Species‐pair associations, null models, and tests of mechanisms structuring ecological communities

Abstract

Biotic interactions and niche processes are fundamental determinants of community structure and species co‐occurrence. Most studies of species co‐occurrence have focused on negative association patterns (segregation presumed to arise from competition), often ignoring positive aggregations, although positive and negative associations may arise from multiple mechanisms. We used a pairwise approach to identify co‐occurrence patterns of 76 fish species across almost 9500 lakes, followed by a meta‐analytic approach to compare the co‐occurrence pattern of each species pair across watersheds and determine their cumulative species associations. Biological information relating to species’ phylogeny, habitat preferences, and diet was used to group species into relevant subsets in order to test community assembly processes of competition, predation, and habitat filtering. We found consistent non‐random patterns of co‐occurrence in nearly half the species pairs and more extremely aggregated than segregated species pairs. Observed co‐occurrence patterns indicated the importance of shared habitat requirements and predation, rather than competition, driving positively aggregated and negatively segregated species associations, respectively. Our meta‐analytic approach incorporating important biological attributes permitted the testing of specific mechanisms of community assembly, providing novel insights into the major determinants of fish community structure and their generality across a vast set of lakes.