COS 119-1 - Homogenization and differentiation of stream fish communities in three dimensions of diversity

Abstract: Biotic homogenization is the reshuffling of communities due to nonnative introductions and the extirpation of unique species, leading to decreased beta diversity. Changes in beta diversity can provide evidence for changes in community assembly processes. More information on community assembly processes can be gained by decomposing beta diversity into biodiversity’s three components, taxonomic, functional, and phylogenetic diversity and examining relationships between these dimensions. North American stream fishes represent an ideal system to study community assembly processes due to a high degree of habitat heterogeneity, biodiversity, and phylogenetic diversity. In addition, stream fishes are confined to stream networks, which provide historic and anthropogenic dispersal barriers. This study sought to (1) examine homogenization/differentiation of stream fishes within three dimensions of beta diversity and to (2) determine if key beta diversity dimension relationships have changed over time to infer changes in community assembly processes. Modern fish community occurrence data was acquired from a continental scale stream fish community dataset for the conterminous United States. Historic fish communities were created using the NatureServe’s Digital Distribution Maps of the Freshwater Fishes of the Conterminous United States. Fish occurrence data from these datasets were aggregated to the HUC8 subbasin scale, which represented discrete communities. Phylogenetic data were acquired from a super tree created for all species in the analysis and trait data were acquired from FishTraits database. Missing trait values were imputed based on phylogenetic relationships. Beta diversity was calculated at each time step as the pairwise dissimilarity between each unique pair of communities using the Sorensen Dissimilarity Index for taxonomic beta diversity and PhyloSor Index for functional and phylogenetic beta diversity. A Bayesian correlation model was used to estimate mean beta diversity values and to calculate the correlations between each beta diversity dimension at each time step. Derived quantities tested if mean beta diversity values and correlation coefficients differed over time. Preliminary results showed an increase in mean beta diversity values in all three dimensions of diversity, indicating that freshwater fish communities have differentiated. In addition, we observed a decrease in key beta diversity relationships over time, indicating a change in community assembly processes. In conclusion, this study suggests that key community assembly processes have changed over time, resulting in a reshuffling of stream fish diversity.