Estimating Deep Soil Salinity by Inverse Modeling of Loop-Loop Frequency Domain Electromagnetic Induction Data in Semi-Arid Region Merguellil (Tunisia)

Accumulation of salts in irrigated soils can be detrimental not only to growing crops but also to groundwater quality. Soil salinity should be monitored reg-ularly and appropriate irrigation with the required leaching rate should be applied to avoid excessive salt accumulation in the root zone, thereby im-proving soil fertility and crop production. We combined two frequency domain electromagnetic induction (FD-EMI) mono-channel sensors (EM31 and EM38) and operated them at different heights and with different coil orientations to monitor the vertical distribution of soil salinity in a salt-affected irrigated area in Kairouan (central Tunisia). Multiple measurement heights and coil orientations were used to enhance depth sensitivity and thereby improve salinity predictions from this type of proximal sensors. The resulting mul-ti-configuration FD-EMI datasets were used to derive soil salinity information using inverse modeling based on a recently developed in-house laterally constrained inversion (LCI) approach. The collected apparent electrical con-ductivity (ECa) data were inverted to predict the spatial and temporal distri-bution of soil salinity. The results highlight several findings about the distri-bution of salinity in relation to different irrigation systems using brackish water, both in the short and long term. The expected transfer of salinity from the surface to deeper layers was systematically observed by our FD-EMI surveys. However, the intensity and spatial distribution of soil salinity varied between different crops, depending on the frequency and amount of drip or sprinkler irrigation. Furthermore, our results show that the vertical transfer of salinity is also influenced by the wet or dry season. The study provides in-sights into the effectiveness of combining two different FD-EMI sensors EM31 and EM38 for the monitoring of soil salinity in agricultural areas, contributing to the sustainability of irrigated agricultural production. The inversion ap-proach provides a more detailed representation of the of soil salinity distri-bution over spatial and temporal scales across different depths and irrigation systems compared to the classical method based on soil samples and labora-tory analysis which is a point-scale measurement. It provides a more extensive assessment of soil conditions at depths up to 4 m with different irrigation systems. For example, the influence of local drip irrigation was imaged and the history of a non-irrigated plot was evaluated, confirming the potential of this method.