Spatial Modeling of Gully Erosion Using Different Scenarios and Evidential Belief Function in Maharloo Watershed, Iran

The main purpose of the present study was to model gully erosion susceptibility by Evidential Belief Function (EBF) data-driven technique via different scenarios in Maharloo Watershed, Fars Province, Iran. So, according to extensive field surveys, the locations of the head cut, end cut, and also the boundary of the gully locations were identified and the gully inventory map was prepared. Then, different spatial layers such as: elevation, slope degree, plan curvature, TWI, distance from rivers, distance from roads, drainage density, slope aspect, lithology, annual mean rainfall, NDVI, land use, and soil characteristics (pH, clay percent, electrical conductivity (EC), and silt percentage), were identified as effective factors on the occurrence of gullies and their maps were prepared and classified in the GIS software. In the next stage, the correlation among each agent and gully erosion positions was considered using the EBF algorithm. Finally, gully erosion spatial maps were prepared and evaluated using ROC curve. The accuracy of the maps prepared using EBF algorithm by three scenarios was 0.833, 0.756, and 0.809 for the head cut, end cut, and polygon of gully locations, respectively.