With expert views ranging from faith that earthquakes could be predicted with certainty to disbelief that they could be predicted at all, confusion reigned in the field of seismic risk. Apparently even state-of-the-art science was unable to shed much light on the question of “how safe is safe enough” when it came to building nuclear power plants in Japan.
The nuclear accidents at Three Mile Island in 1979 and Chernobyl in 1986 stalled nuclear development in many places. Japan wasn’t one of them. There, construction continued full throttle. During the five years after Chernobyl, Japan added five new reactors, with several more under construction.
A map of Japan showing the location of its nuclear facilities. International Nuclear Safety Center, U.S. Department of Energy
The accident at Three Mile Island Unit 2 resulted from a series of equipment malfunctions, design flaws, and operator errors. Eventually about half of the fuel core melted. Nearly 150,000 people fled their homes. In the end, radiation releases from the reactor were small, but the accident symbolized to the world that nuclear power was not the safe form of energy generation its backers claimed.
Seven years after the accident at Three Mile Island, engineers at the Soviet-designed Chernobyl Nuclear Power Plant in the Ukraine were conducting a safety test on the Unit 4 reactor. After a sudden power surge, operators attempted an emergency shutdown, but power spiked instead, rupturing the reactor core and setting off a series of explosions. The graphite used to moderate fission in the core ignited, sending a massive plume of radiation over large areas of the Soviet Union and Europe. Ultimately, more than 350,000 people evacuated and resettled; a large area around the plant has been declared uninhabitable.
The Japanese government and news media portrayed Chernobyl as a man-made disaster caused by poorly trained operators and the structural defects of old, Soviet-built, Soviet-maintained equipment. The message was clear: such an event could not happen in Japan. But during the country’s own nuclear disaster in 2011, Japan’s Yomiuri Shimbun noted, in understated fashion: “[T]his assessment was optimistic.” “We have to recognize,” the newspaper declared in an editorial in April, “that there is no perfect technology”—a difficult admission in a country that prided itself on sophisticated engineering and exacting standards.
From the beginning of the era of nuclear power, the Japanese public had been repeatedly assured by government regulators, plant owners, and the media that it was inherently safe. Eventually, this view was generally accepted, despite the periodic eruption of scandals revealing shortcomings in the competence and integrity of those in charge of nuclear power production.
Headlines scattered over the decades built a disturbing picture. Reactor owners falsified reports. Regulators failed to scrutinize safety claims. Nuclear boosters dominated safety panels. Rules were buried for years in endless committee reviews. “Independent” experts were financially beholden to the nuclear industry for jobs or research funding. “Public” meetings were padded with industry shills posing as ordinary citizens. Between 2005 and 2009, as local officials sponsored a series of meetings to gauge constituents’ views on nuclear power development in their communities, NISA encouraged the operators of five nuclear plants to send employees to the sessions, posing as members of the public, to sing the praises of nuclear technology.
The utilities and regulators used the same tactic in the summer of 2011 as local governments debated allowing the restart of reactors shut down after the Fukushima Daiichi accident. Utility employees were encouraged to send anonymous e-mails supporting a restart or to show up at meetings and deliver the message in person. When the collusion was revealed, public opinion turned hostile.
Through it all, the nuclear industry in Japan remained largely unchallenged, insulated by official reassurances that the nation’s elaborate oversight system was functioning and the industry’s overall performance was beyond reproach, despite a few bad apples.
Challenge was unlikely to come from within the government. Nuclear energy in Japan has been described as “national policy run by the private sector.” Regulators and the regulated cohabit peacefully in a “nuclear village” whose mission is to promote nuclear power. When it comes to who is watching over whom, the lines often are blurred. A regulator today can become a utility employee tomorrow and vice versa.
It’s known as a “revolving door” in the United States. In Japan, moving from a public to a private sector job is called amakudari, “descent from heaven.”1 Bureaucrats know that when they are ready to retire, often at an early age, a comfortable job could be waiting in the industry they once regulated. There’s little incentive to rock the boat.
Shortly after the Fukushima accident, Japan’s Yomiuri Shimbun reported that thirteen former officials of government agencies that regulate energy companies were currently working for TEPCO or other power firms.
Another practice, known as amaagari, “ascent to heaven,” spins the revolving door in the opposite direction. Here, the nuclear industry sends retired nuclear utility officials to government agencies overseeing the nuclear industry. Again, ferreting out safety problems is not a high priority.
The ties between government and industry go beyond sharing personnel. Japan’s Asahi Shimbun reported in early 2012 that twenty-two of eighty-four members of the Nuclear Safety Commission and two of its five commissioners had received a total of about $1.1 million in donations from the nuclear industry over a five-year period ending in fiscal 2010. One-third of the NSC members on committees overseeing nuclear operations had received such donations. Prior to becoming chairman of the NSC in 2010, Haruki Madarame, then a professor at the University of Tokyo, had received about $49,000. Critics of nuclear power rarely were on the receiving end of such largesse. The “nuclear village” looked after its own.
On December 26, 2004, a magnitude 9.2 earthquake struck off the western coast of Sumatra, triggering a devastating tsunami across the Indian Ocean that killed more than 286,000 people in fourteen countries. The quake, like the one that struck Fukushima seven years later, was a subduction earthquake. And, like the 2011 quake, it caught seismologists by surprise. When the earthquake hit, a subcommittee of the NSC was in the midst of a long overdue review of Japan’s seismic standards for nuclear plants. The review, initiated after the 1995 Kobe earthquake, had been inching along for eleven years.
A major issue before the subcommittee was how to determine the maximum earthquake plants had to be able to withstand. The old standard required that every plant be able to survive a nearby magnitude 6.5 earthquake, which in Japan was fairly common. The new standard would replace that generic criterion with limits tailored to the seismic risks at each plant site. The utilities were concerned that the revised standard could effectively increase the earthquake magnitude their plants would have to withstand and require them to make costly seismic retrofits.
When the new guidelines were ultimately approved in September 2006, critics called them too vague. Among the opponents was Katsuhiko Ishibashi, a subcommittee member, who said the guidelines were full of loopholes. Angry at the subcommittee’s unwillingness to consider more rigorous standards, he resigned.
As Japan’s largest utility, TEPCO reaped enormous benefits from the blurred lines between government and industry. The company’s operating record raised plenty of warning signs about the dangers of such incestuous practices, however. And they were apparent well before March 2011.
In 2000, Kei Sugaoka, a nuclear inspector working for GE at Fukushima Daiichi, noticed a crack in a reactor’s steam dryer, which extracts excess moisture to prevent harm to the turbine. TEPCO directed Sugaoka to cover up the evidence. Eventually, Sugaoka notified government regulators of the problem. They ordered TEPCO to handle the matter on its own. Sugaoka was fired.
While TEPCO was ostensibly dealing with the matter, Fukushima Daiichi continued to operate. Then, in late summer 2002, the company admitted it had been falsifying safety records for years, covering up evidence of cracks in thirteen reactor core shro
uds at all three of its plants. The shrouds are stainless steel cylinders that hold fuel assemblies in place and help direct the flow of cooling water. At a press conference announcing the cover-up, government regulators declared that public safety was not threatened. They had reached that determination not from their own inspections, but from TEPCO’s assurances.
When the cover-up became public, TEPCO’s chairman and president resigned. The new chairman declared the falsification of records “the gravest crisis since the company was established.” Despite the scandal, both former executives were retained as advisors to TEPCO. Theirs weren’t the last heads to roll at the company.
• • •
In January 2007, TEPCO admitted to more falsified records, involving a total of about two hundred incidents dating back many years. (A newspaper headline the next day read: “Not Again!”) It was a year of catharsis for Japan’s nuclear industry. Six other utilities revealed their own unreported safety problems. NISA had directed the utilities to own up to any such issues, hoping to head off public opposition to new reactor construction. Whether the outpouring of misdeeds helped or hurt NISA’s cause is not clear.
Among the incidents TEPCO divulged were problems at its Kashiwazaki-Kariwa reactors on Japan’s west coast. Kashiwazaki-Kariwa is the world’s largest nuclear generating station, capable of producing 8,212 megawatts of electricity, nearly double Fukushima Daiichi’s output of 4,696 megawatts. In one instance, during a routine government inspection at Kashiwazaki-Kariwa, operators discovered that a component of the emergency core cooling system wasn’t working. Workers made adjustments in the control room to make it appear the pump was functioning.
According to TEPCO’s explanation, the utility wasn’t the problem; the rules were. “[F]alsification occurred because passing the inspections became the objective,” the company told the government. TEPCO admitted only to handling data “inappropriately,” not to lying to inspectors. For example, the utility failed to report that control rods came loose more than once in the reactor cores at Fukushima Daiichi, in one instance triggering a criticality accident that lasted seven and a half hours. Technically, TEPCO acted legitimately. Reporting the criticality mishap, for instance, was not required under Japan’s safety rules at the time.
YES, THE PAPERWORK MATTERS—A LOT
Fudging figures, doctoring reports, and creative accounting may seem minor, though deplorable, offenses. But they are not minor. When it comes to nuclear plant safety, the value of accurate, complete, and reliable paperwork cannot be underestimated.
In the United States, NRC inspectors audit only about 5 percent of the activities at nuclear plants, according to senior managers at the commission. Most of these audits involve reviewing the records of tests and inspections performed by plant workers. The NRC inspectors themselves witness only a very small fraction of actual tests and inspections. If safety inspectors could not trust a plant’s paperwork, they would have to personally observe many more activities than they do now to gain confidence that their assessment of the plant’s safety was a reflection of reality. To put this another way: when workers feel free to prepare fictional accounts of tests and inspections, nuclear safety assurances begin morphing from nonfiction to fiction as well.
Then came the earthquake that damaged Kashiwazaki-Kariwa. It occurred on the morning of July 16, 2007, with a rupture along the fault that a court had ruled in 2005 didn’t exist. The plant had been built to withstand a smaller quake; this one, with a magnitude of 6.8, created ground motion two and a half times greater than that plant had been designed for. The episode raised questions about TEPCO’s response to such crises that would prove all too relevant in 2011.
When the quake struck, three of the seven reactors were at full power; a fourth was in the process of starting up. The rest were out of service for refueling and maintenance. All four operating reactors automatically shut down. A fire broke out in an electrical transformer. Because it was a national holiday, the plant was shorthanded. Emergency response crews were difficult to assemble. The quake had damaged the plant’s own fire extinguishing system. City firefighters were pulled away from local emergencies and sent to the plant; their arrival was delayed by an hour because of quake damage.
At the time, Prime Minister Shinzo Abe criticized TEPCO for being too slow in reporting problems at Kashiwazaki-Kariwa. “Nuclear power can only operate with the public’s trust,” he told reporters. The International Atomic Energy Agency sent inspectors to the shaken plant and concluded that the damage was “less than expected.” The agency did, however, recommend a reevaluation of the seismic situation at Kashiwazaki-Kariwa, especially the existence of active faults beneath the site. Later in 2007, TEPCO reported a fourteen-mile-long (twenty-three kilometers) active fault in the seabed eleven miles from Kashiwazaki-Kariwa. The company said it had known about the fault since 2003, but did not report its findings then because TEPCO staff didn’t believe the fault would produce an earthquake large enough to threaten the reactors.
Immediately after the 2007 earthquake, the utility took all seven of the Kashiwazaki-Kariwa reactors out of service to check for damage and upgrade their seismic resistance.2 Without 20 percent of its generating capacity, TEPCO posted its first loss in twenty-eight years, totaling $1.44 billion. Its stock value dropped 30 percent.
To bolster confidence in the wake of the Kashiwazaki-Kariwa debacle, a new TEPCO president was named in 2008. Masataka Shimizu took over, replacing Tsunehisa Katsumata, who became chairman. The front office shuffle was getting almost routine: Katsumata had taken the top job in the wake of the 2002 falsification scandal; now Shimizu, a forty-year veteran of the company, stepped in. Both men had spent their entire careers at TEPCO. Shimizu made cost cutting a high priority, and within two years had returned TEPCO to the black, exceeding his target of $615 million in cuts. His secret? Reducing the frequency of inspections.
Direct structural damage is not the only danger that earthquakes can pose to Japan’s coastal nuclear plants. They can also cause devastating tsunamis, as became obvious in March 2011. Yet predicting tsunamis is a similarly inexact science. Given the uncertainties, developing standards for tsunami protection posed another conundrum for regulators and plant owners.
On tsunami protection, TEPCO took its cue from government, and Tokyo seemed in no hurry. The lengthy safety review that led to the 2006 earthquake guidelines did not address tsunamis. Proposed tsunami guidelines were moving through their own separate review process at a pace almost as slow as the tempo at which the quake standards had progressed. The nuclear industry also had a strong presence in these deliberations.
Just as the science of seismology had evolved in the years since Japan’s first nuclear plants were built, tsunami research, a newer field, was gaining ground rapidly. And Japan also provided an ideal laboratory: the Sanriku Coast alone had experienced four destructive tsunamis in a little more than a century, including the 1896 quake and wave that had a run-up height of nearly 125 feet.
Those tsunamis had struck to the north of Fukushima Daiichi. When construction began on the reactors in the late 1960s, engineers dismissed the likelihood that the plant location might be vulnerable. Based on the worst historical tsunami on record at the Fukushima site—resulting from a 1960 earthquake in Chile—the reactors were designed to withstand a tsunami with a maximum height of about ten feet (3.1 meters). TEPCO was so confident of this data point that the company actually lowered the height of the bluff where the plant was to be built by more than eighty feet (twenty-five meters). That made it easier to deliver heavy equipment to the site and to pump cooling water into the reactors. The company also said the excavation would enhance earthquake protection by placing the reactors on bedrock—albeit bedrock far below the natural elevation.
So confident was TEPCO that in 2001 it submitted to NISA a single-page tsunami plan that ruled out the possibility of a large tsunami hitting the plant and causing damage. The company provided no data to support its conclusions, and NISA apparent
ly asked for none. “This is all we saw,” a NISA official told the Associated Press, which located the TEPCO document a decade after it was submitted. “We did not look into the validity of the content.”
NISA lacked authority to question the accuracy of TEPCO’s presented data in any event, because tsunami plans from utilities were voluntary at this stage. TEPCO provided its information in advance of new tsunami guidelines being developed by the Japan Society of Civil Engineers. That group called for protective measures to be based not only on historical tsunami data, but also on wave heights calculated by numerical models, taking uncertainties into account. In the one-page plan for Fukushima Daiichi, TEPCO provided its projection using this method: a wave no higher than nineteen feet (5.8 meters), generated by a magnitude 8.0 quake modeled after one that had occurred in 1938. Accordingly, the utility made some modifications, including raising the motors of the seawater intake pumps. (In 2009, TEPCO further refined the calculations and raised its estimate of the maximum tsunami height to about twenty feet.)
The prevailing wisdom, based on the historical record, was that earthquakes larger than magnitude 8.0 would not occur in the offshore trench near Fukushima, although they had occurred further north off the Sanriku Coast. However, in 2002, the Headquarters for Earthquake Research Promotion said it was possible that earthquakes similar to the 1896 Meiji Sanriku earthquake, which had an estimated magnitude of 8.3, could occur anywhere along the trench as far south as the Bousou peninsula, well below Fukushima Prefecture.
TEPCO paid no attention to that prediction until 2008, when experts warned that the utility’s tsunami assessments could have underestimated the potential size of earthquakes off the Fukushima coast. As a result, TEPCO conducted a calculation assuming that an earthquake comparable to Meiji Sanriku occurred in that area. Based on this model, the utility now predicted a tsunami up to thirty-four feet (10.2 meters) high near the plant’s seawater intake pumps. A wave that large could sweep inland and reach a run-up height of more than fifty-one feet (15.7 meters) around Units 1 through 4 at Fukushima Daiichi. But TEPCO did not consider these results realistic, maintaining that the undersea faults in the area were not the type capable of causing large tsunami-generating earthquakes.
Fukushima: The Story of a Nuclear Disaster Page 7