The hard rock near the quake's epicentre close to Lyttelton may have compounded the effect of the tremor by reflecting greater seismic activity towards the city.

Seismologists say the unlikely combination of depth, size and proximity to a populated region all contributed to the devastation.

The location of the epicentre close to the Port Hills, which are largely basalt rock, may have also played a role in the catastrophic damage.

Geologist Hamish Campbell of GNS Science said seismic activity travels in waves, so it could be compressed, refracted or reflected like an optical waveform.

While the waves could be dispersed or absorbed by some properties, they could also rebound off hard surfaces.

"We suspect that the epicentre was probably on the north side of the Port Hills, where a huge amount of energy would have been literally 'pinged' off the basalt rock, almost behaving like a mirror," said Dr Campbell.

"You can imagine an explosion going off and energy going out towards Christchurch city, but a lot of energy also hitting the hard rock at depth, then being reflected, bounced back and compounding the effect."

This event, called seismic lensing, could explain the hotch-potch damage to the city and suburbs.

"It'll partly be why there are some parts of Christchurch bizarrely much worse affected than others. A lot depends on that nature of the ground immediately beneath the building, but you have to remember that these waves can behave in strange ways.

"It could be that the Pyne Gould building, for instance, might have conformed to all the building standards but was hit by an absolute rogue wave."

Dr Campbell cited the example of a Californian earthquake which struck inland, but mysteriously brought down several buildings in Santa Monica, a huge distance from the original faultline.

Seismologists are still digging through a wealth of seismic data to understand how a magnitude 6.3 quake produced the largest recorded ground-shaking in New Zealand.

The quake - which is an aftershock of the September tremor - reached 9 on the Modified Mercalli intensity scale, which runs to 12.

This intensity destroys most buildings of a low standard, and even hurts post-1980s buildings which are specially designed to withstand earthquakes.

"What the instruments are telling us is that it was a very energetic earthquake for its size," said Geonet project director Ken Gledhill.

Dr Gledhill estimated that the previously unknown faultline ran east-to-west from Taylors Mistake to Halswell. It was roughly 3km to 12km deep and 17km long.

The aftershocks that were occurring along this faultline were tailing off more quickly than the September quake, but would still continue for months.

There have been more than 120 shocks since Tuesday's tremor, the greatest of which was magnitude 5.7.

A team of scientists failed to find a surface rupture in the hills above the epicentre, which meant it was more difficult to understand the nature of the fault and the size of the "stress drop" - the difference between the stress across a fault before and after an earthquake.

Professor Euan Smith, from Victoria University, predicted that the fault had been "well-cemented" after lying quiet for thousands of years.

"So when it went, it went with a particularly violent bang."

While the energy of September's Darfield quake was the equivalent of a car slamming on its brakes sharply, Tuesday's Lyttelton quake appeared to be a longer, heavier release of tension.

Satellite images should soon determine how much the earth moved, and the size of the stress drop.