A Landscape-Based Classification of Fish Assemblages in Sampled and Unsampled Lakes

We related fish species patterns and landscape-scale environmental data from 216 Michigan lakes to identify repeatable types of fish assemblages, identify environmental factors related to assemblage types, and classify fish assemblages in unsampled lakes. Multivariate regression tree modeling of fish species abundances identified six assemblage types that were explained by degree-days during the ice-free period, lake surface area, and mean lake surface temperature. Warmwater species dominated southern lakes, while coolwater and coldwater species had higher abundances in northern lakes. Coolwater species were present in large southern lakes, whereas warmwater species were excluded from northern lakes that had low mean surface temperatures or low degree-days. These results suggest that patterns of lake fish assemblages are shaped by differences in climate as well as lake-specific differences in surface temperature regimes and in vertical availability of coldwater and coolwater habitats. Because we related fish patterns to readily available landscape-level data, our approach can be used to characterize fish assemblages in all lakes across broad geographic extents.