A comparison of current and reconstructed historic geographic range sizes as predictors of extinction risk in Australian mammals

Comparative studies of extinction risk in vertebrate taxa often find that a small geographic range size is the strongest predictor of a high rate of species decline. This suggests that narrowly distributed species are more vulnerable to human impacts, which may have implications for the predictive use of comparative extinction-risk models in conservation planning. However, this association is potentially circular because many species that have suffered substantial declines now have small geographic ranges, making it difficult to separate the role of range size as a predictor of extinction risk from its role as a response to human impact. Here we use data for Australian mammals to compare models of extinction risk that include current geographic range size with models that include historic range sizes reconstructed for the period before European settlement. We find that current range size is a strong predictor of a species' IUCN Red List classification. However, when historic range sizes are used, range size is non-significant and life-history traits assume primary importance in the model. Models that include current range size also tend to underestimate levels of latent extinction risk (the discrepancy between a species' current extinction risk and that predicted from its biological traits), giving misleading predictions of the species and regions with greatest potential for future species declines. The results suggest that there is circularity in the use of current range size to predict rates of species decline, and that species with inherently small distributions are not necessarily the most vulnerable to human impact.