Electrical structure of the central Cascadia subduction zone: The EMSLAB Lincoln Line revisited

Abstract

The EMSLAB experiment was an ambitious onshore–offshore magnetotelluric (MT) transect of the Cascadia subduction zone. When completed (1985–1988), it was the largest experiment of its kind. Modeling and inversion capabilities at the time were, however, not sufficiently sophisticated to handle a fully regularized inversion of the data, including the seafloor data and bathymetric constraints, with the main final model presented based on trial and error forward modeling of the responses. Moreover, new data collected as part of the Earthscope USArray program are of higher quality due to improvements in instrument technology, and augment the original EMSLAB data set, presenting an opportunity to revisit the structure in this part of the subduction system. We have integrated the original wide-band MT data as well as several long-period stations from the original EMSLAB data set and invert these in conjunction with EMSLAB seafloor responses and new Earthscope data on land. This new composite data set has been analyzed in several ways, within a two-dimensional geometry in which conductivity is assumed to be invariant along a strike direction roughly coincident with that of the subduction zone. We have solved for fully smooth regularized models, as well as solutions that allow discontinuities in conductivity along the top surface of the descending slab. Finally, we have tested specific features in the EMSLAB model, notably a moderately shallow (~30km depth) forearc conductor. A feature similar to this shallow conductor is a consistent and required feature in our new inversion models, but the new models highlight the connection between the slab and what is interpreted to be an accumulation of aqueous fluids in the deep crust. The depth (~40km) at which the conductor intersects the slab suggests that the fluids are released by the transition of hydrous basalt to eclogite at upper greenschist facies and higher metamorphic grade. The nose of the mantle wedge has a conductivity consistent with a dry peridotite composition and thermal models of the system. At a depth of around 80km the mantle intersecting the slab shows a slight increase in conductivity. This increase is not sufficient to require the presence of melt, but a conductor indicative of melt can be inserted into the model at this depth without compromising the fit.