Abstract
This paper aims to design and integrate a reservoir operation module into the SAC-SMA hydrological model to simulate flood flows in the Ban Chat reservoir region in Vietnam. The SAC-SMA model is automatically calibrated using a series of stages, including sensitivity analysis by the Sobol method and parameter optimization by the Harmony Search algorithm, while the reservoir simulation is generated based on inter-reservoir operation rules. Optimization of SAC-SMA configuration, validated through two significant flood events in 2023 and 2024 in the area, yields a Nash–Sutcliffe Efficiency value of approximately 0.70-0.79, with peak flow errors of 10.2% for the peak flood season and 16.3% for the late flood period. Additionally, the introduced reservoir operation module effectively simulates the actual management procedures of the Ban Chat reservoir. Computation of the 2024 flood event demonstrates minimal discrepancies in reservoir water levels and total outflow. Although the module showed limited accuracy in reproducing downstream releases during the 2023 flood due to complex multi-objective operations in the Red River system, the coupling demonstrates strong potential for simulating and forecasting reservoir operations.