Global search inversion for electromagnetic induction data using layered models

Abstract

Rigid boom electromagnetic surveys that use coil-coil configurations are often used to obtain information about the subsurface conductivity. Semi-analytic solutions help to simulate electromagnetic induction measurements for a large number of horizontally layered models, which can then be stored and used as a lookup table. This procedure is performed once and then used to find the corresponding model that produces the best data fit, eliminating the need for running numerous simulations in every minimization step of an inversion scheme for large field datasets. We apply this methodology to a numerical example and field data acquired in The Netherlands. Our results from both cases using the global search demonstrate its ability to estimate electrical conductivity distributions in two-layered models in a fast and accurate manner. Furthermore, we apply the workflow using a lookup table based on low induction number approximation-derived measurements. The outcome of implementing this methodology using the low induction number lookup table shows poor accuracy in the electrical conductivity estimations for both the numerical example and the field data in comparison to the semi-analytical approach.