1.Hunan Water Resources and Hydropower Survey,Design,Planning and Research Co,Ltd,Changsha;2.Hohai University
为揭示洞庭湖中枯水期水情变化特征及其驱动因素，采用长短期记忆神经网络模拟洞庭湖出湖流量及湖区水位，通过情境模拟开展水情变化归因分析。洞庭湖1992—2019年9—10月出湖流量大幅减少，主要受长江流量降低的影响。洞庭湖中枯水期水位主要呈下降趋势，其中9—10月平均水位在西洞庭湖、南洞庭湖降幅约1 m，在东洞庭湖降幅约2 m。地形变化对中枯水期水位主要起拉低作用，长江和流域四水流量变化在9—10月起拉低作用、在12月至次年3月起抬升作用，其中对东洞庭湖水位的影响相对更为显著。研究结果可为洞庭湖中枯水期水资源管理和湿地保护提供参考。
This study investigated the water regime changes and the associated driving factors of Dongting Lake in the normal and dry seasons. The long short-term memory network modeled the outflow and water level of Dongting Lake, and the driving factors of water regime variation were detected using scenario simulation. From 1992 to 2019, the outflow of Dongting Lake in September and October decreased dramatically, caused by the reduced flow of the Yangtze River. The water level of Dongting Lake during the normal and dry seasons mainly lowered. The average water level in September and October dropped by 1 m in West and South Dongting Lake, while it dropped by 2 m in East Dongting Lake. Morphologic changes mainly lowered the water level during the normal and dry seasons. Flow changes of the Yangtze River and the four tributaries of the Dongting Lake basin dropped the water level in September and October and elevated the water level from December to March of the following year. The impact of flow changes on the water level of East Dongting Lake was relatively more significant. The results provided implications for water resources management and wetland protection of Dongting Lake in the normal and dry seasons.