Prediction of River Flow Using Time Series and Comparison with Different Methods

Maintaining and improving the quality of living standards of human life on the water resources is most important. It is necessary to make a good planning to protect, develop and use of water resources. The most important step is to determine the existing potential of water resources to use and the potential to likely have in the future. Some mathematical prediction methods have been used to determine expected potential to have in the future. In this study, mathematical models using. Daily flow data of Columbia river, which was measured in the USA between 1950 and 1960 were formed. In the stage of modelling Adaptive Network Based Fuzzy Logic Inference System Method(ANFIS), Artificial Neural Network Nonlinear Autoregressive Models (NAR) and Autoregressive Moving Average models were used. Models are tested by statistical criteria and Fuzzy Logic Inference System model prediction was found better than NAR and ARIMA methods