PROBING MANTLE RHEOLOGY WITH FINITE-FAULT EARTHQUAKE-CYCLE MODELS AND THE SOUTHERN CALIFORNIA GPS VELOCITY FIELD

Viscoelastic relaxation of the lower lithosphere may cause surface velocities near active faults to vary with time between large earthquakes. Given Holocene slip rates, the history of large earthquakes, and high-precision PBO GPS data, this may provide a way to query the rheology of the lower lithosphere and asthenosphere. Interseismic surface velocity perturbations for a range of lithosphere-asthenosphere configurations and rheologies may be computed using earthquake cycle models, then subtracted from the current GPS velocity field. Ideally, if block model-inferred slip rates based on the adjusted GPS velocity field match Holocene rates, the lower lithosphere rheology and structure assumed in the earthquake-cycle model are admissible. For example, the GPS-Holocene slip rate discrepancies for the Mojave and San Bernardino segments of the SAF might be resolved if the mantle effective viscosity is 10¹⁸- 10¹⁹ Pa s (Chuang and Johnson, 2011; Hearn et al., 2013). If such resolution is impossible, changes in slip rate over timescales longer that the interseismic interval (but shorter than 10⁴years) must be considered.

To explore this idea further, I have developed a suite of finite-fault earthquake-cycle models incorporating layered viscoelastic structure and a lithosphere-scale viscous shear zone extending the SAF downward into the mantle asthenosphere. A range of characteristic earthquake magnitudes for the SAF is explored. For M_w = 7.9 characteristic earthquakes, an interseismic perturbation (or “ghost transient”) comparable in magnitude to that produced by model M1 of Hearn et al. (2013) is produced for a range of mantle viscosities (< 10¹⁹ Pa s). Relaxation of a shear zone with either depth-dependent or Burgers (time-dependent) viscosity per unit width can explain postseismic deformation typical of large strike-slip earthquakes. For smaller (M_w = 7.0 or 7.5) characteristic earthquakes, the long-wavelength interseismic perturbation is usually small (< 0.5 mm/yr), even if several such events are modeled simultaneously along adjoining segments. In this case geologic and geodetic slip rates based on the unadjusted GPS field should match, unless slip rates are fluctuating over timescales longer than the interseismic interval.