Factors Influencing Stream Fish Species Composition and Functional Properties at Multiple Spatial Scales in the Sand Hills of the Southeastern United States

Abstract: An understanding of how fish communities differ among river basin, watershed, and stream reach spatial scales and the factors that influence these differences can help in the design of effective conservation programs and the development of reference models that appropriately represent biota under relatively undisturbed conditions. We assessed the heterogeneity among fish assemblages in first- to fourth-order stream sites from four river basins (Savannah, Chattahoochee, Cape Fear, and Pee Dee rivers) within the Sand Hills ecoregion of the southeastern USA and compared it with the heterogeneity associated with watershed and stream reach spatial scales. Fifty-five species of fish representing 15 families were collected by electrofishing, with the most speciose families being Cyprinidae, Centrarchidae, Percidae, Ictaluridae, and Catostomidae. Constrained ordination identified clearly demarcated species assemblages among river basins as well as subbasin environmental variables that affected fish species composition; the amounts of variance attributable to basin, watershed, and stream reach spatial scales were roughly equivalent. Prominent differences occurred between Gulf of Mexico coast and Atlantic coast river basins, but differences among Atlantic coast basins were also apparent. Key variables at the watershed scale included watershed size, relief, extent of anthropogenic disturbance, and forest cover; key variables at the stream reach scale included instream habitat quality, proximity to a larger stream, and stream width. Fish assemblage collective and functional properties were more strongly influenced by variables acting at watershed and stream reach spatial scales than by differences among basins. Species richness peaked at intermediate levels of habitat quality as a likely result of biotic homogenization, indicating that the least disturbed sites within the region do not necessarily possess the highest species richness. Failure to consider this may lead to the overrating of moderately disturbed sites and the underrating of minimally disturbed sites, thus contributing to false conclusions about fish assemblage integrity. Received September 15, 2015; accepted December 15, 2015 Published online April 19, 2016