Statistical evaluation of sentinel-2 spectral indices as environmental indicators of shallow soil resistivity for soil corrosion assessment in a semi-arid region of northwestern Nigeria

Abstract Soil corrosion poses a significant threat to buried engineering infrastructure, particularly in semi-arid environments where variations in soil properties and surface environmental conditions influence electrochemical processes. This study integrated Sentinel-2-derived environmental indices with shallow-layer electrical resistivity and soil pH measurements to evaluate environmental factors associated with soil corrosion susceptibility in Sabon Gida, Katsina State, northwestern Nigeria. Sentinel-2 Level-2 A imagery acquired during the 2026 dry season was processed to derive the Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Bare Soil Index (BSI), and Salinity Index (SI). Spectral index values were extracted at 44 Vertical Electrical Sounding (VES) locations, where shallow-layer resistivity and soil pH data were available. Descriptive statistics, Pearson correlation analysis, and multiple linear regression were employed to investigate the relationships between the remotely sensed environmental variables and field-measured soil properties. The results indicate that the study area is characterized by sparse vegetation cover, low surface moisture, limited bare-soil exposure, and relatively higher salinity-index values that may indicate spatial variations in salinity-related environmental conditions during the dry-season image acquisition period. Strong correlations were observed among the Sentinel-2-derived indices, whereas only weak relationships were found between the spectral indices and shallow-layer resistivity. Multiple linear regression analysis yielded an R² value of 0.197, indicating that approximately 19.7% of the variation in soil resistivity was explained by the model. Although the salinity index was identified as the only statistically significant predictor, the overall regression model was not statistically significant (F = 1.867, p = 0.123). These findings suggest that the remotely sensed variables provide only limited explanatory power with respect to shallow soil resistivity. Nevertheless, the results indicate that Sentinel-2-derived environmental indices provide useful information for characterizing surface environmental conditions associated with soil corrosion but should be used as complementary indicators rather than direct substitutes for field-based geophysical investigations. The study demonstrates that integrating remote sensing with geophysical measurements provides a more robust framework for preliminary soil corrosion susceptibility assessment and offers valuable support for infrastructure planning and environmental monitoring in data-scarce semi-arid regions.