Growth rates of hot extremes: insights from observation-constrained model projections

The increasing frequency of extremely hot events poses significant societal and scientific challenges due to their adverse impacts on human and natural systems, compounded by their unpredictable nature. Climate models are essential tools for investigating the amplification mechanisms of extremes and anticipating their changes under continued greenhouse gas warming, conveying vital information for decision-making. Future projections, however, remain limited by inherent uncertainties stemming from the range in the emission scenarios, inter-model structural differences, and internal climate variability. Furthermore, despite progress, models still struggle to accurately capture observed regional trends, particularly in extreme events.
Observational constraint theories, which link past and future behavior of physical observables, provide an effective paradigm to address model deficiencies and reduce uncertainties, though they often rely on empirical, region-specific relationships. Here, we show that future changes in the probability of hot extremes, and their uneven spread across global land areas, critically depend on the historical mean properties of thermal distributions and their evolution. Among these, historical variability plays a central role in shaping the rate of increase in hot extremes, either amplifying or suppressing the effects of regional background warming. Thus, the accuracy of historical simulations can substantially impact projected hot-event probabilities. Based on this, we develop an analytical approach to seamlessly combine global-scale observations with model outcomes, aiming for more reliable projections.
Despite uncertainties in both simulation and observational data, our results suggest that hot extremes may grow faster than models imply across much of the land surface. In regions like the Euro-Mediterranean and Southeast Asia, observation-constrained growth rates of hot extremes could nearly double bare model projections, even under moderate levels of warming. Crucially, exceeding the 2 °C global warming threshold could push highly vulnerable areas, such as the Amazon and tropical lands, into uncharted climate conditions where extremes become routine. These findings provide a more realistic foundation for assessing future risks and highlight the urgent need for strengthened adaptation and mitigation efforts to prevent rapid, potentially irreversible climate shifts.