Bivariate copula framework for drought risk assessment under climate change and practical water supply systems

Abstract This study assesses the water supply system of Seomjin River basin in South Korea under future extreme drought conditions. Hydrological modeling and probabilistic risk analysis were combined to assess drought resilience. Natural runoff was generated with the SWAT model and linked to the K-WEAP to simulate water supply–demand interactions under CMIP6 climate projections. Four management schemes were examined by varying dam operation rules (firm versus deficit release) and allocation priorities (domestic-first versus equal sharing). A bivariate copula framework was employed to characterize nonlinear reliability–vulnerability relationships, derive joint return periods using both “AND” and “OR” exceedance criteria, and evaluate conditional vulnerability under fixed reliability thresholds. Findings indicate that supply strategies yielded contrasting outcomes across sub-basins: deficit releases lowered vulnerability in some districts but intensified it elsewhere, reflecting the complexity of inter-basin transfers. “AND” analysis emphasized infrequent but high-impact compound failures relevant to infrastructure design, while “OR” analysis captured recurrent shortages suitable for early-warning applications. Conditional vulnerability further revealed trade-offs between urban reliability gains and heightened risks in transfer-dependent regions. These insights highlight the importance of flexible, region-specific allocation strategies and demonstrate the value of copula-based indicators in adaptive drought management.