Four months on, scientists still sifting through the aftermath of the Kaikoura Earthquake say the 7.8 event could be the most complex ever recorded.
GNS Science experts shared their latest findings yesterday in a live public presentation, showing how a total 21 faults had ruptured in the November 14 midnight quake.
Earthquake geologist Dr Kate Clark said of those 21 faults activated, 14 had ruptured violently enough to displace land by more than a metre.
One of the most dramatic examples was along the Kekerengu Fault in Marlborough, where the land offset was as much as 12m, and in some places created walls of raised-up earth across the countryside.
That compared with the largest amount of offset recorded in New Zealand, the 18m horizontal land displacement created by the 8.2 Wairarapa Earthquake in 1855, and the most recent comparable event, the rupturing of the Greendale Fault in the September 2010 Darfield quake, which produced a maximum offset of 5m.
The Kaikoura Earthquake also eclipsed the 30-kilometre rupture distance of the Greendale Fault; once the 7.8 quake struck at its epicentre near Culverden in Canterbury, it tore across a distance of 180km, setting off the 21 faults in the process.
"We haven't done a detailed analysis of comparing every quake that we know of, but we think it's probably approaching or exceeding a world record, in terms of fault surface rupture complexity."
Trying to make sense of what happened remained a huge task for scientists; a team of 50 had been out around the earthquake zone making field measurements.
Clark said those measurements, along with other data, observations made by drones and LiDAR remote sensing surveys, would help inform new topographic models to better measure the fault offsets.
While the field work was expected to wrap up within months, it could still take years to complete a full analysis of the event.
"It was an incredibly complex earthquake. The more we are working on it, the more we are appreciating the complexity of it."
Meanwhile, GNS scientist Dr William Power said tsunami gauge data recorded during the event could also better inform tsunami models.
Power said gauge readings taken at Kaikoura at the time of the quake showed that, in the 25 minutes immediately after, the tidal level dropped by 2.5m, "a classic natural warning sign for a tsunami".
In the 15 minutes following that, the tide rose 4m from the lowest point, before a series of waves were recorded over several hours.
Notably, he added, the tide level changed by a metre in the event; reflecting the fact that gauge had been hauled up by that much along with a large stretch of the coastline it sat on.
Elsewhere, the tsunami registered on gauges long after, showing up in readings at Wellington's Queens Wharf and Castlepoint an hour after, and at the Chatham Islands four to five hours following the event.
Power estimated the tsunami's run-up height at Little Pigeon Bay, where a historic cottage was wrecked, reached 3m.
The force of the tsunami was enough to lift the unoccupied cottage from its foundations and flood it with water a metre deep, leaving debris stuck to its walls.
"An adult would have survived, but an infant or someone with limited mobility would have been in grave danger."
Power and his colleagues had since used tsunami debris, including washed up kina and seaweed, to help determine the tsunami's run-up height.
GNS scientists were also busy trying to map the thousands of landslides created by the quake.
Engineering geologist Dr Sally Dellow said landslides had so far been mapped across an area of 5000sq km, about half of the total 10,000sq km area where landslides had occurred.
Around 200 landslide dams had also been created by hillsides blocking rivers, of which 11 had been identified as still posing risks.
Dellow pointed out one major landslide that came down near Ohau Point, where, had the quake struck at midday, tens to hundreds of people may have been stopped to view the local seal colony.
"We dodged a bullet at this place because the earthquake occurred at midnight."
The latest aftershock probability forecast - due to be updated next week - showed a 72 per cent chance of an average one event measuring between magnitude 6.0 and 6.9 over the next year, and a near certain probability of between five and 26 quakes - or an average 14.4 - measuring between 5.0 and 5.9.
However, these numbers were decreasing over time, in line with the "decay" seen in typical aftershock sequences.