Download Development of the next generation Australian National Earthquake

Proceedings of the Ninth Pacific Conference on Earthquake Engineering
Building an Earthquake-Resilient Society
14-16 April, 2011, Auckland, New Zealand
Development of the next generation Australian National Earthquake
Hazard Map
T.I. Allen, D.R. Burbidge, D. Clark, A.A. McPherson, C.D.N. Collins, and M.
Leonard
Earthquake Hazard Project, Geoscience Australia, Symonston ACT, Australia.
ABSTRACT: Geoscience Australia (GA) is currently undertaking a process of revising
the Australian National Earthquake Hazard Map using modern methods and an updated
catalogue of Australian earthquakes. This map is a key component of Australia’s
earthquake loading standard, AS1170.4. Here we present an overview of work being
undertaken within the GA Earthquake Hazard Project towards delivery of the next
generation earthquake hazard map.
Knowledge of the recurrence and magnitude (including maximum magnitude) of historic
and pre-historic earthquakes is fundamental to any Probabilistic Seismic Hazard
Assessment (PSHA). Palaeoseismological investigation of neotectonic features observed
in the Australian landscape has contributed to the development of a Neotectonic Domains
model which describes the variation in large intraplate earthquake recurrence behaviour
across the country. Analysis of fault data from each domain suggests that maximum
magnitude earthquakes of MW 7.0–7.5±0.2 can occur anywhere across the continent. In
addition to gathering information on the pre-historic record, more rigorous statistical
analyses of the spatial distribution of the historic catalogue are also being undertaken.
Earthquake magnitudes in Australian catalogues were determined using disparate
magnitude formulae, with many local magnitudes determined using Richter attenuation
coefficients prior to about 1990. Consequently, efforts are underway to standardise
magnitudes for specific regions and temporal periods, and to convert all earthquakes in
the catalogue to moment magnitude.
Finally, we review the general procedure for updating the national earthquake hazard
map, including consideration of Australian-specific ground-motion prediction equations.
We also examine the sensitivity of hazard estimates to the assumptions of certain model
components in the hazard assessment.
1 INTRODUCTION
The Australian National Hazard Map is an essential component to the Australian Standard 1170.4
which outlines actions for seismic design (Standards Australia 2007). The current hazard map on
which these seismic loading provisions are based was developed by McCue et al. (1993). The McCue
et al. (1993) map was constructed using modified source zones of Gaull et al. (1990), based on an
additional 6 years of earthquake data and about a hundred independent site hazard estimates. Since
these maps have been produced, many more earthquakes have occurred within Australia, augmenting
the catalogue to provide improved information on the likely recurrence of Australian earthquakes. In
addition, researchers now have a better understanding of the attenuation of earthquake ground-shaking
throughout the Australian crust. Furthermore, recent advances in palaeoseismological and neotectonic
activities have, for the first time, been able to quantify the maximum earthquake magnitude, Mmax,
likely to occur across different tectonic “domains”. Consequently, simply updating these
aforementioned hazard inputs could significantly improve the map and hence the standard.
Paper Number 207
Here we present an overview of work being undertaken within Geoscience Australia’s (GA’s)
Earthquake Hazard Project towards delivery of the next generation earthquake hazard map. Although
other activities are ongoing (Burbidge et al. 2010), this paper will focus on neotectonic studies,
catalogue improvement and ground-motion prediction.
2 PALAEOSEISMOLOGY AND NEOTECTONICS
Most probabilistic seismic-hazard assessments (PSHAs) require an estimate of Mmax – the magnitude
of the largest earthquake that is thought possible within a specified area – and are highly sensitive to
the choice of Mmax (e.g., Mueller 2010). By virtue of a fortuitous combination of climatic conditions,
geology and geomorphology, Australia boasts arguably the richest Quaternary faulting record of all
global Stable Continental Regions’ (SCRs) crust (Clark et al. in prep). Herein we explore the
hypothesis that Australia contains enough different geologic settings and enough samples within each
setting to indicate Mmax. Clark et al. (2010) propose a revised domain division of the Australian SCR
crust based upon geological, geophysical and neotectonics data, a simple analysis of Australian
neotectonic data from which preliminary Mmax estimates are derived, and a schema for pooling
domains that might be helpful in guiding the choice of appropriate ground motion models.
Six source zones (domains) spanning continental Australia have been proposed by Clark et al. (2010)
based upon geological, geophysical and neotectonic data (Fig. 1a).. In principle, each source zone
contains neotectonic faults that share common recurrence and behavioural characteristics, in a similar
way that source zones are defined using the historic record of seismicity. Over 200 fault scarps
captured in GA’s Australian neotectonics database have been used to produce preliminary estimates of
the maximum credible magnitude earthquake (Mmax) across the SCRs of Australia (Clark et al., 2010).
This was done by first grouping the scarps according to the spatial divisions described in the recently
published neotectonics domain model and calculating the 75th percentile scarp length for each domain
(Fig. 1b). Using this scarp length, Mmax was estimated by averaging the maximum magnitude predicted
from a range of different published relations (Wells and Coppersmith 1994 and others). Results range
between MW 7.0–7.5±0.2 across all of the domains. This analysis suggests that potentially catastrophic
earthquakes are possible Australia-wide.
3 REVISION OF CATALOGUE MAGNITUDES
The calculation of Australian earthquake magnitudes has been the topic of several focused workshops
and reports in the past (e.g., Denham 1982; McGregor and Ripper 1976), which have produced
recommendations for the calculation of earthquake magnitudes in Australia. The primary advances in
the development of Australian-specific magnitude formulae occurred in the mid 1980’s through to the
early 1990’s where much progress was made in developing Australian-specific magnitude formulae
which consider the attenuation properties of the Australian crust. It is well-documented that prior to
the development of Australian specific magnitude formulae, that the Richter (1935; 1958) local
magnitude equation – originally developed for southern California – was almost exclusively used to
calculate earthquake magnitudes throughout Australia (e.g., Leonard 2008). Figure 2 shows the
relative difference in the magnitude formulae between Richter and respective Australian ML equations
presently used to calculate magnitudes. As can be seen, there are considerable differences between
Richter and the Australian formulae, particularly at larger source-receiver distances. Below we discuss
attempts to reconcile in these differences in order to revise local magnitudes to be regionally
consistent with current observatory practice. Methods for converting revised catalogue magnitudes to
MW are also discussed.
2
Figure 1. (a) Neotectonic domains (Clark et al. in prep) with fault scarps from the Australian neotectonics
database overlain, (b) Box and whisker plot for fault length, which is a proxy for Mmax (Clark et al., 2010). Boxes
denote 75th and 25th percentiles, central point indicates median value, and whiskers define 90th and 10th
percentiles. D1 - Archaean Craton and non-reactivated Palaeo-Proterozoic; D2 – Sprigg Orogeny; D3 Reactivated Proterozoic; D4 - Eastern Australian Phanerozoic; D5 - SE Australian Rifted Crust; D6 - Extended
Continental Crust; D7 – extended passive margin crust.
3.1 Local Magnitude Corrections
Allen (2010) describes a method to correct catalogue ML’s to be consistent with contemporary
magnitude calculations for specified regions across Australia. The basic technique corrects magnitudes
using the difference between the Richter and local magnitude curves (Fig 2) at a distance determined
by the nearest recording station likely to have recorded the earthquake. For brevity we only discuss
results for western and central Australia (WCA) herein. Earthquakes within the WCA polygon (as
defined in Allen, 2010) were extracted from the Australian earthquake catalogue and local magnitudes
for these events were recalculated for pre-1990 earthquakes using the above procedure assuming the
Gaull and Gregson (1991) magnitude equation. Figure 3 indicates histograms of the residuals of the
3
catalogue ML’s minus the revised ML’s from the present study. When we consider all earthquakes
within the WCA polygon, we observe a small bimodal relationship of magnitude residuals, with the
largest peak at zero and a secondary peak at approximately 0.75 magnitude units (Fig.3a). This figure
shows that magnitudes for most earthquakes do not change significantly as a result of the present
revisions. However, if we only consider earthquakes that have catalogue (i.e. Richter) magnitudes ML
≥ 4.0, we observe a large secondary peak at approximately 0.75 magnitude units (Fig. 3b), with some
residuals up to one unit in magnitude. This figure suggests that many moderate-magnitude pre-1990
earthquakes occurred in remote locations at distances far from the nearest recording instrument.
Figure 2. Comparison of several Australian-specific –log A0 curves minus the Richter (1935; 1958) –log A0
curve: GS86 = Greenhalgh and Singh (1986); GG91 = Gaull and Gregson (1991); MLM92 = Michael-Leiba and
Malafant (1992); and WGW94 = Wilkie et al. (1994), updated using Wilkie (1996) coefficients. Most of the
Australian local magnitude –log A0 curves are similar to the Richter coefficients between 50 ≤ Repi ≤ 180 km. It
is well-acknowledged that the Richter (1958) curve underestimates attenuation (and magnitude) in southern
California at distances less than Repi 30 km (e.g., Bakun and Joyner 1984). This finding appears consistent with
the Australian –log A0 curves presented above.
Allen (2010) considered epicentres located within the southwest seismic zone of Western Australia
and noted that earthquake magnitudes in this region generally did not change markedly. The primary
reason for this is that the SWSZ had continuous monitoring for much of the period considered.
Consequently, most epicentres in the SWSZ were located within about 180 km from the nearest
seismic recorder. As previously mentioned the Australian-specific local magnitude equations were
generally similar to the original Richter (1958) equation in the distance range between 50 ≤ Repi ≤
180 km.
4
Figure 3. Histograms indicating residuals of catalogue magnitudes minus revised magnitudes based on
methodologies herein for (a) all of WCA region and (b) for the WCA region for ML (Richter) ≥ 4.0.
3.2 ML to MW conversion factors
Probabilistic earthquake hazard assessments rely on a catalogue of earthquakes that provide an
estimate of the recurrence of different-sized earthquakes per unit area. Commonly, these catalogues
are composed of disparate magnitude types calculated by different methods. These differences are
rarely considered by practitioners, particularly in stable continental settings. However, to obtain a
more reliable estimate of earthquake recurrence, it is prudent to convert all earthquake magnitudes to a
consistent magnitude type. Moment magnitude MW, has become the most commonly used magnitude
scale in earthquake hazard assessments because it provides a more physical measure of an
earthquake’s size. Consequently, most modern ground-motion prediction equations (GMPEs) are
calibrated to MW. The methodology above describes the harmonisation of Australian catalogue
magnitudes for ML only and does not consider MW. Consequently, we require conversion factors for all
magnitude types in the Australian earthquake catalogue.
Many equations have been developed to convert other magnitude scales to MW, most commonly from
surface-wave magnitude MS and body-wave magnitude mb (e.g., Scordilis 2006). However, providing
a universal conversion equation from local magnitude ML to MW cannot be easily achieved without
consideration of the local magnitude equation itself. Since magnitude is fundamentally a “source” term
(i.e., a measure of the energy release at the earthquake source), the divergence of –log A0 curves
(Fig. 2) at near-source distances effectively means that earthquakes of different source sizes can be
assigned the same ML. Consequently, MW conversion equations are heavily dependent on specific local
magnitude formulae.
Owing to the relative complexity in calculating moment magnitude relative to other magnitude types,
there are very few estimates of MW for Australian earthquakes. Figure 4 shows the conversion
equations used to revise local magnitudes to moment magnitude for the Australian earthquake
catalogue for: 1) eastern Australia, and 2) WCA. These local magnitude revisions and subsequent MLMW conversions will lead to a more homogeneous catalogue that can be incorporated into hazard
estimates.
Figure 4. ML to MW developed for Eastern Australia and Western and Central Australia, respectively. These
equations have been used to convert the catalogue of historical Australian earthquakes to moment magnitude for
the respective regions. Note, the eastern Australian conversion is not defined beyond ML 5.6.
5
4 AUSTRALIAN-SPECIFIC GROUND-MOTION PREDICTION EQUATIONS
There are now a handful of GMPEs that have been developed specifically to estimate ground-motions
from Australian earthquakes. In this section we indicate several GMPEs which are under consideration
for use in eastern Australia only. Allen (2010) shows a comparison of 5% damped response spectra for
a moment magnitude MW 6.5 earthquake at a range of rupture distances for several GMPEs, including
two models developed for southeastern Australia (McPherson and Allen 2006; Somerville et al. 2009).
In general, the GMPEs developed for eastern Australia appear most consistent with the western North
American GMPE of Chiou and Youngs (2008) rather than those from the stable continental interior of
eastern North America. This observation was also noted by Somerville et al. (2009).
Secondly, a selection of GMPEs are evaluated against recorded ground-motion data from the 6 April
2009 MW 4.5 Korumburra, Victoria earthquake, which was well recorded by permanent seismometers
throughout Victoria (Fig. 5). As would be expected, the eastern Australian GMPEs model the
observed ground-motion best overall. However, the faster attenuation rate approximated by
McPherson and Allen (2006) at longer-periods appears more consistent with recorded smallmagnitude data at shorter hypocentral distances. Note that some of the models used – including
Somerville et al. (2009) – are tested outside of the magnitude and distance threshold for which they
were developed. Finally, we note that the Chiou and Youngs (2008) GMPE developed for western
North America appears generally more similar to recorded data from the MW 4.5 Korumburra
earthquake (up to Rhyp = 150 km) than those models developed for the intraplate eastern North
America.
5 CONCLUSIONS
The Geoscience Australia Earthquake Hazard Project is currently undertaking a process of revising the
Australian National Earthquake Hazard Map using modern methods and an updated catalogue of
Australian earthquakes. A draft hazard map is planned to be released for stakeholder comment by mid2011. Once finalised, this map will become an essential input to the national earthquake loading
standards, AS1170.4.
Herein, we have briefly summarised some of the activities being undertaken by GA and other
researchers to provide an updated National Earthquake Hazard map for Australia. These include:
1. The development of a domains model that groups neotectonic faults according to common
recurrence and behavioural characteristics.
2. Estimation of Mmax based on faults scarp lengths in the neotectonic record
3. Revision of local magnitudes ML in the Australian earthquake catalogue from Richter
(1935;1958) to Australian-specific ML formulae (e.g., Gaull and Gregson 1991; Michael-Leiba
and Malafant 1992)
4. Conversion of local magnitudes to moment magnitudes MW leading to more homogeneous
catalogue that can be incorporated into hazard estimates
5. Development of Australian-specific GMPEs and incorporation of other modern global GMPEs
In addition to the aforementioned activities, GA is currently investigating several options for
earthquake source model inputs that will describe the statistical distribution of earthquakes across the
continent (e.g., Leonard, 2010). It is hoped that new insights into the characteristics of earthquake
occurrence and ground-motion attenuation in Australia will yield lower uncertainties on hazard
values than previous estimates.
6
Figure 5 (previous page). Comparison of several GMPEs against response spectra (black lines) recorded at
several stations for the 6 March 2009 MW 4.5 Korumburra, Victoria, earthquake. Solid and dashed lines indicate
east-west and north-south horizontal components respectively: AB95 = Atkinson and Boore (1995); T97 = Toro
et al. (1997); AB06 = Atkinson and Boore (2006); MA06 = McPherson and Allen (2006); CY08 = Chiou and
Youngs, (2008); S09 = Somerville et al. (2009). Rcd represents closest distance to rupture.
6 ACKNOWLEDGEMENTS
We would like to acknowledge Hadi Ghasemi and Jonathan Bathgate for providing internal reviews of
this manuscript. The Earthquake Hazard Project publishes with the permission of the Chief Executive
Officer of Geoscience Australia.
REFERENCES:
Allen, T. I. 2010. The influence of attenuation in earthquake ground-motion and magnitude estimation:
implications for Australian earthquake hazard. In Proceedings of the 2010 Australian Earthquake
Engineering Society Conference. Perth, Western Australia.
Allen, T. I. in prep. A review of historical Australian earthquake magnitude catalogues: reconciling earthquake
magnitudes for use in hazard analysis. In Geoscience Australia Record 2011/?? Canberra.
Allen, T. I., T. Dhu, P. R. Cummins and J. F. Schneider. 2006. Empirical attenuation of ground-motion spectral
amplitudes in southwestern Western Australia. Bull. Seism. Soc. Am. Vol 96(2), 572–585.
Bakun, W. H. and W. B. Joyner. 1984. The ML scale in central California. Bull. Seism. Soc. Am. Vol 74(5),
1827-1843.
7
Burbidge, D. R., T. Allen, C. Collins, A. McPherson, B. Drummond, M. Leonard, D. Clark, P. Cummins, R.
Cuthbertson, M. Dentith, B. Gaull, G. Gibson, M. Leiba, D. Love, K. McCue, P. Somerville, J. Rynn and J.
Wilson. 2010. Future directions for the National Earthquake Hazard Map for Australia. In Geoscience
Australia Record 2010/04. Canberra: Geoscience Australia.
Chiou, B. S.-J. and R. R. Youngs. 2008. An NGA model for the average horizontal component of peak ground
motion and response spectra. Earthq. Spectra Vol 24(1), 173–215.
Clark, D., A. McPherson and C. Collins. 2010. Mmax estimates for the Australian stable continental region
(SCR) derived from palaeoseismicity data. In Proceedings of the 2010 Australian Earthquake Engineering
Society Conference. Perth, Western Australia.
Clark, D., A. McPherson and C. D. N. Collins. in prep. Australia’s seismogenic neotectonic record: a case for
heterogeneous intraplate deformation. In Geoscience Australia Record 2011/?? Canberra.
Denham, D. 1982. Proceedings of the workshop on Australian earthquake magnitude scales. Bur. Min. Res. Rec.
1982/29, Canberra, Australia.
Gaull, B. A. and P. J. Gregson. 1991. A new local magnitude scale for Western Australian earthquakes. Aust. J.
Earth. Sci. Vol 38(3), 251-260.
Gaull, B. A., M. O. Michael-Leiba and J. M. W. Rynn. 1990. Probabilistic earthquake risk maps of Australia.
Aust. J. Earth. Sci. Vol 37(2), 169-187.
Greenhalgh, S. A. and R. Singh. 1986. A revised magnitude scale for South Australian earthquakes. Bull. Seism.
Soc. Am. Vol 76(3), 757-769.
Leonard, M. 2008. One hundred years of earthquake recording in Australia. Bull. Seism. Soc. Am. Vol 98(3),
1458–1470.
Leonard, M. 2010. Analysis of the Spatial Statistical properties of continental seismicity. In Proceedings of the
2010 Australian Earthquake Engineering Society Conference. Perth, Western Australia.
McCue, K., G. Gibson, M. Michael-Leiba, D. Love, R. Cuthbertson and G. Horoschun. 1993. Earthquake
Hazard Map of Australia – 1991. Canberra: Australian Geological Survey Organisation.
McGregor, P. M. and I. D. Ripper. 1976. Notes of earthquake magnitude scales. Bur. Min. Res. Rec. 1976/56,
Canberra, Australia.
McPherson, A. and T. Allen. 2006. An improved understanding of earthquake groundshaking in Australia. In
Proceedings of the 2006 Australian Earthquake Engineering Society Conference. Canberra.
Michael-Leiba, M. and K. Malafant. 1992. A new local magnitude scale for southeastern Australia. BMR J. Aust.
Geol. Geophys. Vol 13(3), 201-205.
Mueller, C. S. 2010. The influence of maximum magnitude on seismic-hazard estimates in the central and
eastern United States. Bull. Seism. Soc. Am. Vol 100(2), 699–711.
Richter, C. F. 1935. An instrumental earthquake magnitude scale. Bull. Seism. Soc. Am. Vol 25(1), 1-32.
Richter, C. F. 1958. Elementary Seismology. San Francisco: W H Freeman & Co, 578 pp.
Scordilis, E. M. 2006. Empirical global relations converting MS and mb to moment magnitude. J. Seismol. Vol
10, 225–236.
Somerville, P., R. Graves, N. Collins, S.-G. Song, S. Ni and P. Cummins. 2009. Source and ground motion
models for Australian earthquakes. In Proceedings of the 2009 Australian Earthquake Engineering Society
Conference. Newcastle.
Standards Australia. 2007. Structural design actions, part 4: Earthquake actions in Australia. Standards
Australia.
Wells, D. L. and K. J. Coppersmith. 1994. New empirical relationships among magnitude, rupture length,
rupture width, rupture area, and surface displacement. Bull. Seism. Soc. Am. 84(4), 974-1002.
Wilkie, J., G. Gibson and V. Wesson. 1994. Richter magnitudes and site corrections using vertical component
seismograms. Aust. J. Earth. Sci. Vol 41(3), 221-228.
Wilkie, J. R. 1996. Earthquake source parameters, Victoria, Australia. PhD Thesis, La Trobe University,
Melbourne.
8