Improving Models for Coseismic And Postseismic Deformation from the 2002 Denali, Alaska Earthquake

Given the multi-decadal temporal scale of postseismic deformation, predictions of previous models for postseismic deformation resulting from the 2002 Denali Fault earthquake (M 7.9) do not agree with longer-term observations. In revising the past postseismic models with what is now over a decade of data, the first step is revisiting coseismic displacements and slip distribution of the earthquake. Advances in processing allow us to better constrain coseismic displacement estimates, which affect slip distribution predictions in modeling. Additionally, an updated slip model structure from a homogeneous model to a layered model rectifies previous inconsistencies between coseismic and postseismic models. Previous studies have shown that two primary processes contribute to postseismic deformation: afterslip, which decays with a short time constant; and viscoelastic relaxation, which decays with a longer time constant. We fit continuous postseismic GPS time series with three different relaxation models: 1) logarithmic decay + exponential decay, 2) log + exp + exp, and 3) log + log + exp. A grid search is used to minimize total model WRSS, and we find optimal relaxation times of: 1) 0.125 years (log) and 21.67 years (exp); 2) 0.14 years (log), 0.68 years (exp), and 28.33 years (exp); 3) 0.055 years (log), 14.44 years (log), and 22.22 years (exp). While there is not a one-to-one correspondence between a particular decay constant and a mechanism, the optimization of these constants allows us to model the future timeseries and constrain the contribution of different postseismic processes.