Viscoelastic finite-element models of the earthquake cycle along plate-boundary faults

• Main Author: Vaghri, Ali
• Language: English
• Published: University of British Columbia 2011
• Online Access: http://hdl.handle.net/2429/37136

In this thesis I describe three related projects on the mechanics of plate bounding faults and fault systems. All of the projects involve viscoelastic earthquake cycle modeling, using the finite element method, and investigating how surface deformation is affected by viscoelastic relaxation or viscous shear zone creep. In the first project I develop test models to study how contrasts in material properties across a fault can affect surface deformation. I model contrasts in elastic plate thickness and in viscosity profiles across a strike-slip fault. I find that an offset between the surface trace of a fault and its position at depth (where it creeps), due to a non-vertical dip, may make interseismic deformation look asymmetric. Rheological contrasts do not yield dramatic asymmetry for models with realistic viscosity profiles and the sense of asymmetry in viscoelastic models actually reverses during the interseismic interval. In the second project, I model a set of parallel strike-slip faults, which allow relative motion of the North American and Pacific Plates in northern California. Geologic information allows me to take the timing of large earthquakes on these faults into account in the modeling, and I compare modeled surface deformation with GPS data from the region. I find that to explain the GPS velocity field, I must incorporate the dip of the faults into my model. The required dip for the Green Valley Fault in particular is consistent with new, double-difference hypocenter data from the region. The final project involves modeling Cascadia Subduction Zone earthquake cycles. This subduction zone has been modeled in the past, but always using a kinematic device (“backslip”) for modeling slip on the subduction interface. I developed a detailed profile model of the Cascadia Subduction Zone, based on data from the Lithoprobe project. Since the Subduction Zone has a curved profile and its slip is driven by relative motion of the model boundaries, the models produced unrealistic uplift along the North America plate boundary. I tested several approaches for preserving the geometry of the subduction interface. I suggest that a new implementation of split nodes in the finite-element method would be required to make this correction work.