John Elliott, Simon Lamb, Zhenhong Li, Barry Parsons
On the 4th of September 2010, a large magnitude 7 earthquake struck the South Island of New Zealand, west of the largest city of the Island, Christchurch. COMET+ researchers, in collaboration with field geologists in New Zealand, are measuring the ground displacements due to this event in order to investigate the rupture on this previously unrecognised fault.
In the early hours of the 4th September, a magnitude 7.0 earthquake occurred on the South Island of New Zealand with significant damage to property but fortunately no loss of life. The epicentre as determined by the USGS is located 50 km west Christchurch (the main population centre in the South Island, pop. ~1/3 million). Initial seismological solutions indicated an east-west striking right-lateral strike-slip fault that occurred relatively shallowly in the earth's crust. Aftershocks recorded by the New Zealand GeoNet (Figure 1) support an east-west striking fault with a fault rupture zone about 50 km long.
New Zealand is part of the boundary between the Pacific Plate in the east and the Australian Plate in the west. To the north, the Pacific Ocean subducts under the Australian plate whereas in the south the converse is true. Between these two regions is the South Island of New Zealand where the oblique collision has formed the Southern Alps which is cut by a series of NE-SW striking strike-slip faults. This region is well known for large strike-slip earthquakes and seismicity, with many events of a similar size occurring in the last 150 years of significant settlement. An unusual feature of this event is that it did not occur on the main faults that exist 100 km to the north on the Marlborough Fault Zone, nor on the main Alpine Fault 100 km to the north-west bounding the southern Alps, but instead occurred on the low-lying Canterbury Plains. Furthermore the orientation of this event is not aligned with the main strike of the faults in this region which are more north-east to south-west, nor is it associated with any clear topographic feature in the landscape.
A few days after the earthquake, field observations of ground rupture and offset features were made. Large offsets of up to 4.5 m were observed, shifting roads, field fences and tree lines out of alignment (Figure 2). These observations provide a very useful constraint of the fault models developed using the satellite radar as they provide an important ground truth as well as a detailed map of the fault trace. Ground cracks and significant displacements over 1 metre were observed to have occurred over a distance of 20 km.
InSAR Satellite Observations
Using radar measurements from the Japanese ALOS satellite (JAXA), we were able to construct interferograms of ground displacement due to the earthquake (Figure 3). This contour map of surface motion indicates that the ground north of the fault moved over one metre away from the satellite, and the ground south of the fault (where the InSAR measurements get closer to the fault rupture) moved towards the satellite by almost 2 metres. We will use further satellite radar observations to constrain our fault models better as more data becomes available.
Preliminary Fault Model
These satellite observations are used to construct a model of the slip that occurred on the fault at depth. A simple single-segment, uniform slip fault is used to represent the earthquake rupture to initially constrain its orientation and dip angle. Our preliminary model suggests an east-west striking, near-vertical fault which is predominately strike-slip but with a thrust component. Using this simple model, we then calculate the slip on a series of 1 km wide patches to represent the outline of the fault underground (Figure 4). This picture indicates that the slip is mainly confined to the upper 10 km of the Earth's crust and extends for about 40 km. The peak slip of about 8 m occurred at 4 km depth. The uppermost portion of the fault shows that significant surface slip occurred over a distance of 30 km with an offset up to 4 m in the centre which supports the field observations.
Figure 1: Aftershock distribution in the 11 days following the main shock as recorded by the GNS Geonet (http://magma.geonet.org.nz/resources/quakesearch/) of New Zealand. The aftershocks' magnitude is shown by the size of the circle and the circles are coloured by the number of days after the main shock. The main focal mechanism determinations (black and white beachballs) suggest a predominately strike-slip event on a near-vertical fault. A clear east-west alignment of aftershocks is observed to occur over a distance of about 50 km.
Figure 2: Ground displacement along the fault rupture clearly showing the right-lateral sense of motion as the once straight fence line is offset abruptly by about 4 metres at a very sharp discontinuity in the field. (Photo courtesy of Dr. Simon Lamb).
Figure 3: Satellite radar observations of ground displacement towards and away from the satellite. The interferogram is a contour map of ground motion with each colour cycle of blue through green to red indicating almost 12 cm of ground motion away from the satellite. Therefore the ground surface north-west of Christchurch has moved more than 8 colour cycles (fringes) away from the satellite indicating over a metre of ground motion along the satellite's look direction (which is 39 degrees from the vertical). Therefore this side of the fault has moved to the right, towards Christchurch. The ground surface to the south-west has moved 17 fringe cycles (almost 2 metres) towards the satellite and away from Christchurch. The SAR images used in this research were provided by the Japanese Aerospace Exploration Agency (JAXA) under a JAXA AO project. The ownership of PALSAR data belongs to METI (Ministry of Economy, Trade and Industry) and JAXA
Figure 4: The slip distribution due to the New Zealand earthquake as determined by InSAR observations. The fault is orientated approximately east-west on a near vertical plane. The fault is broken up into many small 1 km square patches and the slip determined of each using elastic dislocation models. The slip appears constrained to the uppermost part of the earth's crust with significant slip predicted at the surface of up to 4 m. The fault rupture length is about 40 km whereas the fault width is about 10 km.