Postseismic deformations at Parkfield

Co- and post-seismic deformations due to the 2004 Parkfield earthquake (First-year Master research project)

in collaboration with Sylvain Barbot (Earth Observatory of Singapore) & Jean-Philippe Avouac (Caltech)

The 2004 Parkfield, CA earthquake was a long-expected event which ruptured the San Andreas fault between the creeping segment, to the North and the locked one to the South. The presence of multiple geodetic and seismic instruments surrounding the epicenter allowed to record in detail many aspects of the rupture and subsequent motion on and surrounding the fault.

The postseismic relaxation following the 2004 event was dominated by afterslip, which cumulated a total equivalent geodetic moment that is comparable to the moment released coseismically. A relatively large postseismic moment is not generally observed at other locations and implies a large area to partake in stable creep. Using all available data, including continuous GPS and synthetic aperture radar interferograms (InSAR), we tested whether some of the data may be explained by a viscoelastic relaxation of a weak lower crust or whether the strong inferred postseismic moment is only explained by stable (aseismic) afterslip on the fault plane surrounding the coseismic rupture.

We documented a strong tradeoff between inferred viscoelastic flow and deep afterslip below the coseismic rupture. Assuming that time series of predicted viscoelastic flow following the 2004 rupture is separable in space and time, we divise a linear inversion scheme that resolves the relative contribution of afterslip and viscoelastic flow as a function of time. We find that a mechanically viable model that accounts for viscoelastic flow in the lower crust implies less deep afterslip (with an equivalent geodetic moment of 5.8e17 N.m after three months, assuming a uniform shear modulus of 30GPa), as opposed to models that consider afterslip as the single mechanism of postseismic relaxation (which imply a cumulative afterslip geodetic moment of about 9.2e17 N.m for the same period). A viable model, assuming viscoelastic relaxation in the lower crust and upper mantle requires the joint occurrence of some afterslip on the down-dip extension of the fault that ruptured coseismically and viscoelastic flow in the deeper substrate.

Related publication: Bruhat et al. [2011].