Adaptive habitat use in size-structured populations: Linking individual behavior to population processes

The inclusion of flexible behavior in population models can be viewed as part of the development of a mechanistic population/community framework based on individual-level concepts. Based on short-term experiments and analyses of the dynamics of nonstructured models, trait-mediated indirect effects resulting from flexible behavior have been suggested to have major effects on population and community dynamics. We use physiologically structured population models (PSPMs) to account for and assess the importance of flexible behavior in population models where individuals differ substantially in size. We investigate a consumer-resource system in which a size-structured consumer feeds on two resources that are present in two habitats differing in predation risk (pelagic open water vs. littoral vegetation). We compare two flexible rules that individuals might use to trade off foraging gains and size-dependent mortality risks in the two habitats. One rule maximizes the ratio of growth over mortality, and the other optimizes the expected instantaneous biomass increase.As a result of strong population feedback, the dynamics and habitat use are qualitatively the same for both rules when open-water mortalities are low, and indeed differ little from a case in which individuals use the two habitats in a fixed proportion. At higher open-water mortalities, the consumers using the two flexible rules show high-amplitude population cycles, whereas consumers using a fixed rule cannot persist. Overall, our results suggest that the behavioral rule used by individuals may be of minor importance, whereas the population feedback is of major importance. In nonstructured models, increased persistence induced by flexible behavior co-occurs with destabilization, as we find here, but the mechanisms behind this pattern are different, driven by cohort cycles rather than by the paradox of enrichment. We argue that extrapolating from experimental data which describe short-term experiments may overemphasize or misrepresent the long-term, population-level effects of flexible behavior. The inclusion of behavior as a proper individual state in physiologically structured population models remains a fundamental issue to solve.