Unwrapping the triptych of climatic, social and energy-market uncertainties in the operation of multipurpose hydropower reservoirs

Abstract

The ongoing debate about hydropower and its role in the transition towards renewable energy development is strongly associated with its sustainable character, social and environmental footprint, and potential benefits. Multipurpose hydropower reservoirs play a significant dual role as they serve both water and energy uses. However, their operation and management are subject to inherently uncertain processes, future challenges, and externalities originating from climate, society and the energy market. In this context, we contribute an uncertainty-aware optimization methodology that supports operators in accounting for the cascade effects of three main uncertain drivers, i.e., rainfall, water demands, and energy scheduling. To describe climatic and energy-market uncertainties, stochastic approaches are followed to generate 2000 years of daily synthetic rainfall and electricity price data, respectively. To account for human-oriented procedures, i.e., water and energy targets, statistical analyses of historical abstractions are employed to fit copula-based relationships, in which the desirable releases for energy production depend on day-ahead electricity prices. Finally, a toolbox is established, offering insights for decision-making regarding the estimated profits, their expected changes, and the associated risk due to climate or market-oriented shifts. Our approach is demonstrated in a multipurpose reservoir in Greece, Plastiras, which is affected by rapidly increasing socioeconomic conflicts. Our findings indicate that an improved understanding of uncertainty can lead to more efficient operation policies and shield operator from misleading perceptions and their often-detrimental effects. Specifically, in the high electricity price era the expected profits are substantially unstable, since a suboptimal trade-off between irrigation and energy production may be risky and costly.