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 andsustainable agricultural development of semiarid regions. The objective of this study was to achievethe best estimation of electrical conductivity variables from salt-affected soils in a southMediterranean region using Sentinel-2 multispectral imagery. In order to realize this goal, a test wascarried out using electrical conductivity (EC) data collected in central Tunisia. Soil electricalconductivity and leaf electrical conductivity were measured in an olive orchard over two growingseasons and under three irrigation treatments. Firstly, selected spectral salinity, chlorophyll, water,and vegetation indices were tested over the experimental area to estimate both soil and leaf EC usingSentinel-2 imagery on the Google Earth Engine platform. Subsequently, estimation models of soiland leaf EC were calibrated by employing machine learning (ML) techniques using 12 spectralbands of Sentinel-2 images. The prediction accuracy of the EC estimation was assessed by using kfold cross-validation and computing statistical metrics. The results of the study revealed thatmachine learning algorithms, together with multispectral data, could advance the mapping andmonitoring of soil and leaf electrical conductivity.