Application of grey model toward runoff forecasting

A reliable forecasting model is essential in real-time flood forecasting fo r reducing natural damage. Efforts to develop a real-time forecasting model over the past two decades have been numerous. This work applies the Grey m odel to forecast rainfall and runoff owing to the model's relative ability to predict the future using a small amount of historical data. Such a model significantly differs from the stochastic and deterministic models develop ed previously. Ten historical storm events from two catchment areas in nort hern Taiwan art? selected to calibrate and verify the model. Results in thi s study demonstrate that the proposed models can reasonably forecast runoff one to four hours ahead, if the Grey error prediction method is further us ed to update the output of the model.