A global early warning system for predicting exposure of biodiversity to extreme heat

Effective biodiversity conservation during periods of rapid environmental change requires identifying the species and regions at greatest risk. However, most predictions focus on biodiversity shifts several decades ahead, offering limited guidance for immediate conservation action. Here we develop an early warning system for biodiversity based on 9-month seasonal weather forecasts combined with species-specific historical temperature limits. In May 2024, we predicted that >3,500 vertebrate species (out of 30,585 species), including >1,250 species of conservation concern, would be substantially exposed to extreme temperatures 1–2 months in advance on average. Mexico, Sub-Saharan Africa and the Himalayas were predicted to face the highest threats during this forecast period. Early evidence suggests that exposure in these regions likely negatively impacted some species populations. Such advance notice can enable rapid actions to monitor and mitigate these extreme events for sensitive species and regions.