Analysis of flood events in the Three Gorges Reservoir Region using the coupled VIC‐CaMaFlood hydrological model

Abstract

In recent years, the Three Gorges Reservoir Region (TGRR) has been increasingly affected by climate change, with frequent flood events posing challenges to regional flood control. The NO.1 flood of the Daning River in 2024 further highlights the severity of flood risks in the region. This study applied the coupled VIC‐CaMaFlood hydrological model to simulate and forecast flood events in the TGRR. The VIC model simulated runoff generation, while the CaMaFlood model performed hydrodynamic routing, simulating river streamflow, water level variations, and floodplain inundation to enhance flood forecasting accuracy. High‐resolution remote sensing data and hydrological observations were integrated for model calibration and validation, yielding a Nash‐Sutcliffe efficiency (NSE) of 0.81, a relative error (Re) of −2%, and a coefficient of determination ( R ² ) of 0.93, confirming the model's suitability for large‐scale flood simulations. The calibrated model successfully reconstructed 12 typical flood events from 2015 to 2018, with a relative error in peak streamflow ranging from −5.0% to 40.0% and a deviation in peak timing ranging from −5 to 5 days. Additionally, it accurately reproduced the NO.1 flood of the Daning River in 2024, capturing the peak flood process during July 10–11 and the secondary water level rise on July 14 caused by backwater effects from the Yangtze River. The findings demonstrate that the VIC‐CaMaFlood model provides a reliable framework for flood forecasting in large basins. However, further improvements are needed in parameter calibration, the integration of machine learning techniques, and the representation of human activities to enhance prediction accuracy. This study offers valuable insights for flood risk assessment, emergency response, and water resource management in complex hydrological systems.