Sayonara, Japan? The Chances For A Big Eruption In The Next 100 Years

Two Kobe University volcanologists say (PDF) that Japan faces a roughly 1% chance of a devastating eruption on or near southwestern Kyushu in the next 100 years. Yoshiyuki Tatsumi and Keiko Suzuki-Katama modeled their research on a magnitude 8.4 eruption that formed the 12.4-mile-wide Aira Caldera on the Japanese island of Kyushu some 28,000 years ago.

A simple diagram of subduction, by Mike Norton.

Simple subduction, imaged by Mike Norton.

About a month ago, we looked at news of this research. Last week I posted some guest videos about predicting volcanic eruptions, including an interesting Foreign Press Center of Japan briefing on eruption prediction.

So, before we get started…why are there volcanoes in Japan anyway?

Japanese volcanoes and the Ryukyu Arc

In essence, the islands of Japan have formed where the Pacific seafloor is sinking down into the Earth, but it isn’t possible to look at the area in terms of simple subduction.

Five of Earth’s tectonic plates are interacting here – the Eurasian, Amur, Okhotsk, Pacific, and Philippine Sea plates – which makes things complicated.

Instead of asking why there’s no Japan Plate, let’s instead look at all the arcs of volcanoes that this geological plate jamming has made, in particular, the Ryukyu Volcanic Arc.

The Ryukyu Arc is down there in the lower left.  (Image source)

The Ryukyu Arc is down there in the lower left. It actually starts on land (southern Kyushu) and goes about 250 more miles southwest under the sea. (Oregon State)

Subduction zone volcanoes come in arcs because they run parallel to the nearby subduction trench. Technically, they form a a volcanic front on the surface that marks the border between the active trench zone and the nonvolcanic fore-arc area.

That’s as complicated as I want to get. For what we want to look at today, it’s enough to know that Aira Caldera is located on the Ryukyu Arc. So are a lot of other volcanoes that have had huge eruptions. What’s so special about that arc?

Many parts of Japan are deforming rapidly, geologically speaking, because of the complex tectonic setting. The Ryukyu Arc is one of the rare parts of Japan where the rate of strain is much slower. It’s possible that this slower deformation allows much more magma than usual to accumulate underground.

I have read some online sources that say that the area around the Ryukyu Arc is actually stretching. I don’t know and so will just stick with Tatsumi and Suzuki-Katama, the two Kobe University experts, who call it a lower rate of strain.

Space aliens?  (Source)

Space aliens? (Source)


Of note, they don’t use the word “supereruptions,” though I will. In a formal scientific paper they have to say something like “catastrophic caldera-forming eruptions” instead. Aira and Kikai, another Ryukyu Arc volcano, have had such eruptions – in the range of M7 or higher.

That “M” isn’t always the same thing as Volcanic Explosivity Index. It measures the total mass of material a volcano has erupted, while VEI measures the eruption’s severity. Close enough, though – we’re talking about eruptions that would be a VEI 7 or 8.

Yeah, such an event is rough on the people living nearby. But were there people around during those prehistoric supereruptions?

Yes, there were.

The Jomon Culture

The earliest evidence for humans in what is now Japan dates back almost 35,000 years. Little is known about this culture, other than that they used stone tools. They were there for Aira’s M8.4 eruption some 28,000 years ago.

While there is no evidence that this supereruption changed their tool-making style, it might have led to the rise of two distinct cultures in Japan, one in the northeast and another in the southwest. This duality is still present in Japanese culture today.

In any event, by 4000 BC, a more advanced culture – the Jomon people – lived on Kyushu and elsewhere in the Japanese archipelago. There were an estimated 20,000 of them altogether.

On southern Kyushu, the Jomon decorated their pottery using shells and roller patterns. Then one day, in approximately 4350 BC, the sea exploded. It was like this Auckland Museum simulation of a volcanic eruption in the harbor there, only supersized:


Kikai, one of the Ryukyu Arc volcanoes, a little over 60 miles south of Aira and offshore, had an M7 eruption. Today’s geologists call it the Akahoya eruption. It blew some 36 square miles of material straight up into the air, leaving behind a 12-mile-wide caldera. Massive pyroclastic flows ran out over the sea and covered much of southern Kyushu. Ash fell as far north as Hokkaido, covering basically all of Japan. The first tsunami waves (PDF), up to 100 feet high, began pounding Kyushu within hours.

This eruption was not as big as Aira’s M8.4, but the human beings it affected were more advanced and therefore more vulnerable than the neolithic people who had seen Aira’s wrath. After Kikai, there was no more shell and roller decorated pottery on Kyushu. Today we don’t know if the eruption wiped out all of the Jomon on southern Kyushu or if the survivors just moved out of the area.

The modern era

Eventually the vegetation came back and people returned to Kyushu. What is now the nation of Japan took shape. It was fortunate enough never to have to deal with eruptions like those at Kikai and Aira as the centuries passed, though of course there were plenty of “normal”-sized and sometimes devastating volcanic eruptions.

In the 20th century, two seemingly very different scientific disciplines – volcanology and higher mathematics – developed. They would eventually lead to, among other things, the 2014 paper on volcanic risk that we are interested in today.

Mount Aso is another Ryukyu Arc sleeping giant.  These are the volcanic cones in the center of its caldera.  (Image: Miya.m)

Mount Aso is another Ryukyu Arc sleeping giant. These are some of the volcanic cones in the center of its caldera. (Image: Miya.m)

In terms of volcanoes, Japanese authorities established a volcano observatory on Mount Asama in central Honshu, near Tokyo, in 1911; it only monitored earthquakes. Seventeen years later, Kyoto University researchers began systematic studies on Aso caldera, on Kyushu and near Aira. A Tokyo University volcanologist began collecting geophysical data at Mount Asama in 1933.

During World War II, Usuzan Volcano on Hokkaido formed a new dacite dome, which was named Showa-Shinzan. Studies of this process “formed the foundations of volcanic activity observational studies in Japan,” according to the Japan Meteorological Agency’s Toshitsugu Fujii.

From 1959 on, more volcanological observatories were established. In the 1970s, the first national plan for the prediction of volcanic eruptions appeared as activity increased at Sakurajima – a cone in Aira Caldera and one of the most active volcanoes in the world. I think Japan’s scientists and emergency planners are working now under the seventh of these national plans.



And what of mathematics? Well, the focus we are interested in involves extremes. Attempts to deal with extremes by using numbers began in the 18th century, when Bernoulli looked at ways to predict the age of the oldest survivor in a group. From the 1920s through the 1940s, theoretical math explored largest order statistics (r-LOS). People started getting practical with the mathematics of the very big and the very small (formally known as extreme value theory [PDF]) in engineering applications in the late 1950s, but the use of extreme value statistics really didn’t become common until the 1980s. EVT proved its worth first in environmental applications, as well as reliability testing. Today it’s used in finance, insurance, telecommunications, sports, microarrays, and – yes – volcanoes.

Now we are ready to look at the implications of the paper by Tatsumi and Suzuki-Katama, who stand on the shoulders of all those earth scientists and mathematicians.

Japan’s risk of a supereruption

They say that Japan faces a roughly 1% chance of a supereruption in the next hundred years. What does that mean for scientists, for planners, for laymen like you and me?

Back around 4350 BC Kikai had what has turned out to be the last M7 eruption in Japan and one of just a handful of known M7 eruptions in the last 12,000 years. Aira’s M8.4 happened way back, some 28,000 years ago.

Depending on your optimism-pessimism balance, these supereruptions could seem too ancient to worry about or their ages might make you wonder if either volcano is overdue for another big one.

Well, we don’t know one way or the other.

Volcanoes operate on Earth’s time scale, not our own. On the human time scale, there hasn’t been a supereruption in all of recorded history. Geologists study these megacaldera eruptions only through the mess they leave behind – ash deposits, the remains of pyroclastic flows, tsunami deposits, and so forth. Thus, everything they say about such events always comes with the understanding that it’s the best interpretation of field evidence to date, but nobody was there to observe it as it happened, so the story might change as new evidence appears.

And nobody has any idea of what the precursors are for a supereruption, either. It would be nice if they were big and started showing up years before the actual blast, but for all we know, such a huge amount of magma might just suddenly blow.


It’s also true, that on the geologic time scale, the catastrophic eruptions at Kikai and Aira happened just yesterday and are likely to happen again. Kikai’s eruption was the smaller of the two, yet it appears to have had a stronger effect on humans living nearby, because those humans had a more advanced society.

What about the tens of millions of people who live in an ultra-technological state there now? What if scientists could tell that another M8.4 eruption was coming – could they evacuate all of Japan?

Well, just talking about it can make you crazy. Yoshiyuki Tatsumi and Keiko Suzuki-Katama have tried to get a handle on it, quantifying the risk so everyone can discuss it and start making some rational plans.

The two volcanologists used the Aira Caldera M8.4 eruption because it has been well studied. They found that if it happened again, and the worst happened, Japan would face this:

“cm” means how much ash would fall, “million” refers to the total number of people affected in the zone of each dotted line.

– “ignimbrite” means pyroclastic flow, and yes, it does cover most of northern Kyushu

– “cm” means the average amount of ashfall

– “million” refers to the total number of people affected in the zone of each dotted line.


Of note, Mr. Toshitsugu Fujii, in that FPCJ press briefing video we saw last week, said that 10 cm of ash is enough to kill plants.

When the green world dies, it will take a very long time before people can live there again.

So, is it going to happen? Tatsumi and Suzuki-Katama used extreme value theory to show that there is a roughly 99% chance that this very powerful and yet very rare event won’t happen in the next 100 years. However, given the consequences, that 1% chance is enough to give us pause.

What to do now?

Mr. Fujii also described the current state of Japan’s volcano monitoring and prediction efforts (you can also check them out here [PDF]).

Because Tatsumi and Suzuki-Katama focused on Japan, I am ignoring global effects in this post. Just on a society level, can any country – not just limited to Japan (Aira or Kikai, perhaps) or the US (Yellowstone or possibly Long Valley) – cope with a supereruption?

What is the best response to such a catastrophic event – a suspension of civil society for an indefinite period? Rather than running the risk of having an extended period of martial law turn into permanent authoritarian rule (which is possible anywhere in the world, including the United States and other democracies), should we instead accept a high casualty rate but maintain social resilience by relying on the overall effectiveness of millions of individual decisions?

Which is more important: saving lives in the short run or saving civil society in the long run?

I don’t wish to be supine in the face of one of the worst challenges our planet can offer, but seriously, what can we do ahead of time if we haven’t yet decided what sort of world we wish to build after the apocalypse?

Practically speaking, humanity isn’t likely to do anything at all until it happens – whenever and wherever that may be. Afterwards, we will do what the Jomon people did: collect the survivors and rebuild our lives and our culture with whatever we’ve managed to keep intact after the event.

We will have also collected scientific data on what happened just before, during, and after the first supereruption in human history. Yet will it have meaning in a post-supereruption world where the next such event might not happen for tens of thousands of years?

Well, now we’re getting into the question of what science is worth in human society, and I don’t want to go there. We are what we are, and we will have science and society and all of the rest of the things that make us human beings. A supereruption will interrupt all that, but it won’t be the end for us. As a species, we will undoubtedly reel from the local and global effects, but afterwards we will carry on as close to normal as we can – just like the Jomon did.

Unlike the Jomon, though, we are aware of the risk. So, what should we be doing now?

The past is the key to the present

Let us look to the past. Very few detailed studies have been done on the social impacts caused by prehistoric caldera eruptions. Not a lot of evidence has come down to us over the millennia, but some has and Japanese tephrochronologists and other specialists are studying it.

It’s a good start. We are likely to succeed sooner in piecing together our own human history with supereruptions than in trying to see through miles of the Earth’s crust in attempts to accurately predict the next catastrophic event (though this should also be pursued intensively). Knowing what people did when they faced such a terrible event tells us what might work for us when the need arises to build a post-supereruption human world.

The only control we can realistically exercise in all this is over ourselves. Tatsumi and Suzuki-Katama have opened the way for everyone – scientist, emergency planner, politician, and layman – to start viewing the epitome of disaster in rational terms. I congratulate them.

Now then…what’s our next move?



Categories: Sunday morning volcano

Tags: , , , , ,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: