Storm Boris' rainfall: Robust increases at moderate warming levels, large uncertainty at higher warming

Understanding the influence of climate change on environmental extremes is essential to inform adaptation and mitigation policies. Here, we employ an event-focused storyline methodology to evaluate changes in total precipitation linked to Storm Boris, which struck Central Europe in September 2024. Our study addresses both historical (attribution) and future (projection) changes, and builds on our previous work by exploring how results vary with the stringency of applied dynamical constraints. Simulations are conducted with the global CMIP6 coupled climate model AWI-CM1, in which the winds observed before and during Storm Boris—including the jet stream evolution—are imposed (nudged). Such simulations are performed with those same winds under a range of climate states: preindustrial, present-day, and possible future states with 2, 3, and 4 °C of global warming relative to preindustrial levels. We test two nudging regimes: (1) a "weak constraint," nudging only synoptic- and planetary-scale winds from ERA5 in the free troposphere to allow partial dynamical adaptation, and (2) a "strong constraint," imposing winds across all vertical levels and scales to fully inhibit dynamical changes.

Both approaches successfully represent the event, with the strongly constrained setup yielding higher present-day precipitation totals, yielding a present-day rainfall closest to observations. The intensification of accumulated rainfall from preindustrial to present-day is robust, showing increases of 7% (weak constraint) and 4% (strong constraint). For up to +3ºC of global warming, both methods display roughly linear increases in total rainfall. However, at +4ºC, the results diverge: under weak constraints, precipitation changes are minimal or slightly negative relative to present-day, whereas under strong constraints, they continue to increase linearly. These differences are due to thermally-induced dynamical adaptations allowed under the weak constraint. It remains unclear whether these responses represent actual physical responses or are influenced by methodological limitations, and whether similar divergence would be observed in other extreme events. Then, these discrepancies reinforce the need to study a broader set of events and to adopt multi-method approaches to project extreme precipitation changes.