Is the asthenosphere electrically anisotropic?

Two techniques for resolving the conductance of an electrically conductive asthenosphere are presented. The first technique combines observatory electromagnetic data in the period range of the daily variation and longer, which provides penetration depths of 400 km and deeper, with laboratory data of the conductivity of upper mantle materials. The conductance can be estimated by this technique, but the geometry - e.g. the exact depth - is not resolved. The second technique extends the classical magnetotelluric (MT) tensor estimation and decomposition to the period range 1000-30 000 s. If an electrical asthenosphere exists, then in this period range the electromagnetic field penetrates into it (1000 s) but also through it (30 000 s). The concept of regional strike directions which has been successfully used to obtain crustal anisotropy directions is again employed at these long periods. The surprising result is that in the depth range of the asthenosphere a strong directional dependence of the conductance again occurs. Although large arrays will be necessary in order to prove that this is really conductivity anisotropy in the asthenosphere, first field results from arrays 100-200 km in extent are encouraging. The concept of a superposition of crustal and upper mantle anisotropy explains to some extent why earlier magnetotelluric determinations of the asthenosphere had little success: only in the fortunate case where a MT experiment was performed not parallel to the direction of the crustal conductor, but parallel to the asthenospheric one, is the second structure resolved. Electrical anisotropy in the uppermost mantle supports the concept of intracrystalline water, which allows for a contribution of hydrogen diffusity to the conductivity of olivine: this diffusity is highly anisotropic with respect to the crystal axis, and if the crystals are partly aligned, the conductivity measured with MT is also direction-dependent. (C) 2000 Elsevier Science B.V. All rights reserved.