A Kiwi-led team of scientists has detonated hundreds of kilograms of dynamite deep below the North Island to reveal the clearest picture yet of what happens beneath tectonic plates.
An important breakthrough, led by New Zealand scientists and published today in the leading journal Nature, has transformed our understanding of the Earth's tectonic plates.
It shows for the first time how they are gliding on a distinct layer of "soft" rock, only 10km thick and weak enough to allow the plates to shift many centimetres per year.
While the idea that the Earth's surface was made up of at least 16 tectonic plates, moving at different rates, is well established, until now it has not been clear what actually shifts them around.
"To work this out requires an understanding of what happens at the bottom of a tectonic plate," said Professor Tim Stern of Victoria University, who worked alongside GNS Science and universities from the US and Japan on the research.
Previously, it had been difficult to get the necessary images at such great depths using the usual method of recording natural earthquake waves.
"But by generating our own seismic waves using higher frequency dynamite shots we were able to see how they became modified as they passed through different layers in the earth," he said.
"This, along with some new techniques in seismic reflection processing, allowed us to obtain the most detailed image yet of an oceanic tectonic plate."
The create their waves, the team used 1200 portable seismographs placed in a
long line between the Wairarapa and Kapiti coasts.
By setting off explosive charges in the ground and recording the echo with seismic instruments, the scientists were able to build a three-dimensional picture of geological layers to a depth of about 100km below the lower North Island.
The initial focus was the top 50km of subsurface to see how the Pacific and Australian plates interact.
However, the exceptional quality of the data enabled the scientists to resolve structures to twice that depth.
The oil industry already used reflection soundings to delineate oil and gas
reservoirs, typically in the depth range of three to 10 km.
The scientists first conducted an experimental survey using 12 large dynamite
blasts drilled 50 m into hard bedrock, and a large array of portable seismographs stretched out across the southern neck of New Zealand's North Island.
The raw data revealed echoes from the base of the subducting Pacific plate 100 km down, that were unexpectedly sharp.
However, industry-standard data processing was not producing a convincing image of the deep slippery layer.
Trying a simple noise-reduction technique from digital photography called Median Filtering, the reflections from the layer then popped out clearly.
They ultimately found that the thinner layer beneath the plate appeared to contain pockets of molten rock that made it easier for the plates to slide on.
"This means that the plates can be pushed and pulled around without strong resistance at the base," said Professor Stern, who worked alongside university colleagues Professor Martha Savage and Drs Simon Lamb and Rupert Sutherland.
"A weak slippery base also explains why tectonic plates can sometimes abruptly change the direction in which they're slipping. It's a bit like a ski sliding on snow."
Understanding this boundary between the base of cold, rigid tectonic plates and the underlying hot, convecting mantle underneath, Professor Stern said, was central to our knowledge of plate tectonics and the very formation and evolution of our planet.
The data enabled the scientists to deduce that the Pacific Plate is 73km thick beneath the lower North Island, thinner than some earlier predictions.
They then identified a distinct 10km thick layer of soft rock beneath the Pacific Plate.
They called this a "decoupling channel", as it isolated some of the movement of the rigid lithosphere of the Pacific Plate from the flowing asthenosphere below.
Such a channel, sandwiched between the lithosphere and asthenosphere, helped to explain some of the elusive nature of this deep part of the tectonic plate system.
"The data makes a compelling case that the material in the channel is rich in either water or molten rock, which provides a slippery base for a moving tectonic plate," project leader Dr Stuart Henrys, of GNS Science, said.
"Our data is not able to distinguish the sandwich filling, but we were able to tell that the base of the lithosphere is a sharp and distinctive boundary.
"This means the plates can move around without much resistance at their base.
It also explains why tectonic plates can sometimes change direction abruptly."
Whether what the scientists discovered beneath New Zealand holds true for all of the Earth's tectonic plates is yet to be seen - and answering the question would require further investigations.
Professor Tim Stern answers some quick questions from the Herald.
Dynamite seems a wonderfully novel approach. How did you use it - and has this been shown to work in previous studies or experiments?
The use of either dynamite or air guns in water is a standard method for seismic exploration used in the oil and gas industry.
But their target depth is usually 3-8 km , whereas we imaged to 100 km depth.
We did this by modifying the experimental design compared to what the oil and gas industry do.
We use massively larger shot sizes - 500 kg of dynamite in the bottom of 50 m deep bore holes. Also, we had a much longer line of listening devices (seismographs).
So, in the oil industry type of work they would typically have 48 or 96 seismographs spread over, say, three km of offset to listen to their shots - we had 877 seismographs spread over a distance of 85 km.
Previous methods to look at the base of a plate at these depths - the LAB or Lithosphere Asthenosphere Boundary - have been based on recording earthquakes waves that originate many hundreds or thousands of kilometres away that then pass through the LAB and up to the surface.
This only produces a low resolution image of the LAB.
The dynamite shots produce much higher frequency energy - 14 Hz - and allow us to explore the LAB in much more detail.
How exactly were the waves monitored and modelled? Did you draw on GeoNet instruments?
We had a massive set of imported, lightweight seismographs that we borrowed from the USA and Japan.
The ability to mobilise these massive sets of detectors to anywhere to in the world made these experiments possible.
Along our ~90 km long line from Glendhu Rocks in the Wairarapa to Queen Elizabeth Park on the Kapiti coast we had 877 instruments but there would be only be about three Geonet stations close enough to the line to be of use.
So we do use them, but they are not critical.
In essence the shot is let off on GPS minute mark and all the recorders are also linked to GPS timing so the arrival time of all the echoes are timed to an accuracy of milliseconds.
Then the data is taken back and processed with computer code that creates images from many superposed echoes... it's just like cat-scanning the human body but on a 100 km length scale rather than 100 mm length scale.
Going into this study, what was your hypothesis? And at what moment did it become clear what you had discovered?
We had no idea we would get as deep as we did. We had planned to just image the plate interface or the top of the plate which is at depth of about 15-30 km beneath the lower North Island. That study was written up and published in 2013.
But we happened to look further down the records and found these deeper echos, or reflections. So you could say this is a serendipitous finding.
How has this work changed our understanding of tectonic plates and why is this significant for future research? Could a clearer picture of the workings of plates allow for better prediction or risk modelling?
The main contribution is to understand what drives plate movement. This study won't have a direct effect in understanding earthquakes, but it is important to note that the whole frame work of plate tectonics revolutionised our understanding of where and when earthquakes take place.
What our study has done has added a little more to the plate tectonic framework, and in that sense this will contribute to our understanding of earthquakes in the broadest sense.
Is there anything to suggest that the plate mechanisms are isolated only to this plate - or is it expected that the very same mechanics - the "slipping" layers you have identified - exist in plates everywhere?
We addressed this in the last paragraph of our paper. We note a similar, but thicker channel seen beneath a much younger portion of the Pacific plate, but we also note that a possible channel exists at the base of a plate just offshore from Norway.
Why was New Zealand an ideal testing ground for this experiment?
The eastern North Island presents a specially favourable location for this study because the oceanic Pacific plate comes and dips beneath the country at such a shallow angle - 12-15 degrees - and it's very shallow at 15-30 km deep.
In comparison, for example, the Pacific plate beneath eastern Honshu in Japan is more like 60-80 km deep and dipping at 40 degrees.
The New Zealand geometry is ideal for a reflection experiment because you can only image shallow dipping or flat surfaces with the method.
The other place we could image the LAB is in the middle of the ocean.
But there, you could not detonate 500kg dynamite shots in the sea without causing some concern from a variety of organisations.
So, yes, I think we have a special land-based laboratory to study the oceanic lithosphere here that is possibly unique.
What has been the international reaction to this breakthrough?
We've had a lot of interest from overseas groups. The fact that Nature is publishing this in their main journal - there are many spin offs from Nature, like Nature Geoscience - shows they see this as a fundamental discovery of broad interest.
They also commissioned a News and Views for it, which is only done for a few of their articles.
Is there any follow-on work from this that you will be pursuing next?
We want to understand how such a channel would form. We use the analogy of a ski on snow and the thin melt layer that facilitates the ski to move in a frictionless manner. The analogy is not exact, but it's useful.