Towards improved prediction of compound heat and drought extremes by CWRF downscaling

Given their profound socio-economic impact and increasing occurrence, compound heat and drought extremes (CHDEs) have become a focal point of widespread concern. Numerous studies have attempted to reproduce and predict these extremes using general circulation models (GCMs); however, the performance of these models in capturing extreme events remains controversial. This study presents an improved historical simulation of CHDEs over China by using the regional Climate-Weather Research and Forecasting model (CWRF) to downscale the projections of two GCMs that participated in the Coupled Model Intercomparison Project Phase 6. The CWRF downscaling improved GCMs in capturing the thresholds of extreme hot and extreme dry conditions and demonstrates a better agreement with observations in the temporal trends and spatial patterns of extreme heat and extreme drought events. The performance of CWRF downscaling to reproduce CHDEs also surpasses that of GCMs, with an even greater enhancement compared to univariate extreme events. The improvement is particularly pronounced in sub-humid areas, which is primarily attributed to the enhanced simulation of temperature-precipitation coupling relationships by CWRF downscaling. This superiority is found to be associated with the finer land surface processes and land-atmosphere interaction processes of CWRF. This study highlights the important role of land-atmosphere interactions in shaping CHDEs and the efficacy of using regional climate models to reduce uncertainty in extreme event simulations.