Risk-Informed Multiobjective Optimization of Reservoir Operation

Droughts present persistent and severe challenges to the security of regional water supplies, particularly in arid and semiarid regions such as northern China. Traditional reservoir operation models that prioritize water supply reliability or economic efficiency often fail to adequately address the risks posed by extreme drought events. In this study, we develop a novel risk-informed multiobjective reservoir operation model that incorporates three key performance indicators: reliability, resilience, and vulnerability (RRV). This model aims to improve drought response and enhance the overall stability of the water supply system. It is applied to a multisource water supply system composed of the Nierji Reservoir and various water-user sectors. Unlike traditional models, this approach explicitly balances the trade-offs among supply reliability, recovery capability, and water shortage during drought periods. Comparative analyses with conventional strategies (CSs) under both a six-year consecutive dry period and a representative single dry year demonstrate the superior performance of the RRV-based model in drought management. Specifically, the model reduces the average supply disruption duration from 8–10 to 4–6 ten-day intervals, increases water supply reliability to 90%, decreases the maximum single-event shortage depth to 22 × 106 m3, and lowers the average water shortage to 221 × 106 m3. Agricultural water shortages are reduced, although slight increases occur in other sectors. The results highlight resilience as the most influential objective in the model, and its inclusion or exclusion can be adjusted based on different drought response priorities. This study presents a novel and adaptive framework for reservoir operation under drought conditions, offering practical implications for improving the resilience and efficiency of regional water resource systems in the context of climate change.