Search

Reliable projections of extreme future precipitation are fundamental for risk management and adaptation strategies. Convection‐permitting models (CPMs) explicitly resolve large convective systems and represent sub‐daily extremes more realistically than coarser resolution models, but present short records due to the high computational costs. Here, we evaluate the potential of a non‐asymptotic approach, the Simplified Metastatistical Extreme Value (SMEV) to provide information on the future change of extreme sub‐daily return levels based on CPM simulations. We focus on a complex‐orography area in the North Eastern Italy and use three 10‐year time periods COSMO‐crCLIM simulations (2.2 km resolution) under RCP8.5 scenario. When compared to a block r‐maxima approach currently used in similar applications, the proposed approach shows reduced uncertainty in rare return level estimates (about 5%–10% smaller confidence interval) and can improve the quantification of future changes from CPM simulations. We evaluate these changes and their statistical significance in return levels for 1–24 hr durations. The changes show an interesting spatial organization associated with orography, with significant positive changes located at high elevations. These positive changes tend to increase with increasing return period and decreasing duration. Because SMEV can separate the roles of event intensity and occurrence, it allows for physical interpretations of these changes. We suggest that non‐asymptotic approaches permit the quantification of change in rare extremes within available CPM runs. Short duration heavy rainfall may lead to various natural hazards like floods and landslides. Expected change in extreme precipitation due to global warming is a major concern. However, we still cannot quantify these changes because typical climate models cannot reproduce extreme precipitation accurately. The few models that can are very computationally expensive so that we have too few simulations for properly quantifying changes in extremes using traditional statistical methods. Here, we show how to use a new statistical method to quantify extremes from short model simulations. This method is more accurate than currently used methods and may help provide additional insights on the reasons underlying the observed changes. This method could represent a new tool in the hands of the climate research community. Examining the simulations of one model over North‐Eastern Italy, we report an increase in extreme precipitation in mountainous areas and a non‐significant decrease in the low elevation areas. Future changes in extreme precipitation are estimated from a convection‐permitting climate model using a non‐asymptotic statistical approachThe method allows to evaluate the significance of the future changes in return levels and to link them to the changing processesSignificant increase in return levels is generally found in the mountains, higher at the short durations (1–3 hr) and for rarer return levels Future changes in extreme precipitation are estimated from a convection‐permitting climate model using a non‐asymptotic statistical approach The method allows to evaluate the significance of the future changes in return levels and to link them to the changing processes Significant increase in return levels is generally found in the mountains, higher at the short durations (1–3 hr) and for rarer return levels

The future increase in extreme precipitation is likely to have a severe impact on society due to flooding. Previous research has shown the improved representation of precipitation in convection‐permitting models (CPMs), but until now it has not been possible to quantify uncertainties in future changes at convective (<5 km) scales. Here we analyze the first‐ever ensemble of convection‐permitting climate projections run within the UK Climate Projections project. We find that the CPM ensemble shows a significantly stronger intensification of summer hourly precipitation compared to the driving 12 km ensemble, with increases above the Clausius‐Clapeyron relation and natural variability dominating the ensemble spread for extremes. Results suggest that the climate change signal across different CPMs may converge thanks to the more realistic representation of the local storm dynamics. We conclude that CPMs offer the promise of reducing uncertainties for extreme summer precipitation projections due to the better and explicit representation of convection. Accurate predictions of future changes in precipitation extremes are crucial for developing effective adaptation measures to avoid severe impacts on society due to flooding. State‐of‐the‐art kilometer‐scale models, called convection‐permitting models, give a much better simulation of how rainfall varies hour by hour and thus provide credible projections of future changes in hourly rainfall extremes. Up to now it was not possible to evaluate uncertainties in the future changes in extremes at convective (<5 km) scales since only single simulations were available due to the high computational costs. Thus, the ensemble of 12 local (2.2 km) projections, run within the UK Climate Projections project, represents a key step forward for climate prediction. We find that the 2.2 km ensemble shows a significantly stronger future increase of summer hourly precipitation compared to the coarser resolution (12 km) ensemble. The climate change signals are found to be consistent across the 2.2 km ensemble simulations, with differences largely explained just by natural variations in the climate from year to year. This means that convection‐permitting models could reduce the uncertainties in future changes in extreme summer precipitation and thus allow more robust and effective adaptation policy. Convection‐permitting model (CPM) ensemble projects an intensification above the Clausius‐Clapeyron relation for extreme precipitationClimate change signal for extreme precipitation appears to converge in the CPM ensemble, with uncertainty dominated by natural variabilityThe correct representation of convection may lead to reduced model uncertainties in future changes in extreme precipitation