Revealing water supply-demand responses to varying inflow frequencies in inter-basin water diversion projects using a non-iterative deterministic sampling algorithm

Study region The Eastern Route of the South-to-North Water Diversion Project (ER-SNWDP), China, the world's largest cascade pumping station system. Study focus Inter-basin water diversion projects (IWDPs) are key engineering measures for alleviating regional water scarcity, but their water supply scheduling involves a high-dimensional nonlinear optimization problem under reservoir operation constraints. This study developed a universal framework to minimize the water supply-demand gap (WSG) in IWDPs by optimizing multi-reservoir water supply scheduling. A non-iterative deterministic sampling algorithm, the Orthogonal Latin Neighborhood Algorithm (OLNA), was proposed to efficiently determine optimal scheduling strategies. This methodology was applied to the ER-SNWDP and the water supply-demand responses to varying inflow frequencies were also revealed. New hydrological insights for the region The universality and efficiency of OLNA were verified to be superior to mainstream heuristic algorithms in solving multi-year multi-reservoir scheduling problems under varying hydrological conditions. It was found that the WSG of the ER-SNWDP has the potential to be reduced by 1.28 × 10 10 m 3 under 75% inflow frequency. Escalating water demand during high inflow frequency (hydrological droughts) exacerbates supply-demand imbalances despite optimized strategies. A Limit Equilibrium Point (LEP), the transition from sufficient water supply to water shortage ( i.e. , WSG =0), was found to decrease as the inflow frequency increased but to increase with higher initial reservoir storage. These findings provide guidance for water regulation and drought-risk mitigation in IWDPs.