Characterization of the Parkfield Transient: Evidence of Aseismic Stress Transfer


Analysis of more than a decade of high quality data, particularly those from the two-color electronic distance meter (EDM) in the Parkfield, California area reveals a significant transient in slip-rate along the San Andreas fault. The rate increase is 3.3 ± 0.9 mm/yr. during 1993.0 to 1999.0. The most reliable fault creep instruments show a comparable increase in slip rate, suggesting that the slip is localized to the fault and breaks the surface. The EDM and creep data can be simply modeled by a 10km-long vertical right-lateral fault centered 2 km southeast of the central EDM monument CARR, extending from the surface to 9 km depth, and possessing a spatially uniform slip rate of 3.3mm/yr. The existence of such a transient slip event requires that there has been either a reduction in frictional resistance on the fault or an increase in loading stress (and partial release by the slip). The EDM data alone are primarily sensitive to slip on the fault itself, and thus do not provide a good diagnostic. The more sensitive diagnostic for loading/release is the sense of shear strain away from the fault. The presense of two borehole tensor strainmeters, which show a previously reported strain transient with the same time history (Gwyther et al., 1996), provides a way of distinguishing between these two cases. We find that the calculated strains from the above displacement model, corresponding to the stress-release case, are incompatible with the observed borehole shear strains. Both data sets can be satisfied, however, by adding a zone of higher transient right-lateral slip further to the north, which would physically correspond to stress loading from the north. This model makes physical sense, given the occurrence of four significant (M > 4) events within this northern zone during the first two years of the transient. Based on their magnitude, the slip on these faults is of order 6cm and with fault dimension of about 2km. All four events occurred at about 9km depth. If we consider a model of northern slip over a depth range of 2km to 9km depth (above the depths of the events and below the surface, since no transient was observed on northern creepmeters) and extending over a 15km length of fault, then a slip rate of 12mm/yr. over 5 years provides a good fit to both data sets. With this value, the coseismic slip and accumulated slow slip are comparable. Support for this magnitude of northern slip is found in microearthquake data. Based on the recurrence interval of repeating microearthquakes, Nadeau and McEvilly (1999) infer a slip-rate increase of about 10mm/yr. in the northern zone, beginning in 1993, comparable to the value we have used. We thus interpret the Parkfield transient as a composite event involving stress transfer: a northern event dominated by the occurrence of four seismic events and associated slow slip, which serves to load and ultimately trigger a slip event further to the south.

Meeting Name

AGU Fall Meeting (1999: Dec. 1, San Francisco, CA)


Geosciences and Geological and Petroleum Engineering

Document Type

Article - Conference proceedings

Document Version


File Type





© 1999 American Geophysical Union (AGU), All rights reserved.

This document is currently not available here.