Assessing and explaining future changes on sub-daily precipitation extremes using an ensemble of convection-permitting models

Anticipating and understanding the future evolution of intense precipitation events is crucial for improved risk management, especially in regions with mountainous terrain and urban areas vulnerable to natural disasters from extreme weather. Convection-permitting climate models (CPMs) operating at kilometer scales realistically depict convective precipitation mechanisms and complex terrain, enhancing the description of sub-daily extreme precipitation. However, their computational demands restrict simulations to short time periods (10-20 years), and limit the availability of ensemble members, hindering the evaluation of extreme event change and associated uncertainty. This study employs an innovative non-asymptotic extreme value approach, proven effective in estimating rare return levels with reduced stochastic uncertainty even from short datasets, and which can help in providing insights on the changing processes. We apply the Simplified Metastatistical Extreme Value distribution (SMEV) to estimate the projected changes in future extreme sub-daily precipitation in a region characterized by complex terrainspecifically, the North Italy area encompassing both lowlands and the Italian Alps. Our analysis focuses on an ensemble of 9 CPMs from the CORDEX-FPS project, with a spatial resolution of 3 kilometers. We investigate three time periods: historical (1996-2005), near future (2041-2050), and far future (2090-2099) under the RCP8.5 emission scenario. We estimate return levels up to a 1% yearly exceedance probability (100-year return time) for precipitation durations from 1 to 24 hours. Their future change is evaluated at each grid point, conducting a permutation test to assess the statistical significance of the changes. Results indicate a general increase in extreme precipitation across the domain and all durations, with spatial patterns of significant changes varying with durations, time period, and location. A pronounced increase occurs in some of the mountainous areas: at short durations in Eastern Alps, and across all durations in the northern Apennines. The western Alps and surroundings show moderate and not-significant change. Leveraging SMEV's ability to separate precipitation intensity distribution from event occurrence, we also examine the change in distribution parameters to interpret the shift in return levels in term of changes in thermodynamics (linked to temperature and water vapor content) and atmospheric dynamics controls. Interestingly, thermodynamics seems to be driving significant changes at short durations, while small-scale local dynamics contribute across all durations. Differences emerge between the Eastern Alps and Northern Apennines, with the latter showing a stronger intensification of intense versus moderate extreme events. These findings provide valuable insights towards quantifying and understanding the future changes in precipitation extremes, benefiting stakeholders involved in risk management and design of adaptation measures. This study was carried out within the RETURN Extended Partnership and received funding from the European Union Next-GenerationEU (National Recovery and Resilience Plan NRRP, Mission 4, Component 2, Investment 1.3 D.D. 1243 2/8/2022, PE0000005).