Detection of a locked zone at depth on the Parkfield, California, segment of the San Andreas fault

The Parkfield, California, segment of the San Andreas fault is transitional in character between the creeping segment of the fault to the northwest and the locked Carrizo Plain segment to the southeast. The rate of shallow fault slip decreases from 25–30 mm/yr northwest of the epicenter of the 1966 Parkfield earthquake to zero at the southeastern end of the 1966 rupture zone. Data from a network of trilateration lines spanning the San Andreas fault near Parkfield and extending to the Pacific coast near San Luis Obispo shed light on the rate of fault slip at depth since the 1966 earthquake. In this study, average rates of line length change and shallow fault slip were inverted to determine the slip rate at depth on the Parkfield fault segment. The fault is taken to be a vertical surface with unknown distribution of strike-slip displacement in an elastic half-space. A striking result of the inversions is that all solutions providing acceptable fits to the data exhibit a locked zone essentially coincident with the rupture surface of the 1966 Parkfield earthquake. The data require that the locked zone extend nearly as far north as the 1966 epicenter; however, the vertical extent of the locked zone is not well resolved. Over much of the Parkfield segment the fault is slipping faster at the earth's surface than it is at seismogenic depths. In order to fit the trilateration measurements it is necessary to include a component of contraction normal to the trend of the San Andreas. The inversion results suggest a spatially uniform normal strain of −0.06 μstrain/yr. The orientation of the contraction is compatible with geologic and seismic evidence of active folding and reverse faulting in the region. The magnitude of the contraction is consistent with convergence rates inferred from global plate motion models.