Ecophysiological constraints outperform environmental predictors in forecasting climate-driven extinction risk of a Cerrado endemic lizard

Abstract

Introduction Forecasting species responses to climate change using only environmental predictors may underestimate extinction risk because it ignores physiological constraints. Methods We compared species distribution models for Micrablepharus atticolus built with environmental predictors, thermal performance, hours of activity, and a hybrid combination of these predictors. We projected habitat suitability to 2050 and 2070 using 12 global circulation models, three representative concentration pathways, and three land-use scenarios. Results Model accuracy, quantified using the True Skill Statistic (TSS), was highest for the thermal performance model, followed by hours of activity, environment-only, and the hybrid model. The thermal performance model predicted larger current suitable areas and substantially higher future extinction risk than the environment-only model. Discussion For M. atticolus , and likely for species with similar ecological and physiological traits, ecophysiology-based models provide more reliable TSS-based predictions than models based only on environmental predictors and show that ignoring ecophysiological information may underestimate climate-driven extinction risk.