Quantifying multidimensional drought propagation risks under climate change: A vine-copula Bayesian factorial analysis method

Quantitative assessment of multidimensional drought propagation risks under climate change is essential to systematically identify and predict multiple influences of various droughts on natural environment and human society. In this study, a vine-copula Bayesian factorial analysis (abbreviated as VCBFA) method is developed through incorporating vine copula, Bayesian model averaging (BMA), and factorial analysis (FA) within a general framework. VCBFA is applied to quantifying drought propagation risks from meteorological (MD) and agricultural (AD) to hydrological drought (HD) in the Aral Sea Basin (abbreviated as ASB). The results show that the VCBFA method can quantitatively assess the propagation risk among different droughts and effectively manage multi-source uncertainties under considering the climate change impact. Key findings include: (1) drought risk is more likely to propagate when MD and AD exhibit higher severity levels and longer time scales compared to HD; (2) factors such as temperature and precipitation patterns, snowpack melting rates, and hydrological flow pathways increase the susceptibility of the upper ASB to HD propagation compared to the lower ASB; (3) under SSP1-2.6 and SSP5-8.5, the propagation probabilities for ASB would be projected to increase by 0.24 and 0.25 compared to the historical levels; (4) human activities such as irrigation, land cover change result in that AD (42%) would have a higher contribution to risk propagation than MD. This research highlights the interconnectedness of MD, AD and HD, along with a stronger correlation and propagation of multiple drought types under global warming.