Evaluation and Projection of Precipitation and Precipitation Extremes in the Source Region of the Yangtze and Yellow Rivers Based on CMIP6 Model Optimization and Statistical Downscaling

Abstract

This study evaluated the performance of a high-resolution statistical downscaling (HSD) approach integrating optimal global climate models (GCMs) and quantile delta mapping (QDM) for the source region of the Yangtze and Yellow rivers, and then projected regional precipitation extremes. The GCMs captured the general precipitation pattern, but the results indicated systematic overestimations, particularly in eastern parts of the region, with deviations reaching 304.8% for winter. The HSD approach improved the spatial correlation coefficients (SCCs) and reduced the biases for mean precipitation and precipitation extremes, outperforming the GCMs with SCCs for annual precipitation of up to 0.87 and reduction in bias by 35%–60% in the simulation of extreme indices. Future projections revealed substantial reduction in consecutive dry days and pronounced increase in the annual total precipitation on wet days, annual count of wet days (precipitation ≥ 1 mm), and annual count of days with heavy precipitation (precipitation ≥ 10 mm) over the source region under different emission scenarios. Specifically, the latter demonstrated accelerated growth with enhanced greenhouse gas concentration, increasing by 14.5%, 39.9%, and 57.3% under shared socioeconomic pathway (SSP)126, SSP245, and SSP585, respectively, by the late 21st century. The findings of this study highlight the need for enhanced flood risk management strategies over the source region of the Yangtze and Yellow rivers to address the prospect of increased precipitation, and emphasize the critical role of coupling GCM optimization and QDM downscaling in generating reliable, high-resolution climate projections over regions of complex terrain.