Effects of instrument orientation on small-loop electromagnetic induction surveys of localized 2D conductive targets

Frequency-domain electromagnetic induction (EMI) systems, composed of two coplanar small coils separated by a fixed distance (EMI or SLEM), enable the rapid detection of a great variety of near-surface structures. One coil generates a controlled, primary magnetic field and the other records the variations of the induced field while the instrument is moved over the studied area. The most usual acquisition configuration corresponds to horizontal coils, with the instrument axis parallel to the prospection lines. Usually, the interpretation is based on the direct visualization of the plan-views of the data measured at each frequency. In addition, to characterize the subsoil structure in-depth, 1D inversion methods are generally applied. The aim of this work is to analyse how the system orientation affects the ability of the method to detect localized, 2D conductive structures, buried at shallow depths, and the possibility of adequately characterizing these targets through 1D inversions. We performed a survey at a test site that contains two known structures of this type, buried in almost perpendicular directions. We performed parallel prospection lines in the direction of each structure, employing, aside from the usual configuration described before, other configurations that included horizontal and vertical coils, with the instrument axis parallel and perpendicular to the lines. For comparison, we also performed a geoelectric dipole–dipole line crossing one of the targets. The features of the anomalies observed in the graphs of the EMI apparent conductivity data strongly depend on the instrument orientation. In the horizontal coil configurations, a decrease of the apparent conductivity is observed just over the targets. Besides, each vertical configuration practically detects only the target aligned with the plane of the coils, as an important positive anomaly. Through numerical simulations, performed using a 2D forward modelling method, we demonstrate that these features are indeed 2D effects associated with the localized character of the studied conductive objects. Then, we applied to the data a 1D inversion method and drawing together the results generated pseudo 3D models of the subsoil. We found that the models obtained for the vertical coil configurations provide better results. They detect the targets as conductive structures and provide a rather good estimation of their depths. Finally, we compare the EMI results with the image obtained from the 2D inversion of the geoelectrical data and analyse the causes of the observed differences