Ask an Expert: where does liquefaction come from?

00:56, Aug 03 2011
Freeville School principal John Leonard considers the liquefaction silt which poured into the school grounds.

From what depth does the liquefaction silt come and how thick is the layer? Second, I am wondering whether covering liquefaction-prone land with a heavy, impermeable layer (eg, asphalt, concrete) makes liquefaction worse. There has been talk of putting houses on land that did not suffer bad liquefaction (such as the Rawhiti golf course). Is the placement of roads and housing likely to change the way that the land responds to earthquakes with respect to liquefaction? Does a solid surface such as a road act like a plunger and aggravate liquefaction, sucking more of it to the surface? - MARTINE CARTER

Christchurch soils are highly variable and often change significantly, even within short distances. Hence the depth of liquefaction and the thickness of the liquefied layer may vary substantially between different locations. In some cases, only the shallow 2 to 3 metres of the deposit have liquefied. In other cases, the depth of liquefaction reached down to 7 meter or 8 meter and was even deeper in extreme cases. In general, larger thickness of the liquefied layer results in more severe manifestation of liquefaction on the surface. Also, the pressure of the water spurting on the surface during or immediately after liquefaction increases with increasing depth of liquefaction.

Covering the ground surface with asphalt will not affect much the development of liquefaction, especially if the liquefied layer is several metres thick. Similarly, one or two-storey houses are not going to dramatically change the land response with respect to liquefaction. Any structure, however, does interact with the surrounding soils and influences their response to some extent. Roads, in particular, are different from the surrounding native soils and may provide easier pathways for liquefied silt-water mixtures.

There were areas in Christchurch where liquefaction was more severe on the roads compared with adjacent properties. Pipe breaks and collapse of cavities made the effects of liquefaction on roads even worse.

Associate Professor, department of civil and natural resources engineering, Canterbury University


Assume there had been a huge hole dug at Charing Cross, 1 kilometre wide and 10.5km deep. How far down does the shingle go? What is under it? Just solid rock? Is it all homogenous or of different materials? How hot is it down there? Assume we are filming down in the hole (at the hypocentre) at 4.35am on September 4. What would we see? - ANDREW EARL

How far down does the shingle go? We can't be entirely sure of the depth of these deposits without drilling a well down through them, but throughout most of the eastern Canterbury Plains it is fairly safe to assume gravels, sand and silt deposited by rivers such as the Waimakariri and Rakaia extend to a depth of at least several hundred metres.

What is under it? Just solid rock? Under the shingle are layers of conglomerate (hardened/ cemented shingle), siltstone, sandstone and limestone with a total thickness of 1km to 2km and ranging in age from about one to two million years at the top to about 100 million years at the base. This sequence of rocks rests on harder, older and much more deformed Torlesse basement rocks comprising, mainly, siltstone and sandstone that comprise most of the basement throughout the South Island, and with increasing depth they get cooked, pressurised and transformed into a metamorphic rock called schist.

Is it [the shingle] all homogenous or of different materials? The shingle is certainly comprised mainly of gravel, but there are layers of silt and sand interbedded throughout.

How hot is it down there? Rather warm. In this part of the South Island temperature increases with depth at about 30 to 35 degrees Celsius for every 1km, so at 10km to 11km the temperature would be about 300C to 380C (the pressure would be fairly intense as well).

Assume we are filming down in the hole (at the hypocentre) at 4.35am on September 4. What would we see? As well as having cameras down there, it would be great to have the hole kitted out with pressure sensors, fluid gauges, thermometers and microphones. We'd want to start filming well before the main shock to make sure we catch the important precursory phenomena leading up to the full-on rupture.

Wide-angle shots would be great, but I'd hope the director would also be calling for close-ups (even down to the microscopic level) of the fault zone to capture crack formation, their subsequent coalescence into through-going plains, and to document the grain and frictional property changes going on during these critical pre- climatic moments. Once things started to unzip in earnest, the noise would be phenomenal. The pressure waves would hit you like a freight train, over and over again, then fade away.

Brine-type fluids may start to flow through cracks and one side of the hole would violently get displaced across the fault by several metres relative to the other side of the hole. We'd need the camera to record how fast this shift happened, and if it grew after the main shock, and over what time period.

Geologist, GNS Science

With one of the February aftershocks I was in my garden and heard the rumble of a quake to the west of me coming towards me. When I looked at the quake map, the quake was near New Brighton Beach, so the rolling went towards the quake. Is this usual? - KEVIN BARTLETT

It's not that usual. What you possibly experienced was a side- reflection of slow seismic waves off the side of the Port Hills. We have observed some examples of such reflections in the seismic records, especially for some of the September aftershocks. The waves are effectively bouncing off the sub-surface expression of the hills and the sides of the Christchurch sedimentary basin. Such effects could be extreme in small valleys, such as Heathcote Valley.

Seismologist, GNS Science

As the last two main quakes and many recent aftershocks seem to have been located either on or near Banks Peninsula, has the volcanic peninsula anything to do with causing this year's quakes? My question is divided into several parts:

Are the small faulting areas breaking up in the volcanic strata or are they in the original bedrock?

I presume the demarcation between the underside of volcanic structure and the original bedrock is at about the same depth as the depth of Pegasus Bay is below sea level, which is not all that deep. So if the small aftershocks that are occurring in and around the peninsula are somewhere between 5km and 10km deep, then the rock fractures must be in parent bedrock and not in volcanic basalt.

If this is so, are these newly awakened short fault areas under extra pressure (and therefore more vulnerable to fracture) because of the considerable weight of the volcanic rock, soil and biomass of the peninsula pressing down on the original rock?

I heard that the bedrock around the volcanic vents associated with the ancient volcanoes may have weakened the bedrock immediately around them and these areas could be a source of quakes. So far the peninsula's aftershocks seem to be located mostly away from what we have been told are the vents for the volcanoes, unless you count the Port Hills Fault as being close to the ancient Lyttelton volcanic vent-tube.

Was the September 4 quake and associated aftershocks the "detonator" that set off the next two series?

You are correct in that most aftershock activity is originating at greater depth, in the previously faulted greywacke basement and younger bedrock units overlying this basement, but beneath the volcanic rock formations forming Banks Peninsula. In the southwest part of Lyttelton volcano, the greywacke and other rock units underlying the volcanics are exposed in the lower slopes from Charteris Bay to Gebbies Pass and across towards Governors Bay. This provides an important constraint on the depth of the volcanic rocks beneath Lyttelton volcano; ie, close to sea level.

Our recently completed seismic reflection surveys, both onshore and offshore, have revealed the presence of a clear horizon that can be correlated to the projected continuation of the volcanic slopes forming the present-day topography of the volcanic cones. As we trace this horizon in the subsurface away from the areas immediately adjacent to the volcanoes, it progressively flattens out into what we infer to be like a ring-plain of volcanically derived sedimentary strata eroded from the cones. The depth of this horizon beneath the surface ranges to more than 500m below the seafloor in Pegasus Bay, and similarly several hundred metres deep beneath the Canterbury Plains, and progressively deepening away from the immediate vicinity of the peninsula slopes.

The peninsula volcanic rock will be adding a small component of weight acting vertically down on the rock formations and faults located at greater depth, especially where there is high relief on the peninsula today. However, the earthquake sequence has been following a series of pre-existing or "inherited" older faults in the deeper rock formations, and the earthquake triggered fault movements reflect the regional tectonic setting adjacent to the main active plate boundary zone through South Island, rather than the role the volcanic pile has on influencing stress conditions there.

Undoubtedly, beneath the volcanoes and in proximity to the volcanic centres the older bedrock units will have been intensely fractured and altered by the magma intruding up through the crust before erupting on to the surface. However, the major earthquakes and the clusters of aftershocks are not closely linked with the areas where the volcanic centres are, so there is no direct association that we can detect.

Your last question is with regard to the role the September 4 Darfield earthquake might have played in triggering the subsequent February 22 and June 13 events. The answer is yes it did.

The subsequent larger "aftershocks" are triggered earthquakes following the September 7.1. The September quake imposed stress changes caused by the ground movement in the upper crustal rock formations surrounding the fault rupture, especially projecting eastward from the tip of the fault rupture in September. This in turn created stress concentrations on a number of faults to the east, and closer to the city.

The Port Hills Fault rupture of February 22 will in turn have contributed also to the June event. In summary, the crustal distortions driven by the September event are progressively being accommodated by fault slip on a network of faults extending to the east and southeast of Christchurch.

PROFESSOR JARG PETTINGA Head of geological sciences, Canterbury University

* Email your questions to

The Press