We present prestack time‐migrated multichannel seismic images along two cross‐plate transects from the Juan de Fuca (JdF) Ridge to the Cascadia deformation front (DF) offshore Oregon and Washington from which we characterize crustal structure, distribution and extent of faults across the plate interior as the crust ages and near the DF in response to subduction bending. Within the plate interior, we observe numerous small offset faults in the sediment section beginning 50–70 km from the ridge axis with sparse fault plane reflections confined to the upper crust. Plate bending due to sediment loading and subduction initiates at ~120–150 km and ~65–80 km seaward of the DF, respectively, and is accompanied by increase in sediment fault offsets and enhancement of deeper fault plane reflectivity. Most bend faulting deformation occurs within 40 km from the DF; on the Oregon transect, bright fault plane reflections that extend through the crust and 6–7 km into the mantle are observed. If attributed to serpentinization, ~0.12–0.92 wt % water within the uppermost 6 km of the mantle is estimated. On the Washington transect, bending faults are confined to the sediment section and upper‐middle crust. The regional difference in subduction bend‐faulting and potential hydration of the JdF plate is inconsistent with the spatial distribution of intermediate‐depth intraslab seismicity at Cascadia. A series of distinctive, ridgeward dipping (20°–40°) lower crustal reflections are imaged in ~6–8 Ma crust along both transects and are interpreted as ductile shear zones formed within the ridge's accretionary zone in response to temporal variations in mantle upwelling, possibly associated with previously recognized plate reorganizations at 8.5 Ma and 5.9 Ma.

We report results from a wide‐angle controlled source seismic experiment across the Juan de Fuca plate designed to investigate the evolution of the plate from accretion at the Juan de Fuca ridge to subduction at the Cascadia margin. A two‐dimensional velocity model of the crust and upper mantle is derived from a joint reflection‐refraction traveltime inversion. To interpret our tomography results, we first generate a plausible baseline velocity model, assuming a plate cooling model and realistic oceanic lithologies. We then use an effective medium theory to infer from our tomography results the extent of porosity, alteration, and water content that would be required to explain the departure from the baseline model. In crust of ages >1 Ma and away from propagator wakes and regions of faulting due to plate bending, we obtain estimates of upper crustal hydration of 0.5–2.1 wt % and find mostly dry lower crust and upper mantle. In sections of the crust affected by propagator wakes we find upper estimates of upper crustal, lower crustal, and upper mantle hydration of 3.1, 0.8, and 1.8 wt %, respectively. At the Cascadia deformation front, we find that the amount of water stored at uppermost mantle levels in the downgoing JdF plate is very limited (<0.3 wt %), with most of the water carried into the subduction zone being stored in the oceanic crust.

No abstract

The Rainbow massif, an oceanic core complex located in a nontransform discontinuity on the Mid‐Atlantic Ridge (36°N), is notable for hosting high‐temperature hydrothermal discharge through ultramafic rocks. Here we report results from a 9 month microearthquake survey conducted with a network of 13 ocean bottom seismometers deployed on and around the Rainbow massif as part of the MARINER experiment in 2013–2014. High rates (~300 per day) of low‐magnitude (average ML ~ 0.5) microearthquakes were detected beneath the massif. The hypocenters do not cluster along deeply penetrating fault surfaces and do not exhibit mainshock/aftershock sequences, supporting the hypothesis that the faulting associated with the exhumation of the massif is currently inactive. Instead, the hypocenters demarcate a diffuse zone of continuous, low‐magnitude deformation at relatively shallow (< ~3 km) depths beneath the massif, sandwiched in between the seafloor and seismic reflectors interpreted to be magmatic sills driving hydrothermal convection. Most of the seismicity is located in regions where seismic refraction data indicate serpentinized ultramafic host rock, and although the seismic network we deployed was not capable of constraining the focal mechanism of most events, our analysis suggests that serpentinization may play an important role in microearthquake generation at the Rainbow massif.