What Drives the Most Extreme Fire Weather in Europe? A Large-Ensemble Perspective

Recent European wildfire seasons reinforced the urgency of understanding fire weather extremes, characterized by hot, dry, and windy conditions. While previous studies have primarily focused on individual events or anthropogenic attribution, the physically plausible range of fire weather extremes and the contribution of internal variability remain less well explored. We address this gap by using large ensemble climate simulations to comprehensively sample internal variability and to quantify the underlying conditions of extreme and most extreme fire weather - atmospheric circulation, soil moisture and sea surface temperatures - at different spatial and temporal scales. We analyze fire weather extremes during the fire season (May-October) in Southwest Europe using the CMIP6 MPI Grand Ensemble (MPI-GE, 50 members) and ERA5 reanalysis. We evaluate temperature and humidity conditions during extreme events in MPI-GE against ERA5 to assess the representation of climate conditions associated with observed extremes in the model, thereby providing physical grounding for unprecedented events simulated in the ensemble.

In MPI-GE, the most extreme fire weather events are more persistent than in ERA5, with durations of up to two months, and can exhibit spatial extents up to 20% larger than the largest ERA5 events. Despite extending beyond the observational record, the distributions of event duration and spatial extent remain comparable between MPI-GE and ERA5, supporting the use of large ensembles to investigate rare high-impact events. In both datasets, extreme events occur under compound atmospheric conditions characterized by enhanced vapour pressure deficit, depleted soil moisture (SM), and positive geopotential height at 500 hPa (Z500) anomalies. Differences in Z500 anomaly strength and spatial pattern among the most extreme events in MPI-GE are small, indicating that increasing severity does not arise from fundamentally different atmospheric regimes. Positive adjacent North Atlantic sea surface temperature anomalies are detectable already in spring, particularly for the most severe events, while multi-year SM conditions vary substantially, showing that both seasonal drying and persistent multi-year drought can lead to extreme fire weather. Overall, our results indicate that the most severe fire weather events arise from amplified thermodynamic preconditioning within similar large-scale atmospheric conditions, highlighting potential predictability of extreme fire weather risk and opportunities for improved preparedness.