Estimation, forecasting and extrapolation of river flows by artificial neural networks

Cigizoglu H. K.

HYDROLOGICAL SCIENCES JOURNAL-JOURNAL DES SCIENCES HYDROLOGIQUES, vol.48, no.3, pp.349-361, 2003 (SCI-Expanded) identifier identifier


River flow forecasting and estimation provide basic information on a wide range of problems related to the design and operation of river systems. In this study the applicability of artificial neural networks (ANNs) to forecasting, estimation and extrapolation of the daily flow data belonging to the rivers in the East Mediterranean region of Turkey was investigated. Throughout the study a multi-layer perceptron network was used as the ANN structure. In the forecasting part of the study predictions one day and six days ahead were investigated. The extrapolation ability of ANNs, the prediction beyond the calibration range, was the next concern of the study. The ANNs were then implemented to investigate their generalization ability, i.e. the prediction of a different time series with the trained model. River flow estimation using data from nearby stations was the final application of the ANNs. From the graphs and statistics it is apparent that a neural network solution can provide a tighter fit to the data than conventional models.