No species left behind: borrowing strength to map data-deficient species.

Abstract

We lack the data needed to detect and understand biodiversity change for most species, despite some species having millions of observations. This unequal data coverage impedes conservation planning and our understanding of biodiversity patterns. The 'borrowing strength' approach leverages data-rich species to improve predictions for data-deficient species. We review multi- and joint-species distribution models that incorporate traits and phylogenies (termed 'ancillary information') and highlight how they could improve data-deficient spatial predictions. When ancillary information is informative of niche similarity, it has immense potential to improve estimates for data-deficient species distributions and address the Wallacean shortfall. While no statistical method can replace data-collection efforts, approaches discussed in this review offer an important contribution toward closing existing data gaps.