Inferring crustal viscosity from seismic velocity: Application to the lower crust of Southern California

We investigate the role of composition on the viscosity of the lower crust through a joint inversion of seismic P-wave (Vp) and S-wave (Vs) velocities. We determine the efficacy of using seismic velocity to constrain viscosity, extending previous research demonstrating robust relationships between seismic velocity and crustal composition, as well as crustal composition and viscosity. First, we calculate equilibrium mineral assemblages and seismic velocities for a global compilation of crustal rocks at relevant pressures and temperatures. Second, we use a rheological mixing model that incorporates single-phase flow laws for major crust-forming minerals to calculate aggregate viscosity from predicted mineral assemblages. We find a robust correlation between crustal viscosity and Vp together with Vs in the α-quartz regime. Using seismic data, geodetic surface strain rates, and heat flow measurements from Southern California, our method predicts that lower crustal viscosity varies regionally by four orders of magnitude, and lower crustal stress varies by three orders of magnitude at 25 km depth. At least half of the total variability in stress can be attributed to composition, implying that regional lithology has a significant effect on lower crustal geodynamics. Finally, we use our method to predict the depth of the brittle-ductile transition and compare this to regional variations of the seismic-aseismic transition. The variations in the seismic-aseismic transition are not explained by the variations in our model rheology inferred from the geophysical observations. Thus, we conclude that fabric development, in conjunction with compositional variations (i.e., quartz and mica content), is required to explain the regional changes in the seismic-aseismic transition.