Geoscience Reference
In-Depth Information
assuming a
b
-value of 1, this corresponds to a M
w
6 or greater earthquake approximately
one every 5-10 years.
The above calculation, while highly simplified, can be compared with the observed rate
of large earthquakes in the CEUS. The rate of M
w
7 events is difficult to estimate,
so I instead consider the rate of M
w
6 events. For example, since 1850 there have
been approximately six CEUS earthquakes with magnitudes close to M
w
6, as well as
the 1886 mainshock: the largest aftershock of the 1886 Charleston earthquake (Talwani
2011 Mineral, VA, earthquake (e.g.,
http://neic.usgs.gov/neis/eq depot/2011/eq 110823
se082311a/se082311a l.html
,
last accessed 19 Dec. 2012). The observed rate of M
w
6
events between 1850 and 2013 has thus been one every 23 years, an obviously rough
estimate of long-term rate but one that is within a factor of 2-4 of the ballpark estimate.
If one instead considers the expected rate of M
w
5.8 events based on the above model,
since several of the historical earthquakes were slightly smaller than M
w
6, the expected
rate is higher (one every 3-6 years) than the ballpark estimate. I note, however, that (1) the
discrepancy is significantly smaller than any model that proposes to account for M
w
7.5
earthquakes every
500 years given best-available constraints on strain rate, and (2) both
the calculation and the estimation of observed rates are highly simplified and uncertain.
For example, the list of observed earthquakes since 1850 excludes five events that occurred
in Canada but within the region defined in
Figure 12.4
.
Other complications abound. For
(e.g., folding, or pressure solution) could account for a higher percentage of overall strain
release in a low strain rate region than an active plate boundary, reducing the level of
strain released seismically. For example, using calculations based on a simple model, they
show that, for a strain rate of 5
10
−
8
/yr, permanent deformation will reduce accrued
strain on the order of 30% over 3,000 years. For this and other reasons, a quantitative
assessment of strain accrual is highly uncertain. The key point, however, is that even a
low level of distributed strain accrual can generate a significant (and distributed) overall
hazard. That is, even a strain rate as low as 0.5
×
10
−
9
/yr would be sufficient to produce as
many or more large earthquakes as have been observed in historical times, if the strain is
distributed over a region covering roughly 10% of the CEUS. (While significant non-zero
strain rates have only barely been observed to-date in the NMSZ and Wabash region, a rate
on the order of 0.5
×
10
−
9
/yr would be below the current level of detection in other CEUS
×
regions.)
12.3.4 Other intraplate regions
A growing body of geological evidence reveals that many faults in low strain rate regions
generate large earthquakes that are clustered on timescales of a few centuries to perhaps a
few millennia, with much longer intervening periods when the fault might be considered
