[From the U.S. Government Printing Office, www.gpo.gov]
QH 76.5 -M3 C33 1988 Computerized Monitoring & Management of Nontidal Wetlands COASTAL ZONE INFORMATION CENTER 4,1, j, t of 4 if vector for one 4 ,w,mplete edangle March '!P@, 1988 110 v or c rdinates a in Lambert nfortnalproj. 4 .......... Wetland Polygon Outlines NWI digitized interpreted: vectors 01/15/88 Federalsburg Quadrangle t 0' r'l n f nl Maryland Department of Natural Resources Nontidal Wetlands.Division Salisbury State College A Computer-Based Assessment and Monitoring System for Non-tidal Wetlands Dr. K.-Peter Lade Salisbury State College Salisbury, Maryland 21801 COASTAL ZONE NT Mit MATION CENTER Prepared for: Coastal Resources Division Tidewater Administration Maryland Department of Natural Resources Annapolis, Maryland 21401 March, 1988 Final Report Preparation of this document was funded in part by NOAA, Office of Ocean and Coastal Management, and by the Maryland Department of Natural Resources. 1. INTRODUCTION Background downstream waters, andgovide flood control benefits." uide- The Non-tidal Wetlands Division of lines for Protecting Non-tidal the Water Resources Administration, Wetlands in the Critical Area, Maryland Department of Natural Resources, 1987:1) is in the process of developing manage- ment strategies for the State's non-tidal Although public attention has been wetlands. The project described in this focused on the dramatic decimation of report was completed to demonstrate the wetlands in general, it is generally capabilities of a computer-based system in conceded that nontidal wetlands' are at fulfilling some of the monitoring and least in equal danger of declining unless assessment functions required as part of guidelines for management are developed the non-tidal wetlands program. and followed. Of this total resource, The decision to develop a com- many of the State's non-tidal wetlands puterized system is based on several are widely scattered and often small in considerations demanding a computerized areal extent. An automated system for @pproach; principally the expectation that tracking the health and disposition of increasing amounts of diverse data will these wetlands would prove to be need to be incorporated into management particular useful in developing appropriate decisions and the necessity to investigate strategies for their management. both . written documents and graphic A second consideration was based on mapping data. Further, ecological and the need to comply with new regulations. legal aspects associated with non-tidal Thus, wetland management govern the structure of the numeric and @.raphic data bases. "The Chesapeake Bay Critical Although it is not possible here to explore Area criteria require that non- these underlying factors in depth, a brief tidal.wetlands be identified and highlighting of some of the more important afforded protection by local considerations governing non-tidal wetland jurisdictions. Two types of management might help to place this protection measures are specified. project into its appropriate context. First, a minimum 25-foot setback First among the considerations taken or buffer around the identified into account while developing the com- wetlands is to be established puterized monitoring and management within which new development system was the nature of the resource activities, or other activities that itself. may disturb the wetland, are prohibited. Second, local "Non-tidal wetlands are valuable jurisdictions are to protect the areas for plant, fish, and hydrologic regime (e.g., the flow wildlife habitat, are vital to ot water into and from the maintenance of the qualify and wetland) by minimizing land productivity of adjacent or disturbances in the wetland 1 drainage area. Only under information on the vegetation of coastal certain circumstances, are non-tidal wetlands, and the location, alterations to non-tidal wetlands extent, and values of different types of to be permitted by local vegetation could be interactively manipu- jurisdictions, but mitigation of lated by computer. The management the alteration must be under- system to.emerge with the help of the taken. (Guidelines for Protect computerized capability was intended to ing Non-tidal Wetlands in the permit the utilization of a large variety Critical Area, 1987:1) of data types, including aerial phot n digital forZ raphy, satellite imagery (i' in This mandate would require that an soils information, NWI mapping data, and active record be maintained of all non- USGS digital line graph data. tidal wetlands and tracking potential The turnkey system provided as the disturbances. A detailed delineation of principal product of this project is both non-tidal and tidal wetlands was described below, followed by a review of conducted by the U.S. Department of the the tasks undertaken to test the system. Interior, Fish and Wildlife Service as part With the installation of the computerized of its mapping of national wetlands and mapping system the Non-tidal Wetlands published as the National Wetlands Division will gain a si@nificant capability Inventory (1983). ' These 1:24000 scale for managing its growing data base and maps were "prepared primarily by stereo- maintain the ability to constantly update scopic analysis of high altitude aerial that data base through the inclusion of photographs. Wetlands were identified on both visual and tabular data. tl@@ photographs based on vegetation, visible hydrology, and geography in accordance with Classification of Wetlands and Deep-Water Habitats of the United States ... (Cowardin, et al. 1977)" (from annotation on NWI maps). In 1987 DNR contracted with Fish and Wildlife to digitize these maps for the State to permit their eventual use in a computer-based geographic information system. This digitized data set forms the basic inven- tory to which all additional computerized efforts were linked. Goals From November 20, 1987 to November 30, 1987, Salisbury State College, under contract to the Coastal Zone Management Program of DNR, developed a computer- based software and hardware system for use in the management of non-tidal wetlands. This project was funded, in part, by the Office of Coastal Zone Management of the National Oceanic and Atmospheric Administration, United States Department of Commerce. The purpose of the project was to develop a means by which detailed 2 11. System Components - Hardware Overall Concept and System Design All microcomputers used in this project require the addition of an The system designed for use by the arithmetic coprocessor (INTEL's 8087, non-tidal wetlands program is niicrocom- 80287, 80387); a minimum of 640 kilobytes puter based using standard off-the-shelf of ram (random access memory), console nen Compatibility with the IBM- display device, one or more floppy disk compo ts@ t. drives, a mouse, and either the PC-DOS PC standar was a primary requiremen The additional graphics capabilities needed or MS-DOS Operating Systems (version to drive a large-screen monitor act 2.0 or higher). Practical considerations independently of the function of the also dictate the availability of one or microcomputer to the extent that other more mass storage devices (fixed disk, activities may occur on the microcomputer, removable cartridge, optical disk). including data base management, word- Because of the very heavy storage processing, and page composition. requirements for image data, large- capacity off-line mass storage is desir- able' The system supports large-capacity Computing Platform hard disks (70 megabytes and higher), I nine-inch open reel tape drives, and The software to be d&scribed below optical disk (200 megabytes and higher). has been tested and found to run success- Although 80386 machines were used fully on the following microcomputers: in software development, and one such machine is being used by the non-tidal 1. IBM PC/XT class computers wetlands program to run the software, it IBM PC should be noted that such machines are IBM PC/XT more susceptible to bus timing conflicts. IBM Portable If the bus speed exceeds 8 megahertz Leading Edge some hardware may not run properly, and if the CPU speed exceeds 12 megahertz, 2. IBM AT class computers careful attention must be given to IBM AT determining that all components, including Fivestar 286 base memory, are capable of supporting Standard AT the higher speeds. Compaq 286 3. INTEL 80387 class computers Graphics Platform Fivestar 386 PC Limited 386 Graphics functions are provided through -peripheral equipment capable of Based on the compatibility tests completed producing analog rgb output. The as part of this project it is felt that most following hardware combinations have IBM compatible nuicrocomputers running been tested and found satisfactory: INTEI2s 8088, 8086, 80286, or 80386 cpu's will adequately host the software. 1. High Resolution Devices (1024 x 1024 3 or larger) Once data has been input into the Vectrix Pepe & Monitronix monitor system, stored, and manipulated in some Vectrix Presto & Monitronix monitor fashion, there is usually a requirement to Number Nine Pro-1280 & Sony 1030 produce some physical representation monitor thereof Although one might consider 2. Medium Resolution Devices data storage to be a form out data Vectrix 384PC & Electrohome monitor output, we will consider only hardcopy AT&T Targa-16 & Electrohome devices here. monitor (or Sony Multiscan) The following devices have been tested and found satisfactory for the The software does not require data to be production of tabular and graphic output: organized to match the resolution of the display device. The graphics displays act 1. Tabular/Text Only as "windows" on any data set, permitting all dot-matrix rinters the same data set to be as easily manipu- laser printers p(preferred) lated on a medium resolution system as on ink jet printers a high resolution system. thermal transfer printers 2. Graphics/Text Output thermal transfer (e.g. Calcomp Plot Data Input Devices master) color inkjet (e.g. Tektronix 4060) Since data input is a major require- solid color ink (e.g. Howtek Pixel ment of any computerized system, con- master) siderable attention will be paid to this monochrome laser (e.g. HP Laserjet task in the discussion of the software Series 11) below. The following devices are sup- ported for data input: There are many considerations regarding the number of colors that can be 1. Standard DOS-supported devices produced, the resolution capabilities of floppy disk drive individual printers, and the ability to removable hard disk (e.g. Bernoulli emulate pen plotters, that must be taken Box) into account. As a rule of thumb, cartridge tape resolution should be at least 200 dots/in- ch in either monochrome or color, color 2. Non-standard devices dithering must be possible to produce optical disk drive (may emulate a hundreds of colors, and hardware DOS device) limitations governing output size must be 9 inch open reel drive defeatable if a printer is to be considered CD ROM satisfactory. In all cases special drivers mini/mainframe communications link were written to extend and enhance the with file transfer capabilities of printers tested for this project. Data Storage Devices Data storage utilizes the same hardware types as Data Input Devices. See above. Output Devices 4 Ill. System Components - Software Overall Concept and System Design tion of subtle shades of hundreds of colors from palettes of millions of colors A wetlands assessment and monitoring is a characteristic of high-end analog rgb system useful to state and local govern- display systems and an important require- ments was enhanced to encourage monitor- ment where photogrammetric considera- ing individual and cumulative impacts to tions are involved. estuarine and freshwater wetlands from Among the various features of the activities occurring within the wetlands extended MIPS are an intuitive interface themselves, and from adjacent land uses. made possible by the almost exclusive use The current system uses data from the of C-language program modules which National Wetlands Inventory, and from directly address the appropriate micro- classified LANDSAT satellite imagery computer dependent hardware. Software (Thematic Mapper digital data). Capabili- functions are menu-driven (see Supple- ties include interactive display and ment: User's Guide), but support ot a map ing of wetlands types and their graphics pointing device (most often the spatfal distribution. Through software mouse) permits point-and-shoot routines developed in previous projects and to be supported as well. Additionally the significantly enhanced as part of the mouse is used for manipulating geometric current project, NWI wetlands data can be shapes (e.g. various elastic box routines joined to existing files containing informa- used to define active areas within the tion on wetlands permits, watershed land full screen). When used in conjunction use maps, and other tabular data. In with pop-u functions the mouse permits addition, digitized soils data and aerial control of such functions as color photography can be included as data layers balancing, interactive three-dimensional into the management system. Geometrical- modeling (wire-frame and solid render- ly registered base maps can be used to ings), drawing (including on-screen overlay vegetation data, land use, land digitizing), and windowing. cover and soils. Video digitized data, although not registered, can be manipu- lated through vector overlay to enhance Data Import Procedures the interpretative capabilities of the full system. Getting data into a computerized The management and monitoring system is potentially among the costliest system developed as part of this project parts of building and maintaining the derivesfrom a commitment to manipulate system. Most GIS's (Geographic Informa- maps and images. Hence, the core system tion Systems) depend on layers of is given the acronym MIPS for Map and digitized data, where each layer is image Processing System. The graphics carefully matched to a known base map. support aims to present maps and images Consistency of the geographic projection, in full color. Thus natural true-color accuracy of the digitized layers, and screen images may be stored in those reliability of the attribute listings instances where the original data was in associated with digitized layers require the form of a photograph. The preserva- careful planning and data preparation. 5 Regardless of the sophistication of a software to help the user not be dis- computerized Image Processing or Geogra- tracted during data analysis. phic Information System, there is no The preceding discussion and substitute for careful data preparation. examples assume that existing data is Given the fact that a great deal of being imported and that this data resides data has already been collected, often on a device supported by MIPS. Since all either at considerable cost, or under common forms of media can be directly circumstances that are never likely to be accessed through software, there is replicateld, it was our aim to create the considerable power to this system. It most flexible procedures possible for data doesn't matter whetl@er the data is on import. Dius we recognize that it would diskette, optical disk, magnetic tape or be important to provide means for CD Rom. If the devices are available as introducing both new and historical data. part of the microcomputer system, they Further it was anticipated that data serve as an active shared resource.Thus collected for other purposes might be both DLG and Landsat satellite data, usefully integrated inm the system. A supplied by USGS on 9 inch open reel variety of procedures were developed to tape, requires no special procedures to exploit known data structures associated import, a stumbling block for many other with extant software. For example, using microcomputer based systems. If an open MIPS menu selection techniques, it is reel tape drive is not always available or possible to easily import data that were convenient, the software provides a originally prepared in a variety of method for reading nine inch tapes onto different formats; e.g.: optical disks and then using the optical disk as an alternate input medium. 1. SSURGO (Soil Conservation Service) New data can be created by using 2. MOSS (U.S. Fish & Wildlife) either live video or a digital scanner for 3. EDIPS/TIPS (Landsat MSS & TM) the creation of rasters. A digitizing 4. DFX (Autocad) table is used for creating vector files. 5. DLG (USGS) Since there is increasing support for the 6. DEM (USGS) digitizing capabilities of AutoCad, MIPS 7. SPOT @Spot Image) c@rrently imports and exports DFX files 8. NASS National Agricultural with AutoCad. This assures compatibility Statistical Survey) not only with DFX files in general, but also with COGO software that utilizes the Generic import procedures to permit DFX file structure. inclusion of byte, ascii, and binary data Among the most powerful of data are also included, as is support for some input capabilities is the ability to of the newer file formats, including tiff interactively scan, in full. color and at up rio, and targa. to 300 dpi resolution, maps, photogra hs, The system user is insulated, for the transparencies, and virtually any reEec- most part, from the details of data import. tive or transparent media. Although The data import menu allows selection by current technology limits the total size of name of a foreign data set (see Sup- reflective material to 11" by 17", and the plement: User's Guide). Once selected, the total size of transparent material to 8.5" software opens an appropriate new or by 11" when scanned on equipment existing data file, skeletonizes the file if costing under $10,000, there is rapid necessa , and writes the converted data development occurring in this area. T'he to the 7ile. For raster data, histograms software support in MIPS bf the Howtek are automatically produced the first time Scanmaster demonstrates some of the best images are displayed. For vector data, qualities of the software when used in a automatic fit-to-screen and coordinate management context. orientation routihes are accessed by the All functions of the scanner are 6 clearly shown on a graphic representation, the manipulation, on-screen, of calipers, a previewing of the entire scanning surface protractor, an elastic box, an elastic or any subsection is possible, and a circle, and a user-definable polygon. scanned image can be brought directly to Thus in the hypothetical example, the the graphics monitor and/or stored on disk calipers can be used to measure the for additional analysis. linear distance between the proposed construction site and the edge of the wetland. The protractor can be used to Data Manipulation Procedures project an angle, where a change in orientation of a linear feature may serve. @ince there are a large number of to mitigate a negative impact on the ways in which data can be manipulated, wetland. The elastic box and circle may Appendix A: System User's Guide, should be used to define an area of known size be consulted for details. To illustrate the and then to interactively move the box graphics capabilities of the system, one about on the screen to explore alternate pop-up utility will be considered here: placements for a proposed construction MEASURE. that has a known areal extent. Finally There is a frequent need to deter- the user-definable polygon can be used to mine the location, extent, size, and other measure the amount of wetland loss, if physical characteristics of wetlands and any, that would occur as a result of associated natural and cultural features. various mitigation strategies. In a typical case, the investigator may Quick, easy, almost effortless use of need to quickly determine the proximity of powerful computer techniques can help to a wetland to a proposed construction site augment photogrammetric and engineering and to respond to a permit request that skills that most managers and their staff mi@ht affect the status of the wetland. bring to the job of monitoring such Using MIPS it is possible to either bring natural resources as non-tidal wetlands. to the screen a stored photographic image, A second data manipulation process map, or other image appropriate to requires the use of text data. For examining the potential impact on a given example, the National Wetlands Inventory wetland posed by a construction permit digital data allow each wetland to be request. If no such photograph or map located spatially (initially by Lambert has been stored, or if a new photograph Conformal, but through conversion by or map is available a rapid scan can create Latitude/Longitude, State-Plane, UTM, the needed screen image. and other coordinate systems). The data Once on the screen, the pop-up also allow each wetland point, linear or utility MEASURE allows the analyst to polygon to be given a type attribute and quickly perform several useftil operations. measure of length or area. These If the scale of the image is known (as in statistical and descriptive attributes of the case of satellite data) the image may wetlands are best handled by a relational be directly calibrated. If the scale is not data base. For the computer-based non- known (as in the case of most photog- tidal wetlands management system dBase raphy), on-screen measurement of known III+ was chosen as the appropriate features (e.g. distance between two road relational data base. The decision to use intersections) may be used for calibration dBase III+ was made partly on the purposes. Calibration can be in either capabilities of the software and partly on metric or english systems. Linear the basis of its wide distribution in many .measurements maybe in one system and locial, state, and federal agencies. areal. measurements can be reported in Using dBase III+ it is possible to either the same or the alternate system of query the data base by quad and county, measurement. determine the types of wetlands that Using the mouse, MEASURE allows exist within the entity considered, sort 7 the data by wetland type, and report rotations requires that the vectors and statistics about the wetlands. It is further rasters be made to "fit" each other. possible to link other types of data to the Enhancements to MIPS now allow NWI data, including permits and regula- floating vectors over images and then tions files, field data from water quality rubbersheeting the vectors to the image. testing stations, geological and environ- This is done interactively with the mouse mental measurements, and many other by matching identifiable points on the kinds of data that become available. The vector overlay to identical points on the fl6dbility of a relational data base is that raster. The computer then uses a least- it permits the regular addition of not only squares fit algorithm to recompute the new types of data but specialized data vectors and rep@ot them over the image. bases that can then be linked to existing The process is iterative. When a data bases. Appendix A: System User's satisfactory fit has been achieved, the Guide, provides examples and directions for results are displayed and the calculations utilizing dBase III+ to access and manipu- stored. late the NWI data sets. Individual elements in the overlay file Finally, a goal of this project was to may then be selected for display, explore the capability of the system to permitting the investigator to view the merge raster and vector data in- the form distribution of a single class of wetland of overlaying the NWI vector data over types, for example. In a further refine- base maps, such as quads, aerial photog- ment of the software it will be possible raphy and satellite data. This overlay to directly access the dBase III+ informa- capability differs somewhat from the tion by pointing at vectors or vector current GIS technique of overlaying classes on-screen. It will also be possible multiple vector files. The simultaneous to overlay multiple vector files over investigation of raster and vector data by multipleraster images, thereby setting up on-screen overlay and manipulation is the necessary basis for a modeling system possibly one of the most important new that incorporates the best of vector-based computer capabilities that will substantially GIS systems currently available with the contribute to the effectiveness of manage- best of faster-based systems. By ment and monitoring tasks for natural providing this active ra@rge. of the two resource managers, geographers, and others most important graphic-oriented data currently involved in GIS development. types currently in use by planners, Fulfilling this task posed the greatest managers, and academicians, it will challenge to this project and also con- become possible to more ftilly exploit the sumed considerable more time and resour- wealth of field data that has accumulated ces than had originally b@en anticipated. and to suggest more productive ways in A technical discussion of the methods wbich additional data may be acquired to used to manipulate a vector graphics plane address specific management and monitor- alongside a raster graphics plane can be ing needs. found in Appendix C: Technical Details- As a final note in concluding this Software. Briefly, an image is displayed brief overview of data manipulation to the screen as a raster. Then a vector procedures, it should be noted that the file is accessed and the vectors plotted on computerized system described here top of the screen image. If all data overcomes many of the restrictions that conformed rigidly to a specified base map, are still currently hampering the efforts the process of at least presenting a merge of other systems. of vector and raster data would be First, there is no requirement that all complete. However, the inclusion of data conform to one common scale and photography, the probability that at least projection. Data may be imported in some data sets are at different scales, many ways and then handled interactively. different projections, and different angular This process emulates the manner in 8 which data has been traditionally handled in either a generic form or in a systems- by photogrammetrists, but replaces some of specific form to a similar or a different the tedium of either mentally adjusting for image processing or GIS system. This is variations in scale and projection or using done by selecting appropriate expqrt mechanical devices, such as the zoom forms from the menu and then writing transfer scope. the data out to disk or tape. MIPS Second, consistent vector overlays allows extraction of rasters from within can be projected over user-selectable base rasters if a subset of a larger data set is maps and images. This extends to the use desired. Vector interchange is often best of inherently distorted images, such as done through an industry-standard or well aerial photographs, giving the analyst an understood file format such as DFX, extremely powerful means by which he can TIFF, or TARGA file formats. Although bring highly detailed interpretive vector- it is relatively straightforward from a based data (e.g. DLG) information to his software implementation point of view to evaluation of imagery. transfer generic or specific versions of Thirdly, multiple vector overlays and byte raster data or vector co-ordinate interactive attribute manipulation con- data, the increasing commitment to arc- stitute the core of a new generation GIS node data with associated attribute files approach. The result is a system com- d'oes pose new challenges. This task will mitted to intuitive data exploration and require a significant effort if a smooth would suggest interactive modeling as a integration with the existing system is to system function to parallel the more be achieved. traditional raster classification schemes Transfer and storage of data follow- used in image-only data analysis. In all ing manipulation may require no more likelihood, proper development of such than a disk-save of a screen-image that software would also require a stronger has been operated upon. Although many comgFting platform. Currently the 80386 image processing systems allow effortless mac ines offer considerable promise, but saving of screen raster images, few parallel processors, specialized hardware, provide a sufficiently generic form of and UNIX-based operating systems might such a screen save to permit display of a ultimately prove a desirable alternative to saved image on different hardware. MIPS the current PC/DOS environment. To be is addressing that problem currently by successful, a session should perform supporting, through software, screen- calculations rapidly enough so that the restoration techniques that identify the analyst can constantly interact with the original save format and then convert, if data rather than 'waiting for protracted necessary, the save files to allow them to computations to complete. The microcom- be displayed on incompatible hardware. puter is a worthy alternative to the mini Data that might have been altered by or mainframe system only if it is fast and filtering or classification can be saved as interactive. a new element within - the original file, thereby becoming appended to the original data set. Storage is therefore Data Export Procedures automatic. Export would be similar to import except that some consideration The same devices and procedures used must be given to the way in which to import data may also be used to export computed histograms, color lookup tables, data. Since data expqrt may be for the and other computed information is passed purpose of either transferring existing data to external systems. Currently data to another system or for transferring or export is being tested first for software storing manipulated data, there are two compatibility within the same computing somewhat different procedures to follow. platform. Export to other software Data on the system may be exported packages running on micro or mini or 9 mainframe systems is also scheduled for ftiture. testing. In this and other projects for the Often, it is desirable to save the Maryland Department of Natural Resour- results of an analysis session so that a ces, Salisbury State College has demon- hardcopy can be made at a future time. strated the utility of optical disk Since 'a separate section on hardcopy cartridge storage as a means of ensuring techniques follows, discussion of this task that vast amounts of data can be will be deferred and the reader is invited accessible to the microcomputer user. to investigate the relevant section. The non-tidal wetlands project has A final note of caution should be benefitted from this work by being able inserted at this time. Vector and arc-node to access these large databases directly data require relatively little storage space. without the need to support either tap@ For example, an entire quad of NWI data drives, or requiring a link with a mini ay take up less than two megabytes of based system for data storage. space on a disk. Text data also requires in comparatively little space. Large data bases may exceed ten or twenty mega- Analysis Procedures bytes, in some cases perhaps more. But typically even fairly complex dBase files are under five megabytes in size. Image Appendix A details the methods to be data, on the other hand, particularly of used tor analyzing data using the high resolution, can occupy such vast Computerized Non-tidal Wetlands Manage- amounts of space that until recently very ment System. This section presents a little serious consideration had been given concept-oriented overview of all analysis to microcomputers because of their limited procedures currently implemented. storage capabilities. As indicated previously, MIPS at is A single nine inch color photo, core is an image processing software scanned at 300 dpi, for example, can package which has been significantly occupy over 21 megabytes of data storage. enhanced to include vector handling In compressed form it will still occupy capabilities. Since the core systems is over 7 megabytes of storage. Aerial devoted to image processing it is coverage for the state of Maryland to explicable that many of the analysis determine wetland. types and distribution procedures begin with a means of can easily account for several thousand displaying an image on-screen. photos in one acquisition. It is clear that There are two basic ways of display-. storage capacities far exceeding those of ing an image. The first is to access the the floppy and fixed disk are mandatory. original data and using density slicing or The solution has been the optical other algorithms present a monochrome or disk. A new technology, there are still no color rendition of that data. Figures 3-7 standards and only recently have some (in Supplement: Figures, Charts, Tables) manufacturers been able to make their illustrate this process by showing single- drives and media be DOS transparent. channel grey-tone images of satellite data This means that the user notices no (SPOT and LANDSAT), an aerial photo- difference between using an optical disk graph, a soils map, and a tax map. All with over a hundred megabytes of storage are common forms of data types likely to per side and a Bernoulli cartr 'idge, for be of interest to a wetlands manager. A example, which is limited to a total of 20 second way of displaying an image is to megabytes per cartridg@. Although the 250 load a previous screen save. This megabyte optical disk is currently available presumes that once data has been in sufficient quantity to constitute a viable translated into a screen image, the result storage medium, the gigabyte drive and is then saved in compressed form. The disk are the likely stars of the immediate advantage to working with screen saves is 10 that they load very rapidly since no new combined with undistorted vector data. calculations have to be performed. The The ability to easily n@anipulate disadvantage is that some of the informa- rasters and vectors is an important tion inherent in the data is lost due to capability. But of even eater impor- compression. A hardcopy output from a tance is the ability to perrorm measure- screen save is identical to that of an ments on these data, to enhance them original data extraction if the screen is using CAD primitives, to . generate acted upon. However, higher resolution hardcopy products rapidly, and to include output can be achieved by using the tabular data as part of a photogrammetric original data rather than a screen save. analysis at a cost and speed that cannot Regardless of whether the on-screen be achieved in any way other than image has been generated by reading the through the use of a computer and the data or loading a screen save, the next type of software made available through step is to perform some investigatory or the implementation of the non-tidal analytical task. The MEASURE pop-up wetlands management system. utility has already been described and Analysis goes beyond visual inspec- exemplifies productive interaction with tion, however, and MIPS provides more stored data. If color lookup tables are powerful tools than have been described supported by the hardware it is also thus far. A particularly useful feature possible to enhance an image so as to that has great future potential is the highlight certain natural features. In generation of three-dimensional displays. Figure 8 (Supplement: Figures, Charts & MIPS can manipulate three raster planes Tables), sediment loads are clearly visible to provide the equivalent of a two- as exaggerated color differences. dimensional color image and a single Vector data may also be manipulated elevation plane to fit the image data into on-screen. The NWI data sets have been a three-dimensional matrix. Three prroduceF in both paper form and digital dimensional modeling can take advantage 0 m. igures 9 and 10 (in Supplement: of digital elevation data and thermal Figures, Charts and Tables) compare the emission data as well as other kinds of paper and digital products. It is clear data that would aid in the analyst in that if the data is presented in too small understanding his imagery better. a space, important details are lost to the Wetland analysis might use edaphic observer. Even when the scale is changed variation as measured by in-field moisture so that a one-to-one match is achieved sensors or topographic elevation as with a TS minute topographic quad, the provided by DEM data sets to better paper NWI map continues to contain so reflect the condition of wetland place- much information that it challenges the ment. user to use it effectively. The computer- Three dimensional modeling proceeds drawn vector renditions, however, provide in two steps. The first is the creation of better detail by separating the base map a wire-frame that reacts instantly to layer from the interpretation layer and by changes introduced by the analyst. Once allowing interactive zooming to focus on again the mouse is used to move on- any detail, no matter how small. screen sliders, eliminating the need to Combining base map and vector data introduce numbers for elevation, rotation, together allows the interpreter to evaluate scaling, and other required inputs. The the vector data against various base maps, numerical values are, however, always chosen to suit the particular purpose of reported on the menuing iponitor in the analysis. The power of the computer response to mouse movements. is realized when base map and vector data The second step is the creation of at different scales are combined and the solid model with, or without, hidden matched to a common scale, and when line removal. The data may be sampled inherently distorted base map data is to permit a more rapid display of the 11 solid model. A complex, large, three- Plotters and laser printers have provided dimensional solid may take up to fifteen reliable and effective means by which minutes to draw. Although this would presentation graphics can be created. seem to be excessively long, mini and More complex is the process of creating mainframe computers require an even high quality hardcopy output of analog longer time to perform this task than an rgb screen images which are of photogra- 80386 PC with an arithmetic coprocessor. phic quality. Two processes currently are Since the data can always be sampled and used to produce good quality prints of the wire frame is available for rapid analog color images: color ink jet and manipulation, the longer time required to color thermal transfer. produce a finished product is still consis- Considerable effort has been ex- tent with the philosophy of demanding a pended to provide the non-tidal wetlands rapid response from the microcomputer. management system to produce good- For individuals familiar with tradi- quality hardcopy output. , Software tional image processing techniques, the control of color separation output software offers a suite of tools that may characteristics is fairly sophisticated, be used to operate on raster data. giving the user control over almost all Included are the ability to compute the variables that would affect the quality of correlation between rasters, to compute a the hardcopy product. Additionally, the convolution on a single raster using user is given the opportunity to accurate- filters, a s.emiautomated interpretation of ly scale his hardcopy output so as to rasters using preidentified features, and provide standard.map, product capabilities. several predefined index calculators. The Where the maximum size of the paper last of these includes the ability to that a hardcopy device can handle is less calculate a normal difference vegetation than the required map or image, the index, a transformed vegetation index, and software presents a multi-page graphic a leaf area index. Arithmetic and on-screen. Using techniques associated algebraic manipulations can also be with page composition, the software performed on rasters. automatically generates multiple page Vector data can be analyzed by output which can then be panelled, if category where vectors are classed as they desired, to produce a map of appropriate are in the case of the NWI digital data size and scale. set. T'hus subsets of larger data sets can More details concerning the specific be extracted and further overlays made to procedures involved in producing hardcopy allow determination of the value of output is provided in Appendix k More combining several criteria in analyzing the than a dozen printers are currently relation of . different attributes to each supported and there is a strong interest other. in supporting new printing technofogy as In the future the expanded ability to it becomes available. deal with multiple vector files, to edit these files through graphics techniques, and perform computations on combination Compatibility with GIS and Other vector planes will provide the essential Management Structures GIS tools to permit dynamic modeling. During the course of this project, a significant revision of the file structure Output Capabilities used by MIPS was undertaken to allow easier manipulation of raster data and Practical applications of image also the inclusion of vector data in a processing and GIS capabilities require common file format. that the results of an analysis can be Currently the file structure is a graphically shared or included in a report. superset of the enhanced DLG structure 12 @ ished by USGS. Thus a single file, Wue.tified by the system as a raster/vec- tor file (.rvf) may contain groupings and subgroupings of - data types that are inherently different in their individual structures. By maintaining a single file for both raster and vector data it is possible to investigate the relationship among different data layers more effec- tively with the additional benefit that the user is not required to maintain a list of compatible files. To aid the analyst in deciding on which file groupings to investigate, the system now supports an extended labeling feature that permits desc@ptive labels to be attached to individual file elements as well as the file itself. This may also serve to identify the file element as an import or export data set. The software maintains its own labels to * indicate whether the file element is byte raster, compressed raster, binary, ascii, vector, or arc-node. Import from other image processing and GIS systems requires only knowledge of the file structure used by that system or of the structure of the file exchange format. MIPS will directly import known structures and provides a generic import utility for previously undefined structures that do not appear on the menu. Export to these external software packages is the inverse of import and requires no special knowledge other than the desired export. label or format. Export routines for attribute files linked to arc-node files have not yet been developed. This is a high priority item since true compatibility with other GIS systems necessitates carrying along the attribute files with the vector or arc-node files. 13 IV. Data Types Thus far a number of data types have often fly low-level aircraft and acquire been identified in conjunction with the oblique 35 mm color photography. The description of features of the software. USGS and NOAA sponsor acquisition of Below each of these data. types is con- high level aerial photography (e.g. NHAP). sidered individually and a brief discussion Individuals may also acquire photography is included to indicate the utility of each that would be useful to the non-tidal type in addressing the needs of wetlands wetlands program. Regardless of the monitoring and management. nature or origin of photography, it can be brought into the system through scanning and storage on optical disk. Aerial, Photography Large frame (9" by 9") aerial Landsat Satellite Digital Data photography is among the most useftil of all imagery in the delineation of wetlands The U.S. has rovided satellite and wetland boundaries. coverage since 1972 Jmost parts of the The photography may use a variety of world. The instruments have included films and filters. Natural color film with return beam vidicon's, multi-spectral a yellow filter has been shown to be very scanners, and thematic mappers. The desirable both as a resource for making highest resolution available is the 28 wetland spatial extent and species type meter cell size achieved by the thematic determinations, as well as for day-to-day mapper. Further, with seven spectral reference when a permit request requires bands, the thematic mapper is able to viewing the area. provide valuable data to the natural Very popular also is color infrared resource manager because the bands go (CIR) photography. The value of CIR well into the visible infrared and thermal photography lies in its ability to clearly portions of the electroma&netic spectrum. delineate shore from water, to be sensitive Since Landsat data is digital, import to subtle variations in moisture content, into the system only requires knowledge and to permit biomass extraction through of the structure of the specific data one computerized means. has acquired. There are many data Wetland photos flown for the State formats and most are supported by MIPS of Maryland are usually either at a scale to allow automatic extraction through of 1:24000 or 1:12000. Both paper and menu selection. transparent products are customarily generated. Transparencies require special viewing equipment, but provide a richer SPOT Satellite Digital Data variation in tonal gradations than paper prints and are particularly desirable as In a move to commercially exploit the input into a computerized data base market for earth-resources satellite data, through scanning. the French government has launched the Other photographic products are SPOT satellite which is capable of available. The Soil Conservation Districts imaging at 20 meters resolution multi- 14 spectrally (three bands) and at 10 meter attempt to remedy the situation through resolution panchromatically (single wide software development, this project has band). sponsored the conversion of raster to The sensors aboard the satellite can vector data through automated techniques. be "pointed" so that acquisition of data for Algorithms for converting 8-bit data to a particular ground area is ncit strictly binary data, thresholdin raster and limited by the path of the satellite. Both vector editing, and line t1ii-ning are well multispectral and panchromatic data have under way. been acquired in digital, form by Salisbury Prior to the actual conversion of State College as part of other contracts. soils map to vector files it is possible to It is clear that the addition of this scan the soil maps and use them as a important data type is valuable from the base map for projecting NWI vector data. standpoint of wetland monitoring. Since the soils delineations are made over Satellite data, especially high resolution aerial photography, the ability to float satellite data, can be used not only to vectors over the base map and fit the identify wetlands but also to assess the vectors to the base map is a necessary surrounding land use and land cover. prerequisite to using the soils data in this form. Soils The intimate association of soils with Digitized NWI Maps specific wetland types is significant. The U.S. Fish and Wildlife Service Hence: conducted an inventorx of the wetlands of the United States using aerial photog- The presence of undrained raphy commencing in 1979. All wetlands hydric soil is one of the three were classified according to the Service's major criteria used to define official system: "Classification of Wetlands wetlands. Hydric soils are and Deepwater Habitats of the United either: (1) saturated at or near States" (Cowardin et al. 1979). The 1he soil surface with water that National Wetlands Inventory (NWI) is virtuall lacking free oxygen conunitted to establishing a wetland for sigNicant periods during database for the country, in both map the growing season or (2) and computer forms. The initial emphasis flooded frequently for long was on map production and the original periods during the growing NWI maps are available at the standard season." (Atlas of National U.S. Geological Survey topographic map Wetlands Inventory Maps of scale of 1:24000. Chesapeake Bay, vol. 4, p. 4) The Chesapeake Bay inventory was completed in 1984 and wetland maps have Figure 11 (in Supplement: Figures, been produced for the entire Bay area. Charts and Tables) lists a preliminary list In 1987 U.S. Fish and Wildlife completed of hydric soils for Maryland. Remapping digitizing the Maryland maps. These are of Maryland's soils would be required in available on 9 inch open reel tape in many areas to account for natural changes MOSS vector format. All digital data is since the last county soil maps were being compressed, read onto optical disk, prepared'. and imported into the non-tidal wetlands Soils data remains one of the most management system. detailed data types available for natural Appendix F: NWI Map Preparation resource assessment. Since most soils data describes the method of preparation for has not been digitized, the value of the the original large-scale 1:24000 maps, the data cannot be fully realized. In an collateral data sources used, and photo 15 interpretation problems encountered. non-tidal wetlands, it becomes important In addition to the digitization of the to gain control over an extensive litera- Maryland wetland maps, U.S. Fish & ture if enforcement and control are to be Wildlife also provided wetland acreage seriously contemplated. summaries by quad and boundary files for There already exist a variety of county boundaries by quad where quads approaches towards searching text for were transected. specific purposes. At this point it is anticipated that dBase may be used to index reports, findings, published legal Tabular Data decisions, and similar text data. At some point it might become desirable to As indicated above, one of the explore a fuller support of text data benefits achieved by digitizing the National manipulation. Wetland Inventory Maps was the creation Legal concerns are only one of of wetland acreage statistics. This several that might be taken into account information has, been provided in both when mitigation is attempted. Assuming paper form and on magnetic tape by Fish that there is need to frequently mitigate & Wildlife. Initially these data will serve the impact of natural and cultural forces as a primary reference set, . allowing on non-tidal wetland stability, it might manipulation using dBase III+ of the also be desirable to build a database of statistics by quad and wetland category. scientific information on the study of As the conversion from MOSS to wetlands. Microcomputers have demon- MIPS is completed a duplicate set of strated their ability to easily handle the statistics will be generated, defining both task of searching and organizin@, such area coverage by acre (or any other data. With the increasing reliability of system of measurement), as well as linear optical character recognition (OCR) coverage (in the form of perimeter or software and digital scanners, it may line-length statistics). become desirable to build up an extensive By using the on-screen graphics collection of machine-searchable published utilities it will be possible to not only scientific data. regenerate the U.S. Fish & Wildlife statistics, but to modify those data as a function of modifying the wetland vector (to become arc-node) data. Finally, the linking of a database to the vector files will allow a large col- lateral data base to be established. This Will become the basic engine that will ultimately @rive the management utilities by accessing case studies, published literature, legislation and published regulations, and ancillary raster/vector data. Text @egal reviews demand access to an extensive collection of case histories, court decisions, and findings. In the promulgation of laws governing the protection of such natural resources as 16 V. Conversion of Data to Computerized Form Until recently the task of achieving preted imagery and hand drawn mylars compatibility between software and data are common forms of data of interest not has fallen to those who prepare data. The only to wetland managers but resource result has sometimes been a repetitive managers in general. Raster to vector collection of the same basic data to test conversion through hand digitization is a out new software and to demonstrate the proven method of creating vector files. ti superiority" of each new system over the However, the cost of producing these old. files can be quite considerable. Automa- The danger inherent in customizing ted vectorization is particularly desirable data so that it fulfills software needs is where data of historical importance might that it ceases to be universally usable. be of value if vectorized. For example, Database, spreadsheet and word processing shoreline erosion maps and older land software have led the way in showing that use/land cover maps may be valuable in there is a better way. The first attempt providing comparative data. If they can was to establish a "standard" exchange be introduced in a cost-effective manner format. There are ni@ny and few can into a computerized GIS through automa- claim anything beginning to approach ted data conversion, such maps may universal interchange. The second attempt readily form the basis for longitudinal was to provide conversion utilities. modeling studies. Increasingly this appears to be the correct Complex maps containing numerous solution. Just as a typist is reluctant to hand-drawn annotations, overlain over re-enter the same set of numbers or the photography, such as the NWI maps, may same document again and again to suit the prove too difficult to vectorize by needs of different software to be found automated means. Often the task of within the bureaucracy, so, likewise, is the removing extraneous annotation through resource manager loath to conunit resour- vector editing and supplying information ces to the collection of data in a some- about vectors is as time-consuming and what different form than it had been labor-intensive as redigitizing by hand.,_ collected previously. Thus the choice of whether to attempt MIPS uses the second approach automated digitization rests with the outlined above. By converting bidirection- individual and must be determined on a @lly to and from a composite file form it case-by-case basis. is possible to accommodate the complex Maps and photographs have tradition- file structures associated with both vector- ally been the mainstay of resource oriented GIS systems and the multidimen- managers in their attempt to visualize the sional rasters common to image processing. potential impact of both natural and Another consideration in data cultural forces on the resources they are conversion is the degree to which hard- charged to monitor and manage. With ware can be expected to perform transla- the advent of scanning instruments, tion functions as opposed to manual entry whether mounted on aircraft or satellites, systems. A good case-in-point is the the traditional photo product and conversion of raster to vector data. interpreted resource map declined in Thematic maps, optically merged inter- importance. 17 There is little question th.at scanner include text data into an automated data, whether collected bys@assive sensors system. Published data can, in many (e.g. multispectral scanner or b ctive cases, be directly imported as a com- sensors (e.g. SLAR and LIDAR@, as of puter-readable text file through the use enormous value in resource monitoring. of optical character recognition software. However there are two problems that The cost of performing this translation cannot be easily overcome. First, much of has dropped dramatically over the last the scanner data, =lly that from several years and should continue to do satellites, id - nt resolution so while the sop@istication with which .provi es i e for specific monitoring of localized the translation is carried out will resources. The scanners are especially continue to increase as well. useful in collecting data synoptically, but Ultimately it should be possible to less useful in such tasks as species typing blend many different kinds of data into a and monitoring of heterogenous resources. seamless automated system. Querying Further, airborne sensors tend to produce relational data bases has already proved vast quantities of data that require valuable and will increase in value as the considerable effort to process. The cost link between microcomputer and minicom- and effort associated with cleaning and puter. databases is made easier. Most correcting Daedalus data, for example, may promising here is the common use of SQL be more than a particular project can (@tructured Querying Unguage), which warrant, will allow microcomputers to directly Until recently, photographic and map access mainframe databases without the products could not be easily converted into need to translate from one structure to computer-readable form. Drum scanners another. were both expensive and slow. Optical scanners lacked sufficient resolution to produce good results. The emergence of new flatbed scanners capable of rapidly convertin@ photographs, transparencies, and published maps into di ital form has raised the interest in these florms of data. The major limitations are storage require- ments and overall size restrictions for the media to be placed on the scanners. The continuing interest in vector- based GIS's has created a new 'wealth of data. . The NWI mapping project has already been briefly discussed. The commitment by the USGS to prepare digital line graphs and digital elevation models has contributed substantially to the extant database of vectorized data. Increasingly it is possible to manipulate this data without regard to original scale other than to recognize the inherent limitations of the data. 'Me work currently being carried out in merging these vector data sets with rasters will produce powerful new tools for both management and monitoring of a large variety of natural resources. Finally there is the potential to 18 V1. Manipulation and Analysis of Individual Data Types Rasters The digitized NWI data is much more useful, in many ways, than the original Using the image processing core map products. Even scanning the NWI (MIPS), the non-tidal wetlands management maps would not significantly enhance system is capable of acting on individual their value, since their very complexity rasters for the purpose of either data and extensive annotation make it difficult extraction or analysis. to separate out individual wetland types The best example is the use of digital and categories. data, either satellite data or scanned Vector files can also be linked to photography. , Withoul the aid of the attribute files so that each vector object, computer, the analyst must depend on his whether a point, linear, or polygon, can experience and ability to interpret subtle be linked to a large number of associated spectral variations. The computer can be attributes. Thus items of information used to dramatically highlight those subtle about vectors can be manipulated more differences and ease the work of the easily than items of information about analyst. In the case of the system raster cells. described here that highlighting is done through manipulation of the color lookup tables. The computer can also be used to Tabular Data quantify cell data. The measurement routines described earlier are a good Tabular data can be manipulated example of an application of the com- independently of any other kind of data. puter's ability to quantify raster data. Relational data bases are the most Traditional image processing techni- popular method for manipulating tabular ques such as filtering, density slicing, data. However, it is also possible to contrast stretching, and edge enhancement index tabular files and then use search may also be applied to single rasters. algorithms to gain information contained These techniques are especially useful in within those files. making determinations about land use and Further, tabular data may also be land cover in areas where wetlands presented in a number of ways. Charting abound. programs allow an analyst to view his data in a number of different ways, each possibly suggesting different strategies in Vectors a management context. Increasingly data exploration is enhanced by using the One of the appeals of vector data is computer to chart in response to what-if that objects of any size can be accurately scenarios that can be set up by the represented at virtually any scale based on appropriate software. a single@ digitization effort. Unlike In the case of non-tidal wetlands, rasters, wbich, when enlarged, tend to tabular data that might be the basis for break up into blocky representations, analysis would include such items as vectors are redrawn whenever they are acreage statistics, measures of distance moved to a screen or output device. and proximity between wetlands and other 19 known features, and change percentages regulations, and other documents that the for resource inventories. Rapid manipula- manager might wish to consult in tion of tabular data and easy presentation determining potential impacts on par- of these data through graphing represent ticular resources. The ability to include important tools in the management of text files in a management and monitor- natural resources, especially sensitive ones ing system is likely to be increasingly such as non-tidal wetlands. valuable as the size of the text database increases. Text Text data can often be handled similarl to tabular data. However, it is often Yesirable to organize text by topic area and then provide the analyst with an ?pportunity to review the text on-screen in much the same way that he might use printed documentation. Whereas tabular data must be readable at the element level, text data can be treated as an image and presented to the viewer in image form. Tbus it is possible to scan text without concern for conversion into ascii form. ne resulting image can then be treated in the same way as a photograph or map. Although it might appear that it would be preferable to use optical character recognition techniques to automatically convert printed documents into true text files, there are some important considerations that would suggest the need for an alternative. When documents have been typFset it is far more difficult to reco&nize individual characters since proportional fonts and extensive kerning may impose fairly serious limitations on the success with which these documents can be scanned. In cases where text is heavily mixed with graphics it might again be desirable to simply store an image rather than an ascii document. And in some cases the quality of the original may be such that ocr software would simply not be able to cope adequately with the document. How, text is used in a management environment would depend on the specific arrangements decided upon by those involved in maintai the system. Mention has been maderunf the possible use D of legal landmark decisions, published 20 V1. Manipulation and Analysis of Multiple Data Types Comparison of Multiple Rasters greater detail of the sharper resolution of the RBV. SPOT data may also be treated There are many occasions when it is in similar fashion. 10 meter panchromatic desirable to compare rasters to each other data acquired at the same time as 20 or to manipulate more than one raster meter multispectral data can be combined simultaneously. One obvious case would be to achieve the joint benefits of spectral the creation of a false color infrared and spatial detail. composite from separate data channels Insertion of one raster into another when using scanner data. Both Landsat raster may also be considered a variation and SPOT satellite data, for example, of this category. This technique may be provide the opportunity to composite useful in merging graphic*s' or map data channels so that the result is a color with satellite imagery or photography. composite. Generally raster overlaying is Another case in which multiple accomplished in two steps. First one of rasters might be of interest is the the two rasters is resampled to match the comparison of similar data from different scale of the reference raster. Secondly, seasons or years. For example, a study of a slip-slide technique is used to move one erosion processes might be. meaningful only raster over another so that an exact fit if two or more years of data can be can be achieved. examined simultaneously. Likewise the effect of episodic events, such as storms, can best be understood through a before Overlaying of Rasters and Vectors and after comparison involving the manipulation of multiple rasters. The technique of floating vectors Again, non-tidal wetlands are subject over rasters has already been discussed. to disturbance through a variety of There are other occasions, however, when activities and a multi-temporal comparison the analyst may wish to manipulate a of different rasters may best illustrate the combined raster/vector image. The effect of these activities. easiest way to accomplish this is to simply hardware zoom the combined set. The advantage is speed and reliability, Overlaying of Rasters but as the level of zoom increases, the vector lines begin to loose their precision Overlaying of rasters is an extension and may, in fact, introduce errors into and special case of comparing . multiple the representation of a particular feature. rasters. In this case we assume that the The more desirable alternative would overlaying is done for the purpose of be to zoom the raster plane and redraw resampling one or more rasters belonging the.vector plane simultaneously, maintain- tb different data sets to a common scale. ing the scale relationship. Since it is For example Landsat MSS and Landsat currently possible to use an elastic box RBV data were sometimes merged to to redraw the vectors and to use either provide the combined benefits of multi- hardware zoom or resampling to zoom the spectral information of the MSS with the rasters, software implementation to permit 21 the two processes to occur simultaneously When multiple vectors are overlaid will likely occur int he near future. their associated attributes may also be examined. Conversely, specific combina- tions of attributes may be extracted and Overlaying of Multiple Vectors their associated vector objects displayed. These superimposed vector objects may Because of their relatively modest then be subjected to set theory analysis space requirements, multiple vector files, in order to reduce them to common or files containing multiple levels of data, objects.' serve as the basic data form for most MIPS is currently being modified to current GIS's. Superimposition of vectors, permit linking a number of different vector clipping, boolian operations on databases, beginning with dBase III+, to combined vectors, and generation of the vector planes within combined multiple vectors in one plane based on raster/vector files. Because the raster/- criteria derived from an examination of an vector file structure provides an automa- associated database, are powerful proced- ted link between rasters and vectors, the ures that increase the inherent value of potential for increasing the power of this registered vector data. system beyond current GIS's clearly Although polygon format data storage exists. might appear to be the most likely manner in which vector data can be handled by a computer, the arc-node format is ultimate- Extraction of Text Data from Multiple ly more powerful since it allows assigning Quads exclusive boundaries to adjacent polygons and requires less effort to support where In completing this overview of the changes might be made to the data set on strategy bein@ laid out for the manipula- a periodic basis. tion of multiple data types, it is neces- Further, the support of multiple sary to reconsider the place of text and storage schemes for vector data creates tabular data. the possibility to interface with other Text data can be supplied either in software packages. Cogo and Cad/Cam image or in character form and linked to packages are increasing in popularity and graphically derived information through add-in routines to make them suitable to an external database. A more powerful either mapping or specific -resource approach, however, is to allow multiple management needs will likely become more text files to associate with multiple widespread in the future. T'he ability to vectors and quads. @andle different vector formats is as . Thus, ideally, it would be possible to important as the ability to handle different query a text database by example and raster formats. relate the querying strategy to a graphic Tagging objects defined as polygons representation. dBase IV is designed to is not only possible, but the basic permit this kind of querying by example technique used by all geographic informa- activity. The release of dBase IV should tion systems. The alternative is to define occur sometime in the first quarter of a raster cell of a known size. This 1988. alternative has shown to be cumbersome, The computerized non-tidal wetlands often inaccurate, and difficult to manipu- management system should be able to late. Vector tagging is done in the form take full advantage of the enhanced dbase of associating attributes with vector IV capabilities. The first step would be objects. These object most commonly are to test the querying by example feature. polygons, although they could also be This should be followed by a structured linears or points. Multiple attributes can linking to the graphics file structure. be assigned through the use of a database. 22 Linkage of Graphic Wetland Representation to Tabular Data Tabular data usually consists of a calculated extraction from a database. Acreage reports, for example, are often tabulated by political jurisdiction to give some indication as to the total resource component being considered. Projections of increase or loss, threat or damage to a resource may also con- stitute a tabulation. Such tabulations may be derived in a number of ways, although it is anticipated that for the non-tidal wetland data most of the tabulations will be derived from the areal computations performed on specific wetland groups by quad and county. Customarily, representations of tabular data are in the form of a table or graph. Additionally, the tabulation may be reflected by highlighting a representative Tap. This latter feature is of particular interest to resource manarrs, since it provides a quick visual confirmation of a geographically known phenomenon. The non-tidal wetlands management system contains a fairly extensive set of routines to provide graphic output capabilities, including the ability to enhance, highlight, and otherwise show a particular aspect of interest on a map or segment thereof Tabular data, inasmuch as it reflects operations performed on the graphics planes, can be used to visually alter that graphics plane for better illustrative effect. A number of drawing functions are available, as are lettering and other annotation features. Numerous fonts, and such drawing primitives as sized rec- tangles, circles, and polygons provide further flexibility in creating an attractive visual and output product that reflects tabular data. 23 V11. Conclusions T'he work described in this report ities is the ability to "view" data in a represents an attempt to assemble a form familiar to the photogrammetrist, to combined hardware and software system make measurements, enhancements, that builds on past efforts and attempts to annotations, and data layer overlays. The provide a platform for future development. easy, intuitive interface is another major Past efforts have focused on the feature of the software, requiring a fairly evaluation of various remote sensing short learning curve for casual use. techniques and data sources for general Powerful features such as three-dimen- resource management and @specifically for sional modeling and raster filtering non-tidal wetland- evaluation. It has require some knowledge of the nature of become clear that the greater the integra- the data and some skills in remote tion of a varied database, the more sensing, but it is expected that every-day powerful the tools are that can be use of the system will be easy, intuitive, developed to manage this database. and productive. Thus, the . creation of the com- During the course of testing the puterized non-tidal wetlands monitoring hardware and software it became clear and management system is dedicated to the that the system would mature well beyond concept that every conceivable form of the scope of the non-tidal wetlands data that might be of interest to the project. Some of the capabilities wetlands manager should find a lace developed as incidental to the Froject within one or more of the related @ata- have since been tested as part o other bases that constitute the total system. projects and are reported on elsewhere. Map products, digital products, photog- For the present it is appropriate to raphy, vectorized plots, text and tabular return to a discussion of the resource data can all be integrated under the that the project was to help serve: non- system being implemented under this tidal wetlands. project. There are both sound ecological and What is provided' here is both a legal reasons for protecting non-tidal concept and a capability. The concep@ is wetlands. In Maryland, all local jurisdic- one of using a computer to organize tions are required to afford protection to diverse data, present it in a meaningful non-tidal wetlands that meet any of the manner,,whether through text or graphics following criteria (as published in or both, and allow.the analyst or manager "Guidelines for Protecting Non-tidal to use the computer as an interaction tool. Wetlands in the Critical Area", 1987:4): To this end state-of-the-art hardware and software have been coupled in what we 1. All palustrine non-tidal wetlands believe, is a unique and exciting system classified as Aquatic Bed, that is applicable not only to non-tidal Emergent, Scrub-shrub, ot' wetlands management concerns but to a Forested ... of one acre or broad spectrum of resource management larger in size that are shown on tasks. The hardware capability is matched the National Wetlands Inventory by a software capability. Among the most Maps or which may be found by important and innovative of these capabil- onsite survey; 24 2. Such palustrine wetlands not installed as part of the non-tidal wetlands shown on the Maps that are program, should be seen as a tool. In hydrologically connected to the future it is hoped that refinements, streams, tidal waters, or tidal enhancements, changes, and other wetlands, and applications will be found to broaden the 3. Such palustrine wetlands not effectiveness of this tool and that his shown on the Maps that have project represents a beginning to a special importance to fish, solution for non-tidal wetlands monitoring wildlife, or plant habitat. and management, not a definitive and final solution. Guidance Paper No. 3 of the Chesa- peake Bay Critical Area Commission, "Guidelines for Protecting Non-tidal Wetlands in the Critical Area", 1987, provides information on the identification of non-tidal wetlands and the means to be used to establish required setbacks. The paper also discusses the circumstances and conditions under which alterations to a wetland may be allowed, and guidance on appropriate mitigation measures and techniques. It is hoped that the com- puterized system as it becomes imple- mented will enhance the purpose of the Guidance Paper and expand the capabilities of the analyst in a fashion consistent with the goals of the guidance paper. The computerized system provides, at best, a concept for handling large amounts of diverse but related information, and a capability to explore this database quickly and productively. What the system does not provide is the intelli&ence to -interpret the information it contains. There is no substitute for the experience and skill of the human manager. . Although some expert-system capabilities are inherently a part of the expanded system, such capabilities pale when compared to the sophistication of the experienced wetlands manager or field knowledge of the dedicated naturalist. The system described here is a tool, not a solution, and to that end it can be used with great success or abused to the detriment of the resource being managed. This is not intended to be a caveat, but rather a challenge. Tools can be sharpened and honed, applied to new tasks or picked up regularly to perform old and familiar tasks. The computerized system which has been described here and which has been 25 Vill. Appendices [provided as Supplements] A. System User's Guide B. Technical Details - Hardware C. Technical Details - Software D. Reference Data Sets (on optical disks) E. dBase III+ files (on Bernoulli cartridges) F. National Wetlands Inventory Map Preparation 26 3 6668 14109 5026