Using regional climate models to simulate extreme rainfall events in the Western Cape, South Africa

This study evaluates the capability of regional climate models (RCMs) in simulating extreme rainfall events over Southern Africa, and in reproducing the characteristics of widespread extreme rainfall events (WERE) in the Western Cape (South Africa). We obtained simulation datasets of nine RCMs from the Co-ordinated Regional Downscaling Experiment (CORDEX) and compared them with observation datasets from the Global Precipitation Climatology Project (GPCP) and Tropical Rainfall Measuring Mission (TRMM) as well as with the reanalysis dataset (ERAINT) that forced the simulations. A self-organising map was used to classify the WEREs into 12 nodes. The contribution of each RCM to each node was compared with the contributions of GPCP, TRMM and ERAINT. Using ERAINT dataset, we analysed the synoptic-scale atmospheric condition associated with each node. The results show that, in simulating the spatial distribution of the extreme rainfall event over Southern Africa, only four RCMs perform better than the forcing reanalysis (ERAINT) while two RCMs perform worse than the reanalysis. All the RCMs underestimate the threshold of extreme rainfall over Western Cape, poorly simulate the inter-annual variability of the WEREs (r 0.3), but correctly reproduce the maximum frequency of the WEREs in Autumn (March-May). The WEREs in the Western Cape may be broadly grouped into four synoptic rainfall patterns. The first pattern links WEREs with tropical rainfall activities (tropical-temperate troughs); the second pattern shows isolated WEREs; the third and fourth patterns link WEREs with rainfall activities in the mid-latitudes (frontal systems) and over the Agulhas Current, respectively. While most RCMs overestimate the frequency of the first pattern, they all underestimate the frequency of the second pattern but simulate the frequencies of the third and fourth patterns well. The results of this study should help in improving the RCM simulations over Southern Africa and in downscaling impacts of climate change on extreme rainfall events over the region.