Comment on “Cooperative constrained inversion of multiple electromagnetic data sets” (Michael S. McMillan and Douglas W. Oldenburg, 2014, Geophysics, 79, no. 4, B173–B185)

Abstract

McMillan and Oldenburg (2014) develop a method for cooperatively inverting multiple electromagnetic data sets to produce a “consistent” 3D resistivity/conductivity model with improved resolution. Their data set includes airborne-time-domain electromagnetic data, controlled-source audio-frequency magnetotellurics (with inline electric-field and orthogonal magnetic-field measurements for frequencies ranging between 2 and 2048 Hz), and DC resistivity data. Because (1) conductivity/resistivity is known to be frequency-dependent especially in the presence of disseminated metallic particles (Wong, 1979; Marchant et al., 2013), (2) the range of frequencies investigated by McMillan and Oldenburg (2014) is rather substantial ( ∼ four orders of magnitudes), and (3) their case study includes the presence of metallic bodies (sulfide and gold mineralizations) in altered materials, it is legitimate to wonder if some of the basic assumptions used in McMillan and Oldenburg (2014) are valid or not. I believe that the approach used by McMillan and Oldenburg (2014) is attractive to simultaneously invert the resistivity obtained using different geophysical tools characterized by different, and therefore complementary, sensitivity maps. That said, it has, however, some pitfalls regarding two basic assumptions that have not been discussed by the authors and that make their most fundamental assumptions incorrect for this type of application. The first of these pitfalls is related to the frequency dependence of the electric conductivity/resistivity of rocks containing metallic ores and therefore the fact that the electric conductivity measured at different frequencies is not the same. This point has been, unfortunately, ignored in McMillan and Oldenburg (2014). I will show explicitly that the presence of 20–25 vol.% or more of metallic particles can be responsible for a difference between the DC resistivity and the high-frequency resistivity that can reach several orders of magnitude. In the case of altered rocks, the same situation arises at even lower metallic contents. The second issue is that McMillan and Oldenburg (2014) …