Karen R. Felzer, Rachel E. Abercrombie, and Göran Ekström

Annual Meeting, Eastern Section Seismological Society of America, 2002.
 

It has long been known that aftershocks may have their own aftershocks, known as secondary aftershocks.  The rate at which aftershocks trigger their own aftershocks provides information about how faults respond to multiple stressing, an important constraint on the physics of aftershock triggering.  Specifically, for example, if aftershocks are triggered by static stress changes, then all of the stress changes that a fault feels will contribute to it's final rupture time.  Stated differently, how a fault population reacts to a new stressor will be dependent on regional seismic history.  Therefore, aftershocks should produce their own aftershocks at a different rate than other earthquakes do.   Alternatively, if aftershocks are triggered by dynamic stresses, which are only temporary, the effects of different earthquakes may not be additive.  Instead, it is possible that aftershocks are triggered by the dynamic waves changing fault frictional properties, and that once the fault properties have been altered, additional dynamic waves will have little or no effect.  Therefore, if aftershocks are caused by dynamic stresses, then aftershocks may be expected to trigger their own aftershocks at the same rate that other earthquakes do.

Measuring the rate of secondary aftershocks may be difficult, however, since the early part of aftershock sequences can be quite active and dense.  We deal with this problem by delaying measurements for a short time after the initial mainshock; about 30 days after an M 5 mainshock in California, for example, it becomes possible to routinely identify secondary aftershocks.  We measure aftershock rates for a number of mainshocks in California which are not obviously themselves aftershocks of another earthquake, and for a number of earthquakes which do meet our criteria for being aftershocks.  We find that the rate at which aftershocks produce their own aftershocks is statistically indistinguishable from the rate at which other earthquakes produce aftershocks, after relative magnitudes are corrected for.  Or stated alternatively, we find that to the degree of accuracy we can measure, aftershocks always decay according to the modified Omori's Law, and the regionally averaged Omori's Law parameters remain constant, independent of time or recent seismic history.  This finding does not in itself prove dynamic triggering because there are so many unknown parameters and specifics involved in doing calculations with the static stress triggering hypothesis that exactly how much variation it should cause from a constant Omori's Law cannot be calculated; that is, its signal might be lost within our error bars.  However, this finding is consistent with the dynamic triggering hypothesis, and the test may be completed with more data and refined predictions of what we should expect under the static stress hypothesis.