Artefacts of isotropic inversion applied to magnetotelluric data from an anisotropic Earth

Abstract

2-D isotropic approaches are standard for magnetotelluric (MT) data modelling and inversion. Unfortunately, the real subsurface structure is not isotropic everywhere and one should be aware of the possible consequences of applying an isotropic inversion to data from an anisotropic Earth. The work presented herein was motivated by discovering an unusual mid- to lower-crust conductor that appeared to be bent downward into the lithospheric mantle, compared to the neighbouring terranes, when applying routine 2-D isotropic inversion. One major difference between the terranes is the presence of the Okavango giant mafic dyke swarm (northeastern Botswana), beneath which the conductor is imaged to be in the lithospheric mantle rather than the lower crust. The limited width of the dykes makes them an anisotropic feature rather than a normal 2-D structure at MT scale. To examine the possible effects of the dykes, synthetic data were generated from an 1-D model, accounting for the dyke swarm by using an anisotropic block, and then inverted isotropically. The synthetic tests showed that the normal 2-D decomposition and strike analysis techniques are not removing these large scale anisotropic effects, and that an isotropic inversion result obtained in the presence of an anisotropic structure has to be treated with caution. The comparison of the synthetic data with the presented case history strongly suggests that the lithospheric depths of the conductor is an artefact, and it is most likely located in the lower-crust, as everywhere else in the area. Focused anisotropic inversion of the dyke swarm area supports the assumption of a crustal anisotropic structure related to the Okavango Dyke Swarm that affects the isotropic inversion results. © 2011 The Authors Geophysical Journal International © 2011 RAS.