Ground Motion and Seismic Source
Aspects of the Canterbury Earthquake
Sequence
Brendon A. Bradley,
a)
M.EERI, Mark C. Quigley,
a)
Russ J. Van Dissen,
b)
and Nicola J. Litchfield
b)
This paper provides an overview of the ground motion and seismic source
aspects of the Canterbury earthquake sequence. Common reported attributes
among the largest earthquakes in this sequence are complex ruptures, large dis-
placements per unit fault length, and high stress drops. The Darfield earthquake
produced an approximately 30 km surface rupture in the Canterbury Plains with
dextral surface displacements of several meters, and a subordinate amount of ver-
tical displacement, impacting residential structures, agricultural land, and river
channels. The dense set of strong ground motions recorded in the near-source
region of all the major events in the sequence provides significant insight into
the spatial variability in ground motion characteristics, as well as the significance
of directivity, basin-generated surface waves, and nonlinear local site effects. The
ground motion amplitudes in the 22 February 2011 earthquake, in particular, pro-
duced horizontal ground motion amplitudes in the Central Business District
(CBD) well above those specified for the design of conventional structures.
[DOI: 10.1193/030113EQS060M]
INTRODUCTION
The 2010–2011 Canterbury earthquake sequence includes the 4 September 2010 M
w
7.1
Darfield earthquake (e.g., Gledhill et al. 2011) and three subsequent earthquakes of
M
w
≥ 5.9, most notably the 22 February 2011 M
w
6.2 Christchurch earthquake (e.g., Kaiser
et al. 2012) that resulted in 185 fatalities. All of the earthquakes occurred on previously
unknown faults. The Darfield earthquake was the only event in which a surface rupture
was generated (Figure 1a), causing significant damage to houses (Figure 1c), roads,
power poles, and agricultural land, among others (Quigley et al. 2012, Van Dissen et al.
2011). Ground shaking in the Darfield earthquake resulted in widespread liquefaction in east-
ern Christchurch and in isolated areas throughout the region (Cubrinovski et al. 2010) and
substantial damage to unreinforced masonry structures (Dizhur et al. 2010).
The M
w
6.2 Christchurch earthquake caused significant damage to commercial and resi-
dential buildings of various eras (Buchanan et al. 2011, Clifton et al. 2011, Kam et al. 2011).
The severity and spatial extent of liquefaction observed in native soils was profound and was
the dominant cause of damage to residential houses, bridges, and underground lifelines
Earthquake Spectra, Volume 30, No. 1, pages 1–15, February 2014; © 2014, Earthquake Engineering Research Institute
a)
University of Canterbury, Private Bag 4800, Ilam, Christchurch, New Zealand
b)
GNS Science, P.O. Box 30368, Lower Hutt, Wellington, New Zealand
1
(Cubrinovski et al. 2011a). Rockfall and cliff collapse occurred in many parts of southern
Christchurch (Massey et al. 2014, Dellow et al. 2011). The 13 June 2011 M
w
6.0 earthquake
caused further damage to previously damaged structures and severe liquefaction and rock-
falls, and similarly for the M
w
5.8 and M
w
5.9 earthquakes on 23 December 2011. Several
additional smaller aftershocks have also induced localized surface manifestations of lique-
faction (e.g., Quigley et al. 2013), rockfall, and building damage. This paper provides a sum-
mary of seismic sources and ground motion characteristics of the Canterbury earthquake
sequence in order to provide context for subsequent papers in this special issue on structural,
geotechnical, and lifeline performance.
CHARACTERISTICS OF SEISMIC SOURCES
The Canterbury sequence occurred within ∼30 km thick continental crust in a relatively
low strain rate region at the periphery of the Pacific–Australian plate boundary deformation
zone in New Zealand’s South Island. The local geology consists of Mesozoic greywacke
Figure 1. (a) Epicenter locations for M
L
≥ 3.0 events from 4 September 2010 to 10 February
2013 (data from www.geonet.org.nz). Projected surface locations of major blind faults in yellow
and subsurface locations in transparent white (from Beavan et al. 2012) and location of mapped
surface ruptures in red (from Quigley et al. 2012). Loc ations of selected strong ground motion
stations as shown; full station names appear in Table 1. (b) Partial avulsion and related flooding of
the Hororata River in the Darfield earthquake. Mapped Greendale fault trace from Duffy et al.
(2013) in red; black arrows and “U” (up) and “D” (down) denote relative movement across fault.
Blue arrows denote river flow direction. (c) Greendale fault traces running through residential
property. Note lack of dwelling coll apse despite being situated directly on the surface fault
rupture.
2 BRADLEY ET AL.
bedrock variably overlain by a 1–2 km thick package of Late Cretaceous to Neogene sedi-
mentary and volcanic rocks and Pliocene to Quaternary alluvial gravels that locally exceed
1 km in thickness. GPS-derived principal horizontal contraction in the region occurs at
16 nanostrain/year with an azimuth of 110–120° (Wallace et al. 2007).
The 4 September 2010 M
w
7.1 Darfield earthquake was a complex event, beginning
on a steep reverse fault and involving the rupture of at least 7 fault segments including
EW-striking right-lateral faults, NE-striking reverse faults, NNW-striking left-lateral faults,
and NW-striking normal right-lateral faults (Beavan et al. 2012, Elliott et al. 2012).
The largest moment release resulted from the right-lateral rupture of the Greendale fault
(equivalent to a M
w
6.9 7.0 earthquake), which was the only fault to generate a surface
rupture (Figure 1a). Maximum subsurface slip was concentrated at depths of 2–6km
(Beavan et al. 2012) and may have exceeded 7 m over a strike length of ∼7 8km(Elliott
et al. 2012). The combined subsurface fault length is inferred to be ∼48 km (Beavan et al.
2012). The inferred rupture extents of other blind faults (Figure 1a) that ruptured in the
Darfield earthquake range from 0.5–1 km depth (Beavan et al. 2012), suggesting that rup-
ture likely ceased near the base of the Pliocene (∼1km) or Quaternary (∼0.5 km) sedi-
mentary deposits (Jongens et al. 2012). The 29.5 0.5 km–long Greendale fault surface
rupture had a maximum surface displacement of 5.3 m (Quigley et al. 2012). Surface dis-
placement measurements in the central Greendale fault above areas of maximum inferred
subsurface slip typically range from 4–5 m (Figure 2), indicating an apparent decrease in
coseismic slip toward the surface. Steps in the fault surface slip gradients occur in fault
trace step-overs and where other blind faults project to intersect the Greendale fault
(Figures 1 and 2).
The 22 February 2011 M
w
6.2 Christchurch earthquake involved the rupture of 2–3
blind faults (Figure 1a) with reverse and right-lateral displacements (Beavan et al.
2012). Inferred rupture extents were ∼0.5 km depth below surface, suggesting rupture
termination in Miocene volcanic rocks. Maximum coseismic slip was 2.5–3 m at depths
of 4–6km(Beavan et al. 2012, Elliott et al. 2012). The 13 June 2011 M
w
6.0 earthquake
likely involved an intersecting ENE-striking reverse-right lateral fault and NW-striking
left-lateral fault with ∼1km–deep rupture extent and maximum subsurface slip of
<1m(Beavan et al. 2012). The 23 December 2011 M
w
5.8 and M
w
5.9 earthquakes rup-
tured 1–2 largely offshore, NE-striking reverse-right-lateral, blind faults with maximum
slip of >1.4 m occurring at depths of 2–5 km and rupture extents of ∼1kmdeep (Beavan
et al. 2012).
Large surface-slip-to-surface-rupture length was reported for the Greendale fault by
Quigley et al. (2012), and large subsurface-slip-to-subsurface-fault length ratios were
reported for the Christchurch earthquake source (Beavan et al. 2012, Elliott et al. 2012),
implying that large slip-per-unit fault length may be a characteristic of some of the faults
in this region. Using surface rupture data and an elliptical fault displacement model, Quigley
et al. (2012) computed a static stress drop of 13.9 3.7 MPa for the Greendale fault rupture
in the Darfield earthquake. Using InSAR-derived fault models, Elliot et al. (2012) computed
stress drops of 6–11 MPa for individual fault segments in the Darfield earthquake and
14 MPa for the Christchurch earthquake. These results are consistent with reported stress
drops from earthquakes on other faults with long recurrence intervals near the periphery
GROUND MOTION AND SEISMIC SOURCE ASPECTS OF THE CANTERBURY EARTHQUAKE SEQUENCE 3
of plate boundary deformation zones (e.g., 7–12 MPa in Landers-Hector Mine earthquake
sequence; Price and Bürgmann 2002). For comparison, Fry and Gerstenberger (2011) cal-
culate the apparent stress of the Darfield and Christchurch earthquakes to be ∼16 MPa and
∼4 MPa , respectively.
GREENDALE FAULT SURFACE RUPTURE AND ENGINEERING AND
LAND-USE IMPLICATIONS
Field mapping and surveying, combined with LiDAR data, was used to define the
Greendale fault surface rupture trace (Figure 2). The maximum right-lateral surface displace-
ment was 5.3 m. Vertical displacement was typically on the order of tens of centimeters in
flexure and bulging, but at several fault bends, vertical displacement reached 1–1.5 m.
Perpendicular to fault strike, surface rupture displacement was distributed across a ∼30 m
to 300 m wide deformation zone, largely as horizontal flexure (Figures 2c and 2e). On aver-
age, 50% of the horizontal displacement occurred over 40% of the total width of the defor-
mation zone, with offset on observable discrete shears, where present, typically accounting
Figure 2. (a) LiDAR hillshade image o f a typical section of the Greendale fault surface rupture.
(b) Photo showing along-strike variation of surface rupture deformation zone width (the two bare
fields are each ∼40 m wide, and total right-lateral displacement is ∼4.5 m). (c) Plots of cumu-
lative strike-slip surface rupture displacement and histograms of displacement distribution at two
representative sites across the Greendale fault, illustrating that surface rupture deformation is
widest, and more evenly distributed, at step-overs (profile 38), and narrowest and more spiked
where rupture comprises a single trace (profile 39). (d) Net surface rupture displacement along the
Greendale fault (after Quigley et al. 2012). (e) Width (horizontal distance) measured perpendi-
cular to fault strike over which it takes to accumulate 50% and 100% of the total dextral surface
rupture displacement at 40 sites along the Greendale fault (after Van Dissen et al. 2011).
4 BRADLEY ET AL.
for less than about a third of the total displacement. Characterisations of fault displacement,
such as those depicted in Figure 2c, are relevant for both planning fault avoidance set-back
distances (e.g., Villamor et al. 2012) and for designing surface rupture–resilient buildings and
infrastructure (Bray and Kelson 2006, Rockwell et al. 2002).
About a dozen buildings, mainly single-story houses and farm sheds, were affected by
surface rupture, but none collapsed. This was largely because most of the buildings were
relatively flexible, resilient timber-framed structures, and also because deformation was dis-
tributed over a relatively wide zone. There were, however, notable differences in the respec-
tive performances of the buildings. Houses with only lightly reinforced concrete slab
foundations suffered moderate to severe structural and nonstructural damage. Three other
types of buildings performed more favorably and far exceeded life-safety objectives: one
had a robust concrete slab foundation that was stronger than the surrounding soil, another
had a shallow-seated pile foundation that isolated ground deformation from the superstruc-
ture, and the third had a structural system that enabled the building to tilt and rotate as a rigid
body. This third building suffered very little internal deformation, was straightforward to re-
level, and demonstrated, serendipitously, the potential for a high degree of post-event func-
tionality for certain types of buildings in relation to, in this case, distributed surface fault
rupture (Van Dissen et al. 2011).
In 2003, the Ministry for the Environment (MfE), New Zealand, published best practice
guidelines for mitigating surface fault rupture hazard (Kerr et al. 2003, Van Dissen et al.
2006). A key rupture hazard parameter in the MfE guidelines is fault complexity. For a
given displacement, the amount of deformation at a specific locality is less within a distrib-
uted rupture zone, than it is within a narrow zone. Surface rupture displacement on the
Greendale fault was typically distributed across a relatively wide zone of deformation. Build-
ings located within this distributed zone of deformation were subjected to only a portion of
the fault’s total surface rupture displacement, and no building within this zone collapsed. This
provides a clear example of the appropriateness of the MfE’s distributed fault complexity
parameter, at least for Building Importance Category 2a buildings (i.e., residential structures)
and with respect to life-safety.
CHARACTERISTICS OF EARTHQUAKE-INDUCED GROUND MOTIONS
Table 1 provides a summary of the near-source ground motions resulting from the major
events of 4 September 2010, 22 February 2011, 13 June 2011, and 23 December 2011. The
largest ground motions in central Christchurch occurred during the 22 February 2011 Christch-
urch earthquake primarily as a result of its close proximity to the earthquake source. Severe
ground motions were observed at numerous strong motion stations over the multiple events.
Peak accelerations of up to 1.41 g and 2.21 g were recorded at HVSC in the horizontal and
vertical directions, respectively. In the CBD (i.e., CBGS, CHHC, CCCC, and REHS stations),
PGA values ranging from 0.37–0.52 g were observed in the 22 February 2011 event.
GROUND MOTION INTENSITY IN THE CENTRAL BUSINESS DISTRICT (CBD)
Figure 3 illustrates the pseudo-acceleration response spectra of four strong motion sta-
tions (CCCC, CHHC, CBGS, REHS) located in the CBD region during the aforementioned
four events. Despite their geographic separation distances (relative to their respective
GROUND MOTION AND SEISMIC SOURCE ASPECTS OF THE CANTERBURY EARTHQUAKE SEQUENCE 5
