Reconnaissance scale mapping of salinity in three‐dimensions using EM38 and EM34 data and inversion modelling

Abstract

AbstractIn irrigated areas, the spectre of soil salinisation is ever‐present. To obtain baseline data to map salt, or the soil electrical conductivity of a saturated soil paste extract (ECe, dS m−1), we use electromagnetic (EM) data from a survey of EM38 and EM34 data with EM inversion software (EM4Soil). This is because the collected apparent electrical conductivity (ECa, mS m−1) can be inverted to estimate true electrical conductivity (σ, mS m−1) and correlated with ECe. To do this we use a quasi‐two dimensional (Q‐2D) model along a pseudo‐transect, which includes seven soil sample locations, to make a linear regression (LR) calibration between σ and ECe (i.e., ECe = 0.0122 × σ − 0.0535 [R2 = .71]). From the joint inversion of interpolated EM38 and EM34 ECa across 40,000 ha, a quasi‐three dimensional (Q‐3D) model was then used to predict ECe. The model was validated using 30 soil sample locations scattered across the area. The agreement between predicted and measured ECe was moderate (Lin's concordance = 0.59) and conclude that useful information on a reconnaissance scale can be obtained and indicates where more detailed information may be collected to confirm areas of moderate salinity. To better predict ECe, more ECa could be collected on smaller grid spacings (i.e., 0.5 km and 1 km in irrigated and dryland areas, respectively). To improve resolving depth of ECe we recommend including ECa from a single‐frequency multiple‐coil array DUALEM‐421.