Machine-learning-constrained projection of bivariate hydrological drought magnitudes and socioeconomic risks over China

Abstract

Climate change influences the water cycle and alters the spatiotemporal distribution of hydrological variables, thus complicating the projection of future streamflow and hydrological droughts. Although machine learning is increasingly employed for hydrological simulations, few studies have used it to project hydrological droughts, not to mention bivariate risks (referring to drought duration and severity) as well as their socioeconomic effects under climate change. We developed a cascade modeling chain to project future bivariate hydrological drought characteristics in 179 catchments over China, using five bias-corrected global climate model (GCM) outputs under three shared socioeconomic pathways (SSPs), five hydrological models, and a deep-learning model. We quantified the contribution of various meteorological variables to daily streamflow by using a random forest model, and then we employed terrestrial water storage anomalies and a standardized runoff index to evaluate recent changes in hydrological drought. Subsequently, we constructed a bivariate framework to jointly model drought duration and severity by using copula functions and the most likely realization method. Finally, we used this framework to project future risks of hydrological droughts as well as the associated exposure of gross domestic product (GDP) and population. Results showed that our hybrid hydrological–deep-learning model achieved > 0.8 Kling–Gupta efficiency in 161 out of the 179 catchments. By the late 21st century, bivariate drought risk is projected to double over 60 % of the catchments mainly located in southwestern China under SSP5-85, which shows the increase in drought duration and severity. Our hybrid model also projected substantial GDP and population exposure by increasing bivariate drought risks, suggesting an urgent need to design climate mitigation strategies for a sustainable development pathway.