A Multi-Scale Framework for the Analysis of Fish Metacommunities
Pedro R. Peres-Neto and Graeme S. Cumming
Abstract.—One of the central goals of aquatic ecology is to understand how natural and human-induced processes control fish distributions. Species assemblages are regulated by both local and large-scale processes, but most studies have failed to tackle the full range of relevant geographic scales at which fish communities are organized. As a result, we have little understanding of how key variables such as local and regional habitat filters and the dynamics of dispersal, together with local colonization and extinction events, interact to structure freshwater fish metacommunities. Although dispersal in metapopulations can be modeled using a variety of statistical and stochastic patch-occupancy methods, these frameworks are designed for single species and preclude the analysis of metacommunities. We introduce a synthetic quantitative framework that considers local and regional factors, as well as spatial predictors and patch isolation, to estimate the relative importance of different factors in determining the distribution of fish metacommunities. The framework is illustrated using a data set of fish distributions from northern Wisconsin. Our results suggest that isolation is of relatively low importance in structuring fish metacommunities in Wisconsin, with both local and regional habitat playing a more important role. More generally, our framework offers some powerful analytical tools while also highlighting some of the challenges that lie ahead. Questions of phenotypic plasticity and individual dispersal remain difficult to resolve in a largely statistical framework, and the difficulties associated with quantifying microhabitat features over very large landscapes may make it hard to adequately test for local influences. Nonetheless, careful selection and analysis of multiscale, spatially explicit variables and their relationships to community composition can offer valuable insights into the forces that structure metacommunities.