Intraplate seismic hazard: Evidence for distributed strain and implications for seismic hazard - Intraplate Earthquakes - page 318

dormant. Examples (see Coppersmith [1988] and Crone et al . [2003] for summaries) include
the Meers fault in Oklahoma (Crone and Luza, 1990 ) , the clear scarp of which is inferred
to have been generated by two M

3,000 years, with
no evidence of prior events over 100,000 years or more (see Crone et al ., 2003 ) , and the
Cheraw fault in Colorado, which has produced three large earthquakes between 8 and 25 ka
BP.
Especially compelling geological evidence for temporally clustered seismic activity on
individual faults has been documented in Australia. A number of active faults in Australia
reveal behavior similar to that of the Meers fault (e.g., Crone et al ., 2003 ; Clark et al ., 2012 ) ,
including the Roopena fault in South Australia, the Hyden fault in Western Australia, and
the Lake Edgar fault in southwest Tasmania. To quote Clark et al .( 2012 ) , “A common
characteristic of morphogenic earthquake activity in Australia appears to be temporal
clustering. Periods of earthquake activity comprising a finite number of large events are
separated by much longer periods of seismic quiescence.” In eastern Utah, in a region east
of the Basin and Range province, where a precise strain rate estimate is not available but
thought to be low, Schwartz et al .( 2012 ) summarize evidence that the entire neotectonic
history of the Bear River fault consists of twoM6-6.5 surface-rupturing normal earthquakes
during the late Holocene.
In low strain rate regions of China, the especially long historical record reveals that large
earthquakes have occurred throughout a broad region, with no evidence for clustering or
even repeated large events in any individual source zone over the past

7 earthquakes during the past

700 years (Liu
et al ., 2011 ) . One key caveat is that paleoseismological investigations that might document
recurrence intervals have apparently not been undertaken at the source zones discussed by
Liu et al .( 2011 ) . However, with a historical record extending back to 1300, it is unlikely
that individual source zones in low strain rate regions in China are characterized by the
same type of clustering documented in the NMSZ and CHSZ. It is not clear why low strain
rate regions in China do not (apparently) generate the same type of clustering that appears
to be common in other regions.
One possible explanation for clustering is a feedback mechanism between an effectively
embedded finite fault in the upper crust and a viscoelastic lower crust (e.g., Kenner and
Segall, 2000 ) . In such a setting, an initial episode of strain release can trigger a sequence
of large events as post-seismic motions in a deep viscoelastic layer repeatedly reload an
individual fault in the brittle crust. A simplified analytical model predicts that the recurrence
interval will scale linearly with the viscosity of the weak lower crust (Kenner and Segall,
2000 ) . One might conjecture, then, that differences in degree of clustering in different
regions are due to differences in lower crustal viscosity. Lower crustal viscosity is poorly
constrained on a global basis. However, absolute plate motions are well established, and
are known to vary significantly among intraplate regions ( Table 12.1 . ) It has been shown
that plate motions are driven primarily by lateral forces (e.g., Forsyth and Uyeda, 1975 ;
Conrad and Lithgow-Bertelloni, 2002 ) , but drag on the bottom of plates influences plate
motion as well, and is stronger under continents than oceans (Forsyth and Uyeda, 1975 ) .
A relatively fast plate motion might therefore be expected to correlate to some extent with