Geodetic Constraints on the Parkfield Seismogenic Zone: Implications for the Earthquake Prediction Experiment

Sylvain D. Barbot, Piyush S. Agram, Nadia Lapusta, Jean-Philippe Avouac, & Mark Simons

In Preparation 2012, SCEC Contribution #1586

Accurate relocation of microseismicity on the Parkfield segment of the San Andreas Fault in California reveals a clustering of repeating micro-earthquakes with a highly organized spatio-temporal distribution of hypocenters. In general, areas devoid of microseismicity are suggestive of zones with somewhat homogeneous friction properties, either aseismic or seismogenic. The presence of the hypocenters of the 1966 and the 2004 Mw 6 Parkfield earthquakes in a large seismically quiet zone at about 7 km depth indicates that the seismic streaks may delineate the contours of the seismogenic zone. Among various inferences of slip from seismological data, some place most of the coseismic slip in the area of the fault circumscribed by background seismicity. However, inversions for coseismic slip from geodetic data from various investigators show a less obvious correlation between slip and the distribution of micro-earthquakes. Before 2004, the SCIGN GPS network at Parkfield is more sensitive to 1 cm of slip at depths shallower than 6 km than to 50 cm of slip in the deeper seismogenic zone, to which only two far-field stations are most sensitive. We test a couple of methods to improve the resolution of slip inversions from geodetic data. One is to optimize the sampling of the fault plane to reduce the tradeoffs between adjacent slip patches; another is to assimilate data from the interseismic period and require complementary slip distributions of interseismic creep and coseismic slip. Using the proposed methods, we reconcile inversions for coseismic slip from geodetic and seismological data and show that GPS and InSAR are compatible with the interpretation that the streaks of microseismicity circumscribe the seismogenic zone at Parkfield. This result suggests that the San Andreas Fault at Parkfield is partitioned into areas of stable and unstable friction delineated by enduring clusters of seismicity and that the transition occurs over a distance characterized by a concentration of microseismicity. The augmentation of the GPS network after the 2004 event allows us to better constrain fault slip, but more coverage is needed at strategic points to detect anomalous, and perhaps precursory, fault slip in the seismogenic zone during the current interseismic period before any future Mw 6 Parkfield rupture. We find the optimal spatial distribution of the additional GPS stations that will be required to resolve the seismogenic zone geodetically.

Barbot, S. D., Agram, P. S., Lapusta, N., Avouac, J., & Simons, M. (2012). Geodetic Constraints on the Parkfield Seismogenic Zone: Implications for the Earthquake Prediction Experiment. Journal of Geophysical Research, (in preparation).