SCEC Award Number 16048 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Tuning Dynamic Models to Estimate Varying Fault Friction, Asthenospheric Traction, and Lithospheric Stress
Name Organization
Peter Bird University of California, Los Angeles
Other Participants
SCEC Priorities 2d, 3d, 4b SCEC Groups CME, WGCEP, SDOT
Report Due Date 03/15/2017 Date Report Submitted 02/23/2017
Project Abstract
Two kinds of thin-shell finite-element model of neotectonics exist: (1) kinematic models which fit geodetic and geologic data, but use little physics; and (2) dynamic models that use realistic rheologies and stress-equilibrium, but don’t match kinematic data. A dynamic model can be tuned by: (A) adjusting horizontal shear tractions on the base of the lithosphere; or (B) adjusting the friction of each fault. I adjusted a Shells dynamic model to better match long-term flow and slip-rates from the NeoKinema deformation model of UCERF3. In method (A), 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 microplates. At best, the model with fault friction 0.04 reduced RMS fault slip-rate errors from 4.87 to 2.98 mm/a; however, required basal shear tractions were implausible (to 100 MPa) and also implausible in pattern. In method (B), the Shells solution was revised 100 times, and each time the effective friction of 1000 fault elements was adjusted up or down based on the current sense of slip-rate error. Over 60% of fault elements reached the lower friction limit of 0.01. RMS error in slip-rates fell from 5.69 to 1.55 mm/a. My interpretation 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 higher effective friction and serve as nucleation sites. All results are tentative because the planned Community Thermal Model and Rheology Model were not yet available.
Intellectual Merit This project represents an advance in dynamic modeling of neotectonics because it presents a new method of tuning a dynamic model to desired kinematics (by adjusting fault friction).
The specific results for southern California are important because they imply that ~60% of the area of active faults has extremely low effective friction, implying extreme dynamic weakening during earthquakes. This phenomenon has been suggested before, based on laboratory experiments and idealized models of earthquake rupture, but this is the first time that dynamic weakening has been shown to be pervasive and even typical during southern California earthquakes.
Broader Impacts This new method of dynamic modeling can easily be exported to other geographic regions using the open source code.
It is possible that fault slip-rates from this new kind of dynamic model will be used in future seismic-hazard estimates.
Exemplary Figure Figure 1. Ribbon-map of estimated effective friction coefficients on southern California faults, from Tuned_SHELLS_for_SCEC model TS2016005. 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 lithostatic normal stress during earthquakes; thus, very low values suggest that dynamic weakening is common on those faults.