Authors : Seyed Hamed Shakib, Hamid Shoja Rastegari and Ali Rezvani Mahmouei
Abstract: Rainfall-Runoff modeling is considered as one of the major hydrologic processes and it is essential for water resources management. In this study, runoff by the corresponding precipitation was forecasted using the ANFIS (Adaptive Neuro-Fuzzy Inference System) in Chehel-Chai basin, Golestan province, Iran. Given that the aim of this study is to predict the runoff using its corresponding precipitation and also the stations had a proper spatial and height distribution toward the basin, this basin was used in order to obtain the daily rainfall by Digital Elevation Model (DEM). Using topographic maps, the DEM Model was prepared in GIS Software and using the model, the physiographic parameters of the basin was accurately calculated. Using the DEM, digital model of the precipitation was prepared in SPSS Software. In these systems three different input combination including the same day rainfall, same day and the day before rainfall, same day and the day before and two days before rainfall were used. Due to lack of other hydrological data in this basin stations, only the precipitation and its other combinations were selected as inputs so the basin runoffs could be forecasted using a single input in ANFIS. Therefore in the ANFIS network, different types of functions were used; Gaussian, triangular and bell. The results show that the best simulation occurred in a triangular function with a combination of input with two delays. In the final step the rainfall-runoff data were classified into 10 deciles. The results show that in most deciles the model has a high performance. But the ANFIS Model was not effective in forecasting the runoff in high deciles due to the non-repetition of the model parameters of the extent year in the network education and calibration period.
Seyed Hamed Shakib, Hamid Shoja Rastegari and Ali Rezvani Mahmouei, 2016. Use of ANFIS for Rainfall-Runoff Predictions (Case Study: Chehel-Chai Watershed, Golestan Province, Iran). Journal of Engineering and Applied Sciences, 11: 3185-3192.