Protected Areas Serve as Climate Refugia by Reducing Thermal Stress in Ectotherms

Abstract

Protected areas (PAs) are increasingly expected to buffer biodiversity from climate change, yet their effectiveness in reducing physiological stress for ectothermic species remains poorly understood. Here, we tested whether PAs serve as climate refuges for Australian Lampropholis skinks by combining species-specific physiological data with biophysical modelling. Using three key thermal metrics - thermal safety margin (TSM), hours above critical thermal maximum (CTmax), and daily activity hours - we simulated environmental and physiological conditions across thousands of locations inside and outside PAs under present-day, +2 °C, and +4 °C warming scenarios. Our results show that PAs consistently buffer skinks from heat stress by increasing TSMs, reducing exposure to temperatures above CTmax, and maintaining higher activity time under current and climate change scenarios. These benefits were strongest at low latitudes, where thermal stress is most intense, and varied among species depending on their physiological traits and ecological specialisation. Species with broader thermal tolerances experienced greater benefits from protection. Our findings demonstrate that PAs can mitigate climate-driven physiological stress in ectotherms, especially in tropical regions. As climate change intensifies, incorporating mechanistic models into conservation planning can guide the strategic design and management of PAs to support species persistence under climate change.

Data and Code Availability

data.zip: Contains the output data from the biophysical simulations. For each Lampropholis species, files include results under three climate scenarios (present, +2 °C, +4 °C), with metrics such as thermal safety margin, hours above CTmax, and daily activity hours calculated per coordinate.

scripts.zip: Includes all R scripts used to run the analysis. This covers microclimate and ectotherm model simulations (using NicheMapR), extraction of spatial coordinates, model fitting (Bayesian models using brms), and generation of figures and tables included in the manuscript and supplementary material.

coord_lampro: CSV file with the spatial sampling coordinates used to run the simulations, including information on species identity, protection status (inside vs outside PAs), and latitude/longitude.