Seismic Imaging of Scatterer Migration Associated with the 2004 Parkfield Earthquake Using Waveform Data of Repeating Earthquakes and Active Sources

Abstract We used borehole seismic records of four repeating-earthquake clusters and two explosions to investigate the coseismic changes of the scattering wave field of the 28 September 2004 M 6 Parkfield earthquake. We found systematic changes in P-wave and S-wave coda recorded from repeating events that occurred before and immediately after the earthquake. Applying the coda wave interferometry technique allowed us to determine that the observed changes are caused by a localized change in the scattering field. We further developed a technique based on decorrelation indexes calculated from running time windows to locate migrating scatterers. Synthetic tests showed that the technique is relatively insensitive to changes in background velocity of the medium and source location and thus can be applied to records of loosely colocated clusters. We found a localized change of material property within the fault zone at ∼3 km depth beneath the Middle Mountain area. The change is shown most clearly in the P-to-S scattering mode of the active source data and both the P-to-S and S-to-S scattering modes of the repeating earthquakes data, suggesting that the observed scattering property change is a result of charge or discharge of fluids in fractures caused by the 2004 Parkfield earthquake. The same scatterer(s) was found to respond to the 1993 aseismic stress/strain transient event, acting like an in situ stress–strain meter at seismogenic depth. Four-dimensional time-lapse imaging of the scattering wave field thus provides an effective way to monitor the subsurface stress–strain field