عنوان مقاله [English]
One of the most important factors in sustainable development of watersheds is suitable water resources availability in terms of quantity and quality. This study considered artificial neural network (ANN), wavelet analysis and ANN combination (WANN), multi linear regression (MLR) models for monthly Sulfate (SO4) and Chloride (Cl) modeling in the Dez River Bamdezh station and investigates the effects of data preprocessing on model performance using discrete wavelet. For this purpose, wavelet analysis and ANN model, observed time series of river discharge and SO4 and Chloride were decomposed at different scales by wavelet analysis. Then, total effective time series of discharge and SO4(Cl) were imposed as inputs to the neural network model for prediction of SO4(Cl) in one-month- ahead. Results showed that the WANN model performance was better in prediction rather than the ANN and multi linear regression models. The wavelet analysis model produced reasonable predictions for the extreme values. This model dropped the mean absolute percentage error for the multi linear regression model and the ANN model for the Chloride modeling from 0.84 and 0.64, respectively, to 0.52, and SO4 modeling from 1.7 and 0.95, respectively, to 0.63. Furthermore, the model could be employed to simulate hysteresis phenomenon for SO4 modeling, while multi linear regression method is incapable in this event.