Surprise findings: What scientists just learned about the major tsunami threat to NZ’s north

Surprise findings have changed what scientists have long assumed about New Zealand’s major tsunami threat from the north.

And for once, the news isn’t so bad, with updated simulations suggesting the quake-making potential of the vast Hikurangi-Kermadec subduction zone is somewhat lower than they’ve long feared.

Around the world, subduction zones have been responsible for some of the planet’s biggest cataclysms: most recently the devastating 2004 Sumatra and 2011 Tōhoku “megathrust” earthquakes and tsunamis.

But much has been unclear to scientists about the worst that might be unleashed from the sprawling margin stretching thousands of kilometres above New Zealand’s northeast, where the Australian and Pacific plates collide.

“Finding geologic information about ‘great earthquakes’ is very challenging with a lot of uncertainty and, historically, we haven’t seen any great earthquakes either,” said GNS Science seismologist Dr Bill Fry, a co-author of a just-published study detailing the findings.

“The resulting lack of information has left us struggling to know what the next great earthquake might look like and how we should prepare for it.”

Some of the biggest events in the zone have struck around Raoul Island, the largest and northernmost of the main Kermadec Islands.

That included a huge magnitude-8.2 quake in January 1976 and an 8.1 event in March 2021 – both of which hit shortly after slightly smaller ones.

The Hikurangi-Kermadec subduction zone, stretching for thousands of kilometres to New Zealand's northeast, marks the boundary where the Pacific and Australian tectonic plates collide.

Large as those big quakes were – the latter sparked widespread warnings and evacuations in New Zealand and the Pacific – earlier models have suggested the very biggest in the region might be able to reach 9.5, equivalent to that recorded near Chile in 1960.

But now, scientists aren’t so sure.

In the latest study, earthquakes simulated using state-of-the-art models maxed out at 9.17 – still nonetheless the size of the quake that struck off Japan in 2011.

That owed to hidden “stress barriers” found at key junctions in the subduction zone which acted like brakes, preventing the fault from slipping all at once.

“These barriers stop the earthquakes from rupturing the whole subduction interface, which could prevent us from facing the previously worst-imaginable earthquakes,” Fry said.

“They suggest that the maximum credible earthquake is smaller than previously thought – this is great news.”

The insights mark a step change in how scientists have used earthquake simulators, many of which have relied on uniform assumptions about how stress is spread across large fault zones.

But real-world data has shown these don’t quite capture the chaotic nature of earthquakes, and often turn out repetitive patterns.

In the latest study, led by GNS’ Yi-Wun Mika Liao, variability was added to the stress models, making the simulations more realistic by mimicking the complex conditions beneath the Earth’s crust.

“Variability in stress helps us better understand how often, where, and how big future earthquakes might be,” Fry explained.

“Significantly, it helps us estimate the maximum credible size of future earthquakes.”

On March 5, 2021, hundreds of people evacuated homes, hostels and businesses at Paihia and Waitangi and headed to high ground at Waitangi Golf Course when a tsunami alert sounded. Photo / Peter de Graaf

That wasn’t the only fascinating new finding.

The improved models also managed to replicate those “doublet” events near Raoul Island in 1976 and 2021, offering a glimpse into why this region might be prone to rapid-fire quakes.

Fry said earlier studies had shown the “megathrust” fault around Raoul Island was a strong one – but variation in its structure mattered, too.

“Our results have suggested that the geometry of the fault relative to the forces acting on it might be the reason for the strength and the frequent, closely timed earthquakes.”

Fry said the team had been left surprised and encouraged by how much complexity their models had been able to capture.

It meant they could use the same approach to model earthquakes here in New Zealand, while also making better estimates of tsunami risk.

“This increased knowledge helps us prepare for the next great earthquake and tsunami.”

This image shows the location of a major 8.2 quake in the Kermadec Islands that struck on March 5 - just hours after an unrelated 7.2 event near the East Cape. Image / USGS

The study, published in Geophysical Journal International, comes months after scientists warned a major, quick-fire Hikurangi tsunami could kill more than 22,000 and injure nearly 26,000 – even assuming three-quarters of people could evacuate in time.

The risk is also greater than we might think, with recent studies giving one-in-four odds of a magnitude-8.0 event occurring beneath the North Island within the next 50 years.

In the event of a sustained, strong earthquake with potential to cause a tsunami, people in vulnerable coastal areas were urged to heed the “long, strong, get gone” message – and move immediately to the nearest point of high ground, or as far inland as possible.

Jamie Morton is a specialist in science and environmental reporting. He joined the Herald in 2011 and writes about everything from conservation and climate change to natural hazards and new technology.