Species distribution models rarely predict the biology of real populations

Species distribution models (SDMs) are widely used in ecology. In theory, SDMs capture (at least part of) species' ecological niches and can be to make inferences about the of suitable habitat for species interest. Because suitability is expected influence population demography, have been estimate a variety parameters, from occurrence genetic diversity. However, critical look at ability predict independent data across different aspects biology lacking. Here, we systematically reviewed literature, retrieving 201 studies that tested predictions against assessments occurrence, abundance, performance, Although there some support (~53% depending on how was assessed), predictive performance these declines progressively mean fitness, At same time, observed higher success among evaluated single versus multiple species, pointing possible publication bias. Thus, limited accuracy reported here may reflect best-case scenario. We discuss limitations provide specific recommendations their use applications going forward. emphasize SDMs, especially when inform conservation decisions, should treated as hypotheses with rather than stand-ins parameters seek know.