Key points are not available for this paper at this time.
Abstract Extreme precipitation events can cause catastrophic flooding and damage, with climate change projected to intensify associated risks. For reliable impact assessments, it is crucial to account for rare and unprecedented events that produce the most severe societal and economic consequences. Climate emulators provide an efficient means to explore such possibilities, generating large ensembles at a fraction of the computational cost of Earth system models. We introduce a framework for emulation of annual maximum daily precipitation that produces spatially resolved time series conditioned on global mean temperature trajectories, developed within the MESMER-X framework. The emulator, trained on CMIP6 models, faithfully replicates their key statistical properties, natural variability, spatial patterns, and global warming response. Dedicated treatment of arid-region behavior, where even annual maxima can approach zero, and of heavy-tailed tropical extremes ensures consistency with the input data. These developments enable large-ensemble probabilistic exploration and rapid assessment of heavy precipitation events across warming trajectories, supporting regional- to global-scale climate risk assessments and policy-relevant questions linking emission scenarios to changes in precipitation extremes.
Pierini et al. (Wed,) studied this question.