Multi‐decadal and ontogenetic trophic shifts inferred from stable isotope ratios of pinniped teeth

Identifying and characterizing top predators’ use of trophic resources provides important information about animal ecology and their response to changing conditions. Information from sources such as stable isotopes can be used to infer changes in resource use as direct observations in the wild are difficult to obtain, particularly in the marine environment. Stable carbon and nitrogen isotope values were recovered from the canine teeth of grey seals collected from haul outs in the central North Sea in the 1970/1980s (n = 44) and 2000s (n = 25), spanning a period of marked ecosystem changes in the region. Extracting material deposited during juvenile and adult life‐stages, we reconstructed a multi‐decadal record of δ 15 N and δ 13 C variation. Using established correlations between stable isotope ratios and sea bottom temperature we created a proxy for baseline isotopic variability to account for this source of temporal change. We found 1) a significant long‐term decline in juvenile grey seal δ 15 N values, suggesting trophic position has decreased over time; 2) a decline in adult δ 15 N values and contraction in stable isotopic niche space after the North Sea regime shift, signifying both a decline in trophic position and change in foraging habits over the 20th century; and 3) evidence for dietary segregation between juvenile and adult animals, showing juvenile individuals feeding at a lower trophic position and in more nearshore areas than adults. Our results demonstrate the efficacy of mining archived biological samples to address ecological questions and imply important ontogenetic and long‐term shifts in the feeding ecology of a top predator. Long‐term changes in grey seal trophic dynamics may be partly in response to well documented ecosystem changes in the North Sea. Such indirect monitoring of marine predators may have utility when set in the context of ecosystem assessments where paucity of long‐term monitoring data is prevalent.