Professor of Biology
Nuclear Policy Program
Adlai Stevenson College
University of California at Santa Cruz
Santa Cruz, California
95064 U.S.A

15 October 1987





This paper reports a quantitative, site-specific analysis of two nuclear accident scenarios aboard military vessels in a Canadian port. Conventional methodology used by the U.S. Nuclear Regulatory Commission (NRC) to regulate the U.S. civilian nuclear industry is combined with generally conservative assumptions (that is, assumptions that tend to understate the likely impact of an accident) to evaluate the consequences of hypothetical nuclear accidents aboard military ships in the port of Esquimalt and the adjacent city of Victoria, British Columbia, Canada. The results are used as a basis for policy evaluations on the issue of port visits. The results also bear upon the proposed acquisition by the Canadian Armed Forces of  10  to  12  nuclear-powered submarines. Although this analysis has been undertaken for a U.S. military propulsion reactor, comparable effects would be anticipated from a similar accident entailing a British Trafalgar class or a French Rubis class submarine.

The first accident scenario analyzed is incineration of a single nuclear warhead in a ship-board fire. Such an accident would produce a radioactive cloud containing plutonium-239 , which would be carried toward the northeast, directly over Esquimalt/Victoria, by the most probable prevailing winds.

Prompt fatalities have not been considered; instead casualties calculated here would take the form of latent cancer fatalities and genetic defects. The greatest contamination would occur nearest the accident site, although both air and ground contamination would remain well above the NRC limits up to  50  kilometers from the accident site and beyond. Casualties would be concentrated within  5  kilometers of the accident, but could extend out to several tens of kilometers from the accident site. Under unfavorable meteorological conditions the effects of such an accident could be experienced as far away as Vancouver.

The second accident scenario analyzed is a hypothetical nuclear-reactor accident aboard a ship berthed at Esquimalt. The core inventory of a  100  megawatt (thermal) naval propulsion reactor fueled by highly enriched uranium metal is derived from calculations on research reactor fuel performed with the ORIGEN computer code. Release fractions consistent with existing accident histories and radionuclide properties are assumed, and consequent releases to the atmosphere are calculated for  15  radionuclides comprising an estimated  94.8  to  98.1  percent of the projected health detriment for three exposure pathways (cloudshine, inhalation exposure and groundshine). Ingestion and resuspension pathways are ignored under the assumption of early evacuation and decontamination.

Calculated downwind air concentrations of the radionuclides following a four hour propulsion reactor accident, as well as ground deposition, exceed the aforementioned federal U.S. NRC limits by hundreds to thousands of times. Prompt casualties are possible close to the accident but have not been considered here.

The high annual casualties from continued long-term habitation of the city indicate the need for decontamination prior to rehabitation.

Although SHORT-TERM casualties under the generally conservative assumptions of this analysis are relatively low, both accidents modeled would cause from hundreds to thousands of LONG-TERM casualties unless the contaminated urban areas were both evacuated and decontaminated. Rapid evacuation would appear impossible in the absence of effective emergency response plans (see below). The most significant impact, however, could be economic. U.S. Government studies indicate that decontamination could cost tens of billions of U.S. dollars and take months to complete, during which time the local economy would be largely terminated. These cost estimates omit the on-site costs of clean-up, and they omit "indirect" losses from the termination of local economies and ripple effects on provincial and national economies. The ecological and economic impacts of such an accident on surrounding salt water bodies have not been considered here but could also be significant.

The risk to the Canadian public from these accidents is the product of the CONSEQUENCES and PROBABILITY of the accident. Although the consequences can be estimated under idealized accident conditions as assumed above, the probability of each accident requires information that is not within the public domain. Such information includes the number of nuclear warheads aboard ships in port, the frequency and intensity of shipboard fires, the fire resistance of nuclear warheads, and the accident history and operating characteristics of naval propulsion reactors, In the absence of this is information, the probability of the accidents modeled cannot be calculated, and hence the risk associated with port visits or stationing of nuclear powered submarines cannot be assessed accurately.

Emergency preparedness for a nuclear accident in Canadian ports is inadequate to cope with the scale of possible accidents analyzed here. Civilian regulatory bodies exercise no licensing nor oversight authority over the technical aspects of U.S. military reactors and weapons. Port visits by U.S. nuclear-powered or nuclear-capable vessels are conducted under the U.S. General Statement of Assurances (Appendix I). This document does not mention emergency preparedness, is ambiguous on the issues of liability/compensation in the event of an accident, and omits any consideration of nuclear weapons accidents, even though such accidents have occurred and are featured in U.S. military emergency preparedness plans. Emergency preparedness for nuclear accidents is allocated to Canadian authorities, who have assigned such responsibility to the Department of National Defence (DND). DND emergency preparedness plans are not, however, in the public domain. Without public knowledge of emergency preparedness plans, it is therefore not clear how public participation in a time of actual emergency could be implemented. U.S. studies indicate that emergency preparedness is effective only when specific plans adequate to real emergencies are designed, publicized and exercised periodically.

Publicly-available information on DND emergency procedures suggest:

  1. evacuation zones extend only to  609  meters from the accident site;

  2. Nuclear Emergency Response Teams (NERTs) are responsible for seeking information about the type of hazard and containing any radioactive material released; and

  3. responding to nuclear weapons accidents is not a part of NERT planning.

These emergency procedures are ineffectual in that:

  1. contamination and casualty zones could extend to several kilometers from the accident site, as demonstrated in the present analysis, rendering a  609  meter evacuation zone meaningless;

  2. the U.S. General Statement of Assurances explicitly prohibits the boarding of U.S. military vessels for the purpose of obtaining technical information, and

  3. U.S. Department of Defence directive  5230.16  permits concealing nuclear weapons accidents when they occur.

It is not clear, therefore, how NERTs could identify, let alone contain, nuclear materials released in an accident. Indeed NERTs are not alerted by visits of nuclear-capable vessels, and are not situated in several ports visited by such vessels.

The findings of this report provide a technical basis for seven policy recommendations. The principle recommendation is:

If the Canadian public and government decide nonetheless to proceed with port visits and/or acquisition of nuclear powered submarines, a number of additional recommendations follow. These are:

[ Accident Possibilities at Gentilly-2 ]

[ Findings on CANDU Safety ]

[ Reactor Accidents Sub-Directory ] [ COMPLETE DIRECTORY ]

Since March 27th 1996, there have been over
100,000 outside visitors to the CCNR web site, plus

(counter reset July 2nd 1998 at midnight)