Probing 1-D electrical anisotropy in the oceanic upper mantle from seafloor magnetotelluric array data

Electrical anisotropy in the oceanic upper mantle can only be imaged by seafloor magnetotelluric (MT) data, and arguably provides important clues regarding the mantle structure and dynamics by observational determinations. Here, we attempt to probe the electrical (azimuthal) anisotropy in the oceanic mantle by analysing recent seafloor MT array data from the northwestern Pacific acquired atop 125-145 Ma seafloor. We propose a method in which an isotropic 1-D model is first obtained from seafloor MT data through an iterative correction for topographic distortions; then, the anisotropic properties are inferred as deviations from the isotropic 1-D model. We investigate the performance of this method through synthetic forward modelling and inversion using plausible anisotropic 1-D models and the actual 3-D bathymetry and topography of the target region. Synthetic tests reveal that the proposed method will detect electrical anisotropy in the conductive upper mantle or electrical asthenosphere. We also compare the performance of the proposed scheme by using two rotational invariant impedances and two topographic correction equations. The comparison reveals that using different rotational invariants and correction equations provides relatively consistent results, but among the rotational invariants, the sum of squared elements (ssq) impedance yields better recovered results for topographically distorted data than the determinant impedance. An application of the method to seafloor MT array data sets from two areas in the northwestern Pacific reveals the possible presence of two layers of electrical anisotropy in the conductive mantle (<100 Omega-m) at depths of similar to 60-200 km. The anisotropy is estimated to be more intense in the shallower layer for both areas. On the other hand, the estimated anisotropic azimuth (defined as the most conductive direction) and the depth to the interface between the two layers are different between the two array areas separated by a small horizontal distance of similar to 1000 km in spite of their similar seafloor ages. The most conductive directions are aligned neither with the current absolute plate motion direction nor with the fastest direction of seismic azimuthal anisotropy. The inferred electrical anisotropy features may result from array-scale (similar to 1000 km) mantle dynamics, such as small-scale convection, which might affect the electrical and seismic properties differently, although there remains the possibility that some portions of these features are explained by laterally heterogeneous mantle structures.