ISTS42 Program/Agenda

Sea turtles are marine ectotherms and have commonly been considered capital breeders who use energy stores accumulated at an earlier time for reproduction because the energetic costs associated with storage and use of body reserves prior to reproduction are lower. Only in several green turtle (Chelonia mydas) populations, there have been reports of feeding by gravid females during reproductive period, indicating they all are not simply capital breeders, but some may use the energy gained concurrently for reproduction. However, it is still unknown whether such feeding is voluntary or opportunistic, how they feed on prey, and how they allocate the time spent on energy intake and energy-saving behaviors. Herein, we deployed videos, head-mounted acceleration, and GPS loggers on nine gravid green turtles nesting on Ishigaki Island, Japan, to monitor their feeding behavior during the reproductive period. Our results clearly demonstrate that gravid turtles are voluntary benthic herbivores with clear diel rhythms twice a day. Most of the benthic feeding occurred around the algae/seagrass meadows and in the vicinity of the resting sites, indicating that gravid turtles need almost no energy to shuttle between the feeding and resting sites. Moreover, they incur little energy for prey searching because of the immobility of algae/seagrass. These advantages in terms of the cost of energy intake may allow them to feed voluntarily. Our results indicate that gravid green turtles employ mixed capital–income breeding strategy that females mostly rely on capital, but partially use the concurrent intake for breeding when abundant food are available.

The western Mediterranean Sea has traditionally served as a primary foraging ground for juvenile loggerhead turtles from three distinct regional management units: the Northwestern Atlantic, the Mediterranean, and Cape Verde. Nesting activities were only sporadic. Recently, nesting activity on the beaches of Spain, France and Italy has increased, due to warmer sand and water temperatures during the summer months. Genetic analysis has revealed an admixture of parents from Mediterranean and Atlantic origins on the nesting beaches of the western Mediterranean, but the foraging grounds used by those adults are unknown. This study has integrated satellite telemetry data from seven adult females nesting in Spain with stable isotope analysis of carbon (C), nitrogen (N), and sulfur (S) in the epidermis of hatchlings from 16 nests found in Spanish beaches and 58 dead individuals stranded along the coasts of two distinct basins (the Balearic Sea and the Algerian Basin), to identify those foraging grounds. Both satellite telemetry and stable isotope analysis indicated that the majority of the adult females nesting in Spain foraged oceanically in the Algerian Basin. Nevertheless, one of the satellite-tagged females spent 3.5 months foraging also in the central Mediterranean Sea and two hatchlings exhibited distinct stable isotope signals, tentatively associated with the foraging of their mothers in shallow coastal areas of the central Mediterranean Sea. In any case, the dominance in this new population of oceanic foragers using the Algerian Basin is noteworthy, given the prevalence of neritic foraging in adult females nesting in the central and eastern Mediterranean Sea and the Algerian Basin’s oligotrophic nature. It should be noted, however, that the Algerian Basin is the foraging ground used by most of the juvenile loggerhead turtles of Atlantic origin reaching the Mediterranean. In any case, the use of the Algerian Basin as a foraging ground for adult loggerhead turtles exposes them to bycatch from drifting longlines. In the 1990s, Spanish drifting longlines captured as many as 20,000 loggerhead turtles annually, mostly at the Algerian Basin, and posing a significant threat to these turtles in the region. Over the past decade, a dramatic reduction in bycatch levels has occurred in the western Mediterranean Sea driven by operational changes in the Spanish longline fleet, which shifted its focus from bluefin tuna to swordfish as their main target species due to new regulations. The authorities should refrain from increasing the bluefin quotas imposed on longliners in the Algerian Basin unless the fishery avoids using squid as bait and agrees to deploy hooks deeper than 25 m to reduce loggerhead turtle bycatch. Otherwise, the ongoing colonization of the beaches in the western Mediterranean Sea would be jeopardized by the incidental capture of these oceanic adults.

Leatherback sea turtles (Dermochelys coriacea) travel thousands of kilometers to temperate feeding habitats off the northeast United States and Atlantic Canada to forage on gelatinous zooplankton. Their seasonal (May – November) presence in coastal Massachusetts waters coincides with several active fisheries and a peak in vessel activity. This overlap results in an increased risk of entanglement and vessel strike, which is supported by leatherback strandings in the region. To assess leatherback behavior relative to these risks, we deployed custom biologger tag packages on 13 free-swimming, feeding leatherbacks and one disentangled leatherback during September and October 2023. Tags were attached to the turtles with suction cups using a carbon fiber pole or by hand from small (<8 m) research vessels. Tagging locations included waters south of Nantucket Island (n=7), Nantucket Sound (n=6), and Cape Cod Bay (n=1). Two of the tag packages detached prematurely (<10 min) and were collected and redeployed on different turtles on the same day. Of the remaining 12, we successfully recovered 11 tag packages using satellite and VHF telemetry while one tag package could not be located due to satellite tag failure. We collected 196 hours of archival data with 7 overnight deployments, and individual attachment durations ranged from 1.8 to 29.8 hours (mean ± SD: 17.6 ± 9.6). On average, turtles spent just over 20% of their time at the surface, with percent surface time higher south of Nantucket (mean ± SD: 24.9 ± 2.8) and in Cape Cod Bay (23.0) compared to Nantucket Sound (mean ± SD: 17.2 ± 2.4), and surface bouts were brief across all individuals (mean 1.7 min). Turtle dive durations were also brief (mean ± SD: 6.2 ± 2.0 min) and average dive depths were relatively shallow (<10 m), with the exception of the disentangled turtle that had an average dive depth of 19.2 m. Preliminary assessment of sensor data showed that foraging turtles had lower nocturnal activity levels (measured as the turtle’s overall dynamic body acceleration (ODBA)) compared to daytime hours, but the disentangled turtle’s activity level was consistent day and night. The disentangled turtle also dove deeper than the foraging turtles, but this may be due to study site bathymetry, with Cape Cod Bay having deeper water access compared to Nantucket Sound and the waters just south of Nantucket. Further analysis of the accelerometer data and, for some tags, corresponding video, will provide additional insight into leatherback foraging behaviors that may increase their risk to human activities. Future field work will focus on increasing data collection from disentangled turtles.

Adult and sub-adult western Pacific leatherback turtles (Dermochelys coriacea) forage in diverse ecoregions of the Indo-Pacific region, including temperate nearshore waters off central California, USA. The greatest leatherback densities in this region are found in oceanographic retention areas between about Bodega Head (38.3°N) and Monterey Bay (36.6°N). Aerial surveys have been conducted regularly since 1990 to assess the abundance, distribution, and habitat of leatherback turtles that forage within shelf waters during summer and fall. These surveys have documented an 80% decline in the abundance of leatherbacks since 1990, corroborating a similar decline of adult female leatherbacks at Birds Head Peninsula, Indonesia, and revealed interannual variability in the availability of foraging habitat and leatherback abundance. In this region, leatherbacks are vulnerable to vessel strikes and entanglement in fisheries. Species distribution models have successfully been used to assess the risk of entanglement or ship strikes to whales, and to assess potential interactions between leatherback turtles and offshore fisheries throughout the California Current Ecosystem and central North Pacific. However, the resolution of the leatherback models (10-25 km) is too coarse for management in the small nearshore foraging area off central California.

To support management and conservation of leatherback turtles within this key foraging ground, we have developed a dynamic, fine-scale (1-km) spatial density model for leatherback turtles from a subset of the aerial survey data (2002-2022). The surveys were conducted using standard line-transect methods, at 170-185 km hr^-1 airspeed and about 200 m altitude, in twin engine, high-wing aircraft outfitted with bubble windows and a belly port. Potential habitat covariates that were available at a sufficiently fine-scale resolution included bathymetric depth (ETOPO-1, 1 arc-minute), sea surface temperature (Multi-scale Ultra-high Resolution Sea Surface Temperature (SST) at a daily 1-km resolution), and standard deviation of SST within 9×9 km and 25×25 km regions (proxy for frontal regions and mesoscale features). Year was included to account for the previously documented population decline. Models were developed using well-established methods within a generalized additive modeling framework that estimates animal density as a function of habitat covariates within a line-transect framework. Corrections for detection parameters including effective strip half-width, ESW, and the probability of detecting a leatherback on the transect line, g(0) were included in the model to provide absolute estimates of animal density and abundance.

The selected model included four significant variables: bathymetric depth (p<0.001), SST (p=0.002), the standard deviation of SST at a 25×25 km scale (p<0.001), and year (p<0.001). Depth and SST exhibited unimodal relationships, with density peaks at 40-70 m and 14-16°C, respectively. The standard deviation of SST was positively correlated with density. Modeled spatial density maps for the peak foraging season (September – October) successfully captured interannual variability in the location and extent of high-density regions, as was observed during the 2002-2022 aerial surveys. Thus, these dynamic models can help assess leatherback presence and distribution while foraging off central CA, providing a new tool for managing risks to leatherbacks and optimizing future survey effort in areas most likely to support leatherbacks.

Habitat complexity and productivity are two key elements that dictate species richness and resilience in coastal ecosystems. Protected green turtles (Chelonia mydas) form large foraging aggregations in coastal southern California (USA) ecosystems that are often fueled by seagrass (Zostera marina) and macro-algae. Seagrass ecosystems in southern California and around the world are adversely impacted by anthropogenic threats such as habitat destruction and contamination, which may impair the ecological functions of these diverse ecosystems. Seagrasses synthesize essential amino acids (AA_ESS) differently than co-occuring micro and macroalgae, which creates unique AA_ESS d¹³C profiles or ‘fingerprints’. These ‘fingerprints’ avoid isotopic alteration as they move through food webs. Many animals including green turtles cannot synthesize AA_ESS de novo, requiring that they obtain AA_ESS from their diet. Thus, AA_ESS d¹³C profiles of green turtles can elucidate the importance of eelgrass and marine algae in supporting nutrient flow in these systems. From 2018–2019 we analyzed skin from 24 green turtles residing in two established southern California foraging habitats; San Diego Bay (SDB; n=14) and Seal Beach National Wildlife Refuge (SBNWR; n=10). We conducted d¹³C analysis of individual AA_ESS (Ile, Leu, Phe, Thr, Val) on green turtle skin and potential food sources (n= 25) including, seagrasses and macro-algae. Linear discriminant analysis (LDA) was applied to green turtle skin and habitat AA_ESS d¹³C profiles to model primary production influence in mediating nutrient flow in green turtle trophic interactions. In SDB, LDA indicates 79% (n= 11) of sampled green turtles obtain the majority of their AA_ESS from seagrass-based trophic pathways and 21% obtain AA_ESS from macro-algae-based trophic pathways. In SBNWR, most green turtles (60%) obtained the majority of their AA_ESS from macro-algae-based food chains with the other 40% utilizing seagrass-based pathways. We found that green turtle size and age class influenced their tendency to feed in eelgrass vs macro-algae trophic pathways. For example, 67% of all turtles assigned to seagrass pathways were sexually mature adults. Moreover, every turtle larger than 69.0 cm straight carapace length had an assignment to seagrass nutrient pathways. These findings underscore the value of green turtles as sentinels of trophic pathways and ecosystem structure, and suggest that as green turtles grow and mature in southern California, they increasingly rely on seagrass nutrient pathways. At a time when green turtle numbers in southern California are increasing, their strong reliance on seagrass ecosystems highlights the value of this habitat resource for endangered species. Thus, stemming the ongoing degradation of this habitat type throughout this region should be a conservation priority.

In the eastern Pacific, along the coast of the United States and Mexico, green sea turtles (Chelonia mydas) are now commonly found foraging in bays, lagoons, and other interesting inlets as the population continues its impressive recovery. Once at extremely low numbers in the 1980s, the green turtles from this East Pacific population found in northern foraging grounds are reflecting a shift in demographics. There appears to be an increasing presence of smaller and younger turtles in both known, established foraging habitats – such as San Diego Bay and Seal Beach National Wildlife Refuge – and also in new and sometimes unexpected habitats - such as Mission Bay and the San Gabriel River.

As a part of ongoing research, we have continued quantifying demographic parameters and variable vital rates, such as somatic growth rates, time to maturity, and reproductive longevity, to inform effective management of this population. Building on this work, we have now characterized size- and age-distributions in distinct habitats – oceanic vs. neritic – by using a combination of methods including skeletochronology, sequential stable isotope analysis, in-water mark-recapture, and community scientist sightings.

Previous age analysis (by skeletochronology) of 65 green turtles recovered along the U.S. west coast showed that, based on observed minimum sizes of turtles in nearshore foraging grounds of ca. 50 cm curved carapace length (CCL), green turtles in the eastern Pacific spend approximately 5 years in the oceanic zone (a.k.a. the ‘lost years’) before recruiting to nearshore habitats. Here, we extended upon this study by applying stable isotope analysis (SIA) of bone growth layers to estimate age- and size-at-settlement to neritic habitats based on a change in chemistry (distinct stable nitrogen isotope values) in growth layers of humerus bones.

A subset of the 65 aged humeri were sampled for SIA, and of the 33 turtle bones that recorded an ontogenetic shift (21 shifted + 12 recent recruits), the mean ± SD age-at-settlement was 7.4 ± 6.1 yrs (range: 1-33 yrs) and size-at-settlement was 52.6 ± 12.8 cm CCL (range: 28-98 cm CCL). We also evaluated long-term habitat use and foraging patterns based on the SIA results (stable nitrogen and carbon isotope values), and confirmed that once settled to nearshore foraging habitats, turtles remained largely consistent in what and where they ate. Most turtles showed specialization in their diet, yet some exhibited variability over time.

Finally, we present a few detailed case studies of several individual turtles by combining lab, field, and community science data to better understand the life history and behavior of green turtles in Southern California. Collectively, the results present a more in-depth look at the life history and habitat use patterns of green turtles that are now more commonly encountered in Southern California waterways.

Algal-dominated reefs of Liuchiu Island is regarded as the primary foraging habitat of hundreds of green turtles in Taiwan. The island provides easy access to these turtles via snorkeling or SCUBA diving, which has attracted many domestic and international tourists. It has also significantly raised awareness of conservation concerns. To better understand the abundance of sea turtles around the island and how they utilize this habitat, we conducted intensive seasonal surveys at 2 hotspots over a 2-year period. We captured images of turtles during the survey, and then used a photo-ID method to identify individuals and track their occurrence rates. In total, we accumulated around 4260 sightings and identified more than 500 unique individuals. 70% of individuals had a high repetitive sighting rate as well as high fidelity to their foraging habitat. Shifting of individuals among foraging sites has been observed but rare. The number of individuals recorded with our intensive systematic survey method was three times higher than a single time survey. We believe these findings are important towards improving our knowledge on green sea turtle population dynamics and habitat fidelity, as well as our ecological understanding of these precious animals. Further, the methods applied in this study have potential in other fields concerning habitat monitoring and community-based projects with limited resources.