ESR 25:151-164 (2014)  -  DOI: https://doi.org/10.3354/esr00616

ABSTRACT: Incomplete knowledge about local foraging ecology of green turtles hampers their conservation management in Japan, where stocks have only partially recovered from heavy exploitation in previous centuries. We used stable isotope ratios of δ¹³C and δ¹⁵N for turtle carapace scutes, where successive layers contain a chronological record of diet assimilated over a period of years. Turtles were sampled at 2 geographically separate foraging grounds in Japan: the temperate Main Islands (n = 32) and the sub-tropical Nansei Islands (n = 42). Site fidelity was inferred for the majority of turtles at each site (81 and 64% resident turtles) because isotope data indicated diets consistent with food taxa at the respective sites. Immigrant turtles (previous diet outside their current site) were few (n = 4) at the Main Islands site but numerous (n = 14) at the Nansei Islands site, where they were significantly smaller than residents. An ontogenetic shift (Main Islands to Nansei Islands) was inferred for many of the immigrants on the basis of isotope evidence and body size. These immigrants corresponded to a size cohort that was relatively scarce in Main Islands foraging grounds according to previous studies. Bayesian mixing models, used to estimate proportional components of diet, showed varying degrees of imbalance between seagrass and algae and indicated that hypothetical consumption of non-trivial amounts of animal matter was plausible. The latter represented a hypothetical diet component for study turtles since animal matter was rarely found in stomach contents. Potential ambiguity and other issues that constrained inference from mixing models are discussed.

KEY WORDS: Green Turtle · Foraging site fidelity · Ontogenetic shift · Diet composition · Stable isotope analysis · δ¹³C · δ¹⁵N · Chelonia mydas