Hydrological drought attribution analysis of six rivers in China by the coupled model of machine learning and hydrological model

Under the combined effects of climate change and human activities, the evolution of hydrological drought in river basins has become complex. Using monthly runoff data from six major Chinese rivers, a coupled machine learning-hydrological model was applied to simulate runoff changes, and the standardized runoff index (SRI) was used to quantify the contributions of climate and human factors to hydrological drought. The results show that: (1) the coupled model outperformed single models, especially in environmentally complex basins. (2) On a monthly scale, human activities were the primary driver of hydrological drought in the Upper Yangtze River Basin from January to March, May, September, and October. Climate change dominated the monthly drought evolution in the source regions of the Yellow River, Upper Pearl River, Middle-Upper Songhua River, Upper Huaihe River, and the source region of Lancang River in April, June–August, and October. (3) Seasonally, both factors influenced the Upper Yangtze, while climate change dominated other basins (except the Lancang in spring and summer). (4) Overall, climate change was the main driver in most basins, while human activity dominated in the Lancang River Basin.