ISTS42 Program/Agenda

Seagrasses are foundation species for highly productive ecosystems that support fisheries vital to global food security. Green turtles (Chelonia mydas) are specialized grazers that maintain grazed seagrass areas by repeatedly cropping nutritious new growth. Effects of green turtle grazing were largely absent from Caribbean seagrass ecosystems prior to green turtle population increases in recent decades. As population recovery continues, seagrass meadows are returning to more natural levels of grazing. Because canopy height is substantially reduced, the habitat value of grazed areas has been compared to unvegetated substrate, with attendant predictions about loss of fisheries production. Though seagrasses and associated fauna have coevolved with green turtles, green turtle culling has been proposed as a management approach to protect globally declining seagrasses. As such, improved understanding of how green turtles affect seagrass ecosystem function is critical to inform management.

Using remote underwater video (RUV) systems, we investigated effects of green turtle grazing on fish abundance and diversity at two seagrass meadows on the east coast of Eleuthera, The Bahamas. As part of a concurrent study, we mapped the meadows using ground-truthed aerial imagery. Green turtles at both sites maintained mosaics of ungrazed and grazed seagrass, with edges at the interface of these structurally distinct patches. We generated RUV deployment points in a stratified random design with five replicates per microhabitat (bare sand, ungrazed seagrass, edge, grazed seagrass) per site. At each point, we conducted two 4-hour RUV deployments during daytime high tides in June–July 2018. From 283 hours of 4K footage, we extracted relative abundance (MaxN) as maximum number of each fish species observed in a single frame and species richness (S) as number of unique species observed. To evaluate variation in fish abundance and diversity among microhabitats, we (1) summarized microhabitat structural characteristics (PCA of primary producer and grazing intensity metrics); (2) constructed accumulation curves of MaxN and S for each site x microhabitat (vegan R package); (3) evaluated site and microhabitat as predictors of MaxN, S, and Simpson’s Diversity Index (PERMANOVA); and (4) compared fish assemblage composition among microhabitats (NMDS ordination; PERMANOVA), both for taxa and functional entities assigned based on body size and diet (rfishbase R package).

Habitat selection by fish is the product of a food-risk tradeoff—maximizing foraging success and minimizing predation risk—mediated by habitat structure. Given grazing does not result in uniformly simplified seagrass structure, structural characteristics of grazed areas are distinct from unvegetated substrate, and edge represents heterogeneous habitat that may optimize the food-risk tradeoff, we predict fish abundance and diversity will be highest at edge, intermediate in grazed and ungrazed seagrass, and lowest in bare sand. We predict fish assemblage composition at edge will include dominant taxa from both grazed and ungrazed seagrass, and composition will differ between grazed seagrass and bare sand. Our study provides new insight into effects of green turtle grazing on habitat value for seagrass-associated species and contributes to an improved understanding of how seagrass ecosystems functioned before overexploitation by humans led to green turtle ecological extinction.

Characterising diet is important for elucidating trophic webs and understanding ecological roles. Changing environmental conditions are influencing trophic webs and foraging behaviours in the oceans, necessitating technologies that can effectively and rapidly monitor diet. In the sea turtle, whose diet is now recognised to be more variable than once thought, it is important to understand and optimise diet analysis among species, life stages or geographies with varying diet complexities and environmental conditions.

Sea turtle diet analysis has traditionally been conducted via observational studies and gastrointestinal tract analysis, with stable isotope approaches becoming frequent in the last two decades. While these methods are still valuable, molecular approaches are allowing researchers to overcome some of their associated limitations (e.g., morphological stomach content analysis can underrepresent gelatinous prey). DNA metabarcoding is a particularly promising tool with the potential for facilitating wide-scale, non-invasive diet studies in sea turtles.

To assess the potential of dietary DNA metabarcoding analysis in sea turtles, we conducted a systematic review of DNA-based diet studies on marine vertebrates. Overall, studies suggest that DNA metabarcoding can improve the taxonomic resolution of prey species identified, particularly in combination with traditional methods, and its incorporation in diet analysis can lead to the identification of a greater number of prey taxa. However, there were consistent biases and considerations to be made for optimising results and designing future studies (e.g., primer selection, presence of host DNA, inability to distinguish secondary/incidental prey). Dietary DNA metabarcoding studies in sea turtles are limited (n=3 since 2021), but we expect further studies could reveal underrepresented prey species in sea turtles, particularly in terms of harder-to-study diets (e.g., consumption of gelatinous species) or life stages (e.g., immature individuals from remote locations).

This review calls for the scaling up of sea turtle dietary DNA metabarcoding studies and the optimisation of methodology. Multi-faceted, comprehensive approaches to dietary analysis will help to characterise variations in sea turtle diet and effectively monitor changing trophic ecology in response to environmental changes such as rising sea temperatures and displacement to alternative foraging grounds.

The invasion of Halophila stipulacea across Caribbean seagrass meadows has caused concern amongst management agencies due to its high resilience to disturbance, including costly removal strategies. Investigating how large megaherbivores like green sea turtles influence seagrass ecosystem dynamics can help these agencies predict how the invasive seagrass will progress or decline over time. Additionally, the drastic change in green turtle foraging habitat, from Syringodium filiforme and Halodule wrightii-dominated beds to dense, monotypic H. stipulacea, could impact their future habitat use and resource partitioning, information that conservation and management agencies use to implement protective guidelines for this endangered species. This project, part a subsection of a larger Master’s thesis, encompasses a fine-scale tracking study of green turtles’ movement patterns in Brewers Bay, St. Thomas to investigate their foraging selectivity in the mixed-species seagrass beds. A fine-scale positioning system (FPS) acoustic receiver array was deployed across the ~1.5 km² of the bay, which includes seagrass, coral reef, and sand/rock benthic habitat. 17 juvenile green sea turtles were tracked with acoustic transmitters that provided positions with an estimated accuracy of ± 2 meters. The native and invasive seagrass composition was mapped in the highest trafficked daytime area to pair with the turtles’ foraging locations. Turtle movements were linked to seagrass composition within the sampling grid using resource selection functions (RSF) to estimate turtle selection towards each seagrass species in Brewers Bay. Turtles actively selected the two native species present, with no selection towards the invasive seagrass despite its high abundance. This coincides with findings from similar studies suggesting that green turtles are foraging preferentially in native grasses, allowing H. stipulacea to thrive without top-down pressure. Interestingly, three individuals utilized foraging areas in deeper water with monotypic invasive seagrass, which was outside the sampling grid and not suitable for analysis in this study. This pattern of space use has not been observed in past studies of Brewers Bay and could be evidence of turtles beginning to recognize H. stipulacea as a viable food source. These results highlight the need to better understand H. stipulacea as a dynamic factor in green turtle foraging, as developing foraging pressure on the invasive seagrass may help combat its continued spread.

The social behaviors of marine turtles outside of courtship and reproduction are an emerging topic of study in marine turtle biology. Nevertheless, the underlying factors driving variation in social behaviors remain relatively unknown. This study aimed to determine which factors influence social interactions between resident juvenile green turtles (Chelonia mydas) in Brewers Bay, St. Thomas, U.S. Virgin Islands. Specifically, we investigated relationships between the type of interaction, the habitat in which it occurred, and the relative body size of the turtles. We collected videos of juvenile green turtles interacting through opportunistic underwater snorkeling surveys via handheld cameras. Snorkelers followed and recorded turtles in rocky substrate and seagrass meadows during different times of day to record any social interactions. We expanded upon interaction ethograms from past behavioral studies, categorizing interactions as aggressive or passive and then further grouping behaviors into specific types of turtle contact and association. Aggressive interactions can best be understood as dominant behaviors like physical altercations, while passive interactions are understood to be interactions without a display of dominance or possibly acting with curiosity. Juvenile green turtles in Brewers Bay display a range of social behaviors, largely mirroring those documented in past behavioral studies. These included aggressive behaviors such as biting, contact, and displacement and passive behaviors such as inspection and head touching. We discuss behavioral patterns in relation to previous studies on green turtles as well as loggerheads (Caretta caretta) and hawksbills (Eretmochelys imbricata). Our study adds to the understanding of marine turtle sociality by providing evidence of environmental variables influencing potentially complex social behaviors that are contrary to a widely accepted paradigm that reptiles are generally nonsocial.

The increasing use of animal-borne bio-loggers has revolutionized the study of marine megafauna, yet there remains a distinct paucity of research concerning the behavioral and physiological impacts of handling stress and bio-logger attachment. Here, we used 55 animal-borne cameras (referred to as TurtleCam to assess short-term (up to 3 h) changes in the behavior and breathing rate of juvenile green turtles in The Bahamas immediately following capture and bio-logger deployment. The turtles were pursued by boat in shallow waters and snorkelers grabbed them at the base of the front flippers and brought them onboard the boat. Onboard, the turtle was tagged if needed, the carapace size was measured, and the TurtleCam was attached within 15-30 min. A VHF radio transmitter (MOD-050-2, Lotek, USA) and a buoyant foam were attached to a dive camera (DiveCamera+, Paralenz, Denmark), which it was deployed by using 5-min epoxy (KwikWeld, USA) to affix three pieces of 4x4 cm plastic mesh with a galvanic timed releases (AA2, International Fishing Limited, New Zealand) attached. The animal was later released back into the water within 100 m of its original encounter location. Equivalent data were collected from non-handled animals via Unoccupied Aerial Vehicles (UAVs; model DJI Mavic 2 Pro) by flying the unit at an altitude of 15 m, in the same general creek systems and at similar times as TurtleCam surveys. UAVs would record for a minimum of 10 min with the turtle within the center of the image. This was repeated up to three times per day per creek, avoiding using the exact same creek on the exact same day as TurtleCam surveys, to ensure that no filmed individuals were handled by us in the past 24 hours. Finally, we investigated the effect of body-size by comparing the responses of these 55 turtles above and below 500 mm Straight Carapace Length (SCL), where 28 were below 500 mm SCL and 27 were above 500 mm SCL. The animal-borne camera footage revealed that after release turtles spent a high proportion of their time swimming (70 to 80 %), but this showed a continual decrease over time, alongside an increase in resting and feeding behavior until plateauing 90 mins post-release. While this may suggest that turtles have resumed typical behaviors within 90 mins of handling, the mean apnea durations (used as an indicator of dive capacity) increased steadily until the end of the 180 min sampling period. Furthermore, mean apnea durations increased faster for larger turtles even though UAV data suggests similar mean apnea duration under normal conditions. In conclusion, while the effects of handling stress on the behavior of juvenile green turtles may largely diminish within 90 mins, we propose that other physiological stressors may still be affecting metabolic rate and dive capacity for several hours longer.

STORM (Sea Turtle for Ocean Research and Monitoring) is an international, multidisciplinary, telemetry-based research program to study the five species of sea turtle living in the west tropical Indian Ocean, while also collecting key in-situ observations of ocean properties. With over 110 sea turtles (juveniles, nesting females, males) equipped with Argos and GPS environmental tags (depth, temperature, salinity, fluorescence) from January 2019 to January 2023 (including 75 tags released in the Mozambique Channel from January to March 2021 during the 2021 nesting season), the STORM dataset is probably one of the most remarkable of its kind ever obtained in this part of the world.

The STORM program is centered around three interconnected and highly complementary components aiming at: i) studying the ecology of the 5 species of sea turtle inhabiting the South-Western Indian Ocean (SWIO) region, with emphasis on their interactions with the oceanic environment, ii) studying the surface and in-depth physical properties of the tropical Indian Ocean, particularly in the vicinity of tropical cyclones, and iii) implementing educational and communications actions aimed at stepping up the science-society dialogue on the subject of biodiversity conservation.

After a brief overview of STORM's main objectives, we will present some results of the research work carried out to date on sea turtle ecology (ecological connectivity, migration corridors, feeding areas, navigation processes), based on the analysis of tracking data collected by individuals equipped from the French islands of Europa, Tromelin and Réunion (loggerhead, green and olive ridley turtles), iSimangaliso Wetland World Heritage Site (South Africa; loggerhead and leatherback turtles), Moheli National Park (Comoros; green and hawksbill turtles) and Aldabra World Heritage Site (Seychelles; green turtles). Finally, as the sea turtle nesting season in the SWIO corresponds to that of the tropical cyclone season (November - April), we will also briefly assess the behavior of animals trapped in the immediate vicinity of tropical cyclones encountered during their migration in the open sea.

When hatchling loggerhead turtles (Caretta caretta) from Melbourne Beach, FL, U.S.A. leave their nesting beaches, they embark on a multi-year pelagic migration, exploiting a series of regional geomagnetic fields as navigational markers to help them remain within the North Atlantic Subtropical Gyre (NASG). While the geomagnetic field provides ubiquitous spatial cues across earth’s surface, it is not fixed in time; as molten iron in earth’s core moves, geomagnetic signatures gradually drift. Thus, to use inherited regional geomagnetic cues in navigation, the loggerhead population presumably must update responses to regional geomagnetic fields that exist along their migratory route. Here, we investigate the timescale of these updates in orientation responses.

To test innate responses to geomagnetic fields, hatchling turtles were allowed to swim in a magnetic coil system on the beach in Melbourne FL, where they are magnetically displaced to a location that exists off the coast of Puerto Rico. In a previous orientation experiment conducted in 2007, hatchlings from the 2007 cohort responded to a magnetic field that existed near Puerto Rico by swimming in a northeasterly direction (n=22), a response that aligns with their migratory route around the NASG. In this experiment, we test how the 2023 cohort of hatchlings respond to the magnetic field that 1) currently exists near Puerto Rico in 2023 (n=49), 2) previously existed near Puerto Rico in 2007 (n=23), and 3) previously existed near Puerto Rico in 1983 (n=20). Hatchlings from the 2023 cohort responded to the 2023 magnetic field in Puerto Rico by swimming in a northeasterly direction. By contrast, the 2023 cohort of hatchlings responded to the 2007 magnetic field by orienting strongly in the northwesterly direction, a response that is significantly different from both the 2023 response and the original 2007 response. This finding suggests that hatchling loggerheads can update their orientation responses to geomagnetic fields and can do so in under one generation time. The 2023 cohort responded to the 1983 magnetic field by orienting Northeast, which is also congruent with the present location of the drifted 1983 magnetic field. Because the population updates their response within the 16 years between the 2007 and 2023 experiments, it appears that hatchling loggerheads must rely at least in part on environmental influences to set or alter their geomagnetic instructions. Future behavioral work will continue to investigate the evolutionary mechanisms that underlie the ability to quickly update magnetic instructions.