HR: 17:45h
AN: T24B-08 [Abstracts]
TI: Earthquake source parameters of repeating microearthquakes at Parkfield, CA, determined using the SAFOD
Pilot Hole seismic array
AU: * Imanishi, K
EM: imani@ni.aist.go.jp
AF: Geological Survey of Japan, AIST, AIST Tsukuba Central 7 1-1, Higashi 1-Chome, Tsukuba, 3058567
Japan
AU: Ellsworth, W L
EM: ellsworth@usgs.gov
AF: U.S. Geological Survey, MS-977 Office 1-650-329-5020 345 Middlefield Road, Menlo Park, CA 94025
United States
AB:
We determined source parameters of repeating microearthquakes occurring at Parkfield, CA, using the SAFOD Pilot Hole seismic
array. To estimate reliable source parameters, we used the empirical Green's function (EGF) deconvolution method which
removes the attenuation effects and site responses by taking the spectral amplitude ratio between the spectra of the two
colocated events. For earthquakes during the period from December 2002 to October 2003 whose S-P time differences are less
than 1 s, we detected 34 events that classified into 14 groups. Moment magnitudes range from -0.3 to 2.1. These data were
recorded at a sampling rate of 2 kHz. The dataset includes two SAFOD target repeating earthquakes which occurred on October
2003. In general, the deconvolution procedure is an unstable process, especially for higher frequencies, because small
location differences result in the profound effects on the spectral ratio. This leads to large uncertainties in the
estimations of corner frequencies. According to Chaverria et al. [2003], the wavetrain recorded in the Pilot Hole is
dominated by reflections and conversions and not random coda waves. So, we expect that the spectral ratios of the waves
between P and S wave will also reflect the source, as will the waves following S wave. We compared spectral ratios calculated
from the direct waves with those from other parts of the wavetrain, and confirmed that they showed similar shapes. Therefore
it is possible to obtain a more robust measure of spectral ratio by stacking the ratios calculated from shorter moving
windows taken along the record following the direct waves. We further stacked all ratios obtained from each level of the
array. The stacked spectral ratios were inverted for corner frequencies assuming the omega-square model. We determined static
stress drops from those corner frequencies assuming a circular crack model. We also calculated apparent stresses for each
event by considering frequency dependent attenuation, where the average difference between the observed and the calculated
omega-square model was assumed to represent the path and site effects. The estimated static stress drops are high, mostly in
excess of 10 MPa and with some above 50 MPa. It should be noted that the highest value is near the strength of the rock.
Apparent stresses range from 0.4 to 20 MPa, at the high end of the range of those reported by other studies. According to an
asperity model [e.g., McGarr, 1981; Johnson & Nadeau, 2002], the small strong asperity patch is surrounded by a much weaker
fault that creeps under the influence of tectonic stress. When the asperity patch ruptures, the surrounding area slips as it
is dynamically loaded by the stress release of the asperity patch. If so, our estimated source dimensions seem to correspond
to the size of the area surrounding the asperity patch, and the stress drops of the asperity might be much higher than our
estimations. Although this is consistent with the hypocesis of Nadeau and Johnson [1998], it is unlikely that the stress
drops exceed the strength of the rock. We should re-examine the asperity model based on the results obtained in this study.
DE: 7200 SEISMOLOGY
DE: 7209 Earthquake dynamics (1242)
DE: 7215 Earthquake source observations (1240)
SC: Tectonophysics [T]
MN: Fall Meeting 2005