Estimating average rootzone salinity from electromagnetic induction (EM-38) measurements

Abstract

Calibration relationships are required to convert electromagnetic induction (EM-38) measurements (EM, and EMh) to average rootzone salinity (EC(0-0.6m)) in order to estimate likely production losses from salinity. This paper compares two calibration approaches and discusses factors which affect the form of the calibration relationship. Also, coefficients to adjust EM readings from field temperature to 25°C are estimated by modelling the EM response to expected temperature profiles. Calibration relationships determined using multiple linear regression, (ECe(0-0.6m) = bv EM,+ bh EMh +C) have similar prediction errors as those determined by applying the modelled coefficients procedure (Slavich 1990) to estimate average rootzone ECa (ECa(o-o.~m)) from EM, and EMh (ECe(0-0.6m) = SECa(0-0.6m) + I). However, the coefficients of the ECe(0-0.6m) V S ECa(0-0.6m) equation were more readily interpreted in terms of soil texture than those of the multiple regression approach. Using temperature-corrected EM data, the relationship between the slope (S)and the mean profile saturation percentage was similar to published relationships determined using four electrode measurements on defined soil volumes and textures. The intercept term (I) appears to be influenced by field variation in water content, texture and salinity levels below 0.6 m, the depth of interest in this paper. In heavy clay soils with high water-table salinity, the calibration relationship became nonlinear at high salinity.