If 2016's Kaikoura Earthquake happened tomorrow, scientists would immediately be able to identify regions with potentially damaging shaking, thanks to new capability. Photo / Mike Scott
New Zealand now has one of the best earthquake response systems in the world, thanks to a years-long, multi-million-dollar programme enabling scientists to model shaking and impacts in real-time. GNS experts explain to Jamie Morton how it works.
It was one of the most complex earthquakes ever recorded, if not the most complex.
Erupting with a midnight mainshock in rural Canterbury, the Kaikoura Earthquake of November 14, 2016, tore across an incredible distance of 180 kilometres, burying highways, buckling railways and raising up an entire coastline.
Such was its force - scientists later estimated it packed the equivalent energy release of 400 atomic bombs detonating - that it set off some 21 faults, 14 of which ruptured violently enough to displace land by more than a metre.
As we all recall from those dramatic scenes around Wellington and Kaikoura, much of the worst damage was seen far from its Culverden epicentre: highlighting the immense difficulty that comes with trying to predict where a big quake’s impacts will hit hardest.
Yet, if the same event unfolded tomorrow, our scientists would be much better prepared to model these effects, in the first critical hours after an earthquake.
“We would anticipate being able to immediately identify regions with potentially damaging shaking including, in this particular case, identifying the type of strong shaking that may have affected taller buildings in Wellington,” GNS Science seismologist Dr Anna Kaiser told the Herald.
That’s how far New Zealand’s capability has come in just a few years, thanks to a sprawling, $13 million programme that’s helped deliver us one of the best earthquake response systems in the world.
Straddling a major plate boundary, our country faces two of its biggest natural threats in tsunamis and quakes, of which between 100 and 150 of some 20,000 annually recorded prove big enough to be felt.
The urgent need to upgrade our science tools has only been heightened by recent research estimating a 75 per cent chance of our largest on-land system - the 600km-long Alpine Fault - triggering a potentially-disastrous rupture within the next 50 years.
In roughly that same period, scientists have also estimated a 26 per cent chance of our major fault zone - the Hikurangi Subduction Zone - unleashing a quake of magnitude 8.0 or larger beneath the lower North Island.
Three years ago, GNS scientists began work on a programme funded by the Ministry of Business, Innovation and Employment (MBIE) to provide earlier and more accurate information about quakes as they struck, with an ultimate goal of saving lives and recovering faster.
Dubbed Rapid Characterisation of Earthquakes and Tsunami, or RCET, the effort brought to bear a range of new science to fully understand a given quake’s complex, three-dimensional nature.
That would mean scientists could, within minutes, better understand its ability to cause widespread shaking, trigger tsunamis and landslides, and damage infrastructure like roads and buildings.
Previously, Kaiser explained, our response to quakes has been based on basic estimates of where they started - or their hypocentres - along with their magnitudes.
“We have also been able to assess the level of earthquake shaking at locations where we have seismic instruments.”
In small-to-medium sized quakes, seismic waves typically travel through the earth and radiate out from an epicentre, allowing scientists to map the severity of shaking and pinpoint where it’s been most damaging.
But in larger quakes, or those measuring over 6.5, things can get trickier.
That’s because ruptures could start from an epicentre but then extend along a fault - or a raft of them, as we saw in 2016 - and sometimes for hundreds of kilometres.
In such cases, a rupture could build up as it unfolded and release a huge amount of energy far from where it began - meaning that the strong shaking was unlikely to be tidily dispersed around the epicentre.
To tackle this complexity, scientists working on the RCET project have been developing sophisticated new tools that draw on a combination of new data processing and modelling methods.
Some of them are already running in real time, giving scientists an instant stream of information about a quake’s rupture, along with its size, direction and shaking level.
“Within seconds to minutes after a big earthquake, we now have ways to estimate the full extent of the earthquake in 3D,” Kaiser said.
“This first estimate is rough, but very valuable in improving our understanding of what has happened and the potential impacts.”
It also meant scientists could double-check their understanding of a quake’s tsunami-making potential, and also determine magnitudes of major ruptures without having to rely as much on global estimates.
“With the RCET advances in mapping the 3D earthquake and its shaking, we have a clear path forward towards providing real-time estimates of impacts and losses from big events,” Kaiser said.
RCET has also changed the way our scientists now forecast tsunamis.
Earlier, when a big quake hit, they’d use uncertain magnitude estimates to select a forecast from a large database of pre-computed tsunami scenarios.
To account for the messy uncertainty around the approach, they’d also add a “buffer” that effectively doubled the height of forecasted waves, to make up for the danger of a possible underestimate of magnitude.
While international agencies like the United States Geological Survey (USGS) could provide an earthquake analysis without such a buffer, these magnitude estimates could take between 45 and 60 minutes to arrive: in which time a “regional-source” tsunami could’ve already hit.
Before RCET, however, New Zealand still had a crucial tool in its Deep-ocean Assessment and Reporting of Tsunami (Dart) buoy array.
The buoys, which have proved invaluable in recent tsunami scares, form part of the tsunami early warning system for big earthquakes that strike off our country’s eastern and northern coasts.
“With more robust magnitudes available earlier, we can move to better forecasts quicker for regional-generated tsunami, including for those smaller borderline cases,” said GNS scientist Dr Bill Fry, who’s been leading the RCET programme.
“That means less ‘over-forecasting’ that leads to precautionary over evacuation. We’ve also implemented a computational way to use data recorded from our NZ Dart array to improve our forecasting,” Fry said.
“Overcoming this challenge is allowing us to focus our efforts on our ultimate goal of time-dependent forecasts of tsunami inundation, before the waves hit.”
When ground shaking was strong and long-lasting, Fry stressed that those near the coast should still heed authorities’ “long-strong, get-gone” advice and move to higher ground.
The next step in the programme was the release of “Shaking Layers” through GeoNet: a project to provide shaking intensity maps minutes after an earthquake stronger than 3.5.
These new tools would combine ground motion recordings across GeoNet’s seismic networks with model predictions to estimate shaking at any given point in the country.
“In the three years we’ve been going, we’ve already developed and tested the first new tools that we’ll need to support better response and recovery from our next big New Zealand earthquake and southwest Pacific tsunami,” Fry said.
“In practice, this looks like better information underpinning emergency response, more accurate and quicker tsunami forecasts with less over-forecasting, and better science advice for recovery decision-making.”
The programme comes as a large team of scientists have published data mapping nearly 900 faults capable of generating moderate to large quakes.
This wealth of new information helped inform the recently-updated National Seismic Hazard Model which, compared with previous estimates of seismic hazard, showed an increased risk of ground-shaking from future quakes in places such as Blenheim, Wellington, Napier and Gisborne.
Meanwhile, Fry said New Zealand was still exploring the potential use of earthquake early warning systems, having already introduced the US-developed ShakeAlert tech.
“For now, we use this tool to rapidly map earthquake rupture,” he said.
“Early testing suggests it can be used as a capable earthquake early-warning algorithm in Aotearoa, opening up the possibility for further research to explore its full potential, in parallel to research on other components of earthquake early warning systems.”
As for being able to actually pick quakes months or years before they struck, scientists still considered that a virtual impossibility.
“To date research has not yielded reliable or practical methods to predict earthquakes,” Fry said.
“Scientists can only forecast their likelihood.”
New Waihirere Domain plan includes improved facilities; pool removal set for months.