THE NATIONAL ACADEMIES
Advisers to the Nation on Science, Engineering, and Medicine
National Academy of Sciences
National Academy of Engineering
Institute of Medicine
National Research Council
DIVISION ON EARTH AND LIFE STUDIES
Board on Radiation Effects Research
Dr. James M.Smith
Radiation Studies Branch
Centers for Disease Control and Prevention
4770 Buford Highway, NE Mailstop F35 Atlanta, Georgia30341–3742
May 28, 2002
Dear Dr. Smith:
As you know, the Centers for Disease Control and Prevention (CDC) asked the National Research Council’s Committee on an Assessment of CDC’s Radiation Studies from DOE Contractor Sites to evaluate a draft report titled A Risk-based Screening Analysis for Radionuclides Released to the Columbia River from Past Activities at the U.S. Department of Energy Nuclear Weapons Site in Hanford, Washington, prepared by the Risk Assessment Corporation (RAC). That report was submitted to CDC by RAC in partial fulfillment of the conditions set forth in the following statement of work:
The Hanford Environmental Dose Reconstruction (HEDR) Project developed the Columbia River Dosimetry Code (Farris et al., 1994) to calculate radiation doses for hypothetical individual users of the Columbia River at various locations. Initially, the HEDR Project considered all radionuclides released from the Hanford Nuclear Site between 1944 and 1972. Ultimately, doses were calculated for five radionuclides: 24Na, 32P, 65Zn, 76As, and 239Np. The water concentrations of these radionuclides used in the dose calculations were provided by the CHARIMA computer code (Walters et al., 1994).
The five radionuclides listed above were selected by the Technical Steering Panel of the HEDR Project from the initial list of released radionuclides on the basis of a series of scoping or screening calculations (Napier, 1993). The radionuclide exposure pathways considered in the dose calculations were also selected on the basis of scoping calculations. Hoffman et al. (1998) later suggested that 131I, 60Co, and 90Sr should also be considered in a Hanford individual dose
2001 Wisconsin Avenue, NW, Washington, DC 20007 USA 202–334–2232 (telephone) 202–334–1639 (fax) E-mail:national-academies.org
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 1
THE NATIONAL ACADEMIES Advisers to the Nation on Science, Engineering, and Medicine National Academy of Sciences National Academy of Engineering Institute of Medicine National Research Council DIVISION ON EARTH AND LIFE STUDIES Board on Radiation Effects Research Dr. James M.Smith Chief, Radiation Studies Branch Centers for Disease Control and Prevention 4770 Buford Highway, NE Mailstop F35 Atlanta, Georgia30341–3742 May 28, 2002 Dear Dr. Smith: As you know, the Centers for Disease Control and Prevention (CDC) asked the National Research Council’s Committee on an Assessment of CDC’s Radiation Studies from DOE Contractor Sites to evaluate a draft report titled A Risk-based Screening Analysis for Radionuclides Released to the Columbia River from Past Activities at the U.S. Department of Energy Nuclear Weapons Site in Hanford, Washington, prepared by the Risk Assessment Corporation (RAC). That report was submitted to CDC by RAC in partial fulfillment of the conditions set forth in the following statement of work: “Task Description The Hanford Environmental Dose Reconstruction (HEDR) Project developed the Columbia River Dosimetry Code (Farris et al., 1994) to calculate radiation doses for hypothetical individual users of the Columbia River at various locations. Initially, the HEDR Project considered all radionuclides released from the Hanford Nuclear Site between 1944 and 1972. Ultimately, doses were calculated for five radionuclides: 24Na, 32P, 65Zn, 76As, and 239Np. The water concentrations of these radionuclides used in the dose calculations were provided by the CHARIMA computer code (Walters et al., 1994). The five radionuclides listed above were selected by the Technical Steering Panel of the HEDR Project from the initial list of released radionuclides on the basis of a series of scoping or screening calculations (Napier, 1993). The radionuclide exposure pathways considered in the dose calculations were also selected on the basis of scoping calculations. Hoffman et al. (1998) later suggested that 131I, 60Co, and 90Sr should also be considered in a Hanford individual dose 2001 Wisconsin Avenue, NW, Washington, DC 20007 USA 202–334–2232 (telephone) 202–334–1639 (fax) E-mail:national-academies.org
OCR for page 1
assessment (IDA) process for the Columbia River. The objective of the draft RAC report is to perform screening calculations that can be used to evaluate that recommendation. Specifically, RAC was required by CDC to perform the following: Review all the available HEDR Project documents related to the published Columbia River dose calculations and select the best available information related to the quantities of each of the eight radionuclides listed above that were released to the Columbia River between 1944 and 1972. The contractor shall NOT develop any new information on estimates of the radionuclide releases to the Columbia River without the approval of the project officer. Review risk-based screening limits that could be used by CDC as decision criteria for choosing radionuclides for further consideration in the development of an IDA code for the Columbia River pathway. Develop a screening method that accounts for all potential pathways of exposure for each of the eight radionuclides listed above. Organ-specific health risk, not just radiation dose, should be the end point of the screening calculations. All mathematical models and parameter values selected for use in the methodology should be carefully justified and thoroughly referenced. After the screening method developed has been approved by the project officer, perform screening calculations for the eight radionuclides listed above. Screening calculations for a small number of additional radionuclides may be proposed, but no screening calculations will be performed for any additional radionuclides without the concurrence of the project officer. As a result of the screening calculations, formulate for CDC’s consideration recommendations with regard to the inclusion of further radionuclides and pathways in future Columbia River individual dose calculations.” The National Research Council’s committee was asked by CDC to address the following: Has the screening method been carried out appropriately and completely? Are the data sources assessed and used appropriately? Is the environmental transport model presented in the report adequate and applicable for screening calculations for the Columbia River situation? Are the exposure pathways complete and appropriate? Are the screening risk estimates and associated uncertainties applied and presented adequately? Are the discussion and conclusions of the results complete and reasonable? At the initial meeting of the National Research Council’s committee in Washington, DC, on January 14, 2002, representatives of RAC—John Till, Helen Grogan, Edward Liebow, and Arthur Rood—were present. They enlarged on the draft report and the methods they used in arriving at the conclusions set forth in the report, and they responded to questions raised by the committee. Also present at the meeting were Charles Miller, chief of the Environmental
OCR for page 1
Dosimetry Section of the Radiation Studies Branch, and Judy Qualters, acting chief of the Risk Analysis and Communication Section. Summary of the Committee’s Letter Report: Before addressing the six questions enumerated above, we would like to set out briefly the context in which our review must be seen and our general impression of the report. Our review of a report like the one before us cannot be complete in the strictest sense. The committee has neither the resources nor the time to reproduce every element in a report that could have been years in the making. Accordingly, it must focus its attention on matters that are central to the reliability of the report, such as the adequacy and appropriateness of the study design, the completeness and accuracy of the data under analysis (including assurance of their quality and accuracy), and the validity of the conclusions reached. In general, the committee has not focused its attention on the details, the “nitty-gritty”, of the implementation of the study design, although such details are just as important as features of the report previously enumerated; and indeed, in this case the committee, perhaps serendipitously, located serious errors in some of this “nitty-gritty” work. A further consequence of the examination by the committee of some of the “nitty-gritty” work was the location also of many minor, often trivial, errors. The committee lists every error located, even trivia, for two reasons—first, several of the mistakes, despite being trivial, illustrate the larger theme of lack of quality control (see Appendix B); and second, to provide complete documentation of the committee’s findings, particularly because there are instances where the committee does not have the resources to determine the importance of the error documented. Careful scrutiny of implementation is not only much more time-consuming but demands access to information that the committee rarely had at its disposal; and even in this particular case, the committee has not thoroughly reviewed the implementation. That limitation could be somewhat alleviated if CDC routinely archived, in electronic form, all the data and computer programs used in the generation of a report. The ready availability of such information not only would facilitate the committee’s work but could be used by others interested in the report. The committee strongly recommends that CDC establish such archives. Overall, the draft Columbia River report appears to be a well-written account and to describe a method that is adequate and has plausible results. To its credit, RAC has done substantially more work than was required of it by the statement of work. However, when the committee sought to examine the basic data used to derive RAC’s findings, it encountered many discrepancies between RAC’s values and designations and those in the HEDR documents, which are the source of the basic information. For example, as a result of an error in the computer code, six of the eight reactors are misidentified and the reactors are placed in the wrong locations, and they start and stop releasing at the wrong times (see Appendix B for details). Similarly, it appears that the reactor data—power, energy production, and water flow rate—were often incorrectly transcribed. The full impact of those errors on the findings set out in the draft report is not clear; the committee could not redo all facets of RAC’s analysis. However, even if the impact is found to be small, the presence of so many errors reflects poor quality control and a failure of oversight in RAC and CDC. This situation urgently needs remedying. As matters stand, the findings lack credibility, and the report is unacceptable. The committee believes that a careful and thorough revision of the report is needed.
OCR for page 1
Specific issues that the committee was asked to address A. Has the screening method been carried out appropriately and completely? RAC has developed a novel screening method that is appropriate for the type of study under review and that could have broader application. It consists of a two-step approach. In the first step, called “initial screening”, conservative assumptions are used for all pathways and radionuclides. The corresponding lifetime risks of cancer morbidity resulting from each pathway and each radionuclide are then estimated and compared with a cutoff value of 10-4. Through that process, radionuclides that give rise to almost no excess risk for any pathway, even under extreme circumstances, are eliminated. That led to the elimination of 51Cr, 122Sb, 45Ca, and 90Y. In the second step, three exposure scenarios were defined to represent the most-exposed river users (native Americans, local residents, and migrant workers). The corresponding lifetime risks of cancer morbidity resulting from each pathway and each radionuclide were then estimated with less conservative parameter values than for the first step and compared with the total risk for each exposure scenario. Any radionuclide for which the risk is less than 1% of the total risk for any exposure scenario is eliminated. That led to the elimination of 60Co. The screening method is appropriate, although the committee believes that the implementation was flawed because of incorrect computer codes (see Appendices A and B). However, the application of the second step of the method implies that all parameter values are realistic, or at least conservative to the same degree. As written on page 10 of the RAC report: “To allow a relative ranking it is important that the parameter values used to characterize the exposure pathways and the radionuclides are selected in a consistent manner to avoid biasing the results. Assigning realistic values is preferred because it is difficult to define parameter values with the same degree of conservatism consistently.” That statement notwithstanding it does not seem to have been followed for the fish pathway. In the RAC report, it is assumed (1) that the radionuclide concentrations in fish are proportional to the concentrations in water, presumably at the points of intake for water consumption; (2) that the nonresident fish, as well as the resident fish, stay in place long enough so that the concentrations in them reach equilibrium; and (3) that there is no delay between fish catch and consumption. As a result, the radionuclide concentrations in nonresident fish, such as salmon, are overestimated to a much greater degree than those in resident fish. The three exposed groups consume resident and nonresident fish in different proportions, so the risks attributable to fish consumption do not present the same degree of conservativeness among the three exposed groups. Also, because some of the salmon are consumed months after being caught, the concentrations of short-lived radionuclides are overestimated to a greater degree than those of long-lived radionuclides. Finally, in Table 14 of the RAC report, some of the parameter values (related to sediment ingestion and dermal contact exposure) are more conservative for the migrant-workers exposure scenario than for the initial screening. That is not consistent with a uniform method.
OCR for page 1
B. Are the data sources assessed and used appropriately? The data on the different nuclides selected for analysis in general were appropriate and properly assessed except that the data on activity in fish were never compared with activity projected by the models and no explanation for the source of some of the important radionuclides—such as 239Np, 69Zn, 79As, and 51Cr—was presented. However, because fish consumption is the most important pathway for most radionuclides, it is critically important that RAC compare the radionuclide concentrations measured in fish with those predicted. Such comparisons could support the validity of the bioconcentration factors given in Table 13 of the RAC report and the validity of the other assumptions used to predict radionuclide concentrations in fish. Since two-thirds of the salmon were smoked, the contribution of short-lived radionuclides was overstated. Similarly, inasmuch as the bones of fresh-caught salmon were not eaten, that exposure pathway was also overestimated. Moreover, Pacific salmon do not eat during their upstream migration, so it seems unlikely that they would reach equilibrium concentrations for all radionuclides. With regard to the radionuclide-release estimates, it would be useful to demonstrate that the primary list of 23 radionuclides is exhaustive, and it would be important to explain why some fission-product release estimates are based on 90Y data. However, the statement of work specified that the contractor was not to modify or redo the source-term data, so it is not clear whether some of the 76As and 65Zn data should have been questioned. As for data sources, page 27 states that “Monthly release quantities to the Columbia River for a subset of the radionuclides examined from all reactors are provided in Heeb and Bates (1994) and compiled in a Microsoft Excel spreadsheet. Unfortunately, release quantities were not segregated by reactor. For the HEDR study, this was not a major issue because the model domain extended ~480 km (~300mi).” That statement is misleading. The Source Term River Release Model (STRRM) used by Heeb and Bates (1994) evaluated monthly releases from individual reactors, taking account of available reactor-specific information. Their document listed only the sum of these releases over all eight reactors. However, that document also made the statement (pages 3–13 to 3–14) about the input and output files of STRRM that “The size of the files makes it impractical to include printouts in this document. Readers interested in the information contained in either the input or output files may obtain electronic copies in ASCII format from: [the technical steering panel].” In view of the desire to have reactor-specific information, the RAC should have located and used complete electronic copies of the output files. C. Is the environmental transport model presented in the report adequate and applicable for screening calculations for the Columbia River situation? The environmental transport model used to predict the radionuclide concentrations in water and sediments might be adequate, but it needs further justification with regard to its treatment of sediment transport. The committee does not believe that the implementation of the model has been adequately verified, and it identified an error in the implementation (in the
OCR for page 1
function “kernal”) that probably invalidates both the calibration and the results in the draft report. It would have been useful to indicate more clearly the justification for developing such a model, which does not appear to be superior to the CHARIMA model used by HEDR. D. Are the exposure pathways complete and appropriate? We believe that all important pathways have been considered, although many of the exposure parameters used would be inappropriate if the native American exposure scenarios turn out to not have existed. The committee notes that RAC indicated during its briefing that for the Native American Exposure Scenarios, pathways were identified on the basis of “historical evidence suggesting the pathway represents an opportunity for actual (rather than hypothetical) human exposures” (Presentation to committee, January 14, 2002, by Edward Liebow, Environmental Health and Social Policy Center, Seattle, WA). Other sources indicate that native Americans were removed from the Hanford site and that access to the river was prevented (see http://www.hanford.gov/doe/culres/native.html). One source cited by RAC for many of the parameters for these scenarios (Harris and Harper, 1997) stated that “Because our purpose was to reflect actual Columbia River usage as it recently existed and will exist again if access is regained,…” (page 792), suggesting that the exposures described might not have existed during the period covered by the RAC report. In view of such information, RAC or CDC should carefully examine the relevance of the native American scenarios examined in the report. Furthermore, how the pathways have been analyzed might not be wholly appropriate. The fish ingestion pathway in particular seems to have so many conservative aspects that the results might not be credible, and the isotopes screened in and out might change if a less extreme set of assumptions is used. In addition, the fish ingestion pathway analysis indicates that about 90% of the risk arises this way; actual data on concentrations of radionuclides in fish, which exist, should have been used to validate, or at least benchmark, the model instead of depending entirely on analysis to determine the major effect factor. To go from water-sample analysis through the food chain into fish and into risk when the concentration in the water varies from place to place and time to time and when what the fish fed on and where is unknown yields a less than credible result that merits a “reality check” against the available data. E. Are the screening risk estimates and associated uncertainties applied and presented adequately? Although the data indicate that about two-thirds of the fish consumed by the native Americans were eaten dried, this “holdup” factor was not incorporated into any of the risk screening estimates. The failure to incorporate it would tend to make all estimates too high, which is not a serious problem for the first stage screening, given that conservative estimates were desired in any case. But it would have a differential effect among the radionuclides. Including a holdup factor would reduce short-lived radionuclides (for example, 67Zn and 76As) much more than long-lived ones. That might change the relative rankings of radionuclides substantially in the second-phase screening. It is therefore recommended that a holdup factor be incorporated into at least the second-phase risk estimates.
OCR for page 1
On the whole, the screening risk estimates were applied and presented adequately. However, it would have been helpful to present more information on the locations that are considered for the three exposure scenarios (in this regard, adequate maps are needed) and on the description of the risk that is considered. For example, it is not clear whether the incidence of lung cancer or leukemia is assumed to present the same risk as the incidence of thyroid or skin cancer and whether age dependence is taken into account. The uncertainties appear to have been estimated in an inadequate manner. Because the fish pathway is the most important for most radionuclides, special attention should have been given to it. It is likely that the most important uncertainties are attached to the selection of the bioconcentration factors and to the validity of the assumptions used to estimate the radionuclide concentrations in the two types of fish. Neither of those is taken into account in the uncertainty analysis presented in the draft RAC report. Finally, to aid the reader, the discussion of assumptions and uncertainties should be brought together and summarized in one place at the end of the report and in the executive summary, the latter in terms understandable by the general public. F. Are the discussion and conclusions of the results complete and reasonable? In the original HEDR analysis, 19 radionuclides—including 65Zn, 69Zn, and 69mZn—were screened (Napier, 1993). That screening process revealed 65Zn to be among the five radionuclides with the highest potential risk. In the later risk-based screening conducted by RAC, the same 19 radionuclides and four others were screened. As a result of the phase 1 and phase 2 screening processes, the highest 18 radionuclides were ranked by their percentage contribution to the total risk for three scenarios: local residents, native Americans, and migrant workers. 65Zn ranked either fifth or sixth in its contribution to total risk. The surprising result was that 69Zn ranked second or third, accounting for 19–23% of the total risk posed by all 18 radionuclides that passed the phase 1 screening. That is a very high rank for a radionuclide that did not make the top five in the HEDR dose calculations. This unexpected result and the one for 95Zr provide an excellent opportunity for useful expansions of the discussion section, which is very brief. The reader needs to know something about why those results occurred in this procedure and not in the HEDR analyses and to see information on the factors that led specifically to the result that 69Zn was the second-highest contributor to the calculated total risk. Have consistent assumptions been applied for all the radionuclides examined? The early section of the RAC report needs to provide a context for its dose and risk estimates so that those who are not radiation specialists have a frame of reference to understand the results. A suggested way to do so would be to compare the Columbia River “effective dose” estimates with the levels of natural background radiation that the population would receive during the same period (or during their lifetime) and to compare the estimated risks from the two sources. Another way that RAC could provide more realism in discussing the risks would be to perform calculations of parallel scenarios that used approximate central estimates of the population exposures based on the median or modal values for the various parameters that went into its models and present them as an alternative, pointing out the greater likelihood that most people’s exposures would have been closer to the central estimates than to the upper bound estimates.
OCR for page 1
In summary, the draft RAC report appears to be a well-written account and to describe a method that is fairly adequate. However, the implementation of that method was flawed in several ways: (1) There are a number of discrepancies between RAC’s values and designations and those in the HEDR documents, which are the source of the basic information. (2) The environmental-transport model does not seem to have been adequately verified. (3) The treatment of the most important exposure pathway for most radionuclides, the consumption of contaminated fish, is weak because of some simplistic assumptions, the lack of comparison between predicted and measured radionuclide concentrations in fish, and degrees of conservatism that differ among radionuclides and among exposed population groups. (4) The treatment of the uncertainties is somewhat arbitrary. For those reasons, the findings in the draft RAC report are suspect, and the committee believes that a careful and thorough revision of the report is needed. If you desire further elaboration on the comments above or in the accompanying appendixes, please do not hesitate to call or write either Dr. Isaf Al-Nabulsi or me. Sincerely yours, William J.Schull Chairman