Vous êtes sur la page 1sur 25

This article was downloaded by: [University of Tokyo]

On: 17 November 2013, At: 19:20


Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK

New Zealand Journal of Geology and Geophysics


Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tnzg20

The Mw 6.2 Christchurch earthquake of February


2011: preliminary report
a

A Kaiser , C Holden , J Beavan , D Beetham , R Benites , A Celentano , D


c

Collett , J Cousins , M Cubrinovski , G Dellow , P Denys , E Fielding , B Fry , M


a

Gerstenberger , R Langridge , C Massey , M Motagh , N Pondard , G McVerry , J


a

Ristau , M Stirling , J Thomas , SR Uma & J Zhao

GNS Science , Lower Hutt , New Zealand

e-GEOS/ASI , Matera , Italy

Land Information New Zealand , Wellington , New Zealand

University of Canterbury , Christchurch , New Zealand

School of Surveying , University of Otago , Dunedin , New Zealand

JPL/Caltech , Pasadena , California , USA

GFZ , Potsdam , Germany

Land Information New Zealand , Christchurch , New Zealand


Published online: 11 Jan 2012.

To cite this article: A Kaiser , C Holden , J Beavan , D Beetham , R Benites , A Celentano , D Collett , J Cousins , M
Cubrinovski , G Dellow , P Denys , E Fielding , B Fry , M Gerstenberger , R Langridge , C Massey , M Motagh , N Pondard , G
McVerry , J Ristau , M Stirling , J Thomas , SR Uma & J Zhao (2012) The Mw 6.2 Christchurch earthquake of February 2011:
preliminary report, New Zealand Journal of Geology and Geophysics, 55:1, 67-90, DOI: 10.1080/00288306.2011.641182
To link to this article: http://dx.doi.org/10.1080/00288306.2011.641182

PLEASE SCROLL DOWN FOR ARTICLE


Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of
the Content. Any opinions and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied
upon and should be independently verified with primary sources of information. Taylor and Francis shall
not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other
liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or
arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

New Zealand Journal of Geology and Geophysics


Vol. 55, No. 1, March 2012, 6790

The Mw 6.2 Christchurch earthquake of February 2011: preliminary report


A Kaisera*, C Holdena, J Beavana, D Beethama, R Benitesa, A Celentanob, D Collettc, J Cousinsa, M Cubrinovskid,
G Dellowa, P Denyse, E Fieldingf, B Frya, M Gerstenbergera, R Langridgea, C Masseya, M Motaghg, N Pondarda,
G McVerrya, J Ristaua, M Stirlinga, J Thomash, SR Umaa and J Zhaoa
a

GNS Science, Lower Hutt, New Zealand; be-GEOS/ASI, Matera, Italy; cLand Information New Zealand, Wellington, New Zealand;
University of Canterbury, Christchurch, New Zealand; eSchool of Surveying, University of Otago, Dunedin, New Zealand; fJPL/Caltech,
Pasadena, California, USA; gGFZ, Potsdam, Germany; hLand Information New Zealand, Christchurch, New Zealand
d

Downloaded by [University of Tokyo] at 19:20 17 November 2013

(Received 21 July 2011; final version received 1 November 2011)


A moment magnitude (Mw) 6.2 earthquake struck beneath the outer suburbs of Christchurch, New Zealands second largest city,
on 22 February 2011 local time. The Christchurch earthquake was the deadliest in New Zealand since the 1931 Mw 7.8 Hawkes
Bay earthquake and the most expensive in New Zealands recorded history. The effects of the earthquake on the regions
population and infrastructure were severe including 181 fatalities, widespread building damage, liquefaction and landslides. The
Christchurch earthquake was an aftershock of the Mw 7.1 Darfield Earthquake of September 2010, occurring towards the
eastern edge of the aftershock zone. This was a low recurrence earthquake for New Zealand and occurred on a fault
unrecognised prior to the Darfield event. Geodetic and seismological source models show that oblique-reverse slip occurred
along a northeastsouthwest-striking fault dipping southeast at c. 698, with maximum slip at 34 km depth. Ground motions
during the earthquake were unusually large at near-source distances for an earthquake of its size, registering up to 2.2 g (vertical)
and 1.7 g (horizontal) near the epicentre and up to 0.8 g (vertical) and 0.7 g (horizontal) in the city centre. Acceleration response
spectra exceeded 2500 yr building design codes and estimates based on standard New Zealand models. The earthquake was
associated with high apparent stress indicative of a strong fault. Furthermore, rupture in an updip direction towards
Christchurch likely led to strong directivity effects in the city. Site effects including long period amplification and near-surface
effects also contributed to the severity of ground motions.
Keywords: Canterbury earthquake sequence; Christchurch earthquake; strong ground motion; earthquake source model;
liquefaction; landslides

Introduction
A moment magnitude (Mw) 6.2 earthquake struck on 22
February 2011 (NZST) at shallow depth almost directly
below New Zealands second largest city (Christchurch,
population c. 377000) at 12:51 in the middle of the working
day. Extreme ground shaking was experienced, with recorded ground accelerations up to 2.2 g near the epicentre.
The impacts of the Christchurch earthquake were severe,
including 181 fatalities and collapse of some central city
buildings including office buildings and iconic heritage
structures. Liquefaction was widespread and lateral spreading, flooding and subsidence affected the eastern suburbs.
Rockfalls and slope failures impacted on hillside residential
areas causing further fatalities and rendering several hundred residences unsafe. In total, it is estimated that c. 900
buildings, mostly in the central business district (CBD), and
c. 10 000 residential homes may have to be demolished. This
was the most destructive earthquake to impact New Zealand
since the 1931 Mw 7.8 Hawkes Bay earthquake (Dowrick
1998). Total repairs are estimated to cost NZ$1520 billion,
*Corresponding author. Email: a.kaiser@gns.cri.nz
ISSN 0028-8306 print/ISSN 1175-8791 online
# 2012 The Royal Society of New Zealand
http://dx.doi.org/10.1080/00288306.2011.641182
http://www.tandfonline.com

making the Christchurch earthquake the most expensive in


New Zealands history.
The Mw 6.2 Christchurch event was the largest aftershock (to date) of the Mw 7.1 Darfield earthquake on 4
September 2010 (Gledhill et al. 2011). Prior to the Darfield
event, the Canterbury Plains had been an area of relatively
low seismicity for New Zealand since records began (Fig. 1).
Furthermore, no active faults had been mapped within c. 25
km of Christchurch. The Darfield earthquake was a low
recurrence event, producing a c. 30 km long surface rupture
across a Holocene and late Pleistocene age alluvial surface
along the previously unmapped Greendale Fault (Fig. 2;
Quigley et al. 2010, Quigley et al. 2011; Barrell et al. 2011).
Following the Darfield event, aftershock activity was
particularly concentrated at the eastern end of the Greendale
Fault (see green circles in Fig. 2), but the aftershock
sequence was initially somewhat less productive than forecast by time-dependent hazard models (Gerstenberger et al.
2005). Five and a half months after the main shock, the Mw
6.2 Christchurch earthquake occurred towards the eastern
fringe of the aftershock zone and c. 6 km southeast of

Downloaded by [University of Tokyo] at 19:20 17 November 2013

68 A Kaiser et al.

Figure 1 South Island of New Zealand and recorded seismicity


over a 10-year period until March 2010 (all recorded earthquakes
of magnitude  3; GeoNet catalogue; http://www.geonet.org.nz/
earthquake/resources/). Major active faults (GNS Active Faults
Database: http://data.gns.cri.nz/af/) are shown as black lines and
are referenced in the inset map (MFS  Marlborough Fault
System; thick black lines indicate subduction margins). Inset map
is shaded according to topography (warm colours indicate higher
elevation).

Christchurch city centre (Fig. 2). The earthquake occurred


on a previously unmapped northeastsouthwest-striking
fault and did not rupture the surface. Coulomb Failure
Stress (CFS) loading of the fault from the September
Darfield earthquake is estimated to be B0.1 MPa, a
relatively small increase in loading compared to elsewhere
in the region (1 MPa). Several strong aftershocks of
magnitude 5 followed the Christchurch earthquake, including a Mw 6.0 event on 13 June 2011 that produced high
horizontal accelerations (c. 2 g) at the southeastern edge of
the city (epicentre near the suburb of Sumner, Fig. 3).
The Christchurch earthquake was far more destructive
than the Darfield main shock, primarily due to its proximity
to Christchurch. Globally, other large devastating earthquakes that have occurred close to urban centres have
included 2010 Port-au-Prince, Haiti (Mw 7.0), 1995 Kobe,
Japan (Mw 6.8), 1976 Tangshan, China (Mw 7.6) and 2003
Bam, Iran (Mw 6.6). However, the Christchurch earthquake
was notable in that extremely high accelerations were
observed. Recorded peak ground acceleration (PGA)
reached up to 2.2 g (vertical) and 1.7 g (horizontal) at
Heathcote Valley (HVSC) c. 2 km from the epicentre and up
to 0.8 g (vertical) and 0.7 g (horizontal) in the central city
(Fig. 3). Vertical accelerations were particularly strong and
rich in high-frequency energy. Peak accelerations were the
highest recorded in a New Zealand earthquake and among
the highest recorded worldwide; a similar analogue globally
is the 2008 Mw 7.2 Iwate-Miyagi, Japan earthquake with
recorded near-field vertical PGA  3.9 g (Suzuki et al.
2010). In this context it must also be noted that, in
comparison to other historical New Zealand earthquakes

Figure 2 The Mw 6.2 Christchurch earthquake (red star) in the context of the Dareld main shock (green star) and aftershock sequence up
until 04 September 2011. Surface traces of active faults are shown in red (GNS Active Faults Database: http://data.gns.cri.nz/af/; offshore
fault data: P. Barnes, NIWA). Surface projection of buried ruptu associated with the Mw 7.1 Dareld earthquake (Beavan 2010a; Holden
2011), the Mw 6.2 Christchurch earthquake and the preliminary result for the Mw 6.0 June aftershock are shown as yellow dashed lines.
Coordinates are New Zealand Map Grid (m).

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 69

Figure 3 Map of the Christchurch urban area showing maximum peak ground accelerations (vertical and horizontal vector components)
recorded at GeoNet national and regional network seismic stations (labelled) and temporary low-cost Quake-Catcher Network (QCN;
Cochran et al. 2011) accelerometers. Acceleration recorded at selected GeoNet stations is plotted below for the three orthogonal components.

(and most earthquakes globally), the Christchurch earthquake was exceptionally well recorded in the near-field by
the dense national network (GeoNet, including the regional
CanNet network; Avery et al. 2004; Petersen et al. 2010)
and temporary low-cost accelerometers (Quake-Catcher
Network; Cochran et al. 2011). The strong-motion data
provide a rare and globally important near-field dataset for
seismological studies.
We present a preliminary report summarising the main
characteristics of the 22 February 2011 Christchurch Earthquake. We present focal mechanism solutions and preliminary source models derived from geodetic and strong-motion

data. We analyse and discuss features of the strong ground


motions and examine the major effects of the earthquake on
the built environment of Christchurch. Finally, we analyse
the earthquake briefly in the context of current seismic
hazard models.

Tectonic setting and past seismicity


In the central South Island of New Zealand, oblique
continental convergence of 38 mm/yr is accommodated
between the Pacific and Australian plates (DeMets et al.
2010). The plate boundary is delineated by the Alpine Fault

Downloaded by [University of Tokyo] at 19:20 17 November 2013

70 A Kaiser et al.
c. 140 km to the west of Christchurch, which links two
subduction zones of opposite polarity to the north and south
(Fig. 1). Up to three-quarters of the relative plate motion is
taken up in a narrow zone along the Alpine Fault, with
dextral and reverse slip rates up to 25 mm/yr and 10 mm/yr,
respectively (Sutherland et al. 2006; Norris & Cooper 2007).
In the northern South Island, plate motion is largely taken
up by the strike-slip faults of the Marlborough Fault System
(MFS); slip rate on the Hope Fault at the southern end of
the MFS is c. 20 mm/yr (Cowan 1991; Van Dissen & Yeats
1991; Langridge & Berryman 2005). The zone of active plate
boundary deformation has widened eastwards into the
Canterbury Plains during the Quaternary (Forsyth et al.
2008). The strain rate within the Canterbury block is
estimated from GPS-derived velocity fields to be nearly
uniaxial contraction of c. 1610 9 per year in a WNW
direction (Wallace et al. 2007). This corresponds to c. 2 mm/
yr over the 120 km distance between the Porters Pass Fault
in the foothills of the Southern Alps and the offshore limit of
deformation.
While active faults have been mapped in the foothills of
the Southern Alps to the west of the Canterbury Plains
(reviews in Pettinga et al. 2001; Stirling et al. 2008),
thicknesses of Quaternary alluvium deposited by gravelladen braided river systems may mask evidence of low-rate
active tectonic structures beyond the range front (Forsyth et
al. 2008). Evidence for active faulting and folding in the
Quaternary has been identified beneath the generally flatlying plains from high-resolution active-source seismic
investigations, outcrops and subtle geomorphic features
(e.g. Jongens et al. 1999; Estrada 2003; Finnemore 2004;
Forsyth et al. 2008; Dorn et al. 2010). However, active
tectonic structures in the immediate vicinity of Christchurch
were largely unknown prior to the Darfield Earthquake and
aftershock sequence, with the closest known active faults
located c. 25 km to the north of Christchurch (e.g. the
Springbank and Pegasus Bay faults described in Barnes
1996; Forsyth et al. 2008).
Despite the Canterbury regions relatively low seismicity
levels prior to the Darfield earthquake, several historical
events have previously produced low-to-moderate ground
shaking in Christchurch. Large magnitude 67 earthquakes
have occurred in the Southern Alps and foothills to the west
and north of the region in the past 150 years, for example
1888 magnitude (M) 7.1 North Canterbury, 1901 M 6.9
Cheviot, 1929 M 7.0 Arthurs Pass, 1944 M 6.7 Arthurs
Pass and 1995 M 6.2 Cass (Cowan 1991; Doser et al. 1999;
Abercrombie et al. 2000; Gledhill et al. 2000; Pettinga et al.
2001). Moderate-sized events have also occurred in the
Christchurch region, most notably a shallow earthquake in
1869 c. 10 km from Christchurch city centre and an event
further south in 1870 located near Lake Ellesmere. Both of
these events occurred on unknown (buried) faults
and produced shaking of intensity MM7 in Christchurch
(Pettinga et al. 2001).

Tectonic and seismological evidence outlined above


suggests that the Darfield main shock and Mw 6.2 Christchurch earthquake were relatively rare events for New Zealand
with long recurrence intervals. The Darfield earthquake
rupture was in fact complex, occurring on intersecting and
subsidiary blind thrust faults as well as the dominant east
west strike-slip Greendale Fault (Beavan et al. 2010a; Holden
et al. 2011). Faults in the foothills of the Southern Alps
generally strike northeastsouthwest, representative of the
current stress field, whereas older Early Cretaceous faults
found offshore are oriented eastwest (Wood & Herzer 1993).
The Canterbury region has an anomalous crustal structure,
where a transition to older subducted crust of the Hikurangi
Plateau at an ancient (c. 100 Ma) plate boundary is inferred at
c. 10 km depth beneath greywackes underlying the sedimentary deposits of the Canterbury Plains (Forsyth et al. 2008;
Reyners 2011). The orientation of the Greendale Fault and
other faults in Canterbury has suggested that the recent
earthquakes may have occurred on reactivated eastwesttrending Cretaceous faults at this ancient subduction margin
that are present in the upper crust (Reyners 2011). The extinct
basaltic shield volcano of Banks Peninsula outcropping to the
south of Christchurch (Fig. 2) may have also played a role in
concentrating the stress field following the Darfield earthquake (Reyners 2011).

Earthquake source characterisation


Focal mechanism solutions
High quality three-component seismic broadband data
recorded by the GeoNet network were used to calculate
regional moment tensor (RMT) solutions for the 22
February Christchurch earthquake and the large number
of aftershocks. A brief description of the moment tensor
method is given here, and Ristau (2008) provides a detailed
overview of the RMT method in New Zealand. RMT
solutions are routinely calculated by GeoNet for New
Zealand earthquakes with M ]3.54.0 using code developed at the University of California, Berkeley Seismological
Laboratory (Dreger & Helmberger 1993; Pasyanos et al.
1996; Dreger 2003). RMT analysis utilises only regional
waveform data (source-receiver distances of Bc. 1000 km)
and velocity models specific to the source region to calculate
synthetic Greens functions. The observed waveforms and
synthetic Greens functions are then inverted to obtain the
moment tensor elements. The focal mechanism and Mw are
derived from the moment tensor elements.
RMT solutions for the Mw 6.2 Christchurch earthquake
and 73 aftershocks with Mw 3.66.0 are shown in Fig. 4. The
RMT solution for the Christchurch earthquake is an obliquereverse faulting mechanism and the trend of aftershocks (red
dots in Fig. 2) is consistent with the NESW-striking fault
plane dipping to the southeast. The fault orientation is
further constrained from geodetic inversions presented in

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 71

Figure 4 Focal mechanisms derived from regional moment tensor


(RMT) solutions calculated for the Christchurch earthquake
sequence (lower hemisphere projections are plotted). The Mw 6.2
Christchurch earthquake is an oblique-reverse faulting mechanism
and the aftershocks (greyscale according to the type of mechanism)
are mainly strike-slip or oblique-reverse faulting. The largest
aftershock was a Mw 6.0 strike-slip event on 13 June 2011, located
c. 4 km east of the 22 February Christchurch earthquake epicentre.

the next section. Fig. 5 shows waveform fits and changes in


variance reduction and focal mechanism with depth. The
observed waveforms and Greens functions were bandpass
filtered at 0.010.033 Hz for the 22 February event, and
typically in a sub-range of 0.020.1 Hz (e.g. 0.020.05 Hz;
0.030.1 Hz) for other aftershocks depending on the signalto-noise ratio. For all events, solutions are calculated over a
range of depths to find the depth with the largest variance
reduction between the observed and synthetic waveforms.
Focal mechanisms of aftershocks are mainly strike-slip or
oblique-reverse faulting. P-axes are either consistent with the
Mw 6.2 event and the regional P-axis trend for crustal events
in the South Island or else rotated clockwise and similar to
the P-axis of the Mw 7.1 Darfield earthquake. Centroid
depths are almost all very shallow at around 28 km. Most of
the aftershocks are located close to the 22 February
Christchurch earthquake; however, a few trend SE through
Banks Peninsula. The largest aftershock (13 June, Mw 6.0)
was a strike-slip event c. 4 km east of the 22 February event.
Preliminary geodetic and strong-motion data modelling are
currently being conducted to distinguish between the possible
conjugate fault planes (NNWSSE or ENEWSW) and
determine the best-fit fault plane parameters.

Geodetic inversion
A variety of GPS and satellite interferometric radar data are
available to constrain the ground displacement caused by
the Christchurch earthquake. For the preliminary model

described here, we use campaign GPS data collected


between 28 February and 14 April 2011 from 57 sites that
were also occupied following the September 2010 Darfield
earthquake (Beavan et al. 2010a). We also use continuous
GPS data from five sites operated by GeoNet for Land
Information New Zealand (LINZ) and eight sites in
Christchurch operated by private companies. We use
differential interferometric synthetic aperture radar (DInSAR) from the Italian Cosmo-SkyMed (CSK) satellite, both
from an ascending track with images collected on 19
February and 23 February, and a descending track with
images collected on 20 February and 16 March. We present
one of the DInSAR images in Fig. 6. The GPS data are not
shown here; a complete set of GPS data including tabulated
numerical values can be found in Beavan et al. (2011).
The GPS data are processed by standard methods (e.g.
Beavan et al. 2010a) constraining regional Australian and
Pacific sites to their IGS05 coordinates at the epoch of survey.
The resulting daily coordinate and covariance files are
inverted simultaneously for coordinates and displacements
using ADJCOORD software (Bibby 1982; Crook 1992). Plate
tectonic and reference frame motion in the c. 6 months
between surveys is accounted for by solving in the inversion
for the rigid-body translation and rotation of three far-field
sites nominally on the Pacific Plate: CHAT (Chatham Island),
KAIK (Kaikoura) and WAIM (Waimate).
The single-look complex CSK ascending data were
processed using the SARscape software (http://www.
sarmap.ch/), while the descending CSK data were
processed from raw data with the JPL/Caltech ROI_pac
software. The topographic contribution to the interferometric phase was removed using a 3 arcsec digital
elevation model (DEM) from the Shuttle Radar
Topography Mission (SRTM). The DInSAR interferogram
phases were then filtered using a weighted power spectrum
technique (Goldstein & Werner 1998), unwrapped using a
minimum cost flow algorithm (Chen & Zebker 2002) and
projected to a geographic grid using the SRTM DEM.
GPS data provide a 3D coseismic displacement vector at a
set of points, while DInSAR provides coseismic ground
displacement in the line-of-sight from the ground to the
satellite at points throughout the interference pattern,
provided the pattern remains coherent. The interference
pattern is however incoherent in the processed interferograms
(e.g. Fig. 6) throughout substantial parts of Christchurch,
presumably because the reflective properties of the ground
changed due to ground failure and/or building damage.
The displacement data were modelled in two steps
(Arnadottir & Segall 1994) to solve for the location and
geometry of the fault plane and the amount and direction of
slip. First, we used non-linear inversion software disloc99
(Darby & Beavan 2001) to solve for the best-fitting uniformslip rectangular fault plane in an elastic half-space, using the
GPS data only. We found that it was only necessary to
constrain the lower depth of faulting to obtain a solution for

Downloaded by [University of Tokyo] at 19:20 17 November 2013

72 A Kaiser et al.

Figure 5 RMT solution for the 22 February Christchurch earthquake. Map shows the location of the earthquake, best-t solution (lower
hemisphere projection) and the stations used to calculate the solution. Bottom shows the waveform ts of each station for the best-t solution
bandpass ltered at 0.010.033 Hz. Inset shows the variance reduction versus depth and the change in focal mechanism with depth. The best
t is at a depth of 4 km and the parameters for the best-t solution are indicated in the gure.

the other eight parameters describing the fault. We then


extended the fault plane a few kilometres in all directions
and performed a linear inversion for slip on this fault plane
using both the GPS and DInSAR data, again assuming an
elastic half-space model. We also performed a grid search in
the vicinity of this model, varying the location, strike and
dip of the fault to find the lowest chi-squared solution. We
used inversion software based on Jonsson et al. (2002) as
described by Beavan et al. (2010b). The resulting fault
location and slip distribution is shown in Fig. 7.

The modelled fault (Fig. 7) runs from near Cashmere at its


southwest end towards the Avon-Heathcote Estuary and a
few kilometres offshore, with a strike of 588 and a dip of 698 to
the southeast. The main patch of slip, with maximum
magnitude c. 2.4 m, is centred at a depth of c. 4 km beneath
the estuary and is a mix of reverse faulting and right-lateral
strike slip. The slip on the southwestern part of the fault, with
a maximum of c. 1 m, is predominantly right-lateral strike slip.
The modelled solution is equivalent to an earthquake of
moment magnitude 6.3. The geodetic data were collected

Christchurch earthquake, February 2011 73

Unwrapped residual

40
20

-20

LOS, mm

-40

10 km

Central
Christchurch

Ground displacement
away from satellite

ction

t dire

Fligh

Downloaded by [University of Tokyo] at 19:20 17 November 2013

10 km

n
ectio
k dir = 36
o
o
l
ar
gle
Rad nce an
e
d
i
c
n
I

Ground displacement
towards satellite

Figure 6 The coloured image shows an interferogram derived from X-band radar images acquired on 19 and 23 February 2011 by the Italian
Cosmo-SkyMed satellite. Each colour cycle (fringe) represents 1.55 cm of ground displacement in the direction from the ground to the
satellite, so the total line-of-sight displacement between the western edge of the image and central Christchurch is about 25 cm. The order of
the colours indicates whether ground displacement is towards or away from the satellite, as indicated in two regions by white text. The image
becomes incoherent in central and eastern Christchurch, presumably due to ground and building damage. The inset shows the residual
between the unwrapped interferogram and the model prediction. Residuals close to the fault are up to about 945 mm or 93 fringes in the
observed interferogram; we expect that at least some of the negative residuals are due to non-tectonic ground subsidence. As well as the data
from this ascending track, where the satellite is ying north-northwest, we use data from a descending track with images acquired on 20
February and 16 March. The Cosmo-SkyMed original data, acquired in 2011, are a copyright product of the Italian Space Agency (ASI) and
are distributed by e-GEOS, an ASI/Telespazio Company.

some weeks after the earthquake so will contain a contribution from aftershocks, in particular the magnitude 5.8 and 5.9
events on 22 February. Beavan et al. (2011) use additional
geodetic data to generate a more complex model consisting of
two faults: one fault similar to that shown here and a second
smaller fault coincident with the major aftershocks.
A point of interest for the Christchurch recovery is the
amount of uplift and subsidence caused by the earthquake.
In Fig. 7A we show contours of vertical displacement
predicted by the fault model. These are the uplift and
subsidence that would be predicted by the model if the

ground were composed of competent rock; there may be


additional subsidence in regions where ground failure such
as liquefaction and lateral spreading has occurred. The
model shows subsidence of the suburbs north and northwest
of the estuary by up to 1015 cm, but uplift of the estuary
itself by up to 40 cm. These values are sensitive to some
details of the modelling such as the chosen top depth for the
faulting (here 1 km), the smoothing applied in the inversion
and the elastic half-space assumption. However, based on
studies such as that of Amoruso et al. (2004), we are
confident they are correct to within about 925%. These

Downloaded by [University of Tokyo] at 19:20 17 November 2013

74 A Kaiser et al.

Figure 7 Geodetic source model of the Christchurch earthquake. A, The model fault location and slip magnitude. The fault dips to the
southeast with its top edge at 1 km depth. Filled black circles show the near-eld GPS stations contributing to the solution with observed
(blue arrows with 95% condence uncertainty ellipses) and modelled (red arrows) displacements. For the light blue arrows, the uncertainties
are greater than 100 mm and are not shown; these sites contribute little to the solution. The red-and-white star shows the epicentre as located
using data available in March 2011. The grey contours show uplift and subsidence in millimetres predicted by the model. The black square
labelled Chch shows central Christchurch; C and AHE show Cashmere and the Avon-Heathcote Estuary respectively. B, The slip
distribution on the model fault plane for the hanging wall relative to the footwall is shown by the coloured image and slip vectors. The redand-white star shows the estimated hypocentre projected onto the fault plane.

ground movements, even without additional subsidence due


to ground failure, will exacerbate flooding problems in the
low-lying suburbs north and northwest of the estuary.

Strong-motion data inversion


To solve for a detailed slip and time distribution of
earthquake rupture, we inverted data from 13 strong-

motion sites of the GeoNet network, from both the


national network and the regional Canterbury network
CanNet. The source-station distance ranged 220 km.
Seismograms were bandpass filtered between 0.1 and 0.5
Hz and integrated once into velocity data. We modelled
one fault plane with a 598 strike and 698 southeast dip as
defined by the geodetic inversion solution (see above). The
fault plane area is 20 20 km2, divided into 400 1 1 km2

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 75


subfaults. We inverted for one elliptical rupture area
distributed randomly within the fault plane and with
variable rupture velocity, slip and rake. We focused on
fitting the onset as well as the amplitude of the near-source
stations. Our best slip distribution (Fig. 8) is characterised
by a rupture area of c. 12 6 km2 dimension, with a
maximum slip of 3.6 m at 3.5 km depth on the main fault
plane occurring after 2.5 s. The rupture initiates at around
10 km depth and propagates updip reaching a depth of c.
500 m. The rupture velocity obtained from the inversion is
3.0 km/s. The inverted average slip direction (or rake
angle) is 1338 and is consistent with the slip model derived
in the geodetic solution (Fig. 7).
The moment value calculated for that solution is
3.721018 Nm, equivalent to a magnitude 6.3 earthquake.
However, this solution only fits the early part of the
seismograms and suggests that a secondary source is
involved in the earthquake process. This solution is
consistent with detailed strong-motion data inversions
based on data filtered up to 1 Hz discussed in Holden
(2011).

Strong ground motions


Data analyses
Accelerograms were recorded by 129 of the 280 GeoNet sites
operational at the time of the Christchurch earthquake (Fig. 9,
Table 1). This strong-motion dataset will be important
globally for many reasons, particularly as it provides rare
densely spaced observations of near-source ground motions
(Fig. 3). It is also the most significant dataset recorded in
New Zealand since strong-motion recording began here in
the early 1960s. Since then, there have been 61 New Zealand
records with PGA greater than 0.3 g. Sixty-nine percent of
these are from the 22 February Christchurch earthquake
and aftershock sequence, 23% from the Darfield Earthquake and immediate aftershocks and just 8% from other
earthquakes. In addition, two building response records
exceeding 0.3 g PGA were recorded in the Christchurch
event (0.84 and 0.46 g) on multiple floors of the Christchurch Police Station. Of particular importance are the 17
ground and structural records from within 10 km of the fault
plane (Tables 1 and 2). To give a New Zealand perspective
to the importance of that data, nearly all of urban

Figure 8 Slip and rupture time distribution derived from strong-motion data inversion. A, Location and magnitude of slip. The slip
distribution is characterised by a patch of high slip (maximum 3.6 m) occurring north and updip of the hypocentre with an oblique-reverse
mechanism. B, Slip and rake history for the fault plane; slip is shown in colours and rake is represented by black vectors for each grid cell.
Distances are in kilometres and rupture time iso-contours (white) are in seconds. C, Observed (black) and synthetic (red) strong-motion
seismograms for the closest stations used in the inversion. Accelerations were bandpassed in the frequency range 0.010.5 Hz using a centred
Butterworth lter, then integrated once into velocity data. Values above the traces are the absolute peak velocity in m/s.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

76 A Kaiser et al.

Figure 9 Locations of GeoNet strong-motion sensors at the time of


the Christchurch earthquake, with peak ground acceleration
(PGA) levels indicated where strong-motion recordings were
obtained. The concentric circles show epicentral distance.

Wellington Region is within 10 km of one or more active


faults. The recorded ground motions, combined with building construction and ground observations, will provide a
unique opportunity to explain and then reduce levels of
damage in future large earthquakes.
Routine processing has been carried out on the asrecorded strong-motion acceleration time histories; raw and
processed data are available on the GeoNet website (www.
geonet.org.nz). Here, we have re-filtered records within 100
km epicentral distance using a high-pass initiation frequency
of 0.1 Hz (compared to 0.25 Hz used in standard GeoNet
processing) to better preserve the genuine long-period
ground motions (data available at: ftp://ftp.geonet.org.nz/
strong/processed/Proc/2011/02_Christchurch_mainshock_
extended_pass_band/). This re-filtering did not greatly affect
the peak accelerations, but resulted in better estimates (and
quite large increases) for the peak velocities and displacements of many of the records.
Near-source peak horizontal accelerations (at distances B5 km from the fault) were in fact stronger in the
Christchurch event than the Darfield main shock (Fig. 10;
analysis of Darfield earthquake data in Cousins & McVerry
2010). The horizontal peak ground motions at greater
distances from the fault plane were however stronger
in the Darfield earthquake, in keeping with its larger
magnitude.
Attenuation was clearly stronger in the Christchurch
earthquake than in the Darfield earthquake, as expected for
the approximately one magnitude step difference between

the two events. Not only are the expected motions stronger
at a given distance for a larger magnitude event, but the
ratio of amplitudes relative to a smaller event is expected to
increase with distance because of saturation effects near the
source. As depicted in Fig. 10, differences in attenuation
rates are more pronounced for displacements and velocities
than for accelerations. This is commensurate with expectations given that accelerations are generally governed by
high-frequency motions that die away relatively rapidly with
distance compared to low-frequency motions that are
generated to a greater extent in larger events.
A crucial point is that most of central and eastern
Christchurch was within 5 km distance from the fault source
of the Christchurch earthquake, compared with the distance
range of c. 20 km for the Darfield event. Comparisons of
data from central Christchurch sites (CCCC, CHHC,
CBGS, REHS, D09C; site descriptions in Table 1) show
that the peak accelerations in the central city were about a
factor of 2.5 times stronger in the Christchurch earthquake
than in the Darfield event. However, the peak velocities in
the Christchurch earthquake were only 1030% larger, and
the peak displacements were 40% smaller.
Figure 11A compares response spectra of recorded
motions at four sites within 1.5 km of the Christchurch
CBD and spectra from the New Zealand design standard
NZS1170 for return periods of 500 years and 2500 years
(Standards New Zealand 2004). Design spectra in NZS1170
are based on the larger horizontal component. Since the
difference between the weaker and stronger horizontal
components was greater in the Christchurch earthquake
than in most earthquakes, it is particularly important to
perform the comparison for the Christchurch earthquake in
terms of this larger component rather than the geometric
mean of the two horizontal components (used in US design
practice). Comparison to the geometric mean component
could give the misleading impression that the CBD motions
were closer to the 500-year design levels that are required for
normal-use structures in Christchurch.
The shapes of actual horizontal spectra for sites close to
the Christchurch CBD (Fig. 11) are in general deficient with
respect to the code spectral shape for very short spectral
periods up to about 0.30.4 s, but are stronger than code
shapes around 0.751.8 s and again around 2.74 s. The
displacement demands associated with the 3.5 s peak are
much greater than those associated with the short-period
peaks. Peaks are comparable to those observed during the
Darfield earthquake (Cousins & McVerry 2010), although
the long period peak is at 22.7 s in the Darfield motions. In
fact, the Darfield earthquake spectra exceeded the spectra of
the Christchurch earthquake in this period band, despite
generally being considerably weaker at other periods. The
long-period peaks in spectra in both earthquakes are
suggestive of amplification due to deep soils ( 500 m
thick) and/or basin effects below central Christchurch.

Christchurch earthquake, February 2011 77


Table 1 Peak horizontal ground motions recorded within 100 km fault distance in the Christchurch Earthquake of 22 February 2011. The
ground subsoil categories are as dened in NZS1170.5 (Standards New Zealand 2004). Ground motions are calculated from 0.1 Hz high-pass
ltered data; peak values are the strongest motions in any direction in the horizontal plane.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Epicentral
distance (km)

Fault distance
(km)1

PGA
(mm/s/s)

PGV
(mm/s)

PGD
(mm)

Site
code

6
6
5
11

1.1
2.3
2.8
3.7

4050
7238
4759
8102

500
1232
682
864

130
578
238
566

CMHS
PRPC
CCCC
NNBS

9
7

3.8
3.8

2569
3215

483
806

454
341

HPSC
D09C

3.8

4031

808

410

D08C

3.8

5968

1428

593

D10C

6
7
2
9
8
8
11
5
12
14
17
18
18
15
23
24
28
31
35
40
38
43
47
49
56
57
58
61
59
69
68
74
73
80
85
88
91
97
98
95

3.9
3.9
3.9
4.7
4.7
5.1
6.5
7.1
8.7
11
12
13
13
14
17
19
25
26
30
35
35
39
40
43
52
52
53
55
55
61
62
64
68
73
79
82
86
87
90
91

4481
2793
14 780
6297
7225
3405
2910
9799
2440
1850
1206
2225
1598
1448
2222
1909
2631
1013
868
704
837
686
325
665
354
668
733
633
484
915
908
664
897
392
693
482
252
213
142
306

825
528
982
735
969
768
340
486
477
346
126
200
173
72
224
83
150
105
93
67
71
55
37
56
30
54
43
50
38
56
67
42
81
48
43
28
28
26
17
26

265
269
230
249
313
368
195
182
236
146
91
119
91
28
116
51
53
47
37
39
24
32
25
18
23
13
17
12
20
15
13
9
10
16
13
9
10
7
8
10

CHHC
D06C
HVSC
CBGS
REHS
SHLC
RHSC
LPCC
PPHS
SMTC
TPLC
CACS
LINC
MQZ
KPOC
ROLC
SWNC
SLRC
ASHS
DSLC
CSTC
SBRC
AMBC
DFHS
RKAC
SHFC
HORC
OXZ
DORC
WAKC
SPFS
SCAC
KOWC
LSRC
ADCS
CSHS
WSFC
CECS
WIGC
LTZ

Name of recording site

Site class

Christchurch Cashmere High School


Pages Road Pumping Station
Christchurch Cathedral College
Christchurch North New Brighton
School
Hulverstone Drive Pumping Station
Christchurch Police Station
(basement)
Christchurch Police Station (5th
floor)
Christchurch Police Station (13th
floor)
Christchurch Hospital
Christchurch 20 Moorhouse Avenue
Heathcote Valley Primary School
Christchurch Botanic Gardens
Christchurch Resthaven
Shirley Library
Riccarton High School
Lyttelton Port Company
Christchurch Papanui High School
Styx Mill Transfer Station
Templeton School
Christchurch Canterbury Aero Club
Lincoln Crop and Food Research
McQueens Valley
Kaiapoi North School
Rolleston School
Swannanoa School
Selwyn Lake Road
Ashley School
Dunsandel School
Cust School
Southbridge School
Amberley HDC
Darfield High School
Rakaia School
Sheffield School
Hororata School
Oxford
Dorie
Waikari
Springfield Fire Station
Scargill
Kowai
Lake Sumner Road
Ashburton District Council
Castle Hill Station
Westerfield
Cheviot Emergency Centre
Waiau Gorge
Lake Taylor Station

D
E
D
E
E
D
na
na
D
D
C
D
D
DE
D
B
D
DE
D
D
D
B
E
D
D
D
D
D
D
D
D
D
D
C
D
B
D
C
D
B
D
C
D
B
D
C
C
B

1
Fault distances are from the upper edge of the rupture plane, which was modelled as a straight line between points ( 43.580, 172.617) and ( 43.508, 172.775)
(WGS84) in accordance with the fault plane determined from geodetic inversions.

Fault distance (km)1

Vertical (g)

Horizontal (g)

Horizontal (g)

3.9
2.3
3.8
7.1
1.1
3.7
2.8
4.7
3.9
4.7
5.1
3.8
3.8
3.8

2.20
1.89
1.07
0.50
0.85
0.75
0.80
0.52
0.60
0.36
0.49
0.36
0.46
0.84

1.68
0.67
0.22
0.91
0.35
0.81
0.48
0.72
0.35
0.55
0.32
0.27
0.39
0.48

1.27
0.60
0.29
0.96
0.40
0.59
0.38
0.37
0.36
0.45
0.35
0.30
0.36
0.45

2
6
9
5
6
11
5
8
6
9
8
7
6
6
1

Site code
HVSC
PRPC
HPSC
LPCC
CMHS
NNBS
CCCC
REHS
CHHC
CBGS
SHLC
D09C
D08C
D10C

Name of recording site

Site class

Heathcote Valley Primary School


Pages Road Pumping Station
Hulverstone Drive Pumping Station
Lyttelton Port Company
Christchurch Cashmere High School
North New Brighton School
Christchurch Cathedral College
Christchurch Resthaven
Christchurch Hospital
Christchurch Botanic Gardens
Shirley Library
Christchurch Police Station (basement)
Christchurch Police Station (5th floor)
Christchurch Police Station (13th floor)

C
E
E
B
D
E
D
D
D
D
DE
D
na
na

Fault distances are from the upper edge of the rupture plane, which was modelled as a straight line between points (43.580, 172.617) and (43.508, 172.775) (WGS84) in accordance with the fault
plane determined from geodetic inversions.

Figure 10 Comparison of peak horizontal ground motions from


the Mw 6.2 Christchurch and Mw 7.1 Dareld earthquakes.
Ground motions are calculated from 0.1 Hz high-pass ltered
data. A, Peak ground acceleration. B, Peak ground velocity. C,
Peak ground displacement.

Figure 11B compares spectra of the recorded motions at


the four sites near the CBD with those estimated for an
oblique-mechanism magnitude 6.2 earthquake located at a
distance of 4 km from the fault according to the McVerry et
al. (2006) model often used in New Zealand. The model
spectra are shown at the 50- and 84-percentile level (median,
and one standard deviation above the median). The
50-percentile spectrum corresponds approximately to the

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Epicentral
distance (km)

78 A Kaiser et al.

Table 2 Strongest accelerations recorded at GeoNet stations during the 22 February earthquake (vertical and two orthogonal horizontal components).

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 79

Figure 11 Central city acceleration response spectra from the Christchurch earthquake. A, Recorded 5% damped acceleration response
spectra at four central city sites (CHHC, CCCC, CBGS, REHS; locations shown in Figure 3; geometric mean of the four sites shown as thick
red line) compared to the elastic spectra for 500 and 2500 yr (dashed and solid black lines, respectively) from the New Zealand design
standard NZS1170 (Standards New Zealand 2004). B, Recorded spectra are compared with the 50- and 84-percentile estimated spectra for a
magnitude 6.2 oblique mechanism earthquake at 4 km distance (solid and dashed black lines). Models follow McVerry et al. (2006).

NZS1170 500-year spectrum, and is considerably weaker


than the spectra of the recorded motions.
In summary, ground motions recorded in central
Christchurch generally considerably exceeded even 2500year design motions, let alone the 500-year or 1000-year
design spectra used for most buildings in the CBD. The
ground motions were also considerably higher than those
estimated based on the magnitude and distance of the
earthquake at near-source distances, although they were

somewhat less at larger distances ( 10 km). This is further


illustrated in Fig. 12, where comparisons of recorded
spectral acceleration with increasing source distance at 1 s
period are made with estimates for an oblique crustal event
from the national attenuation model (McVerry et al. 2006).
Also of note for the Christchurch earthquake was that
vertical accelerations were greater than horizontal accelerations at near-source distances and significantly richer in
high-frequency (short-period) energy (Fig. 3; Table 2).

Downloaded by [University of Tokyo] at 19:20 17 November 2013

80 A Kaiser et al.

Figure 12 Recorded spectral accelerations (at 1.0 s period) during the Christchurch earthquake compared to the NZ median for a magnitude
6.2 crustal oblique mechanism earthquake from the national attenuation model for Site Class D (McVerry et al. 2006).

Factors contributing to the high accelerations


A number of factors contributed to the high near-source
accelerations experienced in Christchurch during the 22
February event. The bulk of the strong accelerations can be
explained by the proximity of the fault plane and a few other
dominant factors. These include: (1) the high apparent stress
of the event; (2) directivity of the rupture including
the effects of high rupture velocity; and (3) site effects
including decoupling of near-surface materials under high
accelerations.
The Christchurch earthquake had an energy magnitude
(Me 6.7) larger than its moment magnitude, as did
the Darfield main shock (Fry et al. 2011a; Me provided by
the USGS http://earthquakes.usgs.gov). This indicates the
source fault may be strong, with a high level of friction
holding the two sides together. By using teleseismic determinations of moment, we have calculated apparent stress
(Ta) assuming bulk regional properties for density and
shear-wave velocity; more details can be found in Fry &
Gerstenberger (2011). Typically, high Ta is indicative of
spatially small yet strong asperities rupturing with high
velocity and yielding strong high-frequency waves in the
near field. Ta for the Christchurch earthquake was 4.1 MPa
(compared to 15.9 MPa for the Darfield Earthquake). These
values are remarkably high compared to global averages.
Shallow events worldwide have a spread from about 6 MPa
to 0.04 MPa, with an average of c. 0.5 MPa (Choy et al.
2001). Crustal thrusting events from the Choy et al. (2001)
dataset typically have lower Ta than strike-slip events. Some
of the largest Ta observations come from shallow oceanic
earthquakes on transform faults that are oblique or normal
to nearby plate margins. These events have average Ta of
around 3 MPa. Multiple workers have also analysed stress
drops for shallow crustal earthquakes in Japan with
estimates of average Ta around 1.1 MPa (Oth et al., 2010;
Baltay et al., 2011). By re-examining previously published

estimates of radiated energy and correcting the measurements for errors produced by limited band-width sampling,
Ide & Beroza (2001) find average Ta for continental crustal
earthquakes is around 1 MPa with a range between about 0.
1 and 10 MPa. In all cases, the Canterbury earthquakes
remain among the highest-stress events.
Geological and tectonic factors leading to the high
apparent stress in the Canterbury events have been identified
by Reyners (2011) and Fry et al. (2011a). The Christchurch
earthquake may have occurred on a reactivated Cretaceousage fault in the upper crust similar to those found offshore
(Wood & Herzer 1993). These faults exist in greywacke and
schists in the upper crust that are inferred to be tightly
welded to underlying stiff oceanic plateau rocks associated
with a former subduction margin (Reyners 2011). Alternatively, the earthquake may have involved reactivation of a
fault formed during the 126 Ma emplacement of the Banks
Peninsula volcanic rocks to the south of the city. Reyners
(2011) suggests that faults associated with both the anomalous crustal structure described above and in Banks
Peninsula volcanic rocks are likely to be inherently stronger
than those in average crust, and that rupturing them would
subsequently release more radiated energy. Furthermore,
faults in lower seismicity regions or immature faults are
often associated with higher friction along the fault
plane and therefore higher-stress earthquakes (Kanamori
& Allen 1986).
Kinematic and geodetic modelling shows that the maximum displacement on the fault occurred at shallow (34
km) depth, with most of the energy directed northwestwards towards the city. Almost all seismic stations at nearsource distances were located on the footwall side of the
fault, suggesting that the high accelerations shown in Fig. 12
could be associated with directivity effects. Numerical wave
simulations using a combined discrete wavenumber and
boundary integral method (Bouchon 1979; Bouchon et al.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 81


1996) were carried out to model potential source effects in
Christchurch. We assume a 9 8 km (length width)
rectangular thrust fault dipping 658 SE and striking 708,
rupturing with 3.2 km/s rupture velocity and 1208 rake (in
accordance with preliminary source modelssee above).
Ground motion (displacement) is computed across a
5050 km square area. The fault is embedded in a halfspace medium underlying a 1 km thick layer (Fig. 13A) and
slip is equal to 1.5 m at the initial rupture point. Computed
ground motions show vertical displacements in the range
0.851.2 m and corresponding accelerations of 0.52 g on
the footwall side, close (within 2 km) to the projection of the
faults upper end on the free surface. We believe that the
high accelerations can be largely explained by the combined
effect of directivity and the sudden change of polarity of the
motion as the rupture reaches the upper end of the fault,
close to the free surface. Fig. 13B shows a series of snapshots
of the radiated wave field at the surface as the rupture
propagates upwards along the fault. The rupture energy
arrives at the free surface at c. 4 s. The plot at 5 s shows the
instant when the vertical radiation in the vicinity of the fault

trace on both the hanging wall and the footwall sides are of
about the same intensity. At 6 s, the polarity of motion flips
over. At this point the rupture has ended and the radiated
energy is high with the remaining plots showing further
wave propagation in the layered medium. Ground motions
from computed synthetic seismograms (e.g. Fig. 13C)
reproduce the main features of the observed data.
The computed seismograms do not however explain all
features of the recorded ground motions, i.e. the strong
accelerations and higher frequency content in the vertical
component observed at many near-field stations (see accelerograms in Fig. 3). A strong frequency dichotomy was
observed where near-source vertical accelerations were rich
in high-frequency (short-period) energy with maximum
energy at 1015 Hz near the epicentre, in marked contrast
to the dominant lower frequencies (longer periods) generally
observed on the horizontal components. This effect is linked
to the fact that the water table under these sites is quite high,
reaching up to the near surface. Near-surface reflections off
the water table likely amplify energy on the vertical
component by trapping vertically resonating energy in the

Figure 13 Numerical simulations of ground motion. A, Schematic of input fault model and velocity structure; fault strikes at 708. B,
Snapshots from preliminary numerical model of the Mw 6.2 Christchurch earthquake up to 5 Hz on three components (x, y, z are positive
north, east and upwards; modelled area is 50 50 km). C, Computed seismograms for a footwall station shown in A, i.e. at 5 km from the
rupture initiation (or about 1 km from the fault trace projection on the free-surface). It shows the three components of displacement (top)
and acceleration (bottom).

Downloaded by [University of Tokyo] at 19:20 17 November 2013

82 A Kaiser et al.
uppermost metres of the soil column, whereas high-frequency horizontal energy is attenuated in the shallow
subsurface within unconsolidated saturated (and sometimes
liquefied) material (Yang and Sato 2000; Fry et al. 2011b).
Fry et al. (2011b) effectively reproduce this effect by
introducing a shear-wave attenuation factor in the upper
few metres of the stratigraphic column into numerical
simulations of ground motion.
In addition, most of the high-acceleration vertical
records recorded in Christchurch exhibit sharp positive
spikes and broader negative troughs. These features result
in asymmetric records where maximum accelerations in the
upward direction exceed accelerations in the downward
direction (e.g. vertical Z component at HVSC and Pages
Road Pumping Station (PRPC) in Fig. 3; Fry et al. 2011b).
Similar characteristics are found in other global recordings
of strong, shallow earthquakes, most notably the 2008 Mw
6.9 Iwate-Miyagi Nairiku earthquake in Japan (Aoi et al.
2008; Yamada et al. 2009). The asymmetry has been
attributed to a trampoline or slapdown effect (Aoi et al.
2008; Yamada et al. 2009) involving decoupling of nearsurface materials, analogous to spalling that was first
observed in underground nuclear explosions (Eisler &
Chilton 1964). In the trampoline model of Aoi
et al. (2008), the asymmetry in ground acceleration arises
from decoupling of near-surface materials during highamplitude downward acceleration. This occurs when the
tensile forces, that arise on an interface or within a granular
material from downgoing particle oscillation as waves pass,

are larger than its tensile strength. The result is an


approximate free-fall of the material at near-gravitational
acceleration. In this model, the high upwards accelerations
are caused by the compressional response of the granular or
layered media to the stress of the upgoing wave oscillation.
Yamada et al. (2009) suggest the large positive accelerations
are further enhanced by slapdown as downgoing upper soil
layers collide with underlying layers that are returning
upwards during the following earthquake wave cycles.
When the upgoing and downgoing layers collide, highfrequency energy is generated.

Effects on the built environment


Liquefaction
In all three major Canterbury earthquakes (Mw 7.1 Darfield,
Mw 6.2 Christchurch, Mw 6.0 June 13), widespread liquefaction occurred in urban areas of Christchurch (and the
town of Kaiapoi 17 km north of Christchurch CBD in the
first two events) causing extensive damage to residential
properties, water and wastewater networks, high-rise buildings and bridges. The liquefaction was manifested by
massive sand boils and large amounts of sand/silt ejecta
and water littering streets, residential properties and recreation grounds. Nearly 15 000 residential houses and properties were severely damaged due to liquefaction and lateral
spreading, more than half of those beyond economic repair.
The most severe liquefaction in Christchurch occurred

Figure 14 Areas of observed liquefaction in Christchurch due to the 22 February Mw 6.2 earthquake (coloured areas) and the Mw 7.1
Dareld main shock (white contours) based on drive-through reconnaissance and surface manifestation of liquefaction visible on aerial
photographs (Cubrinovski & Taylor, 2011).

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 83


during the Mw 6.2 Christchurch earthquake; however,
damage in Kaiapoi was greatest in the Darfield main shock.
The extent of liquefaction during the 22 February
earthquake in Christchurch is outlined in Fig. 14 (Cubrinovski & Taylor 2011), with comparison to the Darfield
main shock. The distribution of liquefied areas reflects the
combined effects of two important factors for liquefaction
triggering: the soil resistance to liquefaction (a measure of
the capacity of soils to sustain cyclic loading) and the
severity of ground motions (measure for the seismic load or
demand) produced by the two quakes. The suburbs most
severely affected by liquefaction in Christchurch were along
the Avon River to the east and northeast of the CBD
(Avonside, Dallington, Avondale, Burwood and Bexley).
The soils in these suburbs are predominantly loose fluvial
deposits of liquefiable clean fine sands and sands with nonplastic silts. The top 56 m are in a very loose state, with a
CPT cone resistance qc of c. 24 MPa. The resistance
typically increases to 712 MPa at depths between 6 and 10
m; however, lower resistances are often encountered in areas
close to wetlands (Gerstenberger et al. 2011). The more
extensive liquefaction observed in these areas during the 22
February earthquake is consistent with the fact that the
seismic demand specific to liquefaction (i.e. the cyclic stress
ratio or CSR, combining the effects of amplitude and
duration of strong ground shaking) was about 1.52.0 times
higher during the February event as compared to the
September 2010 earthquake. Conversely, at the southwest
end of the city in Hoon Hay and Halswell, more extensive
liquefaction occurred during the 2010 Darfield earthquake
when seismic demand (CSR value) was comparatively
larger.
Liquefaction led to large settlements of many houses
including differential settlements that caused foundation
and structural damage. The large ground movements and
cracks and fissures in the ground also caused extensive
damage to buried lifelines, roads and bridges. The waste-

water and potable water pipe networks and electrical


networks were hit particularly hard. Up to c. 40% of the
Christchurch wastewater network still had limited or no
service nearly one month after the earthquake (M. Christison, personal communication 2011).
The largest damage to land and the built environment
was caused by liquefaction-induced lateral spreading along
the Avon River and close to wetlands; lateral spreading over
a 1 km long area also caused severe damage along the
Kaiapoi River and Courtenay Stream during the Darfield
main shock (Cubrinovski et al. 2010). Lateral spreading
displacements reached anywhere from a few tens of centimetres to 23 m at the river banks and extended inland as
far as 200300 m from the waterway, severely damaging
infrastructure and residential properties within the affected
zone (Fig. 15).
The severe ground motions produced by the 22 February
2011 earthquake caused liquefaction within the CBD which
was particularly extensive along the Avon River. A relatively
large number of high-rise buildings were affected by the
liquefaction. Christchurch soils within the CBD are highly
variable and include surface layers of liquefiable sands and
silts, soft peaty soils and underlying layers of medium-dense
sandy gravels, medium-dense and dense sands. High-rise
buildings therefore often rest on complex foundation
systems. Such difficult foundation conditions when combined with the extremely severe ground motions and induced
soil liquefaction led to large differential movements and
settlement of buildings. For example, Fig. 16 shows large
differential settlements of about 0.4 m between the foundations at the end columns of a six-storey building on shallow
foundations. In preliminary inspections of approximately 20
high-rise buildings within the CBD, a tilt of 0.30.6 degrees
was typically observed for buildings resting either on rafts,
shallow foundations or deep foundations with different
lengths of piles.

Figure 15 Lateral spreading along the Avon River (photo taken post 22 February earthquake).

Downloaded by [University of Tokyo] at 19:20 17 November 2013

84 A Kaiser et al.

Figure 16 Differential settlement of a multi-storey building on shallow foundations in liqueed area of CBD (photo taken post 22 February
earthquake).

Landslides and topographic effects in the southern Port Hills


The proximity and high accelerations of the Mw 6.2
Christchurch earthquake triggered land movement, the
collapse of cliffs and rockfalls in the southern Christchurch
Port Hills suburbs that lie up to 500 m high on the northern
flanks of the extinct volcano forming Banks Peninsula
(Hancox et al. 2011; Dellow et al. 2011). Coastal erosion
and the quarrying of rock have produced steep cliffs at the
base of the hills. At least five people were killed by falling
rocks in this area (three in the Sumner-Redcliffs area and
two walking on Port Hills foot tracks). Several hundred
homes were evacuated because they were close to the foot or
top of dangerous cliffs or on cracked and unstable steep
slopes. In addition to slope instabilities, severe shaking
damage to houses was observed on ridge tops in Mt
Pleasant, Redcliffs and Scarborough, indicating that topographic effects amplified ground motions. By comparison,
slope failure and damage observed after the more distant
Darfield main shock was much more limited.
Following the Christchurch earthquake, the GeoNet
landslide response team and the Port Hills Geotechnical
Group carried out aerial reconnaissance, assessed ground
damage and set up mass-movement monitoring stations.
Four main types of earthquake-triggered mass movements
were identified and mapped: rockfalls, shallow landslides,
deep-seated landslides and tension cracks/rents (Fig. 17).
Rockfalls made up the largest number of mass movements
and were responsible for the five fatalities as well as
substantial damage to properties, roads and other infrastructure. During the earthquake, sections of cliffs, rocky
outcrops and slopes shed much rock debris (Fig. 18A, 18B).
The rockfalls, some of which bounced and rolled long
distances to smash through houses, ranged from single
boulders to large masses of rocks. Localised shallow landsliding and failure of retaining walls, fill slopes and homes

(e.g. Fig. 18D) was found over a large area and was caused
by the strong earthquake ground shaking. However, this
type of damage did not necessarily pose ongoing slope
instability issues in the area. By contrast, many slopes
showed deep tension cracks and rents that indicated sections
of slope with potential for further collapse. Deep-seated
ground movement of large areas was indicated by clusters of
large deformation features including cracks and bulges (e.g.
Fig. 18C). Deep-seated landslides, together with tension
cracks, caused the most damage to the ground and therefore
to houses and infrastructure on the hills. These features
tended to be on or very close to cliff tops and convex breaks
in slope. During the earthquake, the topographic position
and the morphology and geology of the slopes gave rise to
variations in ground shaking, leading to localised areas of
very heavy damage often associated with breaks in slope.
Ground damage decreased rapidly away from these breaks
in slope.

Performance of Christchurch buildings


The building response and degree of damage incurred from
the Christchurch earthquake was strongly influenced by the
building characteristics, in addition to the level of shaking
experienced and the extent of liquefaction at the site. The
building stock in Christchurch consists of unreinforced
masonry buildings, timber buildings, reinforced concrete
frame buildings and tilt-up industrial buildings. We present
preliminary observations of damage to these four typical
building types.
Overall, the nature and severity of building damage
observed across Christchurch was consistent with the
damage generally expected for ground-shaking and liquefaction intensities recorded during the 22 February earthquake.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 85

Figure 17 Map showing the distribution of mass movements, including rockfall density. Data were collated from eld information collected
by the GeoNet landslide response team and Port Hills Geotechnical Group on behalf of Christchurch City Council as well as aerial
photographs taken after the earthquake by New Zealand Aerial Mapping. Coordinates are in New Zealand Map Grid (m).

Damage to masonry buildings including churches (e.g.


Fig. 19) was widespread across the city. These buildings are
usually rigid and characterised by short periods of vibration.
Particularly severe damage occurred to unreinforced masonry buildings that had been classified as earthquake-prone
prior to the Canterbury earthquakes. Such buildings had
inadequate strength and displacement capacity to match the
earthquake demand. Residential and commercial buildings
with unreinforced brick masonry walls also typically suf-

fered partial collapse. This failure mode could be attributed


to the lack of adequate connection between the wall and the
diaphragm. A small number of fatalities resulted from
failure of masonry buildings.
The bulk of Christchurch residential properties are
timber houses. Damage to timber houses (e.g. Fig. 20) can
be broadly correlated with either shaking intensity or
liquefaction and lateral spreading. Some houses near
the epicentral region suffered shaking damage to walls and

Figure 18 Examples of mass movement damage. A, Landslide at Redcliffs School. B, Rockfall damage to residential house. C, Tension
cracks (marked by arrows) indicating an incipient deep-seated landslide. D, Localised failure of retaining wall and ll in Mt Pleasant.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

86 A Kaiser et al.

Fig. 19 Out-of-plane failure of unreinforced masonry (brick/stone) walls.

roof tiles. Lateral spreading was responsible for particularly


severe damage to some residential houses further away from
the epicentral region.
The majority of the building stock in the CBD was
reinforced concrete buildings, with variation in vintage and
number of stories. A large number of casualties in the
Christchurch earthquake resulted from the collapse of two
reinforced concrete office buildings. Older (pre-1976) reinforced buildings typically demonstrated brittle failure due
to lack of ductile response (e.g. the car park building in Fig.
21A). Damage to critical structural elements was also
typically observed in buildings built during 19761992 that
lacked modern ductile detailing (e.g. Fig. 21B). Modern
reinforced concrete buildings (post-1992) generally performed comparatively better in terms of life-safety, typically
sustaining moderate damage. However, staircase damage
was severe in some modern buildings and a few staircases
collapsed. Many buildings also suffered damage due to
pounding impacts from adjacent buildings where there was
insufficient seismic gap between buildings (e.g. Fig. 21C).
Apart from structural damage, damage to contents and to
architectural and service components of buildings was
extensive. The failure of facades, suspended ceiling and

other service components resulted in life-safety issues for the


occupants and significantly contributed to economic loss
from business interruption.
Industrial buildings with tilt-up panel construction
generally performed reasonably well, although in some
cases structural connections between pre-cast panel elements exhibited some damage. In addition, pallet-rack
storage systems within industrial buildings were damaged
often to the point of collapse in buildings located near the
epicentre.

The Christchurch earthquake in the context of seismic hazard


models
The Mw 6.2 Christchurch earthquake was to some extent
accounted for in the distributed seismicity component of
pre-existing probabilistic seismic hazard (PSH) models
(Stirling et al. 2007, 2008). Distributed seismicity models
use the spatial distribution of historical earthquakes to
forecast the size, rate and recurrence of moderate-to-large
earthquakes away from the known faults, and therefore
allow for the possibility of future earthquakes on unknown
sources. The Poissonian (time-independent) national seismic

Figure 20 Residential house damage due to A, severe shaking and B, lateral spreading.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 87

Figure 21 Damage to central Christchurch reinforced concrete buildings. A, Collapse of a at slab car park building. B, Cracking, spalling
and damage to reinforcement in beam. C, Pounding damage to the building on the left.

hazard model (e.g. Stirling et al. 2007) and the regionallybased Canterbury seismic hazard model (Stirling et al. 2008)
identified three classes of earthquakes as dominating the
hazard of the city: moderate-sized (about magnitude 56.5)
earthquakes at close distances to the city; large regional
earthquakes (about magnitude 77.5) on faults beneath the
Canterbury Plains and foothills of the Southern Alps; and
great earthquakes (about magnitude 8) on the distant Alpine
Fault. The Christchurch earthquake was clearly a close-by
moderate-sized earthquake, while the Mw 7.1 Darfield
earthquake was in the category of a large regional earthquake. The notable aspect of the Christchurch earthquake
was the very strong shaking in the near field relative to that
predicted for the size of the earthquake (e.g. Fig. 11, Fig. 12).
Comparison of response spectra for the earthquake to
uniform hazard spectra generated from the national seismic
hazard model show the recorded Christchurch spectra to be
equivalent to the 10000 year return period spectra from the
PSH model. In other words, the ground motions were of

very low probability according to the national seismic


hazard model. Ongoing efforts are focused on better
quantifying the factors that contributed to the high ground
motions in order to assess the possible implications for
future earthquakes in the Canterbury region, and by
extension other comparable areas in New Zealand and
worldwide.
Time-dependent earthquake forecast models such as
Short Term Earthquake Probability (STEP) model (Gerstenberger et al. 2005) were implemented after the occurrence of the Darfield earthquake, and forecast an
earthquake of the size of the 22 February earthquake with
a high probability. In the week prior to the 22 February
earthquake, STEP estimated an approximate 25% chance of
a magnitude 6 or greater earthquake occurring in the general
aftershock zone of the Darfield earthquake in the next year.
Similar time-dependent models are now providing hazard
input to the planning and design associated with the rebuild
of Christchurch (e.g. Gerstenberger et al. 2011).

Downloaded by [University of Tokyo] at 19:20 17 November 2013

88 A Kaiser et al.
Conclusions
The 22 February 2010 Christchurch earthquake was
the most serious to strike New Zealand since the 1931 Mw
7.8 Hawkes Bay earthquake, causing severe ground shaking
and widespread destruction in the city of Christchurch.
Regional moment tensor solutions show a moment
magnitude of 6.2 with a combination of right-lateral
strike-slip and reverse faulting. The trend of aftershocks is
consistent with the northeastsouthwest-oriented fault
plane. Preliminary geodetic data inversions derived from
GPS and InSAR data constrain the fault geometry, revealing a fault plane dipping at 698 to the southeast and striking
northeastsouthwest at 588. The geodetic model suggests
slip on the southwestern part of the fault was more
dominantly strike-slip. Strong-motion source models show
a shallow rupture with average rake of 1338 (i.e. right-lateral
displacement slightly greater than reverse displacement),
consistent with the geodetic model. All preliminary source
models suggest that maximum slip occurred at 34 km depth
with an amplitude of 2.4 m inferred from the geodetic model
and 3.6 m from the strong-motion model.
Earthquakes of this type were to some extent accounted
for in the national seismic hazard model through the
distributed seismicity component. However, the severe
ground motions experienced in Christchurch were of low
probability according to the model.
Recorded ground motions ranged up to 2.2 g (vertical)
acceleration and 1.7 g (horizontal) near the epicentre and up
to 0.8 g (vertical) and 0.7 g (horizontal) in Christchurch
CBD. Comparisons of recorded spectral accelerations with
predictions for equivalent earthquakes from the New
Zealand national attenuation model show that horizontal
ground motions were larger than expected within c. 5 km of
the fault. Furthermore, response spectra in the CBD
exceeded 2500 yr design standards and building codes
currently used in the city. This was also evident for long
period (up to 4 s) motions amplified by the deep sediments
below the central city. Vertical accelerations were higher
than horizontal accelerations throughout much of the eastern and southern city and vertical spectra were significantly
richer in high frequencies. Other factors contributing to the
high accelerations include directivity of the rupture towards
Christchurch, site effects including a trampoline effect from
decoupling of near-surface materials and a high-stress
source indicative of a strong source fault.
Liquefaction in Christchurch was even more widespread
in the Christchurch earthquake than in the Darfield main
shock with suburbs to the east and northeast of the CBD
severely affected. It resulted in foundation and structural
damage to buildings. Large ground displacements caused
further damage to lifelines and infrastructure. Rockfall and
landslides impacted on the southern hillside suburbs close to
the epicentre causing fatalities, damage to residential dwellings and evacuation of properties. Building damage also

correlated with shaking intensity and building type, with


unreinforced masonry buildings particularly affected.
The extensive seismic and observational data collected
during and after the 22 February 2011 earthquake will
provide valuable information to expand our knowledge of
seismic hazard. The data will also have implications for the
rebuilding of Christchurch and future hazard mitigation
elsewhere in New Zealand and worldwide.
Acknowledgements
The very important set of strong-motion records used in this
research was made possible by the GeoNet project sponsored by
EQC, GNS Science and LINZ. The existence of the regional CanNet
component of the network came about because of the vision and
persistence of Dr John Berrill in promoting the concept. Additionally, we thank the global Quake-Catcher Network and our volunteer
seismometer hosts in Christchurch for additional low-cost accelerometer data used in Fig. 3. We thank GeoNet, Trimble Navigation
NZ Ltd, Geosystems NZ Ltd and Global Survey Ltd for providing
continuous GPS data, and Kirby MacLeod and Linda Alblas for
their assistance with the post-earthquake GPS surveys. CSK
original data is copyright 2011 Italian Space Agency; part was
provided by e-GEOS, an ASI/Telespazio company, and part was
provided under CSK AO PI project 2271. Part of the geodetic
research was performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National
Aeronautics and Space Administration. The GeoNet landslide team
and Port Hills Geotechnical Group carried out the comprehensive
field surveys of mass movements that have been used in this report.
Thanks to Russ Van Dissen and Pilar Villamor for comments on this
manuscript. We also thank the Editor, Euan Smith and an
anonymous reviewer for their constructive reviews.

References
Abercrombie RE, Webb TH, Robinson R, McGinty PJ, More J,
Beavan RJ 2000. The enigma of the Arthurs Pass, New
Zealand, earthquake 1. Reconciling a variety of data for an
unusual earthquake sequence. Journal of Geophysical
Research 105(B7): 1611916127.
Amoruso A, Crescentini L, Fidani C 2004. Effects of crustal
layering on source parameter inversion from coseismic geodetic data, Geophysical Journal International 159: 353364,
doi: 10.1111/j.1365-246X.2004.02389.x
Aoi S, Kunugi T, Fujiwara H 2008. Trampoline effect in extreme
ground motion. Science 322: 727730.
Arnadottir T, Segall P 1994. The 1989 Loma Prieta earthquake
imaged from inversion of geodetic data. Journal of Geophysical Research 99: 2183521855.
Avery HR, Berrill JB, Coursey PF, Deam BL, Dewe MB, Francois
CC, Pettinga JR, Yetton MD 2004. The Canterbury University strong-motion recording project. Proceedings of the
13th World Conference on Earthquake Engineering, Vancouver, British Columbia, 1-6 August, Canadian Association for
Earthquake Engineering paper no. 1335.
Baltay A, Ide S, Prieto G, Beroza G 2011. Variability in earthquake
stress drop and apparent stress. Geophysical Research Letters
38(6): L06303.
Barnes PM 1996. Active folding of Pleistocene unconformities on the
edge of the Australian-Pacic plate boundary zone, offshore
North Canterbury, New Zealand. Tectonics 15: 623640.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

Christchurch earthquake, February 2011 89


Barrell DJA, Litcheld NJ, Townsend DB, Quigley M, Van Dissen
RJ, Cosgrove R, Cox SC, Furlong K, Villamor P, Begg JG,
Hemmings-Sykes S, Jongens R, Mackenzie H, Noble D, Stahl
T, Bilderback E, Duffy B, Henham H, Klahn A, Lang EMW,
Moody L, Nicol R, Pedley K, Smith A 2011. Photographic
FeatureStrike-slip ground-surface rupture (Greendale Fault)
associated with the 4 September 2010 Dareld earthquake,
Canterbury, New Zealand. Quarterly Journal of Engineering
Geology and Hydrogeology 44: 283291.
Beavan RJ, Samsonov S, Motagh M, Wallace LM, Ellis SM,
Palmer NG 2010a. The Dareld (Canterbury) earthquake:
geodetic observations and preliminary source model. Bulletin
of the New Zealand Society for Earthquake Engineering 43(4):
228235.
Beavan RJ, Samsonov S, Denys P, Sutherland R, Palmer NG,
Denham M 2010b. Oblique slip on the Puysegur subduction
interface in the 2009 July Mw 7.8 Dusky Sound earthquake
from GPS and InSAR observations: implications for the
tectonics of southwestern New Zealand. Geophysical Journal
International 183(3): 12651286; doi: 10.1111/j.1365-246X.
2010.04798.x.
Beavan J, Fielding E, Motagh M, Samsonov S, Donnelly N 2011.
Fault location and slip distribution of 22 February 2011 Mw
6.3 Christchurch, New Zealand, earthquake from geodetic
data. Seismological Research Letters 82: 789799.
Bibby HM 1982. Unbiased estimate of strain from triangulation
data using the method of simultaneous reduction. Tectonophysics 82(12): 161174.
Bouchon M 1979. Discrete wave number representation of elastic
wave eld in three-space dimensions. Journal of Geophysical
Research 87(B7): 36093614.
Bouchon M, Schultz CA, Toksoz MN 1996. Effect of threedimensional topography on seismic motion. Journal of
Geophysical Research 101: 58355846.
Chen CW, Zebker HA 2002. Phase unwrapping for large SAR
interferograms: statistical segmentation and generalized network models. IEEE Transactions on Geoscience and Remote
Sensing 40: 17091719.
Choy GL, Boatwright JL, Kirby S 2001. The radiated seismic
energy and apparent stress of interplate and intraplate earthquakes at subduction zone environments: Implications for
seismic hazard estimation, U.S. Geol. Survey Open-File
Report 01005: 10 p.
Cochran ES, Lawrence JF, Kaiser A, Fry B, Chung A, Christensen
C 2011. Comparison between low-cost MEMs and traditional
accelerometers: A case study from the M7.1 Dareld Earthquake aftershock deployment. Annals of Geophysics: in press.
Cousins WJ, McVerry GH 2010. Overview of strong-motion data
from the Dareld Earthquake. Bulletin of the New Zealand
Society for Earthquake Engineering 43(4): 222227.
Cowan HA 1991. The North Canterbury earthquake of September
1, 1888. Journal of the Royal Society of New Zealand 21: 1
12.
Crook CN 1992. ADJCOORD: A Fortran program for survey
adjustment and deformation modelling. NZGS EDS Rep. 138,
DSIR Geology and Geophysics, Lower Hutt, New Zealand.
Cubrinovski M, Taylor M 2011. Liquefaction map of 22 February
earthquake from drive-through reconnaissance. Canterbury,
University of Canterbury.
Cubrinovski M, Green R, Allen J, Ashford S, Bowman E, Bradley
B, Cox B, Hutchinson T, Kavazanjian E, Orense R, Pender M,
Quigley M, Wotherspoon L 2010. Geotechnical reconnaissance of the 2010 Dareld (Canterbury) Earthquake. Bulletin
of New Zealand Society for Earthquake Engineering 43(4):
243320.

Darby DJ, Beavan RJ 2001. Evidence from GPS measurements for


contemporary interplate coupling on the southern Hikurangi
subduction thrust and for partitioning of strain in the upper
plate. Journal of Geophysical Research 106(B12): 30881
30891.
Dellow G, Yetton M, Massey C, Archibald G, Barrell DJA, Bell D,
Bruce Z, Campbell A, Davies T, De Pascale G, Easton M,
Forsyth PJ, Gibbons C, Glassey P, Grant H, Green R,
Hancox G, Jongens R, Kingsbury P, Kupec J, Macfarlane
D, McDowell B, McKelvey B, McCahon I, McPherson I,
Molloy J, Muirson J, OHalloran M, Perrin N, Price C, Read
SAL, Traylen N, Van Dissen R, Villeneuve M, Walsh I 2011.
Landslides caused by the 22 February 2011 Christchurch
earthquake and management of landslide risk in the immediate aftermath. Bulletin of the New Zealand Society for
Earthquake Engineering 44(4): in press.
DeMets C, Gordon RG, Argus DF 2010. Geologically current
plate motions. Geophysical Journal International 181: 180.
Dorn C, Green AG, Jongens R, Carpentier S, Kaiser AE, Campbell F, Horstmeyer H, Campbell J, Finnemore M, Pettinga J
2010. High-resolution seismic images of potentially seismogenic structures beneath the northwest Canterbury Plains,
New Zealand. Journal of Geophysical Research 115: B11303,
doi:10.1029/2010JB007459.
Doser DI, Webb TH, Maunder DE 1999. Source parameters of
large historical (19181962) earthquakes, South Island, New
Zealand. Geophysical Journal International 139: 769794.
Dowrick DJ 1998. Damage and intensities in the magnitude 7.8
1931 Hawkes Bay, New Zealand, earthquake. Bulletin of the
New Zealand Society for Earthquake Engineering 31: 139
163.
Dreger DS 2003. TDMT_INV: Time domain seismic moment
tensor INVersion. In International Handbook of Earthquake
and Engineering Seismology, Lee WK ed. Boston, Academic
Press: 81B, 1627.
Dreger D, Helmberger DV 1993. Determination of source parameters at regional distances with single station or sparse
network data. Journal of Geophysical Research 98: 8107
8125.
Eisler JD, Chilton F 1964. Spalling of the earths surface by
underground nuclear explosions. Journal of Geophysical
Research 69(24): 52855293.
Estrada B 2003. Seismic hazard associated with the Springbank
Fault, North Canterbury Plains. MSc thesis, University of
Canterbury, Christchurch, New Zealand, 193 p.
Finnemore M 2004. The application of seismic reection surveying
to the characterisation of aquifer geometry and related
tectonic deformation, North Canterbury. PhD thesis, University of Canterbury, Christchurch, New Zealand: 202 p.
Forsyth PJ, Barrell DJA, Jongens R (compilers) 2008. Geology of
the Christchurch area, scale 1:250000. Institute of Geological
& Nuclear Sciences Limited, 1:250000 Geological Map 16, 1
sheet  67 p, Lower Hutt, New Zealand. GNS Science ISBN
978-0-478-19649-8.
Fry B, Gerstenberger M 2011. Large apparent stresses from the
Canterbury earthquakes of 2010 and 2011. Seismological
Research Letters 82: 833838.
Fry B, Benites R, Reyners M, Holden C, Kaiser A, Bannister S,
Gerstenberger M, Williams C, Ristau J, Beavan J 2011a.
Extremely strong shaking in the New Zealand earthquakes of
2010 and 2011. Eos Transactions. American Geophysical
Union 92(41): 349360.
Fry B, Benites R, Kaiser A 2011b. The character of accelerations in
the Christchurch Mw 6.3 earthquake. Seismological Research
Letters 82: 846852.

Downloaded by [University of Tokyo] at 19:20 17 November 2013

90 A Kaiser et al.
Gerstenberger M, Wiemer S, Jones LM, Reasenberg P 2005. Realtime forecasts of tomorrows earthquakes in California.
Nature 435(7040): 328331.
Gerstenberger M, Cubrinovski M, McVerry G, Stirling M,
Rhoades D, Bradley B, Langridge R, Webb T, Peng B,
Pettinga J, Berryman K, Brackley H 2011. Probabilistic
assessment of liquefaction potential for Christchurch in the
next 50 years. GNS Science Report 2011/15: 30 p.
Gledhill K, Robinson R, Abercrombie R, Webb T, Beavan J,
Cousins J, Eberhart-Phillips D 2000. The Mw 6.2 Cass, New
Zealand earthquake of 24 November, 1995: Reverse faulting
in a strike-slip region. New Zealand Journal of Geology and
Geophysics 43: 255269.
Gledhill K, Ristau J, Reyners M, Fry B, Holden C 2011. The
Dareld (Canterbury, New Zealand) Mw 7.1 Earthquake of
September 2010: A Preliminary Seismological Report. Seismological Research Letters 82(3): 378386.
Goldstein RM, Werner CL 1998. Radar interferogram ltering for
geophysical applications. Geophysical Research Letters 25:
40354038.
Hancox G, Massey C, Perrin N 2011. Landslides and related
ground damage caused by the Mw 6.3 Christchurch Earthquake of 22 February 2011. New Zealand Geomechanics
News 81: 5367.
Holden C 2011. Kinematic source model of the February 22nd 2011
Mw 6.3 Christchurch earthquake using strong motion data.
Seismological Research Letters 82: 783788.
Holden C, Beavan J, Fry B, Reyners M, Ristau J, Van Dissen R,
Villamor P, Quigley M 2011. Preliminary source model of the
Mw 7.1 Dareld earthquake from geological, geodetic and
seismic data. In: Proceedings of the Ninth Pacic Conference
on Earthquake Engineering, 1416 April, Auckland, New
Zealand, paper no. 164.
Ide S, Beroza GC 2001. Does apparent stress vary with earthquake
size? Geophysical Research Letters 28(17): 33493352.
Jongens R, Pettinga JR, Campbell JK 1999. Stratigraphic and
structural overview of the onshore Canterbury Basin: North
Canterbury to the Rangitata River. N. Z. Open File Pet. Rep.
PR4067, Natural Hazards Research Centre, Department of
Geological Sciences, University of Canterbury, Christchurch,
New Zealand: 31 p.
Jonsson S, Zebker H, Segall P, Amelung F 2002. Fault slip
distribution of the Hector Mine earthquake estimated from
satellite radar and GPS measurements. Bulletin of the
Seismological Society of America 92: 13771389.
Kanamori H, Allen C 1986. Earthquake repeat time and average
stress drop. In: Boatwright J, Das S eds. Earthquake Source
Mechanics, Maurice Ewing Series 6. Washington D.C, American Geophysical Union. Pp. 227235.
Langridge RM, Berryman KR 2005. Morphology and slip rate of
the Hurunui section of the Hope Fault, South Island, New
Zealand. New Zealand Journal of Geology and Geophysics
48(1): 4357.
McVerry GH, Zhao JX, Abrahamson NA, Somerville PG 2006.
New Zealand acceleration response spectrum attenuation
relations for crustal and subduction zone earthquakes. Bulletin of the New Zealand Society for Earthquake Engineering
39(1): 158.
Norris RJ, Cooper AF 2007. The Alpine Fault, New Zealand;
surface geology and eld relationships. In: Okaya D, Stern T,
Davey FJ eds., A Continental Plate Boundary: Tectonics at
South Island, New Zealand, Geophysical Monograph Series,
175: 157175. AGU, Washington, D. C.
Oth A, Bindi D, Parolai S, Di Giacomo D 2010. Earthquakes
scaling characteristics and the scale-(in)dependence of seismic

energy-to-moment ratio: Insights from KiK-net data in Japan.


Geophysical Research Letters 37(19): L19304.
Pasyanos ME, Dreger DS, Romanowicz B 1996. Towards real-time
determination of regional moment tensors. Bulletin of the
Seismological Society of America 92: 9981014.
Petersen TK, Gledhill K, Chadwick M, Gale N, Ristau J 2010. The
New Zealand National Seismograph Network. Seismological
Research Letters 82: 920.
Pettinga JR, Yetton MD, Van Dissen RJ, Downes G 2001.
Earthquake source identication and characterisation for the
Canterbury region, South Island, New Zealand. Bulletin of the
New Zealand Society for Earthquake Engineering 34(4): 282
317.
Quigley M, Villamor P, Furlong K, Beavan J, Van Dissen R,
Litcheld N, Stahl T, Duffy B, Bilderback E, Noble D, Barrell
D, Jongens R, Cox S 2010. Previously unknown fault shakes
New Zealands South Island. Eos, Transactions, American
Geophysical Union 91(49): 469471.
Quigley M, Van Dissen R, Litcheld N, Villamor P, Duffy D,
Barrell D, Furlong K, Stahl T, Bilderback E, Noble D 2011.
Surface rupture during the 2010 Mw 7.1 Dareld (Canterbury)
earthquake: implications for fault rupture dynamics and
seismic-hazard analysis. Geology 40(1): 5558.
Reyners M 2011. Lessons from the destructive Mw 6.3 Christchurch, New Zealand, Earthquake. Seismological Research
Letters 82(3): 371372.
Ristau J 2008. Implementation of routine regional moment tensor
analysis in New Zealand. Seismological Research Letters, 79:
400415, doi: 10.1785/gssrl.79.3.400.
Standards New Zealand 2004. NZS1170.5:2004, Structural Design
Actions, Part 5: Earthquake actionsNew Zealand.
Stirling MW, Earthquake Hazards Team 2007. Updating the
national seismic hazard model for New Zealand. Paper 072
In: Conference proceedings, 8th Pacic Conference on Earthquake Engineering, 57 Dec 2007, Singapore. Singapore:
8PCEE.
Stirling MW, Gerstenberger MC, Litcheld NJ, McVerry GH,
Smith WD, Pettinga J, Barnes P 2008. Seismic hazard of the
Canterbury region, New Zealand: new earthquake source
model and methodology. Bulletin of the New Zealand Society
for Earthquake Engineering 41: 5167.
Sutherland R, Berryman KR, Norris RJ 2006. Quaternary slip rate
and geomorphology of the Alpine fault: Implications for
kinematics and seismic hazard in southwest New Zealand.
Geological Society of America Bulletin 118: 464474.
Suzuki W, Aoi S, Sekiguchi H 2010. Rupture process of the 2008
Iwate-Miyagi Nairiku, Japan, earthquake derived from nearsource strong-motion records. Bulletin of the Seismological
Society of America 100(1): 256266.
Van Dissen R, Yeats RS 1991. Hope Fault, Jordan Thrust, and
uplift of the Seaward Kaikoura Range, New Zealand.
Geology 19(4): 393396.
Wallace LM, Beavan J, McCaffrey R, Berryman K, Denys P 2007.
Balancing the plate motion budget in the South Island, New
Zealand using GPS, geological and seismological data.
Geophysical Journal International 168: 332352.
Wood RA, Herzer RH 1993. The Chatham Rise, New Zealand. In:
Balance PF ed. South Pacic Sedimentary Basins. Amsterdam,
Elsevier. Pp. 329349.
Yamada M, Mori J, Heaton T 2009. The slapdown phase in highacceleration records of large earthquakes. Seismological
Research Letters 80: 559564.
Yang J, Sato T 2000. Interpretation of seismic vertical amplication observed at an array site. Bulletin of the Seismological
Society of America 90(2): 275285.

Vous aimerez peut-être aussi