How New Zealand actually works, the earthquakes

The earthquakes, the sequence that remade Christchurch

At 4.35am on 4 September 2010, a fault nobody had mapped tore open under the plains west of the city. Christchurch called it "the earthquake", until the following February, when a smaller one under its own edge killed 185 people and made the first the lucky one. What followed was never a single earthquake: it was a sequence, thousands of breaks on a web of faults the plains had hidden, and it is still arriving, slowly, smaller, but measurably, today.

The first shake, the fault nobody had named

The Greendale fault did not appear on any hazard map. It lay under farmland west of the city, invisible on the surface, and at 4.35am on a Saturday it ruptured for about 30 kilometres and released a magnitude 7.2, at the time, one of the strongest earthquakes New Zealand had recorded. Chimneys came down across the city, brick facades fell in the centre, power and water went out in the west, and the ground under Kaiapoi and the estuary edge slid and liquefied. Nobody died. The hour helped, the streets were empty, and so did the geometry: the rupture ran west to east and spent its sharpest energy under paddocks — GeoNet, the Darfield earthquake.

The city read it as a test passed. Building codes had worked; modern houses had flexed and stood; the brickwork that failed was the brickwork everyone already worried about. What the record showed instead was a warning the city could not yet read: an earthquake that big does not settle the crust, it rearranges the stress, and the faults near it pick up the load.

The rupture it opened on is mapped now, GNS and the councils traced its surface break in the weeks after, nearly 70 kilometres of mapped deformation across the plains. This site draws those traces at /faults, the line on the ground the shake came out of, and the older zones the survey already knew.

The second, 22 February 2011

At 12.51pm on a Tuesday, five and a half months after Darfield, a second fault broke, a magnitude 6.2 centred under the Port Hills' edge, about 5 kilometres deep and within kilometres of the central city. Smaller in magnitude by an order of energy, worse in every way that mattered: close, shallow, and at lunchtime on a working day. The acceleration under the city was among the strongest ever recorded anywhere. The CTV building and the Pyne Gould Corporation building collapsed; the cathedral's spire fell; cliff faces came down on Sumner, Redcliffs and Lyttelton. One hundred and eighty-five people died. Most of the city's brick heart came down with them, and the cordon that went up around it stayed for years — GeoNet, 22 February 2011.

Their names are on the wall the city built for them, the Canterbury Earthquake National Memorial, Oi Manawa, on the Oxford Terrace bank of the Ōtākaro. Smaller stones remember their own corners of it: the 22 February memorial near the CTV site, the St Luke's memorial on Manchester Street, the Medway footbridge memorial in Richmond, and the interment site where some of the dead rest together.

Two softening effects had run the intervening months. The buildings had already taken September's shake and hundreds of aftershocks, cracked walls do not get stronger with rest. And the ground itself, in the drained-swamp suburbs of the east, failed again and worse: liquefaction on a scale the September survey had only sketched. That half of the story has its own page, the ground that remembered it was swamp.

What an aftershock sequence is

A mainshock does not relieve a fault system; it resets it. The stress the break removed from one stretch of crust lands on the faults around it, and they answer over months and years in the pattern seismologists call an aftershock sequence: a burst, then a decay, punctuated by second shocks big enough to seed their own smaller sequences. Canterbury's catalogue reads exactly so: more than 923 shakes of magnitude 3 or greater in September 2010 alone, then the February burst, then the June 2011 doublet (a 5.9 and a 6.0 within ninety minutes of each other), the December 2011 doublet the week before Christmas, and after that the long, ragged fade: dozens a year, then a handful, then the occasional reminder that the tail is not finished — GeoNet's catalogue.

The count on this site's /sequence page stands at over 3,657 earthquakes of magnitude 3 or more, 3 of them magnitude 6 or greater, 39 of them 5 or greater, and it is still an open file, updated each build from GeoNet's catalogue. Below magnitude 3 the instruments have logged tens of thousands more.

Why it was so violent for its size

Magnitude measures the energy at the source; what a city feels is decided between the source and the surface. February stacked every term against Christchurch: the break was close (epicentres are measured in kilometres from Cathedral Square, not tens), it was shallow (energy with almost no distance to spend itself in), and the rupture ran up and toward the city. Then the ground did its own multiplying, soft, saturated soils amplify shaking the way firmer gravels don't, which is part of why the eastern suburbs shook worse than the western ones on the same morning — GNS Science, Canterbury sequence research.

None of that required a bigger earthquake. It required a closer one, which is the uncomfortable arithmetic the sequence left: Christchurch's building stock was engineered against the known Alpine Fault and the Hope Fault, and what arrived was a web of smaller faults nobody had mapped because they never broke the surface.

Was February an aftershock?

The honest answer is "in the sense that matters." It was not a re-break of the Greendale fault, it was its own rupture on a previously unmapped fault under the Port Hills' south-eastern edge. But it belonged to September: the Darfield rupture moved the region's stress around, and the February fault is where a share of it landed. Seismologists write it as one sequence for exactly that reason, and the Royal Commission that examined the building collapses treated the February event within the sequence's own record — Canterbury Earthquakes Royal Commission.

What it changed

The city's map was redrawn in the years after. About 7,000 residential properties, most of Brooklands, much of Bexley, the reaches of the Ōtākaro/Avon that took the lateral spread, were declared residential red zone, bought out, and cleared: whole suburbs whose answer to "what stood here" is now "a street, until 2012". Flat-land suburbs were graded into technical categories, TC1, TC2, TC3, that still govern how a house's foundations are engineered there, while a 2019 study graded the whole district for the next big shake, the vulnerability map of what the model expects, not what the record shows. The central city spent years behind a cordon while thousands of buildings came down, most of the heritage core among them, and the machinery that decided what rose again is its own page: the rebuild. The flooding rules that followed, set where the ground itself had dropped, are on the flooding page.

Where it leaves the city

Christchurch is the rare city whose founding question got asked twice. What stood on this ground before the city, the swamp, the pa sites, the surveyors' first sections, was the record's original question. What stood here before the earthquakes is the one a whole generation now asks first, over car parks that used to be banks and gravel lots that used to be theatres. The sequence is why the two questions sit on the same site.

Last reviewed 2 October 2026, sources: GeoNet earthquake catalogue and Canterbury sequence record, GNS Science sequence research, Canterbury Earthquakes Royal Commission findings, this site's sequence, liquefaction, faults, red zone, grades and before pages. Corrections: /corrections.