Predicting the impacts of climate extremes on plant communities is a central challenge in ecology. Physiological traits may improve prediction of drought impacts on forests globally. We perform a meta-analysis across 33 studies that span all forested biomes and find that, among the examined traits, hydraulic traits explain cross-species patterns in mortality from drought. Gymnosperm and angiosperm mortality was associated with different hydraulic traits, giving insight into the relative weights of different traits and mechanisms in mortality prediction. Our results provide a foundation for more mechanistic predictions of drought-induced tree mortality across Earth’s diverse forests.
Drought-induced tree mortality has been observed globally and is expected to increase under climate change scenarios, with large potential consequences for the terrestrial carbon sink. Predicting mortality across species is crucial for assessing the effects of climate extremes on forest community biodiversity, composition, and carbon sequestration. However, the physiological traits associated with elevated risk of mortality in diverse ecosystems remain unknown, although these traits could greatly improve understanding and prediction of tree mortality in forests. We performed a meta-analysis on species’ mortality rates across 475 species from 33 studies around the globe to assess which traits determine a species’ mortality risk. We found that species-specific mortality anomalies from community mortality rate in a given drought were associated with plant hydraulic traits. Across all species, mortality was best predicted by a low hydraulic safety margin—the difference between typical minimum xylem water potential and that causing xylem dysfunction—and xylem vulnerability to embolism. Angiosperms and gymnosperms experienced roughly equal mortality risks. Our results provide broad support for the hypothesis that hydraulic traits capture key mechanisms determining tree death and highlight that physiological traits can improve vegetation model prediction of tree mortality during climate extremes.