Postseismic Deformation Following the 1906 San Francisco Earthquake

S. J. Kenner and P. Segall, Stanford University

Abstract

We have re-evaluated triangulation data from northern California following the 1906 San Francisco earthquake, thereby increasing the temporal and spatial resolution of postseismic deformation following that event. We have calculated uniform shear strain-rates and average station velocities at Pt. Arena using data from 1906-1907, 1929-1930, and 1973-1975 and for the Pt. Reyes-Petaluma arc using data from 1929-1930, 1938-1939, and 1960-1961. With the addition of recent geodetic data, we infer an effective relaxation time for long-term, postseismic deformation following the 1906 earthquake of 36 ± 16 years. The Pt. Arena data are satisfactorily fit with accelerated afterslip at depth along the San Andreas fault plane. For a 10 km deep, 25 km wide afterslip zone the average slip-rate between 1906 and 1930 is 11.2 cm/yr. Between 1929 and 1975 it is 4.6 cm/yr. Deformation in the Pt. Reyes-Petaluma arc is clearly asymmetric with respect to the trace of the San Andreas fault, especially between 1929 and 1939. After inverting for the range of acceptable accelerated deep afterslip and horizontal detachment models, a detailed analysis using geologically reasonable geometries revealed that both model types have trouble explaining the spatial variations in the deformation field through time. In particular, accelerated deep afterslip models cannot reproduce the breadth of the observed deformation field to the northeast of the San Andreas fault. As a result, more complex, time-varying deformation mechanisms are required to explain the Pt. Reyes-Petaluma arc observations.

Figure 1: General map showing the location of the Pt. Arena and Pt. Reyes-Petaluma networks. The trace of the 1906 San Francisco earthquake is shown by the dashed line. Dark circles denote stations which were used in the northern California primary network inversions.

Figure 2: Evolution of uniform shear strain-rate with time since the 1906 San Francisco earthquake at Pt. Arena (circles) and Pt. Reyes (triangles). Vertical error bars give 1s uncertainties. Horizontal error bars give the time between the first and last surveys included in the inversion. The solid line is the best-fitting exponential regression to the data. Strains have been rotated 20° (clockwise) and 10° (clockwise), respectively, from north so that g1 represents the right-lateral shear strain-rate across that San Andreas fault.

Figure 3: Engineering shear strains within the Pt. Reyes-Petaluma arc between (a, b) 1929 and 1939 and (c, d) 1938 and 1961. Model results are shown for a purely secular prior model (solid), a model with postseismic afterslip slip between 25 and 50 km on the San Andreas fault (dashed), and a model that includes a horizontal detachment at 15 km depth that extend 45 km to the northeast of the San Andreas fault (dash-dotted). Slip-rates in the accelerated afterslip model are 4.0 and 2.8 cm/yr for the 1929-1939 and 1938-1961 periods, respectively. Slip-rates in the detachment model are 4.3 and 3.0 cm/yr for the 1929-1939 and 1938-1961 periods, respectively. Here g1 represents the right-lateral shear strain-rate across that San Andreas fault (N35°W) and g2 represents shearing associated with unilateral extension or compression perpendicular to the trace of the San Andreas fault (tension positive). Vertical error bars give the 1s uncertainty in the magnitude. Horizontal error bars give the lateral extent, in the fault perpendicular direction, of the subnetwork in which the calculation was made.

Reference: Kenner, S.J., and P. Segall, Postseismic Deformation Following the 1906 San Francisco Earthquake, Journal of Geophysical Research, v. 105, p. 13,195-13,209, June 2000.