Effects of neutrality, geometric constraints, climate, and habitat quality on species richness and composition of Atlantic Forest small-mammals

Abstract

Aim To compare the fit of models of climate, habitat quality, neutral processes, and geometric constraints to species richness and composition of small mammal assemblages. Location The South American Atlantic Forest biome. Methods Using neutral models and mid-domain effect models, we simulated species spread in a spatially explicit array of grid cells representing the Atlantic Forest domain. We compared empirical patterns of species richness and composition with predictions of the neutral and mid-domain effect models. We also modeled individual species responses to climatic conditions and forest integrity, a measure of habitat quality. Results Habitat quality was the single best predictor of local species richness (α-diversity), but was a poor predictor of local species composition and of the decay in species similarity with distance (β-diversity). The neutral and mid-domain models generated very similar predictions, and were better predictors of species composition than of species richness. Climate variables were also strongly associated with overall species composition, but not with species richness. Main Conclusions The species richness of small-mammal assemblages in the Atlantic Forest is best explained by variation in habitat quality. In contrast, the composition of small-mammal assemblages is best explained by models of limited dispersal (neutral and mid-domain) and effects of climate on local species composition. Collectively, these results suggest that regional patterns of species richness may be uncoupled from patterns of species composition. Both species richness and composition should be considered when evaluating the predictions of neutral and mid-domain effect models, and of correlations of community structure with climatic or habitat variables.