SCEC Award Number 19186 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Modeling fault branches in 3D using an earthquake simulator and fully dynamic methods: investigating numerical and physical effects
Investigator(s)
Name Organization
David Oglesby University of California, Riverside Keith Richards-Dinger University of California, Riverside James Dieterich University of California, Riverside Christodoulos Kyriakopoulos University of Memphis
Other Participants Baoning Wu, Graduate Student, UC Riverside
SCEC Priorities 1e, 2e, 5a SCEC Groups FARM, CS, SDOT
Report Due Date 04/30/2020 Date Report Submitted 04/30/2020
Project Abstract
Our project is to compare the behavior of earthquake simulations on branched faults between two different modeling methods: RSQSim, a long-term, quasi-dynamic modeling method, and 3D dynamic rupture models. The former method allows one to model the evaluation of stress over multiple earthquake cycles, leading potentially to more realistic initial conditions for such earthquake models. The latter method more rigorously solves the dynamic equations, but requires the input of (usually ad hoc or homogeneous) initial stress patterns. Our goal is to determine how close the two modeling method results are for models with similar initial conditions, including both homogeneous pre-stress and evolved long-term stresses from RSQSim. In the process, we will address whether the propagation of rupture at a branch is dominated more by the heterogenous pre-stress field (which will be the same for both methods), or the dynamic stress field, which in principle could be different due to the different numerical approximates made. Our preliminary results, which compare RSQSim results to the 3D dynamic results of the SCEC/USGS Spontaneous Rupture Code Verification Project, imply that the two methods produce qualitatively similar results: inputs that produce through-going rupture on the main branch for the Verification Project models also produce through-going rupture for RSQSim, and inputs that lead to rupture primarily of the branch give similarly congruent results. We plan to continue this work for different branching angles and for heterogeneous prestress conditions from multi-cycle RSQSim models.
Intellectual Merit This proposed research is directly in line with many of the Basic Questions and Research Priorities of SCEC5. The project directly investigates “how the stress transfer among fault segments depends on time, at which levels it can be approximated by quasi-static and dynamic elastic mechanisms” (P1.e.). It also helps to “determine the probability of rupture propagation through structural complexities” (P2.e.). Finally, it is directly in line with developing “earthquake simulators, which are earthquake rupture simulation codes that allow the development of long-term earthquake catalogs based on approximations to rupture physics and earthquake interaction, and encode the current understanding of earthquake predictability” (P5.a.). In addition to these specific goals, the proposed work also will contribute to SCEC’s Earthquake Gates initiative, as a fault branch is one of the quintessential earthquake gates where rupture path is decided.
Broader Impacts The proposed work will have broader impacts on the evaluation of seismic hazard in California and beyond. Earthquake simulators such as RSQSim may well form part of the foundation of future Uniform California Earthquake Rupture Forecasts (UCERFs). As with all modeling methods, it is important to investigate whether, given the approximations it makes, its results are useful—specifically, whether its behavior at geometrical complexities produces realistic behavior in real-world conditions of heterogenous stress. Comparison with a 3D dynamic modeling code, which makes different approximations, can help in this regard. The results will have implications for how to incorporate earthquake simulators into seismic hazard evaluation.
Exemplary Figure Figure 3. Comparison of Verification results (top panels) and RSQSim result (bottom panels for the left-lateral problem 15. Multiple curves in Verification results correspond to results for different modelers and methods.