Earthquakes Cannot be Predicted / Official Japanese Quake Forecasting System is Flawed

If you’ve been watching the news the last few days, chances are you’ve heard about that researchers are now saying that Tokyo is facing a 70% chance of a 7.0 magnitude earthquake in the next four years:

The preliminary calculations conducted by a team from the university’s Earthquake Research Institute were based on intensified seismic activity in the area after the Great East Japan Earthquake on March 11.

The findings are more dire than a similar estimate by the central government’s Headquarters for Earthquake Research Promotion, which states there is a 70 percent chance a quake of the same scale will hit the southern Kanto region, including the metropolitan area, within 30 years.

Sounds pretty scary. Should we believe the forecast? Not necessarily.

In April of 2011, Nature published “Shake-up time for Japanese seismology” by Professor Robert J. Geller of Tokyo University (Japanese version available here). Geller argues that earthquakes cannot be reliably predicted.

When one compares the Japanese government’s national seismic hazard map to the actual areas where large earthquakes have occurred in the last 30 years, one notices the problem:

“Although such maps may seem authoritative, a model is just a model until the methods used to produce it have been verified. The regions assessed as most dangerous are the zones of three hypothetical ‘scenario earthquakes’ (Tokai, Tonankai and Nankai; see map). However, since 1979, earthquakes that caused 10 or more fatalities in Japan actually occurred in places assigned a relatively low probability. This discrepancy — the latest in a string of negative results for the characteristic earthquake model and its cousin, the seismic-gap model — strongly suggests that the hazard map and the methods used to produce it are flawed and should be discarded.”

Very general predictions are possible:

“If global seismicity and the historical record in Tohoku had been used as the basis for estimating seismic hazards, the 11 March Tohoku earthquake could easily have been ‘foreseen’ in a general way, although not of course its particular time, epicentre or magnitude.”

But the Japanese government continues to rely on methods that attempt – and apparently fail – to predict earthquakes of specific magnitudes taking place in specific regions.

“All of Japan is at risk from earthquakes, and the present state of seismological science does not allow us to reliably differentiate the risk level in particular geographic areas. We should instead tell the public and the government to ‘prepare for the unexpected’ and do our best to communicate both what we know and what we do not. “

Much of Geller’s article focuses on the Japanese government’s official Tokai earthquake prediction. However, the general message is clear: there is no reliable method for forecasting earthquakes.

The latest forecast for Tokyo does not appear to represent a sudden new breakthrough in the field of earthquake predictions. What’s new about it is that heavy emphasis is placed on the increase in seismic activity that occurred in the 6 months following the March 11th disaster. [For a detailed analysis of the forecast - see this post. ]

Some of the articles about the new Tokyo prediction have presented it alongside the highly questionable Tokai earthquake prediction. Both predictions are based on methods that have not been verified as accurate.

Don’t trust the hype. Big earthquakes do happen in Japan, but experts do not know where and when the next big one will occur. If you live in Japan, be prepared for earthquakes, even if you’re outside of what have been deemed high risk areas.

[gigantic hat tip to Steve]

—

Akihabara News – Gadgetry from Japan (Subscribe)
Dannychoo.com – Your portal to Japan (Subscribe)

Ap

Recordings show that in the Kanto area every huge earthquake hits on a somewhat regular basis
The last one was in 1923 (before that in 1855 and 1703, those are the ones I know of), so there is a new one to be expected.
There were also three other quakes with magnitude of 5 or higher since the Tokugawa area started.

http://www.japanprobe.com James

And the Kanto area will no doubt be hit by a bit earthquake some time in the future too.
But “70% chance within 4 years” isn’t the same as a general prediction.

http://www.jlist.com Peter Payne

Good article, am tweeting
Anonymous

It’s a given that earthquakes will happen in Japan, and that they will happen fairly regularly. But I don’t see any reason why the govt publishes earthquake forecasts. I think it’s does more harm than good, putting people at risk for thinking they’ll be safe, and at the same time opening themselves up to blame for not forecasting earthquakes that happen in “low risk” areas which end up causing major damage i.e. 1995 Hanshin Earthquake.
Anonymous

The fact that these models haven’t proven good for predicting earthquakes is well known and accepted in Japan. The next “big one” has long been forecast as being centered on Shizuoka south of Tokyo, and they have been having regular tsunami and earthquake drills for the expected quake there for the last 10 years now. PM Kan even cited the “94%” probability of there being a quake centered on the Hamaoka nuclear plant in Shizuoka in the next 10 years as a justification for shutting it down – a prediction made based on similar modelling.

Of course, no one ever even talked about a Tohoku earthquake, and indeed, nobody even realized there was a faultline beneath Kobe in 1995 (or in Christchurch in 2010/11).

However, fact remains that around four fifths of Japan’s population and industry are located in areas that regularly experience quakes, and are understood to face a near 100% certainty of large quakes and tsunamis occurring sometime. The central government needs to figure out where to spend money as a priority for earthquake and tsunami prevention measures, and businesses and insurance companies in Japan of course face ruinous costs from large quakes, and so the insurance industry, government research bodies and university labs are dedicating almost unlimited funds into developing models that can try in some way to at least give a broad risk assessment of the likelihood of quakes for certain regions for policy purposes, and developing short term predictive technologies (such as alpha wave detectors, which are now used nationwide in Japan, but only are relevant in about half of quakes).

Whenever these figures are released, the academics releasing them are often quizzed on why these models failed to predict nearly all the biggest quakes of the last few decades. The basic answer I think is that they know, and keep trying to build in more statistical and monitoring data, and to improve research into quakes (and nowhere even spends a fraction of the resources Japan does on doing this) and they are always doing their best to improve a science that is somewhat akin to herding cats.

Point is, for all the shortcomings of models used being well known, policy makers, insurers, and the public still have no choice but to use the best information available. Fact is that for the last 1000 years, a major quake devestating Tokyo once each generation has occurred and so it is a near certainty that it will happen again. Given the disturbing similarities of the 3/11 quake to the Indian Ocean quake/tsunami, and the subsequent quakes in Indonesia further down the faultline, the assessment is that the already highly stressed faultlines south of Tohoku in Tokyo and Shizuoka are now even under more pressure.

I agree, the track record of these models isn’t great, and attaching a number like “94%” or whatever that they did with Shizuoka seems borderline ridiculous. But in a country like Japan, it isn’t really a basis for discarding the use of models altogether. The country needs to constantly improve disaster readiness, and only has limited funds to allocate for preparedness at any given time, so useless or not, Japan doesn’t really have any option to go on using the best available models until they can be improved, or something better comes along.
Ted

If I recall correctly, this Tokai earthquake prediction was the main justification for shutting down the nuclear reactor in Shizuoka last year. I wouldn’t go so far as to call it a conspiracy, but perhaps the govt. has a vested interest in focusing on the danger to the Tokai region, rather than coming out and saying that the entire country is a seismological time bomb (and therefore having to explain why the other reactors aren’t being shut down)? Just a thought…

Earthquakes Cannot be Predicted / Official Japanese Quake Forecasting System is Flawed

Recent Comments

Further Reading

Bikini Atoll: Lessons For Fukushima?

Complaints From Loony Parents Cancel Event: Okinawan Children Not Allowed to Play With “Radioactive” Aomori Snow

Remembering the International Media’s Post-Fukushima Hysteria: Radiation Scare Stories Ignored Real Victims of 3/11

A Look Inside Osaka Prison

Dead Shark Found in Yoyogi Park

Japanese Fart Battle Scroll

Dinosaurs Wear Traditional Japanese Clothing

Canada Rejects Refugee Application From Japanese Who Feared Radiation

Ishinomaki: Then & Now

Radioactive Seafood? Miyagi Fishing Industry Works to Prove Safety of Their Fish

Categories

Archives

Subscribe

Japan Blogs

Korea Blogs

Linkroll

Other Asia Blogs