SCEC Award Number 20149 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Building a viscous mantle rheology model for Southern California constrained by tomography and postseismic deformation
Investigator(s)
Name Organization
Kaj Johnson Indiana University Thorsten Becker University of Texas at Austin
Other Participants Jacob Dorsett (IU graduate student)
Simone Puel (UT Austin graduate student)
SCEC Priorities 1c, 1e, 3b SCEC Groups SDOT, CXM, Geodesy
Report Due Date 03/15/2021 Date Report Submitted 07/09/2021
Project Abstract
 We continue to develop 3-D heterogeneous viscoelastic static Green's functions (GFs) for deformation in southern California for the SCEC Community Rheology Model. We have tested viscoelastic mantle models using GPS-derived postseismic displacements following the 1999 Hector Mine and 2019 Ridgecrest earthquakes. In the mantle, we estimate viscosities using laboratory derived (power-law) creep laws where spatial temperature variation is inferred from seismic velocity anomalies. The resulting mantle viscosities vary by orders of magnitude across California and the western US. We construct viscoelastic Green's functions using the mantle viscosity model using finite element code PyLith. To model postseismic deformation, we use boundary element approach with the pre-computed viscoelastic Green’s Functions and model afterslip on the fault with a rate-strengthening friction law. Afterslip and mantle flow are fully coupled. We impose coseismic slip models derived from geodetic data for the Ridgefield and Hector Mine earthquakes.

Major findings:
Postseismic mantle flow is evident in geodetic data throughout southern California following the Mw 7.1 2019 Ridgecrest earthquake. Best-fit models suggest a postseismic viscosity of 1–5×1017 Pa s at 60–100 km depth and steady-state viscosity lower than average for the region.
3D steady-state viscosity model predicts observed pattern of 20-year post-Hector Mine displacements fairly well but overpredicts displacements on the east side of the rupture.
Intellectual Merit For Ridgecrest postseismic, we used 110-day cumulative displacements following the Mw 7.1 July 5 2019 Ridgecrest earthquake to investigate the early transient phase of postseismic deformation. We generated a suite of effective postseismic viscosity models based on the 3D steady-state viscosity model and power-law and Burgers formulations. Our results suggest that compared to our long-term reference viscosity model, a transient viscosity reduction is required to fit the geodetic signal. Stress-dependent, power-law type transients appear to not capture these reductions fully, and a egionally constant, Burger’s type viscosity reduction appears more likely. Best-fit models require a low effective postseismic viscosity of the order of 1-5 × 10^17 Pa s at 60 – 100 km depth, from which we estimate a permissible range of steady-state (pre-earthquake) effective viscosity depth profiles with viscosity of 1-8 × 10^18 Pa s at 60-100 km depths, which tends towards the lower end of previous estimates.

We are able to largely fit cumulative postseismic displacements following the 1999 Hector Mine and 2010 El Mayor-Cucapah earthquakes with the steady-state viscosity model. The magnitude and orientation of observed and models postseismic displacements are in good agreement everywhere except east of the Mojave, suggesting the steady-state viscosity model represents the actual viscosity structure reasonably well except east of the Mojave. Further investigation and modeling is required to reconcile the misfits to the east.
Broader Impacts This project helped support the research of a graduate student at Indiana University (Jacob Dorsett) and UT-Austin (Simone Puel). This work contributes to the development of the SCEC Community Rheology Model.
Exemplary Figure Figure 4. (a). Example time series fit for station SHOS. Blue dots are data after removing linear trend. Green curves and light green bands show mean and two standard deviation range of modeled random walk. Black curve shows mean logarithmic postseismic curve. Shades of pink show 1,2, and 3 standard deviations. (b) Cumulative 20-yr observed and modeled postseismic displacements after Hector Mine (1999-2020). (c) Cumulative 10-yr observed and modeled postseismic displacements after El Mayor-Cucapah (2010-2020).