Testing the applicability of linear Maxwell mantle flow models for interseismic deformation
Abstract
Lake unloading studies and some postseismic deformation studies in the western US infer an upper mantle average effective viscosity of order 10^18 to 10^20 Pa s. Assuming a linear Maxwell viscosity, this corresponds to relaxation times of 2.5-250 years. For earthquake repeat times of several hundred years, models with linear Maxwell mantle predict significant amounts of relaxation of the mantle between earthquakes. As a consequence, model surface velocities at any point in a plate boundary region can vary with time by as much as 3-5 mm/yr over the interseismic deformation period. We have been investigating the consequence of neglecting this time-dependence in block-like models of interseismc deformation in southern California using GPS data and viscoelastic earthquake cycle models. Chuang and Johnson (2011) show that slip rate estimates of the Mojave section of the San Andreas fault and the Garlock Fault are ~2 times higher using viscoelastic block models (with mantle flow) than using elastic block models (no mantle flow). Yet, it is suggested from laboratory experiments that mantle flow is better explained with nonlinear flow laws than a linear Maxwell flow law. Earthquake cycle models show that flow in a nonlinear, stress-dependent mantle is more localized beneath a strike-slip fault than flow in a linear viscous mantle. Therefore earthquake cycle models with linear Maxwell mantle flow may overestimate the amount of long-wavelength relaxing flow during the interseismic period. The advantage of the linear models is that earthquake cycles with Maxwell rheology are easy to implement in semi-anlaytical models. But how good of an approximation is the Maxwell model? Can the model predictions be ruled out based on observations? We use interseismic and postseismic GPS data sets along the North Anatolian fault in western Turkey to test the applicability of linear Maxwell earthquake cycle models. Hearn et al. (2009) showed that the postseismic transient following the 1999 Izmit and Duzce earthquakes are best explained with afterslip in an elastic crust overlying a flowing mantle with viscosity of about 2-5 x10^19 Pa s. Hearn et al. (2009) claim that this viscosity range for linear flow is inconsistent with localized deformation observed in the same region before the 1999 earthquakes. However, this reasoning is largely based on model results of infinitely long strike-slip faults. We examine this problem using 3D earthquake cycle models and investigate whether or not linear Maxwell mantle flow can be ruled out based on pre- and post-earthquake data sets.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2011
- Bibcode:
- 2011AGUFM.T31E..06J
- Keywords:
-
- 8159 TECTONOPHYSICS / Rheology: crust and lithosphere