مقايسه پيش بيني دبي جريان ماهانه رودخانه كرخه با استفاده از روش هاي سري زماني و هوش مصنوعي

Comparing monthly discharge forecasting for Karkheh river by using time series and artificial Intelligent traits

Mentioning the importance of being aware of the rivers flow in future, in this research Karkheh river flow forecasted in two stations called Jelogir and Payepol by time series method with Auto - regressive moving average technique and artificial Intelligent with artificial neural network. Data set and information collected from these two hydrometric stations from 1958:10 2003:9 to periods. Data set and information of 1958:10 to 1999:9 were used to compare the methods. Data set and information of 1999:10 to 2003:9 were used to study the forecasting power. To compare the forecasting error of various methods, MAE, RMSE and MAPE criterias were used. Obtained results showed that Artificial Neural Network has lower error due to time series method in forecasting of monthly flow of Jelogir station. But comparing to change of the flow regime in Paypol station, time series method had lower error in prediction of monthly flow in this station.