Comparing resistivity models from 2D and 1D inversion of frequency domain HEM data over rough terrains: cases study from Iran and Norway

ABSTRACT

Frequency domain helicopter-borne electromagnetic (FHEM) surveys have been used as an effective tool for the exploration of underground resources for as long as airborne electromagnetics (AEM) has existed. Large FHEM data sets are commonly interpreted with very fast 1D inversion algorithms that often numerically adequately fit the data sets, even though they yield incorrect results near 2D/3D geological structures. The present study aims to compare 1D and 2D inversion algorithms when applied to the reconstruction of geologically complex regions. We have developed a 2D inversion algorithm incorporating the Levenberg–Marquardt least-squares approach regularised through spatial constraints to retrieve 2D electrical resistivity models associated with arbitrary surface topography. The approach uses a 2D finite element frequency domain solution and a tailored triangular meshing algorithm based on the Ruppert’s Delaunay refinement for the forward modelling. We illustrate how rough topographic effects obscure the FHEM response and affect recovered resistivity models through numerical experiments. We also demonstrate the influence of acquisition frequency and resistivity structure on the topographic effect. In the Appendix, we discuss the FHEM footprint concept from a 2D perspective to assess how 2D effects affect and bias 1D inversion results. Complex 2D synthetic scenarios are presented to compare 1D and 2D inversion in various settings. Two field cases from Norway and Iran are presented to show the model improvements with 2D inversion. For the Norwegian case, the 2D FHEM inversion aligns well with a model retrieved from ground-based electrical resistivity tomography. We show the bias imposed on the 2D inversion of the data set from Iran by improper system calibration.

KEYWORDS:

• AEM
• HEM
• frequency domain
• 2D
• numerical modelling
• inverse theory
• electrical properties

Acknowledgments

First and foremost, we acknowledge the Norwegian Geological Survey (NGU) for giving us access to the Byneset ERT models and FHEM data. The Geological Survey of Iran is also appreciated for providing FHEM data. Professor Esben Auken is acknowledged for his comments and permitting use to use the efficient software package AarhusInv. The authors are grateful to the Institute of Geophysics, University of Tehran, Norwegian Geotechnical Institute (NGI) and HydroGeophysics Group, Department of Geoscience, Aarhus University for all their supports. We also thank the editors and reviewers for their constructive suggestions and meaningful comments.

ORCID

Hossein Ali Ghari http://orcid.org/0000-0001-6299-976X

Behrooz Oskooi http://orcid.org/0000-0003-3065-194X