Mapping soil salinization in the Shirvan plain based on Sentinel-2 data

In this study, we used topographic indices derived from digital elevation model (DEM), SAR and vegetation indices generated by Sentinel-2 to map soil salinization in the Kurdamir, located in the central and northern Shirvan plain of Azerbaijan Republic, and evaluated the potential of multi-source sensors to predict soil salinity. This paper reflects the results of assessing the degree of salinity of Kurdamir soils based on remote sensing technologies. Classification of soils by salinity degree was performed using indices calculated from data in Sentinel-2 spectral channels. Based on a comparative analysis of machine learning and remote sensing methods, it was found that they provide similar results for soil salinity assessment, when the study area does not cover a large area of land. Using the soil electrical conductivity (EC) values of soil sampling sites as the target variable, and the optimal environmental factors as the predictive variable, we constructed three soil salinity inversion models based on classification and regression tree (CART), and extreme gradient boosting (XGBoost). This work has important implications for understanding soil salinity dynamics, and could be used in further research on land and agricultural management at the regional level.