Using GPS Derived Shear Strain Rates in Southern California to Constrain Fault Slip Rate, Locking Depth, and Residual Off-Fault Strain Rates

Strain rate fields within strike-slip regimes often possess complexity associated with along-strike  slip  rate  variations.  These along-strike  slip  rate  variations  produce dilatational  components  of  strain  rate  within  and  near  the  fault  zones  and  within  the adjacent block areas. These dilatation rates do not directly reflect the slip rate magnitude on the strike  slip  fault,  but  rather  the  relative  change in  along-strike  slip  rate.  Displacement rates measured  using  GPS  observations  reflect  the  full  deformation  gradient  field,  which may involve significant dilatational components and other off-fault deformation. Thus, using displacement rates  to  infer  slip  rate  and  locking  depth  of  major  strike-slip  faults  may introduce  errors  when  along-strike  slip  rate  variations  are  present.  On the  contrary,  true locking depth and slip rate can be obtained from the pure strike-slip component of shear strain rates. In this study we investigate the use of shear strain rates alone (obtained from the full displacement rate field of the SCEC 4.0 velocity field in southern California) to infer fault slip rate and locking depth parameters along the San Andreas (up to 37° N) as well as the San Jacinto fault zones. Such an  analysis is  critical  for  accurate  estimation  of  off-fault  strain  rates  outside  of  the major shear zones.

We conducted benchmarking tests to determine if accurate shear strain rates  can  be  obtained  from  a  synthetic  fault  slip  rate  field  possessing  the  same  station spacing as the SCEC 4.0 dataset. The synthetics were derived using Okada’s [1992] elastic dislocation routine (Coulomb  3.2).  These displacement  rates  were  interpolated  using  bi-cubic  Bessel  interpolation  to  infer the full  horizontal  velocity  gradient  tensor  field,  along with model uncertainties.  To test realistic conditions, along-strike slip  rates were put into the elastic dislocation model and model displacements were output at the true GPS station spacing in southern California from the SCEC 4.0 dataset. The modeled strain rate field shows negligible strain rate artifacts in most regions and both the shear strain and dilatation rates obtained  from  the  bi-cubic interpolation  were  well-resolved.  The inferred shear strain rate field was then  inverted, using a simple screw dislocation  forward model  for the best-fit  fault location,  fault locking depth,  and  fault  slip  rate.  Model parameter estimates were well resolved,  both  near  and away from fault slip rate transitions (± 1 km for fault locking depth; ± 1-2 mm/yr for fault slip rate). Test results to date show the method can resolve slip rates and locking depth within the zones of along-strike transition. Results to date from this methodology applied to southern California using the SCEC 4.0 GPS velocity field show  remarkably  well-resolved  and  prominent  shear  strain  rate  bands  that follow  both  the San Andreas and San  Jacinto  fault  systems.   The shear strain rates  reflect dramatic  along-strike  slip  rate  variations,  found  in  many  previous  studies.  However, fault locking depths are generally shallower than previously published results. Residual off-fault strain rates, not associated with the major strike-slip faults, appear to accommodate ~30% of the total Pacific-North American plate relative motion.