ML072060598

NMFS-SEFSC-144: Assessment Update for the Kemp'S Ridley and Loggerhead Sea Turtle Populations in the Western North Atlantic

ML072060598
Person / Time
Site:	Oyster Creek
Issue date:	11/01/2000
From:	Baker D, Mineta N US Dept of Commerce, National Oceanographic and Atmospheric Administration
To:	Office of Nuclear Reactor Regulation
Davis J NRR/DLR/REBB, 415-3835
Shared Package
ML072060321	List: ML072060325 ML072060330 ML072060334 ML072060337 ML072060339 ML072060342 ML072060344 ML072060345 ML072060373 ML072060377 ML072060382 ML072060384 ML072060391 ML072060397 ML072060400 ML072060403 ML072060405 ML072060406 ML072060409 ML072060414 ML072060439 ML072060441 ML072060543 ML072060545 ML072060550 ML072060555 ML072060558 ML072060561 ML072060564 ML072060570 ML072060573 ML072060592 ML072060594 ML072060596 ML072060598 ML072060601 ML072060603 ML072060606 ML072060609 ML072060613 ML072060616 ML072060618 ML072060621
References
NMFS-SEFSC-444
Download: ML072060598 (132)

Text

NOAA Technical Memorandum NMFS-SEFSC-444 ASSESSMENT UPDATE FOR THE KEMP'S RIDLEY AND LOGGERHEAD SEA TURTLE POPULATIONS IN THE WESTERN NORTH ATLANTIC A Report of the Turtle Expert Working Group November 2000 U. S. Department of Commerce National Oceanic and Atmospheric Administration National Marine Fisheries Service Southeast Fisheries Science Center 75 Virginia Beach Drive Miami, FL 33149 CAL ' ll-ýZ000

10 ATM08p,_

NOAA Technical Memorandum NMFS-SEFSC-444 ASSESSMENT UPDATE FOR THE KEMP'S RIDLEY AND LOGGERHEAD SEA TURTLE POPULATIONS IN THE WESTERN NORTH ATLANTIC A Report of the Turtle Expert Working Group U. S. DEPARTMENT OF COMMERCE Norman Y. Mineta, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION D. James Baker, Administrator NATIONAL MARINE FISHERIES SERVICE Penelope H. Dalton, Assistant Administrator for Fisheries November 2000 Unpublished reports are used for documentation and timely communication of preliminary results, interim reports, or special-purpose information, and have not received complete formal review, editorial control, or detailed editing.

NOTICE The National Marine Fisheries Service(NMFS) does not approve, recommend or endorse any proprietary product or material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material herein or which has as its purpose any intentto cause directly or indirectly the advertised product to be used or purchased because of NMFS publication.

Correct citation of this report is:

Turtle Expert Working Group. 2000. Assessment Update for the Kemp's Ridley and Loggerhead Sea Turtle Populations in the Western North Atlantic. U.S. Dep. Commer. NOAA Tech.

Mem. NMFS-SEFSC-444, 115 pp.

Copies of this report can be obtained from:

National Marine Fisheries Service Southeast Fisheries Science Center 75 Virginia Beach Drive Miami, FL 33149 (305) 361-4285 or National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 (703) 605-6000 (800) 553-6847 (rush orders) ii

Table of Contents List of Tables .......... ...................................... vi List of Figures ......... ........... .................. .. 0viii Preface ................................... ................... x Executive Summary....... ................................ .... xii Turtle Expert Working Group Members .......................... xiv Part 1. Kemp's Ridley Status Report .......... 11..............

Summary .......................................................... 1 History of Nesting at Rancho Nuevo ................................. .

New Nesting Information ..................................... 2 Mdxico - Update of nest numbers 1966-1998 ........................... 2 Nesting recordsfrom South Texas ..................................... 2 Life Stage Definitions and Geographic Locations ..................... 3 Post-hatchlingstage ............................................... 3 Benthic immaturestage (20-60 on) ................................. 3 A dult stage ............ .............. ............................ 3 Status and Trends ............. ................. .................. 4 Nesting data .............. ...................................... 4 Nests vs. hatchlings released 8, 10, and 12 years earlier..... I............. 4 Strandings............................................ 4 Strandings "hotspots" ........................... ................... 5 Changes in size distributions ............................ .5 In-water surveys .................................................. 6 Conclusionsfrom nesting and strandingsdata ........................ 8 Nest protection in Mdxico - Future strategies......................... 9 Population size and trends from a deterministic age-based model .............. 9 Model modificationsfrom the 1998 report .... ....................... 9 Model form

....................................... ................. 10 Parameterization............................. ................... 11 Nesting data used ....................................... 11 Reproduction

.................................................. 12 Sex ratio (all age classes)

................... .................................... 13 Age and growth (see Appendix 1) ........................ ..... 13 NOAA Tech. Mem. NMFS-SEFSC-444 ii

7 1 Y, . .

Benthic immature mortality from strandings catch curve estimates (Zb) ................................................. 13 M odel Results .............................................. 14 Model fits to observed nests ................................. 14 Sensitivity of model results to changes in small benthic immature Zb

..................................................... 15 Model fits with an empirically-derived estimate off ............. 15 Best-fit model projections ................................. 16 Literature cited ................................................... 16 Part 2. Loggerhead Status Report ................................... 40 Sum m ary ............................................................. 40 Stock Definition and Geographic Range .................................... 40 Life stage definitions .............................................. 42 Status and Trends ..................................................... 43 Nesting beaches.............................................. 43 Northern Subpopulation .................................... 43 South Florida Subpopulation .......... .................... 43 Florida Panhandle Subpopulation .................... ...... 44 Nest count correlations across states and regions ............... 44 Yucatan nesting subpopulation ............................. 44 Loggerhead nesting elsewhere in the Gulf of Mexico ............. 44 Strandings.................................................. 45 Trends ............................................... 45 Strandings "hot spots" ................................... 45 TED effect analyses ........................................ 46 In-water surveys .............................................. 46 Sex ratio .............................................. .... 48 Size distributions ............................................. 48 Mean size of stranded loggerheads 1986-87 vs. 1996-97 .......... 48 Size distributions from the St. Lucie Power Plant .............. 49 Populationsize estimates ........................................... 49 Conclusionsfrom nest numbers, strandings,and in-watersurveys ......... 49 Literature Cited ....................................................... 51 Part 3. Anthropogenic Mortality Sources ............................. 71 Sum m ary ................................................... ......... 71 Mortality Sources for Pelagic Stage .................................... 71 Pelagiclonglinefisheries .......................................... 71 Marine debris .................................. e ................ 72 Mortality Sources for Benthic Feeding Stages .............................. 72 Trawls ......................................................... 72 iv NOAA Tech. Mem. NMFS-SEFSC-444

Gil*:*nets .......................................... ,,.......... .. 73 Gill nets ........................................................ 73 Hook and line................................................ 74 Power plants ................................................ 75 Marine debris.........................75 Literature Cited ....................................................... 76 Part 4. Take Limits ........................................... 80 Introduction .......................................................... 80 Three Interim Strandings Limits (ISL) calculations: Methods and Limitations... 81

1) Running average........................................... 81

2) Slope method .................................................. 82 Slope calculation ........................................... 82 Setting the ISL ........ .................................... 82 Caveats of the slope method ................................. 82

3) PotentialBiologicalRemoval (PBR)

....................... ....... 83 PBR calculations for sea turtles ............................ 83 Application to adult loggerheads

............................................. 84 Caveats of PBR ......................................... 84 Summary of results 1996 - 1998 ....................................... 84 Research recommendations for future evaluation of take limits ................ 85 Literature Cited ....................................................... 85 Part 5. Research Recommendations ................................. 92 Appendix 1. Von'Bertalanffy Growth Models for Wild Kemp's Ridley Turtles: Analyses of the NMFS Miami Laboratory Tagging Database ............................................. 94 Appendix 2. Sea Turtle Strandings ................................. 103 NOAA Tech. Mem. NMFS-SEFSC-444 v

List of Tables Table 1. Nests counted at Rancho Nuevo, North Camp and South Camp 1966 - 1998.

Table 2. Small turtle strandings (20-30 cm SCL) and hatchlings released 1 - 4 years prior to stranding year, 1986-1997.

Table 3. Kemp's ridley strandings, 1986-1997.

Table 4. Kemp's fidley strandings trends.

Table 5. Life stages defined in the age-based model for Kemp's ridleys.

Table 6. Mortality rate parameters defined for the age-based model.

Table 7. Hypothetical changes in hatching rate expected with increasing numbers of nests, as applied to density-dependent projections of the population model.

Table 8. Sum of squared residuals (observed - expected nests) for 3 models fit to various time series.

Table 9. Parameters for model best-fits with small benthic immature z=0.5.'

Table 10. Natal origin of foraging-ground loggerhead sea turtles. Table values are proportional contributions.

Table 11. Loggerhead sea turtle nests in the U.S., 1989-1998.

Table 12. Mean annual percent change in loggerhead nest numbers in the U.S. 1989-1998, based on slopes calculated from In-transformed data.

Table 13. Correlations (Pearson) of loggerhead nests from index beaches at Cape Island, Cape Romain National Wildlife Refuge, S.C., and Hutchinson Island, Fla. with nest totals (excluding each index beach) from each state and region, 1989-1998.

Table 14. Correlations (Pearson) of U. S. loggerhead nests from all states and regions, 1989-1998.

Table 15. Loggerhead strandings by region.

Table 16. Loggerhead stranding trends, vi NOAA Tech. Mem. NMFS-SEFSC-444

Table 17. Trends in loggerhead abundance at the St. Lucie Nuclear Power Plant, St. Lucie, Florida (Quantum Resources, 1995, unpublished data), over various time intervals to compare with nests and strandings.

Table 18. Observed size distributions for loggerheads.

Table 19. Pearson correlation coefficients between loggerhead nests and strandings in the southeastern U. S. and Gulf of Mexico and loggerhead captures atthe St. Lucie Power Plant , 1989-1997.

Table 20. Total shiimping vessels and vessels operating in the Campeche and Tamaulipas fleets, 1982-1995.

Table 21. Hatchling production and strandings levels used to calculate slopes.

Table 22. PBR and ISL estimates for adult loggeirheads, 1998.

Table 23. Summary of Incidental Stranding Limits (ISL) and total strandings for 1996-98.

NOAA Tech. Mem. NMFS-SEFSC-444 vii

List of Figures Figure 1. Nesting beaches of the Kemp's ridley sea turtle. Primary nesting beach is Rancho Nuevo.

Figure 2. Nests and hatchlings released from Rancho Nuevo, North Camp and South Camp, 1966-1999 (hatchlings for 1999 are approximate).

Figure 3. Nests counted at Rancho Nuevo, North Camp and South Camp 1989-1998.

Figure 4. Kemp's ridley nest projection if current trends continue.

Figure 5. Scatter plots of nests counted vs. hatchlings released 8, 10 or 12 years later.

Figure 6. Size frequency distributions of stranded benthic immature Kemp's ridleys in 1986-1987 and 1996-1997, pooled across all statistical zones (Appendix 2).

Figure 7. Best fit model estimates of nests.for 10 years to maturity, Zb = 0.5, with and without the post-1990 multiplier as a fitted parameter.

Figure 8. Model fits for 8, 10 or 12 years to maturity with post-1990 multiplier included.

Figure 9. Changes in best-fit parameter estimates for pelagic immature and large turtle instantaneous mortality with changes in the small benthic immature mortality rate, estimated from a catch curve of strandings.

Figure 10. Effect of small benthic immature mortality estimate on the post-1990 multiplier.

Figure 11. Model estimated population size in 1998, 10 years to maturity model with a range of small benthic immature survival rates.

Figure 12. Model fits and residuals for pre-1990 small benthic immatureZb = 0.53, post-1990 Zb

= 0.34. A. Model with best-fit post-1990 multiplier for large benthic immature and adult turtlesf= 0.52. B. Post-1990 multiplierffixed at 0.65 for all benthic immatures and adults.

Figure 13. Model projections with Zb = 0.5, assuming mortality rates remain constant after 1990.

Figure 14. Model projection of nesting female numbers with and without a reduction in nest survival with nest density (10 years to maturity, Zb = 0.5).

Figure 15. Loggerhead nesting areas, Western North Atlantic.

Vill NOAA Tech. Mem. NMFS-SEFSC-444

Figure 16. Loggerhead nests for the Northern Nesting Subpopulation, 1989-1998. A. Nests counts by state. B. Long-term nesting records for Cape Island, S.C.

Figure 17. Loggerhead nests for the South Florida Nesting Subpopulation, 1989-1998. A. Nests counts by region. B. Long-term nesting records for Hutchinson Island, Fla.

Figure 18. Loggerhead turtles captured by the cooling water intake pipe at the St. Lucie Nuclear Power Plant, St. Lucie, Fla., 1977-1998.

Figure 19. Size distributions and cumulative frequecies for loggerhead turtle strandings. A.

Pooled data for 1986-1987. B. Pooled data for 1996-1997.

Figure 20. Size distributions of loggerheads captured at the St. Lucie Nuclear Power Plant, 1977-1998, represented as proportion of total captures.

Figure 21. Slope method for ISL, showing strandings level required to achieve desired slope of m=0.069.

Figure 22.. Effect of the recovery factor (F) on PBR for a population of 2500.

NOAA Tech. Mem. NMFS-SEFSC-444 ix

Preface At the behest of the National Marine Fisheries Service (NMFS), the Southeast Fisheries Science Center (SEFSC) established The Turtle Expert Working Group (TEWG) in February 1995. The charge to the TEWG was first identified in the Endangered Species Act Section 7 Consultation Biological Opinion for Shrimp Trawling in the southeastern United States under the Sea Turtle Conservation Regulations issued in November 1995. Specifically, NMFS was required to select a team of population biologists, sea turtle scientists, and life history specialists to compile and examine information on the status of sea turtle species. The team should attempt to identify: a) the maximum number of individual sea turtles of each species that can be taken incidentally to commercial fishing activities without preventing the recovery of the species; b) the maximum number of individuals that can be taken incidentally to commercial fishing activities without jeopardizing the continued existence of any. listed sea turtle species; and c) the number of stranded sea turtles occurring in each statistical zone that indicates incidental takes are occurring at levels beyond those authorized. Further clarification was received in a letter of appointment from Mr. Rolland Schmitten, Assistant Administrator, NMFS.

Since the first meeting in June, 1995, there have been, several subsequent formal meetings which have resulted in the publication of status reports for the loggerhead turtle, Caretta caretta, and the Kemp's ridley, Lepidochelys konpii (Turtle Expert Working Group. 1998. An Assessment of the Kemp's Ridley (Lepidochelys konpii) and Loggerhead (Carettacaretta) Sea Turtle Populations in the Western North Atlantic. NOAA Tech. Mem. NMFS-SEFSC-409. 96 pp.) The approaches pursued to evaluate the status and condition of these stocks are considered preliminary and exploratory and we consider this a work in progress. The current report continues with this preliminary stock assessment effort and constitutes an update to the previously published report. It is the TEWG's desire that the results published in these reports be interpreted with caution and within the extent of the assumptions both with the data and with the methodologies applied. We wish to note that the exploratory modeling for the Kemp's ridley was completed only in the context of attempting to understand the life history of this species and to extract population level parameters which duplicate empirical data and numerical trends in abundance. This effort is considered a first step in exploring the use of stage- or age-specific models to estimate and evaluate mortality in sea turtles.

The initial charge to the group included a request to develop a method which would establish take levels especially for the commercial shrimp trawl fishery. Actual removals would be established at levels that would not jeopardize the recovery of these species as stipulated and defined by the individual species' Recovery Plans. Determining the impacts of removal on population recovery in the absence of information on population abundance proved difficult.

The TEWG pursued the use of strandings data in developing annual interim stranding limits (ISL's), recognizing that this proxy for mortality is biased low. These ISL's.were estimated annually for 1996, 1997, and 1998. These levels were then monitored by NMFS on a regional basis for each of these two species. These ISL's do not constitute take levels per se and are simply a proxy for take which provides guidance relative to average strandings v. an unusually X NOAA Tech. Mem. NMFS-SEFSC-444

high incidence either over the course of a year or over a short period of time (a so-called "stranding event"). The pursuit of new methods for estimation of ISL's has beendiscontinued, although the methods used are provided in this and previous reports to NMFS.

It is clear to the TEWG that continued work towards developing estimates of take which do not negatively impact recovery is limited in meaning without a clear understanding of the status and condition of these stocks. It is the latter effort which will be continued. The preliminary modeling work reported here will continue to focus data collection efforts and we are confident that future assessment teams can make progress as more data become available. The intention of this report is to guide research efforts to the completion of empirical population estimates to fully evaluate the status of these stocks and evaluate their condition relative to current and future management activities and recovery goals.

Much hard work and deliberation have resulted in all of these reports to NOAA. I personally want to acknowledge all the contributions of data and hard work that were necessary to finalize this report. Considerable amounts of data were contributed that were not collected by the NMFS and these data cannot be used without the explicit permission of the originator.,

Dr. Nancy Thompson, NMFS/SEFSC NOAA Tech. Mem. NMFS-SEFSC-444 xi

Executive Summary o The Kemp's ridley population appears to be in the early stage of exponential expansion. While the number of females nesting annually is estimated to be orders of magnitude less than historical levels, the mean rate of increase in the annual number of nests has accelerated over the period 1987 to 1999. If assumptions that estimated age to sexual maturity and age specific mortality rates are correct, preliminary analyses suggest that the intermediate recovery goal of 10,000 nesting females by the year 2020 is achievable.

o It appears that the South Florida loggerhead subpopulation, which is the largest in terms of nesting females, is stable or may be increasing; the status of the Florida Panhandle and Yucatdtn subpopulations is unknown. Current trends for the Northern subpopulation indicate this subpopulation has declined since the 1970's and may have stabilized, but it is unlikely that the recovery goals for this subpopulation will be achieved.

o Empirical estimates of the number of turtles that can be removed from the population above natural mortality without preventing or impeding recovery cannot be made at this time with the data available. Stranding limits in the form of Interim Strandings Limits (ISL) were developed for each species and are used as a proxy for mortality. Relationships between strandings levels, stock size, and mortality rates are unknown.

o ISL values initially were derived using a risk averse approach for both species based on multi-year averages of stranding totals for each species. For the loggerhead turtles, 5-year running averages were estimated and, depending on the year or region, standard deviations were added or subtracted to this average to provide the ISL. This approach recognizes the uncertainty of the status of this species and accounts for differences in status between subpopulations. Five-year running averages were estimated for the Kemp's ridley, which were then multiplied by one half the current rate of increase observed in nests. In this way, it is recognized that the rate of increase in strandings for this species should not be greater than the population rate of increase.

Additional methods used in 1998 for ISLs included Potential Biological Removal (PBR) for loggerhead adults and a method for Kemp's ridleys which attempted to match the trend in strandings with the trend in hatchling production over the past 12 years, minus I standard deviation for uncertainty. No method for setting strandings limits was completely satisfactory to all Group members.

o Significant data gaps exist which limit the pursuit of complete age specific assessments for these species and subpopulation-specific assessments for the loggerhead turtle.

Recommendations for research to obtain these data are included. Most notably, for both species empirical estimates of age and age specific survivorship are critical. Continued work to define subpopulations and rates of mixing of subpopulations on foraging areas of the loggethead turtle is also needed.

xii NOAA Tech. Mem. NMFS-SEFSC-444

o Estimates of mortality and growth must be developed and based on empirical data collected in the field. This includes estimates of catch and mortality associated with fisheries and other human activities throughout the range of these species. The only way to develop accurate estimates of catch and mortality in fisheries is to provide observer coverage over a statistically valid portion of the fishing effort throughout the range of these turtles.

NOAA Tech. Mem. NMFS-SEFSC-444 xiii

Turtle Expert Working Group Members (as of October 1998)

Crouse, Deborah Division of Endangered Species U.S. Fish and Wildlife Service 4401 N. Fairfax Dr., Rm 452 Arlington, VA 22203 Crowder, Larry B.

Duke University Marine Laboratory 135 Duke Marine Laboratory Road Beaufort, NC 28516 Epperly, Sheryan P.

Southeast Fisheries Science Center NOAA, National Marine Fisheries Service 75 Virginia Beach Drive Miami, FL 33149 Gabriel, Wendy L.

Northeast Fisheries Science Center NOAA, National Marine Fisheries Service 166 Water Street Woods Hole, MA 02543 Harris, Michael Georgia Department of Natural Resources 2070 Highway 278 Social Circle, GA 30025 Henwood, Terry Pascagoula Laboratory NOAA, National Marine Fisheries Service P.O. Drawer 1207 Pascagoula, MS 39568 Heppell, Selina S.

Duke University Marine Laboratory 135 Duke Marine Laboratory Road Beaufort, NC 28516 xiv NOAA Tech. Mem. NMFS-SEFSC-444

present address:

Department of Fisheries and Wildlife Oregon State University 104 Nash Hall Corvallis, OR 97333.

MArquez-Milan, Rene Programa Nacional de Investigacion de Tortugas Marinas Centro Regional de Investigacion Pesquera Playa Ventanas S/N, A.P. 591 Mananillo, Colima, Mrxico, 28200 Murphy, Sally South Carolina Department of Natural Resources P.O. Box 12559 Charleston, SC 29422 Teas, Wendy Southeast Fisheries Science Center NOAA, National Marine Fisheries Service 75 Virginia Beach Drive Miami, FL 33149 Thompson, Nancy Southeast Fisheries Science Center NOAA, National Marine Fisheries Service 75 Virginia Beach Drive Miami, FL 33149 Witherington, Blair Florida Fish and Wildlife Conservation Commission Florida Marine Resources Department of Environmental Protection 9700 South AlA Melbourne Beach, FL 32951 Observers:

Conant, Therese Office of Protected Resources NOAA, National Marine Fisheries Service 1315 East-West Highway Silver Spring, MD 20910 NOAA Tech. Mem. NMFS-SEFSC-444 XV

MacPherson, Sandra L.

U.S. Fish and Wildlife Service 6620 Southpoint Dr., South Suite 310 Jacksonville, FL 32216 Possardt, Earl U.S. Fish and Wildlife Service University of West Georgia Department of Biology Carrollton, GA 30118 Schmid, Jeffrey R.

Cooperative Institute for Marine and Atmospheric Sciences University of Miami 4600 Rickenbacker Causeway Miami, FL 33149 Schroeder, Barbara Office of Protected Resources NOAA, National Marine Fisheries Service 1315 East-West Highway Silver Spring, MD 20910.

xvi NOAA Tech. Mem. NMFS-SEFSC-444

Part 1. Kemp's Ridley Status Report Summary The Kemp's ridley sea turtle declined precipitously from tens of thousands of nesting females in the late 1940's to around 300 nesting females in 1985. Conservation efforts by Mdxican and U. S. agencies have aided this species by eliminating egg harvest, protecting eggs and hatchlings, and reducing at-sea mortality through fishing regulations. From 1985 to 1999, the number of nests observed at Rancho Nuevo, Tamaulipas, and nearby beaches has increased at a mean rate of 11.3% per year. Current totals exceed 3000 nests per year, allowing cautious optimism that the population is on its way to recovery. At current rates of increase, the population should reach 10,000 nesting females around 2020, the target given in the Kemp's ridley Recovery Plan (USFWS and NMFS, 1992). An analysis of number of nests counted versus hatchlings released suggests that there has been an increase in survivorship from hatchling to maturity in the past 5-6 years. An age-based population model was constructed to calculate potential recovery times,_population size, and reasonable take limits. The model was based on growth curves of recaptured juveniles, a mortality estimate from strandings ofjuvenile turtles, and fitted mortality rates for other life stages. A reduction in mortality of the benthic feeding stages was required to achieve a reasonable model fit; without the mortality reduction, there was no rapid increase in the predicted number of nests in recent years. The model projections suggest that the population will continue to increase rapidly if the mortality estimates are reasonable and remain fairly constant over the next two decades. However, because of uncertainties in the -

mortality rates, the range of population size estimates was large (approximately 8-fold). Thus, the model was unable to adequately calculate the number of turtles in the population nor reasonable take limits. Continued effort to estimate mortality rates through mark-recapture studies may reduce the range of possible population sizes.

History of Nesting at Rancho Nuevo Historic information indicates that tens of thousands of Kemp's ridley sea turtles nested at Rancho Nuevo, M6xico, in the late 1940's (Hildebrand, 1963). Before then and until the late 1960's, egg harvesting occurred and nearly all eggs were removed (R. MNlrquez-M,personal communication). In 1966, a turtle camp was established in Rancho Nuevo which protected nests along approximately 25 km of the beach from April to July of each year. Over the past two decades, poaching on protected beaches has been reduced to nearly zero eggs per year. From 1966-1977, 100-400 nests per year were transplanted to a protective corral while the remaining nests were left "in situ": The "in situ" nests experienced high predation and therefore low egg survival. In 1978, a joint program between the U.S. and M6xico was initiated which increased the length of the beach sampled and number of nests protected. The proportion of nests left in situ was reduced to <10%, thereby greatly increasing survival of the eggs. During the first years of the joint U.S./M~xican conservation effort (1978- 1983), at least 10 aerial surveys of the nesting areas were completed from the U.S M~xico border southward to Veracruz. The observation of Kemp's ridley tracks outside of the Rancho Nuevo area were few and scattered (MWrquez-M, 1994). However, in recent years, more nests were observed north and south of NOAA Tech. Mem. NMFS-SEFSC-444 I

Rancho Nuevo and as a result the nesting beach surveys were extended in 1990 and 1991, respectively (Figure 1). The northern camp, called Tepehuajes/Ostional, covered 45-50 km of beach. In 1991, another camp was established to the south of Rancho Nuevo in Barra del Tordo, increasing total beach coverage to over 100 km. Since 1996 more small camps have been added and now over 120 km of beach is included in this protection, covering nearly all of the state of Tamaulipas to Lechuguillas, Veracruz. Because new surveys added since 1996 may inflate the population growth rate estimate, the Working Group decided to include Rancho Nuevo, North and South camps only in subsequent analyses and model runs. Analysis performed on the nests and hatchlings released from the Rancho Nuevo beach alone rarely altered the results qualitatively, and there is reason to believe that females nesting at Rancho Nuevo moved north and south following a hurricane in 1990 (R. Byles,personal communication).

New Nesting Information Mixico - Update of nest numbers 1966-1998 The nesting data presented in the previous report (TEWG, 1998) for years 1978-1995 were taken from draft reports of the U.S. Fish and-Wildlife Service - Joint U.S.-Mrxico Program. We have updated this information (Table 1). The previous data were considered incomplete, as they were grouped over a season for some camps, and in some cases included data from extra camps. Some minor changes were also made to nesting data from 1966-1977 because of information provided by the Mexican Instituto Nacional de Pesca. Some discrepancies with U.S. Fish and Wildlife Service reports occur because of uncounted nests that were lost by natural.

predation, robbed, or left "in situ" and the exclusion of nests that were not discovered until after hatching and those that were donated to the U.S. Padre Island-Galveston experiments. The current nest tally also omits duplication of numbers of nests already included in the numbers protected in corrals or boxes. The annual number of nests and hatchlings released given in Table I were used in all of the trend and model analyses presented in this report.

Nesting recordsfrom South Texas Isolated Kemp's ridley nests were documented along the south Texas coast from 1948-1994, with a total of thirteen nests being found during this forty-six year period. A primary goal of the joint U.S.-M6xico conservation project was to re-establish a nesting colony at Padre Island National Seashore, and thousands of hatchlings were "imprinted" to that area (Woody, 1985).

From 1995-1998, Kemp's ridley nesting in south Texas showed a marked increase with 4, 6, 9 and 13 nests being documented in each of these years, respectively. Of these thirty-two nests, six were confirmed by tags to have been laid by turtles which were part of the headstfirt experiment and were imprinted at Padre Island National Seashore, seven turtles were examined and found to be "wild" turtles, and the remaining nineteen turtles were not examined for tags. (Shaver and Caillouet, 1998).

2 NOAA Tech. Mem. NMFS-SEFSC-444

Life Stage Definitions and Geographic Locations Post-hatchlingstage We have defined the post hatchling stage as the time the hatchlings enter the water to age 2 years, and have assigned turtles that are 5-20 cm straight-line carapace length (SCL) to this stage. Post-hatchlings disperse from M~xican beaches of the Gulf of Mexico, presumably by major oceanic currents (Collard and Ogren, 1990), and development occurs in the epipelagia.

The presence of juvenile turtles along both the Atlantic and Gulf of Mexico coasts of the U.S.,

where they are recruited to the coastal benthic environment, indicates that post hatchlings are distributed in both the Gulf of Mexico and Atlantic Ocean. Our stage-length of two years is an average based on evidence from three sources: 1) growth curves derived from skeletochronological analysis (Zug et al,. 1997) and mark-tecapture data (Schmid and Witzell, 1997), 2) correlation analysis of hatchlings released and the number of strandings of small turtles (20-30 cm SCL) 1, 2, 3 and 4 years later (Table 2), and 3) preliminary data from 6 dead stranded turtles that had been tagged as hatchlings at Rancho Nuevo and were recovered I and 2 years after release (C. Fontaine, NMFS-Galveston, personal communication). Variability in growth rates among individuals likely means that the actual length of time spent in the post-hatchling pelagic stage varies from 1 - 4 or even more years.

Benthic immaturestage (20-60 cm)

We have assigned turtles that are 20 to 60-cm SCL,to the benthic immature stage, representing the size class that has recruited to thecoastal benthic habitat from the pelagic post-hatchling stage (Ogren, 1989). The 6 stranded turtles originally tagged on the nesting beach ranged from 22-28.5 cm SCL, suggesting that turtles of this size are fully recruited to the benthos (C. Fontaine, NMFS-Galveston,personalcommunication). We further divided this stage into small (20-50 cm SCL) and large benthic immatures (50-60 cm SCL) based on the frequency of strandings used in the mortality estimates (see modeling section below). Benthic immature turtles are found in coastal habitats of the Gulf of Mexico and the entire U.S. Atlantic coast. The location and size classes of dead turtles recovered by the Sea Turtle Stranding and Salvage Network suggests that benthic immature developmental areas occur in many areas along the U.S.

coast and that these areas may change given resource quality and quantity. Growth models from mark-recapture data suggest a 7-9 year duration for the benthic immature stage (Schmid and Witzell, 1997; Appendix 1).

Adult stage Based on size data collected on the Mexican nesting beach since the mid-1960's (Mfrquez-M, 1994 and unpublished data), we have defined adults as turtles larger than 60 cm SCL. The distribution of dead turtles >60 cm SCL is limited to the southeast U.S. coast and the Gulf of Mexico and are essentially absent from strandings records for the northeast U.S. coast.

Recent tag returns indicate that individuals from the Atlantic tagged as immature animals do nest NOAA Tech. Mem. NMFS-SEFSC-444 3

on Mkxican Gulf of Mexico beaches and therefore do move between these water bodies (Schmid and Witzell, 1997; Witzell, 1998).

Status and Trends Nesting data Since 1985, the lowest nesting year (740 total nests), the number of Kemp's ridley nests counted at Rancho Nuevo, North and South Camps has increased approximately 11.3% per year (r-squaie = 0.94, P<0.001, 95% C.I. slope = 0.096 - 0.130; Figure 2). Increase in hatchling production from 1985-1998 was slightly less, 9.5% per year (r-square = 0.955, P<0.001, 95% C.I.

slope = 0.082 - 0.108). We assume that little nesting occurred North and South of Rancho Nuevo prior to 1990 (see History of Nesting at Rancho Nuevo, above), so nests at the North and South camps are included in the totals. Like Rancho Nuevo, the number of nests counted at North and South Camps has increased since census efforts began in 1990 and 1991 (Figure 3).

The annual rate of increase for nests at Rancho Nuevo only, 1985 - 1999, is 7.9% per year (r-square = 0.81, P<0O.001, 95% C.I. slope = 0.056 - 0.102). It is uncertain whether the current rate of increase will continue. Also, the large number of nests counted in 1998 was unprecedented and may or may not indicate an increase in the population growth rate (preliminary data from the 1999 nesting season suggests that fewer nests were observed this year). However, we can project the number of nests with the.confidence intervals predicted by the exponential trend (Figure 4).

Given 2.5 nests per female, if the population continues to grow at 9.6 - 13% per year it will reach the target of 10,000 nesting females around 2014 - 2025.

Nests vs. hatchlings released8, 10, and 12 years earlier When we plotted the number of nests observed against the number of hatchlings released 8, 10 or 12 years earlier, we found an increase in the number of nests per cohort for the past 5-7 years, regardless of time lag (Figure 5, a-c). Heppell et al. (1999) found a similar result when they restricted the analysis to Rancho Nuevo nests only. Several large cohorts produced in the first years of the joint U.S.-M6xico recovery effort do not correspond with larger nesting numbers. While a "build-up" of females is expected with increasing cohort sizes, the graphs suggest that increased hatchling production alone is not responsible for the current increase in nest numbers. An increase in hatchling-to-adult survivorship and/or adult female survival in the late 1980's to early 1990's maybe responsible.

Strandings Complete strandings information by size class, fishing zone and year are provided in Appendix 2. Figure A2. i shows the statistical zones for which sea turtle strandings are reported.

Monitoring effort is not directly comparable between zones but has been reasonably consistent over this period. Figure A2.2 a, b, and c depict the Kemp's ridley strandings reported by statistical zone, by year, 1986-1997. These data do not include incidental captures, post-hatchlings, cold-stunned, or head-started animals. Over the 12-year period 1986-97, the Gulf of 4 NOAA Tech. Mem. NMFS-SEFSC-444

Mexico accounted for 69% of total strandings, with the southeast and the northeast accounting for 25% and 6% respectively (Table 3). Strandings were variable, but lower on average in the 1980's and have increased approximately 70% over this time frame (Table 4). The increase has been confined to the southeast and Gulf of Mexico. Strandings of Kemp's ridleys have not increased in the northeast.

Strandings "hotspots" The western Gulf of Mexico (zones 13-2 1), where shrimp fishing effort is consistently high, has accounted for the highest proportion of Kemp's ridley strandings (Figure A2.2),

particularly of adults. There is no survey effort in zones 15 and 16, due to inaccessibility of shoreline, and coverage is low in zones 13 and 14. The lack of data from these zones may or may not reflect a lack of strandings. Over this time frame, zones 17-21 have accounted for 44.8% of all Kemp's ridleys, including 75% of all adults, stranded over the entire east Coast and Gulf of Mexico. There appears to have been a shift in the strandings of adults over this time frame, with adults being more common in zones 18 and 19 prior to 1992, and more common in zones 20 and 21 from 1994 to the present. In 1993, in zones 13 and 14, a large group of small Kemp's stranded in late May-early June. In the eastern Gulf of Mexico (zones 1-12, partial 24-25), survey coverage is low in zones 1, 3, 6, and 7 due to inaccessibility and zone 2 has very little land mass. Zones 4 and 5 account for most of strandings in the eastern Gulf. Red tide events have also been reported periodically from these zones (Foote et al., 1998).

Along the southeast U.S. coast, monitoring coverage is low in zones 24 and 25.

Shrimping effort is generally high in zones 29-34. Zones 30 and 31 account for the majority of Kemp's ridley strandings in this region. In the noitheastern U.S. Kemp's ridley strandings are rare, except for cold-stunning events in the winter months, which are not included in this analysis. No adult ridley strandings have been reported from this region.

Systematic surveys for sea turtle strandings do not occur in M~xico but a few (<10),

almost always adult-sized Kemp's ridleys strand each year at Rancho Nuevo during the nesting season. In 1998, 30 Kemp's strandings were recorded, but the area being monitored for nesting increased in recent years (from 25 km in the 1980's to 120 km in 1998), so it is unclear whether this represents an actual increase.

Changes in size distributions Kemp's ridleys stranded in 1986-98 ranged from about 10-70 cm. Because the population appears to be growing rapidly, we expected to see a shift in the size distribution towards smaller turtles. We found no evidence of a shift in the mean size of Kemp's ridleys stranded over this period, but the size distribution of stranded Kemp's ridleys was significantly different in the 1996-97 compared to 1986-87 (Kolmogorov-Smirnov Test, P<0.0001; Figure 6).

Many more large benthic immature ridleys stranded in the 1990s relative to the 1980s. While analysis of stranded turtles can potentially provide information about shifts in size distributions, NOAA Tech. Mem. NMFS-SEFSC-444 5

and therefore population trends, strandings may not be indicative of actual population trends because of biases associated with fishing mortality.

Fishery independent, in-water studies of sea turtles have been carried out at multiple sites in the southeastern U.S. Although the objectives and target species of these studies vary, the primary goal has been to characterize regional aggregations of sea turtles. Some studies have been conducted for many years (10+), others have only recently been initiated and still others have been intermittent or aperiodic. Mark-recapture surveys of Kemp's ridleys reveal important information about size distributions and growth rates of immature turtles (Appendix 1). Ogren (1989) first suggested that aggregations of smaller Kemp's ridley turtles were increasing in U.S.

waters as a result of the protection at the nesting beach. However, no quantitative data were available to support this claim. Schmid (1998) noted a higher percentage of smaller size classes captured from 1986-95 when compared to the size distribution of Kemp's ridleys reported in 1955 by Carr and Caldwell (1956). Furthermore, the seasonal mean carapace lengths of Kemp's ridley turtles from the former commercial turtle fishery were 5 - 10 cm greater than those recorded in recent tagging studies. The differences may be indicative of an increased number of smaller turtles in this area or could be the result of size bias in the former turtle fishery. The mean carapace length of Kemp's ridley turtles captured at Cape Canaveral, Florida from 1986-91 (37.0 cm, n = 107, Schmid, 1995) was slightly smaller than that reported from 1978-84 (38.6 cm, n = 40, Henwood and Ogren, 1987). However, this difference may not be significant and could be the result of sampling error owing to dissimilar sample sizes. Reynolds and Sadove (1998) reported no significant change in the mean carapace lengths of Kemp's ridley turtles captured in New York waters from 1986-96. Despite possible temporal shifts in size classes, there are no quantitative data that conclusively demonstrate demographic shifts in the population.

In-water surveys In-water studies targeting Kemp's ridleys either have been or are being conducted at several sites along the U.S. Gulf of Mexico and Atlantic coasts. These sites are: Sabine Pass (Tx.), Calcasieu Pass (La.), Big Bend area (Fla.), Cedar Key (Fla.), Ten Thousand Islands (Fla..),

Pamlico Sound (N.C.), Chesapeake Bay (Va.), and Long Island Sound (N.Y.). There are numerous difficulties in designing in-water studies to ensure that standardized catch per unit effort (CPUE) methods are employed such that data can,be compared not only within site but also site-to-site. Environmental conditions, including weather, depth strata, current patterns, visibility, water temperature, by-catch, and investigator experience, can greatly affect turtle capture rates. Biological conditions, such as prey abundance and habitat availability, complicate sampling strategies that may be based on random distribution of the target species. Changes in effort, variability in temporal sampling, and different sampling methodologies-make analyses of population trends particularlydifficult.

Evaluating trends in abundance based on in-water survey data would be statistically unreliable at this time without some method of standardizing the amount of effort. Because most surveys are restricted spatially or temporally, surveys may monitor local availability at the time of the survey rather than the relative abundance of the population as a whole or a life stage of the 6 NOAA Tech. Mem. NMFS-SEFSC-444

population. High variability in survey indices due to patchy distributions means that any subsequent survey estimates will probably continue to have high variances. Trawl surveys may be a feasible means to monitor status of benthic stages of the population, but are not the only method. The Working Group recommends the design and implementation of a pilot survey, bearing in mind the patchy nature of population distribution.

The only in-water trawl surveys that appeared to have a potential for documenting long-term changes in Kemp's ridley abundance were the NMFS data base from the 1970's and early 1980's (reported in Henwood and Stuntz, 1987) and recent trawl surveys conducted by the Gulf and South Atlantic Fisheries Development Foundation. A comparison of these data was conducted by the Foundation (1998; Jamir, 1999).

There were some differences in survey design between the two studies. The original NMFS observer program was conducted to document the rate of incidental capture and mortality for all turtle species that might encounter shrimp trawls. The survey approach was very simplistic; observers were placed aboard cooperative shrimp trawlers, the vessel fished as it normally would, and all turtle encounters were recorded. The Foundation study was conducted as a direct result of 1996 Congressional language requiring an investigation of alternatives to TEDs.

The Foundation more or less attempted to repeat the original Henwood and Stuntz trawl surveys, with a primary intent of demonstrating that offshore shrimp trawlers did not take turtles and therefore, should be exempt from the TED regulations. The Foundation effort (134 sea days nearshore Gulf, 463 days offshore Gulf and 125 days Atlantic) clearly reflects the emphasis of this study, but the distribution of effort does not preclude some comparisons between the two surveys. The Foundation survey approach used observers to document take and the vessel was asked to fish as it normally would (except for 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> tow time restrictions in nearshore water to reduce turtle mortality).

Preliminary analysis of the two data sets indicated large increases in CPUE for Kemp's ridley and loggerhead turtles when compared to the original Henwood and Stuntz estimates. For Kemp's ridley turtles, as much as a 35-fold increase in CPUE was calculated for some areas. A number of factors must be considered to explain these large changes in CPUE. Most importantly, at the time of the original NMFS work, Kemp's ridley population levels were very low, and the data were 99.9% zeros (tows where Kemp's were not caught) with a few ones where a rare ridley was taken. Also, the fleet was not using TEDs so Kemp's ridleys on the shrimping grounds were probably extremely rare due to high mortality rates. Thus, the CPUE calculations from the late 1970's and early 1980's were composed of rare event captures of a highly depleted stock. The Foundation stiidy, on the other hand, was conducted after almost 10 years of full implementation of TEDs, 30 years of nesting beach protection and 10 Wars of increases in nests and number of hatchlings produced at Rancho Nuevo. These factors should result in an increased number of turtles available for capture, and the Foundation data supported these expectations. Foundation observers were able to catch more turtles with much less effort which resulted in huge changes in CPUE.

NOAA Tech. Mem. NMFS-SEFSC-444 7

While the observed increases in CPUE are impressive, caution should be exercised in interpreting these results. When dealing with small CPUE values (as with the old NMFS estimates Kemp's ridley), any increase in number of turtles caught at a given level of effort or decrease in effort with the same number of turtles caught can result in some "multi-fold" increase in CPUE estimates. The Foundation study indicates that there are now Kemp's ridleys in nearshore waters in measurable quantities, and these turtles were not there or were present in very low abundance 15 years ago. What CPUE will not explain in this case is how much the total population has increased. To illustrate this point consider a comparison of CPUE from an overfished stock where 95% of the fish have been taken from the habitat, with CPUE from that same habitat after ten years of protection has allowed 100% reoccupation of that habitat. Now consider that this habitat is only one of many occupied by this species, the rest of which have never been fished. A twenty-fold increase may have occurred in the overfished habitat, but that tells us nothing about the total population increase.

The Foundation catch rates of sea turtles indicate that future in-water trawl surveys could produce enough captures to allow monitoring of population abundance and trends in a cost effective manner. If such surveys were conducted, a randomized sampling design should be used to cover the entire coastline and not just the shrimping grounds where active fishing is occurring.

This is an important consideration to note because while the NMFS and Foundation data sets allow estimation of CPUE on the shrimping grounds, nothing can be said about turtle abundance in areas or times when shrimp fishing is not occurring.

Trawl surveys are probably the best currently available means of obtaining information on the in-water abundance of sea turtles. The Foundation study, while suffering from less than optimal effort allocation, provides some good insights regarding the effectiveness of TED regulations. The increases in CPUE observed in this study are the strongest evidence to date that Kemp's ridleys are coming back on the shrimping grounds where they were almost eradicated two decades ago. The CPUE values reported are encouraging and believable, although they do not imply multi-fold increases in the total population. What they do suggest is multi-fold increases in areas where the species was severely depleted and kept down as a result of shrimp trawling pressure. Kemp's ridleys now appear to be repopulating their former habitat.

Conclusionsfrom nesting andstrandingsdata There is evidence from nest numbers and in-water surveys that the Kemp's ridley population is increasing. There has apparently been an increase in survivorship to maturity and the proportion of large benthic immatures in the strandings records has increased, suggesting that juvenile mortality has decreased. The number of strandings in the Gulf of Mexico increased sharply in 1994 and 1995 but remained fairly constant (about 450 turtles per year) thereafter. It is reasonable to believe that some of the recent increase in strandings is due to an increase in population size. However, because of the time lag required for each cohort to reach maturity and recruit to the nesting population, there is not a clear connection between changes in strandings or in-water surveys and changes in the number of nests, nor is there an easy way to estimate the 8 NOAA Tech. Mem. NMFS-SEFSC-444-

effect of changes in juvenile fishing mortality on the nesting female population. We attempted to elucidate this relationship using a population model, described below.

Nest protection in Mixico - Futurestrategies Since 1978, nearly all Kemp's ridley nests have been moved to protected "corrals" for incubation, thereby greatly reducing the loss of eggs and hatchlings to erosion and predation. As the population continues to increase and expand north and south, the proportion of nests that can be completely protected will decrease. From 1987 to 1999, the number of nests reported has increased annually from less than 750 to approximately 3400 for Rancho Nuevo, Tepehuajes and Barra del Tordo camps. Based on the relative importance of these camps for nesting and the number of workers and equipment that can be accommodated, it is estimated that the total number of nests that can be protected in corrals is about 5,000. This 5,000 total includes about 4,000 from Rancho Nuevo and about another 1,000 at the north and south camps. Another 1,000 nests could be protected "in situ" with wire screens. Any remaining nests will have to remain where laid with increased protection through increased beach monitoring with additional beach workers. In addition, the number of beach workers required each year is determined by the nesting behavior of the turtles. In some years, the turtles aggregate along specific areas of the beach and fewer beach monitors are required as compared to years when nests are scattered.

While efforts will be made to protect as many nests as possible, it is realistic to anticipate that the current nest survival rate of approximately 65% will not continue indefinitely.

Population size and trends from a deterministic age-based model Model modificationsfrom the 1998 report Models aremost useful when they are updated with new data and modified according to the best information available. The demography of sea turtles is, generally, poorly understood, and new information must be incorporated into model form and parameterization whenever possible. Changes to the model that we developed in 1995 are the results of new information and discussion:

1) The small benthic immature mortality rate is based on a re-analysis of strandings information, using a variety of cohort-based approaches to account for changes in cohort size and variance in strandings levels, rather than pooling year classes. This mortality rate is applied to age classes 2 -

6, based on the growth curve from mark-recapture data (Appendix 1).

2) The mortality rate of large benthic immatures is now equal to that of adults, rather than small benthic immatures, based on the poor fit of the catch curves to strandings of age classes greater than 6 years and the argument that large immature turtles (>50 cm SCL) are likely to have a lower mortality rate (and more similar to the adult mortality rate) than smaller turtles.

3) The number of nests used in the model fitting exercise has been updated according to Table 1.

NOAA Tech. Mem. NMFS-SEFSC-444 9

4) The change in nest survival with nest density has been updated according to recent discussions with the INP and camp managers.

5) The number of years used to calculate the fitted parameters (pelagic immature and adult survival and the post-1990 mortality decrease) has been varied to achieve the model that best-fits both past and recent trends in the nesting data.

These changes have altered the fitted parameters, expected growth rates, and population size estimates of the models. We have done a sensitivity analysis similar to that shown in the 1998 report. These models represent our understanding to date of Kemp's ridley population dynamics, but will undoubtedly change as more information becomes available to run more complex simulations.

Modelform A deterministic age-based model was implemented with the short-term objective of providing a framework to organize and evaluate available data and parameter estimates. The development of a model to track stock status in terms of abundance and mortality at age, to project stock abundance levels conditional on current stock status, and to provide management advice will require reliable estimates of input parameters which are currently unavailable.

The number of female hatchlings produced in each year (Table 1) are projected forward as a cohort. Each cohort passes through four different life stages (Table 5), where age at sexual maturity ( is determinedby a growth curve (Appendix I). All animals aged or higher are assumed to undergo mortality at the rate of mature individuals and reproduce every other year.

The model is used to estimate three parameters through a least-squares analysis that compares nests predicted by the model to nests observed each year (Table 1). Instantaneous total mortality is estimated for two age groups: pelagic immature (Zp), and the combined large benthic immature-adult stages (Z,,). A mortality rate for small benthic immature turtles (Zb) is based on empirical data (strandings, see below) independent of the model. The mortality rate of the large benthic immature stage was assumed to be more similar to that of the adult stage than to that of the small benthic immature stage. This is because large benthic immature animals are rare in strandings in the Gulf of Mexico and are assumed to be distributed in different areas from small benthic immature animals. This makes the large benthic juveniles potentially less vulnerable to mortality than small benthic immature animals. A third parameter reflecting a post-1990 decrease in mortality (/) was also estimated. This multiplier was applied to the small benthic immature and large benthic immature-adult mortality rates after 1990 to address a lack of fit (a trend from maximum negative residual values in 1989 to maximum positive value in 1998) which arises when a single mortality rate per life history phase is applied over the duration of the 1966-1999 time series (Figure 7, in Model Results below). The Excel 7.0 Solver algorithm was used to estimate the parameters by comparing predicted numbers of nests to observed values.

10 NOAA Tech. Mem. NMFS-SEFSC-444

Thus, five different mortality rates are obtained over the time and.age range of the model (Table 6) where Zi, Z., andf are estimated using the model fitting exercise and Zb is input or estimated independently. The post-1990 multiplier (/) is a number less than 1.0, representing a proportional decrease in Zb and Z.. Mortality rates are not re-estimated annually. Observational data consist of annual numbers of nests at Rancho Nuevo and.north and south camps (1966-1999) and annual hatchling production (1996-1998), which serves as cohort size for each year (Table 1). Input parameters included adult female remigration rate, mean number of eggs per nest, female age at maturity, sex ratio in the population, and survivorship of benthic immature turtles (see below). The temporal resolution is annual, with time periods beginning 1 July, corresponding with average time of hatchling production and nesting. The number of adults is modeled without age structure and no maximum age is set.

Starting population sizes were derived from estimates of nests and hatchlings in 1966 (Table 1). Past analyses have shown that the model is relatively insensitive to the nest estimate used in 1966, although this value does affect the model estimate of adult survival (Appendix 2 and 3; TEWG 1998). The estimated number of nests in 1966 was converted to an estimate of nesting females in 1966 by applying rates of nests per female and remigration rate. That pool of nesting females was decremented by adult mortality rates between 1966 and the year the 1966 cohort recruited to the nesting population. This assumes that hatchling production before 1966, which would have supported recruitment to the nesting population in the intervening years, was effectively zero, and consequently that there was no recruitment to the nesting population in 1966-1976.

The number of female hatchlings is based on an assumption of a 1:1 sex ratio for hatchlings produced, and of no other surviving hatchlings from sources, other than monitored nesting beaches at Rancho Nuevo, North and South Camps. The ratio between hatchlings produced and subsequent contribution of a cohort to the nesting population is assumed to be constant for all cohorts recruiting by 1990, because age-specific mortality rates are fixed for all ages over that period, after which the effect of the mortality multiplier is incorporated. Similarly, reproductive parameters, including remigration rates, number of nests per female, and growth rates (i.e., stage lengths and age at maturity) are assumed constant over the 1966-1999 series.

Parameterization Nesting data used We considered the modeled population to include nesting and hatchling production at Rancho Nuevo as well as the North and South Camps.. Prior to the mid 1980s, aerial surveys documented few crawls outside Rancho Nuevo; it seems unlikely that many hatchlings were produced given the small numbers of nests and that there was no protection afforded to the eggs or hatchlings. So we assume that hatchlings estimated at Rancho Nuevo led to nests subsequently detected at Rancho Nuevo and elsewhere. Given that Kemp's ridleys nest numbers reached their lowest numbers in the mid-1980s, one might expect their range to expand with NOAA Tech. Mem. NMFS-SEFSC-444 I1I

increases in the number of females in the nesting stage. Nesting surveys began at North and South Camp in 1990 (Table 1).

In the 1998 report, the model was fit to years 1978-1995, representing all years for which there were complete nest counts. Nesting information is now available through 1999 (Table 1).

We wanted to produce a model that fit the time series well but could also be verified by recent nest counts and used to roughly estimate future population growth. The number of nests observed in recent years is more variable than in the 1980's (Figure 2), and the number of years used to fit the model can affect the fitted survival rates, the post-1990 multiplier, and the model projections.

We ran a sensitivity analysis to see how the addition of each new year of data would affect the model fit for the model fit years (1978-199x), early years (1978-1995), recent years (1996 to 1999) and for the entire available time series (1978-1999) (Table 8). We used Zb = 0.5 for the analysis (Zb does not affect the residuals; TEWG, 1998).

Reproduction For the model fitting exercise, nests deposited per female and years between nestings (remigration interval) were needed to convert nests to adult females, and sex ratio was needed to convert females to total population size. Cohort sizes for years 1966 - 1998 were the number of hatchlings released from Rancho Nuevo, North and South camps, so additional reproductive parameters, such as eggs per nest and hatch success, were not needed until we projected population size into future years. For model projections, we examined the effects of reduced nest survival with nest density, as might be expected as more nests are left in situ (Table 7). We considered published and unpublished information in developing the following assumptions and estimates for reproductive parameters:

1. The number of nests per female per season was calculated as 3.075 in a physiological and ultrasonographic study by Rostal (1991). Pritchard (1990) estimated 2.31 nests per female per season, and a recent tagging study using PIT (passive integrated transponder) tags (R. MNbrquez-M., personal communication) estimated 1.8 nests per female per season: We used 2.5 nests per female per season, which is close to the mean of means (2.4 nests per female per season) of these studies.

2. Adult female remigration rate to the nesting beach was estimated from mark-recapture data (Mdirquez-M et al. 1989). We updated these data with current information (R. M~irquez-M.,

personal communication). Roughly 20% of the turtles nest annually, 60% nest every 2 years, 15% nest every 3 years, and 5% every 4 years. The weighted mean remigration rate was 2.05 years, which we truncated to 2 years and used in the model. However, we recognized that estimation of remigration rate must take into account adult mortality rate (Frazer 1989) and the probability of encountering nesting females (Pritchard, 1990), but neither of these were known.

3. A mean number of eggs per nest of 100.8 was noted by U.S. Fish and Wildlife Service and Natiohal Marine Fisheries Service (1992), which we truncated to 100 eggs per nest.

12 NOAA Tech. Mem. NMFS-SEFSC-444

.4A A

4. For population projections beyond 1999, hatch success (i.e., egg survival) for nests in corrals was set at 65%, based On the mean of [annual hatchlings released/(nests* 100 eggs/nest)] for years 1990-1998 (Table 1). Hatch success was reduced step-wise to simulate a reduction in nest protection with nest density. Protection of nests from predators and poachers at Rancho Nuevo nesting beaches is effort-dependent, and can change with time due to changes in numbers of nests laid and availability of resources to protect them. Under previous plans of the Instituto Nacional de la Pesca (INP) and U.S. Fish and Wildlife Service (USFWS), an agreed upon maximum of 5,000 nests would be protected in corrals, with a maximum of an additional 1,000 nests protected in situ by screening. Nests in excess of 6,000 would be left in natural surroundings with no protection, except perhaps predator control and beach patrols. For our model projections, the hatch rates were imposed on the model according to such best approximation protection scenarios (Table 7).

Sex ratio (all age dasses)

For the model exercise in our first report we assumed a sex ratio of 1: 1 (TEWG, 1998).

Since then two studies in the northern Gulf of Mexico have provided support'for that assumption (Coyne and Landry, 2000; Gregory and Schmid, in review). Although no significant differences were found from a 1:1 ratio across months and sampling locations, females did comprise the majority of turtles sampled in these areas.

Age and growth (see Appendix 1)

Age at maturity for Kemp's ridley turtles.was estimated by Mfirquez-M (1972) as 5-6 years. The U.S. Fish and Wildlife Service and National Marine Fisheries Service (1992) thought that estimate too low, but offered no better estimate. Zug et al. (1997) estimated Kemp's ridley maturity between 11 and 16 based on skeletochronological age estimates for stranded turtles and an assumed mature size of 65 cm SCL. Caillouet et al. (1995) developed a growth model for known-age head-started Kemp's ridleys and estimated a 10 year age at maturity, assuming size at maturity of 60 cm SCL. Schmid and Witzell (1997) estimated age to maturity at 8 - 13 years based on the recapture data of wild subadults in Florida and the minimum and mean carapace lengths of nesting females. An analysis of the NMFS Miami Laboratory tagging database (Appendix 1) indicates an 8-12 year age to maturity estimate using the criteria of Schmid and Witzell. Given the variability in estimates, we assumed the actual age at maturity for Kemp's ridley turtles to be within 7 - .15 years. Because size at first nesting and growth rates vary, we ran preliminary model runs with 8, 10 and 12 years to maturity to test the model's sensitivity to this parameter. We assumed "knife-edge" maturity (all individuals become mature at the given age) because there is no information available on the proportion mature at age or size.

Benthic immature mortality from strandings catch curve estimates (Zb)

Catch curves were constructed both by cohort and by standing stock based on length composition of strandings data and updated growth curves. Strandings at length by month and year were re-indexed for consistency with aging by cohort slicing: the "Ridley year" i was NOAA Tech. Mere. NMFS-SEFSC-444 13

defined from 1 July, calendar year i to 30 June calendar year i+1. An updated von Bertalanffy growth curve (Appendix 1) was used to assign age based on length, assuming a I July birthday.

Year class was assigned by subtracting (integer) age from Ridley year.

To generate catch curves based on cohorts, observations were weighted to adjust. for the potential interannual variability in strandings based on variable effort. A weighting factor for each year was calculated as the ratio of strandings in year x /strandings in 1991, the year of the lowest strandings, and each observation was transformed by dividing by this weighting factor.

Thus, one individual stranding in year 1991 would be reflected as a 1, while an individual stranding in other years would be reflected as a number less than 1, with larger reductions in observations from years with high total strandings.

Weighted observations were then categorized by age and Ridley year, natural log transformed (In), and a catch curve fitted for each cohort. The catch curve began with the age corresponding to the highest left-most point in the ln(frequency)-age plot, and continued to include the oldest age before an inflection point was detected by eye. For most plots, the regression line fit ages 2-6. Older age classes had low and variable catches, so we decided to apply the calculated mortality rate to ages 2-6 only and apply the adult mortality rate (Z,) to immature turtles aged 7 - (a-l).

A similar set of catch curve analyses was generated using standing stock year, with weighting by hatchling production for a cohort, instead of weighting by strandings levels by standing stock year. These analyses gave qualitatively similar results but were considered less reliable because of variance in strandings levels. Overall, the range of Zh was 0.3 to 0.8, with a mode of 0.5.

As outlined in the-Model Form section above, pelagic immature and large benthic immature and adult instantaneous mortality rates (Zp and Za) were obtained through least-squares fit of the number of nests predicted by the model and observed nests. In general, pelagic immature mortality correlated negatively with benthic immature mortality to achieve the "correct" number of nesting females (see sensitivity analysis below).

Model Results Model fits to observed nests We considered 4 factors in the model fitting exercise:

1) age at first reproduction

2) which nesting years to include

3) small benthic immature mortality

4) presence or absence of the post-1990 multiplier.

The lowest sums-of-squares for 1978-1999 were naturally for models fit to the entire time series, but often these models had poor fits for earlier years (Table 8). The lowest sum-of-squared 14 NOAA Tech. Mem. NMFS-SEFSC-444

residuals for models fit to fewer years were: 1978-1998 for the 8 years to maturity model, 1978-1996 for the 10 years to maturity model, and 1978-1995 for the 12 years to maturity model. We used the mortality rates fitted for these years for the model projections. The number of years used to fit the model primarily affected the post-1990 multiplier,f, which ranged from 0.4 to 0.7, and the model's future exponential rate of increase, which ranged from 12-17% per year. Without the post-1990 reduction in mortality for benthic immature and adult turtles, the model cannot fit the observed trend in nests, regardless of age at maturity or the number of years used to fit the model (Figure 7, example is for 10 years to maturity, Zb = 0.5, model fit to nests observed from 1978-1996).

. For simplicity, Figures 8(a, b, and c) show only the best fits and residuals for models with 8, 10 or 12 years to maturity, respectively, and small immature Zb = 0.5. All models shown include the post- 1990 multiplier. The best fit parameters for pelagic immature mortality (Zp),

large immature/adult mortality (Za), and the post-1990 multiplier are different for each age at maturity (Table 9). Recall that the fitting algorithm chooses these three parameters to minimize the sum-of-squares difference between the model nest count and the observed nest count. In order to achieve the rapid rate of increase observed in nest counts over the past decade, the age 12 model requires a lower post-1990 multiplier, corresponding to a more drastic reduction (50%)

in the mortality rate of benthic feeding juveniles and adults. The resulting survival rates and, in turn, the population growth rate after 1990, are higher in the age 12 to maturity model (see Best-fit Model Projections).

Sensitivity of model results to changes in small benthic immature Zb Changes in small benthic immature mortality have virtually no effect on the model fits, because increases or decreases in one mortality rate are compensated for by the others. To illustrate this, we plotted the pelagic immature, small benthic immature and large benthic immature instantaneous mortality rates given by the best fit models for small benthic immature Zb (pre-1990) ranging from 0.3 to 0.8 (Figure 9). While most of the compensation occurs as a trade-off between mortality in the pelagic phase and mortality in the subsequent small benthic immature stage, there is also an increase in the post-I1990 multiplier with increasing small benthic immature mortality (Figure 10). Recall that the multiplier is a proportional change in mortality, so higher multipliers result in smaller changes in annual survival pre- and post- 1990.

The mortality compensation results in a very similar predicted number of nests, but very different population size estimates (Figure 11). Because we are trying to fit three parameters with the current model, there are a large number of feasible solutions to the mortality schedule. Thus, until empirical estimates can be derived for the post-1990 multiplier and pelagic immature, large benthic immature and adult mortality, it would be imprudent to use this model to estimate population size.

Model fits with an empirically-derived estimate off We estimated the post-1990 change in small benthic immature mortality by comparing the strandings catch curves for cohorts that stranded before and after 1990. If pre- and post-1990 NOAA Tech. Mem. NMFS-SEFSC-444 15

estimates of Zb are included in the model, (Zb = 0.52 and 0.34, respectively), the fit and model outputs are similar to the small benthic immature mortality= 0.5 run with 10 years to maturity (Figure 12 a). The model predicted a greater post-1990 reduction in large immature and adult mortality (f= 0.52) than the reduction estimated empirically by the catch curve coefficients of small benthic immatures (0.34/0.52 = 0.65). It is not known if this is the result of imprecision of Zb estimates or if there is a larger post-1990 mortality decrease for larger size classes of turtles.

When the post-1990 multiplier was fixed at 0.65, the model had a lower fit and biased residuals (sum-of-squared nest residuals 1978-1999 = 791,064; Figure 12 b).

Best-fit model projections In spite of the large differences in total population size, the model runs with small benthic immature Zb ranging from 0.3 to 0.8 showed a relatively small range of predicted population growth rates, measured as annual proportional increase and assuming that mortality rates are constant through time (range 12 - 16% per year). In general, models with lower Zb (-

higher annual survival for 2-6 year-olds) predicted slower population growth rates (these models also had less reduction in mortality after 1990). Because of their similar post-1990 survival rates, the age 8 at maturity and age 10 at maturity models showed similar population trajectories (Figure 13). The age 12 at maturity model, with its higher post-I 990 survival rate estimates, predicted a faster growth rate despite the later age at first reproduction. The time required to reach 10,000 nesting females varied somewhat in the sensitivity analysis (range 12 - 16 years for the age at maturity = 10 years model). A reduction in hatch success as the nest population increased greatly reduced the expected population growth rate for all of the models, but not until after the 10,000 nesting female mark was reached (example shown is 10 years to maturity, pre-1990 small benthic immature Zb = 0.5; Figure 14). These model projections are for comparative purposes only - they are not meant to forecast population size through time. The models do not include environmental variance and we do not know which, if any, of our models most accurately matches the mortality schedule and life cycle of the Kemp's ridley. As more information on survival and growth is obtained from field studies, we will be able to reduce the number of unknowns and produce more predictive models.

Literature cited Caillouet, C. W., Jr., Fontaine, C. T., Manzella-Tirpak, S. A., and Williams, T. D. 1995.

Growth of head-started Kemp's ridley sea turtles (Lepidochelys kempii) following release.

Chelonian Conserv. Biol. 1:231-234.

Carr, A. and D. K. Caldwell. 1956. The ecology and migrations of sea turtles: 1. Results of field work in Florida, 1955. Amer. Mus. Nov. 1793:1-23.

Collard, S. B. and Ogren, L. H. 1990. Dispersal scenarios for pelagic post-hatchling sea turtles.

Bull. Mar. Sci. 47:233-243.

16 NOAA Tech. Mem. NMFS-SEFSC-444

Coyne, M. S. and A. M. Landry. 2000. Plasma testosterone sexing criteria and sex ratio of the Kemp's ridley sea turtle (Lepidochelys kenpii). U.S. Dep. Commer. NOAA Tech. Mem.

NMFS-SEFSC-436:286 Foote, J. J., N.C. Park, and J. M. Sprinkle. 1998. An increase in marine turtle deaths along the central coast of Florida (1995-1996); is red tide the culprit? U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-415:169-171.

Frazer, N. B. 1989. Nesting cycles in sea turtles: typical, but not cycles. U.S. Dep. Commer.

NOAA Tech. Mem. NMFS-SEFC-232:61-64.

Gregory, L. F. and J. R. Schmid. in review. Stress responses and sex ratio of wild Kemp's ridley turtles (Lepidochelys kempi) in the northeastern Gulf of Mexico. Gen. Comp. Endocrinol.

Gulf & South Atlantic Fisheries Development Foundation, Inc. 1998. Alternatives to TEDS:

final report. Report to NOAA, Contract 50WCNF606083, Tampa, Fla., 19 pp.

Henwood, T. A. and L. H. Ogren. 1987. Distribution and migrations of immature Kemp's ridley turtles (Lepidochelys kempi) and green turtles (Chelonia mydas) off Florida, Georgia, and South Carolina. Northeast Gulf Sci. 9:153-159.

Henwood, T. A. and W. E. Stuntz. 1987. Analysis of sea turtle captures and mortalities during commercial shrimp trawling. Fish. Bull. 85:813-817.

Heppell, S. S., L. B. Crowder and T. R. Menzel. 1999. Life table analysis of long-lived marine species with implications for conservation and management. Amer. Fisheries Soc. Symp.

23:137-148.

Hildebrand, H. H. 1963. Hallazgo del area de anidacion de la tortuga marina "lora" Lepidochelys kempi (Garman) en la costa occidental del Golfo de M~xico. Ciencia 22:105-112.

Jamir, J. V. C. 1999. Revisions to the estimates of incidental sea turtle capture aboard commercial shrimp trawling vessels. Unpublished report. Gulf & South Atlantic Fisheries Foundation, Inc., Tampa, Fla. 17 pp.

M~rquez-M, R. 1972. Resultados preliminares sobre edad y crecimiento de ]a tortuga lora, Lepidochelys kempi (Garman). Mem. IV Congd. Nac. Ocean. (M6xico):419-427.

Miirquez-M., R. 1994. Synopsis of biological data on the Kemp's ridley sea turtle, Lepidochelys kemnpi (Garman, 1880). U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-343, 91 pp.

NOAA Tech. Mem. NMFS-SEFSC-444 17

Mdirquez-M., R., A. Villanueva-O., and P. M. Burchfield. 1989. Nesting population and production ofhatchlings of Kemp's ridley sea turtle at Rancho Nuevo, Tamaulipas, M~xico, p. 16-19. In: Caillouet, C. W., Jr. and A. M. Landry, Jr. (Eds.), Proceedings of the First International Symposium on Kemp's Ridley Sea Turtle Biology, Conservation and Management, Texas A & M University, Sea Grant College Program, TAMU-SG-89-105.

Ogren, L. H. 1989. Distribution of juvenile and subadult Kemp's ridley turtles: results from the 1984-1987 surveys. P. 116-123. In: Caillouet, C. W., Jr. and Landry, A. M., Jr. (Eds.),

Proceedings of the First International Symposium on Kemp's Ridley Sea Turtle Biology, Conservation and Management. Texas A&M University Sea Grant College Program, TAMU-SG-89-105.

Pritchard, P. C. H. 1990. Kemp's ridleys are rarer than we thought. Marine Turtle Newsletter 49:1-3.

Reynolds, D. P. and S. S. Sadove. 1998. Wild captures of sea turtles in New York: species composition shift examined. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-415:254-256.

Rostal, D.. C. 1991. The reproductive behavior and physiology of the Kemp's ridley sea turtle, Lepidochelys kempi (Garman, 1880). Ph.D. Diss., Texas A&M University, College Station, Tx., 138 pp.

Schmid, J. R. 1995. Marine turtle populations on the east-central coast of Florida: results of tagging studies at Cape Canaveral, Florida, 1986-1991. Fish. Bull., U.S. 93:139-151.

Schmid, J. R. 1998. Marine turtle populations on the west-central coast of Florida: results of tagging studies at the Cedar Keys, Florida, 1986-1995. Fish. Bull. 96:589-602.

Schmid, J. R. and W. N. Witzell. 1997. Age and growth of wild Kemp's ridley turtles (Lepidochelyskenipi): cumulative results of tagging studies in Florida. Chelonian Conserv. Biol. 2:532-537.

Shaver, D.J. and C.W. Caillouet, Jr. 1998. More Kemp's Ridley turtles return to south Texas to nest. Marine Turtle Newsletter 82:1-5.

Turtle Expert Working Group. 1998. An assessment of the Kemp's ridley (Lepidochelys kempii) and loggerhead (Carettacaretta)sea turtle populations in the Western North Atlantic.

U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-409, 96 pp.

U.S. Fish and Wildlife Service and National Marine Fisheries Service. 1992. Recovery plan for the Kemp's ridley sea turtle (Lepidochelys kemnpii). National Marine Fisheries Service, St.

Petersburg, Fla., 40 pp.

18 NOAA Tech. Mem. NMFS-SEFSC-444

Witzell, W. N. 1998. Long-term tag returns from juvenile Kemp's ridley turtles. Marine Turtle Newsletter 79:20.

Woody, J. B. 1985. International efforts in the conservation and management of the Kemp's ridley sea turtle (Lepidochelyskenpi). p. 1-3. In: C. Caillouet and A. Landry (Eds.),

Proceedings of the First International Symposium of Kemp's Ridley Sea Turtle Biology, Conservation and Management. Texas A&M University at Galveston, TAMU-SG 105.

Zug, G. R., Kalb, H. J., and Luzar S. J. 1997. Age and growth on wild Kemp's ridley sea turtles Lepidochelys kempii from skeletochronological data. Conserv. Biol. 80:261-268.

NOAA Tech. Mem. NMFS-SEFSC-444 19

Table 1. Nests counted at Rancho Nuevo, North Camp and South Camp 1966 - 1998. Nests and hatchlings released are approximate for 1966 - 1977. Total nests includes estimates (1966 -

1977) or counts (1978 - 1998) of nests not moved to protective corrals. Total hatchlings =

hatchlmigs released from corrals only. **Hatchling count not yet determined for 1999.

Year Rancho North Camp South Camp Total Nests Total Hatchhngs Nuevo Hatchlings Released per Released Nest 1966 5991 5991 29100 4.9 1967 5519 5519 24100 4.4 1968 5117 5117 15000 2.9 1969 4018 4018 28400 7.1 1970 3017 3017 31400 10.4 1971 2012 2012 13100 6.5 1972 1824 1824 14600 8.0 1973 1642 1642 23500 14.3 1974 1466 1466 23500 16.0.

1975 1266 1266 11100 8.8 1976 1100 1100 36100 32.8 1977 1036 1036 30100 29.1 1978 1012 1012 48009 47.4 1979 1019 1019 63996 62.8 1980 900 900 37378 41.5 1981 962 962 53282 55.4 1982 842 842 48007 57.0 1983 795 795 32921 41.4 1984 861 861 58124 67.5 1985 740 740 51033 69.0 1986 752 752 48818 64.9 1987 742 742 44634 60.2 1988 840 0 3 843 62218 73.8 1989 905 11 0 916 66802 72.9 1990 800 204 0 1004 74339 74.0 1991 862 178 74 1114 79749 71.6 1992 932 254 84 1270 92116 72.5 1993 847 251 101 1199 84605 70.6 1994 1126 329 110 1565 107687 68.8 1995 1438 334 149 .1921 120038 62.5 1996 1255 444 219 1918 114842 59.9 1997 1514 358 347 2219 141770 63.9 1998 2409 431 642 3482 167105 48.0 1999 2308 273 776 3357 ** **

20 NOAA Tech. Mem. NMFS-SEFSC-444

Table 2. Small turtle strandings (20-30 cm SCL) and hatchlings released 1 - 4 years prior to stranding year, 1986-1997.

Hatchlings released Year Strandings 1 year earlier 2 years earlier 3 years 4 years 1986 105 51033 58124 32921 48007 1987 30 48818 51033 58124 32921 1988 24 44634 48818 51033 58124 1989 27 62218 44634 48818 51033 1990 28 66802 62218 44634 48818 1991 25 74339 66802 62218 44634 1992 24 79749 74339 66802 62218 1993 94 92116 79749 74339 66802 1994 121 84605 92116 79749 74339 1995 40 107687 84605 92116 79749 1996 41 120038 107687 84605 92116 1997 38 114842 120038 107687 84605 Correlation 0.2603 0.4047 0.1419 0.2833 Table 3. Kemp's ridley strandings, 1986-1997.

1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Total N.E.U.S. 27 12 24 10 II 12 20 25 25 16 14 16 212 S.E.U.S. 32 51 116 43 94 48 30 79 95 91 114 92 885 Gulf of 246 98 75 96 153 105 106 223 430 288 255 304 2379 Mexico Total 305 161 215 149 258 165 156 327 550 395 383 412 3476 NOAA Tech. Mem. NMFS-SEFSC-444 21

Table 4. Kemp's ridley strandings trends. Percent of total strandings represented by each location is shown in parentheses.

1986-88 total .1995-97 total Percent change N.E. U.S. 63 (9%) 46 (4%) -27%

S.E. U.S. 199 (29%) 297 (25%) +49%

Gulf of Mexico 419 (62%) 847 (71%) +102%

Total 681 1190 0.75 Table 5. Life stages defined in the age-based model for Kemp's ridleys.

Life stage Component ages Pelagic immature 0 through I 2 through 6 Small benthic immature 7 to (-1) (e.g., 9 if 10)

Large bent+ic immature Adult+ (e.g., 10+)

22 NOAA Tech. Mem. NMFS-SEFSC-444

Table 6. Mortality rate parameters defined for the age-based model.

Stag Pre- 199 Post-199(

Pelagic immature Z_ Z Small benthic immature (age 2-6' Zh J*Z_

Large benthic immature Za f*Z_

Adul Z: f'Z' Table 7. Hypothetical changes in hatching rate expected with increasing numbers of nests, as applied to density-dependent projections of the population model.

Range in nests Protection regime  % survival to hatchling release 0-5,000 Corral 65 5,001-6,000 in situ with screening 50 6,001-10,000 in situ with predator control 30

>10,000 in situ with no predator 20 control NOAA Tech. Mem. NMFS-SEFSC-444 23

Table 8. Sum of squared residuals (observed - expected nests) for 3 models fit to various time series. Small benthic immature Zb ` 0.5.

8 years to maturity fit to: model fit yrs 1978-1995 1996-1999 1978-1999 (1978-199x) 1995 126,679 126,670 2,149,482 2,276,151 1996 209,839 185,056 429,312 614,368 1997 235,948 221,965 712,188 934,153 1998 526,370 166,135 428,933 595,069 1999 561,440 171,175 390,264 561,440 10 years to maturity fit to: model fit yrs 1978-1995 1996-1999 1978-1999 (1978-199x) 1995 60,823 60,823 1,354,351 1,415,174 1996 86,669 77,980 360,795 438,776 1997 89,642 85,029 435,837 520,867 1998 214,790 79,748 742,805 822,554 1999 429,941 73,829 356,111 429,941 12 years to maturity fit to: model fit yrs 1978-1995 1996-1999 1978-1999 (1978-199x) 1995 161,174 161,174 589,762 750,937 1996 165,800 163,847 923,858 1,087,706 1997 166,084 164,770 997,803 1,162,573 1998 447,024 208,158 720,433 928,592 1999 578,015 165,934 412,081 578,015 24 NOAA Tech. Mem. NMFS-SEFSC-444

Table 9. Parameters for model best-fits with small benthic immature z=0.5.

Pelagic immature Large benthic immature Post-1990 Age at maturity annual survival and adult annual survival multiplier (I)

(age 7+)

pre-1990 post-1990 pre-1990 post-1990 8 0.24 0.24 0.85 0.91 0.56 12 0.37 0.37 0.86 0.92 0.50 10 0.31 0.31 0.85 0.91 0.56 10 -w/out post- 0.38 0.38 0.83 0.83 1.0 1990 multiplier NOAA Tech. Mem. NMFS-SEFSC-444 25

Figure 1. Nesting beaches of the Kemp's ridley sea turtle. Primary nesting beach is Rancho Nuevo Lou iisi an a 1'*Z

. 0 100 200 300 400 600 KiJorr

- ,Nesting Survey Areas GUL 7 1. Padre Island OF

2. North Camp MX 2 3. Rancho Nuevo 0 '- 3 4. South Camp

>4 5. Veracruz

> "6. Campeche C-Y

~Yuc U 5

I In arl ds c (

.A---'

z0 Figure 2. Nests and hatchlings released from Rancho Nuevo, North Camp and South Camp, 1966-1999 (hatchlings for 1999 are approximate).

z m

7000 250,000 6000 200,000 5000 150,000

• 4000

[ -: Nests - hatchlings Z 3000 100,000 2000 50,000 1000 0 0 aD 00 C) (D aDO C) (10 aDO (N (D wD (D( CD -l r- ir- I'-_ N.- 0D 0D 00 00 CDO) 0) 0)0a) M

0) 0) 0)0) 0) 0)0)0M0) 0) 0)0M0) 0) CM0) M NJ,

N) 00 Figure 3. Nests counted at Rancho Nuevo, North Camp and South Camp 1989-1998.

2500 U] [

• Rancho Nuevo 2000 y= 6 71. 1 1 eO.1343X p2 = 0.8818 j-U 1500 A North Camp CD

0) (Tepehuajes) 1000 U © .. __- y = 160.17e0,1614x z U U ----- " R2 = 0.903 C

__-4 0 South Camp 500 H

fl (Barra del Tordo)

C) y = 54.629e0 .2214X R2 = 0.8852 0

1991 1992 1993 1994 1995 1996 1997 1998 1999 7

~11 f-fl CA m

CO cm A

A A

Figure 4. Kemp's ridley nest projection if current trends continue.

zT 0ý 35000 -I-I I 30000 2 0I -- " ------

250004 ---------------------------------- 1-------------- It

/

/ I 10,000 nesting 20000 --

- ----- - -- - -- - - -- --- - -- - - - --- - -7 ---- -- -/- - - - - -

~females if 2.5 z / nests/female SNests observed

ý-------------------- -------;-----------

500---

1---

/

o-~ ~ ~

/

/--

-- -- -f ------ ----- lower 95% CI 10000~~

4 slope = 9.6% per

/ If year 5000 -- - upper 95% CI slope = 13% per year o 1r _W 19~85 1995 2005 2015 2025 t'Q

Figure 5. Scatter plots of nests counted vs. hatchlings released 8, 10 or 12 years later. Numbers denote nesting year.

A. 8year lag 4000 3500 3000 2500 U, *97 2000 *95 go z

1500

• 74

• 94

+76 75 0 +93

• 8 2 1000 L*9

• 82,*5 8 +87 500 0

0 10,000 20,000 30,000 40,000 50,000 80,000 70,000 80,000 90,000 B. lOyear lag 4000 3500

+ 98 *99 3000 2500

• 97 a, 2000 091t 95 1500 *94

• 92 +91 1000 #7s*93*

• go tW4 , uo + 88 *89

+8t*tJH 500 0

0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 Hatcblings C. 12year lag 4000 3500 499 *98 3000 to 2500 U, "97

-, 2000 z . 95 *96 1500 .94

• 92 1000 84 *79 04+.*82`' 88 go9 993 93 * *91 i

+888 500 U

0 *oflo0 20,000 30 POO 40 J3oo 50)300 60)00 70)300 Hatchlings 30 NOAA Tech. Mem. NMFS-SEFSC-444

0z Figure 6. Size frequency distributions of stranded benthic imnature Kemp's ridleys in 1986-1987 and 1996-1997, pooled across all statistical zones (Appendix 2).

H C

C)

C)

A. 1986-1987 pooled strandings R 1996-1997 pooled strandings z

'-Ti 100 100% 120 100%

90 0

-Il 80 80% C 100 80% 2ET CA 0 70 60 C

. 60 60%

C- C-4,60 8 P 50 L 40 LL 40%

30 40 20%

o40%C 20 20% h) 20 MD 10 0 0% 0 E CC)

E E E E E E E '-E E E E E E E E E E E E E E E C) CD) U-) CD C) L CD LO) CD LC) CDU) CD C) CD C) CD U-) CD LD C)

I n 0D L CD In CD In) In CD CD In C ) LO c D LC) sD y n CIDn CnD(C - -l Cl Cr 0') 0) :T Inr In (C (0 (D CN(7- r) C'-) 'C '-I - C3IC ) (0 (0

W Figure 7. Best fit model estimates of nests -for 10 years to maturity, Zb = 0.5, with and without the post-1990 multiplier as a fitted parameter.

7000 -

6000 - -- Observed 5000 - 4. & model fit with post-1990 mortality reduction

-model fit without post-i 990 4000- mortality reduction Z 3000 2000 z0 1000

-3 C

C 0 SI I I 1 I I I I I I I I I I I I I I I I I I 1 1 i I I I I I I I CD B

1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 z

-ri CA CA m

-t CA cm

Figure 8. Model fits for 8, 10 or 12 years to maturity with post-1990 multiplier included.

A 8yearsto rturity Residuzls- 8 yeas to maturty 3503- 600)---------------

3D03 - ObtG-enedrests It' ----------

25CC 2 bdelttt C 13 z 15Ct- W 0--------

0 -- -- -- -- -- -- -- -- ---

.. 30 0 o a a a 'a a a a 'a a

• 50 et,dft B. l0yearsto naturity Residuals - 10 years tosaaturity 8000 450C3-V 000---------------------------

36rC0 - Olsertet rests A a400--------------------- -----

2550D 10000 *200-----------------------

2i0c , ..- - I -Id C

1500-

.'200--------------------------

loci) 500

-400. .. ....

0-

c. 12 yearstorrAturity Residuals - 12 years tomatuuity 200 -

i[10 ... .

- ...- Ouse Medkest C 2M -- A--tmA 0 ' C ""

150D - V 0 urn ---- t Sal aaa a' a" a...a/ a* a a r- a 2co NOAA Tech. Mem. NMFS-SEFSC-444 33

Figurc 9. Changes in best-fit parameter estimates for pelagic immature and large turtle instantaneous mortality with changes in the small benthic immature mortality rate, estimated from a catch curve of strandings.

1.6 pelagic imm.

0 1.2 j---------------------- -------------- ----------------------

E 0

0.8 - -- - -- -- -- -- -- -- -- -- -- - -- - -- -- - -- ------- -* -sm benthic imm.

CD 0.4 ------

0L X'-- - -x---- x---- X-.- ...

-- X -

-- -x ------ x ....

---... X .... X -- x--Ig benthic 0 I I I I I . I I I 1 1 imm.

0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 and zT Small benthic immature mortality (ages 2-6) adults 0l

Figure 10. Effect of small benthic immature mortality estimate on the post-1990 multiplier.

zi 0.65

'4-0.6 CA C)

E 0.55 0

0) 0)

0.5 C-0.45 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 Small benthic immature mortality (ages 2-6)

Figure 11 . Model estimated population size in 1998, 10 years to maturity model with a range of small benthic immature survival rates. Small benthic immatuies should be read with the right axis.

10000 120,000 N 100,000 8000 -

80,000 CL C. 6000 - L------------------------------------------------c 0

C- _

X

- 60,000 -- -- adult E 4000 IU 40,000 - i-Ig z0 "n W

benthic 2000 20,000

- 7, -sm benthic 0 0 Cr' 0.3 0.4 0.5 0.6 0.7 0.8 Small benthic immature mortality (ages 2-6) cm)

Ak 4A

z Figure 12. Model fits and residuals for pre-1990 small benthic immature Zb = 0.53, post-1990 Zb = 0.34. A. Model with best-0 fit post-1990 multiplier for large benthic immature and adult turtlesf= 0.52. B. Post-1990 multiplierf fixed at 0.65 for all

1:..

benthic immatures and adults.

H 0

0 0

A. Residuals 4000 z 3500 800 -- - - - - --

~'~1 3000 0 0 2500 0* 400

~-r) <D2000 -- nests Wbsserved x z .... A .. Model fit 1500 "u 200 0

1000 500 0 -200 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 1978 1981 1984 1987 1990 1993 1996 1999 B. Residuals 4000 3500 800 3000 2500

'0 2000 nEsts observed Z .... Model fitO 1500 , 200 -- - - - - - - - - - - - -- - - -

1000 500 0 01 S-200 an CD CN Q0 CC) 0D N D co r-cc co O C 0 0 M M ) 0)

M M 0M 0 0) 0) 0) 0 0) 0 " 0) 0)M 1978 1981 1984 1987 1990 1993 1996 1999

00 Figure 13. Model projections with Z,, = 0.5, assuming mortality rates remain constant after 1990.

40,000 35,000

-10 years to maturity 30,000

- - - - - - - - - - - - - - - - - - - - - - - - - ------------ ---- r

-- i-- 12 years to E 25,000 maturity 40-0) 20,000 .... ...8 years to maturity z ---------- 3*

15,000 ------- -- -- --

10,000 z

5,000 H

C) 0 II I I I I I I I I I I I I I I I I I I I M 00 C) N4 I- (D 0 a) 0 0*-- (N

.T-- .t-(0 T*-

CO T-0 N

0) o0 00 C

0) 0 o 0 0 0 0- 0 0 0 0, oN oI CN (N ("N (N (N N (N (N (N It O00

.0 r.n

z0 Figure 14. Model projection of nesting female numbers with and without a reduction innest survival with nest density (10 years to maturity, Zb = 0.5).

H z"

30,000 -

25,000 -

ct V."

c-fl . 1 cm? CD

.A ,i 20,000 - -- 4---nest 4A E - - :---- * - ---- - survival all I-0* 15,000 - years =

U,- 0.65 a,

z 10,000 -----

---

- -e

--- ý - - - - - - - - - - - - --- nest survival

-- - - -- - - -- - - -W-

---

- - - -- - - -- - - -- - - -- - - decreases 5,000 with density 0 I I I I I I I I I I I I I I I 1 I I 1 I CO 0 N q (0 cO 0 N  ;:" (,0 CO 0

0) 0 0 0 0 0. T- T-r- I N o)

IT*- 0N", 0N4 0CN N N', 0C 0C\,I -\40 0N', 0C4 0N N 00'4

Part 2. Loggerhead Status Report Summary The loggerhead sea turtle was listed under the Endangered Species Act as threatened in 1978. At the time of listing the species in the North Atlantic was considered to be a single population, but genetic evidence acquired since has indicated the existence of distinct subpopulations which form the basis of management units. The trends of each subpopulation must be assessed separately and the subpopulations may need to be listed separately under the ESA to afford the smaller or declining subpopulations more protection. The TEWG recommends that the life history stages of each loggerhead subpopulation be monitored, and acknowledges that the mix of subpopulations on the foraging grounds makes population-specific assessments complicated. In-Water surveys must be conducted to provide data to evaluate the changes in the age structure of the subpopulations.

Stock Definition and Geographic Range The loggerhead sea turtle (Carettacaretta)occurs throughout the temperate and tropical regions of the Atlantic, Pacific, and Indian Oceans (Dodd, 1988). Its range of habitat includes open ocean waters, continental shelves, bays, lagoons, and estuaries. Loggerheads in the Western North Atlantic nest on high-energy, sandy beaches between the latitudes of 18' and 350 North. Worldwide, nesting on warm temperate beaches is much more common than nesting in the tropics. In the western North Atlantic there are at least 4 genetically distinct loggerhead nesting subpopulations, based on mitochondrial DNA (Bowen et al., 1993; Bowen, 1995; Encalada et al., 1998) (Figure 15)ý and possibly there are more (Francisco et al., 1999):

1. The Northern Nesting Subpopulalion, occurring from North Carolina to northeast Florida, about 290 N. (approx. 7500 nests in 1998).

2. The South Florida Nesting Subpopulation, occurring from 290 N on the east coast to Sarasota on the west coast (approx. 83,400 nests in 1998). This is the largest loggerhead nesting assemblage in the Atlantic and is the second largest in the world (Ross, 1982).

3. The Florida Panhandle Nesting Subpopulation, occurring at Eglin Air Force Base and the beaches near Panama City, FL (approx. 1200 nests in 1998).

4. TheYucatAn Nesting Subpopulation, occurring on the eastern Yucatfin Peninsula, Mexico (Mftrquez-M, 1990) (approx. 1000 nests in 1998).

Historically, nesting has been minor elsewhere in the western North Atlantic, except in Honduras, Nicaragua, Columbia, Venezuela, Bahamas, Cuba, Jamaica, and Puerto Rico (Sternberg, 1981). The present size of these nesting assemblages is unknown and individuals from these nesting areas have not been assayed for genetic affinity. As a whole, nesting by the 40 NOAA Tech. Mem. NMFS-SEFSC-444

loggerhead subpopulations of the northwest Atlantic account for more than 35-40% of the global nesting activity of this species (Ross, 1982).

Cheloniid sea turtle species worldwide have highly structured nesting assemblages (Bowen et al., 1992; Bass, 1994; Bowen et al., 1994; Broderick et al., 1994). Maternally mediated gene flow based upon mtDNA analysis among identified populations is low (Bowen et al., 1993; Bowen, 1995). Natal homing is a dominant force shaping this phylogeographic structure. The nesting assemblages probably represent distinct genetic entities, but other research, including analysis of nuclear DNA (inherited from both parents) is necessary to address stock definition. Analysis of nuclear DNA of green turtles in both the Pacific and in the Atlantic found moderate rates of male-mediated gene flow (Karl et al., 1992), but still, populations within ocean basins were structured. Nevertheless, given the high site fidelity of nesting females to their natal beach and low gene flow between nesting assemblages, most Western North Atlantic loggerhead nesting assemblages are vulnerable to extirpation. Should an assemblage be extirpated, regional dispersal will not be sufficient to replenish the depleted nesting assemblage within thousands of years. This has been demonstrated amply through the loss of assemblages worldwide; they have not been re-established (e.g. Bermuda, Cayman Island, Alto Velo). Thus, both genetic evidence (mtDNA) and the historical record indicate that recolonization of extirpated nesting assemblages does not occur on contemporary time scales (Bowen et al., 1994).

For conservation and management, mtDNA data indicate that nesting assemblages must be considered independent demographically. This conclusion holds even if males should prove not to be philopatric to their natal site, because the production of progeny depends on female nesting success. For this reason, we use the term "subpopulation" to describe the nesting assemblages, consistent with the IUCN definition of the term.

Like other sea turtles, loggerhead post-hatchlings occur in pelagic habitats for a number of years, eventually settling into Coastal habitats as benthic immatures, at approximately 40-60 cm SCL in the U.S. (Carr, 1987). Genetic analyses indicate that turtles from several subpopulations mix on foraging grounds. Table 10 summarizes the estimated proportional contributions of the four identified subpopulations to foraging-ground assemblages, as well as the proportional contributions of Mediterranean haplotypes. The proportional contributions from Chesapeake Bay, South Carolina, Georgia (in part), and the Mediterranean were determined prior to the completion of a more thorough survey of Atlantic nesting beaches and may therefore change by 10% or more, but the qualitative conclusion of mixtures on these foraging grounds remains valid (B. Bowen, personal communication). Because South Florida is the largest Nesting Subpopulation contributing to these foraging grounds, the samples are often dominated by turtles of this haplotype. However, the results of these analyses, conducted at multiple foraging sites along the Atlantic seaboard, indicate that mixing of subpopulations is not random and that the contributions of the larger South Florida Subpopulation increase along a general north-south gradient (northeast U.S. to Florida Bay). Analyses of the genetic identities of loggerhead turtles inhabiting foraging grounds of the Gulf of Mexico are underway but not yet available. Preliminary results from analysis of loggerheads stranding in Texas indicate the Northern Subpopulation may account for up to 10% of the animals stranding there, but most NOAA Tech. Mem. NMFS-SEFSC-444 41

come from the S. Florida Subpopulation, along with some from the Yucatdtn, Mexico Subpopulation (A. Bass, personalcommunication).

Life stage definitions We have defined the life history stages for loggerhead turtles as: pelagic immature from the time a hatchling leaves the nesting beach until it appears on coastal benthic feeding grounds at around 40-60 cm straight carapace length (SCL); benthic immature from settlement to 92 cm SCL; and adult which is 92 cm SCLV. Pelagic immature turtles originating from Atlantic nesting beaches have been observed in the eastern Atlantic near the Azores and Madeira (Maigret, 1983; Carr," 1986, 1987; Bjorndal et al., 1994; Bolten et al., 1994) and in the Mediterranean Sea (Laurent et al., 1998). A growth curve derived from a length-frequency analysis of pelagic immatures in the Azores indicates that loggerheads may stay in the North Atlantic Gyre for 7-12 years, depending on .the size of recruitment to the benthic stage (Bjorndal et al., 2000). It has been suggested that turtles that originate from Gulf of Mexico beaches also utilize the North Atlantic Gyre and are distributed to it via the Loop Current of the Gulf of Mexico and Florida Current (Collard and Ogren, 1990). Strandings records indicate that when pelagic immature loggerheads reach 40-60 cm SCL they recruit to coastal inshore and nearshore waters of the continental shelf throughout U.S. Atlantic and Gulf of Mexico. Benthic immatures have been found from Cape Cod, Massachusetts, to southern Texas, and occasionally strand on beaches in northeastern Mexico (R. Mdrquez-M.,personal communication). Large benthic immature loggerheads (70-91 cm) represent a larger proportion of the strandings and in-water captures (Schroeder et al., 1998) along the south and western coasts of Florida as compared with rest of the coast, but it is not known whether the larger animals actually are more abundant in these areas or just more abundant within the area relative to the smaller turtles. Benthic immature loggerheads foraging in northeastern U.S. waters are known to migrate southward in the fall as water temperatures cool (Epperly et al., 1995; Keinath, 1993; Morreale and Standora, 1999; Shoop and Kenney, 1992), and likely migrate northward in spring. Given an estimated age at maturity of 21-35 years (Frazer and Ehrhart, 1985; Frazer et al., 1994), the benthic immature stage must be at least 10-20 years long. The adult life history stage, 92 cm straight carapace length, has been reported throughout its U.S. range and throughout the Caribbean Sea. Nesting occurs primarily from North Carolina southward to Florida with assemblages recorded in the Florida Panhandle and on the Yucatdin Peninsula (Figure 15). Non-nesting females are reported from the entire area and little is known about the distribution of adult males which are seasonally abundant near nesting beaches during the nesting season.

'Appendix 4 of our first report (TEWG, 1998) details how we arrived at the decision to use 92 cm SCL to distinguish benthic immature and adult turtles. It was based on an arithmetic mean which may be an inappropriate measure of central tendency in non-normally distributed data, which some nesting beach data appear to be. Also, we have not considered the possibility that the average size of neophyte nesters maybe subpopulation-specific.

42 NOAA Tech. Mem. NMFS-SEFSC-444

Status and Trends Nesting beaches Northern Subpopulation The addition of nesting data from the years 1996, 1997, and 1998, did not change the assessment of TEWG that the number of loggerhead nests in the Northern Subpopulation is stable or declining (Figures 16 and 17). The annual number of recorded nests for the period 1989-1998 ranged 4370-7887 nests/year (Table 11). No trends are detectable for North Carolina, South Carolina or Georgia during that period; nests in northeast Florida may be increasing, although the data are too variable to detect a significant trend (Table 12). A longer time series of nesting data for a portion of the Northern Subpopulation comes from Cape Island at the Cape Romain National Wildlife Refuge, South Carolina (Figure 16B). There, the number of recorded nests for the period 1975-1998 ranged 579-2654 nests/year. Nesting at Cape Island has declined an average of -2.7% per year over the 24-year period (r-square=0.284, P = 0.007), but nearly all of that decline occurred during the 1970's. Since 1989, the Cape Island nest numbers have varied annually with no significant trend (1989-1998 slope = -1.4% peryear, r-square = 0.023, P =

0.675). Annual nesting at Cape Island was found to correlate well with nesting in North Carolina, South Carolina (remaining beaches), Georgia and the Northern Subpopulation as a whole for the period 1989-1998, but did not correlate as well with nesting activity of the South Florida and Florida Panhandle Subpopulations (Table 13). The residuals from the regression model correlated well with the residuals from northern nesting beaches and southeast Florida, but not the SW Florida nor the Panhandle residuals.

South Florida Subpopulation The addition of nesting data from the years 1996 through 1998 did not change the assessment of TEWG that the number of loggerhead nests in the -South Florida Subpopulation is stable or increasing. The annual number of recorded nests for the period 1989-1998 ranged 48,531-83,442 nests/year (Table 11), with an average rate of increase over the time series of 3.6% per year (r-square = 0.461, P = 0.031, Table 12). The time series suggests a 4-year cycle in nesting, although the time series is too short to confirm this (Figure 17A). A longer time series of nesting data for a portion of the South Florida Subpopulation comes from Hutchinson Island, Florida (Figure 17B). There, the number of recorded nests for the period 1981-1998 ranged 3121-8214 nests/year and had a mean rate of increase of 4% per year (r-square = 0.746, P <

0.001). In more recent years (1989-1998), Hutchinson Island nests have increased 2.7% per year but the trend is not significant because of data variability (r-square= 0.308, P = 0.096). Variation in nest counts at Hutchinson Island seem to support the 4-year cycle hypothesis for recent years but not earlier years in the time series (Figure 17B). The nest count was found to correlate extremely well with nesting in the remaining parts of the southeast Florida region for the period 1989-1998, but did not correlate as well with the nests counted at northern beaches nor those in the Florida Panhandle (Table 13). However, the residuals from the In-transformed regression correlated fairly well with the residuals from all areas except the Panhandle.

NOAA Tech. Mem. NMFS-SEFSC-444 43

Florida Panhandle Subpopulation The number of loggerhead nests recorded for the Florida Panhandle Subpopulation is increasing, but much of this increase is due to expanded beach monitoring. The length of beach monitored in the Panhandle increased from 186.5 km in 1989 to 359.9 km in 1997. Thus, the status of this Subpopulation cannot be determined at this time.

Nest count correlations across states and regions Table 14 gives the Pearson correlation coefficients for loggerlead nest counts in all U.S.

areas. A second set of correlations was calculated with In-transformed data residuals to remove the linear trend of each time series. State nest counts within a nesting subpopulation generally correlate well with each other and the subpopulation totals. However, the three nesting subpopulations are less well correlated with each other over the 10 year time period. Nests in North Carolina and Georgia may follow the 4-year pattern observed in southeast Florida (Figures 16 and 17), but South Carolina nests are much more variable. The Florida Panhandle Subpopulation is poorly correlated with the other two subpopulations, likely because much of the recent increase in nest counts for that area are due to an increase in monitoring effort. However, the Panhandle nests and SW Florida nests are highly correlated.

Yucatin nesting subpopulation There are few nesting surveys for loggerheads in Mexico. There are 8-12 turtle camps operated each year in central Quintana Roo on the Yucat6in Peninsula. In 1998, 1052 nests were recorded at 12 camps on Quintana Roo beaches (Xcaret, 1999). Nesting of this Subpopulation is believed to extend into Belize and nesting appears to be stable or increasing (R. Mdrquez-M.,

personalcommunication).

Loggerhead nesting elsewhere in the Gulf of Mexico Nesting of loggerheads throughout the Gulf of Mexico is sporadic outside of Florida and few areas conduct. systematic surveys of nesting beaches (S. MacPherson, personal comtnunication). The first estimate of nesting in the Gulf of Mexico outside of Florida indicated that between 1960 and 1962 a reproductive assemblage of approximately 100 loggerheads nested on the remote beaches of the'offshore islands of Louisiana, Mississippi, and Alabama (Ogren, 1978). Aerial surveys in 1977 of the U.S. Gulf beaches west of Florida indicated that nesting had declined (Ibid). In recent years about 21-31 nests were found annually at Bon Secour National Wildlife Refuge in Alabama (1994-1998) and a couple are found each year on Dauphin Island (1995-1997) (Mays and Shaver, 1998; S. MacPherson,personal communication). Up to 9 nests werefound annually at the Gulf Islands National Seashore in Mississippi 1990-1998 and less than 10 nests were found annually at the Breton National Wildlife Refuge in Louisiana (1989-1990) (Ibid). There are few records of loggerheads nesting on the Texas Coast: 18 since 1977 (D. Shaver, personal communication). Eleven of the 18 recorded nests occurred in 1996-1998 and all but one was on the southern Texas coast. Loggerheads also nest occasionally in the 44 NOAA Tech. Mem. NMFS-SEFSC-444

Mexican states of Tamaulipas, Campeche, and Yucatdin. During this decade the annual nest counts for these areas were: 1-5 nests at Rancho Nuevo and Ostionales in Tamaulipas, about 50 nests in Campeche, and about 100 nests in Yucatdin (R. Mdirquez-M.,personal communication).

The genetic identity of these scattered nests is unknown.

Strandings Complete strandings information by size class, statistical zone and year are provided in Appendix 2. As with the analysis of strandings of Kemp's ridley turtles, the loggerhead strandings used by the TEWG excluded incidental captures, post-hatchlings, cold-stunned or head-started animals. Figure A2.1 in Appendix 2 shows the statistical zones for which sea turtle strandings are reported. Monitoring effort is not directly comparable between zones but has been reasonably consistent over this period. Figure A2.3 a, b, and c depicts the loggerhead strandings reported by statistical zone, by year, 1986-1997. There is no survey effort in zones 15 and 16, due to inaccessibility of shoreline, and coverage is low in zones 13 and 14. In the eastern Gulf of Mexico (zones .1-12, partial 24-25), survey coverage is low in zones 1, 3, 6, and 7 due to inaccessibility and zone 2 has very little land mass. The lack of data from these zones may or may not reflect a lack of strandings. Along the southeast U.S. Atlantic coast, coverage is also low in zones 24 and 25. In the northeastern U.S. Atlantic, survey coverage is less rigorous.

However, high human densities along the coast in this area suggests most strandings will get reported. This is not true for inshore waters, such as the Chesapeake Bay and Pamlico and Core Sounds of North Carolina, where strandings likely go unreported.

Trends Table 15 shows the loggerhead strandings by legion. Over the 12-year period 1986-97, the southeast accounted for 63% of the strandings totals, with the Gulf of Mexico and northeast Atlantic accounting for 21% and 15%, respectively. Strandings in the southeastern U.S. and the Gulf of Mexico declined in the early 90's, particularly 91-93, but have increased since then. The combined total for 1995-97 was 13% higher than the combined total for 1986-88 (Table 16). In particular, strandings in the northeast more than doubled over that time frame, increasing the relative importance of mortality in the northeast (from 10 to 19%). The northeast is the only region which has shown significant trends throughout the 1986-1997 period. However, since 1991 all three areas have shown significant increasing trends in strandings, 10-14% per year.

Overall, loggerhead strandings showed an average annual increase of 10.5% per year for the period 1991-1997.

Strandings "hot spots" While loggerhead strandings occur throughout the eastern U.S., the largest proportion of them occur along the southeast U.S. Atlantic coast (Figure A2.3, Appendix 2). Strandings are high throughout the area, with Zone 30 highest in most years and Zone 26 and the southern segment of Zone 36 lower in most years. Adults strand throughout the region. Strandings are lower in the Gulf of Mexico. Zones 4 and 5 have somewhat higher stranding numbers; of NOAA Tech. Mem. NMFS-SEFSC-444 45

particular concern is the proportion of adult loggerheads stranding in these zones. Zones 18, 19 and 20 also have higher strandings but with few adults. In the northeast U.S. Atlantic loggerhead strandings are concentrated in Zones 36 and 37 with fewer adults.

Based on NMFS estimates shrimping intensity is much higher in the western Gulf of Mexico than in the east. It is concentrated nearshore in the .west; more than half the shrimping effort is the 0-1 8m depth zone. In the eastern Gulf, effort peaks in the 18-37m zone (>60% of the effort occurs at these depths) (Nance, 1992, 1993; McDaniel et al., 2000). Loggerhead strandings are consistently highest in zones 4-5 (West Florida) and 18-20 (Texas). Strandings in West Florida include a high proportion of adults whereas strandings in Texas are exclusively juveniles. Given the relatively low shrimping effort in the eastern Gulf, the impact on loggerheads (particularly adults) appears high.

Systematic surveys for sea turtle strandings do not occur in Mexico but a few (<10) loggerhead strandings have been recorded at Rancho Nuevo in recent years. These have been primarily large benthic immatures (>63 cm CCL), except for one 95 cm CCL animal in 1997.

TED effect analyses Biweekly strandings data from South Carolina and Georgia (1980-97) were analyzed to determine the significance and magnitude of TED effects in reducing strandings (Royle and Crowder, 1998). The methods used were similar to those applied to the S.C. data (1980-93) in which strandings were shown to be reduced by 44% (Crowder et al., 1995). TEDS significantly reduced strandings over the period 1980-97 by an estimated 40% in South Carolina and 58% in Georgia relative to strandings estimates without TEDs (Royle and Crowder, 1998). A significant "TED effect" on strandings is detectable through the time series analysis of biweekly data in spite of the increasing trend in annual strandings. Recent analysis of the Georgia data utilizing Poisson regression and taking into account shrimp landings indicates strandings in Georgia (per unit shrimp catch) were reduced 37% when TEDs where in use compared to periods when TEDs were not in use (Royle, 2000).

In-water surveys Fishery independent, in-water studies of sea turtles have been carried out at multiple sites in the Gulf of Mexico and U.S. Atlantic. The primary goals and objectives of these studies vary, and the species targeted vary as well. Some studies have been conducted for many years (10+),

others have only recently been initiated and still others have been intermittent or aperiodic. There are numerous difficulties in designing in-water studies to ensure that standardized catch per unit effort (CPUE) methods are employed such that data can be compared, not only within a site but also site-to-site. Environmental conditions including depth strata, current patterns, visibility, water temperature, and investigator experience can greatly affect catch rates. Biological conditions, such as prey abundance, complicate sampling strategies that may be based on random distribution of the target species. Changes in effort, variability in temporal sampling, and different sampling methodologies make analyses of population trends particularly difficult. At 46 NOAA Tech. Mem. NMFS-SEFSC-444

this time, only two studies enable relatively long-term comparisons of CPUE on a within study site basis2 . These two sites are the nearshore waters off the south central east coast Florida at the St. Lucie Nuclear Power Plant (Quantum Resources, 1995, unpublished data) and the Indian River Lagoon along the central east coast of Florida at Sebastian Inlet (Ehrhart et al., 1996). The TEWG does not feel that changes in CPUE evaluated with recent data from the naked net survey conducted by the Gulf & South Atlantic Fisheries Development Foundation, Inc. (1998; Jamir, 1999) indicate a real 8-fold increase in population size in the southeast U.S; because of the differences in sampling as discussed in the Kemp's report above, but that it does indicate there has been an increase of an unknown magnitude.

Ehrhart et al. (1996) analyzed loggerhead CPUE data from his central Indian River Lagoon study site for three combined sampling periods (1983-1985, 1988-1990, and 1993-1995).

The turtles were captured by gill nets strung across the lagoon, With CPUE measured as turtles.

caught per kilometer net hour. Ehrhart concluded that although there was some variation between the three spans of years, CPUE was not statistically different (Kruskal-Wallace = 1.244, df= 2, P = 0.5369). Captures of loggerheads at the Indian River Lagoon site are immature turtles and hence the CPUE statistics would not be affected by influxes of adult turtles during the nesting season.

The intake pipe of the St. Lucie Nuclear Power Plant captures sea turtles each year, depositing them relatively unharmed in an enclosed lagoon. The species and size of each turtle is recorded, after which the turtles are tagged and released. Because the intake pipe operates continuously, the capture effort may be considered relatively constant. Turtles of all size classes have been caught, but it is not known whether the St. Lucie turtles repiesent an unbiased sample of the nearby turtle population. Over the entire time series (1977-1998), the exponential trend in total loggerheads captured at St. Lucie has a slope of 4.2% per year (r-square = 0.40, P = 0.002)

(Table 17). However, the early part of the time series shows little overall growth and a potentially cyclical pattern in abundance (Figure 18). In recent years, the trend has increased to 17% per year (r-square = 0.76, P = 0.005), although the number of loggerheads captured in 1997 dropped dramatically. The TEWG (1998) reported significant increases in numbers for adult loggerheads( 92 cm SCL) at St. Lucie which track increases in the nesting population, as determined by nesting surveys. These trends continue when post- 1995 data (through 1998) are included (Figure 18, Table 17). Increased numbeis of benthic immature turtles captured in 1996 and 1998 resulted in a significant increasing trend for these turtles as well, which was not detected in the 1977-1995 time series due to data variance (TEWG, 1998). Adult and immature captures were generally well correlated across years (Pearson's r 0.65, P<0.05), without evidence of a time lag.

2Recent examination of SEAMAP trawl data collected by South Carolina Department of

. Natural Resources indicates that those data may also be a source of trend information. The extensive trawl survey is based on a stratified random sampling design in nearshore waters between Cape Hatteras and central Florida and has been conducted seasonally since 1989.

NOAA Tech. Mem. NMFS-SEFSC-444 47

The two studies which provide the longest and most consistent time series of data do not agree in their assessments of trends in the population of benthic immature loggerheads. Because juvenile abundance and capture probability are highly variable, it will be difficult to detect trends without long time series from multiple sites. Both of the surveys discussed here consist of samples that are likely to be dominated by the South Florida Subpopulation, thereby giving little indication of trends in benthic immature loggerheads originating from the Northern Subpopulation. A survey of benthic immature turtles in Pamlico and Core Sounds, North Carolina, will provide more information regarding trends in the different subpopulations. More comprehensive surveys are needed to monitor the status of benthic stages of the population and multiple, index area studies that yield statistically valid CPUE data are urgently needed. Trawl surveys may be one method of accomplishing this goal. The Working Group recommends the design and implementation of a pilot trawl survey, bearing in mind the patchy nature of population distribution (see Research Recommendations).

Sex ratio Studies of loggerheads along the Atlantic coast suggest that these populations exhibit a female bias. Laparoscopy of 47 turtles examined during fall 1997 in North Carolina found 68%

were female and 32% were males (Braun-McNeill et al., 2000). Serum androgen sexing of turtles captured in waters 20 '.C in the same area, 1996-1998 indicated 72% female, 27% male, and 1% unknown (n= 114) (Braun-McNeill et al., in press). Wibbels et al. (1987) reported a female bias (66%), based on serum testosterone titres, across Atlantic coast sites sampled during the 1980's: Chesapeake Bay (52%), Cape Canaveral (60%), Hutchinson Island (74%), and Indian River (58%). It can be speculated that female-biased sex ratios are likely given the predominance of turtles from the South Florida Nesting Subpopulation, whose nest incubation temperatures are generally higher than the pivotal temperature required for an even sex ratio in loggerheads (Mrosovsky, 1988).

Size distributions Mean size of stranded loggerheads 1986-87 vs. 1996-97 The mean size of benthic immature loggerheads that stranded from 1986-1997 varied from region to region with the largest turtles in the eastern Gulf (about 80 cm SCL) and somewhat smaller turtles elsewhere (about 60-70 cm SCL). One might expect that a growing population would experience a shift in size distribution towards smaller turtles. We found no evidence of a shift in the mean size of loggerheads stranded over ihis period, but the size distribution of stranded loggerheads was significantly different in 1996-97 compared with 1986-87 (Kolnogorov-Smirnov Test, P<0.0001) (Figure 19). Benthic juvenile strandings in the 55-65 cm SCL size range increased in abundance in the 1990s relative to the 1980s.

48 NOAA Tech. Mem. NMFS-SEFSC-444

Size distributions from the St. Lucie Power Plant The size composition of loggerheads captured at the St. Lucie nuclear power plant has varied from year to year with some long-term trends (Figure 20). In the 1980's, the proportion of small immatures captured decreased, becoming as low as 50% in 1989 and 1990. Since 1991, this proportion has stabilized at around 70%. The proportion of adult loggerheads in the capture records reached a peak of 17% in 1990 and has fluctuated since 1991 (5-13.5%). The average proportions for 1991-1998 are 69% small immatures, 22% large immatures and 9% adults.

Populationsize estimates Nesting data collected on indexnesting beaches 1989-1998 represent the best dataset available to index the population size of loggerhead sea turtles. These data provide annual estimates of the number of nests laid each year and indirectly provide estimates both of the number of females nesting in a year (based on an average of 4.1 nests per nesting female, Murphy and Hopkins (1984)) and of the number of adult females in the entire population (based on an average remigration interval of 2.5 years, (Richardson et al., 1978). The total number of nests laid along the U.S. Atlantic and Gulf coasts, 1989-1998, has ranged from 53,016-89,034 annually, representing, on average, an adult female population of 44,780 (Table 11) [(nests/4.1)

• 2.5]. On average, 90.7% of the nests were from the South Florida Subpopulation, 8.5% were from the Northern Subpopulation, and 0.8% were from the Florida Panhandle Subpopulation.

The ratio method used by TEWG (1998) to estimate minimum population size of post-pelagic loggerhead tuitles in nearshore U.S. waters required assumptions about the accurate representation of loggerhead size classes in the stranding data. A recent, cursory appraisal of capture data and of TED design information leads the TEWG to believe that this assumption may not hold true. Data from the naked-net study conducted by the Gulf & South Atlantic Fisheries Development Foundation, Inc. (1998; Jamir, 1999) show a significantly smaller (Kolomogrov-Smirnov Test, P<0.005) proportion of larger loggerheads (large immatures and adults) than appears in the stranding data (Table 18). A size bias in the stranding data might be expected if large loggerheads are less likely to be excluded by TEDs than small loggerheads and because of this, they incur higher mortality and aie more likely to strand. A-recent examination of loggerhead morphometrics and TED exit opening dimensions shows that loggerheads greater than approximately 70 cm may not be readily excluded by TEDs (Epperly and Teas, 1999). Also, nesting female loggerheads that occur in nearshore habitats may be more prone to incidental capture and repeat captures, thereby increasing their mortality rate and probability of stranding.

In light of this information, we have chosen not to recalculate population size estimates using the ratio method.

Conclusionsfrom nest numbers, strandings,and in-water surveys In a species with along age to maturity, such as the loggerhead, nesting trends alone may give an incomplete picture of population status. It is conceivable for a population with no new recruitment to the benthic juvenile stage to continue to show increases in nesting for a number of NOAA Tech. Mem. NMFS-SEFSC-444 49

years as benthic juveniles from past cohorts mature. Conversely, a population could continue to show declines in nesting over time due to losses of adults while the immature population is increasing. Thus multiple lines of evidence must be considered in order to determine true population status.

The loggerhead recovery plan established recovery goals primarily in terms of nesting activity. While the loggerhead recovery plan was drafted before the mitochondrial DNA evidence of distinct nesting subpopulations became available, goals in terms of numbers of nests per season were given for North Carolina, South Carolina and Georgia, roughly encompassing the range of the Northern Subpopulation, with a goal for Florida of an increasing adult female population.

Taken together, the data reviewed by the TEWG would suggest that the South Florida Subpopulation of loggerheads is increasing, while no trends are evident in the Northern Subpopulation in recent years. Long term trends at Hutchinson Island, an index nesting beach for the South Florida Subpopuiation, has increased by approximately 4% per year since 1981. While stranding data and trends cannot be attributed to subpopulations, the size of the South Florida Nesting Subpopulation taken together with the genetic evidence that animals from this Subpopulation disperse along the entire Atlantic coast suggest that it predominates in the strandings. Strandings have increased in recent years and there is evidence of a shift in size class distribution toward smaller turtles, supportive of an increasing population. Again, the in-water data are not conclusive, but apparent increases at the St. Lucie Power Plant, where in 1999 an estimated 69% were from the South Florida Subpopulation (Witzell et al., in review), also support the increasing population hypothesis. Apparent increases in nesting in the Florida panhandle area are confounded by increases in monitoring, making interpretation uncertain.

Correlation analysis of the overlapping time series of available data (1989-1997) suggest that, in general, nest counts and strandings do not correlate well (Table 19). High correlation coefficients do occur between time series that are rapidly increasing (e.g., Panhandle nests and northeast strandings), and these correlations disappear when the residuals are compared. The notable exception to the lack of correlation is between Northern nests and adults stranding in the Southeast Region (r = 0.822, P=0.01 1), a relationship that holds in the residuals of the In-transformed data (r = 0.822, P = 0.007). Adult strandings in this region were also well correlated with nest counts in each .ofthe 4 Northern Subpopulation States. Further analysis of strandings correlations by size were not possible because a much higher proportion of stranded turtles in other regions are reported without measurements (SE region reporting average is 52.5%, range 45-59%). Southeast strandings and Gulf of Mexico strandings were well correlated (r=0.762, P=0.0 17) but failed to correlate with northeast strandings. St. Lucie captures of adults and immatures correlated fairly well with South Florida nests and Southeast strandings, although the correlations were not significant at the P<0.05 level (raw data fiurn the St. Lucie captures also correlated with Panhandle nests, but the residuals did not correlate). As more years of data are collected, more detailed time series analyses may reveal which sources of loggerhead population data correlate through time and are likely to be reasonable proxies for population size distributions and trends.

50 NOAA Tech. Mem. NMFS-SEFSC-444

Literature Cited Bass, A.L. 1994. Conservation genetics of hawksbill turtles, Eretmochelys imbricata, in the Caribbean and western Atlantic. M.S. Thesis, Louisiana State University, Baton Rouge, 55 pp.

Bass, A.L., S.P. Epperly, J. Braun, D.W. Owens, and R.M. Patterson. 1998. Natal origin and sex ratios of foraging sea turtles in the Pamlico-Albemarle Estuarine Complex. U.S. Dep.

Commer. NOAA Tech. Mem. NMFS-SEFSC-415:137-138.

Bass, A. L., S-M. Chow, and B. W. Bowen. 1999. Temporal variation in loggerhead strandings from Georgia. Final report to Georgia Department of Natural Resources. Unpublished report. Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, 14 pp.

Bjorndal, K.A., A.B.. Bolten, J. Gordon, and J.A. Camifias. 1994. Carettacaretta(loggerhead) growth and pelagic movement. Herp. Rev. 25:23-24.

Bjorndal, K.A., A.B. Bolten, and H.R. Martins. 2000. Somatic growth model of juvenile loggerhead sea turtles: duration of the pelagic stage. Mar. Ecol. Prog. Ser. 202:265-272.

Bolten, A.B., K.A. Bjorndal, and H.R. Martins. 1994. Life history model for the loggerhead sea turtle (Carettacaretta)populations in the Atlantic: Potential impacts of a longline fishery. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SWFC!201:48-55.

Bolten, A.B., K.A. Bjorndal, H.R. Martins, T. Dellinger, M.J. Biscoito, S.E. Encalada, and B.W.

Bowen. 1998. Transatlantic developmental migrations of loggerhead sea turtles demonstrated by mtDNA sequence analysis. Ecol. Applic. 8:1-7.

Bowen, B.W. 1995. Tracking marine turtles with genetic markers. BioScience 45:528-534.

Bowen, B.W., J.C. Avise, J.I. Richardson, A.B. Meylan, D. Margaritoulis, and S.R.

Hopkins-Murphy. 1993. Population structure of loggerhead turtles (Carettacaretta)in the northwestern Atlantic Ocean and Mediterranean. Sea. Conserv. Biol. 7:834-844.

Bowen, B.W., N. Kamezaki, C.J. Limpus, G.R. Hughes, A.B. Meylan, and J.C. Avise. 1994.

Global plylogeography of the loggerhead turtle (Carettacaretta)as indicated by mitochondrial DNA haplotypes. Evol. 48:1820-1828.

Bowen, B.W., A.B. Meylan, J.P. Ross, C.J. Limpus, G.H. Balazs, and J.C. Avise. 1992. Global population structure and natural history of the green turtle (Chelonianiydas) in terms of matriarchal phylogeny. Evol. 46:865-881.

NOAA Tech. Mem. NMFS-SEFSC-444 51

Braun-McNeill, J., S.P. Epperly, D.W. Owens, and R.M. Patterson. 2000. Sex ratios of foraging sea turtles in the Pamlico-Albemarle Estuarine Complex, North Carolina, U.S.A. U.S.

Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-436:121-122.

Braun-McNeill, J., S.P. Epperly, D.W. Owens, and R.M. Patterson. in press. Sex ratios of immature sea turtles: does water tempeature make a difference? In: Proceedings of the 1 9 "hAnnual Sea Turtle Symposium, U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC.

Broderick, D. C. Moritz, J.D. Miller, M. Guinea, R.J. Prince, and C.J. Limpus. 1994. Genetic studies of the hawksbill turtle: evidence formultiple stocks and mixed feeding grounds in Australian waters. Pac. Conserv. Biol. 1:123-131.

Carr, A. 1986. Rips, FADS and little loggerheads. BioSci. 36:92-100.

Carr, A. 1987. New perspectives on the pelagic stage of sea turtle development. Conserv. Biol.

1:103-121.

Collard, S.B. and L.H. Ogren. 1990. Dispersal scenarios for pelagic post-hatchling sea turtles.

Bull. Mar. Sci. 47:233-243.

Crowder, L.B., S.R. Hopkins-Murphy, and J.A. Royle. 1995. Effects of turtle excluder devices (TEDs) on loggerhead sea turtle populations. Copeia 1995:773-779.

Dodd, C. K., Jr. 1988. Synopsis of the biological data on the loggerhead sea turtle Caretta caretta (Linnaeus 1758). U.S. Dep. Int. Fish Wild]. Serv. Biol. Rep. 88(14), 110 pp.

Ehrhart, L.M., W.E. Redfoot, and D.A. Bagley. 1996. A study of the population ecology of in-water marine turtle populations on the east-central Florida coast from 1982-96.

Comprehensive Final Report to NOAA the National Marine Fisheries Service.

Unpublished report. Department of Biology, University of Central Florida, Orlando, 164 PP.

Encalada, S. E., K. A. Bjorndal, A. B. Bolten, J. C. Zurita, B. Schroeder, E. Possardt, C. J. Sears, and B. W. Bowen. 1998. Population structure of loggerhead turtle (Carettacaretta) nesting colonies in the Atlantic and Mediterranean as inferred from mitochondrial DNA control region sequences. Mar. Biol. 130:567-575.

Epperly, S.P., J. Braun, and A. Veishlow. 1995. Sea turtles in North Carolina waters. Conserv.

Biol. 9:384-394.

Epperly, S.P. and W.G. Teas. 1999. Evaluation of TED opening dimensions relative to the size of turtles stranding in the Western North Atlantic. U.S. Dep. Commer. NMFS SEFSC Contribution PRD-98/99-08, 31 pp.

52 NOAA Tech. Mem. NMFS-SEFSC-444

Frazer, N.B. and L.M. Ehrhart. 1985. Preliminary growth models for green, Chelonia mydas, and loggerhead, Caretta caretta,turtles in the wild. Copeia 1985:73-79.

Frazer, N.B., C.J. Limpus, and J.L. Greene. 1994. Growth and age at maturity of Queensland loggerheads. U.S. Dep. Commer: NOAA Tech. Mem. NMFS-SEFSC-351:42-45.

Francisco, A.M., A.L. Bass, and B.W. Bowen. 1999. Genetic characterization of loggerhead turtles (Carettacaretta)nesting in Volusia County. Unpublished report. Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville, 11 pp.

Gulf & South Atlantic Fisheries Development Foundation, Inc. 1998. Alternatives to TEDs.

Final Report to NOAA, Contract 50WCNF606083. Unpublished report. Tampa, Fla., 19 pp.

Harris, M.J. and C.G. Maley. 1990. Loggerhead Sea Turtle Nesting on Ossabaw, St. Catherines, Sapelo, and Little St. Simons Islands, Georgia in 1989. Project E-4, Study I Unpublished report. Georgia Department of Natural Resources, Brunswick, 28 pp.

Jamir, T.V.C. 1999. Revisions to the estimates of incidental sea turtle capture aboard commercial shrimp trawling vessels. Gulf & South Atlantic Fisheries Foundation, Inc.

Unpublished report. Tampa, Fla., 17 pp.

Keinath, J.A. 1993. Movements and behavior of wild and head-started sea turtles. Ph.D. Diss.

College of Wiliam and Mary, Gloucester Point, Va., 206 pp.

Karl, S.A., B.W. Bowen, and J.C. Avise. 1992. Global population genetic structure and male-mediated gene flow in the green turtle (Chelonia mydas): RFLP analyses of anonymous nuclear loci. Genetics 131:163-173.

Laurent, L;, J. Lescure, L. Excoffier, B. Bowen, M. Domingo, M. Hadjichristophorou, L.

Kornaraki, and G. Trabuchet. 1993. Etude g~n~tique des relations entre les populations mditerrandenne et atlantique d'une tortue marine (Carettacaretta)Al'aide d'un marqueur mitochondrial. C.R. Acad. Sci. Paris 316:1233-1239.

Laurent, L, P. Casale, M.N. Bradai, B.J. Godley, G. Gerosa, A.C. Broderick, W. Schroth, B.

Schierwater, A.M. Levy, D. Freggii, E.M. Abd El-Mawla, D.A. Hadoud, H.E. Gomati, M.

Domingo, M. Hadjichristophorou, L: Kornaraky, F. Demirayak, and Ch. Gautier. 1998.

Molecular resolution of marine turtle stock composition in fishery bycatch: a case study in the Mediterranean. Molecular Ecol. 7:1529-1542.

Maigret, J. 1983. Repartition des tortues de mer sur les cotes ouest Africaines. Bull. Soc. Herp.

1983:22-34.

NOAA Tech. Mem. NMFS-SEFSC-444 53

Maley, C.G. 1995. Loggerhead Sea Turtle Nesting on Ossabaw, St. Catherines, Sapelo, and Little St. Simons Islands, Georgia in 1994. Project E-5, Study I Unpublished report.

Georgia Department of Natural Resources, Brunswick, 16 pp.

Maley, C.G. and M.J. Harris. 1991. Loggerhead Sea Turtle Nesting on Ossabaw, St. Catherines, Sapelo, and Little St. Simons Islands, Georgia in 1990. Project E-4, Study I.

Unpublished report. Georgia Department of Natural Resources, Brunswick, 33 pp.

Maley, C.G. and M.J. Harris. 1992. Loggerhead Sea Turtle Nesting on Ossabaw, St. Catherines, Sapelo, and Little St. Simons Islands, Georgia in 1991. Project E-4, Study ILUnpublished report. Georgia Department of Natural Resources, Brunswick, 37 pp.

Maley, C.G. and M.S. Murphy. 1993. Loggerhead Sea Turtle Nesting on Ossabaw, St.

Catherines, Sapelo, and Little St. Simons Islands, Georgia in 1992. Project E-4, Study.

Unpublished report. Georgia Department of Natural Resources, Brunswick, 23 pp.

Maley, C.G. and M.S. Murphy. 1994. Loggerhead Sea Turtle Nesting on Ossabaw, St.

Catherines, Sapelo, and Little St. Simons Islands, Georgia in 1993. Project E-5, Study I.

Unpublished report. Georgia Department of Natural Resources, Brunswick, 22 pp.

Mdirquez-M., R. 1990. FAO Species Catalogue, Vol. 11. Sea turtles of the world, an annotated and illustrated catalogue of sea turtle species known to date. FAO Fisheries Synopsis, 125, 81 pp.

Mays, J.L. and D.J. Shaver. 1998. Nesting trends of sea turtles in National Seashores along Atlantic and Gulf Coast waters of the United States. Final Report for Natural Resources Preservation Program, Project Number 15. Unpublished report. Padre Island National Seashore, Corpus Christi, Tx., 61 pp.

McDaniel, C.J., L.B. Crowder, and J.A. Priddy. 2000. Spatial dynamics of sea turtles abundance and shrimping intensity in the U.S. Gulf of Mexico. Conserv. Ecol. 4:15 [online] URL:

http://www.consecol.org/vol14/iss1/art15.

Meylan, A., B. Schroeder, and A. Mosier. 1995. Sea turtle nesting activity in the state of Florida. Fla. Mar. Res. Publ. 52:1-51 Morreale, S.J., and E.A. Standora. 1999. Vying for the same resources: potential conflict along migratory corridors. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-415:69.

Mrosovsky, N. 1988. Pivotal temperatures for loggerhead turtles (Caretta caretta)from northern and southern nesting beaches. Can. J. Zool. 66:661-669.

Murphy, T.M. and S.R. Hopkins. 1984. Aerial and ground surveys of marine turtle nesting beaches in the southeast region, U.S. Final report to National Marine Fisheries Service 54 NOAA Tech. Mem. NMFS-SEFSC-444

Southeast Fisheries Center. Unpublished report. Contract No. NA-83-GA-C-00021, 73 pp.

Nance, J.M. 1992. Estimation of effort for the Gulf of Mexico shrimp fishery. U.S. Dep.

Commer..NOAA Tech. Memo NMFS-SEFSC-300, 12 pp.

Nance, J.M. 1993. Effort trends for the Gulf of Mexico shrimp fishery. U.S. Dep. Commer.

NOAA Tech. Memo NMFS-SEFSC-337, 37 pp.

Norrgard, J. 1995. Determination of stock composition and natal origin of a juvenile loggerhead turtle population (Caretta caretta)in Chesapeake Bay using mitochondrial DNA analysis.

M.A. Thesis. College of William and Mary, Williamsburg, Va., 47 pp.

Ogren, L. 1978. Survey and reconnaissance of sea turtles in the northern Gulf of Mexico.

Unpublished report. National Marine Fisheries Service, Panama City, Fla., 8 pp.

Quantum Resources. 1995. Annual environmental operating report for 1994. Unpublished report. Florida Power and Light Company, Juno Beach, Fla., 49 pp.

Rankin-Baransky, K. C. 1997. Origin of loggerhead turtles (Carettacaretta)in the western North Atlantic as determined by mt DNA analysis. M.S.Thesis, Drexel University, Philadelphia, Penn., 50 pp.

Rankin-Baransky, K., C.J. Williams, A.L. Bass, B.W. Bowen, and J.R. Spotila. in review. Origin of loggerhead turtle (Carettacaretta)strandings in the northwest Atlantic as determined by mtDNA analysis. J. Herpetology.

Richardson, T.H., J.I. Richardson, C. Ruckdeschel, and M.W. Dix. 1978. Remigraton patterns of loggerhead sea turtles (Carettacaretta)nesting on Little Cumberland and Cumberland Islands, Georgia. Fla. Mar. Res. Publ. 33:39-44.

Ross, J.P. -1982. Historical decline of loggerhead, ridley, and leatherback sea turtles, pp. 189-195. In: Bjorndal, K.A. (editor), Biology and Conservation of Sea Turtles. Smithsonian Institution Press, Washington, D.C.

Royle, J.A. 2000. Estimation of the TED effect in Georgia shrimp strandings data. Unpublished report, U.S. Fish and Wildlife Service, Laurel. Md., 1I pp.

Royle, J.A. and L.B. Crowder. 1998. Estimation of a TED effect from loggerhead strandings in South Carolina and Georgia strandings data from 1980-97. Unpublished report. U.S.

Fish and Wildlife Service, Laurel, Md., 12 pp.

Schroeder B.A. 1994. Florida index nesting beach surveys: are we on the right track. U.S. Dep.

Commer. NOAA Tech. Mem. NMFS-SEFSC-351:132-133.

NOAA Tech. Mem. NMFS-SEFSC-444 55

Schroeder, B.A., A.M. Foley, B.E. Witherington, and A.E. Mosier. 1998. Ecology of marine turtles in Florida Bay: Population structure, distribution, and occurrence of fibropapilloma. U.S. Dep. Commer. NOAA Tech. Mem. SNFS-SEFSC-415:265-267.

Sears, C.J. 1994. Preliminary genetic analysis of the population structure of Georgia loggerhead sea turtles. U.S. Dep. Commer. NOAA Tech. Memo NMFS-SEFSC-351 :135-139.

Sears, C.J., B.W. Bowen, R.W. Chapman, S.B. Galloway, S.R. Hopkins-Murphy, and C.M.

Woodley. _ 995. Demographic composition of the feeding population of juvenile loggerhead sea turtles (Carettacaretta)off Charleston, South Carolina: evidence from mitochondrial DNA markers. Mar. Biol. 123:869-874.

Shoop, C.R. and R.D. Kenney. 1992. Seasonal distributions and abundance of loggerhead and.

leatherback sea turtles in the waters of the northeastern United States. Herpetol. Monogr.

6:43-67.

Sternberg, J., compiler. 1981. The worldwide distribution of sea turtle nesting beaches. Center for Environmental Education, Washington, D.C., 10 pp.

Turtle Expert Working Group. 1998. An assessment of the Kemp's ridley (Lepidochelys kempii) and loggerhead (Carettacaretta) sea turtle populations in the Western North Atlantic.

U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-409, 96 pp.

Wibbels, T., D.W. Owens, Y.A. Morris, and M.S. Amoss. 1987. Sexing techniques and sex ratios for immature loggerhead sea turtles captured along the Atlantic coast of the United States. U.S. Dep. Commer. NOAA Tech. Rep. NMFS-53: 65-73.

Witzell, W.N., A.L. Bass, M.J. Bresette, D.A. Singewald, and J.C. Gorham. in review. Origin of immature loggerhead sea turtles (Carettacaretta)from Hutchinson Island, Florida:

evidence from mtDNA markers. Fish. Bull.

Xcaret. 1999. Programa de Protecci6n de Tortugas Marinas en el parque Eco-Arqueol6gico Xcaret. Unpublished report. Xcaret, Cancun, Mexico, 78 pp.

56 NOAA Tech. Mem. NMFS-SEFSC-444

z0 Table 10. Natal origin of foraging-ground loggerhead sea turtles. Table values are proportional contributions.

• H Nesting Subpopulation South Western North Atlantic Atlantic Mediterranean Foraging Area Life History Stage(s)

N.C. to Ca* N Northeast South NW Yucatin, Florida Florida Florida Mexico Brazil Greece Northeast U.S.' '* Primarily Benthic Immature 82 0.25 0.59 0.00 0.16 0.00 0.00 ci.

Chesapeake Bay*' Primarily Benthic Immature 63 0.54 0.46 m¢' Pamlico and Core Sounds, N. C? Benthic Immature 97 0.32 0.64 <0.01 0.03 0.01 0.00 ci- S. Carolina*' Benthic Immature

. 33 0.50 0.50 Georgia*' Benthic Immature 97 0.59 0.41 Georgia' Benthic Immature 192 0.24 0.73 0.03 0.00 0,00 3

Hutchinson Island, Flcrida' Benthic Immature 109 0.10 -0.69 0.20 Florida Bay' Benthic Immature and Adult 51 0.08 0.84 0.08 Mediterranean Sea*', Unknown 59 0.57 0.43 Azores and Madeira"' Pelagic Immature 183 0.19 0.71 0.00 0.11 0.00 0.00 Western Mediterranean" Pelagic Immature 59 0.02 0.45 0.53 Eastern Mediterranean" Pelagic Immature 52 0.02 0.47 0.51 Eastern Mediterranean"' Benthic Immature and Adult 58 1.00""

• Studies were cond uctcd before full complement of gcnetic analyses of Atlantic nesting b eaches were availablc.

• Includes rookeries of Turkey and Cyprus, in addit ion to Greece.

Rankin-Baransky 1997; ` Norrgard, 1995; ' Bass et al., 1998; Sears et al., 1995; ' Sears, 1994;' Bass et al., 1999; 'B. Schroeder, unpublished data; Laurent et al., 1993; 'Bowen, 1995, "Bolten et al., 1998, ' Laurent et al., 1998, ':Witzell et al., in revie'; Rankin -Baransky et al., in review.

Table 11. Loggerhead sea turtle nests in the U.S., 1989-1998. Minor nesting occurring north of North Carolina and west of Florida is not included. Data are from Meylan et al. (1995), Schroeder (1994), Maley (1995), Maley and Murphy (1993, 1994), Maley and Harris (1991, 1992), Harris and Maley (1990), Florida Depart ient of Environmental Protection (unpubl.

data), Georgia Department of Natural Resources (unpubl. data), South Carolina Department of Natural Resources (unpubl.

data) and North Carolina Department of Environment, Health and Natural Resources (unpubl. data). Nesting female estimates are based on an average of 4.1 nests/female in all subpopulations.

Year Mean Area 1989-1998 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Northern Nesting Sub population:

North Carolina 457 806 931 729 485 1021 662 776 568 861 730 South Carolina 2444 4491 3657 3943 2757 4136 2959 3892 2169 4265 3471 Georgia 691 1085 1209 1054 - 475 1375 1028 1111 816 1067 991 N.E. Florida 778 1355 952 759 809 1355 1149 1142 893 1359 1055 Total 4370 7737 6749 6485 4526 7887 5798 6921 4446 7552 6247 Nesting Females 1066 1887 1646 1582 1104 1924. 1414 1688 1084 1842 1524 South Florida Nesting Subpopulation:

z0 S.E. Florida 46295 62071 63416 59677 50618 64410 71394 68174 56266 74988 61731 S.W. Florida 2236 3085 3959 4186 3836 5395 6973 6461 6746 8454 5133 Total 48531 65156 67375 63863 54454 69805 78367 74635 63012 83442 66864 z Nesting Females 11837 15892 16433 15576 13281 17026 19114 18204 15369 20352 16308 Florida Panhandle Nesting Subpo ulation:

Total 113 174 287 351 560 772 928 891 1133 1188 640 Nesting Females 28 42 70 86 137 188 226 217 276 290 156 Total U.S. 53014 73067 7441 1 70699 59540 78464 85093 82447 68591 92182 73751 c.L Nests Total U.S, 12930 17821 . 18149 17244 14522 19138 20754 20109 16730 *22483 17988 Nesting Females

Table 12. Mean annual percent change in loggerhead nest numbers in the U.S. 1989-1998, based on slopes calculated from In-transformed data. Bold F statistics are significant at P < 0.05; variance in nest counts from year-to-year and possible cyclical H nesting reduced the power to detect trends for many areas. Data are for calculations of minimum number of nests and may not be accurate for assessments of trends due to yearly changes in nesting beach survey effort. See text for a discussion of trends in loggerhead nesting numbers for each subpopulation.

North South N.E. Total S.E. S.W, Total Florida z Carolina Carolina Florida Northern Florida Florida S. Florida Panhandle

-i~1 Percent change per 1.7% -0.1% 1.5% 2.9% 0.8% 2.9% 13.2% 3.6% 2.6% 3.6%

fl-I

-n year 1989-1998 SE of slope 0.031 0.030 0.036 0.026 0.027 0.014 0.014 0.014 0,025 0.015 r-square 0.036 0.000 0,021 0.136 0.012 0.344 0.919 0.461 0.932 0.424 F 0.297 0.002 0.175 1.261 0.095 4.197 90.805 6.840 109.060 5.895

Table 13. Correlations (Pearson) of loggerhead nests from index beaches at Cape Island, Cape Romain National Wildlife Refuge, S.C., and Hutchinson Island, Fla. with nest totals (excluding each index beach) from each state and region, 1989-1998.

Bold correlations are significant at P < 0.05. A. Raw nest numbers. B. Residuals from In-transformed regression analysis. Raw nest numbers from Cape Island, S.C. and Hutchinson Island, Fla. were not significantly correlated (r = 0.4667, P = 0.174), but their regression residuals were highly correlated (r 0.7602, P 0.01),

North South Georgia NE Northern SE Florida SW S. Florida Florida Carolina Carolina Florida Subpopulation (excluding Florida Subpopulation Panhandle (excluding (excluding Hutchinson (excluding Cape Is.) Cape Is.) Is.) Hutchinson Is.)

A. Raw nest numbers Cape Island. S.C. 0.8132 0.8274 0.8412 0.6966 0.9017 0.5615 -0.0251 0.4629 -0.2197 Hutchinson Island, Fla. 0.5307 0.5107 0.5750 0.6511 0.6213 0.9547 0.7225 0.9456 0.5389 B. Regression residuals Cape Island, S.C. 0.8898 0.8247 0.9047 0.7809 0.9470 0.8813 0,6312 0.8719 -0.0769 Hutchinson Island, Fla. 0.6501 0.6470 0.6663 0.6281 0.7305 0.9414 0.8206 0.9395 0.1378 z

0

Table 14. Correlations (Pearson) of U. S. loggerhead nests from all states and regions, 1989-1998. Bold correlations are significant at P < 0.05. A. Raw nest numbers. B. Residuals from In-transformed regression analysis.

North South Georgia N.E. Northern S.E. S. W. S. Florida Florida zT Carolina Carolina Florida Total Florida Florida Total Panhandle 0ý A. Raw nest numbers North Carolina 1.000 South Carolina 0.8'18 1.000 Georgia 0.935 0.733 1.000 N.E. Florida 0.694 0.683 0.649 1.000 Northern Total 0.920 0.963 0.865 0.799 1.000 S.E. Florida 0.688 0.617 0.702 0.742 0.723 1.000 S.W. Florida 0.271 0.101 0.285 0.459 0.232 0.758 1.000 S. Florida Total 0.635 0.543 0.650 0.717 0.65.8 0.992 0.832 1.000 Florida Panhandle 0.103 -0.091 0.096 0.354 0.042 0.570 0.960 0.665 1.000 B. Regression residuals North Carolina 1.000 South Carolina 0.855 1.000.

Georgia 0.905 0.706 1.000 N.E. Florida 0.695 0.693 0.633 1.000 Northern Total 0.940 0.967 0.844 0.787 1.000 S.E. Florida 0.828 0.807 0.822 0.722 0.873 1.000 S.W. Florida 0.709 0.525 0.758 0.431 0.635 0.863 1.000 S. Florida Total. 0.824 0.790 0.826 0.713 0.863 0.999 0.876 1.000 Florida Panhandle 0.103 0.012 0.029 -0.037 0.040 0.108 0.447 0.115 1.000 ON

Table 15. Loggerhead strandings by region.

1986 1987 1988 1989 .1990 1991 1992 1993 1994 1995 1996 1997 Total N.E.U.S. 93 175 166 134 207 158 206 221 227 299 274 346 2506 Atlantic S.E.U.S. 828 1179 904 896 1091 606 714 558 777 859 1127 902 10441 Atlantic Gulf of 278 371 298 392 288 199 176 186 302 245 415 376 3526 Mexico Total 1199 1725 1368 1422 1586 963 1096 965 1306 1403 1816 1624 16473 Table 16. Loggerhead stranding trends. The annual rate of change for each region is based on a regression slope of In-transfonmed strandings, with the model fit (r-square) given in parentheses.

Slope values in bold are significant at P < 0.05.

Annual percent change 1986-88 1995-97 'Percent 1986-1997 . 1989-1997 1991-1997 change (r-square) (r-square), (r-square)

N.E.U.S. 434 919 + 112% +8.9% +10.0% +11.5 Atlantic (10%.) (19%) (0.801) (0.846) (0.911)

S.E.U.S. 2911 2888 - 1% -0.8% . +1.5% +9.1%

Atlantic (68%) (60%) (0.017) (0.030) (0.659)

Gulf of Mexico 947 1036 + 9% 0.0% +3.1% +13.9%

(22%) (21%) (0.000) (0.068) (0.759)

Total 4292 4843 + 13% 0.0% +3.1% +10.5%

(0.017) (0.138) (0.835) 62 NOAA Tech. Mem. NMFS-SEFSC-444

Table 17. Trends in loggerhead abundance at the St. Lucie Nuclear Power Plant, St. Lucie, Florida (Quantum Resources, 1995, unpublished data), over various time intervals to compare with nests and strandings. Bold values are significant at the P<0.05 level.

Annual percent increase 1977-1998 1989-1998 1991-1998 (r-square) (r-square) (r-square)

All loggerheads 4.2% 14.0% 17.0%

(0.401) (0.783) (0.764)

Adults( 92cm SCL) 12.3% 11.2% 25.3%

(0.605) (0.2282) (0.546)

Large Benthic Immatures (70- 5.5% 10.1% 17.0%

91.9 cm SCL) (0.377) (0.467) (0.687)

Small Benthic Immatures (<70 3.6% 17.0% 16.5%

cm SCL) (0.297) (0.890) (0.805)

Table 18. Observed size distributions for loggerheads in zones 29-32 of the southeast U.S.' May 1997 - May 1998 (except January and February).

Data source <70 cm SCL 70-91.9cm SCL Strandings Data (n=395) 77% 15%

Naked Net Captures (n="

'Gulf & South Atlantic Fisheries Development Foundation, Irc. (1998), Jamir (1999)

NOAA Tech. Mem. NMFS-SEFSC-444 63

Table 19. Pearson correlation coefficients between loggerhead nests and strandings in the southeastern U. S. and Gulf of Mexico and loggerhead captures at the St. Lucie Power Plant, 1989-1997. Bold coefficients are significant at P < 0.05. A. Raw catch data. B. Residuals from ln-transformed regression analysis.

Northern South Panhandle NE US SE US Gulf of SE US Adult SE US St. Lucie St. Lucie nests Florida nests Strandings. Strandings Mexico Strandings Immature all adults nests Strandings Strandings A. Raw data Northern nests 1.000 S. Florida nests 0.604 1.000 Panhandle nests -0.146 0.542 1.000 NE US Strandings -0.128 0.530 0.933 1.000 SE US Strandings 0.263 0.294 0.166 0.311 1.000 Gulf of Mexico Strandings -0.121 -0.026 0.300 0.261 0.762 1.000 SE US Adult Strandings 0.790 0.549- 0.219 0.250 0.592 0.310 1.000 SE US Immature Strandings 0.268 0.186 0.041 0.220 0.977 0.727 0.556 1.000 St. Lucie All .0.072 0,645 0.735 0.658 0.501 0.464 0.412 0.334 1.000 St. Lucic Adults 0.089 0.458 0.399 0.370 0.655 0.522 0.433 0.518 0.886 1.000 z0 St. Lucie Immatures 0.065 0.665 0.785 0.701 0.458 0.442 0.395 0.289 0.995 0.836

> B. Ln-transfo rned residuals 0

Northern nests 1.000 S. Florida nests 0.855 1.000 Cs Panhandle nests 0.285 0.322 1.000 NE US Strandings 0.126 0.218 -0.048 1.000 z SE US Strandings 0.251 0.223 -0.728 0.366 1.000

-TI Gulf of Mexico Strandings -0.138 -0.256 -0.824 -0.165 0.766 1.000 SE US Adult Strandings 0.822 0.534 0.009 0.286 0.556 0.237 1.000 SE US Immature Strandings 0.237 0.194 -0.761 0.385 0.988 0.775 0.541 1.000.

St. Lucie All 0.193 0.319 -0.182 -0.068 0.532 0.393 0.330 0.430 1.000 St. Lucie Adults 0.005 0.068 -0.275 0.202 0:633 0.509 0.369 0.555 0.885 1.000 St. Lucie imnatures 0.249 0.389 -0.118 -0.172 0.450 0.316 0.294 0.341 0.984 0.790

z 0 Figure 15. Loggerhead nesting areas, Western North Atlantic.

H CD rT U tUNITED STATES

-r

.A .

C"

• -. . *.a *.h *i

., . . ,.' - ,.%,"7... *-

.. w s'*

.*....J,. -*, , .. '- * ,*.....

..... > .*:*, *.q z'*  ; ,' L :*

. .:. * *i r :.. *, ' , *" pq

• ..? n...

. ,.:."... , ... . . *: e.'% -, a,, , "'

1 Figure 16. Loggerhead nests for the Northern Nesting Subpopulation, 1989-1998. A. Nest counts by state. B. Long-term nesting records for Cape Island, S.C.

5000 o0 4500 4000 U aSouth 0 3500 Carolina

3000 ------ N.E. Florida C 2500

.- 2000 -A---- Georgia 0

1500 2

-A-A000.A. i-- ".-A -.-4-.- North

- - K ..-- 77 Carolina (A

(D 500- A z

0 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 3000 U,

C

-i 2500 -

2 0

0 2000 4 C.

U WI 1500 -

WI a'

C

'U 1000 -

a' r

a' w

w 500 -

0

-J 0

IC) C- ) - CO nCJ I- - CO If, C-r- - C t to (0 mo CO tO' CD to to to to to

0) to o ton to to to to 66 NOAA Tech. Mem. NMFS-SEFSC-444

Figure 17. Loggerhead nests for the South.Florida Nesting Subpopulation, 1989-1998. A. Nest counts by region. B. Long-term nesting records for Hutchinson Island, Fla.

A.

90000 - .. .. .... ... . .... ....... 9oo°10000 80000 70000South-6000 -00

"*- 60000 east 0

.1 50000 7000 w

--. .0 South-40000 6000 west 0,

w30000 5000 20000 -

10000_A 4000 0 -- . 3000 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 B.

9000 8000

-J LA-In 7000 C

0 6000 In C

U 5000 I

4000 ----------------------------------------------

in In a) 3000 ----------------------------------------------

C (U

1) 2000 ----------------------------------------------

L a) 03 1000 03 0

-J 0

-" Mn 4t '00O",

ýD 0n 0 0 - " MI en 4 , \0 rO CO, OA O, OT, O, S-S O\ O\ - O\ O\ OM cc i Oh ONM OF O- O NOAA Tech. Mere. NMFS-SEFSC-444 67

0 Figure 18. Loggerhead turtles captured by the cooling water intake pipe at the St. Lucie Nuclear Power Plant, St. Lucie, Fla.,

1977-1998.

o 300- c<70 cm a-6.

-' 250

-- 70-91.9 cm z

j 200 j

-*-92+ cm 150 a) ~. 100 z0 o 50

-3 ~0 C)

.I-I'- CF r Mo O I,- M T- CO) 0 N-0 I- CO CO COD cO CO 0) 0) 0) 0)

C) 0) 0) 0) 0) 0a) .0) 0) 0) 0) 0) 0)

• T- T- ',- T*- 'I-- .T-'-* - T- T.- T-

-J z

"11 Cr Cr rTj t

Cr cm A

A A

C© Figure 19. Size distributions and cumulative frequencies for loggerhead turtle strandings. A. Pooled data for 1986-1987. B.

> Pooled data for 1996-1997.

H A.1986-1987 pooled strandings B. 1996-1997 pooled strandings ri 50 -100360 500 A 10M0%

CA90% 90% 0 40o 8oo% C 400o 80%o fl 7

7%

t o70%

70%~

w300- 0*< 60% C03 00 c - 60ý%

0.)0 50%

'200 -40%~ LI.L 200

. 40% In C 30% --30%~

100- 20% 100 20%~

10% 210% D E E E E E E E E E E E E E E E E E E E E

• 0 0 0 0 0 0 0 0 () 0 C 0 0 0 0 0 0 0 0 Li)

Co LO

• q- LW)

Li) L)

(.0 LV)

N- Li)

CC Lf) 0 LI) 0 IC)

- Li)

( IC L I') LO) L) L 1 1") L1) Li L)

I. IT I? 1 - -

0 IC 000 04 oD CD CD CD 6 6D C") 4t (D N- 00 0) 0 0i 0 0 Co LO (D - 00 0) 0 00 oD 1-7(

,ON

C-4Figure 20. Size distributions of loggerheads captured at the St. Lucie Nuclear Power Plant, 1977-1998, represented as proportion of total captures. Time series of total captures is included (right y-axis).

1 800 0.9 700 0.8 600 0.7 -i C 0.6 500 2.

1 I 0

0 0.5 w

.... . .. .... I- 400 0 C

a. 0.4 300 L.

.......... 0, 0.3 . . . .. . . . ... . . . .. . ..

- 200 m 92+ cm CD 0.2 0 70-91.9 cm Y0 0.1 .... ... .......

F-100 El <70 cm n m 1I I I I I I I I I I I I I I I I I I I - 0 O3 Total 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 captures Cr nA 41

Part 3. Anthropogenic Mortality Sources Summary A number of anthropogenic mortality sources have been identified for Kemp's ridley and loggerhead sea turtles (National Research Council, 1990; NMFS & USFWS, 1991; USFWS and NMFS, 1992) but few, outside drowning in bottom trawls, have been quantified with any degree of confidence. While they still cannot be quantified, new information in recent years leads the TEWG to be particularly concerned about longline fisheries and coastal gillnet fisheries, and about marine debris and pollution, mortality sources that primarily affect the pelagic immature stage. A more thorough assessment of anthropogenic mortality sources is provided in the first TEWG report (1998).

Mortality Sources for Pelagic Stage Important sources of mortality occur in the pelagic phase of loggerhead turtles and are not revealed by numbers of stranded loggerheads. Likewise, this pelagic mortality is not accounted for in estimates of population size and mortality presented in TEWG (1998). This pelagic phase is approximately a decade in length (Bjorndal et al., 2000; Bolten et al., 1994).

Pelagic longlinefisheries The loggerhead model (Crouse et al. 1987; Crowder et al. 1994) indicated that the pelagic juvenile stage is likely the second most sensitive stage with respect to population growth rates, after the benthic juvenile stage. Pelagic loggerhead juveniles are believed to circumnavigate the North Atlantic, via the Atlantic Gyre, over a period of several years and have been reported from the Azores, and the Mediterranean as well as the eastern Caribbean. During this transit they are exposed sequentially to a series of longline fisheries that primarily target swordfish and tuna.

Aguilar et al. (1995) estimated the Spanish swordfish longlining fleet in the Mediterranean, one of many fleets working in the region, alone captures more than 20,000 juvenile loggerheads (killing as many as 10,700) annually. About half of these are from southeastern U.S. beaches (Bowen et al.,

1993; Laurent et al., 1993, 1998). Bolten et al. (1994, 1998) reported that virtually all of the pelagic immature loggerheads taken in the Azorean longline fleet are from western North Atlantic nesting subpopulations. In one study on the Mexican tuna longline fishery in the Gulf of Mexico, which included the capture of a Kemp's ridley, sea turtle mortality was 33% (Ulloa Ramirez and Gonzdiles Ania, 2000), comparable to the mortality rates reported by Aguilar et al. (1995). Estimated'bycatch of marine turtles by the U.S. Atlantic tuna and swordfish longline fisheries, based on observer data, was significantly greater than reported in logbooks (Johnson et al., 1999; Witzell, 1999). Observer records indicated that an estimated 4808 loggerheads and no Kemp's ridleys were captured by the U.S. fleet 1992-1997 of which 21 were released dead (Johnson et al., 1999). Logbooks indicated that loggerheads readily ingested hooks (Witzell, 1999).

NOAA Tech. Mem. NMFS-SEFSC-444 71

Marine debris An additional source of mortality that has not been adequately assessed is the ingestion of anthropogenic debris by pelagic turtles. Preliminary indications are that approximately 15% of pelagic post-hatchling loggerheads from Florida beaches have ingested plastics and approximately 46% have ingested tar within the first few weeks of pelagic foraging (n=168) (Witherington 1994, in review).

Mortality Sources for Benthic Feeding Stages Trawls Pursuant to Endangered Species Act Amendments of 1988, the National Research Council Committee on Sea Turtle Conservation assessed the status of the U.S. sea turtles (National Research Council, 1990). Incidental capture in shrimp trawls was identified as the major cause of sea turtle mortality associated with human activities, killing more sea turtles than all other human activities combined. An estimated 500-5000 benthic immature and adult Kemp's ridley annual mortalities and an estimated 5,000-50,000 loggerhead mortalities were attributed to shrimp trawling (Ibid.),

prior to implementation of TED regulations in 1989. This compares to 75-750 and 500-5000 estimated mortalities due to all other known human causes, respectively. The Council identified other fisheries, particularly the winter trawl fishery for summer flounder off North Carolina and Virginia as a possible source of turtle losses and estimated those as 50-200 per year. Among the findings of the National Research Council Committee on Sea Turtle Conservation (1990) were:

(1) mortality of benthic immature and adult stages must be reduced to prevent extinction and effect recovery, under the condition that large numbers of hatchlings continue to be produced; (2) shrimp trawling kills more benthic immature and adult sea turtles than all other human causes combined; (3) in Texas and South Carolina, strandings increased with the onset of shrimping activities and decreased with the closure of Texas waters, indicating that 70-80% of the sea turtles stranding during the shrimping seasons were killed in shrimp trawls; and (4) shrimping can be compatible with sea turtle conservation if adequate controls are placed on trawling activities, especially the mandatory use of TEDs at most times and places.

Federal regulations requiring TEDs in trawls used by offshore (seaward of the COLREGs demarcation line) shrimp trawlers longer than 25 ft were published in 1987 (Federal Register, Vol.

52, No. 124, p. 24247, 28 June 1987). However, TEDs were first required in shrimp trawls beginning in May 1989, but were used only sporadically until May 1990, and then only seasonally (McDonald, 1990; Crouse et al., 1992; Henwood et al., 1992; Crouse, 1993; Weber et al., 1995).

Full implementation of seasonal TED requirements became effective essentially in May 1991.

Beginning in December 1992, year-round use of TEDs was required by most trawlers operating in 72 NOAA Tech. Mem. NMFS-SEFSC-444

Southeastern U.S. waters (57 FR p. 48861, Sept. 8, 1992) and this regulation was fully implemented by December 1994 (57 FR p. 18446, Apr. 30, 1992; 57 FR p. 57348, Dec. 4, 1992). In the three years before Federal Regulations (1986-1988), U.S. sea turtle strandings averaged 33% higher than in TED years 1991-1993 (Crouse, Crowder and Heppell unpubl. as cited by Crowder et al., 1995).

The halt in the decline in numbers of nests at Rancho Nuevo can be attributed to a combination of factors, including a decrease in shrimp fishing effort in M~xican waters when the U.S. shrimping fleet was prohibited from shrimping there in the late 1970s, and the overall decline and deterioration of the Mexican shrimping fleet. Threats to mating and nesting adult Kemp's ridleys from shrimping off Rancho Nuevo were reportedly high until the U.S. and Mexican governments negotiated a bilateral agreement in 1976. Under this agreement US. shrimping in Mdxican waters was phased out through 1979 (Iversen et al., 1993). After 1979, U.S. vessels continued to shrimp off Mdxico, although illegally and at reduced levels, through the mid-1980s, when the U.S. enforced the Lacey Act.

The reduction in shrimping effort caused by the departure of the U.S. fleet from Western Gulf waters of M~xico was furthered by the subsequent decline and deterioration of the Mexican fleet. In 1995, the Mexican shrimp fleet in the Gulf was approximately 660 vessels, many of which were not actually fishing. There has been a decline in the number of vessels operating in the Campeche fleet, which currently represents about half of the operating M~xican vessels, but the Tamaulipas fleet appears to be increasing (Table 20). Additionally, since 1978, Waters out to 4 nm along approximately 14 km of the beach at Rancho Nuevo have been closed to fishing during the nesting season. However, this closure has not been strictly enforced, and until a few years ago, beach workers reported observing 10-20 trawlers operating 2-5 mi off the beach at night. Since 1995, Mexican waters of the Gulf of Mexico are closed to shrimping in concert with the Texas Closure, from May 15 through July 15, through binational agreement.

A study of the Mid-Atlantic winter trawl fishery for summer flounder identified the fishery as a significant source of sea turtle mortality, estimating that 1063 turtles, mostly Kemp's ridleys and loggerheads, were caught in a 4 month period and that 89-191 were killed (Epperly et al., 1995, 1996). Consequently, since 1992, TEDs have been required in the fishery south of Cape Charles, Virginia. At first these regulations were implemented through emergency and interim rules and were seasonal requirements (57 FR p. 53603, Nov. 12, 1992; 57 FR p. 60135, Dec. 18, 1992; 58 FR

p. 4088, Jan. 13, 1993; 58 FR p. 8554, Feb. 16, 1993). They became year round requirements (58 FR p. 48797, Sep. 20, 1993; 59 FR p. 19584, Mar. 7, 1994), and subsequently were codified in a final rule (61 FR p. 1846, Jan. 24, 1996) requiring TEDs in all summer flounder trawls operating south of Cape Charles, Va. There is a temporary reprieve of the regulation automatically occurring each year north of Oregon Inlet from Jan. 14-Mar. 15, when waters predictably are very cold.

Gill nets Turtles are particularly susceptible to entanglement and drowning in gill net gear, especially when nets are set and left unattended. There are a number of gill net fisheries currently operating along the mid- and southeast U.S. Atlantic coastline that are known to incidentally capture NOAA Tech. Mem. NMFS-SEFSC-444 73

loggerhead turtles. Although all or most nearshore gill netting in state waters of South Carolina, Georgia, Florida, Louisiana, and Texas is prohibited by state regulations, gill netting in other states' waters and in federal waters does occur. Of particular concern are the nearshore and inshore gill net fisheries of the mid-Atlantic operating in New Jersey, Delaware, Maryland, Virginia, and North Carolina state waters and/or federal waters offshore thereof. Annual peaks in loggerhead strandings on ocean beaches and just inside the mouths of bays and inlets regularly occur in early summer and late fall, coinciding with increased gill netting activity and observers have documented lethal takes of both loggerhead and Kemp's ridley turtles in these fisheries (D. Christensen, personal communication). In December 1999 the large mesh gill net fishery for southern founder in Pamlico Sound, N.C. was closed to fishing for 30 days, based on elevated levels of stranded Kemp's and loggerhead turtles in inshore waters and the documentation of takes by the fishery (64 FR, p. 70196, Dec. 16, 1999). On May 12, 2000 the large mesh gill net fishery along eastern North Carolina and Virginia, north of Cape Hatteras, including the ocean, was closed for 30 days because of elevated strandings off northern North Carolina, mostly loggerhead turtles, observers' documentation that the large mesh gill net fisheries for monkfish and dogfish do take turtles (D. Christensen,personal communication), and the fact that 4 of the stranded turtles were entangled in large mesh gill nets consistent with these fisheries (65 FR p. 31500, May 18, 2000). NMFS made a determination that the level of mortality was "severely impacting the northern nesting subpopulation of loggerheads",

the first time management has identified a subpopulation of sea turtles as a management unit.

Additionally, the gill net fishery targeting sharks in federal waters offshore Georgia and Florida warrants close observation as these fisheries are known to take loggerheads (Trent et al., 1997) and likely could take Kemp's ridleys. Observers are required in this fishery during certain times of the year; however, it is unclear whether periods of expected highest turtle abundance are adequately covered.

NMFS has initiated discussions with coastal states of the mid-Atlantic in an attempt to address these take issues, however, little progress has been made (T. Conant,personal communication). There is concern by vessel captains over liability issues, and many have refused to take observers on board (S. Epperly,personalcommunication) despite the fact that many are classified as Category 2 fisheries under the Marine Mammal Act and are required to take observers if asked.

Hook and line Loggerhead and Kemp's ridleys are known to bite a baited hook, frequently ingesting the hook. Hooked turtles have been reported by the public fishing from boats, piers, the beach, banks, and jetties (Cannon et al., 1994; J. Braun McNeill, personal communication; A. Cannon personal communication; S. Epperly,personalcommunication) and from commercial fishermen fishing for reef fish and for sharks with both single rigs and bottom longlines (S. Epperly, personal communicaton). Necropsies of turtles have revealed hooks internally (Witzell and Teas, 1994),

which often were the cause of death (W. Teas,personal communication). An investigation of injuries and mortalities related to fish hook ingestion is underway at the National MarineFisheries.

Laboratory in Galveston, Texas (T. Fontaine,personalcommunication) and NMFS currently is exploring adding questions about encounters with sea turtles to intercept interviews of recreational 74 NOAA Tech. Mem. NMFS-SEFSC-444

fishermen conducted by the Texas Parks and Wildlife Department and under the auspicies of the Marine Recreational Fishery Statistics Surveys conducted throughout the Gulf of Mexico and along the Atlantic Coast (T. Fontaine, personalcommunication; T. Conant, personal communication).

NMFS also is considering questioning recreational fishermen aboard headboats throughout the southeast U.S. Atlantic and the Gulf of Mexico to quantify their encounters with sea turtles, also (J.

Braun McNeill, personal communication).

Powerplants Power plants are known to entrain loggerhead and Kemp's ridley sea turtles at the intake canals to their cooling systems. In some cases, the turtles may become trapped on intake screens and drown if they are unable to surface. The cumulative effect of such mortalities is unclear, but is believed currently to be relatively small. Some plants, such as the Brunswick Steam and Electric Plant in Southport, N.C., placed permanent diversion screens over the mouths of their intake canals to keep turtles away from the intake screens of the plant but this has not been 100% effective (W.

Pollard, personalcommunication). Sea turtle entrainment has also been reported at four other power plants in eastern Florida (National Research Council, 1990) and recently at the Crystal River plant in western Florida. This latter facility is of particular concern given its proximity to aggregations of subadult Kemp's ridley turtles in the eastern Gulf of Mexico (Schmid and Barichivich, submitted). Sea turtle monitoring at the Crystal River Energy Complex increased substantially in 1998 and 40 live takes (37 were Kemp's ridleys) and 5 lethal takes (all Kemp's),

were observed that year (NMFS, 1999), mainly from February-May. The turtles were subadults, with carapace lengths of 21-55 cm. Through March 1999, 4 subadult Kemp's ridleys were removed from the intake bars and released alive (D. Bruzek, personal communication).

In some cases the turtles pass unharmed through intake pipes into holding ponds, where they are then captured and released. The best documented situation of this type of interaction is the St.

Lucie Nuclear Power Plant (Florida Power and Light Co.) where, in recent years, capture rates have exceeded 200 loggerheads annually and the annual capture of Kemp's ridleys is generally less than 6 (Quantum Resources, 1995; J. Gorham,personalcommunication).

Marine debris Of 1710 turtles necropsied between 1980 and 1992, 197 (11.5%) had ingested debris, including plastic pieces and balloons (Witzell and Teas, 1994): a greater proportion of loggerheads were affected than were Kemp's ridleys, and in both species the percentage impacted by digested debris was highest in the Gulf of Mexico. Studies conducted in the western Gulf of Mexico reported debris ingestion rates of 51% for loggerheads and 34% for Kemp's ridleys (summarized in Witzell and Teas, 1994). Apparently Kemp's ridleys in the northeast U.S. consume little or no plastic debris (Sadove and Morreale, 1990; Burke et al., 1994).

NOAA Tech. Mem. NMFS-SEFSC-444 75

Literature Cited Aguilar, R., J. Mas and X. Pastor. 1995. Impact 6f Spanish swordfish longline fisheries onthe loggerhead sea turtle Carettacarettapopulations in the Western Mediterranean. U.S. Dep.

Commer. NOAA Tech. Mem. NMFS-SEFSC-361:1-6.

Bjorndal, K.A., A.B. Bolten, and H.R. Martins. 2000. Somatic growth model of juvenile loggerhead sea turtles: duration of the pelagic stage. Mar. Ecol. Prog. Ser. 202:265-272.

Bolten, A.B., K.A. Bjorndal, and H.R. Martins. 1994. Life history model for the loggerhead sea turtle (Carettacaretta)populations in the Atlantic: Potential impacts of a longline fishery.

U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SWFC-201:48-55.

Bolten, A. B., K. A. Bjorndal, H. R. Martins, T. Dellinger, M. J. Boscoito, S. E. Encalada, and B.

W. Bowen. 1998. Transatlantic developmental migrations of loggerhead sea turtles demonstrated by mtDNA sequence analysis. Ecol. Applic. 8:1-7.

Bowen, B.W., J.C. Avise, J.I. Richardson, A.B. Meylan, D. Marganitoulis, and S.R.

Hopkins-Murphy. 1993. Population structure of loggerhead turtles (Carettacaretta)in the northwestern Atlantic Ocean and Mediterranean. Sea. Conserv. Biol. 7:834-844.

Burke, V.J., S.J. Morreale, and E.A. Standora. 1994. Diet of the Kemp's ridley sea turtle, Lepidochelys kempii, in New York waters. Fish. Bull. 92:26-32.

Cannon, A.C., C.T. Fontaine, T.D. Williams, D.B. Revera, and C.W. Caillouet, Jr. 1994.

Incidental catch of Kemp's ridley sea turtles (Lepidochelys kempi), by hook and line, along the Texas Coast, 1980-1992. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-341:40-42.

Crouse, D. 1993. Victory! TEDs required in all U.S. shrimp trawls. Marine Turtle Newsletter 61:3-5.

Crouse, D.T., L.B. Crowder, and H. Caswell. 1987. A stage-based population model for loggerhead sea turtles and implications for conservation. Ecol. 68:1412-1423.

Crouse, D. T., M. Donnelly, M. J. Bean, A. Clark, W. R. Irvin and C. E. Williams. 1992. The TED experience: claims and reality. Center for Marine Conservation, Environmental Defense Fund and National Wildlife Federation, Washington, D.C., 17 pp.

Crowder, L.B., D.T. Crouse, S.S. Heppell, and T.H. Martin. 1994. Predicting the impact of turtle excluder devices on loggerhead sea turtle populations. Ecol. Applic. 4:437-445.

76 NOAA Tech. Mem. NMFS-SEFSC-444

Crowder, L. B, S. R. Hopkins-Murphy, and J. A. Royle. 1995. The effects of turtle excluder devices (TEDs) on loggerhead sea turtle strandings with implications for conservation.

Copeia 1995:773-779.

Epperly,S.P., J. Braun, A. J. Chester, F. A Cross, J.V. Merriner, and P.A. Tester. 1995. Winter distribution of sea turtles in the vicinity of Cape Hatteras and their interactions with the summer flounder trawl fishery. Bull. Mar. Sci. 56:547-568.

Epperly, S.P., J. Braun, A. J. Chester, F. A Cross, J.V. Merriner, P.A. Tester, and J.H. Churchill.

1996. Beach strandings as an indicator of at-sea mortality of sea turtles. Bull Mar. Sci.

59:289-297.

Henwood, T., W. Stuntz and N. Thompson. 1992. Evaluation of U.S. turtle protective measures under existing TED regulations, including estimates of shrimp trawler related turtle mortality in the wider Caribbean. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-303, 15 pp.

Iversen, E. S., D. M. Allen and J. B. Higman. 1993. Shrimp capture and culture fisheries of the United States. Halsted Press, New York; N.Y., 247 pp..

Johnson, D.R., C. Yeung, and C.A. Brown. 1999. Estimates of marine mammal and marine turtle bycatch by the U.S. Atlantic pelagic longline fleet in 1992-1997. U.S. Dep. Commer.

NOAA Tech. Mem. NMFS-SEFSC-418, 70 pp.

Laurent, L., J. Lescure, L. Excoffier, B. Bowen, M. Domingo, M. Hadjichristophorou, L. Kornaraki, and G. Trabuchet. 1993. Etude g&ndtique des relations entre les populations m~diterran~enne et atlantique d'une tortue marine (Carettacaretta)Ai1'aide d'un marqueur mitochondrial. C.R. Acad. Sci. Paris 316:1233-1239.

Laurent, L, P. Casale, M.N. Bradai, B.J. Godley, G. Gerosa, A.C. Broderick, W. Schroth, B.

Schierwater, A.M. Levy, D. Freggii, E.M. Abd El-Mawla, D.A. Hadoud, H.E. Gomati, M.

Domingo, M. Hadjichristophorou, L. Kornaraky, F. Demirayak, and Ch. Gautier. 1998.

Molecular resolution of marine turtle stock composition in fishery bycatch: a case study in the Mediterranean. Molecular Ecol. 7:1529-1542.

McDonald, D. 1990. Shrimpers arrested for non-compliance as efforts to enforce TED regulations intensify. Marine Turtle Newsletter 51:10-12.

National Marine Fisheries Service and U.S. Fish and Wildlife Service. 1991. Recovery plan for U.S. population of loggerhead turtle Caretta caretta. National Marine Fisheries Service, Washington, D.C., 64 pp.

National Research Council. 1990. Decline of the sea turtles: causes and prevention. National Academy Press, Washington, D.C., 259 pp.

NOAA Tech. Mem. NMFS-SEFSC-444 77

Quantum Resources. 1995. Annual environmental operating report for 1994. Unpublished report.

Florida Power and Light Company, Juno Beach, Fla., 49pp.

Sadove, S.S. and S.J. Morreale. 1990. Marine mammal and sea turtle encounters with marine debris in the New York Bight and the northeast Atlantic. U.S. Dep. Commer. NOAA Tech.

Mem. NMFS-SWFSC- 154:562-570.

Schmid, J.R. and W.J. Barichivich. submitted. Kemp's ridley turtle, Lepidochelys kempii. In:

Turtles of Florida. Chelonian Conserv. Biol. Monogr.

Trent, L., D.E. Parshley, and J.K. Carlson. 1997. Catch and bycatch in the shark drift gillnet fishery off Georgia and East Florida. Mar. Fish. Rev. 59:19-28.

Turtle Expert Working Group. 1998. An assessment of the Kemp's ridley (Lepidochelys kempii) and loggerhead (Carettacaretta)sea turtle populations in the Westem North Atlantic. U.S.

Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-409, 96 pp.

Ulloa Ramirez, P.A. and L.V. Gonzdlez Ania. 2000. Incidence of marine turtles in the Mexican longline tuna fishery in the Gulf of Mexico. U.S. Dep. Commer. NOAA Tech. Mem.

NMFS-SEFSC-436:110.

U.S. Fish and Wildlife Service and National Marine Fisheries Service. 1992. Recovery plan for the Kemp's ridley sea turtle (Lepidochelys kempii). National Marine Fisheries Service, St.

Petersburg, Fla., 40 pp.

Weber, M., D. Crouse, R. Irvin, and S. Indicello. 1995. Delay and denial: a political history of sea turtles and shrimp fishing. Center for Marine Conservation, Washington, D.C., 46 pp.

Witherington, B.E. 1994. Flotsam, jetsam, post-hatchling loggerheads, and the advecting surface smorgasbord. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-351:166-168.

Witherington, B.E. in review. Ecology of neonate sea turtles inhabiting debris lines near the Gulf Stream front. Mar. Biol.

Witzell, W.N. 1999. Distribution and relative abundance of sea turtles caught incidentally by U. S.

longline fleet in the western North Atlantic Ocean, 1992-1995. Fish. Bull. 97:200-211.

Witzell, W.N. and W.G. Teas. 1994. The impacts of anthropogenic debris on marine turtles in the Western North Atlantic Ocean. U.S. Dep. Commer. NOAA Tech. Mem. 355:1-21.

78 NOAA Tech. Mem. NMFS-SEFSC-444

Table 20. Total shrimping vessels and vessels operating in the Campeche and Tamaulipas fleets, 1982-1995.

Campeche Fleet Tamaulipas Fleet Year Total Operating Total Operating 1982 605 578 1983 575 507 1983 544 481 1985 512 357 1986 488 354 1987 469 368 unknown 124 1988 450 324 unknown 116 1989 423 294 unknown 139 1990 412 259 unknown 134 1991 356 225 unknown 145 1992 398 293 unknown 169 1993 403 325 unknown 163 1994 347 268 unknown 154 1995 352' 292 unknown 199 NOAA Tech. Mem. NMFS-SEFSC-444 79

Part 4. Take Limits Introduction One of the initial charges to the TEWG was to determine maximum take levels for each species. This is the maximum number of turtles that can be removed by human activities (i.e., above losses due to natural mortality) without jeopardizing recovery goals specified in each Species Recovery Plan. While allocation of take is a regulatory action clearly beyond the scope of the Group, a scientific assessment of the population-level effects of current and maximum allowable anthropogenic mortality to achieve recovery goals is needed. In general, it was agreed that in order to meet recovery goals, "take" should not increase at a rate equal to or greater than the population growth rate for either species. However, reliable indicators of take levels and population growth were generally unavailable. The Group determined early on that observer coverage on shrimp vessels was an inadequate measure of fishinig mortality; less than 1% of shrimp vessel trips include observers and most observers are on offshore vessels that catch few turtles, and there are few observer programs designed to document turtle take by other fisheries. Thus, strandings became the Group's proxy for take. Because the following key information was contentious or lacking, no method for setting strandings limits was completely satisfactory to all Group members:

Populationsize. We do not have a reliable estimate of the number of turtles that are susceptible to fishing mortality in a given year. While there is a fairly good estimate of nest numbers, many assumptions must be made to translate nests into adult females, adult males, and juveniles of both sexes. For loggerheads, an approach that utilized the ratios of males:females and adults:juveniles in the strandings records gave estimatef of 183,627 -

316,113 turtles greater than 40 cm SCL (Table 3; TEWG, 1998). Because we have a time series of cohort sizes for Kemp's ridleys through records of hatchlings released at Rancho Nuevo, a population model was developed to determine total population size annually (see Part 1., above). However, the estimate is completely dependent on unknown mortality rates, and varies by 10-fold. Both methods of estimating population size suffer from a lack of corroborating empirical estimates - we do not even know if we are within the correct order of magnitude.

Proportionof total mortality representedby strandings. In order to estimate population size with the ratio method and to set take limits based on strandings, assumptions must be made about the relationship of strandings and the turtle population. There are two issues: whether the size distribution of stranded turtles is representative of the population as a whole; and what proportion of total mortality is represented by stranded turtles. The probability that a dead turtle will wash up on a beach is undoubtedly highly variable, depending on distance from shore, water temperature, wind and currents, and possibly body size. Fishing mortality may be size-selective and is also variable. The group agreed that a majority of turtles that

• are killed in fishing gear fail to strand, but an actual proportion could not be agreed upon because existing data are not adequate to evaluate this.

80 NOAA Tech. Mem. NMFS-SEFSC-444

Populationdistribution. Turtles of different sizes are not .evenly distributed in time and space. Migrating turtles may be exposed to different levels of risk from fishing mortality as their overlap with fishing fleets varies through the year. Juvenile as well as adult loggerheads from genetically distinct subpopulations appear to overlap disproportionately in space, but the genetics of turtles caught by various fishing fleets has not been determined.

This has profound implications for the ratio method of population size estimation and for setting take limits based on strandings.

Several other unknowns also impeded development of take limits, including age at maturity (to determine time lags). Any one of these data deficiencies could have been surmounted with careful analysis of the best information available; the lack of consensus about take limits primarily rose from the unknown consequences of "stacking" so many uncertainties. Setting a take limit or estimating the impact of take is necessarily a quantitative exercise that requires a substantial amount of information and a good understanding of how data uncertainty will affect the results (Mangel, 1993; Taylor et al., 2000). Although three methods to set strandings limits were used at various times by the TEWG, each was presented as an Interim Stranding Limit (ISL) to be.used for a single.

year, rather than an agreed-upon method to be incorporated into management.

Three Interim Sirandings Limits (ISL) calculations: Meihods and Limitations

1) Running average-In 1996 and 199 7, Kemp's ri dley ISLs were calculated according to a 5 year running average of strandings + 1 standard deviation, which was then multiplied by one-half the observed rate of increase in the nesting population. The observed rate of increase in nests was estimated as the slope of the natural log of nest levels over the past 5 years, transformed to a linear (proportional) scale (TEWG, 1998). Stranded Kemp's ridley turtles greater than 10 cm, including those from cold-stunning events, were used in these analyses. In 1997, the stranding data were partitioned according to geographic regions (western Gulf of Mexico, zones 13 - 21; eastern Gulf, zones 1712 and partial zones 24 and 25; southeastern U.S. Atlantic, partial zones 24 and 25,26-35, partial zone36; northeastern U.S. Atlantic, partial zone 36, 37 - 44). Similar methods were used for calculating loggerhead ISLs in 1996 and 1997 with the exception of the nesting multiplier and the inclusion of cold-stunned turtles. During both years, stranded loggerhead turtles greater than 40 cm were partitioned among the four regions. In 1997, the standard deviation was not added to the running averages of loggerheads on the Atlantic coast in order to give more conservative ISL estimates.

Furthermore, log-normal means were investigated for both species in 1997, but the NMFS Southeast Regional Office, when presented with two sets of ISLs, chose to continue with arithmetic means. In 1998 the Group agreed that the running average method was inadequate for calculating ISLs, primarily because this method does not account for changes in fishing mortality, growing concern about the status of the Northern Subpopulation of loggerheads, and the large number of loggerheads stranding and the size distribution of those animals.

NOAA Tech. Mem. NMFS-SEFSC-444 81

..... * ' ,~~~~~~~V

. *. e" '":' "' ** * ' "" ' :" : " ;" * "

2) Slope method For Kemp's ri dleys, it appeared that in many years the rate of increase of strandings exceeded the observed rate of increase of nests and hatchlings at Rancho Nuevo. A new method was introduced in 1998 that worked on the key assumption that to maintain and enhance population growth that rate of increase of strandings should be lower than the population rate of increase. This method relied on the estimated slopes of hatchling production and strandings and set the ISL for the following year to obtain a desired slope. It was used on a trial basis to calculate an ISL for Kemp's, as a full analysis of potential changes in population size and strandings on the ISLs had not been attempted. As a measure of population growth rate, the group agreed to use hatchling production 2 years prior to the strandings year. Best estimates from correlation studies suggest that turtles spend two years at sea (see Life Stage Definitions in the Kemp's Ridley Status Report, above). This is corroborated by the new catch curve analysis, which shows peak strandings age of 2-3 years old in most strandings years (see Parameterization in the modeling section of the Kemp's Ridley Status Report, above). The strandings years used in the analysis were 1986-1997 for all regions, and did not include headstarted, TED-tested, or cold-stunned turtles and also excluded post-hatchlings (< 10 cm).

Slope calculation Hatchlings released 2 years prior and strandings were In-transformed to calculate and compare slopes (Table 21). Slopes and regression statistics-were calculated with Excel 5.0. Over the 12 year time period the two slopes were nearly identical (m = 0.8 for hatchlings and m = 0.78 for strandings). Strandings were much more variable than hatchlings, but the regression was still significant at the 0.05 level (p = 0.028). The standard error of the hatchling slope was 0.011.

Setting the ISL Strandings in 1998 should not exceed a value that maintains or lowers the slope to a desired level. To set the ISL conservatively, the group agreed that the desired strandings slope should be I standard error below the slope observed for hatchling production. Using a minimization program in Excel (Solver), the number of allowable strandings in 1998 was calculated to give a strandings slope of 0.069 over the years 1986 - 1998 (0.069 = observed slope of 0.08 - 0.011, the standard error). That point estimate was 334 turtles (Figure 21).

Caveats of the slope method The group agreed that this was an interim method in need of further research, but was a more appropriate and conservative estimate than the running average method. Several concerns were raised. First, the goal of the method must be determined: do we want to reduce the rate of strandings increase to assure that the population will continue to increase rapidly, or is the "conservative factor" of reducing the slope by I standard error unnecessary? Regardless of whether the average slope of the strandings is greater than the population slope, if the slopes are not statistically different from each other a reduction in the slope may not be needed. One problem with 82 NOAA Tech. Mem. NMFS-SEFSC-444

determining statistical differences between the two slopes is that their variances are not equal (p =

0.35, one-tailed F test), which make standard slope comparisons (t-tests) invalid. Second, setting the strandings limit for a single year may not be feasible if the desired slope is very different from the observed strandings slope. This will depend on the number of years used in the analysis; 12 years may be inappropriate for Kemp's because it includes hatchling production years when the population was declining and spans more than the current predicted generation time (10 years).

3) PotentialBiologicalRemoval (PBR)

Potential Biological Removal (PBR) was developed for marine mammal stocks (Barlow et al., 1995), which share many life history traits and data restrictions with sea turtles. PBR is based on the concept that human-caused mortality of a protected mammal stock should not exceed one-half the potential net productivity rate of the population, adjusted by a recovery factor (F) that varies from 0 to 1. It is a simple equation that requires a minimum population size estimate (Nmin), the maximum rate of increase predicted (or measured) for a population (Rmax), and pre-determined risk criteria for the recovery factor:

PBR = Nmiin x (0.5 x Rmax') xF Wade (1998) ran simulations for a number of marine mammals to determine appropriate recovery factors for endangered and threatened populations. The default F values are: widespread populations with little risk of depletion, F = 1.0, threatened species F = 0.5, and endangered species F = 0.1. It has been determined that a stock experiencing a level of human-caused mortality exceeding PBR could become depleted (Wade 1998). Although data do not exist to conduct comparable simulations for sea turtle stocks, there are enough life history similarities between the two taxa (relatively low fecundity, advanced age at maturity, longevity, etc.) that the majority of the TEWG felt the PBR concept could be applied to sea turtle stocks. Discussions with the developers of the marine mammal PBR methodology have suggested that with modification, such as calculating separate PBRs for each life stage, PBR is potentially applicable to sea turtles.

Modifications should account for the large difference in the reproductive value and abundance of different life stages and the extreme delay in maturity for some species.

PBR calculations for sea turtles The PBR estimate varies depending on each of the three required parameters. As an example, Figure 22 shows changes in PBR with various recovery factors when Rmax is set at 0.12 (i.e., observed rate of increase for Kemp's ridleys, 1985-1998) or 0.04 (observed rate of increase for loggerhead nests in South Florida, 1989-1998). Note that these rates of increase are probably below Rmax because of incidental mortality. For a population of 2500 turtles, total human-induced take would range from 10 to 1000 individuals. The lowest PBRs occur for very slow-growing populations, where additions to the population (recruitment) each year are only a small proportion of the total population size. While PBR numbers can be very low if conservative recovery factors are used, it is important to realize that PBR represents a conservative estimate of allowable take that will not prevent population recovery or cause a population decline (Wade, 1998). Also, if PBR is NOAA Tech. Mem. NMFS-SEFSC-444 83

used to determine ISLs, additional multipliers are required to reduce total take by the proportion of animals killed that are likely to be counted as strandings.

Application to adult loggerheads Because loggerheads have such a long lag time between hatching and maturation, the slope and running average methods were judged to be inadequate for protection of the threatened Northern Subpopulation by a majority of the Group in 1998. PBR was used to calculate the number of adult loggerhead strandings allowed in southern Atlantic fishing zones. Strandings for all of the zones within an area were used to determine sex ratios of adults. Also, nesting data were combined for the S. Florida and the Florida Panhandle subpopulations. PBR for adult loggerheads was calculated assuming the minimum population size was the estimated size of the adult populations in U.S. waters, the maximum rate of increase was 5%, and the recovery factor was 0.5, given the species is listed as threatened. It was assumed that strandings represent, at best, 25% of the at-sea nearshore mortality (Murphy and Hopkins-Murphy, 1989). Thus, the ISL was set as 25% of the calculated PBR for each arealsubpopulation (Table 22) The result was a mean of 9 adult strandings in the Northern Subpopulation zones and 98 adult strandings in all remaining areas. In addition to the assumptions of the ratio method which are listed in TEWG (1998), this method assumes that adult individuals of the Northern Subpopulation are found in zones 29-44 only, or are outside U.S.

waters. ISLs for turtles <92 cm SCL (and unmeasured) were calculated by region using the running average method.

Caveats of PBR PBR has yet to be evaluated for sea turtles, contains a number of simplifying assumptions, and the default recovery factors are considered arbitrary by some researchers. A minority opinion to the ISL recommendations for 1998 strongly criticized the use of PBR and the emphasis on separating the loggerhead ISLs by subpopulation. None of the parameters in the PBR equation have been fully evaluated, including the strandings:deaths ratio required to set ISLs.. Finally, PBR does not include much of the important biological information available for sea turtles, such as life history and sensitivity of various life stages.

Summary of results 1996- 1998 Strandings of Kemp's ridleys during 1996-1998 for comparison to the ISLs set were 393, 458, and 412, respectively fTable 23). Strandings of Kemp's ridleys in 1996 and 1997 were less than the ISLs set for those years. However, strandings of Kemp's ridleys in 1998 (412) exceeded the lSL of 334 animals by nearly 25%.

Total strandings of loggerheads during 1996-1998 exceeded total ISLs by 12-82% each year (Table 23). In 1996 ISLs were exceeded in all regions except the northeast U.S. and in 1997 ISLs were exceeded in all regions. In 1998 ISLs for juvenile loggerheads were exceeded in all regions except the Eastern Gulf of Mexico. The number of adults stranding in zones 29-44, the area where all adults are presumed to be from the Northern Subpopulation, was extraordinary (83) exceeding 84 NOAA Tech. Mem. NMFS-SEFSC-444

ISL (9) by 922%, even exceeding the number of adults stranding in other zones where they are presumed to be from the much larger (nearly IOx) S. Florida Subpopulation and from the Florida Panhandle Subpopulation.

Research recommendations forfuture evaluation of take limits Take limits are a contentious issue because data on sea turtles are extremely limited and scientific opinions vary on the importance of conservative recovery factors. Better estimates of means and uncertainty in population size, the strandings:deaths ratio, and population distribution will help, but the level of allowable risk must be established first. As long as sea turtle populations appear to be recovering, it is unlikely that disagreements about risk aversion will be resolved. The distinctness of the loggerhead subpopulations is critical, as the nests of Northern Subpopulation beaches do not appear to be increasing. Because the goals stated in the Recovery Plan specify nest increases for these areas, thoughtful evaluation of management priorities is warranted.

A simulation analysis is 'Underwayto examine the effects of population change and variance, number of years analyzed, and parameter uncertainty, in the ISLs calculated using the slope method,.

running averages, and PBR (S. Heppell, personal communication). It mayalso be possible to use the Kemp's ridley age-based population model for projections through time to see how the population might be affected by these take levels, but that will require an estimate of the strandings:total mortality proportion. A future meeting of population modelers familiar with a variety of methods for analyzing sea turtle and fishery data may also be beneficial, particularly if a working simulation model could be provided to test various take limits.

Literature Cited Barlow, J, S.L. Swartz, T.C. Eagle, and P.R. Wade. 1995. U.S. Marine Mammal Stock Assessments: Guidelines for Preparation, Background, and a Summary of the 1995 Assessments. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-OPR-6, 73 pp..

Mangel, 1993. Effects of high seas driftnet fisheries on the northern right whale dolphin Lissodelphis borealis. Ecol. Applic.3:221-229.

Murphy, T. and S.R. Hopkins-Murphy. 1989. Sea Turtle & Shrimp Fishing Interactions: a summary and critique of relevant information. Center for Marine Conservation, Washington, D.C., 52 pp.

Taylor, BL., P.R. Wade, D.P. DeMaster, and J. Barlow. 2000. Incorporating uncertainty into management models for marine mammals: Conserv. Biol. 14:1243-1252.

Turtle Expert Working Group. 1998. An assessment of the Kemp's ridley (Lepidochelys kempii) and loggerhead (Carettacaretta) sea turtle populations in the Western North Atlantic. U.S.

Dep. Commer. NOAA Tech. Mem. NMFS-SEFSC-409, 96 pp.

NOAA Tech. Mem. NMFS-SEFSC-444 85

Wade, P. 1998. Calculating limits to the allowable human-caused mortality of cetaceans and pinnipeds. Mar. Mammal Sci. 14:1-37.

86 NOAA Tech. Mem. NMFS-SEFSC-444

Table 21. Hatchling production and strandings levels used to calculate slopes.

Year Hatchlings produced two Ln Strandings Ln(strandings) years earlier (hatchlings) 1986 58124 10.970 306 5.724 1987 51033 10.840 162 5.088 1988 48818 10.796 216 5.375 1989 44634. 10.706 150 5.011 1990 62218 11.038 258 5.553 1991 66752 11.109 165 5.106 1992 74339 11.216 156 5.050 1993 76238 11.242 327 5.790 1994 92116 11.431 550 6.310 1995 84210 11.341 396 5.981 1996 107687 11.587 382 5.945 1997 120038 11.696 412 6.021 slope 0.080 0.078 standard error 0.01 I 0.030 p-value <0.001 0.027 Lower 95% C.I. 0.056 0.010 Ul1ner 95%=CI1 0. 0.146 NOAA Tech. Mem. NMFS-SEFSC-444 87

Table 22. PBR and ISL estimates for adult loggerheads, 1998.

NORTHERN SUBPOPULATION INPUT DATA Number Stranded Animals >=92 cm, May-August Est. No. Estimated Est. Adult of Nests Est. Prop. of Nesting Number of Pop. In Year Z. 29-44 Females Males Unknowns Females Females Adult Males U.S. Waters PBR ISL 1989 4372 3 5 17 0.46 1066 1252 2318 29 7 1990 7746 13 6 17 0.60 1889 1274 3163 40 10 1991 6749 10 6 19 0.56 1646 1308 2955 37 9 1992 6487 - 9 12 13 0.46 1582 1888 3471 43 11 1993 4528 6 13 17 0.40 1104 1638 2742 34 9 1994 7899 22 16 20 0.55 1927 1565 3492 44 11 1995 5798 18 15 28 0.52 1414 1282 2696 34 8 1996 6940 18 23 29 0.46 1693 1953

• 3646 46 11 1997 4220 18 22 16 046 1029 1188 2217 28 7 Mean 6082 0.50 1483 1498 2982 37 9 SOUTH FLORIDA SUBPOPULATION AND FLORIDA PANHANDLE SUBPOPULATION INPUT DATA Number Stranded Animals >=92 cm, May-August Est. No. Estimated Est. Adult of Nests Est. Prop. of Nesting Number of Pop. In Year Z. 1-28 Females Males . Unknowns Females Females Adult Males U.S. Waters PBR ISL 1989 48644 8 13 15 0.43 11864 15692 27556 344 86 1990 65330 14 7 14 0.60 15934 10623 26557 332 83 1991 67662 11 9 15 0.53 16503 14719 31222 390 98 1992 64214 .15 10 12 0.57 15662 11933 27595 345 86 1993 55012 12 16 10 0.45 13418 16575 29992 375 94 1994 70387 11 8 20 0.54 17168 14715 31883 399 100 1995 79508 6 12 5 0.37 19392 33081 52473 656 164 1996 75526 19 9 19 0.61 18421 11957 30378 380 95 1997 63193 12 13 11 0.49 15413 16294 31707 396 99 M ean 65497 0.51 15975 15454 31429 393 98 88 NOAA Tech. Mem. NMFS-SEFSC-444

z0 Table 23. Summary of Incidental Stranding Limits (ISL) and total strandings for 1996-98. ISL methods for loggerheads in 1998 are from the majority report.

'H Region Size Cold-stunned Total Cs C) Year Species (zones) classes' included Calculation method ISL strandings 1996 L. kempii All > 10cm yes (Arithmetic mean + I std dev)*multiplier 586 393 Cs C. caretta Western Gulf > 40 cm no Arithmetic mean + 1 std dev 164 206 (13-21) z Eastern Gulf > 40 cm no Arithmetic mean + 1 std dev 127 204 (1 - 12, part 24 & 25)

-n Southeast US coast > 40 cm no Arithmetic mean + 1 std dev 803 1131 rJ) (part 24 & 25<- 35, part 36)

Northeast US coast > 40 cm no Arithmetic mean + 1 std dev 269 244 Cd) (part 36, 37 - 44) 19972 L. kempii Western Gulf > 10 cm yes (Arithmetic mean + 1 std dev)*multiplier 311 228 Eastern Gulf > 10 cm yes (Arithmetic mean + I std dev)*multiplier 98 80 Southeast US coast > 10 cm yes (Arithmetic mean + I std dev)*multiplier 117 93 Northeast US coast > 10 cm yes (Arithmetic mean + I std dev)*multiplier 113 57 C. caretta Western Gulf > 40 cm no Arithmetic mean + 1 std dev 192 170 Eastern Gulf > 40 cm no Arithmetic mean + 1 std dev 174 197 Southeast US coast > 40 cm no Arithmetic mean without std dev 792 892 Northeast US coast > 40 cm no Arithmetic mean without std dev 233 311 1998 , L. kcenpii All > 10 cm no Slope method 334 412 C. caretta S. Fla & Panhandle subpop.s 92 cmi no PBR estimate 98 59 adults (1 - 28)

Northern subpopulation 92 cm no PBR estimate 9 83 (29- 44)

C. caretta Western Gulf 40 - 91 cm no Arithmetic mean without std dev 129 170 juveiniles & unmeasured Eastern Gulf 40 - 91 cm no Arithmetic mean without std dev 131 123

& unmeasured Southeast US coast 40- 91 cm no Arithmetic mean -2 std dev 380 904

& unmeasured Northeast US coast 40 - 91 cm no Arithmetic mean - 2 std dev 97 343

& unmeasured 1.Unmeasured turtles were included.

2

.In 1997, methods based on both geometric and arithmetic 5-yr averages were used. However, the NMFS Southeast Regional Office chose to use ISLs based on the arithmetic means.

Figure 21. Slope method for ISL, showing strandings level required to achieve desired slope of m=0.069.

6.5 T 6.3 + x E 6.1 -

0_ x 0 5.9 -

4" x ..... Strandings 5.7 - x old slope = 0.078 X

new slope = 0.069

(_ 5.5 -

e- 5.3 -

Cm I- 5.1 - x x 05 Cz n 4.9 -

4.7 -

4.5-1984 1986 1988 1990 1992 1994 1996 1998 2000 (A-m7

z0 Figure 22. Effect of the recovery factor (F) on PBR for.a population of 2500.

160 z aL 140 0 120 E

100 0

80 0

60 (m

0 40 CL 20 0~

.0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Recovery Factor (F)

Part 5. Research Recommendations Progress towards developing realistic population level models is dependent upon obtaining empirically derived parameters which reflect and incorporate the life history of each species. To this end, the information required includes empirically estimated survival rates for each life hi story stage defined by age. The assignment of ages and stages requires the application of age-length keys developed from growth models, which are used to develop the age structure for the population. The evaluation of changes in the age structure of the population can be completed by examining shifts in size frequency distributions over time and between areas. Shifts in abundance and the age structure must be evaluated in the context of migration in and out of areas identified as developmental or important foraging areas. This information can only come from research on turtles in the water and requires the accumulation of mark-recapture information over several years in multiple areas of known utilization. These in-water areas must be sampled to ensure that growth models, abundance indices, and size/age frequency distributions can be obtained. The distributions for these must be described with some measure of variability so that trends and shifts can be evaluated. Problems with data collected from strandings to estimate age and growth and survivorship have been demonstrated and underscored the need to conduct long term in-water studies that are comparable between years and between areas.

In-water surveys will provide data to evaluate changes in the age structure of populations.

This trend analysis should enable us to evaluate progress towards recovery as defined in the individual recovery plans. Currently, data on the number of nesting females or nests is used to evaluate recovery and to ensure recovery. The timeline to the fully recruited benthic stage is shorter than that for returning to the nesting beach; especially for loggerhead turtles. Tracking recovery requ ires tracking juvenile turtles to ensure recovery rather than waiting the many years to maturity, which may be too late to reverse any declines.

Strandings data continue to provide information about unusually large mortality events that can be attributed to either a natural event or anthropogenic activity. These data also continue to provide information on sex ratios, maturity, breeding condition, and size frequencies. Strandings can be used to determine population demographics and to estimate mortality rates, but only if they are representative of the live turtle population. This assumption must be verified, as it continues to limit the interpretation of results that rely on these data. The relationship between the location and occurrence of strandings and environmental factors such as tides, currents, water temperature, resource condition and quantity must be determined to fully understand where the mortality occurred relative to where the stranding is reported. Stranding protocols have been recently standardized to provide comparable information between areas and time periods; accumulation of these data over several years may provide trend information if strandings are representative of the living population.

Research must continue to pursue ways to age turtles and to validate these ages. Body parts, such as the humerus, ear bone, or other hard structures that may harbor growth rings should be carefully preserved for all specimens collected by the Sea Turtle Stranding and Salvage Network, particularly body parts from known-age turtles. Validation of ages determined by humerus or sclera 92 NOAA Tech. Mem. NMFS-SEFSC-444

growth should be attempted through the use of hard part markers. The use of internal wire coded tags placed in small Kemp's ridleys should be continued to provide aging information. However, the success of this effort can only be ensured through the distribution of tag detectors throughout the range for this species and for use with both dead and live turtles.

While estimating total mortality rates empirically is an immediate and critical requirement, the partitioning of mortality by cause is imperative for the management of these species if recovery goals are not being met. When mortality can be partitioned and identified by cause, then management measures can be implemented to eliminate or reduce controllable anthropogenic sources of mortality. Interactions with fisheries gear can be fully evaluated and quantified with the placement of observers at a statistically valid sampling level to measure incidental take and mortality by this component on a fishery by fishery basis. There is no substitute for the direct observation of take and mortality as evidenced by the effort to quantify these in the commercial shrimp trawl fishery of the southeast U.S. Repeated mortality incidents occur where high levels of strandings are reported coincident with anecdotal observations of nearshore trawling activities, but the lack of observer coverage on nearshore vessels prevents verification and quantification of a causal relationship between trawling activity and subsequent stmndings.

Genetic studies must be continued to fully understand and quantify the stock structure of the loggerhead turtle within U.S. Atlantic waters. Currently, the TEWG describes sub-populations based on nesting location as supported by recent genetic studies. However, these studies are limited to female turtles and do not provide any measure of fidelity by males. Foraging ground studies provide information on the mixing of subpopulations. The extent to which cross over by males occurs likely will determine the actual stock structure for this species. However, the TEWG notes that the natal homing of female turtles, as verified by existing genetic data, suggests that a nesting assemblage could be lost for ecological time periods if a subpopulation declines to extinction.

The pursuit of stock assessments is contingent upon these recommendations to enhance existing research programs and develop new ones that can determine the status and condition relative to recovery of these species. Limitations of the existing data and time series continue to be the critical impediment to completing these analyses. Fine-tuning of data that have not been collected for the purpose of providing empirically based population parameters will not further the efforts towards a stock assessment. Peer reviewed results are certainly preferable and ideal for inclusion in such stock assessments. It is the conclusion of the TEWG that further pursuit of stock assessment analyses is unwarranted until these data and results are available to the group.

NOAA Tech. Mem. NMFS-SEFSC-444 93

Appendix 1. Von Bertalanffy Growth Models for Wild Kemp's Ridley Turtles:

Analyses of the NMFS Miami Laboratory Tagging Database Jeffrey R. Schmid and Amy Woodhead NationalMarine Fisheries Service Southeast Fisheries Science Center 75 Virginia Beach Drive Miami, FL 33149 Prepared for The Turtle Expert Working Group Originally submitted:

November 17, 1997 Modified and resubmitted:

October 21, 1998 A number of von Bertalanffy growth models have been developed for the Kemp's ridley turtle, Lepidochelys kempi. Zug and Kalb (1989), Zug (1990), and Schmid (1995, 1998) prepared von Bertalanffy growth models for wild, subadult Kemp's ridley turtles, but the range of carapace lengths and sample sizes for these studies were insufficient to accurately describe growth parameters. Caillouet et al. (41995) provided a growth model for head-started Kemp's ridley turtles released off the Texas coast and subsequently recaptured in the Gulf of Mexico. Zug et al. (1997) recently refined their skeletochronological growth model by increasing the sample size and the size range of turtles. Schmid and Witzell (1997) provided von Bertalanffy equations for the combined data of tagging studies in Florida (Schmid, 1995, 1998) and subsequent recaptures at Ranch Nuevo, Mexico. This paper presents von Bertalanffy growth model analyses for Kemp's ridley turtles using the data available from the National Marine Fisheries Service (NMFS) Miami Laboratory Cooperative Marine Turtle Tagging Program.

Materials and Methods Tagging records for wild Kemp's ri dley turtles were compiled from several sources: NMFS Miami laboratory tagging database' 2'3 4 5 67 8 9' , Cape Canaveral from 1986 to 1991 (Schmid, 1995),

and Cedar Keys from 1986 to 1995 (Schmid, 1998). Carapace measurements were converted to metric units, as necessary, and transformed to standard straight-line carapace length (SSCL; nuchal notch to posterior end of postcentral scute) using the regression equations of Teas (1993), Schmid (1998), and Schmid and Witzell (1997). Records for a NMFS Galveston headstart turtle, a negative 94 NOAA Tech. Mem. NMFS-SEFSC-444

growth rate (-5.6 cm/yr), and an exceedingly large growth rate (493.1 cm/yr) were omitted from the analyses. Means are followed by +/- one standard deviation.

The von Bertalanffy growth interval equation (Fabens, 1965) was fitted to Kemp's ridley recapture data with a non-linear least squares regression procedure (SAS Institute, 1989) as described by Schmid and Witzell (1997). Growth models were constructed using different intervals of time between initial tagging and recapture, in order to minimize the effects of measurement errors on short-term recaptures. Furthermore, separate growth models were calculated for Kemp's ridley turtles tagged in the Gulf of Mexico and turtles tagged in the Atlantic Ocean.

Results and Discussion Ninety-six recaptures were recorded for Kemp's ridley turtles tagged in cooperative NMFS tagging projects. Of this total, fifty-eight Kemp's ridley turtles were initially tagged in the northern Gulf of Mexico (Fla. - 31, Ms. - 1, La. - 4, and Tx.- 22), and thirty-eight were initially tagged along the Atlantic coast (Fla. - 18, Ga. - 1, N.C. - 3, Va. - 4, and N.Y. - 12). Nine Kemp's ridley recaptures were at or near Rancho Nuevo, Mexico (Table 1); four of these were initially tagged in the northern Gulf (La. - I and Tx. - 3) and five on the Atlantic coast (Va. - 1, N.C. - 1, and Fla. -

3). In the Gulf, a maximum recapture interval of 3.7 years was recorded for a turtle tagged in Texas and recoveredin Mississippi. In the Atlantic, turtles tagged in Florida and North Carolina were observed nesting at Rancho Nuevo, Mexico 8.3 and 8.5 years after last capture.

Kemp's ridley turtle carapace lengths ranges from 22.0 to 65.9 cm for all initial captures and from 22.0 to 66.7 cm for all recaptures. A mean growth rate of 6.7 +/- 6.2 cm/yr (range = 0.0 - 29.1 cm/yr) was calculated for the entire database. Carapace lengths for turtles tagged in the Gulf ranged from 22.2 to 65.8 cm for initial captures and 22.4 to 66.7 cm for recaptures. Carapace lengths for turtles tagged in the Atlantic ranged from 22.0 to 65.9 cm for initial captures and 22.0 to 66.0 cm for recaptures. The mean growth rate for Kemp's ridley turtles tagged in the Gulf (7.5 +/- 6.2 cm/yr) was significantly larger ( 2 = 4.3, df= 1, p =0.04) than that of turtles tagged in the Atlantic (5.5 L 6.2 cm/yr).

The fundamental assumption of the von Bertalanffy model is that growth rate steadily decreases with increasing size and age. Straight-line regression of individual growth rates on the mean of initial and recapture carapace lengths produced negative slopes for all records, Gulf records, and Atlantic records (Fig. 1). This suggests that Kemp's ridley turtle growth rates decreased with- increasing mean carapace lengths, but correlation coefficients for the regressions were low owing to considerable variation in growth rates for 20 - 50 cm turtles.

There was very little change in the estimates of von Bertalanffy growth parameters with increasing recapture intervals, although there were increases in the standard errors of the estimates and the residual mean square errors (MSE) of the growth models (Table 2). The model for all tag records had the lowest MSE for the recapture interval data treatments and was therefore considered the most appropriate growth model (Dunham, 1978). The fitted von Bertalanffy growth equation for all recaptures combined was:

NOAA Tech. Mem. NMFS-SEFSC-444 95

Lt = 70.7 (1 - 0.9378 e-0. 2 0 3 5 t) (1)

The estimated asymptotic length for this model is larger than the mean size (64 cm converted SSCL) and slightly smaller than the maximum size (72.5 cm converted SSCL) for nesting females observed at Rancho Nuevo, Mexico (Burchfield et al., 1988).

The fitted yon Bertalanffy growth equation for Kemp's ridley turtles tagged in the Gulf of Mexico was:

Lt = 7.1.1 (1 - 0.9381 e-0 20 9 5 t), MSE = 6.6244, (2) and the equation for turtles tagged along the Atlantic coast was:

Lt = 73.2 (1 - 0.9399 e-0 166 5 t), MSE = 4.9517. (3)

The estimated asymptotic length for turtles captured in the Atlantic is slightly larger than that of the Gulf. This observation was probably the result of the different growth rates between the two areas and the inverse relationship of the von Bertalanffy parameters in the absence of larger individuals (Knight, 1968; Schmid and Witzell, 1997).

Estimating age to maturity has been the primary application of the von Bertalanffy growth models derived for Kemp's ridley turtles. The growth curve for all recaptures combined (equation 1) is illu strated in Figure 2. By using the mean size of nesting females (64 cm) as the size at maturity, estimates of age to maturity ranged from 11-12 years for all tagging records, 12-13 years for Atlantic records, and 10- 11 years for Gulf records. Minimum size of nesting females (56.0 cm converted SSCL; Burchfield et al., 1988) has also been used as an estimate of size at maturity and resulted in age to maturity estimates as early as 7-8 years for all tagging records, 8-9 years for Atlantic records, and 7-8 years for Gulf records. Another application of von Bertalanffy models has been to estimate the number of years required for turtles to grow within a specified size range (Bjorndal et al., 1995; Schmid and Witzell, 1997). Accordingly, the duration of Ogren's (1989)

Kemp's ridley coastal-benthic subadult stage (20 - 60 cm) would be 7 - 8 years for all tagging records and the Gulf records, and 8 - 9 years for the Atlantic records.

Literature Cited Bjomdal, K. A., Bolten, A. B., Coan, A. L., and Kleiber, P. 1995. Estimation of green turtle (Chelonia mydas) growth rates from length-frequency analysis. Copeia 1995:71-77.

Burchfield P. M., Byles, R., Mongrell, J. V., Bartlett, M., and Rostal, D. 1988. Report on Republic of Mexico/United States of America conservation effort on behalf of Kemp's ridley sea turtle at Playa de Ranch Nuevo, Tamaulipas, Mexico, 1988. Gladys Porter Zoo, Brownsville, Tx.,

83pp.

96 NOAA Tech. Mem. NMFS-SEFSC-444

• • ,k

• • ,* .. * * * *.*i*,::

• * ., .* *,? !* * * *; ******* *,, , . ,

• o ...** i 1.5 , i,:*. ). , ...........

Caillouet, C. W., Jr., Fontaine, C. T., Manzella-Tirpak, S. A., and Williams, T. D. 1995. Growth of head-started Kemp's ridley sea turtles (Lepidochelys kempii) following release: Chelonian Conserv. Biol. 1:231-234.

Dunham, A. E. 1978. Food availability as a proximate factor influencing individual growlh rates in the iguanid lizard Sceloporous merriami. Ecology 59:770-778.

Fabens, A. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289.

Knight, W. 1968. Asymptotic growth: an example of nonsense disguised as mathematics. J. Fish.

Res. Bd. Canada 25:1303-1307.

Ogren, L. H. 1989. Distribution of juvenile and subadult Kemp's ridley turtles: results from the 1984-1987 surveys. p. 116-123, In: Caillouet, C. W., Jr. and Landry, A. M., Jr. (Eds.),

Proceedings of the First International Symposium on Kemp's Ridley Sea Turtle Biology, Conservation and Management. Texas A&M University Sea Grant College Program, TAMU-SG-89-105.

SAS Institute Inc. 1989. SAS/STAT User's Guide, version 6, fourth ed., vol. 2. SAS Institute Inc.,

Cary, N.C. 846pp.

Schmid, J. R. 1995. Marine turtle populations on the east-central coast of Florida: results of tagging studies at Cape Canaveral, Florida, 1986-1991. Fish. Bull., U.S. 93:139-151.

Schmid, J. R. 1998. Marine turtle populations on the west-central coast of Florida: results of tagging studies at the Cedar Keys, Florida, 1986-1995. Fish. Bull. 96:589-602.

Schmid, J. R. and W. N. Witzell. 1997. Age and growth of wild Kemp's ridley turtles (Lepidochelyskempi): cumulative results of tagging studies in Florida. Chelonian Conserv.

Biol. 2:532-537.

Teas, W. G. 1993. Species composition and size class distribution of marine turtle strandings on the Gulf of Mexico and southeast United States coasts, 1985 - 1991. U.S. Dep. Commer.

NOAA Tech. Mem. NMFS-SEFSC-315, 43pp.

Zug, G. R. 1990. Estimates of age and growth in Lepidochelys kempii from skeletochronological data. U.S. Dep. Commer. NOAA Tech. Mem. NMFS-SEFC-278:285-286.

Zug, G. R. and Kalb, H. J. 1989. Skeletochronological age estimates for juvenile Lepidochelys kemnpii from Atlantic coast of North America. U.S. Dep. Commer. NOAA Tech. Mem.

NMFS-SEFC-232:271-273.

NOAA Tech. Mem. NMFS-SEFSC-444 97

Zug, G. R., Kalb, H. J., and Luzar S. J. 1997. Age and growth on wild Kemp's ridley sea turtles Lepidochelys kempii from skeletochronological data. Biol. Conserv. 80:261-268.

Text Footnotes J Bartol, S. Virginia Institute of Marine Science, P.O. Box 1346, Gloucester Point, VA 23062-1346.

2 Branstetter, S. Gulf and South Atlantic Fisheries Development Foundation, Inc., Lincoln Center, Suite 997, 5401 West Kennedy Boulevard, Tampa, FL 33609.

3 Corcoran, G. J.L. Scott Marine Education Center & Aquarium, P.O. Box 7000, Ocean Springs, MS 39566-7000.

4 Epperly, S. Southeast Fisheries Science Center, Beaufort Laboratory, 101 Pivers Island Road, Beaufort, NC 28516-9722.

5 Landry, Jr., A.M. Texas A & M University at Galveston, Department of Marine Biology, Galveston, TX 77551-5923.

6 Martin, E. Ecological Associates, Inc., P.O. Box 405, Jensen Beach, FL 34958-0405.

7 Nelson, D. A. U.S. Corps of Engineers, Waterways Experiment Station, 3909 Halls Ferry Road, Vicksburg, MS 391 80.

8 Rudloe, A. Gulf Specimen Marine Laboratories, Inc., P.O. Box 237, Panacea, FL 32346.

9 Rudloe, J. Panacea Institute of Marine Science, P.O. Drawer AB, Panacea, FL 32346.

io Sadove, S. Puffin Consulting, Inc. P.O..Box 361, Jamesport, NY 11947; Okeanos Ocean Research Foundation, Inc., 431 E. Main Street, Riverhead, NY 11901-2556.

98 NOAA Tech. Mem. NMFS-SEFSC-444

Table 1. Recoveries of Kemp's ridley turtles initially tagged in the U.S. and subsequently recaptured at or near Rancho Nuevo, Mexico.

Date Initial Recapture Years at Tagging Location Tagged SCL SCL Large Atlantic Virginia' 6/19/89 46.6 64.9 6.9 Florida 2 7/24/89 54.8 64.9 4.8 Florida3 1/11/89 42.8 63.8 8.3 Florida3 9/30/90 36.2 62.1 5.6 North Carolina4'

• 10/26/89 40.5 63.0 8.5 Gulf of Mexico Texas5 11/15/92 33.9 35.0' 0.1 Texas5 5/1/94 56.0 65.8 3.0 Louisiana 5 8/13/94 65.8 66.2 0.7 Texas 6,* 3/5/98 62.2 63.0 0.2

• Not included in von Bertalanffy growth models.

Bartol, S. Virginia Institute of Marine Science, P.O. Box 1346, Gloucester Point, VA 23062-1346.

2 Martin, E. Ecological Associates, Inc., P.O. Box 405, Jensen Beach, FL 34958-0405.

3 Schmid and Witzell (1997) 4 Epperly, S. Southeast Fisheries Science Center, Beaufort Laboratory, 101 Pivers Island Road, Beaufort, NC 28516-9722.

5 Landry, Jr., A.M. Texas A & M University at Galveston, Department of Marine Biology, Galveston, TX 77551-5923.

6 Branstetter, S. Gulf and South Atlantic Fisheries Development Foundation, Inc., Lincoln Center, Suite 997, 5401 West Kennedy Boulevard, Tampa, FL 33609.

Recaptured in Laguna Madre de Tamaulipas, Mexico.

NOAA Tech. Mem. NMFS-SEFSC-444 99

Table 2. Estimated values of asymptotic length (a), intrinsic growth rate (k), and means square error (MSE) from non-linear regression of von Bertalanffy growth interval equation for Kemp's ridley turtles (one asymptotic standard error in parentheses).

Data Treatment n a k MSE All tag records 96 70.7 cm 0.2035 5.9436 (2.7) (0.0251)

All tag records > 90 days 62 70.9 cm 0.2005 8.7846 (3.3) (0.0305)

All tag records> 180 days 50 71.2 cm 0.1953 9.4246 (3.6) (0.0318)

All tag records > 365 days 20 69.7 cm 0.2237 19.5439 (5.1) (0.0572)

NMFS Miami tag records' 58 70.7 cm 0.2135 7.1282 (3.4) (0.0325)

NMFS Miami tag records > 90 days 38 71.0 cm 0.2098 10.3059 (4.2) (0.0390)

NMFS Miami tag records > 180 days 32 70.9 cm 0.2103 11.8087 (4.5) (0.0424)

NMFS Miami tag records > 3 65 days 13 69.7 cm 0.2343 26.3726 (6.5) (0.0794)

Excuding the tag records used in Schmid and Witzell (1997).

100 NOAA Tech. Mem. NMFS-SEFSC-444

Figure I. Kemp's ridley turtle growth rates plotted by the mean of the initial and recapture carapace lengths for a) all records, b) Gulf of Mexico records, and c) Atlantic Ocean records.

a) All tag records 40 - . ...... ..

y = -0.1 338x + 12.077 "302 F= 0.0447 E030

.20 10 o &44 0 10 20 30 40 50 60 70 80 Mean SSCL (cm) b) Gulf of Mexico records 40 .J y.---0.1 996x + 15.5721 20 1~ 4 n10 .

0 0

0 10 20 30 40 50 60 70 80 Mean SSCL (cm) c) Atlantic Ocean records 40 1 V.y = -0.0785x + 8+63651 2

R = 0.0195 5 30

'20 10 4 4 0

• i  : i 0 10 20 30 40 50 60 70 80 Mean SSCL (cm)

NOAA Tech. Mem. NMFS-SEFSC-444 101

Figure 2. Growth curve for the combined database of all Kemp's ridley turtle recaptures. Dashed portion of the curve represents extrapolation outside the size range of the database. Numbers on the cuive indicate the estimated age at 5 cm intervals.

80

')n A5 Coastal-benthic 70 I

- -- -- -- - Adult 12.1 F

60 50 7

40 Coastal-benthic Subadult CL 0 30

/4ý 1 8 j 1 1) 20 Epipelagic 10 Juvenile 4 00 0 I I I I I I I 0 2 4 8 10 12 14 16 18 20 22 24 26 28 30 Age (yrs) 4ý-

Appendix 2. Sea Turtle Strandings Figure A2.1. Statistical zones along the U.S. Atlantic and Gulf of Mexico coasts.

36 00*.

46oo' ]j '

NC \o 35 ME 34 NH 41 42°O0' 34 48060 VA OCEANV 27 27 32 38 30LAATLANT 294C A MD 39 NJ LAG 28 38200' LAL GULF OF MEXICO 26300 99*00' 773 0000 0' 91 *00' 87200' 83500' 9,00' NOAA Tech. Mem. NMFS-SEFSC-444 103

Figure A2.2a. Kemp's ridley strandings in the U.S. Gulf of Mexico 104 NOAA Tech. Mem. NMFS-SEFSC-444

z0 Kernps Ridley Strandings, by Year and Zone

~YEAR RECION-Gulf of Mexico C9D C-c

-n2 91w

// , .

99 as~d 0 i1 7Q0 0 :

/7 IT/z -7= Zf I .-

I 3 4$ 5 6 7 8 9 10 11 12 13 14. 1 16 17 18 19 20 21 UNMEASURED - SM.IMMATURE ZONE LG.IMMATURE - ADULT C

LP

Figure A2.2b. Kemp's ridley strandings in the Southeast U.S Atlantic.

106 NOAA Tech. Mem. NMFS-SEFSC-444

0 z

H C)

C)

C) Kemps Ridley Strandings, by Year and Zone REGION-.Southeast U.S. Atlontic z

YEAR 97 Cr)

(/2

-n

(/2 n

A A

A 24 25 28 27 28 29 30 31 32 33 34 35 ZONE

= UNMEASURED -Z LG.IMMATURE

- SM.IMMATURE - ADULT

Figure A2.2c. Kemp's ridley strandings in the Northeast U.S Atlantic.

108 NOAA Tech. Mem. NMFS-SEFSC-444

0z Kemps Ridley Strandings, by Year and Zone

~YEAR REGION-Northeast U.S. Atlantic 97*Z

/ '

C// /

94 f /l 38 .7 38 39, 40 4 4 71 ZONE

= UNMEASURED - SM.IMIMATURE -, LGJMIMATUJRE - ADULT C

Figure A2.3a. Loggerhead strandings in the U.S Gulf of Mexico.

llO NOAA Tech. Mem. NMFS-SEFSC-444

z 0CD C)

Loggerhead Strandings, by Year and Zone YEAR REGIOGIfGulf of Mexico Y, **1-//

7 - / /÷/

96 1

-~~ 941Z C ~/C C/c CC C n n/n7 I 2 3 4 5 6 7 a 2 10 11 12 13 14 15 16 17 1. 19 20 21 ZONE

=-- UNMEASURED mm SM.IMMATURE =LG.IMMATURE m ADULT

Figure A2.3b. Loggerhead strandings in the Southeast U.S Atlantic.

112 NOAA Tech. Mem. NMFS-SEFSC-444

z0

,H

. Loggerhead Strandings, by Year and Zone Z * ~~YEAR REGION-Soufheast U.S. Atlantic

-n

" 96 0'1

~9 93 92

,7 "

B6 i caImIa/aI 7 24 25 28 27 28 29 .30 31 32 33 34 35 ZONE ED UNMEASURED = LG.IMMATURE M SM.IMMATURE - ADULT

Figure A2.3c. Loggerhead strandings in the Northeast U.S Atlantic.

114 NOAA Tech. Mem. NMFS-SEFSC-444

z 0

Loggerhead Strandings, by Year and Zone CD REGION =dlJorfheast U.S. Atlantic YEARs/

97 /_/ a*, / , '* /* -

cn ri) 0T 36 37 38 39 40 41l 42 ZONE" UNMEASURED I LG.IMMATURE

- SM.IMMATURE MADULT