SCEC Award Number 22090 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Stochastic Branching Process Modeling of Earthquake Occurrence
Investigator(s)
Name Organization
Ilya Zaliapin University of Nevada, Reno Yevgeniy Kovchegov Oregon State University Yehuda Ben-Zion University of Southern California
Other Participants Graduate student (Natalie Bladis) at UNR; graduate students (TBN) at OSU and USC
SCEC Priorities 2e, 3d, 1d SCEC Groups EFP, Seismology, FARM
Report Due Date 03/15/2023 Date Report Submitted 03/24/2023
Project Abstract
 The project develops an earthquake modeling framework based on recent results in the theory of branching processes. The expected products include an improved stochastic branching model of earthquake occurrence, a methodology for model estimation and inference, and a suite of tools for operational analysis of seismicity and extracting information about the current state and dynamical changes of earthquake flow, including signals that reflect preparation processes of large earthquakes. The main framework of the proposed studies is the Invariant Galton-Watson (IGW) branching process model of earthquake occurrence. The project continues the PIs efforts in cluster analysis of seismicity and tracking generation of earthquake-induced rock damage and evolving localization of background seismicity (SCEC projects #19063 and #20065). The research used the updated version of the high-quality relocated catalogs in California. The project trained graduate students and facilitate the cross-disciplinary collaboration among UNR, OSU, and USC.
Intellectual Merit The project developed a novel model for earthquake occurrence based on the theory of Invariant Galton-Watson stochastic branching processes. The model expands the well-established self-exciting branching process modeling paradigm by addressing several well-known deficiencies in the existing earthquake modeling frameworks and providing a low-parametric tool for analysis of observed seismicity. The project developed statistical inference for the proposed model, based on theoretical distributions of the offspring numbers, earthquake cluster size, and cluster combinatorial depth (number of aftershock generations). The model estimation has been shown robust with respect to the minimal magnitude of analysis and to various errors of earthquake offspring identification. The analysis of the observed seismicity of southern California contributes to better understanding of the regional earthquake dynamics.
Broader Impacts The project results have an impact on research areas outside of the immediate project scope. The project develops a novel method for modeling earthquake occurrence, and contributed to the general theory of random self-similar trees that has applicability beyond seismology.
Exemplary Figure Figure 1: Invariant Galton-Watson (IGW) process fit to the empirical cluster sizes (number of earthquakes) in southern California. The empirical survival function of cluster size (green) and its IGW fit (red). Catalog of (Hauksson et al. 2012, extended) with M > 2 during 1981-2019.