Measuring Climate Impact on Extinction Risk in Amphibians

The increasing relevance of climate change as a threat of species extinction is a pressing concern, as highlighted by the recent IUCN Red List assessment for amphibians. Based on the concept of the climate niche, i.e. conditions required for a stable population, recent studies have estimated the dramatic implications of different climate change scenarios on species. However, there is an ongoing discussion in the community which measures are best suited to quantify climate change impacts on extinction risk, given the available data.
In this study, we provide a consistent framework to evaluate three published measures on historical changes in extinction risk. The compared measures include a classical bioclimate envelope approach, an ensemble of species distribution models, and the average climate change within species' ranges. We train an advanced statistical model (random forest) to predict changes in extinction risk between 1980 and 2021 in 6,288 amphibian species. This analysis is controlled for factors such as geographical range area, human pressures, and other external threats.
We find that two measures based on the climate niche do not predict historical changes in risk, when other factors are controlled for. Also, we find that predictions of risk, based on average climate change, can be misleading when applied to future scenarios. These findings highlight the limitations and inherent uncertainties of predicting climate impact for a high number of species, given the standard datasets and tested methods.