Salinity Properties Retrieval from Sentinel-2 Satellite Data and Machine Learning Algorithms

The accurate monitoring of soil salinization plays a key role in the ecological security and sustainable agricultural development semiarid regions. objective this study was to achieve best estimation electrical conductivity variables from salt-affected soils south Mediterranean region using Sentinel-2 multispectral imagery. In order realize goal, test carried out (EC) data collected central Tunisia. Soil leaf were measured an olive orchard over two growing seasons under three irrigation treatments. Firstly, selected spectral salinity, chlorophyll, water, vegetation indices tested experimental area estimate both EC imagery on Google Earth Engine platform. Subsequently, models calibrated by employing machine learning (ML) techniques 12 bands images. prediction accuracy assessed k-fold cross-validation computing statistical metrics. results revealed that algorithms, together with data, could advance mapping conductivity.