Considering dispersal costs to understand fish community dissimilarity in a high-gradient basin of North America

Freshwater fish metacommunities are best understood when considering the dendritic structure of riverine networks. The dendritic structure imposes restrictions to dispersal associated with the connectivity. Many structures restrict the movement of fish even more, such as dams and the terrain slope (dispersal costs). We investigated the influence of environmental predictors and dispersal costs on the beta diversity of freshwater fishes from the Upper Tennessee River using Generalised Dissimilarity Modelling (GDM). In addition, we tested the effects of asymmetrical dispersal costs (high costs for upstream dispersal) on the nestedness of native and non-native fishes. Environmental predictors were more important than dispersal costs for explaining the overall fish dissimilarity in the GDM models, with the turnover capturing most of the explanation compared to the species richness difference. Dam heights were the most important dispersal cost variable in the GDM, mainly for species turnover. Overall dissimilarity of the native fish subset was better explained by environmental and dispersal variables than non-native fish (20.03% vs. 8.41%). Considering the native species assemblage subset, dispersal cost related to maximum channel slope between pairs of sites and watercourse distance increased the nestedness of adjacent upstream–downstream sites, whereas those attributed to dams reduced the nestedness. The results support that dams increase overall dissimilarity in the native fish metacommunity of the Upper Tennessee basin, but also reduce the nestedness of adjacent sites. Our findings improve the knowledge on how mechanisms and processes associated with dispersal costs in watersheds under the effects of dams create patterns of dissimilarity and nestedness.