SCEC Award Number 21086 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Developing earthquake simulators for UCERF4
Name Organization
Bruce Shaw Columbia University
Other Participants
SCEC Priorities 5a, 4c, 5b SCEC Groups EFP, FARM, CS
Report Due Date 03/15/2022 Date Report Submitted 03/25/2022
Project Abstract
We report here on progress made in the last year. Two papers were published,
one paper was submitted. The first one, ``Toward Physics-Based Nonergodic PSHA: A
Prototype Fully Deterministic Seismic Hazard Model for Southern California'' ({\sl Milner, Shaw, et al., 2021}) was published in the Bulletin of the Seismological Society of America.
The second one, "An Earthquake Simulator for New Zealand'' ({\sl Shaw, et al., 2022}) was also published in the Bulletin of the Seismological Society of America.
A third paper on rupture plausibility filters was submitted for publication, ``Enumerating Plausible Multifault Ruptures in Complex Fault Systems with Physical Constraints'' ({\sl Milner, Shaw, and Field, 2022}).
Intellectual Merit Earthquake simulators have the potential to contribute to our understanding of earth- quake hazard in a number of different ways. One way, demonstrated in recent work, is in directly estimating earthquake rupture forecasts, which, when combined with ground motion models, give hazard estimates which compare very well with traditional statistical model estimates [Shaw et al., 2018]. A second way is to use the models to help develop and constrain assumptions in the statistical models. This approach was used in use UCERF3 to support magnitude dependent coefficient of variation in the time dependence of elastic rebound recurrence [Field et al., 2015]. A third way being actively explored is to use the simulators as source models for ground motion studies [Milner et al., 2021]. All three approaches have shown much promise.
In this work we aimed to use the simulators as a guide and test bed for a key ingredient in the UCERF machinery: helping define the rupture set. In UCERF, the rupture set is the set of possible ruptures that go into the inversion, which then solves for the probabil- ities of the ruptures given the data constraints. In UCERF3, this rupture set was chosen through a series of filters which included constraints on jump distances, rake changes, azimuth changes, cumulative azimuth changes, and a local Coulomb measure, among other criteria. Subsequent study has identified a number of shortcomings with this approach, and opportunities to improve upon it. This work aimed to use the complex multi fault ruptures spontaneously emerging from the earthquake simulators to help guide an improved process for developing and testing the rupture sets for a next generation UCERF4. New criteria and a comparison with the simulator were developed.
Broader Impacts Helps improve rupture set for UCERF4, the leading methodology for fault-based multifault rupture seismic hazard analysis.
Exemplary Figure Figure 7: Rupture plausibility figures. (a). The new filters use various measures of coulomb failure interactions to improve the rupture set. This illustrates those coulomb interactions. 3D view looking north of faults from the Third Uniform California Earthquake Rupture Forecast (UCERF3) model broken up into 2 km x 2 km patches for Coulomb calculations. In this example, eight subsections of the Garlock fault are used as sources (green) and Coulomb stress changes are computed to all other patches (with contributions summed across all source patches). Receiver patches are colored by their sign with darker colors indicating greater amplitude, and subsection outlines are colored by the sum across all receiver patches (red is positive, blue negative). This shows the Coulomb-preferred co-rupture direction of the left-lateral Garlock Fault connecting to the Mojave section of the right-lateral San Andreas Fault (SAF). Coastlines are overlaid in black. (b). Improvement in how new proposed rupture criteria match RSQSim rupture set. The rate at which ruptures from the RSQSim comparison model fail the plausibility filters used in UCERF3 (solid gray line) and proposed here (solid black line), as a function of magnitude. This excludes the minimum number of subsections per cluster filter, which is common to both models and was not intended to gauge rupture plausibility. Failure rates are also given for rupture subsets where up to one (dashed lines) and two (dotted lines) subsections are removed from an end of a participating fault section; this shows that the majority of failing ruptures are largely compatible except for one or two incompatible subsections. From [Milner, Shaw, and Field, submitted, 2022].