Inferring Foraging Areas of Nesting Loggerhead Turtles Using Satellite Telemetry and Stable Isotopes

In recent years, the use of intrinsic markers such as stable isotopes to link breeding and foraging grounds of migratory species has increased. Nevertheless, several assumptions still must be tested to interpret isotopic patterns found in the marine realm. We used a combination of satellite telemetry and stable isotope analysis to (i) identify key foraging grounds used by female loggerheads nesting in Florida and (ii) examine the relationship between stable isotope ratios and post-nesting migration destinations. We collected tissue samples for stable isotope analysis from 14 females equipped with satellite tags and an additional 57 untracked nesting females. Telemetry identified three post-nesting migratory pathways and associated non-breeding foraging grounds: (1) a seasonal continental shelf-constrained migratory pattern along the northeast U. S. coastline, (2) a non-breeding residency in southern foraging areas and (3) a residency in the waters adjacent to the breeding area. Isotopic variability in both delta C-13 and delta N-15 among individuals allowed identification of three distinct foraging aggregations. We used discriminant function analysis to examine how well delta C-13 and delta N-15 predict female post-nesting migration destination. The discriminant analysis classified correctly the foraging ground used for all but one individual and was used to predict putative feeding areas of untracked turtles. We provide the first documentation that the continental shelf of the Mid-and South Atlantic Bights are prime foraging areas for a large number (61%) of adult female loggerheads from the largest loggerhead nesting population in the western hemisphere and the second largest in the world. Our findings offer insights for future management efforts and suggest that this technique can be used to infer foraging strategies and residence areas in lieu of more expensive satellite telemetry, enabling sample sizes that are more representative at the population level