The landslide occurred in a suburb of the South Island of Dunedin. The remarkable picture below, from the Hancox paper, shows the state of the site three days before failure:
On the left side of the image (the east) lies a quarry from which about 300,000 cubic metres of sand had been removed. Running across the centre right of the image is a large array of cracks that were opening up as the landslide moved. These cracks clearly extend into the suburb, although they become less easy to discern in this area. In fact cracking was first noted in some of the houses in 1972 or even before. Through time the cracks grew until in 1978 they defined the rear scarp of the landslide as seen above. The cracks caused several water main breakages in late 1978 and early 1979.
The slide was extensively monitored in the period before final failure. The data suggests that the rate of movement was steadily increasing through time, with rates as high as 10-15 cm/day being noted in the week before failure!
Final failure began at about 9 pm on 8th August. It lasted about 30 mins, during which time a large block moved forward by about 50 m, leaving a graben structure behind that was about 16 m deep. This is very clear in the image below. The slide was a deep-seated translational block slide that covered an area of about 18 hectares. It was about 800 m, 400 m long and up to 40 m deep. The average movement was about × 400 m, up to 40 m thick). According to Hancox (2007) it slid down a 7°-dip slope at an average speed of about 1.7 m/min. The rate of movement was sufficiently slow that no-one was killed, although many people needed to be rescued and, of course, lost their homes.
So why did the slope fail? Well, the first key factor is that the site was susceptible to failure under natural conditions. The materials were dipping in the same direction as the slope and were weak and susceptible to sliding. There were ancient landslide deposits on the site that point to previous instabilities, well before humans could have played a major role. Second, the removal of the sand from the quarry removed support from the slope, making it far more likely to fail. In fact, Hancox suggests that the slope only needed groundwater to increase by 0.5 m for movement to start.
The slide was probably triggered by increased water levels in the slope. This is likely to have come from two sources. First, the few years before the landslide were wetter than had been the previous 20 years or so. Second, Hancox (2007) suggests that there was a leaking water main that may have been allowing 5 million litres to enter the slope each year.
New Zealand is of course a landslide prone environment, but there can be little doubt that many lessons have been learnt from this failure. By modern standards it seems amazing that the houses were still inhabited when cracks as large as those shown in the first photograph were developing, and movement rates of 10 cm per day were being recorded. The authorities in New Zealand are clearly using the anniversary of the landslide to remind people of the need to remain vigilant.
Hancox, G.T. 2008. The 1979 Abbotsford Landslide, Dunedin, New Zealand: a retrospective look at its nature and causes. Landslides 5: 177-188. Journal page.