Applying Fuzzy Linguistic Method to Predict River Runoff

The runoff forecasting is an important factor in river basin management. In this paper, we use fuzzy linguistic rules to solve the river runoff prediction problem. The runoff relationship between the upstream and downstream runoffs has the fuzzy property in the hydraulic engineering, and this kind of relation creates an opportunity for the prediction of downstream runoffs, which offers a foundation to the decision-making for hydraulic engineering design, operation and management. Based on the fuzzy linguistic rules existed in input and output variables, in this paper an attempt was made to construct a forecasting model with the fuzzy logic and applied the gradient descent method in optimizing model’s parameters. The modelling results in Zipingpu station show that the fuzzy logic forecasting model (FLFM) has promising potential for providing reliable water supply forecasts.