By Richard Black
Environment correspondent, BBC News, Japan
Prof Tobin is coordinating the search for precursors of quakes
Some rules have no exceptions, and this is one: whenever there is a big earthquake, and especially if lives are lost, some journalist asks, "when will we be able to predict them?"
It is a question that many earthquake scientists shy away from, and with good reason.
Say it might be possible, and you risk being held to making it possible; wail about how hard it is, and you may see a gap develop in your funding.
To Harold Tobin, it is an understandable question and an important question, but its time has not yet come.
"I think we're at a point in our science now when we have to ask a question that comes before that one," he says.
And that question is: "Are there in fact any precursory signals for earthquakes or not? We don't know the answer."
More reliably than pride going before a fall or calm coming before a storm, there has to be something, or some set of things, that indicates a high likelihood of a quake coming - otherwise, prediction cannot be.
Down the years and across the planet, many potential precursors have been suggested and studied, from physical phenomena such as changes in local magnetism to the movements of animals and birds.
But none has shown anywhere near enough reliability to be thought of as a serious contender.
Professor Tobin on predicting the 'big one'
As co-chief project scientist of the Nankai Trough Seismogenic Zone Experiment (NanTroSEIZE), Prof Tobin is well placed to co-ordinate a search for such precursors.
The Nankai Trough off the eastern coast of Japan produces big earthquakes - often accompanied by a tsunami - about once every 100 years. The 1,400-year written record maintained by coastal communities means it is already one of the most documented and studied of the Earth's seismogenic subduction zones.
As the six years of its operational life progress, NanTroSEIZE will deploy instruments deeper into a seismogenic - earthquake-generating - region than ever before, down to 6km (4 miles) below the Earth's surface.
Seismic surveys of the Nankai Trough have mapped strata in close detail
The Japanese government, the major funder for the project if you take into account the fact that it built and supplies the state-of-the-art research ship Chikyu, is deploying a network of sensors on the sea floor around the Nankai Trough that will yield yet more data on what is happening in the zone.
If there is a precursor signal or set of signals, NanTroSEIZE is well placed to find it.
The expedition's initial years have been a combination of drilling for real results and drilling to prove technology, notably the riser system that is the key to probing deeper than a couple of kilometres.
But Masa Kinoshita, NanTroSEIZE's other co-chief project scientist who is based at the Japan Agency for Marine-Earth Science and Technology (Jamstec), says there have already been some surprising findings from the first, relatively shallow holes.
"We drilled through two active faults, although they're too shallow to cause seismic waves," he relates.
"We found the fault zone was quite narrow, quite focused, like 2mm or something like that. It's good to know something new anyway."
Rock cores extracted from the region also showed that some of the ground was under stress in directions that were unanticipated.
THE NANTROSEIZE PROJECT
Project will extract rock cores from the Nankai Trough, an area that has regularly generated major earthquakes and tsunamis
Scientists will be able to analyse the seismogenic, or earthquake-producing, region, where one of the Earth's tectonic plates slides under another, in unprecedented detail
They hope to learn what triggers earthquakes here and in other areas that are geologically similar, including the Sumatran fault that caused the devastating 2004 tsunami - potentially leading to ways of predicting them
Projects such as NanTroSEIZE often use the term "ground-truthing" as their rationale - and in this case, it may literally be true.
Events in the oceanic subduction zones that cause the world's great earthquakes are modelled with computer programs - but unless you have accurate data from the fault zone to put into the model, what comes out is likely to be well, you probably know the old saying about computers - "Rubbish in, rubbish out" is the polite version.
"Without this information at the fault - to know the properties or strength and the stress - we cannot have important information about predictions," says Dr Kinoshita.
The rock samples taken from NanTroSEIZE holes are examined to give information on parameters such as composition and density, which again can be plugged into computer models.
The zone has already been intensively surveyed with sonar, with scientific ships pulling huge arrays of hydrophones - underwater microphones - that produce intricate images of rock layers under surface.
They give scientists a picture of what the rock might be like - but without drilling, you never know for sure.
In science, as in other walks of life, an ever-present danger is taking what we know to be true in one situation and assuming it reflects the generality.
Computer images give an idea of the formation of undersea rocks
So NanTroSEIZE alone cannot answer the general question of how to predict earthquakes - or, if you prefer Harold Tobin's version, of whether they are predictable at all.
Lisa McNeill from Southampton University in the UK, who will be taking a leading role in a NanTroSEIZE expedition later this year, has also worked on the Sunda Trench near Sumatra that ruptured with such devastating force in December 2004.
"The two zones are very similar in that in both situations you've got two tectonic plates converging towards each other," she says, "and these are the situations where you generate very large earthquakes, as we saw in Sumatra and the Indian Ocean.
"In Sumatra, we've got areas that have literally just released that strain, while in the Nankai Trough that happened in the 1940s.
"In both cases, we're probably looking at repeat times of hundreds of years; so they are at slightly different stages after the earthquake, and that would be an interesting comparison."
Deep drilling in the Sunda Trench presents more logistical problems than the Nankai Trough. But another intriguing comparison may come from a hole that was recently completed into the San Andreas Fault in California - not a subduction zone, but a fault zone nevertheless.
Currently, though, some of the ideas are running beyond what technology can deliver.
"We expect the temperature at the [Nankai Trough] fault is as high as 150C or 170C," says Dr Kinoshita.
"We don't know any electronics that can survive, and there is no available instrument at this moment to make this crustal deformation measurement possible; so we are now developing simpler but more robust instrument."
There is another issue beyond what triggers a subduction zone earthquake - namely, what determines whether a quake triggers a tsunami?
It, too, is a vital question, given the catastrophic damage that tsunami can wreak.
Answers may lie in the accretionary prism - a structure on top of the sea-floor rock, as that is what is actually in contact with the bottom of the ocean.
"An accretionary prism is simply the deep sea sediments that were deposited on top of the ocean floor and then rafted towards the subduction zone by plate tectonics," explains Harold Tobin.
"As that happens, the plate that's on top - in this case the Japanese islands - simply scrapes that layer of sediment and rock off like it was the blade of a bulldozer.
"The reason it's important is that the uplift of the sea floor, which causes the tsunami, is actually determined by the shape and geometry of the faults that run through the accretionary prism."
Sonar scans of the Nankai Trough show the accretionary prism is riddled with strata and slippages, indicating an earlier history of traumatic upheavals; but what that means for generation of tsunami here is another matter.
From a simple question - can we predict earthquakes? - a host of other questions branches off, like the multifarious channels of a megasplay fault.
The next four years of NanTroSEIZE will follow some of those trails and see where they lead - but how close it takes us to the "big one" is, for the moment, impossible to predict.
This page is best viewed in an up-to-date web browser with style sheets (CSS) enabled. While you will be able to view the content of this page in your current browser, you will not be able to get the full visual experience. Please consider upgrading your browser software or enabling style sheets (CSS) if you are able to do so.