Comparing the Performance of Fuzzy Based Models in Stream Flow Forecasting on Lighvan River

River discharge forecasting and analyzing its influencing factors is one of the most important issues in water resources management. In this research, fuzzy based models (Fuzzy Inference System (FIS) and Adaptive Neuro-Fuzzy Inference System (ANFIS)) is used for performance of river flow forecasting process. Three parameters of precipitation, temperature, and daily discharge of the Lighvanghai basin is used for daily river flow forecasting in Lighvan river. On initial preprocessing of data, the randomness of data studied by return points test. Then, for determination of optimum lags of input parameters correlogram of data was studied. Finally, prediction was performed in two parts of synthetics and discharge of previous days. Assessment of prediction results by using various values as Nash-Sutcliff value showed that ANFIS model could predict discharge of these river with high exact (CNS=0.998) and low dispersion rather than FIS model (CNS=0.993). Also, in prediction of daily discharge through assessment of two designed parts, temperature had no significant effects, but precipitation of current day is more effective than discharge of two days ago.