Hot extremes increase faster than models suggest: evidence from observation-constrained projections

The increasing frequency of extreme hot events poses major societal and scientific challenges due to their adverse impacts on human and natural systems, compounded by their unpredictable nature. Climate models are essential for identifying the mechanisms that amplify extremes and for anticipating long-term changes that inform decision making, yet their accuracy is limited by internal variability, structural uncertainties, and systematic biases. Observational constraint approaches that link past and future behavior of physical observables offer a promising way to address these limitations, though they often rely on region-specific empirical relationships.

Here, we show that future changes in hot event probabilities and their uneven spread across global land areas depend critically on the historical properties of temperature distributions. In particular, historical variability controls the growth rates of probabilities, either amplifying or dampening the effects of regional background warming, with important implications for climate-change projections. Building on this insight, we develop a universal analytical framework that combines observational evidence with model output to provide more robust assessments of future changes. Results indicate that hot event probabilities may increase faster than suggested by models alone across much of the land surface. In large areas, including the Euro-Mediterranean and Southeast Asia, observation-constrained increases could exceed model-based estimates by nearly a factor of two, even at low levels of global warming. Surpassing the 2 °C warming threshold could push highly vulnerable regions, such as the Amazon and other tropical land areas, into uncharted climate conditions where extreme heat becomes routine.

These findings support more realistic evaluations of future risk and underscore the need for strengthened mitigation efforts to prevent rapid and potentially irreversible climate shifts.