Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Stable Isotope Tracking of Endangered Sea Turtles: Validation with Satellite Telemetry and δ15N Analysis of Amino Acids

  • Jeffrey A. Seminoff ,

    jeffrey.seminoff@noaa.gov

    Affiliation National Oceanic and Atmospheric Administration–National Marine Fisheries Service, Southwest Fisheries Science Center, La Jolla, California, United States of America

  • Scott R. Benson,

    Affiliation National Oceanic and Atmospheric Administration–National Marine Fisheries Service, Southwest Fisheries Science Center, La Jolla, California, United States of America

  • Karen E. Arthur,

    Affiliation Department of Geology and Geophysics, University of Hawai’i at Manoa, Honolulu, Hawai’i, United States of America

  • Tomoharu Eguchi,

    Affiliation National Oceanic and Atmospheric Administration–National Marine Fisheries Service, Southwest Fisheries Science Center, La Jolla, California, United States of America

  • Peter H. Dutton,

    Affiliation National Oceanic and Atmospheric Administration–National Marine Fisheries Service, Southwest Fisheries Science Center, La Jolla, California, United States of America

  • Ricardo F. Tapilatu,

    Affiliations Marine Laboratory, The State University of Papua, Manokwari, Papua Barat Province, Indonesia, Department of Biology, University of Alabama, Birmingham, Alabama, United States of America

  • Brian N. Popp

    Affiliation Department of Geology and Geophysics, University of Hawai’i at Manoa, Honolulu, Hawai’i, United States of America

Stable Isotope Tracking of Endangered Sea Turtles: Validation with Satellite Telemetry and δ15N Analysis of Amino Acids

  • Jeffrey A. Seminoff, 
  • Scott R. Benson, 
  • Karen E. Arthur, 
  • Tomoharu Eguchi, 
  • Peter H. Dutton, 
  • Ricardo F. Tapilatu, 
  • Brian N. Popp
PLOS
x

Abstract

Effective conservation strategies for highly migratory species must incorporate information about long-distance movements and locations of high-use foraging areas. However, the inherent challenges of directly monitoring these factors call for creative research approaches and innovative application of existing tools. Highly migratory marine species, such as marine turtles, regularly travel hundreds or thousands of kilometers between breeding and feeding areas, but identification of migratory routes and habitat use patterns remains elusive. Here we use satellite telemetry in combination with compound-specific isotope analysis of amino acids to confirm that insights from bulk tissue stable isotope analysis can reveal divergent migratory strategies and within-population segregation of foraging groups of critically endangered leatherback sea turtles (Dermochelys coriacea) across the Pacific Ocean. Among the 78 turtles studied, we found a distinct dichotomy in δ15N values of bulk skin, with distinct “low δ15N” and “high δ15N” groups. δ15N analysis of amino acids confirmed that this disparity resulted from isotopic differences at the base of the food chain and not from differences in trophic position between the two groups. Satellite tracking of 13 individuals indicated that their bulk skin δ15N value was linked to the particular foraging region of each turtle. These findings confirm that prevailing marine isoscapes of foraging areas can be reflected in the isotopic compositions of marine turtle body tissues sampled at nesting beaches. We use a Bayesian mixture model to show that between 82 and 100% of the 78 skin-sampled turtles could be assigned with confidence to either the eastern Pacific or western Pacific, with 33 to 66% of all turtles foraging in the eastern Pacific. Our forensic approach validates the use of stable isotopes to depict leatherback turtle movements over broad spatial ranges and is timely for establishing wise conservation efforts in light of this species’ imminent risk of extinction in the Pacific.

Introduction

Elucidating patterns of migratory connectivity for broad-ranging animals is central to defining spatio-temporal management priorities for these species. Knowledge about movements and foraging area destinations can discern the relative value of discrete regions for resource acquisition, and when examined at a population level, can provide insights into demographic consequences of habitat selection patterns [1][3]. However, inherent challenges of directly observing long-distance migratory behaviors often require creative application of existing tools to infer prevailing movement patterns [4], [5].

Satellite telemetry has been a primary tool to track long-distance movements of vertebrates from around the world [1], [6], [7], but the high cost of tags, satellite tracking data acquisition, and analysis (up to US $5K per animal) coupled with large body size requirements for tracked animals have precluded its application for many vulnerable species and age classes. These challenges have often resulted in extremely limited samples sizes, which thereby limit the value of telemetry as a tool for revealing migratory variability within focal populations. A new approach to rapidly assess the movements of large cross sections of imperiled populations could greatly enhance conservation of such animals.

Stable isotope analysis (SIA) is a relatively low cost, complementary approach to satellite tracking of animal movements. Isotopic compositions of consumer tissues integrate information from foraging environments [8], and thus, when an animal moves among spatially discrete food webs that are isotopically distinct (often termed ‘isoscapes’), stable isotope values of its tissues can provide unambiguous information about its previous location [4], [5], [9][14]. An additional advantage is that the stable isotope tracking method does not require initial marking of individuals, and thus every ‘capture’ (i.e. sampled animal) provides information on prior whereabouts.

Stable isotope animal tracking has been employed successfully for a variety of terrestrial and aerial species, due in large part to the well-understood systematic patterns of stable isotope variations in meteoric water and plants across continents [5], [13], [14]. In marine systems, however, patterns of isotopic abundances are poorly resolved for most ocean regions. Spatial gradients in stable isotope values at the base of the food web have been described on very large scales [15], [16], but few regional maps are available [17] and little is known about temporal shifts in isotope abundances in natural systems [17], [18]. For high-order consumers, the elaboration of foraging habitats is further obscured by the substantial difference in stable isotope values between their tissues and those of basal producers [8], [18][20]. Clearly, isotopic tracking of marine species would benefit from a greater understanding of spatiotemporal isotopic patterns in marine fauna as well as from clarification about the baseline influence on isotopic compositions of tissues of higher-order marine consumers [21].

Integrative studies that combine satellite tracking with stable isotope analysis are needed to validate the isotope tracking method [20], [22]. However, confounding such studies is the inability of bulk tissue analysis to distinguish the influence of a consumer’s trophic status from the effect of baseline (i.e. primary producer) isotope values of the ecosystem within which that animal resides [22], [23]. The application of compound-specific isotopic analyses of amino acids (CSIA-AA) can elucidate the differential impacts of these factors and thus can substantially enhance the value of stable isotope measurements for tracking marine species [21], [24]. This is possible because some AAs, such as phenylalanine, retain the isotopic composition of source nitrogen at the base of the food web, whereas other AAs, such as glutamic acid, are significantly enriched in 15N as they move through the food web [25], [26]. Baseline and trophic information can therefore be obtained from a consumer’s tissue without need for analyses of prey items or of basal food web samples [24], [26]. To date, however, this technique has not been validated for most taxa [27].

Marine turtles are an ideal taxon for such integrative studies due to their broad ranging movements, residence in discrete foraging areas, and fidelity to specific nesting beaches, where they can be easily sampled and equipped with telemetry devices [6]. Further, the extremely slow isotopic replenishment rates (i.e. turnover) of soft tissues of turtles [28], [29] allows for prevailing isotope regimes at foraging areas to be traced long after a subject’s departure. Leatherback turtles (Dermochelys coriacea), in particular, are highly mobile with migrations often spanning entire ocean basins [3]. They forage exclusively on gelatinous prey such as scyphomedusae and pyrosomes [30][33] that commonly show isotopic values that reflect regional food webs [34], [35]. Supported by previous studies that confirm the value of SIA for studying their life history [30], [36][38], leatherbacks are an ideal species in which to explore the value of stable isotope tracking to depict foraging areas over broad scales.

Pacific leatherbacks are considered at imminent risk of extinction [39] due to long-term harvest of eggs and adults and incidental capture in fishing gear [3], [39]. Knowledge of foraging areas and migratory connectivity in the Pacific, where these threats persist, is critical to the sustained management of leatherbacks. Satellite telemetry efforts have shown that leatherbacks nesting in Papua Barat Province, Indonesia inhabit multiple foraging areas throughout the Pacific [40]. However, despite these insights there is no information about the extent to which an individual’s putative foraging area influences the isotopic compositions of its body tissues, nor is their information about patterns of individual philopatry to specific foraging areas. If adequately resolved, the isotope tracking method may provide a means to determine foraging regions for a large number of Pacific leatherbacks and thus provide information vital to the development of effective management strategies.

Here, we build upon previous satellite tracking efforts [40] and examine the migratory strategies of Indonesian leatherbacks using nitrogen isotopic compositions of bulk tissue, the results of which are validated with compound specific nitrogen isotopic analysis of amino acids and satellite telemetry. Because of the apparent variability in nitrogen isotope abundances in the eastern vs. western Pacific [15], [16], and the fact that individuals from this stock forage in both broadly separated regions [40], we hypothesized that the stable nitrogen isotope ratios in leatherback skin tissue collected at the Indonesian nesting beach would reflect the foraging region used by individual turtles prior to the nesting season. If δ15N values of the tissues of turtles from these separate regions are sufficiently different, our approach may foster a transition from satellite telemetry to stable isotope tracking as a means to decipher animal movements in the Pacific and across ocean basins in general.

Results

Leatherbacks that were satellite tracked in this study departed their Indonesian nesting beach en route to two broad regions (Figure 1): (i) the ‘western Pacific’ (n = 8), with primary destinations in the Sulawesi, Sulu, and South China Sea region and the North Pacific Transitional Zone (NPTZ), and (ii) the ‘eastern Pacific’ (n = 5), with the prevailing terminus centered in the California Current Large Marine Ecosystem. Tracking data were acquired for an average 312±131 days per turtle, with six turtles tracked for more than one year (maximum tracking duration  = 519 days; Table S1). Most turtles, but not all, were tracked long enough to distinguish their particular foraging area. Putative foraging regions were corroborated by results from previous tracking efforts [40] that applied state-space models to show that the majority of these turtles had commenced area-restricted search behavior (i.e., assumed foraging) by the end of their respective tracking. There were two abbreviated tracks, with durations of 46 d and 95 d; however, the orientation of departure for both turtles, coupled with the consistency of this departure trajectory among numerous other turtles tracked to the western Pacific [40] indicated they were moving to western Pacific foraging grounds.

thumbnail
Figure 1. Map of satellite-tracked post-nesting movements of 13 leatherback turtles from Jamursba-Medi, Papua Barat Province Indonesia, overlaid with stable nitrogen isotopic values for skin tissue of nesting females from the same population.

Satellite-tracked turtles were studied in 2007 and 2010; non-tracked turtles (n = 65) were skin sampled in 2005–2007 and 2010. Blue and red track lines depict turtles within the ‘high δ15N’ and ‘low δ15N’ groups, respectively. Tracks with numbered termini depict movements of leatherbacks that were analysed for CSIA-AA (Table 2).

https://doi.org/10.1371/journal.pone.0037403.g001

Leatherback epidermis (stratum corneum, hereafter referred to as ‘skin’) samples (n = 78) had δ13C values ranging from –20.9 ‰ to –15.1 ‰ and δ15N values from 9.7 ‰ to 17.7 ‰ (Table 1, Figure 2). Bulk skin δ13C values were normally distributed about a single mean (–17.3±1.0 ‰), whereas bulk skin δ15N values had a strongly bimodal distribution with distinct ‘low δ15N’ (range = 9.7–12.4 ‰, n = 36) and ‘high δ15N’ (range = 12.9–17.7 ‰, n = 42) groups (see inset Figure 1, Figure 3a). When linked with telemetry results, we found that all turtles tracked to the eastern Pacific were within the high-δ15N group, whereas all but one turtle moving to western Pacific sites were in the low-δ15N group (Figure 1). However, despite the apparent spatial influences on δ15N values, these approaches alone could not rule out trophic influences on the measured δ15N disparity.

thumbnail
Table 1. Summary of leatherback tissue sampling and δ13C and δ15N values for turtles nesting in Jamursba Medi, Papua Barat, Indonesia.

https://doi.org/10.1371/journal.pone.0037403.t001

thumbnail
Figure 2. Scatterplot of stable carbon vs. stable nitrogen values for 78 leatherback turtles sampled at the Jamursba-Medi nesting beach in Papua Barat Indonesia.

https://doi.org/10.1371/journal.pone.0037403.g002

thumbnail
Figure 3. Leatherback turtle skin nitrogen stable isotopic compositions.

a, nitrogen isotopic composition of skin from leatherback turtles known to feed in the eastern and western Pacific; b, nitrogen isotopic composition of glutamic acid and phenylalanine in skin tissue from leatherback turtles feeding in the eastern and western Pacific; c, trophic position calculated using the isotopic compositions of glutamic acid and phenylalanine [17], presented as a weighted average based on AA measurement error. Error bars represent weighted standard deviation of trophic position means; n = 3 each. While there was a significant difference in skin δ15N between eastern and western Pacific foragers (t-test: p<0001), there was no significant difference in trophic position between these two groups (t-test: p = 0.682).

https://doi.org/10.1371/journal.pone.0037403.g003

Using CSIA of phenylalanine (a ‘source’ AA, sensu16), we found average δ15Nphenlyalanine (± SD) values of 7.27±0.03‰ and 4.17±0.05‰ for eastern and western Pacific foragers, respectively (Figure 3b). Combining δ15Nphenlyalanine values with δ15N of glutamic acid from the same samples (Table 2, Figure 3b) indicated that there was no difference in the trophic position for these two foraging groups (weighted average TPEP ± s.d.  = 2.36±0.01, TPWP ± s.d.  = 2.35±0.01, P = 0.682; Figure 3c), and that the difference in bulk skin δ15N values was instead driven by the differences in baseline nitrogen isotopic composition between the two regions.

thumbnail
Table 2. δ15N values (‰) of source and trophic amino acids in leatherback turtles known from satellite tracking to migrate to the western and eastern Pacific Ocean.

https://doi.org/10.1371/journal.pone.0037403.t002

We also examined the nitrogen isotopic compositions of a wide array of other trophic (i.e., alanine, valine, leucine, isoleucine and aspartic acid) and source (i.e., glycine, tryosine and serine) amino acids to determine if they too were consistent with interpretation of the results from only glutamic acid and phenylalanine as per previous marine ecosystem studies using CSIA-AA [41]. The δ15Nglutamic acid values were highly correlated with the average δ15N values of the trophic amino acids and the δ15Nphenlyalanine values were highly correlated with the average δ15N values of the source amino acids (Figure 4). In addition, the average differences in the δ15N values of eastern and western Pacific trophic and source amino acids averaged ca. 3 ‰, which is similar to the difference in means between the bulk skin high-δ15N and low-δ15N groups (Table 2).

thumbnail
Figure 4. Plot of the δ15N value of glutamic acid versus the average δ15N value of the trophic amino acids alanine, valine, leucine, isoleucine and aspartic acid (red), and the δ15N value of phenylalanine versus the average δ15N value of the source amino acids glycine, tyrosine and serine (blue).

The slopes of the regression lines in this figure are not different at the 95% confidence interval. The satellite-tracked movements of each turtle are shown in Figure 1.

https://doi.org/10.1371/journal.pone.0037403.g004

The disparate bulk skin δ15N values between known (i.e. satellite-tracked) eastern and western Pacific foragers combined with CSIA-corroboration that these differences were due to baseline isotope values in each region allowed us to characterize the foraging area destinations for each individual based on its skin δ15N value using a Bayesian mixture-model analysis. Among the six models considered (Table 3), all models performed well and no model stood out as the best according to our model selection method. Models 1 through 4 performed well in terms of the mean of the differences (near zero - unbiased), but Models 5 and 6 performed well with respect to the variability in the differences (good precision). Estimated variances of the two normal distributions were different (Figure 5), indicating two-variance models (Models 2, 4, and 6), were more appropriate than the one-variance models (Models 1, 3, and 5). One turtle (72489; Table S1) exhibited an eastern Pacific δ15N value (i.e. ‘high’ δ15N: 15.05‰) but was tracked to the western Pacific (Figure 1). Consequently, the results of the models that used the known foraging locations were influenced by this turtle, and resulted in more-precautionary foraging region assignments. Because of the likelihood that the proportion of foraging destinations changes annually, we use Model 4 (known foraging destinations, time-independent) and Model 6 (unknown foraging destinations, time-dependent) (Table 3) to assign foraging region to each turtle.

thumbnail
Table 3. Definitions of four models used to estimate the underlying isotopic signatures of the Eastern and Western Pacific Ocean and to assign each individual to one of the two foraging grounds.

https://doi.org/10.1371/journal.pone.0037403.t003

thumbnail
Figure 5. Marginal posterior distributions of δ15N values for the two assumed foraging grounds (bell-shaped curves) and individual assignments to one of the two foraging grounds (filled circles in horizontal arrangement) with color-coded probabilities of belonging to the eastern Pacific foraging area.

Panel A is from Model 4, whereas the Panel B is from Model 6. For both models, posterior distributions depicted in red are from the ‘low δ15N’ (i.e. putative western Pacific foragers), whereas those in blue represent the ‘high δ15N’ group (i.e., putative eastern Pacific foragers).

https://doi.org/10.1371/journal.pone.0037403.g005

The two estimated distributions of δ15N values (Figure 5) indicated that there were two isotopically distinct foraging regions for these turtles. All turtles that were tracked to the eastern Pacific were assigned to the distribution with the higher mean, whereas all but one turtle (72489; Table S1) that were tracked to the western Pacific foraging area were assigned to the distribution with the lower mean (Figure 1). Results of Model 4 included nine ambiguous assignments using our cut off point of 0.8 (Table 4); these nine turtles had intermediate skin isotopic compositions and thus could not be assigned to either region. Model 6, which did not incorporate known foraging areas, however, assigned all turtles to one of the two foraging grounds with probability greater than 0.8 (Table 4).

thumbnail
Table 4. Parameter estimates and results of mixture model assignments for eastern Pacific (EP) and western Pacific (WP) foraging.

https://doi.org/10.1371/journal.pone.0037403.t004

The proportion of turtles that foraged in the eastern Pacific was estimated from Model 4 to be approximately 0.57 annually (median, 95% posterior interval: 0.43–0.74). Using Model 6, which did not use the telemetry data, the annual proportion of turtles that use the eastern Pacific ranged from 0.3 to 0.6, although the small sample sizes resulted in large uncertainty (Table 4).

Discussion

Here we present the first study integrating satellite tracking with stable isotope analysis of relevant bulk tissues and amino acids to study marine turtle movements and stable isotope ecology. Isotope spatial gradients have been proposed as a viable means to track disparate foraging groups of marine turtles [42], [43] and seabirds [22], [23]; however, these studies were unable to decipher the influence of baseline control (i.e. spatial effect) vs trophic variation on isotope values. By integrating the CSIA approach, we demonstrate that differences in δ15N values of bulk skin from leatherbacks are not the result of variation in trophic position between groups, but instead due to differences in the baseline values of primary producers between two broad regions in the Pacific. Although control on tissue δ15N values of marine animals by the biogeochemistry of a region has been demonstrated based on AA-CSIA on local and global scales for other organisms [24], [44], it has not been demonstrated in marine turtles or other highly migratory species, nor has this observation been corroborated via satellite telemetry.

The application of CSIA to a broad group of trophic and source amino acids indicated that leatherbacks from the eastern and western Pacific foraged at virtually identical trophic levels (TPAA ∼2.4 based on the δ15N values of glutamic acid and phenylalanine; Figure 3c). This is a rather low trophic position for a large marine mega-vertebrate, yet not surprising considering that leatherback turtles consume gelatinous prey [3], [32], [33]. In the eastern Pacific, leatherbacks forage largely on the sea nettle (Chrysaora fuscescens) [40, S. Benson, unpubl. data], but in the western Pacific their diet is less clear, although there too leatherbacks are known to forage on gelatinous prey [33]. Additional empirical dietary data on leatherbacks would help clarify the precision of CSIA-AA for establishing a species’ trophic status. Regardless, the similarity in trophic status between leatherbacks foraging in the two ocean regions as determined by the isotopic compositions of a wide array of trophic and source amino acids signals an important step forward in our understanding of the isotope ecology of this critically endangered species.

The fact that leatherbacks are believed to inhabit foraging regions for the entirety of their non-breeding life phases [2], [40] suggests that the isotope profiles of these regions were able to fully integrate into leatherback body tissues prior to the commencement of their return migration to the nesting beach. However, it is likely that some individuals access multiple spatially disparate and ephemeral foraging sites within a broad region during non-reproductive periods. In the Atlantic, for example, leatherbacks may forage in the northeastern Atlantic during the first year of their migration and then cross the Atlantic to forage along the Canadian coasts or vice versa [45], [46]: leatherbacks are also known to undertake extensive seasonal north-south movements in the western North Atlantic during non-breeding periods [47]. Indeed among tracked leatherback turtles in the Pacific, leatherbacks have been shown to forage along the central California coast during late summer and fall, and move into more tropical latitudes of the eastern Pacific during winter periods [40]. In the Pacific, however, there are no data indicating that a leatherback would forage both in the eastern and western Pacific during a single inter-breeding cycle [40]. Thus, considering the elevated δ15N values throughout the eastern Pacific due to influences of denitrification, we believe that even if a turtle accessed multiple foraging hotspots in the region, prey in these areas would all have elevated δ15N values.

Isotopic turnover in adult turtle skin is believed to be on the order of four to six months [35], [36], and thus longer than most pre-nesting, foraging area-to-nesting beach migratory durations of Indonesian leatherbacks [40]. This suggests that upon arrival at the nesting beach, the δ15N of leatherback skin continues to reflect the isotopic conditions in the distant foraging areas, with the isotopic values most closely reflecting conditions at the last foraging site accessed by a turtle prior to the nesting migration. Further, even for the longest of migratory durations, the skin δ15N values are largely, if not exclusively, derived from foods consumed at distant foraging areas because marine turtles generally do not forage during migratory movements [3], unlike other taxa such as seabirds [48].

Interestingly, we found little difference in δ13C values of bulk skin between putative eastern and western Pacific foragers. This was unexpected considering that carbon isotope values of tissues have been used to track a variety of marine animals across large ocean regions [20], [48], [49]. Lack of bulk skin δ13C variability suggests that carbon isoscapes are more uniform in areas where leatherbacks foraged, despite the well-documented oceanic-neritic gradient in δ13C values of marine primary producers [50]. Perhaps the lesser magnitude of δ13C variation is due to the fact that leatherbacks foraged primarily in coastal regions in both eastern and western Pacific foraging grounds, with less use of offshore regions [2], [3], [40]. Nevertheless, we would expect turtles foraging in the relatively eutrophic waters of the eastern Pacific to diverge in their skin δ13C values from those turtles foraging in the shallower, oligotrophic waters of the Sulu Suluwesi and South China Sea regions due to the disparate oceanography of these two broad regions [51]. Although the reason(s) for low δ13C variability are unclear, our data indicate that nitrogen isotope values of skin are more effective than carbon for tracking the broad-scale movements of leatherback turtles, and perhaps other high seas-dwelling vertebrates in the Pacific. To further resolve spatial patterns in δ13C we encourage additional isotopic studies that focus on satellite-tracked Pacific leatherbacks.

The nitrogen isotope dichotomy in leatherback skin δ15N values was likely due to differences in nitrogen biogeochemistry in the eastern versus western Pacific [15]. The influence of nitrogen cycling on the δ15N values in marine turtle tissues has been suggested previously [34], [52]; however, so far no studies have examined δ15N in tissues of turtles from two N cycling regimes in the same ocean basin. In the present study, leatherbacks tracked to the western Pacific remained in an area dominated by N2-fixation, where source nitrogen has a lower isotopic composition [16], [53]. Turtles moving to the eastern Pacific – presumably to take advantage of gelatinous prey blooms in the California Current Large Marine Ecosystem [40] – enter a region characterized by high phytoplankton δ15N values in surface waters [54] caused by denitrification in the eastern Tropical Pacific [15], [55] and advection of this water mass northward [56]. Significantly, leatherbacks nesting in Pacific Costa Rica, all of which remain in the eastern Pacific for their entire post-nesting migrations [2], did not show bimodality in δ15N values of nesters, and instead had bulk tissue δ15N values more in line with our eastern Pacific group (mean δ15Nyolk  = 14.7±2.0 ‰ [37]), underscoring the uniformity in δ15N values among leatherbacks foraging throughout the eastern Pacific, and the role of nitrogen cycling regimes in driving observed patterns of marine turtle nitrogen isotopic compositions.

We acknowledge that the nitrogen isoscape spatial heterogeneity is likely more complex than the bimodal ‘eastern/western’ scenario present here. Indeed, previous simulations of δ15N patterns in surface waters of the Pacific based on N cycling regimes [15], [16], [57] revealed high-δ15N hotspots in the mid latitudes of the northeastern and southeastern Pacific, along with areas of low predicted δ15N spanning much of the North Pacific. Although empirical studies of δ15N values of baseline producers collected in-situ are needed, these models underscore the variability in δ15N values of surface waters across the Pacific. Two factors likely contributed to the dichotomous δ15N results. First, despite the long distance of satellite tracks, leatherback movements were restricted to relatively few areas of the Pacific, perhaps reflecting only a subset of all Pacific N isoscapes. For example, based on extensive previous satellite tracking efforts [40] it is probable that any east-bound turtle ultimately moved into the same general area along the Pacific coast of North America (central California USA to British Columbia Canada) to forage. Thus, most if not all eastern Pacific foragers likely integrated the composition of an isotopically uniform region. For those remaining in the western Pacific, there was more variation in foraging destination, with turtles moving to the North Pacific Transitional Zone or to the Indo-Pacific regions of the South China and Sulu Suluwesi seas. Yet despite the great distances that separate these areas, the entire region is characterized by a relatively uniform N cycling regime, with N-fixation as the dominant process [15], [16], [57]. Second, the bimodality in δ15N values of leatherback skin may result from the inability of skin - a slow turnover tissue - to capture the variability of δ15N as turtles move through different nitrogen isoscapes during foraging and reproductive migrations. Had a faster turning over tissue such as blood plasma been studied, perhaps a more complex isoscape would have been observed. However, such a tissue would likely have turned over too fast, thereby precluding the conservation of foraging region δ15N influences during the longest of nesting migrations [40].

Considering the integrative nature of this study, a primary caveat is that the relationship between leatherback skin δ15N values and satellite tracking destinations is partially hindered because isotopic compositions provide information on foraging area before nesting and satellite tracking pinpoints areas used after nesting. Effectively, these two techniques have a time lag between them. However, the fact that our Bayesian foraging area assignment tests (Model 6) correctly predicted the foraging destination of 12 of 13 (92%) of the tracked turtles underscores the value of the multi-tool approach despite this temporal disconnect. Nevertheless, to alleviate this issue in the future, we recommend that turtles are equipped with transmitters in foraging areas and tracked to nesting beaches, thus the satellite track and nesting beach skin sample would be reflective of the same time period. Further, if at the time of tagging, a tissue sample was collected, we could better understand the changes in isotopic composition during a nesting migration [sensu 58].

Conservation Value of Leatherback Isotope Tracking

Our forensic approach provides a solution to identifying spatial use patterns across vast areas for vulnerable, migratory marine species, and signals a promising, cost-effective technique for determining the spatial affinities of other highly migratory marine animals. We agree that the spatial resolution of the isotope tracking approach is orders of magnitude less than that of satellite tracking. However, satellite telemetry may cost up to US$ 5K per tracked animal, whereas the bulk tissue SIA and CSIA-AA for the same animal together cost < US$ 200. The substantial savings underscores the potential for examining a large number of individuals from a population of interest. It would also be instructive to employ additional cost-effective analysis of biomarkers such as trace elements [59] to determine if greater spatial resolution of leatherback migratory strategy can be gained by a larger suite of biogeochemical approaches.

Despite the apparent lesser spatial resolution afforded by stable isotope tracking, this technique has important conservation value for leatherback turtles. For example, the concordance of individual turtles’ δ15N groupings with distinct post-nesting migratory destinations suggests that most leatherbacks returned to the same ocean region from where they originated prior to nesting. This apparent segregation of turtles to eastern and western Pacific locations provides evidence of foraging area philopatry among most leatherback turtles nesting in Indonesia. Although a novel concept for leatherbacks, philopatry to specific foraging regions illustrates the potential for within-population variation in environmental influences on reproduction and other life history traits. In light of the spatial mosaic of anthropogenic impacts occurring throughout the Pacific, this finding also suggests that segments of this nesting population experience differential impacts from human pressures.

The decline of Pacific leatherbacks has been attributed to egg harvest at nesting beaches [39], [60] and incidental capture (bycatch) in marine fisheries [61], [62]. Commercial fisheries are ubiquitous in both regions, although stringent fisheries management in U.S. waters of the eastern Pacific–which includes a special leatherback management zone (i.e. Leatherback critical habitat designation; [63])– compared to the more widespread presence of unregulated artisanal net and longline fisheries in the western Pacific suggest that these two areas present strongly different conservation challenges for leatherbacks. By establishing the foraging region for a large cross section of the nesting population of leatherbacks, we are able to separate potential differential influences affecting different portions of the population. Further, by exploring the relative proportion of putative eastern vs. western Pacific foragers over larger time scales, discrepancies in return migration rates (i.e. survivorship) between the foraging groups may be revealed and perhaps will allow for the design of more spatially appropriate management recommendations.

Materials and Methods

Study Site

This research was conducted at Jamursba-Medi, a 18-km long nesting beach along the northern Bird’s Head coast in Papua Barat Indonesia (0° 22’ S, 132° 33’ E). This nesting beach hosts the largest remaining nesting aggregation for leatherback turtles in the Pacific Ocean, and has been the site of extensive monitoring and research efforts for more than a decade [64]. The nesting season extends from April to September, peaking in July [64].

Tissue Sampling and Preparation

Epidermis skin samples (i.e., stratum corneum; 0.10–0.25 g wet mass) were collected from 78 leatherbacks in 2005, 2006, 2007, and 2010 upon each turtle’s first witnessed nesting at the beginning of the nesting season with a razor from the dorsal surface of a hind flipper and stored in NaCl solution until transport to the lab. The top layer of the skin (stratum corneum) was separated from the underlying tissue (stratum germinativum), rinsed with deionized water, finely diced with a scalpel blade, then freeze-dried at –50°C for 12 h in a lyophilizer (BenchTop K, VirTis, SP Industries, Gardiner, NY, USA). Lipids were removed from skin samples using an accelerated solvent extractor (Model 200, Dionex, Bannockburn, IL, USA) with petroleum ether as the primary solvent. Following lipid extraction the samples were freeze dried at –50°C for 3 h to remove any residual solvent. Sub-samples of prepared tissue (0.6–1.0 mg) were weighed with a microbalance and packed in tin capsules for mass spectrometric analysis.

Bulk Skin Stable Isotope Analysis

Bulk skin δ13C and δ15N stable isotopic compositions were determined using an on-line C-N analyzer coupled with an isotope ratio mass spectrometer (Finnigan ConFlo II/Delta-Plus). Approximately 1.0 mg of each sample was loaded into sterilized tin capsules and analyzed by a continuous-flow isotope-ratio mass spectrometer in the Stable Isotope Laboratory at Scripps Institution of Oceanography, La Jolla, California, USA. We used a Costech ECS 4010 elemental combustion system interfaced via a ConFlo III device (Finnigan MAT, Bremen, Germany) to a Deltaplus gas isotope-ratio mass spectrometer (Finnigan MAT, Bremen, Germany). Sample stable isotope ratios relative to the isotope standard are expressed in the following conventional delta (δ) notation in parts per thousand (‰):(1)where Rsample and Rstandard are the corresponding ratios of heavy to light isotopes (e.g.15N/14N) in the sample and standard, respectively. Rstandard for 13C was Baker Acetanilide (C8H9NO; δ13C = –10.4) calibrated monthly against the Peedee Belemnite (PDB) limestone formation international standard. Rstandard for 15N was IAEA N1 Ammonium Sulfate ((NH4)2SO4; δ15N = +0.4) calibrated monthly against atmospheric N2 and USGS Nitrogen standards. All analytical runs included samples of standard materials inserted every 6 to 7 samples to calibrate the system and compensate for any drift over time. Hundreds of replicate assays of reference materials indicated measurement errors of 0.06‰ and 0.12‰ for carbon and nitrogen, respectively. Samples were combusted in pure oxygen in the elemental analyzer. Resultant gasses were reduced to N2 and CO2 gasses and passed through a series of thermal conductivity detectors and element traps to determine percent compositions. Acetanilide (71.09% C; 10.36% N) was used for calibration.

Satellite Telemetry

Deployments were concentrated after the peak of each nesting season (n2007 = 12, n2010 = 1) to maximize the amount of post-nesting movement data. We used a combination of satellite-linked transmitters on leatherback turtles during the study period (Table S1), including the following platform transmitter terminals (PTT) models: Wildlife Computers (Washington, USA) MK10 (n = 1) and SPLASH tags (n = 1); Telonics (Arizona, USA) ST20 (n = 4); and Sea Mammal Research Unit (SMRU, U.K.) Satellite Relay Data Loggers (SRDL) (n = 7). All PTTs featured a salt-water switch that suppressed transmission while submerged and were duty-cycled to optimize battery life (6 h on, 19 h off), reporting positions via System Argos (Landover, Maryland) every 1–3 d. Only the best position for each day with an Argos location quality code of ‘LC0’ or better (between 250 m and >1.5 km accuracy) was included in the tracks.

Before attaching transmitters we visually assessed each turtle for signs of injury or compromised health and only selected turtles that appeared to be in normal condition. To minimize impacts on nesting females, we waited until turtles began laying eggs, well into the nesting process when females enter a trance-like state, before sampling and attaching the transmitters. During handling we measured curved carapace length (CCL) and width (CCW), applied PIT and flipper tags [65], and attached the transmitters.

The 13 turtles tracked in this study include the 12 turtles tracked in 2007 [40] for which skin samples were available, and an additional tissue-sampled leatherback tracked in 2010. In 2007, we attached PTTs to leatherback turtles see 40 with the aid of a flexible harness that consisted of soft nylon webbing with flexible polyvinyl tubing over the shoulder straps and a corrodible pin designed to release the harness within 1824 months 40. In 2010, the single transmitter was attached to the nesting female with a direct attachment approach [see 66].

Compound Specific Stable Isotope Analysis of Amino Acids (CSIA-AA)

Skin samples from six satellite-tracked turtles (three that moved to the eastern Pacific and three that moved to the western Pacific) were analyzed for δ15N analysis of amino acids (Table S2). Although this technique is more expensive than δ15N analysis of bulk tissue, careful sample selection can yield important information that will enhance understanding of bulk isotope data sampled from a larger sample pool.

Samples were prepared for CSIA-AA by acid hydrolysis followed by derivatization to produce trifluoroacetic amino acid esters [67] using standard methods [41], [68]. Stable nitrogen isotopic composition was determined using a Delta XP mass spectrometer interfaced with a Trace Gas Chromatograph [68]. Measured isotopic compositions were corrected relative to known values for internal reference material - norleucine and aminoadipic acid. Samples were run in triplicate and standard deviations for each sample averaged 0.4‰ (range: <0.1 to 2.6‰). Average δ15N for source and trophic AAs was calculated with S.D. as follows:(2)

To confirm consistency in the δ15N measurement accuracy between the Scripps Institution of Oceanography mass spectrometer (used for bulk skin analyses) with that from the mass spectrometer at University of Hawaii we conducted bulk analyses on the same tissue samples as those from six individuals. These samples were analyzed for bulk carbon and nitrogen stable isotope composition, and the nitrogen isotopic composition of specific amino acids (Table S2).

Calculation of Trophic Position

The fractional trophic positions of leatherback samples were calculated using the measured δ15N values of glutamic acid and phenylalanine [Eq. 3; 26]. There are two assumed factors when using AAs to calculate trophic position - the trophic enrichment factor (TEF), and the so-called Beta value (β). The TEF is the difference between glutamic acid and phenylalanine at each trophic step (+7.6‰) and the β value (+3.4‰) is the difference between glutamic acid and phenylalanine δ15N values in marine plants [26]. Error associated with the trophic position calculation was determined by propagation of error using the uncertainly in β and the trophic enrichment factor [26] and the measured reproducibility for glutamic acid and phenylalanine for each sample.

The nitrogen isotopic composition of AAs from a wide range of marine primary producers, herbivores and omnivores has been derived with an equation using the isotopic composition of glutamic acid and phenylalanine to estimate fractional trophic position in the consumers [Eq. 3; 26]. The relationship between the δ15N values of glutamic acid and phenylalanine has been found to provide a good indicator of trophic position (TPAA) in a range of species including various zooplankton and fish [26], [68], but has yet to be tested in marine reptiles such as turtles.(3)

Foraging Region Assignment

To model the difference in isotope values between the two foraging grounds, we conducted a Bayesian mixture-model analysis [69, pp. 220–223]. We assumed that leatherback turtles from the nesting beach had come from one of the two foraging regions that have different isotopic characteristics, and that isotope values in turtle tissues could be modeled with normal distributions using different means for each foraging region. Turtles were treated as independent samples. We looked at two basic models for the underlying isotopic distributions for the two foraging grounds: (1) two means and a common variance and (2) two means and two variances.

Isotope samples from turtles with telemetry data were treated in two ways: (1) without using the known foraging destinations from satellite telemetry results, and (2) using the foraging ground destinations. The proportions of turtles that use the two foraging grounds were estimated as either time independent or dependent. When they are treated as time dependent, these annual proportions were treated as independent of each other.

With the treatment of means and variances, telemetry data, and proportions of foraging grounds, a total of six models were fitted to the data (Table 3). Appropriateness of models was examined using posterior simulation, in which the observed data were compared to observable (or simulated) data [70]. We simulated a dataset of the same size for each set of parameters from the joint posterior distribution. Both the observed and each simulated dataset were sorted from the smallest to largest values and the sum of the differences between the datasets computed. The distribution of sums of differences was used to determine the model fit. OpenBugs (version 3.2.1) [71] was used to obtain the joint posterior distribution of the parameters through Markov chain Monte Carlo. Four independent chains of 105,000 steps with randomized starting points were used to sample from the joint posterior distribution. The first 5000 steps were used to tune the sampling algorithm and discarded for computing summary statistics. For the remaining 100,000 samples of each chain, every five steps were retained to reduce the serial autocorrelations. Posterior samples of 80,000, therefore, were used to compute the summary statistics for the means and variance of normal distributions of the stable isotope values for the two foraging grounds, the proportion of turtles that belong to each foraging ground, and the assignment of each turtle to a foraging ground with associated probability. For all parameters, we used vague prior distributions within appropriate parameter spaces. To make an inference about the proportion of turtles that forage in the eastern or western Pacific Ocean, we arbitrarily defined the probability 0.8 as the cut-off point.

Supporting Information

Table S1.

Summary of satellite transmitter deployments and body size information for leatherback turtles nesting at Jamursba-Medi, Papua Barat, Indonesia.

https://doi.org/10.1371/journal.pone.0037403.s001

(DOC)

Table S2.

Bulk δ15N and δ13C and compound specific δ15N (CSIA-AA) values for glutamic acid (Glu) and phenylalanine (Phe) from leatherback turtle skin samples collected in Jambursba Medi, Indonesia.

https://doi.org/10.1371/journal.pone.0037403.s002

(DOC)

Acknowledgments

We thank L. Hess, E. LaCasella G. Lemons, R. LeRoux, B. Lyon, D. Prosperi, and G. Serra-Valente for assistance; the Light Stable Isotope Lab (University of Florida) the Department of Geology and Geophysics Stable Isotope Biogeochemistry Lab (University of Hawai’i), and J. Curtis and E. Gier for assistance with stable isotope analyses; and J.W. Hayes, T.T. Jones, W. Perrin, M. Tiwari, K. Van Houtan, and B.P. Wallace for comments on the manuscript. All skin tissue samples collected in 2005 and 2006 were exported from Indonesia under CITES permit 16825/IV/SATS-LN/2006 and imported to the USA under CITES import permit 06US844694/9. Samples collected in 2007 were exported under CITES permit 02222/IV/SATS-LN/2008 and imported under 07US844694/9. Samples collected in 2010 were exported under permit 16825/IV/SATS-LN/2006 and imported under 06US844694/9. This is SOEST contribution number 8644.

Author Contributions

Conceived and designed the experiments: KEA SRB JAS BNP. Performed the experiments: KEA SRB PHD JAS. Analyzed the data: KEA SRB TE JAS. Contributed reagents/materials/analysis tools: SRB PHD JAS RFT. Wrote the paper: KEA TE BNP JAS.

References

  1. 1. Block BA, Johsen IK, Jorgensen SJ, Winship AJ, Shaffer SA, et al. (2011) Tracking apex marine predator movements in a dynamic ocean. Letters to Nature 475: 86–90.
  2. 2. Shillinger GL, Palacios DM, Bailey H, Bograd SJ, Swithenbank AM, et al. (2008) Persistent leatherback turtle migrations present opportunities for conservation. PLoS Biol 6: e171.
  3. 3. Musick J, Lutz PL (1997) The Biology of Sea Turtles. London: CRC Press. 347 p.
  4. 4. Hobson KA, Wassenaar LI (2008) Tracking Animal Migration with Stable Isotopes. London: Academic Press. 157 p.
  5. 5. Rubenstein DR, Hobson KA (2004) From birds to butterflies: animal movement patterns and stable isotopes. Trends Ecol Evol 19: 256–263.
  6. 6. Godley BJ, Blumenthal JM, Broderick AC, Coyne MS, Godfrey M, et al. (2008) Satellite tracking of sea turtles: Where have we been and where do we go next? Endang Spec Res 4: 3–22.
  7. 7. Block BA, Teo SLH, Walli A, Boustany A, Stokesbury MJW, et al. (2005) Electronic tagging and population structure of Atlantic bluefin tuna. Nature 434: 1121–1127.
  8. 8. DeNiro MJ, Epstein S (1981) Influence of diet on the distribution of nitrogen isotopes in animals. Geochim Cosmochim Acta 45: 341–351.
  9. 9. Newsome SD, Clementz MR, Koch PL (2010) Using stable isotope biogeochemistry to study marine mammal ecology. Mar Mam Sci 26: 509–572.
  10. 10. Minami H, Ogi H (1997) Determination of migratory dynamics of the sooty shearwater in the Pacific using stable carbon and nitrogen isotope analysis. Mar Ecol Prog Ser 158: 249–256.
  11. 11. Burton RK, Koch PL (1999) Isotopic tracking of foraging and long-distance migration in northeastern Pacific pinnipeds. Oecologia 119: 578–585.
  12. 12. Reich KJ, Bjorndal KA, Bolten AB (2007) The ‘lost years’ of green turtles: using stable isotopes to study cryptic lifestages. Biol Lett 3: 712–714.
  13. 13. Marra PP, Hobson KA, Holms RT (1998) Linking winter and summer events in a migratory bird by using stable-carbon isotopes. Science 282: 1884–1886.
  14. 14. Miller NG, Wassenaar LI, Hobson KA, Norris DR (2011) Monarch butterflies cross the Appalachians from the west to recolonize the east coast of North America. Biol Lett 7: 43–46.
  15. 15. Somes CJ, Schmittner A, Galbraith ED, Lehmann MF, Altabet MZ, et al. (2010) Simulating the global distribution of nitrogen isotopes in the ocean. Global Biogeochem Cycles 24: GB4019.
  16. 16. Deutsch CA, Gruber NP, Key RM, Sarmiento JL, Ganachaud A (2011) Denitrification and N2 fixation in the Pacific Ocean. Global Biogeochem Cycles 15: 483–506.
  17. 17. Faure , G (1986) Principles of Isotope Geology. 2nd Edition. New York: John Wiley and Sons, Ltd.
  18. 18. Boecklen WJ, Yarnes CT, Cook BA, James AC (2011) On the Use of Stable Isotopes in Trophic Ecology. An Rev Ecol Evol Systemat 42: 411–440.
  19. 19. Post DM (2002) Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83: 703–718.
  20. 20. Phillips RA, Bearhop S, McGill RAR, Dawson DA (2009) Stable isotopes reveal individual variation in migration strategies and habitat preferences in a suite of seabirds during the breeding period. Oecologia 160: 795–806.
  21. 21. Graham BS, Koch PL, Newsome SD, McMahon KW, Aurioles D (2010) West JB, Bowen GJ, Dawson TE, Tu KP, editors. In: movement Understanding, pattern , editors. and process on Earth through isotope mapping. New York: Springer. pp. 299–318.
  22. 22. Jaeger A, Lecomte VJ, Weimerskirch H, Richard P, Cherel Y (2010) Seabird satellite tracking validates the use of latitudinal isoscapes to depict predators’ foraging areas in the Southern Ocean. Rapid Commun Mass Spectrom 24: 3456–3460.
  23. 23. Votier SC, Bearhop S, Witt MJ, Inger R, Thompson D, Newton J (2010) Individual responses of seabirds to commercial fisheries revealed using GPS tracking, stable isotopes and vessel monitoring systems J Appl Ecol 47: 487–497.
  24. 24. Popp BN, Graham BS, Olson RJ, Hannides CCS, Lott MJ, et al. (2007) Dawson T. Siegwolf R, editors. Stable Isotopes as Indicators of Ecological Change. New York: Academic Press. pp. 173–190.
  25. 25. McClelland JW, Montoya JP (2002) Trophic relationships and the nitrogen isotopic composition of amino acids in plankton. Ecology 83: 2173–2180.
  26. 26. Chikaraishi Y, Ogawa NO, Kashiyama YTakanoY, Suga H, et al. (2009) Determination of aquatic food-web structure based on compound-specific nitrogen isotopic composition of amino acids. Limnol Oceanogr Methods 7: 740–750.
  27. 27. Evershed RP, Bull ID, Corr LT, Crossman ZM, van Dongen BE, et al. (2007) Compound-specific stable isotope analysis in ecology and paleoecology. In: Michener RH, Lajtha K, editors. Stable Isotopes in Ecology and Environmental Science. Malden, MA: Blackwell Publishing Ltd 480–540:
  28. 28. Graham WM, Pagès F, Hamner WM (2001) A physical context for gelatinous zooplankton aggregations: a review. Hydrobiologia 451: 199–212.
  29. 29. Pitt KA, Connolly RM, Meziane T (2009) Stable isotope and fatty acid tracers in energy and nutrient studies of jellyfish: a review. Hydrobiologia 616: 119–132.
  30. 30. Dodge KL, Logan JM, Lutcavage ME (2011) Foraging ecology of leatherback sea turtles in the Western North Atlantic determined through multi-tissue stable isotope analyses. Mar Biol 158: 2813–2824.
  31. 31. Davenport JH, Balazs GH (1991) ‘Fiery bodies’ – Are pyrosomas an important component of the diet of leatherback turtles? British Herpetol Soc Bull 37: 33–38.
  32. 32. James MC, Herman TB (2001) Feeding of Dermochelys coriacea on Medusae in the Northwest Atlantic. Chelon Conserv Biol 4: 202–205.
  33. 33. Fossette S, Gleiss AC, Casey JP, Lewis AR, Hays GC (2011) Does prey size matter? Novel observations of feeding in the leatherback turtle (Dermochelys coriacea) allow a test of predator −prey size relationships. Biol Lett. https://doi.org/10.1098/rsbl.2011.0965
  34. 34. Reich KJ, Bjorndal KA, Martınez del Rio C (2008) Effects of growth and tissue type on the kinetics of 13C and 15N incorporation in a rapidly growing ectotherm. Oecologia 155: 651–663.
  35. 35. 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 2007: 534–542. pp. 534–542.
  36. 36. Wallace BP, Seminoff JA, Kilham SS, Spotila JR, Dutton PH (2006) Leatherback turtles as oceanographic indicators: Stable isotope analyses reveal a trophic dichotomy between ocean basins. Mar Biol 149: 953–960.
  37. 37. Seminoff JA, Jones TT, Hastings M, Eguchi T, Jones D (2009) Stable carbon and nitrogen isotope discrimination in soft tissues of the leatherback turtle (Dermochelys coriacea): insights for trophic studies of marine turtles. J Exp Mar Biol Ecol 381: 33–41.
  38. 38. Caut S, Guirlet E, Angulo E, Das K, Girondot M (2008) Isotope analysis reveals foraging area dichotomy for Atlantic leatherback turtles. PLoS ONE 3(3): e1845.
  39. 39. Spotila JR, Reina RD, Steyermark AC, Plotkin PT, Paladino FV (2000) Pacific leatherback turtles face extinction. Nature 405: 529–530.
  40. 40. Benson SR, Eguchi T, Foley DG, Forney KA, Bailey H, et al. (2011) Large-scale movements and high-use areas of western pacific leatherback turtles, Dermochelys coriacea. Ecosphere 2: art84.
  41. 41. Sherwood OA, Lehmann MF, Schubert CJ, Scott DB, McCarthy MD (2011) Nutrient regime shift in the western North Atlantic indicated by compound-specific δ15N of deep-sea gorgonian corals. Proc Natl Acad Sci 108: 1011–1015.
  42. 42. Hatase H, Takai N, Matsuzawa Y, Sakamoto W, Omuta K, et al. (2002) Size-related differences in feeding habitat use of adult female loggerhead turtles Caretta caretta around Japan determined by stable isotope analyses and satellite telemetry. Mar Ecol Prog Ser 233: 273–281.
  43. 43. Zbinden JA, Bearhop S, Bradshaw P, Gill B, Margaritoulis D, et al. (2011) Migratory dichotomy and associated phenotypic variation in marine turtles revealed by satellite tracking and stable isotope analysis. Mar Ecol Prog Ser 421: 291–302.
  44. 44. Dale JJ, Wallsgrove NJ, Popp BN, Holland KN (2011) Nursery habitat use and foraging ecology of the brown stingray Dasyatis lata determined from stomach contents, bulk and amino acid stable isotopes. Mar Ecol Prog Ser 433: 221–236.
  45. 45. Fossette S, Hobson VJ, Girard C, Calmettes B, Gaspar P, et al. (2010) Spatio-temporal foraging patterns of a giant zooplanktivore, the leatherback turtle. J Mar Syst 81: 225–234.
  46. 46. Eckert SA, Bagley D, Kubis S, Ehrhart L, Johnson C, et al. (2006) Internesting and postnesting movements and foraging habitats of leatherback sea turtles (Dermochelys coriacea) nesting in Florida. Chelon Conserv Biol 5: 239–248.
  47. 47. James MC, Myers RA, Ottensmeyer CA (2005) Behaviour of leatherback sea turtles, Dermochelys coriacea, during the migratory cycle. Proc Royal Soc B: Biol Sci 272: 1547–1555.
  48. 48. MacKenzie KM, Palmer MR, Moore A, Ibbotson AT, Beaumont WRC, et al. (2011) Locations of marine animals revealed by carbon isotopes. Nature Sci Rep 21: 1–6.
  49. 49. Cherel Y, Hobson KA (2007) Geographical variation in carbon stable isotope signatures of marine predators: a tool to investigate their foraging areas in the Southern Ocean. Mar Ecol Prog Ser 329: 281–287.
  50. 50. Tagliabue A, Bopp L (2008) Towards understanding global variability in ocean carbon-13. Glob Biogeoch Cycles 22: GB1025.
  51. 51. Garrison TS (2009) Essentials of Oceanography. Pacific Grove, California: Brooks Cole Publishers.
  52. 52. Pajuelo M, Bjorndal KA, Alfaro-Shigueto J, Seminoff JA, Mangel J, et al. (2010) Stable isotope dichotomy in loggerhead turtles reveals Pacific-Atlantic oceanographic differences. Mar Ecol Prog Ser 417: 277–285.
  53. 53. Montoya JP (2007) Michener R, Lajtha K, editors. Stable Isotopes in Ecology and Environmental Science. Malden, MA: Blackwell Publishing Ltd 176–201:
  54. 54. Rau GH, Ohman MD, Pierrot-Bults A (2003) Linking nitrogen dynamics to climate variability off central California: a 51-year record based on 15N/14N in CalCOFI zooplankton. Deep Sea Res II 50: 2431–2447.
  55. 55. Voss M, Dippner JW, Montoya JP (2001) Nitrogen isotope patterns in the oxygen-deficient waters of the Eastern Tropical North Pacific Ocean. Deep Sea Res I 148: 1905–1921.
  56. 56. Castro CG, Chavez FP, Collins CA (2001) Role of California Undercurrent in the export of denitrified waters from the eastern tropical North Pacific. Global Biogeochem Cycles 15: 819–830.
  57. 57. Berman-Frank I, Cullen JT, Shaked Y, Sherrell RM, Falkowski PG (2001) Iron availability, cellular ion quotas, and nitrogen fixation in Trichodesmium. Limnol Oceanor 46: 1249–1260.
  58. 58. Minami H, Minagawa M, Ogi H (1995) Changes in stable carbon and nitrogen isotope ratios in sooty and short-tailed shearwaters during their northward migration. Condor 97: 565–574.
  59. 59. Gemperline PJ, Rulifson RA, Paramore L (2002) Multi-way analysis of trace elements in fish otoliths to track migratory patterns. Chemomet Intel Lab Syst 60: 135–146.
  60. 60. Crowder L (2000) Leatherback’s survival will depend on an international effort. Nature 405: 881.
  61. 61. Lewison RL, Crowder LB, Read AJ, Freeman SA (2004) Understanding impacts of fisheries bycatch on marine megafauna. Trends Ecol Evolution 19: 598–604.
  62. 62. Eckert SA, Sarti LA (1997) Distant fisheries implicated in the loss of the world’s largest leatherback population. Mar Turt Newsl 78: 2–7.
  63. 63. Federal Register (2010) Endangered and Threatened Species: Proposed Rule To Revise the Critical Habitat Designation for the Endangered Leatherback Sea Turtle. 75(2): 316.
  64. 64. Hitipeuw C, Dutton PH, Benson SR, Thebu J, Bakarbessy J (2007) Population status and internesting movement of leatherback turtles, Dermochelys coriacea, nesting on the northwest coast of Papua, Indonesia. Chelon Conserv Biol 6: 28–36.
  65. 65. Bolten AB (1999) Techniques for measuring sea turtles. In: Eckert KL, Bjorndal KA, Abreu-Grobois FA, Donnelly M, editors. Research and Management Techniques for the Conservation of Sea Turtles. Gland, Switzerland: IUCN/SSC Marine Turtle Specialist Group Publication No 4. pp. 110–114.
  66. 66. Fossette S, Corbel H, Gaspar P, Le Maho Y, Georges J-Y (2008) An alternative technique for the long-term satellite tracking of leatherback turtles. Endang Spec Res 4: 33–41.
  67. 67. Macko SA, Uhle ME, Engel MH, Andrusevich V (1997) Stable nitrogen isotope analysis of amino acid enantiomers by gas chromatography/combustion/isotope ratio mass spectrometry. Anal Chem 69: 926–929.
  68. 68. Hannides CS, Popp BN, Landry MR, Graham BS (2009) Quantification of zooplankton trophic position in the North Pacific Subtropical Gyre using stable nitrogen isotopes. Limnol Oceanogr 54: 50–61.
  69. 69. Congdon P (2001) Bayesian statistical modeling. West Sussex, England: John Wiley and Sons, Ltd.
  70. 70. Gelman A, Carlin JB, Stern HS, Rubin DB (2004) Bayesian data analysis. 2nd edition. Boca Raton, FL: Chapman and Hall/CRC Press.
  71. 71. Lunn D, Spiegelhalter D, Thomas A, Best N (2009) The BUGS project: Evolution, critique, and future directions. Stat Medicine 28: 3049–3067.