3D Inversion of Helicopter-borne Electromagnetic Data - A Cut-&-Paste Strategy

We present a cut-&-paste strategy for the 3D inversion of helicopter-borne frequency-domain electromagnetic data. Standard interpretation procedures often involve laterally constrained stitched 1D inversion techniques to create pseudo-3D models that are largely representative for smoothly varying conductivity distributions in the subsurface. Pronounced lateral conductivity changes may, however, produce significant artefacts that can lead to serious misinterpretation. Still, 3D inversions are numerically very expensive. Our approach therefore restricts the full 3D inversion to those parts of the survey where the 1D inversion actually fails. Using a cut-&-paste technique, those regions are extracted, separately inverted in 3D, and finally re-introduced into the original model. We apply a Gauss-Newton inversion scheme using a staggered-grid finite-difference forward operator that takes into account the dielectric response. An explicit representation of the Jacobian matrix serves for optimum computational performance. We introduce a three-way tensor quantity which facilitates the matrix assembly of the forward operator as well as the efficient calculation of the Jacobian. Finally, we deliver the proof of concept for the inversion using a synthetic and a field data set from the Cuxhaven tunnel valley in Germany.