We use surface wave measurements to reveal anisotropy as a function of depth within the Juan de Fuca and Gorda plate system. Using a two-plane wave method, we measure phase velocity and azimuthal anisotropy of fundamental mode Rayleigh waves, solving for anisotropic shear velocity. These surface wave measurements are jointly inverted with constraints from shear wave splitting studies using a Markov chain approach. The resolved structure is consistent with previous SKS studies, but our inversions provide key missing information that holds clues to the vertical distribution of strain and spreading center processes. Anisotropy of the Juan de Fuca plate interior is strongest (~2.4%) in the low velocity zone between ~40-90 km depth, with ENE direction driven by relative shear between plate motion and mantle return flow from the Cascadia subduction zone. In disagreement with measurements, weak (~1.1%) lithospheric anisotropy in Juan de Fuca is highly oblique to the expected ridge-perpendicular direction, perhaps connoting complex intra-lithospheric fabrics associated with melt and/or off-axis downwelling. In the Gorda microplate, strong shallow anisotropy (~1.9%) is consistent with inversions and aligned with spreading, and may be enhanced by edge-driven internal strain. Weak anisotropy with ambiguous orientation in the low velocity zone can be explained by Gorda’s youth and modest motion relative to the Pacific. Deeper (≥90 km) fabric appears controlled by regional flow fields modulated by the Farallon slab edge: anisotropy is strong (~1.8%) beneath Gorda, but absent beneath the Juan de Fuca, which is in the strain shadow of the slab.

Seismic deployments in the Alaska subduction zone provide dense sampling of the seismic wavefield that constrains thermal structure and subduction geometry. We measure P and S attenuation from pairwise amplitude and phase spectral ratios for teleseismic body waves at 206 stations from regional and short-term arrays. Parallel teleseismic travel-time measurements provide information on seismic velocities at the same scale. These data show consistently low attenuation over the forearc of subduction systems and high attenuation over the arc and backarc, similar to local-earthquake attenuation studies but at 10´ lower frequencies. The pattern is seen both across the area of normal Pacific subduction in the Cook Inlet, and across the Wrangell Volcanic Field where subduction has been debated. These observations confirm subduction-dominated thermal regime beneath the latter. Travel times show evidence for subducting lithosphere much deeper than seismicity, while attenuation measurements appear mostly reflective of mantle temperature less than 150 km deep, depths where the mantle is closest to its solidus and where subduction-related melting may take place. Travel times show strong delays over thick sedimentary basins. Attenuation signals show no evidence of absorption by basins, although some basins show signals anomalously rich in high-frequency energy, with consequent negative apparent attenuation. Outside of basins, these data are consistent with mantle attenuation in the upper 220 km that is quantitatively similar to observations from surface waves and local-earthquake body waves. Differences between P and S attenuation suggest primarily shear-modulus relaxation. Overall the attenuation measurements show consistent, coherent subduction-related structure, complementary to travel times.

Small-scale convection beneath the oceanic plates has been invoked to explain off-axis non-plume volcanism, departure from simple seafloor depth-age relationships, and intraplate gravity lineations. We deployed thirty broadband OBS stations on ~40 Ma seafloor in the equatorial Pacific, in a region notable for gravity anomalies measured by satellite altimetry elongated in the direction of plate motion. P-wave teleseismic tomography reveals alternating upper mantle velocity anomalies on the order of ±2%, oriented parallel to the gravity lineations. These features, which correspond to ~300-500 ˚K lateral temperature contrast, and possible hydrous or carbonatitic partial melt, are strongest between 150 and 260 km depth, indicating rapid vertical motions through a low-viscosity asthenospheric channel. Coherence and admittance analysis using new multibeam bathymetry soundings substantiates the presence of asthenospheric density variation, and forward modelling predicts gravity anomalies that qualitatively match observed lineations. This study provides observational support for small-scale convective rolls beneath the oceanic plates.

The East African Rift System provides a rare location in which to observe a wide scope of rifting states. Well-defined active narrow rifting in the Main Ethiopian Rift (MER) transitions to incipient extension and eventually pre-rifted lithosphere through the northwestern flank of the Ethiopian Plateau (EP). Although the MER is well studied, the off-axis region has received less attention. We develop Rayleigh wave phase velocity maps, Ps receiver functions, and H-κ; stack surfaces, and jointly invert these data using a trans-dimensional, hierarchical Bayesian inversion algorithm to create shear velocity profiles across the MER and EP. All shear velocities observed are slower than the PREM global average, a reflection of the elevated temperatures that persist from plume impingement. In the EP, we find a shallow mantle slow shear velocity lineament parallel to the MER axis, amidst otherwise faster shear velocities. The crust is shallow in the MER, but also in the northwestern EP flank, with thicker crust found throughout the plateau caused by crustal underplating and flood basalt emplacement. Shear velocities more reduced than the already low regional average, in concert with surficial volcanic features, geodetic observations, and slow P- and S-wave anomalies, support off-axis extension in the Ethiopian plateau, requiring reevaluation of the localization of continental breakup in the narrow MER.

Little has been seismically imaged through the lithosphere and mantle at rifted margins across the continent-ocean transition. A 2014-2015 community seismic experiment deployed broadband seismic instruments across the shoreline of the eastern North American rifted margin. Previous shear-wave splitting along the margin shows several perplexing patterns of anisotropy, and by proxy, mantle flow. Neither margin parallel offshore fast azimuths nor null splitting on the continental coast obviously accord with absolute plate motion, paleo-spreading, or rift-induced anisotropy. Splitting measurements, however, offer no depth constraints on anisotropy. Additionally, mantle structure has not yet been imaged in detail across the continent-ocean transition. We used teleseismic S, SKS, SKKS, and PKS splitting and differential travel times recorded on ocean-bottom seismometers, regional seismic networks, and EarthScope Transportable Array stations to conduct joint isotropic/anisotropic tomography across the margin. The velocity model reveals a transition from fast, thick, continental keel to low velocity, thinned lithosphere eastward. Imaged short wavelength velocity anomalies can be explained by edge-driven convection. We also find layered anisotropy. The anisotropic fast polarization is parallel to the margin within the asthenosphere. This suggests margin parallel flow beneath the plate. The lower oceanic lithosphere preserves paleo-spreading-parallel anisotropy, while the continental lithosphere has complex anisotropy reflecting several Wilson cycles. These results demonstrate the complex and active nature of a margin which is traditionally considered tectonically inactive.