Developing reservoir operation strategy under the mechanism-data dual-driven framework: Coordinating the water demand characteristics of multi sectors

Reservoir operation is facing increasing challenges in balancing water demands across multiple sectors, making it crucial to develop effective operation strategy to improve the utilization efficiency of limited water resources. To address this, this study proposes a mechanism-data dual-driven framework for developing reservoir operation strategy. At the mechanistic level, considering the water demand characteristics of multi sectors, the localized AquaCrop-OSPy is coupled with the reservoir operation model to construct Model I. And the model aims to maximize agricultural yield. Simultaneously, Model II is established that the reservoir operates according to the standardized operation policy. At the data-driven level, based on diversified information, the light gradient boosting machine is used to extract reservoir operation strategy, and the shapley additive explanations (SHAP) is applied to analyze the decision-making mechanisms of the strategy. Using the Dongzhuang reservoir in the Yellow River Basin as an example. The results demonstrate that in long-term simulations, Model I can reduce water shortages in the ecological sector from 0.12 % to 0.04 %, in the domestic-industrial sector from 2.20 % to 0.64 %, and increases agricultural yield by 7.79 % compared to Model II. The extracted operation strategy extracted demonstrates significant overall benefits. Among them, agricultural yield is 2.60 % higher than Model II, but 4.82 % lower than Model I. SHAP analysis reveals that the end of month water level of the reservoir is influenced by the directional heterogeneity of multiple-factor. Among them, the water level at the beginning of month has the greatest impact, while the monthly precipitation has the least impact. • A new framework for developing reservoir operation strategy is proposed. • A reservoir operation model is constructed to coordinate multi-sector water demand. • The reservoir operation strategy that improves crop yields is developed. • The strategy is influenced by the directional heterogeneity of various factors.