Intraplate seismic hazard: Evidence for distributed strain and implications for seismic hazard - Intraplate Earthquakes - page 311

using the same techniques that Nuttli ( 1973a , b) used to estimate magnitudes for the New
Madrid earthquakes. The intensity values determined by Bollinger ( 1977 ) have provided
the basis for later investigation using increasingly modern methodology. Johnston ( 1996 )
estimated M w 7.3

0.26 for the Charleston earthquake. Bakun and Hopper ( 2004 ) report
a preferred M w value of 6.9. Published M w values for the Charleston earthquake have thus
been relatively consistent: the U.S. National Hazards Mapping project currently assumes
a range between M6.8 and 7.5, with highest weight given to a value of 7.3 (Frankel et al .,
2002 ) .
Comparing the intensity distributions of the 1886 mainshock with the best characterized
intensity distribution from the 1811-1812 New Madrid sequence (the December main-
shock), Hough ( 2004 ) concludes that these events had comparable magnitudes. Based on
the results of Hough and Page ( 2011 ) , this suggests M w

+ / −

6.8 for the Charleston mainshock,
consistent with the estimate of Bakun and Hopper ( 2004 ) . This estimate is also consistent
with the results of Leon et al .( 2005 ) , who analyze liquefaction observations and estimate
magnitudes of Charleston paleoearthquakes of 6.5-7.0.

12.2.3 Historical earthquakes: recurrence rates

To estimate the seismic moment release rates for the NMSZ and CHSZ, it is necessary to
estimate both the size of the large historical events, which are assumed to be characteristic
earthquakes for their respective source zones, and their average recurrence rates. Sand
blows generated by prehistoric earthquakes have been identified in both the NMSZ and
Charleston regions. Although it is likely that future investigations will improve estimates
of recurrence times, studies to date have produced compelling evidence of prehistoric
earthquakes comparable to those in historic times. Paleoseismic investigations suggest a
repeat time on the order of 400-500 years over the past 2,000-3,000 years for both the
New Madrid sequence and the Charleston earthquake (Talwani and Cox, 1985 ; Talwani
and Schaeffer, 2001 ; Tuttle et al ., 2002 ) ; they also suggest that the NMSZ tends to produce
prolonged sequences with multiple, distinct mainshocks, the magnitudes of which are
comparable to those of the 1811-1812 events (e.g., Tuttle and Schweig, 1996 ; Tuttle et al .,
2002 ] .
As summarized by Grollimund and Zoback ( 2001 ) , this documented rate of Holocene
NMSZ activity cannot be representative of a long-term rate (e.g., Schweig and Ellis, 1994 ) ,
because extensive seismic reflection data reveal small cumulative offsets in post-Cretaceous
embayment sediments (e.g., Hamilton and Zoback, 1981 ) . Van Arsdale ( 2000 ) concluded
from a combination of trench and reflection results that Holocene slip rates on the Reelfoot
fault are at least four orders of magnitude higher than during the Pleistocene. Pratt ( 2012 )
suggests that pre-Holocene seismic activity involved faults other than those currently active
in the NMSZ. However, in all attempts to model NMSZ seismogenesis, including that of
Grollimund and Zoback ( 2001 ) , it is taken as an incontrovertible constraint that NMSZ
activity has “turned on” at some point in the relatively recent geological past, most likely
within the Holocene.