Evaluation of Fuzzy Logic in Verifying of Runoff Curve Number Estimation Based on Remote Sensing of Gorganrood Basin

The main aim of this study is to estimate the Curves Number based on remote sensing and fuzzy methods for use in precipitation-runoff modelling. For this purpose, Landsat-7 satellite images were used for the second degree basin of Gorgan River, northern Iran. By making the corrections and satellite images processing using ENVI, the NDVI index of vegetation and the type of vegetation were extracted. Then, using soil map, obtained from the National's Resources Comprehensive Plan, and land slope map the type of vegetation, the CN of the sub-basins with fuzzy logic and non-fuzzy approaches calculated. To evaluate the calculated CN, a rainfall-runoff event was selected and by simulating the rainfall-runoff using HEC-HMS model, the runoff simulation was performed. The results showed that the calculated CN involving Landsat-7, using fuzzy logic simulated the runoff more accurately than non-fuzzy. So that the coefficients of determination between the observed and simulated hydrographs, obtained from non-fuzzy, fuzzy Gaussian and fuzzy trapezoidal functions were obtained equal 0.76, 0.94 and 0.84, respectively. Also, the maximum flow rate of the simulated hydrograph compared to the observed hydrograph in the fuzzy state has a difference of less than 2 m3/s compared to the non-fuzzy with a difference of 5 m3/s.