Slow-slip earthquakes may provide the key for forecasting the world's biggest shakes

Thursday, 19 May 2016

Engineers recover a Tohoku University seafloor absolute-pressure gauge after one year on the seafloor along the plate boundary offshore of Gisborne.

Findings from a multi-million dollar international study off the Gisborne coast may one day help predict the kind of 'megathrust' earthquake and tsunami that could devastate parts of New Zealand. PAUL GORMAN reports.

The aquamarine Pacific Ocean lapping at the sunburned coastal hills north of Gisborne hides secrets that belie its tranquility.

Geophysicist Stuart Henrys, of GNS Science, with some of the ocean-bottom seismometers prior to their deployment off the coast of Gisborne in 2014. They typically spend about 12 months on the seafloor recording tremors before being retrieved and replaced by fresh instruments.

Scientists are now beginning to expose some of those secrets and the underlying turmoil of the northern segment of the Hikurangi Trench, which marks the plate boundary where the Pacific Plate is being subducted beneath the Australian Plate, on which the North Island sits.

A multi-million dollar investigation involving scientists, engineers and technical staff from New Zealand, Japan and the United States has concluded that areas where slow-slip earthquakes (SSEs) occur on this offshore subduction zone can also generate large, tsunami-producing earthquakes.

Seafloor absolute-pressure gauges lashed to the deck of the R/V Roger Revelle. It is important to strap down the instruments, especially during high seas such as those sometimes experienced around the coast of New Zealand.

It also recorded, for the first time, details of the movement of the seafloor during a slow-motion earthquake in September-October 2014 and revealed how shallow these quakes can occur, to within possibly less than 2km of the sea bed.

It is hard to overestimate the importance of this research to life and livelihood in New Zealand. Last year, scientists found proof that central New Zealand could be counting down to a highly damaging 'megathrust' earthquake, generated in the southern part of the Hikurangi Trench, offshore Wairarapa, Wellington and Marlborough.

Those researchers, from GNS Science, Geomarine Research in Auckland and the University of Texas, calculated that in the past 1000 years two subduction quakes of at least magnitude 7 or 8 occurred - one between about 880 and 800 years ago and the other between 520 and 470 years ago.

Earlier research suggested a magnitude 8.9 megathrust subduction quake could cause more than 3000 deaths and 7000 injuries in the Wellington region and cost around $13 billion.

Project leader Dr Laura Wallace, a research scientist at the Institute for Geophysics at the University of Texas at Austin, said finding a relationship like this between slow and normal earthquakes might offer a chance for 'better forecasting' of major plate-boundary quakes and tsunami in the future.

She said in a few cases, SSEs had been shown as possible precursors for giant, 'megathrust' quakes. Slow-motion earthquakes appeared to have occurred on the undersea plate boundary southeast of Japan a couple of months before the March 2011, magnitude 9.0 Tohoku-Oki quake which generated a deadly and highly destructive tsunami.

'This experiment offshore Gisborne is the first time we have been able to measure detailed centimetre-level movement of the seafloor. Our results showed that slow-slip occurred on the same part of the plate boundary offshore Gisborne that has generated large earthquakes and tsunami in the past.'

In a slow-slip earthquake, the Earth's crust moves a few to tens of centimetres over weeks rather than rupturing suddenly in seconds as it does in a normal quake. SSEs have been detected by GeoNet GPS equipment along the Gisborne and Hawke's Bay coasts and also in the Kapiti and Manawatu regions.

The latest work near Gisborne involved more than 20 researchers from GNS Science, the University of Texas, Columbia University, the University of California - Santa Cruz, the University of Colorado - Boulder, Kyoto University, Tohoku University and the University of Tokyo.

Wallace estimated the cost of the project, including ship time, was close to $6 million, with funding from the US, Japan and New Zealand.

Seismometers and absolute-pressure gauges from US and Japanese universities were deployed up to 80km offshore of Gisborne in 2014 and left for a year to record earthquakes and vertical motion at different depths and locations around the plate boundary.

Wallace said the experiment 'worked much better than we expected'.

'The results are incredibly exciting. We found between 1.5 and 5.5cm of uplift of the seafloor offshore Gisborne during the September-October 2014 slow-slip event.

'The results show the slow-slip events off Gisborne occur at very shallow depths, to within 2km of the seafloor and it is possible that the slip events go all the way [up] to the seafloor. The Gisborne slow-slip events are the shallowest well-documented slow-slip events on Earth.

'One of the really important aspects about the plate boundary offshore Gisborne is that much of the plate motion appears to be accommodated by slow-slip earthquakes. That's a big contrast with the southern Hikurangi margin, beneath the Wairarapa and Wellington, where the plate boundary seems to be locked and building to a big earthquake in the future.'

Wallace said it would be wrong to assume slow-slip quakes relieved crustal stresses and made big quakes a more remote possibility. The magnitude 7.2 quake offshore of Gisborne in 1947 and its tsunami showed the area was able to generate large events too, although it was impossible to know if it was preceded by a slow-slip quake.

The 1947 quake might have been what scientists called a 'tsunami earthquake'.

'These tend to cause larger tsunami than you would expect given their magnitude. So with a magnitude 7.2, you wouldn't expect a really big tsunami, but this actually caused an tsunami. These tsunami quakes tend to have longer period motions, so people onshore won't feel as much strong shaking, making it difficult for them to know that a large quake and possible tsunami have occurred,' she said.

Another major marine project is planned for 2018 when the International Ocean Discovery Program will drill two holes into the sea bed off Gisborne and place sensors inside them.

Continuous monitoring of slow and normal quakes there would help scientists firm up any correlations and theories linking SSEs, megathrust quakes and tsunami. Extending the GeoNet network offshore would also help with that, Wallace said.

Forecasting the 'great' quakes - defined as magnitude 8 and above - remains the ultimate aim of many seismologists.

GNS Science principal scientist and geophysicist Dr Stuart Henrys said 'traditional efforts' to forecast quakes were based on a model in which subduction faults slowly and steadily accumulated stress from tectonic forces until they failed in a large 'felt' earthquake, possibly a megathrust quake.

'We now know that faults don't behave that way and non-traditional earthquakes (like slow slip) occur throughout the earthquake cycle, leading to more complex patterns of stress accumulation.

'Understanding and monitoring the time changes of the system of slow-slip may provide a valuable tool to forecast future damaging earthquakes. This not only applies to offshore Gisborne and the East Coast, North Island, region but also globally.'

Wallace said there was a need to better understand the link between slow-slip events and damaging quakes.

'We've seen a few large quakes preceded by slow-slip events. The [July 2013] Cook Strait sequence occurred at the time of a slow-slip event beneath the Kapiti region, so there may be some relationship.

'We need to better understand the relationship between slow-slip events and large, damaging quakes. Once we can fully understand the influence of slow-slip events on seismicity rates, that offers a chance for better quake forecasting.'