Two Contrasting Models of Lithospheric Structure
The subduction of the Juan de Fuca plate beneath North America changes markedly along the length of the subduction zone, notably in the angle of subduction, distribution of earthquakes, volcanism, geologic and seismic structure of the upper plate, and regional horizontal stress. To investigate these characteristics, we conducted detailed density modeling experiments of the crust and mantle along two transects across the Cascadia subduction zone. One crosses Vancouver Island and the Canadian margin, and the other crosses the margin of central Oregon. Both density models were constructed independently to a depth of approximately 50 km. We gathered all possible geologic, geophysical, and borehole data to constrain the density calculations. The final densities for the Oregon and Vancouver lithosphere models were obtained from gravity inversions.
Our results confirm that the downgoing slab of the Cascadia subduction zone dips significantly steeper beneath Oregon than beneath Vancouver Island, lending support to the idea that the Juan de Fuca plate is segmented from north to south. In addition, our gravity models indicate that the mantle wedge beneath western Oregon (i.e., below the western Cascades) is lighter than the mantle beneath the Canadian continental crust. This low density agrees with the low mantle velocities observed in the mantle and the present day extensional regime of the Pacific Northwest.
A gravity low at the deformation front of the Oregon margin, absent along the Vancouver margin, can be explained by the different bathymetry of the two regions and by the depth to the top of the subducting Juan de Fuca plate. If the accretionary prisms along these profiles were modeled with equal densities, a density inhomogeneity in the lower part of the models would be necessary. Thus that the density of the accretionary prism for the Vancouver profile must be approximately 0.1-0.2 g/cm3 greater than that for Oregon. A density difference within the accretionary prisms also agrees with other data. We note that the volume of accreted sediments is approximately twice as large along the Vancouver profile than along the Oregon profile, and the prism reaches a greater depth (approximately 20 km as compared with 12 km for the Oregon profile). This implies that the sediments within the accretionary prism at Vancouver Island are at a higher metamorphic grade, and therefore have higher densities.
We find that a substantial part of the coastal gravity maxima for both lines is caused by increasing density with depth in the subducting plate. In the proposed model, the maximum possible density of the slab was used to satisfy constraints for the average density of the near coastal crust for both profiles. If a density increase with depth is not introduced into the model, very high densities would be required for the near surface coastal and continental crustal blocks.