What is the Origin of the High Q Anomaly at the Boundary Between the Locked and Creeping Segments of the San Andreas Fault?

Thousands of microearthquakes have been recorded since 1987 by the borehole seismic network at Parkfield, providing us with a comprehensive data set that has been used to characterize the nature of wave propagation in the San Andreas Fault Zone at Parkfield. An analysis of this data, combined with numerical modelling, has shown that fault zone guided waves (FZGW) are generated within a fault zone that is 100 to 200 m wide at seismogenic depths, with a shear wave velocity that is 20-40% lower than that of the adjacent unfaulted media. Amplitude guided wave tomographic inversion results show that the FZGW are most effectively generated in a well-defined part of the FZ that plunges to the northwest through the region of highest moment release and separates locked and slipping sections of the fault (as determined from both geodesy and microearthquake recurrence rates). This feature coincides with a sharply-bounded region of high Q within the fault zone, inferred from finite-difference modeling of the observed wave field. We use a particle-based numerical modelling scheme, the Discrete Particle Scheme (DPS), to investigate the origin of the localized zone of FZGW generation and high Q anomaly at Parkfield. This zone may delineate the NW edge of the M6 asperity, with the high Q arising from dewatering due to fracture closure and/or fault-normal compression and/or changes in fracture orientation due to the complex stress field at the boundary between the creeping and locked zones of the San Andreas Fault. These processes can be modeled in the DPS, which captures the dilation or closure of fractures due to changes in the orientation or magnitude of the regional stress field. Seismic waves can then be propagated through the model to investigate the effect of changing fracture populations on the wavefield. We propose to use changes in the attenuation and velocity of the wavefield to image the fracture distribution and hence the spatial distribution of stress within the fault zone.