Beyond Attributing Precipitation Extremes: Warming Partly Counteracts Climate Change-Driven Flood Increases

Anthropogenic climate change is known to intensify extreme precipitation events, which is generally assumed to follow approximately the Clausius-Clapeyron relation at 6% per K of warming. Whether the resulting flood hazards scale proportionally, however, remains uncertain. Here, we employ nudged climate model simulations to reproduce observed extreme precipitation events by constraining large-scale atmospheric circulation. We then conduct runoff simulations in which observed weather data are adjusted according to the nudged climate simulations, under both observed and pre-industrial greenhouse gas emissions. This approach allows us to isolate the thermodynamic contribution of climate change while avoiding uncertainties in potential large-scale circulation changes. We quantify both extreme precipitation and resulting flood hazards for more than 100 European flood events with observed impacts between 1981 and 2024.

Our simulations show that extreme precipitation intensity would have been on average only about 1% lower without climate change for floods that occurred before 2000, increasing to 5% in the most recent decade (2015-2024). Floods, however, intensify more substantially: climate change-driven precipitation changes would suggest a 7% reduction in peak flows without climate change. Yet higher temperatures have led to an increase in evaporation and thus drier antecedent soil conditions, offsetting some of this precipitation effect. Taken together, the net climate change impact on floods reduces to approximately 3% across events for the recent decade, which is no statistically significant climate change effect. These moderate changes on average mask a high variability across events, and sometimes even a high spatial heterogeneity within the same event; some floods would have been 20% less intense without anthropogenic climate change, with negligible mitigation effects from warming, while for others the flood hazard would have even been larger under pre-industrial conditions due to the combined climate change effects. This underscores the importance of conducting event-specific flood attribution studies to identify highly relevant changes at the local level.

Our results highlight that attributing precipitation extremes alone is not a good proxy for estimating climate change-induced changes of flood hazards. In particular, explicitly simulating floods and accounting for competing temperature effects provides a more complete picture. Reconciling these effects is crucial for public engagement with climate attribution results, and essential for future flood risk assessments and estimates of climate-change-related losses.