Forecasting wildlife die-offs from extreme heat events

Extreme heat events pose increasing challenges to biodiversity conservation worldwide, yet our ability to predict the time, place and magnitude of their impacts on wildlife is limited. Extreme heat events in Australia are known to kill thousands of flying-foxes (Pteropus spp.), and such die-offs are expected to become more frequent and widespread in the future under anthropogenic climate change. There is a growing need for predicting when and where such heat-related die-offs would occur, to facilitate short-term wildlife management and conservation actions. In this study, we used gridded hourly air temperature forecasts [Australian Community Climate and Earth-System Simulator (ACCESS-R) Numerical Weather Prediction (NWP) model] from the Australian Bureau of Meteorology to predict flying-fox heat-related mortality based on an empirically determined threshold of 42.0 degrees C. We tested the accuracy and precision of this model using a twofold evaluation of the ACCESS-R NWP forecast air temperature during a recorded extreme heat event with in situ air temperature measurements and interpolated weather station data. While our results showed a slight discrepancy between the modelled and measured air temperatures, there was no significant difference in the forecast's accuracy to predict die-offs during an extreme heat event and the overall summer period. We evaluated the accuracy of mortality predictions based on different air temperature thresholds (38.0, 40.0, 42.0 and 44.0 degrees C). Our results revealed a significant probability of flying-fox mortality occurrence when forecast air temperature was >= 42.0 degrees C, while the 24- and 48-h forecasts accurately predicted 77 and 73% of the die-offs, respectively. Thus, the use of 42.0 degrees C forecast air temperature from the ACCESS-R NWP model can predict flying-fox mortality reliably at the landscape scale. In principle, the forecaster can be used for any species with known thermal tolerance data and is therefore a promising new tool for prioritizing adaptation actions that aim to conserve biodiversity in the face of climate change.