Real-time flood forecasting using time-varying parameter hydrological model: case study for Ta Trach reservoir

Flood forecasting for reservoir operation is a complex and challenging subject. It is, however, fundamental for minimizing damage and maximizing economic efficiency in reservoir management. Currently, real-time flood forecasting represents an essential trend on a global scale. This study introduces a real-time flood forecasting approach using a time-varying parameter hydrological model, applied to forecast inflows to Ta Trach reservoir in the historical flood season in 2020. The model dynamically updates parameters to reflect basin conditions in every time steps. Notably, the study’s method achieves high accuracy with Nash–Sutcliffe Efficiency values of 99.32, 95.7, and 89.14% for 1-h, 3-h, and 6-h lead times, respectively. Results surpass traditional fixed parameter and artificial intelligence models. Moreover, requiring only updated rainfall and inflow data, the model is computationally efficient, compatible with existing infrastructure in research area. With these advantages, the method presented in the study has opened a new approach and is suitable for broader applications in flood flow forecasting.