Integrated scheduling–assessing system for drought mitigation in the river–connected lake

Abstract

Lakes are important inland surface water resources and have great influence on the ecological environment as well as the surrounding residential life. However, global lake water resources showed a depleting tendency over the past decades because of the climate change and human activities. To mitigate the drought of lakes linked to a regulated main river, this study proposes an integrated scheduling–assessing system (ISAS) based on the machine learning methodology for a large river–lake system controlled by upstream reservoirs. Closely calibrated to observational data, the ISAS was applied to the middle Yangtze River to mitigate the Poyang Lake drought. The results show that the drought situation in the downstream lake could be improved through the reservoir optimal operation. For the Poyang Lake case, the lowest lake level is not obviously improved, while the starting data of the drought could be delayed by 12, 11, and 17 days, comparing to the conventional scheme in typical dry, normal, and wet years, respectively. Moreover, the duration of the drought could be 20, 19, and 21 days less. It is illustrated that accelerating the reservoir filling speed and decelerating the emptying speed is beneficial to alleviate the drought situation of downstream river-connected lakes.