DESLEY BLANCH : It was the monster earthquake which struck north-eastern Japan in 2011 that spurred one geophysicist into action to create a much more accurate map of the world’s earthquake hotspots.
PROFESSOR DIETMAR MULLER : It was that earthquake that kicked us into action, because the weekend when that earthquake happened, I thought – well, on Monday some journalists are going to call me up and ask me why did that earthquake happen and I didn’t know at that time.
DESLEY BLANCH : Meet Dietmar Muller who is lead researcher and Professor of Geophysics at the University of Sydney.
The new map pinpoints the regions on the seafloor which are at higher risk of generating powerful earthquakes and by which it’s hoped will help reduce earthquake damage and loss of life.
PROFESSOR DIETMAR MULLER : So I started looking at some maps and I noticed that very close to the region where this earthquake happened, a huge scar on the ocean floor, in fact a double scar called the Kashima fracture zone was located and I thought -- wait a minute, that must have something to do with it.
DESLEY BLANCH : Well, can we have a first quick lesson in where and how earthquakes originate, which seem to be in zones where one techtonic plate slips under the other. Is that correct?
PROFESSOR DIETMAR MULLER : That’s right and these are the earth’s subduction zones where oceanic techtonic plates slip underneath so-called overriding plates, and they are mostly continental. They can sometimes also slip under other pieces of oceanic plates and the friction at this interface is what causes a lot of great earthquakes.
DESLEY BLANCH : And referring back to that Japanese earthquake your research has found zones that are at much higher risk. So what did you discover?
PROFESSOR DIETMAR MULLER : Well, we found that those regions where large scars on the ocean floor called fracture zones meet subduction zones, that these areas seem to be at much higher risk. The reason for this seems to be that there are very large submarine ridges associated with these scars. They can be as tall as two to three kilometres. So these are the areas that our data analysis identified as particularly at risk for great earthquakes.
DESLEY BLANCH : Now your evidence is statistical and it comes from data that’s been kept on 1500 earthquakes and your figures showed that most of the bigger magnitude earthquakes came from these intersection regions. So do you yet understand why these meeting points become hot spots?
PROFESSOR DIETMAR MULLER : The actual physics of earthquake generation is extremely complex and non-linear and we are still far away from understanding exactly what happens deep in the earth at these regions. But we do have a conceptual model, and the conceptual model is that’s where these very large ridges are being subducted--where they meet subduction zones, that’s where we generate excess friction and this added friction between these ridges that are grinding against the overriding continent, that’s where stress accumulates over a long time, can be hundreds or thousands of years, and when this stress is eventually released, that’s when you generate a great earthquake.
DESLEY BLANCH : Now, you went back over 1500 earthquakes, that’s a lot of analysis and to do this analysis of all this data you used a recently developed data mining method that’s previously only been used to match internet users to consumer goods. I find that intriguing. But did you need to modify the technology in any way to fit your need?
PROFESSOR DIETMAR MULLER : It was actually quite straightforward to adapt that methodology to our task, because if you think about it, these very large data sets of earthquake data they’re not unlike the data that are usually mined on the internet, lots of users are making certain choices, buying different goods, so it was quite straightforward to adapt this methodology to our problem.
DESLEY BLANCH : So what makes your method so much more accurate than past information?
PROFESSOR DIETMAR MULLER : Our method is not really more accurate it’s just that we asked a different question from the questions that had been asked before.
There is actually an interesting disconnect in the earthquake research community. Most seismologists who analyse earthquakes are very much focussed on continental areas which is of course where people live who are affected by these earthquakes.
Now because we have a background in marine geophysics we had a look at the ocean floor that’s going down instead of just what is happening on the continent and, there’s a marine geophysics community – you know people who go out on research vessels and map these submarine features – and that’s a very different community from the seismologists who usually analyse earthquakes.
I suppose we are one of the few groups that straddles these two areas and we were able to connect up these different data sets and ask questions that hadn’t been asked before. Namely – is there globally a connection between these fracture zones – these oceanic fracture zones – these large scars on the ocean floor where they meet subduction zones.
DESLEY BLANCH : And your results were quite interesting, because you were more accurate than anything else. What were some of those figures?
PROFESSOR DIETMAR MULLER : Yes, in a way our results provides additional insights, because most earthquake hazards maps are based on the so-called instrumental earthquake record that starts at about 1900 which is when seismographs were first deployed to measure earthquake data.
Of course, these very largest earthquakes they can have very long recurrence times, so they can be hundreds of thousands of years. And, of course, if there was an earthquake that happened over 1,000 years ago, where the data collected after 1900 won’t tell you anything about that time and so this is precisely what happened in Japan. There had been an earthquake over eleven hundred years ago but the historical records documenting that earthquake had not been used to help generate better hazard maps.
Now our approach picks all these areas where there hasn’t necessarily been a large earthquake in the last 100 years, but by gaining the insight that there is a fundamental connection between places where these very large scars on the seafloor are going down into subduction zones, we can then alert the community to these regions and say, well, perhaps we should look at all these regions and just scrutinise the historical records, not only look at earthquakes from the last 100 years.
DESLEY BLANCH : Well, our Asia and Pacific listeners are living around the Pacific Ring of Fire and around the Sunda Trench off Indonesia, and there’s Sumatra and Japan, as we’ve been talking about, these are all high earthquake activity areas. So what will your map mean for them?
PROFESSOR DIETMAR MULLER : That’s right. Those regions, for example, the Indian Ocean region offshore Sumatra is particularly densely populated by fracture zones on the ocean floor and this is one of the reasons why this area is so earthquake-prone.
Similarly, there are some very large fracture zones associated with subduction around Japan, and also a bit further up north, and along the Eurasian margin towards Kanchatka. And similarly on the other side of the Pacific there are lots of fracture zones intersecting with the margin of Chile and Peru.
DESLEY BLANCH : What this map means for them that it opens up a new avenue to assess long term earthquake risk in these regions, so we are no longer just reliant on the instrumental record from the last 200 years, but we have now identified a whole bunch of bands along the world’s subduction zones that we should fundamentally pay more attention to in terms of scrutinising all data including historical data that is available to us and to try to better assess the long-term earthquake risk.
Of course we cannot say exactly when and where the next great earthquake will occur but we have a pretty good idea which regions are fundamentally more at risk than others, because we found that amongst the 15 largest earthquakes on record, 13 are associated with these special techtonic niche environments, so that’s about 87 per cent.
Now if earthquakes were randomly distributed along all of the world’s subduction zones this number should just be 25 percent, so statistically it’s a very significant association.
DESLEY BLANCH : Professor Dietmar Muller is a geophysicist at the University of Sydney.
EXTERNAL LINKS: EarthByte Group GPlates