Current motion on faults of the San Andreas system in central California inferred from recent GPS and terrestrial survey measurements
The San Andreas fault system of California comprises dominantly right-lateral strike-slip faults and forms part of the Pacific-North American plate boundary. This fault system has been studied extensively using geological and geophysical methods since it was first brought into prominence by the 1906 M = 8(^1)(_4) San Francisco earthquake. Observations of surface deformation thought to define an earthquake deformation cycle have been inferred from terrestrial and space-based geodetic methods. The observed relative motion in these networks has also been used to constrain the distribution of motion across the plate boundary. Sites in three profiles extending across the fault system in the San Francisco bay region were measured up to 7 times between March 1990 and February 1993 using the Global Positioning System (GPS). The data were processed using the Bernese V3.2 software. The GPS data were combined with trilateration and VLBI data to create a spatially dense sample of the deformation field in the region. Approximately 35±3 mm/yr of fault-parallel (N33ºW) shear is distributed across a deforming zone that increases in width northwards from 60 to 100 km and in style from fault-concentrated deformation in the south to near-linear trends in the north. No systematic convergence upon the fault is observed. Both two- and three-dimensional models of dislocations in an elastic half-space were used to model the deformation and to investigate the effects of structural complexities such as a low-rigidity fault zone, the depth to which surface creep extends, geometrical complexities of the fault system and along-strike variations in slip rate. The models produce a remarkably close fit to the deformation despite such a rheologically simple Earth structure. Approximately half of the observed deformation is accommodated along faults to the east of the San Andreas fault. A zone of concentrated deformation across the San Andreas fault zone in the north of the region may be the result of a 1-2 km wide low-rigidity fault zone there. Surface creep rates, although highly variable, appear to increase to the south. An increase in depth of the surface creep zone to the south may also accompany this. The variations in slip rate at depth along strike are consistent with connectivity between the major faults of the system. Quasi-steady slip on discrete fault planes or shear zones may occur down to 2-3 times the seismogenic depth and deformation rates are probably almost constant throughout much of the earthquake cycle. The present earthquake potential calculated from the estimated slip rates indicate that several fault segments may have an earthquake potential equivalent in magnitude to the "characteristic" earthquake assumed for that segment. The estimates of relative motion indicate that deformation across the San Andreas fault system, plus that observed to the east of the Sierra Nevada mountains, can account for all of the Pacific-North American plate motion rate.