Intraplate seismic hazard: Evidence for distributed strain and implications for seismic hazard - Intraplate Earthquakes - page 309

deformation, but rather is a steady-state process that drives both background seismicity as
well as occasional large earthquakes.) In either case, any inference of rupture length for
a historical earthquake based on present-day microseismicity patterns is clearly uncertain.
A defensible alternative constraint on rupture length, for example, discussed in detail by
Mueller et al .( 2004 ) and Hough and Page ( 2011 ) , is that NM3 rupture was bounded by
the intersection with the northern and southern limbs of the NMSZ ( Figure 12.3 ) rather
than the significantly longer rupture length assumed by a number of other researchers (e.g.,
Johnston and Schweig, 1996 ; Frankel et al ., 2012 ) . I discuss this issue at more length in a
later section.
The Charleston earthquake of September 1, 1886 - 9:50 p.m. LT on August 31, 1886 -
was the primary event in an apparently more typical earthquake sequence: a single large
mainshock preceded by a small number of foreshocks and followed by a conventional,
although spatially distributed, aftershock sequence (Dutton, 1889 ; Seeber and Armbruster,
1987 ) . As discussed by Hough ( 2004 ) , the overall felt extent of the mainshock was similar
to that of the principal 1811-1812 New Madrid earthquakes, although far better sampled
by extant archival accounts. As summarized by Talwani and Dura-Gomez ( 2009 ) and Dura-
Gomez and Talwani ( 2009 ) , available geophysical data combined with precise locations
of present-day seismicity reveals a complex fault system including the northeast-striking
Woodstock fault, an oblique right-lateral strike-slip fault with a

6 km long antidilational
left step through which the Sawmill Branch fault is the most active. As discussed by Dura-
Gomez and Talwani ( 2009 ) , detailed contemporary accounts point to the involvement of
multiple faults, likely including the Woodstock fault, in the 1886 Charleston mainshock.
Proposed rupture scenarios for this earthquake have been less detailed than those discussed
above for the 1811-1812 New Madrid sequence. Recent microseismicity does not appear
to delineate the historical mainshock rupture, and extensive liquefaction, as documented
initially by Dutton ( 1889 ) and later by Talwani and Schaeffer ( 2001 ) is distributed over a
broad source zone.

12.2.2 Historical earthquakes: magnitudes

The magnitudes of the principal 1811-1812 earthquakes are of critical importance for haz-
ard assessment and efforts to understand intraplate seismogenesis. As discussed by numer-
ous past studies (see Johnston and Schweig, 1996 , for summary), inferred rupture scenarios
and the size and distribution of liquefaction features provide some constraint on magni-
tude. There is no question that both the 1811-1812 mainshocks and the 1886 Charleston
earthquake generated extensive liquefaction (e.g., Fuller, 1912 ; Obermeier et al ., 1990 ;
Saucier et al ., 1991; Tuttle et al ., 2002 ; Dutton, 1889 ; Talwani and Cox, 1985 ) . Widespread
and enormous sand blows such as those in the NMSZ provide prima facie evidence for
large magnitudes. However, such magnitude estimates are not well constrained (e.g., Pond
and Martin, 1997 ; Stein and Newman, 2004 ) . Recently, Holzer et al .( 2011 ) presented
preliminary results from a new method to estimate peak ground acceleration (PGA) from
cone penetration test soundings, concluding that PGA values in the liquefaction zone of