Disentangling the drivers of local species richness using probabilistic species pools

Aim Local richness is a result of both regional richness and local site‐specific factors. We quantify the effects of different regional (dispersal, environmental filtering) and local (habitat selection) community assembly processes in the Cape reeds (Restionaceae) of the Cape flora and test if community assembly processes vary spatially. Location Southern Africa, Cape Floristic Region. Taxon Angiosperms, African Restionaceae (restios). Methods We calculate local probabilistic species pools, based on local species richness, dispersal rates and environmental filtering at a 4 km2 resolution. We then compare the relative contribution of the processes shaping the species pool spatially. Finally, we correlate observed species richness (based on 916 circular, 10 m diameter releves) with the local species pools to test the explanatory power of additional local edaphic variables that could act as filters at very small scales. Results There is substantial spatial variation in the local probabilistic species pools. Within the central part of the Cape, the local pools are mostly influenced by environmental filtering, while along the periphery dispersal limitation dominates. Observed species richness is lower than the predicted species pool size, which forms an upper boundary of the potential species richness for a local site. Local, site‐specific edaphic factors help to explain the remaining differences between local species pool size and α‐richness. Main conclusion The analytical framework for inferring local probabilistic species pools provides a powerful tool for including the influence of regional species richness on variations in releve α‐richness. This allows us to disentangle the various processes that operate at different spatial scales, and ultimately determine local restio species richness in the Cape Floristic Region.