The hidden fault that caused the February 2011 Christchurch earthquake

In September 2010, Christchurch was shaken by the magnitude 7.1 Darfield earthquake, caused by movement along faults west of the city on the Canterbury Plains. This earthquake produced a visible rent across the landscape that allowed scientists to directly measure the movement of the longest fault segment, the Greendale Fault. However, the violent magnitude 6.3 earthquake that devastated Christchurch on 22 February 2011 was caused by movement along a fault that does not appear to have broken the surface. Scientists have thus had to rely on measurements using a variety of techniques and instruments to determine its location and the nature of its movement.

Based on data from GPS stations, satellite radar images, seismographs and strong-motion recorders, the fault that caused the 22 February earthquake lies within about six kilometres of the city centre, along the southern edge of the city. The fault rupture (Fig. 1) was about 14 kilometres long, and extends east-northeast from Cashmere to the Avon-Heathcote estuary area. The fault plane extends a few kilometres offshore, but not much fault movement occurred beneath the ocean.

Figure 1:  This Google map image shows the fault plane (rectangular area) across the southern part of Christchurch and northern Port Hills. Colours on the fault plane indicate the amount of slip between the two sides of the fault (see Fig. 2). The contour lines indicate the amount (in mm) the land has risen (blue contours) or subsided (red contours) due to the slip on the fault. The white line is the contour where there was no change in height. The red, green and yellow coloured symbols show some of the GPS stations whose displacements were used to derive the fault slip model.

Figure 1: This Google map image shows the fault plane (rectangular area) across the southern part of Christchurch and northern Port Hills. Colours on the fault plane indicate the amount of slip between the two sides of the fault (see Fig. 2). The contour lines indicate the amount (in mm) the land has risen (blue contours) or subsided (red contours) due to the slip on the fault. The white line is the contour where there was no change in height. The red, green and yellow coloured symbols show some of the GPS stations whose displacements were used to derive the fault slip model.

The top of the fault lies at a depth of about a kilometre beneath the surface, and the rupture extends down along the fault plane for about seven kilometres. The fault is not a vertical cut through the earth, but rather it dips towards the south at an angle of about 65 degrees from the horizontal. The main part of the fault thus lies beneath the northern edge of the Port Hills.

Movement of land on either side of the fault plane was mixture of vertical motion and sideways slip (Fig. 2). The block of land south of the fault slid up the fault surface by as much as 2.5 metres on the section of fault near the Avon-Heathcote estuary. This raised part of the Port Hills and part of southern Christchurch. This type of fault motion is called reverse faulting. The deeper parts of the fault, and the westernmost 5 to 6 kilometres of the fault slipped predominantly horizontally by a few tens of centimetres. Relative to the fault, the land north of the fault shifted eastward while the land to the south of the fault shifted westward.

Figure 2:  Diagram showing the amount and direction of slip of the rock between the two sides of the fault. The rock on the south side of the fault has moved up and westward by as much as 2.5 metres relative to the rock on the north side of the fault. Fault slip comes to within one kilometre of the ground surface. The greatest movement was upward and toward the northwest under the Avon-Heathcote estuary area. The red star shows the location where the fault rupture started.

Figure 2: Diagram showing the amount and direction of slip of the rock between the two sides of the fault. The rock on the south side of the fault has moved up and westward by as much as 2.5 metres relative to the rock on the north side of the fault. Fault slip comes to within one kilometre of the ground surface. The greatest movement was upward and toward the northwest under the Avon-Heathcote estuary area. The red star shows the location where the fault rupture started.

Because the fault doesn’t break the surface, the land overlying the top of the fault has been slightly folded, with the south side warped upward and the north side down (see Fig. 1).

The land has gone up as much as 40 centimetres around the western side of the Avon-Heathcote estuary. The Port Hills have gone up by varying amounts, from about 5 centimetres under Lyttelton Harbour to a maximum of about 25 centimetres at the base of the hills near the Heathcote valley.

As a direct result of the fault slip, the Bexley, Aranui, Wainoni, Avondale and New Brighton areas have gone down, mostly by less than 15 centimetres. Central, northern and northeastern Christchurch have also gone down, but generally by less than 5 centimetres. However, there may be additional subsidence as a result of ground compaction and liquefaction during the strong shaking.

The fault rupture started with a small amount of slip between the two sides of the fault at about 6 km depth. Over the next few seconds the rupture spread upwards and towards the northwest, with the amount of slip increasing with time. The direction of movement—up and towards the northwest—focussed the energy of the earthquake towards Christchurch and helps explain the severe damage in the city. The suburbs of Heathcote and Redcliffs lie above the fault, resulting in heavy damage and extensive rockfalls.

A number of techniques were used to determine the fault’s position and movement.

  • Radar images taken from satellites before and after the earthquake were analysed. The images can be combined to show the total amount of shift of the ground surface, both vertically and horizontally, caused by the earthquake (see Fig. 3).
Figure 3:  Image indicating ground displacement made by combining satellite radar images taken before and after the earthquake. The coloured image shows an “interference pattern” derived from X-band radar images taken on 19 and 23 February 2011 by the Italian Cosmo-SkyMed satellite. Each colour cycle represents 1.5 centimetres of ground displacement, so the total displacement between the western edge of the image and central Christchurch is about 25 centimetres.

Figure 3: Image indicating ground displacement made by combining satellite radar images taken before and after the earthquake. The coloured image shows an “interference pattern” derived from X-band radar images taken on 19 and 23 February 2011 by the Italian Cosmo-SkyMed satellite. Each colour cycle represents 1.5 centimetres of ground displacement, so the total displacement between the western edge of the image and central Christchurch is about 25 centimetres.

  • Position data from GPS stations (Fig. 1) before and after the earthquake was analysed, using measurements from both existing continuous GPS stations and temporary stations installed after the earthquake.
  • Movement was modelled using the ground shaking recorded during the earthquake by “strong-motion” seismometers. These instruments are designed specifically for recording strong ground shaking.
  • The general region of fault slip is outlined by the aftershocks of the 22 February earthquake. There is still additional work to do to locate these aftershocks precisely, at which time they may provide additional detail on the fault rupture. Link to the latest aftershocks map.

This article describes our understanding of the Christchurch earthquake fault as at the 8th of April, about 7 weeks after the earthquake. We expect that details of the fault location and slip distribution will be updated in the future, as we incorporate more data and use more sophisticated data analysis techniques.

Acknowledgements
For discussions and sharing of early results: Pierre Briole, ENS, France; Marcello de Michele, BRGM, France; Eric Fielding, JPL, USA; SARmap group, Switzerland; INGV, Italy; GSI, Japan; Shaun Levick, Caroline Holden, Bill Fry, Stephen Bannister, Martin Reyners, all at GNS.

For CSK satellite radar data: e-GEOS, an ASI/Telespazio company, especially Andrea Celentano. For processing of the radar data: Mahdi Motagh.

For GPS data: LINZ, especially Josh Thomas and Dave Collett; GeoNet; Geosystems/Trimble New Zealand; Global Survey; Andersen & Associates, especially Brent George; Christchurch City Council; Otago University.