DMG Elder, IF McCahon, M Yetton - Soils & Foundations Ltd
Technical Abstract
In considering the earthquake hazard in Christchurch it is useful to apply the law of precedence: the past is the best indicator of the future. In the first 80 years of the city's history, four large earthquakes significantly damaged the growing settlement, one seriously. Any one of these four events today would cost the city millions of dollars in direct damage and could result in major disruption to the local economy. The largest of these events was virtually under the city, with an epicentre close to New Brighton. It is nearly 70 years since the last large event (1922, Motunau earthquake, intensity VII in the city). When can the next similar event of this size be expected? In this report, the first detailed seismic hazard assessment of the city, we have attempted to answer this question by adopting current seismic hazard analysis techniques and applying them specifically to Christchurch.
Seismic hazard analysis involves three components: a seismicity model (a model of earthquake occurrence probability in regions close enough to affect the city); an attenuation model (approximating energy loss and wave modification as the seismic waves travel to the basement rock under the city) and a site response model (predicting the changes to the earthquake waves as they propagate up through the gravels, sands and silts underlying the city).
The seismicity model developed here makes use of the traditional Gutenberg Richter occurrence relationship (log N = a - bM). In the common case when b is close to one, an approximate ten-fold reduction in earthquake occurrence occurs with each step up the Richter magnitude scale. Thus by knowing the number of relatively frequent small earthquakes, the average recurrence interval of more infrequent larger events can be predicted. The basic model has been refined for the central and northern South Island by subdividing the area into 10 seismicity zones. In each zone the maximum credible earthquake has been inferred from the length (and, where known, the displacement per event) of the known active faults in the zone.
Probabilistic information is obtained from the number of earthquakes historically recorded in the zone over a given period of time. For magnitudes less than 6, the recorded instrumental data is from 1964-1988; for magnitudes greater than 6 and less than 6.5, 1940-1988; and magnitudes greater than 6.5, 1840-1989. It should be noted that even the period 1840-1989 is much shorter than the return period for major earthquakes on many faults near Christchurch and, while being the maximum record available, this time span is still relatively short.
Analysis indicates that potential exists for relatively rare but very large earthquakes (approximately magnitude 8) along the Alpine fault, which essentially marks the western edge of the Southern Alps. More frequent moderate to large earthquakes (around magnitude 6-7.5) can be expected in the Canterbury Plains foothills and North Canterbury area, and less frequent moderate earthquakes under the Canterbury Plains and Christchurch itself. The attenuation model predicts that the damage in the city from these three types of event are likely to be similar. Of the four serious earthquakes in the early city history, three occurred in the foothills and North Canterbury region (the Amuri, Cheviot and Motunau earthquakes) and one virtually beneath the city (the New Brighton earthquake).
An important component of this study has been to consider the additional effect at Christchurch of the deep, relatively soft sediment underlying the city (the site response model). This creates major changes in the nature of the earthquake shaking by modifying the ground acceleration, velocity and displacement at any frequency. In some areas of the city the earthquake vibrations are amplified. As a result the overall average hazard for the city increases when compared to areas on bedrock, (for example most of Banks Peninsular) by approximately 0 to 2 intensity units, or by 0 to 1 unit when compared to areas on 'average ground' (comprising shallow sediment). Within the city distinct local variation results in particular from gradational changes in the top 30m of sediment.
On average the calculated probabilities for various intensities of shaking in Christchurch are as follows:
| Modified Mercalli Intensity | Approximate Expected Effect | Average Return Period |
| Intensity VI | Minimal property damage. | 7 years |
| Intensity VII | Some property damage, loss of life unlikely. | 20 years |
| Intensity VIII | Significant property damage, loss of life possible. | 55 years |
| Intensity IX | Extensive property damage, some loss of life. | 300 years |
| Intensity X | Catastrophic property damage, major loss of life. | In excess of 6,000 years |
These probabilities indicate that Christchurch has an overall seismic hazard level comparable to Wellington for medium intensity earthquake shaking. However this level of hazard is lower than that in Wellington for very large catastrophic events.
Section III of the report considers surface ground damage which may occur associated with an earthquake. The greatest concern for Christchurch, located near a saturated, sand and silt rich, prograding coastline, is the potential for liquefaction. This phenomenon occurs when the tendency for loose granular materials to compact during earthquake shaking results in a pore water pressure increase, and reduction or total loss in strength. This may cause subsidence, foundation failure and damage to services. Analysis shows that large areas of the city are underlain by sands or silts which, if sufficiently loose, would be highly susceptible to liquefaction. Although insufficient soil testing has been carried out to characterise densities in all areas, extensive investigation has been done in the central city. Some silts and sands in this area are loose and extremely vulnerable to liquefaction.
It should be noted that the historical earthquakes which the city has experienced do not appear to have generated significant liquefaction in the city areas occupied at the time. However, when considering the magnitude and location of these four historical earthquakes, available analysis models indicate they were probably too small to initiate the process. The lack of past evidence does not exclude the possibility of liquefaction in the future.
Figure S3 (only in full research paper) summarises the available information regarding areas of potential liquefaction and areas of expected seismic wave amplification. It is obvious from examination of the figure that a very large part of the eastern city is potentially subject to liquefaction while amplification effects are pronounced in the areas to the north of the central city and in scattered south-western areas.
Consideration of the likely effect of a large earthquake in the hill areas suggests damage by landsliding is likely if a large earthquake coincides with the winter wet period. Wet or saturated conditions only exist each year for a relatively short time (two to four months), but assuming wet conditions exist, the probability of significant damage from soil slope movement in the hill areas is estimated to have a return period of around 300 years. Houses below steep hillslopes in rural catchments are generally most at risk. However liquefaction induced landsliding in alluvial materials along the lower reaches of the Avon Heathcote rivers, and around the margins of the estuary, may be a more significant hazard.
Chapter 11 of the report considers briefly the potential damage to the city in terms of the impact on physical structures. One consequence of the effect of the deep alluvium beneath the city is to reduce the structural response at high frequencies and amplify the lower frequency shaking. This will probably subject mid to high rise structures to levels of resonant shaking exceeding those anticipated by present design methods. Residential buildings may be shielded from resonance effects, however the anticipated large amplitude ground motion may cause inertial effects. Damage to contents, heavy furniture and fittings, hot water cylinders and chimneys is likely.
Possibly the most significant physical impact on the city may be damage to water, sewer and power supply services. With the depths of relatively soft alluvium under the city, the strains experienced by pipelines are expected to be high, with corresponding high pipe stresses and pipe joint displacements. If liquefaction occurs the sewerage reticulation system and treatment station could be severely damaged. Liquefaction could also affect main electrical supply nodes. Inertial loads on electrical heavy equipment at substations are likely to be larger than presently assumed for design with a corresponding high incidence of power failure. Subsidence or displacement of the very soft reclaimed land at Lyttelton is likely (little of this land existed at the time of previous large historical earthquakes) and the associated damage to oil tanks could have serious environmental consequences.
We have not attempted an in-depth lifelines study for Christchurch, or included economic or sociological analysis in this report. In addition to the need for this type of work, we recommend further action from the engineering profession including a review of the current seismic loadings code, local seismic design practices and building stock. We suggest site specific studies for the Lyttelton tank farm, Bromley sewerage ponds, pumping stations, substations, hospitals, civil defence facilities, airport and key bridges. Major areas of further research include studies of sand density variations and susceptibility to liquefaction across the city; continued paleoseismic evaluation of adjacent active faults, particularly the Alpine Fault, and further investigation of the deep sediments below the city.