“Risk = Probability × Consequences”: Probability, Uncertainty, and the Nuclear Regulatory Commission's Evolving Risk Communication Rhetoric

Abstract

This article examines the rhetoric employed by the Atomic Energy Commission and its successor, the Nuclear Regulatory Commission, to communicate the risks of nuclear power to legislators and the public. Close reading of official and unofficial documents demonstrates the importance of developing an effective risk-communication strategy in anticipation of danger rather than in response.

Keywords:

• nuclear power
• rhetoric
• rhetoric of technology
• risk communication
• technical communication
• technology

Notes

The NRC's WASH- and NUREG-series reports use varying page numbering systems. NUREG-0980, cited here, begins page numbering anew in each chapter and appendix; therefore page 1–33 is the 33rd page of chapter 1. Citations throughout this essay reflect each document's numbering system.

The world's first electricity-generating nuclear reactor, the Experimental Breeder Reactor outside of Arco, Idaho, suffered a partial core meltdown in November 1955 that resulted in no casualties.

An unpublished 1965 update to the report more than doubled that estimate, to a new total of $17 billion, and increased the estimated death total more than eightfold (Fuller, 1976/1984, p. 137).

In an unpublished 1965 update to the Brookhaven report, AEC Deputy Director of Regulation Clifford Beck wrote, “The possibility that such accidents might occur cannot be excluded, and there has been accumulated some evidence that a few failures may have almost occurred which could have resulted in more serious accidents” (as cited in Fuller, p. 153).

The coverage required has increased with each successive renewal of the Act, with the current industry-paid liability coverage surpassing $12 billion (see NRC, 2011).

“An event tree is a logic method for identifying the various possible outcomes of a given event which is called the initiating event”; initiating events for a reactor accident might include a rupture in the coolant system or a sudden loss of power to key systems (NRC, 1975b, p. 42). Fault trees work backward from a specific failure and estimate the probable cause for that failure (p. 45).

Using this methodology, one reactor operating for 100 years and 100 reactors operating for one year both constitute 100 reactor years of experience.

Seife (2003) notes that many engineers are resistant to probabilistic analysis because it embraces uncertainty rather than working to eliminate it. The concept of uncertainty is foundational to risk assessment and communication, however: If all potentialities were known and accounted for, there would be no risk to communicate. Although the WASH-1400 team provided firm numbers in an attempt to quantify risk, they admitted that the numbers are projections and not calculations of actual frequency. They noted that risks “had to be estimated, rather than measured, because although there are about 50 such plants now operating, there have been no nuclear accidents to date resulting in significant releases of radioactivity in U.S. commercial nuclear power plants” (NRC, 1975a, p. 1).

The 2007 NRC fact sheet on PRA refers to the single number as a “point estimate,” noting that although such numbers are convenient for communicating and comparing risk, a distributed range offers a more complete picture of the understood risk.

In a commentary on the report, a special review committee criticized NUREG-1150 for having “implied greater insight into the processes assumed to be taking place than was justified” (NRC, 1990b, p. 14).

At the time of the SOARCA documents’ writing, limited information on the Fukushima meltdowns was available to the writers. However, they assure readers that “continued U.S. plant operation and NRC licensing activities present […] no imminent risk” (NRC, 2012a, p. 14).

Although the document is repeatedly referred to as a “brochure,” it is clearly intended for online distribution: It is 62 pages long and liberally sprinkled with hyperlinks as well as color graphics that would be quite costly to reproduce in bulk.

Additional information

Notes on contributors

David Reamer

David Reamer is an associate professor of English and Writing at the University of Tampa, where he teaches courses in technical writing, media writing, publication management, and academic writing. His work has appeared in Rhetoric Review, Programmatic Perspectives, FYHC: First-Year Honors Composition, and the edited collection Stories of Mentoring: Theory and Praxis.