VARIATION IN TISSUE STABLE ISOTOPES, BODY SIZE, AND REPRODUCTION OF WESTERN PACIFIC LEATHERBACK TURTLES. A Thesis. Presented to

Transcription

1 VARIATION IN TISSUE STABLE ISOTOPES, BODY SIZE, AND REPRODUCTION OF WESTERN PACIFIC LEATHERBACK TURTLES A Thesis Presented to The Faculty of Moss Landing Marine Laboratories San José State University In Partial Fulfillment of the Requirements for the Degree Master of Science by Deasy N. Lontoh May 2014

2 2014 Deasy N. Lontoh ALL RIGHTS RESERVED

3 The Designated Thesis Committee Approves the Thesis Titled VARIATION IN TISSUE STABLE ISOTOPES, BODY SIZE, AND REPRODUCTION OF WESTERN PACIFIC LEATHERBACK TURTLES by Deasy N. Lontoh APPROVED FOR MOSS LANDING MARINE LABORATORIES SAN JOSÉ STATE UNIVERSITY May 2014 Dr. James T. Harvey Dr. Michael H. Graham Dr. Jeffrey A. Seminoff Moss Landing Marine Laboratories Moss Landing Marine Laboratories NOAA Southwest Fisheries Science Center

4 ABSTRACT VARIATION IN TISSUE STABLE ISOTOPES, BODY SIZE, AND REPRODUCTION OF WESTERN PACIFIC LEATHERBACK TURTLES by Deasy N. Lontoh Western Pacific leatherback turtles migrate to geographically distinct foraging regions. Turtles that nest between April and September in Bird s Head nesting beaches in Papua Barat, Indonesia migrate to the South China Sea (SCS), North Pacific Transition Zone (NPTZ), and Northeast Pacific (NEP). The primary objective of this study was to compare body size and reproduction of turtles that foraged in these distinct regions, as inferred by their skin stable isotope values. Skin samples and reproductive output data were collected during the 2010 and 2011 nesting seasons. Foraging group composition of the nesting population varied between the two seasons. Leatherbacks that foraged in the NEP had greater body size and more years between consecutive nesting seasons than did others that foraged in the SCS and NPTZ. Leatherbacks that foraged in the NPTZ laid more clutches and had fewer years between consecutive nesting seasons than did others that foraged in the SCS and NEP. Within-population variation in body size, reproductive output, and breeding periodicity likely reflect variation in energetic gains and returns associated with each foraging region. To examine relationships between nitrogen and carbon of skin and fasting duration, two skin samples were collected from 53 turtles during separate nesting emergences in Skin stable nitrogen values increased with sampling interval but stable carbon did not, indicating that stable nitrogen values can be used to detect nutritional stress in fasting leatherback turtles.

5 ACKNOWLEDGEMENTS This research project would not have been possible without funding and support from the Bird s Head Leatherback program of the State University of Papua, U. S. NOAA Southwest Fisheries Science Center, Moss Landing Marine Laboratories, Dr. Earl H. Myers and Ethel M. Myers Oceanographic and Marine Biology Trust, PADI foundation, Friend s of Moss Landing Marine Laboratories Signe Memorial Scholarship, and Packard Research and Travel Grant. I sincerely thank members of my committee: Jim Harvey, Mike Graham, and Jeff Seminoff. Jim, thank you for leaving encouraging phone messages when I was in the field, for always giving good advice, and for patiently guiding me through this process. Thank you Mike for always being available, for teaching me about discriminant analysis, and for drawing my attention to bubble plots. Jeff, you have generously shared your isotope lab and expertise with me. Thank you for your guidance and encouragement over the years. In addition, I would like to thank Scott Benson, who has been an amazing mentor. Thank you for teaching me valuable lessons in the field and at Moss Landing. I always value your insights. Ricardo Tapilatu and the Bird s Head Leatherback crew made it possible for me to collect data and samples in the field. Ricky, thank you very much for helping me with many aspects of my research and for teaching me how to get various permits for my samples. I thank each one of the Bird s Head Leatherback crew members: William Iwanggin, Barakhiel Heri Nugroho, Hengki Wona, Emma Sabarofek, Erick Sembor, v

6 Natalia Mansoben, Alosius Numberi, Oktofianus Raweyai, Riki Mayor, Bertha Matatar, and Rehmus Bonepay. Friends, thank you for your hard work, for teaching me how to live and work on a remote nesting beach, and most importantly, for treating me like a family. Thank you Betuel Samber of Papua Natural Resource Agency, Indonesia Institute of Science (LIPI), and Indonesia Ministry of Forestry for helping me obtain appropriate permits. Many people from the Southwest Fisheries Science Center helped me with permitting, processing samples, and providing analytical advice. I thank Erin LaCasella for working with me on getting CITES import permit for my samples. Garrett Lemons and Brad MacDonald, thank you for teaching me how to process my samples. I thank Tomo Eguchi for introducing me to linear mixed model and R, and for having the wisdom to steer me away from Bayesian inference because the timing was not right. I thank Peter Dutton for his sage advice and encouragement. Thank you Manjula Tiwari for your wisdom in the field. I thank Moss Landing Marine Labs community for enriching my graduate experience. My fellow travelers in the Vertebrate Ecology Lab, thank you for supporting me in this journey. I am truly grateful for our friendship. I thank my parents, John and Lina for always believing in me. Thank you my mother, Ning, for being my inspiration. Lastly, my husband. Kevin, you have endured months of separation and you always called back when we lost satellites. Thank you for being a patient and constant companion in this journey. vi

7 TABLE OF CONTENTS LIST OF TABLES... ix LIST OF FIGURES... x INTRODUCTION... 1 Chapter I Within-population variation in tissue stable isotopes, body size, and reproductive output is linked to foraging destination of western Pacific leatherback turtles...3 ABSTRACT... 4 INTRODUCTION... 5 MATERIALS AND METHODS Study site Skin stable isotopes of satellite-tracked leatherbacks Skin sample collection, preparation, and analysis Stable isotope data analyses Body size and reproductive output data collection Body size and reproductive output data analyses RESULTS Stable isotopes Body size and reproductive output DISCUSSION Stable isotopes Body size and reproductive output LITERATURE CITED Appendix 1. Skin δ 34 S of satellite-tracked Dermochelys coriacea sampled in the Bird s Head peninsula, Papua Barat, Indonesia and their foraging destinations Appendix 2. Descriptive statistics of reproductive output variables of Dermochelys coriacea nesting in Papua Barat, Indonesia between April and September of 2010 and Appendix 3. Relationship between egg mass and diameter, and their correlation coefficient Appendix 4. Relationships between hatchling mass, straight carapace length (SCL), and straight carapace width (SCW), and their correlation coefficients vii

8 Chapter II Detecting nutritional stress in leatherback turtles using stable isotope values of skin...72 ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED Appendix 1. Comparison among ordinary and generalized least squares models used in investigating relationship between changes in δ 15 N and fasting duration by plot of residuals versus fitted values and AICc value Appendix 2. Comparison among ordinary and generalized least squares models used in investigating relationship between changes in δ 13 C and fasting duration by plot of residuals versus fitted values and AICc value CONCLUSIONS LITERATURE CITED viii

9 LIST OF TABLES Chapter I Table 1. Summary of sampling and skin stable isotope values (δ 15 N and δ 13 C) of satellite-tracked Dermochelys coriacea Table 2. Estimated clutch frequency models...29 Table 3. Coefficients from the optimal linear model predicting Dermochelys coriacea clutch frequency Table 4. Coefficients from the optimal linear mixed-effect model predicting egg mass of Dermochelys coriacea...32 Table 5. Coefficients from the optimal linear mixed-effect model predicting hatchling mass of Dermochelys coriacea...34 Table 6. Coefficients from the optimal linear mixed-effect model predicting emergence success of Dermochelys coriacea clutches Table 7. Coefficients from the optimal linear mixed-effect model predicting hatching success of Dermochelys coriacea clutches Table 8. Coefficients from the optimal linear mixed-effect model predicting remigration interval of Dermochelys coriacea...39 Table 9. Variation in remigration interval of Dermochelys coriacea detected nesting in three seasons in the Bird s Head peninsula, Papua Barat, Indonesia ix

10 LIST OF FIGURES Chapter I Figure 1. Foraging regions of Dermochelys coriacea that nest during April to September in Jamursba Medi and Wermon beaches on the northwest coast of Papua, Indonesia... 8 Figure 2. The northwest coast of Papua Barat, Indonesia, showing Dermochelys coriacea nesting beaches where skin samples and reproductive output data were collected Figure 3. Skin δ 15 N and δ 13 C of satellite-tracked Dermochelys coriacea and of turtles sampled during 2010 and 2011 nesting seasons Figure 4. Foraging group composition of the nesting population during 2010 and 2011 nesting seasons based on skin δ 15 N and δ 13 C of Dermochelys coriacea...27 Figure 5. Curved carapace length versus width of Dermochelys coriacea classified to distinct foraging regions Figure 6. Curved carapace width versus estimated clutch frequency of Dermochelys coriacea classified to distinct regions Figure 7. Clutch size and number of yolkless eggs obtained from multiple clutches of individual Dermochelys coriacea...31 Figure 8. Curved carapace width versus egg mass of Dermochelys coriacea sampled during 2010 and 2011 nesting seasons Figure 9. Boxplots of Dermochelys coriacea egg mass by year and foraging region Figure 10. Egg mass versus hatchling mass of Dermochelys coriacea clutches incubated at different locations in Jamursba Medi in during 2010 and 2011 nesting seasons Figure 11. Incubation duration versus hatchling mass of Dermochelys coriacea clutches incubated at different locations in Jamursba Medi during 2010 and 2011 nesting seasons Figure 12. Incubation duration versus emergence success of Dermochelys coriacea clutches incubated in Jamursba Medi hatcheries during 2010 nesting season x

11 Figure 13. Incubation duration versus hatching success of Dermochelys coriacea clutches incubated in hatcheries in Jamursba Medi during 2010 nesting season Figure 14. Frequency distribution of remigration intervals of Dermochelys coriacea classified to distinct foraging regions Figure 15. Environmental conditions before 2010 and 2011 nesting seasons as indicated by monthly Multivariate ENSO Index (MEI), Northern Oscillation Index (NOI), and Southern Oscillation Index (SOI) Chapter II Figure 1. The northwest coast of Papua, Indonesia, showing Dermochelys coriacea nesting beaches where skin samples were collected Figure 2. Relationship between sampling interval and δ 15 N and δ 13 C of Dermochelys coriacea skin xi

12 INTRODUCTION The Bird s Head peninsula in Papua Barat, Indonesia hosts the largest remaining nesting aggregation of leatherback turtles (Dermochelys coriacea) in the Pacific. Jamursba Medi and Wermon beaches in Bird s Head are the primary nesting beaches, and they contain approximately 75% of total western Pacific leatherback nesting activity annually (Dutton et al. 2007). Although considered the most robust among western Pacific populations, the Bird s Head nesting population has been declining. The number of nests in Jamursba Medi has decreased 78% during the past 27 years, and it has decreased 63% in Wermon since monitoring began in 2002 (Tapilatu et al. 2013). Post nesting, Bird s Head leatherbacks migrate to geographically distinct, productive regions throughout the Pacific. Leatherbacks that nest during boreal summer (April - September) migrate to the South China Sea and temperate North Pacific whereas leatherbacks that nest during austral summer (October - March) migrate to Indonesian Seas and temperate South Pacific (Benson et al. 2011). These regions vary in distance from the nesting beach, latitude which influences water temperatures, habitat type, and productivity. Energetic costs and returns associated with foraging in these distinct regions influence how rapid leatherbacks accumulate sufficient energy for reproduction and how much they allocate for reproduction. Advancement in tracking technology has greatly increased our ability to link breeding and non-breeding areas of seasonal migrants (Webster et al. 2002, Godley et al. 2010). This technology allows investigations on how non-breeding habitat quality 1

13 influences abundance, survival, and productivity of migratory vertebrates (Harrison et al. 2011, Zbinden et al. 2011, Vander Zanden et al. 2014). Stable isotopes are increasingly used to study migratory connectivity because they offer a less invasive, more costeffective method than satellite telemetry. They also have been used to detect nutritional stress in fasting animals, but with mixed results (Hatch 2012). Both applications of stable isotopes were investigated in this study. The overall objective of Chapter I was to compare body size and reproductive output of western Pacific leatherbacks that foraged in distinct regions in the Pacific. Specific objectives were: 1) to determine if inferring foraging regions of nesting turtles using skin stable isotope values was possible, 2) to determine foraging group composition of the nesting population, and 3) to compare body size, within-year reproductive output, and breeding periodicity of the foraging groups. The objective of Chapter II was to evaluate the use of skin stable isotopes to detect nutritional stress in fasting leatherbacks. 2

14 Chapter I Within-population variation in tissue stable isotopes, body size, and reproductive output is linked to foraging destination of western Pacific leatherback turtles 3

15 ABSTRACT Reproduction of migratory vertebrates is linked to the quality of non-breeding habitat. Western Pacific leatherback turtles (Dermochelys coriacea) that nest in Papua Barat, Indonesia during boreal summer forage in geographically and oceanographically distinct regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). To examine reproductive consequences of foraging in these distinct regions, skin samples, body size and reproductive output data were collected during 2010 and 2011 nesting seasons. Foraging region of nesting turtles was inferred based on their skin stable isotopes. The NEP and NPTZ foraging groups were most abundant in 2010 whereas the SCS group was dominant in Body size and reproductive output varied greatly among individuals. Leatherbacks that foraged in the NEP were larger and nested less frequently than others that foraged in the NPTZ and SCS. Leatherbacks that foraged in the NPTZ laid more clutches and nested more frequently than others that foraged in the NEP and SCS. Egg mass was greater in 2010 than in 2011, which might be linked to ocean conditions before the nesting season. Incubation environment influenced hatchling phenotypes, emergence success, and hatching success instead of maternal foraging region. Energetic demands and returns associated with foraging in the NEP, NPTZ, and SCS regions likely shaped the observed within-population variation in body size, reproductive output, and breeding periodicity of western Pacific leatherback turtles. 4

16 INTRODUCTION Populations of migratory vertebrates are regulated by events occurring during breeding and non-breeding seasons. Factors that affect per capita survival and reproduction have the greatest impact on population dynamics, of which quality of habitats is one (Newton 2004). Habitat quality refers to the ability of the environment to provide conditions appropriate for individual and population persistence (Hall et al. 1997, p. 178). This term broadly encompasses physical and biological forces that influence survival and reproduction of individuals (thereby abundance) such as availability of resources (e.g., prey items, mates, nesting sites), species interactions (e.g., predation, competition, parasitism), environmental variability, and anthropogenic stressors (e.g., pollution, bycatch) in a habitat. Demographically, habitat quality can be defined as the expected per capita population increase from a given habitat, and best measured by abundance, survival, and reproduction (Johnson 2007, p. 491). For long-lived vertebrates, abundance and reproduction are easier to measure than survival. Breeding abundance and reproduction have been linked to habitat quality during the non-breeding season, particularly for vertebrates relying on stored energetic capital for reproduction ( capital breeders in Drent & Daan 1980). During nonbreeding seasons, capital breeders forage to accumulate sufficient energy reserves for reproduction (Naulleau & Bonnet 1996, Madsen & Shine 1999). Energetic incomes (e.g., prey abundance and caloric content) and expenditures (e.g., thermoregulation, 5

17 food search and handling) associated with the non-breeding habitat determine how fast they achieve breeding body condition and how much they allocate for reproduction. Numerous studies indicated that greater non-breeding habitat quality led to better body conditions and survival, earlier arrival to breeding areas, greater reproductive output, better reproductive success, and greater breeding abundance (Weimerskirch et al. 2003, Norris & Marra 2007, Harrison et al. 2011). More evidence of links between non-breeding habitat quality and abundance, productivity, or survival of marine turtles are emerging as long-term data of nesting populations become available (Solow et al. 2002, Saba et al. 2007, van Houtan & Halley 2011) and tracking methods advance (Webster et al. 2002, Godley et al. 2010). Marine turtles migrate over long distances between nesting and foraging areas (Hughes et al. 1998, James et al. 2005, Eckert et al. 2006, Benson et al. 2007b, 2007c, 2011) and rely primarily on stored energetic capital for reproduction (Plot et al. 2013). Energy gained through foraging in non-breeding habitat is stored as subcutaneous and visceral fat, which can be mobilized to fuel metabolic processes such as vitellogenesis (Kwan 1994). Vitellogenesis is a process through which protein and lipid are mobilized to develop oocytes or yolk precursors (Guraya 1989), and marine turtles complete this process before migrating to the nesting beach (Hamman et al. 2002). Good body condition and sufficient fat stores may be involved in inducing vitellogenesis, but further research on factors influencing this process is needed. At the nesting beach, turtles lay multiple clutches of eggs during 6

18 the nesting season and return to nest typically every two or three years. Reproductive output and breeding frequency vary among species (van Buskirk & Crowder 1994). Within species, variation in body size and reproductive output is associated with variation in nesting location, feeding habit, ocean productivity, and foraging region (e.g., Tiwari & Bjorndal 2000, Saba et al. 2008, Hatase & Tsukamoto 2008, Wallace & Saba 2009, Zbinden et al. 2011, Vander Zanden et al. 2014). All indicate an influence of habitat quality. The Bird s Head peninsula in Papua Barat, Indonesia hosts the largest nesting aggregation of western Pacific leatherback turtles (Dermochelys coriacea). Jamursba Medi and Wermon beaches host approximately 75% of total western Pacific leatherback nesting activity annually (Dutton et al. 2007). Although considered the most robust, the Bird s Head nesting population has been declining 5-12% per year (Tapilatu et al. 2013). The Bird s Head leatherbacks migrate to geographically and oceanographically distinct regions throughout the Pacific (Benson et al. 2011). Leatherbacks that nest during boreal summer (April to September) migrate to the South China Sea and temperate North Pacific, whereas others that nest during boreal winter (October-March) move to Indonesian seas and temperate South Pacific. They migrate to these distant regions to feed on seasonally abundant gelatinous preys (e.g., Scyphomedusae; Eisenberg and Frazier 1983, Shenker 1984). I studied leatherbacks that nested during boreal summer and foraged in the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS; Fig. 1). 7

19 Figure 1. Foraging regions of Dermochelys coriacea that nest during April to September in Jamursba Medi and Wermon beaches on the northwest coast of Papua, Indonesia: the South China Sea (SCS), North Pacific Transition Zone (NPTZ), and Northeast Pacific (NEP). Map created based on satellite tracks in Benson et al. (2011). The NEP, NPTZ, and SCS regions vary in distance from Bird s Head nesting beaches, latitude, water temperature, habitat type, productivity, variability, and presumably prey availability and quality. The NEP is the furthest foraging region (~10,000 km) located along an eastern boundary current at temperate latitudes. Seasonal variability in primary productivity is influenced by wind-driven upwelling (Huyer 1983, Strub et al. 1987, Lynn et al. 2003), whereas interannual variability is affected by El Niño Southern Oscillation (ENSO, McGowan et al. 1998). Annual 8

20 primary production ranges from 0.10 to 1.60 g C m -2 d -1 with a maximum during spring (Longhurst 2007). The NPTZ is located mid-distance from Bird s Head nesting beaches (~6,500 km) between 32 and 42 N latitudes. It is bounded by the subarctic and subtropical fronts (Roden 1970). The NPTZ has a sharp transition in surface chlorophyll-a concentration known as the Transition Zone Chlorophyll Front (TZCF), operationally defined as 0.2 mg m -3 surface chlorophyll-a isopleth (Polovina et al. 2001). The NPTZ is a persistent, basin-wide feature that moves north and south seasonally and interannually (Bograd et al. 2004). Annual primary production ranges from 0.15 to 0.75 g C m -2 d -1 with a maximum production during spring and a lesser maximum between September and November (Longhurst 2007). El Niño events were associated with increased primary production (Karl et al. 1995) and a more meandering TZCF, which created convergence areas thereby enhancing foraging opportunity of marine predators (Polovina et al. 2001). The NPTZ provides important foraging habitats for marine vertebrates such as Thunnus alalunga, Caretta caretta, Phoebastria immutabilis, Phoebastria nigripes, Callorhinus ursinus, Mirounga angustirostris, and Lamna ditropis (Laurs & Lynn 1977, Polovina et al. 2000, 2001, Kappes et al. 2010, Ream et al. 2005, Weng et al. 2005, Simmons et al. 2007). The SCS is closest (~2,500 km) to Bird s Head nesting beaches located in tropical latitudes. Monsoon winds drive seasonal changes in SCS circulation pattern 9

21 (Wyrtki 1961, Qu 2000) and primary production of SCS (Liu et al. 2002, Lee Chen 2005). Annual primary production ranges from 0.25 to 0.80 g C m -2 d -1, with a maximum during winter (Longhurst 2007). Interannual variability in the SCS is associated with ENSO (Chao 1996, Rong et al. 2007, Liu et al. 2004, Wang et al. 2006), and interannual variation in primary production is less than seasonal variation (Liu & Chai 2009). These differences in foraging habitat quality influence energy budget and allocation of western Pacific leatherback turtles. In this study, I examined reproductive consequences of foraging in these distinct regions. Foraging region of nesting leatherbacks in this study was inferred based on stable nitrogen and carbon isotope values of their skin. Relative to satellite telemetry, which has been primarily used to track movement of migratory vertebrates (Godley et al. 2008, Block et al. 2011), stable isotope provides a cost effective, less invasive method of tracking (Hobson 2008, Ceriani et al. 2012, Seminoff et al. 2012). The spatial specificity of stable isotope tracking is lesser than satellite tracking (Webster et al. 2002) but sufficient for the purpose of this study. Stable isotopes are naturally occurring forms of the same element with different numbers of neutrons in the nucleus. Stable isotope values are predictably integrated into animal tissues. In tissue with slow turnover rate, such as leatherback skin, isotopic values are conserved for four to six months (Seminoff et al. 2007, Reich et al. 2008). Determining breeding or non-breeding locations of seasonal migrants based on tissue stable isotopes is possible because there are spatial gradients in stable isotope 10

22 values or isoscapes. Seminoff et al. (2012) found that satellite-tracked leatherbacks that foraged in the eastern Pacific had greater skin δ 15 N than others that foraged in the western Pacific. Using compound-specific stable isotope analysis of amino acids (CSIA-AA), they controlled for baseline influences on δ 15 N values and showed the dichotomy in skin δ 15 N values was not a result of turtles foraging at different trophic levels but rather due to different baseline δ 15 N values of eastern and western Pacific. Within an ocean basin, δ 13 C values track productivity, with greater values associated with rapid algae growth (Goericke & Fry 1994, Popp et al. 1998). Productive nearshore areas generally have greater δ 13 C values than less productive open ocean habitats (Rau et al. 1982, Rubenstein & Hobson 2004). I hypothesized based on these isotopic gradients that greater skin δ 15 N would distinguish leatherbacks that foraged in the NEP from others that foraged in the SCS and NPTZ regions and lesser δ 13 C would further distinguish leatherbacks that foraged in the NPTZ from others that foraged in the SCS (hereafter NEP, NPTZ, and SCS foraging groups). I hypothesized that foraging group composition of the nesting population would vary between years. Composition of the nesting population would be driven by timing of breeding of foraging groups. Timing of breeding would differ among groups because energetic demands and returns were unique among regions. I examined the effect of foraging region on body size, within-year reproductive output, and breeding periodicity. I hypothesized the NEP and NPTZ foraging groups would have greater body size than the SCS groups because greater body size confers 11

23 survival benefits in cooler and more distant temperate regions. Benson et al. (2011) found satellite-tracked western Pacific leatherbacks that migrated to temperate regions had greater body size than others that migrated to tropical regions. I hypothesized that the larger NEP and NPTZ foraging groups would have greater reproductive output by having more eggs (e.g., greater clutch size) or heavier eggs. In addition to providing thermoregulation advantage, larger body size allows greater energy storage for reproduction. A positive correlation between body size and clutch size had been widely reported for Caretta caretta and Chelonia mydas (Hirth 1980, Frazer & Richardson 1986, Hays & Speakman 1991) and chelonians (Congdon & Gibbons 1985, Elgar & Heaphy 1989). Bjorndal & Carr (1989) and Wallace et al. (2007), however, reported that body size explained little variation in clutch size of C. mydas and no variation in clutch size of eastern Pacific D. coriacea. I predicted that incubation environment and egg mass would have a greater influence on hatchling phenotypes, emergence success, and hatching success than maternal foraging region. Numerous studies indicated that temperature, moisture content, sand particle size, and gas exchange influence embryo development and clutch survival (McGehee 1990, Booth & Astill 2001, Reece et al. 2002, Booth 2006, Wallace et al. 2006b, Tapilatu & Tiwari 2007). I used incubation duration, which has a negative relationship with sand temperature (Chan & Liew 1995, Houghton & Hays 2001, Matsuzawa et al. 2002), as a proxy for incubation temperature. Remigration interval is the number of years between consecutive nesting seasons ( breeding 12

24 periodicity ). It reflects the amount of time to complete migration, which makes up a great percentage of total reproductive cost (Wallace et al. 2006b) and the amount of time to accumulate sufficient energy reserve for reproduction. Because they migrate over a greater distance, I hypothesized the NEP group would have greater remigration intervals. Additionally, I predicted these turtles would have a more variable remigration intervals because the NEP is only available seasonally and conditions are more variable there than the NPTZ and SCS. Understanding the influence of foraging habitat quality on survival and reproduction, both of which determine population growth, will help management of different components of this declining population. The ability to distinguish foraging groups will help evaluate how the different groups contribute to population growth and how spatially explicit threats might impact the overall population. MATERIALS AND METHODS Study site Body size and reproductive output data and skin samples were collected from 207 leatherback turtles nesting at the Bird s Head peninsula on the northwest coast of Papua Barat, Indonesia ( S, E; Fig. 2). Jamursba Medi and Wermon beaches are the primary nesting sites (Hitipeuw et al. 2007, Tapilatu et al. 2013). Jamursba Medi is composed of Wembrak, Batu Rumah, and Warmamedi beaches. They form a continuous beach spanning 18 km of coastline. 13

25 Wermon beach is located approximately 30 km east of Jamursba Medi and spans six km. Samples and data were collected opportunistically from early June to mid August 2010 and from late June to end of September Figure 2. The northwest coast of Papua Barat, Indonesia, showing Dermochelys coriacea nesting beaches, Jamursba Medi and Wermon, where skin samples and reproductive output data were collected (Tapilatu et al. 2013). Skin stable isotopes of satellite-tracked leatherbacks I obtained skin δ 15 N and δ 13 C values of 31 leatherbacks that were satellitetracked between 2003 and 2011 (Table 1) and δ 34 S values from 13 of them (Appendix 1). In 2011, I also sampled skin from a turtle that was tagged in California in 2004 (Table 1). I assumed this turtle exhibited strong fidelity to the NEP region as documented by Benson et al. (2011) and considered its δ 15 N and δ 13 C values associated with the region in my analyses. 14

26 Table 1. Summary of sampling and skin stable isotope values (δ 15 N and δ 13 C) of satellite-tracked Dermochelys coriacea. No. Sampling date δ turtles 15 N range ( ) δ 13 C range ( ) References July to to , July to to , July to to , July to to , 3 9 June August July August Benson et al. (2007b) 2 Benson et al. (2011) 3 Seminoff et al. (2012) to to , 2, 3 Skin sample collection, preparation, and analysis Skin of tracked and untracked females were sampled, prepared, and analyzed similarly. Skin was sampled from the dorsal axial region of a hind flipper during egglaying trance or nest covering. Sampling surface was scrubbed with alcohol to remove algae, and a small sample of the epidermis (<10 x 10 x 1 mm, g wet mass) was obtained using a single edge razor. Betadine was applied to the sampling surface following sampling. Samples were preserved in 70% ethanol or in a saturated salt solution in 2-ml cryogenic vials; neither preservatives affected isotope values (Barrow et al. 2008). Skin samples were rinsed with deionized water, finely diced with a scalpel blade, then freeze-dried at -50 C for 12 hours in a lyophilizer (BenchTop K, VirTis, 15

27 SP Industries, Gardiner, NY, USA). Lipids were removed from samples using an accelerated solvent extractor (Model 200, Dionex, Bannockburn, IL, USA) with petroleum ether as the primary solvent. Sub-samples of prepared tissue ( mg) were weighed with a microbalance and packed in tin capsules for mass spectrometric analysis. Stable carbon and nitrogen ratios were determined by combusting the subsamples of prepared tissue in a Costech ECS 4010 interfaced via a ConFlo III device (Finnigan MAT, Bremen, Germany) to a Deltaplus isotope-ratio mass spectrometer (Finnigan MAT, Bremen, Germany) in the Light Stable Isotope Mass Spectrometry Lab in the Department of Geological Sciences at the University of Florida, Gainesville. Sample stable isotope ratios were compared with isotope standards and expressed following conventional delta (δ) notation in permil ( ): δ x E = R sample R standard where x is the atomic mass of the heavier isotope of element E, and R is the ratio of heavy to light isotopes ( 15 N/ 14 N, 13 C/ 12 C, and 34 S/ 32 S). The international standard for 15 N was atmospheric nitrogen and for 13 C was Vienna Pee Dee Belemnite. All analytical runs included reference materials, USGS 40 L-glutamic acid (δ 15 N = -4.52, δ 13 C = ) and USGS 41 L-glutamic acid (δ 15 N = 47.57, δ 13 C = ), interspersed regularly among samples to calibrate the system and to compensate for any drift through time. Reference materials were calibrated monthly 16

28 against international standards. The analytical error was 0.10 for δ 15 N and 0.12 for δ 13 C. To determine stable isotope ratios of sulfur (δ 34 S), tissue sub-samples were combusted with an elemental analyzer (ECS 4010, Costech Analytical, Valencia, CA). SO2 gases were separated with a 0.8m gas chromatograph column (100 C) and analyzed with a continuous flow isotope ratio mass spectrometer (Delta PlusXP, Thermofinnigan, Bremen; Brenna et al. 1997) in the Stable Isotope Core Laboratory at the Washington State University. Final determination of δ 34 S was based on collection of ions 64 and 66. The laboratory used a dual reactor configuration (Fry et al. 2002) with the second reactor full of quartz chips to buffer 18 O contribution to the SO2. Approximately 5 mg of niobium pentoxide was amended to each sample to improve combustion. No correction for oxygen isotope contribution was made. Sulfur isotopic ratios are reported relative to VCDT (Vienna Cañon Diablo Troilite) by assigning a value of -0.3 to the reference material silver sulfide (IAEA S-1; Coplen & Krouse 1998). At least three primary isotopic reference materials were interspersed with samples for calibration. The analytical error was 0.05 for δ 34 S. Stable isotope data analyses Skin δ 15 N and δ 13 C values of satellite-tracked leatherbacks (Table 1) were used as predictors of foraging group membership in a direct linear discriminant analysis (LDA). Skin δ 34 S values were not used because a limited sample size was processed; values were included in Appendix 1. Equal prior probability was assigned for each 17

29 group. The resulting discriminant functions were used to classify turtles sampled in 2010 and 2011 into the three foraging groups. Leave-one-out cross validation was performed to assess how well classification coefficients performed, and the proportion of turtles correctly assigned was compared with a null expectation. Proportion of each foraging group in the nesting population was estimated for 2010 and 2011 seasons based on LDA classified-turtles with greater than 70% group membership probability. A Chi-squared test was performed to test the null hypothesis of no difference in foraging group composition between 2010 and Body size and reproductive output data collection For each turtle, Passive Integrated Transponder (PIT) tag number, date, sector number, and nesting status (nesting or non-nesting) were recorded. To identify individuals, a PIT tag was inserted into the shoulder muscle of each turtle following Dutton and McDonald (1994). Curved carapace length (CCL), defined as maximum curved length from the tip of the first bony ridge alongside the midline to the distal carapace tip, and curved carapace width (CCW), defined as curved distance at the widest part of the carapace from side ridge to side ridge were measured using flexible measuring tape (± 0.1 cm). Turtles were remeasured upon subsequent encounters during the nesting season. Data collected from clutches included: clutch frequency, clutch size, number of yolkless eggs, egg mass and diameter, hatchling mass and body size, emergence success, and hatching success. Clutch frequency was estimated because long beaches and limited monitoring 18

30 by personnel likely resulted in undetected nesting. Clutch frequency was estimated similar to Frazer & Richardson (1986), using emergence data collected throughout the nesting seasons. It was estimated for each individual using the following equation: Internesting interval mode was used instead of the mean because of the high likelihood of missing an emergence. The internesting interval mode was 9 days. I assumed that a leatherback detected at Jamursba Medi and Wermon for the first time was laying her first clutch and that a nesting emergence occurred close in time to a non-nesting emergence. Because clutch frequency of one was unlikely, clutch frequency was estimated only for individuals seen more than once (ovipositing or not) during the nesting season. Date of last emergence - Date of first emergence Internesting interval mode +1 Eggs were counted, weighed, and measured before the clutch was relocated into a hatchery or before a female covered her nest. In 2010, clutches were relocated into Wembrak west, Wembrak east, and Warmamedi hatcheries as part of an investigation on factors influencing hatching success (Tapilatu et al. unpublished). The shaded and fenced-in hatcheries protected clutches from predators and extreme sand temperatures (Tapilatu & Tiwari 2007). Nearly all sampled clutches were relocated in the hatcheries in For these clutches, eggs were counted, weighed, and measured before burying them in the hatcheries. In 2011, sampled clutches were either left in situ or relocated to areas above high water mark when they faced risk of tidal 19

31 inundation. Eggs were counted as they dropped from the cloaca, weighed, measured, and then placed back into the nest before the turtle covered her nest. I counted the total number of yolked and yolkless eggs, weighed randomly selected yolked eggs and 10 yolkless eggs using a digital pocket scale (± 0.1 g) and measured their diameter using a digital caliper (± 0.01 mm). Hatchlings were measured after they emerged from nests. In the hatcheries, five to ten days before expected hatch dates, a wire-mesh enclosure was placed above each nest to capture emerging hatchlings. In 2011, each clutch was protected against predation by a flexible circular mesh (~ 5 cm mesh size) approximately 1.5 meter in diameter. The mesh was secured with wooden stakes above the nest. To prevent hatchlings from getting entangled, I monitored the nests closely five days before expected hatching dates and removed the mesh when hatchlings were discovered at depth less than 40 cm from the surface. Straight carapace length and width (SCL and SCW) of randomly selected hatchlings were determined using a digital caliper (± 0.01 mm). Hatchling mass was determined using a digital pocket scale (± 0.1 gram). Sampled hatchlings were released together with hatchlings from the same clutch immediately after sampling. Emergence and hatching successes were computed for each clutch in the hatcheries in Emergence success was calculated by subtracting the number of large eggshells (>50%) by the number of live and dead hatchlings remaining in the nest after hatchlings had emerged and dividing it by clutch size. Hatching success 20

32 was calculated by dividing the number of large eggshells by clutch size. Number of years between consecutive nesting seasons (remigration interval) was calculated based on recapture data collected between 2003 and 2011 for leatherbacks from which a skin sample was collected. Body size and reproductive output data analyses Only leatherbacks with greater than 70% group membership probability were included in the analyses to balance the need for including only correctly classified individuals in the analyses and having sufficient sample size to capture variability within and among foraging groups. The two satellite-tracked turtles sampled in 2010 and 2011 were considered to have 100% membership probability. The null hypothesis of no difference in mean CCL and CCW among foraging groups was tested using multivariate analysis of variance (MANOVA). Mean CCL and CCW were used for turtles measured multiple times during a nesting season. Year and its interaction with foraging group were included in the analysis but removed when their effects were non-significant. Histograms of CCL and CCW indicated they were normally distributed, and residual plots indicated equal variances. Pillai s trace test statistic was chosen because it is considered most robust when assumption of equal covariances among group is not met. Clutch frequency was modeled using general linear models with foraging group, year, their interaction, and CCL or CCW as predictors. Optimal model was selected based on Akaike Information Criterion (AIC) values (Burnham & Anderson 21

33 2002). Standardized residual plots were examined to determine normality of errors and homogeneity of error variance. Clutch size, egg and hatchling measurements, emergence and hatching successes, and remigration interval were modeled using linear mixed-effect (LME) models to account for correlated observations from an individual and nested structure of the data. I followed top-down model selection approach outlined in Zuur et al. (2009). I started with the most complex model (less than optimal model) where the fixed component contains all predictors and their interactions. I then identified the optimal structure of the random component from the most complex model by including female identity as a random factor and fitting various variance structures. Once the optimal structure of the random component was found, I identified the optimal structure of the fixed component. Throughout the process, likelihood ratio tests and AIC values were used to select the most optimal model. The final model was refitted using restricted maximum likelihood estimation and coefficients were estimated. Repeatability or intraclass correlation coefficient defined as correlation among measurements within group (in this study, within females) were determined using an LME-based method (Nakagawa & Schielzeth 2010). To help evaluate the fit of the final LME model, I reported variance explained by fixed component alone (marginal R 2 ; R 2 LME(m)) and variance explained by fixed and random components together (conditional R 2 ; R 2 LME(c)). These were computed following Nakagawa & Schielzeth (2013). LME modeling was done using package nlme (Pinheiro et al. 22

34 2013) in R version (R Core Team, 2013). In the fixed component of the most complex model predicting clutch size, I included foraging group, year, CCL or CCW, number of days since first emergence, and their two-way interactions. An extreme outlier, a clutch size of 19, was excluded from the analysis to meet the assumption of normally distributed errors of LME analysis. Because egg mass and diameter were positively correlated (Appendix 3), I focused on modeling egg mass only. The fixed component of the most complex model for egg mass included foraging group, year, CCL or CCW, number of days since first emergence, and their two-way interactions. Two extreme outliers, egg mass means of 54.9 and 46.7 grams, were excluded from the analysis to meet assumption of normally distributed errors of LME analysis. Hatchling mass, SCL, and SCW were positively correlated (Appendix 4) so I selected to model hatchling mass only. Foraging group, year, CCL or CCW, egg mass, incubation duration, and clutch location nested within year were included in the fixed component of the most complex model. Wembrak west, Wembrak east, and Warmamedi hatcheries were nested within 2010, and Wembrak, Batu Rumah, and Warmamedi beaches within Foraging group, hatchery, and incubation duration were included in the fixed component of the most complex model for emergence and hatching sucesses (2010 only). Remigration interval was compared among foraging groups using LME modeling with foraging group in the fixed component and female identity in the 23

35 random component. RESULTS Stable isotopes Skin stable isotope values of satellite-tracked leatherbacks clustered by foraging destinations (Fig. 3a). The δ 15 N values were from 9.78 to 17.73, the δ 13 C values from to , and δ 34 S from to (Table 1, Appendix 1). Means of skin δ 15 N and δ 13 C were (95% CI from to16.04 ) and ( to ) for satellite-tracked leatherbacks that foraged in the NEP, (11.42 to ) and ( to ) for leatherbacks that foraged in the NPTZ, and (9.89 to ) and ( to ) for leatherbacks that foraged in the SCS. Isotopic ranges of the SCS foraging group were less than those of the NEP and NPTZ groups. Skin δ 15 N and δ 13 C values were weakly correlated (r = -0.22, p = 0.001), but neither was correlated with δ 34 S values (r = , p = with δ 15 N; r = 0.010, p = with δ 13 C). Skin δ 15 N and δ 13 C values of satellite-tracked leatherbacks were used as predictors of membership in the NEP, NPTZ, and SCS foraging groups in an LDA. Two discriminant functions were calculated with a combined χ 2 (4) = (p < ). After removal of the first function, there was still a strong association between groups and predictors, χ 2 (1) = 9.281, p = The first function separated 24

36 the NEP from the NPTZ and SCS groups, and the second function separated the NPTZ from the SCS group. The two discriminant functions accounted for 73% and 27% of the between-group variability. Leave-one-out cross validation indicated the discriminant functions correctly classified 23 of 31 satellite-tracked leatherbacks (74.2%), which was different from the null expectation of 33.3% (p = ). Skin was sampled from 207 leatherbacks (mostly untracked females) during 2010 and 2011 nesting seasons, and analyzed for δ 15 N and δ 13 C. Only one sample failed to provide reasonable results. Skin δ 15 N values ranged from 8.71 to and δ 13 C values from to Skin δ 15 N and δ 13 C values of leatherbacks sampled in 2010 overlapped largely with NEP and NPTZ isotopic ranges (Fig. 3b) whereas those sampled in 2011 mostly overlapped with NPTZ and SCS ranges (Fig. 3c). The NEP and NPTZ groups were dominant in 2010 with proportions of 0.48 and 0.38, respectively, whereas the SCS group was dominant in 2011 (0.43; Fig. 4). Composition of the nesting population differed between the 2010 and 2011 seasons (χ ,2 = , p = 0.001), especially for the NEP and SCS groups (Bonferroni adjusted p < 0.05). 25

37 Figure 3. Skin δ15n and δ13c of satellite-tracked Dermochelys coriacea (a), and of turtles sampled during 2010 (b) and 2011(c) nesting seasons. Foraging regions: the Northeast Pacific (NEP; blue circles), North Pacific Transition Zone (NPTZ; orange squares), and South China Sea (SCS; red triangles). Leatherbacks sampled in 2010 are depicted as (+), and 2011 as (x). 26

38 Figure 4. Proportion of leatherback turtles nesting in boreal summer of 2010 (black) and 2011 (white) from the three foraging regions based on skin δ 15 N and δ 13 C of Dermochelys coriacea. Foraging regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). Only turtles with greater than 70% group membership probability are included. Body size and reproductive output Body size differed among leatherbacks classified to distinct regions (MANOVA, Pillai s trace, F4,338 = , p = ), with the NEP foraging group having greater CCL and CCW than the others (Fig. 5). Mean (± 1 SE) CCL and CCW were ± 0.9 cm and ± 0.6 cm for the NEP group, ± 1.0 cm and ± 0.6 cm for the NPTZ group, and ± 1.0 cm and ± 0.7 cm for the SCS group. 27

39 Figure 5. Curved carapace length versus width of Dermochelys coriacea classified to distinct foraging regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). Only turtles with greater than 70% group membership probability are included. Estimated clutch frequency varied greatly among individuals. The optimal linear model included CCW (F1,127 = , p = 0.030) and foraging group (F2,127 = , p = 0.019; Table 2) but explained only 9% of variation. Clutch frequency was greater for the NPTZ foraging group than the NEP and SCS groups, and increased with CCW (Table 3, Fig. 6). At the group-specific mean CCW, mean clutch frequency was 5.6 for the NEP group, 6.1 for the NPTZ group, and 4.8 for the SCS group. 28

40 Table 2. Estimated clutch frequency models. The change in AIC (Δ AIC) is relative to the optimal model. The null model includes the intercept term only. FG = foraging group, CCW = curved carapace width, df = degrees of freedom, R 2 = multiple R- squared. Model AIC Δ AIC df R 2 FG + CCW FG + CCW + Year FG CCW Null model Table 3. Coefficients from the optimal linear model predicting Dermochelys coriacea clutch frequency. The model includes curved carapace width (CCW) and foraging regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). Estimates for the NPTZ and SCS foraging groups are differences from the model intercept. Variable Estimate SE t-value p-value Intercept (NEP) NPTZ SCS CCW

41 Figure 6. Curved carapace width versus estimated clutch frequency of Dermochelys coriacea classified to distinct foraging regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). Clutch frequency was estimated for turtles seen more than once during 2010 and 2011 nesting seasons. Only turtles with greater than 70% group membership probability are included. Clutch size varied greatly among individuals. The optimal LME model for clutch size included a constant in the fixed component and female identity in the random component (R 2 LME(m) = 0, R 2 LME(c) = 0.62). I found a female s later clutches did not have fewer eggs than her earlier clutches, and observed no relationship between clutch size and body size or number of yolkless eggs (Fig. 7). The intraclass correlation coefficient was 0.62, which indicated strong repeatability or consistency in clutch sizes of individual leatherbacks. 30

42 Figure 7. Clutch size (greys) and number of yolkless eggs (oranges) obtained from multiple clutches of individual Dermochelys coriacea. Darker colors indicate clutches laid earlier during the nesting season by that individual. The optimal LME model for egg mass included CCW (F1,74 = 4.240, p = ) and year (F1,74 = 6.163, p = ) in the fixed component and female identity in the random component (R 2 LME(m) = 0.10, R 2 LME(c) = 0.68). Egg mass increased with CCW and was greater in 2010 than in 2011 (Table 4, Fig. 8). The SCS foraging group indicated the greatest decrease in egg mass (6.6 grams) followed by the NEP group (3.0 grams), and the NPTZ group (1.0 gram; Fig. 9). The intraclass correlation coefficient was 0.64, which indicated that egg masses of clutches laid by the same individuals were strongly correlated. 31

43 Table 4. Coefficients from the optimal linear mixed-effect model predicting egg mass of Dermochelys coriacea. The model includes year and curved carapace width (CCW) in the fixed component and female identity in the random component. Estimate for 2011 is a difference from the model intercept. Variable Estimate SE df t-value p-value Intercept (2010) CCW Figure 8. Curved carapace width versus egg mass of Dermochelys coriacea sampled during 2010 (black) and 2011 (grey) nesting seasons. Lines drawn based on coefficients from the optimal linear mixed-effect model for egg mass (Table 4). 32

44 Figure 9. Medians, quartiles, and ranges of Dermochelys coriacea egg mass by year and foraging region. Foraging regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). Only turtles with greater than 70% group membership probability are included. The optimal LME model for hatchling mass included egg mass (F1,21 = , p < ) and incubation duration (F1,21 = 5.336, p = ) in the fixed component and female identity and clutch location in the random component (R 2 LME(m) = 0.27, R 2 LME(c) = 0.56). Hatchling mass increased with egg mass and incubation duration (Table 5, Fig. 10, 11). Intraclass correlation coefficients were 0.40 for female identity and 0.36 for clutch location. Given the same egg mass and 33

45 incubation duration, hatchlings from Warmamedi hatchery tended to be heavier than those from Wembrak west (Fig. 10, 11). Table 5. Coefficients from the optimal linear mixed-effect model predicting hatchling mass of Dermochelys coriacea. The model includes egg mass and incubation duration in the fixed component and female identity nested within clutch location in the random component. Variable Estimate SE df t-value p-value Intercept Egg mass Incubation duration Figure 10. Egg mass versus hatchling mass of Dermochelys coriacea clutches incubated at different locations in Jamursba Medi during 2010 and 2011 nesting seasons. Line drawn based on coefficients from the optimal linear mixed-effect model for hatchling mass (Table 5). Incubation duration is held constant at mean value. 34

46 Figure 11. Incubation duration versus hatchling mass of Dermochelys coriacea clutches incubated at different locations in Jamursba Medi during 2010 and 2011 nesting seasons. Line drawn based on coefficients from the optimal linear mixedeffect model for hatchling mass (Table 5). Egg mass is held constant at mean value. The optimal LME model for emergence success included incubation duration (F1,26 = , p = ) in the fixed component, female identity in the random component, and exponential error variance structure (R 2 LME(m) = 0.17, R 2 LME(c) = 0.39). For clutches that hatched, emergence success increased with incubation duration (Table 6, Fig. 12). Emergence successes of clutches laid by the same individuals were weakly correlated (intraclass correlation coefficient = 0.26). 35

47 Table 6. Coefficients from the optimal linear mixed-effect model predicting emergence success of Dermochelys coriacea clutches. The model includes incubation duration in the fixed component, female identity in the random component, and exponential error variance, ε i N 0,σ 2 e 2 ( centered incubation i )( 0.13) ( ); centered incubation duration is calculated by subtracting mean incubation duration from incubation duration of each sample. Variable Estimate SE df t-value p-value Intercept Incubation duration Figure 12. Incubation duration versus emergence success of Dermochelys coriacea clutches incubated in Jamursba Medi hatcheries during 2010 nesting season. Line drawn based on coefficients from the optimal linear mixed-effects model for emergence success (Table 6). The optimal LME model for hatching success included incubation duration (F1,24 = , p = ) and hatcheries (F2,24 = 9.425, p = 0.001) in the fixed 36

48 component, female identity in the random component, and exponential error variance structure (R 2 LME(m) = 0.20, R 2 LME(c) = 0.31). For clutches that hatched, hatching success increased with incubation duration, and was greater for clutches incubated in Warmamedi hatchery than in Wembrak west hatchery (Table 7, Fig. 13). Hatching successes of clutches laid by the same individuals were weakly correlated (intraclass correlation coefficient = 0.13). Table 7. Coefficients from the optimal linear mixed-effect model predicting hatching success of Dermochelys coriacea. The model includes hatchery and incubation duration in the fixed component, female identity in the random component, and exponential error variance, ε i N 0,σ 2 e 2 centered incubation i ( )( 0.19) ( ) ; centered incubation duration is calculated by subtracting mean incubation duration from incubation duration of each sample. Estimates for Wembrak east and Warmamedi hatcheries are differences from the model intercept. Variable Estimate SE df t-value p-value Intercept (Wembrak west hatchery) Wembrak east hatchery Warmamedi hatchery Incubation duration

49 Figure 13. Incubation duration versus hatching success of Dermochelys coriacea clutches incubated in hatcheries in Jamursba Medi during 2010 nesting season. Lines drawn based on coefficients from the optimal linear-mixed effect model for hatching success (Table 7). The NEP foraging group had greater remigration intervals than the NPTZ and SCS groups. Mean (± 1 SE) remigration interval was 4.0 ± 0.3 years for the NEP foraging group, 2.3 ± 0.4 years for the NPTZ group, and 2.9 ± 0.4 years for the SCS group (Table 8, Fig.14). Foraging group explained 31% of variation in remigration interval (R 2 LME(m) = 0.31, R 2 LME(c) = 0.73). Remigration intervals of individuals were strongly correlated (intraclass correlation coefficient = 0.60). Five of eight leatherbacks that were seen nesting in three seasons had variable remigration intervals (Table 9). 38

50 Table 8. Coefficients from the optimal linear mixed-effect model predicting remigration interval of Dermochelys coriacea. The model includes foraging region in the fixed component and female identity in the random component. Foraging regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). Estimates for the NPTZ and SCS regions are differences from the model intercept. Variable Estimate SE df t-value p-value Intercept (NEP) < NPTZ SCS Figure 14. Frequency distribution of remigration intervals of Dermochelys coriacea classified to distinct foraging regions: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). Only turtles with greater than 70% group membership probability are included. 39

51 Table 9. Variation in remigration interval of seven Dermochelys coriacea detected nesting in three seasons in the Bird s Head peninsula, Papua Barat, Indonesia. All turtles had greater than 70% group membership probability. Remigration intervals (years) No. of turtles Foraging group 2, 2 2 NPTZ, SCS 2, 3 2 SCS 3, 2 1 NPTZ 3, 4 2 NEP DISCUSSION Stable isotopes Inferring the foraging region of nesting leatherbacks using stable isotopes ( stable isotope tracking ) was possible because each region had distinct isotopic profiles. Distinct biogeochemical processes determine isotopic values of primary producers, which are predictably integrated into consumer tissues such as skin of leatherback turtles. Using a combination of satellite and stable isotope tracking, I found that leatherback skin δ 15 N and δ 13 C values were linked to the NEP, NPTZ, and SCS foraging areas (Fig. 3a). Using compound-specific stable isotope analyses of amino acids, Seminoff et al. (2012) demonstrated that greater skin δ 15 N of eastern Pacific foragers (equivalent to the NEP foraging group in this study) was not a result of those leatherbacks foraging at a higher tropic level, but instead the result of differring δ 15 N values among basal producers. The δ 15 N values were likely associated with distinct nitrogen cycling regimes greater δ 15 N with denitrification and lesser δ 15 N with nitrogen fixation (Montoya 2007). In the eastern Pacific coast, algae 40

52 productivity is fueled by denitrified water from the eastern tropical Pacific advected poleward by California undercurrent (Liu & Kaplan 1989, Altabet et al. 1999, Voss et al. 2001). A large percentage of global pelagic denitrification occurs in the eastern tropical Pacific (Cline & Richards 1972, Codispoti & Richards 1976). In the oxygendepleted zone, nitrate serves as an alternate electron receptor for bacterial respiration (denitrification). Lighter nitrogen isotope ( 14 NO3 - ) is preferentially used in the denitrification process, leaving 15 N-enriched water (Cline & Kaplan 1975, Liu & Kaplan 1989, Voss et al. 2001). In contrast, in the highly stratified SCS and oligotrophic NPTZ, primary productivity is fueled by molecular nitrogen (Karl et al. 1997, Shiozaki et al. 2010) made biologically available by nitrogen-fixing prokaryotes (diazotrophs), such as Trichodesmium spp. (Minagawa & Wada 1986, Saino & Hattori 1987, Capone et al. 1997). The organic form of nitrogen resulting from microbial fixation is only slightly more depleted in 15 N than atmospheric N2 (δ 15 N = 0 ; Wada 1980, Wada & Hattori 1991, Carpenter et al. 1997), resulting in lesser δ 15 N integrated into tissues of primary producers and consumers. The NPTZ and SCS foraging groups both had lesser skin δ 15 N, but they were distinguishable by their δ 13 C values (Fig. 3a). Unlike Seminoff et al. (2012), I found there was enough variation in skin δ 13 C values to distinguish the two groups, which was likely caused by inclusion of greater number of satellite-tracked leatherbacks in this study (n = 31 versus n = 13 in Seminoff et al. 2012). As with δ 15 N, skin δ 13 C values reflect δ 13 C of primary producers. Algae δ 13 C is generally a function of their 41

53 growth rates, which is influenced by availability of dissolved phosphate, aqueous CO2, and cell geometry (Goericke & Fry 1994, Bidigare et al. 1997, Pancost et al. 1997, Popp et al. 1998). Algae growth may be slower in less productive waters of the NPTZ than in productive neritic waters of the NEP and SCS, resulting in lesser δ 13 C being integrated into skin of leatherbacks that foraged in the NPTZ. The SCS foraging group had considerably less isotopic ranges than the NEP group (Fig. 3a), likely reflecting the amount of isoscape structure in these regions. Although leatherback foraging in these regions spans similar latitudinal range (Benson et al. 2011), leatherbacks in the NEP encounter a more diverse isoscape than those in the SCS. In the eastern North Pacific, δ 15 N and δ 13 C baselines decrease by 2-3 from temperate to high latitude areas (Saino & Hattori 1987, Goericke & Fry 1994, Aurioles et al. 2006). In addition to using coastal water shelf (<200m), leatherbacks also use offshore areas in central and northern California (Benson et al. 2011), which would have lesser δ 13 C baseline. In contrast, leatherback foraging was only limited to nearshore areas of the SCS. Despite having greater sample size, not all satellite-tracked leatherbacks were correctly classified. Given variation in isotope values within each foraging region, an even greater sample size might be needed to improve correct assignment rates. Foraging group composition of the nesting population varied between 2010 and The nesting population consisted of 48% of the NEP foraging group in 2010 and 27% in These percentages are similar to 33-60% estimated by Seminoff et 42

54 al. (2012) during the and 2010 nesting seasons. Inclusion of misclassified leatherbacks and differential detection of foraging groups could contribute to this variation in nesting population composition but it could also arise from differential timing of breeding of foraging groups driven by habitat quality. Only leatherbacks with greater than 70% group membership probability were considered in the estimation. Although minimized, it is possible that some incorrectly classified leatherbacks were included, which could distort estimates of group percentages. Spatial and temporal separation in nesting activities among foraging groups could lead to differential detection thus resulting in one group being overrepresented, but there was no evidence for this. All three foraging groups were detected throughout the sampling periods at all study areas. Another possible explanation is that variation in nesting population composition reflected differences in timing of breeding of the foraging groups. Foraging habitat quality will greatly influence leatherbacks timing of breeding. Favorable foraging conditions in a particular region might lead to many leatherbacks reaching breeding body condition, resulting in similar timing of breeding. Because energetic demands and gains vary by region, the amount of time required to accumulate sufficient energy will differ among groups. At this point, the number of favorable foraging seasons required by each foraging group to reach breeding body condition is unknown so I cannot predict when a particular group would predominate at the nesting beach. Prey response to favorable or poor ocean climates (e.g., prey 43

55 quality) also will influence timing of breeding. For example, the NEP foraging group requires more productive seasons than the NPTZ and SCS groups because of the greater energetic demand. But the potential energetic returns might be greater in the NEP than in other regions during favorable ocean climates thus greatly reducing the amount of time to accumulate sufficient energy reserves. Having knowledge on breeding body condition threshold and how each group responds to regional variation in foraging conditions would lead to better understanding of patterns in abundance at the nesting beach. Body size and reproductive output The NEP environment might select for larger leatherbacks. Leatherbacks use their large body size ( thermal inertia ; Paladino et al. 1990) and behaviors to thermoregulate (Bostrom et al. 2010), which allow them to forage in cooler waters. In the NEP, greater body size enables leatherbacks to store greater reserves for fueling the long trans-pacific migration and weathering unfavorable, cooler, foraging conditions. Productivity and water temperature in the NEP are more variable than in the NPTZ and SCS (Longhurst 2007). When water temperature decreases in the NEP during winter, leatherbacks migrate to the distant Eastern Equatorial Pacific (EEP; Benson et al. 2011). In the NPTZ, the fronts shift interannually but they persist year-round (Polovina et al. 2001, Bograd et al. 2004), and leatherbacks could select favorable temperature and prey regimes within the Transition Zone as C. caretta turtles do (Kobayashi et al. 2008). In the SCS, leatherbacks move shorter distances 44

56 between areas northeast of Palawan and Sunda shelf near Borneo likely to adjust to monsoon-driven seasonal access to prey (Benson et al. 2011). Having greater body size to store more reserves in the seasonally limited and highly variable NEP foraging environment would ensure survival during unfavorable foraging seasons. Body size and foraging region influenced the number of clutches laid in a nesting season. Larger individuals laid greater numbers of clutches, which indicated that large body size might lead to greater access to resources. The NPTZ foraging group laid greater numbers of clutches in a season. Considering they had the shortest remigration interval of two years, they might be contributing greater reproductive resources into the population over their lifetimes than the NEP and SCS groups. These results, however, must be interpretted with caution. There was great amongindividual variability. Although the influence of body size and foraging region on clutch frequency was statistically significant, the two factors explained only about 9% of variation in the data. Finally, although estimating clutch frequencies was better than relying on observed clutch frequencies, the true clutch frequencies were still likely underestimated. Clutch size and egg mass varied greatly among individuals. Despite a large sample size, I did not have enough power to detect trends in clutch size and egg mass associated with foraging region. Great among-individual variability and intraclass correlations, which characterized clutch size and egg mass data in this study, decreased power in multilevel analyses (Hedges & Rhoads, 2010). Including 45

57 individual- or cluster-level covariates that explain considerable between- and withinvariances can increase power (Hedges & Rhoads, 2010) but covariates I used (year and body size as measured by CCL and CCW) explained little variation in the data as indicated by low marginal R 2. Mass, body condition index, girth, and thickness of neck or hind flipper may be better indicators of nutritional status (Davenport et al. 2009, 2011, Harris et al. 2011, Plot et al. 2013), and thus be better at explaining among-individual variability in clutch size and egg mass. Clutch size and egg mass were strongly repeatable in clutches laid by the same females, indicating heterogeneity in individual quality. The existence of high-quality, typically older and more experienced, individuals in populations has been documented in numerous taxa (e.g., Forslund & Pärt 1995, Cam & Monnat 2000, Lescroël et al. 2009). These individuals have greater probability of reproduction, reproductive performance, and survival (Cobley et al. 1998, Weladji et al. 2008, Hamel et al. 2009) thus contributing disproportionately greater reproductive resources into the population. Age and experience of leatherbacks in this study were unknown thus their influence on reproductive output could not be examined. It is important to evaluate whether reproductive performance is repeatable across multiple nesting seasons, not just within a season. Long-term longitudinal studies that track leatherback reproduction and survival during various conditions will provide insights on how individuals respond to changes in their habitat quality. Variation in egg mass might be related to ocean climate. The Multivariate 46

58 ENSO Index (MEI), Northern Oscillation Index (NOI), and Southern Oscillation Index (SOI) showed negative or neutral anomalies, indicative of lesser productivity 12 months before the 2010 nesting season, and positive anomalies before the 2011 season (Fig. 15). In a less productive year, leatherbacks seemed to invest more per clutch (heavier eggs). In a study to examine the impact of dietary quality on reproductive allocation, Warner et al. (2007) reported a similar observation. Female Amphibolurus muricatus with a poor-quality diet produced lesser number of clutches of heavier eggs. Wallace et al. (2006b) reported that albumen comprised 63% of leatherback eggs and explained 33% of variation in egg mass. Greater egg mass in 2010 might be due to increased albumen provisioning, which Wallace et al. (2006b) suggested might facilitate water absorption and offer fitness benefits to hatchlings. In a more productive year, conversely, leatherbacks might lay more numerous but smaller eggs, which would undoubtedly increase their fitness (probability of producing more hatchlings). Limited by my ability to detect all nesting emergences, however, I detected no year effect in clutch frequency and no tradeoff between number and mass of eggs that would have supported this hypothesis. The decrease in egg mass was least for the NPTZ foraging group and greatest for the SCS group (Fig. 9), which might indicate lesser interannual variability in the NPTZ foraging habitat and greater interannual variability in the SCS. Longer remigration intervals of the NEP group might be buffering the effect of a less productive year on egg mass. 47

59 Figure 15. Environmental conditions before 2010 and 2011 nesting seasons (shaded) as indicated by monthly Multivariate ENSO Index (MEI; solid line), Northern Oscillation Index (NOI; dashed line), and Southern Oscillation Index (SOI; dotted line). The MEI values were multiplied by -1 to make the values have the same sign as NOI and SOI values. MEI monthly values were obtained from NOAA Earth System Research Laboratory s Physical Science Division ( psd/enso/mei/table.html), and NOI and SOI from the Pacific Fisheries Environmental Laboratory s Live Server Access ( As expected, incubation environment rather than maternal foraging region affected hatchling phenotypes, emergence success, and hatching success. Clutch location (in situ or hatchery) and incubation duration affected hatchling mass and hatching success, and incubation duration affected emergence success. Poor hatching success at Wembrak has been linked to greater sand temperatures (Tapilatu & Tiwari 2007). Investigation into effects of incubation environment on hatching success and hatchling sex ratio in Bird s Head nesting beaches is ongoing (Tapilatu et al. unpublished). 48

60 Similar to leatherbacks in other populations in the world, Bird s Head leatherbacks typically return to nest within two and three years, but there was considerable variation among foraging groups (Fig. 14). Additionally, remigration interval varied within individuals (Table 9). These variations were likely influenced by fluctuations in foraging habitat quality, a link that had been suggested by previous investigators (e.g., Carr & Carr 1970) and supported by recent studies. Recent research indicated that variation in remigration interval, and consequently annual nesting population, is associated with food habits (Hatase & Tsukamoto 2008), trophic status (Broderick et al. 2001), foraging areas (Vander Zanden et al. 2014), ocean productivity (Wallace et al. 2006a), winter sea surface temperatures (Solow et al 2002, Chaloupka et al. 2008, van Houtan et al. 2010), and climate variability (Saba et al 2007, van Houtan et al. 2010). Because each foraging group was capable of returning to the nesting beaches in two and three years, I could not rule out that some of the longer remigration intervals resulted from undetected nesting seasons. Despite this possibility, I will discuss how habitat quality of each foraging region might explain the observed variation in remigration interval. Greater remigration intervals of the NEP foraging group might be a consequence of foraging in the furthest region with limited foraging season. From Bird s Head nesting beaches, leatherbacks travel for months to reach the NEP region, and their journey back takes approximately 5-7 months (Benson et al. 2011). They forage between early summer and late fall (Starbird et al. 1993, Benson et al. 49

61 2007a), when dense aggregations of scyphozoan jellies develop in retention areas along the coast within upwelling shadows (Shenker 1984, Graham 1994, Benson et al. 2007a). Leatherbacks were more abundant in nearshore areas in years when Northern Oscillation Indices were positive (Benson et al. 2007a, 2011), which were typically associated with increased zooplankton production (Schwing 2002). In the winter, colder sea surface temperatures and limited prey availability had been hypothesized to drive leatherbacks to overwinter in warmer waters of the EEP (Benson et al. 2011). Although foraging in EEP was estimated to occur in only 6% of daily locations (Benson et al. 2011), leatherbacks could be consuming Pyrosoma spp., which had energy density 5-10 times greater than scyphozoan jellies (Davenport & Balazs 1991, Jones et al. unpublished). Stomach contents of by-caught Lepidochelys olivacea foraging in similar latitudes consisted of primarily Pyrosoma spp. (Davenport & Balazs 1991, Polovina et al. 2004). Great variability in remigration interval within the NEP foraging group likely reflect interannual variability in foraging conditions as well as individual ability. One turtle, confidently assigned to the NEP foraging region (0.97 membership probability), returned to Bird s Head nesting beaches in 2009 and 2011, which indicated it was able to accumulate sufficient energy during the 2010 foraging season. Nearshore areas of the NEP, however, were not particularly productive in Monthly NOI values fluctuated considerably between negative and positive anomalies throughout the year (Fig. 15). It is unclear how this particular individual 50

62 was able to reach breeding body condition at distant NEP in such a short time. Most leatherbacks that foraged in the NEP, however, had three, four, and five years between consecutive nesting seasons. These longer remigration intervals were consistent with observations by Benson et al. (2011) who had recaptured individuals in central California in one to three subsequent years, indicating a minimum of two to four foraging seasons in the NEP. In contrast, leatherbacks that foraged in the NPTZ region had shorter (1-3 years) and less variable remigration intervals (Fig. 14), which indicated greater and less variable annual energetic return associated with foraging in this region. The NPTZ foraging group expended less on migration than the NEP group and possibly less on thermoregulation and foraging as well. Leatherbacks arrive at this middistance region within 3-7 months post nesting and occupy mostly warmer waters ( C; Benson et al. 2011). I have little information on leatherback prey items in this region, but a leatherback by-caught in this region had its stomach full of Pyrosoma colonies (Davenport & Balazs 1991). Additionally, Pyrosoma spp. are one of six most common prey items of C. caretta foraging in the same region (Parker et al. 2005). To what extent this nutritious prey was a component of the leatherback turtle diet is unknown, and knowledge on the ecology of Pyrosoma is limited. Greater feeding rates and short generation times among other adaptations, however, allow pelagic tunicates such as Pyrosoma to reach great densities in unpredictable and patchy environments (Alldredge & Madin 1982). The NPTZ foraging group 51

63 exhibited the least foraging activity (17% of daily locations; Benson et al. 2011) yet they are able to reach breeding body condition relatively fast. Foraging on reliable patches of nutritious prey in mid-distance region with favorable thermal regime may explain shorter and less variable remigration intervals for leatherbacks that foraged in the NPTZ. Leatherbacks that foraged in the SCS also had shorter and less variable remigration intervals than those foraging in the NEP. To reach the SCS, leatherbacks migrate for 1-2 months and they occupy warmer waters ( C; Benson et al. 2011). They forage year-round and exhibit greater amounts of foraging activity than leatherbacks in the other regions (59% of daily locations; Benson et al. 2011). Little is known about what leatherbacks eat in the SCS, which limits my ability to infer energetic gains. The SCS has the least primary productivity among the three regions (Longhurst 2007) and likely supports lesser jelly biomass. The SCS region, however, is surrounded by densely populated countries whose activities can lead to localized jelly blooms (Purcell 2012). Heat dumping rather than heat retention is the primary thermoregulatory challenge in the SCS. This may not be energetically costly, however, because they are consistently exposed to a similar thermal regime throughout their lives, whereas leatherbacks that move to the temperate regions have to acclimate to much cooler water. Shorter migration distance to the SCS might lead to leatherbacks attaining breeding body condition faster and allowing shorter remigration intervals. 52

64 Apparent reproductive benefits associated with foraging in the NEP region were not detected. On contrary, results of this study indicated leatherbacks that foraged in the NEP region might be at a disadvantage because they seemed to reproduce less often than those in the NPTZ and SCS regions. If there are no reproductive benefits, why do leatherback turtles use the distant NEP foraging region? Gaspar et al. (2012) used ocean currents to model dispersal of hatchlings from the New Guinea nesting beaches, including Jamursba Medi and Wermon. Their models revealed that variability in ocean current determined the partitioning of hatchlings into different dispersal areas, all of which, except the Indian Ocean, were used by adult leatherbacks. He suggested that these juveniles return to previously known foraging areas to maximize their chances of survival instead of searching for new areas (the learned migration goal hypothesis). In other words, they likely did not select a particular foraging region after sampling all or a number of them. In adults, fidelity to the NEP foraging region is well documented (Benson et al. 2011). Reproduction is only one function to which leatherbacks allocate energy based on energetic income and expenditures associated with a particular foraging region. Lesser frequency of breeding exhibited by leatherbacks that foraged in the NEP region might be balanced with earlier maturity, greater survivorship, or longer life span, all of which ultimately determine lifetime reproductive output. In this study, the link between non-breeding habitat quality and reproduction of the western Pacific leatherback turtle was examined. Foraging region of nesting 53

65 leatherbacks was inferred based on their skin stable isotope values, which was validated with satellite tracking. The ease and cost-effectiveness of stable isotope tracking allowed me to obtain the required large sample size. The use of distinct foraging regions resulted in intrapopulation variation in body size, clutch frequency, and remigration interval, but not in clutch size and egg mass. Leatherbacks that foraged in the NEP were larger, and had greater and more variable remigration intervals than leatherbacks that foraged in the NPTZ and SCS. In contrast, leatherbacks that foraged in the NPTZ laid more clutches during a nesting season and nest every two years on average. Variation in distance from the Bird s Head nesting beaches, water temperature, prey abundance and quality likely affected energy budgets of individuals, which resulted in the observed intrapopulation variation. The multilevel modeling approach permitted inclusion of multiple observations from the same females and quantified correlation among them. Strong repeatability in clutch size and egg mass in multiple clutches laid by the same females indicated heterogeneity in individual quality that led to great among-individual variability. The decrease in egg mass in 2011 and variation in remigration interval within individuals and foraging regions indicated that individuals responded to interannual variation in foraging habitat quality. The Bird s Head leatherback population is subjected to a mosaic of threats by foraging in distinct regions, but having multiple foraging groups might buffer the rate of population decline due to foraging habitat loss. Nevertheless, loss of one foraging 54

66 habitat would affect population growth in a complex way through reduced recruitment. The growth and dynamics of the western Pacific leatherback population depends not only on fecundity but also on growth rate, age or size at maturity, probability of survival, and longevity. Future studies that quantify and examine how foraging habitat quality influence these parameters will provide insights into the population dynamics of western Pacific leatherback population. 55

67 LITERATURE CITED Alldredge AL, Madin LP (1982) Pelagic tunicates: unique herbivores in the marine plankton. Bioscience 32: Altabet MA, Pilskaln C, Thunell R, Pride C, Sigman D, Chavez F, Francois R (1999) The nitrogen isotope biogeochemistry of sinking particles from the margin of the Eastern North Pacific. Deep-Sea Res Pt I 46: Aurioles D, Koch PL, Le Boeuf BJ (2006) Differences in foraging location of Mexican and California elephant seals: evidence from stable isotopes in pups. Marine Mammal Sci 22: Barrow LM, Bjorndal KA, Reich KJ (2008) Effects of preservation method on stable carbon and nitrogen isotope values. Physiol Biochem Zool 81: Benson SR, Forney KA, Harvey JT, Carretta JV, Dutton PH (2007a) Abundance, distribution, and habitat of leatherback turtles (Dermochelys coriacea) off California, Fish Bull 105: Benson SR, Dutton PH, Hitipeuw C, Samber B, Bakarbessy J, Parker D. (2007b) Post-nesting migrations of leatherback turtles (Dermochelys coriacea) from Jamursba-Medi, Bird s Head Peninsula, Indonesia. Chelonian Conserv Bi 6: Benson SR, Kisokau KM, Ambio L, Rei V, Dutton PH, Parker D (2007c) Beach use, internesting movement, and migration of leatherback turtles, Dermochelys coriacea, nesting on the north coast of Papua New Guinea. Chelonian Conserv Bi 6:7 14 Benson SR, Eguchi T, Foley DG, Forney KA, Bailey H, Hitipeuw C, Samber BP, Tapilatu RF, Rei V, Ramohia P, Pita J, Dutton PH (2011) Large-scale movements and high-use areas of western Pacific leatherback turtles, Dermochelys coriacea. Ecosphere 2: art84 Bidigare RR, Fluegge A, Freeman KH, Hanson KL, Hayes JM, Hollander D, Jasper JP, King LL, Laws EA, Milder J, Millero FJ, Pancost R, Popp BN, Steinberg PA, Wakeham SG (1997). Consistent fractionation of 13 C in nature and in the laboratory: Growth-rate effects in some haptophyte algae. Global Biogeochem Cy 11: Bjorndal KA, Carr A (1989) Variation in clutch size and egg size in the green turtle nesting population at Tortuguero, Costa Rica. Herpetologica 45:

68 Block BA, Jonsen ID, Jorgensen SJ, Winship AJ, Shaffer SA, Bograd SJ, Hazen EL, Foley DG, Breed GA, Harrison AL, Ganong JE, Swithenbank A, Castleton M, Dewar H, Mate BR, Shillinger GL, Schaefer KM, Benson SR, Weise MJ, Henry RW, Costa DP (2011) Tracking apex marine predator movements in a dynamic ocean. Nature 475:86 90 Bograd SJ, Foley DG, Schwing FB, Wilson C, Laurs RM, Polovina JJ, Howell EA, Brainard RE (2004) On the seasonal and interannual migrations of the transition zone chlorophyll front. Geophys Res Lett 31: L17204 Booth DT (2006) Influence of incubation temperature on hatchling phenotype in reptiles. Physiol Biochem Zool 79: Booth DT, Astill K (2001) Incubation temperature, energy expenditure and hatchling size in the green turtle (Chelonia mydas), a species with temperature-sensitive sex determination. Aust J Zool 49: Bostrom BL, Jones TT, Hastings M, Jones DR (2010) Behaviour and physiology: the thermal strategy of leatherback turtles. PloS ONE, 5:e13925 Brenna JT, Corso TN, Tobias HJ, Caimi RJ (1997) High-precision continuous-flow isotope ratio mass spectrometry. Mass Spectrom Rev 16: Broderick AC, Godley BJ, Hays GC (2001) Trophic status drives interannual variability in nesting numbers of marine turtles. Proc R Soc B 268: Burnham KP, Anderson DR (2002) Model selection and multi-model inference: a practical information-theoretic approach. Springer-Verlag, New York, NY, p van Buskirk J, Crowder LB (1994) Life-history variation in marine turtles. Copeia 1994:66 81 Cam E, Monnat J (2000) Apparent inferiority of first-time breeders in the kittiwake: the role of heterogeneity among age classes. J Anim Ecol 69: Capone DG, Zehr JP, Paerl HW, Bergman B, Carpenter EJ (1997) Trichodesmium, a globally significant marine cyanobacterium. Science 276: Carpenter EJ, Harvey HR, Fry B, Capone DG (1997) Biogeochemical tracers of the marine cyanobacterium Trichodesmium. Deep-Sea Res Pt I 44: Carr A, Carr MH (1970) Modulated reproductive periodicity in Chelonia. Ecology 51:

69 Ceriani SA, Roth JD, Evans DR, Weishampel JF, Ehrhart LM (2012) Inferring foraging areas of nesting loggerhead turtles using satellite telemetry and stable isotopes. PLoS ONE 7:e45335 Chaloupka M, Kamezaki N, Limpus C (2008) Is climate change affecting the population dynamics of the endangered Pacific loggerhead sea turtle? J Exp Mar Biol Ecol 356: Chan EH, Liew HC (1995) Incubation temperatures and sex-ratios in the Malaysian leatherback turtle (Dermochelys coriacea). Biol Conserv 74: Chao SY, Shaw PT, Wu SY (1996) Deep water ventilation in the South China Sea. Deep-Sea Res Pt I 43: Cline JD, Kaplan IR (1975) Isotopic fractionation of dissolved nitrate during denitrification in the eastern tropical North Pacific Ocean. Mar Chem 3: Cline JD, Richards FA (1972) Oxygen deficient conditions and nitrate reduction in the eastern tropical North Pacific Ocean. Limnol Oceanogr 17: Cobley ND, Croxall JP, Prince PA (1998) Individual quality and reproductive performance in the Grey-headed Albatross Diomedea chrysostoma. Ibis 140: Codispoti LA, Richards FA (1976) An analysis of the horizontal regime of denitrification in the eastern tropical North Pacific. Limnol Oceanogr 21: Congdon JD, Gibbons JW (1985) Egg components and reproductive characteristics of turtles: relationships to body size. Herpetologica 41: Coplen TB, Krouse HR (1998) Sulphur isotope data consistency improved. Nature 392:32 32 Davenport J, Balazs GH (1991) Fiery bodies - Are pyrosomas an important component of the diet of leatherback turtles. British Herpetological Society Bulletin 37:33 38 Davenport J, Plot V, Georges JY, Doyle TK, James MC (2011) Pleated turtle escapes the box shape changes in Dermochelys coriacea. J Exp Biol 214: Davenport J, Fraher J, Fitzgerald E, McLaughlin P, Doyle T, Harman L, Cuffe T (2009). Fat head: an analysis of head and neck insulation in the leatherback turtle (Dermochelys coriacea). J Exp Biol 212: Drent RH, Daan S (1980) The prudent parent: energetic adjustments in avian 58

70 breeding. Ardea 68: Dutton P, McDonald D (1994) Use of PIT tags to identify adult leatherbacks. Marine Turtle Newsletter 67:13 14 Dutton PH, Hitipeuw C, Zein M, Benson SR, Petro G, Pita J, Rei V, Ambio L, Bakarbessy J (2007) Status and genetic structure of nesting populations of leatherback turtles (Dermochelys coriacea) in the western Pacific. Chelonian Conserv Bi 6:47 53 Eckert SA (2006) High-use oceanic areas for Atlantic leatherback sea turtles (Dermochelys coriacea) as identified using satellite telemetered location and dive information. Mar Biol 149: Eckert SA (2002) Distribution of juvenile leatherback sea turtle Dermochelys coriacea sightings. Mar Ecol Prog Ser 230: Eisenberg JF, Frazier J (1983) A leatherback turtle (Dermochelys coriacea) feeding in the wild. J Herpetol 17:81 82 Elgar MA, Heaphy LJ (1989) Covariation between clutch size, egg weight and egg shape: comparative evidence for chelonians. J Zool 219: Forslund P, Pärt T (1995) Age and reproduction in birds hypotheses and tests. TREE 10: Frazer NB, Richardson JI (1986) The relationship of clutch size and frequency to body size in loggerhead turtles, Caretta caretta. J Herpetol 20:81 84 Fry B, Silva SR, Kendall C, Anderson RK (2002) Oxygen isotope corrections for online δ34s analysis. Rapid Commun Mass Sp 16: Gaspar P, Benson SR, Dutton PH, Réveillère A, Jacob G, Meetoo C, Dehecq A, Fossette S (2012) Oceanic dispersal of juvenile leatherback turtles: going beyond passive drift modeling. Mar Ecol Prog Ser 457: Godley BJ, Blumenthal JM, Broderick AC, Coyne MS, Godfrey MH, Hawkes LA, Witt MJ (2008) Satellite tracking of sea turtles: Where have we been and where do we go next. Endang Species Res 4:3 22 Godley BJ, Barbosa C, Bruford M, Broderick AC, Catry P, Coyne MS, Formia A, Hays GC, Witt MJ (2010) Unraveling migratory connectivity in marine turtles using multiple methods. J Appl Ecol 47: Goericke R, Fry B (1994) Variations of marine plankton δ 13 C with latitude, 59

71 temperature, and dissolved CO2 in the world ocean. Global Biogeochem Cy 8:85 90 Graham WM (1994) The physical oceanography and ecology of upwelling shadows. PhD dissertation. University of California, Santa Cruz. Guraya SS (1989) Ovarian follicles in reptiles and birds. Springer-Verlag, Berlin, p Hall LS, Krausman PR, Morrison ML (1997) The habitat concept and a plea for standard terminology. Wildlife Soc B 25: Hamann M, Limpus CJ, Whittier JM (2002) Patterns of lipid storage and mobilisation in the female green sea turtle (Chelonia mydas). J Comp Physiol B 172: Hamel S, Côté SD, Gaillard JM, Festa-Bianchet M (2009) Individual variation in reproductive costs of reproduction: high-quality females always do better. J Anim Ecol 78: Harris HS, Benson SR, Gilardi KV, Poppenga RH, Work TM, Dutton PH, Mazet JAK (2011) Comparative health assessment of western Pacific leatherback turtles (Dermochelys coriacea) foraging off the coast of California, J Wildlife Dis 47: Harrison XA, Blount JD, Inger R, Norris DR, Bearhop S (2011) Carry-over effects as drivers of fitness differences in animals. J Anim Ecol 80:4 18 Hatase H, Tsukamoto K (2008) Smaller longer, larger shorter: energy budget calculations explain intrapopulation variation in remigration intervals for loggerhead sea turtles (Caretta caretta). Can J Zool 86: Hays GC, Speakman JR (1991) Reproductive investment and optimum clutch size of loggerhead sea turtles (Caretta caretta). J Anim Ecol 60: Hedges LV, Rhoads C (2010) Statistical power analysis in education research. U. S. Department of Education. Washington DC: National Center for Special Education Research. Institute of Education Sciences Hirth, HF (1980) Some aspects of the nesting behavior and reproductive biology of sea turtles. Am Zool 20: Hitipeuw C, Dutton PH, Benson S, Thebu J, Bakarbessy J Population status and internesting movement of leatherback turtles, Dermochelys coriacea, nesting on the northwest coast of Papua, Indonesia. Chelonian Conserv Bi 6:

72 Hobson KA (2008) Applying isotopic methods to tracking animal movements. In: Hobson KA, Wassenaar LI (eds) Tracking animal migration with stable isotopes. Academic Press, Oxford, UK, p Houghton JDR, Hays GC (2001) Asynchronous emergence by loggerhead turtle (Caretta caretta) hatchlings. Naturwissenschaften 88: van Houtan KS, Halley JM (2011) Long-term climate forcing in loggerhead sea turtle nesting. PLoS ONE 6:e19043 Hughes GR, Luschi P, Mencacci R, Papi F (1998) The 7000-km oceanic journey of a leatherback turtle tracked by satellite. J Exp Mar Biol Ecol 229: Huyer, A (1983) Coastal upwelling in the California current system. Prog Oceanogr 12: James MC, Ottensmeyer CA, Myers RA (2005) Identification of high-use habitat and threats to leatherback sea turtles in northern waters: new directions for conservation. Ecol Lett 8: Johnson MD (2007) Measuring habitat quality: a review. Condor 109: Kappes MA, Shaffer SA, Tremblay Y, Foley DG, Palacios DM, Robinson PW, Costa DP Hawaiian albatrosses track interannual variability of marine habitats in the North Pacific. Prog Oceanogr 86: Karl DM, Letelier R, Hebel D, Tupas L, Dore J, Christian J, Winn C (1995) Ecosystem changes in the North Pacific subtropical gyre attributed to the El Nino. Nature 373: Karl D, Letelier R, Tupas L, Dore J, Christian J, Hebel D (1997) The role of nitrogen fixation in biogeochemical cycling in the subtropical North Pacific Ocean. Nature 388: Kobayashi DR, Polovina JJ, Parker DM, Kamezaki N, Cheng I, Uchida I, Dutton PH, Balazs GH (2008) Pelagic habitat characterization of loggerhead sea turtles, Caretta caretta, in the North Pacific Ocean ( ): Insights from satellite tag tracking and remotely sensed data. J Exp Mar Biol Ecol 356: Kwan D (1994) Fat reserves and reproduction in the green turtle, Chelonia mydas. Wildlife Res 21: Laurs RM, Lynn RJ (1977) Seasonal migration of North Pacific albacore, Thunnus alalunga, into North American coastal waters: distribution, relative abundance, 61

73 and association with Transition Zone waters. Fish Bull 75: Lee Chen YL (2005) Spatial and seasonal variations of nitrate-based new production and primary production in the South China Sea. Deep-Sea Res Pt I 52: Lescroël A, Dugger KM, Ballard G, Ainley DG (2009). Effects of individual quality, reproductive success and environmental variability on survival of a long-lived seabird. J Anim Ecol 78: Liu KK, Chao SY, Shaw PT, Gong GC, Chen CC, Tang TY (2002) Monsoon-forced chlorophyll distribution and primary production in the South China Sea: observations and a numerical study. Deep-Sea Res Pt I 49: Liu KK, Kaplan IR (1989) The eastern tropical Pacific as a source of 15 N-enriched nitrate in seawater off southern California. Limnol Oceanogr 34: Liu Q, Jiang X, Xie SP, Liu WT (2004) A gap in the Indo-Pacific warm pool over the South China Sea in boreal winter: seasonal development and interannual variability. J Geophys Res 109: C07012 Liu G, Chai F (2009) Seasonal and interannual variability of primary and export production in the South China Sea: a three-dimensional physical biogeochemical model study. ICES J Mar Sci 66: Longhurst A (2007) Ecological geography of the sea. Second edition. Academic Press, Amsterdam, Netherlands, p Lynn RJ, Bograd SJ, Chereskin TK, Huyer A (2003) Seasonal renewal of the California Current: the spring transition of California. J Geophys Res 108:3279 Madsen T, Shine R (1999) The adjustment of reproductive threshold to prey abundance in a capital breeder. J Anim Ecol 68: Matsuzawa Y, Sato K, Sakamoto W, Bjorndal K (2002) Seasonal fluctuations in sand temperature: effects on the incubation period and mortality of loggerhead sea turtle (Caretta caretta) pre-emergent hatchlings in Minabe, Japan. Mar Biol 140: McGehee MA (1990) Effects of moisture on eggs and hatchlings of loggerhead sea turtles (Caretta caretta). Herpetologica 46: McGowan JA, Cayan DR, Dorman LM (1998) Climate-ocean variability and ecosystem response in the northeast Pacific. Science 281: Minagawa M, Wada E (1986) Nitrogen isotope ratios of red tide organisms in the East 62

74 China Sea: a characterization of biological nitrogen fixation. Mar Chem 19: Montoya JP (2007) Natural abundance of 15N in marine planktonic ecosystems. In: Michener R, Lajtha K (eds) Stable isotopes in ecology and environmental science. Blackwell publishing, Malden, p Nakagawa S, Schielzeth H (2010) Repeatability for Gaussian and non-gaussian data: a practical guide for biologists. Biol Rev 85: Nakagawa S, Schielzeth H (2013) A general and simple method for obtaining R 2 from generalized linear mixed-effects models. Methods Ecol Evol 4: Naulleau G, Bonnet X (1996) Body condition threshold for breeding in a viviparous snake. Oecologia 107: Newton I (2004) Population limitation in migrants. Ibis 146: Norris DR, Marra PP (2007) Seasonal interactions, habitat quality, and population dynamics in migratory birds. Condor 109: Paladino FV, O'Connor MP, Spotila JR (1990) Metabolism of leatherback turtles, gigantothermy, and thermoregulation of dinosaurs. Nature 344: Pancost RD, Freeman KH, Wakeham SG, Robertson CY (1997) Controls on carbon isotope fractionation by diatoms in the Peru upwelling region. Geochim Cosmochim Ac 61: Parker DM, Cooke WJ, Balazs GH (2005) Diet of oceanic loggerhead sea turtles (Caretta caretta) in the central North Pacific. Fish Bull 103: Pinheiro J, Bates D, DebRoy S, Sarkar D and the R Development Core Team (2013) nlme: Linear and Nonlinear Mixed Effects Models. R package version Plot V, Jenkins T, Robin J, Fossette S, Georges J (2013) Leatherback turtles are capital breeders: morphometric and physiological evidence from longitudinal monitoring. Physiol Biochem Zool 86: Polovina JJ, Howell E, Kobayashi DR, Seki MP (2001) The transition zone chlorophyll front, a dynamic global feature defining migration and forage habitat for marine resources. Prog Oceanogr 49: Polovina JJ, Kobayashi DR, Parker DM, Seki MP, Balazs GH (2000) Turtles on the edge: movement of loggerhead turtles (Caretta caretta) along oceanic fronts, spanning longline fishing grounds in the central North Pacific, Fish 63

75 Oceanogr 9:71 82 Polovina JJ, Balazs GH, Howell EA, Parker DM, Seki MP, Dutton PH (2004). Forage and migration habitat of loggerhead (Caretta caretta) and olive ridley (Lepidochelys olivacea) sea turtles in the central North Pacific Ocean. Fish Oceanogr 13:36 51 Popp BN, Laws EA, Bidigare RR, Dore JE, Hanson KL, Wakeham SG (1998) Effect of phytoplankton cell geometry on carbon isotopic fractionation. Geochim Cosmochim Ac 62: Purcell JE (2012) Jellyfish and ctenophore blooms coincide with human proliferations and environmental perturbations. Ann Rev Mar Sci 4: Qu T (2000) Upper-Layer circulation in the South China Sea. J Phys Oceanogr 30: Rau GH, Sweeney RE, Kaplan IR (1982) Plankton 13 C: 12 C ratio changes with latitude: differences between northern and southern oceans. Deep-Sea Res Pt A 29: Ream RR, Sterling JT, Loughlin TR (2005) Oceanographic features related to northern fur seal migratory movements. Deep-Sea Res Pt II 52: Reece SE, Broderick AC, Godley BJ, West SA (2002) The effects of incubation environment, sex and pedigree on the hatchling phenotype in a natural population of loggerhead turtles. Evol Ecol Res 4: Reich KJ, Bjorndal KA, del Rio CM (2008) Effects of growth and tissue type on the kinetics of 13 C and 15 N incorporation in a rapidly growing ectotherm. Oecologia 155: Roden GI (1970) Aspects of the Mid-Pacific Transition Zone, J Geophys Res 75: Rong Z, Liu Y, Zong H, Cheng Y (2007) Interannual sea level variability in the South China Sea and its response to ENSO. Global Planet Change 55: Rubenstein DR, Hobson KA (2004) From birds to butterflies: animal movement patterns and stable isotopes.trends Ecol Evol 19: Saba VS, Spotila JR, Chavez FP, Musick JA (2008) Bottom-up and climatic forcing on the worldwide population of leatherback turtles. Ecology 89: Saba VS, Santidrián-Tomillo P, Reina RD, Spotila JR, Musick JA, Evans DA, 64

76 Paladino FV (2007) The effect of the El Niño Southern Oscillation on the reproductive frequency of eastern Pacific leatherback turtles. J Appl Ecol 44: Saino T, Hattori A (1987) Geographical variation of the water column distribution of suspended particulate organic nitrogen and its 15 N natural abundance in the Pacific and its marginal seas. Deep-Sea Res Pt A 34: Schwing FB, Murphree T, Green PM (2002) The Northern Oscillation Index (NOI): a new climate index for the northeast Pacific. Prog Oceanogr 53: Seminoff JA, Bjorndal KA, Bolten AB (2007) Stable carbon and nitrogen isotope discrimination and turnover in pond sliders Trachemys scripta: insights for trophic study of freshwater turtles. Copeia 3: Seminoff JA, Benson SR, Arthur KE, Eguchi T, Dutton PH, Tapilatu RF, Popp BN (2012) Stable isotope tracking of endangered sea turtles: validation with satellite telemetry and δ 15 N analysis of amino acids. PLoS ONE 7:e37403 Shenker JM (1984) Scyphomedusae in surface waters near the Oregon coast, May August Estuar Coast Shelf S 19: Shiozaki T, Furuya K, Kodama T, Kitajima S, Takeda S, Takemura T, Kanda J (2010) New estimation of N2 fixation in the western and central Pacific Ocean and its marginal seas. Global Biogeochem Cy 24:GB1015 Simmons SE, Crocker DE, Kudela RM, Costa DP (2007) Linking foraging behaviour of the northern elephant seal with oceanography and bathymetry at mesoscales. Mar Ecol Prog Ser 346: Solow AR, Bjorndal KA, Bolten AB (2002) Annual variation in nesting numbers of marine turtles: the effect of sea surface temperature on re-migration intervals. Ecol Lett 5: Starbird CH, Baldridge A, Harvey JT (1993) Seasonal occurrence of leatherback sea turtles (Dermochelys coriacea) in the Monterey Bay region, with notes on other sea turtles California Fish and Game 79:54 62 Strub PT, Allen JS, Huyer A, Smith RL, Beardsley RC (1987) Seasonal cycles of currents, temperatures, winds, and sea level over the northeast Pacific continental shelf: 35 N to 48 N. J Geophys Res 92: Tapilatu, RF, Dutton PH, Tiwari M, Wibbels T, Ferdinandus HV, Iwanggin WG, Nugroho BH (2013) Long-term decline of the western Pacific leatherback, 65

77 Dermochelys coriacea: a globally important sea turtle population. Ecosphere 4: art25 Tapilatu RF, Tiwari M (2007) Leatherback turtle, Dermochelys coriacea, hatching success at Jamursba-Medi and Wermon beaches in Papua, Indonesia. Chelonian Conserv Bi 6: Tiwari M, Bjorndal KA (2000) Variation in morphology and reproduction in loggerheads, Caretta caretta, nesting in the United States, Brazil, and Greece. Herpetologica 56: Vander Zanden HB, Pfaller JB, Reich KJ, Pajuelo M, Bolten AB, Williams KL, Frick MG, Shamblin BM, Nairn CJ, Bjorndal KA (2014) Foraging areas differentially affect reproductive output and interpretation of trends in abundance of loggerhead turtles. Mar Biol 161: Voss M, Dippner JW, Montoya JP (2001) Nitrogen isotope patterns in the oxygendeficient waters of the Eastern Tropical North Pacific Ocean. Deep-Sea Res Pt I 48: Wallace BP, Saba VS (2009) Environmental and anthropogenic impacts on intraspecific variation in leatherback turtles: opportunities for targeted research and conservation. Endanger Species Res 7:11 21 Wallace BP, Kilham SS, Paladino FV, Spotilla JR (2006a) Energy budget calculations indicate resource limitation in eastern Pacific leatherback turtles. Mar Ecol Prog Ser 318: Wallace BP, Sotherland PR, Santidrián-Tomillo P, Reina RD, Spotila JR, Paladino FV (2007)Maternal investment in reproduction and its consequences in leatherback turtles. Oecologia 152: Wallace BP, Sotherland PR, Santidrián-Tomillo P, Bouchard SS, Reina RD, Spotila JR, Paladino FV (2006b) Egg components, egg size, and hatchling size in leatherback turtles." Comp Biochem Phys A 145: Wang C, Wang W, Wang D, Wang Q (2006) Interannual variability of the South China Sea associated with El Niño. J Geophys Res 111:C03023 Warner DA, Lovern MB, Shine R (2007) Maternal nutrition affects reproductive output and sex allocation in a lizard with environmental sex determination. Proc R Soc B 274: Webster MS, Marra PP, Haig SM, Bensch S, Holmes RT (2002) Links between 66

78 worlds: unraveling migratory connectivity. Trends Ecol Evol 17: Weladji RB, Loison A, Gaillard JM, Holand Ø, Mysterud A, Yoccoz NG, Nieminen M, Stenseth NC (2008) Heterogeneity in individual quality overrides costs of reproduction in female reindeer. Oecologia 156: Weng KC, Castilho PC, Morrissette JM, Landeira-Fernandez AM, Holts DB, Schallert RJ, Goldman KJ, Block BA (2005) Satellite tagging and cardiac physiology reveal niche expansion in salmon sharks. Science 310: Weimerskirch H, Inchausti P, Guinet C, Barbraud C (2003) Trends in bird and seal populations as indicators of a system shift in the Southern Ocean. Antarct Sci 15: Wyrtki K (1961) The physical oceanography of the southeast Asian waters. NAGA report 2:1 195 Zbinden JA, Bearhop S, Bradshaw P, Margaritoulis D, Newton J, Godley BJ (2011) Migratory dichotomy and associated phenotypic variation in marine turtles revealed by satellite tracking and stable isotope analysis. Mar Ecol Prog Ser 421: Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed effects models and extensions in ecology with R. Springer, New York, NY, p

79 APPENDICES Appendix 1. Skin δ 34 S of satellite-tracked Dermochelys coriacea sampled in the Bird s Head peninsula, Papua Barat, Indonesia and their foraging destinations: the Northeast Pacific (NEP), North Pacific Transition Zone (NPTZ), and South China Sea (SCS). All turtles were part of a movement study by Benson et al. (2011). PIT tag Sampling date Sampling location δ 34 S ( ) Foraging destination A 23 July 2007 Jamursba Medi NEP A 26 July 2007 Jamursba Medi NEP A 22 July 2007 Jamursba Medi NPTZ A 23 July 2007 Jamursba Medi NPTZ A 22 July 2007 Jamursba Medi SCS A 23 July 2007 Jamursba Medi SCS A 26 July 2007 Jamursba Medi SCS A 27 July 2007 Jamursba Medi SCS A 29 July 2007 Jamursba Medi NEP? A 29 July 2007 Jamursba Medi NPTZ? A 22 July 2007 Jamursba Medi Unknown A 26 July 2007 Jamursba Medi Unknown A 27 July 2007 Jamursba Medi Unknown 68

80 Appendix 2. Descriptive statistics of reproductive output variables of Dermochelys coriacea nesting in Papua Barat, Indonesia between April and September of 2010 and One measurement represents an individual. Variable Mean ± SD n Estimated clutch frequency ± Clutch size 83.5 ± No. yolkless eggs 31.3 ± Egg mass (g) 76.8 ± Egg diameter (mm) ± Yolkless egg mass (g) 23.0 ± Yolkless egg diameter (g) ± Hatchling mass (g) 42.1 ± Hatchling straight carapace length (mm) ± Hatchling straight carapace width (mm) ± Calculated for turtles seen more than once during the nesting season. 69

81 Appendix 3. Relationship between egg mass and diameter, and their correlation coefficient. 70

82 Appendix 4. Relationships between hatchling mass, straight carapace length (SCL), and straight carapace width (SCW), and their correlation coefficients. 71

83 Chapter II Detecting nutritional stress in leatherback turtles using stable isotope values of skin 72

84 ABSTRACT Fasting is a behavioral adaptation that allows an animal to allocate time and energy to other activities such as migration, reproduction, or hibernation rather than acquiring food. Generally, organisms adapt to periods of food deprivation by mobilizing fat and minimizing protein loss. Measuring changes in body composition is most reliable for detecting nutritional stress but this method requires sacrificing subjects, thus making it unsuitable for studies on wild animal population. Stable isotopes offered a nondestructive alternative for detecting nutritional stress, but field and laboratories studies indicated mixed results. Most consistently reported is enrichment of 15 N in tissue and excreta (increase in δ 15 N) with prolonged nutritional stress. The potential use of stable isotopes to detect nutritional stress in nesting leatherback turtles (Dermochelys coriacea), which primarily rely on endogenous reserve, was examined. Two skin samples were obtained from the same leatherbacks (n = 53) during separate nesting emergences in Skin δ 15 N increased gradually with sampling interval (days fasting), which indicated that leatherback turtles were transitioning from fat to protein catabolism. Skin δ 13 C, however, did not change significantly. There was great among-individual variability in changes in skin δ 15 N and δ 13 C, which suggested variation in initial body composition. I demonstrated that skin δ 15 N is a promising tool for detecting nutritional stress in nesting leatherback turtles. Future studies should include other indicators of nutritional stress to help interpret changes in stable isotope values. 73

85 INTRODUCTION Many vertebrates are able to tolerate periods of prolonged fasting. Distinct from starvation when lack of food limits feeding, animals forgo opportunities to feed during fasting (Castellini & Rea 1992, McCue 2010). Bears, seals, and penguins routinely fast during their annual cycles, and these prolonged fasts occur during energetically expensive activities such as migration, reproduction, and molting (Mrosovsky & Sherry 1980, King & Murphy 1985, Cherel et al. 1988). Generally, vertebrates adapt to long periods of food deprivation by mobilizing fat and conserving protein (Felig 1979, Goodman et al. 1984, Le Maho et al. 1988, Robin et al. 1988). Fasting can progress to starvation when food or a signal to resume feeding are lacking. Starvation can be divided into three sequential phases marked by changes in physiology or by the use of primary oxidative fuels (Le Maho et al. 1981, Robin et al. 1987, Cherel et al. 1988, Castellini & Rea 1992, Hervant & Renault 2002, Caloin 2004). During rapid phase I, body mass and resting metabolic rate (RMR) decrease rapidly and fat is mobilized. Phase II is marked by lesser body mass loss, RMR, and nitrogen excretion, constant protein utilization, and greater use of fat. Phase III, a critical phase, is marked by slow depletion of fat and protein. The loss of essential proteins can potentially cause death. This late phase is reversible by resumption of feeding. An internal signal associated with increased protein use induces behavioral changes such as locomotion and food search before lethal depletion of body reserves (Cherel et al. 1987, Cherel & Le Maho1991, Le Maho et al. 1988, Koubi et al. 1991). 74

86 Body composition changes during food deprivation. McCue (2010) reviewed these changes and the different ways animals survive starvation. Various tissues and organs contain carbohydrate, fat, or protein depots. With increased fasting duration, loss of mass may occur in some tissues and organs but not others. Tissues with high metabolic costs such as gastrointestinal tissues can decrease greatly in mass (McNurlan & Garlick 1981, McMillan & Houlihan 1989, Croom et al. 1999) whereas heart, brain, and kidney may be preserved (Goodman et al. 1984, Merkle & Hanke 1988, Cherel et al. 1992, Soengas et al. 1996). The summed decrease in mass of tissues and organs equals the total body mass lost. Although measuring changes in body composition is best for detecting nutritional stress, this method requires sacrificing the organism and often is not suitable for studies on wild populations. Instead, researchers developed non-lethal ways to detect nutritional stress by monitoring loss of body mass, plasma metabolites, and stable isotopes in tissues or breath. The purpose of this study was to document changes in stable nitrogen and carbon isotope values in skin of western Pacific leatherback turtles (Dermochelys coriacea) associated with prolonged fasting during a nesting season. Understanding how prolonged fasting affects tissue stable isotope values will help researchers using this technique to be cognizant of such potential effect. Similar to bears, seals, and penguins, leatherback turtles routinely fast during their annual cycles. They migrate great distances between foraging region and nesting beach (Hughes et al. 1998, James et al. 2005, Eckert et al. 2006, Benson et al. 2007, 2011). During migration and nesting periods they forgo 75

87 feeding and rely primarily on endogenous reserves accumulated during the foraging period (Caut et al. 2008, Plot et al. 2013). Recent studies indicated that leatherbacks may forage during the nesting season (Southwood et al. 2005, Fossette et al. 2008, Casey et al. 2010), but this foraging is presumed to be of lesser magnitude than foraging in the nonbreeding areas. Plot et al. (2013) reported that leatherbacks nesting in French Guiana lost approximately 11% of their initial body mass during the 71-day nesting season. They also recorded a significant decrease in plasma concentrations of glucose, triacylglycerides, and urea, which indicated these leatherbacks relied on stored fat during the nesting season. Stable nitrogen and carbon isotopes in tissues have been used increasingly to detect nutritional stress, but field and laboratory studies indicated mixed results (Hatch 2012). Enrichment of 15 N in tissue and excreta (increase in δ 15 N) associated with nutritional stress is most consistently reported in studies of vertebrates (Hobson et al. 1993, Cherel et al. 2005, Gaye-Siesseger et al. 2007, Castillo & Hatch 2007, McCue & Pollock 2008), but enrichment of tissue and enrichment of excreta do not occur together. As McCue (2007) suggested, there are two competing hypotheses on how nutritional stress affects δ 15 N values. Hobson et al. (1993) observed enrichment in liver and muscle of fasted wild Ross Geese (Chen rossii), and suggested that excreta would be 15 N depleted. During prolonged fast, 14 N is preferentially excreted without being replaced by protein input, leaving tissue enriched (Gannes et al. 1997). In contrast, McCue & Pollock (2008) observed no changes in tissue δ 15 N of fasted reptiles while excreta became 15 N enriched. 76

88 They suggested two pools from which nitrogen can be recycled: labile (circulating in the forms of protein and amino acids) and nonlabile (in tissues). As labile sources of nitrogen become depleted in prolonged nutritional stress, 15 N-enriched nonlabile pool contributes more to the excreted nitrogen (enriched excreta). Due to the slow turnover rate of skin (~ 4-6 months; Seminoff et al. 2007, Reich et al. 2008), I predicted no change or a gradual increase in skin δ 15 N and no change in skin δ 13 C with fasting duration. MATERIALS AND METHODS Skin samples were collected from leatherback turtles nesting at the Bird s Head peninsula on the northwest coast of Papua Barat, Indonesia ( S, E; Fig. 1). Jamursba Medi (18 km long) and Wermon (6 km long) beaches are the primary nesting sites for the western Pacific leatherback turtle (Hitipeuw et al. 2007, Tapilatu et al. 2013). Samples were opportunistically collected between early June and mid August of

89 Figure 1. The northwest coast of Papua, Indonesia, showing Dermochelys coriacea nesting beaches where skin samples were collected (Tapilatu et al. 2013). Two skin samples were collected from the same individual at two different nesting emergences. Skin was sampled from the dorsal axial region of left and right hind flippers during the egg-laying trance or nest covering. Sampling surface was scrubbed with alcohol to remove algae, and a small sample of the epidermis (<10 x 10 x 1 mm, g wet mass) was obtained using a single edge razor. Betadine was applied to the sampling surface following sampling. Samples were preserved in 70% ethanol or in saturated salt solution in 2-ml cryogenic vials; neither preservatives affected isotope values (Barrow et al. 2008). Skin samples were rinsed with deionized water, finely diced with a scalpel blade, then freeze-dried at -50 C for 12 hours in a lyophilizer (BenchTop K, VirTis, SP 78