HR: 09:45h
AN: S31A-08    [PDF]
TI: After the lightning and before the thunder; non-Omori behavior of early aftershocks?
AU: * Vidale, J E
EM: vidale@ucla.edu
AF: IGPP and ESS, UCLA, LA, CA 90095
AU: Cochran, E S
EM: cochran@moho.ess.ucla.edu
AF: IGPP and ESS, UCLA, LA, CA 90095
AU: Kanamori, H
EM: hiroo@gps.caltech.edu
AF: Seismo Lab, Caltech, LA, CA 90025
AU: Clayton, R W
EM: clay@gps.caltech.edu
AF: Seismo Lab, Caltech, LA, CA 90025
AB: Mainshock rupture is immediately followed by the receding drumbeat of aftershocks, which have been known to obey Omori's law for nearly a century. Unfortunately, the initial few minutes of aftershocks are difficult to observe due to the mainshock coda, network clipping, and the confusion of nearly simultaneous events. Yet this interval holds valuable clues about how stress-induced ruptures may be delayed, often in fits and starts, and the friction laws that govern temblors. With the advent of continuous recording on dense networks, we can pry open a window to reveal aftershock activity 1 to 5 minutes after moderate and large earthquakes. We high-pass-filter 100-sample-per-second three-component recordings to minimize reverberating coda, and select stations for low noise and impulsive arrivals to reveal patterns of P and S arrivals. We have examined close-in recordings of six earthquakes recorded on TriNET in southern California (M4.0 2001-10-28 Compton, M4.7 2002-09-21 Yorba Linda, and M7.2 1999-10-16 Hector Mine) and on Hi-Net in Japan (M7.3 2000-10-06 Tottori, M5.1 2003-05-11, and M5.4 2003-5-17). Visual examination of high-passed seismograms typically detects several times more events in the first few minutes than are recorded in the best catalogs. Calibration of observed amplitudes with the magnitudes of those events that are in the catalogs allows all detected events to be assigned a magnitude. Preliminary results suggest a deficit of aftershocks by roughly a factor of two from 100 to 300 sec after the mainshock relative to 1/t rate decline with time. This deficit is comparable to the one tentatively found for much smaller M1 to M2 (Rubin, JGR, 2002, p 3-10). This pattern is consistent with the existence of such an interval in rate-and-state friction models prior to the 1/t aftershock decay rate that (Dieterich, JGR, 1994, interval b). We expect that capturing early aftershock rates will aid in the understanding of the transition from propagating rupture to sporadic aftershock activity.
DE: 7209 Earthquake dynamics and mechanics
DE: 8164 Stresses--crust and lithosphere
SC: Seismology [S]
MN: 2003 Fall Meeting