Convection‐Permitting Models Offer Promise of More Certain Extreme Rainfall Projections.

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. Plain Language Summary: 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. Key Points: 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 [ABSTRACT FROM AUTHOR]