Estimating Soil Properties Distribution at a Restored Wetland Using Electromagnetic Imaging and Limited Soil Core Samples

Electromagnetic imaging (EMI) provides an efficient approach for characterizing variations in soil physicochemical properties at a high spatial resolution. While EMI has been widely used to estimate variations in soil properties in agricultural, geotechnical, and contaminated sites, limited applications have been reported for wetlands. This study assesses the use of EMI for estimating soil property distributions at a restored wetland in northwestern Ohio, USA. We acquired spatial distribution of soil apparent electrical conductivity (ECa) and apparent magnetic susceptibility (MSa) via EMI over a 162,000 m² restored wetland using an EM-38-MK2 instrument towed behind a utility terrain vehicle equipped with a differential ground positioning system. We collected twenty-two undisturbed soil samples and analyzed them in the laboratory for soil moisture (SMC), organic matter (SOM), porosity, bulk density, and texture. A least square linear regression model was used to compare the correlation between each soil property with measured ECa and MSa and subsequently, ECa was used to predict the distribution of SMC and SOM using the statistical model validated using the leave-one-out technique. We observed strong correlations between soil texture, SMC, and SOM, with ECa; with SOM showing a slightly dominant control. This study shows that ECa can predict the distribution of SMC and SOM in wetland soils to an accuracy of ~ 67–70% for these datasets. The spatial ECa patterns matched the USDA soil map for the site. This study validates the potential of extending EMI for characterizing wetland soil properties, improving sampling plans, and interpolating soil property estimates to unsampled regions.