There have been a
number studies that have shown that the strong shaking in large earthquakes is
nonlinear (i.e that a significant portion of the
energy input is not transmitted, but instead lost to heating and damage to
rocks). Using a moving window cross-correlation analysis on the waveforms of
repeating earthquakes I have identified significant decreases in seismic
velocity coincident in time with the Mw 6.9 Loma Prieta Earthquake.
We find that the
magnitude of the velocity changes caused by Loma Prieta (as observed by NCSN
stations) correlates very nicely with the strength of shaking at the site.
This provides compelling
evidence that the velocity reductions observed at the NCSN stations were caused
by the strong shaking of the Loma Prieta earthquake damaging the near surface.
This is detailed in:
Rubinstein, J. L.
and G. C. Beroza (2004), Evidence for
Widespread Nonlinear Strong Ground Motion in the Mw6.9 Loma Prieta Earthquake
, Bulletin of the Seismological Society of America, Vol. 94, No. 5, pp.
1595–1608.
Susceptibility to Nonlinear
Strong Ground Motion
Six months after
the Loma Prieta earthquake, while the velocity reductions caused by Loma Prieta
were still healing, we observe another drop in seismic velocity, coincident
with the Ml 5.4 Chittenden Earthquake (the largest aftershock of Loma Prieta).
These velocity reductions are surprisingly large for such a small earthquake. The magnitude of the velocity reductions caused by Chittenden
correlate quite nicely with those caused by Loma Prieta.
This suggests that
once earth materials are damaged, they are more susceptible to further damage
than they would be in an undamaged state.
This study is
detailed in:
Rubinstein, J. L.
and G. C. Beroza (2004), Nonlinear
Strong Ground Motion in the Ml5.4 Chittenden Earthquake: Evidence that
Preexisting Damage Increases Susceptibility to Further Damage, GRL, v.
31, L23614, doi: 10.1029/2004GL021357.
Depth
Constraints on Nonlinear Strong Ground Motion
The 2004 M6
Parkfield Earthquake presented an excellent opportunity to study strong ground
motion. The wealth of data (strong ground motion, near surface, and borehole seismomters) allows for extensive analysis of the strong
ground motion and its effects. Our preliminary results show that near surface
stations (NCSN) have large S delays coincident with the Parkfield Earthquake
and borehole stations (HRSN) show no change in S Delays coincident with the
event. This suggests that delays are only accumulating very near surface (the
average depth of HRSN stations is 150m), and therefore nonlinearity is also
limited to the near surface as well.
The results of
this study are detailed in:
Rubinstein, J.L.
and G.C. Beroza (2005), Depth
Constraints on Nonlinear Strong Ground Motion from the 2004 Parkfield
Earthquake, GRL, vol. 32, L14313, doi:10.1029/2005GL023189.
Nonlinear Strong Ground
Motion and Fault Zone Damage
We have also studied the 2003 M8.0 Tokachi-Oki earthquake. Like the studies above, we observe velocity reductions caused by this earthquake. Similarly, we find the velocity changes to scale with site amplification and strong ground shaking. Unlike the previous studies, though, we find a path dependence of these velocity reductions. For those paths that cross the rupture zone of the earthquake we observe larger velocity reductions than those paths that don’t (yet we still do see velocity reductions for these paths as well).
This suggests that velocities were reduced in two locations: 1) at the near surface near our seismometers and 2) near the rupture zone. For those velocity reductions that occur near the rupture zone, it is unclear where in depth they are. We expect that the velocity reductions are either on the rupture zone itself where they were caused by the motion of the earthquake OR they are at the near surface and result from the strong shaking of the earthquake damaging the near surface sediments (i.e. nonlinear strong ground motion). The results of this study are detailed in:
Rubinstein, J.L.,N. Uchida, and G.C. Beroza (2007), Seismic Velocity Reductions Caused by the 2003 Tokachi-Oki Earthquake, JGR, vol. 112, B05315, doi:10.1029/2006JB00444.
Remaining Questions About Nonlinear Strong Ground Motion
The above studies
leave a few questions unresolved. Stay
tuned for future work which attempts to address them:
1) How do the
velocity reductions we observe compare to more standard measures of
nonlinearity in strong ground motion (i.e. strong ground motion analysis)?
2) How does the
strength of the velocity reductions (and therefore nonlinearity) caused by
medium magnitude earthquakes (like Parkfield and Chittenden) compare to the
strength of those caused by larger earthquakes (like Loma Prieta and
Tokachi-Oki)?
3) Can we see
increased susceptibility to damage, as a result of preexisting damage, anywhere
else (i.e. the lab or other locale?)