A simulation-based approach to forecasting the next great San Francisco earthquake

Department of Physics, Boston University, Boston, Massachusetts, United States
Proceedings of the National Academy of Sciences (Impact Factor: 9.67). 11/2005; 102(43):15363-7. DOI: 10.1073/pnas.0507528102
Source: PubMed


In 1906 the great San Francisco earthquake and fire destroyed much of the city. As we approach the 100-year anniversary of that event, a critical concern is the hazard posed by another such earthquake. In this article, we examine the assumptions presently used to compute the probability of occurrence of these earthquakes. We also present the results of a numerical simulation of interacting faults on the San Andreas system. Called Virtual California, this simulation can be used to compute the times, locations, and magnitudes of simulated earthquakes on the San Andreas fault in the vicinity of San Francisco. Of particular importance are results for the statistical distribution of recurrence times between great earthquakes, results that are difficult or impossible to obtain from a purely field-based approach.

Download full-text


Available from: Bruce D. Malamud,
  • Source
    • "We show the first quantity – the cumulative distribution of expected interoccurrence intervals between m > 7 earthquakes in the EMC region in Fig. 4 (the EMC region and fault model is shown in Figs. 1 -3). We fit the observed distribution to a Weibull function (Rundle et al. 2005; Shcherbakov et al. 2005), for which the fraction of recurrence intervals ∆t r that are less than ∆t r is "
    [Show abstract] [Hide abstract]
    ABSTRACT: In this manuscript, we introduce a framework for developing earthquake forecasts using Virtual Quake (VQ), the generalized successor to the perhaps better known Virtual California (VC) earthquake simulator. We discuss the basic merits and mechanics of the simulator, and we present several statistics of interest for earthquake forecasting. We also show that, though the system as a whole behaves quite randomly, (simulated) earthquake sequences limited to specific fault segments exhibit measurable predictability in the form of increasing seismicity precursory to large m > 7 earthquakes. Based on this, we develop an alert based forecasting metric and show that it exhibits significant information gain compared to random forecasts. We also discuss the long standing question of activation vs quiescent type earthquake triggering. We show that VQ exhibits both behaviors separately for independent fault segments; some fault segments exhibit activation type triggering; other fault segments are better characterized by quiescent type triggering. We discuss these aspects of VQ specifically with respect to faults in the Salton Basin and near the El Mayor-Cucapah region in southern California USA and northern Baja California Norte, Mexico. We also discuss several proposed extensions of the VC model and related technologies that will be operated in conjunction with VC to produce composite earthquake forecasts.
    Geophysical Journal International 10/2015; 203(3):1587-1604. DOI:10.1093/gji/ggv320 · 2.56 Impact Factor
  • Source
    • "Virtual Quake (VQ) is a boundary element, fault type earthquake simulator based on slider-block dynamics and Green's function stress transfer. VQ is a generalization of the perhaps better known Virtual California (VC) earthquake simulator, originally developed by Rundle [1988], in which the fault parameterization has been generalized and simplified to facilitate modeling of any arbitrary fault system[Heien and Sachs, 2012; Heien et al, 2014; Rundle et al, 2005; Sachs et al, 2012]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We introduce a framework for forecasting earthquakes using Virtual Quake (VQ), the generalized successor to the perhaps better known Virtual California earthquake simulator. We briefly introduce the VQ simulator, including its availability to research organizations and statistics relevant to earthquake forecasting applications. We discuss contemporary, regional type, forecasts and also show that forecasts can be significantly improved by partitioning catalogs along fault sections.
  • Source
    • "This is supported by numerous studies of static stress triggering, as well as the existence of a persistent stress shadow (inhibition of earthquakes) following the 1906 earthquake (Stein, 1999). Long-term seismicity rates, fault interactions, and patterns of moderate to large earthquakes are products of models of fault systems controlled by the mechanics of crustal deformation over long time periods (Ward and Goes, 1993; Rundle et al., 2005, 2006a). Understanding how such systems evolve in time under a relevant set of governing physical laws is a needed critical step toward reliable earthquake forecasting. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Earthquake simulation on synthetic fault networks carries great potential for characterizing the statistical patterns of earthquake occurrence. I present an earthquake simulator based on elastic dislocation theory. It accounts for the effects of interseismic tectonic loading, static stress steps at the time of earthquakes, and post-earthquake stress readjustment through viscoelastic relaxation of the lower crust and mantle. Earthquake rupture initiation and termination are determined with a Coulomb failure stress criterion and the static cascade model. The simulator is applied to interacting multifault systems: one, a synthetic two-fault network, and the other, a fault network representative of the San Francisco Bay region. The faults are discretized both along strike and along dip and can accommodate both strike slip and dip slip. Stress and seismicity functions are evaluated over 30,000 yr trial time periods, resulting in a detailed statistical characterization of the fault systems. Seismicity functions such as the coefficient of variation and a- and b-values exhibit systematic patterns with respect to simple model parameters. This suggests that reliable estimation of the controlling parameters of an earthquake simulator is a prerequisite to the interpretation of its output in terms of seismic hazard.
    Bulletin of the Seismological Society of America 06/2009; 99(3):1760-1785. DOI:10.1785/0120080253 · 2.32 Impact Factor
Show more