Repeating earthquakes and prospecting for temporal change in rock properties associated with geodetic deformation at Kilauea Volcano, Hawaii

Temporary seismic deployments on Kilauea Volcano from Nov. 1999 to June 2000 and Feb. to Aug. 2007 provided dense observations of earthquakes during two east rift zone intrusions (Feb. 23, 2000 and June 17-21, 2007). We identify clusters of earthquakes based on waveform similarity. Waveform pairs with cross correlation coefficients greater than 0.8 between 1 and 10 Hz in 1999/2000 and 3-10 Hz in 2007) are grouped together. Clusters are separated when a stack of grouped waveforms is cross correlated with another stack and the inter-cluster cross correlation coefficient is less than 0.7. Many earthquakes during the 1999/2000 deployment had similar waveforms. Site W33 on the south flank recorded the largest number of earthquakes and also had the most pairs of similar earthquakes. The maximum cross correlation coefficient between a pair of earthquakes was 0.99, and the largest cluster contained 193 earthquakes. Rift zone sites (e.g. W37 and W38) had very few similar earthquakes, suggesting site effects might play an important role in the ability to identify clusters. Further analysis will help determine if focal mechanism variations or more widely separated locations contribute to the reduced similarity at these sites. Seismicity rates increased during the Feb. 2000 intrusion, but the waveforms of these earthquakes are similar to those that were occurring before and after the intrusion. Thus, the stress change due to the intrusion appears to have increased activity on continually active clusters of earthquakes (also shown by Rubin et al. [JGR, 1998] during the 1983 intrusion). Also during the 1999/2000 deployment, a M5 earthquake occurred on the south flank. Active clusters of smaller flank earthquakes were “turned off” by the large earthquake, while new clusters activated after the earthquake in nearly the same locations, but with slightly different waveforms. Despite having significantly more observed earthquakes, repeatability of waveforms during the 2007 temporary deployment was very low. Background noise levels were higher during this deployment, possibly due to elevated tremor levels or the near-coast locations of the 2007 deployment. The maximum cross correlation coefficient between earthquake pairs at site SEQA1 is 0.97, but the largest cluster, has only 19 earthquakes. Relocating additional earthquakes plus analyzing the waveforms for changes in seismic velocity and shear wave splitting will allow us to relate the seismic observations to deformation and stress changes derived from geodetic data.