Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
SCIENCE AND ITS LIMITS: THE REGULATOR'S DILEMMA 11 original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution. rate in a large population of exposed mice. How many mutations will occur in a population of mice exposed to 100 millirems of radiation? Here the mutations, if induced at all by such low levels of exposure, are so rare that to demonstrate unequivocally an effect with 95 percent confidence would require the examination of many millions of mice. Though in principle this is not impossible, in practice it is. Moreover, even if we could perform so heroic a mouse experiment, the extrapolation of such findings to humans would still be fraught with uncertainty. Thus, the effects of very low level insult in human beings are rare events whose frequency again is beyond the ability of science to predict with accuracy. When dealing with events of this sort, science resorts to the language of probabilityâthat is, instead of saying that this accident will happen on that date or that a particular person exposed to a low-level insult will suffer a particular fate, it tries to assign probabilities for such occurrences. Of course, where the number of instances is very large or where the underlying mechanisms are fully understood, the probabilities themselves are perfectly reliable. In quantum mechanics there is no uncertainty as to the probability distributions. But in the class of phenomena being discussed here, even though the likelihood of an event's happening or of a disease's being caused by a specific exposure is given as a probability, the probability distribution itself is very uncertain. One can think of a somewhat fuzzy demarcation between what I have called science and trans-science: the domain of science covers phenomena that are deterministic, or the probability of whose occurrence can itself be stated precisely; trans- science covers the domain of events whose probability of occurrence is itself highly uncertain. "Scientific" Approaches to Rare Events Despite the difficulties, science has devised mechanisms for estimating, however imperfectly, the probability of rare events. For accidents, the technique is probabilistic risk assessment (PRA); for low-level insults, a variety of empirical and theoretical approaches have been used. Though probabilistic risk assessment had been used in the aerospace industry for a long time, it first sprang into public prominence with Norman C. Rasmussen's Reactor Safety Study in 1975 (U.S. Nuclear Regulatory Commission, 1975). Probabilistic risk assessment seeks to identify all sequences of subsystem failures that may lead to a failure of the overall system; it then tries to estimate the consequences of each system failure so identified. The output of a PRA is a probability distribution, P(C); that is, the probability, P, per reactor-year (RY), of a consequence having magnitude C. Consequences include both material damage and health effects. The
SCIENCE AND ITS LIMITS: THE REGULATOR'S DILEMMA 12 original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution. probability of accidents having large consequences is usually less than the probability of accidents having small consequences. A probabilistic risk assessment for a reactor requires two separate estimates: first, an estimate of the probability of each accident sequence and, second, an estimate of the consequencesâparticularly the damage to human healthâcaused by the uncontrolled effluents released in the accident. An accident sequence is a series of equipment malfunctions or human miscalculations: a pump that fails to start, a valve that does not close, an operator confusing an "on" with an "off" signal. For many of these individual events, we have statistical dataâfor example, enough valves have operated for enough years so that at least in principle we can make pretty good estimates of the probability of failure. But uncertainties still remain, since we can never be certain that we have identified every relevant sequence. Proof of the adequacy of PRA must therefore await the accumulation of operating experience. For example, the median probability of a core melt in a light-water reactor (LWR), according to the original Rasmussen report, was 5 Ã 10-5/RY; the core melt at Three Mile Island's number 2 reactor (TMI-2) occurred after only 700 light-water reactor- years. However, TMI-2 differed from the reactors treated by Rasmussen and, in retrospect, one could rationalize most of the discrepancy between the Rasmussen estimate and the seemingly premature occurrence at TMI-2 (Rasmussen, 1981). Since TMI-2, the world's LWRs have accumulated some 1,500 years of reactor operation without a core melt. This performance places an upper limit on the a priori estimate of the core-melt probability. Thus, if this probability were as high as 10-3/RY (as had been suggested by D. Okrent, 1981), then the likelihood of surviving 1,500 reactor-years would not be more than 22 percent; or, we can say with 78 percent confidence that the core-melt probability is not as high as I in 1,000 reactor-years. With 500 LWRs on line in the world, should we survive until 2000 without another core melt, we could then say with 95 percent confidence that the core-melt probability is not higher than 1 in 3,000 reactor-years. In the absence of such experience, one is left with rather subjective judgments. Although the Lewis critique (U. S. Nuclear Regulatory Commission, 1978) of Rasmussen's study asserted that it could not place a bound on the uncertainty of PRA, Rasmussen has argued that his estimate of core-melt probability might be in error by about a factor of 10âthat is, the probability may be as high as 1 in 2,000 reactor-years or as low as 1 in 200,000 reactor-years. As we see, we can, after 1,500 reactor-years of operation without a core melt, say with about 50 percent confidence that Rasmussen's upper limit (1 in 2,000 reactor-years) is not too optimistic. And if we survive to 2000 without a core melt, the confidence level with which we can make this
SCIENCE AND ITS LIMITS: THE REGULATOR'S DILEMMA 13 original typesetting files. Page breaks are true to the original; line lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as the authoritative version for attribution. assertion rises to 95 percent. Our confidence in probabilistic risk analysis can eventually be tested against actual, observable experience. But until this experience has been accumulated, we must concede that any probability we predict must be highly uncertain. To this degree our science is incapable of dealing with rare accidents, but time, so to speak, annihilates uncertainty in estimates of accident probability. Unfortunately, time does not annihilate uncertainties over consequences as unequivocally as it does frequency of accidents. A large reactor or chemical plant accident can cause both immediate, acute health effects and delayed, chronic effects. If the exposure either to radiation or to methyl isocyanate (MIC) is high enough, the effect on health is quite certain. For example, a single exposure of about 400 rems will cause about half of those exposed to die. On the other hand, in a large accident there will also be many who are exposed to smaller doses, indeed to doses so low that the dose-response is indeterminable. At Bhopal, 200,000 people who were exposed to MIC recovered. We cannot say positively whether or not they will suffer some chronic disability. The worst accident envisaged in the Rasmussen study, with a probability of 10-9/RY, would lead to an estimated 3,300 early fatalities, 45,000 early illnesses, and 1,500 per year delayed cancers among 10 million exposed people. Almost all of the estimated delayed cancers are attributed to exposures of less than 1,000 millirems per year, a level at which it is very difficult to estimate the risk of inducing cancer. Similarly, the critique by the American Physical Society (1975) of the Rasmussen study attributed an additional 10,000 deaths over 30 years among 10 million people exposed to cesium-135 from a large accident. The average exposure in this case was 250 millirems per year, again, a level at which our estimates of dose-response are extremely uncertain. Has the nuclear community, particularly its regulators, figuratively shot itself in the foot by trying to estimate the number of delayed casualties resulting from these low-level exposures? In retrospect, the Rasmussen study would have been on more solid ground had it confined its estimates only to those health effects that resulted from exposures at higher levels, where science makes reliable estimates. For the lower exposures the consequences could have been stated simply as the number of man-rems of exposure of individuals whose total exposure did not exceed, say, 5,000 millirems, without trying to convert this number into numbers of latent cancers. Thus, health consequences would be reported in two categories: (1) for highly exposed individuals, the number of health effects; (2) for slightly exposed individuals, the total man-rems, or even the distribution of exposures accrued by the large number of individuals so exposed. Perhaps some scheme such as this could be adopted in reporting the results of future
