Precipitation event profiles in a sub-hourly convection-permitting climate model ensemble

Apart from the rainfall depth, the impact of an extreme precipitation event is influenced by its temporal profile, including the timing, magnitude, and duration of the peak intensity, which often occur on sub-hourly time scales. It is therefore crucial to accurately represent this time scale in climate models to increase the confidence in projected climate change signals of extreme precipitation.

High-resolution climate projections at the convection-permitting (CP) scale have been shown to improve the representation of precipitation intermittency, intensity, and diurnal cycle, and this greatly improves their representation of extreme precipitation at sub-daily time scales. However, previous studies of CP simulations have often been limited to hourly model outputs, and little is known about their representation of sub-hourly extreme precipitation.

Our study investigates sub-hourly precipitation in the KIT-KLIWA ensemble - a CP climate model ensemble over Germany with a resolution of 2.8 km. It is driven by 3 CMIP5 GCMs that are coupled to the regional climate model COSMO-CLM. We use a novel event-based approach to compare modelled extreme precipitation events at a temporal resolution down to 5 mins with station and radar observation networks in Germany for the historical period (1971-2000).

Our results show the benefit of using an event-based analysis for the understanding of modelled precipitation biases in CP climate model simulations. Moreover, we find that key features of the temporal precipitation event profiles - including the 5-min peak intensity and the timing of the bulk precipitation - are reproduced by the CP climate model simulations.