Strength Variations of Southern California from Rheological and Geodynamical Approaches

Understanding the rheological structure of the lithosphere is important for inferring loading rate on faults and their potential downdip extensions. To this end, we compare viscosity estimates from geodynamic models and predictions of lithosphere rheology. At each point on a grid covering Southern California, we first produce deviatoric stress estimates averaged from the surface to 100 km depth obtained by modelling variations of crustal structure (gravity potential energy) and effective viscosity. This geodynamic model is evaluated on the basis of surface geodetic data. In a complementary approach, we generate a strength envelope at each grid point based on various community products provided by the Southern California Earthquake Center (SCEC) and associated researchers. For example, we use the thermal model of Shinevar et al. (2018) and crustal thickness variations from Shen and Ritzwoller 2016). At each depth, the lowest of the stress associated with dislocation or diffusion creep is retained. Eight alternative rheological models are developed, that consider either wet or dry rheologies, a uniform grain size (1mm) or a grain size tied to a piezometer, and a maximum allowed stress of 300 MPa or 30 MPa. We use the flow laws of feldspar from Rybacki et al. (2006) is used in the crust and that of Hirth and Kohlstedt (2004) for olivine in the mantle. At this point, lithological variations in the crust are neglected, although we find evidence in our results that they are probably important. The strength envelop is integrated with depth for various strain rates to produce an effective rheology of the lithosphere. We then determine the strain rate associated with the geodynamically-inferred stress and use it to produce a viscosity estimate from the rheological model

 

The first result of this analysis is that the effective rheology of the lithosphere is highly non-linear (effective stress exponent between 10 and 30). Therefore only a limited range of stress is expected at any given location. In general, that stress is quite high so that the viscosity estimates from the geodynamic model can be explained only when using the weakest crust and mantle rheologies. Although certain viscosity variations are consistent between the models (strong Sierra Nevada block, weak Salton trough area), in other places they are not. In particular, the Great Basin block appears strong in the geodynamic model but weak in the rheological model due to high temperatures there. This study shows that there are likely important rheological variations due to mineralogy. It is expected that the Great Basin is underplated by gabbroic rocks that are not included here but would increase the effective viscosity of the rheological model. Mineralogical variations would also allow a great variety of accessible stress at each location. Finally, it is also likely that the stress does not reach failure at every depth, which is the basis for considering low saturation values for the strength envelop.