind, A. 18 Griffin, 0. M. 15 Rickard, J. A. LUNCHEON Berryman, M. S. 25 Hannan, P. J. 3 Rodriguez, G. I11 Bigler, A. 17 Inderbitzen, A. L. 22 Romine, B. H. 21 Blockwick, T. 6 Junz, H. S. 18 Ross, H. L. 6 Bolle, D. M. 16 Kaufman, R. 27 Ross, J. 12 Brainard, E. C., 11 28 Kelly, P. 6 Schroeder, K. 21 Brundage, W. L., Jr. 3 K itsos, T. R. 18 Schuler, F. M., Jr. 8 Bybee, R. W. 6 Knecht, R. 12 Searle, W. F., Jr. 4 Chamberlain, S. G. 20 Koch, C. 24 Shapiro, M. 1 2 Clark, P. 1 2 Kreps, J. BANQUET Stang, P. 1 2 Cobb, W. R. 6 Leitzell, T. L. 24 Steinmueller, M. 5 Dauphinee, T. M. 7, 13 Levering, S. R. 18 Stirling, D. 1 2 DiBona, C. PLENARY Maudlin, L. Z. LUNCHEON Venzke, C. 24 Doelling, N. 9 McCormick, M. E. 26 Votawv, C. W. 14 Dykstra, J. 24 McGuinness, T. 22 Weddig, L. 24 Earle, M. D. 28 Menard, H. W. 18 Weissman, D. E. 1 6 Eddy, C. P. 1 8 Moore, J. N. 10 White, J. J. 29 Edgerton, G. 1 Morgera, S. D. 2, 23 Williams, C. W. 1 2 English, G. 29 Negroponte, J. 24 Esterly, H. H. 1 7 Nordhaus, W. D. PLENARY xx Exhibitors BOOTH AANDERAA INSTRUMENTS, INC. 217 ADVANCED CABLE AND ASSEMBLY 279 AMETEK CORP. 232-33 APPLIED MICROSYSTEMS LIMITED 227 BENTHOS, INC. 319-20 BIRNS AND SAWYER, INC. 230-31 BOOK EXAMINATION CENTER 288 CAYMAN-BELIZE INFORMATION CENTRE 268 COMPASS PUBLICATIONS, INC. 274 CONFERENCE BOOK SERVICE 301 CUBIC WESTERN DATA 277-78 DAMES AND MOORE 244-45 DEFENSE MAPPING AGENCY, WASHINGTON, D.C. See U.S. Navy DEL NORTE TECHNOLOGY, INC. 266-67 DIGICOURSE, INC. 269 EASTERN INSTRUMENTATION CORP. 262 ENCYCLOPAEDIA BRITANNICA 302 ENGINEERING SERVICE ASSOCIATES 200-16, 218-224 AMF Sealink J-TEC Associates Appleton Machine Lucker Manufacturing Co. D.G. O'Brien, Inc. Nova Scotia Research Foundation Diving Unlimited International Ocean Applied Research Environmental Devices Corp. Philadelphia Resins Franklin Electric Technology Development Corp. Giannini Instruments Tension Member Technology Huntec'70, Ltd. U.S. Phosmarine Hydracarta Corp Vector Cable Co. Hydrobios Sabins Corp. ENVIROMARINE SYSTEMS, INC. 318 ENVIRONMENTAL RESEARCH AND TECHNOLOGY 243 EOTECH 342 GANS AND PUGH ASSOCIATES, INC. 306-08 E.G. & G., Inc. Memodyne EPC Labs Rochester Corp. Harvey - Lynch Subsea Systems, Inc. Hydrolab GENERAL INSTRUMENT CORP. 256 GEOSERVICES, INC. 246-47 GLENAIR, INC. 258-59 GREAT LAKES ENVIRONMENTAL 228 GUILDLINE INSTRUMENTS, INC. 324 xxi HELLE ENGINEERING, INC. 336 HYDRONAUTICS, INC. 346 INFORMATION HANDLING SERVICES 303 INNERSPACE TECHNOLOGY, INC. 255 INTERNATIONAL UNDERWATER CONTRACTORS 260-61 JMR INSTRUMENTS, INC. 298 KLEIN ASSOCIATES, INC. 253-54 LOCKHEED CORPORATION 249, 250-52 MAGNAVOX GOVERNMENT AND INDUSTRIAL ELECTRONICS CO. 275-76 MAR, INC. 335 MARINE ENGINEERING/LOG 297 MARSH-MCBIRNEY, INC. 312 MCCLELLAND ENGINEERS, INC. 341 MOTOROLA, INC., GOVERNMENT ELECTRONICS DIV. 311 U.S. NAVY 325-8, 331-33, 337-39, 343-44 David Taylor Naval Ship R&D Naval Ship Engineering Center Naval Ocean Research and Development Activity Naval Research Laboratory Naval Oceanographic Center Office of Naval Research NOAA 290-91 NAVIGATION COMMUNICATIONS SYSTEMS 229 OCEAN INSTRUMENTS, INC. 272-73 OCEANIC SERVICES GROUP 239, 240-42 Cortland Line Loos and Co. EMP Industries Shape General Oceanics Telstar Electronics Hydro Search OIL SPILL INTELLIGENCE REPORT 348 OTIS ENGINEERING CORP. 323 PAROSCIENTIFIC, INC. 257 PERRY OCEANOGRAPHICS, INC. 236 PLESSEY ENVIRONMENTAL SYSTEMS 270-71 PREFORMED LINE PRODUCTS CO. 235 PUBLISHERS FOR CONVENTIONS, INC. 234 RADAR DEVICES, INC. 309 ROCKWELL INTERNATIONAL 225-26 SCHONSTEDT INSTRUMENT CO. 296 SEA DATA CORP. 329 SEA GRANT-MIT 345 SEABOARD INTERNATIONAL EQUIPMENT CO. 263 SIMRAD, INC. 299 SIPPICAN CORP. 304-5 SONATECH, INC. 287 SUB SALVE INDUSTRIES, INC. 248 TA INSTRUMENT CASE CO. 330 TRW SYSTEMS GROUP 264-65 TELEDYNE HASTINGS-RAYDIST 294 TELLUROMETER, USA 300 xxii TETRATECH INC. 317 TRACOR MARINE AND R.F. BUSBY ASSOCIATES 295 UNIVERSITY OF SOUTHERN CALIFORNIA 292 VIKING INDUSTRIES, INC. 237-38 WALL ROPE WORKS 347 WESMAR, OFFSHORE SYSTEMS DIVISION 321 WESTERN GEAR CORP. 282-83 WESTINGHOUSE ELECTRIC CORP. 284-86 ED WINSTON AND CO. 313-15 C.A. Richards and Associates Neil Brown Instruments Edo Western Sonotech J.W. Automarine Varian of Canada Kintec, Inc. xxIili EVOLUTION OF BUOY ELECTRONICS AND TELEMETRY J. C. McCall, E. G. Kerut - NOAA Data Buoy Office G. Haas, R. Mueller - Sperry Support Services National Space Technology Laboratories NSTL Station, MS 39529 Abstract cessor technology and suitable communication satel- In November 1967, the National Data Buoy Devel- lites, enables a fully automated and reliable approach opment Project was established by the United States for the acquisition of remote marine environmental Coast Guard. The primary purpose of this organization data on a synoptic basis. was to verify and standardize the design and perfor- 2. Current NDBO Programs mance of environmental data buoys by establishing an extensive network of sophisticated buoys capable of de- Moored Buoy Payloads livering a wide variety of data, supplemented by smal- ler buoys of simpler design for special user needs. I Three types of payloads are currently being main- rained and operated by NDBO: the Prototype Environ- 1972, user needs were reevaluated and the program was redirected to be more responsive to these needs. mental Buoy (PEB), and the Phase I and Phase I pay- Consequently, the goal of establishing a network of loads. Of the 18 deployed systems previously men- multipurpose buoys was discarded. In January 1973, tioned, 9 are PEBs, 4 are Phase I, and 5 are Phase II the project was designated the NOAA Data Buoy Office Payloads. In addition, development has been completed (NDBO) and was sanctioned by the National Weather on a low-cost Continental Shelf Buoy Payload (CSBP) Service as a prime user of data. based on advanced technology. Figure 1 shows the cur- rent NDBO deployment locations. A major NDBO mission is to develop the technolo- gy required to gather and report environmental data A major requirementis the continuous acquisition from marine areas on an accurate, reliable, and syn- and telemetry of near-real-time synoptic environmen- optic basis. The provisioning of this data would enable tal data to the weather forecasting community, includ- improved weather forecasting, including warnings of ing the National Weather Service (NWS). These data, impending catastrophies. The essential ingredient in along with other environmental data, are used for en- meeting the objective involves the deployment of a num- gineering evaluation and analysis and for the genera- ber of buoys moored in various gulfs and oceans. tion of monthly inputs to the National Oceanographic Data Center (NODC) and the National Climatic Center This paper will review the evolution of sensors, (NCC) for archival purposes. electronics, and communication hardware applicable to NDBO programs. We will also present an overview of our major programs so that buoy technology can be bet- ter appreciated in terms of the NDBO missions. 1. Background Since June 1972, NDBO has deployed environmen- tal reporting data buoys in various gulf and ocean re- - gions to provide synoptic data for weather reports and for scientific data archives. As of June 1978, 18 . moored buoys have been reporting environmental data ' on a routine basis. Accurate and reliable buoy sen- sors, electronics, and communications are required to properly support NDBO's mission of gathering and re- porting environmental data from the marine areas of : the world. The current complement of buoy electron- -.o ics was determined by a series of hardware evolutions, 43 which in turn were brought about by our continued quest / \ =- for the most accurate and reliable equipment that tech- o nology could offer. The evolution of the data buoy hardware, together with the availability of micropro- Figure 1. NDBO Ocean Data Station Locations Fig ue1DOOenDt tto oain Prototype Environmental Buoy (PEB) Payload. PEBs are currently in service in deep ocean The Prototype Environmental Buoy (PEB) design was areas, including the Gulf of Alaska, North Pacific, and based on concepts proven during initial NDBO pro- Atlantic. grams. Although initially tailored to meet the needs of I the weather community, its measuremPhase I and Phase II Payloads - The Phase II Pay- the weather community, its measurement capability was expanded to encompass surface wave data and sub- load program is an outgrowth of the Phase I Payloads surface water tenmperatures down to 300 meters. for small buoy systems that were developed in the 1972- 73 time period. These buoys were configured as Figures 2 and 3 show 10 and 12-meter diameter spheres and as horizontal and vertical cylinders, with hulls, respectively. Both types of hull support PEB typical diameters ranging between 4-1/2 to 5-1/2 feet. payloads. Meteorological sensors are located at the None of the small buoy systems achieved significant suc- 10-meter level. Measured values are telemetered via cess with regard to survivability and operability in the radio link to shore every 3 hours in a self-initiated, open sea. However, the data acquisition payloads were internally programmed mode. For special needs, quite reliable. e.g., during abnormal weather disturbances, the on- board weather data acquisition and reporting can be The Phase II Payloads are second-generation pay- changed to hourly operation by command from shore to loads that evolved from the small buoy technology. report hourly. Present and previous data frames can These payloads are presently integrated into the 12- also be acquired on demand via radio link. The on- meter hull and into an existing intermediate size hull, board acquisition and timing is controlled by a special- NOMAD, shown in Figure 4. (The NOMAD hulls were purpose computer, and batteries provide the power re- developed in previous U.S. Navy programs.) The Phase quired for up to 3 years of buoy operation. The pay- II Payload provides meteorological environmental pa- load is equipped with a dual HF/UHF communication rameters and surface wave data. The on-board data system. system. processing and data rf link are very similar to that of the PEB payload, employing nonprogrammable hard- The redundant rf links were implemented to en- ware and both HF and UHF satellite communications. sure a reliable and orderly transition from HF to UHF Five Phase II Payloads are in service in both deep ocean satellite communications. The conversion to UHF com- and continental shelf areas. Additional buoys are oper- munications is occurring as quickly as overall system ational using the earlier Phase I Payloads. considerations permit, in order to derive the advantage of this more reliable mode of communications. Continental Shelf Buoy Payload (CSBP) - The CSBP is in production and became operational this sum- mer. The overall system will draw extensively upon .iJ' -' proven hardware and system designs. For example, the entire meteorological sensor suite will be identical j Figure 2. Ten-Meter Discus Head Buoy Figure 3. Twelve-Meter Discus Bull Buoy ! 7~~~~~~~~ Generally, the drifting buoys are used for various scien- tific investigations, including such projects as arctic ice dynamics, Lagrangian ocean current measurements, and oceanographic and meteorological measurements at sea. These buoys are deployed worldwide. Ice Drifting Buoys - The ice drifting buoys are capable of monitoring environmental quantities in the deep polar regions to further the understanding of the the arctic ice and its environment. These buoys are generally equipped with pressure and temperature sen- sors, and use a UHF orbiting satellite relay data link for telemetry and position fixing. The daily changes in position provide the measurement of ice movements needed to assess the ice dynamics. Figure 5 shows an air-deployable ice buoy. Lagrangian Drifting Buoys - The primary purpose of this buoy is to achieve Lagrangian tracking of water parcels while simultaneously measuring basic meteoro- logical parameters. Tracking is enhanced by the use of a drogue which couples the surface buoy to a particular water layer depth. The drogue also reduces the wind drift of buoys. The meteorological sensor suite general- ly measures wind speed, air temperature, and air pres- Figure 4. NOMAD Hull sure, and is located less than 2 meters above the design to that used in the Phase II Payload. In addition, data waterline. Communications and position fixing depend upon the operation of a UHF transmitter on the buov input ports will be provided for future sensor additions,f including a wave measurement system and a multi-ele- ment thermistor line. The buoy communications will ~ ~ A' consist of the UHF transmitter and associated elec- tronics similar to that used on Phase II and PEB Pay- The CSBP will rely entirely on a UHF-satellite communications relay link, which will eliminate the need for the large buoy antennas required on previous i NDBO HF systems. However, the CSBP will feature a program-controlled microprocessor instead of the non- programmable special-purpose computers employed on previous payloads. This element, the Intel 8080 micro- processor, will acquire sensor data and digitally process and format the data. The microprocessor also controls the reception and transmission of data via the UHF radio set. The NOMAD hull, along with other hulls of oppor- AM6 tunity, will be used in conjunction with the CSBPs to acquire and report environmental data in various de- ployment areas. Twenty-four CSBPs will be procured, with de- ployment starting in the latter part of 1978 in the con- tinental shelf and deep ocean areas. Drifting Buoys Unlike moored buoys, whose prime purpose is the reporting of synoptic data to weather forecasting groups, drifting buoys provide information to the scientific com- munity, usually for nonsynoptic purposes. Occasionally, there is a need for near-real-time receipt of data (e. g., the use of a buoy to track oil spills as was done in the Figure 5. Air Deployable Ice Buoy case of the ARGO MERCHANT, which broke up on the Georges Bank just prior to New Year's Day in 1977). 3 which sends data to a specialized electronics package on 3. Evolution of Hardware board a polar-orbiting satellite. The satellite in turn relays the data to a ground station. Relative motion be- Sensors tween the satellite and buoy produces a Doppler effect on the frequency of the rf link. This frequency shift is Figure 7 presents an overview of the sensor re- processed along with certain satellite orbital parame- quirements for the three diverse NDBO payloads, i.e., ters to derive buoy position. Deep Ocean, Continental Shelf, and Drifting Buoys. FGGE Participation - NDBO will participate in the These sensors can conveniently be grouped into two First GARP Global Experiment (FGGE) scheduled from major categories: Meteorological and Oceanographic. June 1 through June 30, 1979. We have developed a TIROS Meteorological Drifting (TMD) buoy for the FGGE application. MOORED BUOY CONTINENTAL DRIFTING PAYLOADS SHELF BUOY The TMD buoy is small, free-drifting buoy capable PAYLOADS PAYLOADS of measuring barometric pressure and sea-surface temperature in the open-ocean environment. The buoy also monitors its internal temperature and battery voltage. In operational use during the Global Weather Experiment, the TMDs will acquire data on a periodic METEOROLOGICAL METEOROLOGICAL METEOROLOGICAL DATA DATA DATA (PARTIAL) basis and transmit via a UHF radio frequency up-link to the TIROS-N and NOAA-A polar-orbiting satellites, which are scheduled for launch, during 1978. SUBSURFACE OCEAN SUBSURFACE OCEAN OCEAN CURRENT TEMPERATURE DATA TEMPERATURE DATA TRACKING DATA Figure 6 shows the design configuration of the TMD buoy. The mechanical characteristics are as follows: WAVE & WAVE WAVE SPECTRA ICE MOTION Overall Length 10 feet (3.05 m) Figure 7. Sensor Requirements Vs. Buoy Payloads Maximum Diameter 27 inches (68. 6 cm) Spar Diameter 8 inches (20.3 cm) Meteorological Sensors - The meteorological pa- Deployed Weight 204+ pounds (92. 7 kg) rameters presently measured by the buoys include wind speed and direction, atmospheric pressure, and air Total Buoyancy 503 pounds (228. 6 kg) temperature. The initial design for wind speed and direction sensors consisted of a vortex-shedding ane- mometer and a wind vane integral with the anemometer. This device had a fluxgate magnetic compass within the base of the anemometer and provided a direct readout of wind direction with respect to magnetic North. A subsequent configuration consisted of a propeller ane- 1-SB~ l l l l lmometer and a wind vane integral with the anemometer. This design features an external magnetic compass, and requires an additional stage of wind direction res- olution (i. e., from wind direction with respect to the buoy to wind direction with respect to North). Both systems are currently installed on operational buoys. Air pressure transducers of the variable-capaci- tance type have been used exclusively by NDBO to mea- sure atmospheric pressure. The pressure-actuated member in these barometers is a diaphragm that de- fleets in accordance with the applied air pressure. This very small movement causes a change in the ca- pacitance of the sensing capacitor. Although the basic system has remained the same throughout NDBO's vari- ous programs, the atmospheric pressure sensors have been improved to provide enhanced long-term stability and less sensitivity to temperature and vibration. These improvements have been brought about by the Figure 6. TIROS Meteorological Drifting Buoy utilization of more advanced mechanical and electronic 4 elements and a packaging reconfiguration that isolated of the ocean sensor packages would neither accept nor the pressure capsule from the transducer electronics. retain a charge. Both platinum wire elements and thermolinear In reviewing the results of this experiment, the thermistor elements have been used by NDBO to sense following conclusions were drawn: air temperatures. Perhaps the major change for both the platinum and thermistor assemblies has been an * The inductive coupling concept is a valid improvement in shielding techniques used to reduce the technique for the transmission of infor- effect of incident and reflected solar and infrared radi- mation. ation. In addition, wherein previous units were power or fan aspirated, the newer designs are naturally wind 0 The NDBO-configured ocean sensor sys- aspirated. The air temperature transducer on some tem is not operationally feasible due to buoys is packaged integrally with the wind sensor pack- its high degree of complexity. ages. A separate wind-vaned shield arrangement is used on other buoys. * Reliable measurements of current speed and direction cannot be obtained from a Sensors for measuring solar and infrared radia- surface-moored buoy due to motion of tion, precipitation, and dew point were developed for the mooring line and the interference of the Engineering Experimental Phase (EEP) buoy sys- both the sensor and mooring line. tems. These sensors are not in use on the present more simplified operational buoys, but a request for In view of the limitations associated with the ocean these measurements has recently been made to NDBO sensor package, NDBO sought alternative solutions. by the National Weather Service. The mooring line/data line sensor package was sub- sequently replaced by a decoupled configuration in The meteorological sensors are installed at the 10- which a subsurface float isolated the ocean sensors meter level on the large 10- and 12-meter diameter from wave action. Currently, the inductively coupled discus hull buoys and at the 5-meter level on the upline telemetry is being replaced by acoustic links. NOMAD buoys. Originally, the EEP discus buoys had Ocean temperature is the only parameter being mea- meteorological sensors at both the 5- and 10-meter sured and reported operational at this time, but de- levels, but the lower sensor platform is no longer used. velopmental efforts are underway on the remaining oceanographic measurands prior to their incorpora- Subsurface Ocean Sensors - The development of tion as reporting parameters. subsurface ocean sensors was perhaps one of NDBO's most ambitious undertakings. The result of this effort Follow-on systems used thermistor strings to was the delivery of several inductively coupled ocean measure water temperature. Developed directly by sensors capable of being attached to a combination NDBO, the thermistor string is constructed of PVC mooring line/data line. Each sensor was packaged tubing and nylon tees filled with oil. Individual wires within a cylindrical pressure housing with hemispheri- for each thermistor run through the tubing. Breakouts cal end caps, and each contained individual oceano- for the thermistor are provided at the tees, and the graphic transducers and secondary power cells. The thermistors are potted and screwed into the tee. total package was 20-1/2 inches long and 12-1/2 inches in diameter. Wave and Wave Spectra Data Sensors - In the area of wave data, NDBO has experienced an orderly transi- In June 1975, NDBO deployed a 12-meter hull buoy tion from simple systems providing only wave height in the Gulf of Mexico. In addition to the standard and period, to microprocessor systems providing wave meteorological instrumentation, five inductively cou- spectrum outputs. The first wave measurement sys- pled ocean sensors attached to a mooring line/data line tem was essentially a buoy motion measurement device were included on this buoy. The ocean sensor packages for obtaining experimental data on the severity of the were installed at depths of 50, 100, 200, 350, and 500 environment within the buoy hull. The measurement meters, to measure conductivity, temperature, pres- device was enclosed in a gyro-stabilized platform from sure, and current direction. The 50- and 100-meter which acceleration and pitch and roll angles were de- packages were additionally equipped with current rived. Problems with the stabilized platform led to a transducers to measure current speed. strapped-down system. This system consumed con- siderably less power and was quite reliable, but fal- The buoy was retrieved in late January 1976. Ex- tered when it came to producing high-quality data. tensive damage was noted on the mooring line/data line at the oceanographic sensor package attachment points. Although an improved version of the strapped-down The extent of the damage was such that complete sepa- system was on the drawing boards, the requirement for ration of the packages would have occurred in a matter spectral data as well as wave height and period led to of weeks. Also, it was found that the internal batteries the parallel development of two other systems. One of these systems produces spectra using time-series in- mitted digitally to shore. puts to generate a covariance function on board a buoy microprocessor. At the shore processing site, the co- Special drift tests were conducted to determine variances are converted to displacement spectral den- the dynamic performance capability of the buoy-Omega sity by performing the Fourier transform on the covar- system. The results showed that, in processing the iance function. Integration of the spectra produces the digitized data on shore, the computer was unable to moments from which the wave height and period are maintain correct Omega lane counts. This required calculated. This system produces data of excellent manual manipulation of data. An rms error of 5.81 quality. kilometers was determined. The other wave measurement system resulted from The satellite position scheme used the Navys a requirement for a system that could generate coarse Transit system, otherwise termed the Navy Naviga- Transit system, otherwise termed the Navy Naviga- spectral density and reasonable wave height and period tional Satellite System (NNSS). The NNSS receiver de- data on board a platform without a sophisticated proces- signed for buoy installation has several important dif- sor. In this system, time-series inputs are derived ferences from standard commercialunits. The two from a strapped-down accelerometer. Twelve fre- prominent features are automatic operation and low quency-sensitive analog filter circuits receive the power consumption. The receiver begins a frequency time-series inputs, after which each output is search upon command and locks onto the satellite signal squared and smoothed. At the shore station, the when the satellite rises above the horizon. The Doppler data is processed with a matrix and generates a count begins at the appropriate time; at each 2-minute spectrum for acceleration, velocity, displacement, mark thereafter, a Doppler word is completed and and estimates of the average value of the displace- loaded into memory and a new Doppler count is started. ment spectral density in the band pass of each of This sequence continues until the signal is lost or eight the 12 analog channels. Doppler words are obtained. The buoy-NNSS system digitizes the Doppler data and transmits it to shore for The next area is the development of wave direc- processing. tional spectra. Recent demands by various govern- ment agencies for directional spectra has prompted Dynamic tests were performed using drifting plat- NDBO to pursue this development vigorously. An ex- forms and the on-board NNSS system. Only limited perimental system has been developed and tested that ephemeris data were used, with an attendant decrease will provide directional spectra in 15' directional in- in position accuracy. In addition, uncompensated buoy in position accuracy. In addition, uncompensated buoy crements. The experimental system calculates auto- crements. The experimental system calculates auto- motion added to the overall position accuracy error, to and cross-covariances of heave acceleration and wave yield a total rms error of slightly over 4 kilometers. slope in two orthogonal directions for a number of time displacements. Similar to the routine for nondirectional The most recent development underway is a stand- The most recent development underway is a stand- spectra, these functions are transmitted to shore where alone NNSS buoy receiver-microprocessor. This sys- alone NNSS buoy receiver-microprocessor. This sys- the spectra are constructed via the Fourier transform. tem will be capable of extracting ephemeris data di- tern will be capable of extracting ephemeris data di- The output data is three dimensional and includes dis- rectly from the satellite, computing on-board posi- rectly from the satellite, computing on-board post- placement spectral density, frequency, and direction. tion, and retransmitting processed position data to tion, and retransmitting processed position data to Position-Location Sensors - Position-location sen- shore. This will eliminate the need for long-haul wire- sors are required to track ocean currents, monitor ice line ephemeris updates and will reduce the length of motion, and to detect drift of moored buoys. Several the retransmitted message, since processed rather than raw data will be transmitted. approaches have been advanced and implemented to raw data will be transmitted. solve the overall position location problem. Initially, a hyperbolic very low frequency Omega system and a More recently, a Buoy Transmit Terminal (BTT) satellite position-fixing system were designed and was designed to operate on small data collection plat- satellite position-fixing system were designed and forms and in concert with the NIMBUS-F satellite Ran- evaluated for determining NDBO buoy position.doAcesMaumntStm(RS) Ingerl dom Access Measurement System (RAMS). In general, RAMS provides the capability for data collection and The Omega receiver, designed for unattended location determination for a large number of austere buoy operation, incorporated several features that were platform configurations. The BTT is particularly ap- platform configurations. The BTT is particularly ap- unavailable in general commercial units. Perhaps the plicable to scientific programs requiring non-real- plicable to scientific programs requiring non-real- most outstanding feature included an automatic syn- time reporting of low volume data from low-cost drift- chronizer capable of tracking the shoreline transmittert g om time-multiplexed sequence. The buoy receives a unique ing buoys. side frequency to synchronize a function generator to Platform position is derived from the relative the switching sequence of the composite Omega signal. motion between the platform and the satellite, which This synchronization is usually performed manually in can be measured by Doppler frequency shift patterns. commercial receivers, Up to four lines of position are All measurements and processing are centrally located, digitized by the on-board Omega system and retrans- 6 either in the satellite itself or at the earth tracking sta- only memories to correct either programming errors tion, thereby enabling the use of relatively simple sys- in the digital computer or memory failure or both. tems on the small platforms. The location accuracy for buoy applications is specified at 5 kilometer rms, Second-generation systems have had buoy data which has been verified by extensive testing and deploy- processors specifically designed for buoys. By sim- ments. plifying and incorporating data processors of the non- programmable "hard-wired" type, relatively fail-safe The BTT consists of two major functional sub- systems have been achieved. Greater simplicity has systems: a digital logic section and a solid-state trans- been achieved along with significant reduction in power mitter. The digital section interfaces with the sensor consumption. data, converts and formats the data into a proper trans- mission format, and encodes the data for PSK trans- Current and future NDBO payloads will rely on mission. The transmitter portion consists of a stable programmable microprocessors for the ultimate in oscillator circuit, a multiplier-phase shift key circuit, reliability, without the need to set excessive limits for and a power amplifier circuit. Prominent BTT charac- on-board processing. teristics include a power output of 2.4 watts, flexible sensor input interface options, data rates of 100 bits Communications per second, and 300 milliwatts average power. It is packaged to fit within a 4-inch pipe, 21 inches long. There are two general categories of communication requirements for NDBO programs. The first require- A BTT has also been developed to operate with ment is brought about by the timeliness feature that the TIROS-N and NOAA-A polar orbiters. The prom- is required for various weather forecasting groups, in- inent performance characteristics are similar to the eluding the National Weather Service (NWS). The NDBO units developed to operate with NIMBUS, except for weather message is approximately 10 seconds long when changes in bit-rate transmitter frequency and frame- transmitted at a 75 bit per second rate, and includes message format. All BTTs developed to date for the air pressure and temperature, wind speed and direction, TIROS-N will be used in FGGE and are installed in sea surface temperature, and wave data. NDBO buoys TIROS meteorological drifters. Additional TIROS-N are deployed in the Atlantic Ocean, Gulf of Mexico, BTTs are scheduled for drift detector application on North Pacific, and Gulf of Alaska. These diverse geo- moored buoys. graphic locations require over-the-horizon data links. Data Processing. The second category of communication require- ments is governed by the needs of the scientific com- The electronics associated with NDBO's first munity. These applications generally do not require buoys were developed with little experience in advanced the timeliness feature of the first type, although pro- state-of-the-art hardware design. Initially, the major s tate-of-the-art hardware d e sign. Initially, the major cessed data occasionally is desired in a matter of hours. requirement was that the buoys have extensive flexi- bility in order to perform a spectrum of engineering An important example of user applications would be and data acquisition tests necessary for establishing the oceanographers, who have for many years mapped state of the art of buoy technology. Initially designed ocean currents through Lagrangian tracking techniques. to meet what now appears to have been unrealistic data These oceanographers are interested in obtaining posi- buoy requirements, these complex payloads were ex- tion fixes from drifters that follow ocean currents. The pensive and not very reliable. By eliminating some of length of the position-fixing message is about 1 second. the excessive processing associated with flexibility and Another example of scientific users would be investiga- options, and by setting more realistic data accuracy tors concerned with wave data. Fleet Numerical Weath- requirements, the cost of the second-generation pay- er Central (FNWC) currently utilizes NDBO's wave loads was about one-third the cost associated with the spectral data from moored buoys as calibration points initial NDBO payloads. In addition, reliability was en- for their formulation of prediction models used to fore- hanced by the reduction in system complexity. cast sea states for the entire northern hemisphere. Work is underway to extend the wave measuring capabil- The initial buoy payloads had data processing sys- ity to drifting buoys. The length of the spectral wave tems constructed around a general purpose minicom- data is approximately 40 seconds. There are other puter (NOVA), which acted as a controller and data important applications of buoys for scientific uses, in- processor for the overall operation of the entire buoy eluding oil spill tracking and the First Worldwide Global including sensors, communications, and power systems. Weather Experiment. However, general data processors of the off-the-shelf variety used in buoy payloads have not performed well. The approaches applicable to the NDBO communica- Also, the minicomputers were relatively high average tion requirements are limited because of link range dis- power devices. Some improvement was experienced tances, buoy size and power limitations, available fre- by the incorporation of permanently programmed read- quencies, and the availability of cooperative satellites. The buoy-to-shore ranges are generally beyond the line 7 of sight. Discounting techniques such as meteor-burst elements throughout the system, including three separ- systems and tropospheric scatter, the fulfillment of the ate power amplifiers, antenna couplers, and receivers over-the-horizon link requirements quickly reduces to for each frequency of operation. The second-genera- either conventional HF or satellite communications. tion systems used common elements coupled with dy- namic switching. One-hundred watt amplifiers are High Frequency (HF) versus Satellite Communications employed almost exclusively on all NDBO platforms using HF communications. HF communication is a mature technology that has been extensively applied in long-range buoy communica- The HF antennas are basically variations of ver- tions. This technology is the mainstay of long-range tically polarized dipoles. Where existing masts can be military, government, international, and amateur oper- utilized as efficient radiators, they are used as antennas ations. However, when theoretically compared to satel- in a disc-stub configuration. Figures 2 and 3 show a lite communications, HF has decided disadvantages. model of such an arrangement. On platforms where no Since HF wavelengths are an order of magnitude longer suitable masts exist, a vertical whip approximately 21 than applicable UHF wavelengths, larger antennas are feet long is used. required for HF. Long-range HF paths are based upon real-time ionospheric support conditions and require op- UHF - The first-generation NDBO/Geostationary Op- timum frequencies for specific ranges. A reliable HF erational Environmental Satellite (GOES) unit was derived system entails working in multiple frequency bands to from a land-based design used to provide the communica- account for the dynamic diurnal/seasonal variations of tion interface between a remote land-based platform and the ionosphere. Overall, the dependence of HF com- the geostationary spacecraft. The land-based design was munications upon good ionospheric support for each developed from preliminary design tradeoffs with empha- buoy-to-shore path establishes an approximate year- sis placed on modulation, modulation rate, and frequency round 95-percent upper limit for basic communication band of operation. The results of this study indicated reliability. that the most productive link operation would be from 400 to 500 MlHz; and that, with regard to the modulation Satellite communications, employing frequencies scheme, differential Phase Shift Keying (PSK) would be above 400 MHz, propagate through the ionosphere and moderately superior over other schemes possessing are generally not subject to ionospheric effects. Once similar cost and complexity of technique. To facilitate established, the satellite link can operate over a dedi- timing and to take advantage of the spacecraft's ultra- cated channel. Since satellite communications utilize stable Cesium frequency standard, the earth station unit frequencies possessing one-tenth the length of cor- was designed to track spacecraft signals continuously, responding HF communications, the antennas can be thus providing a precise clock reference. The earth sta- built to one-tenth the length of corresponding HF an- tion unit can be adapted to receive an entire matrix of tenna elements. The single-band operations at UHF commands sent from the GOES spacecraft. enables these antennas to be naturally tuned, and eliminates the need for frequency-matching networks. NDBO modified an existing earth station unit to This eliminates a mode of failure and also reduces adapt it to buoy use. The major modifications included power requirements. Finally, a single ground station an increase in output power to compensate for loss of pointed antenna gain, the incorporation of an autorustic can adequately support a synchronous satellite over an pointed antenna gain, the incorporation of an automatic search and acquisition demodulator, the incorporation earth quadrant, whereas a network of shore stations is search and acquisition demoduator, the incorporation required for reliable HIF communications support for of a self-timed-only mode of operation, DC power the same area, conversion, and an appropriate mechanical redesign for ocean environment suitability. These comparisons were a convincing argument for choosing satellite communications, and formed the HF versus UHF Satellite Performance Comparisons basis for NDBO's quest for satellites and associated spectrum resources. The subject of HF versus UHF performance com- parisons is extensively treated in references 1 and 2. Communications Hardware Suffice it to say that, when properly designed and im- plemented, both HF and UHF can achieve composite HF - The hardware currently used in NDBO buoys link reliabilities in excess of 95%. However, UHF belongs to a set of configurations derived by hardware satellite communications does provide a clear advan- evolution. All configurations are capable of multiple- tage over HF. The following is a capsule summary of frequency operation, generally employing three and the comparisons: occasionally four frequencies to obtain optimum fre- quency utilization for the prevailing ionospheric condi- � When HF ionospheric conditions are tion. The configurations use common transmitters and favorable, and for links below the receivers, in conjunction with dynamic switching, to 3000-kilometer range, HF and UHF achieve multiple-frequency operation. The first-gener- performances are comparable. ation system employed three independent and separate * When HF ionospheric conditions are 2.0- unfavorable, and for multiple-hop -OriginalBuoyQuality links in excess of 3000-kilometers, r UHF performance is superior to that of HF perhaps by a factor as much as 2 to 1. BUOY REQUIREMENTS * UHF link reliability is not affected by 1.-B R - message lengths. Under good iono- spheric conditions, HF can handle both long, i.e., greater than 30 seconds, and 5 short messages with relative ease. How- ever, as the HF link becomes marginal, reliability of the longer messages de- creases considerably. AIR BAROMETRIC WIND WIND TEMPERATURE PRESSURE SPEED DIRECTION (�C) (mb) (mIs) (Deg x 10) Evaluation of Performance Figure 8. Evolution of Buoy Quality. Up to this point, we have addressed hardware and system evaluation in our quest for enhanced perfor- mance. Actual performance, as measured by system 4. Conclusion Mean Time Between Failures (MTBF), amount of data acquired and disseminated, and the associated quality of that data have all shown a respectable improvement. technology and changing requirements has been accom- plished by logical and effective evolution of our buoy electronics. It is our goal to further evolve our elec- tronic systems to provide the most productive and reli- and ending July 1975, an average of four moored buoys able vehicle for meeting the needs of our users and po- were on station at any one time. During that time peri- tential users. od, the composite buoy MTBF was 36 days, calendar 1976 was a transitional year. During calendar year 1977, an average of 15 buoys were on station and the composite buoy MTBF was 150 days. 1. Haas, G., T. L. Livingston, R. P., Mueller, R. J. Roten, "Remote ocean engineering environ- Quantity of Acquired and Disseminated Data mental data acquisition, " presented at the Ocean '77 Conference. Between 1972 and 1975, an average of 2200 synoptic weather messages per buoy per year were acquired and 2. Livingston, T. L., G. Haas, R. Mueller, "Data delivered to the user community. During 1977, an aver- Buoy communications experiences: HF vs. UHF," age of 2550 synoptic weather messages were similarly presented at the 1977 RTCM Assembly* acquired and disseminated by NDBO moored buoys. In addition to weather messages, a total of approximately 22,000 wave spectra and 8,000 bathymetric reports were acquired and disseminated by our buoys during 1977. Data Quality The accuracy of our disseminated data has improv- ed steadily since program inception. Our present buoy requirements are shown on Figure 8. Also shown is our initial buoy quality compared to our present quality performance. By evaluating our on-board buoy sys- tems, we have succeeded in achieving better than one order of magnitude improvement in air temperature and wind speed and better than a two-fold improvement in barometric pressure and wind direction. 9 NEW MOORING DESIGN FOR A TELEMETERING OFFSHORE OCEANOGRAPHIC BUOY Paul D. Higley and Arthur B. Joyal Raytheon Environmental and Oceanographic Services P.O. Box 360, Portsmouth, R.I. 02871 Telephone 401-847-8000 Abstract telemetered to shore. These buoy systems were to be A new buoy system has been developed for the . deployed year round in the heavily fished and severe A new buoy system has been developed for the U. S. Bureau of Land Management to gather, process, and environment of Georges Bank. Water depths for the buoy Bureau of Land Management to gather, process, and sites vary from 40 to 200 meters. Frequent storms, telemeter oceanographic data from the Georges Bank sites vary from 40 to 200 meters. Frequent storms, area of the New England outer continental shelf high waves, and strong tidal and storm drive currents, area of the New England outer continental shelf. characteristic of Georges Bank, necessitated a stable The buoy system incorporates a taut-moored sub- durable mooring mechanical system design with a reli- surface sphere 6 m below mean water connected to an able data link through the air-sea interface. The neces- anchor by an instrument string using an oil-filled synthe- sity to pierce this air-sea interface through violent tic rubber hose as the strength member. Three main tic rubber hose as the strength member. Three main storms, and the need for continuous on-board processing, sensor packages on the string measure current speed were two major factors governing the design of the buoy components, magnetic reference heading, conductivity, mooring system. and temperature. The bottom sensor package is con- nected to the anchor by an acoustic release. Sensor assemblies selected to be incorporated in The anchor consists of a 4500 kg free-flooding stor- these buoy systems include electromagnetic current age battery which supplies power to the system for up to meters, fluxgate magnetometer compasses, inductive four months. The air-sea interface is pierced with a conductivity probes, thermistor temperature probes, 12.2 m fiberglass mast attached to the subsurface sphere. and quartz crystal pressure gauges. A microprocessor Mounted on top of the mast is the antenna for trans- located within the buoy samples the sensor and controls mitting data via the GOES satellite telemetry system. the magnetic tape data loggers. The telemetry system Operational systems were deployed in the fall of 1977. utilizes the Geostationary Operational Environmental Satellite (GOES). (For description of the instrumentation on this buoy system refer to Real Time Oceanographic INTRODUCTION Data from Georges Bank by R.N. Lobecker and others in In 1976, the U.S. Bureau of Land Management ini- the proceedings of Oceans t78.)1 tiated a physical oceanographic study as part of the envi- DESIGN CRITERIA ronmental studies program for the New England outer Several platforms have previously been designed to continental shelf. The Georges Bank region of the outer Several platforms have previously been designed to gather oceanographic data from the continental shelf. shelf had been delineated as a potential site of petroleum ger eor t om the onita el However, efforts to combine these platforms with a tele- resources. Since a minimum of oceanographic data merry link to shore have resulted in cumbersome, existed for this area a field program was initiated to enin o relae ste n arom obtain necessary data. This data collection program expensive, or unreliable systems. A new approach to was divided into three parts; surface currents studies the buoy design was needed to provide a long-life data by aerial tracking of droques, regional hydrography link through the air-sea interface which could withstand by aerial tracking of droques, regional hydrography the rigors of the Georges Bank area. studies on a series of hydrography cruises, and sub- the rigors of the Georges Bank area. surface current studies using instrumented buoy sys- The environmental criteria to be met in the buoy temrns. mooring design are called out below:2' 3, 4, 5 Each of these buoy systems was to measure current 1. Water depths to vary from 40 to 200 meters. speed and direction, temperature, and conductivity at three levels. In addition, temperature every 10 meters 2. Surface current speeds to range from 1 meter second at the 40 meter stations to 0.5 meters through the water column and bottom pressure were to the 0 meter stations t t per second at the 200 meter stations, with bot- be measured. In-situ processing was to include calcu- r se d e et o th o tom speeds 75 percent of these values. lations of the salinity, the cross products of current with salinity and temperature, and 30-minute averages 3. The buoy must operate in 6 meter waves and of all data. Data was to be stored on-board as well as survive 12 meter waves. 10 4. Systems must operate in 20 m/s wind loading and 14. The system must be designed to be deployed and survive 40 m/s wind loading. retrieved during adverse weather conditions with- out damage. 5. System performance should not deteriorate during icing conditions. SYSTEM DESIGN These environmental criteria plus the need for a Central to the buoy system design is a taut-moored stable instrumentation platform, a reliable telemetry subsurface sphere, nominally 6 meters below the surface link, and long-term deployments, generate a number of and providing a net upthrust of 1100 kg. Mounted atop demanding requirements for the buoy mooring design. this sphere is a fiberglass mast piercing the sea surface and serving as a stable platform for the telemetry antenna. 1. The data line through the air-sea interface must The tension member between this subsurface sphere and be secure from the excessive flexing normally e cunred o te i ling n ay the anchor is a polyester fiber reinforced rubber hose. encountered in buoy systems utilizing slack This hose is oil-filled and contains electrical conductors moors or surface following buoys. to the instruments and the main battery in the anchor. 2. The telemetry antenna must be maintained above The anchor, which also serves as the power supply, con- the sea surface and must not be deflected by wind tains lead acid storage cells in an oil-filled, free flood- and currents such that its performance deterior- ing, steel housing. Figure 1 shows a pictorial view of ates. The antenna beam pattern selected for use the system. with the satellite telemetry system will accom- The surface piercing mast concept provides a semi- modate a deflection from the vertical of no more modate a deflection from th an 20more rigid member with a low frontal area and a low drag coefficient. This approach minimizes the transmission 3. Sufficient buoyancy must be provided to limit of wave-induced motions to the instrument platform and system depression caused by current drag. provides a limited flexing conduit for the data cable through the air-sea interface to the telemetry antenna. 4. The drag form of this buoyancy must be as small as possible to limit wave induced motions and the The use of a hose system concept satisfies many of drag form of all major components must be the basic mooring design objectives. The large outside minimized. diameter produces relatively low strumming frequencies even in high currents. Resonance excursions at these 5. System resonant frequencies should be out of the strumming frequencies are prevented by the damping wave spectrum and strumming of the mooring effect of lumped masses (sensor packages and thermistor line tension member must be minimized when housing) along the hose mooring line. The hose provides exposed to significant currents. a limited elastic elongation to minimize storm or deploy- 6. A compliant elastic mooring line to survive ment-induced surge loading while isolating the strength storm induced surges loadings during extreme member from the electrical conductors. Extra cabling wave conditions is a necessity. is coiled into the hoses to accommodate hose stretch. The hose is oil-filled and sealed, and all cabling and 7. Electrical cabling should be mechanically separ- connectors are potted to provide two-layer protection ated from the mooring line strength member to from sea water. Throughout the system design, redun- permit energy absorbtion to eliminate detri- dant seals were incorporated increase reliability. mental effects on the cabling along with isolating strumming motions from the cabling. The instrument string is attached to the anchor- battery via an acoustic release and a breakaway power 8. Electrical cabling must be protected from sea cable. This enables separate recovery of the anchor water intrusion by a double seal. and minimizes the possibility of damage to instrumenta- 9. Above water components must resist ice build up. tion during retrieval. A separate pickup buoy moored to a bottom tether line enables recovery of the anchor- 10. Corrosion protection must be sufficient to pre- battery, as well as providing a secondary recovery sys- I vent deterioration of the system over long-term tem should the acoustic release fail. In addition to this deployments. pickup buoy, three marker buoys are set in the area. 11. Antifouling protection must be sufficient for All four surface buoys are equipped with zenon strobe long-term deployments. lights and radar reflectors to increase visibility to local traffic. 12. A secondary recovery system must be provided trafc. to minimize the possibility of loss. MAST/ANTENNA ASSEMBLY 13. The system must be highly visible to the sub- The antenna/mast assembly provides for the follow- stantial fishing and shipping interests known to ing functions: traffic the area. * It serves as the antenna platform designed loads. With the stainless steel end fitting, the completed mast weighs in at 118 kg. * It carries and protects the RF transmission cable SPHERE ASSEMBLY * Most importantly, it penetrates a dynamic ocean surface. The sphere, located 6 meters below the surface to avoid ocean surface dynamic effects, required certain While satisfying these functions stated above, the design parameters to be considered: mast design required the incorporation of the following characteristics: 0 Highly reliable penetration seals * Smallest possible profile to minimize the drag * Smallest possible diameter to reduce drag effects effects of current from currents * Resistance to the long term effects of ultraviolet � Corrosion resistance light � Maximum possible net buoyancy * Resistance to extremes of temperature, air and ~~~~~~~~~water 0~� Sufficient internal volume to contain the system electronics � Watertight a Capability of being launched, retrieved and deck * Sufficient structural stength to withstand the bend- handled under poor sea conditions without being ing moment induced by storm conditions damaged * Flexibility to crack and flake off ice buildup. 0 Should the sphere break the surface under condi- tions of 12.2 m waves, the sphere structure and The end result was a mast design having a 14 cm waves, the sphere structure and attachment areas must have the capability to outside diameter at the sphere and a 9 cm outside diame- resist shocks and surges. ter at the antenna end and a uniform taper over the 12.2 meter mast length. (See Figure 2. ) This geometry The resulting design is a low alloy/low carbon steel represents an extreme slenderness ratio, but does in sphere weldment fabricated from spherical elements, fact satisfy the required design characteristics. skin thickness being 0.48 cm. A 46 cm diameter access hatch is provided to install and service the various elec- Materials employed for the mast construction con- tronic assemblies. The hatch sealing surfaces have sisted of fiberglass cloth and polyester resin wound been fabricated from 316 and 316L stainless steel. The around a urethane foam core. To minimize the tooling hatch flange is welded to the sphere skin with a doubler investments, the mas was fabricated in approximately plate (1 cm thick carbon steel) employed to uniformly equal segment lengths of 3 meters. This approach redistribute the resulting hydrostatic loads and to mini- resulted in the need for these segments to be joined. A mize the effects of the stress concentration factors good deal of attention was directed towards developing a resulting from the relatively heavy flange section and joint design which would give an overall greater struc- tural strength than the adjacent mast segments. This objective was achieved by specifying a long taper on adja- The net result is a sphere assembly displacing 1814 cent ends, and employing a epoxy/glass system which kg, having a net buoyant force of 1100 kg an air weight of produced a chemical bond and mechanical interlock of approximately 544 kg. relatively high strength. A 2 meter rigid standpipe is welded to the bottom of The antenna mounting flange has a wave deflection the sphere to lower the pivot point to just above the top taper to limit wave impact loading. A neoprene gasket sensor package. The location of this pivot point com- seal is used to provide a low pressure watertight seal. bined with the 1100 kg net upthrust provide the righting The lower mast end has a 316L stainless steel welded moment necessary to maintain the mast within the fitting, 61 cm long and on which the fiberglass has been required deflection from the vertical. wrapped. This fitting serves as a means to rapidly Corrosion protection for the sphere assembly con- attach the mast to the subsurface sphere and uses an"O" sists of standard techniques for undersea equipments. A ring to provide a watertight seal at this interface. Cen- 316 stainless steel was selected for the unwelded com- tral to the mast assembly is a fiberglass conduit through ponents in order to provide a reasonable resistance to which the RF cable passes. The RF connectors on eachn order to crevice corrosion. In order to avoid sensitizing the end of this cable are sealed to provide a secondary seal stainless steel components which were scheduled to see in the event the mast seals leak. Following receipt of int tthe mast fr om the fabricating vendor , each mast is welding heat, a 316L (L for low carbon) alloy was speci- thes fbycing lngdi a ads fied, thus avoiding the formation of carbide precipitates. structurally tested by cycling longitudinal axial loads The carbon steel sphere skin was protected with a multi- and the application of a number of bending cycles to the layer coating system consisting of a primer, zinc rich 12 layer, epoxy overcoat, and a final anti-fouling paint Mechanical physical properties of the hose assembly selected for galvanic compatibility with the alloys were specified as follows: employed in fabrication. � Working load--1588 kg Electronic assemblies mounted within the sphere * Ultimate tensile load-6804 kg include the following: * Elongation-0.5% stretch per 454 kg load * Microprocessor * Life expectancy-3 years minimum in operation. * Power Distribution Center In order to compensate for the hose elongation as the � Telemetry Transmitter/Receiver hose is loaded in tension, a method was developed to wind * Magnetic Tape Recorder the electrical cables in a helical coil and install them into the hose. Enough extra cable is installed to prevent cable * Three Battery Packs. breakage up to the hose elastic limit. A stuffing tube is All internal assemblies were designed and tested to installed at each hose junction to limit oil migration withstand shock and vibration induced by the rigors of between sections and strain relieve the cables. small ship environments, deck handling, launching and retrieval. Figure 3 shows the instrumentation in the nect the hoses to the sensor cages. Interruption of the sphere. Also, the assemblies were designed to pass hose at the mid-depth sensor serves to halve the hose through a relatively small hatch and to be rapidly length, thus simplifying handling. To be galvanically installed or removed for at-sea servicing and mainte- compatible with the stainless steel sensor cages, the nance. Humidity control is accomplished by purging the junction boxes have been fabricated from 316L and oil internal volume with a dry nitrogen gas and installing filled for the same reasons metioned for filling the hoses. dessicant. Figure 4 shows the sphere and mast ready Pressure-tight electrical bulkhead receptacles are pro- vided on the junction boxes to enable mating of the neo- SENSOR AND MOORING STRING ASSEMBLY prene molded cable assemblies which bring the cabling past the sensor assemblies. The internal side of the Each mooring string has three sensor assemblies; receptacles within the junction boxes are potted with a near the bottom, mid-depth and near the surface. Each dielectric urethane potting compound in keeping with the sensor assembly measuring temperature, salinity, cur- system design philosophy of providing redundant seals. rent speed and direction is mounted in a stainless steel cage. This cage passes mooring tensile loading across Dual 1.9 cm diameter pins are employed to couple the sensor while providing a minimum of current flow sensor cages to hose junction boxes, permitting quick disruption. The cage also provides sensor protection shipboard assembly. Cable connectors to junction boxes during deck handling, launching, and retrieval operations. are threaded to allow connections to be made rapidly. During launching and retrieval, this time is quite critical Electrical power cabling running from the anchor/ as the operations must be smooth and continuous to pre- power supply to the sphere and instrumentation cabling vent damage to equipment, particularly in heavy seas. from the sphere to each sensor assembly and the remote temperature sensors is contained within the rubber hose. This design provides excellent longevity and surviv- A light mineral oil in the hose dampens vibrations and ability in the dynamic ocean environment while maintain- serves to pressure equalize the hose assemblies, thus ing a minimum weight, maximum corrosion resistance, reducing the possibility and extent of a seawater leak. ease of fabrication, and good deck handling characteris- This oil also has good dielectric properties and further tics. serves to protect the electrical quality of the system. ANCHOR/POWER SUPPLY Each hose is 10 meters long and has terminations of either 316L stainless steel or 5084 aluminum. Aluminum electrical power to the system for an uninterrupted electrical power to the system for an uninterrupted has been used wherever possible to maximize the buoyant period of up to four months as well as serving as the period of up to four months as well as serving as the force available at the sphere. force available at the spher e. anchor. The anchor-battery consists of two sections of The hose has a 7.6 cm outside diameter and a 0.64 two volt cells. One section provides �26 volts and the cm wall. The hose is constructed with a Nitrile inner other 412 volts. The cell used is a photovoltaic station- core covered by six layers of unidirectional polyester ary cell designed for long-term slow discharge at cold yarns. Four of these yarns are parallel to the axis of temperatures. This lead-acid cell uses a lead calcium the hose and two are laid with opposing bias angles. alloy to limit battery gassing and provide a low self dis- These yarns are locked onto the metallic termination charge. Cell capacity limitation is acid starvation in nipple by a steel wire wrapping. The entire hose has a this application so a 1.300 specific gravity electrolyte is vulcanized neoprene cover. The final hose design has an used to increase available power. A glass fiber is actual tested ultimate tensile strength of 8,900 kg. wrapped on the plates to improve cell life at this high 13 specific gravity. Each cell provides 1100 A-hr. at a from 860 to 1100 kg., and buoy depressions ranging from 2000 hour discharge rate. The 38 series-wired cells are 0. 3 to 1.2 meters. A most critical result of this analysis completely filled with electrolyte, capped with a de-en- was the forecast of the antenna angular deflection. The trainment device, and vented to the top of the battery antenna beam pattern is omni-directional in the horizontal case via a plastic tube. These cells are mounted in the plane with a 40�-wide beam centered 400 above the hori- circular steel case pictured in Figure 5. In air weight is zontal in the vertical plane. less than 4400 kg. while in-water weight is 2585 kg. This accommodates a satellite elevation of 400 above As the assembly is lowered, seawater is allowed to the horizon. As a result, the antenna must be held within enter a built-in compartment to compensate for trapped 200 of the vertical to maintain telemetry communications. gas and temperature volumetric changes. Once on the The output of the static analysis, combined with the mast bottom, battery outgassing results in some of this sea flex angle analysis, yields a worst case antenna angle of water being forced out. Upon raising the unit from the 180 at the 40-meter station during exposure to 1 m/s bottom, the trapped gas expands and the remaining sea- currents and 40 m/s winds. water is expelled. If the cells have undergone severe Two types of dynamic models were employed to deep discharge and excessive gassing occurs, some oil T te on o e r moyi sy investigate the response of the instrument mooring sys- is forced out. The amount of lost oil is limited by inter- te th ri of ensr ent co in s tem to a matrix of environmental conditions? These nal ducting so that the trapped gas itself is vented. This models were exercised with various station depths, cur- models were exercised with various station depths, cur- insures that the battery posts and wiring continue to be rent profiles and wave spectra. Basic mooring design rent profiles and wave spectra. Basic mooring design protected in oil. For servicing, means were provided to parameters were varied to optimize buoy mooring parameters were varied to optimize buoy mooring pump out any remaining entrapped sea water and replace design. The first model simulates the torsional charac- design. The first model simulates the torsional charac- it with oil. ~~~~~~~~it with oil. ~teristics of the mooring system to investigate possible The design allows for continuous operation at a tilt errors induced in current vector calculations due to buoy angle of 45�, should the container come to rest on an rotation. The second model investigates heave, surge, uneven bottom and short-term angular tilts in excess of and pitch. Both models utilize a finite-element, lumped 900 to allow for handling and retrieval dynamics. Evi- parameter approach to yield a set of coupled non-linear dence has been observed that during retrieval sufficient second order ordinary differential equations. Numerical drag is produced to, in fact, tilt the container 90�. The integration is employed to solve these equation sets for container structure is fabricated from a low alloy, low motion response. Forces considered include: hose ten- carbon steel protected against seawater and acid corros- sion, hydrodynamic drag, inertia, virtual inertia, buoy- ion by a multi-coat epoxy system. ancy, and weight with dynamic excitation from current and wave effects. Wind drag was inserted as a static ACOUSTIC RELEASE-FLOTATION ASSEMBLY ACOUSTIC RELEASE-FLOTATION ASSEMBLY input. Current was assumed as a constant for each case The instrumented sensor string and hose assemblies and wave conditions inputted as either a regular or an are connected to the anchor/power supply via an acoustic irregular sea. Sea states were inputs as an empirical release system. During retrieval, activation of the wave energy spectrum based on the Pierson-Moskowitz acoustic release results in separation of the sensor hose model.10 All wave motion amplitudes were assumed to assembly from the anchor-power supply. A breakaway decay exponentially with depth. power cable connector is used to allow electrical separa- Torsional modeling predicted small mooring dis- tion. Flotation chambers affixed to the acoustic release placements with resonant periods varying from 50 to 157 placements with resonant periods varying from 50 to 157 bring the bottom end to the surface for simplified retrie- seconds. Heave/surge-pitch modeling predicted no val of the instrument string. Additional flotation cham- problem motions with the exception of the 40-meter depth bers are attached along the string, where necessary, to station in the regular period survival sea conditions. In compensate for the additional weight, and aid in handling this case, hose tensions reached 4500 kg (50% of ultimate this case, hose tensions reached 4500 kg (50%/ of ultimate during deployment and retrieval operations. aiu during deployment and retrieval operations. strength) and could approach a slack condition. Various STATIC AND DYNAMIC ANALYSES inputs were changed and the model rerun. Optimization of this design involved a reduction of the net upthrust of A series of static analyses were performed to opti- the subsurface sphere to soften the buoy response. mize the instrumented buoy mooring design.6, 7, 8 The tesbufc peet sfe"tebo epne mize the instrumented buoy mooring design.6 , The The modification reduced hose tensions to 2300 kg. and critical parameters of interest included instrument pack- the slack conditions was eliminated. age tilt, hose tension, system depression, horizontal displacement, and antenna angle. Current speeds up to SYSTEM PERFORMANCE REVIEW 1.25 m/s and winds up to 40 m/s were inputs to the Design of the buoy system was initiated in September analyses, as well as hose elasticity, drag coefficients, andsystes welshts ela dr imentsi 1976, and a prototype mooring was deployed on Nantucket and system weighs and dimensions. Shoals in March 1977. A full-instrumented engineering The results of these analyses predicted sensor pack- model successfully operated off Block Island in Septem- age tilts ranging from 30 to 70, hose tensions ranging ber 1977. One buoy system was installed in 84 meters 14 of water at Georges Bank in November 1977 and retrieved 8. Weigel, R. L., "Oceanographic Engineering". after a one month deployment. A second buoy system Prentice-Hall, Inc., 1964. was deployed in 84 meters of water at a different site in 9. General Electric, Re-entry and Environmental Sys- December 1977 and retrieved after four months opera- tion in April 1978. During this four month deployment, several major storms passed over the buoy site, gener- 10. Pierson, W.J., Jr., and Moskowitz, L., "A Pro- ating significant wave heights in excess of 7.5 meters posed Spectral Form for Fully Developed Wind Seas and wind speeds over 35 m/s. Following this deployment, Based on the Similarity theory of S.A. Kitaigorodskii", the system was extensively inspected for wear and cor- Journal of Geophysical Research, Vol. 69, No. 24, rosion. Critical wear points, such as the pivot below 1964, pp. 5181-5190. the sphere, experienced insignificant wear. There was no system deterioration evident from this 4-month deployment. At the time of this writing (June, 1978), three buoy systems are installed and operational on Georges Bank. Two of these buoys are at 84 meter depth stations and one buoy is installed at a 200 meter depth station. ACKNOWLEDGEMENTS The buoy systems were developed by the Raytheon Company, Oceanographic & Environmental Services, under contract AA550-CT6-53 with the U. S. Bureau of Land Management. Many individuals, not listed as authors, contributed to the development effort, among them: Robert Lobecker, Ron Franklin, Les Stroup, David Cook, Earl Lind, Gary Milano, Richard Neidich, Robert Day, and Captain Robert Hanson. Dynamic modeling was contributed to by the General Electric Re-entry and Environmental System Division. Technical monitoring of the contract for the Bureau was accom- plished by Kenneth Berger and Eiji Imamura. REFERENCES 1. Lobecker, R.N., Milano, G., Franklin, R. and Lind, E., "Real Time Oceanographic Data From Georges Bank", Proceedings Oceans 1978. 2. Beardsley, R.C. and Butman, B., "Circulation on the New England Continental Shelf; Response to Strong Winter Storm", Department of Meteorology, M. I. T. and Woods Hole Oceanographic Institute, 1974. 3. Bumpus, Dean F., "Review of the Physical Ocean- ography of Georges Bank", Woods Hole Oceano- graphic Institute, 1975. 4. Neu, H.A., "Extreme Wave Height Distribution along the Canadian Atlantic Coast", 1972. Ocean Industry. 5. Neu, H.A., "Wave Climate of the Canadian Atlantic Coast and Continental Shelf', 1970, A.O. L. Report 1971-10 December 1971, Bedford Institute. 6. McCormick, M.E. "Ocean Engineering Wave Mechanics", John Wiley and Sons, 1973. 7. Berteaux, H., Buoy Engineering, Wiley Inter- science, 1976. 15 5.5m 31 1 TEtEMETRY I,,,|~~~~~ H~~RECEIVER & 6.5m TRANSMITTER TAPE RECORDER ELECTROMAGNETIC CURRENT SENSOR RUIY BATTERIES SUSSURFACE MOORIN RUY DUOATA PROCESSOR ADD CONTROLLER LOCAL POWER TEMPERATURE DISTRIBUTION TEMPERATURE 'UNIT SENSENSOR SENSOR UI N I T SO lam \ REMOTE _ MICOEPTH / SENSORS \/UNIT HOSE/CABLE ASSEMBLY REACOUSTC FLOATATION BOTTOM UNIT BOTTOM . / PRESSURE SENSOR POWER ANCHORI TOPICKUP BUOY CABLE MAIN BATTERY BATTERY NOTTOSCALE Figure 1. System Configuration 16 Figure 2. MAST Figure 5. Anchor/Battery Figure 3. Sphere Electronics Figure 4. Sphere/MAST Deployment 17 A SUCCESSFUL METEOROLOGICAL DRIFTING BUOY F.A. Reid Senior Engineer Hermes Electronics Limited P.O. Box 1005 Dartmouth, Nova Scotia B2Y 4A1 Abstract The development of a successful meteor- ological drifting buoy has been undertaken. Buoys which measure sea surface temperature and barometric pressure have been built, tested and successfully deployed. A descrip- tion of this buoy is given, with details ofW the sub-systems. 1. Introduction 1.6 ~~~~~~~~~2w The development of a successful meteor- ological drifting buoy was started in 1976 as part of the Canadian Ocean Data System project, a contract held by Hermes Electronics Limited. The buoys were to act as free drifting plat- forms for the measurement of barometric pressure and sea-surface temperature. The M O DIFIED SPAR pressure was to be measured to an accuracy of SA � 1.0 millibar (mb) (including all sources of error) over the range 900 to 1050 mib. The temperature was to he measured with an accuracy of � 0.500 over the range -50C to +40Cc. The buoys were to use the Nimbus-RAMS system for Figure #1. Prototype Hulls data telemetry and position fixing. The buoy was to last 12 months in 1000 water and 6 A truncated cone hull shape has characteris- months at 0'C. tics which the above tests proved desirable. The hull material was changed to aluminum for ease of There are a number of problem areas which construction, strength and cost considerations. had to be thoroughly investigated before a buoy This hull has been tested extensively and has which met the above specifications could be become the basis of several drifting buoy config- produced. The process has taken two years and urations. has resulted in the production of approximately 100 buoys at Hermes Electronics for use in the As the hull is now made of aluminum, a radio First CARP Global Experiment (FGGE). transparent cover must be provided for the antenna. This cover must be extremely robust as 2. Hull it must withstand high impact loading during handling and launch. The antenna cover is con- The earliest prototype hulls were con- structed of RFP and seals to the aluminum hull via structed of a fibre reinforced plastic. The a flat seal above the water line. The antenna two that showed most promise were a 2.Om spar cover also forms an integral part of the baro- and a 1.6m modified spar, as shown in metric pressure inlet port which is described Figure #1. Several tests including full size below. tank tests, launch and mooring tests were con- conducted with these buoys. The results indi- 3. Electrical System cated that, while both types would suit the requirement, the spar was only marginally The electrical system of the drifting buoy stable for good communication and the modified has four parts, namely: Battery and switch; spar would be too expensive to produce. Sensors; Data formatter and transmitter; Antenna. 18 ________________ square wave with a period Proportional to pressure. ~~~~PRESSURE PORT This output is averaged and linearized over the PRESSURE PORT PRESSURE PORT pressure range to give a count which is linear with INLET TUBE W~~~~ATER TRAP pressure. FRP ATENNACOVERThe data formatter and transmitter is manu- ~~- -~ ANTENNA factured by Handar. The unit accepts the input FLAT SEAL- ~~~~~~~data, formats and transmits the data in a form *~~~~~PESR compatible with the Nimbus-RAMS system. The units TRANSDUCER ~~~~~~~~have proven to be very reliable, with over 18 y ~~~~~~~months of continuous operation in the field. * ~~~~~~~WATERLINE JTRANSMITTER 4 . Pressure System * ER~~~~CNRENSATEON TRPTEMPERATURE IThe heart of the meteorological drifting buoy SENSOR is the accurate measurement of barometric pressure. PRESSURE PORT ' 'To make measurements to the accuracy required over BRAIN TUBE ~~~~~ALUMINUN 1 year on an inexpensive drifting buoy meant that several novel solutions to the problems of dynamic A ~~FOAM FILLER pressure error, water entrapment, and water vapor inflow had to be found. ANTI-FOULING BATTERY ~~~~The movement of air across the pressure port PA]INT inlet will give rise to a dynamic pressure which adds to the static pressure and causes an error. The magnitude of this error depends on the velocity of the air and the geometry and attitude of the ITOTAL WEIGHT - Okg pressure inlet. The pressure inlet on these drift- ing buoys is on top of the conical antenna cover. Due to this conical section there is a low pressure -~~~~~ ~~area at the apex. The present inlet compensates for this low pressure by causing a dynamic pressure increase. The present port assembly has been Figure #2. Drifting Buoy tested extensively in the wind tunnel and has given the following results: less than I mb dynamic The battery chosen for the drifting buoy was error at all speeds and attitudes up to 30 in/sec1 an alkaline dry-cell system with a nominal voltage less than 0.3 mb error at speeds of less than 20 of 13V and a capacity of 160 Ah. This system was in/sec. With the design presently in use no attempt chosen because of its energy-density, low temper- is made to keep water from flowing through the p ~~~~ature performance and its wide availability. The inlet port, as significant pressure gradients could battery capacity is sufficient to operate the develop across any membrane structure. Instead a buoys for 1 year at lOWC and 6 months at OOW. drain tube is provided which runs directly from the Field tests have shown that the present design inlet to an exit below the water line. The pres- meets these requirements. sure for the barometer is tapped off this drain tube, spirals upwards, into a desiccant container The switch used for the power supply is and then to the barometer (Figure #3). The purpose magnetically operated, and does not require the of the spiral is to provide a suitable volume of integrity of the hull to be broken for actuation. trapped air to allow the buoy to submerge to a The switch is normally closed and is held open depth of 10 metres, without any water reaching the when the magnet is placed in close proximity to top of the spiral. it. This allows the buoy to be launched without the necessity of opening the hull to switch on; The desiccant container is used as a dryer to and more importantly, if the magnet is retained by remove any water vapour present in the system. the packing crate, the buoy cannot be launched in the off state. S. Testing The temperature sensor used on the drifting The first 30 buoys produced utilizing this * ~~~~buoy is bonded to the inside of the aluminum hull design have been the object of extensive testing just below the water line. This allows the during the last 18 months. measurement of the sea-surface temperature without * ~~~~~the problems of sealing the sensor against the sea Twenty of the buoys were deployed in the for one year. The sensor itself, is a linearized Atlantic, Pacific, and Indian oceans. Figure #4 thermistor configured in a bridge to give a linear gives a map of their deployments and drifts. Of voltage versus temperature characteristic. the twenty buoys only four have had premature electrical or mechanical failure and two have The pressure sensor is a Paroscientific Digi- lasted in excess of 18 months. It is very diff- quartz pressure transducer model 230A. This unit icult to measure the data quality from the field was chosen for its excellent accuracy and long test buoys but in those instances where a suitable term stability. The output of the transducer is a reference existed the data quality was to spocifi- 19 cation. Further studies were conducted in controlled situations to study the affects of fouling on the L' buoys and pressure system. The results indicated that there should be no significant fouling of the buoys anywhere in the world's open oceans. As a design goal the buoys were to be cap- able of being launched from any type of ship, by an untrained crew without means of a crane or with the ship stopping. To verify this, buoys were dropped from a height of 20m into the water, both horizontally and vertically without damage. The packing crate was designed such that the magnet is retained with the crate when the buoy is unpacked. I'Lcz A rigourous suite of environmental tests were undertaken to bring to light possible failure modes. No major faults were discovered but a numer of minor modifications were incor- porated as a result of these tests. 6. Conclusions A drifting buoy which meets the design goals set out for has been built and extensively tested. The present design has been used as the basis for the Canadian drifting buoy contribution to FGGE. In addition, in excess of 50 drifting buoys of slightly different configuration but of the same general design have been supplied to various users in the last 12 months. Figure #3. Pressure System i~~~~~~~~~~~~ BUOYS �~~~~~~~~~~~~~~~~~MY17 4~~~~~~~~ BUOYS ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~MAY 1977 Figure~~~~~~~A 1977 3rftn BUOY DpomnS ~~~~~~20BOSMY17 Acknowledgements The development of these drifting buoys was undertaken as part of the Canadian Ocean Data System contract * 0SZ5-0106 under the supervision of John Brooke. The author would especially like to acknowledge the guidance given by Dr. Jim Elliot, Bedford Institute of Oceanography and Dr. John Garrett, Institute of Ocean Sciences, throughout the entire project. 21 AN ANALOG WAVE SPECTRUM ANALYZER SYSTEM FOR OPERATIONAL BUOYS K. E. Steele Dr. A. S. Hananel NOAA Data Buoy Office Sperry Support Services National Space Technology Laboratories National Space Technology Laboratories NSTL Station, Mississippi 39529 NSTL Station, Mississippi 39529 Abstract The other two payloads had limited data acquisition capabilities. One was the Magnavox In 1975, the NOAA Data Buoy Office (NDBO) Limited Capability Buoy (LCB) type, which had a began deploying a data acquisition system iden- digital interface for sensor subsystems.6 The tified as the Prototype Environmental Buoy (PEB) other type was one that had already been con- payload on their operational buoys. This pay- ceived by General Dynamics for the Prototype load could report environmental data from remote Environmental Buoy (FEB).7 The PEB-type payload ocean areas to shore through radio telemetry, had analog interfaces and 12 channels available but could only receive data inputs in the form in which to connect the output(s) of a wave of analog voltages. Digital data could not be measurement system. accommodated. NDBO subsequently developed a wave measurement system specially designed to NDBO sought to develop a wave measurement interface with the PEB payload. This Wave Spec- system that would take maximum advantage of the trum Analyzer (WSA) system consisted of 12 par- 12 channels of the PEB payload, yet at the same allel analog filters, each of which produced an time could be used with the Magnavox LCB type. output voltage related to the spectral density Such a system was developed, and the basic sys- in the vicinity of the center frequency of the tem hardware has been described in detail in filter. Duplicates of this WSA system have another paper. The system consists of 12 par- since been deployed on a number of operational allel analog bandpass filters that are fed by deep-ocean buoys and have produced wave data the voltage output of a strapped-down accelero- routinely. meter as shown in Figure 1. 1. Introduction 1 Early in 1973, NDBO initiated an effort to develop an accurate, reliable wave measurement Data ~Channel 2 system capable of measuring one-dimensional wave ~ IP spectra for use on operational buoys. The first step in this effort was the December 1973 de- Chan e13 ployment of an experimental system in the Atlan- tic to measure significant wave height and period.1 Much was learned from this effort but the data produced was not of good quality.n Channel5 December 1974, an experimental wave measurement system was deployed which was designed to pro- Channelt duce wave spectra data every 3 hours.3 This e- Other system demonstrated the feasibility of the Oathe routine acquisition and teletype relay of D accurate near-real-time spectral wave data, but unfortunately proved to be unreliable due to a number of design problems, and was thus unsatis- than factory as an operational system. NDBO had planned to use three types of buoy data acquisition and reporting systems, or pay- loads. Each system ne eded a reliable spectral wave measurement system. One of these types, the Magnavox Phase II payload4, was a techni- tChan1 cally advanced system with digital interfaces . for the relay of data from sensor subsystems. A wave data analyzer for this system has been developed and is described elsewhere.5 Figure 1. WSA SYSTEM SCHEMATIC 22 Each filter produces an output voltage de- Given that eq. (1) is the proper weighting pendent on the voltage spectral density of the function for the effect of the kth filtering input in the vicinity of tge bandpass of the circuit on the input voltage spectrum, it will filter. It has been shown that the weighting be shown how the outputs of K separate and inde- factor for the kt4 of K (normally 12) filters pendent analog filtering circuits can be used as is, in volts/volt , a system to estimate the original sea heave dis- placement spectrum. In Section 2, a mathemat- gk(f) = Ik(f)l' = ical expression for the count received at shore from the kth filter is developed. Section 3 =--g~(f -Af)2 shows how the count from a single filter can _ -g -2k~ (* k (1) sometimes be used to estimate the spectral den- 7 (fd 2 -2 2 2+ sity at the center frequency of that filter. A 4Z (f2- k' Afk) method for using the set of K counts from the K filters to determine K estimates of the spectral densities at the K center frequency is presented ~~~~~~~~~where ~in Section 4. Computations based on the tech- i = ~- nique of Section 4 are shown in Section 5. Sec- tion 6 presents conclusions about the perfor- g (volts/volt) = gain of the output buffer mance of these systems based on operational use. of the filtering circuit Section 7 acknowledges those who helped in this (the same for each filter- effort. ing circuit, magnitude = 2.0) (volts/volt2) = gain of the squaring cir- 2. Mathematical Analysis of the WSA cuit of the filtering cir- cuit (the same for each The objective of the WSA system is to esti- filtering circuit, magni- mate the sea displacement spectrum, S(l)(f) tude = 0.1) [m2/Hz], which is usually zero or negligibly small below some lower frequency limit, Fi(Hz); gk(volts/volt) = gain of the kth filter, i.e., I which is usually different S( (f) 0, f < F(2) for each filter and is determined by hardware adjusetmermintsed Estimates of S( )(f) at particular frequencies ~ad~j~ustments are to be made through the use of a number, K, of analog bandpass filters. f (Hz) = frequency at which the of analog bandpass filters. weighting factor is to be computed The original sea spectrum of displacement, S _f(f), has an associated sea velocity spectrum, (Hz) -ter frequency of the kth (f) [(m/sec)2/Hz], and a sea acceleration fk (Hz)ilter, which is usually spectrim, S(3)(f)[(m/sec2)2/Hz]. In compact ifileret w h eh ualte form, the relationship between these three spec- different for each filter and is determined by hard- tra ran be written as ware adjustments r w aStes (r)(f) = (2rf)2(r-i) S(1)(f) (3) Gk(jf) = modified linear system transfer function, related where r and i take the values 1 (displacement), to the primary system trans- 2 (velocity), and 3 (acceleration). fer function as follows: However, the acceleration spectrum seen by the accelerometer on board the buoy, S(3)(f), Gk = ~ 'g k(J differs from the true sea acceleration spectrum Afk (Hz) = noise equivalent bandwidth S(3)(f) by the factor of the frequency-dependent of the kth filter, which is hull transfer function, w(f). Thus, usually different for each filter and is determined by (3)(f) = w(f) . S(3)(f) (4) hardware adjustments There may be a slight difference between the true vertical acceleration of the discus buoy hull and the buoy acceleration vector projected I- I G(jf)1 df onto the axis of the accelerometer, which has its ~2-g~~~~2 J axis fixed parallel to the mast, but NDBO's exper- _-~~~~~ ~ience with other spectral wave measurement sys- (f . Af )2 tems has shown that S(3)(f) in eq. (4) is prac- 1=/2 | -2_2+ df. tically the same as the spectral input to the (f2_fk)2+ (f . Afk)2 linear single-axis accelerometer.9 23 The acceleration spectrum is converted by t = time at which the sth sample is the accelerometer to a voltage spectrum; i.e., s taken S(3)(f)[volt2/Hz] = K2[volt2/(m/sec2)2] s = 1, 2, , S = the number of -(3) 2)2/Hz] samples taken during single data S (f)[(m/sec2)2Hz] (5) acquisition period where K is the slope of the accelerometer linear S = total number of samples taken calibration curve. from each filter during a data acquisition period. If the input voltage to the kth filter, which has a spectral content of 5(3)( as Different sets of values for c, B, R and S given by eq. (5), is identically zero, then the correspond to each payload type. output of the filtering circuit should ideally be zero. Also, when the sea state is stationary with a non-zero S( (f), the output voltage of TABLE I. WSA INTERFACE CHARACTERISTICS the filtering circuit should ideally have a FOR PEB AND LCB PAYLOADS steady, constant value. In the actual hardware, however, these ideal performances are not real- PAYLOAD ized. There is usually an offset voltage, Ek PEB LCB k, AID CONVERTER ( GENERAL at the output even when the input is grounded to CHARACTERISTICS SYMBOL UNITS DYNAMICS) (MAGNAVOXI zero. This offset voltage may be positive, zero, or negative. Thus, the actual performance of Min-Voltae -r/ Va - the filtering circuit is given by Count Out Ror ont 4096 t'=(3) +Mf)'-/Max Voltage I Volts +10 +10 Ek(t) = Ek(t) + gk( f) S (f) df (6)orO nt 0O AID Count for Zero where gk(f) is given by eq. (1), S(3)(f) is Voltage In lavgl c Count 2047.5 0 given by eq. (5), and Ek(t) is the time-depend- AID Sensitivity Count ent offset voltage of the kth filter. The time- Inominalvaluel B Volt -204.8 409.6 dependent voltage Ek(t) is input to the analog- Total Number of to digital (A/D) converter. (See Figure 1.) Samples Taken S Integer 4096 1 Interval Between When the input to the A/D converter is Samples Sec 0.125 - grounded, the output count of the converter will be c. This count depends on the type of payload Total Duration of to which the WSA is connected, as described in Sampling Sec 512 Table I. The A/D transformation for any WSA system is represented in the following equation: A number S of samples are taken from the Cks [count] = c [count] + B [count/volt] A/D converter and averaged to produce the count sent to shore by the payload; i.e., Ek(ts) [volts] (7) S * Ek(ts) [volts] (7) Ck (count) = 1 (8) S Z Cks except that s=l Cks = 0 if c + B . Ek(ts) < O Substituting for Cks from eq. (7) into eq. (8), Cks = R if c + B Ek(ts) > R the count sent to shore can be written as where 1 Ck = c + B Ek(ts ) . (9) Ck = count at the output of the A/D k s converter from filter k at the time sample number s is taken Using eq. (6), the average voltage into the A/D c = count when zero voltage is input converter is given by S S B = sensitivity of the A/D converter 1( (t S L BEk(t) S Ek(ts) Ek(ts) = input voltage to A/D converter s=l s=l from filter k at time t R = upper limit count of the A/D + gk(f) (3)(f) . df converter o k 24 or or Xck = B -g ~g j2 -2 ek =ek + gk(f) S)(f) df (10) Fk+l (15) I (f.Af)2 w( f) (2f)2(3i) S(i)(f) df s T_2 (f2j-f2)2 + (f.Af)2 1lC F 4 where ek = Ek(t) is the average actual s=l where k = 1, 2, ...,K and i = 1, 2, 3 voltage and The limits of integration have been narrowed ~s~~~~ ~~due to the negligibility of the product ek E Ek(ts) (11) w(f) � (2rf)2(3-i) . s(i)(f), when f < 1 and s=l f > FK+1. FK+1 is that frequency above which is the average offset voltage coming out of the there is no significant contribution to the kth filter into the A/D converter. integral, due to either the hull filtering or the absence of wave energy. The significance Inserting eq. (10) into eq. (9) yields of the "K+1" subscript will be demonstrated in Section 6. Ck - (c + Bek) = B J gk(f) . S( )(f) df . (12) 3. Narrow Band Approximations of Spectral Density If (c + Be ) > 0, then the definition kf (c + Bek) 0, then the definition An estimate of spectral density at the cen- ter frequency of the kth filter can be made from -C ~ -=~ -c+Be~~ ~the output count of the kth filter alone, if the Ck = c + Bek (13) kth filter is very narrow relative to the spec- tral variation about its central frequency fk' yields a value Ck, which is the count received If the assumption is made that the expres- at shore when the accelerometer is not being (3-i) (i) subjected to any acceleration; i.e., when sion w(f) (2ff) S(f) in the inte- ~~~~~~i=(3) _ aue grand of eq. (15) is slowly varying within the S (f) E 0 for all f. Thus, Ck is measured range of non-negligible values of the filtering after the buoy is integrated and is at rest at function gk(f), then eq. (15) can be approxi- dockside reporting counts. mated as If (c + Bek) < 0 , as can occur in the case k = Ck k B. .K k (fk) of the LCB payloads, then Ck cannot be measured (2W )2(3-i) (i) (16) directly and must be computedfrom eq. (13) with k k 1k measured values of c, B, and ek. Also, it must in which the definition of the noise equivalent in which the definition of the noise equivalent be computed when the system is installed under bandwidth given in Section 1 has been invoked. conditions that preclude having the buoy at rest. Solving eq. (16) for the spectral density, Ck and Ck can now be combined into one pa- S(i)Y (Ckk) .(17) rameter, as follows: (3-i) B.W.F.K2.91 .w(fk). (2-rfk)2 'Af Bk 2 k k Ck Ck k = B gk(f) . S(3)(f) df . (14) This linear relationship between an estimated �0~~~~~ ~~value for S(i)(fk) and the count Ck sent from the buoy to shore sometimes allows a reasonably This ck value can be computed at shore either accurate determination of spectral density at from the count Ck sent from the operational buoy frequency fk from the kth filter alone, without from the count Ck dentefrominhed operational buyeq and the count Ck determined in advance by eq. using all the K filters as a system. The valid- (13), or by taking direct readings. It will not ity of this determination depends on the narrow- be necessary to compute Ck, as will be shown. ness of the filters; but making the filters extremely narrow would introduce instabilities into the filters, and would require an imprac- tically large number of filters if no gaps were Substituting in eq. (14) for g (f) from to be left in the spectrum. Therefore, it is =Sz(3)~~~ k ~ not desirable to make the filters extremely eq. (1), and for S'3'(f) from eqs. (3), (4) and narrow banded, and eq. (17) is only to be used (5), we obtain: occasionally for quick estimates of spectra. 25 4. Computation of the Spectral Density Matrix TABLE II. INTERVAL DESCRIPTION Interval Interval Boundaries Interval Center When the bandwidths of the K filters are Number F. IHz) F(Hz) Width Frequency not extremely narrow, a considerable contribu- j J. (Hz) f(Hz) tion to the output of a filter may come from the skirts of the filter due to spectral densities 1 0.045 0 055 0.01 0.05 in other parts of the frequency range away from 2 0.055 0.065 0.01 0.06 the bandwidth of the filter. Figure 2 shows the 3 0.065 0.075 0.01 0.07 12 filters of the WSA system, spaced in such a 4 0.075 0.085 0.01 0.08 way that 12 contiguous bandwidths cover the fre- 5 0.085 0 095 0.01 0 09 quency range from 0.045 to 0.360 Hz. This inter- action between filter ranges is derived herein 6 0. 095 0.105 0.01 0.10 and put into computational form. 7 0.105 0.14 0.035 0.1225 8 0. 14 0.18 0.04 0.16 1.25 9 0. 18 0.22 0.04 0.20 10 0.22 0.26 0.04 0.24 liOG 1 fl '\ IAl 1A A,11 0.26 0.30 0.04 0.28 0/75 z 0 1 Il I~ 12 0.30 0.36 0.06 0. 33 With the intervals defined in this manner, Eo 0 11 1 % I \ eq. (15) can be rewritten, with no approxima- I~IIYI' I!11 ~ y i/v \ tions, as 0.25 ck I j+i 0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 j=l . FREQUENCY (Hz) Figure 2. NORMALIZED FILTER RESPONSE CURVES (f'Afk)2 w(f) . (2f)2(3-i) FOR A 12-CHANNEL WSA f2S )(f) � df (20) T2(f2- )2 + (f.6f)2 Let the total interval F! < f < Fk+l (outside of which the integrand of eq. (15) is If the interval widths, df., are made reasonably assumed to be negligible) be divided into K small, the function S((f) can be considered, intervals, such that the left-hand boundary of to some degree of approximation, to be a con- each interval is F. and the right-hand boundary stant within each of the K intervals. Further- is Fj+i, with j = 1, 2, 3, ..., K. The jth more, the function w(f) can be assumed to behave 3J+1 linearly within each interval, without noticeable interval has a width of loss in accuracy; i.e., f=F -J+ (18) [w(f)]j = + W. + Wf (21) and a center frequency of The constants W. and W. are defined as follows: fj = (1/2)(Fj+. + Fj) (19) ~ ~ j~+1 J. j - w -+l j = J W = w. - F.W. (22) The intervals F. were chosen so that the f J J 3 3 widths 6f. are smallest when the integrand is significantly large and varying fast, and larg- where wj and wj+l are the values of the hull est when the integrand is either small or vary- transfer function at the lower and upper bound ing slowly. Furthermore, these intervals were frequencies of the jth interval. chosen so that f. a f. for all j; i.e., the j i . .. .Thus, using eq. (21) and denoting by center frequency of the interval is identical to the center frequency of the corresponding filter. S () - S(i)(f), the constant value of the The total number of intervals must be the same s(i) j as the total number of filters, for reasons that within interval j, eq. (20) will be discussed subsequently. Nevertheless, is transformed into the intervals are distinct from the filters and K must be identified by a different subscript. (i) (3) ck = kX. . S (23) Table II lists the intervals normally used j=1 on all NDBO buoys. The number of filters and intervals is K = 12. 26 where In principle, there is no need for ASi) to be A . = - 2 g jladded, because corrections for all known offsets kj B 9Kgk |have already been made. However, we reasoned FjJ that, after all known corrections had been made, some residual small random error would still exist in the spectra. If this error were nega- ~(f-Af )2 -(W.+W_.f) /f)2(3-i) tive, the system would fail to report very small (fAfk)2 (Wj+f) (2f )2i spectral energies which, because they are small, 2 df (24) would not be of much interest to users. But, if .3(f2 -f2)2 + (f.Af )2 the error were positive, the system would report (2)2+(4 k ksmall energy spectra that did not exist. Equation (23) yields K different counts, ASMi) was thus included to desensitize the each of which is generally a function of the K system. The negative values used to date have spectral densities. However, each spectral den- be en selected by best engineering judgment, but been selected by best engineering judgment, but sity as a function of the K counts is needed. they should ideally be selected by calibration This is obtained by solving eq. (23) in its against a satistically valid set of ground-truth matrix form: spectra. With eq. (31), the spectral densities in = ( i .]*S . (25) eq. (30) can be redefined to yield Solving eq. (25) for the spectral densities yields j = jk k (32) *-(i)TA -1 k=1 Si) = Aki c .(26) Using the spectral density values of eq. (32), kJ~ ~the spectral moments M2i) are computed as _Mi) (i follows: Denoting the inverted matrix of kj by xk i.e.,lows: K -1 [Aki HSM .i) 6f (33) r(~)1 (27) 2(i-1) 2 j (33 kj ] [jk , j =1 the solution can be written as The spectral moments M0 and M2 are computed in til _ rx~i~l . | t . t28) terms of the displacement spectral density by jk ck (28) using eq. (3): It is now obvious why the number of inter- ( St.) 6f. (34) vals must be the same as the number of filters; j=l -(i) eq. (25) can be solved only if the matrix Akj is invertible (i.e., only if it is a square (2) matrix in which k and j both take K values). M S( * 6ff j=l Using the summation notation, eq. (28) can be written in the fashion of eq. (23): K si) - ) Ck . (29) = > (2rf)2 * 6f. (35) XM . J 29 J jk k2f)'Si f j=1 k=1 Recalling from eq. (14) that ck = C - Ck, we These moments are needed for the calculation of k ~~k k9 the following: can further expand eq. (28): K K ( 4 M36) H1/3 =4~- si) = E jk .(30)) Ajk jk k=l k=lT -f37 T2 = 2M (37) The second term in eq. (30) is the correction to the spectral density calculation based on the total count received at shore. A direct empir- where ical correction, ASi), was added to this cor- H = significant wave height = average rection so that H1/3 rection so that 1/3 height of the highest 1/3 of all waves K ;(i) AS Ei) (i (31) T2 = average zero-up-crossing wave period. j j - jk~k k=l 27 5. Computations The last printed line in Figure 3 gives the spectral moments, the significant wave height The computations leading to the determin- H1/3' and the average zero-up-crossing wave ation of the wave spectra and the spectral mo- period T The displacement spectrum is plotted ments are summarized in Figure 3. The constants listed in the figure include the following: versus frequency in Figure 4. The noise equiva- lent spectrum, which is an approximation of the K = K wave spectrum, is plotted in Figure 5. B=B G3B = g G2B = g KB = CBO = c BUOY - EBO3 DATA TIME - 1/28/78 9: 0 LATITUDE - 0.560000E 2 LONGITUDE = 0.147900E 3 NUMBER OF FILTERS (K) = 12 A-TO-D SENSITIVITY (B) = -0. 204800E 3 GAIN OF OUTPUT BUFFER (G3B) = 0.200000E 1 GAIN OF SQUARING CIRCUIT (G2B) = 0.100000E 0 SLOPE OF ACC. LINEAR CAL. CURVE (KB) = 0.510000E 0 ZERO VOLTAGE COUNT (CBO) = 0.204750E 4 GAIN(K) OPTION (I=COMPUTE, 2=INPUT) = 2 OFFSET OPTION (1=COUNTS, 2=VOLTS) = 1 * FILTER DATA: Ck c+ Bek Ck Ck s(i)(fk) FILTER CENTER BAND- GIB OFFSET MEASURED COUNT MAX DISP NOISE EQUIV. SPECTRAL POTENTIOMETER SETTINGS K (K) FREQ. WIDTH GAIN VOLTS COUNTS VOLTS COUNTS DIFF. SPECTRA ACC. VEL. DISP. P1 P2 P3+P4 1 1 0.0500 0.010 196.07 -0.042 2056.0 3.630 1304.0 -752.0 51.334 0.1811 1.8345 18.5877 0.9506 0.0400 0.1047 2 2 0.0600 0.010 69.44 -0.007 2049.0 0.764 1891.0 -158.0 197.372 0.3014 2.1209 14.9230 0.3367 0.0400 0.1257 3 3 0.0700 0.010 52.36 -0.017 2051.0 0.618 1921.0 -130.0 187.378 0.4333 2.2399 11.5792 0.2539 0.0400 0.1466 4 4 0.0800 0.010 52.36 -0.007 2049.0 2.156 1606.0 -443.0 109.838 1.4657 5.8010 22.9595 0.2539 0.0400 0.1676 5 5 0.0900 0.010 52.36 -0.051 2058.0 2.742 1486.0 -572.0 68.571 1.8768 5.8690 18.3536 0.2539 0.0400 0.1885 6 6 0.1000 0.010 52.36 0.012 2045.0 3.103 1412.0 -633.0 44.989 2.0582 5.2136 13.2061 0.2539 0.0400 0.2094 7 7 0.1225 0.035 28.18 0.100 2027.0 3.748 1280.0 -747.0 19.707 2.3423 3.9538 6.6740 0.4782 0.1400 0.2566 8 8 0.1600 0.040 26.18 0.012 2045.0 3.865 1256.0 -789.0 6.865 2.4130 2.3876 2.3624 0.5077 0.1600 0.3351 9 9 0.2000 0.040 26.18 0.012 2045.0 3.792 1271.0 -774.0 2.812 2.3168 1.4671 0.9291 0.5077 0.1600 0.4189 10 10 0.2400 0.040 26.18 -0.027 2053.0 2.932 1447.0 -606.0 1.356 1.9042 0.8374 0.3683 0.5077 0.1600 0.5027 11 11 0.2800 0.040 26.18 0.012 2045.0 2.507 1534.0 -511.0 0.732 1.9898 0.6429 0.2077 0.5077 0.1600 0.5864 12 12 0.3300 0.060 21.37 0.012 2045.0 1.223 1797.0 -248.0 0.380 1.8511 0.4306 0.1001 0.6217 0.2400 0.6912 *INTERVAL DATA: Wj W. F. j ij Fj j fji) s i) TRANSFER LEFT LEFT CENTER OFFSET SPECTRAL DENSITY MATRIX SPECTRAL DENSITY EMPIRICAL OFFSETS INTERCEPT SLOPE TRANSFER FREQ. K FREQ. WIDTH ACC. VEL. DISP. ACC. VEL. DISP. ACC. VEL. DISP. 0.984 0.601 1.0111 0.0450 1 0.0500 0.0100 -1.190 -11.613-113.926 0.0149 0.1836 0.8185 0.0000 0.0000 -1.2800 0.981 0.657 1.0171 0.0550 2 0.0600 0.0100 3.656 26.230 185.992 0.1441 1.0191 6.5261 0.0000 0.0000 -0.6200 0.978 0.709 1.0237 0.0650 3 0.0700 0.0100 7.761 40.142 206.132 0.0051 0.0929 0.4436 0.0000 0.0000 -0.3300 0.970 0.815 1.0307 0.0750 4 0.0800 0.0100 6.935 27.381 107.449 1.5721 6.2776 24.7491 0.0000 0.0000 -0.2000 0.961 0.919 1.0389 0.0850 5 0.0900 0.0100 6.910 21.591 67.122 1.9371 6.0625 18.7973 0.0000 0.0000 -0.1200 0.955 0.975 1.0481 0.0950 6 0.1000 0.0100 6.753 17.626 45.499 2.0222 5.3360 13.8380 0.0000 0.0000 -0.0800 0.943 1.096 1.0578 0.1050 7 0.1225 0.0350 6.609 11.220 18.612 2.5415 4.3188 7.1460 0.0000 0.0000 -0.0400 0.933 1.164 1.0962 0.1400 8 0.1600 0.0400 6.616 6.487 6.249 2.6583 2.6085 2.5084 0.0000 0.0000 -0.0100 1.133 0.054 1.1428 0.1800 9 0.2000 0.0400 6.425 4.043 2.505 2.6514 1.6633 1.0217 0.0000 0.0000 -0.0100 1.753 -2.765 1.1449 0.2200 10 0.2400 0.0400 6.835 2.995 1.303 2.0040 0.8749 0.3792 0.0000 0.0000 0.0000 3.050 -7.752 1.0343 0.2600 11 0.2800 0.0400 7.738 2.555 0.838 2.2930 0.7432 0.2396 0.0000 0.0000 0.0000 3.377 -8.842 0.7242 0.3000 12 0.3300 0.0600 21.331 5.037 1.177 2.0618 0.4835 0.1121 0.0000 0.0000 0.0000 0.000 0.000 0.1937 0.3600 MATRIX : M(0)= 1.0745, M(2)= 0.6055, M(4)= 0.6539, HI/3= 4.1464, T= 8.370 Figure 3. SAMPLE WSA PRINTOUT 28 25 25 N_ 20 -0 20 N~~~~~~l I 5~~~~~~~~~ lo 10 .05 0.15 0.20 0.25 0.30 0.35 0.05 0.10 0.15 0.20 0.25 0.30 0.35 f (Hz) f (Hz) Figure 4. DISPLACEMENT SPECTRUM Figure 5. NARROW BAND APPROXIMATION OF THE DISPLACEMENT SPECTRUM 6. Summary and Conclusions NDBO plans to perform calibrations and data quality evaluations of the system with the newly When the work described herein was initia- available Waverider Analyzer Satellite Communi- ted, a wave measurement system was needed for cator (WRANSAC), which circumvents many of the the PEB-type payloads that were to be produced logistical problems we have had with conventional and used on NDBO's 10-meter discus hull buoys. Waveriders.13 Such calibration would allow the These payloads were not to have the capability magnitudes of the negative empirical offsets, to accept digital inputs, and were not to be which we believe to be larger than necessary, to programmable internally to perform any function be reduced. beyond averaging samples. Changing the internal structure or interfaces of these payloads to It should be noted that, in the Pacific accommodate digital data was not an option. Ocean, where swells occasionally appear at fre- quencies below 0.45 Hz, the conditions of the NDBO conceived two types of wave measurement mathematical analysis are violated, and the systems that could be interfaced to the analog resulting spectra are distorted. channels of the PEB payload - an analog wave height and period measurement system, and the Since the WSA systems went into service, system described in this paper. We believed that NDBO has developed a system for the estimation the latter would be considerably better, despite of one-dimensional wave spectra. It is intended its inferiority to digital sampling systems such to replace the WSA systems as soon as resources as the ones NDBO now has on some of its more allow; however, even with the limitations im- advanced buoys. We therefore had the WSA systems posed by budgets and the state of the art of designed, fabricated, integrated, and deployed other related systems, the current system has on PEB buoys. Spectra, wave height, and wave performed its function well. Our experience period have been generated successfully with com- with this early system is hereby documented for puter-produced matrices since 1975. a better understanding of its functions and limitations, and its important contribution to During the initial deployments of the PEBs, the development of better wave measurement systems. NDBO simultaneously acquired several hours of Waverider spectral data at the deployment sites. The agreement between these two sources of spec- 7. Acknowledgements tra was satisfactory, and as close as could be expected given the design limitations of the WSA. Mr. Frank Remond of Sperry Support Services designed and led the development of the WSA hard- The remoteness of NDBO's deep ocean sites ware which this paper describes mathematically. impedes the acquisition of the large quantities He provided equation (1) to the authors, and par- of comparison data needed to establish, with ticipated in the development of the mathematical statistical validity, either the hull transfer description of the system. He also read the function or the error bounds of the spectra final version of this paper for correctness. produced. The main problem is the reluctance of ships to standby for days while data is Mr. Billy Graham of Sperry programmed the taken. In the absence of large quantites of mathematics and was essential to making the cal- comparison data, we have had hindcast evaluations culations work correctly. of the WSA data performed by an NDBO contrac- tor.10,11,12 Those interested are encouraged Mrs. Deborah Hojem of Sperry edited the to read these evaluations, which found the paper. The unusually difficult and tedious typ- data to be satisfactory as an approximate ing was done by Mrs. Linda Bridges and Ms. Edna estimate of spectra. Bender, also of Sperry. 29 References 1. Michelena, E.D., K.E. Steele, C.S. Niederman, 12. Hubert, W.E., "Evaluation of Spectral Wave and J. H. Hinchman, "A Reliable Wave Measure- Data Acquired by NOMAD Buoys," Ocean Data ment System for NOAA Data Buoys," Proceedings Systems, Inc., Final Technical Report of the Tenth Annual Conference of the Marine (Task Order No. Eight), April 1977. Technology Society, 1974. 13. Adamo, L.C., K.E. Steele, and E.L. Burdette 2. Steele, K.E., E.D. Michelena, and J.M. Hall, "A GOES Reporting Waverider Buoy," Proceed- "Wave Data Available from NDBO - Current ings of the MTS-IEEE Conference, The Ocean and Planned," Proceedings fo the Interna- Challenge (OCEANS 78), Washington, D.C., tional Symposium on Ocean Wave Measurement Sept. 6-8, 1978. and Analysis (WAVES 74), Vol. 1, p. 417, Sept. 1974. 3. Steele, K.E., J.M. Hall, and F.X. Remond, "Routine Measurement of Heave Displacement Spectra from Large Discus Buoys in the Deep Ocean," Proceedings of the First Combined IEEE Conference on Engineering in the Ocean Environment and the Annual Meeting of the Marine Technology Society (OCEANS 75), Sept. 1975. 4. "Equipment Operations Manual for Phase II Payload System," prepared under Contract NAS 13-16 for the NOAA Data Buoy Office by Magnavox Government & Industrial Electronics Company, July 1976. 5. Steele, K.E. P. Wolfgram, A. Trampus, and B.S. Graham, "An Operational High Resolu- tion Wave Spectrum Analyzer System for Buoys," Proceedings of the OCEANS 76 Confer- ence, Washington, D.C., Sept. 13-15, 1976. 6. "Practical Experience with Buoys," NOAA Data Buoy Office, November 1973. 7. "Technical Manual for the Model 313 Data Acquisition and Control System for the Pro- totype Environmental Buoy," prepared for the NOAA Data Buoy Office by General Dynamics Electronics Division, Report No. TM-75-PEB-3, July 1975. 8. Remond, F.X., "Ocean Spectrum Measurement with Analog Filters," Proceedings of 22nd International Instrumentation Symposium, San Diego, Calif., May 1976. 9. Burdette, E.L., "The Comparison of Discus- Buoy Wave Spectra Produced by Body-fixed and Vertically Stabilized Accelerometers," Proceedings of the MTS-IEEE Conference, The Ocean Challenge (OCEANS 78), Washington, D.C., Sept. 6-8, 1978. 10. Hubert, W.E., "Evaluation of Spectral Wave Data from EB-41," Ocean Data Systems, Inc., Technical Report 6, January 1976. 11. Hubert, W.E., "An Operational Evaluation of PEB Buoy Systems," Ocean Data Systems, Inc., Final Technical Report (Task Order No. Six), December 1976. 30 A GOES-REPORTTNG WAVERIDER BUOY L. C. MAamo Kenneth E. Steele Ernest L. Burdette Louis C. Adamo, Inc. NOAA Data Buoy Office Computer Sciences Corp. P. O. Box "L" NSTL Station, MS 39529 NSTL Station, MS 39529 Solana Beach, CA 92075 DURATION WOULD BE REQUIRED TO OBTAIN THE TRANSFER FUNCTION OF A DATA BUOY HULL. THIS IS A NEW INSTRUMENT FOR MEASURING AND REPORTING PROHIBITIVELY EXPENSIVE, AND THERE ARE OTHER OCEAN WIND WAVES HAS BEEN CONSTRUCTED, TESTED AT DIFFICULTIES. SEA, AND PUT TO USE IN FIELD STUDIES OF THE A WAVERIDER EOUIPPED WITH INTERNAL DATA TRANSFER FUNCTION OF OPERATIONAL WAVE MEASURING PROCESSING CAPABILITY AND A SATELLITE SYSTEMS ON LARGE DATA BUOYS. THE PROTOTYPE TANSAC WAS INTEGRATED FONLABROM AN EXISTING COMMUNICATIONS LINK COMPARABLE TO THE LONG RANGE LINK ON THE DATA BUOYS WOULD OVERCOME ALL THESE MICRO-COMPUTER, A SATELLITE RADIO TRANSMITTER, AND D INK ON THE DATA BUOYS WOULD OVERCOME ALL THESE A WAVERIDER. TWENTY MINUTES OF WAVE DATA ARE DIFFICULTIES. IN 1976, A FEASIBILITY STUDY (REF. 1) WAS CARRIED OUT TO EXAMINE THIS POSSIBILITY. PROCESSED AND PRODUCE HOURLY REPORTS OF MEAN WAVE 1) WAS CARRIED OUT T N THIS POSSIBILITY. HEIGHT, MAXIMUM CREST, MINIMUM TROUGH, AND 150WING CONCLUSIONS WERE OBVIOUS IN LATE LAGS OF THE AUTOCORRELATION FUNCTION. THE INTERNATIONAL NETWORK OF SYNCHRONOUS INTEORNATIONAL NETWORK OF SYNCHRONOUS A RELAA. A PROTOTYPE WAVERIDER ANALYZER SATELLITE METEOROLOGICAL SATELLITES PROVIDES A RELAY COMMUNICATOR (WRANSAC) COULD BE BUILT CAPABILITY PERMITTING WRANSAC TO BE USED ALMOST ANYWHERE IN THE WORLD OCEAN. THE DEVELOPMENT OF MAKING EXTENSIVE USE OF EXISTING HARDWARE THE WRANSAC PROTOTYPE IS DESCRIBED AND DATA FROM AND SOFTWARE WITH A MINIMUM F DEVELOPMENT; B. THE PROTOTYPE COULD BE EXPECTED TO THE DECEMBER 1977 DEPLOYMENT IS PRESENTED. PROVIDE ONE-DIMENSIONAL SPE CTRAL WAVE DATA PROVIDE ONE-DIMENSIONAL SPECTRAL WAVE DATA IN THE SAME FORMAT AS THE MOST SOPHISTICATED 1. INTRODUCTION OPERATIONAL DATA BUOYS AND WITH ACCURACIES CLOSE TO THE WAVERIDER'S SPECIFICATIONS; THE NOAA DATA BUOY OFFICE (NDBO) HAS BEEN C. SERVICE LIFE COULD BE EXPECTED TO BE SIX OPERATING SEVERAL DIFFERENT WAVE MEASUREMENT MONTHS WITH HOURLY REPORTS; SYSTEMS INSTALLED ON A VARIETY OF DATA BUOY HULLS. D. THE RELIABILITY OF THE COMMUNICATIONS FOR THE MOST PART, THESE SYSTEMS REPORT IN LINK AS A FUNCTION OF ANTENNA PATTERN, REAL-TIME, AND THE REPORTS ARE WIDELY DISSEMINATED TRANSMITTER OUTPUT, SATELLITE ELEVATION AND BY THE NATIONAL WEATHER SERVICE. NDBO HAS HAD A SEA CONDITONS COULD BE BEST EVALUATED IN REOUIREMENT FOR A HIGH-QUALITY WAVE MEASUREMENT THE FIELD. SYSTEM TO DETERMINE THE TRANSFER FUNCTION OF EACH OF THE SEVERAL HULL TYPES IN USE, TO VERIFY PROPER THE WRANSAC WOULD HAVE OTHER APPLICATIONS. OPERATION OF THE WAVE SYSTEMS ON NEWLY DEPLOYED ITS EASE OF DEPLOYMENT WOULD ALLOW ONE OR MORE DATA BUOYS, AND TO PROVIDE A REFERENCE MEASUREMENT UNITS TO BE DEPLOYED TN THE PATH OF TROPICAL BY WHICH THE QUALITY OF THE DATA PRODUCED BY THE STORMS FOR SPECIAL STUDIES. TT COULD BE USED AS DIFFERENT TYPES OF HULLS AND MEASUREMENT SYSTEMS AN OPERATIONAL SYSTEM FOR OBTAINING WAVE SPECTRA CAN BE JUDGED OVER AN EXTENDED PERIOD OF TIME. WHERE SHIP TRAFFIC AND VANDALISM WO�LD NOT INTERFERE. IT COULD ALSO BE USED AS A BACKUP BETWEEN 1973 AND 1978, THE WAVERIDER INSTRUMENT ON THE HIGH-SEAS IN THE EVENT OF THE (TRADEMARK OF DATAWELL BV, HAARLEM, THE FAILURE OF AN OPERATIONAL DATA BUOY. NETHERLANDS) ACCELEROMETER BUOY WITH ITS OUTPUT RECORDED ON AN INSTRUMENTATION FM MAGNETIC TAPE IN MID-1077, NDBO CONTRACTED WITH LOUIS C. RECORDER, WAS USED TO MEET THESE REQUIREMENTS. ADAMO, INC. TO BUILD A PROTOTYPE (REF. 2). THE OVERALL, THESE ATTEMPTS WERE NOT SATISFACTORY FOR PROTOTYPE WAS BUILT, TESTED IN THE LABORATORY, AND A NUMBER OF REASONS. THE PRINCIPAL DEFICIENCY TESTED AT SEA IN DECEMBER 1977. ALL TESTS STEMS FROM THE COMBINATION OF THE WAVERIDER'S INDICATED IT PERFORMS ACCORDING TO THE LIMITED TELEMETRY RANGE (50 KM) AND THE HIGH-SEA EXPECTATIONS OF THE FEASIBILITY STUDY. AVAILABLE LOCATIONS OF NEARLY ALL THE OPERATIONAL DATA EVIDENCE INDICATES THAT COMMUNICATIONS ARE BUOYS. BASICALLY, TO TAKE THE REQUIRED DATA WITH ABSOLUTELY RELIABLE WITH SATELLITE ELEVATIONS A WAVERIDER ALONGSIDE AN OPERATIONAL DATA BUOY, ABOVE 40, ALTHOUGH THE SYSTEM MAY NOT BE OPTIMUM REOUITRES THAT A LARGE SHIP STAND BY FOR A WITH RESPECT TO MINIMIZING POWER CONSUMPTION. RELATIVELY LONG TIME, HOPING FOR A VARIETY OF SEA CONDITIONS, WHILE THE NDBO TECHNICIAN MONITORS THE FOLLOWING SECTIONS DESCRIBE THE SYSTEM, DATA BEING ACOUIRED BY THE WAVERIDER. IT APPEARS LABORATORY TESTS, THE SEA TEST, TEST DATA, NOW THAT A MEASUREMENT PERIOD OF AT LEAST ONE WEEK CONCLUSIONS, AND OPERATIONAL RESULTS TO DATE. 31 2. SYSTEM DESCRIPTION THE TWA WAS DESIGNED FOR THE LARGER NDBO BUOYS, AND HAS BEEN RELIABLY FUNCTIONING THE WRANSAC STRATEGY WAS TO MAINTAIN THE ABOARD VARIOUS OF THEM SINCE OCTOBER 1976. IT WAS FEATURES INHERENT IN THE WAVERIDER WHILE DESCRIBED BY STEELE, ET AL (REF. 3). OVERCOMING ITS TWO LIMITATIONS OF SHORT TELEMETRY RANGE AND NEED FOR A SEPARATE EXTERNAL DATA THE WDA WAS DESIGNED TO ACCEPT AN ANALOG ACQUISITION SYSTEM; AND TO UTILIZE AVAILABLE VOLTAGE FROM EITHER AN ACCELERATION OR HARDWARE AND SOFTWARE. THE RESULTANT PROTOTYPE DISPLACEMENT TRANSDUCER. ALTHOUGH ACCELERATION INSTRUMENT HOPEFULLY WOULD BE CAPABLE OF CARRYING TRANDUCERS HAVE BEEN EMPLOYED ON VIRTUALLY ALL OUT MOST OF THE TASKS ENVISIONED FOR WRANSAC AS NDBO BUOYS, THIS DISPLACEMENT OPTION WAS INCLUDED WELL AS PROVIDING INFORMATION ON THE RELIABILITY SO THAT THE WDA COULD BE USED WITH THE OUTPUT OF A OF THE COMMUNICATION LINK UNDER VARIOUS DATAWELL HEAVE SENSOR, A WAVERIDER, OR A WAVE CONDITIONS. THE WRANSAC PROTOTYPE WAS INTEGRATED STAFF, WITHOUT REDESIGN. FROM THE FOLLOWING MAJOR ELEMENTS: +TWAVERTDER/FL-6900 SERIES; THE 0.oM-DIAMETER AN AMPLIFIER IS INCLUDED AT THE WDA INPUT TO BTUOY PRODUCED BY DATAWELL; OPTIMIZE THE TRANSDUCER SIGNAL LEVEL AT THE 44+AVE DATA ANALYZER (WDA); A SPECIALIZED ANALOG-TO-DIGITAL CONVERTOR. THE SIGNAL PASSES MICRO-COMPUTER SYSTEM, DESIGNED BY MAGNAVOX, THROUGH A SIX-POLE BUTTERWORTH LOW-PASS FILTER WHICH SAMPLES AN ANALOG SIGNAL, CALCULATES WITH HALF-POWER FREQUENCY AT 0.5sZ. THE FILTER THE FIRST 150 LAGS OF THE AUTOCORRELATION OUTPUT IS SAMPLED AT A RATE OF 1.5 HZ TO PRODUCE FUNCTION, AND STORES THE RESULTS IN A RANDOM 8-BIT COUNTS. RELATIVE TO THE UNITY GAIN AT VERY ACCESS MEMORY; LOW FREQUENCIES, THE GAIN OF THE FILTER AT 1.OHZ +GOES DATA TRANSMIT TERMINAL (GOTT); A DATA IS 0.000244. THUS, THE ALIASING ALLOWED BY THE ACQUISITION SYSTEM AND RADIO TRANSMITTER, WDA AT THE HIGHEST FREOHENCY OF INTEREST, 0.5 HZ, DESIGNED BY AMERICAN ELECTRONICS LABORATORY, IS INSIGNIFICANT. WHICH SELF-INIT IATES REPORTS TO THE SYNCHRONOUS METEOROLOGICAL SATELLITES A MICRO-PROCESSOR IN THE WDA FORMS THE (AMERICAN, GOES-I, SMS-2; EUROPEAN, METEOSAT; EOUIVALENT OF AUTOCOVARIANCES FROM THE RAW SOVIET, GOMS; JAPANESE, GMS); SAMPLES. THE COMPLETE RECORD IS NEVER STORED. +CHU ASSOCIATES INC. MICROWAVE ANTENNA MODEL THE SUMS NEEDED FOR AUTOCOVARIANCES ARE FORMED CA-3140; DURING THE 2/3-S BETWEEN SAMPLES. SAMPLING +SOFTWARE ON SHORE DESIGNED BY NDBO TO CONTTNUES FOR TWENTY MINUTES, FOR A TOTAL OF IR00 PROCESS THE AUTOCORRELATION FUNCTION OF SAMPLES. THESE SUMS AND OTHER WAVE DATA ARE ACCELERATION OR DISPLACEMENT INTO AN ESTIMATE FORMATTED INTO 155, 16-BIT WORDS AND STORED IN A OF THE SPECTRAL DENSITY FUNCTION OF SEA BUFFER IN THE TDA. FROM THE BUFFER, THE INTERFACE SURFACE DISPLACEMENT. TO THE GDTT RELAYS THE WORDS UPON REQUEST BY THE GT)TT, TO THE GDTT. THE GDTT TRANSMITS THE DATA A SYSTEM BLOCK DIAGRAM TS PRESENTED TN FIGURE 1. VIA THE SATELLITE TO WALLOPS ISLAND AND THEN TO THE NATIONAL SPACE TECHNOLOGY LABORATORIES (NSTL). THE MAIN PROBLEMS SOLVED IN THE COURSE OF INTEGRATION WERE: I. PROVIDING ENOUGH POWER FOR RAW DISPLACEMENT SPECTRA, Sd (K), ARE THE WDA WHITCH WAS DESICGNED FOR USE ON MUCH LARGER CONSTRUCTED AT NSTL FOR FREQUENCIES F = K XEF, K BUOYS; AND 2. DEVELOPING A DIGITAL INTERFACE EQUALS 0, 1, 2, ... M, IN WHICH AF = 0.005HZ BETWEEN THE WDA AND THE GDTT. AND M = 150. THESE DISPLACEMENT SPECTRA ARE CONSTRUCTED IN THE SAME MANNER AS FOR THE LARGE THE POWER PROBLEM WAS SOLVED BY SELECTING THE NDBO BUOYS, EXCEPT THAT NO MULTIPLICATION BY LARGEST LITHTIUM CELLS AVAILABLE IN PRODUCTION AS FREQUENCY-TO-THE-MTNUS-FOURTH-POWER IS NECESSARY THE POWER SOURCE FOR THE PROTOTYPE. THESE CELLS TO TRANSLATE ACCELERATION SPECTRA TO DISPLACEMENT (POWER CONVERSION INC. MODEL 660-5AS) HAVE AN SPECTRA. A SMOOTH ESTIMATE OF BUOY DISPLACEMENT ENERGY DENSITY OF 300WH PER KG, A NEARLY CONSTANT POWER SPECTRAL DENSITY TS OBTAINED BY HANNING: VOLTAGE, AND CAN TOLERATE SURGES DURING DISCHARGE. THE DIGITAL INTERFACE WAS DESIGNED AND FABRICATED Sd(l) = n.5 sd(1) + 0.25 Sd (2) BY DATA/WARE DEVELOPMENT, INC. UNDER SUBCONTRACT. THEIR DESIGN IS IMPLEMENTED WITH CURRENT C/MOS SdfR) = 0.25 (K-I) + 0.5 S (K) + 0.25 Sd(K+I), TECHNOLOGY. A CASSETTE TAPE RECORDER WAS INCLUDED Sd = . IN ORDER TO EXPEDITE CHECKOUT AND TO PROVIDE THE K = 2, 3, . . (M -1) CAPABILITY FOR RECORDING COMPLETE SAMPLES AND CALCULATED RESULTS FOR ABOUT 100 RECORDS, FOR THOSE APPLTCATTONS TN WHICH THE AUTOCORRELATION S (M = 0.5 Sd(M - 1) + 0.5 SO(M) FUNCTTON TS INSUFFICIENT. 32 THROUGH THE APPLICATION OF A SYSTEM POWER TRANSFER 3. LABORATORY TESTS FUNCTION B(K) AND A NOISE CORRECTION FUNCTION N(K), AN ESTIMATE OF WAVE DISPLACEMENT POWER THE LABORATORY TESTS WERE CONDUCTED IN TW( SPECTRAL DENSITY W(K) IS PRODUCED: PARTS: FIRST, A CHECKOUT OF THE WDA, INTERFACE, GDTT, TAPE RECORDER AND SATELLITE COMMUNICATIONS; W(K) = (SA(K) - N(K))/B(K), K 1,2, . . . . SECOND, A COMPLETE SYSTEM TEST INCLUDING THE WAVERIDER. TO ACCOMPLISH THE FIRST CHECK TWO THE WRANSAC IS SO QUIET THAT N (K) IS SET TO 0. SPECIAL PROGRAMS WERE WRITTEN AND LOADED INTO THE CALCULATION OF MOMENTS Mi AND PERIOD T ARE ALSO WDA'S PROGRAMMABLE READ-ONLY MEMORY. THESE MADE. (REF. 4) PROGRAMS CAUSED THE WDA TO LOAD PARTICULAR j i PATTERNS OF NUMBERS INTO ITS RAM TO BE READ VIA Mi= E (2rK6F) W(K) bF i =0, 2 THE INTERFACE TO THE CASSETTE, RECORDED, AND THEN katM1/2 TRANSMITTED. COMPARISON OF THE DATA RECORDED AND 8H1 = 4 oM/2 RECEIVED VIA THE SATELLITE LINK VERIFIED CORRECT OPERATION OF THE INTERFACE. WITH THE NORMAL T = 21 0 PROGRAM INSTALLED IN THE WDA, SINUSOIDAL INPUTS "Ol 2 )1/2 FROM A SIGNAL GENERATOR WERE INPUT TO THE WDA, THE SYSTEM POWER TRANSFER FUNCTION COMPRISES PROCESSED AND STORED ON THE MAGNETIC TAPE. THE THE EFFECT OF THE WAVERIDER HULL TRANSFER MAGNETIC TAPE RESULTS WERE COMPARED WITH THE FUNCTION, THE WAVERTDER DOUBLE-INTEGRATOR CHARACTERISTCS OF THE INPUT WAVE TO VERIFY THE FILTERING FUNCTION, AND THE BUTTERWORTH FILTER IN OPERATION OF THE MICRO-COMPUTER AND ITS SOFTWARE. THE WDA. FINALLY, THE AUXILIARY HARDWARE WAS INSTALLED THE PRINCIPAL SPECIFICATIONS OF THE WRANSAC IN THE WAVERIDER AND AN END-TO-END SYSTEM TEST WAS PROTOTYPE ARE GIVEN BELOW. RUN. THIS INCLUDED THE STANDARD TEST FOR ATTITUDE OF THE WAVERIDER PLATFORM, AND THE TEST OF VERTICAL SENSITIVITY PERFORMED ON A VERTICAL TEST * * * STAND WHICH RAISES AND LOWERS THE WAVERIDER WITH A NEARLY SINUSOIDAL MOTION OF APPROXIMATELY 3-S PHYSICAL PERIOD AND 40-CM AMPLITUDE. A COMPUTER PROGRAM 170KG WAS WRITTEN TO SIMULATE THE WDA SOFTWARE ON A 0.9M DIAMETER GENERAL PURPOSE COMPUTER. THIS PROGRAM WAS USED 1.5M/S MAXIMUM OPERATING SURFACE CURRENT TO PROCESS THE DIGITAL DATA RECORDED ON THE 4300 REPORTS AT HOURLY INTERVALS CASSETTE TAPE. SENSOR THE RESULTS WERE COMPARED WITH THE RESULTS OF WAVE HEIGHT THE WDA'S CALCULATIONS, ALSO STORED ON THE TAPE, MINIMUM: P-P NOISE (1HZ BANDWIDTH AND THOSE SAME RESULTS RECEIVED AT NSTL AFTER EOUIVALENT TO 2CM) TRANSMISSION THROUGH THE GOES. EXACT AGREEMENT MAXIMUM: 40M WAS FOUND. THE SENSITIVITY OF THE SYSTEM AS ACCURACY: +/- 4% BETWEN 0.035rZ AND 0.5HZ CONFIGURED FOR INITIAL DEPLOYMENT IS 12.25 DIGITAL COUNTS PER METER VERTICAL DISPLACEMENT. FULL ACCELEROMETER LINEARITY RANGE IS 256 COUNTS OR APPROXIMATELY PLUS/MINUS 10 NON-LINEARITY RECTIFICATION: METERS. <0.002M/S/S FOR 6M/S/S AMPLITUDE HORIZONTAL SENSITIVITY <3% OF VERTICAL SENSITIVITY 4. SEA TEST MICRO-COMPUTER INPUT BANDWIDTH: 0 TO 0.5HZ AFTER COMPLETION OF THE LABORATORY TESTS, DIGITIZATION: 8 BITS 1WRANSAC WAS DEPLOYED TOGETHER WITH A STANDARD OUTPUT: 155 WORDS, EACH 15 DATA BITS WAVERTDER OFF LA JOLLA, CALIFORNIA, ON 7 DECEMBER INCLUDING, MEAN, MAX CREST1977. THE LOCATION WAS 32 51.'N, 117015.9'W IN MINL TROUG, MEAN,150 LAXOES 18M DEPTH OF WATER. THE TWO BUOYS WERE 150M MIN TROUGH, 150 LAGS SAMPLTNG- 1.5RZ FOR 20 MINUTES/RECORD APART. DEPLOYMENT WAS ACCOMPLISHED FROM A SMALL BOAT LOWERED FROM SCRIPPS PIER WITH THE RECORDER CAPACITY COOPERATION OF THE SHORE PROCESSES GROUP AT 104 RECORDS, EACH COMPRISED OF 1800 SAMPLES SCRIPPS INSTITUTION OF OCEANOGRAPHY. ACTUAL AND 155 WORDS OF CALCULATED RESULTS DEPLOYMENT INCLUDING TIME SPENT TN A OUICK SURVEY OF THE BOTTOM IN THE DEPLOYMENT AREA WAS ABOUT GDTT THREE HOURS. THE BUOYS WERE RETRIEVED ON 5 RADIATED POWER: 40W 1 400MHZ JANUARY 1078. REPORT INTERVAL: JUMPER PROGRAMMABLE FOR 1-12HRS IN 1-HR INCREMENTS OUTPUTS: ASCII ENCODED, ]00 BITS/S 33 THE FM RECORDINGS ENCOUNTERED A TYPICAL DURING THIS TEST, UNUSUAL WEATHER CONDITIONS SAMPLE OF PROBLEMS. AS MENTIONED ABOVE, A NEARBY OCCURRED IN THE SAN DIEGO AREA, INCLUDING THE TRANSMITTER CAUSED RADIO INTERFERENCE DURING ABSENCE OF NORMAL WINTER STORMS, AND THE BUSINESS HOURS. THIS IS SIMILAR TO THE OCCURRENCE OF UNUSUAL CALMS AND STRONG EASTERLY INTERFERENCE FROM A VESSEL'S COMMUNICATION WINDS. AS A RESULT, THE TWO BUOYS DID NOT TRANSMITTER WHEN THE WAREP IS USED AT SEA. THE EXPERIENCE A WIDE RANGE OF WAVES APPROACHING FROM STRONG EAST WINDS FILLED OUR OFFICES WITH SAND AND SEAWARD, AND THERE WAS ONLY A SHORT FETCH DUST, CONTAMINATING THE RECORDER AND TAPES. SHOREWARD OF THE BUOYS. THE DIFFERENCE IN LOCATION OF THE BUOYS AND THE PRESENCE OF LA JOLLA THE RECORDS SELECTED FOR COMPARISON WITH SUBMARINE CANYON COULD BE EXPECTED TO INTRODUCE WRANSAC WERE TAKEN DURING THE HOLIDAY WEEK 27-31 SIGNIFICANT DIFFERENCES IN WAVES AT THE TWO DECEMBER, AT 2-HR INTERVALS, WHEN INTERFERENCE WAS LOCATIONS OVER SOME OF THE WIND CONDITIONS THAT AT A MINIMUM. THE TIME HISTORIES OF THE SURFACE OCCURRED. DURING THE DEPLOYMENT AN ARRAY OF SYSTEMS AND THE NEARBY PRESSURE ARRAY ARE SHOWN IN PRESSURE SENSORS, PART OF THE CALIFORNIA COASTAL FIGURE 2. (REF. 5) A SAMPLE OF THE SPECTRA FROM ENGINEERING DATA NETWORK, WAS OPERATING OFF THE WRANSAC (W) AND THE WAVERIDER (E) IS SHOWN IN END OF SCRIPPS PIER. DUE TO THE COMPLICATED FIGURE 3. THESE SPECTRA, TAKEN AT A TIME OF TOPOGRAPHY IN THE AREA, DIRECT COMPARISON BETWEEN RELATIVELY RAPID CHANGE IN TOTAL ENERGY, THE TWO SYSTEMS TS NOT STRAIGHTFORWARD. THE ILLUSTRATE THE CLOSE CORRESPONDENCE AT FREQUENCIES SHOALING COEFFICIENT FOR DEEPWATER WAVES OF 2.8 BELOW 0.15HZ, AND THE GENERALLY ELEVATED VALUES OF SECONDS PERIOD IS ALREADY REDUCED TO 0.95 AT THE THE WAVERIDER AT HIGHER FREQUENCIES. THE WRANSAC DEPTH OF THE BUOY INSTALLATION. THERE ARE NO WIND PLOT BEGINS AT 0.02HZ, WHILE THE WAVERIDER PLOT DATA NOR WIND OBSERVATIONS AVAILABLE AT SCRIPPS BEGINS AT 0.03HZ. THE LOW FREQUENCY DISCREPANCIES PIER DURING THE TEST. THE STANDARD WAVERIDER DATA ARE WELL-WITHIN THE 80T CONFIDENCE LIMITS WAS RECORDED IN OUR OFFICES USING A DATAWELL WAREP ESTIMATED FROM A CHI-SQUARE DISTRIBUTION, USING AND AN INSTRUMENTATION FM TAPE RECORDER. THE THE DEGREES OF FREEDOM MENTIONED ABOVE. WAREP AUTOMATICALLY CONTROLLED THE INTERVAL AND DURATION OF THE RECORDINGS TO SYNCHRONIZE WITH THE FIGURE 4 PRESENTS THE TIME HISTORY OF A WRANSAC'S SAMPLE PERIOD. DUE TO THE RUGGED SINGLE SPECTRAL ESTIMATE (0.13HZ) FROM THE WRANSAC OVERLAND TRANSMISSION PATH, THE TELEMETRY SYSTEM. WAS OPERATING NEAR ITS LIMIT. AS A RESULT, RECORDS MADE DURING REGULAR BUSINESS HOURS WERE LINEAR REGRESSIONS WERE COMPUTED FOR FIVE OFTEN INTERRUPTED BY A NEARBY POWERFUL PAIRS OF SERIES, INCLUDING THOSE PLOTTED IN OFTN NY A NEARBY POWERFUL TRANSMITTER. FIGURES 2 AND 4, PLUS SPECTRAL ENERGY DENSITIES AT THREE OTHER FREQUENCIES. THE REGRESSIONS WERE NOT THE FM TAPES WERE ANALYZED AT NSTL. SPECTRA FORCED THROUGH THE ORIGIN. THE RESULTS ARE WERE CALCULATED FROM ABOUT 1000, 1HZ SAMPLPRESENTED IN THE FOLLOWING TABLE. A RESOLUTION OF 0.01HZ (50 LAGS). DISPLACEMENT RESOLUTION WAS 205 COUNTS PER METER. THESE SPECTRAL ESTIMATES HAVE AT LEAST 36 EQUIVALENT REGRESSIONS: 27 DECEMBER - 31 DECEMBER 1977 NUMBER OF DEGREES OF FREEDOM. NO ANTI-ALIASING FILTER WAS USED DURING DIGITIZATION OF THESE TAPES CORREL AND THE RESULTING NOISE IS VARIABLE FROM PARAM SERIES NBR RANGE SLOPE INTRCPT COEFF AND THE RESULTING NOISE IS VARIABLE FROM RECORD-TO-RECORD AND REEL-TO-REEL. NO NOISE CORRECTION HAS BEEN APPLIED TO DATA PRESENTED HEREIN, EXCEPT TO IGNORE ESTIMATES AT 0.01 AND [CM] 0.02HZ. STG HT E VS P 8 35-83 0.53 7.34 0.80 [CM] FOR THE COMPARISON PLOTS SHOWN HEREIN, ENERGY DENSITY [CMCM/HZ] SPECTRA BASED ON THE WRANSAC REPORTS WERE 0.05HZ E VS W 40 50-700 1.04 16.36 0.87 CALCULATED USING ONLY 75 LAGS; THUS IN THE COMPARISONS THESE SPECTRA HAVE 48 DEGREES OF 0.09HZ E VS W 40 200-3000 0.90 132.82 0.93 FREEDOM. PRESENTATIONS OF WRANSAC DATA EXCLUSIVELY ARE BASED ON 150 LAGS AND HAVE 24W 40 600-6300 0.85 130.00 0.93 DEGREES OF FREEDOM. 0.17HZ E VS W 40 300-3000 0.60 343.48 0.81 5. SEA TEST DATA NOTES: E DESIGNATES WAVERIDER; W, WRANSAC; THE AVAILABLE DATA INCLUDE 30 DAYS OF HOURLY P, PRESSURE ARRAY. REPORTS FROM WRANSAC, THE COMPLETE RECORD OF SAMPLES FOR THE FIRST 104 REPORTS, FM RECORDINGS CONSIDERING THE SMALL RANGE OF SEA STATES AND OF THE WAVERIDER RECORDS MADE AT EITHER 2- OR 4-HR ENERGY DENSITIES REPRESENTED IN THE DATA, AND THE INTERVALS THROUGHOUT THE PERIOD, AND CORRESPONDING PROBLEMS WITH THE FM RECORDINGS, THE CORRELATIONS STRIP-CHART RECORDS FOR MOST OF THE FM RECORDINGS. BETWEEN THE TWO INSTRUMENTS ARE QUITE HIGH. ONLY A FEW SAMPLES OF THESE DATA ARE PRESENTED HEREIN. THE HIGH DENSITY OF THE WRANSAC DATA IN THE FREQUENCY VS. TIME PLANE PROVIDES UNUSUAL DETAIL 34 AS SHOWN IN FIGURES 5 AND 6. THE DATA PRESENTED THIS LONG TEST SHOWED THAT, IN THE ABSENCE OF IN FIGURE 5 WERE OBSERVED DURING THE PERIOD OF HIGH OR VARIABLE SEA STATES, MANY DAYS OF HOURLY RECORD-BREAKING SANTANA WINDS. THESE ARE SOUTHERN DATA ARE REOUIRED TO PRODUCE A STATISTICALLY CALIFORNIA'S FOHN WINDS WHICH ARE GENERALLY RELIABLE HULL TRANSFER FUNCTION. JUST EXACTLY HOW NORTHEAST, BUT STRONGLY INFLUENCED BY LOCAL MANY DAYS ARE NEEDED IS A MATTER OF JUDGEMENT, BUT TOPOGRAPHY. THESE WINDS REACHED THEIR PEAK DURING TWO IS TOO FEW AND FORTY-TWO IS MORE THAN THE PERIOD INCLUDED IN FIGURE 5, DYING AWAY TO A NECESSARY. AT EITHER EXTREME THE DURATION IS CALM AT THE END OF THAT PERIOD. WITH THE LONGER THAN IS CONVENIENT FOR A SHIP TO STAND BY. INSTRUMENTS ONLY ABOUT 1000M FROM SHORE, NO LOCAL IT IS THIS FACT THAT MAKZS THE WRANSAC SO VALUABLE WAVES WERE GENERATED, YET THOSE APPROACHING FROM TO NDBO IN ITS OPERATIONS. SEAWARD WERE STRONGLY ATTENUATED. FIGURE 6 PRESENTS DATA OF A MORE TYPICAL NATURE FOR AS SHOWN IN REFERENCE 1, THERE IS ROOM FOR DECEMBER. THE HIGHER FREQUENCY MAXIMUM RESULTS IMPROVEMENT IN THE PROTOTYPE: THE COMPUTER AND FROM THE TYPICAL LAND BREEZE, SEA BREEZE REGIME. GDTT CAN BE SIMPLIFIED, POWER CONSUMPTION CAN BE THE TWO LOW FREQUENCY RIDGES AT 0.12HZ AND 0.05HZ REDUCED, AND MORE FLEXIBILITY CAN BE INTRODUCED ARE SWELL FROM DISTANT STORMS. THE FORMER IS INTO THE DATA PROCESSING. SUCH CHANGES SHOULD ASSOCIATED WITH A GULF OF ALASKA SYSTEM, WHILE THE RESULT IN A LESS EXPENSIVE, MORE RELIABLE LATTER APPEARS TO HAVE A SOURCE IN THE NORTHWEST INSTRUMENT, SUITABLE FOR WIDE APPLICATION. PACIFIC. MOST OF THE FEATURES IN FIGURE 6 WOULD BE VERY POORLY REPRESENTED WITH SAMPLING LESS ACKNOWLEDGMENTS DENSE THAN WRANSAC PROVIDED. THE WORK DESCRIBED HEREIN WAS PERFORMED UNDER 6. CONCLUSIONS AND RECENT RESULTS NDBO CONTRACT NO. 03-7-038-720 AND COMPUTER SCIENCES CORP. P.O. NO. A-016. THE AUTHORS THE CONCLUSIONS OF THE FEASIBILITY STUDY OF WERE ASSISTED BY MR. FRANK REMOND OF SPERRY THE WRANSAC CONCEPT HAVE BEEN PUT TO THE TEST BY SUPPORT SERVICES. A MAJORITY OF THE CREDIT FOR CONSTRUCTING, TESTING, AND USING A PROTOTYPE THE SMOOTH TECHNICAL PROGRESS OF THIS WORK TS DUE WRANSAC INSTRUMENT. THE RESULTS HAVE BEEN TO MR. RICHARD KEELE, OF DATA/WARE DEVELOPMENT, EXTRAORDINARILY SATISFYING WITH THE PROTOTYPE INC., WHO WAS RESPONSIBLE FOR THE DIGITAL OPERATING AS EXPECTED IN EVERY WAY. CONCERNS INTERFACE. DATAWELL, BV PROVIDED INVALUABLE ABOUT RELIABLE COMMUNICATIONS FROM A SMALL BUOY TECHNICAL INFORMATION ON THE WAVERTDER AS WELL AS ARE DIMINISHING. WHILE NO SPECIFIC TESTS HAVE THE MODIFIED TOP HATCH TO MATE WITH THE MICROWAVE BEEN CONDUCTED, 1700 TRANSMISSIONS OF ABOUT 4.2 ANTENNA. DEPLOYMENT FROM SCRIPPS PIER WAS MADE MILLION BITS HAVE BEEN COMPLETED, WITH SEAS UP TO POSSIBLE THROUGH THE COOPERATION OF THE SHORE 3M SIGNIFICANT HEIGHT FROM 300N LAT. ONE THOUSAND PROCESSES GROUP AT SCRIPPS INSTITUTION OF OCEANOGRAPHY. THE PREPARATION OF THIS PAPER WAS FIFTY-THREE REPORTS WERE EXAMINED, INCLUDING THE OCEANOGRAPHY. THE PREPARATION OF THIS PAPER WAS HIGHEST SEA STATES, AND FIVE REPORTS WERE FOUND TO SUPPORTED BY THE AUTHORS' RESPECTIVE CONTAIN TRANSMISSION PARITY ERRORS, WHILE 13 ORGANIZATIONS. REPORTS WERE MISSING DUE TO GROUND STATION REFERENCES PROBLEMS. THIS IS GOOD EVIDENCE THAT THE PROTOTYPE IS PROBABLY OVER-POWERED. 1. ADAMO, L. C., "AN EVALUATION OF THE FEASIBILITY OF INTEGRATING THE WAVE DATA SINCE COMPLETION OF THE DECEMBER 1977 TEST, ANALYZER AND GOES DATA TRANSMIT TERMINAL WITH WRANSAC HAS BEEN USED TO ESTABLISH THE HULL POWER THE WAVERIDER BUOY SYSTEM TO REPORT THROUGH TRANSFER FUNCTION OF TWO OPERATIONAL NDBO BUOYS. THE GOES DATA NETWORK." NDBO CONTRACT IN MARCH 178 TT WAS TETHERED TO A 12-METER 01-6-038-901, APRIL 1077. (UNPUBLISHED) DIAMETER DISCUS HULL LOCATED AT SITE 42002 (26N LAT, 93.5W LONG) IN THE GULF OF MEXICO. AS A SHIP 2. ADAMO, L.C., "FINAL REPORT OF THE DEVELOPMENT STOOD BY, 43 HOURLY REPORTS FROM WRANSAC WERE TESTING, AND DELIVERY OF A PROTOTYPE RECEIVED AT SHORE ALONG WITH THE SPECTRAL WAVE GOES-REPORTING WAVE MEASUREMENT SYSTEM KNOWN DATA FROM THE DISCUS. THESE REPORTS LED TO THE AS WRANSAC." NDBO CONTRACT 03-7-038-720, FIRST HULL TRANSFER FUNCTION BASED ON SYSTEMATIC FEBRUARY 1978. (UNPUBLISHED) AT-SEA MEASUREMENTS. THIS CALIBRATION OPERATION HAS UNDOUBTEDLY IMPROVED THE QUALITY OF DATA FROM 3. STEELE, K. E., A. TRAMPUS, P. A. WOLFGRAM, NDBO'S 12-METER DISCUS HULLS. THE SHIP STOOD BY B. S. GRAHAM, "AN OPERATIONAL HIGH AT SITE 42002 TO PROTECT AGAINST THE LOSS OF RESOLUTION WAVE DATA ANALYZER SYSTEM FOR NDBO'S ONLY WRANSAC. BUT THIS USE OF WRANSAC BUOYS." PUBLISHEO IN THE PROCEEDINGS OF SHOWED THAT 43 REPORTS ARE NOT REALLY ADEQUATE TO OCEANS '76. (CORRECTED 9 DECEMBEP 1076) ACCOMPLISH A SATISFACTORY CALIBRATION, PARTICULARLY WHEN THE SEA STATE IS UNCHANGING OR 4. GODA, YOSHIMT, "ESTIMATION OF WAVE STATISTICS OUIET. FROM SPECTRAL INFORMATION." WAVES '74. AMER. SOC. OF CIVIL ENG., NY 1074. WRANSAC WAS LEFT UNATTENDED, TETHERED TO THE NOMAD-HULL BUOY AT SITE 42003 (26N LAT, P6W, LONG) 5. CALIFORNIA COASTAL ENGINEERING DATA NETWORK IN APRIL 1978. HIGH SEAS INTERFERED WITH THE MONTHLY SUMMARY REPORT NO. 25, CALIFORNIA SCHEDULED RECOVERY, RESULTING IN A SIX-WEEK TEST. DEPARTMENT OF NAVIGATION AND OCEAN A RELIABLE HULL TRANSFER FUNCTION FOR THE NOMAD AT DEVELOPMENT AND USA CORPS OF ENGINEERS. THIS SITE HAS BEEN PRODUCED WITH THE EXCELLENT DECEMBER 1977. (UNPUBLISHED) WRANSAC DATA ACQUIRED. 35 0.7 -,- ..- I .I ____ WRANSAC N. I rnh ----WAVERIDER UHF . / TRANSMITTER 2* I DIGITAL W 'I T ALLO ~~~~~~~~C.3 INTERFACE u'a. 0.3 $ILANO D WAVE DATA DIGITAL 0ESS.2 OEN mT BA~~~~~~~~~~~~~ASTTE L" BAT ~~~RECORDER IIA 0.1- ANALOG c o IGTA ROUBES G IN INTEGRATOR - 0.0 liii lii 00 12 00 12 00 12 00 12 00 WAVE ENERGY 27XII 28XII 29XII 30XII 31XII SPECTRA TIME (Z) Figure 4. TIME HISTORIES OF SPECTRAL DENSITY (0.13 Hz) Figure 1. WRANSAC SYSTEM BLOCK DIAGRAM 0.50 I I I I SPECTRALDENSITYLEVELSAREIN WRANSAC METERS2- HERTZ-1 90 - WAVERIDER H PRESSURE ARRAY 0.40- oo L- 0.01 I~~~~~~~~ a- I " Io0.30- IN' m 70 ,, - , 0.05 i-60- - us00 0.20- Dj 40 _ /'~'~�v.C 0.10- -------- 0.5*--- 0)~~~~~~~~~~~~~~~~~1 00 12 00 12 00 12 00 12 00 0.00 27XII 28X11 29XI 30XII 31XII 6z 12z 18z j ,z 12z 18z Bz 12z 18;z O 12z TIME (ZI 12118177 12119177 12/20177 12/21/77 Figure 2. TIME HISTORIES OF SIGNIFICANT HEIGHT Figure 5. CONTOURS OF EQUAL SPECTRAL DENSITY FROMWRANSAC 0.7 0.50 N - SPECTRAL DENSITY LEVELS ARE IN METERS2 - HERTZ-1 0I.6 - BUOY WAVE HT. WAVE PER. 0.40- 0.5 - i EB99 0.8M 6G1A I WROI 0.8M G.AS ii = ~~~~~~~~~~~~~~~~~~~~0.01 0.4 - >0.30- 0.05"-- 0.3 ~ ~ - LU 0.20- 02 I uj~~~~~~~~~ IAJ02.... I 'Z Ii~ - . \ C 0.1 N 01 0- 1__1,_1__ ~~~0.05-- .00 .050 .100 A SO .200 .250 .300 .350 .400 .450 .500 0.00 , FREQUENCY (Hz) 12Z 18Z 6Z 12z 12/31177 1/1178 1/21Z7 1/3/178 Figure 3. WAVE SPECTRAL PLOTS WROl-EB99 SOB �2050 Figure 6. CONTOURS OF EGUAL SPECTRAL DENSITY FROM WRANSAC 36 WAVE DIRECTION MEASUREMENT BY A SINGLE WAVE FOLLOWER BUOY W. I. Sternberger, L. R. LeBlanc, F. R. Middleton Department of ocean Engineering University of Rhode Island Narragansett Bay Campus Narragansett, Rhode Island 02882 Abstract complicated, and expensive spatially distributed This paper describes a system for measuring arrays of wave measurement devices. Using wave the direction of ocean wave propagation, as well as staffs on a structure requires spatial separation wave height and period. The system employs a con- on the order of the wave length of interest in ventional wave follower buoy, modified so as to order to obtain reliable results. provide compass, pitch and roll signals in addition to the vertical component of buoy acceleration. The buoy used in this study was a Model WF- The modifications are indicated in the paper, and 100 Wave Follower Buoy manufactured by Coastal the buoy was tested in two sea state conditions in Data Service, Inc., (licensed to Hydro Products, Narragansett Bay. The pitch, roll, compass, and San Diego), on loan to the University of Rhode heave acceleration signals were all processed on Island for purposes of the present development. the computer in the laboratory to provide a variety It was necessary for coastal Data service to mod- of results. One test condition was ideal in that ify their standard buoy, by the addition of tilt the wind and wave conditions were changing steadily and compass sensors, to make it possible to mea- * ~~throughout the data recording process. The second sure propagation direction. Some internal modifi- test was exactly opposite, with small waves and a cations were required to provide space for the consistent wind direction. The results presented necessary additional sensors. in the paper include the wave height power spectral density, a polar tilt histogram for the buoy motion, This paper will describe these modifications, and a family of polar plots of the average tilt explain the operation, and show at-sea test re- angle in each 50 sector of azimuth of the buoy sults from a particular wave field. The results pitch/roll. will include the usual power spectral density 1. Introduction ~plots and the directional pattern of the dominant wave frequency bands that were present at the time This paper is one of a series of papers on adpaeo h xeiet the subiject of the development of wave follower2.Ter buoys. The series commenced in 1975 with OTC paper 2424 on the mathematical analysis of wave data mak-Thwaeflorbuyhihassdints ing optimum use of fast computer techniques. Next, sTudy wav followed buoy whatich wacs use in tisvre was OTC paper 2597 in 1976 devoted to the problem studulm.I is thesge somtall pitactsing and rlinverte of sectrl copenstionof aceleatio dat inthe buoy which makes it possible to sense the pro- order to produce a true wave height power spectral paaindrcinoyh aiu oiatwv density (PSD). Finally, a preliminary set of re- f aaindreqcytionds ofThe priariou hdrodnminan feave sults, of wave direction measurements, was pre- tures of the WF-100 buoy assembly were given in sented in 1978 in OTC paper 3180. it is the pur - detail in the 1976 OTC paper mentioned earlier. pose of the present paper to elaborate upon the The most important feature for the present discu- measurement of the direction of propagation of wave sini h ivre edlm ofgrto mode(s) present in an open-sea field. wherion isthe centerted hoizntalum dofgragtisocae whearei the centter of therzna das semly.te En many important ocean operations, either on na h otmo h seby the open sea or near shore, the direction of prop- With this arrangement, when a wave passes by agation is often of more importance than the significant wave height, or the power spectrum. the free-floating buoy, the wave surface slope pre- This is clearly the case, for instance, for storm sents a pitch/roll driving moment to the buoy wavs srikng becha harbor channel, a break- assembly. Since the buoy is axisymmetric, pitch waves striking a beach, ~~and roll really apply to orthogonal tilt angles water or any sort of offshore structure. Up to relative to an arbitrary fore-and-aft direction of the present time, anyone needing wave incidence the buoy. The only sensors required in addition direction data has been forced to deploy complex, to the high sensitivity accelerometer are two good inclinoneters and a magnetic compass fixed to the buoy hull. The tilt angle magnitude and its mag- Illustrations at end of paper netic heading are correlated with the slope of a 37 wave and its direction of propagation. The commercial magnetic compass performed satisfactorily and it provided an analog output There is rarely interest in individual wave voltage produced by a low-torque pot connected to slope characteristics. The data user is much more the compass card. Finally, the output signals interested in wave direction statistics, because from the complete buoy system included four analog this is more relative to wave forces on structures, voltages -- buoy vertical acceleration, pitch, or slamming effects on beaches and breakwaters. roll, and magnetic heading of the lubber line. Since there is no apparent standard method for displaying directional spectra, the tilt data from To make use of these four analog signals for the experiment are displayed in different ways. this experiment they were delivered by a tether wire to the small surface vessel which carried In an ideal model instance, with sinusoidal the tape recorders and other support equipment. waves passing a buoy, the magnitude of the wave The four signals were each applied to an individ- slope on the front (before a crest) of a wave would ual FM magnetic tape channel on the same recorder be the same as on the back (after a crest) of a so the signals were synchronized. wave. In the real world, with non-symmetric waves, the magnitude of the slope is higher and the dura- Since the end result of this study was to be tion of the slope shcrter on the front of an in- wave propagation direction, there was no specific dividual wave. Another source of roll asymmetry need to generate an accurate wave height PSD. In- is the relatively steady roll force applied to the stead, the acceleration PSD was transformed by exposed portion of the buoy in an actual wind dividing through by W4which was an easy way to driven wave field. For these reasons it is not produce a crude PSD for wave height which was simple to predict the roll behavior of a given valid for most of the energetic wave frequency buoy on an open water wave situation. In fact,, band. the one safe prediction is that tilting in the direction of propagation on the front of a wave, This wave height PSD was then used as a guide as well as the reverse tilt on the back side of for processing the tilt data. The data will be the wave will occur with most small surface buoys. shown later, but for the present discussion, there are many processes that might be of some use with This sort of rocking behavior, enhanced some- the tilt data. For example, to a first approxima- what by the roll resonance of the buoy would at tion, the two tilt signals constitute a Cartesian first appear to present a 1800 ambiguity problem. form of displacement from the vertical. One can In most wave measurement operations, the local readily convert this to a polar form, with the geography, wind data, and other conditions make it magnitude corresponding to the tilt angle and the quite simple to eliminate the ambiguity if the polar angle giving the direction of the tilt rela- wave direction display fails to do so itself. tive to magnetic North. 3. Equipment and Processes 4. Results and Data As indicated earlier, the buoy used in this There were several ideas for ways to handle study is a Coastal Data Service Model WF-100 mod- the tilt and bearing data to get the most out of ified to incorporate two inclinometers and a mag- it. One was to generate a polar histogram (10 netic compass. The WF-100 is thoroughly described intervals) of events which corresponded to the and shown in photos in the 1976 OTC paper. The magnitude of the tilt exceeding some arbitrary hull shape has changed to more of an ellipsoid of threshold angle. The purpose of the threshold is revolution than the spherical shape shown in the to simply remove a great deal of the purely ran- paper. dom tilting motion that any surface buoy encoun- ters in surface waves. A sample polar histogram The inclinometers first used were commercial produced on the computer for the total 15-minute units that proved to be unsuitable for two rea- data set is shown in Figure 2 simply because it sons. They displayed resonant behavior and were is interesting. In this case, an event was de- sensitive to inclination at right angles to their fined as a tilt exceeding an arbitrary threshold normal response axes. For this reason, the two value, within a 10 azimuth bin and the radial orthogonal inclinometers were unavoidably cross- scale is arbitrary corresponding to the number of coupled, so that another solution had to be found. events. The solution came from a special inclinometer de- sign which provided two analog output signals For comparison's sake, a polar histogram for (DC voltage) almost linearly proportional to the the data corresponding to the calm sea conditions pitch and roll angles. (Figure 1). Pitch is de- is presented as Figure 3. Both figures are fined as a tilt in the direction of the arbitrarily plotted versus the same scales with thresholds of established lubber line on the buoy. 1l . Because of the apparent lack of activity in the calm sea case, it was decided that further Calibration of the tilt sensor is performed analysis be limited to the rough sea condition in the lab with the aid of a tilt jig. Static data. tilts of known magnitude, along each of the orth- ogonal planes, are correlated to the D.C. output What is interesting to note about the histo- of the unit under test. grams is the fact that more occurances of large tilt happen into the wind rather than with the wind. It is hypothesized that this occurs as a 38 result of the extended duration of the back of the by the eye, the wind was mainly out of the heading wave. When data sampling occurs at a fixed rate, 1650 and shifting back and forth to a considerable an extended tilt duration results in a proportion- extent. The sea was quite mixed also, and occas- ally larger incidence count. ional large troughs came at the vessel from unex- pected directions. The reason for picking this This discussion will be limited to one field position near Prudence was that we would have an operation in Narragansett Bay, under rough storm opportunity to expose the buoy to quite different conditions (see Figure 4). The locations marked sea conditions. 1, 2, and 3 correspond to the beginning of each of three continuous five-minute data recording inter- It is interesting to compare the "event" pat- vals. The test site was selected because it was tern of Figure 2 with the tilt angle magnitude obvious that the support vessel would drift roughly pattern of Figure 9. Both were produced from the Worth by Northwest in the wind, and drift would be same collection of data -- the total 15-minute out of the lee of Prudence island. This seemed run. In the former case, there is no averaging ideal because the combination of the steady waves across adjacent bins, which process would have plus those refracted by Prudence Island would be produced a smoother pattern. In such a case, re- changing as time progressed. In addition, the latively small rocking components of tilt produce wind was not steady from one direction and the the large number of counts in the 10 bins and the result was a highly confused sea, as far as corresponding large spikes in the polar pattern. observers were concerned. Now referring to Figure 9, the number of tilt All of the data was brought back to the com- events in a given polar interval (here 5 0 of azi- puter lab for processing, first for the PSD of wave muth) is given no weight at all. A single large height which is plotted in Figure 5. This is the buoy roll in a particular 5O sector will cause a PSD for the entire 15-minute data run and was pro- large average value of tilt angle or a large ra- duced by the tor4 transfer function mentioned dius in this figure. The asymmetric character of earlier. This rough PSD plot is sufficient to show surface waves is important here. Because the the maximum wave height power point at about 0.43 slope is greater on the front of the wave than it Hz. The corresponding period is about 2.32 seconds is on the back, the pattern of Figure 9 indicates which appeared consistent with visual observations a large average tilt tendency of the buoy in the at the time. The lower local peak in the PSD at direction of propagation. about 0.725 Hz. is known to be related to the 1-1/3 second natural heave frequency of the buoy. These differences discussed relative to the Compensation for this non-wave feature is tilt count and the tilt magnitude do not consti- thoroughly discussed in the 1976 paper, and it is tute any sort of problem with the data, but rather of minor importance in the present study. they display a useful range of information that can be extracted from the same data. Since the Figures 6 - 9 are polar plots of the average accelerometer output signal is always available, magnitude of the tilt angle within each 50 azimuth it is a trivial matter to detect whether a parti- interval after removing the purely random compon- cular tilt of the buoy is associated with the ent of tilt. This random component is the tilt front or the back of an individual wave crest by of the particular 50 interval which has the small- the polarity of the accelerometer signal. est average tilt. 6. Conclusions and Suggestions The question naturally arises as to which dom- inant wave bands are influencing this polar plot The application of a single wave follower the most. The processing was set up to perform a buoy to determine both the PSD of a wave field and numerical filtering process on the wave data at the direction of propagation of dominant wave the dominant frequency 0.425 Hz + 0.05 Hz at the modes was successfully demonstrated. No attempt 3 db down points. The objective was to isolate was made to exhaustively treat the data gathered this dominant mode to determine its own propagation in the experiment, but rather to show in simple direction. Figures 6, 7, and 8 are respectively ways that the data is good. The proposition that the 1st, 2nd, and 3rd five-minute segments, while a single buoy having the proper pitch-roll-heave Figure 9 is the entire 15-minute data interval. properties can be so effective in observing essential wave characteristics has been proven. 5. Interpretation of Data The buoy can actually be handled by two men Figure 4 shows the sites marked 1, 2, and 3 without the aid of lifting gear on the vessel. It corresponding to the three sequential five-minute can just as well be launched from a small dinghy data intervals. They are situated some 1500 and because of its light weight. The data in this ex- 2000 feet apart, indicating a wind (and tidal periment was transferred from the buoy to the re- current) drift on the order of five to seven feet corder on the support vessel by means of an elec- per second (3 knots) roughly in the direction of trical cable containing a wire for each of the the wind. Each time interval finds the vessel and four analog signals. There is some extra power buoy further out of the lee of Prudence Island and required to energize the compass and the two tilt in a different position relative to the waves be- sensors, but there is no reason to prohibit tele- ing refracted by the island. observers on the metering the data to shore to avoid the need for a vessel agreed that refracted waves could be seen support vessel. 39 As mentioned in the introduction, this paper board microprocessor is being designed which can is the third in a series dated back to the 1975 easily pick out such wave parameters as a signi- OTC meeting. The developmental process is continu- ficant wave height, wave period, etc., to be up- ing, and present efforts are maintaining the focus dated and delivered over the telemeter link. Such on buoy deployment during extreme sea events features as these can be selected, changed, im- (storms). A control clock has been used which can proved, etc., by simple program changes in the detect the onset of a storm and adjust the data microprocessor. Remote command of an unattended sample length and period to suit the user. An on- buoy, well off shore, is well within the realm of possibility in the near future. N Wind Fig. 3 Polar Histogram Test 2, 10 bins BATTERIES Fig. 1 Buoy Instrument Package �3 02 N~~~~~~~~~~~~~~~~ 01 is � Prudence NARRAGANS ETT BAY Wid wind ~~PICUT NaUutical Wile ISLANM Fig. 2 Polar Histogram Test 1, 1� bins. Fig. 4 Test Site 40 .. N Wind Find q Fig. 8 Average tilt in 5 bins Fig. 7 Av'erage tilt in 5 bins Wind 3rd 5-minute filtered Fig. 6 Average Tilt in 5 bins 2nd 5-minute filtered 1st 5-minute filtered N 3000 - 1000 1 300 - 10.... 4. 30- 10- in Frequency (Hz) 3- o Fig. 9 Average tilt in 5 bins 0.4 0.6 0.8 Total 15-minutes filtered Fig. 5 Wave Height PSD A DIGITAL SURVEY FATHOMETER�PROVIDING A CHART DISPLAY AND DIGITAL OUTPUT Mr. J. Russell Hogan Raytheon Company P. 0. Box 360 Portsmouth, Rhode Island 02871 Abstract The Digital Survey Fathometer� (DSF- 600) is a portable, dual frequency recording depth sounder with a permanent chart display, digital output, and features automa tic pha sepl changing of the chart display. Operating at frequencies of 40kHz and 200kHz (frequency switch selected), the DSF-600 provides high resolution, quality bathymetric profiles in water depths to 310 fathoms. By interconnect- ing it with position finding equipment and a suitable digital tape recorder or data logger, depth and position data is generated in perma- nent digital form for precision bathymetric survey operations. The DSF-600 system is offered with several options, allowing the The DSF-600, consisting of an analog chart user to choose the combination that best meets display and a visual digital display, is capable his surveying needs. of highly accurate measurement of water depths up to 310 fathoms or 620 meters, depending upon the model utilized. Raytheon manufactures two ver- sions of the DSF-600, one measuring the depth in feet or fathoms dependent upon a switch position 1. Introduction selection, or the other version measuring depths (j;N in meters. The DSF-600 Digital Survey Fathometer'- is a highly accurate, medium water recording The unit operates with a 40kHz or 200kHz depth sounder with a digital display, specific- transducer depending upon the desired depth and ally designed for precision bathymetric survey its resolution. The 200kHz transducer operation operations. Tide, draft, sound speed and depth is limited to 50 fathoms or 100 meters in depth. information can be stored in permanent digital form by interconnection with a digital tape Depth information is continuously updated recorder or data logger. A choice of high or at a maximum rate of 600 soundings/minute and low frequencies is available, as well as a com- permanently recorded on electrosensitive chart bination of the two. Also available are a paper. Pulses of acoustic energy are transmitted narrow beam transducer for increased bottom into the water and depth measurements are based definition, a DC inverter for portable operation upon the propagation time of the return echoes. and a remote digital depth display for ease of navigation by the boat operator. In order to visually and permanently store depth measurements, electrosensitive paper is The DSF-600 has several features that passed beneath an electrical pen that marks the make it unique in depth sounding equipment: paper only when electrical energy from the trans- ducer is present. The recorder's electrical * TVG Circuitry marking pen is time synchronized with the pulses * Bottom Acquisition Gating of transmitted acoustic energy and with the digi- tal processing circuitry so that return echoes * Automatic Phasing Mode printed on the paper correspond to a specific * Test Mode digitized depth measurement. 42 The system is fully automatic. The selected, is used to reset the depth counter cir- receiver incorporates time varying gain and very cuitry. The timing relationship between the KEY little manual gain and threshold adjustments are and T. signals are automatically compensated for required after initial set-up. With the phase depending upon the depth range. For example, if select switch in the automatic position, phase the depth being recorded is increasing beyond the selection is automatically changed with varying limit of the phase that it is printing within, depths. The depths are divided into 6 different the electronics automatically select the next range phases for purposes of recording upon the deeper phase, by generating the transmit pulse chart display. A range phase is divided into earlier with respect to the KEY pulse. See 60 foot or 60 fathom increments. This is delin- Figure 2, Key Pulse and Phase Timing. eated later in the paper. Automatic phasing, as it is called, is dis- The scale switch selects "~feet"r or cussed later in the paper. The transmit signal, "fathoms" for the depth scale, and also selects generated from "To", is applied to the trans~- the transmitted pulse width. A pulse width of mitter card which pulses the transducer causing 0.1 millisecond (ins) is used for 0-310 foot a burst of acoustic energy to be sent through depths, while a ins pulse width is used for the water. The transmit pulse will occur simul- 0-310 fathom range. Probing of deeper depths taneously with "To"' if both the tide and draft require more acoustical energy be transmitted, thumbwheels are set at zero. Any tide or draft thus a longer pulse is needed. settings other than zero will cause "To" and transmit to occur at different times. If this A detection mode switch selects either condition exists, the depth shown in the digital the first or peak return echo within an auto- display will indicate the depth beneath the matically set bottom acquisition gate. The transducers, plus any adjustments made by the first return (digital) signal is generated at tide or draft settings. See Figure 3, Example the beginning of the return echo (analog) sig- of the Tide/Draft Adjustment Feature. nal, whereas the peak return signal (digital) is generated when the return echo has reached Return signals are sensed by the transducer its maximum signal level. The threshold is amplified in the 200/40 TVC receiver, and passed adjusted for minimum signal strength required to both filter receiver cards. The frequency in for digital processing. use determines which filter card will amplify. The filtered signal is then amplified in the Post Other information, such as tide, draft Detection receiver which, effectively, is a low and speed of sound values can be entered digi- pass filter, rectified to a DC level, and passed tally via thumbwheel switches so that the depth to the digital electronics to be digitized and can be accurately computed for visual and chart processed. The received signal is also applied display. to a high voltage print amplifier which boosts the signal up to a level to energize the stylus Another feature is the shallow water and print through the chart paper. This event limit setting. Any depth measurement less than pulse samples a highly accurate depth counter and the value set for the shallow water limit, will stores the processed depth information into cause an audible alarm to be sounded and the memory latches. digital display to read zero. This depth information is then sent along the data bus to the digital depth display, com- 2. System Description puter I/O and remote display connectors. Other The sstemis snchrnize andinitateddigital inputs transmitted along the data bus from the key pulse signal. This signal is an are Draft, Tide and Speed of Sound information. once veryrevoltion All data is time multiplexed and presented in electrical pulse generated oneeeyrvlto arallel in BCD format on the bus. of the stylus belt. See Block Diagram, Figurep 1. The belt contains a slotted hole, and when 3 V ici it aligns with the keying light and phototran- 3 V ici sistor, the light signal, converted to current,Aunqefarempodisteievry is amplified and sent to the main electronicsAunqefarempodisteievry card age o reet te neessay cicuity toing gain (TVC) circuit. As the name implies, the card age o reet te neessay cicuity tooutput gain of receiver circuit is proportional initialie operaton of th system.to the elapsed time between the transmitted pulse The digital electronics generates all of and return echo signal. The farther the trans- the vtal imin sigals ssenial or popermitted signal travels in water and, consequently, sythem perorance timing signals essnitialto poes the return echo signal, the more the echo signal operation and time synlzhronizes the recorder's is attenuated. It is known as transmission loss. prining tylu to he sstemelecronis. AAs distance travelled is directly proportional to ptrit o ing styulse "to the isyse elronics. frmAh the elapsed time, the echo attenuation increases star~~~~~~~~tofl npus," isp ouefrm te with time. Thus, the TVC circuit compensates for KEY signal. This signal, which changes its tim he return echo signal strength. relation to KEY in accordance with the phase 43 The transmission lossI may be considered 5. Automatic Phasing Mode as a sum of a loss due to spreading and a loss due to attenuation. The spreading loss varies Another feature in the unit is the auto- with range according to the logarithm of range m atic phasing mode. This allows the DSP-600 to and is expressed in db per distance doubled. operate in a "hands-off" mode and does not re- Attenuation loss, on the other hand, varies quire operator assistance after initial equipment linearly with range and is expressed as db per turn-on and receive sensitivity adjustment. unit distance. The recording chart is divided into six Knowing the speed of sound in water, the different phases (depth ranges) for purposes of transmission loss in db can be expressed as a condensing the depth recordings, so that the function of time as opposed to that of range or depth range from 0 to 310 fathoms can be printed. distance. Thus, the time differential between The phase and corresponding depth ranges are tab- the transmitted signal and the return signal ulated below: corresponds to a specific transmission loss (db). This loss (db) is compensated for in the TVG cir- Phase , Depth cuitry by adding the appropriate gain to the input amplifier so that its output signal is at I 0 - 60 ft/fins the same level for any depth. For example, a 2 50 - 110 ft/fins transmitted sound signal takes 100 milliseconds 3 100 - 160 ft/fins (ns) round-trip to reach a depth of 240 feet 4 150 - 210 ft/fins (assume 4800 ft/sec speed of sound) and it is 5 200 - 260 ft/fins assumed the attenuation loss is 65db/kyd, the, 6 250 - 310 ft/fins total propagation loss is 54db. Therefore, the amplifier gain would be set for +54db so as to Note: (1) Each phase has 10 feet/fathoms obtain a 0db output signal. of overlap. (2) Phase and Range are automatically 4. Bottom Acquisition Cate printed on the record. An automatic bottom acquisition gate is Each phase has a maximum depth capability. generated from the depth information. This When this maximum value is approached, the depth gate is used to lock around and track the depth. value will be printing near the bottom portion Any return signals that fall outside this gate of the chart paper. If this maximum is exceeded, are ignored by the digitizer. If three consec- the system will change phase and the depth value utive transmit signals occur and the depth will print near the upper of the chart. The max- information received falls outside the auto imum (analog) depth recording of the system is bottom acquisition gate, the system enters an 310 fathoms. See Figure 4, Chart Recording. "'alarm" mode which fully opens the gate, allow- ing for any analog return signal to be processed. As previously discussed in the System The alarm warning buzzer will sound, the selected Description, the KEY pulse signal synchronizes scale indicator will flash, and the digital depth the printing stylus to the electronics. As the display will read zero. depth range increases or decreases by 60 feet or fathoms, depending upon the position of the scale During this alarm mode, a bottom acqui- switch, the phases will change. It is accomn- sition gate will be generated only after three plished by generating the transmitting phase consecutive return signals have been detected pulse early (2Oins which corresponds to 50 feet and digitized. The width of this gate is deter- in depth) with respect to the next lower adjacent mined by the actual depth measurement of the phase number pulse. See Figure 5, Automatic third return signal and the scale and phase Phase Timing. selected. Either the first or peak return echo that falls within this gate can be processed. Phase I pulse is delayed from the KEY pulse depending upon the setting of a chart zero poten- A bottom reference mark is printed on the tiometer, located on the front panel. In order chart at a fixed distance above the bottom. It to compensate for stylus wear, this potentiometer is printed at the lead edge of the tracker gate. enables the operator to adjust the calibration The distance mark varies as a function of depth. marks to agree with the preprinted scale marks As the depth increases, the tracker gate becomes on the chart paper. wider proportionally. The KEY pulse is generated approximately The phasing can be accomplished manually every lO0ms (feet-scale position) or every 600ins or automatically. The manual selection of Phase (fathom-scale position). For this explanation, I through 6 can be thought of as a back-up oper- "feet" position will be used. ation to the automatic phase mode. 44 The DSF-600 would be limited to four (4) Tests by Bedford Institute were conducted phases without special circuitry, as Phase 5 on board one of the "Baff in's" launches. Primary transmit pulse would occur close to the previous interest was in reliable operation at 40kHz in KEY pulse and sometimes coincident due to the deep water with vessel speeds of up to 12 knots. asynchronous condition between KEY and Phase pulses. Phase 5 and Phase 6 pulses are made The Canadian Coast Guard's test platform coincident with Phase I and 2, respectively so was a 55-foot shallow draft survey boat, and that the depth range could be expanded. In interest is in shallow water channel depth veri these two phases (5 and 6), the next KEY pulse fication at 200kHz. Very fast changes in depth is ignored or gated out of the circuits so that are encountered at the channel's edge, and a the echo return from 310 foot depth (correspond- prime requirement is that the digitizer should ing to l3Oms in time from transmit pulse to track these slopes accurately. Results of the return echo) can be printed upon the chart dis- demonstration showed that the DSF's performance play in proper depth position. Note that the was superior to that of other digital sounders Range depth is automatically printed on the under test. record. 9. Summary and Conclusions 6. Display Only Mode The DSF-600 has many unique features that In the event that the water depths ex- other similar equipments do not offer: ceed the value of 310 fathoms, it becomes necessary to operate the unit in the DISPLAY * Automatic Phasing ONLY mode. This mode is selected when the S Bottom Acquisition Gating DSF-600 is to be used just as a digitizer or when depth information exceeds 310 fathoms. S TVG Circuitry The unit has satisfactorily digitized depths S Self Test Mode in excess of 600 fathoms. 0 Reliable operation with a MEBF 7. Test Mode which can be in excess of 1200 hours. The test mode allows the operator to0Cosrcefreninmta self-test the operation of the digitizer. In 0Construions for oeanviroinmentals this made, the operator checks the printed condion opn ocaun-ginhess(qulifedt calibration marks on the chart paper. Correct a nd openlaunck, vibralified teop operation of all controls can also be verified. MLSecsokviratureon, temidiy) As the phase switch is changed from one S Provisions made for bulkhead, rack phase to another (i.e., Phase I to Phase 2 to or table mounting. Phase 3, etc.), preselected depths will be dis-5Hosdion-eclghwgt played and printed upon the recorder for each aluminum cabinet. of the six phases. The automatic phasing has a testing pro- cedure also. The AUTO TEST DEPTH potentiometer is turned slowly and depths from o to 310 Reference fathoms (feet) can be tested and printed upon 1 rcRbr . Picpe fUdrae the recorder. 1 rcRbr . Picpe fUdrae Sound", pp. 94. S. Performance Capability The DSF-600 has been building a reputa- tion in the hydrographic survey community for performance. Proof was given during field demonstrations, with conditions established and performance observed by potential users of the equipment. Those who have observed the DSF in oper- ation include representatives of the Bedford Institute of Oceanography, the Canadian Coast Guard, the U.S. Army Corps of Engineers, Texas Instruments, Inc., the N.O.A.A., and the U.S. Naval Oceanographic Of fice. 45 FIGURE 1. DSF-600 BLOCK DIAGRAM 200/40 POST- PRINT TRANSMITTER TVGDETECTION L X \ TR~gH S TYLUS DISPLAY o PAPER v I ENCODER , C- TAKE-UP MOTOR LO 200/40 kHz ENCODER ~TRANSDUCER CHART MOTOR TRANSDUCER I ' 11 ~ ~DRIVE KEY SIGNAL FIGURE 2. KEY PULSE AND PHASE TIMING KEY L- [ PHASE1 PHASE 2 J]|~~~~~~ | ~PHASE 3 50 50 FIGURE 3. EXAMPLE OF TIDE/DRAFT ADJUSTMENT FEATURE WATER SURFACE TRANSDUCER ADJUSTMENT MADE 10 FOR TIDE/DRAFT 10 DEPTH BENEATH TRANSDUCER IN THIS EXAMPLE THE "DIGITAL DEPTH DISPLAY WILL READ 12." BOTTOM 46 FIGURE 4. CHART RECORDING CHAQT 669224 DI PTH IN FEMT/FATHOMS DSF-00 O f R 2~14 RAYTHEON: DSF-6GO FREQUENCY: 40k1-k SCALE: FATHOMS RANGE PHASES THREE AND FOUR: LOCATION: CONTINENTAL SHELF *BOTTOMATCFERENCEINGRK PRINTED 5 F A T HOMS ABOVE FATHO~~~~~~MS OVRLPI EAC~~~~~~~~H PHASE3 ~~- iPHASE AN4CL FIGUE 5. A U TOATI PHASE TMN 47~ L~K- A DETERMINISTIC AND RANDOM ERROR MODEL FOR A MULTIBEAMI HYDROGRAPHIC SONAR SYSTEM Richard D. Angelari, PhD National Oceanic and Atmospheric Administration National Ocean Survey C6Sl Test and Evaluation Laboratory Sensor Test Branch Rockville, Maryland 20852 Abstract this study. The expanding and increasing affluent world population has brought about a growing need for commodities and energy on a global scale. Two major marine activities affected by this are shipping and mining. The shipping industry, because of its up to 30 meter draft superships, and the mining industry (partic- ularly the hydrocarbon sector), both require timely and accurate bathymetric charts for safe and economical operation. To meet thes~ needs the Bathymetric Swath Survey System (BS) is BEAMS being developed. The BS3 is a 21 beam mapping sonar designed to accurately and efficiently portray the seabed directly beneath and to the port and starboard side of an underway hydro- graphic ship. This paper examines the effects of 11STARBOARD BEAMS four major inherent error sources on system accuracy. These sources are uncompensated ship roll, sound velocity errors, pulse stretching bm and noise. Figure 1. Bathymetric Swath Survey System (BS3) 1. Introduction 2. Error Sources The BS3 is a 21-beam sonar system designed Error sources will be designated as follows: to measure a swath of ocean bathymetry during hydrographic missions. Figure 1 depicts the Type I error (uncompensated roll error) is swath geometry. Basically, the system operates that portion of ship roll not sensed by the ship in two modes, the transmit mode and the receive roll sensor (due to its inherent limitations) and mode. During one transmit mode, the portside therefore uncompensated by the BS3 processing bottom is insonified by an acoustic pulse, system. This error will be symbolized by A(p . followed by a portside receive mode. The receive mode utilizes 11 contiguously directed 5� beams, Type II error (sound speed error) is error one vertical beam, and 10 portside beams. The resulting from less than perfect estimates of portside operation therefore consists of a total sound speed in water. This error is expressed as swath coverage of SSO. These operating modes a ratio of the measured velocity to true are then repeated for the starboard side. Depth velocity m C* - C measured C true sensed by each beam is determined by measuring the go-return time of the acoustic pulses and Tye III error (pulse stretching error) transforming this time interval to slant ranges to the bottom. These slant ranges in turn are It will be demonstrated in the discussion to then translated to depths by a simple trigo- follow, although the return echo is time delayed nometric operation. Errors due to ship's motion and attenuated, it is not a replica of the and environmental factors are for the most part idealized pulse, i.e., having infinite slope at its sensed by the BS3 signal processing subsystems leading and trailing edges. It is, in fact, and corrected for, however, limitations of the attenuated and in addition, it is rounded and sensing elements, electronics, and noise may con- stretched. This distortion results in delaying tribute errors. The establishment of the error the indicated time of arrival of the return echo bounds due to these sources is the objective of and is therefore an error source. 48 U.S. Government work not protected by U.S. copyright. Type IV error (noise related error) 3. Mathematical Error Model error is similar in effect to pulse stretching error since it causes premature indication (in- stead of delayed as for pulse stretching) of the Type I, II, III, IV Error Model - Depth error can pulse arrival time and is therefore considered be expressed as a % quantity as follows: an error of the same rank. Figure 2a shows a ship's hull together with the necessary geometry True Depth- Measured Depth x 100% to establish the error model. Figure 2b depicts True Depth the pulse stretching effect. The inner beam (1) (vertical beam) receives a relatively narrow pulse where e% is the error in % relative to the with relatively straight edges. The outer beam true depth. return pulse is much more rounded and stretched. The slope of the leading edge decreases as the To establish the contribution to depth error due pulse width increases. This becomes apparent by to error sources, a functional relationship between the following argument. Consider a beam pointing these sources and depth error in % is desired, at an angle Op degrees from the vertical as in- that is: dicated. eBW is the beam width angle. The go= F0AP'C*14 (2) center of the beam strikes the bottom at D, the inner edge at A, the outer edge at C. AB re- presents a wave front. The inner edge arrives This relationship is: (Angelari [1]) at the bottom first whereas the outer edge does so at some time later equal to the time of * COSOp + xOewcosdDTAN[eALate(,/ propagation along XP. Thus points A&B on the LiorII )rte1 COSr6�AO- eRA]l incoming wave front which traveled together at IC (rI launch time arrive back at the receiver sep- arated from each other by the propagation time to traverse distance 2X. The pulse has there- Equation (3) is the mathematical error model for fore been stretched. This stretching decreases the BS3 system having Type I, II, III, and IV. the edge slopes of the return pulse. error sources. 4. Error Analysis Figure 2a Deterministic versus Random Error Sources: Central to the analysis are the concepts of AIlI~NG deterministic phenomena and random phenomena. ANGLED Deterministic phenomena are those that can be de- scribed by an explicit mathematical relationship. Random phenomena are those that can be described s"~ *~lAIDonly in terms of probability statements and statistical averages. Figure 2b _,,~rESHGLD L EVEL The error model, Equation (3), may represent ~-~ T}. I~rtOAP~weither deterministic or random system error de- pending on the nature of the error sources. If the error sources AOp C*,andA are deterministic then this equation is the BS3 Deterministic Error Since the detection threshold level is fixed, Model (DEM). If the sources are random then this pulse detection time will vary as the pulse equation represents the Random Error Model (REM). stretches. As a first approximation, it will be Both types may exist simultaneously, for this assumed that the error is proportional to the reason both will be considered in this study. stretched pulse (Tsp), i.e., detection time * error is ~ Tsp where 4 is a positive fractional Estimates of the expected ranges of error sources number less tan I and is hereon designated as are: IA an and is are: -~~amp +1' to +3' the timing error coefficient. ~ can also take p - - on negative values; this follows since the = +0.1 to +0.3 stretched pulse is in a noisy environment and premature triggering can occur if noise peaks - -- exceed the threshold level for any appreciable time prior to the true arrival time of the re- turn echo. These will be denoted as the specified error sources (SES) and will be designated as follows: AOp=+l', $=+0.l and C*=l+.01 "lo error" SES �X represents the distance BC. 49 Dp =+3�, A =+0.3 and C* =+1.02 "hi error" SES time the excursion is in the negative regions (3rd and 4th quadrants) meaning that the more The SES will first be considered as deterministic serious type error is as probable as the less errors and will be used in the DEM analysis to serious (positive regions) error. Point q is the follow. For the REM analysis "lo error" (i,e., quiescent operating point and represents a bias due AOp =+1�, A =+0.1 and C* =1.01) will be specified to error in estimating sound velocity with no as the statistical descriptors and represent the other sources present. In this case, since the 1 a values for the uncompensated roll, sound estimate of sound speed was correct, the static velocity, pulse stretching and noise error system bias error is zero.* sources. 5. Deterministic Error Model Ott Plots of the error model (eq 3) for ranges of , X error source magnitudes will be informative. To decrease the number of plots the outer beam - 50 and the inner beam 9q,=00 only will be A - plotted and only the maximum roll specified for -' ~u5 RJSOL OL Mul [DEGREES the BS3 (20�) will be treated. Thus a total of four pointing angles will be considered. These are shown in Table I. ,- Table I POINTING ANGLES EVALUATED IN THIS STUDY 1. Op = 0� vertical beam -20 2. Gp = 20� (vertical beam + 20� roll) 3. Op = 50� outer beam 4. Gp = 70� (outer beam + 20� roll) -a-.---------- - Plots of the BS3 system model error due to the Figure 3 0p=70�, C*=1l.00 SES are shown in Figures 3 through 8, 6% is plotted along the ordinate, LOp along the abscissa, A asa parameter and for fixed values Case 2. The Vertical Beam with 0� Roll, 0p=0, of C*. The rhomboids in these figures delimit C*=l (Figure 4) the error bounds for "lo error" (mnop) and "hi error" (rstu) sources. = I 6. Discussion of Error Curves e Case 1. The 50� beam with + 20� roll, (0p = 70�.) ' C* = 1.00 (Figure 3) Figure 3 is the error diagram for the outermost beam (0w,=500) with a positive 20� roll, i.e., a p _, Op =70�. In addition, it is assumed here that -r-- r -2 3 2 u 4-5 the sound speed has been correctly estimated, i.e. UNCOMPENSATED BRGIANGIE ERMOP C* =1.00. The rhomboid mnop represents the -s operating error region for "lo error", i.e., for AOp =+1� and ; +0.1. From this diagram it is concluded that the measured depth error will not be greater than between 6.7% positive (point n) or 7.�% negative (point p) for "lo" excursions. - Thus a true depth of 100 feet would be indicated as being anywhere between 93.3 feet and 107.8 feet. For "hi error," Op=�3�, 4=�0.3, the range of error is between 17.9% (s) and approx- Figure 4 Op=0, C*=l.00 imately -30% (u). For this case the true depth would be indicated as being anywhere between 82.1 feet to 130 feet. Clearly the more serious errors are the ones having the greatest negative magni- tudes since hazards at the true depth could be indicated to be deeper than they are, i.e., 100 * It is not meant here that the BS3 error is zero feet true depth may be indicated to be 130 feet when C*=1 but rather the MODEL bias error is zero. deep. The diagram also shows that for 50% of the It is expected that an overall system bias error ~~-4~~~~~~~~ ~exists and will show itself when the BS3 field .4is the fixed array beam angle, Op is the tests are completed. tests are completed. pointing angle 50 The rhomboid area has diminished greatly indi- 33 cating a greatly reduced error mode of operation. This is to be expected since the return echo in the vertical beam experiences the least pulse =5 stretching (see Figure 2b). Note that for this case the system error is almost independent of the w ,0 magnitude of 4. The maximum error, (for "hi error") is less than +0.5%. The rhomboid de- generates to a point tthe q point) for 4=0 and L0p=o and since C*=l (for this case) C"-% is zero, Ae p -- i.e., an errorless system. _5 _O � :P:_P-UGRS.. 3-U-M-] Case 3. The vertical beam with 20� roll i.e., Op= R 0�*-l-.00 (Figure 5). \ The rhomboid area has increased such that "lo -10 error" and "hi error" are approximately +1% and +2% respectively. - - a, !E Ie='S~ Figure 6 op=5b0�, &-=.62 t I z 20 5 --A i -y -1A:0 5 ,,,, a ~~~~~~~~~~~~~~~~~~~~~~~~~,- ,.s6ep -IS~~~~~~~~~~~~~~~~~~~~Ap . .. -5 -4 -3" -2 '"i"'2 ", " 41"5 ; a -~- - m q: UNCP'MPENSATED:ROLLtAGLEJDEGREESI-Ap --5.5 ~~~~~~~~~~~~~-232~~~~~-5 Figure 5 Op=20�, C*=l.00 Case 4. The 50� beam with 0� roll i.e.. QD=50? -15 C*=1.02 (Figure 6). -23 This figure depicts system error with 0� roll and an overestimation of sound speed of +2%. The Figure 7a Op=0�, C*=1.02 quiescent point has shifted negatively by 2% this is a systematic error. The "lo" rhomboid is now A passing note is made to indicate the role played almost entirely in the negative quadrants, this by the beam-width angle OBWAs can be verified by is the most undesirable error since this would reference to equation (3), a reduction of beam- indicate a potential hazard to be deeper than its width from the present 5� to 1� will reduce sensi- true depth. tivity to A (pulse stretching and noise effect). Case 4. The vertical beam with 0 roll. C*=1.02 This in effect would have a marked improvement on ~~~~~~(Figure 7). ~system performance. Since decreased beam-width requires larger transducer elements, among other The rhomboid are significantly decreased, the factors, a trade-off study would be necessary quiescent point is again at -2%. Uncompensated should such an improvement be sought. The high- rolls of magnitude up to 3� contribute little lighted rhomboid (Figure 8) represents "lo error" error, here again this is due to the near neg- region for the outer beam with 200 roll (p=700) ligible effect on the vertical beam of pulse C*= 00 and OBW =1. This figure represents the stretching as previously discussed (Fig. 2b). same eagnitude error sources of figure 3 but with Figure 7b shows the biasing effect in the vertical a 5 to 1 reduction in beam-width (i.e. from 5 to due to error in estimating sound speed by 10). +5% (C*=I+.05). The case for 0% error (C*=l.00) is included for emphasis. 51 ~~~1D'5~~~~~~: t15 1 0 P I - 1 ' 2--a45 -2- UNCOMPENSAED ROLL ANGLE [DEGREESI Aep n o I -( -? P a 1-5 -B Figure 9 Bar-line graph summarizing "lo","hi" error ranges . The dot represents the quiescent * operating point (q point) -Is V. Random Error Model Pigure3 7b oP~ =0%, C*-l.00, 1.05 and f.95 For this model the error sourcesa3p, A and C* will be considered random variables normally distri- buted with zero mean and standard deviations equal to the SES "lo" error values. Formally stated this ~230~~~ ~becomes: v atp= N[0, al= N[0, o>2) = N 0,~ � N[00, (0.1) ] o * g C*= N I, pcjj N[O, (1)2] (where a tilda (-) indicates a random variable) Equation 3 may now be replaced by the corresponding Mi Aep9--- prandom error model. ( N CD3MPNSEO 2ROLL ANGLOERREE( - COSP (eBwCOSEpTAN[e9n-e+,Bw/a X100 XlO 1 cos[eep] cos0e e - rBw/21 (4) The probability density function (pdf) fg) cor- responding to 4, once fdund will completely des- cribe the REX.' Unfortunately, the density- function may be very difficult or even impossible to determine directly from Equation 4. Equation 4 _-23 \\\\\ 'will be expanded about the zero point as a truncated Taylor series, a more tractable expression results, for small error excursions this becomes Figure 8 Effect of decreasing beamwidthl from % % % 5� to 1�0 for the Outermost beam with -p 7� (ep, cp, c* = pAOp + q$ + rc* (5) C* =1.00 (compare with Figure3 ) where now Op, 4 and C* are realizations of the rvs and represent small excursions and Figure 9 is a graphical display (in bar-line p (5a) graph form) summarizing the salient information of a this study. This figure displays the following q = (5b) information for Op=70�, 50�, 20�, and 0� for r = ac C*=1.02, 1.01, 1.0, 0.99, and 0.98. ac* (5c) 1. The location of the operating point q(dot) are partial derivatives and for small excursions 2. The error in % of true depth due to a varia- tion in 4 and Op, for "lo error" (bar ends) 3. The error in % of true depth to a variation in 4 and 9p, for "hi error" (line ends) 4. The operating area in the upper (clear areas) and lower (shaded areas) regions. @ the tilda (-) will be omitted it is understood we are dealing with random variables (rvs) 52 are essentially constant. These partials are de- 64a rived directly from Equation 4. A tabulation of . .9..a . l a l a a ha 5 IX a p, q, and r for each of the 11 pointing angles is .4 ,l Hiif -i i i' I given in Table II. 3i;11 llit ii Table II ii Op� p q r ., 0 0 -0.38 -1.00 5 -0.15 -1.15 -1.00 10 -0.31 -1.92 -1.00 i iii . . ._.. 15 -0.47 -2.72 -1.00 1 = = m ..e- ;= r,_ ;= 20 -0.64 -3.56 -1.00 f5tip . iil .fiil.d ifjiifIf 25 -0.81 -4.46 -1.00 * .-- 30 -1.01 -5.43 -1.00, a n 35 -1.22 -6.50 -1.00 40 -1.46 -7.72 -1.00 Figure10 Plot of the REMi on probability paper 45 -1.75 -9.13 -1.00 50 -2.08 -10.82 -1.00 8. Conclusions and Recommendations Since a linear combination of Gaussian terms is Two error models for the BS3 have been developed. itself Gaussian, Either one or both may apply depending on whether the error sources are primarily deterministic or (6) random or both. If the sources are deterministic s(%) = N (PE(%) ,2(%)] then the accuracy and precision of the BSs may be markedly improved through calibration and correct- where De(%) = 0(7a) tive processes. If the sources are random then o2 2 42r 7b) only statistical bounds may be defined for the 2e(%) = p2 qtsp2 + q2 j2 + r2o2c*7 system i.e., accuracy may be specified at any The two descriptive parametersVe(%) and a2t (%) desired confidence level. completely define the RE4I. Thus, probability omplef the BS3 erpror being between + is: It should be borne in mind that, although four of the BS~ error being between +6 is: P [-(%) < (%) < -(%)] +d(%) error sources considered in this study are major fs Caexp(- E2(%)E de(%) sources, other sources are present and can only i i 2 serve to further degenerate system performance. -2 d)of 202e5%) These other sources are believed to be minor so that their contributions are believed to be small. Integral (8) is a well tabulated function and Since the specified error sources (SES) magnitudes may readily be evaluated, However, a more con- given in Section 4 are estimates, laboratory/field venient method is to evaluate it graphically measurement experiments designed to establish the using normal probability paper. As an example ranges/statistics of the error sources are consider the error predicted by the REM when the recommended. outermost beam angle is pointing at 50�, i.e. Op=500. From Table II: p = -2.08 q = -10.82 r = -1.00 Suhstituting into Equation 7b s(%) = (-2.08)21 1 ]2 + (-10.82)2 [0.1 ]2 + (_1)2 [ 1 Acknowledgement (the quantities in the square 1 ] brackets are the"lO"SES) =-2(%) = 6.59 The author would like to thank Mr. Richard Ribe from which of NOAA's Test and Evaluation Laboratory for or (%) = 2.75 I evaluating the partia]sp, q, and r (eq 5a, 5b, 5c) and for generating Table II. also II li (%) - 0 References Figure10 is a plot of the REM (equation 4), it is seen that the 50' beam when pointint at 500 may be in error by up to +5.2% of depth at 95% CL.' [l]Angelari, Richard D. (National Ocean Survey, Similarly, the vertical beam, Op = 0", may be in National Oceanic and Atmospheric Administration, error by +1.9% of the true depth at the 95% CL. $as. Department of Commerce, Rockville M. D.)1978 Op = 10�, 200, and 30� plots are also shown. "System Error Model of the Bathymetric Swath Intermediate pointing angle error plots may be Survey System (BS3)" 77pp (unpublished manuscript) established by calculating e % using Equation 7b and values of p, q, and r for the desired Op. The plots also indicate the lo error values. 53 PORTABLE ACOUSTIC TRACKING SYSTEM (PATS) Robert W. Gill William C. Gardner Applied Physics Laboratory Naval Facilities Engineering Command University of Washington Chesapeake Division 1013 N.E. 40th Street, Seattle, WA 98105 Washington Navy Yard, Washington, D.C. 20390 Abstract (j) to include a radio frequency link which A portble aoustictrackng sysem ha beenwould permit monitoring, plotting, and A portble aoustictrackng sysem ha beencontrolling the system at points up to S developed for the Navy to track underwater con- mi from the tracked object struction teams in water depths to 300 ft. The tracking system can be deployed and operated by the (k) to include nonmagnetic, diver-carried divers and is intended for use in remote areas and acoustic transponders. work. A self-calibrating system provides a geo- graphically justified track for up to eight divers. Operation Diver position coordinates are computed from acous- tic measurements at periodic intervals and dis- The deployment and utilization of the PATS, as played for control purposes. The positions are conceived, will generally proceed as follows. The also recorded for future reference. PATS, along with the members of the underwater construction team, will be deployed to the work site where the system will be removed from portable shipping containers. The PATS receiving/plotting System Requirements modules will be set up either on the beach or aboard a diver-support platform.' The portable The basic requirements governing the design tracking array, complete with cable and surface and construction of PATS were that the system be buoy, will be loaded aboard a small, maneuverable portable and readily usable at remote field sites boat such as a Zodiac or Boston whaler, and taken by Navy underwater construction teams. Specifi- to the deployment site. The array will then be cally, PATS was designed: lowered to the sea floor by its own electromechan- ical umbilical cable. After the array has been (a) to be deployable from boats as small as implanted on the sea floor and a surface buoy with 20 ft in length a radio frequency transmitter has been deployed, (b) to be capable of continuously tracking ~~~~~~~~~~~~~~~~~~~the small boat, with a pinger attached, will trav- (b) o becapale o cotinuuslytracingerse over the array while being tracked from shore up to eight targets simultaneously by either a radio frequency positioning system or theodolites. Once the array's center and orienta- (c) to contain no single subassembly with a tion have been established with respect to geodetic weight greater than 60 lb bench marks or major landmarks, the divers can begin work. Each diver will carry a synchronous (d) to be suitable for operation in water pinger attached to a backpack (see Figure 1). depths of up to 200 ft Suppose that the divers are searching for a (e) to be suitable for use in a relatively cable, pipeline, or other object. They will be noisy nearshore environment continuously tracked, and the track will be plotted in X-Y coordinates as well as depth. Once the (f) to have a range of approximately I mi object has been located, the divers will toggle a "mark" switch on the pinger assembly which will (g) to have a circular probable error of cause the plotter to annotate the recording. As 5 ft the divers follow the contour of the object, they will continue to toggle the "mark" switch at each (h) to include plotting equipment to provide significant feature. In this way, the path or a permanent record of target track contour of the object will be mapped. The scale on the X-Y plotter can be selected to match the scale (i) to be suitable for immersion periods of of charts or maps of the area so that the completed 60 days on station recording can be overlayed directly on the chart. 54 System Description The PATS system allows the positions of up to eight divers to be accurately determined at inter- vals as short as I s. The Cartesian X,Y,Z coordi- nates are listed on a line printer, and the posi- tions are displayed on an analog plotter and on a liquid-crystal display; this display can also show the positions in polar coordinates (range and bearing) for guiding the divers within the tracking volume. The system employs synchronous tracking. In such a system, a tracking projector emits a pulse at a precisely known time. If the sound velocity in the medium is known, measurement of the acoustic transit time to receiving transducers in a tracking array allows an accurate determination of range. These transit times will be referred to as timing numbers. If the locations of the receiving trans- ducers are known in some coordinate system, the Figure 1. Diver with backpack unit. position of the tracking projector can be deter- mined in that coordinate system. This requires measurement of the time (range) from the tracking projector to three or more separated transducers. In situations where a buried cable is being track- Three transducers can determine the position except ed, the divers will use PATS in conjunction with for the image ambiguity (i.e., the location above underwater magnetometers. PATS will contain or below the plane of the transducers); addition of chargeable batteries which can be changed on site a fourth transducer resolves that ambiguity. to accommodate lengthy operations. Upon completion of the underwater task, two people aboard a small The PATS system accomplishes its objectives boat will retrieve the surface buoy and array for through the interaction of three major subsystems replacement in the transportable shipping containers. (see Figure 2): DATA RECEIVER AND COMPUTER X-Y MAGNETIC ~~~DIVER DISPLAY PLOTTER RECGNETEC ELECTRONICS PAN~~~~~~~~~~ECODEL PACKAGE PROCESSOR - BUOY RECEIVER PROCESSOR TRANSMITTER UNIT (RPT) DIVER XDCR ~~ELECTRONICS -. J PACKAGE Figure 2. Block diagrom of' underwater electronics. (a) a tracking array containing a micro- the diver unit is 190 dB re I liPa at I m. At 2000 processor and acoustic and radio telem- yd, for example, the approximate transmission etry systems losses would be 11 dB for absorption and 66 dB for spreading; the signal level at the tracking trains- (b) a receiver-processor-transmitter (RPT) ducer would thus be 113 dB re I iaPa. This means unit containing synchronizing, display, that the level of the background noise within the and computation systems bandwidth of the receiver must be less than 87 dB re I piPa. Some representative ambient source (c) diver units containing synchronous acous- levels within the 2-kHz bandwidth of the tracking tic transmitters. transducers are:1 Each of these systems is operated by detachable, Beaufort wind-force 4........30 dB rechargeable batteries. Time for array deployment Sea state 3.............41 dB or retrieval is about I h. Battery power is suf- Mode-rate precipitation .......76 dB ficient for about 24 h of continuous operation of the RPT unit and the tracking array, and 8-h opera- A "moderately moving" small boat emits about tion of the diver unit. An external battery can be 67 dB; therefore, the support boat can be as close used to extend the operation of the tracking array. as 300 yd from the tracking transducer, even when the divers are at maximum range. The underwater tracking portion of PATS (Fig- ure 3) consists of two bottom reference transducers It was assumed that the divers would operate, and a floating tracking array of four buoyant, sub- at most, I h after synchronization; therefore the surface transducers deployed in the form of a oscillator stability requirement is �3 parts in square with 250-ft sides. The two transducers on 1 07. The azimuthal error depends on the frequency opposite corners of the square are about 40 ft and the baseline of the tracking transducers. At higher than the others, forming four "tilted" 31.25 kHz, a baseline of, say, 167 ft would result triangles. This configuration allows the resolu- in an error of 5 ft at 5000 ft from the transducer. tion of the image ambiguity. The two reference Because of acoustic considerations, it was decided transducers on the bottom are about 750 ft apart, to use buoyant transducers located at mid-depth. one on each side of the square. The use of buoyant transducers necessitated the continuously updated calibration survey. The PATS tracking cycle consists of 10 sub- cycles, each approximately I s (1.048576) in Diver Unit length. During the first subcycle, 250-kl-z "sur- vey" pulses are transmitted, in turn, from each To allow divers to swim over cables with an tracking transducer. In this way, the shifting operational metal detector (such as the Mk 14 positions of the buoyant tracking transducers are ordnance locator), the diver unit had to be as updated every 10 s. The next eight periods are nonmagnetic as possible. Therefore, plastic reserved for transmissions from the diver units. A encapsulated transistors, a plastic-encased oscil- different diver unit may transmit in each period, lator, air-core transformers, and gel-cell batter- the same unit in all periods, or any other combina- ies were used. The unit weighs 12 lb in air, and tion; the diver units must be preset to these is 2 lb negatively buoyant, 26 in. long, and 4 in. conditions before water entry. The last period in diameter. Although the "mark" switch requires contains sequential transmissions from both ground pressing, it is on a cord at the diver's wrist so reference transducers. that the diver is essentially hands free. The diver units transmit at 31.25 k~z. The The battery provides about 8 h of operation transmissions consist of a 1.024-ms pulse for and is detachable to allow charging. The diver tracking purposes followed by a pulse that is unit must be connected to the RPT unit to be delayed in proportion to the output of a depth turned on or off or synchronized. To protect the transducer. This allows an accurate depth meas- pinger, the diver assembly contains a pressure urement at extreme ranges where the normal depth transducer that reduces the power when the water calculation would be degraded due to multipath depth is less than 6 ft. The diver unit is housed propagation. To mark a specific spot on his track, in a clear, plastic (Lexan) cylinder to allow the diver depresses a switch that elongates the monitoring of the switches that control the trains- tracking pulse to 4.096 ins. Emergency conditions mission period, a neon light that indicates the are indicated by locking this switch down, which transmitter is transmitting, and a liquid-crystal elongates both the tracking pulse and the depth display of a number of internal parameters. pulse. In normal operation, the array and RPT would The acoustic receivers incorporate automatic be activated approximately 1/2 h before the dive. gain control (AGO) so that the receiver gain is At this time, the RPT is synchronized to the array always the maximum possible in the particular noise and the diver units turned on and synchronized to environment, and a time-varied gain so that trains- the RPT. Feedback from the diver unit allows missions from divers slightly beyond the 1.048576-s monitoring of synchronization drift. The diver period will not affect the following period. The unit can be hung from the side of the boat for a PATS design goal was a signal-to-noise ratio of position check. System operation is now fully 26 dB at the receiver. The transmitting level of automatic. 56 RCVG STATION,--RCVG STAION, RCVG STAION,~-'TYPICAL RCVG STATION ASHORE TENDER OR SUPPORT CRAFTCOPTR AD TELEMETRY IVER ~ ~ ~ ~ DIE 2 1~~DVE Figure 3. Artist 's sketch of system. 57 Tracking Array period and then switches to receive to conserve power and allow radio control of the array. Cur- The original concept was that almost all of rently, the only function implemented is the power the computational ability would be located in the on/off control. To do this, a small part of the tracking array. However, the size of the survey circuit must be active at all times. and tracking programs became larger than expected, increasing the current drain because of the added The acoustic tracking frequency is 31.25 kHz programmable read-only memories (PROMs). Further, to achieve long range and the signals are ampli- it was thought that, at least in the first version, tude modulated and frequency shift-keyed. Since a the more complex parts of the system should be shorter distance is involved, the survey frequency easily accessible. Therefore, the survey and is higher (250 kHz) to provide better resolution; tracking programs were moved to the RPT unit. The the survey frequency is amplitude modulated. recent availability of low-drain PROMs and micro-Tharyisdigetowhtndhempc processors will allow a return to the originalTharyisdigetowhtndhempc concept when the system no longer requires modifi- of a small boat. The flotation collar that houses cation. the radio antenna is highly visible to assist in location. Short-time operation (I day) is pro- The microprocessor portion of the array elec- vided by an internal battery and extended opera- tronics is based on the Intel 8080.2 This is an tion by an external, field-replaceable battery. eight-bit parallel processor with instruction times between 4 and 20 pjs at our clock rate of Receiver-Processor-Transmitter 1 MHz. The microprocessor can run at a clock rate of 2 MHz, but a slower rate allows the use of CMOS The RPT unit can be operated on the beach, on gating and logic, which reduces the current drain. board the divers' boat, or at the site of a more The limiting factor for the current drain now is powerful computer where the RPT data can be sub- the RE transmitter which draws 4 W and has a 50% jected to more sophisticated processing. The RPT duty cycle. unit itself is battery-operated; a small, battery- driven inverter is used for the auxiliary printer A precision, 4-MHz, temperature-compensated and plotter which are powered by alternating cur- oscillator provides the drive for a binary timing rent. Since current drain is not as severe a chain. The resulting period is 1.048576 s. Some problem in the RPT as it is in the tracking array, thought was given to using decade counters so that the processor is an Intel SBC 80/10 single-board the period would be exactly I s and could be syn- computer.3 chronized with WWV of the National Bureau of Stan- dards. It was decided, however, that the storage The RPT, in addition to transforming the of the additional data would overburden the system. timing numbers into X,'Y,Z coordinate positions, The period is divided into 65,536 parts by repre- can provide information to the following detach- senting the timing number with a 16-bit digital able devices: (a) an analog plotter, (b) a more number. The resolution of the timing number is powerful computer, and (c) a line printer. The thus about 16 lis, or 2 in. in water, so that the internal liquid-crystal display can show coordi- system resolution is limited by the acoustic nate information for a selected diver or, alter- jitter. Data are strobed from the timing chain natively, the range and bearing of a diver to any into holding latches since the Intel 8080 is not other diver or the ground reference points, so capable of servicing three interrupts in 16 ps. that divers can be easily vectored to other points The timing chain, through gating, also provides on the range. The original timing numbers are re- the drive for the four survey transmitters and the tained on a tape cassette so that a permanent two ground reference transmitters, the drive for record is available for replay. Replay is accom- the RE transmitters, and, with a PROM, the TVG- plished on the RPT unit by changing one switch, AGC curve for the acoustic receivers. and transferring the data cassette from the tape writer to the tape reader. The radio telemetry link between the array and the RPT unit is a 4-W, 139.55-MHz radio with a Software 3-kllz bandpass. While only one frequency is now in use, two may be required if two systems are The tracking array program gathers timing used within 5 mi of each other. numbers for 10 s and determines whether a valid (three receiving transducers out of four) mark/ The elemtry ransisson cnsiss oftheemergency pulse has been transmitted by a diver. The elemtry ransisson cnsiss oftheThis information, along with the system status, is data gathered in the previous ten 1-s periods and reformatted and sent to the radio transmitter. is 180 eight-bit bytes. Transmission is asynchro- The program requires 1250 bytes of memory. nous and is modulated at a rate of 500 Hz. The beginning of the data burst occurs precisely at A test unit has been designed and constructed the beginning of the 10-s period to allow synchro- that uses data from an externally-generated tape ization of other units to the array. One hundred cassette to simulate the acoustic pulses that seventy-six bytes represent timing numbers; two would be received during a diver operation. In bytes contain mark and emergency information, and this way, the entire processor, software and hard- two bytes contain system status information (buoy ware, was extensively tested before actual data moisture alarm and battery low voltage alarm). were available. The array transmits about one-half of the 10-s 58 The survey and tracking programs for the RPT Other Uses unit were written in FORTRAN on a minicomputer and were thoroughly tested before being transferred to Although the PATS will initially be used by the microcomputer. These tests involved a simu- the Naval Facilities Engineering Command, other lated track with 100,000 points. No errors were uses are possible. For example, operations in- found when signals from all four tracking trans- volving underwater oil platforms require precise ducers were received; the maximum error was 1% information about the location of structural im- when one transducer was allowed to drop out ran- perfections. Inquiries have been made concerning domly. First, the survey program is initialized the use of PATS to study the habits of sharks. In and performs a survey (i.e., constructs the matri- this case, placement of the "diver" unit on the ces for transforming the timing numbers into shark would clearly be the responsibility of the coordinate positions, using the spacings between user. the buoys and from the buoys to the surface). Six valid surveys are required before proceeding Acknowledgments to the tracking portion of the program. To be valid, the surveys must contain no missing spac- This work was supported by the Naval Facili- ings and the redundant spacing must agree with ties Engineering Command under Contracts N62477- the calculated distance. Depths are averaged to 7S-C-0364 and N00017-74-C-1208. compensate for wave action. The first survey requires about 1.8 s, and subsequent surveys take about 1.3 s. Next, the tracking program starts. References The initial tracking computation requires recep- tion of signals from all four transducers; the 1. G.M. Wenz, "Acoustic ambient noise in the first three timing numbers are used in one compu- ocean: spectra and sources," J. Acoust. tation and the last three in another. These Soc. Am. 34: 1936-62, December 1962. values are then compared to resolve the image ambiguity, and the correct two are averaged. 2. Data Catalog 1977, Intel Corporation, pp. After the initial computation, computations with 10-28 to 10-32. only three transducers may be made based on pre- vious coordinate data. The entire tracking com- 3. T. Rolander, "SBC 80/10 - System 80/10 putation for all 10 periods requires about 5 s. Single Board Computer Applications," Intel Total memory requirements for the survey and Corporation Application Note AP-26, 1977. tracking programs are about 10,000 bytes. 59 SWEPT CARRIER ACOUSTIC UNDERWATER COMMUNICATIONS A. Zielinski L. Barbour, Graduate Student Faculty of Engineering and Applied Science Faculty of Engineering and Applied Science Memorial University of Newfoundland Memorial University of Newfoundland St. John's, Nfld., A1B 3X5 St. John's, Nfld., A1B 3X5 CANADA CANADA Abstract Both factors are time dependent, resulting in a complex, randomly time varying communication This paper describes a novel idea of reducing channel. The transmitted acoustic signal is also multipath interference in underwater acoustic attenuated due to spreading and absorbtion of the communication systems. The proposed idea is based acoustic energy by water. on the observation that the time separation be- tween direct and multipath waveforms can be util- 2. A Model for Underwater Acoustic ized to suppress the multipath interference. A Communication Channel proposed system employs a carrier with instanteous frequency swept periodically in time over a broad It is assumed, that in addition to multipath, frequency range. This frequency range is much there exist a direct acoustic path between trans- larger than the bandwidth occupied by a message mitter (T) and receiver (R) as shown schematically signal. On the other hand the reciprocal of in Fig. 1. sweeping rate is much smaller compared to the message bandwidth. The carrier is modulated by a message resulting in a time varying instantaneous narrow-band spectrum of the transmitted signal. ? XSURFACE PATH The received signal is demodulated by a band-pass T R tracking receiver and suppression of the out-of- DIRECT PATH band multipath spectrum is accomplished. A simple model for the multipath channel has been BOTTOM PATH proposed and related to the system operation and performance. Some results and conclusions from SEA FLOOR tests performed on the prototype are presented. FIG.1 MULTIPLE UNDERWATER ACOUSTIC LINK 1. Introduction In line with increasing interest in ocean A complete underwater communication system research and industrial offshore development there can be represented by a model shown in Fig. 2. is a growing demand for an efficient, reliable and high speed underwater sonic communication system. Speech communication between divers, remote 1--------------I control and telemetry for underwater operations, (t) i diverse oceanographic measurements are but some examples of data to be transmitted. T7I~ ~i~T +I + TRCIj PATH EE Underwater communication by means of the TRNSMITTER RECEIER acoustic waveform encounters however several prob- | H' l lems related to the characteristics of the under- water channel [1]. With sufficient transmit power l for an adequate received signal level, the principal obstacle against reliable transmission is believed to be the multipath interference [2] . FIG. 2 AN UNDERWATER COMMUNICATION SYSTEM MODEL to be the multipath interference [2]. The physical phenomina responsible for multi- path are primarily: In the transmitter, a message signal, x(t), is - multiple acoustic paths resulting transferred by a suitable modulation into a desired from surface and bottom reflections frequency range. The electric signal is then con- verted into an acoustic signal by a transmitting - Ray bending due to water temperature transducer represented by a filter T (j). The variation. acoustic signal, s(t), is transmitted through an 60 underwater channel, converted back into an electric whereas the random character of H (jw) can be signal by a receiving transducer, TR(jw), and introduced by considering H (ft) and i (w,t) as applied to a receiver front filterK(ju), which random processes for any particular w. is followed by a demodulator. x'(t) represents the demodulated message signal. The underwater channel A simplified probabilistic model for multi- is simulated by a parallel connection of the direct path which assumes an infinite number of scatterers and indirect path. A time delay To in the direct with random and uniform phase gives a Rayleigh 0 -jW-ro probability density function (pdf) for Hm(wt) path is represented by filter e-T� and is related a nd uniform pdf for )(ft) [5]. and uniform pdf for ijm(w,t) [5]. to the geometrical distance S between the trans- mitter and the receiver and to the sound velocity c. Attempts have been made to obtain a more = (1) adequate model for H (jw) by relating statistical T -~~~~~~~~~~~~1 o c properties of the modeling filter to the actual The filter, H (jw), represents attenuation and physical conditions. Venetsanopoulos [3] absorbtion of the acoustic energy in the water and developed, for example, a statistical model is assumed to have a real transfer function, i.e. related to surface roughness. K (jw) = CH () Finally, n(t) in Fig. 2 represents an �0~~~ 0 ~~~~additive ambient noise which will not be consid- Based on emperical data [4] and assuming spherical ered in this paper. spreading one can conservatively approximate H (au) by: 0 When the delta pulse is applied at the input of the underwater channel model, i.e. s(t) = 6(t) H (2Tf)O = -(20 log S + 36.4 x lO5fS) (2) then the received signal r(t) will be as shown Ho( fdB = -(20 log S + 36.4 x 10-5rs) (2) schematically in Fig. 3. for 20kHz < f < 100 kHz r(t) _ _ . , ~~~~~~~~~~~~~~~~r~t} where Ho(2rf)dB is expressed in dB refering to l1barDIRECT PATH d(t) S is the distance between transmitter and t MULTIPATH receiver in meters l f is the frequency of the transmitted signal t in kHz. rm t Clearly, H (X) is the transfer function of a low I T pass filter. FIG. 3. IMPULSE RESPONSE OF THE UNDERWATER A time delay, T , in the indirect path of CHANNEL MODEL. Fig. 2 is related to the shortest multipath dis- tance between transmitter and receiver and it is The direct path impulse response, d(t), has a assumed that finite width due to the low-pass nature of the filter H (JW). The multipath signal, m(t), is the T > T(3) impulse response of the time varying random filter m o H (ju). Because of the finite energy carried by A filter H (jl) represents the combined effect of tle multipath signal all multipath signals. It is basically a low pass filter due to the similar considerations as for him E {m(t)} = 0 H (jw) filter. Longer distances for multipath t + imply higher attenuation and therefore a smaller bandwidth of Hm(jw) compared to Ho(j) should be where E { denotes expected value. expected. m 0 The multipath duration time, T, can be arbitarily defined as the value of t for which E {m(t)} is The random and time varying nature of multi- neglegible (assuming time origin at T ). Physically path suggests that the modelling filter, H (jw), T + T is the time limit for the meaningful multi- m m should be a time variant, random filter described path signal. in probabilistic terms rather the deterministic *transfer function H (jw). When a harmonic signal (carrier) ~~m~~ ~s(t) = Acos W t = Re {Ae c }; _ < t < With this approach, the time dependance can c(t) = Acos w t = Re Ae c < < be explicitly indicated by writing (5) is applied at the input of the underwater channel Hm(i) = H (,t) eif(W't) model, then the output signal is given by: m m .bWt W T where X (fit) = I H Qw) |(4b) r(t) = Re { Ae3 c [e c ' (jW ) + and m(w,t) = L Hm(jw) (4e) (U c ( +6) = Re { M(t)e c De 61 where M(t) = A H m(W ,t) (7a) The receiver front filter K(jw) is assumed to ~~~m ~c~~ ~have a band-pass frequency characteristic, a(t) = -W T + lm(wc ,t) (7b) centered at the carrier frequency w , bandwidth c m mcAB and impulse response duration b. c D = A Ho(~c) (c D=AoH (wdc (7c) cIn sequel we introduce a instantaneous and = -w T (7d) possibly a time varying spectrum, F(w, t), as c o Incthe o not n a p , (t) determined by a finite observation window b. In the phasor notation, a phasor, M(t)ej'(t), represents the tine varying multipath signal, m(t), In particular the instantaneous spectra of while phasor, De , corresponds to the direct the direct signal, d(t) and multipath signal, m(t), signal, d(t). are denoted by Fd(w, t) and F (, t) respectively. aydt) A phasor, R(t)e t), representing the For a sufficiently high carrier frequency received signal, r(t), is a geometrical sum of the i.e., for others as shown in Fig. 4, where 2 << b (13) -- < < b ~ ~~~~~~(13) R(t) eV(t) = M(t)eja(t) + Dej (8) c the instantaneous spectra, F (w, t) and F (w, t) are narrowband and essentially time independent WI as shown schematically in Fig. 5a, where Fd(w, t) / t)eJ(t) is in fact a single frequency component wc' Rtfieiit)/ / /Dejp IFd(w, AR0 ~d (wt) Fm(.,,t) 1Tm-rk ],fi FIG.4 PHASOR REPRESENTATION FOR THE DIRECT I I.'. | - - AND MULTIPATH SIGNALS. c ( l (a) ~ :k ~ (b) For constant values of D and ~, the direct FIG. 5 INSTANTANEOUS SPECTRAOF THE DIRECT AND signal, d(t), has only one frequency component, w c MULTIPATH SIGNALS On the other hand the multipath signal which is both amplitude and angle modulated: i.e. M(t) = 4(t) Cos [to t + cx(t)] (9Suppose now that the carrier frequency, wto is m(t) = M(t) cos [~ct + a(t)] (9) chagn'iery ntmie c changing linearly in time, i.e. occupies a finite frequency band. Because M(t) and (t) = kt, -- < t < (14) a(t) are in practice slow varying functions of time, c the frequency band occupied by m(t) is narrow and where k is a proportionality constant. centered at w . Similarly, the received signal, r(t), is both amplitude and frequency modulated: The change in carrier frequency is followed by r(t) = R(t) cos [t + Y(t)] (10) changes of the instantaneous spectra, Fd(w, t) and r(t) = R(t) cos [tot + y(t)] (10) (o ) c ~~~~~~~~Fm (w, t). The instantaneous frequency of the received signal If the rate of change is slow enough, F(, t) is: ~~~~~~~~~~~~~~If the rate of change is slow enough, F d(w, t) is:d will remain approximately a single frequency com- (t) dy + ponent located at w (t - T ). The spectrum, it dt +c(1 F (t, t), due to the time Relay, T > T , is The presence of interfering amplitude and angle lagging Fd(w, t), and due to multipath duration T. lagging mFd(w, t), and due to multipath 2uration T, is broadened proportionally to T as shown in modulation of the received carrier reflects i raee rprinlyt ssoni modulation of the received carrier reflects Fig. 5b. Clearly, if the receiver band-pass filter difficulties encountered in underwater communi- with bandwidth AB could track a slow change in the cation. carrier frequency w (t), then the multipath signal 3. Swept Carrier Communication Concept spectrum would be outside of the filter bandwidth resulting in principle, in a complete rejection of A situation is considered in which the carrier the multipath signal. of eq. (5) is transmitted through the channel model Any narrowband modulation of the carrier by of Fig. 2. ~~~~~~~~~~Any narrowband modulation of the carrier by ~~~~~~~of Fig. 2. ~a message signal would broaden both Fd(w, t) and F (o, t) spectra but with a proper choice of a It is assumed that the receiving transducer m characteristic is constant over a broad system parameters the frequency separation between characteristic TR(j) is constant over a broad Fd(t, t) and F (w, t) could be maintained. The frequency range, i..e. d m... ~~frequency range, i.e. ~main practical problem in the implementation of this idea arises from the limited frequency range TR(jw) = 1 for wmin < w < Wmax (12) over which the carrier frequency can be swept. 62 Water attenuation and transducer characteristics as a periodic function of time. A more convenient confines the sweeping range to a certain limit geometrical model for finding the locations of Fd(t, m) and F (t, m) can be derived from Fig. 6a wu i < w (t) < w mx(15) and is shown in Fig. 6b. Here the time origin is - c - c -max assumed to be: -T . The multipath spectrum width For continuous operation therefore, X (t) has to and its location are obtained as the projection be a periodic function of time, with a period T of a segment: w T ce(t) = f(t) = f(t - T) (16) M = 2B sin -2 (19) Because f(t) can always be represented by a Fourier which is rigidly attached to the rotating phasor Series we limit our attention to the harmonic T frequency sweep: j(s(t - T - ) (20) Be iC(t) = 0 + B cos Wst (17) where = T - T m o where i - central frequency From geometrical consideration the multipath 2i = lwwpgrucspectrum central frequency cm is given by w =-- - low sweeping frequency wST Ts Tn= +cs (sT- w- (21) Sm = o + B cos ( st - sT - -) (21) and w B = m . oand m -B min and bandwidth, AM by: c + B = T c T 0max ~~~~~o mw~ax ~AM = 2B sin (w t - msT - - ) sin 2 The time-frequency relationship for the instant- (22) aneous spectra, Fd(c, t) and F (w, t) are now more The receiver's tracking filter of bandwidth complex then that shown previously in Fig. 5 for AB follows the location of the instantaneous freq- the linear sweep. To show this relationship it is uency of the direct signal, i (t). As seen from convenient to write the departure of the instant- Fig. 6b, the direct and multipath signal spectra aneous carrier frequency, ID (t), from the central move along the frequency axis periodically in time. frequency, o, as: The multipath spectrum, F (i, t),falls periodically jW t within the tracking filter bandwidth AB. The W (t) - W = B cos (t) = Re {Be (18) multipath signal supression results as an effect c o a of two mechanisms: Jmt for periods of time in which AB and AM The projection of the rotating phasor, Be , do not overlap f in which AB and AM * ~~~~~~~~~~~~~~~~~~do not overlap, F (in, t) is outside of onto the real axis therefore indicates the trans- the receiver filtmr bandwidth and ideally mitted carrier instantaneous frequency. is completely rejected by the filter. is completely rejected by the filter. Similarly, the instantaneous frequency lo- - for periods of time in which AB and AM cation of the direct signal is given by: overlap, multipath supression is still R.ci e (t - T ) possible, if ~~~~Ref { ~~BeJ~ 0} ~AM > AB The instantaneous frequencies of the earliest and In this situation only a portion of the latest multipath signals are represented by phasors multipath energy falls within the jBe s - m and B js(t Tm - T) receiving filter bandwidth. Be and Be respectively. The location of the instantaneous Any narrowband modulation of the swept spectrum of the multipath signal lies then between c arrier by a message signal will not appreciably carrier by a message signal will not appreciably the projections of the above phasors as shown in alter the geometrical model shown in Fig. 6b. If Fig. 6. modulation is applied, the instantaneous spectrum of the direct signal will occupy a finite frequency Fdi~~l"~t_) ~~i ~band, and the multipath spectrum will be broader / FmIOD\/~CttiIOf / than that without modulation. �|fo Af /mli t{ m E4. System Parameter Considerations \~"("~~ ~~~~~~ \\| \The swept carrier communication concept des- _ \ X~vD(|~'o~m) < >~<>cribed earlier is applicable to any narrowband ~B.J~,~-%~-%-rm) B h)modulation applied for a message transmission. The ~~~(a) (b) most obvious and simple choice would be either the narrowband frequency or amplitude modulation. The FIG. 6 PHASOR REPRESENTATION FOR F,(,I) AND F;,4) message signal, x(t), could be either an analog or digital signal of a limited spectrum. Without these specific assignments it is difficult to pro- As can be seen from Fig. 6a, the multipath spectrum, pose a system performance criterion for parameter Fm(t, w), changes both its position and width, AM, optimization. 63 Certain general observations can be made how- The slowly varying instantaneous carrier frequency ever at this point. Considering the geometrical is recognized by the functional block "Inst. model of Fig. 6 it seems obvious that increasing carrier". The output signal of which the sweeping frequency range w - w . will allow better separation between The dIrect and c multipath spectra and therefore will result in is then used to tune the central frequency of the better multipath supression. However, the maximum band-pass tracking filter. The output of the sweeping range is limited by the characteristics tracking filter is passed through a demodulator, of the transmitting and receiving transducers and the output of which represents the transmitted by water attenuation. message signal x'(t). Parameters, T and T, are determined by a An experimental system has been designed and particular transmission situation and by multipath tested. Narrowband frequency modulation was duration. The possibility of optimization lies in chosen for message transmission. the proper choice of a sweeping function, f(t) shown in eq. (16) or, if sweeping is harmonic, by The prototype has the following parameters: the proper choice of a sweeping frequency ws. - frequency sweep range : 65kHz - 95kHz If, for example, it is required that (fmin fmax) AM < W - cW for all t (24) - central frequency, f : 80kHz - sweeping frequency, : 5Hz < f < 50Hz then the sweeping frequency must satisfy f - variable s wT M = j 2B sin < 2B - Message signal : 300Hz < f <1800Hz 2 frequency, fm or w T < n (25) sor XT < a - (25) - Receiving filter : 4kHz bandwidth, AB For the maximum frequency separation between direct and multipath spectra, the angle ~ in Fig. Back to back tests of the system were con- 6b should be: ducted and excellent results were obtained for the w T designed parameters' range. W = ) T + s (26) s (2 To assess the system performance in a multi- Both conditions (25) and (26) can be satisfied path situation a harmonic interfering signal has only if been added to the unmodulated transmitted swept carrier. The signal w(t), proportional to the T < 2T (27) instantaneous carrier frequency is shown in Fig. 8. and the optimum sweeping frequency in this case is: s T (28) s+ T2 For large T, the multipath spectrum will occupy the entire bandwidth ( . min, m) and the system min. max performance can be improved by narrowing the tracking filter bandwidth. 5. Receiver Structure and Experimental Results The block diagram of a possible receiver configuration is shown in Fig. 7. r(t) SAND- DEMODU- x(tI Fig. 8 - Recovered Sweeping Signal, w(t) The maximum and minimum values of w(t) correspond wet) =kalw,1) to the carrier frequencies of 65kHz and 95kHz respectively. The interfering tone has a freq- INST. uency of 80kHz and small distortion can be ob- served in the corresponding level of w(t). A similar test has been conducted for modu- FIG. 7 BLOCK DIAGRAM OF A RECEIVER lated carrier. Fig. 9 shows a transmitted and 64 demodulated message signal, x(t) (upper trace) and x'(t) (lower trace) respectively. The period- ical distortion seen is again caused by an inter- fering tone of frequency 80kHz. Fig. 11 - Recovered Sweeping Signal,w(t) More extensive full scale ocean tests are to be carried out in the near future. Fig. 9 - Transmitted and Demodulated Conclusions Message Signal Limited tests have also been carried out in a small Based on limited experimental results the pro- water tank. Due to the dimensions of the tank, posed swept carrier communication coccept seems to very strong and long lasting multipath was encount- be a promising way of message transmission in a ered. Fig. 10 shows the received signal structure nultipath situation. The generality of the concept (lower trace) for the transmitted harmonic burst allows different modulation - demodulation (upper trace). schemes to be considered for further investigation. Full evaluation of the swept carrier communications concept shall be possible only after extensive ocean tests. Summary A model for multipath communication is des- cribed. Based on this model, a novel swept carrier communication concept has been proposed and dis- cussed. Several observations concerning system parameters have been made. Some experimental results from tests carried out on the system proto- type are presented. References 1. Lytle, D.W.: "Characteristic Problems of the Underwater Channel", IEEE International Fig. 10 - Tank Reverberation Conference on Communications, Seattle, Fig. 10 - Tank Reverberation Washington 1973, V2 pp. 38/1-3. Even under this severe condition, the tracking signal w(t) was well defined as shown in Fig. 11. 2. Chauncey, S. Miller; Bobman, Carl E: "An Experiment in High Speed Underwater Telemetry", IEEE Tnt. Conference on Engineering in the Ocean Environment, New Port, R.I. 1972, pp. 34-38. 3. Venetsanopoulos, A.N.: "Stochastic Filter Modeling for the Sea-Surface Scattering Channel", J. Acoust. Soc. Amer., V. 49, N 4, PI 1, April 1971, pp. 1100-7. 4. Urick, R.J.: "Principle of Underwater Sound for Engineers", McGraw-Hill Book Co., 1975. 5. Schwartz, M.: "Information transmission, modulation, and noise", McGraw-Hill Book Co., 1970. 65 SMALL SUBMERSIBLE ACOUSTIC COMMUNICATIONS SYSTEM DESIGN Salvatore D. Morgera Raytheon Company Submarine Signal Division Portsmouth, RI 02871 Abstract the mother ship and up to depths of about 20 kft. It is desirable for future Navy oceanographic and military A design is presented for a system providing highly desirable for future Navy oceanographic and military reliable command and control acoustic communications operations to equip the submersibles with sophisticated active and passive sonar equipment and have it operate between a mother ship and a number of small, fast sub- active and passive sonar equipment and have it operate at ranges up to 100 nmi from the mother ship and at mersibles. The small submersibles may be employed speeds of up to 10 kn. speeds of up to 10 kn. for underwater mining, exploration, bottom mapping, or military surveillance. Modulation and coding design Figure 1 illustrates the scenario of interest. A is presented; the techniques discussed provide multiple surface ship with an acoustic transmit array, generally protection against multipath and fading, high reliability, at a shallow-to-intermediate depth, communicates with acceptable transmitted signal total time duration, sim- a submersible equipped with an acoustic receive array, plicity, and economy. The required decision point generally at an intermediate-to-deep depth, the objec- signal-to-noise ratio (SNR) for Rayleigh fading condi- tive being to provide command and control information tions is derived for the modulation and coding design. to the submersible. The acoustic information transfer Particular attention is paid in the receive signal pro- is treated here as strictly being from ship-to-submer- cessing to the initial doppler and message doppler varia- sible and the amount of message information that must tion problems inherent in a scenario with mobile end be transmitted is small. We assume that any data points. A Figure-of-Merit calculation is provided for gathered by the submersible can be retrieved at a later typical geometrical and environmental parameters. It time, either physically or via EM means, although a is shown for a realistic source level that the required submersible-to-ship communications link utilizing data SNR can be achieved at long range with extreme end compression is possible. Since there may be more point motion. than one submersible, the message must contain desti- nation information as well as the actual command and 1. Introduction control information. It is anticipated that the above message information can be conveyed by an 8-bit block. The explosive growth of undersea vehicles, which followed development of North Sea oil and gas discov- The minimum operational range is established as eries of the late sixties and early seventies, continues. 100 nmi. This operational range capability permits a While the diver still dominates commercial underwater substantial operating radius about the surface ship. The activities, the role of manned submersibles and reliability of the communications system is described by remotely controlled vehicles progressively increases. two (conditional) error probabilities: The largest US submersible commitments at present 1) Pr { Mi * mi I mi }, the probability that a message are in the academic and military communities. The mi destined for the ith submersible is not received and Navy has a total of seven manned submersibles and seven unmanned vehicles in operating condition at the moment. They are based on both coasts and in Wash- ington, DC; their largest concentration is in the San Diego area. Most of the Navy vehicles have a capability for TV viewing and video taping. Most have a manipulator and are constructed to carry 35 mm or 70 mm still cameras with a light strobe. Configuration, dimensions, weight, complexity, and depth capabilities vary from vehicle to vehicle, as does work instrumentation and vehicle atti- tude monitoring instrumentation. At present, Navy vehicles operate within a relatively confined radius of Fire . Small Submersible Scenario 66Figure. Small Submersible Scenario 66 decoded correctly as Mi, and 2) Pr I Mj = mj I mi; a surface ship projector (source) depth S = 30 ft (9.1 m), i * j }, the probability that a message mi is received receiver array depths R = 300 ft (91.4 m) and R = 1000 ft and decoded as a valid message by the (incorrect) jth (304. 8 m), and frequency fc = 200 Hz to obtain the fol- submersible. The communications system is designed lowing interval estimates for L and B,t to conform to the following error probabilities per L e [0. 5, 2.4] s (3) message, Pr MiM mi | mi m =10-6 (1) B [. 02, 0.2] Hz. (4) The lower limit on L is valid for the important range Pr iM. = I Jm -12 ij 92,,N increment 100 - 200 nmni; the upper limit on L obtains j j i' s only at the shorter ranges on the order of 10 nmi. The lower limit on B applies to volume dispersion (i. e., purely refracted propagation paths) and an end point where relative motion of 30 kt at a range of 100 nmi; the upper N N 8 limit on B includes the effects of surface and bottom s NT < 2 scattering with the standard deviation of the boundary undulations assumed to be 1.4 m (SS3) and 0.2 m, The quantity Ns is the number of submersibles within respectively. The upper limit on B is believed to be operational range and NT is the command and control quite pessimistic. information. The communications system's basic design philoso- 2. Modulation and Coding Design phy is to use minimum signaling elements well suited to the acoustic communication channel and any existing The modulation and coding design proposed here The modulation and coding design proposed here equipment, further improve the bit error rate perfor- mance by diversity to achieve a moderately low bit * Multiple protection against multipath and fading error rate, and employ error correcting coding to obtain the final error rates. The proposed callup com- * Low (conditional) error probabilities munications system employs 4-ARY FSK (M = 4), a [cf. Equations (1) and (2)] simple modulation technique which is efficient in terms * Acceptable transmitted signal total time duration of the available bandwidth W, facilitates recognition of fades and receiver AGC adjustments, and allows an � Simplicity acceptable transmitted signal time duration. Multiple * Economy. protection against multipath and fading is provided, first, by use of a long-time duration T for the minimum The last point above is a matter of extreme practical signaling element (tonal or chip), i. e., we choose, importance; the proposed modulation and coding design can be easily implemented utilizing existing communi- T = 16s >> L (5) cations system technology. and second, by the redundancy offered by 5-fold fre- Several considerations, foremost among them being quency diversity with the diversity-to-diversity channel the capability of existing communications system hard- tonal spacing much greater than the fading correlation ware, the minimum operational range requirement of width of 1/L. The transmitted tonal frequency width is 100 nmi and associated standard path loss to the approximately 1/T; whereas, the receive frequency receiver, and the relative absence of commercial and resolution Af is equal to military traffic lines, have led to a signaling center frequency fc in the range [175, 225] Hz and a signaling Af T- = .125 Hz > B (6) bandwidth W of 50 Hz, as shown in Figure 2a. The These selections leave sufficient unused bandwidth to geometry of surface ship and (any one) submersible, depicted in Figure 1, allows multiple reflections, the down-link end point relative motion. The long tonal caustics (surface ship projector in surface duct), and time duration and the redundancy act to increase the ray path refraction. The severity of these effects ray path refraction. The severity of these effects SNR at the decision point for the transmission bits. depends (environmentally) on the ocean depth, season, The modulation and diversity desin parameters are The modulation and diversity design parameters are nature of the boundary inhomogeneities, and vertical and depicted in Figure 2b. horizontal sound velocity gradients. In general, there will be degradations suffered by the transmitted signal, due to multipath cancellation and fading; the source of these degradations is conveniently described by the t Bottom class and sound velocity profile (SVP) medium time dispersion L and frequency dispersion B. employed were those associated with the acoustic NISSM II and other theoretical models were utilized with province MGS3(57N, 40W) and the summer season. 67 ~~~~~~~~~~I I ~ to be maintained at 10-6 [cf. Equation (1)], the prob- ability of a codeword error, from Equation (8), must 1WCSON: *1 ,c (175,225] Hz be held to, -7 2a. SIGNALING CENTER FREQUENCY AND BANDWITH Pr { Mi * mi I mi 5 x 10 2= 1,2 (9) The false message error probability of Equation (2) is related to the probability that both codewords DIVERSITY CHANNELS DIVERSITY CHANNELS (cli, c2i) contain the same error, i.e., I I Pr{ Mj . mi;i j = Pr { M. = mj mi;ij }. w~~~~~~~~~~~~~~~~~~~~~~~~~~~ - -IOHz IC liT WCDORE ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Pr$M2 M j~ i*i} Pr {[ M J2j = mj mi; ij} 1 <(N)[1 Pr {Mj=mi mi; M DIVERSITY 4-ARY DIVERSITY j = 1,2 (10) CHANNEL I CHIS CHANNEL2 , I0 DOPPLER I I SHARD = id � L.? *-.Z 1_r A 1 1 1 I The upper bound is obtained by assuming that each code- , , O --) ;' _- O*word has N nearest neighborsttwhich are equally likely T ~L ~ to decode in error. The bracketed complementary I H '' 5 : probability is numerically equal to that of Equation (9). NOTES: For a particular 8-bit block of information bits, there I. THERE ARE 80 SPECTRAL SINS PER 81VERSITY Z TIME DURA RATIO N OF MINIMUM SGNAGDIVERSITY are eight nearest neighbors; thus, N = 8 and, (DENOTED BY SHADED AREAS) T:61i 2b. MODULATION DESIGN PARAMETERS Pr { Mj = m. I mi; i * j } 3.1 x 10-14 (11) Figure 2. Modulation and Diversity Design J which is substantially less than the error probability The low (conditional) error probabilities of Equa- requirement of Equation (2). An upper bound on the tions (1) and (2) may be met by an encoded message probability that any one or more vessels have a false format consisting of two codewords, each carrying the message is given by, eight information bits. Upon reception, the codewords N are decoded separately and the eight information bits 1 - [I - Pr mi ; i 4 j ] s . accepted as valid only if both codewords decode identi-12 cally. Mathematically, we may express the ith encoded Pr { Mj. = m i; i * j} <1.0 x 10-2 message xi, in the following manner, x. 4[f1I(m.). 2 (MiA= (c .'(7 by invoking the Union bound and assuming that N = 32. x. -&[f (m.), f (m)]= (c ., c .) (7) 1i = f lm ) 2 I 1l 21 We now turn to the detailed code structure. The code must be capable of correcting at least two errors wherse f () f2a () denotae linear transformations on the so that received bits of moderate quality will not affect message mi1 and (Cli, c2i) are the two aforementioned message condcat ated. the p itwof a men ned proper response, and must admit economical and simple codewords concatenated. The probability of message encoding and decoding. The latter requirement demands error may be written as, that the codeword length be as short as possible. The Pr { M mi I mi } = 2 Pr { Mi .= m. i m. number of linear code alternatives that satisfy the 1i 1i 1| 1i ~ Ql~~ w Ipractical requirements is small; namely, we have the - Pr Mi = M i2 (21, 11; 2, 6) difference-set code and the (24, 12; 3, 7) L-Pr {M2 i * m mi} (overallparity check) extended Golay code [1]. The difference-set code corrects up to two errors in the 2 Pr { Mi mi mi } = 12 21-bit codeword and is relatively simple to decode; the (8) fact that it is a Projective Geometry cyclic code also guarantees that it can be (1-step) majority-logic where Mli and M2i represent the received and decoded decoded. The Golay code corrects up to three errors messages associated with the codewords cli and c2i, and is capable of detecting the existence of a fourth respectively. Since the probability of message error is error in the 24-bit codeword. The Golay code is mod- error in the 24-bit codeword. The Golay code is mod- erately hard to decode; however, much recent work on efficient implementation and block synchronization t We assume that the occurrences of codeword errors techniques make it the more desirable choice. To use are statistically independent, justified by the fact that the codewords will be separated in time by greater than the average fade duration 1/B. tt In the Hamming metric sense. 68 the Golay code, we shorten it to a (20, 8; 3, 7) code in this is a most reasonable figure for the operational order to accommodate the eight information bits. Note range requirement, as shown in Section 4. that the shortened code retains the same distance and error correcting features of the original code ]. In concluding the modulation and coding design, we must mention the total transmitted signal time duration. Assuming the codeword error of Equation (9), we We transmit two codewords for a total of 40 bits with may compute the error rate per coded bit for the the time duration of a minimum signaling element equal shortened Golay code, and then the required SNR per to 16s; thus, the total time duration is 10. 7 min. We tone at the decision point. The interleaving and una- must also allow, at most, 1.0 min for a preamble seg- nimity count on the transmission bits acts to guarantee ment to alert the receiver and provide real time bit the statistical independence of coded bit errors; there- synchronization and 2.0 min propagation time for a fore, the coded bit error rate P may be found from the 100 nmi range. expression, Pr $M~i �, I n MJ 5 - 10~-7 3. Receive Signal Processing The receive signal processing is that commonly 20 20 n 20 - n associated with M-ARY FSK supplemented by doppler / ( P (20 -P) processing to accommodate the down-link end point n = 4 relative motion. It is required that the initial doppler be estimated and that any doppler variations be esti- 4 (2) P 2= 1, 2 (12) mated throughout the transmitted signal time duration of approximately 10.7 min. and To aid in the initial doppler estimation, a two-phase P 5 3.2 x 10 (13) preamble is proposed. Each phase would consist of a predefined burst of no more than five tonals with time The transmission bit error rate p may be determined The transmission bit error rate p may be determined duration T = 16s; thus, the entire preamble time dura- from the expression, tion is a mere 2T = 32s. The preamble and (encoded 3 2 and modulated) message time allocations are shown in P=p +3p (1 - p) Figure 3. The preamble tonals for each phase are - -3 x14 located utilizing the processing shown in Figure 4; tonal detection/location unanimity is required for the two and phases.t The initial doppler estimate so obtained is -2 then employed to assign a total of 20 phase-difference frequency trackers to the anticipated message spectral Finally, the transmission character error rate Pe for bins (four per diversity, 5-fold diversity). Shortly M = 4 FSK signaling is related to the bit error rate of after the transmitted signal onset, 15 of the trackers Equation (15) by, may be disabled, with the remaining five tracking the encoded and modulated message bits. = 2(M- 1) e M -2 _ 5.0 x 10 (16) PHASE I PHASE 2 CODEWORD CODEWORD 2 (DO0ITS) (SITS) The character error rate is directly related to the required SNR at the decision point. For M = 4 FSK ' I signaling, ideal detection, and 5-fold frequency diver- sity, the shortened Golay coding scheme requires the '.~ 325 Z ] ~ 10.7 mln , following SNR value per tone for (Rayleigh) fading conditions [3], PREAMBLE TIME TRANSMITTED SIGNAL TIME DURATION S (DURATION (ENCODED AND MODULATED MESSAGE) 4.5 dB (17) Figure 3. Preamble and Message Time Allocations The total required SNR is, of course, 11.5 dB. Under fading conditions, an additional allowance of 2. 0 dB is necessary to allow for non-ideal filters, timing error, waveform distortion in the medium, and AGC and noise t The same reliability philosophy is employed here as shifts in the receiver. The required S- per tone is then is in using two codewords to transmit the message and 6. 5 dB to meet the (conditional) error probabilities; requiring that both be identically decoded. 69 As shown in Figure 4, the tracked encoded and 4. Performance Prediction modulated message bits are diversity combined, Golay decoded, and thresholded, resulting in the desired 8-bit block of information bits. A matter of concern is FOM = Ls - [LA - (DI - DA)] the capability of the frequency trackers to deal with the expected frequency slew rate. Exceedingly worst-case [LN - (LSNR - DN)] -(DT + DS) CPA dynamics calculations indicate that the slew rate s (19) is bounded as follows. where LA is the ambient noise spectrum level, DA is s 1.7 x 10-3 Hz/s (18) the degradation to the theoretical acoustic receive array The slew rate of Equation (18) is computed using a directivity index DI, LSN is the self noise level, LSNR is the theoretical acoustic array discrimination against relative end point range rate of 30 kn, a closest point of approach of 10 nmi, and a frequency of 200 Hz. The self noise, N is the degradation to the LSNR, DT is the detection threshold, and DS is the processing sys- simple phase-difference tracker algorithm envisioned tem loss. The symbol ) denotes power level addition. for the receive signal processing is capable of a slew rate of 3.1 x 10-3 Hz/s at a SNR of 3 dB per spectral The total background noise spectral level LN bin (a spectral bin Af = .125 Hz; cf. Figure 2). We (bracketed term of Equation (19)) has been measured note that this SNR figure is less than the 4. 5 dB of at the broadside beam output of a representative Equation (17), required for acceptable (ideal) detection/ receive array for several speeds and a frequency of decoding. No problem is then anticipated with tonal 200 Hz.t The values obtained are shown in Table 1. tracking during the transmitted signal time duration. Table 1. Total Background Noise Spectral Level for Several Speeds, 200 Hz Speed (kn) LN (dB//lFPa/-Hz- ) raRDWARE OFTWARE - - - - - - - - - 3 57. 0 6 58.8 ACOUSTIC RECEIVE ARRAY INPUTS l l 8 62. 0 10 68.0 l l BPF l l 15 79. 1 E | SGA l= | | The transmitted source level per tone is taken to be, o H l l l lXLS = 188.0 dB//lpPa . m (20) ,.-*e o..[ ___.| implicitly assuming an omnidirectional projector. The value of Equation (20) is also selected as a value achievable by existing equipments.t t The detection l | DOsEMMPOODX | | threshold per tonal may be calculated from the relation, I ,64H m DT = MDL + 10 log (Af) - RG (21) ?NN AK where Af = 0. 125 Hz is the receive frequency resolu- | ' I ' 2 | tion, MDL is the minimum detectable level per tonal, I -" 25j " and RG is the redundancy gain. The MDL is obtained from Equation (17) plus 2. 0 dB to account for non-ideal detection; thus, the MDL = 6. 5 dB per tone. We assume NORM-----n_ DIVERSITY RG = 1. 0 dB, nominally associated with a redundancy COMBINE I factor of from 2 to 4. Inserting the above values into l ?L-" -] Equation (21), we obtain, EST COPP DECODE DT = -3. 5 dB (22) HHRES3OLO | t The ambient noise spectral level included in the total background noise spectral level is consistent with that -L__- ___ _of the acoustic province MGS 3 (57N, 40W), S=30 ft (9. im), R=300 ft (91. 4m), and the summer season. Figure 4. Receive Signal Processing Functional Diagram ItfMore precisely, the total required source level LS + 10 log M is achievable by existing equipments. 70 Contribution to the factor DS principally arises from 5. Conclusions operator loss. We assume the operator loss is zero; A design has been presented for command and con- accordingly, we set DS = 0 dB in the ensuing calcula- sinas been resh and trol acoustic communications between a surface ship and tions. ~~~~~~~~~~tions. ~a number of small fast submersibles. The design offers Based on the above discussion and the data pre- a reliable command and control capability out to the sented in the form of Table 1 and Equations (20) and desired range of 100 nmi for submersible speeds up to (22), we compute the FOM values of Table 2. 10 kn, in addition to a simplicity and economy of imple- Table 2. FOM and Signal Excess at 100 nmi mentation. It is felt that the results presented here will for Several Speeds, 200 Hz prove useful to the command and control of commercial Speed (kn) FOM (dB) SE (dB), 100 mni submersibles devoted to mining, exploration, and bathy- metry tasks and to military submersibles dedicated to coastal protection and enemy surveillance activities. 3 134.5 15.5 6 132.7 13.7 References 8 129.5 10.5 10 1213, 5 64.5 [1] Peterson, Wesley W. and E.J. Weldon, Jr., ~~~15 112.4 - 6.6 ~Error-Correcting Codes, The MIT Press, Cam- bridge (1972). To complete the picture, transmission loss calcula- [2] tions have been made utilizing the FACT and NISSM II Berlekamp, Elwyn R., Algebraic Coding Theory, models with coherent multipath combination and the McGraw-Hill, New York (1968). geometrical and environmental parameters previously [3] mentioned. The peak transmission loss in a 50 nmi Nuttall, Albert H., "Error Probability Character- window centered about 100 nmi was computed as istics for Orthogonal Multiple Alternative Communi- approximately 119 dB. The signal excess SE at 100 nmi cation With D-Fold Diversity," NUSC/NL TR 4769, is also shown in Table 2. We see that it appears possi- 19 June 1974. ble to provide reliable command and control capability out to the desired range of 100 nmi for submersible speeds up to approximately 10 kn. 71 FREQUENCY SHIFT ARRIVAL TIME DETERMINATION OF ACOUSTIC SIGNALS by Edward W. Early John E. Ehrenberg Roundney K.B. Liem Applied Physics Laboratory, University of Washington 1013 N.E. 40th Street, Seattle, Washington 98105 Abstract under the particular ambient noise condition with- out saturating the receiver. Since the receiver is used in a synchronized system, a time-varied In the demodulation of underwater acoustic gain is incorporated to compensate for the de- signals, timing determination is usually based on creased signal level as the distance increases. the leading edge of a pulse, and is therefore sub- ject to errors that are functions of pulse rise time and signal level. This paper describes a system using a phase-locked loop demodulator to 0 Z determine the time of arrival of a frequency- BANDPASS E shifted pulse. In this system, timing is shown to AMPLIFIER be essentially independent of signal strength, and timing jitter is shown to be low. TVG RAMP --M X GAIN GENERATOR M TAMPLIFI ER GENERATOR CONTROL LIMIT & FILTER Introduction Many underwater acoustic systems use pulse modulation techniques in which detection occurs at the leading edge of the pulse. This method leads to timing errors that are functions of the pulse rise time and the received signal level, in addi- tion to timing perturbations due to a low signal- to-noise ratio. In a system developed by the THRESHOLD LOW PASS LOCKEDLI DETECTOR FILTER LOOP Applied Physics Laboratory to track Navy divers, DETECTOR the total error had to be kept within �S ft even though the signal strength might vary widely. Because it was recognized that the medium itself might introduce considerable errors, it was sought DIGITAL to keep receiver timing errors to a minimum. INTEGRATION I NTEGRATION OUTPUT The acoustic system developed to fulfill that CLOCK CLOCK requirement uses a pulse in which the carrier fre- FREQ. quency is shifted several kilohertz midway through its length. At the receiver, the signal is ampli- Figure 1. Functional block diagrcm of receiver. fled and presented to a phase-locked loop (PLL) which detects the frequency transition. By proper choice of bandwidth and pulse length, the midpoint The amplified signal is passed to the detec- of this shift can be made relatively independent tor channel where the timing information is de- of signal strength. The remaining timing pertur- ciphered. A clipper removes the lower amplitude bations are then a function of signal-to-noise noise. The signal is then band-limited in the ratio only. phase-locked loop and the transition edge is de- termined. The output of the PLL is then low-pass ~The Receiver ~filtered. Two criteria are imposed on the pulse after this step. It must meet threshold and mini- A functional block diagram of the receiver is mum pulse width requirements. The latter is shown in Figure 1. An acoustic pulse in the water determined by digital integration. is picked up by the receiver hydrophone and ampli- fied together with the in-band noise. The noise The in-band noise that is not removed by the component is then integrated and used to adjust clipper is generally of shorter duration than the the gain of the receiver to the maximum possible pulse and would be removed by the minimum pulse- width requirement. 72 Receiver Application of the PLL n Ct dy(t) 4 In the receiver just described, a phase- locked loop is employed to detect a 1-ms frequency- dt shifted pulse centered at 31.25 kHz. The pulse is transmitted at 34.48 kHz for twenty cycles and The slope of the output is approximately equal to switched to 28.57 kHz for twelve cycles, which the size of the frequency step fA divided by the completes the 1-ms period. rise time tr. Therefore The phase-locked loop is tuned to a free-run- 6t c ning frequency of f0 = 31.25 kHz, and the capture = n r n range is set to include the high and low frequency t A 2Wf extremes of the pulse. During the reception of a A frequency-shifted pulse, the PLL locks onto the high-frequency portion within a few cycles. At the frequency transition, the PLL follows the The variance of the timing error is frequency shift almost instantaneously. This results in a negative ramp in the output of the 1 2 phase detector. In the presence of noise and , (6a reverberation, this frequency transition is the A most stable timing edge of the signal. Note that the PLL is always locked during the frequency where on is given by Eq. 2. The standard devi- transition. ation of the resultant timing error is Timing Error Due to Noise '3 The phase-locked loop functions as an FM 3t + 2Wf~ *vk + 2WfAexp[-R] ' (7) discriminator. The output of the low-pass filter following the loop can be written as For the system being considered here, W = 3.3 kHz, Y(t) = s(t) + n(t) , (1) B =l kHz, and fA 6 kHz. where s(t) is the noiseless output and n(t) is additive noise. The additive noise at the output of the discriminator can be attributed to two effects. When the signal-to-noise ratio is high, X'_NOISELESS the additive noise causes fluctuations in the OUTPUT envelope of the input signal. These fluctuations OISY in turn result in noise at the output of the OUTPUT discriminator. Secondly, the input noise can also TIMING THRESHOLD result in some spurious zero crossings in the input signal. The spurious zero crossings pri- marily occur when the signal-to-noise ratio is low, and cause noise spikes at the output of the discriminator. An expression for the variance of the output noise that accounts for both effects isl = w2 + 2Wf~exp!I~R)Figure 2. Effect of noise on the output On + 2Wf exp[-R] (2) of the FM discriminator. $ BRA where fA is the magnitude of the frequency step, R Experimental Analysis of the Receiver is the signal-to-noise power ratio out of the bandpass filter, and B and N are the bandwidths of In determining the timing errors of the re- the bandpass and low-pass filters, respectively. ceiver, we used the experimental setup shown in Figure 3. The sync pulse triggered a frequency- The effect of noise on the output of the FM shifted pulse and, at the same time, started the discriminator is shown in Figure 2. The additive time interval counter. The receiver processed the noise causes an amplitude error En which in turn pulse and generated a digital signal to stop the results in a timing error et. If the signal-to- counter. In order to compare the frequency-shift noise ratio is large, the following approximations method of time determination and the leading-edge can be used: method, the PLL was temporarily converted to an amplitude detector for another series of timing dy(t) = error tests. 6 ~~~~~~~~(3) t dt n The results of the tests using the frequency- P or shifted pulse and the PLL detector are tabulated 73 in Table I and the results using a single frequen- the leading-edge detection method, the timing ncy burst and an amplitude detector are tabulated error is 336 Vs over the 40-dB range of S/N ratios in Table II. shown whereas with the frequency-shift method it is 10 ps (see Tables I and II). In theory, X is constant as a function of the S/N ratio for the frequency-shift method. When timing errors are FREQUENCY |DECADE important and the dynamic range of the received GENERATOR ATTENUATOR signal is large, the improvement shown by use of the frequency-shift method would be very signifi- cant. The difference between the leading-edge average timing and the frequency-shift average timing is due to the fact that the frequency-shift RANDOM 6-kHz DECADE detection point comes at approximately 600 ps ENERO LBANDPASSE ATTENUATOR _ whereas leading-edge detection occurs at the start GENERATOR FILTER of the pulse. 1,000 - START TIME STOP COUNTER C 900 - SYNC. PULSE 800 - 800- GENERATOR , FREQUENCY SHIFT Figure 3. Experimental setup for determination :i 700- * LEADING EDGE of timing errors. IE Table I. Frequency-shift detection. 600 - S/N Ratio X 0 (dB) (Us) (Us) 500 - Z 10 942 18 < 13 933 4 16 929 6 20 927 3 400 - 30 934 1 40 934 1 50 933 1 300 - Table II. Leading-edge (amplitude) detection. S/N 0' ' l ' l Ratio X a0 10 20 30 40 50 (dB) (Is) (Ps) S/N RATIO (dB) 10 672 47 13 589 38 Figure 4. Mean arrival time. 16 491 25 20 410 20 30 367 17 Figure 5 plots the standard deviations of the 40 346 11 timing errors with the two methods. There is good 50 336 6 agreement between the frequency-shifted experi- mental data and theory except at the higher S/N The average time from the trigger pulse to levels where the receiver has become saturated. the arrival time X for these two situations is It should be noted that a I-ps standard deviation plotted in Figure 4. These data show a signifi- is insignificant in most underwater acoustic sys- cant minimization of the timing error when the tems. The leading-edge data show a considerably frequency-shift detection method is used. With greater timing deviation. 74 100 - 70 - @ FREQUENCY SHIFT 50 - \ r LEADING EDGE 30 - El~~~[ \ El [] 10 - 7-. I�� 0.7 0.5- 0.3~~ b 3-X * ~~~0.7 - O0.5 - * ~ 0.3 -\ THEORETICAL PLL PERFORMANCE AS CALCULATED FROM EQ. 7 0.1 I I I 0 10 20 30 40 50 S/N RATIO (dB) Figure 5. Standard deviation of timing errors. The foregoing presentation has shown the decided improvement of the PLL detector over the leading-edge detector--both in the reduction of mean timing error and in the reduction of timing deviations (jitter). Acknowledgment This work was supported by the Naval Facili- ties Engineering Command under Contracts N62477- 75-C-0364 and N00017-74-C-1208. Reference 1. R.E. Ziemer and W.H. Tranter, Principles of Communications, Houghton-Mifflin Co., 1976. 75 SHIPBOARD EVALUATIONS OF DRAG COEFFICIENTS FOR THE MARINE REGIME C. W. Fairall, G. E. Schacher, K. L. Davidson, T. M. Houlihan Environmental Physics Group Naval Postgraduate School Monterey, California 93940 constant at higher velocities. Cardone (1970) Abstract has proposed a velocity dependence of C based Abstract ~~~~~~~~~~~~~~~~~10 on Wu's work and Charnock's (1955) relation for A considerable amount of overwater dissi- aerodynamically rough flow. Cardone proposes pation and wind profile data obtained from that a minimum in C occurs at wind velocity shipboard measurements were interpreted rela- tive to the representation of the drag coeffi- U10 6 rnsec due to the transition from smooth cient. Drag coefficients were interpreted to rough flow. The velocity at which this mini- relative to wind speed, wave and hydrostatic mum occurs is related to the minimum free surface stability influence. It was noted that at all ocean wave phase velocity. On the otherhand, wind speeds and particularly at low wind speeds, Kraus (1972) lists five investigations that showed drag coefficient values are larger when large no significant velocity dependence for C In waves are present. Likewise, it was noted that in unstable conditions drag coefficient values addition, Cardone showed that ClO should be larger increase as neutral conditions are approached. for unstable conditions than for stable conditions. This is in agreement with the data of DeLeonibus (1966) and Hasse (1968) but contrary to the re- 1. Introduction sults of Smith (1970). This paper will examine both of these issues. Since it is difficult to Drag coefficient is one of the important see how drag coefficient could be a strong func- quantities for characterizing the air-sea tion of sea state and also be, on the average, interaction. A number of calculations of drag independent of wind velocity (in other words, a coefficient were made from values of the rate lack of dependence on wind velocity implies a of turbulence dissipation, e, measured during lack of dependence on sea state), an examination research programs on optical propagation, fog, of wave height effects on drag coefficient is and turbulent transport over the ocean. The also presented. data was taken during shipboard operations in the Pacific and Atlantic oceans and the Mediter- 2. Theory ranean Sea. The bulk of the data is from two cruises: The first was the "Cooperative Exper- A thorough treatment of the boundary layer iment in West Coast Oceanography and Meterology, equations exists in Lumley and Panofsky (1964), 1976" (designated CEWCOM 76) aboard the RV Businger (1971) or Kraus (1972). The Reynolds Acania, primarily a marine fog investigation off stress, r, is represented by southern California. CEWCOM 76 was organized 2 by the Naval Postgraduate School and the Naval Tip -<ulwl> = U2 Ocean Systems Center. The second cruise was aboard the USNS HAYES (designated HAYES 77), where u' is the fluctuating horizontal wind organized by Naval Research Laboratory to study velocity, w' is the fluctuating vertical wind aerosols and turbulence in the Atlantic and velocity, U* is the friction velocity and p is Mediterranean. A small amount of data taken at the density of air. The height dependence of the anchor in Monterey Bay from the RV Acania using mean horizontal wind velocity U and the rate of a bow-tethered buoy (designated BUOY 77)' that turbulent energy dissipation s are permitted measurements closer (- 1 meter) to the sea surface is also included in this report. KZ DU f (R.), (2) Atmospheric turbulence and the air-sea U* 2z M interaction have been under investigation for a KZs number of years. Wa (1968) has compiled drag = f (R.) coefficient data from investigations back to U*3 1876 and compared them to laboratory work. He found the oceanic ten meter drag coefficient where K = .35 is the Von Karman constant. Z is (C10) to increase with increasing wind velocity the height above the surface, Ri is the Richard- upt bu 10 =15/sadtorai up to about U lo=M 15 m/sec, and to remain son number, and f and fE are stability 76 corrections, equal to 1 in neutral conditions. The mean data from all levels were logged with a The Richardson number describes the atmospheric mini-computer data acquisition system. Averaging stability in terms of temperature and velocity times were usually 20 to 30 minutes. In addition, gradients, routine weather observations were made including estimations of wave height. R. = -~ (_v/_Z Ra = 2 V (4) 4. Data Analysis Velocity fluctuation data from the FM tape where g is the acceleration of gravity, T is the was evaluated using a fourier spectrum analyzer. absolute temperature, and 8 is the virtual poten- Assuming k = 2Jf/U, Equations 4 and 7 yield V tial temperature. In near neutral conditions 1/3 1/2 U, = (2TrKz/U) (f u(f)/c) 1, (8) U* = (:KZO11 (5) where f is the frequency in Hertz. A real-time F and the drag coefficient at height Z is analysis during the HAYES 77 cruise was obtained by measuring the RMS value of the velocity fluctu- C = (U/U)2 (6) ations in a frequency band specified by a lower frequency limit, f�, and an upper frequency limit, and is normally taken at a reference height of 10 f . This can be related to the spectral density u meters. Calculations of C10 reported herein are by based on determination of U* from e using the Kolmogorov representation of the one dimensional f k dk = <u' >. (9) velocity power spectral density Cu(k), o Since we are measuring the RMS of u' between fre- c~u (k) = ar2/3 k513 (7) quency limits, we must evaluate the integral between these same limits, where a = .52 and k is the wave number. This is an indirect method of obtaining drag k coefficient and several authors have found that u (k) dk = <u'2> (10) it gives 20% to 40% higher results for C than k� those from flux or direct stress measurements. This discrepancy is attributed to a breakdown Again using k = (2irf/U) we can write the friction of the "local isotropy" assumption inherent in velocity as Equation 7. Unfortunately, small ships are non- stable platforms and a high frequency velocity ..1/3 /(3a) ]l/2u (.11) fluctuation determination of t (via Eqn. 7) is U, =L -2/3 -2/3 EMS. the only practical method of avoiding the plat- z u form motion problem. Equation 8 is referred to as the spectral method 3. Instrumentation and Equation 11 as the RMS method. During HAYES 77 an f = 2 HZ and f = 200 HZ were selected. The velocity fluctuations were measured with A comparison of the two techniques for the HAYES cylindrical hot film sensors using a Thermo- 77 data is given in Fig. 1. Systems type 1054 constant temperature amemometer. The velocity signal was amplified, filtered, and Since the HAYES 77 and CEWCOM 76 cruises were recorded on FM tape. On the HAYES 77 cruise the moving ship operations, U was not measured di- IQ data was real-time analyzed in a manner described in Section 4. Mean temperature was measured with m ents For with 76 pedic 2 is an ments made with the ship stopped (Fig. 2 is an an HP model 2850 B quartz thermometer, accurate to .010C. The relative humidity was measured with eame) were used For HAYErpola s measurements were effected. For HAYES 77, U1 was Hygrodynamics hygrosensors (Dunmore type lithium 10 chloride), accurate to about 3% RH. Mean winds calculated from the relative wind speed and were measured using Thornthwaite cup anemometer direction and the ship's speed (hand recorded from units, accurate to about 4%. One of each of the a convenient readout). Both methods lead to con- instruments was operated at each level, with an siderable errors in U10. The BOUY 77 data is much additional quartz thermometer for sea surface more accurate since the system is not moving and temperature. The measurement heights for the the sensors are closer to the surface where the three cruise were as follows: turbulence and gradients are greater. Fig. 3 Height in Meters shows the single measurement standard deviation, CRUISE LEVEL I LEVEL II LEVEL III LEVEL IV a , for C10 (calculated from U, averaged over all levels) from the BUOY 77 data and the combined HAYES 77 and CEWCOM 76 data. It is interesting BUOY 77 1.0 3.0 6.6 that for the BUOY 77 data, the c based values for HAYES 77 10.5 21 U, are lower than the profile values (Fig. 4). HAYES 77 10.521 Except for Fig. 4, corrections for stability were 77 not made since data occurs predominately for Drag coefficient (C10) versus wind speed neutral conditions. Fig. 5 shows the probability distribution of the Richardson number (at Z = 10 10 meters) for 523 values obtained over the three tion which characterize a smooth flow regime, C10 cruises. decreasing with increasing U, at low wind speeds 5. Results and rough flow regime, C10 increasing with in- creasing U10, at higher wind speeds. The minimum Average values of C10 for the three data 0 value was near 55 m sec a d C10 alu wa ner 55 msecand the distribu- sets are shown as a function of wind velocity in 10 Fig.6and are combined in Fig. 7 to compare with tion agreed reasonably well with a representation by Cardone (1969). Cardone's curve. The error bars represent the expected error for the average (t CT/ V, N = A C10 versus U10 representation in which C10 number of points). The minimum value of C10 oc- values were grouped in three wave height cate- curs at U1 5 5. m/sci odareetwt curs at U 10 ~ 5.5m/sec in good agreement with gories indicated that at all wind speeds and par- Cardone's proposed minimum of 6 m/sec. The ticularly at low wind speeds the 010 values are average for all 445 data values is C0 larger when larger waves were present. 1.55 x lo and a/ IN-= .05 x I1-3. A C10 versus V10 representation in which C10 A part of the normal operation weather obser- were grouped in two stability categories indicat- vations were "visual" estimates of predominate ed that in unstable conditions, 0.10 values wave height categories and the C10 vs U10 increase as neutral conditions are approached. analysis (Fig. 8) was repeated. Kraus (1972) argues that there should be little, if any, sea Acknowledgements state effect on C10 on the grounds that the wind The authors wish to recognize the technical interacts primarily with the high frequency part and operational contributions of Jim Corbin and of the spectrum. The measured rapid increase of Lyn May. This work is supported by NAVSEA PMS C10 for decreasing velocity with larger waves 405 and NAVAIR. is definitely statistically significant, although it could be due to increasing violation of the References local isotropy assumption. 1. Brocks, K. and Krugermeyer, L., 1970: The One of the more interesting roughness of Sea Surface. dat istheshft f te010 vU10 crewt data is the shiftofthe vsU curvewithInstitute for Radiometeorologie and Maritime stability. Despite theoretical predictions, Meteorologie Report No. 14, Hamburg University. lower values of C0 were found for unstable con- 2. Businger, J. A., Wyngaard, J. C., Izumi, Y. ditions (Fig. 9). In this graph, U* elements and Bradley, E. F., 1971: Flux-Profile Rela- used to determine C10 were corrected for stability tionships in the Atmospheric Surface Layer. using f (Ri) from Champagne et al (1977). J. Atmos. Sci. 28, 181-9. This unexpected result in relative magnitudes 3. Cardone, J. J., 1969: Specifications of the of neutral and unstable C10 values also occurred Wind Distribution in the Marine Boundary in a considerable number of overwater data de- Layer for Wave Forecasting. New York Univer- scribed by Deleonibus (1971). He argued convinc- sity, Scientific Report CSL-TR69-1, University ingly that differences from predictions were due Heights, New York. to wave influence. 4. Champagne, F. H., Friehe, C. A., LaRue, J. C., 6. Summary and Conclusions 1977: Flux Measurements, Flux Estimation Techniques and Fine Scale Turbulence Measure- A considerable amount of overwater dissipa- ments in the Surface Layer Over Land. tion (c) and wind profile data obtained from J. Atmos. Sci. 34, 515-530. shipboard measurements were interpreted relative to the representation of the drag coefficient. 5. Charnock, H., 1955: Wind Stress on a Water Analyses procedures used were those which are Surface. Quart. J. Roy. Meteorol. Soc. SI, practical in view of inherent difficulties in 639. shipboard measurements of both fluctuating and mean parameter. Drag coefficients were interpre- 6. DeLeonibus, P. S., 1966: Momentum Flux Obser- ed relative to wind speed, wave and hydrostatic vations from an Ocean Tower. U. S. Naval stability influence. The small scale velocity Oceanographic Office (Unpublished) Washingta; fluctuation data used to define e was not con- D.C. sidered to be influenced by ship motion. Reason- able definitions of hydrostatic stability as well as visual estimates of wave conditions were used in the interpretations. 78 7. Hasse, L., 1968: On the Determination of the Vertical Transports of Momentum and Heat in the Atmospheric Boundary Layer at Sea. Hamburg 05 Geophysikalishe Einzelschrifen, Heft II. 0.5 - 8. Kraus, E. B., 1972: Atmosphere - Ocean Interaction. Oxford, Clarendon Press. 04 - 0 9. Lumley, J. L. and Panofsky, H. A., 1964: The Structure of Atmospheric Turbulence, New York, 0 3 Inter-Science. 10. Smith, S. D., 1970: Thrust-Anemometer E0.2 Measurements of Wind Turbulence, Reynolds, * Stress and Drag Coefficient over the Sea. J. Geophys. Res. 75 (33), 6758-70. 11. Wu, Jim, 1969: Froude number Scaling of Wind- Stress Coefficients. J. Atmos. Sci., 26, 0 0. 02 03 0.4 05 408-13.O 408-13. 0 0,1 0.O.I 2 0.3 0.4 0.5 U* SPECTRUM (m/sec) 12. DeLeonibus, P. S., 1971: Momentum Flux and Wave Spectra Observations from an Ocean Tower, J. Geophys. Res., 76, 6506-27. Figure 1. A comparison of the RMS vs the SPECTRUM method for measuring U* during HAYES 77. The error bars represent i a/ i. 10.0 9/24/76 9.0- 8.0- 7.0- 60 - E 4.0 _O3.0 2.0- 1.0- I I I I I I I I I 0 2 4 6 8 10 12 14 16 18202224 LOCAL TIME (hrs) Figure 2. Mean wind velocity at Z = 10 meters vs local time for the day 9/24/76 during CEWCOM 76. 79 30 66 JHAYES 77 2 -CEWC0M 76 2.0 - x BOUY 77 4.0 - 3.0 - r o0 2.0 - 'o I.O - ~~~~80b 40 5 2 1.0 - 41 5 if 60 1 X- I A2~ 3-1.5 -1.0 -0.5 0 0.5 1.0 Ii IiI I 0 2.0 4.0 6.0 8.0 10.0 R. U10 (m/sec) Figure 3. Standard deviation, a , vs U . .' Figure 5. The probability density of occurrence 10e number of values averageof Richardson number, p(Ri) vs R., The number of values averaged all data combined. are shown for each point. U.-WZ0C~t3 E Z-mZ'3.Om a Profile U.-KZOU/az 6.0 - A Z.5m (U.) o HAYES 77 2/15 2/16 2/17 Profile .120 0- -level I .120 level 2.0586 .18 - a level3 .093 x CEWCOM 76 .6-0 a-aR40 - .16- 3 0 3 A 4 .10 ox AOU~ ~~~~~~~~~ 0 *. .12- x A 0 oA f .0 A 2xUAA~ .0- e 00 .04 - x 3 .06- cI ve .0- :4- - A .0 - .00 -J---J-- -j 15 16 17 14 15 16 17 1I 14 15 16 17 18 19 TIME(hrs) 0 2.0 4.0 6.0 8.0 10.0 Figure 4. Friction velocity, U*, vs local time U (m/sec) for BUOY 77. IC) Figure 6. Drag coefficient vs wind velocity for each data set. The error bars represent � a/A. 80 6.0 - -CARDONE 5.0 - 4.0 - 3.0- o 2.0- - 740 3521 o 7757 5 4 86 55 0 , I ' I i I , I , I 0 2.0 4.0 6.0 8.0 10.0 6.0- 0-0. IRi 0.1 UiO (m/sec) - 05<R<-0i 5.0 - Figure 7. Drag coefficient vs wind velocity, all data. The error bars represent 4.0 - * o/+I. The number of values averaged are shown for each point. '6 3.0 - 0 co 2.0- 6.0 - oW< ft 10 \ -----xW = 4ft 5.0 - - 0 2.0 4.0 6.0 8.0 10.0 4.0- \ UO (m/sec) 3.0- \ O. \r, Figure 9. Drag coefficient vs wind velocity for different categories of 0 2.0 - 1 atmospheric stability. I.O - 0 0 2.0 4.0 6.0 8.0 10.0 U 0O (m/sec) Figure 8. Drag coefficient vs wind velocity for different categories of wave height, W. The error bars represent � a/Vi. 81 THE MECHANICAL PROPERTIES OF CHITOSAN MEMBRANES R. B. Clark B. L. Averbach Graduate Research Assistant Professor of Materials Science Department of Materials Science and Engineering Department of Materials Science and Engineering Massachusetts Institute of Technology Massachusetts Institute of Technology Cambridge, Massachusetts 02139 Cambridge, Massachusetts 02139 ABSTRACT in weak acetic acid and have investigated the effects of the deacetylation variables on the The mechanical properties of chitosan mechanical properties of the resulting films. membranes have been investigated as a function of Although the properties of chitin membranes several processing variables. Chitosan, the have been studied, there has been very little deacetylated derivative of chitin (N-acetyl-D- research on the mechanical properties of glucosamine), was processed from blue crab shells chitosan films. Chitosan membranes are from Virginia. Various chitosans, which had been potentially useful as a wrapping material, and deacetylated under different conditions, were these data on the mechanical properties may dissolved in weak acetic acid and then cast into serve as basic information for such applications. clear, flexible membranes about 50 pm thick. The stress-elongation curves exhibited a fairly constant elastic modulus, a yield point, and a 2. RAW MATERIAL high tensile strength (avg. 135 MPa). The mechanical properties were not greatly affected Nine samples of chitosan were produced from by the deacetylation procedures investigated the bottom shell of blue crabs by Mr. Peter here, although a small decrease in yield strength Perceval at the Hunt Crabmeal Company of Hampton, and an increase in elongation were observed at Virginia. Samples A through F were demineralized the highest temperature used (1000C). in 5% HC1 for 16 hours at room temperature and deproteinated in 2% NaOH for 2 hours at 850C. For samples G, H, and J, the time of demineralization 1. INTRODUCTION was reduced to 5 hours, and the temperature of deproteination was lowered to 650C. There were Chitin is the structural component in the no significant differences in properties due to shells of crustaceans. Chitosan is formed by the these changes. deacetylation of chitin, but chitosan is not a unique product because of variations in the raw Table I shows the variations in deacetylation materials and in the processing procedures. parameters among the samples. Samples A, B, and C Chitin itself varies from one crustacean species were made from the same batch of chitin and were to another and even from part to part in the same identical except for the time of deacetylation. animal. Also, the literature indicates a wide Samples D, E, and F were processed to check the variety of processing procedures, and these effects of NaOH concentration. The last three, variations result in chitosans with different G, H, and J, differed mainly in the temperature properties. In the present work we utilized only of deacetylation. The chitosans used in this the bottom shell of the blue crab from Virginia. study were low in impurity content. Calcium Each phase of the processing was carefully averaged 0.02% (by weight) and never exceeded controlled. First, the crushed shells were soaked 0.06%. Residual ash values ranged from 0.02% to in dilute HC1 to remove minerals, primarily 0.17% and averaged 0.1%. calcium carbonate. Next, a weak NaOH solution was used to remove the intrinsic protein. The resulting chitin was deacetylated in strong NaOH 3. EXPERIMENTAL PROCEDURE solutions at elevated temperatures. This last step, deacetylation, is the most difficult. If Eight grams of chitosan were mixed with the conditions are too mild, the product is 784 grams of distilled water. After the addition insoluble, due to residual chitin. If the of 16.00 grams of acetic acid, the solution was conditions are too severe, the chitosan polymer stirred for 10 minutes and allowed to stand for chains are degraded. This work addresses the at least 12 hours. Large insoluble particles effects of such degradation on some of the were filtered out with nickel mesh (#40). The properties of chitosan. solution was maintained at 200C, and viscosity measurements were made using spindle #1 of a Chitosan is an excellent film former. We have Brookfield LV viscometer. The solution was then cast membranes from various chitosans dissolved centrifuged through a TFE filter having 30-60 pm 82 TABLE I CHITOSAN SAMPLES Viscosity* Mw+ Sample Deacetylation procedure (N.s/m2) (106) A 50% NaOH, 2 hrs., 100'C .652 1.012 B 50% NaOH, 3 hrs., 100�C .476 .997 C 50% NaOH, 4 hrs., 1000C .297 .732 D 41% NaOH, 4 hrs., 100�C .148 .634 E 45% NaOH, 3 hrs., 100�C .142 .635 F 50% NaOH, 3 hrs., 100�C .332 .886 G 50% NaOH, 4 hrs., 75�C 2.070 1.298 H 50% NaOH, 3 hrs., 90�C 1.010 1.094 J 50% NaOH, 3 hrs., 100�C .354 .887 * Solution containing 8 gms chitosan, 16 gms acetic acid, and 784 gms of water. + Weight average of the molecular weight. pores. (Sample D had so many insoluble particles 4. RESULTS AND DISCUSSION that this step was omitted.) After a thorough stirring, the solution was poured into a level Figure 1 shows the characteristic shape of glass tray (29 cm square) and dried with forced the stress-elongation curve of thin chitosan air. When the film was dry (approximately films. The two almost linear regions were also 40 hours later), it was soaked in 1 M NaOH and noted by Joffe and Hepburn for aged regenerated then rinsed with distilled water to neutrality. chitin films (1). The sharpness of the yield The wet membrane was placed on a celluloid sheet point seemed to improve with the uniformity of and dried with forced air. The result was a film thickness. Some of the thinner films also clear, flexible membrane about 50 pm thick. exhibited a slight yield drop just past the yield point, as shown in Figure 1. The mechanical testing procedure closely followed ASTM Method D882. The films were The modulus of elasticity was difficult to conditions at 23� + 3.0�C in 30-50% relative determine accurately because of the relatively humidity for at least 40 hours prior to testing. high strain rate employed. 194 strips had an Test strips 1.27 cm in width and about 16 cm in average value of 6300 MPa with a standard length were cut with a razor blade. 223 specimens deviation of 710 MPa (based on minimum rather were obtained from 23 films. The thicknesses than average areas). Joffe and Hepburn reported were measured to + 1 pm with a dial micrometer, values close to 2000 MPa; Yaku and Yamashita and the minima were marked with a felt-tip pen. obtained moduli near 1200 MPa for acetylated (Due to the presence of insoluble particles, chitosan films (2). sample D had to be measured optically in a microscope, resulting in a loss of accuracy.) The tensile strength varied widely from The strips were examined for flaws and then strip to strip but showed no significant placed in rubber-faced grips with a 10 cm length differences among the batches. The average between grips. Tensile tests were made at a value, 135 MPa (standard deviation: 18 MPa), strain rate of 0.5 per min. at room temperature was based only on the 166 samples that broke in 30-50% relative humidity. The films were at or near their marked minima. This result observed during testing to see if yield and compares favorably with the value of Muzzarelli failure occurred at the minima and if flaws et al of 69 MPa for chitosan films (3). One of influenced the results. us (B.L.A.) earlier reported a tensile strength of 152 MPa for a chitosan film prepared in a manner similar to the present method (4). However, these values are all below the 590 MPa reported for horseshoe crab chitin by Rutherford and Austin (5). 83 Figure 1. Stress-elongation curve Table II lists the yield strengths of the characteristic of chitosan membranes samples. Examination of the results for (strain rate: 0.5 per min.). samples A - C reveals no clear trend of yield strength with respect to deacetylation time. Thus, polymer degradation, as judged by viscosity and molecular weight (Table I), had no effect on 1 50 - the yield strength of samples varying in time of deacetylation. For samples deacetylated with varying NaOH concentrations (D - F), no trend could be established since the high yield stress of D was probably due to the inaccurate thickness measurements. On the other hand, temperature variations 0ioo- appeared to have a slight effect on the yield O [O -strength. This is shown graphically in Figure 2. EL The yield strength remains constant for the lower temperatures but drops between 90� and 100�C. '-'~~~~~~~~~~~ gThis slight trend however is not comparable to c ~ ~~~~n I ~~~~~the large shift in molecular weight illustrated in Figure 3. |~-~~~~~~~~~ ~The total elongation of the chitosan strips ()50 tis shown in Table III. Samples A, B, and C again show the ambiguous effect of deacetylation time. More study will be required to confirm an increased elongation for chitosan deacetylated for 3 hours (sample B). The apparent trend due to NaOH variation (samples D - F) should be viewed with caution since the insoluble particles in D probably caused premature fracture. It is I ~I I {felt that the 8.2% elongation of sample D is not O10 20 30 characteristic of the material. Elongation (%) TABLE II YIELD STRENGTHS OF CHITOSANS Average yield Standard Number of usable strength deviation Sample test strips* (MPa) (MPa) A 13 96.4 6.5 B 20 93.5 4.0 C 18 102 3.5 D 25 116 7.7 E 30 98.4 6.7 F 23 95.0 5.1 G 25 103 6.7 H 31 104 4.0 J 27 97.0 5.3 * Includes only strips that yielded near measured minima. 84 11~~~~~~~~0 - -- ~~6. ACKNOWLEDGEMENTS This work is a result of research sponsored by the MIT Sea Grant Program, supported by NOAA's 105"'~~~~~ - ~~Office of Sea Grant, U.S. Department of Commerce, " 105- under grant number 04-7-158-44079. The U.S. i0 t 'government is authorized to produce and distribute ~~~_~~ ~~~ ~ ~~ > --reprints for governmental purposes notwithstanding CD 100- 1 any copyright notation that may appear hereon. The authors wish to thank Mr. Peter Perceval, C(~ ~ ~~~~D and Mr. William P. Hunt of the Hunt Crabmeal --'~~~ It ~ Company, for their close cooperation in the n 9 5 processing of the chitosan. Thanks are also due 95 -Dr. Wayne Bough and Dr. Arnold Wu of the Marine :- Extension Service, University of Georgia,for the ._ molecular weight data and for advice on the 9C1)~~~- techniques of viscosity measurements. 90-- t_� I I I I 70 80 90 100 Deacetylation Temperature 1.4 (�C) Figure 2. The effect of deacetylation temperature on the yield strength of chitosan membranes. 1.2 - (D0 0 Samples G, H, and J indicate that increased deacetylation temperatures caused larger 3 -.0 - elongations. This trend is graphically depicted in Figure 4. The reason for increases in elongation with lower molecular weights is unclear. Microstructural analysis of these samples should aid in intrepreting these data. 0.8 - The NaOH concentration during deacetylation appeared to be critical for the solubility of the resulting chitosan. For the blue crab l l l l chitosan, 45% NaOH was the minimum solution 7 strength required. 70 80 90 100 Deacetylation Temperature 1. CONCLUSIONS (0C) The deacetylation variables studied had little effect on the mechanical properties of blue crab chitosan membranes. However these Figure 3. The effect of deacetylation temperature same variables caused substantial changes in on the molecular weight of blue crab chitosan. viscosity and molecular weight. It therefore appears that variations in viscosity and molecular weight do not greatly affect the mechanical properties of chitosan films. It should be noted that these conclusions only apply to the rather narrow range of deacetylation variables studied here. 85 TABLE III ELONGATION OF CHITOSAN STRIPS Average elongation Standard Number of usable to fracture deviation Sample test strips* (%) (%) A 12 15 6.4 B 18 25 7.5 C 18 14 7.2 D 28 8.2 4.2 E 23 19 8.4 F 21 24 7.4 G 28 9.1 5.8 H 30 19 9.4 J 27 25 9.6 * Excludes strips that broke at flaws. 7. REFERENCES 1. I. Joffe and H. R. Hepburn, Observations on Regenerated Chitin Films, Journal of 3 - Materials Science, 8 (1973), p. 1751. 0 P' 2. F. Yaku and K. Yamashita, A Method for *'--~~~~~~~~~~~ /the Production of Chitin Film, Japanese 1 20- Patent No. 48-19213, June 12, 1973. O 0P ~~~~~~~__ 3. R. A. A. Muzzarelli, A. Isolati, and EJ_-~~~~~ /A. Ferrero, Chitosan Membranes, Ion O 10 - -1 g Exchange and Membranes, 1 (1974), p. 193. 10_ - 4. B. L. Averbach, Film Forming Capability o _of Chitosan, Proceedings of the First International Conference on Chitin and A\, I ~~I I I Chitosan, (to be published). 70 80 90 100 5. F. A. Rutherford and P. R. Austin, Marine Deacetylation Temperature hitin Properties and Solvents, Proceedings of the First International Conference on (�C) Chitin and Chitosan, (to be published). Figure 4. The effect of deacetylation temperature on the elongation at break of chitosan membranes. 86 SEDIMENT AS A CAUSE OF MACROCYSTIS GAMETOPHYTE NON-SURVIVAL L. A. Volse, Associate Professor Department of Applied Science and Engineering U.S. Coast Guard Academy New London, Connecticut 06320 Abstract Many hypotheses have been formulated to ex- plain the alleged kelp bed diminishment. They in- Survival of Macrocystis spp. gametophytes clude sea urchin proliferation at sewage outfalls is a function of sediment depth on the surface (North, 1970-74), near-extinction of the sea otter to which they attach. Relative survival varies (Estes, 1974), increased water turbidity near as an exponential function of calculated mean sewage outfalls (Neushal, 1976), presence of toxic depth in all cases. If spores attach to a chemicals or pesticides (Barilotti, 1976), and surface prior to sediment settlement, no survival sediment settlement on rock substrate (Devinny, is noted after four days at depths above 500 1977). microns. If spore inoculation occurs after The purpose of this paper is to outline settlement, no survival is noted for calculated certain experiments conducted to test the latter mean depths above 40 microns. Microscopic hypothesis that sediment deposits (whatever their examination of silt dispersal patterns showed origins) might be a direct cause of Macrocystis spp that below 40 micron depths the percentage of non-survival (Volse, 1977). The studies described non-occluded surface was also an inverse ex- herein deal only with the effects of varying the ponential function of mean sediment depth. We sediment depths under laboratory conditions. infer that motile zoospores cannot pass through Timing of spore inoculations (i.e. whether spores even the finest silt particles. Comparatively were inoculated prior to or after sediment settle- better survival is noted with attachment prior ment) proved to be a most important parameter in to deposition. all of the experiments conducted. 1. Introduction 2. Experimental Methods Considerable concern over the disappearance Procedures for Macrocystis spp. laboratory and fluctuation in density of giant kelp culture have been developed by Neushal (1963), (Macrocystis spp.) beds along the southern North (1970) and others. By modifying existing California coastline has been shown recently. techniques, a standard procedure was developed by Fisherpersons (sic), sports divers, environmen- the author to provide for statistical analysis talists and others value this algae as a vital throughout the course of his experiments. Zoo- part of the complex marine ecosystem. spores from Macrocystis spp. sporophylls, through Commercially, the giant kelp plant is a chilled seawater temperature shock, were released principal source of algin, a highly efficient to form a spore solution. These spores were thickening and gelling agent. Kelp is also a incubated under controlled conditions of tempera- good source of potash and other useful chemicals ture and lighting for a standard period of four (Limbaugh, 1955). More recently, concerns over days. Various combinations of sediment with and limitations of food and energy resources have without spore solution went into an array of petri pointed towards kelp as a promising panacaea, dishes, each containing a microscope slide in owing to it's rapid growth characteristic. The conjunction with an appropriate amount of nutrient- former Energy Research and Development Adminis- rich seawater. Temperature was kept at 80�C tration's Ocean Farm Project is an example of throughout all experiments. Four days after in- one of these later schemes. oculation of zoospore solution, slides were re- Giant kelp beds apparently have diminshed moved. sediment washed clean, and a slide cover put both in size and density as the human population in place. Under a light microscope of appropriate of the coast of west America has grown. Although power to insure statistical significance, counts evidence is confusing and difficult to evaluate, of surviving gametophytes were made. A typical North (1970-74) and others contend that Macro- appearance of these germlings after four days cystis tends to flourish along coastlines with incubation is shown in the microphotograph of less dense populations, compared with it's figure 1. growth off heavily populated regions such as the Los Angeles basin. 87 U.S. Government work not protected by U.S. copyright. statistical testing. Empirical equations fit data selected by the method of least squares. Curves of the types y = a + mx, y = e mx, and y = mxa were investigated and the type selected which showed the best correlation coefficient, even when significant fits with lesser correlation might be found. 3. Gametophyte Survival as a Function of Sediment Depth The totality of experiments established a relationship between GSI and mean calculated sediment depth. Best fit for all data proved to be given by an equation of exponential type, plotting on semi-logarithmic paper as a negatively sloped straight line. Using statistical tables based on Figure 1. Macrocystis pyrif era germlings four quantity of data and correlation coefficients ob- days after spore inoculation. Note original 5 tamed, indicates 99+ percent probability of such micron dia. spore with 30 micron long germ tube data relationship being valid. offshoot. Consolidated data from Tables i and 2 gave the best equation for the case where spore solution was introduced simultaneously with or prior to sediment To measure a particular inoculation of paeeta spores' ability to attach to substrate and -lcmeta survive conditions imposed experimentally, a SI 3.eE. measure, called gametophyte survival index (GSI) was devised. One petri dish of each experimental arra wasdesinatd asa "cntrl" dsh, ithwhere GSI is in percent and sediment depth, d, is array was designated as a "control" dish, with no sediment to be introduced therein. Four days in microns. For 37 data pairs, correlation co- following inoculations, germling counts were made efficient is -0.777. on both control and experimental slides. GSI is In cases where spores were allowed to attach to slide prior to sediment introduction, the defined as the ratio of a field count for an ex- perimental sedimentary slide to a field count for expirical efficien is of th a t of the exponential coefficient is 1/200th that of the control slide without sediment. Setting Eq. 1 With data from 23 observations (Table 3), control slide GSI at 100, GSI on a scale from 0 best equation is through 100 becomes a relative measure of -0.0044d gametophyte survival under each of the imposed invironmental conditions. Amounts of sediment calculated to settle Amouns ofsedient alcuatedto sttlewith a correlation coefficient of 0.792. Plot of compactly at its specific gravity were weighed with correspon efficical into plastic bottles along with volumetric measurement of sea water media (enough for dupli- equations are shown in Figure 2. cate samples). In cases where sediment was added concurrently with spore inoculation (pro- Tin Slime Layers viding opportunity for sediment to settle prior to spore attachment), each dish received 30-ml of An examination of several slides before sedi- well-shaken media-spore-sediment suspension. In cases where sediment was added 24 h later, only methdbnwaedofpirogrmngcus 20 wlhofmedia-sre sotiment first addedate, only revealed that, owing to gradation of particle sizes 20mlofmedia-shoe sollowing f t went adintiona 1(i.e. according to Figure 3, less than 10 percent dish.Theollofing mdia anadsedimtifonallod are smaller than 10 micron diameter) slides with suspension of media and sediment followed. Sourceniof Macrocysti spp. zoospores wascalculated mean depths of up to 20 microns were not Source of Macrocystis spp. zoospores was from sporophylls obtained from the base of plant fully occluded by sediment. Table 4 indicates the growing in the vicinity. When immersed in observed relationship between calculated mean sedi- chilled seawater media at 800C, temperature shock ment depth and the observed percentage of field which appeared as sediment-free. A plot of these released spores giving a golden-brown color to data is given in Figure 3, for which the empirical the solution. A one-ml amount of spore solution euto went into the sediment-media bottles, and after -0.127d careful shaking, inoculation into petri dishes 10.23 was then carried out volumetrically. To guard against mistakes and to provide a e uard a t mstakes o xperoid a was obtained. PSF is the observed percentage of better statistical basis, each of the experi- field that was sediment-free and dis the calculated ments was run in duplicate. Germlings were mentscwasrunt in ducthe fieldofeterasOx were 1mean sediment depth (up to about 40 microns). With couned i th fied o eiter a500 or 25xdata from 24 observations during one experiment, a light microscope, depending upon the quantities to be visually counted. Nine random counts were regression coefficient of r= -0.966. This is far made on each experimental slide. Data was greater than r(crit. 1%) = -0.462, showing an compiled and is shown in Tables 1, 2, and 3. extremely high statistical significance. At mean Regression analysis is the basis of all calculated sediment depth of 40 microns or greater, 88 trt Tht I: sediment coverage proved to be complete in all 1QQT cases viewed. The prime factor involved here seems to be particle occlusion of a surface. Long -n Th# 57 t 3 = H(1976) remarked that germlings have been observed $60 6t-r-## #t- -EF W-5t~ = __=E XE~ii5Eto attach themselves to small particles and con- 5_, _ _---=0=.=F tinue their growth under laboratory conditions. 4 0i -~_g -=~ ~[2_-_ Such attachment, although possible, would seem to 30,V:V 7i1-_ Hs E3L -J _ _ F provide anuncertain fate for the gametophyte, + ~F~ -~~_ar, -:_ : :.'~ =:% g r::<'!_:' gbeing subject to whatever water currents nature provided. It seems apparent that such attachment m t--'ht-tt Wh=- ~~~-XSL::tttt ==W-~- -=it==-- to fine particles would prove futile insofar as 44tH 1T z 7 =P~t == +4! -T4=Fultimate sporophyte survival is concerned. l = - i njY 1t4 - -.+ 5 - r' r, hr -- 5--P----P--P-T P f - 2 1 = t X ~ ~ ~ ~I- ti�;; t 6- 4e 4ri 04 t tL;-~t_ Ltir V`-_ - - - - __ - tN r i7 i j o tINV 2f Ft 0W Ff WMt d�L Er - T4 v tr~%r~rr~4 Th tfl~ K-40 l U0 200 300 400 5 00 600 700 ffi$= F T:I_~f~i~tt= $- w t - :Rt. ra:, ti t fi- - t Tr- f SEDIMENT DEPTH, d (microns) _ * = f -'�N4 V TflfT ThtVMt-- F t f _ 4 P r rf-#. - -ITu P- Tc KPYI Figure 2. Effects of sediment on survival of to n.$ L-t_ Ptt Macrocystis pyrifera gametophytes, including � 3 i-WI FE#0,A .,T& WV L- ;:4 A effect of timing of spore inoculations. Data c 3 - taken from Tables 1-3.- t 1t � 1 CALCULATED SEDIMENT DEPTH, d A ) '4Zrti ._. L 1 .::--4: ._J7;: l a i-, - .Figure 4. Relationship between calculated mean : -; I F-i Ili : - , :_ l1_i t ' :, l . ...... ;: .. ....... L-:.: ;; _ sediment depth and percentage of field observed as sediment-free. Note the similarity of re- lationships between GSI and PSF vs. d. &I40: : ;I ' ; , We note from Equation 3 that the percentage ....X :.. . -- *-- l L... u ~~ -i - -1iof field sediment-free (PSF) is an inverse exponential function of sediment depth, d, as is P 30 - h>_ _ ___i__ ___ - the gametophyte survival index (GSI) of Equation 1. -.-.----- .- ...--1-. --1- -1----- I While the equations are not identical, comparisons ..... t t......'*'- '-i 'I" ' over much of the range examined (Figure 3) are _ii20 - consistant with a simple explanation: the spore -- --1-- -1 I produces a germling attached to the slide only if 10ii-F15liIi: i--ht% .... iit falls on exposed glass. If the spore falls on Epj10 �--W|~---F- F - � -8 9 4 ----i--t--sediment particles, it either fails to develop or -t H|- --!--t -- -|- I~lfT-_ l- g -T-sZ is lost when the slide is rinsed. Because a spore K- ;-2-1-l- ~: ~ l~t 1 attached to sediment is unlikely to survive under natural conditions, experimental survival rates arwot-80 s'. r e- On cLn' n (- - probably are similar to those under appropriate *" GRAIN DIA. (MICRONS circumstances in the ocean. 5. Summary and Conclusions Figure 3. Sediment grain size based on hydro- meter analysis of sediments. All particles are Macrocystis pyrifera gametophyte survival is smaller than 74 micron diameter. inversely related to sediment depth. In cases of sediment introduction either before or along with zoospore inoculation, the best empirical equation 89 corresponding to 37 data pairs (Fig. 2) is TABLE I -- EFFECT oPVARNO DEPI OP aOwslorrons sAmS oN wnoern sunn,, (SBI~ IN PLACE PR.IOR 1 SFRPE INOCULATI(N) Date . e- sediment No. so Nean geoling d. de. Gaetophyte suvival -0.o d odepth, d cotts/ coont per of ma. index, 3GS -0. 088d . . . o un't, (mc o s o s ,felon onono lsStnonsl nitero. fiend toast (tercent) GSI = 34.3 e Eq. 1 ,16 0 68 28.5 19 9 1.50 0.9a 20. 9 9 O1.11 0.93 2.0 02 o n-sn ~~~~~~~~~~~0.00 in 0.11 O.U ~~~2.0 where GSI is gametophyte survival index in per- 1n//76 6 9s i 250 too 9� - 2 5 :0 'W cent and d is mean sediment depth in microns. In9 . 27: 5~ ~~~~~9. 7.5 26.3 10 9 1 . cases where spore inoculation occurred 24 hours a ~~~~~~~~~~~2.2 98.6 a 20 n 0.4 20 h prior to sediment placement, the best empirical 4 1 n 0. 4O//7 - 1.78 1.30 equation corresponding to 23 data pairs (Fig. 2) .2/8/7, 12 o 9 ~ is 0 0. 10.0 t0.5 10.6 16.8 lO 9 505 06 20 9 3.22 2.6n a, 6 1.2 -0.0044d Notes. () Tulation s ....osists of compilation of data from to sepaate experiments, each having a four day culture period. GSI = 54.4 e Eq. 2 (2) Sdiment particle. ae less tha 74 mire d.a. Excellent statistical correlation exists for both equations showing the relationship between vari- ~ables to be highly significant. STABLE 2 -- EDPC O VAR4�N DEM OF tOtSTOJ SA0 s 0,tT O rt CtOtM E SURVAL ables to be highly significant. (SEDIMENT AND SPORF IXED AT TIM OF IN.CO0ATION} Examination of slides made prior to counting te M.ea...edisent No. of emn gerling Sd. d. Gaetophyte .rvia of depth, d coants/ ot per of moan inda, BIT procedures revealed that field coverage by parti- rant .....o ot .... fiend on ..ne..I cmcount courcnt) 12/15/76 0 4 io.S cles is variable, increasing with sediment depth. 5 9 7.80 3.90 15.9 According to a plot of twenty data pairs (Fig. 4), 05 2.12 :o 20 l O 9 6.33 2012 12.9 the best empirical equation is 20 . 1. . i/S/Of 01.00 4-St itO~~1.7 -0.127d n2o 9 .o 5.O 4.7 20 h.7 n.1 o 15.27 . PSF=103 e Eq. 3 0 OSS 1/3O/77 0 0 .01 0 0 1.00 9.o33 2 9O0 9. 0.05 12.1 where PSF is percentage of surface sediment- 2 9 4 .3 9 free and d is sediment depth in microns. We can 1/29/77 3 25 4 0 9, 5 0 2.00 1 3.2 conclude that although Equation 3 is not identi- 1.045 4. cal with Equation 1, comparison over most of the 01 3 6 range examined is consistent with the explanation toes (1) Tabulationconsists of compilation of data from four experiments, each with & four day culture period. that a spore produces a germling attached to a (2) Sedisoo paice are ess than 7 i di. slide only if it falls on exposed glass. By comparison of Equation 1 and Equation 3, TABLE 3 -- EFFECT OF VARYIN DOPs OF CNSTRUCTION SAD SEnIoENT ON GAMETOPTEP SUR-VIVAL and by rounding off the exponential exponent to (sEDnos ssTnCc ODnC tot i R SMEt eICLTOIOi) 0.1 in both cases, we can conclude that an te M .dea ming .s. de.. OMnes.pspte arvhvnh nf deptl, d c..... i .... pr S .... ofen ideh s.. arrpoximate linear relationship exists between uo .... fieed oot orce 11/12/76 o s lB 5,4 .5 GSI and PSF, up to a calculated sediment depth 0 29 1:.9 0.. 19 ~ ~ ~ ~ ~ ~ ~ ~ ~~~~4. of some 40 microns. It appears that the diameter 2 h . of particles at these low depths does not control .67 .11 2.1 12//7 0.69 0 .67 1.1 the relationship, but rather the degree of distri- 2//7 0 30 9.00 19 9 2.50 1.00 20.4 hO9 7.70 2.309 6. bution of these particles in occluding the slide. n r 3:0 7 4 .0oo 3.0 01 6.0 19 9 6.22 2.77 69.I When sediment is in place prior to zoospore 5 3 .oa0 6.9 92 9 5.a9 2.32 65. arrival, presence of any sized particle is 9 .6 2. 19.6 456 9 1 .44 2..13 i.o sufficient to impede attachment of zoospore to i4 9. 0.37 1.2 iS~ ~ ~~~~o. 04 0.50 4.0 50//5/70 0 7 3.0 50.7 5CO substrate. Even though spores may manage to s 9 h.0 6h 32.7 0 23.7 0.3 4. attach themselves to loose sediment particles, 1.9 6 '. the ultimate instability of the particle seems to o 11.3 436 23. 20 9 2~~~~1.1 S.1 43.1 preclude gametophyte survival. Noes, (1) Tabulaion cosists of compilation of data from three experiment, each wih a four day culture period. Whenever spores attach themselves to a sur- (2) Sedimen prt esa lstan 74 mon da. face prior to sediment introduction, gametophyte survival chances are greatly increased. We infer TABLE 4 ICSHP LCU D AN S D AN TAtLN 4 -- Rt~hsE ttptStlttL.ATIO NSP TED n M ttIN~ SEeTtDPsTH A00 that the effect of sediment in this case lies not PRCENAGo OF FTELD CLEAR OF SnIMoN fa�oahoumd moan Peruentaie of field in interfering with the germling's continued sedimen Ier e ed sedomens b.- nheredmdim...t- denth.l(dl f ree./SFI attachment, but rather in preventing nutrients or 1 o it oxygen from being absorbed by the attached gameto- g phyte. o ' 5 00 60 5 t O O0 5~~~~~~3 40 10 4O 10 40 'O 35 1a 10 20 10 20 'O 10 'O 10 20 I0 2~~~~~~~ o Notes: (1) Obsertvaions made dunin experiment of 12/8/76. (2) Consrction and sand sediment with p-ticle diaeter 90~es tha 74ion 90 References 1. Barilotti, D. Craig, "An overview of kelp forest biology," paper presented at environ- mental symposium for teachers. University of California, La Jolla, California, 2 April, 1976. 2. Devinny, Joseph S., personal communication, University of Southern California, Environ- mental Engineering Department, Los Angeles, California, 1977. 3. Estes, James A. and Plmifano, John, "Sea otters: their role in sea urchin nearshore communities," Science, 1974: 1058-1060. 4. Limbaugh, Conrad, Fish Life in the Kelp Beds and the Effects of Kelp Harvesting, La Jolla: University of California Institute of Marine Resources, IMR Ref. 55-59, 1955. 5. Long, Ann, personal communication, Cali- fornia Institute of Technology, Marine Laboratory, Corona del Mar. 1976. 6. Neushul, M. and Eaxo, Frana T. "Studies on the giant kelp Macrocystis i., growth of young plants," American Journal of Botany 50, 4 (1963):349-353. 7. Neushul, Michael, personal communication, University of California, Santa Barbara, California, December, 1976. 8. North, Wheeler J., Kelp Habitat Improvement Project. Annual Report, 1969-1974, 5 vols., Pasadena, California. Institute of Tech- nology, 1970-1974. 9. Volse, Louis A., Effects of Sediments on Macrocystis pyrifera Gametophytes, doctoral dissertation, University of Southern Cali- fornia, Los Angeles, California, June 1977. 91 NAVY RESEARCH ON MARINE BORERS AND THE LABORATORY CULTURING OF LIMNORIANS Kathleen M. Parrish and John D. Bultman Naval Research Laboratory Washington, D.C. 20375 ABSTRACT The Naval Research Laboratory's investi- gation of the protection of wood in marine ser- vice which began with an in depth study of --- creosote, is now concentrating on a study of the active antiborer constituents of naturally bioresistant woods. This work is now focussing on woods which are resistant to limnorians, marine crustacean isopods, because most current protective measures are inadequate against these pests. In order to study the effect of the natural wood extractives on the physiology of these marine borers under controlled con- ditions, a laboratory culture of Limnoria tri- . punctata has been established using a closed recirculating aquarium system. The usual in- vertebrate closed aquarium culture techniques have been modified to fit our needs. This in- cludes the redesign of the biological filter bed, and its location outside each aquarium. A full description of our closed recirculating aquarium system and the procedures for its maintenance are given. Fig. 1 Wooden structures in San Francisco Bay INTRODUCTION destroyed by marine boring organisms: A - The ravages of wood-destroying organisms railroad trestle, B - ferry slip, and C - have been known to man since the time of municiple warf and pier (1). Photo Aristotle. The Greeks and the Romans mention courtesy of the American Wood-Preservers wood-borers in their writings, and according to a Assn. German biologist, the Bible refers to borers inshows the typical hour-glass effect cau- Figure 2 shows the typical hour-glass effect cau- the passage, "As the worm the wood, so an ugly sed by limnorians as they damaged piling support- wife destroyeth her husband". The Phoenecians and the Vikings and others charred the hulls of ing a pier, and Fig. 3 shows a fender timber their ships and smeared them with a mixture of thoroughly infested by teredinids. The outer sur- their ships and smeared them with a mixture of arsenic, sulfur, pitch, and animal fat to prevent face of the wood gives little indication of the borer attacks. The galleys of Troy were similar- extent of the internal damage. Although it is ly sheathed with lead, and in the 16th century difficult to assess the annual expenditure by the preventive measures were needed when the wooden Navy for repair or replacement of biodamaged wood, the most recent estimate (2) places the cost at dikes of Holland were being destroyed. Today, wooden-hulled ships are no longer important militarily or in world commerce so for NAVY INTEREST IN WOOD PROTECTION them borer damage is no longer a pressing econo- mic problem. However, man is still plagued by Navy use of Wood the depredation of fixed harbor installations by marine boring organisms and also by fungal in- fection. Figure 1 shows the extreme damage available as a construction material. Also, wood wrought by these animals in the San Francisco is still the cheapest, and frequently the most suitable construction material for many purposes because of its unique physical properties such as strength-to-weight ratio and resiliency. Because 92 U.S. Government work not protected by U.S. copyright ration, the Navy has a vested interest in wood preservatives and preservation techniques. Shortcomings of Present Preservatives The Navy is interested in wood treatments because current preservatives for wood in marine service are not entirely effective. The most widely used preservative treatment is pressure impregnation with whole creosote or creosote/coal tar, but in tropical and subtropical waters creo- soted timbers are often destroyed in a few years by the gribble Limnoria tripunctata. Also, creo- kl1 Btf~ I sote can be metabolized by the bacterium Pseu- domonas creosotensis (4). This microbe degrades t t he creosote, thus making the wood vulnerable to marine borers. Other treatments with mixtures of copper, chromium, and arsenic (CCA, ACA), with or without subsequent pressure impregnation with creosote, are very effective against borers but ~~~~Fig. 2 Limo ig howng hethe treating process embrittles the wood (5) Fig. 2 Limnorian-damaged piling showing the characteristic damage pattern produced by these organisms in the inter-tidal zone (1). Photo courtesy of the Ameri- can Wood-Preservers Assn. The potential for environmental pollution has become a vital consideration in the use of wood preservatives. Creosote and other wood protect- ants are becoming increasingly suspect, and they are scheduled for review by the Environmental Protection Agency (EPA) as materials possibly hazardous to the environment. Also, creosote is classified as an oil by EPA and its use is subject to the same stringent environmental regulations as other oils. In addition, current attitudes toward copper, chromium, or arsenic in natural situations may prevent their use as wood protect- ants in the future, even if the treatment techni- que is improved to prevent embrittlement. Because of its large investment in wood, the need for more effective treatments, and the po- tential environmental problems associated with the current wood preservatives, it is important to the Navy to seek new wood protectants which will eliminate these problems and also reduce the annual maintenance costs associated with wooden marine structures. MARINE BORER RESEARCH The Naval Research Laboratory (NRL) has been engaged for many years in research on materials protection in the natural environment. Early work by Sweeney and Price (6,7) on wood-preser- Fig. 3 Fendrtmbethrouhlyinfsvation involved an exhaustive study of creosote Fig Fnde tmers. Theouerosur iest as a marine wood preservative. Currently, re- little indicatio n of the oextent of gsearch on wood protection is based on a study of linternatl indamcag te. naturally bioresistant woods. This includes the isolation of their active constituents, a deter- of~ thseatrbue teayse areunt-mination of the effects of these compounds on the of these attributes the Navy uses large quanti- ties of wood. It has in service about 1.5 million physiology of the target organisms, and a deter- mination of the feasibility of using these com- utility poles, over 6 million crossties, innumer- pounds directly as wood protectants or models able dolphins and camels, 1000 piers and wharves, and over 180 miles of fender system constructed fo ch oo n tean s se for such woods in the Panamanian forest and else- of this material (3). Since all of this wood must be treated to protect it from biodeterio- where culminated in the evaluation of a large 93 an economically favorable anti-limnorian could be number of woods for their inherent resistance to isolated from one of these woods, it might be pos- wood-destroying organisms (8,9). The most resist- sible to incorporate it into creosote, or use it in ant wood in all environments to all wood-destroy- conjunction with the dibutylbenzylphenol to provide ers evaluated (marine borers, termites, fungi) complete borer protection. To study the effects was Dalbergia retusa (cocobolo). A phytochemical of extractives from these limnorian-resistant examination of the heartwood was performed to woods on the physiology of limnorians we have isolate and characterize the antiborer constitu- established a laboratory colony of Limnoria tri- ent, subsequently identified as obtusaquinone (10). punctata to provide the necessary experimental This compound was active against teredinids but animals. not pholads or limnorians. Research (11) showed that obtusaquinone interfered with the normal BIOLOGY OF LIMNORIANS formation of shell by metamorphosing teredinid larvae. This is accomplished by disrupting the Limnorians, or gribbles as they are commonly formation of the tanned protein matrix, the con- called, are small isopods belonging to the class chiolin network, upon which mineralization takes Crustacea. They average about 0.5 mm in diameter place. x 3 mm, the male being slightly smaller than the female. The body (Fig. 5) can be divided into The ineffectiveness of obtusaquinone toward three main segments which are the head, thorax, pholads and limnorians and the prohibitive cost of and abdomen. The head possesses six sets of appen- its commercial production led to a manipulation dages. These include the mandibles and first pair of the molecular structure of obtusaquinone to see of maxillae which are equipped with sharpened, if a better compound could be synthesized. The chitinized tips and strong muscles for chewing. extensive alteration of the obtusaquinone molecule The thorax contains seven pairs of walking legs produced a series of benzylphenols, one of which ending in claws which are used to grasp the wood. was found to be very active against the molluscan During the breeding season a brood pouch develops borers (teredinids, pholads). Susceptible wood between the second and fifth swimming legs. Ferti- treated with this compound was free of damage by lized eggs are deposited here and develop into teredinids and only very lightly attacked by embryos and larvae; young adults immediately begin pholads (Fig. 4). Preliminary studies indicate boring at right angles to the parental burrow. that this dibutylbenzylphenol will also be an The abdomen bears six pairs of legs called pleo- anti-termitic. Its effectiveness against mollus- pods. These pleopods are equipped with gills can borers and its easy, economical synthesis on which are constantly in motion, causing a circu- a commercial basis, make this compound a good lation or water within the burrow. This water is candidate as a wood protectant for marine ser- enriched with oxygen which is adsorbed and used vice at least for these organisms, and steps will by the animals for metabolism. be taken to obtain its certification for this use. ocdermolo brood pouch flh gs 50 p"roopod pleopods Fig. 5 Morphological representation of a limno- Fig. 4 X-ray photograph of pine sapwood disks. ran, showing the whole animal, lateral Those of the rleft contain 0.9% dibutyl- view. (From Marine Boring and Fouling untreated controls. The damage was caused sion of the University of Washington sion of the University of Washington predominantly by teredinids. Press. 1959). However, the problem of inadequate protection Limnorians are continually active and once a against wood damage by limnorians remains, and new burrow is begun, the animals are usually en- will remain even if creosote continues to be used closed within the wood in 4-6 days. The burrows, as a wood preservative. Consequently future work which are just beneath and parallel to the wood at NRL will be more strongly oriented toward an surface, are marked at regular intervals by small investigation of those woods which were particu- breathing holes not large enough for the animals larly resistant to attack by limnorians (8). If to pass through. These holes serve as passages 94 for oxygen-containing water which is circulated culturing depended upon using planktonic larvae through the burrows by the beating action of the instead of obtaining them from fertilized eggs pleopods. under controlled laboratory conditions. Most re- cently, Loosanoff and Davis (21), and Walne (22) No wood debris accumulates at the anterior employed techniques which enables them to culture end of a burrow, and all wood fragments scraped successfully the larvae of bivalves. However, away are ingested and pass through the alimentary their laboratories were located at the seashore canal (12). Limnorians were once thought to bur- and their culture systems utilized running sea- row into wood solely for protection, however sub- water. Turner, a leading authority on the labora- sequent studies have indicated that they are tory culture of limnorians and teredinids, also capable of utilizing the wood scrapings as food. uses natural seawater in her system (23). The ad- The wood is hydrolyzed by a cellulase which is vantage of open culture systems is that the animals produced by the borers (13) rather than by cellu- are supplied with a continuous flow of fresh sea- lase-producing microbial symbionts in the gut. In water and the used water containing the waste pro- fact, Ray (14) showed that no microorganisms are ducts of their metabolism, mostly toxic nitrogen- present in the alimentary tract of these animals, eous organic compounds, returns to the sea. Scanning electron microscope studies by Sleeter, et al (15) have confirmed these observations. There are also disadvantages associated with They also showed that there are no microorganisms open systems. These include problems with the at the anterior ends of the burrows where the mechanics of the system, corrosion of pipes with boring activity of the animals is apparently concomittant release of undesirable metallic com- rapid enough to prevent the colonization of the ponents, fluctuations in salinity and temperature newly-formed wood surface. Limnorians require of the water, and clogging of filters by storm- nitrogen for amino acid and protein synthesis, generated silt. and wood is low in this element. Thus if limno- rians utilize wood as a food, their nitrogen Closed Recirculating Aquaria requirement must be satisfied from another source. Marine fungi have been considered because these Laboratories not located near the seashore nitrogen-rich microorganisms are so ubiquitous; must use closed aquarium systems which are more another suggested source is the pool of dissolved complex to maintain. These systems, however, nitrogenous compounds (ammonia, nitrites, nitrates) offer advantages such as ease and convenience of in the water (16). monitoring and sampling, continuous visual obser- vation, freedom from severe variations in salinity Lignicoulous marine fungi (or bacteria) may and temperature, preferential selection of biota have some relationship to limnorians other than cultured without danger of contamination by plank- nutritional. It has been hypothesized that be- tonic forms, and independence from such natural fore marine borers can penetrate a wood surface effects as algal blooms and increasing levels of it must be softened by fungal activity, and it has pollution encountered in the near-shore environ- been reported that limnorians are unable to attack ment. wood in the absence of marine fungi (17), however Many laboratories and commercial aquaria use it was demonstrated subsequently that limnorians recirculating tanks for culture and display pur- are capable of growing in fungus-free wood (18,19). poses. The design parameters for monitoring these The consensus is that marine fungi are not neces- tanks have been presented by several investigators sary for limnorian attack on wood, although a (24-26). Some of the standard techniques were fungal-softened wood surface would certainly in- modified and unique features were incorporated to crease the rate of attack. accommodate the special problems encountered in culturing our limnorians. LABORATORY CULTURING OF LIMNORIANS Aquarium System at NRL Laboratory culturing of experimental animals Limnoria tripunctata are maintained in five is desirable if controlled research is to be con- 30-gallon, all-glass aquaria; tight-fitting lucite ducted because culturing allows the researcher covers minimize evaporation. Artificial seawater ready access to a population of animals with a is made up according to Kester, et al (27) in 45- known history. The organisms can be watched liter batches. Each batch of water is adjusted to closely, under conditions most like those encoun- a final salinity of 30.00/00 with distilled water, tered in nature, and their patterns of growth and aerated until the pH is steady at approximately reproduction can be observed to determine if lab- 8.2, and stored at 30C without filtering. Figure oratory inbreeding is causing adverse genetic 6 presents a schematic of the recirculating cul- effects, ture system and Fig. 7. shows three tanks contain- ing limnorian-infested pine panels. A small mag- Open Seawater Aquarium Systems netic-drive centrifugal pump circulates water from the tank via 3/4 inch polyethylene tubing through Culturing of marine crustaceans and bivalves an in-line filter then to the base of the biologi- is not new. In 1927 Wells (20) reared the Ameri- cal filter bed (BFB). The water passes upward can oyster Crasseostrea virginica from artifi- through the filter and back to the tank. The flow cially fertilized eggs to the settlement and shell rate ranges from 12 to 60 liters/hr depending on -building stage. However as late as the 1950's the buildup of detritus within the BFB. Three air 95 side. The BFB consists of a plexiglass cylinder about 6 feet tall and 6 in. in diameter, filled with 40 pounds of crushed oyster shells. The shells are supported off the bottom of the cylin- der by a perforated plexiglass plate. The usual i14 - t l I BFB for closed water aquaria is normally contained .s~tr8�i 4 = inside the tank and is used in conjunction with a subgravel filter/airlift system. This system could not be used in our laboratory because the s=t ~ i P & boring activity of the limnorians produces frass (a combination of powered wood and fecal material) at a rate of about 1 inch/month. If the filter bed were contained inside the 2 - aquarium, the frass would clog the interstices be- tIl q~~5~ BSI tween the oyster shells, preventing efficient water flow. The resulting stagnation would upset the efficient biological filtration of the water and could eventually lead to the death of the Fig. 6 Schematic view of our closed-system animals. Wood panels used for culturing the borers aquaria for maintaining limnorian cul- are suspended in the tank (see Fig. 7) from glass tures. Water from the tank is moved by hooks inserted into one end and hung from a plexi- an all-plastic centrifugal pump (1) glass framework which rests on the bottom of the through an in-line filter containing aquarium. These panels are riddled with burrows charcoal and plastic wool (2) which re- in about 6 months' time as is shown in Fig. 8. A moved larger particulate matter, to the control pine panel is also shown at lower right. bottom of the vertical biological filter bed (3) of crushed oyster shells, and then back into the tank. The water is oxygenated and the intra-tank circula- tion improved by air from an air pump (4) which is routed through an air-lift tube (5). Fig. 8 Attacked from both sides, these pine panels were reduced from sound wood (A) to spongy masses after 6 months' exposure in our laboratory colony of Limnoria tripunctata. Water Conditioning A newly established aquarium goes through a conditioning period of about 1 month before the culture animals are introduced. During this time the new filter system is exposed to the excretory Fig. 7 Three laboratory culture tanks containing products, e.g. urea, of hardy marine organisms limnorian-infested pine panels suspended such as hermit crabs, brittle stars, urchins, or by glass hooks from horizontal support small fish. The conditioning process is accelera- rods. The vertical columns are the bio- ted by the addition of some bacterial-contaminated logical filter beds containing the crushed oyster shells from an operating aquarium. The oys- oyster shell. ter shell inoculum provides the beginning of a rapidly increasing population of heterotrophic pump-operated airlift tubes located at the cor- bacteria which became suspended in the water ners of each tank provide adequate intratank column and attached to the shell surfaces in the water circulation and keep the water saturated BFB. The organic nitrogenous compounds' (urea) with oxygen. The in-line filter which prevents produced by the 'conditioning' animals is rapidly large particles from reaching the BFB, consists converted by these bacteria into ammonia which of a plexiglass cylinder containing activated reaches a high concentration in the water shortly charcoal with plastic filter fiber on the upstream after the animals are introduced into the system. 96 This mineralization reaction is the first stage FUTURE RESEARCH in the biological removal of toxic metabolites (Fig. 9). To reiterate, dibutylbenzylphenol holds pro- ORGANIC MATTER mise as a preservative against the molluscan borers which makes it a candidate as a creosote replacement should the continued use of creosote be banned. However, this benzylphenol is not ASSIMLATION OF NITRATE SWAOFCATION effective against limnorians. To provide this SY PLANTS AND BACTERIA \missing protection, future work at NRL will be NITROGEN more strongly oriented toward an investigation of FIXATION those woods from the earlier study (8) which were IOENITRIFICATION t 4 particularly resistant toward limnorians; the NO -+140 NO H interest in teredinids will be maintained. Coco- J3 2 ' 1 bolo (Dalbergia retusa), guayacan (Tabebuia guayacan), and cedro espino (Bombacopsis quinata) are already being examined for their anti-limno- \ 1" / 2rian heartwood component; lignum vitae (Guajacum officinale), zaragosa (Conocarpus erectus), and NITROWCTER WITI080110N^3 quira (Platymiscium pinnatum) will be added to the list. The extractives from these woods will be evaluated as anti-limnorians, and a study of the effects of the active compounds upon limnorian �2p #O2'~"~physiology will be made to determine which bio- chemical systems are being disrupted; relation- ships between chemical structure and biological activity will be sought. Animals from our labora- nitrogenous metabolites which accumulate tory culture will be used for this purpose. in the water. (From Spotte, S.H. 1970. Fish and Invertebrate Culture. John Wiley and Sons, Inc.). REFERENCES The high concentration of ammonia stimulates 1. Hill, C.L. and C.H. Kofoid. (Eds.). 1927. a rapid increase in the number of the supplied Marine borers and their relation to marine con- nitrifying autotrophic bacteria Nitrosomonas sp. struction on the Pacific coast. Final report and Nitrobacter sp. which perform the oxidation of the San Francisco Bay marine piling commi- of ammonia to nitrite and nitrite to nitrate, ttee. National Research Council/American Wood- respectively. The oxidation of ammonia to ni- Preservers Assn. San Francisco, CA. trate is the second step in the biological fil- trate is the w seond step in the biological fil- 2. Fischer, E.C., L.S. Birnbaum, J. DePalma, J.S. tration of the water. The final stage in this process is the reduction of nitrate to nitrous Muraoka, H. Dear, and F.G. Wood. 1975. Survey process is the reduction of nitrate to nitrous oxide or elemental nitrogen. This denitrifi- r eport: Navy biological fouli ng and biodeteri ration. Report NUC TP 456. Naval Undersea cation step can be carried out by autotrophic or heterotrophic bacteria and can occur aerobically or anaerobically. 3. Page, R.Z. 1972. Federal Committee on Wood Protection, Minutes of the 8th. Meeting, June Within 2 weeks to a month the system is con- 6-7. Naval Facilities Engineering Command, ditioned and the concentrations of ammonia, Washington, D.C. nitrite, and nitrate have stabilized; there is no., R.W. Drisko, and H. Hoch longer a lag time between successive steps in the nitrogen conversion process because all the nitro- 1961. Pseudomonas creosotensis sp., a creo- gen intermediates are being processed simultane- sote-tolerant marine bacterium. Appl. Micro- ously, and the bacterial colony is in equilibrium with the system. At this time the limnorians 5. Eaton, M.L., J.A. Erelicharz, and T. Roe. can be introduced. Mechanical properties of pressure-treated marine piles - Results of limited full-scale The buildup of toxic products (ammonia, testing. Naval Civil Engineering Laboratory nitrite, nitrate), oxygen concentration, pH, and Technical Report (in press). salinity Aremocnitored weekly. The concentration of toxic substances is controlled further by the monthly replacement of 25% of the water in the 1958. Coal tar cre osote studies. Part 2 - tanks with fresh salt water. The in-line filters are cleaned twice weekly, and the tank bottoms resistance to marine borer attack. Corrosion are vacummed monthly. Also, the BFB is dismant- led and cleaned with salt water bimonthly. 7. Price, T.R. and T.R. Sweeney. 1956. Marine borer control. Part VII - Evaluation of petro- leum-ether-soluble fractions and distillation fractions of creosote. Formal Report 4851. Naval Research Laboratory, Washington, D.C. 97 8. Southwell, C.R. and J.D. Bultman. 1971. 21. Loosanoff, V.L. and H.S. Davis. 1963. Marine borer resistance of untreated woods Rearing of bivalve molluscs. In, F.S. over long periods of immersion in tropical Russel (Ed.). Advances in Marine Biology, waters. Biotropica 3(1) 81-107. Vol. I, pp 1-136. Academic Press, New York, N.Y. 9. Bultman, J.D. and C.R. Southwell. 1976. Natural resistance of tropical woods to 22. Walne, P.R. 1964. The culture of marine terrestrial wood-destroying organisms. bivalve larvae. In, K.M. Wilbur and C.M. Biotropica 8(2): 71-95. Yonge (Eds.). Physiology of Mollusca, Vol. I, pp 197-210. Academic Press, New York, 10. Jurd, L., K. Stevens, and G. Manners. 1972. N.Y. Quinoid constituents of Dalbergia retusa heartwood. Phytochemistry 11: 3287-3292. 23. Culliney, J.L., P.J. Boyle, and R.D. Turner, 1975. New approaches and techniques for 11. Waite, J.H. and K.M. Wilbur. 1976. studying bivalve larvae. In, W.L. Smith and Dalbergia polyphenols and shell formation M.H. Chanley (Eds.). Culture of Marine in mollusks: Preliminary results. In, Invertebrate Animals. pp 257-271. Plenum J.D. Bultman (Ed.). Proceedings of a Work- Press, New York, N.Y. shop on the Biodeterioration of Tropical Woods: Chemical Basis for Natural 24. Saeki, A. 1958. Studies on fish culture in Resistance. Naval Research Laboratory, filtered closed-circulation aquaria. Bull. Washington, D. C. Japanese Soc. Sci. Fisheries 23(11): 684-695. 12. Oliver, A.C. 1962. An account of the 25. Kelley, E.E. Ideal configurations for semi- biology of Limnoria. J. Inst. Wood Sci. closed circulating aquarium systems. 9:32-91. American Society of Ichthyologists and Her- petologists, 50th Meeting. 13. Ray, D.L. and J.R. Julian. 1952. Occurrence of cellulase in Limnoria. Nature 26. Goldizen, V.C. 1970. Management of closed- 169: 32. system marine aquariums. Helgolanderwiss Meersunters 20: 637-641. 14. Ray, D.L. 1959. Nutritional physiology of Limnoria. In, D.L. Ray (Ed.). Proceedings 27. Kester, D.R., I.W. Duedall, D.N. Connors, of the Friday Harbor Symposium on Marine and R.M. Pytkowicz. 1967. Preparation of Boring and Fouling Organisms. pp 46-60. artificial seawater. Limnol. Oceanog. 12 University of Washington Press. (1) 176-179. 15. Sleeter, T.D., P.J. Boyle, A.M. Cundell, and R. Mitchell. 1978. Relationships between marine microorganisms and the wood-boring isopod Limnoria tripunctata. Mar. Biol. 45: 329-336. 16. Ray, D.L. 1959. Marine fungi and wood borer attack. Proceedings of the American Wood-Preservers Assn. 55: 147-154. 17. Schafer. R.D. and C.E. Lane. 1957. Some preliminary observations bearing on the nutrition of Limnoria. Bull. Mar. Sci. Gulf Carib. 7: 289-296. 18. Kohlmeyer, J., G. Becker, and W.D. Kampf. 1959. Versuche zur kenntnis der ernahrung der holzbohrassel Limnoria tripunctata und ihre beziehung zu holzzerstorenden pilzen. Z. angew. Zool 46: 457-489. 19. Ray, D.L. and D.E. Stuntz. 1959. Possible relation between marine fungi and Limnoria attack on submerged wood. Science 129: 93-94. 20. Wells, W.F. 1927. Report of the experi- mental shellfish station. Rep. N.Y. St. Conserv. Dep. No. 16, 1-22. 98 THE MOTILTIY OF THE CELLS AND THE EXTRACELLULAR MATRIX OF LABYRINTHULA -- A COLONIAL ORGANISM FOUND ON ZOSTERA MARINA D. Solomon, S. Sher and E. Bell Room 16-820A Department of Biology Massachusetts Institute of Technology Cambridge, Massachusetts 02139 Cells move within the matrix corridors at speeds up to 200 P per minute (7,19). Cells have no visible means of locomotion, i.e. no Labyrinthula is a saprophytic colonial cilia, flagella or podia (11). Cells without marine protist found on Zostera marina. It con- matrix are stationary. Changes of direction sists of cells which move only within an elastic with or without reversal of cell polarity and matrix thought to be synthesized and secreted movements past stationary cells have been documen- by the cells themselves. Although they have no ted by time lapse microcinematography using phase- visible means of locomotion, that is, no cilia, contrast and Normaski illumination. The decom- flagella or podia, the cells exhibit movements position of food organisms such as yeast, bac- at velocities up to 200 14/min. The extracell- teria, and diatoms occurs extracellularly but only ular matrix also exhibits motility. We have when the matrix is in direct contact with food found, with the help of time lapse cinemi- (11, 19). The lytic enzymes capable of such cropraphy, lamellipodia-like extracellular activity are presumed to be localized in the processes which exhibit rapid translocation matrix which in addition to acting as a conduit and appear to form the extracellular corridors for a vegetative cell also may serve as a within which the cells move. Indirect anti- nutrient reservoir. The matrix does not appear body staining shows antimyosin specific rails to be a site of protein synthetic activity (16, extending through the extracellular corridors 19). New branches of the matrix in the form and branching into the lamellipodia-like of fine filaments have been observed forming and processes. The rails appear to correspond to fusing at distances of 200 microns from a veg- regions of birefringence observed with polarized etative cell (2). light. Our observations have led us to con- struct a model which attempts to explain Labyrinthula motility. METHODS AND MATERIALS A. Culture of Labyrinthula INTRODUCTION Labyrinthula was isolated from free-floating strands of Zostera marina gathered in shallow water Labyrinthula is a eukaryotic colonial protist in Sippiwisset, Mass. during late May of 1977. suspected of being the etiological agent of the The strands were transported to M.I.T. in local wasting disease of Zostera Marina, reported to sea water and refrigerated at 4�C for 48 hours have occurred in temperate waters of the Atlantic before plating. Thirty-six segments (3 cm. in from 1930-34 (15). Out of convenience it has been length) were plated on 1% SSA (11) medium con- classified phylogenically as a marine slime mold taining the following antibiotics: 0.25 grams/ although it has no true affinity with the liter of streptomycin sulfate and 0.25 grams/ Mycetozoa (13,19). In its vegetative state, an liter penicillin G sodium salt (ICN Pharma- extracellular, anastomosing, network of unit- ceuticals, Inc.). The agar-sea water solution membrane-bound corridors encloses fusiform cells was autoclaved and cooled to 50�C before adding 10-15 microns long (5,11,12). horse serum and antibiotics. Labyrinthula growth was observed after 48 hours, extending an average The vegetative cells may have as many as of 3 mm from each of two segments. Colonies were 15-20 electron-dense invaginations on the cell excised with a micropipette and plated on fresh surface. These are called "bothrosomes" and are 1% SSA with antibiotics. This procedure was re- thought to be membrane synthesizing centers, but peated twice before plating on 1% without anti- a role for the "bothrosome" in membrane synthesis, biotics. Since then, labyrinthula cultures have matrix production, nutrition, locomotion or nter- been maintained at room temperature by transfer- cellular communication is still undemonstrated (11). ring every two weeks 1 cm blocks excised from 1% SSA which include the leading edge of a colony. 99 A marine yeast, Rhodotorula sp., was isolated 0.05 ml of 1% horse serum-sea water was from Zostera marina gathered from the same source. added to each cover slip. The dish was covered Twelve segments of Zostera (3 cm. in length) were and incubated from 8-48 hours before fixation in plated on MV (14) medium. Cubes (0.5 cm.) of agar order to permit the colony to extend beyond the medium containing Rhodotorula sp. and other organ- aggregation of agar. Cover slips were then gently isms were transferred to YM (14) acid medium, rinsed in 3.3% saline, pH 7.4, and fixed in 10% pH 3.5, which selects for Rhodotorula. They were Formalin phosphate-buffer saline (PBS) for 5 min. then incubated at room temperature for 14 days Fixed cover slips were immersed for 5 minutes in and transferred by pasteur pipette to YPGS (14), the following solutions; PBS),50% acetone-water, a maximum growth medium. Rhodotorula cells were 100% acetone, 50% acetone-water, and PBS. The grown by inoculating YPGS medium and incubating coverslips were incubated with ten lambda of at room temperature for 4 days with gentle, rabbit antimyosin antibody at 370C for 60 sterile aeration (1 liter air/min./liter of minutes. The cover slips were rinsed and incubated medium). The culture was subsequently centri- with twenty lambda of goat antirabbit antibody (GAR) fuged at 300xg for 15 minutes at 4�C, decanted and conjugated with fluorescein isothiocyanate for 60 resuspended in 3.3% saline, pH 7.4, and autoclaved minutes at 370C. Cover slips were rinsed in PBS for 20 minutes. Cell counts were made with a and mounted inverted on glass slides with a non- hemacytometer. fluorescent mounting medium (Aqua-Mount , Lerner Laboratories). Examination was with a Zeiss We found that the addition of 10 (5) auto- incident-fluorescence microscope. Microphotographs claved Rhodotorula sp. in 0.1 mm saline solution were taken with a Zeiss 35 mm leaf-shutter camera. layered on the petri plate (100 mm.) prevented the encystment of Labyrinthula. RESULTS B. Microcinematography A. Culture of Labyrinthula Time-lapse films of moving cells and matrix The species of Labyrinthula isolated from were taken with a 16 mm. Arriflex camera modified Sippiwissett resembled type LX reported by Watson for time-lapse sequencing. Exposures of 1/2 sec. in cell size, cell form, cell division and colony at intervals ranging from 1/2 to 5 seconds were morphology. Growth varied considerably despite made on Kodak Plus-X Reversal film. A Zeiss attempts to establish uniform conditions. "Slime- standard microscope was used with a 5X Zeiss way" width and cell concentration were related to ocular and 20X and 40X Zeiss phase contrast #2 colony age and moisture content of the medium as objectives. Colonies were prepared by inoculating others have previously reported. a Sykes-Moore chamber filled with a solution of 1% horse serum in 0.2 mp. millipore-filtered, Maximum growth occurred in colonies autoclaved sea water with 100 microliters of a inoculated with ldtcells of live Rhodotorula sp. 0.1% Labyrinthula culture. From a 2 mm. diameter inoculum, colony growth could cover a 100 mm petri plate in 85 hours at 25�C -- a radial increase averaging 0.56 mm. per C. Polarization Microscopy hour and an increase in area of 2500 fold. Polarization photomicrography was carried Autoclaved Rhodotorula sp. induced growth in out with a Zeiss Universal microscope fitted with colonies which had aggregated but which had not Zeiss polarizer and analyser, O1X Zeiss ocular, undergone a noticable change in cell morphology 20X and 1OOX oil Nikon HI rectified objectives, by forming encystment structures. Two months and a 0.00006mm mica compensator. Photographs after the addition of autoclaved yeast in these were taken with a Zeiss 35mm leaf-shutter micro- colonies growth was retarded or ceased completely. scope camera on Tri-X film pushed to 1600 ASA with However, the further addition of 10 cells of live, Dynafine. Specimens were prepared by culturing as distinguished from autoclaved, Rhodotorula sp. colonies on 25 mm diameter A.H. Thomas Co. cover induced growth approaching the maximum obtainable slips following the protocol described under in our cultures. antibody staining. Cover slips were mounted inverted in a Sykes-Moore chamber. Colony morphology varied in relation to the agar (Bacteriological Grade AgarAgar, Fisher Laboratories) or agarose (Agarose Powder, Bio-Rad) D. Antibody Staining concentrations. In 1% agar or agarose colony growth was three-dimensional from the start. Cover slip cultures of Labyrinthula were prepared by inoculating 12 mm diameter, #1 cover slips with 0.01 ml of 0.1% SSA medium drawn from the leading edge of a colony between two and six days old. Inoculated cover slips were placed on an open-cell urethane spacer in an 100 mm petri dish with approximately 10 ml of double-distilled water. 100 B. Matrix and Cell Motility Time-lapse films of expanding vegetative region of the lamellipodium as the beginning colonies revealed the presence of lamellipodia- of a new "slimeway", the presence or absence of like extracellular processes not previously reported. vegetative cells appeared to have no effect on The lamellipodia exhibited cell-independent motility. the developing morphology or motility of the They were observed to extend perpendicularly from lamellipodia. existing "slimeways", with a sheet-like form varying from 10 to 70 microns in width and 5 to 90 When the lamellipodium reached an existing microns in length (Fig. 1). As the lamellipodium "slimeway", fusion was observed to occur. On advanced (Fig. 2) vegetative cells were observed to fusion the lamellipodium, with its wide fair like travel into its distal portion only, but not into leading edge, formed a concave triangular the fan-like process itself (Fig. 3). Frequently, "slimeway" junction (Fig. 4), and vegetative such cells were observed to enter the new "slimeway", cells were observed to travel through the stop 15 to 60 seconds, and without turning around, junction into other regions of the colony. reverse direction and exit. Using the distal Fig. 1. Phase contrast micrograph Fig. 2. Sequence of lamellipodia of Labyrinthula vegetative cells and morphogenesis. Formation of anastomosed extracellular matrix. Lamellipodium lameelipodium on the right. extending to the right of the "slimeway." Bar equals 10 pm - Fig 3. Expansion of lamellipodium Fig. 4. Formation of a new "slimeway" to the left of the "slimeway.' in the left lamellipodium. C. Antibody Staining of Labyrinthula Colonies Indirect fluorescent antibody staining of formalin-fixed colonies by rabbit antimyosin antibody, specifically stained a branched network of extracellular cables extending throughout the extracellular matrix (Fig. 5). These tables correspond, in part, to the phase-contrast observed "slimeways" along which the vegetative cells move (Fig. 6). Cables were also observed throughout regions of the extracellular matrix where phase-contrast or Nomarski images of r"slimeways" were hardly visible. These thin cables were observed in high concentrations at "slimeway" junctions and in regions of the leading edge of the extracellular matrix of the colony. Within the lamellipodia, the stained cables are fan-like and are observed to converge as the ribs of a fan in a direction opposite to the leading edge of the lamellipodium. In some slimeways the antibody staining was observed to be in the form of two lateral rails which diverged to follow the perimeter of Fig. 5. Antimyosin antibody stained a cell and either converge or closely approach Labyrinthula cells and matrix. Bright each other after passing the cell. Between some fusiform shapes are vegetative cells. cells separated by less than 30 micron the Bar equals 10 pm separation between rails was uniform. A double-blind evaluation of the antijyosin staining and (Table 1) suggests that the stained structures contain myosin. TABLE I. DOUBLE-BLIND EVALUATION OF INDIRECT ANTIBODY STAINING AND CONTROLS (n=4). INHIBITOR 1st STAIN 2nd STAIN AVE. INTENSITY PBS RAM GAR + + + Myosin RAM GAR + Myosin RAM GAR + + PBS Myosin GAR - Myosin NRS GAR - PBS PBS GAR - PBS NRS GAR - PBS NRS GAR - PBS PBS GAR - Fig. 7. Antimyosin antibody stained "slimeway" junction. Note seperate cables GAR = Goat Antirabbit; NRS = Normal Rabbit Serum extending from the cell on the right. PBS = Phosphate buffer Saline; RAM = Rabbit Antimyosin lyosin and rabbit antimyosin antibody were pre- incubated at 37�C for 30 minutes. 102 D. Polarization Microscopy Observations with rectified polarization optics revealed the presence of birefringence within the "slimeways" as identified by phase contrast microscopy, and within regions of lamellipodia seen poorly with phase-contrast microscopy. Figures 8 and 9 show the bifringence of slimeways were taken on the + and - side of extinction. The pattern of birefringence appears to correspond to that of the antimyosin antibody cabling reported in this paper. Rapid changes (1-2 seconds) in sign and intensity of birefringence were observed in conjunction with and independent of cell movement. Certain birefringent cables have been observed to diverge as they approach vegetative cells, and converge behind the cells like the antimyosin stained cables. In addition birefringent zones, 3-4 microns in length, have been observed to travel along the length of the "slimeway" without apparent relation to the translocation of cells. As yet we have no evidence of what is responsible for the phenomena. DISCUSSION We have reported the existence in Labyrinthula of highly motile extracellular lamellipodia and a network of extraceilular cables or rails which stain with antimyosin antibody. The presence of lamellipodia at the leading edge of a colony as well as at points within the colony suggest that they play an important role in colony expansion and intracolonial slimeway development. The lamellipodium appears to be homologous with the leading edge of the animal tissue cell which extends on a substrate in advance of the cell during translocation. The putative presence of a network of fine myosin cables in the lamellipodium is suggestive that the contractile protein is implicated in the movement of the organism. Similarly the presence of putative myosin cabling in the slimeways suggests a mechanism for the motility of labyrinthula cells which possess no visible means of locomotion. Although we suspect the presence of myosin in Labyrinthula by demonstrating a protein band from colony extracts which comigrates with chicken gizzard Figs. 8 & 9. Polarization micrograph myosin during thin layer acrylamide gel of living cells and matrix on the + & - electrophoresis, we have not yet been able to sides of extinction. Matrix exhibits show the presence of actin. Nonetheless we distinct birefringance. Bar equals speculate that active filaments on the surface of 10 pmn the cell interacting with myosin cables of the slimeway could constitute the driving force ACKNOWLEDGMENTS needed for cell movement. Equally fascinating in the interaction between cells and extracellular The authors thank Peter Yanover and Dr. matrix are the signals and controls needed to L. Marek for their help in time lapse micro- establish direction and speed of cell movement. cinematography and, Charlotte Merrill for her Understanding the basis for Labyrinthula motility, technical assistance. Support for this project both cellular and extracellular, may have important was provided in part by the M.I.T. - U.R.O.P. ecological significance. If Labyrinthula is the fund. agent responsible for periodic eel grass decline we may discover factors which permit it to overun a host with which it is normally in balance. 103 REFERENCES 1. Amon, T.P. and Perkins, F.P. 1968. Structure 15. Renn, C.E. 1935. A Mycetozoan Parasite or of Labyrinthula Sp. Zoospores. J. Protozool. Zostera Marina Nature. April 6, 1935: 544- 15(3):543-546. 543. 2. Bell, E. 1969. Labyrinthula, a E.D.C. film 16. Solomon, D.J. 1977. Personal films on aby- with D. Porter as consultant. rinthula motility. 3. Bonner, J.T. 1967. The Cellular Slime Molds 17. Sykes, E.E. and D. Porter 1973. Nutritional (2nd ed.) Princeton Univ. Press, Princeton, Studies of Labyrinthula Sp. Mycologia. Vol. 205 pp. LXV, No. 6, pp. 1302-1311. 4. Hofer, H.W., Pette, D., Schwab-Stay, H. and 18. Watson, S.W. 1957. Techniques for the isol Schwab D. 1972. Enzyme Acitivty Patterns and ation of Labyrinthula and Thraustochytrium Mitochondria - cytoplasmic Relations in Pro- in pure culture. J. Bact. 73:368-377. tozoa. Comparative Study of Tetrahymena pyriformis, Allogromia laticollaria, and 19. Watson, S.W. 1957. Cultural and Cytological Labyrinthula coenocytis. J. Protozool. 19(3): Studies on Species of Labyrinthula. Ph.D. 532-537. thesis, Univ. of Wisconsin. 164 pp. 5. Hohl, H.R. 1966. The Fine Structure of the 20. Young, E.L. 1943. Studies on Labyrinthula; Slimeways in Lab. J. Protozool., 13:41-43. the etiological agent of the wasting disease of eel grass. Amer. J. Bot. 30:586-593. 6. Hollande, A. and M. Enjumet. 1955. Sur l'evolution at La Systematique des Labyrin- thula Algeriensis nov. sp. Am. Sci. Nat. Zool. 17:357-368. 7. Quick, J.A. 1974. A New Marine Labyrinthula with Unusual Locomotion. Trans. Amer. Micros. Soc. 93(1):52-61. 8. Perkins, F.O. 1970. Formation of Centriole and centriole-like structures during meiosis and mitosis in labyrinthula sp. J. Cell Sci. 6:629-653. 9. Perkins, F.O. and Amon, J.P. 1969. Zoo- sporulation in Labyrinthula; an electron microscope study. J. Protozool. 16:253-256. 10. Perkins, F.O. 1972. The ulstrastructure of holdfast, "rhizoids" and "slimetracks" in thraustochy-triaceous fungi and Lab spp. Arch. Mikrobiol. 84:95-118. 11. Porter, D. 1969. Ultrastructure of Labyrin- thula. Protoplasma 67:1:19. 12. Porter, D. 1972. Cell division in the marine slime mold, Labyrinthula sp. and the role of bothrosome in extracellular membrane pro- duction. Protoplasma 74:427-448. 13. Porter, D. 1974. Phylogenic considerations of the Thraustochytriaceae and Labyrinthu- lacalae. Veroff. Inst. Meeresforsch. Bremerh. Suppl. 5:19-44. 14. Porter, D. Unpublished methods and mater- ials of Labyrinthula and Marine Yeast culture. 104 A NON-ACOUSTIC LONG DISTANCE UNDERWATER COMMUNICATIONS SYSTEM Lt. R. G. Williscroft N. C. MacLeod National Oceanic and Atmospheric Administration Technology Development Corporation NOAA Ship Surveyor S-132 Underwater Systems 1801 Fairview Avenue East 155 Moffett Park Drive, Bldg. C Seattle, Washington 98102 Sunnyvale, California 94086 Abstract direct projection acoustic systems are severely limited by waterborne noise An underwater communications system is (Shilling, Werts, Schandalmeier, 1976). described which utilizes Technology Development Corporation's (TDC) Hydrocom The most common acoustic system uses a coupled to a Radio Retransmission Buoy. modulated carrier frequency between 8 kHz Hydrocom is nonacoustic, utilizing instead and 50 kHz in a single side band (SSB) an electric current field which is not suppressed carrier mode (Martin and Adams, adversely affected by thermoclines, rever- 1970). Amplitude modulation (AM) systems beration, or interferometer effects. It are possible, but half the output energy is can be operated in noisy and confined lost to the carrier. Frequency modulation environments where acoustic systems are (FM) systems, while producing excellent severely degraded and even when objects, results, are practical only over very short such as reefs or man-made structures, lie distances because the broad-band carrier between the communicators. The Radio frequency deviation requirements neces- Retransmission Buoy enables telephone sitate the use of ultrasonic frequencies clarity communications to exist between which are quickly absorbed in sea water divers or a submerged vehicle and a ship (Shilling, Werts, Schandalmeier,1976). SSB several miles distant. Suitable radio system at the low frequency end of the patch or satellite relay will extend the spectrum are characterized by relatively communications link to any place on Earth, long range (900 to 1,000 meters) and high enabling real-time laboratory overview of noise, while those at the high end of the distant onsite underwater inspection or spectrum experience comparatively little data gathering. noise but have one-fifth the range for the same output power (Martin and Adams, 1970). 1. Introduction All acoustic systems are subject to masking by waterborne noise and suffer from The history of underwater communications transmission losses due to absorption, has been dominated by hardwire systems for scattering, refraction, and multiple-path the tethered diver and by acoustic systems transmission. According to Woods and for the free-swimming diver (Shilling, Lythgoe (1971), absorption, which increases Werts, Schandalmeier, 1976). A hardwire exponentially with frequency in sea water, system is comparable to a telephone, sets a practical limit of about 0.5 MHz to including a transducer at both ends, ampli- low power transmission. Scattering is most fication methods, and transmission line. serious in turbid water or in the presence It will not be further discussed here. of large concentrations of fish, with the Acoustic systems fall into two categories: greatest losses occuring when the deep direct and modulated. scattering layer (DSL) rises near the sur- face at night time. For divers, communi- The simplest acoustic system is direct cation along nearly horizontal paths is projection in which the voice is projected most seriously affected by refraction. A into the water using an "underwater loud- temperature gradient of 0.07 degrees speaker" and is usually received by the Celsius per meter is sufficient to counter diver without any special processing. The the effect of pressure on the speed of simplest of these is a waterproof diaphragm sound in sea water (Woods and Lythgoe, through which the diver speaks. More com- 1971). Thus, normal near-surface tempera- plicated versions employ simple amplifi- ture variations can strongly affect cation devices at the transmitting, and acoustic communication ranges. Multiple sometimes even the receiving, end. All 105 reflections off the sea surface, rocky The quality of a signal transmitted via bottom, and interposing natural structures, ECF is subject to the limitations of the such as reefs and plants, and man-made equipment being used instead of environ- structures lead to signal distortion and mental considerations. The result is a fading. near-telephone clarity and intelligi- bility which surpasses any acoustic system. The current state of underwater communi- cations is aptly summed up by Penzias and Goodman (1973): "None of these (current 3. System Description underwater communication systems) is satis- factory .. "General. The Hydrocom Retransmission System (RS) consists of two primary com- The system discussed in this paper is ponents. The first is a standard Hydrocom nonacoustic and is unaffected by the losses Diver Unit, which provides voice communica- described above. In fidelity it exceeds tion between divers and which allows them acoustic systems and possesses a nominal to locate each other using the directional line-of-sight range to the horizon which characteristics of the ECF. The second is can be extended to any place on Earth with a small surface buoy which couples Hydrocom off-the-shelf equipment. Diver Unit electronics to a Motorola HT-220 Handi-Talki UHF transceiver. 2. Electric Current Field Diver Unit. The HRS-1 Diver Unit con- sists of a main housing which contains the An electric current field (ECF) is electronics and batteries, a microphone and generated when an electric potential is bone conduction earphone, ankle electrode, applied across two electrodes immersed in a and connecting wire. The housing is con- conducting medium. In normal sea water structed of 16-inch long, 3-inch diameter (salinity near 33 o/oo), this field propa- stainless steel tubing, capped with Delrin gates at a speed of approximately 3x106 plastic end caps sealed with 0-rings. it meters per second. The ECF spreads in a is rated to 300 feet. pattern similar to the magnetic field sur- rounding a magnet and produces a voltage The top cap contains the microphone- gradient which is directly proportional to earphone connector, electrode connector, the amount of current. A second pair of charge plug receptacle, and power/sensi- electrodes introduced into the ECF, as tivity switch. The internal electronics shown in Figure 1 , will have a voltage and power chassis are connected to the top induced across them. The voltage varies as cap. the cosine of the angle formed by the base- lines of the two sets of electrodes. Thus, The bottom cap contains the ground a second pair of electrodes oriented in the electrode and a large screw which holds the ECF, as shown in Figure 2, will experience unit together. essentially no induced voltage. The microphone and earphone are connect- In a communication system using ECF, a ed to the main housing with a single water- voice signal is impressed across the trans- proof connector. The microphone is pro- mitting electrodes, and a suitable means of tected with a special micropore bladder detecting and amplifying the received which allows it to equalize to 300 feet, signal is connected to the receiving and it can be installed in almost any full electrodes. For a given angular orienta- or half face mask. The earphone is a tion of the transmitting and receiving rubber-covered bone conduction unit. Bone electrode pairs (antennas), the voltage at conduction is used to eliminate the distor- the receiving electrodes is directly tion problems caused by the water-tympanic proportional to: membrane interface. (a) The output power (amount of current The ankle electrode is attached to the put into the field; ankle with a Velcro fastener and is con- (b) The length of the transmitting base- nected to the top cap by a simple insulated line (distance between transmitting wire. electrodes); and (c) The length of the receiving baseline The entire unit is provided with a (distance between the receiving bracket for mounting to a standard SCUBA electrodes). backpack (see Figure 3). and is inversely proportional to: (a) Water conductivity; and (b) The cube of the range between trans- mitting and receiving antennas. iO6 Retransmission Buoy. The HRS-1 Retrans- D-Mode. In the D-mode, two or more mission Buoy consists of a 24-inch long, divers are equipped with HRS Diver Units. 5-inch diameter stainless steel tube with Communication range is generally less than Delrin plastic end caps sealed with 400 feet and is optimum between divers who O-rings, an electrode with 150 feet of have the same general orientation to the connecting wire, a 39-inch antenna, and a water column, i.e., both horizontal or float collar (see Figure 4). The internal vertical. A diver wishing to communicate electronics couple a standard HRS-1 Diver uses a "throwaway" word to activate his VOX Unit transceiver to a Motorola HT-220 and then speaks normally. The other divers Handi-Talki UHF transceiver in a bi- will hear him speak. To shorten the trans- directional simplex arrangement. The buoy mission range (to avoid interference with a automatically receives any Hydrocom signal nearby group, for instance), the ankle within its effective range and retransmits electrode need only be moved nearer th it on whatever UHF channel the internal backpack. To enable all other divers to Handi-Talki is set; and it also automa- home in on one diver under low visibility tically retransmits underwater any signal conditions, that diver activates his trans- it receives on the UHF receiver. mitter and talks continuously, i.e., count- ing, while holding himself horizontally Electronics. HRS-1 electronics consist oriented. The other divers place them- of a transmitter, receiver, and selves horizontally and rotate through the voice-operated transmitter (VOX) circuits. azimuth until they detect a transmission The 30-watt transmitter is powered by a null. They then swim towards the null. 24-volt nickel cadmium battery pack. The receiver incorporates automatic gain con- Retransmission Buoy is placed in the water trol (AGC) circuitry, which maintains near- Retransmission Buoy is placed in the water ly constant volume in the diver's ear. The cate among the mselves in the 0-mode , but VOX circuit is a two-stage, two-level ate among themselves in the D-mode, but device. An initial "throwaway" word, such all transmissions are simultaneously as "Hi," is used to trigger the first retransmitted by the buoy on the selected stage. The first stage threshold is set UHF channel for surface reception by any stage. The first stage threshold is set high enough so that hard breathing noises portable or fixed radio capable of high enough so that hard breathing noises or bubbles do not trigger the transmitter. receiving on the divers is usu ally made trans- Once the first stage VOX trips, the voice mission to the divers is usually made with Once the first stage VOX trips, the voice a hand-held or other portable unit. The a hand-held or other portable unit. The may be returned to a normal speaking communication range between surface unit level. The second stage VOX then takes and divers is limited to the normal range over and continues transmission until the of UHF units (usually li ne of sight to the diver stops speaking or breathing for of UHF units (usually line of sight to the diver stops speaking or breathing for horizon). Homing on the buoy can be approximately one second. effected in a similar manner to D-mode The system is normally in the receive homing. mode. When the VOX circuit is triggered by DR-Mode. In the DR-mode, two HRS the diver's voice, the receiver is disabled Retransmissio n Buoys are used. Each buoy Retransmission Buoys are used. Each buoy and the transmitter activated. is associated with a group of divers. In the HRS-1 Retransmission Buoy, the Although the two groups of divers can be widely separated, a diver from one group receiving Handi-Talki squelch circuit acti- widely sepa rated, a diver from on e group vates the VOX circuits, and the received can communicate with a diver from the UHF signal is coupled to the Hydrocom within radio range of each othe buoy s are transmitter. When the buoy receives an wi thin radio range of each other. This underwater signal, that signal activates mode is a logical exte nsion of the the Handi-Talki transmitter and is then DRS-mode, wherein a surface unit is con- broadcast on the selected UHF channel . trolling two sets of divers on a coordi- (The HRS-2 Retransmission Buoy incorporates nated underwater project. a tone generator in the interface to effect more positive control over the switching extension of the DRS-mode in which the process.) surface unit is patched by some suitable means to a very distant receiver. The 4. System Operation two-way communication link can be effected by a standard radio patch or even by The HRS operates in four modes. D-mode satellite relay. The essential innovation here is that the signal from the diver (diver to diver) is the basic Hydrocom in here is that the s ign al from the diver action and preceded HRS. DRS-mode (diver arriving at the surface unit is already in retransmit short range) is the principal electromagnetic form and requires no HRS mode. DR-mode (diver to diver retrans- further processing for relay. mit) and DRL-mode (diver retransmit long range) are special applications of the DRS-mode. 107 5. System Potential 6. Summary The Hydrocom Retransmission System Diving operations require a speech com- utilization potential falls into three main munication system that can be carried by areas: communication, monitoring, and con- the diver without hampering mobility and trol. Communication is primary and has safety; can provide easy two-way intelligi- been discussed in some detail already. ble speech; is reliable, effective, effi- Beyond the obvious appeal of enhanced local cient, flexible, and simple to operate; and communications lies the possibility of has a reasonable signal propagation range. direct laboratory control of distant The Hydrocom Retransmission System (HRS) instrument emplacement and maintenance. meets these requirements and, in areas of The value of real-time system feedback can intelligibility and range, far exceeds never be overestimated, and the HRS brings anything available today. HRS provides this benefit to the field of diver-emplaced near-telephone clarity communications over instrumentation. unlimited range. It also provides the diver with a simple-to-operate homing ECF has been shown to be responsive to system. objects placed within the field (U.S. Navy, HRS can be adapted for applications in 1972). This principle of "electric anomaly remote surfaced and submerged traffic detection" has applications ranging beyond monitoring and control, accurate, inexpen- submarine and swimmer intrusion detection sive underwater object location and plani- for the military. Applications range from metric survey, real-time telemetric and remote fish counting for a marine biologist feedback analysis of remote underwater to a relatively inexpensive system for the data, and various anti-collision devices. marine traffic manager who continuously monitors shipping location and flow within ~~~~~~~~~~~his area. ~7. References his area. In June 1977, TDC Underwater Systems Martin, W. R. and R. L. Adams, personnel demonstrated the feasibility of "Physiologic Factors in the Design of telemetry transmission utilizing ECF. Underwater Communication," Marine Applying this capability to HRS results in Technology Society Journal, 4T.33S the unique ability to apply real-time (May-June 1970). sophisticated computer analysis to distant Penzias, W. and M. W. Goodman, Man underwater events. A pilot study has also Beneath the Sea, p. shown that ECF techniques can effectively Wiley-Interscience, New York (1973). be applied to acoustic net and to azimuth- Shilling, I. W., M. F. Werts, N. R. dip-range planimetric survey methods. Due Schandalmeier, "Underwater Communications," to its greater in-water velocity and its in I. W. Shilling, M. F. Werts, N. R. freedom from environmental degrading, a Schandalmeier, eds., The Underwater survey technique based upon ECF is in- Handbook, Plenum Press, New York (1976). herently more accurate than acoustic U.S. Navy, Phase I Final Report, Bridge methods, and its design simplicity results Defense System Feasability Model, Contract in considerably lower manufacturing costs. N61339-71-C-0230 (Confidential) (February 1972). HRS bidirectionality is an additional Woods, J. D. and J. N. Lythgoe, factor. Any underwater system which would "Apparatus and Methods for the Diving benefit from direct or feedback control is Scientist," in J. D. Woods and J. N. a potential area for HRS application. The Lythgoe, eds., Underwater Science, pp. automatic positioning of a floating plat- 1-31, Oxford University Press, London form over a submerged reference point has (1973). been demonstrated at TDC using scale models. The marine traffic monitoring system discussed above could be expanded to include a passive navigating system with shipboard readouts, enabling bridge per- sonnel to monitor the vessel's channel position. An automatic pilot could be incorporated into such a system. With suitable shipboard alarms, bridge pylons and submerged obstructions could be marked with an ECF for an effective anti-collision system. 108 LIFT LINE DYNAMICS - A TECHNIQUE TAILORED FOR THE USER W. S. Pope C. S. Albro Principal Research Engineer Research Engineer Battelle, Columbus Laboratories Battelle, Columbus Laboratories 505 King Avenue 505 King Avenue Columbus, Ohio 43201 Columbus, Ohio 43201 Abstract problem, we became concerned that the mariner who would have to use a previously selected A technique is proposed as an aid to the line, and possibly make some on-site decision marine salvage master at sea and the ocean based on changing circumstances, was not being engineer. The aid may be used to find line considered. size, factors of safety, maximum operating conditions, salvage object parameters and Suppose, for instance, that a line has other useful information relating to the been selected for a specific lift task; e.g., lifting or lowering of heavy objects at sea, lift a specific object from a specific ship A computer program based on Arthur D. Little, in a maximum depth and sea state. Further Inc. equations for maximum dynamic line force suppose that when the expedition arrives on is used to generate a number of parametric station: the object has tipped over and filled curves for a specific ship lift point and line with mud, or slid into deeper water, or the type which are then collapsed into one simple seas are higher than anticipated, or the equation--the key formula--for dynamic factor primary line is damaged and a backup employed, of safety. This is combined with the easily or any of a host of occurrences which could obtained static factor of safety, and the change the parameters of the lift. The results displayed on a set of nomographs and salvage master would then have to decide charts, whether to proceed, abort, or change the situation enough to proceed. The proposed The paper explains the need for such a method will help in making that decision, and scheme, what the basic equations mean, their many others, concerning the selection, use limitations, and how the charts and nomographs and limitations of lifting lines. are used. 3. How the Technical Aid is Used 1. Introduction The salvor would have in front of him a The lifting or lowering of objects at sea set of charts and nomographs which would relate from a floating platform has always been a the lift parameters. His job would be to risky and expensive business. There seems to determine the numbers for the parameters for be no widely accepted design method for selec- which values are available or can be estimated, tion of line and figuring safety factors. In- enter the charts and determine the remaining stead a wide variety of rules of thumb, unknown factor for the lift. All of the heavy manufacturers' recommendations and computer analytical work has been "front end loaded" by programs are employed. This paper illustrates the engineers preparing the charts, so that the an approach to simplifying the problem of effort in actually using them is brought to a matching the characteristics of the line, plat- minimum number of steps. form, and lifted object. Each set of charts is prepared for: In this paper we have proposed a straight- forward, rational technique for line selection o A specific ship and lift point (e.g., based on best engineering information and a stern "A" frame on an offshore analysis, which avoids ad hoc computer solu- supply boat, as in our example) tions for each and every lift situation. The results are presented in such a way that they o A specific type of line (e.g., Samson may be used by the salvage master at sea, as 2-in-1 Power Braid, also used in our well as the ocean engineer in the design example). office. Thus each set becomes part of the ship's equip- ment, usable indefinitely, as long as the 2. The Problem Facing the Salvor hydrodynamic characteristics of the ship are not altered. After analyzing a number of specific lift cases, and looking at various aspects of the 109 4. Preparation of the Charts In Ref. 1, the sensitivity of the solu- tions for dynamic stress to the degree of The information needed to prepare the compliance with some of the above assumptions charts consists of: was examined. Liu(3) looked at the sensitiv- ity of the solution to varying values of each o Ship Response Characteristics - re- independent variable, and in Ref. 4 reported sponse amplitude operator (RAO) is on an actual at-sea test conducted by Muga to used in this analysis verify analytical predictions based on these equations. o W o Sea State Formulation - the Bretschneider sea state spectrum was chosen for our In general, the limitations are reasonably example, for Sea States 3, 4, and 5 well met in normal practice, with certain spe- 'a cific exceptions, in which case caution must o Line Characteristics - modulus (average be exercised in interpreting the stress results. rxw value at 20% of breaking strength), In particular, the non-linearities involved in density the line modulus, object damping and hoisting speed have more pronounced effects on stress o Generalized Salvage Object Charac- near resonance, when the drag to inertia affects teristics - size, shape, weight, ratio is either very low or very high. density .0 The Arthur D. Little equation can be o Lift Operation Characteristics - line expressed as follows: diameter and maximum length. I =2 ,2 2 2 This information is input to a computer program, )11 which increments the variables so as to map (tan i + tan w')] regions of maximum line load for the expected ranges of all variables. where w' is the dimensionless frequency The output is then plotted parametrically T' is the dimensionless load displacement U1 and examined for best fit to simple power function relationships. The result is a 2' is the dimensionless tension of the relatively simple formula for total dynamic 0 to end o f line safety factor as a function of three parameters: a "nominal" dynamic safety factor, a sea state and i', U', ( and ~ are functions of the parameter, and an object parameter. Finally, line and load properties and/or the static and dynamic safety factors are combined into a total safety factor. All parameters are displayed in nomograph form This, and a companion equation, which calculates for ease of manipulation. the maximum tension anywhere in the line, can be rearranged to provide a response function, 5. The Dynamic Model RI U The mathematical model used to compute o dynamic stresses was the Arthur D. Little, which represents the maximum line Inc.(1,2) boundary value equation, derived tension, D, wherever it occurs in a given line, from the basic equation of motion of a line divided by the excitation amplitude, Uo, as a segment. The following limitations apply to function of frequency. Figure 1 is a series of its use. power spectral density plots which illustrate the following sequence of steps used to arrive at 1. A single, constant diameter cable the values for significant dynamic tension. 2. Constant cable modulus of elasticity 3. Cable hangs vertically under lift 1. The amplitude of response of the lift point point on the ship to unit wave height 4. Ship exciting force is sinusoidal, or is determined; this is the RAO. can be represented as a sum of sine waves 2. A sea way power spectral density is 5. Damping on the cable is neglected constructed for the sea state of 6. Rate of raising or lowering is small interest. This is S(H,n); the area relative to the exciting velocity relative tothe excitig velocityof this curve represents significant 7. Sea state is statistically stationary o f his cur vere ofi1/3 and ergodic ~~~~~~~~wave height H1/3 (average of 1/3 and ergodic highest waves). 8. The ship RAO function remains un- changed over time 3. The lift point excitation spectrum 9. The damping force on the object is U2 will be the product of the linearized. (H W) RAO spectrum and the sea state 110 FIGURE 1 spectrum. (As noted, the area under the U2(H) curve is the square of the significant excitation, U 1/3 (RAO) \(The response equation for dynamic 14t~~ @ / \ ~~~~~force involves a linearization of .�cc~~~~~ -44O~~ /the normal'4elocity squared" drag WX~~~~~~~~~ S ) ~term in the equation of motion of 0 o the load. U1/3u becomes a synthetic velocity, used in the response equa- tion, to represent the same energy dissipation as the velocity squared term in the equations of motion.) >/ \Area = H 4. Finally, the product of the lift m (2H4 u) I 1/3 point excitation spectrum U21/3(H,) 'S(2H,w apd the line dynamic response function R is the Dynamic Force Spectrum a Xa X X \ ~~~~~~~~D ( the square root of whose Xo~~~~~~~~ Z \ ~area is the significant line tension, W~~~~~~~~~~~~~~~ U,0P~~~~~~~ / f\D 1/3. A calculation is made for dynamic force -- ~~~~for incremental values of line length from 0 to 20,000 ft. Each answer is compared to the one in memory and replaces it if the new answer is higher. Thus, each combination of lift O U2 o (HU) parameters (ship characteristics, sea state, object parameters and lift line parameters), X43~~ / \ ~~~~~produces one answer which represents the high- est value of significant dynamic force en- countered in the line, at a length not speci- ofi~~ X X \ ~~~~fied, during a transit between surface and .O XArea = / 20,000 ft. 2 U 1/3 These results may also be presented in 1/3 X-,4~~~~~~~~~ /terms of a probability to exceed a set point dynamic load in a given time. However, it was felt that the approach of presenting simple forces and factors of safety has more universal ~~~~~~~~~O B ~appeal. O -,4 04-' X R2/ 6. Results and Discussion iW W C@~~~~~~~ a / ~For each support platform, the total factor a0 of safety can be expressed as: - 1 B B B T I l + I s W D D a f D~~~~ a-ff~~~~~~ n .where B = breaking strength of line, lb 2~~~~~~,~~~~~Wo 2 = wet weight of object (dry weight DI minus buoyancy) (U,W) vU,~ a J ho' \ ~~~~D = dynamic line force, lb Ao~~~~~~~ / \ ~Our curve fitting exercise resulted in the o 2 Ae= /\following equation for fD: tArea =i3 f r1 a,,Z bf 0| / ~D = K ~U w1~~ ]a[KzZ] Wfd] ~Frequency ~~ 0 Frequency where, K = constant differences in max. load between nylon and power braid are very small U w = significant excitation amplitude oven we raid aretery so 00 times frequency of highest energy oe-ierne fprmtrs that both may be used on these graphs, input if the appropriate breaking strength a,b = exponents, derived from curve tables are consulted. fitting o The particular relations derived here d = nominal dynamic safety factor were for nearly buoyant synthetic d WD line; the basic dynamic equations would still be valid for metal or with, WD = dynamic weight, a factor account- Kevlar lines, but more curve fitting Ding for wet weight plus added mass Kva iebtmr uv itn effects dueto weightrplusaied mand would be needed to accommodate a more entrpecsdu atoetraruned and in-marked dependence of the dynamic load entrapped water, around and in o ie-egt side the object. This derived equation could in fact be used by the salvors and engineers*, however, 7. Technical Aid Example we felt that a nomograph presentation would be (See Figure 2) more accessible and also allow easier manipu- To illustrate the use of the charts and lation to find answers other than fT Vnomographs, take the following situation: A worked example is shown in Figure 2 along with the nomographs and chart required; s trengthetic lin, b additional information, which would be included to help in estimating values for CD and CM,o Spherical Salvage Object - has been omitted for brevity. Wet weight, W0 = 15,000 lb Four points are worth noting: Mass coefficient, CM = 1.5 I + added mass coeff Total volume, V = 310 cf o The salvor need not be concerned with Dral volume, C = 22 sf Drag Product, GDA = 22 sq.ft. critical length. This is so because computations for stress are made at all lengths, but only the greatest Find: what is highest sea state which can stress is retained. For neutrally be tolerated for a total factor of safety of 3? buoyant nylon line (our test case) the critical length usually occurs Nomograph A (line (1)) within about 1200 feet of the surface, so that even much deeper lifts will 1. Connect V and CM; read added weight WA = still be governed by the higher loads 30,000 lb as the object passes through this critical region. 2. Dynamic weight WD = Wo + WA = 15,000 lb + 30,000 lb = 45,000 lb o "Wet" weight (weight in air minus weight of displaced water) and en- Nomograph B (line (2)) closed volume of object are the key parameters used in the nomographs and 3. Connect B and fT, extend to read total line they are most readily estimated by force, D + W, = 34,000 lb the salvor. 4. Dynamic force D = (D + Wo) - W, = 34,000 lb - o It is not necessary to make an estimate 15,000 lb = 19,000 lb of E, the line modulus, as this has been done in the initial calculations. Nomograph C Breaking strength is needed, but is a much more readily available number. 5. Connect WD and D, extend to find WD/D Noteworthy in this regard is that the 2.37 (line (3)) 6. Extend line at WD/D = 2.37 to the right, * A computer run was made as a check in which parallel to Z-axis (line (4)) the dynamic and total safety factors were calculated from the A. D. Little equation and 7. Connect WD and CDA, find Z = 4.8 (line (5)) the derived equation. For the 270 distinct data points, the standard deviation between 8. Go to sea state chart, intersection of WD/D= the two was 5.2% for f and 2.7% for f2.37 and Z = 4.8 (line (6)) gives max D ~ T operating sea state =3 112 Optional Iteration 1. Salvor adds a light disk 5.3 ft in diameter a distance above object; this provides addi- tional added weight effect (line (7), nomo- graph A) in the amount of 5000 lb (the disk acts as if it were a sphere of water of the same diameter whose volume is 80 cf). 2. The effective drag-area will be doubled. 3. In nomograph C, connect new WD = 50,000 lb with new CDA product = 44 to find new Z = 9.0 (line (8)). 4. Connect new WI with previous D (unchanged from requirements imposed in Nomograph B) extend to find WD/D = 2.63, line (9). 5. Intersect new Z and new W /D to find increased max sea state >2. In addition to the examples cited here, any of the other parameters besides sea state may be treated as unknowns and solved for by use of the nomographs, as long as the remain- ing factors are known or can be estimated. References 1. Arthur D. Little Inc., Tech Report No. 1370863, Stress Analysis of Ship-Suspended Heavily Loaded Cables for Deep Underwater Emplacements, Cambridge, Mass, August 1963. 2. Arthur D. Little, Inc., Memorandum Report No. 3030365, Some Design Considerations for Deep Water Suspension Lines, Cambridge, Mass, March 1965. 3. Dynamic Tension Analysis of a Simple Lift System - a Computer Method, by C. L. Liu, Naval Civil Engineering Laboratory, Port Hueneme, California, ASME Paper No. 71- UnT-7, 1971. 4. Naval Civil Engineering Laboratory, Technical Report No. R703, "Dynamic Stress Response of Lifting Lines for Oceanic Operations", Port Hueneme, California, 1970. 5. Computation of Relative Motion Effects in Offshore Supply Operations, by J. R. Paulling, P. D. Wood, University of California, OTC Paper No. 2634, 1976. 113 v C DTWO~~~~xIOOO) B(XIOOO) 100 TWxI00 -5. X1oO 4 5. --800- .800 -200 -4-560 .600 ..4.0 -10 100 - 40 40 80 . .3.5 -.20 --40 -.3.0 2 - -2.52 - -~~~~~~~~~~~~~~~~1 400 10 80 2. 60() 80 8()-2. 688 --40 - 4 1.5 -200 *-2 0 (7) - 2 -. 1 .~~~~~~~~~~~~~~~~~~400 10 -1.0 -- Noorp1 Nomograph ANmorhB CDA ZWDxOO DxIO) DD -..200 - 300 7 DX00 D6lO) W/ ~200 .5 9- --2 Sea State 3 --100 10..I0 -80 .60 .60 -40 .15 45 ..8 --10 (9) 10 08 -- 68) --30I (5) ~--40 .3 --10 -- 2 50 --40 (6) --8~~~~~~~-5 6 -- 1 -.60 ~-60 1 *- ... 2 4 6 ~10 26 48 8~80L00 2b)0 i0 Nomograph C Z FIGURE 2 114 EXTERNAL WORK OF BREATHING OF DEMAND-AIR UNDERWATER BREATHING APPARATUS* J.J. Grodski Defence and Civil Institute of Environmental Medicine 1133 Sheppard Avenue West P.O. Box 2000 Downsview, Ontario, M3M 3B9 Canada Abstract The purpose of this paper is to show how practical, automated calculations of the external Underwater breathing apparatus can be work of breathing can be obtained. Additionally, effectively evaluated by measuring the external performance of various demand-air underwater work of breathing. This paper reviews the exist- breathing apparatus is examined by relating the ing and the proposed standards for demand-air experimentally determined external work of breath- breathing apparatus and recommends improvements to ing to the Bentley-Reimers limit. A modification the Bentley-Reimers criterion. An automated meth- of the Bentley-Reimers criterion consistent with od of collecting data on the external work of the U.S. Military Specifications6&7 (Mil-Specs) breathing is also described along with test re- and the current technology, is proposed for the sults of various demand-air breathing apparatus. shallow end of the depth scale. The test and evaluation data in this paper sup- port the proposed, modified Bentley-Reimers 2. The External Work of Breathing criterion. The cumulative product of the pressure 1. Introduction and the volume of air moved at each instant is equal to work: The life support systems designer is continuously attempting to provide divers with low work = pdV (1) breathing-resistance life support equipment. The ultimate goal, of course, is to reduce the exter- Fig. I shows the pressure-volume (p-V) loop of a nal work of breathing to zero. Numerous investi- breathing cycle. The area within a p-V loop is gators have attempted to define acceptable exter- proportional to the respiratory work, which can be nal respiratory workload standards for men wear- determined by simultaneous plotting of the vari- ing respiratory apparatus. Silverman et al.1 ables p and V in appropriate units and measuring conducted the first comprehensive study of the the resulting area. The measurement is often per- physiological effects of resistive breathing at formed manually but such a procedure is laborious. atmospheric pressures. Cooper2 reviewed the work The same may be accomplished quickly and with a of Silverman and co-workers and recommended stand- high degree of accuracy by performing a programmed, ards for maximum permitted total rate of respir- numerical integration using a suitable calculator. atory work done on a breathing apparatus. More recently, Bentley et al.3 investigated the tol- erance to the external respiratory work of miners at atmospheric pressures. They proposed that the standard should use the level of the to- EXHALATION tal external respiratory work per unit ventilation Doak CAp),,, at which 90 percent of the tested population does not experience respiratory discomfort. Their ex- -WORK perimental data showed that the level corresponded to 0.17 kg.m/liten of air breathed. Others pro- posed ekther higher or lower acceptance criteria. :AMBIENT Reimers summarized these proposals and suggested 'VOLUME that Bentley's criterion should be adopted for ILXXVN testing underwater breathing apparatus and that this criterion should remain unaltered for pres- sures (depths) up to 1000 feet of seawater (few). This Bentley-Reimers limit of 0.17 kg.m/liten of Peak (AP).. ventilat ion has been used in recent underwater breathing apparatus evaluations at the Defence INHALATION and Civil Institute of Environmental Medicine TIDAL VOLUME- (DCIEM) (unpublished) and the U.S. Navy Experi- mental Diving Unit5 (NEDU). Figure 1. Pressure-volume loop of a breathing cycle and its features *DCIEM Publication No. 78X18 115 This latter approach was used in this work. Addition of equations (2) and (3) yields an equation for the peak total pressure difference: The total respiratory work (internal plus external, if any) is calculated from measurements (Ap) T[cm 2 01 = 11.5 + 0.1425 x (depth[fsw]) (4) of the differential pressure between the thoracic cavity and the external ambient pressure. The ex- It can be readily shown that under steady flow con- ternal work of breathing for demand-type of under- ditions, the sum of the volume-average respiratory water breathing apparatus is determined from the and expiratory pressure difference equivalent to pressure differences between the airway (mouth) the Bentley-Reimers limit of 0.17 kg.m of work per pressure and the hydrostatic reference pressure at liter of ventilation is 17 cm H 0 Bentley et al.3 the level of the centre line of the second stage discuss implications of the various waveforms on diaphragm. In dry testing, the ambient pressure this conversion. A set of breathing simulator is used as the reference. Wet testing of other tests conducted at DCIEM on two different single- than the demand type of apparatus involves a dif- hose regulators provided a direct measurement of ferent approach4. The meaning and significance of the relationship between the sum of the peak in- the external respiratory work of breathing is re- spiratory and expiratory pressures and the external cognized by both engineers and physiologists. work of breathing. Fig. 2 shows the spread of the data with both parameters normalized to the zero 3. Respiratory Impedance Standards depth (surface). The slope of an equivalent dimen- sional relationship was found to be constant, ap- Mil-Specs6'7 for demand-air, double hose proximately .009 [(kg.m/f)/cm H20]. Therefore, and single hose SCUBA regulators in depths from the Bentley-Reimers limit is equivalent to the sum the surface to 200 fsw provide the single, pre- of the peak inspiratory and expiratory pressures of sently accepted breathing impedance standard. The 18.9 cm 20 and the pressure dependence of the com- mechanical inhalation and exhalation peak pressure bined specifications shown in equation (4) has an limits specified by the Mil-Specs may be described equivalent form involving the external respiratory by the following equations: work per unit ventilation: (AP)EXH[cm H20] = 5.0 + 0.075 x (depth[fsw]) (2) work[kg.m/i] = 0.1035 + 0.00128 x (depth[fsw]) (5) (Ap)INH[cm H20] = 6.5 + 0.0675 x (depth[fsw]) (3) or in metric units, The peak pressures are readily measured and hence work[kg.m/Z] = 0.1035 + 0.0042 x (depth[msw]) (6) the standard has been used in numerous equipment evaluations. In fact, because of the lack of gen- The line representing this equation intercepts the eral standards, the Mil-Specs were applied to Bentley-Reimers limit at the pressure equivalent to other types of underwater breathing apparatus for 51.93 fsw or 15.83 meters of seawater (msw) as which they were neither intended nor well suited. shown in Fig. 3. At the shallow depths, the Mil- Moreover, the physiological significance of the Specs standard is more severe than the Bentley- peak pressures is obscure and the need to tighten Reimers proposal. Moreover, recent material and the spandard is well known8. The Reimers pro- technological improvements incorporated into the posal eliminates these difficulties and provides a standard, acceptable in both physiological and (fw) 00 engineering terms. It is, therefore, interesting 0.2 et - to make a comparison of the Mil-Specs standard to the Bentley-Reimers criterion. . BENTLEY-REIMERS en 0.17 ~ BENTLEY-REIMERS Oi .17 .. O-SCUBAPRO MARK V " *EQUIVALENT OF THE U.S. X-POSEOMN CYCLON 300 X MIL-SPEC STANDARD (~~~~~~~~AOD~EPrTH ) X/, 0.__ PROPOSED AN m ~~~~~~~~~~~~~~~~~MODIFICATION .ASIMFACE TOTAL LB 0 - LL. z / z I (PDEPT : ' : : / ~~~~PSURFACE/]TA -s - s N in w TOTAL LLI a aGO 0.5 1.0 1' 2.0 25 DEPTH (mew) Figure 2. Normalized total external respira- Figure 3. Bentley-Reimers acceptance cri- tory work vs normalized sum of terion and its proposed modifi- mechanical peak pressures cation 116 equipment warrant a tightening of the Mil-Specs standard. It is therefore proposed that the BREATHING 8 PORT ON Reimers proposal4 be modified for demand-air VELOCITY [MACHINE TEST CHAMBER breathing apparatus- within the depth range of 0 to RANSDUCER 20 msw (65.6 fsw), to include a change (Fig. 3) '\ CHAMBER described by the following equation: PRESSURE I EXHALE work[kg.m/i] = 0.1 + 0.0035 x (depth[msw]) (7) I Q [- This modification is not significantly more res- INHALE trictive than the corresponding range of the Mil- MOUTHPIECE Specs standard, and further diminution of the limit - . ... is possible. It should be recognized that the pro- vision of optimal breathing conditions in this [POSIT IDON shallow range is particularly important because it DEODULATRANSDUCER -TO affects the largest diving population. TRANSDUCER ELECTRICAL 4. The Experimental System FILTER FILTER PENETRATOR The hyperbaric unmanned respiratory test- TO ANALOG ing facility (HURTF) at DCIEM has been used for the RECORDERS measurement of external work of breathing on under- water breathing apparatus. The facility includes I I DVM I DVM a hnperbaric chamber, a breathing simulator sys- tem schematically shown in Fig. 4, and analytical PLOTTER ition system, schematically shown in Fig. 5. The HP breathing machine uses the reciprocating action of IBI [II a piston moving within a cylinder to simulate the 16K ventilatory action of the human lungs. TAPE CALCULATOR MEMORY STORAGE During the experiments, the momentary piston position and the pressure difference be- Figure 5. Schematic of the fast data acqui- tween the mannequin mouth pressure and the refer- sition system ence pressure were measured concurrently. The resulting piston position and the piston velocity waveforms, with the velocity of the driving motor preset to a constant value throughout the cycle, are shown in Fig. 6. The waveform of the flow at > the mouthpiece, affecting the pressure difference, n is related to the piston velocity waveform. The O discontinuity in the sinusoidal pattern of the 1TtliI l, h Ih!'.iii iI Gl;i:i piston velocity waveform is due to the backlash I existing in the drive linkage. A linear trans-l l: ducer (SCR Division of Moxon Inc., Model Dll10) O I I I I il!I and a pressure transducer (Validyne Eng., Model DP15), equipped with a diaphragm rated at 1 psi, Ol I I I :I I I I !i i were used to measure respectively the piston posi- tion and the pressure difference (Fig. 5). The INHTLE ! EXHALE I HYPERMaC TEST CHAMER O I I 1 f i1lF o ' eliil Il I I i i 1 ,il:1lliit Hdiii / E~~~o44 N EQul.I~ C O h " 0ili liii I I I ihl ~: i ii I liii iii4iT I /--" / � 711i : ,J-m :l; ::l I I I 1 :1 iil;. - I .. Jif .lt /_~~ | / r WETI POT l liil:l!lii : l: I iil !g!lfii1if!liFl liL EL ,: i | I !! 1: I I ;i I ilWl ii i 41 [:i2,,.;_:JF EREATHING ] & I -ENVIROMENTAL Figure 4. Schematic of the Me Fure V iir configuration ocity waveforms of DCIEM breath- pxnmlm V I, I O glilG ii~Iti~f~pIffifflRf |1 ) f \ /. | * 3 SEC Z Figure 4. Schematic of the hyperbaric Figure 6. Velocity of the driving motor, chamber-breathing simulator piston position, and piston vel- configuration ocity waveforms of DCIEM breath- ing machine 117 transducer was powered from a carrier demodulator particular total, inhalation, and exhalation data (Validyne, Model CD19). Routing of the measured points in Fig. 8 with the corresponding p-V loops signals through low-pass filters (Krohn Hite, Mod- in Fig. 7. el 3342) elimintated noise from the signals before Fig. 9 shows examples of p-V diagrams ob- they were fed to analog recorders or t8 the digi- tained with a double hose regulator using experimen- tal data acquisition system. The latter consisted tal conditions similar to those described earlier. of two fast digital voltmeters (HP-3437A) control- A comparison with Fig. 7 indicates that the peak led from a calculator (HP-9825A) with a 16K memory pressures and corresponding work of breathing asso- and equipped with a string - advanced programming ciated with the double hose regulator are larger and a plotter - general I/0 - extended I/0 ROMs. than with the single hose regulator. This is in This two channel system acquired data at a rate of agreement with the well known facts, further con- up to 100 data pairs per second when the voltmeters firmed by a comparison of data shown in Fig. 8 and were interfaced to the calculator using a single Fig. 10. A summary of the external work of breath- HP-IB interface. The use of two HP-IB interfaces ing data for the double hose regulator for depths up increased the acquisition rate to a maximum of 500 to 60 msw is contained in Fig. 10. The portion of data pairs per second with the total capacity of the work associated with inhalation is high at the 1750 .data pairs. The acquired data was stored on 20.. a magnetic tape for subsequent analyses and management. In the analysis, a complete inhala- tion/exhalation cycle was selected from the re- 15.. cording and the integration was performed using the digital data. A digital plotter (HP-9872A)was used for the graphical display of the resulting 10. diagrams. *-e.f4ma 5diagrams. Experimental Data d 625. . maw The combined experimental and data acqui- 5 . ---4----ms -- sition systems were used to obtain data on the ex- 40.1.0 msw "- s - ternal work of breathing of a variety of under- t - water breathing apparatus for pressures up to the 2 VOLUE maximum depth of 60 maw (197 fsw). The tested equipment was dry throughout the tests. -5 Fig. 7 shows examples of p-V dia- _ .. '* ,' / grams obtained with a single hose regulator at .. \ _= three different test conditions. Two diagrams were obtained respectively at atmospheric pressure / and 60 msw (197 fsw), with the breathing machine 15.. ; creating a sinusoidal demand of 409 respiratory -/.. minute volume (RMV), corresponding to the respira- tory demands10 of an average man performing heavy work at surface atmospheric pressures. This RMV Figure 7. Examples of p-V diagrams obtained was attained by selecting the tidal volume (VTv) with a single hose regulator of 2t and the respiratory rate of 20 breaths per minute (bpm). This combination of VTV and the 2 00 (fsw) 200 respiratory rate was specified as the test condi- L 20 tion in the Mil-Specs standards; it also corres- I: / ponds to the second of three test conditions re- 3ENTLEY-AEIMERS commended by Reimers4. The third p-V diagram, 0m 7 enclosing the largest area, was obtained at the / pressure equivalent to 40 msw and 62.59 RMV, cor- responding to the human respiratory demands10 at the onset of the severe work level. This RMV set- IX ting, attained at 25 bpm is the highest demand t w setting recommended by Reimers4.* m 1 The external work of breathing was . I calculated from these and additional p-V traces. Fig. 8 shows the effects of pressure and the simu- o / AJ lated level of RMV on the work per unit ventila- l ; O tion. The inhalation and exhalation components of the external work of breathing for 401 RMV are al- < z so shown. The external respiratory work per unit / ventilation at this ventilation level is well be- low the Bentley-Reimers standard; however, at W 62.51 RMV, it crosses the standard in the vicinity of 37 msw (121 fsw). This is in agreement with DEPTH (mew) expectations expressed by Reimers4: "most open circuit SCUBA regulators will not support a 62.5/ Figure 8. External work of breathing data min minute volume except at very shallow depths". for a single hose regulator It is interesting to identify and correlate the surface and varies little with the depth, while the Tests conducted on a Band Mask, operated work of exhalation is initially low but markedly with a 3/8" I.D., 300 ft long umbilical and sup- increases with the depth. The total work of plied with air at 150 psi over the bottom pressure breathing at 404 RMV is relatively high and it in- (psiob), resulted in p-V diagrams and the external tercepts the Bentley-Reimers limit in the vicinity work of breathing data shown respectively in Fig. of 50 msw (164 fsw). The 604 RMV curve (obtained 11 and Fig. 12. The work of breathing associated at 30 bpm and VTV of 22) shows that the larger with the use of the Band Mask is relatively low. respiratory demands are associated with an unac- However, at 62.54 RMV, the work deviates signifi- ceptably high level of respiratory work. cantly from the 602 RMV results at depths greater than about 25 msw (82 fsw), and becomes excessive 29 v6O ,48.7msw around 50 msw (164 fsw). 15.. 28..~~~~~~~~~~,.. 15 / 231.. �'/ . 40,t 0msw \ 8' 1, / ,'E 15.. g - 5 ... ...\ t � 0 .' Van ,' |~~~~~~~~~~~~~~~ ~~~~~~~~~~ VOLUM*'InE Cb,' i /401, Om..w \ -ua~) . - .. 40 l, 60 msw ~'-m. W_1s l"-la~~~~~~~~~~~~~... .......... - .... -23.. -Z.. ~ _ _-1-. .A... .t Figure 9. Examples of p-V diagrams obtained with a double hose regulator -28 Figure 11. Examples of p-V diagrams obtained with a Band Mask IL20 too &2 0 100 awW) 200 3. 34 .20 10 fw 0 E .17 Ate BENTLEY-REIMERS E B BENTLEY-REIMERS 0.17 w / w~ ~ ~ ~~~~~~~~~~~~~~~~~ ', XHALj4ML-4a RIV4 2~ 2 o 0 s N R * E 8 W s s as 8 ws so3 DEPTH <mow) DEPTH (mew. Figur 10.Extenal ork o brethin dat Figure 12. External work of breathing data for a double hose regulator 119 0 ~ ~ ~ ~ ~ ~ ' / ,, ,.� ~~~~~~~~~~~~~~Figure 11. Ext ramlesor of bratin datiague1Etrnalmsr ofbrethingedat for a double hose regula~wtorfo a Band Mask o ~'oo~~~~~~~~~~~~1 Results of tests conducted with a helmet, b. an automated method of collecting data capable of operating in the demand mode, are shown on the external work of breathing has in Fig. 13 and Fig. 14. The helmet was supplied been described; with air at 170 psiob, using the same umbilical which was used in the Band Mask tests. The regula- tor stage at the mouthpiece of the helmet was eq- "4OMew uipped with a venturi assist. Effects of this fea-. ture are readily discernible in the inspiration half of the p-V diagrams shown in Fig. 13. They .. . indicate that at depth, the system provides a posi- in.. tive pressure in the mouthpiece once an initial de- 8 l .' _ % / mand is created and a continuous gas flow condition exists. This effective assistance yields a nega- . ,2 ' 1, \ tive external work of inhalation at depths greater 5. II than about 9 msw (30 fsw) and reduces the total external respiratory work below the exhalation ex- ternal work of breathing (Fig. 14). The positive 2 T VOUBE (b pressure during the inhalation also reduces the in- ternal work of inhalation. This contribution is not identified in the data presented in Fig. 14. However, its effect would reduce the combined in- ternal and external (total) work of breathing. None of the tested equipment was subjec- ted to 22.5i RMV, the lowest respiratory demand le- vel recommended by Reimers4. However, it is expec- ted that at such a low respiratory demand level, all tested apparatus would readily satisfy the modified Bentley-Reimers standard. Figure 13. Examples of p-V diagrams obtained with a helmet operating in the An examination of figures showing the ex- air demand mode ternal work of breathing per unit ventilation for the four types of demand-air breathing apparatus tested, shows that the modified Bentley-Reimers 0 ioo (few) 200 criterion is a suitable standard within the range of pressures tested using 409 RMV (20 bmp and 29 V ). Interestingly, the older technology double 0.7 BENTLEY- REIMERS hose regulator failed to satisfy the modified Bentley-Reimers criterion at depths greater than -.. 50 msw (164 fsw). E When the greatest recommended demand4 of * ..OTA -. 62.5Z RMV was used, the Bentley-Reimers criterion was not satisfied at the various depth ranges. The 0.1 available physiological data does not clearly pro- vide the probabilities for occurrence of such de- w mands at depth. Therefore, further investigation of the statistical aspects of divers' physiological W demands and external work of breathing limits at U. high demand rates should be conducted. It is proposed that the modified Bentley- O Reimers criterion be adopted by researchers in this 3 0.0 * . : field as the external work of breathing standard to J evaluate demand-air breathing apparatus. Secondly, < it is suggested that the conditions and require- ments associated with the maximal ventilation of 00 divers are further investigated. Summary 0 0 0 0 0 0 0 to I t K) Q In conclusion, the following points are DEPTH (msw) submitted: Figure 14. External work of breathing data a. a suitable method of evaluating under- for a helmet operating in the water breathing apparatus is to measure air demand mode the external work of breathing; 120 c. the modified Bentley-Reimers criterion represents a suitable standard to evalu- ate all types of demand-air breathing apparatus in the 0-60 msw range, using an RMV of 40i; and, d. the upper limit RMV for demand-air breath- ing apparatus and its suitability to the modified Bentley-Reimers criterion re- quires further investigation. Acknowledgements Encouragement from LCdr B.A. Ridgewell, ex- perimental contributions of Mr. R.E. Clark, P. Eng. and Mr. R.S. MacLean, and programming efforts of Sgt D.J. Stewart, Mr. P.K. Tam and Miss S.M. Brennan are gratefully acknowledged. References 1. Silverman, L., et al. Arch Ind. Hyg. and Oc. Med. 3, 461 (1951) 2. Cooper, E.A., J. Appl. Physiol. 15, 1053 (1960) 3. R.A. Bentley, et al., Proceedings, Second Con- ference on Portable Life Support Systems, held at Ames Research Centre, Moffett Field, California, May 11-13, pp. 295-303, (1971), NASA SP-302 4. Reimers, S.D., Report 19-73, Navy Experi- mental Diving Unit, U.S. Navy (30 Jan 74) 5. Middleton, J.R., Report No. 9-77, Navy Experimental Diving Unit, U.S. Navy (May 77) 6. U.S. Military Specifications: Regulator, Air, DemandsDiver's; MIL-R-19558 (SHIPS), 14 Aug 56, superceded by MIL-R-19558A (SHIPS), 17 May 60. 7. U.S. Military Specification: Regulator, Air, DemandSingle Hose, Non-Magnetic, Diver's; MIL-R-24169 (SHIPS), 3 Dec 65, superceded by MIL-R-24159A (SHIPS), 22 Mar 67. 8. The Second Undersea Medical Society Workshop, Respiratory Limitations of Underwater Breath- ing Equipment, held at Harvard School of Public Health, 14-15 Mar 73. 9. Grodski, J.J., The Breathing Simulator Sys- tem at DCIEM, DCIEM Technical Report, DCIEM, Toronto, Canada, Jul 78. 10. Kao, F.F., An Experimental Study of the Path- ways Involved in Exercise Hypernoea Employing Cross Circulation Techniques. The Regulation of Human Respiration, (Eds. Cunningham, D.J.C. and Lloyd, B.B.), Blackwell, Oxford, (1963) 121 UNDERWATER REPAIR OF NUCLEAR REACTORS Mr. Leon V. Manry Dr. John D. Peters The Ocean Corporation Energy and Resources Department 7929 Glenmont System Development Corporation Houston, Texas 77036 7929 Westpark Drive McLean, Virginia 22102 Abstract A man in a dry reactor vessel is exposed to radi- ation from the full 360 deg. of the vessel wall. An innovative approach to maintenance and repair A worker under water receives only a small dose inside reactor vessels without dewatering has of radiation, coming primarily from the wall been accomplished by the utilization of underwater nearest to him. technicians. The presence of water provides shielding to the workers thereby reducing both The rate of radiation falls off dramatically as the radiation dosage and the dose rate and per- the diver backs away from the wall. Measurements mits much longer stay time in the repair area. taken before the underwatertechnician enters the This provides several distinct benefits: (1) the vessel can establish how far he must stay away number of people exposed to radiation during a from theworking area to receive the minimum prac- repair operation is reduced by an order of mag- tical dose of radiation for a given time period. nitude or more; (2) the overall duration of the (See Table 1). repair operation is usually shortened drastically; (3) quality standards are more easy to maintain; and (4) the individual radiation dosage is reduced RADIATION PROFILE SURVEY OF REACTOR VESSEL to unprecedented levels for repair operations. Present experience with underwater technicians DISTANCE indicates clearly that virtually all kinds of jobs DEPT FOM VESSEL NORTH WEST SOUTH EAST can be accomplished under water. This includes WALL SIDE SIDE SIDE SIDE determination of the problem, repair, and quality assurance check up which are all done with greater Reactor safety because of the shielding effect of the Flange 200 250 100 60 coolant. 1' 60 90 30 25 O' 2' 10 50 10 - Introduction 3' 5 10 10 - 1' 90 150 30 250 Repairs on nuclear reactors and their associated 5' 2 50 40 10 3' 30 20 5 equipment customarily have been done either with 20 5 remotely manipulated tools or by men working in 1' 100 200 150 400 radioactive areas for short periods of time. Both o10 2' 50 150 30 - procedures have disadvantages. 3' 30 40 10 - Exposure is less when remotely operated tools are 1' 4500 10000 6000 10000 used, but the methods are so specialized that 15' 2' 2500 000 2000 - used, 3' 1~~~~~~~~~~~~~~~~~~~~~~~20oo00 200oo0 existing tools have to be modified, or new ones fabricated each time. While these tools are 1' 6000 10000 6000 12000 highly maneuverable, repair work using them tends 20' 2' 4000 10000 3000 - to be less efficient than when a worker is at 3' 900 3500 1000 the spot where he can use a variety of tools, take measurements, make observations, and take Northwest Cnat1500 1500 1500 I - instructions from support personnel. nubber Cotc 1500 1500 1500 Snubber I' 500 - 500 Level 2' 100 - 100 Traditional practice has been to dewater the 3 10 - 30 reactor vessel when workers have to go inside for the repair operation, or for inspections. A l' 900 1000 1000 1000 25' 2' 200 500 200 - new practice is to use underwater techniques 3 ' 100 150 100 - without dewatering the vessel. This provides decisive advantages in safety, time, and cost. Survey Instrument: Most important is the reduction in radiation Rad Cun: nderwater probe. exposure because the water attenuates the radia- tion emanating from the vessel itself. Table 1 122 Data collected to date show radiation doses re- tor of 4.8. This coolant was not expected to ceived by underwater technicians doing inspection present any difficulty, but to avoid any possi- and maintenance inside reactor vessels have been bility of radioactive contamination or of inges- kept one or two orders of magnitude below what tion of the coolant, the divers were required to individuals would obtain in a dewatered vessel. have total water-tight integrity. Therefore, This is in spite of the fact that underwater tech- the suits and helmets they worehad to be totally nicians spend 10 to 15 times as much time inside dry. Toeliminateany possibility of backflow of the reactor vessel; in fact, the underwater tech- coolant into the helmet, the helmet exhaust was nicians pick up less radiation than surface sup- not allowed to go into the water but was diverted port workers. to the surface through a return air hose. This also alleviated any possibility of exhaust bubbles contaminating the air. Nuclear Underwater Services This was the crew's first nuclear diving operation, Complete underwater nuclear service can be pro- and they received two weeks of training for the vided in different ways. One example is the job. Three days on location were spent in plan- collaborationbetweenThe Ocean Corp. of Houston, ning, preparation, and gaining familiarity with Texas and System Development Corporation (SDC) the procedures for getting into and out of the with offices in McLean, Virginia. SDC provides vessel. Concurrently, procedures were developed the engineering design capability needed to re- for doing the work. Diving operations began on place inoperativemechanical systems inside reac- the fourth day and continued over the next several tor vessels. This includes the design and manufac- weeks without incident or difficulty. ture of tools that can be used efficiently while the diver remains at a safe distance from vessel A total of 18 diving days were spent on this job walls, and the running of engineering analyses and the underwater working time was 55 hours and on parts that are excessively worn or damaged. 4 minutes. Thirty hours and 30 minutes of this Ocean Corporation has extensive experience in diving time was used for the actual removal, re- nuclear reactor maintenance. Their divers have placement, torquing, and quality inspection of the a total of 3475 minutes at work inside reactor shims. Theremaining time was used for subsidiary vessels. operations, such as grinding, removal of debris, setting and removing dummy shims, vacuuming, The divers used as nuclear underwater technicians alignment, quality control, and inspection of other are college-trained, have extensive diving exper- parts of the vessel. ience, understand engineering terminology, and can manipulate engineering measuring devices and It had been estimated that the job (performed in the equipment. They are especially trained for the usual manner) would take three or four months and work inside reactor vessels. use as many as 200 to 300 boilermakers. Although any such estimate is open to speculation, there un- Nuclear underwater technicians inspect, repair, or doubtedly was dramatic saving in time, personnel, replace parts inside the reactor vessels. They and cost by use of underwater technicians. are used for reconfiguring spent fuel storage racks, repairing fuel transfer equipment, repairing leaks in fuel storage pool liners, and for almost any conceivable operation that has to be performed in Health Physics Aspects a radioactive environment where water can be used for shielding. A major concern was the radiation dosage to which the diving personnel might be exposed. Health physics personnel supervised all operations, and The First Job a number of detectors of different typeswereused on the divers, both in and out of the water. Technicians from Ocean Corp. conducted their first During the diving operations, the divers received dives in a commercial nuclear reactor during the less radiation dosage than the workers on the period of December 1973 through March 1974. The surface. No diver received a radiation dose even major work required was replacement of the shims approaching the three-month limit of 2500millirem on six snubbers near the bottom of the nuclear set by the contractor. (See Table 2). vessel, a region of high radioactivity. These underwater technicians have worked at sev- At a preliminary meeting between the prime con- eral nuclear power plants, the latest being the tractor and the owner, it was agreed that scien- Three Mile Island Plant near Harrisburg, Pennsyl- tifically trained divers, capable of operating vania. Here they entered the reactor vessel to with scientists and engineers and understanding replace a pintal pin pusher tool which was stuck. precision operations, were desired for this job. Attempts at repair using remote tools and other The primary coolant inside the vessel was water methods had been unsuccessful. The underwater which contained 1900 ppm of boron with a pH fac- technicians did the job in 2 hours and 54 minutes. 123 DOSE SUMMARY Diver's - Ocean Corporation DIVER TIME UNDERWATER DOSE - QTR ALLOWABLE REMAINING 1 227 minutes 560 2500 1940 2 263 862 " 1638 3 269 892 " 1608 4 347 784 " 1716 5 50 351 2149 6 27 " 909 " 1591 7 454 681 " 1819 8 * 0 97 " 2403 9 169 967 " 1533 10 359 823 " 1677 11 * 0 181 2319 12 359 909 " 1591 13 275 679 1821 14 288 741 1759 15 217 1211 1289 Total 3304 minutes � These two gentlemen spent no time underwater, but were on the surface at all times. NOTE: A-1 of the divers spent time on the surface in addition tO their diving time. Table 2 The diver who spent the longest time (1 hour and 54 minutes) in the vessel received a total full body count of approximately 75 milliremfor that entire period of time. The other diver, who had been down for a period of one hour, took a whole body dose of only 54 millirem. These values are so low it is hard to justify other methods for such repairs. The radiation hazard is almost completely eliminated if tools are devised that permit the diver to maintain a reasonable dis- tance from the vessel wall. Time is Vital Experience has shown that when a utility needs divers for emergency repairs or inspection, time is vital. This makes it important that divers are prepared to initiate diving operations as rapidly as safety permits. Divers must be accus- tomed to working inside the tanks that approximate the configuration, depth, and diameter of a reac- tor vessel. In addition, diving gear modified for nuclear operations must be kept ready for use. Underwater operations have proved to be a prac- tical way to reduce radiation hazards for main- tenance and repair personnel in nuclear power plants. In addition, repair and inspection times are shorter, which leads to lower costs. 124 UNDERWATER STUD WELDING Professor Koichi Masubuchi Dr. Hironori Ozaki Mr. Jun-ichi Chiba Department of Ocean Engineering Kawasaki Heavy Industries Mitsui Engineering and Massachusetts Institute of Technology Co., Ltd. Shipbuilding Co., Ltd. Cambridge, Massachusetts 02139 Kobe, Japan Tokyo, Japan can be installed in a water-tight enclosure and the welding gun can be remotely activated to complete welding. An initial work was conducted during the Abstract 1974/75 academic year. A U.S. patent on under- water stud welding gun(#3,989,920) was granted in Stud welding is a simple process which requires November, 1977.(3) no skill of the operator. Experiments were conducted to determine whether underwater stud welding could During the 1976/77 academic year further be made. Most experiments were made with mild- experiments were conducted on underwater stud steel studs 3/4 inch in diameter. Base plates used welding. Most experiments were made with mild- were mild steel and HY-80 steel, 1/2 and 1 inch steel studs 3/4 inch in diameter. Base plates (6.4 and 25.4 mm) thick. It was found that stud used were mild steel and HY-80 steel, which is a welding could be made successfully under water. In low-alloy, quenched-and-tempered steel with a the mechanical and metallurgical examinations, minimum specified yield strength of 80,000 psi including tensile tests, bend tests, examinations (56 Kg/m2). HY-80 steel is widely used for of macro-structures, and micro-hardness tests, submarine hulls and other structures. underwater welded specimens exhibited properties comparable to those of welds made in air. This study was conducted at the Massachusetts Institute of Technology under the direction of An effort is being made to develop an inte- Professor K. Masubuchi. Dr. H. Ozaki, who was a grated stud welding unit which can be used for research associate at M.I.T., is a research various applications in underwater fabrication and engineer, Welding Research Laboratory, Kobe Works, repair. Kawasaki Heavy Industries Co., Ltd., Kobe, Japan. Mr. J. Chiba, who was a graduate student at M.I.T., is an engineer, Mitsui Engineering and Shipbuilding Co., Ltd., Tokyo, Japan. 1. Introduction 2. General Description of Stud Welding During the last several years researchers at the Massachusetts Institute of Technology have Stud welding is an arc welding, in which the conducted an extensive study on underwater welding. stud serves as the electrode and a "stud welding The research effort was supported by the National gun" as the electrode holder. The heat for welding Sea Grant Office of the National Oceanic and the stud is developed by passage of current through Atmospheric Administration, Department of Commerce. an arc from the stud to the workpiece. The study covered various welding processes including shielded metal arc, gasmetal-arc, flux-shielded metal arc as well as stud welding which is discussed were are to dicharge stud welding, in this paper. Several reports, theses, and papers tud ding and in the stud weldingy have been published.(112) ~that differ mainly in the method of power supply. have been published. (1,2) For arc stud welding a d-c power source (a motor generator, a transformer-rectifier, or a storage Stud welding is an arc welding process in which bator, a user-recpower or apstor a metal stud (or a similar part) and a work piece battery) stud wer sua for eaectr are hatedand elte by a eletricarc rawndischarge stud welding is a low-voltage, electro- are heated and melted by an electric arc drawn sai trg ytmuulycagdb - between them. Then the two pieces are brought together rapidly under pressure to form a welded current. joint. Although stud welding is widely used in Of the fifty or so years that stud welding various construction work in air; there have been has been around, only in the last 15 has there no, orea very few if any, published articles been an upsurge in the variety of applications of reporting the use of underwater stud welding. this welding process. Capacitor discharge stud welding is the fast production process used in onuringdthecoursewofathe w eldingc research e rs notthin-gage metal fabrication. It is suitable for on underwater lding has.T.grearche nticed welding studs approximately 1/4 to 1/2 inch (6.4 that stud welding has a great pocentich to 12.5 mm) in diameter to a base plate up to water use. Stud welding is a simple process which aprxmtl0.inhOm)tic.Acsu approximately 0.4 inch(10 mm) thick. Arc stud requires no skill of the operator. The welding unit welding can be used for broader applications. Some 125 of the fields that utilize stud welding include The power house, or stud welding machine, is appliances, automotive, construction, industrial a transformer-rectifier type welding power supply equipment, and shipbuilding. which can be used for both stud welding and ordinary shielded metal arc using covered electrodes. 3. Welding Experiments ~An arc stud welding gun designed for air use In the study conducted during the 1974/75 was used in the experiment. Since the gun was not academic year, experiments were conducted using a made water tight, the gun was not immersed in water.* capacity discharge type stud welding equipment available at M.I.T. at that time. Underwater welds In conducting a welding experiment under water, were successfully made using small studs 1/8 inch the specimen was placed in a tank containing tap (3.2 mm) diameter. Results were reported in a water. Limited experiments using actual sea water thesis by Zanca. (4) revealed little difference in weld appearance from During the 1976/77 academic year, attempts ta aeuigtpwtr were made to weld larger studs up to 3/4 inch(19mm) Mild steel studs were used for both mild-steel in diameter. The KSM Fastening Division of Omark and HY-80 steel base plates 1/2 inch and I inch Industries was kind enough to loan, at no cost to (12.5 and 25.4 mm) thick. The stud diameter was MITa complete set of arc stud welding equip- 3 .64 inch (19. mm), adthe wedbseimter was2.5ic ment. This paper presents results obtained in 064ic 1. m.Tesu a .5ic experiments using this equipment. Further details (70 mm) long before welding, and 2.5 inch (64 mm) are given in a thesis by Chiba. (5) after welded. Welds were made under three conditions: Figure 1 shows schematically the experimental set-up. 1. Air welding with dry surface 2. Air welding with wet surface 3. Underwater welding (I inch(25.4 mu) deep). POWER HOUSE c'In an initial study, it was found that arc C ~~~~~~ ~initiation under water was practically impossible. Even in welding on a wet surface it was found to be * ~~more difficult to initiate the arc than welding on a dry surface. C ( 11 ~ ~~~~ An effort was made toward finding ways of 4 ~~~~~~initiating a stable arc under water. A stable arc 3 ~~~~~~could be effectively initiated when a small amount of steel wool or aluminum foil was placed between the tip of the stud and the workpiece. Steel wool was used throughout the experiment. Studs could be successfully welded in sea water.. It was found that successful welds could be made in water 4 inches (100 mm) deep. WELD CABLE ~~~~~~~~~The optimum welding time for an air weld and CONTROL CABLE a wet surface weld were experimentally found to be 0.6 to 0.9 second. For underwater welding (I inch STDWELDING GUN (25.4mm) deep) it was 0.7 to 0.85 second.. In both GRUND CABLE cases, the welding current was set to 1750 amperes, which was recommended by the manufacturer for air welding. when the welding time was too short (0.5 sec.), an extremely "cold" was the result. A welding time of 0.9 second gave an excessively "hot' weld. The welding time had asignificant influence on the weld quality. For each welding condition, the welding machine was set to produce optimum results. Test specimens were taken from welds WATER ~made under optimum conditions. '7 ~~~~4. Mechanical and Metallurgical Tests SUPPORT 4.1 Tensile Tests Tensile specimens were prepared by welding FIGURE I ARC STUD WELDING SET-UP Currently an effort is being made to develop a water tight stud welding gun which can be used completely immersed in water. 126 studs on both sides of the plate. The specimen was joint could be quickly examined. Studs that frac- set between the threaded chucks of a Baldwin tured in the weld metal or the heat-affected zone hydraulic testing machine, which provided a slow, were considered unsuitable for usage. This type of steady tensile force to the test specimen. The test is easily carried out at the job site by an test on each specimen was run until either stud operator, and can be used to test weld studs when fractured, giving the ultimate strength of the any change of welding conditions is made. weld, or the weld fractures, indicating an insuffi- cient fusion in the weld joint. Figure 2 shows an example of a specimen sub- jected to the bend test. The specimen fractured Table 1 gives the tensile test results of the along a thread of the stud. The bend test results welded stud specimens. Two specimens were prepared showed that welds were sufficiently strong. from both air welding and wet-surface welding. Three specimens were prepared for underwater welding (1 inch (25.4 mm) deep). All of these specimens were welded under optimum conditions. All of the specimens, except #3 (wet-surface weld) fractured at the threaded portion of the studs, giving the ultimate strength of the stud. Specimen #3 fractured in the stud along the heat-affected zone of the weld. The fractured surface revealed insufficient fusion. TABLE 1 TENSILE TEST RESULTS OF STUD WELDED SPECIMENS MAXIMUM TENSILE SPECIMEN CONDITION LOAD STRENGTH lbs per stud ksi (metric ton) (kg/mm2) FIGURE 2 APPEARANCE OF BEND TEST RESULT I AIR WELD 2.82 x 104 76.74 UNDERWATER 1 in. DEPTH---MILD STEEL BASE PLATE (12.79) (53.97) 2 AIR WELD 2.65 x 104 72.12 4.3. Macroscopic Examinations (12.02) (50..72) * 4 Macroscopic examinations were made of sections 3 WET SURFACE 2.58 x 10 70.08 prepared from some welded samples to studyhow welds WELD (11.68) (49.29) were made. As an example, Figure 3 shows the cross 4 WET SURFACE 2.72 x 104 74.02 section of an underwater stud weld made on a mild- (12.34) (52.07) steel plate. The width of the welded zone was about the same for both underwater and air welds. 5 UNDERWATER This was rather surprising, because a smaller weld I in. DEPTH 2.70 x 104 73.48 zone was expected for underwater welds due to the (12.25) (51.69) quenching effect. 6 UNDERWATER 24 1 in. DEPTH 2.70 x 104 73.48 (12.25) (51.69) 7 UNDERWATER I in. DEPTH 2.72 x 104 74.02 (12.34) (52.07) � -- HAZ portion of the stud side fractured t From the tensile test results, it was found that underwater stud welding with the quality of an air weld could be achieved by carefully applying a steel wool arc initiator. A 3/4 (19 mm) diameter stud can hold a load of approximately 10 to 12 tons. 4.2 Bend Tests A simple bend test was devised to provide FIGURE 3 CROSS SECTION OF WELDED STUD information about weld quality. A steel pipe four feet long was used for bending studs. Although no UNDERWATER I in. DEPTH---MILD STEEL BASE PLATE bending load was recorded, this simple bend test was adequate in that the soundness of the welded 127 Even under careful examinations, no voids, On the basis of the results obtained in this cracks or slag inclusions were found in weld research we believe that the arc stud welding can areas of all the specimens examined. Fused metals be used successfully for a variety of underwater apparently solidified with no visible defects. We applications. concluded that studs can be successfully welded to both mild-steel and HY-80 steel plates For example, it is technically possible to develop an integrated stud welding system which has 4.4 Microstructures and Microhardness the following features: 1. Remote control Examinations of microstructures were made of some specimens after the macroscopic examinations. 2. Automatic stud feeding Details are described in Chiba's thesis. 3. Multiple stud heads Examinations of microstructures were carried out by employing magnifications of x100 to x1000. 4. Waterproofed No cracks were found in the welded specimens made either in air or under water. The welded specimens were subjected to micro- hardness surveys using a Wilson Tukon Tester. A vertical traverse was made along the center of the weld. A 500 gram load was used to relate the Knoop hardness number (KILN) to the Vickers hardness number. The maximum hardness numbers at the heat- affected zone were as follows: Maximum Hardness, Vickers Mild steel Air 325 Underwater 340 HY-80 Steel Air 380 ... 3 Underwater 400 1 5. Conclusions and Possible Applications The purpose of this research project was to 4 develop a method of joining a large diameter stud to a metal object under water. Major conclusions 7 12 drawn in the research are as follows: (1) The arc stud welding system can be easily used under water provided that a proper arc initiator, such as steel wool or aluminum to 6 foil, is used. power source (2) No sophisticated system, such as a dry cham- ber for water evacuation from the welding site or gas shielding equipment, is needed in order to obtain a weld with good quality. (3) In the mechanical and metallurgical examina- 10 tions, including tensile tests, bend tests, examinations of macrostructures and micro- structures, and micro-hardness tests, ITEM underwater welded specimens exhibited prop- 1. Bushing 7. Spindle erties comparable to those of welds made in air. 2. Stud distributor 8. Stud 3. Feeding pipe 9. Ferrule (4) An HY-80 steel plate can be successfully stud welded using a mild-steel stud. Neither 4. Solenoid 10. Workpiece Neither cracks nor significant porosity, 5. Spring 11. Magnet which are responsible for the loss of mechanical properties, were found in the 6. Hook 12. Stud feeder weld area. 13. Water pump FIGURE 4 INTEGRATED UW STUD WELDING UNIT 128 Figure 4 shows the comceptual design of an References integrated stud welding unit. The unit has four stud heads and can attach four studs to the work- 1. Brown,A..J., Brown, R..T., Tsai, C. L., piece in a consecutive order within two seconds. and Masubuchi, K., Report on "Fundamental Each stud (3/4 inch (19 mm) diameter mild steel, Research on Underwater Welding", Report No. for example) can hold a load of 8 tons, thus a ITSG 74-29, M.I.T. Sea Grant Program, 32-ton capacity can be achieved with this four- September, 1974. head unit. 2. Tsai, C. L., Ozaki, H., Moore, A. P., Zanka, In welding operation, stud welding unit can .MraaSndasbciK, be held in position by electric magnets. The unit L. of New S.,rovd Tchiq K., is energized by pressing a button either bya diver " Dev elding", Neproved Ts for on site or by remote control switches. New studs Unde rat Prog", Aprt 1977. can be fed automatically through a stud feeder after each welding is completed. A water pump may 3 Masubuchi, K., and Kutsuna, M., "Underwater be used to create a dry welding environment in the Welding Gun", U.S. Patent No. 3,989,920, arc zone to ensure weld quality in some critical November 2, 1976. operations. 4. Zanca, L. M., "Underwater Stud Welding", M.S. One example of many expected applications of TeiMITSpebr 95 such a development is to lift a sunken ship in ocean salvage operations. Figure 5 shows a 5. Chiba, J., "Fundamental Study of Underwater schematic presentation of such operations. The Arc Stud Welding", M.S. Thesis, M.I.T., May, welding unit can even be operated by a special 1977. rescue submersible. It is hoped that such an integrated system can be developed in the near future Support ship Remote control J submersible with ~~~~ihwel ding unit Sunken object FIGURE 5 APPLICATION OF INTEGRATED WELDING SYSTEM FOR OCEAN SALVAGE OPERATION 129 DEVELOPMENT OF A DESIGN BASELINE FOR REMOTELY CONTROLLED UNDERWATER WORK SYSTEMS Robert L. Wernli NAVAL OCEAN SYSTEMS CENTER San Diego, California 92152 Abstract tion and repair operations. Basic components of the work package (Fig. 1) include two simple outer manipu- lator arms without elbow functions that act as The Work Systems Package (WSP), following two "grabbers" or restraining/holding arms to steady the years of at-sea testing and evaluation, has com- vehicle or hold small work pieces. pleted an extensive laboratory time-motion analysis. This analysis, designed to simulate remote opera- tion, provided a controlled evaluation of the sys- tem. Instrumentation was installed to provide time- motion and power consumption data on system compo- nents. The work tasks were divided into numerous subtasks and behavior motions which would be repre- TVCAMERA sentative of the WSP's diverse capabilities. A TL computer analysis was performed on the data to pro- vide a quantitative description of the WSP, its capabilities and the parameters effecting them. WINCH Thus, the effect of system modifications on mission time and Dower requirements can be easily simulated to assist the designer in system optimization. TOR The operational evaluation of the WSP has, for the first time, provided a quantitative data base on which to base future underwater work system HYDRAULIC RABBERS designs. POWER UNIT 1. Background The Work Systems Package (WSP), under the direction of the Naval Seark Systems Package WSP under, was Figure 1. WSP as it would appear mounted to the direction of the Naval Sea Systems Command, was manned submersible ALVIN designed, fabricated and is undergoing operational testing at the Naval Ocean Systems Center (NOSC) in San Diego. As part of the Deep Ocean Technology A centrally located seven-function manipulator arm (DOT) project, the WSP program was initiated in can select, interchange and operate a variety of February of fiscal year 1972, by NOSC working in hydraulically-powered, explosively-actuated or elec- conjunction with Battelle Institute, Civil Engineer- trically-actuated tools. Included in the tool stor- ing Laboratory and the David Taylor Naval Ship age box are tools to perform cable cutting, synthetic Research and Development Center/Annapolis. The line cutting, nut torquing, jacking, prying, wire Work Systems Package (WSP) is designed to provide bine cutting, nut torquing, jacking, prying, wire a versatile work capability when mounted as a unit ping an d stud dr iving. An electrically-driven on the Navy's Cable Controlled Underwater Recovery hydraulic pump unit supplies the power to most tools. Vehicle (CRVUWS) unmanned orcable controlled submersible Electric power is supplied to the system from a self- ~System (nanbeotrledsbmribecontained battery package. Control of all operations vehicles, and the ALVIN, SEACLIFF, and TURTLE manned and functions is provided through a multi-plexed vehicles. In addition, it can be positioned and telemetry circuit from the vehicle. Pressure toler- controlled by divers or operated independently from ant electronic and hydrau lic components operate at a surface support ship for operations at shallow full ambient pressure in oil filled, pressure com- depths without the need for a submersible. pensated enclosures. The system was designed to accomplish a com- Upon completion of assembly, checkout and pre- plete work task on the ocean floor without the liminary tests, the WSP was mated to the CURV III necessity of resurfacing for tool interchange. The WSP underwent Potential tasks include salvage, recovery, installa- 130 U.S. Government work notprotectedbyU.S.copyright. six weeks of operational testing at the Navy's San Simultaneous activation of both joints results in a Clemente Island test facility in fiscal year 1976. linear motion of the wrist pivot axis along a line Such tests as underwater docking with a submerged passing through the intersection of the shoulder test fixture, tool exchanges and operation, object rotate and pivot axes. The combination of the chain identification and recovery, and a simulated flight drive and linear extend features provides a linear recorder recovery were successfully completed. feed capability for the operation and interchange of tools and bits. The superior ooerability of the system, the short-time (2 to 2-1/2 minutes) required for remote Each tool of the WSP is stored in its assigned tool exchanges underwater and the successful per- location in an extendable tool box. Stiff nylon formance of a complicated recovery sequence requir- brushes provide the restraint required for retention ing exchange and operation of nine different tools and the compliance necessary for tool exchanges. and bits in 2-1/2 hours, achieved, and in many cases, Hose reels are eliminated from the system by running surpassed original design goals. hydraulic lines down the manipulator, through a hydraulic slip ring in the hand to two quick-discon- Following the successes at San Clemente Island, nect fittings designed to mate underwater with each the WSP was flown to NOSC's Hawaii Laboratory for tool. Mechanical guides along both the tools and interface and testing with the Remote Unmanned Work the manipulator hand ensure proper alignment, while System (RUWS). The WSP and RUWS were mated and interlocking notches on the tool secure it in the operated in the RUWS test pool in preparation for hand when gripped. Through the linear extend fea- support of the Large Object Salvage System (LOSS) ture of the manipulator, tools can be gripped easily operational demonstration at the Naval Coastal and extracted from the tool box for work operations. Systems Laboratory (NCSL) during the fiscal years 1976 and 1977. The WSP/RUWS was flown to Panama Tools which require bits, such as drills, taps, City, Florida, where it successfully completed sup- sockets and saw blades, are equipped with a special port of the LOSS operations. The systems easily quick-disconnect chuck assembly. Bits are obtained performed such tasks as midwater docking, cable by deflecting this chuck against the bit holder on cutting, stud driving, messenger line attachment, the tool box, moving the tool forward until the bit and air hose attachment using quick-disconnects. is fully inserted, and pulling the bit laterally out of its clip. This last action releases pressure on The WSP was then returned to San Diego where the tool chuck, thus locking the bit in the tool. it was prepared for an extensive laboratory evalua- Bits are replaced by following the reverse procedure. tion of the operating characteristics of the system. Figure 2 illustrates an acquired saw bit just after This laboratory evaluation is the subject of this it has been pulled laterally out of its clip. report. 2. Test Description Test Setup The WSP was set up in the laboratory at NDSC, San Diego. A frontal work plane test fixture, the primary work area for which the WSP was designed, was set up exactly as the test fixture used during the San Clemente Island tests with CURV III (refer- ence (1)). This would allow comparison of lab tests results to those acquired at sea. To simulate at-sea conditions, the operator was isolated from viewing the work area directly. View- ing was provided by the two low-light level TV cam- eras and the two monitors located on the WSP control console. The two cameras, one located approximately in the center of the system and the other on the upper right side above the manipulator, provided the dual perspective necessary for the performance of tool exchanges and work operations. Both cameras could be remotely moved in pan and tilt from the control console by means of position control joy- sticks. The manipulator is a seven degrees-of-freedom, rate-controlled, hydraulic manipulator, which is controlled through the activation of discrete action joystick (toggle) switches. Through utilization of a special chain drive within the manipulator, the plane of the wrist joint remains constant in refer- Figure 2. WSP manipulator extracting cutoff wheel ence to the horizontal, irrespective of the opera- from the tool box. Two TV cameras and winch are tion of either the shoulder pivot or elbow joints, visible in the upper left of the photo. 131 The primary analysis of the tests was directed it is possible to determine the impact on these towards a description of the manipulator, tools, vehicles and the dive time or number of dives required control functions and their interactions. to perform a given salvage mission. Task Requirements Test Subjects The objective of the laboratory testing was to The subject pool consisted of four operators obtain sufficient data to permit accurate prediction one with 100 hours experience in operating the WSP of work completion time and power consumption for during previous testing and major WSP sea trials, typical WSP salvage missions. Through applications and the others with only limited experience (less of this data base, subsystems which would receive than 10 hours each) in operating either laboratory the greatest benefit through a design modification or operational manipulator systems. Data obtained would be identified and more advanced work systems from the experienced subject were utilized to vali- could be reliably designed. date the laboratory test data as a representative performance of the WSP in underwater conditions. To achieve this goal , the tasks performed in Data collected on the four operators were evaluated the laboratory had to be representative of those and their performance with the WSP was averaged to performed at sea during the San Clemente Island and provide realistic estimates of time and power para- LOSS operations. This would allow verification of meters. The operators performed each of the 14 tasks the laboratory data in salvage mission projections. repeatedly for 10 trials to ensure the existence of The tasks would also be representative of those per- an adequate data base free of significant learning formed at Electric Boat (References 2, 3 and 4) effects. Data representative of learning were elimi- which would be used in future time-motion analysis. nated from the final data pool to represent perfor- mance by experienced operators as nearly as possible. The resultant Laboratory Test Plan consisted of 14 tasks that collectively utilized a representa- 3. Test Results tive sample of all too] types and their various pre- cision requirements. Operator Workload Allocation The 14 tasks selected for evaluation in the The first task of the data analysis was to laboratory included the following: identify and eliminate learning effects from the data base. During projection of slavage scenarios, elimi- 1. Sample retrieval 8. Brushing nation of the learning effects will provide more 2. Acquire tool 9. Hooking reliable results. 3. Replace tool 10. Valve turning 4. Acquire bit 11. Unbolting Figure 3 shows the learning curves resulting 5. Replace bit 12. Sawing through 10 trials of tool acquisition. It can be 6. Cut rope sample 13. Drilling seen that the naive subjects approached the profi- 7. Cut cable sample 14. Tapping ciency of the experienced operator during the test- ing period. Since these data points were taken Data Requirements throughout the entire test period (as opposed to sequential operation of primary tasks), the effect To provide accurate projections of salvage of total system learning can be indicated. missions, the data taken must be subdivided into subtasks small enough to be representative of all system operations. Therefore, each of the 14 tasks TOOL ACQUISITION was subdivided into four subtasks, or Behavior 15.00 EXRINDSUJC Elements, which consisted of various combinations *AvEXERIGEFNC~EDSUBJECTS of the basic actions (travel, tool use and align- 12.00- AEAEFRNIESBET ment) required to perform the task. All work and LU tool capabilities required on realistic 14SP salvage 9.00 - missions can be represented by sequentially perform- ' ing a series of these tasks and subtasks. Each of the categories of travel, alignment and tool use 60 were subdivided by the requirement for force control, ~30- ____________________ high or low precision and specific tool function.0 In addition, data were taken on the amount of 0.000 a 9 I time the manipulator operated or was idle and the TRIALS amount of time the TV pan and tilts were utilized. This would provide a more accurate time motion analysi of th operatr's wokload.Figure 3. Subject learning curves for tool Data were taken also on the power consumption acquisition. of the system during the performance of each of the subtasks. This was of primary importance since the WSP is designed for installation on battery operated Since erratic behavior is expected when learn- manned boats with limited dive times. When the 30 ing is occurring, a good measure of learning is the VDC and 60 VDC power consumption of the WSP is known, variability of variance associated with each task 132 mean. If no learning occurs, the operator's perfor- possibly without complicated or expensive system mance would be relatively stable. Figure 4 is a modifications. In comparison, the first area usually comparison of the total variance occurring in trials addressed for modifications is the manipulator, four through 10 with the overall variance occurring which could result in a more expensive, complicated, in trials one through 10. A 40 percent reduction less reliable system, although it may do the job in variance is found for all subjects and all tasks faster and more accurately. But, according to the for trials four through 10, thus substantiating the data, this decrease in time will apply primarily to assumption that task learning occurs in trials one manipulator operation, 33 percent of the time spent through three. The analysis, therefore, was focused and not necessarily across the time of the entire on mean task time for trials four through 10. mission by that same amount. Since several areas are interrelated, adjustments to the data, based upon subtask modifications, will allow a more accu- ~~~~~~80- ~~~ 60.0rate evaluation of potential system modifications ,. and concentration on areas where maximum benefit will _______________ be realized. < < 40 - I< 35.5 40% REDUCTION 20 - F 35~~~~~~~~~~~~~~~~~~~.50 0 40 -__ TRIALS 1-10 TRIALS 4-10 OZ <to c UJ Z5. < -J - uEO OL)-O Figure 4. Comparison of total variance for all OPERATION WITHOUTTOOLS(%) a m AJ tasks - all subjects. IAVERAGE OPERATION | TIME 50 33 17 / 100 The resultant task performance times averaged 50 / IDLETIME 2 0' I Z across subjects appear in Table 1. LOW SPEED PUMP DUTY TIME (3) / 33 17 TOTAL POWER CONSUMPTI ON 32 27 14 27 Z w - U OPERATION WITH TOOLS (%) F E ,O AVERAGE OPERATION *_ a, QoI~J Hj as M TIME 37 30 11 22 100 *2 z >Q 2a LOW SPEED PUMP - ,. W* IDLE TIME 37 (22) ' LOW SPEED PUMP 1. SAMPLE RETRIEVAL 1.58 .27 5.5 84 DUTY TIME 30 11 2. ACQUIRE TOOL 4.32 2.95 13.4 191 HIGH SPEED PUMP 3. REPLACE TOOL 2.74 1.83 8.3 112 DUTY TIME (4) // 22 4. ACQUIRE BIT 2.58 1.43 12.1 117 TOTAL POWER (10) 5. REPLACE BIT 2.93 1.90 9.0 127 CONSUMPTION 17 18 6 26 23 6. CUT ROPE 2.02 .30 6.6 125 (1) LIGHTING = 0.75 KW 7. CUT CABLE 1.30 .42 4.3 66 (2) LOW SPEED PUMP IDLE = 1.55 KW 8. ~ BRUSH 2.~39 . ~509 8.2 15 {~(3) LOW SPEED PUMP DUTY = 2.00 KW (4) HIGH SPEED PUMP DUTY = 3.97 KW (ON-OFF ONLY) 9. WINCH HOOK 5.45 2.17 17.2 239 9. WINCH HOOK 5.45 2.17 17.2 239 'IT IS ASSUMED THE MANIPULATOR IS NOT BEING MOVED DURING TOOL 10. VALVE OPEN 1.03 .26 3.3 51 ACTIVATION. 11. UNBOLT 1.66 .89 5.4 79 12. SAW ALUMINUM 7.61 3.12 26.0 542 13. DRILL HOLE 2.25 .63 7.6 118 14. TAP HOLE 2.50 1.15 8.2 119 14. TAPHOLE 2.50 1.15 8.2 119 Table 2. Operational time distribution (percent). * ASSUMES THE UTILIZATION OF ONLY ONE TV MONITOR * ASSUMES THE UTILIZATION OF 750 WATTS FOR UNDERWATER LIGHTING: Power Allocation ONE 250WATT LIGHT ON THE TV BRACKET AND ONE 500WATT SIDE LIGHT The power analysis on the system was broken into two sections. The first was the 30 VDC power required by the Command/Control electronics; the Table 1. Summary of WSP laboratory results for second was the 60 VDC power required by the lights total task time and power consumption. and hydraulics. The WSP runs on 60 VDC batteries, either its own or those of a manned submersible. Table 2 gives a breakdown of the operator's Therefore, the impact of the task or mission to be time based on the data taken. Two of the most performed on the battery supply of the vehicle is striking results indicated by Table 2 show the important. Also, in the case of a manned boat, the amount of time the operator spends contemplating 30 VDC power is provided by the life support system the situation or problem and the amount of time power of the vehicle. Therefore, adequate data in spent operating the TV pan and tilts. Efforts to this area are also required. Table 1 provides the reduce these times would have considerable impact, power data required for the performance of any of 133 the tasks. The power data have been derived to rep- resent operation on a manned vehicle with one TV camera and a lighting load of 750 watts. O1-ich NYLONLINE Table 2 shows the allocation of this power to ELECTRICAL CABLE the system. It was found that the overall power consumption was directly proportional to the task times and that the manipulator and tool operation used 75 percent of the power, while the lighting FLIGHT accounted for 25 percent of the power consumption. \ RECORDER By utilizing the data of Table 2, areas of high power consumption can be identified and reduced. - For example, the time spent by the hydraulic system- . idling accounts for 32 percent of the power consump- tion. The power loss in this area could be reduced by increasing the efficiency of the hydraulic sys- ALUMINUM BOLT tem during idle or by reducing the operator decision time. Data Validation SEQUENCE OF OPERATION 1. EXTRACT THE DRILL MOTOR AND A 1-INCH DRILL BIT Utilization of the laboratory data to predict 2. DRILL ACCESS HOLES IN THE ALUMINUM COVER TO ALLOW salvage scenarios for WSP in an ocean environment SPREADERINSERTION requires that proof be developed to show that the 3. EXTRACT THE SPREADER, INSERT INTO THE ALUMINUM SKIN AND requires that proof be developed to show that the OPENTHESKINTOALLOWINSERTIONO F THE JACK OPEN THE SKIN TO ALLOW INSERTION OF THE JACK laboratory and at-sea performances are equivalent. 4. REPOSITION THE VEHICLE TO ALLOW USE OF THE JACK Development of such proof is possible via compari- 5. EXTRACT THE JACK, INSERT, AND SPREAD APART THE ALUMINUM son of the data base collected in the laboratory RIBSALLOWINGREMOVALOFTHE"FLIGHTRECORDER" a. EXTRACT THE IMPACT WRENCH AND SOCKET AND REMOVE THE %-INCH and the data collected during the San Clemente BOLT FROM THE "FLIGHT RECORDER" Island and LOSS support operations and sea trials. 7. ATTACH A BUOY-LINE TO THE "FLIGHT RECORDER" AND REMOVE Since the experienced subject is common to all the ITFROMTHETESTFIXTUREUSINGTHEMANIPULATOR test series, direct comparison of his data provides ATT. ACHE CABLE-CUTTER AND CUT THE ELECTRICAL CABLE the required validation. 9. EXTRACT THE SYNTHETIC LINE-CUTTER AND CUT THE 1-INCH NYLON LINE ATTACHED TO THE "FLIGHT RECORDER" RELEASING IT TO Comparison of the data bases may occur at two FLOATTOTHESURFACE levels: (1) comparison of the performance of specific tasks and (2) comparison of the completion Figure 5. Simulated "flight recorder" recovery times for entire salvage scenarios scenario. Two specific tasks were replicated in the laboratory and the at-sea data bases. These are Predicted Salvage Missions tool acquisition and tool replacement. Data indi- cate the average performance times and standard deviations recorded are statistically identical at now be used to predict salvage mission time and a 99 percent level. power requirements. With this in mind, ten differ- rent work scenarios were evaluated to demonstrate The completion of three major salvage scenarios the WSP's salvage capability. A list of the scenar- during WSP sea trials allows comparison of the pre- ios and their respective time and power requirements ~~~~during WSP sea trialsare shown in Table 3. dicted completion time with that actually recorded at sea. Two scenarios were completed during the San Clemente Island sea trials: salvage of a simulated TOTAL 30 VDC so VDC flight recorder (Figure 5) and recovery of a test WORK STAN DARD VEHICLE VEHICLE box. One scenario was completed during the LOSS SCENARIO DESCRIPTION TIME DEVIATION POWER POWER (min) (min) (KwHr/Wat-Hr) (KwHr/Watt-Hr) support operations: connection of an air hose and . CUT OPENING, 216 17 0.70 12.3 messenger line. Completion time for these three THINMETAL 2. CUT OPENNG, 265 17 0.A9 16.3 scenarios was predicted by sequentially adding 265 17 0.89 16.3 appropriate task and subtask data recorded in the 3. REMOVE ELECTRONICS 110 11 0.35 5.3 laboratory. 4. PADEYE DRILL/TAP 29 6 0.09 1.2 FASTENERS S. PADEYE CONVENTIONAL 122 12 0.40 5.6 The data indicate that the actual scenario FASTENERS times are accurate to within the standard deviation 6. RIG CABLES 65 6 0.21 3.2 calculated. 7. OPERATE SEA VALVES 46 5 0.14 2.0 8. LOSS SUPPORT 10 3 0.03 0.4 Therefore, the laboratory task and subtask OPERATIONS 9. SCI FLIGHT RECORDER 79 10 0.25 3.6 data effectively represent the general performance I O. SCI TEST BOX 38 7 0.12 1.7 capabilities of the WSP system at sea. Further, RECOVERY the prediction of large salvage scenarios (in excess of 30 minutes) with laboratory data provides an Table 3. Estimated work time and power consumption accurate prediction of work completion times for for ten salvage scenarios. for ten salvage scenarios. actual undersea missions. 134 A detailed description of the tasks and their analy- Depending on the type of work to be performed, sis can be found in Reference 5. For simplicity, the benefit of the different controllers can be only one task is shown here. That task, the "simu- easily seen. Results indicate a potential savings lated flight recorder recovery," which was performed of 30 percent of salvage time through the use of the at SCI, is shown in Figure 5. Performance time pre- harness position controller. The savings in power dicted for this scenario was 89 minutes (maximum) may not be quite as high since, although the time compared to 91 minutes recorded at-sea, which sub- required to complete the task is less, the power con- stantiates this technique. sumed per unit time will be higher. One must also consider the working environment of the operator System Modification Studies when considering control devices. It is not an easy task to use a position controller when viewing Controller Evaluations through a manned vehicle viewport as opposed to a more spacious control room topside in a tethered One of the applications of the WSP data base is vehicle configuration. The designer is now given an the capability to evaluate modifications to critical indication of the proper design path in achieving an subsystems. One such area is the manipulator con- efficient work system. This path could be easily troller. verified through hands-on testing of the most proba- ble modification candidate without the need of test- Control options include pushbutton fixed rate ing all of the other options. (WSP current design), pushbutton variable rate, joy- stick fixed and variable rate, discrete position Automatic Manipulator Control control, and harness (master-slave) position control, Experimental data collected at Electric Boat The use of minicomputer programmed control of Division, (References 2, 3 and 4) were used in the manipulators and robots in industry has increased projections. rapidly. Such automation has increased accuracy and decreased work time in the performance of repeti- The pushbutton fixed rate controller is taken tive tasks. Therefore, it becomes a viable area for as the baseline data since its operation was similar application to the WSP. In an effort to determine to the controller used on the WSP. The resultant its benefit, a series of tests was performed using increase or decrease in subtask completion time for a microprocessor controller for the WSP manipulator. other control options is indicated relative to the The controller was programmed to go through a series baseline data. Adjustments to the baseline for the of tool and bit exchanges which it accomplished with ten salvage scenarios is presented in Table 4. excellent time reductions and near perfect alignment Projection of the effectiveness of each of the con- and target location. The time reductions, Table 5, trol options listed in Table 4 is made by adjusting are shown as compared to both experienced and the subtasks of individual tasks by the Electric inexperienced operators. Preliminary adjustments to Boat Data. the data base were performed using only the experi- enced operator times of Table 5 and resulted in an 18 percent time reduction. It can be concluded that the data acquired with the programmer tests will yield even more dramatic results. UA W be LU LUZ OPERATORS PRO- REDUCTION ~'~~~~~ ~z I Z_ Q_ O F- TASK INEXP. EXP. GRAMMER INEXP. EXP. M< ~ ACQUIRE TOOL 5.18 2.12 0.9 820/ 57% M< M 0g -o REPLACE TOOL 3.24 1.42 1.31 59% 8% IL in M* n L - Q 00 DC ACQUIRE BIT 3.02 1.23 1.00 33% 17% 1. CUT OPENING, 1 .93 1.34 1.18 1.53 .93 REPLACE BIT 3.56 1.30 0.74 79% 43% THIN METAL 2. CUT OPENING, 1 .94 1.48 1.28 2.43 .99 THICK METAL 3. REMOVE ELECTRONICS I 1 1.29 1.13 1.88 .77 Table 5. Comparison of WSP task times (minutes) 4. PADEYE DRILL/TAP 1 .97 1.24 1.07 1.69 .66 under direct operator control and computer control. FASTENERS 5. PADEYE CONVENTIONAL 1 .96 1.27 1.11 1.72 .70 FASTENERS Automatic Camera Control 6. RIG CABLES 1 .91 1.08 .97 1.37 .58 7. OPERATE SEA VALVES 1 1.02 1.26 1.11 1.98 .74 With the inclusion of position feedback sensors 8. LOSS SUPPORT 1 .90 -1.10 1.00 1.40 . on the manipulator, the possibility of automatic OPERATIONS 9. SCl FLIGHT RECORDER 1 .94 1.16 1.04 1.56 .61 camera positioning exists. Microprocessor control 10. SCI TEST BOX RECOVERY 1 .92 1.15 1.03 1.53 .58 could be easily applied to have the pans and tilts 1.0 .95 1.24 1.09 1.71 .70 move the cameras so that they will follow the manipulator hand exactly. This would enable the operator to concentrate on the task and not have to Table 4. Normalized scenario completion times stop work operations to move or adjust the cameras. utilizing various control types. Calculating the reduction in salvage scenario completion times is straightforward. If it is assumed that all present camera control times will 135 be eliminated by automatic positioning, completion times may be calculated by subtracting the predicted camera operation time from the predicted scenario completion time. The resultant data indicate that an average 8 percent reduction is achieved across all the salvage scenarios. 4. Conclusions It has been shown that the data base which has been generated for the WSP will be a valuable too] to those individuals working on advancing the state- of-the-art in undersea work systems. The system designer now has an indication of the areas where he should concentrate and the potential benefits achiev- able. The laboratory tests performed with the WSP have also yielded a great deal of practical data on performing remote work in the areas of viewing, sen- sors, tool and control requirements and much more. For a more detailed discussion of the actual work tasks and recommendations, the reader is referred to the laboratory test report, Reference 5. The test- ing also has shown areas where more research is required to answer the question of how man is to work remotely in the sea. Hopefully, these and other studies will continue the quest for that answer. References; 1, Wernli, R. W., WSP/CURV III Operational Testing at San Clemente Island - 1975 Work Systems Package (WSP) Program, Naval Undersea Center, NUC TN 1699, ~July 1976. 2. Pesch, A. ~J., G. R. Simoneau and R. G. Hill, Final Report- Operator Performance in Undersea Manipulator Systems, Electric Boat Division, General Dynamics, Groton, Connecticut, 1972. 3. Pesch, A. J., R. G. Hill and W. F. Klepser, Capabilities of Operators as Divers Versus Submersible Manipulator Controllers in Undersea Tasks, Electric Boat Division, General Dynamics, Groton, Connecticut, June 1970. 4. Pesch, A. ~J., W. F. Klepser, R. G. Hill and G. R. Simoneau, Operator Performance with Alternate Forms of Unilateral Position Control for Undersea Manipulators, Electric Boat Division, General Dynamics, Groton, Connecticut, September 1972. 5. Bertsche, W. R., K. P. Logan, A. J. Pesch, R. L. Wernli (Principal Investigator), Evaluation of the Design and Undersea Work Capability of the Work Systems Package, Final Report, Naval Ocean Systems Center, NOSC TR 214, 1 April 1978. 136 COASTAL DEMOGRAPHIC CHANGES IN THE UNITED STATES, 1950-1974 Niels West Department of Geography and Marine Affairs University of Rhode Island Kingston, RI 02881 Abstract Two objectives are sought in this paper. The first seeks to describe the demographic changes The present paper compares the demographic between urban and rural counties from a functional developments which have characterized the coastal approach. The rationale for this line of inquiry zone in conterminous United States. The population parallels recent research in the suburbanization changes in the coastal counties are analyzed for process, which attempts to account for the increases the two census periods 1950-1960 and 1960-1970, and that suburban counties have experienced and which population estimates made in 1974. The coastal are often directly related to the decline of many counties are grouped into three classes: coastal central cities. The second objective documents urban, non coastal urban, which are part of coastal regional changes within the coastal environment. SMSAs and coastal rural. Three objectives are This line of inquiry parallels non-coastal demograph- sought. The first describes the demographic ic research, which has analyzed the higher growth changes which have taken place nationally during rates in the so-called sunbelt states compared to the past 24 years. The second objective analyzes changes occurring in the rest of the United States.2 the demographic differences in the three types of The coastal zone, as presently defined, covers settlements discussed above. The final objective 209,167 square miles, or 7 percent of the 48 con- compares regional, urban/exurban differences in tiguous states. In 1950, this area had a population the demography within the coastal zone. The find- of slightly less than 30 per cent of the national ings of these are discussed in light of the urban/ population, a proportion which grew moderately dur- suburban migration and competing uses for the ing the 1950 and 1960 decades. nation's shoreline. The United States population grew by 18.4 per cent during the 1950-60 decade, a rate which declined significantly during the following census period 1. Introduction which continued into the 1970's. Assuming similar trends between 1974 and 1980 to those prevailing Considering the recent legislative emphasis on during the first four years of the decade, it is economic, environmental, and social problems occur- estimated that the national population will grow by ring in the coastal zone, it is surprising that the approximately 10 per cent during the present decade. vast majority of papers concerning this region have The demographic changes in the coastal zone dealt with either the physical impacts of proposed stabilized during the early 1970's. Slightly more coastal modifications or the legal and instituticnal than one-third the total national (conterminous) interpretations of the C.Z.M.A. of 1972. population now reside in this area. This trend This paper is descriptive in the sense that parallels the national growth, except the changes it traces the demographic changes which have taken which occurred in the coastal zone were more varied. place in the coastal zone during the past quarter During the 1950's, the coastal population grew by century. The coastal zone, as defined here, con- more than 29 per cent compared to 18 per cent for sists of 286 counties which border on the coast, the nation as a whole, This development continued estuaries, or tidewaters of conterminous United during the 1960's, although at a much reduced rate States. Thus, neither Alaska nor Hawaii have been and converged with the national demographic develop- included in this analysis. Nor have the 85 coun- ment in the early 1970's. In 1974, the population ties surrounding the Great Lakes been included, growth in the coastal zone had actually dropped to The 286 counties have been divided into three below the national rate (3.4 and 3.9 per cent re- groups. The first group consists of 116 counties spectively). The comparative density figure for located on the shore, and is part of Standard Metro- the coastal zone was about four times higher than politan Statistical Areas (SMSA). The second group the national average, although it should be recog- consists of 37 counties, which are part of coastal nized that great variances in county densities SMSAsbut do not actually border on the shoreline, exist within both the coastal zone and the nation. The third group consists of 133 counties which are both coastal and rural (i.e., not located within 2. Functional Analysis coastal SMSAs). 1This approach differs from that used by Spangler , who looked at coastal demograplic The importance of the coastal zone for urban changes but based his analysis on U.S. Coast Guard and marine transportation is reflected in the propor- regions. tion of the total coastal zone area classified rural, 137 SUMMARY STATISTICS FOR CONTERMINOUS AND COASTAL UNITED STATES 1950 - 1974 TOTAL PERCENT TOTAL PER CENT PER CENT YEAR CONTERMINOUS CHNE DENSITY COASTAL OF TOTAL CAG EST U.S. POPULATION CHNEU.S. POPULATION U.S. POPULATION CAC EST 1950 150.697.361 50.6 44.434.425 29.5 212.4 1960 178.464.236 18.4 59.9 57.065.131 31.9 28.4 272.8 1970 202.143.083 13.2 67.9 68.089.035 33.6 19.3 325.5 1974 210.196.000 3.9* 70.7 70.433.000 33.5 3.4 336.7 *Projected population growth rate estimated at 9.9 per cent assuming similar trends between 1974/80 as prevailed between 1970/74. **Projected population growth rate estimated at 8.6 per cent assuming similar trends between 1974/80 as prevailed between 1970/74. Table I coastal urban, and inland urban. Nearly half the cities is located close to the waterfront, which coastal zone of 209,228 square miles (48 per cent) reflects the important functional ties between the (Table II) is classified non-urban, a significant port facility and the rest of the city. With in- portion of which is used for private (primary and creasing emphasis on industrial and business devel- secondary) residences. Kimmelnan et.al, Inasp- opments in the suburbs, it is clear that counties arate study, concluded that 32 per cent of the located at some distance from the CBD and the water shoreline is residential, while another 59 per would benefit relatively more from immigration than cent is undeveloped. The balancS of 9 per cent the central city counties and counties with large is in public recreational usage. urban Dort facilities. The urban planning and ur- The next largest group consists of the coastal ban geographical literature is replete with examples urban counties which comprise approximately 42 documenting population changes between suburban per cent of the total area. In reflecting upon counties and inner cities, although empirical stud- these figures, it should of course be noted that ies emphasising coastal demographic changes are very not all areas classified as urban coastal necess- few. arily are urban from a functional point of view. There is at least one additional notion which Much of the New York waterfront, from a functional may contribute to our understanding of the compara- point of view, is coastal rural because of the tively lower growth rates which coastal urban coun- extensive areas included as part of the Gateway ties have experienced. This relates to the dramatic National Park, yet all five Boroughs have been changes which most North American waterfronts 'have classified coastal urban. The inland urban coun- undergone. Because of containerization and the ties, which are part of coastal SMSAs, made up the vastly increased demand for space required for mar- balance comprising about 9 per cent of the total shalling yards, much of the traditional waterfront coastal zone. is now vacant and awaiting urban renewal develop- Of the three functional groups, the inland mental4 In cities with former active urban port urban counties experienced the most rapid growth facilities, only a few of the previously busy piers during the 1950's and 1960's with declining rates have been retained and in some instances refurbished in the early 1970's. Assuming similar trends dur- for use by a growing cruise industry. The vast maj- ing the latter part of the present decade, overall ority of urban piers are -now often abandoned. Ware- growth in the inland urban counties is likely to housing and manufacturing which used to occupy much fall from nearly 36 per cent during the 1960's to of the inner city waterfront have most often moved little more than 11 per cent for the present decade. with the shipping lines and freight forwarders to In light of the attractiveness of the coastal the larger fully integrated container facilities, zone in terms of private residences and businesses, most of which are located peripherally to the CBD. it is interesting that the category classified as It is suggested that these factors have contributed urban coastal grew at a rate significantly lower significantly to the somewhat lower growth rates than either inland urban or coastal rural. Two which coastal urban counties have experienced. Be- possible explanations may account for these develop- cause of rapid changes along the urban waterfronts, ments and are briefly outlined below. The first it is expected that much of the abandoned real es- is related to the suburbanization process and con- tate will become the focus for urban renewal efforts. current stabilization or even stagnation of the In fact, some very interesting urban waterfront central business districts (CBD) of many older renewals are currently underway or have been comple- cities. The CBD of most North American coastal ted in Boston, Baltimore, Seattle, and New Orleans, 138 COMPARATIVE DEMOGRAPHIC ANALYSIS OF COASTAL URBAN, INLAND URBAN, AND COASTAL RURAL COUNTIES 1950 - 1974 1950 1960 1970 1974 Total Urb. Coastal Pop. 37.446.298 47.812.621 56.539.285 58.068.900 Per cent of 84.3 83.8 83.0 82.4 Total Coastal Population 42.5 Population Land Per Cent Change 27.6 18.3 2.7* Density 420.6 537.1 635.2 652.3 Total Urb. Inland Pop. 3.803.404 5.155.803 6.553.340 6.863.800 Per Cent of Population 8.6 9.0 9.6 9.8 Total Coastal Land 9.2 Population Per Cent Change 35.6 27.1 4.7 Density 198.3 268.8 341.6 357.8 Total Rural Coastal Pop. 3.184.723 4.096.707 4.996.410 5.500.300 Per Cent of Total Coastal Population 7.1 7.2 7.3 7.8 Population Land 48.3 Per Cent Change 28.6 21.9 10.1** Density 30.5 40.6 49.5 54.4 *Projected 1970/80 growth rate assuming similar trends for balance of decade = 6.8 er cent * Projected 1970/80 growth rate assuming similar trends for balance of dacade = 11.8 per cent Projected 1970/80 growth rate assuming similar trends for balance of decade = 25.2 per cent Table II In some respects, the coastal rural counties tal amenities available in rural or exurban counties provided the real surprise. As a group, these located peripherally to the larger metropolitan counties grew slightly faster than the coastal ur- areas.5 ban counties during the 1950's and increased their In this context, it should be noted that the rates during the 1960's. Of even greater interest, population density in the coastal zone is signifi- however, has been the ability of these counties to cantly below those of both the coastal urban and continue to grow during the early 1970's at rates inland urban counties as well as below the mean two to three times those of the two urban classes. density values for the county as a whole. (Tables 1 In fact, if the rates prevailing between 1970 and and 2.) 1974 continue during the rest of the decade, these Several generalized conclusions can be drawn counties will grow twice as fast as the inland from the analysis of the three functional groupings urban counties and nearly four times the rate of identified in this part of the analysis. Not sur- the coastal urban counties. While it is still too prisingly, the rate of growth in the coastal zone early to formulate and test hypotheses related to followed the overall slowdown which has been ob- this phenomenon, Barry, a few years ago, suggested served for the rest of the country, i.e. a much more that "inter-urban peripheral" growth may be a re- rapid dempgraphic growth during the 1950's which sult of the movement of businesses from the core of continued into the 1960's and has tended to converge cities to exurban areas located in relatively with the national rates in the 1970's. close proximity to the urban centers. The advan- The second conclusion concerns the temooral tage of this strategy is to be located close enough changes which have occurred within each of the groups. to the services (airports, stock exchange, etc.) Surprisingly, the non-coastal counties located within available in large metropolitan areas and yet far coastal SWSAs grew significantly faster in the early enough away to obtain real and relative environmen- part of the study period, but also showed the fastest 139 relative slowdown in growth. The urban coastal Since all classification is intended to minimize counties also grew, although not as rapidly as the differences within each class (region) while maxi- non-coastal urban counties. Several notions were mizing differences between classes, a chi square test advanced relating urban overall changes to the was performed on the six regions and the three specific case of coastal urban counties which should types of counties tabulated (coastal urban, coastal be looked at in greater detail than has been pos- rural and inland urban). The chi square is a non- sible here. parametric statistic, testing for differences bet- In some ways, the most surprising group was w een two or more variables. The observations are cast the rural coastal counties which exhibited growth in a contingency table or matrix where one axis rates similar to the urban coastal group, except represents the value of the ohservation(county)on that rural coastal counties have been projected one variable, while the other axis represents the to grow in the short term (to the end of the 1970 value of the county on the remaining variable. The decade) by 25 per cent. higher the chi square statistic, the lower the probability that the differences between the two variables would have occurred by chance. In the 3. Regional Coastal Analysis social sciences, levels of confidence of .05 or .01 are the usual accepted probabilities. The previous analysis compared coastal pop- The 286 counties were classified by region (as ulation by county with comparable sets on the nat- predetermined) and the coastal urban, inland urban, ional level. The ensuing analysis divides the and coastal rural trichtomy. The test statistics coastal zone into six sub-regions. Regional and proved to be highly significant (alpha = .0001), functional differences will be noted. Finally, which suggests that the six regions as defined are some of the key demographic variables will be tes- valid. ted for purposes of detecting statistically signif- The six regions vary in size from approximately icant differences among the six regions. 60,000 square miles (Southwest) to less than 16,000 The 286 counties were divided into six regions square miles (New England). Not surprisingly, the defined arbitrarily as New England, Mid-Atlantic, population density is almost perfectly inversely Southeast, Gulf, Southwest and Northwest. Each correlated with the size. region was defined on the basis of state boundaries. The temporal changes in the population showed The regions and states included appear in Table 3. interesting shifts, which again reflects at least some of the changes in the settlement patterns which U.S. COASTAL DEMOGRAPHIC REGIONS have characterized the nation during the third quarter of the 20th century. The gross population changes suggest a regroup- ing of the six regions for purposes of analyzing Regions Nos. of States regional population shifts between 1950 and 1974. Counties The smallest region, comprising the Pacific North- west, neither increased or decreased its share of the total coastal population. (Table IV). New Engl. 28 Maine, New Hampshire, Mas- sachusetts, Rhode Island and REGIONAL POPULATION SHIFTS Connecticut190-97 New York, New Jersey, Mary- Mid-Atl. 54 land, Delaware, Pennsylvania and Washinton, D.C.REGIONS 1950 1960 1970 1974 * Virginia, North Carolina, Southeast 109 South Carolina, Georgia, New England 13.7 12.1 11.4 11.3 Florida Gulf 48 Alabama, Louisiana, Texas Mid-Atlantic 45.1 41.3 39.0 37.7 Southwest 19 California Southeast 9.7 12.3 13,9 15.1 Northwest 28 Oregon, Washington Gulf 7.9 8.4 8.5 8.8 number of c~~~~~~~~~ounthwest 1ncude 20n 21is 21gio *The n u b r o Sounthwest incude in. this regio varies from 107 to 109 as Virginia Beach was not Northwest 5.8 5.5 5.7 5.6 created until the mid-fifties and Chesapeake until 1963. Furthermore, Nansenond was deleted as a TbeI county following the 1970 Census. All changes TbeI occurred in Virginia. The remaining five sub-regions can be assembled Table III into two, those which reduced their share and those which increased their share of the coastal zone Regionalizing a given area represents an population. Both the New England and Mid-Atlantic attempt to spatially classify the coastal zone. regions reduced their shares considerably during 140 the past 24 years. For New England, the reduction urbanin both 1950 and 1974. These figures suggest was less than half of that which the Mid-Atlantic relatively few changes in the proportion of coastal experienced. The latter region maintained a 45 urban population to the total regional population. per cent share of the total coastal population in Both the Gulf and Northwest regions remained re- 1950 which dropped to less than 38 per cent by 1974. markably stable as the urban coastal population Of the three remaining regions, the Southeast only varied between 68 and 78 percent. increased its share by more than 5 per cent -- from The rural coastal population in the sunbelt slightly less than 10 percent in 1950 to more than states declined or remained stable, whereas the 15 per cent by 1974. The Southwest followed with comparable figures for the New England and Mid- a 3.7 per cent advance from 17.8 per cent to 21.5 Atlantic regions experienced slight increases per cent. The net increase for the Gulf region's between 1950 and 1974. The decline in coastal share of the coastal population was minor, slightly rural populations for the three southern regions is less than one per cent, and could as easily have been related to the overall growth which these regions discussed together with the Northwest. have experienced, most of which appears to have It is clear that the migration to the sunbelt occurred in urban counties. Of potentially greater states, which has been in evidence for some time, interest, however, are the increases in the rural is reflected in the coastal migration pattern as coastal population which the Mid-Atlantic region well. The two regions which lost in their share has experienced and which may well be related to of the coastal population (New England and Mid- the population increase of rural counties at the Atlantic) also are the most densely populated sec- expense of population losses in urban counties. If tions of the coast. Conversely, the three southern the equalization process between urban and rural regions which gained were characterized by sig- counties can be further verified, it would appear nificantly lower densities. The scale of the pre- to lend further support to Berry's notion of "inter- sent analysis is too coarse to determine the max- urban peripheral growth" as applied to the coastal imum number of people a given area can support, zone, which was discussed briefly above. It should, but it is possible that the declining shares of of course, be borne in mind that population density the New England and Mid-Atlantic regions may be is only one --albeit an important -- consideration affected by the relative scarcity or cost of land influencing regional growth. Other factors may in close proximity to the shore. This would be of even greater importance which should be in- appear to be particularly relevant in the case of vestigated, if further attempts are to be made to the Mid-Atlantic region. Newling, in working with identify regional growth. population projections for New Jersey's municipal- The chi square test was used to test if the ities, suggested that the "filling in" process of population densities for the four time periods were a given region would follow the logistic curve. statistically significant. All counties were cast The shape of the curve would be contingent upon in four 2 x 6 contingency tables where the density the size of the initial population, density, and variable for each of the four time periods was nom- the area available (zoned) for residential growth.6 inalized. Counties with mean densities below the Assuming the validity of the Newling model and overall coastal zone mean were separated from the applying it not in the context of forecasting, but counties above the coastal zone mean. Four sets merely to account for past growth, it is suggested of contingency tables were cast, one for 1950, that the Mid-Atlantic Region may be approaching 1960, 1970 and 1974 and all proved to be highly the upper asymptote of the logistic curve. In statistically significant. The chi square verifies this context future increases are likely to be a the qualitative conclusions which have been dis- result of a "filling in" process of existing empty cussed above. areas. Additional growth may occur as a result of changes in local zoning ordinances from a relatively low density of one, half and quarter acre zoning, 4. Summary (R-40, R-20, or R-10) to significantly higher densities, possibly available under PUD or other This paper has addressed the demographic multi-family zoning. changes occurring in the coastal zones of the 48 In order to test this notion, it would be contiguous states. Significant differences exist necessary to analyze municipal population changes between the growth and development of coastal zone within each of the 286 counties included in the ana- and the nation. It was found that the population lysis. Because of the magnitude of the data re- in the coastal counties grew at significantly quired, it was not possible to provide answers to higher rates during the 1950's and 1960's, Three these important questions in the present paper. different types of counties were recongized, coast- A regional demographic breakdown of the three al urban, inland urban and coastal rural. Again classes of counties was made for the four periods demographic differences were observed. Inland ur- 1950-1974. (Table V). While significant variations ban counties were found to grow at rates faster exist in the urban coastal/urban inland/rural divi- than both coastal urban and coastal rural counties. sions, further inspection of the table suggests Unable to test statistically, a few hypotheses were that the proportion of total population classified advanced which relate coastal demographic develop- remained remarkably stable throughout the 24 years. ments to the urban/suburban changes characterizing Both the Northwest and Southeast had significantly much of the eastern older cities. To test regional fewer people classified as urban coastal in 1950 differences, the coastal zone was divided into six compared to any of the four remaining regions. On regions. Temporal and functional differences were the opposite scale, the vast majority of the people noted among the six subcoastal regions. On the in the Mid - Atlantic and Southwest were classified basis of temporal and functional demographic 141 REGIONAL, FUNCTIONAL AND TEMPERAL DEMOGRAPHIC 2. Borchert, D.J. and J.D. Fitzsimmons, "Pop- PER CENT CHANGES 1950 - 1974 ulation Redistribution in the United States, 1950-1975," paper presented at the annual meeting of the Assoc. of Amer. Geographers, 1950 1960 1970 1974 New Orleans, LA., April 10, 1978. o Coastal Urban 95.7 95.7 95.6 95.1 3. Kimmelman, B. (et.al.), Studies in Environ- Inland Urban ment Vol. 5 - Outdoor Recreation and the Rural 4.3 4.3 4.4 4.9 Environment, Office of Research and Develop- ment, E.P.A., Washington, D.C., 1974. Coastal Urban 84.5 81.6 79.3 78.4 �< Inland Urban 13.6 16.2 18.1 18.6 4. Moss, M.L. "The Challenge of Urban Waterfront Rural 1.9 2.2 2.6 3.0 Revitilization," paper presented at the �z ~~~~~~~~~~~~~Princeton University Conf. on the Urban Cosa Ubn 5. 655 6. 689Coast and Energy Alternatives, May 17-18, 1978. m Coastal Urban 56.7 65.5 67.7 68.9 ) Inland Urban 14.0 10.3 9.4 8.8 5. Barry, BJ.L. "The Geography of the United *Rural 29.3 24.2 22.9 23.3 States in the Year 2000," Ekistics, (1970), 29, 341-345. See also his Growth Centers in Coastal Urban 77.6 78.4 78.1 77.6 the American Urban Systems, Vol. 1, Ballinger Inland Urban 8.9 8.7 9.1 9.8 Publishing Co., Cambridge, MA., 1973. o Rural 13.5 13.0 12.8 12.6 6. Newling, B. "Population Projections for New Coastal Urban 97.9 97.8 98.2 98.0 Jersey to 2000," City College Publications, Coastal Urban 97.9 97.8 98.2 98.0 New York, 1968. ~ Inland Urban Rural 2.1 2.2 1.8 2.0 fn 4 Coastal Urban 70.0 70.9 70.8 68.5 Inland Urban 5.8 6.6 8.5 9.7 Rural 24.2 22.4 18.7 21.8 Table V developments occurring, three coastal geographic trends were recognized. The Northeast, comprising New England and the Mid-Atlantic regions, was losing its proportionate share of the total coastal population. Urban populations were not growing at nearly the same rate as the coastal rural areas, which appears to increase at a f aster rate than in any of the remaining four subregions. The sunbelt regions proportionately grew the fastest reflecting the overall regional migration pattern which has been in evidence for some years. The coastal rural areas tend to decline, probably in response to the rapid urban demographic develop- ment which has occurred in this region. The third development consists of the Northwest, which is the most stable in demographic terms. While population growth has been experienced in this region, it has maintained its overall relative share of the coastal zone population as well as its share of functional land-uses. References 1. Spangler, M.B. "Projections of Socio-economic Trends in the Coastal Zone," Marine Techn. Soc. Jour.(Jily/Aug., 1972), 6,4,21-24. 142 GREAT LAKES SHORELINE EROSION - WESTERN LAKE MICHIGAN J. M. Pezzetta J. Robert Moore University of Petroleum and Minerals Institute of Marine Science Dhahran, Saudi Arabia University of Alaska Fairbanks, Alaska 99701 (3) nearshore hydrodynamics, particularly waves and alongshore currents; (4) meteorological effects, especially wind fields, Abstract storm tracks, and their frequency, intensity and duration; Three coastal sites of varying topographic and (5) the effects of the freeze-thaw process and geologic characteristics were selected along the shore ice on the shoreland materials; Wisconsin shoreline of Lake Michigan for monitoring (6) the influence of artificial protective struc- the processes and rates of shoreland erosion. Perio- tures on nearshore processes and shoreline configu- dic (seasonal) plane surveys revealed that the mean ration. recession rates varied from 0.05 ft (0.015 m)/TDM The development of emergency as well as long (thirty-day month) to 2.7 ft (0.082 m)/TDM, the lat- term coastal zone management programs should be ter having occurred during the fall/winter of 1975/ structured, therefore, in a manner which seeks to 76 along a high, steep, clayey bluff near Port Wash- address these specific factors. Shoreland zoning ington. A sinuosity factor S, invoked as a quantita- and setback allowances must be developed within the tive measure of the planimetric geometry of the context of well-defined regulations that consider bluff crest, provided an indirect evaluation of the not only the rates of bluff recession but also the compositional uniformity or variability of the mechanics and variability of erosive processes in bluff deposits. While high lake levels have accele- the coastal environment. In addition, the effective- rated the processes which induce failure and slum- ness of different types of erosion control measures ping, particularly along those shoreline segments must be evaluated in terms of both engineering/eco- having narrow beaches, the entire mechanism of nomic considerations and land use management prac- bluff undermining and collapse depends largely on tices. Historical evidence has shown that a piece- the direct and indirect effects of precipitation meal approach to shoreline protection, more often including rainfall, sheet 6 channelled runoff, than not, serves only to enhance rather than ameli- water table fluctuations, frost action & shore ice. orate the problem. Innovative plans of an integra- tive nature that maximize the protection at a rea- sonable cost must be explored and developed (cf. 1. Introduction Brater, et al, 1975). A necessary part of any program which seeks A pervasive water resource problem in the to optimize the effectiveness of any plan of shore- Great Lakes basin which has received considerable line protection must rely on a substantial and ade- attention in recent years is the extensive erosion quate data base derived from a critical examination of the coastlines. Shoreland segments that are par- and evaluation of the physical characteristics and ticularly vulnerable to the forces of erosion tend dynamic processes of the coastal environment, parti- to suffer severe and rapid slumping during periods cularly along shoreline reaches that are rapidly of high lake elevations. By eliminating the pro- succumbing to erosive activity. tection afforded by a wide beach face, high lake Available information on the erosional charac- levels enable the erosive activity to be focussed teristics of the western coast of Lake Michigan directly at the toes of the bluffs, terraces, or generally has been of a qualitative nature (e.g. dunes. As a consequence, wave-induced undermining Striegel, 1968; U.S. Army Corps of Engineers, 1971). and bank failure are substantially accelerated dur- Striegel's work is especially detailed in its des- ing such times. However, field observations have cription of the lakeshore and includes one of the demonstrated quite clearly that bluff erosion and few quantitative assessments of bluff recession for recession are not entirely a function of lake level Lake Michigan. For the period 1874 to 1946, he cal- fluctuations and that other, sometimes subtle, en- culated the recession rate in Racine County to be vironmental phenomena play a very significant, of- approximately 1.2 ft (36.6 cm)/year or 0.099 ft ten dominant role in this complex process. (3.02 cm) per thirty-day month (TDM). Several interrelated factors, both natural and The principal purpose of this study is to es- man-made, which merit consideration in evaluating tablish rates of shoreline recession along select the nature and degree of coastal erosion are: segments of the western coastline of Lake Michigan (1) the natural geometry of the shoreline and its and to evaluate the sediment distribution patterns orientation to prevailing wind-induced waves; in the immediate offshore waters adjacent to the (2) the morphology, geology and competency of the onshore monitoring sites. shorelands and inshore zone; This initial report addresses primarily the 143 bank recession problem and to a lesser extent the warranting detailed monitoring, both onshore and relationship between nearshore hydrodynamics and offshore. Of those 15 stations, the three discus- sediment dispersal in the littoral zone. sed in this report were chosen to provide specific examples of erosional characteristics of shoreline segments showing varying geological properties and 2. Description of the Study Area susceptibilities to bank undermining as well as to demonstrate actual or potential hazards to ripar- Excluding the shores of Green Bay, the Wiscon- ian owners. sin section of the Lake Michigan coastline is ap- The general locations of these sites are shown proximately 288 statute miles (463 km) in length in Fig. 2. These three principal stations are des- and extends from Rock Island just off the tip of cribed as follows: M-2 in Sevastopol Town Park, Door County peninsula to the Illinois state line northeast of Sturgeon Bay; M-7 in Sandy Bay, south- (Fig. 1). ern Kewaunee County; and M-ll in Lake Park, Port The entire coastal environment falls within a Washington. physiographic region of the state known as the At each site a steel rod or 'pin', 1/2" OD by Eastern Ridges and Lowlands (Fig. 2), the topogra- 50" long (1.27 cm x 127 cm), was driven into the phic features of which were shaped by Pleistocene ground until some seven or eight inches (18 to 20 glaciers. The relief of the coastal bluffs ranges cm) were left protruding above the surface. These from a few feet above the lake surface near the pins served as bench marks from which the depar- southern limits of the survey to a maximum of 164 tures to the bluff edge were measured. The actual ft (50 m) above the lake (744 ft or 227 m above spot on the ground where the pin was driven was mean sea level) just north of Milwaukee. The shore- chosen on the basis of the field of view offered lands fall away to the north and south of this re- the transit operator, the scope of the bluff crest gional high. that could be measured from the pin and possible Except for the rocky northern portions of hazards arising from loss of the marker by bank Door County, the lakeshore consists essentially of slumping. Generally the pins were set at some con- till deposits in the form of end and ground mor- venient distance back from the crest edge but in aines and to a lesser extent of ancient lake basin positions that provided good bearings along approx- deposits the latter which have been reworked into imately 100 to 300 ft (30 to 91 m) of the shore dunes and low terraces of fine to medium grade bluff. At some sites good setback was not possible sand (Thwaites, 1956 and Thwaites and Bertrand, due to the proximity of the highway which is loca- 1957). ted within-15 ft (4.6 m) of the bluff scarp. In Beneath the cover of glacial drift the western such cases the marker was 'tied in' to other per- coastline of Lake Michigan is underlain almost en- manent structures nearby, e.g. utility poles, tirely by sedimentary rocks of Silurian age, prin- buildings, etc. cipally the mid-Silurian Niagaran dolomites which The number of bearings and distances to the dip at a low angle (approximately 10) eastward be- crest edge that were measured as well as the neath Lake Michigan. A narrow subcrop of Upper spread of the sight lines were determined largely Devonian shale and dolomite is found in the area on the basis of the crest geometry. An irregular between the cities of Sheboygan and Milwaukee. The configuration generally required a greater 'spray' principal surface exposures of bedrocks occur of bearings than one with a smoother outline. along the shoreline north of Jacksonport and extend Samples of the bluff materials were obtained across the straits to Washington and Rock Islands. along the faces of the banks. In addition beach There in the immediate offshore region are rocky sediments were collected near the toe of the bluff ledges and platforms often strewn with glacially at locations directly in line with the survey pin. derived boulders and cobbles. Bottom sediments were obtained from the nearshore Ten principal streams drain the watershed of zone in the vicinity of the shore stations. The this coastal environment (Fig. 2). Although the textural properties of the field samples were de- landscape through which these short rivers and termined by means of standard sedimentological creeks flow is generally rolling to quite subdued, techniques. some streams enter the lake through channels which Each of the three sites was surveyed a minimum have deeply incised the coastal bluffs. A hydrolo- of three times, with each survey coinciding approx- gical study conducted on three creeks near one of imately with a change in season. A fourth survey the monitoring stations (M-7) indicates that the was conducted at Station M-ll, Lake Park because sediment load contributed to the lake by fluvial the rapid rate of bank failure there warranted discharge may be quite low (Pezzetta, 1974). There- further observation. fore, the modern nearshore lacustrine deposits are The recession rates calculated for each site derived principally through the erosion of the till are mean values of several measurements scaled banks of which the shorelands are composed. from data plotted on accurate base maps. Using the total elapsed time in days between surveys the rates of bank recession were evaluated in terms of 3. Field and Laboratory Methods a thirty-day month (TDM). A sinuosity factor was invoked as a measure of During the summer of 1974 a reconnaissance the changes in crestline geometry. This factor is survey was undertaken of the entire coastline to defined as S = Lb/Lsl , where S is the ratio bet- identify a tractable program of study that would ween the actual length of the bluff crest (L, ) include some of the more problematic sections of and the length of a straight line (L ) which joins the shoreline. Seventy-six coastal sites were visi- the distal points of the surveyed section. The ted on that survey of which 15 were selected as sinuosity thus offers a numerical means for repre- 144 AF~~~~~~~~~~~~NTRMARTN 196 U~~~~ eoo,. S. IAMSLIILLINOIS D# Figure 2.Th Gratpae showing lctoso stuyarea. -2 M-7 ~~~~~~~~AFTE MARIN. aln1heWsoni hoeie 145~~~~02 4 os senting the general configuration of the bluff the nature and rate of plant succession over sand crest with time. Therefore, S indicates the degree dunes would be of particular interest to coastal of irregularity or smoothness of the crestline zone planners and managers, especially if an effec- after each survey. The limiting value for S is tive yet natural process of shoreline stabilization unity, in which case the sinuosity, and hence, the is desired within this type of'shoreline environment. shape of the bluff edge is reduced to a straight The texture of the nearby littoral deposits line. It should be noted that the sinuosity of the (Mn = 2.52@, 0.18 mm) is quite similar to that of bluff crest is not to be equated with the actual the dune sands. These fine sands which comprise the shoreline configuration at the water's edge except principal offshore sediment facies grade eastward in the case of vertical cliff shorelines. Field ob- (i.e. lakeward) into slightly coarser fractions. servations have shown that a smooth and regular Therefore, the fine clastics derived from the coas- shoreline is often found adjacent to an irregular, tal dunes tend to blanket the inshore zone proximate receding bluff crest. to the monitoring site. Subsequent dispersal of the The sinuosity also provides an indirect means sediments is effected by coastal lake currents which for assessing the nature and intensity of the ero- transfer the deposits both north and south of this sional processes and the changes in composition of location. the shoreland materials as the banks recede. A high- ly irregular crestline (S>>1) indicates that the (ii) Sandy Bay, Kewaunee County (M-7): This station processes of bank erosion operate in such a manner is located along a moderately high, steeply-dipping as to accelerate the development of landward-direc- bluff whose crest lies about 25 to 40 ft (7.6 to ted cusps or hollows while lakeward protruding ele- 12.2 m) above the lake surface.(Fig. 4). ments or promontories recede at a lesser rate. In Between the first and second surveys the mean this study it has been observed that cusp formation recession rate for the entire site was determined tends to occur in slumping bluffs of highly vari- to be 0.05 ft (0.015 m)/TDM. After the third survey able composition and texture, e.g. coarse, very it was observed that the rate had increased by a poorly sorted glacial till. factor of two, i.e. up to 0.10 ft (0.03 m)/TDM. Con- commitant with the change in recession rate, the sinuosity had decreased from 1.135 to 1.099. However 4. Observations at Stations M-2, M-7, and M-ll the greatest amount of slumping ( 2 to 4 ft., 0.6 to 1.2 m) had taken place near the distal ends of (i) Sevastopol Town Park, Door County (M-2): This the survey line where the bluff is characterized by site is located in a sand dune environment immedi- cusps and promontories. In the intervening, straigh- ately south of Cave Point, a rocky bluff northeast ter segment only minor erosion had taken place, ge- of Sturgeon Bay. The survey pin was set next to the nerally less than 1 ft (0.03 m). Hence, the rapid forest cover some 74 ft (22.6 m) from the dune slumping of the promontories in effect had gradually crest (Fig. 3). The crestline of the shorelands at reduced the irregularities of the crestline toward this location was more difficult to delineate than a smoother configuration. the bluff edges of sites composed of clayey till Where the crestline is cuspate in form the de- deposits. A distinctive change in slope along the posits tend to be comprised of poorly sorted layers lakeward side of the dunes is not generally present of variegated sand and silt with numerous stringers because the wind rounds off the tops of the dunes. and lenses of gravel scattered throughout. The However, at the base of the sand bluffs, where wave straighter part of the bluff was somewhat more uni- action had removed part of the dune face, a well- form in composition with higher proportions of finer developed, steeply-dipping scarp was observed (see clastics. The overall texture of the bluff deposits profiles in Fig. 3). The dune face above the scarp ranged from 0.184 (0.88 mm) to 6.200 (0.014 mm). dips lakeward at about 350, the maximum angle of Sampling of the bottom sediments in the immedi- repose of loose sand. However, from the dune crest ate offshore zone was rendered quite difficult by back towards the survey pin the dune surface dips the presence of a shallow platform of bedrock and lakeward at much shallower angles that ranged from indurated glaciolacustrine clays which extends out 70 to 11-1/2�. into the lake at this location. Waves sweeping over The seasonal mean bluff recession rate at this this shelf-like part of the lake floor preclude the site was remarkably uniform with values of 0.61 ft retention of unconsolidated sediments along this (0.19 m)/TDM for late fall, 1975 and 0.57 ft (0.17 surface. The littoral zone, however, was characteri- m)/TDM for late spring, 1976. After the first sur- zed by a narrow band of well-sorted, fine to medium vey, the sinuosity of the dune crest decreased from sand (2.304, 0.20 mm to 1.920, 0.26 mm) derived from 1.176 to 1.118. Hence the processes of erosion had the coastal slump blocks by the sorting action of partially straightened out the crest edge. Between waves. Lag deposits of coarse textured, fairly well- the second and third surveys the bluff continued rounded gravels comprise the blanket sediments of to recede at approximately the same rate with no the beach face. perceptible change in sinuosity. The textural uni- formity of the dune sediments (Mn = 2.22p, 0.22 mm; (iii) Lake Park, Port Washington (M-ll): The survey a~ = 0.44) is an important factor which contributes bench mark for this monitoring site consists of a to this regular type of bluff recession. fiducial mark impressed into the concrete base of Stabilization of the dune sands was only par- a light stand located in a small urban park at the tially developed between the crest edge and the north side of the city of Port Washington (Fig. 5). forest cover immediately to the west of the survey The surface of the bluff near the survey area is monument through plant overgrowth. Along the bluff smooth and even with a slight rise to the north. crest vegetation ranged from sparse to absent. It The mean elevation of the bluff crest ranges from should be noted here that a critical examination of 82 to 97 ft (25 to 30 m) above the elevation of the 146 N 5 12 7.-127� 29~) 9' 163 09') Feel 20 75.1'i 12620 i M7 Sandy Boy 1 Mile North of the Kewaunee Nuclear Power Plant -rr B F rlr1~~~~~~~~~~~~~~~~~~~~'50) 73.5'(169'50 D Kewaunee County h` c ~~~~~~~~~~~~M-2 Sevastopol Town Park. LAKE Door County P 0 Elapsed Tine Between First and Second Snoeys8 =27 Days El~p..d TI.. B . ? . . . . ~ J ' Second and Third Survey8s 229 Days First and Second Sureeyeal S= 3 e LAKE Se.ond eeddThnld 3s r.e..=19$D.Y- 50 00 aloft Pr-tfle A' 4'!..~~~~~~~~~~~~~~~~,, IA 12~~~~~~~~~~~~~~~~~~~I 12?02'48') 25lr 0.-- 958;~ . P.~Z~rL9 205 31"20 2aB6' .L8'500 Bluff Prof~le A E E -~lls 18.5"1154 30) 1 15.i'-linr 00') 11.13 235.9 nwf700l ''-ias 5. 340Y I~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ C so' Figure 3. Station M-2, Sevastopol Town Park, Door Figure 4. Station M-7, Sandy Bay, Kewaunne County. County. Bluff Profile A ......e S M'P.'t Won- ta Elapsed :Tim Betw.een FirSt and Second Suanyse 132 Days MA I Second and Third Sarneyse a34 Loke Pork - ~~~~~~~~~Thlied and Faneth SorooyS=56 Port Washington, Ozoukee County 06d9' 173' 2t'i 81.71 LM~ ~~~~~~ 54') 80,9' (10~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~72'..' co~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0 a20 Figure 5. tation M-l, Lake Pak in PortWashington Figure 6 Bluff Profiles A n ,Sato l Ozaukee~~~~~~~~~~~~~~-9 County. lake (Fig. 6). At the time of the survey the pro- consist of two distinctive facies, one of silt dis- jected horizontal distance between the crest edge tributed close to shore, and the other of sand dis- and the waterline ranged from 82 ft (25 m) at point persed in somewhat deeper water. This pattern of C to 95.5 ft (29 m) at point D. The upper 25 to 40 sedimentation implies that the inshore zone tends ft (7.6 to 12.2 m) of this bluff is an extremely to be flooded by a preponderance of fine clastic steep to vertical face. The surface of the slumped debris derived principally from the nearby clayey debris which accumulates at the base of the bluff, till bluffs by rapid erosion of the banks. however, assumes more moderate slope angles (see profiles A and C). The inherant danger which such a steep bluff 5. General Observations and Conclusions poses to park visitors prompted the City Park Com- mission to install a snow fence along the crest as Careful field observations have established a precautionary measure. The site, therefore, of- that shoreline erosion involves a complex interac- fers a dramatic example of a serious environmental tion of a number of interrelated and sometimes hazard which has arisen from rapid erosion of a subtle environmental factors. While no single cause coastal bluff within an urban setting. has been identified in this study, meteorological The seasonal variation in the sinuosity of the effects must play a significant role in the over- bluff crest is given as follows: all process. In particular, precipitation and wide S = 1.038, summer of 1975 ranging temperature fluctuations act upon the bluff 1.029, fall of 1975 materials to alter substantially their cohesive 1.168, winter of 1976 properties and thereby induce instability of the 1.138, spring of 1976 banks. Eventual failure of the bluffs stems from The mean rate at which the bluff edge receded oversteepening which, in turn, is related to the between the first and second surveys was 0.40 ft manner in which the bluffs respond to erosional (0.12 m)/TDM. During the same period slumping had processes. resulted in a minor adjustment in the crestline Slumping of large segments of clayey till, in geometry toward a slightly straighter configuration particular, is generally preceded by the develop- (S decreased from 1.038 to 1.029). The amount of ment of vertical cracks which gradually widen and material dislodged during that interval, however, cause 'peeling away' of slices of the bluff face. was not uniformly distributed along the bluff. These openings usually persist over a period of Rather, the largest segment with a maximum surface several seasons. In summer, rain and runoff which thickness of approximately 7 ft (2 m) broke away enters the fissures saturates the deeper porous from the face at point A within the southern end layers, weakens the cohesiveness of the clays, and of the section while the least amount of erosion in some cases causes washout of sand lenses. These occurred at point D near the northern end where processes, operating individually or collectively, only one-half to one foot (0.15 to 0.30 m) of mass serve to undermine and oversteepen the banks to movement had taken place. such an extent that often overhangs of matted, During the late fall of 1975 and the winter root-filled soil layers 6 to 12 in. thick (15 to of 1976 a very remarkable change occurred at this 30 cm) develop at the crest edge. In winter, the site when a large slice of the bluff some 17 ft freezing of pore water and its accompanying expan- (5.2 m) thick slipped down the sheer face of the sion causes further enlargement of the cracks and bluff. The mean recession rate for that interval fissures. This process of seasonal dilation thus of time was determined to be 2.7 ft (0.82 m)/TDM, facilitates the further entry of rain and meltwater or 6.75 times the mean rate calculated for the pre- into the banks. vious summer. At the same time the sinuosity in- When bank failure occurs, the oversteepened creased very markedly from 1.029 to 1.168. This segments either slide or cascade down the bluff event left behind at the southern end of the survey face and accumulate at the base in the form of section a large, projecting element or promontory moderately steep surfaces, somewhat akin to moun- which subsequently slumped down the bluff during tain talus slopes. If further movement of the dis- the late winter and early spring of 1976. In so lodged debris does not take place, grass sod doing, the sinuosity decreased somewhat to a value carried down with the slump may continue to grow of 1.138. in place and help to stabilize temporarily the The thick glacial deposit of which this bluff loosened mass. In some places, however, water seep- is composed consists mostly of very poorly sorted, ing out of the bluff face may offset the stabili- fine to very fine clayey sand with mean grain sizes zing effect of plant growth by keeping the slump that range from 2.624 (0.16 mm) to 3.85~ (0.066 mm). soft and plastic and, hence, susceptible to further The southern segment tends to contain a higher pro- rapid movement. portion of clayey sediments which may account for As noted above, slump block separation first its somewhat greater resistance to slumping. How- occurs in the form of vertical fractures near the ever, eventual failure does take place in this tops of the bluffs. The development of curved slip material due to the oversteepening effect produced surfaces is more complicated and arises from a sub- by undermining. sequent accretionary process in which the dislodged The inference to be drawn here is that while and slumped debris plays an important part. If a changes in the rates of bluff recession reflect sufficient mass of sediment accumulates at the base variations in the intensity of erosional processes, of the bluff, it acts as a sort of pedestal the top the geometric response (i.e. sinuosity) of the re- of which temporarily supports the base of a newly ceding crestline apparently is controlled largely developing slice in the upper part of the bluff. by the textural composition of the bluff materials. As the toe of the block which has already slumped The offshore deposits adjacent to this site is carried away, either through wave action or by 149 sheet and rill wash, its supportive effect is sub- sediments derived from the banks with the coarser stantially diminished. This then permits another components of the beach. Wave action subsequently segment to break away completely from the top of restores the equilibrium texture of the beach de- the bluff and to slide down over the remaining part posits by winnowing out the finer fractions. of the previous one. The process eventually results The degree to which lake ice causes damage to in the accumulation of a series of imbricated slump the shoreline has not been fully established along blocks whose repose angles rapidly decrease in a this part of Lake Michigan. Depending on the seve- lakeward direction. In time erosion smooths over rity of the winter season, it has been noticed that the irregularities and the surface so formed as- a well-developed ice foot as much as 80 to 90 ft sumes a concave curvature over which subsequent (24 to 27 m) wide may form along parts of the lake- slumps can slide. shore, particularly in bays and coves where protec- It is apparent that if the toe of the slumping tion from the wind mitigates the movement of the bluff exists at or very close to the waterline, the water. As long as the ice remains fast to the off- undermining effects of wave action, particularly shore bar, it acts as a sturdy barrier in shield- storm waves, will be considerable. However, if the ing the beach zone from the pounding action of beach face is quite wide, the wave energy will be waves. However, during spring breakup, and especi- dissipated before impingement at the base of the ally with the onset of strong onshore winds, the bluff can occur. In the latter case, it is evident dislodged ice floes can inflict considerable da- that the processes which induce erosion of the mage to vulnerable segments of the shoreline. bluffs are quite independent of lake level fluctua- In conclusion, caution should be exercised in tions. directly relating changes in shoreline recession The significance of variations in lake levels rates to lake level fluctuations alone. Rather, the insofar as erosion of the shorelines is concerned perspective should be one which includes a consi- lies, therefore, in the fact that higher or lower deration of the total spectrum of geological, lake stages reflect corresponding changes in the meteorological, and nearshore lacustrine phenomena amount of precipitation which falls over the Great which influence the coastal regions of the Great Lakes basin. Above-normal precipitation not only Lakes. fills the lake basins above their long term mean levels, and thereby reduces the width of the pro- Acknowledgments tective beach faces, but it also has a more perva- sive effect on the land surface. Under such cir- This project was funded (in part) by the Office cumstances the bluff materials acquire a greater of Sea Grant, NOAA, U.S. Department of Commerce, amount of moisture than they would under conditions through an institutional grant awarded to the of normal precipitation either through direct ab- University of Wisconsin, Madison, Wisconsin. sorption of rain and meltwater or as a consequence of upward migration of the local watertable. The References excess moisture reduces the cohesive properties of the bank deposits and thus renders them more sus- Brater, E., Armstrong, J.M. and McGill, M. (1975). ceptible to failure. Michigan's Demonstration Erosion Control Program, The efficacy of this process was demonstrated Update Evaluation Report, Michigan Department of quite convincingly in a similar type of study con- Natural Resources, Lansing, Michigan, 53 pp. ducted along the south shore of Lake Superior (Edil, et al, 1975). There severe erosion of the steep Edil, T.B., Pezzetta, J.M. and Wolfe, P.R. (1975). banks of red glacial till had been observed in lo- Sediment and Erosion Control in the 'Red Clay' Area cations quite removed from the influence of the of Western Lake Superior Basin, A Demonstration lake even given the highest recorded changes in the Project. Interim Report submitted to the Ashland elevation of the lake surface. In one locale con- County Soil and Water Conservation District, 65 pp. siderable bank failure had taken place some 60 ft (18.3 m) upstream from the mouth of a deep, dry Pezzetta, J.M. (1974). Sedimentation Off the Kewau- gully that dissects the high, abrupt shorelands in nee Nuclear Power Plant, University of Wisconsin that area. Instability of these bluffs had develop- Sea Grant College Program, TR No. 221, 59 pp. ed, therefore, principally through moisture seepage which was manifest high within the upper faces of Striegl, A.R. (1968). Shoreland and Floodplain the gully walls where prominent scars revealed the Zoning along the Wisconsin Shore of Lake Michigan, effects of massive slumping. Wisconsin Department of Natural Resources, Division The climatic extremes which characterize the of Environmental Protection, 50 pp. Great Lakes region also contribute to the processes of shoreline erosion. Winter observations have Thwaites, F.T. (1956). Wisconsin Glacial Deposits- shown that in spite of freezing air temperatures, Map. Wisconsin Geological and Natural History Sur- sufficient radiant energy from the sun is absorbed vey, Madison, Wisconsin. by the dark-colored, eastward to southeastward fac- ing banks to cause melting of a thin (1/8 to 1/4 in; Thwaites, F.T. and Bertrand, H. (1957). Pleisto- 0.3 to 0.6 cm) surface layer of the frozen bluff cene Geology of the Door Peninsula, Wisconsin, face sediments. Hence, around midday mud slurries G.S.A. Bull., Vol. 68, pp. 831-879. would begin to flow down the bank slopes and often run out across the beach face to the waterline. The U.S. Army Corps of Engineers (1971). Great Lakes quantity of sediment transferred from the bluffs to Region Inventory Report, National Shoreline Study, the lake in this manner is usually quite small. How- U.S. Army Corps of Engineers, North Central Dis- ever, the mechanism effectively mixes fine, clayey trict, Chicago, Illinois, 221 pp. 150 ENVIRONMENTAL IMPACT ASSESSMENT OF SHORELINE BULKHEADING: A HISTORICAL APPROACH Michael E. Penney Department of Marine Affairs-Geography University of Rhode Island Kingston, Rhode Island 02880 Abstract for making regulatory decisions in 1974. The act has a preservationist tone that is modified by One limitation to the environmental impact acknowledging a need to "accommodate necessary assessment of incremental developments, such as economic development" (1, p. 13). Counties, private bulkhead-fill projects, is the lack of cities, and towns may adopt a wetlands zoning historical data on the rate and extent of develop- ordinance under provisions of the act and adminis- ment. Analysis of a historical series of aerial ter it through an appointed wetlands board. The photographs provided a baseline for demonstrating board is charged with determining its jurisdiction the need for and effectiveness of regulation of over specific development sites and with conducting shoreline bulkheading. Projections of pre- a balancing test of the "anticipated public and regulation hulkheading rates into the regulatory private benefit" and detriment of the proposed period make wetlands regulation seem highly activity. If the benefits outweigh the detriments effective in reducing the overall rate of bulk- and the activity does not ''violate or tend to heading in a Virginia county. However, this violate the purpose and intent" of the act, a bulkheading appears to be spatially contagious permit can be granted (1, p. 20). and it is suggested that the causes of this con- The guidelines are to assist the boards in tagion need to be carefully examined by decision- making correct decisions. They assign a system makers unless lowering the rate of activity is of relative values to various wetland communities the only objective. and detail development criteria tending to preserve the intent of the act. The valuation scheme recog- nizes the importance of wetland productivity and habitat and the functions of erosion control, 1. Introduction flood buffering, and water quality control. Devel- opment criteria include establishing a need for Analysis of a historical series of aerial the wetland alteration for purposes of protecting photographs can reveal the rate and spatial existing property or gaining water access when arrangement of many coastal activities. For other alternatives are not available and when privately constructed shoreline bulkheading, the environmental value of the specific site is such information can be used to construct a low. with and without regulation scenario. The scenario In many respects, the Virginia guidelines are for a Tidewater Virginia county is interesting in similar to those developed for Corps of Engineers light of the established state and federal guide- "Section 404" regulatory programs. Of central lines for making regulatory decisions about wet- importance to this study are the concepts of cumu- lands alterations. Knowledge of the historic rate lative impacts and "anticipated" structures in the of bulkheading allows projection of amounts of Corps guidelines. In making regulatory decisions, bulkheading into the period of regulation. These consideration should be given to projections are useful in constructing several measues o reguator effetiveess. eterina-the probable impact of each proposal in measues o reguator effetiveess. eterina-relation to the cumulative effect created tion of the spatial distribution of bulkheading by other existing and anticipated struc- channelward of and along the shoreline allows inferncesabou theamouts ofwetlnds estryed.tures or work within the general area. inferncesabou theamouts ofwetlnds estryed.[3, 320.4(a) (2) (iv)] These estimates are also useful in constructing measures of regulatory effectiveness. Finally, "Probable impact" seems to refer to the uncertain private bulkheading is spatially contagious, and nature of the assessment process; "cumulative it can be inferred that case by case regulation effect" to the total effect of many activities of this type of activity can never be completely or alterations; and "anticipated structures or effective. It should be emphasized that the type work" to projections or predictions of alteration of analysis used here is applicable in other geo- levels. These factors suggest three overlapping graphic areas and for other types of activities, measures of regulatory effectiveness. One would but the specific results are not. be based on whether regulation considered environ- The Commonwealth of Virginia passed a mental and developmental values in proper balance. Wetlands Act in 1972 and promulgated guidelines Another might consider whether regulation slowed 151 the rate of development and detrimental impacts to shoreline is eroding at an average of less than 3 some specified levels. A third might be to compare feet per year. Independently repeated house counts patterns and levels of development before and after and bulkhead measurenents differed by less than regulation. The final two measures can be 15%. approached with the type of analysis presented An existing wetlands inventory of the county in this paper. The first involves definitions of was examined (12). Of the total 6,991 acres environmental and developmental values, a contro- present, only 1,163 were adjacent to the studied versial subject only touched on here. shoreline. The individual wetland units were The environmental valuation and impact assess- highly skewed toward small fringe or pocket marshes ment literature is voluminous, and much of the most with an average size of approximately 8 acres. The controversial work is oriented toward valuation of small size and proximity to developed areas of wetlands. The literature also reflects a growing these wetlands would seem to have made them candi- controversy over the balancing methodology used dates for fill over this historical period. in regulatory decision making. Wetland "preser- vationists" cite the need for regulation in face of rapid disappearance of wetlands in many areas. 3. Results Those oriented toward "balanced" development cite familiar property values for filled wetlands. But House counts and bulkhead measurements were neither side has expended much effort in even des- analyzed to determine general bulkhead construction cribing the typical development. An excellent trends and spatial distribution on this shoreline. overview of this controversy is available through General trends were determined using standard least reference to Gosselink, Odum, and Pope (4); Pope square techniques for bulkheads and houses over and Gosselink (5); and Odum and Skej (6) on the time. The spatial arrangement was analyzed using "preservationist"~ side. Walker (7 and 8) and the Wald-Wolfowitz runs test. Results of both Shabman and Batie (9) present views that seem efforts are only summarized here, but details are oriented toward "balanced" development. Lynne available on request. (10) and Huettner (11) review valuation methodology Summary data for bulkhead measurements and from an economic standpoint. The problematic house counts are presented for the 815 cell shore- nature of regulatory decisions is implicit in all line in Table 1. Simple exponential curves fit of these references, but none of them deal expli- house and bulkhead data well (r larger than .99) citly with the problems associated with slow, with acceptable levels of significance (p less incremental alterations perpetrated by private than 5%). Estimates of the amount of bulkheading landowners. Many of these developments are for added between 1972 and 1974 were obtained using purposes of erosion control, land acquisition, the fitted exponential curve and a straight line and aesthetic enhancement and the developmental values may be as difficult to determine as wet- lands values are. In some areas, incremental Table I developments may account for a large percentage of regulatory decisions. This paper describes Summary House and Bulkhead Data one type of incremental development, bulkheading, in one area. 1937 1953 1963 1974 2. Method Total bulkhead A linear grid, scaled to 500 feet, was regis- length in feet 720 5750 24005 53545 tered on aerial photographs from 1937, 1952, 1963, and 1974 of York County, Virginia. Grid cells were Total number also located and numbered on a large scale master of houses 209 367 688 1171 map and on data sheets. Locations of houses within 500 feet of the shoreline were interpretively transferred to the master map for each year of photography, and then counted. Interval scale interpolation between 1963 and 1974. These esti- measurements of bulkheads were made on each set mates are 14,382 and 5,371 feet added for the of photographs and the data recorded on the sheets. exponential and straight line approximations, The mean distance of the bulkheads from the toe of respectively. The reason for obtaining these very the upland bank was about 15 feet and highly skewed different estimates is for later comparison to to smaller distances. Due to gaps in the photo- estimates of the amount of bulkheading actually graphic record, measurements and counts were added between 1972 and 1974 under the Virginia limited to about 80% of the shoreline. Wetlands Act as reported in Jones (13). Of the studied shoreline of 815 grid cells Detection of spatial patterning of bulk- only, two short sections are in nonresidential heading was essential to this study, but the usage. One 1500 to 2000 foot, 3 to 4 cell, section approach was kept simple. If the spatial distri- contains an oil refinery embedded in residential bution on the developed shoreline for each year is shoreline. Another 2000 foot section contains a random, the implication would be that each grid Coast Guard training center similarly surrounded cell is equally likely to contain a bulkhead. If by residential housing. One 6 to 8 cell section, the spatial distribution is nonrandom or clustered, with one bulkhead and five houses in 1974, is the implication would be that some grid cells are eroding at over 3 feet per year. The remaining more likely to contain a bulkhead than others. 152 Similarly, the spatial distribution of developments Also, bulkhead additions, which have an average on the entire shoreline is significant to this length of less than 250 feet, are nonrandomly study. distributed on the space available for bulkheading. A set of Wald-Wolfowitz runs tests were made For each test, the probability of incorrectly of the hypothesis that bulkheading is randomly rejecting the hypothesis of randomness is less distributed on the developed shoreline, and that than .00003. The table also illustrates that the developments are randomly distributed on the entire actual average bulkhead run length is longer than shoreline. For this test, runs were defined as the predicted average run length. adjacent grid cells of identical, say house pre- sence, elements. Runs of each element could be any number of grid cells long. Runs of one element 4. Discussion are separated by runs of a different element, say cells with no houses. The runs test allows calcu- Jones states that rough statewide estimates lation of the expected number of runs and the of yearly amounts of wetlands destroyed before the variance under the hypothesis of randomness, Virginia Wetlands Act (app. 450 acre/year) and knowing only the total number of runs actually after (app. 25 acre/year) make regulation under present and the number of cells containing each the act appear to be very effective in that wet- type of element. The results of the test show lands destruction was reduced to 5.5% of the that the developed shoreline was clustered along original value (13, p. 50). Pre- and postact the entire 815 cell shoreline with a probability estimates of amounts of bulkheading added and of less than .00003 of incorrectly rejecting the wetlands destroyed from analysis of aerial photo- hypothesis of random spatial distribution for both graphs provide an interesting contrast. 1963 and 1974. Linear (5,371 foot) and exponential (14,382 Runs tests for the spatial distribution of foot) estimates of the amount of bulkheading that bulkheads on the developed shoreline are more would have been added between 1972 and 1974 with central to this study and are presented in Table 2. no regulation are contrasted with Jones's esti- The results show that bulkheads are nonrandomly mates of 1,300 to 1,400 feet from permit data distributed on the 1963 and the 1974 shorelines after regulation began. These contrasts suggest (no tests were performed for the 1937 and 1952 that actual amounts of bulkheading added under data). regulation were from 7.5% to 26% of the amounts expected (for the exponential and linear estimates, respectively) had the historical trends continued. Table 2 Neither of these estimates make regulatory effec- tiveness appear quite as high for bulkheading in Bulkhead Runs Test Results this county as the measure suggested by Jones for amounts of wetlands destroyed statewide. The amount of wetlands destroyed in York 1963 1963* 1974 County can also be estimated, assuming that the 53,545 feet of bulkheading present in 1974 was installed 15 feet from the toe of the upland bank Number of Cells (the average value) in a continuous fringe marsh at least 15 feet wide. With these assumptions, the amount destroyed would be about 22.4 acres. Bulkheaded 69 87 189 Considering that 1,163 acres remain on this seg- ment of shoreline, the amount destroyed is about Unbulkheaded 276 290 290 1.9% of that originally present. Since 56% of the bulkheading present was added between 1963 and 1974, approximately 12.5 acres were destroyed Number of Runs during that period, or about 1.1 acres/year. Jones estimates that .14 acres/year were destroyed in this county during the first two years of regu- Predicted 111.4 134.9 226.7 lation (13, p. 20). The implication is that regu- lation reduced wetlands loss to about 12.7% of the Actual 74 102 173 amount expected had historical trends continued. But the model used to estimate wetlands losses in this study is unrealistic. Since wetlands are not Run Length a continuous fringe in this county and bulkhead to bank distances are skewed to smaller than 15 feet, the estimate of 1.1 acres/year is probably large Predicted 1.2 1.3 1.7 and actual/expected loss of 12.7% too small. Whatever estimate is chosen, regulation Actual 1.9 1.7 2.2 appears to have been very effective in reducing the historic rate of bulkheading and wetlands Actual/ loss. If historic bulkheading practices had Predicted 1.6 1.3 1.3 continued (keeping the assumption about a con- tinuous wetland fringe), approximately 12% of the current wetlands acreage present would have been Additions of bulkheading to 1963 shoreline. destroyed with complete bulkheading. If regula- tion is just a rate-controlling process, it is 153 likely that bulkheading will still become complete 5. Pope, R. M., and Gosselink, J. G. A Tool for at a later date. If regulation is to be more than Use in Making Land Management Decisions a rate-controlling process, some attention would Involving Tidal Marshland. Coastal Zone have to be given to the spatial pattern of bulk- Management Journal, 1973, 1, 65-74. heading in the regulatory decision-making process. Over the historical period covered by this 6. Odum, E. P., and Skej, S. S. The Issue of study, most bulkheads have not only been installed Wetlands Preservation and Management: A Second close to the bank but closer to each other than View. Coastal Zone Management Journal, 1974, random placement would lead us to expect. The 1, 151-163. implication of this spatial clustering is that the decisions of private homeowners to install a bulk- 7. Walker, R. A. Wetlands Preservation and head have somehow been influenced by the previous Management on Chesapeake Bay: The Role of decisions of their neighbors. Three common reasons Science in Natural Resource Policy. Coastal for bulkheading are erosion control, land acquisi- Zone Management Journal, 1974, I, 75-101. tion, and aesthetic improvement. And it is easy to see how any one of these might be contagious 8. Walker, R. A. Wetlands Preservation and in a neighborhood on a short stretch of shoreline. Management: A Rejoinder--Economics, Science, For instance, erosion might be naturally high or and Beyond. Coastal Zone Management Journal, exacerbated by existing bulkheads. Or the change 1974, 1, 227-233. in appearance and increase in amount of property from one bulkhead might be attractive to a neigh- 9. Shabman, L. A., and Batie, S. S. Estimating boring homeowner. However, current state and the Economic Value of Natural Coastal Wetlands: federal guidelines make it clear that proposed A Cautionary Note. Research Report AE 30, alterations must fulfill some real need and/or August 1977. Department of Agricultural result in minimal environmental damage. In many Economics, VPI, Blacksburg, Virginia. situations, the individual need may be unclear and the incremental damage seem minimal; but as this 10. Lynne, G. D. Incommensurables and Tradeoffs study shows, there can be a natural tendency for in Water Resources Planning. Water Resources bulkheads to be contagious. Detection of conta- Bulletin, 1977, 13, 93-105. gious activities in specific areas could lead regulatory agencies to a closer examination of 11. Huettner, D. A. Net Energy Analysis: An the cause of these activities. Economic Assessment. Science, 1976, 192, This study has used historical bulkheading 101-104. data to make inferences about regulatory effec- tiveness in one locality in Virginia. By all the 12. Silberhorn, G. M. York County and Town of measures used, regulation seems to have reduced Poquoson Tidal Marsh Inventory. Special Report bulkheading and fill activity to small percentages No. 53 in Applied Marine Science and Ocean of what projections of historical trends would lead Engineering, Virginia Institute of Marine us to expect. But these levels of regulation may Science. Gloucester Point, Virginia, 1974. not satisfy everyone. For those who feel regula- tory decisions should be more closely examined, 13. Jones, J. C. Local Environmental Management-- discovery of spatial clustering of bulkheads, or A Case Study: The Virginia Wetlands Acts, other activities, could lead to closer examination 1972-1974. Unpublished masters thesis, of the reasons for private bulkheading. Although College of William and Mary, Williamsburg, the results of this study are applicable only to Virginia, 1976. one area, the method could be used with photo- graphic and other historical data to develop care- ful estimates of regulatory effectiveness. References 1. Laws of Virginia Relating to Wetlands and Subaqueous Lands. (Issued by the Marine Resources Commission, Newport News, Virginia.) Charlottesville, Virginia: The Michie Company, 1975. 2. Wetlands Guidelines. (Issued by the Marine Resources Commission, Newport News, Virginia, 1974.) 3. Regulatory Program of the Corps of Engineers. 42 FR 37122-37164. July 19, 1977. 4. Gosselink, J. G., Odum, E. P., and Pope, R. M. The Value of the Tidal Marsh. Pub. No. LSU-SG- 74-03, Center for Wetland Resources, Louisiana State University, Baton Rouge, 1974. 154 DRIFTING BUOYS FOR OCEAN DATA COLLECTION Dr. John Garrett Institute of Ocean Sciences P.O. Box 6000 Sidney, B.C. V8L 4B2 Canada Two polar orbiting satellites carrying the Argos equipment are planned to be in operation at Abstract the same time, so that observations can be obtained from any point at least four times per day. How- Development of low cost drifting buoys is now ever the satellite orbits are not separated by 900 to the point where they may be considered a proven so that these observations will not be equally tool for ocean data collection. Recent experiments spaced in time. Also because of the low altitude in preparation for the First GARP Global Experiment (830 km) of the orbits data will only be received have demonstrated lifetimes in excess of a year. from a band approximately 5000 km wide centred on An operational data processing and dissemination the satellite sub track. Hence the age of the service will soon be available. Although collec- most recent data available at any given instant tion of ocean current data continues to be of will vary considerably depending on the location interest, buoys intended for other purposes are of the platform involved relative to the current discussed, using examples from experiments conduc- location of the satellite. As an example of this, ted by the Canadian Ocean Data Systems project. the age of the data expected to be available up to a transmission time of 0100 GMT, for the Southern Hemisphere, is shown in Figure 1. The number of passes per day permitting determination of loca- tions depends on the latitude of the platform and Techniques for locating and recovering data from remote platforms using polar orbiting satellites have matured to the point where they may be considered operational rather than experi- mental. Although they depend on the complexity of the sensor system the total costs for each day of operation of simple meteorological buoys are comparable to the cost of a radiosonde or an XBT. Thus buoys of this type can be expected to receive increasing use for routine data collection from presently data sparse areas of the ocean. The common elements to all such systems are the satellite equipment and ground data processing and dissemination system, and the platform trans- mitters. The sensors, hull characteristics and use of data will be different from system to sys- tem depending on the objectives involved. 2. Satellite Data Location and Retrieval System Although several experimental systems have been used up to now, such as IRLS, Eole, and Nimbus-6 RAMS, an operational system is planned for the period 1979-1985. The Argos system, as it c 1 2 is called, will be carried aboard a series of operational meteorological satellites beginning L K 1111 with Tiros N, and followed by NOAA-A through NOAA- G. This system will be supported by a data pro- cessing and dissemination service called Service 4 6 7 Argos, which plans to have observations available - t within 6 hours after the time they are transmitted. m This service is operated by the Centre National d'Etudes Spatiales of France on a cost recovery Figure 1. Age of latest data for transmissions up basis. The following is a summary of information to 0100 hours GMT, in hours. available from Service Argos. 155 varies from 6 at the equator to 10 to 12 at �600 temperature and barometric temperature were con- and 26-27 at �900. The location accuracy will be structed for a pre FGGE test as part of the better than �5 km. Canadian Ocean Data Systems programme, by Hermes Electronics Ltd. of Dartmouth, Nova Scotia. These The data transmitted from the buoy or plat- buoys were designed to be launched from ships of form are stored on the satellite until the satel- any type without use of cranes or without the ship lite passes over one of three ground stations, stopping, i.e. they were to be simply dropped into where it is transmitted to the ground and sent to the water from deck level. The battery life was NOAA in Suitland, Maryland, U.S.A. There the to be 6 months in water temperatures of O0C and Argos data is separated from the data from other one year in water temperatures of 100C. Eighteen experiments on the satellite, and is relayed to of the buoys were deployed in the Southern the Service Argos Centre at Toulouse, France. In Hemisphere and two in the North Atlantic between Toulouse the platform locations are calculated, January and May 1977. the telemetry data translated to physical values as requested by the user, and the data is stored The life histories to present are shown in or disseminated as requested by the user. The Figure 2. Half of the buoys had ceased to trains- data is available in a variety of forms such as mit after 366 days of operation. Of these, two4 telex, magnetic tape, direct line computer to had been deliberately turned off, two had exhausted computer transfer, or mailed listings. their batteries after nearly a year, two had failed after being washed ashore and four were premature In order to operate with the satellite each electronical or mechanical failures. Another buoy buoy or platform must have a suitable transmitter. has since exhausted its battery after 394 days, To insure against disruption of the whole system by while the 9 still transmitting have been transmit- poorly designed or constructed transmitters the ting for an average of 450 days (to June 1, 1978). transmitters must undergo an acceptance and certi- If the two buoys which were switched off are ex- fication process administered by Service Argos. cluded we are left with a population of 18 buoys The transmitters already accepted and certified of which the ninth failed after 394 days. Eleven typically weigh I kg and have an average power of the 20 have never gone ashore: the fifth consumption of 60 to 100 my, including trans- failure among those occurred after 366 days and missions. The total power requirement for a given the sixth after 390 days. It is somewhat difficult platform will of course depend on the requirements to apply classical failure analysis techniques to of the sensors and signal conditioning. this sample to estimate Mean Time Before Failures because of the question of how to treat buoys At the same time that the data received which have gone aground. It seems clear that at aboard the satellite is being stored in the on least half of the buoys can be expected to be board memory it is rebroadcast on the satellite transmitting at the end of a year, which might VHF beacon. With a suitable receiver it is pos- thus be taken as an upper limit to the time over sible to receive this data directly and to extract which their cost can be amortized. the sensor messages and convert them to physical values. This will minimize the time delay in Fot many buoy applications the buoy will be acquiring the data, but data can only be acquired considered to have gone out of service when it runs from those platforms which are in view of the aground. The rate at which this occurs will depend satellite while the ground station is also in view on the locations at which the buoys are launched. of the satellite. This limits the effective The 20 FGGE prototypes were generally launched in radius of reception of such a station to less than mid ocean regions and thus might provide an estim- 4000 km. Also determination of the location of ate of the grounding rate to be expected from a platforms requires independent knowledge of the network covering the whole ocean. It is in fact a satellite orbit. conservative estimate since the replenishment of such an array would usually take place on the up- 3. Reliability and Costs stream edges, so that the buoys used for replenish- ment could expect much longer than average life- The costs of obtaining data in this way in- times. Ten out of the twenty prototypes had volve capital costs for the buoy or platform plus either failed or gone aground after 213 days, which operating costs associated with the data process- might be used as an estimatE. of the minimum time ing and dissemination. In order to estimate the over which the cost can be amortized. cost-effectiveness it is often interesting to estimate the capital cost per unit time or o'bserv- At the present time the simplest transmitters ation by dividing the capital costs for a platform cost less than $2,000, complete with antenna, by the time the platform can be expected to serve while complete FCCE type buoys with barometric or the number of data points it can be expected to pressure and temperature sensors, cost less than provide. For many drifting buoy applications the $10,000. Thus the average capital cost per day buoys are considered expendable so that the amort- for such a buoy is less than $50 per day and ization period depends on the reliability of the probably less than $30 per day. To this must be buoys. added the cost of data location and retrieval. Service Argos estimated that to recover their A recent experiment provides an indication of costs they would have to charge a maximum of $20 the reliability presently achievable. Twenty per day for calculating the buoy location and drifting buoys measuring sea surface temperature converting the telemetry values to physical 156 Buoy Id 0460 f Failure 0275 1 Grounding 0274 -- -' Fatal Grounding 1327 -j 4DSwitched Off 0242 1310 0720 .. 0232 0426 1645 1504 0570 0740 1602 0676 1543 1361 0640 0546 0352 I ~ ~I i I iI 0 1 2 3 4 5 6 x100 Days Figure 2. Histories of FGGE Prototypes to June 1, 1978. parameters. However this figure was based on an be much more useful and interesting than might be estimate of the minimum number of platforms which expected. For a typical spar buoy drogued at 20 m would be operating so that if more platforms are or 30 m it is very difficult to determine from the operating the cost will come down. There will be track when or whether the drogue has failed a cost associated with getting the data from (Kirwan, 1978). The average speeds are entirely Toulouse to the user, but in any case the maximum consistent with those to be expected from currents. total cost is $70 per buoy day, with the probable cost more in the range of $40 to $50 per buoy day. The Hermes FGGE prototype buoys mentioned This is of course comparable to the cost of an XBT earlier carried barometric pressure transducers. or a radiosonde. For several periods during the experiment three or more buoys were distributed in such a way that the 4. Measurements pressure measurements could be used to make an estimate of the vector geostrophic wind. The To date the principal use of satellite track- fluctuations in the north and east components of ed drifting buoys has been to measure ocean cur- the wind could be correlated with the fluctuations rents or the motion of pack ice or icebergs. in the north and east components of the buoy Results of such investigations have been presented velocities. In 19 out of 24 records of between by Cresswell (1978), Kirwan, McNally, Reyna and 36 days and 201 days in length the correlation be- Mervell (1978) and Martin and Gillespie (1976). tween the wind component and the buoy velocity component parallel to it was higher than would be Elementary calculations of the probable significant at the 95% level. Only 7 out of 24 effects of wind drag and current shear between the correlations between the wind component and the surface and the depth of the drogue, if any, velocity component directed at 90� to the left of suggest that errors of several centimeters per the wind were significant at this level, although second should be expected. More complex calcula- it should be noted that all but 5 of the 24 were tions including wave effects are even less re- positive, indicating a tendency to the left of assuring (Nath, 1977a). A few careful model the wind. (These buoys were in the Southern experiments (Nath 1977b, Vachon 1977) and a few Hemisphere.) The correlation coefficients were careful full scale tests (Saunders 1976, Vachon typically 0.25 to 0.50, i.e. the fluctuations in 1977) tend to ~upport them. The effects of vector the wind explained 25% to 50% of the fluctuations errors of this magnitude could be serious as in the buoy motion. The interesting feature is pointed out in Kirwan, McNally, Chang and Molinari that the regression coefficients for the signifi- (1975). cant correlations ranged between 0.01 and 0.02, i.e. the buoy speed was between 1% and 2% of the However, field data from actual buoys seem to geostrophic wind speed. 157 This is surprisingly low for a buoy which is undrogued and which has nearly half of its cross sectional area above the water, with a total length of less than 2 meters. The motion also in- cludes not only the buoy response but the response of the surface layer of the ocean in which the buoy is imbedded. On the other hand the geostro- phic wind probably overestimates the surface wind so that the response to the surface wind could be higher. The only conclusion that can be drawn is that during periods of light winds even an un- drogued buoy provides reasonably good measurements of water velocities. Whatever the credibility of drifters as indic- ators of water motion, they can also be used as ra reliable inexpensive platforms to obtain other types of information. The first large scale example of this will be provided during the FirstA CARP Global Experiment, when an array of 300 drifting buoys provided by 8 different countries will provide barometric pressure and sea surface temperature measurements from the oceans of the Southern Hemisphere. This array will be deployedt by the cooperating ships of a number of countries between October 1978 and June 1979. Figure 3. Prototype "Fast 'Drifter" for thermal Other possible surface level sensors would be structure transect measurements using wind speed and direction and air-sea temperature thermistor chain. difference, which would then permit an estimate of the air-sea heat exchange. Also useful and prac- tical would be a wave gauge of some type, with have been constructed, again as part of the processing circuits on the buoy to reduce the data Canadian Ocean Data Systems Project at Hermes to a few statistical values for transmission as Electronics. These buoys, shown in Figure 3, are described by Clayson (1975). This could be con- designed to be blown along by the action of the bined with satellite remote sensing of wave wind, and hence have a maximum surface area. They direction to reduce some of the ambiguities and will be deployed in the North Pacific during the uncertainties of the satellite measurements. summer of 1978. oceanographers of course are interested in 5. Conclusions subsurface data. Many thousands of XBTs are drop- ped annually to try to monitor just the thermal The operational use of satellite tracked structure. One obvious alternative is drifting drifting buoys will be possible with the advent of buoys equipped with thermistor chains. Here a an operational data processing and dissemination choice presents itself. If the buoy does not move service. Expendable buoys can be cost effective with respect to the water column the changes as their reliability permits amortization of their recorded in the temperature structure will reflect capital cost over periods of 200 to 300 days. A change in the structure of the water column, which variety of measurement capabilities have already might be due to internal waves, to heating and been demonstrated for such buoys and more will be cooling or to a variety of other causes. If the developed soon. Thus we should expect expendable ocean temperature field has significant variabil- drifting buoys to become an increasingly common ity on space scales of 100 Ion a buoy array with a tool for various measurements from ocean areas. resolution of 100 km will be required to delineate this field, i.e. 100 buoys in an area 1000 km square. The other alternative is to deliberately Akoldeet make the buoy move with respect to the water, so Akoldeet that the thermistor chain provides a transect or The buoys described were constructed as part a section. Then the number of buoys required toofteCndaOcnDtaSsmprjtwhh resove te tmpertur fied dpend onthe imehas been managed by John Brooke and Jim Elliott at scale as well. If the time scale of ocean temper- the Bedford Institute of Oceanography and Graham ature variability is a month, and if the buoys Smith at Hermes Electronics Ltd. Andy Reid of move 300 km in a month, only one third as many Hermes Electronics was primarily responsible for buoys would be required as for buoys fixed the drift buoys. The data were obtained using the relative to the water. ~~RAMS on the Nimbus 6 satellite with the help and4 To examine the feasibility of this idea at cooperation of the NASA Nimbus-6 team. the same time as conducting reliability tests of a new type of thermistor string a few prototypes 158 References Clayson, C.H. (1975) Wave Data Telemetry via the Nimbus-F Satellite; Proceedings of Conference on Instrumentation in Oceanography held at University College of North Wales, Bangor, 23-26 September 1975. Cresswell, G.R. (1976) A Drifting Buoy Tracked by Satellite in the Tasman Sea; Aust. J. Mar. Freshwater Res., 27, 251-62. Cresswell, G.R. (1978) The Trapping of Two Drifting Buoys by an Ocean Eddy., Deep Sea Research (in press). Kirwan, A.D., G.J. McNally, E. Reyna, W.J. Merrell (1978) The Near Surface Circulation of the North Pacific: J. Phys. Oceanog. (in press). Kirwan, A.D., G. McNally, M.S. Chang, R. Molinari (1975) The Effect of Wind and Surface Currents on Drifters, J. Phys. Oceanog., 5, No. 2, pp. 361-368. Martin, Pat, C.R Gillespie (1976) Arctic Odyssey - Five Years of Data Buoys in AIDJEX: Proc. Symposium on Meteorological Observations from Space June 1976, Philadelphia Pennsylvania. COSPAR, IAMAP/IUGG, WMO, AMS. Nath, J.H., (1977a) Laboratory Validation of Numerical Model Drifting Buoy-Tether-Drogue Systems, Final Report to NOAA Data Buoy Office on Contract 03-6-038-128. Nath, J.H., (1977b) Laboratory Model Tests of Drifting Buoy and Drogue, Engineering Report for NOAA Data Buoy Office on contract 03-6- 038-711. Saunders, P.M. (1976) Drifting Buoy Lagrangian Test, Final Report of Woods Hole Contract NAS 13-5 with NOAA Data Buoy Office. Vachon, W.A. (1977) Current Measurements by Lagrangian Drifting Buoys - Problems and Potential : Proc. IEEE Oceans '77, pp. 46 B1 - 46 B7. 159 A STABLE PLATFORM FOR MICROSCALE STUDIES OF VELOCITY, TEMPERATURE AND CONDUCTIVITY Bill P. Johnson Scripps Institution of Oceanography La Jolla, California 92093 Abstract To verify the behavior of the instrument body, vertical and horizontal accelerometers were In the past ten years, several investigators have constructed instruments to measure small scale fitted to measure an y r esidualmotion s. These processes in the oceans. In attempting to describe devices are balance beam type accelerometers and thus are also useful for tilt measurements. In these processes, insufficient attention has been order to relate these measured motions with the given to the problem of the sensors or the record- order to relate these measured motions with the ing of body motion relative to the media which is rotation of the instrument, due to the autorota- being measured. Many investigators chose to tion of the wing assembly, a magnetometer of the beig masued.Many investigators chose to limit their studies to scales which are unaffected flux-gate type was also installed. This device by sensor motion or to make certain assumptions measures rotational displacement relative to the eartsmantcfedWigpthwsslcd about frequencies and amplitudes of the sources of earth's magnetic field. Wing pitch was selected such that the instrument would have a fall rate of noise. When attempting to resolve microscale from 5 to 15 cm/sec. Later, computer modeling by phenomena, these problems become acute if not ab- otsee Laer conrmdlic Mortensen and Lange (1977) confirmed the applica- solutely limiting. An instrument was developed at bility of the design approach and experimentally Scrips Intituion f Dcenogrphy n 194 tobility of the design approach and experimentally Scripps Institution of Oceanography in 1974 to determined data agreed favorably with the computer attempt to advance the state-of-the-art in ability model. The d evelopment of the VMSR occurred at model. The development of the VMSR occurred at to accurately resolve these small scale processes. It has at its heart a free falling body which em- approximately the same time as the Osborn velocity ploys a large overmass and uniform drag character shear probe (Osborn, 1974) which employs the air- istics to stabilize its motion during data foil probe developed by Siddon (1974). The Osborn acquisition. In addition, several body motion instrument is more closely tailored to cross flow parameters are carefully recorded. The sensors velocity shear measurements while the VMSR mea- pa~~~~~~~~~~~~~sraetes ainre ca trefullyt raec ordd h sensors are selected for optimum resolution in time and sures finer scale turbulent rates of strain. space and very careful examination of sampling techniques were undertaken. Early prototyping of the VMSR employed an FM data acquisition which was selected because of its compatibility with an existing data transcription HISTORY - VELOCITY MICROSTRUCTURE RECORDER (VMSR) system. Several test cruises were conducted in the San Diego Trough and the instrument was used In 1973 development was begun on a new device on its first cruise in the Gulfo expedition in the which capitalized on the experience gained at SIO Gulf of California in 1974. Prior to the Mixed with microstructure recorders. While temperature Layer Experiment (MILE)(August, 1977) in the Gulf and conductivity micro and fine scale experiments of Alaska, a digital data acquisition system was had been done by several researchers (Gregg and installed which stores data on a cassette recorder Cox, 1971),small scale velocity measurements had The MILE requirement of working in the surface not been previously carried out. It was obvious wave field led to the installation of an hydraulic that in order to measure water velocity independent wing deployment system to positively lock the of body motion, a very stable platform would be wings in position. Since horizontal velocity required. A large spar type apparatus was chosen fluctuations are sensed on one of the wing tips, so that a large overmass and vertical righting any "flopping" of the wings would introduce spur- moment could be employed. A 25 cm outside dia- ious velocity signals. This system was removed in meter 7075-T6 aluminum tube with 1.4 cm wall thick- preparation for the POLYMODE Local Dynamics Exper- ness was fitted with hemispherical end caps to iment. A conductivity cell and platinum resistance achieve a high ratio of displacement to weight. thermometer were then added. The instrument is For descent rate control, the tube was fitted with shown in Figure 1. light weight carbon fiber reinforced "wings" which could be retained alongside of the tube during THE SENSORS free fall to the desired depth and then deployed in a horizontal mode. This configuration allowed The method used on the VMSR for velocity the fitting of a heavy 26 kg lead weight as low as measurement is the classic "hot wire anamometry". possible on the instrument to provide a very stable During the development of this instrument, several fall characteristic. It was hoped that this con- types of thermistors were evaluated. Cox et al. figuration would yield the least perturbation in (1969) used several glass thermistors in parallel the presence of velocity shear layers as the in- for obtaining temperature measurements. However, strument descended in the ocean. the unbacked "flake" was later thought to be 160 capable of faster response times. The flake con- In addition to the differentiated velocity sists of thermistor material with leads attached and temperature gradients, 3 axis accelerometers, and coated with a thin plastic dialectric mater- conductivity, platinum resistance temperature, ial. These flakes were used also as early velo- magnetometer and pressure, several engineering city sensors. Sizes ranging from 20 x 20 x 2mills parameters are sampled. These include supply vol- to 100 x 200 x 4 mills were used with coatings tages, reference voltages and wing deployment ver- ranging from .2 to 1 mill. Response characteris- ification. The parameters and their resolution tics of some of these thermistors were reported and accuracies are shown in Table 1. by Mortensen (1976). All of the flake variations were extremely fragile and the thin plastic coat- DATA COLLECTING SYSTEM ing was easily destroyed by impact with oceanic biota. An additional problem of the flakes was The electronics package located in the large the relatively high noise level assumed to be as- tube is preprogrammed to sample the pressure sociated with the method of lead attachment. The gauge frequency until the instrument has descended recent availability of substrate backed flakes to the desired depth. When trigger depth is led to the use of BeO backed flakes for fast tem- reached the wings are deployed into the horizontal perature measurements and alumina backed flakes position and recording is commenced. for velocity measurements. These units use a gold masking process which seems to reduce the The velocity and temperature signals from the noise problem considerably. We are currently thermistors are actively filtered at 25 Hz, the evaluating these units with a thin glass coating Nyquist frequency (sample frequency/2) with an 18 on the thermistor material with an uncoated re- db per-octave roll-off and appropriately scaled verse side. These units are very rugged, can be for a + 10 volt range. The remaining signals are made very small (50 x 50 x 10 mills) and show sampled with a 2.5 Hz sample frequency and under some promise. programming control are multiplexed to a 12 bit analog to digital converter. Passive unheated thermistors are used for tem- perature gradient measurements with one located on Six channels are sampled at a 50 sample per a nose cone at the lowest point on the bottom of second rate with one of the 6 channels subcommuta- the instrument and 1 located on a wing tip. Ther- ted so that 10 parameters are sampled at 50 sam- mistors heated to 400C above water temperature ples per second. The serial bit stream thus with a constant current are located 1 cm distant emanating from the analog to digital converter is from the temperature measuring thermistor for the passed in alternate seconds of time to 1 of 2 3600 velocity gradient measurement. An additional vel- bit memory banks. At the end of 1 second of data ocity sensor is mounted on a 2nd wingtip. collection, data storage is switched from one mem- ory bank to the other and the cassette tape re- A 3000 psi digiquartz pressure transducer is corder is started and then receives the previous mounted on the lower end cap. This device has a 1 second's 3600 bits of data. The writing on tape nominal 40 kHz output which is period counted requires about 1/4 of a second so that the tape (4096 periods of a 2.09 MHz clock) to allow an deck is idle 75% of the time. In order that 8 cm resolution in 2000 m of water. An NBIS Inc. start-stop transients of thetape recorder do not 3 cm conductivity cell is used along with a plati- induce noise in the high gain gradient circuits, num resistance thermometer (PRT) mounted on a sep- all signals are passed to and from the tape via arate pressure case on the lower end cap. The optically isolated electronics and the tape re- conductivity cell and PRT are mounted in the same corder has its own battery pack. horizontal plane 3 cm apart. Response time of the PRT is 300 ms while the conductivity cell has a At the end of a predetermined elapsed time 3 cm length. At a fall rate of 10 cm/sec these the recording process is stopped, the ballast time constants are approximately matched. weight is jettisoned via an explosive squib and the instrument becomes positively buoyant and The circuit used for the conductivity cell ascends to the surface. A leak detector in the (Fig. 3) is transformerless and thus avoids many lower end cap can terminate the recording process of the problems usually encountered with conducti- and jettison the ballast also. Hydrostatic back- vity measurements. A 100 kHz signal is divided by up release systems are fitted to protect against 10 and used to excite two of the four electrodes electronic system failures. with a precision constant square wave current. A track and hold amplifier samples the voltage pre- We have been able to successfully use enter- sent on the remaining two electrodes at a point tainment quality cassettes with this recorder and 40% and 80% on the square wave excitation wave thus are able to record greater than 10 mega bits form. This scheme does not require any correc- of data on a 120 minute cassette. At the sample tion for the reactance of the cell. The plus and rate of 6 channels x 50 samples per second or 300 minus voltage is rectified and fed to an optically x 12 bits per second,we sample 2700 seconds. At a coupled amplifier and the low impedance signal is fall rate of 12 cm/sec, we sample 325 meters of then fed to the main pressure case and hence to water column. The data acquisition system is the data acquisition system. The disadvantage of shown in block form in Figure 2. this circuit is that it requires considerable DATA RECOVERY power (0.8 watt) and is relatively expensive to construct. The development of the micro-computer within the last five years has had as great an impact on 161 oceanography as the development of the transistor Office of Naval Research under Contract in the late 1940's. Fully operational computers N00014-75 C-0152. can now be hand-carried aboard ship. We use the Tektronics 4051 Graphic Display unit as the basis REFERENCES of our data handling system. This unit has inter- faced to it a cassette reader, line printer, hard Cox, C. S., P. W. Hacker, B. P. Johnson and copy unit and also is interfaced to an IBM com- T. R. Osborn, 1969, 1969, Finescale of tem- puter when used on land. This unit uses a higher perature gradient. Marine Temperature level language (Basic) so that sophisticated pro- Measurements Symp., Trans. Mar. Tech. Soc., gramming can be accomplished in 10-20% of the time June, 95-104. that would be required for a mini-computer system. The entire system costs less than $15,000 and Gregg, M. C. and C. S. Cox, 1971, Measurements of maintenance costs have been less than $1,000/year. oceanic microstructure of temperature and electrical conductivity. Deep-Sea Res., 18, With the Tektronics system we are able to 925-934. play our recorded cassettes into the graphic unit displaying analog and/or spectral plots and Mortensen, A. C., 1976, Optimized system for high numeric variance data all within minutes after the resolution measurements of oceanoc tempera- instrument has been recovered from a launch. The ture and velocity. PhD Thesis, Univ. of Ca., simplicity and versatility of this system has San Diego, 168 pages. allowed greater attention to the VMSR itself. Mortensen, A. C. and R. E. Lange, 1977, Design SOME COMPARATIVE DATA criteria for wing-stabilized free-fall vehicles. Deep-Sea Res., 23(12),1231-1240. Figures 4 and 5 show data taken on a pre- Polymode shakedown cruise and on the MILE experi- Osborn, T. R., 1974, Vertical profiling of velo- ment respectively. The MILE launch was a near city microstructure. J. Phys. Oceanogr., 4, surface deployment and clearly shows the effects 109-115. of the surface wave field on the motional behavior of the instrument. Any attempt to resolve the measured velocity fluctuations without complete understanding of the motions would obviously be futile. The San Diego Trough data shows the instru- ment's response to a velocity shear layer. The Table 1 fluctuations of vertical velocity measured at the Sample lowest point on the instrument are encountered 1.2 Parameters Interal Resolution Accuracy rotational periods later by the wings at the highest point on the instrument and where virtually A V vert. .02 sec .1 /sec/sec 2 mmsec/sec all of the drag or lift exists. This introduces A V horiz. .02 sec .2 mm/seclsee 1 mm/se/sece a perturbation in the fall rate and in the instru- A T vert. .02 sec 10 p�C NA ment tilt (inserts). The instrument falling at a Thoris. .02 sec 2Oo~C NA 12 cm/sec with a spatial separation between nose sensor and wing of 250 cm and rotating at 18 z aec. .02 sec 100 P 0 100 P a seconds per rotation will encounter the shear x acc. .02sec mI tilt 1Im tilt layer 250/12 = 31.3 seconds or 1.15 rotational periods later. This notional behavior is used to y acc. .02 se 1 tilt 1 tilt correct the velocity and temperature data recorded. PRT .2 sec .04�C .04�C Cond. .2 sec 5 p U uk CONCLUSIONS Rot. .2 sec .80 .8� The Velocity Microstructure Recorder is a Press. .2 see 8 cm 20 m useful tool for accurately describing the velocity Eng. .2 sec 1/4096 and temperature microstructure fields in the ocean. To date more than 100 launches have been made with this instrument. The single most perplexing pro- Laboratory calibrations for absolute conductivity are very difficult. blem remains the ability to manufacture reliable Oceanic calibrations are used and are dependent upon the quality of the temperature and velocity sensing probes whose standards. response characteristics are both fast enough and adequately understood. Additional work on this problem is underway. ACKNOWLEDGEMENTS My thanks to Dr. R.E. Lange for his guidance on this project. William Fleck, Helmut Moeller and Jack Lucas have given diligent laboratory and at sea support. This work was supported by the 162 Radio Antenna 4 -Refrievo/Balo/e I j S A' --Floshing Light *@j=! -~Upper End Cop a Shroud - 5I- Flotation Level Proes Velocily - Temperoature PRESURE \ SRG~ OPEPSS R pR Pressure Casing 10o Acquisition System N Figure 2 , N T ELECTRODES j -7 *-Wing Retoiner Ring 5 - Leed Bollast (approx 28 Kg's)PLE - Velocity Probe -' "Sr --Lower End Cop (with Ihru-hull connectors shown) -.--t .-Vacuum Chomber (fracture pin housing shown) a O Pressure SensorE -Protective Ring I- Disposable Bollast Weight TRA /--- F |Nose Probe Cone Ve locity -Temperoture Probes VMSR Conductivity, Platinum Temperoture Sensors Conductivit Cirit Figue- 1 Figure 3 VEL. NORTH- EAST- NORTH- VEL. NORTH- EAST- NORTH- GRAD. SOUTH TILT TILT GRAD. SOUTH TILT TILT A ._, 4 I: 2. K �, TILT K' ,. TILT IOcmec Icm//sec il !* '\ L Figure 4 Figure 5 163 MARINE INSTRUMENTATION: AN ASSESSMENT OF TECHNOLOGY VERSUS NEEDS Stephen H. Koeppen, Paul C. Etter Thomas H. Hesselbacher, Anton L. Imderbitzen MAR, Incorporated 1335 Rockville Pike Rockville, Maryland 20852 Abstract presents recommendations based on these con- clusions. This paper summarizes a recently completed study undertaken to provide NOAA's Office of2.Apoc Ocean Engineering with an up-to-date informa-2.Apoc tion base on: (1) the needs of government, Marine Environmental Data Needs academia, and industry for marine environ- mental data, and (2) the state-of-the-art of As mentioned in the Introduction, the scope oceanographic instrumentation in support ofofacringthnedfrmrnenvo- theseidenifie dat nees. Tis iformtionmental data was limited to a review and analysis was ultimately used to assess the present of data needs studies conducted over the previous state of oceanographic sensor technology. ten years and documented in report form. The reasons for limiting the scope in this manner 1. Introduction were to avoid needless duplication, to identify areas where data needs information is lacking, One of the functions of NOAA's Office of and to identify problems and pitfalls encountered Ocean Engineering is the assessment of the tech- in conducting data needs surveys. nology and programs related to ocean data acquisition systems and instrumentation. In The approach to the needs part of the study order to carry out this assessment effectively, consisted of a number of well-defined, though it is necessary for OOE to be aware of: (1) interrelated tasks. First, a literature search the needs of government, academia, and industry for data needs reports was conducted via tele- for marine- environmental data and (2) the state- phone calls to cognizant individuals in the marine of-the-art of marine sensors and instruments- community combined with a computer search on the tion. It was the objective of the present study NOAA OASIS system of the Oceanic Abstract and to provide OOE with a documented, up-to-date in- NTIS data bases. Reports which were the results formation base in these two areas and assess the of actual surveys as well as other reports docu- present state of marine instrument technology menting data needs of various users were consi- based on this information. dered. Telephone calls were attempted to oneI hundred ninety-four individuals resulting in suc- The study consisted of three parts. The cessful contact with one hundred sixty-one, The first part was a review and analysis of previous names of the individuals contacted are listed in studies of marine data needs completed over ap- Ref. I by organization. As a result of these proximately the past 10 years. No new survey was calls, over one hundred documents were received undertaken as it was felt necessary to first of which forty-six were found to be applicable summarize and document what had been done to date. to data needs. The computer search turned up an A ranking system was developed and applied to additional seven documents. A cursory analysis these results to arrive at the twenty most import- of these fifty-three reports identified thirteen taint marine parameters. The second part was a which contained quantitative information worthy questionnaire survey of instrument production and of a detailed analysis. A closer analysis development organizations to obtain an up-to-date resulted in two of these reports being ruled out summary of the state-of-the-art in instrumenta- as "representative" of marine data needs. The4 tion for measuring these twenty parameters. The eleven reports used in the detailed analysis are third part was an identification of deficiencies listed as Ref s. 2-12. in marine instrumentation with recommendations for correcting them. Based on the cursory analysis of the reports, a classification scheme was developed to class- The following two sections discuss the ap- ify marine environmental data needs as a frame- proach, the results, and the conclusions drawn work for the detailed analysis. The classifica- with regard to data needs information, instru- tion scheme was developed with the goal in mind ment technology, and deficiencies. Section 4 of ultimately comparing data requirements with 164 instrument characteristics to determine deficien- the characteristics usually associated with "data * ~~cies in the present state-of-the-art. Conse- needs". Although the operational considerations quently, the approach taken was to choose classi- are extraneous to pure "data needs", it is these fication categories which would lend themselves types of needs which are quite often expressed in * ~~to a direct comparison with instrument design a survey intended to identify instrument deti- specifications. The result was a scheme con- ciencies. * ~~tamning six major categories of characteristics: In order to prioritize data needs, a weight- * Marine Parameter ing scheme was developed with which to compute * Application an importance score for each marine parameter encountered in the reports. Four assessment cri- * User teria were chosen: application, user, volume of data needed and number of requests for data on a * Geographic Area given parameter. Within the application cate- * Data Specifications gory, the first four applications listed in a Operatinal ConsierationsTable 1 were favored over the last two because * Operatinal Consierationsbenefits from their use of ocean data are more The rationale for this choice of major categories immfvoediate. Wthin other users sicaegoy Governmen is that the first four categories should tend to wsfvrdoe h te sr ic oen deteminethe equiemens inthe ast wo cte-ment formulates the legislative mandates to which gotrmines whic reuiemntsain the charatrstic direcatly the other users respond. A discussion of the go~rielwihcnainthed cactorinstrumen deirectl detailed weighting scheme for each of the four Theselateg orieswr uthrumroent downgn. assessment criteria may be found in Ref. 1. shown in Table 1. The data specifications areSteofheAtfInruntTcolg For the state-of-the-art survey of the twenty top ranking parameters, a two-page questionnaire Table I Breakdown of Six Major Categories of Marine Environmental Data Needs Characteristics MARINE PARAMETER USER DATA SPECIFICATIONS OPERATIONAL CONSIDERATIONS a Physical * Government * Range *a Measurement Technique * Chemical * Industry * Accuracy * Biological * Academia a Resolution - Remote * Geological * Temporal - In-Situ * Meteorological - Field - Sampling Interval - Lab - Averaging Period - Experiment Duration * Physical Spatial~~~ Size * Spatial -W~~~~~Feight APPLICATIN GEOGRAHIC ARRA- Sampling Interval APPLICATION_ G E O G R A PHIC _ AREA * Electrical * Vertical -PwrRqieet * Operational a Deep Ocean *Horizontal -PwrRqieet * Forecasting a Continental - Resolution a Support a Regulatory Shelf * Monitoring * Estuary * Vertical - Deployment Platform * Baseline * Lake a Horizontal - Data Recording and * Research * River - esrmn xetProcessing a Arctic! / esrmetEtn - Auxiliary Equipment Antarctic a Vertical e Horizontal * Limitations - Temperature -Shock -Pressure - Tow Speed - Other 165 was designed. Accompanying this questionnaire * Fifty-one parties returned the postcard was a stamped, return-addressed postcard to saying they had no relevant information. obtain a quick indication of who planned on res- ponding and how many questionnaires we would * Twenty-four parties sent brochures and/ receive. A list of production and development or specification sheets from which sixty- sources was prepared for three categories: fede- nine questionnaires were filled out by ral laboratories, academic/research institutes, MAR. and industrial manufacturers. Names for the list were obtained through contacts made during the a Forty-one questionnaires were filled out data needs study, the Sea Technology Buyers Guide/ and returned by the remaining twenty-five Directory 1977-78, and personal knowledge. The parties (five questionnaires were not final mailing list consisted of six federal lab- applicable). oratories, eighteen academic/research institu- tions, and one hundred sixty-four industrial manu- * More than thirty-three additional ques- facturers for a total of one hundred eighty-eight tionnaires were promised (as per post- potential sources. card) but never received. 3. Results and Conclusions Thirty-eight organizations were contacted 3. Results and Conclusions by phone as a follow-up to the mailing. Marine Environmental Data Needs Some of these responded with question- naires or literature, others promised The eleven data needs reports analyzed con- t u int tained information on ninety-seven different The response of each organization that a question- marine parameters. The top twenty, resulting naire was sent to is given in Ref. 1. One hundred from the assessment scheme, are listed in Table 2 and five questionnaires thus formed the data base in order of descending score. The list includes for the state-of-the-art assessment. eight physical parameters, seven chemical para- meters and five meteorological parameters. NoTehogyAssmn biological or geological parameters made the topTehogyAssmn twenty (the top biological parameter, coliform, In the technology assessment phase of the ranked twenty-fifth and the top geological para- study, the documented marine environmental data meter, bottom geomorphology, ranked fortieth). needs were compared with the results of the These rankings tend to indicate either: (1)in tueaiosae-fh-rt(OAsrvy that ithi theoverll maine ommuity heseIt must be stressed that this comparison suffered two categories are not very important, or (2) from two major limitations which may have a sig- the published reports do not adequately repre- nificant impact on the credibility of the results sent the entire marine community. The latter (it was beyond the scope of the effort to assess conclusion is the most probable. the degree of credibility afforded by the study; rather, only an attempt to point out potential Table 2 shortcomings in the analysis was made). These two limitations are as follows: Top Twenty Marine Parameters The first involves the nature of the data Temperature Wave Direction needs expressed by the users. Specifically, it Salinity Wave Height ~~must be recognized that there are inherent pro- Salinth T idesHih blems in precisely specifying data requirements, Barometric Pressure Nitrate as discussed in some detail in Ref. 1, and that Wave Period Nitrite there is no assurance that these stated needs do Current Velocity Silicate not reflect the more commonly-measured, and thus Air Temperature Phosphate familiar, parameters rather than those which Water Surface Humidity ~actually may be required. The second limitation TepeateSurfae p Humdt concerns the instrument capabilities based on WidTeloiyLwMolecuatrWe ight manufacturer's specifications. Since there are Disslvd Velciy d r w olcuarbWeigh no universally accepted procedures or techniques Disolvd 2 Hyrcrosavailable for instrument calibration and inter- comparison, no attempt was made to validate manu- State-of-the-Art Survey facturer'sa stated instrument capabilities. Of theone hndredninet-one uerie sentIn spite of these limitations, it was possible out (thre soures wenred sentety-on quesinaries)sen to make a number of recommendations regarding out thre sorceswer sen tw quetionairs): marine data requirements, state-of-the-art analy- * Ninety-one parties did not respond sis procedures and instrument development. The at all. ~~~~~~~~detailed conclusions supporting these recommenda- tions may be found in Ref. 1. 166 4. Recommendations . It is recommended that a centralized informa- tion system be established in OOE listing all * Although our SOTA survey indicated that commercially available instruments and on- instrumentation to measure the parameters going development projects. The system also listed below were acceptable to meet the should include a reference file of performance stated needs, follow-up discussions indicate test data. there are problems in the area of reliability and dynamic performance. Therefore, it is * It is recommended that no additional data recommended that more performance informna- needs surveys be conducted until the SOTA tion be sought for these areas before any analysis procedure can be more firmly esta- definitive direction is taken toward new blished and made more credible. A well development. defined and comprehensive classification scheme for marine instrumentation and means - temperature - current velocity of verifying instrument performance relative - salinity - air temperature to standardized performance criteria must be - depth - water surface included in any future SOTA analysis procedure - barometric temperature to make it credible and worthwhile. pressure - wind velocity - wave period - wave direction Acknowledgements - wave height - tides - humidity The authors would like to acknowledge the * Those requirements exceeding state-of-the- cooperation and assistance of the many indi- art cpabilty fo theaboveparamtersviduals contacted during the literature search shoul be eamine in reate detal toportion of this study. In addition, we would like oudbetexaminedi getheir veailty . to thank the organizations responding to the determine their validity. instrument state-of-the-art survey for taking the * Istrumentation development should be con- tm od o sidered for in-situ or field measurements of the following parameters: Ref ertnces - dissolved - phosphate 1. Koeppen, S.H., P.C. Etter, T.H. Hesselbacher oxygen - pH and A.L. Inderbitzen, "Marine Instrumentation: - nitrate - low molecular An Assessment of Technology Versus Needs", - nitrite weight hydro- MAR TR No. 207, Final Report, NOAA Contract carbons No. 7-35281, May 1978. * . ~~~Developmental emphasis in dissolved oxygen 2. Eskite, W.H., Jr., and R. DeRyche, 1977, sensors should be directed at improving "Requirements for Marine-Related Data which reliability. may be obtained by Satellite Remote Sensing in the Period 1980-1985 and Beyond" (draft). * *~~I Developmental emphasis in sensors to measure U.S. Department of Commerce, NOAA. Unpu- the nutrients (nitrate, nitrite, silicate, blished Manuscript. phosphate) should be placed on reliable, automated in-situ capabilities. Present 3. Baugen, Dean, P., 1972. "Survey of Opera- measurement techniques involve wet-chemical tional and Research Requirements for Arctic * ~~~~methods in the laboratory, either afloat or Data". Prepared for National Data Buoy * ~~~~ashore. Center, NOAA, under Contract NAS 8-26886. University of Washington, Applied Physics * Sensors for in-situ or field measurements Laboratory, APL-UW 7225, November 1972. of pH need improvements in reliability. 4. Intergovernmental Oceanographic Commission * Further definition of hydrocarbon measure- and World Meteorological Organization, 1977. ment requirements, in general, is required "Integrated Global Ocean Station System, before any specific assessment of low General Plan and Implementation Programme, molecular weight hydrocarbon data require- 1977-1982". IOC Technical Series 16, WMO ments can be made. No. 466. World Meteorological Organization, 37 pp. * Efforts should be made to establish uniform performance criteria, perhaps through 5. National Aeronautics and Space Administration, standardized performance specifications, 1977. "Proposed NASA Contribution to the for each suite of sensors used in the mea- Climate Program", November 1977 Goddard surement of any of the twenty parameters Space Flight Center. identified in this study. Standard spec- ifications should be developed and matched 6. Rose, F.W., 1977, "Determination of Design against test results sought from sources and Operational Criteria for Offshore Faci- other than manufacturers. lities", in Satellite Applications to Marine 167 Technology, Conference Proceedings, November 15-17, 1977, New Orleans, Louisiana, AIAA. 7. U.S. Department of Commerce, NOAA, 1976. "Measurement Requirements for Marine Water Quality Data", a collection of survey responses on file at the Office of Ocean Engineering. 8. U.S. Department of Commerce, NOAA, 1974. "Ocean Instrumentation". Prepared for the Interagency Committee on Marine Science and Engineering, Federal Council for Science and Technology in response to the National Ocean Policy Study, U.S. Senate, November 1974. 9. U.S. Department of Commerce, NOAA, Data Buoy Office, 1976. "Proceedings of the Ocean Profiling Workshop", Bay St. Louis, Mississippi, June 2-4, 1976. National Technical Information Service, Springfield, Virginia, 182 pp. 10. U.S. Department of Commerce, NOAA, National Weather Service, 1977. "Shipboard Environ- mental Data Acquisition System" (draft), NWS Headquarters (W16), Silver Spring, Maryland. 11. U.S. Department of the Navy, Office of Naval Research, 1969. "National Data Program for the Marine Environment", Volume I, July 31, 1969. Prepared by System Development Cor- poration under Contract No. N00014-67-C-0559. 12. U.S. Department of the Navy, Naval Ocean- ographic Office, Marine Sciences Department, 1964. "Table of Proposed Oceanographic Measurement Requirements". Informal Manuscript Report No. 0-15-64, 16 pp. 168 A SALINITY CHAIN FOR ESTUARINE USE E.L. Lewis R.B. Sudar Frozen Sea Research Group Frozen Sea Research Group P.O. Box 6000 P.O. Box 6000 9860 West Saanich Road 9860 West Saanich Road Sidney, B.C. V8L 4B2 Sidney, B.C. V8L 4B2 TWO IDENTICAL ELECTRODE WELLS Abstract I ZTRW TRERMISTOR WELL Preliminary results are presented for sensors with good long-term stability allowing salinity determination from measurements of electrical conductivity and temperature. A chain of these sensors, multiplexed onto a single electronic driving board within PICTORIAL VIEW the recorder unit, allows time series measure- ment of salinity/temperature profiles to within 5 � 0.02� 0/00 and � 0.010 C over a period of a month. The length of the chain is not limited [I by any electrical considerations. For a long time there has been a require- ment for a moorable sensor chain to record salinity and temperature values in coastal regions. A major limitation in past attempts _ ( WELL) to fill this need has been lack of stability (1) PVC BODY (4) STAINLESS STEEL ELECTRODES in the conductivity sensor cell constant caused (2) PVC TREADEDo RING (5) PVC CONNECTOR by biological activity, dimensional changes, or (3) PVC SPACERS (6) GLASS TEUING corrosion. Another problem in chain construct- ion was that most conductivity sensors employed an inductive principle which made it difficult to place them more than a few feet from their associated electronics. Thus a chain would be assembled from a number of individual sensors complete with electronics rather than permitting particularly if the electrodes carry any current them to be multiplexed onto a single electronics due either to the finite input impedance of the board, a much cheaper construction. In contrast amplifier or because they partially shunt the sea- the four electrode method of measuring conduct- water circuit. A very high input impedance is of- ivity, such has been utilized in the Guildline fered by Dauphinee's circuit and the voltage elec- * CTD system and the Neil Brown microprofiler, trodes are recessed to stay out of the current path. typically operates at frequencies where the in- Nevertheless variations in current density over ductive effects associated with long lines sep- the current electrodes as a result of changing arating sensor and electronics no longer are metal/sea water interface conditions can alter the significant. sensed voltage difference. A development of this system used for the Guildline Bench salinometer is My modifying electrode design it is described by Dauphinee and Klein (1975). All four possible to minimize the consequences of cor- electrodes are placed in side arms away from the rosion or dimensional changes in the immediate main current path so that current flow along the path interval containing the voltage electrodes is vcntofoure electrodes sythemselescrbe In tahiee four electrode system described by Dauphinee almost independent of local changes at the elec- (1968) two "voltage" electrodes sense a potential trode surface. Another factor in cell design has difference between fixed points along a current been the need to obtain a high enough cell imped- path between two "current" electrodes, the cur- a nce in typical sea water to allow use of easily rent being varied to keep this voltage constant. obtainable electronic circuit components within The current then provides a measure of the cell their normal operating range. This has usually re- conductance. Changes in current flow path due quired considerable physical constrictions in cur- to corrosion etc. will effect the sensed voltage rent flow which in turn have created a problem in 169 CONDUCTIVITY Vrtf CELL 2 r R ell TIMING VOLTAGE V0 L\ I; i oSLi i I CHOPPER voROVREF Vo. _H RCELL O .----'-. FIG.2 Conductivity Circuit (Schematic) cell flushing as different water masses move past nating current between the two pairs of current the sensor in the chain. electrodes through the external sea water so that the potential sensed by the voltage electrodes is A design to minimize these problems is shown forced to be equal to half the reference voltage. in Figure (1) and comprises two electrode wells The current required to do this is measured across each containing three electrodes in the form of the series resistance R and the output chopper washers. The two outer current electrodes in times the sampling of this voltage so that trans- each well are short-circuited to give an approx- ients associated with the square wave current imation to a field free space between them and drive through the sea water have decayed away. A the voltage electrode is placed therein centrally. typical operational frequency for this system is This ensures no current flow through the voltage 250 Hz. Power consumption is approximately 7 electrode and the cell impedance is insensitive milliamps from a 12 volt DC (� 6 volt DC nominal) to variations in the interelectrode spacing. In single unregulated source. Readings can be taken one constructional form the main body of the cell after 5 seconds following application of power. is made of PVC plastic and the washer electrodes The DC output voltage is directly proportional to with appropriate PVC spacers are dropped into the conductance of the cell in the range 0 to 20 cylindrical holes machined in its surface. The mmhos. The constant of proportionality is normally spacer/electrode assembly is then set up with an determined by substituting standard resistors in epoxy bond while being compressed by a PVC thread- place of the cell and the repeatability of such ed bushing. Dimensions outside the electrode calibration is within + 1 pmhos or � 0.004 mmhos/cm assembly are critical and two small glass tubes equivalent for the cells used. A limitation of are inserted in each well to be coaxial with the this system is that both the power source and the electrode rings; the glass to PVC bond is made circuitry must float electrically, being "grounded" with silicon rubber. The impedance of the sea to the sea through the conductivity cell electrodes water circuit appears to have only minor depend- only. Any leakage resistance must exceed a min- ence on current electrode corrosion, a fact att- imum of 100K ohms. ributable to the symmetry of the thin disc geometry. A third well contains a thermistor that is used in It was considered that results from tests on a conventional Wheatstone bridge to determine re- one cell alone would be of questionable value as sistance and hence temperature. variations in calibration might be discovered due to individual mechanical problems, electrical mal- The electronics to drive the conductivity functions etc. In addition if readings from three sensors is contained on one 5 x 13 cm printed systems were to change simultaneously, this prob- circuit board and has a component cost of about ably would indicate a variation in the environment $50.00. It is shown schematically in Figure (2). rather than in the cells themselves. Thus three The circuit is a variation of that described by lots of electronics were connected to three cells Dauphinee (1968). A timing generator produces and tests made by interchanging them so that each switching pulses for a reference voltage and out- cell would read the same with any set of electron- put sampling. The reference voltage is applied ics and that each set of electronics would produce to an operational amplifier which drives an alter- the same value for a standard resistor. Once this 170 100' ) 2 1 0 * 60" 2- -- - - - - X -2-' � , "" W *L 2 --Z-- u40" ~~~~~~~~~2 2 - w-'( ~~~~~~~~~0 >" 2" U a.n - - 0 . - 3 , , 100- <S . . LJ 80- 2 ZLfo 0~~~~~~~~~~ 0 02 4 6 8 11214 14~~~~~~~~c (O 40-- TIME (days) 0 20-- FIG.3 Difference between salinity readings from . conductivity - temperature systems and bath W 0 b_ N salinity determined by bottle samples. a: 100- WI had been done the three cells were placed in an 1. 3 insulated tank of sea water which was being con- z tinually mixed by a magnetic stirrer. Samples 60' were withdrawn from this bath at weekly intervals for analysis in a bench salinometer when it was 40- found that the salinity was changing very slowly due to evaporation. At the beginning and end of every test period a standard resistor was used as a check of the entire measurement and recording 8 -6 -42 0 2 4 6 system. Figure (3) shows results from a typical SALINITY -AVERAGE SALINITY(o/,Qx102) calibration test in the laboratory. The sea water in the laboratory tank was The sa waer i thelaboator tan wasFIG.4 H~istograms of the salinity of one cell minus sterilized to prevent biological growth . When the ave rag e salinity of one cell all the cells were placed in the sea they were fitted with a light shield and a pad containing an anti- measuring the same water mass. The three foulant containing tributyl tin. Previous lab- cases shown correspond to; 1. poor flushing due to fouling of the entrance port, 2. im- oratory experiments had shown that this compound proved fl ing of th e entranc e proved flushing due to widened entrance in sea water does not contribute to conductivity ports, 3. perfect flushing in a laboratory values at the level of interest. This confirms the results of Wff et al (1975).stirred bath. All readings were taken at the results of Waff et al (1975). 30 minute intervals. Figure (4) section 1, shows a histogram of the salinity values obtained from one of the three cells after four weeks of an eight week period of which contained the measured salinity of bottle continual immersion in the sea. A wide spread of samples from the tank. values is noticeable. At this stage a diver went to inspect the installation which was in 6 meters test s a ches near shore of water near the pier off shore from this in- tests a chain was manufactured consisting of five stitute. Although the cells were not taken from the water, his replacement of the antifoulant pads FFor test purposes the chain was wound into a circle with a more permeable version was sufficient to of approximately Ia diameter so that the five cells protruded radially inwards and should be in the improve cell flushing which was the cause of the wa tr mass. As a suld an in the spread in readings. The results just after his visit are shown in section (2) of Figure (4). Sub- entire assembly dropped from the buoy and fell to sequently a storm churned up the sediment surround- the seabed at a depth of 186 meters. The initial ing the pier and the values of salinity varied results had appeared good but, upon recovery, it widely from one cell to the other. At this stage was noted that a shift in calibration of approx- the instrument was recovered, still keeping the imately 0.08 V/0 salinity had occured in four of cells beneath sea water in a bucket, and was trans- the five cells. This shift appears to be per- ported back to the laboratory. After a simple manent and is probably associated with small di- flushing with fresh water all cells and gave values mensional changes, perhaps movement of the glass in a range of � 0.02 0/00, Figure 4 (section 3) tube in the rubber as a result of the pressure. Tests were started on the effects of pressure on 171 the cell constant using a pressure vessel but at present are inconclusive in that at 300 psi an in- stability was introduced in the cell almost cer- tainly the result of water penetration between some components of the assembly. This cell eventually recovered, presumably after the injected water emerged. At the time of writing, June 1978, the chain is back in the sea and results taken over a period of one week are encouraging. Work is proceeding making cells of the same basic design using different constructional techniques with a view to eliminating the pressure sensitivity. Data from the chain will be collected over longest periods. Ropefully results can be reported during the verbal presentation of these studies at the conference. References 1. Dauphinee T.M. "Some applications of D.C. and square wave A.C. techniques to undersea measurements". Paper #68-635 Instrument Society of America Annual Conference 1968. 2. Dauphinee T.M. and H.P. Klein "A new automated laboratory Salinometer" Sea Technology 16 1975 pp 23-25. 3. Waff W.B., Wolfgram P.A., and J.P. Rohling "Development of a water quality instrumentation package for long term operation from buoys and other unattended marine platforms. Paper #OT C 2299 Seventh Annual Offshore Technology Corn- ference Dallas Texas 1975 pp 529 - 533. 172 DEVELOPMENT OF A SMALL IN-SITU CONDUCTIVITY INSTRUMENT A.M. Pederson M.C. Gregg Applied Physics Laboratory Applied Physics Laboratory and University of Washington and Department of Oceanography Sea-Bird Electronics University of Washington 4735 W. Mercer Way Seattle, Washington 98195 Mercer Island, Washipgton 98040 Abstract The instrument was originally developed by the Applied Physics Laboratory, University of Washing- The initial development and testing of a small tont and is now commercially available from Sea- in-situ conductivity instrument was reported at Bird Electronics (under license from the University OCEAN 73.1 Since then, extensive operational ex- of Washington) as Model SBE-4 (Figure 1). perience has been obtained with a number of these instruments on moorings and on free and tethered platforms that move vertically and horizontally. This experience is reported here, as well as de- sign improvements, and calibration and maintenance procedures. Also discussed are cell flushing characteristics and the problems of matching con- ductivity measurements spatially and temporally to temperature measurements. 1. Introduction For details of instrument theory of operation and construction the reader is referred to Refs. 1 and 2. Briefly, the instrument's sensing element is a three-electrode (platinum) flow-through type cell used as a two-terminal device by holding the outer electrodes at the same potential. The cell resistance R2 between the center electrode and the Figure 1. Conductivity meter, SBE Model 4-01. outer pair is used as one of the bridge components in a patented high quality Wien bridge type oscil- 2. Platinization lator, whose frequency of oscillation is given by the familiar relationship A spongy coating of platinum black is deposited on the cell electrodes by electrolysis, using a so- 2 1 W -= (1) lution of platinum chloride and lead acetate in 1R2C1C2 hydrochloric acid. The process is called platini- zation, and results in greatly reduced electrode RI, C1 and C2 are fixed stable bridge components, impedance due to the increased effective surface and R2 = K/a; a is the electrical conductivity of area. the fluid within the cell in Siemens per meter,* and K is the cell constant (determined by cell ge- The cell used in the instrument is purchased ometry). Thus, from Beckman Instruments with platinization of about 7 coulombs per cm2 of electrode area. Ac- a = R1K C1C2 w= f.2) cording to Ref. 3, this should be sufficient to ensure low electrode resistance R., and high elec- trode capacitance Cs (hence minimum Zs). However, Nominal parameter values are R1 = 450 0, K = 2000, Beckman uses a much higher current density while and C1 = C2 = 0.033 pF, resulting in k1 4 x 10-8. platinizing than is recommended in ASTM Standard *Multiply S/m by 10 to obtain mmho/cm. tPartially supported by the Office of Naval Re- search and the National Science Foundation. 173 D1125-64 (reapproved 1971). We have experimented at about -1% per �C and +1% per �C, respectively. with platinizing cells at the lower ASTM current The magnitudes of Rs and Cs were found to be depen- density (1 to 4 mA per cm2 of electrode area), and dent on the electrode material, as well as on the have found no significant difference in performance chemical composition of the electrolyte. None of the compared with cells platinized at higher currents. work reported in Refs. 3 and 4 used salt water as We do, however, routinely add another 3 coulombs/cm2 the electrolyte--hence the results cannot be applied to all new cells at the ASTM current density, re- directly to electrodes used for the measurement of versing the polarity on the electrodes every 10 to ocean conductivity. There is some evidence that Rs 20 s to minimize hydrogen formation. is an order of magnitude larger5 and that Cs is some- what smaller for platinum electrodes in salt water 3. Static Calibration and Curve Fitting than in potassium chloride (potassium chloride was used for much of the work in Refs. 3 and 4). For All SBE-4 conductivity instruments are cali- well-platinized electrodes of the size used in SBE-4 brated at the Northwest Regional Calibration (100 mm2), it is estimated that Rs is a few tenths Center (NRCC), operated under the direction of the of an ohm and Cs is a few hundred microfarads. National Oceanographic Instrumentation Center (NOIC), a NOAA organization. Methods of calibra- tion used by NRCC are described in Ref. 2. The +0.025-- absolute accuracy of the calibration bath is better than 0.01%, (corresponding to = 0.001 S/m), with long term stability and repeatability probably on the order of �0.003%. Calibrations are done over a temperature range of 0 to 300C at two different O.015-- salinities to reveal any temperature effects of the instruments. (a) 0.010-- In the following discussion we will be speak- 0 35.0%o ing of residuals (Ao). 0.005-- A 247%o \ac = 0, - a ,(3) SBE-4, S/N 10 C E �--' MAY'77, CALIBRATION where 0C is the measured conductivity of the cali- bration bath and a is the conductivity indicated by the SBE-4 instrument. If the simple theoretical b (b relationship a = klf2 is used, we find AO as shown in Figure 2a. Note the relatively poor fit, parti- cularly at higher conductivities. This was the -0.005 method reported in Ref. 1. It can be shown that by the fact that the RC oscillator deviates slightly from the theoretical relationship w2 = 1/(R1R2CiC2) ( due to parasitic capacitances in the circuitry. If �-t the oscillator conformed perfectly to the theoreti- cal relationship, the instrument residuals would be -0o00 as shown in Figure 2b. We see that there is still 2 3 4 5 a concave upward curvature, as well as an apparent temperature effect (at 4.3 S/m there is about 15�C difference in bath temperature between the 24.7L Figure 2. Conductivity residuals for (a) a = kf2;- and 35.0%,baths). Some of the temperature effect (b) a = k2]2, oscillator frequency ad- could be due to temperature sensitivity of the RC justed to theoretical; and (c) a = ad- oscillator (< �1 Hz over 0�C to 200C, corresponding (a - bfl)2 + cT2 to �0.0008 S/m at 0 = 4.3 S/m). However, this is not large enough to account for the observed effect, and it could be in the wrong direction (all instru- ments show similar characteristics, with the low If we now assume that Rs and Cs are the same salinity residuals positive with respect to the for each of the three electrodes, we can compute the high salinity residuals). frequency of oscillation to be In Ref. 4, Jones and Christian showed that R1C C electrode polarization impedance is essentially a 1 - 1.5 K + 1.5 C small resistance (Rs) in series with a large capac- (- + 1.5 R )C (4) itance (Cs), that Rs is inversely proportional to 2 Cs K fl/2, Cs is approximately inversely proportional to R. (+ 1.5 R )C2 f0.4,* and that Rs and Cs are temperature dependent as where C1 = C2 = C. Equation 4 can be solved for a *Work reported in Ref. 3 showed that Cs is very to yield nearly independent of frequency for well- platinized electrodes. 174 CY Kw2RC2 (1 + 3R2 R. C2) A+ 1S2 R.C2 requires that the frequency response functions of 1I(I R1C slR 1 I +.SW2RC2I I(5) the sensors be known. Pressure sensors essentially 2 2) [RI C 2 2 respond instantaneously to pressure changes, and (1 + 6Rs w R1C ) C - RRs (1 - .5 cause no difficulties. The response of temperature 1%s����i�~�'~I":' �2�'�'ll sensors to a change in ambient temperature is lim- ited by the time required for diffusion through Since Rsw2R1C2 << 1, it can be shown that the thermal boundary layer around the probe, the 2 insulating coating or protective sheath, and through 6oC 1.5 = _ -1.5 ar (6) the temperature sensitive material. The response of Rs * R2 *R K conductivity sensors is generally limited by the flushing time of the cell, although the thermal mass and of the cell body can also cause an effective lag due to transfer of heat between the cell body and the 1 .R CaS 2 sample volume of water within the cell. In some da 1 2 (7) applications on slowly moving vehicles, a pump has Cs KC been used with the SBE-4 instrument to give a s flushing time of 130 ms. Laboratory measurements and computations indicate that for speeds through It can also be shown that the effects of Rs and Cs, the water > 1 m/s the free-flow flushing time (in considering their temperature and frequency depen- seconds) is 0.25/speed, where speed is in m/s. dence, are sufficient to approximately explain the There is an additional delay time of 0.23/speed shape of Figure 2b. due to the cell volume that precedes the first electrode. The observed effects of electrode impedance and oscillator deviations from theoretical are rel- For probes moving in linear gradients, the atively small, and can be eliminated from the static effect of the lag of the probes is to produce a calibration data by an equation of the form: constant offset in the measurements. For up and down casts of a CTD,* or a cycling towed body, the f 2 offsets are in opposite directions, and produce a = 2+ cT , (8) separation of the two temperature-salinity (TS) (a - bf)2 diagrams. where T is in �C and the constants a, b, c, and n Due to the existence of finestructure in the are obtained by solving to fit the equation to the profiles, which is defined as changes in the gradi- calibration data. A program for the HP97 has been ents over scales from about 0.5 m to tens of meters, written to do this--at least four calibration points simple lag corrections are not adequate to recon- are required with one of the points at a different struct the true profile from the data. This is il- salinity than the other three. The relationship of lustrated by considering a linear gradient with a the empirically derived equation 8 to equation 2 can step change superimposed. The difference in the best be seen by using the binomial expansion on phase and amplitude response of the temperature and equation 8 to obtain conductivity sensors to this transient produces a salinity spike, which is not removed by a lag cor- f2 [ 2b fn + 3 n2 1 rection. To correct for transients such as this, a - 1 + 2a + f . . + cT . (9) digital filters based on accurate determinations a ( of the frequency response functions of both tem- perature and conductivity sensors must be used to All terms are small with respect to f2/a2 (the term deconvolve the data. To date, this has not been expected from the theoretical relationship depicted done, although the problem is being worked on. The by equation 2). The expansion terms beyond those principal difficulty has been obtaining the trans- listed are insignificant. The fn and f2n terms cor- fer functions of the sensors. The values published rect fer the nonperfect performance of the oscilla- by many manufacturers are not adequate since they tor and the additional curvature, while the cT2 term were obtained by plunging the probes from air into brings the curves for the two different salinities water and thus do not simulate the boundary layers into alignment. Figure 2c shows the quality of fit present on probes as used in the ocean. typically obtained by this method. The fit is clearly within the limits of accuracy and repeata- The problem of salinity spiking can be illus- bility of the NRCC facility. trated with a section of MSRt data from the MILE** cruise. Figure 3 shows the vertical profile of 4. Dynamic Calibrations measured temperature, with salinity and density com- puted from conductivity, temperature and pressure The above procedure produces an equation that adequately relates instrument output frequency to conductivity under static conditions; i.e., when the instrument has come to equilibrium with the * Conductivity-temperature-depth instrument. environmental parameters of interest (conductivity, temperature, and pressure). In order to be of practical use, it must be possible to compute sali- slowed by helicopter-like rotors to a descent nity and density from a conductivity-temperature- rate of about 80 mm/s. pressure data set under nonstatic conditions. This **Mixed Layer Experiment 175 3.00 5.00 7.00 9.00 1100 13.00 the spiking problem is much worse. It is interest- 0.� 8 ing to note that data taken during MILE with other ~MS~~R el8~ ~CTD systems also exhibit spiking of several tenths -25.0 of to, even though some of the conductivity cells are much smaller than the SBE-4 cell. -50.0 5. Cell Maintenance -75.0 Since the SBE conductivity instrument output -100-0 frequency is determined by the cell resistance, it ~~T <Y5 AD ~~~is important that the electrodes be kept clean. ~-i1~~~25o~0t~ ~Cell electrodes contaminated with oil, biological LI -20growth, or other foreign material will cause the CLPY,~~ ~ ~ ~ ~ > ~instrument to indicate a water conductivity lower W- -150.0 than it really is. As with other types of conduc- L -175.0 tivity cells (including inductively coupled types), ~~~~~~-175- ~0 2 Xfouling within the cell can reduce the effective cross-section area, resulting in the same effect. -200.0 If the cell is allowed to dry between usage, -225-0 lljsalt crystals may form on (and in) the platinized surface of the electrodes. The next time the in- -250-0 l strument is used, it will take some time before these crystals are dissolved--in the meantime, 2750 \ \ ~~~electrode impedances may be affected. We therefore ~~~~~~-275-~0 ~keep the cells filled with distilled water between use; this is easily done with a length of Tygon 322 50 33.00 33.50 34.00 tubing looped between the ends of the cell. An additional important benefit of keeping the cell 24.50 25.00 25050 26E00 26.50 27-.00 ends closed is that airborne contamination (of which there is an abundance on most research yes- Figure 3. Temperature, salinity and density pro- sels) can not enter the cell. For long-term file from MSR drop 81.. file from MSR drop 81. storage, we have recently begun using a special storing solution (MT-A-10*) designed to prevent the for water with only a moderate amount of high wave- growth of micro-organisms, a possible source of number activity (a well-mixed surface layer with a cell contamination smooth transition into the thermocline). Figure 4 is an expanded section of Figure 3, where a salin- For applications where it is not convenient to ity spike can now be seen in an area of sharp tem- keep the cells filled with water, such as work in perature gradient. For stormy sea conditions with the polar regions, we recommend flushing the cells violent overturning at the base of the mixed layer, with distilled water after each use. 5.00 7.00 9.00 11: 00 1300 6. Cell Cleaning -25-0 M -D ,a- 6. Cell Cleaning M PR B1 When checks against water samples or other evi- dence indicates that a cell needs cleaning, we fill -30.0 * SPIKING the cell with a 4% solution of MT-6* and let soak S D ~T /for 30 min, then flush with warm fresh water for S T I1 min. ~~~~~-35-~~~~~0 lIf severe fouling has occurred (visible depos- U) [35-0its on the inside of the cell walls), we fill the LJ cell with a 4% solution of MT-6 and place the en- LI lltire instrument in an ultrasonic bath for 15 min, -E 40.0 \/then flush for 5 min with warm fresh water. We tried a number of different cleaning tech- * niques on several conductivity instruments in De- -~~45- ~ ~~~ ~~~0 )cember 1977, using a calibration bath to carefully determine the effects of each cleaning process on the instruments. Unfortunately, we had no badly -50.0 32.50 32-70 32.90 33.10 33-30 33.50 24 00 24-50 25-00 25-50 26 00 Figure 4. Expanded section of data from MSR drop *Available from General Chemical Corporation, 4710 81. S Street N.E., Kenilworth, Washington, D.C. 20027. 176 contaminated cells to work with, so our test re- rms over a 2.5-Hz bandwidth. Since this is con- suits were more an indication of whether the clean- siderably less than the observed noise level, Fig- ing process itself would adversely affect the in- ure 5 represents the inherent noise of the instru- strument calibration. The tests indicated that ment. either MT-6 or a non-ionic detergent such as Triton X-100 is a satisfactory cleaning agent, the residual How well salinity can be resolved is of ulti- chemicals being easily flushed away with a small mate interest to the oceanographer. The system amount of water. In fact, little or no adverse noise in the conductivity, temperature, and pres- effects were observed when there was no flushing sure channels limits the sensitivity of the salin- after cleaning other than that due to normal bath ity data, and that of other derived parameters such stirring. as N2 (the stability frequency) and the gradient Richardson number.7 The variance of the salinity 7. Resolution and Noise Level noise, Vars, is given by To obtain maximum resolution, some form of ,Ds\2 2 aS\2 VS\2 period counting should be used. In a typical appli- Vas = \T V arT + Var + VarP . 10) cation, the time interval generated by 1408 periods of the SEE-4 signal is measured by a S-mHz clock.. . ..... of the SBE-4 signal is measured by a 5-mliz clock. The partial derivatives are the sensitivities of A 5-Hz sample rate is used, since the count is com- salinity to temperature (6S/6T 0.8), con ductivit y pleted in less than 0.2 s. The resolution (least 1) t peure (S/6 . c1t count) varies from 5.6 x 10-6 S/m at a = 2.5 S/m (6S/Sa = 10), and pressure (6S/6P = 4.2 x 10-8). cut o 1.6 x 10-5 S/m at a = 5.0 S/m. The variance of conductivity, Varo, is just the square of the rms noise computed from Figure 5 In Figure 5 the noise level of the instrument (Var� = (1.1 x 10-5 S/m)2). The variance of tem- is shown by the spectrum of a quiet section of MSR perature, VarT, is derived from Figure 6, a plot data taken in a fjord The spectral level of the of the spectral level of the companion SBE temper- data taken in a fjord. The spectral level of the auesno rmtesm S rpfo hc noise is about 5 x 10-11 (S/m)2/Hz, corresponding ature sensor from the same MSR drop from which Figure 5 was obtained. The temperature spectral to an rms value of 1.1 x 10-5 S/m over a 2.5-Hz noise level is = 6 x 10-9 eC2/Hz or 1.2 x e - Ic bandwidth. For period counting, the theoretical ms over a 2.5-Hz bandwidth. Thus VarT = (1.2 x ~~~~~~~~~~~~~~rsoe .-zbandwidth. ThuoVrT=1.x spectral level due to the least count is given by 0 C)2. Similar data are used to compute Vap (least count)2/6 (see Ref. 6). Because the least 27 ascals2 When these values are used in count value for these data is 8.6 x 10-6 S/m, the atn e e in corresponding spectral level due to quantization equation 10, we find only is 1.2 x 10-11 (S/m)2/Hz, or 5.5 x 10-6 S/m 9 -8 -12 Var = 9.2 x 10 + 1.2 x 10 + 1.3 x 10 S T a P (11) -~~~~~~~~~~~~~~~~~- Io 5 = 2.12 x 10-8 - -6~~ \ ~~~~The rms noise level in salinity is therefore (2.12 x lo0 -6_\ SBE - ~4-0l~ 10-8)1/2, or 1.5 x 10-4%o. S/N 18 I-5: i7 \10 N - SBE-3-01 IN~~~ -8 \' l~~~~~~~Io-6 S/N 425 a 10 8 ~~~~~~~~~~~~~i-9 10 - 9N 1o 101.0'''"'l''''"'l'''l'"'l '' .... 10-9 , , , i,,,,i',,-,,,,8',i,,,,i,',''I"" 1-3 -o2 10 0 10 101 10-3 l~ 10 10� 10 10-3 10-2 10-i 10- 10 -3 2 -I i0� i0I FREQUENCY (HZ) FREQUENCY (HZ) Figure 5. Conductivity spectra from MSR drop 50. Figure 6. Temperature spectra from MSR drop 50. 177 8. Circuit Modifications a subsequent drift of +0.0005 S/m during 5 months of non-use. Again, no cleaning was performed and the Our first two conductivity instruments were cell was kept filled with distilled water between built for use with SPURV* in 1973,8 and since that use. time about 30 more have been built for a number of different organizations in a variety of applica- Four SBE-4 instruments were used on a towed CTD tions. Initially, we had some problems with long- system in tests in Puget Sound and during the MILE term stability due to galvanic currents flowing be- cruise, again without cleaning. For reliability, tween the instrument case and the cell electrodes, the towed system had two pairs of temperature and and because of improper cell cleaning procedures. conductivity instruments. Calibrations for the two In the spring of 1977 we devised a method of iso- conductivity instruments used in the Puget Sound lating the electrodes from the rest of the circuitry tests showed drifts in Ao of -0.0005 and +0.0005 S/m with large capacitors, thus eliminating the galvanic during the 2-1/2 month period. Another calibration current problem. The modification causes a small 5 months later showed additional drifts of -0.0005 shift in the instrument calibration constants, pri- and +0.001 S/m. Before and after calibrations for marily because of a l-megohm resistor in parallel the two conductivity instruments used during about with the cell resistance now required to bias the 140 h of towing during MILE showed drifts in au of input stage FET. However, the effects on instru- +0.001 and -0.01 S/m. The excessive drift in the ment calibration due to temperature coefficients or one instrument apparently occurred catastrophically aging of the isolation capacitors are negligible. prior to cruise operations, since the offset between the two instruments was relatively constant through- Galvanic currents through the electrodes are out the cruise. The drift was caused by an elec- now believed to be the cause of the -0.002 S/m drift tronic failure rather than by electrode contamina- observed in the 87-h bath test reported in Ref. 1. tion. A similar bath test performed in December 1977 after dc isolation of the electrodes showed almost no de- Five SBE-4 instruments were used during the tectable drift during an 89-h test of two instru- summer of 1977 on a 9-m vertical array which was ments (+0.0001 and -0.0001 S/m). towed in near-surface estuarian waters of Puget Sound.9 Calibrations by NRCC showed drifts in au The galvanic current problem first arose when of less than �0.002 S/m for all instruments during it was found necessary to hard-wire the instrument the 6-month period between calibrations. The cells case to electronic ground in certain RF-noisy en- were kept filled with distilled water during periods vironments to prevent spurious modulation of the of non-use. output frequency. Recently, we have added RF chokes inside the instrument in the power, ground, and In yet another application, an SBE-4 instru- signal lines, as well as additional filtering on the ment was used in Puget Sound tests and during MILE internal regulated supply voltage. These changes on the MSR system. Calibrations at NRCC before and have made the circuit immune to stray RF fields, and after these experiments showed a drift in Aa of it may no longer be necessary to hard-wire the cases about -0.002 S/m during the 4-month period. A sin- to electronic ground in noisy environments. How- gle point calibration 3 months later showed an addi- ever, to be on the safe side, we still recommend tional drift of -0.002 S/m. case grounding unless there are compelling reasons for having the whole instrument floating electri- Note that, due to the form of equation 3, only cally. positive changes in au can be attributed to elec- trode contamination. Since a number of drift trends 9. User Experience in no were negative, we tend to believe that instru- ment drift is dominated by electronic drift rather Our best results have been obtained during the than cell contamination. There are exceptions, of past year, after dc isolation of the electrodes. course. Several years ago an SBE-4 instrument was We had a conductivity meter on each of two moorings used on a near-surface mooring for 30 days off the at 1100 m depth for 25 and 8 days, respectively, mouth of the Columbia River. After that period of during MATE.t Calibrations at NRCC before and after time, visible organic growth had contaminated the the MATE cruise showed drifts in ao of +0.001 and interior of the cell, the instrument output was -0.0005 S/m for the 2-1/2 month period between cali- noisy, and Au had shifted by about +0.03 S/m. brations. When a single point calibration check was made 5 months later, additional drifts of +0.0005 10. Summary and -0.0005 S/m were observed. These results were obtained with no cleaning, but with the cells kept Since first reported,l extensive operational filled with distilled water between use. experience has been obtained with about 30 of the instruments on moorings and on free and tethered A third SBE-4 instrument was used on SPURV platforms moving both vertically and horizontally during about 20 h of operation at depths to 1500 m through the water. This experience has shown typi- during MATE. Before and after calibrations for this cal long-term stabilities of �0.0003 S/m per month. instrument showed a drift in Au of -0.0007 S/m, with While fouling remains a potential problem, evidence suggests that it is not a serious problem except in near-surface moorings or in heavily polluted waters. *Self-Propelled Underwater Research Vehicle. tMid-ocean Acoustic Transmission Experiment. 178 Dynamic calibration, including flushing char- 3. G. Jones and D.M. Bollinger, "The measurement acteristics, is an area that needs further work to of the conductance of electrolytes," Journal reduce salinity "spiking." This is a problem com- of the American Chemical Society, Vol. 57, pp. mon to all conductivity sensors. In this instru- 280-284, February 1935. ment, flushing time can be controlled with an ex- ternal pump--not a viable technique with some types 4. G. Jones and S.M. Christian, "The measurement of cells. of the conductance of electrolytes," Journal of the American Chemical Society, Vol. 57, pp. Since the instrument output is an FM signal, 272-280, February 1935. resolution is dependent on the counting method used. In a typical application, resolution is 1 x 5. R.G. Paquette, "A modification of the Wenner- 10-S S/m at five samples per second, and the spec- Smith-Soule salinity bridge for the determina- tral level of instrument noise is 5 x 10-11 (S/m)2/ tion of salinity in sea water," University of Hz, corresponding to an rms noise level of 1 x 10-5 Washington Oceanography Department, Technical S/m over a 2.5-Hz bandwidth. Report No. 61, Ref. 58-14, August 1958. Acknowledgments 6. J.D. Irish and M.D. Levine, "Digitizing error from period and frequency counting techniques," The preparation of this paper was partially Deep Sea Research, Vol. 25, pp. 211-219. supported by the Office of Naval Research (by the oceanography section of ONR through the Oceanic 7. M.C. Gregg, "The effect of bias errors and sys- Microstructure Program, and by the acoustic sec- tem noise on parameters computed from T, C, P tion of ONR through the MATE program). We are in- and V profilers," submitted to Journal of debted to Larry Murdock of NRCC for consistently Physical Oceanography. good calibrations, to Eric Aagaard (APL-UW) for in- strument mechanical design and flushing character- 8. H.R. Widditsch, "SPURV--The First Decade," istic computations, and to Bob Francois (APL-UW) Applied Physics Laboratory, University of for measurements of cell flushing characteristics. Washington, APL-UW 7215, October 1973. References 9. G.B. Gardner and J.D. Smith, "Turbulent mixing in a salt wedge estuary," Hydrodynamics of 1. A.M. Pederson, "A small in-situ conductivity Estuaries and Fjords, Elsevier Scientific instrument," Proceedings of OCEAN 73, IEEE Publishing Co., Amsterdam, 1978. International Conference on Engineering in the Ocean, pp. 68-75, September 1973. 2. M.C. Gregg and A.M. Pederson, "High resolution salinity measurements," Instruments and Methods in Air-Sea Interaction (Text from NATO Advanced Study Institute), Ustaoset, Norway, April 1978 (in press). 179 EVALUATION OF METAL-FILM TEMPERATURE AND VELOCITY SENSORS AND THE STABILITY OF A SELF-PROPELLED RESEARCH VEHICLE FOR MAKING MEASUREMENTS OF OCEAN TURBULENCE James D. Irish and Wayne E. Nodland Applied Physics Laboratory University of Washington4 1013 N.E. 40th, Seattle, WA 98105 Abstract Introduction Metal-film temperature and velocity sensors are In experimental oceanographic research, many being added to the Applied Physics Laboratory's different types of measurements and techniques are Self-Propelled Underwater Research Vehicle (SPURV) required to sample properly the physical processes to extend its measurement capability to turbulence occurring in the ocean. Measurements are required scales. Before making field measurements, the sen- as a function of both time and space. The Applied sors are being evaluated in the laboratory and the Physics Laboratory (APL) of the University of vibration level of SPURV is being measured to eval- Washington has developed and utilized unmanned, uate its stability as a platform from which to make free-swimming submersible vehicles to make horizon- turbulence measurements. The vibration is measured tal spatial measurements. These vehicles have been along the three major vehicle axes (vertical, hori- given the acronym SPURV for Self-Propelled Under- zontal, and longitudinal) using accelerometers. water Research Vehicle. Development began in 1958; Field measurements were made with the accelerometers since that time, the navigation and data-collection and the records integrated to obtain velocity spec- systems in the prototype and later vehicles have tra which are compared with typical measurements of been modified to meet different needs. These needs ocean turbulence. The vehicle's motion is below include studies of thermal structure (McKean and the turbulence spectra when ocean turbulence dissi- Ewart, 1974; Ewart, 1976a) and dye diffusion (Ewart pation is above 10-4 ergs/cm3/s, except at a multi- and Bendiner, 1974a) and support of acoustic meas- ple of the motor rotation frequency where the urements (Ewart, 1976b; Ewart, Irish and Olson, vibration appears as a strong peak. 1977). Electronics for remote operation of the metal- The two current versions of SPURV are desig- film temperature and velocity sensors were devel- nated SPURV I and SPURV II. Further information oped. To evaluate the sensors and electronics, on the SPURV I vehicles can be found in a report test facilities were built for both static and by Widditsch (1973). The lower-frequency tempera- dynamic calibrations. A temperature-controlled ture and conductivity sensors carried on the SPURVs bath is used for static temperature calibrations. were developed at APL and are described by Pederson A submerged jet mounted within the temperature (1969 and 1973). A fluorometer for diffusion meas- bath is used for static velocity calibrations. A urements was also developed at APL for SPURV, and thermal plume tank is used for dynamic temperature its use is described by Ewart and Bendiner (1974b). response calibrations of both the temperature and Ile are currently increasing the measurement capa- velocity sensors. The dynamic velocity calibration bility of SPURV II to include water temperature is made by oscillating the sensor within the static and water velocity sensors with a sufficiently jet with a shaker pot. high frequency response to resolve turbulence and dissipation scales. While developing the new sen- Results of these tests show that the sensor sor package, we are also investigating the stabil- frequency response can be adequately determined, ity of SPURV II as a platform for making the and that it extends to high enough frequencies to high-frequency measurements, studying the behavior resolve turbulence fluctuations. The temperature of the individual sensors, and comparing the re- sensor's noise and drift are low enough that good sults with expected signal levels in the ocean. temperature measurements to turbulence scales can be made. However, poor stability and high noise in Vehicle Motion Compared with the velocity sensor make the measurements unreli- Expected Turbulence Levels able, and more work is being done in an attempt to improve the sensor. Knowledge of the vehicle's motion (vibration) is critical for interpreting the high-frequency velocity and temperature measurements. It is impossible to differentiate a true high-frequency velocity signal from a high-frequency vibration 180 that moves the sensor. Also, vertical motion of 100 the temperature sensor in a temperature gradient SPURV II will produce a false temperature signal. ACCELERATION To measure the vehicle motion, a three-axis 977-PUGET SOUND complement of force-balanced accelerometers N 1O -- RUN 01-3.5 M/S (Sundstrand, Model QA-1100-T2) was mounted as far forward as possible in the new SPURV II configura- tion (see Figure 1). The accelerometers were in- CROSS dividually calibrated on a shaker table at the 2 University of Washington. Also a self-calibration e 10 - VERTICAL was done by injecting a signal into the accelerom- E ....--LONGITUDINAL eters' self-test input according to the manufac- turer's specifications. The results of the two O methods agreed. RECOVERY RADIO ANTENNA HYDROPHONE FOR LINK AND FLASHING LIGHT WHTH MOTHER SHIP l PROPULSION MOTOR 0 AND CONTROL \\ SURFACE DRIVERS \ CONTROL ELECTRONICS BATTERIESI/ DATA RECORDING ELECTRONICS ETNER SECTION HOUSING HIGH FREQUENCY TAPE RECORDER PRESSURE CASES FOR ACCELEROMETER AND HIGH FREQUENC Y SENSORS 0 02l' ( I AR PACK3AG METAL-FILM TEMPERATURE AND VELOCITY SENSORS FREQ (Hz) Figure 1. The SPURV II vehicle for turbulence measurement. Figure 2. Acceleration spectra from three-axis accelerometer. Accelerometer measurements were made during the summer of 1977 to evaluate the new vehicle Another series of measurements was made with configuration and to compare the results with SPURV II traveling at four different speeds between expected turbulence signals. The signals from 3.5 m/s and 1.6 m/s. The acceleration spectra were the accelerometers were passed through a bandpass integrated to obtain velocity spectra for compari- filter (8 Hz to 1400 Hz) before recording. All son with measurements of oceanic turbulence. The resulting spectra were corrected for the sensor results for the vertical axis are plotted for all and filter responses. speeds in Figure 3. The overall vibration de- creases as vehicle speed decreases. The peaks Figure 2 shows three acceleration spectra with labeled 1, 2, 3, and 4 are due to forced oscilla- SPURV II traveling at 3.5 m/s. The longitudinal tions. The peaks labeled 2 occur at the basic vibration was less than the vertical motion or motor rotational frequency of 41 Hz at the highest horizontal cross-axis motion. Several single and speed and decrease to 22 Hz at the lowest speed. multiple peaks are seen in the record which are The peaks labeled I are at half the motor fre- due to natural vehicle vibration modes and to quency, decreasing from 21 to 11 Hz. The peaks forced oscillations induced by the propulsion labeled 4 (not completely resolved in frequency) motor. The largest amplitude peak occurs at occur at the commutator slot frequency, and the 547 Hz, and is related to the motor speed. The peaks at 3 are at one-half the slot frequency. motor commutator slot frequency is 28 times the The peaks at 3 are the major contamination in motor frequency, or about 1130 Hz, which is double our turbulence measurements. The peaks labeled the frequency at which this peak occurs. It is A through F do not change frequency with motor not clear why a large vibrational peak should speed, and are natural frequencies in the vehicle appear at a sub-harmonic of the commutator slot itself. frequency. It probably is not an electrically induced signal because it does not immediately Since the total vibrational energy decreases disappear when power to the motor is shut down, with velocity, the vehicle would provide the best but decays rather slowly as the vehicle slows down. background for the high-frequency measurements when traveling at the slowest velocity for which it is stable (about 2 m/s). To provide a worst- case estimate, we will use the 3.5 m/s spectra for comparison with expected levels of oceanic turbu- lence. 181 The metal-film velocity sensor is sensitive to longitudinal motions of the vehicle and fairly 2 insensitive to vertical and cross vibrations. ~~~~~~I n The apparent velocity vector (mean vehicle veloci- 08 __ IjA 3 ty plus turbulent fluctuations) is very close to the longitudinal axis. We will use the longi- tudinal spectrum at 3.5 m/s for comparison with -- I �t,14 expected turbulence levels. -- ' .|lTurbulence theory predicts that at high wave- D numbers the energy density depends only on the \ Y CI lkinematic viscosity v and the rate of energy dis- -~10 |~~~I(~~~ 1 Isipation �. With this assumption, dimensional lo -- l lliarguments lead to a so-called Universal Turbulence l ll | ~~Spectrum. Assuming v = 0.015 cm2/s (which is I f 'I1 .F typical of sea water at 8�C), we used the method rQ~ F | E l~of Stewart and Grant (1962) to graph a family of N '.i. I - ~I ~~~I _~ p I. Icurves for different dissipation rates. Figure 4 loE ~ 10~~~~~~~~~ Ashows these universal spectral curves for dissi- ---3.5 M/S " pation rates between 10-4 and 10 ergs/cm3/s. l-2 i8M/S ad -12 2.0M/S UNIVERSAL TURBULENCE SPECTRA E - ~~~1.6 M/SE\ \\ \ 3-- 16- \o SPURV II \ VERTICAL AXIS I 4--ACCELEROMETER TEST 10 - I PUGET SOUND -1977 RUN 04 8 -8 t 1015, , 1O LONGITUDINAL I IO~~ ~ 2l� 103 14) 3SPECTRUM AT' I 10 0 10 I0 3.5m/s - , FREQ (Hz ) DISSIPATION, E Figure 3. Velocity spectra from vertical acceler- UNITS ERGS/C-S gI ometer at four vehicle speeds. Peaks 1 10 -l through 4 are due to forced multiples of IC E the propulsion motor frequency. Peaks A V =0015cm2/s through F are natural vehicle vibrations. (WATER AT 35%' (WTE AND 351oo Vertical motion of the temperature sensor in 0 a vertical temperature gradient will contaminate the temperature measurements. The amount of con- 4, tamination was estimated by integrating the verti- -12 l cal velocity spectrum to get displacement, and then 0 I multiplying by the vertical temperature gradient 0.1 C 1O O10 1000 to get an equivalent temperature spectrum. Since theory assumes turbulence on these scales is iso- tropic, vertical and horizontal measurements should WAVENUMBER be equal. Our vibrational contamination is more than 40 dB below the vertical temperature-gradient C p m spectra taken by Gregg (1977) in the North Pacific thermocline, and more than 50 dB below his measure- Figure 4. Comparison of SPURV vibrations with theo- ments beneath the equatorial undercurrent (Gregg, retical dissipation rates between 10-4 1976). Our estimated contamination is more than and 10 ergs/cm3/s and rates measured by 60 dB below horizontal spectra taken by Nasmyth Nasmyth (1970) and Grant, Stewart and (1970) beneath the thermocline off Vancouver Island. MoiZZlliet (1962). The light dashed line Therefore, we do not expect that vibrational effects is the locus of points at which the max- will be observable in our temperature measurements. imum dissipation exists. E represents the measurements by Nasmyth; = rep- resents the measurements by Grant, et al. 182 Values of � between 2 x 10-4 and 3 x 10-3 ergs/ related to velocity by the so-called King's Law: cm3/s were found by Nasmyth (1970) in the Pacific 2 Ocean off Vancouver Island at depths between 60 E = [A + B(pu)n](t - te) (1) and 200 m (lower shaded region). Grant, Stewart and Moilliet (1962) made measurements in a tidal channel with high Reynolds flow and found values d ereiA, B is n velcitants is the fluid of s between 0.1 and 0.6 ergs/cm3/s (upper shaded a t is temperaty, te is the sensoround region). Values of a in the ocean are expected to temperature , and te iswth oce, be around 10-4 to 10-3 ergs/cm3/s, with energeticthveoiysnraculymsrsdniy be aoundl0- to o-~ergscm3s, wth nergtic temperature. Since p, u, and te vary with time, region approching n a ofi erg/m3/s.the velocity sensor actually measures density fluctuations, velocity fluctuations, and environ- When compared with the longitudinal vibration mental temperature fluctuations. spectrum at 3.5 m/s, turbulent signals with � down to 10-4 ergs/cm3/s are for the most part above the c alowrte the vel sensortue level of the vibration-induced velocity. The vi- bration peak at one-half the commutator slot fre- calibration bath so that the ambient temperature quency affectc one results only at a specific can be controlled during calibration runs. A frequency. The effects of this peak on estimates typcal static calibration at four different of dissipation can be eliminated by the technique overheat settings is shown in Figure S. A reason- described by Stewart and Grant (1962) to fit meas- able fit to these curves is given by urements to the Universal Turbulence Spectrum. 2 2 0.26 E2 [-8x10-2 0.32(pu)0 ' 26] Also, cross-spectral techniques using the simul- = + ' (2) taneous accelerometer measurements could be used to subtract the coherent vibrational effects from o i e the measurements (Bendat and Piersol, 1971). We are continuing to investigate these forced vibra- where ti is the initial temperature when the bridge tional peaks and trying to reduce their effects. is balanced and T. is the overheat temperature of the sensor. A comparison of the measured turbulence spec- tra and vibrational spectra taken concurrently will give signal-to-noise ratios from which con- fidence estimates can be calculated. The lack of HOT FILM CALIBRATION an error estimate is a deficiency found in much SENSOR *7488 of the turbulence work published to date. 4- 20-78 Hot-Film Velocity Measurements TANK WATER TEMP=12.3� In order to resolve the velocity spectra for dissipation rates up to 1 erg/cm3/s, at vehicle PROBE OVERHEAT SETTINGS speeds of 3.5 m/s, the frequency response of the O =1.O8 velocity sensor must extend to beyond 500 Hz. To --' obtain a better than 10-dB signal-to-noise ratio IO- - A =:1.06 - - 2 < out past the maximum in the dissipation spectrum, 0=1.04 the sensor noise should be below 10-9 (m/s)2/cpm. Metal-film velocity sensors such as used by Grant, > 81.02 Stewart, and Moilliet (1962) have the required - frequency response. For our study we used Thermo W[w Systems Incorporated type 1230 NACL, conical (_ platinum-film sensors. These sensors consist of < 6. P a 1.5-mm diameter quartz rod whose end has been .D ground to a 40= cone. A band of platinum is plated 0 o around the end of the cone, and is the resistance element. The resistance is typically 5 to 6 4--130 4 with a temperature sensitivity of about 8 to [ w 0 9 x 10-3 Q/OC. The sensors are coated with a thin o quartz coating to insulate the film from salt 6D water. - 2-- . ._ The sensor electronics were built by APL and the sensor is contained in one leg of a dc bridge. The bridge is remotely balanced at operating depth to the ambient water temperature using a low- voltage bridge excitation. The circuit is then WATER VELOCITY W S) switched to the feedback configuration typically used for constant-temperature hot-film anemometry Figure 5. Static calibration of hot-film velocity (Weidman and Browand, 1975; Fremuth, 1977). A sensor at four overheat settings. probe overheat of approximately 44�C is used. The electronics maintain the sensor at constant tem- perature; the bridge voltage required to do this is 183 Because the gain of the feedback amplifier Metal-Film Temperature Measurement in the constant-temperature anemometer circuit is limited by circuit stability requirements, the To resolve the dissipation scales of tempera- bridge operates off-balance, with the amount of ture (which are on the order of 1 cm) at a maximum off-balance dependent on the output bridge voltage. vehicle speed of 3.5 m/s, a sensor is required To obtain the true sensor overheat, the overheat that responds to frequencies of several hundred settings must be corrected by the amount of the hertz. Lueck, Hertzman and Osborn (1977) have bridge unbalance. measured the response of a number of thermistors made for use in the ocean. They find that ther- Equation 2 was used to estimate the effects on mistors do not meet this high-frequency requirement. the velocity measurement of typical variations in The problem is the slow response time of the insu- fluid density and temperature encountered in the lating coating required to isolate the element ocean environment. For the North Pacific thermo- electrically from salt water. Metal-film sensors cline at constant depth, typical density variations exhibit a higher frequency response than thermis- would produce an output signal equivalent to a tors and appear to be the best solution. We use velocity variation of about 10-3 m/s, which is the same type of conical sensor for temperature considered insignificant for the purposes of our measurements as for the velocity measurements. measurements. However, the effects of temperature variations in the medium are much more severe; a The sensor electronics were designed and built 0.10C rms variation in temperature (which is typi- by APL for utilization in SPURV. The sensor is in cal of internal wave effects) would produce an one leg of an ac bridge operated at 6 kHz and bal- equivalent velocity signal of approximately anced close to the mean ocean temperature during 3 x 10-2 m/s rms and could be a significant part the operation. Temperature changes at the sensor of the sensor output. produce a corresponding resistance change in the metal film and a change in the unbalanced output The typical sensitivity to velocity at a mean from the bridge. This output is amplified, band- speed of 3.5 m/s is 0.3 V/m/s. Because of the non- pass filtered, and synchronously detected. In linear relationship between bridge output and SPURV, the signal is digitized to 14 bits at a velocity, this sensitivity increases with decreas- sampling rate of 900 Hz. Sensitivity is about ing velocity (at 2 m/s, DE/Du = 0.5 V/m/s). Thus, 1 x 10-0C temperature change per bit. The SPURV to achieve maximum sensitivity, it would be best to circuitry is self-ranging in two bands spanning operate at the slowest vehicle velocity. The esti- 00 to 29�C. The electronics also have self-test mated sensor noise level at 3.5 m/s is 2 x 10-9 features to check their operation before launch. (m/s)2/cpm, which is barely good enough to resolve the dissipation peak at a = 10-4 ergs/cm3/s. At Static calibration of the metal-film tempera- lower vehicle velocities the signal-to-noise ratio ture sensors is done in the APL temperature-con- would improve since the sensitivity to velocity trolled bath. The temperature of the bath is meas- fluctuations would increase. ured with a platinum resistance thermometer and a Mueller bridge. The thermometer is standardized Dynamic calibration is accomplished by oscil- in a triple-point cell. Comparisons show that lating the hot-film sensor in the mean flow with a calibrations made in our tank are within 2 x 10-3�C shaker pot. We have not made this measurement as of calibrations done by the Northwest Regional yet, but similar measurements by others indicate Calibration Center under the direction of the that sensor response is fairly flat past 500 Hz, National Oceanographid Instrumentation Center. The so we are not concerned about frequency response static calibration measurements are made at various limitations. We are concerned about the dynamic bridge currents, and the measured sensor resistance response of the velocity sensor to turbulent tem- is extrapolated back to a zero-current resistance. perature fluctuations. To evaluate this effect, The extrapolated resistances are fit with a second- measurements will be made in the plume tank de- order polynomial. A typical calibration is scribed in the following section. R(Q) = 5.492854 Q + 8.8468 x 10 Q/0C � T The major problem with these sensors is sta- (3) bility. Several of them have developed pinhole - 3.7 x 10 C leaks, or the quartz coating over the thin film has flaked off. Also any "goop" such as oil or where T is temperature in degrees Celsius. The living organisms in the water can coat or stick to resistance change as a result of bridge current is the sensor. Any change in the boundary layer typically 2 x 10-6 Q/mA2. The current in the around the sensor will drastically affect the bridge circuit creates a small amount of self- calibration. heating which makes the sensor slightly velocity sensitive, but the effects are negligible. Probe We are considering using an acoustic velocity resistances have tended to increase about 0.02 0 meter to obtain measurements with the required low during the past year, which is about a 20C drift in frequency and dc stability. This meter would also calibration. Minor variations in resistance have enable in-situ calibrations of the thin-film sensor been observed between weekly calibrations which are in the region of spectral overlap. Our experience probably due to changes in the contact resistance to date with hot-film sensors does not make us con- when the sensors are disconnected and reconnected. fident that we can make accurate measurements of This low-frequency drift is not detrimental, since turbulent velocity fluctuations without the acous- the low-frequency and mean temperature will be tic current meter. measured with a thermistor-controlled Wien bridge 184 oscillator (Pederson, 1969) which has good low- -0.90 'C RISE frequency stability. The two sensors will overlap one decade in frequency to aid in intercalibration. Dynamic calibrations to obtain the frequency response of the metal-film sensor to temperature are done in our plume tank. A thermal plume acts as a thermal impulse when a sensor is moved hori- zontally through it. The response function is determined from the sensor's output as it passes through this impulse. For this measurement, the sensor and electronics are evaluated as one unit. The plume (Figure 6) is generated by heating ?DIS I five Manginin resistance wires spaced 3 mm apart. About 100 W are dissipated in the wires, generating a plume with a 4'C rise in temperature and a width LEAST COUNT 0012'C of 0.25 to 0.5 cm. 22-- -4.03'C RISE Figure 7. Response of metal-film sensor when shot through plume at 3.3 m/s. 21-- Fabula (1968) studied the response of metal- film temperature sensors, and determined that the frequency response can be modeled by LLJ 20--H(f) = Exp[-A(fTf/D) 1/2] (4) 1~20'- ZD - o34 Cm where f is the frequency, D is the thermal diffu- od sivity of the fluid, and A is a length scale re- Lw lated to the thickness of the boundary layer over the metal film which, in turn, is a function of LLo 19-- velocity. The faster the sensor speed, the smaller H- A becomes and the higher the sensor frequency response extends. Figure 8 shows a plot of the measured response at 3.3 m/s and the theoretical curve. Hughes (1968) measured the response at slower speeds and his response is also shown. When the sensor is operating on SPURV, the temperature signal is band-limited to frequencies below our ) ~ 450-Hz Nyquist frequency. Figure 8 indicates the :2 -: 0 : 2 response at the Nyquist frequency would be about 16 dB down from its response at low frequencies 17Y- cm (i.e., below 10 Hz). This is the best response obtainable with the present state of sensor devel- Figure 6. Spatial plot of temperature through opment. The response of the metal-film sensors thermal plume. begins to roll off at lower frequencies than the response of coated thermistors. However, the response of the metal-film sensor decreases at a The metal-film sensor is mounted on the end of slower rate so as to be better at high frequencies. a pneumatic piston and rammed horizontally through Because of this low-frequency attenuation, response the plume at speeds between 1 and 3.5 m/s. The corrections must be made over all frequencies. speed is measured over a 10 cm distance by two photo-transistor/LED pairs (the plume being cen- Figure 9 compares the noise of our metal-film tered between the pairs). temperature sensor with some typical horizontal temperature spectra. The SPURV measurements were The output of the metal-film sensor's electron- taken in the North Pacific (Ewart, 1976a) and are ics is recorded on SPURV's high-frequency recorder. averages over long horizontal transects regardless For this measurement, the signal is digitized to Of energy level in each block. A single energetic 8 bits at a sample rate of 10,800 Hz. A plot of patch is thus averaged out. The Grant et al. the digitized output is shown in Figure 7. The (1962) data are from selected patches of high- frequency response function is the transform of the turbulence regimes, and hence are higher than the high-frequency sensor response divided by the SPURV results. The two Nasmyth (1970) curves are transform of the plume. also patches selected for high energy. 185 ..I ....I.... .. The electronic and sensor noise is nearly white, and its spectral level is 6.8 x 10-10 �C2/Hz. When this level is divided by the frequency re- . * .sponse function of the sensor shown in Figure 8, * -0.091630 . it forms a lower bound on the measurements of environmental temperature fluctuations (see Fig- LL U/ \_ MEAN TEMP 17.50C ure 9). In the more energetic regions shown, we -, z *E VELOCITY= 3.3M/S will be noise limited only at the highest frequen- Z \cies. I-0 ain Therefore, in turbulent patches where the E5 a= / >signal levels are high, good spectra can be meas- < ' ured, but at the speeds of SPURV the extended fre- HUGHES (1968) quency response of the metal-film sensors is MEAN TEMP=10.940C \ \ required. There will also be quiet regions that VELOCITY = 1.77 M/S have energy levels much below the sensor noise 0.1 ok Nlevels, and these will not be resolved. Again, as v01 .1. .'..... . ...1 , l , , .4i.'.... with the accelerometer results, there is a window 0 5 10 I 20 25 in which good estimates of temperature spectra can be made. Knowing our noise level and the sensor response will allow us to eliminate those measure- 4F REQ(X-H) ments that are sensor limited, and determine a confidence level for the rest. Figure 8. Frequency response for metat-film tem- perature sensor. References Bendat, J.S., and A.G. Piersol, 1971. Random Data: 00 Analysis and Measurement Procedures, Wiley- Interscience. HORIZONTAL Ewart, T.E., 1976a. Observations from straight isobaric runs of SPURV, Preliminary report for SPECTRA the IAPSO/IAMAP PSll Poster Session of the 1-2 f~i~llli_ _ XJOA in Edinburgh. Ewart, T.E., 1976b. Acoustic fluctuations in the open ocean--a measurement using a fixed path, --= Up---- ~6 l,~ D4:I/ji#Di~ tJ. Acoust. Soc. Am., 60(1): 46-59. NASMYTH -4E i4 B1 iiPA~TCHES Ewart, T.E., and W.P. Bendiner, 1974a. Dye diffu- sion in the deep ocean, Transactions of the 55th Annual Meeting of the AGU, 55(4): 316. Ewart, T.E., and W.P. Bendiner, 1974b. Techniques for estuarine and open ocean dye dispersal 166 ~_ -_- l / /// measurement, Rapp. P.-v. Reun. Cons. int. Explor. Mer., 167: 201-212. 10 -7 _ - _ i////l/////nl/ll/l//~ Ewart, T.E., J.D. Irish, and L.O. Olson, 1977. The MATE Experiment, A Preliminary Report, l // t Applied Physics Laboratory, University of 1cTI18 ~I A //// IIWashington. Fabula, A.G., 1968. The dynamic response of towed IO--9 thermometers, J. Fluid Mech., 34(3): 449-464. METAL FILM SENSOR NOISE WITH Fremuth, P., 1977. Frequency response and elec- -110 RESPONSE CORRECTIONS tronic testing for constant-temperature hot- 2IOl3 I IOO2 10' 1 2 103 wire anemometers, J. Phys. E. Sci. Instrum., 10: 705-710. WAVENUMBER Grant, H.L., R.W. Stewart, and A. Moilliet, 1962. (cpm ) Turbulence spectra from a tidal channel, J. Fluid Mech., 12(2): 241-268. Figure 9. Comparison of metal-film thermometer noise level with measurements of temper- Gregg, M.C., 1976. Temperature and salinity micro- ature in the ocean. structure in the Pacific equatorial undercur- rent, J. Geophys. Res., 81(6): 1180-1196. 186 Gregg, M.C., 1977. A comparison of finestructure Pederson, A.M., 1973. A small in-situ conductivity spectra from the main thermocline, J. Phys. instrument, Proceedings IEEE Intern. Conf. Oceanog., 7(1): 33-40. Engineering in the Ocean Environment, pp. 68- 75. Hughes, B.A., 1968. Appendix to Fabula, 1968. Stewart, R.W. and H.L. Grant, 1962. Determination Lueck, R.G., O. Hertzman, and T.R. Osborn, 1977. of the rate of dissipation of turbulent energy The spectral response of thermistors, Deep near the sea surface in the presence of waves, Sea Res., 24: 951-970. J. Geophys. Res., 67(8): 3177-3180. McKean, R.S., and T.E. Ewart, 1974. Temperature Weidman, P.D. and Browand, F.K., 1975. Analysis of spectra in the deep ocean off Hawaii, J. Phys. a simple circuit for constant temperature Oceanog., 4(2): 191-199. anemometry, J. Phys. E. Sci. Instrum., 8: 553- 560. Nasmyth, P.W., 1970. Oceanic Turbulence, Ph.D. Dissertation, University of British Columbia. Widditsch, H.R., 1973. SPURV--The First Decade, APL-UW 7215, Applied Physics Laboratory, Uni- Pederson, A.M., 1969. An accurate low cost temper- versity of Washington. ature sensor, Transactions of Marine Tech- nology Society Temperature Measurements Sym- posium. 187 IMPLICATIONS OF THE JAPANESE EXPERIENCE IN AQUACULTURE DEVELOPMENT FOR THIRTY-THREE FOOD-SHORT COUNTRIES " Yoshiaki Matsuda Woods Hole Oceanographic Institution Woods Hole, Massachusetts 02543 aspects. Abstract The purpose of this paper is to suggest a Aquaculture development in Japan has been flexible model for revitalizing aquaculture de- limited by the following factors: supply and velopment strategies, emphasing the 33 food-short demand, environmental suitability, technical countries in 1973 (1). Assuming that experience capability, legality, experience, infra- in developed countries has some applications to structure, existence of leadership, economic developing countries, the history of aquaculture feasibility and social welfare incentives. in Japan is reviewed. The limiting factors for Similar limiting factors were examined for aquaculture development in Japan are then tested thirty-three food-short countries. Fifty- for the 33 food-short countries. Finally, a two- five aquaculture variables were tested using dimensional aquaculture development model is de- multiple regression analysis combined with veloped as a key to cope with technology transfer Pearson correlation analysis. Results imply problems. that aquaculture development in these countries are affected by the same limiting factors as 2. Methodology in Japan. These findings are the basis for a two-dimensional aquaculture development model The methodology used in this study is based which may be helpful in coping with future primarily on a combination of descriptive, cor- '-echnology transfer problems. relation and multiple regression analyses. Aqua- culture development in Japan is described first in order to explore limiting factors in aqua- 1. Introduction culture development and the growth pattern. The hypothesis that aquaculture activities in the 33 This study is designed to devise an ef- food-short countries are not limited by these fective aquaculture development strategy for factors described for Japan is tested using developing countries where there are food short- Pearson correlation and multiple regression an- ages. alyses. Based on these findings, an aquaculture development model is suggested. The rapidly increasing popularity of aqua- culture and the debates which issue from the (l) study area. The countries considered world-wide aquaculture development scene ne- were Japan and the 33 food-short countries re- cessitate a re-evaluation of present schemes to ported in 1973 (1). These include Algeria, transfer technology related to aquaculture (13). Angola, Bangladesh, Bolivia, Cameroon, Central Aquaculture has the potential for developing African Republic, Chad, Ecuador, El Salvador, into an important vehicle for solving some of Ethiopia, Gambia, Guinea, Guyana, Haiti, India, the employment, income, and nutrition problems Indonesia, Iran, Iraq, Kenya, Mali, Mauritania, in food-short countries. FAO and other UN Niger, Nigeria, Philippines, Saudi Arabia, Sene- agencies, the World Bank, the Asian Development gal, Syrian Arab Republic, Tanzania, Uganda, Bank, Southeast Asian Fisheries Development Upper Volta, Yemen Arab Republic, Yemen Demo- Center (SEAFDC), International Center for cratic, and Zaire. Living Aquaculture Resource Management (ICLARM), USAID, U.S. Peace Corps, OXFAM (England), Inter- (2) Data. There is an abundance of litera- national Development Research Center in Canada ture on Japanese aguaculture. The most useful (IDRC), International Foundation for Science data are Drew 1951 (2), Inaba 1964 (9), Japanese in Sweden (ISF), Japan International Coopera- Federation of Laver and Shell Fisheries Coopera- tion Agency (JICA), and German Volunteers are tives (IFLSFC) 1969 (10), Matsuda 1977 (12), some of the organizations promoting aquaculture Miyaji 1974 (15), Oshima 1973 (16) and Suisansha in developing countries (11). In addition, Fisheries Yearbook (20). national development plans for aquaculture have been proposed in many developing countries (4, Aquaculture data for the 33 food-short coun- 6,7). However, strategies now in use are in- tries were compiled from Bell and Canterbery 1976 flexible and show a bias toward technological (1); FAO 1974 (2); FAOa,b,c, and d, 1975 (4,5,6,7), 188 *Prepared with funds from the Pew Memorial Trust and by the Department of Commerce, NOAA Office of Sea Grant under Grant #04-7-158-44104, and the Marine Policy and Ocean Management Program of the Woods Hole Oceanographic Institution. Pillay 1972 (17); 1973 (18) and 1976 (19) and in- point scale (0: unacceptable; 1: poor potential formation from the World Bank, SEAFDC, USAID, U.S. acceptance; 2: satisfactory potential acceptance; Peace Corps, OXFAM, and JICA (11). and 3: already cultured). Selected variables are summarized in Appendix For the environmental suitability variable, Tables 1-2. Included were 1 dependent and 54 in- environmental acceptability scores were adopted dependent variables. Per capita aquaculture pro- from Bell and Canterbery. Each country's total duction in 1975 (kg) was selected as the dependent score consists of the sum of environmental accepta- variable because most of these countries have few bility scores of the 14 species measured by a 1.2- imports or exports of these products and this point scale (0: no environmental acceptability, reflects consumption as well as production. The and 1.2: the highest environmental acceptability). total aquaculture production in 1975, based on Pillay 1976 (19), was divided by population For technical capability, three variables estimates for 1975 in the UN Demographic Year- were used: amount of international assistance, book 1975 (4). number of research units, and number of papers presented at any one of the 3 FAO Symposia on Eleven basic variables are chosen as inde- Aquaculture. International assistance is an pendent variables. It should be noted that data external factor affecting technical capability, from private sectors were not readily available; while research units are an internal factor. The thus this analysis is biased toward government number of papers reflects, at least roughly, the activities. These 11 variables include consumer degree of research activity. International assist- acceptability, environmental suitability, amount ance to aquaculture is measured by a score of 1 of papers presented at any one of the 3 FAO Sym- if exist, 0 otherwise for each variable: aqua- posia on Aquaculture, experience, degree of inter- culture joint ventures, technical assistance to national relations, degree of political activities, aquaculture from UN, related agencies, and so average rates of return on investment (%), average forth, during the 1967-1973 period. The number of rates of return on operating costs (%), and research units is the total number of research average rates of return on gross income (%). institutions presenting papers to any one of the 3 FAO Symposia on aquaculture in Asia (17), Latin A preliminary analysis suggested that the America (3) and Africa (5). Another variable is last three economic variables had no significant the total number of papers presented at any one of correlation with dependent variables at the 0.05 the 3 FAO Symposia. The use of numbers for these level mainly due to insufficient data (Table 1). activities may not reflect the true relative im- Therefore, we deleted these three variables from portance because one project might be much larger further analysis. In addition, there were inter- than others. However, in this instance each pro- actions and overlapping among the rest of the 8 ject is weighted equally because sufficient data independent variables (Table 2). Thus, twenty- is not now available to provide a better indicator. eight two-way interactions of these 8 variables and 15 three-way interactions were added as inde- The degree of political activities (which re- pendent variables (Appendix Table 1). present legal structure, infrastructure and social welfare programs)takes account of government sub- Consumer acceptability was taken to re- sidies, research institutions, extension service, present the equilibrium of present supply and cooperatives, joint ventures, government farms, demand although the data in this category is in- stocking and conservation practices, and inter- sufficient. Consumer acceptability scores were national assistance to aquaculture. adopted from Bell and Canterbery. Total scores for each country are the sum of consumer accepta- International assistance to aquaculture, bility scores of 14 aquaculture species Clndian fisheries trade (imports and exports) data for carp, channel catfish, tilapia, eel, rainbow 1975 with FAO Yearbook of Fisheries Statistics trout, mullet, milkfish, yellowtail, Panaeus 1975 (8), presentation of papers at any one of the shrimp, oyster, mussels, walking catfish, macro- 3 FAO Regional Workshops on Aquaculture Planning brachium, and blue-green algae) measured by a 4- in Africa (4), Asia (6) and Latin America (7) in 189 1975, and participation (1 if participated, 0 other- reliable results. wise) to any one of the 3 FAO Symposia on Aqua- culture were used to indicate interest on the part 3. History of Aquaculture in Japan of the countries' leaders in aquaculture ("Leader- ship"). Japanese history reflects chronological phases in the history of aquaculture: precondition (be- Experience may not be a product of long his- fore Edo era), phase I (Edo era: 1600-1868), phase tory; it may be a function of learning during a II (1868-1945), and phase III (after World War II) relatively short period. However, data is limited (12). During the precondition period, common carp on such short-term experience. Therefore, we used (Cyprimus carpio L.) was first raised and became historical records as a proxy for experience (1: popular among the nobles. During the phase I old practice - before 1900; 2: moderately old period, oyster (Osterea gigas T.), laver (Porphyra practice - 1900-1945; 3: recent practice - after tenera k.), fancy carp (Cyprinus carpio L.) and World War II; and 4: negligible practice). goldfish (Carassius auratus L.) appeared as aqua- culture species. During the phase II period, Profits are a measure of economic feasibility salmon (Onkorhynchus spp.) and trout (Salmo as an incentive for the private sector. Economic gairdnerii irideus G.), eel (Anguilla japonica T. feasibility is represented by average rates of et S.), pearl (Pinctada martensii D.), "Ayu" fish return on investment (%), on operating costs (%) (Plecoglossus altivelis T. et S.), and yellowtail and on gross income (%). However, data were a- (Seriola quinqueradiata T. et S.) were added. All vailable for only six of the 33 countries (Central other aquaculture species were introduced during African Republic, India, Indonesia, Nigeria, the phase III period. Philippines and Senegal). These data were a mixture of survey results and proposed budgets, Some knowledge of methods of raising carp in so an average of such rates of return was figured ponds was apparently introduced from China by 100 for each of these six countries. A.D. (2). In 724 the Emperor Shomu emphasized the Buddisht precept against killing and changed (3) Techniques. Techniques employed are a Japanese eating habits for the next five centuries combination of Pearson correlation, and multiple until Shinran, a new Buddism leader in the 13th regression analyses using the SPSS system at the century, advocated that people might eat fish Woods Hole Oceanographic Computer Center. One of because eating fish was beyond the Buddhist pre- the serious problems of the regression analysis is cept against killing (16). During the 9th and multicollinearity, which means that some or all of 10th centuries, common carp were raised as pet the independent variables are very highly inter- fish among the nobles and developed as moving correlated, resulting in a reduced reliability in treasures in landscape gardening. the relative importance indicated by the partial regression coefficients. In order to eliminate the Despite the early introduction of fish culture effects of multicollinearity, Pearson correlation into Japan, the rearing of fish was of relatively and multiple regression analyses were combined in little importance and its development was modest such a way that one selected the most important before 1600. This is attributable to limited independent variable from the Pearson correlation supply and demand, quasi-legal prohibition of analysis, checked the correlation coefficients eating fish, and the lack of strong interest between the most important variable and the rest of among the countries' leaders in aquaculture de- independent variables through Pearson correlation velopment, in addition to a lack of infrastructure analysis, selected variables of low correlation co- and experience. efficient (r (.5) with the most important variable, and analyzed a set of the most important variable During the phase I period (1600-1868), aqua- and its low-correlated independent variables with culture was regarded as a potential local in- the dependent variable through multiple regression dustry by some local lords, and they attempted to analysis. If more than two independent variables transplant various species in various parts of were ranked as most important, the same procedure Japan. After much trial and error, a common carp was repeated until the analysis was completed. culture settled in Niigata and Nagano prefectures, This method was adopted as it gaves the most oyster at Hiroshima Bay, laver and clams (Corbi- 190 C. et V.) which constituted most of the relatively cula leana p.) in Tokyo Bay, and goldfish at Yama- high aquaculture production during the 1941-1945 to-kooriyama. Despite the favorable climate, the p dvle period. This is a clear example of a government's development of aquaculture as a whole was slow duringpmens p eriod bcaqu e asa hlite inas t ability, through its legal powers, to wipe out all during this period because of limited infrastruc- past efforts toward aquaculture development, as ture, experience and technical capability, and a well as to promote its development. relatively low demand for the cultured species. Immediately after the war the aquaculture During the phase II period (1868-1945) the country, after long isolation, adopted some as- industry entered a period of recovery during the pects of modern Western civilization with energy general reconstruction, and began phase III. Owing to General MacArthur's intervention, the and enthusiasm. The government dispatched people pearl and oyster industries recovered quickly (16), to America and Germany to learn techniques per- and the Fisheries Agency Act was passed in 1948. taining to aquaculture (16). Salmon and trout hatcheries wer abualtureFishery Section wan tt Again the legal climate became favorable for aqua- hatcheries were built, a Fishery Section was es- and the infrastructure re- culture development, and the infrastructure re- tablished in 1881, many laws concerning aqua- tablishe d i n 1881, ma ny laws conersning aqua- covered quickly. By 1960 the experience of aqua- culture were legislated, and numerous relatedchnical educational and res earch institutions were e s- capability increased rapidly. Fixed net culture tablished. Corresponding to government efforts, developed in 1952, and artificial seedlings for developed in 1952, and artificial seedlings for commercial aquaculture developed among private laer became possible in 1958 (JFLSFC 1969). The laver became possible in 1958 (JFLSFC 1969). The ysectors. (9,15w and 16). Eel, pearl, Ayu, and demand for aquaculture species was also high, re- yellowtails were added as aquaculture species, sulting in good profit potential. The following suiting in good profit potential. The following due mainly to the enthusiasm of individuals. decades, which were characterized by economic decades, which were characterized by economic Aquacultur production in Japan during th growth, resulted in the rise of living standards Aquaculture production in Japan during the 1912-1975 period is shown in Figure 1,. The first and increased demand for aquaculture products, national aquaculture conference was held in 1913 except for pearls and oysters. In 1962 the Inland national aquaculture conference was held in 1913, Sea Fish Farming Center was established and and subsidies for aquaculture became available in 1918. Raft culture of oysters developed in floating net culture developed for laver, a popu- in ~~~19182~ . ~ afdlar magazine "Fish Culture" was first published 1924. The Perfectural Experiment Station Act was legislated in 1929. in 1964, and a low temperature seed storing net legislated in 1929. for laver was developed in 1965. A high quality of leadership toward aqua- culture development in both public and private fowever, the aquaculture industry has en- countered environmental problems due to urbaniza- sectors played an important role. Increased ex- *perince and technical capability, as well as tion and water pollution as well as natural cata- perience and technical capability, as well as a favorable legal environment, encouraged the strophe, Abnormally warm weather (laver), ty- growth of aquaculture. Further more, the demand phoon, diseases (laver, goldfish, eel, yellowtail and scallops), shortage of fingerlings (yellow- for aquaculture products increased due to rapid (yellow population growth and urbanization and to efforts tail and eel), red tide (yellowtail, pearl, to expand markets (9,16), particularly in the eel oyster), declines of demand (pearl and oyster) and pearl industries. By 1930 all the basic con- and reclamation (laer) have been some of the limiting factors for aquaculture development ditions for a large expansion of aquaculture had following the war. Thus, Japanese aquaculture developed, and an infrastructure existed which production as a whole does not show a smooth would lead to economically sound aquaculture. growth curve; instead, it follows a modified combination of the Sigmoid curve and the J- This growth was curtailed by Japan's in- volvement in international wars during the 1930s shaped growth curve. and 1940's. Pearls and eels were regarded as In 1973, the Japanese Diet dealt with issues luxury goods, and were prohibited. Carp, laver pertaining to pollution and on July 1, 1977, the and some other aquaculture species were also Japanese government declared a 200-mile economic treated in the same way, with the exception ofpanese government declared a 200-mile economic zone. There now exists a new climate favorable grass carp (Ctenopharyngodon idellus C. et V.) for aquaculture. and silver carp (Hypophthalnlchthye moritrix 191 Despite the importance of environmental In summary, throughout the history of Japan, suitability for aquaculture production in general, aquaculture development has been limited by supply it was not highly correlated with per capita pro- and demand, environmental suitability, technical duction. This might result from inadequate data. capability, legality, experience, infrastructure, In fact, most of the selected species for the en- existence of leadershipeconomic feasibility and vironmental suitability contribute neither to pre- social welfare incentives. Aquaculture has not sent aquaculture production nor to consumption. followed a smooth growth curve. Although, environmental suitability value is positively correlated with the values for con- 4. Analysis of the Material for the 33 Food-Short sumer acceptability, international assistance, and Countries degree of international relations there is no obvious explanation for this. It maybe that an (1) Preliminary Results environment suitable for selected species maybe of a general enough type to attract the interest Pearson correlation analysis with per capita of international agencies. Nevertheless, im- aquaculture production in 1975 reveals several provement of data on environmental suitability is significantly correlated independent variables essential for future analysis. (S = 0.001): consumer acceptability, amount of international assistance, number of research Among technical capability variables, the units, number of papers, degree of international amount of international assistance has the highest relations, degree of political activities, and correlation with per capita production. Inter- experience; and one less significant variable national assistance is intended to provide tech- (S = 0.033), environmental suitability (Table 1). nology for these countries to increase per capita The high correlation among these independent vari- production in response to critical nutrition prob- ables results in multicollinearity problems lems. The largest recipient of international as- (Table 2). sistance to aquaculture during the 1967-1973 per- iod was the Philippines, which recorded the Among these independent variables, consumer highest per capita aquaculture production in 1975 acceptability is the only one subject to changes (2.94 kg). Further, the amount of international in demand. Although the data include species which assistance is correlated with all other seven in- are not now consumed in the 33 food-short coun- dependent variables. In particular, high correla- tries, consumer acceptability is highly correlated tions are observed with the number of research with per capita production. Consumer accepta- units, degree of international relations, degree bility is likely to reflect present consumption of political activities, and experience. High patterns in such a way that the higher the pre, correlation with experience is explained by the sent per capita consumption, the larger the con- fact that one main purpose of technical assistance sumer acceptability.Because external trade is projects has been to improve existing aquaculture negligible at present, per capita consumption and management. Therefore, most international as- per capita production can be interchanged. sistance has been practiced in Bangladesh, India, Further, consumer acceptability is significantly Indonesia and the Philippines where aquaculture correlated with 7 other independent variables is an ancient practice. Thus, all these high (r>.7): international assistance, number of correlations are reasonable. research units, degree of political activities, and experience. Since international assistance, The next highest correlation with per capita number of research units, and degree of political production was .7864 with the number of research activities are highly production oriented vari- units. This high correlation is partly attri- ables, it is possible that effects of these vari- butable to one inclusion of the Philippines where ables on production are transformed into effects there are seven research units (the highest) and on consumer acceptability. Experience, or a the highest per capita production. Although most past history in aquaculture, certainly effects of these research units have been supported by consumer acceptability because consumer taste and international assistance, internal activities are preference are products of their past. partially also important for production. Internal research activities provide fingerlings where they 192 are not easily available, advice, better aqua- Results from the multiple regression analysis, culture management, and support through extension combined with Pearson correlation analysis, are services. Research activities might affect total summarized in Table 3. production directly and per capita production indirectly. For example, in India where total After eliminating multicollinearity problems, aquaculture production, supported by the same the final results indicate that a three-way in- number of research units as in the Philippines, is teraction of the consumer acceptability, amount of the highest (494,OO0t), per capita production in international assistance, and the number of papers 1975 was relatively small (.825Kg) because of presented at any one of the 3 PAO Symposia on Aqua- India's huge population. Further, the number of culture has the most significant relationship to research units is highly correlated with number per capita aquaculture production in 1975. The of papers, degree of international relations, square of multiple correlation (R2) is .85794 which degree of political activities and experience. implies that the result is highly reliable. The The direct correlations among these variables are regression coefficient is large (B = 2443) so that obvious. However, these correlations could be the impact of the interaction of these three limi- directly or indirectly affected by the following ting factors on per capita aquaculture production logic: the longer the experience, the higher the would be large. consumer acceptability, which in turn leads to larger total production, and more active political Despite relatively high correlations with per and research activities. As political and re- capita production, the final result does not in- search activities increase, international rela- dlude any direct effect of the 8 basic independent tions become more important. variables, because of high interactions and/or overlap among these. Since the result is mechani- Number of papers presented at international cally determined by regression analysis, assuming meetings had the least correlation with per capita linearity, we could not simply say that other var- production because of its indirect effects. Number iables are insignificant. Instead, if this inter- of papers has a high correlation with degree of action variable is deleted from the data, we could political activities. The logic behind this is expect a completely different second best answer that the more active the political activities, the which might be some other combination. p ~~more active the research activities, which in turn leads to larger number of papers. In summary, per capita aquaculture production in these 33 food-short countries is limited by Degree of international relations is correla- consumer acceptability, environmental suitability, ted with per capita production, but only a little. amount of international assistance, number of re- The high correlation with degree of political acti- search units, number of papers presented to any vities is because external incentives have en- one of the 3 FAD Symposia on aquaculture, degree hanced internal activities. of international relations, degree of political activities, and experience. These limiting fac- Degree of political activities and experience tore interact with and/or overlap each other and are also correlated with per capita production. can explain per capita aquaculture production up High correlation between degree of political acti- to the oint that the square of multiple correla- vities and experience is reasonable because his- tion (R ) reaches .85794, based on our present torical experience has influenced aquaculture acti- data. These results for the 33 food-short coun- vities internally. tries imply that aquaculture development is also limited by one or a combination of the same fac- Despite the fact that most data on independent tors as prevail in Japan. variables is limited to government activities, significant correlations with the dependent vari- 5. Discussion able, per capita production, imply that the role of government in per capita aquaculture production The evaluation of ecom~omic feasibility for is important. the thirty-three food-short countries has suf- fered from insufficient and/or inadequate data (2) Results for economic variables in these countries. There 193 is mch rom fr th impovemnt o dat. Neer- capability, legality, experience, infrastructure, is mch rom fr th impovemnt o dat. Neer- existence of leadership, economic feasibility and theless, economic feasibility seems to be critical social welfare incentives. The results for the in these countries. According to PAO (4), failures 33 food-short countries indicate that present per of some of the ill-conceived programs in the past capita aquaculture production as an indicator for have continued to remain a major constraint in con- aquaculture development is a function of an inter- vincing farmers and investors of the economic via- action of the consumer acceptability, amount of bility of aguaculture. There are indeed problems international assistance and the number of papers of increasing input costs, exploitation of middle- presented at any one of the PAO Symposia on aqua- men, perishable products and remoteness of pro- culture; and all the other basic independent vari- duction areas from the market, which directly ables could also he used, with a declining degree affect the economic feasibility. of correlation. These results imply that aqua- culture development in these countries is also Based on the previous analyses, together with limited by one or a combination of the same sort the importance of economic feasibility, we could of factors as prevail in Japan. one possible assume that aquaculture development measured by development from our analyses could be a hypo- per capita production in these food-short coun- thetical aquaculture development model to cope tries is limited by one or combinations of nine with technology transfer problems in these food factors in any given circumstance: supply and short-countries. This model would imply the demand, environmental suitability, technical floig capability, legality, experience, existence of floig leadership, infrastructure, economic feasibility i) Until industry reaches the "take off" and social welfare incentives which interact with stage, the growth of aquaculture is a one another. Assuming that the Japanese experience slow, step-by-step process. is also applicable to a growth model in these countries, a hypothetical aquaculture development ii) The elimination of any one of the model could be posited. The model consists of limiting factors will contribute to functional and growth dimensions. the growth of aquaculture, but not much. The functional dimension emphasizes the im- portance of the existence of leadership as iii) Nevertheless, the best way to develop lubrication function, infrastructure as a trans- aquaculture is to eliminate the limiting portation function, social welfare as an incentive factors one by one. for the public sector, and of economic feasibility as an incentive for the private sector. Supply Although further research is needed for the and demand, environmental suitability, technical improvement of data, economic feasibility, ex- capability, legality and experience become crucial perience, social welfare, links among limiting as basic limiting factors for aquaculture develop- factors and methodology, the proposed two-dimen- ment only when leadership exists, and infrastruc- sional aquaculture development model -- function ture, economic feasibility and social welfare in- and growth -- is likely to be adequate to predict centives are sound and adequate. when, how, and why aquaculture can develop in any The growth dimension is regarded as a com- given circumstance or place. bination of the Signoid shaped growth curve in the long run and the J-shaped growth curve in the short Akoldmn run which occurred in aquaculture development in The author wishes to acknowledge Mr. M.N. Japan. if each limiting factor is removed, the Mistakido at PAO and his colleagues for providing growth curve should be a perfect combination of me the variable information; Drs. Paul M. Eye, H. these two curves. However, each limiting factor Burr Steinbach, Kenneth 0. Emery, Robert W. Morse, may become so critical that despite increasing Susan B. Peterson, Leah J. Smith and Woollcott efforts the growth curve would stagnate as in Smith, at the Woods Hole oceanographic Institution India or decline as in Indonesia (18 and 19). for constructive criticism of the manuscript. Aquaculture could even be eliminated. For example, Acknowledgment is also extended to Mrs. Yoko regardless of an early introduction of aquacultural Matsuda, Kaleroy L. Hatzikon, Ann Martin and methods, aquaculture activity is negligible in Lynda E. Davis for their kind cooperation and Guyana, Iran, Iraq, Niger, Saudi Arabia, Syrian assistance. I would like to thank the Institu- Arab Republic, and upper Volta (3 and 5). Another tion's Marine Policy and Ocean Management Program classic example is the abandoned ponds built by which made this study possible. colonial governments in many parts of Africa (5). Modern versions of this disintegration could re- References sult from disease and competition, urbanization. and pollution, as occurs in Japan (12). (1) Bell, F.W. and E.R. Canterbery, Aquaculture for the Developing Countries: A Feasibility 6. Summary and Conclusions Study. Ballinger Publishing Company. In any given circumstance, aquaculture de- Cmrde ascuet,17,p.26 velopment in Japan has been limited by one or a combination of the following factors: supply and demand, environmental suitability, technical 194 (2) Drews, R.H., "The Cultivation of Food Fish (16) Oshima, J., A Geographical Study of Aqua- in China and Japan: A Study Disclosing, culture in Japan, University of Tokyo Press, Contrasting National Patterns for Rearing Tokyo, 1973, pp. 455 (in Japanese). Fish Consistent with the Differing Cultural Histories of China and Japan," Ph.D. Dis- (17) Pillay, T.V.R., ed., Coastal Aquaculture in sertation, University of Michigan, 1951, the Indo-Pacific Region, Fishing News (Books) pp. 250. Ltd., London, 1972, pp. 497. (3) FAO, Working Papers on Symposium on Aquacul- (18) Pillay, T.V.R., "The Role of Aquaculture in ture in Latin America, Montevideo, Uruguay, Fisheries Development and Management," J. Fish. 26 November - 2 December 1974. Res. Board Can., 1972, 30(12): 2202-2217. (4) FAOa, Aquaculture Planning in Africa. Report (19) Pillay, T.V.R., "Overview of Aquaculture of the First Regional Workshop on Aquaculture Development During the Last Ten Years," FAO Planning in Africa, Accra, Ghana, 2-17, July Technical Conference on Aquaculture, Kyoto, 1975. ADCP/REP/75/1, pp. 114. Japan, 26 May - 2 June, 1976. (5) FAOb, Working Papers on Symposium on Aqua- (20) Suisansha, Fisheries Yearbook 1968, Tokyo 1968; culture in Africa, Accra, Ghana, 30 September 1971, 1971; and 1976, 1976 (in Japanese). - 6 October 1975. (21) United Nations, U.N. Demographic Yearbook (6) FAOc, Aquaculture Planning in Asia. Report 1976, 1977. of the Regional Workshop on Aquaculture Planning in Asia, Bangkok, Thailand, 1-17 October 1975. ADCP/REP/75/2, pp. 154. (7) FAOd, Aquaculture Planning in Latin America. Figure 1. Aquaculture Production in Report of the Regional Workshop on Aquaculture Japan during the 1912-1975 Planning in Latin America, Caracas, Venezuella, Period (Unit: 1,000t) 24 November - 10 December 1975. ADCP/REP/76/ 3, pp. 173. 00 (8) FAO, Yearbook of Fisheries Statistics 1975, Vols. 40-41, 1976. PHASE I PHASE m l' (9) Inaba, S., Practical Eel Culture, Midori Shobo, Tokyo, 1959, pp. 281 (in Japanese). 800- (10) Japanese Federation of Laver and Shell Fisheries Cooperatives, "The First Twenty WORLD WAR I - Years of Japanese Federation of Laver and Shell Fisheries Cooperatives," 1969, pp. 8 600- (in Japanese). (11) Matsuda, Y., "International Assistance on Aquaculture in Developing Countries," Un- published paper, Woods Hole Oceanographic Institution, 1977, pp. 28. 400 400- (12) Matsuda, Y., "History of Aquaculture in ! Japan," Woods Hole Oceanographic Institution. I Unpublished paper, 1977, pp. 56. (13) Matsuda, Y., "The Growth of Aquaculture in | Developing Countries: Potentials, Patterns, 200 and Pitfalls," Fisheries, 1978, in press. ! (14) Ministry of Agriculture and Forestry, Fisheries Statistics of Japan 1975, 1977, *! pp. 77. I i 1920 1940 1960 YEAR (15) Miyaji, D., A Story of Ayu, Iwanami-shinsho 386, Iwanami-shoten, Tokyo 1974, pp. 227 (in Sources: Ministry of Agriculture and Forestry, Japanese). Fisheries Statistics of Japan 1975, 1977(14) and Suisansha, Fisheries Yearbook 1968, 1968; 1971, 1971; and 1976, 1976(20). 195 Table 1. Pearson Correlation with Per Capita Aquaculture Production in 1975 for the Thirty-Three Food-Shortages Countries Categories Independent Correlation Aquaculture Variables Coefficient (Indicators) (r) Cases Significance Supply and Demand 1. Consumer acceptability scores .7653 27 .001 Environmental Suitability 2. Environmental suitability scores .3965 29 .033 Technical capability 3. Amount of international assistance .7881 33 .001 4. Number of research units .7864 33 .001 5. Number of papers .5540 33 .001 Existence of leadership 6. Degree of international relations .5550 33 .001 Legality-infrastructure- social welfare 7. Degree of political activities .6172 33 .001 Experience 8. Experience -.6553 33 .001 Economic feasibility 9. Average rates of return on investment .6762 6 .140 10. Average rates of return on op. costs. -.4594 6 .359 11. Average rates of return on gross income -.4849 6 .326 Table 2. Correlation Matrix for the Eight Basic Independent Variables CA ES IA RU P DIR DPA E Consumer acceptability (CA) 1.0 .52 .76 .80 .60 .68 .80 -.74 Environmental suitability (ES) .52 1.0 .55 .47 .42 .61 .47 -.40 Amount of international assistance (IA) .76 .55 1.0 .80 .65 .73 .76 -.78 Number of research units (RU) .B0 .46 .80 1.0 .88 .70 .85 -.73 Number of papers (p) .60 .42 .65 .88 1.0 .61 .76 -.59 Degree of international relations (DIR) .68 .61 .73 .70 .61 1.0 .75 -.72 Degree of political activities (DPA) .80 .47 .75 .85 .76 .75 1.0 -.80 Experience (E) -.74 -.40 -.78 -.73 -.59 -.72 -.80 1.0 Table 3 Emperical Results ** Significant at the 0.01 level Y' = 3.812x102 + 2443 X X2X3 where Y' is an estimated per capita aquaculture production in 1975 (kg), X1 is the consumer acceptability, X2 is the amount of international assistance, X3 is the number of papers; Pearson correlation coefficient between Y (per capita aquaculture production in 1975) and X1X2X3 is .92625 in which 33 countries are included; F value for regression coefficient (B) is 187.2187**; multiple correlation coefficient (R) is .92625; R2 is .85794; and F value for multiple correlation coefficient (R) is 187.2187.** 196 Apendix Table 1: Aquaculture Variables Used Item Variables 1 Consumer acceptability* 2 Environmental suitability* 3 Amount of international assistance on aquaculture: 1967-73* 4 Number of research units concerning aquaculture in 1975* 5 Number of papers presented at the 3FAO Symposia on Aquaculture in Asia, Africa and Latin America* 6 Degree of international relations on aquaculture* 7 Degree of political activities* 8 Aquaculture experience* 9 Average rates of return on investment (%)* 10 Average rates of return on operating costs (%)* 11 Average rates of return on gross income (%)* 12 Item 1 x Item 2 13 Item 1 x Item 3 14 Item 1 x Item 4 15 Item 1 x Item 5 16 Item 1 x Item 6 17 Item 1 x Item 7 18 Item l x Item 8 19 Item 2 x Item 3 20 Item 2 x Item 4 21 Item 2 x Item 5 22 Item 2 x Item 6 23 Item 2 x Item 7 24 Item 2 x Item 8 25 Item 3x Item 4 26 Item 3x Item 5 27 Item 3 x Item 6 28 Item 3 x Item 7 29 Item 3 x Item 8 30 Item 4x Item 5 31 Item 4 x Item 6 32 Item 4 x Item 7 33 Item 4 x Item 8 34 Item 5 x Item 6 35 Item 5 x Item 7 36 Item 5 x Item 8 37 Item 6 x Item 7 38 Item 6 x Item 8 39 Item 7 x Item 8 40 Item 1 x Item 3x Item 4 41 Item l x Item 3x Item 5 42 Item 1 x Item 3x Item 6 43 Item l x Item 3x Item 7 44 Item I x Item 3x Item 8 45 Item 3x Item 4x Item 5 46 Item 3x Item 4x Item 6 47 Item 3x Item 4x Item 7 48 Item 3x Item 4x Item 8 49 Item 4x Item 5x Item 6 50 Item 4x Item 5x Item 7 51 Item 4x Item 5x Item 8 52 Item 5x Item 6x Item 7 53 Item 5x Item 6x Item 8 54 Item 6x Item 7x Item 8 55 Per capita aquaculture production in 1975 (Kg)** * Basic independent variables. ** Dependent variables. 197 Appendix Table 2: Data Matrix for the Thirty-Three Food-Short Countries ~~~~~~~~~~~~~~~,~~~~~~~~~~~w C o ~~~~ ~~~~~ ~~4J 4J34 I-H g ~~~~~~ ~~~~4-) cc ~ 0) (d43 4' 04 -& 4"' 0 0 004 'r( CIX H �� a, o OP~~~~~~~~~~~~ ~ ~ >, H~~~~4 43 OLn 2 H4 H 0 ) 43 J4 4J 0)4 r, 0)2 0 aH 0 44 9a4 4 40 '4-4 e 1- -H 43> HM a )0 a 0 0 -Z~ Oo ow 0a 00 H4 r 043 ) 0H 04 0' w -i a) eo $ -H a)- .1 -4 aH w 44 -Ha (n 4-40)D U 0 0)0 a)0-- Q) 4. 4.304- N4Q B4H 004 00 0 0430 0-H 0 43 00) 4-'0OP4 4302 -I a)) 0-H Z -H M0o 431 0) m f a eo, 04wm 43 0 J 4343r' 44 '44 0)0%:-H a)- -IH , 043 00 04r 4> 04 t- 0 4 C: I ~00EA 0 0)0$~43 0 > 04 02 0M 0 04 roe *43 0-1 r- 4 C 00)0) 0-H a) 00 ) -03 --H43 00 -H 002W -H13 4J O43H : 443 04 0) P > )430En a) O 00 00 S3 00 0 0)003 0)0 a) ) >0) >a) >0 0)00 Algeria 9 8.0 0 0 0 0.0 0 4 n~a. n.a. n.a. 0 Angola 16 10.2 0 0 0 0.0 0 4 n.a, n~a. n.a. 0 Bangladesh 31 11.8 4 1 0 2.0 3 1 n.e. n.a. n.a. 0.995 Bolivia 23 4.4 0 1 1 0.0 2 2 n.a. n.a. n.e. 0.248 Cameroon 23 12.8 4 1 1 1.0 4 3 n.a, n.a. n.a. 0.007 Central Afr.Rep. n.a. 5.0 2 2 5 3.0 3 2 174.7 196.8 58.8 0.023 Chad 17 5.0 0 0 0 1,0 0 4 n.e. n.a. n.a. 0 Ecuador 22 13.2 2 2 4 3.5 4 2 n.a. n.a, n.e. 0.147 El Salvador 27 13.5 1 2 2 3.0 5 3 n.a. n~a. n.e. 0.301 Ethiopia 17 11.3 0 1 1 1,0 3 4 n.e. n.e. n.e. 0 Gambia n.e. n.e. 0 0 0 1.5 0 4 n.a, n.a. n~e. 0 OD' Guinea 16 13.0 0 0 0 0.0 0 4 n.a. n.e. n.e. 0 Guyana n.e. n~e. I 0 0 2.0 1 3 n.a. n.a, n~e. 0 Haiti 18 12.1 0 0 0 0.0 0 4 n.e. n~e. n.e. 0 India 33 13.0 5 7 11 3.0 7 1 20.8 152.6 49.3 0.825 Indonesia 35 12.0 4 3 1 4.0 4 1 35.7 93.2 45.0 1.101 Iran 9 10.9 2 0 0 1.0 0 3 n.a. n.a. n.e. 0 ,Iraq 10 7.0 2 0 0 1.0 0 3 n.a. n.a. n.e. 0 Kenya 16 13.8 2 1 0 3.0 3 2 n.e. n.e. n.e. 0.029 Mali 16 5.0 0 0 0 1.5 0 4 n.e. n.e. n.e. 0 Mauritania n.a. n.e. 0 0 0 0.0 0 4 n.e. n~a. n~e. 0 Niger 16 7.0 1 0 0 1.0 2 3 n.a. n.a, n.a, 0 Nigeria 22 13.0 3 2 6 3.0 4 2 7.7 22.0 18.0 1.191 Philippines 37 13.0 7 7 6 4.0 7 1 371.6 149.8 46,3 2.940 Saudi Arabia 11 4.8 0 0 0 0.0 0 3 n.e. n.a. n~a, 0 Senegal n.a, n.e. 2 1 1 2.0 5 3 14.9 1650.0 93.0 0.046 Syrian Arab Rep, 10 6.0 0 0 0 0.0 1 3 n.e. n.e. n.e. 0 Tanzania 18 13.0 2 2 3 4,0 4 2 n.a. n.a. n,a. 0.098 Uganda 17 5.0 2 1 2 2.0 4 3 n.e. n.e. n~e, 0.060 Upper Volta 17 5.0 0 1 1 0.0 0 3 n,a. n.e. n.e. 0 Yemen Arab Rep, 9 4.8 0 0 0 0.0 0 4 n.e. n~a. nae. 0 Yemen Dam. n.e. 4.8 0 0 0 0.0 1 4 n.e. n~a. n.e. 0 Zaire 17 5.0 1 1 1 1.0 4 2 n.e. n.a, n~e. 0.201 *Basic independent variables **Dependent variables OPTIMAL MANAGEMENT OF GLOBAL TUNA RESOURCES Man-Loong Chan, Ph.D.* J.E. Buchanan W. Garling J. Sasser, Jr. Tetra Tech, Inc. Tetra Tech, Inc. 3700 Mt. Diablo Blvd. 1911 Ft. Meyer Drive Lafayette, CA 94549 Arlington, Virginia 22209 *currently affiliated with Systems Control, Inc., 1801 Page Mill Road, Palo Alto, CA 94304 easily be modified. The model includes five ver- sions, each of which incorporates a different ver- Abstract sion of the Population Dynamics Module. A global bionomic model capable of evaluating man- The remaining portion of this paper consists of two agement policies for the world tuna resources is sections. Section 2 presents a description of the developed. The model integrates the biological model and its uses and limitations. Section 3 des- and economic forces into an analysis framework. cribes the input requirements and the output gen- It is made up of two major modules: a Population erated by the model. Dynamics Module (PDM) and an Economic Module (EM). 'The model has been coded and tested on a computer. 2. Description of the Model The PDM tracks the variations in the populations of the different tuna stocks throughout the world. General Description of the Model The EM computes the profit-maximizing fishing efforts of the various countries at different The Global Tuna Systems Model is made up of two fishing grounds, given the available standing major modules: a Population Dynamics Module (PDM) crops and a world ex-vessel price of tuna. This and an Economics Module (EM). The PDM tracks the computation of fishing efforts is accomplished variations in the populations of the different tuna using a linear programming model. The model as stocks throughout the world. These variations are currently formulated is particularly useful for due to reproduction, natural mortality, recruit- evaluating the various tuna management alterna- ment, fishing mortality, and migration. tives. It is useful as a screening tool to reduce the numerous tuna management alternatives (and The Economics Module (EM) computes the profit- their combinations) into a smaller set of options maximizing fishing efforts of the various countries for more detailed assessment. at different fishing grounds, given the available standing crops and a world ex-vessel price of tuna. This amount of fishing effort is then translated 1. Introduction into landings, the sum of which comprises the world supply tuna. The demand for tuna by various This paper describes a global bionomic model for consuming nations is next computed. If the supply evaluating management policies for the world tuna matches the demand at the ex-vessel price, then resources. It highlights the functions and opera- this price and quantity will constitute the equil- ting characteristics of the model, its uses and iru uniis tewsteaoecmua limitations, and the attendant data needs. tions are repeated with a different ex-vessel price until approximate equality is achieved between The model, entitled the Global Tuna System Model, supply and demand quantities. has been developed to achieve several goals. It should be capable of delineating the interplays The interrelationship between the PDM and the EM between biology and economics. This is accom- is llustrated in Figure 1. As indicated in the pihd by simulatingthbilgcldnmcoffgrthrarfievrisofhePpain pi~,uthe bilgcldyaicoffe~ ren therearefv vestocks in the wopuldatinde the iffeent una tocs inthe orld andre-Dynamics Module: Generalized Production Model, presenting the economic forces governing the Cohort Model, Schaefer's Model, Leslie Matrix different tuna fisheries throughout the world. Model, and the Birth-Death Process Model. Since It should also be able to evaluate tuna popula- the Shaefer's Model is a special case of the tion dynamics models as management tools. The . Generalized Production Model, in essence there are model should also serve as a framework for organi- only four versions. Each of these four versions zing data bases to be used in the evaluation of is integrated into a specific version of the tuna management alternatives. Economics Module. To accomplish these goals, the model has been de- Formulations of the Model veloped and coded in modular forms to accomodate modifications in the future. It is conceptually The mathematical formulations of the model are Correct but has not yet been fully verified. How- summarized below. It should be noted that the ever, as more data become available and a better population dynamics are formulated in terms of bio- understanding of dynamics is gained, the model can numbers. Translating these into units of biomass 199 aBmt = migration mortality coefficient of species j transferring from PICK A SET OF , fishing ground n to R TUNA PRICES R = transit time to migrate from fishing ground n to t It should be noted that Schaefer's Model is a LP MODEL TO DESSUPPLY . special case of this model, with m._ being set FISHING ssu DETERMINE OPTIMAL DEMAN OF TUNA > equal to two (Schaefer, 1954; 1957).3z FISHING EFFORTS TCH? b. Cohort Model ............. i STANDING CROP INFORMATION ES This is an age-specific population dynamics model with the capability of accommodating up to ten age ,E CR LE BIRTH- . EVALUATE THE classes. It is a variation of the Murphy catch ,ENERUTLIZEO MATRIX DEATH UNITED STATEST , equation (Tomlinson, 1970; Murphy, 1965; 1966). PROORi E~EDFORTUNA IMOGEL MODEL MODEL PROCESS NEED FOR TUNA Migration and fishing mortality are also incorpor- . I . :ODEL:T I~: MD IMPORT . ated into the model. oo......................... .................. .............................. I POPULATION DYNAMICS MODULE ECONOMICS MODULE FIGURE 1 INTERRELATIONSHIP BETWEEN POPULATION DYNAMICS -M (m) N m)(t)- jQ N (m)(t) MODULE AND ECONOMICS MODULE j 9 tr < tc jt m<j can be easily accomplished by multiplying with the (m appropriate average biomass per fish. d() 6j Njn )(t)- N t) dNi(t) jnk jn jj Population Dynamics Module dt a. Generalized Production Model (m) +q (mj ik2. i,k The generalized production model is a vari- (t) (m) (m) ation of the Bernoulli's equation (Chapman, 1967; ijk ij jk Pella and Tomlinson, 1969). In particular, it T also includes the effects of fishing mortality ijk ] and migration. c dP.j(t) mjQ it = H~g Pjz(t) - K. P j(t) (1)M J J Jdt HjR PjR~t) KJa P with N.j (0) = w, i NL t) , spawner-recruit -q EZ Vik fijkr(t) Pm(t) (m=2 relationship i,k ik 3 (ik) it ijkp and Nj.(t+) Nji(t_) h - ' (t- =i AnR) -JR hlwhere 8 E RjmR PjR (t-Ang) - Sjtt Pj ) where(t) where t = time in half years Nhm)(t) = population size (in bionumber) of age P R(t) = bionumber of species j at J class m of species j at fishing fishing ground R at time t ground R at time t HQK.Qm = constants (i) M = natural mortality coefficient Vk = number of fishing vessels of i t type i in the fleet of F (m) = fraction of fishing effort applied country k at fishing to age class m ground R g (.) = spawner-recruit relationship, giving fijk (t) = fishing effort in fishing r = agnumbe r of recruitm fijkk(t) days of boat type i, country t= age of recruitment, i.e., age at days of boat type i, country jQ which the fish first become catchable k, catching species j at t be fishing ground R at time t t+ ginning of year t qjz = catchability coefficient t_ = end of the year (t-l) Ti = conversion factors that ~ijk2 translate fishing efforts c. Leslie Matrix Model into units of fishing days The Leslie Matrix Model (Leslie, 1945) is a vector with respect to a standard difference equation with a ten-dimensional vector, size vessel and type accommodating ten different age classes. The Leslie matrix relates the population size of the 200 stock at one time period to the next time period N t) ber of as shown below: j, 1 2 bionuber of stock j at the 2th fishing ground N (t+l) = A (t) N (t) where Pr {Njt(t+h) = i+1INjH(t) = i) t =time = a + i h+8(h) N (t) = vector of population size in bionumber) hk) (h) of different age classes at time t A (t) = Leslie matrix (10xlO) Pr Nj(t+h) = i-INjt(t) 0 fl f2 f3 ....f' = �0 fl~ f2 f3* fg = (bjt + ijt) h+8(h) PO O O O O pO 0 0 0 . O Pr { Njk(t+h) = i[Nj,(t) = i} 0 pi 0 0 ....0 r =1 -[aj +b +i(.j+p)] h+8(h) o o O . ..P 0_ 0 _ with lim e(h) + 0 h+ 0 fi = fecundity rate of age class i p. = probability of a fish of age i sur- viving to the age (i+l) h = very small time increment = 1 - 0.5 (Mj() + qj() _ Vikg a = immigration rate i,k f() F fecundity rate per spawner ijkt J / bjg = emigration rate iJkk d. Birth-Death Process Model j M + q V f Mit + qik ikg fi k i The Birth-Death Process Model is a variation of ijkT the stochastic process formulation used in epidem- ilogic research (Karlin, 1969). The state of the Economics Module stochastic process N. (t) represents the number of individuals in a stock j at fishing ground To determine the profit-maximizing fishing efforts Q. Its transition from one state to another, the following linear programming (LP) model is representing the population size fluctuation, is used: governed by probability measures related to a Poisson process (Papoulis, 1965). Then the ex- maximize pected population size M(i~t) at time t, given an (P q1 Vikl P1 - bik)fijk_ F- initial mean population size of i, is described i,j,k T by the following differential equation: ijkQ subject to (i) iQ()-bjQ(X) ( )k < min { Bj+ Xj pi dMC~(t (a \ \ Ci)Z q~2.Vikt ~ ~ mm B., Qj. dt it - I1?i j Mjl2(t) ik T bdtt)~ 3tj~ 9, jk ijkZ (i) qjk Vik fijk P < Cikk with M() = i Tik where fijkf > 0 for all i,j,k,t 0where jk Mji~t) : N= ( E NjPik(t) where k=0 Pj = ex-vessel price per ton of tuna species j bk = variable cost per ton of tuna landing Pk(t) = Pr {NjN(t) = kNj(0) = by country k of boat type i Fik = fixed cost per vessel 201 B.j = standing crop of species j at fishing The current model as formulated is also capable of ground Z considering the management alternative of adopting Q a = catch quota for species j at fishing 200-mile fishing zones, except that the memory ground Z core and computer time requirements are prohibi- ik = fleet capacity of boat type i, country tive. The equivalent number of fishing grounds k, at fishing ground must be doubled, and the complete model run using VikW = fleet size of boat type i, counting k at least monthly time intervals instead of the at fishing ground current semi-annual intervals. Moreover, the cost functions must be modified for each month. This LP model is solved for a different fishing ground each time, assuming the fleet sizes V 's The model also has other uses. It serves as a are dispatched to each region based on an framework to guide the data collection procedure investment function. Summing up the landings from needed for evaluating management strategies. It each fishing ground gives the world's supply of specifies data requirements, and thus data collec- tuna for a given set of ex-vessel prices PI's. tion schemes can be designed to minimize extrane- J ~ ~ ous efforts. The following investment function is based on research conducted by Huppert (1976): The model can also serve as a screening tool to reduce the numerous management alternatives (and For the first five years: their combinations) into a smaller set of options. This smaller set of options can next be evaluated V(t+l) = V(t) + C(t) t = 1, 2, 3, 4 in detail, taking into account political, insti- tutional and other considerations. However, the where model should not be construed as the ultimate V(t) = number of vessels in the fleet at only authority on what the optimal management time t policy should be. C(t) = change in the number of vessels for the time period t Limitations of the Model After five years, The current formulation does not give a truly optimal solution. The assignment of fleet sizes V(t) + 20 if A(t) > 0.2 to various fishing grounds is pre-specified at the V(t) + lOOA(t) if O<A(t) < 0.2 beginning of the season. Ideally this should be part of the optimization problem and should be V(t+l) = 0.95 �(t) if -O.l<A(t) < 0 computed concurrently with the fishing effort 0.75 V(t) if -0.2<A(t) <-0.1 optimization. 0.5 V(t) if A(t) <-0.2 The current formulation is not truly optimal for another reason: The LP model is run for one where 5 region at a time. The mathematical formulation, 'a (t-i) of course, is set up to provide the global optimum. A(t) = ZE 3 V(t-i) Unfortunately the large computer core size re- i=2 quirement forces the adoption of suboptimal solutions. -ff (t) = profit for year t The current model also has other limitations. On the biological side, the catchability coefficient For this study, the following demand function is does not take the density-dependency into account. does not take the density-dependency into account. based on the research of Bell (1969). As a result the landing value may be inaccurately n D = cca - n p + -y kn I estimated. The unit stock also need to be properly defined. where An D is the natural logarithm of tuna demand The constants cited in the literature for the by a country, An p the natural logarithm of tuna different population dynamics models may not price, Zn I the natural logarithm of per capita always be appropriate if the unit stock does not disposable personal income in a country, and correspond to the geographical specifications of a,SB, are constants. the model. Uses of the Model An added dimension to the unit stock concept is the migration effect. The incorporation of This Global Tuna System Model is useful for evalu- migration may influence the constants used in ating the following management alternatives: defining the population dynamics. Consequently, a setting of total catch quota on different species new estimation scheme may be in order to properly at the various fishing grounds; setting of differ- discern the interaction between unit stock and ent countries' landing quota at various fishing migration. grounds; regulations on fleet capacity expansions; policy affecting the tuna demand; tuna fleet in- The other major limitation associated with the vestment policies; and regulations on uses of biological aspect of the model is that many of fishing gears. the required parameters are age-specific. 202 Obtaining the data to estimate these parameters growth equation (1938) and the weight-length may be difficult. relationship; mortlaity rates associates with migration; spatial distribution of tuna species; Another major limitation related to the biological catchability coefficient; fishing power conversion aspect of the model is the lack of reliable esti- factors; spawner-recruit relationships; and mates of the fishing power conversion factors. fecundity rates. Such estimates should be able to take into account the catchability of different stocks by The output consists of two groups of information. the various types of fishing gear. The first group pertains to the standing crop. It shows the biomass and bionumbers of different On the economical side, one major limitation of tuna stocks at various fishing grounds. the model is that perfect competition has been assumed in the model. But the real world is an The other group of output is related to the imperfect competition market. A handful of tuna- catches. A tabular output summarizing the tuna consuming or -producing nations can affect the catches by various types of boats owned by dif- market drastically. Consequently, the Economics ferent producing nations is given for each fishing Module should be modified in the future to accom- ground at each half of a year. At the end of all modate this consideration. these tabular outputs is a summary of the world's tuna market indicating the total catches in metric This model also assumes that the objective of all tons segregated by species, countries, and types the producing nations is to maximize the total of fishing vessel. profits. But each nation may have different ob- jectives (e.g., equity, employment, foreign ex- change), so that the objective in this model may References need to be modified for the proper management of the fishery. Chapman, D.G., 1967. "Statistical Problems in the Optimum Utilization of Fisheries Resources", The adoption of this objective also has the im- Int. Stat. Inst. Bull., Vol. 42, No. 1, pp. plicit assumption that an international tuna 268-290. management council exists which has jurisdiction over all the global tuna resources. In this case Huppert, D., 1976. Personal communication. the "council" decides that profit maximization is the best for all nations involved. In reality Karlin, S., 1921. A First Course in Stochastic such a "council" does not truly exist. Conse- Processes, Academic Press, 4th ed. quently, the Economics Module may benefit from a modified formulation. Leslie, P.H., 1945. "On the Use of Matrices in Certain Population Mathematics", Biometrika, The model also utilizes a relatively primitive Vol. XXXIII, Part III, November, pp. 183-212. formulation of the demand function. Interactions among the demands for other tuna species have not Murphy, G.I., 1965. "A Solution of the Catch been fully taken into account. Other important Equation", J. Fish Res. Board of Canada, Vol. 22, considerations like the price of substitutes have pp. 191-202. not been incorporated. Moreover, the demands of nations other than those of the United States Murphy, G.I., 1966. "Population Biology of the cannot be accurately estimated due to the lack of Pacific Sardine (Sardinops Caevulea)", Proc. supporting data bases. These demands were rather Calif. Acad. Science, Vol. 34, p. 84. crudely extrapolated from the United States demand in this project. Future refinement of this demand Papoulis, A., 1965. "Probability, Random formulation is in order. Variables and Stochastic Processes", McGraw-Hill. 3. Input and Output Pella, J.J. and P.K. Tomlinson, 1969. "A Generalized Stock Production", Inter-Am. Trop. There are two major groups of input data: fishery Tuna Comm. Bulletin, Vol. 13, pp. 419-496. and population dynamics related. The fishery re- lated inputs include: number of boat types, Tomlinson, P.K., 1970. "A Generalization of the species, tuna-producing nations and fishing Murphy Catch Equation", J. Fish. Res. Board of grounds; definition of boat types, species, tuna- Canada, Vol. 27, pp. 821-825. producing nations and fishing grounds; fishing fleet size; variable and fixed costs associated von Bertalanffy, L., 1938. "A Quantitative with each type of fishing vessel; changes in Theory of the Organic Growth", Human Biology, fleet sizes; and management alternatives (e.g., Vol. 10, pp. 181-213. catch quota, legal size limitations, fishing gear limitations). The population dynamics related inputs include the following categories of information (Age-specific data may be required for some versions of the model.): Initial population size; natural mor- tality rate; constants for the von Bertalanffy 203 A MICROCOMPUTER-BASED ECHO-INTEGRATION SYSTEM FOR FISH POPULATION ASSESSMENT Raynard Y. Kanemori and John E. Ehrenberg Applied Physics Laboratory University of Washington 1013 N.E. 40th, Seattle, WA 98105 ABSTRACT using minicomputers. 3,4 Although minicomputer- based systems have great storage and processing power, they are very expensive and normally data Abundance estimates obtained with acoustic can not be processed in real time because of their echo-sounding equipment play an important role in size. A tape recorder is used to collect the fish the management of fish stocks. The common tech- data in the field and the data are then processed nique presently used is to process the signal from on the minicomputer-based echo-integration system the echo sounder with a digital computer or to at the laboratory. record the analog signal for later processing. With the advent of microprocessor technology, A microcomputer-based echo-integration system an echo-integration system can be built that has has been developed for fish population assessment. the same processing power as a minicomputer-based It is a relatively inexpensive, field-oriented, system. A microcomputer-based echo-integration portable instrument that can process data in real system with limited capabilities has previously time from any existing echo sounder. The signal been developed by Wiebe and Stevens.5 The system from the echo sounder is detected, digitized, discussed here is an extension of past work on squared, and averaged for various depth ranges. echo-integration systems done at the University of The averaged squared echo level is then scaled to Washington. The system calculates the averaged provide a measure of the density of the fish squared echo voltage levels for up to ten differ- school at the selected depth ranges. The system ent depth ranges. The digital integrator output automatically tracks the bottom depth and prevents for a given depth range, Rl to R2, is integration of the return from the bottom.N 1. INTRODUCTIONRlR il Recently, there has been an increase in the where V)2 is the squared sample values of the de- use of acoustic techniques for fish population tected echo signal and N is the number of samples assessment. The two common acoustic techniques processed for the depth layer. The output is used to estimate fish abundance are echo counting later scaled to provide an estimate of the density and echo integration. In echo counting, the den- of the fish school at the selected depth ranges. sity is obtained from the acoustic return by a The scale factor relating the integrator output to processor that counts the number of fish echoes. fish density, A, is split into two parts. That Echo counting provides a good abundance estimate is, for low-density fish populations. However, it can not be used for high-density fish populations be- X = A-B-I RlR2 cause the echoes from the individual fish overlap. The echo-integration method does not have this weeAi osatta eed nprmtr limitation. The basic principle behind echo inte- thatere noAet epnet is a constant thatdensonprmts gration is that the echoes from individual fish de ed ontetm-aidgincaatrsiso add icoheentl.' I othe wors, te avragethe sounder and the depth range being processed, value of the squared signal scattered from a fish and IRI,R2 is the digital integrator output for school is proportional to the number of fish. An the depth range from RI to R2. echo-integration system calculates the average squared echo sounder output for various depth lay- 2. HARDWARE DESIGN ers. The averages are then scaled to provide estimates of the fish density. A block diagram of the hardware for the echo-ntegationsys-microcomputer-based echo-integration system is A number of different echo-nintFigure Teanalog signa-poesn tosytems have been feeound. Anaog bechoinacrterand hardware includes an amplifier which increases tioneystm have inaeqnfuadtoe dynamccurange, cmared wihthe amplitude of the input signal from the echo dgtaleycho-einaeutegdnmcratingssem Sompaedigita l sounder, and an envelope detector which consists dgtlecho-integration systems. have benimplmeted of a rectifier and a low-pass filter. These cir- echo-ntegatio sysems hve ben iplemntedcuits were implemented with operational amplifiers. 204 --- - - - - - - - - - - -I ~I ~ MICROCOMPUTER-BASED ECHO-INTEGRATION SYSTEM I ------- I ANALOG SIGNAL PROCESSING SYSTEM 1 I I AL-WV LOW-PASS SPMAPLE- AID S S NDER FEC TIFILTER HOLD CON VERTE4 -------------- S ITRIGGER INCONTROL ,BOTTOM PULSE DATA I LOST- TM990/ I BOTTOM -lOOM I ALARM MICRO- ALPHANUMERICIl COMPUTER I DISPLAY I I | PRINTER TRIGGER TKEY- TRIGGER OUTI :SCO~PE I SCOPE I DETECTED OUTPUT Figure 1. Block diagram of the microcomputer-based echo-integration system hardware. The signal from the analog signal-processing that the echo return from the bottom is not inte- hardware is digitized by a sample-hold integrated grated. A bottom alarm is set off when the bottom circuit and a 12-bit analog-to-digital converter. pulse is lost. The signal is sampled at a rate of 7.4 kHz by an asynchronous clock. The converter has a short- 3. SOFTWARE DESIGN cycled conversion time of 6 Us because only 10- bits are being used. This leaves about 129 Us for A block diagram of the software for the the real-time integration process. microcomputer-based echo-integration system is shown in Figure 2. The initialization of the The microcomputer used in the present system software starts the interactive parameter-entry is a TM 990/100M, which is a single-board computer program. This program asks the user to input the based on the 16-bit TMS 9900 microprocessor. The parameters for the integration program through the board has a maximum of 512 words of Random Access keyboard. The questions include (1) the duration Memory and 4000 words of Erasable Programmable of the run, which is set by the number of pings; Read-Only Memory as well as programmable serial (2) the normalizing constant, which converts the and parallel input/output interfaces. If more final integrated value to fish density (this capability is needed, the memories and I/O can be constant depends on the target strength and acous- expanded. The TM 990/100M runs at a clock rate of tic system parameters such as source level, gain, 3 MHz. There are 15 external maskable interrupts. transducer directivity, and receiving response); The board also has hardware multiply and divide. (3) the threshold to eliminate unwanted noise; (4) the depth intervals; and (5) the time-varied gain The input devices consist of a keyboard to corrections. Up to 10 depth intervals can be input the parameters for the integration program processed with this system. The integration and a number of switches that interrupt the com- program begins when the last parameter is entered. puter to perform various tasks required by the This program strobes the sample-hold to begin the user. These parameters and tasks are discussed in data acquisition process. It then checks the A/D the software section of this paper. converter to see if the sample has been converted. Digitization takes 6 Us. After the sample has The output devices consist of a portable been digitized, the data are transferred to the printer which lists the parameters and final in- microcomputer. The sample is squared and accumu- tegration results and a 16-character, alphanumeric, lated. The accumulation is done using 32-bit light-emitting diode display which is used for the double-precision addition. The depth of the same purpose. An oscilloscope displays the signal acoustic return is determined by keeping track of that is being integrated and the depth ranges that the number of samples processed following the are being processed. The bottom depth is tracked trigger pulse. The number of samples is directly automatically by the microcomputer software so proportional to the depth. For a 7.4-kHz sampling rate, a depth interval of 1 m contains 10 samples. 205 INTEGRAITION BOTTOM INPUT INTERRUPT INPUT/OUTPUT PROGRAM TRACKI PARAMETER ROUTINES ROUTINES ENTRY ROUTINE DATA SIGNAL -PAUSE SWITCH PRINTER S ACQUISITION AVERAGING ALARM RRESET BUTTON DRIVER 'B CONSTANT SWITCH D CORRECTION PARAMETER CHANGE CONSTANT SWITCH DENSITY CALCULATION *-FORCE OUTPUT CALCULATION SWITCH ALPHANUMERIC KEYBOARD ROUTINE -DUMP SWITCH DISPLAY DRIVER -DIAGNOSTIC SWITCH VER FLOATING POINT PACKAGE Figure 2. Block diagram of the microcomputer-based echo-integration system software. The bottom is tracked automatically by the microcomputer. A bottom pulse from the echo sounder sends an interrupt signal to the micro- computer to begin the bottom-tracking routine. A time window is created around the interrupt sig- nal, and the next bottom pulse is checked to see if it is within the time window. If it is, the window is centered about the present bottom pulse. If a bottom pulse falls outside the window, the integration process stops and the microcomputer waits for the next trigger pulse. If the bottom pulse is outside the window again, an alarm is sounded. After all the samples are accumulated, the average value is calculated for each depth range by dividing the accumulated totals by the number of samples. The results are then scaled to pro- vide a measure of the density of the fish school. Figure 3. Thze prototype unit. Each switch on the front panel has a task. Figure 3. The prototype unit. If a switch is turned on, it will send an inter- rupt to the microcomputer. The microcomputer will go to the routine selected by the switch and per- for mass storage of the digital data for later form the task. A pause switch is used to stop the processing. Floppy disk systems, cassette tape integration temporarily. The diagnostic switch is recorders, and bubble memory boards are some of used to check the functioning of the keyboard, the alternative storage devices. Another future printer, display, and memory. The force output improvement will be to calculate the time-varied switch is used to stop the integration at any time gain constant and the final density in real time. and print the results. The parameter change This could be done by hardware built around an switch is used to change parameter values in the integrated-circuit floating-point processor, or by integration program. The dump switch is used to software written in FORTRAN and using the stand- reset the accumulators for each depth layer to alone FORTRAN library. If the software approach zero. is chosen, additional memory must be incorporated into the present prototype. 4. CONCLUSIONS The microcomputer-based echo-integration sys- A prototype unit has been constructed, and is tem appears to have much potential. This system shown in Figure 3. The unit is currently being should provide an accurate, inexpensive, and port- tested in the laboratory with tape-recorded data. able method for the real-time estimation of fish As soon as the laboratory testing is completed, stock size. the unit will be ready for field testing. Several improvements can be made in the future to increase the capabilities of this sys- tem. One improvement might be to provide a method 206 5. ACKNOWLEDGMENTS 3. R.E. Thorne, "A New Digital Hydroacoustic Data Processor and Some Observations on The research described in this paper was Herring in Alaska," J. Fish. Res. Board sponsored in part by the Washington Sea Grant Can., 34:2288-2294, 1977. program under National Oceanic and Atmospheric Administration Grant 04-7-158-44021. The authors 4. P.H. Moose, J.E. Ehrenberg and J.H. Green, would like to thank Dr. Gregory Zick of the Uni- "Electronic System and Data Processing versity of Washington Department of Electrical Techniques for Estimating Fish Abundance," Engineering for the use of his microprocessor Proc. IEEE Conf. on Eng. in the Ocean, development system during the software development. Vol. 1, pp. 33-36, 1971. REFERENCES 5. A.J. Wiebe and D. Stevens, "Digital Integrator Enhances Echo Sounder's Accuracy" in Applying 1. J. Ehrenberg and D. Lytle, "Acoustic Tech- Microprocessors, L. Altman and S. Scrupski, niques for Estimating Fish Abundance," IEEE eds., McGraw-Hill, pp. 145-146, 1976. Trans., Vol. GE-10, pp. 138-145, 1972. 2. H.W. Lahore and D.W. Lytle, "An Echo Inte- grator for Use in the Estimation of Fish Populations," Washington Sea Grant Publ. (70-1), 1970. 207 PROSPECTS FOR THE DEVELOPMENT OF A SWATH FISHING VESSEL Mark S. Rice Jack Harmon Advanced Design Staff Systems Engineering and Analysis SEACO, Incorporated SEACO, Incorporated Kailua, Hawaii 96734 Kailua, Hawaii 96734 ABSTRACT The modern fishing fleet has adopted multiple fishing techniques for survival. A single vessel is equipped to deep trawl, shallow trawl, long-line and even set pots for lobster and crabs. In this way, today's fisherman is no longer restricted to the fortunes of the traditional one-technique fishing boat. Data are presented on the variable requirements for the Central Pacific industry and the type of vessel that is needed. The introduction into modern ship design of the Navy's SWATH* concept provides the industry with an advanced craft that has a wide expanse of usable deck and work spaces and extraordinary stability for safer, more effective deployment of equipment. The numerous features of the SWATH design are discussed, and aspecific, though necessarily preliminary, design is presented for an advanced Semi-submerged Fishing Vessel (SSFV) that will be constructed in the early 1980's. FIGURE 1. SEACO's SSFV Concept INTRODUCTION REQUIREMENTS FOR MULTI-SPECIES FISHING VESSELS SEACO, Incorporated has initiated a program in Hawaii that will culminate in the construction of a Semi-submerged In seeking a profitable return on investment, the modern Fishing Vessel (SSFV). The program calls for the cooperation of fisherman faces both economic and legislative restrictions. The several state, federal and commercial concerns in developing economic restrictions are generated by the market via the and building a vessel that can economically fish the Central influence of intermediates in the marketing structure; the Pacific. The program is unprecedented in that it involves a modern fisherman can do little to influence either the cooperative government/commercial effort to match an intermediates or the consumer. Furthermore, he has limited Pacific. The progrm is unprecedentedintermediates or the consumer. Furthermore, he has limitedI c ead vanced technologyr to present fishing needs. control over legislative restrictions since they often stem from environmental limitations or international political pressures. The present schedule calls for design of the SSFV to be To remain profitable in the face of these variable restrictions, the completed in 1979. Construction of the first of five initial vessels modern fishing vessel must be VERSATILE-the vessel's return is to begin in early 1980. Construction will commence on on investment can no longer depend on a single species of fish or a subsequent vessels at four to six-month intervals. Inquiries have single fishing technique. As noted by Blount and Schaefers been received from both scientific and industrial organizations (Reference 1) more than a decade ago: that are interested in the unusual advantages of this type of "There is a growing trend in the USA for combination- vessel. type fishing vessels to be built in areas other than the Pacific coast. This trend has been brought about by The SSFV is an outstanding example of the military the realization that by building combination-type technology transfer to the civilian sector. The SEACO advanced vessels, fishermen are able to harvest several types of fishing concept, Figure 1, is a close relative to the Navy's Stable fish, each of which may require a different type of Semi-submerged Platform (SSP) KAIMALINO; its intro- gear. This provides for year-round use of vessels and duction into other industries is long overdue. The benefits to removes the risk that one-purpose boats always face, be derived from such a vessel are discussed in numerous Nairy dependence on a single type of fish, the abundance and publications and are summarized in this report as they pertain to availability of which may fluctuate greatly." the fishing industry. Since the Blount and Schaefers article, the availability of many commercially important species has declined and today the economic pressures in many fisheries are more severe than at *Small Waterplane Area Twin Hull any time in the past. In contrast, recent legislative action has given the U.S. fisherman an unparalleled opportunity to compete with foreign boats in U.S. coastal waters via the 208 recently enacted 200-mile territorial limit. The result of the relaxation in foreign competition is manifest in several present POUNDS VALUE conditions, namely: ISLAND AND SPECIES CAUGHT (DOLLARS) * Increased speculative capital for new vessel design and Total landings 11,891,275 7,503,964 construction * Increasing competition among U.S. vessels, and ISLAND * Growth of larger conglomerate fishing organizations Hawaii 2,361,850 1,754,979 Thus, there are two sources of pressure in today's U.S. Maui 750,414 277,209 commercial fishing industry, both of which suggest that Lanai 6,990 3,178 technological advances are imminent, namely: Molokai 35,761 27 175 Oahu 8,388,893 5,255,543 * Increased economic pressure promoting versatility Kauai 347,367 185,880 * Increased potential profits resulting from 200-mile restrictions SPECIES The Central Pacific fishery represents a special case of the Sea catch, total2 11,871,076 7,486,272 general U.S. fishing industry. That is, the 200-mile territorial Aku (Skipjack) 6,891,039 2,911,061 limit, when applied to the Hawaiian chain alone, results in more Ahi (Yellowfin) 1,723,128 1,463,376 than 800,000 square miles of rich fishing grounds that are now Akule (Bigeye Scad) 746,857 442,211 under U.S. jurisdiction. Although the region is in a temporate Ahi (Bigeye) 425,537 845,188 climate, sea conditions are generally bad throughout the year. Opelu (Mackerel Scad) 291,337 231,236 The fishing within this region is characteristic of deep open- Striped Marlin 230,412 113,243 ocean fishing. Species sought commercially within this region Opakapaka (Snapper) 147,505 165,195 range from crustacea to pelagic. Fishing in the vicinity of shoals Ono (Wahoo) 132,105 90,448 and reefs is commonplace while the long-line is a principle Mahimahi (Dolphin) 119,332 145,010 technique in open-ocean work. A general summary of species Ahipalaha (Albacore) 104,942 93,311 sought and landed in Hawaii is given in Figure 2. This total is Ulua (Jack) 93,771 74,542 estimated to be a very small percentage of the landings that the Ulaula Koae (Onaga) 80,543 162,219 region could ecologically support (Reference 2). Furthermore, Pacific Blue Marlin 73,582 22,108 this list of fish that are presently soughtonacommercial scale by Uku (Snapper Family) 69,832 68,321 U.S. vessels does not include significant pelagic catches. The Kawakawa (Bonito) 66,669 18,215 exploitation of open-ocean catches is currently being carried out Limu 62,211 36,873 only by foreign vessels that are operating under quotas Hapuupuu (Grouper) 57,478 66,920 established by the U.S. The U.S. vessels capable of the extended Kahala (Amberjack) 40,434 21,822 open-ocean fishing simply do not exist. Rather, the Hawaiian Weke (Goldfish) 37,086 21,673 fishing fleet consists of a myriad of old designs that are Ulaula (Ehu) (Snapper) 34,310 58,411 interspersed with a handful of more modern vessels, all of which Crab (Kona) 26,577 55,900 are under 125 feet in length. Pond Catch, total 20,199 17,692 Based on a knowledge of the resource, the shortcomings of 1Shown separately for all species over 35,000 pounds or $50,000. existing fishing vessels, and the current economic situation, one is able to project a conceptually optimum vessel for work in the 2Includes species not shown separately. Central Pacific. The major characteristics of such a vessel are discussed below. Source: Hawaii State Department of Land and Natural Resources, Division of Fish and Game, Commercial Fish Landings for 1. Range Fiscal Year July 1975 through June 1976 (October 1976). The principal fishing port in the Central Pacific is Honolulu. From this port, a vessel may transit 1500 miles to the FIGURE 2: Commercial Fish Landings by Species and Island: N.W. to the nearest refueling port, Midway Island. To the Year Ended June 30, 1976, reproduced from the southwest. American Samoa represents the closest island of Hawaii State Data Book - 1977. commercial significance at a range of more than 2200 miles. To the west, Guam and the scattered islands of the Pacific Trust vessel of some minimum size to weather the sea conditions that Territory are all more than 1500 mile from Honolulu. Since the will be encountered. Similarly, the vessel must have a catch straight-line transits to these ports exceed, in all cases, 1500 holding capacity to make a 3,000 to 4,000 mile voyage profitable miles, the working range of vessels fishing in these areas should unless a mother ship is to be used for the transport of fuel, stores, be at least 4000 miles at design speed or 2000 miles at design crew, etc. The upper limit to vessel size will be governed both by speed with an additional range of 4000 miles at fishing speeds. the density of the species sought (operating costs per unit time Corresponding to these distances are deployment durations that must be well below the wholesale value of the catch brought on range from three weeks to two months, depending upon fishing board per unit time) and by the capital available for vessel techniques and the catch found. construction. Thus, an upper and lower limit to vessel size may be defined with relative ease while optimum size remains a 2. Displacement Tonnage function of numerous economic and ecological considerations. Displacement tonnage can be roughly correlated with 3. Speed overall vessel dimensions. The optimum dimensions for a multi- species, Central Pacific fishing vessel will be related to the The optimum average speed for a Central Pacific fishing economics of the vessel and its catch. It will be necessary to have a vessel is a function of five basic competing variables as shown in 209 Figure 3. It is apparent that drag will escalate as a function standard hull configuration. Such assumptions are routinely of the velocity squared via the basic equations for vessel drag and made since basic hull shape/design has been standardized propulsive power: through many years of convergent design practices. Although size may vary from region to region, the basic design shapes of D = C, '/2 pV2A fishing hulls are remarkably similar throughout the United States. To some extent, the catamarab hull was a new approach where: D =total drag to hull design for fishing vessels. Unfortunately, the catamaran has three basic characteristics that have limited its acceptance Cd total drag coefficient as a commercial fishing vessel, namely: p =the mass density of sea water * The increased capabilities do not offset increased costs V =velocity * The concept's sharp motions make deck work hazardous A =characteristic area of the vessel a It suffers from structural problems that result from P = (DV)/77 ~~~~~~wave-impact loading In this paper, a radically different but already proven design is where: P = power discussed with respect to its applicability as a fishing vessel. = propulsive efficiency THE SMALL WATERPLANE AREA TWIN HULL (SWATH) CONCEPT 4X The predecessors of the modern SWATH concept were early platforms designed without propulsion for stability where 3 ~~~~~men worked in waves. From this initial concept, the idea evolved 21 fLbrplsSot in two directions, the large drilling platform and higher-speed C., o C-t-truco and Capioul vessels such as the D UPL US (Netherlands), the K-4IMALINO SX Cooso Fuel (U. S. Navy), and the recent Mitsui vessel, MARINE ACE 52LR (Japan). These vessels may take various forms as shown in PER ~~~~~~~~~~~~~~~Figures 4 and 5. MIEL ~~~~~~~~~~~~~~~~~~~~~~~~~~25 0 5 10 IS 20 AVERAGE SPEED OF THE VESSEL(IN KNOTS, (AVERAGED OVER 20 YEAR LIFE) FIGURE S. Semi-Qualitative Illustration of Fishing Vessel Revenue to Cost Ratio The labor costs, expressed in terms of dollars per mile, will decrease as the average speed of the vessel increases. Similarly, the initial capital cost of the vessel amortized over the life of the vessel will decrease with increasing speed to the point where it costs proportionately larger amounts of initial capital to achieve the higher average speeds. The revenue (or value of catch) per FIGURE 4. SSP KAIMALINO U. S. Navy Prototype mile will be relatively independent of speed except as increased SWATH Ship speed may lead to increased fishing efficiency (as in the case of trawling for fish that can escape the trawl). The curves shown inJ Figure 3 are qualitative and thus represent only the general The U.S. Navy's introduction to the semi-submerged ship trends for each variable. The specific analysis for an individual concept came from Dr. Thomas Lang of the Naval Ocean vessel will be a function of many variables that characterize the Systems Center (NOSC). The concept combines two torpedo- vessel and its operation; however, these curves do illustrate the shaped lower hulls and a box-like upper hull using four need for an iterative approach in determining fishing vessel streamlined struts for support. The vessel was designed by the revenue to cost ratios. NOSC Hawaii Laboratory and is currently home-ported in Kaneohe, Hawaii. Although this vessel was designed for military Although there are many other variables to consider in use and does not have many design features that would make it optimizing a Central Pacific fishing vessel, they may all be suitable as a fishing vessel, it does have certain basic considered subordinate to the three principle variables: speed, characteristics (Reference 3) that are representative ofa considerations such as: crew size, propulsion type, storage paragraphs. method, etc. However, the discussion to this point has assumed a 210 4. Internal Load The SWATH concept has been criticized as one that is load limited; in fact, if high operating speeds are required, then any design becomes load limited due to sustained limitations on the practical size of power plants that may be utilized. However, drilling platforms are not considered to be load-constrained vessels and moderate-speed SWATH vessel designs share this attribute. If, for example, we consider the displacement of two cylinders with a radius of 6.5 feet and a length of 80 feet as compared with two cylinders with a radius of 8.5 feet and a length of 104.6 feet, one sees a displacement change of 211 short tons (171 short tons as compared with 381 short tons) or 123 percent. Thus, a very large change in displacement is characteristic of relatively small changes in lower hull dimensions. For moderate speed vessels, the additional power required is a function of increases in the projected frontal area of the vessel. The drag penalty is not a strong function of displacement, and, as a result, moderate-speed SWATH vessels are not payload constrained from a design standpoint. FIGURE 5. Trident "DUPLUS" Ship Concept As the price of petroleum fuels increases, it is increasingly critical that the energy expended be minimized. Along these lines, the SWATH vessel may be conceptualized to tailor the hull 1. Stability and Seaworthiness 1. Stability and Seaworthiness geometry to the load onboard. For instance, flexible boundaries may allow fuel compartments to be used for catch storage, or Perhaps the greatest advantage offered by the SWATH collapsible appendages to the hull may further increase the concept is its seaworthiness and stability under all weather, speed and heading combinations. As with a monohull, the storage capability (at a corresponding penalty in drag). In this s peed andheadingcombinations.A with a mon themanner, the hull retains its minimal drag characteristics under natural period of the vessel in a given seaway will depend upon all loading conditions. Some interesting alternatives include its external loading from deck gear, over-the-side loads, etc. towin g devices of minimum dr ag shapes such as a semi- However, it may be stated without exception that the natural submersible barge or devices that can be stow ed onboard when submersible barge or devices that can be stowed onboard when periods of SWATH vessels in roll, pitch and heave are much periods of SWATH vessels in roll, pitch and heave are much not in use. Still other alternatives include the Soviet's factory longer than corresponding quantities for similar monohull ship concept or some form of transport/mother vessel which will vessels. Furthermore, the motions are highly regular with near- allow the SWATH vessel to remain on-station. allow the SWATH vessel to remain on-station. critical damping. Thus, the motions of a SWATH vessel in a seaway are in complete contrast to the motions of a catamaran 5. External ading which experiences short-period, sharply-damped motions. 2. Static StabiliThe SWATH concept has the capability to compensate for 2. Static Stability list created by off-axis external loads. This is accomplished with integral ballast tanks in each submerged hull. These When Dead-In-the-Water (DIW), a SWATH vessel has compartments may contain fuel in appropriate bladders, the stability attributes that have made the semi-submerged r efrigerated cargo or ballast air/water. By employing a drilling platform a success. Waves pass between the upper and KAIMALINO style ballast s ystem, the entire vessel may be lower hulls with little change in the vessel's displacement. The KIAIOsyeblatsseteetr eslmyb lower hulls with little change in the vessel's displacement. The raised and lowered in the water to meet specific freeboard beam of the SWATH vessel is not tied to overall displacement as in a monohull; thus, the beam may be a large fraction of the requirements. If, for instance, a stern trawl is to be brought onboard, the vessel may be deballasted to ensure that the cod end length, giving the vessel similar response to head and beam seas. does not rupture during recovery. does not rupture during recovery. The lower hulls and control surfaces provide large drag in Due to the spacing of the lower hulls, a moon pool or open vertical motions. This drag is the source of damping that makes w ell is easily i ncorporated in the SWATH design. Such a pool can well is easily incorporated in the SWATH design. Such a pool can SWATH motions gentle. Upper hull sections may be faired to SWATH motions gentle. Upper hull sections may be faired to be located at the intersection of the fore-and-aft and side-to-side dissipate the energy of wave impact should amplitudes exceed axes of motion. It may be used in recovering objects and/or the critical clearance between the design waterline and the fishing equipment when seas are rough and relative motion fishing equipment when seas are rough and relative motion upper hull. When DIW, the SWATH ship will turn broadside to between the SWATH and incoming objects must be minimized. between the SWATH and incoming objects must be minimized. wind and/or waves in a manner similar to that experienced by a monohull. Due to the large beam and characteristics of the hulls, 6. Survivability 6. Survivability the broaching problems of a SWATH ship are much less severe than those of a monohull of equivalent displacement. The SWATH concept entails a relatively large number of integral compartments. In the event of hull rupture, it is unlikely 3. Dynamic Stability that a significant loss of buoyancy would result. If both lower hulls should flood, the vessel will be supported by the upper hull. The stabilityaa of a SWATH vessel when underway in a Given a SWATH vessel of 100-125 feet in length, it is anticipated seaway is a function of both the basic semi-submerged hull that a vessel so damaged still could survive in sea state 7. that a vessel so damaged still could survive in sea state 7. design and the control that may be imparted by submerged control surfaces. Such control surfaces are hydraulically actuated and may be controlled manually or by electrical signals 7. Maneuverability generated by simple heave, pitch and/or roll gyros. Although not The characteristically large beam of SWATH vessels necessary for moderate speeds, these surfaces improve the Tecaatrsial ag emo WT esl neesse'ary for moderate speeds, these surfaces improve the allows a wide propeller spacing. This results in the potential for vessel's response beyond that attained by the hull configuration large turning moments due to the aplication of differential alone. ~~~~~~~~~~~~~~large turning moments due to the aloolication of differential alone. 211 thrust between propellers. A SWATH vessel can easily pivot on decks nor a large sheer in the deck. There is a large deck area its own axis or, while underway, can be maneuvered by that provides more open space for crew members on extended propellers alone. When turning at high speeds the SWATH voyages. Since the enclosed area of the upper hull is large and vessel exhibits a natural tendency to bank into turns in contrast free of equipment/machinery, there is ample area for recreation. to the motion of turning monohulls. These phenomena make work on deck relatively safe at any speed or sea state. 10. Deck Versatility 8. Speed and Powering The box-shape upper hull structure of the KAIMALINO is conducive to a wide variety of deck configurations. Since there is The drag on a SWATH hull is similar to that on a monohull no sheer in the deck, cables and lines may be buried in the deck to of similar displacement until the vessels are compared in a prevent injury from line parting. Deck equipment may be strong seaway. The small waterplane area has the advantage of mounted on semi-permanent mounts, rails, or may be fixed to the reduced drag in large seas. The extent of this drag reduction deck. The large open deck area makes fairleading much easier. relative to a monohull is not well documented for actual SWATH Trawl drums and net handling gear may be located a safe vessels although a large amount of model data exists (Refer- distance from the ramp (where surges have taken the lives of ~~~~~~~~~~~ence 4). ~many fishermen). Cranes, catheads, winches and miscellaneous ence 4). deck hardware may be mounted on pallets that interface directly Propulsion on existing SWATH vessels includes gas with the ship. In this manner, a vessel may come alongside and turbines and diesels. Propellers are located at the aft end of the change her entire fishing configuration in a short period. lower hulls, thus providing a very favorable cavitation number. Similarly, entire container modules may be "plugged" into the The propulsive efficiency is further increased by the symmetric deck. Such modules could fit within the upper hull or could rest boundary layer that is drawn into the propeller. Under average on deck. As demonstrated by the U.S. Navy's SWATH vessel, the sea states SWATH propeller efficiencies will be 5% to 10% better concept also provides an excellent helicopter platform at sea, a than monohull propeller efficiencies. An indirect benefit of useful tool for modern fish spotting. SWATH powering in a seaway is the capability of maintaining design speed in high sea states. Monohull vessels thus have three AN SSFV DESIGN FOR THE basic problems in a seaway with regard to speed and powering, CENTRAL PACIFIC namely: Through studies of the literature and discussions with (1) The propeller efficiency decays as sea states rise due fishing interests in Hawaii, a program to determine specific to the entrained air that is drawn into the propeller. design is being formulated for a Central Pacific SSFV. Final conceptualizations of candidate SSFV vessels will be dependent (2) The overall drag of the vessel increases significantly as on the outcome of the design studies; however, specifications for the wave loading on the hull increases, two preliminary concepts are shown in Figure 6. The 300 ton concept has been illustrated in Figure 7. This design has been (3) Severe heave, pitch and/or roll motions may force a generated around the 15 knot design speed that is presently reduction in speed since the wave-induced loads on the considered to be economically optimum for many Central Pacific hull could cause structural failure. fishing vessels. Also shown in Figure 6 are the anticipated costs to build and operate both the 300 and 450 ton vessels. The costs have been expressed in terms of dollars per mile such that they A SWATH ship is relatively immune from these problems. may be compared with the generalized argument shown in The depth of the propeller centerline is deeper than on a Figure 2. Although this method of estimating operating conventional vessel of equal displacement; this produces a more economics has a number of inaccuracies, it provides a favorable cavitation number. In addition, the lower propeller conceptually straightforward estimate of the breakeven point centerline and the reduced vessel motions greatly reduce the risk for the vessel. This projection for a 450 ton SSFV yields a of the propeller emerging from the water. The drag of the breakeven point of 18�/lb. of catch while the projection for the SWATH ship is not a strong function of sea state; rather it is a 300 ton SSFV yields an estimate of 12�/lb. It is evidentthatthese direct consequence of the small waterplane area of the struts. estimates are a strong function of the initial assumptions. Thus, Reduced motions and initial structural design characteristics the analysis of the economic viability is subject to revision as make speed reductions in most sea states unnecessary. Only these initial assumptions become established fact. extremely large or steep waves might impose a speed reduction and even then, adjusting the attitude by shifting ballast can As a vessel ages, the fuel, labor and miscellaneous costs compensate for the condition without slowing, will increase with inflation. However, the construction and capital costs have been amortized over the vessel's life. If the Thus, SWATH vessel offers little improvement in drag retail cost of fish follows inflationary trends, then the overall displacement relative to a monohull on a flat sea although it does ratio of revenue to cost will improve with vessel age since the exhibit superior drag/displacement characteristics in a seaway. initial construction/capital cost remains constant. ECONOMIC CONSIDERATIONS 9. Habitability A rule of historic inflation dictates that no new design can The habitability of a SWATH vessel is superior to any be as inexpensive as one that currently exists. And, this is true monohull of equivalent displacement or even several times except in those rare cases involving a breakthrough in larger. This characteristic is a result, in part, of the reduced technology or an inordinate saving through a value engineering magnitude and longer period of roll, pitch and heave motions. achievement. And while the SSFV indeed represents a break- Comparative studies between several monohulls and a SWATH through in technology, it undoubtedly is going to suffer from vessel have demonstrated a significant reduction in seasickness the inflationary cost of tooling up for a new design. of personnel on SWATH vessels (Reference 3). The geometry of the upper hull provides an increase in the usable space, a Informal discussions with some of the European reduction in the ambient humidity, and does not have multiple Shipyards indicate that they believe the cost of a semi- 212 submersible design to be in the same ballpark with that of a similar-sized monohull. This is possible when one considers the design economies of the easy-to-fabricate shapes (round lower hulls and box-shaped upper hull) with the complex monohull components. It is highly likely that a production line of semi- submersible vessels will enjoy additional economies over a similar line of monohulls; obviously the lower hulls and struts lend themselves to certain mass production techniques. There are additional economies (operational) due to the : - small number of crewmen. The semi-submersible vessel, with its single operator concept, can be operated at-sea with a much smaller crew than its monohull counterpart. The major economies of the SSFV, however, are most likely to be found in its greater operational effectiveness. This vessel will put to sea under conditions where other vessels cannot - - - - - leave port. More important, when it reaches the operating site, the SSFV will be able to go into an operational state, despite most : ., : weather conditions that it might encounter. Most of the present Hawaiian fleet puts to sea about 80% of the time (during the summer, perhaps 99%), but often circles the operating area waiting for the environment to improve. It has been roughly estimated that the overall operating time at the site is in the order of 50%. FIGURE 7. A 300-Ton SSFV SPECIFICATION 300 TON CONCEPT 450 TON CONCEPT In contrast, the SSFV should beabletoputtosea 99%of the time, year-round, and begin operations immediately upon Displacement Tonnage 300 long tons 450 long tons arrival at the operating site. These factors are admittedly crude Payload 140 long tons 220 long tons but provide some insight into the anticipated improvements in Range 2000 nmi @ 14 kts. 4000 nmi @ 15 kts. operational economies. More exact relationships will result from Design Speed 16 knots 17 knots the economic studies soon to be carried out in conjunction with Length Overall 96 feet 112 feet the University of Hawaii Sea Grant Office. Beam 52 feet 62 feet Crew 8 12 SUMMARY Fuel Cost per Mile' $4.80/nmi $5.56/nmi Labor Cost per Mile 2 $4.20/nmi $5.60/nmi This paper has addressed briefly the dilemma faced by the Const. & Outftg. Cost $2.5 million $4.0 million modern fisherman in the United States. The specific situation in Amort. Cost per Mile 4 $5.56/nmi $13.88/nmi the Central Pacific has been expressed. SWATH technology Misc. Cost per Mile $4.00/nmi $5.00/nmi provides an exciting solution to most of the problems that exist. A Total Cost per Mile $18.56/nmi $30.04/nmi specific design for this region has been shown with detailed substantiations for its effectiveness. Plans are underway for construction of five SSFV's. Since this paper has been written as plans are in the formative stages, the descriptions are somewhat This estimate is based upon a shaft horsepower of 2500, a fuel con- sumption of 0.4 lbs/hp-hr, a speed of 15 knots and a fuel cost of general. However, studies are presently underway to examine $0.55/gallon. specific technical, economic and operational alternatives. The ie. 2Oh xI1 gal $0.5 x I hr - results of these studies will be presented in future reports to i.e. 2500hp x h x 6.6 lb x g-non 15mi nine preserve the cooperative and open nature of the SSFV program. hp-hr 6.6 lb gallon 15 ml mile 2 Labor estimates are based on: 12 crew members, an average 8-hr working day, an average rate of pay of $7.00/hr, and an average speed of 15 knots (this speed covers in-port time as well). 8 hrs x $7.00 1 day - $5.60 ths:12me xman-day hs 120 mi ml * ~ ~ ~ ~ ~ ~~ ~-thus:a 12 men x 1 x $70a 561. Blount, L. H. and E. A. Schaefers, "Recent U.S. Combination Fishing Vessels," in Fishing Boats of the World:3, Jan-Olof Total vessel cost estimate is based on a bare hull cost of $3.6 million Traung (editor), Fishing News (Books), Limited, London, ($4.00/lb of displacement) and an outfitting cost of $1.4 million. England, 1967. 4The amortization is based onan initialcostof $5,000,o000on a 9 percent loan over 20 years with a resulting finance charge of $4,725,000. The 2. Comitini, S., "An Economic Analysis of the State of the average life span speed of the vessel is assumed to be 4.0 knots. Hawaiian Skipjack Tuna Fishery, Univ. of Hawaii Sea * ~~~~~~~~~~~~~~~~~~~Hawaiian Skipjack Tuna Fishery," Univ. of Hawaii Sea Grant Program, UNIHI-SEAGRANT-TR-76-01, Nov. 1977. i.e. $9,725,000 ,x 1 hr 1 day x year - $13.88 20 years 4.0 kts 24 hrs 365 days mile 3. Hightower, J. D. and R. L. Seiple, "Operational Experiences 5 This estimate includes insurance and drydoeking costs as well as with the SWATH Ship SSP KAIMALINO," presented at mis cell aneous berthingfeets, etc. AIAA/SNAME, Advanced Marine Vehicle Conference, April 1978. FIGURE 6. Specifications for Two SSFV Concepts 4. Lamb, R. G., "The SWATH Concept: Designing Superior Operability into a Surface Displacement Ship," David Taylor Naval Research and Development Center, Report No. 4570, Dec. 1975. 213 DEVELOPMENTS IN THE LAW: THE FISHERY CONSERVATION AND MANAGEMENT ACT OF 1976 Eugene R. Fidell LeBoeuf, Lamb, Leiby & MacRae 1757 N Street, N.W. Washington, D.C. 20036 Legislative action was also required in Abstract connection with the Governing International Fishery Agreements (or GIFAs)(5) since it The Fishery Conservation and Management became clear early on that certain of the Act of 1976 has now been in effect for over procedural steps necessary for the issuance 1-1/2 years, and has been the subject of of foreign fishing permits could not be com- attention and action by the Congress, the pleted in time for the onset of the new Executive Branch and the Federal Courts. A regulatory regime under the FCMA.(6) It was summary of the legal concerns that have arisen unfortunate that all the pieces of the puzzle and that merit further attention is presented. could not be in place at the appointed hour, but Congress ultimately did, in approving the first round of GIFAs, make allowance for the Introduction unusual situation, and granted relief from these procedural niceties (which included the require- The Fishery Conservation and Management ments that license fees be paid in advance, that Act of 1976 was signed into law by President permits be on board the vessel, and that certain Ford on April 13, 1976, although its important waiting periods be observed by the Regional substantive provisions did not go into effect Fishery Management Councils before permit until March 1, 1977.(1) As a result, we issuance).(7) Congress also gave express currently have only about eighteen months of recognition to the special circumstances of the practical experience under this legislation. Reciprocal Fisheries Agreement with Canada,(8) a Nevertheless, there have already been many matter that has since been the source of some significant legal developments with respect international friction and even litigation, as I to the FCMA, and my purpose today will be to will shortly note. Here Congress was faced describe these. I shall organize these ob- with a departure from the framework of the servations in the following categories: statute, since the Canadian agreement was deemed (1) legislative developments; (2) litigation; not to be a GIFA.(9) A special act was felt to (3) administrative aspects. Because there be necessary because the agreement had pro- is more to tell than time permits, much of visions that were not consistent with the FCMA. this treatment will necessarily be rather (10) summary, but I shall be happy to entertain Congress has not, however, felt moved yet to questions afterwards. address some of the thorny technical issues that inhere in the statute. These issues have to do with such matters as the jurisdiction of the 1. Legislative Developments federal courts to provide judicial review of NOAA actions under the FCMA, elements of proof Changes in the FCMA have been proposed for in a forfeiture action, and the scope of reserved a variety of reasons. Perhaps the most state powers in light of the extremely confused important is the recognition that the statute provisions of � 306 of the law.(1ll) For its part, as originally enacted failed to make adequate NOAA has expressed interest in fine-tuning on provision for the possible loopholes in the some of these items,(12) but the idea has failed area of foreign investment and joint ventures. to catch fire on the Hill, even though other This subject received a considerable amount of measures have received detailed attention, attention both on the Hill(2) and within including hearings. Happily, the Congress has NOAA,(3) and I assume that most of this audience taken steps to alter the statutory requirement is aware of the present legislative situation on that all applications for foreign fishing permits this point. An effort has been made to tidy be reproduced in their entirety in the Federal things up with respect to the treatment of Register.(13) Perhaps Congress' reluctance is foreign fish processing vessels,(4) although due to the fact that there are few votes to be we should not underestimate the creativity of gained by such legislative action, but it is to businessmen in seeking new ways to do business be hoped that the 96th Congress will do better in light of changing legislative mandates. than its predecessor in addressing the technical legal issues, rather than leaving them to the courts for resolution. 214 I might perhaps close this legislative constitutional because there was no probable overview with the observation that the FCMA has cause, the defendant was convicted and fined had some unusual impacts on other legislation, $200. The other criminal prosecution involved both foreign and domestic. Domestically, you the masters of a Mexican fishing vessel seized may be aware that the 1977 amendments to the by the Coast Guard off Texas(19) and a Soviet Clean Water Act extended certain United vessel seized off Massachusetts.(20) The States marine pollution controls to the area Mexican master pleaded guilty and was given a covered by the FCMA.(14) Note, incidentally, suspended prison term of three months, while that those amendments do not simply push the the Soviet skipper was given a suspended nine- pollution controls out to a 200-mile limit; month sentence. they actually extend the Clean Water Act There have been a number of judicial further to the extent that the FCMA covers proceedings to forfeit foreign fishing vessels. marine resources beyond 200 miles. Thus, These have included the Soviet vessel Taras because the FCMA protects anadromous fishes of Shevchenko,(21) the Mexican vessel Adriana,(22) United States origin beyond the Fishery Con- the Taiwanese vessel Highly No. 301,(23) the servation Zone, or living resources of the Japanese vessel Sachi Maru No. 22,(24) and the continental shelf where the shelf extends Spanish vessel Costa de Noruega.(25) In addition, beyond the FCZ,(15) the Clean Water Act, too, part of the catch of another Soviet fishing has a broader sweep. vessel was seized shortly after the FCMA took On the foreign legislative front, Canada effect.(26) has amended its Fisheries Act to provide for More interesting and more important than issuance of summonses in lieu of arrest for these isolated fishing violation cases are the minor offenses.(16) Quite plainly this was cases in which litigants have sought to obtain inspired by the streamlined "parking ticket" direct judicial review of administrative actions concept found in � 311(c) of the FCMA, and under the FCMA. shows quite nicely how one nation may learn The first such action was brought by Maine from another in the development of laws. to obtain review of the 1977 herring quotas. Basically, the state alleged that these quotas were a rerun of the earlier ICNAF quotas, and 2. Litigation that the Commerce Department had taken improper factors into consideration when developing the Fishermen have been suing one another for quotas under the FCMA. The case was at first centuries over the damnedest things. They are remanded to the trial court by the First Circuit probably as litigious as any group in American Court of Appeals, with instructions to require society, but nonetheless, there has been sur- Secretary Kreps to state the reasons for her prisingly little litigation to date under the decision with respect to the quota.(27). There- FCMA. Only a few cases have reached the after, Maine appealed again, and on the second Courts of Appeals, and no case directly in- appeal the Court of Appeals in Boston sustained volving the FCMA has yet come to the Supreme the district court's approval of the Secretary's Court. Domestically, the paucity of cases action.(28) The Court's opinions are important may be because the industry does not have because they show that the Commerce Department a deep-seated tradition of doing things together, may take into account relations with foreign whether nationwide, regional, or even between countries even though the "cardinal aim" of the neighboring ports. On the foreign front, statute is to help "the development of a United fishing industries in other countries have States' controlled fishing conservation and failed to assert legal rights in our courts, management program designed to prevent over- presumably because they continue to rely upon fishing and to rebuild depleted stocks...."(29) the Department of State and their own govern- The Court went out of its way to underscore the ments as their spokesmen in dealings with the breadth of the Secretary's discretion in setting regulators. Finally, because the FCMA provides quotas. for administratively-assessed civil penalties Other litigation has been directed at a as an alternative to immediate recourse to the variety of NOAA and NMFS actions under the new federal district courts(17)--and because NOAA law. For example, two civil actions were filed has been slow in developing the civil penalty to challange the policy announced by NOAA with cases to the point where litigation might ensue respect to foreign fish processing ventures with- --the structure of the Act and the nature of its in the FCZ. One such case was filed in the implementation have both served to reduce the district court in Washington, D.C., by a variety flow of litigation--at least for the moment. of West Coast fishing companies, (30) while the As of June 1978 there had been only a hand- other case was commenced in the district court ful of court cases under the Act. Only three in Seattle by a different group of firms.(31) of these were criminal prosecutions, and in fact Both actions seek to invalidate the NOAA policy one of the three cases had little to do with under which fish caught by U.S. vessels, but fishing. Rather, it was a prosecution in the processed by foreign vessels, are treated as United States District Court for the District part of the U.S. share, rather than deducted from of Rhode Island on a charge of assaulting a the total allowable level of foreign fishing.(32) boarding officer.(18) A special agent of NMFS Both cases also raise questions as to whether was assaulted while trying to perform a routine NOAA properly complied with the Administrative dockside boarding of a United States vessel. Procedure Act in adopting the policy statements Despite a contention that the boarding was un- on this issue. 215 Beyond this recent activity, cases have documented steamboats in the landmark case of been filed in the district court in New Jersey Gibbons v. Ogden,(41) was also applicable to regarding the validity of the Fishery Management vessels licensed for the fisheries. More Plan for surf clams and ocean quahogs;(33) in recently, however, the court took a step backward the district court for Massachusetts regarding in holding that Montana could have big-game the validity of the Fishery Management Plan for license fees that discriminated against nonresi- Atlantic groundfish;(34) in the Western dents,(42) despite a rather long line of cases District of Washington regarding the continua- that had seemed to cast doubt on such measures. tion of Canadian fishing for salmon prior to Whether the doctrine will have an impact on the Congressional approval of the appropriate administration of the FCMA remains to be seen. international agreement;(35) and in the same district court with respect to the validity of the Fishery Management Plan for salmon.(36) 3. Administrative Aspects Maine has also sued to challenge the structure and operation of the Regional Councils.(37) Turning to the Executive Branch role in Little purpose would be served by going administration of the FCMA, it must be said that through the details of each of these cases, things have gotten off to rather a slow start. It is perhaps more important to identify the Hundreds of notices of violation have been given main themes that are being struck in the various out by NOAA and the Coast Guard, but the process lawsuits. These are: has there been com- of bringing these cases to fruition has been pliance with the Administrative Procedure Act exceedingly slow. This is perhaps the most severe with regard to rulemaking, and to what extent shortcoming in the administration of the Act, for does that law apply to particular agency the lack of enforcement teeth can bring the entire actions under the FCMA? Has a proper Environ- regulatory effort into disrepute. mental Impact Statement been issued? Have the More broadly, question may be raised as to proper factors established in the FCMA for whether administration of the Act is being con- management decisions been taken into account ducted evenhandedly as between foreign and and the analysis spread upon the record? To domestic vessels. As Senator Magnuson has ob- the extent that statutory obligations exist with served, "[c]onservation is not a one-sided affair, respect to enforcement, have these been ob- and foreign and domestic fishermen must be subject served? If emergency action is taken, was it to the same enforcement rules. To create a double necessary? standard for speculative 'foreign policy' reasons Obviously, this list is not exhaustive, and will undermine our entire conservation effort." as further litigation arises, entirely new (43) I confess that I have in the past more or theories will be advanced. One issue which I less assumed that there might be some discrimina- expect will be the basis for much litigation tion against foreign fishing vessels given the is the basic one of whether the biological/ fact that the FCMA built upon a statutory scheme scientific data base actually supports the that essentially barred the door to foreign management decisions that have been made. fishing. I am happy to report that the statistics Congressman Ruppe, for example, has stated that to date do not appear to bear out such a supposi- "[t]he available biological data on which to tion. The figures for the period ending base fishery management plans is inadequate and December 31, 1977 are inconclusive, although they the development of a sufficient data base will can be read as suggesting that foreign fishing require an extensive research program."(38) vessels tend to get off more easily in the en- Unless there is a sound, articulable basis for forcement process. However, this may also be the management plans, they will be subject to because there are in fact fewer foreign violations. question and judicial invalidation. For this In any case, during that period there were 225 reason it is critical that NOAA's research reports of serious domestic violations, and only efforts be directly linked to the FMP- 133 reports of serious foreign violations. Domes- development process, and that plans disclose, tic vessels were given citations for minor offen- in intelligible terms, the scientific basis ses 129 times, and foreign vessels received 329 upon which they are founded. such citations.(44) These data would have to be Other judicial actions should perhaps also analyzed to determine whether foreign or domestic be mentioned. The Supreme Court has been quite vessels were boarded disproportionately to their active in the natural resources field of late, numbers. including the fisheries law area in particular. Administrative action of an important kind In Massachusetts v. Westcott,(39) the Court occurred last Spring in the context of the ongoing held that Massachusetts could not bar a U.S.- negotiations with Canada. After the negotiations documented fishing vessel from fishing in seemed to be falling apart, a decision was made to waters of Massachusetts even though the vessel bar Canadian fishing from United States waters, was homeported in Rhode Island. Similarly, including Canadian recreational fishing on the in Douglas v. Seacoast Products, Inc.,(40) the seacoasts and recreational and commercial fishing Court struck down Virginia's efforts to on the Great Lakes.(45) Despite the fact that establish its own citizenship requirement for the FCMA's predecessor, the Bartlett Act, had been vessels seeking to fish in its waters; the construed to apply to the Great Lakes,(46) the Court held that the Enrollment and Licensing Administration concluded that the Great Lakes Act of 1793, which had previously been applied were not covered by the FCMA's prohibition on to defeat New York's efforts to control U.S.- foreign fishing on the ground that the term 216 "fishing" in the statute must be read in light In any event, the opportunities for change, of the definition of "fish", which in turn may whether on the administrative, judicial or be read as limited to marine fishes.(47) The legislative side,are manifold. It will be a result was that Canadian recreational fishing test of the wisdom and skill of the drafters was permitted to go forward on the Lakes.(48) of the FCMA to see how well their brainchild Whether the decision is right or wrong, it withstands these changes. shows quite eloquently how much discretion is vested in the administrative agencies con- References cerned with implementation of the new law. Another administrative aspect has to do 1. 16 U.S.C. �� 1801 et seq. (1976) ("FCMA"); with the flow of information, such as the see FCMA �� 201(a), 204(a). rendering of interpretations like the one 2. See, e.g., 200-Mile Fishery Oversight-- just mentioned. I am very pleased to state Joint Ventures, Hearings on Joint Ventures: that NOAA has, at my urging, adopted a H.R. 2564, H.R. 4165 and H.R. 4166 Before liberal policy with respect to the public the Subcomm. on Fisheries and Wildlife availability of its General Counsel's Conservation and the Environment of the opinions. Without these opinions, many House Comm. on Merchant Marine and Fish- observers of the fishing scene have had to rely eries, 95th Cong., 1st Sess., ser. no. on rumor and guesswork to learn the details of 95-19 (1977) ("Joint Ventures Hearings"). the Government's legal position on important 3. See NOAA Proposed Interim Policy, 43 Fed. questions. With the opinions, all affected Reg. 5398 (1978); NOAA Policy Statement, parties will be able to guide their actions U.S./Foreign Fishing Ventures Within Fish- with a far better understanding of Government ery Conservation Zone, 43 Fed. Reg. 20532 policy than was available in the past. NOAA (1978). is to be applauded for this welcome change. 4. See Fishery Conservation and Management Act Which leads me to my concluding observation. Amendments, S. Rep. No. 95-935, 95th Cong., The lack of regular access to agency General 2d Sess. (1978). Counsel opinions is only a small part of a 5. FCMA � 201(c). larger problem. Information flow as a whole 6. See generally International Fishery Agree- under the FCMA has, in my view, been haphazard ments: Hearings on H.J.Res. 240, H.R. 2081, at best, and this void has not been properly H.R. 3753 and H.R. 4954 Before the Subcomm. filled by the trade press. What is badly needed on Fisheries and Wildlife Conservation and is a regular information and document service the Environment of the House Comm. on Mer- that will present important textual materials chant Marine and Fisheries, 95th Cong., of general interest to the fisheries industry, 1st Sess., ser. no. 95-1 (1977). recreationists, environmental groups and 7. Fishery Conservation Zone Transition Act government officials, in a convenient form � 3(1), 16 U.S.C. � 1826 (1976). with a minimum of delay and expense. It is my 8. H. Doc. No. 95-90, 95th Cong., 1st Sess. hope that industry will support development (1977); Pub. L. No. 95-73, 91 Stat. 283. of such a service at an early time, both in 9. See, e.g., H.R. Rep. No. 95-783, 95th Cong., its own interest and as a public service. 1st Sess. 4 (1977). 10. See Letter of Transmittal from the President to the Congress, Feb. 28, 1977, in H. Doc. Conclusion No. 95-90, supra note 8. 11. See Memorandum of the author on Technical There have been major legal developments Amendments to Fishery Conservation and on every front since the Fishery Conservation Management Act of 1976, in Joint Ventures and Management Act took effect last year. The Hearings, supra note 3, at 281-82. pace of these developments will accelerate in 12. Letter from Dr. Robert M. White, Adm'r, NOAA, the next year or two, as the shakedown period to Rep. Robert L. Leggett, Mar. 21, 1977. under this important legislation continues. 13. See FCMA � 204(b)(4). While specific predictions can be hazardous, 14. See Clean Water Act � 311(b), 33 U.S.C.A. one can at least suggest that important � 1321(b) (West Supp. 1978). changes could come about as a result of 15. FCMA � 102. President Carter's ongoing interest in govern- 16. See 25-26 Eliz. 2 ch. 35, � 18 (Can. July 14, ment reorganization, whether that interest 1977), amending Fisheries Act � 61 (Can.). manifests itself as a Department of Oceans, an 17. FCMA � 308; see generally Fidell, Enforcement overall Department of Environment, or as part of the Fishery Conservation and Management of the new approaches being taken with respect Act of 1976: The Policeman's Lot, 52 Wash. to protection of our borders. Other changes L. Rev. 513, 546-50 (1977). can be anticipated as the law in general 18. United States v. Champlin, Crim. No. 77-67 continues to evolve. For example, the Supreme (D.R.I. Nov. 21, 1977). Court's recent decision with respect to 19. United States v. Hernandez-Lopez, Crim. No. administrative inspections(49) could have an B-77-302 (S.D.Tex. Aug. 8, 1977). The indi- impact on enforcement of the Act, as could a cation in [1977] NOAA Annual Report, Fishery possible future decision by the Court on the Conservation and Management Act of 1976, at question whether lethal force may be used in 22 (1978), that a six-month sentence was im- pursuing a nonviolent fleeing criminal.(50) posed appears to be erroneous. 217 20. [1977] NOAA Annual Report, supra note 19, at 22; N.Y. Times, May 3, 1977, at 20, col. 4. 21. United States v. Russian F/V Taras Shevchenko, Civil No. 77-1086-M (D.Mass., filed Apr. 21, 1977). 22. United States v. F/V Adriana, Civil No. B- 77-179 (S.D.Tex., filed July 29, 1977). 23. United States v. F/V Highly No. 301, Civil No. A77-181 (D. Alas., filed Sept. 2, 1977). 24. United States v. F/V Sachi Maru No. 22, Civil No. A78-34 (D. Alas., filed Feb. 24, 1978). 25. United States v. Spanish F/V Costa de Noruega, Civil No. 78-762 (D.N.J., filed Apr. 12, 1978). 26. United States v. 17.56 Metric Tons, More or Less, of Fish, Civil No. 77-1086-M (D.Mass., filed Apr. 21, 1977). 27. Maine v. Kreps, 563 F.2d 1043 (lst Cir. 1977). 28. Maine v. Kreps, 563 F.2d 1052 (lst Cir. 1977). 29. 563 F.2d at 1049. 30. Tom Lazzio Fish Co. v. Kreps, Civil No. 78- (D.D.C., filed May 1978). 31. Pacific Seafood Processors Ass'n v. Kreps, Civil No. C78-313S (W.D. Wash., filed May 23, 1978). 32. Cited supra note 3. 33. Sea Harvest, Inc. v. Kreps, Civil No. 78- 514 (D.N.J., filed Mar. 15, 1978). 34. Ferrara v. United States, Civil No. 78-420-G (D.Mass., filed Feb. 21, 1978). 35. Washington Trollers Ass'n v. Vance, Civil No. C78-248V (W.D. Wash., filed Apr. 24, 1978). 36. Nevaril v. Kreps, Civil No. C77-358S (W.D. Wash., filed May 13, 1977). 37. Maine v. Kreps, Civil No. 77-176-SD (D.Me., filed Aug. 3, 1977). 38. 124 Cong. Rec. H2629 (daily ed., Apr. 10, 1978) (remarks of Mr. Ruppe). 39. 431 U.S. 322 (1977). 40. 431 U.S. 265 (1977). 41. 9 Wheat. 1 (1824). 42. Baldwin v. Fish and Game Comm'n of Montana, 46 U.S.L.W. 4501 (U.S. May 23, 1978). 43. Sen. Comm. on Commerce, News Release No. 95- 52 (Apr. 13, 1977). 44. [1977] NOAA Annual Report, supra note 19, at 18-19. 45. See N.Y. Times, June 6, 1978, at 1, col. 1; 124 Cong. Rec. H4928 (daily ed. June 5, 1978) (remarks of Mr. Latta). 46. See Fidell, Ten Years Under the Bartlett Act: A Status Report on the Prohibition on Foreign Fishing, 54 B.U.L. Rev. 703, 716-17 (1974). 47. See FCMA �� 3(8), 3(10), 307(2). 48. Washington Post, June 11, 1978, at A21, col. 1. 49. See Marshall v. Barlow's, Inc., 46 U.S.L.W. 4483 (U.S. May 23, 1978). 50. See N.Y. Times, June 1, 1978, at A19, col. 1 (reporting 8th Cir. decision sustaining dam- age action by parent of fleeing nonviolent felony suspect killed by police gunfire). 218 INPUT-OUTPUT ANALYSIS FOR SALT MARSH BIOPRODUCTIVITY * Thomas M.. Lesohine Leah J. Smith Woods Hole Oceanographic Institution Woods Hole Oceanographic Institution Woods Hole, Massachusetts 02543 Woods Hole, Massachusetts 02543 some of the important biological interactions with Abstract the total input requirements and potential eco- nomic outputs of the system. Input-output analysis is applied to a salt marsh system to describe possible bio-economic In a sense, nitrogen is the "currency" being effects of adding nitrogen fertilizer. A passed back and forth among tidal water, grasses, simple model is developed to trace exchanges of herbivores and other compartments in the marsh as nitrogen among five compartments of the marsh biological activity takes place. Much ecosystem model: sediments, Spartina root/rhizome bio- analysis over the past 20 years has focused on the mass, Spartina aboveground biomass, tidal flow of energy and nutrients through similar bio- water and shellfish. The resulting matrix can logical systems (among frequently cited energy be used to show relationships in the system among flow models are H.T. Odum (6) and J.M. Teal (9)). controllable inputs and outputs with potential Several recent authors have noted the similarity economic value. Some costs of fertilizing an of these flows to the flow of money through a experimental plot are given, but no direct national economy, and a mathematical theory of comparison can yet be made between costs and flows in ecosystems equivalent to the economic likely economic benefits. input-output analysis of W. Leontief (5) has been developed (Hannon (3) and Finn(l)). Gosselink et al (2) have used this analogy directly in 1. Introduction placing a dollar value on an acre of salt marsh, determined by measuring annual net primary pro- ductivity, a proxy for annual energy flow, and Recent research an salt Marshes; at Woods Hole has emphasized both qualitative anc quan- converting energy units to dollars. Our goal is titative aspects of the flow and intercon- to link the costs of marsh fertilization to the version of nitrogen as it passes among the potential harvest of fish and shellfish relatively various biological compartments of salt marsh far removed in position in the food web from these systems. It has been established that the initial nutrient inputs, and it is natural for us artificial addition of nitrogen-based fertilizer to use an input-output format. can enhance the productivity of biological systems within such salt marshes (10), although Productivity of the grasses which comprise the possible bio-economic benefits of salt marsh the major component of the tidally borne detritus fertilization - enhanced fish and shellfish in these marshes, mainly two species of Spartina,is production - are relatively far removed from enhanced by fertilization. This detritus in turn initial perturbations on the system such as is the basic food source for the marsh's food webs. Primary production both in salt marshes (10) fertlizaion.The trucure f ths smll-and in coastal waters (7) is nitrogen limited. A scale bio-economic system is in many ways similar and in oaseal te (7) of nitrogen to that of economic systems which have been des- through marsh food webs has been done in Great cribed with input-output analysis. Other eco- truhmrhfo eshsbe oei ra noicwith input-oaches havbnalysi. Othr descre Sippewissett Marsh of Falmouth, Massachusetts, and nomic approaches have been used to describe production in manufacturing or agricultural in Buzzard's Bay, with which the marsh has tidal systems. Production functions of various kinds exchange. A complete annual budget of nitrogen either describe such systems in engineering inputs and outputs to and from Great Sippewissett terms or make assumptions about the contribution has been prepared (10). and substitutability of various factors of pro- ductin (pinciallylabo andcapial).Most*Prepared with funds from the Pew Memorial Trust duction (principally labor and capital). Most of these systems, however, assume a high degree and by the Department of Commerce, NOAA Office of of control over all inputs and outputs. In a Sea Grant under Grant #04-7-158-44104, and the salt marsh used as a site for enhanced pro- Marine Policy and Ocean Management Program of the ductivity, many of the assumptions underlying Woods Hole Oceanographic institution. standard economic production functions do not hold. In addition, the complex and vital roles of biological interactions require different treatment than a standard manufacturing system. The input-output approach enables us to combine 219 The salt marsh performs an important '�onver- "back calculation" through a flow model to estimate sion" activity on the nitrogen which enters it. the nitrogen requirements of meeting a desired lev- The marsh intercepts a flow of ground water rich el of demand for marsh outputs expressed in terms in dissolved nitrates, its principal natural nu- of biomass output of fish or shellfish. This gives trient source, and converts it to a mixture of a way of estimating the feasibility of marsh ferti- particulate nitrogen and disssolved organic lization for the purpose of enhancing bioproduction and inorganic nitrogen compounds, comprising a as well as the probable costs of fertilization. net annual export to the Bay. The particulate The method is general and can be applied in any nitrogen is mostly grass detritus while a large situation in which an appropriate flow model is portion of the dissolved nitrogen export is available. ammonium. The effect of the marsh is thus to manufacture particulate nitrogen food for har- 2. A Linear Model for Bioproduction vestable shellfish which is then available to man, and for small mollusks and other benthic If a national or regional economy is composed creatures available in turn to commercially of n sectors i = 1,...,n, then the total cash flow valuable fish. All are dependent on the through sector j (i.e. the value of j's total pro- tidally innundated habitat the marsh provides. duction) can be expressed as the sum of the cash Were the marsh not there, the impact of the ni- flows to all other sectors from j (the value of trogen from the groundwater would be primarily sales made by j to other sectors for the purposes on the phytoplankton in the Bay, and through of their production) and the value of sales to them on the grazing food webs less directly tied consumers by j (final demand). Letting x. be the to resources harvestable by man. value of j's total production, xij be the 3value of sales to i by j for production by i, and yj be the The effects of nitrogen-based fertilizer final demand for j's production, we can express applications on the biomass yield of Spartina this formally as alterniflora, the dominant grass in the marsh, can be dramatic: an application of 8.4 g/m2/ (1 j = ~x.j + yj. week, done biweekly over a seven month period, i 13 was shown to double, net annual aboveground pro- duction of S. alterniflora in experimental plots, If we assume further that xi. is proportional to while an application of 25.2 g/m /week resulted xi, then Equation (1) can be written in a near tripling of biomass yield (11). Thus, the simple application of fertilizer can increase X 2. +j y or the food supply for marine organisms, making it tempting to regard fertilized salt marshes as open aquaculture systems and to view salt marsh (3) X XA + Y fertilization in economic terms. in vector notation, where X and V are n dimen- sional row vectors and A is an nxn matrix of co- It is unlikely that commercial aquaculture efficients. based on salt marsh fertilization would be prac- tical, since little control can be exerted over the natural biological processes in a marsh. on usual assumptions of economic input-output anal- the other hand, lack of controllability is a ysis, Replacement of the economic sectors with virtue in that the costs incurred in maintaining compartments in an ecosystem model and the cash a salt marsh as an open aquaculture system are flows with flows of energy or nutrients generates low. This is especially true if we adopt the the equivalent interpretations of Hannon (3) and viewpoint that fertilization, if it proves prac- Finn (1). In the specific example we will pre- tical, should be conducted locally by towns for sent below, Equation (1) will be interpretable the benefit of local residents and in-shore fisher- as a statement that the total nitrogen flow men who harvest marsh-dependent fish and shellfish. through compartment j is equal to the sum of the The explicit costs are thus limited to maintenance direct flows from j to all other compartments and of the fertilization program if harvesting is the flow from j out of the system, after Finn. assumed free. It is possible that secondarily It is not necessary to assume that the system treated sewage effluent could be used as ferti- modeled is conservative (i.e. that the inputs lizer in some areas, reducing costs even further balance the outputs for each compartment), though and providing an additional tertiary sewage treat- in a technical sense the computations require ment benefit (8). Though the market value of that inputs and outputs do balance. It is easy the potential increase in harvest may be small, enough, however, to balance the flows artificially an important component of its total value is re- by the addition of sources and sinks without creational and difficult to quantify. changing the nature of the basic result. It should also be noted that the substitution used We will show how "biomass flow" through the to derive Equation (2) assumes that a change in marsh can be indirectly charted by an analysis of total compartment flow leads to proportional the flows of nitrogen which result as initial changes in the flows which enter that compart- nitrogen inputs are converted biologically into ment from other compartments. Were the methods higher organisms. In the sections that follow of Isard (4) to be cast in this framework, then we describe our input-output methodology in more the flows could be thought of as flows of gootor detail and develop some cost estimates for marsh fertilization. We show how it is possible by 220 materials directly, rather than of the nutrients or energy stored in such materials. The model described by Equation (3) has two simple equivalent forms of interest: (4) X(I-A) = Y, and (5) X = 2(I-A)-, where Equation (5) is understood to mean the equality holds given that (I-A)-1 exists. The first form expresses the outflow from the various compartments in terms of the total flow through them, while the second makes it possible to solve for the total flow necessary for the system to produce an outflow of a given intensity. This latter interpretation is the one of greatest interest to us and also the formulation of Figure 1 greatest use in input-output economics. In our system we will consider marsh shellfish pro- A three compartment ecosystem with flows. Plus duction (measured as nitrogen content in average signs indicate compartments which are accumulating annual mussel growth) as an outflow from the excess flow. model's shellfish compartment. By expressing the total nitrogen flow through the compartment to unbalanced and that compartments two and three are which fertilizer is an input (marsh sediments) as each accumulating an excess of five units of flow, a function of its nitrogen input, we can then The system's flows aresummarized ininput-outputform estimate the increase required in nitrogen in- in Table 1. Here each column represents the out- flow (and hence the required fertilization rate) flows from the compartment at its heading, both to to support an increase of a given size in shell- the compartments at the row headings and out of fish production. the system completely. Sinks are added to the appropriate columns so that total inflows and out- in the economic literature the quantity flows balance. Rows provide similar information (I-A)-1 is called the Leontief inverse of the with respect to inflows for compartments at their matrix A. In a sense, it expresses the multi- plier effect that a change in final demand for one sector's goods will have on the other sectors, Sour- In- as a result of the direct and indirect cash flows which that sector's changed production will in- duce throughout the economy. This is easy to (1) 0 1 0 0 14 15 see in the simple one-dimensional case: If we replace Y by Y + Ay in Equation (5) and solve for (2) 5 0 2 0 3 10 the new level of total cash flow V which will result, we find (3) 10 0 0 0 0 10 XI = (Y+AY) (1-A)-i=Y(1-A)- + AY(1-A)- = Sinks 0 5 5 0 X + AYO1-A)-I Out- flows 0 4 3 10 In the special case in which AY = $1! XI =X+fs (I-A)-1, so that (1-A)1 represents exactly the Total 15 10 10 10 total new cash flows required to sustain the new level of production. In the nitrogen flow model elements of the nxn matrix (I-A) essentially total the direct and indirect outflows coupling Table I the compartments at the row and column headings which result when a unit change is made in the outflow from the compartment at the row heading. Specifically, the element in the (i,j) position Figue 4x4 malrxAonitng wit the crepningter-A) represents the portion of outflow from i which compartmental and balancing flows of Table has been through j. after division by the row totals, is shown inl Figure 2, along with the corresponding (I-A) 3. A, Flow Model for the Great Sippewissett Marsh matrix. Direct calculation via Equation (5) shows that if the outflows vector 7 is replaced by 7' Figure I below illustrates a hypothetical = (0,4,6,10), in which the outflow from compart- th~ree-compartment ecosystem with intercompart- ment 3 is doubled, a total flow requirement from mental flows, inflows and outflows of specified compartment I of about 18 units results, a 20% magnitude. Note that the flow in the figure is increase in the 15 units total flow it supplies to the rest of the system in Figure 1. If 221 intercompartmental flows in the system are propor- wet weight. Flows in the marsh are summarized in tional to flows into the donor compartments, then Table 2 below, in the format used for Table 1. an appropriate manipulation of the magnitude 14 The flow diagram is omitted, but it can be easily inflow to compartment 1 could produce the desired drawn from the information in the table. total requirement from that compartment. In- ~~~~~~~~~A: 0 .07 0 0 (1) (2) (3) (4) (5) flows .5 0 .07 02 ~~~~~.5 0 .2 0 ~(1) 0 4200 0 490 470 5970 1 0 0 0 (2) 6990 0 0 0 0 1270 (3) 0 1210 0 0 0 0 0 .5 .5 0 (I-A):1.05 .0736 _.014 (4) 1270 0 1160 0 70 26250 (I-A) :1.05 .0736 ,0147 0 .735 1.06 .210 0 (5) 0 0 0 730 0 0 oj ~~Out- 1.05 .0736 1,0 0 15 1.05 .0736 1,015 0flows 2870 0 0 31600 70 .893 .563 .612 I Table 2 Figure 2. Normalized matrix of flows A and corresponding Annual nitrogen flows for Great Sippewissett matrix (I-A)-1 for the system of Figure 1, I Marsh in kg N yr , summed over the nitrogen in all major forms found in the marsh. Balancing The detailed marsh model, together with its flows are omitted. Compartments are: (1) application to the problem of linking shellfish Sediments, (2) Belowground biomass (3) Above- production to marsh grass fertilization, is quite ground biomass (4) Tidal water (5) Shellfish. similar to the example above. The model traces the annual exchanges of nitrogen (in units of kg The central portion of the matrix I-A)-l is N yr 1) among five compartments in a schematic presented in Figure 3 below. As explained above, flow model of the marsh. The compartments are [l) the elements of this matrix summarize the inter- Sediments, (2) Spartina root/rhizome biomass, (3) compartmental transfers which lead to the various Spartina aboveground biomass, (4) Tidal water, outflows from the system. This kind of informa- and (5) Shellfish. The path of a nitrogen input tion can provide important clues about the specific via fertilization would be roughly through the relationships between flows in the system. With five compartments in the order listed, though there respect to the problem we have posed, the element is a fair amount of internal cycling and flow to in the (5,1) position is especially significant. and from the outside environment in the model. It tells us that, given an outflow of a specified The nitrogen in all major forms of nitrogen magnitude from compartment 5, only about 10% of found in the marsh (particulate organic, dis- that outflow has been through compartment 1. solved organic, dissolved inorganic) is included. This stands in marked contrast to the situation Inputs to the system from outside include the ni- in our hypothetical example, in which no flow trogen in groundwater, precipitation and tidal could enter compartment 3 without first passing inflow, and bacterial, algal and rhizomal fixation. through compartment 1 (See Figure 1). Thus, Outputs include the nitrogen in tidal outflow, based on the model, it would be difficult, for denitrification, volatilization, and losses to example, to double shellfish production in the deep sediments. All are summed over a year, In marsh (outflow from compartment 5) by fertili- most cases data used are averages of several years. zation alone (i.e. by increasing the magnitude Details of how various values were calculated are of the inflow to compartment 1). omitted here, but are available upon request. Fertilization, the nitrogen in groundwater and (1) (2) (3) (4) (5) precipitation and that fixed by surface bacteria and algae are lumped into a single exogenous in- (1) 1.48 .564 .005 .128 .063 flow to compartment 1. The inflow value used in the model excludes fertilization values, however, (2) 1.25 1.48 .004 .108 .053 so that the model actually describes the marsh in its natural state. Fertilization is thus viewed as (3) 1.25 1.48 1.00 .108 .053 a manipulation of this inflow similar to that des- cribed for the input to compartment 1 in the hypo- (4) .102 .074 .036 1.01 .006 thetical system. Likewise, the nitrogen in net annual average growth of shellfish, as estimated (5) .102 .074 .036 1.01 1.01 for the marsh's natural population of ribbed mussels (Modiolus demissus), is listed as the sole exogenous output from compartment 5. This is done Figure 3 in an attempt to approximate a potential sustained -1 yield harvest. This figure, 70 kg N yr'1, cor- The central portion of the matrix (I-A) , as com- responds to roughly 8500 kg mussel meat production, puted for Great Sippewissett Marsh nitrogen flows. 222 This result is not surprising in terms of The alternative of obtaining nitrogen from the nitrogen flow model we have presented. In- secondarily treated sewage effluent would depend spection of Table 2 shows, for example, that no on acceptance by the local community as well as flow can enter the shellfish compartment without on the presence of a nearby sewage treatment first passing through compartment 4, tidal water. plant. Under these circumstanices, the nitrogen This compartment in turn is strongly driven by the itself could be provided free of charge to the large annual tidal exchanges it has with nearby marsh, but the labor and capital costs would be coastal waters, signified by the large inflows higher to handle the greater volume required. In and outflows shown for this compartment in the contrast to the 45% nitrogen content of some com- table. A related factor which must be considered mercial fertilizers, secondarily sewage effluent is the fact that dissolved organic nitrogen (DON), has only about 15 parts per million of nitrogen, whose function is not well understood in the and an application rate comparable in nitrogen marsh, makes up a significant portion of the in- content would be 1.25 million liters per week per flows and outflows in our model. In fact DON hectare (instead of 25 kg N/ha/week of commercial makes up 62% of tidal inflows, 58% of tidal out- fertilizer). The feasibility of using this source flows, 45% of groundwater flow, and more than of nitrogen is clearly very dependent on parti- half of the total inputs and outputs to and from cular characteristics of the site. the marsh (10). DON apparently plays no role in shellfish feeding and in fact may pass through 5. Summary and Conclusions the marsh with very little interconversion, since total inputs and outputs of this species Input-output analysis can be a useful tool very nearly balance (10). The model presented for assessing the economic potential of biological here must be regarded as a first pass attempt systems over which man can exert little control. at analytical assessment of the potential for It makes possible the linkage of controllable enhanced shellfish production via marsh fertili- inputs to outputs of potential economic value, zation. A fairer test would probably involve a while allowing for explicit consideration of the model in which all DON is subtracted from the complex biological processes which ultimately flows. Final judgement on the ultimate practi- govern the system's output of such commodities. cability of marsh fertilization must thus be This seems particularly true of salt marshes, withheld until further analysis is made. where research has shown that fertilization can lead to dramatically increased yields of marsh 4. Costs of Fertilization grass biomass, and hence grass detritus which is potential food for shellfish. Estimating the cost of nitrogen fertiliza- tion of a salt marsh is difficult because the We have used an input-output format to des- equipment for applying the fertilizer is not cribe nitrogen flow through a New England salt standard. Rather than provide engineering cost marsh which has been intensively studied by estimates of large - scale systems, we will in- biologists over the past few years. While much dicate the approximate cost of the system now has been written about the use of input-output used to fertilize a 2 to 3 ha section of marsh, analysis to integrate environmental and economic This system pumps the fertilizer through spray concerns, attempts to do so are often limited by irrigation pipes to all sections of the plot. the lack of a detailed data base for the analysis. K� ~ The capital cost of the system is approximately We believe the problem of linking a marsh's bio- $29,000, excluding design costs. Depending on mass yield of fish and shellfish to fertilization remoteness, type of foundation required and other presents a valuable opportunity to test in detail site - specific factors, this cost could be redu- this extended input-output methodology in a ced by $5,000 or more. Vandalism has been a prob- setting in which all links in the system being lem in maintaining the equipment, so a ten-year considered are under active study. life is estimated. Assuming that the operator of the system can check it during periodic visits The input-output model presented here re- to the marsh for other purposes, and that approxi- presents a distillation of data collected in field mately the present level of vandalism continues, and laboratory studies. While the present model annual maintenance cost is estimated to be $3,000 is far from definitive in answering the questions per year or less.1 The operating costs are then we have presented, it illustrates a method which labor and fertilizer. If fertilizer is applied can be used to examine other similar ecological- for six months of the year, labor requirement economic interactions. Future work should lead to will be about one-half time during that sit months detailed analysis of the fate of the various flows ($3,000 if annual full-time salary rate is in the marsh system and provide some strong indi- $12,000) to apply fertilizer daily and check the cations of the extent to which similar systems equipment. Commercial fertilizer containing 45% lend themselves to this type of analysis. nitrogen is estimated to cost $1,000 per year(25kg N/ha/week for 25 weeks = 650 kg of nitrogen in fertilizer). This is the high-level applica- tion rate; researchers feel most of the increase in growth of Spartina could be obtained with only one-third as much fertilizer2. This would de- crease fertilizer cost but not labor and capital upkeep costs. 223 Acknowledgements 11. Valiela, I., J.M. Teal and W. Sass, "Production and dynamics of salt marsh Several people must be acknowledged for help- vegetation and the effects of experimental ful criticisms and access to unpublished data used treatment with sewage sludge," The Journal in the preparation of this paper. These include of Applied Ecology, Vol. 12, 1975, pp. 973- Ivan Valiela and Thomas Jordan of the Boston 982. University Marine Program, Jim Morris of the Marine Biological Laboratory Ecosystems Center, John Teal and Peter Kallio of Woods Hole Oceanographic Institution. Special thanks go to Jack Finn of the MBL Ecosystems Center for his assistance in the computer analysis of our results. 1Peter Kallio, personal communication, June 1978. John Teal, personal communication, June 1978. References 1. Finn, J.T., "Measures of ecosystem structure and function derived from analysis of flows," The Journal of Theoretical Biology, Vol. 56, 1976, pp. 363-380. 2. Gosselink, J.G., E.P. Odum and R.M. Pope, "The value of the tidal marsh," The Center for Wetland Resources, La. State University, Baton Rouge, 1973, SG-74-03. 3. Hannon, B., "The structure of ecosystems," The Journal of Theoretical Biology, Vol. 41, 1973, pp. 535-546. 4. Isard, W., Ecologic-Economic Analysis for Regional Development, The Free Press, New York, 1972. 5. Leontief, W., "Quantitative input-output relations in the economic system of the United States," The Review of Economics and Statistics, Vol. 18, 1936. pp. 105-125. 6. Odum, H.T., "Trophic structure and productivity of Silver Springs, Florida", Ecological Monographs, Vol. 27, 1957, pp. 55-112. 7. Ryther, J.H. and W.M. Dunstan, "Nitrogen, phosphorus, and eutrophication in the coastal marine environment," Science, Vol. 171, pp. 1003-1013. 8. Smith, L.J. and J.E. Huguenin, "The economics of waste water - aquaculture systems," IEEE QCEAN '75 Proceedings, 1975, pp. 285-293. 9. Teal, J.M., "Energy flow in the salt marsh ecosystem of Georgia," Ecology, Vol. 43, 1962, pp. 614-624. 10. Valiela, I. and J.M. Teal, "Inputs, outputs and interconversions of nitrogen in a salt marsh ecosystem," Paper presented at the First European Ecological Symposium, Ecolo- gical Processes in Coastal Environments, Norwich, U.K., 1977. 224 UNDERWATER PROFILING OF ICEBERGS USING SUBMERSIBLES B.P. Sukhov NORflCO Limited P.O. Box 8833 St. John's, Newfoundland, Canada AlB 3T2 Abstract 3. Using a side scan sonar operating in a verti- cal mode. This technique was used with suc- A new method is proposed for iceberg draft cess by R.Q. Robe in 1974 [4]. An attempt to tethered submersible as an underwater platform scattered acoustical waves and a scanning sonar equipped with a depth meter, an up-looking echo is also worth mentioning here [5,6]. Both a sounder or an up-looking scanning sonar. Block side scan sonar and a scanning sonar were also diagrams and instrumentation for measurement and used for underwater profiling of either ice- processing of data are presented. Basic measure- bergs [7] or pack ice [8]. The accuracy of nent equations and accuracy of measurement are the above methods proved to be very low. analyzed by two methods which give comparable Therefore, there is a need to propose new results. techniques. 2. Basic Principle of Iceberg Drpft 1. Introduction Measurement and Underwater Profiling Icebergs transported by the Labrador Current A. Echo Sounder Mode of Operation from the northern Baf fin Sea are a great danger for oil well heads and other equipment installed An underwater platform is equipped with a on the sea bottom such as pipes and cables. depth sensor, up-looking and down-looking echo Since icebergs may ground and destroy equipment sounders and moves under the iceberg (Fig. 1 and installed on the bottom, the offshore oil indus- 2). The down-looking echo sounder measures a dis- try requires statistical iceberg draft data which tance from the platform to bottom, CD = h .The can be used to estimate the probability of a break det esrmaursapafr' epth 2'i.e. or disaster. Iceberg draft is also of interest for 2'itneAa ploin cosudrmaue shipping and other marine activities and the a vertical distance from platform up to the ice- effect it may have on drift, groundings and de- berg, i.e. distance CB. Iceberg draft, d, in terioration. point B immediately above the platform could be calculated as d = h 2-h. Measurement of these The measurement of the physical properties parameters during a2platform movement gives a of icebergs is also of great interest in con- cross-section profile of an iceberg's underwater nection with their use as a source of fresh water and energy. Projects for utilizing ice- bergs are very dependent upon the success of long distance towing operations. Knowledge of the iceberg's draft, underwater profile, volume and mass is of great fundamental importance to such operations. Regarding the iceberg draft measurement problem, there are three methods which have been da used with varying degrees of success. h 1. Measuring the water depth close to a ground- h ed iceberg. It was first done by H. Rodman ' in 1890 [1] and later by W.B. Dowson in 1907 in the Strait of Belle Isle [2]. 2. Dragging a cable with a depth sensor attached underneath the iceberg by means of two ships passing on either side of the berg [3]. Fig. I BASIC PRINCIPLE OF MEASUREMENT 225 A' - II I Current l I max Fig. 2 ECHO SOUNDER MODE of OF OPERATION part. Maximal measured difference, d = h - 2. a towed body An' is determined as d for the lowestmpaoint , 3. a manned submersible d.e. an iceberg draft. 4. a military submarine 5. aneerpu tthere o unmanned submersible, tethered or untethered. B. Sonar Mode of Operation Analysis with all pros and cons has been given An underwater platform is equipped with a in a previous paper with the conclusion that a most depth sensor and an up-looking sonar scanning in suitable platform for iceberg measurements is an 1 a plane perpendicular to the direction of platform unmanned, tethered submersible (Fig. 2). Mode of movement (Fig. 3). Here the platform is in point operation for both the manned and unmanned submer- C, and it moves in a direction perpendicular to sible has been described as well [9]. the plane of the drawing. Scanning angle sensor gives the angle a between a vertical line AD and 3. Instrumentation a sonar beam CF. p is the distance between the platform and the iceberg measured by sonar. Thus, A block diagram of the system suitable for the the iceberg draft in point F can be calculated as: echo sounder mode of operation is shown in Fig. 4. The wide view recorder sends a trigger signal to the d = h2 - hi = h2 - P cos a clock and generators of both the up and down-looking echo sounders. The clock starts to measure time. Every sonar scan gives a cross-section semi- Echo signals from both the iceberg and bottom stop profile, i.e. a profile up to the deepest point. the clock and as a result, its output is hi - the For full profile, we need a second run at the other distance from the submersible to the iceberg and side of the deepest point E. Again a measurement5h3 - the distance from the submersible to the bottom. for the deepest point represents an iceberg draft D ata from the sound velocity sensor is used for dis- d ar tance measurement corrections. The sign invertor changes the parameter hi into - hi and puts it in For in situ measurements, instruments and the operational amplifier-summator. The second sensors should be delivered under the iceberg. input of the operational amplifier is the submer- There are various ways in which this could be sible depth h2, i.e. the distance from the submer- accomplished: sible to the surface. Its output is hs + (-hu) = d, the iceberg draft at the point immediately above 1. a scuba diver the submersible. So three parameters, d, h3 and h2 226 vertical position, I, the transducer rotates in the _7 - - ~- horizontal plane and looks forward. This mode of A:' ~~~~~~~~operation is used in approaching the iceberg for a rough estimation of the iceberg draft and underwater * d profile. After lowering the submersible to its dmax working depth with sufficient clearance below the dh: iceberg the pilot rotates the tilt mechanism to the -B1 horizontal position, 1I, so that the transducer -"F ~~~~~~scans in a vertical plane perpendicular to the GI&,~~ movement direction of the submersible. The following data is fed into the processor; h3 ~~~~~~~~~~the distance to the iceberg, p, from a transducer, an angle, a, from an angle sensor and the submer- ID ~~~~~~~~sible position relative to the surface, the depth /I Z / h2' ~~~~~from the depth sensor. It calculates the draft, Fig. 3 SONAR MODE OF OPERATION volume and mass of the iceberg as in the previous case. A top view plotter gives an iceberg's pro- file in the form of isoplanes or isolines of equal go into the processor which inputs them into a side depth. view recorder. Both the recorder and the processor work in the following sequence: 4. Accuracy Analysis 1. A trigger signal forces the recorder's pen Let us now investigate the accuracy and to place a mark on the paper strip which possible errors of the proposed modes of operation. represents the surface. The same signal is This is a typical case of indirect measurement; inserted in the processor. that is, iceberg draft cannot be measured directly but is calculated after the other two or three 2 . After a certain time proportional to the parameters have been measured. local iceberg draft, the processor sends a signal to the recorder and the pen makes a Basically, the accuracy of the echo sounder second mark which represents a profile line mode depends on accuracy of the depth meter and the of the underwater part of the iceberg. echo sounder to be used. 3. After a time proportional to h , the distance Total error of method might be found from the between the surface and the submersible, the following simple calculation: processor sends another signal to the record- er. its pen makes a third mark which repres- d =(h2 �A h 2) (hI � A h) I ents the track of the submersible. h2 ~Ah 2-h ~Ah = 4. After a time proportional to h ,the distance2- 2 1- 1 from the submersible to the bot'om, the pen h2 -h ~Ah 2 A hi makes a fourth mark.2 1 2 1 h2 -h (A h As a result, we have on the recorder diagram 2 1- h 1 a series of underwater profiles and the track of h2- hI+ A d the submersible between the iceberg and the bottom2 1- (Fig. 5). IA dl = A h2j + JA h11 The processor calculates the form, volume and where A h2- absolute error of depth meter mass of the underwater part of the iceberg. Be-2 sides this, with the help of a keyboard on the A hI- absolute error of echo sounder ship, data concerning the above water portion of1 the iceberg (width and height) can be imputed to A d -absolute error of a measured the processor. It then calculates a volume and berg's draft mass of the entire iceberg and constructs,on the CRT screen and top view plotter the underwater Practically, A d is equal to the sum of profile of the iceberg consisting of isodepth A h + A h lines or isodepth planes. The position of both I 2' the ship and submersible relative to the iceberg Now let us discuss an error of the second could be represented on both the CRT screen and mode of operation, i.e. a submersible equipped top view recorder using data from a positioning with a scanning sonar. The problem could be system. approached by two methods -an analytical method and a direct method. A block diagram for the sonar mode of opera- tion is shown in Fig. 6. A continuously rotating transducer is installed on a tilting mechanism which can be in either of two positions. In the 227 | UP | I | UNIT ABOVE h, WWATER hi -hi h +(-h)=d WATE h3 o SUPPLY SIDE VIEW h RECORDER ~~~~~~DOWN h3PROFILER UNDER WATER PROFILE VOLUME MASS Fig. 4 BLOCK DIAGRAM FOR THE ECHO SOUNDER MODE A. Analytical Method JAAI = JABI + IACDI + IADCj + IACADI Let us consider the basic equation Remember: AC = Ap d = h2 - P cos a C = p d + A d = h2 + A h2 - (p � A p) cos (a + A a) D = cos a This can be represented in the following form AD = A(cos a) A =B-CD or A + a A = B � A B - (C + a C) (D + A D) Surface This last term can have four possible forms (C + AC) (D + AD) = CD + ACD + ADC + ACAD (C + AC) (D - AD) = CD + ACD - ADC - ACAD h2 (C - AC) (D + AD) = CD - ACD + ADC - ACAD th Track of (C - AC) (D - AD) = CD - ACD - ADC + ACAD , ubmersible Therefore h A �+ AA = B + AB - [CD + A(CD)] A(CD)max CD + ADC + ACAD Fig. 5 RECORDER DIAGRAM FOR THE ECHO SOUNDER MODE 228 FORWARD, - -- -, I - AI, II - -. - lSONAR POSITION UP/ ' ' I INDICATOR DOWN hr ) PROCESSOR DEPTHI2V mI BOARD TOP VIEW PLOTTER DRAFT PROFILE VOLUME MASS Fig. 6 BLOCK DIAGRAM FOR THE SONAR MODE We need to find A(cos a). So cos (a + An) Let us find Ad as a difference between the = cos An + sin a sin As. But cos Aa Z 1, measured and true values of draft. because An = 0 and cos (a + An) = cos a + sin a sin An. So that Isin a sin AaI is the modulus of + Ad = d - d absolute error of cos a and the equation for Aeas true becomes: - Ad = d - d true meas lAdl = Ah2 + AP cos a + p sin a sin An + Ap sin a A positive error + Ad reaches a maximum when h2 is maximum and p cos a is minimum, i.e. h + sin hA An2 and (p -Ap) cos (a + na) correspondingly. A relative error is equal to A negative error - Ad reaches a maximum when h2 is minimum and p cos a is maximum, i.e. h2 - Yd = + Ad x 100% 2 and (p + Ap) cos (a - An) correspondingly. So d therefore we have: +Admax [(h2 + Ah2) - (p - Ap) cos (a + Au)] B. Direct Method - [h2 - p cos a] According to the basic equation, d = h - p cos a, the error for an iceberg draft is a ?un- Adax [h2 - P cos a] - [(h2 - Ah2) - ction: (p + Ap) cos (a - Au)] Ad = f(Ah2, , Ap, ) 229 5. Conclusion Calculations have been taken using real parameters of modern depth meters, echo sounders and sonars available on market. These calcula- tions show that both methods give results which are in good agreement and the accuracy of iceberg draft measurement and profiling using a submer- sible in both modes of operation would be approxi- mately � 1%. References 1. H. Rodman, "Report of Ice and Ice Movements in North Atlantic Ocean", U.S. Navy Hydrographic Office, 1890, No. 93 2. W.B. Dawson, "The Currents in Belle Isle Strait", Department of Marine and Fisheries, Ottawa, 1907. 3. E.H. Smith, "International Ice Patrol", The Meteorological Magazine, The Meteorological Office, Air Ministry, 1925, Vol. 60, No. 718, pp. 229 - 236. 4. R.Q. Robe, "Height to Draft Ratios of Icebergs", POAC '72, Third International Conference on Port and Ocean Engineering under Arctic Condi- tions, 1975, Vol. 1, pp. 407 - 415. 5. J.M. Berkson, E.S. Clay, T.K. Kan, "Mapping the Underside of Arctic Sea Ice by Back- scattered Sound", Journal of the Acoustical Society of America, 1973, Vol. 53, pp. 777 -781. 6. T.K. Kan, C.S. Clay, J.M. Berkson, "Sonar Mapping of the Underside of Pack Ice", Journal of Geophysical Research, 1974, Vol. 79, No. 3, pp. 483 - 488. 7. C.P. Benedict, "Underwater Profiling of Icebergs", POAC '73, Second International Conference on Port and Ocean Engineering Under Arctic Conditions, 1973, pp. 601 - 607. 8. T. Pousi, M. Luukkala, "A Narrow Beam Sonar to Measure the Submarine Profile of an Ice Ridge", POAC '75, Third International Con- ference on Port and Ocean Engineering Under Arctic Conditions, 1975, Vol. 1, pp. 389 - 406. 9. B.P. Sukhov, "Measurement of Iceberg Draft", First International Conference on Iceberg Utilization, Ames, Iowa, 1977. 230 MONITORING SYSTEM FOR A NEW OCEAN CABLE PLOW K. Yamamura Y. Iwamoto Y. Shirasaki Research & Development Research & Development Research & Development Laboratories Laboratories Laboratories Kokusai Denshin Denwa Co., Kokusai Denshin Denwa Co., Kokusai Denshin Denwa Co., Ltd. (KDD) Ltd. (KDD) Ltd. (KDD) 1-23, 2-chome, Nakameguro, 1-23, 2-chome, Nakameguro, 1-23, 2-chome, Nakameguro, Meguro-ku, Tokyo, Japan Meguro-ku, Tokyo, Japan Meguro-ku, Tokyo, Japan Abstract and cable system construction cost. Because the buried cable will be protected against damage On the continental shelves, it is desira- caused by fishing gear, and also will not inter- ble to bury ocean cables under the sea bottom fere with fishing activities and moreover the from the point of view of coexisting with fish- cable to be buried will be fairly cheaper than ing activities. A new type of cable plow with the armored cable. seven blades developed by KDD is designed to KDD has conducted investigations into a have a capability of burying the cable to a cable burying plow3 since 1962, because the plow depth of 70 cm under the sea bottom at a towing would be indispensable for constructing cable speed up to about 3 knots. Since the plow is systems on the vast continental shelves around equipped with no power mechanism and is rela- Japan and Southeast Asia. tively small in size and light in weight, the The cable burying system is composed of a effective burying works are achieved by adequate cable plow and a cable ship, so that the cable operation of the cable ship. This paper de- plow must be designed to be suitable for the scribes the cable plow monitoring system for performance of the cable ship. The new cable obtaining enough information for the operation plow developed by KDD was designed to adapt the of the cable ship. A new ocean telephone cable cable ship KDD-Maru (4300 tons). system was constructed between Japan and China As shown in Figure 1, the cable burying in 1976. KDD succeeded in burying the cable plow with seven blades is relatively small in about 700 km across the continental shelf. size (7 m length) and light in weight (6.5 tons). The plow is composed of a balancing wing and a digging blades. The balancing wing is 1. Introduction for stabilizing the plow attitude, and the dig- ging blades is for digging a trench in the sea The ocean telephone cable system plays an bottom. important role in international telecommunica- The pin-joint coupling of the balancing tion network together with the satellite com- wing with the digging blades enables the plow to munication system. get over small obstacles on the sea bottom and Five international cable systems are now to be towed stably even at sandwave area. working in Japan, among which Japan-China Cable Systeml,2,3 was constructed in summer 1976. Since the cable route crossed the vast conti- nental shelf in the East China Sea, the cable Balancing wing of about 700 km which corresponded to about 80% of the total cable length was buried success- fully under the sea bottom. Digging blades The continental shelf abounds in fishing resources, where fishing activities are carried out extensively. Most of cable troubles which have occurred on the continental shelf area in Cable-leading the past were caused by fishing activities, equipment especially trawling4. To protect the cable against the damage the following two methods have been adopted: (1) protecting the cable by armor wire (2) burying the cable under the sea bottom In the area where the latter method is applicable, it is more desirable to bury the ocean cable under the sea bottom from the point of view of coexisting with fishing activities Figure 1 Cable plow 231 The plow towed by the cable ship is de- the number of components, duplication as much signed to bury several hundred kilometers of as possible, and simplicity in mechanical the cable to a depth of 70 cm under the sea configuration. bottom at a speed up to about 3 knots. The cable and repeaters are smoothly guided to the plow through the lots of guiding rings which 2. Monitoring System form a path from the ship to the plow, and are laid on the bottom of the trench dug by the plow. Figure 2 shows the operating configura- Configuration of monitoring system tion of the cable burying works. As the cable is fed from the ship to the plow through the The cable plow monitoring system is mainly guiding rings which are supported by the main consists of the following: (1) sensors on the towing rope and auxiliary rope, a complicated plow, (2) multiconductor cable, and (3) monitor- control system for feeding the cable is un- ing equipment on board as shown in Figure 3. necessary. The sensors on the plow are designed for The cable burying system developed by KDD 300 m of water depth, because the operating is relatively simple, but has sufficient func- depth is less than 200 m. Most of the sensors tion for burying the cable and repeaters. But are installed in the balancing wing of the plow the plow does not have any active mechanism to be protected against various hazard on board such as an attitude stabilizer and a plowshare or on the sea bottom, which consequently does suspension, accordingly, the effective burying not impose any restrictions on the handling of works are achieved by adequate operation of the the plow. cable ship, for example, towing speed, towing Each signal detected by each sensor is rope length. transmitted to the monitoring equipment on The important information required for board through the multiconductor cable. Multi- adequate operation of the cable ship and the plex transmission technique which is usually plow has to be given from the plow monitoring adopted for transmitting the lots of signals is system. But the plow structure is not suitable not adopted in this system, because simplifi- for equipping the sensors to monitor the condi- cation of the monitoring system, facilitation tion of the plow. In the plow the space for of the system check, and reduction of the the sensors and the type of the sensors are influence of a partial trouble of transmission subject to restriction, because the sensors cable such as broken wire are thought more which may restrict the plow handling on board important. or the plow lowering and recovery will degrade The monitoring equipment on board displays the work efficiency excessively. several kinds of information as to the plow on Within these constraints, the cable plow the panel-meters, pen recorders and also dis- monitoring system was satisfactorily developed plays the plow working on the sea bottom on a to obtain enough information as to the plow on model plow simulator, by processing the signals the sea bottom for carrying out the burying transmitted from the sensors. works. To avoid an interruption of the burying Monitoring items works due to a trouble of the monitoring sys- tem, high reliability and long life are attain- The number of monitoring items as to the ed by adopting the following basic principles: plow is ten as follows, which is decided in minimum interdependence of one functional consideration of the structure of the plow and element upon another, use of reliable compo- its working conditions on the sea bottom. nents supplied on the market, minimization of Cable ship KDD-MARU Ocean tug- boat Winch for auxiliary rope / = - .- -- Tua ine ,-__ Winc for main Monitoring ,ea ,eVI towing rope equipment Cable Main towing rope Main towing guiding ring rope Cable Auxiliary rope Cable plo Auxiliary rope Side lookinm ____ ~ ~ sonar fish _ Sea bottom Cable Figure 2 Cable burying works 232 MULTI CONDUCTOR (13) tension of auxiliary rope SENSORS CABLE MONITORING EQUIPMENT On the other hand, an ocean tug-boat which ~ Cable plow Sea Cable ship ' keeps the cable ship on the cable laying course tows a side-looking sonar fish for surveying F Pitch L obstacles such as rock, sunken ship. I Sensor The plow attitude is monitored in items | SesRoll D isplay (1), (2), and (3) mentioned above which monitor ensor Unit the inclination angle of the plow to a horizon- IFSensoier ~ tal plane, because stable attitude of the cable plow is necessary to dig trench in the sea | Sensor |Junction Slip Measurement Model plow bottom efficiently. See uRingcUnir SFlipimulator Item (4) is not only useful to grasp the I Sensor Auxil )a r op Winch plow attitude intuitively, but also is used to Sensor I MAuxilindury rope compute the dug trench depth and the cable buri- I Sensor cable) D al depth. I-Louad Pnocessing The operation of the cable press roller CI ell ni which presses the cable into the trench bottom is monitored in item (5). This item is also used to compute the cable burial depth and Figure 3 Configuration of monitoring system moreover it gives the information whether the repeater passes through the plow smoothly. Whether the balancing wing touches the (1) pitch angle of the balancing wing (pitch surface of the sea bottom or not is monitored angle) in item (6). (2) roll angle of the plow (roll angle) Whether the plow is set on the sea bottom (3) pitch angle of the digging blades of the correctly or upside down is monitored in item plow (inclination angle) (7). This information is especially important (4) bend angle between the balancing wing at the beginning of the cable burying works and the digging blades (bend angle) when the plow is hanged down by the main towing (5) operating press angle of the cable press rope and set on the sea bottom. The plow is roller (press angle) designed to be set on the sea bottom correctly, (6) information whether the balancing wing it is apprehended, however, that the plow is set touches the surface of the sea bottom or upside down where the sea bottom is muddy or not (touch bottom) when tide runs very strong. (7) information whether the plow is set on Tension required to dig trench is monitored the sea bottom correctly or upside down in item (8). On the other hand, item (12) shows (correct setting). the tension which is sum of item (8), weight of (8) towing rope tension guiding rings, fluid resistance to cable and (9) dug trench depth guiding rings, and frictional force caused be- (10) cable burial depth tween the main towing rope and the stern chute In addition to the items mentioned above, of the cable ship. Accordingly, the results the following items as to the ship board fa- obtained from items (8) and (12) are sometimes cilities are also monitored: much different. (11) tension, speed, and length of the cable The depth of the trench dug by the plow to be buried and the cable burial depth are shown in item (9) (12) tension of main towing rope and item (10) respectively, which are computed Pitch sensor Looad cell Sen d sensor Press angle senso- , " /~ d , .. gging, blades/ - '-..~-Cable I I.- 0 I-- _~-burial depth Aux ctry 'rI I I\nchliation sensor = Cable press roller ,lTouch platns - _ Balancing wing / Roll sensor Figure 4 Appearance of the sensors on the plow 233 from the results obtained from items (1), (2), (3), (4) and (5). N 3. Sensors on the Plow SWING AMPLITUDE - 60 Figure 4 shows an appearance of the sen- sors on the plow. A pendulum method is adopted in the sensors that detect pitch angle, roll 6'- 50 angle, and inclination angle. To obtain high confidence of the monitoring system, the check I and calibration of it are carried out thorough- 4 40 ly on board before the plow is hanged down to the sea bottom. But the plow is placed upside down on the deck for reasons of its handling, 2 3�0 accordingly, these sensors should be applicable throughout 360'. o - The sensor is composed of a synchros, an o � aluminum thin circular plate which has an o' o' 2o' 3o' adequate small weight on one point of its edge, TRUE MEAN ANGLE 5 and a magnet as shown in Figure 5. Since the mechanical torque of the synchros is small, the required length of the pendulum, that is, the radius of the circular plate, is short for obtaining required accuracy, accordingly the sensors are composed small in size. Figure 6 shows an equivalent circuit of synchros. When ac voltage is applied to the rotor winds, the voltage across the external stator legs are is true mean angle, and Dg is swing amplitude expressed by the following function of rotor of the pendulum. The error due to such input- shaft anglef output non-linearity of synchros is decreased by adopting a magnetic damping which damps the Ebc = Eb - Ec=- -v/3Em rsin swing amplitude of the pendulum by eddy-current where Em is the maximum rms voltage that is in- effect on the aluminum plate. The accuracy of duced from the stator neutral to any of the the sensor is designed better than +0.5'. external connections. As the sensor is The digging blades has no space for the placed horizontally and is set to make the sensor of inclination angle in it, then the angler of equation (1) zero, the voltage Ebc sensor is installed in the balancing wing. therefore corresponds to the inclination angle Figure 8 shows the sketch of the sensor. As of the sensor to a horizontal plane. But, shown in the figure, the axis of rotation of because of a relation of sinusoidal function the digging blades is combined with the case of between the synchros output Ebc and the angle synchros (stator) through a shaft A, therefore , the average output Ebc measured by electric the pitch of the digging blades is transferred circuits does not agree with the true mean to the synchros. The spur gear A attached co angle P when the swing amplitude of the pendu- the shaft A turns the spur gear B attached to lum is large. Figure 7 shows the error com- the axis of the potentiometer. As the case of puted theoretically, where Af is the error, fB the potentiometer is fixed at the sensor housing, Stotor Eb - r----------- - --3 t Spur lXnet ' c =t;3g~E S haft A Ec circular Spur. LA Digging Watertight Potentiometer blades '-Aluminum thin . ,, circular plate housing Balancing wing � I-Weight Figure 5 Sensor for Figure 6 Equivalent circuit Figure 8 Inclination and bend sensor pitch angle of synchros 234 that is, the balancing wing, the potentiometer Eca is positive. Consequently the information can detect the bending angle between the bal- whether the plow is set on the sea bottom cor- ancing wing and the digging blades. This rectly is obtained by monitoring the polarity potentiometer is required to have such charac- of the voltage Eca. teristics as long rotational life and high The towing rope tension is detected by a resolution. watertight load cell (max. 30 tons). A potentiometer is also used in the sen- The dug trench depth and cable burial sor of the cable press roller for detection of depth are computed with the above-mentioned an operating angle of the cable press roller. various angles. These depths are able to be As the sensor can be composed small in size, computed by two methods, which are intended to it is installed near the axis of rotation of increase the reliability in monitoring the the press roller. depth. The fundamentals of the measurement is To detect whether the balancing wing shown in Figure 9. The following relationship touches the surface of the sea bottom, it is is established among pitch angle 0l, bend angle advisable for the sensor to be non-touch type 02, and inclination angle 9, from the point of view of handling and relia- = 92 - 91 (3) bility. But it is very difficult for our plow to install such non-touch type sensor as an In Figure 9, point A shows a jointing point be- ultrasonic or optical sensor. Therefore a tween the balancing wing and the digging blades, following touch-type sensor is developed as point B shows the axis of rotation of the cable shown in Photo. 1. The sensor is composed of press roller, point C shows the top of the a touch-plate mechanism and a potentiometer, seventh plow-blade, and point G shows the point the former touches the sea bottom and the where the balancing wing touches the surface of latter detects the rotational movement of its the sea bottom. The dug trench depth DI is plate. In order to protect the sensor against given by; various damages caused by bump, and to compose duplication, two touch-plates with enough flexibility are installed in each narrow space between the balancing wing and the digging weeA I C=LA 3 n between the balancig in adethe digging Using the relationship of equation (3), D1 is given by; D1 = L2cos(92-01) - Llsin(92-0l) - L3 (4)' Touch-plate On the other hand, the cable burial depth D2 is given by; D2 = L4cos (9+a) - LI sin G + L5 - L3 (5) or D2= L4cos(G2-91+t) -Llsin(92-91)+L5-L3 (5)' where BE= L4, L5 is radius of the cable press roller and a is operating angle of the cable press roller (LCBE= a). As mentioned above, D1 and D2 are computed from 9, a or 01, 02,a � The measurement error of DI, D2 is less than �2 cm, the main cause of this error is due to a Digging blades little looseness of the joint between the bal- ancing wing and the digging blades. As the parts on the market which are mass Photo. 1 Touch sensor produced are adopted in the sensors, the cost of the sensors is low. The detecting mechanism is simplified to decrease the parts which have The information whether the plow is set on possibility of trouble. the sea bottom correctly is obtained from a part of signal of the pitch or roll angle sen- sor. In Figure 6, another voltage across the external stator legs is expressed by the fol- lowing function Balancing wing Cabl dess Diggingblades[ Radius Ls) Eca EC Ea<3 Ensin 6 0') (2) ea -'- T[ T- ~__~ -- .,.~, / /bottom where2 Eca<O ; -120� <p<60� Cable Eca>O ; 60 < <240o C Cable' When the plow is set on the sea bottom correct- Trench bottom ly, the angle p is generally in the range from -30� to +300, therefore the voltage Eca is negative. On the other hand, when the plow is set upside down on the sea bottom, the voltage Figure 9 Fundamentals of the depth measurement 235 4. Multiconductor Cable The computing circuits are designed to suit the approximate formulas (6) and (7) The multiconductor cable is used to trans- respectively in order to simplify the circuits. mit the monitoring signals from the cable plow The computation error is, however, less than a to the monitoring equipment on board, on the few percentage which is not a serious problem other hand, it plays a part in supporting the practically. cable guiding rings together with the main tow- Under normal conditions, the frequency of ing rope as shown in Figure 2. Consequently troubles of monitoring circuits is very low. the cable suffers a stretching and bending It is important, however, that the monitoring force repeatedly at the stern chute of the circuits function properly under the very cable ship and the cable-leading equipments of severe conditions for a long time while the the plow. Figure 10 shows the armored multicon- burying works are carried out successively. ductor cable. A breakdown tension of the cable Therefore, additional circuits for check and is about 15 tons. The cable is designed tor- calibration of the monitoring circuits are pro- sionless structure and contains thirty-five vided on this unit in order to obtain high conductors. The cable tension usually is kept confidence by carrying in-service testing less than 2 tons to prevent breaking down of periodically. the conductors. %Tr, t i.ple tiNsted d Display unit conductor Twin iwmsted conduc/o, hlaroprene jacket Measurement unit Tension members Polyethylene jacket Figure 10 Multiconductor cable Data processing unit 5. Monitoring Equipment on Board The monitoring equipment on board which is mainly composed of measurement unit, display unit, data processing unit and model plow simu- lator, is designed for easy operation and high reliability. Photo. 2 shows this equipment. The basic functions of the measurement unit are to receive the signals transmitted from the sensors on the plow, to process the received signal so as to show the working conditions of the plow on the sea bottom, to transmit the processed signals to the display unit and model plow simulator, and to check and calibrate throughout the monitoring circuits. The dug trench depth and the cable burial depth are computed by analog computing circuits equipped in this unit, which are mainly com- posed of several operational amplifiers. The formulas (4) and (5) are expressed in following formulas (6) and (7) respectively when the angles 0, a are less than 100 which is true in Model plow simulator ( Bridge ) the real case. D1 1L2 - L1 -L3 (6) D2 _ L4 + L5 - L10 - L3 (7) Photo 2. Monitoring equipment on board 236 At the display unit, the panel-meters 7. Conclusion indicate important information such as pitch angle, roll angle, bend angle, trencb depth, This paper describes the plow monitoring burial depth, and towing rope tension. The system for the new ocean cable plow developed operation of the touch sensor and information by KDD. of correct setting are displayed on indication The monitoring system is designed to pro- lamps. vide enough information for the effective cable The signals of pitch, roll and bend angle burying works. are transferred to the model plow simulator The sensors on the plow are designed to unit, in which the model plow is actuated to have enough reliability against the hazards of simulate precisely the plow attitude, in servo the ocean and severe working conditions on mechanism. The steering and speed control of board or on the sea bottom. the cable ship and adjusting of the towing rope The construction of Japan-China Cable Sys- length will be carried out adequately by refer- tem in which the cable of about 500 km was ring the model attitude. Displaying with such buried successively has been the first of its model is much helpful for the ship operator to kind in the world. grasp the plow attitude than with meters or two dimensional display unit such as CRT. A microcomputer is equipped in the data 8. Acknowledgement processing unit to process the information which is important as a cable laying work re- The authors would like to thank Dr. S. cord. In this unit, average value, maximum Shimura, Dr. Y. Nakagome, Dr. H. Kaji, and Dr. value, and minimum value of the data detected H. Yokoi for their guidance and encouragement. by sensors are computed and printed out peri- They also wish to thank Dr. T. Ezoe, Mr. K. odically. Automatic data correction which Kobayashi, Mr. H. Sugimoto, and Mr. K. Shirai, decreases the error caused by drifts of charac- for their helpful cooperation, who worked on teristics of measurement circuits is adopted the development of the cable plow and the aux- in this unit. iliary rope. Appreciation is also expressed to Mr. M. Yamashita for his helpful cooperation in the work which is closely relating to the 6. Cable Laying Works contents of this paper. Japan-China Cable System was constructed in summer 1976. This cable system was designed References to have a capacity of 480 voice-grade circuits by employing CS-5M submarine cable system which 1. M. Itano, "Burying Work of the Japan-China had been developed in Japan. Submarine Cable," ITU Telecommunication Since the cable route crossed the vast Journal, Vol. 44, No. 7, 1977. continental shelf in the East China Sea, spe- cial countermeasures were taken to solve 2. S. Shimura, 0. Kameda, "Japan-China Sub- problems arising from cable laying in shallow marine Cable," Journal of IECE of Japan, water over long distance. Their major points Vol. 61, No. 5, 1978. were adoption of temperature controlled varia- ble-gain repeaters and cable burying under 3. K. Suzuki, F. Kinoshita, "New Submarine the sea bottom. Cable Burying System," Journal of IECE of The cable laying works were carried out Japan, Vol. 61, No. 5, 1978. dividing into four times. In second works which succeeded first works carried out at the 4. H. A. Baxter, R. E. Mueser, "The Develop- offshore area of China, the cable of about ment of Ocean Cable Plows," IEEE Trans. on 500 km including 40 repeaters and equalizers Communication Technology, Vol. COM-19, -was buried satisfactorily under the sea bottom No. 6, 1971. day and night for six days successively. In this area, water depth changed from 40 m to 140 m gradually. The cable burying speed was average 2.0 knots (max. 4.2 kt, min. 0.4 kt), and the cable burial depth was from 50 cm to 130 cm. The plow monitoring system which was adopted in this works worked perfectly during all construction works without any troubles. By virtue of burying the cable, no trouble has ever occurred since the system was con- structed. 237 DISPLAY AIDS FOR REMOTE CONTROL OF UNTETHERED UNDERSEA VEHICLES* W. L. Verplank Man-Machine Systems Laboratory Department of Mechanical Engineering Massachusetts Institute of Technology Cambridge, Massachusetts 02139 low resolution picture of 80 K bits and a channel Abstract capacity of 10K bits/sec., there would be at most, one picture every 8 seconds (S = 8 seconds). A "predictor" display superimposed on slow- scan video or sonar data is proposed as a method The effects of trying to navigate with just to allow better remote manual control of an u this pictorial information are illustrated in tethered submersible. Simulation experiments show Figure 1. good control under circumstances which otherwise make control practically impossible. 6C~a 1. Introduction Ii S JTpu cw Untethered, unmanned submersibles have been limited to automatic control on simple pre-pro- - grammed or target-seeking trajectories. More precise navigation and obstacle avoidance will require increasingly sophisticated automatic con- cii IN a ] UIS -flt [][ trol and/or direct control from the surface, t{ez ske. Direct human control through a sonic communication channel will be difficult because of the low band- width and the signal travel time. Probably the Figure la, Effect of delays from transport most productive approach will be a combination of time (T) and scan time (S). elementary automatic control such as is possible with some present-day tethered submersibles (e.g., The picture from I is received T + S seconds altitude or depth and heading control) plus after it is taken; the first operator response is display aids which make control easier for the received by the vehicle at least T seconds later, operator. This paper proposes a display aid for a total delay of 2T + S seconds. While the which is particularly applicable to the problems operator is looking at the still picture from of time-delay and slow frame-rate. the commands he is sending are actually moving the vehicle from 1' to 2', as illustrated in 2. The Problem Figure lb. For remote control, there are two sources of difficulty with sonar communications: time-delay fced VkLT and slow-frame-rate. Round trip time-delay is the time for a command to travel to the vehicle and / the first indication of response to travel back. At a minimum this will be two times the distance /kde n divided by the speed of propagation, 2T. * For example, T = 1 second at about 5,000 feet. t" ric Pictorial information from television camera / or obstacle avoidance sonar will be further delayed because of limited channel capacity. Assuming a This work was supported in part by ONR Contract Figure lb. Positions of vehicle at times in Fig. la. N00014-77-C-0256. The untethered vehicle control problem was suggested in discussions with the Harbor Branch Foundation and the M.I.T. Office of Sea Grant. 238 3. Predictor Display Predictor displays were first used for submar- ine control (Kelley, 1968). NASA considered pre- dictor dispalys for remote control of unmanned SaV Wd lunar roving vehicles, (Arnold, 1963) but sent tnt btt men instead. ,Ct-,F- The predictor display proposed here presents -AT)- a symbol superimposed on the slow-frame-rate . ..... . . and time-delayed picture from the vehicle's television camera. The symbol responds instan- taneously and continuously to the operator's commands predicting "future` positions of the vehicle. For example, refering to Figure lb, when is complete the predictor symbol would show Figure 3. Delays associated with the position 1' . Before the next picture from predictor calculation El arrives, the symbol will be moved, in res- ponse to the operator's commands, to position 2' Adaptive Estimation The position of the vehicle is computed from Another feature that could be built into the a local model of the vehicle responses and the local model of the vehicle is some estimate of the operator's commands u(t), as shown in Figure 2. disturbances (such as current), The current model as well as the vehicle model could be updated on the basis of the mismatch between predicted and measured vehicle position. slow ~~~~~~~~4. A Dmonstration Experiment In order to explore the effects of the pre- dictor display, an interactive simulation was Q) v~~~~~~ICLE ~~~~~written on an Interdata 70 computer and Imlac graphic display. A random terrain was generated and displayed in perspective, updated every 8 seconds, to simulate the pictorial information. A moving predictor symbol was generated respresent- in the vehicle as a square in perspective. Two straight ridges were added to the random terrain to serve as a test course. (Figure 4). Figure 2. Predictor display superimposed on pictorial data Pictorial or Map Displays The predictor symbol may prove useful both on pictorial displays (superimposed on television or obstacle-avoidance sonar) and on map-like position diaplays. Map displays would avoid one difficulty of pictorial displays,which is loosing the pre- dictor symbol when it moves out of the field of view of the camera (for example, moving sideways or backward). Auxiliary Position Data Figure 4. Computer-generated terrain If position data is available from transponders with predictor symbol or locator beacons, it could be used to update the vehicle model. With just the pictorial data, The simulated vehicle was controlled by the the open-loop prediction would have to span an operator with a spring-centered 2-degrees-of- interval of (at least) 2T + S to (at most) freedom joystick. The dynamic response of the 2T + 2S seconds. With auxiliary feedback the vehicle was simple integration with forward speed open loop estimate will only need to span the proportional to forward-back position of the stick delay of that auxiliary data (at minimum 2T). The and turn-rate proportional to left-right position signals and corresponding delays are shown in of the stick. The vehicle was always the same Figure 3. (u(),command vector; x(-), vehicle height above the terrain (simulating automatic location data). altitude hold). No disturbances such as currents were simulated. Also, it was found important to 239 have a good detent and dead-zone on the stick to Only with extremely slow speed was it possible avoid inadvertant commands. to keep track of the ridge. Approximately five minutes and 40 pictures were required to traverse A stationary "table" was drawn to indicate just one of the ridges (half the course). This where the next picture was to come from while the is shown in Figure 7. "1real-time" predictor continued to move in response to the operator's commands (Figure 5). Dotted lines were added to this table to indicate the field of view. This reduced the considerable confusion about how the picture was expected to change, and served as a guide for keeping the vehicle within its own field of view, which is the best strategy for using this kind of predictor on the pictorial display. I Figure 5. Predictor plus "table" showing from where next picture will come Results ~~~~~~~~~~Figure 7. Success at slow speed with no predictor With the predictor symbol, practically con- A typical path, without the predictor, is tinuous motion was possible. A typical path is shown in Figure 6. The dotted lines represent shown in Figure 8. The course was completed in + I terrain-unit from the ridge. The circles 3 minutes and 23 pictures. represent the vehicle's position every 2 seconds. V's represent the field of view of each picture sent. Quite often there is no movement between successive dots (2 sec.) or successive pictures (8 seconds). 24~~~~0~ Request Mode A typical path in request mode (using the predictor) is shown in Figure 10. Compared to One unexpected finding from these experiments periodic mode, the time is about the same but the number of pictures used is one-half to one-third; icallysevery eghat rathecon sending the picture pvelocities are higher but there is a wait for 10 ically every eight seconds, sending the picture scnsa ahpcuei ae n et only upon the operator's request reduces the total number of pictures necessary, and encourages On an actual vehicle, probably both modes a "move and wait" strategy which avoids confusion. should be available with the request mode used when The difference is illustrated in Figure 9. move-and-wait strategy is appropriate (for precise positioning based on pictorial feed-back, and when environmental disturbances are small). Periodic (I�WC .aa~d mode is probably more appropriate for less precise J- r- I navigation and continuous motion when the pre- []I AX ~dictor symbol can be relied upon. ! ! /t\~~1 /f ! / Another trade-off that should probably be ._ built into the pictorial feedback is variable frame-rate/resolution. In a more dynamic and uncertain environment (i.e., larger bandwidth disturbances or target motion) sampling rate Figure 9a. New picture every 8 seconds: will want to be higher at the expense of resolu- "periodic mode" tion. 6. Conclusions and Recommendations For the conditions studied (T = 1 sec., S = 8 sec.) manual control is not feasible without display aids such as the predictor symbol. The ftRiA request mode is preferred as it seems to avoid confusion and reduce the number of pictures !V/f t /v f ! necessary. Fab If -1 t i-m e The present results are at best preliminary. We studied only very simple vehicle dynamics and only one set of delay conditions. Further study Figure 9b. New picture upon request: with laboratory simulation can investigate: ?'request mode" 1) more realistic vehicle dynamics, In the periodic mode (Figure 9a) a short move starting with the receipt of picture RI will not 2) environmental uncertainties such as drift, be reflected in the next picture, 9 , as the o rator might expect; instead he has to wait for 3) a broader range of delay conditions and W. In request mode (Figure 9b), the wait for pictorial confirmation is minimized. 4) various degrees of partial automation, Also, the predictor displays (both pictorial and map) could be used on existing tethered vehi- cles to simulate untethered operation and eval- uate the potential for untethered operation. References 0 ol i� I Kelley, C. R., Manual and Automatic Control, Wiley, New York, 1968 Arnold, J. E., and Braisted, P. W., Design and Evaluation of a Predictor for Remote Control Systems Operating with Signal Transmission Delays. NASA TN D2229, 1963. Figure 10. Typical path in the request mode 241 MAJOR MARINE LEGISLATIVE ACTIONS IN THE 95TH CONGRESS Alfred W. Anderson* Office of General Counsel National Oceanic and Atmospheric Administration 6001 Executive Boulevard Rockville, MD 20852 little difference between a bill and a joint resolution. A joint resolution from the House Abstract is labeled "H.J. Res." A joint resolution from the Senate is labeled "S.J. Res." The legislative process provides safeguards frThe legislate docr esys provides s Concurrent and simple resolutions are matters for the American democratic system. It provides affecting the operations of the House and Senate. a method whereby the people can govern themselves and direct their destiny and, to an extent, deter- Bills introduced in the House are placed in mine their fate. ~~~~~~Bills introduced in the House are placed in mine their fate. In marine, as in other affairs, the "hopper." Bills introduced in the Senate are the "hopper." Bills introduced in the Senate are the U.S. has been far from perfect, but we have introduced on the floor and may or may not be introduced on the floor and may or may not be the chance each year to legislatively reconsider read for the Members. Once introduced, a bill is our marine problems. ~~our marine problems. ~referred to the appropriate committee/subcommittee 1. Introduction for hearings and study. If it meets the approval of the subcommittee/committee, it is reported to The legislative process was designed to be the floor of the respective house. Once a bill The legislative process was designed to be passes in one house, it moves to the other house a practical safeguard of the American democratic and goes through the same process. Ifdifferences and goes through the same process. If differences system. The fact that a proposal cannot become arise, a bill goes to a Conference Committee arise, a bill goes to a Conference Committee law without approval of both houses of Congress where, hopefully, the differences are eliminated and consideration by the President is a tremen- ~~~~dous safeguard. ~or one of the two houses must accept the other's dous safeguard. version. Once differences are resolved, the com- promise bill must then be passed on the floor of Traditionally, legislation originates with a the House and the Senate. Upon passage by both Member of Congress. Yet, constituents (either as houses, the bill is checked for accuracy, printed, individuals or corporate entities or other asso- and signed by the Speaker of the House and the and signed by the Speaker of the House and the ciations) transmit proposals for legislation to President of the Senate. The bill is then President of the Senate. The bill is then Congressmen who may introduce them as they see delivered to the President for action. delivered to the President for action. fit. Many laws originate this way. Executive communications from a Cabinet mem- If the President approves, he signs the bill and it becomes law. By constitutional provision, ber or the President transmitting draft bills to a bill may become law without the President's the Speaker of the House and the President of the iga e I the Presidents signature. If the President objects to a bill, Senate are significant sources of legislation. s tur i t resit object o he must return it to Congress with his objection An executive draft bill is referred to the stand- within 10 days (Sundays excepted) after it has within 10 days (Sundays excepted) after it has ing committee(s) having jurisdiction over the ing committees) having jurisdiction over the been presented to him. If he does not return it subject matter, and the Chairman(men) usually and both houses are still in session, the bill and both houses are still in session, the bill introduce the bill in the form it was received. automatically becomes law. If Congress has adjourned, and the President does not act, the bill Proposed legislation with the most signifi- does not become law; thus, the "pocket veto." If cant effect upon the general population are the "Bills" and the "Joint Resolutions." The bills Congress is in session and the President vetoes a initiating in the House carry the prefix "H.R." bill, it is returned with his objections to the initiating in the House carry the prefix "H.R." h os n e house in which it originated. The House and the while bills originating in the Senate carry the Senate may then pass the bill with two-thirds of prefix "S." the Members present and override the President's veto. If the bill is initially approved by the Joint resolutions may originate in the House Jorinthe Srelnate- oint Ther isve r President or if the bill becomes the law of the or in the Senate--not jointly. There is very land over the objections of the President, it is -.... - ---- ---- ----ready for enrollment and publication as a binding * This paper represents remarks to be made statute. at the September 1978 Meeting of the Marine Technology Society. The views expressed are An enacted bill is sent to the Administrator personal and do not necessarily represent the of the General Services Administration for assign- views or position of the U.S. Government. ment of a Public Law (P.L.) number, paginated 242 for the Statutes at Large covering that session P.L. 95-36, June 1, 1977, authorized appro- * ~~~of Congress, and published as a "slip law." priations thru September 1980 for the administra- tion of the Deepwater Port Act of 1974. After printing by the Government Printing * ~~~Office (GPO), the slip laws are available to P.L. 95-53, June 22, 1977, authorized appro- * ~~~officials and the public immediately. They may priations for fiscal years 1978-80, to carry out be obtained by annual subscription or individual the Commercial Fisheries Research and Development purchase from the Superintendent of Documents, Act of 1964. Funds were provided the Department GPO. They are printed in sequence in the Stat- of Commerce (NOAA) for apportionment to the States utes at Large and later appear in the United to carry out the purposes of the 1964 Act, and States Code. additional funds were authorized to be available to States where there was a commercial fishery 2. Summary of Significant Legislation failure or serious disruption affecting future production due to a resource disaster arising from P.L. 95-6, February 21, 1977, the Fishery natural or undetermined causes. In addition, the p ~~~Conservation Zone Transition Act, gave congres- Secretary of Commerce was authorized funds for use sional approval to certain international fishery by States to develop new commercial fisheries. agreements negotiated in accordance with the Fishery Conservation and Management Act of 1976. P.L. 95-58, June 29, 1977, authorized appro- The Act officially approved the international priations for the National Sea Grant Program for fishery agreements between the United States and the fiscal year 1978. Bulgaria, Rumania, the Republic of China, German Democratic Republic, the Union of Soviet Socialist *P.L. 95-63, June 5, 1977, the National Republics, and the Polish People's Republic, all Advisory Committee on Oceans and Atmosphere concerning fisheries off the coast of the United Act, established qualifications for individuals States. appointed to the National Advisory Committee on Oceans and Atmosphere (NACOA) and authorized Appropriate amendments were made to the appropriations for the Committee for fiscal year Fishery Conservation and Management Act of 1976 1978. The Act established a committee of 18 concerning the definition of the term "governing members. Section 3 defined membership qualifi- international fishery agreement," action by the cations, terms, and duties. This section also Councils, permits, and permit fees. In addition, provided that the President designate one of the it repeals the Northwest Atlantic Fisheries Act members as Chairman and one as Vice Chairman. of 1950. The Act required the Committee to undertake the continuing review of the national ocean policy, Several other laws approved other U.S. fishery coastal zone management, and the status of the agreements as follows: P.L. 95-8, March 3, 1977, marine and atmospheric science and service European Economic Community, Japan, Republic of programs of the United States, and advise the Korea, and Spain; P.L. 95-73, July 26, 1977, Secretary of Commerce with respect to the carry- Canada; and P.L. 95-219, December 28, 1977, ing out of the programs administered by NOAA. The Mexico. Act provided for an annual report, compensation and travel expenses, interagency cooperation and P.L. 95-26, May 4, 1977, made supplemental ap- assistance, and repealed the Act of August 16, 1971, propriations for the fiscal year ending September which established the original advisory committee. 30, 1977, to the United States Fish and Wildlife Section 8 authorized appropriations in the amount Service and the National Oceanic and Atmospheric of $520,000 for the fiscal year ending September Administration (NOAA) for operations, research, 30, 1978. and facilities as well as for coastal zone manage- ment and for the Coastal Energy Impact Fund. The P.L. 95-85, August 2, 1977, authorized appro- Coast Guard also received supplemental appropria- priations for the Department of Transportation. tions for the Pollution Fund. Included are funds for U.S. Coast Guard opera- tions, construction, improvements, and R&D among P.L. 95-33, May 26, 1977, authorized appro- other activities. priations for fiscal years 1978-80, to carry out the Atlantic Tunas Convention Act of 1975. The P.L. 95-86, August 2, 1977, appropriated funds Act of 1975 was amended to define the term for the Departments of State, Justice, Commerce, "fisheries zone" as the waters included within and the Judiciary for the year ending September a zone contiguous to the territorial sea of the 30, 1978. The Act appropriated more than $600 United States, in which the inner boundary is a million for operations, research, and facilities line coterminous with the seaward boundary of of NOAA; more than $50 million for the Coastal each coastal State, and the outer boundary is a Zone Management Fund; and $115 million for the line drawn in such a manner that each point on it Coastal Energy Impact Fund. In addition, the Act is two hundred nautical miles from the baseline appropriated funds for expenses of the Marine from which the territorial sea is measured; or Mammal Commission, as authorized by Title II of similar zones established by other parties to P.L. 92-522. the Convention to the extent that such zones are recognized by the United States. P.L. 95-86 also appropriated funds for the 243 Maritime Administration for, among other activi- Environmental Protection Agency to end the dumping ties, ship construction, research and development, of sewage sludge into the oceans or into waters operations, and training. described in section 101(b) of P.L. 92-532 as soon as possible after November 4, 1977, but in no case P.L. 95-91, August 4, 1977, the Department of may the Administrator issue any permit or any re- Energy Organization Act, established a Department newal thereof after December 31, 1981. in effect, of Energy with, among other things, an Assistant this ends dumping of sewage waste after December Secretary with environmental responsibilities. 31, 1981. The Act defined the term "sewage sludge" as any solid, semisolid, or liquid waste generated P.L. 95-99, August 15, 1977, authorized appro- by a municipal waste-water treatment plant the priations for activities of the National Science ocean dumping of which may unreasonably degrade Foundation. Among the funds authorized for fiscal or endanger human health, welfare, amenities, or year 1978, slightly more than $210 million were the marine environment, ecological system, or for Astronomical, Atmospheric, Earth, and ocean economic potentialities. Sciences. P.L. 95-194, November 18, 1977, extended the Agriculture Act of 1977, provided price and income 1967 relating to the reimbursement of seized com- protection for farmers and assured consumers an mercial fishermen until October 1, 1978. It also abundance of food and fiber at reasonable prices. authorized the Secretary of Commerce in a new Title XIV contained declarations of needs in agri- section of the Act to make loans to the owners or cultural research and extension. Although the Act operators of U.S. vessels for damage or fishing is primarily designed for fresh water systems, it gear loss valued at $2,000 or more which the has possible application for coastal areas. In Secretary has determined was caused by a vessel section 1403, the Department of Agriculture is of a foreign nation operating within the fishery designated as the lead agency for food and agri- conservation zone established by the Fishery cultural sciences, and in section 1404, Congress Conservation and Management Act of 1976. It defined aquaculture as an agricultural science. directed the Secretary to investigate each inci- Section 1414 authorized appropriations. dent of loss for which a loan was made under this section. If tbe Secretary determines that the P.L. 95-136, October 18, 1977, authorized owner or operator who received the loan was not appropriations for fiscal year 1978 for the at fault, the Secretary may cancel repayment of Departments of Commerce and the Interior; and such loan and refund to such owner or operator for fiscal years 1978-83 for the Marine Mammal any principal and interest pryments made prior to Commission to carry out the Marine Mammal Pro- the date of such cancellation. If the owner or tection Act of 1972. In addition, it amended the operator was at fault, the loan shall be repaid Act of 1972 to make it unlawful for any person or within a reasonable time as determined by the vessel or other conveyance to take any species of Secretary. The law authorized the Secretary, with whale incident to commercial whaling in waters the assistance of the Attorney General and the subject to the jurisdiction of the United States. Secretary of State, to take appropriate action to collect for loss of funds under this Act. In addition, Congress declared: (1) the navi- gable waters of Puget Sound to be fragile; (2) P.L. 95-212, December 19, 1977, amended the that Puget Sound and the shore areas are threat- Endangered Species Act of 1973 (P.L. 93-205) and ened by domestic and international traffic of authorized appropriations of $10 million for tankers carrying crude oil; and (3) that it is section 6 of the Act for fiscal year 1977 and a necessary to restrict such tanker traffic in the sum not to exceed $16 million for the period Sound in order to protect the navigable waters, October 1, 1977, thru September 30, 1981. TheI the natural resources, and the shore area imme- authorization is to be shared by the Department diately adjacent thereto from environmental harm. of the Interior and the Department of Commerce. The law prohibits any officer, employee, or other The funds are to be used by the States for an official of the Federal Government from issuing, active program for the conservation of endangered renewing, granting, or approving any permit, or threatened species. license, or other authority for constructing, renovating, or modifying a terminal, dock, or P.L. 95-217, the Clean Water Act of 1977, other facility affecting the waters of Puget December 27, 1977, amended the Federal Water Sound, which may result in any increase in the Pollution Control Act of 1972. Move than 45 volume of crude oil capable of being handled by pages of amendments were made to the original any such facility (measured as of October 18, Act. Many amendments concerned authorization 1977,) other than oil to be used in the State of approval and authorization extensions. Some Washington. concerned State Jurisdiction (Water Rights), a Clearinghouse for Alternative Treatment Informa- P.L. 95-153, November 4, 1977, amended the tion, and Training Grants to name a few. Others Marine Protection, Research, and Sanctuaries Act concerned an accelerated program of estuarine of 1972 to authorize appropriations to carry out studies, oil spill liability out to 200 miles, the provisions of such Act for fiscal year 1978. marine sanitation devices, aquaculture, dredge The Act provided authorizations for Titles I, and fill permit hearings, and sludge disposal. II, and III. In addition, it required the It has a requirement for a study to be completed 244 by January 1, 1979, on the effects of seafood The results, findings, and information regard- processing which disposes of untreated natural ing ocean pollution research and development and wastes in marine waters. monitoring programs conducted by the Federal Government are to be disseminated to other Federal P.L. 95-219, December 28, 1977, approved the departments and agencies and to other interested fisheries agreement with Mexico. It also amended persons by the Administrator. section 2 of Reorganization Plan No. 4 of 1970 (relating to NOAA) to provide a General Counsel Appropriations not to exceed $5 million for and five Assistant Administrators, one of whom fiscal year 1979 are authorized. shall be for Coastal Zone Management and one of whom shall be for Fisheries. 3. Pending Legislation P.L. 95-257, April 7, 1978, amended title XI, The 95th Congress could have another four to section 1104, of the Merchant Marine Act of 1936 five months to act at the time this paper goes to to permit the guarantee of obligations for fi- press. As a result, it is impossible to accurately nancing fishing vessels in an amount not to exceed list all of the legislation passed and signed into 87-1/2 percentum of the actual or depreciated law for the 95th Congress. Consequently, it is actual cost of each vessel. worth while to list a few of the bills pending at this time and later at the MTS Meeting indicate P.L. 95-263, April 17, 1978, authorized the their progress as of that date. The following are President to issue a proclamation designating the some of the important bills listed by title, number week of April 16-22, 1978, as "National Oceans and description: Week." The President urged all Americans to become aware of the oceans and to learn about Deep Seabed Mining - H.R. 3350 and S. 2053, to their proper use. promote the orderly development of hard mineral resources in the deep seabed, pending adoption of P.L. 95-269, April 26, 1978, the Rivers and an international regime relating thereto. Harbors Improvements Act, amends the Acts of August 11, 1888, and March 2, 1919, pertaining Superfund - Oil Spill Liability - H.R. 6803, to projects for the improvement of harbors and S. 2083, and S. 2900, to provide a comprehensive rivers by contract. The 1978 Act authorizes system of liability and compensation for oil spill private contract for dredging if the Secretary of damage and removal costs. the Army determines that private industry has the capability to carry out the work at reasonable NOAA Organic Act - H.R. 9708 and S. 2224, to prices. The federally owned fleet for dredging establish a national ocean policy and set forth and related work may be reduced in an orderly the missions of NOAA. manner to a point where only emergency and national defense work can be accomplished "in Sea Grant Authorization - H.R. 10822 and house" if private industry demonstrates the H.R. 10623, to authorize appropriations to carry capability to carry out the required projects. out such programs for FY 79-80. A two-year study is to be conducted to deter- Marine Protection, Research, and Sanctuaries mine the necessary size of the federally owned Act - H.R. 10661 and S. 2767, to authorize appro- fleet. The cost of private contract for subject priations for FY 79-80. work may not exceed 25 percent of the cost of Government work on the same project. NACOA Authorization - H.R. 10823, to authorize appropriations for FY 79-80. P.L. 95-273, May 8, 1978, the National Ocean Pollution Research and Development and Monitoring Fishery Conservation and Management Act Act, established a program of ocean pollution re- Authorization - H.R. 10732, to authorize appro- search, development, and monitoring. It directed priations for FY 79-80. the Administrator of NOAA, in consultation with the Director of the Office of Science and Tech- Endangered Species Authorization - H.R. 10883 nology Policy and other Federal officials, to and S. 2899, to authorize appropriations to carry prepare a comprehensive 5-year Plan for the over- out the Endangered Species Act of 1973 during all Federal effort in ocean pollution research FY 79, 80, and 81. and development and monitoring. Marine Mammal Protection Act Authorization - The Administrator is further directed to H.R. 10730, extends appropriations for FY 79-81. establish a comprehensive coordinated and effec- tive ocean pollution research and development and Fishermen's Protective Act Authorization - monitoring program. H.R. 10878, would : (1) amend section 7(e) of the Fishermen's Protective Act of 1967 relating to the The Administrator is authorized to provide reimbursement of seized commercial fishermen; and financial assistance in the form of grants or (2) amend section 10 of the Act dealing with com- contracts for research and development and moni- pensation for vessel or gear loss due to foreign toring projects or activities which are needed to fishing activities. meet the priorities set forth in the Plan. 245 Aquaculture-S. 1043, S. 2218, and H.R. 9370, to provide for the development of aquaculture in 5. Summary and Conclusions the U.S. The 95th Congress, 1st and 2nd Sessions, pro- Polar Living Marine Resources Conservation Act duced 278 public laws from January 1, 1977, thru of 1978 (Krill bill), H.R. 12668, to provide funds May 21, 1978. More than 20 were marine or ocean for the construction of a research vessel for the oriented. Undoubtedly, additional laws affecting Arctic and Southern Oceans. the oceans will be enacted prior to the end of the 2nd Session. Tanker Safety - S. 682, to amend the Ports and Waterways Safety Act of 1972. Other than the appropriations Acts, there are several that deserve special consideration. For Liquid Natural Gas (LNG) Terminal Siting and example, the July 1977 National Advisory Committee Safety Act - S. 2273, would confer on the Secretary on Oceans and Atmosphere Act (P.L. 96-63) restruc- of Energy jurisdiction over construction permits tured and redirected the Federal Advisory Committee and operating licenses for LNG facilities. considering marine and atmospheric affairs. The members were not confirmed until January 1978, and Shoreline Erosion bills - H.R. 10015, H.R. 3775, it is likely that significant studies and/or recom- H.R. 10825, and S. 2514, to authorize the Secretary mendations of the Committee will not be issued of Commerce to make annual grants to assist States before the latter part of 1978. in carrying out public shoreline erosion manage- ment measures in the Great Lakes region. The Food and Agriculture Act of 1977 (P.L. 95- 113) is interesting. Section 1404 defined aqua- National Weather Service Act - H.R. 8763, to culture as an agricultural science. This appears, require the improvement of weather programs of at first glance, to be in conflict with Federal NOAA and affirm Federal responsibility for the freshwater and marine fisheries policies. However, provision of effective weather and relatedservices from analysis of the intent of Congress and inter- to assure, to the maximum extent possible, that all departmental arrangements, the authorities in available Federal resources are utilized in a co- aquaculture of the Department of the Interior and ordinated manner for weather related research, of the Department of Commerce have not been development, and technology. altered. National Climate Program Act - S. 1980, Public Law 95-153, amended the Marine Protec- S. 2092, and H.R. 6669, to establish a national tion, Research, and Sanctuaries Act of 1972 and climate program. provided more than authorizations. Importantly, it directed the Administrator of EPA to cease National Manned Undersea Science and Technology issuing or renewing permits for dumping of sewage Act - S. 2285, would establish a program for the sludge into ocean waters and defined the term enhancement of the U.S. capability in manned "sewage sludge." This Act could be the beginning undersea science and technology, of the end of ocean dumping of harmful sludge. Fishery Conservation and Management Act The Rivers and Harbors Improvements Act of 1978 Amendments - S. 3050, would authorizethe Secretary (P.L. 95-269) has a potential for directing addi- of Commerce to issue permits under certain condi- tional funds to qualified private contractors for tions to allow American-caught fish to be sold to carrying out projects for improvement of rivers foreign fish processing vessels if such catches and harbors (other than surveying, estimating, and exceed the American processing capability. gagings). In addition, it may reduce the Army Engineer's fleet but, at the same time, ensure its Navigation Development Act - H.R. 8302, would: modern and technical ability. The reduction may (1) impose a fuel tax on inland waterway barges; create a need for additional contract services in (2) authorize approximately $2.4 billion for the field and contracts for more modern equipment. harbor and waterway development; and (3) establish 4 an international commission for further study of Another Act with potential impact is the the feasibility of a sea-level canal in Panama. National Ocean Pollution Research and Development and Monitoring Act (P.L. 95-273). It formally 4. Prospective Legislation establishes a program of ocean pollution research, development, and monitoring. The Administrator As of May 20, 1978, there are otherlegislative of NOAA, in consultation with the Director of the topics that are likely to be formalized as bills Office of Science and Technology Policy in the and considered in the 95th Congress. An example Executive Office of the President, is designated is the Anadromous Fish Conservation Act authoriza- as the leading Administrator for pollution tion . There are obviously others that will be research. The Administrator is directed to seek introduced prior to the presentation of this paper and accept assistance provided by other Federal before MTS in September. The more important of agencies engaged in pollution research. Section 6 these will be described at that time. provides for grants and contracts for research and 246 development and monitoring projects needed to support the objectives of the Act. This appears to be a major move ahead in R&D. The major effect of these marine oriented laws will be seen in 1979 and ensuing years, first in the budget process, next in the work load, and hopefully, in environmental and natural resource protection. Therefore, the results of the 95th Congress probably will not be fully appreciated for several years. References 1. Anderson, Alfred W., Unpublished Manuscript, "Legislative and Other Authorities for Marine Affairs," January 1978. 2. Johnson, Frances, Unpublished Print, "Chronological List of Legislative Authorities" (NOAA 1977). 3. Public Laws 95-1, January 19, 1977, thru 95-284, May 21, 1978, Government Printing Office, Washington, D.C. 4. Zinn, Dr. Charles J., Revised and Updated by Willett, E. F., "How Our Laws Are Made," Document No. 93-377, U.S. Government Printing Office, Washington, D.C., 1974. 247 THE ECONOMICS AND TECHNOLOGY OF AN INTEGRATED OCEAN SURVEILLANCE SYSTEM FOR CIVIL ENFORCEMENT Norman L. Stone Ocean Systems Division Sanders Associates, Inc. 95 Canal Street Nashua, New Hampshire 03060 Abstract weaknesses - all weather ability, resolution, range, revisit time, endurance or time on station Ocean surface surveillance for civil en- to name a few. None can tell precisely what type forcement is more complex in some respects than of fish is in a hold from a distance. military surveillance because of the require- ment for very detailed information on what the ship is doing and has done and because of the legal requirement for hard evidence of any violation. The partnership between patrol air- craft and surface ships will probably continue to be the keystone but unmanned platforms such as anchored buoys can help. Sensors such as high resolution radar, IR/UV, LLTV and ESM com- bined with on-board data retrieval and storage and real time data communications can aid the task by helping to optimize resource alloca- tions. A command and control scheme using tight data links and interactive displays can pro- vide timely all-source information. Costs can be quite reasonable compared to the economic benefits. The man is still indispensable, and in the case of on-board observers on foreign vessels competitive with standoff methods. 1. Introduction Surveillance for enforcement in the new 200 mile economic zone and in the territorial and contiguous zones is in many ways more com- plex and difficult than military ocean surveil- lance. About 2 million square miles are in- cluded within the US 200 mile economic zone, over Figure I The Falcon Guardian in its US Coast half of which are off Alaska and Hawaii. Guard Surveillance Role Until some new breakthrough comes, the(PogrpFacnJt word "enforcement" when it comes to a surveil- Generally speaking, the high altitude plat- lance system must be understood in the sense forms have the advantages of wide area coverage that electronics can't enforce anything. The and synoptic viewing. Today's enforcement systems man on the scene - from a ship in most cases - include combinations of surface ship and aircraft is the one who finally does "enforcement." patrol together with inputs from military agencies. What maritime surveillance technology can do is Domestic vessels operating in the vicinity of locate potential violators with more or less foreign ships have a vested interest in enforcement accuracy, classify them by type and even by name and are also a source of surveillance information. and perhaps indicate something about what the suspect ship is doing or intends to do. The Automatic tracking from shore is possible by technical tools we have to work with include the means of automatic encoding and transponder re- surveillance platform, sensors, and interpre- transmission of LORAN location. The requirement tation and data storage devices and the command for detailed analysis of the cargo and even, in and control and communication links that inter- the case of fisheries enforcement, of fishing gear, connect them. records, and even fishing techniques goes far be- Eachof te cobinaion of latfrm, en-yond what standoff sensors can do from a distance sahor and commbications hasit sltfrenth sand in today's art. Retention of hard evidence, mostly 248 photographic is a necessary requirement for overfishing beyond the optimum harvest - an prosecution of violators. economic benefit to all countries. It also pro- vides priority to the fishing fleet of the sover- @4 4~ seign country. Military budgets for Ocean Surveillance /Ut -~ /~.~/ �'v~ / ~ / /usually exceed civil budgets and some quite com- plex ocean surveillance systems have evolved. AIRCRAFT 105 3.50 60 , 1/ODAT Many of the sensors, and particularly the data IMANN ER) handling, display and communications and command HELICOPTER 2H55X 10'- 255 1/R0IIDAY and control techniques developed for the military SHIP 3 OTIS02 1000 200 501 DAYS have important civil applications. This paper (DISPLAC EMENT HULL) describes a "integrated" system. RADAR SATELLITE 3 X 107 2800 5 X 103 4 HRS (MULTIPLE LOWN ORHT) 3. How the Particular Need Affects ANCHOREDBJUOY 3X 122 G0 104 CONSTANT FIXED the Surveillance System AREA 3.1 Fisheries Enforcement SHORE STATION 2 X 104 250 5 X 10O CONSTANT FIXED (OTH. SF HPF AREA LOR1AN TRANS. Fisheries enforcement is in many ways more complex and difficult than ASW because of the 'ONE COAST GUARD ESTIMATE FOR PRESENT COSTS AVERAGES25/NM2/YEAR. need to make detailed examination of the catch Figure 2 - Surveillance Platform Comparison including species and the date caught. No remote sensor or surveillance can provide complete in- The chief method of surveillance will doubt- formation, but combinations of sensors used less continue to be a partnership between patrol synergistically with a modern data link and com- aircraft and surface ships, with the bulk of the mand and control can make the job easier more thorough and more cost effective. stand-off large area search capability in the g aircraft. Equipped with high resolution radar, infra-red or low light level TV for night view- ing, microwave radiometry, photography and means of electronic intercept, and the dual advantages of speed and altitude, the wide area covered by an aircraft is the key to efficient enforcement. HERRING /INDOW Anchored buoys can provide early alerting in- / 7 NM formation for drug law enforcement. Satellites .. .. and shore stations have potential roles, but cost or lack of all-weather imaging with high COR .. 0.. resolution in the first case and short seaward .. .. range except for HF DF or OTH radar in the '.- second case leaves air search as today's primary wide area surveillance tool with radar as the key long range sensor. Typical airborne surface search radars have a 50 nm detection range on a 100 m2 trawler target in sea state 4-5. . . I 2. The Type of Enforcement Affects Surveillance Figure 3 -Potential Patrol Aircraft Route Covering Herring Window in Sub Area Several of the world's coastal nations have 5 - New England already established a 200 mile economic zones with primary emphasis on fisheries. The U.N. Much more than observation from a distance is Conferences on Law of the Sea will probably re- required to answer many fisheries enforcement re- sult in a world wide agreement on establishment lated questions. of such zones. Is there a valid permit? Is the catch proper The traditional interests of coastal na- in terms of species, quota and minimum size? Is tions also remain. These include controlling the authorized fishing gear being used and in- customs and drug traffic, ocean bottom mineral cluding trawl or seine type and net mesh size, and exploitation, search and rescue, weather and is it being employed as authorized (bottom or mid- ice patrol, and environmental or pollution con- water trawl)? Is there an excessive incidental trol of oil spills, hazardous chemicals and catch of other species aboard? Shelf bottom fish? wastewater and dumping. Has the vessel kept proper navigation and opera- The underlying reasons for civil enforce- tions records? ment are economic or protection of life. The Other questions fortunately can often be cost of surveillance and enforcement must bear answered to some degree by remote observation. some relationship to the gains the enforcing What is the identity and nationality of the country obtains by so doing. Fisheries en- corcementry fobinstanc, is dintendedstorpreven- ship? Is or was there a conflict with fixed gear? forcement, for instance, is intended to prevent 249 Modern technology offers much, but it is or more boardings per ship in 1-3 months or a clear that stand-off sensor cannot tell a red minimum of once per quarter. One Coast Guard hake from a silver hake or a 520 mm haddock from officer estimates that a single ship can conduct a 520 mm cod in the fish hold or reefer of a ship. 2 thorough boardings and inspections per day. Last year US and foreign fishermen took more than Given the number of US Coast Guard ships available 5 million metric tons in the coastal zone. for fisheries duty, (about 45 high and medium en- durance cutters plus 70 smaller ships), a limit of several hundred ships could be handled in a month if that was all they had to do which is not the case. Last year the US Coast Guard conducted about 1500 fisheries enforcement boardings. Long range surveillance can give clues as to activity and location, and what to look for so as to optimize the use of ship time. In the New England/mid Atlantic area in 1978 US Coast Guard expects that 1400 aircraft flight hours and 900 ship days would fulfill sighting and boarding standards in areas 1-5 and other potential violation zones. US and Canadian officials report that ships of some countries, parti- cularly the USSR, are usually very coQperatYve. 3.2 Customs and Smuggling The requirement here centers on the drug traffic. Quite large quantities of drugs such as Figure 4 - The Taras Shevchenko Being Boarded marijuana are imported to the US from South in March 1977 (UPI Photo) America and other areas by ship. Marijuana by nature is bulky, easing the search, but hard drugs such as cocaine or heroin are easily concealed and present a most difficult ship search problem. An estimated $5 billion worth of drugs at street value came into the US by ship in 1976. The patterns of operation of the boats, areas and time of opera- tion, routes followed and advance information from informants are important clues for surveillance. A history of offense by a given ship as is also the case in fisheries .enforcement, is an important en- forcement clue. For this reason, the ease of access to historical records that modern digital data storage and retrieval techniques can supply is important. Drug traffickers may rendezvous with smaller vessels to transfer cargo for the run into shore, providing another clue to activity by their pattern of operations. Radio intercept may be of value in some cases, but aircraft radar, direct visula observation, informants, and possibly anchored deep water buoys are the surveillance Figure 5 - The Taras Shevchenko was Stopped tools. The US enforces customs requirements only Again and Seized in April 1977 for to 12 miles, but information obtained hundreds of Fisheries Violations and Brought miles away in the passages between the Carribean to Boston (USCG) Islands or off the North Coast of South America and the Yucatan Channel may be vital to the Fishing fleets of a given nation usually close-in task. Cooperation by the trafficker can- follow traditional patterns of fishing as to time not be expected and they routinely monitor US Coast and location and type of fish. They are often Guard frequencies for advance warning of intercept. grouped fairly closely in the area where the fish are. They often exchange catch information The ships often simulate flaglessness and con- by radio, usually VHF. These characteristics at ceal or mislead as to their identity. Night sur- least tend to bound the search area. Typical veillance or boardings are virtually not conducted numbers of foreign fishing vessels in the US with the advantage going to the trafficker. Northwest Atlantic fisheries in 1977 were 50-90 Anchored buoys with recorders aboard and HF or VHF in groups of 10 to 30, often grouped by national- transmitters may be of value in certain areas to ity. Alaska reached a peak of nearly 700 foreign extend the time and area of surveillance at fishing vessels in June 1977. US Coast Guard moderate cost. guidelines requireboarding by a party that often The US emphasis on fisheries enforcement as includes a fisheries inspector at the start and compared to drug law enforcement is about 4 to 1 end of the permit period, and random boardings in resources and time in favor of fisheries. during the fishing period resulting in perhaps 2 Foreign fishing vessels receive the bulk of 250 attention but the emphasis on drug enforcement is increasing. In 1977 there were about 8200 fish- eries related flight hours compared to about 1700 general law enforcement flight hours and about 30 vessel seizures for all causes. 3.3 Other Surveillance Objectives Search and rescue, pollution control, mineral exploitation, weather and ice patrol and general law enforcement remain important civil surveillance objectives. All of their ends must be served by a surveil- lance system. SAR will often be a cooperative mat- ter using radio distress signals such as the US Coast Guard Distress Alerting Location System or emergency beacons. As with fisheries and customs, environmental standoff surveillance techniques must be supplemented with physical presence. For in- stance, an oil spill may be pinned to the offend- Figure 7 AIL AN/APS-128 Airborne Surveillance ing ship by photography of the slick, but if time has passed a physical sample must be obtained for fingerprinting. In 1977 there were about 1200 4.2 Electro-Optical and aerial oil pollution patrols and 800 ocean dumping Visual surveillance missions conducted by the US Coast Guard. Photo, visual and LLTV/IR/UF techniques in clear weather provide the best resolution of all 4. Sensors the sensors. Oil spill detection by multi- spectral sensors is possible and night detection and classification means are an essential part of a comprehensive surveillance system. Photo- graphs or stored IR/UV images together with super- imposed location data are often key evidence in fisheries violations and oil spill or dumping cases. A drawback in all of these cases is their RADAR Ei IRBAND 50-200 NO No No YES 10D20' reduced effectiveness in bad weather such as LLTV 0.4-.7p M I-5 YES YES YES NO 1-2' FLIR 3-2RpM I-5 YES YES YES NO 1-2' heavy fog or low cloud cover. Passive microwave ACOUSTICS 20HZ-20KRZ 5. YES PARTIAL SOME YES N/A radiometry may help supplement these sensors at EssVoF close ranges. MICROWAVE 2-12 GHZ 50-200 YES PARTIAL NO YES N/A HF, VHF, UDF COMMS 50-200D YES PARTIAL SOME YES N/A 4.3 Electronic Signal Intercept (ESM) 2-50OMHZ MICROWAVE RADIOMETE R 36 GHZ 1-2 - NO NO NO 10-20' PHOTO 0.2-30p M 1-20 YES YES YES N O 0.5" VISUAL 0.4-0.7p M 1-20 YES YES YES NO 0.5" TRANSPONDERS 0ERR I AND 50-200 YES YES No YES N/A Figure 6 - Typical Airborne Standoff Sensor Characteristics 4.1 Radar The primary wide area search sensor will probably continue to be aircraft radar. I-band search radars or SLAR's are usually a part of any airborne sensor suite. With a 100-200 Kw peak power, a typical radar will weigh about 300 pounds and have a range of 40-50 miles on a 10-50 m2 fishing boat and 100 miles or more on a 103 m2 freighter in sea state 4/5. Range is reduced as sea state increases due to background sea clutter. Figure 8 - Sanders' SR-200 Low Cost ESM System Maximum ranges in low sea states require high Covers 2-18 GHz flight which is inconsistent with detailed con- tact investigation which requires low attitudes. Using scanning or wideband receivers from As satellite and active OTH radar techniques patrol aircraft, direction, and sometimes iden- become economical, they may play a part. tity or intention can be determined depending on 251 intercept equipment and the sophistication of the foreign ship and where deliberate interference the target vessel. In some cases, it may be with enforcement carries a heavy penalty. possible to determine ship name from signal exter- nals alone much as an expert manual telegrapher Today's civil surveillance equipment is not could recognize frequent contacts by their fist. usually built with countermeasures in mind. Those The use of ESM depends on the fishing vessel using techniques employed by the military would have to its radar (if any) and ship to ship communications be used should active resistance to surveillance its radar (if any) and ship to ship conumunications. bcm rbe.Lwpoaiiyo necp Radio and radar silence can reduce the usefulness become a problem. Low probability of intercept of these too but it remains an important part of adar using coded pulse wide band transmissions and of these too but it remains an important part of EMaddrcinfnigwt ihsedsto total surveillance. ESM and direction finding with high speed set on and analysis capability are two approaches. These ~~~4.4 Acoustics ~techniques will add considerably to the cost of surveillance equipment. Air dropped free floating sonobuoys or per- 6. Description of an Integrated manently stationed deep water anchored buoys can 6. Description of an Integrated provide additional clues as to identity, location, System and possibly type of activity. Engine acoustics transmitted through the water and depth sounder The system uses the patrol aircraft with all- or fish finder emissions can be interpreted in weather and night imaging equipment as the basic some cases to tell the number and type of engines, sensor platform. Examples of suitable patrol air- possibly sizoe and speed, and country of origin, craft are the Lockheed P-3, CP-140 & 5-3, the A library of past contacts is necessary for best Fokker F-27 Maritime, the Falcon-Jet Guardian, and A library of past contacts is necessary for best teHwe-idlyNmo n 4 osgadr use of this method. In some areas, the sound of the Hawker-Siddeley Nirod and 748 Coastguarder. a trawl drogue against the bottom will give evi- Tightly knit and reliable communication links to a trawl drogue against the bottom will give evi-shrantohenfcigufcehpsreom dence of illegal methods or depth of catch, shore and to the enforcing surface ships are com- denceof ilegalmethds ordept of ctch.bined with automatic data storage, analysis and Anchored buoys appear applicable to drug traffic bied with automatic data storage, analysis and detection in distant straits and passages aid retrieval. possibly at egress from point of origin. 4.5 Transponders or IFF . A possible identity sorting method is to is- sue a low cost portable transponder with each fishing permit. At VHF or microwave upon air- craft interrogation this would be coded to pro- vide ship identity, permit number, and location on the aircraft CRT graphic display superimposed on the target skin return. It is a relatively inexpensive means of sorting. 5. Evasion and Countermeasures .. Active enforcement may lead to active op- position or inverse surveillance. Among the more obvious techniques are use of - -------- false or unclear identities, radio and radar silence, misleading communications, and deception Figure 9 - Enforcement System Functional Diagram as to intent. Radio and radar silence on the part of violators would eliminate radio intercept or The surveillance sensor suite on the aircraft ESM effectiveness. Installation of radar detec- consists of microwave surface search radar, 1FF tion equipment on the suspect ship will warn of interrogation, an electronic intercept and HF VHF the coming approach of a search aircraft at 200 UHF DF equipment from 2-500 MHz and ESM such as mile ranges and more - time enough to change the Sanders SR-200, Figure 8 covering 2-18 GHz, course, enter or leave an area, cease an illegal night vision devices, and photography with super- imposed date, location and intercept identity in- fishing procedure, or conceal or dump contraband imposed date, location and intercept identity in- overboard. One remedy is sparing use of the air- formation tied into the aircraft omega navigation borne search radar. Enforcement communications system. An inexpensive floppy disk storage system may likewise be intercepted and narcotics smug- working with an alphanumeric CRT and keyboard per- glers sometimes use the technique by monitoring mits the aircraft surveillance operator to retrieve Coast Guard frequencies. Very little can be done and store timely information on registration per- to minimize radar cross-section of larger steel mits, prior records, such as previous inspections ships, but the deliberate use of ships made of low or offenses. A modem connected to an HF or VHF reflection materials can reduce radar visibility. digital data link feeds the floppy disk and CRT This is more applicable to the drug trade. It is display so that the aircraft stored record can be unlikely that active countermeasures such as radar updated remotely from shore (to enter or modify or counications jamming would be eloyed, new permits in flight) or the aircraft operator or communications jamming would be employed, cnetrnwdt notesse nsoe n particularly in fisheries where continuity of the can enter new datainto the system on shore. One right to use the area is of great importance to current problem is maintaining timely registration 252right to use the area is of great importance to 252 permit data during extended patrols. Because the aircraft works closely with sur- face patrol ships, a two way digital data exchange over a data link to identical ship-board storage and retrieval equipment is part of the sys tem. Single sideband secure HF and VHF voice is also included. As might be expected, the major delay times in timely enforcement actions requiring command decisions are not primarily in the data link, but in the decision making process ashore. For in- stance, on the high seas where no offense has been committed as in drug or fisheries cases in- volving highly migratory species where there is no right to board, cooperation of the suspect ship's country of registry may be required and the decision chain may be quite lengthy before receiving a statement of "no Objection" throughFiue1-ThsypclSnrsdtinGahc the Coast Guard Commandant. Some fishing en- Fgr 0Ti yia adr dtn rpi forcement actions may require US Statement De- Display Used in the AN/FYK-14 System partment consultation. According to Gerry Studds, is Appropriate for Enforcement a Massachusetts Congressman, the National Security Command Centers Council became involved in a seizure last April, Tedt ol eatmtclysoe n e refuingperisson o size n ovios volaor. trieved in a disk file. An operator can auto- Anchored deep water buoys such as used in matically transmit edited or updated data to any synoptic oceanographic data collection have a desired area center or ship or aircraft at sea so potential role in customs enforcement by pro- that the corresponding display can appear on both viding early detection and location of ships the on-site aircraft or ship and the area center. using remote channels between islands. For in- By this means data time late, particularly on stance, logical installations of such buoys routine matters such as permit initiation or might be made in the Florida Straits and Windward modification will be minimized. Fortunately, most Passage between Cuba and Haiti, and in the Nova enforcement matters do not require the instant Passage, Yucatan Straits and waters around Puerto response times that some military situations do. Rico. Equipped with tape recorders the anchored Slower speed often permits economy of equipment. buoys can be read out upon interrogation from air- craf, shp orshor or ven elf nitited ponIt is estimated that a moderate size low cost the triggering of an acoustic threshold. Ab- cmue ftePP1 yewt 0mgbtds breviated data are transmitted over an HF link memory and editing graphic displays similar to those frequncy eleced fr curentRF popagtionshown in Figurel10 could amply handle all enforcement conditions. requirements for data storage and retrieval for an area such as the US East Coast. A smaller dupli- Two way data from local area command centers cate of the central setup would be installed at and the central command and control location would each of the local area or regional headquarters. be combined with military and other types of sail- ing reports. Voluntary reports, however, are not 7. The Dollar Value of Coastal particularly useful for exact location. Except Surveillance and Enforcement in selected high density harbor areas we have not begun to approach the tight positional control Enforcement only has a point if it has some attained by airways systems. economic or other value translatable into a benefit. At the central command and control information As one measure of value some percentage of the value on location and readiness status of surveillance of the products obtained from and carried over the forces, operations in progress, records of recent coastal areas could be used using the assumption pasthistrylink wih oter ivilageciesandthat enforcement prevents or reduces economic damage military will be maintained. Graphic CRT dis-insmreaontthttlcrgvlu.Sh playssuchas tose epiced i Figre 1 aremeasures may apply to minerals obtained (oil, etc) used to display wide area data or to zoom in on adfseispout.Ohrmaue r h a paticuar stuaion.For nstncecomptervalue of the customs tariff income on shipped pro- retrieved data on fishing permits on a given date ductsreasional unpollued (fisheries) and tevleo in a particular area might be spotlighted in rcetoa noepeevdadtevleo response~ to ageltm ur rmasrelac keeping shipboard undesirable imports (such as aircraft upon spotting a vessel displaying an ilgldus u fteUA unfamiliar permit number. Chemical Hazard In 1977 the Coast Guard seized 24 ships and information can also be accessed. about 900,000 pounds of marijuana with a retail 253 value of nearly $300 million. Experienced Coast Guard officers estimate that about 10 times that amount escapes them. The 1975 US fisheries catch of 2.2 million metric tons had an estimated landed value of $97l million dollars. Foreign ships in US waters landed more than the US catch. oil spill surveillance and fingerprinting as a minimum could lead to financial recovery of cleanup costs from the offenders and is a most effective deterrent against deliberate in- fractions based on some Coast Guard experiences. It would take a very detailed knowledge of the relationship between the intensity of enforce- ment and the reduction of violations and the optimum fisheries harvest to draw a specific con- clusion about the value of enforcement. The sheer dollar size of the fishing and drug traffic and the number of violators caught suggest that the economics will probably support enforcement at or above present levels. Lest we get too fond of technological solu- tions keep in mind that a shipboard observer pro- gram which could cost about $5 million megs for a 100 man program would put an observer on nearly every foreign fishing boat off New England. Compared to the initial cost of a single surveil- lance aircraft ($5 meg or more), to which operat- ing costs must be added, this does not sound like a bad return. 8. Summary and Conclusions Aircraft with an integrated surveillance suite and in a partnership with surface patrol ships will continue to be the primary enforce- ment means. Technology has not developed ap- proaches that give enough detail to eliminate the requirement for boarding and close-up inspection but it can provide valuable resource allocation information. Modern data processing and command links will permit even better resource alloca- tion. Enforcement surveillance costs are probably more than justified by the economic benefits ob- tained, but a more rigorous estimate of the re- lationship between enforcement, surveillance and economic return needs tc be made. Acknowledgements The kind help of CORE Streeper and Crowell of the US Coast Guard Headquarters in providing information on Enforcement Operations and Plans is gratefully acknowledged. 254 CULTURAL RESOURCE MANAGEMENT IN THE COASTAL ZONE Daniel Koski-Karell The Karell Institute P.O. Box 1616 Arlington, Virginia 22210 Abstract which mandate the protection of such This paper discusses cultural re- resources located on State-owned sub- sources in the coastal zone, touching merged lands. upon their definition and character. It also cites several legislative author- Suggestions on how to effectively ities relating to their protection and manage these resources as part of exist- management. Lastly, a possible means ing Coastal Zone Management programs to integrate cultural resource manage- will also be made. It appears that ment into the coastal zone program is integration into the planning process suggested. is the best recourse. This would pre- vent the possible delay of a project during its construction phase, for if it is discovered that construction 1. introduction threatens a significant cultural resource, work may be halted. Case studies will The cultural resources of the be used as examples to illustrate how coastal zone represent a significant cultural resources can be integrated aspect of our country's heritage and into the planning process, and, con- national patrimony. Because of their versely, what can and has resulted location, and the fact that they are a should they be ignored. non-renewable resource, it is apparent that a managerial concern under the 2. Cultural Resources Coastal Zone Management program is re- quired. However, cultural resources Perhaps a definition of the term on land and underwater normally do not "cultural reore is in order. Es- have a political constituency, and are sentially, it means an object or assem- often ignored even though their protec- blage of them (which may include people) tion is mandated under various Federal that can provide a society with informa- and State laws. tion relating to its origins and develop- ment. As human beings are both immersed This paper discusses several aspects in, and tireless producers of cultural of cultural resource management in the resources, one can understand that the coastal zone. One of these is an intro- total sum of these objects and assem- duction to its cultural history ranging blages is immense. However, cultural from Pleistocene times to the recent resources vary in significance, and it past. Particular emphasis is given to is this key concept--"significance"-- the maritime aspect--especially ship- that usually determines their disposi- wrecks from 1500 to the present. Atten- tion if a decision must be made. Of tion is also called to the studies seek- course, the significance of a cultural ing to identify sensitivity areas and resource is often a subjective decision, known locations of ship losses since the one that is sometimes tempered by econ- time of European contact. omic and political considerations. Significance is very often related to, A review of applicable Federal and and a result of, the input received various State laws will follow. The from interested people or groups. National Environmental Policy Act and the Archeological and Historic Preserva- The earliest cultural resources in tion Act will be discussed as they re- the United States for which there is late to cultural resources in the coastal firm evidence date from the Ice Ages of zone. Several states also have laws the Pleistocene. It appears that 255 big-game hunting peoples crossed the must be preserved or have any unavoid- land bridge linking northeastern Asia able impact mitigated through scientific with Alaska, and populated the New excavation or recording. The Coastal World. At that time the sea level was Zone Management Act reinforces this le- much lower than today. It is likely gal tradition by including cultural, that these peoples lived on lands now historic, and esthetic values. submerged, in addition to the uplands where evidence from their culture is Several States also have laws pro- normally found today. As the millenia tecting cultural resources. Some exam- passed, and the glaciers melted, these ples are Virginia, South Carolina, prehistoric Americans shifted their Florida, and California. In such cases, areas of habitation as the coastal zone the State usually claims title to mater- was displaced by the rising sea level. ial of archeological and historical The Eropen colnizaion o thevalue located on State-owned lands, The Eropen colnizaion o theincluding submerged lands. In such New World, which began in earnest after cases, unauthorized disturbance of such the year 1500, was marked by a concen- material is prohibited, Delaware is tration of activity along the coastal one example Of a State which even seeks zones of the hemisphere. Settlements to protect archeological resources on and farms were established there. privately-owned lands. Shipping soon became the preeeminent means of commercial transport and Over the past few years there has communications. In the course of time, been developing an effort to incorporate a tremendous number of vessels were lost the protection and enhancement of cul- due to storms, warfare, and navigational tural resources into development pro- errors. Study of their distribution jects, both on land and underwater. This reveals that the greatest number met has resulted both from the legislated their fate less than a mile (1.6 kin) requirements, and from unfortunate from shore. Their remains rest there Project delays resulting from inadequate to this day, with exact locations being measures to determine the presence or largely unknown. Many of the early absence of cultural resources before settlements and farms were ultimately construction. An example of this latter abandanoned, also. These cultural case involved a sewage treatment plant resources are now being rediscovered which was scheduled to be built in Vir- during archeological survey activities, ginia in 1975. Construction had almost and sometimes during the construction of begun when it was determined that the a coastal facility. facility would destroy the site of an Indian village dating from 1200 A.D. Currently the Bureau of Land Man- The significance of this site was such agement of the U.S. Department of the that the start of construction was de- Interior is conducting under contract layed and an emergency salvage excava- two intensive cultural resource studies tion conducted. If the project managers of the Atlantic Continental Shelf. The had not decided to sponsor this excava- goal is to develop a management tool to tion, the project would have been halted utilize in conjunction with OCS develop- during construction by court order, re- ment efforts. When complete, they will sulting in considerable delays and addi- both assist in managing submerged tional expense. cultural resources and in developmental planning to predict which areas are A more practical approach is being likely to contain those of significance. taken by the Army Corps of Engineers in their project to improve the shipping 3.Legislation channel from Baltimore to the entrance 3. ~~~~~~~to the Chesapeake Bay. Already a pre- There is a long history of Federal liminary cultural resources reconnais- legislation relating to cultural re- sance has been conducted to determine if sources. For example, the 1906 Antiqui- there is evidence to warrant an arche- ties Act provides protection for cultur- ological survey of the several areas to al resources over a certain age on Fed- be affected. If that survey discovers erally-owned lands. Also, the National significant cultural resources, there Environmental Policy Act specifically is adequate time to make minor modifica- includes important historic and cultural tions of the project design or to conduct aspects of our national heritage. The emergency archeological excavations to 1974 Archeological and Historical Pres- mitigate any adverse impact. ervation Act directs that in any Feder- ally-licensed activity or program or Federal construction project, archeolog- ical and historical data and resources 256 14. Planning 5. Conclusion To some managers involved in the The goal of cultural resource man- coastal zones cultural resources may be agement in the coastal zone is to both viewed as a low priority problem among protect and enhance the cultural patri- many others. However, a more efficient mony of our nation and to avoid ununeces- approach is to view these particular re- sary delays in development projects. sources as one of the irreplacable ele- The optimum approach for project managers ments of the coastal zone, one to be is to prevent problems from developing managed as well as possible within the by doing what is necessary, correctly, overall program. The optimum way to the first time around. deal with this particular aspect is to integrate it into the planning process. The ownership of cultural resources No reasonable person wishes to see a is diffused throughout our society. They needed project delayed on account of constitute a part of the cultural patri- cultural resources. However, such de- mony that we all share. For this reason lays have happened in the past, and their protection ha's been mandated by could very well occur in the future. It several Federal and State laws. This is much better to determine if any sig- protection must be weighed, however, nificant resources are present in a pro- against the needs of our present and ject area long before the construction future society, and decisions as to phase is scheduled to begin. disposition made accordingly. If threatened, significant cultural re- For coastal zone managers, the sources can still be removed and re- chief point of contact concerning cul- corded to preserve their inherent data. tural resources is the'State Historic By working in concert with cultural Preservation Officer in each state or resource managers, the managers of the territory. By coordinating with his of- coastal zone can both protect the heri- fice, as well as with the state arche- tage of our nation and facilitate ologist, the best recourse concerning a coastal zone development efforts. planned project can be determined. There are currently on-going survey and plan- ning activities within most state his- toric preservation programs. They can often provide prompt answers to ques- tions involving areas of the coastal zone. To date, most inventories have been oriented towards cultural resources on dry land. If a project is to affect submerged lands it is likely to be neces- sary to enter into a program of area- References specific background studies and survey. 1. Act for the Preservation of American Thera are three general phases in- Antiquities, 1906; (PL 59-209; volving a determination of whether sig- 3 4 ta.2516USC41) nificant cultural resources are located34Sa.2516U.C41) within a given area. The first of these 2. Archeological and Historical is background study and reconnaissance. Preservation Act, 1974; (PL 93-291; The second is intensive survey and eval- 88 Stat. 174; 16 U.S.C. 469,k). uation. Third, if necessary, is an emergency excavation and recording effort 3. Coastal Environments, Inc., Cultural to mitigate any adverse impact upon the Resources Evaluation of the N~orther-n resources in the project area. Evidence Gulf of Mexico Continental Shelf; from each phase is used to determine the Interagency Archeological Services, necessity for conducting the following Department of the Interior; one. If properly implemented early inWa h ntD.;197 the planning process, these studied willWahntD..197 provide guidance for any necessary design 4. Ditton, Robert B., John L. Seymour, modifications or management decisions. and Gerald C. Swanson, Coastal If not implemented, the unwelcome possi- Resources Manaaement. Be'vond bility of delays, litigation, and re-BuecryanthMrk; straiing oders ouldadverely afectLexington Books; Lexington, MA; the progress of a coastal zone project. 1977. 257 5. Iroquois Research Institute, Archeological and Historical Inves- tigations for Energy Facilities: A State of the Art ReDort; Office of Energy Systems, Federal Power Commission; Washington, D.C., 1977. 6. Koski-Karell, Daniel, "Underwater Archeological Surveys and the Proton Magnetometer," Quarterly Bulletin. Archeological Society of Virginia 30(4):192-193; June 1976. 7. Koski-Karell, Daniel, Cultural Resources Reconnaissance for the Baltimore Harbor and Channels Advanced Engineering and Design Study; Baltimore District, U.S. Army Corps of Engineers; Baltimore, MD; 1978. 8. National Environmental Policy Act, 1969; (PL 91-190; 91 Stat. 852; 42 U.S.C. 4321-4347). 258 THE IMPACT OF THE 1978 GENEVA SESSION OF UNCLOS III ON OCEANl DEVELOPMENT Lewis Alexander Marine Affairs Program Francis Cameron Washburn Hall Dennis Nixon University of Rhode Island Kingston, RI 02881 Abstraci Part XI were amended the developed countries would be unable to accept a new oceans treaty. The other After four and a half years of effort, the four negotiating groups addressed non-mining issues. third United Nations Law of the Sea Conference The subject matters of the seven NGs were as follows: has still not produced a draft treaty that could be open for signature. The Spring 1978 Geneva NG - 1 --System of exploration and exploitation session brought a new will to negotiate on critical of deep seabed minerals areas of the text, and made significant progress towards consensus. However, the result of the NG - 2 --Financial arrangements for deep seabed many compromises necessary to reach this agree- mining ment may tamper its acceptance by the world community. NG - 3 --Composition and voting of the Council of the Seabed Authority NG - 4 --Access of landlocked and geographically The Seventh Session of the Third U. N. Con- disadvantaged states to fisheries with- ference on the Law of the Sea (UNCLOS 1II) met in in the economic zones of other states Geneva, Switzerland from 28 March through 19 May, in the same region or subregion 1978 with 142 delegations in attendance. The first two weeks of the Session were taken up in NG 5 --Settlement of disputes relating to the deciding the designation of a President. Hamilton exercise of the sovereign rights of Shirley Amerasinghe of Sri Lanka had served as coastal states in the exclusive econ- Conference President since the beginning of UNCLOS omic zone III in 1973, as well as heading the Sri Lanka delegation to the U. N. But between the Sixth NG 6 --Definition of the outer limits of the and Seventh sessions of UNCLOS III there was a continental shelf, and the question change of government in Sri Lanka, and Ambassador of revenue sharing Amerasinghe was subsequently removed from that country's delegation. So the question arose, NG 7 --Delimitation of naritime boundaries could a person not accredited by his own country between adjacent and opposite states to the United Nations serve as President of a and the settlement of disputes thereon. major U. N. Conference? Ultimately, it was voted by the delegates to the seventh session that Mr. In addition, Committees II and III worked dur- Amerasinghe should continue in the Presidency, but ing the last five weeks of the session to try and by this point one-quarter of the time allotted identify other parts of the ICNT text in which for the seventh session had gone by. And with proposed revisions might receive substantial sup- the Presidency issue settled, a third week was port. In the work both of the Committees and of taken up in deciding on the organization of work the NGs, they were bound by Paragraph 10 of the for the session, leaving only five weeks for ser- document adopted on 13 April, 1978 concerning ious negotiations. organization of work. The paragraph reads in part The work of the session was ultimately en- "Any modifications or revisions to be made in the trusted to seven negotiating groups (NGs) on out- Informal Composite Negotiating Text should emerge standing hard-core issues. The first three of from the negotiations themselves. ... (and should not these concerned Part XI of the Informal Composite be adopted) unless presented to the Plenary and Negotiating Text (ICNT) on deep seabed mining. found, from the widespread and substantial support The United States and other industrialized prevailing in Plenary, to offer a substantially countries had objected strenuously to Part XI, as improved prospect of a consensus." it emerged after the sixth session in New York City during the summer of 1977. It was claimed 1. Suggested Revisions of the ICNT that the text leaned far too much toward the positions advocated by the Group of 77, repre- Negotiating Group I issued a report recommend-~ senting the developing states, and that unless ing changes or additions affecting eight articles 259 of Part XI of the ICNT, as well as two paragraphs er, the continental shelf issue may be part of a of Annex II. The changes are generally favorable general "package" involving also the rights of to the positions of the United States and other land-locked and geographically-disadvantaged industrialized countries. These texts do not rep- states, the failure to reach consensus within NG-6 resent a consensus, and were not discussed either is quite unfortunate. by Committee I or the Conference Plenary, but they There was also no agreement in Negotiating are thought to offer a substantially improved Group 7 on the wording of Articles 74 and 83, prospect of achieving a consensus at some subsequent concerning the delimitation of maritime boundaries session of UNCLOS III. One important feature of between opposite and/or adjacent states. Clearly, the proposed changes is that transfer of technology this is a major problem for a large number of would no longer be a condition of obtaining a con- countries, and its resolution will impact fisheries tract from the Seabed Authority. However, new and mineral production problems. A good example language in the text, requiring transfer of tech- is the existing disagreement between the United nology in certain cases, is strongly opposed by States and Canada with respect to lateral maritime U. S. ocean mining interests. boundaries. Tentative progress was, however, made Negotiating Group 2 made progress on the on the question of the extent, if any, of the third question of financial arrangements, a topic which party settlement procedures for boundary delimi- had not been negotiated previously. There was tation disputes. some redrafting of Articles 170 through 175 of the Most of the changes that occurred at the last ICNT, as well as of paragraph 7 of Annex II and of negotiating session concern procedural matters or paragraphs 9 and 10 or Annex III. The result was recommendations for change, rather than revisions that the new texts are considerably clearer than based on a consensus of the countries represented. was that of the ICNT on complex financial matters, Therefore, much of the substantive material in the although the revised texts offer little in the way text remains the same. The only major revisions of specific figures and percentages. Many consider have resulted from a reaction to external events detailed financial planning for the Seabed Authority such as the grounding of the Amoco Cadiz, illus- too speculative without more data regarding the trating that the major catalyst for agreement will anticipated profits from seabed mining activities, be similar "crisis" situations, or unilateral Like the report of NG-1, this report also was not responses to perceived "crisis" situations, such considered by Committee 1 or by the Plenary. as the establishment of exclusive fishery zones. Negotiating Group 3 debated strenuously the The Amoco Cadiz incident led the United States question of the composition of and voting procedures and France to reopen the draft treaty text on marine of the Council, but no consensus was reached on pollution even in the face of fears that this would these issues. eventually undermine the consensus already reached, Negotiating Group 4 issued a report amending and actually could weaken the text. However, the paragraph 2 of Article 62 of the ICNT, and sub- French specifically called for changes in the Text stantially modifying Articles 69 and 70. In the which would allow coastal states to intervene at case of the latter article it was suggested that the early stages in tanker casualties and to the term "geographically disadvantaged state" be strictly regulate vessels entering their coastal replaced by "states with special geographical waters. The United States generally stated the characteristics". The Article then goes on to need to clarify the provisions on a coastal state's spell out the criteria for being so designated. right to protect itself against pollution in the Although there appeared to be strong support for territorial sea, and to strengthen enforcement in the revised text of NG-4, during discussions in the economic zone. the Plenary, it became apparent that the issues The existing text established a combination of the land-locked/gds were becoming bound up with of port state, coastal state, and flag state juris- that of the outer limits of the continental shelf diction over marine pollution. The proposed re- and the procedures for revenue sharing with respect visions to the Text generally strengthen the to continental margin exploitation beyond 200 miles rights of coastal states to prevent vessel source from shore. pollution in their waters. There was broad agree- Negotiating Group 5 issued a text amending ment on a revision that would allow coastal states Article 296 of the ICNT, and added a new article, to employ special measures to protect rare or fragile as well as a general provision in the Convention ecosystems as well as habitats of depleted, threat- on abuse of rights. The basic concept that NG-6 ened, or endangered species of marine life. Con- agreed upon was compulsory conciliation with respect sensus was also reached on the promotion of IMCO's to various types of fisheries disputes within the recommended ship routing systems designed to economic zone, particularly on the allocation of minimize the threats of tanker casualties, and on surplus stocks to foreign nations. The text is the adoption of international rules and standards seen as a considerable improvement over dispute on prompt notification to coastal states of mari- settlement provisions in the ICNT, and should form time accidents. There was less agreement on the a basis for serious negotiations at subsequent more controversial issue of the coastal state's UNCLOS III sessions. right to adopt discharge standards that are stricter Negotiating Group 6 was unable to reach a than the international standards for vessels passing consensus on defining the outer limits of contin- through their territorial sea. ental shelf jurisdiction, due largely to a new Also in Committee III the United States and proposal put forward by the Soviet Union that there several other countries sought in vain to bring be a distance limit of 300 nautical miles on national about some amendments of the ICNT articles on rights to seabed resources. Since, as noted earli- marine scientific research, but the only "improve- 260 ments"obtained were small changes in the provisions "sufficiently improved over (last year's) Informal on scientific research with respect to deep seabed Composite Negotiating Text" so as to "substantially mining. From the viewpoint of the United States enhance the prspect of consensus." Yet MarneDubs, Scientific Community, substantial restrictions of the Kennecott ocean mining consortum, has still exist on oceanic research, stated emphatically that the new articles and the The fisheries provisions of the text remain revised old ones are "worse than they were before." basically unchanged, and most of the unilateral Dubs has charged that the Carter administration is fishery management declarations, such as the attempting "to establish a new adult partnership Fishery Conservation and Management Act, corres- with the third world by simply acceding to its ponds substantially with the UNCLOS III text. This demands. . ." neglects the fact that the Law of the Sea nego- What will be the cost of consensus for U.S. tiations were largely responsible for shaping ocean interests? The discussions which took place national legislation. However, while there is concerning the Preamble and Final Clauses indicate widespread agreement on the basic right of the the potential difficulties ahead. Many states coastal state to manage fisheries within the expressed the view that even after four years of economic zone, there is still no better under- negotiations, discussion of the clauses was still standing of the fishery management concepts to premature. However, others argued on behalf of an be used within that zone. The concept of optimum extended preamble incorporating the Conmon Heritage yeild is still being defined tithe United States' concept and endorsing the New International Economic Exclusive Fisheries Zone, as well as what the Order along with related "guiding principles." harvesting capacity of the coastal state is and Although a consensus of delegates may approve of thus of the surplus available for allocation to the New International Economic Order, the United foreign nations. These problems obviously are States Senate is unlikely to be as sympathetic. the same for similar provisions in the ICNT. Al- This reflects the disenchantment of United States though these concepts are still in the process ocean development interests with the deep seabed of definition, there was agreement at the Geneva mining provisions of the text, as well as the session on dispute settlement mechanisms for general frustration felt by other United States dealing with the problems of coastal state conser- ocean interests with the evolution of the UNCLOS IIL vation obligaions, as well as allocation of surplus One possible "safety valve" for the United fishery resources. States, if the treaty ever reaches the ratification An additional problem presented in the ICNT stage, is the use of reservations. However, there is the question of what fishing rights land-locked was only limited support for a liberal reservations and geographically disadvantaged states have in policy at the Geneva session. The Philippines and the exclusive economic zones of their coastal Syria supported the concept because in their view, neighbors. There is agreement on the fact that it would encourage nations to sign which might these "disadvantaged" states should have some type otherwise not do so. Peru, Chile, and Uruguay of preferential rights, but disagreement on the argued that reservations should be restricted to nature and extent of these rights. Little has "non-fundamental" provisions of the convention; changed in regard.to highly migraory ecies, this proposal creates the obvious problem of gain- sucn as tuna, or i n regarc to mnaaromousc ' PcIesuc h a or Hoen re ve, ting consensus on which articles are "fundamental" soecies, such salmon. However, the primary rignt o the state of origin in regulating or "non-fundamental". Japan spoke out most and harvesting salmon was strengthened in the strongly against reservations stating that it latest Geneva session. The trend set in estab- would destroy the "package deal" concept which lishing unilateral fisheries management scheneswill has served as the foundation for the UNCLOS continue whether or not a successful treaty negotiations. emerges from the Law of the Sea Conference. How- The necessity of permitting reservations in ever, the tuna and salmon fisheries, as well as a massive endeavor such as UNCLOS must be realized. other stocks that cross between different manage- The permissive use of reservations would encourage ment jurisdictions, indicate that bilateral and adherence to the treaty and bring us as close to multilateral agreements will be critical for universal adherence as could be reasonably expected. future fisheries management. An adverse effect, of course, is that reser- Did the Geneva session bring us any closer vations will confuse the legal relationships among to a Law of the Sea treaty? Negotiations, after the parties which ratify the treaty. The problem the initial procedural problems had been overcome, at this point, however, is salvaging something were intense. The difficult issue of how to manage worthwhile from all the effort that has been devoted the revisions of the ICNT was solved by the a- to UNCLOS III. Without the use of reservations, doption of the "collegial"system, which gave the legal relationships will become even more confusing President and the Chairmen of the main committees, some nations will be forced to renounce the entire acting together as a team, collective responsibility treaty over only a few issues. Would it not be for the revisions. Serious discussions on the more advisable to have as many nations as possible Preamble and Final Clauses were begun. The issue agree to the treaty rovisions which are compatible of reservations received serious consideration as with their national goals, and at least provide well. a common basis upon which individual negotiations The net result is that the Geneva negotiations can begin? did bring us closer to a treaty -- but in a form The impacts of the Geneva session of UNCLOS ITt that is unlikely to be favorably received by the are several: first, the delegates were able to United States Senate. Ambassador Richardson re- reach consensus on a number of articles in the ported that the deliberations produced a text negotiating text; second, a serious effort to 261 work out the final details was agreed to take place in New York from August 21 to September 8; third, the text produced has created enough irreconcilable differences that a policy on reservations must be approved before the treaty is open to signatures. A liberal reservations policy may be the only means by which the treaty can achieve some measure of the significance its drafters had hoped for. U. S. interests might be served by reservations in at least three cited areas: (1) transfer of technology in seabed mining; (2) revenue sharing in the proceeds of exploitation of the resources of the continental margin beyond 200 miles; and (3) marine scientific research within the extended econimic zone. Since there is a substantial possibility that the provisions of the treaty concerning these areas would not receive the advice and consent of the U. S. Senate, it is incumbent upon the U. S. negotiators to ensure that reservations to the treaty will be specifi- cally provided for in the final clauses. If the final draft of the treaty specifically forbids the use of reservations, U. S. interests, and those of many other states, might be better served through the development of customary law. UNCLOS has had a tremendous educational in- fluence in the world community. However, it remains highly unlikely that one comprehensive document could be signed and ratified by all members of the community of nations without reservations. In areas where consensus was truly reached, the revised ICNT will provid~e a valuable starting point in deliniating accepted bounds for state practice in the oceans. 262 FORCED MOTIONS OF A CABLE SUSPENDED FROM A FLOATING STRUCTURE Raymond L. Bisplinghoff David 0. Libby Tyco Laboratories, Inc. Simplex Wire and Cable Company Tyco Park P.C. Box 479 Exeter, New Hampshire 03833 Portsmouth, New Hampshire 03801 Figure 1 Cable Support System Abstract Recent and proposed applications for power cables in the ocean require that the cable be suspended from floating structures which move due to the wave action. Tension changes in the Floating cable due to these motions may be a substantial Structure percentage of the static tension under some con- Support Buoy ditions. Limiting values of the tension are necessary for proper design of cable strength members to withstand simultaneous flexing and tensioning. A method is described for calculating these tensions. In-plane and out-of-plane motions of the cable are analyzed separately. The basic equations of motion of the cable are CableCal developed and lowest-mode solutions are derived. 1. Introduction Power cables for OTEC plants and otherI recent applications in the ocean environment1 require that long lengths of cable be suspended from floating structures to the ocean floor. Anhr Lines for Buoy This is an unusual operating condition for a power cable because it is simultaneously subject to a relatively high tension, a variety of pos- sible motions, and substantial electrical stresses. In order to properly design a power cable for this environment, it is necessary to estimate the loads on the cable and the amount of bending to which it might be subjected. In order to suitably limit tension and con- Four kinds of loading were identified which trol bending, while accommodating a maximum of might have significant effects on tension and/or movement of the structure, an intermediate sup- bending: port system was chosen. A schematic of it is shown in figure 1. A buoy supports the cable 1) static tension, 2) current drag, 3) at a distance from the structure somewhat strumming effects, and 4) forced motions. greater than the water depth. This distance and the length of cable are adjusted to avoid either For cable design purposes, relatively simple cal- over-tensioning the cable or depositing part of culations of the static configuration, and the the cable on the ocean bottom during movement-of effects of current drag seemed adequate. Esti- the structure. The buoy is placed deep enough mates of the relative effects of strumming were to avoid most wave action, and its anchoring made by judicious application of techniques such system is designed to limit its movements to as those described in Ref. (2). Although these negligible amounts. The attachment of the cable techniques are not yet developed for catenary to the structure is assumed to be at approxi- cables with substantial sag, some approximations mately the same depth as the buoy. using the basic methods led to an estimate of 263 maximum strumming amplitude and resulting drag Referring to the coordinate system exhibited increase. Existing work on forced motion of in Fig. 2, small vibratory displacement of the slack cables did not seem to go far enough or cable during forced motion may be expressed in lend itself to routine use for estimation of terms of the normal modes of vibration, as follows: cable motion and increased tension due to that motion. It was therefore necessary to develop an approximation technique for estimating maxi- mum values of the cable tension and movement resulting from forced motions of one end. c/ t) (2) 2. Defining the Problem i) ( (3) A cable suspended from a floating structure is forced into motion when the structure moves in response to waves. For structures large Where are normal coordinates and enough to require a large power cable, it can be are the x, y and z components, re- initially assumed that the cable has a negli- spectively, of thei-th normalized eigenfunction or gible effect upon the motion of the structure. mode shape of the cable in water. The latter sat- As a first approximation, the motions of the isfy the orthagonality property structure will be assumed to cause simple har- (4) monic movements of the end of the cable. Fre- quency and amplitude will be specified to be where consistent with the motions of the structure in s= length of cable a given sea condition. mass per unit length of cable including the virtual mass of water Only the lowest modes of motion of the cable will be considered unless the excitation The normal coordinates ~;t) are computed by frequency is substantially different from the solving the differential equation frequency of the lowest mode. This simplified ,-.4o /-b'I analysis allows calculation of cable loads and M. W M 14 (5) motions accurately enough to establish bounds J-Ij - on the loads for cable design. where ct = frequency of j-th natural frequency 3. Formulation in Terms /ql5 =) *hd- gener- of Normal Coordinates alized mass of j-th natural mode and where r'-14 J are generalized forces in the j-th natural mode. The latter are formulated as prescribed forces L derived from external disturbances designated by superscript D and forces arising from the motion of the cable designated by superscript M. F D, Fyand F5 are force components per unit length a T of cable due to prescribed disturbances and FM, w h FM and F5M are force components per unit length y resulting from the motion. In the following sections of the paper, it is assumed that the in-plane and out-of-plane v motions of the cable are uncoupled and can be treated separately. Figure 2 Cable in Static Configuration 264 4. Out of Plane Forced Motion and where AO = density of sea water If it is assumed that the in-plane and D = cable diameter out-of-plane motions of the cable are uncoupled, V = kinematic viscosity of sea water we may express the out-of-plane motion by: The differential equation defining the dis- (F .Yt) = . � �6y) ~(%) (6) placement of the first mode of out-of-plane 'I'~"~~~~ ~motion becomes: Where i,(4) are defined by Eq (5) in which: Whr Qc)ae + ,n = g A S (14) 4j e J ago lads (7) where 2- - 2.25 7,--- j - /- $ /1 P~~~~ ~(8) In Eq (14), the normal coordinate g, represents red Jphysically the angular motion of the swinging cable in its first mode. The soluction of ---.q M Eq (14) may be written as A4 =f- A js ) (15) The natural frequencies c- and corresponding Where the maximum angular excursion of the cable mode shapes P; for the out-of-plane motions in the "jump-rope" mode is are defined by the homogeneous integral c ("q_ I equation .- , w 7 - w.)+( )�/' w=v Cts )tA Q*(iC , (10) The additional dynamic cable tension induced by the swinging motion is computed from the maxi- Where C2(S,6%) is an influence function. In mum inertial load per unit length, practice, the eigenfunctions and eigenvalues are obtained from a lumped parameter approxi- P = M ' (16) mation of Eq (10) expressed as $ =z C (11 Eq (16) may be substituted into the catenary Adc~z C j j't (11) equations to compute the added cable tension. Or in matrix form $I m =Lc t i(12) Recently, this method was applied1 to a LU)t~ 29 ~~~~cable approximately 600 feet (180 m) long, 6 Where is an n x n matrix of influence inches (150 mm) in diameter, and weighing 16 coeffecients and is a diagonal matrix of pounds (7.3 kg) per foot in water. The analysis lumped masses. showed that static tension in the cable was approximately 10,000 pounds (4500 kg). When the In the present analysis, an engineering end was forced into harmonic motion at a period approximation is obtained by taking the lowest of about 14 seconds and an amplitude of 20 feet, mode of out-of-plane motion. This is a (6 m) the additional tension in the cable was pendulous motion in which the cable swings less than 1000 pounds. At shorter periods the laterally in the manner of a "jump rope". tension was less. The cable end defined by the coordinates (L, 0, O) is attached to the floating structure 5. In-Plane Forced Motion and is assumed to be displaced harmonically in the z direction according to The displacements u and v of in-place forced motion, assumed to be uncoupled from w-(Qo,%) = A S,,, t lateral motion, are defined by Where A is the frequency and A is the amplitude _, (17) - (", Y, t.) =) .(..1.7, j) + of motion of the floating structure. f~~zt = g ysi )~t (18) The terms of Eq (5) may be reduced with Where H,'+) are defined as before by Eq (5) in appropriate simplifications to which (46 t , A s(19) i,_. i I- -- - ' I The natural frequencies, c; and the corresponding Where Cf is a viscous damping coefficient mode shapes defined by $ j v are computed defined by from the coupled integral equations Cc _- L-'z~- z'-./'�f~ OJ~ D = z <sA)4, 0t)0C +(sII) i (s) ' (dt)d 't (22) 265 ) c'sj/) +c Jt9 Ot 265 Which may be reduced through finite difference The same cable described in the previous approximations to the lumped parameter form section can be characterized for in-plane motions I!'T rc'l[ i{{ by a mass matrix: 3 [E~~~~~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C 4rIc b43 3 Li' The influence coefficient matrix in Eq (23) may L o 3s be calculated numerically by techniques such as Where the units are slugs. The lumped masses, those described in Ref. (3). which include the entrained mass of water, are equally spaced 195 feet apart. When the support In the present analysis, an engineering points of the cable are 500 feet apart in the approximation is obtained by taking only the static condition, the frequency of the first mode lowest mode of in-plane motion. This mode is of in-plane motion is computed by representing the cable by a three 0.29 rad./sec. mass system with mass matrix IM, 0 H and the normalized mode shape is H'] '~~~~~~~~~~ / ~ ~ ~ ~ ~ ~ ~ l~.00 0.505~ ~~~~~ C ~~~~~~~~~~~0.601 /0~r~ .l1 As in the out-of-plane analysis, the cable 0.801I end, defined by the coordinates (L, 0, 0), is attached to the floating structure and is The additional tensions induced by a motion assumed to be displaced harmonically in the X of the end of the cable at a period of 14 seconds direction according to and an amplitude of 20 feet (6 m) are less than 1500 pounds (680 kg). Analysis of behavior at shorter periods has been limited but tensions Where o( is the frequency and B the amplitude appear to be lower. of motion of the floating structure. The appropriate terms in Eq (5) are com- 6. Recommended Additional Effort puted from puted from2 Two particular assumptions used in simpli- At �($L," t )1 +! ~fying the above analysis are subject to review and further development. The present model -L</~ 7 K /'lL' n4 assumes a linear drag term at the outset of the analysis. An initial assumption of a drag term proportional to velocity squared, followed by a s'.t rdr//a~\ - 1I4J,~OI i 1'> 4 linearization at a later stage of mathematical -C1 f2 L~YJ; s9~AaT/;J/JL.J' '71 development might provide a more accurate picture of cable motions, but the present analysis Where AL; is a segment of cable associated with probably produces conservative results. The the i-th mass for purposes of computing the assumption that one end of the cable remains force due to damping and and and are fixed may not be an accurate one for some appli- computed at the location of the i-th mass. cations. Significantly more work is necessary to accurately analyze motions of a second, movable The appropriate differential equation de- end. fining the first mode of in-plane motion is 4, 2.- , .o. [,-,] a;-i.t (27) .f7cC:�, 0;;It) ~ 7. Conclusions Where c- ... � The analysis of forced-end motion of a 4c. 7M ,- suspended cable provides an estimate of motions and tensions in the cable. For many cables, the The solution of Eq (27) is expressible as accuracy of this analysis will be sufficient for &;.I* co, -/ (28) design purposes because substantial safety fac- tors in strength will be utilized. In addition, the limits of motion are often of more interest M ~r /L~4 F~v, ~/~ ~rthan the details of that motion. MANt WY 6 - ~ thisLimited experience in the application of this method indicates that a multiplier of 1.5 The additional cable loads induced by the to 2 sheuld be applied to the static tension to first in-plane mode of motion are computed from allow for these dynamic effects before a safety the inertial loads on each mass by means of the factor is applied to cable strength. This figure formulas, may not apply to cables of much different sizes, = 3 c O however. 266 No attempt has yet been made to analyze the phasing of the two motions. An in-plane addition of the two tensions was sufficient for previous work. Incorporation of the two motions into one analytical model may be necessary if detailed analysis of the com- bination of the two motions is desired. References 1. Pieroni, C. A., B. S. Clark, and D. 0. Libby, "A High Voltage Power Cable Installation Concept between Offshore Structures," Proceedings of the 10th Offshore Technology Conference, Paper Number 3325, May 1978. 2. Skop, R. D., 0. M. Griffin, and S. E. Ramberg, "Seacon II Strumming Predictions," Naval Research Laboratory Memorandum Report 3383, October 1976. 3. Dominguez, R. F., and C. E. Smith, "Dynamic Analysis of Cable Systems," Journal of the Structural Division, American Society of Civil Engineers, Vol. ST8, August 1972. 267 EFFECTS OF LONG TERM TENSION ON KEVLAR ROPES: SOME PRELIMINARY RESULTS Thomas P. Bourgault NAVAL UNDERWATER SYSTEMS CENTER NEWPORT, RI 02840 Abstract Another discrepancy in the available data was the lack of actual environmental testing, and before This paper is a summary of the preliminary the NUSC and WHOI0 data became available, there was results obtained thus far in the NUSC Kevlar Test essentially no at-sea test results. Therefore, in Program. The purpose of the program is to invest- order to undertake a program which was both real- igate the documented strength reduction of a istic of actual usage and at the same time con- number of Kevlar ropes when subjected to a variety trolled for variable isolation, NUSC consulted of loads in the ocean environment. with DuPont and the rope manufacturers to lay out the entire test plan.3 From these meetings, the Long term tests, both in-air and in-sea water following six categories were identified as being were initiated with periodic residual testing to the most important considerations in an effort to establish the effects of load, time and environ- determine the effects of long term tension on ment on over 400 rope samples. Kevlar ropes. Final and more conclusive results will be These considerations include: available upon completion of the 12 month test period in November 1978. 1) fiber to be used to build the samples, 2) rope constructions to be tested, 3) duration of the test, Introduction 4) static load levels, 5) environment, and At the Oceans 76 Conference, the Naval 6) terminations to be used. Underwater Systems Center presented data showing extensive strength loss of lKevlar after four months All LKevlar rope samples acquired for the test in an ocean mooring system.' At about this same program use Kevlar 29 as the basic building compo- time, the Woods Hole Oceanographic Institution nent. This fiber was specially fabricated by produced data supporting these findings which DuPont for the test program and fully characterized together posed serious questions as to the by them prior to shipment to the manufacturers. lifetime and reliability of using Kevlar in the This not only assured the quality of the fiber to ocean environment. 2 be used in the test program but also established a basis for comparison should the results indicate a Prior to these reports, all documented testing basic fiber problem. done by DuPont and others on the basic Keviar fiber showed no significant strength loss when Both the NUSC and WHOI0 data was comparable subjected to various load levels both in air and although the rope constructions (and manufacturers) in water. Therefore in an effort to resolve this tested by each were different. Therefore, an dilemma, the Naval Underwater Systems Center effort was made to include in the sample group as acquired funding to initiate a comprehensive test many different constructions as was economically program to investigate this phenomena. feasible. The actual sample groups and rope con- structions selected are discussed later in this This report describes the test program con- report. ducted and some preliminary results that indicate no "significant" loss in strength of Kevlar ropes The duration of the test program is one year after both the three and six month test periods. with the initial control testing being followed by residual testing at three, six and twelve months. In this way, the time history effect of load on the Test Program various samples could be determined. After researching all existing information The static load levels selected were 0, 1200, along with the results of various test programs, 2100 and 3000 lbf. All samples were specified at it was found that the majority of documented data 6000 lbf so that these load levels would correspond was based on the testing of individual Kevlar to 0, 20, 35 and 50% of the rated breaking strength. fibers and not from finished rope samples. In this way, the effect of increasing loads would be 268 U.S. Government work not protected by U.S. copyright. determined along with any aging problem when con- , EXPOSURE MONTHS TIME ~~~~~~~TOTAL NO. sidering the zero load samples. LOAD T 3 6 12 OF SAMPLES HISTORY- (10/77) (1/78) (4/78) (10/78) The environment itself was an important con- sideration since most of the fiber testing had 0 LOAD -AIR 5 3 3 3 14 been done in air and the NUSC and WHOI data SEA WATER - 3 3 3 9 followed extensive exposure at sea. Therefore, 1200Ibf to determine any environmental effects, analogous -AIR 3 3 3 9 in-air and in-sea water testing was done. -SEA WATER - 3 3 3 9 2100 Ibf Finally, since general usage ropes are -AIR - 3 3 3 9 supplied with manufacturers terminations, it was -SEA WATER - 3 3 3 9 decided to acquire all samples pre-terminated 3000 Ibf effective ~-AIR - 3 3 3 9 utilizing each manufacturers most effective - 3 3 3 9 method. In this way, the cutting and reterm- SEA W A T R _____ inating of long samples which could possibly 77 affect the data during residual testing was t3 SPARES eliminated. 80 Test Samples Figure 1. Test Program and Sample Breakdown The five sample groups included in the test program are defined below: Test Set-Up A. U.S. Steel: 3 x 19 construction, torque balanced oceanographic wire rope with a poly- Subjecting 400 samples to four separate loads ethelene jacket and terminated at both ends with in two different environments was accomplished by swaged sockets. attaching the various samples of each group in series for each particular load and exposure time. B. Philadelphia Resins Corp.: 1500 denier braided construction with a polyester jacket and In air, the sample lengths were suspended terminated at both ends with epoxy filled wire horizontally (anchored at one end) and tensioned rope sockets. by a snatch block arrangement with the pre- scribed weight. C. Wall Rope Works: 1500 denier uniline (parallel) construction with a polyester jacket For the in-water test, steel buoys were and spliced terminations at both ends with heavy fabricated (with the buoyancy of the floats being thimbles. equal to the required tensions) and used in sub- surface mooring systems in Narragansett Bay. The big D. Cortland Line Co.: 15000 denier counter/ advantage of implementing the test in this way is helical construction with a polyester jacket and that, at the conclusion of each exposure period spliced terminations at both ends with thimbles. complete systems could be recovered and tested without interrupting any other test apparatus. E. Cortland Line Co.: 15000 denier braided construction with a polyester jacket and spliced terminations at both ends with thimbles. Test Results The U.S. Steel wire rope samples were in- The environmental testing was initiated in cluded in the test program as an additional con- October 1977. At the time of this writing, both trol and also as an up-to-date comparison of steel the three month and six month test periods have versus Kevlar in similar applications. been completed with only the twelve month data left to finish the test program. The total number of test samples, 400 (80 per sample group) is significant since this translates All the sample testing throughout the program into 800 individual terminations. Therefore, due took place at the U.S. Coast Guard R&D Center at to the large number of terminations, the test pro- Avery Point in Groton, CT. The testing was done gram becomes an investigation of not only the by the author and Mr. Carl Jones of DuPont so that effects of load etc. on Kevlar but also the any discrepancies or unexpected phenomena could be reliability of terminations and the ability of the analyzed when they occurred. manufacturers to apply them. The breakdown of the 400 samples and the entire test program is All samples from the various manufacturers given in figure 1. were specified at 6000 lbf breaking strength. However the delivered samples ranged from 5000 to 10,500 lbf. Because of this, the expected loading 269 percentages of 20, 35 and 50% were variable TEST STATIC I AVERAGE I CHANGE IN depending upon the sample group. MENVIR ONMENT PERIOD L S It (MONTHS) STfb!) I TbH The tensile testing procedure for all the 0 0 6276 - samples is identical. Each one is exercised from 0 6293 +0.3 0 to approximately 1500 lbf ten times prior to des- 6 0 6053 -3.6 truction. The cross head speed is held constant 3 1200 6347 +1.1 at 12 in/min. Finally, it should be noted that a AIR 6 1200 5650 -10.0 5% change in strength is attributed to scatter 3 2100 5800 -7.6 which is based on the evaluated standard deviations 6 2100 5867 -6.5 of all the samples. 3 3000 - The results obtained thus far in the test pro- 6 3000 6047 -3.6 gram are presented in figures 2 thru 6. They are 3 0 5733 -87 given individually by sample group to avoid con- 6 0 5593 -10.9 fusion when comparing results. SEA 3 1200 6060 -3.4 WATER 6 1200 5367 -14.5 All of the 3000 lbf in-water moorings failed early in the test due to premature sample termin- 3 2100 5500 -12.4 ation failures. Because of this, and the lack of 6 2100 * spare samples, this portion of the program has been deleted. * data not yet available Figure 3. Philadelphia Resins Braid Results The steel wire rope, as expected, has lost The Philadelphia Resins 1500 denier braided none of its strength at this point in the program. Kevlar shows a reduction in strength in virtually The lack of in-water data is a result of corrosion all of the tests. However, the tendency of the which developed due to a galvanic couple between in-air tests is for the reduction to decrease with the galvanized terminations and the large bare increasing load. A possible explanation is that steel buoys. These samples are being reterminated elongation, alignment and load sharing of the and the data will be published when available. fibers is accomplished more efficiently at the higher loads. This argument does not hold true however, for the in-water samples and the absolute numbers are significantly higher which possibly TEST STIC AVERAGE CHANGE IN indicates that sea water does contribute to overall ENVIRONMENT PERIOD LOAD STRENGTH STRENGTH strength reduction. (MO NTH) (I b P (I b)(0/o)strength reduction. (MONTHS) (Ibf) (bf) (/O) O 0 6926 - Initially in the test there was some prob- 3 0 7170 +3.5 lens with the Philadelphia Resins potted termin- 6 0 7253 +47 ations. After some investigation, it was found that the solvent used to clean the Kevlar had not 3 1200 7347 +6.1 AIR 6 12�00 7293 +453 been given sufficient time to evaporate thereby causing inefficient bonding between the fiber and 3 2100 6868 -0.8 epoxy. (Note: Philadelphia Resins had, for safety 6 2100 7613 +9.9 purposes changed solvents from MEK to less volatile ~~~~~3 3000 *~Xylene. However, because of this problem, they 6 3000 7387 +6.7 have returned to using MEK exclusively.) 3 0 7413 +7.0 6 0 7107 +2.6 SEA 3 1200 * The Wall Rope uniline construction showed much scatter in the data and at this point in the 3 2100 * program, no definite trends can be established. 6 2100 * : The data from the final test period will hopefully yield more insight into this construction. * data not yet available One very important characteristic was noted Figure 2. U.S. Steel Wire Rope Results however, with the uniline in the 3000 lbf in-air (3 month) set-up. Shortly after this test was initiated, a Philadelphia Resins sample failed (due to the termination/bonding problem noted above) causing all the samples in this string to go slack. The sample was replaced and the weight reapplied but within two days a uniline sample failed midspan. Again the sample was replaced and the test continued. However, this phenomena 270 occurred four more times and it is now believed TEST STATIC AVERA CHANGE IN that the uniline construction is susceptible to ENVIRONMENT PERIOD LOA STENH STRENGTH (MONTHS) ~?bf (Ibf) Ioo shock damage caused by the instantaneous release of tension. This finding is important when 3 0 7358 3 0 7680 +4.4 considering a mooring system under load that is 6 0 7710 +4.8 recovered by the actuation of an acoustic release which in effect would subject the rope to the same 3 1200 7360 0.0 AIR 6 1200 7385 +0.3 shock release damage as that observed in the test set-up. If this is the case, then reusing the 3 2100 6760 -8.1 rope could be catastrophic even after extremely 6 2100 7260 -1.3 short exposure periods. 3 3000 * 6 3000 7380 +0.3 3 0 6987 -5.0 a ~ ~ ~ ~ ~~~~~~~~~~~~~~~6 0 6547 -I 1.0 ENVIRONTEST STATIC AVERAGE I CHANGE IN 6 0 6547 -11.0 ENVIRONMENT PERIOD LOAD STRENGTH ISTRN TH SEA 3 1200 6680 -9.2 (MONTHS) I(Ib) (bfl Ibf) , WATER 6 1200 7380 +0.3 0 0 4946 -- 3 0 4946 - - 3 2100 6708 -8.8 3 � 5193 +5.0 6 2100 6 0 5340 +8.0 3 1200 4733 -4.3* data not yet available AIR 6 1200 4993 +1.0 Figure 5. Cortland Counter/Helix Results 3 2100 4693 -5.1 6 2100 4560 -7.8 3 3000 6 3000 440 -12 ~6 3000 4400 -10.2 ~ The Cortland 15000 denier braid tested per- 3 0 5140 +3.9 formed extremely well in the program with no sys- 6 0 5400 +9.2 ten failures caused by premature sample or term- SEA 3 1200 5173 +4.6 ination problems. However, very little data, as WATER 6 1200 4780 -3.4 shown in figure 6, could be used to evaluate the 3 2100 4625 -6.5 overall performance of this rope. The reason for 6 2100 * - this was that almost all samples failed "within" the termination during residual testing. Therefore, * data not yet available this data did not reflect the actual rope strength and was not accepted as being valid. Figure 4. Wall Rope Uniline Results Cortland Line supplied two different construc- TEST STATIC AVERAGE CHANGE IN tions to be evaluated in the test program. The ENVIRONMENT PERIOD LOAD STRENGTH STRENGTH 15000 denier counter helix construction has showed (MONTHS) (IbfI (Ibf) ) a similar tendency as the Philadelphia Resins data, 0 0 10666 - that being in this case, a decrease in the strength 3 0 - reduction with time under load. Also similar is the greater strength loss observed in sea water 3 200 for this construction. AIR 6 1200 - 3 2100 * - As noted previously, the Wall Rope uniline 6 2100 10380 -2.7 showed no indication of sea water being detri- mental. This data reaffirms the fact that sea 3 3000 10190 -4-5 water does not affect the strength of the Kevlar fiber.4' 5 However, this does not exclude the 3 0 10240 -4.0 possibility that in some constructions, sea water 6 0 9660 -9.4 does have a detrimental effect due to increased SEA 3 1200 fiber abrasion etc. when wet. WATER 6 1200 *- - 3 2100 - 6 2100 - * data not yet available ** termination failures Figure 6. Cortland Braid Results 271 Conclusions References The preliminary results obtained thus far in 1. Bourgault, T. P., "Design and Performance of the test program have provided no definitive trends a Two-Stage Mooring for Near Surface Measure- attributable to all Kevlar ropes. The environ- ments," Proceedings of the Marine Technology mental testing in sea water has been conclusive Society Oceans 76 Conference, Marine Technology in that it appears that this medium has little, if Society, Washington, DC, 1976. any, affect on the basic Kevlar fiber. The effects of load and exposure time have also varied through- 2. Walden, R. G., "Tests of Kevlar as a New out the sample groups which indicates that a major Mooring Material," Polymode News, No. 15, variable affecting residual strength is rope 17 September 1976. construction. Based on the data generated thus far, this does not appear to be a major factor, 3. Bourgault, T. P., "The Kevlar Cable Test however, the final results from the 12 month test Program," NUSC TM No. 77-2004, NUSC, Newport, period should provide additional insight into the RI, 18 January 1977. extent of this problem. 4. Swenson, R. C., and Ferer, K. M., "Aramid Fiber The test program has reaffirmed the need for for Use as Oceanographic Strength Members," additional termination research and has also NRL Report 8040, NRL, Washington, DC, uncovered a load release problem with the parallel 5 November 1976. construction. 5. Ferer, K. M., "Effects of Long Term Tension on Kevlar Aramid Fiber," NORDA Technical Note 8, Naval Ocean Research and Development Activity, National Space Technology Laboratories Station, MS, May 1977. 272 PERFORMANCE/FAILURE ANALYSIS OF ACOUSTIC ARRAY CONNECTORS AND CABLES AFTER 6-10 YEARS OF SERVICE Colin J. Sandwith Applied Physics Laboratory University of Washington 1013 N.E. 40th Seattle, WA 98105 Abstract One of the arrays, Array 8, was a buoyant type such as that shown in Figure 1; the other two, A performance/failure analysis was conducted Arrays 9 and 6, were rigid types. The major dif- on three acoustic arrays recovered from a depth of ference between these arrays is the foundation of 3300 ft at a naval tracking range off St. Croix in the structure (rigid vs buoyant). The major dif- the Virgin Islands. The analysis included bench ferences between the electrical-acoustical systems tests of electronic components and examination of (hereafter referred to as the harnesses) are the connectors, cables, seals, and housings with elec- diameter of the cable seals and the materials used tron microscopes, X-ray diffraction, and optical in the cable jackets. Array 8 had been in service microscopes to determine the effectiveness of for 10 years and Arrays 9 and 6 for 6-1/4 years. A design features and materials. One array had been complete performance/failure analysis was conducted submerged for about 10 years, the others for on Array 8 whereas a less rigorous analysis was 6 years. conducted on Arrays 9 and 6 because they had been in service for a shorter period. All the arrays Cables with polyethylene jackets performed had been installed at 3300-ft depth on the ocean satisfactorily whereas cables with polyurethane bottom a few miles off St. Croix in the Virgin jackets exhibited leaks and visible circumferential Islands. The water at this site has an average cracks approximately 1/4 in. long. Of 42 double oxygen content of 4 ml/l, a salinity of 35.00 %, O-rings examined, 8 of the first O-rings leaked; and a temperature of 5.6�C.3 half of these were on one array. The second O-ring on all seals was effective. Arcing, apparently due A performance/failure analysis uses all back- to lightning, produced a low resistance in one of ground information on design, fabrication, testing, the Morrison connector seals by creating a carbon implantment, operation, recovery, and shipping to path of charred teflon. Certain packing seal explain the results observed and determine the glands and backup washers in the Morrison seals of cause or causes of failures. Since a number of one array sustained permanent deformation because contractors and agencies were involved to varying of improper assembly. degrees in each one of these tasks, the background information was somewhat fragmented. For example, Introduction the performance of the systems just before recovery is uncertain, and there are conflicting reports on The Applied Physics Laboratory of the Univer- the coating systems, if any, used on the housings. sity of Washington (APL-UW) conducted a performance/ failure analysis on three acoustic tracking arrays General Procedure for the Naval Facilities Engineering Command (NAVFAC). This analysis was part of an effort to NAVFAC and subcontractors recovered the arrays document the effects of long-term deep submergence in the fall of 1976, detached the harnesses from on seals, electrical cables, and housings.1 The the superstructures, and packed and shipped them to results will be used to increase array life by APL. The packed harnesses were stored in the improving design. In addition, reduced data will shipping crates in a dry, cool place for approxi- be incorporated into an APL reference manual on mately 6 mo. awaiting approval for the analysis. seals and connectors.2 seals and connectors.2 On approval, the components of all three arrays were unpacked and identified. Identification was A primary objective of the analyses was to facilitated by the color-coded vinyl tape that had locate all failures--including those in the elec- been used to designate the position of the hydro- tronic systems--and determine the cause of each phones during assembly and implantment. The com- failure. This paper will be limited to tests on ponents were then tagged to maintain identity and the cables, seals, and housings and their results. their as-received condition was documented with The extensive tests conducted on the electrical- photographs and notes. acoustical system will be the subject of another paper. 273 The following is a synoptic description of the Results examination program carried out on Array 8, which was subjected to a complete performance/failure Array 8 analysis. Array 8 was planted in December 1966 and First, the corrosion and other damage on recovered in September 1976. The specific reasons housings, cable clamps, fasteners, transducer caps, for recovery, other than general malfunction, are and strain-relief components was noted visually and not known by the author. However, a tracking range recorded. Observations included the type of cor- status report dated 9 December 1973 shows the rosion damage (uniform, pitting, crevice or leach- following entry: "Y hydrophone bad (?) on ing) and a description of its extent (deposit Array 8." thickness, depth of penetration, shape of the deposit, distribution of products, color of depos- The packing of the Array 8 harness was the it, and structure of deposit). The corrosion best of the three harnesses received and probably mechanism, such as galvanic coupling, crevices, produced the least shipping damage. Array 8 was concentration cells, fretting, coating failure, and the only array that had styrofoam in front of its selective leaching, was also identified if mean- hydrophones to protect them from shipping damage. ingful and possible. Little attempt was made to The most notable packing damage was to the inter- evaluate the corrosion damage in terms of reduced connection cables and cable clamps, and was caused strength or increased friction, or to identify the by the low bending radius of the cables. The corrosion product. extent of the damage could not be quantified. Interconnecting cables between transducers and The design and the material specifications for junction chambers were disconnected by pushing the Array 8 were extracted from engineering files. Morrison-type cable seals2 out of the junction Material specifications are summarized in Table I. chamber and separating the cable from the hydro- No documentation was found on painting or other phone housing at the double O-ring joint at the protective coatings. Morrison seal housing. Hydrostatic tests, resis- tivity tests, visual examinations, and dimensional checks were carried out on the cable and seal Table I. Array 8 Material Specifications assemblies to find leaks and determine the effect of pressure on resistivity in the cable jacket and Morrison cable seals. If defects were found, Description Material Plating further tests with a scanning electron microscope Housing, Seal Steel AISI Zinc, 0.0005-in. were conducted to determine the cause. The re- (Preamp) C1010-1020 thick,* QQ-Z-325 sistance of each cable and seal assembly was sys- Class 2, Type 2 tematically measured with a megohmmeter (General Radio Co., Type No. 1862-B) both in air and in Housing, Junction Steel AISI Same as above Chamber C101S water after approximately 1 day at 1500 psig to test for changes caused by time and exposure. Housing, Preamp Steel AISI Same as above Where megohmmeter readings were not possible be- C1Ol1-1020 cause of low resistance paths due to voltage break- down, the resistances were measured with an Caps, End, Junction Steel AISI Same as above ohmmeter. Chamber CIOIO-1020 Packing Gland Commercial Neoprene N/A* After the Morrison seals were pushed out, the Non-Conductive, seal housings were examined to determine the depth 40 � 5 duro of water penetration and the type and extent of corrosion damage to the housing surface. In cer- Washer Backup Commercial Neoprene N/A tain uses, the housings were also sectioned for the Non-Conductive, tam uses, the housinos were also SO~~~~~8 � S duro examination. Insulator Laminated Nylon, N/A Each cable seal that exhibited low resistivity Fabric Base, was then systematically disassembled to locate the Phenolic Resin low rpisstance paths and determine their cause. staine and/o cut) Washer, Thrust Plastic Sheet, Glass N/A The size, condition (distorted, stained and/or cut)Epoxy and resistivity of the seal glands and backup Resin washers were measured. Evidence of leaks such as corrosion products and water penetration was re- Cable, Coaxial Polyurethane Jacket N/A corded and the cause was determined if possible. Interconnecting 0.25 in. 0.D. Cable, Coaxial Polyethylene Jacket N/A Each hydrophone assembly was disassembled 1.0 ine 0.t. further to perform an analysis on the housings, Signal seals and electronic-acoustical systems. The conditions of the boot and the O-ring seal were documented. *No additional coatings specified. 274 External Condition electrical tape. Electrical tape was used in numerous places around the housings, cables, and Junction Chamber: The corrosion and calcar- structures and performed well. eous deposits were thinner (1/16 in. to 3/16 in.) on the signal cable end plate of the junction Heavy encrustations (<1/4 in. thick) of a chamber than on the interconnecting cable end plate combination of calcareous deposits and hydrated (3/4 in.) of the junction chamber. At the O-ring forms of iron-oxide compounds were attached to the points on the junction chamber, the deposit was strain-relief housing. Penetration of corrosion approximately 1/4 in. thick. Light scraping with a into the low-carbon steel was estimated to be screwdriver exposed bare metal on the side of the 1/16 in. housing, where a 3/8-in. thick deposit was found. The scraped surface had wide shallow pits no deeper Cables: The condition of the interconnecting than 1/16 in. cables appeared to be good on first visual exami- nation. Subsequent electrical and hydrostatic Hydrophone Assembly: Corrosion and calcareous tests, which are discussed later, revealed several deposits on the hydrophone housings were thinner defects. and more tenacious than those on the junction box (1/16 in. thick or less). The hydrophone housings Internal Condition were made of a slightly different alloy than the housing on the junction chamber, and may have been System Operation: Performance tests were made painted. on each hydrophone assembly. During these tests, the power to the hydrophone and the hydrophone The stainless-steel cap screws used on both output signal were routed through the main signal the junction chamber and the hydrophone housings cable. By starting with the main signal cable and performed perfectly. Upon drying, the corrosion systematically making electronic and acoustic and calcareous products on the sides and the bottom measurements through to the acoustic transducer, of the hydrophone housings had become powdery and defects and malfunctions could be identified and easy to remove compared to the crusted products on located, and the performance of working components the junction chamber. However, the corrosion and could be evaluated. Selected results from these calcareous deposits on the metal top of the hydro- tests are given in Table II and described in the phone housing were much harder and tenacious like following paragraphs. those on the junction chamber. The main signal cable could not be used to Transducer Boots and Retaining Wires: The power the hydrophones because of a discontinuity in transducers' rubber boots and retaining wires the conductive shield. Examination indicated that performed exceptionally well in this environment. the shield had parted because of thinning induced They were practically free of marine growth and by plastic straining. The shield was apparently corrosion products, except for slime and micro- continuous (at least intermittently) at the time of fouling. recovery because that conductive shield was used to fire an explosive bolt. Fasteners: The stainless-steel, cap-screw type fasteners used on Array 8 were more resistant The junction chamber was opened at the main to corrosion damage than the fasteners used on The junction chamber was opened at the main either of the other two arrays. The low-carbon signal cable end in order to power the hydrophones steel structures in Array 8 apparently were sacri- t hrough the Ju nction chamber electronics. By fically corroded to protect the stainless-steel chamber leads, it was determined that hydrophones fasteners, which were used in all cases except on chamber leads, it was determined that hydrophones the cable clamps. The fasteners used on the cable C X and X-Y were operational and that hydrophones clamps were made of carbon steel and had deterio- was dead agrees with the 1973 tracking that hydrophone Y rated beyond use. was dead agrees with the 1973 tracking range report. Main Cable Strain-Relief and Hose Clamps: Main Cable Str ain-Relief had re a nd Hose Clamps: The problem in the Y hydrophone was traced to The strain relief had reached the end of its use-Y preampli- ful life. The metal and nonmetal components used fier was opened like a clam shell by milling to attach the cable to the housing had failed lengthwise on both sides so that the electronics because of corrosion. The tightening bolts in could be examined. Figure 2 shows the sectioned the two stainless-steel hose clamps, used to com- housing, the potted preamplifier, and the hydro- press a sleeve around the cable, failed completely. The This type of failure would be expected of stain- staininge assembly less the cable assembly. The less-steel hose clamps exposed to such extended staining in the center of the housing is due to marine service. The bolt is often made of 400 welding and the staining on the ends is probably marine service. The bolt is often made of 400 series stainless steel, the threaded portion of. due to corrosion produced by sealed-in volatile the clamp constitutes a crevice, and the clamp strap is usually made of 300 series stainless Close examination indicated that the cable for steel. Therefore, the bolt and threaded portion th e Z hydrophone was stretched at the p oint where are anodic, and are sacrificed by galvanic corro- the hydrophone housing. Tests on the sion to protect the large, cathodic clamp strap. cable indicated that the stretching had parted both the center conductor and the shield. The Z hydro- The stainless-steel hose clamps behaved most phone was found to be operational when power was satisfactorily when they were covered with vinyl applied beyond the damaged cable section. 275 Table II. Summary of Performance/FaiZure Analysis on Array 8 Hydrophone C X Y Z X-Y Tilt Ex. Bolt Morrison Cable Seals Corrosion Depth Corydrosion Depthnd) 0.80 0.80 0.80 0.80 0.80 1.15 1.40 (Hydrophone End) Cause Deforma Deform Deform Deform Deform Deform Cable Corrosion Depth 0.80 0.97 0.79 0.76 0.91 1.07 0.86 (Junction Chamber End) Cause Deform Deform Deform Deform Deform Deform Deform Cable, Interconnect (RG 58) Condition --- --- --- Cut --- Cut Cut Damage 2 leaks --- --- Separation 2 leaks --- --- Cause Aging --- --- Stretch Cracks --- --- (Leaching) Seals, O-Ring Hydro Transducer OK OK OK OK OK OK NO Hydro Cable OK OK OK OK OK 1st Leak Cuts Junction Chamber, Main Cable End (lst OK, 2nd backup cut) Junction Chamber, Hydro End (Ist OK, 2nd OK) Transducer Impedance OK OK OK OK OK --- --- Corrosion ND ND ND ND ND None --- Boot OK OK OK OK OK --- --- alst seal glands deformed by incorrectly installed cable clamp sleeve and excess of metal shims bleak in polyurethane jacket O-Ring Seals: The 0-ring seals in both junc- plate, were sectioned to illustrate the distribu- tion chamber end plates were well lubricated and tion and nature of corrosion products. The sec- not leaking even though a defect existed. The tioned junction chamber end plate shown in Figure 4 0-ring backup gland of the second (inboard) 0-ring graphically illustrates the depth of seawater on the signal cable end of the junction chamber had penetration. been cut (sheared) as the end cap was pushed into the cylinder. (See Figure 3.) The cutoff section The sectioned Morrison seal housing with the (about 3/4 in. long and 1/16 in. deep) of the double 0-ring joint to the tilt mechanism housing 0-ring backup had then been pushed into the first is shown in Figure 5. The distribution and loca- (outboard) 0-ring groove. This defect did not tion of corrosion products and knowledge about affect the successful operation of the first Morrison seal design lead to the conclusion that a 0-ring. leak in the cable jacket allowed seawater to sur- round the seaward backup washer, but that the seal The double 0-ring piston seal on the bottom glands on either side blocked further leaking. plate of the tilt sensor housing worked perfectly. The surfaces on the piston part of the seal were Electrical and mechanical tests were conducted very clean and excess lubricant was evident. How- to detect leaks in the cables or seal assemblies ever, the cable seal housing assembly had to be and to measure the seal resistivity. The procedure pressed out of the bottom plate because the first was to disconnect each interconnecting cable, with 0-ring had leaked. The decrease in the inside its seal assemblies intact, from the array harness. diameter of the housing due to corrosion products The section tested included the Morrison seal from caused interference between the piston and the the junction chamber end, the interconnecting housing wall. Dense, tenacious and hard corro- cable, and the Morrison seal with its housing from sion products had collected from the seawater side the hydrophone assembly. A typical interconnecting all the way up to the second 0-ring. The cause of Morrison cable seal assembly is shown in Figure 6. leakage could be a defective 0-ring or lack of suf- The test setup is shown in Figure 7. ficient squeeze on the 0-ring. Insulation breakdown megohmmeter measurements Cables and Cable Seals: The depth that sea were made between all connector pin and housing water and corrosion products had penetrated the combinations, first in air and then in water at 0, Morrison cable seals was measured on the seal 500, 1000, and 1500 psi; pressure was maintained at housings. The results are shown in Table II. The 1500 psi for approximately 24 hours and the meas- Morrison seal housings on the junction chamber, urements repeated, both at 1500 psi and after which were made by boring holes through the end reducing the pressure back to zero. The results are given in Table III. 276 Table III. Megohmreter and Resistivity Values: Typical or Representative Values for Array 8 Interconnecting Cables and Cable Seal Assemblies Hydrophone Explosive Bolt C X Y Z X-Y Tilt J.C: End Bolt End Air Hydro (end) c-Sa 17b 24 ND --- 9 ---- 0.4 C-H 17 24 ND --- 9 --- --- 0.45 S-H 120k/0 c 200k/0 ND --- 80k/O --- --- 180k/ Junction Chamber (end) C-S 17 24 40 72 9 37 60 --- C-H 13 24 ND 58 9 37 60 --- S-H lOOk/0 80k/0 ND 200k/0 50k/0 40k/0 90k/O --- Water P=O, t=od C-S 17 28 40 82 18 40 S0 0.60 C-H 17 28 40 130 18 40 50 0.60 S-H 120k/0 2b00k/0 200k/0 1 18k/0 100k/0 90k/0 180k/0 P=1500, t=0 C-S 33 35 50 140 22 ND ND 0.45 C-H 33 35 45 170 22 ND ND 0.50 S-H 2k/0 220k/0 200k/0 1 50k/0 ND ND 140k/O ac = center conductor; S = shield, H = housing. bIf a megohmmeter reading could not be made because of breakdown at the 500 Vdc megohmmeter setting, the voltage was reduced until a satisfactory reading could be obtained. When it was necessary to use a voltage other than 500, the megohmmeter or resistance reading and the voltage are indicated as follows: reading/voltage. CAll readings are 1000 M unless otherwise indicated. The letter k indicates 1000 S. dp = pressure; t = time (hours). Leaks in the interconnecting cables were often All of the interconnecting cable sleeves on detected by measuring a low resistance between the Array 8 were installed backward. The typical shield and the water in the test chamber shown in interconnecting cable clamp, cable sleeve, metal Figure 7. The location of the leak was determined shims, and Morrison seal assembly in Figure 6 by one or both of two methods: (1) injecting air show the distortion and gap produced by the im- into the cable jacket and observing bubbles emit- properly installed cable sleeves. ting from the leak, or (2) withdrawing the cable slowly from the water and detecting a change in the All of the seaward backup washers showed measured resistance between the shield and the substantial reductions (by factors of 104 to 106) water. Several leaks were located by these in voltage breakdown resistance, except for the methods. seal at the sensor end of the explosive bolt cable. The low resistance of the backup washers was prob- Figure 8 shows an example of typical cracks ably caused by adsorption of corrosion products through the urethane cable jacket. Figures 9a and the absorption of seawater admitted, in part, and 9b are a series of photomicrographs taken with by the improperly installed cable sleeves. The a scanning electron microscope to determine the absorption was symptomatized by swelling. The material microstructure and deduce the probable diameter of the seaward backup washers increased by failure mechanism. Figure 9b shows extensive 1.2% and the thickness by as much as 32%. The microcracking. The cable jacket material was badly diameter of the other backup washers tested in- deteriorated and embrittled probably because of creased by 1.2% and the thickness by about 11%. leaching of binder or resin. The seal glands in the Morrison seals showed Cable Seal Packing Glands and Backup Washers: a dimensional change due to creep and a substantial Each Morrison seal assembly was completely dis- reduction in hardness. The diameter of the seals assembled, and the size of each individual packing decreased by 1.4% and the thickness increased by gland and backup washer was measured to determine 2%. The measured hardness was consistently 15% the amount of dimensional change (creep) that had below the mean hardness specified in the design taken place under service conditions. The hard- drawings. ness, voltage breakdown resistance, and appearance of each item were also determined. The results are shown in Table IV and discussed briefly here. 277 Table IV. Cable Seal Packing Gland and Backup Washer Properties Cable Seal Components -- (Inboard) (Seawater) --Design Backup Backup Backup Backup Washer Seal Washer Seal Washer Seal Washer Seal Hydrophone End: Diameter (in.) 0.761 + 0.005 0.781 + 0.005 Avga 0.768 0.769 0.769 0.770 0.769 __- c Maxa 0.772 0.770 0.776 0.772 0.772 -_- c Mina 0.768 0.767 0.767 0.768 0.763 __~ c Thickness (in.) 0.094 + 0.005 0.250 + 0.005 Avg 0.104 0.225 0.103 0.257 0.123 ---c Max 0.107 0.258 0.108 0.260 0.125 ---c Min 0.098 0.250 0.099 0.255 0.120 ___c Hardness 80 + 5 40 + 5 Avg (duro) 68 38 68 38 64 39 Max 70 40 70 40 70 40 Min 65 35 66 35 60 37 Voltage Breakdown Resistanceb None None Avg 432 170 29 34 10-3 172 Max 800 250 40 800 7 d 200 Min 240 120 12 180 10-5 5 Junction Chamber End: Diameter 0.761 + 0.005 0.781 + 0.005 Avg 0.769 0.770 0.770 0.770 0.771 0.769 Max 0.768 0.772 0.772 0.772 0.775 c Min 0.770 0.769 0.768 0.769 0.765 0.769 Thickness 0.094 � 0.005 0.250 + 0.005 Avg 0.104 0.255 0.104 0.255 0.125 _ _c Max 0.110 0.257 0.109 0.260 0.130 ___c Min 0.100 0.252 0.100 0.252 0.120 _ _c Hardness 80 + 5 40 + 5 Avg 70 40 69 40 67 41 Max 72 43 70 42 70 42 Min 65 39 65 40 65 39 Voltage Breakdown Resistanceb None None Avg 380 174 260 258 1 x 10l3 500 Max 800 240 700 300 17 x 10-3 900 Min 160 140 52 170 6 x 10-3 220 Average for all (7) identical cable seal components bMegohmmeter readings taken across seal or washer with clamp contacts touching both faces and separated by approximately one-third the seal diameter. Megohmmeter voltage was 500 Vdc unless otherwise indicated. All readings are 1000 megohms. Deformed @ 50 Vdc Array 9 tracking range status report indicated that hydro- phones C and Z were bad. The material specifica- The analysis performed on the harnesses from tions for Array 9 are summarized in Table V. Arrays 9 and 6 was less extensive than that per- formed on Array 8. No hydrostatic tests were The crating procedures and the shipping damage performed on the cable and seal assemblies and only received by Array 6 and Array 9 appeared to be defective seals were completely disassembled and equal. Because neither had styrofoam in front of analyzed. the acoustic transducer, some of the observed transducer damage may have been produced during Array 9 was a rigid design planted in June shipping. In Array 9, two transducers (C and Z) 1970 and recovered in September 1976. The array were exposed because of cuts on the boots. It is was recovered because of malfunction, but the unlikely that the cuts in the boots were solely the specific defects are not known by the author. The 278 Table V. Material Specifications for Arrays 9 and 6 Description Material Plating Housing Seal (Preamp) Steel, AISI, Zinc,* Electrodeposited C1010-1020 Cold QQ-Z-325, Class 2, Type II Finish Housing, Junction Carbon Steel, Same as above Chamber AISI MT 1015 Housing, Preamp, Steel AISI, Same as above Cap, End C1010-1020 Junction Chamber Steel AISI, Same as above C1010-1020 Cap, End (Seal Junction Steel, Cold Finish Same as above Chamber) AISI, C1018 or leaded C1018 Pin, Connector Brazz or Bronze Nickel Plate QQ-N-290,* Class 2, Type V Packing, Gland Butyl, Chloriniated N/A* 30 S5 duro Washer, Backup Teflon N/A Insulator Teflon N/A Washer, Thrust Plastic Sheet N/A NEMA G 10 Cable, Coaxial Polyethylene N/A Jacket 0.3 in O.D. Cable, Coaxial Signal Polyethylene N/A Jacket 1.0 in O.D. 'No additional coatings specified results of shipping. However, at least one trans- (gravity effects) of the seal during service, the ducer of Array 9 appeared to have been jammed into seal-to-housing diametral interference, and the the side of the crate. None of the hydrophones on electrical potential (grounding and stray current) any of the arrays had transducer cap protection. of the housings. ExternaZ Condition The results of the resistivity tests indicate that the cable seal assemblies in Array 9 performed The UQC interconnecting cable was severed better than those in Array 8. However, this con- about 20 ft from the hydrophone. clusion is qualified by differences in seal mater- ial, period of service, cable performance, and The corrosion and calcareous deposits on the assembly errors. junction chamber and housings of Array 9 were substantially less and their distribution was more Cable Seal Packing Glands and Backup Washers: uniform than those on Array 8. Megohmmeter measurements of the Morrison seal assembly for the Z hydrophone, as received in the No indications of paint were found. junction-chamber end of the interconnecting cable, showed a low (160 kQ) resistance between the center Internal Condition conductor and the shield conductor. The 160-kG path became intermittent after the seal assembly Cables and Seals: The resistivity of the was pushed out of the junction chamber housing. interconnecting cables and seal assemblies for Slight flexing of the cable seal assembly produced hydrophones C, Y and Z and the UQC was measured in erratic changes in resistance readings, which air before the Morrison seals were pushed out of varied between 16 kG and 50 MQ at 500 Vdc. Dis- the junction chamber end-plate. Because many of assembly, inspection, and examination with a scan- the cables were parted or cut, each cable and cable ning electron microscope revealed that the low seal assembly was measured individually. The resistance path was through a black residue on two results of these tests are shown in Table VI. The adjacent teflon washers in the Morrison seal. The Morrison seals in the hydrophones exhibited higher resistance of the black residue between the hole resistance readings than those in the junction for the center conductor and the hole for the chamber. Some factors that may have contributed to shield conductor was 350 kQ on both of the two the noted difference in behavior are the attitude adjacent insulating washers. At the site of the 279 Table. VI. Sumnary of Performance/Failure Analysis on Array 9 C X Y Z X-Y UQC Seals, O-Ring Hydro Trans OK OK OK OK OK OK Hydro Cable OK OK OK 1st Leak OK 1st Leak JC Hydrophone .--- - - - - - JC Main Cable OK - --- --- --- - OK Seals, Morrisona Hydrophone (end) --- --- --- --- --- --- CS --- --- 50 kM CH X -- -- 20 kM SH 30 M --- 10 M 3 M --- 13 M Junction Chamber (end) -- --- --- Lightning ---- CS 40 kM --- 40 kM 160 k --- 50 kM CH 40 kM --- 40 kM 5 M --- 18 kM SH 10 M --- 3 M 5 M --- 8 M JCCS Failure .--- --- - Arcing --- --- Lightning Corrosion Depth Hydrophone --- --- 1.4 in. 0.85 in. --- --- Cause --- --- Crevice-Corrosion --- --- JC --- --- 1.4 in. 1.4 in. --- --- Cause Crevice-Corrosion Cable, Interconnector: (RG 58) Condition Cut Intact OK Good Intact Intact Damages --- --- - Cut --- --- Cause ----- ---? --- --- Transducer Exposed, Exposed, Impedance Plot Exposed Straight OK OK Corrosion ND ND ND ND ND ND Boot Cut OK OK Cut OK OK Fasteners: Severe corrosion damage Housings: Hydrophone, Poor; JC, ND; Morrison Main, OK Clamps: Severe corrosion damage aVoltage breakdown and resistance measurements: megohmmeter measurements with 500 Vdc indicated by km = 109 Q; M = 106 Q @ 50 Vdc; VOM measurements with 3 Vdc indicated by k = 103 Q. black residue, the center conductor pin and the neoprene used on Array 8. However, this conclusion shield pin were coated by a thin layer of black is qualified by the fact that the service period liquid. The liquid was not identified. for Array 8 was ten years whereas that for Array 9 was six years, and the fact that the reversed cable Figures 10a and lOb show SEM photomicrographs sleeves in Array 8 produced substantial seal dis- of the residue and damaged teflon between the two tortion and increased the susceptibility to leak- conductor holes on the teflon washer. The residue ing. appears to be charred teflon created by lightning or arcing. The damage occurred at the point of Array 6 closest approach between the center conductor and the shield. Array 6, a rigid design, was planted in June 1970 and recovered in September 1976. The specific The post-service properties of the Morrison reasons for recovery are not known to the author. seal packing glands can be summarized as follows: The range status report of 9 December 1973 lists avg. diameter = 1.034 in.; avg. thickness = the following entry for Array 6: "Z bad." 0.375 in.; avg. hardness = 32 durometers; voltage breakdown resistance = - (i.e., exceeds capacity Examination of Array 6 was limited to evalua- of instrument); and appearance, good. The design ting the external corrosion and mechanical damage, diameter, thickness, and hardness are 1.040 � determining the acoustical and electrical perform- 0.005 in., 0.374 � 0.005 in. and 30 � 50 duro- ance of individual hydrophones, and recording the meter, respectively. general condition of the interconnecting cables and the seals. The post-service properties of the packing seal glands of Array 9 indicate that the chlori- nated butyl performed better than the commercial 280 Some damage could be observed to the trans- The corrosion and calcareous deposits on the ducers as the hydrophone assemblies sat in the housing of the hydrophone were not as extensive nor cradles in the shipping crates. The X-Y transducer as deep as those on the other two arrays. was missing. The only part remaining was the base of the boot under the stainless wire. The Z trans- The results of the tests conducted on the ducer was still in the boot, but crushed. The Y Array 6 are summarized in Table VII. transducer was also in the rubber boot, but the boot was almost completely severed and the element was crushed. Table VII. Summary of Performance/Failure on Array 6 C X Y Z X-Y UQC Seals, O-Ring Hydro Trans OK OK OK OK OK OK Hydro Cable OK lst Leak 1st Leak 1st miss.(1) lst 1st leak JC Hydro OK 2nd OK leak JC Shore OK Seals, Morrison ND ND ND ND ND ND Cable Interconnect (RG 58) Condition Cut Intact Intact Cut Intact Intact Corrosion ND ND ND ND ND ND Cause ND ND ND ND ND ND Transducer Impedance OK OK Crush Crush Gone OK Corrosion ND ND ND ND ND ND Boot OK OK Crush Cut Gone OK Cut Base Cut Fasteners: Corrosion; Heavy Housings: (Hydrophones and JC uniform light) Clamps Torn Torn Sleeve Sleeve (1)lst O-Ring missing Observations 4. A small piece of the backup O-ring in the end- plate seal on the main cable end of the 1. Interconnecting cables with polyurethane junction chamber on Array 9 had been cut off jackets exhibited visible circumferential during assembly and become lodged in the first cracks about 1/4 in. long at several places. (seaward) O-ring groove. The first O-ring was These cables exhibited many additional leaks effective. which were easily located by checking the electrical conductivity to seawater and by S. The first, or seaward, O-ring was missing from observing bubble escapement. the interconnecting cable seal housing on hydrophone Z of Array 6. The second inboard 2. Arcing, probably due to lightning, occurred at O-ring was effective. the closest approach (approximately 1/16 in.) between the center conductor and the shield in 6. Of 42 double O-ring seals examined, 8, or the rigid teflon washers of the Morrison seal approximately 20%, of the first O-rings in the junction chamber of Array 9. This leaked. Four of the six first O-rings in the arcing produced a low-resistance path of what cable seal housings on the hydrophones of appeared to be charred teflon. Array 6 leaked. The second O-ring on all these seals was effective. 3. The cable sleeves under the cable clamps for the interconnecting cables on Array 8 were 7. The deepest penetration of corrosion products installed backward. This error produced some (yellow and black deposits) in the Morrison of the gland deformation that led to leakage seals was 1.4 in. This corrosion was due to past the first gland on many of the Morrison an interaction between seawater, the lubri- seals. cant, the housing, and the gland materials. 281 8. The first gland of all the Morrison seals on 3. On recovery of all arrays, either an expert in Array 8 was adversely affected by deformation performance/failure analysis should be on hand due to installing the cable sleeves back- pr a procedure should be prescribed and fol- ward, and to the use of too many shims; i.e., lowed during the recovery, packaging and overcompression of the Morrison seals, shipping to verify the as-recovered condition of all arrays. 9. It is not always possible to differentiate between damage sustained during use and phys- 4. Continuous or regular monitoring of the ical damage sustained during recovery, pack- shield-to-seawater resistance and center ing, and shipping. conductor-to-seawater resistance should be made at the range. Failures or changes in 10. Especially in Morrison seals, dark to black array behavior could then be traced either to stains and corrosion products were found in climatological conditions such as lightning the vicinity of large solder joints in the storms or to operational conditions such as center conductor and the shielding. The voltage or current overloads. color, location, and distribution of the products suggest they were caused by incom- 5. Stricter quality control measures should be plete removal of the solder flux. instituted for the assembly and installation of Morrison seals, solder joints within the 11. Analysis of test data from Array 8 did not seals, cable clamps, and sleeves. indicate that the high pressure on the Morrison seals glands produced any significant 6. There should be greater insulation capacity decrease in the resistance between the con- and greater separation between the center ductors and the housings. conductor and the shield than is presently provided in the Morrison seal designs used in 12. All seaward backup washers in the Morrison Arrays 9 and 6. seals on Array 8 showed substantial reductions (by factors of 104 to 106) in resistivity, 7. All O-rings should be heavily coated with except for the seal at the sensor end of the O-ring lubricant and all O-ring grooves and explosive bolt. The low resistivity was spaces of the 0-ring seals should be filled caused primarily by deformation which contri- with lubricant at the time of installation. buted to the absorption of seawater and cor- rosion products. 8. Double 0-ring or even triple O-ring seals should be used where a long-term (several 13. The backup washers in the Morrison seals on years or more) O-ring seal between seawater Array 8 increased in volume. The diameter of and electronics is required. the seaward washers increased by 1.2% and the thickness by 32%. The diameter of the other 9. Quality control measures during installation backup washers tested increased by 1.2% and and assembly of O-rings should be improved. the thickness by 11%. To prevent cutting and shearing of 0-rings during assembly, more alignment fixtures or 14. The packing seal glands in the Morrison seals alignment procedures should be adopted. on Array 8 showed a dimensional change due to creep, and a reduction in hardness. The 10. A test and evaluation study should be con- diameter of the seals decreased by 1.4% and ducted on elected and candidate materials for the thickness increased by 2%. The measured Morrison seals, spacers, and washers to iden- hardness was consistently 15% below the mean tify the materials that have the best resis- hardness specified in the design drawings. tivity, dimensional stability, and corrosion resistance. Recommendations 11. A cathodic protection system for the array 1. A comprehensive performance/failure analysis harnesses should be designed and used. program for acoustic arrays should be insti- tuted. Each recovered array should be ana- 12. Protective hoods should be put over all hydro- lyzed to determine why it failed and what phones immediately after recovery to prevent improvements should be made. further damage. 2. The value of performance/failure analyses on Acknowledgment arrays, as with any instrument system, can be significantly increased by maintaining a This work was supported by the Naval Facilities complete record on design, fabrication, ini- Engineering Command, Chesapeake Division, Washington tial performance, installation, operation, Navy Yard, Washington, D.C., under Contract N00017- recovery, handling, and storage. 74-C-1208. 282 References 1. NAVTORPSTA Report 1332, "Analysis of Array Components Recovered at St. Croix, VI," by M.M. Abo, Naval Torpedo Station, Keyport, Washington, May 1977. 2. Reference Manual on Interference Seals and Connectors for Undersea Electrical Applica- tions, prepared by the Applied Physics Lab- oratory of the University of Washington for the Naval Facilities Engineering Command, Chesapeake Division (available from NTIS, 5285 Port Royal Road, Springfield, Virginia 22161; ref. Publ. ADA 036841), June 1976. ~ 7.. .. 3. Technical Note 1424, "Inspection of Objects _ _ Retrieved from the Deep Ocean - AUTEC Acoustic Array," by J.F. Jenkins, Civil Engineering Laboratory, Port Hueneme, California, DN 487017, February 1976. Figure 2. Sectioned hydrophone housing, potted preampZifier and acoustic transducer assembly Rx INTERCONNECT CABLE & PREAMPLIFIERS, FIVE ON 30-FT CENTERS INTERCONNECT �-1 ij(HYDROPHONE ASSEMBLY) CABLES 4 JUNCTION CHAMBER TILT MECHANISM STRAIN RELIEF Figure 3. Double O-ring seal of main signal cable end of junction chamber, show- EXPLOSIVE 14i 1 ing cut piece of O-ring backup gland SEPALMTION I t \as it was deposited in the 1st O-ring ELECTRIC groove CABLE Figure 1. Buoyant array, St. Croix 283 Figure 4. Seawater corrosion products in (left to right) X hydrophone, explosive bolt and tilt cable seal housings Figure 6. Typical interconnecting-cabZe seal assembly, Array 8 I !fiU 'l'"" l!!''' :''...... . . I , v-l!' ill~-~lll~i...l~~~'lnallll~lllllll 8 9 19 II . - i9 2 Figure 5. Sectioned Morrison seal housing from the tilt mechanism housing 284 TEST HOUSING FOR SHIELD O MORRISON SEAL HOUSING MORRISON SEAL FROM CENTER CONDUC / WITH DOUBLE O-RING JUNCTION CHAMBER SEAL TO HYDROPHONE HOUSING PRESSURE GAUGE A CLAMP / lPUMP TEST----- PRESSURE WATER CHAMBER / /////////////////// Figure 7. Hydrostatic test apparatus Figure 8. Cracks in urethane jacket opened by bending the cable Figure 9a. Circumferential cracks through urethane cable jacket, SEM (76X) 285 Figure 9b. Magnification of a region in Figure 9a showing microcracks in deteriorated urethane cable jacket, SEM (1,500X) Figure lob. Magnification of Figure 10a, SEM (50X) Figure loa. Arcing damage between conductor holes in teflon washer in Morrison seal, Array 9, SEM (20X) 286 FLOW-INDUCED TRANSVERSE MOTIONS OF A FLEXIBLE CABLE ALIGNED WITH THE FLOW DIRECTION R. J. Hansen C. C. Ni Naval Research Laboratory Naval Research Laboratory Washington, DC 20375 Washington, DC 20375 developed turbulent, annular flow in the test section. The downstream end of the rigid rod Abstract was attached to the upstream end of the flexible cable. The downstream end of the cable was either Experimental studies of the motions induced free or fixed on the test section centerline. The in a long, flexible cable by axial flows have flexible cable was terminated a minimum of 2m been conducted in a large,blowdown water channel from the downstream end of the test section to A cable with a diameter of 1.59cm and avoid any flow irregularities caused by the exit a length of 9m was employed. Experiments were of the flow into the downstream liquid reservoir. conducted over a fluid velocity range of 4.6 to The dynamic behavior of the cable was monitored 9.1 m/s. Both free and fixed downstream termina- photographically by three motorized 35mm cameras. tions of the cable were utilized, the former One was focused on the entire flexible member, the simulating a towed cable and the latter a second on an approximately 2m segment near the mooring cable. Measurements of the drag middle of the member, and the third on a similar coefficient of the cable were also made and length of cable near its downstream extremity. compared with previously made predictions. Because of the horizontal orientation of the test section, a nearly neutrally buoyant test cable was employed. It consisted of a braided 1. Introduction polypropylene rope (1.59cm diameter) through which a 6.4mm diameter water-filled Tygon tube had been Flexible cables aligned with the flow are threaded. The experimentally determined relation- present in a variety of systems used in the ship between force and elongation for this com- ocean environment, including mooring lines posite test piece is shown in Figure 1. aligned with the current and tow cables aligned with the tow direction. The present paper 0.15 - describes a laboratory-scale experimental study of the flow-induced transverse motions of such cables. Experimental results are presented for a 1.59cm diameter braided polypropylene cable having a length of approximately 9m. Results are reported for a cable fixed at both ends (simulating a mooring line) and for a cable which is free at the downstream end (as is a z towed cable.) 05 2. Experimental Facility and Procedures The experiments were conducted in a large, O blowdown water channel. The facility is 0 5 10 15 20 described in detail elsewhere (1, 2.) It was LOAD APPLIED AT THE END OF THE CABLE (kg) designed to subject the flexible cable to a fully-developed turbulent, annular flow of Figure 1: Tension - elongation relationship for constant velocity. The facility consists of the cable upstream and downstream liquid reservoirs, an air tank for pressurization of the upstream When the downstream end of the cable was free, liquid reservoir, and a 15.5m long, 19.1cm it had a hemispherically shaped downstream diameter transparent test section. A constant termination in the present work. Previous exper- flow velocity through the test section was imental studies have shown that the flow-induced realized by holding the total head (pressure motions of flexible cylinders with length-to- plus gravity) constant at the bottom of the diameter ratios on the order of 50 or less are upstream reservoir. A 1.59cm diameter rigid affected by the shape of this downstream rod was mounted on the centerline of the first extremity (3), whereas more recent work with 4m of the test section to produce a fully- much larger length-to-diameter ratios show no 287 U.S. Government work not protected by U.S. copyright. significant dependence on this parameter (1). by the flow was used along with the results of The flow-induced motions with a fixed downstream Figure 1 to obtain the drag coefficient on the termination were studied in the present work by cable. The assumption was made in the calculation attaching this end to a rigid rod in much the that the flow exerts a uniform shear stress over same way as at the upstream end. the entire length of the cable. Thus, the tensile force, Fo, on the upstream extremity of the cable is twice the value obtained from Figure 1 as the 3. Experimental Results force corresponding to the total elongation caused by the flow. The drag coefficient, CD, defined A. Free downstream end by the following equation, is shown in Figure 3. Typical photographs of the center region of F the cable with a free downstream end at a bulk C 0 (1) flow velocity of 7.52 m/sec and of the downstream D pUI~aL region over the bulk velocity range of 4.60 to ANNULA 7.0 ANLRFO 7.52 m/s are shown in Figure 2. + F- -.- 6.0- + ~~~~~~~~5...-- .A......FLOW ++ +4 3.0 - INFINITE MEDIUM 2.0- 1.65 3.6 4.0 4.5 5.0 5.5 6.0 Rax104 Figure 3: Measured drag coefficient of the cable and theoretically predicted drag coefficients of cylinders with smooth surfaces in annular flow and in an infinite medium. Here P denotes liquid density and a and L the cable radius and length, respectively. Also shown in Figure 3 are two analytical predictions for the drag coefficient of a smooth cylinder in annular flow (4) and one for a smooth cylinder in an infinite medium (5). The mean positions of the cable with a free downstream end at four of the five bulk flow velocities at which extensive data was taken are shown in Figure 4. The vertical scale is -20 z CABLE DIAMETER UNSTRETCHED CABLE LENGTH 8.32 (hi) - (O Figure 2: Photographs of the cable with a free end (from top to bottom the center region of the DISTANCE FROM THE FIXED END(m) cable at a bulk fluid velocity of 7.52 m/s and 6 4 2 o ~ I S I I ~I I the downstream end region at 7.52, 6.70, 5.85 8 I ~~'' 7.52m/s 5.09, and 4.60 m/sec). 7.52 r/s ~ // ~~4.60 m/s .~~ The flow is from right to left, and the bulk flow -,85 .n/s I -. 0 velocity in the test section, hereafter denoted \ 09 rm/sF by V, is approximately 0.88 times the maximum ' value, U. The actual displacements are 0.75 20 times those recorded by the cameras because of the curvature of the pipe wall. Transverse motions are readily apparent, as well as a change Figure 4: Mean position of the cable with a free in cable length with flow velocity. downstream end (relative to the test section The change in length of the cable produced centerline). 288 magnified 120 times compared to the horizontal can exist with increasing flow velocity: (a) the scale for the sake of clarity. The displacements entire cable is in tension, the tensile force of the mean positions of the cable from the test varying monotonically from a maximum at the section centerline aredue to slight deviations upstream end to a minimum at the downstream end; from neutral buoyancy. These displacements and (b) a zero-tension condition exists at the the associated angles of the cable to the mean downstream end; (c) the zero-tension condition flow are very small, so their effect upon the is between the two ends of the cable, with flow-induced motions of the cable are thought to observable slack downstream of this point. This be negligible. range of conditions may be anticipated from a The rms deviation of the cable position from force balance on the cable, which takes the form the mean, a, is a measure of the magnitude of the flow-induced motions in the cable. Figure 5 F(s) = + 2 r a (7-x) (2) 2' 10 - Here F(x) denotes the force in the cable at a distance x from the upstream end, F. the tensile force in the cable in the absence of flow, rw 0 the wall shear stress exerted by the flow on the E cable exterior, and L the cable length. Also the assumption is made in deriving equation 2 Z that F(x) is greater than or equal to zero in the whole cable or if not that the cable can support 4 compressive as well as tensile forces. Since w the first of these assumptions is satisfied when oS the tension is zero at the downstream extremity of the cable, equation 2 may be used to obtain o FREE END Of THE CABLE the following relationship between F. and other 7.0 Om FROM THE UPSTREAMED END parameters at this critical condition: 0 . F, ( P U) (C D) (17 a L) 5 ~~~6 7 8 SULK FLOW VELOCITYV Wmis) Experiments were conducted at bulk fluid velocitites of 4.60, 5.64, and 5.97 m/s to realize the range of conditions which has been identified. Figure 5: RMS deviation of the cable with a free The mean position of the cable in the test section end from the mean location. at these velocities is shown in Figure 6. shows a over the range of bulk flow velocities -- studied and at two axial locations: the free end DISTANCE FROM UP-STREAM END OF THE CABLE (m) and 7.Om from the fixed upstream end. An 8 6 4 2 average of 30 measurements of cable position were used to obtain each data point. The ampli- tude of the flow-induced motions is weakly dependent on the flow velocity. Larger values at 4.60 and 6.71 m/s are probably attributable to the existence of in-phase motions of the dominant zV4_ frequency components at these velocities. At all N" ' V=460m/i velocities the values of a decrease with increas- / 59 ing distance from the free end. This trend is apparent in Figure 5. \ The frequency content of the flow induced motions of the cable with a free end could not be . K \ determined quantitatively from the photographic J I .! records. The frame rate of the cameras (3 per second) was too low to preserve frequency infor- _ _L mation according to the Nyquist criterion. Visual observation suggested that these motions -15 were of predominantly low frequency (on the order of 10 Hz or less), however. Subsequent studies of the transverse motions of large L/D, smooth- Figure 6: Mean position of the cable with a surfaced cylinders in axial flow have shown a fixed downstream end (relative to the test section continuous spectrum of vibration frequencies, centerline). with the spectral intensity largest in the 3-10 Hz range (1). The zero tension condition at the downstream end occurred at a bulk velocity of approximatedly B. Fixed downstream end 5.2 m/s, corresponding to a value of Fs from equation 3 of approximately 4.2 kg. Slack in the When the downstream end of a cable in axial downstream region of the cable is apparent at flow is fixed in space, three distinct conditions higher velocities. 289 The value of a at three axial locations is less (3). As flow velocity is increased adjacent shown in Figure 7 for the three fluid velocities. to such small LID cylinders, static divergence is replaced by progressively higher oscillatory 5 modes. No circumstances similarly dominated by a A -- 4.27m a single mode of deformation could be identified X----- 7.Olm for the large L/D flexible cable. E Further experimental studies would be helpful in more completely characterizing the flow-induced b ~~~~~~~~~~~~~mot ions of long, flexible cables and identifying i 3 ~~~~~~~~~~~~~the dominant mechanisms by which they are 3 generated. Measurements of the transverse motions should be made which allow determination of their > ~~~~~~~~~~~~~frequency content at selected locations along the W x * cable. Companion two-point measurements should be W ~~~~~~~~~~~~~conducted to ascertain the dominant propagation velocities and directions of transverse distur- bances as a function of frequency. The optical a---- 7.85m displacement sensor technique utilized in recent I ~~~flexible cylinder studies (1) would be well suited 4.5 5.0 s.5 6.0 for such experiments. BULK FLOW VELOCITY, V (rn/a) 5. Conclusions Figure 7: EMS deviation of the cable with a fixed downstream end from the mean position at the three Transverse oscillations are excited in long distances from the upstream end indicated, cables by axial flows. These motions are largest near the downstream and when this end is free. A trend of increasing amplitude of the flow-induced For the fixed end case the motions are a function motions with fluid velocity is observed. As in the of the tension in the cable in the absence of free downstream end configuration the motions flow. These motions should be of importance in appeared to be of predominantly low frequency, towed or moored systems with components sensitive but the frame rate of the cameras was too low to to low-frequency and small-amplitude oscillations. make a quantitative determination of frequency spectrum as a function of fluid velocity and Acknowledgement: This work was supported jointly axial position. by the Office of Naval Research (Code 222) and the Naval Research Laboratory. 4. Discussion References The results reported herein show that axial flows can excite transverse motions of towed and 1. Ni, C.C. and R.J. Hansen, "An experimental moored cables. These motions are largest in study of the flow-induced motions of a flex- conditions of low cable tension, such as near the ible cylinder in axial flow," (in preparation, downstream extremity of a cable with a. free end 1978). or, if both ends are fixed, a cable with small tension in the absence of flow. Even in these 2. Hansen, R.J. and C.C. Ni, "An experimental circumstances the motions are of the order of a study of flow-induced motions of flexible cable diameter or less over the range of para- cables and cylinders aligned with the flow meters considered in the present work. Moreover, direction," ASMdE Preprint 76-WA/FE-15, 1976. the motions appear to be of predominantly low frequency. Thus, the transverse motions induced 3. Paidoussis, M.P., "Dynamics of flexible by axial flows should in general be of engineering slender cylinders in axial flow; part II: significance only for cables or cable systems used experiments," Journal of Fluid Mechanics, in conjunction with instrumentation packages or Vol. 26, 1966, pp. 737-751. other equipment which is sensitive to low- frequmency smal-mpistudibrations.4 Knudstewen, the and He. atzTase, Mclawuill BokDo.ynamcs, frequmercy, dsml-pitud ibrations. exis Knudsen J.N. and He. atzTase, McrwHlluidC. Dynamc., flow-induced motions of the large LID cables New York, 1958. studied in the present work and those reported previously for flexible cylinders with smooth 5. White, F.M., "An analysis of axisymmetric exterior surfaces. Static divergence (a station- turbulent flow past a long cylinder,"1 ASMdEI ary bending deformation) is not present in the Transactions: Journal of Basic Engineering, fixed or free downstream end cables over the range Vol, 94, 1972, pp. 200-206. of parameters studied. Static divergence is present above a critical flow velocity near the free downstream end of a large LID cylinder with a smooth exterior surface (1). It is also evident above a critical flow velocity in flexible cylin- ders having relatively small LID ratios of 50 or 290 THE STATE OF TECHNICAL'DATA ON THE HYDRODYNAMIC CHARACTERISTICS OF MOORED ARRAY COMPONENTS J. H. Pattison P. R. Rispin David W. Taylor Naval Ship David W. Taylor Naval Ship Research and Development Center Research and Development Center Bethesda, Maryland 20084 Bethesda, Maryland 20084 Abstract similar systems. Data from selected sources were extracted and arranged according to component shape A survey is presented of the available and major grouping of hydrodynamic characteristics, hydrodynamic data on components of suspended- including hydrodynamic load coefficients, added array mooring systems. The data are used by the mass and damping coefficients, strumming, and system designer in calculating the forces and responses in waves. motions induced in the system by the dynamic As to the state of the data, complete data ocean environment. The components include a were found for components with simple geometric variety of shapes of packages which contain shapes, such as cylinders and spheres. However, oceanographic sensors and other moored array data for more complex shapes were incomplete. In components. The data include hydrodynamic load the sections which follow, the quality and quantity coefficients, added mass and damping coefficients, of data are assessed. Suggestions are given for and strumming effects which result from surface research to improve the quality and quantity of waves and subsurface currents. The various con- certain data. ponents are catalogued according to types and shapes of components. Types include sensor and array elements, mooring lines, and buoys. 2. Hydrodynamic Parameters The quality and quantity of data are assessed. Suggestions are given for research to improve In the ocean environment, moored array com- the quality and quantity of certain data. ponents are acted upon by different forces, in- cluding gravity, tensions from the mooring lines, hydrostatic forces, and steady and unsteady hydro- dynamic forces. All the forces affect the re- sponses of moored array components to the environ- ment. This paper concentrates on the hydrodynamic forces. 1.. Introduction As discussed by Morgan (Reference 1), steady hydrodynamic forces induced by a steady current on Various organizations involved in ocean a body may include both friction and gravity com- research develop suspended array systems for such ponents and are relatable to the inertia of the F ~~~applications as the measurement of environmental moving fluid by the following Newtonian force parameters, weather prediction, and the tracking coefficients: of military and commercial shipping. Designers of the moorings for these systems require a CD knowledge of the hydrodynamic environment, in- Drgcecet D (1 2)P(2 cluding surface wave spectra and subsurface (12p D currents to be expected at the site of interest. They also need to know the responses of the Lift coefficient C = L and (2) system's components. When the environment is L (1 2PUA known many of these responses may be determined from the hydrodynamic characteristics of the Momn ofiin 3 various components. This paper explores the Momet1cofficent M =L state of technical data on the hydrodynamic char-(l2U A acteristics of oceanographic sensors and mooring components and indicates their use in determining where and are the characteristic areas for the responses of such systems. drag an lift ksuch as frontal area and planform In an investigation for the Naval Facilities area), D and L are drag and lift forces, k is a Engineering Command, the published and unpublished characteristic length of the body, M is pitching literature was searched to establish the present moment, U is flow velocity, and p is water density. level of technical knowledge of experimental and These coefficients typically depend on angle of analytical procedures and of data on the hydro- attack and on the Froude number, P n = U/,N gt and dynamics of components of moored arrays and Reynolds number, R n= Uk/v where v is the kine- 291 U.S. Government work not protected by U.S. copyright. matic viscosity and g is the acceleration due to particles generated by waves is at its maximum at gravity. the surface and diminishes below the surface. When The typical case of a sensor or array corn- water particles move less than a significant length ponent oscillating in pure translation, as shown of the moored component, the effect of waves can be in Figure I may be used to illustrate unsteady neglected. The relationship between the amplitude forces. In this case, the unsteady forces can be of water particle motion in deep water and the described by an equation of motion of the form: surface wave amplitude is shown in the following equation: m'i + ci+ kz =F(t) (4) 2 A(h) = A 0exp(-w b/g) (7) where c is the damping factor, k is the spring constant provided by the mooring lines, m' is the where A(h) and A 0are single amplitudes of the virtual mass (mass of the body m plus the added water particle and waves respectively, h is the mass of water ma ), z is the displacement of the depth of submergence and w is the wave frequency component from an equilibrium position, and the in rad/sec. dots represent time derivatives. Similar equa- One empirical formulation of surface wave tions exist for other translational and rotational amplitude is that of Bretschneider, (Reference 2) modes of motion. Added mass and damping data are 32.4 H often presented as functions of amplitude and 52 frequency of motion. Also, spring constants are o 2 W3 exp 4 ,4) feet(8 functions of the configurations in which the com- a a ponents are moored. The above oscillating system has a natural frequency given by: where H 95is the significant wave height in feet4 and T Is the average period of the wave in =f secongs. o 3\r (5 It'follows from the above discussions that the sources of some hydrodynamic forces on mooring components depend on the depth at which the com- ponents are submerged. When a component is on or t ~~~~~~near the surface, surface waves are the primary Extensible 4Heave Motion z(t) source of hydrodynamic loading. When the com- Mooring Line BP ponent is submerged, current is usually the main source of hydrodynamic loading. Changes of water depth also affect hydrodynamic characteristics. T ~~~~~~ When a component is submerged in intermediate or shallow water where the speed of the current is 1 ~~~~~~high, the strumming force induced by the shedding Current - ~ jor vortices is important. It is not an important I ~~~~~~factor, however, when the component is deeply sub- � ~~~~~~merged and the speed of the current is low. d ~~~~~~Thus, the nature of the hydrodynamic charac- -A Hd ~~~~~~teristics of a mooring system is related to the location of the array and to the depth of sub- mergence of each of its components. The importance of the 6epth of submergence on the hydrodynamic Figure I Cylindrical Component Subject characteristics of the types of component con- to Typical Flow and Motions sidered here is indicated in Table 1. With the above considerations, the designer may use hydrodynamic data summarized below in various mathematical models to predict the per- In a steady current, cylinders, spheres and formance of a moored system in the ocean environ- other shapes often are subjected to transverse or ment. Details on various models are given in lift forces due to vortex shedding at a frequency References 3, 4 and 5. defined by: us f n (6) 3. Survey of Data Available 5 z For convenience in the assessment of hydro- where S is the Strouhal number, which is a func- dynamic data, moored array components may be tion ofnReynolds number. When the frequency f divided into three general groups according to approaches f , vortex shedding excites the con-s component shape. The first group contains basic inorsnne0hspeoeocle shapes of packages and housings which enclose ponent itreoac.Tiphnmnncledoceanographic sensors and other subsurface corn- strumming, can lead to structural failure or sensor overload. ponents. A second group contains basic shapes The wave effects on moored components also of various mooring lines. A third group contains must be considered. The motion of the water various shapes of surface and subsurface buoys. 292 TABLE 1. IMPORTANCE OF VARIOUS Basic Shapes for Packages and Housings HYDRODYNAMIC CHARACTERISTICS The basic shapes include the following generic shapes: circular cylinders with flat, hemispherical, and conical ends; spheres; oblate p" Forces,. I) ai seruinU g stF- n Sn and prolate spheroids; and general streamlined -rye0 LtL (drag, lift) ,sshapes such as torpedo shapes. greIn deep sea where A C C C A C C Of all the shapes considered, the hydro- law crre"texists idynamic characteristics of prolate spheroids and [nrteedffte and 0A.8C- AAAtorpedo-shaped bodies have been most extensively law subnrnceA where C A A A A higher current exits investigated. Thus, data for these bodies are pwed surfacewhere SD A A A B the most reliable for ocean engineering applica- and surface where 8 A A A g A A wove effect exist tions. where The existing data for finite length cylindri- A: Predoaiant paraweter cal bodies tend to be sketchy and piecemeal. A B: loportant parameter measure of engineering judgment is recommended in C: Nonimportant parameter R.: WReynolds uer the ra of nea t o vicous foces selecting the hydrodynamic data for these com- F Frmude Number: the ratio of inertia to gravity forces ponents. Sn: Strouhal Number the period parameter of vortex shedding However, finite length cylinders are often used to house oceanographic sensors. In a typical situation, as shown in Figure 1 these sensors are subjected to current normal to the cylinder and to mooring line induced motion along the cylinder. For the three groups, the availability of The current induces a normal drag on the cylinder the data is summarized in a matrix form in which may be determined from a plot of drag pro- Table 2. The letter symbols indicate whether portionality CD/C shown in Figure 2 as a func- D Dow n in Figure 2 as a func- the available data are ample, adequate, or tion of fineness ratio L/d. Equation (1) gives inadequate for each combination of parameter the form for the drag coefficients C and C and type of component. Ample indicates that where the area A dl and d is the iameter and many sources of data are available to charac- � is the length of the cylinder. Typically, terize the parameter in question. Adequate C = 1.2 for the nonvibrating cylinder. As to indicates that enough sources are available to the forces experienced in the heave motion, added choose from. Inadequate indicates that either mass is one input to the equation, Equation (4). too few sources or no sources are available. It is shown as a function of fineness ratio and an acceleration modulus in Figure 3, from Refer- ence 6. The added mass is given as a ratio to the mass of the displaced water TABLE 2. AVAILABILITY OF HYDRODYNAMIC DATA p 2 md = pd. (9) Range of Flow Added Dynamlo Type of Parameters Forces Mass and Responses Components Cylinders B C 8 B B C C Spheres A A * A A B Spheroids A B 8 B a 8 C j Ellipsoids C C C 8 B C 1.0 h Other StreMllne 8 A 8 8 B C H Shapes :-I 8 0.9 - Rigid Cylindr, A A A A A A < Bare Cable A A A B 8 a Z .8 Chain A CC C C C HC Cable airings: Ef Fringe l Rbbon B BC C C B to C . 7 Streamlined A C C C C Other ABC CC C 0 0.6 Caheled Comine d Co 0.5 _ General Floating A 8 B 0.4 1 I r � Specific Sur"face A O0 10 20 30 40 Buoys BOyISO BSoyn A 8 C FINENESS RATIO: L/d Buoy ApperndMee A C C A Ml i S Ade4 C Inede �* notapplicabie Figure 2 - Drag Coefficients on Cylindrical Bodies in Normal Flow 293 100 iaeOf all the shapes considered, the most data aeavailable for rigid circular cylinders which 2 .are considered simulations of flexible cables. - ~ ~~ U/ad 5 (top line) ' Adequate data are available on bare cables of cir- 4 cular cross-section. There is a need, however, to 3 determine the hydrodynamic characteristics of chains 2 *and fairings. Particularly, the added mass and I (bottom) damping coefficients for chains and fairings are -0 10- needed for the design and dynamic analysis of we . moored systems. o Bare cables are often used in mooring and - ~~towing applications. The hydrodynamic character- * ~~istics of cables are quite different from those of * - ~~~~~~~~~cylinders of infinite length. One reason is that cables are flexible. A long cable may respond to ~~~ 1.0- ~~~~~~~~a flow in many different modes of vibration, W ~~~~~~~~~~~~~depending on the tension and flow speed. For . ~~example, the drag coefficient of a bare cable in a flow speed of 2 knots may differ from that of a smaller bare cable in a flow speed of 3 knots, even if the Reynolds numbers are the same for both cases. The tension, mass, and length of the cable 0.1.must be considered in determining the proper drag 0.01 0.1 1.0 2 coefficient for the cable in the flow. Skop, Griffin and Ramberg (Reference 7) have FINENESS RATIO t/d developed a mathematical model which accounts for the interdependence between tension, mass, and length of cable and drag on the cable. In many Figure 3 -Added Mass on Circular Cylinders cases, the model yields drag coefficients which in Axial Motion as a Function of Fineness are too high. For example, on a long cable the Ratio lid and Acceleration Modulus U2/ad vibration induced by vortex shedding does not lock on to one distinct mode. A better fit to sea trial data is obtained if the maximum amplitude distri- 2 ~~~bution in the model is reduced to 15 to 25 percent The acceleration modulus is defined as U 2/ad of its computed value, and this reduced amplitude where a is the half amplitude of a sinusoidal is used in the remainder of the computations. acceleration. In the absence of data on damp- Average drag coefficients of 1.7 and 1.9 are ing on cylinders in axial motion, damping forces obtained on bare cable at sea. are assumed to be quasi steady functions of the Published data are available for other drag computed at any instant of time from mooring line shapes. See Reference 8. D(t) = 1/2pz 2 (t) d2 CD' 10 Various Shapes for Buoys Buoys are considered to include floating where for a blunt cylinder C D. 0.83. bodies, surface and submerged buoys, and appendages Except for surface buoy applications, which of buoys. As components of moored systems, the are treated in a later subsection, the typical buoys experience wave, current, and wind loads. component sensors usually are located in inter- The important hydrodynamic characteristics are the mediate and/or great depths of submergence. As damping, drag, and added mass of the buoys, and shown in Table 1, the dominant hydrodynamic char- the response of buoys in waves. Strumming is not acteristics for this range of depth are the steady an important factor because the drag and wave force, added mass, and strumming force. Since forces are larger than the strumming force induced the strumming force is not considered important by current. in the case of streamlined bodies, ample data Data available on basic shapes discussed in exist on the steady forces and added mass for the last two sections are general enough to be sensor and array components in deep and inter- applied to wide ranges of sizes and shapes of mediate submergence. Only limited data are moored array components and mooring lines. By applicable to the case of shallow submergence contrast, the data available on buoys apply to where the strumming and wave effects are impor- specific shapes and sizes of buoys and may not be taint. easily generalized. As a sample of data on buoys, Figure 4 com- Basic Shapes for Mooring Lines pares the drag characteristics in smooth water of various buoy shapes and sizes listed in Table 3. Typical shapes for mooring lines include Here, for the same buoyancy (payload carrying bare, smooth jacketed, and faired cable and chain. capability) hemispheres and disks have lower drag The cable includes wire rope and other construc- than do spheres. As another sample, Figure 5 tions, such as electromechanical cable. shows the heave response of a spar buoy whose 294 characteristics are given in Table 4. See also Reference 9. 8 EXPERIMENTS W ' 7 - ~THEORY- OAO=0.128m SPHERES: 6 - A�0lO2m o A oH 5 o B At; 4 - vELLIPSOID V HEMISPHERE H 2 _ A DISK 1 o 0 .5 (See Table 3) I H 0.5 . �l O .05 .10 .15 .2 .25 .3 .35 .4 .45 .5 0.4 0.4 �- 4 FREQUENCY OF REGULAR WAVES (Hz) 0.3 - � 0.2 Figure 5 - Heave Response of Spar Buoy Described in Table 4 03 0.1 -_ i X I I i , CZa TABLE 4. FULL-SCALE CHARACTERISTICS OF SPAR BUOY 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1.0 FOR WAVE RESPONSES SHOWN IN FIGURE 5 FROUDE NUMBER F BASED ON FLOAT DIAMETER d Buoy Radius 0.2667 m EQUATION: F = U/ X overall Lengt 5.600 Mnas 1182 kgn Draft 5.155 Figure 4 - Drag Characteristics of Spheres and Other Shapes at a Vertical Force to Displacement Ratio of 0.75 Posltion fG 1.735 above Baseline Posltion of 5 2.578 above Basellne TABLE 3. PHYSICAL CHARACTERISTICS 0.843 OF SURFACE FLOATS SHOWN IN FIGURE 4 Moenst of Inertia 0.304 G � Center of gravity 8 - Center of buoyancy 5. Summary and Conclusions |oOd ioaneter Sleight in Air Oisplacennt in. c lb 51 lb N The hydrodynamic characteristics of moored BShnreA 27.75 70.5 69.1 307 409.1 1820 Altinusa array components such as buoys, cables, fairings, Sphere 8 25.00 56.0 17.5 78 335.3 1492 Rubber and sensors are available in published and un- Sphner C 12.00 30.5 14.2 53 32.7 145 Laintde tI.od published literature. These data are presented Eolipoid 9.50 24.1 17.8 79 32.4 144 ULamnatd Wood, in tables and figures on basic component shapes, 19-inch long e -finah1e 0lon.g on basic mooring line shapes, and on buoys and isin 20.65 12.23 10.3 46 115.6 336l5.inch .digh are presented in terms of hydrodynamic coefficients, Olsk la.so 47.0 1S.o 71 58.2 259 Lanated ioodg. added mass and damping, strumming, and wave force 6-inch highand response of buoys. For the basic component shapes, there are adequate data for the steady flow forces on the shapes either in the submerged or floating con- dition. Adequate inertia and damping coefficients also can be found or calculated with existing methods. Inadequate data exist for strumming forces, body responses to strumming excitation, and the effects of interference from appendages, attachments to mooring lines, or other nearby components on the hydrodynamic characteristics of the basic component shapes. 295 For the basic mooring line shapes, there are 6. Holler, R., "Hydrodynamic Effect of Harmonic adequate data on the drag and strumming forces Acceleration," Naval Air Development Center but inadequate data on the itiertia and damping Report AESMTM-29-71, 1971. characteristics. In addition, the hydrodynamic data on chains are inadequate. It should be 7. Skop, R. A., 0. M4. Griffin and S. E. Ramberg, noted that the data for cables are either "Seacon II Strumming Predictions," Naval obtained from rigid cylinders or from short Research Laboratory Memorandum Report 3383, cables. The hydrodynamic characteristics of long October 1976. mooring cables in water are being investigated as parts of ongoing research programs. 8. Folb, R., "Experimental Determination of For the available buoys, there are adequate Hydrodynamic Loading for Ten Cable Fairing data sources for the flow forces and the inertia Models," David W. Taylor Naval Ship Research and damping coefficients. Data are inadequate and Development Center, Report 4610, November on the effects of appendages on the hydrodynamic 1975. characteristics of buoys. A systematic evaluation of the literature 9. Dern, J. C., "Unstable Motion of a Spar on basic shapes of moored array components Buoy," 9th Symposium of Naval Hydrodynamics, indicates a lack of data on the effects of vibra- ONR, 1972. tion on their hydrodynamnic characteristics. A component, being a part of a moored system, is always in a state of motion. As a result, the flow forces of drag and lift, and the inertia coefficients on a vibrating body are different from those on a body stationary in the flow. An experimental program is needed to evaluate the effects of vibration on the hydrodynamic characteristics of streamlined bodies. The effect of the length of a flexible cable on the hydrodynamic characteristics of cables and cable fairings is not fully resolved. Analytical and experimental studies are necessary to investigate the responses of moored systems to various environmental conditions. The effects of the interference of other components on the hydrodynamic characteristics of moored array components are not yet known. An experimental program is suggested to study these effects. References 1. Morgan, W. B., "Testing of Hydrofoils and Propellers for Fully Cavitating or Ventilated Operation," 11th International Towing Tank Conference, 1966. 2. Berteaux, H. O., BUOY ENGINEERING, John Wiley & Sons, Inc., New York, 1976. 3. Albertson, N. D., "A Survey of Techniques for the Analysis and Design of Submerged Mooring Systems," U.S. Naval Civil Engineer- ing Laboratory, Technical Report R-815, August 1974. 4. Dillon, D. B., "An Inventory of Current Mathematical Models of Scientific Data- Gathering Moors,"~ Hydrospace-Challenger, Inc., Report TR 4450 0001, February 1973. 5. Berteaux, H. 0. and N. K. Chhabra, "Computer Programs for the Static Analysis of Single Point Moored Surface and Subsurface Buoy Systems," Woods Hole Oceanographic Institu- tion, Technical Report WHOI-73-22, March 1973. 296 MOORING COMPONENT PERFORMANCE KEVLAP MOORING LINES G.A. Fowler R. Reiniger Atlantic Oceanographic Laboratory Atlantic Oceanography Laboratory Bedford Institute of Oceanography Bedford Institute of Oceanography Dartmouth, Nova Scotia Dartmouth, Nova Scotia Canada B3L 14A2 Canada B3L 4A2 Abstract that the use of Keviar would extend operational R ~~~~~~~~~~life to 12 months or more and allow current meter Kevlar ropes used as oceanographic mooring data to be collected throughout a 5000 m water lines potentially offer the advantages of cor- column. rosion resistance, low stretch combined with high strength and low specific gravity. Tests com- pleted at sea to date on a variety of line con- 2. Test Program structions to evaluate the performance of the material in the ocean environment, have resulted Extensive work has been carried out in labor- in ambiguous but dissapointing results. Strength atory conditions to evaluate the performance of degradations have been noted that reduce its Kevlar as a mooring material (Horn et al., 1977). desirability as a long term mooring line material. However, before it can be considered as an opera- More testing must therefore be completed to deter- tional mooring material information on its per- mine the reasons for the noted strength reductions formance at sea must be collected. Therefore, a so that methods or eliminating the problem may be program to assess Kevlar under typical environ- found. mental and operational conditions was initiated. In addition to assessing material performance 1. Introduction it was important to learn methods of handling the material and adapting it for use in on-going field The scientific use of data obtained from programs. Termination techniques which would give oceanographic moorings is influenced by overall good strength conversion results while being mooring configuration which in turn depends upon quickly and reliably completed by field personnel the performance of its individual components. In were thus accorded prime importance in the study. F ~~~particular the choice of line material plays a key role in the determination of the maximum mooring Because of the nature of the tests run, two length and operational life of the system. Kevlar restrictions on the results obtained must be mooring lines offer definite advantages such as mentioned. Firstly, Kevlar samples or longer high strength, low stretch corrosion resistance lines were placed in moorings with loading factors and low specific gravity over metallic lines as determined by the scientific program. No attempt described by Swenson (1975) and Ferer and Swenson was made to include or assess the effect of (1976). ~~~~~~~~~~transient forces such as those encountered in the anchor last setting technique employed or experi- This paper describes some of the work that enced during recovery. Secondly, the number of has been conducted at the Bedford Institute of samples, was small so that some concern with Oceanography to evaluate Kevlar 29 as a potential statistical significance exists. mooring material so that moorings could be set without restriction of location for extended As is the case with conventional materials periods of time. The mooring line sought was a the construction of the strength member is of direct replacement for the 3 x 19, 5 mm (3/16" major importance. For Kevlar it is even more nominal) stainless steel wire presently used in critical in that it determines, in large measure, the Institute's field programs. The wire is the type of termination that should be employed. prepared by hand-splicing terminations, which are However, concern over the type of termination to potted in zinc for cathodic protection and, be used played no part in the initial purchase of typically, experience 6-month exposure on taut- material. Four Kevlar rope constructions with line subsurface buoy moorings. It was anticipated breaking strengths in the desired order of 2000 kg (4400 lb) have been acquired during the past two _______________ ~~~~~~years for evaluation in operational use as sum- marized in Table 1. R Kevlar is the registered trade mark of El duPont de Nemours and Co. Rope samples were first terminated and tested in the laboratory to establish a tested rating 297 Table 1. Types of Kevlar Mooring Lines Evaluated Type Construction Impregnation Diameter (mm) Jacket A i x 19 Polyurethane 5.2 (0.205") Extended polyurethane B 3-layer braid Polyurethane 6.1 (0.240") Braided dacron C 7 x 7 Polyurethane 5.3 (0.210") Braided bonded nylon D Single braid Wax 6.4 (0.250") Braided nylon NOTE: In addition to the rope types outlined above two other samples, of parallel construction, were acquired but never used at sea because of very light or nonexistent impregnation and very loose jackets. rather than the manufacturer's listed breaking including varying the number of tucks, up to 15, strength. Test lengths were then subjected to and varying the tension in the rope being spliced limited exposure on an uninstrumented mooring were tried with no observed improvement. A very before inclusion on an operational system. compact tight splice is obtained if each strand of However, before this stage was accomplished it was the bitter end is given a full 10 tucks around a necessary to achieve a measure of competence in standing strand rather than to tuck each bitter the termination of each rope type. Originally end strand once in sequence, but scatter improve- mechanical terminations as recommended by the ment was minimal. The braided ropes do not lend manufacturer were tried but were discarded in themselves to conventional splicing techniques but favour of splicing techniques. For field work a method described by the Wall Rope Co. involving where the rope must be wound on winch drums and the weaving of a grip was used with good results. passed over sheaves the mechanical terminations In the case of the three-layer braid rope, a time were felt to be too rigid while splicing was more of close to two hours was required to complete attuned to the material. Similarly, potted termi- each termination. For our operations this is nations were not used operationally because of excessive. inflexibility, erratic results obtained and the fact that they were not considered to be field With a bonded braided jacket one of the most installable. time-consuming parts of the termination procedure is the jacket removal in the area to be termi- Kevlar lines constructed in the same manner nated. A soldering iron with a 370�C (700�F) fine as conventional wire rope (Types A and C) a tip makes this job easier and also reduces the simple "liverpool" (or long splice) with a minimum possibility of damage to the strength member. of 10 tucks and covered with a mastic adhesive lined heat shrinkable tubing was found to produce Several moorings (Table 2) were set at dif- a quick and robust termination. However, initial ferent loading factors (fraction of breaking tests with this termination gave results with wide strength) the exposure periods were dependant scatter. Various methods of completing the splice primarily on the scientific operation and there- fore left in place for periods ranging from 3 to Table 2. Kevlar Moorings and Duration Type Depth (i) Mooring Length (m) Exposure (month) Loading Factor (%) A 240 200 2� 12.5 B 240 200 3 12.5 170 130 3 12.5 5180 1180 6 19 6o 30 14 12.5 C 60 30 3 12.5 170 130 3 12.5 54oo 1400 6 25 54oo 1400 6 25* D 6o 30 3 12.5 * This mooring includes a 750 m length which had experienced a previous 6-month exposure. 298 14 months. Figure I shows a typical deep water mooring on which Kevlar materials have been exposed. Current meter records on these deep water moorings have shown mean currents of 4 to SUB-SURFACE BUOY 10 cm/s; however, maximum speeds were in general 2-DISC 1Q00 LBS. between 30 to 40 cm/s. Also, the records have shown occasional peaks as high as 60 cm/s over a time period of i to 2 months. SOM. 7X7 KEVLAR After the moorings were recovered samples were cut from the line to produce terminated 4000 M._ ,, specimens 1.5 m long. Exposed terminations were cut so as to produce suitable test specimens AANDERAA C.M. thereby allowing comparison between material with -4 an exposed termination and material taken from the 43 body of the rope. Samples were taken from una- braded sections of the rope and terminations which 739 M. 7 X 7 KEVLAR showed obvious damage were not included in the test results. 4- 17" GLASS BALLS 3. Mooring Line Test Results 3 Tests of the I x 19 construction rope (Type A) were discontinued after one trial at sea 4762 M. 7 X 7 KEVLAR since it exhibited excessive torque imbalance. It was also found that the polyurethane jacket was extremely prone to damage which made the choice of AANDERAA C.M. test samples difficult and produced questionable test results. 150 M. 7?X?7 KEVLAR The three-layer braid rope (Type B) stood up well at sea, in part because of the durability of the braided dacron jacket. Although the dacron jackets are tough they do add to the diameter of the assembly, 28% as opposed to 12% for braided 6-17" GLASS BALLS nylon jackets in the rope size tested. However, its use was also discontinued when strength losses in the terminations reached the unacceptable levels shown in Table 3. Breaks were mostly 3 located right at the end of the exposed termin- ations. However, when companed with the results 20 M. 7 X 7 KEVLAR later obtained with other rope types these strength losses were not significantly higher. The long term mooring (14 months) was placed RELEASE before tests on the Type B material was completed and when it was recovered test results contra- dicted the results previously obtained. The variability of test values constitutes a major difficulty in the evaluation of the material. It should be noted that the latter mooring was set and recovered in an inshore area without the need 482 M. 3 X 19 STAINLESS ST. for spooling the rope on a winch either during WIRE setting or recovery. The 7 x 7 construction (Type C) has the advantage of almost zero torque imbalance and can be easily terminated with conventional techniques but with the reservation of erratic results noted 3 above. Break strength data shown in Table 3 5420 M._ (ANCHOR) indicate that degradation can be appreciable. The 12-month data is obtained from so few samples that it must be considered with caution, also the noted degradation in the rope of 18% on the 3-month mooring is questionable since the four tests Figure 1. Typical Deep Water Oceanographic were split with two very high and two very low Mooring. breaks. These comments aside, there is evidence of a negative effect of exposure which is not 299 Table 3. Strength Degradation of Kevlar Rope and Terminations after Mooring Explosure Type Tested Break Strength Exposure No. of Samples Strength Reduction % kg month (Deviation as % of Sample Means) Rope Termination B 1910 (4200) 3 4 9 i6 19(10) 6 4 9(2) 6 29(17) 14 2 20(4) 5 14( 7) C 1920 (4230) 3 4 18(12) 5 13(12) 6 21 9( 7) 13 18( 8) 12 2 20(0.5) 1 23 D 2430 (5350) 3 2 6( 3) 4 20( 9) readily explained. The nylon jacket used on this extent as the wire with the result that greater rope abraded much more readily than the dacron snap loading can develop. jacket employed in Type B. In addition, the latex material which bonds the jacket to the strength member softened after submergence and may affect 5. Conclusion the re-use of this material. Of the Kevlar mooring lines which have been The results from one shallow mooring trial examined for possible use as an operational only are available on the single braid rope (Type material in direct replacement of metallic lines, D) but it appears to be following the same trend none can be judged as offering the desired quality as the three-layer braid (Type B) with excessive of assured extended operational life. strength losses at the end of the terminations. Further tests at the 6- and 12-month time periods The testing procedure employed, that of will have to be completed before any conclusion placing the material in the ocean environment can be reached on the usability of this material/ complete with all the shipborne hazards, may have termination combination. caused the contrast with results obtained in the laboratory. But since the material must even- tually be used in that role it is essential 4. Operations at Sea that confidence in its performance at sea be gained. Kevlar mooring lines are well adapted to use at sea with the material's lightness, flexibility The engineering and scientific benefits which and resistance to twist being the salient 'deck' would accrue from the use of Kevlar in oceano- advantages. Extra care, however, must be taken to graphic moorings certainly justify further experi- avoid abrasion of the material or jacket parti- mentation. However, at present there is no eco- cularly during the recovery of a mooring where the nomic justification for use of the material, potential for fouling is extreme. Redesign of considering the price differential over metallic some mooring components, back-up buoyancy packages lines, can be made. for example, will have to be carried out to elim- inate sharp edges which contribute to mooring line abrasion. Slight modification of the anchor last References mooring technique is required in that the ship must be slowed from the speed normally used when 1. Swenson, R.C., "The cable development deploying an all metallic mooring. The lighter programs for suspended sensor applications," Kevlar lines do not form a catenary to the same Naval Underwater Systems Center, New 300 London, CT, 22 September 1975, NUSC Technical Report 4915. 2. Ferer, K.M. and R.C. Swenson, "Aramid fibre for use as oceanographic strength members," Naval Research Laboratory, Washington, D.C., November 5, 1976, NRL Report 8040. 3. Horne, M.H., P.G. Riewald and C.H. Zweben, "Strength and durability characteristics of ropes and cables from Kevlar Aramid fibres," MTS-IEEE Ocean '77, Pp. 24E-1 to 24E-12. 4. Walden, R., "Tests of Kevlar as a new mooring material," Polymode News No. 15, September 1976. 5. Ferer, K.M., "Effects of long term tension on Kevlar Aramid fibre," Naval Ocean Research and Development Activity, May 1977, NORDA Technical Note 8. 301 SPECIFYING AND USING CONTRA-HELICALLY ARMORED CABLES FOR MAXIMUM LIFE AND RELIABILITY ALBERT G. BERIAN THE ROCHESTER CORPORATION Abstract Some of this mishandling is negligence but many cable failures have been attributed to a misunder- The importance of complete information re- standing of E-M cable characteristics and limita- garding system parameters in the specification tions. This constitutes another communication gap of a cable is discussed. System parameters to between the cable handling personnel and the cable be covered include cable performance and mission manufacturer. Another possible reason for mis- profile, environmental consideration, verifica- handling is lack of knowledge of the interaction tion testing, the cable handling system and between the cable and handling system components. connection features. The impact of these param- eters on the material and construction features The corrosion effect of salt water on the of the cable is discussed in terms on the effect material of the strength member and the associated on life and reliability. returement criteria should be evaluated in estab- lishing the mission objective of the cable. The mechanical effect of various cable Although the diameter reduction of isolated steel handling operations are related to life and wires in stagnant and flowing salt water have been reliability. Successfully used spooling proce- test determined, the application to the E-M cable dures, the trade-offs for zero rotation torque composite is very difficult. The effects of balance in and the major factors affecting current flow in metallic strength members has been flexure fatigue life are studied. The effects observed but not quantitively established. Cath- of corrosion in a salt water environment together odic protection of exposed metallic strength with successful techniques currently being used members is being successfully used for large metal to minimize this corrosion are also studied. masses, but it has not been studied as a corrosion preventative technique for steel contrahelically armored cables. 1. Introduction The major limiting factors for contra- The importance of communicating a thorough helially(C-) a ro r e cabes sed n te mainedesciption of the performance, environment, henvical CH romoecalsuedn th marie: handling system and connections to the C-H cable environment are: ~~~~~~supplier is shown to be necessary in order to - misapplication obtain the optimized design. - mishandling - corrosion 2. The Contra-Helically Armored Cable in order of importance. The bases for misappli- Specification cation of a cable include: (1) incomplete speci- fication, including a description of the mission Specification Structure and Coverage and ancillary cable handling system and (2) compromizes in cable design due to late consid- Information necessary for the proper specification eration in the system design. Misapplication, of E-M cables includes: therefore, is primarily due to lack of an - desired electrical and mechanical perfor- information exchange between the systems design- mance and their interaction during the er and E-M cable designer or applications engi- objective cable mission life including neer. storage, shipment and operations. A cable which is properly designed for a - life cycle objective sysem anbe amgedbymishandling such as: - environmental impact including pressure, system can be damaged by ~~~temperature, temperature shock, fluids - using improper spooling techniques (salt water, oils, etc.), mechanical - using an incorrect sheave groove and/or (strumming, vibration) pitch diameter - type of system and major mechanical loading - overrunning characteristics (Ref. 1) - overloading - characteristics of the E-M cable handling 302 and deployment system such as in some load measuring heads and type of - electrical and mechanical terminations traction winch or capstan. When the system is - test verification of the design committed to the use of particular equipment some compatability provisions can sometimes be made in Performance vs. Construction Specification the cable design. In many applications the systems designer Terminations becomes inclined to specify the major construction Compatability of potting and molding materials charaterisics o thedesird E-Mcable TheCompatability of potting and molding materials characterisvtics of the desired E-M cable. The with cable insulations and jacket compounds should primary motivation for this approach is usually be considered. The use of polyurethane as a mold- timesaving because the use of a construction speci- fication can shorten the initial procurement cycle ing and potting compound has become very popular : ~~~~~~~~~~~~~~~~becuase of its versatility� However, a bond of The disadvantage of this approach is firming of th becuase of its versatility. However, a bond of polyurethane to some commonly used insulating specification before the cable manufacturer has an polyurethane to some commonly used insulating opportunity to incorporate experience relative to compounds such as polyethylene and polypropylene is appropriate materials and constructions for meeting difficult. the specification elements as discussed in Ref. 2. To obtain the multiple pressure carriers which To obtain the multiple pressure carriers which The by-passing of the performance specifica- may, as discussed in Ref. 4, include sealing of the i ~~~~The by-passing of the performance specifica- conductor insulation, cable jacket and connector tion also avoids the opportunity for an effective tra de-off study wherein reliabiity fand life are body to the potting/molding compound, those materi- trade-off study wherein reliability and life are als having proven bonding properties should be evaluated for various costs, diameters and weights als having proven bonding properties should be of the E-M cable, ~~~~~~specified and selected. of the E-M cable. The several types of available specified and selected. strength member designs, as discussed in Ref. 3, Armor terminations can usually be adapted to encompass a variety of configurations and metal ansteghmbrdsi.e di fre any strength member design. The desire for field and coating techniques, providing wide latitude of ay seng e he se f e repairability has encouraged the use of epoxy selection for the cable designer. potted or locking cone assemblages. Test Verification of Performance Capabilities 3. Cable Handling The strength member properties of major con- cern are usually breaking strength and flexure fatigue life over sheaves. Other characteristics The life and reliability of a C-H armored which are important to determining usage limits cable can be adversely affected by the handling hinclude stress-strain and load-rotation. secnario which encompasses all facets of equipment include stress-strain and load-rotation,�eetonadue selection and use. Whenever possible the test procedure should incdeciver tsti the nes pcelre sou Spooling Tension: This is the operation of trans- include comparative testing of the new cable or concionc ne cmparevi testi the n cabme oferring cable from a reel which may be the shipping construction with one previously tested in the same reel to a tension drum. As discussed in Ref. 5 the equipment using the same procedures to permit geometric objective of proper spooling onto a correlation of the new cable performance with that geometric objective of proper spooling onto a tension winch drum is to uniformly wind the cable of a cable having some operational history. ofi acbl hvig om opraioal hitoy.so that maximum diametral support is provided to all turns. Contrahelically armored cable has a Conditioning of the cable specimen is an essential part of a test specification to insure greater crushing resistance when all components essential part of a test specification to insure are tightly held in a stable, mutually supportive that the strength member elements have their proper e tightly held in a stable, mutually supportive position such as occurs when the cable is under stable relative relationship for maximum mutual position such as ours when the cable is under support. This condition can be verified by the tension. The problem of proper spooling onto a storage drum whereby the load tensions of the cable specimen exhibiting a repeatable stress-strain *characteristic. A tyical decline in the construc- are assumed by a traction winch or capstan is chaactrisic.A typical decline in the construc- tional stretch, or trpe, i relatively noncritical. In the many systems using tional stretch, or that which is non-repeatable, is ~~~~illustrated in Exhibit 1. tensioning winches, however, proper spooling tech- nique is mandatory in order to avoid cable damage. When several layers are spooled onto a winch drum ~~~~~Handling System ~under tension the drum and inner cable layers can be damaged under the high compressive force. A A major function of the strength member of E-M proper spooling procedure will provide for suffi- cables is mechanical protection of the electrical e spooling o ure i foryfi cient spooling tension to insure a uniformly firm, components. As discussed in Ref. 2 the outer metal crush-resistant cross-section but not high enough crush-resistant cross-section but not high enough strength member types provide the best protection to impose a crushing load on the winch drum. to impose a crushing load on the winch drum. while the synthetic fibers and center strength member designs offer minimum protection. The : ~~~~~~~~~~~~~~~~~~This idealized tension will vary from a maxi- limits of this mechanical protection can be exceed- This idealized tension will vary from a maxi- mum at the layers near the drum to a minimum on the ed by improper design of handling equipment or by outer layers. The maximum tension will generally use of improper procedures. outer layers. The maximum tension will generally use of improper procedures. A not exceed 50% of the cable breaking strength. As a generally applied rule, the spooling tension Major characteristics of the handling system jor cartanter s to the hanli sp ti em should equal expected service tension profile which of importance to the cable specification are bend diameters (winch drums and sheaves), reverse l~ends will include the sum of loads due to buoyed cable diameters (winch drums and sheaves), reverse bends wihpyodadfito rda ocs weight, payload and friction or drag forces. 303 A spooling tension schedule which applies to a sheave to cable diameter ratio (Did) was increased vertical cable deployment is illustrated in Exhibit from 15 to 25 times. The relative service life for 4. Note that the approximate slope of II�% of cable various Did =T ratios can be approximated by: breaking strength per 1000 ft. of cable deployed can also be expressed as about 90% of the cable T2 weight per 1000 ft. in air. A procedure which has2 proven satisfactory is to reduce the tension by �~T the slope value every 500 ft.I When T is increased from 12 to 25 the service life Grooved Sheaves would be increased by: The use of these sleeves has been successful be- 625 =27 ie cause they provide: Z_=27tie successful calaerspacingorm the fairst and The similarily of wire rope and C-H armor construc- succssfu layrs t for thebasi fortions suggests that these relationships may also spooling with a double break; i.e. two one- apply to C-H armored cables. Insufficient data are half pitch offset movements per drum available at this time to verify this assumption. revol ution. - support for the bed (inner) layer by Tru aac increasing the bearing surface area and Tru aac therefore reducing the bearing pressure, P ,The generic term "torque balance" is an expressed as: incomplete description of a cable characteristic. P= 2T Additional information required is the tolerance in B Dd ~~~~~~ terms of the amount of rotation at a tensile load and is expressed in degrees per foot. PB= bearing pressure, lb./sq.in. The construction effect of decreasing the T = cable tension, lb. rotation characteristic of a C-H armored cable is D = groove pitch diameter, in. illustrated in Exhibit 6. In this example, differ- d = cable diameter, in. ent armor assemblies were applied over the identi- cal core. The armor construction of Cable B is a Spooling with a Double Break on an Ungrooved Drum standard 24 x 24 wire which is used on many logging To ahiee te to 12 p t c hoffet oveentpercables. To improve the torque ratio, R , from 2 00 drmTotachionth two filler ofsetrpae noemeded Their to 1.04, the metal ratio of the outer akd inner lenthshul betaIon, the drume ctiparcufrne aneded thei armor layers had to be reduced from 1.6075 to diaetengrthikes should be aboth 0rum tircmfesec n h 0.9683 and it was accomplished by an increase in tecbediameter, oreticness spooled b bu .5tie the inner armor wire diameter and decrease in the the cable dameter beig spooled.outer armor wire diameter. The resultant rotation As shown in Exhibit 5 the drum is layed out by decrease from 18.28 degrees per foot to 0.67 scribing a line across the drum at 90 degrees to degrees per foot was obtained by trading off other the flange at a point about 2 in. behind the cable advantages, such as: entry hole. - the larger outer armor wires of cable B are more resistant to snagging and wear The cable is started onto the drum against the - the larger cable B outer armor wires have a flange and one spacer is inserted after a 180 greater preform spring constant and will degree rotation. The spooling is continued with therefore tend to be more stable and remain the coils tight so that the -I breaks start at the tightly wrapped over a longer service life. scribed lines. ~~~~~~~~The advantage of near zero rotation balance is At the opposite flange the second spacer is greater kink resistance and elimination of the located 180 degrees from the first. The last cable damaging consequences. The mechanism of kink for- coil should fit snugly against the drum flange and mation is initiated when a large amount of torsion- filler. As spooling continues the cable will enter al energy, developed by the rotation and unbalanced the valley formed by the bed layer and proceed in torque, is suddenly released by a step decrease in the opposite direction. tension. This condition can occur when a load is allowed to strike a fixed surface (a condition Sheave and Drum Diameters termed over-running) or the tension is suddenly A rule of thumb for establishing the minimum rlae ttespoted pitch diameter is to multiply the diameter of the When the load is slowly released from an largest wire by 400. A cable having the largest unbalanced cable the slow counter-rotation will wire of 0.030" dia. in the armor would have, by pri h oqei h ro ie ooecm this rule, a minimum bend diameter of 12 inches. pemtthe highu inther-armor frirestionn iniiaeareturn The consequence of bending over a smaller diameter rttion tog ither-rorfictiona angua positiot eun. is a reduced flexure fatigue life, oaint h rgnlaglrpsto. Compartive ests n wir ropehas sown aThe sudden load decrease prohibits initiation 2C5 m3a5atimes ineasts in faigue life has thew of this return process and loops form in the cable- 2.5-35 ties icreae infatiue lfe a thethe reverse process of coiling a cable. There is 304 a tendency of developing one loop or coil per rota- ted with zinc ions and therefore becomes neutral tion of the cable. If the cable is allowed to with respect to the steel. Over a long time remain unloaded or the load was reapplied very period, however, these ions will tend to dissipate slowly, the normal return rotation could occur. and leave a normal salt water corrosion condition. However, upon a sudden reapplication of tension the diameter of these coils or loops decrease to the The sea water corrosion of unprotected steel point where the armor wires are permanently has been measured as a surface reduction of: deformed. In extreme cases the very tight inner - .003' to .008" during the first year armor can form a separate loop which emerges be- tween the outer armor wires. - .001" to .005" thereafter Fatigue Life over Sheaves Alternate materials to GIPS for corrosion WhenC-H rmoed cbleis fexed eiher verresistance include stainless steel, Inconel and sheave C-H arothed handleing fequpent, ethearmovr beryllium copper. The stainless steels have proven wires experience high rubbing contact forces. smwa iapitn;freape h olwn Evidence of wires becoming repositioned is in the are data for tunnel attack on samples of stainless crackling sound heard when these cables are winched steels are 2340 feet depth in a North Atlantic at high tension. Operational conditions which harbor. increase the amplitude of this relative armor wire AS egho unlAtc t ~~ motion, the bearing pressures and the friction all AS egho unlAtc tend to increase wear and shorten the flexure 302 6 inch fatigue life. 304 2 inch 306 0.5 inch The major factors which contribute to this increased armor wire wear include: Some cable users have reported that crevice - dimetr ofsheves nd rums th smalercorrosion of stainless steels can be greatly - dimetr ofsheves nd rums th smaler reduced by using zinc anodes and some others diameters increase the motion amplitude and rpr ucs ihml te nds bearing pressure. rpr ucs ihml te nds - winching under a high load; the bearing The high cost of Inconel and other highly pressure between armor wires increases with corrosion resistant materials has prevented their increased cable tension. broad use. - lubrication; as for any rubbing metals, good lubrication will decrease friction and 5. Conclusions reduce wear. - contamination; mud, sand, rust or other The reliability and life cycle cost of oceano- abrasive materials trapped between armor graphic systems using contra-helically armored wires will increase wear. cables can be greatly improved by considering cable - wnchngspeed; the higher speeds will and connection requirements during system design. aceeaeteefc fohrwerindching . The cable and connector specification should conditions. In extreme cases, individualinldaloprtnlevomnalndhdig armor wires may yield because of having system information to permit selection of optimum insufficient time to relocate to a stable materials and concepts. position in the armor assembly. The selection of handling system components should be made with consideration of the effects on 4 Coroio the mechanical action of armor wires. The internal uorrosiun ~~~~armor friction and wear will increase with larger The most commonly used armoring material is and faster movements within the armor and with galvanized improved plow steel which is usually decreased lubrication. Bend diameters, winching coated with a corrosion preventive compound. This speeds, contamination and lubrication are major compound can be flushed from the cable after a points of consideration in maximizing reliability short service life, leaving the galvanized coating after the optimum design has been selected. unprotected. By corrosion and abrasion, the gal- vanized coating, in turn, may be removed and 6. References corrosion of the basis steel is initiated. To avoid the source of much corrosion, the 1. Stange, Win. F., "New Synthetic Cables and saltwate traped etwen aror lyers fluhingTerminations for Underwater Instrumentation with fresh water or spraying with a lubricant/ S s esIA 97Ana ofrne inhibitor will greatly extend the cable life. 2. Berian, Albert G. and Felkel, Edward M., Additional protection for GIPS can be obtained "Armor Designs Offer a Wide Range of E-M Cable by jacketing overall. This technique protects the Properties", Oceans 75 IEEE/MTS Annual Confer- armor even if punctured. It appears that the water ence. which accumulates under the jacket becomes satura- 305 3. Berian, Albert G., "Specifying Electro-Mechani- cal Cables", Sea Technology, pp. 21-22, July 1975. 4. Berian, Albert G., "Compatability of Underwater Cables and Connectors", Oceans 76 IEEE/MTS Annual Conference. 5. Berian, Albert G., "Design and Handling Factors in the Reliability and Life of Electrical Wire- lines", Inter-Ocean '76, June 1976, Dusseldorf, West Germany. 6. Lilly, Joseph C., "The Mathematical Relation- ship Between the Bending Radius of a Cable and the Length of Cable Lay" ROTATION fill,, I , A. TENDS TO LOOSEN LHL' OUTER ARMOR ? >~//// ~/ I ' I ROTATION Stain Third z ' X~~Second First B. TENDS TO TIGHTEN LHL* OUTER ARMOR EXHIBIT I *Left Hand Lay TYPICAL DECREASING CONSTRUCTION STRAIN OF E-M CABLES EXHIBIT 2 EFFECT OF SPOOLING ARRANGEMENT ON Cid ARMOR Shipping Spool Tension Capstan Length Tension Winch Drum Counter Head EXHIBIT 3 TYPICAL E-M CABLE SPOOLING SET-UP 306 Maxismou Tension 20 1 - )8 ' / ~~~~~~~~~~~~~~~~~~~~~Cable Entry Hole ~L~~~ ~5 , 1t o i at6~~~~~~~~~~~~~~ is 20 25 30 ACable Diameter Deployed Cable Length in 1000 Feet EXHIBIT 4 SPOOLING TENSIONS FOR A TENSION WINCH -. .[54d Filler EXHIBIT 5 SPACER AND DRUM LAYOUT FOR DOUBLE-BREAK SPOOLING EXHIBIT 6 COMPARISON OF CONSTRUCTIONS WITH DIFFERENT TORQUE RATIOS, RT CABLE A CABLE B O.D. .680" .675" Outer Armor N d 36 wires .050" dia. 24 wires .071" dia. Inner Armor Nd? 22 wires .065" dia. 24 wires .056" dia. Pitch diameter, Muter, D In. .630" .604" Pitch diameter, Inner, DP? In. .515" .471" Wire diameter, outer, d P i 25.00 x 10-4 50.41 x 10-4 lire diameter, iner, d i 42.25 x 10-4 31.36 x 10-4 . Metal ratio: NG d2 0.9683 1.6075 Torque ratio: R1 = Nod2o Do sin go 1.04 2.00 NidLi Di sin /i Ultimate Tensile Strength, UTS, lb. 31,450 34,300 Rotation at 16,000 lb. deg/ft. 0.67 18.28 e at 16,000 lb. in/in. 0.82 0.93 Weight in air, Wa, lb./1000' 675 665 Weight in water, Ww, lb./1000' 518 512 UTS UTS (ft.) 64,710 65,960 wI THE DEVELOPMENT AND TESTING OF CURRENT METERS FOR LONG-TERM DEPLOYMENT ON THE CONTINENTAL SHELF K. E. Heldebrandt L. S. Trest Texas Instruments, Inc. Sperry Support Services P. 0. Box 60J.5 National Space Technology Laboratories Dallas, Texas 75222 NSTL Station, Mississippi 39529 E. D. Michelena NOAA Data Buoy Office National Space Technology Laboratories NSTL Station, Mississippi 39529 Abstract the established user needs for ocean current measurements. NDBO's objective was to develop Measurements of ocean currents to support current measurement systems for direct applications U.S. continental shelf construction and environ- in the mixed layer and high survivability situationI mental pro ects have been established as a major along the continental shelf. Accomplishment of the user need., Regimes of interest range from the program objective was based on the judicious selec- surface to 300 meters and from concentrated, short- tion of the more promising state-of-the-art current term measurements to sparse, long-term measurements sensors, and on a coordinated plan of laboratory of up to 3 years. The timeliness, reliability, and testing, field testing, and improvements. availability of the data are frequently more im- portant to the user than absolute data accuracy. 2. NDBO Current Meter Test Plan One of the more critical needs of the offshore industries is for measurements of current during Major elements of the NDBO plan for on-going design events, such as severe storms. Near-real- testing and development of a current measurement time data is also highly desirable during quieter capability are as follows: periods to aid on-site construction and to permit ~Slcin rcrmnadlbrtr timely environmental monitoring in coastal and tesetion, prcurement, sensaorstat hvrvousy estar reios.The MOAA Data Buoy Office (NDBO) ts fcretsnosta aepeiul has initiated a development and test program to pro- songo rms ntsso culue vide current measurement systems to meet these needs. * Modification and adaptation of current sen- NDBO's objective is to develop this capability using sors to facilitate at-sea testing on a sub- existing state-of-the-art current sensors and data surface array designed to decouple the buoys. A parallel effort is also underway to sensors from the effect of surface waves. develop an underwater acoustic data link for long- term at-sea testing and operation of the current *Addition of a remote capability to control sensors. This paper presents an overview of the and monitor the status of current measure- NDBO current meter development and test program, ment subsystems (CMSSs) from the surface including the implementation plan, significant during short-term at-sea testing. features of the hardware design, and the results a Short-term at-sea testing to identify of testing conducted to date. defects in design and operational perfor- mance. 1. Introduction a~~~ Correction of defects, retesting, and mod- A detailed study in 1975 by the Current Meter iiaino M~ o erra-iedt Review Panel within the NOAA Office of Marine transmission to facilitate long-term at- Technology (OMT) established the urgent need for sea testing. measurement of currents in the mixed layers of the * Addition of full capability for control and ocean and on the continental shelf, to support near-real-time transmission of current test marine surveying and offshore construction projects. data to shore. Major obstacles to satisfying this need were the inability of available current measurement systems * Long-term at-sea testing to evaluate relia- to operate reliably and to deliver accurate meas- bility and determine operational degradation. urements under the influence of the dynamic ocean a Post-deployment calibration of CMSSs in the environment within these regions. The Working laboratory, correction of defects, and re- Conference on Current Meters held at the University testing. of Delaware in January 1978, confirmed many of the findings of the 1975 study; the conference also 3. Test Implementation found that neither the need for ocean current data nor the number of associated problems seemed to Selection, Procurement, and Laboratory Tests of have diminished. Current Meters. Three ocean current sensors were initially selected for test - the Cushing Engi- In 1976, NDBO initiated a current meter oeering, Inc. (CEI) electromagnetic current meter, development and test program to satisfy many of the Marsh-McBirney, Inc, (MMBE) electromagnetic 308 current meter, and the Neil Brown Instrument Sys- magnetic tape. tem, Inc. (NBIS) acoustic current meter. Each type employs a different technique to sense and measure current velocity. The CEI and MMB sensors use the electromagnetic principle with orthogonal electric probes, a concept that has been in use for several years. However, the tech- niques of application are different. The CEI uses a large 7-inch cylindrical flow transducer, where- as the MMB uses a 4-inch spherical transducer. The NBIS type uses an acoustic path-length concept in which the relative time of travel of two simultan- eous signals is used to determine current velocity. The basic considerations in the selection of these models over other available designs were their lack of moving parts in the sensor, their potential for accurate measurements in the mixed layer, potential for high reliability and survivability in extreme environmental conditions, and the favorable early test results of the transducer concepts. Three of each type of current meter were pro- cured by NDBO for testing. The original design of each was modified to meet a common electrical and mechanical interface specification and to provide a capability to respond to data rate and status request commands from external controls. Each CMSS measures and records mean values of east and north ocean current vectors, as well as the time the mean was taken and the operational status of the CMSS at that time. Data is sampled at a 1 Hz rate, and a 5-minute mean is computed either every 5 minutes or every 3 hours, depending on the external data rate command. Other sampling and data rates are available by adjustment of in- ternal controls. The cassette tape recorder is standard in all CMSSs, and the recording format has been standardized. Each type has an internal data recording capacity of over 1 year at normal data rates, and is designed to operate to a depth of 300 meters and with a mooring line tension of up to 10,000 pounds. The CEI CMSS (designated Model LCD 612 In Situ Velmeter3) is shown in Figure 1. The pressure hous- ing is 64 inches long, with an outside diameter of 7 inches; the total length is 89.5 inches. It Figure 1. CEI MODEL LCD 612 IN-SITU VELMETER weighs 95 pounds in air. This design has a cylin- drical flow sensor large enough to house all sub- systems. The sensor incorporates an alternating RAT electromagnet and two orthogonal pairs of flow detection electrodes located near the center of the housing and attached flush with the outside container. Figure 2 is a functional block diagram of the basic unit. C ERE Each pair of electrodes senses a voltage pro- portional to the component of velocity perpendicu- lar to the plane determined by the axis of the electrode pair and the sensor longitudinal axis of symmetry. These voltages are conditioned and resolved into east and north components of cur- P ~rent by a vector rotator using the output of a EMFss fluxgate compass which is proportional to the angle of the electrode pair relative to magnetic north. The current components are then processed Figure 2. BLOCK DIAGRAM, GUSHING ENGINEERING, into mean values and recorded incrementally on INC., IN-SITU VELMETER 309 Figure 3 shows the MMB CMSS (designated Model 585 Adaptive Recording Current Meter4). The pres- sure housing is 42 inches long, with an outside diameter of 7 inches; the total length is 60 inches. It weighs 95 pounds in air. This sensor has a spherical electromagnetic water velocity sensor with two orthogonal pairs of flow detection A N A ^ R..'A electrodes. A digital geomagnetic compass meas- ures the orientation of the current meter with 1! I p ~io respect to the magnetic meridian. A microprocessor- based data acquisition system resolves the current vectors into east and north components of current, computes the means of these components, and trans- fers the mean values to an incremental tape unit 00N0 for recording. A functional block diagram of the system is shown in Figure 4. Figure 4. BLOCK DIAGRAM MARSH-McBIRNEY, INC., MODEL 585 ADAPTIVE RECORDING CURRENT METERS The NBIS CMSS (designated ACM-1 Vector-Aver- aging Acoustic Current Meter5) is shown in Figure 5. Its pressure housing measures 32 inches long, with an 8-inch outside diameter; the total length is 48 inches. It weighs 75 pounds in air. This design employs an acoustic phase shift detection scheme for correlating the time of travel of two acoustic beams to the velocity of the fluid medium along the beam path. Two pairs of orthogonal trans- ducers measure velocity components relative to the instrument. The transmitted acoustic beams are focused on a reflecting plate, which then returns the signals to the receiving transducer. A func- tional block diagram of this system is shown in Figure 6. Voltages proportional to the current velocity components are resolved by a trigonometric resolver using the output of a magnetometer compass. The compass output is pro- portional to the angle between the instrument hous- ing and magnetic north. These current components are averaged and recorded incrementally on magnetic tape. A series of investigative tests of elements of the CMSS were planned and implemented prior to delivery of the completed units to NDBO. The three types of current velocity transducers were flow tested to determine their characteristics. The CEI transducer was tested by NDBO in the U.S. Geological Survey Flow Facility at the National Space Tech- nology Laboratories (NSTL). The MMB and NBIS trans- transducers were tested for NDBO in the Naval Ship Research and Development Center flow basin by the NOAA Test and Evaluation Laboratory (T&EL). The initial NDBO test plan specified that all CMSS types would undergo the same test sequence; however, a delay in the delivery of the CEI units, caused by technical difficulties during factory testing, necessitated a revision in the plan where- by only the MMB and NBIS CMSSs would undergo this test sequence. Upon completion of assembly and subsequent testing by the manufacturers, the MMB and NBIS units were transferred to the NOAA T&EL CURRENT METER for a series of evaluation tests to determine the performance and suitability of the instruments for field deployment and testing. Laboratory tests 310 tests were performed on selected CMSSs. Vibration testing was conducted in accordance with MIL-STD-167B and temperature testing was in accordance with MIL-STD-801C. The NBIS CMSS was also subjected to a combined pressure-tension test at maximum specified hydrostatic pressure and tension. No leakage or Ii observable defects occurred as a result of any of the environmental tests. The results of these tests indicated that the CMSSs were capable of >2 ~~~~~~~~~~~~operating within the requirements of the NDBO per- formance specification under these controlled con- ditions, and that they should be suitable for field I deployment and testing. The six CMSSs were subse- quently shipped to Texas Instruments, Inc. (TI), for implementation of the current meter field test program. ~> Control and Status Monitoring to Facilitate Testing. In order to enable long-term at-sea testing of cur- rent meters, NDBO began a parallel effort in 1976 to develop a method to remotely control and monitor the status of current meters under test. This was the first step toward the ultimate goal of trans- ferring test and operational data in near real time from a subsurface current meter to a surface plat- ~A form where the data could be retransmitted to shore by UHF satellite link. ~~t~~pi~u~ An at-sea test of an underwater acoustic communication system to control and monitor from began in December 1976 off Jacksonville, Florida. Frequent boat trips to the site were required to execute the test. Difficulties in obtaining ship support and weather-related delays impacted sched- ules during the test. Part of the equipment was lost while attempting to recover the test hardware for removal of recorded test data. However, the Figure 5. NBIS MODEL ACM-1 VECTOR AVERAGING test met the objective, and reliable performance ACOUSTIC CURRENT METER was achieved under a wide variety of test condi- tions during the 10-month test period.8 During this test period, the Applied Physics Laboratory of the University of Washington conducted a current measurement system concept study for NDBO to define the preliminary requirements for current measurements in an operational scenario. This study indicated the feasibility of a Current Meas- urement System (CMS) similar in concept to that ~anmn~ ~ being tested by TI, but with an acoustic telemetry subsystem that could transmit data to the surface in near real time.9 Beyond the potential increase in value to users of operational data received in near real time, the data could have great value during the current meter testing. Access to data before retrieval of the cassette tapes from the - CMSSs could prevent loss of recorded date in case of catastrophic loss of the CMSSs. It could also Figure 6. BLOCK DIAGRAM, NBIS, INC. VECTOR eliminate the need for frequent sea trips to AVERAGING ACOUSTIC CURRENT METER monitor status during testing. Frequent retrieval of the moored CMS array to recover the test data was costly and had been difficult in the past. were conducted, as well as water flow tests using the Naval Ship Research and Development Center NDBO evaluated both concepts in reference to flow basin. 6,i the immediate requirements for testing the current measurement subsystems, and to meeting future oper- The evaluation included steady flow cali- ational goals. It was concluded that the TI con- brations on the four cardinal headings, directivity cept, with the addition of a near-real-time data response tests in both the horizontal and vertical capability, would provide the best technical planes, and compass calibrations. Environmental approach in the near term. TI was subsequently 311 contracted by NDBO to integrate, test, and eval- TRANSMISSION TO uate a prototype CMS, including CMSSs selected NCSC SHORE FACILITIES from the current meter test program and a TI- developed real-time acoustic telemetry subsystem. INSTRUMENT The tasks to be accomplished by TI include the HOUSE preparation and field testing of CMSSs, a parallel effort to design and test the acoustic telemetry / subsystem, integration of these subsystems with a FLIGHT DECK - - - ' /1I subsurface mooring and buoy data system, and 15mT II installation and testing of the complete system I DECk0 I on a data buoy for approximately 1 year in the Gulf of Mexico. 40 FT Field Testing of CMSSs. In preparing the three MEAN JJ WATER - -- MMB and three NBIS CMSSs for field testing, func- LEVEL "T tional testing and inspection were conducted to I' verify that NDBO performance and interface require- NEI ments were being met. Electrical interfaces, , Itl I WEEK TEST including status request, status reply, high data rate on, and high data rate off, were checked for NSI proper operation with a test set that simulated 103 F. ! 0 FMONTH .'TEST an acoustic telemetry unit. Operational procedures I ES were developed and performed on each CMSS. Execu- NI_ MMS tion of these procedures resulted in the isolation of several front end problems, particularly on the MMB units, which had not been identified during previous testing. All identified problems that ! were expected to impact field testing were cor- BOTTOM-- rected prior to entering the next test sequence. CONDUIT The first planned at-sea test phase of the CMSSs program has been underway at the Naval Figure 7. TEST SETUP AT DEEP TOWER Coastal Systems Center (NCSC) in Panama City, Florida, since April 1978. Tests accomplished to units, as shown in Figure 8. Severe crevice cor- date have included a 2-week preliminary at-sea rosion was observed between one of the titanium test to establish functional operation, followed rods and the lower aluminum casting on one of the by a 3-month extended test to evaluate short-term NBIS units (Figure 9). A plot of the measured performance. current magnitudes by the NBIS units is shown in Figure 10, which shows good correlation between The CMSSs were installed at the NCSC Stage 1 the three CMSSs. Directional correlation was deep-water tower, which is located 12 miles off- also good. shore. Figure 7 shows the test setup. Each CMSS was hardwired to an acoustic telemetry interface simulator to provide interface characteristics identical to the actual acoustic telemetry. This permitted the data rates to be changed during testing and accommodated the linkage of the status monitor through the NCSC tower-to-shore telemetry system for routine monitoring of the CMSSs at the NCSC shore station. The compasses in each of the CMSSs were caged prior to deployment to eliminate the effects of local electromagnetic disturbances. The CMSSs were carefully aligned with the known direction of the tower, so that current directions could be measured. An MMB cylindrical, electro- magnetic current sensor was installed at each test level on the tower subsequent to the 2-week test, Figure 8. FOULING OF MARSH-McBIRNEY CMSS to provide reference current data. No suitable reference was available for the 2-week test. The failure of the MMB units was traced to a Upon recovery of the units following the 2- subtle error in the software which was manifested week test, several equipment anomalies were noted. by the zeroing (caging) of the compasses. Bench The MMB units failed to record data during the tests were ineffective in uncovering the problem entire test, but all NBIS units measured and re- before deployment. This problem was corrected corded data throughout. The severe environment and the three MMB units were reinstalled on the under the tower had resulted in the beginning of tower, along with two NBIS units, for a 1-week fouling on all unprotected surfaces of the MMB retest. The crevice corrosion on the NBIS unit 312 Acoustic Monitoring Reporting Capability. The acoustic telemetry subsystem that is now under development by TI will be used to facilitate long- term testing of the CMSSs. It consists of an Underwater Acoustic Subsystem (UASS) that inter- faces directly with the CMSS, and a Surface Monitor- ing Reporting Subsystem (SMRS) that interfaces directly with the digital data processor of the buoy data system. A functional diagram of the acoustic subsystem and its interfaces is shown in Figure 11. Each CMSS will be connected to a UASS to provide for control and real-time data transfer to the surface. Each SMRS will be capable of communicating with up to 16 UASSs within a radius Figure 9. CORROSION OF NEIL BROWN CMSS of at least 1 nautical mile. 30Ii -Serial No. 100 d / I ---Serial No. 200 I!~~~~~~~~~ ~I ~f], . Serial No. 300 , ~20 ~ b~" 01 Ii I~~~~~~~~~~~ 8; L ~~~~~~---- ---- C) I Figure 11. FUNCTIONAL DIAGRAM OF CMSS AND ACOUSTIC TELEMETRY SUBSYSTEM SHOWING THEIR 4122 423 4 4/2 j 4226 INTERFACES DATE Figure 10. CURRENT MAGNITUDE - NBIS CMSS DURING 2-WEEK NCSC TOWER TEST Long-Term Test of Current Measurement System. Preparations for the long-term at-sea testing of the modified CMSSs are underway, with deployment was found to be caused by direct contact between scheduled for February 1979. The subsurface moor- the titanium rod and the aluminum base. This was ing that will support the CMSSs and interconnected corrected by installing the proper insulator. UASSs is planned to be deployed in the northern Gulf of Mexico approximately 10 miles from South Upon recovery of CMSSs after the retest, the Pass at the mouth of the Mississippi River. A following was observed. All MMB units had suc- 20-foot NOMAD boat-shaped buoy, containing the cessfully recorded data, but the magnitudes of the SMRS and a standard NDBO Data System, will be measurements were grossly in error. The acoustic moored nearby. The water depth at this location reflecting plates from both NBIS units were found is 600 feet. This site was chosen for relative to be missing. However, the units had recorded ease of logistics and access. data, although erroneous. At this time, these failures are still under investigation. We believe TI will coordinate all aspects of the long- that their origins can be found quickly and corrected. term test, including the planning, integration, and Fogcorrection of the problems discovered testing of all subsystems, deployment and recovery during the. preliminary tests, two of each CMSS type operations, maintenance and servicing, and data will be installed as shown in Figure 7 for'a 3- reduction and analyses. Figure 12 illustrates the month extended test. In addition to the periodic data flow during the long-term test. monitoring of CMSS status during this test, obser- vations of fouling, corrosion, and general appear- ance of the CMSSs will be made. At the conclusion of the 3-month tower test, the performance of all CMSSs will be evaluated, and units with adequate performance potential will be modified for real- time data transfer and integration into the long- term current measurement system test. 313 4. Marsh-McBirney, Inc., Technical Manual for ARC 585 Adaptive Recording Current Meter, 1978. 5. Lawson, K. D., N. L. Brown, D. H. Johnson, and ],o (frC~yWR. A. Mattey, "A Three-Axis Acoustic Current Meter for Small Scale Turbulence," ISA, 1976, I~ ~~pp. 501-508. 6. Appell, G. F., "Performance Assessment of Advanced Ocean Current Sensors," NOAA T&EL unpublished presentation of the OCEANS 77 Conference, Los Angeles, California, 1977. 7. Appell, G. F., "A Review of the Performance of an Acoustic Current Meter," NOAA T&EL 9%4~C c Ji~i~ summary of laboratory tests on NBIS current meters, 1977. 8. Texas Instruments, Inc., "Long Term Testing of Acoustic Telemetry System, Test Report", Figure 12. CURRENT MEASUREMENT SYSTEM AT-SEA May 15, 1978, NASA Contract NAS 13-25. TEST-DATA FLOW 9. Okerlund, J. H., "Operational Current Measure- ment System, Final Report", June 20, 1977, 4. Conclusions Applied Physics Laboratory, University of Washington, NOAA Data Buoy Office Contract The development of an ocean current measure- 03-6-038-732. ment capability for the mixed layer and continental shelf regions of the oceans is not a straightforward endeavor, and requires considerable resources and a great deal of patience. The success of the program requires a good technical plan and strict adherence to sound technical disciplines and a methodical approach. Experience has shown that there are no shortcuts to an operational ocean current measurement capability. NDBO's approach to development testing and refining state-of-the-art current sensors, and to adapting them for system operation in the actual environment, together with the application of the principles stated herein, should accomplish the desired objective. This development approach is expected to provide reliable current sensors to measure and internally record ocean currents in a conventional manner; to provide reliable current sensors interfaced to a standard buoy data sub- system to deliver ocean current data via satellite to users in near real time; and to provide this information over an extended (relative to present systems) period of time. References 1. NOAA Office of Marine Technology, unpublished report of the Current Meter Review Panel, 1975. 2. NOAA Office of Ocean Engineering, "Minutes of the Working Conference on Current Measurements," Sponsored by NOAA OOE and the Delaware Sea Grant College Program, University of Delaware, Newark, 1978. 3. Cushing, V., "Electromagnetic Current Meter," Proceedings of the OCEANS 76 Conference, Washington, D.C., Sept. 13-15, 1976, pp. 25C-1 - 25C-17. 314 CHARTING OF OUTER CONTINENTAL SHELF SURFACE CURRENTS BY AERIAL TRACKING OF TRACERS Timothy L. Flynn David O. Cook Aero-Marine Surveys Raytheon Company, Oceanographic & P.O. Drawer 1230 Environmental Services New London, CT 06320 P.O. Box 360 Portsmouth, RI 02871 Abstract in estuaries and the nearshore area. Several investiga- As part of a physical oceanographic study of the New England tors2 3,4,5 have used time-lapse photography to docu- OCS sponsored by the U.S. Bureau of Land Management, techni- ment the motion of drogues or other surface current ques have been developed for acquisition of synoptic surface cur- followers in relation to fixed references such as the rents data over large areas. Dye-emitting Lagrangian tracers, shoreline or moored buoys This method was extended equipped with radio location beacons, are seeded in patterns and tracked by aircraft for several days. offshore using a vessel with known position as a refer- 6 The aircraft-deployed tracers consist of 132 cm-long cylinders, ence . Another aerial technique involves stereo photog- containing a fluorescent dye mixture, which are weighted to float raphy of floating targets to measure instantaneous surface vertically. Four color coded arms joined to the cylinder near its current velocities7 8. These previously-employed aerial top provide for unique visual identification, and frequency-audio tracking techniques have had limited offshore application coded VHF radio location beacons are mounted in the top of because of the lack of fixed reference points. Also, selected tracers. relocation difficulties caused by drogue or target dis- Tracer tracking is accomplished with a Cessna Skymaster air- craft with VHF homing equipment, vertically mounted Hasselblad persion have limited the duration of aerial Lagrangian cameras, a radiometer, and an advanced Loran C unit. The photo- studies to a tidal cycle. graphic tracer data is subsequently digitized in the lab and computer-processed into tabular and graphical displays. This paper describes a new approach to aerial measurement of Lagrangian currents in the offshore In a typical study, between 60 and 180 tracers are deployed in several patterns dispersed over an area of approximately 25,000 region. Development of this approach was prompted by km2. A cargo aircraft such as a Beechcraft G-1 8 is used for tracer the U. S. Bureau of Land Management' s need to acquire deployment, and relocation of tracers takes place with the Sky- data on surface currents on the New England outer con- master for several days thereafter. Results to date have yielded tinental shelf in connection with potential oil and gas interesting data on regional circulation, diffusion, and circulation development there. The study area, Georges Bank, within a warm-core eddy. encompasses over 50,000 km2. Synoptic surface current data within this area were required over a period of several days in each season; spatial scales of interest A knowledge of surface currents is critical to pre- were regional (several hundred km), local (20 to 50 km), dieting potential environmental impacts from offshore and small scale (2 to 5 km). Special interest was activities such as oil development and waste disposal. directed at frontal zones on the northern and particularly Floatable wastes and most spilled oil remain at the water the southern flanks of Georges Bank. surface where their transport is controlled by surface To meet this requirement, a new surface current currents and direct wind stress. Eulerian data can not tracer was developed and a Cessna skymaster aircraft be readily extrapolated to the surface because flow at was outfitted with an advanced Loran C unit for position- was outfitted with an advanced Loran C unit for position- these two levels is frequently decoupled. Lagrangian ing, aerial cameras, a radiometer, and recording equip- ing, aerial cameras, a radiometer, and recording equip- methods represent the only viable method for measuring ment. Up to 180 tracers have been deployed over the currents in the upper few meters of the water column study area with tracking continuing for several days. study area with tracking continuing for several days. Several Lagrangian techniques are available for study- Relocation of the dye-emitting tracers is facilitated by ing surface currents. Drift bottles/cards represent the equipping selected units with radio beacons. Data is simplest approach, but provide only indirect data on flow processed using specially-developed computer software. trajectories. Shallow-drogued buoys have traditionally This new approach to the measurement of surface been tracked by ships and more recently by satelliterr Logistics and expense tend to limit the number of buoys design, survey aircraft, field logistics, and data reduc- which can be simultaneously tracked. tion in the remainder of this paper. tion in the remainder of this paper. Aerial photography has provided a means of tracking large numbers of surface current markers concurrently 315 2. Tracer Design rotochute during air deployment and are color-coded to pro- vide for unique visual tracer identification in the water. The Lagrangian surface current tracers required for These arms, which are 20 cmwide and 60 cm long, lie a few aerial tracking on the outer continental shelf had to sat- cm below the water surface after the tracer is deployed. isfy the following criteria: Within the outer tube is an inner tube filled with a (1) They mustrespond to water motion inthe surface mixture of fluorescein dye, polyethylene glycol, and layer, a s defined physicallyeand biologically, and oleic acid. Wave-induced vertical motion of the tracer and a check valve at the tube bottom cause water to be (2) They have to be visible and uniquely identifiable forced upward in the space between the inner and outer in aerial photographs taken from altitudes of 150 tubes. Rising water entrains the fluorescein dye through to 600 m. holes in the inner tube, and this bright green dye is discharged through vent holes near the water surface. (3) They must be relocatable after intervals of 1 to With moderate sea conditions, the dye supply will last three days. To achieve a uniform rate of dye dissolu- (4) They have to be air-deployable, and sufficiently tion, the ratio of dye to oleic acid must be adjusted for rugged to remain intact in the water for at least seasonally-varying water temperatures. The purpose of two weeks. the green fluorescein dye is to assist in visual relocation of the tracers. Existing Lagrangian devices did not meet all of the above criteria, and this led to development of a new A VHF homing beacon placed within the exposed tracer. The development work was undertaken in con- portion of the tube in selected units is the most impor- junction with Eotech Corp. (Ft. Lauderdale, Florida) tant aid to tracer relocation. These beacons are similar and involved several increments of field testing off to those which serve as downed aircraft locators, and for Florida and in New England waters. this study we have been assigned six frequencies each of which have up to six audio codes. Depending on battery The first consideration in the design process was The first consideration in the surcdelsign process wan as type, the beacons will operate up to 2 weeks and can be definition of the surface layer. Physical oceanographers aerially-detected from 30 to 80 km distance. often consider surface waters to lie above the thermo- cline. The surface layer can be more strictly defined Susceptibility of the tracers to transport by wind as the upper few meters of the water column which is acting directly on the portion exposed to air was inves- susceptible to wind forcing and diurnal heating and tigated in a series of field tests. In each test, displace- cooling. However, recent field studies9 have demon- ment of a tracer from a "zero leeway marker" -a dye- strated the existence of intense shear in the upper meter emitting device floating in the upper 30 cm of the water of the water column which affects oil and other pollutants column with minimal exposure at the surface-was borne at the water surface. These studies identify documented as a function of time by aerial photography. the surface layer as a three-dimensional body whose The rate of displacement varied from 2 to 4 cm/s with a thickness ranges from centimeters to meters depending wind speed varying between 5 and 8 m/s. This downwind on environmental conditions. This thickness decreases motion relative to the U. S. Coast Guard "zero leeway with increasing stability of the water column; conversely, marker" was less than one percent of the wind speed. the thickness increases with homogenization of tempera- During the Georges Bank studies, it has been ture and salinity by mixing. observed that the tracers move downwind faster than dye From a biological standpoint, the upper few cm to patches diffused in the upper few meters of the water 1 m of the water column corresponds to the neustonic column but slower than aluminum powder at the water layer. This zone of intense sunlight is selectively inhab- surface. These observations coupled with test results ited by various species, such as meroplankton, at certain lead us to conclude that the tracers effectively reflect stages of their life cycles and by zooplankters on a diur- motion of the upper meter of the water column, as nal basis. intended. In view of these physical and biological considera- 3. Data Collection Platform tions, it was decided that the Lagrangian tracer should respond to currents in the upper meter of the water Aircraft column. The aircraft used for tracking and relocating sur- Thetracerdevelopedforthe surface current study (Fig- face current tracers is a Cessna Skymaster model ure 1) consists of a 132 cm long, 9 cm diameter polyethylene 337-G (Figure 2). It is a high wing monoplane with one tube with a flotation collar near the top with four radiating 210 hp Continental engine on the forward end of the flatpolyethylenearms. Thetubeisballastedtofloatverti- fuselage and another on the aft end. The tail is a twin cally similar toa sparbuoy, withthe upper 20 cm exposed boom type with dual rudders connected by a single hori- above the water surface. The four arms serve as a zontal stabilizer and elevator. 316 The aircraft is the civilian equivalent of the military Apollo series of manned space flights. The MK-70 has a 70 02 observation and light attack aircraft. It has a gross mm distortion-free lens anda reseau plate with a 5x5 grid weight of 2100 kg, a cruise speed of 313 km/h and a calibratedtowithin several microns. Bulk film magazines range of approximately 2224 km with reserves. It is have beenfittedto the cameras to allow loading a standard 30 equipped with full civilian instrumentation for instrument m roll of uncut film. This provides up to 500 exposures with- flight in bad weather including dual VHF communications, out reloading. dual VOR navigation receivers, distance measuring equipment, a marker beacon receiver, an ILS receiver, Radiometer a radar altimeter, multi-mode radar, a flight director, A Barnes PRT-5S radiometer is employed to sense a horizontal situation indicator, and a fully coupled auto- sea surface temperature beneath the aircraft. Incoming pilot. It also has dual head automatic direction finders, infrared radiation in the 9.5 to 11.5 micron range is and for tracer homing, it has VHF direction finder units. compared with an internal reference of known tempera- The main camera hatch is located under the right ture, yielding an output of surface water temperature forward seat and contains two 70 mm cameras and a versus time. Data has been logged on a strip chart radiometer. Aft on the port side is a periscope which recorder annotated with position information. affords forward as well as vertical visibility for the Data Panel photo-operator. A strip chart recorder for radiometric data is located amidships. Behind the seats in the aft Flight data has, for the first year of the program, end of the cabin are a photo data panel, life raft, survival been recorded on a photographic data panel originally kit, and life preservers. built for Dr. W. Richardson at Nova University. The panel resembles an aircraft instrument panel and con- The aircraft has accommodations for a crew of three or tains a gyro-horizon for vertical reference, a direction four, dependingonthe season, inthe survey configuration. gyro, a wet compass, a magnesyn compass, a pressure Normal crewwouldbe a pilot and navigator in the front and altimeter, a dual Loran C display, a 24 hour sweep a photo-operator and scientist-observer (during colder second hand clock, and a digital second counter. Dial months) intherear. Duetotheengineandwingconfigura- faces are photographed by a standard Hasselblad 500-EL tion, each crew positionhas excellentvisibility tothe sides camera mounted in the unit and synchronized with the and downward. two vertical Hasselblad MK-70 cameras. The data panel is currently being replaced with a digital data logger Navigation System which will also accept the radiometer input and record The offshore positioning system is a critical element all data on magnetic tape cassettes. to the experimental design of the surface currents pro- gram and much effort has been devoted to obtaining the 4. Logistics best possible performance. During the first year pro- Study Design gram, a Simrad Model 123 Loran C system specially programmed for airborne use was utilized. This unit The surface current field studies have been executed tracks up to four rates concurrently and provides a in two modes relating to the time scales of the trans- visual update once per second of any two of these rates. ports under investigation. Of primary interest is the A coordinate convertor displays and updates the equiva- regional circulation pattern of the overall study area. lent latitude and longitude. Because the scale of non-tidal motion, such as wind ; The first year program had numerous instances of drift, is usually greater than that of tidal motion, estab- lishing the positions of Lagrangian tracers once a day temporary Loran C outages either due to weather or precipitation static build-up on the aircraft. Occasional is generally sufficient to resolve regional circulation over periods of several days. By contrast, small scale 10 microsecond jumps have also occurred when the air- over periods of seve ral days. By contrast, small scale circulation in the vicinity of frontal zones requires craft is maneuvering; if properly documented, these can higher frequency sampling of tracer movement. In the be corrected. Much of this problem relates to the poor p res ent study a rea, such processes as convergence, existing Loran C coverage of the Georges Banks area, shearing, and Langmuir circulation require observa- a situation which will improve considerably this summerr ation w a E C ta. Itions of tracer locations at 15 to 30 minute intervals. when a new East Coast net goes into operation. In the absence of outages of jumps, accuracy of the Loran C Regional circulation has been addressed with large positions is rated at � 100 m which is considerably better scale surface current studies. These involve seeding than Loran A or Omega. large numbers of tracers-between 60 and 180 have been used-along a series of patterns distributed around Aerial Cameras Aerial Cameras Georges Bank. A total of 10 to 12 line and occasionally Hasselblad MK-70 photogrammetric cameras are "L" shaped patterns, lying mostly at the margins of the used to determine tracer positions. These 70 mm for- Bank and normal to isobaths, have been used to investi- mat cameras were originally developed for NASA for the gate the seasonal surface currents climatology. The 317 patterns are 20 to 30 km long and have tracers spaced every A time budget is established prior to each day of fewkm. Occasionalclosely-spaced"clusters"oftracers tracer relocating which allows all patterns to be visited. are deployed along key lines to study dispersion. During a The time at each pattern is then used to navigate from season, thepatternsareoccupiedintwogroups inback-to- radio beacon to radio beacon, with visual searching con- backlarge scale studies focusing onthe easternandwestern fined to the cluster locations. Excessive dispersion halves of the Bank. A typical season's deployment scheme normally results in abandonment of cluster searching is shown on Figure 3. Tracking of the deployedtracers then after the third day. The relocation rate for beaconed takes place dailyfor severaldays. These large scale stud- clusters is quite high-80 to 90 percent-on the first two ies constitute the core of the program. days of relocating, and then begins to drop off as the beacons weaken and tracers disperse. The tracer Small scale studies completed in a single day have location overflights are long, with individual flights been used to examine local circulation near shear zones, irregular topography, and other areas of interest. One to two short lines 2 to 4 km long are seeded with tracers, The field operations are scheduled for fair weather dye, and aluminum powder every 100 to 300 m. Positions periods, although weather encountered over Georges of these Lagrangian devices are monitored every 15 to 30 Bank has included squalls, rain, snow, and high wind. minutes. Poor weather in the form of precipitation or low clouds is unfavorable to a mission from several standpoints. Radiometry and photography are precluded, the Loran C Deployment of tracers during the large scale studies signal may be degraded, and the flight schedule must be involves two aircraft: the Skymaster survey airplane more conservative because of safety factors. Operating which navigates to the proper locations and obtains initial experience has led to fairly accurate forecasting of positions for tracers, and a larger cargo aircraft which weather over the study area and aborted missions have actually deploys tracers. A Beechcraft Model G-18 been kept to a minimum. cargo aircraft has been used quite successfully for deployments-it can accommodate large numbers of 5. Data Processing tracers, has a speed and maneuverability compatible Following the field data acquisition phase of a sur- with the Skymaster, and has the necessary range to com- face currents study, films from the data camera and plete the usual 1100 to 1300 km deployment track. Once aerial cameras are developed. Aircraft flight tracks are the two aircraft navigate to the start of a deployment plotted for each day of operation, and any necessary pattern, they fly down the line in formation with the corrections to the Loran C rates are implemented. Beechcraft at 225 m and the Skymaster following at Loran C positions associated with each tracer sighting 380 m. The Skymaster directs the Beechcraft to deploy are then plotted and converted to latitude-longitude tracers when critical Loran C intercepts are reached. The coordinates. These coordinates correspond to a point physical act of tracer deployment is performed by two on the aerial photographs whose precise position depends technicians stationed at the open rear door of the Beech- on upon aircraft tilt and other factors. Tracers appearing craft. Tracer "clusters" were initially deployed by on each frame of aerial film are digitized with respect to throwing out nine tracers in rapid succession; the present selected reseau marks, and these digitized locations approach is to wrap these in bundles of four and five using along with tracer identification, aircraft orientation water soluble polyvinyl alcohol so the subgroups begin (roll, pitch, heading, altitude), time, and latitude- dispersing from point sources. The Skymaster following longitude coordinates for each frame are placed on longitude coordinates for each frame are placed on the Beechcraft obtains initial positions on the deployed punched paper tape. tracers and observes that radio beacons are performing properly. The tape is processed on a CDC CYBER 174 com- puter which uses photogrammetric equations to calculate Tracer relocation overflights are normally scheduled the latitude and longitude of each tracer photographed. for days 2, 3, 5, 7 or days 2, 4, 6, 8 of a large scale study, weather pemitig.TThe accuracy of the calculated tracer positions relates study, weather permitting. The program has evolved to several potential error sources. These sources, away from large numbers of unbeaconed tracers, which along with the uncertainty in tracer position associated are increasingly difficult to relocate as time passes and with them, include the following: 1) resolution of air- originally-straight lines become distorted, towardsfewer craft tilt indicator + 13 m: 2) resolution of aircraft numbers of beaconed tracers. Logistics are constrained altitude indicator, 15 m- 3) camera film distortion, by available daylight and fuel reserves, and time is + 3 m; 4) lack of synchronization between triggering of simply not available for extensive visual searching. The the cameras and Loran C updating, + 26 m; and 5) Loran radio beacons can be sensed from 30 to 80 km away C accuracy, � 100 m. The total errorl0 of + 105 m is depending on aircraft altitude and, with practice, the dominated by accuracy of the Loran C navigation system. pilot can navigate the aircraft directly over them. At This degree of accuracy is acceptable in view of the present, unbeaconed tracers are reservedforthe clusters. 318 normal 5 to 30 km transport exhibited by tracers between 6. Conclusions daily sightings. Techniques have been developed to enable collection Outputs of the computer processing include the of synoptic data on surface currents in offshore regions following data products: 1) a listing of tracer identifica- having areas of many thousand square kilometers. These tion, position, time, date, photographic frame, and techniques include use of a new Lagranglan tracer and of extrapolated speed since the last sighting of the tracer specialized aerial deployment and tracking method- in question; 2) a map showing successive tracer locations ologies. Results can be used to define a seasonal sur- connected by straight lines and annotated with data on face current climatology and investigate transport tracer identification, time, date and extrapolated speed; processes in eddies. The surface current capability and 3) a map showing the area encompassed within each constitutes a significant tool for assessing potential aerial photograph to aid in constructing photomosaics impacts of offshore activities such as oil development (applicable where individual photographs are overlapping). and ocean dumping. The radiometer data is digitized and computer-pro- cessed separately, producing a listing of time and tem- perature and a map showing the aircraft track line The surface current technology described in this annotated with observed water temperatures. paper was developed under U. S. Bureau of Land Manage- ment contract AA550-CT6-53 with the Raytheon Company. The type of surface current data which can be The field program is being executed by Aero-Marine acquired by aerial tracking of the Lagrangian tracers is Surveys, subcontractor to Raytheon. The LaGrangian illustrated on Figures 4 and 5. Figure 4 shows observed tracers were developed in conjunction with Eotech Inc. tracer transport in the Georges Bank vicinity during who then fabricated production units. Many persons companion studies in the winter of 1978. The first study, covering the eastern half of the Bank, extended from covering the eastern half of the Bank, extended from particular, we acknowledge the efforts of Kenneth Berger Jan. 31 to Feb. 3; the second study which concentrated B (BLM), Ted Crosby, Gary Davis, Dan Frye, Barbara on the western half lasted from Feb. 20 to 25. Succes- (Grnbe) Flynn, William Kielhorn, John McMicken, sive positions of individual tracers are connected with James Richardson, Harry Williams, and Priya straight lines. Wickramaratne. This figure illustrates the substantial area which can be synoptically studied using the methods described in References this paper. The tracer transport reflects two factors: 1. Cromwell, T., Montgomery, R., and Stroup, E., a wind-induced surface current caused by moderate west "Equatorial undercurrent in Pacific Ocean revealed to northwest winds occuring during both studies which is by new methods", Science, Vol. 119, 1954, pp. superimposed upon a topographically-controlled subsur- 648-649. face circulation pattern. The subsurface circulation is directed clockwise around the north, east and southern 2. Duxbury, A. C., "Currents at the mouth of the flanks of Georges Bank. Tracer paths which are in Columbia River", Photogrammetric Engineering, apparent conflict in the southwestern portion of the Bank Vol. 33, 1967, pp. 305-311. actually reflect the presence of a warm water filament. 3. Cook, D. O., "Evaluation of estuarine circulation by This filament was observed both by aerial radiometry aerial photography of current drogues", Preprints, and by hydrographic measurements from a concurrent 7th Annual Marine Technology Society Conference, cruise. 1971, pp. 663-670. Figure 5 illustrates the Lagrangian transport Figure 5 illustrates the Lagrangian transport 4. Assaf, G., Gerard, R., and Gordon, A. L., "Some observed during November 15 to 20, 1977 when a warm 4. Ass af, G., Gerard, R., and Gordon, A. L., "Some mechanisms of oceanic mixing revealed in aerial core Gulf Stream eddy was impinging on the southern photographs", Journal of Geophysical Research, flank of Georges Bank. Note the marked clockwise Vol. 76, 1971, pp. 6550-6572. rotation of tracers within the eddy; these moved as much as 90 km in a single day. Those tracers at the eddy 5. Klemas, V., Davis, G., and Wang, H., "A cost- margin and beyond exhibited much lower rates of trans- effective satellite-aircraft-drogue approach for port. The airborne radiometry measurements made in studying estuarine circulation and shelf waste dis- conjunction with the tracer tracking provided valuable persion", IEEE Ocean '75 Conference Proceedings, information on daily changes in eddy position. 1975, pp. 751-760. The surface current data exemplified by Figures 4 6. Ramey, E. H., "Measurement of ocean currents by and 5 has obvious implications from a scientific stand- photogrammetric methods", U. S. Coast and Geo- point as well as for the prediction of the transport of detic Survey Technical Memorandum 5, 1968, 18 pp. spilled oil and other pollutants. 319 7. Cameron, H. L., "The measurement of water current 9. Gordon, A. L. and Gerard, R. D., '"Wind drift sur- velocities by parallax methods", Photogrammetric face currents and spread of contaminants in shelf Engineering, Vol. 18, 1952, pp. 99-104. waters", U.S. Coast Guard Office of Research and Development Center, Report No. CG-D-5-75, 1973, 8. Keller, M., "Tidal current surveys by photogram- Development Center, Report No. CG-D-5-75, 1973, 64 pp. metric methods", U.S. Coast and Geodetic Survey Technical Bulletin 22, 1963, 20 pp. 10. Fisher, A., "The Mathematical theory of probabil- ities", McMillan Co., New York, Vol. 1, 1915, p. 106. F~~~~~~~~~~~~~~~~ .a', _ -RADIO BEACON - -FLOATATIOR COLLAR a- A2~ -IEENTIFICATION ARMS ,.). -DYE CANNISTER . C i FOOT VALE\ Figure 1. a - side view of the Lagrangian tracer; b - appearance of the tracer after deployment. 70� 0 6 89 � 67� 66� L ~~~o 42' L --- * 41' O 00 7. m o p e~~~~I _60- -JANWI I) I197v LIFE RAFT CAMERA BAY .0519 Figure 2. The Cessna Skymaster C-337 survey aircraft, Figure 3. Deployment locations of beaconed tracers in N700AM, owned by Aero-Marine Surveys, Inc. the Georges Bank study area during the winter 1978 large scale surface current studies. Not shown are clusters of eight unbeaconed tracers deployed around four of the beaconed units during each study. . . ~~\ Figure 4. Observed motion of surface current tracers Figure 5. Observed motion of surface current tracers during the winter 1978 large scale studies. Positions in the vicinity of warm core eddy Q during the November where tracers were observed are connected by straight 1977 large scale study. Observed tracer positions are lines. connected by straight lines. 320 AN UNDERWATER TOWED ELECTROMAGNETIC SOURCE FOR GEOPHYSICAL EXPLORATION B. E. Tossman D. L. Thayer The Johns Hopkins University The Johns Hopkins University Applied Physics Laboratory Applied Physics Laboratory Johns Hopkins Road Johns Hopkins Road LueMaryland 20810 Laurel, Maryland 20810 W. A. Swartz The Johns Hopkins University Applied Physics Laboratory Johns Hopkins Road Laurel, Maryland 20810 Abstract The magnetic dipoles are produced by 6 ft diameter multi-turn loops built into Low frequency electromagnetic aft vertical and horizontal stabilizing methods are used in geophysical explor- surfaces. The electric dipole is pro- ation by detecting the magnetic field duced by driving nose and tail electrodes distortion between a transmitter and with low frequency current. The UCS also receiver produced by locally conductive contains an instrumentation cylinder bodies. Both ground and airborne sys- which includes vehicle altitude and depth tems are in current use. One airborne sensors, tracking transducer, and relays multi-coil aerial electromagnetic pros- for directing the input power to either pecting system houses the transmitter the magnetic or electric dipoles or a and receiver in a 30 ft bird while high-power dissipation non-field producing another reduces the system to an imped- dummy load. ance measurement using a single super- conducting coil. Hydrodynamic towing requirements for this electromagnetic source were that it It is possible to extend low fre- be stable in roll, pitch, and yaw to quency airborne geologic prospecting better than -*5 degrees oscillatory motion techniques into the ocean environment with average values within 5 degrees of by employing an underwater towed source straight and level flight. Position was of electromagnetic radiation and a to be stable to within *-5 ft in both receiving magnetic or electric field horizontal and vertical directions. detector. The receiver can be simul- These hydrodynamic characteristics were taneously towed on an auxiliary cable, achieved passively with the body being boom mounted on the towing platform towed at depths up to 500 ft and speeds (surface or subsurface), or land or of 5 to 7 knots. In addition, the body underwater based. was to be constructed of mostly nonmetal- An underwater towed body suitableliannomgeccmpets for geophysical exploration applications This paper concentrates on the design has been constructed and employed in details of the UCS including its mechani- underwater propagation studies of low cal, hydrodynamic, and magnetic design frequency electromagnetic fields. characteristics and the electrical system * ~~Called the Underwater Calibrated Source consisting of the high current drive (UCS), it is 13 ft long weighing 840 power system and shipboard monitoring pounds in air with a capability of pro- system for attitude and depth detectors. ducing magnetic dipoles in vertical and horizontal directions, and a horizontal electric dipole. Powered by a current feedback controlled high-power modified sonar amplifier, the UCS can produce 9700 amp-turn-meter2 of magnetic dipole or 200 amp-meters of electric dipole at 50 amps at frequencies up to 200 Hz without significant attenuation from coil inductance. 321 UNDERSEA STEREOSCOPIC TELEVISION R. B. Fugitt and R. S. Acks REMOTE OCEAN SYSTEMS, INC. P.O. Box 26626 San Diego, California 92126 Abstract be particularly helpful when used with remotely operated manipulators or other related work The enhanced visual depth perception pro- systems (Ref 4, 5), and for greater control vided by stereoscopic television offers during precise maneuvering of undersea vehicles. improved efficiency for many undersea tasks Extensive testing performed at the Naval Ocean involving inspection, positioning, manipu- Systems Center has shown significant improve- lation, or remote vehicle control. Several ment in terms of accuracy and reduced task techniques exist for implementation of stereo- times when stereo television was used for con- scopic television. A two-camera system with a trol of underwater manipulators (Ref 6, 7). polarization coded display has been developed The purpose of this report is to discuss for undersea use. A video multiplexing tech- the theory and alternative implementations of nique is employed allowing signal transmission stereoscopic television and to describe a on a single coaxial cable and stereo documen- system which has been developed for offshore tation using a standard single-channel video applications. recorder. Tests of the undersea stereoscopic system have demonstrated its effectiveness in providing accurate, realistic depth perception 2. Methods of Depth Perception and a sense of spatial awareness not possible with conventional closed circuit undersea There are many mechanisms by which visual television systems. "depth" or object distance can be perceived when using only one eye or camera. The most important of these are: 1. I n t r od uction ~~~1. Occlusion of objects by those located One of the primary limitations of pre- nearer to the observer. sently available undersea television systems 2. Subtended angle of objects of known is lack of display realism. Experienced size. operators of manned submersibles almost in- 3. Relative sizes of objects of known variably prefer direct vision through view- size. ports to the use of closed circuit television. 4. Nearness to the horizon of objects, This preference exists because underwater e.g., an airplane in the sky or boat television systems generally provide no color on the ocean. information, they often suffer from poor 5. Position with respect to objects of contrast and resolution and they provide known location, e.g., a book located little or no visual depth perception and on top of a desk. orientation information. 6. Focusing of the eye or camera. The technology exists to alleviate these 7. Apparent motion of stationary objects shortcomings, but it has been applied only to with respect to a moving observer. a limited extent, partly due to cost and S. Color degradation and loss of resolu- system complexity. Backscatter reduction tion with distance. techniques exist for improving underwater 9. Shadow relationships with respect to visibility, and the use of head-coupled stereo- artificial lighting. scopic television has been demonstrated in the 10. Rate of angular displacement of ob- laboratory and on remotely controlled undersea jects with known velocity. vehicles (Ref 1, 2, 3). At present, color 11. Loss of perspective with distance. television is available for undersea use and is being utilized in an increasing number of Unfortunately, when working in the under- applications. Although desirable for any appli- water environment most of these depth cues are cation from a human factors standpoint, the lost. The distance at which objects can be greater realism and resultant sense of "opera- seen is severely restricted, and no horizon is tor presence" provided by such improvements will apparent. In general, the operator of a sub- 322 mersible vehicle or work system will be dealing general, at an angle between 6 and with objects that are unfamiliar in size and 12 degrees. perspective. In most cases, therefore, the These rules in effect state that the geo- only available single-eye or single-camera cues metry of the cameras and the viewer should be will be occlusion, focus, color degradation, such that the observer has the illusion of shadow relationships and apparent motion. In actually being present at the camera location. many cases, some or all of these remaining cues Rules 1 and 2 are absolutely necessary for may also be absent or may be too subtle to be superposition of the two images. Rules 3 and of any use. 4 can be varied under certain conditions to en- Stereoscopy is one of the most powerful hance the stereoscopic illustion by distorting depth cues, but it requires the use of two the scene's dimensions. eyes or two camera images. This mechanism of Variation from rule 5 results only in an depth perception is discussed in greater apparent scale change which is equal for height, detail in the following section. width and depth dimensions. Consider, for ex- ample, a situation in which two large sharks are being observed at 30 and 40 feet, respectively. 3. The Mechanism of Stereoscopy If the station points are separated by the standard 2-1/2 inches, very little parallax Because our left and right eyes are spaced between the sharks will be noted. If, however, approximately 2-1/2 inches apart, each receives the stereogram is based on a 25-inch effective a slightly different image. The amount by which interocular spacing, the visual angle subtended the optical axes of the eyes of any observer by each shark will be unchanged; however, paral- converge to superpose the right and left images lax will be greatly increased, allowing the of any object is a function of the distance from relative distances to be judged more easily. the observer to that object. Parallax between The pair of sharks in effect appears stereo- objects at different distances is instinctively scopically to be at a distance of 3 and 4 feet, interpreted as an indication of their relative respectively, and reduced to 1/10 actual size. distances. This is the mechanism of stereo- This technique is often used with long-range, scopic vision which literally interpreted aerial, stereoscopic photography for enhance- means "seeing solid." ment of terrain features. Stereoscopy is the technique or art of The intersection of the camera axes is creating an illusion of stereoscopic presence by necessary to create the illusion that the scene means of a stereogram, which is actually a set is being viewed through a window or frame locat- of two images, one provided for each of the ed between the observer and the scene. Just as observer's eyes. Every artist and photographer every two-dimensional photograph, painting or knows a painting or picture has a particular television image has a border, so must a stereo- "station point" from which it should be observed, gram. This effect is very noticeable in View- In a stereogram, two such station points must be master stereoscopic slides. The border actual- simulated. ly has a depth effect and seems to float in rTo faithfully conserve the stereoscopic space a few feet before the observer's eyes. illusion requires attention to the following The frame contributes nothing to the illusion of geometric considerations. depth, but does make it easier to rationalize the stereoscopic illusion with the limited 1. Each image must be properly oriented fleld-of-view. The stereoscopic frame, if the about the optical axis of the observer's illusion is properly handled, becomes an eye so horizontal lines appear to be immediate reference when an observer first looks horizontal. into the viewer and helps him focus on the scene. 2. The vertical position of both images If handled improperly, however, the stereoscopic must be identical, e.g., the horizon distance to this illusionary window is greater images should superpose in the eyes than that to the observed objects. This is a of the observer, disturbing illusion, and it is difficult to 3. Each object in each view must sub- rationalize why objects disappear from view when tend the same angle to the observer's they cross the border of the frame. Since the eye as it would if viewed from the borders of the separate images are used to form appropriate station point. the stereoscopic frame, the amount of eye con- 4. The horizontal spacing of the two vergence necessary to superpose the borders images of a distant object should be should be equal to the convergence angle of the such that an eye convergence angle of camera's axes. With a 2-1/2-inch camera sepa- 0 degrees is required to superpose ration, the apparent distance from the observer them, i.e., infinity should appear at to this frame would then be equal to the dis- infinity. tance from the cameras to the intersection point 5. Station point or camera separation of their axes. In practice, an apparent dis- should be about 2-1/2 inches to tance of 12 to 24 inches to the stereoscopic approximate the average interocular frame is both comfortable to the observer and is spacing. usually within the range of the nearest objects 6. The camera's optical axes should con- to be observed. This requires a camera conver- verge to a distance closer than the gence angle between 6 and 12 degrees. The set of nearest object of interest and, in rules proposed in this section describe how the 323 mechanism of stereoscopy is most effectively re- (or prisms) are used to superimpose the views produced in artificial form. This can be summar- and form the image at a comfortable viewing ized by the following statement: The geometry of distance from the observer. the camera(s) and viewer should provide the ob- A single camera underwater system of this server with an undistorted illusion of stereo- type is described more fully in Reference 8. scopic presence at the camera's location. Although underwater tests showed significant reduction in task time and operator errors, the system suffered from off-axis image distortion 4. Implementation of Stereoscopic Television and a narrow horizontal field of view (17 degrees). The general principles discussed in the pre- An alternative approach is to use two camn- vious section can be employed in a wide variety eras which readily provide larger fields of of stereoscopic television systems, each with its view and higher resolution. Figure 2 shows a own applications, advantages, and disadvantages, display technique using two side-by-side tele- Several different techniques can be used for sepa- vision monitors. A preferable approach which ration of the stereoscopic images. Among then are allows freedom of operator movement is shown in sequential switching (electronically or mech- Figure 3. Since no viewing hood or electrical anically) between views, color filter and po- connections to the glasses are required, the larization techniques, and optical separation observer can easily divert his attention to and superposition using lenses, prisms, and other controls or displays, and can move freely mirrors. about when required. An undersea stereoscopic A simple approach which requires only one television system based on this technique has camera and one monitor is shown in Figure 1. been developed, and is described more completely in the following section. OBJECT ~~K~~OBJECT TEVISION MIRRORSMOIR ALENS LEFT RIGHT CAMERA CAMERA TV IMAGE TUBE TELEVISION MONITOR LEFT RIGHT MONITOR J MONITOR OBSERVERSTEREOSCOPIC VIEWER Figure 1. Single Camera Stereoscopic TelevisionPRS A system of mirrors placed in front of an ASSEMBLY undersea camera form a side by side stereo- OBSERVER scopic image pair on the television monitor. This image pair is then viewed with an optical system which presents the appropriate image to each eye of the observer. Lenses and mirrors Figure 2. Two Monitor Periscope Display 324 tape recorder. To re-construct the stereo POLARIZING ~image, the signal is re-played and viewed using LEFT P O LTERIZN the stereoscopic display. MON~~~~ILTER To accomodate stereo television on systems MONITOR ~~~~~~~~presently limited to one camera due to cable, penetrator, or data channel constraints, an - ~~~~~~~~undersea power and video multiplexing unit was also developed. With the addition, both video ___________________channels are sent over a single coaxial cable or data channel. The undersea multiplex unit SEMI- IRIGHT is small in size (2-3/4 inches OD by 9-1/2 TRANSPARENT MONITOR inches long) .and can easily be used by divers MIRROR or on small undersea vehicles. IMAGE REVERSED - - ~~~~~~~~6. Applications Many undersea and offshore applications of conventional closed circuit television could benefit from the depth perception and sense of * - ___________spatial awareness which is obtained with stereo- scopic television. These include: POLARIZATION ANALYZER 1. Positioning tasks using manipulators or 21... other work systems. 2. Precise control of remotely manned vehicles. OBSERVER 3. Inspection and video tape documentation (diver or vehicle). Figure 3. Polarization Coded Stereoscopic Display 4. Enhanced optical search and detection. 5. Real time and recorded stereoscopic bottom mapping and site surveys. 5. Undersea Stereoscopic Television System 6. Subsea equipment positioning in dril- ling and production operations. The basic stereoscopic television system 7. Mating of structures in offshore consists of two underwater television cameras, construction a camera mounting fixture, the stereo display 8. Internal pipeline inspection. and viewing glasses. 9. Shipboard or platform monitoring For most applications the cameras are sep- tasks (reels, etc.) * ~~~arated by approximately 2-1/2 inches to provide realistic depth perception. With appropriate Figure 4 illustrates the display for a selection of the camera and monitor fields of two-camera stereoscopic system manufactured by view, no camera convergence or separation ad- Sub Sea Systems, Inc. This system has been * ~~~juatment is required to view objects ranging in distance from approximately one foot away to * ~~~infinity. For other special applications such as long range object detection, sea floor stereo * ~~~mapping or drill pipe re-entry, an appropriate wide baseline between cameras can be used for exaggerated depth perception. The stereoscopic display utilizes a geo- metry similar to that shown in Figure 3. The * ~~~images are linearly polarized and superimposed using a semi-transparent mirror. An angle- * ~~~dependent light attenuating film placed in front of the side monitor allows the observer to see only the superimposed stereo pair. The observation glasses use polarization filters oriented such that the appropriate image is presented to each eye of the observer. Since separate monitors are used for each camera, differences in the camera images (size, distortion, position, etc.) can be adjusted at * ~~~the monitors to provide a matched, undistorted stereo image pair. For stereoscopic documentation, video multiplexing circuitry was developed which allows both right and left camera signals to be recorded on a standard single-channel video Figure 4. Stereoscopic Display 325 successfully employed for many of the above stereoscopic television can significantly im- applications. At present, it is being used prove system effectiveness by providing opera- in a diver-held version and on the remotely tors with an accurate sense of depth perception. controlled vehicles TELESUB 1000 (Remote Ocean This three dimensional awareness of objects Systems Inc.) (Fig. 5) and Seabug I (UDI and their positions relative to the remote cam- Operations, Ltd.) for pipeline inspection in eras is particularly useful in situations where the Gulf of Mexico. accurate manipulation or precise vehicle con- trol is required. References 1. Fugitt, R. Bruce, "Advanced Television Sys- tems for Undersea Use," 1971 IEEE Confer- ence on Engineering in the Ocean Environ- ment, Sept 1971. 2. Fugitt, R. Bruce, "Small Remotely-Manned Vehicles," 1973 IEEE International Confer- ence on Engineering in the Ocean Environ- ment, Sept 1973. 3. Fugitt, R. Bruce, "Design and Operation of Two Remotely Manned Undersea Vehicles," IEEE and Marine Technology Society Conver- ence on Engineering in the Ocean Environ- ment, Sept 1975. 4. Uhrich, Richard W., "Terminus Controlled Deep Ocean Manipulator," 1973 IEEE Confer- ence on Engineering in the Ocean Environ- ment, Sept 1973. 5. Uhrich, Richard W., "Manipulator Develop- ment at the Naval Undersea Center," Naval Undersea Center, San Diego, CA, NUC TP 553, January 1977. 6. Pepper, Ross L., Robert E. Cole and David C. Smith, "Operator Performance Using Con- ventional or Stereo Video Displays," 1977 Proceedings of the Society of Photo-Optical Figure 5. TELESUB 1000 Remotely Instrumentation Engineers, August 1977. Controlled Vehicle 7. Uhrich, R. W. and R. Bruce Fugitt, "Inves- tigation of Control and Viewing Techniques for use with Remote Undersea Manipulators," 7. Conclusions Naval Ocean Systems Center, San Diego, CA NOSC TR 233, 1 May 1978. Many present and future undersea vehicles 8. Fugitt, R. Bruce and Richard W. Uhrich, and work systems do or will depend on closed "Underwater Stereoscopic Television and circuit television to provide visual feedback. Display Realism," Naval Undersea Center, Improved display realism through the use of San Diego, CA, NUC TP 358, July 1973. 326 THE COMPARISON OF DISCUS BUOY WAVE SPECTRA PRODUCED BY BUOY-FIXED AND VERTICALLY STABILIZED ACCELEROMETERS E. L. Burdette Computer Sciences Corporation National Space Technology Laboratories NSTL Station, Mississippi 39529 state number. Abstract A 12-meter discus buoy was instrumented to True vertical acceleration may be measured measure wave spectra by the NOAA Data Buoy Office by a vertically stabilized gimballed accelerom- (NDBO) with both buoy-fixed and vertically stab- eter. Vertical stabilization has been achieved ilized accelerometers feeding a data acquisition by gyroscopic means4 and by an inertial system system which, on a 3-hour cycle, acquired and equivalent to a long-period pendulum. The latter stoed 15mintetime series of the two sensor approach is employed by Datawell in the widely otus.Theboyradpoed an opertedinut used Waverider buoy.5 The acceleration sensor otheutlati Ocea buor wseveralmonths and operate-i from this buoy, with analog double integration duedwae dtantiOcan forsveral ragmoondtions.ndThis circuitry, is marketed as a separate item, the duce wav dat ove a rnge f coditins. his Datawell Heave Sensor. paper describes the sensors and data acquisition system. The data are compared on the basis ofNDOoeasbuyinteAltcad their acceleration spectra; phase and coherenceNBOoeasbuyinteAltcad funcionsfor he to tie seies re aso eam-Pacific Oceans, and Gulfs of Alaska and Mexico. indfeunctios fore penthed twhic tindeicsate tatea Buoys moored at these stations have on-board fo.Rsut re frequenies where significantwae eneryt, data acquisition and reporting systems, commonly fomfeasuredce bywherbote outputificathwae buory-i referred to as payloads. A payload comprises fiedaccelroeteda be used buoy theouputaof theaboy meteorological and oceanographic sensors, elec- oftevricalxtaiied accelerometer.cnb sdtorpaeta tronics for the acquisition, processing, and of th vericall stailize accleromterformatting of data from these sensors, and a communication system to relay the formatted data 1. Introduction ~~to shore on a 3-hour synoptic cycle. Each oper- 1. Introduction ~~~ational buoy is normally equipped with a wave The masurmentof te acelertionof ameasurement subsystem, which provides a displace- Th~ ~ ~ ~ ~ ~ ~ ~ ~~~ tsecta measuremaen forh incelusaion in th floating surface platform has been used as a 6 etseta.7tmt o ncuini h technique for obtaining wave data for over 20 synoptic report. Two types of wave systems6' yer.Time series of buoy acceleration may be are used by NDBO, each of which has a buoy-fixed, prcseyooti aedipacmnsptrs. single-axis accelerometer mounted parallel to in either of two ways:thbuymsasaoinseor The decision to use a buoy-fixed rather * Analog or digital double integration than vertically stabilized sensor was based on of theceeaiontm e sresit a an NDBO in-house theoretical analysis, which displacement time series from wihaindicated that such a sensor mounted on a discus spectal esimat is otaind dirctlybuoy would satisfactorily approximate the Output *Transformation of an estimate of the of a gimballed vertical acceleromter. This acceleationspectum ino dislace-theoretical result was tested experimentally by mceeratb iiiongec spectral dninty ipae comparing displacement spectra from such systems F m~~~~vlen by thecorepoding fachsetora denit with spectra from a Waverider deployed nearby. vanglarfeuency to he fourthspondngfatr.o Results of the comparison showed no apparent anguar fequecy t thefourh poer.difference between the spectra. A more thorough Early experiments obtained acceleration data test was planned for the Experimental Environmental from sinle-ais aceleometr montedon aResearch Buoy (XERB), which is described in detail shrtpriod sphgeraiscacclperometer mounted onaby Steele, et al.8 A buoy-fixed accelerometer shortperid sphrica pendlum1or montedand a vertically stabilized accelerometer were rigidly parallel to the symmetry axis of a surface isaldo h EB iesre fsno following buoy.2 Either approach results in the i saldonuthepu;tsiwee secries onbadfo senalsisra measurement of the vector sum, or resultant, of otuswr eoddo or o nlssa gravitational and wave accelerations (including shore. horizontal components). Tucker3 has shown that use of the resultant acceleration can cause re- 2. The XERB Buoy duced sensitivity in the displacement spectrum to low-frequency waves, as compared to systems The primary goal of the testing associated that measure true vertical acceleration of a with the XERB buoy was the determination of the foat. The error was shown to increase with sea feasibility of obtaining sea surface directional 327 spectral estimates in near real time from a large supplied to a 32-channel multiplexer (mux) input (12-meter diameter, 91,000 kg), remotely operated to a 12-bit analog-to-digital converter (A/D) discus buoy. For these tests, NDBO used an which scaled linearly over the range +10 to -10 experimental payload, which is based on a 16-bit volts. The A/D was interfaced to a Data General minicomputer. The payload acquired and processed NOVA 4001 minicomputer, which controlled mux data into auto- and cross-covariances, which were channel selection and A/D conversion. A real- transmitted to shore for computation of the direc- time clock provided the time of year in a buffer tional spectral estimate. available to the minicomputer. The minicomputer was also interfaced to a Kennedy Model 1600 incre- A completely separate payload was also in- mental 7-track recorder, which produced a 556 bpi stalled on board XERB to confirm the proper oper- tape in a format compatible with most 7-track ation of the experimental system and to report computer system tape drives. The experimental meteorological data from the site. The latter payload was operated on a 3-hour cycle based on system was a standard NDBO operational payload midnight Greenwich Mean Time (GMT). Immediately of the Phase II type, manufactured by the following the completion of directional wave data Magnavox Corporation. A Wave Data Analyzer (WDA) processing at approximately 21 minutes after the subsystem of this payload produced a covariance synoptic hour, data were acquired and written on function estimate at the synoptic hour, which tape by the minicomputer. All active channels was broadcast to shore for processing into a were sampled essentially simultaneously by the one-dimensional spectral estimate. minicomputer in a single scan of the mux, which occurred at 0.5-second intervals. Tape writing A secondary goal of the XERB project was to took place between scans, with count values for provide a test bed on which to mount various each sensor channel from a scan recorded as a sensors. Digitized time series of sensor data single tape frame. In addition to the sensor would be recorded on the buoy, periodically re- data, each frame began with a two-character covered during service visits, and processed at synchronizing word and ended with a check word, shore. The time series would form a large data the sum of all data values in the frame. Mul- base, thereby allowing sensor performance eval- tiple frames were recorded within a physical uation, sensor comparison, and analysis of buoy record on the tape. A total of 1800 frames of motion, data (15 minutes) were recorded per synoptic report. 3. Description of the Experiment Although each six-bit tape data character had a hardware parity bit that was attached by Acceleration Sensors. The buoy-fixed acceleration the recorder and checked by the tape reading sensor was a Kistler triaxial servoaccelerometer unit, the checksum proved to be valuable in of the closed-loop, force-balance type. The flagging tape write/read errors. The most transducer was oriented with one axis parallel common of these was the loss of a character, or to the buoy mast and one toward the buoy bow. "byte dropout". On shore, data from the on-board Sensitivity of the mast-parallel axis was 1.02 tape was processed into engineering units prior V/m/sec. to storage in the data base. In the event of a tape error as evidenced by the checksum, all data The vertically stabilized accelerometer was for the offending frame were discarded. Each a Datawell heave sensor (40 sec), manufactured by missing frame was replaced by interpolation be- Datawell BV, Haarlem, The Netherlands. Vertical tween valid data points in the series. stabilization is achieved inertially by mounting the transducing element in a double gimbal with Deployment. The XERB was deployed on March 26, 1977, fluid damping. This arrangement is mechanically in the Atlantic Ocean at 30 24'N, 80008'W, approx- equivalent to attaching the transducer to a damped imately 100 km east of Jacksonville, Florida, on spherical pendulum with a 40-second period. Since the edge of the continental shelf. A single-point the longest-period ocean surface wave observed mooring was used to secure the buoy in 200 meters in nature is in the 30-second range, the pendulum of water. is not excited by the wave motion and remains vertically stabilized. In fact, the manufacturer Upon deployment, the experimental payload claims that response of the sensor to horizontal began recording data, with operation taking place acceleration in the wave period range is less on a 3-hour cycle based on midnight GMT. Reliable than 3 percent of the sensitivity to vertical data was obtained for approximately 3 weeks, until acceleration, which is 1.00 V/m/sec. hardware problems caused this payload to cease operation. Regular data recording resumed in Both sensors were installed below the deck June 1977 and continued for approximately 4 weeks. in a center trunk on the buoy symmetry axis. The At this point, serious payload problems necessi- vertically stabilized sensor was mounted with the tated retrieval of the buoy, ending this phase manufacturer's prescribed sensing point approx- of the project. imately at the waterline. The buoy-fixed sensor was mounted approximately 0.5 meter below this Sea states during the data collection period point and 0.3 meter off the buoy centerline. were light. Significant wave heights ranged from 0.2 to 2.1 meters, and winds ranged from calm to Data Recording. In its deployed configuration, 12 m/sec. 11 analog signals from 5 motion sensors were 328 4. Analysis of Data 5. Results Once the acceleration time series were Spectral comparisons of data from the two stored in engineering unit form in the data base acceleration sensors were made on 20 different format, a number of standard computer programs data records taken from the March-April 1977 could be used to process the data for analysis. period. Time series comparisons were also made Using these programs, the acceleration data were on selected records from this group. Figure 1 compared through comparison of the time series compares the two time series for a period of 1 data by means of plots, and through spectral minute. Note that agreement is quite good, analysis. The auto-covariance of each time series except for an offset in one of the sensors rela- was computed out to m =~100 lags, from which a tive to the other. raw spectral estimate, S (fk) was obtained for 1.0 frequencies P ' f = k/2mAt (1) 0.8 k 0.6 - where k = 0, 1, 2, ...,m; At = 0.5 second, and p = 1, 2 denotes the particular time series (cf. Bendat and Piersol9). The p were smoothed by Hanming to obtain an acceleration spectral estimate, �p(fk), having 36 equivalent degrees of freedom. . A cross-covariance function estimate was -0lt K also computed and used to produce a raw estimate = -.4-V of the cross-spectrum, from which a smoothed L i estimate 0.6 - i I S12 (fk) C12(fk) - i Q12fk) (2) :2.V- was obtained. A relative phase function 0 10 20 3) 5o 60 1 [Q,2(~~~~~~~~~~~~~~~~~k ~~~TIME (SEC) Q 12(f k) (f k) =tan 1(3) Figure 1. TIME SERIES OF ACCELERATION FROM BUOY- 12k I f 02 ~FIXED MAST-PARALLEL ACCELEROMETER AND VERTICALLY STABILIZED ACCELEROMETER. DATA ACQUIRED MARCH 26, 1977, 0700 GMT was also computed for each frequency. The phase function estimates the phase angle of time series Figure 2 compares the power spectrum of 2 with respect to time series 1 at frequency f acceleration from the gimballed accelerometer k with the spectrum from the buoy-fixed acclerometer. Data to compute the spectra shown in Figure 2 were series is defined by acquired on March 26, 1977, 0700 GMT. The same i2 comparison is made in Figure 3 for waves of higher frequency. Data to produce the latter spectra Y2 2 ( 12 k =(4) were acquired on April 2, 1977, 1500 GMT. Agree- 51(fk~ S2(fk) ment in both cases is excellent and is typical of the entire data set. This function is analogous to the square of a correlation coefficient computed for each fre- A displacement spectral estimate quency, fk' as described in Jenkins and Watts10. The coherence function is a measure of the linear D(2fk) = C2sfk)4 dependence of time series 2 on time series 1 at the frequency fk, and has the property that if was obtained from each acceleration spectrum and the two time series are related by a constant used to compute significant wave height via linear parameter, yj2(fk) will be unity. Simi- larly, when the two series are completely unre- H 4JTF (6) lated, the coherence function will be zero. If 1/3 0 the coherence function is between zero and unity, one or more of the following conditions is satis- The spectral moment, Mo, was obtained by fied: numerical integration of the displacement spec- trum. The displacement spectral density below a Uncorrelated noise is present in the two frequency fo was set to zero as a noise correction time series procedure, to prevent the noise level values of * The relationship between the time series acceleration spectral density from contaminating is nonlinear the wave height estimate. The significant wave height values shown in Figures 2 and 3 were com- e One or both of the time series is influ- puted using fo = 0.07 Hz. enced by an extraneous signal. 329 O.76 0.3 1 .03 0.7 ~ ~~~ ~~ ~ ~~~~~~~~~~~~~ \ _atadell A ccelerometer 1 5~ ~ ~~~~(*b Ginbled) - 0.6 H 1.4m l ----- Kistler Acceleroceter - J\n 4 tA Hl/l; (Buoy- Ficed) 0.4 0.3 I - .04~I 0.2 0.3 0.4 05 0.1 0.2 0 .3 N~ D 4 050 0.1 0.2 0.3 0.4 0.5 0 0.1 0.3 0.4 0.5 FRE(.ENCY (Hi) FREO ENC� (HI) Figure 4. COHERENCE BETWEEN BUOY-FIXED ACCELER- Figure 2. ACCELERATION SPECTRA FROM BUOY-FIXED ATION AND VERTICAL ACCELERATION TIME AND VERTICALLY STABILIZED ACCELEROMETERS SERIES AS A FUNCTION OF FREQUENCY. COMPUTER FROM DATA ACQUIRED MARCH 26 DATA ACQUIRED APRIL 2, 1977, 1500 GMT 1977, 0700 GMT 3.0 Further evidence of this conclusion is presented in Figure 5, which shows y2(fk) versus ~~2.5~~~~~~ A~~S I(fk). Note that coherence remains high until the acceleration spectral density diminishes _(G..wdel lc)ere) below 0.01 m2/sec4/Hz, at which level the coher- 2lHI/3 'I, 1ence drops rapidly to zero. The solid curve of 2.0 | l liKistlerAlcceler'oeter Figure 5 represents a "mind's eye fit" of the I I H~(uoy-IFixed) data. It is interesting to note that the curve En s 1/3=l does not pass through the origin, but intercepts IUs~~ \ ~~the abscissa at �l(fk) = 0.002 m /sec4/Hz. This lo ~~/ V \~intercept value is a rough measure of the level 1.0 / of truly random noise in the buoy system. O.lJ\'I . I 1.0 0.5. 5~~~~~~~~~~~~~~~~~~~~~~~~~~~.- 0 0.1 0.2 0.3 0,4 0.5 FREQUENCY (Hz) ' / 0. Figure 3. ACCELERATION SPECTRA FROM BUOY-FIXED .- AND VERTICALLY STABILIZED ACCELEROMETERS t 0. COMPUTED FROM DATA ACQUIRED APRIL 2, / 1977, 1500 GMT 7.-. �73 'D | atd Points Fal linq lear Curve 0.3-I .Ano blous Points Figure 4 shows the coherence between the Seen at f<0. 03 He o.2- two time series, yi2(fk), versus frequency, com- puted for the data of April 2, 1977, 1500 GMT, ol- I corresponding to Figure 3. Comparison with the . spectrum of Figure 3 indicates that coherence 0 ' between the gimballed and buoy-fixed accelerom- 0. 0.02 0.03 0.04 0. 05 0.06 eters approaches unity at frequencies for which CCEEION SPECTRAL DENSITY (2/s4/) the power spectrum shows a significant level of acceleration. Figure 5. COHERENCE BETWEEN BUOY-FIXED ACCELER- ATION AND VERTICAL ACCELERATION TIME SERIES PLOTTED AS A FUNCTION OF VERTICAL ACCELERATION SPECTRAL DENSITY. DATA ARE A COMPOSITE OF THREE TIME PERIODS 330 The phase of buoy-fixed acceleration relative beyond the academic, since an inertially stab- to vertical acceleration is shown in Figure 6, ilized sensor is more expensive than the single- which is a plot of 012(fk) versus frequency. The axis, buoy-fixed accelerometer by a factor of data points represent a combination of data from from 5 to 10, depending upon the particular the several records. Points for which the coher- single-axis sensor chosen. The buoy-fixed sensor ence function was less than 0.95 were excluded is also smaller by a similar factor. Although from the plot in order to minimize the effect of this is not a great concern for a large discus noise. A fit to the linear function buoy, constraints in handling the Datawell sensor are a concern when that accelerometer is employed. 0(f) = mf + b (7) The Datawell sensor remains an obvious choice was made by the method of least squares for the in some applications; e.g., on board a non-surface- plotted data. Estimates of m = -0.138 deg/Hz and following platform and in situations where an b = 0.232 degree were obtained with a correlation analog displacement signal is required. In NDBO's coefficient of 0.75. The least squares line is experience with the heave sensor as an accelerometer shown in Figure 6 as a solid line. A phase shift and with the Datawell Waverider which uses the same that is a linear function of frequency implies a sensor, no failures of the intricate suspension constant time difference between the two time have occurred when handling was conducted per series. In this case, the buoy-fixed accelerom- Datawell instructions for these devices. eter lags the vertical accelerometer by approxi- mately 38 msec. The phenomenon appears consis- tently throughout the data; an explanation of a 7. Suggestions for Further Investigation delay of this magnitude has not been found. In view of the excellent agreement between the buoy-fixed and vertically stabilized acceler- +5 ometers for relatively light sea states, the degree of agreement in heavier sea states should be investigated. - ~~~~~~EQUENCY (H) NDBO has introduced several boat-shaped hull 0.1 0.2 0.3 0.4 0.5 NOMAD buoys into its network of operational buoys. These buoys have been equipped with wave systems that use a buoy-fixed accelerometer as the system sensor. Preliminary results of high seas, in situ comparisons between displacement spectra from :. %..j . NOMADs and a GOES-reporting Waverider buoy have ' ' shown remarkable agreement. 8. Acknowledgements -10 The preparation of this paper and the work described herein were supported entirely by the Figure 6. PLOT OF PHASE OF BUOY-FIXED ACCELER- NOAA Data Buoy Office through National Space ATION RELATIVE TO VERTICAL ACCELERATION Technology Laboratories Technical Work Request PLOTTED AS A FUNCTION OF FREQUENCY DB-3T01. The contributions of Mr. K.E. Steele, who conceived and instigated the comparison experiment, and Dr. A. Trampus, under whose auspices the spectral analysis computer programs 6. Conclusions were developed, are also gratefully acknowledged. Results of the data analysis described herein, of which the data in Figures 1 through 6 are typi- cal, confirm the agreement between the buoy-fixed and vertically stabilized accelerometers. However, some care should be exercised in extending the References excellent agreement seen at the low to moderate sea states to higher sea states for which no 1. Tucker, M.J., "A Shipborne Wave Recorder", sensor comparison data is available. This is Transactions of the Institute of Naval particularly true in view of Tucker's theoretical Architecture 98, London, p. 236, 1956. result, which predicts that the error obtained by the two sensors should increase with the sea state 2. Marks, W., and R.G. Tuckerman, "A Telemeter- number. ing Accelerometer Wave Buoy", Proceedings of the International Conference on Ocean Wave However, the data summarized herein strongly Spectra, p. 281, Prentice-Hall, Tnglewood suggest that the buoy-fixed accelerometer is as Cliffs, New Jersey, 1963. accurate a sensor as the inertially stabilized vertical accelerometer for the purpose of esti- 3. Tucker, M.J., "The Accuracy of Wave Measure- mating wave displacement spectra from a large ments Made With Vertical Acceierometers", discus buoy. This conclusion has implications Deep Sea Research 5 p. 185, 1959. 331 4. Gaul, R.D., and N.L. Brown, "A Comparison of Wave Measurements From a Free-Floating Wave Meter and the Monster Buoy", Transactions of the Second International Buoy Technology Symposium, Marine Technology Society, Washington, D.C., 1967. 5. Verhagen, C.M., P.L. Gerritzen, and J.G.A. von Breugel, "Operation and Service Manual for the Waverider", Datawell, Haarlem, The Netherlands, Doc. No. 5095-67020. 6. Steele, K.E., and A.S. Hananel, "An Analog Wave Spectrum Analyzer System for Operational Buoys", Proceedings of the MTS-IEEE Conference, The Ocean Challenge (OCEANS 78), Washington, D.C., Sept. 6-8, 1978. 7. Steele, K.E., P.A. Wolfgram, A. Trampus, and B.S. Graham, "An Operational High Resolution Wave Data Analyzer System for Buoys", Pro- ceedings of the OCEANS 78 Conference, Washington, D.C., Sept. 13-15, 1976. 8. Steele, K.E. E.L. Burdette, and A. Trampus, "A System for the Routine Measurement of Directional Wave Spectra from Large Discus Buoys", Proceedings of the MTS-IEEE Conference, The Ocean Challenge (OCEANS 78), Washington, D.C., Sept. 6-8, 1978. 9. Bendat, J.S., and A.G. Piersol, Measurement and Analysis of Random Data, John Wiley and Sons, Inc., New York, 1966. 10. Jenkins, G.M., and D.G. Watts, Spectral Analysis and Its Applications, Holden-Day, San Francisco, 1968. 11. Adamo, L.C., K.E. Steele, and E.L. Burdette, "A GOES-Reporting Waverider Buoy", Proceed- ings of the MTS-IEEE Conference, The Ocean Challenge (OCEANS 78), Washington, D.C., Sept. 6-8, 1978. 332 A SMALL, WIDE-BAND, LOW-FREQUENCY, HIGH-POWER SOUND SOURCE UTILIZING THE FLEXTENSIONAL TRANSDUCER CONCEPT J. A. Pagliarini R. P. White Ocean Systems Division Ocean Systems Division Sanders Associates Sanders Associates Nashua, New Hampshire 03060 Nashua, New Hampshire 03060 Abstract The flextensional transducer offers the following advantages: Flextensional transducers have wider bandwidths, lower operating frequencies, o Good impedance match to water load. and higher power handling capabilities than any other types of transducers of o Greater bandwidth for a given size comparable size. They make possible a (3:1 advantage). new class of underwater sound sources having much smaller size while retaining o High efficiency. all the other operating characteristics of conventional sound sources. The o 'High power density per unit volume. flextensional concept appeared several years ago. The concept is based on the o High power density per unit weight * ~~vibrational modes of an oval cylindrical (30 watts per pound). shell. Mutual coupling between these modes results in potential bandwidths of o Reduced size and weight. several octaves. This paper presents the design concept and performance predictions o System robustness (i.e., bandwidth for a small, low-frequency sound source. and power margin). Test data on an experimental model are presented and compared with predicted o Easily implemented ceramic prestress. values. o Ruggedness. o Reliability. 1. Introduction 2. Background Requirements for transducers of small * ~~size, low frequency and wide bandwidth all Flextensional technology has existed favor a flexural mode of vibration. The for more than a decade. Elastic wave flextensional concept offers improved theory for shell deformation and vibration radiation efficiency and greater depth was developed by Dr. Royster and extended range compared to other flexural modes by C. Brigham. W. Toulis performed early such as the flexural bar or flexural disc. work in developing hardware for low freq- It offers a 3:1 bandwidth advantage at uency high power applications. The coup- frequencies in the 100 H~z to 5 kHz range, ling between ceramic stack length mode and even greater bandwidths if higher and shell membrane modes represents rela- order modes are incorporated. tively complex elastic wave theory which resists reduction to practice. * ~~~The flextensional transducer consists of a shell which, because of its shape Recently developed engineering design and dimensions, has one or more resonant equations formulated through a combined vibrational modes in the band of interest. theoreticallempirical approach utilizing * ~~A piezoelectric bar is used to excite the the consultant services of G. Brigham of desired modes. There are four classes of Wheeler Industries now permit greater flextensional transducers, distinguished accuracy in the prediction of transducer by their basic shapes, as shown in Figure performance characteristics. These 1. The subject of this paper is the Class equations also enable the engineer to IV flextensional transducer. extrapolate the design parameters to match other performance requirements. 333 This effort has resulted in Figure 2 and the nodal positions for the a high degree of confidence in perform- first flexural mode are shown. As can be ance predictability of Class IV flex- seen, the nodes are located at 90-degree tensional designs. intervals, with opposite-phased displace- ment in adjacent quadrants. A perfectly The flextensional transducer has circular shell in flexure thus has poor proven itself in a wide variety of radiating characteristics, since equal applications over the past ten (10) years areas of shell surface are moving outward including the following: and inward at the same time. If, however, the cross section is modified to an oval � Project MIMI - ONR/University of Miami shape, as shown in the same figure, the DeepStar Propagation Program at NUSC nodes are displaced toward the ends of the shell. As a result, there are unequal o AFAR - Range - Acoustic Communications areas moving outward and inward, and the Program shell is an efficient radiator. The � UDL - Mobile Acoustic Communications amplitude of vibration of the larger Program (diaphragm) areas is also greater than that of the end area, further enhancing the The AFAR and UDL applications are radiating efficiency. very important data gathering efforts in the field of acoustic communications. The flextensional transducers utilized in these programs have demonstrated conclusively their reliability, ---"N efficiency and broadband response. The salient features of the flextensional design are: * Minimum size and weight. * Low transducer Q (3:1 advantage over conventional techniques). * Greater power handling capability per unit volume (3:1 over conventional techniques). \ . ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~(NCo~5 o System robustness (performance margin against bandwidth and power ,.U levels). CLASS I CLASS II ----- Fig. 2: Fundamental flexural modes of a ring and an ellipse Some means must be provided to drive CLASS III CLASS IV the shell into this mode of vibration in response to electrical signals. This is Fig. 1: The four classes of accomplished by inserting a piezoelectric flextensional transducers ceramic bar between the ends of the shell, as shown in Figure 3. The bar is made up of ceramic slabs, individually electroded 3. Design Concept and polarized, and bonded together with epoxy. The electrodes are connected so that all the slabs are driven in parallel The Class IV flextensional trans- in the k~3 mode. This mode has the high- ducer has a shell which approximates an est electroacoustic coupling coefficient, oval shaped cylinder. A cross-sectional resulting in maximum transfer of energy view of a circular cylinder is shown in to the mechanical motion. The use of 334 many slabs in parallel also results in a low electrical impedance, which facili- tates amplifier matching.I 1IN' DIAflIRACM S-AI NCTIN F'RINEL IN I Fig.~~~~~RW 3:BscFetnialCocept~ The electric field applied~~~~~~~~ across each ceram~~icsabzrdcsasri nFg 4N /A~ TpcalClasI etni na the~~~~~~~ dieto _niae.Ti ocsTasue the~~~AAI edofteseltdelcwic in urnprouce a eflctin i te Wen esinin ths tpe f fex diaphragms. The amplitude of deflection tensional transducer the lengthfof th of the diphragms i greater han thatstack is ade slighly greate than th of the shell ends by the r~~atikfmjrlnt fte pnn ewe h hl sufacsithg.r litl Bu-fpasein Fetension.Bavalue mustbecept raitinfothe end.lecasofthis, witineomeld copresived stress teac ceramicsla rducer s an eficetradiatr rainFgs to avi signiicant Closs IFexnsofa The amltdeirpuaswlstecoupiong odeindiffcated. This lorcestrasue antr rouesa step-plectween the Wherai stcdopessiveniasgconistyen wit fleximu and the diaphragms, also gresutsigodperathang dethan d power is mde slighl reablefrta h 0 ac~oustic impedandshyte rai fmajrlntch tof the watermaimum bandwidnth. hl with fior axesven trasuer vovlum. The rai-ndsert thesack the shell mustbexaddo atinguarea of sheoiahrgs cosrion dealsformc apllwsaced insehyrauicn Copressiv bandfoc gre~~~~~~~~~~auses thedsanc btweeno the shell ends. s i eurdbcuehdottcpesr a~~~~~~~~~~t inreasel, alowigistfte raition of onthediprgsaswlashhdiv acou~~~~~~~~~~stack Befoerg occusefrton, the diprg ees ends tofpac theceaistk ~~~~~~~~~~~stracks aihvry ite oat-fpae d with tepsoxy. BAftaler mstacekep ~~~~~~~~~~~~israition frmthe end.rBcaue oftis, wti removmned frompesv sthes T~~~~~~~~~~~hel dahagmpiuestpu, aswll a h owping th epin efcs. hel loenst area step-up between copres the crmcstack . Thpesie transducostent withmxiu acou~~~~~~~hnesetialycoimpleteancepmtc for thewatr o aiumbnwdh edcausers tohpevn fodistnce betee the shell ed sak Beoeintertior.teedso h stc3r cae ih px.Afe5tc 4. Performance Predictions The predicted impedance, admittance, and constant-voltage transmitting response curves for a unit of the above design are shown in Figures 5, 6, and 7. In addition, the predicted constant-voltage trans- mitting response with series tuning is shown in Figure 8. I 1.0 0.6 0.8 1.0 1.2 1.4 U) RELATIVE FREQUENCY 4000. o 0.8 Fig. 7: Predicted Constant-Voltage Transmitting Response (TVR) c 5Q00" 0.7 0 100 200 300 Resistance OHMS Fig. 5: Predicted Impedance 2.5 . . 0.. 6 0.8 1.0 1.2 1.4 RELATIVE FREQUENCY 0 2.0 . 2 Fig. 8: Predicted TVR with Series Tuning -I 0.7 5. Test Results !.5 !.5 :Measurements of the above performance 1.5 0 0 .5 1.0 1.5 parameters were made on several units at G, Millimhos Sanders' acoustic calibration facility. In addition, measurements of power linear- ity and directivity were made. The results of a typical unit are shown in Figures 9- 16. The measured values closely match the predicted values. The impedance and ad- mittance curves, as well as the response measurements, show a mechanical Q of approximately 3, which represents a 33 per cent bandwidth. This exceeds by approx- imately 3 to i widths achievable by other transducer types such as the flexural disc. Power linearity is demonstrated up to a level of at least 192 dBu PA at I meter. 336 1.0 V400- 0.9 .1 12o 1.1 ) to. 8-10'- / 4500 - o l 0.7 X: 0.6 0.8 1.0 1.2 1.4 0 100 200 300 RELATIVE FREQUENCY Resistance OHMS Fig. 9: Measured Impedance Fig. 12: Measured TVR 2.5" 0.9 - o /0.8 1.2 r 2.0- 1.0 0H0.7 0 1.5 : : ' 0.6 0.8 1.0 1.2 1.4 0 0.5 1.0 1.5 RELATIVE FREQUENCY G, Millimhos Fig. 13: Measured TVR, with Series Tuning Fig. 10: Measured Admittance 200 " 1.2 190 1.1 100" 0.9 Li 180 0 I I W I X 0 100 200 300 0 10 Relative Input,dB Resistance OHMS Fig. 14: Measured Power Linearity Fig. 11: Measured Impedance with Series Tuning 337 6. Conclusion In conclusion, a transducer having wider bandwidths, lower operating frequency and higher power handling capability than conventional types of comparable size has been achieved by means of the flextensional concept. References +~-~~~~ ~ 1. Brigham, G.A., Design Parameters of the Class IV flextensional transducer, Wheeler Industries Final Report under Sub-contract NFZAC N95 105939M. 2. Royster, L.H. and G.A. Brigham, Present status in the design of flextensional - Xunderwater acoustic transducers, J. Acoust. Soc. Am., Vol. 46, p76, 1969. 3. Rutledge and R.A. Royster, Approximate model for predicting operational para- meters of the Class IV Flextensional Underwater Transducer J. Acoust. Soc. Am., Vol. 50, p82(A), 1972. Fig. 15: Measured Pattern, XY Plane 4. Brigham, G.A., An equivalent circuit (scale - ldB per division) for oval flextensional transducers, J. Acoust. Soc. Am., 51, 82(A), 1972. 5. Brigham, G.A., Vibrations of oval rings, J. Acoust. Soc. Am., Vol. 42, p. 1203, 1967. *' ' 6. Brigham, G.A., Analysis of the Class IV Flextensional Transducer by use of wave mechanics, J. Acoust. Soc. Am., Vol. 56, pp 31-39, 1974. 7. Brigham, G.A., Lumped parameter analysis and equivalent circuit of the Class IV Flextensional Transducer, NUSC Tech. Report, July 1973. Fig. 16: Measured Pattern, XZ Plane (scale - ldB per division) 338 AN ANALYSIS OF THE DISTRIBUTION OF MEASURED BUBBLE PULSE PERIODS OF EXPLOSIVE SOURCES Richard J. Hecht Louis A. Mole Underwater Systems, Inc. Underwater Systems, Inc. 8121 Georgia Avenue 8121 Georgia Avenue Silver Spring, Maryland 20910 Silver Spring, Maryland 20910 Abstract Signal, Underwater Sound (SUS), which is a stand- ard U.S. Navy item. Manufacturing specifications Small explosive charges are often utilized for the SUS were obtained from the Naval Weapons as acoustic sources in the measurement of signal Station, and are summaried in Table 1 -9 The propagation in the ocean. Of particular im- SUS used in this experiment nominally contain portance when used as investigative tools is the 1.8 lb of TNT and were preset to detonate at consistency of the generated power spectrum. A pressures equivalent to 60 and 300 ft of depth. measure of consistency is the degree of variation The detonation is initiated by mechanical means in the first bubble pulse period, defined as the when the design pressure is attained. The SUS interval from the onset of the initial shock is manufactured under rigorous military specifi- wave to first bubble pulse maximum. To deter- cations so that the performance is consistent mine the degree of consistency that can be ex- from shot to shot. pected from a series of detonations, an experi- ment was conducted in which over 1300 charges TABLE 1 7-9 were detonated at nominal depths of 60 and 300 SUS MANUFACTURING SPECIFICATIONS feet. Using the pressure time histories, the SUS Type bubble pulse periods were first determined and Mk Zl Mk 82 then analyzed from two viewpoints: first in re- Lowest actuating pressure 22 psi 120 psi lation to the charge specifications, and second Highest actuating pressure 31 psi 147 psi with regard to the possible physical reasons for Detonation depth tolerance + 10 ft + 30 ft the shapes of the distributions. The effects of Average descent velocity 15 ft/s 15 ft/s charge weight, detonation depth, temperature and Minimum Explosive (TNT) 1.65 lb 1.65 lb wave height on the observed bubble pulse period Booster (Tetryl) 0.07 lb 0.07 lb * ~~~distribution were analyzed. 3. Experiment 1. Introduction The data was acquired during an exercise in The general features of the acoustic wave- the Northeast Pacific Ocean in September 1973. * ~~~forms generated by underwater explosives are well The SUS shots were launched from the shooting known, having been analyzed in detail by a num- ship over a period of a few days as the ship fol- ber of individualsi-5. The acoustic pressure lowed a northerly course from Latitude 29001' N versus time signature of an underwater explosive to 46045' N. The 60 and 300 ft depth SUS were source is a sensitive function of the charge launched from the ship so as to detonate at two weight and the depth of the explosion. The bub- minute intervals. A transducer suspended from ble pulse period of an underwater explosion is the shooting ship was used to receive the SUE a function of both the explosive yield and the signals, which were recorded on analog magnetic detonation depth, and, therefore, can be ef- tape in the FM mode, along with IRIG B time code fectively used to monitor the consistency of the a nd voice communication. Table 2 shows the num- resultant acoustic source spectrum and the de- ber of each type of SUS launched and the number tonation depth. The bubble pulse period is de- that detonated. fined as the time from the beginning of the shock wave to the maximum of the first bubble pulse6. TABLE 2 The intent of this paper is 1) to present an an- TABULATION OF SUE STATISTICS * ~~~alysis of an experiment it, which 1300 explosives SUE Type were detonated at depths of 60 and 300 ft, 2) to n 61 MkB82 compare the performance of the explosive sources Nominal detonation depth 60 ft 300 ft * ~~~with the manufacturer's specifications, and Number dropped 696 695 F ~~~3) to compare the variations in the data with Number processed 654 655 recorded environmental changes. Number not recorded 16 14 Duds 26 26 2. SUS Specifications The explosive used in this experiment was a 339 4. Data Processing versely proportional to depth to the five-sixths power, the corresponding bubble pulse variation The signal processing was performed using a is +10%. Correspondingly, for the Mk 82, three minicomputer that was programmed to analyze the hundred foot SUS, a specification tolerance of recorded data in real time. A block diagram of + 30 ft is permissible, thereby resulting in a the processing system is presented in Figure 1. + 8% variation in bubble pulse period. Because For each shot, successively, the computer deter- of the insensitivity of the bubble pulse period mined: 1) the occurrence of an individual explo- to yield variations, it must be concluded that sive signal, 2) the approximate detonation time, the variations are largely controlled by the 3) the bubble pulse period, and 4) the error depth specifications. Examining Figures 2 and 3 bounds. After these calculations were made, the it is evident that all but a few uf the points computer presented the results to the operator for the bubble pulse periods lie within these for evaluation as to acceptance or rejection. specifications. The analog to digital sampling rate of the SUS Figure 4 is a plot of the observed bubble signal was derived from the 8th harmonic of the pulse period versus time for the Bk 61 SU5. An time code carrier, which provided a time res- examination of the figure shows that the majority olution on the bubble pulse period of 0.125 ma. of the values lie between 125 and 130 ins, which in addition, this processing procedure removes are equivalent to 52.2 and 56.4 feet in depth. any speed variation present due to the analog This range of values, in which the majority of tape recording or playback process. A quality the points lie, is comparably demonstrated in control check on the automatic processing indi- Figure 2. If we were to set the acceptance cated consistent agreement to within 0.15 me when criteria for these SOS to be within this range of compared to manually determined bubble pulse values, then the depth variation is + 2%, which periods. is the same percentage as that for the yield. Because the overwhelming majority of the shots' bubble pulse periods fall within this small win- S. Data Statistics dow, it is evident that the variation in depth and yield is actually much smaller than the speci- After completion of the processing, the data fications indicate. Using this criterion was analyzed statistically to determine the con- (bubble pulse period between 125 and 130 me) for sistency of the events. Figure 2 presents a selection of the acceptable SUS for propagation histogram of the bubble pulse periods for the loss measurements, the estimated depth variation Bk 61 SUS. The median bubble pulse period is of the source is on the order of plus or minus 127.5 ms which, for a yield of 1.8 lb of TNT, cor- 2 feet. As a matter of interest this is compara- responds to a detonation depth of 55 ft. The mean ble, if not superior, to the depth control that value of the period is 126.9 ms with a standard can be maintained for a towed continuous source. deviation of 3 ma. The distribution exhibits a The observed bubble pulse period versus time strong skewness to the shorter bubble pulse peri- for the Bk 82 SUS is shown in Figure 5. An exami- ods, the coefficient of skewness being -1.3. nation of the range of values reveals the same Figure 3 presents the bubble pulse periods spread of data as was observed for the Bk 61 505. for the Bk 82 SUS. The median bubble pulse is However, there is an obvious trend, of a de- 42.8 me which, for 1.8 lb of TNT, corresponds to creasing value in the bubble pulse period, during a detonation depth of 29i ft. The mean value of the experiment. As the ship transits northward, the data is 43.2 ms with a standard deviation of the SUS appear to detonate a deeper depths. This 1.3 ins. The distribution can be considered to be trend is discussed in the next section. Although symmetric, since the coefficient of skewness is the range of the majority of the points (shown in only -0.40. Note that the data for the two SUS Figure 3) appears to be 4 me, if the trend (shown types are similar in that the mean and median in Figure 5) is taken into account the range is values correspond to shallower depths than the ex- only about 2.5 me. If this spread of values is pected nominal values. used as the criterion for SUS selection, the same Since the bubble pulse period is dependent on narrow range of source depths is achieved as for4 both the explosive weight and the detonation the Bk 61 shots. It should be noted, however, depth, the relative effects of the construction that the uncertainty of 0.15 me due to automatic specifications for each variable were examined processing increases the range for these deeper with regard to their effect on the bubble pulse shots more than it actually is. To improve the4 period. The explosive specification is at least data accuracy at deeper depths, both the sampling 1.65 lbs of TNT plus 0.07 lb of booster. The rate and, correspondingly, the recording band- maximum explosive weight is controlled by the width would have to be increased. "f ill line" of the explosive container, which is considered in this analysis to be 1.85 lb 8. Since both the peak pressure output and the bubble pulse 6. Environmental Factors period are proportional to the cube root of the yield, a variation of + 2% in the bubble pulse In order to obtain a better understanding of period can be expected for these small variations the data shown in Figures 4 and 5, hourly averages in yield. of the data were determined. These results are For the Bk 61 sixty foot SUS, a depth speci- shown in Figures 6 and 7 for the Bk 61 and Bk 82 fication tolerance of + 10 ft is permissible. shots, respectively. The straight lines on each Therefore, since the bubble pulse period is in- of the graphs represent the linear least squares 340 fit to the data, which emphasizes the increasing Acknowledgements depth trend of each curve. Several observations can be made about the data when plotted in this This analysis was made possible by the coop- manner. Both data sets demonstrate a trend to eration of many people, including K. Lackie and lower depths, although the indication is much V. Sprague from NAVOCEANO, who acquired the data stronger for the Mk 82s shown in Figure 7. The and made it available to us for processing. We data of Figure 6 also demonstrates a larger vari- would also like to thank Y. McGann and R. Johnson ation in hourly averages during the last thirty for discussions regarding SUS specifications and hours of the experiment. The data in Figure 4 performance and P. Bucca for supplying environ- also show a greater spread of values during this mental information. The signal processing was period. These results appear to correlate quite performed under the auspices of the LRAPP Office well with sea state, which is plotted in Figure 8 of ONR under Contract N00014-73-C-0484. along with the data of Figure 4. The wave height is the sum of swell and wind wave height as re- corded from the bridge of the shooting ship. With References waves sixteen feet in height, it is possible for the bubble pulse period to be affected by the 1. Cole, R. H., Underwater Explosions, Princeton corresponding dynamic pressure changes of as much University, Princeton, N.J. (1948). as 4 psi. As noted in the previous section, the sharp 2. Weston, D. E., "Underwater Explosions as downward trend in values shown in Figure 5 is also Acoustic Sources," Proc. Phys. Soc. Lond. present in Figure 7. One possible explanation for 76, 233-244 (1960). this may be the temperature decrease in the ocean that occurred as the ship transited northward. 3. Slifko, J. P., "Pressure Pulse Characteristics Figure 9 shows the time history of the temperature of Deep Explosions as Functions of Depth and during this period. At the 300 ft depth, the Range," Naval Ordnance Laboratory, NOLTR temperature decreases by about 120 C at a linear 67-87 (1967). rate during this period. A possible, but somewhat speculative, cause for the increase in detonation 4. "Manual of Test Procedures for Signal Under- depth may be that more pressure is required to water Sound (SUS)," Vast, Inc., STF-TP-102 activate the firing mechanism as the "0" ring (1970). seals become colder. Although the temperature decreases in a similar fashion at the sixty foot 5. Bedford, N. R., S. K. Mitchell and M. S. Wein- depth, the actions of the waves at the surface stein, "Determination of Source Depth from appear to make the temperature effect less the Spectra of Small Explosions Observed at noticeable. Long Ranges," J. Acous. Soc. Amer. 60, No. 4, 825-828 (1976). 7. Conclusion 6. Gaspin, J. B. and V. K. Shuler, "Source Levels of Shallow Underwater Explosions," Naval Ord- The intent of this paper has been to demon- nance Laboratory, NOLTR 71-160 (1971). strate that SUS explosive charges may be used, with a high degree of confidence, as acoustic 7. "Experimental and Development Specification, sources. Specifically, the results demonstrate SUS," Bureau of Naval Weapons, Dept. of the that 1) the SUS are within manufacturing specifi- Navy, XWS-4660(RU) (1963). cations, 2) if the ranges of the bubble pulse periods that constitute good shots are properly 8. "Explosive Section, Underwater Sound Signal," defined, depth uncertainties may be reduced by Bureau of Naval Weapons, Dept. of the Navy, 25% or more, and 3) the environment can cause MIL-E-23271(WEP) (1962). perturbations in the observed bubble pulse peri- ods. An appreciation of these factors and their 9. "Arming and Firing Mechanism," Bureau of Naval interrelationships is of importance to anyone Weapons, Dept. of the Navy, XWS-4669(RU) planning to utilize SUS as acoustic sources. (1963). 341 16- 14- TAPE RECORDER ~ 12- 4, 12- VOICE TIME CODE FREQUENCY AID - CHANNECH AHANNEL PROCESSOR INPUT 0 DIGITAL 8- COMPUTER R 0SEVTN ~ 6- OUTPUT DATA TAPE STORAGE 4- OPERATOR I CONTROL 2- Figure 1. Block Diagram of Bubble Pulse Period Real Time Computer Processing System 38 40 42 44 46 48 Bubble Pulse Period (MS) Figure 3. Distribution of Bubble Pulse Periods for the Mk 82 SUS (300 foot) 12- 10- O 6- 4- 2- 120 130 140 Bubble Pulse Period (MS) Figure 2. Distribution of Bubble Pulse Periods for the Mk 61 SUS (60 foot) 342 140 120- - I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 02 110~~~~~~~~~~~~ 2200 2300 2400 2500 2600 220 2200 240 2000 Ti..me of ..Occurrence (Doy ad Hour) Figure 4. Diagram of Measured Bubble Pulse Figure 6. Diagram of Hourly Averages of Bubble Period Value Versus Time for the Pulse Period Versus Time for the Mk 61 (60 foot) Mk 61 (60 foot) 45 - . - * ' . . o..*. 'V*.4. . -.t.- . :. ot.. � -~.~.o _;o *o-. ... .; _ � �.. 40. *.... . .. a5 - 1~~ ~ ~ ~ ~ ~ ~ ~ ~~ ~~~20- � � ' 30~~~~~~~~~~~~~~~~~~13 2200 2300 2400 22500 2600 Time of Occurrence (DSy and Hoer) Figure 5. Diagram of Measured Bubble Pulse Period Value Versus Time for the Mk 82 (300 foot) 343 ~~~~~26. ~~0- .*~ Surface Value * 31s 2 *~~~~~~ 60 ft. Value *-.lt'* � . 300 F. Value 22.0- '! *3e 44. 14.0- ss t o Figure 7. Diagram of Hourly Averages of Bubble Pulse Periods Versus Time for the Mk 82 (300 foot) 6.0 2200 2300 2400 2500 2600 Observation Times (Day and Hour) Figure 9. Time History of Temperature Values Measured as Ship Transited Northward 140 0 12 120 - * ** *16 A Pulse Period Values Versus Time for the Mk 61 (60 foot) 3440 2200~~~~~~~~~~o 23oo 2400 2500 260oo ~~~~~~~~~~~OsrainTimeso curec (Day and Hour) ~~~~~~~~Figure 9. Diarm o ae Heightor aof Tmperasured Bublues Pulse P~~~~easriod ValusVrusShime foransthed k614 (6 foot t~o.. "..:'' '-:3"'4 STATISTICAL MEASURES OF BEAM OUTPUT NOISE DUE TO SHIPPING M. Moll R. M. Zeskind W. L. Scott Bolt Beranek and Newman Inc. Bolt Beranek and Newman Inc. Bolt Beranek and Newman Inc. 1701 North Fort Myer Drive 1701 North Fort Myer Drive 1701 North Fort Myer Drive Arlington, Virginia 22209 Arlington, Virginia 22209 Arlington, Virginia 22209 Abstract densities of the various types of ships on those routes; and the transmission loss to the array In a band of rather low frequencies, almost site as a function of range and direction. all of the ambient noise in the ocean is gener- The averaged noise power at the output of a ated by ship traffic. If the output of a beam- beamformer in a specified passband resulting from former in this band of frequencies is squared ship traffic in an acoustic basin* maybe expressed and averaged, the result would be found to as fluctuate in a random manner. This quantity can be regarded as a random variable Y characterized m n ij by a probability density function. Y = E E Z S i Z B 1) This paper presents an algorithm for the i=l j=l k=l ijk ijk ijk calculation of the probability density function of Y, which is organized according to shipping where m is the number of routes in the basin, routes in an acoustic basin and the types of n is the number of ship types, ships on those routes. A.. is the number of ships of type j on One of the inputs required by the algorithm route i; Aij is a random variable, is a set of source characteristic functions, one S is the source intensity of the kth for each type of ship. Additional required in- ship of type j on route i; it is a puts include the beam pattern of the array, random variable that is statistically transmission loss functions, ocean route enve- independent of the source intensity lopes, and average numbers of ships on those of any other ship, routes. Zijk is the intensity transmission ratio from ship ijk to the receiving point, Bijk is the gain for a plane-wave arriving 1. Introduction at the array from ship ijk. In a band of rather low frequencies, almost Various first-order probability measures can be all of the ambient noise in the ocean is gener- derived from (1); these measures depend on the ated by ship traffic. The performance of sonars statistical measures of the random variables in operating in this band is limited by this compo- the summand. nent of background noise; therefore, it is nec- essary to measure or predict certain statistical measures of beam noise to support system design 2. Characteristic Function for Y and deployment decisions. The probability measures of beam noise The probability density function for Y can be depend on many factors: array configuration, obtained from its characteristic function orientation, and location (including depth), and season. Thus it would be both expensive Ay(w) = E[expjwY] (2) and time-consuming to attempt to obtain the required beam noise measures for a large number where E denotes the expectation operator. Substi- of sites experimentally. Furthermore, if ship- tuting Cl) in (2) gives ping routes andtraffic densities were to change because of strategic or economic causes, theM nAi measurements would have to be repeated. There ( = m is, therefore, considerable incentive for the y l= jm l kAl ijkZijk ijk development of a capability of predicting statistical measures of array beam noise. Input data required for the prediction in- clude the following items: array location, *The acoustic basin in which the array is located orientation, and plane-wave directivity pattern; is defined by bathymetry (e.g., land masses and acoustic source characteristics of various types high underwater ridges). The intensity of scund of ships; ocean route envelopes, and the traffic from sources outside of the basin is negligible. 345 A.. m n lj -= E E t X expjSij.kZijkBijk (4) where aij is the average number of ships of type j i=l j=l k=l in an increment of route i. Substituting (ll) into (10) gives m n =T t T .(u) C(5)m i=l j=l i ij(u) = exp(-aij) a {aij dgfGi(g)fdqfQi(q;g) where x Sj [Z (g,q)b(g,q)]}a a! (12) ij() = E expjWSijkZijkBiJ] (6) The sum is the power series expansion for the expo- nential of the quantity in braces; thus In the sequel it is assumed that C = exp(-a.) expaa fdgf (g)fdqf (qg) ij() = exp(aij) exp{aijgfGi(g)dQi(q;g) Zi = z(ik. (7) x k' [wz (g,q)b(g,q)]} (13) Bijk = b(G ijkQ ijk (8) Bijk = b(Gijk,' ijk3 The cumulant function is the logarithm of the characteristic function: where z( , ) is a deterministic transmission ratio function; b( , ) is a deterministic gain function for a plane-wave; Gijk is the earth-centered angle ij() ij Gi()dgfQi(qg) measuring the longitudinal coordinate (in the direction of the nominal route) of ship ijk; and x Sj [ Z (g,q)b(g,q)] - 1} (14) Qijk is the earth-centered angle measuring the transverse coordinate (normal to the nominal route) of ship ijk. It is assumed that all ship coordi- If the distributions of the longitudinal co- nates are random variables that are mutually in- ordinates G ij are uniform over the route, then dependent. Substituting (7) and (8) in (6) and performing the expected-value operation gives fGi(g) = P', � < g < i = 0, elsewhere (15) ij(w) aiJ (a)[dgfG (g)fdqfQ (q;Sg)fdsfSj(s) a where &i is the length of route i. Substituting x expjusz(g,q)b(g,q)] (9) (15) in (14) yields where pij(a) = P(Aij = a); a = 0, 1, 2, 3, ... ij = k ri dgdqf (qg) fG ( ) is the probability density function i for G..k ; j = 1, 2, ..., n; k = 1, 2, 3, ...X USj [w z (g,q)b(g,q)] Pjikij (6) f ( , ) is the probability density function for Qijk depending on the variable 13 Qi for Qij*k depending on the variable where p is the earth's radius, and ki. is the average number of ships of type j per unit length g; j =1, 2, ..., n; k = 1, 2, 3, ,. of route i. In (16) the variables of integration are the route variables g and q; an alternative The inner integral of (9) can be expressed as a variables erati charctersti function set of variables is the pair of observation vari- characteristic function; thus ables related to the range and bearing of point from the sensor. ij(W) = p (a) {fdgF.(g)fdqfQ(q;g) The characteristic function is the exponential 13j a= 0 ij a IgGi Qi qg(9@jof the cumulant function; thus i. (u>)= exp 'ip(j ). Substituting this result in (5) gives the charac- teristic function of Y: x $SJ [Q z(g,q)b(g,q)]}a (10) m n where CSJ( ) is the source characteristic function Cy(0) = Z exp pij( ) for ships of type j. i=l j=l In the sequel, it is assumed that A-i is a m n Poisson random variable; thus, its probabilities = exp E x .. ) (17) are given by i=l j=1 pij(a) = exp(-aij)aij , a 1 , 2, 3, ...ll) The second of these equivalent forms is easier to calculate. 346 3. Probability Density Functi6n for Y 4. Source Characteristic Functions Once the characteristic function (Dy(w) has It is shown in (10) that the calculation of been found, the corresponding probability density the characteristic function of Y requires a source function fy(y) can be obtained via the inverse characteristic function $Sj( ) for each type of transformation. Because fy(y) is real, the in- ship. In particular, each of these characteristic verse transformation is given by functions pertains to the same frequency band as the ambient noise prediction. The required source 1 characteristic functions can be derived from sta- fy(y) = f /(w) cos wy de tistical information produced by J. C. Heine And L. M. Gray. [1] SHeine and Gray round it expedient to divide + - f I(w) sin wy dw (18) the world's merchant fleet into three types of T0 vessels: fishing, small merchant (ships over 500 long tons and less than 700 feet long), and large where R(w) is the real part of the characteristic mrhn sisoe 0 etln) where ~~~~~~~~~~~merchant (ships over 100 feet long). function cOy(w), and 1(w) is its imaginary part. For each type of vessel, four noise generating Simplification of (18) can be achieved by ex- components were characterized: (1) the broadband ploiting the fact that Y can only take on positive spectrum that is related to the collapse of pro- values; therefore, fy(y) = 0 for y < 0. For y > peller cavitation bubbles; (2) the narrowband 0, (18) becomes spectrum (fundamental and harmonics) at the pro- Io peller blade rate which are due to the modulation f (-y) = f R(w) Cos wy dw of the cavitation spectrum by the spatially vary- F iT 0 ing inflow to the propeller; (3) the narrowband spectrum (fundamental and harmonies) at the diesel firing rate which are related to the vibratory Tr_ 1 1(w) sin wy dw = 0 (19) forces of the main propulsion system; and (4) the narrowband spectrum at the ship's electric plant frequency. The use of (19) in (18) yields alternative, freque nc y. simpler expressions for the density function-:h eut fHieadGa eei h form of histograms for the source pressure level and source frequency distribution for the funda- f (y) = 2 f w) cos wy du ' (20) mental narrowband components. The levels of F Tr 0 harmonics relative to the fundamentals were speci- fied by equations. = 2 f IM sin Wy dw (21) A procedure was devised to convert the Heine- - s Gray narrowband data to the distribution of source level in a specified 1-Hz band. If the fundamen- Thal is greater than 1 Hz, then no more than one single integral for y > 0, and either (20) or component of the harmonic set can fall in a 1-Hz increment of frequency. Let (21) can be used. This result will reduce the computational load since only the real part of C be the event that no component falls qy(w) need be calculated and stored, or Vice 0 in the specified 1-Hz band. versa. A numerical approximation technique was C be the event that the fundamental falls developed to evaluate the special form of the 1 in the specified 1-Hz band. inverse transform given by (20). It is assumed that R(w) is negligible for all w greater than C be the event that the first harmonic some wN. Assume R(w) has been computed at N+l 2 falls in the specified 1-Hz band, and equally-spaced points in w, where the first sample point is at w = 0, and N is an even inte- so on. ger. The range of the integral in (20) is With the stipulation that the frequency of the divided into sub-intervals, each containing three sample points. Using a quadratic approximation amental isclusive; than is, only on ts for R(w) in each sub-interval, the integral over aemtal xlsv;ta s nyoeo h events can occur in a single performance of the each sub-interval can be evaluated analytically. This results in a numerical approximation to (20) experint. Now let U be the source pressure level in a of a form that can readily be implemented on a of aformthatcanreadly b imlemeted n aspecified 1-Hz band from a particular narrowband digital computer. This numerical approximation spec e thz rad from a wel las the has the desirable property that the selection of furcentat rdiate t ot a l aslty the nterampl intrvalis dctatd ony byfundamental. According to the total probability the iExampintervaltionof(2)hos dcthated ifonv- btheorem for mutually exclusive events, the prob- RMe. Examination of (20) shows that if conven- aiiydsrbto ucinfrUi tional numerical integration techniques were employed, the selection of intersample intervals would depend on the value of y. Fu(u) = 1 P(U < uICi)P(Ci) (22) 347i= 347 where P( ) denotes the probability of the argu- component source levels. For example, ment. The corresponding probability density the mean value of the source intensity function is of the kth component is n f Cu) = d F (U) = Z fu(UICi)P(Ci (23) u10 U U U I m. du O fui) (24) TU_ i=O 1Ui The source level density for the fundamental is 3. Determine the low-order moments of the derived directly from the corresponding histo- source intensity from the low-order gram; the density is therefore a discontinuous moments of the component source inten- function. For a given harmonic, the same function sities. is employed, displaced by the amount the harmonic level is exceeded by the fundamental level. The 4. Determine the values of the parameters probability that the fundamental falls in the of the characteristic function so that specified band is readily determined from the its moments match the calculated moments. frequency histogram for the fundamental. The same histogram suffices for the first harmonic, The advantage of this procedure is that it avoids except that the frequency scale is doubled, and the excessive oscillations of the characteristic so on. functions that would be introduced by the direct As stated previously, the broadband compo- utilization of the discontinuous density functions. nent of the radiated noise is related to the collapse of propeller cavitation bubbles. From a consideration of significant differences in 5. Computer Programs the distributions of peak levels and associated frequencies of the continuous cavitation spectrum, The algorithms for calculating the prob- Heine and Gray defined eight broadband classes, ability density of Y have been programmed in each with its characteristic spectrum. The dis- FORTRAN for a CDC 6000 digital computer. The tribution of these broadband classes among types algorithms are implemented as three separate of ships (fishing, small or large merchant) were programs and associated data files, all repre- given by histograms. sented in Figure 1. These data provide a basis for deriving the probability density function for the broadband source levels for each tpye of ship. First, for a given 1-Hz band, the source level for each RADIAD NOIS broadband class is obtained from the correspond- DATA FOR EACH S ing spectral density functions. Then, the prob- TYPE OF SHIP INFORMATION LOSS DATA ability of occurrence of each type of ship is / OTHER F ENCY DEPENDENT determined from the histogram showing the dis- INPUT INF tribution of the broadband class among the types MAION PROGRAM of ships. These data were than employed to con- TO COMPUTE struct discrete probability distributions of GEOMETRY DATA source levels for each type of ship. Finally, these discrete probabilities were distributed over 10 dB increments of source to obtain STORED OF GEOMETRY I J TRANSMISSION density functions for broadband source levels. DATA LOSS DATA For each ship, the total radiated intensity is the sum of those of the four components. For a ship selected at random from a population of a PROGRAM TO COMPUTE CHARACTERISTIC FUNCTION given type, the source intensity is a random � y variable that is the sum of the source intensi- Y ties of the four component random variables. Heine and Gray have concluded that the latter are mutually independent. The procedure for obtaining the source characteristic function for a given type of ship is outlined below. PROGRAM TO COMPUTE PROBABILITY DENSITY FUNCTION OF Y 1. Select a characteristic function with at least two parameters that Yy applies to a non-negative random variable. RESULTS 2. Calculate the low-order moments of the source intensity of each com- ponent. These can be readily obtained from the discontinuous FIGURE 1 Data Flow of Ambient Noise Model. density functions fk( ) for the 348 The input information is divided into three of the value of the density variable. Finally, types. Transmission loss data at a given sensor since the distribution of the ships on the sea location is read into the computer and stored an surface is stated in terms of ocean routes in- a file as a table of transmission loss as a func- stead of static densities in incremental areas tion of range and bearing. One such transmission facilitates modification of the programs for cal- loss file is created for each frequency of in- culations of second- [2, Sections 1.3 and 2.0] or terest. Sensor, route and ship traffic informa- higher-order statistical measures; e.g., autoco- tion, which is independent of frequency, is read variance of Y. into the computer. This frequency-independent information is used as the input to a program which computes such quantities as route segment References lengths, earth-centered angles between routes and sensor, and other geometric -parameters needed 1. J. C. Heine, L. M. Gray, "A Statistical Model in the computation of the characteristic function for the Acoustic Source Levels of Merchant (Dy~y). The outputs of this geometry program are Ships," Bolt Beranek and Newman Inc. Report stored on a file for later use as inputs to the No. 38i~4, July, 1978. characteristic function program. The characteristic function of beam noise 2. M. Moll, R. M. Zeskind, F. J. M. Sullivan, power is computed via the algorithm derived in "Statistical Measures of Ambient Noise: Section 2. This program reads in radiated noise Algorithms, Programs, and Predictions," data for each type of ship and other frequency- Bolt Beranek and Newman Inc. Report No. 3390, dependent input information such as beam pattern June, 1977. parameters. Beam steering angle and array orien- tation are also entered at this point. The characteristic function program reads in the data Acknowledgements stored in the geometry file and the appropriate transmission loss file. The program calculates The development of the ambient noise pre- the mean and variance of received average noise diction programs was supported by the Naval power as well as its characteristic function and Analysis Program, Code 431, Office of Naval Re- stores the results on a file to be used as input search, The Undersea Surveillance Division, Code to the final computer program. 320, Naval Electronic Systems Command, and the The final program computes the probability Acoustics Division, Code 8160, Naval Research density function of Y using the numerical algo- Laboratory. rithm described in Section 3. The mean, variance and probability density function for Y are printed out as the final result. The structure of the program provides several advantages. The geometric data need be computed only once for a given sensor location and route structure. Since it is independent of frequency and beam steering angle, the same geometry data can be used many times by the characteristic function program. Furthermore, the transmission loss data need be entered only once for a parti- cular sensor location. 6. Discussion Algorithms and programs have been described for predicting the probability density function for the averaged power Y of the noise output of a beamformer caused by ship traffic. The density is obtained from a calculation of the characteris- tic function. The programs are capable of utiliz- ing realistic input information regarding ship source characteristics, ocean routes and traffic, and acoustic transmission loss. The assumption that the number of ships of a given type on a given route is a Poisson random variable allows considerable simplification in the calculation of the characteristic function. The relevance of the Poisson variable to this application has been pre- viously examined [2, Appendix B]. The algorithm for the inverse transformation to obtain the prob- ability density function of Y permits the sampling of the characteristic function to be independent 349 A WIDEBAND SONAR EQUATION L. W. Brooks Binary Systems, Inc. P.O. Box 1105 Salida, Colo. 81201 Introduction frequency domain; thus to make quantita- tive use of these models it is easier to work in that domain. Of course, if The controlled impulse generator developed correctly, both time and fre- (CIG) is a technique developed by EDO quency domain representations will be Corp. for generating nearly arbitrary equivalent. pressure signatures in the far field of an electro-acoustic projector 11]. As sketched in Figure 1, the technique Target Response consists of driving the projector with an equalized voltage which compensates its electro-acoustic transfer function. Accordingly, let p Ct) be the time Using this approach, very short signals varying, on-axis, far-field, plane-wave (on the order of 150 wsec) can be pressure corrected to 1 yd and measured generated. Such short pulses have the in pPa. I.e. for a square pulse of potential to provide extremely good duration T, the source level would be: range resolution for advanced mine hunting, swimmer detection and torpedo T sonars. A correct evaluation of the SL = (log) CIG sonar performance must include the fact that the sonar environment is 0 frequency dependent; thus the CIG pulse Next, let T denote the complex becomes increasingly distorted as a Fourier transform of PT(t) Thus; function of range. Urick 12] gives a very simple P = p te2rTft dt (2) technique for evaluating transient T jT signals by replacing the source level in the sonar equation with a more gen- In direct analogy to the usual sonar eral term which is the transmitted equation, the complex Fourier transform energy flex density divided by the dura- of the pressure at the receive 'ransducer tion of the echo. Unfortunately, Urick is; does not give a method for computing the echo length. In this paper, a wideband vT(epp1f) 27TifT sonar equation is developed which can V'f = 2 1 i f be used to estimate the pulse distortion R = r210' over modest ranges. Expressions will be given for the receiver response to (3) the target, ambient noise, self noise, and reverberation. In Equation (3), V (B,p,f) is the one-way complex voltage paTtern of the transducer, it is possible to develop a wide- a is the propagation attenuation, and p(f) band sonar equation in the time domain. is the Fourier transform of the target In this approach, the receiver output is impulse response for a stationary target. expressed as the five-fold convolution We assume p(f) is normalized so that the of the transmit waveform, the environ- target strength is given by: mental impulse response, the target impulse response, the receive transducer 2 impulse response and the receiver filter TS = 101ogfp(f) (4) impulse response. However, most sonar models (propagation loss, target strength, etc., see Urick [2]) are expressed in the transform of (3Y' that is the time representation of the pressure pulse 350 at the transducer for a stationary We note that if the target is a target. We assume that the effect of point target, the hydrophone response is a moving target is to produce a time white, target is on the beam axis and the scaling of the received pulse. That is, absorbtion is negligible, (9) becomes following Deley [3], the received pres- proportional to a wideband cross ambigu- sure pulse is written, ity function between the p(t) and w(t) (the inverse transform of W(f)) 13]. p(t) = a P (t-T(t)) Equation (9) is general and applies PR) = Pr tR to any waveform. However, here we are interested in baseband CIG pulses whose -~ p~~o((t-T)S ) ~(5) - B PR,~~ ((inshape is similar to that sketched in Figure 1. Such short pulses have essen- (t-T) ithsinlrtnetially no capability to resolve time where PR (-)is the signal returned scale differences (i.e. Doppler). Thus by a stationary target, T(t) is the (9) can be simplified to (9) can be simplified to: range delay of that part of the signal returning at time t, and B is the time scale factor given by: Z(T-To) =fP(f)TE(f)W*(f)e27if(T-To)df _0~~~~ 2V ~~~~-o (10) = 1 - T(t) 1 (6) c where TE(f) is the "environmental trans- fer function": In (6), V is the radial target velocity. Note that (5) is normalized so that the VT(8,1,f)VR(O,4,f)P(f) received pulse energy is independent of TE(f) = 2 lr(11) the motion. In the frequency domain (5) r 210'lr is written: In (10), the peak signal level p (f) = P2f/') V (0,1,f/O) occurs when l= T. and She peak output R Or2 10'lr T signal power is iZ(0) l p (f/s) e2ifT (7) Response to Ambient Noise Now, if R(f) represents the complex, on-axis transfer function of the receive The ambient and self noise response transducer and VR(0,4,f) is its complex may be derived by modeling the noise beam pattern, the Fourier transform of field as a collection of independent the electrical signal at the transducer plane waves arriving at all possible output is: angles and frequencies. In this case, one obtains the standard result for PUM/~) the power spectral of the noise at the Pe(f) 2p lar VTO (2,,f/)VR( T Rf) receiver transducer terminals as, [2], Sr210.r 2,ffifT p(f/$)R(f)e (8) No (f ) 2 N (f) = - R(f) (12) AG(f) Finally, if the receiver is matched to a spectral shape W(f) with a time scale factor O., and a round trip delay where No(f) is the noise spectral density To, the output is given by: at the output of an isotropic hydrophone and AG(f) is the receive transducer noise gain (see [2] pg. 32). Thus the average Z(TTo;S,~o) noise power at the receiver output is given by: VT(6,0,f/O)VR( , ,f) =P P(f/O) T lr2 T 2lOo lr1 N =] [G(f) IR(f)121W(f)12 df = O j~tf/6) 210.1ar N =| JR(f)| |W~f)| df (13) -~ * 2urif'' J AG(f) �P(f/O)W (f/SO)e TQdf - (9) 351 Reverberation included here and can be found in [5]. The result is: Reverberation is modeled as scatter- 1 ing from a collection of similar P P(f)R(f)W f) 2 "targets" having independent, random S rR f amplitudes, phases and positions. This model is a special case of the one used Cr by Middleton 14]. Strictly speaking, Ss (f)2co (fs )df since the scatterers are not geometri- s cally identical, the impulse response should be considered a random function; (16) however, it is adequate for our purposes to model the scatterers as having where Ss(f) is the surface scattering identical impulse response functions coefficient versus frequency, e is the except for amplitude and phase. With elevation angle to the surface, and this assumption we have that the �(f,e ) is the equivalent two-way azimuth reverberation power at the receiver beam pattern defined by: output can be written [5], T 2 (fs) =f vT( ,p,f)vR(eI ,f)2 d PR(TO) E ElakZR(Tk-O)1 (14) - (17) all scatters Evaluation for a Torpedo Application wpre a is the complex amplitude of the k scattering point, T is its position, and Z is given by Equation (10), eval- The key equations which give the adZR isgvnby Equation (10), eval- uated for the transform of the reverber- target, noise, and reverberation levels ation impulse response, i.e. for p (f). at the receiver output (Eqs. (9), (13), The above equation may be evaluated by (16)) must be evaluated numerically; counting the average number of scatters however, this is not difficult since at each delay, azimuth, and elevation; only a single variable of integration thus, is involved. In our particular applica- tion a nine-point Gaussian quadrature formula was found to be adequate. ( To ) =fff|ZR(I - To ) | | a|n (T ,8,~)dTded~ P' JJJ[R0 a nT dTded1 The example chosen concerned a 1 ft diameter aperture, a 150 psT (15) pulse, and a peak pressure of 10 IPa (corrected to 1 yd). Figure 2 shows 12 a typical CIG spectrum, and Figure 3 where al is the average amplitude shows the effect of increasing range squared of the scattering points and on the received signal spectrum (at n is the average number per unit delay, the transducer). For that calculation, azimuth, and elevation. the target was assumed to have a constant target strength with frequency of 2 dB, thus the major effect is Equation (15) while reasonably attenuation which reduces the frequency general, is not particularly useful at which the peak energy occurs as the since it involves a five-fold iterated range increases. integral. It can be simplified by noting that a) for the short CIG pulses Figures 3 and 4 show the results IZ (T-To) I is approximately a Dirac of numerically evaluating (9), (13) delta function allowing the integration and (16) for a square pass receiver. over T to be evaluated, b) for surface The ambient noise model versus fre- or bottom reverberation n(T,e,O) is a quency, No(f) was taken from Urick [2]. Dirac delta function in elevation angle The torpedo self noise was assumed to 0, allowing the integration over 0 to be equivalent to sea state 2 at search be evaluated, and c) with the two previ- speed and to increase 15-20 dB at ous assumptions the only remaining attack speed. The model for surface functions which depend on azimuth angle scattering strength versus frequency are the transmit and receive beam required to evaluate (16) was taken patterns, thus the integral over 0 can from Audreyeva (see 15]). be replaced by an equivalent beamwidth. The details are too lengthy to be 352 Note that for such a short pulse, the sonar is ambient or self noise limited at all ranges of interest. Also, the signal-to-noise ratio is bet- ter in the higher frequency band (Fig. 4) than in the lower one, even though the lower band corresponds to maximum received signal energy at 3 kyd Csee 160- 12 Fig. 2). The reason, of course, is P=1 pPa that ambient and self noise are lower 1 in the higher frequency band, and m 140 T=lOpsec enough so to compensate the loss in signal energy. 120- References 4jU 00) w c� 100- [1] C. Mazzola "Gen. of Wideband I Pulses ..." EDO Rept. 10991, W EDO Corp., College Pt. N.Y. 11356, 80- July 7, 1976. s / [2] R. J. Urick Principles of Underwater 60 , , , Sound, McGraw Hill, 1975. 0.1 1.0 10 100 0.1 1.0 10 100 [3] G. W. Deley "Waveform Design", in Freq. in kHz Radar Handbook, M. I. Skolnik ed. McGraw Hill, 1970. Fig. 2. CIG Pulse Energy Spectrum (Measured at 1 yd) 14] D. Middleton, "A Statistical Theory of Reverberation, Part I and II" IEEE Trans. Vol IT 13, pp 372-414, 1967. [5] L. Brooks, L. Palmer, and F. Rees, "Wide Band Transient Detection" BSI Rept WB77-12 Binary Systems, Inc. 10750 Columbia Pike, Silver Spring, Md. 20901, Dec. 30, 1977. 43 40- 12 H P1 =10 pPa T=150 sec 20-TS=2 dBd H m/ Equalizing Projector � Filter D4 3 -20- // \R=.75Kyd s-i -20- e (t[ Wf t ~ -U 40- / /R=l.5Kyd _ a,-40-/ 70 ~eCt)~ C)> ~ _a-60- R=3.OKyd A___ -80 Input Equalized Far Field 0.1 1.0 10 100 Signal Voltage Pressure Freq. in kHz Fig. 3. Received Energy Spectral Density for a Specular Target Figure 1. The CIG technique 353 70 Receiver Band: 7-14 kHz Receiver Band: 12-19 kHz 60- m NkTarget \ Target TS2 dB TS=2 dB '~ 50 Self Noise -1~~u~~ ~Self Noise ttack Speed \ Self Noise P4. -ttack Speed 40- > Ambient SS3 U 30 ,, Ambient SS3 30' 20- ~ Self Noise f @ 20- \ Search Speed\ )O~~~~~ /1'~~~~ \ Self Noise Surface earch Speed Reverberation Surface 10- V = 6-10 m/su Reverberation V = 6-10 m/s 0 1.0 2.0 3.00 1.0 2.0 3.0 Range in Kyd. Figure 4. Received Power Levels Versus Range 354 SYNTHETIC APERTURE IMAGING USING ACOUSTICAL HOLOGRAPHY W. S. Gan Acoustical Services Pte Ltd 29, Telok Ayer Street Singapore 1 Republic of Singapore Abstract there is a single point reflector located at (x ,yr) and that the transmitted sig- We develop the mathematicalven basisy for synthetic aperture imaging using acoustical holography in a form suitable s(t) = A sin(( t). w(t) for application to underwater viewing. s(t) = Ao sn(o t) w(t) Synthetic aperture technique by itself is a high resolution ultrasonic imaging sy- w(t) = 1, 0 : t - T stem. By using acoustical holography we can improve the resolution even further. We preform demodulation and low band-pass otherwise (1) filtering on the holographic signal. where is the center frequency of the This is followed by heterodyne detection w on the signal. Then we have the super- ultrasound pulse position of complex images using two fre- T is the pulse duration, which is quencies or double exposure acoustical holographic interferometry. We then com- normally an integral number of half cy- cles. pare the resolution obtained after applying cles. The reflected signal is a function of the holographic technique with that before applying. We finally outline a system to transducer position and is given by test the above principle. r(x (xTtt) = Ai sino(t - .t -Tr ) 1. Introduction dr(XT) =(xr xT)2 + Y' r T) Synthetic aperture technijup is first used in microwave(radar) but has r 2.dr(xT) (2) been used in sonar systems as an effec- c tive ~eans to get high azimuth resolu- tion. In this paper we apply acoustical where c is the velocity of sound holography to a synthetic aperture system. A/A is the attenuation due to ab- The purpose is to improve the resolution sorption, scattering cross-section, and in the range direction and to achieve su- r nsc erinecin perresolution2 We make use of the method The holographinsducer inefficien bcy. of Sato et al and apply it to the syn- The holographic signal is given by thStic aperture ~ystem of M. L. Dick et h = s(t) + (3) al . Sato et al 's method is based on h(xTt) = s(t) + r(xT,t) (3) the utilization of superposition of com- plex images obtained by multiple frequen- For demodulation, we make use of both cies of waves. The superposition is per- quadrature components. So demodulation formed after the nonlinear heterodyne o- yields peration. Uc(xTvt) = h(xT,t) . Cos( Wt) 2. Demodulation and Low- pass Filtering = I si O(2t - t - sin We assume that a small transducer moves in a straight line parallel to the x-axis. The y-axis is vertical with the trasnducer located at y = 0 and pointing (0 r)- w(t - tr) + A parallel to the y-axis. The transducer - at any time is located at (XT,O). Assume 355 might be obtained when a very short dura- sin(2 10ot) . w(t) tion of pulses are used can be obtained by using actual pulses with a relatively long us(xT,t) = r(xT,t) sin(wot) duration. Su(xT t) = r(xT9t) * sin(JOt) We first use two frequencies for the interferometrical holography and we let = A1I cos ( 0(2t r + cos the two frequencies used be I)l and O2' 2 ( (4) Then our eqn. (5) becomes w(t + l(XTt) = h(XTt). cosUlt . e.Jlt -jo(t) + h(xT,t). cosD)2t. e4J2t+ jo(t) (coS2(iot - 1). w(t) usl(xT,t) = h(xT,t). sinTl)t e. wlt-j(t) This is followed by low-pass filtering to remove the sum frequency from both Uc(xT + h(xt) sin e (t) t) and us(xT,t). This filtering yields) .in2t (6) Vc'(XTt) = - A1 sin(Uor) . wl'(t -r) The next step is to perform heterodyne pro- cessing on (6). Then we have + Ao sin(2oJt). wl'(t) Ccl(xTt) = U cl(xT,t) s*(t)> T (7) 7- c v S(xTt) = Al cos(W o ). w 1'(t - ) where Tc is the integrating time. Eqn.(7) 8s TO 7I r I r contains the two-frequencies acoustical ho- lographic interference fringes. t We now use four frequencies for the *io A(coos - 1). W1\ interferometrical holography and we let the 2~ (5) two frequencies used be 4C1, W&21 LJ3 and where wl'(t 4 )r) and wl'(t) are modifica- ) 4. Then our eqn. (5) becomes tions of the delayed original window, w(t) Uc2(xTt = h(x ,t) . cos( t. ej1t e (t) and the original window, due to the low- pass filter. ~~~~~~pass filter. ~+ h(x ,t). cos t. eJW2t+j (t) 3. Utilization of Multiple Frequencies T 2 of Waves + h(x T,t). cos 3t. eJW3t-j (t) When the synthetic aperture technique is used, the azimuth resolution is im- j W t+ j9(t) proved as much as we desire by increasing + h(xt).cos 4t ej 4 (t) the aperture to be synthesized. On the other hand, the resolution in range direc-t) = h(XTt) .sint-j (t) tion is left to the utilization of very us(x short pulsed waves.When the pulse width is reduced extremely to a width as narrow as several wave lengths, the detection of the + h(xTt). sinW2t. e2t phase and amplitude of the reflected waves becomes very difficult. This is the re- + h(x . sint-j (t) sult of the decrease of the resistance to Tt). s 3 noises, besides the difficulty of genera- tion of complete signals of such a short. ,W t+ jQ(t) duration. We now make use of multi-fre- + h(xT,). sin4t. e 4 quencies acoustical holographic interfero- metry to get high range resolution without decreasing the pulse width. The idea of adecreasing the pulse width. The idea of The next step is to perform heterodyne pro- acoustical holographic interferometry is cessing on (). Then we have the linear superposition of complex images obtained by multiple frequencies of waves. C (x And it can be shown that by synthesizing cc2 2(T,t) s*(t) T (9)c the spectrum of multiple frequencies(spec- tral synthesis) of waves, the images which Equation (9) contains the four-frequencies 356 acoustical holographic interference frin- ges. For both the two-frequencies and the four-frequencies cases, the reconstructed images are given by the Fourier transfor- mation of eqns. (7) and (9). 4. Comparison of System Resolution before and after the Application of Holography It is very difficult to show mathema- tically the amount of improvement in reso- lution due to the application of multiple frequencies acoustical holographic inter- ferometry. Qualitatively, by applying ho- lography to a synthetic aperture sonar sy- stem, we are combining aperture synthesis of synthetic aperture technique with spec- tral synthesis of multiple frequencies ho- lography to obtain superresolution. The improvement in resolution is in range re- solution due to the utilization of very short pulsed waves when the pulse width is reduced to a width as narrow as several wavelengths. 5. System Development 2 We shall follow the work of Sato et al for the design of our system but with some modifications. This will be to in- clude in electronic instrumentation for demodulation and low-pass filtering works. References 1. Cutrona, I.J., Vivian, W.E., leith, E. N., Hall, G.O., "A high-resolution ra- dar combat-surveillance system," IRE Transactions on Military Electronics, MIL-6, 19I6T, p.-II-9 2. Sato, T., Ikeda, 0., Ohshima, H., and Fujikura, H., "A few effective prepro- cessings in synthetic aperture sonar system," Acoustical Holography, Vol.7, 1977, pp.5697 .-.-. ... 3. Dick, M.L., Dick, D.E., McLeod, F.D., and Kindig, N.B., "Ultrasonic Synthetic Aperture Imaging," Acoustical Hologra- phy, Vol.7, 1977, pp.377...... 357 ENVIRONMENTAL CONSIDERATIONS FOR SITING AN OCEAN THERMAL ENERGY CONVERSION EARLY OCEAN TESTING PLATFORM AT FOUR PROPOSED AREAS M.D. Sands Pat Wilde Interstate Electronics Corporation Lawrence Berkeley Laboratories Oceanic Engineering Division University of California Anaheim, California 92803 Berkeley, California 94720 Abstract discharge, effects of impingement and entrainment, biocide effectiveness and This presentation summarizes the environmental toxicity to nontarget biota, working fluid considerations for siting OTEC-1, the Early release effects, and worker safety. Also, the Ocean Testing Platform. As you may know, the international, Federal, state, and local legal concept of Ocean Thermal Energy Conversion implications of siting must be considered. (OTEC) is not new to the alternative energy When all relevant data is at hand the source arena. d'Arsonval first proposed the predictive process for assessing environmental use of the thermal gradient between warm impact is underway. surface and deep cold ocean waters in the late nineteenth century. In the 1950's, the French government considered OTEC, but, after 1. Introduction developing some equipment, abandoned the program prior to testing. The Oceanic Engineering Division of Interstate Electronics Corporation is under contract to With the oil embargo in the early 1970's and the Department of Energy (DOE) to prepare the the resulting oil price increases, the United Environmental Impact Assessment for the States Federal Government, in 1975, esta- 1-megawatt early ocean test platform. In the blished the Energy Research and Development EIA, the issues of safety and health, legality Administration to conduct energy-related of siting and site specific impacts are under research. Presently there are several solar evaluation. The potential impacts under study research projects under evaluation, one of include toxic element effects, biocide release which is OTEC, and the Early Ocean Testing impacts, thermal mixing effects, point and Platform for a prototype evaluation has a nonpoint source outfall, impingment/entrain- projected test date in early 1980. The ment and recommendations for OTEC-1 environ- ultimate use of OTEC may be to provide mental and system monitoring. Briefly, this commercial electricity generation in place of presentation will summarize the history of the use of fossil or nuclear fuels, or to OTEC, OTEC-1 design aspects, the study sites, create floating plant-ship operations for and report some of the progress to date. industrial production of aluminum or fertilizers. 2. OTEC History As a result of the implementation of the The OTEC concept uses the thermal gradient National Environmental Policy Act of 1969, all that naturally exists in the oceans to government-funded activities must consider condense and evaporate a working fluid. The potential environmental consequences of minimal temperature difference required is activities and prepare an Environmental Impact approximately 20�C which limits the Assessment (EIA). If significant impacts may geographical region that can be used, to occur, an Environmental Impact Statement (EIS) between 35 degrees north and south latitudes. must be prepared bringing environmental considerations into the decision-making Ocean Thermal Energy Conversion was first process. The considerations in assessing proposed by d'Arsonval in the late 19th impact for OTEC-1 first require a detailed century, but the technology available at that description of the physical system design. time was inadequate to develop the concept Included in the design description are the further. Georges Claude, between 1929-1930, depth of intake and discharge pipes, volumes attempted to use an open cycle pilot plant for discharged, and applicable safety regulations electrical power production. Before the cold and procedures. The detailed site descriptive water pipe was destroyed, Claude did get a information including the biological, chem- small power output and prove the concept. ical, physical oceanographic, and meteor- Later, in the 1950's, the French government ological data must be gathered from all supported a partially government owned company available sources. Particular study areas to construct an OTEC plant on the African include the effects of metallic element Ivory Coast. While several subsystems were discharge, ocean water thermal mixing and designed, built and tested, the work was never 358 completed. J. Hulbert Anderson and J. H. Biofouling Control: The prevention of Anderson, Jr. first proposed the construction microfouling on the heat exchanger and of a 100 megawatt plant using the closed cycle condensers surface is a critical aspect for approach in 1966.(1) OTEC-1 operation. Buildup on the order of 50 microns will reduce the heat duty 15 to 25 OTEC was one of the six solar technologies percent.(5) To restrict this buildup, selected by the National Science Foundation's mechanical methods (Amertap and M.A.N. Research Applied to National Needs (RANN) brushes) in conjunction with chemical methods program in the early 70's. In January 1975, (chlorine) will be used for biofouling the Energy Research and Development Agency control.(4) (now the Department of Energy) was formed and 4. OTEC-1 Siting Criteria assumed the lead role for program development. The site that is selected for OTEC-1 testing Since the initial budget allocation of $85,000 must be located within 200 nautical miles off in fiscal year 1972, the program has grown the United States and its properties.(6) The dramatically to the present funding level for minimum temperature differential must be fiscal year 1978 in excess of 35.3 million greater than or equal to 200C for five dollars.(3) continuous months with an overall annual temperature spread of 16�C minimum. The maximum allowable currents shear or velocity 3. OTEC-1 Physical Description gradient must not exceed 1 knot in 50 meters vertically and 1 knot in 200 meters Two types of systems can be used for OTEC horizontally. Additionally, the water sea power generation--the open or closed cycles. state cannot exceed Beaufort No. 4 for 9 The open cycle utilizes seawater as the months continuous.(6) working fluid where low-pressure steam is evaporated from the warm-water flow. The In terms of extreme conditions, OTEC-1 will closed cycle employs a working fluid with good operate in seas with a significant wave height heat transfer characteristics such as ammonia, of 4.9 meters, 62 cm/sec surface currents and freon or propane. The Rankine power cycle 15 meters/sec winds. Both the platform and the includes a pump, condenser, turbine and retracted, detached or modified deployed cold evaporator. The working fluid is converted to water pipe will survive 12.5 significant wave a vapor (power is generated by expansion heights, 1.0 meter/sec surface currents and through the turbine) and condensed by the cold winds of 51 meters/sec. water at the condensers. OTEC-1 is not scheduled to have a turbine, or produce a net The study sites designated for this study by energy gain, but to demonstrate and evaluate the Department of Energy include Keahole various technological concepts.(4) Point, Hawaii; Punta Tuna, Puerto Rico; New Orleans, La; and the West Coast of Florida. While the platform for the 1-megawatt plant Keahole Point is approximately 3.9 kilometers has not been selected, the physical operations offshore in 960 meters of water. The bottom specifications have been outlined. Presently, sediments are basically volcanic sands and two concepts have been suggested to supply the silts. The mean flow through the site is cold water resource to the platform: a between 10-30 cm/sec. The vertical delta T platform-supported sectional pipe string over the year ranges from 19.8 to 22.50C.(7) (Figure la) or a separately deployed and positioned spar buoy/riser pipe (Figure lb) Punta Tuna is approximately 4.6 kilometers which interfaces with the platform through a offshore in 1,200 meters of water. The bottom "quick-disconnect" umbilical. The disconnect sediments are predominately silts and sands mechanism will separate the buoy from the high in carbonates. The mean flow through the platform in the event of an emergency. site is also 10-30 cm/sec; vertical delta T over the year ranges from 20.9 to 23.9�C.(7) Water Supply: The assumption was made that the cold water pipe will deliver a seawater The New Orleans site is approximately 89 flow from the 1000 meter depth to the kilometers off the southeast pass and 138 condensers at a flow rate of 257,000 liters kilometers off Mobile Point, Ala. The bottom per minute (68,000 gal/min). The warm water sediments are primarily terrigenous silts and pumps will deliver 227,000 liters per minute clays. The annual mean flow through the site (60,000 gal/min) from the surface waters. is from 35 cm/sec to variable. The vertical delta T over the year ranges from 16.6 to Discharge: The warm and cold water may be 25.00C.(7) mixed and discharged through a common header with extensions over the stern at a depth of The west coast of Florida site is approxi- 40 meters. The combined flow of 484,000 mately 250 kilometers off Cape San Blas and liters per minute will be discharged at a rate 272 kilometers off St. Petersburg. The bottom of 308 cm/sec. The method of discharge sediments are composed primarily of silts and presently is undefined. In this evaluation, siltey clays. The flow through the study site both a horizontal and vertical discharge were is 100 cm/sec and variable. The vertical considered. delta T range for the year is 19.7 to 15.0�C.(8) 359 PLATFORM SUPPORTED MOORED SPAR BUOY BARGE STRUCTURE BARGE STRUCTURE WARM WATER R esidue INTAKE MIXED OUTFALL MIXED OUTFALL (40m) fi COLD WATER RISER COLD WATER RISER 1000m 10 (a) ~~~~~~~~~~~~~~~~(b) Figure 1. Alternative Design Concepts for Cold Water Residue 5. Environmental Considerations It is recommended that a similar study, specifically applicable to OTEC-l, be The considerations for siting OTEC-1 that must undertaken that will lay the groundwork for be addressed include health and safety on the the early ocean test platforms. platform as well as in the surrounding vicinity, legal or licensing issues of the The Federal agency involvement in licensing is platform, and the potential environmental an important siting consideration. Research consequences that may result from site beyond the territorial sea has been subject to occupation. control by the Federal government. Because of the large number of agencies involved in Health and Safety: The safety aspects for licensing at the Federal and State level, OTEC-1 are similar to various operational Knight et al. (1977) suggested the designation procedures inherent in most ocean-going of a lead agency to coordinate the various vessels. In preparing an evaluation of these approvals required from the administrative issues, at least four department level bodies involved. agencies and fourteen non-governmental organizations will be involved in an advisory If designation of a lead agency could be made, or regulatory function.(4,8,9) it would reduce the lead time required to receive permits and would lay the groundwork Legal Considerations: Several complex for future OTEC activities. Congressional national and international legal ramifications approval is required for agency designation. surrounding OTEC-1 operations have been Candidate agencies include (but are not discussed by Knight, Nyhart and Stein (1977) limited to): Army Corps of Engineers, U.S. who published a comprehensive discussion of Coast Guard, or the Department of Commerce the legal, political, and institutional (principally NOAA).(11) aspects of Ocean Thermal Energy Conversion (under the auspices of the American Society of Environmental Consequences: The main International Law).(10) Their report covered consideration for the assessment of potential many of the complex issues of OTEC in general. environmental impacts are dependent upon the 360 near field dispersion of the plume. Once Nutrient Redistribution isopleths are prepared for each contaminant introduced to the environment, applicable OTEC-1 site occupation is unlikely to toxicity data for resident organisms can be significantly increase primary productivity compared. near the site. After mixing with the surface waters and discharge at the outfall, most of Preliminary evidence obtained from mathe- the nutrients will be transported to matical model predictions and expected density approximately 90 meters depth where light is gradients indicate that the 1-megawatt expected to be the limiting factor. horizontal mixed discharge water would mix in the water column 50 meters below the discharge Biocide Release pipe.(ll,l2) It is expected that background diurnal fluctuations of up to 1oC would Although chlorine has been specificied for obscure much of the thermal impact at the OTEC-1, the method or rate of application is bottom of the surface mixed layer with the not known at this time. The EPA N.P.D.E.S. major impact observed within a few hundred Chlorine regulations for cooling water meters of the horizontally discharging jet. restrict the release of chlorinated waters to no more than 2 hours per day with a 30-day If discharged vertically, an entirely dif- average not to exceed 0.2 mg/l. It is ferent plume behavior will prevail, expected that the main influence of release Essentially the discharged mix will be will be on the invertebrates and eggs and injected into the top of the main thermocline. larvae stage of fish. The toxicity or Given an initial momentum effect, the bioassay data available for the oceanic discharging jet will reach approximately 60 species at these sites is very limited. meters below the discharge pipe before Studies are underway to further evaluate this significant entrainment, thus effectively release. injecting the discharge into the upper portion of the main thermocline. This would, of Impingement/Entrainment course, minimize feedback to the warm water intake, but put the discharge at or below the OTEC-1 will circulate millions of liters per surface mixed-layer depth at each of the four day and organisms that reside in the region sites during summer conditions.(l2) will be subjected to impingement and entrainment effects. The most serious The environmental impacts that are under impingement problem is expected to be with evaluation include the impacts that may result those fish which have the lowest avoidance from trace element leaching from the heat capabilities, the coelenterates. OrganLsms exchangers, condensers and the platform entrained in the cold water pipe will suffer surface, the effects of biocide release on 100 percent mortality as a result of the target and nontarget biota, the effects of pressure and temperature change and physical nutrient redistribution to the photic zone and abuse, while the warm water intake mortality the effects of organism impingement on the should be much less. (14) Field programs intake screens and the entrainment of underway at the OTEC regions will provide the organisms in the system. (13) qualitative and quantitative biological data necessary to further evaluate these effects. Trace Element Release The heat exchangers, pump impellers, metallic Acknowledgement piping and bottom paints are sources of toxic metallic ions. The pump impellers and This OTEC study is funded by the United States metallic piping are minor sources; the Department of Energy under contract number greatest release will come from the heat EG-77-C-06-1033. exchangers and the bottom paints. The heat exchanger materials likely to be used on OTEC-1 are aluminum and titanium. These References materials are relatively nontoxic in the marine environment. (1) Griffin, Owen M. August 1977. POWER FROM THE OCEANS' THERMAL GRADIENTS. Sea The Hughes Mining Barge, a proposed OTEC-1 Technology Volume 18, No. 8 p. 11, platform, is painted with the antifouling paint Kelkote 109AF. The active agent in the (2) W.S. Deptartment of Energy, March 1978. paint is cuprous oxide and is applied at a Environmental Development Plan Ocean thickness of 6 mils. The amount of release Thermal Energy Conversion. DOE/EDP/0006 from this surface will vary with the age of (Available NTIS) p. 48. the coatings and rate of flow past the hull, The release of 1 kilogram copper would result (3) Mulcahy, Michael. Aug. 1977. OTEC - FROM in a2 loading to the marine environment of 2 85,000 TO 35 MILLION IN SIX YEARS, Sea mg/m or 0.2 mg/l, assuming the copper is Technology Volume 18, No. 8, p. 16. confined to a 1-cm film over the surface. (14) 361 (4) Sands, M.D. February 1978, "Progress Report for the Environmental Impact Assessment Program for the 1-Megawatt Early Ocean Test Platform", to appear in Proceedings Fifth Anual Conference on Ocean Thermal Energy Conversion, TN. Veziroglu, Editor; CERI/UM. (5) Bell, Kenneth J. October 1977. The Effect of Fouling Upon OTEC Heat Exchanger Design, Construction and Operation. Presented at the Biofouling and Corrosion Symposium, coordinated by Battelle Northwest Laboratories and sponsored by the Energy Research and Development Agency. (6) US. Department of Energy, Request for Proposals for System Integration Contractor for OTEC-1, received January. 1978. (7) OTEC-1 Site Environmental Profile Enclosure A (Revision, January, 1978) DOE Working Document. (8) United States Coast Guard. 1973. Rules and Regulations for Cargo and Miscellaneous Vessels, subchapter 1. Department of Transportation, Washington, D.C. 179 pp. (9) Occupational Safety and Health Administration. 1976. General Industry Safety and Health Standards, Department of Labor, Washington, D.C. 649 pp. (10) Knight, G., Nyhart, J.D., Stein, R.E. 1977. Ocean Thermal Energy Conversion, Lexington Books, Lexington, MA. 230 pp. (11) Personal communication, K. Bathen, 31 January 1978. (12) Bathen, K. 1975. An Evaluation of Oceanographic and Socio-Economic Aspects of a Nearshore Ocean Thermal Energy Conversion Pilot Plant in Subtropical Hawaiian Waters. Department of Ocean Engineering, University of Hawaii, Honolulu, Hawaii. 271 pp. (13) Lewis, L. (ed.) Proceedings, OTEC Environmental and Resource Assessment Workshop, June, 1977. (14) U.S. Department of Energy. (in press.) ENVIRONMENTAL IMPACT ASSESSMENT FOR OTEC-1, Interstate Electronics Corporation, Anaheim, California. 362 TECHNOLOGY CONSIDERATIONS IN THE DESIGN OF A COMMERCIAL OFFSHORE ENERGY CONVERSION (OTEC) PLANT William W. Rogalaki, Jr. and Robert J. Scott Dr. Julio G. Giannotti Gibbs & Cox, Inc. Giannotti & Buck Associates, Inc. 2341 Jefferson Davis Highway Crestwood Building, Suite 40O2 Arlington, Virginia 22202 5711 Sarvis Avenue Riverdale, Maryland 20840 Abstract The hull forms considered in the optimization process consisted of six generic shapes - ship or The design and construction of an OTEC plant barge, vertical cylinder or disc, spar, submersible, slated to produce electricity for commercial con- semi-submersible, and sphere. In addition, a range sumption in the near future presents a number of of electric power plant outputs was considered, in unique problems and is forcing the advancement of this case, 50 to 500 megawatts (MW), to bracket what the state of the art in ocean structures in a number was considered at the time a reasonable variation in of f ields. The successful deployment of a oem- size. The upper limit of 500 MW was chosen because mercially viable OTEC plant is dependent upon ident- it was felt that based on an initial assessment of ifying these technology areas and developing a pro- the state-of-the-art (SOA) in large ocean platform gram to insure that any potential program obstacles construction, deployment, and operation it repre- F ~~~are adequately addressed. sented the upper limit with respect to size, i.e., approximately one million long tons. The evaluation process utilized a figure of merit (FOM) based on an economic benefit to cost relationship and resulted 1. Introduction in the final selection of a J400 MW ship and 400 MW semi-submersible for further development. At the The concept of OTEC utilizing thermal differ- same time, Lockheed Missiles and Space Company and entials in the ocean to generate electrical power M. Rosenblatt and Son, Inc., who were engaged in appears to be one of the more viable methods of parallel efforts, each made their recommendations as 6 ~~~ transforming solar power into usable forms; how- to size and configuration, and were asked to develop ever, a number of ocean engineering related problems further the ship and spar (Lockheed), and the sphere need to be overcome to reduce the risks inherent and spar (Rosenblatt). with the concept to acceptable levels. This paper will discuss some of the higher risk aspects asso- The ship and semi-submersible designs by Gibbs ciated in the development of the platform and cold & Cox, Inc. underwent a series of iterations until water pipe (CWP) for an OTEC commercial power plant they converged to the point where feasibility was that utilizes a closed Rankine cycle power system. demonstrated and first-cut cost estimates could be made. Table 1 contains the principal character- The material discussed herein is based on a istics of both the ship and semi-submersible while study done by Gibbs & Cox, Inc. for the U. S. Figure 1 is an inboard profile of the ship. Department of Energy (Reference (1)). This study, one of three parallel efforts, had as its objectives Table I the selection of the optimum plant output and hull OTEC Commercial Plant configuration for a Commercial OTEC plant and the Principal Characteristics subsequent conceptual design of the two leading candidates. During this process a number of tech- Ship Semi-Submersible nology areas requiring further development were identified, the more significant of which are Length (ft.) 620 436 addressed in the subsequent sections of this paper. Beam (ft.) 300 436 Depth (ft.) 85 328 2. OTEC Commercial Plant Concepts Draft (ft.) 145 215 Displacement 311,000 451,000 To establish a point of reference for the (full load)* discussion of technology requirements, a brief Displacement 126,000 139,000 outline of the baseline designs used in the hull (light ship)* selection process mentioned in the introduction Nominal Power Output 400 400 along with the two final conceptual designs will be (MW) presented in this section. * Displacements are given in long tons; I long ton =2,240 lbs. 363 LEGEND Size and Configuration. Perhaps one of the E Evaporator M Pump Motor most intimidating aspects of the OTEC commercial C Condenser T Turbine CWP Cold Water Pipe G Generator plants discussed in the previous section is their WP Warm Water Pump PC Pump Conversion size. As can be seen from comparing the displace- CP Cold Water Pump ments in Table 1 with those in Table 2 the proposed OTEC plants are structually more massive than any other floating platform, and this fact combined with their unusual configurations (especially when one _~ ~ 'M Tconsiders the 3000 foot long 100 foot diameter CWP ~ -w~-~ ~ LiJ i , hanging below the platforms) places them in the :~- r r forefront of the SOA in ocean structures. To fur- Zr C :'. C ther emphasize this point Figure 2 compares the C ~ profiles of the largest existing semi-submersible (ODECO's Ocean Ranger) and the Gibbs & Cox, Inc. semi-submersible. Table 2 - so 106 Current SOA in Floating Structures Type Dimensions Displacement Figure 1 400 MW OTEC Plant, Ship Configuration (Length, Beam, (Light Ship) Inboard Profile Draft) The ship is basically a barge-like structure Tanker 1360 x 261 x 145 78,000 divided into quadrants, with each quadrant contain- (Batillus) ing two 50 MW power plant modules. Drill Rigs 399 x 262 x 152 28,000 Warm water is taken in through screens by the CSemi- Submersible) eight warm water pumps, is pumped to eight evapor- ators (one per module) mounted veritoally below the Oil Production hull, and is discharged back to the ocean at or near Platform 100 diam. 445 high 20,000 the bottom of the mixed layer through individual (Moored Spar) discharge pipes (one per heat exchanger). The ammonia vapor is piped to demisters and then to the turbines which are directly coupled to the gener- ators. The turbine exhausts discharge to eight condensers which receive their cooling water via the cold water pipe and cord water pumps, and the con- densate is then pumped back to the evaporators. Each of the eight power modules is totally indepen- dent and cannot be cross connected. Power con- - version equipment, control spaces, and personnel spaces are located on the main deck in deckhouses. The semi-submersible version is of the ver- _ tical column type, and consists of four legs each containing two power system modules with the evapor- ... Rage ators, condensers, and warm and cold water pumps mounted externally in a vertical orientation. The columns are joined underwater by a cruciform struc- ture which also serves to support the CWP, and above water by the upper cross structure - a platform that 400 MM Seni-Sub, InOTnEC Plant also houses the power conditioning equipment, con- trol spaces, and living quarters for the crew. From Figure 2 - Semi-Submersible Size Comparison a volumetric efficiency standpoint, the semi-sub- mersible was inferior to the ship, but it was hoped that it would have superior seakeeping character- Aside from the production difficulties (which will istics as manifested in loads in the CWP. be discussed in a subsequent section) associated with such unconventional platforms, the revolution- 3. Major Technology Areas ary nature of the step up to the OTEC Plants as opposed to the normal evolutionary approach taken by the marine industry presents one of the more sign- Areas having a significant impact on the suc- ificant problems with respect to risk and the SOA. cess of the OTEC concept will be discussed in the The various regulatory agencies (ABS, U.S. Coast following paragraphs with respect to their position Guard) and most insurance companies tend towards the relative to the current SOA. conservative, and although no insurmountable techni- 364 cal road blocks solely as a result of size and the OTEC program where both scale model testing and configuration are foreseen, the advancement of the full scale experiments on test platforms are expec- state of art must be pushed forward in a manner ted to be carried out. In the meantime the results analogous to that in the offshore oil industry. obtained from these simulations should be used with caution pending a thorough validation of the theory CWP. The CWP is without a doubt the most involved. difficult technical issue facing the OTEC ocean system (non-power plant) designers both from the In most studies conducted up to now the pre- standpoint of determining the design loads and con- diction of the CWP dynamic loads and bending struction/deployment. A structure with a length of stresses has been carried out using linear frequency 3000 feet and an inside diameter in the vicinity of domain computer simulations. These tools are fairly 100 feet is unique in the maritime field, and the economical to use but they are somewhat limited in complications in establishing design criteria and that the non-linearities of the platform and CWP methods are formidable. The pipes developed by response are not adequately considered. Such non- Gibbs and Cox, Inc. with the assistance of Giannotti linear behavior is best represented in the time and Buck Associates are stiffened mild steel shells, domain, Reference (2), but unfortunately this is an but other materials such as concrete, glass rein- expensive technique. The general consensus is that forced plastic, and rubber are being considered to the frequency domain tools should be used to conduct reduce the wave and current induced loads in the parametric estimates of long term platform behavior CWP's. In the following paragraphs some of the in order to allow one to define the worst or critical problems involved in establishing those loads will design conditions. Having accomplished this, the be discussed. next step would be to conduct a few specific simu- lations with the time domain tool corresponding to The wave induced motions of the platform in the events associated with the worst case con- six degrees of freedom give rise to dynamic bending ditions. In this manner the structural response can and vertical stresses which could cause structural be determined with due consideration given to the failure as a result of either single extreme stress effect of non-linearities without jeopardizing the excursions or cyclic loadings. Superimposed on validity of the design calculations. these loads is the action of the ocean current velocity which varies with depth and gives rise to For this OTEC study a frequency domain simula- steady state bending of the CWP; therefore, a con- tion developed by Paulling, Reference (3), was used. siderable amount of effort during the initial stages The dynamic system under consideration consists of a of the OTEC platform studies has been devoted to the floating platform with the CWP suspended below it. prediction of these structural loads. A linearized frequency domain analysis of the response of the coupled system of platform and pipe The analysis of steady state current induced to incident waves is performed by the simulation. drag forces acting on the CWP is fairly straight Responses in random seas are obtained by superimpos- forward subject to the availability of empirical ing the response to a series of regular waves of data on drag coefficients at high Reynolds numbers different frequencies and amplitudes. Results of and vortex shedding at appropriate Strouhal numbers. the analysis include the motions of the platform and Studies aimed at developing this information are a series of points along the pipe, forces at the currently underway and preliminary results have been connection between the pipe and platform, and stres- used in various OTEC platform design studies. De- ses due to axial force, bending and shear in the tailed information summarizing this data is given in pipe. Reference (1). A typical output of this analysis is shown in The dynamic CWP loads and stresses induced by Figure 3 where the significant bending stresses in ocean wave action require rather involved analyses the CWP are plotted against the axial position along of the complex interaction of waves, platform and the pipe for the ship configuration in random beam CWP. As in any structural dynamics problem the seas. parameters which affect the magnitude and distri- bution of the loads and the response are geometry, The Paulling simulation is applicable to most material, structural arrangements, boundary con- conventionally configured platforms such as the ship ditions, mass distribution and the characteristics and semi-submersible developed as part of the cur- of the dynamic loads (time history, frequency, mag- rent study. Difficulties arise when more unusual nitude, distribution, etc.). In the case of stan- forms such as the spar, sphere, and cylinder need to dard ship and ocean platform applications many of be analysed since in those cases the simulation the analytical tools for predicting motions and loses validity because of its inability to accurate- loads induced by the seaway have been validated with ly calculate the hydrodynamic coefficients for such experimental data; however, the simulation of the geometries. Another frequency domain simulation OTEC platform/CWP configuration has not. This will developed by Garrison, Reference (4), presently hopefully be accomplished during the next phases of 365 MILD STEEL CWP must crythe entire load of the pe.The shpand BEA SEAS 5. semi-submersible designs utilize a universal style HINGED NEUTRALLY BUOYANT CEP joint containing two separate bearing assemblies, WALL THICKNESS - each of which carries the entire weight of the pipe. 0.2IFT . The present 501 in bearing design is very close to .- SOD - 0~.4 FT that required for such a device. Similar units were .0? ~~~~~~~developed for the stabilization assembly on the 1 ,ooo Glomar Explorer, and it is the opinion of several that ~ ~ ~~~~~~ bearing manufacturer that limited development would 1,500 result in a satisfactory bearing design for the OTEC plant; however, the use of heavier materials such as 21SDO 0 I 1I concrete for the CWP could result in unacceptable 4 ~~~~~~~~~~~loads on a compliant joint. Such a situation would require considerable research ahd development in 2,500 bearing design or the addition of expensive buoyancy 3,0 to the pipe to reduce bearing loads. 0 10 20 50 40 so so DYNAMIC SENDING STRESS ABOUT LONGITUDINAL Position Keeping System - The environmental AXIS OF SHIP IN K0I. conditions at some of the potential OTEC sites - namely peak current velocities up to 6 knots, Figure 3 -Typical Results, CWP Loads Simulation combined with the large projected underwater areas of the plants result in drag loadings that require position keeping systems beyond the present SOA. exists which does handle arbitrarily shaped bodies Although the watch circle requirements for the and can be used as an input to the CWP portion of the plants are not extremely stringent, the drag loading Paulling simulation; however the previously men- on either the 400 MW ship or semi-submersible plant tioned limitations still exist, i.e., no full scale in a current of only 2 knots is approximately two or model test verification. In addition the current million pounds requiring the equivalent of 67,000 models do not include all external effects on the shaft horsepower to hold the platform stationary., CWP such as secondary loads from the current induced This coupled with the extreme depths at some of the pressure distribution (steady state and time depen- sites under consideration (4650 feet at New Orleans) dent) and internal flow. and the long operating life of the plant (40 years) presents some difficulties with respect to position To adequately reduce the risk in motions and keeping technology. loads predictions for the CWP it will be necessary to develop an all-encompassing simulation in both Two basic systems are under consideration to the frequency and time domain. Their output could meet the position keeping requirements of the OTEC be used in one or more of the standard structural commercial plant; i.e., active (thrusters) and analysis programs to develop an optimum structural static (mooring lines and anchors). Non directional design. The development of these,' models is not active units analogous to a conventional ship's considered to be beyond present day capabilities, propulsion plant with sufficient power to maintain but it is considered an essential advancement of the thOTCpasonttinrersnlyviab. SQA with respect to the overall development of the In addition 2000 to 3000 horsepower rotatable units CWP. The simulations will then have to be verified that could be mounted in the required quantities are with a concerted program of testing instrumented also state-of-the-art, but with both these models and partial and full scale pipes on either approaches the parasitic power drain on the plant test platforms or limited scale OTEC plants prior to electrical output would be quite high reducing dras- commitment to the large expenditures required for tically the power available to the grid or an on- the commercial plant. site energy intensive industrial process. This CWP/Hull Interface - The method of attaching last h eodcoc h ttcsse. the OWP to the platform itself also presents some The combined environmental factors and vessel unique problems in itself. Considering that the OWP characteristics mentioned previously make any developed for the two OTEC commercial plants by proposed mooring system beyond the present SOA; Gibbs and Cox, Inc. weighed approximately 20,000 however, in most respects the required extrapolation tons in water and that vertical dynamic loads add an of current technology is not extreme. The 3000 to additional 2000 tons, it can be seen that the load 5000 foot water depths at the OTEC sites are within imposed on the platform structure in the vicinity of the present capabilities of mooring systems, but the the OWP is quite high. problem arises when the loads imposed by the OTEC The ajorprolem rise whn a omplantplants are compared to those of conventional drill The mjor roblm arses wen acompiant rigs. A number of systems, including ones with joint is considered. In many conditions a device mooring lines composed of a series of linked pipe that decouples the CWP and hull in rotation reduces sections fabricated from high strength steels, have the wave induced bending moments in the upper been proposed by different groups, and the one de- regions of the pipe, and the device can be con- veloped for the Gibbs and Cox, Inc. commercial figured to facilitate OW? deployment. The critical plant designs by Alan C. McClure Associates of item in this case is the bearing assembly since it 366 Houston, Texas is seen to be feasible albiet expen- The move exotic materials under consideration sive. It relies on conventional components to make for the CWP present similar problems. I f GRP or up an unconventional system, and it consists of a 14 reinforced rubber are chosen for the commerical point moor using wire rope for the anchor rode and plant CWP, it will involve a multi order of magni- deadweight anchors. The critical problem of long tude increase in the size of any structure made of life has been mitigated by the provision for, per- these materials with a commensurate level of risk. iodic replacement of the wire, and as a logical extension of the present SOA the associated risk is The second issue, that of the dimensional * ~~~considered to be minimal. restrictions of most existing fabrication facilities and the lack of sheltered deep water (with the Construction and Deployment - The critical exception of Puget Sound), will require the advance- aspects of the large size and unusual configuration ment of the SOA in offshore construction. Even with of the OTEC plants and the CWP can be distilled down such facilities as Offshore Power System's in to the difficulties involved in their fabrication, Jacksonville, Florida, the excessive drafts of the * ~~~assembly and deployment. Almost all construction OTEC plants will necessitate a large amount of fab- facilities in the United States are configured for rication, assembly, and deployment work at the conventional ships of generally conventional propor- operational sites. This is especially true of the tions and deal almost exclusively with steel as a CWP. For example the Gibbs and Cox, Inc. ship is * ~~~building material. On the other hand, the OTEC projected to be built in halves in a conventional plants, as should be realized by now, are rather building dock with the halves being joined afloat. unconventional, and a number of materials have been The heat exchangers and seawater discharge pipes proposed for their construction - steel and concrete will be added in deep water ( > 300 feet) along with for the hulls, and steel, aluminium, concrete, the CW? which will be fabricated on site. glass-reinforced plastic (GRP), and reinforced rub- ber for the CWP's. Any heavy construction done offshore is dif- ficult, dangerous and expensive, and the magnitude The primary issues with most existing fabri- required to place the proposed OTEC plants in cation facilities are the building dock dimensions operation exceeds anything yet accomplished in the and depth of water. Both the ship and semi-submers- offshore oil industry. The techniques and equipment ible by Gibbs & Cox, Inc., Inc. , the spar and ship by required to manipulate and assemble large sub- Lockheed and the spar and sphere by M. Rosenblatt assemblies on site must be developed to the point and Son, Inc. exceed the width and/or draft of all that the risk involved is reduced to acceptable U.S. shipyards unless significant facilities im- levels. provement efforts are undertaken or techniques that involve fabrication in sections with final assembly The magnitude of the problem can be seen when at sea are utilized. The problem is compounded by one considers the CWP. If built of steel it will in the scarcity of deep water (50 feet or greater) all probability be fabricated on an assembly barge immediately adjacent to the shoreline which would on site from a series of sub-assemblies, and will permit much of the out of dock assembly work to take grow downwards into the water from this barge as place in sheltered waters. sections are added at the top. When complete, the entire Pipe, 3000 feet long and weighing approx- The first issue, that of materials technology imately 20,000 tons, will be lowered and moved in shipbuilding, is of course, closely linked with bnahteOE ltomweei ilb tahd the material chosen for the OTEC plant's hull and The entire effort is envisioned to take about nine CWP. As stated above the U.S. shipbuilding industry months and during the process severe environmental is almost exclusively oriented to the production of conditions may be encountered. steel structures. The largest concrete floating structure built in this country to date is an LPG production barge for ARCO which is 461' long, 136' Environmental Considerations - The deployment in beam, and 56' in depth and displaces 65,000 tons of OTEC plants raises a number of domestic and full load and 29,000 tons light ship, Reference (5). international environmental problems which must be Larger concrete offshore structures have been built, addressed during the planning and design phases of the Ninian Central oil storage platform which dis- any future commercial plant. places approximately 600,000 tons with ballast for example, but most of these have been constructed in In order to deal with these possible environ- dedicated facilities oriented totally towards con- mental impact problems, the Department of Energy in crete construction. Additionally these structures cooperation with several academic and industrial sit on the bottom and are not subjected to the full organizations has been sponsoring extensive investi- spectrum of loads that a surface platform such as gations aimed at answering some of the more critical OTEC would be. The point here is that while concrete questions. These studies can be grouped into two does not pose any insurmountable technical problems, separate but related areas of study, i.e., near it would require an extensive investment to convert field and far field. units of the shipbuilding (or offshore construction) industry over to concrete production. This means Near field flow computations are of fundamental * ~~~that new production procedures must be developed and importance to the OTEC program for two main reasons. design criteria established by regulatory agencies First, they determine the inflow temperatures as a which implies a degree of technical risk. function of design and operational parameters and 367 of the ambient conditions. Second, they are a basic followed. The revolutionary nature of many of the input to far field environmental impact computations advancements in the SQA may be reduced to an since they determine the redistribution of heat, evolutionary process if such a program is developed salinity, horizontal momentum, biological species, and adhered to, using present level offshore tech- and nutrients between different ocean layers and the nology as a foundation. The approach being taken by possible addition of biocides and pollutants to the U.S. Department of Energy follows the philosophy certain layers. The near field flow problem has and will hopefully culminate in the commercial oper- been studied In sufficient depth by several investi- ation of a number of these large plants in the near gators. Their findings are summarized in Reference future. Ci); however, what is needed now is an effort to merge the different analyses and experimental results into a consistent methodology or model which References can then be used uniformly in OTEC plant siting and design. 1. "Ocean Thermal Energy Conversion (OTEC) Plat- The far field problem has to do with the form Configuration and Integration - Final impact of one or more operating OTEC plants on a Report,"l Gibbs & Cox, Inc., Report 18351-10 specific region of the ocean. Here one must look at (W-10,000, 10,010, 10,020), June 1978. the oceanographic processes in a much larger scale than in the case of the near field problem discussed 2. Tate, M. B. and Perini, L. L., "Dynamic Loads previously. These decisions go beyond the details Induced by Severe Storms in Elastic Cold W~iter of the individual plant design as they affect the Pipes Attached to OTEC Ships by Fixed and economic, social, political and environmental struc- Hinged Connections," Paper presented at the ture of the area of concern. Reference Cl) summar- Fifth OTEC Conference, Miami Beach, Florida, izes some of the most important far field studies February, 1978. which have been conducted so far under the OTEC program, and as with the near field case, there is a 3. Paulling, R. J., Jr., "A Linearized Dynamic need to synthesize the various analyses into a Analysis of the Coupled OTEC Cold Water Pipe single prediction model for a number of potential and the 11MB-i Barge System," Report prepared OTEC sites. for Morris Guralnick Associates, Inc., San Francisco, CA, August, 1977. In actuality the question of environmental problems associated with OTEC will probably not be 4. Garrison, C. J. "Hydrodynamics Interaction of fully addressed and identified until the time that Waves With a Large Displacement Floating Body," environmental impact statements are filed and some Naval Postgraduate School Report NPS-69 Gm operational experience has been gained with some of 77091, September, 1977. the proposed test plants. 5. Anderson, A. R., "A 65,000-ton Prestressed 4. Minor Technology Areas Concrete Floating Facility for Offshore Storage of LPG," Marine technology, Volume 15, Number 1, January 1978 The previously discussed areas represent the ones most critical to the success of the OTEC concept from an ocean system standpoint, but a num- ber of others with an inherently lesser degree of technological risk also exist. Some of these include the seawater pumps used to pump the warm and cold circulating water through the evaporators and condensers and the maintenance aspects of a ocean structure with a 40 year operational life. In these cases the SOA is very close to that required, and its extension to that point is not anticipated to pre- sent any degree of difficulty. As an example the warm water pumps will have to pump about 2.25 million gal/min of water per evaporator. If multi pumps are used then current axial flow designs are sufficient while a single pump of that size does not represent a difficult extension of the present SOA. 5. Summary None of the technical issues discussed in this paper are seen as a road block to the successful deployment of an OTEC commercial plant in the 400 MW sierneprovidedve wechnlorgydanizedn progicali prgsiz e tehoogrdancemn provieg wl rganied loical 368