A coupled model applied to complex river–lake systems

Abstract

ABSTRACTSimulating river–lake systems while accounting for climate change and human activity is challenging, particularly when dealing with the complex interactions between rivers and lakes. In this study, we developed an integrated model coupling long short-term memory (LSTM), a two-dimensional hydrodynamic model (NewFlood), and the time-variant gain model (TVGM). We applied and validated this model using the Yangtze River–Poyang Lake system. The coupled model accurately and efficiently simulated the water levels, streamflow, and water exchange between the Yangtze River and Poyang Lake. The average Nash-Sutcliffe efficiency coefficient of the coupled model reached 0.92, indicating high simulation accuracy. The coupled model holistically accounts for the impacts of climate change and human activities on the spatiotemporal dynamics of river–lake systems, providing a robust scientific basis for quantifying the relative contributions of different factors influencing river–lake interactions.