A geoelectric model of the Cascadia subduction zone

Results of the EMSLAB geoelectric experiment aimed at studying the Cascadia subduction zone are discussed. A 2-D interpretation of magnetic variation and magnetotelluric data was conducted in a hypothesis-testing mode. Three hypotheses were examined: (1) fluid saturation of the lower continental crust and the absence of continental asthenosphere; (2) fluid saturation of the lower continental crust and development of a continental asthenosphere; and (3) a subvertical zone of melting penetrating the continental lithosphere. The first stage of the interpretation involved experiments on smoothed inversion with the REBOCC program. These experiments showed that, under the complex conditions of the Cascadia subduction, simultaneous inversion of TE and TM modes yields an intricate, geophysically meaningless alternation of conducting and nonconducting zones with a poor minimization of the misfit. The most interesting result was obtained from the autonomous inversion of the TE mode (tippers and longitudinal impedance phases). The resistivity section comprises three subvertical zones reaching mantle depths: the oceanic (conducting), coastal (nonconducting), and Cascadia (conducting) zones. Notwithstanding its simplified nature, this result outlines the development of oceanic and continental asthenosphere. The second stage of the interpretation employed the II2DC and IGFMT2D programs using models with a fixed geometry of blocks. The inversion was conducted on a relatively detailed grid allowing, during the misfit minimization, a free choice of crustal and mantle structures corresponding to various hypotheses on the structure of the Cascadia subduction zone. The interpretation algorithm consisted of a sequence of partial target inversions taking into account variations in the sensitivity and robustness of components of the TE and TM modes. In our opinion, this type of algorithm yields the most reliable and detailed results under conditions of the Cascadia subduction. The interpretation included four successive levels: (1) inversion of tippers; (2) inversion of longitudinal impedance phases; (3) inversion of transverse apparent resistivities and transverse impedance phases; and (4) generalization and geological interpretation. Tippers played the main role, because with decreasing frequency they become free from the distorting effect of near-surface heterogeneities. A new geoelectric model of the Cascadia subduction zone was constructed. Its main distinctive features are as follows: (1) at depths of up to 40 km, a conducting upper part of the downgoing plate containing fluids of oceanic and possibly dehydration origin is clearly identifiable; and (2) the continental section contains a conducting crustal layer and a conducting asthenosphere that are connected by a subvertical conducting zone of wet melting confined to the High Cascade volcanic arc. The reliability of the model is confirmed by tests. The model agrees well with the modern concepts of the Cascadia subduction zone.