The Cascadia subduction zone is extremely complex in the western Washington region, involving local deformation of the subducting Juan de Fuca plate and complicated block structures in the crust. It has been postulated that the Cascadia subduction zone could be the source for a large thrust earthquake, possibly as large as M9.0. Large intraplate earthquakes from within the subducting Juan de Fuca plate beneath the Puget Sound region have accounted for most of the energy release in this century and future such large earthquakes are expected. Added to these possible hazards is clear evidence for strong crustal deformation events in the Puget Sound region near faults such as the Seattle fault, which passes through the southern Seattle metropolitan area. In order to understand the nature of these individual earthquake sources and their possible interrelationship, we have conducted an extensive seismotectonic study of the region. We have employed P-wave velocity models developed using local earthquake tomography as a key tool in this research. Other information utilized includes geological, paleoseismic, gravity, magnetic, magnetotelluric, deformation, seismicity, focal mechanism and geodetic data. Neotectonic concepts were tested and augmented through use of anelastic (creep) deformation models based on thin-plate, finite-element techniques developed by Peter Bird, UCLA. These programs model anelastic strain rate, stress, and velocity fields for given rheological parameters, variable crust and lithosphere thicknesses, heat flow, and elevation. Known faults in western Washington and the main Cascadia subduction thrust were incorporated in the modeling process. Significant results from the velocity models include delineation of a previously studied arch in the subducting Juan de Fuca plate. The axis of the arch is oriented in the direction of current subduction and asymmetrically deformed due to the effects of a northern buttress mapped in the velocity models. This buttress occurs under the North Cascades region of Washington and under southern Vancouver Island. We find that regional faults zones such as the Devils Mt. and Darrington zones follow the margin of this buttress and the Olympic-Wallowa lineament forms its southern boundary east of the Puget Lowland. Thick, high-velocity, lower-crustal rocks are interpreted to be a mafic/ultramafic wedge occuring just above the subduction thrust. This mafic wedge appears to be jointly deformed with the arch, suggesting strong coupling between the subducting plate and upper plate crust in the Puget Sound region at depths >30 km. Such tectonic coupling is possible if brittle-ductile transition temperatures for mafic/ultramafic rocks on both sides of the thrust are assumed. The deformation models show that dominant north-south compression in the coast ranges of Washington and Oregon is controlled by a highly mafic crust and low heat flow, allowing efficient transmission of margin-parallel shear from Pacific plate interaction with North America. Complex stress patterns which curve around the Puget Sound region require a concentration of northwest-directed shear in the North Cascades of Washington. The preferred model shows that greatest horizontal shortening occurs across the Devils Mt. fault zone and the east end of the Seattle fault.