Ecological niches as stable distributional constraints on mammal species, with implications for Pleistocene extinctions and climate change projections for biodiversity

ABSTRACT Aim Theoretical work suggests that species’ ecological niches should remain relatively constant over long‐term ecological time periods, but empirical tests are few. We present longitudinal studies of 23 extant mammal species, modelling ecological niches and predicting geographical distributions reciprocally between the Last Glacial Maximum and present to test this evolutionary conservatism. Location This study covered distributional shifts in mammal species across the lower 48 states of the United States. Methods We used a machine‐learning tool for modelling species’ ecological niches, based on known occurrences and electronic maps summarizing ecological dimensions, to assess the ability of ecological niches as modelled in one time period to predict the geographical distribution of the species in another period, and vice versa. Results High intertemporal predictivity between niche models and species’ occurrences indicate that niche conservatism is widespread among the taxa studied, particularly when statistical power is considered as a reason for failure of reciprocal predictions. Niche projections to the present for 8 mammal taxa that became extinct at the end of the Pleistocene generally increased in area, and thus do not support the hypothesis of niche collapse as a major driving force in their extinction. Main conclusions Ecological niches represent long‐term stable constraints on the distributional potential of species; indeed, this study suggests that mammal species have tracked consistent climate profiles throughout the drastic climate change events that marked the end of the Pleistocene glaciations. Many current modelling efforts focusing on anticipating climate change effects on species’ potential geographical distributions will be bolstered by this result — in essence, the first longitudinal demonstration of niche conservatism.