Reducing lake water-level decline by optimizing reservoir operating rule curves: A case study of the Three Gorges Reservoir and the Dongting Lake

Abstract

Water levels of a lake are critical for functions such as navigation, water supply and ecological services; however, the lake water level can be reduced significantly by upstream reservoir refill operations. In this study, the reservoir operating rule curves were redesigned to satisfy the lake’s water demand. Variations in lake water level in response to reservoir releases upstream are simulated by using a river-lake model, which is simple and less time-consuming than hydrodynamic models. In the river-lake model, inflow to the lake and outflow from the lake are derived by using a series of regression models of upstream reservoir releases. Then optimization operation model is developed to minimize the lake water-level decline with the restored water level as the benchmark, and to maximize the benefits of hydropower generation simultaneously. The NSGA-II algorithm is used to solve the optimization model, where a moving-average filter is embedded to smooth the operating rule curves. China’s Dongting Lake and its upstream reservoir, the Three Gorges Reservoir were selected as a case study. The proposed river–lake model is effective for simulating the lake water level variations, with a Nash–Sutcliffe Efficiency coefficient of 0.94. With the derived optimal operating rule curves, the lake water-level decline caused by conventional reservoir operation during the refill period is reduced by 5.0%, and the hydropower generation and hydropower reliability are improved by 3.9% and 8.3%, respectively. Therefore, the proposed method is an effective tradeoff between lake restoration and enhancing the benefits of hydropower.