University of Tabriz Journal of Geography and Planning 2008-8078 Articles in Press 2026 01 03 Analysis of the Relationship Between Vegetation Indices and Soil Erosion Using the Revised Universal Soil Loss Equation (Case Study: Firozabad Watershed, Ardabil) Analysis of the Relationship Between Vegetation Indices and Soil Erosion Using the Revised Universal Soil Loss Equation (Case Study: Firozabad Watershed, Ardabil) 21054 10.22034/gp.2026.69329.3456 FA Mousa Abedini Professor, Department of Physical Geography, Faculty of Social Sciences, University of Mohaghegh Ardabili, Ardabil, Iran. AmirHesam Pasban Ph.D Student, Department of Physical Geography, Faculty of Social Sciences, University of Mohaghegh Ardabili, Ardabil, Iran, Iran Aylar Khayati Noshahr Ph.D Student, Department of Physical Geography, Faculty of Social Sciences, University of Mohaghegh Ardabili, Ardabil, Iran, Iran Journal Article 2025 09 25 The aim of the present study is to analyze the relationship between vegetation indices and soil erosion using the RUSLE model in the Firozabad watershed, Ardabil Province. For this purpose, Sentinel-2 satellite images from 2024 were employed, and six vegetation indices, including NDVI, SAVI, MSAVI, EVI, GNDVI, and VARI, were extracted. These indices were calculated using Google Earth Engine and ArcGIS, and their spatial maps were generated. At the same time, the RUSLE model was applied to estimate soil erosion intensity. Subsequently, by overlaying the layers, a soil erosion map at the watershed scale was prepared. Furthermore, in order to analyze the correlation between vegetation indices and soil erosion intensity, Pearson’s correlation and linear regression analysis were conducted in SPSS software. The results showed that soil erosion values across the watershed ranged between 2.46 and 98.73 tons per hectare per year. Pearson’s correlation analysis indicated that soil erosion had the highest positive and significant correlation with NDVI and EVI (0.38). The MSAVI and VARI indices also showed positive and significant correlations with values of 0.29 and 0.26, respectively, while SAVI demonstrated a negative and significant correlation with a value of –0.22. This pattern suggests that increasing greenness, as measured by NDVI and EVI, contrary to initial expectations, was associated with increased soil erosion. In contrast, the SAVI index, which accounts for the soil effect, reflected the protective role of vegetation cover in sparsely vegetated areas. The linear regression analysis also confirmed the correlation results; the highest coefficients of determination (R²) between soil erosion and NDVI and EVI were 0.148 and 0.146, respectively, indicating that these two indices explained soil erosion variations. The VARI and MSAVI indices, with values of 0.13 and 0.08, and the GNDVI and SAVI indices, with values of 0.070 and 0.074, showed lower explanatory power. The aim of the present study is to analyze the relationship between vegetation indices and soil erosion using the RUSLE model in the Firozabad watershed, Ardabil Province. For this purpose, Sentinel-2 satellite images from 2024 were employed, and six vegetation indices, including NDVI, SAVI, MSAVI, EVI, GNDVI, and VARI, were extracted. These indices were calculated using Google Earth Engine and ArcGIS, and their spatial maps were generated. At the same time, the RUSLE model was applied to estimate soil erosion intensity. Subsequently, by overlaying the layers, a soil erosion map at the watershed scale was prepared. Furthermore, in order to analyze the correlation between vegetation indices and soil erosion intensity, Pearson’s correlation and linear regression analysis were conducted in SPSS software. The results showed that soil erosion values across the watershed ranged between 2.46 and 98.73 tons per hectare per year. Pearson’s correlation analysis indicated that soil erosion had the highest positive and significant correlation with NDVI and EVI (0.38). The MSAVI and VARI indices also showed positive and significant correlations with values of 0.29 and 0.26, respectively, while SAVI demonstrated a negative and significant correlation with a value of –0.22. This pattern suggests that increasing greenness, as measured by NDVI and EVI, contrary to initial expectations, was associated with increased soil erosion. In contrast, the SAVI index, which accounts for the soil effect, reflected the protective role of vegetation cover in sparsely vegetated areas. The linear regression analysis also confirmed the correlation results; the highest coefficients of determination (R²) between soil erosion and NDVI and EVI were 0.148 and 0.146, respectively, indicating that these two indices explained soil erosion variations. The VARI and MSAVI indices, with values of 0.13 and 0.08, and the GNDVI and SAVI indices, with values of 0.070 and 0.074, showed lower explanatory power.