Predicting of daily Khazir basin flow using SWAT and hybrid SWAT-ANN models

Abstract

Abstract Soil and Water Assessment Tool (SWAT) and Artificial Neural Network (ANN) models have been successfully used in streamflow forecasting problems. In this research SWAT model and developed model called (hybrid SWAT-ANN) were applied for forecast daily streamflow for Khazir River at Asmawa and Khanis gauge stations which lies in Kurdistan region in Iraq. In order to calibration and validation of the models, the recorded daily flow which available for twelve years (2004–2015) was divided into two sets. The output of the SWAT model was compared with the recorded data of the streamflow, the residual time series values, which resulted from the comparison, was solved by using the ANN approach. Finally, hydrological model using combined SWAT and ANN tools were obtained, which called hybrid SWAT-ANN model. The root means square error (RMSE) and coefficient of determination (R2) for Asmawa and Khanis stations using hybrid SWAT-ANN model were found to be more efficient than SWAT model in forecasting the daily streamflow of the basin.