Mapping of Potential Soil Erosion Hazard using Fuzzy Logic in Gharnave Waetershed, Golestan Province

The assessment of soil erosion hazard can identify critical regions in a watershed and prioritize management plans and soil conservation measures. This study aims to present a new approach for mapping potential soil erosion hazard using fuzzy logic and GIS in the Gharnave watershed, Golestan Province. Three main factors affecting soil erosion processes including slope (%), soil erodibility and rainfall erosivity were prepared as input layers of the fuzzy model. The slope factor map (%) was obtained from a Digital Elevation Model (DEM) with a spatial resolution of 30m, provided by ASTER global DEM data. The rainfall erosivity factor raster map (R) was generated using interpolated annual R values at 14 stations and utilizing the Ordinary Kriging (OK) Geostatistical method and Gaussian semi-variogram model. Then, the fuzzy maps of the three layers were made by defining appropriate membership functions. These maps are combined by applying fuzzy rules in the MATLAB software. This led to a fuzzy based soil erosion hazard map. In addition to that, a potential soil hazard map using the RUSLE model was created as a reference map for the accuracy assessment purpose. The accuracy assessment undertaken between the two maps showed an overall accuracy of 58% and a Kappa coefficient of 0.76 for the very high (VH) hazard class. This implies that there is an acceptable agreement between the results obtained from fuzzy logic and RUSLE models.