A test of the abundant‐center hypothesis for stream fishes

Abstract

The abundant‐center hypothesis (ACH) provides a conceptual model for predicting range‐wide distributions of species abundance, suggesting that abundance peaks in the center of the geographic range and declines towards range edges. Empirical studies testing the ACH and its subsequent derivations predominantly occurred in terrestrial systems and reported mixed support. Moreover, none of these models consider the possibility of multiple geographic areas of elevated abundance (which we refer to as ‘abundant cores'). Naturally dispersal limited species may exhibit multiple abundant cores, requiring refinement of the ACH. We used fish species abundances from 29 206 community monitoring surveys and weighted geospatial kernel density estimation to identify the number of abundant cores for 64 freshwater fish species. We regressed the number of abundant cores against range size and body size to test if larger geographic distributions and body sizes contain more abundant cores than smaller distributions and body sizes. The two predictors are surrogates for evolutionary age and dispersal ability, respectively, because older species are generally associated with larger ranges, and large‐bodied fishes have greater dispersal ability than small‐bodied fishes in dendritic networks. For studied species, 43 exhibited multi‐core distributions, and 21 exhibited a single‐core distribution. Species range size, but not body size, was significantly and positively associated with the number of abundant cores. The ACH was not a good descriptor of the abundance patterns of most stream fishes we studied, suggesting that an abundant center model may not be well‐suited for freshwater fishes. Recent geo‐climatic events in evolutionary time have isolated populations of the same species by a matrix of unsuitable habitat and/or hard dispersal barriers, providing the basis for multi‐core distributions. Biogeographic and ecological mechanisms likely underpin observed multi‐core patterns, and our work indicates that the ACH and related concepts still present opportunities for testing and refinement.