A better understanding of sea turtle spatial ecology is critical for the continued conservation of imperiled sea turtles and their habitats. For resource managers to develop the most effective conservation strategies, it is especially important to examine how turtles use and select for habitats within their developmental foraging grounds. Here, we examine the space use and relative habitat selection of immature green turtles (Chelonia mydas) using acoustic telemetry within the marine protected area, Buck Island Reef National Monument (BIRNM), St. Croix, United States Virgin Islands.
Space use by turtles was concentrated on the southern side of Buck Island, but also extended to the northeast and northwest areas of the island, as indicated by minimum convex polygons (MCPs) and 99%, 95%, and 50% kernel density estimations (KDEs). On average space use for all categories was<3 km2 with mean KDE area overlap ranging from 41.9 to 67.7%. Cumulative monthly MCPs and their proportions to full MCPs began to stabilize 3 to 6 detection months after release, respectively. Resource selection functions (RSFs) were implemented using a generalized linear mixed effects model with turtle ID as the random effect. After model selection, the accuracy of the top model was 77.3% and showed relative habitat selection values were highest at shallow depths, for areas in close proximity to seagrass, and in reef zones for both day and night, and within lagoon zones at night. The top model was also extended to predict across BIRNM at both day and night.
More traditional acoustic telemetry analyses in combination with RSFs provide novel insights into animal space use and relative resource selection. Here, we demonstrated immature green turtles within the BIRNM have small, specific home ranges and core use areas with temporally varying relative selection strengths across habitat types. We conclude the BIRNM marine protected area is providing sufficient protection for immature green turtles, however, habitat protection could be focused in both areas of high space use and in locations where high relative selection values were determined. Ultimately, the methodologies and results presented here may help to design strategies to expand habitat protection for immature green turtles across their greater distribution.