Bridging Ensemble Robustness and Process Realism: Event-Based Convection-Permitting Modeling of Extreme Events in Emilia-Romagna, Italy

Convection-permitting regional climate models (CPRMs) are increasingly recognized as essential for realistically simulating extreme rainfall, particularly for events driven by mesoscale dynamics. However, their application to climate change studies faces a fundamental dilemma: while CPRMs are required to explicitly resolve deep convection and associated rainfall extremes, their high computational cost severely limits ensemble size, constraining robust estimation of changes in rare events and associated uncertainties.

In this study, we address this challenge using an event-based downscaling framework with convection-permitting simulations performed using the HARMONIE-Climate (HCLIM) model at kilometre-scale horizontal resolution. The analysis focuses on the Emilia-Romagna region in northern Italy, which has recently experienced devastating extreme rainfall events. These events were primarily associated with cut-off low systems and Mediterranean cyclones, leading to persistent moisture transport, quasi-stationary convection, and widespread flooding. The non-hydrostatic dynamics and advanced microphysics of HCLIM enable an improved representation of rainfall intensity, spatial organization, and temporal evolution compared to convection-parameterized regional climate models.

To assess future changes in extreme rainfall, we combine targeted HCLIM simulations with the SMHI Large Ensemble, consisting of 50 ensemble realizations based on EC-Earth3 under present-day and warming scenarios. The ensemble will be used for the events for CPRM downscaling, 

bridging the gap between ensemble robustness and process-level realism.

Results suggest that, under warming, cut-off low systems tend to weaken, leading to a reduced intensification of the most extreme rainfall events. This behavior is consistent with a weakened meridional temperature gradient over Europe, which is critical for the formation and maintenance of cut-off lows over the Mediterranean region. Ongoing analyses aim to further quantify associated rainfall changes and disentangle thermodynamic and dynamical contributions. Overall, the proposed framework provides a computationally feasible pathway for assessing future changes in high-impact rainfall extremes using CPRM, supporting climate risk assessment in regions vulnerable to extreme events.