C' '6--//7>y i .<.< °\ c % c % Vso^ / Environmental Impact Statement for the Hurricane Amelioration Research Project Rockville, Maryland February 1 978 U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration \ .«P«!JP»»^ ^^: ..TiSTcCC a 5 Environmental Impact Statement (EIS) for the Hurricane Amelioration Research Project Rockville, Maryland February 1 978 U.S. DEPARTMENT OF COMMERCE Juanita M. Kreps, Secretary National Oceanic and Atmospheric Administration Richard A. Frank, Administrator FOREWORD The Hurricane Amelioration Research Project is a proposed experiment to be directed by the National Oceanic and Atmospheric Administration (NOAA) collaboratively with the Mexican Secretariat of Agriculture and Hydrology. The primary goal of the experiment is to test the hypothesis that maximum surface winds in hurricanes can be reduced 10 to 15 percent or more by seeding the proper clouds in specified portions of the storms with freezing nuclei (silver iodide) . SRI International (formerly Stanford Research Institute) prepared the bulk of this report during September 1977 under contract to NOAA. The report presents the results of an analysis of the environmental effects of performing the experiment in the eastern North Pacific off the west coast of Mexico. The analysis covers the environmental effects of dispensing silver iodide and of any resulting changes in the hurricanes; it does not cover environmental effects of the deployment and operation of project aircraft. XI ACKNOWLEDGMENTS NOAA gratefully acknowledges the contributions of the following: ° SRI International staff members, T. P. Miller, E. C. Barrera, C. M. Bhumralkar, S. J. Everett, R. D. Daughters, R. E. Freeman, B. R. Holt, C. A. Kroll, E. M. Liston, R. W. Mack, J. W. Maresca, W. Viezee and R. A. Zink, for preparing Sections I through VII. S. Brand of the Environmental Prediction Research Facility, U. S. Naval Post Graduate School, Monterey, California, for making available various studies dealing with hurricanes and typhoons, o A. F. Gustafson and W. J. Denney of the Weather Service Forecast Office, Redwood City, California, for providing valuable information on eastern North Pacific hurricanes. iii 11 CONTENTS FOREWORD LIST OF ILLUSTRATIONS viii LIST OF TABLES x i I INTRODUCTION I- 1 A. Background I- 1 B. Hurricanes 1-2 C. Study Approach 1-3 D. Summary of Findings 1-3 E. Contents of Report 1-5 II DESCRIPTION OF THE PROJECT II- 1 A. Overview II- 1 B. Seeding Hypothesis II- 1 C. Aircraft Operations and Seeding Strategies .... II- 6 D. Project Activities II- 9 1. Deployment and Support II- 9 2. Seeding Eligibility Rules 11-10 III PHYSICAL AND BIOLOGICAL SETTING Ill- 1 A. Location and General Description Ill- 1 B. Physical Environment Ill- 1 1. Geology Ill- 1 2. Soils Ill- 2 3. Hydrology Ill- 3 4. Climate of the Experimental Areas and Their Vicinity Ill- 4 5. Oceanography Ill- 9 C. Biological Environment Ill- 9 1. Vegetation Ill- 9 2. Fauna 111-13 IV IV HURRICANES IV- 1 A. General IV- 1 B. Formation of Tropical Cyclones IV- 2 C. Climatology IV- 4 1. Occurrence IV- 4 2. Movement IV- 7 3. Recurvature IV- 7 4. Duration IV-16 5. Size IV-17 •6. Intensity IV-18 D. Typical Characteristics of a Mature Hurricane . . . IV-18 1. Wind Field IV-21 2. Circulation IV-21 3. Pressure Field IV-23 4. Temperature Field IV-23 5. Cloud Distribution IV-23 6. Rainfall Distribution IV-25 7. Severe Weather Phenomena IV-25 E. Natural Variability of Hurricanes IV-26 F. Prediction of Hurricanes and Their Movement .... IV-26 G. Storm Surge and Waves IV- 31 1. Storm Surge IV-31 2. Waves IV- 34 V SOCIOECONOMIC SETTING V- 1 A. Geographic Overview . . . • V- 1 B. The Regions V- 1 C. Settlement Patterns V- 3 D. Settlement Pattern Along Coastal Lowlands V- 8 E. Population Characteristics V-ll F. Housing Characteristics V-13 G. Economic Activity V-18 H. Governmental, Legal, and Institutional Structures . V-23 1. Status of Entities that Lie Within the Experimental Area V-23 2. Relevant U.S. Domestic Legislation V-24 3. Relevant Legislation of Mexico V-25 4. Relevant Treaties V-25 VI EXPECTED RESULTS OF EXPERIMENT , A. General , B. Background 1. Previous Seeding Experiments .... , 2. Data Analysis Method 3. Numerical Hurricane Modeling C. Effects of Seeding on Hurricane Intensity (Maximum Windspeed) , 1. Experiments on Hurricane Esther, September 16 and 17, 1961 , 2. Experiments on Hurricane Beulah, August 23 and 24, 1963 3. Experiments on Hurricane Debbie, August 18 and 20, 1969 4. Experiments on Hurricane Ginger, September 26 and 28, 1971 5. Effect of Seeding on Windspeed Profiles 50 to 200 Nautical Miles from Hurricane Center . 6. Natural Versus Induced Variability of Intensity. D. Effects of Seeding on Hurricane Motion E. Effect of Seeding on Size of Storm F. Effects of Seeding on Precipitation Pattern and Amount G. Duration of Changes in Hurricane H. Effects of Seeding on Associated Severe Storms and Large Scale Environment 1. Severe Storms (Tornadoes, Thunderstorms) . . . . 2. Large Scale Environment I. Effects of Seeding on Surge and Waves 1. Surge 2. Waves VII PROBABLE ENVIRONMENTAL IMPACT OF EXPERIMENT A. General B. Impacts of Silver Iodide and Other Air- Launched Material 1. General 2. Natural Environment 3. Human Environment C. Impacts of Hurricane Changes 1. General Considerations 2. Likelihood of Impact Occurrence VI- 1 VI- 1 VI- 1 VI- 1 VI- 3 VI- 7 VI-10 VI-10 VI-10 VI-12 VI-12 VI-14 VI-14 VI-16 VI-24 VI-25 VI-30 VI-31 VI-31 VI-31 VI-32 VI-35 VI-38 VII- 1 VII- 1 VII- 1 VII- 1 VII- 1 VII- 3 VII- 3 VII- 3 VII- 7 VI 3. Natural Environment 4. Socioeconomic Environment • • VIII POTENTIAL ALTERNATIVES TO PROJECT A. Do Not Perform B. Postpone C. Alternative Experimental Areas D. Alternative Modification Techniques IX IRREVERSIBLE AND IRRETRIEVABLE COMMITMENTS OF RESOURCES APPENDICES Upper Limits for Concentrations of Silver Iodide .... Operational Application of the Land Fall Criterion . . . REFERENCES ■ * : VII- 8 VII-11 VIII-1 VIII-2 VIII-2 VIII-3 VIII-5 IX-1 A-l B-l R-l vii ILLUSTRATIONS II- 1 Eastern Pacific Experimental Area II- 2 II- 2 Schematic Cross Section of Hurricane II- 4 II- 3 Hypothetical Cross Section of Hurricane Before and After Seeding II- 5 II- 4 Examples of Aircraft Employment Timetable II- 8 III- 1 Typical August Weather Pattern in Eastern North Pacific Ill- 7 III- 2 Typical Hurricane Tracks during the "High Season" (16 July-15 September, 1965-74) Ill- 8 III- 3 Vegetation Patterns in Coastal Regions 111-10 IV- 1 Average Annual Number of Tropical Storms in Each Cyclone Development Area IV- 3 IV- 2 Monthly Frequency Distribution of Eastern Pacific Hurricanes (1966-1976) IV- 5 IV- 3 Annual Variation in Eastern Pacific Hurricane Occurrence IV- 6 IV- 4 Tracks of Eastern North Pacific Tropical Cyclones, August through October 1976 IV- 8 IV- 5 Frequency Distribution of Tropical Cyclone Track Direction in the Eastern Pacific (1956-1971) IV- 9 IV- 6 Hurricanes and Tropical Storms Moving in North- easterly Direction During 1965-74 IV-12 IV- 7 Hurricanes and Tropical Storms Moving in Northeasterly Direction During September (1965-74) IV-13 IV- 8 Hurricanes and Tropical Storms Moving in Northeasterly Direction During October (1965-74) . . . IV-14 IV- 9 Tropical Storms and Hurricanes Crossing the Western Coast of Mexico (1971-1967) IV-15 IV-10 Maximum Wind Speeds of Eastern Pacific Hurricanes (1972-1977) IV-19 IV-11 Percent of Eastern Pacific Hurricanes with Peak Winds Less than Indicated Speed (1972-1977) . . . IV-20 viii IV-12 Components of a Mature Typhoon IV-22 IV-13 Relation Between Minimum Storm Pressure and Maximum Wind IV-22 IV-14 Vertical Cross Section of Temperature Anomalies in Hurricane IV-24 IV-15 Schematic Cross Section of Winds, Clouds, and Pressure of Typical Hurricane IV-24 IV-16 Algorithm for the NHC72 Prediction Model IV-30 IV-17 Storm-Surge Profile at Coastline in Northern Hemisphere IV-32 IV-18 Lines of Equal Relative Wave Height for Slow Moving Typhoon or Hurricane (in Northern Hemisphere) . IV-36 V- 1 Study Regions V- 2 V- 2 Population Distribution V- 6 V- 3 Areas Below 200 Meters Elevation V- 9 VI- 1 Data Collected Before, During, and After Seeding to Determine the Validity of the Hypothesized Sequence of Events , VI- 4 VI- 2 Observed and Predicted Track of Hurricane Alma (1962) as Forecast by a Numerical Model VI- 8 VI- 3 Radial Wind Profile of Hurricane Daisy VI-11 VI- 4 Hurricane Debbie Windspeed Profiles Recorded on August 18, 1969 VI-13 VI- 5 Hurricane Debbie Windspeed Profiles Recorded on August 20, 1969 VI-13 VI- 6 Observed Winds (12,000 ft. Altitude) at Greater Distances from Center VI-15 VI- 7 Frequency Distribution of Changes in Maximum Winds of Unseeded Hurricanes VI-17 VI- 8 Storm Track and Forecast Vector Movement for Hurricane Debbie, 1969 VI-20 VI- 9 Estimates of Modified Track for 10 Hours of "Normal" Heating Rate in Symmetric Pattern VI-22 VI-10 Estimates of Modified Tracks for Extreme Heating Rate in Asymmetric Patterns VI-23 VI-11 Water Budget of Composite Western North Pacific Typhoon VI-27 VI-12 Cumulative Rainfall Along East-West Line During Passage of Northward Moving Storm VI-29 VI-13 Geographical Distribution of Hurricane Tornadoes (1948-1972) VI-33 ix VT-14 Geographical Distribution of Typhoon Tornadoes (1950-1971) VI-34 VI-15 Profiles of Observed Windspeed and Computed Open Coast Surge for Hurricane Debbie (1969) VI-37 B-l Samples of 431. Forecasts in the Atlantic Ocean (1965-1974) B-3 B-2 Hypothetical Land Area in the Path of a Hurricane . . B-4 TABLES III- 1 Climatic Statistics for Selected Locations Ill- 5 IV- 1 Recurving Tropical Storms and Hurricanes in the Eastern Pacific IV-11 IV- 2 Average Time Tropical Cyclones Spent in Each of Three Stages ..... IV-16 IV- 3 Circular Cloud Diameters (Degrees Latitude) of Tropical Cyclones in the North Pacific IV-17 IV- 4 Tropical Cyclone Prediction Models Available on a Routine Basis for the Atlantic Region IV-29 V- 1 Population of Western Mexico by States and Regions — Urban-Rural Breakdown V- 5 V- 2 The Population of Major Cities on the West Coast ... V- 7 V- 3 Population Distribution of the Coastal Lowlands in the Western Mexican States and Regions V-10 V- 4 Mexican Western States Working Population Breakdown by Position Held V-12 V- 5 Western Mexican State's Working Population Breakdown by Activity V-14 V- 6 Western Mexican States — Total Number of Housing Units and Materials Used in the Construction of Roofs and Floors V-16 V- 7 Western Mexican States — Total Number of Housing Units and Materials in Construction of Walls V-17 V- 8 Ranking of the Top Five Western Mexican States in Production of Selected Crops, Forestry, and Fishing V-19 VI- 1 Results of Experiments in Seeding Hurricane Clouds Near the Eyewall VI- 2 VI- 2 Methods of Data Analysis VI- 6 VI- 3 Forecast Errors for Hurricane Debbie (August 1969) . . VI-19 VI- 4 Sample Calculations of Hurricane Significant Wave Heights in the Outer Region of the Storm .... VI-40 VII- 1 Damage Criteria for Forests VII- 5 VII- 2 Ocean Bound Freight Movement from Mexico VII- 16 VIII-1 Annual Average Number of Hurricanes/Typhoons VIII- 3 xi I INTRODUCTION A. Background Project STORMFURY is a scientific experiment designed to explore the structure and dynamics of Western Atlantic tropical cyclones and the potential for their beneficial modification. The experiment began as a joint operation of the United States Weather Bureau and the Depart- ment of Defense (Navy) . Initial experiments using silver iodide to seed a hurricane in 1961 were sufficiently encouraging that Project STORMFURY was formally organized by the Weather Bureau and the Navy. The project has continually evolved and is today an effort directed by the National Oceanic and Atmospheric Administration (NOAA) with cooperation from the National Aeronautics and Space Administration (NASA) and the National Science Foundation (NSF) . Four Atlantic hurricanes have been seeded since 1961 with silver iodide crystals dispensed from aircraft with pyrotechnic devices developed by the U.S. Navy. The results were sufficiently promising in terms of reducing the peak hurricane winds that NOAA has made plans to expand the area of operations to include the eastern North Pacific so that the number of candidate hurricanes will be larger, thus providing more opportunities for seeding experiments. The experimental areas for Project STORMFURY and the Hurricane Amelioration Research Project, as the currently planned operations are called, include portions of the western North Atlantic and the eastern North Pacific, respectively. This report concerns the eastern North Pacific operations, which will be based at La Paz, Mazatlan, and Acapulco. The area of operations will be within 600-nautical mile circular sectors outward from these cities. Hurricanes will be selected for seeding on the basis of a set of eligibility rules designed to 1-1 minimize the probability of an altered hurricane affecting a populated land mass. The seeding of a huricane will be performed in the daytime over an 8-hour time span, during the experimental period of the project, 1 2* June through October. ' B. Hurricanes Tropical cyclones with windspeeds of 64 knots or more are generally called typhoons in the western North Pacific and hurricanes elsewhere. (In Japan, tropical cyclones are called typhoons if the winds are 34 knots or more.) The characteristics of hurricanes (and typhoons) are quite variable. The storms can be from 100 to several hundred miles across. The region of maximum wind, which in rare cases exceeds 200 knots, may be located 5 to 40 nautical miles from the storm center (point of lowest atmospheric pressure) . The center of the storm is called the "eye", a region of relative calm. Storms typically stay in the hurricane stage for several days, although this period can range from a few hours to a few weeks. Hurricanes generally travel at an average speed of 8 to 20 knots. They have been known to stay in the same general area for several hours and in other cases to travel several hundred miles in 12 hours. A phenomenon associated with hurricanes, the storm "surge" or rise in the mean level of the ocean, is due primarily to the reduction in atmospheric pressure and to the effects of wind and ocean bottom topography. In the open ocean the average water level may rise approximately 1 to 2 feet. This effect is separate from the wind driven wave action, which may result in waves over 50 feet high. As the hurricane approaches land the surge can increase considerably because of the dynamic interaction with the ocean bottom and the coastline. On a gradually sloping shore the average level of the water may rise over * References, when not specifically mentioned in the text by author and year, are numbered consecutively for each section in the report and are listed on the final pages of the report. 1-2 15 feet, with high waves on top of that. For small islands with relatively steep shores, such as those in the eastern Pacific experi- mental area, the surge height is about the same as in the open ocean. C. Study Approach The environmental analysis was conducted by an interdisciplinary team. Initial work consisted of the preparation of a description of the Hurricane Amelioration Research Project, with emphasis on the possible outcomes of the experiment, and an outline of the major characteristics of the natural and socioeconomic environment in and around the experimental area. A preliminary identification of possible impacts was then made by project participants to ensure that no significant impacts were overlooked and to assist in the selection of impacts for more detailed analysis. Such analyses were then carried out to the level of detail required by their importance and justified by the quality and coverage of available data. D. Summary of Findings The probable environmental impacts associated with the Hurricane Amelioration Research Project are of two types: direct and indirect. The direct impacts are those arising from the deployment of the aircraft, equipment, and personnel to and from Mexico; support of personnel and aircraft in Mexico; operation of aircraft and equipment during the project; and dispensing of the silver iodide. The indirect impacts include any effects arising from any changes brought about in the hurricane by the seeding operations. Following are brief summaries of the major findings of the environ- mental analysis (which did not cover deployment, support, and aircraft operations) . ° The dispensing of silver iodide will have no discernible effect on air quality, water quality, soil, or biological species because of the small quantity that will be dispensed, the large area over which it will be dispensed and then dispersed, and the low reactivity of silver iodide with the relevant biological species. 1-3 The state of knowledge of hurricane modification and the natural variability of these storms is such that there is a broad range of uncertainty as to the specific effects of seeding on tropical cyclone dynamics and on the associated surge. The preponderance of opinion, which is generally supported by theoretical and observational studies, including many computer simulations, and by the existing but limited experimental results, is that: - Any effects of seeding on the hurricane's structure, windspeed, and movement will probably disappear within 18 hours after the seeding stops. The hurricane will then most likely have virtually the same characteristics, be moving at very nearly the same speed, and be in very nearly the same place as if it had not been seeded. - During the 18 hours after seeding, the peak winds, the peak surge, and the peak waves will most likely decrease. At greater distances from the storm center the winds and surge may increase somewhat . - The speed of the storm along its track should be unaffected by the seeding. According to model calculations, the track itself may be altered somewhat during seeding and for a few hours thereafter. This deviation may amount to 10 to 20 nautical miles for short periods. This is well within the observational noise level for determining the true track of the storm. - There should be little or no change in rainfall associated with the hurricane as a result of seeding except for a slight change in the local rainfall pattern during the actual seeding operation, which will take place over the open ocean. The expected behavior of a seeded hurricane, taken together with the eligibility rules that determine which hurricanes may be selected for seeding, indicates that hurricane seeding will not pro- duce any significant physical impacts outside the experimental area. Within the experimental area there is a small probability that a seeded hurricane will affect a populated area while it is still in an altered state. If such an even does occur and the mainland or an island is struck with peak winds, the chances are that the winds and surge, and hence the damage, will be significantly less than they would have been without seeding. There is also a small probability that an altered hurricane will pass near the mainland or an island at distances at which the peripheral winds and surge will have increased somewhat. These effects may cause a slight increase in damage. Since the eligibility rules can be changed as additional information is developed — before the experiment, based on modeling results, and during the experiment, based on actual measured results — the selection of hurricanes can probably be adjusted to reduce these probabilities without appreciably affecting the number of hurricanes eligible for seeding. 1-4 It should be stressed that the Hurricane Amelioration Research Project is a scientific experiment with specific restrictions on conditions under which it is to be conducted, as defined by the seeding eligibility rules. This assessment of the environmental impacts associated with the Project was made with those restrictions in mind. If the experiment is successful and the results indicate that hurricane damage can be decreased significantly by cloud seeding, an operational hurricane amelioration program to take advantage of its potential benefits to society will be recommended. Initiation of an operational program would require a revised environmental impact statement that reflects the less restrictive conditions under which it would operate. E. Contents of Report The following section of the report presents a brief description of the Hurricane Amelioration Research Project; its underlying hypothesis; and an outline of project activities including (1) seeding operations and strategies, and (2) seeding eligibility rules. Section III is a description of the physical and biological setting of the experiment. Section IV summarizes the properties of hurricanes, including their formation, climatology, and major characteristics. Also included is a description of hurricane variability, hurricane prediction, and storm surge. Section V describes the socioeconomic characteristics of the relevant regions of Mexico, including the governmental, legal, and institutional structures. Also presented is a brief summary of previous experience with hurricanes on the western coast of Mexico. Section VI describes the probable results of the experiment in terms of the dispersal of silver iodide and the resulting changes in the hurricane The island of Clipperton in the experimental area belongs to France, 1-5 and surge. Included are changes in intensity (maximum windspeed) , motion, size, precipitation, associated severe storms, and surge. The probable duration of such changes is also discussed. Section VII describes the probable environmental impacts of the major aspects of the experiment: dispersal of silver iodide and changes in hurricane and surge. Potential impacts are considered for both the natural environment (water quality, air quality, flora, fauna, soils, and topography) and the human and man-made environment (health, socioeconomic activities, and land use and management). Section VIII discusses potential alternatives to conducting the Hurricane Amelioration Research Project, and Section IX briefly identifies resources irreversibly and irretrievablly committed to the Project. The Appendices contain detailed information on the concentrations and impacts of silver iodide used to seed the storms and a discussion of the land-fall criteria. 1-6 II DESCRIPTION OF PROJECT A. Overview The primary goal of the Hurricane Amelioration Research Project is to test the hypothesis that the maximum winds in hurricanes can be reduced by at least 10 to 15 percent by seeding the proper clouds in specified portions of the storms with freezing nuclei (silver iodide) . The portion of the project being described here will be based at La Paz, Mazatlan, and Acapulco. Figure II-l shows the general location of the experimental area. Hurricanes will be selected for seeding on the basis of the eligibility rules described at the end of this section. In each experiment, there will be about 16 flights over a period of 32 hours involving 5 different aircraft. Approximately 875 pounds of silver iodide will be dispensed in 5 separate seeding flights over an 8-hour period. Aircraft monitoring of storm characteristics, beginning about A hours prior to the first seeding, will be nearly continuous throughout the entire 32-hour experimental period. The experimental results will also be monitored by satellite. B. Seeding Hypothesis The proposed experiments are designed to produce a redistribution of the wind energy concentrated near the storm's center so that the maximum winds will be decreased. At low altitudes warm, moist air flows in a spiral toward the storm's center. As this air flows into the storm it acquires additional heat and moisture from the ocean. Much of this air flows upward near the center of the storm into the eyewall, a band of clouds encircling the relatively calm eye. Most of the storm's kinetic energy arises from the increased buoyancy of the air * Descriptions of project activities are derived from. Reference 1. The descriptions and the analyses here and elsewhere in the report are limited to operations based on the west coast of Mexico. II-l -I < a. < < 1- Z LU S a: LU a. X % CN LU o Q CJ — LL o CJ CO < "■ Q. VI ■o c z a cc wl LU s 1- < o a LU >o £ o _C «- •o — c <0 LU "2 (0 (X 3 c a Jfc 2 LL II-2 resulting from the heat released as the water vapor in the rising air changes to liquid or to ice. The air, already moving in a slow circular pattern because of the rotation of the earth, speeds up as it spirals inward, maintaining much of its angular momentum. As it approaches the storm's core, the air reaches destructive speeds before moving upward and then away from the storm's center at high altitudes (see Figures II-2 and II-3) . Seeding provides a means for changing the location of the areas where most of the vertical motion occurs. If silver iodide particles are introduced into a cloud containing supercooled" water droplets, the droplets will freeze and release heat (latent heat of fusion) into the surrounding air. This additional heat in the ascending air of the cloud further increases its buoyancy and causes more vigorous vertical motion. As the air rises, it expands and cools. Additional water vapor then condenses to liquid or ice, releasing much greater quantities of latent heat . Since hurricanes have many clouds which contain large' quantities of supercooled water, the potential exists for the addition of significant amounts of heat by seeding certain regions of the hurricane. Experiments and theoretical studies indicate that, in many areas outside of the eyewall, there are clouds that can be made to grow, sometimes explosively, by seeding. Causing such growth would greatly increase the upward motion of the air in these regions, with the result that much of the vertical air motion would occur at larger radii. This means that the inward spiraling air would not reach such high speeds before it begins its upward flow. Model experiments suggest that such a sequence of events would weaken the old eyewall and create a new eyewall at a larger radius with reduced maximum winds. The new eyewall may be stable for several hours. * Liquid water below 32° (0°C) . II-3 EYgVKAWi SUPERCOOLED WATER CLCUOS N OUTFLOW LAYER "^ Source: Aoaotea from Reference 2 FIGURE 11-2. SCHEMATIC CROSS SECTION OF HURRICANE II-4 BEFORE SEEDING AFTER SEEDING Source: Adapted from Reference 3 FIGURE 11-3. HYPOTHETICAL CROSS SECTION OF HURRICANE BEFORE ANO AFTER SEEDING (Not to Scale) II-5 The following sequence of events is hypothesized to occur when the seeding experiment is carried out: 1. Clouds are seeded outward (away from the storm center) from the external edge of the hurricane eyewall. 2. Supercooled water in the seeded cloud freezes, latent heat of fusion is released, the buoyancy of the upper portion of the cloud increases, and increased vertical motion results in increased condensation rates and cloud growth. 3. The seeded cloud reaches the outflow level, providing a conduit for the major vertical mass transport at a larger radius . 4. The old eyewall circulation weakens as the vertical mass transport is concentrated in the seeded clouds, and the subsidence (downward air motion) in the eye decreases. 5. The maximum wind speeds are reduced due to the above effects and to the decreased temperature gradients. 6. The pressure field adjusts to the wind and temperature fields. 7. Finally, the storm returns to its natural state, as determined by the synoptic environment (i.e., general atmospheric and oceanic conditions surrounding the hurricane) . C. Aircraft Operations and Seeding Strategies On the basis of numerical model simulations and the hypotheses for hurricane modification, certain data have been identified as critical to the evaluation of the modification experiment. Collection of these data requires a carefully coordinated program and aircraft with sophisticated instrumentation. Approximately 20,000 self-consuming silver iodide flares will be dropped during an 8-hour period in each storm selected for experimenta- tion. The flares will be ejected at about 30,000 feet altitude and should burn out by the time they reach 12,000 feet. II-6 A total of five instrumented aircraft will participate in experimental operations. An example aircraft employment timetable is illustrated in Figure II- 4. This diagram contains inserts illustrating the number of aircraft and crews required, as well as the basic monitoring pattern superimposed upon a simulated radar presentation. The periods of monitoring by each aircraft are indicated by blocks; the small illustratations directly above or below these blocks depict the position of the basic pattern to be flown during that time period. In summary, the experimental operations will consist of Primary low-level monitoring at a 5000-foot altitude from 4 hours prior to seeding to 12 to 20 hours after seeding. (This is required because of the magnitude and frequency of the natural variability of hurricanes) . Middle level monitoring, including basic cloud physics measure- ments before, during, and after seeding (at 20,000 to 24,000 feet). Seeding at 26,000 to 30,000 feet for periods of up to 6 to 8 hours. Upper level (40,000 feet) monitoring before, during, and after seeding. In addition, nearly continuous radar monitoring, as well as measurements of sea surface temperature and its variations with depth in areas surroundinj and beneath the hurricanes, will also be obtained. The question of the most appropriate method and location of seeding has undergone lengthy study. Three strategies have been considered: Identify and selectively seed actively growing cloud cells. Fly along the first rainband intersecting the eyewall clouds and seed massively within that near rainband. Seed on a radial pattern in a sector of the storm which is generally convectively active. Each one of these strategies has its problems. In regard to the first, it is questionable whether it is practical to identify and selectively seed a significant number of actively growing cells in a short period of time in a hurricane environment. The massive seeding pattern required for II-7 V) 5 z UJ IO - \— "o Q. IS O z z u. < ir en 2 CE CM CM < < < o < ■Si m a. IX < z a. UJ -J 33 < LU 2 z LU 2 u O a si = 5 5 LU ~ o 3 S ? < 2 o 5 a § 5 .a 51 LU D a 439j ;o spuosnom- 1H3I3H II-8 the second strategy may not be operationally feasible since the number and sizes of active cells in bands is unknown. As regards the third pattern, it has been calculated that by flying a radial patttern, possibly only a relatively few ice nuclei will be distributed in the clouds immediately outside the radius of maximum winds . This means that the seeding might not have the desired effect. A panel of experts has reached a consensus on a mixed seeding strategy. P r o je ct aircraft should seed (subject to operational limitations) from within the clouds along a band, with the seeding rate accelerated when. the instruments indicate active cells are being penetrated. The seeding goal should be to inject as much seeding agent into the band, and especially into the active areas, as nature and technology will permit. The panel has discounted the possibility of overseeding. Since the convective clouds suitable for seeding are largely concentrated in one or two quadrants, any major seeding effort should be asymmetrical rather than symmetrical. Some factors which have not yet been clarified satisfactorily may have a very important bearing on future seeding strategies. For example, the bright band observed in radar pictures of typhoons and hurricanes indicates that there may be large quantities of ice crystals in the inactive clouds planned to be seeded. If true, then it is possible that artificial freezing nuclei not deposited in updrafts may become mixed with the natural ice crystals in the environment of the convective towers and thus have very little influence on later developments. In other words, seeding may provide only a small increase in the concentration of p otentially active freezing nuclei. Experiments in seeding actively rising cumulus clouds (not in typhoons or hurricanes) at about 20,000 feet did not indicate an increase in buoyancy but rather the development of heavy showers. If this is the dominant reaction of hurricane clouds to seeding, then the basis of the • Hurricane Amelioration Research Project experiments becomes uncertain. D. Project Activities 1. Deployment and Support Projected requirements for deployment and for support of equipment and personnel are similar to those outlined in the Appendix of Reference A. II- 9 2. Seeding Eligibility Rules Strict guidelines for the conduct of the Hurricane Amelioration Research Project have been defined. The criteria to be used to determine whether a hurricane is eligible for seeding experiments in the eastern North Pacific region include: The hurricane must be in the mature (stable) state, with maximum winds of 65 knots or more. The hurricane must lie within an operational area of 600-nautical miles radius around La Paz, Mazatlan, or Acapulco, and it must remain in this area for at least 12 (daylight) hours. The predicted movement of the hurricane must indicate that there is a probability of 10 percent or less that the storm center will be within 50 nautical miles of a populated land mass during the 24-hour period after seeding begins. An additional operating constraint will be that the seeding aircraft will not be within 50 nautical miles of a populated land mass during actual seeding operations. The first criterion is a scientific requirement. To properly test the hurricane modification hypotheses, only those storms which have attained the intensity of a hurricane will be seeded; the requirement of mature stage ensures a more or less steady state hurricane. The second criterion is based on operational constraints set by the readiness time and range of the available aircraft. Both safety and scientific requirements are the basis of the third criterion. Seeding before land masses influence the hurricane's behavior will avoid introducing additional complexities into the experiment. More importantly, the third criterion has been designed to prevent a seeded storm from affecting populations before it returns to its "natural" state. This criterion is based on the experience of an earlier seeding experiment (Hurricane Debbie, 1969), as well as numerical modeling experiments which show that a seeded storm will begin to return to its natural state 6 to 18 hours after the seeding has ended. It is evident that successful implementation of the third criterion depends on accurate forecasting of tropical cyclone tracks. Project personnel feel that this particular rule is very conservative and provides a significant margin 11-10 of safety. And it is suggested that as forecasting techniques improve, forecast errors generally will diminish and hence the margin of safety implicit in the criterion will increase. (An example of the application of this criterion is included as Appendix B) . 11-11 Ill PHYSICAL AND BIOLOGICAL SETTING A. Location and General Description The experimental area consists of the ocean within 600 nautical miles (1110 km) of Acapulco, Mazatlan, and La Paz (and at least 50 nautical miles from populated land). The area is roughly bounded by the Mexican coast to the east, Clarion and Clipperton Islands to the west, Guadalupe Island to the north, and an equatorial zone to the south (Figure II-l in previous section) . The experimental area contains only three populated islands or island groups, and one of these, the Islas Tres Marias, lies only 60 nautical miles from the mainland. For the purposes of this experiment, this island cluster is effectively part of the mainland. This leaves the Revilla Gigedo Island cluster, which contains one inhabited island, Socorro, on which a naval base is to be constructed, and the island of Guadalupe, which contains a naval base with a population of 100 to 150, as the only inhabited land in the area that could otherwise be seeded. B. Physical Environment 1. Geology Geographically, the study area consists almost wholly of young oceanic crust overlain by 2 to 4 kilometers of water. The study area itself is relatively simple, notable primarily for a series of major east-west trending fault zones, a paucity of islands and seamounts, and a pronounced trench that borders the continental shelf from the Gulf of California to Panama. In contrast, the known structure of the uplands along the study area's eastern border is complex. It probably resulted from a complex series of division and movements of the landmasses now comprising the peninsula of Baja Californa (Anderson, 1971), and similarly complex but little studied movements on the Mexican mainland. It is III-l currently a zone of marked seismic and volcanic activity (Heirtzler, 1968; Eardley, 1962). I The features most important to the proposed project are the narrowness of the continental shelf (rarely more than 50 km) , which tend to minimize the size of the storm surge, and the coastal ranges, which generally reach 2000 to 3000 meters elevation within 50 km of the coast. These coastal ranges minimize penetration of hurricanes into the interior but probably increase the severity of flash flooding and the frequency of landslides caused by heavy rainfall. Topography is also the most significant geologic feature of the coastal islands, most of which are characterized by rocky cliffs 5 to 50 meters high, small northerly beaches, and upland plateaus (Wilson, 1922). The oceanic islands, Guadalupe, Socorro, and Clarion, are similar and have southern anchorages. Clipperton Atoll, much of which stands only 1 to 2 meters above sea level with a maximum elevation of 20 to 25 meters , has no coastal cliffs (Doran, 1959). 2. Soils Few detailed soil surveys have been conducted in Mexico. The maps available are generalized indications of the soils expected to be present, rather than soils actually observed (Dregne , 1968; FAO, 1975). However, for this study, most areas can be expected to be susceptible to severe erosion due to lack of vegetal cover or to steepness of the terrain. Desert soils can generally be expected to possess surface layers of gravel that retard erosion; however, the volume of runoff tends to be quite high and its erosive power is correspondingly high. On the other hand, the woodland soils south of Mazatlan, which are derived from volcanic and metamorphic rocks on steep slopes (FAO, 1975), probably are easily eroded wherever the native vegetation has been removed because of the fine texture of the surface layers of the soil. Soils on the oceanic islands apparently have not been character- ized but probably resemble those of the mainland and the high islands in the western Pacific, except for the soils of Clipperton. Clipperton's soils probably are alkaline, excessively drained, and nearly devoid of III-2 humus, as are the atolls to the west (Wiens, 1962). Doran (1959) refers to the surface as coarse sand and rubble, apart from a 2-hectare area of volcanic rock. The soils of Guadalupe are derived from volcanic rocks and thus, although dry, are probably fertile. Agriculture, apart from goat or sheep herding, probably is not feasible because of lack of dependable water sources (Wilson, 1922). 3. Hydrology . The mainland bordering the study area is characterized by numerous small drainages, marked seasonal stream flow, and few permanent water supplies, particularly north of Mazatlan. The principal water supplies of the central portion of the Baja California peninsula and the eastern shore of the Gulf of California, both of which are exceedingly arid, are the groundwaters in the coarse sediments in the beds of the larger streams, occasional pools in the larger potholes in the upper reaches of these streams, and an occasional spring (Wilson, 1922). These sources are supplemented by permanently flowing streams in the southern half of the mainland's coast, roughly between Mazatlan and the Guatemalan border. However, stream flow data are not readily available for these rivers, except for the Rio Grande de Santiago, which lies at the junction of the desert and thorn forest-savanna regions at the southern end of the Gulf of California. Although the Rio Santiago is one of the larger streams of western Mexico, it is probably representa- tive of these short, small rivers with respect to its vast range in discharge, which varies nearly 200 fold during most years. Extremes between 1969 and 19 72, for example, ranged from a minimum of 16 m-Vs to a maximum of 4000 m-Vs (UNESCO, 1974). Irrigation has become increasingly important in recent decades, and a series of irrigation works stretch along the coast south of the Gulf of California. Surface water quality is probably low throughout the region because of high desert runoff and severe erosion. Based on extrapolations from southern California (CRFSC, 1970) , it is plausible to expect sediment loads ranging from the more than 15,000 ug/ml (ppm) in the most arid zones III-3 to perhaps 300 yg/ml in the least arid zones. Concentrations of dissolved solids probably range from 2500 yg/ml in the north to perhaps 350 yg/ml in the south. Exceedingly high concentrations of dissolved solids are to be expected locally, particularly in the northern deserts. The coastal islands generally have no permanent water supplies. The large oceanic islands have small permanent water supplies in springs and Clipperton has a groundwater lens with low-quality freshwater (Wilson, 1922; Doran, 1959). 4. Climate of the Experimental Areas and Their Vicinity Desert conditions generally prevail in Baja California and in northwestern Sonora on the eastern side of the Gulf of California. These areas are in the subtropical high-pressure belt that is characteristically dry around the world. Rainfall is irregular, and crops cannot generally be produced without irrigation. Almost no rain falls from February through May, and individual months may have no rain. Go.ing southward .along the Pacific slope of the Sierra Madre Occidental toward Mazatlan, rainfall increases considerably to nearly 1 to 3 inches per month in July, August, and September. At several locations, precipitation in these 3 months accounts for nearly 70% of annual rainfall. In summer, the convective rising of moist air accounts for local showers and thunderstorms. During summer, the average temperature ranges from 70° to 100°F, with maximum temperatures above 100°F almost daily in the northern desert areas. During winter, average temperatures range from 50° to 80°F. Mean and extreme climatic conditions for several locations in February and August are shown in Table III-l. Southward from Mazatlan, the Pacific coast of Mexico has a tropical savanna climate with moderate to heavy summer rainfall and a distinct dry period of 2 to 4 months during the winter. The rainfall is largely concentrated in the June to October wet season. Average annual rainfall ranges from about 30 inches at Mazatlan to 58 inches at Acapulco. The average temperature during summer ranges from 70° to 90°F, and during winter from 60° to 80°F. Inland from the fringes of coastal lowlands, a climate typical of tropical highlands prevails in the plateaus. At III-4 30 U > < O O OOO O O OC O OH in -J" ^T in rt ift ifl fN i*-or> Ps m OO O Ps pH a «i © «-< o «» inn cm r^so en tn m rv. g » l 4 n ps n cn ei o en -i CO OS w O vC O m en »H OO «T 91 O — , ^ r-t rs. —i in n (N O <—> CN OOO o o o o o m n >o -j- rs. oo en «•* o (M p-i o cm " n sr in «N m -* i f-4 o^** n oo O N rs. o.O •j — * in n rs m -^ CO Ps CO Ps CO 00 CO iO O Ps CS- oia en so en en cm in oo en en o x o n pi ^ m « ri so en o\ ~* cn o cn enoo enps co coco «T p*< rs ■& -4 O mi —> O 3> f"s p"l yQ fS. ^y o en en en en in h in 33 —> — I vO Ps O "J Li — = O CJ " rs 5 -i Z - ■:< 3 c *J -J = C "N Z. — * ^ 2 ^ > V S S V! ~ — * o 2 3 — ecu S — 3 3 < = S CJ b J £ s as «3 z sc -j < in © III-5 elevations of 3000 to 8000 feet and removed from direct marine control, the plateau areas heat rapidly by day and cool rapidly at night. Rainfall from showers and thunderstorms remains heavy during the summer, and it is similar in amount and seasonal distribution to the coastal lowlands. The entire area is dominated by the easterly trade winds and the proximity of the Intertropical Convergence Zone (ITCZ) during summer. The eastern North Pacific ranks second only to the western North Pacific in the development of tropical storms and hurricanes (typhoons). Hurricanes are concentrated in June through October. August has the highest frequency, with two to three of the annual average of seven to eight hurricanes. Hurricane activity is greatly influenced by the ITCZ, between the northeasterly trade winds of the Northern Hemisphere and the south- easterly trade winds of the Southern Hemisphere. In the eastern Pacific, this zone reaches its most northerly mean position (10°-15°N) in August and September (see Figure III-l). Hurricanes generally begin as tropical cyclones south of Acapulco and then move west-northwest (see Figure III-2). Conventional and satellite data show that eastern Pacific hurricanes are relatively small, with moderate intensity: 75 percent dissipate before reaching the 145°W meridian. These characteristics and the predominant west-northwest tracks are attributed, in part, to the area of warm (>80°F) water over which hurricanes form, an area restricted by the proximity of very cold water on either side, and to high-speed tropospheric westerly winds (see Figure III-l). Of the 66 hurricanes that occurred from 1965 through 1974, 7 hit the Pacific coast of Mexico and 4 of these moved into Baja California. The coastal area most affected by tropical cyclones extends from Acapulco (16.5°N) to Navojoa (27°N) , where 12 tropical cyclones made landfall from 1965 through 1974. 1 1 1-6 u, O -1 < to a z O -J < to Q 2 I OC ai H < 3 N O UJ Z s ai o z UJ a i UJ > Z o u -I < u & I h- X UJ Fil jj! i tii cc UJ I- < 5 UJ a P -i < H < OJ a z 3 H i/> UJ 5 UL O 3 a < > < a UJ z T j -> o H as < UJ CC H UJ < a. 2 2 Ul X H o UJ CC U < 3k X < u. a: "3 -j < 3 Ul •t w 1 Ul Q Z o CO Z LL £ O UJ a a -1 > Ul > UJ 1 CC < a 2 03 r> z 2 1- Ul CC D a III-7 z o a 3 o X a — • x at X 1- o o o »• z £ 'A a — o ,« * »• < s> i os - p* 1- w* LU jj < » u i OS (/J 3 « ,« -3 ifl »• = , o C -J ' •" 3 < >" i (J 3 «■ ~J 3 a. 3 39 > a « M "2 H> C ^ 5 •3 «' 3 3 o 9 g 01 #• OS » 3 ci g p<1 3 ■% E III-8 5. Oceanography Three surface water masses are found in the vicinity of the experimental region: equatorial Pacific water, subarctic North Pacific water, and eastern North Pacific water. 1 There are four major current systems: the North Pacific Current, the California Current, the North Equatorial Current, and the Equatorial Countercurrent. 1 ' 2 » 3 The waves in the experimental area are generally small. The waves are highest during the winter, being less than 5 feet about 80 to 90% of that season. The highest' waves and strongest currents in the experimental region are caused by the passage of tropical storms. C. Biological Environment 1. Vegetation The vegetation patterns of Baja and the mainland reflect the rainfall in this area, which is seasonally to permanently arid (Figure III-3) . The northern portion of the Baja peninsula, which receives fairly dependable winter rains, is covered at low elevations by the hard, prickly leaved shrubs or chaparral characteristic of southern California. At higher elevations, the chaparral is replaced by stands of oak and pine that contain small quantities of marketable pine. East of this zone is an extremely arid Colorado River delta marked by hot springs, mud volcanoes, salt flats, and sparse desert vegetation. Portions of this arid zone are irrigated with waters from the Colorado River, but the remainder currently has little economic value. To the south of the chaparral zone lies the arid Vizcaino sector of the Sonoran Desert, portions of which have no rain for periods of 2 to 5 years. The shrub communities throughout the deserts of the Baja peninsula are generally less than 1.5 meters high, and on any given site are comprised of 2 to 20 species. However, because the species composition changes markedly with changes in location, drainage, and soil type, the vegetation for the desert as a whole is quite diverse, and particularly so with respect to the growth forms of the plants. Small leaved, evergreen shrubs, III-9 z o a Ui ac -4 2 o o to z oc Ui fc I z o g a LU > LU ac D a - £ 111-10 deciduous shrubs, leaf succulents, and stem succulents abound. Trees are few and restricted to the wettest sites and the heartier species- palms, cottonwoods, and willows (Shreve, 1951; Shelford, 1963; Wilson, 1922). The desert along the eastern shore of the Gulf of California is similar to, but wetter than, the western coast in the northeast, and the shrubs are consequently larger. The average height of the vegetation is greatest in the foothills of Sonora, reaching 5-6 meters, although heights of 10-12 meters occur in the plains of central Sonora (Shreve, 1951). Shrubs and small trees, such as cottonwood, willow, and mesquite dominate the floodplains where subsurface water is available, but sage and similar small shrubs dominate the drier alluvial plains where ground- water is absent. These in turn are replaced by smaller species and by leafy succulents on the shallow, rocky soils of the steeper mountain slopes. The lowland deserts merge gradually along their eastern edges with the shrub and grass communities, designated as mesquite-grassland in Figure III-3. Along their southern edge, the lowland deserts merge with the thorn forests, which are comprised of small, multistemmed trees with compound leaves and numerous, often massive, thorns* These forest or shrub communities are characterized by a fairly dependable, but brief rainy season from July through September. Both annual and perennial herbs are numerous and in the wetter areas are sufficiently abundant to support occasional fires* (Shelford, 1963). The thorn forest is replaced along its southern edge by savannas and winter deciduous tropic forests. Although savannas are variable, they are typically coastal and occur in the flatter terrain. They range from mosaics of grasslands and trees to grasslands with scattered trees or grasslands with ribbons of forest. The latter are characteristic of level uplands transected by steep walled valleys; such topographic discontinuities inhibit the spread of fire (Shelford, 1963). Deciduous forests lie in a belt inland from the savannas and occupy the portions of the mountains steep enough to inhibit fire and agricultural land clearing, and hot enough to sustain tropical species. *Fire is an important causal agent in vegetation distribution. III-ll The forests characteristically contain one or two layers of trees, and on the driest sites, where the tree canopy is sufficiently open, a dense shrub layer. These species typically produce flowers and fruits during the winter dry season and grow vegetatively during the wet seasons. Oak and oak-pine forests inland of the deciduous tropical forest dominate the upper slopes and all but the highest and wettest elevations, which are covered with the cloud forests (temperate rain forests) and at extremely high elevations, alpine meadows. Oak forests occur at 1000 to 2300 meters, pines and alpine meadows above 3800 meters. The cloud forests are distinguished more by climate than by the species composition of the major trees, but possess numerous species of vines and epiphytes, and a lush herbaceous understory. The oak and pine forests are comprised of numerous and relatively small trees (9 to 10 meters) and are relatively open with fairly dense understories of grasses, other herbs, and shrubs (Shelford, 1963). The vegetation of the offshore islands generally resembles that of the adjacent mainland, although portions of the larger islands such as Guadalupe and Cedros once possessed forests on their uplands. These forests were declining rapidly in the 1920s because of heavy grazing by goats (Wilson, 1922), and are probably nonexistent now. Several species have been eliminated from the flora of Guadalupe, a few of which have persisted on the goat-free island of Outer islet at the southern end of Guadalupe (Carlquist, 1965). The vegetation on the islands of Clipperton and Clarion is comprised largely of beach species (Carlquist, 1965; Doran , 1959), and San Benedicto is probably similar because these are the species that are most likely to have invaded the island after it was covered with ash and lava during an eruption in 1952 (Carlquist, 1965). Socorro, however, supports a herd of sheep and probably has an herbaceous flora comprised of grasses, introduced grassland weeds, and possibly shrubs of Mexican origin. Roca's flora is probably similarly impoverished. 111-12 2. Fauna The wild animals of greatest interest to the proposed project are the aquatic mammals found in Guadalupe and the coastal islands near the peninsula of Baja California. The marine mammal fauna of these islands is exceptionally diverse and is comprised of species that are recovering from a period of near extermination in the late 1800s. These species include the Gudalupe fur seal, the northern elephant seal, the California sea lion, the harbor seal, and the gray whale (Bartholomew, 1967; Wilson, 1922). The Guadalupe fur seal and the elephant seal both persisted as small groups of perhaps 20 to 100 individuals on Guadalupe and have been slowly increasing in number in recent years. The elephant seal has migrated northward and now breeds as far north as Ano Nuevo Island near Santa Cruz, California. The ■ Guadalupe fur seal has recovered more slowly and consisted of only 400 to 600 individuals as late as 1965, with only 1000 in 1970 (U.S. Fish and Wildlife, 1973). It breeds only on Guadalupe but occasionally is seen on other islands. The sea otter, which was once abundant in this region may occasionally visit these islands, probably numbers less than 2000 (U.S. Fish and Wildlife, 1973). The gray whale, which breeds in the lagoons on the central coast of Baja during the winter, has made a successful recovery in the eastern Pacific but is still of concern as a species because of its uncertain population status in the western Pacific (U.S. Fish and Wildlife, 1973). The terrestrial species of greatest concern are the Sonoran pronghom and the peninsular bighorn, both of which are restricted to northern Mexico and adjacent California. The Sonoran pronghom, which probably numbers less than 1000, differs in size and coloration from related sub- species, which are widely dispersed in western North America. Similarly, the peninsular bighorn, which now numbers about 100, possibly less, is a supspecies of the rocky mountain bighorn (U.S. Fish and Wildlife, 1973). A few reptiles are restricted to Cedros, but the faunas of the offshore islands generally consist of species found on the mainland (Savage, 1967). 111-13 The remainder of the terrestrial fauna is comprised of more widely spread and more numerous species. The largest of the wild game in the desert region are deer, pronghorn antelope, cougars, and coyotes, all of which are found throughout western North America. So are the smaller mammals, such as the ground squirrel, the white-spotted mouse, the jack rabbit, the kangaroo rat, even though they are often represented by subspecies not found elsewhere. The reptiles of the deserts are less broadly distributed than the mammals, but nonetheless occur over very large areas. The forest fauna broadly resembles that of the desert but the species often differ (e.g., white-tail rather than mule deer). In addition, the number of mammal species tends to increase (e.g., the mountain lion and coyote are joined by the jaguar, ocelot, and wolf). The marine fauna is diverse, and largely tropical. Corals extend into the Gulf of California, but the reefs in the Gulf are com- prised of relatively few species (DiSalvo and Odum, 1974). Diversity increases southward, attaining higher levels of diversity in Panama than are found in the Caribbean at the same latitude (Porter, 1974). The structure of these reefs is broadly similar to that of the more inten- sively studied reefs of the western Pacific, with a steeply sloping reef front populated by massive, compact corals; a broad flat upper surface at or just below low tide level; and an interior slope with more delicate corals bordered to the landward side by a lagoon. The lagoon may be covered with seagrasses, and the landward edge is often bordered by one or more zones of mangroves. The reefs are replaced to the north by beds of kelp and other seaweeds which, like the corals, provide cover and food for numerous fishes. In summary, the number of species increases southward along the coast in both terrestrial and marine communities, and the number of species present per unit area tends to be high in all but the driest of the desert areas. However, the only animals that are rare enough and exposed enough to the potential effects of the proposed seeding to be of significant concern are the marine mammals of Guadalupe and the islands along the Pacific coast of the Baja peninsula. (See Section VII. B.) 111-14 IV HURRICANES A. General At present there is no common international terminology for tropical cyclones of differing intensities. Even the general term "tropical cyclone" is subject to various interpretations. The diversity of classification systems currently in use poses serious problems in accurate interpreta- tion or comparison of tropical cyclone forecasts and statistics prepared by various national meteorological services. In this report, the term tropical cyclone is used in its broadest sense and is defined as a cyclone developing over tropical or subtropical oceans, having a definite closed surface circulation (wind flow). The intensity will be described using the classifications defined by the U.S. meteorological services: Tropical Depression: A weak tropical cyclone having highest sustained windspeeds less than 34 knots. Tropical Storm: A tropical cyclone having highest sustained windspeeds of 34 to 64 knots. Hurricane or Typhoon: A tropical cyclone with highest sus- tained windspeeds of 64 knots or more.* Tropical cyclones of this intensity in the North Pacific Ocean are generally called typhoons west of 180° and hurricanes east of 180° longitude. In these definitions, sustained winds are averages over 1 minute; extreme surface gusts in typical cyclones may be 30 percent to 50 percent higher than the reported sustained surface winds. Tropical cyclones develop principally in eight oceanic regions of the globe: eastern North Pacific, western North Pacific, north Indian Typhoons with winds of 130 knots or more are sometimes referred to as supertyphoons. To illustrate classification system differences, tropical cyclones with winds greater than 35 knots are called typhoons by Japanese meteorologists. 1 IV-1 Ocean (Bay of Bengal and Arabian Sea), south Indian Ocean, western South Pacific, Australian area, and North Atlantic (including the Caribbean and the Gulf of Mexico). Figure IV-1 shows the average annual number and percentage of the global total of tropical cyclones (or tropical storm or greater intensity) occurring in each storm development area. Over half of the average global total of 80 tropical storms per year occur in the North Pacific (38 percent in the western North Pacific and 17 percent in the eastern North Pacific). Observations from meteorological satellites during the past few years have established that the frequency of tropical storms and hurricanes in the eastern North Pacific is much higher than previously thought. B. Formation of Tropical Cyclones Research has established that a number of conditions are required for a pre-existing disturbance to develop into a tropical storm. The disturbance must be located: Over a large ocean area with a water surface temperature greater than 26° or 27°C (80°F) More than 5° in latitude from the equator In a region of small vertical wind shear through the troposphere (i.e., only small changes in wind direction or speed with altitude, up to about 35,000 to 40,000 feet). The large monthly and annual variations in the incidence of hurricanes observed in all areas can be related to variations in sea surface tempera- ture and vertical wind shear about their climatological norms. However, in spite of general agreement among meteorologists on the physical environment necessary for storm formation, there is considerable IV-2 Crt 2 cr o H CO -J 3 < O LU 0. CC O < Sfi LU LU o s — It « O W -l S UJ 2 > 3 lu Z Q -1 LU < Z 3 O 2 (J Z J < > O ?! _1 LU W < > z H- < — o r* _ 1 »— < > 33 O LU o X «— D u. o 9 U g H 3 LU z LU a o s cr X jj a. ^ 10 IV- 3 disagreement about the synoptic situations (general weather patterns) in which they actually form. C. Climatology 1. Occurrence The average monthly frequency for hurricanes in the eastern North Pacific is indicated in Figure IV-2 . Similar to the western Pacific, the highest frequency occurs in August, but the secondary September maximum is greatly reduced. In contrast with the western North Pacific, no tropical storms are observed from November to April. One analysis (Renard and Bowman, 19 76) shows that from 1965 through 1974, 45.5 percent of the 145 tropical storms reached hurricane intensity and that the 15-day period with the highest frequency of hurricanes in the eastern Pacific was August 16 through August 31, during which 16 out of 66 hurricanes were recorded in the 10-year period. The highest frequency of typhoons in the western Pacific was August 24 through 28 as analyzed over a 24-year period (Reference A) . Figure IV-3 shows the variability in the annual occurrence of hurricanes from 1966 to 1976; 1971 and 1974 were years of peak activity. A word of caution is necessary about comparing hurricane data for the eastern Pacific during different years. For example, for 1947 through 1953, an average of 5 cyclones per year were recorded as having reached tropical storm intensity. The average number cited doubled to 10 per year from 1954 through 1961. Rosendal (1962) attributes this increase to improved aerial reconnaissance, increased shipping, and better For example, while some view the development process as a progressive intensification of an easterly wave, ' 4 others 5 attribute the majority of storm developments to intensification of low level vortices which form along troughs. Gray 2 ' 6 attempted a reconciliation of the above differences by analyzing composite wind data of 312 disturbances which later intensified into tropical storms. His analysis supports the concept of storm formation at the surface position of the trough only in the Pacific region. The North Atlantic-Caribbean area exhibits the most complicated processes of any development region. IV-4 o a > o z x UJ O < — 3 S a. < at < 2 Z ac LU H 9 < LU lL o z g H «M. ^ D a 0) * VJ CO 03 o 5 ]vT Z > • o V3 lO o z LU z LU 3 > 3 a < a LU cc 5 LL. 3 > > I LU CC -J o LU h- LL z CJ 21 o < < 2 3. UJ 3 CN > > -j — ■r Z LU o D 2 g u_ 9 U 3 io hinow a3d S3NvoiaanH do asawnN aovaaAv IV- 5 O SO 3 V CN < > J z JJ X (J o o 'JJ z < c 'J < < z Q x < > z z < > X 3 IV- 6 communications. With the advent of the meteorological satellite in 1960, detection capability greatly increased. Using data coverage from the TIROS satellites, Sadler (1964) showed that for August to October 1962, only 5 of 22 probable tropical storms detected by the satellites were reported by ships and other conventional means. Since 1966, the excellent satellite coverage has supplied most of the cyclone fixes and intensity data for the bulletins and advisories. Therefore, from 1966 on, statistics on tropical cyclone activity in the eastern North Pacific are much more reliable. 2. Movement Generalized mean tracks of hurricanes for the eastern North Pacific are available in various climatological publications (e.g., Crutcher and Quayle, 1974; Hansen, 1972). Mean tracks, however, do not reflect the large variability of storm tracks. If tracks of storms covering a long period are plotted on the same map, they are not only complex but greatly interfere with one another (see, for example, Figure III- 2) . Consequently, it is often preferable to use actual tracks occurring in shorter periods, such as a few months or a specific year. Cyclone tracks in the eastern North Pacific during August, September, and October 19 76, are shown in Figure IV-4. Tropical cyclone tracks in the eastern Pacific Ocean are reasonably uniform in direction in comparison with those in the western Pacific (track directions are shown in Figure IV-5, based on data for 1956 through 1971). Eighty percent of the tracks' directions fall within o + 20 degrees of 290 (approximately west -northwest) . Tropical cyclones of hurricane intensity occur predominantly east of 130 W. Excursions westward of this longitude are confined almost exclusively to the "high o o season" of 16 July through 15 September. The region between 15 and 20 N o eastward of 120 W is that most heavily traversed by hurricanes. 3. Recurvature Recurving cyclone tracks are those westward tracks that turn clockwise until there is a component of motion toward the east. From IV- 7 :£L ; -> ' i i i S * « > a s s | { 2 * | » if MUl i a I is 1 1.1 a w * 1 «: M . i u mi < u 0. o oe H u u. x oc H UU ■^ e m < O p Z H V z8 O o: _ Z LU X . h a if) O 2 = S o • iu £ 3 > u. X O H 3 CO H '.U > LU < 52 X <= 5 oe i- < Ul K «T < > M 5 IU > (X _i D a z u. o s a a w 3 3 VI IV-8 56r 48 - 40 U < IX 32 - 24 - 33 a 2 H 1 1 1 ! 1 1 - MEAN 298° MODE 290° - - - - - — / — / 1 1 I 1 1 240° 260° 280° 300° 320° 340° 360° DIRECTION - degreus N 020° 040° Source: Hansen. 1972 FIGURE IV-5. FREQUENCY DISTRIBUTION OF TROPICAL CYCLONE TRACK DIRECTION IN THE EASTERN PACIFIC (1956-1971) IV- 9 1956 through 19 71, only 14 percent of the 191 tropical storms and hurricanes in the eastern North Pacific recurved. Statistics for 1965 through 19 74 show that 48 tropical storms and hurricanes out of 145 (i.e., 33 percent of the 10-year sample) recurved. This observation of increased frequency of recurvature is most likely due to the more complete information available from satellite observations. Table IV-1 presents data for recurving storms in the eastern North Pacific, using the most recent (1965 through 1974) records. As indicated, tropical storms and hurricanes in the eastern North Pacific are most likely to recurve at the beginning and at the end of the season. During September and October, the probability of recurving exceeds 50 percent. The main danger of recurving tropical cyclones in the western North Pacific lies in their sometimes rapid acceleration after recurvature. Recurving cyclones in the eastern North Pacific do not generally follow this behavior. They are characterized by slower mean speeds (7 to 8 knots) before recurvature and little if any acceleration after recurvature. Figure IV-6 shows the frequency and mean speed of tropical storm and hurricane movements toward the northeast* for 1965 through 19 74. After recurvature, the highest direction frequencies toward the mainland are found east of 115°W and south of 25°N. The mean speed increases with latitude. Figures IV-7 and IV-8 show similar data for September and October, when the percentage of storms that recurve in the eastern North Pacific is highest (see also Table IV-1). Northeasterly moving storms north of 25°N occur in September only. South of 25°N, storms moving toward land occur most frequently in the area west and southwest of Mazatlan in September and October. For 1965 through 1976, when satellite observations provided more complete information on tropical storm activity, the number of storms recorded that hit the main coastline of Mexico is more reliable. Figure IV- 9 shows the areas where most of the storms crossed land during the latter portion of this period. The highest incidence of landfall occurs near 20°N. The numbers include all storms traveling within 22° on either side of northeast. IV-10 CU > M 3 CJ 01 as eg Eh CJ cj < z; « w CO < Be] en < a ■z < CO S CO < O cs I m ON c o s u 3 CJ 0) 1) CJ •H u U 3 u C 3J c^ o ON CM ON 00 m m OS CM CN CN 0> 00 o CN ON 00 r— | CO m on ON CN ON cj OS co -3 - 3 !-i 0) -3 u 4-1 a 0) CO cu ,3 J >, 3 u § H 60 n _i 3 3 OJ "*S — ^ < CO o as IV-11 \a 00 Z o H O 41 | 41 > 5 3 5 > oc M V} < 3 o vO LU mm 3 X H» e O z z • 5 3 O" «i T3 i 3 s a z ,5 3 > o 2 CO 5 cc O 1 £ 3 — > 9 <•* o 3 co -J < o O t/s lO -1 ^ UJ (£) c ■■3 Z < O 1 a u a s 3 '.O IT e z "3 *■ T3 a -j % c X G v» ■3 3 e CO 3 e > £ UJ s a *~ s a O IV- 12 uo a O £ § s s i M C a 2 "3 00 *■* 5 vt X i) a 3 i— 3) 00 SI -J w < a. U CL D o *™"* cr - ?■• V r* a to z u3 < o» 00 ■*•" LU cc Z LU < 33 — U ^ 2 PTE kno _) 'JJ c 00 — s o 8 IV- 13 io CO URING peed s IRECTION D e is average s s 5) lEASTERLY D are; lower figur 3 8 £ 1 X i» 3 VING IN NO ruber in grid s Q 2 STORMS W ure is total 3 o < - *■ 3 a ? 0> x £ c i- i o ? i Z f> 3 .fl o 4M ■3 to as s LU J* z *•* •3 < a: a CO 9 > 3 LU 9 X a 3 3 3 3 a o IV- 14 a> 8 X ill S UL O H w < O u z £ UJ H- vs Ui 3 ui O z V5 C/l 3 s o ^ o , Si en fca 9 *7 UJ J (T in o h- 3 2 X V) Si 2 9 ID < -J < S £ o ai 3 a. O 3 H (O > IT* S si Z a CL X o > o 5 2 X 3 UJ H — 9 e ! i 3 5 3 Ifl cs 3 a es u. o n IV- 15 4. Duration Table IV-2 summarizes for 1965 through 1974, the life of the average tropical cyclone in its various stages of intensity. Note that hurricanes have an average life history of 2.6 days, nearly the same as a tropical storm. The hurricane area of the eastern North Pacific is Table IV-2 AVERAGE TIME TROPICAL CYCLONES SPENT IN EACH OF THREE STAGES (Eastern North Pacific, 1965-74) Time in Stage Documented As Tropical Cyclone Category Depression Storm Hurricane Tropical depresssion 100% 2.9 days Tropical storm 39% 61% 1.9 days 2.9 days Hurricane 22% 43% 35% 1.7 days 3.3 days 2.6 days Source: Renard and Bowman, 1976. near dissipating influences that greatly complicate daily storm intensity problems. Cold water that upwells along the coast from about latitude 24°N northward is spread to the west-southwest by prevailing currents. Westerly winds in the upper troposphere, which are related to the Hawaiian branch of the subtropical jet, are northwest of the ridge line that is usually oriented east-northeast to west-southwest near the edge of the cold water. A tropical storm dissipates rapidly when it moves under the margin of a westerly jet, that is toward the equator because the jet favors upper convergence, entrance of cold air aloft, subsidence, and drying. (See Figure III-l, Chapter III.) The dissipation of many eastern North Pacific hurricanes has been attributed to the upper part of the storms being sheared off by the high-speed westerly winds of the troposphere (Figure III-l, Chapter III). IV-16 5. Size The mean radius from the eye of a tropical cyclone to the outermost closed isobar of the surface pressure analysis has been commonly used as the measure of western North Pacific tropical cyclones. However, this method yields large variations in apparent size because of sparse data and subjective analysis of surface pressure. The concept of Mean Circular Cloud Diameter (MCCD) was introduced to define a parameter for estimating maximum tropical cyclone winds from satellite photography. The MCCD of a tropical cyclone in the eastern North Pacific is more accurately determined than measurements relying on the outer closed surface isobar analysis because of the good satellite pictures available daily. Using MCCD to describe tropical cyclone size is suggested as an improved climatological measurement. The MCCDs of 40 recent tropical cyclones of the eastern North Pacific were measured from satellite pictures and compared with 24 tropical cyclones of the same period from the western North Pacific. The results are summarized in Table IV-3. Table IV-3 CIRCULAR CLOUD DIAMETERS (DEGREES LATITUDE) OF TROPICAL CYCLONES IN THE NORTH PACIFIC Diameter Range Mean Di ame t e r Standard Deviation 40 eastern North Pacific 1.5-5.0 3.3 0.5 24 western North Pacific 2.7-7.2 4.7 1.0 22* western North Pacific 2.7-6.5 4.5 0.6 Gloria (1969) and Joan (1970) were excluded from this sample. Their respective diameters of 6.5 and 7.2° place them in a category of cyclones not found in the eastern North Pacific. Source: Hansen, 1972. IV-17 The standard deviation of the western North Pacific sample was increased considerably by the inclusion of two very large typhoons: Gloria (1969) 6.5° and Joan (1970) 7.2°. By excluding these two cyclones from the sample, the standard deviation of 1° was reduced to 0.6°, comparable with the variability found for the eastern Pacific sample. The difference in mean values of the two samples objectively supports the observation that eastern North Pacific tropical cyclones are indeed smaller than tropical cyclones in other areas. The mean eastern North Pacific cyclones typically covers an area only one half that of the tropical western North Pacific cyclone. 6. Intensity Reliable data on maximum windspeeds for tropical storms and hurricanes in the eastern North Pacific are available only for 1972 through 1976, when good satellite data coverage enabled classification of storm intensity and maximum wind speed. Figures IV-10 and IV-11 summarize the maximum windspeed information for the 42 hurricanes that occurred during that period: 18 hurricanes (43 percent) were associated with maximum winds of 100 knots or more, and more than half occurred after July. Storms with maximum winds exceeding 110 knots can occur during any month of the hurricane season. D. Typical Characteristics of a Mature Hurricane The mature stage of a hurricane is that period during the storm when it is at or near its maximum intensity. This stage lasts from a few hours to a week or longer. It is characterized by relatively constant pressure at the center, while the atmospheric pressure in surrounding areas is gradually falling. The area affected by strong winds may increase slightly. The area of clouds and weather associated with the hurricane is usually largest, and the eye is most clearly defined in the mature stage. Because the meteorological elements are not distributed uniformly throughout all sections of the hurricanes, it is customary to describe the storms in terms of the four quadrants separated by the line along IV-18 JUNE-NOVEMBER AUGUST. SEPTEMBER, OCTOBER 30 90 100 110 MAXIMUM WINO SPEED - 64 knots) is about 60 nautical miles, but it can vary from as small as 15 nautical miles to in excess of 100 nautical miles. The innermost portion of the hurricane is the eye, in which windspeed diminishes rapidly as the point of lowest pressure is approached. The size of the eye as determined by the radius of eyewall clouds varies from less than 10 to 30 to 40 nautical miles. Although it is usually depicted as circular, the eye sometimes becomes elongated, whereas at other times it is diffuse with a double-structured appearance. 2 . Circulation The circulation of intense storms extends upwards to around 14 to 15 km (45,000 to 50,000 feet). The circulation of air within hurricanes may generally be divided into three layers. The inflow layer extends from the surface to about 3 km (10,000 feet) and contains pronounced radial motion of air towards the center; the midlayer, from 3 km to 8 km (10,000 to 25,000 feet), is characterized by mostly tangential airflow with little or no radial flow; and the outflow layer, in which maximum outflow occurs, at about 12 km (40,000 feet). The vertical shear (change) in wind is generally quite small up to about 6 km (20,000 feet); therefore, measurements of windspeed within this altitude can be extrapolated to indicate surface windspeed. Air flows through the inflow layer, rises primarily in the eyewall cloud, and finally flows outward from the storm IV-21 Souro: Rrf»nmot 7 FIGURE IV-12. COMPONENTS OF A MATURE TYPHOON IVVAJ "^ T .... ! ...„. i 1 — i — r t ! : 1 - \ ^•. - 980 •\ >k> _ J3 £ • Ui •\\ • • "" 1 960 I \ 900 * *U i 1 ; i i , ! ! I ! \ ' \ .,..1 40 60 50 !00 120 140 MAXIMUM SUSTAINED SURFACE WINC- Knots Source: ntfartnc* 3 FIGURE IV-13. RELATION BETWEEN MINIMUM STORM PRESSURE AND MAXIMUM WIND IV- 2 2 top and sinks some distance away. A small sinking motion also occurs inside the eye. 3. Pressure Field From the time a storm enters the deepening stage (period when the central pressure is falling) until it dissipates completely, it has a clearly defined low-pressure center that extends from the surface to the upper levels of the storm. In the northern hemisphere the pressure gradient is generally greatest to the right of the center and weakest to the left. ' The magnitude of the pressure gradient may vary from 0.5 to 2 millibars /mile and more; total pressure drops of 50 mb in 50 miles are not uncommon. Minimum surface pressures below 950 mb frequently occur in mature hurricanes. There is a close relationship between the central mimimum surface pressure and maximum sustained surface winds within the storm (Figure IV-13) . 4. Temperature Field Intense hurricanes convert heat energy to potential energy and potential energy to kinetic energy. Their primary energy source is the release of latent heat of condensation that occurs in the eyewall and spiral rainbands. Maximum warming occurs in the upper levels, with temperatures being 10°C or more above normal (Figure IV-14) . Although the mid-level maximum temperature gradients are concentrated in a narrow band, very small gradients are noted within the eye itself. 5. Cloud Distribution As shown by radar and satellite pictures, the major cloud systems in hurricanes have a spirally banded structure. Progressing inward, the amount and vertical development of cumulus clouds increases until the bands merge into a nearly solid mass of clouds which forms around the eye of the storm. The cloud distribution within the eye varies greatly from case to case; on some occasions the eye can be almost free of clouds, but generally small amounts of low, middle, and high clouds are present. The maximum height of cumulonimbus clouds in hurricanes is related to the intensity, and it is not uncommon for these heights to exceed 15 km (50,000 feet), especially in the eyewall clouds. Figure IV-15 presents a IV-23 W 100 80 60 40 20 20 40 60 80 100 £ RAOIAL (DISTANCE FROM £Y£ IN NAUTICAL MILES Sourcs: Reference 9 FIGURE IV-14. VERTICAL CROSS SECTION; OF TEMPERATURE ANOMALIES IN HURRICANE (Difference in °C. Relative to Mean Tropical Atmosphere) y//////////?y / /////^^ £|pSO>il-E Source: Reference 10 FIGURE IV-15. SCHEMATIC CROSS SECTION OF WINDS. CLOUDS. AND PRESSURE OF TYPICAL HURRICANE IV-24 schematic cross section of wind, clouds, and pressure for a typical hurricane. 6. Rainfall Distribution Very heavy rainfall is generally associated with mature hurricanes. A typical mature hurricane moving at a speed of 12 knots can produce about 11 inches of rain in 48 hours at a location on the track of the storm. Precipitation is concentrated in the inner core, and the distribution extends farther to the right than to the left of the direction of motion. Rainfall can change considerably in case of recurving hurricanes; it is concentrated in the right front quadrant of storms moving across water onto land. The correlation between rainfall amount and minimum sea level pressure (over a period of 24 hours) is quite strong for recurving hurricanes to the west of a land mass. Because of a lack of data it is difficult to draw a definitive picture of rainfall distribution associated with hurricanes over the ocean. 7. Severe Weather Phenomena Thunderstorms and tornadoes occur in association with hurricanes. The typical distribution of thunderstorms is not known because of inadequate data. Thunderstorms are generally observed in outer regions of hurricanes and less frequently in the inner core, perhaps because they are obscured by the mass of clouds. Tornadoes associated with hurricanes are typically generated in the right front quadrant. They are directly related to the storm's intensity as it strikes land. The majority of the hurricane-related tornadoes occurring over the United States were within 100 nautical miles of the coast. (See Figure VI-13) . There appears to be very little correlation of tornado occurrence with storm velocity, direction, or time of day. No Pacific hurricanes. time of day. No such information is available for eastern North IV-25 E. Natural Variability of Hurricanes Hurricanes display large natural variations in strength and struct' For example, changes in maximum windspeeds of 10 percent or eye diameter changes of 10 nautical miles or more in a 6-hour period are not uncommoi These transient variations may be associated with either distinct struc- tural changes (such as a double eyewall structure changing to a single eyewall) or interaction with synoptic scale features such as an upper level cold low and passage over a cold underlying sea surface. In addition, storms have been observed to undergo drastic changes when the> pass over land areas — both large continental land masses and relatively smaller islands in the middle of oceans. This natural variability makes it difficult to determine whether changes in seeded hurricanes are indue by seeding, by natural forces, or by a combination of influences. F. Prediction of Hurricanes and Their Movement Hurricane warnings and protective measures depend on an inter- locking chain of weather and oceanographic prediction. Prediction begins with detection and tracking of a hurricane seedling or rain cloud cluster. Next is the prediction of the development of the seedling into a storm with sustained winds of gale force (34 knots or more). Then, (1) the growth and intensification of the circular wind system (commonly referred to as the hurricane vortex) , (2) the movement and landfall of the hurricane center as it approaches populated land areas, and, finally (3) the storm surge which often inundates the coastline as the storm passes over land must be predicted. The most important meterological tool for detecting and tracking storm systems is the weather satellite. The polar orbiting satellite provides successive sightings of storm systems that make it possible to track and estimate their growth. The geostationary satellite, positioned 22,300 miles above the equator, normally provides storm surveillance pictures once every 27 minutes. Movie loops are compiled from these pictures to show the motion of clouds, and often the precise movement of the storm center. IV-26 Direct probing of the vortex by reconnaissance aircraft is used to gather information that is used diagnostically to determine asymmetries and the state of development of the storm system. It can also be used to determine environmental influences which may reflect what is about to happen in the life cycle of the storm. Reconnaissance information is also very useful for numerical simulation of storm movement and behavior. However, the intricate interactions between the individual convective cells of the storm and the larger scale environment are still not known fully, and thus presently, the task of predicting development and growth of the vortex is a matter of diagnostic rather than prognostic reasoning. At Weather Service Forecast Office, Redwood City, California, several objective techniques (described below) are used to aid forecasters in the preparation of storm warnings. These techniques are based on extrapolation of previous storm positions, "steering" by the large scale circulation pattern, and statistical computation that is a weighted average of selected storms from previous years. However, skillful forecasting of tropical storms with the aid of these techniques also requires accurate location information for each tropical cyclones, particularly for the 12 or 24 hours prior to the forecasts. Errors in the direction of the storm's motion of only a few degrees can produce very large errors in 72 hours. Although aerial reconnaissance has been less frequent in eastern Pacific hurricanes than in western Atlantic storms, over 50 operational reconnaissance missions have been conducted in the area and 3 research flights flown in mature Pacific hurricanes since 1973. In addition, satellites have been observing that area for more than 15 years, and with the presence of GOES in recent years, pictures are available at one-half hourly intervals. Storm locations determined from high-resolution satellite products, particularly in the mature stage required for STORMFURY operations, are quite accurate and comparable to standard aircraft reconnaissance fixes. Techniques have also been developed using satellite products to estimate the intensities of hurricanes. These estimates have proven to be quite good statistically, but relatively large errors are made occasionally. IV-27 In recent years the large number of Atlantic area tropical cyclones with anomalous motion characteristics have highlighted inherent weaknesses in the purely statistical prediction systems. This has prompted the development of dynamical and statistical-dynamical models. However, it is not likely that significant improvements will be forthcoming over the next several hurricane seasons. In the interim, statistical forecasting systems probably will be modified so as to weight persistence and climatology commensurate with the ability to provide storm positioning errors in an operational environment. The National Hurricane Center (NHC) Miami, Florida, uses output from a number of prediction models as objective guidance preparatory to issuance of tropical cyclone advisories.* A brief description of the techniques that have been in operational use at NHC over the past several hurricane seasons is given in Table IV-4. Each of these systems is completely computerized and requires the specification of one or more parameters involving storm position and storm motion as partial data input each time the program is run. The ability of statistical tropical cyclone displacement models to improve forecast skill, often referred to as reduction of variance, is derived from predictors systematically selected from climatology, persistence, or from some type of environmental data that imply a synoptic steering environment. The method by which these three predictor classes are utilized by the NHC72 prediction model is illustrated in Figure IV-16. The final NHC 72 forecast is a composite derived from inputs from two subsystems, identified in Figure IV-16 as CLIPER (acronym for climatology and persistence) forecast and a synoptic forecast. Appropriate weighting factors are applied to the subsystem forecasts to arrive at a final NHC72 forecast. * National Weather Service Eastern Pacific Hurricane Center at Redwood City, California used the techniques of NHC to compile digitized eastern Pacific tropical cyclone tracks (Baum and Rasch, 1975). IV-28 Table IV-4 TROPICAL CYCLONE PREDICTION MODELS AVAILABLE ON A ROUTINE OPERATIONAL BASIS FOR THE ATLANTIC REGION 1. NHC6 7 — A stepwise screening regression model using predictors derived from current and 24-hour-old 1000, 700, and 500-mb data. It also uses persistence in the early forecast periods. 2. S ANB AR 1 ^ — A filtered barotropic model using input derived from the 1000 to 100-mb pressure-weighted winds. Requires the use of "bogus" data in data-void areas. System modified in 1972 15 so that the wind field near the storm conforms to the current storm motion. 3. HURRAN 1 6 — An analog system using a data base of 703 storms dating back to the year 1886. 4. CLIPER 17 — A stepwise multiple screening regression system using predictors derived from climatology and persistence. 5. NHC72 18 — A modified stepwise multiple screening regression system which combines NHC6 7 and CLIPER into a single model. 6. NHC73 19 — A system similar in concept to NHC72 except that it uses the so-called "perfect-prog" method to introduce numerical prognostic data into the prediction scheme. 7. MOHATT 20 — The Navy's modified HATRACK scheme, which uses a geos trophic steering concept applied to heavily smoothed analyses and prognoses produced by the Fleet Numerical Weather Central, Monterey, California. A "bias" correction is applied to the forecast after observing initial 12-hour performance. IV-29 GEOPOTENTIAL HEIGHT DATA* 24 HRS AGO OBSERVED GEOPOTENTIAL HEIGHT DATA* O uj Z X > o CO U. > ■ a w 3 M S is < * 2 9 1— CM <2 X u z o a • a. is ID a £ a "3 u X en UJ < a. <-> — iu -J cr u o u. o a. INITIAL STORM LOCATION 3 a • t> STORM MOTION 12 HRS AGO a o a Q 2 ? X "3 INITIAL STORM MOTION c a i o s a * ai a i a s a. r*. £ z CC O UL 2 I H cc o o CD > UJ cc 3 a IV-30 G. Storm Surge and Waves Hurricane winds have two major effects on the ocean: (1) increased mean water level, referred to as the "wind tide" or "storm surge" and (2) increased wave length. 1. Storm Surge Generally, surge has three distinct stages: the initial stage, or forerunner; the hurricane stage; and the final stage, or resurgence. 42 The hurricane stage is the sharp rise in the water level that, in the Northern Hemisphere, occurs to the right of the storm track. This rise, which occurs as the hurricane approaches land, lasts for 2 to 5 hours. Figure IV-17 presents a storm surge profile for a hurricane in the Northern Hemisphere. It schematically indicates the height to which the surge will rise at various distances along the coastline on either side of landfall — the point where the center of the hurricane crosses the coastline. The figure applies to a coastal region with a gradually sloping ocean bottom. The deviation of the mean water level from normal depends on: 21 (1) Characteristics of the storm, such as direct wind stress, the atmospheric pressure difference, wave setup, rainfall, storm track, storm movement, and storm size. (2) Topography and basin geometry, such as depth, embayments, estuaries, barrier islands, islets, and length of coastline (3) Initial state of the ocean, such as the tide level. The storm surge depends heavily on offshore topography, with surge highest along shallow continental shelf regions. For a wide, shallow-shelf topography in which the length of the shoreline is large in comparison to the diameter of the hurricane, such as the Gulf of Mexico coast, the surge may be as high as 9 meters (30 feet). 21 With some notable exceptions, the continental shelf regions along the Pacific Mexican coastline are generally steep and less than 20 nautical miles. Two large embayments on the Pacific Ocean side of the Baja peninsula, the Gulf of Tehuantepec, the continental shelf behind the Islas Marias, and IV- 31 1H9I3H 39«nS ao a i 9 3 I < 3 5 Ul S UJ X CL I UJ X Z QC LU X / 3 / 1 • f o < a. O O | '. r °* •:.. 7 < I ffl *■/' - " o •.[ ' * *% x ■'• UJ ^L ■•> ■•-. V H /-.. ...•••^•••••'' 's • -./ M o • V < • : - • '••.. « : r\ o . • w 3 / < Q 3. /" V I -j UJ '/' / >*" s -1 "^ < r I > u 2 3 a < *5 LU CO 3 3 s L 1 < £ z a s \ > < z \ 4 2 • \ »- «/» LU s S O z u o o 3 2 >* 3 < 1 CO Z g a LU s > Q LU r a S %> V-2 The physiographic regions also serve to distinguish between the major population groups, settlement patterns, and economic activities along the Pacific Coast. The rapidly urbanizing, newly settled Northwest region, for example, is experiencing agricultural and economic prosperity. It contrasts with the economic underdevelopment of the two southern regions. A traditional, indigenous culture, a dense rural population — particularly in the South Pacific region — and subsistence agriculture have tended to isolate the two southern regions from the mainstream of economic development in the country. The three regions correspond roughly with the distribution of hurricane landfall points in the eastern Pacific. As Figure IV-9 indicates, the Central Coast region has been exposed to as many as nine storms since 1971, as opposed to only five in the South Pacific region. For the Central Coast region, the relatively high incidence of hurricane activity is tied to the region's location within the genesis area of eastern Pacific storms. In the Northwest region, hurricane incidence is often a result of recurvature. C. Settlement Patterns Since pre-Columbian times, Mexico's population has been concentrated in the intermountain valleys and basins of the south-central regions, where plentiful rain permits subsistence agriculture. In the Northwest, however, rain has been scarce, and on the northern Gulf Coast the tropical climate has bred disease. Thus, the native population was sparse and characteristically nomadic. The Spanish did little to alter this settlement pattern; they merely administered existing centers of indigenous population. In the central regions of the country where Spanish influence was dominant, urban centers were superimposed on indigenous settlements. Minimal Spanish urbanization occurred in the northern and southern regions of the coast. In the arid Northwest , the few towns the Spanish established existed only to consolidate the frontier territory. In the Central Coast region, however, links with the Valley of Mexico have been stronger, largely because of their common physiography. With greater cultural and economic integration, the native settlements of the \ V-3 Central Coast states have tended to disappear. The contrast between a modern, urban society and an indigenous, rural society is most evident in the states in the central region of the country. The 19 30s and 1940s saw the country develop extensive programs of irrigation and road building to integrate less well-developed regions and to tap fallow agricultural land. Under the presidency of Miguel Aleman (1946-52), a national department was set up to manage the nation's scarce water resources. Dams were built along the major rivers of the northern coastal lowland; almost overnight, the Northwest became one of Mexico's most productive agricultural regions and has since undergone the most rapid population growth of any region in the country. Since World War II, all regions of the country have experienced a percentage decline in rural population, which has been moving to the cities and towns. The rate of rural out-migration has been lower in the South Pacific region because no large cities exist there. Urbanization in the Northwest region, on the other hand, has not only been the fastest of all regions, but all political subdivisions of the Northwest are predominantly urban. According to the 1970 census (see Table V-l) , the population of the Northwest region increased by more than 50 percent between 1960 and 1970, with urban population increasing by 87 percent (compared with a national average urban growth of 64 percent). In contrast, the population of the South Pacific increased by only 27 percent over the same period, with an urban population increase of even less — 18 percent. The Central Coast states have stayed close to national averages, maintaining a 10-year overall population increase of 32 percent. The 1976 populations of the Northwest, Central Coast, and South Pacific regions are estimated at 4.9, 9.4, and 4.6 million, respectively, for a total of 19 million inhabitants. See population distribution in Figure V-2. The Northwest has 13 cities with a population of 50,000 or more, as opposed to 8 in the Central Coast and 3 in the South Pacific* If the * See Table V-2, V-4 c * at -o Oi r*. as o> -^ PH it aj i-> CO j-J u as d «» o — a ai -a a as K so u o H !*• < OS <-l —I to O ^ U CO M =) x o e o o u Cs o « a. c o as 8«5 •- p* in p*. *a> -m ao -T vD < vO < o OS —" sO o " * —• »3* r^ 00 tN O J — (~ vC — X r"* \0*T(N — CO N fl 00 -J fsj N M <*1 CO J^ »■« 00 ON« sO — C-J in «£ — iC^IX n fl ^ m ^ ji -* O © o^ ••a* s* in c«4 o r» O o in P*» Os o o m in m ^ sr -J C 1/1 - CO u"i rs. ^o t>4 0\ SO MM ^ v£) oo « An W cm «wt ^ *a» rn -*» n *■ . OS 1* o u 5) •* fl X - £ nj 41 s J= 5 H -O - 2 b- a U 3 -3 ** -3 3 JJ O U U ■ « "3 a -r* .* m c c >> u £. C O c — vO a 0> »< « fl — ^ 51 3 2 f- s o t/I r>» u c> v m a. M — o CSI m m O \£ S o o oo r-- — — O O ST r*» sD as vO o — -. as en — in fs -c CN -" iflO ^ O 30 vT f^ ?1 CO- KINO e — — — C C 5£ o a 3S — — w C " *•* ^ w "J U — 5 - u — 3 is JJ SB u-l — r^ r^ o Ml OS ^J in ^ CT> OS ?n t» w 3 >>© — vO • = as -3 o — V 3 V-5 z c D £ Z o a a. O a. UJ OS 3 a 7-6 Table V-2 THE POPULATION OF MAJOR CITIES ON THE WEST COAST (in thousands) 1976* Rankin gf City Population Northwest Region 2136.4 37 Ensenada, B.C. 111.8 57 La Paz, B.C. 67.5 9 Mexicali, B.C. 348.8 7 Tijuana, B.C. 393.4 35 Tepic, Nay. 117.3 17 Culiacan, Sin. 238.3 48 Los Mochis, Sin. 87.3 28 Mazatlan, Sin. 155.8 29 Ciudad Obregon, Son. 152.9 52 Guaymas, Son. 76.0 14 Hermosillo, Son. 259.0 63 Nogales, Son. 59.4 56 San Luis Rio Colorado, Son. 68.9 Central Coast 2764.8 65 Colima, Col. 53.0 12 Acapulco, Gro. 276.7 62 Iguala, Gro. 60.6 60 Ciudad Guzman, Jal. 62.8 2 Guadalajara, Jal. 1900.4 20 Morelia, Mich. 215.8 36 Uruapan, Mich. 115.7 50 Zamora, Mich. 79.8 South Pacific 278.1 53 Tapachula, Chis. 75.6 49 Tuxtla Gutierrez, Chis. 84.6 34 Oaxaca, Oax. 117.9 Coastal Regional Totals 5179.3 * Estimates. t National ranking based on population. Source: Anuario Estadistico de los Estados Unidos Mexicanos 1968-69, prepared by the Secretaria de Indus tria y Comercio, 1970. V-7 growth rates of those cities is considered, the Northwest has been increasing steadily at a regional rate of 72 percent over 10 years. In the Central region, on the other hand, the variety of its coastal economy is evident in the contrast between the growth rates of Acapulco and Colima — 255 percent versus 12 percent . As opposed to the well- integrated urban growth of the Northwest, where major cities are centers for processing and guiding the economic activity of the hinterland, Acapulco' s spectacular growth has not been equaled in the hinterlands. D. Settlement Pattern Along Coastal Lowlands Because ocean surge and flooding from excessive rain often tend to be the most destructive part of hurricane damage, the settlements in the coastal zone below elevation of 200 meters were examined." From Figure V-3, it is clear that only a small portion of the territory in the study zone lines below this elevation — a situation altogether different from the Mexican and U.S. Gulf Coast. There, the coastal lowlands extend for hundreds of miles inland in places. The Pacific coast, in contast, has only a narrow zone of coastal lowlands. The population in this narrow strip totals on the order of 5 million inhabitants (see Table V-3) . More than 60 percent of this population is in the Northwest region, where the coastal lowlands extend farthest inland and are most intensively farmed. Within the lowlands of the Northern region, 63 percent of the population lives in the 29 cities of more than 10,000 inhabitants. The rural lowlands population in the region, which comprises a third of the area total, is grouped in nearly 7000 settlements. The profusion of small settlements along the coast is particularly notable in the Territory of Baja California; there, a rural population of nearly 60,000 is scattered among 1500 settlements, averaging 40 people per settlement. This population dispersion makes hurricane * The survey does not, in fact, strictly coincide with the 200-meter elevation. Certain areas such as the entire peninsula of Baja California, were included because of their importance to hurricane damage evaluation. V-8 < > UJ tr UJ H LU o o CM 5 o _l UJ S3 in < UJ cc < 01 > UJ 5 a 7-9 3 i in N •» W> 9> -NO •& -* m -• — a = v =-a — a 3 — °k 9. z < a _: - 3 "5 i 5 31 ■J -~i 51 -s z ** 2 2 Si = a -i ;n 1— 3> l"1 ^ « 5 s X X n 3 O r^ ?-* *» j ^ !V| ^ f* rx _ — ~T ^> >T •c X — i »T 3 J- r> -T — x o IN* ^r 3v n n •g ! X -C — >T in :n — ii _ ^•1 " 5 2 Ml m a. i r*i r-» — • r* o>> »T in — 3> ^ f* in ^» ^ rsi ^ so .e iT «» 3* 3 3>i !• — " X in "■ P* X C ^ ^ N O ^ r* -C ~x >y — r*. >n O in <0 a « "J 14 a a 3 a 5 _ » r"5 2 < c li. o 2 a 5/5 -r' Ji I 3 — Si — >C\ -> >< a — ai HI f^. s0 ^ rt z? *0 r^> o *D X 3 O c?i in 3* n a X >r — ^ f"n »y OL-im x » m •» a ?>:n -j .c — in — xr-j^ cm ^ jo r^ in X r^ r-» in n ^y p ^- as — — >» — a -j k a o> »• x e»4 ■— a <"■ »> ^T — 3« (M X <■ rM ^ - x s fsj — X X :vi ^ io it ji r» — r-4 X ■~* ^ .-N — x a n x a m i*- ^ ? — a o in i" ^ u a n -J .3 ■a a a JS — ^ s a w a "; J a — u w s I a u V a W a *~ *j •* — — - a — a ■a in cx*"* g - / " x >^CC^-~ — X — ^ a X 5 — O — N N ^ r- a x ^ vT ^ a a xa X v. y. r-io warning and evacuation procedures particularly difficult; on the other hand, it argues against really massive disasters. The population density of the South Pacific coastal zone is 25 inhabitants per square kilometer, the Central Coast's is 19, and the Northwest's is 12. In the Northwest region, the coastal zone that has the highest hurricane frequency is Nayarit, which has 30 inhabitants per square kilometer, one of the highest densities on the Mexican Pacific coast. In the Central Coast, the coastal lowlands of Colima and Guerrero have densities above 30; whereas those of Jalisco and Michoacan have densities of less than 15 people per square kilometer. Furthermore, although the coasts of Guerrero and Colima both have high densities, their settlement patterns differ. The majority of Colima 's population (52 percent) lives in coastal cities of more than 10,000 people, and its housing is predominantly of brick with tile — and sometimes palm — roofs (see Housing Characteristics below). In Guerrero 54 percent of the population lives in settlements of less than 2500 people (Acapulco is the only city of more than 10,000 people). E. Population Characteristics When statistics on the labor force of the coastal regions are examined (see Table V-4) , further differences appear. For example, the northwest is economically diverse: In the most northern border states, urban working and management classes predominate, accounting for a little more than 50 percent of the total labor force; farther south in the same region, farmhands account for more than 30 percent of the population in Sinaloa and Nayarit. The relatively high use of farmhands in these two states, compared with the low use in the more arid state of Baja California, results from the difference in agricultural practices within the Northwest. V-ll -I 3 xj C «"> •» — '"» © X W *rt Cf C* n* i/> © u-\ © >•» .» 9» © — 3 ! »j sC >o — *- .e s *» "> -a © p- <■" as tfl © — « c x « » — « O — ^ o w »M •» »«•»» a OS «* 2? 1 = «- I Si I I: ^- *- < o> — m >T ~J x o> Mi ./", ON ~* — © © ^» O o " x */\ 30 -c lA o X O O <— 1 m X x Ml — ~ "™ "™ ■" IM ■^ (Nl ~ pm >» -» p* 1! , ?*i -I 3> o ^ ^4 » n 3> c r* j 51 r* SO X o* CO 3- © <-\ -T O X - •a 9> -^ ~j >T .£ t~- X fJ*. £ J> o ■Ji 'A m •^ -» r* -■ S z si ■ « , ol = fl ; a! i J-ji! *3- 3 ■o n © — o 9 X p~ © ul S © f« © 3 X p^ *r, (N © p* 3", »/■> — pm «» (N 4rt 3\ N N I" - Cf» m M !N O — o a •» x o» ^ o> N ^ fl (N f N < !S ■^ P^ *ff PM X O m !J §1 w2| fl 3, g s ■M k. o -• — r* 3\ x X *T V£ 9> C 50 ! r» X IX o <7» r* x * «j N l/> in « ipi x n x x o> n X 5'** as! I s s © c a © © U" •• l^ m o »no»^"" — X <~ X r» -» ^ © X ^ ^jr r*. m in — (N r^ — «flF rt x c * — © — _ x \o — vo » -a ol : c s ■— a X - 2 M ^ s i X - X U1 c Z «j ^ o j I .... 1,3 .9 b V-12 Although most agriculture in the Northwest employs modern technology and commercial outlets, the wheat and cotton farming of the North is more highly mechanized than that in the southern part of the region. There, the principal crops of tomatoes, watermelon, and rice require more intensive farm labor. Northern farming is more concentrated in modern latifundios , or plantations; in the southern half smaller plots are more common, collectivized. The South Pacific region contrasts sharply with the northern regions. A small part- of the population is employed as management and working class. Self-employment accounts for 40 percent and collective farming for 14 percent; added to the 24 percent employed as farm hands, the traditional subsistence sector adds up to nearly three-quarters of the labor force. Finally, if the work force is divided into economic sectors, the economic structures of the three regions are shown to differ again. The economic diversity of the Northwest, for example, is unsurpassed elsewhere along the coast: Baja California surpasses the national average in four sectors — manufacturing, construction, commerce, and services. The other states of the region, although diversified, lack the economic breadth of Baja California. This is especially true with regard to the manufacturing sector, in which every coastal state, except for Baja and Jalisco, falls substantially below the national average. Mexico, like most other Latin American countries, has its industry concentrated in two or three major metropolitan centers. Industrialization in the more peripheral, coastal states has been largely limited to processing raw materials. Thus, aside from the two border states of Baja California and Sonora, and the Central Coast state of Jalisco — whose capital, Guadalajara, is Mexico's second largest city — the coastal states employment is concentrated in farming, forestry, and fishing (see Table V-5). F. Housing Characteristics Given modern housing technology, it is generally assumed that coastal urban areas will be safer during a storm because they are structurally V-13 1 c u u 1) 5 > ■11 t= « en o .J >» ej«l mt a\ CM en en CM lA CM CM DM «-c r*» O P"l -* «» IM * - » nT 00 i-* CO ON u 11 -3 1 c/1 3 •w c a C c 7) 3 -H a w CO l en in 01 4) o U e k- a 1) c i T- O O ^ CM *0 O 0"! cH CM rt in -4 iv ct\ cn 0\ %y in N *j OB » m His en o QO -a* m en Ifl N -^ fl ^ ooddo •* cm en >T odod J 0. «1 > o cu SO fll °" If) rv CO c* « u i * 3 E - ^ U o CU 3 a. C/1 c ■H cd ^»- H < s Z < o M X W r § B 6- e/i U 2 i 3 u e *■> o 71 «5 S > e O0 U — I C « — -* 01 «) a n - CO o u. u. >a a o 00 mi •m c u CO -4 CO y .* m O V- 3 e- o ej so —i cm r» O r* en -O m3 cm -t ii-i *T en sO en 1 >, u 00 u «1 C W r». .0 en -» m SO ON as vO cm en 00 ■o vr 3 Id 3 •-t oa co co en o in CTN CO CM mi r» m cjn S 3 -O mi mi P* m> CM .H 10 u C r M m oo en in co O en o O mi *3 Ol I'l cm m d d O m ■c ej> r-N r** oo oo en oo o o o o o o o o o o o o o o o o o o a o o o o S s o o o o o o o o o * H CO CO c -* ml »H a C U CO e ao w u ■J 14M *- 71 -4 as 2 ~ C3 3 fl lis SO w ■■3 T w CO u u B 'J '^ V) 3ti V) CJ u — * CO s «J ^ ™ (0 T C3 J u u w 7) 3- 1 fl q q "3 9 r; — u — ^ rg CO CO _^ — 1 i««i >l g 3 u — H J _ ■J m -•- >c y fj CO -* a ", U 71 u 2S =3 Z cyi «rt C g » c 3 O o CO G Z "J 5/1 3 u n c V-14 sounder than similarly located rural settlements: U.S. statistics indicate that recent hurricanes have caused less loss of life than those occurring at the turn of the century; it is generally concluded, therefore, that increased modernization (for example, urbanization) leads to increased safety. The conclusion is only partly true, however. Although loss of life has been reduced in the United States, damages to physical structures, measured in monetary terms, have skyrocketed. Furthermore, the fewer lives lost probably result more from improved hurricane warning system than from improved physical structures. In Mexico, these factors must be kept in mind when evaluating a coastal region's ability to resist hurricanes. Although rural housing is more easily destroyed by severe storms than modern urban housing, rural housing is often constructed of materials — wood, palms, adobe, earth — that are more easily and rapidly replaced than those of urban housing. Furthermore, some of the rural material is considerably less lethal under severe wind speeds than the brick, concrete, or sheetmetal often used in cities. An exception, however, is the common use of tile roofs in many rural areas of the Mexican coast; tiles are particularly deadly under hurricane conditions. The large concentration of poor settlements in Mexico's coastal cities is one of the greatest sources of potential and actual damage from hurricanes. A survey of Mexico's major coastal cities shows that between 60 and 70 percent of their population resides in high-density, structurally poor housing, with poor drainage and poorly tied-down housing materials. Note, however, that in many of the states along the coast, more than half of the lowland population lives in settlements of less than 2500 (Table V-3). The statistics on the distribution of roof and housing materials in the Central Coast (i.e. , the region most susceptible to hurricanes) indicates, with the exception of Colima, a high incidence of adobe construction with tile roofs (see Tables V-6 and V-7) . Another common material in the same region is brick, particularly in the state of Colima. V-15 Z -: rj <--i r>* — * — ( r-i — rsi *-** M 49 •* «"t *•* «Q r«fc »* — OV — -«© — f">e»J co — 1 o M »rt is n irt >* n n »n c* -* ^ *& r* © «i *rt c* — . — w n *^ oo -r -j" co co o n in m -• r*» J*t Ov *-* O n in r«* m m *» OO n©^©i/iOOO-J© *© © - — CO o o o o — 4 -* — V -< V-16 3 o z o I— I H u =S H en z 8 < Q z < o z M ir> =3 3 c as K M K ** H 8-* W jse »< X *-! »■* k X M a> «T m -3- r- «o irt ch r^ vo -3* «-* as CM r^ CM f» mm u ^3- CM so en •-• ao 1T> CD CM CO ^3 sO r-i CO CM sO en cj *0 1^. CO »T s© u-1 — • in -3 rn sO O »■* sO m -» jj on -3- s© en en CD r-l -a- O sr in _ r^ en -e w co N^nco>i-'iO^ s© on ** o> in en en on so — >© eM en on en ■c — < — < — < CM — i p» en •«a" O >s ■ U mt s e vO »* en -j ON -3- p-t O m «* en u •& CM ■ • w u SI »s! sv; M ?--' s-e j-« »* »< »< »«« »< »«! »si « »s« M O vO . o CM vl m so — cm en \0 r*. CM r-« so m •j o > ~-« -^ •-* co r~ m* M CM CM CM -" ~*t mm SO r^ vO O «i ON ~T SO in sO 00 CM mt r» _T o !-^ -3- -J3 CM ~3 o in 3 so ■* mm CM >. ^* ^4 •ST -—4 en •— « ^T •-H m m4 sO o ON mm sO uO en CM m* «sT _«h ^ • u m a. « M i-S S-! »4 K « ^f M ** fr« i< »4 is! »>« »< « M e. CM r-j lA — * o O CM co . «-* u-> r» CO CO en sO o O O ON CO m OX vfi en (TV ao ON C-d m mm O « 5^ *-? *e »>! {«« S>? J-S »>« M JsJ s-s »< »4 X a sO O — * CD CN ^c" CM r— « ^~ PN O — ^^ co O 30 en CM O eo sT ■3 1 H en © * H ~" l/~l ■J3 •3- O co -3- ^3 CM en ^a* ON m a m on co m ON I H O H I I C aj 4J X JsS 8-« »-t sO O sO O ON O sO O o o o o m o -3- O r-» -« ON — • co — ■ Psl -« in m m mm CO sO in CM CM sO m* Sst SsS »>! O © O r~ O o m o o o H fs f. vO rt so in O co CM —I is! M )s> M K X i^-o oo^o-joono mo soo -romocnoeno -< m o ON en O CO O CO o ' o on o o o ■mm en mm O -" lO sO o « a u s •s a C o M 01 cs -J e u (3 -i -n ^-i >x 9 <« c O o u U o y (9 sU 'J <9 — 1 E U 0) o a ■H u •*H rt u pH CD <-H u w O 3 <8 ^« c sj U -1 2 u o c o H — i a 5 30 01 as 2 1 12 V-17 In Acapulco, brick and concrete construction dominates in about one- third of the housing; the category of "other," or makeshift, material is also common, and is used in about one-quarter of the city's roofs. Wooden houses with palm roofs are common in some states, particulary in southern Baja California and the two states of the South Pacific region. In the former region, 43 percent of the houses have palm roofs. This is evidence not only of the abundance of palms in the territory, but also of the frequency with which hurricanes strike the southern part of the peninsula. In the southern half of the Northwest region, where hurricanes also strike with some frequency, housing materials do not differ significantly from the national average (i.e., about one-third adobe, one-half brick and concrete, and the remainder divided between wood, earth fill, and makeshift materials). The frequent use of palms, tiles, and makeshift materials for roof construction does differ from the national norm. The unusually high concentration of makeshift roof materials throughout the Northwest region — as much as 33 percent in Sonora — seems based on the region's rapid influx of population, A large, in-migrant population, in an arid/semiarid , urbanizing region has to depend a great deal on makeshift materials for housing. If the coastline of Sinaloa is scrutinized, urban and rural housing can be distinguished. Of Mazatlan's 30,000 houses, more than 50 percent are made of brick or concrete, 20 percent of adobe, and 15 percent of wood. Rosario's 6900 houses are comprised of 35 to 40 percent brick and concrete, but adobe and earth fill have an almost equal share. In Mazatlan, tile and adobe roofs are the rule, but in Rosario palm roofs are used in more than a third of the housing. G. Economic Activity In 1975, at least half of Mexico's agricultural output came from the Pacific coast states, including 70 percent of the country's wheat and fish products, and approximately 60 percent of cotton, coffee, and rice (see Table V-8) . V-18 «n CM 1 _ e» o o 1 r. O I un •—• en vO en ej» -* -2 o o \o O 1 1 1 m I -a* i I O en 00 O 00 r>- Ov o» m en eM o» en vo «ar in co 00 O 00 © O c CO *-4 o -• o o u o p-4 ~* o o ■H O aj u • • 0) o> ■■a -J 00 oo so A • co u-i en r-» ON ■* 3 ■u r~ as »■* en 4 o» eM *-4 m ill III vO j co do at r» vO o m ■* en o e oo ex p4 en U-I SI o to u .o US o vO vO o o o o ■J- 1 1 •* 1 1 s» 1 eM eM I o 1 1 1 ~* 1 — 1 * 1 1 1 1 1 1 >/■> m o ■* un o O 1 1 1 o ^r vO -J- vO 1 1 1 1 -o 1 1 .— s id c u o to 4-1 60 3 • VI AJ 11 6 V- ' CO o o o o u-i m -I O I oo en i I O ej» I I en ox I >»• O i"» so -» vO l O vo r-. I eM as eM I u-i •* en en n4 , eM i (•». I eM V4 ■s s o I s •rf "H 01 o — 1 £ E 60 o 4J V %4 <4M O at! •— 1 t— 1 4J fl s C -^ M it n] tg ¥< 4-t IH oa oa Z en 19 o z z (0 -l u H 3 V-19 In 1945, Mexico imported nearly 20 percent of its corn and wheat. Since then, the Mexican government, in an intensive effort, has increased national production of these basic food goods. Between 1950 and 1970, production of wheat increased from 300,000 tons to more than 2.6 million tons, with a quadrupling of yields per hectare. The new wheat technology requires a continuous, assured supply of water; therefore, wheat production has been limited to areas of extensive irrigation, such as in the states of Sonora and Sinaloa. As Table V-8 indicates, almost the entire pro- duction of wheat along the Pacific coast comes from the Northwest region. Wheat is one example of the differences between the irrigated agriculture of the Northwest and the rain fed agriculture of the more southern regions; rice, cotton, and soybeans have also benefited from the irrigation water of the Northwest. The Northwest dominates agricultural production along the Pacific coast. Until 1973, more than half of Mexico's export income came from agricultural crops; of these, the most important are cotton, sugar cane, coffee, and the winter crops, such as tomatoes, strawberries, and melons. With the exception of coffee, which is grown in the southern states of Guerrero, Oaxaca, and Quiapas , these export crops are mostly grown in the Northwest. For example, cotton, which is Mexico's most important export product, has been concentrated in Baja California and Sonora; the irrigation systems of the Northwest, plus the region's contact with the United States, led to investments by large U.S. cotton investors for the cultivation of soybeans. Sugar cane is grown mostly along the coast of Sinaloa and Nayarit , as well as in the interior of Jalisco, Michoacan, and Oaxaca. Sugar refineries are common in these areas. Winter crops are almost exclusively grown in the irrigated farms of Sinaloa and Nayarit — strawberries in the latter; watermelons and tomatoes in the former. In summary, because Mexico's largest share of commercial agriculture comes from the Northwest region, the region is strategically important in the national economy. However, because so much of this agriculture V-20 lies on the irrigated areas of the coastal lowlands, the danger of flooding from severe storms is substantial. Typically, when hurricanes strike the Northwest states, the most severe economic damage is to crops nearing harvest that are destroyed by the flooding of rivers. Because industry in the coastal region is largely limited to the processing of raw materials, its distribution closely follows that of agricultural production. Thus, sugar mills and liquor refineries are concentrated in the coastal areas of Sinaloa and Nayarit, where sugar cane is produced. Similarly, industries that specialize in packaging of fruits and vegetables are almost exclusively located in the Northwest states, especially Sinaloa. Oil and soap industries are also concentrated in the Northwest coastal plains — as well as in Colima — where a great deal of sesame is grown. The tobacco industry is located in Nayarit, where the majority of Mexico's tobacco is produced. Raw materials such as minerals have attracted industry into the southern regions as well. Iron ore, copper, and manganese plants are common, for example, in the mountainous coastlines of Guerrero, Michoacan, and Colima. And salt processing plants are distributed along the Pacific coast, including many of the islands in the Gulf of California. Most industry on the Pacific coast, however, is concentrated in the Northwest where commercialized agriculture can serve as an economic base for industrial expansion. Mexico has extensive maritime resources, but its population eats little fish. Fishing off the Pacific coast has considerable potential for expansion. After the government's extension of its seas sovereignty limits to 200 miles in 1976, Mexico has been faced with the dual challenge of expanding its fishing fleet and processing facilities, and developing a larger domestic market for fish. In this regard, the large investment of 1.58 billion pesos in the Integrated Fisheries Program is aimed at relieving Mexico's dependence on foreign fleets. The construction of 323 craft, including 120 for shrimp, and others for red snapper and other common species is taking place. The V-21 program is expected to create 5500 new jobs. There are also plans to set up 63 distribution outlets for frozen fish; 12 will be located in the Federal District, with the rest throughout the country. During the 1960s and early 1970s, small coastal cities and towns grew rapidly as tourist centers, and the Pacific coastline from Acapulco northward became known as the Mexican Riviera. Acapulco is the largest of these new urban resorts. Little more than a village until after World War II, by 1970 its population had reached 174,000. Its numerous hotels cater to persons of modest as well as lavish income, and it is the first choice of middle-class Mexican vacationers as well as many world travelers . Some 500 miles to the northwest of Acapulco, Puerto Vallarta is a smaller but even faster growing tourist center. It is a primary vacation spot for residents of Guadalajara. Farther north still, the small cities of Mazatlan and Guaymas are growing rapidly in response to the increasing tourist trade, and in various coastal localities entire new tourist centers are planned or under construction. On the Baja California Peninsula alone, in 19 74, the Mexican government had plans to dredge 5 deepwater ports, install 17 desalinization plants, and pave airport runways to accommodate visitors to 26 new tourist centers. Over the past 10 years, foreign tourists have spent 200 billion pesos in Mexico, and this has kept the Mexican balance of payments in equilibrium, at the same time supplying funds for further development. In Acapulco, public investment totals 800 million pesos, including money for the construction of the monumental Convention Center. Some 770 million pesos have been spent on installations at Cancun, on the Carribean coast (more than 120 million pesos on the airport alone) . In the new Ixtapa-Zihuatanejo resort, on the Pacific coast, almost 500 million pesos have been earmarked for the international airport. Similar investments were made in Mazatlan, Cozumel, Manzanillo, Puerto Vallarta, Cabo San Lucas, and other tourist centers. The Baja California Transpeninsular Highway has been completed, opening new areas to tourism. A ferry service links the peninsula of Baja California with the Bahia de Banderas , on the coast of Nayarit. V-22 H. Governmental, Legal, and Institutional Structures 1. Status of Entities that Lie Within the Experimental Area Storm tracks in the eastern Pacific area of interest cross Clipperton Island and Mexico's islands, the Baja California peninsula, and the Mexican mainland. Clipperton Island is a small undeveloped and uninhabited island at 10°20' north latitude and 109°13' west longitude; it was awarded to France in 1935 in international arbitration, and appears to have neither persons nor property subject to storm damage. Mexico is a populous nation whose governmental structure is that of a federal republic, much like the structure of the U.S. government Thirty-one states surround the federal district where Mexico City is located. The present Constitution of Mexico was adopted in 1917, and has been amended many times . The federal government of Mexico has an elective president and congress, and a three-level court system. Each state has its own constitution, elective government, and court. Mexico has claimed jurisdiction over territorial waters to a distance of 12 nautical miles, and has announced "patrimonial" rights (principally fishing* and seabed rights) to a distance of 200 nautical miles. In 1960, the Mexican Constitution was amended to state that national boundaries include reefs, shoals, and submarine plateaus in the ocean surrounding offshore islands, the continental shelf, and the airspace over the country. Mexico maintains a relatively independent hemispheric and international foreign policy position. Mexico's relationships with the United States in particular have been positive, although they reflect a respectful caution. In the United Nations, Mexico has refused a seat on the Security Council because of its opposition to intervention in the affairs of sovereign states. See, for example, Ley de Zona Exclusiva de Pesca de la Nacion, Diario Official of January 20, 1967. V-2 3 2. Relevant U.S. Domestic Legislation a. National Environmental Protection Act of 1969 This law requires all agencies of the Federal government to file an environmental impact statement for "...major federal actions 2 significantly affecting the quality of the human environment." Federal guidelines issued by the Council on Environmental Quality pursuant to the Act specifically designate weather modification activities as coming within the Act. The guidelines particularly mention weather modification activities conducted by the National Oceanic and Atmospheric Administra- 3 tion of the Department of Commerce. The regulation requires subject agencies, inter alia, "to explore alternative actions that will avoid or minimize adverse impacts and to evaluate both the long- and short- range implications of proposed actions to man, his physical and social 4 surroundings, and to nature." A related rule vis-a-vis international operations requires the National Security Council's Under Secretaries Committee (now the Ad Hoc Group on Weather Modification) to review "the international aspects of weather modification generally" and all weather modification activities "affecting other countries or conducted outside U.S. territory." b. Weather Modification Reporting Act This federal legislation provides for the reporting of nonfederal weather modification activities to the Secretary of Commerce. By interagency agreement, the Secretary has secured promises from other federal agencies that they too will comply with the registration mechanism. Although it may not seem sensible to list the statute in that NOAA would, in effect, simply be reporting to itself, the underlying purpose as spelled out in implementing regulations would seem to be applicable to NOAA. Of particular interest is the requirement for the Secretary to notify state officials where a weather modification activity entails possible danger to persons, property, or the environment . V-24 3. Relevant Legislation of Mexico The principal and authoritative source of Mexican law is the Civil Code for the Federal District and Territories of Mexico. It has served as a model for the civil codes of the individual states. 4. Relevant Treaties a. United Nations Charter 8 The Charter is a backdrop for the general workings of international law. b. International Court of Justice This is the judicial arm of the United Nations, and is the progeny of the Permanent Court of International Justice associated with the League of Nations. The Court, which sits at the Hague in the Nether- lands, is open to all members of the United Nations (and other states fulfilling certain conditions), but it is not open to private individuals The Court has jurisdiction over all matters which the parties may refer to it, and over matters specifically provided for in the U.N. Charter or in treaties or conventions in force. Should there be a dispute as to the Court's jurisdiction, the Court itself decides the issue. The Court could, for example, hear a case brought before it by Mexico on behalf of its citizens that claims injury to persons and property resulting from hurricane damage when the hurricane was one seeded as part of the Hurricane Amelioration Research Project. If the U.S. Government believes that the matter falls within the declaration whereby it acceded to the compulsory jurisdiction of the Court, it would participate in the Court's deliberations. c. Convention of the Safety of Life at Sea The United States is signatory to this Convention which seeks to promote the safety of life at sea. The following are included as obligations of the contracting parties: The obligation to warn ships of gales and tropical storms: and V-25 The obligation to issue daily, by radio, weather bulletins suitable for shipping, containing data of existing weather conditions and forecasts. The relevance of this convention for the Hurricane Amelioration Research Project is the fact that, even if seeded hurricanes do not cause untoward effects on land masses, they may temporarily alter the storms to some degree as they are seen by ships plying sea routes in the experimental area. Since the expected effects of seeding, as viewed from the surface, would not be distinguishable from the variability normally encountered in these storms, the existing procedure for warning ships of gales and tropical storms should satisfy the obligation under the Convention. d . Convention of the Intergovernmental Maritime Consultative Organization * 3 The relevance of the convention is similar to that of the above. One purpose of the organization is "to encourage the general adoption of the highest practicable standards in matters concerning maritime safety and efficiency of navigation." ** Again, it would appear that having undertaken this obligation, it would be appropriate for the United States to consider possible impacts the experiment might have on shipping which traverses the area. The Intergovernmental Maritime Consultative Organization itself would seem to be an appropriate place for such matters to be considered. It sees itself as standing ready to advise on "...any matters concerning shipping that may be referred to it by an organ or specialized agency of the United Nations." One such specialized agency which could make a referral is the World Meteorological Organization, which is well aware of the past efforts of Project STORMFURY and the intentions of the Hurricane Amelioration Research Project. e. Convention of the World Meteorological Organization 6 The United States is an active participant in the World Meteorological Organization, one of whose goals is "to further the application of meteorology to aviation, shipping, water problems, agriculture, and other human affairs." 17 This organization, which was V-26 formed in 1947, has been innovative in coordinating worldwide weather forecasting efforts. Because it is aware of the progress and problems associated with weather modification, it would be "...a logical candidate among existing bodies for coordinating (international) weather modification efforts." 18 The organization's interest in the field is reflected in the following statement contained in one of its reports: Before undertaking an experiment on large-scale weather modi- fication, the possible and desirable consequences must be carefully evaluated and satisfactory international arrange- ments must be reached. 19 It should be noted, however, that no formal mechanism presently exists within the World Meteorological Organization which would require reporting and coordination of hurricane seeding experiments with the international community. Nevertheless, it is the intent of the U.S. Government to notify the WMO, at an appropriate time, of its plans to conduct the Hurricane Amelioration Research Project experiments. f . Convention on the High Seas 2 The United States is also a signatory to this Convention, Article 2 of which states the following: The high seas being open to all nations, no State may validly purport to subject any part of them to its sovereignty. Free- dom of the high seas is exercised under the conditions laid down by these articles and by the other rules of international law. It comprises, inter alia, both for coastal and non- coastal States: (1) Freedom of navigation (2) Freedom of fishing (3) Freedom to lay submarine cables and pipelines (4) Freedom to fly over the high seas These freedoms, and others which are recognized by the general principles of international law, shall be exercised by all States with reasonable regard to the interests of other States in their exercise of freedom of the high seas. 21 V-27 In conducting seeding flights over the high seas as part of the Hurricane Amelioration Research Project, the United States will be exercising its "freedom of the high seas." As party to the above Convention, the United States has obligated itself not to conduct its activities in a way that negatively affects the freedom of movement of ships and aircraft of other nations in the area of the experiment. V-28 VI EXPECTED RESULTS OF EXPERIMENT A. General The two direct results of the proposed experiment will be the dispersal of silver iodide into the atmosphere and any resulting changes in the hurricane. Appendix A gives estimates of the upper limits on concentrations of silver iodide. They are generally orders of magnitude below maximum acceptable health and safety levels. This section presents quantitative estimates of the potential hurricane changes caused by seeding based on existing data and indicates the major un- certainties involved. Because no hurricane modification experiments have been conducted in the Pacific, the discussions are based to a large extent on the results of Project STORMFURY experiments in the Atlantic area. The results of simulated seeding in a theoretical model of a hurricane have also been considered. B. Background 1. Previous Seeding Experiments The results of hurricane seeding to date are summarized in Table VI-1, below, for those experiments in which the seeding was performed near the eyewall. (In addition, a hurricane was seeded on October 13, 1947, and Hurricane Ginger was seeded September 26 and 28, 1971. The clouds seeded in these storms were far different and the seedings were done in a different fashion from that for the storms listed in the table; see discussions below.) The natural variability of hurricanes imposes great obstacles to evaluation of results of such modification experiments. Although the results of the first experiment on Hurricane Esther (1961) and the second experiment on Hurricane Beulah (1963) were in the desired direction (decrease in storm intensity) , the magnitudes of the changes were the same VI-1 Table VI-1 RESULTS OF EXPERIMENTS IN SEEDING HURRICANE CLOUDS NEAR THE EYEWALL Name Hurricane Esther Hurricane Esther Hurricane Beulah Hurricane Beulah Hurricane Debbie Hurricane Debbie Silver Iodide App rox. Max. No . of Used* Wind Speed Date See dings 1 (No. /kg) Ch ange (%) Sept . 16, 1961 8/35.13 -10 Sept . 17, 1961 1 8/35.13 + Aug. 23, 1963 1 55/219.96 + Aug. 24, 1963 1 67/235.03 -14 Aug. 18, 1969 5 976/185.44 -30 Aug. 20, 1969 5 978/185.82 -15 Values in column are for total number of units and total kilograms of silver iodide used each day. Test results indicate the smaller seeding pyrotechnic units make more efficient use of the silver iodide. f Pyrotechnics dropped outside seedable clouds. Source: Reference 1. as frequently observed in unmodified hurricanes. The more massive and more frequent seedings of Hurricane Debbie (1969) were followed by greater changes in the maximum windspeed; however, even these were no larger than those sometimes occurring naturally in hurricanes. Also, there have been so few cases of seeding experiments that it is not possible to perform evaluations having statistical significance. Because of this, there has been considerable emphasis on the development of theoretical models of change during modification experiments and to determine how much and in what sequence they should change. The above considerations, as well as experience, suggest the desirability of designing evaluation procedures which use a combination of physical and statistical techniques. These procedures would make it possible to determine, with relatively few cases, whether or not seeding will modify hurricanes. VI-2 The Hurricane Amelioration Research Project has been set up as a scientific procedure with a view toward evaluating and documenting the entire sequence of events rather than just seeding the storm and then making a final measurement to determine whether the intensity changed. This procedure determines whether or not each link in the hypothesized chain of events actually occurs. Figure VI-1 indicates the hypothesized sequence of events, as well as the type of data collected before, during, and after seeding to determine the validity of each hypothesis. Step 5, for example, states that the maximum windspeed is reduced because of decreased temperature and pressure gradients. In order to test the validity of this event, windspeed measurements are made continuously for a 24- to 30-hour period. Monitoring of windspeeds is conducted nearly continuously at the 5000-foot level, occasionally at the lowest levels, and frequently at the middle and upper levels. The correlations of these windspeeds with altitude are estimated and the temperature, water vapor, and pressures are measured frequently. 2. Data Analysis Method Table VI-2 lists the methods of data analysis that will be used to determine the effect of seeding on hurricanes. Besides making direct comparisons of the data collected before, during, and after seeding, the temporal and spatial response characteristics for the basic meteorological elements are determined by using the "variational optimization approach." 3 It is expected that this not only helps establish trends before, during, and after seeding, but also estimates the contributions from the natural events. The digitized radar reflectivities can be used to determine rainfall rates for the seeded and adjacent (unseeded) cloud areas. The observations are supplemented by comparing numerical model predictions for seeded and unseeded cases. The above analysis is reinforced by cause-to- effect -physical analysis procedures. For example, synoptic satellite and aircraft data are used to determine the interaction between the hurricane and the synoptic scale environment. VI-3 SEED CLOUDS OUTWARD FROM THE EYEWALL SUPERCOOLED WATER FREEZES IN SEEDED CLOUDS LATENT HEAT OF FUSION RELEASED BUOYANCY OF UPPER PORTION OF SEEDED CLOUDS INCREASES 1. Upstream, downstream, and seeded clouds sampled for particle state (middle level) . 2. Microwave measurements of total integrated liquid water above the aircraft (middle level) . 3. Vertical velocities (vertical mass flux), temperatures, water vapor, and pressures measured (low -middle -upper levels). 4. Radar reflectivities measured in eyewall, and seeded and adjacent cloud areas (low-middle levels) . TRANSVERSE CIRCULATION ALTERED MAJOR VERTICAL MASS TRANSPORT OCCURS AT LARGER RADII THAN BEFORE SEEDING 1. Mass inflow measured (low level). 2. Mass outflow measured (upper level). 3. Vertical velocities (vertical mass flux), temperature, water vapor and pressures measured (low -middle -upper levels). 4. RHI and PPI radar reflectivities measured (low-middle levels). OLD EYEWALL CIRCULATION WEAKENS AS VERTICAL MASS TRANSPORT IS CONCENTRATED IN SEEDED CLOUDS SUBSIDENCE IN EYE DECREASES 1. RHI and PPI radar reflectivities measured for eyewall, and seeded and adjacent clouds nearly continuously for the period of the experiment (30 hours) (low-middle-level). 2. Vertical velocities (vertical mass flux), temperatures, water vapor, pressures measured (low -middle -upper levels). FIGURE VI- I. DATA COLLECTED BEFORE, DURING, AND AFTER SEEDING TO DETERMINE THE VALIDITY OF THE HYPOTHESIZED SEQUENCE OF EVENTS VI-4 MAXIMUM WIND SPEEDS REDUCED DUE TO DECREASED TEMPERATURE AND PRESSURE GRADIENTS 1. Wind speeds measured nearly continuously for 24-30 hour period of. experiment at 5000 ft. 2. Wind speeds measured occasionally at 1000-1500 ft. 3. Wind speeds measured frequently (middle-upper levels) (correlation with height) . 4. Temperatures, water vapor, and pressures measured frequently (low- middle-upper levels). PRESSURE FIELD ADJUSTS TO WIND AND TEMPERATURE FIELDS 1. Wind speeds, vertical velocities, temperatures, pressures, and water vapor measured (low -middle -upper levels) . STORM STARTS TO RETURN TO ITS NATURAL STATE 6-18 HOURS AFTER FINAL SEEDING 1. Wind speeds, vertical velocities, temperatures, water vapor, and pressure measured (low -middle -upper levels) . 2. Mass inflow measured (low level). 3. Mass outflow measured (upper level). 4. RHI and PPI radar reflectivities measured for eyewall and principal rainbands (low-middle levels) . 5. Principal clouds sampled for particle state (middle level). 6. Total integrated liquid water content above aircraft measured using microwave sensors (middle level) . Source: Reference 2 FIGURE VI- I. (Concluded) VI -5- Table VI -2 METHODS OF DATA ANALYSIS I. Direct comparison of data collected before, during, and after seeding A. Windspeeds, temperature, pressures, water vapor B. Mass transports (vertical-horizontal) C. Cloud particle state (seeded-unseeded) D. Radar reflectivities (seeded-unseeded) 1. Cloud heights 2. Distributions E. Synoptic scale. II. Variational optimization — filter the data to determine the temporal and spatial response characteristics for the basic meteorological elements for selected scales of motion (cumulonimbus, rainband, eyewall, and typhoon) A. Establish trends before, during, and after seeding 1. Estimate contributions from natural events B. Determine magnitude of response for each selected scale of motion 1. Sequence and timing of changes 2. Correlate responses to changing radar structure C. Determine character of signals before, during, and after seeding. III. Digitization of radar reflectivities A. Determine rainfall rates for seeded and adjacent areas B. Integrate (space) total rainfall over the storm C. Integrate (time) total rainfall for selected areas. IV. Comparisons with numerical model predictions A. Seeded B. Unseeded. Source: Reference 2 . VI-6 3. Numerical Hurricane Modeling The proper use of numerical models promises considerable benefit to the design and interpretation of the Hurricane Amelioration Research Project. This is especially true to the extent that the models include new techniques that explicitly incorporate cloud entrainment, the apportionment of vertical mass flux between clouds and their environ- ment, the presence of clouds of various sizes and, to the extent feasible, the effects of freezing and precipitation in the convective elements. Numerical models are divided into two-dimensional (2-D) and three-dimensional (3-D) models. The 2-D models are limited to simulation studies of highly idealized storms. Although gross qualitative comparisons can be made with certain aspects of real hurricanes, model results are representative only of some sort of "average" or "typical" storms. Specifically with regard to the Hurricane Amelioration Project, consider- able uncertainty arises in 2-D simulations of hurricane modification experiments since the effects due to latent heat release are somewhat arbitrary with no explicit representation of the cloud physics. However, despite these problems, these models are very helpful in providing suggestive guidance on the design and interpretation of the experiment's hypothesis and results. Although the computations with 2-D models by Rosenthal and Moss, 4 Estoque, 5 and Sundquist, for example, provide no support for the original hypothesis (storms consistently intensify when seeded on that basis) , other numerical experiments by Rosenthal and Moss 4 have indicated that more effective modification can be obtained if the storm is seeded just radially outward from the eyewall center. The 3-D models have been developed only during the last five to six years. ' 8 > 9 These models are equipped with a finer computational grid and have been tested against the recorded behavior of several past hurricanes The results show that, in some specific cases, the models can predict the intensity and movement of storms with surprising accuracy (Figure VI-2) . The 3-D model developed by Jones 9 has three nested grids which are fully VI-7 45* 40° 35 c 30 c 25° 45° 40° OBSERVED PREDICTED 35' 75 70 c 65< Source '• Reference 10 FIGURE VI- 2. OBSERVED AND PREDICTED TRACK OF HURRICANE ALMA (1962; AS FORECAST BY A NUMERICAL MODEL VI-8 interacting. This system has been tested by means of several numerical experiments, and realistic features have been reproduced. Also, models that are capable of using observed initial data and that allow the storm to move and interact with its environment are under development at a number of institutions. The Modeling Group at the National Hurricane and Experimental Meteorology Laboratory (NHEML) , Miami, Florida, is now developing a three-dimensional tropical cyclone with three nested grids capable of simulating the track, intensity, and inner core structure of tropical cyclones. The model is being designed with the capability to provide accurate simulation of various cloud seeding tactics associated with the modification strategies proposed for Project STORMFURY for hurricane modification. The ultimate test of a 3-D model (or any model) is the ability to make accurate predictions from initial conditions that represent an observed state of the storm. Hurricane models have not reached this level of sophistication. Optimistically, a suitable model for the study of hurricanes and their modification is still an accomplishment for the future. For example, there are certain doubts about the accuracy or "physical fidelity" of the asymmetries in the meteorological fields generated by existing 3-D models. It is not yet established whether the asymmetries which have formed in various simulation experiments are accurate representations of the asymmetries observed in nature or are partly numerical artifacts. Such models do represent, however, the only means by which modification of hurricane tracks by seeding can be investigated ■ quantitatively in any theoretical sense. They are also probably the most appropriate means for improving predictions of intensity changes of actual individual hurricanes. Thus, although the 3-D models do require considerably more computer time than the 2-D models for forecasts of equal length and spatial resolution, their development is an important integral part of the Hurricane Amelioration Research Project. VI-9 C. Effects of Seeding on Hurricane Intensity (Maximum Windspeed) The major goal of the Hurricane Amelioration Research Project is to decrease the peak surface winds of a tropical cyclone by redistributing the kinetic energy away from the region of most destructive winds. Figure VI-3 shows the radial wind profile of an actual hurricane; it can be seen that the region of the strongest winds extends to 25 or 30 nautical miles (45 or 55 km) from the center. While not all hurricanes have radial wind profiles similar to that shown, the modification experiments appearing successful have been on storms with wind profiles such as those shown in the figure, The results of the four hurricane seeding experiments since 1961 are described below. 1. Experiments on Hurricane Esther, September 16 and 17, 1961 On September 16, the maximum windspeeds decreased by about 10 percent, and the radius of maximum windspeed increased very slightly within a 2-hour period. In the experiment on September 17, it is believed much of the seeding material was released inside the hurricane eye in the clear air at a radius smaller than that of the maximum winds. No changes in the characteristics of the hurricane were reported. 2. Experiments on Hurricane Beulah, August 23 and 24, 1963 On August 23, the hurricane was immature and unsteady, and the eyewall was an open semicircle of cloud changing rapidly in position. A sudden change just before the seeding resulted in the silver iodide being dropped into an almost cloud free region. It probably did not enter the active clouds of the eyewall during the 2 1/2-hour monitoring of the hurricane subsequent to the seeding. No changes in the characteris- tics of the hurricane were reported. After the August 24 seeding, radial profiles of pressure-height showed an average decrease in horizontal pressure gradient force of 16 percent in the region from 10 to 40 nautical miles from the hurricane center. Maximum windspeeds decreased by 14 percent, and the radius of maximum wind increased by distances varying between 4 and 10 nautical miles. The strong convective towers forming the eyewall prior to seeding appeared to weaken, followed by reformation of the eyewall at approximately 10 nautical miles greater radius after seeding. VI-10 I 1 1 I i — i — i — i — i — i — i — i — r AUGUST 27, 1958 J L J I L 100 200 DISTANCE FROM CENTER -nm J I 300 Source: Reference l FIGURE VI- 3. RADIAL WIND PROFILE OF HURRICANE DAISY 71-11 3. Experiments on Hurricane Debbie, August 18 and 20, 1969 These were the first STORMFURY experiments conducted in a manner similar to that prescribed for the present hypothesis (described in Section II-B) . These were also the first experiments in which a hurricane was seeded more than one time on the same day. On August 18, and again on August 20, Hurricane Debbie was seeded with silver iodide five times at 2-hour intervals over an 8-hour period. At 12,000 feet, the lowest level at which measured winds were recorded for several hours, the winds decreased 31 and 15 percent, respectively, on August 18 and 20, from 2 hours prior to the first seeding until 5 or 6 hours after the fifth seeding, Figures VI-4 and VI-5 indicate the results. The following items are also of interest: (1) The maximum windspeed decreased by amounts which exceeded the most optimistic estimates of STORMFURY advocates and . exceeded the decreases suggested by the model experiment. (2) The changes seemed to involve a flattening of the -innermost peak of maximum winds and a flattening of the inner area profile so that the residual high winds were spread over a plateau with a poorly defined maximum. (3) The next maxima were generally at a greater radii than the initial maxima. (4) Subjectively, the profiles following the seedings do not look "normal." The inner peaks in hurricane profiles are usually well defined, although other plateaus of fairly high windspeeds are not uncommon in mature storms. (5) On the time scale of hours, the double maximum structure in the wind field appeared relatively stable. (6) After the seeding (outside the inner maximum) , erosion of the inner maximum windspeed occurred. (7) The outer maximum was outside of most of the seedings and did not weaken significantly except in the right quadrant (where the strongest winds were originally found) . (8) These changes are not at variance with those predicted by the hypothesis or suggested by the model experiment. 4. Experiments on Hurricane Ginger, September 26 and 28, 1971 This experiment differed from the above three in that the seedings were performed in the rain sector and not in or near the eyewall. VI-12 100 - WINDSPEEDS MEASURED AT 12,000 FEET ALTITUDE 90 - - 90 BEFORE FIRST SEEDING AFTER THIRD SEEDING 4 HRS AFTER FIFTH SEEDING 20 10 10 20 DISTANCE FROM HURRICANE CENTER -nm 40 HNS (RIGHT) Source: Reference 12 FIGURE VI-4. HURRICANE DEBBIE WINOSPEED PROFILES RECORDED ON AUGUST 18, 1969 100 90 WIOOSPEEDS MEASURED AT 12,000 FEET ALTITUOE BEFORE FIRST- SEEDING 3 HRS AFTER FIFTH SEEDING SSW 40 (LEFT) 20 10 10 20 DISTANCE FROM HURRICANE CENTER -nm 40 NNE (RIGHT) Source: Reference 12 FIGURE Vl-5. HURRICANE DEB8IE WINOSPEED PROFILES RECORDED ON AUGUST 20, 1969 VI-13 The experiment was performed on a poorly defined and diffuse storm, and perhaps should not have been undertaken at all. Thus, four hurricanes have been seeded since 1961. All of these seedings, except those of Ginger , were performed in or near the eyewall. In all but one of these seeded storms, there were indications of a reduction in windspeed. In no case was there an indication of a windspeed increase. Only the two Hurricane Debbie experiments, however, were conducted in a manner closely resembling the present hurricane modifica- tion hypothesis. Analyses of the Debbie data indicate that there is more evidence than just the changes in windspeed that supports the present hypothesis. The time sequences of the wind, radar, and other data suggest, although not conclusively, that a modification to Hurricane Debbie was achieved; however, more of these seedings are required before definitive statistical support can be claimed. 5. Effect of Seeding on Windspeed Profiles 50 to 200 Nautical Miles from Hurricane Center Knowledge of the effect of seeding a hurricane on the windspeed at larger radii (50 to 200 nm) from the storm center is very important because it is quite possible that the outer envelope may acquire higher velocities and thereby influence storm surge effects adversely. However, there are very little, and not necessarily representative data, showing the wind profiles out to 50 nm or more. The only known data from a seeded hurricane are provided by the Debbie (1969) experiment of August 18 and 20. These show that on August 18, while the windspeeds decreased (in the left quadrant) out to 100 nm, they exceeded the preseeding values beyond a 110-nm radius by about 10 to 20 knots (Figure VT-6) . A time history of mean relative radar echo velocities has shown that windspeeds tend to increase at radii of 40 to 50 nm in all quadrants of the storm. i0 The above mentioned observations appear to be consistent with model results, which also predict an increase in winds outside the seeded eyewall region. 6. Natural Versus Induced Variability of Intensity While it is relatively easy to determine how the intensity of a storm changes with time, it is a very difficult problem to determine VI-14 Ik 100 - 80 - T 1 I I I I I I I I I I I I ' I I I 1 3EF0RE SEEDING (AUGUST 18) DISTANCE FROM HURRICANE CENTER -nm 100 BEFORE SEEDING (AUGUST 20) . \ m » , 80 - - o c _ J tjLj^A -* 1 60 *| - Q UJ AFTER SEEDING 1 — a. a z 40 (AUGUST 21) N- 5 20 n 200 ( LEFT ) 150 100 50 DISTANCE FROM HURRICANE CENTER- nm ■ Source : Reference 15 FIGURE VI-6. OBSERVED WINDS (12,000 FT. ALTITUDE) i AT GREATER DISTANCES FROM CENTER (HURRICANE DEBBIE, 1969 ) • VI -15 whether the intensity changes are caused by the seeding. The reason for this is that the natural variability (discussed in a previous section) of hurricanes is nearly the same magnitude as the changes expected to be induced by the seeding experiment. Figure VI-7 indicates the frequency distribution of windspeed changes (over a 12-hour period) in unseeded hurricanes in the Atlantic region. It clearly shows that many storms have large changes that are caused strictly by natural variations. There were various individual cases where the intensity changed up to 30 percent in 12 hours, and up to 50 percent in 24 hours. Thus, evidently, it is impossible to be certain that the expected reduction of windspeed by 15 to 30 percent in a seeding experiment is caused by seeding independently of natural changes. The development of theoretical hurricane models has provided the capability of simulating modification experiments, and these have somewhat supported the Hurricane Debbie experiment results. However, the models do not completely simulate the seeding, and also some theoretical models 5 ' 6 have shown an increase in windspeed as a result of seeding. It may be noted though, that the various theoretical simulations are not comparable since they are based on somewhat different models. In summary, whereas the seeding of a hurricane can be expected to cause a decrease in the intensity of the storm, available evidence — both theoretical and experimental — is inadequate to determine definitely whether the observed changes are induced by the seeding, by natural forces, or by a combination of both, D. Effects of Seeding on Hurricane Motion There is a general concern that hurricane modification experiments might change the tracks of the storms and thus cause damage in an entirely unexpected area. On the other hand, it is also feared that if modification of a storm does alter its track, then the much needed rainfall in a given area might be diverted to another area. In view of the above, it is important to carefully examine the question as to what effects, if any, the seeding of a storm has on its motion. The state-of- the-art in this regard is such that this question still remains largely VI-16 70 60 - 50 - UJ \ J. ••w-Stf 10 - -50% -30% -10% 10% 30% 50% MAXIMUM WIND SPEED CHANGES IN 12 HOURS Source'- Reference 1 1 FIGURE VI- 7. FREQUENCY DISTRIBUTION OF CHANGES IN MAXIMUM WINDS OF UNSEEDED HURRICANES VI-17 unanswered since, again, there have not been enough actual seeding cases to obtain a statistically significant analysis. However, there have been some attempts to make some qualitative and quasi-quantitative assessments based upon analyses of past seeding experiments and numerical model simulations. A direct relationship between tropical storm motion and environ- mental wind flow has been shown to be valid irrespective of the storm latitude, speed, direction of motion, intensity or intensity change (see Section IV). Thus, as the environmental wind field changes, so does the motion of the tropical storm. Analyses by Gray et al. 16 of 10 years of rawinsonde data from 30 stations in the western Pacific indicate that, in general , the tracks of tropical storms are closely related to the flow in the lower (altitude) level. A recent study 17 has indicated that the storm speed appears to be controlled by the 700 mb (3 km) wind flow pattern and the storm direction by the 500 mb (5 to 6 km) flow pattern. Given the scale and magnitude of the changes in windspeed hypothesized and the much larger region in which the flow field is positively cor- related with storm motion, the silver iodide seeding of the inner core of tropical cyclones would seem highly unlikely to have a detectable or significant influence on the motion of these storms. An investigation of the effects of seeding on the motion of the hurricanes seeded since 1961 has recently been completed. It examined the official forecasts made in real time by the National Hurricane Center and recently computed CLIPER forecasts 19 in order to document possible changes in storm motion resulting from the seeding experiments. It found that the synoptic-scale data can explain all of the major changes in direction of the tracks of these storms. For example, in the case of Hurricane Debbie (1969) , which was seeded on two occasions in a manner quite similar to that proposed for the Hurricane Amelioration Research Project, the CLIPER forecast errors during the seeding event of August 18 are relatively small (Table VI-3) . However, the official forecast errors were relatively large for forecasts initiated on August 19 and 20 when the major change in the storm's direction of motion occurred due to the storm approaching the synoptic scale circulation of the westerlies (Figure VI-8) . VI-18 Average Table VI-3 FORECAST ERRORS FOR HURRICANE DEBBIE (August 1969) Official Forecast CLIPER Forecast Date/Time 12-hour 2 4 -hour 12 -hour 24-hour 16/12Z 48 nm 71 nm 12 nm 33 nm 17/00Z 78 160 8 8 17/12Z 44 79 19 19 18/00Z 13 42 16 25 18/12Z 36 29 13 90 19/00Z 60 135 46 128 19/12Z 60 136 13 47 20/OOZ 36 73 23 21 20/12Z 66 208 54 149 21/00Z 97 284 32 177 21/12Z 73 104 37 47 22/OOZ 62 140 52 80 22/12Z 63 149 52 109 23/00Z 62 155 52 152 23/12Z 109 251 18 118 60 134 29 80 Official 24-hr forecast errors for all storms (1969) Mean = 140 nm Std. Dev. = 83 nm Source: Reference 21. VI-19 Source: Reference 21 FIGURE VI- 8. STORM TRACK AND FORECAST VECTOR MOVEMENT FOR HURRICANE DEBBIE, 1969 VI-20 . Although the forecasts for the seeded storms represent a small sample, the analysis of official and statistical forecasts suggests that forecast errors are independent of the time of seeding. No anomalous movements were detected that could obviously be associated with any of the eight seeding events. Also, no obvious quantitative differences were noted for seeded periods versus nonseeded periods in relation to the official forecasts and the results of an objective forecast scheme. In other words, the effect (if any) on the movement of the storm caused by seeding is well within the natural variability range. Although it is impossible to determine with any degree of certainty (because of the availability of only a few cases) that the motion of the storm is or is not affected in some small way by seeding, it appears that large changes in the storm movement cannot be caused by seeding based on the current hypothesis. Numerical estimates of the change in track of seeded storms have been made by simulating seeding in the three-level, three-dimensional model of a tropical cyclone developed at the National Hurricane and Experimental Meteorology Laboratory . 20 > 30 The model was used for a series of numerical experiments in which the vortex was subjected to enhanced heating during the mature stage. This heating was assumed to be applied in either a symmetric or asymmetric fashion, and the magnitude was selected to represent moderate (normal) or extreme (four times normal) heating rates, the latter being greater than would occur in any man-made seeding experiment. The results show that the principal effect of the enhancement upon the track of the vortex is to induce a small oscillation about the mean path. Changes that were noted after the seeding began were of the order of 1 to 6 percent of the total 195-nm (315 km) displacement of the control vortex with the 1 percent result (normal heating in the right front quadrant) possibly being the most realistic. Examples of some of the results are presented in Figures VI-9 and VI-10. The seeding simulation experiments support the observational research that indicates that the motion of a tropical cyclone embedded in an environmental current is primarily controlled by that current and is not greatly modified, and then only temporarily, by the changes of the vortex structure caused by seeding simulation. VI-21 °-1 p-p I f Q o UJ UJ / UJ CQ V) \ V) * / D / o I x O 2 t- < UJ X "_J < c 2 o ce 3S o o JZ i U. — O "3 co ■o 3 o E o X o u ^™ 4> CC O 0) u. CO ^ o 2 < CC cc UJ i- Q < UJ a. u_ o s o 5 cc u. s o 5 _ o o 0) > CO o. UJ If) 1- < 2 . in N. 3 UJ h- 1- t» co < ju UJ CC u> 2 41 "3 „ . o E 0» 0> o vD. Irt 1 — eg 3 o ■ u e > Z 9 UJ • if 3 O O O O- - O cn VI-22 o », °J -.11 H "J UJ 03 vt i VI •*> 1 1 » \ 1 1 f /' \i° _l i \ i \ i * i O OB *^Z »-. \ X \ 2 \Vn o- acqs O a o u. < x 3 S \ irt\ a > O K \l jl Z u. \ t o CC E X w UJ ^ o U. 0J CO < a: CO CO Z cr Q UJ UJ u. i- a g o 2 u u. o CO UJ cr UJ 2 2 '8 < > CO fr i- co UJ < Z 4) 6 l» I w 1 1—1 . o in V E o w o > UJ 3 a Z 5 e « <•• D O CM • « p C 3 O u. o - o C/> VI-23 The preceding discussions indicate that two silver iodide seedings of the inner core of hurricanes (as proposed for the Hurricane Ameliora- tion Research Project) had no significant influence on the movement of the storms. However, this finding must not be considered as a definitive one for the following reasons. The models used for the prediction of storm tracks use various statistically based procedures to identify the probable movement of the storm based on historical or analog data or on statistical screening procedures. These relate the expected movement to conditions prevailing in the large scale environment at the time the forecast is initiated. And experience has shown that no one of the various models is uniquely superior for the majority of forecasts, and that there are significant errors which range from 90 to 175 nautical miles for the 24-hour forecast. There are a few dynamic methods for forecasting storm motion which yield mean position errors of 75 nm at 24 hours, 150 nm at 48 hours, and 300 nm at 72 hours. 23 The main difficulty with these methods is that they are not able to predict the small deviations of storms (from their relatively smooth track) that are observed. These have amplitudes of as much as 20 or 30 nm and periods ranging from a few hours to about two days. The results (regarding the effect of seeding on the environment of storms) based on numerical simulation also require further research because of the limitations of the current numerical models to test the cloud seeding hypotheses realistically. This is particularly true with respect to the representation of the condensation heating processes. E. Effect of Seeding on Size of Storm One of the effects of seeding a hurricane, according to the current hypothesis, would be to displace the maximum wind region outward. It is envisaged that indication of such a displacement would be found in the outward displacement of the storm eyewall as manifest by the precipitation echoes on airborne or groundbased radar. Airborne radar photographs of Hurricane Debbie (seeded on August 18 and 20, 1969) have been used 22 to measure the echo free area within the eye at 5-minute intervals, beginning 1 hour before the first seeding and ending 1 hour after the last seeding. VI-24 Their results show that on August 18, the echo free area (storm eye) increased suddenly at seeding time plus 1 hour and 15 minutes; the increases ranged from 50 percent to three fold. In contrast, airborne radar studies of unseeded storms have indicated that the eye of a mature hurricane generally tends to decrease in size. For example, data from Carla (1961), Betsy (1965) , and Beulah (1967) indicated that the eye decreased from an average of 30 nm in diameter to 23 nm in diameter during the first 24 hours after reaching hurricane intensity. However, much more research of both unseeded and seeded storms is necessary before any definitive conclusions can be drawn about the effect of seeding on the size of the storm. This is particularly true in the case of eastern Pacific hurricanes, which have not been studied so extensively as western Pacific and Atlantic hurricanes with respect to natural variations in their size, duration, and basic structure. However, available information points out that on the average, eastern Pacific hurricanes are relatively smaller and have shorter durations than those formed in the western Pacific and Atlantic. F. Effects of Seeding on Precipitation Pattern and Amount Many areas of the world depend on rainfall from tropical cyclones to alleviate drought, nurture crops, fill reservoirs, restore ground water, and satisfy other water requirements. An important consideration, therefore, is whether or not the beneficial precipitation accompanying tropical cyclones is increased or decreased by modification experiments. The discussion in this subsection examines whether modification of a hurricane could cause significant alterations in the rainfall pattern or amount associated with the storm. The water vapor budget of a mature, steady state typhoon has recently been estimated based on the analysis of 10 years of rawinsonde soundings and rainfall data from several small islands in the western Pacific* The average moisture budget for all typhoons with central pressure less than or equal to 980 mb (i.e., windspeeds > 80 knots) and * Although no comparable study is available for other areas, the results should be generally applicable. VI-25 storm center positions from 4° to 30°N is shown schematically in Figure VI-11. The figure shows that within 2° of the storm center (120 nm) the horizontal convergence of water vapor (moisture received from adjacent radial bands) is twice as large as the local evaporation from the ocean surface. This implies that small changes in evaporation within 2° of the storm center would not have a significant effect on precipitation in the inner regions of the storm because evaporation is not the major contributor to the moisture budget at these radii. The full extent of storm convective activity is usually confined to an area within a radial distance of 6° (360 nm) of the storm center. In this region, 73 percent of the moisture for precipitation is accounted for by local evaporation. Since such a large portion of the moisture in the precipitating regions of tropical cyclones is accounted for By local evaporation, significant changes in total storm precipitation are possible only if windspeeds can be lowered over a large part of the storm within 6° of the center. The hurricane modification experiments, however, are attempting to decrease the winds by 10 to 20 percent in a region covering less than 5 percent of the area within 6° of the storm. Windspeed decreases of this magnitude within 1° of the eye would have very little effect on the total evaporation because this region is very small compared with the total area covered by convective activity. Besides the moisture budget study, satellite techniques have also been used to estimate the time changes of rainfall associated with seeded hurricanes in the Atlantic. Profiles of rainfall along and at right angles to the direction of motion of Hurricane Debbie for August 18 and 20, 1969, were derived from radar data. Comparison of these profiles with those for the days on which no seeding was attempted (August 19 and 21), shows that there are virtually no differences between the two. While it is possible that seeding may have altered the rainfall over localized areas, no changes in the large scale volumetric rain output of Hurricane Debbie were detected. The magnitude of changes in rainfall that might occur due to hurricane modification experiments have been simulated in a recent study. 25 A typical rainfall pattern for an unseeded tropical storm has been altered in VI-26 en o r» o r«. CO d O d "• "•■— o •^ r* CO 6 d d i —•• **-* at o *■ ■■* r«» lO d d d -m. "*~~~ o O o •■* r- to d d d — ■"•" ""~"~ CO m (0 •^ CO d *■* ~ i — * — — r>» ^ p "* n in ~ r> — ■• - — o co o CM O o t- UJ CO o a. s o o E *-• o u> ■o »^ c ■o a CO > ID o Ui C ) cc ■o o cc Z O z Z co _ w- o UJ 3 c < H a cc UJ > 3 2 u a ■o < a. < cc o a. 2 **^ UJ < o CC > u a. u CM 4> CC e o o 5 3 O Z o o X a. o ° o ° 2-3 x cr o cr UJ co UJ UJ co o a. 2 O o o a. o > w CD "a 5 _ o CD "^ 5 e- CD "X s r _ o U. O ml o ■£ .r UJ o Q 3 CD cr UJ < 5 c ... o c o .5 = c CD O 5 ° UJ cc o vi-2 7 several different ways that represent both reasonable and extreme consequences of proposed modification experiments. Figure VI-12 is an example of the results. It indicates the accumulated rainfall along an east-west line for a steady state storm moving at a constant speed, 10 knots, toward the north. It includes the assumptions that the rain- fall rate is decreased by 50 percent in the radial region 20 to 40 km (11 to 22 nm) from the center (eyewall area) , and increased by 20 percent in radial region 40 to 80 km from the center (inner rainband) . These adjustments in the rainfall rates are assumed to persist at the same rate for 20 hours. The accumulated rainfall represents an upper limit to changes in precipitation that might be caused by a modification experiment since modified storms are expected to begin returning to their natural state 6 to 18 hours after seeding. It can be seen that the overall net change in rainfall is very small, and that, even at particular points, the hypothesized seeding effects on rainfall rates that might occur as a consequence of seeding experiments indicate that seeding will produce very little change in the rainfall patterns. Theoretical model simulations have also shown that the seeding will cause a small increase in the total integrated rainfall over the entire storm. However, it will be distributed over a larger area, and the maximum rainfall at a particular point would not be any greater than that for an unseeded storm. In summary, the research to date has indicated that the total rain- fall associated with a tropical storm is basically a function of the synoptic scale environment, and thus any significant changes in the total storm precipitation can occur only by affecting the large scale mass convergence profiles or by lowering the windspeeds over a very large area. Therefore, the hurricane amelioration experiments, which are aimed at reducing and spreading out the small area of extremely intense winds at the core of a hurricane, will not appreciably alter total storm rainfall. VI-28 - 1 CO O z T T" T~ I T 1 o 7 o UJ UJ 0. CO / 2 cc o (0 ^•— ; : .v •■;'.•«•* -^ ' '-.:,v; ;'£ '•' _l ±z •"■ << ' .Jo- •""■' OCX UJUJ 2^ * 3* • ■ >.-■*" *■■'-,. '*•■ o o CM o - * co« 1 + . ' r . : r-f ,:.r CO < a UJ LU O z *^>i^& ' ' •:''■ : ■?■' :• z~~ 2 <>. . - — QZ — ' CJ O i o CO 1 o o CM i 3 CO CO < X ^\ u ^ _l ' -J "^0^ ■ "■• ' - f ' ; 'Y^ o o o 1- < 11 CM T CO u. z < — *£ DC a ^**<*\~ LU_I ^^* '.;>>■•:*.* — _l -» < UJ 01 < 2< ^^5^:-v *'•■];.; ' m . z zee ^W ■ ^"" o a CO CO 3 ^***S^V 5 oc 3 O -J u -J _) a z- < CO z a z < CD Z J: ^ „ 1 < < ac < OS I 1_ _L | | i ^ o o o CO o CM o CO o - o o I UJ u z < CO a a < or o CO o cvj o CO O O CM C9 3 Z AC a: =5 c O o «•■» UJ o z 3 -J E K CO co "*"* UJ 5 2 or i < o co UJ CO C9 z z > o o -J < z a -J or -J < < u. z < or X or o z UJ > U. o < UJ CO < 3 CO 2 CO 3 < o 0. in CM a ac 3 O CO cm UJ or CO u. o o o o O o o CO r- CO in ut to VI-31 start dissipating. This is indicated very clearly by Figures VI-13 and VI-14, which show that severe storms associated with cyclones occur inland. In view of this, and the fact that seeding operations are expected to be confined to mature cyclones that are not likely to affect land/populated areas during seeding operations, the seeding of hurricanes should not have any effect on the hurricane-spawned severe storms. The research literature on cyclone-associated severe storms consists of only about 10 articles that deal with individual case studies or a climatology of a number of cases. Although informative, these studies have not come to grips with the critical environmental process which can be associated with the occurrence of severe storms even in unseeded hurricanes. This, together with the uncertainties associated with the behavior of seeded hurricanes, makes it very difficult to draw definitive conclusions. 2. Large Scale Environment Various studies 16 have clearly documented that, by and large, the large scale circulation controls the hurricane, its movement, rainfall, intensity, and other characteristics. Knowledge gained from observations, theoretical studies, and numerical model experiments indi- cates that the seeding experiments are not expected to destroy hurricanes, change their tracks significantly, or have much effect on rainfall. Hence, it is extremely unlikely that the proposed hurricane seeding experiments will have any noticeable effect on the large scale synoptic or general circulation of the atmosphere or climate. I. Effects of Seeding on Surge and Waves If the experimental hypothesis is correct, and the storm returns to its initial conditions a few hours after seeding, no lasting change in the storm surge is expected. The results of numerical modeling of the hurricane reported above and the limited experimental results suggest that this is true. If, as expected, hurricane size, intensity, wind profile, or track changes temporarily, corresponding changes in the mean water level will occur. VI-32 Source : Reference 27 FIGURE VI-13. GEOGRAPHICAL DISTRIBUTION OF HURRICANE TORNADOES (1948-1972) VI-33 Source: Reference 26 FIGURE VM4. GEOGRAPHICAL DISTRIBUTION OF TYPHOON TORNADOES (1950-1971) VI-34 1. Surge Adverse impacts produced by storm surge will occur if the temporary changes in the hurricane increase the surge level along the Mexican coastline. Model studies 28 ' 29 have been made of possible changes in surge caused by altered pressure distribution, wind profile, and storm size. These changes apply to storms on tracks that are either perpendicular or parallel to the coast for a standard basin (as defined in Section IV) that simulates a wide, shallow continental shelf area. The model (see References 28 and 29 for model assumptions, accuracy, and limitations) suggests that changes in the storm structure will alter the surge pattern. Model calculations 28 suggest that the peak surge may increase as a result of seeding. As a result of these model calculations, the authors cautioned against using maximum windspeed as the sole and direct measure of peak surge; they indicated that consideration of at There are additional reasons for not being too quick to accept the model results as a realistic prediction of the results of the seeding experiment. First, given the complexity of predicting surge, these models are simplistic; among other things, they do not consider the effects of offshore topography. Second, the increase in peak surge height predicted by the model has not been observed in nature. Third, seeding eligibility rules preclude operating on storms close enough to land for any change in the peak surge to have significant effect on the coast. The model calculations also suggest that increased surge in the peripheral region of the storm can occur even if the peak surge is decreased by seeding. However, storm surges at this distance from the storm center are rarely damaging. The calculations are based on the assumption that decreased magnitude of the maximum winds results in in- creased outer winds (see Figure VI-15) . As indicated above, only a VI-35 least one additional parameter is imperative. -, ■ limited set of observations (on Hurricane Debbie) is available to verify this assumption. This figure shows an increased surge level of 1.0 to 1.5 feet in the outer, low windspeed regions of the storm (within 140 nautical miles) with a corresponding decrease of 3.0 to 6.5 feet in the peak surge. Because Hurricane Debbie was somewhat larger than most eastern Pacific hurricanes, the effect described would be confined to a smaller radius for eastern Pacific storms. Many eastern Pacific hurricanes tend to move approximately parallel to the coast. This per- ipheral effect could be of interest for such storms, depending on their distance from the shore when seeded (see Section VII. C. 2.). The actual magnitude of the increase in the surge for eastern Pacific storms is extremely difficult to ascertain from the example above because of the uncertainty of offshore topographic effects, the surface wind speeds, and the size of the storm. For the narrow, steep shelf areas, the surge height would normally be less than for wide shallow shelf areas. For offshore islands with narrow and steep offshore areas, this effect would probably be small. In addition, the model predictions extend outward only to 140 nautical miles and do not cover the region of most interest. Because of limited wind speed data existing for the periods before, during, and after seeding, it is hardly feasible to extrapolate or make reliable quantitative judgments from the curves in Figure VI-15. Without modeling or observations for the Mexican coastline, the results of the surge modeling 8 discussed above, although qualitatively representative, are considered inconclusive and probably high. In summary, changes in storm surge may occur on nearby coasts if changes in hurricane track, speed, wind profile, pressure distribution, or size also occur. Therefore, prediction of changes in storm surge resulting from seeding depends on the ability to predict changes in the hurricane's structure and movement. VI-36 o o o o • • M »»»*-!H9l3H 39«nS E c l 2 o CE U. UJ (J 2 < a ao CM » u c « CC 3 O en CO < o Ui a. o a UJ 3 a. 2 O o a z < a UJ UJ a. CO a CO o a UJ > cr UJ CO oa o CO UJ u. O cr o. io i > UJ cr 3 o UJ CO 09 UJ a UJ z < o cr cr 3 X cr e o cr 3 CO stou*-a33dSaNIM 1»»*-1H9I3H 30HDS VI-37 Because the average radius of eastern Pacific storms (Table IV-3) is 1.65 degrees (about 100 n.m.), and the seeding eligi- bility rules effectively prevent seeding any storm that is close to the shore, the impact of any increased surge on the shore should be minimal . Waves Any changes in hurricane wind field characteristics will alter the magnitude or spatial distribution of the wave field. If seeding reduces peak winds and increases winds in the outer regions (see Figure VI-15) , wave height in the outer portions of the storm will be increased. Any coastal regions or offshore islands directly affected by these outer portions will probably have larger waves than otherwise. It is possible, although unlikely, that the present seeding eligibility criteria could be met and yet the wave height along the shore could increase somewhat. Consider a storm slightly more than 150 n. miles (278 km) offshore and moving approximately parallel to the coast (see Section VII. C. 2). In determining the approximate magnitude of the results of seeding such a storm, the wind field changes that occurred during the seeding of Hurricane Debbie (Figure VI-15) are assumed to be representative of the expected changes. Again, because Hurricane Debbie was somewhat larger than most eastern Pacific hurricanes, the wave height increases would be confined to smaller regions. A simple model 32 for forecasting hurricane wind- generated waves (see References 32, 33 for assumptions, accuracy, and limitations) is used to calculate the significant wave height* in the outer regions of the storm (250 km from the center). The Hurricane Debbie * Average height of the highest one«third of the waves VI-38 wind profiles suggest that the wind speed increased about 5.2 m/s (10 knots) at 250 km (135 n.m.). The wind speed before seeding ranged from 9.3 m/s to 10.3 m/s (18 to 20 knots). The model predictions of the significant wave height are shown in Table VI-4 for wind speeds ranging from 9 m/s (17.5 knots) to 15 m/s (29.2 knots) in 3 m/s (5.8 knots) increments. The model calculation shows that for every 3-m/s increase in the wind speed, slightly less than a 1 meter increase in the deep water significant wave height occurs. The model results should only be considered qualitative, however, because the model does not account for the effects of storm track, shoaling, or dissipation processes. In addition, the magnitude of the wind increase is based solely on the modified wind field of one hurricane modification experiment which may not be representative of seeded storms in general. Nevertheless, the calculation does suggest that small increases in the wave height might be anticipated under some conditions if the storm were to move closer to shore after seeding. Table VI-4 SAMPLE CALCULATIONS OF HURRICANE SIGNIFICANT WAVE HEIGHTS IN THE OUTER REGION OF THE STORM (250 km [135 n. mi.] from Center) Wind Speed Wave Height 9 m/s 1.26 m 12 2.08 15 3.07 The waves propagate out of the storm as well, therefore, even though a modified hurricane regains its initial condition after seeding has been terminated, the swell approaching the coastal regions and offshore islands may differ from that of an unseeded hurricane. However, the change in the wave climate would probably be minor. VI-39 Since, as with storm surge, the magnitude of the wave field at the coast for seeded storms moving parallel to the coast and marginally meeting the eligibility criteria is difficult to predict, the degree of severity of the impact is also difficult to predict. Because the average radius of the eastern Pacific storms is 1.65° (99 nautical miles) and the seeding eligibility rules, in effect, prevent seeding any storm closer to the coastline than about 150 nautical miles, the impact of increased surge and waves on the coast should be quite small. VI- 40 VII PROBABLE ENVIRONMENTAL IMPACT OF EXPERIMENT A. General The effects of the Hurricane Amelioration Research Project in the eastern Pacific consist of: (1) the individual and cumulative impacts of the deployment, support, and aircraft operations; (2) the dispersal of silver iodide; and (3) the changes in the hurricanes and surge. The discussion of impacts in this section is organized around the second and third categories of project activities mentioned above; the first category is outside the scope of the analysis. B. Impacts of Silver Iodide and Other Air-Launched Material 1. General Three types of devices will be launched from aircraft during experimental missions: silver iodide flares for modifying the hurricane; dropwindsondes for measuring wind, temperature, humidity and pressure; and expendable bathythermographs for measuring sea surface temperatures. The latter two devices weigh less than 5 pounds each and are deployed by parachute over the open ocean. They should slowly corrode with negligible environmental effects. Apart from the intended impact on the atmosphere, the impact of the silver iodide (Agl) flares is also negligible. The rationale for this conclusion is presented below. . 2. Natural Environment a. Water Quality 3 Z. Agl dispersal will not significantly affect water quality. Concentrations of Ag equal to or greater than 5 ugAg/L* water constitute a hazard to marine environments, whereas levels less than 1 ug/L present * L = liter VII-1 little risk. 1 Because Agl is quite insoluble, the biological concentra- tion of Ag ions, the source of the greatest hazard, will at worst be on the order or 2 ug/L in the upper few centimeters of surface waters for extremely brief periods. In practice, given the strong mixing of both the atmosphere and the upper hundred meters of the ocean during a hurricane, it is unlikely that oceanic concentrations will ever approach even this relatively low value as a result of seeding. Details of the calculations are given in the Appendix of Reference A. b . Air Quality As a quite conservative estimate, the 20 g of Agl iodide released by each flare can be assumed to be uniformly distributed in a volume of 50 million cubic meters, giving a concentration of Agl in the air of 0.4 ug/m-'. Hence, the maximum concentration of Ag in the air, assuming that all the Ag remains in the air and that none is entrained in the rain, will be 0.18 jg/m-*. This will be approximately 50 times less than the 10 -g/ta* threshold limit for Ag reported by the American Conference of Governmental Industrial Hygienists in 1966. Similarly, the maximum concentration of I (as iodate or iodide) in the air, again assuming no solution or entrainment in the rain, will be about 0.22 ug/m^, in contrast to reported threshold limits of 1000 ug/m . Thus, neither Ag nor I added by the proposed experiments will significantly alter the concentrations of these elements in the atmosphere. c. Land Form and Soil No measurable impacts are anticipated. Assuming Bowen's estimate of 0.1 ygAg/g soil applies to island soils in the study area, and that the bulk density of the soils is 2 g/cm^, there would be some 40 kg Ag/ha in the upper 20 centimeters of soil. Thus, if seeding were to occur over an island, even if an entire flare's combustion products were to be restricted to an area 1 kilometer on a side, the addition to the existing Ag in the soil would be less than 0.003 percent. It is probable that the actual increment would be at least two orders of magnitude smaller because this assumes unrealistically low dispersal of the Agl through the air. VII-2 d. Biota No significant impacts of living matter are anticipated because the expected changes in both I and Ag concentrations in soil, air, and water will be well within the variation that can be expected to occur naturally. 3. Human Environment No impacts on human health, man-made structures, or land uses are expected. Because the maximum concentrations of both Ag and I in air and water are well below known toxic threshold limits, there will be no threat to human health. Because of the extremely low probability of an unburned (or burning) flare landing on the ground, effects on structures or land use are very unlikely. C . Impacts of Hurricane Changes 1 . General Considerations Subject to the numerous uncertainties and qualifications discussed in the previous section, the only hurricane changes which may have any appreciable impact are those of the wind, and possibly, the accompanying surge. During a 6- to 18-hour period after seeding, peak winds are expected to decrease, and peripheral winds are expected to increase somewhat. These changes are expected to occur only in the experimental area or just beyond its boundaries. Adherence to the seeding eligibility criteria is intended to prevent any hurricane in a modified state from affecting any land mass. However, because the criteria are applied to the center of the hurricane, land masses will be affected if the hurricane approaches land near enough for the higher peripheral winds to be experienced on land. As noted in the previous sections, there are very few data on wind changes during actual seeding experiments. Examination of these data does provide some perspective, however, for categorizing the areas of interest. Figures VI-4 to VI-6 in the preceding section indicate the variation of measured windspeed with radius from the center of Hurricane VII-3 Debbie (1969) before and after the modification experiments. These data were obtained by aircraft, mostly at 12,000 feet altitude. The presump- tion is that there is little variation between winds at the altitudes of measurement and winds at the surface. If such is the case, the winds at the surface would have been approximately as shown in the two figures, and estimates of the wind force at sea level can be inferred from them. The dynamic pressure arising from the motion of the air is a determinant of the actual wind force on an object. Since the dynamic pressure varies as the square of the windspeed, a reduction in peak windspeed from 100 knots to 83 knots (as reported in the Debbie experiment of August 20, 1969) would reduce the peak force on a given object to a value of (.83)2 or 69 percent of its original value. If the wind on an object in the outer regions of the modified storm were to increase from say, 30 to 40 knots, the force on the object would change from 9 percent to 16 percent of that for a 100-knot wind. No systematic summary is available of variation with windspeed or dynamic pressure of damage to biota or structures. However > there is sufficient information to make some useful deductions. Damage criteria for forests have been developed (Table VII-1) that "are applicable in most cases." Note, however, that the table does not take into consideration the effects of precipitation and ground saturation on the vulnerability of trees to uprooting. The specific nature of the damage will depend on a variety of conditions, including the types of trees, whether they are in natural or planted stands, in favorable or unfavorable growing conditions, and whether or not they are simultaneously subjected to surge effects. For example, coconuts and Casuarina may tolerate winds up to about 125 knots, while breadfruit and pandanus trunks break at windspeeds of 70 to 80 knots, and perhaps less. The above information indicates that, with respect to trees, changes in higher windspeeds, say, above 50 to 60 knots, could be con- siderably more important than changes at lower speeds. Consequently, a stand of trees affected by the central portion of a hurricane during its modified phase might be beneficially affected by the modification to a VII-4 Table VII-1 DAMAGE CRITERIA FOR FORESTS Damage T yP e Severe Nature of Damage Up to 90 percent of trees blown down; remainder denuded of branches and leaves. (Area impassable to vehicles and very difficult on foot.) Equivalent Steady Wind Velocity (miles per hour) 130-140 (Knots) (113-122) Moderate About 30 percent of trees blown down; remainder have some branches and leaves blown off. (Area passable to vehicle only after extensive clearing. ) 90-100 ( 78- 87) Light Deciduous only; very few trees blown down; some leaves and branches blown off. (Area passable to vehicles.) 60- 80 ( 52- 70) Source: Reference 4. VII-5 significant degree. On the other hand, a stand of trees affected by the outer portion might be subject to only slightly adverse affects (in the absence of significant changes in surge) . There is less clear-cut information on damage to structures. The wind force increases with the square of the windspeed, and the cost of a building generally increases with the level of wind force it is designed to resist. From these considerations, it can be shown that over a considerable range of windspeeds , the level of damage in dollar value should increase faster than the square of the windspeed. One study 5 of hurricane damage in the United States, for example, indicated that the dollar value of damage varies as the fourth power of the wind. Although for a variety of reasons" this study is not directly applicable to damage to island and to Mexican coastal structures, it does indicate the importance of changes in winds at the higher levels. It would appear that, as in the case of damage to vegetation, changes in the higher winds near the center are considerably more important with respect to damage to structures than changes in the lower winds near the edges of the hurricane. Similar arguments can be stated with respect to surge. The uncertainty is higher because there is some question as to whether increases or decreases in wind always lead to corresponding increases or decreases in surge. The reasons include: (1) the types of structures are different; (2) the horizontal and vertical distributions of structures are different; (3) since the study applies to the United States, surge plays a much larger role in generation of damage because of the much higher surges generated by the gradually sloping ocean bottom and the large coastal areas; and (4) as the hurricane intensity at landfall increases, the land area over which the hurricane travels prior to dissipating also generally increases, thus exposing more structures to damage. This is not the case with a small island or a narrow coastal region next to mountains where the entire area is exposed to similar winds. Once a certain level of wind is reached, almost all the structures are destroyed. Typhoon Amy (1971), for example, blew 80 percent of the structures down on Truk with peak winds of 98 knots. VII-6 2. Likelihood of Impact Occurrence No detailed probability analysis is available of a seeded hurricane striking or passing close to the western coast of Mexico or an island while it is in an altered state. Some perspective about this likelihood can be gained, however, by examining the relations between the seeding rules and the hurricane characteristics in the eastern Pacific experimental area. As mentioned in connection with Figure IV-5, 80 percent of eastern Pacific hurricane track directions fall within ±20° of 290° (approximately west-northwest). Most of these hurricanes can thus be seeded since they are moving away from the mainland. (Most of the Pacific coastline of Mexico is at angles between 300° and 330°.) Even if a hurricane were moving parallel to the shoreline, use of the three explicit seeding eligibility rules (page 11-11) and estimated hurricane prediction accuracy would require the hurricane to be a considerable distance offshore for it to qualify for seeding. This "considerable distance" has yet to be determined, in part because Prediction Probability Ellipses for the eastern Pacific are not available. Some preliminary estimates have been made, however, by using an example ellipse for the western North Pacific (Part I.C.2 of Appendix A of Reference A discusses the use of probability ellipses and illustrates the one used here). The estimates are conservative because the prediction of movement of western Pacific typhoons is probably more accurate than that for eastern Pacific hurricanes. The results indicate that the hurricane would have to be on the order of 150 nautical miles offshore to meet the requirement that the probability of its being within 50 nautical miles of land within the next 24 hours is 10 percent or less. At distances, on the order of 150 nautical miles there would be little likelihood that the increased peripheral winds or surge will significantly affect the shoreline. VII-7 3. Natural Environment This section presents a discussion of the impacts to the natural environment that might occur in the unlikely event that a hurricane passed close to or struck the mainland or an island while it was still in a modified state. a. Water Quality Any impacts on water quality arising from a modified storm are most probable on low islands and coastal lowlands, particularly those on the order of 100 nautical miles from the modified storm center where increases in the storm surge and peripheral wind speeds could increase the land area inundated. Inundation by seawater could contaminate major sources of both domestic and agricultural water supplies along the coast, but the impacts should be localized to the coastal fringe. b. Land Form and Soil If the storm behaves as expected following seeding, there is little potential for adverse impacts because islands in the experimental area are rare, generally uninhabited, and possess biota adapted to the periodic occurrence of hurricanes. However, if the effects of seeding are exceptionally persistent and the storm recurves to strike the mainland, the increased diameter of the storm may increase the area of hazard to the heavily populated coastal lowlands. In either case, the effects of storms on land form and soil quality will be most conspicuous along the coasts, where the effects of wind, waves, and flooding will be combined. Effects on the interiors of high islands and continental interiors should be much less conspicuous, but some increase in the serious erosion and flooding hazards associated with all hurricanes might be expected. Consequently, alterations in typical storm effects caused by the proposed seeding experiment may affect — beneficially near the center, adversely near the periphery — the VI 1-8 windward side of the coastal ranges of the Mexican mainland and portions of Baja California if the seeded storm unexpectedly recurves, and if the effects of seeding are unexpectedly large and persistent. Outside the storm center the increases in surge height, coupled with increased windspeed, and wave height, may increase soil erosion or burial By gravel and rubble. 8 * 9 However, these adverse effects of the storm should be reduced near the storm center as a result of seeding, and the net effect should be beneficial. c. Biotic Environment The proposed experiment may cause some loss of wildlife if the storm strikes land and if effects of the seeding are unexpectedly large and persistent. However, these impacts, if they occur, are unlikely to significantly reduce the population size of any terrestrial species. Nor are the proposed experiments likely to affect the grey whale adversely; it breeds during the winter when hurricanes do not occur in the eastern Pacific. Although a seeded storm could alter the undersea topography of the breeding groups of the grey whale, it is unlikely that this impact will be large enough to affect the large lagoons (which have surely been struck by hurricanes in the past) . Similar arguments can be given for potential impacts on the seals and sea lions of the offshore islands, but the logic is at first glance less persuasive with respect to these organisms that are resident year round, restricted geographically, and few in number. However, these animals should be relatively unaffected by changes in storm surge, and they occupy islands that lie within areas excluded from seeding (Figure II-9) . Hence, it seems probable that impacts of a seeded storm on these species will be insignificant. The impacts of seeding on both reef and open ocean biota will be modest and perhaps beneficial over the long term. The principal impact on open ocean systems will be to enlarge the area of intense mixing, and hence the surface water of biological productivity will be increased. The major impacts on reef ecosystems will be to lessen the intensity of impact on those reefs that are hit by the modified storm center but to VI I -9 increase the intensity of wave action on those reefs that are relatively remote from the storm center. Although such intensification will cause more extensive erosion of reefs, this superficial destruction is a normal part of the environmental regime to which reef communities are adapted, and it probably plays an essential role in maintaining a diversity of sites for colonization by "weaker" species of coral and the invertebrate and fish populations that depend on them. For corals that bear the full brunt of a hurricane, destruction of as much as 80 to 90 percent of the coral colonies has been reported for windward reefs. Impacts on corals below the low water level in lagoons, under otherwise similar circumstances, appear to be negligible. In general, the more rapidly growing branching, fragile species are more susceptible to damage than the slower-growing, massive, globular to hemispherical species. 15 The terrestrial impacts of a direct hit by a hurricane are severe for low islands and coastal lowlands, but impacts on the terrestrial fauna appear to be relatively mild. A direct hit of Jaluit Atoll* by Typhoon Ophelia in 1958 provides the best documented analysis of storm damage. This typhoon, which coincided with a spring tide, destroyed an estimated 80 to 90 percent of the vegetation, contaminated the available water supplies, and destroyed most of the food supplies, and all but one of the fishing boats; apparently, however, it did not severely affect the wild, nonhuman fauna, with the possible exception of the rat population.*^ Impacts were most severe on those islands subjected to both heavy winds and inundation, and hence, scour and soil burial; impacts were markedly less on islands subjected to heavy winds without accompanying inundation.^* 11 However, the biotic impacts were not persistent, and significant recovery occurred within 3 years in areas where the soils remained intact, and complete recovery was expected within 10 to 15 Jaluit is in the Marshall Islands VII-10 years. 8 Recovery from a similar storm in British Honduras was somewhat slower, 10 possibly reflecting the greater surge heights and sediment scouring that might be expected along continental shores with well- developed continental shelves. Great variation in impact occurs among islands that are subjected to direct hits by typhoons or hurricanes. The least impacted islands retain a relatively undamaged ground cover, and heavily damaged, but rapidly recuperating tree crops. 13,14 Consequently, it is probable that impacts from increased surge and windspeeds relatively distant from the center of seeded storms might increase the damage. Nevertheless, this damage will probably be no more, and possibly less, than that sustained on the least impacted islands of Jaluit atoll — all of which were near the center of Ophelia. Recovery should be equally rapid, occurring within a few months to a few years. ■ Impacts on mainland or island fauna attributable to change induced by seeding should be minor. Cattle and pets tend to be sheltered with the human populace and should be unaffected by the incremental impact caused by seeding. However, rodents and other small animals may be affected somewhat more. Birds, lizards, and insects survived the Jaluit typhoon without conspicuous affect 12 (although some mortality is to be expected) and presumably would similarly escape from damage from distant effects of hurricane modification. 4 . Socioeconomic Environment a. Experience with Hurricanes Eastern North Pacific hurricanes, although not so large or well-publicized as their Western North Atlantic counterparts, occur with greater frequency. Furthermore, they have traditionally been con- sidered more treacherous. Largely because of their compact size they are harder to track and their wind speeds increase much more rapidly between the storm's periphery and center, leaving little time to prepare for the onslaught of wind and rain. In addition, the severest North Pacific storms tend to occur later in the season than those of the North Atlantic. VII-11 One of the more severe storms on record, the Manzanillo Hurricane of October 1959, is a good case in point. Although its winds were confined to within a 30-mile radius of the eye, wind speeds rose from 65 to more than 135 knots in this short distance. Winds, floods, and landslides caused nearly 1000 deaths in Manzanillo and the surrounding areas; moreover, 6 Mexican merchant ships and 1 naval vessel were sunk.* 6 In 1957, and again in 1965 with Hurricane Hazel, severe late season storms struck in the vicinity of Mazatlan, leaving more than 5000 people homeless in 1965. l7 Baja California is another area that regularly suffers from eastern Pacific hurricanes. In September 1959, a major storm crossed over La Paz in a northeast direction and then recurved in a northwest direction. Severe damage to lives and crops occurred all along the eastern coast of the peninsula, particularly in the small towns of Murlege, La Purissima, and San Ignacio. 18 b. Effects of Hurricanes Damage from hurricanes stems from three primary effects: flooding from heavy rains, storm surge, and hurricane winds. Depending on the location and type of settlements, physical structures, or economic activities, the proportionate impacts of these effects differ. The most severe impact of a hurricane typically falls on the following areas: Urban squatter settlements. Rural settlements. Physical infrastructure: roads, railroads, and communication lines. Subsistence and commercial agriculture. Fishing industry. Indirect effects on agro-industry. In terms of sheer numbers, the first of these — squatter settlements — is the most heavily affected. The destruction of lives and properties to these poor communities is the most marked effect of hurricanes VII-12 that hit the major coastal cities of Mexico. Often constructed of temporary, makeshift materials that are poorly anchored down and seriously lacking in adequate drainage and sewerage disposal systems, these communities are typically left in a shambles by the passage of a hurricane. Most buildings are deroofed, and water stagnates at levels as high as 3 feet for days before draining away. The unsanitary conditions that prevail in these areas, to say nothing of the almost complete destruction of property, reach a critical level; usually, an entire city is threatened with water contamination and epidemics. Furthermore, because these communities are often located along river beds or severely eroded hillsides, they are particularly susceptible to damages from flooding and landslides. Given that squatter settlements usually make up between 30 and 60 percent of the population in Mexico's rapidly growing cities, the hazard posed by hurricanes to coastal populations will continue to increase as the rural poor continue to migrate to large cities. This is particularly true of the Northwest region, with its high rate of urbanization. Even more so, the city of Acapulco, with its growth of more than 200 percent in 10 years, stands as a particularly high risk area were a major storm to strike. Rural settlements, although less concentrated and less susceptible to health hazards, still run serious risks if located alongside rivers, or near the coast below elevations of 5 to 15 feet above normal sea level. Furthermore, their isolation from urban rescue centers places them in greater danger, both before a storm, due to less access to e vacuation and warning sources, and after, due to inaccessibility by rescue operations. Although many rural house roofs are constructed with thatch or palm leaves, many are also made of tiles. Under severe storm conditions', the latter material is not nearly so safe as the former. Furthermore, depletion of forest resources in the Pacific zone, has caused moderate to severe soil erosion in most of the states, particularly Baja California, Sonora, southern Sinaloa, and most of the central Coast. 19 The danger to people and crops from landslides is therefore considerable in the rural areas . VII-13 The economic damage caused by a hurricane is mostly concentrated on agriculture and the fishing industry (if the cost of repairs to physical transport and communications infrastructure are ignored) . North Pacific hurricanes typically strike the Northwest region when crops of cotton, rice, beans, soya, and sesame are vulnerable to damage from wind and flooding. Because many of these crops are close to irrigation canals or rivers, the severe rains that accompany hurricanes often flood these crops. Sometimes crops of a given irrigation region are wiped out, such as the 23,000 hectares of cotton that were destroyed in the Valle de Santo Domingo (Baja California) following Hurricane Doreen in August 19 77. 20 Similarly, large hurricanes seriously damage coastal fishing fleets. Hurricane Hazel in 1965 destroyed 30 percent of the shrimp fleet off the coast of Sinaloa and the shrimp industry was left in a state of crisis. 21 This industry, which produces more than $20 million a year in income for the region — almost completely in foreign exchange — is a critical element of the regional economy. Indeed, serious direct damages to either agriculture or fishing will multiply several times through other sectors because the regional economy is highly dependent on its primary sector as an economic base. Almost all its industry, for example, is of a processing type. c. Impacts Within the Experimental Area Seeding will be limited to storms predicted to have a probability of 10 percent or less of coming within 50 nautical miles of a populated land mass within 24 hours of seeding. The seeding rules may be further modified by local conditions. No direct adverse social impacts from this experiment will occur if the seeding causes no modified hurricanes to strike land masses that are populated or used for social or economic activity . VII-14 If the center of a modified hurricane strikes a populated area, the damage will probably be smaller than that caused by the storm if it had not been seeded. If the area is struck by slightly increased peripheral winds, the damage may be somewhat greater than if seeding did not occur. Taken together, however, the net effect of decreased central winds and increased peripheral winds would be decreased damage because of the substantially greater strength of the winds near the storm center. It is currently believed that hurricane modification will not significantly change the total amount of rain that falls in a particu- lar geographic location during the storm. If this is the case, flooding or damage to crops, soil, and man-made structures will be approximately the same with or without seeding. However, any widening of the rainfall pattern could increase negative impacts by spreading damage to crops and population settlements. Loss of life, destruction of towns, and losses in economic activity, including commercial fishing operations, all occur during hurricanes and tropical storms. However, it will be difficult or impossible to measure the extent to which damage is reduced or increased as a result of the experiment. The program may indirectly benefit the coastal area of Mexico by increasing knowledge about storm activity in the North Pacific. Early identification of hurricane activity will allow improved warning systems and earlier preparation for emergency response. In the short term, however, this potential benefit may not be realized because existing communications systems in much of the coastal region are poor. However, it may encourage the development of a more wide- spread warning and rescue system. d. Impacts on Shipping and Aviation Routes through the Region There is substantial coastal and ocean bound shipping activity along Mexico's Pacific Coast (Table VII-2) . Ships are usually redirected when information is received about hurricanes along the VII-15 Table VII-2 OCEAN BOUND FREIGHT MOVEMENT FROM MEXICO (January-December 1975) Number of Ships Pacific Ports Foreign National Ensenada 64 — Isla de Cedros 67 — San Carlos 13 — La Paz 1 — San Marcos 52 — Guaymas 78 13 Mazatlan 139 1 Manzanillo 219 44 Lazara Cardenas 78 — Acapulco 79 24 Salina Cruz 21 5 Madero 6 1 Source: Navy Secretariat of Mexico General Director of Ports Operation, VII- 16 route. It is not known if further shifting of routes would be deemed necessary because of seeding of hurricanes. Shipping fleets at dock on the Mexican coast are at risk from both modified and unmodified hurricanes. Warnings often come too late to allow shifting of fleets out of the danger zone. However, because few hurricanes will be seeded during the experiment, any impact will be insignificant when compared with the effects of unseeded hurricanes on shipping routes. Aircraft normally avoid regions of known hurricane activity. Since the hurricane modification experiments will be conducted only in areas of known hurricane activity, they should have no impact on the routing of aircraft through the region. VII-17 VIII POTENTIAL ALTERNATIVES TO PROJECT The principal purpose of the Hurricane Amelioration Research Project is to test a hypothesis that seeding a tropical cyclone will reduce its peak winds, and hence reduce damages. Existing knowledge about tropical cyclones falls short of that needed to beneficially modify these storms on an operational basis. This Project is a scientific experiment designed to broaden the knowledge base, but it will not be sufficient to establish an operational modification capability. Rather, it will improve understanding of the dynamics of tropical cyclones and determine whether modification is feasible. Whether, or to what extent, cyclones should be routinely modified are public policy questions concerning the optimal way to adjust to these damaging storms. Other adjustment strategies are possible, in- cluding accepting losses, compensating for losses (with disaster insurance, for example), and reducing the potential losses (by limiting the vulnerable population and property, by improving building standards, and by improving forecasting and warning systems) . Because it is possible to reduce or compensate for potential losses, accepting losses is no longer a viable strategy. However, it is not known whether reduction and compensation strategies are necessarily potentially less effective than a modification strategy, especially in combination. The selection of an adjustment strategy is a public policy question involving issues and problems which are beyond the scope of this experimental project and this environmental impact assessment. Nevertheless, these issues and problems must be assessed prior to a decision on the need for an operational modification program. In the context of the experiment, there are four pertinent alter- natives to conducting the Hurricane Amelioration Research Project: do not perform, postpone, perform in another area, and use other modifica- tion techniques. VIII-1 A. Do Not Perform Pursuing this alternative would be appropriate if a decision were made that modification is not an appropriate way to reduce losses due to tropical cyclones. This alternative would preclude perhaps the only way to determine whether modification is a feasible adjustment strategy; field experiments to acquire knowledge of hurricane behavior and to test modification possibilities are essential to making this determination , A limited number of seeding experiments is unlikely to have signi- ficant cumulative impacts on man and nature. On the other hand, the possible benefits of improved knowledge about cyclones and of a demonstrated modification capability are considerable. Therefore, further research appears warranted. A decision not to perform any further experiments would be based on consideration of political, philosophical, or social issues, rather than scientific ones. The importance of these issues makes it necessary, if experiments are conducted, to give adequate attention to the time, place, and other conditions of the experiments. In addition, the full range of social and political, as well as scientific, factors must be considered in any decision about hurricane damage reduction strategies. B. Postpone Postponing the project would permit the gathering of more information to improve understanding of hurricane dynamics and possibly to resolve certain nonscientif ic issues. In addition, modeling efforts could be advanced to achieve better description of hurricanes and more accurate representation and assessment of modification experiments. It was estimated in October 1974 that "a suitable model for the study of typhoon modification could, optimistically, be available in 3 to 5 years." On the other hand, losses from hurricanes would continue, and the sooner successful modification can be demonstrated, the sooner losses can be reduced, if modification is acceptable as a course of action. Mon- itoring of unmodified hurricanes to obtain data for modeling efforts is VIII-2 essential, but experimental data from modification attempts are also necessary for the development of adequate models. It is felt that the most efficient way to proceed, in terms of resources and from a scientific point of view, is to conduct seeding operations simulta- neously with monitoring of unseeded storms and monitoring of unseeded periods of storms that are seeded. No significant benefit would be gained by delaying actual seeding operations. In fact, a delay could be quite costly in that one season might bring several seeding candidates, while another season might have no suitable storms for experimentation. C. Alternative Experimental Areas Tropical cyclones are common in six major oceanic regions around the world. Table VIII-1 lists these regions and the average annual number of hurricanes or typhoons. Modification experiments are possible only in these regions. To date, all modification experiments have been conducted in the North Atlantic Ocean because there is a considerably larger body of information about unseeded storms in this area than for other areas. In addition, this area is logistically convenient and the only oceanic region in which hurricanes can directly threaten the United States. Table VIII-1 ANNUAL AVERAGE NUMBER OF HURRICANES /TYPHOONS Average Annual Percent of Region Number Total North Atlantic 5.2 13.5 Eastern North Pacific 5.8 15.0 Western North Pacific 17.8 46.1 North Indian 2.2 5.7 Southwest Indian 3.8 9.8 Southwest Pacific and Australian 3.8 9.8 38.6 100.0 Source: Reference 2. VIII-3 The data in Table VIII-1 shows that more than three times as many typhoons occur in the western North Pacific as hurricanes in the North Atlantic. A similar ratio exists between the number of storms which are eligible for seeding experiments under the present eligibility rules. This scientific advantage is somewhat offset by the cost of establishing remote experimental operations. In addition, in the Western Pacific, many more sovereign states could be affected by the experiments. The population within and immediately adjacent to the experimental area is considerably smaller in the Western Pacific. How- ever, many storms which pass through the experimental area ultimately make a landfall on densely populated land masses. Proposed experimenta- tion in this area was not acceptable to certain countries of the region and, accordingly, there are no U.S. proposals for such projects in this area. Historically, the eastern North Pacific has experienced barely one- third the number of tropical cyclones that have occurred in the western North Pacific. However, for the period 1969-1973, there was no significant difference in the average annual numbers of storms eligible for seeding, days at hurricane intensity within a hypothetical experimental area, and possible seeding experiments. From a scientific point of view, Western Pacific typhoons are preferred targets. "* Their intensity (windspeed) at the possible time of experimentation is higher on the average. More intense storms are usually better defined, and a well organized storm is desired for experi- ments. Furthermore, Eastern Pacific storms tend to dissipate more quickly. This would complicate the evaluation of experimental seedings. Historical data also are more extensive for Western Pacific typhoons. Operational considerations include the difficulty of establishing and maintaining field operations and the time available to prepare for a possible experiment. The tropical storm stage can be considered as the period in which decisions about the experiment can be made and mobilization begun. Available data indicate no significant difference between past storms in the Western and Eastern Pacific. VIII-4 The geographic characteristics of the Eastern Pacific are decidedly advantageous compared with the Western Pacific. Only 23 percent of the eligible Eastern Pacific storms struck land, whereas 67 percent of the eligible Western Pacific storms ultimately made a landfall. This is significant because the risk of damage attributed fairly or unfairly to any experiment is reduced at the outset. Furthermore, the international aspects of the experiment involve only one other nation and the population at risk is smaller. In general, the North Atlantic has these same disadvantages with respect to the Eastern Pacific. D. Alternative Modification Techniques The basic postulate of tropical cyclone modification is that the addition of heat to certain regions of the storm will interrupt the inward spiral of air toward the eye and divert some of it to the outflow layer, thereby weakening the eye and reducing maximum winds. In the Hurricane Amelioration Research Project, the heat is added by distributing silver iodide particles in clouds containing supercooled water. The particles serve as nuclei on which water can freeze and in the process release its latent heat of fusion. This heat release initiates a chain of events in which additional heat is released by the condensation of water vapor. The alternatives to this technique are to use other seeding materials or to affect the energy flow in a different way. Silver iodide was the first material selected for use as a seeding agent in weather modification experiment and has become the most widely used agent. Ecological studies indicate that it does not have a significant adverse effect on the environment. Furthermore, the means of dispensing silver iodide are the most highly developed, and available pyrotechnic flares and aircraft dispensers are well suited to hurricane seeding requirements. One alternative modification technique is to apply a monomolecular film to the ocean surface to suppress evaporation and the transfer of heat and moisture from the ocean to the air within the storm. 5 ' 6 This VIII-5 technique is still in the early research stage. Effectiveness of a thin film under high winds and turbulent sea conditions has not yet been demonstrated, nor has an effective dispensing system been developed. Another approach to modification using carbon dust particles has been theoretically studied. 7 The hypothesis is that, when dispersed, the dust will cause localized heating, and the air converging into the eye of the storm will tend to rise before reaching the eye, thereby weakening the storm. It was estimated that to cause heating sufficient to reduce the storm's winds would require 10 to 20 C-5A aircraft at a cost of roughly $500, 000. 6 Further research on the technical and economic feasibility is needed. In the near term, no reasonable alternative to the proposed seeding technique exists. The relative merits of possible alternatives from an environmental point of view are uncertain. VIII-6 IX IRREVERSIBLE AND IRRETRIEVABLE COMMITMENTS OF RESOURCES The Hurricane Amelioration Research Project will require material resources for the production of silver iodide flares, bathythermographs, and dropwindsondes , and petroleum products for aircraft operations and other transportation. These materials will be consumed or expended, and consequently will be irreversibly and irretrievably committed. The financial resources applied to the project will also be irretrievably committed. IX- 1 APPENDIX A UPPER LIMITS FOR CONCENTRATIONS OF SILVER IODIDE A. Products of Combustion Each flare will produce 20 grams of Agl (9.2 grams of Ag, and 10.8 grams of I) . This material has an extremely low solubility in water , about 1570 micrograms per cubic meter (yg/m ) . Agl is stable in air and fresh water. In sea water it will slowly dissolve to form AgCl which is more soluble than Agl. It is possible that there could be products of incomplete combustion of the flare. These could include magnesium iodide (Mgl ) , aluminum iodide (All ) , or unreacted silver iodate (AglO ) . The Mgl and All will react slowly to form magnesium oxide (MgO) and aluminum oxide (A1„0 ) , and either free iodine or iodides. All of these reaction products are innocuous and of insignificant amounts compared with the enormous background amounts in soil and in the ocean. 3 The AgI0„ has a solubility of about 30 grams per cubic meter (g/m ) in fresh water. In sea water it will react to form silver chloride 3 (AgCl) with a solubility of about 0.9 g/m . B. Estimates of Resulting Concentrations of Reaction Products in Air and Water For the purposes of this analysis, it has been assumed that the flares will be dropped at 100 meter intervals (about 1 per second) and that they will burn for 5000 vertical meters. Therefore, the products of combustion of each flare will be distributed in a volume of about 50 million cubic meters (5000 x 100 x 100) . To calculate "worst case" concentrations, it has been assumed that there will be no further dilution of the reaction products beyond this volume. In actual storm conditions the atmospheric turbulence would dilute the products of combustion and A-l might well reduce the following estimated concentration by a factor of at least one thousand and possibly by one million. The maximum concentration of silver in the air, assuming that all the silver is in the air and that none is entrained in the rain, will 3 be 0.18 yg/m . The maximum concentration of iodine (as iodate or iodide in the air, again assuming no solution or entrainment in the rain, will 3 be about 0.22 yg/m . An average value for the amount of water in a nonprecipitating 3 cloud is about 0.5 g/m . With this amount of water present and all of the Agl assumed to be available for solution in the water droplets, the 3 maximum concentration of Agl in the droplets will be 1570 yg/m , as limited by its water solubility. This is equivalent to a silver con- 3 centration of 722 yg/m . The maximum concentration of silver from rain in the ocean is also 3 722 yg/m since the concentration cannot increase once the Agl is dis- persed in the cloud. The background concentration of silver in the 3 ocean is quite variable but it has been measured at around 300 yg/m . C. Effect of a Flare Falling in a Lagoon or on the Ground The criterion for pyrotechnic design is no more than 1 flare per 1000 blowing out or not igniting after it is dropped. It may be further assumed that, because of the Hurricane Amelioration Research Project operating criteria, the probability against there being a lagoon in the drop zone is of a similar order of magnitude. Therefore, the odds of a partially burned or unburned flare dropping in a lagoon are insignificant If a partially burned flare did land in a lagoon, it would dissolve very slowly because of the organic binder in the pyrotechnic mixture. As it dissolved, the AgI0~ would immediately react to form AgCl. Be- cause of the very slow rate of dissolving, it is probable that the AgCl would not precipitate but would remain in solution. A- 2 All lagoons have some internal circulation and ventilation. There- fore, as the flare dissolved, the silver would be diluted by the current. It is possible that the silver concentration in the immediate vicinity of the flare could reach toxic levels for stationary marine life. This is unlikely because of the very slow dissolving rate that is expected. No data exist on the dissolving rate of a flare in sea water, so it is not possible to give a quantitative estimate of the maximum concentration rate. If it is assumed that one flare completely dissolved in a lagoon 3 lkmxlkmx20m deep, the final concentration would be about 0.5 yg/m , which is 600 times less than the average background silver concentration in the ocean. An unignited flare is even less of a problem than a partially burned one because of the waterproof casing around the pyrotechnic, although it is not known how long it would take for this casing to be worn through and then how long it would take for the pyrotechnic to dissolve. Because of the Project's operational criteria, the probability of a flare landing in a populated area is even less than the probability of a flare landing in a lagoon. If a burning flare did land on the ground, there is a possibility of a fire. However, the land mass would be immediately beneath an active hurricane and any fire would probably be extinguished by high winds and/or torrential rains. Because it is insoluble, the Agl can never reach a toxic level unless it is eaten. A- 3 APPENDIX B OPERATIONAL APPLICATION OF THE LAND FALL CRITERION The basic input to the technique used to select hurricanes for ex- perimental seeding is a sample of forecast hurricane positions and their verifications. Various forecast methods (e.g., statistical, analog, and persistence) can be used. We propose to use the official forecasts of the National Hurricane Center. Figure B-l illustrates a sample of 431 such forecasts for hurricanes in the western Atlantic Ocean during the 10-year period, 1965-1974. To prepare the probability ellipses, we "normalize" all forecasts with respect to the direction of hurricane motion (OPQ) , and all verifying positions (dots) are located with respect to the forecast position (point P) . All initial positions of the hurricanes then lie along OP. After all verifying positions are so located, a bi-variate normal distribution is fitted to their pattern using the least-squares method. The resulting probability distribution of figure B-l is typical in that the centroid of the ellipse is somewhat behind the forecast position; that is, actual movement is slightly less than predicted movement. Application of the probability ellipses so constructed to the selection of hurricanes for experimental seeding involves constructing probability ellipses for forecast storm motion on transparent overlays for a specific map projection and scale and overlappiig the appropriate ellipse on the map with point P at the forecast position and OPQ oriented along the direction of hurricane motion. A hypothetical land area plus a surrounding 50 mile buffer zone in the path of the hurricane are shown in figure B-2 The probability of the storm center verifying within 50 miles of the potentially affected land is then determined by obtaining the sum of all probability sectors within this land-plus-bif fer area. In the case illustrated, this accumulated probability is about 5% (18 of the 0.25 unit areas of probability equals 4.5%). B-l The selection criteria then involves a particular forecast technique for a specified time period with eligibility depending upon a given accumulated probability of the forecast verifying within a given distance of land. We propose to use the official NHC forecasts for 24 hours with an eligible storm having less than 10% accumulated probability of verifying within 50 miles of land. B-2 STORM MOTION = approx. 60 n. mi. Figure B-1. Samples of 431 Forecasts in the Atlantic Ocean (1965-1974) B-3 Ql STORM MOTION 50-mile Buffer Zone Area Equal to 0.25% Probability = approx. 60 n. mi. Figure B-2. Hypothetical Land Area in the Path of a Hurricane B-4 REFERENCES* A. T. P. Miller, et al., 'Environmental Impact Assessment of Project STORMFURY-Pacific, Stanford Research Institute, Final Report, Contract No. 03-5-022-98 (April 1976). Section I 1. R. C. Gentry, "Hurricane Modification," in Hess, W. (ed.) "Weather and .Climate Modification," New York, John Wiley and Sons (1974). 2. NOAA, Environmental Research Lab., "ERL Program Plan, Program Development Plan: STORMFURY-Americas" (Draft), Office of Weather Modification, Boulder, Colorado (as augmented by discussions with NOAA personnel) . Section II 1. NOAA, Environmental Research Lab., "ERL Program Plan, Program Development Plan: STORMFURY-Americas" (Draft) , Office of Weather Modification, Boulder, Colorado (as augmented by discussions with NOAA personnel) . 2. R. C. Gentry, "Hurricane Modification," in Hess, W. (ed.) "Weather and Climate Modification," New York, John Wiley and Sons (1974). 3. R. C. Sheets, "Project STORMFURY: Hurricanes — Can Their Destructive Force be Reduced," Mariner's Weather Log (1975). 4. N. E. LaSeur, Summary and Recommendations from Evaluation Sub-group, STORMFURY Design Workshop, National Hurricane and Experimental Meteorological Laboratory, Coral Gables, Florida (July 1974). Section III 1. H. U. Sverdrup, M. W. Johnson, and R. H. Fleming, "The Oceans," Prentice-Hall, Inc., New York, 1087 pp. (1942). 2. G. Dietrich, "General Oceanography," trans, by F. Ostapoff, Inter- science Publishers, New York, 588 pp. (1963). *References, when not specifically mentioned in the text by author and year, are numbered consecutively for each section in the report. Those references mentioned by author and year are listed alphabetically by report section. "+Reference A is used for this document in all sections of this report. R-l Ill 3. G. Neumann and W. J. Pierson, "Principles of Physical Oceanography," Prentice-Hall, Inc., Englewood Cliffs, N.J., 545 pp. (1966). 4. Climatological and Oceanographic Atlas for Mariners, Volume II, North Pacific Ocean, prepared by Office of Climatology and Oceano- graphic Analysis Division, Washington, D.C. (1961). D. L. Anderson, "The San Andreas Fault," Sci. Am., (November 1971); (reprinted in Continents Adrift , W. H. Freeman and Co. , San Francisco, CA. 1972). G. A. Bartholomew, "Seal and Sea Lion Population of the California Islands," In, R. N. Philbride, ed., Proc. of the Symp . on the Biology of the California Islands, Santa Barbara Botanical Gardens (1967) 5. Carlquist, "Island Life: A Natural History of the Islands of the World," Doubleday & Co., New York, 451 pp. (1965). H. F. Cline, Mexico, Revolution to Evolution: 1940-1960 , Oxford Univer- sity Press, New York (1963). CRFSC, "Comprehensive Framework Study, California Region," Appendix V, "Water Resources." Report of the California Region Framework Study Committee for the Pacific Southwest Interagency Committee Water Resources Council (1970) . L. H. DiSalvo and H. T. Odum. "Coral Reefs," In, J. T. Odum, G. J. Copeland & E. A. McMahan, ed., Coastal Ecological Systems of the United States, 372-441 pp. (1974). E. Doran, Handbook of Selected Pacific Islands , Pacific Missile Range, Misc. Publ. No. PMR-MP-59-30 (1959). H. E. Dregne, "Appraisal of Research on Surface Materials of Desert Environments," In, W. G. McGinnies, B. J. Goldman, and P. Paylore, eds.. Deserts of the World: An Appraisal of Research Into Their Physical and Biological Environments , University of Arizona Press, 301-377 pp. (1968). A. J. Eardley, Structural Geology of North America . Second Edition, Harper and Row, New York (1962). FAO, Soil Map of the World . Vol. Ill, Mexico and Central America, FAO- UNESCO, Paris (1975). J. R. neirtzler, "Sea-Floor Spreading," Sci. Am. , (December 1968). (reprinted in Continents Adrift , W. H. Freeman and Co., San Francisco, CA. (1972). R-2 Ill, IV J. W. Porter, "Community Structure of Coral Reefs on Opposite Sides of the Isthmus of Panama," Science, 186:543-545 (1974). R. J. Renard and W, N. Bowman, "The Climatology and Forecasting of Eastern North Pacific Ocean Tropical Cyclones," Technical Paper No. 7-76 (NAVENVPREDRSCHFAC) , Naval Postgraduate School (July 1976). J. C. Sadler, "Tropical Cyclones of the Eastern North Pacific as Revealed by TIROS Observations," J. Appl. Meteor., Vol. 3, 347-366 pp. (1964), J. M. Savage, "Evolution of the Insular Herpetof aunas , " In, R. N. Phil- brils, ed., Proc. of the Symp . on the Biology of the California Islands, Santa Barbara Botanical Gardens (1967). V. E. Shelford, "The Econology of North America," University of Illinois Press, Urbana, 610 pp. (1963). F. Shreve , "Vegetation of the Sonoran Desert," Carnegie Institute of Washington, Publ. 591 (1951). UNESCO. Discharge of Selected Rivers of the World , Vol. Ill, Part II, Studies and Reports in Hydrology No. 5 (1974). U. S. Fish & Wildlife, "Threatened Wildlife of the United States," Bureau of Sport Fisheries and Wildlife, U. S. Department of Interior, Resource Publication 114 (1973). H. J. Wiens , Atoll Environment and Ecology , Yale University Press, New Haven, Conn. (1962). E. W. Wilson. Lower California and Its Natural Resources , Memoirs of the National Academy of Sciences, Vol, XVI, First Memoir (1922). Section IV 1. G. D. Atkinson, "Forecaster's Guide to Tropical Meteorology," AWS Tech. Rept. 240, Air Weather Service (MCA) USAF (April 1971). 2. W. M. Gray, "Global View of the Origin of Tropical Disturbances and Storms," Monthly Weather Review, Vol. 96, pp. 669-700 (Oct. 1968). 3. H. Riehl, "On the Formation of Typhoons," Journal of Meteorology, Vol. 5, pp. 247-264 (1948). 4. M. Yanai, "A Detailed Analysis of Typhoon Formation," Journal of Meteorological Society of Japan, Vol. 39, No. 4., pp. 187-214 (1961), 5. J. C. Sadler, "On the Origin of Tropical Vortices," In Proceedings of the Working Panel on Tropical Dynamic Meteorology, Navy Weather Research Facility, NWRF 12-1167-132, pp. 39-76 (August 1967). R-3 IV 6. W. M. Gray, "Tropical Cyclone Genesis," Atmospheric Sciences paper No. 234, Colorado State University, Fort Collins, Colorado (1975). 7. C. E. Palmer et al., "The Practical Aspect of Tropical Meteorology," Air Force Surveys in Geophysics, No. 76, AFCRC TN-55-220, GRD, Air Force Cambridge Research Center, Bedford, Massachusetts (Sep- tember 1955) . 8. G. D. Atkinson and C. R. Halliday, "Tropical Cyclone Minimum Sea Level Pressure—Maximum Sustained Wind Relationship for the Western North Pacific," FLEWEACEN Techn. Note JTWC 75-1, U.S. Fleet Weather Central Guam, Box 12 COMNAVMARIANAS , F.P.O., San Francisco, California (May 1975). 9. N. E. LaSeur and H. F. Hawkins, "An Analysis of Hurricane Cleo (1958) based on Data from Research Reconnaissance Air Craft," Monthly Weather Review, Vol. 91, pp. 694-709 (1963). 10. T. T. Fujita et al. , "A Model of Typhoons Accompanied by Inner and Outer Rainbands," Journal of Applied Meteorology, Vol. 6, No. 1, pp. 3-19 (February 1967). 11. D. J. Novlan and W. M. Gray, "Hurricane Spawned Tornadoes," Monthly Weather Review, Vol. 102, No. 7, pp. 476-488 (1974). 12. C. J. Neumann, "The Effect of Initial Data Uncertainties on the Performance of Statistical Tropical Cyclone Prediction Models," NOAA Tech. Memo NWS SR-81 , NOAA National Weather Service, Scientific Serv- ices Div., Southern Region, Fort Worth, Texas (March 1975). 13. B. I. Miller, E. C. Hill, and P. P. Chase, "Revised Technique for Forecasting Hurricane Motion by Statistical Methods," Monthly Weather Review, Vol. 96, No. 8, pp. 540-548 (1968). 14. F. Sanders and R. W. Burpee, "Experiments in Barotropic Hurricane Track Forecasting," Journal of Applied Meteorology, Vol. 7, No. 3, pp. 313-323 (1968). 15. A. C. Pike, "Improved Barotropic Hurricane Track Prediction by Adjustment of the Initial Wind Field," NOAA Technical Memorandum NWS SR-66, 16 pp. (1972). 16. J. R. Hope and C. J. Neumann, "An Operational Technique for Relating the Movement of Existing Tropical Cyclones to Past Tracks," Monthly Weather Review, Vol. 98, No. 12, pp. 925-933 (1970). 17. C. J. Neumann, "An Alternate to the HURRAN Tropical Cyclone Forecast System," NOAA Technical Memorandum NWS SR-62 , Fort Worth, Texas, p. 24 (1972). R-4 IV 18. C. J. Neumann, et al. , "A Statistical Method of Combining Synoptic and Empirical Tropical Cyclone Prediction Systems," NOAA Technical Memorandum NWS SR-63, Fort Worth, Texas, p. 32 (1972). 19. C. J. Neumann and M. B. Lawrence, "Statistical-Dynamical Prediction of Tropical Cyclone Motion," NOAA Technical Memorandum NWS SR-69, p. 34 (1973). 20. R. J. Renard, S. G. Colgan, M. J. Daley and S. K. Rinard, "Forecasting the Motion of North Atlantic Tropical Cyclones by the Objective MOHATT Scheme," Monthly Weather Review, Vol. 101, No. 3, pp. 206-214 (1973). 21. U.S. Army Coastal Engineering Research Center, Shore Protection Manual, Vol. I (1973) . 22. I. Isozaki, "An Investigation on the Variations of Sea Level due to Meteorological Disturbances on the Coast of the Japanese Islands (VI)," Papers in Meteorology and Geophysics, Tokyo, Vol. 21(3), pp. 291-321 (October, 1970). 23. , "An Investigation on the Variations of Sea Level due to Meteorological Disturbances on the Coast of Japanese Islands (IV): Storm Surges on the Pacific and Okhotsk Sea Coasts of North Japan," Journal of the Oceanographical Society of Japan, Vol. 25, No. 1, pp. 191-200 (August 1969). 24. M. Miyazaki, "Characteristics of Storm Surges Induced by Typhoons Along the Japanese Coast," Proc. of the WMO Technical Conference, Manila, WMO No. 408 (15-18 October 1974). 25. S. Unoki, "A Note on the Storm Tide of the Japanese Coast," Coastal Engineering in Japan, Vol. 4, pp. 1-10 26. K. Waditi and T. Hirono, "Storm Tides Caused by Typhoons," Proc. of The UNESCO Symposium on Typhoons, pp. 31-38 (9-12 November, 1954). 27. B. R. Bodine, "Storm Surge on the Open Coast: Fundamentals and Simplified Prediction," Tech. Memo No. 35, U.S. Army Corps of Engi- neers, Coastal Engin. Res. Center, Washington, D.C. (1971). 28. C. L. Bretschneider , "Engineering Aspects of Hurricane Surge," Proc. Techn. Conf. on Hurricanes, Amer. Met. Soc, Miami Beach, Florida (1958). 29. D. L. Harris, "Hurricane Storm Surges," In "Proc, Hurricane Sympo- sium, Society for Oceanography, Houston, Texas, Oct. 1966," Sidney A. Stubbs (ed.), pp. 200-228 (1967). 30. , "Characteristics of the Hurricane Storm Surge," Tech. Paper No. 48, U.S. Weather Bureau, U.S. Dept. of Commerce, 139 pp. (1963). R-5 IV 31. N. S. Heaps, "Storm Surges on a Continental Shelf," Phil. Trans. Royal Soc, Series A, 257 ( ), pp. 351-383 (1965). 32. R. A. Hoover, "Empirical Relationships of the Central Pressures in Hurricanes to the Maximum Surge and Storm Tide," Monthly Weather Review, Vol. 85(5), pp. 167-174 (May 1957). 33. C. P. Jelesnianski, "A Numerical Calculation of Storm Tides Induced by a Tropical Storm Impinging on a Continental Shelf," Monthly Weather Review, Vol. 93(6), pp. 343-358 (1965). 34. , "Numerical Computations of Storm Surges without Bottom Stress," Monthly Weather Review, Vol. 94(6), pp. 379-394 (1966). 35. , "Numerical Computations of Storm Surges with Bottom Stress," Monthly Weather Review, Vol. 95(11), pp. 740-756 (1967). 36. , "Bottom Stress Time-History in Linearized Equa- tions of Motion for Storm Surges," Monthly Weather Review, Vol. 98(6), pp. 462-478 (1970). 37. , "SPLASH (Special Program to List Amplitudes of Surges from Hurricanes) : II. General Track and Variant Storm Con- ditions," NOAA Tech. Memo NEWS/TDL-52, 55 pp. (1974). 38. M. T. Miyazaki, T. Ueno , and S. Unoki , "Theoretical Investigations of Typhoon Surges Along the Japanese Coast," Oceanogr. Mag., 13(1), pp. 51-75 (1961). 39. J. W. Nickerson, "Storm-Surge Forecasting," NAVWEARSCHFAC Technical Paper No. 10-71, 44 pp. (1971). 40. P. Welander, "Wind Action on a Shallow Sea: Some Generalizations of Ekman's Theory," Tellus, Vol. 9(1), pp. 45-52 (1957). 41. , "Numerical Prediction of Storm Surges," Advances in Geophysics, 8, pp. 316-379, Academic Press, N.Y. (1961). 42. P. Groen and G. W. Groves, "Surges," The Sea, M. N. Hill (ed.), Vol. 1, Interscience Publishers, New York, pp. 611-646 (1963). 43. A. T. Dodson, "Tides and Storm Surges in a Long Uniform Gulf," Proc. Roy. Soc. London, Series A, 237 ( ) (1956). 44. J. E. Overland, "The Response of Selected Bays and Estuaries to Hurricane Storm Surge: I. Aspects of a Two-Dimensional Shallow 3ay Model," NOAA/NWS/TDL Prelim. Report (1974). R-6 IV 45. , "The Response of Selected Bays and Estuaries to Hurricane Storm Surge: Apalachicola Bay, Florida," NOAA/NWS/TDL Prelim. Report (1974) . 46. J. Proudman, "The Propagation of Tide and Surge in an Estuary," Proc. Roy. Soc. London, 231, pp. 8-24 (1955). 47. , "Oscillation of Tide and Surge in an Estuary of Findp Length," J. Fluid Mech. , 2 ( ) , pp . 371-382 (1957). 48. R. 0. Reid, "Forced and Free Surges in a Narrow Basin of Variable Depth and Width: A Numerical Approach," Tech. Rept . 57-25T, Dept. of Oceanogr. and Meteor., Texas A&M Univ., College Station, Texas, 60 pp. (1957). 49. , and B. R. Bodine, "Numerical Model for Storm Surges in Galveston Bay," J. Waterways and Harbors Div. , Proc. ASCE (WW1) , • pp. 33-57 (1968). 50. S. Unoki and I. Isozaki, "On the Effect of a Dike with Openings on the Storm Surge Caused by the Typhoon," Coastal Engin. in Japan, 6 ( ), pp. 57-65 (1963). 51. S. Brand, J. W. Blelloch, and D. C. Schertz, "State of the Sea Around Tropical Cyclones in the Western North Pacific Ocean," Environmental Prediction Research Facility Technical Paper No. 5-73 (June 1973). 52. M. D. Earle , "Extreme Wave Conditions During Hurricane Camille," J. of Geophysical Research, Vol. 80, No. 3. pp. 377-379 (1975). 53. S. Kikuchi , "Wave Height of Swells and Relative Location of Typhoon in the Neighborhood of Kyushu," J. of Meteorological Research, Tokyo, Vol. 21(6), pp. 309-319 (June 1969). 54. M. M. Patterson, "Hind casting Hurricane Waves in the Gulf of Mexico," Offshore Technology Conference Paper No. 0TC 1345 (1971). 55. S. Unoki, "On the Ocean Waves Due to Tropical Cyclones," Coastal Engineering in Japan. 56. U.S. Army Coastal Engineering Research Center, Shore Protection Manual , Vol. 1 (1973). 57. H. L. Hansen, "The Climatology and Nature of Tropical Cyclones of the Eastern North Pacific Ocean," Thesis, Naval Postgraduate School, Monterey, California (September 1972). R-7 IV, V 58. c. P. Jelesnianski, "SPLASH (Special Program to List Amplitudes of Surges from Hurricanes): I. Landfall Storms," NOAA Tech. Memo, NWS/TDL-46 (1972). Section V 1. 42 U.S.C. Sections 4321-4347. 2. Id., Section 102(2) subsection c. 3. 38 Fed. Reg. 20550 Appendix, Effective date January 28, 1974. 4. Id., Section 1500.2 (b)(3). 5. E. B. Weis and W. H. Lambright, "Policy Determinants of Weather Modification," In American Meteorological Society, 4th Conference on Weather Modification, p. 538 (November 18-21, 1974). 6. An Act to Provide for the Reporting of Weather Modification Activities to the Federal Government, P. 0. 92-205, 15 U.S.C. Sec- tion 330. 7. President of the United States State of the Union Message, February 15, 1973. Regulations effective 15 February 1975, Federal Register (1974). 8. United Nations Charter: U.S. Ratification June 26, 1945 (as amended 1965) 59 Stat. 1031. T. S. (Treaty Series) 993. 9. Id., Chapter XIV, Articles 92-92 and Statute of the International Court of Justice. 10. International Court of Justice, "Compulsory Jurisdiction," Declara- tion by the President of the United States (August 14, 1946), 61 Stat. (2) 1218, T.I.A.S. (1598). 11. Convention on the Safety of Life at Sea, 50 Stat. 1121, 2 Bevans 782 (May 31, 1929). 12. Id., Article 35. R-8 V 13. Convention on the Intergovernmental Maritime Consultative Organiza- tion, 9 U.S.T. 621, T.I.A.S. 4044 (March 6, 1948; signed March 17, 1958). 14. Id., Article 1(a). 15. Id. , Article 1(d). 16. Convention of the World Meteorological Organization; 18 U.S.T. 2795, 2800; T.I.A.S. 6364. 17. Id. , Article 2(d). 18. CM. Hassett, "Weather Modification and Control," In Texas Inter- national Law Journal, Vol. 7, No. 1, p. 108 (1971). 19. World Meteorological Organization, "Second Report on the Advancement of Atmospheric Science and Their Application in the Light of Develop- ments in Outer Space" (1963) . 20. Convention on the High Seas, 13 U.S.T. 2313, T.I.A.S. No. 5200, 1958 (1962). 21. Id. , Article 2. Section V — Bibliography J. Eugene Haas, "Social Aspects of Weather Modification," Bulletin American Meteorological Society, Vol. 54, No. 7, July 1973. National Science Foundation, Human Dimensions of the Atmosphere , U.S. Government Printing Office, Washington, D.C., 1968. Lowell Ponte, "Using Weather Modification as a Weapon of War," San Francisco Sunday Examiner and Chronicle, Sunday Punch, p. 2, Feb. 8, 1976. W.R.D. Sewell, et al. , Modifying the Weather: A Social Assessment , Uni- versity of Victoria, British Columbia, 1973. " Arnold L. Sugg, "Economic Aspects of Hurricanes," Monthlv Weather Review, p. 147 (March 1967). Webster's Geographical Dictionary , Revised Edition, Merriam, Springfield Mass. , 1960. ' R-9 Section VI 1. R. C. Gentry, "Hurricane Modification," in Hess, W. (ed.), Weather and Climate Modification , New York, John Wiley and Sons (1974~7"! 2. R. C. Sheets, personal communication (1975). 3. R. C. Sheets, "Analysis of Hurricane Debbie Modification Results Using the Variational Optimization Approach," Mon. Wea. Rev., Vol. 101, pp. 663-634 (September 1973). 4. S. L. Rosenthal and M. S. Moss, "Numerical Experiments of Relevance to Project Stormfury," NOAA Tech. Memo ERL NHRL-95, National Hurricane Research Laboratories, Coral Gables, Florida (December 1971). 5. M. A. Estoque, "Hurricane Modification by Cloud Seeding," In Proceedings of International Conference on Weather Modificaion, Canberra, Australia, American Meteorological Society, pp. 154-160 (September 1971) . 6. H. Sundquist, "Model Tropical Cyclone Behavior in Experiments Related to Modification Attempts," Tellus, Vol. 24, pp. 6-12 (1972). 7. R. A. Anthes, "Development of Asymmetries in a Three-dimensional Model of a Tropical Cvclone," Mon. Wea. Rev., 100, pp. 461-476 (1972). 8. T. Kirshnamurh, et al., "Florida State Universities Tropical Prediction Model," Tellus, 15, pp. 523-535 (1973). 9. R. W. Jones, "A Preliminary Estimate of a Tropical Cyclone Change Caused by Artificial Enhancement of Convective Heat Release," Project Stormfury Annual Report 1973-74, National Hurricane and Experimental Meteorological Laboratory, NOAA, Coral Gables, Florida (in preparation). 10. A. L. Hammond, "Hurricane Prediction and Control: Impact of Large Computers," Science, Vol. 181, pp. 643-646 (August 17, 1973). 11. R. C. Sheets, "Application of Bayesian Statistics for Stormfury Results," Project Stormfury Annual Report 1969, National Hurricane Research Laboratory, Coral Gables, Florida (May 1970). 12. R. C. Gentry, "The Hurricane Modification Project: Past Results and Future Prospects," Project Stormfury, Annual Report, 1969, National Hurricane Research Laboratory, Coral Gables, Florida (May 1970). 13. P. G. Black, "Use of Echo Velocities to Evaluate Hurricane Modifica- tion Experiments," In. Project Stormfury Annual Report 1970, National Hurricane Research Laboratory, Coral Gables, Florida (May 1971). R-10 VI 14. S. L. Rosenthal, "A Circularly Symmetric Primitive Equation Model of Tropical Cyclones and Its Response to Artificial Enhancement of the Convective Heating Functions," Mon. Wea. Rev., Vol. 99, pp. 414- 426 (1971). 15. H. F. Hawkins, "Comparison of Results of the Hurricane Debbie (1969) Modification Experiments with those from Rosenthal's Numerical Model Simulation Experiments," Monthly Weather Review, Vol. 49, No. 5, pp. 427-434 (May 1971). 16. W. M. Gray, et al. , "Interim Report to NOAA National Hurricane Research Laboratory on CSU Typhoon Research in Support of Project Stormfury," Grant No. 04-5-022-14, Dept. of Atmospheric Sciences, Colorado State Univ., Fort Collins, Colorado (September 1975). 17. J. E. George, "Tropical Cyclone Motion and Surrounding Parameter Relationships," Atmospheric Science Paper No. 241, Colorado State Univers., Fort Collins, Colorado (December 1975). 18. R. C. Sheets, personal communication (1975). 19. C. J. Neumann et al., "A Statistical Method of Combining Synoptic and Empirical Tropical Cyclone Prediction Systems," NOAA Tech. Memo NWS SR-63, NOAA National Weather Service, Scientific Services Div. , Southern Region, Fort Worth, Texas (1972). 20. R. W. Jones, personal communication, 1975, 21. R. W. Knight and R. C. Sheets, "Effects of Seeding on Storm Motion," Project Stormfury, Annual Report 1973-74, National Hurricane and Experimental Meteorological Laboratory, Coral Gables, Florida (in preparation). 22. P. G, Black et al . , "Variations in the Eye Configuration of Hurricane Debbie on 18 and 20 August, 1969," Project Stormfury Annual Report 1969, National Hurricane Research Laboratory, Miami, Florida (May 1970). 23. F. Sanders, "Dynamic Forecasting of Tropical Storm Tracks," Trans. N.Y. Acad. Sci. Ser. II, Vol. 32, pp. 495-508 (1970). 24. W. M. Gray, et al., "Interim Report to NOAA National Hurricane Research Laboratory on CSU Typhoon Research in Support of Project Stormfury," Grant No. 04-5-022-14, Dept. of Atmospheric Sciences, Colorado State Univ., Fort Collins, Colorado (September 1975). 25. R. C. Sheets and R. W. Knight, "Simulations of Hurricane Rainfall Distribution Changes Based on the Stormfury Hypothesis," Project Stormfury Annual Rept. 1973-74, National Hurricane and Experimental Meteorological Laboratory, Coral Gables, Florida (in preparation). R-ll VI 26. T. T. Fujita et al., "Typhoon-Associated Tornadoes in Japan and New Evidence of Suction Vortices in a Tornado Near Tokyo," J. of Meteor. Soc. of Japan, Vol. 50, pp. 431-453 (1972). 27. D. J. Novlan and W. M. Gray, "Hurricane Spawned Tornadoes," Monthly Weather Review, Vol. 102, No. 7, pp. 476-488 (1974). 28. C. P. Jelesnianski and A. D. Taylor, "A Preliminary View of Sccrm Surges Before and After Storm Modifications," NOAA Technical Memo. No. ERL WMPO-3 (May 1973). 29. C. S. Barrientos and C. P. Jelesnianski, "Some Numerical Experiments of the Effects of Hurricane Modifications on Storm Surges," Proc. of the WMO Technical Conference, Manila, WMO No. 408 (15-18 October 1974). 30. S. L. Rosenthal ? Personal Communication (19 September 19 77 j 31. William J. Danney, "Cool Inflow as a Weakening Influence on Eastern Pacific Tropical Cyclones," NOAA Technical Memo, NWS WR-100 , December 19 76, 4 pp. 32. D. B. Ross, "A Simplified Model for Forecasting Hurricane-Generated Wind Waves," Bull. Am. Meterol. Soc . 52(1) (January 1976). 33. V. J. Cardone and D. B. Ross, "State of the Art Wave Prediction Methods and Data Requirements," paper presented at the NOAA Ocean Wave Climate Symposium, Washington, D. C. (July 1977). Section VII 1. National Academy of Sciences, Committee on Water Quality Criteria, Environmental Studies Board, "Water Quality Criteria 1972." '.. Environmental Protection Agency, Office of Air and Water Programs, Office of Air Quality Planning Standards, "Air Pollution Engineer- ing Manual" (Second Edition) (May 1973). 3. H.J.M. Bowen, "Trace Elements in Biochemistry," Academic Press, New York, 24 p. (1966). 4. U.S. Department of Defense, "The Effects of Nuclear Weapons," S. Glass- tone (ed.), P. J. Dolan (assoc. ed . ) , U.S. Atomic Energy Commission (in preparation) . 5. D. I. Blumenstock, "A Report on Typhoon Effects Upon Jaluit Atoll, I. Wind, Wave and Storm Conditions at Jaluit, January 7-8, 1958," Atoll Research Bulletin No. 75:5-20 (1961). R-12 VII 6. D. W. Boyd, R. A. Howard, J. E. Matheson, and D. W. North, "Decision Analysis of Hurricane Modification," Stanford Research Institute (June 1971). 7. U.S. Fleet Weather Central Joint Typhoon Warning Center, "Annual Typhoon Report (s)," augmented by New York Times newspaper accounts (1959-1974). 8. 0. I. Blumenstock, F. R. Fosberg, and C. G. Johnson, "The Re-survey of Typhoon Effects on Jaluit Atoll in the Marshall Islands, Nature 189:618-620 (1961). 9 # W. A. Niering, "Terrestrial Ecology of Kapingamarangi Atoll, Caroline Islands," Ecol. Mongr. 33:131-160 (1963). 10. D. R. Stoddart, "Re-survey of Hurricane Effects on the British Honduras Reef and Cays, October 30-31, 1961," Nature 207:589-592 (1965). IX. D. I. Blumenstock, "Typhoon Effects at Jaluit Atoll in the Marshall Islands," Nature 182:1267-1269 (1958). 12. J. L. Gressitt, "A Report on Typhoon Effects Upon Jaluit Atoll: X. Terrestrial Fauna," Atoll Res. Bull. No. 75:69-73 (1961). 13. H. J. Wiens, "A Report on Typhoon Effects Upon Jaluit Atoll, III. General Description of Storm Effects," Atoll Research Bulletin 75:21-36 (1961). 14. F. R. Fosberg, "A Report on Typhoon Effects Upon Jaluit Atoll: VIII. Flora and Vegetation;' Atoll Research Bulletin 75:51-55 (1961). 15. D. R. Stoddart, " Effects of Hurricane Hattie on the British Honduras Reefs and Cays," October 30-31, 1961," Atoll Res. Bull. 95:1-142 (1963) 16. "North Pacific Hurricanes; Timid or Treacherous," Mariner's Weather Log 11(6) : November 1967. 17. Excelsior , September 27-28, 1965. 18. Excelsior , September 10, 1959. 19. Atlas of Mexico , Bureau of Business Research, University of Texas, Austin, Texas, 1975. 20. Excelsior , August 17, 1977. 21. Excelsior , September 28, 1965. R-13 VII 22. J. E. Haas, "Social and Political Aspects of Planned Weather Modification," _In Bulletin of the American Meteorological Society, Vol. 49, No. 2, p. 158 (February, 1968) . 23. R. C. Gentry, "Modifying the Greatest Storm on Earth — The Hurricane," In Underwater Science and Technology Journal, p. 204 (December, 1970). 24. L.F.E. Goldie, "Liability for Damage and the Progressive Development of International Law," In International and Comparative Law Quarterly (1965) f.n.p. 1264, citing Rosenthal, et al., Catastrophic Accidents in Government Programs, p. 30 (1963). 25. M. A. Estoque, "Hurricane Modification by Cloud Seeding," presented at The International Conference on Weather Modification, Canberra, Australia, p. 157 (emphasis added) (September, 1971). 26. R. C. Gentry, "Modifying the Greatest Storm on Earth — The Hurricane," In Underwater Science and Technology Journal, p. 260 (December, 1970). 27. M. Miyazaki, "Characteristics of Storm Surges Induced by Typhoons Along the Japanese Coast," I_n Proceedings of the WMO Technical Con- ference (Manila, 15-18 October 1974), Geneva: Secretariat, WMO, p. 42 (emphasis added) (1975). 28. R. H. Simpson, "Hurricane Prediction: Progress and Problem areas," In Science, Vol. 181, No. 4103, p. 899 (September 7, 1973). 29. Corfu Channel Case, I.C.J, p. 4, 22 (1949). 30. Trail Smelter Arbitration; see American Journal of International Law, Vol. 35, p. 684 (1941). 31. H. J. Taubenfeld, "International Environmental Law: Air and Outer Space," In Natural Resources Journal, Vol. 13, p. 324 (April, 1973). 32. J. W. Samuels, "International Control of Weather Modification Activities: Peril or Policy?" In Natural Resources Journal, Vol. 13, p. 335 (April, 1973). 33. W. H. Haubert, "Toward Peaceful Settlement of Ocean Space Disputes: A Working Paper," In San Diego Law Review, Vol. 11 p. 745 (1974). 34. J. P. Chamberlain, "Individual Rights and Space Liability," In ABA Journal, Vol. 58, p. 60 (January, 1972). 35. P. St. Amand, quoted in R. F. Taubenfeld and H. J. Taubenfeld, "Some International Implications of Weather Modification Activities," In International Organization, Vol. 23, No. 4, p. 812 (Autumn, 1969). R-14 VII 36. L. Ponte, "Using the Weather As a Weapon of War," In San Francisco Examiner and Chronicle, Sunday Punch, p. 2 (February 8, 1976). 37. Testimony of Congressman Gilbert Gude (Republican-Maryland) Sub- committee on International Organizations and Movements, House Com- mittee on Foreign Affairs, Weather Modification As A Weapon of War, 93rd Congress, Second Session, p. 5 (September 24, 1974). 38. G. MacDonald, "Weather Modification As A Weapon," I_n Technology Review, p. 59 (October/November 1975) . 39. L. Ponte, "Using the Weather As A Weapon of War," In San Franciso Examiner and Chronicle, Sunday Punch, p. 2 (February 8, 1976). 40. E. B. Weiss and W. Lambright, "Policy Determinants of Weather Modi- fication," In American Meteorological Society, Fourth Conference on Weather Modification, Ft. Lauderdale, Florida, p. 539 (November 18-21, 1974). Section VIII 1. S. L. Rosenthal, "Numerical Simulation of Typhoon Modification Ex- periments," In Typhoon Modification, Proceedings, World Meteoro- logical Organization Technical Conference (October, 1974). 2. H. L. Crutcher and R. G. Quayle, "Mariners Worldwide Climatic Guide to Tropical Storms at Sea," NAVAIR 50-1C-61 (March, 1974). 3. NOAA, Environmental Research Lab., "ERL Program Plan FY 74-1, Program Development Plan: Stormfury-Pacif ic 1977-78" (Draft), Office of Weather Modification, Boulder, Colorado (March 25, 1974). 4. H. F. Hawkins and S. M. Imbembo, "Seeding Opportunities for Project STORMFURY in the Western Pacific (1964-1973) and Eastern Pacific (1969-1973)," Project Stormfury Annual Report 1973-74 (in preparation). 5. W. D. Mallinger and T. P. Mickelson, "Possible Use of Monomolecular Films in Hurricane Abatement," Mariners Weather Log, 17, pp. 277-281 (September 1973) . 6. U.S. Department of Commerce, National Oceanic and Atmospheric Admin- istration, Office of Environmental Monitoring and Prediction, "Summary Report: Weather Modification, Fiscal Year 1973," Rockville, MD (December, 1974). 7. W. M. Gray, "Feasibility of Beneficial Hurricane Modification by Cargon Dust Seeding," Department of Atmospheric Science, Colorado State University, Fort Collins, Colo. (1973). R-15 ftU.S. GOVERNMENT PRINTING OFFICE: 19 78-261-238/24 A PENN STATE UNIVERSITY LIBRARIES minium i I A0Q0070TMD275 _■