Ecological bleaching trajectories under severe heat stress are only partially captured by acute heat stress assays

Global marine heatwaves have devastated tropical coral reefs, and further mortality is projected under ongoing climate change. Identifying thermally tolerant coral colonies is therefore a priority for conservation, restoration, and research. Portable acute heat stress assays (e.g., CBASS) enable rapid, standardized estimates of coral thermal tolerance under field conditions. However, it remains unresolved whether such experimentally derived metrics (ED5, ED50, DW) predict bleaching and mortality in situ. Here, we quantified acute thermal tolerance metrics for 2,068 coral colonies across 12 common Indo-Pacific species, six months prior to an unprecedented heat stress event in northeastern Peninsular Malaysia and compared experimentally derived ED and DW values to subsequent bleaching severity and mortality in the field. Experimental thermal tolerance metrics explained only a limited proportion of variation in bleaching outcomes and survival. Predictive power varied among species and was higher in slow-growing species. Our findings suggest that while acute heat stress assays capture substantial variation in coral thermal tolerance, their ability to predict in situ outcomes is context-dependent and diminishes under severe thermal stress. Ultimately, in situ coral bleaching under severe heat stress may reduce the discriminatory capacity of acute assay-derived tolerance metrics.