Fast, laterally smooth inversion of airborne time‐domain electromagnetic data

Abstract

We present a fully developed, fast approximate method for 1D inversion of time-domain electromagnetic data. The method is applied to a helicopterborne transient electromagnetic data set from the Toolibin Lake area of Western Australia using the lateral parameter correlation method to ensure lateral smoothness of the inverted models.The method is based on fast approximate forward computation of transient electromagnetic step responses and their derivatives with respect to the model parameters of a 1D model. The inversion is carried out with multi-layer models in an iterative, constrained least-squares inversion formulation including explicit formulation of the model regularization through a model covariance matrix. The method is 50 times faster than conventional inversion for a layered earth model and produces model sections of concatenated 1D models and contoured maps of mean conductivity in depth intervals almost indistinguishable from those of conventional inversion.To ensure lateral smoothness of the model sections and to avoid spurious artefacts in the mean conductivity maps, the inversion is integrated with the lateral parameter correlation method. In this way, well determined parameters are allowed to influence the more poorly determined parameters in the survey area.Applied to the Toolibin data set, the inversion produces model sections and conductivity maps that reveal the distribution of conductivity in the area and thereby the distribution of salinity. This information is crucial for any remediation effort aimed at alleviating the salinization problems.

Seismic results as a-priori knowledge for airborne TEM data inversion — A case study

“…Ellis, 1998, Cox et al, 2010, are recently available, but for non-complex geology with adequate lateral change 1-D inversion with lateral or spatial constraints deliver appropriate quasi 3-D underground models (Christensen et al, 2009, Viezzoli et al, 2008. Constraints include boundary conditions and delimit changes of values within a defined deviation.…”

“…The inversion was done with 1-D Lateral Constraint Inversion (LCI) using a 5 layer model as well as a smooth inversion with 19 layers (Christensen et al, 2009) for both transmitter moments together (Fig. 2c).…”

“…from seismic lines, affect more TEM soundings in an indirect way than a LCI would do. Due to limited computational power the approximate forward response was used for the inversion process (Christensen et al, 2009).…”

“…A defined number of layers are generally chosen for the models in the entire area, even if there are less or more geological layers. Lateral constraints along the flight lines or spatial constraints with neighboring data points in all direction assure flat models with small lateral variation (Auken and Christiansen, 2004;Christensen et al, 2009;Viezzoli et al, 2008). The resulting resistivity depth sections and maps often allow a direct delineation of hydrological or geological regions via resistivity contrasts.…”

Spatially and Conductivity Log Constrained AEM Inversion

“…This is an approach that has been used successfully for 1D AEM inversion previously, in for example, laterally constrained inversion (Auken et al, 2005), holistic inversion (Brodie and Sambridge, 2006;Brodie, 2010) and spatially constrained inversion (Viezzoli et al, 2008). Christensen et al (2009) use a different three-stage procedure where the soundings are inverted independently in stages one and three, but in the final stage lateral constraints that were derived in the second stage by a parameter correlations technique, are imposed.…”