Soil Salinity Prediction and Its Severity Mapping Using a Suitable Interpolation Method on Data Collected by Electromagnetic Induction Method

Abstract

Salt mining and shrimp farming have been practiced in the Non Thai district and the surrounding areas for more than 30 years, creating saline soil problems. To solve the soil salinity problem, soil salinity prediction and mapping utilizing the electromagnetic induction method (EMI) and spatial interpolation methods were examined in the Non Thai district, Nakhon Ratchasima province, Thailand. The research objectives were (1) to predict soil salinity using spatial interpolation methods and (2) to identify a suitable spatial interpolation method for soil salinity severity mapping. The research methodology consisted of five steps: apparent electrical conductivity (ECa) measurement using an electromagnetic induction (EMI) method; in situ soil sample collection and electrical conductivity of the saturated soil paste extract (ECe) measurement; soil electrical conductivity estimation using linear regression analysis (LRA); soil salinity prediction and accuracy assessment; and soil salinity severity classification and overlay analysis with relevant data. The result of LRA showed a strong positive relationship between ECe and ECa. The correlation coefficient (R) values of a horizontal measuring mode (HH) and a vertical measuring mode (VV) were 0.873 to 0.861, respectively. Four selected interpolation methods—Inverse Distance Weighting (IDW), Ordinary Kriging (OK), Ordinary CoKriging (OCK) with soil moisture content, and Regression Kriging (RK) without covariable factor—provided slightly different patterns of soil salinity prediction with HH and VV modes. The mean values of the ECe prediction from the four methods at the district level varied from 2156.02 to 2293.25 mS/m for HH mode and from 2377.38 to 2401.41 mS/m for VV mode. Based on the accuracy assessment with the rank-sum technique, the OCK is a suitable interpolation method for soil salinity prediction for HH mode. At the same time, the IDW is suitable for soil salinity prediction for the VV mode. The dominant soil salinity severity classes of the two measuring modes using suitable spatial interpolation methods were strongly and very strongly saline. Consequently, the developed research methodology can be applied to conduct soil salinity surveys to reduce costs and save time in other areas by government agencies in Thailand. Nevertheless, to apply the EMI method for soil salinity survey, the users should understand the principle of EMI and how to calibrate and operate the EM device properly for accurate ECa measurement.