SCEC Award Number 13187 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title A Comprehensive Analysis of Recent Transient Activity in Southern California
Investigator(s)
Name Organization
William Holt Stony Brook University
Other Participants Gina Shcherbenko
SCEC Priorities 5, 2, 1 SCEC Groups Transient Detection, Geodesy, SDOT
Report Due Date 03/15/2014 Date Report Submitted N/A
Project Abstract
 Through previous SCEC funding we have developed, tested, and automated a geodetic network processing system for detection of anomalous strain transients in Southern California. These codes are now being incorporated into the Collaboratory for the Study of Earthquake Predictability (CSEP). The modeling procedure determines time-dependent displacement gradient fields from continuous GPS (cGPS) time series [Hernandez et al., 2005, 2007; Holt and Shcherbenko, 2013]. The method has been tested using the SCEC IV Transient Detection Exercise and is able to recover the spatial and temporal distribution of slow events and determine their statistical significance. To date we have detected several transient strain phenomena. These anomalous strains have occurred on a variety of time scales, they involve heterogeneous distributions, and they have impacted stress rates and coulomb stress changes on faults. Using the anomalous strain estimates, along with a model for secular strain accumulation, we have tracked the stress change history following the El Mayor event. We have used this strain change history to infer the Coulomb stress changes on known fault structures throughout southern California. Using a rate/state algorithm that incorporates the influence of all strain changes (co-seismic and post-seismic) constrained by the cGPS data, we generate predicted seismicity rates and compare with observed rates for magnitudes > M 1.5. Seismicity rate patterns are well predicted throughout the last 3 years, but rates are underpredicted by a factor of 3 along the Laguna Salada Fault (northernmost segment)/southern Elsinore FZ. Elsewhere, rate predictions provide a good match with observed, including along the Brawley fault zone, both before and after the swarm of August, 2012.
Intellectual Merit Analysis of the cGPS data have revealed significant transients on a variety of time scales. Some of these are linked with known earthquakes (post-seismic relaxation in El Mayor-Cucapah [Pollitz et al., 2012]; Brawly Swarm, Hauksson et al., 2013), whereas others are linked with unknown, and potentially slow-slip, processes (Parkfield and San Simeon regions) (Holt and Shcherbenko, 2013). We have quantified model estimates of stress changes on fault interfaces [King et al., 1992] in association with these measured strain changes. We are now investigating the relation between the inferred stress tensor changes and seismicity rates (including a comparison with moment tensor style) and we will provide this tool as a data product. Following El Mayor event, the GPS constrained strain evolution shows the following: (1) The Southern San Andreas experiences a reduced rate of right-lateral strike slip strain accumulation between 2010.5 – 2012.5. (2) The San Jacinto Fault has normal rate of right-lateral strike-slip strain accumulation during this time. (3) Before the Brawley swarm of August, 2012, the state of strain evolves to enable unclamping of a left-lateral fault zone there. Furthermore, the Coulomb stress changes on the faults in the region progressively increase toward failure up to the time of the Brawley swarm. (4) Large shear strains accumulate on the Laguna Salada Fault (northernmost segment)/southern Elsinore FZ, and these regions show increased Coulomb stress changes throughout the post-seismic process. Using a rate/state algorithm that incorporates the influence of all strain changes (co-seismic and post-seismic) constrained by the cGPS data, we generate predicted seismicity rates and compare with observed rates for magnitudes > M 1.5. Seismicity rate patterns are well predicted throughout the last 3 years, but rates are underpredicted by a factor of 3 along the Laguna Salada Fault (northernmost segment)/southern Elsinore FZ. Elsewhere, rate predictions provide a good match with observed, including along the Brawley fault zone, both before and after the swarm of August, 2012.

This work fulfills the SCEC Science Objective 5b “Application of geodetic detectors to the search for aseismic transients across southern California” and also supports 2c, 1d, and 1e and 2d. Moreover, the tracking of strain and stress changes over time, along with inferred seismicity rate changes, fulfills the recommendations under Research Strategies in Tectonic Geodesy to: (a) “Improve our understanding of the processes underlying detected transient deformation signals and/or their seismic hazard implications through data collection and development of new analysis tools.”
Broader Impacts The Broader Impact of this work involves the training of a graduate student, Gina Shcherbenko. Gina is currently writing up a paper summarizing the results of Coulomb stress changes and seismicity rate predictions. The development of a data product, enabling automated detection of anomalous strain from CGPS data also constitutes a broader impact.
Exemplary Figure Figure 4. (A) Strain tensor differences plotted on top of contoured shear strain (scalar magnitude of pure strike-slip style). The tensor differences along the southern San Andreas show a left-lateral sense of motion, which implies a reduced rate of right-lateral motion (relative to the reference steady-state model) during the 2 year time period following El Mayor event. (B) Same as (A), but plotted on top of contoured dilatation strain. (C) Coulomb stress change on right lateral faults (vertical) shortly after the El Mayor event. Epicenter locations of all earthquakes for the month of April, 2010 [Hauksson et al., 2012]. (D) Same as (C) but for the total period spanning April 4, 2010 – August 25, 2012 (effects of Brawley swarm on stress changes are not included). The epicenters of earthquakes from the same time period are also plotted. Contour interval is 0.2 bars and maximum stress changes approach 1.6 bars over the two years of strain evolution following El Mayor event.