Crustal Anisotropy in the Cascadia Subduction Zone - Evidence from Teleseismic Receiver Functions

Abstract:

Crustal anisotropy in the overriding North American plate along the Cascadia subduction zone is poorly constrained, mainly due to low levels of seismicity in the subducting Juan de Fuca slab that limit shear wave splitting analysis. Therefore, to investigate along-strike variations in crustal anisotropy in the Cascadia subduction zone, we compute P-to-S receiver functions at 12 broadband seismic stations using the multiple-taper correlation receiver function estimator. We observe P-to-SV converted energy on radial component receiver functions that is consistent with conversions originating at the Moho of the overriding plate and the top of the subducting slab. At stations above the mantle wedge corner we see evidence of an “inverted Moho,” supporting previous studies that suggest a highly serpentinized, low velocity mantle wedge in this region. We also observe P-to-SH conversions on transverse component receiver functions that are consistent with the presence of dipping and/or anisotropic structure in the crust and uppermost mantle. To further constrain the orientation of crustal anisotropy, we compute synthetic receiver functions using both (1) a trial-and-error forward modeling scheme and (2) a neighborhood algorithm inversion. Our results suggest that lower crustal anisotropy is highly complex and variable along-strike, but with orientations reflective of E-W oriented subduction. This work also serves as a comparison for studies that attempt to use shear wave splitting of non-volcanic tremor to constrain crustal anisotropy in Cascadia. Preliminary results for the orientation of crustal anisotropy beneath seismic station GNW on the Kitsap Peninsula, Washington show general agreement with shear wave splitting results.