The LITHOPROBE seismic reflection project on Vancouver Island was designed to study the large-scale structure of several accreted terranes exposed on the island and to determine the geometry and structural characteristics of the subducting Juan de Fuca plate. In this paper, we interpret two LITHOPROBE profiles from southernmost Vancouver Island that were shot across three important terrane-bounding faults—Leech River, San Juan, and Survey Mountain—to determine their subsurface geometry and relationship to deeper structures associated with modem subduction.The structure beneath the island can be divided into an upper crustal region, consisting of several accreted terranes, and a deeper region that represents a landward extension of the modern offshore subduction complex. In the upper region, the Survey Mountain and Leech River faults are imaged as northeast-dipping thrusts that separate Wrangellia, a large Mesozoic–Paleozoic terrane, from two smaller accreted terranes: the Leech River schist, Mesozoic rocks that were metamorphosed in the Late Eocene; and the Metchosin Formation, a Lower Eocene basalt and gabbro unit. The Leech River fault, which was clearly imaged on both profiles, dips 35–45 °northeast and extends to about 10 km depth. The Survey Mountain fault lies parallel to and above the Leech River fault and extends to similar depths. The San Juan fault, the western continuation of the Survey Mountain fault, was not imaged, although indirect evidence suggests that it also is a thrust fault. These faults accommodated the Late Eocene amalgamation of the Leech River and Metchosin terranes along the southern perimeter of Wrangellia. Thereafter, these terranes acted as a relatively coherent lid for a younger subduction complex that has formed during the modem (40 Ma to present) convergent regime.Within this subduction complex, the LITHOPROBE profiles show three prominent bands of differing reflectivity that dip gently northeast. These bands represent regionally extensive layers lying beneath the lid of older accreted terranes. We interpret them as having formed by underplating of oceanic materials beneath the leading edge of an overriding continental place. The upper reflective layer can be projected updip to the south, where it is exposed in the Olympic Mountains as the Core rocks, an uplifted Cenozoic subduction complex composed dominantly of accreted marine sedimentary rocks. A middle zone of low reflectivity is not exposed at the surface, but results from an adjacent refraction survey indicate it is probably composed of relatively high velocity materials (~ 7.7 km/s). We consider two possibilities for the origin of this zone: (1) a detached slab of oceanic lithosphere accreted during an episodic tectonic event or (2) an imbricated package of mafic rocks derived by continuous accretion from the top of the subducting oceanic crust. The lower reflective layer is similar in reflection character to the upper layer and, therefore, is also interpreted as consisting dominantly of accreted marine sedimentary rocks. It represents the active zone of decoupling between the overriding and underthrusting plates and, thus, delimits present accretionary processes occurring directly above the descending Juan de Fuca plate. These results provide the first direct evidence for the process of subduction underplating or subcretion and illustrate a process that is probably important in the evolution and growth of continents.

The seismic structure of the British Columbia continental margin has been investigated using four reversed refraction profiles. The profile across strike extended 350 km from the volcanic arc on the continent to the deep ocean of the Juan de Fuca Plate; the three profiles along strike were located on Vancouver Island, on the continental shelf, and in the deep ocean on the Juan de Fuca Plate. Interpretation of the profile along Vancouver Island yields a well constrained model for the upper crust with velocity increasing from ~5.3 km/s at the surface to ~6.4 km/s at 2 km depth to ~6.75 km/s at 15.5 km depth where the velocity increases sharply to ~7 km/s. The velocity structure of the deep crust and the crustal thickness are poorly constrained. Four possible velocity functions, based on ambiguous first arrivals and (or) secondary phases interpreted as Moho reflections, are presented. The preferred one includes a deep crustal low velocity zone with a crustal thickness of 37 km; models with a constant 7.1 km/s deep crust require thicknesses of 52 km. Preliminary results from the profile across strike show the dip of the basement towards the continent steepens from approximately 1.4° immediately west of the continental rise to ~4° beneath the rise. Sediment velocities increase as the sedimentary layer thickens towards the shelf. The Moho, with velocity near 8 km/s, appears to dip at similar angles in this region; the dip is ~6° from the edge of the shelf to the central portion of Vancouver Island; here there is an abrupt thickening of the continental crust by about 10 km with a flat-lying Moho to the east. This suggests a contact between subducting oceanic Moho and continental Moho. A small positive velocity gradient is required in the mantle.Two short reflection lines, one using explosives and the other a large air gun fired in an inlet, were recorded on a land-based multichannel reflection system. These were run to test the feasibility of obtaining coherent reflections to upper mantle depths in this complex geological environment, and of acquiring deep reflection data in coastal areas with an air-gun source. The preliminary explosion section showed reflections near 4.4 and 6.8 s. The depths of these reflections correspond closely to the 15.5 km crustal refractor and the top of the subducting oceanic lithosphere, respectively. Dip on the deeper reflector is close to that estimated from the refraction profile. Without stacking or velocity filtering, the air-gun recordings on a line adjacent to the explosion profile show arrivals of energy at similar times.