Soil erosion susceptibility mapping of Hangu Region, Kohat Plateau of Pakistan using GIS and RS-based models

Soil erosion is a crucial geo-environmental hazard worldwide that affects water quality and agriculture, decreases reservoir storage capacity due to sedimentation, and increases the danger of flooding and landslides. Thus, this study uses geospatial modeling to produce soil erosion susceptibility maps(SESM) for the Hangu region, Khyber Pakhtunkhwa(KPK), Pakistan. The Hangu region, located in the Kohat Plateau of KPK, Pakistan, is particularly susceptible to soil erosion due to its unique geomorphological and climatic characteristics. Moreover, the Hangu region is characterized by a combination of steep slopes, variable rainfall patterns, diverse land use, and distinct soil types, all of which contribute to the complexity and severity of soil erosion processes. These factors necessitate a detailed and region-specific study to develop effective soil conservation strategies. In this research, we detected and mapped 1013 soil erosion points and prepared 12 predisposing factors(elevation, aspect, slope, Normalized Differentiate Vegetation Index(NDVI), drainage network, curvature, Land Use Land Cover (LULC), rainfall, lithology, contour, soil texture, and road network) of soil erosion using GIS platform. Additionally, GIS-based statistical models like the weight of evidence(WOE) and frequency ratio(FR) were applied to produce the SESM for the study area. The SESM was reclassified into four classes, i.e., low, medium, high, and very high zone. The results of WOE for SESM show that 16.39%, 33.02%, 29.27%, and 21.30% of areas are covered by low, medium, high, and very high zones, respectively. In contrast, the FR results revealed that 16.50%, 24.33%, 35.55%, and 23.59% of the areas are occupied by low, medium, high, and very high classes. Furthermore, the reliability of applied models was evaluated using the Area Under Curve(AUC) technique. The validation results utilizing the area under curve showed that the success rate curve(SRC) and predicted rate curve(PRC) for WOE are 82% and 86%, respectively, while SRC and PRC for FR are 85% and 96%, respectively. The validation results revealed that the FR model performance is better and more reliable than the WOE.