SCEC Award Number 17064 View PDF
Proposal Category Individual Proposal (Data Gathering and Products)
Proposal Title Predictive Model for Earthquake Rupture Extents
Investigator(s)
Name Organization
Glenn Biasi United States Geological Survey Steven Wesnousky University of Nevada, Reno
Other Participants
SCEC Priorities 2e, 5b, 5a SCEC Groups EFP, WGCEP, CISM
Report Due Date 11/30/2018 Date Report Submitted 03/24/2020
Project Abstract
 Earthquake early warning (EEW) systems can quickly identify the onset of an earthquake rupture, but the first few seconds of seismic data only weakly predict the final rupture length. We present methodology for estimating the conditional probabilities of rupture length given a nucleation point from an EEW system. The methodology is based on the empirical observation that bends and steps in a fault are geometric complexities with some probability of arresting rupture, and the assumption that the effect of fault complexities compounds serially as rupture grow in length. In this context, the methods developed here provide a basis to estimate where a rupture may go once it has initiated, and with what probabilities. These probabilities are readily compiled in advance for any given starting subsection in the fault model, in effect covering likely nucleation locations for large earthquakes anywhere in the California fault network. It is suggested that probabilities constructed in this manner could ultimately be used before the event to advise policy about alerting possible rupture extents for earthquakes initiating at points along a fault network. Operationally, pre-compiled probabilities could quickly be accessed by the EEW system when an earthquake has initiated. As an illustrative and immediately useful example, we find that an earthquake that initiates at Bombay Beach on the SE end of the San Andreas fault only reaches San Bernardino about 5% of the time, the point at which modeling suggests a major risk to Los Angeles.
Intellectual Merit A major issue in seismology today is the possibility of developing an Earthquake Early Warning System. A current limitation of seismological observation is that the ultimate extent of a rupture cannot (yet?) be predicted from the initial phase of first-arrival waveforms. We put forth a novel methodology based on empirically derived probabilities of a rupture passing bends and steps in mapped fault trace to probabilistically assess, once a moderate sized earthquake is recognized, the location and ultimate length to which the earthquake will grow along the fault model of the Southern California Earthquake Center .
Broader Impacts The broader impact of the research has the potential to of significant societal impact. The probabilities are readily compiled in advance for any given starting subsection in the fault model, in effect covering likely nucleation locations for large earthquakes anywhere in the California fault network. It is suggested that probabilities constructed in this manner could ultimately be used before the event to advise policy about alerting possible rupture extents for earthquakes initiating at points along a fault network. Operationally, pre-compiled probabilities could quickly be accessed by the EEW system when an earthquake has initiated and serve to save many lives and damage to lifeline infrastructure.
Exemplary Figure Figure 1. Illustration of single fault composed of 9 panels (subsections) illustrating possible rupture extents for an earthquake initiating in central panel S0. The probability of any given rupture is pi.