SCEC Award Number 23104 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Estimating stress state along the San Jacinto and southern San Andreas faults on the eve of past ground rupturing earthquakes and implications for current stress state
Investigator(s)
Name Organization
Michele Cooke University of Massachusetts Amherst
Other Participants Emery Anderson-Merritt
SCEC Priorities 1c, 1d, 2a SCEC Groups SDOT, SAFS, Geology
Report Due Date 03/15/2024 Date Report Submitted 03/20/2024
Project Abstract
 Estimating evolving stresses along a fault system may reveal the conditions that generated previous ground-rupturing earthquakes and provide realistic initial conditions for dynamic rupture models. We use quasistatic forward numerical models incorporating 3D complex configuration of active faults in southern California to estimate shear tractions along the southern San Andreas and San Jacinto fault systems since 1000 CE. These tractions include the interseismic accumulation due to tectonic loading, viscoelastic relaxation of stress within the crust between earthquakes, and coseismic effects of nearby earthquakes. To simulate interseismic shear traction accumulation we use a two-step back slip approach to estimate linear interseismic loading rate and subtract from it the effect of viscoelastic relaxation. We simulate ground-rupturing earthquakes using earthquake timing and extents compiled by Scharer & Yule (2020), testing two different assumptions about coseismic stress release: a complete stress drop, for which we release all of the accumulated shear traction, and an incomplete stress drop, for which we apply a tapered stress drop independent of the accumulated traction. Both models produce average slip per event that is generally consistent with Scharer & Yule (2020), but they produce different traction accumulation histories and pre-earthquake tractions. In the complete stress drop models, maximum pre-quake tractions rarely exceed ~0.5 MPa and minimum tractions increase slightly with rupture length; in the incomplete stress drop models, pre-quake tractions are more variable, and the minimum pre-quake tractions increase distinctly with rupture length. This new modeling approach may provide estimates of shear tractions that are unavailable from direct measurements.
Intellectual Merit In this study, we estimate the evolving and pre-earthquake along-strike shear tractions along the San Andreas and San Jacinto faults since ~1000 CE, and investigate the impacts of uncertainty in earthquake timing and upper crustal viscosity on shear traction estimates. Uncertainties in upper crustal viscosity (η ≥ 1020 Pa-sec) have a greater impact on shear tractions than earthquake timing uncertainties. In the simulations, viscosity less than 1020 Pa-sec results in excessive stress relaxation and average slip per earthquake event that does not match geologic estimates. Estimates of the fault shear tractions through time and over several earthquake cycles reveal potential conditions that preceded previous ground rupturing earthquakes. Our findings show that longer ruptures are associated with greater accumulated shear traction prior to the earthquake. The distributions of pre-earthquake tractions estimated here can provide heterogeneous initial conditions for dynamic rupture models that may be more realistic than either uniform traction assumptions or the resolution of regional stress states onto fault surfaces.
Broader Impacts The pre-earthquake tractions along the San Andreas and San Jacinto faults can be used as initial conditions for dynamic rupture models. Because this product derives from the past earthquake record and explicitly incorporates several processes that contribute to fault stress, it provides a significant refinement over approaches that make assumptions about these processes. The refined pre-earthquake tractions for models that simulate past earthquake events may yield further insights into the conditions that generate damaging earthquakes. This project supports both a UMass PhD student, Emery Anderson-Merritt, who is transgender, and a female PI, Cooke, with deafness.
Exemplary Figure Fig. 4: Pre-earthquake shear tractions from Monte Carlo simulations. A) Models with elastic and viscoelastic (η = 1020) rheology show impact of earthquake timing uncertainty, and B) Models with viscoelastic rheology (limited to η ≥ 1020 Pa-sec) show impact of viscosity uncertainty.