SCEC Award Number 22028 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title A kinematically-tuned dynamic model of southern California neotectonics based on CRM, CTM, CFM, & UCERF3
Investigator(s)
Name Organization
Peter Bird University of California, Los Angeles
Other Participants
SCEC Priorities 1c, 3c, 3e SCEC Groups SDOT, CXM
Report Due Date 03/15/2023 Date Report Submitted 03/08/2023
Project Abstract
 Dynamic models of neotectonics can be tuned by: (A) adjusting the common effective friction on all faults; (B) adjusting horizontal shear tractions on the base of the lithosphere; or (C) adjusting the friction of each fault separately. I built a Shells_v5.0 dynamic model incorporating the lateral heterogeneity of CTM and CRM, and then adjusted it to better match long-term fault offset-rates from the NeoKinema deformation model of UCERF3 that was built on CFM. Method (A) had little success; RMS residual errors were 4.6 mm/a for fault offset rates, 2.6 mm/a for GPS velocities, and 35 for stress directions. In method (B), I defined 31 microplates based on long-term velocities in the NeoKinema solution; then I used an iterated-solution method to adjust basal tractions on 23 of these microplates. At best, this reduced RMS fault offset-rate errors from 4.6 to 3.7 mm/a; however, required basal shear tractions were implausible in magnitude and also in map-pattern. In method (C), the Shells_v5.0 solution was revised 100 times, and the effective friction of each of 1000 fault elements was adjusted based on its current sense of slip-rate error. RMS error in slip-rates fell rapidly from 5.5 to 1.7 mm/a. About 48% of fault elements reached the lower friction limit of 0.01, and another 18% were below 0.21. My interpretation of these results is that ~60% of active fault area in southern California experiences near-total stress-drop in large earthquakes, due to dynamic weakening. Yet, other areas retain high effective friction and serve as nucleation sites.
Intellectual Merit The effective merger of dynamic (physics-based) and kinematic (data-based) models of tectonics gives both practical and theoretical benefits. A tuned quasi-static long-term dynamic model is the best basis for time-dependent studies such as rupture simulators and earthquake-sequence-simulators. It may also be the best kind of deformation model for seismic hazard studies. At the same time, this work is an advance in understanding earthquakes: Although they must nucleate on asperities with high shear stress, much of the ruptures occur on dynamically-weakened fault segments which release virtually all of their shear stress.
Broader Impacts There is long-term potential for improved seismic-hazard models, if tuned dynamic models such as this one are found to be worthy deformation models for computing long-term-average fault slip rates, and also permanent (non-elastic) deformation rates in surrounding continuum (which are potentially seismogenic).
Exemplary Figure Figure 1. Ribbon-map of computed effective friction coefficients on southern California faults, from Tuned_SHELLS_for_SCEC model TS2023001. The upper limit of 0.85 is the assigned friction of all continuum elements (not shown) between the faults. A lower limit of 0.01 was also imposed. A reasonable physical interpretation of these “effective friction” coefficients from a quasi-static dynamic model is that they give the ratios of coseismic shear stress to effective (lithostatic – hydrostatic) normal stress during earthquakes; thus, very low values suggest dynamic weakening.