Robert J. Geller: A Seismologist’s View of Nuclear Safety Issues in Japan — Part I


Robert J. Geller, Professor of Geophysics in the Graduate School of Science of the University of Tokyo

Idea Gallery is a recurring series in which guest authors provide their perspective on issues affecting nuclear energy. Today’s guest author is Robert J. Geller, Professor of Geophysics in the Graduate School of Science of the University of Tokyo.

>>Read Part II of the guest perspective piece from Dr. Geller. 


Twenty-nine months have passed since the Tohoku earthquake and the level-7 accident at the Fukushima Dai-ichi nuclear power station (NPS). Japan’s government, nuclear industry, and public have, alas, not used these 29 months fruitfully.

The basic facts of the Fukushima Dai-ichi NPS accident are well known. The magnitude-9 earthquake on March 11, 2011 generated a mega-tsunami of about 13m at the Fukushima NPS, far in excess of the design basis. After the tsunami hit the NPS all emergency power was lost, as external power became unavailable, and on-site emergency diesel generators were knocked out by the flooding caused by the tsunami. Meltdowns and hydrogen explosions at several reactors ensued, with extensive release of radiation. There were many shortcomings in the handling of the accident by the plant operator, Tokyo Electric Power Co. (TEPCO), and the government. The situation has now been largely stabilized but is not completely under control; full decommissioning will probably take at least 30 to 40 years. Many residents of the area surrounding the NPS have lost their homes and livelihoods and may be facing a lifetime of enforced relocation. These refugees have not yet received adequate compensation.

To move forward it first is necessary to look backward—the factors that caused the Fukushima accident must be analyzed and identified, and systemic flaws must be rectified. We should also analyze the good news—data from NPSs at which accidents did not occur, notably Tohoku Electric Power Company’s Onagawa NPS.

Let’s start with the bad news. Although many public inquiries into the Fukushima accident have been conducted, the following three fundamental issues have not yet been sufficiently addressed.

(1) The safety myth. Starting roughly 50 years ago, when the possibility of introducing nuclear power into Japan was first discussed, proponents repeatedly told the public that nuclear power was “safe.” This reiteration created the public misperception referred to as the “safety myth” (“anzen shinwa” in Japanese). This statement is of course, strictly speaking, incorrect—everything built by humans has some risks. It would have been far better to honestly explain the various risks and the countermeasures being taken to deal with them, but probably proponents were afraid that this would make it impossible to secure public acceptance of NPS construction.

It undoubtedly was widely—albeit largely tacitly—recognized in Japan that if there were no risks associated with nuclear power it would not have been necessary to site NPSs in remote areas and offer municipal governments large subsidies for accepting them. However, once the safety myth became the stated public position of the government and the nuclear industry, it became politically impossible to openly discuss nuclear accident scenarios and countermeasures. This helped to engender the smugness, complacency and lack of preparation that regrettably became apparent in March 2011.

(2) Responsibility for nuclear accidents. An NPS—or any other structure—must be designed to withstand specified levels (the “design basis” events) of earthquake shaking, tsunami height, etc. Engineers obtain advice from geoscience consultants (e.g., geologists, seismologists, soil engineers) when deciding on the design basis for an NPS, but it should be firmly established both by law and public consensus that responsibility is borne by the engineer rather than by the consultant. If this is not done then there is an obvious moral hazard—the engineer can shop around for a geoscience consultant who will set a low design basis, and then use compliance with the consultant’s advice as an excuse to evade taking responsibility for accidents caused by beyond-design basis earthquakes or tsunamis.

I don’t know who did the original consulting on earthquake and tsunami risk at Fukushima in the 1960s, but a committee of the Japan Society of Civil Engineers (JSCE) published a tsunami assessment in 2002 which became the basis for TEPCO’s and the government’s expected tsunami at Fukushima—roughly 6 meters at the shore line.

The members of the JSCE committee were almost all either from the nuclear industry or were so-called “go yō gakusha” (a cynical term for researchers who are perennially appointed to government advisory panels and who almost invariably conform to the government’s position). Neither JSCE nor TEPCO nor the government appears to have made significant efforts to canvass a wider range of expert opinion inside or outside Japan.

After the 2011 Tohoku tsunami caused the Fukushima accident, TEPCO spokesmen, government officials and many geoscientists (basically members of the same general group as the geoscientists who said that preparing for a 6m tsunami would be enough) popped up in the media saying that the actual tsunami (about 13m high at Fukushima) was unforeseeable (“soutei gai” in Japanese). Neither TEPCO officials, nor government officials, nor the geoscience consultants have yet faced any criminal or civil liability for the accident. The issue of responsibility seems to have vanished into thin air.

(3) Geoscience Issues. Science works by formulating hypotheses and testing them against observed or experimental data. Most fail and are rejected. The few that appear to succeed are further tested. Most of these, too, are eventually rejected. The few hypotheses that survive this gauntlet are often modified significantly during the testing and winnowing process. Problems can arise when unvalidated hypotheses are adopted as the basis for public policy without the recognition that they may be on shaky ground.

Over time, many earthquake scientists have proposed a variety of hypotheses regarding possible future tsunamis and earthquakes. Some may seem so intuitively obvious that they are adopted without people even realizing that they’re just hypotheses. One example is the idea that the way to determine the tsunami risk at a particular site is to study the tsunami history at that particular site. When the Fukushima Dai-ichi NPSs were first designed in the 1960s a very low design tsunami basis was adopted because no large tsunamis were known to have occurred at Fukushima in the past 300 years or so. The tsunami generated by the 1960 Chile earthquake observed at Fukushima was therefore used as the basis for the tsunami design standard of several meters.

The problem with using site-specific information to determine the design basis is that the Earth is 4.6 billion years old, and the Japanese islands were formed about 15 million years ago, so data from only the past several hundred years will probably not be fully representative. In the case of Fukushima, even in the 1960s, it should have been obvious to both the engineers and their geoscience consultants that if an earthquake about 10,000 km from Fukushima could generate a tsunami of several meters in height at Fukushima, a similar earthquake immediately offshore Fukushima could generate a much larger tsunami, but this possibility was not reflected in the design of the NPS.

Based on global earthquake data from the past 100 years, some scientists proposed in the 1980s that Tohoku and Sumatra could have relatively large magnitude-8 earthquakes, but not magnitude-9 mega-earthquakes like the 1960 Chile earthquake. This hypothesis was widely accepted in Japan. However, the 2004 M9 quake in Sumatra and the discovery reported by a series of papers in the 1990s and in 2001 of sedimentary deposits from three mega-tsunamis in Tohoku in the past 3,000 years served to disprove the hypothesis that M9 earthquakes could not occur in Tohoku. This evidence was known to TEPCO and the regulators but was not acted on. What evidently happened was a classic case of “cognitive dissonance.” Once TEPCO and the regulators got locked in to one mindset, they proved unable to reopen their minds as new evidence emerged. As a result, they were woefully underprepared for the M9 earthquake and subsequent tsunami on March 11, 2011, with the disastrous consequences we now know all too well.

What went right. The Onagawa NPS of Tohoku Electric Power Company, located on the Pacific coast about 100 km to the north of Fukushima, was built on higher ground than the Fukushima Dai-ichi NPS. The location of seawater circulation motors and pumps was also an important reason that the Onagawa reactors could be shut down safely and that there was no significant safety-related tsunami damage, although there was some non-safety-related damage requiring expensive repairs. Although the earthquake shaking was slightly in excess of the design basis and its duration was several times longer than the design basis, an international inspection team under IAEA auspices reported that there was no significant damage at the plant. The plant operator is still conducting detailed inspections of the units of the Onagawa NPS as a step toward verifying that restarting can be safely implemented. (See presentation by Isao Kato, Deputy Superintendent of the Onagawa NPS, and IAEA report for details.)

Summary: Bill Clinton’s campaign headquarters in 1992 famously focused attention on the key issue by posting a sign “It’s the economy, stupid!” on the wall above each staffer’s desk. Without in any way minimizing the issue of earthquake risk to NPSs, the combined lesson from Fukushima and Onagawa seems clear: “It’s the tsunami, stupid!”

>>In part II of Dr. Geller’s Idea Gallery series, the author discusses the road forward.