I. INTRODUCTION
Since the 1940s, the Atomic Energy Commission and its successor organizations (now the U. S. Department of Energy, DOE) have operated facilities to produce materials for nuclear weapons. In recent years, persons living around those facilities have become increasingly concerned that radioactive emissions from the facilities have affected their health. Under a memorandum of understanding concluded with DOE in December 1990, the U. S. Department of Health and Human Services, through its Centers for Disease Control and Prevention (CDC), has begun a series of studies to assess the possible health consequences of exposure to emissions of radioactive materials from DOE-managed nuclear facilities throughout the United States.
At CDC's request, the Board on Radiation Effects Research in the National Research Council (NRC) Commission on Life Sciences has organized the Committee on an Assessment of CDC Radiation Studies to provide scientific advice to CDC's Center for Environmental Health and Injury Control and to evaluate the quality and completeness of CDC's assessments. The NRC committee's charge is as follows:
-
Review and comment on the design, methods, analysis, statistical reliability, and scientific interpretation of dose-reconstruction studies and related epidemiologic follow-up studies.
-
Recommend ways to strengthen study protocols and analyses so that the scientific validity of the study results can be ensured.
The first specific task requested by CDC was a review of draft reports prepared by the Radiological Assessments Corporation (RAC) pertaining to its efforts to reconstruct environmental doses of radionuclides in the vicinity of the Fernald, Ohio, nuclear facility–the Feed Materials Production Center (FMPC) ( Figure 1 ). The drafts reported on were radionuclide source terms and uncertainties in 1960-1962, and they constitute two of seven tasks in the Fernald dose-reconstruction project. The NRC committee's review of those draft reports was published in 1992 (National Research Council, 1992).
A. FERNALD NUCLEAR FACILITY
In its 1992 report, the NRC committee described the operations of FMPC which was, until 1989, a large-scale government-owned, contractor-operated facility that produced uranium metal products used as feed materials in DOE defense programs. Construction began May 16, 1951, and FMPC produced its first uranium derby * Oct. 11, 1951. The production plants and administrative buildings today occupy 136 acres on a 1,050-acre site. Producing the derbies, ingots, billets, and fuel cores required a series of chemical and metallurgic conversions that occurred in nine specialized plants on the site. These are described in the 1992 NRC report.
After almost 38 years of production of low-enriched uranium fuels (containing 2% or less 235U), production was suspended in July 1989. In 1990, FMPC was changed from a production facility to an environmental restoration site; in October 1990, DOE transferred management responsibility for FMPC from its defense-programs organization to its Office of Environmental Restoration and Waste Management. Since Jan. 1, 1986, the Westinghouse Materials Company of Ohio has operated the facility as prime contract for DOE; before that, National Lead of Ohio was the prime contractor. Westinghouse management changed the name of the Fernald subsidiary in 1991 to Westinghouse Environmental Management Company of Ohio to reflect the restoration mission at the Fernald site. On Aug. 23, 1991, DOE officially changed the name of the facility to the Fernald Environmental Management Project. This report refers to the facility as FMPC.
The Fernald facility is about 27 km northwest of downtown Cincinnati in sparsely populated farm country. The site is about 1 km west and north of the Great Miami River, which runs in a southerly direction ( Figure 1 ). Its site includes a small creek called Paddy's
* A derby is an ingot in the shape of a derby hat. |
Run, which runs through the facility near its western border and flows into the Great Miami River. An FMPC effluent discharge line empties into the Great Miami River. The river's average flow in 1990 was 164 m3/s at a point 16 km upstream of the FMPC effluent line. About 39 km downstream of FMPC, the Great Miami River flows into the Ohio River.
During most of the past 40 years, uranium has been discharged each year in liquids into the Great Miami River and in dust and gases into the air. Although FMPC no longer produces uranium metal, radioactive materials from Fernald and other DOE sites are stored there. Since the early 1970s FMPC has been the federal government's storage site for thorium; in 1990, about 1,100 metric tons of thorium compounds were stored there. There also are four K-65 silos, large partially submerged concrete storage tanks in a waste disposal area. These tanks contain radionuclides in the 238U decay chain, and hence are a source of release of 222Rn.
B. METHODS USED IN THE RADIOLOGICAL ASSESSMENTS CORPORATION REPORT
The purpose of RAC's Fernald Dosimetry Reconstruction Project is to estimate the radiation doses received by members of the public who lived near FMPC between 1951 and 1989. The goal of Task 4 of the project, Environmental Pathways –Models and Validation, is to develop a methodology that can be used to translate the release estimates developed in Tasks 2 and 3 (relating to radionuclide source terms and uncertainties in 1960-1962) into concentrations of radioactive materials in the residential environment. While direct measurements would have been desirable, only gummed-film data were available; therefore, the authors explain mathematical models that describe the movement of radionuclides in the environment around FMPC, using site-specific data whenever possible. To address the concerns of scientists, government officials, and the public, RAC has emphasized simplicity, flexibility, and physical plausibility in its models. RAC also has conducted uncertainty analyses, and attempted to validate its models and provide for a reasonable cost of implementation. The models have been tested through a comparison of uranium release
rates estimated for the years 1960 through 1962 and the concentrations based on measurements in air, in soil and on gummed film.
The RAC model underpredicted the amount of uranium deposited at distances away from the facility as measured on gummed film for the same locations; but by a reasonably acceptable factor of approximately 2. Interpreting the measurements of deposition with gummed film can be difficult because the fractional retention by gummed film under dry-and wet-deposition conditions can vary markedly. The report should note whether these historical measurements were accepted as presented or whether appropriate calibration factors were used. In other words, the uncertainty in the measurement of uranium deposition with gummed-film collectors should be noted. The models used for radon and radon progeny released from the K-65 silos and the validation calculations underpredicted measurements by about a factor of 2, as did comparisons of calculated dose rates for gamma radiation from the silos with measurements of the radiation field. The authors of the Task 4 report suggest that the source of the discrepancy could be mischaracterization of the waste in the silos with respect to the quantity of 226Ra. Estimates of the source term were revised in November 1993, and the new figures supplement some of the model's information. The results of the final report are expected to provide a better match to measured data.
In the RAC study, an assessment domain is defined as the residential areas within an circle 8-km (5 mi) radius around the FMPC production site ( Figure 1 ). It is estimated that more than half of the particulate material released from FMPC to the atmosphere was deposited in this area and that the densities of the radioactive materials deposited at the outer boundary of the assessment domain were a factor of 10 lower than were densities at the site boundary. This pattern of deposition resulted from the relatively dense nature and large size of the particles containing the uranium that were released.
The study's environmental models focus on five environmental pathways through which the released radioactive materials move
-
leaks from storage containers
-
releases to the air
-
air-to-ground deposition
-
ground-to-air resuspension
-
runoff and leaching of radionuclides from soil to surface water and groundwater
Human exposure to the radioactive materials comes from contact with several media and by several methods:
-
direct inhalation
-
direct exposure to γ radiation
-
contamination of food crops and animal forage
-
direct ingestion of contaminated soil
-
ingestion from the drinking water supply
-
irrigation of crops
-
consumption of fish
-
direct γ radiation from exposure to contaminated water
The RAC model estimates that atmospheric pathways dominate, and the transport of uranium particles is believed to have been strongly influenced by the particle-size distribution. The median particle diameters for the outlet ducts of the dust collectors were estimated to be in the aerodynamic range of 5-9 μm. However, the model estimates that about half of the uranium released to the atmosphere emerged from scrubbers in the form of large droplets that rapidly agglomerated into large particles with diameters of 30-150 μm. The RAC authors considered three main issues in their choice of an air dispersion model: wind trajectories, building wake effects, and whether the model could be validated against air-monitoring data. They developed an approach to quantify each source of uncertainty in terms of a probability distribution.
To validate air transport calculations, RAC relied primarily on monthly average uranium concentrations measured in the air at four perimeter stations at the corners of the production area. They also used monthly averages of the amount of uranium deposited on
sheets of gummed film at 22 locations within the assessment domain. They estimated a total deposition of 2 × 105 to 9 × 105 kg uranium from measurements of uranium in soil.
Short-term events, defined as episodic releases (which increased the total uranium release rate by a factor of at least 10 for a period of fewer than 10 days) were given special attention. However, the time increment over which the average release rate is calculated should be specified (hourly, daily, weekly). In the interest of thoroughness and completeness, it is important to treat episodic releases carefully and in a consistent and sound manner. Other sources of exposure from FMPC releases that were considered included the inhalation of radioactive progeny of 222Rn produced by the decay of 226Ra a component of waste materials stored in the K-65 silos, and direct γ-radiation exposure from the silos.
Three model approaches, a simple monthly dilution (MD) model, the National Council on Radiological Protection and Measurements Screening Model (National Council on Radiological Protection and Measurements, 1989), and the GENII Model (Napier et al., 1988), were used to conclude that the major pathways that contribute to radiation dose from uranium in the surface water (originating from the main effluent line to the Great Miami River and from runoff and spills entering Paddy' s Run) were those involving ingestion of drinking water and consumption of local fish. Dose models for exposure to radon and radon progeny were based on those in NCRP Report 78 (National Council on Radiation Protection and Measurements, 1984b). For all other radionuclides, the internal dosimetry was estimated based on the methodology of ICRP Publication 30 (International Commission on Radiological Protection, 1979). Finally, four demographic models were developed to obtain estimates for the location, number, and distribution of age and sex of residents of the assessment domain between 1951 and 1988.
C. COMMITTEE ACTIVITIES
This report describes how the NRC committee performed its task and contains its preliminary analysis, findings, and conclusions. As for this committee's first report (National
Research Council, 1992), the reader is reminded that the RAC's draft report pertains to a portion of the dose assessment FMPC (Task 4) and does not contain any conclusive results. The 1992 NRC report identifies a series of criteria for an ideal dose reconstruction to provide a flame of reference for evaluating the RAC reports. In January 1992, the NRC committee received extensive briefings from RAC personnel on the dose reconstruction study at Fernald. In March 1992, the committee visited Fernald, observed the operation, and was briefed by FMPC (now FEMP) personnel about past and current operations. After the visit, the committee met again with RAC staff and conducted a public meeting to allow citizens to discuss the dose reconstruction activities.
In March 1993, the committee received the draft report, Task 4. EnvironmentalPathways–Models and Validation, a 44-page document supplemented with 23 appendixes that contain an additional 420 pages (Radiological Assessments Corporation, 1993). The committee met in July 1993, and in February and March 1994, to evaluate the methods proposed in Task 4 in the context of the criteria for an ideal dose reconstruction and to write this report. In its review, the committee focused on the source term, atmospheric transport, surface and groundwater exposure pathways, and dose assessment. Special attention was given to an examination of the models used and to RAC 's validation methods and uncertainty estimates. General comments about these issues appear in the text of this report; and comments about specific portions of the RAC report appear in the Appendix.
The Task 4 report is designed to bring together diverse resources to provide a general perspective on dose reconstruction. Overall, the committee judged the document to be thorough in its scope and approach and of high scientific merit. The authors did an excellent job of bringing together and interpreting the technical literature on dose assessment. The presentation is clear, authoritative, and comprehensive. Since the development of methodologies that can be used to translate release estimates into concentrations of radioactivity in the environment where people lived is one of six projected tasks in the Fernald Dosimetry Reconstruction Project, it is appropriate that there is some bridging between this document (on Task 4) and the draft source term document, on Task