Climate change impacts on extreme rainfall frequency prediction

Abstract

Abstract The full integration between the computation of climate change effects and the prediction of extreme rainfall frequency is not yet well developed. In this study, the maximum daily rainfall of 26 stations in the western Kingdom of Saudi Arabia (KSA) are extracted covering an area of 180,000 km2 for processing and analyzing. Discordancy test (Di) showed that some stations are discordant, and the selected study area needs to be subdivided in order to reduce the inherent discordance. The rainfall stations are subdivided into three sub-regions based on a new approach by using L-Skewness parameter value (low, moderate, and high). Five probability distribution functions (PDFs) are evaluated using goodness of fit (Zdist) test and L-moment ratios diagram (LMRD). It was found that for sub-regions A, B, and C, the best fits are generalized Pareto (GPA), Pearson type III (PE3) and generalized extreme value (GEV) PDFs, respectively. Regional growth curves for each sub-region are developed and the predicted extreme rainfall for 100 years' return periods are computed for each station. Finally, climate change impact is evaluated using the emission scenario A2 which is about +40% and the predicted extreme rainfall frequency is computed taking into consideration the climate change impacts.