Harnessing the world's biodiversity data: promise and peril in ecological niche modeling of species distributions

Recent advances allow harnessing enormous stores of biological and environmental data to model species niches and geographic distributions. Natural history museums hold specimens that represent the only information available for most species. Ecological niche models (sometimes termed species distribution models) combine such information with digital environmental data (especially climatic) to offer key insights for conservation biology, management of invasive species, zoonotic human diseases, and other pressing environmental problems. Five major pitfalls seriously hinder such research, especially for cross-space or cross-time uses: (1) incorrect taxonomic identifications; (2) lacking or inadequate databasing and georeferences; (3) effects of sampling bias across geography; (4) violation of assumptions related to selection of the study region; and (5) problems regarding model evaluation to identify optimal model complexity. Large-scale initiatives regarding data availability and quality, technological development, and capacity building should allow high-quality modeling on a scale commensurate with the enormous potential of and need for these techniques.