In this study, it has been investigated whether the SSTs of Mediterranean Sea and Persian Gulf and combination of them are applicable and effective variables as predictors for operational streamflow forecasting, regionally, in Karkheh basin or not. For this goal, the singular value decomposition (SVD) method has been used to determine the effective nodes of Mediterranean Sea and Persian Gulf on the climate of the subbasins of this basin and to produce the most correlated time series of SST with the streamflow of each subbasin. In this research, the best predictors have been detected from several combinations of the appropriate predictors based on the cross-validation analysis for the results of the Generalized Regression Neural-Network model. Results show that autumn SST of Mediterranean Sea and Persian Gulf is prominently effective variables for forecasting the streamflow in all subbasins of Karkheh basin in April and May, respectively. Summer SST of Persian Gulf has been detected as an effective predictor for streamflow forecasting in April and May in snowy regions. Moreover, the results demonstrate that the combination of Mediterranean Sea SST and Persian Gulf SST affects the streamflow in almost all the regions of the basin in April, while the streamflow in May is affected only by Persian Gulf SST. In addition, the north and west regions of Karkheh basin (Garsha and Seimareh subbasins) as well as the east and south regions of this basin (Tang Mashooreh and Karkheh subbasins) have similar pattern of the best predictors for operational streamflow forecasting in all the spring months.

Keywords