HR: 11:20h
AN: S31G-05    [PDF]
TI: Frequency-Dependent Dynamic Triggering
AU: * Brodsky, E E
EM: brodsky@ess.ucla.edu
AF: UCLA Dept. of Earth and Space Sci., 595 Charles Young Dr., Los Angeles, CA 90095
AU: Prejean, S G
EM: sprejean@usgs.gov
AF: Stephanie Prejean, USGS 345 Middlefield Rd., Menlo Park, CA 94025
AB: Dynamic triggering has now been robustly documented for at least 4 M$_w$$>$7 earthquakes. As broadband seismic network coverage becomes denser, the continuing absence of mid-range triggering from M6-7 earthquakes motivates us to ask under what conditions do seismic waves trigger earthquakes. We provide observational evidence here that one important factor in triggering is the frequency of the seismic waves. The key evidence comes from the cases of remote triggering at Long Valley caldera. At least three regional $M_w>6$ earthquakes since 1990 have produced shaking in Long Valley that exceeded the local peak particle velocity (and therefore transient elastic strain) of the 2002 Denali earthquake, yet the Denali earthquake triggered local seismicity and the regional ones did not. For instance, the Northridge earthquake (M$_w$=6.7, $\Delta$=381 km) produced shaking with a peak-to-peak vertical amplitude of 1.4 cm/s in Long Valley and the Denali earthquake (M$_w$=7.9, $\Delta$=3458 km) had peak-to-peak shaking of 1.0 cm/s in the same spot. However, the locally weaker Denali earthquake triggered local earthquakes and Northridge did not. The Denali-triggered events began within 10 s of the beginning of the Rayleigh wave particle motion, which is less than the duration of the Northridge wave train. Therefore, it is difficult to appeal to the difference in duration of the wavetrains. We turn instead to the alternative explanation that long-period waves ($\stackrel{\sim}{>}$10--15 s) are more effective at triggering earthquakes than short-period ones. Under a low-pass filter with a corner at 15 s, Denali produced 2.5 times greater shaking at Long Valley than Northridge. The other examples are also consistent with this threshold. The sensitivity of triggering to long-period waves is consistent with a barrier-clearing model for triggering where local earthquakes are caused by a redistribution of pore pressure on a fault due to the unclogging of local flow paths by the seismically-induced flow. The long-period seismic waves may be more effective at clearing fractures than short-period ones because the longer the period of oscillation, the larger the total volume entrained. The result is that thicker barriers can be removed by long-period waves than short-period ones. The thicker barriers support larger head differences if all else is equal. Therefore, long-period waves may produce larger pressure changes on the fault than short-period ones of the same amplitude. Other potentially frequency-dependent mechanisms, such as rate and state friction, cannot explain the sustained triggering after the end of the seismic waves (Gomberg, JGR, 2001).
DE: 7209 Earthquake dynamics and mechanics
DE: 7223 Seismic hazard assessment and prediction
DE: 7230 Seismicity and seismotectonics
DE: 7260 Theory and modeling
SC: Seismology [S]
MN: 2003 Fall Meeting