[From the U.S. Government Printing Office, www.gpo.gov]
CO ,oastal Zone 'Olt OF Information Center 04, NOAA Technical ftmorandum NWS HYDRO-23 t 'T)@4 ES O@ JUL 1 1975 STORM TIDE FREQUENCY ANALYSIS FOR THE COAST OF PUERTO RICO Francis P. Ho f 'Office of Hy4rology rn ATION CENTLR' Si 1 ver Spring, Md. May 1975 QC 851 X6 H9 no.23 COASTAL loon NAT04AL OCF-AMC AND National Weather ATMOSPHERIC AMNISTRATION Service NOAA TECHNICAL MEMORANDA National Weather Service, Office of Hydrology Series The Office of Hydrology (HYDRO) of the National Weather Service (NWS) develops procedures for making river and water supply forecasts, analyzes hydrometerological data for planning and design criteria for other agencies, and conducts pertinent research and development. NOAA Technical Memoranda in the NWS HYDRO series facilitate prompt distribution of scientific and technical material by staff members, cooperators, and contractors. Information presented in this series may be preliminary in nature and may be published formally elsewhere at a later date. Publication 1 is in the former series, Weather Bureau Technical Notes (TN); publications 2 to 11 are in the former ser- ies, ESSA Technical Memoranda, Weather Bureau Technical Memoranda (WBTM). Beginning with 12, publica- tions are now part of the series, NOAA Technical Memoranda, NWS. Publications listed below are available from the National Technical Information Service, U.S. Depart- ment of Commerce, Sills Bldg., 5285 Port Royal Road, Springfield, Va. 22151. Price $3.00 paper copy; $1.45 microfiche. Order by accession number shown in parentheses at end of each entry. Weather Bureau Technical Notes TN 44 HYDRO 1 Infrared Radiation from Air to Underlying Surface. Vance A. Myers, May 1966. (PB-170- 664) ESSA Technical Memoranda WBTM HYDRO 2 Annotated Bibliography of ESSA Publications of Hycrological Interest. J.L.H. Paulhus, February 1967. (Superseded by WBTM HYDRO 8) WBTM HYDRO 3 The Role of PErsistence, Instability, and Moisture in the Intense Rainstorms in Eastern Colorado, June 14-17, 1965. F.K. Schwarz, February 1967. (PB-174-609) WBTM HYDRO 4 Elements of River Forecasting. Marshall M. Richards and Joseph A. Strahl, October 1967. (Superseded by WBTM HYDRO 9) WBTM HYDRO 5 Meterological Estimation of Extreme Precipitation for Spillway Design Floods. Vance A. Myers, October 1967. (PB-177-687) WBTM HYDRO 6 Annotated Bibliography of ESSA Publications of Hydrometeorological Interest. J. L. H. Paulhus, November 1967. (Superseded by WBTM HYDRO 8) WBTM HYDRO 7 Meterology of Major Storms in Western Colorado and Eastern Utah. Robert L. Weaver, January 1968. (PB-177-491) WBTM HYDRO 8 Annotated Bibliography of ESSA Publications of Hycrometeorological Interest. J. L. H. Paulhus, August 1968. (PB-179-855) WBTM HYDRO 9 Elements of River Forecasting (Revised). Marshall M. Richards and Joseph A. Strahl, March 1969. (PB-185-969) WBTM HYDRO 10 Flood Warning Benefit Evaluation - Susquehanna River Basin (Urban Residence). Harold J. Day, March 1970. (PB-190-984) WBTM HYDRO 11 Joint Probability Method of Title Frequency Analysis Applied to Atlantic City and Long Beach Island, N.J. Vance A Myers, April 1970. (PB-192-745) NOAA Technical Memoranda NWS HYDRO 12 Direct Search Optimization in Mathematical Modeling and a Watershed Model Application. John C. Monro, April 1971. (COM-71-00616) (Continued on inside back cover) 0 0 NOAA Technical Memorandum NWS HYDRO-23 STORM TIDE FREQUENCY ANALYSIS FOR THE COAST OF PUERTO RICO Francis P. Ho Office of Hydrology U.S. DEPARTMENT OF COMMERCE NOAA Silver Spring, Md. COASTAL SERVICES CENTER May 1975 2234 SOUTH HOBSON AVENUE CHARLESTON, SC 29405-2413 COASTAL ZONE INFORMATION CENTE Property of CSC Library UNITED STATES NATIONAL OCEANIC AND National Weather DEPARTMENT OF COMMERCE ATMOSPHERIC ADMINISTRATION Service Rogers C. B. Morton, Secretary/ Robert M. White, Administrator George P. Cressman, Director STORM TIDE FREQUENCY ANALYSIS FOR THE COAST OF PUERTO RICO CONTENTS Page Abstract 1 1 Introduction 1 1.1 objective and scope of report 1 1.2 Authorization 2 2 Summary of historical hurricanes 3 2.1 Introduction 3 2.2 Hurricane tracks 3 2.3 Historical notes 3 2.3.1 September 12-19, 1876 3 2.3.2 August 18-25, 1891 3 2.3.3 August 13-25, 1893 3 2.3.4 August 3-24, 1899 4 2.3.5 August 21-25, 1916 4 2.3.6 September 6-20, 1928 4 2.3.7 September 8-16, 1931 4 2.3.8 September 25-october 3, 1932 5 2.3.9 August 9-19, 1956 - BETSY 5 3 Climatology of hurricane characteristics 7 3.1 Hurricane parameters and frequency 7 3.2 Definition 7 3.3 Data sources 7 3.4 Probability distribution of hurricane central pressure 8 3.5 Probability distribution of radius of maximum winds 9 3.6 Probability distribution of speed of storm motion 9 3.7 Frequency of tropical cyclone tracks 10 3.7.1 Seaward alongshore storms, San Juan 10 3.7.2 Landward alongshore storms, San Juan 10 3.7.3 Landward alongshore storms, south coast 10 3.7.4 Seaward alongshore storms, south coast 10 3.7.5 Landfalling storms, south coast 10 3.7.6 Exiting storms, north coast 10 3.7.7 West coast 11 3.7.8 East coast 11 3.8 Track density method 11 3.8.1 Track frequency 11 3.8.2 Direction probability 12 3.8.3 Ppplication at a point 12 3.8.4 Landfalling frequency on southern coast 12 3.9 Probability distribution of tropical cyclone direction 12 of motion 0 Pane 4 The hurricane surge model 15 4.1 The model 15 4.2 Application of model to Puerto Rico 15 5 Storm tide frequencies by joint probability method 17 5.1 Example--San Juan 17 5.2 West coast 18 5.3 South coast 18 5.4 Extrapolation to uncharted areas 18 5.5 Pressure-wind relationship 19 6 Wave effects 23 6.1 Introduction 23 6.2 Wave problem in Puerto Pico 23 6.3 Wave swash study for extratropical storms 23 6.4 Estimated hurricane effects 24 7 Demarcation of flood prone limits 25 7.1 Objective 25 7.2 Precipitous coastal areas 25 7.3 Flat coastal areas 25 References 26 Picures 27-43 NOTE This report was prepared for the Federal Insurance Administration by the National Weather Service, NOAA, in August 1973. It is being issued in this form for wider distribution. STORM TIDE FREQUENCY ANALYSIS FOR THE COAST OF PUERTO RICO Francis P. Ho Special Studies Branch, Office of Hydrology National Weather Service Silver Spring, Md. 20910 A report on work for the Federal Insurance Administration Department of Housing and Urban Development by National Oceanic and Atmospheric Administration, Department of Commerce ABSTRACT Storm tide height frequency distributions on the coast of Puerto Rico are developed for the National Flood Insurance Program by computing storm tides from a full set of climat6logically representative hurricanes, using the National Weather Service hydrodynamic storm surge model adapted to local conditions. Hurricane parameters needed for the storm tide computation are analyzed and presented. Tide levels for the southern coast of Puerto Rico are shown in coastal profile between annual frequencies of 0.10 and .002. Similar tide levels for San Juan on the northern coast and Mayaguez on the western coast are also presented, and tentative interpolated tide levels of .01 annual frequency for the remaining coasts. Wave effects on the north coast are discussed. Chapter 1 INTRODUCTION 1.1 objective and scope of report The Federal Insurance Administration (FIA), Department of Urban Develop- ment (HUD), requested the National Oceanic and Atmospheric Administration (NOAA) to delineate the 100-year flood line from storm tides on the coast of Puerto Rico. This study is part of the survey of the whole United States by FIA to delineate those areas and communities subject to flooding, required by Section 1360(l) of the National Flood Insurance Act of 1968 and is called a "Type 7" study by the FIA. Over much of the United States this type of delineation, more limited than a flood insurance rate-making study ("Type 15"), has proceeded quadrangle by quadrangle of the U. S. Geological Survey topographic series, from readily available stream flow or tidal data. For coastal Puerto Rico there are no such readily available storm tide data. Therefore, NOAA found it necessary to develop data in the same series of steps used in hurricane zones on the mainland coast for "Type 15" rate-making flood frequency analyses. These steps are: (a) establish the climatological frequency of certain hurricane parameters from past storms, (b) construct smoothed-out depictions of the sea surface bottom from the coast to a 300-foot water depth, (c) compute storm surges for various coastal points from representative climatolog- ical hurricanes with a hydrodynamic model, (d) add the storm surge to astronomical tide at various phases, (e) collect the computed total .storm tide elevations into a flood frequency diagram, (f) read 100-year flood levels from these diagrams, and (g) construct the 100-year flood line on maps, making allowances for wave action on the immediate beach front. A locator map and smoothed off-shore water depths are shown in Figure 1-1. In view of the reconnaissance intent of the study, short cuts were taken wherever possible. However, the basis has been laid for a rate- making "Type 15" study if this should be undertaken in the future. 1.2 Authorization The National Flood Insurance Act of 1968, Title XIII, Public Law 90-448, enacted August 1, 1968, authorizes and directs the Secretary of Housing and Urban Development to establish and carry out a National Flood Insurance Program. The Secretary is authorized to secure the assistance of other Federal Departments or other agencies on a reim- bursement basis in identifying flood-plain areas, including coastal areas. Authorization for this particular study is Project Order No. 4, Agreement No. IAA-H-5-73 dated May 2, 1973, between the Federal Insurance Administration (FIA) of the Department of Housing and Urban Development (HUD) and NOAA. Acknowledgment Three components of NOAA collaborated on the Puerto Rico assignment. The hurricane climatological analysis and computation of tide levels are by the Special Studies Branch, Office of Hydrology, National Weather Service. Definition of the Bathymetry (water depth) and adaptation of the storm surge computation model to Puerto Rico are by the Storm Surge Unit, Techniques Development Laboratory of the Systems Development Office, National Weather Service. Map preparation and coor- dination are by the Coastal Mapping Division, National Ocean Survey. The study was funded by the Federal Insurance Administration, Department of Housing and Urban Development. -2- CHAPTER 2 SUMMARY OF HISTORICAL HURRICANES 2.1 Introduction This chapter summarizes the hurricanes that have moved across the island of Puerto Rico since 1871. A few of the lesser storms are o mitted. Prior to the year 1871, there were three prominently m. noted hurricanes which inflicted severe damage to the island. These are the hurricanes of Santa Ana on July 26, 1825; Los Angeles on August 2, 1837; and San Narciso on October 29, 1867. It is cus- tomary in Puerto Rico to name a hurricane after the saint on whose day it happens to occur. There were no meteoroloaical data available on these storms and records are incomplete and inaccessible. 2.2 Hurricane tracks The tracks of the hurricanes are shown in Figure 2-1. 2.3 Historical notes Brief notes on the history and damages caused by the hurricanes abstracted from published papers follow: 2.3.1 September 12-19, 1876 The center of this storm entered Puerto Rico between Humacao and Yabuca and exited between Rincon and Mayaguez. The lowest barometric pressure recorded was 989.5 nil-, (29.22 in.) at San Juan. This storr was extremely destructive. In the San Juan area, 45 ships were lost and severe darnage was done in the land areas. 2.3.2 August 18-25, 1891 One of the nost disastrous of West Indian hurricanes occurred on August 18-19, 1891. Its center approached Puerto Rico from the southeast, passing through Martinique on August 18th and causing $10 million damage and the loss of 700 lives on that island. The storm entered Puerto Rico on Auqust 19th and exited the north coast later that eveninc. The storm was of snall diameter but of great intensity. This storm was referred to as the Martinique hurricane 2.3.3 August 13-25, 1893 This tropical storm was first detected on 71.ugust 13 southeast of the Ulindward Islands moving in a northeasterly direction. Its center 3 0 entered the southeast coast of Puerto Rico near Punta Guayama in the afternoon of the 16th and exited the island near Isabela. The lowest barometric pressure recorded at San Juan was 988 mh (29.17 in.). 2.3.4 August 3-24, 1899 One of the most destructive hurricanes in Puerto Pican history passed directly across the entire length of the island on August 8. The storm has become known as "San Ciriaco." More than 3,000 lives were lost, mostly from drowning. Property loss was enormous. The storm center entered near Arroyo on the southeast coast at about 8 a.m. and the calm in the "eye," or center, lasted about 15 minutes. The barometric pressure fell to 939.7 mb (27.75 in.). The maximum wind speed at Arroyo was estimated at about 90 knots. A storm tide destroyed almost all of the houses at the port of Humacao on the east coast. Destructive winds and torrential rains accompanied the hurricane in its progress across the island. between 1 and 2 p.m. the center left the island near Aquadilla on the west coast. 2.3.5 August 21-25, 1916 This small size storm moved across Puerto Rico in an east to west direction on August 22. Its center passed just south of San Juan which experienced a maximun wind of about 80 knots, and a minimum barometric pressure of 997 mb (29.44 in.). An estimated million dollars damage was inflicted in Puerto Rico by this storm along its path of about 45 to 50 miles in width. 2.3.6 September 6-20, 1928 The most intense hurricane of the 20th century to strike Puerto Rico was first reported at about 300 miles east of the Leeward Islands. Its center moved west-northwestward and entered the southeast coast of Puerto Pico, near Guayama. The hurricane moved across the island at an average speed of about 11 knots and exited the north shore between Aquadilla and Isabela. A minimun barometric pressure (931.3 mb or 27.50 in.) in the vicinity of Puerto Rico was reported by the stearnship Matura located at about 10 miles south of the island of St. Croix. The lowest barometric pressure recorded at Guayama was 936.3 mb (27.65 in.) Maximum winds of about 130 knots were recorded at San Juan, some 30 miles to the north of the hurricane center. Winds of hurricane force were experienced throughout the island to the north of tine path. Approximately 300 persons in Puerto Rico lost their lives in this storm. hundreds of thousands of people lost their homes during the storm passage. Property and crop losses werc estimated at about $50 million. 2.3.7 September 8-16, 1931 This hurricane which was named after San Nicolas raked the north coast of Puerto Rico on September 10. The storm had almost developed - 4 - into hurricane intensity when it passed to the north of the Virgin Islands. By the tire it reached San Juan, hurricane winds were estimated at about 80 knots and a low barometric pressure of 987.8 mb (29.17 in.) was reported. The hurricane noved in a westward direction alonq the entire north coast of Puerto Rico. Damage caused by the storm was confined to a strip of 5 or 6 miles in width extending from San Juan to Aguadilla. 2.3.8 September 25-October 3, 1932 This hurricane is known as "San Ciprian" in Puerto Pico. Its center crossed the island on September 26, entering the eastern shore near Ceiba in the evening. A barometric pressure of 938 mb (27.70 in.) was recorded on board of the S. S. Jean in the harbor of Ensenada Honda. A minimum pressure of 980.4 irb (28.95 in.) was recorded at San Juan as the center of the hurricane passed some distance to the south, with estinated maximum winds of more than 100 knots. 225 lives were lost in Puerto Rico and 3,000 or more persons were injured in this storm. An estimated 75,000 to 250,000 persons were left homeless after the storm passage. Property damage in the island was estimated at $30 million.- 2.3.9 August 9-19, 1956 - BETSY Hurricane BETSY was first detected on August 9 to the east of the Lesser Antilles. Its center moved westward to west-northwestward and Passed over the island of Guadeloupe, F.W.I., on August 11. Continuing on its west-northwestward course, the hurricane reached the southeastern coast of Puerto Rico on the 12th. The center of hurricane BETSY entered the coast-near Guayaria and exited the northern coast of the island near Arecibo at an average speed of 18-19 knots. A minimum sea level pressure of 983 mb (29.03 in.) was recorded at Guayama shortly after the hurricane entered land. A mininum pressure of 987 mh (29.15 in.) was recorded at Ramey AFB at about 14 n. mi. to the south of the center. Based on this observation, Col6n Ell estimated a central pressure of 973 mb (28.73 in.) for the hurricane as it exited the northern shore. Maximum winds of 100 knots were reported by reconnaissance aircraft on August 10 when the hurricane was located some distance to the west of Puerto Pico. The maximum wind speed of about 65 knots with gusts to 80 knots was recorded at San Juan airport. Sixteen persons lost their lives in the storm in ruerto Rico and the damage inflicted on the island was estimated at $40 million. 5 CIT TER 3 MP CLIMATOLOGY OF 1JURRICA17, CHARACTrRISTICS 3.1 Hurricane Parameters and Frequency The hurricane parameters needed for storm tide computation are: central pressure, radius of maxinum winds, direction of motion relative to the coast, and the speed of forward motion. If track is parallel to the coast, then distance from the coast is needed instead of direction. Probability distributions are required for each of these parareters to evaluate tide frequencies. Also needed is the overall frequency with which hurricanes enter the coast in terns of strikes per rile per year, or sore equivalent unit, and frequency with which hurricanes pass parallel to the coast within certain discrete distances. Storrs exitina the coast are also counted where significant. 3.2 Definition For convenience, a distinction is made in this chapter between "frequency" and "Probability." Frequency is defined as the number 1 of occurrences of some event per year and has dimensions of tine while probability is the fractional part of a total and is dirensionless. 3.3 Data sources The data for probability distributions of the variables central pressure, radius of maximum winds, and forward motion are primarily from Navy reconnaissance flights for the period 1945 throuqh 1972. Only storrs reported to reach hurricane intensity (winds > 64 knots) are used in this part of the study. These flight data are abstracted fror Annual -.roT-)ical Storn Reports of the U. S. Navy [3] and fror- oriainal flight data obtained fror the National Clinatic Center. Colon and Dunn and Miller [ref. 1 and 41 furnished additional values. The statistics on the frequency of storm occurrences and on direction of storw mtion in this report are based on the yearly storm track charts by Cry [21 from 1871-1963 and fron Monthly Weather Review articles between 1964-1972. Poth hurricanes (winds cyreater than 64 knots) and tropical storms (inaxinurn wind 34 to 64 knots) are included in the statistics since the distinction on the track charts is not ,always clear. Storrs classified as "tropical depressions" (less than 34 knots) are not included -7- 0 3.4 Probability distribution of hurricane central pressure Hurricane surges vary directly with the depression of the storm's central pressure below a representative peripheral pressure, other factors being equal. Thus, central pressure is a convenient intensity index. The real driving force for the surge is the stress of the wind on the water, rouqhly proportioned to the square of the wind speed. But the wind speed squared results from its driving force, the pressure depression. To obtain a probability distribu- tion of central pressures, reports over the area bounded by latitude 15-20*N and longitude 65-70*N were examined duringthe period of 1945 throuqh 1972 and are believed complete. A list of hurricanes (wind reported > 64 Kt) and their adopted parameters are shown in Table 3-1. The central pressure of the August 1956 hurricane is from adjusted minimum pressure for Ramey APB after Colon [1] and the central pressure for hurricane EDITH of September 1963 is from Dunn and Miller [4]. Both of these pressure values are lower than the minimum pressure reported by reconnaissance aircraft over the study area. Ficure 3-1 shows the probability distribution of hurricane central pressure based on the Table 3-1 values. To fit a curve to these data that refers to the same statistical population of storms as the track frequency count to be described later, we proceed in the same manner as in an earlier report([5] p. 12). The track frequency count includes both "hurricanes" and "tropical storms." The data points were plotted by using the formula P = (M-0.5)/N, where P = accumulative probability M = "rank" of observation N = number of observations. By track chart count there were 29 tropical cyclones ("tropical storms" plus "hurricanes") through the study area during the 28 study years. Only 17 central pressures are listed in Table 3-1. These are assured to cover the "hurricane" range. N in the formula is then 29 and the pressures are arrayed and plotted as if they were the 17 most intense cases of 29. A curve is then drawn by eye to the data and extended smoothly to cover the "tropical storm" range. The central pressures reported by aircraft for the less intense hurricanes seem. too high to support winds of hurricane intensity. Possibly the reported minimum pressure does not coincide with the actual minimum central pressure due to the uncertainty in determining the exact location of a hurricane center when its eye wall was not well defined (which is often the case in a storm barely reachinq hurricane intensity). For these reasons the fitted curve is pulled to the left of the upper eight points plotted in the diagram. - 8 - 0 For comparison with an adjacent area for which longer quantitative records are available, Figure 3-2 shows, on the same graph, the probability distribution of tropical cyclone central pressures for the Puerto Rico area (18*N) from Figure 3-1 and the southern tip of Florida (25*N) from [6]. There is a difference of about 10 mb if the percentage of storm occurrences is compared level for level. The Puerto Rico curve is comparable to the Palm Beach-Daytona Beach reach of the Florida east coast in [6]. If this 28-year sample is representative, tropical cyclones originating in the Caribbean Sea may not reach their full intensity in the Puerto Rico area. 3.5 Probability distribution of radius of maximum winds In all hurricanes, proceeding from the storm center outward, winds increase from low values at the center of the eye to their most intense velocity just beyond the edge of the eye, then decrease. The average distance from the storm center to the circle of maximum wind speed is called the radius of maximum winds (P) and is adopted as a convenient single number to be used as an index of the size or lateral extent of the hurricane, a factor which affects the surge profile along the coast. Most of the R's in Table 3-1 are the average distance from center to the maximum wind belt, from the aircraft reports. Where the distance range for this band was quite wide, additional_guidance was obtained from the reported radar eye radius, adjusted to R by using the radar eye radius-R-central pressure diagram from Shea [7] - figure 26. ) The probability distribution of the radius of maximum winds for hurricanes in the Puerto Rico area is shown in Figure 3-3. The curve reveals that more than 50% of the storms passing the study area were small in size (R < 12 n. mi.). This is consistent with small size storms generally observed in tropical latitudes. Only about 20% of the storms had R greater than 15 n. mi. The minimum R in the sample of this study was 4 n. mi. 3.6 Probability distribution of speed of storm motion The probability of distribution of the speed of forward motion of hurricanes in the vicinity of Puerto Rico from the data in Table 3-1 Is shown in Figure 3-4. The height of surge on the coast increases with increas1ng storm speed within the range or observed values in the study area [9] because of dynamic effects in the water. Thus, the occasional fast-moving storms, especially if they are large, pose the greatest hazard. Only one hurricane had a speed of 22 knots. The absence of high speed of forward motion is characteristic of low latitudes. An inspection of hurricane tracks within the study area (bounded by 15*-20*N and 65*-70*W) for 102 years reveals only two hurricanes having a forward speed of 22 knots (estimated from 24-hr positions) during the period of 1871 through 1972, thus Figure 3-4 is in general representative of the longer period. - 9 - 3.7 Frequency of Tropical Cyclone Tracks The tide frequency analysis treats three classes of storns separately, i.e., landfallino, exitina, and alongshore sterns, because the dynamic model described in Chapter 4 is set up to handle these separately. It is, therefore, loqical to examine freauency of tropical cyclone occurrences separately according to these three predeterrined categories, and this was done for the north, west, and south coasts. The count methods are explained in the following paragraphs and are summarized in Table 3-2. The predominant track directions are fror the east and ESE (Figure 2-1). 3.7.1 Seaward alongshore storms, Fan Juan L north-south line was drawn on a nap through San Juan. Tropical cyclone tracks ("hurricanes" 'plus "tropical storms") crossing this line were counted for the period 1871-1972. Storns that had crossed land east of the line--that is, storrrs that passed over the northeast corner of the island and then exited the coast--are omitted from the seaward count. The accurulated seaward count is shown in the right half of Figure 3-5. All crossinc-s were from. east to west. 3.7.2 Landward alongshore storms, San Juan Storms that crossed the north-south line through San Juan on the landward side of the city were counted; the accumulated frequency is shown on the left half of Ficure 3-5. 3.7.3 Landward alongshore storrrs, south coast Storms crossinq 66*30' north of the south coast were counted, with the accumulated freauency for the 102 years shovm in the left half of Figure 3-6. 3.7.4 SeaWard alongshore storms, south coast Storms crossing 6611301 south of the Puerto Rico coast were countedf except those striking-the island (on the southwest corner) were omitted. The accurnulated count is shown in the right half of Ficure 3-6. 3.7.5 Landfallinq storms, south coast The frecTuencv of lanefallincy storrs on the south coast was estir:.ated by the track density rethod described in the next section. Fxiting storms on the south coast are so rare they are neglected. 3.7.,6 Exiting storms, north coast Since the north and south coasts are essentially parallel and the - 10 north-south cradient of storm frequency is slight, the exiting storm frequency on the north coast was assumed to be the same as the landfalling frequency on the south coast without making any calculations. Landfalling storms on the north coast are neglected. 3.7.7 West coast "Exiting" storms is the onlv important category for the west coast. An estimate of the frequency was obtained by averaging the "landward alongshore" frequencies for San Juan and for the south coast. 3.7.8 East coast No computations were made for the east coast. 3.8 Track density method The frequency with which hurricane storm tracks landfall on the southern Puerto Pico coast is more complicated than "alongshore" because of the small angle hetween tracks and coast and the varying coastal directions. In order to handle this in a straightforward manner we resort to the track density method described in this section. 3.8.1 Track frequency The frequency of occurrence of tropical cyclones may be expressed as storm trac density at a point. This is defined as the number of storm tracks which cross that point, from any direction, per unit length normal to track per unit time. In concept, on obtains this number by a limit process which may be expressed as: lim (N ) = storm track density at a point, Dt D - 0, t where N is the count of tracks passing through a circle of diameter D in time t. Practically it is necessary to count storm tracks passing through a large enough circle over a long enough period of time to smooth out random fluctuations. This was done by counting tropical cyclone tracks from the sources mentioned earlier over 2.5' longitude squares with the corners cut off to approximate circles. The number of storms passed through each of these 2.5' "octagons" is shown in Figure 3-7 between 15' and 20'N and 60' and 70'W. These values approximate the number of storms passed through a circle of 150 nautical miles diameter per 102 years. A smoothing analysis of these numbers yields the estimate of "point track density." The adopted count for the southern coast of Puerto Rico, by visual analysis of Figure 3-7, is 49. This is equivalent to 3.2 storms per 10 nautical miles per 100 years: 49 x 10 100 3.2 150 102 3.8.2 Direction probability The second part of the track density analysis for each 2.5* octagon is the track direction. Track directions by 15* class intervals were counted for the 102 years for each octagon. Histograms like Figure 3-8 were then constructed for each octagon. The ordinate is the normalized frequency of occurrences (f/lli), where f is the count in a direction class interval, N the total count for all directions for the octagon and i the class interval width (in decrees). A "Beta" distribution function was then fitted to each of these histograms by using a computer program listed in the "IBM Scientific Subroutine Package." An accumulated probability curve is obtained by integrating the "Beta" curve. An example of these plots is shown in Figure 3-8. 3.8.3 Application at a point Finally, the frequency with which storms enter a coast from a particular direction span is a product of the point track density for the region, the fraction of the total storms within the specified direction span, and the sine of the angle between the coast and the storm direction class interval. Example: Given that a portion of the southern coast of Puerto Rico is oriented 90*-270*, the overall track density is 3.2 storms per 10 nautical miles per 100 years and 30% of the storms core from directions between 115* and 135*. (Average direction 125* = 35* to coast.) The computed count of landfalling storms from this direction interval is 3.2 x .3 x sine 35* = .55 storms per 10 nautical miles of coast per 100 years. Carrying out this operation through 180* gives the total landfalling frequency. The exiting frequency can be handled in the same way. 3.8.4 Landfalling frequency on southern coast The accumulated direction of motion curve from Figure 3-8 for the Puerto Rico octagon is replotted in Figure 3-9 on a larger scale. Direction of motion between 115* and 200* is construed as "landfalling" for east-west portions of the southern coast. This 50% of the total storms is grouped into the three class intervals shown by the dotted line in the diagram: 125*, 30%; 150*, 18%; 180*, 2%. Applying the procedure in the example above to each class interval in combination with the track density of 3.2 storms per 10 n. mi. per 100 years gives a total landfalling frequency of 1.1 per 10 n. mi. per 100 years. 3.9 Probability distribution of tropical cyclone direction of motion This is one of the factors required for surge computation for landfalling and exiting storms. It is given by Figure 3-9. - 12 - TABLE 3-1 Hurricane Parameters Peference n Trax. eye -oint cri R T wind diar. Lat Long Storn, Date Name (mb) (n.mi.) (kt) (kt) (n.ri.) Oil ow 1947 Sep 4-21 952 -- 15 120 7 22.3 66.6 1950 Aug 20-Sep 1 PAYER 990 8 11 90 12 17.1 61.0 1950 Aug 30-Sep 16 DOG 962 18 10 120 20 19.6 64.4 1951 Aug 12-23 CHARLIE 978 -- 22 90 22 16.5 72.9 1953 Aug 28-Sep 9 CAROL 929 11 16 130 3 19.2 60.4 1955 Aug 3-14 COT-RUE 952 12 16 120 20 19.5 63.7 1956 Aug 9-19 BETSY 973 14 16 108 16 17.9 66.1 (after Col6n) 1956 Aug 30-Sep 6 ELLA 119183 20 16 95 30 16.8 70.5 1958 Sep 4-12 FIFI 1,000 -- 16 80 30 19.7 60.9 1960 July 9-16 ABBY 992 8 15 85 15 14.6 65.7 1960 Aug 29-Sep 13 DONNA 942 13 13 140 22 16.9 60.0 1963 Sep 23-29 EDITH 978 -- 10 110 25 14.7 62.8 (after Dunn and Miller) 1963 Sep 26-Oct 13 FLORA 954 10 9 130 20 15.5 70.7 1964 Aug 20-Sep 5 CLEO 938 6 15 135 12 16.8 66.7 1966 Aug 21-Sep 5 FAITH 987 15 15 98 10 19.9 65.2 1966 Sep 21-oct 11 INEZ 932 4 14 140 8 16.9 66.6 1967 Sep 5-22 BEULAH 940 12 8 126 10 16.7 66.7 TABLE 3-2 SUMMARY OF TROPICAL CYCLONE TRACK COUN7 PPOCEDURFS PUERTO RICO, 1871-1972 Alonashore Storms Landfalling Storms Exitinq Storms Seaward Landt,.,ard South Coast "Track density" Omit Count tracks that miss Count all tracks method. island crossing that cross 660 30'W. 66- 301U. North Coast omit "Track density" method Count tracks that riss Count all tracks (Same count as "land- island crossing north- that cross north- falling" on south south line through south line through coast.) San Juan. San Juan. VJ' e s tCoast omit Take average of "land- Omit Omit ward alongshore" count for north and south coasts. CHAPTER 4 THE IIURRICA14-E SURGE MODEL 4.1 The model A hydrodynamic model developed by the National Weather Service [8, 9, 10) for the predicting of hurricane surges on the United States rainland coast when a storm is approaching has been employed in all previous coastal flooding frequency reports for the FIA b-1ky 1140AA. This model corputes a complete surface wind- field for the hurricane from the central pressure and the radius of maxirun winds. The asymmetry associated with forward motion of the storTr is also taken into account. The resultina wind stress on the water, and, by application of the appropriate physical equations of motion, the rise of water level on the Continental Shelf and the coast as the hurricane approaches or passes are then computed. The excitation of a wave by the hydrostatic rise of water level due to the diminution of the overlying atrospheric pressure is also taken into account. To accomplish this, calculations of the requisite physical variables are nade at a series of arid points at successive short tire intervals. The necessary approxinations are made to reduce the computation to a quantity that is economical on a large fast corputer. Computed coastal surges have been compared to observed surges in those hurricanes where sufficient data are available. These comparisons are described in the cited reports. Some of the peculiar conditions confronting the use of this model for the island of Puerto Rico and the approximations to take care of then for the nurposes of this study are enumerated below. 4.2 Application of model to Puerto Rico The hurricane surge program* was modified to handle the island of Puerto Rico. Two rain problens were encountered which required fundamental chanqes and development of the dynamic T-odel; these are: (1) hurricanes alona Puerto Rico coasts are smaller than alona the Gulf and r'ast Coasts of the Unitee States (2) the slope of the continental shelf at Puerto Rico is steep. Known as the "SPLASF" program from the acronym of the publication describing it (9]. The grid spacing in SPLASH is too coarse to see the driving forces of small sized storms; hence, the program was modified to shorten grid length from 4 miles to 2.5 miles; this reduction of grid size required a corresponding reduction of the time step for finite-difference computations and the program is more expensive to run. The problem of steep shelf slope is more difficult to solve. The approach was empirical. Experiments were performed to arrive at an optimum slope (as close as possible to the real slope) before the computations became unstable. These experiments were time consuming and were a significant portion of the development work required for the Puerto Rico project. Results appear to show that the coastal surge is insensitive to further increases in shelf steepness beyond a certain value. The modified SPLASH program was run with one-dimensional depth basins for the southern part of Puerto Rico; each basin's depth profile was localized for particular point along the coast. A modified shoaling curve for the southern part of the island was derived. The shoaling curve, with the hurricane climatology for Puerto Rico, described in Chapter 3, was used to estimate surge levels as described in Chapter 5. On the north coast, where the shelf slope is relatively uniform (Figure 1-1), the slope at San Juan was developed as an example for that area. - 16 - C112ASTER 5 STORM TIDE FRFOUFNCIES BY JOINT PPORABILITY 5.1 Example--San Juan The first step in the joint probability method is to divide the hurricane parameter probability distributions into class intervals and read out the rid-point value for each class interval. This is done in Table 5-1. Under forward speed, f, for example, in Table 5-1A, 8 kt. is the rid-point of the lowest 10% of storms from Figure 3-4, 10.9 kt. is the mid-point of the next 20%, etc. Central pressure depression, D, (1011 mb rinus central pressure) and direction of motion, @'L are similarly abstracted from Figures 3-1 and 3-9. The radius of maximum winds in Table 5-1A needs additional explanation. Following the precedent of indications of hurricane behavior in the vicinity of Florida and the 6ulf of Mexico, some tendency is assumed for smaller radius of maximum winds to be associated with the deepest hurricanes. Thus, in Table 5-1A, 70% of the deepest hurricanes are assumed to have an R of 10 n. mi. and 30% 16.5 n. ri. with a trend toward 50% each for the less intense hurricanes as indicated in the table. Thus, Table 5-1A defines 288 different hurricanes (8 x 2 x 6 x 3 = 288) that in the aggregate represent the climatological possi- bilities in the vicinity of San Juan. The probability (fraction of all hurricanes) of each of these is obtained by multiplying the respective parameter probabilities in the table. The sun of the probabilities of the 288 hurricanes, of course, equals 1.0. It has already been determined that the freauency of all exitinq storms is .0011 per nautical mile of coast per year. As the second step, calculations are made with the rodified SPLASH program described in Chapter 4 of the surae profile that would he produced in the vicinity of San Juan by each of the 288 exitina hurricanes, for the shelf slope specified for that region. (Many of the surge profiles are obtained by adjustment of other profiles rather than b-,, conplete surge computer computations.) Then lov. astronorical tide, high tide, and two interrediate tide levels are added to each of the 288 surge profiles, yielding 4 x 288 = 1152 storr tide profiles. (The mean tide ranqe at San Juan is 1.1 feet.) Next, each storm is allowed to exit fror the coast not only at the point most critical for San Juan but at points to the east and west, and the storr, tide profiles shifted along the coast accordingly. 17 As the third step, storrtides were sirilarly corputed for the alongshore storrs fron the data in Table 5-1B. Close in alonashore storr.s account for more than two-thirds of tl,-Ie tides eaual to or exceeding the 100-year value. rinally, surrninq all the possibilities--includino comgination of surges with hich, low, and intermediate astronorical tide and coastal placerent--yields the total tide frequency araph of Figure 5-1. In working out the joint probability of each hypoth- esized storv event, D, f, QL astronomical tide, and coastal place- ment are considered statistically independent, while R is dependent on D as shown in the table. These frequency values are still-water levels on the open coast that would be reasured in a tide quage house or other enclosure, excluding wave action. The wave question is discussed in the next chapter. 5.2 West Coast The same procedure was used at Mayaguez on the west coast using exiting storms only. The results are compared with Can Juan in Figure 5-2. At 14ayaauez, being in the lee of the island, the hurricane wind intensity threat is less than at San Juan, but the slope of the sea bottorn is not as steep and, therefore, is rore critical for production of surges. These factors combine to give the same 100-year return period tide level at Mayaguez as at San Juan, a higher 500-year tide level, and a lower 10-year tide level. 5.3 South Coast The sarie procedure was followed at Cabo Pobo, Mar Negro, and Playa De Humacao on the south coast for landfallina and alongshore hurricanes with the parareters listed in TaLle 5-2. The results are shown in Figure 5-3 for the 10-, 25-, 100- and 500-Year return periods. Interpolations were then rade hetween these three points following the shoaling factor curve that had been worked out (Chapter 4) from comparative surge calculations with standard hurricanes every few niles along the coast. 5.4 Extrapolation to Uncharted Areas The oLjective of this study is to establish the 100-year flood line. A complete tide frequency distribution has been calculated as described for the southern coast of Puerto Pico and selected points on the west and north coasts. The 100-year tide level was inter.T)olated between these points frorn inspection of @-,,ater depth r..a,,-)s. This interpolation is shown by the dashed lines on Figures 5-2 and 5-3. In a definitive "type 15" stud-v additional calculations should be made. 18 5.5 Pressure-wind relationship Part of the process of computing a hurricane surqe frorr hurricane parameters 1,y the SPLASH program [91 is to calculate the wind field from the three parameters central oressure depression, radius of raxiruir. -winds, and one-half the storr. motion vector. The theory and empirical relations for doincy t1lis are described in [81 and [10). The stress of wind on water at each grid point at each tire step is then obtained by multiplying the wind vector by a stress coefficient. It turns out that the maximum hurricane wind speeds calculated by this method are consistently smaller than the raxinum. wind speed estirates from aircraft for the hurricanes in Table 3-1. The latter are presumed to be estivates of the surface wind derived fror. the aPpearance of the sea as seen from the aircraft and from the measured wind at flight level. The maximum wind speeds computed by the SPLASH proqram, usina an averaae storm forward speed of 14 knots, and the aircraft-reported maximum winds are both plotted vs. central pressure for this set of hurricanes in riqure 5-4. Curves are fitted by eye to the two sets of points. For this study we have adhered to the results given by the current version of the SPLASH computer program. The disparity between the two wind-pressure relations in Figure 5-4 is large enough that in a "type 15" study for Puerto Rico the pressure-wind relationship for that region should be reexamined. such a reexamination would involve the total syster., including the atmospheric frictional coefficients which are used in computinq the surface wind field dynarically and the wind-on-water stress coefficient. We note one nrecedent for a disparity between aircraft hurricane winds and Winds from other sources in the Carribean area beinq resolved in favor of the other sources. Reference [111 contains an analysis of the surface wind field in hurricane HAZEL of 1954 in the vicinity of Great Inagua Island, p. 93 and Ficure 12-5 of the report. Ship reports and indirect calculations were criven more weight than aircraft reconnaissance reports. 19 TAFLE 5-1 Tropical Storm Paraneters-San Juan, Puerto Pico A. Exitina Storns Fe = -0011 Prd D Pi R=10 R=16.5 f P f OL PQ 85.0 0.01 0.7 0.3 79.2 .03 .7 .3 8.0 0.1 71.1 .06 .6 .4 10.9 .2 35 0.60 60.3 .10 .6 .4 13.9 .2 60 .36 43.2 .20 .5 .5 15.3 .2 90 .04 25.2 .20 .5 .5 16.4 .2 14.4 .20 .5 .5 20.6 .1 9.0 .20 .5 .5 B. Alongshore Storms Fb Fb R rr L (at sea) (inland) 2.0 0.0098 0.0118 6.5 .0118 .0118 11.0 .0118 .0106 10.0 0.5 15.0 .0118. .0108 16.5 .5 21.7 .0275 .0186 30.4 .0304 .0196 e Svrbols are identified on paq 21. - 20 TPX- Lr 5 - 2 Tropical Storr, Pararreters--South Coast, Puerto Rico Landfalling Storns Fn 0011 Prd D Pi f P f QL PQ R=10 R=16.5 94.0 0.01 0.7 0.3 88.0 .03 .7 .3 8.0 0.1 79.0 .06 .6 ..4 10.9 .2 35 0.60 67.0 .10 .6 .4 13.9 .2 60 .36 48.2 .20 .5 .5 15.3 .2 90 .04 26.0 .20 .5 .5 16.4 .2 15.6 .20 .5 .5 20.6 .1 10.0 .20 .5 .5 B. Alongshore Storms Fh rb R P L (at sea) (inland) 4.3 0.0186 0.0079 13.0 .0205 .0088 21.7 .0235 .0127 10.0 0.5 30.4 .0254 .0137 16.5 .5 39.1 .0274 .0147 47.8 .0314 .0157 Symbols are identified. on raae 21. 21 Lecrend for Tables 5-1 and 5-2 D = Central pressure deficit (mb). Pi = Proportion of total storms with indicated D value. f = Forward speed of storm (hnots). ,Pf = Proportion of total storms with indicated f value. R = Distance from center of storr. to principal belt of maximum winds (nantical miles) . Pr = Proportion of storms with indicated R value. Prd Proportion of storms in D class with indicated value. OL Direction of entry, measured clockwise from- the coast (decrrees) . PQ Proportion of total storms with indicated OL value. L = Effective distance perpendicularly landward or seaward from coast to storm. track (-nautical miles) . Fb = Average number of storms per year that pass at distance L. Fnr Fe Frequency of storm tracks crossing coast, landfallina and exiting, respectively (storr. tracks per nautical rile of coast per year). Notes: (1) Alongshore storns have the same values of D, Pi, f. and Ff as those for landfallinci. 22 CHAPTER 6 W,AVE EFFLCTS 6.1 Introduction Hurricanes as well as winter-type extra-tropical cyclonic wind storms produce high waves. The pounding of breaking waves is responsible for much of the beach erosion and structure damage on the irunediate beach front in storms. Near the coast wave arpli- tudes (trough to crest) are restricted by physical constraints to about 0.7 of the water depth. The Corps of Engineers, Depart- rent of the Army, is making a study of wave height criteria for coastal flood insurance analyses for the FIA. In tidal flood mappinc to date, pendina results of this study, a "velocity zone" has been delineated by NOAA and other agencies along the beach front where serious wave darnage can occur, with the inland limits of this zone estirated subjectively by reference to terrain. 6.2 VIave prohlem in Puerto Rico For several reasons, the wave question is particularly critical in assessina the coastal flood hazard in Puerto Rico. First, much of the coastline is relatively steep and the zone subject to inundation is narrow. In some areas the entire flood zone can be reached by waves. Second, the slope of the sea bottor fror. the shore outward is relatively steep. The effect of this is to limit surge levels, which are highest in shallow water, but to expose the beach to larae waves. As already mentioned, the liniting factor on wave height, once eguilibrium with irind force is reached, is water depth. Third, the long fetch of open water in almost every direction from the island allows storns to impose a full measure of wave-aenerating energy on the sea. 6.3 Wave swash studt A y for extratropical storms Fields and Jordan of the U.S. Geological Survey have surve.ved and analyzed wave darqac'Te on the north coast of Puerto Pico [121 in a cooperative study with the Department of Public Vorhs, Cornonwealth of Puerto Rico. Their data deal with extra-tropical (winter) storr events durinc the 19601s. in each instance storrs far to tne north (north of rermuda) produced swells resultinc in severe 1-.7aves on the north coast. Their report (121 shows raps of "wave swash" elevations (heiaht of water on the coast in waves) in the December 4, 1967 storr, which destroyed nore than 300 beach-front homes between 23 - San Juan and Arecibo. Figure 6-1 is a reprint of their tentative wave swash stage-frequency relation for Arecibo and Figure 6-2 a reprint of their generalized diagram. The report gives specific data on the decrease of wave height inland and up river courses, which is rapid. 6.4 Estimated hurricane effects Strong hurricanes passing north of Puerto Rico could be expected to produce waves comparable to those described in [12]. Hurri- canes have stronger winds but shorter fetch and usually persist less time in a given location than the worst wave-producing extra-tropical storms. The south coast of Puerto Rico is not exposed to waves from extra-tropical cyclones to any important-degree. Waves can be expected in hurricanes. On a frequency basis waves of a given amplitude will occur less frequently on the south coast than on the north coast but may occur in combination with higher surges. - 24 - CHAPTER 7 DEMARCATION OF FLOOD PHONE LIMITS 7.1 Objective The objective of this study is to indicate the approximate limit of special flood hazards along the coast of Puerto Pico on Maps for the Federal Insurance Adminstration. This limit has been delineated directly on existing USGS topographic quadrangles without field surveys. (field surveys would he necessary for a "type 15" study.) Principal guidelines for this, besides the topography, are the derived 100-year hurricane tide levels (Fiqures 5-2 and 5-3) and the wave swash analysis in [12]. 7.2 Precipitous coastal areas In the application of these factors, terrain plays an important role. Much of the coast of Puerto Rico is precipitous. In such areas to delineate tile inland.limit of flooding, an estimate for wave swash was added directly to the 100-year hurricane tide level. Reference (12] served as a guide for the swash estimates, using the higher values cited. This presupposes that that waves associated with hurricanes on both the north and south coasts could be cimparable to the more severe wave swash events described in (12]. The latter are fromr extra-tropical storms. 7.3 Flat coastal areas For the flatter flood plain areas it was assured that waves would be damped out a short distance inland. The figuares and maps in [12) served as a guide for estimating the inland penetration of the wave swash. Beyond this wave penetration zone, the hurricane 100-year tide level controlled the inland limit of special flood hazard. In coastal reaches intermediate between precipitous and nearly flat, interemediate criteria were applied: the hurricane tide level plus a modest increment for wave swash. 25 PJ@. FE PENCE S 1. Col6n, Jos4 A., "Meteorological Conditions over Puerto Pico durincy Hurricane Betsy, 1956," Monthly Weather Review, Vol. 87, No. 2f February 1959, pp. 69-80. 2. Cry, George VI., "Tropical Cyclones of the North Atlantic Ocean," Technical Paper No. 55, U.S. Weather Bureau, Washington, D.C., 1965, 148 pp. 3. U.S. Fleet Weather Facility, "Annual Tropical Storm Report," Jacksonville and Miami Florida, 1945-1972. 4. Dunn, Gordon E., and Miller, banner I., Atlantic Hurricanes, Louisiana State University Press, Baton Rouge, La., 1964, 377 pp. 5. Myers, Vance A., "Joint Probability of Tide Frequency Analysis Applied to Atlantic City and Long Beach Island, N.J.," ESSA Technical Marorandum WBTM HYDRO 11, Environmental.science Services Administra- tion, Silver Spring, Md., April 1970, 109 pp. 6. Ho, Francis P., Goodyear, Hugo V., and SchvA?erdt, Richard V.,., "Some Aspects of Clirratological Characteristics of Hurricanes and Tropical Storms, Gulf and East Coasts of the United States," (paper presented at the Eighth Technical Conference on hurricanes and Tropical Meteorology of the American Meteorological Society, Key Discayne, Florida, May 1973, to be published). 7. Shea, Dennis J., "The Structure and Dynamics of the Hurricane's Inner Core Region," Atmospheric Science Paper No. 182, Department of Atrnos- pheric Science, Colorado State University, Fort Collins, Colorado, April 1972, 134 pp. 8. Jelesnianski, Chester P., "Numerical Computations of Storm Surges With Bottorn Stress," Monthly weather Review, Vol. 95, No. 11, November 1967, pp. 740-756. 9. Jelesnianski, Chester P., "SPLASH (Special Program To List Arplitudes of Suraes From Hurricanes) I. Landfall Storms," NOAA Technical Memorandum NWS TDL-4G, National Oceanic and Atmospheric Administra- tion, Silver Spring, Md., April 1972, 52 pp. 10. Jelesnianski, Chester, and Taylor, A.D., "A Preliminary View of Storm Surges Before and After Storm Modifications," NOAA Technical Menorandurn ERL WtTO-3, National Oceanic and Atmospheric Adrinistra- tion, Boulder, Colorado, May 1973, 33 pp. 11. Graham, Howard E., and Hudson, Georgina N., "Surface Winds Near the Center of Hurricanes (and Other Cyclones)," National Hurricane Research Project, Report No. 39, U.S. Weather Bureau, Washington, D.C., September 1960, 200 pp. 12. Fields, Fred K., and Jordon, Donald G., "Storr-1,11ave Swash Along the North Coast of Puerto Rico," Hydrologic Investigations Atlas HA-430, U.S. Geological Survey, Washington, D.C., 1972. - 26 Z-o Tb'o* 0 6*3 0' Woo. -a- SOFATHO ISABELA AGUADILLA ARECISO SAN JUAN CEIBA UAYAGUEZ 'hAYA VE HUMACAO PUNTA GUAY -fll@@-@.:.: 67*00' PoNCE PARGUERA-i-IeOO' 0 -1600 GUAVA A -L'e0e 11 @-. 0 - fl) FATHOM fl) FATHOM 40 FATHOM ATHO., i.-F L ISAB@Ll 6 00 S LIA@GUERA"-IeOCr IUAVA-A "i 0 Figure -Locator -nap. AUU3-25 AUG. 9-16 AKIB-25 U.S. DEPARTMENT OF COMMERCE 1893 1956 1091 National Oceanic mad Atmosphorto Administrathan NATIONAL WEATHER SERVICE 12 JULY 22- AUG. 2 SEPT. 6-20 JULY 5-13 1901 A -24 8 9 24 AUG.30-SEPT11 1 1896 to AUG. 21-25 isle SEPT. x2-t le76 SEPT B-16 to 00 1931-0-1 .-SANoi MAS 1EQUES SEPT 25- OCT. 3 13 193Z CABO ROJ GUAYAMA ST. CROIX 13 LEGEND @.@ HURRICANE TRACK -4- TROPICIAL STORM 0 STORM POSITION AT 7 Am ExT. 9 STORM POSITION AT 7, PM EST SMOOTHED COASTLINE AUG 3-11 1952 MAJOR STORM (DATE UNOERLINffD) 31 023 Figure 2-1.-- Landfalling hurricanes for the period 1871-1972. 100 90 80- 70- 60- 0 W LL 50- 0 F- z 4 0 LLJ C) x Uj a_ 30 20 10 0- 920 940 960 980 1000 CENTRAL PRESSURE (MB) Figure 3-l.--Probability distribution of tropical cyclone central pressure (based on data listed in Table 3-1) in the Puerto Rico area, 1945-1972. 29 100 80- 1n60- 0 U- SOUTHERN TIP PUERTO RICO 040- OF FLORIDA z w X w CL20 - 00 920 940 960 980 1000 CENTRAL PRESSURE( ME3 Figure 3-2.--A comparison of probability distributions of tropical cyclone central pressure for Puerto Rico area (180N) and the southern tip of Florida (25*N). 30 100 80- 0 60 0 z w C-) Of W40- CL 20- 0. 5 10 15 20 RADIUS OF MAX.W INDS( N,Ml,) Figure 3-3.,17ame as Figure 3-1 but for radius of raximin win0s. 31 - 100 90- Q) 80 70- 060- U- 0 @-50 z 0-40- ,30 20- 10- 0--8 10 12 14 16 18 20 FORWARD SPEED ( KT, Fiqure 3-4.--Same as Piqure 3-1 but for speed of fon-7ard motion. 32 30 Ui z UJ20 D C-) 15-- 0 D 10-- W < 5-- 0 610- 40 20 G 40 60 DISTANCE INLAND N-MI.) DISTANCE FROM COAST( N. MI. 5 --10 15 T Fiaure 3-5- -Accumulative freauency of alonqshore tropical cyclones for San Juan, Puerto Pico, IR71-1972. 0 z PO-- X 0 15-- D C-) I O-- < tA 5- 60 40 2 20 40 60 DISTANCE INLAND(N@j DISTANCE FROM COAST( N. MI.) 5 Q __lo --15 1-ioure 3-6-Sarne as ricure 3-5 hut for southern coast, Puerto Pdco, 1871-1972. 7 0 650 600 0/ 200 44 55 52 PUERTO, RICO A Ln .43 50 54 0 1-5 Figure 3-7.--Freauenc v of tropical cyclone occurrences over 2.5 degree latitude and longitude octagons in the vicinity of Puerto Pico in 102 years. 1871-1972. 2.6 2.4- 2.2- 2.0- 1.0 1.8- .9 1.6 .8 Q94.4 .7 0 0 .6 x 1.0 .5 (n .4 cn .8 (D (k: .6 4 .2 .2 0- 0 210 195 180 165 150 135 120 105 90 75 60 45 D I R ECT I ON ( DEGR EES) Figure 3-8-HistoRram and Beta distribution fit for frequency of tropical .,00/. cvclone tract-, direction for octagon bounded by 17.5-20.0*N and 65.0-67.5*V. Dashed line denotes the accumulated probability distribution of direction of storm motion. - 36 - 100 80- LANDFALLING S T 0 R M S (SOUTH COAST),' I CO EXITING STORMS M (NORTH COAST) W60 0 z W40- W- w ALONGSHORE STORMS 20- 0 60 80 100 12-0 140 160 180 DIRECTION OF STORM MOTION ( DEGREES FROM NORTH) Figure 3-9.--Accuinulative probability distribution of direction of tropical cyclone motion 17.5-20*N 65-67.5*w, from Fiqure 3-8. Dashed lines denote class intervals adopted for tide frequency computation. - 37 15 14 (611 U_ 9 8 C:) w 6 5 < 4 0 3 SAN JUAN PUERTO RICO (NORTH COAST) 0L 10 25 50 100 500 RETURN PERIOD YEARS) TOTAL TIDE FREQUENCY CURVE Figure 5-1, -Tide frequency. San Juan, Puerto Rico. 38 - Figure 5-2-Intemolated total tide frequency for the 10--, 25-, 100-, and 500-year return period, south coast, 'Puerto Rico and interpolated 100-year value for west and east coast. 0 C-) < UJI OW 0 M z LLI N< az 0-1 UJ z Wo < 0 :3 0 W 0 C) cc cro LLJ < C-) 0 C-) 0 z Ljj >- UJ z < < < 0 < 0 <U) Z (L CL 14 SOUTH COAST OF PUERTO R I CO J 121.4 9.8 W10 9.0 8.7 7.4 -T- (D P% Ul 6 5. m - 5. 3 5.2 W 4.5 0 -------0 4- 2.5 2.2 2.8 < 0 i0YEAk F- (0 N 0 z L;j Uj C-) < 0 ;t -i < 0 Cr- D D (D LLI C-) 0 < m Q 0 C-) > < Ld z cr Uj Uj 0:& 14 ... < Cn NORTH COASTOF PUERTO RICO J1 2- @-l 0 - UJ 9.0 W U- 1 8- 76 kE 500 YEAR 6- 5.3 100 YEAR 5.3 w 4- 3.3 2.5 25 YEAR - 0 10 YEAR 0 0 2- 0 2.6 1,8 0- Figure 5-3-Same as Figure 5-2 for north coast. 140 120 - U< 100 - A E3 80- 920 940 960 980 1000 CENTRAL PRESSURE (MB) Fiqure 5-4-raxirum hurricane surface win(' vs. central pressure for storms in Table 3-1. Plotted noints: maxiT-ur ,,ind reported 1-r aircraft, Curve A: eye-fit to these points. Curve B: !lax i rurn ,,.,ind vs. central pre,;sure relationship from SPIAFE rodel [8) For 15 n. ri. and forward speed of 14 kt. 41 - 7.0 6.5 Ld 6.0 z u > 5.5 Lu 5.0 TL 0 Ld < 4.5 Ll 4.0 3.5 0.7 1 2 4 7 10 20 RECURRENCE INTERVAL, IN YEARS '@;tafje-frepiency relation of wave stoish. at Arecibo, longi- bide 66%.1'. Figure 6-l.--Stage-frequency relationship for wave swash at Arecibo. From [121. 7 42 9 8 Estima led 20-year recurrence interval > Ld -j -If 5 W W z 4 LLI 0 < 0 Uj 3 W 10-year recurrence interval \0 0 X 2 .c.,re2in1.rvaI W > 1-year recurrence interval 2 3 4 DISTANCE FROM SHORE TO 120-FOOT DEPTH CURVE, IN THOUSANDS OF METERS Relation ofstage ardfi-equency ofwave s-wash to distance between shoreline and, the 120-foot depth cilt-ve. Figure 6-2-Generalized stage frequencies for wave swash, north coast of Puerto Rico. From [12]. \ 5-year ecuren val 43 (Continued from inside front cover) NWS HYDRO 13 Time Distribution of Precipitation in 4- to 10-Day Storms--Ohio River Basin. John F. Miller and Ralph H. Frederick, May 1972. (COM-72-11139) NWS HYDO 14 National Weather Service River Forecast System Forecast Procedures. December, 1972. COM-73-10517) NWS HYDRO IS Time Distribution of Precipitation in 4- to 10-Day Storms--Arkansas-Canadian River Ba- sins. Ralph If. Frederick, June 1973. (COM-73-11169) NWS HYDRO 16 A Dynamic Model of Stage-Discharge Relations Affected by Changing Discharge. D. L. Fread, December 1973. (COM-74-10818) NWS HYDRO 17 National Weather Service River Forecast System--Snow Accumulation and Ablation Model. Eric A. Anderson, November 1973. (COM-74-10728) NWS HYDRO 18 Numerical Properties of Implicit Four-Point Finite Difference Equations of Unsteady Flow. D. L. Fread, March 1974. NWS HYDRO 19 Storm Tide Frequency Analysis for the Coast of Georgia. Francis P. Ho, September 1974. (COM-74-11746/AS) NWS HYDRO 20 Storm Tide Frequency for the Gulf Coast of Florida From Cape San Blas to St. Petersburg Beach. Francis P. Ho and Robert J. Tracey, April 1975. NWS HYDRO 21 Storm Tide Frequency Analysis for the Coast of North Carolina, South of Cape Lookout. Francis P. Ho and Robert J. Tracey, in press, 1975. NWS HYDRO 22 Annotated Bibliography of NOAA Publications of Hydrometeorological Interest. John F. Miller, in press, 1975. f I i @ @m1milloolill i 000028938