Multi-event capture–recapture analysis reveals individual foraging specialization in a generalist species

Populations of species typically considered trophic generalists may include specialized individuals consistently feeding on certain resources. Optimal foraging theory states that individuals should feed on those resources most valuable to them. This, however, may vary according to individual differences in detecting or processing resources, different optimization criteria, and competitive abilities. White Storks (Ciconia ciconia) are trophic generalists at the population level. Their European population recovery has been attributed to increased wintering in southern Europe (rather than Africa) where they feed upon new anthropogenic food subsidies: predictable dumps and less predictable and more difficult to detect, but abundant, invasive Procambarus clarkii crayfishes in ricefields. We studied the foraging strategies of resident and wintering storks in southwestern Spain in ricefields and dumps, predicting that more experience in the study area (residents vs. immigrants, old vs. young) would increase ricefield specialization. We developed the first multi-event capture–recapture model to evaluate behavioral consistency, analyzing 3042 observations of 1684 banded storks. There were more specialists among residents (72%) than immigrants (40%). All resident specialists foraged in ricefields, and ricefield use increased with individual age. In contrast, some immigrants specialized on either dumps (24%) or ricefields (16%), but the majority were generalists (60%). Our results provide empirical evidence of high individual foraging consistency within a generalist species and a differential resource selection by individuals of different ages and origins, probably related to their previous experience in the foraging area. Thus, future changes in food resource availability at either of the two anthropogenic subsidies (ricefields or dumps) may differentially impact individuals of different ages and origins making up the wintering population. The use of multi-event capture–recapture modeling has proven useful for studying interindividual variability in behavior.