Intraplate seismic hazard: Evidence for distributed strain and implications for seismic hazard - Intraplate Earthquakes - page 315

M w 7 event every 500 years, a rate that is consistent with the historic and prehistoric

moment release rate if the principal 1811-1812 events and earlier mainshocks were in the
range M w 6.7-7; or (2) The principal 1811-1812 events were significantly larger thanM w 7,
in which case deep-seated strain accrual processes in the NMSZ give rise to very low rates
of surface deformation that are close to the level of detectability with available GPS data
(e.g., see Kenner and Segall, 2000 ) . The latter interpretation could be effectively untestable,
at least in the foreseeable future, although increasingly precise results from GPS data will
continue to improve an upper bound on strain accrual. I suggest the former interpretation
is more likely, and better supported by available evidence.

12.3.2 Mechanism of strain accrual

The mechanism for localized strain accumulation in the NMSZ further remains enigmatic.
Several mechanisms have been proposed, including (1) stress concentration associated
with a Paleozoic mafic rift pillow (Grana and Richardson, 1996; Stuart et al ., 1997 ) ; (2)
glacial isostatic adjustment (GIA) following the removal of the Laurentide ice sheet (e.g.,
Grollimund and Zoback, 2001 ) ; (3) localized mantle flow driven by descent of the ancient
Farallon slab (e.g., Forte et al ., 2007 ) ; and (4) isostatic adjustment following the rapid
incision of the Mississippi River Valley (Van Arsdale et al ., 2007 ; Calais et al ., 2011 ) .
As discussed by Talwani (this volume), the predicted magnitudes of stresses generated by
the second and fourth mechanisms are low, which poses a problem for appealing to these
mechanisms as the explanation for Holocene NMSZ activity. However, the explanations are
attractive because they provide a qualitative and to some extent quantitative explanation for
why NMSZ activity has been concentrated in the late Holocene (e.g., Kelson et al ., 1996 ;
Tuttle et al ., 2002 ) . There is no question that GIA contributes a significant level of present-
day strain in eastern North America, not only over the region formerly covered by the
Laurentide ice sheet but also well to the south, in the forebulge region (e.g., James and Bent,
1994 ; Grollimund and Zoback, 2001 ) . Glacial isostatic adjustment is sometimes discounted
as the likely cause of seismic activity because the strain signal is spatially distributed
whereas significant moment release is apparently localized. Also, Wu and Johnston ( 2000 )
discounted the importance of the mechanism for NMSZ activity because earthquakes
triggered by GIA in this location are predicted to have predominantly thrust mechanisms.
However, unlike earlier interpretations of the 1811-1812 sequence (e.g., Johnston, 1996 ) ,
the sequence is characterized by a predominantly thrust mechanism in the interpretations
of Hough et al .( 2000 ) and Hough and Page ( 2011 ) . It is moreover plausible that GIA is a
necessary but not sufficient condition for seismic activity: it has long been proposed that
pre-existing zones of weakness such as failed rifts concentrate strain in intraplate crust
(e.g., Sykes, 1978 ) . Along the St. Lawrence Seaway in Canada, where the present-day GIA
signal is higher than in the New Madrid region and has been measured with GPS data,
Mazzotti et al .( 2005 ) show that Holocene seismic moment release is consistent with the
strain accrual associated with GIA (within the limitations of the short historical earthquake
catalog).