In 1997 we published three papers on the evolution of stresses
in southern California from 1812 to 2025--two in JGR [Deng and
Sykes, 1997a, b] and one as Deng's PhD thesis [Deng, 1997]. In
all three, we included tectonic stress buildup for a large number
of active faults in southern California (e.g. many of those in
the Phase II report). In the calculations of Deng and Sykes [1997a]
we also included co-seismic stress drops for most shocks of M>7
since 1812. We showed that 95% of earthquakes of M>6 of known
or inferred mechanism occurred in regions of southern California
that were calculated to have moved closer to failure in terms
of changes in the Coulomb Failure Function (CFF). We concluded
that it is important to characterize earthquake triggering (like
stress) in a tensorial rather than a scalar sense. All changes
in CFF in this and the other two papers were calculated as a function
of time with respect to a baseline just before the large to great
shock of 1812 on the San Andreas fault.
Deng and Sykes [1997b] expanded that work by including co-seismic
stress drops for 36 shocks of M>6 since 1812 including the
large (M ~ 7.3) Laguna Salada event of 1892 in northern Baja California,
which was not modeled earlier. They calculated CFF and the triggering
history for moderate-, small- and micro-earthquakes of known mechanism
in southern California from 1812 to 1995. They found that more
than 85% of events of M>5 from 1932 to 1995 occurred in areas
that were calculated to have moved closer to failure, i.e. were
areas of positive CFF. Most other shocks of M>5 occurred close
to boundaries between areas of negative and positive CFF. More
than 85% of small (M>3) and micro M>1.8 earthquakes from
1981 until just before the Landers shock of 1992 of San Andreas-type
mechanisms occurred in regions that were calculated to have moved
closer to failure as well (Figure
1). The ratio of positive to negative
CFF at the epicenters of events of M>1.8 is very high for values
of the apparent coefficient of friction between 0.0 and 0.6 but
drops off rapidly for larger values of that coefficient. The highest
percentage of earthquakes occurred in regions where CFF was about
+ 1 MPa (10 bars) with respect to the 1812 baseline (Figure 2).
Hence, our work to date has greatly helped to identify the locations
of future moderate and large earthquakes as well as sites where
such events are unlikely to occur during the next few decades.
About half of the large and great historic earthquakes in southern
California occurred on faults other than the San Andreas itself.
Several of those event were located on so-called slow-moving faults,
i.e. faults with low long-term slip rates. While the stress evolution
for fast-moving faults can be easily estimated using simple models,
the stress interaction between fast- and slow- moving faults is
not fully understood. Earthquakes of the same magnitude on less
active faults, especially those in highly populated areas, can
be more damaging than those on some of the most active faults.
We found that the stress loading on a slow-moving fault is a combination
of a long period of slow increase associated with the fault itself
and one or more short periods of fast change (stress modulation)
associated with nearby fast moving faults. Deng and Sykes [1997a]
calculated changes in CFF as a function of time since 1812 for
the Newport-Inglewood and Landers faults, the sites of the 1933
Long Beach and 1992 earthquakes. By the times of those shocks
CFF increased appreciably along those faults as a result of stress
loading associated with nearby faults with much greater long-term
slip rates. CFF had increased by about 8 bars near the hypocenter
of the Landers earthquake, a significant fraction of the stress
accumulated in the 5000 years since the last similar Landers event.
The addition of the large 1892 Laguna Salada shock to our stress
evolutionary calculations led to an additional stress shadow to
the northwest of its epicenter, an area of faults with low long-term
slip rates. Thus, that area, which includes San Diego, has been
slow to recover from the stress shadow created by the 1892 shock.
Whether faults like the Rose Canyon fault near San Diego have
emerged from the stress shadow or not depends upon the parameters
chosen for the 1892 event.
The calculations we have performed thus far are for a purely elastic
half space. Changes in rheology are likely to be important when
the time scale of interest is years to hundreds of years. Deng
and Sykes (unpublished, 1997) put a large amount of effort in
developing a set of computer codes for Finite Element calculations
in Visco-Elastic Rheology (FEVER1.0) using a fully adaptive, multigrid
method, which consists of both self-adaptive refinement and adaptive
relaxation techniques. The computer codes are designed using modern
computer technology known as Object-Oriented Design, and implemented
in C++. These factors make finite element calculations not only
possible, but economic in terms of both computing time and memory,
with which we think no available commercial or non-commercial
packages can compete. FEVER1.0 can take into account any initial
strain conditions, boundary conditions, material types, temperature,
and fault displacements. The calculation can be performed for
either displacement or stress in a full 3-D space. Deng is now
using the program at Cal Tech as a SCEC post-doc.
Sykes was one of the co-organizers along with Ross Stein and Ruth
Harris of a workshop entitled "Earthquake Stress Triggers,
Stress Shadows, and their Impact on Seismic Hazard" that
was held at Menlo Park on March 21 and 22, 1997. It was co-sponsored
by SCEC and USGS.
Publications 1996-1997
Deng, J., and L. R. Sykes, Triggering
of 1812 Santa Barbara earthquake by a great San Andreas shock:
implications for future seismic hazards in Southern California,
Geophys. Res. Lett., 23, 1155-1158, 1996.
Deng, J., and L. R. Sykes, Evolution of the stress field in southern
California and triggering of moderate-size earthquakes: a 200-year
perspective, J. Geophys. Res., 102, 9859-9886, 1997a.
Deng, J., Stress Evolution and Earthquake Triggering in southern
California, PhD Thesis, Columbia University, 219 pages, 1997.
Deng, J., and L. R. Sykes, Stress evolution in southern California
and triggering of moderate, small, and micro earthquakes, J. Geophys.
Res., 102, 24,411-24435, 1997b.
Caption for Figure 1 (next page)
Fig. 1.
Change in Coulomb Failure Function, CFF, as of 1980 (as calculated
with respect to 1812 baseline). Focal mechanisms superimposed
on each subfigure are for events between 1981 and just before
Landers shock of 1992. a) CFF for vertical, right-lateral
faults striking 321o and mechanisms with strike, dip and rake
within 25o of those calculated and for events of M > 3.0. b)
same as a) except 1.8 < M < 3.0. c) same as a) except
strike = 270o and dip = 30oN. d) same as c) except 1.8
< M < 3.0. Note that most events occur areas of positive
CFF.
Fig. 2.
Histogram of frequency of occurrence of 2167 earthquakes from
Fig. 1b of San Andreas type of mechanisms and 1.8 < M <
3.0. Note that most events occurred areas of where CFF was about
10 bars (1 MPa).
