Main content area

Using ensemble precipitation forecasts and a rainfall-runoff model for hourly reservoir inflow forecasting during typhoon periods

Lee, Kwan Tun, Ho, Jui-Yi, Kao, Hong-Ming, Lin, Gwo-Fong, Yang, Tsun-Hua
Journal of hydro-environment research 2019 v.22 pp. 29-37
hydrograph, hydrologic models, hydrometeorology, rain, runoff, typhoons, uncertainty, Taiwan
This study proposes an integrated hydrometeorological system combining a fully physically based rainfall-runoff model (i.e., kinematic-wave-based geomorphologic instantaneous unit hydrograph model, KW-GIUH) with a numerical weather model (i.e., Taiwan cooperative precipitation ensemble forecast experiment, TAPEX) for performing hourly reservoir inflow forecasting patterns in the Shihmen Reservoir, in Northern Taiwan. If there is an accurate reservoir inflow with enough lead time, it could help reservoir operators to efficiently operate reservoirs in both dry and wet times of the year. Five historical typhoons, having severe impacts on the study reservoir, were used for model calibration, validation, and further application. For accurate reservoir inflow forecasting, it is necessary to understand a series of uncertainties that occur in this system. Therefore, the present study assessed the forecast uncertainty of the peak inflow and cumulative reservoir inflow on the basis of several TAPEX runs. Using the average forecasting result derived from all TAPEX runs could be less uncertain than randomly choosing a run’s result. For example, the results for Typhoon Saola (2013) showed that the uncertainty in the cumulative reservoir inflow ranged −31.3% (first TAPEX run) to 27.0% (sixth run); otherwise, its average forecasting result was only −0.9%. On the basis of the analyzed typhoons, the proposed system can provide accurate forecasts of the magnitude and timing of the peak inflow when rainfall distribution is concentrated during typhoon events. The results demonstrated that the proposed hydrometeorological system can substantially represent reservoir inflow forecasting in the Shihmen Reservoir and provide valuable information for operating reservoirs.