Testing alternative methodologies for modeling species' ecological niches and predicting geographic distributions

Knowledge of world biodiversity remains sparse, with millions of species left to be described, most species' geographic distributions poorly understood and the ecological and evolutionary processes that underpin geographic patterns of diversity still far from resolved. Many large-scale conservation projects, however, depend critically on more complete descriptions of species' distributions and there is increasing interest in incorporating process as well as pattern into biodiversity evaluation. The inferential step that leads from incomplete present knowledge to a explicit prediction of geographic distribution is presently made via diverse methods which have not been tested against each other to establish which would provide the greatest predictive ability for different types of questions and data. We propose a NCEAS working group that will review and compare diverse predictive modeling approaches with the goal of producing an ideal strategy for modeling parameters related to ecological niches and predicting geographic distributions.