Landscape-scale mapping of soil salinity with multi-height electromagnetic induction and quasi-3d inversion in Saharan Oasis, Tunisia

Abstract

Approximately half of Tunisia's irrigated area suffers from soil salinization due to arid and semi-arid conditions, irrigation with low-quality water, shallow and saline groundwater, and poor drainage. To manage salt-affected soils, efficient assessment methods are needed to map soil salinity in irrigated lands and to evaluate the performance of crop, water, and soil management strategies. In this study, apparent soil electrical conductivity (ECa) data were collected using a non-invasive electromagnetic induction (EMI) instrument, EM38, at different dipole orientations and heights from the soil surface, across the large agricultural area of Fatnassa Saharan (an oasis in southern Tunisia). Using a quasi-3d (q-3d) inversion algorithm, the spatial distribution of true soil electrical conductivity (σ) was estimated and the potential for prediction of the electrical conductivity of the saturated soil paste extract (ECe) from inverted ECa data was evaluated, by establishing a linear regression (LR) between σ and ECe collected at 51 reference points. The results indicate that the q-3d inversion of ECa measured at 0 and 0.2 m heights gives the best coefficient of determination (R2) between σ and ECe (R2 = 0.61). Calibration procedure was performed afterward to develop a single calibration equation and an independent validation dataset to evaluate the robustness of the obtained calibration model. The calibration-validation procedure shows moderate strength (R2 = 0.56), good accuracy with the root mean square error (RMSE) of 2.94 dS/m, small bias with mean error (ME) of − 0.56 dS/m, and an acceptable Lin’s concordance correlation coefficient (LCCC) of 0.80. Our study reveals the applicability of using the proposed methodology, based on inversion of ECa data collected at multiple heights and orientations to develop a single regional calibration and to efficiently estimate ECe at any depth across the irrigation district. Also, in the absence of recent devolved multicoil EMI instruments, ECa data measured using a single coil instrument but at different heights can provide sufficient data, required for inversion modelling and quantitative investigation of soil properties. The prediction results reveal that the soil of the oasis is moderately to severely saline which may adversely affect agricultural productivity and the sustainability of crop production.