Dr. Robert Geller shakes up assumptions about earthquake prediction
For over 40 years, the Japanese government has issued seismic hazard maps that indicate which parts of the country are most in danger of a serious earthquake. Currently, the area of greatest interest to forecasters is the Nankai Trough region, a fault zone off the east coast of Japan running along the shores of Shikoku and up to Wakayama Prefecture. According to official forecasts, there is an 80% chance that a major earthquake will strike the region within the next 30 years, a probability that has shaped disaster response considerations.
Some researchers, however, have been critical of the government’s forecast methods. Dr. Robert Geller, professor emeritus of the University of Tokyo in the field of seismology, has published numerous works laying out his position that neither short-term predictions nor long-term forecasts are feasible. In fact, he argues, such predictions may end up being harmful if regions forecasted to have a low earthquake risk are lulled into a false sense of security.
Speaking at the Foreign Correspondents’ Club of Japan, Dr. Robert Geller bluntly titles his talk “The ‘Imminent’ Nankai Trough Mega Quake: Myth or Reality?” His opening is equally blunt: “It’s a myth.”
The Japanese government’s hazard maps go back to the 1970’s, when a scientific report announced a high likelihood of a powerful earthquake striking the Tokai region between Nagoya and Tokyo. In response, the government passed the Large-Scale Earthquake Countermeasures Act in 1978, which mandated regular forecasts of impending earthquakes. In the four decades since, no major earthquake has struck the Tokai region.
Dr. Geller’s contention is that this law was passed before the Tokai quake prediction could be tested and verified, and that subsequent events and research have cast strong doubts on the usefulness of forecasts. “Short-term prediction of imminent earthquakes is impossible, because they are what’s known in physics as a chaos-type process: any small variation in the parameters completely screw up your ability to predict what’s going on,” Dr. Geller explains. “And long-term prediction is impossible because quakes don’t repeat in cycles, there is no statistically significant evidence for cycles.
“We know only a tiny fraction of all that has happened over geological history. The idea that the ‘same earthquake’ repeats in a cycle is complete nonsense. There are zillions of faults buried under the Tokyo area, and the next earthquake to hit Tokyo will probably be on a fault that was a surprise.”
In particular, he points out that the major earthquakes of the past few decades have occurred in areas labeled ‘low-risk’, while those in ‘high-risk’ areas have not had unusual levels of activity. “Looking at the map issued in 2010, the 2011 Tohoku earthquake occurred in a region that was very safe, according to the government,” he says. “And then in 2016 the Kumamoto earthquake also occurred in a place that was supposedly very safe. And last September, an earthquake in Hokkaido, in another ‘safe’ place. And meanwhile the ‘dangerous’ places have all been quiet.”
On the limitations of hazard maps, Dr. Geller says, “if we take larger areas, we can make more reliable probabilistic statements. For example, if we take all of Western Europe, we can say it is much safer than Japan. But as we try to make our area smaller, the forecasts become completely unreliable.
“If you ask me, ‘Can you prove that earthquake prediction will never become possible?’, the answer is no. Science can’t do that. But people have tried it for 130 years, and the more they have tried, the worse and worse the prospects have gotten.”
The problem this presents, according to Dr. Geller, is that low-risk assessments may result in a false sense of security for residents and local officials who end up foregoing preparation measures or making decisions that aren’t grounded in facts. In the years leading up to the earthquake in Kumamoto, for example, local governments in Kyushu had cited hazard maps in their promotions of the region as a safe place to establish businesses because of its supposed low earthquake risk.
Even more ominously, he notes, “if TEPCO seriously thought there might be a magnitude 9 earthquake and tsunami at Fukushima, they would have fixed some of the things they did. They would have put the emergency power generators on top of a hill, instead of in the basements, where they were flooded. If they had had emergency power, then there wouldn’t have been an explosion. I think that’s corollary damage from warning about Nankai all the time.”
As for what can be done instead, Dr. Geller ends with a sobering conclusion: “There have been many quakes all along [Japan] in the past, and will be more in the future, but no one can say from where to where. Present science cannot say what part of Japan is more or less dangerous than other similar parts. I know everyone would like that to be possible, but unfortunately, it is not.”