Michael E. Caron, M. Meghan Miller and Daniel J. Johnson1
Data collected from continuously-recording GPS stations of the Pacific Northwest Geodetic Array (PANGA) provide a key input to investigating partitioning of strain among active crustal faults in the Puget Sound area. Neotectonic deformation in this area is related in large part to northward-directed motion of the Cascadia forearc due to entrainment above the obliquely subducting margin between Juan de Fuca and North America. A rational crustal fault block model for the Pacific Northwest has been constructed, using data from a variety of sources, including extant geologic mapping, aeromagnetic and gravity survey data, digital elevation model imaging and seismic survey data. The current model, which includes all of Washington and Oregon and a portion of southern British Columbia, consists of nine blocks of varying size. In addition to ensuring that each block was bounded by reasonably well-defined crustal faults, additional consideration was given to the need to have at least two GPS stations located on each block in order to properly drive rotations and translation vectors during inversion.
Previous crustal strain partitioning in the Puget Sound area has relied in large part on paleoseismic constraints, although widespread unconsolidated Quaternary sediments and extensive urban development obscure many of the critical structures and direct observation is often difficult or impossible. However, recent marine and land-based high-resolution seismic surveying has allowed formulation of important paleoseismic constraints for several of the more important faults in the Puget Lowlands, including constraints on Quaternary slip rates for the Seattle and South Whidbey Island faults.
Although relatively sparse, sufficient high quality data have recently become available from the PANGA network to allow inversion of these data across a kinematically consistent crustal block model. The inversion process requires input of the basic block model parameters including individual fault segment definitions such as spatial location, strike and dip, locking depth and burial depth. Where deemed appropriate, the model also allows input of a priori constraints such as dip, tensile and strike-slip rates. The other key inputs to inversion are robust velocity solutions for continuously recording GPS stations, including data from the PANGA network in the Pacific Northwest as well as data from several stations on the North American craton used to provide a stable reference frame.
Inversion of a completely unconstrained model results in unrealistically high rates of closure motion across critical crustal faults in Puget Sound, in part due to the sparseness of the velocity field. Consequently, reasonable a priori slip constraints were applied to the Seattle and South Whidbey Island faults, as well as to the less-well-constrained Tacoma fault. With these three faults reasonably constrained, model inversion produces small residual velocity vectors, indicating substantial agreement between model and observed velocities.
An analysis of block model results may allow formulation of key constraints on the way the forearc behaves under strain and the likelihood of catastrophic or damaging release of this strain by seismic rupture. Although robust conclusions are premature, modeled fault slip rates from inversion show, for example, up to 8mm of reverse dip slip on the fault bounding the north side of the Olympic Mountains uplift. This fault is not currently known to have significant Quaternary slip rates. The amount of slip modeled on this fault exceeds modeled (and in part paleoseismically constrained) slip across the Tacoma, Seattle, South Whidbey Island and Devils Mountain faults in Puget Sound. This modeled slip budget excess implies that slip is either being taken up along poorly known (or unknown) faults to the south in the large area of Quaternary cover in the lowlands to the south of the Olympic Mountains uplift, or is being accommodated by folding.
1 Department of Geological Sciences, Central Washington University, Ellensburg, WA 98926