lesser degree, 131I, whose half-lives are 5.3 and 8.1 days, respectively. The initial estimate of the total population dose was 3500 person-rem. The mean dose was about 2 mrem (0.002 rem) per person. The highest dose anyone received was below 100 mrem. These estimates are likely to be too high, since they do not allow for the protective effects of evacuation, nor for protective shielding by buildings in the case of persons staying indoors. The amount of radioactivity in milk due to the incident was trivial.

Using the linear dose-response hypothesis, it can be estimated that less than one cancer and less than one serious genetic effect were induced in the entire population at risk (the existing cancer mortality rate for that population is 3400 per year).

At the time of the committee’s final deliberations, it appeared that the health risks of this accident were negligible (a single automobile accident during the evacuation would have done more damage), but the doubts and fears raised by the incident could have far-reaching consequences.

WASH-1400 In 1975 the Nuclear Regulatory Commission published the results of a study undertaken to estimate the probability of the great variety of accidents that might occur in the operation of nuclear reactors. The Reactor Safety Study (known as WASH-1400 or the Rasmussen Report)133,134 employed decision-tree and fault-tree analysis, based on judgment of risk at each step of a specific sequence of events leading to an accident, and tempered by the operating experience that had been accumulated.

The NRC has recently withdrawn its approval of the absolute probabilities of risk of reactor accidents calculated in the report,135 without commenting whether individually or as a group they are too high or too low. Nonetheless, as the external review group of the NRC has pointed out,136 WASH-1400 is still the most comprehensive attack on the problem of reactor safety, and its approach, collected materials, discussions, and experience are of great importance in providing a base from which the examination of this complex problem can advance.

Reactor accidents that involve “breach of containment” release radioactive materials into the atmosphere and thus may affect the public. The accident at Three Mile Island did not breach containment; its radioactive emissions were released by the operator in the course of controlling the rising pressure within the reactor during its shutdown (necessitated by insufficient cooling).*

WASH-1400 estimated the frequency of accidents of this severity as 1 in 300 to 1 in 30,000 reactor-years of operation. The range has proved to be


Statement 9–9, by J.P.Holdren: Actually, part of the release resulted when a design error caused automatic pumping of contaminated water into an unsealed building.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement