The CAFE Experiment : a joint seismic and MT investigation of the Cascadia Subduction System

Abstract

In this thesis we present results from inversion of data using dense arrays of collocated seismic and magnetotelluric stations located in the Cascadia subduction zone region of central Washington. In the migrated seismic section, we clearly image the top of the slab and oceanic Moho, as well as a velocity increase corresponding to the eclogitization of the hydrated upper crust. A deeper velocity increase is interpreted as the eclogitization of metastable gabbros, assisted by fluids released from the dehydration of upper mantle chlorite. A low velocity feature interpreted as a fluid/melt phase is present above this transition. The serpentinized wedge and continental Moho are also imaged. The magnetotelluric image further constrains the fluid/melt features, showing a rising conductive feature that forms a column up to a conductor indicative of a magma chamber feeding Mt. Rainier. This feature also explains the disruption of the continental Moho found in the migrated image. Exploration of the assumption of smoothness implicit in the standard MT inversion provides tools that enable us to generate a more accurate MT model. This final MT model clearly demonstrates the link between slab derived fluids/melting and the Mt. Rainier magma chamber.

Diagenetic, metamorphic, and hydrogeologic consequences of hydrothermal circulation in subducting crust

“…In the simulation with hydrothermal circulation, dehydration of the subducting mantle occurs under the region of low resistivity in the mantle wedge (~20 km farther landward from the zone of mantle dehydration in the reference simulation). Thus, the cooling of the slab by hydrothermal circulation allows mineral-bound fluid in the subducting mantle to be transported deeper into the subduction zone, resulting in the location of subducting mantle dehydration (and, in turn, the flux of those fluids to the overlying subducting crust; e.g., Hermann et al, 2006;Walowski et al, 2015) that is more consistent with seismic and magnetotelluric observations and/or interpretations (McGary, 2013;McGary et al, 2014).…”

“…The seismically inferred onset of eclogitization in the subducting crust occurs at ~45 km depth (~220 km landward of the trench) (McGary et al, 2014). However, a velocity contrast across the oceanic Moho persists to more than 80 km depth, indicating that some low-velocity (probably hydrated) crustal material persists in the subducting slab to this depth (McGary, 2013). In addition, low electrical resistivities in the mantle wedge (constrained by magnetotelluric data), consistent with ~2% melt by volume (McGary et al, 2014), are likely tied to fluid sources from the subducting slab.…”

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

“…The data are equivocal on whether there is a significant conductor at greater depths around 80-100 km as seen elsewhere in Cascadia (McGary et al, 2012;McGary, 2013) and interpreted as the primary melt source from the slab: a conductor is a consistent but not a required feature of the data. The melt signal is likely obscured by the strong conductor associated with shallow fluid release.…”

“…As such, the magnetotelluric method which measures mantle conductivity is an ideal complement to seismic methods for constraining the structure and generation of fluids at subduction zones. In addition to the EMSLAB survey (Wannamaker et al, 1989), other MT studies have taken place across the Cascadia system including: across southern British Columbia (BC) (Soyer and Unsworth, 2006;Kurtz et al, 1986Kurtz et al, , 1990; through Washington state (McGary et al, 2012;McGary, 2013); and over the broader Cascadia subduction system as a result of the US Array Earthscope program. The latter has collected a fully 3D array (at 70 km spacing) of long-period MT data (Patro and Egbert, 2008;Zhdanov et al, 2012;Meqbel et al, under review).…”