SCEC Award Number 11137 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title Relative Contribution of Afterslip and Viscoelastic Relaxation following the 2004 Parkfield Earthquake
Investigator(s)
Name Organization
Jean-Philippe Avouac California Institute of Technology
Other Participants Michael Aizazis
Sylvain Bernard
1 Undergraduate Student
SCEC Priorities A3, A10, C SCEC Groups FARM, CDM
Report Due Date 02/29/2012 Date Report Submitted N/A
Project Abstract
Funding from SCEC3 in 2011 has contributed to supporting three studies of the earthquake cycle at the Parkield segment of the San Andreas Fault. In a first study, we have looked at the postseismic relaxation following the 2004 Mw\,6 Parkfield, CA earthquake using GPS and Interferometric Synthetic Aperture Radar (InSAR) data~\citep{bruhat+11}. We have shown that the surface deformation is dominantly explained by afterslip around the coseismic rupture with a possible contribution of distributed viscoelastic flow in the lower crust. This study highlights the variations of the degree of localization of deformation at crustal depths. This study has been executed by a visiting undergraduate student, Lucile Bruhat, with the co-mentoring of post-doctoral fellow Sylvain Barbot. In a second study partially funded by SCEC3, we have developed a physical model of the earthquake cycle at Parkfield with rate-and-state friction~\citep{barbot+12a}. The model consists of a spatial distribution of friction property which is tuned to explain a variety of seismological, geodetic and paleoseismic observations. The model generates spontaneous ruptures between longer periods of quasi-static loading and reproduces the average recurrence time, the magnitude and hypocenters of the latest moderate-size earthquakes at Parkfield. In particular, we have found that the change of hypocenter location of the Mw\,6 earthquakes from the north in 1966 to the south in 2004 can occur spontaneously on rate-and-state friction faults due to a complexity in the history of faulting. This study was performed in collaboration with Prof Nadia Lapusta and is to be published in Science this year. Lastly, the SCEC3 founding has allowed us to use InSAR and GPS data, in combination to results from seismological investigations, to check whether or not the seismogenic zone at Parkfield is surrounded by persistent streaks of seismicity~\citep{barbot+12b}. We have found that this scenario is plausible, i.e., not in contradiction with the geodetic data. But more importantly, we have established that augmenting the Plate Boundary Observatory (PBO) GPS network with 20 additional stations - if optimally situated - would allow us to monitor slip in the seismogenic zone with sufficient resolution. These studies provide a self-consistent physical interpretation of the earthquake cycle at Parkfield and motivate further targeted investigations, part of the new Parkfield Special Fault Study Area of SCEC4.
Intellectual Merit We have developed the first fully dynamic model of the earthquake cycle capable of quantitatively reproducing a wide range of observations for the Parkfield segment of the San Andreas Fault. Our study demonstrates the possibility of creating comprehensive physical models of fault zones that integrate geodetic and seismological observations for all stages of the earthquake source cycle. As computational resources and methods improve, more realistic fully dynamic simulations - allowing for a wider range of earthquakes magnitude occurring on a set of interacting faults - will become possible. Such simulations could in principle be used to assess the full range of earthquake patterns that a particular fault system might produce, or assimilate observation about past earthquakes and interseismic loading to assess future seismicity.


Our proposed study is directly relevant to long term goals of SCEC4, such making progress on some Fundamental Problems of Earthquake Physics. The project integrates the goals of the SCEC Science Plan with the many overlapping goals of the EarthScope Science Plan by characterizing the crust and lithosphere of the natural laboratory of Southern California and enhancing models of the rheology of the lithosphere.
Broader Impacts The project has involved the participation of a female undergraduate student, Lucile Bruhat, as part of her undergraduate thesis. Lucile has published an article early in her carrier and has since been accepted for graduate school at Stanford. Post-doctoral scholars Sylvain Barbot and Piyush Agram also participated to the project. We have developed the first physical model that can explain all the periods of the earthquake cycle at Parkfield, including the slow tectonic loading, the details of a dynamic rupture and the transient that follows every rupture. The presented study demonstrates the possibility of creating comprehensive physical models of fault zones that integrate geodetic and seismological observations for all stages of the earthquake source cycle. As computational resources and methods improve, more realistic fully dynamic simulations will become possible. Such simulations could in principle be used to assess the full range of earthquake patterns that a particular fault system might produce, or assimilate observation about past earthquakes and interseismic loading to assess future seismicity. The resulting region-specific models can then be used to study how the entire earthquake cycle is affected by various additional factors, such as static and/or dynamic triggering and human intervention in the form of hydrothermal energy harvesting and CO2 sequestration. Hence such models would provide an indispensable tool for understanding and mitigating seismic hazard around active faults, with important benefits to society.
Exemplary Figure Figure 2

Caption: Model that reproduces the entire seismic cycle at Parkfield. A) Spatial distribution of rate-weakening (white) and rate-strengthening (grey) friction properties (a-b), with rate-weakening values in the seismogenic zone, delineated by the background seismicity. B-F) Snapshots of a Mw 6.0 seismic cycle with rupture nucleating spontaneously to the south, near the Parkfield 2004 hypocenter. Another Mw 6 event nucleates 20 years later.

Barbot et al. 2012