Any citizen with even a casual awareness of the public debate over nuclear power is familiar with the usual talking points, pro and con, regarding this issue: safety, costs, environmental impacts, etc. I will not burden the reader with a rehash of these familiar issues.
Instead, I propose to enrich the debate with some issues with which the general public might be less familiar, all of which issues lead strongly to the conclusion that electric power generation from nuclear reactors should be phased out with deliberate speed and the technology abandoned — permanently.
This essay consists of three sections: First of all, the recent disaster at the Fukushima nuclear plant in Japan urgently brings the science of plate tectonics into the debate, and raises the question of whether the promoters of nuclear power are willing and able to take the long-term implications of that technology into consideration as they select sites for these facilities.
In the second section, we ask whether it is possible to accurately and reliably assess the safety of nuclear reactors. A failed attempt to do so thirty years ago suggests that such an assessment is impossible, not simply because of a lack of scientific knowledge and technological capacity, but more fundamentally, because of the insurmountable inability to anticipate all possible circumstances that might occur in the operation of the plant.
Finally, these and other considerations lead to the conclusion that nuclear power is not economically viable and sustainable without massive government subsidies that are unavailable to its competing technologies.
Fukushima: A Disaster Waiting to Happen
What were Tokyo Electric Power Co. (TEPCO) and General Electric thinking when they decided to site the world’s largest nuclear power complex at Fukushima, on the eastern coast of Northern Japan?
Perhaps they weren’t thinking at all, or at least they were thinking only for the short-term. Myopia is endemic to the corporate mind, which is dedicated to an early return on investment. “In the long-term,” John Maynard Keynes famously remarked, “we’ll all be dead.”
Nonetheless, a disastrous earthquake followed by a tsunami was certain to happen along the eastern coast of Japan. Not a question of if, but of when. That certainty was ordained by the science of plate tectonics and validated in the geological record.
The sword of Damocles hanging over Fukushima is the Japan Trench, located about 100 miles due east of, and parallel to, the coastline where the plant is located.
The trench is a subduction zone, where the Pacific plate dives down under the Okhotsk plate and into the mantle. The Japanese islands, like the Marianas and the Aleutians, owe their very existence to subduction which, as it grinds along, produces great earthquakes and tsunamis.
Tsunamis can be produced by volcanoes and landslides. But they most reliably occur along subduction zones, as the ocean floor during an earthquake is suddenly and violently jolted, causing a pulse of water to move outward and perpendicular to the fault line. The Indonesian tsunami of December 26, 2004, which killed almost a quarter of a million people, was caused by a magnitude 9.1 earthquake along a subduction zone about 100 miles west of Sumatra. Among other noteworthy subduction quakes/tsunamis are the “Good Friday” Alaska earthquake in 1964 (magnitude 9.1), and the Chilean earthquake of 2010 (magnitude 8.8).
And so, because the Japan Trench is parallel to the coast of northern Japan, the tsunami was aimed directly at that coast.
Because of the dynamics of plate tectonics, earthquake/tsunamis are endemic to Japan. For example, in 1923 a magnitude eight earthquake struck central Japan, leveling the city of Yokohama and destroying more than half of Tokyo, at the cost of about 100,000 lives.
The investors of the Fukushima plant knew all this, and yet they went ahead and built a facility that was designed to withstand a magnitude seven earthquake. (The Richter magnitude scale is not linear, it is logarithmic. Accordingly, the energy released in a magnitude nine quake is not two-ninths greater than that of a magnitude seven. It is about a thousand times greater). TEPCO continued to operate the facility, despite warnings from the International Atomic Energy Commission.
To put the matter bluntly, the investors and designers of Fukushima gambled that during the operational lifetime of the plant, there would be no earthquake greater than magnitude seven. They gambled, and the people of northern Japan lost. Economists call this loss an “externality.”
In California two commercial nuclear power facilities, at San Onofre between San Diego and Los Angeles and at Diablo Canyon near San Luis Obisbo, are located along the Pacific coast and near seismically active faults. As a resident of southern California, I must wonder if the operator of that plant, Southern California Edison, like TEPCO in Japan, is likewise gambling with my life and the lives of my neighbors. Heads they win, tails we lose.
And earthquakes and tsunamis are not the only, or even the greatest, threat posed by nuclear power reactors. The Three Mile Island accident was caused by a mechanical failure, and the Chernobyl disaster was caused by human error.
Building a nuclear power complex along a shoreline opposite a subduction zone is risky. That fact is a “known known.” How risky? That is an unknowable unknown. Any attempt to assess the risk, or for that matter the risk associated with any and all nuclear power plants, is almost certain to underestimate that risk. A reliable and accurate assessment of the risk of a failure of a nuclear power reactor is unobtainable, now and forever.
These are bold assertions that I will endeavor to demonstrate below. To do so, we will examine an ambitious and massive attempt, some thirty years ago, to assess the safety of nuclear power plants, and its subsequent spectacular failure to achieve that objective. Because the reasons for that failure remain valid today, this is a tale well worth retelling in the light of the disaster at Fukuyama and in the face of the determination of the Obama Administration, despite that disaster, to proceed with the construction of the first new nuclear power plants in thirty years.
Reactor Safety: The Rasmussen Report Revisited
Concerned about public criticism of their nuclear energy ambitions, the promoters of commercial atomic energy at the Atomic Energy Commission (AEC) initiated in 1972, the “Reactor Safety Study,” which was to become known as “The Rasmussen Report,” after its Director, Norman Rasmussen of the Massachusetts Institute of Technology. In August, 1974, the draft Report was released with much fanfare in a public-relations extravaganza that prompted one newspaper to proclaim: “Campaigners Against Nuclear Power Stations Put to Rout.” Following this triumphant entrance, scrupulous scientific assessment began behind the facade, after which it was all downhill for the Report. The AEC’s successor organization, the Nuclear Regulatory Commission (NRC), quietly withdrew endorsement of the Rasmussen Report in January, 1979.
Rushed into print to provide support for a renewal of the Price Anderson Act (a federally mandated limit of industry liability following a nuclear reactor failure), an eighteen page “Executive Summary” of the final Report was distributed to Congress and the Press in October, 1975, and in advance of the release of the full, 2300 page Report.
Perhaps the most famous item of the Executive Summary was the claim that the chances of being killed by a nuclear power plant “transient” is about equal to that of being killed by a meteorite. This mind-catching statistic has proven to have a longevity far exceeding that of the Report which spawned it. In general, the Summary concluded that:
…The likelihood of reactor accidents is much smaller than that of many non-nuclear accidents having similar consequences. All non-nuclear accidents examined in this study including fires, explosions, toxic chemical releases, dam failures, airplane crashes, earthquakes, hurricanes and tornadoes, are much more likely to occur and can have consequences comparable to, or larger than, those of nuclear accidents.
Closer examination revealed a startling discrepancy between the cheerful reassurances of the Executive Summary and the nine volumes of technical information. In his splendid book, The Cult of the Atom (Simon and Schuster, 1982), based upon tens of thousands of pages of AEC documents pried loose by the Freedom of Information Act, Daniel Ford observes that:
As one moves from the very technical material … to the Executive Summary … a change of tone as well as of technical content is evident. In the “back” of the study, there are cautionary notes, discussion of uncertainties in the data, and some sense that there may be important limitations to the results. The qualifications successively drop away as one moves toward the parts of the study that the public was intended to see. In the months following the study’s completion, the honesty of the official summary … became the most controversial issue.
The reassuring conclusions of the Rasmussen Report were based upon numerous highly questionable assumptions and methodologies. Among them:
- By definition, the report estimated damage and casualties due to anticipated events. There is no clear acknowledgment that all possible significant events were not, and could not be, covered by the study. As it turned out, the near-disaster at Three Mile Island was just one of several “unanticipated” events. And as noted above, a magnitude nine earthquake was not anticipated by the designers of the Fukushima plant.
- In fact, whole categories of failures were excluded from the risk estimates. For example, it was assumed that back-up safety systems would always operate in case of the failure of a primary system. Given this assumption, the risk of a catastrophic accident would be the product of the probability of the independent failure of both systems, and thus highly unlikely. However, this discounted the possibility of a “common-mode failure,” such as that at Browns Ferry, Alabama, in 1975 (soon after the release of the Report), where, due to faulty design, an accidental fire disabled both systems at once — yet another event excluded by the Rasmussen rules. Similarly, the Japanese earthquake and tsunami of March 11, 2011 disabled both the primary and backup safety systems at the Fukushima facility.
- The Report focused on mechanical and equipment failures, and discounted design flaws and “human error,” as if these were in some sense insignificant. Also overlooked was the possibility of sabotage and terrorism.
- The report adopted the so-called “fault-tree” method of analysis, described by the Report as “developed by the Department of Defense and NASA … [and] coming into increasing use in recent years.” Not so. As Daniel Ford reports, “long before [Rasmussen] adopt the fault-tree methods … the Apollo program engineers had discarded them.”  As a retired professor of engineering recently explained to me: “the simulation or probability tree … analyses … are used to locate the weak links in your design, given the possible sources of failure that you know of or can specify… [However, the analyses] are not meant to yield a credible probability of failure, but instead yield at best a lower bound for that probability.” (EP emphasis)
- The “probabilities” assigned to the component “events” in the “fault tree,” leading to a hypothetical failure, were based upon almost pure speculation, since, because the technology was new, the evaluators lacked any precedents upon which to base probability assessments. (Both Rasmussen himself, and his Report, admitted as much). (Ford 138, 141). Thus, because the Report was fundamentally an advocacy document, this gave its pro-nuclear investigators the license to concoct unrealistically low risk assessments.
- These “low risk estimates” in the Executive Summary were startling, to say the least: “non-nuclear events,” it claimed, “are about 10,000 times more likely to produce large numbers of fatalities than nuclear plants.” But the footnote to this statement gave it away, when it added that such “fatalities … are those that would be predicted to occur within a short period of time” after the accident. However, few fatalities due to radiation exposure are “short-term.” In fact, as Physicist Frank von Hipple pointed out, a careful reading of the voluminous technical material would disclose that for every ten “early deaths” conceded in the Summary, the same accident would cause an additional seven thousand cancer deaths. (Ford, 170) This was only one of the several scandalous discrepancies between the “public” Executive Summary and the Technical material in the Report, which led Morris Udall, then Chair of the Subcommittee on Energy and the Environment, to demand a new Executive Summary. The NRC refused.
- The “peer review” of the Report was perfunctory at best. The reviewers were given eleven days to assess an incomplete 3,000 page draft report — a schedule virtually designed to yield invalid assessments. Even so, many of the referees returned withering criticisms, especially of the statistical methods employed by the studies. The findings of this review group were not released by the AEC or the NRC, and the published Report was unaltered by these criticisms.
These and numerous other flaws in the study led one critic to wryly comment that “the chance of the Rasmussen Report being substantially wrong is somewhat more than the chance of your being hit by a meteorite.”
Though the general public was much impressed by the public relations show orchestrated by AEC, informed professional investigators immediately began the erosion of credibility. Among these were the Bulletin of the Atomic Scientists, the Union of Concerned Scientists, and, most significantly, an independent panel set up by the American Physical Society and chaired by Harold Lewis of the University of California, Santa Barbara. Each of these returned severe criticisms of the Report.
All this bad news eventually led the Reactor Safety Study into the halls of Congress. Daniel Ford describes what followed:
In some cases [congressional] members and staff probed the issues [of reactor safety] carefully, prepared detailed follow-up reports, and tried to bring about needed reforms. Congressman Morris Udall’s Subcommittee on Energy and the Environment, for example, held extensive hearings on the validity of the Reactor Safety Study. His protests about the misleading manner in which the report’s findings were presented to the public forced the NRC, in January 1979, to repudiate the results of the study. (p. 226)
And so, at length, the relentless discipline of science and scholarship, combined with a rare display of uncompromising congressional oversight investigation, brought about the downfall of the AEC/NRC “Reactor Safety Study.”
The NRC’s “withdrawal of endorsement” stood in stark contrast to its release, scarcely four years earlier. This time there were no publicity releases, media interviews or press conferences. It was hoped that the announcement would go unnoticed amidst the usual gross output of news out of Washington. Given the widespread public opposition to nuclear power, this expectation was bound to be frustrated.
In the end, the Rasmussen Report was yet another attempt at justification of “the peaceful atom” which backfired on the proponents. Historians looking back on this technological extravaganza may note, with some bewilderment, that however severe the attacks by the critics, commercial nuclear power was, in this case at least, inadvertently done in by its defenders.
Nuclear Power Fails the Free Market Test
Still more substantial objections to nuclear power have been raised by scientists and engineers much more qualified than I am. So I will not repeat them here. (To read these objections, google “Physicians for Social Responsibility,” “Union of Concerned Scientists” “Natural Resources Defense Council” and “The Rocky Mountain Institute“). However, in closing, a few additional concerns are worthy of mention.
(1) First of all, every source of electric power, with the exception of nuclear power, “fails safe.” A failure at a coal-fired plant would, at worst, destroy the plant. But the damage would be localized and short-term. Failures at a wind-farm or solar facility are trivial. However, the damage caused by a nuclear meltdown and radiation release endures for millennia and can render huge areas permanently uninhabitable, as they have in Ukraine and Belarus due to the Chernobyl disaster, and as they likely will in Japan following the Fukushima catastrophe.
(2) Nuclear industry assurances as to the safety of their facilities are flatly refuted by their unwillingness to fully indemnify the casualty and property losses that would result from a catastrophic release of radiation from a nuclear accident. Since 1957, the Price Anderson Act has set a limit on the liability that private industry must pay in the event of an accident. The amount of that limit, originally $560 million for each plant, has been routinely revised, so that as of 2005 the limit is now $10.8 billion for each incident. Clearly, the Fukushima disaster will exact a cost far exceeding that amount. Were such an event to occur in the United States, the cost of such a disaster in excess of ten billion would be born by the victims and by the taxpayers. The contradiction is stark: the nuclear industry and its enablers in the NRC tell the public that nuclear energy is safe. And yet, at the same time, they are unwilling to back up these assurances with a full indemnification of their facilities.
(3) The public has not been adequately informed of the ongoing hazards of nuclear power. For example, the Union of Concerned Scientists report that in the past year, there were fourteen “near misses” among the 104 nuclear plants operating in the United States. And according to the Washington Post (March 24), the Nuclear Regulatory Commission has disclosed that “A quarter of U.S. nuclear plants [are] not reporting equipment defects.”
(4) The widely-heard claim that “nobody in the United States has ever died due to commercial nuclear power” utilizes “the fallacy of the statistical casualty.” Specific cancer deaths due to artificial nuclear radiation are, of course, indistinguishable from cancer deaths due to other causes. Yet epidemiological studies show, beyond reasonable doubt, that some deaths are attributable to artificial radiation. The inference from “no identifiable specific deaths” to “no deaths whatever” is fallacy made infamous by the tobacco industry’s successful defense against suits filed by injured smokers or their surviving families.
(5) The claim that nuclear power is the “safest” source of energy commits the “fallacy of suppressed evidence.” Such a claim pretends that the risk of nuclear power is confined to the radiation risks adjacent to a normally operating plant and immediately following each “event.” Usually excluded from such assessments are deaths and injuries involved in the mining, milling, processing, shipment, reprocessing, storage and disposal of fuel — in short, the entire “fuel cycle.”
(6) Similarly, the claim that nuclear power is the “cheapest” power available is likewise based upon “the fallacy of suppressed evidence.” Specifically, nuclear proponents arrive at this conclusion by “externalizing” (i.e., failing to include) such costs as government subsidies for research and development, the costs of disposing of wastes, the cost of decommissioning of facilities, and, again, the cost of risks to human life, health and property. As noted above, the risk factor is excluded due to the Price Anderson Act and the failure to acknowledge “statistical casualties. Once all these “externalized costs” are included, nuclear power adds up to the most expensive energy source, hands down. Over fifty years of industry research, development and operation have not altered this fact. Meanwhile, as R & D of alternative energy sources progress and economies of scale kick in, the costs of solar, wind, tide, geo-thermal and biomass energy continue to fall. (See UCS, “Nuclear Power Subsidies: The Gift that Keeps on Taking,” and Amory Lovins, “With Nuclear Power, ‘No Acts of God Can Be Permitted’”).
Because of considerations such as these, no nuclear plants have been commissioned since the completion in 1985 of the Diablo Canyon facility along the central coast of California. The Obama Administration is prepared to change all this, as the President has announced $8 billion in federal loan guarantees to allow the building of the first nuclear power plant since Diablo Canyon.
Without this “federal intervention,” along with the Price-Anderson liability cap, no new nuclear plants would be built. The “free market” would not allow it. And yet there are no conspicuous complaints from the market fundamentalists on the right.
Why am I not surprised?
PostScript: My involvement in the Diablo Canyon controversy goes way back. In 1981, a group of local citizens blockaded the Diablo Canyon construction site in an act of civil disobedience, for which, predictably, they were arrested. At the time, I was a Visiting Associate Professor of Environmental Studies at the University of California, Santa Barbara. The defense team asked me to testify as to the “reasonableness” of the protesters’ belief that the Diablo Canyon nuclear reactors posed a significant danger to their community and to themselves. The prosecution objected on the grounds that the defense was asking me to “do the jury’s work.” The judge concurred, and so I was not permitted to testify. My account of this experience and critique of the ruling may be found in my unpublished paper, “A Philosopher’s Day in Court” at my website, The Online Gadfly, The discussion above of the Rasmussen Report is a revision of my unpublished class discussion paper from 1980, “The Strange Saga of the Rasmussen Report”.