Mapping annual soil loss in the southeast of Peshawar basin, Pakistan, using RUSLE model with geospatial approach

Abstract

Soil erodibility is challenging for agricultural soils, sustainable farming systems and water-food security in developing countries. Regional scale precise estimation of soil erosion is vital to save fertile soils. For this purpose, geospatial technology is the researchers’ top choice. The Revised Universal Soil Loss Equation (RUSLE) has the potential plausible estimates of soil erosion in the region. The study area comprises parts from two basins (Peshawar and Nizampur) having distinct land surface characteristics. This difference made the applicability of RUSLE model challenging. The influencing factors were combined using weighted sum in GIS. The study area was categorized into five soil erosion zones: very low (0–0.1 t ha−1 year−1), low (0.1–0.50 t ha−1 year−1), moderate (0.50–0.7 t ha−1 year−1) high (0.7–1 t ha−1 year−1), and very high (>1 t ha−1 year−1). The 1,748 km2 (174800 hectares) study region has an annual soil loss of 0.1 to > 1.0 t ha−1 year−1. The overall accuracy of the estimated soil erosion was found to be 90% with Kappa coefficient of 87.25% and receiver operating characteristics area under the curve (ROC-AUC) score of 0.819. This study’s will help to address the climate change-induced challenges of soil erosion in agricultural lands.

Keywords

• GIS
• remote sensing
• RUSLE
• soil loss
• weighted sum