The electrical conductivity distribution of the Hallandsås Horst, Sweden: a controlled source radiomagnetotelluric study

Abstract

ABSTRACTWe studied the behaviour of controlled source responses of 2½D synthetic models based on controlled source tensor magnetotelluric ( = 1–12 kHz) and radio magnetotelluric ( = 14–250 kHz) data sets collected along profiles on the Hallandsås Horst, Southern Sweden, with a view to study the depth extent of possible near‐surface fractures zones that might intersect a tunnel under construction. Based upon 2.5D forward modelling of the transverse electric (TE) and transverse magnetic (TM) mode radio magnetotelluric data, a simplified 2D model was constructed such that a dipping conductor was continued down to a depth of 140 m, which is the depth of the tunnel. Based on this extended model, we calculated both plane wave and controlled source responses for the frequency range 1–250 kHz corresponding to the source location in the study. Finally, we inverted both data sets under plane wave assumptions. When comparing synthetic controlled source and plane wave data in the frequency range 1–12 kHz, we find that the controlled source tensor magnetotelluric transfer functions generally deviate from the plane wave condition for frequencies less than 6 kHz. The phases are much more distorted than apparent resistivities. The TM mode phases, in particular, are generally distorted for frequencies less than 6 kHz. Regarding the vertical magnetic fields, we find that at all frequencies the real part of tippers are distorted by source effects but the tipper component along the profile direction is much less affected by the off‐profile lying controlled source tensor magnetotelluric source than the tipper component orthogonal to the profile direction. Based upon synthetic model studies of combined controlled source tensor magnetotelluric and radio magnetotelluric responses, we find that determinant data are less distorted than individual TE‐w and TM‐mode data. On the other hand, TE‐ and TM‐responses have greater resolving power and, when used with care (selecting data that are weakly distorted by source effects), the corresponding models can resolve the conductor down to tunnel depths.