<%BANNER%>

Economic Impact of a Mediterranean Fruit Fly Outbreak in Florida


PAGE 1

ECONOMIC IMPACT OF A MEDITERRANEAN FRUIT FLY OUTBREAK IN FLORIDA By RAPHAEL YVES PIERRE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

Copyright 2007 by Raphael Yves Pierre

PAGE 3

To my wife, Alaine and my son, Andy.

PAGE 4

iv ACKNOWLEDGMENTS First and foremost, I am indebted to Drs. P. K. Nair and Michael Bannister for providing me the opportunity of beginning my do ctorate studies in the School of Forest Resources and Conservation. I wish to express my sincere appreciation and gratitude to my entire supervisory committee: Drs. John VanSickle (chair); T homas Spreen (co-chair); Edward Evans; Norman Leppla; Moss Charles; James Seal e; and David Mulkey. They all provided useful criticisms and suggestions, thereby c ontributing greatly to the quality of this dissertation. Drs. John VanSickle and Edward Evans dese rve special mention for allowing me to complete my doctoral studies in the Depart ment of Food and Resource Economics. They granted me a graduate research assistants hip at a time when I was badly in need. Drs. Thomas Spreen, Norman Leppla, and Charles Moss deserve special recognition for their unwavering support and guidance. They gave freely of their time and advice. I have been very fortunate to have their e xpertise available to me. I wish also to recognize tremendous suppor t from Dr. Gary Steck, Dr. David Dean, Mr. Mike Shannon, Mr. Terry McGovern, Mr. R. E. Burns, and Mr. Loren Carpenter. I am very appreciative of the opportunity to have worked with them throughout the research process. They were willing to provi de the crucial information needed to conduct this study.

PAGE 5

v My life in Gainesville would have been dreary and unbearable without a friendly environment sustained by SFRC and FRED budd ies. I am especially grateful for the unwavering friendship of Lurleen Walters Alain Michel and Vony Petit-Frere. I am also lucky to have had the understa nding of a wonderful and supportive friend, Myrtha Jean-Mary. I am truly indebted to her for help and encouragement throughout my doctoral studies at the University of Florida. Finally, I would like to express my love and deepest admiration for my wife, Alaine Jean, and my son, Andy Pierre, to whom this dissertation is dedicated. Their moral support and commitment have inspired me to achieve this accomplishment.

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES..........................................................................................................xii ABSTRACT.....................................................................................................................xiii CHAPTER 1 INTRODUCTION........................................................................................................1 Problematic Situation....................................................................................................1 Problem Statement........................................................................................................3 Hypotheses....................................................................................................................3 Objectives.....................................................................................................................4 Thesis Outline...............................................................................................................4 2 REVIEW OF MEDFLY INTRODUC TIONS AND INFESTATIONS IN FLORIDA.....................................................................................................................6 Biological Profile of the Mediterranean Fruit Fly........................................................6 Overview of the Florida Medfly Detection and Eradication Program.......................10 Overview of the Florida Fr uit and Vegetable Sector..................................................14 Summary.....................................................................................................................18 3 POLICY FRAMEWORK OF THE SANITARY AND PHYTOSANITARY (SPS) AGREEMENT.................................................................................................24 The World Trade Organization Agreement on the Application of Sanitary and Phytosanitary Measures..........................................................................................24 Phytosanitary Protocols fo r the International Moveme nt of Fresh Fruits and Vegetables in Fruit Fly Free Areas.........................................................................33 Regulated Post-harvest Treat ments and Procedures for the Quarantine Control of Fruit Flies................................................................................................................39 Concluding Remarks..................................................................................................47

PAGE 7

vii 4 COST ANALYSIS OF THE MEDFLY DETECTION AND ERADICATION PROGRAM.................................................................................................................54 Specification of the Bayesian Modeling Framework.................................................54 Overview of the Bayesian Decision Process.......................................................54 Definition of the Variables..................................................................................56 Temporal dimension.....................................................................................57 Spatial dimension.........................................................................................58 Cost Function.......................................................................................................60 Future cost of eradication.............................................................................62 Optimization model......................................................................................62 Probabilistic Models...................................................................................................63 Probability of Detection: F (X1,t | X2,t, Z)............................................................63 Probability of Infestation: F(X1,0)......................................................................63 Multiple Trap Sensitivity of McPhail traps: F(X2,t | X1,t Z).............................65 Comparative Sensitivity of Mc Phail versus Jackson Traps................................67 Results........................................................................................................................ .69 Pest Population Projection...................................................................................70 Size and Cost of the Infestation...........................................................................71 Multiple Trap Sensitivities for ML Traps...........................................................74 Probabilities of Detection for ML Traps.............................................................75 Optimal Trap Densities........................................................................................75 Conclusions.................................................................................................................76 5 WELFARE ANALYSIS OF A MEDF LY OUTBREAK IN FLORIDA...................87 Fundamentals of the Partial Equilibrium Model........................................................87 Adaptation of the Spatial Equilibrium Mode l to the Fruit and Vegetable Industry...90 Grapefruit Model.................................................................................................90 Vegetable Model.................................................................................................91 Specialty Model...................................................................................................92 Cost Impact of a Medfly Quar antine Restriction on Florida...............................94 Empirical Results........................................................................................................96 Solutions of the Grapefruit Mode l under a Medfly Quarantine..........................97 Option I (with the market of Japan).............................................................97 Option II (without the market of Japan).....................................................100 Solutions of the Vegetable Mode l under a Medfly Quarantine.........................102 Tomatoes....................................................................................................103 Peppers.......................................................................................................105 Cucumbers..................................................................................................106 Squash........................................................................................................107 Eggplants....................................................................................................108 Watermelons...............................................................................................108 Strawberries................................................................................................109 Solutions of the Specialty Mode l under a Medfly Quarantine..........................109 Aggregate Impacts....................................................................................................111

PAGE 8

viii 6 POLICY RECOMMENDATIONS AND CONCLUSIONS....................................129 Areas for Alternative Policy Measures.....................................................................129 Assessing the level of Medfly Risk in Florida..................................................129 Assessing the effects of entry c onditions on the level of risk...........................130 Improving the single trap sensitivity.................................................................131 Improving the pesticide efficacy.......................................................................132 Summary and Conclusions.......................................................................................132 Limitations of the Study and Sugge stions for Further Research..............................135 APPENDIX A PEST POPULATION STRUCRURE IN MIAMI AND TAMPA...........................139 B GRAPEFRUIT MODEL..........................................................................................141 LIST OF REFERENCES.................................................................................................143 BIOGRAPHICAL SKETCH...........................................................................................158

PAGE 9

ix LIST OF TABLES page 2-1 Major fruit and vegetable crops grown in Florida according to their importance as Medfly hosts.........................................................................................................19 2-2 Costs of Medfly infestations in Florida....................................................................20 2-3 Baseline and emergency budgets for th e Medfly preventi on and detection program ................................................................................................................... 22 2-4 Important fruits: acreage, yiel d, and use (by crop), 2001-02...................................22 2-5 Important vegetables: acreage, yiel d, and utilization, by crop, 2001-02.................23 3-1 Formats for phytosanitary protocols for th e international moveme nt of fresh fruit and vegetable commodities......................................................................................50 3-2 Regulated postharvest treatments, adva ntages, limitations, a nd alternatives under consideration............................................................................................................52 4-1 Distribution of day de grees required by stage..........................................................77 4.2 Average monthly distances flown by diffe rent fractions of Medfly population......77 4-3 Eradication cost equations........................................................................................79 4-4 Hyperbolic tangent approximation of marginal probability function for multiple trapping sensitivity of McPhail traps.......................................................................79 4-5 Coefficients of comparative sensitiv ity and mean daily captures by period............80 4-6 Distribution of the expected populat ion size and generation time per location and per season at 50, 77, 98, and 119 days a of the infestation................................80 4-7 Distribution of the intr insic rates of increase and doubling times of the pest population per location and per season....................................................................81 4-8 Distribution of the infested area, qu arantine area, and eradication cost per location and per season at 50, 77, 98, and 119 days a of the infestation..................81

PAGE 10

x 4-9 Distribution of multiple trap sensit ivities for ML traps under all outbreak scenarios for different trap densities........................................................................82 4.10 Marginal trap sensitivities for ML tr aps under different outbreak scenarios...........83 4-11 Distribution of probab ilities of detection for ML traps under all outbreak scenarios for different trap densities........................................................................84 4-12 Marginal values of probability of detection for ML traps densities under different outbreak scenarios.....................................................................................85 4-13 Optimal trapping density per t ype of trap, location and month...............................86 5.1 Outbreak scenarios and cost implications of a Medfly infestation on the Florida fruit and vegetable industry....................................................................................114 5-2 Baseline annual returns of fresh and processed Florida grapefruit for the 200001 season and changes in the medfly m odel including the market of Japan..........115 5-3 Baseline world FOB revenue for red grapefruit for the 2000-01 season and changes in the Medfly model including the market of Japan.................................116 5-4 Baseline world FOB revenue for white grapefruit for the 2000-01 season and changes in the Medfly model including the market of Japan.................................117 5-5 Baseline annual returns of fresh and processed Florida grapefruit for the 200001 season and changes in the Medfly m odel excluding the market of Japan.........118 5-6 Baseline world FOB revenue for red grapefruit for the 2000-01 season and changes in the Medfly model excluding the market of Japan................................119 5-7 Baseline world FOB revenue for white grapefruit for the 2000-01 season and changes in the Medfly model excluding the market of Japan................................120 5-8 Planted acreage in the baseline and Medfly models by crop and area...................121 5-9 Baseline production for the 2000-01 season and percentage changes in production in the Medfly model by crop and area.................................................122 5-10 Baseline revenue for the 2000-01 season a nd changes in revenue s in the Medfly model by crop and area..........................................................................................123 5-11 Baseline demand for the 2000-01 season a nd percentage changes in demand in the Medfly model by crop and market...................................................................124 5-12 Baseline average prices for the 2000-01 s eason and percentage changes in prices in the Medfly model by crop..................................................................................125

PAGE 11

xi 5-13 Baseline annual returns of fresh and processed specia lty citrus for the 2000-01 season and changes in the Medfly model...............................................................125 5-14 Baseline FOB revenues for specialty citrus for the 2000-01 season and changes in the Medfly model...............................................................................................127 5-15 Aggregate impacts of a Medfly outbreak in Florida with scenarios of threemonth, six-month, and one-y ear quarantine periods..............................................128 5-16 Baseline production and percentage ch anges in crop production in the Medfly model by area.........................................................................................................128 5-17 Baseline revenues by area for the 200001 season and changes in revenue in the Medfly model.........................................................................................................128 6-1 Medfly risk levels in Florida..................................................................................136 6-2 Effects of entry conditions on the risk level...........................................................137 6-3 Air passenger baggage clearance cost im plications of changes in the number of passengers...............................................................................................................137 6-4 Impact of an improvement in single trap sensitivity on the optimal trapping density....................................................................................................................137 6-5 Impact of pesticide efficacy on the level of Medfly risk........................................138

PAGE 12

xii LIST OF FIGURES Figure page 2-1 Female Mediterranean fruit fly................................................................................19 2-2 Male Mediterranean fruit fly....................................................................................20 2-3 Seasonal distribution of Medf ly occurrences in Florida..........................................21 2-4 Distribution of Medfly interceptions in Florida.......................................................21 3-1 Inward shift in import supply resultin g from the imposition of an SPS barrier.......49 4-1 Bayesian decision process........................................................................................78 4-2 Treatment and quarantine areas of an infestation scenario in Miami (October)......79 5.1 Price equilibrium, aggregate demand and supply..................................................113 5-2 Effects of phytos anitary regulations.......................................................................114 6-1 Shifting of the eradicati on cost curve with change in the pesticide efficacy.........136

PAGE 13

xiii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ECONOMIC IMPACT OF A MEDITERRANEAN FRUIT FLY OUTBREAK IN FLORIDA By Raphael Yves Pierre May 2007 Chair: John J. VanSickle Cochair: Thomas H. Spreen Major Department: Food and Resource Economics We evaluated the potential impact of a Mediterranean fruit fly infestation in Florida. We developed a Bayesian decision framework to analyze the costs of Florida Medfly prevention, detection and eradication programs under early versus late detection scenarios. Modeling results s upport the hypothesis that optim al trapping density varies across locations and seasons. Because of the low probability of detecting small Medfly populations, the corresponding optimal trapping densities are high, ranging from 82 to 465 traps per ha for McPhail tr aps and from 9 to 80 traps pe r ha for Jackson traps. It would be extremely costly to maintain su ch high trap densities over a wide area. Alternative solutions lie in the search for an increase in pesticide efficacy and an improvement of the trapping t echnology. Development of more effective female-targeted trapping systems will provide a new dimension to the detection of small Medfly populations.

PAGE 14

xiv Partial equilibrium models were also used to investigate welfare changes for the major fruit and vegetable crops under scenar ios of a 3 mo, 6 mo, and 1 y quarantine period. Our analysis provides insight rega rding the magnitude of welfare changes associated with a Medfly outbreak and/or in festation in Florida. These changes vary across crops, depending on the competitive posi tion of Florida growers for the crop, size of the infested area, and lengt h of quarantine period. Finally, we tested the effects of chan ging the entry conditions on the level of Medfly risk in Florida. Our sensitivity-ana lysis tests showed the increasing number of international passengers entering Florida to be the driving parameter affecting Medfly introduction and establishment in Florida. Additional passenger ba ggage-clearance costs will be continuously needed to keep pace w ith the increasing numb er of international travelers entering Florida. Another way to mitigate the risk of Me dfly introduction into Florida is to encourage and support suppression and eradication activit ies against fruit fly populations in Caribbean basin countries.

PAGE 15

1 CHAPTER 1 INTRODUCTION Problematic Situation Fruit and vegetable production is an important industry in the state of Florida, with an estimated value of $4 billion as farmer cas h receipts, including c onsiderable exports. Florida farmers used a little more than 10 million of the states nearly 35 million acres and made more than a $12 billion direct impact on the states economy (Florida Department of Agriculture and Consumer Services 1998). Citr us sales topped $1.3 billion, accounting for more than 22 % of stat e agriculture sales in 1997. Florida farmers provide more than 70% of the nations citrus and nearly 10% of its vegetables. They produce 80% of the countrys domestically gr own vegetables during the winter months. The Mediterranean fruit fly (popularly known as Medfly1) poses a serious threat to fresh fruit and vegetable pr oduction throughout Florida and s outhern Florida. Since its first detection in Florida in 1929, the pest has been intermittently introduced in the state. Most of its introductions can be traced to accidental or intentio nal (smuggling) human interventions; Florida is the path for large volume of commodities and international travelers (APHIS 1995a, 1999, 2001, 2003a). Because of its wide range of hosts, its explosive reproductive capacity, and its extreme adaptability to adverse ecological conditions (Klassen et al. 1994), Medfly is considered one of the worlds most destructiv e fruit pests and is the subject of strict 1 The molecular data on the genetic variability of the Mediterranean fruit fly confirm that the name Medfly may be inappropriate because the ancestral home of C. capitata is Africa, so it ought to be called Africafly. (Malacrida et al.1998; Hoy 2003).

PAGE 16

2 quarantine and comprehensive control progr ams. Countries are allowed to impose technical barriers to the free movement of fresh fruit and vegetable commodities to protect national production from pests, diseas es, and contaminants; but the Sanitary and Phytosanitary (SPS) Agreement requires th at any regulatory m easure be based on scientific assessment of risks associated with intr oducing these invasive species. Regulatory and eradication2 measures are justified becau se the pest significantly increases production costs; and fruits grow n in Medfly-infested regions cannot be exported to Medfly-free areas (thereby aff ecting national and international trade). APHIS and the State of Florida spend milli ons of dollars each year on research and exclusion3 (risk-assessment studies, clearing import ed cargos, inspec ting and regulating passenger baggage, restrictions on imports) to prevent a wide complex of exotic animal and plant diseases and pests, including Medfly. The Medfly-detection program was designed for early detection of Medfly introdu ctions. Florida is c onsidered Medfly-free thanks to periodical eradication campaigns combining non-chemical and chemical control methods. Prompt aerial applica tions of malathion bait over Florida cities regarded as major entry points of Medfly have been key to the success of these campaigns. However, public concern has been growing about the intensive use of malathion in eradication operations. Some cr itics advocate the ab andonment of all chemicals. Floridas legislation is likely to restrict aerial app lications of pesticides over heavily populated areas (APHIS 1999). Alternative detecti on, prevention, and eradication programs 2 Eradication is a process used to achieve a fruit-fly free area. Trapping surveys are carried out to measure the efficacy of control measures (such as bait sprays, SI T, biological control, and MAT) used to eliminate a pest from an area (IAEA 2003). 3 Exclusion is a process used to minimize the risk of introduction or re-introduction of a pest in a free-area. Trapping survey are carried out to determine the presence of species that are under exclusion measures and to confirm (or reject) the fr ee-area status (IAEA 2003).

PAGE 17

3 involving intensive tr apping and regulatory measures preventive sterile-release programs, and ground-only pesticide applica tions are being studied in a context of increased trade and human travel. Problem Statement Medfly incursion and establishment in Fl orida involve factors beyond the control of APHIS policy makers. Medfly infestations in Florida might become likelier with increased trade and movement of people from infested countries. In such a context, timely detection is crucial to effective management and cont rol of a Medfly population at the earliest possible stage. However, disc overing low populations of wild Medflies in continuous plantings of a host species is exceedingly difficult (C alkins et al. 1984). Optimal control may require greater density of traps to better detect low populations, but maintaining a high trap density over a wide area is extremely costly. Decisions to increase trap density must consider the consequences of failing to detect low populations. Another area of concern is the potential impact of a Medfly outbreak in the Florida fruit and vegetable sector. Economic st udies (APHIS (1993, 1999) conservatively measure producer losses, ignoring the uncertain effectiveness of dete ction and eradication program effectiveness, potentia l changes in retail prices, and costs to consumers. An appropriate modeling framework is needed for a welfare analysis of a Medfly outbreak in Florida. Hypotheses There is a trade-off between early and late detection in cost management of the Medfly-eradication program. Early-detecti on costs are high for trapping, but low for eradication (with a low probability of establis hment). Late detection costs for eradication are high (with a high probab ility of establishment).

PAGE 18

4 Hypothesis 1 Optimal trapping density varies across lo cations and seasons; it minimizes the expected cost of the preven tion and eradication program A Medly outbreak in Florida would probabl y cause a decrease in export markets, an increase in preand post -harvest production costs, and a yield reduction from the infested areas. Hypothesis 2 The magnitude of welfare changes associat ed with a Medfly outbreak will vary from one crop to another, depending on growers competitive position for the crop, size of infested area, a nd length of quarantine period. Objectives Our overall objective was to analyze the cost implications of a Medfly outbreak in Florida for APHIS and the State of Florida in general, and for producers and consumers in particular. We analyzed a variety of out break scenarios. Specific objectives are to: 1. Review the policy framework for using the Sanitary and Phytosanitary (SPS) Agreement 2. Evaluate costs of Medfly detection and eradication programs under different pestdetection scenarios 3. Use estimated probabilities of detection to determine the optimal trapping strategy specific to each location and season of the year 4. Evaluate welfare changes a ssociated with outbreak scen arios of 3 mo, 6 mo, and 1 y quarantine periods 5. Formulate alternative policy measures to reduce the risk of Medfly introduction and establishment in Florida. Thesis Outline The plan of this thesis is as follows. The next chapter pres ents an overview of biological profile of the Medi terranean fruit fly, Florida fr uit and vegetable sector, and

PAGE 19

5 Medfly-detection and eradication programs. The third chapter follows with an analysis of the World Trade Organization (WTO) Ag reement on the application of the SPS measures. This review focuses on phytosan itary protocols and re gulated postharvest treatments for the domestic and internationa l movement of fresh fruit and vegetable commodities. Chapter 4 describes the components of th e Bayesian decision framework: spatialtemporal model of infestation, models of pr obability, and cost minimization model. These models were used to estimate (1) probabilities of detecting a Medfly infestation in Florida under different outbreak scenarios and (2) opt imal trapping densities minimizing total expected cost of prevention and eradicat ion programs. Chapter 5 follows with the description of optimization models used for analyzing welfare changes associated with a Medfly outbreak in Florida. We analyzed scenarios of 3 mo, 6 mo, and 1 y quarantine periods. Chapter 6 deals with the determination of Medfly risk by combining the results of Chapters 4 and 5. Sensitivity-analyses are conduct ed to evaluate the effects of changes in the major factors leadin g to the introduction and establishm ent of Medfly into Florida. The paper concludes with the formulation of alternative policy measures to reduce the risk of a Medfly endemi c situation in Florida.

PAGE 20

6 CHAPTER 2 REVIEW OF MEDFLY INTRODUCTIONS AND INFESTATIONS IN FLORIDA Biological Profile of the Mediterranean Fruit Fly The Mediterranean fruit fly (Ceratitis capitata ) is the most notorious member of the family Tephritidae (Narayanan and Batra, 1960). It originates from sub-Saharan Africa and is now widely distributed in many countri es of the Mediterranean coast. It has no near relatives in the western hemisphere. Adult Medflies are slightly smaller than a housefly and are very colorful. They can be readily distinguished from any native fruit flies of the new world. Females have a characteristic yellow wing pattern and the apical half of the scutellum is black (Figure 2-1) and males can be separated from all other members of his family by the black pointed expa nsion at the apex of the anterior pair of orbital setae (Figure 2-2). Medfly is one of the most destructive ag ricultural pests in th e world (Papadopoulos et al. 2001, 2002). Medfly reduces the yield of fruit and vegetable crops and it affects their quality. Ripened fruits infested with th e Medfly may be unfit to eat, as the female pierces the soft skin and deposits eggs in th e puncture. A female Medfly may lay as many as 800 eggs during her lifetime. The average da ily oviposition rate is 11 eggs (Thomas et al. 2001; Papadopoulos et al. 2002 ). Those eggs produce maggots or wormlike larvae that feed on the pulp of the fresh fruit, drop to the ground, and transform into pupae in the soil or elsewhere. Many researchers report high mort ality in Medflies, particularly in citrus hosts (Back and Pemberton 1918; Bodenheim er 1951). However, high fecundity and

PAGE 21

7 adult longevity generally more than compensa te for high larval mortality (Carey 1982a). If only 10% of larvae survive, the population can double every 10 days. Life expectancy and insect developmen t depend on climatic conditions. Under typical Florida summer weather conditions, th e life cycle of the medfly (which also includes maturation of the pupae into adults) varies from 21 to 30 days (Thomas et al. 2001). Optimum temperatures for the most rapid development and high potential for spread range from 70 to 90o F. Adults emerge in the largest numbers early in the morning during warm-temperature and longer freeze-fre e periods, and emerge more sporadically during cool weather. According to Buyc kx (1994) and Bodenheimer (1963), Ceratitis capitata is confined to regions where cold weather combined with low humidity interrupts development in egg, larval, and pupal stages for less than 100 days. Low temperatures greatly increase the duration of the egg stage. Nevertheless, field studies report that pupae can carry the species thr ough unfavorable conditions such as lack of food, or water, or temperature extremes (T homas et al. 2001). Pupae can develop at temperatures as low as 50o F (10oC), making the Medfly extremely adaptable and a serious pest-control challenge. Ceratitis capitata is a polyphagous and multivoltine tr opical species that has spread to all countries bordering th e Mediterranean region, southe rn Europe, western Australia, Central and South America, and Hawaii. It ha s been recorded in more than 200 different types of fruits and vegetables. Medfly was first discovered in Central America in April 1955 near San Jose, Costa Rica. It then spr ead into Panama, Nicar agua, El Salvador, Honduras, Guatemala, and the southern border of Mexico (Gonzales 1978). Primary hosts of Medfly in Central Amer ica are coffee, tangerine, orange, grapefruit, peach, and

PAGE 22

8 tropical almonds. Medfly populations there main ly vary according to seasonal factors, reaching a peak during the dry season and with the maturation period of the host fruits, particularly coffee (Henning et al. 1972; Eskafi and Kolbe 1987). Low populations coincide with increased rainfall but vary depending on elevation. Much of Florida would maintain high Medfly populations because of both favorable climate and availability of pr eferred host materials throughout the year. Average mean monthly temperatures for the city of Miami (Dade County) are in the optimal range 8 months a year, and above the minimum temperature for Medfly development all year. Almost all major crops produced in Florida (oranges, grapefruit, lemons, tangerines, and tangelos) could be heav ily infested by Medfly (Table 2-1). Some vegetables and melons are classified as occasi onally or rarely infested hosts. Medfly also could be attracted to patches of ornamental trees that provide shelter in otherwise barren areas (Buyckx 1994). Thorough knowledge of the hosts in one country often can help predict the hosts most likely to be in fested in a newly infested country. Movements of flies seemed to be influen ced more by the distribution of hosts than by wind direction (Buyckx 1994; Wakid and Shoukry 1976; Hafez et al. 1973). Flies do not move great distances under favorable c limatic conditions and where fruits and vegetables are available. Movements appear to be restricted to a few hundred meters per week. Long distance flights of ove r 20 km are associated with some behavioral surviving mechanism for Medly populations in areas where fruit suitable for ovipositing are unavailable during certain times (Harris a nd Olalquiaga 1991). Thus, population density and the economic damage would vary from area to area, depending upon the seasonal availability of hosts and the presence or absence of unfavorable climatic conditions.

PAGE 23

9 The biggest factors in dispersal and spread of Medfly populations are fruit transport and trade (Buyckx 1994). Modern roads allow fr esh fruit to be transported throughout the states. International pa ssengers might also bypass detec tion and enter the United States with high-risk Medfly host material. Medf ly surveys conducted by APHIS (1995a) show that the major pathways for Medfly into Fl orida are passenger and crew baggage from foreign countries. Approximately 93% of all high-risk and infested host materials enter Florida at the ports of Miami and Orlando. JF K International Airpor t is the only other airport with a significant volume (4%) of highrisk and Medfly-infested material destined for Florida. The risk of Medfly entering the U. S. vi a Mexico is small, because Mexico is Medfly-free. However, there is a potential risk for Medfly in troduction from infested fruit carried into the United States by travelers from a third country. About 2.17 % of people crossing the Mexican border illegally each year are not of Mexican descent, and most of them (78%) are from Central American c ountries where Medfly occurs (APHIS 1992). Many of these illegal aliens carry food with th em that could be infested with fruit fly. These findings are consistent with hist orical data on Medf ly occurrences and infestations in Florida (Table 2-2). Severa l Medfly occurrences are closely associated with ship and boat traffic in Miam i (1964, 1967, 1984, 1985 and 1988); Tampa (1981); and Fort-Lauderdale (1990). Five of the nine Florida infestations occurred in or around Miami Springs, a small upper-income housing area at the northern edge of Miami International Airport. Seasonal distribution of Medfly occurren ces in Florida (Figure 2-3) seems not be related to colonization potenti al for Medfly, as temperatures and host

PAGE 24

10 availability are favorable all year in southern Florida (APHIS 1992). Seasonal distribution of interceptions from Latin Amer ica is the same as s easonal distribution of Medfly detection in Florida. However, seas onal distribution of inte rceptions from Hawaii (where Medfly is permanently established) ma rkedly differs from seasonal distribution of Medfly detection in Florida (Figure 2-4). These observations strongly suggest that Latin American countries (instead of Hawaii) could be a high-risk source of Medfly infestations in Florida. Overview of the Florida Medfly Detection and Eradication Program Florida is a managed Medfly-free area4 in which the pest has been intermittently eradicated. As argued previously, no limiting na tural factors prevent its establishment in Florida. The State is permanently protected by federal regulatory actions (quarantine inspection, surveillance networ ks, import restrictions) c ontrolling the movement of people and commodities. Some of these meas ures target specific importing countries where the Medfly is established or not completely eradicated. 5 Under GATT and WTO rules, APHIS is compelled to ensure that quarantines are imposed for sound biological reasons, rather than for prot ectionist trade barriers. A cooperative agreement between APHIS and the State of Florida provides early detection of Medfly introductions. Timely de tection of small Medfly populations greatly helps management of this pest (Dowell et al. 1999), but recent studi es show that early 4 Malavasi et al. (1994) distinguish two types of flyfree areas, a natural and a managed free area. In a natural fly-free area, the species naturally does not occur (because of eco logy, host preference, geographical distribution, etc). Mana ged fly-free areas are production zone s from which the target fruit fly has been eradicated. These areas must be permanently protected by regulatory actions. 5 For instance, cooperative partners hip agreements are signed with Me xico and other countries of the western hemisphere (like Guatemala and Costa Rica) with a view to diminishing their pest problems, thus reducing the risks of Medfly introductions into the United States.

PAGE 25

11 detection can be difficult (Papadopoulos et al. 2000). The lure for Medfly (called Trimedlure) is weak and has li ttle ability to attract the f lies to the trap (Scribner 1983). The Jackson trap (the most effective and co mmonly used Medfly trap) can only catch 1 in 2000 Medflies in the area. Current national protocol requires 21 trap s per square mile for high-risk areas, 12 traps for medium-risk areas, and 3 traps for low-risk areas. 6 About 47,404 traps are currently placed in 49 counties in Florida. About 54% of th e traps are concentrated in high-risk areas in the following counties: Pinellas, Hillsborough, Orange, Palm Beach, Dade, and Broward. In FY 2003-04, the total co st of the Medfly detection program was approximately $7 million, with salaries and supplies accounting for 86.8% (Table 2-3). Drawing on the 1997/98 Medfly infestati on, potential budget changes in case of emergency (about $ 398,500 per month of emerge ncy) represent significant increases in employees overtime hours, trav el expenses, and services. The success of any Medfly-eradication program depends on early-detected infestation (an infestation wh ere the area under quarantine is 110 square miles or less). Once detection traps capture one or more Medf ly adults, additional traps are placed to determine whether an outbreak has occurred and/or to limit the outbreak. Drawing on the potential mobility of Medfly adults, this trapping strategy occurs in an 81-square-mile area around the fly find, which is divided into a core area (1 square mile for a single fly capture) and several buffer areas. The whole ar ea is placed under strict quarantine to prevent the movement of any regul ated articles to non-infested areas of the state. All host fruits on the property and thos e properties immediately adja cent are stripped promptly 6 Five ML traps and 16 TML traps are placed in hi gh-risk areas, two ML traps and 10 TML traps in medium-risk areas, and one ML trap and 2 TML traps in low-risk areas.

PAGE 26

12 and disposed of according to APHIS protocol s. However, extensive fruit stripping may stimulate dispersal of gravid females, th ereby making eradication more difficult. Treatment occurs when a Medfly infestation is determined7 to occur. Generally, eradication procedures combine mechanical, chemical, and biological controls. Table 2-2 summarizes the total eradication costs from 1929 to 1997. Eradication costs have been drastically reduced since the early 1960s by the development of trapping technology and because of the intensive use of aircraft during eradication operations. Mass spray applications of malathion bait have been made possible over he avily populated areas, often within minutes after disc overy of an infestation. This approach has been key to the success of these eradication programs, providi ng complete coverage of the epicenter of Medfly infestation. Ground applications continue to be used for the treatment of soil with dieldrin to kill emerging adults and larvae entering the soil. Chemical control is used mo stly to reduce Medfly populations to a low level before the sterile insect technique (SIT) can be use d. Currently used to pr event and eradicate, SIT uses the ability of factory-produced inse cts to disrupt the nor mal mating patterns of wild Medflies. Sterile Medflie s mate with their wild counterparts, resulting in the production of infertile eggs. Preventive steril e release program is more successful when the number of wild flies intr oduced is very low. In Florida, about 125,000 flies per week per square mile are currently released over the high-risk areas8 to provide prevention control. In cases of emergency, another 400,000 flies per week per square mile would be 7 The deliberative process is based on the following: 1) presence of two flies within a three-mile radius; 2) presence of one mated female; or 3) presence of larvae or pupae (APHIS 2003a). 8 Area-wide sterile release would incl ude the following criter ia: areas where Medflies were detected in the past, areas in proximity to ports of entry, and/or urban or suburban areas where frequent movement of imported and exotic Medfly host occurs.

PAGE 27

13 released over each standard block of 100 squa re miles to support er adication operations. In 2004, the cost of the Preventive Rel ease Program (PRP) was roughly $3.34 million (Table 2-3). In an emergency, PRP cost s would increase by about $ 82,000 per block. Sole reliance on SIT for prevention has b een unsuccessful, mostly because program managers cannot maintain the necessary release ratio of sterile to wild fruit flies. For curative purposes, a minimum ratio of 10:1 (ste rile to wild) is required to halt Medfly population growth and achieve complete erad ication (Carey 1982a). Eradication using SIT is ineffective in production areas becau se sterile insects are killed by grower applications of insecticides. Entomologist s have also become aware of behavioral deficiencies of sterile insects versus their wild counterparts. Artif icial conditions of the mass-rearing reduce mating performance, pr oducing Medflies that compete poorly for females (Jang et al. 1994; Jang 2002). Vari ous approaches to improving the mating success of SIT flies have been studied. The future effectiveness of Medfly contro l programs is inextricably bound to public concern about pesticide use and potential adve rse effects. Ample data on environmental impacts (fish kills, invertebrate losses, and human health effects) were collected during the 1997 program in Florida to show that malathion is not safe (APHIS 1999). These issues raised divergent belie fs about whether the benefits of eradication operations outweigh the environmental costs, and whether the risks associated with pesticide use are manageable to acceptable levels. Our study helped elucidate these controversies by examining the probabilities of risk and rela ted economic consequences associated with a Medfly outbreak in Florida.

PAGE 28

14 Overview of the Florida Fruit and Vegetable Sector Florida produces a wide range of fruit and vegetable crops, responding to increased consumer demand driven by population growth and growing concern over a healthy diet. In 2001, Florida farmers used approxim ately 900,000 acres to produce 5.28 billion pounds of fruits and vegetables. These crops provide revenues of $4 billion to Florida growers. Tables 2-4 and 2-5 report the acr eage, yield, and use of Florida fruit and vegetable crops for the 1999-2000 season. Orange s, fresh tomatoes, and grapefruits are the leading crops, accounting for 68.13% of the to tal acreage allotted to this sector. Other important crops include bell pe ppers, cucumbers, tangerines, tangelos, watermelons, and eggplants. Florida is known for its citr us fruits, primarily grown in the central and southern parts of the state. State farmers lead the na tion in the production of oranges, grapefruit, tangerines, and tangelos. About 75% of the na tions oranges are grown in the state: more than 90% was used to make more than 1.5 bi llion gallons of jui ce in 1997. Florida and Brazil are the major competitors, accounti ng for over 20 and 60% of world production, respectively. Florida orange growers suffer par ticularly from large fl uctuations in orange juice prices, for demand is highly sensitive to consumer income (Spreen 2001). Florida also produces 77% of the U.S. dome stic grapefruit and nearly 47% of the world supply. This production is grossly split in half: one half is processed, and the other half is marketed in fresh form. Florida gr apefruit growers are highly dependent on export markets in Europe, Canada, and Japan for fres h grapefruit. In particular, the opening of the Japanese fresh citrus market has resulted in large increases in shipments of fresh white seedless grapefruit to Japan (Spreen et al. 1995).

PAGE 29

15 Tangerine production in Florida has increased from 222.3 million pounds in 1986/87 to 522.5 million pounds in 2002/03. An average of 70 percent of the tangerines produced are sold in the fresh market, with the remainder going to the processing sector for juice, sections, or other uses (USDA 2003). U.S. domestic production is supplemented by tangerine imports mainly from Mexico and Spain. While imports of Mexican tangerines have remained relatively stable over the last six years, imports of clementines from Spain have increased from 33 million pounds in 1994/95 to 119 million pounds in 2000/01. The state also ranks second nationally in the value of its ve getable crops that account for more than 25% of Florida agricultu re sales. Florida wint er fresh vegetables are produced mostly in the sout hern half of the state where adequate conditions prevail (VanSickle et al 1994). Although the state ha s faced a growing array of problems in the winter fresh vegetable industry, growth in vegetable production for the last two decades has been mostly related to the use of hybrid cultivars and improved management practices. In 2002, Florida growers used less than 290,000 acres and received $ 506 million in sales from tomatoes, $ 245 milli on from green peppers, $ 122 million from snap beans, $ 107 million from cucumbers, and $ 105 million from sweet corn. Fresh vegetables generally move from the field to the packing shed fo r packing, pre-cooling, and storage before shipment to wholesale mark ets. Industry sources estimate a total of 60 to 70 packers/shippers throughout the state. Florida growers held a competitive edge over their traditional Mexican competitors in the U. S. markets for field-grown vegeta bles. However, the development of greenhouse technologies has brought recent changes in U.S. vegetable markets. Florida growers are

PAGE 30

16 now facing growing competition with the la rgest greenhouse producing areas in Spain, Italy, France, and Greece. Productivity in Eur opean greenhouses is nearly three fold comparable to Florida field production and product quality is generally higher from greenhouse versus field-produced vegetabl es (Cantliffe & VanSickle 2003). This competition is likely to affect all major vege table crops grown in Florida, like tomatoes, peppers, eggplants, cucumbers, muskmelons and to some degree, watermelons. For instance, U. S. imports of greenhouse tomatoes have grown rapidly, from 43.9 million pounds in 1994 to 395.5 million pounds in 2000, including 224 million from Canada, 76.5 million from the EU, and 96 million from Mexico (Cook 2002). As imports increase, fresh tomato acreage declines in both Florid a and California. Con cern is growing about the impacts of importing greenhouse tomato es on U. S. vegetable industry. Along these lines, an economic evaluation of the potential damage to the Florida fruit and vegetable sector from a Medfly infestation must be approached within a framework of growing competition among th e different economic agents from both within and outside the United States. In addi tion to the potential lo sses associated with yield reduction and increases in preand postharvest costs, Florida growers competitive position would be further weakened through pr ice adjustments due to losses in export markets and shipment restrictions to ot her states. APHIS (1993, 1999) predicts that countries would react according to their Medf ly status and their regulations. While some countries like Mexico, Argentina, and Chile would require treatme nt of Medfly hosts from Florida, others like China, Japan, a nd the Caribbean nations would prohibit the importation of all Medlfy hosts, including marg inal hosts for a number of years. It is

PAGE 31

17 estimated that Florida would lose over 50% of its export markets if Medfly became established in Florida. APHIS and FAO/IAEA have carried out many studies to assess the economic impact of the potential Medfly damage on fr uit and vegetable produc tion, using a partial budget approach (APHIS 1993, 1999; FAO 1995; IAEA 1995; Enkerlin and Mumford 1996). Losses in producers reve nues are estimated, but none of these studies have given consideration to price change s related to changes in output and export markets. All changes in production are measured at curr ent prices and costs to customers are completely ignored. Furthermore, findings from risk assessm ent studies carried out by APHIS (1995b) are not incorporated into the economic anal ysis, which could have better supported and shaped regulatory policy options. Losses in the value of production are grossly estimated at 5% (APHIS 1993, 1999), under the questionabl e assumption that eradication, regarded as a proven technology, can be achieved with complete certainty. The State of Florida would incur an expected eradication9 cost of $4.8 million each year if eradication were successful. Costs to producers would ra nge from $32 million to $300 million, depending on whether eradication is successful. However, Farnsworth (1985) argued that eradication is a two-event combination (eradication feasible and eradi cation not feasible) with a ra nge of probabilities summing to one. Other studies emphasize the uncerta inty of prevention and control program effectiveness and recommend the use of partial equilibrium models to estimate the 9 The expected cost of eradication per year is calculated by multiplying the probability of an outbreak per year by an average eradication cost. The calculation of this probability is based upon the history of Medfly outbreaks in Florida. It is predicted that, given no major changes in APHIS exclusion activities, this probability is about 0.2 per year (or once every 5 years).

PAGE 32

18 economic effects of this program under differe nt outbreak scenarios (Regev et al. 1976; Rendelman and Spinelli 1999; Brown et al. 2002 ). Approaches taken in these studies allow for price changes in the commodities concerned under the assu mption that linkages with other similar commodities are small. Summary Medfly is one of the most destructive ag ricultural pests in th e world. It has been introduced 18 times in Florida, leading to 9 infestations. Five of these infestations occurred in or around Miami Springs, a sma ll upper income housing area located on the northern edge of the Miami International Air port. Without control measures, much of Florida would maintain high Medfly populati ons because of both favorable climate and availability of preferred host materials th roughout the year. Various observations strongly suggest that Latin American countries could be a high-risk source of Medfly infestations in Florida. Florida is a managed Medfly-free area in which the pest has been intermittently eradicated. Millions of dollars are spent annually on exclusi on and detection activities to prevent Medfly establishment in Florida or at least provide ear ly detection of its introductions. Such investments are justified because of the importa nce of the fruit and vegetable industry in Florida. The State leads the nation in the producti on of a wide range of fresh fruit and vegetable crops, thereby re sponding to an increas ed consumer demand driven by population growth and growing concern over a healthy diet. An economic evaluation of the potential damage from Medfly on the Florida fruit and vegetable sector must be approached within a framework of growing competition among the different economic agents from both within and outsi de the United States. Partial equilibrium models can be designed to allow for poten tial price changes in the commodities

PAGE 33

19 concerned and take into account the uncertainty of the detection a nd eradication program effectiveness. Figure 2-1 Female Mediterranean fruit fly Table 2-1 Major fruit and vegetable crops grow n in Florida according to their importance as Medfly hosts Heavily infested Occasionally infested Unknown Lab infestations Grapefruit, tangelo, tangerine, lime, bell peppers, lemon, mango Avocado, eggplants, ripe tomatoes, strawberries Watermelons, snap beans, squash Cucumbers, eggplants Sources: Liquido et al. 1991

PAGE 34

20 Figure 2-2 Male Mediterranean fruit fly Table 2-2 Costs of Medfly infestations in Florida Costs ($ millions) Year Area of detection Counties affected Nominal 1990 1929 Orlando 20 7.5 56.5 1956 North Miami 28 11 50.6 1962 Miami 3 1 4.1 1963 Miami 1 0.3 1.2 1981 Tampa 1 1 1.4 1984 Miami 1 1 1.2 1985 Miami 1 2.2 2.6 1987 Miami 1 1.3 1.5 1990 Miami 1 1.8 1.8 1997 Tampa 5 24 20 Source: Clark et al. (1992)

PAGE 35

21 Figure 2-3 Seasonal distribution of Medfly occurrences in Florida Figure 2-4 Distributi on of Medfly interceptions in Florida 10 15 20 25 30 35 JanFebMarAprMayJunJulAugSepOctNovDec Monthly Distribution Degree Celsiu Average Minimum Temperatures Average Mean Temperatures Average Maximum Temperatures Upper Medfly Threshold Lower Medfly Threshold Upper Medfly Optimal Lower Medfly Optimal Infestations in Miami Source: APHIS 1992 0 5 10 15 20 25 30 JanFebMarAprMayJunJulAugSepOctNovDec Monthly Distribution% of Medfly Intercepti o Medfly Interceptions from Latin America Medfly Interceptions from Hawaii Infestations in Miami Source: APHIS 1992

PAGE 36

22 Table 2-3 Baseline and emergency budgets fo r the Medfly preven tion and detection program Preventive release program Detection program Baseline annual budget ($) Emergency budget a ($) Baseline annual budget ($) Emergency budget b ($) Salaries, benefits & overtime 931,402 5,000 5,090,988 112,000 Travel expenses 27,3 00 2,000 56,625 105,000 Transportation & shipping 168,000 9,600 14,500 30,500 Rents, utilities & communication 250,376 362,149 9,000 Services & repairs 1,006,544 10,120 180,000 116,000 Supplies 892,000 55,000 989,000 21,000 Equipment 67,000 241,000 5,000 Total 3,342,622 81,720 6,934,262 398,500 a These costs are estimated for each additional pr oduction and release of 400,000 flies per week and per block of 100 square miles. b These costs are estimated for each month of emergency. Source: APHIS 2002 Table 2-4 Important fruits: acreag e, yield, and use (by crop), 2001-02 Use Crop Bearing acreage Yield per acre Fresh Processed Total (1,000 acres) (boxes) (1,000 boxes) All oranges 586.9 392 9,524 220,476 230,000 All grapefruit 101,3 461 17,380 29,320 46,700 All tangerines 24 275 4,204 2,396 6,600 Tangelos 9.7 222 696 1,454 2,150 Limes 0.8 188 125 25 150 Lemons 0.9 94 85 Avocados 5.9 156 920 920 Source: Florida Agricultural Statistics Service (2003).

PAGE 37

23 Table 2-5 Important vegetables: acreage, yield, and utilization, by crop, 2001-02 Crops Planted Acreage Harvested Acreage Yield per Acre Production Total Value (acres) cwt (1,000 cwt) ($ 1,000) Tomatoes 43,500 43,500 338 14,688 474,284 Cucumbers 7,500 7,500 386 2,893 56,012 Bell Peppers 17,250 17,100 320 5,469 170,340 Squash 12,000 11,700 135 1,578 44,543 Eggplants 1,800 1,800 257 463 12,501 Watermelons 25,000 23,000 330 7,590 62,238 Strawberries 6,900 6,900 255 1,760 153,472 Source: Florida Agricultural Statistics Service (2003).

PAGE 38

24 CHAPTER 3 POLICY FRAMEWORK OF THE SANI TARY AND PHYTOSANITARY (SPS) AGREEMENT The World Trade Organization Agreement on the Application of Sanitary and Phytosanitary Measures Sanitary and phytosanitary (SPS) restrictions are characterized as a subset of traderelated policies known as tech nical barriers to trade (TBT10). They include all measures adopted by a country to protect human, animal, or plant life and health from risks related to diseases, pests, and dis ease-carrying or causing organisms, as well as additives, contaminants, toxins or dis ease-causing organisms in food, beverages, or feedstuffs. Primary method of protection has been th e development of quarantine protocols exhibiting differing degrees of trade restric tions like complete bans, seasonal and/or geographical bans, postharvest disinfestations procedures (fumigation, cold storage or others), inspection at points of export and import, and even information remedies (Paul and Armstrong 1994; Roberts 1998). Unlike mo st non-tariff barriers, SPS measures are potentially welfare-increasing, for they ma y correct market failures resulting from externalities associated with the moveme nt of agricultural products across national borders (Roberts et al. 1999; Spreen et al. 2002). The World Trade Organization (WTO) Agreement on the Application of SPS measures recognizes all nations sovereign righ ts to enforce health standards on imports, as agricultural trade facilita tes the transportation of potentia lly harmful pests (which can 10 The Technical Trade Barriers (TBT) are in turn characterized as a subset of social regulations, encompassing all measures adopted by a country to achieve health, safety, quality, and environmental objectives (Roberts et al. 1999)

PAGE 39

25 cause widespread destruction when carried into a country). Increasing use of SPS standards and regulations reflec ts in many cases increased concern for public health and the environment. However, it is generally acknowledged that SPS m easures can also be used as disguised protection for domestic ag riculture. Domestic producer groups with a vested interest in a particul ar regulatory outcome are likely to lobby for overly restrictive measures that limit competition from impor ts, by exaggerating either the probability of infestation or the cost impact of infestation (Spreen et al. 2002). One objective of the Uruguay Round of multilate ral trade negotiations, as set out in the Punta del Este Declaration, is to minimi ze the adverse effects that SPS regulations and barriers can have on trade in agri culture. Indeed, government size and power facilitate the enactment and enforcement of regulatory barriers to trade for producer protectionism and/or consumer welfar e purposes (Thilmany and Barrett 1996). The Standards Code defined in the Tokyo Round failed to stem disruptions of trade in agriculture caused by the misu se of technical restrictions (Stanton 1977). The challenge before the negotiators of the new WTO Agreement on the Application of SPS measures was to create a set of rules that would stri ke the proper balance be tween allowing health and environmental protection and disallowing mercantilist regulatory protectionism (Roberts 1998). Toward this end, new substantive and pro cedural disciplines were established to facilitate the decentralized policing of SPS measures. WTO country members are required (1) to apply the same rules to dome stic and imported products and (2) to notify their trading partners of proposed SPS measures that might affect trade. Trading partners are therefore given opportunities to comment on a measure be fore it is adopted. The SPS

PAGE 40

26 Agreement also establishes a dispute-settle ment procedure involvi ng several levels of consultation, such as informal consultati ons, bilateral negotiati ons, recourse to the Committee on Technical Trade Barriers, and co nvocation before a panel of government officials. The cornerstone of the SPS Agreement is f ound in its Article 5 (which deals with the issues of risk and level of protection) [Federal Register 2004a]. Countries are granted the rights to choose their appropr iate level of protection (ALP ) against imported pests and diseases, but their regulations must be dem onstrably based on an assessment of risk and clearly related to the cont rol of the risk. Risk assessment typically involves the identification of the hazard, appraisal of the likelihood and consequences of the hazardous situation, and specification of th e way in which SPS measures would reduce those consequences (Caswell 2000). The SPS Agreement recommends that risk analysts develop a strong understanding of the pest biol ogy and potential pathways leading to its introduction in a new environment (Gray et al 1998). They often have to make use of value judgments, while struggling with data gaps, large uncertainties and the need to extrapolate. In sum, the analysis results in an assessment of the probability of introduction of a pest or di sease (a likelihood model). The different disease outcomes are treated as inputs into the ec onomic model to estimate. Net social welfare is the yardstick used by economists to capture the trade and welfare effects of a regulato ry protection model under pest control situatio ns involving either probabilities of infestation or certain ties (Roberts et al. 1999; Roberts 2000). From the perspective of the importi ng country, changes in net soci al welfare resulting from the imposition of a phytosanitary barrier (Spreen et al. 2002) can be expressed as follows:

PAGE 41

27 H W H EC W EC PS W PS CS W CS W (3-1) where W is aggregate social welfare, PS a nd CS denote producer surplus and consumer surplus respectively, EC acc ounts for the enforcement costs associated with the imposition of the phytosanitary barrier, a nd H is some index of human health. Figure 3-1 examines more closely changes in producer surplu s resulting from the reduction in the aggregate supply (AS) due to a shift of the excess supply (ES) in the exporting country and an increase in the produc tion costs associated with the use of the additional technology required by the SPS barrier. As a result, domestic prices increase, and consequently, domestic production (DS) also increases. The model could also accommodate the prevention program costs and the expected value of government pest eradication expenditures. These economic consid erations allow risk managers to identify the most effective pest management strategi es and to gauge whether a proposed measure meets the criterion of the SPS Agreement that it be the least trade restrictive (Caswell 2000). However, the SPS Agreement, as a trade f acilitator, does not endorse an explicit account of the costs and benefits of a po licys effects on producers and consumers. Rather, it encourages a myopic focus on dire ct risk-related costs of import protocols (Roberts 1998). Consideration of producer su rplus losses (gains) resulting from lower (higher) prices would likely be seen to be in violation of the spirit of the SPS Agreement, because they are costs related to commercial activity, but unrelated to health or environmental protection (Roberts 2000). T hus, under the SPS Agreement, commercial considerations might be appropriately factored into a countrys choice of its single ALP

PAGE 42

28 (Appropriate Level of Protection), but they shou ld not be used as decision criteria for individual risk mitigation measures. Such an approach stands in contrast to the economic paradigm of the U.S. Executive Branch directives requiring agen cies to base their SPS measures on costbenefit analysis (CBA) [USDA 1993]. Under th e risk assessment paradigm, the role of economics is relegated to the calculation of qua ntities of imports to help risk assessors with their job of calculati ng the likelihood and consequences of disease or pest introduction. Primary intent is to reduce the opportunities of practicing mercantilist regulatory protectionism in favor of domestic producer groups. Theref ore, the application of the SPS Agreement calls for the developm ent of international standards for the monitoring of CBA-based regulatory policies. Along these lines, the SPS Agreement seek s primarily to harmonize analytical frameworks for addressing risk. Internationa l standards for phytosanitary measures --such as the use of a systems approach, the establishment of pest free areas --are designed to achieve international harmonizati on of SPS measures. The former standard provides guidelines for the development and ev aluation of integrated measures for pest risk management, while the latter describes the requirements for the establishment and use of pest-free areas (PFA) as a risk ma nagement option for phytosanitary certification of plants and plant products (FAO 1996, 2002). Assignment of a PFA status is normally based on verification from specific trappi ng surveys and, subsequently, appropriate phytosanitary measures are required to ma intain freedom. A more effective pest management can be achieved by combining two or more independent phytosanitary

PAGE 43

29 measures11, like cultural practices, field trea tment, post-harvest disinfestations, inspection, or other procedures. The development of these internationally -accepted standards would contribute to facilitating trade by limiting the use of unjus tifiable SPS measures. Recognizing that the international guidelines may not reflect the preferences and/or needs for externality mitigation of every nation, the WTO Agreemen t also allows country members to set a higher level of protection, which must be ba sed on available scientific evidence. Thus opportunities are given for the expression of pol itical and cultural differences among the country members for evaluating threats to people and the environment. A number of situations are enumerated where national standards may differ from and/or exceed international standards (Article 2 of the SPS Ag reement). Wide discretion is also afforded to national governments in the determination of situations where international standards might be inappropriate (Bre dahl and Forsythe 1989). The first two years of implementation of the SPS Agreement saw a broad-based regulatory review among so me WTO members. High-inco me countries started to question whether the regulatory measures impos ed by their major tradin g partners were in compliance with the new Uruguay Round discip lines. The United States, for instance, identified over 300 questionable market re strictions imposed by 62 countries, which threatened, constrained, or blocked an estim ated $ 5.0 billion of US exports (Roberts 1999). One illustration of these market restri ctions, dating back to 1988, relates to the European Union (EU) ban on the importation of beef from cattle tr eated with growth11 The characteristic of a systems approach is that it requires at least two or more phytosanitary measures that are independent of each other. With independent measures, the probability of failure is the product of the probabilities of all the independent measures. A systems approach may also include any number of measures that are dependent on each other. With dependent measures, the probability of failu re is approximately additive, implying that all dependent measures must fail for the system to fail (FAO 2002).

PAGE 44

30 enhancing hormones. The United States questione d the scientific basis of this regulatory measure, thereby arguing that the ban had b een primarily motivated by the desire of the EU officials to impose a disgui sed restriction on the productivity in the beef sector and on imports from the most competitiv e beef exporters. In return, th e EU rejected these claims, arguing that the ban addressed public anxiet ies vis--vis the consumption of hormonetreated beef and that the SPS Agreement contai ned no disciplines restricting the absolute level of protection that a member may choose. The USA/EU dispute exemplifies a very diffi cult case to resolve, where culture and consumer preferences affect risk assessm ent. Citizen and consumers from country members show different perceptions towards ri sk associated with SPS issues (Bureau and Marette 2000; Caswell 2000; Schuh 2000). For in stance, the U.S. beef exporters have been less willing to accept the fact that a la rge percentage of European consumers may have a cultural aversion to eating beef produced with growth-enhancing hormones or antibiotic drugs. The issue is whether intern ational trade regulations should take into account cultural differences among countries and how to establish a clear dividing line between what is perceived as a food safety i ssue and what could be a mere subterfuge for plain old economic protectionism. The SPS Agreement provides no guidance and offers little scope for incorporating cultural analysis into SPS tr ade issues (Caswell 2000). Wh ile recognizing the EU rights to adopt a precautionary approach on a tempor ary basis, the Panel argued that the ban per se was not in conformity with the SPS Agreement and that there was no sufficient scientific evidence of dange rous effects on human health associated with the consumption of meat products treated with hormones. The hormone case provides some

PAGE 45

31 preliminary indication that the WTO Dispute Se ttlement Body is more likely to base its verdicts on a rational relationship between obj ective scientific assessments and the policy choices made by governments. Another interesting case was the Mexico complaint about the U. S. ban imposed since 1914 on the importation of Mexican Ha ss avocados, on the grounds that the fruit was a host of various fruit flies and that its importation could also lead to the importation of quarantine pests of concern. Mexico conducted in situ expe riments to demonstrate that the Hass avocado fruits attached to the tree are biochemically and ecologically resistant to infestations12 with A. serpentina, A. ludens, and A. striata under natural field conditions and there was only a minimal ri sk associated with importing this crop (Hoeflich 2000). While scientific argument s and evidence may have been a necessary condition for trade liberalization, they were certainly not a sufficient condition. For over 80 years, no trade protocol could be reached prior to NAFTA between Mexico and the United States. The policy process was capture d by Californian producers benefiting from the monopoly over the U. S. avocado market. However, the approval of NAFTA contributed to the end of this monopoly situ ation, providing space for negotiations and an opportunity for science to take part in th e decision-making process (USDA 2001, Federal Register 2004b; Hoeflich 2000). The above cases show the complexity of th e issues involved in the application of the SPS Agreement. Politics, economics and cultu re often play prominent roles in the choices of regulatory measures to address risk management problems (Caswell 2000). In particular, economics may play a strong role in measuring costs and benefits of SPS 12 It is important to underline that, under laboratory and field force conditions, Hass avocado fruits are a good host of A. ludens, an average host of A. serpentina, and a poor host of A. striata.

PAGE 46

32 measures and ranking them on the basis of how much they will improve (or undermine) the social well-being. Ultimately, policy choi ces emerge from the interaction of the demand for measures by various domestic interest groups (including producers, consumers, and processing industries) a nd the supply of barriers available to policymakers. Few people disagree that the SPS Agr eement has contributed to improve international trading relationships among country members and to restore the rule of law in this fractious area (Thornsbury 2002). Tr ansparency of the WTO members has been enhanced. Procedures and rules for dispute set tlement have been established and followed by WTO members. More importantly, regionaliza tion efforts have fostered the alignment of national regulations with international standards for phyt osanitary measures and the compliance with the obligations of the SPS Ag reement. The risk paradigm of the SPS Agreement has also reduced the degrees of freedom for the disingenuous use of SPS measures to restrict imports in response to narrow interest group pressures (Roberts 2000). However, the SPS Agreement still face s the ongoing challenge of how to lay the foundations of the SPS measures on scientific restrictions, while allowing for more flexibility in the use of economic and cu ltural considerations. Empirical evidence concerning the extent of questionable techni cal measures in international agricultural trade is very difficult to asse ss, because of unavailability of comprehensive data sources (Thornsbury 2002). Furthermore, despite several attempts over the years to resolve the USA/EU dispute over the hormone-t reated beef, the ban is still in effect (Pantin et al. 2004), leaving substantial room for questio ning the capacity of the SPS Committee and the Appellate Body of the SPS Agreement to enforce authorized sanctions.

PAGE 47

33 Phytosanitary Protocols for the Internat ional Movement of Fresh Fruits and Vegetables in Fruit Fly Free Areas Tephritid fruit flies are among the most dest ructive agricu ltural pests threatening the sustainability of fruit a nd vegetable industries in many pa rts of the worl d. Millions of dollars are spent annually in fly-free countries to enforce quarantine restrictions with a view to preventing the introduction of exotic fruit fly pe sts and maintainin g their fly-free area status. On the other hand, fruit producer s in fly-infested c ountries invest in expensive post-harvest facilities and treatmen ts in return for gaining and/or maintaining access to export markets. The application of the SPS Agreement has materialized essentially into the establishment of phytosan itary protocols serving as mechanisms for facilitating trade and transferri ng the cost of enforcement of the SPS restrictions onto the exporting countries (Spreen et al. 2002). The agreed-upon phytosanitary protoc ols between importing and exporting countries are inspired by the fly-free producti on model that often us es the concept of the systems approach. A fly-free field may refer to small areas such as a farm, an orchard, or a group of properties. Its delimitation requi res a thorough knowledge of the biological profile of the pest concerned (FAO 1996) a nd a risk management strategy aimed at maintaining pest freedom from specific pa rcels and production areas (Malavasi et al. 1994). However, attention is now moving from a fly-free field to fly-free zone approach (Vijaysegaran 1994; IAEA 1995). The fly-free zone concept encompasses an entire geographic or political entity in wh ich permanent eradication efforts such as intensive use of inspection sites at ports / airports and road stations massive sterile fly barrier, trapping surveys, and regular sprays against fruit flies -result in freedom of all fruit fly

PAGE 48

34 species of quarantine importance (Hendr ichs et al. 1983; Hendrichs 1996). Pest population suppression over large geographica l areas would undoubtedly provide better control and benefit a large number of growers. Apart from the success of an eradication project in Japan, eradication of a fruit fly sp ecies by sterile insect technique (SIT) and/or other means has not been reported in othe r Asian countries (Kawasaki 1991). Major concerns would be the cost of the sterile inse ct technique and the increasing risk of reinfestation. Table 3-1 presents three different form ats of phytosanitary protocols agreed upon between importing and exporting countries. Phyt osanitary requirements for international movement of fresh fruits and vegetables vary according to pest stat us and distribution, number of other key pests, isolation, geogr aphical location, techni cal level of fruit production, economic value of the crop, and chan ges in target markets (Hendrichs 1996). Format I (see column 1of Table 3-1) account s for the basic phyt osanitary protocol providing for detection surveys in and around the designated producti on areas to assure pest freedom against fruit flies of quarantin e importance in the field. This quarantine restriction is the centerpiece of the phytosanitary protocols governing (1) the shipment of fresh citrus fruit from the United States (Flo rida, California, Arizona and Texas) to some specific ports of entry in China and (2) th e importation of watermelon, squash, cucumber, and oriental melon from the Republic of Ko rea to the United States between December 1st and April 30 (Federal Regist er 2003a). Minimum trap densities based on research findings are specified in these protocols, as are the frequency of tr ap servicing and the suitable trap locations for the surveillance of flies of quarantine importance. The National

PAGE 49

35 Plant Protection Organization (NPPO13) of the exporting country must provide monitoring reports on trap surveys and releva nt information on groves, shippers/packers, and storage facilities for annual review of the protocol. Preinspection visits by the NPPO of the importing country to the exporting country are provided for to conduct a review of the certification procedures. Travel expenses (i.e. transp ortation, lodging, and a per diem allowance) for all trips during the first two years of the agreement to designated groves, shippers/packers, and storage facilitie s are borne by the exporting country. Certification procedures are based on negative trapping, providing scientific evidence that the concerned pest s do not occur in designated areas and that this condition has been officially maintained. Exported fr esh commodities must be transported under closed conveyance and kept separated from packed commodities from non-designated areas. The NPPO of the exporting country sh all perform a strict inspection of export shipments and ensure that exported products ar e free of quarantine pests. No post-harvest treatment is required in the protocol. In the ev ent of detection of a live fly on arrival, the importing country shall immediately notify the exporting country about suspending the importation of fresh commodities from the desi gnated grower or grove, shipper/packer and storage facility, and the shipment shall be returned, re-exported or destroyed. The suspension of the phytosanitary ce rtificate shall be maintained until the relevant cause is identified and appropriate co rrective actions are taken. Format II (see column 2 of Table 3-1) diffe rs from Format I in that the former involves a more stringent and complex certi fication process, reflecting the dominant 13 Each member country is required to form a National Plan Protection Organization (NPPO), which accounts for the governments official service to discharge the functions specified by the International Plant Protection Convention (IPPC), deposited in 1951 with FAO in Rome, and subsequently amended (NAPPO 1998; Federal Register 2002).

PAGE 50

36 position of countries like Japan and South Kor ea (which are regarded as lucrative markets for fresh products) [Vijaysegaran 1994; Si mpson 1993; Reiherd 1992]. Cases of this format are exemplified in the phytosanitary pr otocols for the shipment of Florida citrus fruit to Japan and South Korea (APHIS 2003b). Th e cornerstone of this protocol format is the requirement that the designated producti on area be surrounded by a buffer zone of 1.5 mile, which should not contain any preferred host plants of the pest concerned. Where a preferred host plant does occur in the buffer zone, ground or aerial bait spray shall be applied at 7 to 10-day interval s beginning 7 days prior to ha rvest and continuing until the end of harvest. Also, the minimum size of th e designated area is required to be 300 acres. Such a phytosanitary restriction is designed to stimulate the development, among fruit growers, of a concerted fly population manage ment strategy over a significantly large area, so that the potential number of flies moving into orchards from neighboring orchards is largely re duced. (Hendrichs 1996). Early season certification criteria for grapefruit shipped during August 1st to December 20 are less restrictive. Such proced ures are based on the proven resistance of early season citrus fruit to Caribfly infestation, which is regarded as a fly-free period (Simpson 1993). Nonetheless, th e standard certification proc edures appear to be very stringent, requiring an early su rveillance of the Caribbean fl ies (trap servicing beginning 30 days prior to harvest) and a high trap density (15 traps per square mile). Format II also allows for a second certif ication procedure referred to as bait spray with the following requirements: 1) the minimum size of the designated area must be 40 acres (16 hectares) surrounded by a 300-f oot wide buffer zone, 2) the buffer zone must be free of preferred host plants, 3) the designated area must be at least mile from

PAGE 51

37 areas where numerous host plants are present, 4) traps are located in the designated area and 300 feet buffer zone at the density of 15 traps per square mile with trap servicing beginning 30 days prior to harvest, and 5) aer ial bait sprays are appl ied at the beginning of the harvest period, consisting of a mixtur e of 2.4 ounces of 91 percent malathion and 9.6 ounces of protein hydrolyzate bait per acr e. Where the designa ted area is located within mile of numerous preferred hosts, ae rial bait sprays shall be applied earlier, beginning 28 days before harvest at 7 to 10day intervals until the end of harvest. In the event of a Medfly outbreak, countries like Japa n, South Korea, and South Africa take a more drastic approach by prohi biting the importation of all Medfly hosts including marginal hosts for a number of years (USDA1999). Japan, in particular, will not issue any phytosanitary cer tificate for any Medfly host commodity from the infested country, even if a quarantine treatment approved by the country of origin is applied. Such phytosanitary restrictions are justified by th e seriousness of the Medfly attacking over 300 fruit and vegetable commodities. Ne vertheless, other countries like Mexico, Argentina, and New Zealand adopt a more flexible appr oach based on the fly-free production model. The phytosanitary regula tions include the establishment of a quarantine area around a radius of 17 miles fr om the epicenter of the outbreak, intensive trapping surveys, and the treatment of Medf ly hosts according to agreed-upon treatment schedules. The third and last protocol format (column 3 of Table 1) provides an illustration of the peculiar use of a systems approach based on the notion of low pest prevalence. This strategy reflects the delicate pos ition of the United States --as both net exporter of some fresh agricultural commodities and net importer of others --switching from its long-

PAGE 52

38 standing practice of only recognizing entire countries as free or not free of a particular disease to a regi onalization regulation (Ahl a nd Acree 1993; Federal Register 2003a). Another significant factor is the fact that the U.S. territory is not free of fruit flies. By allowing the importation of fresh fruit commodities in the mainland from Hawaii where the Mediterranean fruit fly has be en established, APHIS finds itself forced to abide by the principle of na tional treatment and, therefore, to apply the same rules to imported products. These considerations suppor t the tendency towards the adoption of risk management strategies combining va rious preand post-harvest actions and treatments sequentially so that they can pr ovide acceptable statistic al probabilities of quarantine security (Armstrong 1991). Two cas es fall within this systems approach framework: the importation of clementines, mandarins, and tangerines from Chile (where the Mediterranean fruit fly and other quarantin e pests of concern are known to occur) under a series of complementary phytosanit ary measures. The requirements include a test program of certifica tion of low prevalence, a pos t-harvest processing, and phytosanitary inspection (F ederal Register 2004a) the importation of fresh Hass variety avoca dos from Mexico into the United States, using pest risk mitigation strategies comb ining two major tactics: (1) limiting the geographical distributio n of avocados to 19 States and the District of Columbia within the Unites States and (2) allowi ng a 4-month shipping season each year (Federal Register 2004b). The efficacy of the systems approach is based on the combination of complementary measures acting independently to ensure an a ppropriate level of phytosanitary protection (FAO 2002). For instance the phytosanitary rest rictions act in a fail-safe manner, so that redundant safeguards are built into the process. If one mitigation measure is not completely successful, the ot her will ensure that the risk of pest introduction is insignificant. Furthermore, a ny pest detection or irregularity would result

PAGE 53

39 in immediate actions to eliminate the pe st risk, including par tial cancellation of phytosanitary certificates or to tal prohibition of imports. It is worth noting that the phytosanitar y protocols between exporting and importing countries have been products of continuous and intense negotiati ons. Disputes over SPS standards go beyond the eviden ce of scientific phytosanit ary restrictions. The most important issue is often examination of the implications these scientific protocols would have on the safety of domestic producers. In fact, policymakers tend to place greater weight on producer rather than consumer welf are. Cases dealing with food safety issues are even much more difficult to resolv e as producer groups could easily mobilize consumers to back their claims for more re gulatory protectionism. What really makes possible the establishment of these phytosanitary protocols is the opportunity provided by the SPS Committee to air grievances over unjusti fied measures when bilateral technical exchanges reach an impasse. Despite fierce opposition from domestic interest groups, regulatory agencies are often fo rced to arrive at some acceptable arrangement with their trading partners by fear of retaliatory measures and / or unnecessary reciprocal phytosanitary barriers on domestic exports. Regulated Post-harvest Treatments and Pr ocedures for the Quarantine Control of Fruit Flies This section turns to the analysis of the regulated postharvest treatments that are often required for allowing unrestricted movement of tephritid fruit fly host commodities in domestic or international commerce. No qua rantine treatment is universally applicable to all products or all quara ntine pests (Mitchell and Ka der 1985). Each treatment has some inherent problems and limitations. Agri cultural research and regulatory agencies have developed various tests to evaluate pot ential quarantine treatments against fresh

PAGE 54

40 commodities infested with different fru it fly life stages. Two major tests --confirmatory test for quarantine security and small test for efficacy --are commonly used for the approval of a postharvest treatment schedule. Qu arantine security refers to the level of confidence that the quarantine treatment w ill disinfect quarantine pests from the host commodities (Armstrong and Couey 1989). Specif ically, it involves compliance with the phytosanitary requirements defined by the NPPO of the importing count ry to ensure that the quarantine pest of concern cannot become established in any geographical area where it does not already exist. One problem in quarantine treatment devel opment is the lack of availability of appropriate statistical criteria that ca n guarantee quarantine s ecurity (Couey 1983). Although the probit 914 mortality at the 95% confidence level remains the quarantine security statistics most commonly used in th e post-harvest technol ogy literature, it does not properly indicate the risk of a pest spec ies spreading into non-in fested areas (Landolt et al. 1984). Alternatives to the use of pr obit 9 mortality have been proposed, but not fully developed, such as the probability of ma ted pairs of quarantine pests in a shipment of a host commodity (Landolt et al. 1984), the host/pest relationships and natural infestation rates (Couey and Chew 1986). Commercially applied treatments are also monitored for efficacy. The term efficacy describes a quarantine treatment that adequately disinfests pest organisms at the required level of quaran tine security without advers ely affecting the commodity 14 The probit 9 statistics at the 95% confidence infers that no more than 3 individuals from a population of 100,000 will survive a quarantine treatment, which is a mortality rate of 99,997%. The treated population must equal 100,000 or more target organisms in three or more tests with no survivors. So, the treated populations are derived for the survivors of the control population by the formula: (A/B)C= estimated treated population, where A = the population of survivors from the controls, B=the control weight (untreated commodity), and C= the treated commodity weight (Spitler and Couey 1983; Armstrong et al. 1984).

PAGE 55

41 (Armstrong and Couey 1989). While APHIS as th e regulatory agency is not liable for damages caused by the quarantine treatment, ot her Federal and State agencies like the Food and Drug Administration (FDA), the E nvironmental Protection Agency have primary responsibility for ensu ring that approved control tr eatments are not harmful to the commodity, workers, the consumer, or the environment (Federal Register 2002). The number of approved fumigants and fumigation schedules has drastically diminished over the last two decades, due to environmental problems, health concerns, and lack of research. Following the cancella tion of ethylene dibromides registration by the U.S. Environmental Protection Agency (F ederal Register 1984), the methyl bromide (MB) has remained the most widely used fumigant for horticultural commodities because of low cost, ease of application, relative sa fe usage, rapid disp ersion throughout the fumigation chambers, and rapid penetration into the commodity (Mitchell and Kader 1985; Armstrong and Couey 1989). Approved MB concentrations, durati ons, and temperatures depend upon the commodity and the fruit fly species to be cont rolled. Strawberries infested with Ceratitis capitata can be safely transported from th e quarantine area after MB fumigation schedules at 15oC or above with 48 g/m3 for 3 h (Armstrong et al. 1984). MB dosages required to achieve 99.9968% kill (probit 9) of Anastrepha suspensa infestations of grapefruits at 21o-24oC are proven to be 40 mg/liter for a 3-carton load and 56 mg/liter for a 12-carton load fumigated in 0.8 m3 chambers (Table 3-2, column 1). However, final acceptance requires tolerance and residue tests under the variety of conditions encountered in processing and shipme nt (Benschoter 1979). Following successful cucumber fumigation schedules at 19oC or above with 32 g/m3 for 4 h, the results of

PAGE 56

42 phytotoxicity tests revealed very minor phyt otoxic effects that cannot affect the marketability of the product (Armstrong and Garcia 1985). However, the safety of MB fumigation ha s been called into question because of reports of carcinogenicity on laboratory animal s. MB residue levels in papaya, tomatoes, bell peppers, bananas, and eggplants treated ag ainst Ceratitis capitata are showed to range from 1.7 to 42 ppm, depending on the met hod of fumigation, the commodity, and the method of storage (Baker 1939; Seo et al, 1970, 1971). Therefore, al ternate treatments may be needed if agriculture and food s upplies are to be protected. One potential candidate fumigant that is receiving attenti on is phosphine, which offers some advantages in terms of rapid diffusion throughout the lo ad without a re-circu lating system, rapid dissipation of very low residues, and tole rance to fumigation by avocados, bananas, tomatoes, and bell peppers (which are injure d by fumigation with methyl bromide) [Seo et al. 1979]. The main disadvantage of phosphine fumigation is, however, the long treatment time (2 to 4 days instead of 2 hours), which makes it less promising for perishable commodities. Intensive refrigerated fumigation facilities would be needed at a very high cost to keep perishable comm odities under refrigeration during that long exposure time and, consequently, avoid unaccept able deterioration (Mitchell and Kader 1985). A combination of MB fumigation and co ld treatment seems to be a more economically feasible alternative. This techni que was used in Califor nia to disinfect fresh stone fruits during a Medfly out break. Treatment schedules list ed in APHIS quarantine treatment manual specify fumigation with 32 g/m3 for 2, 2.5, or 3 h, followed by cold storage for a minimum of 3 days to a maximu m of 11 days at temperatures ranging from

PAGE 57

43 a low of 0.55oC to 13.33oC (Code of Federal Regulations 2003; Spitler and Couey 1983). The cold portion of the treatment is often c onducted aboard ship, while in transit from infested areas. Shipments of cold treated fr uits are certified upon compliance with strict requirements for temperature monito ring in cold storage facilities. Cold treatments alone have been accepted by USDAAPHIS as quarantine treatments for 14 fruits and vegetables fr om 48 countries subject to infestations by Ceratitis capitata, Anastrepha ludens, and other species of Anastrepha (APHIS 1976). Treatment schedules are very severe (Table 3-2, column 3) involving fruit exposure to temperatures below 5oC for extended durations (from 10 to 16 days). The highest temperature listed in the Plant Protecti on Quarantine Manual (APHIS 1976) is 2.2oC. Probit analyses also predicte d that 16-20 days would be required for fruit held at temperatures ranging from 2.8 to 6.6oC (Burditt and Balock 1985). Such quarantine requirements severely limit the use of cold treatments, since most tropical fruits are damaged by extended storage below 10oC. Under cold treatment conditions, a certain percentage of fruit always exhibits chilling injury symptoms. In pa rticular, susceptibility of grapefruit to low temperatures is proved to vary with season and fruit location on the tree. Fruit on the outer canopy are more suscepti ble to chilling injury than those harvested from the interior of the tree. Reducing losse s in grapefruit shipme nts involves strict compliance with a series of basic requirem ents such as: avoidi ng prolonged degreening, ensuring proper application of fungicides, using stable water wa x, and providing proper warming before fumigation and after completi on of cold treatment (Ismail et al. 1986). Recent studies also show that preconditioning at warm temperatures increase fruit tolerance to cold treatments (Mitchell and Kader 1985). For instance, papayas infested

PAGE 58

44 with eggs and larvae of D. dorsalis, D. cucurbitae, and C. capitata are successfully treated after immersion in 49oC water and exposure to 8o or 9oC for 10 days (Couey et al. 1984). Nevertheless, the use of combined heat and cold treatments requires knowledge of the biological profile of the insect, the lethal effects of the standard heat treatment for decay control, cold treatment, and appropriate quality control procedures. Papaya fruits are required to be picked between colorbreak and one-quarter ripe Unripe papayas are unlikely to be infested with fruit flies, as containing sufficient quantities of benzylisothiocyanate to deter oviposition and to re duce survival of eggs and larvae whenever ovipostion occurs (Seo et al. 1982; Seo and Tang 1982). Hot-water treatment also may be used al one to disinfect fruit. Exposure to hot water at 45oC for 20 min or at 55oC for less than one min wo uld destroy all immature stages of D. dorsalis, D. cucurbitae or C. capitata (Armstrong 1982), but it would easily damage many fresh commodities (Table 3-2, column 2). Sinclair and Lingren (1955) found that navel oranges, lemons, and a vocados were very easily damaged by the standard vapor heat treatment. One in six unripe papayas is damaged after exposure to microwave until a central temperature of 45oC, followed by a double-dip in 48.7oC and 24oC during 20 min respectively (Hayes et al. 1984). Cases of tolerance to heat treatments are reported to very few commodities. Damage to valencia oranges and grapef ruit treated with standard heat vapor15 or the quick run-up can be avoided (or reduced) if the fruits are hydrocool ed after treatment. Satisfactory results have also been obtained with mangoes immersed in 50o-55o C water 15 The vapor heat treatment consists of gradually warming infested fruit for several hours (approach time), then increasing the fr uit temperature to 43oC and holding that temperature fo r 8 h (holding time). The quick run-up requires a short pre-heat period to a sp ecified temperature, then a gradual warming to 47oC, similar to the approach time in the standard vapor heat treatment (Baker 1952; Ba lock and Kozuma 1954).

PAGE 59

45 for 15 minutes (Sharp and Spaldi ng 1984) or vapor heated at 46oC for 160, 220 and 280 minutes (Mitcham and McDonald 1993). Concern is, however, growing about the feasibility of heat treatment due to the relatively high cost of application. Vapor heat cost per kg of fruit is estimated at $ 0.35 ag ainst $ 0.027 and 0.26 for methyl bromide fumigation and irradiation re spectively (Federal Register 2002, 2003b; Pszczola 1992). Other alternatives to fumigation for insect quarantine treatments have also been explored, such as the controlled atmosphe re and irradiation treatments. Forced air treatments (Table 3-2, column 4) consist of exposing fresh horticultural commodities to moderately low levels of O2 ( 2%) and /or high levels of CO2 ( 50%). Although modified atmosphere is cost competitive to chemical fumigation (Soderstorm et al. 1984) and leaves no chemical residues on the fru it, the possibilities fo r commercial use are much less certain. Most fresh fruits and vegetables cannot tolerate such extreme atmospheres for prolonged storage periods (Armstrong and Couey 1989; Smilanick and Fouse 1989; Yahia et al. 1991; Ke and Kade r 1992). Keitt mango is, however, reported to be very tolerant to insecticidal O2 and/or CO2 atmospheres for up to 5 days (Yahia 1993). Further research work is needed to de termine the level of tolerance of other mango cultivars and the level of mortality of important quarantine insects. Radiation is by far the most publicized quarantine treatment, mainly because of renewed interest in this treatment as a poten tial alternative to th e use of chemicals. Treatment schedules (Table3-2, column 5) involve exposing the pr oduct to a radiation source for a time period sufficient for it to ab sorb a required dose level of gamma or X rays. Successful use of this procedure is ba sed on the determination that an undesirable organism will be inactivated at a dose leve l that is tolerated by the host commodity

PAGE 60

46 (Mitchell and Kader 1985). For safety c oncerns, the Food and Drug Administration (FDA) establishes a dose limit of one kilogray (Federal Register 2002), which is by far less than all irradiation doses contained in APHISs rule. This quarantine policy is consistent with study findings indicating that most insects are sterilized by doses below 0.75 kGy. Unlike the heat, cold, and fumigation treatmen ts that generally kill the pest, what really matters in irradiation is the treatmen t dose required to break the life cycle of the insect, so that it cannot become establishe d in a new uninfested area (Rigney 1989; Baker 1939). Treatment efficacy is based on prevention of adult emergence. The killing of fruits flies in order to minimize fruit damage from feeding insects is of secondary importance. While the criterion of quarantine security may be fully achieved w ith a low irradiation dose, the marketability of the fruit is likely to decrease, due to the potential presence of large numbers of fruit fly eggs and/or larvae in the fruit. Achieving a lethal effect on fruit fly eggs or larvae would require doses in excess of 1 kGy, which would be damaging to many fruits. Research findings indicate that irradi ation provides acceptable quarantine treatment for various fresh commodities infested with fr uit flies. For instance, probit 9 quarantine security is reached with doses below 150 Gy for grapefruit and mangoes infested with Caribbean fruit flies, causing acceptable levels of phytotoxicity to the fruit (von Windeguth 1986, 1987). Phytotoxicity tests at doses ranging from 50 to 1500 Gy indicated that no observable damage occurr ed at levels between 50 and 500 Gy when Arkin carambolas were irradiated at 25oC and then held for 9 days at the same temperature (Gould and von Winderguth 1991). Furt hermore, irradiation of foods offers

PAGE 61

47 consumers many advantages the most important of which could be safe transport of produce from insect quarantine areas and repl acement of less safe chemical fumigants (Bruhn et al. 1987; Schutz et al. 1989). However, there are a number of considerations that dictate extreme caution in projecting radiation as an alternative to fumigation for in sect quarantine treatment. First, consumer acceptance is critical to the applica tion of irradiation and the realization of the advantages it offers. Consumers have expr essed growing concerns about the safety of irradiated foods, due to negative advertisements sponsored by opposition groups. Although consumers show a higher level of c oncern for chemical sprays and pesticide residues than for food irradiati on, it is still hard to sell irradiation to the American public (Schutz et al. 1989; Bruhn et al. 1986). Secondl y, most fresh fruits offer no promise for commercial irradiation because either alte rnative procedures are cheaper and more effective or radiation injury is excessive. Irra diation is unlikely to compete at present cost with chemical inhibitors or to substitute fo r refrigeration. In the future, feasibility of irradiation will depend on whether the phase-out of methyl bromide as a soil fumigant results in an increase in its unit cost of production (Federal Register 2002). Concluding Remarks This literature review on the SPS Agr eement emphasized the new substantive and procedural disciplines established to ach ieve harmonization of SPS regulations and minimize their adverse effects on agricultural trade. Countries are granted the rights to choose their appropriate level of protection (ALP) against im ported pests and diseases, but their regulations must be supported by sc ientific criteria. Thus, the SPS Agreement encourages the use of the least trade rest rictive measures, thereby making provision for

PAGE 62

48 the inclusion of the costs of prevention a nd control programs as the major factor in regulatory decisions. The application of the SPS Agreement ha s materialized esse ntially into the establishment of agreedupon phytosanitary protocols between exporting/importing countries. These protocols serve primarily as mechanisms for f acilitating trade and transferring the cost of enforcement of SPS re strictions to the expor ting countries. In the event of a Medfly outbreak in Florida, it is anticipated that th e reactions of export markets vis--vis Florida fresh fruit and vegetable commodities would greatly differ, ranging from additional certification, quarant ine treatment, to prohibition. Florida growers would be unlikely to reach a tradi ng arrangement with countries like Japan and South Korea, which are regarded as lucrativ e markets for fresh citrus. Negotiations would be very tense as to whether their total prohi bition would affect Fl oridas total fruit and vegetable production or the portion of produc tion localized in the quarantine area. Other countries would allow for the importa tion of fresh fruit and vegetables from Florida under very strict phyt osanitary restrictions. Given the seriousness of the Medfly attacking over 200 species of fruits and ve getables, the phytosanitary restrictions are expected to be more stringent that those contained in the Caribbean fruit fly-zone certification protocol, including intensive trapping surveys, establishment of quarantine areas around 17 miles from the outbreak epicente r, and regulated postharvest treatments. Treatment schedules would vary across commo dities, including fumigation (strawberries and cucumbers), cold storage (fresh citrus ), combination fumigation and refrigeration treatments (avocados), and vapor heat treat ment (mangoes). Florida growers would be reluctant to applied some of APHIS-approve d quarantine treatments to commodities --

PAGE 63

49 such as bell peppers, eggplants, tomatoes --because either they are uneconomical or cause excessive fruit injury. More research work needs to be done on the technical and economic feasibility of the regul ated postharvest treatments. Price ES'DS ES AS' AS DSource: Spreen et al. 2002Quantity Figure 3-1 Inward shift in import supply resu lting from the imposition of an SPS barrier

PAGE 64

50Table 3-1 Formats for phytosanitary protocols for the internat ional movement of fresh fruit and vegetable commodities Format I ( column 1) Format II ( column 2) Format III ( column 3) 1. Properties Sites *Must be free of fruit fly, monitored by MCPhail traps. Groves, packing ho uses, and storage facilities must also be certified 2. Certification Procedures *Based on negative trapping -Annual notifications of the exporting country to the importing about the trapping surveys in the areas of production *Phytosanitary inspection -Pre-inspection visits (one per year) by the authorities of the importing country to supervise the process of certification at the expense of the exporting country. -Citrus fruit must be free of live insects and mites of quarantine concern and identified with labels indicating place of production, grower, shipper, and storage facility -Citrus must be transported in sealed container and kept separated from unapproved shipments -Exports restricted to specific ports of entry in the importing country 1. Production Sites *Must be free of fruit fly, located at specific distances from preferred hosts and surrounded by a buffer zone 1.15 mile. Packing facilities must be located 3 mile from infested area 2. Certification Procedures *Based on negative trapping -For early season, minimum size of production area is 300 acres. Trap surveys with 2 traps per sq. mile. If buffer zone has preferred hosts, ground or aerial applications at 7-10 day intervals, beginning 7 days prior to harvest -For standard season, mini mum size of prod. Area is 300 acres. Trap surveys with 15 traps per sq. mile. If buffer zone has preferred hosts, ground or aerial applications at 7-10 day intervals, beginning 7 days prior to harvest *Based on bait spray procedures -For early season, minimum size of production area is 40 acres. Trap surveys with 15 traps per sq. mile. If buffer zone, has preferred hosts, ground or aerial applications at 7-10 day intervals beginning 7 days prior to harvest. -For standard season, mini mum size of prod. Area is 40 acres, must be at least mile from preferred host. Trap surveys with 15 traps per sq. mile. Aerial applications at 7-10 day intervals, beginning 28-30 days prior to harvest 1. Productions sites *May be a pest free area or low prevalence area 2. Certification Procedures *Based on the use of systems approach combining two or several independent phytosanitary measures *Case 1 combining three sequential measures -Low prevalence Certification Test of low prevalence based on a random sample of fruit undergoing a washing process. If no pests are found al ong the process, the production site is certified -Post-harvest treatment Products treated in accord ance with the agreed treatment manual. Shipments must be accompanied by documentation indicating the type of treatment adopted -Phytosanitary Inspection Based on biometric sampling with an acceptance level of zeros infested units. *Case 2 combining two major tactics -A geographical ban: limiting the importation to some specific geographic areas -A seasonal ban: limiting the importation during some specific periods of the year

PAGE 65

51Table 3-1 Continued Format I ( column 1) Format II ( column 2) Format III ( column 3) 3. Measures to be taken if detection is found *Immediate notification to the exporting country and establishment of a regulated quarantine area with a radius of 17 miles from the center of the pest outbreak *If a Mexican fly is found, the shipment must be treated but no citrus fruit from the affected grove will be exported 4. Exceptional measures in case of a Medfly infestation *Increase the trapping density to one trap per square km and immediate suspension of citrus fruit from the regulated quarantine area with a radius of 17 miles from the center of the outbreak. Shipment of fruit transited through this quarantine area must be in sealed container 3. Measures to be taken if detection is found during inspection *Immediate notification to the importing country if one fly is found as a result of the trap survey *Withdrawal of the production area from the protocol season if three files are found. Reinstatement is conditioned to intense spraying after investigation 4. Exceptional measures in case of a Medfly infestation *Fruit & vegetables considered to be hosts of Medfly fly are prohibited entry from countries to be infested with this pest. Such commodities remain prohibited regardless of whether they have met the entry requirements of any other country 3. Measure to be taken if detection is found *If one live fruit fly is intercepted along the inspection process, the consignment will be reshipped or destroyed and a prohibition will be placed on further imports of the host material until corrective action is undertaken 4. Exceptional measures in case of a Medfly infestation Increase the trapping density to one trap per square km; establishment of a quarantine area found a radius of 17 miles from the center of the outbreak; treatment of all regulated articles according to approved schedules. Shipment of fruit transited through this quarantine area must be in sealed container Source: http://excerpt.ceris.pur due.edu/doc/ctrylist.html

PAGE 66

52Table 3-2 Regulated postharvest treatments, advantag es, limitations, and alternat ives under consideration Fumigation Heat Treatment Cold Treatment Controlled Atmospheres Irradiation Regulated schedules *MB fumigation at 21o C or above at NAP with 3248g/cubic meter MB for two to four hours. *MB fumigation at NAP with 32g/cubic metric MB for 3 hours at 21oC for citrus infested with Medfly. *MB fumigation at 27oC and NAP for 2 hours for citrus infested with Caribfly and Mexican fruit fly. *Vapor-heat treatment of citrus at 43oC for 8 hrs. *Hot-water preheat im mersion of papayas, bananas, & mangoes at 42oC for 20 min to disinfect against Medfly, Melon, Oriental & Caribflies. *Cold treatments at temperatures ranging from 0oC to 2.2oC for a minimum of 10 days and a maximum of 16 days respectively. Such treatments are used against Medly, Caribfly and other quarantine fruit flies *Exposure of fresh commodities to low levels of oxygen and high levels of carbon during some period of time *Limit dose of one KGy established by Food and Drug administration for disinfestations *Approved doses varies between 150-250 Gy across different fruit fly species *Dose varying between 50-150 Gy to achieve probit for grapefruit and mango infested with Caribflies Advantages *MB is the preferred fumigant for horticultural commodities be-cause of low cost, ease of application, and relative safe usage. *Papayas & mangoes are resistant to heat damage. *Heat treatments have the merit of effective fungicidal & insecticidal action and no chemical residue. *Cold treatments are very practical when used during transit from producing areas to distant markets *Cost competitive to chemicals *Fruit treated with forced air are free of chemical residue *Provides quarantine security by prevention of adult emergence *Strawberries and papayas show high poten tial for commercial irradiation because they tolerate high doses without excessive ham

PAGE 67

53Table 3-2 Continued Fumigation Heat Treatment Cold Treatment Controlled Atmospheres Irradiation Limitations *Growing health & safety concerns about the use of MB *Valencia oranges & grapefruits are resistant to vapor heat injury but other citrus are easily damaged. *Hot water treatment kills fruit fly eggs but a low percentage of larvae can survive if fruits begin to ripen. *High potential fruit damage and relative high cost of application *Use of cold treatments is limited because most fresh commodities are damaged by extended storage below 10oC *Very few fresh fruits can tolerate extreme atmosphere for extended storage periods *Consumer reluctance about the safety of irradiated foods. Hard to sell to the American public *Limited economic feasibility of irradiation while requiring high capital investments *Irradiation shows little promise for perishable commodities because of excessive injury Alternatives under investiga-tion Fumigation at 30oC, with decreased fumigation time and MB concentration is very promising due to losses in MB phytotoxicity and reduction in inorganic residues *Research on potential candidates for MB replacement *Satisfactory results are obtained through combining heat treatment with MB fumigation of stone fruits and papayas *Need further research word on the economic feasibility of combined heat and fumigation treatments *Preconditioning at warm temperatures reduce cold storage injury on some citrus like grapefruit *Proven technical feasibility of combined heat and cold treatments to disinfect papayas of Medfly and other fruit flies *Keitt mango is very tolerant to insecticidal atmospheres for up to 5 days. Further research work is needed to test other mango varieties and level of pest mortality *Need further research work on consumer acceptance of irradiated food *Need further research work on the economic feasibility of combined cold and irradiation treatments

PAGE 68

54 CHAPTER 4 COST ANALYSIS OF THE MEDFLY DETECTION AND ERADICATION PROGRAM Specification of the Bayesian Modeling Framework Overview of the Bayesian Decision Process This sub-section discusses the basic pr inciples behind Bayesian statistical inference. The problem is to use historical information about Medfly interceptions and trap sensitivity with a view to (1) computing the probabi lity of detecting a Medfly infestation into Florida at a given time and (2 ) determining the optimal trap density that minimizes the expected cost of APHISs pr evention, detection, and eradication program. Such an approach to statistical inference is called Bayes theorem, which combines some prior distribution and available data to fo rm a posterior or revised distribution. The underlying principle of this theory is that a ll uncertainties are desc ribed by probabilities: unknown parameters have probability distributio n both before the data are available and after the data have been obtained (Cox a nd Hinkley 1974). While the theory does not offer a formal guarantee of objective truth a bout the system under study, it does at least ensure some kind of internal consistency among related decisions by the same individual. The decision is called a Bayesian decision (o r the optimal decision) if the action chosen minimizes average (expected) lo ss that is associated with the costs of a wrong decision (Wonnacott & Wonnacott 1977) Let X1,., Xn denote a sample distribution in dexed by a continuous parameter While, in classical statistic al estimation, it is appropria te to treat the parameter as a

PAGE 69

55 single fixed value, the Bayesi an theory allows for treati ng it as a random variable to account for our knowledge and uncertainty rega rding the parameters value (Press 1989; Morgan and Henrion 1990). The function g () is called prior probability mass function, since it is determined before observing X in the current experiment; that is, is based on previous practical experience a nd understanding. The posterior or revised di stribution of for X1,,.., Xn is given by: d g L g L hx x x x x xN N N) ( ) ,...., ( ) ( ) ,...., ( ) ,....., (1 1 1 (4-1) where the first term in the numerator L(x1,..,xN|) is a likelihood function indexed by the parameter The conditional probability h (|x1,.,xN) is called posteri or probability mass function, given the current data, since it is determined after observing the current data set. The prior distribution provides a dditional information into the analysis and allows for a gain in logical clarity. Figure 4-1 summarizes the logic of a Ba yesian decision. The framework analyzes the trade-off between early versus late detec tion in terms of costs of the Florida Medfly prevention, detection and er adication program. Early de tection costs are high for trapping, but low for eradication (with a low probability of establishment). Late detection costs for eradication are high (with a high probability of establishment). The major components of the framework --past data on Medfly in terceptions, single trap sensitivity, prior probability of infestation, multiple trap sensitivity --are supportive of the optimal decision relative to the trapping strategy. The pr obability of detection f (x1| x2, z) is a posterior probabilit y, as its computation revises th e prior probabilities to reflect

PAGE 70

56 the observational information available on trap density and sensitivity. If the objective is to minimize the expected cost of eradication, th en this cost has to be computed for each possible trap density per location and seas on of the year. The optimal (or Bayesian) decision is to select the tra pping strategy associated with th e lowest expected prevention and eradication costs. Definition of the Variables We assumed the introduction, into Florida, of a small lot of mangos originating from a hypothetical Latin American country and containing approximately 1,000 Medfly eggs. We considered two co rrelated non-normal variables X1,t and X2,t: 1. X1,t stands for the total Medfly populati on present at a given time and some location in Florida 2. X2,t accounts for the number of Medflies captu red in the trapping system put in place by APHIS. The amount of informati on provided by trapping reflects both the density of traps and the single trap sensitivity. The objective of this model was to build a spatially and temporally demographic picture of a hypothetical Medfly infestation in Tampa and Miami. Sp atial distribution of Medfly host plants was assumed to be unifo rm throughout the regions of interest. The mate-finding process was also considered a cr ucial determinant of the potential for pest colonization in natural settings (Allee 1931; Prokopy and Hendrichs 1979). Along these lines, the infestation model was initialized with a cohor t of 50 ovipositing females resulting from a twenty-percent survival of preadult Medflies with a sex ratio of 1:1 (i.e. 100 female adults and 100 male adults). E gg production covered a 10-day period at a fixed rate of 11 eggs /female /day. Survival rates varied across stag es; egg survival is 0.4, larval survival is 0.5, pupal su rvival is 1. Only 20% of the newborn Medflies survived to adulthood.

PAGE 71

57 Temporal dimension The infestation model is inspired by th e time-specific life-table approach to population growth, which is grounded on the assu mptions of constant recruitment rate and steady mortality rate (Southwood and He nderson 2000). The rate of insect population growth is basically dependent on temperature and is predicted with time. Physiological time is commonly expressed in terms of day-degrees (Do) or hour-degrees (ho), being the cumulative product of total time x temper ature (above the threshold) [Hughes 1962; Hardman 1976; Atkinson 1977]. A minimu m temperature exists below which no measurable development takes place. The de velopmental zero for Medfly was found to be 54.3oF (Shoukry and Hafez 1979). Thus, th e number of day degrees, Do, accumulated above the developmental threshold for a life stage is computed as follows: ) ( ) ( ) (F TH F AV Fo o o oD (4-2) where AV stands for the average daily temperature in Fahrenheit and TH is the developmental threshold for Medfly in Fahrenheit. This temperature model is also used by APHI S to predict the entire life cycle of a Medfly population and to guide program ac tions (eradication treatments, length of trapping activities, and regulatory functions). About 328o C-day degrees (590.4o F-days degrees) must be accumulated before one lif e cycle has been comp leted (APHIS 2002). Table 4-1 shows the number of day-degrees required by one bug to transition from one stage to the next stage. Temperature da ta (minimum and maximum daily average temperature) used to simulate the growth development process are collected from the National Oceanic and Atmospheric Admini stration, United States Department of Commerce.

PAGE 72

58 The model design includes tw o locations (Miami or Tampa) at three different months of the year (February, June, or Oct ober) to allow for both regional and seasonal differences in the State. The analysis is car ried out over a 200-day period, using an Excel spreadsheet that provides information on the structure and the size of the pest population over time. Under each scenario, the instantane ous rate of population change, r, and the doubling time, DT16, are computed, using the following equations (Carey 1982b): m l et t t rt 01 (4-3) r DTe) 2 (log (4-4) where t is the age in days, lt is the probability of surviving to age t, mt is the number of female offspring produced at age t. The parameter, r, is key to predict the size of the Medfly infestation at different points of time, using the finite version of the Malthusian equation: e N N rt t 0 (4-5) where Nt and N0 stand for the size at time t and th e initial size of the population, respectively. Spatial dimension Spatial dispersion and movement is as im portant as birth and death rates for the population dynamics of insects and is a ma jor determinant of the boundaries of the infested region (Papadopoulos et al. 2002, 2003). The underlying assumptions for modeling spatial distribution of the infestation are that (1 ) the Medflies are considered 16 The parameter, DT, designates the time (in number of days) needed for the pest population to double.

PAGE 73

59 relatively weak dispersers and (2) the vast majority of th e population does not disperse very far (Soria and Cline 1962; Katiyar and Valerio 1963; Nadel and Guerrieri 1969; Serghiou and Symmons 1974; Wakid and Shoukry 1976; Plant and Cunningham 1991; Katsoyannos et al. 1998; Papadopoulos et al. 200 0; Papadopoulos et al. 2001). Table 4-2 gives the proportion of pre-adu lt and/or adult Medflies disper sed to different distances from the epicenter. Random dispersal is the dominant strategy of spread, occurring over an expected radius of 0.113 mile per month and thereby re sulting in the scattering of 29% of the original population and the expa nsion of the infested area. Cases of long distance flights in se arch for food and/or oviposition sites (Bateman 1972; Harris and Olalquiaga 1991) and of human conveyance ove r distances ranging from 24 to 100 miles (Williamson 1983) are also incorporated into th e model, as leading to potential outbreaks in new areas. Nevertheless, the probabilities of facing new Medfly outbreaks away from the initial epicenter are subject to the so-cal led Allee-effect, that is, the opportunity of finding a mate. If the number of adult female s moving away from th e epicenter is less than one, the outbreak potential is considered insignificant. This restriction is based on the assumption that the chances for the males to attract the females would be very low when the size of the lek formed by the males is too small (Carey 1982b). The model implies that the density of pests is highest at the epicenter and decreases in space as small portions of the pest populat ion move away from the epicenter. The notion of the epicenter of a pest infestation suggests that there is a point from which the pest infestation originates (Mangel et al. 1984). Figure 4-2 illustrates the spatial dispersion pattern of a hypotheti cal pest infestation occurrin g in Miami during the month

PAGE 74

60 of October at Day 50. Each 1-by-1-mile square parcel17 accounts for the infested unit area, while the number inside each square parcel stands for the number of pre-adult and/or adult females present in the square par cel. The size of infestation is the sum of the 1-by-1-mile square parcels and the quaranti ne area can be derived through delineating a 5-mi-wide buffer perimeter around all infested core areas. Cost Function The optimal decision (or Bayesian decision ) was made on the basis of the optimal trap density that minimizes the expected cost of the Medfly prevention, detection, and eradication program. We assumed that the co st function of the emergency program was approximated by the log functional form: )2 1 11 1 1 1 0ln (ln ln ln ) 1 ( ) ln(x xt tA A C (4-6) where C stands for the emergency program cost per ha, A is the infested area in ha units, and X1,t is the pest population density per ha. Th e size of infestation and the quarantine area are the key parameters used to calc ulate the total emergency program costs by summing (1) the cost of the emergency detecti on cost (DC), (2) the cost of the curative release of sterile f lies around all buffer perimeters (RC) and (3) the costs of weekly malathion applications (AC) over the infested area under scenarios of low, moderate, and high pesticide efficacy.18 The total emergency program co st (TC) under each scenario is computed, using the following equations: 17 Each 1-by-1-mile square parcel can be associated with what is called a section in the Township and Range system developed by the Federal government. Each township comprises 36 sections. 18 Pesticide efficacy is defined here as the propor tion of bugs that will not survive the weekly spray treatment. Thus, low, moderate, and high pesticide e fficacies correspond to 70%, 80%, and 90% of bugs killed during the spraying operation, respectively.

PAGE 75

61 DC=WEEKLY EMERGENCY DETECTION COST (# SPRAYING WEEKS +THREE LIFE CYCLES) (4-7) RC= (QUARANTINE AREA/100) STERILE RELEASE COST (# SPRAYING WEEKS +THREE LIFE CYCLES) (4-8) AC = AREA INFESTED APPLICATION COST UNIT # SPRAYING WEEKS (4-9) TC = DC + RC + AC (4-10) The spreadsheet models used to compute the number of spraying weeks required to eradicate the pest population take into account both the proportion of individuals surviving a spray treatment and the new adult females emerging during eradication operations. The infestation rates applied unde r spraying conditions are extremely low, ranging from 3 to 6 eggs per female over a three-day oviposition period. The Medfly population is considered totally eradicated in the model when the sum total of ovipositing females is less than 1. We computed the OLS estimates of the parameters for the cost function under each outbreak scenario (Miami or Tampa, February, June, or October). These estimates are shown in Table 4-3. The F statistic was used to test whether the regression coefficients are different in the different periods (Februar y, June, and October) or it is appropriate to pool the data and estimate a single equation for the entire period from February to October. The resulting F ratios are highly signi ficant, with calculated values for F [6,204] approximating 390.5533 and 440.2952 for Miami and Tampa respectively. So, consistent with our expectations, these results reject the null hypothe sis that the coefficient vectors are the same for the three periods in each location.

PAGE 76

62 Future cost of eradication As stated above, there is a trade off between early versus late detection. Maintaining high trap densities over a wide area to discover low, patchy populations of wild flies is extremely costly. On the other ha nd, if the trapping system is ineffective and fails to detect the pest population at early stages, APHIS managers would have to face the weighted values of the future cost of erad icating a growing pest population. Therefore, the future cost of eradicating the pest at time t, FU(X1,t), can be expressed as follows: T t s t t s t s s tx x x x d x x d C z f z f FUx x x x x x x x1 1 1 1 2 1 2 1 0 1 150 1 0 1 98 1 50 1 119 1 98 1..... ) ( ) 1 ( ) 0 ( ( ) ( (4-11) where F(X1,s,X2,s=0|z) and F(X1,s,X2,s=1|z) are the multiple trap sensitivities for X2,s =0 and X2,s=1, respectively. C(X1,t ,A) stands for the cost function of the emergency program. Optimization model To find the optimal trap density, the mi nimization problem can be defined as follows: ) 1 ( ) ( ) ( ..........1 2 1 1z f C FU PZx x x x Mint t Z (4-12) where Z is the variable accounting for the trap density and P is the detection program cost per trap unit. The calculation of P is base d on the assumption that 47,404 traps are placed in Florida over 4,490 square miles. The aver age annual price of a trap is roughly $168.22. The value of the objective function is com puted under all outbreak scenarios for the values of Z lying between 1 and 500. The optim al trapping density for each scenario is the one that minimizes the objective function.

PAGE 77

63 Probabilistic Models Probability of Detection: F (X1,t | X2,t, Z) The purpose is to calculate the probability of detecting the pr esence of a Medfly infestation at a given point of time, on the basis of prior information available (probability of infestation, single trap sensitivity of the trapping system, and number of traps). As shown in Figure 4-1, the probability of detection, F(X1,t | X2,t, Z) is defined as follows: F (x1,t| x2, Z) = Probability of detecti ng a Medfly population in Florida at a given time t, given the number of traps (Z) and trap sensitivity ) ( ) (2 2 1Z F Z Fx x xt (4-13) The term F(X1,t, X2,t | Z) in the numerator is a joint density function and, therefore, is the product of the multiple trap sensitivity [F (X2,t | X1,t, Z) ] and of the probability of infestation at t = 0, [ F (X1,0) ]. It can be calculated as follows: ) ( ) ( ) (0 1 1 2 2 1x x x x xF Z F Z Ft t (4-14) The term F(X2 | Z) in the denominator is a summation that encompasses all possible outcomes of the trapping system, including th e probabilities that the pest is present but not trapped, and the probabilities that the presence of pest is detected with different numbers of Medfly adults captured in the trapping system. x x x xt td Z F z F, 1 0 2 1 2) ( ) ( (4-15) Probability of Infestation: F(X1,0) This model of probability examines routes of Medfly introduc tion and uncertainties regarding fly survival and effectiveness of exclusion activities. It is assumed that opportunities of Medfly introduction into Flor ida are continuously increased through high volume of international travel, agricultural industry demands, and international trade agreements (USDA 1999). International air pa ssenger baggage is c onsidered the highest

PAGE 78

64 risk pathway for Medfly into Florida. About 8.8 million air passengers arrive in Florida per year, of which 5 million originate from Medfly infested countries (APHIS 1999). Weighted passenger with Medfly infested mate rials to Florida is estimated in 1995 at 6.6 x 10-10 based on baggage surveys carried out by Plant Protection and Quarantine from July, 1993 through September 1994 (APHIS 1994). Using the general framework developed by Ba ker et al. (1993), th e probability of a Medfly infestation in Florida can be calculated as follows: ) 1 (1 ) (0 1 p XNF (4-16) where N stands for the number of passengers pe r year arriving from countries where Medfly occurs, p is the proportion of weighted passenge rs with high risk materials, and is the probability that a single infested unit leads to an establishment. The parameter, p, is considered the infestation level associat ed with air passenger baggage clearance. To be 95% confident that the infestation level is no more than p, the number of passengers (n) to inspect can be calcul ated (Couey and Chew 1986) as: ) 1 log( ) 95 0 1 log(p n (4-17) Under the assumption that the number of survivors per infested unit follows a Poisson distribution, the parameter is defined (Whyte et al. 1996) as follows: ) 2 1 (2e e (4-18) where is the average number of pests present per infested unit, is the proportion of individuals surviving to reproduce, and is the suitability of conditions for the pest. Thus, the probability of a Medfly infestation is written as

PAGE 79

65 ]) 2 1 [ 1 (2 0 1 0 1 1 ) (0 1e e p XX X FN (4-19) where X(1,0) (X=) stands for the number of ovipositi ng females per infested unit. As stated above, all outbreak s cenarios are built up unde r the assumption that the infestation is initialized with a small lot of mangos containing 1000 eggs. The number of eggs per infested mango is assumed to be 100 with a ten-percent probabil ity of surviving to adulthood and a 1:1 sex ratio. Multiple Trap Sensitivity of McPhail traps: F(X2,t | X1,t Z) The multiple trap sensitivity, F(X2,t | X1,t Z), stands for the probability of capturing one or several flies by trapping, given the occurrence of an in festation at a given time and some location in Florida. It is a function of the size of the adult population and the number of traps. Our study used data on different population levels of a uniform distribution and age class with various numbers of McPhail traps per unit area (Calkins et al. 1984). The conditional probability of de tecting low to moderate populations of Anastrepha suspensa in citrus groves in Central Fl orida was described by a polynomial approximation of the cumulative distribution fu nction (cdf), which was constrained to the zero-one range by a hyperbolic tangent f unction (Taylor 1984, 1990). The functional form of the conditional cdf is written as: )) , ( tanh( 5 0 5 0 ) (2 1 1 2Z H Z Fx x x xt t (4-20) and the associated conditional probability density function (pdf) is computed by taking the derivative of Equation 4-20. Thus, the conditional pdf is: ) , ( ( ) , ( 5 0 ) (2 1 2 2 1 1 2secZ H Z Z fx x h x x H x xt t j t (4-21)

PAGE 80

66 where H(.) is a polynomial function, Hj is the partial derivative of H(.) with respect to x2, tanh(.) is the hyperbolic tangent, and sech(.) is the hyperbolic secant. By restricting the polynomial approximation to a quadratic func tion, Equation 4-21 can be extended as follows: ) ( ) 2 ( 5 0 ) (2 33 2 2 22 2 1 11 2 23 1 13 2 1 12 3 2 2 1 1 0 2 2 22 23 1 12 2 1 2secZ a x a x a x a x a x x a a x a x a a h x a a x a a x xt t t t t tZ Z Z Z Z f (4-22) The parameters characterizing the pol ynomial approximation are estimated by least-square error method. The minimiza tion problem is expressed as follows: N i i t i TT x x T x xZ f Z f Min1 2 1 2 2 1 2) )) 1 ( ( ) 1 ( ) 0 ( ( ..... (4-23) subject to: 000 10 34 ,.... ....... 0 ) 0 (, 1 1 1 2 x x x xt t tz f (4-24) 000 10 34 ..... ....... 0 ) 1 (, 1 1 1 2 x x x xt t tz f (4-25) 0, 1 11 13 2 12 1 12x a a x a a xt tz H (4-26) 0233 2 23 1 13 3z a x a x a a ztH (4-27) where f(x2=0, x1|z) and f(x2=1,x1|z) are the respective theoretical values of the probability derived from the hyperbo lic tangent function, Ti are the probabilities provided in the dataset, and H/ X1,t and H/ Z are the partial derivatives of H with respect to X1,t and Z. The last two constraint s (Equations 4-26 and 4-27) ar e imposed to satisfy the requirements that the probability density f unction be monotonic and increasing over the

PAGE 81

67 range of values of X(1,t) lying between 34 and 10,000 and Z varying between 1 and 500. All correspondent proba bilities are also constrained to be positive (Equations 4-24 & 425). The least-square error estimates provided by the Excel solver are given in Table 44. All constraints and optimality conditions are satisfied. The value of the standard error of the estimate accounts for the measure of the goodness of fit of the estimated regression line. Also, the Kolmogorov-Smirnov test statistics do not reject the null hypothesis that the theoretical va lues of the probability dens ity functions (derived from the hyperbolic tangent function) and the valu es of probability provided in the dataset follow the same distribution. Comparative Sensitivity of McPhail versus Jackson Traps Results obtained in experiments with McPhail traps can be properly extrapolated, using estimates of comparative trap sensitivity. Literature reported significant differences between McPhail and Jackson traps in th eir performance of capturing Ceratitis capitata in terms of probability of detecting small populations, number of flies captured, and proportion of females (females/[males + females]) captured (Heath et al. 1997; Katsoyannos 1994; Katsoyannos et al. 1999a, 19 99b; Papadopoulos et al. 2001). McPhail traps (ML traps) are baited with a dry synthe tic multilure/liquid protein and, such as, are used to capture both females and males of a number of pest tephritid species (Newell 1936; IAEA 2003). On the other hand, trimed lure compounds are typically placed in Jackson traps (TML traps) that are effectiv e in attracting males, but they are weakly attractive to females (Beroza et al. 1961; Nakagawa et al. 1970; Harris et al. 1971). Casana-Giner et al. (2001) argued that no attractant for female C. capitata was comparable to the male C. capitata captures of TML traps, which may be due to a higher

PAGE 82

68 response to odor-attraction of th e males than females in C. capitata Development of trimedlure compounds would contribute to enha nce the probability of Medfly control, improve trapping strategies, and re duce costs of trapping systems. We used capture data from experiments of sterile insect releases conducted by APHIS in Tampa to develop estimates of co mparative sensitivity of ML and TML traps by making the assumption that the experimental conditions (type of crop, density of crop, tree fruit distribution, trap location or positio n) were the same (Drummond et al. 1984). The full dataset covers 1700 obser vations of daily captures from 2nd to 27 February (20 counts), from 31 May to June 30 (23 counts), and from 1st to 30 October (22 counts). Trap counts were expressed as the mean num ber of flies caught per trap per day and transformed to [ln(catches +1)] to stabilize their variances before analysis (Katsoyannos et al. 1999a; Cohen and Yuval 2000; Papadopoulos et al. 2001). Using the data for the entire sample, February, June, and October, we obtained four estimated OLS regressions, postulated as follows: u catchesML catchesTML ) 1 ln( ) 1 ln( (4.28) where stands for the coefficient of comparative trap sensitivity. We used the F statistic for testing whether the unrestricted regressi ons (as opposed to the restricted or pooled regression) for the three periods were systema tically different. A probability level of 0.05 was used for all statistical tests. Results of the regression mode l (Table 4-5) show variati ons in mean daily captures and coefficient of relative sensitivity across seasons and types of traps. Both TML and ML traps were more effective in February, wh en the life cycle of the pest is longer and

PAGE 83

69 fly dispersal ion is restricted. Nevertheless, TML traps outperformed ML traps in total captures of male C. capitata during all periods. Mean dail y captures in TML traps are 5.80 times greater than in February, 9.84 times greater than in October, and 7.65 times greater in June. The t-test statistic supported the hypothesis that the co efficients of comparative sensitivity estimated in the OLS regressions are statistically different from zero. Confidence intervals for the estimated coefficients are shown in Table 4.5. On the basis of the F test, we strongly re jected the null hypothesis of hom ogenous slope coefficients. The test sequence is naturally halted, as th e regression models (Equation 4.28) postulated do not contain any intercept. These findings are supportive of the conclusion that the coefficients of relative sensitivity vary syst ematically across seasons and types of traps. These coefficients are incorporated into th e hyperbolic tangent function as augmentation factors (Moss et al. 2004) to investigate th e effects of trap sensitivity improvement on probability of detection and optimal trap density. Consider ( ), the augmentation factor, being the technol ogical change. The multiple trap sensitivity of TML trap can be expressed as follows: ) ( ) (, 1 2 1 2Z ZX X F X X Ft t ML t t TML (4-29) Results The findings from the Bayesian modeling fr amework are outlined in this section. Our study provides estimates of the coloniza tion potential of Me dfly populations in Florida, probabilities of detecting low populatio ns at early stages, expected costs of the prevention and eradication program, and optim al trapping densities across different locations and seasons.

PAGE 84

70 Pest Population Projection Results of the temporal model of infesta tion (Tables 4-6 and 4-7) show variations in generation time across locations (Miami or Tampa) and seasons (February, June, and October). In the event of a Fe bruary infestation in Miami or Tampa, the expected time for F1 generation to reach the ovipos itional stage is 44 and 64 days respectively. At 50 days of the infestation, for instan ce, the projected size of th e pest population is 860 and 200 Medflies, respectively. The in stantaneous rate of change in population size is 0.0812 and 0.0587 for Miami and Tampa, respectively, implying that the pest population will double approximately every 8.53 and 11.80 days, resp ectively (Table 4-7). These results follow from the fact that average mean monthly temperatures in both Miami and Tampa are below optimal levels for Medfly development during the months of February and March. However, as the temperature increases during th e months of April and May, the length of the life cycle will be drastically reduced to 22 or 23 days for the fourth generation. The instantaneous rate of populat ion change in population will also increase to 0.1005 and 0.0988 in Miami and Tampa, respectively. As a result, the colonizat ion potential at 119 days will increase with a population si ze approximating 38,781 and 22,990 Medflies, respectively. As expected, the colonization potential of a Medfly population is much higher during the summer months, when average m ean monthly temperatures for Miami and Tampa are in the optimal range for Medfly development. The time to the ovipositional stage averages 21 days and the projected doubling time is 4.65 days and 4.83 days for Tampa and Miami, respectively. For instance, at 98 days of the infestation, the pest population size in Miami and Tampa is e xpected to be 363,390 bugs and 532,267 bugs, respectively. Differences in population si ze and structure between Tampa and Miami

PAGE 85

71 (Appendix A) spell the sensitiv ity of the Medfly development to changes in weather conditions. As opposed to a February infestation, th e October infestation starts under more favorable weather conditions, and, therefore, with a moderate colonization potential. At 50 days of the infestation, initial pest popul ations in Miami and Tampa are expected to attain a size of 1960 bugs and 1300 bugs, respec tively. The first two generations will increase on average by approximately 11. 06% and 9.33% each day, respectively. However, the duration of the egg, larval, and pupal stages will be considerably increased by lower temperatures during the months of December and January, causing a significant reduction in the colonization potential of the pest population. In Mi ami, generation times increase from 29 days for the second genera tion to 52 days for the fourth generation, while the intrinsic rate of population increase decreases up to 0.0777. In Tampa, the drop in the intrinsic rate of populat ion increase is even more seve re. The development of larval and pupal stages is expected to be almost st opping, thereby causing a so rt of stagnation in the pest population size. Size and Cost of the Infestation The results of the spatial model of the infe station (Table 4-8) show variations in infested areas, quarantine areas, and total eradication costs under different outbreak scenarios. Note that the erad ication cost is hereby estim ated under the assumption of a ninety-percent pesticide efficacy. None of the outbreak scenarios simulated in the spatial-temporal model can be considered early-detected infestations (w hich, according to the APHIS criteria, are supposed to spread over a quarantine area eq ual to or less than 110 square mile). Maintaining such a target involves detecting th e pest population and starting to spray it at

PAGE 86

72 less than 44 days. The reason for this is that, at 50 days, th e initial parent cohort under all outbreak scenarios (including both seasonal and geographical variations) is expected to start ovipositing, implying that the eradica tion will not be completed around the time required for all F2 eggs to emerge as adults. By this time, the infested area averages 22 square miles large with a quarantine area ranging from 121 to 421 square miles. The February infestation in Tampa is expected to be confined to a two-square-mile area, while the June infestation in Miami or Tampa will spread over a 32 square mile area and the total regulated area will encomp ass a 421-square-mile area. Nevertheless, any infestation detected at 50 days can be considered a moderate outbreak as the maximum distan ces flown by the flies are ex pected to be less than 4 linear miles and the expected pr oportion of adults moving away from the epicenter is low (< 10%). Total cost of eradica tion of a 50-day old infestation is expected to vary across locations and seasons, from $ 2.06 million for a February infestation in Tampa, $ 6.1 million for an October infestation, to $ 26.3 million for a June infestation in Miami. When spraying starts three weeks later (i.e., at 77 days of the infestation), the cost of eradication of a February or October infestation in Miami will approximately quadruple compared to the costs of the 50-da y old infestation in the same location and season. The 77-day old pest population in Miam i is expected to spread over a 21 and 24 square-mile area for the February and Octobe r infestations, respectively. However, the rate of increase in eradication cost is mu ch less in Tampa, approximating 48% and 110% for a 77-day-old infestation occurring in Oc tober and February, respectively. The Tampa infestation is expected to spread over a 12-square-mile area and less than 250 square miles will be placed under strict regulation. Theses results reflect well the sub-optimal

PAGE 87

73 weather conditions prevailing in Tampa for th e Medfly development, especially during the winter months. A 77-day-old infestation must be consid ered a serious outbreak when it occurs during the summer months eith er in Miami or Tampa. Av erage distances flown by the flies are expected to be greater than 12 linea r miles and the projected proportion of adults moving away from the epicenter will be around 12%. Such an infestation is very likely to spread over two to four coun ties. The expected treatment area covers 115 square miles large and 1,055 square miles are expected to be under strict regulation. The Tampa infestation will be more costly than the Miami one. The reason for this relative difference is based entirely on the size and the structure of the F1 generation in the two locations. The highest proportion of the pe st population in egg, larval, and pupal stages in Tampa spells the difference in eradication costs. As shown in Table 4-8, it will cost twi ce or three times more to eradicate the F1 and F2 generations of an October infe station than those of a Februa ry infestation. The reversal of this situation is observed for the F3 and F4 generations. At 98 days and more, whilst eradication costs of an Oct ober infestation in either Miami or Tampa increase at a decreasing rate, those of a February infestati on tend to increase at an increasing rate. In Miami, the projected eradication cost of a 98-day or 119-day old infestation will be approximately $ 147 million and $ 1.9 billion, respectively. At 119 days, for instance, the February infestation will spread over a 300-s quare mile area with high risks of facing additional outbreaks in remote areas during eradication operations. A 98 or 119-day old infestati on is even more serious when it occurs during the summer months in either Tampa or Miami. The expected quarantine area encompasses

PAGE 88

74 approximately 6,800 and 8,000 square miles resp ectively. Eradication costs are expected to be extremely expensive, approximating $ 2.9 billion and $ 7.7 billion in Miami and Tampa, respectively. It is likely that su ch infestations become out of control. Multiple Trap Sensitivities for ML Traps The estimates of multiple trap sensitivity fo r the ML trap (Table 4-9) are extremely low, thereby confirming how difficult it could be to detect low Medfly populations at early stages. As expected, the highest trap sensitivities are reported for the pest population during the summer months. For instan ce, the sensitivity of a trapping system to a 50-day-old infestation varies across trap densities from 4.95 x 10-6 for a density of one trap per square mile in Miami to 5.84 x 10-4 for a density of 21 traps per square mile in Tampa. The chances of detecting a 119-da y old infestation are higher, with trap sensitivities lyi ng between 1.35 x 10-4 and 1.11 x 10-2. On the other hand, the lowest trap sensitivities are found for a 50day old infestation during th e months of February and October. For instance, the chances of detec ting a 50-day old infesta tion in Tampa are ten times lower in February than in June. The marginal values of trap sensitivity (Table 4-10) are all positive within the interval ranges of our data set, varying from 7.08 x 10-7 to 6.31 x 10-4 with a clear tendency to increase with trap density. The higher the trap density, the higher will be the marginal trap sensitivity. For instance, for a given pest population in Miami (October), the marginal trap sensitivity to a 50-da y old infestation varies from 3.88 x 10-6 to 2.74 x 10-5 for trap densities of 2 and 21 traps per s quare mile, respectively. The direction of changes in marginal trap sensitivity is also the same with changes in pest population size. For a given trap density in Tampa (June), the marginal trap sensitivity varies from 5.7 x

PAGE 89

75 10-6 for a 50-day old infestation to 1.59 x 10-4 for a 119-day old infestation. These results strongly suggest that there is a positive gain in increasing trap density. Probabilities of Detection for ML Traps The computed values for the probability of detection (Table 4-11) differ from the trap sensitivities in that th e former are conditional to (1) the probability of infestation, F(X1), estimated at 0.005371 for the State of Fl orida and (2) to all possible outcomes in the trapping system, ranging from X2=0 to X2 Nevertheless, the results relative to the probability of detection follow the same pattern as those for the multiple trap sensitivity. The reported probabilities vary across trap densities from 3.23 x 10-6 to 2.24 x 10-4. The highest probabilities of dete ction are found for the pest in festation occurring during the summer months, while the chances of detecti ng a pest infestation during the month of February are extremely low. Furthermore, all marginal values of probability of detection are positive, ranging from 4.77 x 10-7 to 2.11 x 10-4. Like for the marginal trap sensitivity, the general tendency is for the ma rginal values of the probability of detection to increase with trap density and pest population size. Ne vertheless, at some trap densities, i.e. at a density of 12 traps per square mile, the marginal values of probability of detection show a slight decrease with an increase in trap density. Optimal Trap Densities The optimal solutions regarding trapping de nsity are presented in Table 4-13. The lowest optimal trap density for ML traps is 82 traps per ha and is reported for a June infestation occurring in Tampa. A total of 184 traps per ha are to be placed in Miami during the month of February to achieve the optimal solution. The highest optimal trap density for ML traps (465 trap s per ha) is found for a Tamp a infestation starting in

PAGE 90

76 October. These results are s upportive of the hypothesis that the optimal trapping density varies across locations and seasons. The optimal solutions for ML traps greatly differ from those for TML traps that are found to be more effective for capture of male C. capitata Optimal trapping densities for TML traps range from 9 to 80 traps per ha, w ith the highest optimal trap density being reported for a February infesta tion occurring in Tampa. A total of 25 traps per ha are to be placed in Tampa during the month of Oct ober to achieve the optimal solution. The lowest trap density (9 traps per ha) is found fo r a June infestation in Tampa. These results highly suggest that emphasis should be placed more on improving trap sensitivity rather than on increasing trap density. Conclusions The objectives of this chapter were to (1) compute the multiple trap sensitivities for ML and TML traps and the probabilities of dete cting a Medfly infestat ion in Florida at its early stages and (2) determine the optimal tr apping density that can minimize the total expected cost of the Medfly prevention, detection and eradication program. Our study shows that the colonization pot ential greatly varies across se asons and locations and that none of the outbreak scenarios simulated in th e spatial-temporal model can be considered early-detected infestatio ns. The chances of detecting Medf ly populations at early stages are extremely low. Sensitivity of ML traps to a 50-day-old infestation varies across trap densities from 4.95 x 10-6 for a density of one trap per square mile in Miami to 5.84 x 104 for a density of 21 traps per square mile in Tampa. Because of significant progress made in developing more potent lures for male C. capitata TML traps are expected to be on average 7.76 times more sensitive than ML traps.

PAGE 91

77 The results relative to the probability of detection follow the same pattern as those for the multiple trap sensitivity. The reported pr obabilities vary across trap densities from 3.23 x 10-6 to 2.24 x 10-4. The highest probabilities of detection are found for the pest infestation occurring during th e summer months, while the chances of detecting a pest infestation during the month of February are extremely low. Optimal trapping densities also vary acro ss locations and seas ons, ranging from 82 to 465 traps per ha for ML traps and from 9 to 80 traps per ha for TML traps. These results strongly suggest that emphasis s hould be placed on improving the multiple trap sensitivity, which is, reportedly, increas ing with trap density and population size. Table 4-1 Distribution of da y degrees required by stage Day Degrees Required to Transition (oF) Transitional Phase Minimum Maximum Average Egg / Larvae 33.8 47.9 40.85 Larvae / Pupae 153.9 219.45 186.67 Pupae / Pre-adults 308.6 436.8 372.7 Pre-adults / Adults 596 608.3 603.1 Source: APHIS 2002 Table 4.2 Average monthly distances flown by different fractions of Medfly population Means of spread Fraction of population moving away from the epicenter Average distances flown (linear mile) 0.15 0.125 0.06 0.435 0.042 0.75 0.027 1.25 Random dispersal 0.018 1.75 0.0015 12 0.0006 24 0.00042 48 0.00030 96 Long distance flight & human conveyance 0.00018 108

PAGE 92

78 Figure 4-1. Bayesian decision process

PAGE 93

5-mi-wide buffer perimeter 868 201 46 80 27 65 37 56 29 43 18 25 45 Figure 4-2 Treatment and quarantine areas of an infestation scenar io in Miami (October) Table 4-3 Eradicati on cost equations Tampa Miami Coefficients Pooled data February June October Pooled data February June October 0 (se) 9.8013 (0.456) 10.32897 (0.0015) 7.072238 (0.0164) 8.774308 (0.0000) 9.6896 (0.440) 9.626233 (0.0108) 8.802522 (0.0002) 8.234525 (0.001) 1 (se)a 0.3409 (0.032) 0.917753 (0.0137) 1.52183 (0.0118) 0.702589 (0.0114) 0.5854 (0.029) 0.85166 (0.0187) 1.061598 (0.0052) 0.822641 (0.0193) 11 (se) 0.0634 (0.075) 0.011916 (0.0298) -0.04039 (0.0279) -0.04655 ((0.0345) 0.0241 (0.072) 0.046724 (0.0548) -0.02198 (0.012) -0.00538 (0.0377) Residual sums of square 196.983 3.77434 9.76437 0.58212 181.18 9.770366 1.40914 3.330146 F[6,204] 440.2952 390.5533 a se=standard error of the parameter Table 4-4 Hyperbolic tangent approxima tion of marginal probability function for multiple trapping sensitivity of McPhail traps Coefficients a a0= -3.74367 ; a1= 0.25243; a2= 1.053351; a3 = 0.68775; a12= -0.04322; a13= -0.02082; a23= -0.07316; a11=0.0114269; a22=0.69595; a33=-0.01823 Standard error of the estimate Residual sum of squares = 0.00024037 Standard error = 0.0000160247 Kolmogorov-Smirnov test b D value = 0.08 P-value = 1.00 a These coefficients refer to the conditional probability density function described in Equations 4.21 & 4.22, which is approximated by a quadratic function. The size of the adult population, X, and the number of traps, Z, are the two major variables of this function. bKolmogorov Smirnov test determines if two datasets differ significantly by using the maximum vertical deviation between their curves. The maximum difference in cumulative fraction is D=.45. With a D value estimated at 0.08 for a p-value =1.00, the KS test supports the hypothesis that the values of the probabilities in the dataset and the theoretical values of the probability density functions follow the same distribution.

PAGE 94

Table 4-5 Coefficients of comparative se nsitivity and mean da ily captures by period Pooled Data February June October Trap counts 65 20 23 22 Untransformed means of daily captures TML trap ML trap 43.4 5.9 84.4 11.6 21.5 3.5 29.0 3.3 Coefficient of comparative sensitivity of TML to ML traps (in log space) confidence intervals standard error t value Residual sum of squares 1.928979 1.62 2.23 0.152154 12.67778 36.17546 1.75722 1.01 2.77 0.354838 3.451957 4.92351 2.034088 1.20 2.86 0.400214 2.936439 10.55651 2.286278 1.48 3.09 0.387087 3.211114 10.6455 Calculated F[2,64] value 12.30973 Table 4-6 Distribution of the expected popul ation size and generation time per location and per season at 50, 77, 98, and 119 days a of the infestation Miami Tampa Expected b population size Generation c time Expected population size Generation time Month Day (Medflies) (Days) (Medflies) (Days) 50 860 44 200 64 77 2,211 33 1,420 33 98 20,570 27 1,584 25 Feb. 119 38,781 23 22,990 22 50 4,741 24 5,588 23 77 75,044 21 91,948 21 98 363,390 21 532,267 21 June 119 1,448,303 21 2,206,223 21 50 1,960 26 1,300 29 77 14,300 29 2,200 52 98 25,531 41 14,179 91 Oct 119 138,243 52 14,300 39 a Medfly populations at 50, 77, 98, and 119 days roughly correspond to the first (F1), second (F2), third (F3), and fourth (F4) generations, respectively. b Population size is expressed in terms of numbers of female pre-ad ults or adults and its calculation takes into account the survival probab ilities at all stages. c Generation time refers to the time required for one generation of eggs to reach the ovipositional age.

PAGE 95

Table 4-7 Distribution of the intrinsic rate s of increase and doubli ng times of the pest population per location and per season Miami Tampa Month Day a Intrinsic rate of increase Doubling time (days) Intrinsic rate of increase Doubling time (days) 50 0.0812 8.53 0.0587 11.81 77 0.0812 8.53 0.0587 11.81 98 0.1005 6.89 0.0813 8.53 Feb. 119 0.0805 8.61 0.0989 7.01 50 0.1484 4.67 0.1490 4.67 77 0.1249 5.55 0.1263 5.49 98 0.1108 6.26 0.1144 6.06 June 119 0.1008 6.87 0.1042 6.65 50 0.1240 5.59 0.1222 5.67 77 0.0974 7.12 0.0644 10.76 98 0.0769 6.36 0.0718 9.65 Oct 119 0.0777 7.96 0.0551 12.57 a Medfly populations at 50, 77, 98, and 119 days roughly correspond to the first (F1), second (F2), third (F3), and fourth (F4) generations, respectively. Table 4-8 Distribution of the infested area, quarantine area, and eradication cost per location and per season at 50, 77, 98, and 119 days a of the infestation Miami Tampa Infested area Quarantine area Eradication cost Infested area Quarantine area Eradication cost b Month Day (square mile) ($1,000) (square mile) ($1,000) 50 4 182 2,327 2 121 2,066 77 21 327 10,325 11 208 4,345 98 81 860 146,824 39 504 27,991 Feb 119 299 1,943 1,960,186 140 936 373,674 50 32 421 26,265 32 421 24,825 77 115 1,055 289,511 115 1,055 315,602 98 250 4,276 1,574,211 445 5,714 5,355,057 June 119 365 6,764 2,907,668 554 8,228 7,702,316 50 12 252 6,091 6 192 4,913 77 24 360 22,453 13 240 7,281 98 48 740 61,384 13 240 7,281 Oct 119 49 758 99,271 15 263 15,882 a Medfly populations at 50, 77, 98, and 119 days roughly correspond to first (F1), second (F2), third (F3), and fourth (F 4) generations, respectively. b The costs for pesticide applications are part of the total eradication cost and are estimated under the assumption of 90percent pesticide efficacy

PAGE 96

82Table 4-9 Distribution of multiple trap sensitivities for ML traps under all outbreak scenarios for different trap densities 50 77 98 119 50 77 98 119 Trap density (per mi 2) Tampa February Miami February 1 0.000000584 0.000000877 0.00000184 0.00000871 0.000001 0.00000192 0.00000949 0.000012 2 0.000002 0.00000294 0.00000594 0.0000262 0.00000334 0.00000618 0.0000285 0.0000355 5 0.00000966 0.0000138 0.0000265 0.000107 0.0000155 0.0000275 0.000115 0.000142 10 0.0000305 0.0000425 0.0000787 0.000296 0.0000475 0.0000814 0.000318 0.000388 12 0.000041 0.0000568 0.000104 0.000384 0.0000634 0.000108 0.000413 0.000503 16 0.0000652 0.0000894 0.000161 0.000578 0.0000995 0.000167 0.000621 0.000753 21 0.0001 0.000137 0.000243 0.000846 0.000152 0.000251 0.000907 0.001097 Tampa June Miami June 1 0.00000553 0.0000287 0.0000732 0.000188 0.00000495 0.0000265 0.0000599 0.000135 2 0.000017 0.0000823 0.000203 0.000506 0.0000153 0.0000763 0.000168 0.000367 5 0.0000707 0.000315 0.000742 0.001771 0.000064 0.000293 0.000618 0.001305 10 0.0002 0.000832 0.001897 0.004382 0.000182 0.000776 0.00159 0.003265 12 0.000261 0.001068 0.002414 0.005526 0.000238 0.000998 0.002027 0.004131 16 0.000396 0.001577 0.003512 0.00793 0.000361 0.001474 0.002959 0.005956 21 0.000584 0.002266 0.004977 0.011088 0.000534 0.00212 0.004205 0.008368 Tampa October Miami October 1 0.00000332 0.00000401 0.0000132 0.0000129 0.00000365 0.0000104 0.0000117 0.0000421 2 0.0000104 0.0000125 0.000053 0.000039 0.0000383 0.0000114 0.0000311 0.000119 5 0.0000448 0.000053 0.000155 0.000153 0.0000487 0.000125 0.000139 0.000447 10 0.000129 0.000152 0.000423 0.000416 0.00014 0.000345 0.00038 0.001166 12 0.00017 0.000199 0.000547 0.000539 0.000184 0.000447 0.000492 0.001491 16 0.000261 0.000304 0.000817 0.000805 0.000281 0.000671 0.000737 0.002188 21 0.0003888 0.000451 0.001188 0.001171 0.000418 0.000979 0.001074 0.003126

PAGE 97

83Table 4.10 Marginal trap sensitivities for ML traps under different outbreak scenarios 50 77 98 119 50 77 98 119 Trap density (per mi 2) Tampa February Miami February 2 0.000000708 0.000001032 0.00000205 0.00000874 0.00000117 0.00000213 0.00000951 0.00001175 5 0.00000255 0.00000362 0.00000685 0.0000269 0.00000405 0.00000711 0.0000288 0.0000355 10 0.000004168 0.00000574 0.0000104 0.0000378 0.0000064 0.00000108 0.0000406 0.0000492 12 0.00000525 0.00000715 0.00001265 0.000044 0.00000195 0.0000133 0.0000475 0.0000575 16 0.00000605 0.00000815 0.00001425 0.0000485 0.00000903 0.0000148 0.000052 0.0000625 21 0.00000696 0.00000952 0.0000164 0.0000536 0.0000105 0.0000168 0.0000572 0.0000688 Tampa June Miami June 2 0.000005735 0.0000268 0.0000649 0.000159 0.00000518 0.0000249 0.00005405 0.000116 5 0.0000179 0.0000775 0.0001797 0.000422 0.0000162 0.0000722 0.00015 0.0003127 10 0.00002586 0.0001034 0.000231 0.000522 0.0000236 0.0000966 0.0001944 0.000392 12 0.0000305 0.000118 0.0002585 0.000572 0.000028 0.000111 0.0002185 0.000433 16 0.00003375 0.000127 0.0002745 0.000060 0.0000308 0.000119 0.000233 0.000456 21 0.0000376 0.000378 0.000293 0.000632 0.0000346 0.000129 0.0002492 0.0004824 Tampa October Miami October 2 0.00000354 0.000004324 0.0000129 0.0000127 0.00000388 0.0000104 0.0000115 0.0000384 5 0.0000115 0.0000135 0.0000386 0.0000382 0.0000124 0.0000313 0.0000348 0.0001093 10 0.00001684 0.0000198 0.0000536 0.0000526 0.0000183 0.000044 0.0000482 0.0001438 12 0.0000205 0.0000235 0.000062 0.0000615 0.000022 0.000051 0.000056 0.0001625 16 0.00002275 0.00002625 0.0000675 0.0000665 0.0000243 0.000056 0.00006125 0.0001743 21 0.0000254 0.0000294 0.0000742 0.0000732 0.0000274 0.0000616 0.0000674 0.0001876

PAGE 98

84Table 4-11 Distribution of probabilities of detection for ML traps under all outbreak scenarios for different trap densities 50 77 98 119 50 77 98 119 Trap density (per mi 2) Tampa February Miami February 1 0.00000329 0.00000484 0.0000107 0.0000481 0.00000552 0.0000106 0.0000525 0.0000663 2 0.00000418 0.00000615 0.0000124 0.0000548 0.00000698 0.0000129 0.0000596 0.0000742 5 0.00000635 0.00000907 0.0000174 0.0000703 0.00001018 0.0000181 0.0000755 0.0000933 10 0.00000992 0.00001382 0.0000256 0.0000962 0.00001545 0.0000265 0.0001034 0.0001262 12 0.00001151 0.00001595 0.0000292 0.0001078 0.00001781 0.0000303 0.0001159 0.0001413 16 0.00001513 0.00002074 0.0000374 0.0001341 0.00002309 0.0000387 0.0001441 0.0001748 21 0.00002041 0.00002796 0.0000496 0.0001726 0.00003102 0.0000512 0.0001851 0.0002239 Tampa June Miami June 1 0.00003057 0.0001586 0.000405 0.001039 0.00002736 0.0001465 0.0003311 0.0007463 2 0.00003555 0.0001721 0.000425 0.001058 0.000032 0.0001596 0.0003513 0.0007676 5 0.00004645 0.0002069 0.000487 0.001163 0.00004204 0.0001925 0.0004060 0.0008574 10 0.00006505 0.0002706 0.000617 0.001425 0.00005919 0.0002524 0.0005172 0.001062 12 0.00007329 0.0002999 0.000678 0.001552 0.00006684 0.0002803 0.0005692 0.001160 16 0.00009190 0.0003659 0.000815 0.001840 0.00008378 0.0003421 0.0006867 0.001382 21 0.0001192 0.000462 0.001015 0.002263 0.0001089 0.0004327 0.0008581 0.001708 Tampa October Miami October 1 0.00008354 0.00002217 0.0000729 0.0000713 0.00002018 0.0000575 0.0000647 0.0002327 2 0.00002175 0.00002614 0.0000815 0.0000801 0.00002384 0.0000650 0.0000726 0.0002489 5 0.00002943 0.00003482 0.0001018 0.0001005 0.00003199 0.0000821 0.0000913 0.0002937 10 0.00004195 0.00004943 0.0001376 0.0001353 0.00004553 0.0001122 0.0001236 0.0003793 12 0.00004774 0.00005588 0.0001536 0.0001513 0.00005167 0.0001255 0.0001318 0.0004187 16 0.00006057 0.00007055 0.0001896 0.0001868 0.00006521 0.0001557 0.0001710 0.0005078 21 0.00007985 0.00009204 0.0002425 0.0002389 0.00008531 0.0001998 0.0002192 0.0006380

PAGE 99

85 Table 4-12 Marginal values of probab ility of detection for ML traps dens ities under different outbreak scenarios 50 77 98 119 50 77 98 119 Trap density (per mi 2) Tampa February Miami February 2 0.000000477 0.00000065 0.00000113 0.00000332 0.000000729 0.000001156 0.000003572 0.00000395 5 0.000000721 0.000001458 0.000002494 0.00000775 0.00000107 0.000002571 0.000007974 0.00000952 10 0.000000714 0.000002378 0.00000409 0.00000129 0.000000105 0.000000420 0.00001394 0.00001645 12 0.000000797 0.000001064 0.000001805 0.00000578 0.00000118 0.00000192 0.000006276 0.00000753 16 0.000000904 0.00000239 0.000004079 0.00001314 0.00000132 0.000004213 0.00001406 0.00001675 21 0.000001055 0.00000361 0.000006114 0.00001926 0.00000159 0.00000623 0.00002049 0.0000245 Tampa June Miami June 2 0.00000249 0.00000673 0.00000995 0.00000949 0.00000232 0.00000654 0.00001011 0.0000106 5 0.00000363 0.00001741 0.00003146 0.0000526 0.00000335 0.0000164 0.00002733 0.0000449 10 0.00000372 0.00000318 0.0000648 0.000131 0.00000343 0.0000299 0.0000555 0.0001023 12 0.00000412 0.00001465 0.0000304 0.0000633 0.00000382 0.00000139 0.00002604 0.0000490 16 0.00000465 0.00003302 0.0000685 0.000144 0.00000424 0.0000309 0.0000587 0.000111 21 0.00000545 0.0000482 0.0001003 0.0002112 0.00000504 0.0000452 0.0000857 0.0001628 Tampa October Miami October 2 0.00000170 0.00000198 0.00000429 0.00000439 0.00000183 0.00000377 0.00000394 0.00000807 5 0.00000256 0.00000434 0.00001013 0.00001021 0.00000272 0.00000569 0.00000725 0.00001493 10 0.000002505 0.00000731 0.0000178 0.00000173 0.00000271 0.00000602 0.00000645 0.00001711 12 0.0000289 0.00000322 0.00000801 0.00000803 0.00000307 0.00000666 0.00000728 0.00001974 16 0.00000321 0.00000733 0.00001800 0.00001772 0.00000338 0.00000754 0.00000821 0.00002226 21 0.00000372 0.00001074 0.0000264 0.00002608 0.0000402 0.00000881 0.00000963 0.00002604

PAGE 100

86 Table 4-13 Optimal trapping density pe r type of trap, location and month Optimal trap density (# traps per ha) Miami Tampa Month ML trap TML trap ML trap TML trap February 322 55 465 80 June 92 12 82 9 October 184 19 248 25

PAGE 101

87 CHAPTER 5 WELFARE ANALYSIS OF A MEDFLY OUTBREAK IN FLORIDA Fundamentals of the Partial Equilibrium Model Theoretically, partial equilibrium models deal with a competitive economy where consumers and producers are considered pri ce takers (Mass-Colell 2000). Adam Smiths invisible hand acts to bring th e market to the point where the two curves cross, i.e., supply equals demand. The key feature of thes e models is that they do not include all production and consumption accounts in an eco nomy, nor do they attempt to capture all changes in the global economy. It is assume d that consumers expenditures on the goods included in the model account for a small portion of their total expenditures. Three conditions are essential for a competitive equilibrium, corresponding to the requirements that producers optimize, consumer s optimize, and that markets clear at the equilibrium prices. The equilibrium w ill then consist of a production plan yj for each firm, a consumption vector xi for each consumer, and a price vector .*p Below are the mathematical expressions of the requirements: 3. Profit maximization Given the equilibrium price p*, firm js equilibrium output q* must maximize ) max( ........q c p qj j j q (5-1) Y qj jto subject ..... .. (5-2) where Yj is the technology set, cj (qj ) is a cost of produci ng q units, with c(q) > 0 and c(q) >0. Because each competitive fi rm takes prices as given, the first order condition

PAGE 102

88 0 ) ( q c p (5-3) defines the supply function and the invers e supply function can be written as ) (1q pQ (5-4) 4. Utility maximization. Given the equilibrium price p*, consumer is equilibrium vector (xi) must maximize ) ( .....max X Ui i x (5-5) m p xi i ito subject ..... .. (5-6) where m is the budget share. Th e first order condition is ) ( p Uxi (5-7) The aggregate demand for consumption good under consideration is the sum of the corresponding individual de mands across consumers: ) ( p x i ipx) ( (5-8) The inverse demand functi on can be expressed as: ) ( ) (1x x p X (5-9) 5. Market clearing. Figure 5-1 represents the price vector such that aggregate demand for x equals aggregate supply. The equilibrium solutions p*, qj*, and xj* must solve the system of equations: j all for p pq cj j.. .. )).. ( ( *' (5-10) i all for p px Ui i.. ... )... ( ( *' (5-11) ij j ip pq x*) ( *) ( (5-12) 6. Welfare analysis. Economic welfare an alysis specifically in volves assessing how the equilibrium outcome (p*, q*, and x*) of a competitive market adjusts to changes in the environment. Consumer surplus is, by definition, the amount of extra utility consumers enjoy at price p and is represented by the area below the demand curve and above the price equilibrium. Likewise, producer surplus is the extra revenue received by producers at p and is represented by the area below the price equilibrium and above the supply curve. The effects on consumers are measured in terms of changes in the consum er surplus (CS), which is expressed as x x U CSi i i ip ) ( (5-13) A differential change in CS is given by

PAGE 103

89 dx p x p p dCS ) ) ( ( ) ( (5-14) since consumers face new effective price ) ( p and U(x) = p(x) for all goods. Likewise, the effects on producers are captured as follows: ) (q c x PSj j j jp (5-15) dq q c p p dPS ) ( ) ( ) ( (5-16) where PS stands for the producer surplus. The net social welfare (or aggregate social surplus) is the summation of the change in aggregate consumer surplus and the change in aggregate producer surplus. Let S (x, q) be the net social welfare formally defined as follows: ) ( ) ( p dPS p dCS dS (5-17) dx q c x p dS )) ( ) ( ( (5-18) Integrating the equation 5-18 from 0 to x yiel ds to the total value of the net social welfare: xds s c s p x S0)) ( ) ( ( ) ( (5-19) 7. Impact of a Medfly infestation. In th e event of a medfly introduction and/or colonization into Florida, the changes in net social welfare are likely to be associated with a decline in export mark ets. This would follow from the imposition of phytosanitary restrictions on traded goods by importing countries (Figure 5-2). These phytosanitary regulations would increas e the cost of exporting, thus shifting the excess supply schedule to the left (from ES to ES). Prices would rise from P1 to P2, leading to a decreasing demand of traded goods in importing countries. Aggregate welfare is like ly to decline in importing countries. The underlying assumption is that there are no demand-s timulus effects from the phytosanitary restrictions. Effects of a medfly infestation on net social welfare will vary across commodities, depending upon the Floridas in itial market share for each commodity and the geographic and temporal dynamics of the pest population. Define ), (*mxi ), (*mqjand ), (*mp respectively, to be the consumption, produc tion, and price paid by consumers during the quarantine period. Letting ) 0 (*Sand) (*mS, be the levels of net social welfare before and during the quarantine period. The change in S is given by:

PAGE 104

90 ds s c s p x m xm x xS S ) *( ) 0 *( *)) ( ) ( ( )) 0 ( ( )) ( ( (5-20) Adaptation of the Spatial Equilibrium Mo del to the Fruit and Vegetable Industry Grapefruit Model The grapefruit model used was originally developed at the University of Florida in 1991 (Pana) and later modified in 1995 (Spreen et al. 1995). Basically, the model consists of supply and demand components combined in an optimization problem to generate market-clearing prices for a given period. The supply component is represented by an implicit supply function for Florida, which is the unique supply region included in the model. In fact, Florida is th e worlds dominant grapefruit supplier, producing nearly 47% of total production for the 2000-2001season. Two commodities, red seedless and white seedless grapefruit are produced and total production for each commodity is computed, based on the inventory of bearing and non-bear ing trees, the number of trees by age, and a vector of yields. The model allows fo r an endogenous allocation between fresh and processed utilization. The demand side of the model includes c onsumption in the United States, European Union, Canada, and Japan. White seedless and red seedless grapefruit varieties are treated separately in the fresh market. A single market for grapefruit juice is also considered. In sum, the model consists of nine inverse de mand equations, eight for the fresh market and one for the processed market. Simple linear own-price demand equations were estimated for each market (Pana 1991). Quality standa rds greatly vary across fresh markets and such differences are reflected through varia tions in packout rates and market prices. The methodology and mathematical represen tation of the spatia l equilibrium model are outlined in Appendix B. The model uses the profit-maximizing problem to find the

PAGE 105

91 optimal shipments for each commodity (red s eedless versus white seedless) in each market and the equilibrium prices as well. Then, harvesting and marketing costs are deducted from these equilibrium prices to estim ate the fresh and processed on-tree prices. Vegetable Model We hereby used a monthly partial equilibri um model initially developed by Spreen et al. (1995) and, later exte nded by VanSickle et al. (2000). Vegetable crops are allocated in the model across four demand regions of the United States, including the Northeast, Southeast, Midwest, and West. These regi ons are represented by the New York City, Atlanta, Chicago, and Los Angeles wholesale markets, respectively. Transportation costs were included for delivering shipments to e ach of the regional markets based on mileages determined by the Automap software. Average transportation cost of a fully loaded refrigerated truck carrying is estimated at $1.3072 per mile (VanSickle et al., 1994). The model includes the following vegetable crops: tomatoes, peppers, cucumbers, squash, eggplants, watermelons, and strawber ries. All of these crops are listed as preferred or marginal hosts for Ceratitis capitata (Liquido et al. 1991) and, therefore, are regarded by APHIS as regulated articles (Federal Register 2003a). Cropping systems used in each producing area differ in harvesti ng date, crop planted and crop association. Double cropping involves the use of the same unit of land for a sec ondary crop after the primary crop is harvested. Such a practice allo ws growers to save inpu ts used in the first crop. Pre-harvest and post-harvest producti on costs were estimated for each production system and area by Smith and Taylor (2003). Monthly wholesale prices and unloads da ta collected by U.S. Department of Agriculture, Agricultural Marketing Service, Fruit and Vegetable Division, and Market News Branch were used to estimate the parameters of the market demand equations,

PAGE 106

92 using the inverse Rotterdam system. Intercepts of the demand equations were adjusted to reflect aggregate demand (Spr een et al. 1995). The model assumes that the commodities produced in the different produc tion points are perfect substi tutes. Slopes of the demand functions are also assumed to be constant over all quantities. The supply side of the model incorporates an implicit supply function under the assumption of a fixed proportion technology. Th e upward sloping supply curve results, in part, from the increasing transportation co sts as the demand region grows out from a particular supply region (McCarl and Spr een 1980; Peters and Spreen 1989). The methodology and mathematical representation of the model are outlined in Appendix C. Optimal dual solution to this spatial equilibri um model provides market prices in each demand area by month and commodity. Specialty Model The specialty model is developed to analy ze the allocation of the specialty citrus crops (kearly, temples, early tangerines, honey tangerines, and tangelos) in the U. S. market. The supply side of the model is domestically represented by the production of specialty crops from Florida, which is by far the predominant supplier in the U.S. market. Florida production accounts for over 70% of th e U.S. production and is determined in the model through multiplying for each specialty crop bearing acreage by average yield per acre. Domestic production is supplemented by imports mainly from Mexico and Spain. The model assumes that Florida fresh early ta ngerines compete with imports of Mexican tangerines and Spanish clementines in the U. S. market. The imposition of a fixed per-unit tariff is included in the model. It is also a ssumed that they are perfect substitutes and the U.S. consumption of these varieties is driven by the same demand curve.

PAGE 107

93 Simple linear own-price own demand e quations are calculated for the fresh utilization of each specialty crop in the U. S. market. Price and quantity effects obtained from the model are assumed to be the average of effects across different market locations within the U.S. market. The processed market is a residual market for the fruit that are damaged or too small to be sold in the fr esh market (USDA 2003). A fixed-price model is included for the specialty juice used by th e Florida juice processing industry to blend with orange or grapefruit juice for colo ring and sweetening. The model allows an endogenous allocation to the fr esh and processed market. The specialty model can be characterized as a spatial equilibrium problem with an implicit supply. Supplies and demand of specialty crops are equated in an optimization problem to generate market-clearing prices for the 2000-2001season. The demand side of the model is delineated by defining: Z b a P v v v v (5-21) as the inverse demand for commodity v in the U.S. market. Z v is the quantity of commodity v consumed in the U. S. market. If the subscript v refers to early tangerines, Z v includes both domestic production and im ports from Spain and Mexico. The parameters a v and b v account for the intercept and the slope of the inverse demand equation, respectively, and are both as sumed to be non-negative. Consider: 1. Q v Thousand cartons of commod ity v sent to packing house 2. v Packout rates for commodity v 3. E v Quantity of commodity v eliminated in fresh market and sent to processing market: Q Ev v v) 1 ( 4. PR Processing costs in dollars per sse (Single Strength Equivalent) gallons of juice 5. PJ Fixed price of juice pr oduced for blending purposes 6. JU v Quantity of juice of commodity v sold for blending purposes

PAGE 108

94 7. UJ v Juice yield for commodity v in sse gallons per box 8. TP v Total production of commodity v from Florida 9. PK Packing costs incurred in Florida 10. MXEAT Imports of Mexican tangerines 11. CLEM Imports of clementines from Spain 12. TARS Per-unit tariff imposed on clementines imported from Spain 13. TARM Per-unit tariff imposed on Mexican tangerines With these definitions, the quadratic programming model can be written as: ) ( ) ( ) ( .....2 2 1TARS CLEM TARM MXEAT PR PJ MaxQ PK JU Z b Z av v v v v v v v (5-22) to subject .. TP Qv v 5 0 (5-23) CLEM MXEATQ Zv v v (5-24) Q UJ JUv v v v v) 1 ( (5-25) v all forJU Z Qv v v.. ... .... 0 .... , (5-26) The optimal solution to this quadratic programming model provides the equilibrium FOB prices for each commodity in the U.S. ma rket and the optimal allocation between the fresh and processed markets. Cost Impact of a Medfly Quaran tine Restriction on Florida The baseline models described above were modified under outbreak scenarios with three-month, six-month, and one-year quara ntine periods. A characterization of these scenarios is presented in Table 5-1. Scenar io I assumes a three-month quarantine period with a 11% reduction in yields over the ar ea affected. About 30% of total fruit and vegetable production area in Florida is assume d to be marked out within the core area infested. A higher yield reducti on (30%) is associated with a six-month quarantine period (Scenario II) and the size of the production area affected is assumed to be 50%. The

PAGE 109

95 worst case scenario (Scenario III) is associ ated with a situation where APHIS has little control over the eradication process. The quarantine period spreads over one year: the entire fruit and vegetable pr oduction area is affected with a 50% yield reduction. Given the seriousness of the Mediterranean fruit fly, the alternative models assume that all shipments of fresh fr uits and vegetables from Fl orida would be subject to a stringent and complex certification process si milar to the Caribbean Fruit Fly protocol. Quarantine regulations imposed in the Medfly model are guided by the principles of the systems approach, comprising several co mplementary measures: intense trapping servicing, field treatment, and postharvest treatments. Trap se rvicing costs vary with trap density from $10.8 for the low-prevalence cas e (three-month quarantine period) to $72 for the high prevalence case (one-year quaran tine period). It is the same for the costs associated with field treatment for the contro l of the infestation. It is assumed that malathion will be applied once a week during th e time the fruit is susceptible to attack. The number of malathion applications is assumed to double in an Medfly endemic situation in Florida. The Medfly models only incorporate the re current costs that are, by definition, incurred annually in the posth arvest processes. Compliance tr eatments used in the models vary from crop to crop, depending on govern ment regulations, the level of quarantine security offered by the treatment, and its e fficacy. For those commodities with more than one treatment available, the model uses the l east costly treatment. This accounts generally for the treatment that causes the least dama ge to the commodity. For instance, methyl bromide fumigation schedules are adopted as quarantine treatments for strawberries, cucumbers and watermelons infested with Ceratitis capitata Citrus fruits are cold treated

PAGE 110

96 with treatment damage varying from 5% to 15% across commodities (grapefruit, tangerines, tangelo, and oranges). The Medfly models assumed that vapor heat treatment would be used for the certification of peppers, eggplants, a nd squash infested with Ceratitis capitata While this treatment has not proven economical in the past (APHIS 1993), our study offered an opportunity of assessing its economic feasibility as a quarantine treatment. Nevertheless, it is assumed that tomato growers would pr efer paying the cost of removing and dumping ripe tomatoes from the field rather than th e cost of treating them by vapor heat. Tomato fruits would be harvested earlier than usual (i.e., with less co lor) at the cost of a reduced yield so as to reduce the portion of ripe tomatoes. Empirical Results The baseline model for each category of crops was solved using GAMS programming software. The first simulati on was made for the 2000-01 crop year under the assumption that Florida is a Medfly-free area. Then, the baseline model was adjusted to reflect decline in yields, increase in preharvest, and posth arvest production costs associated with a Medfly outbreak/infestation in Florida. Simulation of Medfly models was made under scenarios of three-month, sixmonth, and one-year quarantine periods. In particular, the Medfly grapefruit model was also solved under two different options. The first option assumed the best alternative, that is, Florida growers would manage to negotiate with Japan and agree on a Medfly fl y-free zone certificati on protocol. Thus, the Medfly grapefruit model include s the market of Japan. The second option (the Medfly model without the market of Japan) was then solved where domestic supplies were increased to reflect the decrea se in export markets. The resu lts of the baseline and Medfly models are outlined in the following sections.

PAGE 111

97 Solutions of the Grapefruit Model under a Medfly Quarantine The optimal solution to the baseline m odel provides the equilibrium fresh and processed utilization for each variety (red seed less and white seedless), the equilibrium consumption of fresh grapefruit in both domes tic and export markets, and the equilibrium market prices as well. The baseline soluti on performed reasonably we ll in replicating the observed pattern of prices and consump tion for the 2000/2001 production season. Results of the baseline mode l (Table 5-2) show variations in the optimal allocation between fresh and processed ut ilization for the baseline mode l. While the utilization of red varieties is almost split into half between the fresh and processed markets, approximately 80% of the wh ite grapefruit produ ction goes to the processed market. Roughly 80% and 51% of white seedless a nd red seedless grap efruit shipments, respectively, are exported (Table 5-3 and 54). Exports of white s eedless grapefruit to Japan are extremely crucial for Florida grow ers, accounting for ove r 70% of the total production. FOB fresh pr ices for exports to the Europ ean Union and Japan markets tend to be higher than domestic prices, underlying the lower pack-out ra tes for higher quality fresh fruit intended for the export markets. It is the same for FOB prices of fresh red seedless grapefruit, which are higher than FOB prices of white seedless grapefruit. Option I (with the market of Japan) We consider first the results of the Medfly model that includes the market of Japan, under the assumption that a phytosanitary prot ocol is reached between APHIS and Japan for the certification of shipments of fresh ci trus. Under the scenario of a three-month quarantine period, average fresh on-tree pri ces are expected to fall by approximately 2.90% and fresh utilization of red seedless a nd white seedless varieties is projected to increase by 47,000 and 17,000 boxes, respectively (Table 5-2). These results follow from

PAGE 112

98 a reduction in the volume of fruit abandoned (which compensates for the decline in yields). Processed utilization of red seedless gr apefruit is projected to decline by approximately 16,000 boxes, whereas projected processed utilization of white seedless grapefruit increases by 15,000 boxes. Total lo sses in overall on-tree revenues for white seedless and red seedless grapefruit are estimated at $4 million, with a loss of $3.1 million for red seedless grapefruit and $930, 000 for white seedless grapefruit. A portion of the loss in the overall on-tree revenue for white seedless grapefruit is offset by the gain in the processed on-tree revenue for this variet y. FOB revenues are projected to decline in both domestic and export markets, with total losses approximating $ 1.0 million and $2.7 million for white seedless and red seedless grapefruit, respectively (Tables 5-3 & 5-4). Fall in FOB prices results in an increase in the marketings of fresh red seedless and white seedless grapefruit in the United States, European Union, and Canada. The situation is different in the market of Japan where the incr ease in the FOB prices of fresh red seedless and white seedless grapefruit leads to a re duction in the level of consumption. Patterns of change differ under a six-mont h quarantine scenario. Average on-tree prices for the fresh market of red seedless a nd white seedless grapefruit are expected to increase approximately 8% and 68%, resp ectively. Fresh on-tree revenues for red seedless grapefruit also increase by approximately $3 million, offsetting the reduction in fresh utilization of this variety (Table 5-2) However, fresh on-tree revenues for white seedless grapefruit decline by approximatel y $10 million, as a result of a significant decline in fresh utilization of this variety.

PAGE 113

99 Average processed on-tree prices of red seedless and white seedless grapefruit are projected to increase approximately 137% and 130%. Processed utilization of both varieties will decline by approximately 97,000 boxes and 3 million boxes respectively. For both, the processed on-tree revenues are ex pected to increase, offsetting the decline in processed utilization. Overall on-tree revenues of red seedless and white seedless grapefruit are projected to increase by approximately $23 million and $3 million, respectively. World FOB revenues for red seedless and white seedless grapefruit in the sixmonth quarantine model are expected to e xperience a reduction of approximately $3 million and $33 million, respectively (Tables 5-3 & 5-4). Nevertheless, the impact on FOB revenues will differ across the different ma rkets. In the United States and European Union, the FOB revenues for red seedless and white seedless grapefruit will increase as a result of an increase in FOB prices, which w ill offset the reduction in the marketings of fresh grapefruit. In Japan, th e reduction in the marketings of fresh red seedless and white seedless grapefruit by 523,000 boxes and 3.8 mi llion boxes, respectively, will more than offset the potential gain from the rise in FOB prices. As a result, the FOB revenues for red seedless and white seedless grapefruit ar e expected to decline 7.5% and 54.8% respectively. In Canada, FOB revenue for white seedle ss grapefruit is expected to increase by $203,000, as a result of a 34-percent increase in FOB price. However, FOB revenues for red seedless grapefruit will experience a reduction by approximately $1 million: the increase in the marketings of red seedless gr apefruit will not offset the loss from the fall in FOB prices.

PAGE 114

100 Results of the one-year quarantine model (S cenario III) give a picture of a general collapse of the Florida grapefruit industry The marketings of fresh grapefruit are systematically blocked in the United States, European Union and Japan. Only the market of Canada will survive, with significant reduction in total fresh sales of red seedless and white seedless grapefruit by approximately 713,000 boxes and 104,000 boxes, respectively (Table 5-2). FOB pr ices of red seedless and white seedless grapefruit in this market will rise to $15.78 and $18.78 per cart on, respectively (Tables 5-3 & 5-4). World FOB revenues will be drastically reduced. The significant reduction in fresh utilizati on will more than offset the potential gain from the rise in average fresh on-tree prices Fresh on-tree revenues for red seedless and white seedless grapefruit are projected to d ecline 94% and 99%, respectively, compared to their levels in the baseline model. In the processed market, on-tree revenues for red seedless and white seedless grapefruit wi ll increase 197% and 307%, respectively. Overall on-tree revenue for red seedless gr apefruit will declin e by approximately $ 80,000, whereas total on-tree revenues for white seedless grapefruit will increase by approximately $ 10,000. Option II (without the market of Japan) We now turn to the analysis of the resu lts of the Medfly gr apefruit model without the market of Japan (Tables 5-5 to 5-7). Under a three-month quarantine scenario, the pattern of changes in fresh on-tree prices is the same as in the previous case (Medfly model with the market of Japan), but the magnit ude of the fall in ontree prices is bigger in the Medfly model without the market of Japan. An average 40percent decrease in fresh on-tree prices (Table 55) is reported in the Medfly model without the market of Japan, as opposed to a 3-percent decrease in the Medfly model with the market of Japan.

PAGE 115

101 This result reflects the loss of Japanese export markets a nd a glut on the domestic market. Unlike the previous case, fresh utilization of red and white grapef ruit is expected to decline, as a result of a large increase in th e volume of fruit abandone d. As a result, fresh on-tree revenues of red seedless and white seedless grapefruit decline 45% and 84%, respectively. Overall on-tree revenues for red seedless and white seedless grapefruit also decrease by approximately $ 23 million and $ 11 million, respectively, in spite of significant gains in proc essed on-tree revenues. The direction of the change in FOB prices and revenues for red seedless grapefruit is the same as in the previous case. FOB pr ices per carton in the United States, Canada, and European Union decrease to $ 7.85, $ 7.62, and $ 8.96, respectiv ely, leading to an increase in the marketings of red seedless grapefruit (Table 5-6). Losses in FOB revenues for this variety are estimated at $ 5.8 million, $ 2.1 million, and $ 2.8 million in the United States, Canada, and European Union, respectively. World revenues for white seedless grapefruit also decline 79%, from $ 80 million to $ 17 million (Table 5-7). However, FOB revenue for white seedless gr apefruit in each individual market is projected to increase, as a result of a signi ficant increase in FOB prices, which will more than offset the reduction in fresh utilization. In the six-month quarantine model (Scenar io II), FOB prices of white seedless grapefruit in the United States, Canada, and European Union increase to $ 9.24, $ 8.28, and $ 10.21, respectively, leading to an aver age 12-percent increase in FOB revenues in comparison with the baseline solution. FOB prices of the red seedless grapefruit also increase in comparison with their low levels in the three-month quarantine model, but

PAGE 116

102 still remain below the optimal solution in th e baseline model. As a result, FOB revenues for red grapefruit in the United States and Canada decline by approximately by $ 0.5 million and $ 1.9 million, respectively. European Union is the only market where FOB revenue for red seedless grapefruit is not expe cted to change in spite of the increase in FOB prices. Fresh on-tree prices per box of red and wh ite seedless grapefruit increase to $ 8.21 and $ 8.46, respectively (Table 5-5), in comp arison to the optimal solution in the threemonth quarantine model. This increase is not enough to offset the reduction in the marketings of fresh grapefruit. Fresh on-tree revenues for red seedless and white seedless grapefruit decline by approximately $ 38 million and $ 31 million, respectively. In the processed market, average on-tree prices of red seedless and white seedless grapefruit continue to increase by approximately 291% and 272%, respectively, leading to significant gains in the processed on-tree re venues. The overall on-tr ee revenues of red and white seedless grapefruit are expected to decline by $ 4.3 million and $ 7.1 million, respectively. Under a one-year quarantine scenario, resu lts of the Medfly model without the market of Japan are identical to those of th e Medfly model with the market of Japan. World FOB revenues for red seedless and white seedless grapefruit are expected to decline by approximately $ 223 million and $ 79 million, respectively (Tables 5-6 & 5-7). Only the marketings of fresh grapefruit in Canada will survive, with significant reductions in total fresh sales. Solutions of the Vegetable Model under a Medfly Quarantine The baseline solution to this model provides the equilibrium consumption of each commodity in every month in each demand re gion, the optimal level of shipments from

PAGE 117

103 each supply point, the optimal production of each cropping system by production area, and the quantity of each commodity produced in each supply by month. The baseline solution performed reasonably well in replicat ing the observed pattern of shipments and acres planted for the 2000/2001 season. Results of the baseline mode l (Table 5-8) show the di stribution of acreage planted by cropping system in each producing area. Ac reage planted in Florida is 104,863 acres, accounting for approximately 40% of tota l acreage planted in all producing areas included in the model. Approximately 35% of total vegetable production is produced in the State of Florida. Florida is the second ma jor producer of tomatoes after Mexico, with a production of 56.5 million cartons (Table 5-9). Florida is the only watermelon producing area in the model and is also the l eader in the production of peppers, squash, and eggplants. Results of the Medfly model are outlined in the following sections. The analysis is made on the assumption of an infestation starting in Palm Beach and Southwest (three-month quarantin e scenario), then spreading over the West Central and the Dade County (six-month quarantine scenar io), and leading to a Medfly endemic situation over the whole State of Florida. Tomatoes Under a scenario of a three-month qua rantine period in Palm Beach and in Southwest Florida, acreage planted to tomatoes in Florida is expected to increase as a result of a significant increase in the acr eage planted in Dade County by 20,452 acres (Table 5-8). Palm Beach C ounty will stop producing commercia l quantities of tomatoes and total acreage of tomatoes in all producing areas is also expected to decline. Tomato production in Florida will increase 20% (Table 5-9) and Florida growers will increase their shipping point revenues by approximately $ 100 million (Table 5-10) as a result of

PAGE 118

104 an increase in their market share. These results reflect the competitive advantage held by Florida in tomato production, which is due to higher yields per acre and lower marketing costs in all producing areas. Mexican growers will suffer the greatest loss in tomato shipping point revenues with a loss of $52. 4 million. The overall tomato production will increase by 0.1%, but average wholesale prices in demand markets will increase approximately 1%, from $8.64 to $8.72 per cart on (Table 5-12). Quantities of tomatoes consumed will only decline in the area of Los Angeles (Table 5-11). A six-month quarantine period is considered a serious blow to tomato production in Florida. Total acreage of tomatoes in Flor ida is expected to decline by approximately 22,000 acres (Table 5-8), compared to the tota l acreage in the baseline model. Most of Florida production will concentrate in West Central Florida. Dade County will stop producing commercial quantities of tomatoes and total acreage in the Southwest and Palm Beach will decline by approximately 20, 000 acres and 1,000 acres, respectively. As a result, total tomato producti on in Florida will decline 76 % (Table 5-9), and Florida growers will suffer a loss in their ship ping point revenues estimated at $293 million (Table 5-10). Alabama will stop producing commercial quantities of tomatoes and Virginia will suffer a loss of $5.7 million in their shipping point revenues. All other producing areas will increas e their tomato production a nd shipping point revenues. Increase in production will be small in Califor nia and South Carolina, estimated at 3.2% and 4.9%, respectively (Table 5-9). Mexico wi ll experience the greatest benefit with an expansion of its acreage of tomatoes and an increase in production by 37%. Mexican growers will gain market share and shipping poi nt revenues. Nevertheless, overall tomato production will decrease by 9.1%, leading to an increase in average wholesale prices by

PAGE 119

105 9.2% (Table 5-12). Quantities of tomatoes consumed will decrease in all demand markets. In a Medfly endemic situation (one-year quarantine period), the economic impacts on Florida tomato production will be much greater. Dade County, Palm Beach and Southwest Florida will stop producing commercia l tomatoes. West Central Florida will be the only producing point in Florida with 10,846 acres. Florida growers will lose 90% of their market share and suffer a loss of $377 million in their shipping point revenues. On the contrary, Mexico will increase further its production and market share. The total revenues that Mexican grower s receive for tomatoes are expected to increase by $225 million. Nevertheless, average wholesale pri ce is expected to increase by approximately 13% as a result of a significant decline in overall tomato production. The level of tomato consumption in demand markets will d ecrease by an average 10% (Table 5-11). Peppers Under a three-month quarantine scenario, acreage of peppers will increase by 43 acres in West Central Florida, while acreag e of peppers in Palm Beach will decline by 1971 acres (Table 5-8). Total pr oduction in Florida will decl ine 16.3% as a result of a reduction in total acreage planted to peppers by 1928 acres. Texas is also expected to decrease acres of peppers from 12,680 acres to 9,962 acres. Mexico will increase acreage by 1908 acres. Overall pepper acr eage and production are expect ed to decline 6.16% and 8.8%, respectively (Table 5-9), while the aver age wholesale price is projected to increase 5.93%. The impacts of a six-month quarantine scenar io are even bigger for Florida. West Central Florida will stop producing commerci al quantities of peppe r and Palm Beach will decrease acres of peppers from 7,715 acres to 1,756 acres. As a result, pepper production

PAGE 120

106 in Florida will decline 91.8% (Table 5-9) and growers will suffer a loss of $151 million in their shipping point revenues (Table 510). Mexico and Texas will increase their production of peppers by 28% and 54%, respec tively, and will gain significant market share. Like in the previous case, overall pepper production will decline by approximately 25%, leading to a 12-percent increase in aver age wholesale price. Quantities of peppers consumed in Atlanta, Los Angeles, Chicago, and New York will decrease 36.1%, 9.4%, 21.9%, and 41.7%, respectively (Table 5-11). Under a one-year quarantine scenario, co mmercial production of peppers will be completely blocked in Florida. Texas and Me xico will be the only producing points in the one-year quarantine model. This shift account s for a loss of $171million in shipping point revenues for Florida, while Texas and Mexico increase their shippi ng point revenues by $17.4 million and $71.2 million, respectively. Average wholesale price increases 12.8%, leading to an average 23-percent decline in the equilibrium consumption in demand markets. Cucumbers Cucumber production in Florida is expected to decline under all scenarios. The rate of decline varies with the temporal dynamics of the pest population, from 31.6% in the three-month quarantine model to 92.6% in the one-year quaran tine model. It is the same for the losses in Floridas shipping poin t revenues, approximating $5 million, $12 million, and $23 million in the three-month, si x-month, and one-year quarantine models, respectively. As cucumber pr oduction decreases, Mexico ga ins more market share and shipping point revenues. Sh ipping point revenues for Me xico increase by $120,000, $990,000 and $4.1 million under scenarios of a three-month, six-month, and one-year quarantine periods, respectively (Table 5-10) Total shipping point revenues in the

PAGE 121

107 Medfly model are expected to decline under all scenarios, as a result of a decline in the quantities of cucumbers consumed in all demand markets. Average wholesale prices increase 5.6% in the one-year quarantine model. Squash In the event of a three-month quarantin e period in Palm Beach and Southwest Florida, squash production will be completely blocked in Southwest Florida. Double cropping of squash increases in West Cent ral Florida and single cropping increases in Dade County. Total squash acreage in Florida is expected to increase (Table 5-8), but total production will decline 4% (Table 5-9). Loss of squash production in Florida is not offset with increased production in Mexico. Overall squash producti on in the three-month quarantine model declines 1.94%, leading to a 0.8-percent increase in average wholesale prices (Table 5-12). Equilibrium consumption in demand markets declines by an average 1.8-percent (Table 5-11). Florida growers su ffer a loss of $2.9 million in their shipping point revenues, while Mexico increases th eir shipping point revenues by 1.01 million. The six-month quarantine will be disastrous for Florida, where squash production will be completed blocked in all producing ar eas. Mexico is expected to increase squash production from 1518 units to 3335 units. Over all production in the six-month quarantine model declines 43.6%. The average wholes ale price increases 9.6%, leading to a reduction in consumption levels by 43.6% in demand markets. Mexico will gain market share and shipping point revenues, while Flor ida will lose $51 million in shipping point revenues. Mexico is expected to decrease their acreage from 15,959 acres in the one-year quarantine model to 10,886 acres in the one-year quarantine m odel. Overall production in the one-year quarantine model increases by 587 units, compared to the optimal solution

PAGE 122

108 in the six-month quarantine model. Average wholesale prices per carton decline from $15.17 to $14.90, leading to some improvement in the level of squash consumption in all demand markets. Eggplants The impacts are particularly disastrous for Florida. Commercial eggplant production will be completely eliminated in Florida. This accounts for a loss of $56 million for Florida growers. Mexico produc tion increases by approximately 130%, which is largely insufficient to offset the total loss of Florida production. Overall eggplant production will decline 28.6%, leading to an increase in average wholesale price by 7.86%. The level of eggplant consumption in demand markets will decline 71.4%. Watermelons Florida is the only watermelon producing ar ea in the model. Watermelon is grown as a second crop following pepper in West Ce ntral Florida and tomatoes in Southwest Florida. Acreages of watermelon are expected to decline under all scenar ios, as a result of the loss in profitability of the first crop. In the six-month and one-year quarantine models, production of watermelon as a second crop following pepper will be completely eliminated. Furthermore, Southwest Flor ida will stop producing commercial watermelon in the one-year quarantine model. This shift results in a decline in watermelon production, ranging from 12% in the three-m onth quarantine model to 58% in the oneyear quarantine model. Average wholesale pr ices are expected to increase 5.9%, 17.3%, and 33.5% in the three-month, six-month, a nd one-year quarantine models, respectively. Florida growers will lose shipping point reve nues as a result of significant reduction in the level of consumption in the demand market s. For instance, in the one-year quarantine

PAGE 123

109 model, quantities of watermelons consumed will decline by 57.9%, causing a loss of $22.3 million in shipping point revenues. Strawberries Total impacts are not rea lly significant for strawb erries under a three-month quarantine scenario. Acreages of strawberry pl anted in West Central Florida will decrease by 26 acres, while California will increas e production acreage by 14 acres. Overall strawberry production in the three-month qua rantine model will decline 0.01%, but the average wholesale price remains the same. Neith er will there be a change in the level of consumption in the demand markets. However, commercial strawberry produc tion in Florida will be completely eliminated in the six-month and one-year qua rantine models. California is expected to increase production by 20.6%, which is not enough to offset the loss of Florida production. Overall production in these models will decline 2.6%, causing an increase in the average wholesale price by 4.5%. The leve l of consumption in demand markets will decrease by 1.82%. California will increase th eir shipping point revenues by $81 million, but Florida will lose $94 million. Solutions of the Specialty Model under a Medfly Quarantine Over two thirds of total production of k-ear ly, temples, and tangelos are shipped to the processing plant, with the highest percen tage (95%) of proce ssed allocation reported for k-early (Table 5-13). On the contrary, fresh allocation is predominant for early and honey tangerines, accounting for 70% and 57% respectively. In the event of a three-month quarantine peri od, utilization of all specialty crops is expected to decline, resulting in an increase in prices. The decline in the fresh utilization will be highly significant for k-early, temples and tangelos, approximating 49%, 81%,

PAGE 124

110 and 97%, respectively. Average on-tree prices of these specialty crops will increase by 161%, 408%, and 407%, respectively. Neverthele ss, average fresh on-tree prices of early tangerines and honey tangerines are expected to decline 28% and 45% respectively. In particular, the decline in the av erage on-tree price of early tangerines is due in part to the competition with imports of early tangerines and clementines from Mexico and Spain, respectively. Fresh on-tree revenues for temples, tangelo s, early tangerines, and honey tangerines are projected to decline by approxima tely $9000, $363,000, $3.6 million, and $4.4 million, respectively. Regarding the k-early, th e increase in the average on-tree price will more than offset the decline in fresh utiliz ation, leading to an increase in fresh on-tree revenue by approximately $5.3 million. In the processed market, the pattern of ch ange is roughly similar for all specialty crops. Processed on-tree revenues in the thre e-month quarantine model will decline as a result of a reduction in proce ssed utilization. Average proce ssed on-tree prices of early tangerines and honey tangerines will remain unc hanged, but those of tangelos, temples, and k-early will increase by 11%, 5.9%, and 1.2%, respectively. Total on-tree revenues for all specialty crops are expected to decline by approximately $18.5 million. Revenues for k-early, tangelos, early ta ngerines, and temples are projected to decline by approximately $ 7.8 million, $5. 1 million, $4.5 million, and $1.8 million, respectively (Table 5-14). Price increase fo r honey tangerines will more than offset the decline in the marketings of this commodity. As a result, projected FOB revenues for honey tangerines will in crease by $353,000.

PAGE 125

111 The economic impacts of a six-month quarantine period are even bigger. Marketings of fresh temples a nd tangelos will completely cease, as total production of these specialty crops is expected to be se nt directly to the processing plant. Fresh utilization of k-early, early tangerines, and honey tangerines will decline 80%, 72%, and 15%, respectively (Table 5-13). Fresh ontree revenues will also decline by approximately $4 million, $4.5 million, and $619,000 respectively, as the decline in fresh utilization offset the increase in fresh ontree prices. Processed on-tree revenues are expected to decline as a result of the re duction in processed utilization. The projected overall on-tree revenue loss is approximately $67 million, with losses for k-early and early tangerine accounting for $57.8 milli on and $5.8 million, respectively. Adjusted FOB revenues for k-early and early tangerines are proj ected to decline by approximately $41 million and $15 million, respec tively, in spite of significant increase in average FOB prices (Table 5-14). Rega rding honey tangerine, the increase in prices will more than offset the decline in total fresh sales. As a result, FOB revenue for honey tangerine will increase by $1.25 million. The results of the one-year quarantine m odel (Scenario III) gi ve a picture of a general collapse of the specialty citrus industr y. Fresh marketings of all specialty crops but honey tangerines will be completely elim inated. The processed market will survive with significant reductions in total proce ssed utilization. Processed revenue losses are estimated at $168 million, with losses for k-early accounting for $162 million. Overall on-tree revenues for all specialty crops will decline by approximately $195 million. Aggregate Impacts A Medfly outbreak in Florida is expected to have significant aggregate impacts (Tables 5-15 to 5-17) on producer revenues and consumer surplu s. In the event of a three-

PAGE 126

112 month quarantine period in Flor ida, aggregate revenues in th e whole fruit and vegetable sector are expected to decline approximat ely $9 million or $39 million (Table 5-15), depending upon whether Florida growers could negotiate a certification protocol for the exports of fresh grapefruit to Japan. Neve rtheless, projected sh ipping point revenues increase $13.7 million for the Florida ve getable sector (Table 5-16). Vegetable production in all producing areas but Virginia/M aryland is projected to decrease, with production losses in Alabama/Tennessee and So uth Carolina accounting for 46.6% and 32.6%, respectively (Table 5-17). Shipping poi nt revenues for these vegetable producing areas will decline $9.5 million a nd $23.2 million respectively. These results follow from the fact that much of the potential revenue loss associated with yield reduction and prea nd postharvest production cost increase will be offset by the rise in average wholesale pri ces and FOB prices. Consumers surplus is expected to decline $237.6 million (in the Medf ly model with the market of Japan) or $343.4 million (in the Medfly model without th e market of Japan). Consumers surplus losses are due to both a decline in the quantities of products consumed and an increase in the prices paid for those products consumed. Impacts of a six-month quarantine period in Florida are more significant for both consumers and producers. Consumer surplus losses increase to $821.3 million (Option I) or $1.25 billion (Option II) as the general price levels for fresh commodities increase and the consumption levels decrease. Florida ship pers stand to lose $705 .7 million (Option I) or $742.2 million (Option II) in shipping point revenues, as a result of severe production losses and significant increases in prea nd postharvest production costs. However, California, Texas, and Mexico are expected to increase their ve getable production by

PAGE 127

113 13.3%, 28.1%, and 41.4%, respectively. In part icular, Mexico growers will increase their shipping point revenues by $ 294.3 million in th e vegetable sector. They will also gain additional revenues from incr easing their exports of tange rines to the U.S. market. An endemic Medfly situation (Scenario II I) is considered a coup de grace to the whole fresh fruit and vegetable industry in Fl orida. Vegetable production is expected to decline 91.4%. The grapefruit a nd specialty citrus industry is unlikely to survive without the fresh market. This scenario will result in a $ 1.03 billion decline in Florida shipping point revenues. Total consumer surplu s loss will amount to $ 1.75 billion. Alabama will also stop producing commercial quantities of vegetables and Virginia will suffer some production loss estimated at 8%. However, California, Texas and Mexico will further increase their production, market share, and shipping point revenues. The overall production level will decline, as total fruit a nd vegetable production loss in Florida from an endemic Medfly situation cannot be completely offset in the short term. C(*), q(*) Consumer surplus p(x) = p(q) Producer surplus P(*), x(*) x(p) = q(p) Figure 5.1 Price equilibrium, aggregate demand and supply

PAGE 128

114 P D1 S1 ED ES" S2" ESD2 S2 P2 P1 Importing countryInternational marketExporting country Figure 5-2 Effects of phy tosanitary regulations Table 5.1 Outbreak scenarios and cost im plications of a Medfly infestation on the Florida fruit and vegetable industry Outbreak Scenarios Scenario I Scenario II Scenario III Characterization Quarantine Period 3 months 6 months One year Production area Affected 30% 50% 100% Yield reduction 11% 30% 50% Quarantine Control ($ per 1,000 cartons) Trapping Servicing 10.8 36 72 Field Treatment 560 560 1120 Cold storage Treatments 1900 1900 1900 Transport ripe Tomatoes 4250 4250 4250 $ / per 1000 kg MB fumigation 27 27 27 Vapor heat 200 200 200

PAGE 129

115 Table 5-2 Baseline annual re turns of fresh and processed Florida grapefruit for the 200001 season and changes in the medfly m odel including the market of Japan Fresh Fruit Processed Fruit Grapefruit Variety Utilization On-tree Price On-tree Revenue UtilizationOn-tree Price On-tree Revenue Total Ontree Revenue (1,000 boxes) ($ per box) ($ 1,000) (1,000 boxes) ($ per box) ($ 1,000) ($ 1,000 Baseline Solution Red 11,886 10.27 122,069 14,891 0.98 14.593 136,662 White 3,851 10.02 38,587 14,325 1.09 15,614 54,201 Scenario I a Red 11,933 (0.4)d 9.97 (-2.92) 118,972 (-2.54) 14,875 (-0.10) 0.98 (0.00) 14,577 (-0.10) 133,549 (-2.27) White 3,868 (0.4) 9.73 (-2.90) 37,640 (-2.45) 14,340 (0.10) 1.09 (0.00) 15,631 (-0.10) 53,271 (-1.71) Scenario II b Red 11,297 (-4.95) 11.07 (7.82) 125,103 (2.48) 14,797 (-0.63) 2.32 (136.63) 34,314 (135.14) 159,417 (16.65) White 1,711 (-55.56) 16.78 (67.50) 28,717 (-25.57) 11,169 (-22.03) 2.52 (130.91) 28,112 (80.04) 56,829 (4.84) Scenario III c Red 715 (-93.98) 21.58 (110.11) 15,429 (-87.36) 6,211 (-58.29) 6.97 (611.40) 43,303 (196.73) 58,732 (-57.02) White 31 (-99.19) 27.55 (174.9) 854 (-97.78) 9,058 (-36.76) 7.00 (542.84) 63,469 (306.48) 64,323 (18.67) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-year quarantine period. d The numbers in parentheses refe r to percent changes in the parameters; they can be positive or negative.

PAGE 130

116Table 5-3. Baseline world FOB revenue for red grapefruit for the 2000-01 season and ch anges in the Medfly model including the market of Japan Baseline Solution Scenario I a Scenario II b Scenario III c FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue $ per carton 1,000 cartons $ 1,000 $ per carton 1,000 cartons $ 1,000 $ per carton 1,000 cartons $ 1,000 $ per carton 1,000 cartons $ 1,000 USA 9.22 11,784 108,601 9.03 (-1.98)d 11,958 (1.47) 107,981 (-0.57) 9.73 (5.6) 11,294 (-4.16) 109,891 (1.19) 0 0 0 Canada 9.22 2,142 19,741 8.81 (-4.4) 2,186 (2.05) 19,259 (-2.44) 8.44 (-8.5) 2,226 (3.92) 18,787 (-4.83) 15.78 (71.2) 1,429 (-33.3) 22,542 (14.2) European Union 10.42 6,050 63,605 10.28 (-1.43) 6,118 (1.12) 62,893 (-0.27) 10.97 (5.21) 5,800 (-4.13) 63,626 (0.88) 0 0 0 Japan 14.36 3,796 54,526 14.75 (2.7) 3,603 (-5.1) 53,114 (-2.53) 15.41 (7.3) 3,273 (-13.8) 50,436 (-7.50) 0 0 0 World FOB Revenue 245,933 243,277 (-1.1) 242,740 (-1.3) 22,542 (-90.8) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario II: one-year quarantine period. d The numbers in parentheses refer to percent changes in the pa rameters; they can be positive or negative.

PAGE 131

117Table 5-4 Baseline world FOB revenue for white grapefruit for the 2000-01 season and ch anges in the Medfly model including the market of Japan Baseline Solution Scenario I a Scenario II b Scenario III c FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue $ per carton 1,000 cartons $ 1,000 $ per carton 1,000 cartons $ 1,000 $ per carton 1,000 cartons $ 1,000 $ per carton 1,000 cartons $ 1,000 USA 6.71 1,816 12,189 6,22 (-7.4)d 1,881 (3.6) 11,700 (-4.01) 9.92 (47.8) 1,394 (-23.2) 13,828 (13.44) 0 0 0 Canada 6.71 166 1,114 6.05 (-9.9) 172 (3.6) 1,041 (-6.6) 8.96 (33.5) 147 (-11.4) 1,317 (18.22) 18.76 (179.47) 62 (-62.65) 1,163 (4.4) European Union 7.83 283 2,217 7.32 (-6.5) 292 (3.18) 2,137 (-3.6) 10.95 (39.8) 227 (-19.8) 2,486 (12.13) 0 0 0 Japan 11.90 5,437 64,619 11.96 (0.6) 5,390 (-0.9) 64,464 (-0.24) 17.64 (48.43) 1,654 (-69.6) 29,177 (-54.8) 0 0 0 World FOB Revenue 80,139 79,342 (-1.0) 46,808 (-41.6) 1,163 (-98.5) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario II: one-year quarantine period. d The numbers in parentheses refer to percent changes in the pa rameters; they can be positive or negative.

PAGE 132

118 Table 5-5 Baseline annual returns of fresh and processed Florida grapefruit for the 200001 season and changes in the Medfly m odel excluding the market of Japan Fresh Fruit Processed Fruit Utilization On-tree Price On-tree Revenue Utilization On-tree Price On-tree Revenue Total on-tree Revenue Grapefruit Variety (1,000 boxes) ($ per box) ($ 1,000) (1,000 boxes) ($ per box) ($ 1,000) ($ 1,000) Baseline solution Red 11,886 10.27 122,069 14,891 0.98 14,593 136,662 White 3,851 10.02 38,587 14,325 1.09 15,614 54,201 Scenario I a Red 11,064 (-6.9)d 6.02 (-41.38) 66,605 (-45.43) 13,304 (-10.70) 3.54 (261.22) 47,096 (222.73) 113,701 (16.80) White 1,041 (-72.96) 5.90 (-41.12) 6,142 (-84.1) 9,739 (-32.01) 3.78 (246.8) 36,813 (135.8) 42,955 (-20.74) Scenario II b Red 10,192 (-14,25) 8.21 (-20.1) 83,676 (-31.45) 12,706 (-14.70) 3.83 (290.8) 48,664 (233.47) 132,340 (-3.16) White 939 (-75.61) 8.46 (-15.60) 7,944 (-79.41) 9,670 (-32.50) 4.05 (271.60) 39,164 (150.82) 47,108 (-13.10) Scenario III c Red 715 (-93.98) 21.58 (110.11) 15,429 (-87.36) 6,211 (-58.29) 6.97 (611.40) 43,303 (196.73) 58,732 (-57.02) White 31 (-99.19) 27.55 (174.9) 854 (-97.78) 9,058 (-36.76) 7.00 (542.84) 63,469 (306.48) 64,323 (18.67) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-year quaranti ne period. d The numbers in parentheses refer to percent changes in the parameters; they can be positive or negative.

PAGE 133

119Table 5-6 Baseline world FOB revenue for red grapefruit for th e 2000-01 season and changes in the Medfly model excluding the market of Japan Baseline Solution Scenario I a Scenario II b Scenario III c FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) USA 9.22 11,784 108,601 7.85 (-14.9)d 13,087 (11.10) 102,733 (-5.4) 9.05 (-1.8) 11,942 (1.34) 108,075 (-0.48) Canada 9.22 2,142 19,741 7.62 (-17.35) 2,315 (8.10) 17,640 (-10.64) 7.75 (-15.95) 2,300 (7.37) 17,825 (-9.70) 15.78 (71.2) 1,429 (-33.3) 22,542 (14.2) European Union 10.42 6,050 63,065 8.96 (-14.01) 6,726 (11.17) 60,265 (-4.44) 10.27 (-1.43) 6,141 (-1.5) 63,068 (0.00) Japan 14.36 3,796 54,526 World FOB revenue 245,933 180,638 (-26.55) 188,968 (-23.2) 22,542 (-90.8) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario II: one-year quarantine period. d The numbers in parentheses refer to percent changes in the pa rameters; they can be positive or negative.

PAGE 134

120Table 5-7 Baseline world FOB revenue for white grapefruit for the 2000-01 season and ch anges in the Medfly model excluding the market of Japan Baseline Solution Scenario I a Scenario II b Scenario III c FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) USA 6.71 1,816 12,189 7.88 (17.43)d 1,662 (-8.48) 13,097 (7.45) 9.24 (37.7) 1,484 (-18.28) 13,712 (12.50) Canada 6.71 166 1,114 7.71 (14.9) 157 (-5.42) 1,210 (8.62) 8.28 (23.4) 153 (-7.8) 1,267 (13.73) 18.76 (179.47) 62 (-62.65) 1,163 (4.4) European Union 7.83 283 2,217 8.96 (14.4) 263 (-7.06) 2,356 (6.27) 10.21 (30.4) 240 (-15.2) 2,450 (10.5) Japan 11.90 5,437 64,619 World FOB Revenue 80,139 16,663 (-79.2) 17,429 (-78.25) 1,163 (-98.5) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario II: one-year quarantine period. d The numbers in parentheses refer to percent changes in the pa rameters; they can be positive or negative.

PAGE 135

121 Table 5-8 Planted acreage in the base line and Medfly models by crop and area Acreage (acres) Crop / Area Baseline Scenario I a Scenario II b Scenario III c Tomato Florida Dade 4,408 24,860 0 0 Palm Beach 2,798 0 1,653 0 West Central 11,077 6,211 14,230 10,846 Southwest 20,975 16,425 1,006 0 California 36,408 35,729 37,583 37,252 Alabama / Tennessee 3,448 1,842 0 0 South Carolina 6,923 4,665 7,266 6,960 Virginia / Maryland 6,282 7,758 5,449 5,781 Mexico Sinaloa 34,951 29,298 49,265 53,185 Baja 5,369 6,556 5,972 5,732 Total 132,639 133,344 122,424 119,756 Bell Peppers Florida Palm Beach 7,175 5,204 1,756 0 West Central 10,997 11,040 0 0 South west 0 0 360 0 Texas 12,680 9,962 16,249 16,445 Mexico Sinaloa 13,600 15,508 20,917 23,058 Total 44,452 41,714 39,282 39,503 Cucumbers Florida Palm Beach 6,693 5,204 3,409 0 West Central 0 0 0 705 Mexico Sinaloa 10,076 10,095 10,231 10,724 Total 16,769 15,299 13,640 11,429 Squash Florida Dade 8,081 9,369 0 5,408 Southwest 3,637 0 0 0 West Central 0 2,493 0 0 Mexico Sinaloa 7,265 7,650 15,959 10,886 Total 18,983 19,512 15,959 16,294 Eggplants Florida Palm Beach 5,327 0 0 0 Mexico Sinaloa 2,734 6,294 6,294 6,294 Total 8,060 6,294 6,924 6,294 Watermelons Florida West Central 1,812 1,630 14,230 10,140 Southwest 17,338 16,425 1,006 0 Total 19,150 18,055 15,236 10,140 Strawberries Florida West Central 4,545 4,519 0 0 California South 10,518 10,543 14,470 14,470 North 9,217 9,206 8,029 8,029 Total 24,280 24,268 22,499 22,499 a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-year quarantine period.

PAGE 136

122 Table 5-9 Baseline production for the 2000-01 season and percentage changes in production in the Medfly model by crop and area Production Crop / Area Baseline Scenario I a Scenario II b Scenario III c Tomato (Units) ( % changes ) Florida 56,506 20.0 (75.9)d (91.6) California 39,321 (1.9) 3.2 2.3 Virginia / Maryland 4,272 23.5 (13.3) (8.0) South Carolina 6,853 (32.6) 5.0 0.55 Alabama / Tennessee 2,138 (46.6) (100.0) (100.0) United States 109,090 7.7 (40.3) (49.8) Mexico 73,786 (11.1) 37.0 46.1 Total 182,876 0.1 (9.1) (10.5) Bell Peppers Florida 18,172 (16.3) (91.8) (100.0) Texas 7,735 (21.4) 28.1 29.7 United States 25,907 (17.9) (56.0) (61.3) Mexico 10,282 14.0 53.8 69.5 Total 36,189 (8.8) (24.8) (24.1) Cucumbers Florida 4,016 (31.6) (58.2) (92.6) Mexico 5,572 0.2 1.5 6.4 Total 9,588 (13.1) (23.5) (35.1) Squash Florida 4,395 (4.5) (100.0) (62.5) Mexico 1,518 5.3 119.7 49.8 Total 5,913 (1.9) (43.6) (33.7) Eggplants Florida 7,457 (100.0) (100.0) (100.0) Mexico 3,352 130.2 130.2 130.2 Total 10,809 (28.6) (28.6) (28.6) Watermelons Florida 6,475 (12.1) (31.7) (57.9) Strawberries Florida 12,725 (0.6) (100.0) (100.0) California 53,354 0.1 20.6 20.6 Total 66,069 (0.01) (2.6) (2.6) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-year quarantine period. d The numbers in parentheses refer to negative percent changes.

PAGE 137

123 Table 5-10 Baseline revenue for the 2000-01 season and changes in revenues in the Medfly model by crop and area Revenues ($) Crop / Area Baseline Scenario I a Scenario II b Scenario III c Tomato ( changes in revenues ) Florida 448,227,870 99,845,370 (292,344,865)d (376,920,890) California 286,004,000 (5,335,300) 9,228,300 6,628,900 Virginia / Maryland 42,736,960 10,039,420 (5,666,830) (3,412,580) South Carolina 71,270,170 (23,244,210) 3,528,690 385,930 Alabama / Tennessee 20,477,170 (9,540,260) (20,477,170) (20,477,170) United States 868,716,950 71,764,240 (305,732,655) (393,796,590) Mexico 496,178,430 (52,445,250) 181,606,640 225,859,660 Total 1,364,895,380 19,318,990 (124,126,015) (167,936,930) Bell Peppers Florida 171,026,490 (18,248,230) (150,679,366) (171,026,490) Texas 58,828,370 (12,608,780) 16,556,590 17,469,280 United States 229,854,860 (30,857,010) (134,122,776) (153,557,210) Mexico 102,449,600 14,372,700 55,118,400 71,243,500 Total 332,304,460 (16,484,310) (79,004,376) (82,313,710) Cucumbers Florida 26,579,760 (4,770,520) (11,132,500) (23,121,807) Mexico 64,228,410 120,240 987,010 4,133,530 Total 90,808,170 (4,650,280) (10,145,490) (18,988,277) Squash Florida 51,107,510 (2,797,163) (51,107,510) (29,234,740) Mexico 19,105,090 1,012,360 22,862,580 9,523,530 Total 70,212,600 (1,784,803) (28,244,930) (19,711,210) Eggplants Florida 56,568,750 (56,568,750) (56,568,750) (56,568,750) Mexico 25,895,360 33,727,530 33,727,530 33,727,530 Total 82,464,110 (22,841,220) (22,841,220) (22,841,220) Watermelons Florida 65,883,680 (3,321,801) (8,497,300) (22,345,670) Strawberries Florida 93,912,080 (431,710) (93,912,080) (93,912,710) California 466,867,600 450,500 80,500,700 80,500,700 Total 560,779,680 18,790 (13,411,380) (13,411,380) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-yea r quarantine period. d The numbers in parentheses refer to negative changes in revenues.

PAGE 138

124 Table 5-11 Baseline demand for the 2000-01 se ason and percentage changes in demand in the Medfly model by crop and market Demand Crop /.Demand Areas Baseline Scenario I a Scenario II b Scenario III c ($ / Unit) ( % changes ) Tomatoes Atlanta 41,080 0.6 (9.2) (9.7) Los Angeles 33,871 (0.8) (5.5) (5.8) Chicago 38,744 (0.14) (10.7) (11.8) New York 66,763 (0.06) (10.6) (13.2) Bell Peppers Atlanta 9,955 (7.8) (36.1) (20.5) Los Angeles 5,529 (2.4) (9.4) (7.9) Chicago 11,399 (7.1) (21.9) (21.3) New York 9,305 (15.8) (41.7) (41.1) Cucumbers Atlanta 2,077 (20.4) (36.3) (50.9) Los Angeles 1,988 (1.70) (7.0) (12.8) Chicago 2,943 (14.7) (22.3) (28.2) New York 2,580 (14.2) (27.4) (47.3) Squash Atlanta 1,808 (2.1) (61.3) (38.2) Los Angeles 1,690 (1.5) (1.4) (0.9) Chicago 1,686 (2.3) (48.0) (42.1) New York 728 (1.8) (93.3) (82.9) Eggplants Atlanta 1,673 (67.2) (67.2) (67.2) Los Angeles 2,991 (5.5) (5.5) (5.5) Chicago 3,983 (19.1) (19.1) (19.1) New York 2,162 (48.2) (48.2) (48.2) Watermelons Atlanta 468 (68.6) (100) (100) Los Angeles 1,348 (7.2) (24.7) (51.2) Chicago 1,484 (13.3) (45.8) (95.0) New York 3,174 (5.2) (17.9) (37.1) Strawberries Atlanta 11,088 (0.0) (3.0) (3.0) Los Angeles 16,215 (0.0) (0.7) (0.7) Chicago 17,554 (0.0) (1.6) (1.6) New York 20,649 (0.0) (2.0) (2.0) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-year quarantine period. d The numbers in parentheses refer to negative percent changes in demand.

PAGE 139

125 Table 5-12 Baseline average prices for the 2000-01 season and percentage changes in prices in the Medfly model by crop Average Price Crop Baseline Scenario I a Scenario II b Scenario III c ($ / Unit) ( % changes ) Tomatoes 8.64 0.92 9.72 13.42 Bell Peppers 10.28 5.93 12.54 12.84 Cucumbers 13.21 2.20 3.93 5.60 Squash 13.91 0.80 9.06 7.11 Eggplants 8.55 7.86 7.86 7.86 Watermelons 13.84 5.85 17.27 33.53 Strawberries 11.77 0.00 4.50 4.50 a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-year quarantine period. Table 5-13.Baseline annual returns of fresh and processed specialty citrus for the 2000-01 season and changes in the Medfly model Fresh Fruit Processed Fruit Utilization On-tree Price On-tree Revenue Utilization On-tree Price On-tree Revenue Total on-tree Revenue Specialty crops (1,000 boxes) ($ per box) ($ 1,000) (1,000 boxes) ($ per box) ($ 1,000) ($ 1,000) Baseline solution K-Early 6,405 2,47 15,820 121,686 2.58 313,950 329,770 Temples 287 0.74 212 865 2.53 2,188 2,400 Early Tangerines 2,480 2.64 6,547 1,063 1.83 1,945 8,492 Honey Tangerines 1,165 8.12 9,460 883 2.28 2,013 11,473 Tangelos 563 0.74 416 1,234 2.46 3,036 3,452 Scenario I a K-Early 3,269 (-49)d 6.45 (161) (21,085 (33.3) 115,092 (-5.4) 2,61 (1.2) 300,390 (-4.3) 321,475 (-2.5) Temples 54 (-81) 3.76 (408) 203 (-4.2) 689 (-20.3) 2.68 (5.9) 1,847 (-15.6) 2,050 (-14.6) Early Tangerines 1,571 (-36.7) 1.90 (-28) 2,985 (-54) 673 (-36.7) 1.83 (0.0) 1,232 (-36.7) 4,217 (-50.3) Honey Tangerines 1,130 (-3) 4.46 (-45) 5,040 (-46.7) 856 (-3) 2.28 (0.0) 1,952 (-3) 6,992 (-39) Tangelos 14 (-97) 3.75 (407) 53 (-87.3) 843 (-31.7) 2.73 (11) 2,301 (-24) 2,354 (-32)

PAGE 140

126 Table 5-13 Continued Fresh Fruit Processed Fruit Utilization On-tree Price On-tree Revenue Utilization On-tree Price On-tree Revenue Total on-tree Revenue Specialty crops (1,000 boxes) ($ per box) ($ 1,000) (1,000 boxes) ($ per box) ($ 1,000) ($ 1,000) Scenario II b K-Early 1,269 (-80) 9.35 (279) 11,865 (-25) 99,258 (-18.4) 2.62 (1.6) 260,056 (-17.2) 271,921 (-17.54) Temples 0 (-100) 0 (-100) 0 (-100) 573 (-33.8) 2.73 (7.9) 1,564 (-28.5) 1,564 (-34.8) Early Tangerines 705 (-71.6) 2.98 (12.9) 2,101 (-67.9) 302 (-71.6) 1.83 (0.0) 553 (-71.6) 2,654 (-68.7) Honey Tangerines 990 (-15) 8.93 (9.9) 8,841 (-6.5) 750 (-15.1) 2.28 (0.0) 1,710 (-15.1) 10,551 (-8.0) Tangelos 0 (-100.0) 0 (-100.0 0 (-100.0) 731 (-40.7) 2.74 (11.4) 2,003 (-34) 2,004 (-42) Scenario III c K-Early 0 (-100) 0 (-100.0) 0 (-100.0) 57,641 (-52.6) 2.63 (1.9) 151,596 (-51.7) 151,596 (-54) Temples 0 (-100) 0 (-100.0) 0 (-100.0) 337 (-61) 2.73 (7.9) 920 (-58) 920 (-61.7) Early Tangerines 0 (-100) 0 (-100.0) 0 (-100.0) 0 (-100) 0 (-100.0) 0 (-100.0) 0 (-100) Honey Tangerines 582 (-50) 25.32 (212) 14,736 (-55.7) 441 (-50) 2.28 (0.0) 1,005 (-50) 15,741 (-37.2) Tangelos 0 (-100) 0 (-100.0) 0 (-100.0) 430 (-65.1) 2.74 (11.4) 1,178 (-61.2) 1,178 (-65.9) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario III: one-year quarantine period. d The numbers in parentheses refe r to percent changes in the parameters; they can be positive or negative.

PAGE 141

127Table 5-14.Baseline FOB revenues for specialty citrus fo r the 2000-01 season and changes in the Medfly model Baseline Solution Scenario I a Scenario II b Scenario III c FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue FOB Price Fresh Sales FOB Revenue ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) ($ per carton) (1,000 cartons) ($1,000) K-Early 5.67 12,809 72,627 9.91 (74.8) 6,537 (-49) 64,782 (-10.8) 12.62 (123.5) 2,537 (-80.2) 32,017 (-55.9) 0 (-100) 0 (-100.0) 0 (-100.0) Temples 4.80 573 2,750 8.56 (78.3) 108 (-81.2) 924 (-66.4) 0 (-100) 0 (-100.0) 0 (-100.0) 0 (-100) 0 (-100.0) 0 (-100.0) Early Tangerines 5.75 4,959 28,514 7.63 (32.7) 3,142 (-36.6) 23,973 (-15.9) 9.44 (64.2) 1,410 (-71.6) 13,310 (-53.3) 0 (-100) 0 (-100.0) 0 (-100.0) Honey Tangerines 8.50 2,329 19,797 8.92 (4.9) 2,259 (-3.0) 20,150 (1.8) 10.63 (25.1) 1,980 (-15.0) 21,047 (6.3) 15.63 (83.8) 1,165 (-99.91) 18,209 (-8.02) Tangelos 4.80 1,126 5,405 8.56 (78.3) 28 (-97.5) 240 (-95.6) 0 (-100) 0 (-100.0) 0 (-100.0) 0 (-100) 0 (-100.0) 0 (-100.0) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. c Scenario II: one-year quarantine period. d The numbers in parentheses refer to percent changes in the pa rameters; they can be positive or negative.

PAGE 142

128 Table 5-15 Aggregate impacts of a Medfly outbreak in Florida with scenarios of three-month, sixmonth, and one-year quarantine periods Aggregate Impacts ($) Consumer Surplus Loss Revenue Loss Option 1 d Option 2 e Option 1 Option 2 Scenario I a 237,672,602 343,475,606 8,834,760 38,998,760 Scenario II b 821,344,300 1,248,580,549 705,753,280 742,181,280 Scenario III c 1,748,625,935 1,748,625,935 1,027,090,400 1,027,090,400 a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. cScenario II: one-year quarantine period. d Option 1: Medfly model with the export market of Japan. e Option 2: Medfly model without the export market of Japan. Table 5-16.Baseline production and percentage ch anges in crop production in the Medfly model by area Production Areas Baseline Scenario I a Scenario II b Scenario III c (Unit) ( % changes ) Florida 109,746 (1.3)d (80.7) (91.4) California 92,674 (0.7) 13.3 12.9 Texas 7,735 (21.4) 28.1 29.7 Virginia/Maryland 4,272 23.5 (13.2) (8.0) South Carolina 6,854 (32.6) 4.9 (0.5) Alabama/Tennessee 2,139 (46.6) (100.0) (100.0) United States 223,420 (2.7) (34.2) (39.6) Mexico 94,509 (2.4) 41.4 49.4 Total 317,929 (2.6) (11.8) (13.2) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. cScenario II: one-year quarantine period. d The numbers in parentheses indicate negative changes in production. Table 5-17 Baseline revenues by area for the 200 0-01 season and changes in revenue in the Medfly model Change in revenue ($1,000) Areas Baseline ($1,000) Scenario I a Scenario II b Scenario III c Florida 913,306,070 13,707,240 (664,242,280)d (773,130,400) California 752,871,600 (4,884,800) 89,729,000 87,129,500 Texas 58,828,370 (12,608,780) 16,556,590 17,469,280 Virginia/Maryland 42,736,960 10,039,420 (5,666,830) (3,412,580) South Carolina 71,270,950 (23,244,990) 3,527,910 385,150 Alabama/Tennessee 20,477,170 (9,540,2 60) (20,477,170) (20,477,170) United States 1,859,491,120 (26,532,170) (580,572,780) (692,036,220) Mexico 707,856,930 (3,212,550) 294,302,140 344,487,660 Total 2,567,348,050 (29,744,720) (286,270,640) (347,548,560) a Scenario I: three-month quarantine period. b Scenario II: six-month quarantine period. cScenario II: one-year quarantine period. d The numbers in parentheses indicate negative changes in revenue

PAGE 143

129 CHAPTER 6 POLICY RECOMMENDATIONS AND CONCLUSIONS Areas for Alternative Policy Measures Assessing the level of Medfly Risk in Florida Table 6-1 provides an estimate of the level of Medfly risk in Florida, which is defined as the cumulative expected value of the consequences associated with a Medfly outbreak and/or infestation in Fl orida. As stated in Chapte r 3, Medfly risk encompasses two major cost components: prevention a nd eradication expenditures and producer revenue losses. Expected valu e of Medfly prevention and er adication expenditures varies across locations and seasons. An October in festation in Tampa accounts for the lowest risk case scenario, while the June infestation in this same location stands for the highest risk case scenario with an expected cost approximating $ 4 million. Whereas the summer months are considered the highest risk period for an infestation in Florida, the risk-related costs associated with producer revenue losses ar e very likely to be insignificant, because very few shipments of fresh commodities occur in summer. Expected revenue losses shown in the be low part of Table 6-1 vary with the duration of the quarantine peri od. The three-month, six-month, and one-year quarantine scenarios are associated with 77-day, 98-day, and 119-day ol d infestation, respectively. Expected revenue losses for Option I (incl uding the market of Japan) are $ 39,612, $ 80,240, and $ 655,918 for a three-month, si x-month, and one-year quarantine period respectively. The expected revenue losses for Option 2 are higher than those for Option I, reflecting the importance of Japanese export markets for the survival of the Florida

PAGE 144

130 grapefruit industry. These re sults suggest the need for Florida growers to negotiate with Japan and agree on a Medfly fly-fr ee zone certification protocol. It is noteworthy to recall here that th e SPS Agreement encourages a myopic focus on risk-related costs of import protocols and eradication expenditures. Consideration of producer revenue losses in the definition of th e Acceptable Level Risk (ALR) is likely to be seen as a violation of the spirit of the SPS Agreement. Our research findings show that the producer revenue losses cannot be ignored, because any Medfly infestation in Florida will further undermine the compe titive position of Florida grow ers in the worlds sectors of fresh commodities. The probabilities of a Medfly endemic situation in Florida are low, but the consequences of such an event are very high, threatening the survival of the production of fresh commodities in Florida. This issue emphasizes the debate regarding the incongruence of the SPS Agreement with the cost-b enefit-analysis approach advocated by the U.S. Executive Branch directives. Assessing the effects of entry co nditions on the level of risk Table 6-2 gives a summary of the effects of entry conditions on the level of Medfly risk in Florida. These effects are the sa me on both expected costs of prevention and eradication and producer revenu e losses. One-percent increas e in the average number of bugs present per infested unit leads to approxima tely 0.42-percent increa se in the level of Medfly risk. On the other hand, the ratio of change between the level of risk and the number of passengers (N) or the level of infestation (p) is roughly 1:1: one-percent increase in N or p will result in one-percent in crease in the level of risk. Therefore, the increasing number of international passengers entering Florida can be considered the driving parameter that will impact on th e likelihood of a Medfly introduction and establishment in Florida.

PAGE 145

131 Our study also investigated the relationships between the parameters N and p. As the number of international passengers, N, te nds to increase over time, APHIS managers are expected to increase the staffing level at the inspection points with a view to reducing the level of infestation, p, and, at least, keeping constant the level of Medfly risk in Florida. When N increases by 5%, 4.45 additiona l staff years are needed to keep constant the level of Medfly risk in Florida. The leve l of infestation, p, will be reduced by 4.76% and the passenger baggage clearance costs to achieve this target are estimated at $ 213,274. By the same token, the projected ad ditional clearance cost is $ 426,563, when N increases by 10%. These results suggest that APHIS would have to maintain significant increases in inspection technologies and st affing to keep pace with a continuous and increasing movement of potentia lly infested host material. Improving the single trap sensitivity Table 6-4 presents a summary of the potenti al gains of a technological change in the trap design. Improving the attractiveness of female attractants in the field would contribute to increase the single trap sensitivity ( ) and reduce the optimal trapping density. When, for instance, the single trap sensitivity doubles, the probability of detecting low populations will increase by 201.66% and the optimal trap density will decrease by 50.19%. Another important finding is that the rate of change in the probability of detection increases as the si ngle trap sensitivity increases. When is multiplied by 5, the probability of detecti on will increase by more than ten times, leading to a reduction in the optimal tra pping density by approximately 80%.

PAGE 146

132 Improving the pesticide efficacy Another important area of concern is to check on the effect of pesticide efficacy on the level of risk. The computations of the multiple trap sensitivities, probabilities of detection, and optimal trap densities in Chapter 4 are made under the assumption of 90percent pesticide efficacy, meaning that 90% of pests are killed during each spraying operation. Figure 6-2 shows the shifting out of the curve of the eradication cost function with an increase in pesticide efficacy, indi cating a decline in total eradication cost. Nevertheless, a decline in the pesticide efficacy will lead to a higher eradication cost and, in some circumstances, to higher probability of Medfly establishment in Florida. Table 6-5 shows the impact of a change in pesticide efficacy on the level of Medfly risk in Florida. When the pesticide efficacy decreases from 90% to 80%, the level of Medfly risk in Tampa (October) increases by 23. 31%. The rate of change in the level of Medfly risk varies across lo cations and seasons. With th e same 80-percent pesticide efficacy, the level of Medfly risk increases by 28.13% and 22.33% in Tampa during the months of June and February, respectively. Furthermore, the lower the pesticide efficacy, the higher will be the level of Medfly risk. When the pesticide efficacy drops to 70%, for instance, the level of risk in Tampa during the months of June and February will increase by 61.25% and 40.91%, respectively. Summary and Conclusions The overall objective of this study was to an alyze the cost implications of a Medfly outbreak and/or infestation for APHIS and the State of Florida in general, and for Florida producers and consumers in particular. In Ch apter 2, we reviewed literature on Medfly introductions and infestations in Florida. In Chapter 3, we discusse d the issues regarding the application of Sanitary and Phytosanit ay (SPS) Agreement and the major rules and

PAGE 147

133 quarantine regulations that affect the fruit a nd vegetable sector in the event of a Medfly outbreak in Florida. In Chapter 4, we de veloped a Bayesian modeling framework to examine the trade-off between early and late detection in terms of expected costs of APHISs prevention, detection, and eradicat ion program. Finally, the results of welfare models are presented in Chapter 5. Our Bayesian modeling framework provides support for the h ypothesis that there exists an optimal trapping density that va ries across locations and seasons. Because the computed values of probability of detection of low Medfly populations are very low, the corresponding optimal trapping densities are ve ry high, ranging from 82 to 465 traps per ha for ML traps and from 9 to 80 traps fo r TML traps. It would be infeasible and extremely costly to maintain such high tr ap densities over a wi de area. Alternative solutions lie in the search for an increase in pesticide efficacy and an improvement of the performance of the trapping system. Potentia l gains from improving the trap technology would include increasing the sensitivity of the trapping system in detecting small populations of C. capitata and lowering optimal trap densities. Emphasis should be placed on developing potent synthe tic attractants for female C. capitata Such a development will provide a new dimension to the detection survey. Our spatial equilibrium models provide estimates of potential welfare changes from a Medfly infestation under s cenarios of three-month, sixmonth, and one-year quarantine periods. A Medfly outbreak in Florida is expe cted to have significant aggregate impacts on producer revenues and consumer surplus. In the event of a three-month quarantine period in Florida (Scenario I), aggregate reve nues in the whole fruit and vegetable sector are expected to decline approximately $9 million or $39 million, depending upon whether

PAGE 148

134 Florida growers could negotiate a certification protocol for the exports of fresh grapefruit to Japan. Consumers surplus is expected to decline $237.6 million (in the Medfly model with the market of Japan Option I) or $343. 4 million (in the Medfly model without the market of Japan Option II). Impacts of a six-month quarantine peri od (Scenario II) in Florida are more significant for both consumers and producers. Consumer surplus losses increase to $821.3 million (Option I) or $1.25 billion (Option II), while Florida shippers stand to lose $705.7 million (Option I) or $742.2 million (Opti on II) in shipping point revenues, as a result of severe production losses and signi ficant increases in preand postharvest production costs. An endemic Medfly situati on (Scenario III) is considered a coup de grace to the whole fresh fruit and vegetable industry in Florida. Vegetable production is expected to decline 91.4%. Th e grapefruit and specialty citr us industry is unlikely to survive without the fresh market. This scenar io will result in a $ 1. 03 billion decline in Florida shipping point revenues. Total c onsumer surplus loss will amount to $ 1.75 billion. In addition to the producer revenue losses, the levels of Medf ly risk are also associated with prevention and eradicati on expenditures whose expected values vary across locations and seasons. An October in festation in Tampa accounts for the lowest risk case scenario, while the June infestation in this same location stands for the highest risk case scenario with an expected co st approximating $ 4 million. The expected prevention and eradication cost for a June infestation in Miami is about $ 1 million. Our study also investigated the effects of improved trap sensitivity, change in pesticide efficacy, and entry conditions on the leve ls of Medfly risk in Florida. When, for

PAGE 149

135 instance, the single trap sens itivity doubles, the probability of detecting low populations will increase by 201.66% and the optimal trap density will decrease by 50.19%. The rate of change in the level of Medfly risk resu lting from a change in pesticide efficacy is expected to vary across locations and seas ons. When the pesticide efficacy decreases from 90% to 80%, the level of Medfly risk is expected to increase by 28.13% and 22.33% in Tampa or by 25.31% and 31.65% in Miami during the months of June and February, respectively. The number of international passengers, N, entering Flor ida can be considered the driving parameter that will impact on th e likelihood of a Medfly introduction and establishment in Florida. When N increases by 5%, 4.45 additional staff years are needed to keep constant the level of Medfly risk in Florida. Additional passenger baggage clearance costs would also be needed to keep pace with an increasing number of international passengers entering Florida and to prevent a resulting increase in the level of Medfly risk. Alternative ways of mitigating the risk of Medfly introduction into Florida are to encourage and support suppressi on and eradication activ ities against fruit fly populations Caribbean Basins countri es. Emphasis can also be placed on implementing educational programs aimed at ra ising public awareness of the threat of exotic pests, like the Mediterranean fruit fly. Limitations of the Study and Sug gestions for Further Research Our study provides estimates of the directi on and magnitude of the potential impact of a Medfly outbreak and/or infe station in Florida. The quantitative results of this study may not be taken literally, but they are consis tent with the biological profile of the pest under consideration and with historical data on past interceptions and trap sensitivity. Its primary limitation is the lack of extensive data on different aspects covered in this study.

PAGE 150

136 We often had to make judgment calls to c ope with the risk assessment on the Medfly case. Further research should focus on the im provement of the trapping technology. Field studies can also be done in Me dfly-infested countries to test the sensitivity of different traps to low populations. 0 100000 200000 300000 400000 500000 600000 507798119 DaysEradication Cost o Low Pesticide Efficacity Moderate Pesticide Efficacity Low Pesticide EfficacityFigure 6-1 Shifting of the eradication cost cu rve with change in the pesticide efficacy Table 6-1 Medfly risk levels in Florida Medfly risk ($) Expected prevention and eradication cost Month Tampa Miami October 946 12,281 February 9,333 66,944 June 3,910,186 1,026,099 Expected revenue losses Scenarios Option 1a Option 2b Three-month quarantine 39,612 40,069 Six-month quarantine 80,240 82,444 One-year quarantine 655,918 666,405 aOption 1: Medfly model with the export market of Japan. bOption 2: Medfly model without the export market of Japan.

PAGE 151

137 Table 6-2 Effects of entry conditions on the risk level Number of passengers (N) Number of bugs per infested unit (X(1,0) Level of infestation (p) Change in parameters (% change in risk level) 5% 4.98 2.11 5.01 6% 5.98 2.50 5.98 7% 6.98 2.88 6.98 8% 7.97 3.26 7.99 9% 8.97 3.64 8.97 10% 9.99 3.98 9.96 Table 6-3 Air passenger baggage clearance cost implications of changes in the number of passengers Change in # of passengers (N) Percent change in level of infestation a needed (%) Additional staff years needed Additional clearance costs ($) 5% 4.76 4.45 213,274 6% 5.66 5.34 255,922 7% 6.54 6.23 298,572 8% 7.41 7.12 341,224 9% 8.23 8.01 383,872 10% 9.09 8.90 426,563 a) This is the percent change in the level of infestation required to keep constant the level of Medfly risk as N increases over time Table 6-4 Impact of an improvement in sing le trap sensitivity on the optimal trapping density Probability of detection Optimal trap density Technological change ( ) (% change) = 2 201.66 (50.19) = 3 466.80 (66.48) = 4 781.31 (74.85) = 5 1138.43 (79.66)

PAGE 152

138 Table 6-5 Impact of pesticide effi cacy on the level of Medfly risk Expected Prevention / Eradication Cost ($) Month Tampa Miami High Pesticide Efficacy (90%) October 946 12,281 February 9,333 66,944 June 3,910,186 1,026,099 Moderate Pesticide Efficacy (80%) October % increase 1,166 23.31 14,614 18.99 February % increase 11,416 22.33 83,892 25.31 June % increase 5,010,254 28.13 1,350,899 31.65 Low Pesticide Efficacy (70%) October % increase 1,333 40.91 17,033 38.69 February % increase 13,759 47.43 102,654 53.34 June % increase 6,305,054 61.25 1,748,090 70.36

PAGE 153

139 APPENDIX A PEST POPULATION STRUCRU RE IN MIAMI AND TAMPA Table A-1. Distribution of pest population structur e in Miami per season at 50, 77, 98, and 119 days of the infestation Population Structure Month Day OF Eggs Larvae Pupae AF 50 100 2,200 1,760 ---100 February 77 110 1,210 -------1,980 98 ----------33,880 3,630 119 11,374 97,163 39,930 11,979 1,719 50 660 7,260 7,260 ----1,540 June 77 6,292 69,212 116,063 29,814 17,908 98 42,060 462,656 535,275 73,791 104,351 119 143,921 1,583,131 36,076 81,170 1,096,465 50 200 ------440 1,540 October 77 2,200 ---13,552 17,424 ---98 726 13,310 ------23,474 119 1,815 192,995 423,791 22,361 ---OF = Ovipositing females PAF = Pre-adult females

PAGE 154

140 Table A-2. Distribution of the pest population struct ure in Tampa per season at 50, 77, 98, and 119 days of the infestation Population Structure Month Day OF Eggs Larvae Pupae PAF 50 ----------200 February 77 100 --3,520 880 ---98 1,100 3,630 484 ------119 -----4,840 30,008 6,776 50 770 8,470 9,680 242 1,430 June 77 8,833 97,163 142,683 44,721 15,367 98 55,237 607,601 664,702 122,985 188,603 119 196,190 2,158,083 36,076 81,170 1,744,622 50 200 ---440 1,980 ---October 77 2,200 ------------98 220 32,670 34,848 ---1,980 119 2,200 ---30,976 8,712 ---OF = Ovipositing females; PAF = Pre-adult females

PAGE 155

141 APPENDIX B GRAPEFRUIT MODEL Consider Qv F boxes of grapefruit of variety v grown for the fresh market Qv P boxes of variety v grown for the processed market TPv total production of variety v in boxes PFv growing and harvesting cost for fresh market PRv growing and harvesting cost for processing market Pvj F = avj bvj Zvj inverse demand for va riety v in fresh market j Pvj F price of variety v shipped to destination j Zvj cartons of variety v consumed in market j vj packout rate for variety v in market j E vj quantity of variety v elimin ated for fruit packed in market j Qvj F cartons of variety v intended to be sold in market j jQvj F = Qv F Zvj = vj Qvj F E vj = (1vj) Qvj F PP = Y inverse demand for processing market Y = vJUv jE vj + vJUv Qv P PP price of one gallon of juice in processing market PK vj packing costs per carton PE vj eliminating charge per carton PP processing cost per box Max vj( avj Zvj bvj Zvj 2 ) + ( Y Y2) vPFv Qv F vPRv Qv P vjPK vj Qvj F vj( PE vj + PP) E vj vPP Qv P st .5 jQvj F Qv F Qv F + Qv P TPv Zvj vj Qvj F E vj (1vj ) Qvj F

PAGE 156

142 Y .5 VJ JUv E vj + 5 VJ JUv E vj All variables non-negative The Lagrangian function associated with the quadratic prog ramming model is: L = vj( avj Zvj bvj Zvj 2 ) + ( Y Y2) vPFv Qv F vPRv Qv P vjPK vj Qvj F vj( PE vj + PP) E vj vPP Qv P + v [ Qv F .5 jQvj F ] + wv [TPv Qv F Qv P ] + vj [ vj Qvj F Zvj ] + vj [(1vj ) Qvj F E vj ] + [5 VJ JUv E vj + vJUv Qv P Y ] The first order conditions associated with the Lagrangian function are: (1a) L / Zvj = avj bvj Zvj vj 0 (1b) ( L/ Zvj) Zvj = 0 (1c) Zvj 0 (2a) L / Y = Y Y 0 (2b) ( L/ Y) Y = 0 (2c) Y 0 (3a) L / Qv F = PFv + v wv 0 (3b) ( L/ Qv F) Qv F = 0 (3c) Qv F 0 (4a) L/ Qv P = PRv + JUv wv PP 0 (4b) ( L/ Qv P) Qv P = 0 (4c) Qv P 0 (5a) L/ Qvj F =PK vj v + vj vj + vj [(1vj ) 0 (5b) ( L/ Qvj F) Qvj F = 0 (5c) Qvj F 0 (6a) L/ E vj = PE vj + JUv PP 0 (4b) ( L/ E vj) E vj = 0 (4c) E vj 0

PAGE 157

143 LIST OF REFERENCES Ahl, A.S., and Acree, J.A. 1993. Implications of Regionalization and Risk Assessment for Developing Countries, Revue Elev. Med. Vet. Pays Trop. 46 (1-2):19-22. Allee, W.C. 1931. Animal Aggregation: A Study in General Sociology. University of Chicago Press, Chicago, IL., 334 p. APHIS 1976. Plant Protection and Quarantine Treatment Manual (T107). USDA, APHIS, PPQ. APHIS 1992. A Mediterranean Fruit fly risk assessment: Planning and risk analysis systems. APHIS, USDA. APHIS 1993. An economic impact assessment of the Mediterranean fruit fly cooperative eradication program. APHIS, USDA. APHIS 1994. Medfly Exclusion Evaluation: An Analysis of Potential Pathways for Entry of Medfly into the United States. APHIS, USDA. APHIS 1995a. Medfly Exclusion Evaluation: an Analysis of Potential Pathways for Entry of Medfly into the United States, APHIS, USDA. APHIS 1995b. Risk Assessment: Emergency Regulatory Activities for Medfly. APHIS, USDA. APHIS, 1999. Risk Reduction Strategy: Florida Medfly Program: Environmental Assessment, APHIS, USDA. APHIS, 2001. Fruit Fly Cooperative Control Program: Final Environmental Impact Assessment 2001. APHIS, USDA, 253 p. APHIS 2002. Action Plan for Mediterranean Fru it Fly. Florida Department of Agriculture and Consumer Services, Division of Plant Division (USDA, APHIS, PPQ). APHIS, 2003a. Mediterranean fruit fly acti on plan. APHIS, Plant Protection and Quarantine. APHIS 2003b. Protocol for the Exportation of Fresh Fruit from Florida to Japan. APHIS (USDA). Armstrong J. W. 1982. Development of a Hot-Water Immersion Quarantine Treatment for Hawaiian-grown Brazilian Bananas. J. Econ. Entomol., 75: 787-790.

PAGE 158

144 Armstrong, J. W., and Couey, H. M. 1984. Methyl Bromide Fumigation Treatments at 300C for California Stonefruits Infested with the Mediterranean Fruit Fly (Diptera: Tephritidae). J. Econ. Entomol., 77: 1229-1232. Armstrong, J.W., Schneider, E.L., Garcia, D.L., and Couey, H.M. 1984. Improved Holding Technique for Infested Commodities Used for Mediterranean Fruit Fly Quarantine Treatment Research. J. Econ. Entomol., 77: 553-555. Armstrong, J.W., Schneider, E.L., Garcia, D.L., and Couey, H.M. 1984. Methyl Bromide Quarantine Fumigation for Strawberries Infested with Mediterranean Fruit Fly (Diptera: Tephritidae). J. Econ. Entomol. 77: 680-682. Armstrong, J.W., and Garcia, D.L. 1985. Methyl Bromide Quarantine Fumigations for Hawaii-Grown Cucumbers Infested with Melon Fly and Oriental Fruit Fly. J. Econ. Entomol. 78: 1308-1310. Armstrong, J.W., and Couey, H. M. 1989. Fumigation, Heat, and Cold: In Fruit Flies, their Biology, Natural Enemies, and Control, edited by A.S. Robinson and G. Hooper. Elsevier, p 411-424. Armstrong, J.W. 1991. Postharvest Quarantine Treatments in the Tropics. In Proceedings of First International Symposium of Fruit Flies in the Tropics, edited by S. Vijaysegaran and A. G. Ibrahim. Mardi. Kuala Lumpur, Malaysia, pp 49-59. Atkinson, P.R. 1977. Preliminary Analysis of a Field Populations of Citrus Red Sale, Aonidielia Aurontii (Maskell) and the Measurement and Expression of Stage Duration and Reproduction for Life Tables. Bull. Entomol. Res. 67: 65-87. Back, E.A., and Pemberton, C.E. 1918. The Mediterranean fruit fly in Hawaii. USDA Bulletin, 536. 119 pp. Baker, A. C. 1939. The Basis for Treatment of Products where Fruit Flies are Involved as a Condition for Entry into the United States. USDA Circ. 551: 8 pp. Baker, A. C. 1952. The Vapor-Heat Process. USDA Dept. Agric. Yearbook. U.S. Gov. Printing Office. Washington, D.C. Baker, R.T., Cowley, J.M., and Harte, D.S. 1993. Pest Risk Assessment: A Process for the Assessment of Pest Risk Associated with the Importation of Plants and Plants Products into New Zealand, Lynfield Plant Protection Centre, No. 1, New Zealand Ministry of Agriculture and Fisheries, 16 pp. Balock, J.W., and Kozuma, T. 1954. Sterilization of Papaya by Means of Vapor-Heat Quick Run-up. Special Report No. 7, Fruit Fly Investigations in Hawaii. U.S. Dept. Agric., Entomol. Res. Branch, Honolulu, Hawaii. Barten, A.P., and Bettendorf, L.J. 1989. Price Formation of Fish: an Application of an Inverse Demand System. Eur. Econ. Rev. 33: 1509-1525.

PAGE 159

145 Bateman, M.A. 1972. The Ecology of Fruit Flies. Annu. Rev. Entomol. 17: 493-518. Benschoter, C. A. 1979. Fumigation of Grapefruit with Methyl Bromide for Control of Anastrepha Suspense. J. Econ. Entomol. 72: 401-402. Beroza, M., Green, N., Gertler, S.I., Steiner, L.F., and Miyashita, F. 1961. Insect Attractants: New Attractants for the Mediterranean Fruit Fly. J. Agric. Food Chem. 9:361-365. Bigsby, H. 1994. The Economic Impacts of Horticultural Imports and Associated Pest Risk. Lincoln University, (unpublished document). Bigsby, H., and White, C., 2001. Quantifying Phytosanitary Barriers to Trade. In Interdisciplinary Food Safety Research. Edited by N. H. Hooker & E. A. Murano. CRC Press, Boca Raton, p 69-85. Bodenheimer F.S. 1951. Citrus Entomology in the Middle East, Uitgeverij Dr. W. Junk, S-Gravenhage, 160 pp. Bodenheimer, F.S. 1963. Ceratitis Capitata in the Mediterranean Basin, EPPO, 1963, in Report of the International Conference on the Mediterranean Fruit Fly and San Jose Scale, Grunberg A. ed., Paris. Bredahl, Maury E. and Forsythe, Kenneth W. 1989. Harmonizing Phytosanitary and Sanitary Regulations. The World Economy, Vol. 12, No. 2: 189-206. Brown, C., Lynch, L., and Zilberman, D. 2002. The Economics of Controlling InsectTransmitted Plant Diseases. Amer. J. Agr. Econ. 84(2) : 279-291. Bruhn, C. M., Schutz, H. G., and Sommer, R. 1986. Attitude Change toward Food Irradiation among Conventional and Alternative Consumers. Food Technology: 86-91. Bruhn, C.M., Schutz, H. G., and Sommer, R. 1987. Food Irradiation and Consumer Values. Ecology of Food and Nutrition Vol. 21: 219-235. Burditt, A.K. Jr., and Balock, J. W. 1985. Refrigeration as a Quarantine Treatment for Fruits and Vegetables Infested with Eggs and Larvae of Dacus Dorsalis and Dacus Cucurbitae (Diptera: Tephritidae). Bureau, J. C., and Marette, S. 2000. Accounting for Consumers Preferences in International Trade Rules. In Incorporating Science, Economics, and Sociology in Developing Sanitary and Phytosanitary Standards in International Trade: Proceedings of a Conference, p 170-200. http://www.nap.edu/openbook/03099070902/html The National Academy of Sciences. Buyckx, E. J. 1994. Bioclimatic Effects on the Distribution of the Mediterranean Fruit Fly (Diptera: Tephritidae) in the Maghreb. In Fruit Flies and the Sterile Insect Technique, edited by C. O. Calkins, W. Klassen and P. Liedo. CRC Press, Boca Raton, pp 139-164.

PAGE 160

146 Calkins, C.O., Schroeder, W.J., and Chambers, D.L. 1984. Probability of Detecting Caribbean Fruit Fly, Anastrepha Suspensa (Loew) (Diptera:Tephritidae), Populations with McPhail Traps. J. Econ. Entomol. 77: 198-201. Cantliffe, D.J. and VanSickle, J.J. 2003. Competitiveness of the Spanish and Dutch greenhouse Industries with the Florida Fresh Vegetable Industry. University of Florida, Institute of Food and Agricultural Sciences. Carey, J.R. 1982a. Demography and Population Dynamics of the Mediterranean Fruit Fly. Ecological Modeling 16: 125-150. Carey, J.R. 1982b. A Life Table Examination of Growth Rate and Age Structure Tradeoffs in Mediterranean Fruit Fly Populations. In Fruit flies of Economic Importance, edited by R. Cavalloro. A.A. Balkema/Rotterdam, pp 315-320. Casana-Giner, V., Gandia-Balaguer, A., He rnandez-Alamos, M.M., Mengod-Puerta, C., Garrido-Vivas, A., Primo-Millo, J., and Primo-Yufera, E. 2001. Attractiveness of 79 Compounds and Mixtures to Wild Ceratitis Capitata (Diptera:Tephritidae) in Field Trials. J. Econ. Entomol. 94(4):898-904. Caswell, J. A. 2000. Overview. In Incorporating Science, Economics, and Sociology in Developing Sanitary and Phytosanitary Standards in International Trade: Proceedings of a Conference, p 1-19. http://www.nap.edu/openbook/03099070902/html The National Academy of Sciences Clark, R. A., Steck, G. J., and Howard, V. W. Jr. 1992. Detection, Quarantine, and Eradication of Exotic Fruit Flies in Florida. Florida Department of Agriculture and Consumer Services, Division of Plant Industry. Code of Federal Regulations 2003. Title 7--Agriculture. Chapter III --Animal and Plant Health Inspection Service, Department of Agriculture, Part 301--Domestic Quarantine Notices---Subpart --Mediterranean Fruit Fly. Code of Federal Regulations, Vol. 5, Parts 300-to 399, p 70-78. Cohen, H., and Yuval, B. 2000. Perimeter Trapping Strategy to Reduce Mediterranean Fruit Fly (Diptera:Tephritidae) Damage on Different Host Species in Israel. J. Econ. Entomol. 93(3):721-725. Cook, R. 2002. Emerging Hothouse Industry Poses Challenges for Californias Fresh Tomato Industry. Giannini Foundation of Agricultural Economics. Cox, D.R., and Hinkley, D.V. 1974. Theoretical Statistics. Chapman and Hall, London. Couey, H. Melvin 1983. Development of Quarantine Systems for Host of the Medfly. HortScience 18:45-47. Couey H. M., Linse, E. S., and Nakamura, A. N. 1984. Quarantine Procedure for Hawaiian

PAGE 161

147 Papayas Using Heat and Cold Treatments. J. Econ. Entomol. 77: 984-988. Couey, H.M., and Chew, V. 1986. Confidence Limits and Sample Size in Quarantine Research. J. Econ. Entomol. 79: 887-890. Dowell, R. V., Siddiqui, I.A., Meyer, F., Spaugy, E. L. 1999. Early Results Suggest Sterile Flies May Protect S. California from Medfly. Calif. Agric. 53: 28-32. Drummond, F., Groden, E., and Prokopy, R.J. 1984. Comparative Efficacy and Optimal Positioning of Traps for Monitoring Appl e Maggot Flies (Diptera:Tephritidae). Environ. Entomol. 13:232-235. Edmondson, W.T. 1968. A Graphical Model for Evaluating the Use of the Egg Ratio for Measuring Birth and Death Rates. Oecologia 1: 1-37. Enkerlin, W., and Mumford, J. 1996. Estimation of the Economic Returns of Alternative Components of Management for the Mediterranean Fruit Fly in Israel, Palestine and Jordan. IAEA/FAO-ICCET. Eskafi, F.M., and Kolbe, M.E. 1987. Relation of Trap Catches with Fruit Infestation. USDA/ARS Guatemala Repor t II to APHIS/PPQ. Honolulu, HI: Tropical Fruit and Vegetable Research Laboratory, Agriculture Research Service, U.S. Department of Agriculture. Excerpt, 2004. Sanitary and Phytosanitary Protocols between Importing and Exporting Countries. http://excerpt.ceris.purdue.edu/doc/ctrylist.html Farnsworth, R.L. 1985. An Economic Model to Measure Costs and Benefits of Eradication. In Pest control: Operations and Systems Analysis in Fruit Fly Management, edited by M. Mangel, J.R. Carey, and R.E. Plant. Nato ASI, Series, Vol. 11, pp 67-78. FAO 1995. Economic Evaluation of Damage Caused by, and Methods of Control of, the Mediterranean Fruit Fly in the Maghreb. Report of an Expert Group, FAO, IAEA-TECDOC-830. FAO 1996. International Standards for Phytosanitary Measures: Requirements for the Establishment of Pest Free Areas. FAO, Publication No. 4. FAO 2002. International Standards for Phytosanitary Measures: The Use of Integrated Measures in a Systems Approaches for Pest Risk Management. FAO, Publication No. 4. Federal Register 1984. Ethylene Dibromide: Amendment of Notice to Cancel Registration of Pesticide Products Containing Ethylene Dibromide. Federal Register 49: 14182-14185. Federal Register 2002. Irradiation Phytosanitary Treatment of Imported Fruits and Vegetables. Federal Register, Vol. 67, No. 205, p 65016-65029.

PAGE 162

148 Federal Register 2003a. Importation of Fruit and Vegetables. Federal Register, Vol. 68, No. 243, p 70448-70463. Federal Register 2003b. Fruits and Vegetables from Hawaii. Federal Register, Vol 68, No. 24, p 5796-5800. Federal Register 2004a. Importation of Clementines, Mandarins, and Tangerines from Chile: Proposed Rule. APHIS, Federal Register Document No. 04-6325, p1326213269. Federal Register 2004b. Mexican Hass Avocado Import. Federal Register, Vol. 69, No. 115: 33584-33587. Florida Agricultural Statistics Service 2003. Citrus Summary 2001-2002. Florida Tallahassee, Florida. Florida Department of Agriculture and Consumer Services 1998. Florida Agricultural Facts. Tallahassee, Florida. Gonzalez, R.H. 1978. Introduction and Spread of Agricultural Pests in Latin America: Analysis and Prospects. FAO Plant Protection Bull. 26(2): 41-52. Gould, W. P. and von Winderguth, D.L. 1991. Gamma Irradiation as a Quarantine Treatment for Carambolas Infested with Caribbean Fruit Flies. Florida Entomologist 74 (2): 297-300. Gray, M. G., Allen, C. J., Burmaster, D. E., Ga ge, S. H., Hammitt, J.K., Kaplan, S., Keeney, R. L., Morse, J. G., North, D. W., Nyro p, J. P., Stahevitch, A., and Williams, R. 1998. Principles for Conduct of Pest Risk Analyses: Report of an Expert Workshop. Risk Analysis, Vol. 18, N0. 6. Hafez, M., Abdel M. A.A., Wakid, A.M., and Shoukry, A. 1973. Studies on some Ecological Factors Affecting the Control of the Mediterranean fruit fly, Ceratitis Capitata Wied. In Egypt by the Use of the Sterile Male Technique, Zeitschrift Ang. Entomologie, 73, 230. Hardman, J.M. 1976. Life Table Data for Use in Deterministic and Stochastic Simulation Models Predicting the Growth of Insect Populations under Malthusian Conditions. Can. Entomol. 108: 897-906. Harris, E.J., Nakagawa, J.S., and Urago, T. 1971. Sticky Traps for Detection and Survey of three Tephritids. J. Econ. Entomol. 64:62-65. Harris, E. J., and Olalquiaga, G. 1991. Occurrence and Distribution Patterns of the Mediterranean Fruit Fly in Desert Areas in Chile and Peru, Environ. Entomol. 20, 174-178. Hayes, C.F., Chingon, H. T.G., Nitta, Frederick A., and Wang, W. J. 1984. Temperature

PAGE 163

149 Control as an Alternative Ethylene Dibromide Fumigation for the Control of Fruit Flies (Diptera: Tephritidae) in Papaya. J. Econ. Entomol. 77: 683-686. Heath, R.R., Epsky, N.D., Guzman, A., Rizzo, J., Dueben, B.D., and Jeronimo, F. 1997. Effect of Adding Methyl-Substituted Ammonia Derivatives to a Food-Based Synthetic Atractant on Capture of Mediterranean and Mexican Fruit Flies (Diptera: Tephritidae). J. Econ. Entomol. 90:1584-1589. Hendrichs, J., Ortiz, G., Liedo, P., and Schwarz, A. 1983. Six Years of Successful Medfly Program in Mexico and Guatemala. In Fruit Flies of Economic Importance, edited by R. Cavalloro. A. A. Balkema, Rotterdam, p 353-365. Hendrichs, J. 1996. Action Programs against Fruit Flies of Economic Importance: Session Overview. In Fruit Fly Pests: A World Assessment of their Biology and Management, edited by Bruce A. M., and Gary J. S.. St. Lucie Press. p513-519. Henning, R.G., Shaw, J.G., Kirkpatrick, J.D., Parsons, P.S., and Rhode, R.H. 1972. An Economic Survey of the Mediterranean fruit fly in Central America. Field Report 21. Washington, DC: Economic Research Service, U.S. Department of Agriculture Cooperating with U.S. Agency for International Development. Hoeflich, W. E. 2000. Plant Quarantines and Hass Avocados: Role of Science in Solving Pest Quarantine problems (Hass Avocado Case Study): In Incorporating Science, Economics, and Sociology in Developing Sanitary and Phytosanitary Standards in International Trade: Proceedings of a Conference, p 217227. http://www.nap.edu/openbook/03099070902/html The National Academy of Sciences Hoy, M.A., Jeyaprakash, A., and Nguyen, R. 2001. Long PCR is a Sensitive Method for Detecting Liberobacter Asiaticum in Parasitoids Undergoing Risk Assessment in Quarantine. Biol. Control 22: 278-287. Hughes, R.D. 1962. A Method for Estimating the effects of Mortality on Aphid Populations. J. Anim. Ecol. 31: 389-96. IAEA 1995. A Proposal for Medfly Control or Eradication with the Sterile Insect Technique in the Middle East. IAEA, Vienna, Austria. IAEA 2003. Trapping Guidelines for Area-Wide Fruit Fly Programmes. International Atomic Energy Agency, Vienna, 2003. Ismail, M.A., Hatton, T.T., Dezman, D.J., and Miller W.R. 1986. In Transit Cold Treatment of Florida Grapefruit Shipped to Japan in Refrigerated Van Containers: Problems and Recommendations. Proc. Fla. State Hort. Soc. 99: 117-121. Jang, E.B., Light, D.M., Binder, R.G., Flath, R.A., Carvalho, L.A. 1994. Attraction of Females Mediterranean Fruit Flies to the Five Major Components of Male-Produced Pheromone in a Laboratory Flight Tunnel. J. Chem. Ecol. 20:9-20.

PAGE 164

150 Jang, E.B. 2002. Physiology of Mating Behavior in Mediterranean Fruit Fly (Diptera:Tephritidae): Chemoreception and Male Accessory Gland Fluids in Female Post-mating Behavior. Florida Entomologist 85 (1), 89-93. Katiyar, K.P. and Valerio, S.J. 1963. Estudio sobre la Dispersion y Longetividad de la Mosca del Mediterraneo Ceratitis Capitata Wied. Marca con P32. Turialba 13 : 181184. Katsoyannos, B.I. 1994. Evaluation of Mediterranean Fruit-Fly Traps for Use in the Sterile Insect-Technique Programmes. J. Appl. Entomol. 118:442-452. Katsoyannos, B.I., Kouloussis, N.A., and Carey, J.R. 1998. Seasonal and Annual Occurrence of Mediterranean Fruit Fly (Diptera:Tephritidae) on Chios Island: Differences between two Neighboring Citrus Ochards. Ann. Entomol. Soc. Am. 91: 43-51. Katsoyannos, B.I., Heath, R.R., Papadopoulos, N.T., Epsky, N.D., and Hendrichs, J. 1999a. Field Evaluation of Mediterranean Fruit Fly (Diptera:Tephritidae) Female Selective Attractants for Use in Monitoring Programs. J. Econ. Entomol. 92:583589. Katsoyannos, B.I., Papadopoulos, N.T., Heath, R.R., Hendrichs, J., and Kouloussis, N.A. 1999b. Evaluation of Synthetic Food-Based Attractants for Female Mediterranean Fruit Flies (Diptera:Tephritidae) in McPhail Type Traps. J. Appl. Entomol. 123:607-612. Kawasaki, K. 1991. Eradication of Fruit Flies in Japan. In Proc. Int. Symposium on Biology and Control of Fruit Flies, Okinawa, Japan, 22 pp. Ke, D., and Kader, A.A. 1992. Potential of Controlled Atmospheres for Post-Harvest Insect Disinfestation of Fruits and Vege tables. Postharvest News and Information 3(2):31n-37n. Klassen, W., Lindquist, D.A., and Buyckx, E.J.; 1994. Overview of the Joint FAO/IAEA Division Involvement in Fruit Fly Sterile Insect Technique Programs. In Fruit flies and the Sterile Insect Technique, edited by C. O. Calkins, W. Klassen and P. Liedo. CRC Press, Boca Raton, pp 3-26. Landolt, P. J., Chambers, D. L., and Chew, V. 1984. Alternative to the Use of Probit 9 Mortality as a Criterion for Quarantine Treatments of Fruit Fly (Diptera : Tephritidae) Infested Fruit. J. Econ. Entomol. 77: 285-287. Liquido. J.N., Shinoda, L.A., and Cunningham, R.T. 1991. Host Plants of the Mediterranean Fruit Fly (Diptera: Tephritidae) : An Annotated World Review. Miscellaneous Publications of the Entomological Society of America, No. 77, 1-51. Malacrida, A.R., Torti, C., Gomulski, L.M., Sebastiani, F., Vonviccini, Gasperi, G., and Guglielmino, C.R. 1998. Genetic Aspects of the Worlwide Colonization Process of Ceraratitis Capitata. Genetics 89:501-507.

PAGE 165

151 Malavasi, A., Rohwer, G.G., and Campbell, D. S. 1994. Fruit Fly Free Areas: Strategies to Develop Them. In Fruit Flies and the Sterile Insect Technique, edited by C. O. Calkins, W. Klassen, and P. Liedo. CRC Press, p 165-179. Mass-Colell, A. 2000. Microeconomic Theory. Oxford University Press, 2000. McCarl, B.A., and Spreen, T.H. 1980. Price Endogenous Mathematical Programming Models as a Tool for Sector Analysis. Amer. J. Agr. Econ. 62:87-102. Mitcham, E.J., and McDonald, R.E. 1993. Effects of Quarantine Heat Treatment on Mango Fruit Physiology. Acta Horticulturae 343: 361-366. Mitchell, F. G., and Kader, A. A. 1985. Postharvest Treatments for Insect Control: In Postharvest Technology of Horticultural Crops, edited by A. A. Kader, R. F. Kasmire, F. G. Mitchell, M. S. Reid, N. F. Sommer and J. F. Thompson. Cooperative Extension, University of California, Division of Agriculture and Natural Resources, p 100-103. Monro, H. A. U. 1969. Manual of Fumigation for Insect Control. FAO Agricultural Studies No. 79, Rome. Morgan, M.G., and Henrion, M. 1990. Uncertainty: A Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis. Cambridge University Press. Moss, C. B., Schmitz, T. G., and Schmitz, A. 2005. The New Institutional Economics, Schumpeter, and E-Commerce: In E-Commerce in Agribusiness, edited by T. G. Schmitz, C. B. Moss, A. Kagan, and B. Babcock. Florida Science Source, Inc., p 3-24 Nadel, D.J., and Guerrieri, G. 1969. Experiments on Mediterranean Fruit Fly Control with the Sterile Male Technique. In Sterile Male Technique for Eradication or Control of Harmful Insects. International Atomic Energy Agency, Vienna, pp 97-105. Nakagawa, S., Farias, G.J., and Steiner, L.F. 1970. Response of Female Mediterranean Fruit Flies to Male Lures in the Relative Absence of Males. J. Econ. Entomol. 63:227-229. NAPPO 1998. NAPPO Regional Standards fo r Phytosanitary Measures (RSPM): Surveillance for Quarantine Fruit Flies (In a Portion of a Generally Infested Area). Secretariat of the North American Plant Protection Organization, RSPM # 10, 8 pp. Narayanan, E.S., and Batra, H.N. 1960. Fruit flies and their Control. Indian Council of Agricultural Research (ICAR), New Delhi, 9-12 pp. Newel, W. 1936. Progress Report on the Key West (Florida) Fruit Fly Eradication Project. J. Econ. Entomol. 29:116-120. Pana, Regina 1991. A Model of the World Market for Fresh and Processed Grapefruit.

PAGE 166

152 Unpublished M.S. Thesis, University of Florida. Pantin, D., Snadiford, W., Henry, M., and Preville, C. 2004. Alternative Facing the Caribbean Banana Industry in the Light of the April 1999 World Trade Organizations (WTO) Ruling on the European Unions (EU) Banana Regime. Report Commissioned by WINFA, CPDC, and Oxfam, Caribbean. Papadopoulos, N.T., Katsoyannos, B.I., and Carey, J.R. 2000. Spring and Early Summer Phenology of Ceratitis Capitata (Diptera:Tephritidae) in Northern Greece. In Proceedings, Fifth International Symposium on Fruit Flies of Economic Importance, Penang. Malaysia, CAB Wallingford, UK. Papadopoulos, N. T., Katasoyannos, B.I., Kouloussis, N.A., Hendrichs, J., Carey, J.R., and Heath, R.R. 2001. Early Det ection and Population Monitoring of Ceratitis capitata (Diptera:Thphritidae) in a Mixed-Fruit Orchard in Northern Greece. Journal of Economic Entomology, Vol. 94, no. 4, pp-971-978. Papadopoulos, N. T., Katsoyannos, Byron I., and Carey, J.R. 2002. Demographic Parameters of the Mediterranean Fruit Fly (Diptera: Tephritidae) Reared in Apples. Annals of the Entomological Society of America 95 (5): 564-569. Papadopoulos, N.T., Katsoyannos, B.I., and Nestel, D. 2003. Spatial Autocorrelation Analysis of a Ceratitis Capitata (Diptera: Tephritidae) Adult Population in a Mixed Deciduous Fruit Ochard in Northern Greece. Environ. Entomol. 32 (2): 319-326. Paul, R.E., and Armstrong, J.W., 1994. Introduction. In Insect Pests and Fresh Horticultural Products: Treatments and Responses, Edited by R.E. Paul and J. W. Armstrong. CAB International. Peter, M.A., and Spreen, T.H. 1989. Price Endogenous Mathematical Programming Models and Integrability: An Alternative Approach. Amer. J. Agr. Econ. 71:1342. Plant, R.E., and Cunningham, R.T. 1991. Analyses of the Dispersal of Sterile Mediterranean Fruit Fly (Diptera:Tephritidae) Released from a Point Source. Environ. Entomol. 20 (6): 1493-1503. Press, S.J. 1989. Bayesian Statistics: Principles, Models, and Applications. John Wiley & Sons. Prokopy, R., and Hendrichs, J. 1979. Mating Behavior of Ceratitis Capitata on a FieldCaged Host Tree. Ann. Entomol. Soc. Am. 72: 642-648. Pszczola, D.E. 1992. Irradiated Produce Reaches Midwest Market. Food Technology: 89-92. Regev, U., Gutierrez, A. P., and Feder, G. 1976. Pest as a Common Property Resource : a Case Study of Alfafa Weevil Control. Amer. J. Agr. Econ.: 186-197. Reiherd, C., 1992. Citrus Production Areas Maintained Free of Caribbean Fruit Fly for

PAGE 167

153 Export Certification. In Proc. Int. Symp. On Fruit Flies of Economic Importance, Antigua Guatemala, 1990, edited by P. Liedo and M. Aluja. Springer-Verlag, p 407. Rendleman, C. M., and Spinelli, F. J. 1999. The Cost and Benefits of Animal Disease Prevention: the Case of African Swine Fever in the U.S. Environ. Impact Assess. Rev.: 19: 405-426. Rigney, C.J. 1989. Radiation-Disinfestation of Fr esh Fruit: In Fruit Flies, their Biology, Natural Enemies, and Control, edited by A.S. Robinson and G. Hooper. Elsevier pp 425-434. Roberts, D. 1998. Preliminary Assessment of the Effects of the WTO Agreement on Sanitary and Phytosanitary Trade Regulations. Journal of International Economic Law 377-405. Roberts, D., Josling, E. T., and Orden, D. 1999. A Framework for Analyzing Technical Trade Barriers in Agricultural Markets. Market and Trade Economics Division, Economic Research Service, U.S. Department of Agriculture, Technical Bulletin N0. 1876. Roberts, D. 2000. Sanitary and Phytosanitary Risk Management in the Post-Uruguay Round Era: An Economic Perspective: In Incorporating Science, Economics, and Sociology in Developing Sanitary and Phytosanitary Standards in International Trade: Proceedings of a Conference, p 33-5. http://www.nap.edu/openbook/03099070902/html The National Academy of Sciences Schuh, G. E. 2000. Historical and Social Science Perspectives on the Role of Risk Assessment and Science in Protecting the Domestic Economy: Some Background. In Incorporating Science, Economics, and Sociology in Developing Sanitary and Phytosanitary Standards in International Trade: Proceedings of a Conference, p 23-32. http://www.nap.edu/openbook/03099070902/html The National Academy of Sciences Schutz, H. G., Bruhn, Christine M., and Diaz-Knauf, K. V. 1989. Consumer Attitude toward Irradiated Foods: Effects of Labeling and Benefits Information. Food Technology: 80-86. Scriber, J. 1983. The Medfly in California: Organization of the Eradication and Public Policy. HortScience, Vol. 18(1). Seo, S.T., Balock, J.W., Burditt, A.K. Jr., and Ohinata, K. 1970. Residues of Ethylene Dibromide, Methyl Bromide, and Ethylene Chlorobromide Resulting from Fumigation of Fruits and Vegetables Infested with Fruit Flies. J. Econ. Entomol. 63: 1093-7. Seo, S.T., Kobayashi, R.M., Chambers, D.L., Steiner, L.F., Balock, J.W., Komura, M., and Lee, C.Y.L. 1971. Fumigation with Methyl Bromide plus Refrigeration to Control Infestations of Fruit Flies in Agricultural Commodities. J. Econ. Entomol. 64: 1270-4.

PAGE 168

154 Seo, S.T., Akamine, E.K., Goo. T.T.S., Harris, E.J., and Lee, C.Y.L. 1979. Oriental and Mediterranean Fruit Flies: Fumigation of Papaya, Avocado, Tomato, Bell Pepper, Eggplant, and Banana with Phosphine. J. Econ. Entomol. 72: 354-359. Seo, S.T., Farias, G.F., and Harris, E.J. 1982. Oriental Fruit Fly: Ripening of Fruit and its Effects on the Index of Infestation of Hawaiian Papayas. J. Econ. Entomol. 75:173-178. Seo, S.T., and Tang, C.S. 1982. Hawaiian Fruit Flies (Diptera: Tephritidae): Toxicity of Benzyl Isothiocyanate against Eggs or 1st Instars of Three Species. J. Econ. Entomol. 75: 1132-1135. Serghiou, C., and Symmons, P. 1974. Ecology of the Mediterranean Fruit Fly in Cyprus. U.K. Ministry of Overseas Development-Cyprus Ministry of Agriculture. Sharp, J.L., and Spalding, D.L. 1984. Hot Water as a Quarantine Treatment for Florida Mangoes Infested with Caribbean Fruit Fly. Proceedings of the Florida State Horticultural Society, 97: 335-357. Shoukry, A., and Hafez, M. 1979. Studies on the Biology of the Mediterranean Fruit Fly, Ceratitis capitata. Entomol. Exp. Appl. 26 :33-39. Simpson, S.E. 1993. Caribbean Fruit Fly-Free Zone Certification Protocol in Florida (Diptera: Tephritidae). Florida Entomologist 76 (2): 228-233. Sinclair, W.B., and Lindgren, D.L. 1955. Vapor Heat Sterilization of California Citrus and Avocado Fruits against Fruit-fly Insects. J. Econ. Entomol. 48: 133-148. Smilanick, J.L., and Fouse, D.C. 1989. Quality of Nectarines Stored in Insecticidal LowO2 Atmospheres at 50C and 150C. J. Amer. Soc. Hort. Sci. 114: 432-436. Smith, A.S. and Taylor, T.G. 2003. Production Costs and Commodity Budgets for Selected Florida Vegetables. Publication of the Department of Food and Resource Economics, Florida Cooperative Extension Service, EDIS Document FE436. Soderstrom, E.L., Gardner, P.D., Baritele, J.L., Nolan de L., J.L., and Brandi, D.G. 1984. Economic Cost Evaluation of a Generated Low-oxygen Atmosphere as an Alternative Fumigant in the Bulk Storage of Raisins. J. Econ. Entomol. 77: 457-461. Soria, F., and Cline, J.F. 1962. Etude du Vagabondage de Ceratitis Capitata Wied. En Tunisie a lAide de Radio-Isotope. Ann. Inst. Nat. Res. Agron. 32:79-94. Southwood, T.R.E., and Henderson, P.A. 2000. Ecological Methods. Blackwell Science pp 447-461. Spitler, G.H., and Couey, H.M. 1983. Methyl Bromide Fumigation Treatments of Fruits Infested by the Mediterranean Fruit Fly. J. Econ. Entomol. 76: 547-550.

PAGE 169

155 Spreen, T.H., VanSickle, J.J., Moseley, A.E., Deepak, M.S., Mathers, L. 1995. Use of Methyl Bromide and the Economic Impact of its Proposed Ban on the Florida Fresh and Vegetable Industry. University of Florida, Institute of Food and Agricultural Sciences. Bull. 898(Tech.). Spreen, T.H. 2001. Terrorist Attacks in New York City and Washington, D.C.: Implications for the Florida Citrus Industry. University of Florida, Cooperative Extension Service, Institute of Food and Agricultural Sciences. Spreen, T. H., VanSickle, J. J., and Brewster, C. M. 2002. Trade Distortions in a Free-Trade Zone: the Case of Sanitary and Phytosanitary Restrictions. In Agricultural Globalization Trade and the Environment, edited by Charles B. Moss, Gordon C. Rausser, Andrew Schmitz, Timothy G. Taylor, and David Zilberman. Kluwer Academy Publishers. pp 253-267. Stanton, G. 1977. Implications of the WTO Agreement on Sanitary and Phytosanitary Measures. In: Understanding Technical Barriers to Agricultural Trade, Orden, D. and Roberts, Donna (editors). International Agricultural Research Consortium: St Paul. Taylor, R. 1984. A Flexible Method for Empirically Estimating Probability Functions. West J. Agr. Econ. 9: 66-76. Taylor, R. 1990. Two Practical Procedures for Estimating Multivariate Nonnormal Probability Density Functions. Amer. J. of Agr. Econ., Vol. 72, No 1, 210-217. Thilmany, D. D., and Barrett, C. B. 1996. Regulatory Barriers in an Integrating World Food Market. Review of Agricultural Economics, Vol. 19, No. 1: 91-107. Thomas, M.C., Heppner, J.B., Woodruff,R.E,, Weems,H.V., and Steck,G. J. 2001. Mediterranean Fruit Fly (Ceratitis capitata ): Featured Creatures. Institute of Food and Agricultural Sciences (IFAS), University of Florida, pp 14. Thornsbury, S. D. 2002. Sanitary and Phytosanitary Issues: Where Does the WTO Go from Here. In Agricultural Globalization Trade and the Environment, edited by C. B. Moss, G. C. Rausser, A. Schmitz, T. G. Taylor, and D. Zilberman. Kluwer Academy Publishers, p 269-283. USDA 1993. Guidelines and Discount Rates for Benefit-Cost Analysis of Federal Programs. USDA, Executive Order 12866. USDA 1999. Risk Reduction Strategy: Florida Medfly Program, Environmental Assessment. USDA, APHIS, AMarketing and Regulatory Programs. USDA 2001. Proposed Rule for Mexican Hass Avocado Import Program Expansion: Final Environmental Assessment. APHIS, USDA, 12 pp. USDA 2003. Tangerines: The Easy-Peel Citrus. Economic Research Service, USDA, Fruit and Tree Nuts Outlook / FTS-306.

PAGE 170

156 VanSickle, J.J., Belibasis, E., Cantliffe, D., Thompson, G., Oebker,N. 1994. Competition in the U.S. Winter Fresh Vegetable Indus try. USDA, Economic Research Service, Agricultural Economic Report No 691. VanSickle, J.J, Brewster, C., and Spreen, T.H. 2000. Impact of a Methyl Bromide Ban on the U.S. Vegetable Industry. University of Florida, Institute of Food and Agricultural Sciences, Bulletin 333. Vijaysegaran, S 1994. Fruit Fly Problems in Southeast Asia and Efforts to Meet Them. In Fruit Flies and the Sterile Insect Technique, edited by C. O. Calkins, W. Klassen, and P. Liedo. CRC Press, p 131-164. Vogel, D. 1995. Trading Up: Consumer and Environmental Regulations in a Global Economy. Cambridge, Mass: Havard University Press. Vogel, D. 2000. The Hass Avocado Case: A Political Science Perspective. In Incorporating Science, Economics, and Sociology in Developing Sanitary and Phytosanitary Standards in International Trade: Proceedings of a conference, p 228230. http://www.nap.edu/openbook/03099070902/html The National Academy of Sciences von Windeguth, D.L. 1986. Gamma Irradiation as a Quarantine Treatment for Caribbean Fruit Fly Infested Mangos. Proc. Florida State Hort. Soc. 99: 131-134. von Winderguth, D.L. 1987. Gamma Irradiation as a Quarantine Treatment for Florida Grapefruit Infested with Caribbean Fruit Fly, Anastrepha suspensa (Loew). Proc. Florida State Hort. Soc. 100: 5-7. Wakid, A.M., and Shoukry, A. 1976. Dispersal and Flight Range of the Mediterranean Fruit Fly, Ceratitis Capitata Wied. in Egypt. Z. Ang. Ent. 81: 214-218. Whyte, C.F., Baker, R.T., Cowley, J.M., and Harte, D.S. 1996. Pest Establishment: A Quantitative Method for Calculating the Probability of Pest Establishment from Imported Plants and Plant Products, as a Part of Pest Risk Assessment. NZ Plant Protection Centre Publications No. 4, Ne w Zealand Ministry of Agriculture, 11 pp. Williamson, D.L. 1983. The Medfly in California: Methods of Attack. HortScience 18 (1): 44-45. Wonnacott, T.H., and Wonnacott, R.J. 1977. Introductory Statistics, Third Edition. John Wiley & Sons Press. WTO, 2004a. The WTO Agreement on the Application of Sanitary and Phytosanitary Measures (SPS Agreement): Text of the Agreement. http://www.wto.org/english/tratop_e/sps_e/spsagr_e.html Yahia, E.M. 1993. Responses of some Tropical Fruits to Insecticidal Atmospheres. Acta Horticulturae, 343: 371-376.

PAGE 171

157 Yahia, E.M., Ke, D., and Kader, A.A. 1991. Responses of Bartlett Pears to Insecticidal O2 and CO2 Atmospheres. HortScience 26: 734 (abstr.).

PAGE 172

158 BIOGRAPHICAL SKETCH Raphael Yves Pierre is a Haitian by birth, born on October 21, 1956. He received from the Universite dEtat d Haiti a Bachelor of Science degree in agronomy in 1981 and a Bachelor of Science degree in social work in 1982. He worked for various private development agencies from 1981 through 1997. In January 1998, he earned a fellowship grant from the Inter American Foundation and began his graduate studies at the University of Florida. He completed the requirements for a concurrent master program in Latin American Studies (MA) and in Food and Resource Economics (MS). He received these two degrees in May 2000. He took a hiat us from studying for one year and half and worked as a consultant for various interna tional organizations in Haiti. In August 2001, he accepted a position as a doctoral graduate research assistant at the University of Florida. His doctoral degree was awarded in May 2007.


Permanent Link: http://ufdc.ufl.edu/UFE0010125/00001

Material Information

Title: Economic Impact of a Mediterranean Fruit Fly Outbreak in Florida
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010125:00001

Permanent Link: http://ufdc.ufl.edu/UFE0010125/00001

Material Information

Title: Economic Impact of a Mediterranean Fruit Fly Outbreak in Florida
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010125:00001


This item has the following downloads:


Full Text





ECONOMIC IMPACT OF A MEDITERRANEAN FRUIT FLY OUTBREAK IN
FLORIDA















By

RAPHAEL YVES PIERRE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2007

































Copyright 2007

by

Raphael Yves Pierre




























To my wife, Alaine and my son, Andy.















ACKNOWLEDGMENTS

First and foremost, I am indebted to Drs. P. K. Nair and Michael Bannister for

providing me the opportunity of beginning my doctorate studies in the School of Forest

Resources and Conservation.

I wish to express my sincere appreciation and gratitude to my entire supervisory

committee: Drs. John VanSickle (chair); Thomas Spreen (co-chair); Edward Evans;

Norman Leppla; Moss Charles; James Seale; and David Mulkey. They all provided

useful criticisms and suggestions, thereby contributing greatly to the quality of this

dissertation.

Drs. John VanSickle and Edward Evans deserve special mention for allowing me to

complete my doctoral studies in the Department of Food and Resource Economics. They

granted me a graduate research assistantship at a time when I was badly in need.

Drs. Thomas Spreen, Norman Leppla, and Charles Moss deserve special

recognition for their unwavering support and guidance. They gave freely of their time and

advice. I have been very fortunate to have their expertise available to me.

I wish also to recognize tremendous support from Dr. Gary Steck, Dr. David Dean,

Mr. Mike Shannon, Mr. Terry McGovern, Mr. R.E. Burns, and Mr. Loren Carpenter. I

am very appreciative of the opportunity to have worked with them throughout the

research process. They were willing to provide the crucial information needed to conduct

this study.









My life in Gainesville would have been dreary and unbearable without a friendly

environment sustained by SFRC and FRED buddies. I am especially grateful for the

unwavering friendship of Lurleen Walters, Alain Michel and Vony Petit-Frere.

I am also lucky to have had the understanding of a wonderful and supportive friend,

Myrtha Jean-Mary. I am truly indebted to her for help and encouragement throughout my

doctoral studies at the University of Florida.

Finally, I would like to express my love and deepest admiration for my wife,

Alaine Jean, and my son, Andy Pierre, to whom this dissertation is dedicated. Their moral

support and commitment have inspired me to achieve this accomplishment.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TA BLES .................................................................... ............ .. ix

L IST O F F IG U R E S .... ...... ................................................ .. .. ..... .............. xii

A B S T R A C T .............................................. ..........................................x iii

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

P rob lem atic Situ ation ........... ...................................................... .... .. ............
Problem Statem ent .................. ....................................... .. .. ... ...... ....
H y p o th e se s .............................. ............................................................. ............... 3
O objectives ................................................................. ........ .......... 4
T hesis O outline ................................................................. ..... .......... 4

2 REVIEW OF MEDFLY INTRODUCTIONS AND INFESTATIONS IN
FLO R ID A ............................................................... ..... ..... ........ 6

Biological Profile of the Mediterranean Fruit Fly ........................ ................6
Overview of the Florida Medfly Detection and Eradication Program .....................10
Overview of the Florida Fruit and Vegetable Sector...............................................14
S u m m ary ...........................................................................................18

3 POLICY FRAMEWORK OF THE SANITARY AND PHYTOSANITARY
(SP S) A G R E E M E N T ........................................................................ ...................24

The World Trade Organization Agreement on the Application of Sanitary and
Phytosanitary M measures .............. ......... .... ...........................................24
Phytosanitary Protocols for the International Movement of Fresh Fruits and
V vegetables in Fruit Fly Free A reas ................... ..... ... ...................... .... 33
Regulated Post-harvest Treatments and Procedures for the Quarantine Control of
Fruit Flies ....................... ....... ... .. ... .. .. .. .............3.. 9
C including R em arks .............................................................47









4 COST ANALYSIS OF THE MEDFLY DETECTION AND ERADICATION
P R O G R A M ........................................................................................................... 5 4

Specification of the Bayesian Modeling Framework .............................................54
Overview of the Bayesian Decision Process............. ........... ..............54
Definition of the Variables ........... ..... ......... ................... 56
T em poral dim ension .......................................................... ............... 57
Spatial dim ension ..................................... ............... ..... ..... 58
Cost Function ............... ......... ........ ......... 60
Future cost of eradication................. .......... ....................62
O ptim ization m odel .. ...... ............................................. ..... ........ .... 62
Probabilistic M odels ......................... ... .......... .... .. .......... .............. .. 63
Probability of D election: F (X1,t | X 2,t, Z) ................................. ................ 63
Probability of Infestation: F(X 1,0) ......................................................................63
Multiple Trap Sensitivity of McPhail traps: F(X2,t I Xl,t, Z)...........................65
Comparative Sensitivity of McPhail versus Jackson Traps .............................67
R results ..............................................................................................69
P est P population P rejection ............................................................ ...... ........ 70
Size and Cost of the Infestation................................................................ 71
M multiple Trap Sensitivities for M L Traps ................................. ............... 74
Probabilities of Detection for M L Traps .................................. ............... 75
O ptim al T rap D ensities................................................ ............................ 75
C o n c lu sio n s........................................................................................................... 7 6

5 WELFARE ANALYSIS OF A MEDFLY OUTBREAK IN FLORIDA ................... 87

Fundam entals of the Partial Equilibrium M odel ................................... ............ ..... 87
Adaptation of the Spatial Equilibrium Model to the Fruit and Vegetable Industry ...90
G rapefruit M odel ..................................... .................... ... ......... .... 90
V vegetable M odel ................................................ .......... .....91
Specialty M odel ................. ...................... .. .. ..... .. .......... .......... .. 92
Cost Impact of a Medfly Quarantine Restriction on Florida............................94
E m pirical R results .................................................. ............................ 96
Solutions of the Grapefruit Model under a Medfly Quarantine ..........................97
Option I (with the m market of Japan) .................................. ............... 97
Option II (without the market of Japan)................................................... 100
Solutions of the Vegetable Model under a Medfly Quarantine.........................102
T o m ato e s .............................................................................................. 1 0 3
Peppers ................................. .......................... ... ......... 105
Cucumbers ...................... ........... .... ............... 106
S q u a sh .................................................................................................. 1 0 7
Eggplants ............... .................. ........................... .... ....... 108
Watermelons ............. ... .............................. 108
Strawberries ........... .................... ...........109
Solutions of the Specialty Model under a Medfly Quarantine ........................109
A g g reg ate Im p acts .............................................................................................. 1 1 1









6 POLICY RECOMMENDATIONS AND CONCLUSIONS ................................... 129

A areas for A alternative Policy M measures .................................................................... 129
Assessing the level of Medfly Risk in Florida ...............................................129
Assessing the effects of entry conditions on the level of risk .........................130
Improving the single trap sensitivity .... .......... ....................................... 131
Im proving the pesticide efficacy ............................................ ............... 132
Sum m ary and C conclusions ................................................................. ................ ... 132
Limitations of the Study and Suggestions for Further Research ............................135

APPENDIX

A PEST POPULATION STRUCTURE IN MIAMI AND TAMPA........................139

B G R A PE FR U IT M O D E L ............................................................... .....................14 1

LIST OF REFEREN CES ......... ......... ..... ............... ..................................... 143

BIOGRAPHICAL SKETCH .............. ........... ... .............. 158
















LIST OF TABLES


page

2-1 Major fruit and vegetable crops grown in Florida according to their importance
as M edfly h o sts.................................................. ................ 19

2-2 Costs of Medfly infestations in Florida...................... ....................20

2-3 Baseline and emergency budgets for the Medfly prevention and detection
program ........ ..................................... ............... 22

2-4 Important fruits: acreage, yield, and use (by crop), 2001-02 ..............................22

2-5 Important vegetables: acreage, yield, and utilization, by crop, 2001-02 .................23

3-1 Formats for phytosanitary protocols for the international movement of fresh fruit
and vegetable com m odities ............................................. ............................. 50

3-2 Regulated postharvest treatments, advantages, limitations, and alternatives under
con sideration .........................................................................52

4-1 Distribution of day degrees required by stage ............. ........................................77

4.2 Average monthly distances flown by different fractions of Medfly population......77

4-3 Eradication cost equations......................................................... ............... 79

4-4 Hyperbolic tangent approximation of marginal probability function for multiple
trapping sensitivity of M cPhail traps ............................................ ............... 79

4-5 Coefficients of comparative sensitivity and mean daily captures by period............80

4-6 Distribution of the expected population size and generation time per location
and per season at 50, 77, 98, and 119 days a of the infestation..............................80

4-7 Distribution of the intrinsic rates of increase and doubling times of the pest
population per location and per season ........................................ ............... 81

4-8 Distribution of the infested area, quarantine area, and eradication cost per
location and per season at 50, 77, 98, and 119 days a of the infestation..................81









4-9 Distribution of multiple trap sensitivities for ML traps under all outbreak
scenarios for different trap densities ............................................. ............... 82

4.10 Marginal trap sensitivities for ML traps under different outbreak scenarios ...........83

4-11 Distribution of probabilities of detection for ML traps under all outbreak
scenarios for different trap densities ............................................. ............... 84

4-12 Marginal values of probability of detection for ML traps densities under
different outbreak scenarios ................................. .....................................85

4-13 Optimal trapping density per type of trap, location and month ............................. 86

5.1 Outbreak scenarios and cost implications of a Medfly infestation on the Florida
fruit and vegetable industry ........................................................ ............. ..114

5-2 Baseline annual returns of fresh and processed Florida grapefruit for the 2000-
01 season and changes in the medfly model including the market of Japan ..........115

5-3 Baseline world FOB revenue for red grapefruit for the 2000-01 season and
changes in the Medfly model including the market of Japan..............................116

5-4 Baseline world FOB revenue for white grapefruit for the 2000-01 season and
changes in the Medfly model including the market of Japan..............................117

5-5 Baseline annual returns of fresh and processed Florida grapefruit for the 2000-
01 season and changes in the Medfly model excluding the market of Japan......... 118

5-6 Baseline world FOB revenue for red grapefruit for the 2000-01 season and
changes in the Medfly model excluding the market of Japan.............................1.19

5-7 Baseline world FOB revenue for white grapefruit for the 2000-01 season and
changes in the Medfly model excluding the market of Japan..............................120

5-8 Planted acreage in the baseline and Medfly models by crop and area.................121

5-9 Baseline production for the 2000-01 season and percentage changes in
production in the Medfly model by crop and area ...........................................122

5-10 Baseline revenue for the 2000-01 season and changes in revenues in the Medfly
m odel by crop and area ................................................ .............................. 123

5-11 Baseline demand for the 2000-01 season and percentage changes in demand in
the M edfly model by crop and market ............................................................. 124

5-12 Baseline average prices for the 2000-01 season and percentage changes in prices
in the M edfly m odel by crop ....................................................................... ..... 125









5-13 Baseline annual returns of fresh and processed specialty citrus for the 2000-01
season and changes in the M edfly model ......................... .... ......... ......... 125

5-14 Baseline FOB revenues for specialty citrus for the 2000-01 season and changes
in the M edfly m odel .......................................... ........................ 127

5-15 Aggregate impacts of a Medfly outbreak in Florida with scenarios of three-
month, six-month, and one-year quarantine periods....................... ...........128

5-16 Baseline production and percentage changes in crop production in the Medfly
m odel by area ....................................................................... 12 8

5-17 Baseline revenues by area for the 2000-01 season and changes in revenue in the
M edfly m odel ....................................................................... 128

6-1 Medfly risk levels in Florida ........ .... ............... ............... 136

6-2 Effects of entry conditions on the risk level....... ... ..................................... ..137

6-3 Air passenger baggage clearance cost implications of changes in the number of
passengers ............ ..... ..... ... ......... ......... ................................137

6-4 Impact of an improvement in single trap sensitivity on the optimal trapping
d e n sity .......................................................................... 1 3 7

6-5 Impact of pesticide efficacy on the level of Medfly risk................. ........... 138
















LIST OF FIGURES


Figurege

2-1 Fem ale M editerranean fruit fly ........................................ .......................... 19

2-2 M ale M editerranean fruit fly ............................................................................20

2-3 Seasonal distribution of Medfly occurrences in Florida ............... ..................21

2-4 Distribution of Medfly interceptions in Florida.................... ................21

3-1 Inward shift in import supply resulting from the imposition of an SPS barrier.......49

4-1 B ayesian decision process ...................... .. .. .......................................................78

4-2 Treatment and quarantine areas of an infestation scenario in Miami (October)......79

5.1 Price equilibrium, aggregate demand and supply ..........................113

5-2 Effects of phytosanitary regulations .......................................... ...............114

6-1 Shifting of the eradication cost curve with change in the pesticide efficacy .........136















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

ECONOMIC IMPACT OF A MEDITERRANEAN FRUIT FLY OUTBREAK IN
FLORIDA

By

Raphael Yves Pierre

May 2007

Chair: John J. VanSickle
Cochair: Thomas H. Spreen
Major Department: Food and Resource Economics

We evaluated the potential impact of a Mediterranean fruit fly infestation in

Florida. We developed a Bayesian decision framework to analyze the costs of Florida

Medfly prevention, detection and eradication programs under early versus late detection

scenarios. Modeling results support the hypothesis that optimal trapping density varies

across locations and seasons. Because of the low probability of detecting small Medfly

populations, the corresponding optimal trapping densities are high, ranging from 82 to

465 traps per ha for McPhail traps and from 9 to 80 traps per ha for Jackson traps. It

would be extremely costly to maintain such high trap densities over a wide area.

Alternative solutions lie in the search for an increase in pesticide efficacy and an

improvement of the trapping technology. Development of more effective female-targeted

trapping systems will provide a new dimension to the detection of small Medfly

populations.









Partial equilibrium models were also used to investigate welfare changes for the

major fruit and vegetable crops under scenarios of a 3 mo, 6 mo, and 1 y quarantine

period. Our analysis provides insight regarding the magnitude of welfare changes

associated with a Medfly outbreak and/or infestation in Florida. These changes vary

across crops, depending on the competitive position of Florida growers for the crop, size

of the infested area, and length of quarantine period.

Finally, we tested the effects of changing the entry conditions on the level of

Medfly risk in Florida. Our sensitivity-analysis tests showed the increasing number of

international passengers entering Florida to be the driving parameter affecting Medfly

introduction and establishment in Florida. Additional passenger baggage-clearance costs

will be continuously needed to keep pace with the increasing number of international

travelers entering Florida. Another way to mitigate the risk of Medfly introduction into

Florida is to encourage and support suppression and eradication activities against fruit fly

populations in Caribbean basin countries.














CHAPTER 1
INTRODUCTION

Problematic Situation

Fruit and vegetable production is an important industry in the state of Florida, with

an estimated value of $4 billion as farmer cash receipts, including considerable exports.

Florida farmers used a little more than 10 million of the state's nearly 35 million acres

and made more than a $12 billion direct impact on the state's economy (Florida

Department of Agriculture and Consumer Services 1998). Citrus sales topped $1.3

billion, accounting for more than 22 % of state agriculture sales in 1997. Florida farmers

provide more than 70% of the nation's citrus and nearly 10% of its vegetables. They

produce 80% of the country's domestically grown vegetables during the winter months.

The Mediterranean fruit fly (popularly known as Medfly ) poses a serious threat to

fresh fruit and vegetable production throughout Florida and southern Florida. Since its

first detection in Florida in 1929, the pest has been intermittently introduced in the state.

Most of its introductions can be traced to accidental or intentional (smuggling) human

interventions; Florida is the path for large volume of commodities and international

travelers (APHIS 1995a, 1999, 2001, 2003a).

Because of its wide range of hosts, its explosive reproductive capacity, and its

extreme adaptability to adverse ecological conditions (Klassen et al. 1994), Medfly is

considered one of the world's most destructive fruit pests and is the subject of strict

1 The molecular data on the genetic variability of the Mediterranean fruit fly confirm that the name
"Medfly" may be inappropriate because the ancestral home of C. capitata is Africa, so it ought to be called
"Africafly." (Malacrida et al.1998; Hoy 2003).









quarantine and comprehensive control programs. Countries are allowed to impose

technical barriers to the free movement of fresh fruit and vegetable commodities to

protect national production from pests, diseases, and contaminants; but the Sanitary and

Phytosanitary (SPS) Agreement requires that any regulatory measure be based on

scientific assessment of risks associated with introducing these invasive species.

Regulatory and eradication2 measures are justified because the pest significantly

increases production costs; and fruits grown in Medfly-infested regions cannot be

exported to Medfly-free areas (thereby affecting national and international trade).

APHIS and the State of Florida spend millions of dollars each year on research and

exclusion3 (risk-assessment studies, clearing imported cargos, inspecting and regulating

passenger baggage, restrictions on imports) to prevent a wide complex of exotic animal

and plant diseases and pests, including Medfly. The Medfly-detection program was

designed for early detection of Medfly introductions. Florida is considered Medfly-free

thanks to periodical eradication campaigns combining non-chemical and chemical control

methods. Prompt aerial applications of malathion bait over Florida cities regarded as

major entry points of Medfly have been key to the success of these campaigns.

However, public concern has been growing about the intensive use of malathion in

eradication operations. Some critics advocate the abandonment of all chemicals. Florida's

legislation is likely to restrict aerial applications of pesticides over heavily populated

areas (APHIS 1999). Alternative detection, prevention, and eradication programs

2 Eradication is a process used to achieve a fruit-fly free area. Trapping surveys are carried out to measure
the efficacy of control measures (such as bait sprays, SIT, biological control, and MAT) used to eliminate a
pest from an area (IAEA 2003).

3 Exclusion is a process used to minimize the risk of introduction or re-introduction of a pest in a free-area.
Trapping survey are carried out to determine the presence of species that are under exclusion measures and
to confirm (or reject) the free-area status (IAEA 2003).









involving intensive trapping and regulatory measures, preventive sterile-release

programs, and ground-only pesticide applications are being studied in a context of

increased trade and human travel.

Problem Statement

Medfly incursion and establishment in Florida involve factors beyond the control of

APHIS' policy makers. Medfly infestations in Florida might become likelier with

increased trade and movement of people from infested countries. In such a context,

timely detection is crucial to effective management and control of a Medfly population at

the earliest possible stage. However, discovering low populations of wild Medflies in

continuous plantings of a host species is exceedingly difficult (Calkins et al. 1984).

Optimal control may require greater density of traps to better detect low populations, but

maintaining a high trap density over a wide area is extremely costly. Decisions to

increase trap density must consider the consequences of failing to detect low populations.

Another area of concern is the potential impact of a Medfly outbreak in the Florida

fruit and vegetable sector. Economic studies (APHIS (1993, 1999) conservatively

measure producer losses, ignoring the uncertain effectiveness of detection and eradication

program effectiveness, potential changes in retail prices, and costs to consumers. An

appropriate modeling framework is needed for a welfare analysis of a Medfly outbreak in

Florida.

Hypotheses

There is a trade-off between early and late detection in cost management of the

Medfly-eradication program. Early-detection costs are high for trapping, but low for

eradication (with a low probability of establishment). Late detection costs for eradication

are high (with a high probability of establishment).









* Hypothesis 1

Optimal trapping density varies across locations and seasons; it minimizes the

expected cost of the prevention and eradication program

A Medly outbreak in Florida would probably cause a decrease in export markets,

an increase in pre- and post-harvest production costs, and a yield reduction from the

infested areas.

* Hypothesis 2

The magnitude of welfare changes associated with a Medfly outbreak will vary

from one crop to another, depending on growers' competitive position for the

crop, size of infested area, and length of quarantine period.

Objectives

Our overall objective was to analyze the cost implications of a Medfly outbreak in

Florida for APHIS and the State of Florida in general, and for producers and consumers

in particular. We analyzed a variety of outbreak scenarios. Specific objectives are to:

1. Review the policy framework for using the Sanitary and Phytosanitary (SPS)
Agreement

2. Evaluate costs of Medfly detection and eradication programs under different pest-
detection scenarios

3. Use estimated probabilities of detection to determine the optimal trapping strategy
specific to each location and season of the year

4. Evaluate welfare changes associated with outbreak scenarios of 3 mo, 6 mo, and 1
y quarantine periods

5. Formulate alternative policy measures to reduce the risk of Medfly introduction and
establishment in Florida.

Thesis Outline

The plan of this thesis is as follows. The next chapter presents an overview of

biological profile of the Mediterranean fruit fly, Florida fruit and vegetable sector, and









Medfly-detection and eradication programs. The third chapter follows with an analysis of

the World Trade Organization (WTO) Agreement on the application of the SPS

measures. This review focuses on phytosanitary protocols and regulated postharvest

treatments for the domestic and international movement of fresh fruit and vegetable

commodities.

Chapter 4 describes the components of the Bayesian decision framework: spatial-

temporal model of infestation, models of probability, and cost minimization model. These

models were used to estimate (1) probabilities of detecting a Medfly infestation in Florida

under different outbreak scenarios and (2) optimal trapping densities minimizing total

expected cost of prevention and eradication programs. Chapter 5 follows with the

description of optimization models used for analyzing welfare changes associated with a

Medfly outbreak in Florida. We analyzed scenarios of 3 mo, 6 mo, and 1 y quarantine

periods.

Chapter 6 deals with the determination of Medfly risk by combining the results of

Chapters 4 and 5. Sensitivity-analyses are conducted to evaluate the effects of changes in

the major factors leading to the introduction and establishment of Medfly into Florida.

The paper concludes with the formulation of alternative policy measures to reduce the

risk of a Medfly endemic situation in Florida.














CHAPTER 2
REVIEW OF MEDFLY INTRODUCTIONS AND INFESTATIONS IN FLORIDA

Biological Profile of the Mediterranean Fruit Fly

The Mediterranean fruit fly (Ceratitis capitata) is the most notorious member of the

family Tephritidae (Narayanan and Batra, 1960). It originates from sub-Saharan Africa

and is now widely distributed in many countries of the Mediterranean coast. It has no

near relatives in the western hemisphere. Adult Medflies are slightly smaller than a

housefly and are very colorful. They can be readily distinguished from any native fruit

flies of the new world. Females have a characteristic yellow wing pattern and the apical

half of the scutellum is black (Figure 2-1), and males can be separated from all other

members of his family by the black pointed expansion at the apex of the anterior pair of

orbital setae (Figure 2-2).

Medfly is one of the most destructive agricultural pests in the world (Papadopoulos

et al. 2001, 2002). Medfly reduces the yield of fruit and vegetable crops and it affects

their quality. Ripened fruits infested with the Medfly may be unfit to eat, as the female

pierces the soft skin and deposits eggs in the puncture. A female Medfly may lay as many

as 800 eggs during her lifetime. The average daily oviposition rate is 11 eggs (Thomas et

al. 2001; Papadopoulos et al. 2002). Those eggs produce maggots or wormlike larvae that

feed on the pulp of the fresh fruit, drop to the ground, and transform into pupae in the soil

or elsewhere. Many researchers report high mortality in Medflies, particularly in citrus

hosts (Back and Pemberton 1918; Bodenheimer 1951). However, high fecundity and









adult longevity generally more than compensate for high larval mortality (Carey 1982a).

If only 10% of larvae survive, the population can double every 10 days.

Life expectancy and insect development depend on climatic conditions. Under

typical Florida summer weather conditions, the life cycle of the medfly (which also

includes maturation of the pupae into adults) varies from 21 to 30 days (Thomas et al.

2001). Optimum temperatures for the most rapid development and high potential for

spread range from 70 to 900 F. Adults emerge in the largest numbers early in the morning

during warm-temperature and longer freeze-free periods, and emerge more sporadically

during cool weather. According to Buyckx (1994) and Bodenheimer (1963), Ceratitis

capitata is confined to regions where cold weather combined with low humidity

interrupts development in egg, larval, and pupal stages for less than 100 days. Low

temperatures greatly increase the duration of the egg stage. Nevertheless, field studies

report that pupae can carry the species through unfavorable conditions such as lack of

food, or water, or temperature extremes (Thomas et al. 2001). Pupae can develop at

temperatures as low as 500F (10C), making the Medfly extremely adaptable and a

serious pest-control challenge.

Ceratitis capitata is a polyphagous and multivoltine tropical species that has spread

to all countries bordering the Mediterranean region, southern Europe, western Australia,

Central and South America, and Hawaii. It has been recorded in more than 200 different

types of fruits and vegetables. Medfly was first discovered in Central America in April

1955 near San Jose, Costa Rica. It then spread into Panama, Nicaragua, El Salvador,

Honduras, Guatemala, and the southern border of Mexico (Gonzales 1978). Primary hosts

of Medfly in Central America are coffee, tangerine, orange, grapefruit, peach, and









tropical almonds. Medfly populations there mainly vary according to seasonal factors,

reaching a peak during the dry season and with the maturation period of the host fruits,

particularly coffee (Henning et al. 1972; Eskafi and Kolbe 1987). Low populations

coincide with increased rainfall but vary depending on elevation.

Much of Florida would maintain high Medfly populations because of both

favorable climate and availability of preferred host materials throughout the year.

Average mean monthly temperatures for the city of Miami (Dade County) are in the

optimal range 8 months a year, and above the minimum temperature for Medfly

development all year. Almost all major crops produced in Florida (oranges, grapefruit,

lemons, tangerines, and tangelos) could be heavily infested by Medfly (Table 2-1). Some

vegetables and melons are classified as occasionally or rarely infested hosts. Medfly also

could be attracted to patches of ornamental trees that provide shelter in otherwise barren

areas (Buyckx 1994). Thorough knowledge of the hosts in one country often can help

predict the hosts most likely to be infested in a newly infested country.

Movements of flies seemed to be influenced more by the distribution of hosts than

by wind direction (Buyckx 1994; Wakid and Shoukry 1976; Hafez et al. 1973). Flies do

not move great distances under favorable climatic conditions and where fruits and

vegetables are available. Movements appear to be restricted to a few hundred meters per

week. Long distance flights of over 20 km are associated with some behavioral surviving

mechanism for Medly populations in areas where fruit suitable for ovipositing are

unavailable during certain times (Harris and Olalquiaga 1991). Thus, population density

and the economic damage would vary from area to area, depending upon the seasonal

availability of hosts and the presence or absence of unfavorable climatic conditions.









The biggest factors in dispersal and spread of Medfly populations are fruit transport

and trade (Buyckx 1994). Modem roads allow fresh fruit to be transported throughout the

states. International passengers might also bypass detection and enter the United States

with high-risk Medfly host material. Medfly surveys conducted by APHIS (1995a) show

that the major pathways for Medfly into Florida are passenger and crew baggage from

foreign countries. Approximately 93% of all high-risk and infested host materials enter

Florida at the ports of Miami and Orlando. JFK International Airport is the only other

airport with a significant volume (4%) of high-risk and Medfly-infested material destined

for Florida.

The risk of Medfly entering the U. S. via Mexico is small, because Mexico is

Medfly-free. However, there is a potential risk for Medfly introduction from infested fruit

carried into the United States by travelers from a third country. About 2.17 % of people

crossing the Mexican border illegally each year are not of Mexican descent, and most of

them (78%) are from Central American countries where Medfly occurs (APHIS 1992).

Many of these illegal aliens carry food with them that could be infested with fruit fly.

These findings are consistent with historical data on Medfly occurrences and

infestations in Florida (Table 2-2). Several Medfly occurrences are closely associated

with ship and boat traffic in Miami (1964, 1967, 1984, 1985 and 1988); Tampa (1981);

and Fort-Lauderdale (1990). Five of the nine Florida infestations occurred in or around

Miami Springs, a small upper-income housing area at the northern edge of Miami

International Airport. Seasonal distribution of Medfly occurrences in Florida (Figure 2-3)

seems not be related to colonization potential for Medfly, as temperatures and host









availability are favorable all year in southern Florida (APHIS 1992). Seasonal

distribution of interceptions from Latin America is the same as seasonal distribution of

Medfly detection in Florida. However, seasonal distribution of interceptions from Hawaii

(where Medfly is permanently established) markedly differs from seasonal distribution of

Medfly detection in Florida (Figure 2-4). These observations strongly suggest that Latin

American countries (instead of Hawaii) could be a high-risk source of Medfly

infestations in Florida.

Overview of the Florida Medfly Detection and Eradication Program

Florida is a managed Medfly-free area4 in which the pest has been intermittently

eradicated. As argued previously, no limiting natural factors prevent its establishment in

Florida. The State is permanently protected by federal regulatory actions (quarantine

inspection, surveillance networks, import restrictions) controlling the movement of

people and commodities. Some of these measures target specific importing countries

where the Medfly is established or not completely eradicated. 5 Under GATT and WTO

rules, APHIS is compelled to ensure that quarantines are imposed for sound biological

reasons, rather than for protectionist trade barriers.

A cooperative agreement between APHIS and the State of Florida provides early

detection of Medfly introductions. Timely detection of small Medfly populations greatly

helps management of this pest (Dowell et al. 1999), but recent studies show that early



4 Malavasi et al. (1994) distinguish two types of fly-free areas, a "natural" and a "managed" free area. In a
natural fly-free area, the species naturally does not occur (because of ecology, host preference,
geographical distribution, etc). Managed fly-free areas are production zones from which the target fruit fly
has been eradicated. These areas must be permanently protected by regulatory actions.

5 For instance, cooperative partnership agreements are signed with Mexico and other countries of the
western hemisphere (like Guatemala and Costa Rica) with a view to diminishing their pest problems, thus
reducing the risks of Medfly introductions into the United States.









detection can be difficult (Papadopoulos et al. 2000). The lure for Medfly (called

Trimedlure) is weak and has little ability to attract the flies to the trap (Scribner 1983).

The Jackson trap (the most effective and commonly used Medfly trap) can only catch 1 in

2000 Medflies in the area.

Current national protocol requires 21 traps per square mile for high-risk areas, 12

traps for medium-risk areas, and 3 traps for low-risk areas. 6 About 47,404 traps are

currently placed in 49 counties in Florida. About 54% of the traps are concentrated in

high-risk areas in the following counties: Pinellas, Hillsborough, Orange, Palm Beach,

Dade, and Broward. In FY 2003-04, the total cost of the Medfly detection program was

approximately $7 million, with salaries and supplies accounting for 86.8% (Table 2-3).

Drawing on the 1997/98 Medfly infestation, potential budget changes in case of

emergency (about $ 398,500 per month of emergency) represent significant increases in

employees' overtime hours, travel expenses, and services.

The success of any Medfly-eradication program depends on early-detected

infestation (an infestation where the area under quarantine is 110 square miles or less).

Once detection traps capture one or more Medfly adults, additional traps are placed to

determine whether an outbreak has occurred and/or to limit the outbreak. Drawing on the

potential mobility of Medfly adults, this trapping strategy occurs in an 81-square-mile

area around the fly find, which is divided into a core area (1 square mile for a single fly

capture) and several buffer areas. The whole area is placed under strict quarantine to

prevent the movement of any regulated articles to non-infested areas of the state. All host

fruits on the property and those properties immediately adjacent are stripped promptly

6 Five ML traps and 16 TML traps are placed in high-risk areas, two ML traps and 10 TML traps in
medium-risk areas, and one ML trap and 2 TML traps in low-risk areas.









and disposed of according to APHIS protocols. However, extensive fruit stripping may

stimulate dispersal of gravid females, thereby making eradication more difficult.

Treatment occurs when a Medfly infestation is determined7 to occur. Generally,

eradication procedures combine mechanical, chemical, and biological controls. Table 2-2

summarizes the total eradication costs from 1929 to 1997. Eradication costs have been

drastically reduced since the early 1960s by the development of trapping technology and

because of the intensive use of aircraft during eradication operations. Mass spray

applications of malathion bait have been made possible over heavily populated areas,

often within minutes after discovery of an infestation. This approach has been key to the

success of these eradication programs, providing complete coverage of the epicenter of

Medfly infestation. Ground applications continue to be used for the treatment of soil with

dieldrin to kill emerging adults and larvae entering the soil.

Chemical control is used mostly to reduce Medfly populations to a low level before

the sterile insect technique (SIT) can be used. Currently used to prevent and eradicate,

SIT uses the ability of factory-produced insects to disrupt the normal mating patterns of

wild Medflies. Sterile Medflies mate with their wild counterparts, resulting in the

production of infertile eggs. Preventive sterile release program is more successful when

the number of wild flies introduced is very low. In Florida, about 125,000 flies per week

per square mile are currently released over the high-risk areas8 to provide prevention

control. In cases of emergency, another 400,000 flies per week per square mile would be


7 The deliberative process is based on the following: 1) presence of two flies within a three-mile radius; 2)
presence of one mated female; or 3) presence of larvae or pupae (APHIS 2003a).

8 Area-wide sterile release would include the following criteria: areas where Medflies were detected in the
past, areas in proximity to ports of entry, and/or urban or suburban areas where frequent movement of
imported and exotic Medfly host occurs.









released over each standard block of 100 square miles to support eradication operations.

In 2004, the cost of the Preventive Release Program (PRP) was roughly $3.34 million

(Table 2-3). In an emergency, PRP costs would increase by about $ 82,000 per block.

Sole reliance on SIT for prevention has been unsuccessful, mostly because program

managers cannot maintain the necessary release ratio of sterile to wild fruit flies. For

curative purposes, a minimum ratio of 10:1 (sterile to wild) is required to halt Medfly

population growth and achieve complete eradication (Carey 1982a). Eradication using

SIT is ineffective in production areas because sterile insects are killed by grower

applications of insecticides. Entomologists have also become aware of behavioral

deficiencies of sterile insects versus their wild counterparts. Artificial conditions of the

mass-rearing reduce mating performance, producing Medflies that compete poorly for

females (Jang et al. 1994; Jang 2002). Various approaches to improving the mating

success of SIT flies have been studied.

The future effectiveness of Medfly control programs is inextricably bound to public

concern about pesticide use and potential adverse effects. Ample data on environmental

impacts (fish kills, invertebrate losses, and human health effects) were collected during

the 1997 program in Florida to show that malathion is not safe (APHIS 1999). These

issues raised divergent beliefs about whether the benefits of eradication operations

outweigh the environmental costs, and whether the risks associated with pesticide use are

manageable to acceptable levels. Our study helped elucidate these controversies by

examining the probabilities of risk and related economic consequences associated with a

Medfly outbreak in Florida.









Overview of the Florida Fruit and Vegetable Sector

Florida produces a wide range of fruit and vegetable crops, responding to increased

consumer demand driven by population growth and growing concern over a healthy diet.

In 2001, Florida farmers used approximately 900,000 acres to produce 5.28 billion

pounds of fruits and vegetables. These crops provide revenues of $4 billion to Florida

growers. Tables 2-4 and 2-5 report the acreage, yield, and use of Florida fruit and

vegetable crops for the 1999-2000 season. Oranges, fresh tomatoes, and grapefruits are

the leading crops, accounting for 68.13% of the total acreage allotted to this sector. Other

important crops include bell peppers, cucumbers, tangerines, tangelos, watermelons, and

eggplants.

Florida is known for its citrus fruits, primarily grown in the central and southern

parts of the state. State farmers lead the nation in the production of oranges, grapefruit,

tangerines, and tangelos. About 75% of the nation's oranges are grown in the state: more

than 90% was used to make more than 1.5 billion gallons of juice in 1997. Florida and

Brazil are the major competitors, accounting for over 20 and 60% of world production,

respectively. Florida orange growers suffer particularly from large fluctuations in orange

juice prices, for demand is highly sensitive to consumer income (Spreen 2001).

Florida also produces 77% of the U.S. domestic grapefruit and nearly 47% of the

world supply. This production is grossly split in half: one half is processed, and the other

half is marketed in fresh form. Florida grapefruit growers are highly dependent on export

markets in Europe, Canada, and Japan for fresh grapefruit. In particular, the opening of

the Japanese fresh citrus market has resulted in large increases in shipments of fresh

white seedless grapefruit to Japan (Spreen et al. 1995).









Tangerine production in Florida has increased from 222.3 million pounds in

1986/87 to 522.5 million pounds in 2002/03. An average of 70 percent of the tangerines

produced are sold in the fresh market, with the remainder going to the processing sector

for juice, sections, or other uses (USDA 2003). U.S. domestic production is

supplemented by tangerine imports mainly from Mexico and Spain. While imports of

Mexican tangerines have remained relatively stable over the last six years, imports of

clementines from Spain have increased from 33 million pounds in 1994/95 to 119 million

pounds in 2000/01.

The state also ranks second nationally in the value of its vegetable crops that

account for more than 25% of Florida agriculture sales. Florida winter fresh vegetables

are produced mostly in the southern half of the state where adequate conditions prevail

(VanSickle et al 1994). Although the state has faced a growing array of problems in the

winter fresh vegetable industry, growth in vegetable production for the last two decades

has been mostly related to the use of hybrid cultivars and improved management

practices. In 2002, Florida growers used less than 290,000 acres and received $ 506

million in sales from tomatoes, $ 245 million from green peppers, $ 122 million from

snap beans, $ 107 million from cucumbers, and $ 105 million from sweet corn. Fresh

vegetables generally move from the field to the packing shed for packing, pre-cooling,

and storage before shipment to wholesale markets. Industry sources estimate a total of 60

to 70 packers/shippers throughout the state.

Florida growers held a competitive edge over their traditional Mexican competitors

in the U. S. markets for field-grown vegetables. However, the development of greenhouse

technologies has brought recent changes in U.S. vegetable markets. Florida growers are









now facing growing competition with the largest greenhouse producing areas in Spain,

Italy, France, and Greece. Productivity in European greenhouses is nearly three fold

comparable to Florida field production and product quality is generally higher from

greenhouse versus field-produced vegetables (Cantliffe & VanSickle 2003). This

competition is likely to affect all major vegetable crops grown in Florida, like tomatoes,

peppers, eggplants, cucumbers, muskmelons, and to some degree, watermelons. For

instance, U. S. imports of greenhouse tomatoes have grown rapidly, from 43.9 million

pounds in 1994 to 395.5 million pounds in 2000, including 224 million from Canada,

76.5 million from the EU, and 96 million from Mexico (Cook 2002). As imports increase,

fresh tomato acreage declines in both Florida and California. Concern is growing about

the impacts of importing greenhouse tomatoes on U. S. vegetable industry.

Along these lines, an economic evaluation of the potential damage to the Florida

fruit and vegetable sector from a Medfly infestation must be approached within a

framework of growing competition among the different economic agents from both

within and outside the United States. In addition to the potential losses associated with

yield reduction and increases in pre- and postharvest costs, Florida growers' competitive

position would be further weakened through price adjustments due to losses in export

markets and shipment restrictions to other states. APHIS (1993, 1999) predicts that

countries would react according to their Medfly status and their regulations. While some

countries like Mexico, Argentina, and Chile would require treatment of Medfly hosts

from Florida, others like China, Japan, and the Caribbean nations would prohibit the

importation of all Medlfy hosts, including marginal hosts for a number of years. It is









estimated that Florida would lose over 50% of its export markets if Medfly became

established in Florida.

APHIS and FAO/IAEA have carried out many studies to assess the economic

impact of the potential Medfly damage on fruit and vegetable production, using a partial

budget approach (APHIS 1993, 1999; FAO 1995; IAEA 1995; Enkerlin and Mumford

1996). Losses in producers' revenues are estimated, but none of these studies have given

consideration to price changes related to changes in output and export markets. All

changes in production are measured at current prices and costs to customers are

completely ignored.

Furthermore, findings from risk assessment studies carried out by APHIS (1995b)

are not incorporated into the economic analysis, which could have better supported and

shaped regulatory policy options. Losses in the value of production are grossly estimated

at 5% (APHIS 1993, 1999), under the questionable assumption that eradication, regarded

as a proven technology, can be achieved with complete certainty. The State of Florida

would incur an expected eradication9 cost of $4.8 million each year if eradication were

successful. Costs to producers would range from $32 million to $300 million, depending

on whether eradication is successful.

However, Farnsworth (1985) argued that eradication is a two-event combination

(eradication feasible and eradication not feasible) with a range of probabilities summing

to one. Other studies emphasize the uncertainty of prevention and control program

effectiveness and recommend the use of partial equilibrium models to estimate the


9 The expected cost of eradication per year is calculated by multiplying the probability of an outbreak per
year by an average eradication cost. The calculation of this probability is based upon the history of Medfly
outbreaks in Florida. It is predicted that, given no major changes in APHIS' exclusion activities, this
probability is about 0.2 per year (or once every 5 years).









economic effects of this program under different outbreak scenarios (Regev et al. 1976;

Rendelman and Spinelli 1999; Brown et al. 2002). Approaches taken in these studies

allow for price changes in the commodities concerned under the assumption that linkages

with other similar commodities are small.

Summary

Medfly is one of the most destructive agricultural pests in the world. It has been

introduced 18 times in Florida, leading to 9 infestations. Five of these infestations

occurred in or around Miami Springs, a small upper income housing area located on the

northern edge of the Miami International Airport. Without control measures, much of

Florida would maintain high Medfly populations because of both favorable climate and

availability of preferred host materials throughout the year. Various observations strongly

suggest that Latin American countries could be a high-risk source of Medfly infestations

in Florida.

Florida is a managed Medfly-free area in which the pest has been intermittently

eradicated. Millions of dollars are spent annually on exclusion and detection activities to

prevent Medfly establishment in Florida or at least provide early detection of its

introductions. Such investments are justified because of the importance of the fruit and

vegetable industry in Florida. The State leads the nation in the production of a wide range

of fresh fruit and vegetable crops, thereby responding to an increased consumer demand

driven by population growth and growing concern over a healthy diet. An economic

evaluation of the potential damage from Medfly on the Florida fruit and vegetable sector

must be approached within a framework of growing competition among the different

economic agents from both within and outside the United States. Partial equilibrium

models can be designed to allow for potential price changes in the commodities









concerned and take into account the uncertainty of the detection and eradication program

effectiveness.


Figure 2-1 Female Mediterranean fruit fly


Table 2-1 Major fruit and vegetable crops grown in Florida according to their importance
as Medfly hosts
Heavily infested Occasionally Unknown Lab infestations
infested
Grapefruit, tangelo, Avocado, eggplants, Watermelons, snap Cucumbers,
tangerine, lime, bell ripe tomatoes, beans, squash eggplants
peppers, lemon, strawberries
mango
Sources: Liquido et al. 1991

























Figure 2-2 Male Mediterranean fruit fly


Table 2-2 Costs of Medfly infestations in Florida


Year

1929
1956
1962
1963
1981
1984
1985
1987
1990
1997
Source: Clark et al.


Area of
detection
Orlando
North Miami
Miami
Miami
Tampa
Miami
Miami
Miami
Miami
Tampa
(1992)


Counties
affected
20
28
3
1
1
1
1
1
1
5


Costs ($ millions)
Nominal 1990


7.5
11
1
0.3
1
1
2.2
1.3
1.8
24


56.5
50.6
4.1
1.2
1.4
1.2
2.6
1.5
1.8
20

















































I Infestations in Mani
10
Jan Feb Mr Apr My Jun Jul Aug Sep Cat Nbv ec
Monthly Dstribution
Figure 2-3 Seasonal distribution of Medfly occurrences in Florida


S- Average MninnTerrperatures
..*A .. Average KbanTerrperatures
Average axirumTerperatures
- Uper Kfby Th~reshdd
-a- Lower KIdfly Threshod
--- LUer efIlyOQ rral
- -- Lower Vdfly Cprral




SoLrce: APHS 1992


- Idfly hterceptns from Lasn Amrca
--- MIfly htecepansfromlawa


Souce: APHIS 1992


Jan Feb Mr Apr May in Jul Aug Sep Oct ibv

Figure 2-4 Distribution of M t slr'fi'tfceptions in Florida


.-- ----*----.------- --- ----.--- --- -- --

-A



-A A
S. A.- ... .



A A

mm\
U UU U U











Table 2-3 Baseline and emergency budgets for the Medfly prevention and detection
program


Preventive release program
Baseline annual Emergency
budget ($) budget a ($)


Detection program
Baseline annual Emergency
budget ($) budget ($)


Salaries, benefits 931,402 5,000 5,090,988 112,000
& overtime
Travel expenses 27,300 2,000 56,625 105,000
Transportation & 168,000 9,600 14,500 30,500
shipping
Rents, utilities & 250,376 362,149 9,000
communication
Services & 1,006,544 10,120 180,000 116,000
repairs
Supplies 892,000 55,000 989,000 21,000
Equipment 67,000 241,000 5,000
Total 3,342,622 81,720 6,934,262 398,500
a These costs are estimated for each additional production and release of 400,000 flies per week
and per block of 100 square miles. b These costs are estimated for each month of emergency.
Source: APHIS 2002


Table 2-4 Important fruits: acreage, yield,
Crop Bearing Yield per
acreage acre
(1,000 acres) (boxes)
All oranges 586.9 392
All 101,3 461
grapefruit
All 24 275
tangerines
Tangelos 9.7 222
Limes 0.8 188
Lemons 0.9 94
Avocados 5.9 156


and use (by crop), 2001-02
Use
Fresh Processed
(1,000 boxes)
9,524 220,476
17,380 29,320


4,204

696
125


2,396

1,454
25


920


Source: Florida Agricultural Statistics Service (2003).


Total

230,000
46,700

6,600

2,150
150
85
920






23



Table 2-5 Important vegetables: acreage, yield, and utilization, by crop, 2001-02
Crops Planted Harvested Yield per Production Total Value


Acreage


Acreage


Acre


(acres) cwt
Tomatoes 43,500 43,500 338
Cucumbers 7,500 7,500 386
Bell Peppers 17,250 17,100 320
Squash 12,000 11,700 135
Eggplants 1,800 1,800 257
Watermelons 25,000 23,000 330
Strawberries 6,900 6,900 255
Source: Florida Agricultural Statistics Service (2003).


(1,000 cwt)
14,688
2,893
5,469
1,578
463
7,590
1,760


($ 1,000)
474,284
56,012
170,340
44,543
12,501
62,238
153,472















CHAPTER 3
POLICY FRAMEWORK OF THE SANITARY AND PHYTOSANITARY (SPS)
AGREEMENT

The World Trade Organization Agreement on the Application of Sanitary and
Phytosanitary Measures

Sanitary and phytosanitary (SPS) restrictions are characterized as a subset of trade-

related policies known as technical barriers to trade (TBT10). They include all measures

adopted by a country to protect human, animal, or plant life and health from risks related

to diseases, pests, and disease-carrying or -causing organisms, as well as additives,

contaminants, toxins or disease-causing organisms in food, beverages, or feedstuffs.

Primary method of protection has been the development of quarantine protocols

exhibiting differing degrees of trade restrictions like complete bans, seasonal and/or

geographical bans, postharvest disinfestations procedures (fumigation, cold storage or

others), inspection at points of export and import, and even information remedies (Paul

and Armstrong 1994; Roberts 1998). Unlike most non-tariff barriers, SPS measures are

potentially welfare-increasing, for they may correct market failures resulting from

externalities associated with the movement of agricultural products across national

borders (Roberts et al. 1999; Spreen et al. 2002).

The World Trade Organization (WTO) Agreement on the Application of SPS

measures recognizes all nations' sovereign rights to enforce health standards on imports,

as agricultural trade facilitates the transportation of potentially harmful pests (which can

10 The Technical Trade Barriers (TBT) are in turn characterized as a subset of social regulations,
encompassing all measures adopted by a country to achieve health, safety, quality, and environmental
objectives (Roberts et al. 1999)









cause widespread destruction when carried into a country). Increasing use of SPS

standards and regulations reflects in many cases increased concern for public health and

the environment. However, it is generally acknowledged that SPS measures can also be

used as disguised protection for domestic agriculture. Domestic producer groups with a

vested interest in a particular regulatory outcome are likely to lobby for overly restrictive

measures that limit competition from imports, by exaggerating either the probability of

infestation or the cost impact of infestation (Spreen et al. 2002).

One objective of the Uruguay Round of multilateral trade negotiations, as set out in

the Punta del Este Declaration, is to minimize the adverse effects that SPS regulations

and barriers can have on trade in agriculture. Indeed, government size and power

facilitate the enactment and enforcement of regulatory barriers to trade for producer

protectionism and/or consumer welfare purposes (Thilmany and Barrett 1996). The

Standards Code defined in the Tokyo Round failed to stem disruptions of trade in

agriculture caused by the misuse of technical restrictions (Stanton 1977). The challenge

before the negotiators of the new WTO Agreement on the Application of SPS measures

was to create a set of rules that would strike the proper balance between allowing health

and environmental protection and disallowing mercantilist regulatory protectionism

(Roberts 1998).

Toward this end, new substantive and procedural disciplines were established to

facilitate the decentralized policing of SPS measures. WTO country members are

required (1) to apply the same rules to domestic and imported products and (2) to notify

their trading partners of proposed SPS measures that might affect trade. Trading partners

are therefore given opportunities to comment on a measure before it is adopted. The SPS









Agreement also establishes a dispute-settlement procedure involving several levels of

consultation, such as informal consultations, bilateral negotiations, recourse to the

Committee on Technical Trade Barriers, and convocation before a panel of government

officials.

The cornerstone of the SPS Agreement is found in its Article 5 (which deals with

the issues of risk and level of protection) [Federal Register 2004a]. Countries are granted

the rights to choose their appropriate level of protection (ALP) against imported pests and

diseases, but their regulations must be demonstrably based on an assessment of risk and

clearly related to the control of the risk. Risk assessment typically involves the

identification of the hazard, appraisal of the likelihood and consequences of the

hazardous situation, and specification of the way in which SPS measures would reduce

those consequences (Caswell 2000). The SPS Agreement recommends that risk analysts

develop a strong understanding of the pest biology and potential pathways leading to its

introduction in a new environment (Gray et al. 1998). They often have to make use of

value judgments, while struggling with data gaps, large uncertainties and the need to

extrapolate. In sum, the analysis results in an assessment of the probability of

introduction of a pest or disease (a likelihood model). The different disease outcomes are

treated as inputs into the economic model to estimate.

Net social welfare is the yardstick used by economists to capture the trade and

welfare effects of a regulatory protection model under pest control situations involving

either probabilities of infestation or certainties (Roberts et al. 1999; Roberts 2000). From

the perspective of the importing country, changes in net social welfare resulting from the

imposition of a phytosanitary barrier (Spreen et al. 2002) can be expressed as follows:









W = ACS AW + AW AW AW (3-1)
AW= ACS- +APS- +AEC- +AH (3-1)
ACS APS AEC AH

where W is aggregate social welfare, PS and CS denote producer surplus and consumer

surplus respectively, EC accounts for the enforcement costs associated with the

imposition of the phytosanitary barrier, and H is some index of human health.

Figure 3-1 examines more closely changes in producer surplus resulting from the

reduction in the aggregate supply (AS) due to a shift of the excess supply (ES) in the

exporting country and an increase in the production costs associated with the use of the

additional technology required by the SPS barrier. As a result, domestic prices increase,

and consequently, domestic production (DS) also increases. The model could also

accommodate the prevention program costs and the expected value of government pest

eradication expenditures. These economic considerations allow risk managers to identify

the most effective pest management strategies and to gauge whether a proposed measure

meets the criterion of the SPS Agreement that it be the least trade restrictive (Caswell

2000).

However, the SPS Agreement, as a trade facilitator, does not endorse an explicit

account of the costs and benefits of a policy's effects on producers and consumers.

Rather, it encourages a myopic focus on direct risk-related costs of import protocols

(Roberts 1998). Consideration of producer surplus losses (gains) resulting from lower

(higher) prices would likely be seen to be in violation of the spirit of the SPS Agreement,

because they are costs related to commercial activity, but unrelated to health or

environmental protection (Roberts 2000). Thus, under the SPS Agreement, commercial

considerations might be appropriately factored into a country's choice of its single ALP









(Appropriate Level of Protection), but they should not be used as decision criteria for

individual risk mitigation measures.

Such an approach stands in contrast to the economic paradigm of the U.S.

Executive Branch directives requiring agencies to base their SPS measures on cost-

benefit analysis (CBA) [USDA 1993]. Under the risk assessment paradigm, the role of

economics is relegated to the calculation of quantities of imports to help risk assessors

with their job of calculating the likelihood and consequences of disease or pest

introduction. Primary intent is to reduce the opportunities of practicing mercantilist

regulatory protectionism in favor of domestic producer groups. Therefore, the application

of the SPS Agreement calls for the development of international standards for the

monitoring of CBA-based regulatory policies.

Along these lines, the SPS Agreement seeks primarily to harmonize analytical

frameworks for addressing risk. International standards for phytosanitary measures -

such as the use of a systems approach, the establishment of pest free areas are

designed to achieve international harmonization of SPS measures. The former standard

provides guidelines for the development and evaluation of integrated measures for pest

risk management, while the latter describes the requirements for the establishment and

use of pest-free areas (PFA) as a risk management option for phytosanitary certification

of plants and plant products (FAO 1996, 2002). Assignment of a PFA status is normally

based on verification from specific trapping surveys and, subsequently, appropriate

phytosanitary measures are required to maintain freedom. A more effective pest

management can be achieved by combining two or more independent phytosanitary









measures11, like cultural practices, field treatment, post-harvest disinfestations,

inspection, or other procedures.

The development of these internationally-accepted standards would contribute to

facilitating trade by limiting the use of unjustifiable SPS measures. Recognizing that the

international guidelines may not reflect the preferences and/or needs for externality

mitigation of every nation, the WTO Agreement also allows country members to set a

higher level of protection, which must be based on available scientific evidence. Thus

opportunities are given for the expression of political and cultural differences among the

country members for evaluating threats to people and the environment. A number of

situations are enumerated where national standards may differ from and/or exceed

international standards (Article 2 of the SPS Agreement). Wide discretion is also afforded

to national governments in the determination of situations where international standards

might be inappropriate (Bredahl and Forsythe 1989).

The first two years of implementation of the SPS Agreement saw a broad-based

regulatory review among some WTO members. High-income countries started to

question whether the regulatory measures imposed by their major trading partners were in

compliance with the new Uruguay Round disciplines. The United States, for instance,

identified over 300 questionable market restrictions imposed by 62 countries, which

threatened, constrained, or blocked an estimated $ 5.0 billion of US exports (Roberts

1999). One illustration of these market restrictions, dating back to 1988, relates to the

European Union (EU) ban on the importation of beef from cattle treated with growth-

1 The characteristic of a systems approach is that it requires at least two or more phytosanitary measures that are
independent of each other. With independent measures, the probability of failure is the product of the probabilities of
all the independent measures. A systems approach may also include any number of measures that are dependent on
each other. With dependent measures, the probability of failure is approximately additive, implying that all dependent
measures must fail for the system to fail (FAO 2002).









enhancing hormones. The United States questioned the scientific basis of this regulatory

measure, thereby arguing that the ban had been primarily motivated by the desire of the

EU officials to impose a disguised restriction on the productivity in the beef sector and on

imports from the most competitive beef exporters. In return, the EU rejected these claims,

arguing that the ban addressed public anxieties vis-a-vis the consumption of hormone-

treated beef and that the SPS Agreement contained no disciplines restricting the absolute

level of protection that a member may choose.

The USA/EU dispute exemplifies a very difficult case to resolve, where culture and

consumer preferences affect risk assessment. Citizen and consumers from country

members show different perceptions towards risk associated with SPS issues (Bureau and

Marette 2000; Caswell 2000; Schuh 2000). For instance, the U.S. beef exporters have

been less willing to accept the fact that a large percentage of European consumers may

have a cultural aversion to eating beef produced with growth-enhancing hormones or

antibiotic drugs. The issue is whether international trade regulations should take into

account cultural differences among countries and how to establish a clear dividing line

between what is perceived as a food safety issue and what could be a mere subterfuge for

plain old economic protectionism.

The SPS Agreement provides no guidance and offers little scope for incorporating

cultural analysis into SPS trade issues (Caswell 2000). While recognizing the EU rights

to adopt a precautionary approach on a temporary basis, the Panel argued that the ban per

se was not in conformity with the SPS Agreement and that there was no sufficient

scientific evidence of dangerous effects on human health associated with the

consumption of meat products treated with hormones. The hormone case provides some









preliminary indication that the WTO Dispute Settlement Body is more likely to base its

verdicts on a rational relationship between objective scientific assessments and the policy

choices made by governments.

Another interesting case was the Mexico complaint about the U. S. ban imposed

since 1914 on the importation of Mexican Hass avocados, on the grounds that the fruit

was a host of various fruit flies and that its importation could also lead to the importation

of quarantine pests of concern. Mexico conducted in situ experiments to demonstrate that

the Hass avocado fruits attached to the tree are biochemically and ecologically resistant

to infestations12 with A. serpentina, A. ludens, and A. striata under natural field

conditions and there was only a minimal risk associated with importing this crop

(Hoeflich 2000). While scientific arguments and evidence may have been a necessary

condition for trade liberalization, they were certainly not a sufficient condition. For over

80 years, no trade protocol could be reached prior to NAFTA between Mexico and the

United States. The policy process was captured by Californian producers benefiting from

the monopoly over the U. S. avocado market. However, the approval of NAFTA

contributed to the end of this monopoly situation, providing space for negotiations and an

opportunity for science to take part in the decision-making process (USDA 2001, Federal

Register 2004b; Hoeflich 2000).

The above cases show the complexity of the issues involved in the application of

the SPS Agreement. Politics, economics and culture often play prominent roles in the

choices of regulatory measures to address risk management problems (Caswell 2000). In

particular, economics may play a strong role in measuring costs and benefits of SPS

12 It is important to underline that, under laboratory and field force conditions, Hass avocado fruits are a
good host of A. ludens, an average host of A. serpentina, and a poor host of A. striata.









measures and ranking them on the basis of how much they will improve (or undermine)

the social well-being. Ultimately, policy choices emerge from the interaction of the

demand for measures by various domestic interest groups (including producers,

consumers, and processing industries) and the supply of barriers available to

policymakers.

Few people disagree that the SPS Agreement has contributed to improve

international trading relationships among country members and to restore the rule of law

in this fractious area (Thornsbury 2002). Transparency of the WTO members has been

enhanced. Procedures and rules for dispute settlement have been established and followed

by WTO members. More importantly, regionalization efforts have fostered the alignment

of national regulations with international standards for phytosanitary measures and the

compliance with the obligations of the SPS Agreement. The risk paradigm of the SPS

Agreement has also reduced the degrees of freedom for the disingenuous use of SPS

measures to restrict imports in response to narrow interest group pressures (Roberts

2000). However, the SPS Agreement still faces the ongoing challenge of how to lay the

foundations of the SPS measures on scientific restrictions, while allowing for more

flexibility in the use of economic and cultural considerations. Empirical evidence

concerning the extent of questionable technical measures in international agricultural

trade is very difficult to assess, because of unavailability of comprehensive data sources

(Thornsbury 2002). Furthermore, despite several attempts over the years to resolve the

USA/EU dispute over the hormone-treated beef, the ban is still in effect (Pantin et al.

2004), leaving substantial room for questioning the capacity of the SPS Committee and

the Appellate Body of the SPS Agreement to enforce authorized sanctions.









Phytosanitary Protocols for the International Movement of Fresh Fruits and
Vegetables in Fruit Fly Free Areas

Tephritid fruit flies are among the most destructive agricultural pests threatening

the sustainability of fruit and vegetable industries in many parts of the world. Millions of

dollars are spent annually in fly-free countries to enforce quarantine restrictions with a

view to preventing the introduction of exotic fruit fly pests and maintaining their fly-free

area status. On the other hand, fruit producers in fly-infested countries invest in

expensive post-harvest facilities and treatments in return for gaining and/or maintaining

access to export markets. The application of the SPS Agreement has materialized

essentially into the establishment of phytosanitary protocols serving as mechanisms for

facilitating trade and transferring the cost of enforcement of the SPS restrictions onto the

exporting countries (Spreen et al. 2002).

The agreed-upon phytosanitary protocols between importing and exporting

countries are inspired by the fly-free production model that often uses the concept of the

systems approach. A fly-free field may refer to small areas such as a farm, an orchard, or

a group of properties. Its delimitation requires a thorough knowledge of the biological

profile of the pest concerned (FAO 1996) and a risk management strategy aimed at

maintaining pest freedom from specific parcels and production areas (Malavasi et al.

1994).

However, attention is now moving from a fly-free field to fly-free zone approach

(Vijaysegaran 1994; IAEA 1995). The fly-free zone concept encompasses an entire

geographic or political entity in which permanent eradication efforts such as intensive

use of inspection sites at ports / airports and road stations, massive sterile fly barrier,

trapping surveys, and regular sprays against fruit flies result in freedom of all fruit fly









species of quarantine importance (Hendrichs et al. 1983; Hendrichs 1996). Pest

population suppression over large geographical areas would undoubtedly provide better

control and benefit a large number of growers. Apart from the success of an eradication

project in Japan, eradication of a fruit fly species by sterile insect technique (SIT) and/or

other means has not been reported in other Asian countries (Kawasaki 1991). Major

concerns would be the cost of the sterile insect technique and the increasing risk of re-

infestation.

Table 3-1 presents three different formats of phytosanitary protocols agreed upon

between importing and exporting countries. Phytosanitary requirements for international

movement of fresh fruits and vegetables vary according to pest status and distribution,

number of other key pests, isolation, geographical location, technical level of fruit

production, economic value of the crop, and changes in target markets (Hendrichs 1996).

Format I (see column lof Table 3-1) accounts for the basic phytosanitary protocol

providing for detection surveys in and around the designated production areas to assure

pest freedom against fruit flies of quarantine importance in the field. This quarantine

restriction is the centerpiece of the phytosanitary protocols governing (1) the shipment of

fresh citrus fruit from the United States (Florida, California, Arizona and Texas) to some

specific ports of entry in China and (2) the importation of watermelon, squash, cucumber,

and oriental melon from the Republic of Korea to the United States between December

1st and April 30 (Federal Register 2003a). Minimum trap densities based on research

findings are specified in these protocols, as are the frequency of trap servicing and the

suitable trap locations for the surveillance of flies of quarantine importance. The National









Plant Protection Organization (NPPO13) of the exporting country must provide

monitoring reports on trap surveys and relevant information on groves, shippers/packers,

and storage facilities for annual review of the protocol. Pre-inspection visits by the NPPO

of the importing country to the exporting country are provided for to conduct a review of

the certification procedures. Travel expenses (i.e. transportation, lodging, and a per diem

allowance) for all trips during the first two years of the agreement to designated groves,

shippers/packers, and storage facilities are borne by the exporting country.

Certification procedures are based on negative trapping, providing scientific

evidence that the concerned pests do not occur in designated areas and that this condition

has been officially maintained. Exported fresh commodities must be transported under

closed conveyance and kept separated from packed commodities from non-designated

areas. The NPPO of the exporting country shall perform a strict inspection of export

shipments and ensure that exported products are free of quarantine pests. No post-harvest

treatment is required in the protocol. In the event of detection of a live fly on arrival, the

importing country shall immediately notify the exporting country about suspending the

importation of fresh commodities from the designated grower or grove, shipper/packer

and storage facility, and the shipment shall be returned, re-exported or destroyed. The

suspension of the phytosanitary certificate shall be maintained until the relevant cause is

identified and appropriate corrective actions are taken.

Format II (see column 2 of Table 3-1) differs from Format I in that the former

involves a more stringent and complex certification process, reflecting the dominant


13 Each member country is required to form a National Plan Protection Organization (NPPO), which
accounts for the government's official service to discharge the functions specified by the International Plant
Protection Convention (IPPC), deposited in 1951 with FAO in Rome, and subsequently amended (NAPPO
1998; Federal Register 2002).









position of countries like Japan and South Korea (which are regarded as lucrative markets

for fresh products) [Vijaysegaran 1994; Simpson 1993; Reiherd 1992]. Cases of this

format are exemplified in the phytosanitary protocols for the shipment of Florida citrus

fruit to Japan and South Korea (APHIS 2003b). The cornerstone of this protocol format is

the requirement that the designated production area be surrounded by a buffer zone of 1.5

mile, which should not contain any preferred host plants of the pest concerned. Where a

preferred host plant does occur in the buffer zone, ground or aerial bait spray shall be

applied at 7 to 10-day intervals beginning 7 days prior to harvest and continuing until the

end of harvest. Also, the minimum size of the designated area is required to be 300 acres.

Such a phytosanitary restriction is designed to stimulate the development, among fruit

growers, of a concerted fly population management strategy over a significantly large

area, so that the potential number of flies moving into orchards from neighboring

orchards is largely reduced. (Hendrichs 1996).

Early season certification criteria for grapefruit shipped during August 1st to

December 20 are less restrictive. Such procedures are based on the proven resistance of

early season citrus fruit to Caribfly infestation, which is regarded as a fly-free period

(Simpson 1993). Nonetheless, the standard certification procedures appear to be very

stringent, requiring an early surveillance of the Caribbean flies (trap servicing beginning

30 days prior to harvest) and a high trap density (15 traps per square mile).

Format II also allows for a second certification procedure referred to as "bait

spray" with the following requirements: 1) the minimum size of the designated area must

be 40 acres (16 hectares) surrounded by a 300-foot wide buffer zone, 2) the buffer zone

must be free of preferred host plants, 3) the designated area must be at least /2 mile from









areas where numerous host plants are present, 4) traps are located in the designated area

and 300 feet buffer zone at the density of 15 traps per square mile with trap servicing

beginning 30 days prior to harvest, and 5) aerial bait sprays are applied at the beginning

of the harvest period, consisting of a mixture of 2.4 ounces of 91 percent malathion and

9.6 ounces of protein hydrolyzate bait per acre. Where the designated area is located

within 12 mile of numerous preferred hosts, aerial bait sprays shall be applied earlier,

beginning 28 days before harvest at 7 to 10-day intervals until the end of harvest.

In the event of a Medfly outbreak, countries like Japan, South Korea, and South

Africa take a more drastic approach by prohibiting the importation of all Medfly hosts

including marginal hosts for a number of years (USDA1999). Japan, in particular, will

not issue any phytosanitary certificate for any Medfly host commodity from the infested

country, even if a quarantine treatment approved by the country of origin is applied.

Such phytosanitary restrictions are justified by the seriousness of the Medfly attacking

over 300 fruit and vegetable commodities. Nevertheless, other countries like Mexico,

Argentina, and New Zealand adopt a more flexible approach based on the fly-free

production model. The phytosanitary regulations include the establishment of a

quarantine area around a radius of 17 miles from the epicenter of the outbreak, intensive

trapping surveys, and the treatment of Medfly hosts according to agreed-upon treatment

schedules.

The third and last protocol format (column 3 of Table 1) provides an illustration of

the peculiar use of a systems approach based on the notion of low pest prevalence. This

strategy reflects the delicate position of the United States as both net exporter of some

fresh agricultural commodities and net importer of others switching from its long-









standing practice of only recognizing entire countries as "free" or "not free" of a

particular disease to a "regionalization regulation" (Ahl and Acree 1993; Federal Register

2003a). Another significant factor is the fact that the U.S. territory is not free of fruit

flies. By allowing the importation of fresh fruit commodities in the mainland from

Hawaii where the Mediterranean fruit fly has been established, APHIS finds itself forced

to abide by the principle of national treatment and, therefore, to apply the same rules to

imported products. These considerations support the tendency towards the adoption of

risk management strategies combining various pre- and post-harvest actions and

treatments sequentially so that they can provide acceptable statistical probabilities of

quarantine security (Armstrong 1991). Two cases fall within this "systems approach"

framework:

* the importation of clementines, mandarins, and tangerines from Chile (where the
Mediterranean fruit fly and other quarantine pests of concern are known to occur)
under a series of complementary phytosanitary measures. The requirements include
a test program of certification of low prevalence, a post-harvest processing, and
phytosanitary inspection (Federal Register 2004a)

* the importation of fresh Hass variety avocados from Mexico into the United States,
using pest risk mitigation strategies combining two major tactics: (1) limiting the
geographical distribution of avocados to 19 States and the District of Columbia
within the Unites States and (2) allowing a 4-month shipping season each year
(Federal Register 2004b).

The efficacy of the systems approach is based on the combination of

complementary measures acting independently to ensure an appropriate level of

phytosanitary protection (FAO 2002). For instance, the phytosanitary restrictions act in a

fail-safe manner, so that redundant safeguards are built into the process. If one mitigation

measure is not completely successful, the other will ensure that the risk of pest

introduction is insignificant. Furthermore, any pest detection or irregularity would result









in immediate actions to eliminate the pest risk, including partial cancellation of

phytosanitary certificates or total prohibition of imports.

It is worth noting that the phytosanitary protocols between exporting and importing

countries have been products of continuous and intense negotiations. Disputes over SPS

standards go beyond the evidence of scientific phytosanitary restrictions. The most

important issue is often examination of the implications these scientific protocols would

have on the "safety" of domestic producers. In fact, policymakers tend to place greater

weight on producer rather than consumer welfare. Cases dealing with food safety issues

are even much more difficult to resolve as producer groups could easily mobilize

consumers to back their claims for more regulatory protectionism. What really makes

possible the establishment of these phytosanitary protocols is the opportunity provided by

the SPS Committee to air grievances over unjustified measures when bilateral technical

exchanges reach an impasse. Despite fierce opposition from domestic interest groups,

regulatory agencies are often forced to arrive at some acceptable arrangement with their

trading partners by fear of retaliatory measures and / or unnecessary reciprocal

phytosanitary barriers on domestic exports.

Regulated Post-harvest Treatments and Procedures for the Quarantine Control of
Fruit Flies

This section turns to the analysis of the regulated postharvest treatments that are

often required for allowing unrestricted movement of tephritid fruit fly host commodities

in domestic or international commerce. No quarantine treatment is universally applicable

to all products or all quarantine pests (Mitchell and Kader 1985). Each treatment has

some inherent problems and limitations. Agricultural research and regulatory agencies

have developed various tests to evaluate potential quarantine treatments against fresh









commodities infested with different fruit fly life stages. Two major tests confirmatory

test for quarantine security and small test for efficacy -are commonly used for the

approval of a postharvest treatment schedule. Quarantine security refers to the level of

confidence that the quarantine treatment will disinfect quarantine pests from the host

commodities (Armstrong and Couey 1989). Specifically, it involves compliance with the

phytosanitary requirements defined by the NPPO of the importing country to ensure that

the quarantine pest of concern cannot become established in any geographical area where

it does not already exist.

One problem in quarantine treatment development is the lack of availability of

appropriate statistical criteria that can guarantee quarantine security (Couey 1983).

Although the probit 914 mortality at the 95% confidence level remains the quarantine

security statistics most commonly used in the post-harvest technology literature, it does

not properly indicate the risk of a pest species spreading into non-infested areas (Landolt

et al. 1984). Alternatives to the use of probit 9 mortality have been proposed, but not

fully developed, such as the probability of mated pairs of quarantine pests in a shipment

of a host commodity (Landolt et al. 1984), the host/pest relationships and natural

infestation rates (Couey and Chew 1986).

Commercially applied treatments are also monitored for efficacy. The term

"efficacy" describes a quarantine treatment that adequately disinfests pest organisms at

the required level of quarantine security without adversely affecting the commodity

14 The probit 9 statistics at the 95% confidence infers that no more than 3 individuals from a population of
100,000 will survive a quarantine treatment, which is a mortality rate of 99,997%. The treated population
must equal 100,000 or more target organisms in three or more tests with no survivors. So, the treated
populations are derived for the survivors of the control population by the formula: (A/B)C= estimated
treated population, where A = the population of survivors from the controls, B=the control weight
(untreated commodity), and C= the treated commodity weight (Spitler and Couey 1983; Armstrong et al.
1984).









(Armstrong and Couey 1989). While APHIS as the regulatory agency is not liable for

damages caused by the quarantine treatment, other Federal and State agencies like the

Food and Drug Administration (FDA), the Environmental Protection Agency have

primary responsibility for ensuring that approved control treatments are not harmful to

the commodity, workers, the consumer, or the environment (Federal Register 2002).

The number of approved fumigants and fumigation schedules has drastically

diminished over the last two decades, due to environmental problems, health concerns,

and lack of research. Following the cancellation of ethylene dibromide's registration by

the U.S. Environmental Protection Agency (Federal Register 1984), the methyl bromide

(MB) has remained the most widely used fumigant for horticultural commodities because

of low cost, ease of application, relative safe usage, rapid dispersion throughout the

fumigation chambers, and rapid penetration into the commodity (Mitchell and Kader

1985; Armstrong and Couey 1989).

Approved MB concentrations, durations, and temperatures depend upon the

commodity and the fruit fly species to be controlled. Strawberries infested with Ceratitis

capitata can be safely transported from the quarantine area after MB fumigation

schedules at 150C or above with 48 g/m3 for 3 h (Armstrong et al. 1984). MB dosages

required to achieve 99.9968% kill (probit 9) of Anastrepha suspense infestations of

grapefruits at 210-240C are proven to be 40 mg/liter for a 3-carton load and 56 mg/liter

for a 12-carton load fumigated in 0.8 m3 chambers (Table 3-2, column 1). However, final

acceptance requires tolerance and residue tests under the variety of conditions

encountered in processing and shipment (Benschoter 1979). Following successful

cucumber fumigation schedules at 190C or above with 32 g/m3 for 4 h, the results of









phytotoxicity tests revealed very minor phytotoxic effects that cannot affect the

marketability of the product (Armstrong and Garcia 1985).

However, the safety of MB fumigation has been called into question because of

reports of carcinogenicity on laboratory animals. MB residue levels in papaya, tomatoes,

bell peppers, bananas, and eggplants treated against Ceratitis capitata are showed to range

from 1.7 to 42 ppm, depending on the method of fumigation, the commodity, and the

method of storage (Baker 1939; Seo et al, 1970, 1971). Therefore, alternate treatments

may be needed if agriculture and food supplies are to be protected. One potential

candidate fumigant that is receiving attention is phosphine, which offers some advantages

in terms of rapid diffusion throughout the load without a re-circulating system, rapid

dissipation of very low residues, and tolerance to fumigation by avocados, bananas,

tomatoes, and bell peppers (which are injured by fumigation with methyl bromide) [Seo

et al. 1979]. The main disadvantage of phosphine fumigation is, however, the long

treatment time (2 to 4 days instead of 2 hours), which makes it less promising for

perishable commodities. Intensive refrigerated fumigation facilities would be needed at a

very high cost to keep perishable commodities under refrigeration during that long

exposure time and, consequently, avoid unacceptable deterioration (Mitchell and Kader

1985).

A combination of MB fumigation and cold treatment seems to be a more

economically feasible alternative. This technique was used in California to disinfect fresh

stone fruits during a Medfly outbreak. Treatment schedules listed in APHIS' quarantine

treatment manual specify fumigation with 32 g/m3 for 2, 2.5, or 3 h, followed by cold

storage for a minimum of 3 days to a maximum of 11 days at temperatures ranging from









a low of 0.55C to 13.330C (Code of Federal Regulations 2003; Spitler and Couey 1983).

The cold portion of the treatment is often conducted aboard ship, while in transit from

infested areas. Shipments of cold treated fruits are certified upon compliance with strict

requirements for temperature monitoring in cold storage facilities.

Cold treatments alone have been accepted by USDA-APHIS as quarantine

treatments for 14 fruits and vegetables from 48 countries subject to infestations by

Ceratitis capitata, Anastrepha ludens, and other species of Anastrepha (APHIS 1976).

Treatment schedules are very severe (Table 3-2, column 3), involving fruit exposure to

temperatures below 5C for extended durations (from 10 to 16 days). The highest

temperature listed in the Plant Protection Quarantine Manual (APHIS 1976) is 2.2C.

Probit analyses also predicted that 16-20 days would be required for fruit held at

temperatures ranging from 2.8 to 6.6C (Burditt and Balock 1985). Such quarantine

requirements severely limit the use of cold treatments, since most tropical fruits are

damaged by extended storage below 10C. Under cold treatment conditions, a certain

percentage of fruit always exhibits chilling injury symptoms. In particular, susceptibility

of grapefruit to low temperatures is proved to vary with season and fruit location on the

tree. Fruit on the outer canopy are more susceptible to chilling injury than those harvested

from the interior of the tree. Reducing losses in grapefruit shipments involves strict

compliance with a series of basic requirements such as: avoiding prolonged degreening,

ensuring proper application of fungicides, using stable water wax, and providing proper

warming before fumigation and after completion of cold treatment (Ismail et al. 1986).

Recent studies also show that preconditioning at warm temperatures increase fruit

tolerance to cold treatments (Mitchell and Kader 1985). For instance, papayas infested









with eggs and larvae ofD. dorsalis, D. cucurbitae, and C. capitata are successfully

treated after immersion in 490C water and exposure to 80 or 90C for 10 days (Couey et al.

1984). Nevertheless, the use of combined heat and cold treatments requires knowledge of

the biological profile of the insect, the lethal effects of the standard heat treatment for

decay control, cold treatment, and appropriate quality control procedures. Papaya fruits

are required to be picked between colorbreak and one-quarter ripe. Unripe papayas are

unlikely to be infested with fruit flies, as containing sufficient quantities ofbenzyl-

isothiocyanate to deter oviposition and to reduce survival of eggs and larvae whenever

ovipostion occurs (Seo et al. 1982; Seo and Tang 1982).

Hot-water treatment also may be used alone to disinfect fruit. Exposure to hot

water at 450C for 20 min or at 550C for less than one min would destroy all immature

stages ofD. dorsalis, D. cucurbitae or C. capitata (Armstrong 1982), but it would easily

damage many fresh commodities (Table 3-2, column 2). Sinclair and Lingren (1955)

found that navel oranges, lemons, and avocados were very easily damaged by the

standard vapor heat treatment. One in six unripe papayas is damaged after exposure to

microwave until a central temperature of 45C, followed by a double-dip in 48.7C and

24C during 20 min respectively (Hayes et al. 1984).

Cases of tolerance to heat treatments are reported to very few commodities.

Damage to valencia oranges and grapefruit treated with standard heat vapor15 or the

quick run-up can be avoided (or reduced) if the fruits are hydrocooled after treatment.

Satisfactory results have also been obtained with mangoes immersed in 50o-55 C water

15 The vapor heat treatment consists of gradually warming infested fruit for several hours (approach time),
then increasing the fruit temperature to 430C and holding that temperature for 8 h (holding time). The quick
run-up requires a short pre-heat period to a specified temperature, then a gradual warming to 47C, similar
to the approach time in the standard vapor heat treatment (Baker 1952; Balock and Kozuma 1954).









for 15 minutes (Sharp and Spalding 1984) or vapor heated at 460C for 160, 220 and 280

minutes (Mitcham and McDonald 1993). Concern is, however, growing about the

feasibility of heat treatment due to the relatively high cost of application. Vapor heat cost

per kg of fruit is estimated at $ 0.35 against $ 0.027 and 0.26 for methyl bromide

fumigation and irradiation respectively (Federal Register 2002, 2003b; Pszczola 1992).

Other alternatives to fumigation for insect quarantine treatments have also been

explored, such as the controlled atmosphere and irradiation treatments. Forced air

treatments (Table 3-2, column 4) consist of exposing fresh horticultural commodities to

moderately low levels of 02 (<2 2%) and /or high levels of CO2 (> 50%). Although

modified atmosphere is cost competitive to chemical fumigation (Soderstorm et al. 1984)

and leaves no chemical residues on the fruit, the possibilities for commercial use are

much less certain. Most fresh fruits and vegetables cannot tolerate such extreme

atmospheres for prolonged storage periods (Armstrong and Couey 1989; Smilanick and

Fouse 1989; Yahia et al. 1991; Ke and Kader 1992). Keitt mango is, however, reported

to be very tolerant to insecticidal 02 and/or CO2 atmospheres for up to 5 days (Yahia

1993). Further research work is needed to determine the level of tolerance of other mango

cultivars and the level of mortality of important quarantine insects.

Radiation is by far the most publicized quarantine treatment, mainly because of

renewed interest in this treatment as a potential alternative to the use of chemicals.

Treatment schedules (Table3-2, column 5) involve exposing the product to a radiation

source for a time period sufficient for it to absorb a required dose level of gamma or X

rays. Successful use of this procedure is based on the determination that an undesirable

organism will be inactivated at a dose level that is tolerated by the host commodity









(Mitchell and Kader 1985). For safety concerns, the Food and Drug Administration

(FDA) establishes a dose limit of one kilogray (Federal Register 2002), which is by far

less than all irradiation doses contained in APHIS's rule. This quarantine policy is

consistent with study findings indicating that most insects are sterilized by doses below

0.75 kGy.

Unlike the heat, cold, and fumigation treatments that generally kill the pest, what

really matters in irradiation is the treatment dose required to break the life cycle of the

insect, so that it cannot become established in a new uninfested area (Rigney 1989; Baker

1939). Treatment efficacy is based on prevention of adult emergence. The killing of fruits

flies in order to minimize fruit damage from feeding insects is of secondary importance.

While the criterion of quarantine security may be fully achieved with a low irradiation

dose, the marketability of the fruit is likely to decrease, due to the potential presence of

large numbers of fruit fly eggs and/or larvae in the fruit. Achieving a lethal effect on fruit

fly eggs or larvae would require doses in excess of 1 kGy, which would be damaging to

many fruits.

Research findings indicate that irradiation provides acceptable quarantine treatment

for various fresh commodities infested with fruit flies. For instance, probit 9 quarantine

security is reached with doses below 150 Gy for grapefruit and mangoes infested with

Caribbean fruit flies, causing acceptable levels of phytotoxicity to the fruit (von

Windeguth 1986, 1987). Phytotoxicity tests at doses ranging from 50 to 1500 Gy

indicated that no observable damage occurred at levels between 50 and 500 Gy when

Arkin carambolas were irradiated at 250C and then held for 9 days at the same

temperature (Gould and von Winderguth 1991). Furthermore, irradiation of foods offers









consumers many advantages the most important of which could be safe transport of

produce from insect quarantine areas and replacement of less safe chemical fumigants

(Bruhn et al. 1987; Schutz et al. 1989).

However, there are a number of considerations that dictate extreme caution in

projecting radiation as an alternative to fumigation for insect quarantine treatment. First,

consumer acceptance is critical to the application of irradiation and the realization of the

advantages it offers. Consumers have expressed growing concerns about the safety of

irradiated foods, due to negative advertisements sponsored by opposition groups.

Although consumers show a higher level of concern for chemical sprays and pesticide

residues than for food irradiation, it is still hard to sell irradiation to the American public

(Schutz et al. 1989; Bruhn et al. 1986). Secondly, most fresh fruits offer no promise for

commercial irradiation because either alternative procedures are cheaper and more

effective or radiation injury is excessive. Irradiation is unlikely to compete at present cost

with chemical inhibitors or to substitute for refrigeration. In the future, feasibility of

irradiation will depend on whether the phase-out of methyl bromide as a soil fumigant

results in an increase in its unit cost of production (Federal Register 2002).

Concluding Remarks

This literature review on the SPS Agreement emphasized the new substantive and

procedural disciplines established to achieve harmonization of SPS regulations and

minimize their adverse effects on agricultural trade. Countries are granted the rights to

choose their appropriate level of protection (ALP) against imported pests and diseases,

but their regulations must be supported by scientific criteria. Thus, the SPS Agreement

encourages the use of the least trade restrictive measures, thereby making provision for









the inclusion of the costs of prevention and control programs as the major factor in

regulatory decisions.

The application of the SPS Agreement has materialized essentially into the

establishment of agreed-upon phytosanitary protocols between exporting/importing

countries. These protocols serve primarily as mechanisms for facilitating trade and

transferring the cost of enforcement of SPS restrictions to the exporting countries. In the

event of a Medfly outbreak in Florida, it is anticipated that the reactions of export

markets vis-a-vis Florida fresh fruit and vegetable commodities would greatly differ,

ranging from additional certification, quarantine treatment, to prohibition. Florida

growers would be unlikely to reach a trading arrangement with countries like Japan and

South Korea, which are regarded as lucrative markets for fresh citrus. Negotiations would

be very tense as to whether their total prohibition would affect Florida's total fruit and

vegetable production or the portion of production localized in the quarantine area.

Other countries would allow for the importation of fresh fruit and vegetables from

Florida under very strict phytosanitary restrictions. Given the seriousness of the Medfly

attacking over 200 species of fruits and vegetables, the phytosanitary restrictions are

expected to be more stringent that those contained in the Caribbean fruit fly-zone

certification protocol, including intensive trapping surveys, establishment of quarantine

areas around 17 miles from the outbreak epicenter, and regulated postharvest treatments.

Treatment schedules would vary across commodities, including fumigation (strawberries

and cucumbers), cold storage (fresh citrus), combination fumigation and refrigeration

treatments (avocados), and vapor heat treatment (mangoes). Florida growers would be

reluctant to applied some of APHIS-approved quarantine treatments to commodities -










such as bell peppers, eggplants, tomatoes because either they are uneconomical or

cause excessive fruit injury. More research work needs to be done on the technical and

economic feasibility of the regulated postharvest treatments.




Price


ES' DS



ES ., A



%%% AS









D


Source: Spreen et al. 2002 Quantity

Figure 3-1 Inward shift in import supply resulting from the imposition of an SPS barrier














Table 3-1 Formats for phytosanitary protocols for the international movement of fresh fruit and vegetable commodities


Format I (column 1)

1. Properties Sites
*Must be free of fruit fly, monitored by MCPhail
traps. Groves, packing houses, and storage facilities
must also be certified

2. Certification Procedures
*Based on negative trapping
-Annual notifications of the exporting country to the
importing about the trapping surveys in the areas of
production

*Phytosanitary inspection
-Pre-inspection visits (one per year) by the authorities
of the importing country to supervise the process of
certification at the expense of the exporting country.

-Citrus fruit must be free of live insects and mites of
quarantine concern and identified with labels
indicating place of production, grower, shipper, and
storage facility

-Citrus must be transported in sealed container and
kept separated from unapproved shipments

-Exports restricted to specific ports of entry in the
importing country


Format II ( column 2)
1. Production Sites
*Must be free of fruit fly, located at specific
distances from preferred hosts and surrounded by a
buffer zone > 1.15 mile. Packing facilities must be
located > 3 mile from infested area
2. Certification Procedures
*Based on negative trapping
-For early season, minimum size of production area is
300 acres. Trap surveys with 2 traps per sq. mile. If
buffer zone has preferred hosts, ground or aerial
applications at 7-10 day intervals, beginning 7 days
prior to harvest

-For standard season, minimum size of prod. Area is
300 acres. Trap surveys with 15 traps per sq. mile. If
buffer zone has preferred hosts, ground or aerial
applications at 7-10 day intervals, beginning 7 days
prior to harvest

*Based on bait spray procedures
-For early season, minimum size of production area is
40 acres. Trap surveys with 15 traps per sq. mile. If
buffer zone, has preferred hosts, ground or aerial
applications at 7-10 day intervals beginning 7 days
prior to harvest.

-For standard season, minimum size of prod. Area is
40 acres, must be at least 12 mile from preferred host.
Trap surveys with 15 traps per sq. mile. Aerial
applications at 7-10 day intervals, beginning 28-30
days prior to harvest


Format III ( column 3)
1. Productions sites
*May be a pest free area or low prevalence
area

2. Certification Procedures
*Based on the use of systems approach
combining two or several independent
phytosanitary measures

*Case 1 combining three sequential measures
-Low prevalence Certification
Test of low prevalence based on a random
sample of fruit undergoing a washing process.
If no pests are found along the process, the
production site is certified
-Post-harvest treatment
Products treated in accordance with the agreed
treatment manual. Shipments must be
accompanied by documentation indicating the
type of treatment adopted
-Phytosanitary Inspection
Based on biometric sampling with an
acceptance level of zeros infested units.

*Case 2 combining two major tactics
-A geographical ban: limiting the importation
to some specific geographic areas
-A seasonal ban: limiting the importation
during some specific periods of the year














Table 3-1 Continued
Format I (column 1)

3. Measures to be taken if detection is found
*Immediate notification to the exporting country and
establishment of a regulated quarantine area with a
radius of 17 miles from the center of the pest
outbreak

*If a Mexican fly is found, the shipment must be
treated but no citrus fruit from the affected grove will
be exported


4. Exceptional measures in case of a Medfly
infestation

*Increase the trapping density to one trap per square
km and immediate suspension of citrus fruit from the
regulated quarantine area with a radius of 17 miles
from the center of the outbreak. Shipment of fruit
transited through this quarantine area must be in
sealed container


Format II ( column 2)

3. Measures to be taken if detection is found during
inspection
*Immediate notification to the importing country if
one fly is found as a result of the trap survey

*Withdrawal of the production area from the protocol
season if three files are found. Reinstatement is
conditioned to intense spraying after investigation

4. Exceptional measures in case of a Medfly
infestation

*Fruit & vegetables considered to be hosts of Medfly
fly are prohibited entry from countries to be infested
with this pest. Such commodities remain prohibited
regardless of whether they have met the entry
requirements of any other country


Format III ( column 3)

3. Measure to be taken if detection is found

*If one live fruit fly is intercepted along the
inspection process, the consignment will be
reshipped or destroyed and a prohibition will be
placed on further imports of the host material
until corrective action is undertaken

4. Exceptional measures in case of a Medfly
infestation

* Increase the trapping density to one trap per
square km; establishment of a quarantine area
found a radius of 17 miles from the center of the
outbreak; treatment of all regulated articles
according to approved schedules. Shipment of
fruit transited through this quarantine area must
be in sealed container


Source: http://excerpt.ceris.purdue.edu/doc/ctrylist.html














Table 3-2 Regulated postharvest treatments, advantages, limitations, and alternatives under consideration


Fumigation


*MB fumigation at 210 C
or above at NAP with 32-
48g/cubic meter MB for
two to four hours.

*MB fumigation at NAP
with 32g/cubic metric MB
for 3 hours at 21C for
citrus infested with
Medfly.

*MB fumigation at 27C
and NAP for 2 hours for
citrus infested with
Caribfly and Mexican
fruit fly.

*MB is the preferred
fumigant for horticultural
commodities be-cause of
low cost, ease of
application, and relative
safe usage.


Heat Treatment


*Vapor-heat treatment of
citrus at 430C for 8 hrs.

*Hot-water preheat im -
mersion of papayas,
bananas, & mangoes at
420C for 20 min to di-
sinfect against Medfly,
Melon, Oriental &
Caribflies.


*Papayas & mangoes are
resistant to heat damage.

*Heat treatments have the
merit of effective
fungicidal & insecticidal
action and no chemical
residue.


Cold Treatment


*Cold treatments at
temperatures ranging
from 0OC to 2.20C for a
minimum of 10 days and
a maximum of 16 days
respectively. Such
treatments are used
against Medly, Caribfly
and other quarantine fruit
flies


*Cold treatments are very
practical when used
during transit from
producing areas to distant
markets


Controlled Atmospheres

*Exposure of fresh
commodities to low levels
of oxygen and high levels
of carbon during some
period of time


*Cost competitive to
chemicals

*Fruit treated with forced
air are free of chemical
residue


Irradiation


*Limit dose of one KGy
established by Food and
Drug administration for
disinfestations

*Approved doses varies
between 150-250 Gy
across different fruit fly
species

*Dose varying between
50-150 Gy to achieve
probit for grapefruit and
mango infested with
Caribflies


*Provides quarantine
security by prevention of
adult emergence
*Strawberries and pa-
payas show high poten -
tial for commercial
irradiation because they
tolerate high doses
without excessive ham


Regulated
schedules


Advantages














Table 3-2 Continued
Fumigation

*Growing health & safety
concerns about the use of
MB


Limitations


Alternatives
under
investiga-tion


* Fumigation at 300C,
with decreased fumiga-
tion time and MB
concentration is very
promising due to losses in
MB phytotoxicity and
reduction in inorganic
residues
*Research on potential
candidates for MB re-
placement


Heat Treatment
*Valencia oranges &
grapefruits are resistant to
vapor heat injury but
other citrus are easily
damaged.

*Hot water treatment kills
fruit fly eggs but a low
percentage of larvae can
survive if fruits begin to
ripen.

*High potential fruit
damage and relative high
cost of application

*Satisfactory results are
obtained through com-
bining heat treatment with
MB fumigation of stone
fruits and papayas

*Need further research
word on the economic
feasibility of combined
heat and fumigation
treatments


Cold Treatment

*Use of cold treatments is
limited because most
fresh commodities are
damaged by extended
storage below 10C


*Preconditioning at warm
temperatures reduce cold
storage injury on some
citrus like grapefruit

*Proven technical feasi-
bility of combined heat
and cold treatments to
disinfect papayas of
Medfly and other fruit
flies


Controlled Atmospheres

*Very few fresh fruits can
tolerate extreme
atmosphere for extended
storage periods


*Keitt mango is very
tolerant to insecticidal
atmospheres for up to 5
days. Further research
work is needed to test
other mango varieties and
level of pest mortality


Irradiation
*Consumer reluctance
about the safety of
irradiated foods. Hard to
sell to the American
public

*Limited economic
feasibility of irradiation
while requiring high
capital investments

*Irradiation shows little
promise for perishable
commodities because of
excessive injury


*Need further research
work on consumer
acceptance of irradiated
food

*Need further research
work on the economic
feasibility of combined
cold and irradiation
treatments














CHAPTER 4
COST ANALYSIS OF THE MEDFLY DETECTION AND ERADICATION
PROGRAM

Specification of the Bayesian Modeling Framework

Overview of the Bayesian Decision Process

This sub-section discusses the basic principles behind Bayesian statistical

inference. The problem is to use historical information about Medfly interceptions and

trap sensitivity with a view to (1) computing the probability of detecting a Medfly

infestation into Florida at a given time and (2) determining the optimal trap density that

minimizes the expected cost of APHIS's prevention, detection, and eradication program.

Such an approach to statistical inference is called Bayes' theorem, which combines some

prior distribution and available data to form a posterior or revised distribution. The

underlying principle of this theory is that all uncertainties are described by probabilities:

unknown parameters have probability distribution both before the data are available and

after the data have been obtained (Cox and Hinkley 1974). While the theory does not

offer a formal guarantee of objective truth about the system under study, it does at least

ensure some kind of internal consistency among related decisions by the same individual.

The decision is called a Bayesian decision (or the optimal decision) if the action chosen

minimizes average (expected) loss that is associated with the costs of a wrong decision

(Wonnacott & Wonnacott 1977)

Let X1,......., Xn denote a sample distribution indexed by a continuous parameter 0.

While, in classical statistical estimation, it is appropriate to treat the parameter 0 as a









single fixed value, the Bayesian theory allows for treating it as a random variable to

account for our knowledge and uncertainty regarding the parameter's value (Press 1989;

Morgan and Henrion 1990). The function g (0) is called prior probability mass function,

since it is determined before observing X in the current experiment; that is, 0 is based on

previous practical experience and understanding. The posterior or revised distribution of

0 for X,,....., Xn is given by:


L(XI,"xNO)g(O) (4-1)
h( X,. .... ) -L .... (4-1)
fL(XI,-.,X O)g(O)dO

where the first term in the numerator L(xi,.....,XN|O) is a likelihood function indexed by

the parameter 0. The conditional probability h (0|xi,....,XN) is called posterior probability

mass function, given the current data, since it is determined after observing the current

data set. The prior distribution provides additional information into the analysis and

allows for a gain in logical clarity.

Figure 4-1 summarizes the logic of a Bayesian decision. The framework analyzes

the trade-off between early versus late detection in terms of costs of the Florida Medfly

prevention, detection and eradication program. Early detection costs are high for

trapping, but low for eradication (with a low probability of establishment). Late

detection costs for eradication are high (with a high probability of establishment). The

major components of the framework past data on Medfly interceptions, single trap

sensitivity, prior probability of infestation, multiple trap sensitivity are supportive of

the optimal decision relative to the trapping strategy. The probability of detection f (xi|

x2, z) is a posterior probability, as its computation revises the prior probabilities to reflect









the observational information available on trap density and sensitivity. If the objective is

to minimize the expected cost of eradication, then this cost has to be computed for each

possible trap density per location and season of the year. The optimal (or Bayesian)

decision is to select the trapping strategy associated with the lowest expected prevention

and eradication costs.

Definition of the Variables

We assumed the introduction, into Florida, of a small lot of mangos originating

from a hypothetical Latin American country and containing approximately 1,000 Medfly

eggs. We considered two correlated non-normal variables Xi,t and X2,t:

1. Xl,t stands for the total Medfly population present at a given time and some location
in Florida

2. X2,t accounts for the number of Medflies captured in the trapping system put in
place by APHIS. The amount of information provided by trapping reflects both the
density of traps and the single trap sensitivity.

The objective of this model was to build a spatially and temporally demographic

picture of a hypothetical Medfly infestation in Tampa and Miami. Spatial distribution of

Medfly host plants was assumed to be uniform throughout the regions of interest. The

mate-finding process was also considered a crucial determinant of the potential for pest

colonization in natural settings (Allee 1931; Prokopy and Hendrichs 1979). Along these

lines, the infestation model was initialized with a cohort of 50 ovipositing females

resulting from a twenty-percent survival of pre-adult Medflies with a sex ratio of 1:1 (i.e.

100 female adults and 100 male adults). Egg production covered a 10-day period at a

fixed rate of 11 eggs /female /day. Survival rates varied across stages; egg survival is 0.4,

larval survival is 0.5, pupal survival is 1. Only 20% of the newborn Medflies survived to

adulthood.









Temporal dimension

The infestation model is inspired by the time-specific life-table approach to

population growth, which is grounded on the assumptions of constant recruitment rate

and steady mortality rate (Southwood and Henderson 2000). The rate of insect population

growth is basically dependent on temperature and is predicted with time. Physiological

time is commonly expressed in terms of day-degrees (D) or hour-degrees (h), being the

cumulative product of total time x temperature (above the threshold) [Hughes 1962;

Hardman 1976; Atkinson 1977]. A minimum temperature exists below which no

measurable development takes place. The developmental zero for Medfly was found to

be = 54.3F (Shoukry and Hafez 1979). Thus, the number of day degrees, D,

accumulated above the developmental threshold for a life stage is computed as follows:


D(oF) = AV(F) TH(F) (4-2)


where AV stands for the average daily temperature in Fahrenheit and TH is the

developmental threshold for Medfly in Fahrenheit.

This temperature model is also used by APHIS to predict the entire life cycle of a

Medfly population and to guide program actions (eradication treatments, length of

trapping activities, and regulatory functions). About 3280 C-day degrees (590.4 F-days

degrees) must be accumulated before one life cycle has been completed (APHIS 2002).

Table 4-1 shows the number of day-degrees required by one bug to transition from one

stage to the next stage. Temperature data (minimum and maximum daily average

temperature) used to simulate the growth development process are collected from the

National Oceanic and Atmospheric Administration, United States Department of

Commerce.









The model design includes two locations (Miami or Tampa) at three different

months of the year (February, June, or October) to allow for both regional and seasonal

differences in the State. The analysis is carried out over a 200-day period, using an Excel

spreadsheet that provides information on the structure and the size of the pest population

over time. Under each scenario, the instantaneous rate of population change, r, and the

doubling time, DT16, are computed, using the following equations (Carey 1982b):


1= e rtlm (4-3)
t=O


DT= (1og2 (4-4)
r

where t is the age in days, It is the probability of surviving to age t, mt is the number of

female offspring produced at age t. The parameter, r, is key to predict the size of the

Medfly infestation at different points of time, using the finite version of the Malthusian

equation:


N,=NoNe (4-5)

where Nt and No stand for the size at time t and the initial size of the population,

respectively.

Spatial dimension

Spatial dispersion and movement is as important as birth and death rates for the

population dynamics of insects and is a major determinant of the boundaries of the

infested region (Papadopoulos et al. 2002, 2003). The underlying assumptions for

modeling spatial distribution of the infestation are that (1) the Medflies are considered


16 The parameter, DT, designates the time (in number of days) needed for the pest population to double.









relatively weak dispersers and (2) the vast majority of the population does not disperse

very far (Soria and Cline 1962; Katiyar and Valerio 1963; Nadel and Guerrieri 1969;

Serghiou and Symmons 1974; Wakid and Shoukry 1976; Plant and Cunningham 1991;

Katsoyannos et al. 1998; Papadopoulos et al. 2000; Papadopoulos et al. 2001). Table 4-2

gives the proportion of pre-adult and/or adult Medflies dispersed to different distances

from the epicenter. Random dispersal is the dominant strategy of spread, occurring over

an expected radius of 0.113 mile per month and thereby resulting in the scattering of

29% of the original population and the expansion of the infested area. Cases of long

distance flights in search for food and/or oviposition sites (Bateman 1972; Harris and

Olalquiaga 1991) and of human conveyance over distances ranging from 24 to 100 miles

(Williamson 1983) are also incorporated into the model, as leading to potential outbreaks

in new areas. Nevertheless, the probabilities of facing new Medfly outbreaks away from

the initial epicenter are subject to the so-called Allee-effect, that is, the opportunity of

finding a mate. If the number of adult females moving away from the epicenter is less

than one, the outbreak potential is considered insignificant. This restriction is based on

the assumption that the chances for the males to attract the females would be very low

when the size of the lek formed by the males is too small (Carey 1982b).


The model implies that the density of pests is highest at the epicenter and decreases

in space as small portions of the pest population move away from the epicenter. The

notion of the epicenter of a pest infestation suggests that there is a point from which the

pest infestation originates (Mangel et al. 1984). Figure 4-2 illustrates the spatial

dispersion pattern of a hypothetical pest infestation occurring in Miami during the month









of October at Day 50. Each 1-by-l-mile square parcel17 accounts for the infested unit

area, while the number inside each square parcel stands for the number of pre-adult

and/or adult females present in the square parcel. The size of infestation is the sum of the

1-by-l-mile square parcels and the quarantine area can be derived through delineating a

5-mi-wide buffer perimeter around all infested core areas.

Cost Function

The optimal decision (or Bayesian decision) was made on the basis of the optimal

trap density that minimizes the expected cost of the Medfly prevention, detection, and

eradication program. We assumed that the cost function of the emergency program was

approximated by the log functional form:


In(C)= o+(a,- 1)ln A + l In Xl, + 1(lnA + In )2 (4-6)

where C stands for the emergency program cost per ha, A is the infested area in ha units,

and Xtis the pest population density per ha. The size of infestation and the quarantine

area are the key parameters used to calculate the total emergency program costs by

summing (1) the cost of the emergency detection cost (DC), (2) the cost of the curative

release of sterile flies around all buffer perimeters (RC), and (3) the costs of weekly

malathion applications (AC) over the infested area under scenarios of low, moderate, and

high pesticide efficacy.18 The total emergency program cost (TC) under each scenario is

computed, using the following equations:



17 Each 1-by-l-mile square parcel can be associated with what is called a section in the Township and
Range system developed by the Federal government. Each township comprises 36 sections.

18 Pesticide efficacy is defined here as the proportion of bugs that will not survive the weekly spray
treatment. Thus, low, moderate, and high pesticide efficacies correspond to 70%, 80%, and 90% of bugs
killed during the spraying operation, respectively.









DC=WEEKLY EMERGENCY DETECTION COST (# SPRAYING WEEKS

+THREE LIFE CYCLES) (4-7)

RC= (QUARANTINE AREA/100) STERILE RELEASE COST (#

SPRAYING WEEKS +THREE LIFE CYCLES) (4-8)

AC = AREA INFESTED APPLICATION COST UNIT # SPRAYING

WEEKS (4-9)

TC = DC + RC + AC (4-10)

The spreadsheet models used to compute the number of spraying weeks required to

eradicate the pest population take into account both the proportion of individuals

surviving a spray treatment and the new adult females emerging during eradication

operations. The infestation rates applied under spraying conditions are extremely low,

ranging from 3 to 6 eggs per female over a three-day oviposition period. The Medfly

population is considered totally eradicated in the model when the sum total of ovipositing

females is less than 1.

We computed the OLS estimates of the parameters for the cost function under each

outbreak scenario (Miami or Tampa, February, June, or October). These estimates are

shown in Table 4-3. The F statistic was used to test whether the regression coefficients

are different in the different periods (February, June, and October) or it is appropriate to

pool the data and estimate a single equation for the entire period from February to

October. The resulting F ratios are highly significant, with calculated values for F [6,204]

approximating 390.5533 and 440.2952 for Miami and Tampa respectively. So, consistent

with our expectations, these results reject the null hypothesis that the coefficient vectors

are the same for the three periods in each location.









Future cost of eradication

As stated above, there is a trade off between early versus late detection.

Maintaining high trap densities over a wide area to discover low, patchy populations of

wild flies is extremely costly. On the other hand, if the trapping system is ineffective and

fails to detect the pest population at early stages, APHIS managers would have to face the

weighted values of the future cost of eradicating a growing pest population. Therefore,

the future cost of eradicating the pest at time t, FU(Xi,t), can be expressed as follows:

FU(,,) =

SX1,50X1,98 X1,119 t- (4-11)
SJ J J ( f(, 2 = Oz) f (X, X2= lz)C(1,t)dx ..... d
-X1,0 Xl1,50 X 1,98

where F(Xi,s,X2,s=0|z) and F(Xi,s,X2,s=1 z) are the multiple trap sensitivities for X2,s =0

and X2,s= 1, respectively. C(Xi,t,A) stands for the cost function of the emergency

program.

Optimization model

To find the optimal trap density, the minimization problem can be defined as

follows:

Min ...........PZ +FU(X,) + C(Xlt)f (2 = 1, XZ) (4-12)

where Z is the variable accounting for the trap density and P is the detection program cost

per trap unit. The calculation of P is based on the assumption that 47,404 traps are placed

in Florida over 4,490 square miles. The average annual price of a trap is roughly $168.22.

The value of the objective function is computed under all outbreak scenarios for the

values of Z lying between 1 and 500. The optimal trapping density for each scenario is

the one that minimizes the objective function.









Probabilistic Models

Probability of Detection: F (X1,t I X2,t, Z)

The purpose is to calculate the probability of detecting the presence of a Medfly

infestation at a given point of time, on the basis of prior information available

(probability of infestation, single trap sensitivity of the trapping system, and number of

traps). As shown in Figure 4-1, the probability of detection, F(Xi,t I X2,t, Z) is defined as

follows:

* F (xl,t| x2, Z) = Probability of detecting a Medfly population in Florida at a
given time t, given the number of traps (Z) and trap sensitivity

F(X,7,X2 Z)
F( Z) (4-13)
F(X2 Z)

* The term F(Xi,t, X2,t | Z) in the numerator is ajoint density function and, therefore,
is the product of the multiple trap sensitivity [F (X2,t Xit, Z) ] and of the
probability of infestation at t = 0, [ F (X1,o) ]. It can be calculated as follows:

F(Xl,tX2 Z)= F(X2 Xl,, Z)F(Xo0) (4-14)

* The term F(X2 I Z) in the denominator is a summation that encompasses all possible
outcomes of the trapping system, including the probabilities that the pest is present
but not trapped, and the probabilities that the presence of pest is detected with
different numbers of Medfly adults captured in the trapping system.

F(X2 z)= JF(x,,X2 Z)dX;, (4-15)
0

Probability of Infestation: F(X1,0)

This model of probability examines routes of Medfly introduction and uncertainties

regarding fly survival and effectiveness of exclusion activities. It is assumed that

opportunities of Medfly introduction into Florida are continuously increased through high

volume of international travel, agricultural industry demands, and international trade

agreements (USDA 1999). International air passenger baggage is considered the highest









risk pathway for Medfly into Florida. About 8.8 million air passengers arrive in Florida

per year, of which 5 million originate from Medfly infested countries (APHIS 1999).

Weighted passenger with Medfly infested materials to Florida is estimated in 1995 at 6.6

x 10-10, based on baggage surveys carried out by Plant Protection and Quarantine from

July, 1993 through September 1994 (APHIS 1994).

Using the general framework developed by Baker et al. (1993), the probability of a

Medfly infestation in Florida can be calculated as follows:

F(X,) = (1-- D)N (4-16)

where N stands for the number of passengers per year arriving from countries where

Medfly occurs, p is the proportion of weighted passengers with high risk materials, and

) is the probability that a single infested unit leads to an establishment. The parameter,

p, is considered the infestation level associated with air passenger baggage clearance. To

be 95% confident that the infestation level is no more than p, the number of passengers

(n) to inspect can be calculated (Couey and Chew 1986) as:

log(1 0.95)
n = (4-17)
log(1- p)

Under the assumption that the number of survivors per infested unit follows a

Poisson distribution, the parameter ) is defined (Whyte et al. 1996) as follows:


D = (1+e 2 2) (4-18)

where [t is the average number of pests present per infested unit, c] is the proportion of

individuals surviving to reproduce, and is the suitability of conditions for the pest.

Thus, the probability of a Medfly infestation is written as









F(X,)= 1-(l-pY [l+e-l,o-2e-1,02 ) (4-19)


where X(lo) (X= [4) stands for the number of ovipositing females per infested unit. As

stated above, all outbreak scenarios are built up under the assumption that the infestation

is initialized with a small lot of mangos containing 1000 eggs. The number of eggs per

infested mango is assumed to be 100 with a ten-percent probability of surviving to

adulthood and a 1:1 sex ratio.

Multiple Trap Sensitivity of McPhail traps: F(X2,t I X1,t Z)

The multiple trap sensitivity, F(X2,t I X1,t, Z), stands for the probability of capturing

one or several flies by trapping, given the occurrence of an infestation at a given time and

some location in Florida. It is a function of the size of the adult population and the

number of traps. Our study used data on different population levels of a uniform

distribution and age class with various numbers of McPhail traps per unit area (Calkins et

al. 1984). The conditional probability of detecting low to moderate populations of

Anastrepha suspense in citrus groves in Central Florida was described by a polynomial

approximation of the cumulative distribution function (cdf), which was constrained to the

zero-one range by a hyperbolic tangent function (Taylor 1984, 1990). The functional

form of the conditional cdf is written as:

F(X2 tZ) = 0.5 + 0.5 tanh(H(u' X,, Z)) (4-20)

and the associated conditional probability density function (pdf) is computed by taking

the derivative of Equation 4-20. Thus, the conditional pdf is:

f(x2 Xt,Z)= 0.5*H'(xIu,x2,Z)*sec h 2(H(l,tX2'Z) (4-21)








where H(.) is a polynomial function, H' is the partial derivative of H(.) with respect to x2,

tanh(.) is the hyperbolic tangent, and sech(.) is the hyperbolic secant. By restricting the

polynomial approximation to a quadratic function, Equation 4-21 can be extended as

follows:

f(X2 X1,,Z) = 05 (a2+ a2 X,t +a23z+2a X)
*sec h2(ao+al x, +azX2+a3z+al2 XltX2 + a13XltZ (4-22)
2 2 Z2)
a23X2Z +a ,t + a2X2 +a33ZZ)

The parameters characterizing the polynomial approximation are estimated by

least-square error method. The minimization problem is expressed as follows:

Min..... (f(x =, Z)- (1-T)) +(f(= l,,)-T)2 (4-23)

subject to:

f(x2 = 0, X, z) > 0.......V X,....34 < X, <10,000 (4-24)

f(X2 = 1,, z) > 0.......V .....34 < l,<10,000 (4-25)


X a,+at a2X2+ z+2ax1>O (4-26)

8H
Z a3+ a13 Xl+ a23 X2+ 2a33Z 0 (4-27)
dz

where f(x2=0, xilz) and f(x2=l,xi z) are the respective theoretical values of the probability

derived from the hyperbolic tangent function, Ti are the probabilities provided in the

dataset, and H/OXil,t and 8H/ Z are the partial derivatives of H with respect to X1,, and

Z. The last two constraints (Equations 4-26 and 4-27) are imposed to satisfy the

requirements that the probability density function be monotonic and increasing over the









range of values of X(,t) lying between 34 and 10,000 and Z varying between 1 and 500.

All correspondent probabilities are also constrained to be positive (Equations 4-24 & 4-

25).

The least-square error estimates provided by the Excel solver are given in Table 4-

4. All constraints and optimality conditions are satisfied. The value of the standard error

of the estimate accounts for the measure of the "goodness of fit" of the estimated

regression line. Also, the Kolmogorov-Smirnov test statistics do not reject the null

hypothesis that the theoretical values of the probability density functions (derived from

the hyperbolic tangent function) and the values of probability provided in the dataset

follow the same distribution.

Comparative Sensitivity of McPhail versus Jackson Traps

Results obtained in experiments with McPhail traps can be properly extrapolated,

using estimates of comparative trap sensitivity. Literature reported significant differences

between McPhail and Jackson traps in their performance of capturing Ceratitis capitata in

terms of probability of detecting small populations, number of flies captured, and

proportion of females (females/[males + females]) captured (Heath et al. 1997;

Katsoyannos 1994; Katsoyannos et al. 1999a, 1999b; Papadopoulos et al. 2001). McPhail

traps (ML traps) are baited with a dry synthetic multilure/liquid protein and, such as, are

used to capture both females and males of a number of pest tephritid species (Newell

1936; IAEA 2003). On the other hand, trimedlure compounds are typically placed in

Jackson traps (TML traps) that are effective in attracting males, but they are weakly

attractive to females (Beroza et al. 1961; Nakagawa et al. 1970; Harris et al. 1971).

Casana-Giner et al. (2001) argued that no attractant for female C. capitata was

comparable to the male C. capitata captures of TML traps, which may be due to a higher









response to odor-attraction of the males than females in C. capitata. Development of

trimedlure compounds would contribute to enhance the probability of Medfly control,

improve trapping strategies, and reduce costs of trapping systems.

We used capture data from experiments of sterile insect releases conducted by

APHIS in Tampa to develop estimates of comparative sensitivity of ML and TML traps

by making the assumption that the experimental conditions (type of crop, density of crop,

tree fruit distribution, trap location or position) were the same (Drummond et al. 1984).

The full dataset covers 1700 observations of daily captures from 2nd to 27 February (20

counts), from 31 May to June 30 (23 counts), and from 1st to 30 October (22 counts).

Trap counts were expressed as the mean number of flies caught per trap per day and

transformed to [In(catches +1)] to stabilize their variances before analysis (Katsoyannos

et al. 1999a; Cohen and Yuval 2000; Papadopoulos et al. 2001). Using the data for the

entire sample, February, June, and October, we obtained four estimated OLS regressions,

postulated as follows:

ln(catchesTML + 1) = ln(catchesML + 1) + u (4.28)

where p stands for the coefficient of comparative trap sensitivity. We used the F statistic

for testing whether the unrestricted regressions (as opposed to the restricted or pooled

regression) for the three periods were systematically different. A probability level of 0.05

was used for all statistical tests.

Results of the regression model (Table 4-5) show variations in mean daily captures

and coefficient of relative sensitivity across seasons and types of traps. Both TML and

ML traps were more effective in February, when the life cycle of the pest is longer and









fly dispersal ion is restricted. Nevertheless, TML traps outperformed ML traps in total

captures of male C. capitata during all periods. Mean daily captures in TML traps are

5.80 times greater than in February, 9.84 times greater than in October, and 7.65 times

greater in June.

The t-test statistic supported the hypothesis that the coefficients of comparative

sensitivity estimated in the OLS regressions are statistically different from zero.

Confidence intervals for the estimated coefficients are shown in Table 4.5. On the basis

of the F test, we strongly rejected the null hypothesis of homogenous slope coefficients.

The test sequence is naturally halted, as the regression models (Equation 4.28) postulated

do not contain any intercept. These findings are supportive of the conclusion that the

coefficients of relative sensitivity vary systematically across seasons and types of traps.

These coefficients are incorporated into the hyperbolic tangent function as augmentation

factors (Moss et al. 2004) to investigate the effects of trap sensitivity improvement on

probability of detection and optimal trap density.

Consider Y(y), the augmentation factor, \ being the technological change. The

multiple trap sensitivity of TML trap can be expressed as follows:


FrT (X2, X ,, Z)= wF,(X2, XI' ,Z) (4-29)

Results

The findings from the Bayesian modeling framework are outlined in this section.

Our study provides estimates of the colonization potential of Medfly populations in

Florida, probabilities of detecting low populations at early stages, expected costs of the

prevention and eradication program, and optimal trapping densities across different

locations and seasons.









Pest Population Projection

Results of the temporal model of infestation (Tables 4-6 and 4-7) show variations

in generation time across locations (Miami or Tampa) and seasons (February, June, and

October). In the event of a February infestation in Miami or Tampa, the expected time for

F1 generation to reach the ovipositional stage is 44 and 64 days, respectively. At 50 days

of the infestation, for instance, the projected size of the pest population is 860 and 200

Medflies, respectively. The instantaneous rate of change in population size is 0.0812 and

0.0587 for Miami and Tampa, respectively, implying that the pest population will double

approximately every 8.53 and 11.80 days, respectively (Table 4-7). These results follow

from the fact that average mean monthly temperatures in both Miami and Tampa are

below optimal levels for Medfly development during the months of February and March.

However, as the temperature increases during the months of April and May, the length of

the life cycle will be drastically reduced to 22 or 23 days for the fourth generation. The

instantaneous rate of population change in population will also increase to 0.1005 and

0.0988 in Miami and Tampa, respectively. As a result, the colonization potential at 119

days will increase with a population size approximating 38,781 and 22,990 Medflies,

respectively.

As expected, the colonization potential of a Medfly population is much higher

during the summer months, when average mean monthly temperatures for Miami and

Tampa are in the optimal range for Medfly development. The time to the ovipositional

stage averages 21 days and the projected doubling time is 4.65 days and 4.83 days for

Tampa and Miami, respectively. For instance, at 98 days of the infestation, the pest

population size in Miami and Tampa is expected to be 363,390 bugs and 532,267 bugs,

respectively. Differences in population size and structure between Tampa and Miami









(Appendix A) spell the sensitivity of the Medfly development to changes in weather

conditions.

As opposed to a February infestation, the October infestation starts under more

favorable weather conditions, and, therefore, with a moderate colonization potential. At

50 days of the infestation, initial pest populations in Miami and Tampa are expected to

attain a size of 1960 bugs and 1300 bugs, respectively. The first two generations will

increase on average by approximately 11.06% and 9.33% each day, respectively.

However, the duration of the egg, larval, and pupal stages will be considerably increased

by lower temperatures during the months of December and January, causing a significant

reduction in the colonization potential of the pest population. In Miami, generation times

increase from 29 days for the second generation to 52 days for the fourth generation,

while the intrinsic rate of population increase decreases up to 0.0777. In Tampa, the drop

in the intrinsic rate of population increase is even more severe. The development of larval

and pupal stages is expected to be almost stopping, thereby causing a sort of stagnation in

the pest population size.

Size and Cost of the Infestation

The results of the spatial model of the infestation (Table 4-8) show variations in

infested areas, quarantine areas, and total eradication costs under different outbreak

scenarios. Note that the eradication cost is hereby estimated under the assumption of a

ninety-percent pesticide efficacy.

None of the outbreak scenarios simulated in the spatial-temporal model can be

considered early-detected infestations (which, according to the APHIS criteria, are

supposed to spread over a quarantine area equal to or less than 110 square mile).

Maintaining such a target involves detecting the pest population and starting to spray it at









less than 44 days. The reason for this is that, at 50 days, the initial parent cohort under all

outbreak scenarios (including both seasonal and geographical variations) is expected to

start ovipositing, implying that the eradication will not be completed around the time

required for all F2 eggs to emerge as adults. By this time, the infested area averages 22

square miles large with a quarantine area ranging from 121 to 421 square miles. The

February infestation in Tampa is expected to be confined to a two-square-mile area, while

the June infestation in Miami or Tampa will spread over a 32 square mile area and the

total regulated area will encompass a 421-square-mile area.

Nevertheless, any infestation detected at 50 days can be considered a moderate

outbreak as the maximum distances flown by the flies are expected to be less than 4

linear miles and the expected proportion of adults moving away from the epicenter is low

(< 10%). Total cost of eradication of a 50-day old infestation is expected to vary across

locations and seasons, from $ 2.06 million for a February infestation in Tampa, $ 6.1

million for an October infestation, to $ 26.3 million for a June infestation in Miami.

When spraying starts three weeks later (i.e., at 77 days of the infestation), the cost

of eradication of a February or October infestation in Miami will approximately

quadruple compared to the costs of the 50-day old infestation in the same location and

season. The 77-day old pest population in Miami is expected to spread over a 21 and 24

square-mile area for the February and October infestations, respectively. However, the

rate of increase in eradication cost is much less in Tampa, approximating 48% and 110%

for a 77-day-old infestation occurring in October and February, respectively. The Tampa

infestation is expected to spread over a 12-square-mile area and less than 250 square

miles will be placed under strict regulation. Theses results reflect well the sub-optimal









weather conditions prevailing in Tampa for the Medfly development, especially during

the winter months.

A 77-day-old infestation must be considered a serious outbreak when it occurs

during the summer months either in Miami or Tampa. Average distances flown by the

flies are expected to be greater than 12 linear miles and the projected proportion of adults

moving away from the epicenter will be around 12%. Such an infestation is very likely to

spread over two to four counties. The expected treatment area covers 115 square miles

large and 1,055 square miles are expected to be under strict regulation. The Tampa

infestation will be more costly than the Miami one. The reason for this relative difference

is based entirely on the size and the structure of the Fi generation in the two locations.

The highest proportion of the pest population in egg, larval, and pupal stages in Tampa

spells the difference in eradication costs.

As shown in Table 4-8, it will cost twice or three times more to eradicate the Fi and

F2 generations of an October infestation than those of a February infestation. The reversal

of this situation is observed for the F3 and F4 generations. At 98 days and more, whilst

eradication costs of an October infestation in either Miami or Tampa increase at a

decreasing rate, those of a February infestation tend to increase at an increasing rate. In

Miami, the projected eradication cost of a 98-day or 119-day old infestation will be

approximately $ 147 million and $ 1.9 billion, respectively. At 119 days, for instance, the

February infestation will spread over a 300-square mile area with high risks of facing

additional outbreaks in remote areas during eradication operations.

A 98 or 119-day old infestation is even more serious when it occurs during the

summer months in either Tampa or Miami. The expected quarantine area encompasses









approximately 6,800 and 8,000 square miles respectively. Eradication costs are expected

to be extremely expensive, approximating $ 2.9 billion and $ 7.7 billion in Miami and

Tampa, respectively. It is likely that such infestations become out of control.

Multiple Trap Sensitivities for ML Traps

The estimates of multiple trap sensitivity for the ML trap (Table 4-9) are extremely

low, thereby confirming how difficult it could be to detect low Medfly populations at

early stages. As expected, the highest trap sensitivities are reported for the pest

population during the summer months. For instance, the sensitivity of a trapping system

to a 50-day-old infestation varies across trap densities from 4.95 x 10-6 for a density of

one trap per square mile in Miami to 5.84 x 10-4 for a density of 21 traps per square mile

in Tampa. The chances of detecting a 119-day old infestation are higher, with trap

sensitivities lying between 1.35 x 10-4 and 1.11 x 10-2. On the other hand, the lowest trap

sensitivities are found for a 50-day old infestation during the months of February and

October. For instance, the chances of detecting a 50-day old infestation in Tampa are ten

times lower in February than in June.

The marginal values of trap sensitivity (Table 4-10) are all positive within the

interval ranges of our dataset, varying from 7.08 x 10-7 to 6.31 x 10-4 with a clear

tendency to increase with trap density. The higher the trap density, the higher will be the

marginal trap sensitivity. For instance, for a given pest population in Miami (October),

the marginal trap sensitivity to a 50-day old infestation varies from 3.88 x 10-6 to 2.74 x

10-5 for trap densities of 2 and 21 traps per square mile, respectively. The direction of

changes in marginal trap sensitivity is also the same with changes in pest population size.

For a given trap density in Tampa (June), the marginal trap sensitivity varies from 5.7 x









10-6 for a 50-day old infestation to 1.59 x 10-4 for a 119-day old infestation. These results

strongly suggest that there is a positive gain in increasing trap density.

Probabilities of Detection for ML Traps

The computed values for the probability of detection (Table 4-11) differ from the

trap sensitivities in that the former are conditional to (1) the probability of infestation,

F(X1), estimated at 0.005371 for the State of Florida and (2) to all possible outcomes in

the trapping system, ranging from X2=0 to X2->0. Nevertheless, the results relative to the

probability of detection follow the same pattern as those for the multiple trap sensitivity.

The reported probabilities vary across trap densities from 3.23 x 10-6 to 2.24 x 10-4. The

highest probabilities of detection are found for the pest infestation occurring during the

summer months, while the chances of detecting a pest infestation during the month of

February are extremely low. Furthermore, all marginal values of probability of detection

are positive, ranging from 4.77 x 10-7 to 2.11 x 10-4. Like for the marginal trap

sensitivity, the general tendency is for the marginal values of the probability of detection

to increase with trap density and pest population size. Nevertheless, at some trap

densities, i.e. at a density of 12 traps per square mile, the marginal values of probability

of detection show a slight decrease with an increase in trap density.

Optimal Trap Densities

The optimal solutions regarding trapping density are presented in Table 4-13. The

lowest optimal trap density for ML traps is 82 traps per ha and is reported for a June

infestation occurring in Tampa. A total of 184 traps per ha are to be placed in Miami

during the month of February to achieve the optimal solution. The highest optimal trap

density for ML traps (465 traps per ha) is found for a Tampa infestation starting in









October. These results are supportive of the hypothesis that the optimal trapping density

varies across locations and seasons.

The optimal solutions for ML traps greatly differ from those for TML traps that are

found to be more effective for capture of male C. capitata. Optimal trapping densities for

TML traps range from 9 to 80 traps per ha, with the highest optimal trap density being

reported for a February infestation occurring in Tampa. A total of 25 traps per ha are to

be placed in Tampa during the month of October to achieve the optimal solution. The

lowest trap density (9 traps per ha) is found for a June infestation in Tampa. These results

highly suggest that emphasis should be placed more on improving trap sensitivity rather

than on increasing trap density.

Conclusions

The objectives of this chapter were to (1) compute the multiple trap sensitivities for

ML and TML traps and the probabilities of detecting a Medfly infestation in Florida at its

early stages and (2) determine the optimal trapping density that can minimize the total

expected cost of the Medfly prevention, detection and eradication program. Our study

shows that the colonization potential greatly varies across seasons and locations and that

none of the outbreak scenarios simulated in the spatial-temporal model can be considered

early-detected infestations. The chances of detecting Medfly populations at early stages

are extremely low. Sensitivity of ML traps to a 50-day-old infestation varies across trap

densities from 4.95 x 10-6 for a density of one trap per square mile in Miami to 5.84 x 10-

4 for a density of 21 traps per square mile in Tampa. Because of significant progress

made in developing more potent lures for male C. capitata, TML traps are expected to be

on average 7.76 times more sensitive than ML traps.









The results relative to the probability of detection follow the same pattern as those

for the multiple trap sensitivity. The reported probabilities vary across trap densities from

3.23 x 10-6 to 2.24 x 10-4. The highest probabilities of detection are found for the pest

infestation occurring during the summer months, while the chances of detecting a pest

infestation during the month of February are extremely low.

Optimal trapping densities also vary across locations and seasons, ranging from 82

to 465 traps per ha for ML traps and from 9 to 80 traps per ha for TML traps. These

results strongly suggest that emphasis should be placed on improving the multiple trap

sensitivity, which is, reportedly, increasing with trap density and population size.



Table 4-1 Distribution of day degrees required by stage
Transitional Phase Day Degrees Required to Transition (F)
Minimum Maximum Average
Egg/ Larvae 33.8 47.9 40.85
Larvae/ Pupae 153.9 219.45 186.67
Pupae / Pre-adults 308.6 436.8 372.7
Pre-adults / Adults 596 608.3 603.1
Source: APHIS 2002

Table 4.2 Average monthly distances flown by different fractions of Medfly population
Means of Fraction of population moving Average distances flown
spread away from the epicenter (linear mile)
0.15 0.125
Random 0.06 0.435
dispersal 0.042 0.75
0.027 1.25
0.018 1.75

Long 0.0015 12
distance 0.0006 24
flight & 0.00042 48
human 0.00030 96
conveyance 0.00018 108





























IX fx:x)dx0
Cost-=-P.Z-(fixed-
cost)r v
P:--pricer Cost -of-trapping- surveyed Cost- of-eradication*1
Z:-number-of-traps'

-, t Optimal-dec ionI I **-- -C(X )
where.XI-is-population-size-at-
Alin PZ+C( ~1) (12n Z)* di time-t"





Figure 4-1. Bayesian decision process














I 3- I ?SO I '-r I- ,d buffer







Figure 4-2 Treatment and quarantine areas of an infestation scenario in Miami (October)

Table 4-3 Eradication cost equations
Tampa Miami
Coeffi-
cients Pooled February June October Pooled February June October
data data
Ao 9.8013 10.32897 7.072238 8.774308 9.6896 9.626233 8.802522 8.234525
(se) (0.456) (0.0015) (0.0164) (0.0000) (0.440) (0.0108) (0.0002) (0.001)
A1 0.3409 0.917753 1.52183 0.702589 0.5854 0.85166 1.061598 0.822641
(se)a (0.032) (0.0137) (0.0118) (0.0114) (0.029) (0.0187) (0.0052) (0.0193)
All 0.0634 0.011916 -0.04039 -0.04655 0.0241 0.046724 -0.02198 -0.00538
(se) (0.075) (0.0298) (0.0279) ((0.0345) (0.072) (0.0548) (0.012) (0.0377)
Residual
sums of 196.983 3.77434 9.76437 0.58212 181.18 9.770366 1.40914 3.330146
square
F[6,204] 440.2952 390.5533

a se=standard error of the parameter

Table 4-4 Hyperbolic tangent approximation of marginal probability function for
multiple trapping sensitivity of McPhail traps
ao= -3.74367 ; ai= 0.25243; a2= 1.053351;
Coefficients a a3 = 0.68775; al2= -0.04322; al3= -0.02082;
a23= -0.07316; all=0.0114269;
a22=0.69595; a33=-0.01823
Standard error of the estimate Residual sum of squares = 0.00024037
Standard error = 0.0000160247
Kolmogorov-Smimov test b D value = 0.08
P-value = 1.00
a These coefficients refer to the conditional probability density function described in Equations 4.21 & 4.22,
which is approximated by a quadratic function. The size of the adult population, X, and the number of
traps, Z, are the two major variables of this function. bKolmogorov Smimov test determines if two datasets
differ significantly by using the maximum vertical deviation between their curves. The maximum
difference in cumulative fraction is D=.45. With a D value estimated at 0.08 for a p-value =1.00, the KS
test supports the hypothesis that the values of the probabilities in the dataset and the theoretical values of
the probability density functions follow the same distribution.










Table 4-5 Coefficients of comparative sensitivity and mean daily captures by period


Pooled Data
Trap counts 65


Untransformed means of
daily captures
TML trap
ML trap
Coefficient of comparative
sensitivity of TML to ML
traps (in log space)
confidence intervals
standard error
t value
Residual sum of squares
Calculated F[2,64] value


43.4
5.9


1.928979
1.62< P <2.23
0.152154
12.67778
36.17546


84.4
11.6


21.5
3.5


1.75722 2.034088
1.01 0.354838 0.400214
3.451957 2.936439
4.92351 10.55651
12.30973


29.0
3.3


2.286278
1.48<3 <3.09
0.387087
3.211114
10.6455


Table 4-6 Distribution of the expected population size and generation time per location
and per season at 50, 77, 98, and 119 days a of the infestation


Month Day


Feb.





June


Expected b
population size
(Medflies)
860
2,211
20,570
38,781


4,741
75,044
363,390
1,448,303


liami
Generation c
time
(Days)
44
33
27
23

24
21
21
21


Tampa
Expected Generation time
population size
(Medflies) (Days)
200 64
1,420 33
1,584 25
22,990 22

5,588 23
91,948 21
532,267 21
2,206,223 21


50 1,960 26 1,300 29
Oct 77 14,300 29 2,200 52
98 25,531 41 14,179 91
119 138,243 52 14,300 39
a Medfly populations at 50, 77, 98, and 119 days roughly correspond to the first (F1),
second (F2), third (F3), and fourth (F4) generations, respectively. b Population size is
expressed in terms of numbers of female pre-adults or adults and its calculation takes into
account the survival probabilities at all stages. C Generation time refers to the time
required for one generation of eggs to reach the ovipositional age.


February
20


June
23


October
22











Table 4-7 Distribution of the intrinsic rates of increase and doubling times of the pest
population per location and per season


Miami


Tampa


Month Day a Intrinsic rate of
increase
50 0.0812
Feb. 77 0.0812
98 0.1005
119 0.0805


June


0.1484
0.1249
0.1108
0.1008


Doubling time
(days)
8.53
8.53
6.89
8.61

4.67
5.55
6.26
6.87


Intrinsic rate of
increase
0.0587
0.0587
0.0813
0.0989


0.1490
0.1263
0.1144
0.1042


Doubling time
(days)
11.81
11.81
8.53
7.01


4.67
5.49
6.06
6.65


50 0.1240 5.59 0.1222 5.67
Oct 77 0.0974 7.12 0.0644 10.76
98 0.0769 6.36 0.0718 9.65
119 0.0777 7.96 0.0551 12.57
a Medfly populations at 50, 77, 98, and 119 days roughly correspond to the first (F1),
second (F2), third (F3), and fourth (F4) generations, respectively.


Table 4-8 Distribution of the infested area, quarantine area, and eradication cost per
location and per season at 50, 77, 98, and 119 days a of the infestation


Miami
Month Day Infested Quarantine
area area
(square mile)
50 4 182
77 21 327
Feb 98 81 860
119 299 1,943


421
1,055
4,276
6,764

252
360
740
758


Eradication
cost
($1,000)
2,327
10,325
146,824
1,960,186

26,265
289,511
1,574,211
2,907,668

6,091
22,453
61,384
99,271


Tampa
Infested Quarantine
area area
(square mile)
2 121
11 208
39 504
140 936


421
1,055
5,714
8,228

192
240
240
263


Eradication
cost b
($1,000)
2,066
4,345
27,991
373,674

24,825
315,602
5,355,057
7,702,316

4,913
7,281
7,281
15,882


a Medfly populations at 50, 77, 98, and 119 days roughly correspond to first (F1), second
(F2), third (F3), and fourth (F4) generations, respectively. b The costs for pesticide
applications are part of the total eradication cost and are estimated under the assumption of 90-
percent pesticide efficacy


50
77
June 98
119

50
Oct 77
98
119













Table 4-9 Distribution of multiple trap sensitivities for ML traps under all outbreak scenarios for different trap densities
Trap density 50 77 98 119 50 77 98 119
(per mi 2) Tampa February Miami February


0.000000584
0.000002
0.00000966
0.0000305
0.000041
0.0000652
0.0001

0.00000553
0.000017
0.0000707
0.0002
0.000261
0.000396
0.000584

0.00000332
0.0000104
0.0000448
0.000129
0.00017
0.000261
0.0003888


0.000000877 0.00000184
0.00000294 0.00000594
0.0000138 0.0000265
0.0000425 0.0000787
0.0000568 0.000104
0.0000894 0.000161
0.000137 0.000243
Tampa June


0.0000287
0.0000823
0.000315
0.000832
0.001068
0.001577
0.002266
Tampa
0.00000401
0.0000125
0.000053
0.000152
0.000199
0.000304
0.000451


0.0000732
0.000203
0.000742
0.001897
0.002414
0.003512
0.004977
October
0.0000132
0.000053
0.000155
0.000423
0.000547
0.000817
0.001188


0.00000871
0.0000262
0.000107
0.000296
0.000384
0.000578
0.000846

0.000188
0.000506
0.001771
0.004382
0.005526
0.00793
0.011088

0.0000129
0.000039
0.000153
0.000416
0.000539
0.000805
0.001171


0.000001
0.00000334
0.0000155
0.0000475
0.0000634
0.0000995
0.000152

0.00000495
0.0000153
0.000064
0.000182
0.000238
0.000361
0.000534

0.00000365
0.0000383
0.0000487
0.00014
0.000184
0.000281
0.000418


0.00000192 0.00000949
0.00000618 0.0000285
0.0000275 0.000115
0.0000814 0.000318
0.000108 0.000413
0.000167 0.000621
0.000251 0.000907
Miami June


0.0000265
0.0000763
0.000293
0.000776
0.000998
0.001474
0.00212
Miami
0.0000104
0.0000114
0.000125
0.000345
0.000447
0.000671
0.000979


0.0000599
0.000168
0.000618
0.00159
0.002027
0.002959
0.004205
October
0.0000117
0.0000311
0.000139
0.00038
0.000492
0.000737
0.001074


1
2
5
10
12
16
21


0.000012
0.0000355
0.000142
0.000388
0.000503
0.000753
0.001097

0.000135
0.000367
0.001305
0.003265
0.004131
0.005956
0.008368

0.0000421
0.000119
0.000447
0.001166
0.001491
0.002188
0.003126













Table 4.10 Marginal trap sensitivities for ML traps under different outbreak scenarios
Trap density 50 77 98 119 50 77 98 119


0.000000708
0.00000255
0.000004168
0.00000525
0.00000605
0.00000696

0.000005735
0.0000179
0.00002586
0.0000305
0.00003375
0.0000376


0.00000354
0.0000115
0.00001684
0.0000205
0.00002275
0.0000254


(per mi )


Tampa February
0.000001032 0.00000205
0.00000362 0.00000685
0.00000574 0.0000104
0.00000715 0.00001265
0.00000815 0.00001425
0.00000952 0.0000164
Tampa June
0.0000268 0.0000649
0.0000775 0.0001797
0.0001034 0.000231
0.000118 0.0002585
0.000127 0.0002745
0.000378 0.000293

Tampa October
0.000004324 0.0000129
0.0000135 0.0000386
0.0000198 0.0000536
0.0000235 0.000062
0.00002625 0.0000675
0.0000294 0.0000742


0.00000117
0.00000405
0.0000064
0.00000195
0.00000903
0.0000105

0.00000518
0.0000162
0.0000236
0.000028
0.0000308
0.0000346


0.00000388
0.0000124
0.0000183
0.000022
0.0000243
0.0000274


Miami February
0.00000213 0.00000951
0.00000711 0.0000288
0.00000108 0.0000406
0.0000133 0.0000475
0.0000148 0.000052
0.0000168 0.0000572
Miami June
0.0000249 0.00005405
0.0000722 0.00015
0.0000966 0.0001944
0.000111 0.0002185
0.000119 0.000233
0.000129 0.0002492


0.00000874
0.0000269
0.0000378
0.000044
0.0000485
0.0000536

0.000159
0.000422
0.000522
0.000572
0.000060
0.000632


0.0000127
0.0000382
0.0000526
0.0000615
0.0000665
0.0000732


October
0.0000115
0.0000348
0.0000482
0.000056
0.00006125
0.0000674


0.00001175
0.0000355
0.0000492
0.0000575
0.0000625
0.0000688


0.000116
0.0003127
0.000392
0.000433
0.000456
0.0004824


0.0000384
0.0001093
0.0001438
0.0001625
0.0001743
0.0001876


Miami
0.0000104
0.0000313
0.000044
0.000051
0.000056
0.0000616


2
5
10
12
16
21













Table 4-11 Distribution of probabilities of detection for ML traps under all outbreak scenarios for different trap densities
Trap density 50 77 98 119 50 77 98 119
(per mi 2) Tampa February Miami February


0.00000484 0.0000107
0.00000615 0.0000124
0.00000907 0.0000174
0.00001382 0.0000256
0.00001595 0.0000292
0.00002074 0.0000374
0.00002796 0.0000496
Tampa June


0.00000329
0.00000418
0.00000635
0.00000992
0.00001151
0.00001513
0.00002041

0.00003057
0.00003555
0.00004645
0.00006505
0.00007329
0.00009190
0.0001192

0.00008354
0.00002175
0.00002943
0.00004195
0.00004774
0.00006057
0.00007985


0.000405
0.000425
0.000487
0.000617
0.000678
0.000815
0.001015
October
0.0000729
0.0000815
0.0001018
0.0001376
0.0001536
0.0001896
0.0002425


0.0000481
0.0000548
0.0000703
0.0000962
0.0001078
0.0001341
0.0001726

0.001039
0.001058
0.001163
0.001425
0.001552
0.001840
0.002263

0.0000713
0.0000801
0.0001005
0.0001353
0.0001513
0.0001868
0.0002389


0.00000552
0.00000698
0.00001018
0.00001545
0.00001781
0.00002309
0.00003102

0.00002736
0.000032
0.00004204
0.00005919
0.00006684
0.00008378
0.0001089

0.00002018
0.00002384
0.00003199
0.00004553
0.00005167
0.00006521
0.00008531


0.0000106 0.0000525
0.0000129 0.0000596
0.0000181 0.0000755
0.0000265 0.0001034
0.0000303 0.0001159
0.0000387 0.0001441
0.0000512 0.0001851
Miami June


0.0001465
0.0001596
0.0001925
0.0002524
0.0002803
0.0003421
0.0004327
Miami
0.0000575
0.0000650
0.0000821
0.0001122
0.0001255
0.0001557
0.0001998


0.0003311
0.0003513
0.0004060
0.0005172
0.0005692
0.0006867
0.0008581
October
0.0000647
0.0000726
0.0000913
0.0001236
0.0001318
0.0001710
0.0002192


1
2
5
10
12
16
21


0.0001586
0.0001721
0.0002069
0.0002706
0.0002999
0.0003659
0.000462
Tampa
0.00002217
0.00002614
0.00003482
0.00004943
0.00005588
0.00007055
0.00009204


0.0000663
0.0000742
0.0000933
0.0001262
0.0001413
0.0001748
0.0002239

0.0007463
0.0007676
0.0008574
0.001062
0.001160
0.001382
0.001708

0.0002327
0.0002489
0.0002937
0.0003793
0.0004187
0.0005078
0.0006380














Table 4-12 Marginal values of probability of detection for ML traps densities under different outbreak scenarios


Trap density
(per mi 2)
2
5
10
12
16
21

2
5
10
12
16
21

2
5
10
12
16
21


0.000000477
0.000000721
0.000000714
0.000000797
0.000000904
0.000001055

0.00000249
0.00000363
0.00000372
0.00000412
0.00000465
0.00000545

0.00000170
0.00000256
0.000002505
0.0000289
0.00000321
0.00000372


Tampa February
0.00000065 0.00000113
0.000001458 0.000002494
0.000002378 0.00000409
0.000001064 0.000001805
0.00000239 0.000004079
0.00000361 0.000006114
Tampa June


0.00000673
0.00001741
0.00000318
0.00001465
0.00003302
0.0000482
Tampa
0.00000198
0.00000434
0.00000731
0.00000322
0.00000733
0.00001074


0.00000995
0.00003146
0.0000648
0.0000304
0.0000685
0.0001003
October
0.00000429
0.00001013
0.0000178
0.00000801
0.00001800
0.0000264


0.00000332
0.00000775
0.00000129
0.00000578
0.00001314
0.00001926

0.00000949
0.0000526
0.000131
0.0000633
0.000144
0.0002112

0.00000439
0.00001021
0.00000173
0.00000803
0.00001772
0.00002608


0.000000729
0.00000107
0.000000105
0.00000118
0.00000132
0.00000159

0.00000232
0.00000335
0.00000343
0.00000382
0.00000424
0.00000504

0.00000183
0.00000272
0.00000271
0.00000307
0.00000338
0.0000402


Miami February
0.000001156 0.000003572
0.000002571 0.000007974
0.000000420 0.00001394
0.00000192 0.000006276
0.000004213 0.00001406
0.00000623 0.00002049
Miami June


0.00000654
0.0000164
0.0000299
0.00000139
0.0000309
0.0000452
Miami
0.00000377
0.00000569
0.00000602
0.00000666
0.00000754
0.00000881


0.00001011
0.00002733
0.0000555
0.00002604
0.0000587
0.0000857
October
0.00000394
0.00000725
0.00000645
0.00000728
0.00000821
0.00000963


0.00000395
0.00000952
0.00001645
0.00000753
0.00001675
0.0000245

0.0000106
0.0000449
0.0001023
0.0000490
0.000111
0.0001628

0.00000807
0.00001493
0.00001711
0.00001974
0.00002226
0.00002604









Table 4-13 Optimal trapping density per type of trap, location and month
Optimal trap density (# traps per ha)
Month Miami Tampa
ML trap TML trap ML trap TML trap
February 322 55 465 80


June

October


248