Defining the robust operating rule for multi-purpose water reservoirs under deep uncertainties

摘要

Depending on the concept of optimality, the decision makers responsible for the planning and management of water resources seek to maximize the optimization performance. However, the optimized solutions are vulnerable to failure because decision making in water resources management usually involves factors with deep uncertainties (e.g. runoff conditions, water demand growth, climatic forces, etc.). In this study, we further contribute to the Many-Objective Robust Decision Making framework for defining the robust operating rules of a water supply system of interest. The multi-objective optimization and uncertainty analysis tools were used to reveal the trade-offs between the competing objectives and discover the sensitive factors for the plausible states of the system, respectively. The robust operating rules are demonstrated for the Han to Wei inter-basin water transfer project, which is the most important water diversion project in Shaanxi Province, China. Results show that, although the current operating rules are optimized for the water supply system over the long term, these operating rules cannot deal with the problem of performance degradation under deep uncertainties associated with runoff conditions and water demand growth. The two uncertain factors are the sensitive factors responsible for the failure or success of the system, and robustness of the system may be achieved by reducing the effect of key uncertainties (e.g. reduced water demand growth). Furthermore, there are obvious differences in the robust operating rules across different months, with the key or sensitive months providing the uncertain ranges to likely sustain the success of the system. Finally, the successful frequency of the system derived from the most robust operating rules is 18% higher than that obtained from the current operating rules for alternative states of the world. The robust operating rules offer critical insights into the challenges posed by deep uncertainties and provide a management template for decision making on climate change and complex human activities.