Improved Characterization of Coral Bleaching Patterns From a Percentile‐Based Threshold Model

Abstract Tropical corals are increasingly threatened by anthropogenic global warming and marine heatwaves. However, questions remain about how to best characterize potential heat stress relevant to coral bleaching. The degree heating week (DHW) metric has guided bleaching risk assessments for ∼25 years. Here, we compare the DHW approach, which uses a static anomaly threshold, with a percentile‐based approach developed from a widely used marine heatwave definition. We find that the explanatory power of coral bleaching risk—measured by the equitable threat score—is higher for the percentile‐based approach compared to the DHW, with improvements of up to 41%. This is likely because percentile thresholds better reflect spatiotemporal temperature variability. We further demonstrate how the weather‐forecasting framework can help identify current optimal adapting baselines, a key step toward evaluating ecological risks as global warming challenges the use of fixed versus shifting baselines.