Inversion of Conductivity Profiles from EM Using Full Solution and a 1-D Laterally Constrained Algorithm

Abstract

In highly conductive environments the apparent electrical conductivity [Formula: see text] data generated from electromagnetic (EM) instruments are known to be non-linear. This is particularly the case when high conductivity bodies are present in the subsurface. However, little attention has been given to this issue in the research literature of the environmental and hydrological sciences. In this paper we describe the development of an inversion algorithm, which consists of a 1-D inversion with 2-D smoothness constraints between adjacent 1-D models, whereby the forward response is calculated using the full solution of the induction phenomena. The robustness of the algorithm is evaluated using [Formula: see text] data acquired from two study areas. In the first case study, [Formula: see text] data is acquired with a DUALEM-21 across a golf green in Guelph, Ontario Canada. In the second case study, a DUALEM-421 is used to collect [Formula: see text] across an irrigated field located on a clay alluvial plain of the Lower Gwydir Valley (Australia). The general patterns of modeled true electrical conductivity [Formula: see text], as achieved from our inversion algorithm with the full solution, are shown to compare favorably with the available information and existing knowledge at each site. We also find that the models calculated with the new algorithm compare favorably with those obtained using individual 1-D inversion.

Quantitative Two‐Layer Conductivity Inversion of Multi‐Configuration Electromagnetic Induction Measurements

“…Here, shit s of about 2.5 mS m −1 in the apparent conductivity base level have been observed in low conductivity environments (see also Singleton et al, 2010). At er correction of these shit s the observed results still show a shit compared to the inverted results (Santos et al, 2010b). h is indicates that the inl uence of surrounding conditions (i.e., temperature, solar radiation, condition of power supply, system up-time) on the output of the device is signii cant (Sudduth et al, 2001;Robinson et al, 2004;Abdu et al, 2007;Gebbers et al, 2009) and that the calibration of these data can still be improved.…”

“…The most commonly used forward model is the cumulative response model or local‐sensitivity model (McNeill, 1980). Due to the increased computing power, improved forward models based on the full solution of Maxwell's equation can be used (Lavoué et al, 2010; Santos et al, 2010b). Both models are used in our inversion algorithm and are briefly described.…”

“…The general shape of the synthetic models was reasonably well reconstructed, but due to the smoothed inversion, the boundaries between the layers were smeared out. A follow‐up paper by Santos et al (2010b) used an improved forward modeling tool based on the EM model and the same inversion scheme. Two practical datasets collected with multiple‐array electromagnetic instruments were inverted and showed that using the EM model returns better results when high conductivities are present (Callegary et al, 2007).…”

Inferring soil salinity in a drip irrigation system from multi-configuration EMI measurements using adaptive Markov chain Monte Carlo

“…Preliminary analysis indicated that the electromagnetic forward model, which is based on high induction number assumption, returned more reliable apparent electrical conductivity values than the standard sensitivity curves of McNeill (1980). Furthermore, increased computational power made it possible to characterize the subsurface by utilizing forward models based on the Maxwell equation (Santos et al, 2010). The effective depth of exploration is independent of EC a in a low induction number condition, whereas in high induction number condition an inverse relationship was found between the depth of exploration and EC a (Callegary et al, 2007).…”

“…Several inversion algorithms have been developed for EMI measurements to improve the resolution of subsurface features and the assessment of soil properties (Hendrickx et al, 2002;Santos et al, 2010;Triantafilis and Monteiro Santos, 2013). The majority of these inversion algorithms solve a 1-D earth model for electromagnetic wave propagation.…”