Top of This issue Current issue
The effects of radiation from an incident such as a nuclear blast or plant meltdown are characterized as being affected by "time, distance and shielding". The events this week in Japan make me think that in order to be completely safe, we would have to measure both time and distance in light years and rely on the atmosphere of another planet to provide the shielding. In other words, if I could choose to live on Proxima Centauri 4, it would be a matter of unconcern to me if you fuck up this planet with radiation. If your nuclear power plant and my ass must co-exist, however, on the same tiny planet, I don't think I want you messing around with nuclear power, after all.
Power plants and earthquakes
I adore Google's "Advanced Search" function, which allowed me to exclude the current Japanese events while looking for occurrences of "nuclear plant" and "earthquake". Here's what I found.
The Wikipedia article on the Diablo Canyon nuclear power plant states: "Diablo Canyon is designed to withstand a 7.5 magnitude earthquake from four faults, including the nearby San Andreas and Hosgri faults. Equipped with advanced seismic monitoring and safety systems, the plant is designed to shut down promptly in the event of significant ground motion." The only source given, however, is a 1976 Time Magazine article, which is less sanguine. "All atomic power plants located in earthquake zones must be constructed to withstand the most severe shaking ever recorded in the area. But geologists keep finding new earthquake faults—sometimes after construction has started." The article relates how the Hosgri fault, which is offshore, was not discovered until the billion dollar plant was already underway. Hosgri was thought to have caused a 7.25 magnitude earthquake, while Diablo Canyon had only been designed at the time to withstand a 6.75 magnitude quake. The plant was upgraded after discovery of the fault.
A 1989 report posted on the Nader website is prescient. Written the day after a 6.9 magnitude San Francisco earthquake caused the collapse of a freeway, and a shut down at Diablo Canyon, the report quotes Ken Bossong of the Critical Mass Energy Project in 'Washington. D.C.: "Numerous U.S. nuclear reactors such as Diablo Canyon in California, Indian Point in New York, and Millstone in Connecticut are constructed perilously close to known fault lines." He sketches two scenarios, one in which the containment is breached and the core exposed, and another more likely one in which the temporary pools in which spent fuel is kept are exposed. (The latter has definitely happened in Japan, where information is equivocal on whether reactor containment has also been breached.) Bossong observes rather poignantly: ""Whether any nuclear reactor could withstand a quake that registers as high as 7.5 on the Richter scale is purely hypothetical since the nuclear industry has no experience with a quake of that magnitude."
In a 2004 editorial on SFGate.com,Haydar Akbari notes that a 6.6 magnitude earthquake the prior month in the city of Bam, Iran, in which 40,000 people died, illuminates "the potential danger of the nuclear power plant being constructed in the southern city of Bushehr, which is on the same geological fault line" as Bam. Bushehr itself "has been destroyed three times by earthquakes in recent history (1877, 1911 and 1962)." While Iranian officials and the company building the plant insist that it can withstand up to a 7.2 magnitude earthquake, "there is no guarantee that a temblor of greater magnitude will not strike. If that happens, the immediate and long-term consequences will be larger and more tragic than the Chernobyl nuclear plant disaster in 1986."
A 2007 article in New Scientist reports the siting of a new Russian-designed reactor at a site "rejected in the 1980s because it was prone to earthquakes. The proposed site, at Belene, is near an active fault, according to Gueorgui Kastchiev, the former head of Bulgaria's nuclear plant at Kozloduy - where four of the six Russian reactors have been deemed unsafe and shut down." I particularly loved this coda: "Also, Kastchiev says, the design is untested, and Bulgaria has a poor safety culture." Othrwise, everything's fine!
In a 2008 article soon after the Chinese earthquake, the Telegraph reports "China has a research reactor, two nuclear fuel production sites and two atomic weapons sites in Sichuan province, where the quake struck, between 40 and 90 miles from the epicentre.... China's largest plutonium production reactor is also in the quake zone at Guangyuan." Plutonium is the most lethal type of nuclear emission when inhaled, according to this week's news reports. The article does not give the quake's magnitude, reported elsewhere as 7.9.
A 2009 article from the Phillipines describes the opposition of environmental groups to the recommissioning of the Bataan Nuclear Power Plant, built near fault lines at the orders of dictator Ferdinand Marcos in the 1970's but never operated. One of the organizers of a "no to BNPP" group is quoted as saying that "Tuesday morning’s earthquake near Iba, Zambales is a reminder of the hazards that await the BNPP and the community within the area". That one was 5.3 on the Richter scale.
A Canadian environmental lawyer states in a blog entry from June 2010: "Soon after our office stopped shaking in today’s second earthquake, I started to wonder again about Ontario’s nuclear plants. OPG says we shouldn’t worry: 'There have been no earthquakes greater than a Magnitude 4 within 50 km of the Pickering nuclear power station over the last 100 plus years.' This morning’s earthquake is reported to have been a magnitude 3; the afternoon earthquake is now being reported as 5 or 5.5."
In January, a Greenpeace blog post quoting an Indian thinktank report, said that the "Jaitapur nuclear power plant in the Ratnagiri district of Maharashtra [is] sitting on an earthquake zone".
Probability
The way we play the odds in life is very interesting. Most of us are at least vaguely aware that we risk our lives every day, driving on highways, flying on airplanes, and even walking down New York streets or (a favorite counter-argument of people arguing the safety of technology or against odds-based arguments) taking a shower. (Do a Google search on "odds of slipping in shower" and you will find an unsourced assertion that has spread everywhere that the odds of a slip and fall in the shower are 1 in 2232.)
When we do think about these things, we tend to get very nervous, and to say that the odds of dying in a particular way are "infinitesmally small". However, this statement is not the same as "impossible". The fact that something is statistically unlikely does not translate to "it will never happen", but the opposite: that given enough time, it will happen.
According to a 1935 Science Newsletter article on the odds of being dealt a 13-spade hand at bridge, "if every one of the 10 million bridge players in the U.S. dealt 20 bridge hands 365 days in the year, it would take the whole group of enthusiasts—or would they be lunatics?—just about 8 years, 8 months, and 1 week to deal enough hands to get a single one containing 13 spades." This is based on a calculation that the odds of a 13 spade hand are "one in 635,013,559,600 deals". From a recent source, a more relaxed prediction: "With 20 million bridge players dealing 30 hands a week, we should get one all-spade hand per 20 years." However, what this really means is that if enough humans play bridge long enough, it will happen.
In an MSNBC report: "Each year, at the typical nuclear reactor in the U.S., there's a 1 in 74,176 chance that the core could be damaged by an earthquake, exposing the public to radiation. That's 10 times more likely than you winning $10,000 by buying a ticket in the Powerball multistate lottery, where the chance is 1 in 723,145." The most earthquake-endangered reactor in the country turns out, surprisingly, to be Indian Point, in Westchester, NY. The reason: its not as well-designed; the plants like Diablo Canyon, built near fault lines in California, were designed accordingly, while New York earthquake risks were not as well understood (there is apparently a fault line through Westchester).
When we fly a plane, we know that there is a 1 in x chance it will crash. Because most of us don't fly every day or even every month, we know that the odds are also good that a rare crash, is likely to happen to someone else. Similarly, when we drive, we can always hope that the inevitable fatal crash which occurs the same day will happen on some other highway, to someone else.
However, when we are calculating the odds of an event which may affect the whole population of this tiny planet, like a radiation plume which will go everywhere in the upper atmosphere, the logic (and the huge subjective component) becomes very different.
The 1 in 74,176 chance took just fifty-six years to happen, from the first Russian nuclear plant coming online in 1954, the year of my birth.
Where is it written, by the way, that the next earthquake in any area will not exceed the last in magnitude? Designing a plant to withstand the biggest local earthquake ever does not make any sense. A 9 magnitude earthquake must always happen somewhere for the first time, before it happens the second time.
Ten thousand years
There is an ongoing debate how to warn future humans who may not speak English or be able to read, of the nuclear waste buried at places like Yucca mountain. (Plutonium 239 has a half life of 24,000 years.) According to a Wired article, "the Department of Energy convened a conclave of scientists, linguists, anthropologists and sci-fi thinkers to develop an elaborate system intended to shout 'Danger!' to any human being for the next 10,000 years -- regardless of what language they speak or technology they use." Absurd as this is, they are at least prescient to think about this at all.
If the thinking is that we are at the beginning of a technology we will use--or at least mediate the consequences of--for thousands of years, are we really comfortable with the idea that a magnitude 9 earthquake will hit a power plant once every sixty years, during those millenia? Can we live with that, in terms of the human and financial losses?
Or are we cheerfully optimistic about future safety improvements? In the development of the paper clip, as described by Henry Petroski in The Evolution of Everyday Objects, every glitch is answered by a new improvisation, Of course, paper clips don't kill people when they fail. Everytime a plant has a serious incident, will we design a klugey fix to avoid the same problem next time? If so, do we think that we will be good enough to design a plant that can withstand a 9 magnitude earthquake coupled with a tsunami? On what grounds?
At age 12, as I understood that the growth of population compared to the growth of crops and other resources could lead to a dire "twenty-ninth day" emergency in the future, I asked my father what would happen and he answered: "Oh, people will think of something. They always do." But I believe instead, that there are limits to human ingenuity, and limits to technology, or at least the technology we will be able to discover. Physicists say we will never travel faster than light. The easy technology fixes of our science fiction imagination--the force field you could erect around a melting-down plant, the transporter beam to send the destroyed plant to the far side of the moon--will not be available. In an article some years ago, I compared the increased obesity of American boys to the increased buffness of their G.I. Joe toys. Instead of solving real problems, we seem to wish them away. Bush's "Brownie, you're doing a heck of a job" while corpses washed around during Katrina, resonates with the Japanese government's performance today.
Normal accidents
One of my bibles is Charles Perrow's Normal Accidents, a 1984 book which, sadly, is out of print. Dr. Perrow analyzes those aspects of human nature and business as usual which contribute to or cause technology-related accidents. My favorite section of the book is called "Losing a lake". It relates a 1980 incident when a Texaco rig located in lake Peigneur, Louisiana, accidentally punctured a salt mine 1,300 feet below the lake, which then drained out, pulling in after itself 65 acres of a local tourist attraction, the Rip van Winkle Live Oak gardens. In Perrow's terms, this "briefly created an interactive and tightly coupled system out of two independent linear systems. We do not generally expect," he adds trenchantly, "a Louisiana lake, an oil rig, and a salt mine to suddenly be subsystems of a larger system".
Like most Greek myths which resonate with or provide magical explanations for, human situations, the story of Cassandra is sadly familiar to anyone who has correctly predicted a problem and been ignored. The human herd mentality, vanity and endless capacity for denial result in our refusing to believe valid information, and then acting like its newly discovered when the problem actually happens. One of the morbidly enjoyable tasks involved in writing for the Spectacle since 1995 has been that of searching for and reporting the Cassandras who accurately forecast large problems (and occasionally having the disturbing satisfaction of being one myself).
One such prediction was the statement by Ralph Nader, reported above. Perrow also is very clear what he thinks of nuclear plants. Chapter two of the book is entitled "Nuclear Power as a High Risk System". The first sentence is: "Why haven't we had more Three Mile Islands?" his answer: "we simply have not given the nuclear power system a reasonable amount of time to disclose its potential" (written five years before Chernobyl). Perrow notes that nuclear plants are non-trivial systems in whiuch trivial accidents can be quite destructive.
Perrow predicts Chernobyl by noting that the Soviets haven't bothered building the kind of containment we do. He says an incident in Russia will kill a lot more people than Three Mile Island--as it did.
He tells the now familiar story of Indian Point's siting and design, thirty five miles north of New York City, without consideration of earthquakes, and the fault lines being discovered afterwards. But he mentions two other things not as well known. Westinghouse, the plant operator, once applied to build a nuclear plant in Queens. More significantly, he describes a near-miss accident at Indian Point, where 100,000 gallons of Hudson River water leaked in, "covering the first 9 feet of the reactor vessel". Safety devices, which "expected" hot water, did not report the cold. The flooding was discovered by accident, and panic ensued; one supervisor restarted the reactor, without considering that the cold water may have caused cracking in the vessel; another shut it down. All this happened on a Friday morning, and nobody informed the Nuclear Regulatory Commission til Monday afternoon--nor the public and press until Tuesday, five days into the incident.
We could really lose New York City by accident one day.
Perrow concludes that "a review of some of the serious accidents that have occurred reveals the complexity of the plants, the difficulty of recovery from minor accidents so that they will not become major ones, the unlikelihood that the industry will even learn from the accidents, and the sanguine and casual response of the industry and the NRC."
Should we do it at all?
The idea of progress tells us that things always get better, the path is always upwards. But human history has proved this not true; we have already been through periods of historical darkness, where technology and learning were lost; also, the idea of progress begs a key question (and technological determinism indignantly denies the question) of whether everything that can be done, should be done. Petroski would have been amused (perhaps he was) by an article which appeared in the New York Times not long after the September 11, 2001 attacks, describing various technologies which could have saved lives that day. The most memorable idea was a personal parachute you keep under your desk so you can jump out the window in an emergency. Petroski's idea of "kluge after kluge" leads to the parachute, but it strikes me that I don't think we should create buildings you would need a parachute to escape from. Not unless you are willing to sacrifice the people you effectively coerce into working in them. The saddest stories in the world were those of the people who stayed on in the World Trade Towers after the 1993 attack and died in the 2001 one.
I discussed above the way we always misinterpret the math, understanding "rarely" to mean "never". But there is another aspect of human political culture, which leads us to do the math while denying it exists. The official story is that every drop of human life is precious, and we will never count the cost while saving life. The reality, however, is that almost every major political decision costs a life, somewhere (if you just know how to look at the consequences). Tea Party types howled about "death panels", accusing the new health legislation of fostering bureaucracies that could decide Grandma wasn't worth saving. Immediately on the heels of that, Tea Party-inflected Arizona discontinued Medicaid support for organ transplants, effectively killing people who already had a date scheduled to recieve one. We will spend any amount of fantasy money to to avoid one death, but frequently, not one cent of hard cash from our own pockets.
How does this relate to nuclear reactors? Again, the official story is that every life is sacred, and we will never knowingly endanger anyone. But nuclear plant siting decisions are made in the world of hard currency: we need the energy and are willing to play the odds. Until something better comes along, we will continue using a technology which may cause us to lose some tens or hundreds of thousands of people (or even an entire industrial country) every sixty years or so.
In Perrow's parlance, Indian Point--and the Japanese reactors, and every other reactor on earth--form a tightly coupled system with my ass. The problem is, nobody ever asked me, and I never agreed to it. But its a damn small planet.