Enhancing island biogeography: improving identification of potential species pools via environmental filtering.

A successful theory in ecology is the equilibrium species numbers on islands. Most of it is about the total number of species on "islands" (sensu lato), ignoring complications like species identities, history, and environmental differences. We propose an operational approach for identification of potential species pools, moving from predicting only numbers toward a better understanding of species identities. This procedure includes two main steps: (1) provide an initial pool of species names, and (2) refine that list via niche-based environmental filtering. To provide the initial species pool (step 1), we explored two regions of different sizes. For step 2, we used ecological niche modeling to separate potential colonizers from species that do not have any distributional potential on the island(s) in question. As an example, we used data on butterflies of the Caribbean Islands. We found that, regardless of the size of the pool, our methods were able to refine (reduce) potential species pools without increasing omission errors. Our predicted species pools, however, show apparent "errors" of commission-we interpret these species as representing as-yet unrealized potential colonizers. Overall, this approach allows researchers to refine estimates of species pools, and thereby improve the ability to predict potential local species richness and species identities.