Intraplate earthquakes in Australia - Intraplate Earthquakes - page 25

In general, greater topographic expression associated with faults and fault systems
occurring in extended crust relative to non-extended crust suggests a higher rate of seismic
activity in the extended setting, consistent with observations worldwide (e.g., Johnston,
1994 ; Cloetingh et al ., 2008 ; Mooney et al ., 2012 ) . Using the same reasoning, non-cratonic
crust might be expected to have a higher rate of seismic activity than cratonic crust (cf.
Figures 2 . 4 and 2.5), by virtue of there being no relief generation in cratonic crust. This
distinction, together with the variation in fault character between domains, should be
recognised in attempts to identify analogous systems worldwide.

2.4 Patterns in earthquake occurrence

The record of contemporary seismicity in Australia suggests that earthquake epicentres
are spatially and temporally clustered (Denham, 1988 ; Leonard, 2008 ; Sinadinovski and
McCue, 2010 ) . As discussed in Section 2.1, concentrations of epicentres in the historic cata-
logue are dominated by four “seismogenic zones” ( Figure 2.1 ) (Hillis et al ., 2008 ; Leonard,
2008 ; Sandiford and Egholm, 2008 ) . However, whether the short record of contemporary
seismicity is representative of time periods significantly longer than the observation win-
dow has not been statistically or empirically tested in the Australian context (cf. Kafka,
2002 ) . The persistence of patterns in the short historic catalogue can be assessed at much
longer timescales by comparison with the record of morphogenic seismicity. Evidence
from the paleo-record (Crone et al ., 2003 ; Clark et al ., 2011a , 2012), which essentially
captures events of

M > 5.5, suggests that large earthquake occurrence within Australia
exhibits both spatial and temporal clustering (Section 2.3). Temporal patterns in large SCR
earthquake occurrence may be inferred at the scale of a single fault (Crone et al ., 1997 ,
2003; Clark et al ., 2008 ) , of groups of faults (Leonard and Clark, 2011 ) , and at the domain
scale (Holdgate et al ., 2003 ; Sandiford, 2003b ; Paine et al ., 2004 ; Braun et al ., 2009 ; Clark
et al ., 2011a , 2012).
For example, the distribution of fault scarps in cratonic southwest Western Australia
(D1) ( Figure 2.4 ) , together with the seismogenic strain arguments referred to previously
(cf. Leonard et al ., 2007 ; Leonard, 2008 ; Braun et al ., 2009 ) , infer that seismicity in the
SWSZ represents only the current locus of activity, rather than a zone of long-lived activity
(cf. Sandiford and Egholm, 2008 ) . As most fault scarps in other parts of Australia are not
associated with historic seismicity (e.g., Crone et al ., 2003 ; Clark, 2010 ) , a similar rule
may apply.
There is no precedent in cratonic Australia (neotectonic domains 1 and 3 of Clark
et al., 2012 ) to indicate how long seismicity will persist. However, a large proportion of
contemporary seismicity in the SWSZ is thought to relate to the Calingiri, Cadoux, and
Meckering earthquakes (Leonard, 2008 ) . If aftershock activity relating to surface-rupturing
earthquakes is used as a measure of the longevity of activity in a region, then the work of
Stein and Liu ( 2009 ) suggests that a millennial timescale might be applicable. Liu et al .
( 2011 ) and Liu and Wang ( 2012 ) present evidence for migration of the locus of seismicity