THE NATIONAL ACADEMIES
Advisers to the Nation on Science, Engineering, and Medicine
National Acacemy of Sciences
National Academy of Engineering
Institute of Medicine
National Research Council
COMMISSION ON LIFE SCIENCES
Board on Radiation Effects Research
May 11, 2000
Dr. James M. Smith
Radiation Studies Branch
Centers for Disease Control and Prevention 4770 Buford Highway, NE Mailstop F35 Atlanta, GA 30341-3742
Dear Dr. Smith:
As you are aware, the Centers for Disease Control and Prevention(CDC) asked the National Research Council's Committee on the Assessmentof CDC's Radiation Studies to review and comment on the Draft FinalReport titled “Source Term Calculation and Ingestion Pathway DataRetrieval Evaluation of Materials Released from the Savannah RiverSite”. That document represents the findings of phase II of the contemplatedfive phases involved in the Savannah River Site Environmental DoseReconstruction Project. Aside from its general assessment of thedocument, the committee was asked to address the following six questions:
“Can you comprehend the methods used by reading the report? Can youreproduce the estimates with the information provided?”
“Were the screening methods used by Risk Assessments Corporation forradionuclides and chemicals sufficient?”
“Were the appropriate radionuclides and chemicals selected for sourceterm estimation?”
“Were the methods used to estimate the source term for each radionuclideand chemical sufficient?”
“Are the environmental monitoring data comprehensive and substantivefor uses in future phases of the dose reconstruction?”
2101 Constitution Avenue, NW, Washington, DC 20418 USA
202-334-2232 (telephone) 202-334-1639 (fax) 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 Acacemy of Sciences National Academy of Engineering Institute of Medicine National Research Council COMMISSION ON LIFE SCIENCES Board on Radiation Effects Research May 11, 2000 Dr. James M. Smith Chief , Radiation Studies Branch Centers for Disease Control and Prevention 4770 Buford Highway, NE Mailstop F35 Atlanta, GA 30341-3742 Dear Dr. Smith: As you are aware, the Centers for Disease Control and Prevention(CDC) asked the National Research Council's Committee on the Assessmentof CDC's Radiation Studies to review and comment on the Draft FinalReport titled “Source Term Calculation and Ingestion Pathway DataRetrieval Evaluation of Materials Released from the Savannah RiverSite”. That document represents the findings of phase II of the contemplatedfive phases involved in the Savannah River Site Environmental DoseReconstruction Project. Aside from its general assessment of thedocument, the committee was asked to address the following six questions: “Can you comprehend the methods used by reading the report? Can youreproduce the estimates with the information provided?” “Were the screening methods used by Risk Assessments Corporation forradionuclides and chemicals sufficient?” “Were the appropriate radionuclides and chemicals selected for sourceterm estimation?” “Were the methods used to estimate the source term for each radionuclideand chemical sufficient?” “Are the environmental monitoring data comprehensive and substantivefor uses in future phases of the dose reconstruction?” 2101 Constitution Avenue, NW, Washington, DC 20418 USA 202-334-2232 (telephone) 202-334-1639 (fax) national-academies.org
OCR for page 1
“Do the Phase II results provide adequate information to establisha risk-based screening level for radionuclides?” At a meeting with representatives of CDC and the Risk AssessmentsCorporation (RAC) in Augusta, Georgia, after a visit to the SavannahRiver Site (SRS) on the afternoon of March 30, 1999, the committeewas further charged with reviewing and commenting on “The Community Executive Summary prepared by CDC and RAC for publicinformational purposes”. “Ease of use and utility of the electronic version of the Draft FinalReport”. In the paragraphs to follow, the committee sets out its overall viewof the strengths and weaknesses of the document and then addressesthe eight points enumerated above. General assessment: It would be difficult to read the Draft Final Report and be unimpressedby the care with which RAC has surveyed, collated, and interpretedthe historical information bearing on the radioactive and chemicalreleases from SRS. Clearly, the RAC staff has done an exemplary job.Notwithstanding that, the report is difficult, indeed tedious, toread and is more repetitive than seems necessary. The committee believesthat the report could be improved by more rigorous editing and theuse of internal cross-referencing to avoid some of the repetition. The report refers to doses throughout—in text, figures, and tables—but, rarely specifies their type (such as effective, equivalent,and committed). An effort should be made to ensure that the typeof dose is always clearly specified; otherwise, information takenout of context can be easily misunderstood. That applies to all figures,tables, and text. The committee believes that the report should devote more effortto giving overviews of measured levels and tying together variouscollections of measurement data. For example, for I-131 in milk,chapter 10 presents no graph or table of variations in average levelsby year, but presents only levels measured in one specific year;therefore, the reader cannot form an overall picture of the levels.It would also be helpful if a graph showed whether estimated amountsof I-131 released at SRS in various years correlated with the amountsfound in milk in those years, with appropriate adjustments for contributionsfrom global fallout or fallout from Nevada Test Site activities.
OCR for page 1
Specific issues that the committee was asked to address: Can you comprehend the methods used by reading the report? Can youreproduce these estimates with the information provided? The investigators compiled the data on radionuclides produced andused in the many facility processes, on when and how releases occurred,and on estimated release quantities and release pathways. These includeI-131, H-3, Ar-41, I-129, Cs-137, Sr-90, U isotopes, and Pu-238,239, and 240. They reviewed the measurement procedures for the variousradionuclides and adjusted the reported results for sampling linelosses, chemical form, and so on. They were then in a position toset priorities for targeting radionuclides. They used a publishedmethod, developed by the National Council on Radiation Protectionand Measurements (NCRP, 1996), to assign priorities to various radionuclidesfor further investigation. However, for casual readers who do notdeal with screening often, some of the text in chapter 3 is confusing.It refers to phase 1 screening, level 1 screening, and “our nextlevel of screening”. NCRP (1996), the underlying basis for thesecalculations, refers to possible screening levels I, II, and III.The text needs careful review to ensure that readers can easily understandwhat has been done in each instance and why. For chemicals, the process of evaluation is less transparent becausebefore the late 1970s there were few measurements on chemical spillsor releases. Most of the estimation process therefore was based onprocurement records of quantities used and knowledge of processesand procedures to derive rough estimates of likely quantities releasedto air or water. The investigators assigned toxicity values to thechemicals with respect to carcinogenicity and developmental and reproductiveeffects; the toxicity values were based on published compilationsby the Environmental Protection Agency (EPA), Agency for Toxic Substanceand Disease Registry (ATSDR), American Conference of GovernmentalIndustrial Hygienists (ACGIH), Occupational Safety and Health Administration(OSHA), National Institute of Occupational Safety and Health (NIOSH),International Agency for Research on Cancer (IARC), and the NationalToxicology Program (NTP). For noncarcinogenic hazards, they usedthe reference-dose or reference-concentration transforms to estimatetoxicity in humans from the rodent data. For carcinogenesis, theyused the EPA slope factors, which are generally based on a linearextrapolation from the upper confidence bound on the data-based estimateof risk. The committee notes, however, that for the chemical sourceterms the results are scattered throughout chapters 17 and 18 andthat as a result of this scattering it is difficult to form a unifiedpicture of the relative magnitude of releases of various chemicals. In summary, the methods used for estimating the magnitude of releasesof radionuclides from SRS are reasonably clear, as are the methodsfor assigning priorities to specific radionuclides for further study.For chemicals, the procedures for estimatingquantities of releasesare less well-chronicled, but this is largely because the data thatwould be required for doing so are unavailable. The assignment oftoxicities to each chemical is not well delineated, because it wasa complex procedure that relied on reference sources that used differentdata, criteria, and designations.
OCR for page 1
Were the screening methods used by Risk Assessments Corporation forradionuclides and chemicals sufficient? Radionuclide releases The comments, which follow, concern the initial list of radionuclidesthat were considered, the deficiencies of the NCRP method for screening,and the manner in which the NCRP method was applied. The NCRP method should be described briefly. In particular, it wouldbe helpful to indicate what the end points are: effective dose fora hypothetical member of a critical group in a hypothetical siteand the dose being calculated for a time period of 50 years followingthe release. It should also be noted that releases to groundwaterare not considered. It is important to point out that other methods could have led todifferent results. For example, if doses to future generations areconsidered, the relative role of long-lived radionuclides will bemore important. That is particularly true for radionuclides thatare mobile in the environment, such as C-14. It is important to note that the NCRP method is conservative andthat the degree of conservatism is different between pathways. Thiscan lead to erroneous results regarding the relative importance ofthe radionuclides. Radionuclides near the cutoff should be reevaluatedagainst best estimates. The method used to select the radionuclides to be considered infuture phases of dose reconstruction is not clear. In a preliminaryscreening, the relative importance of all pathways of airborne andwaterborne releases for dose is determined for all radionuclides.The results are illustrated in figures 3-1 and 3-2. However, itwould have been useful to include in those figures all the radionuclidesthat were considered in the screening. For convenience, it wouldalso have been helpful to rank the radionuclides in decreasing orderof importance in tables 3A-2 and 3A-3. Parenthetically, why are thereseparate entries for the various isotopes of ruthenium (Ru-103, Ru-106,and Ru-103,106) in figure 3.1 and table 3A-2? In a later screening,individual pathways are considered; the results of the later screeningshould presumably be used to select the radionuclides to be considered.However, the selection seems to have been made arbitrarily. If thiswas not so, then the text should set out the rationale underlyingthe selection clearly. For completeness in the estimation of releases, the committee believesthat C-14 should be included. C-14 arises from reactor emissionsfrom the same air activation that gives rise to Ar-41 and from plant-separationemissions from activation of impurities in the fuel, target, andcladding materials. Similarly, on the basis of specific activity,it is surprising that Pu-240 is estimated but Pu-238 is not.
OCR for page 1
Table 3A.1 list radionuclides associated with fuel, releases, orwaste at SRS. This list presumably was used to do the screening.Uranium is listed, but it is not clear whether the SRS release estimatesinclude releases from the seven coal-fired power plants that werein operation on the site. In any case, releases from coal-fired powerplants include a series of naturally occurring radionuclides, notablyRa-226, Pb-210, and Po-210, which should have been considered. The most notable discrepancies between the published statementsof the SRS and RAC involve Ar-41, plutonium, and iodine. These discrepanciesare attributed to differences in the estimation procedures, suchas the failure to take into account line losses or, in the instanceof iodine, failure to incorporate releases of organically bound iodine.In the case of Ar-41, the discrepancy might be resolved by inquiringinto changes in “neutron economy” or “reduced neutron leakage” related to changes in fuel loading and fuel managementduring the periods involved. Such information would be found notin the monitoring data, but in the reactor technical-support data.As to the plutonium-release discrepancy, one can estimate the amountof plutonium that might have been deposited in the sample line fromthe existing sampler flow data and assumed deposition factors. Alternatively,the line could be sampled now, and the committee urges that thisbe done. Perhaps, easier still, in as much as the plutonium is soold, surveying the outside of the line for growth in americium mightgive the answer; it would require little more than about 30 minutesof hand calculation to determine the feasibility of this approach.It would also be useful either to examine the existing measurementsof plutonium in soil at SRS or to initiate a sampling program toestablish an inventory of material that has been released. That techniquehas worked extremely well at Rocky Flats, at Hanford, and elsewhere. Large amounts of U-233 were produced at SRS for the Navy program,but this is not mentioned. If the irradiated thorium was processedat SRS, it should still be there. If it was processed elsewhere,it nonetheless was irradiated at SRS. The committee finds the explanation of “f” in the formula at the top of page 3.5 difficult to understand. Whydoes the concentration decrease as the wind frequency increases?The total quantity of material going over a specific spot increasesas wind frequency increases, but how does frequency affect concentration? Page 4.1-2, first paragraph. Is this correct? Is not the transformationrate between hydrogen, deuterium and tritium great enough to precludewhat is discussed here? Page 4.2-33, last paragraph. Same as comment regarding plutoniumabove. Some of the beta-gamma emitters have long enough half- livesthat the same solution might be available. Table 3A. 1 indicates the radionuclides for which SRS release estimatesare available. It would have been helpful to provide the numericalvalues of these releases. Also, the differences between U, U-233,U-235, U-238, and Un-ID-alpha should be described.
OCR for page 1
The committee believes that the use of the NCRP method to screenthe radionuclides to be considered is appropriate. However, the committeefinds the discussion of the screening method on pages 3-3 and 3-4troubling, and notes that if consequences are overestimated nonuniformly,as was apparently done, the answers canbe incorrect. Chemical releases Estimation of the chemical releases from SRS appears more conjecturalthan that of the radioactive releases. Monitoring data are skimpy,and the releases, have of necessity, been estimated more indirectlyfrom purchase orders and knowledge of plant processes. We recognizethat out of necessity RAC was obliged to adopt this indirect approach,and we commend RAC for its effort. However, we fail to understandsome of the methods. For example, if the national average was notused for sulfur in coal, why should it be used for mercury in coal? In summary, the committee cautions RAC and CDC that careful attentionbe given to evaluating the chemical source terms. Oversight and adviceshould be sought to make certain that the sources of chemicals, theweighting of their potential toxicity, and their adverse health effectsare assessed with the care and deliberation that RAC has demonstratedfor radionuclides. Were the appropriate radionuclides and chemicals selected for sourceterm estimation? The RAC team used an orderly process to determine the most importantradionuclides to study in connection with releases to air and surfacewater. In phase I, it used the NCRP (1996) screening model to preparea long list of radionuclides for possible inclusion in its study.In phase II, RAC used the NCRP screening approach to study not onlythe radionuclides released, but also their importance in differentpathways. Results of the analyses (given in tables 3-2 and 3-3) showthe calculated relative importance of various radionuclides in variousexposure pathways. After those analyses, many (but not all) high-ranking radionuclideswere selected for additional analyses of historical releases. Theseanalyses are given in chapter 4 for airborne releases of tritium,radioiodine, beta- and gamma-emitting particles, activation products,and alpha-emitting radionuclides. Chapter 5 provides correspondinglydetailed information on radionuclides released to surface water,including tritium, Cs-137, Sr-90, I-131, activation products, andalpha-emitters. The committee observes that, on the basis of results presented inthe report, RAC's estimates of radionuclide releases are higher thanhistorical estimates by a factor of 2-3 for I-131 and by a factorof 4 for plutonium (Pu-239, Pu-240, and possibly Pu-238). Those changesare not trivial, so the committee sought to assure itself that theyare
OCR for page 1
appropriate. However, given the general levels of agreement seenfor the other emissions, the committee deems it unlikely that otherimportant discrepancies that have not been addressed before willbe discovered. That is particularly true in light of the breadthof historical data that have been identified, used, and made availableon CD-ROM. These remarks notwithstanding, the committee urges thatfurther validation be undertaken before RAC's revised estimates areaccepted. The discussion (see page 4.2-33, lines 21-22) of the measurementsof releases of beta- and gamma-emitting particles to the atmosphere,says that “a cautious approach of multiplying the release estimatesby 4 is recommended.” That is vague. Was the factor of 4 used inthe results presented? Was it used for the alpha-emitting particles?If all cases were treated in the same way, it should be made clear.If they were not, this should be stated explicitly and a reason given. From an operational point of view, it is interesting that most ofthe releases of plutonium, unlike those of uranium, occurred in onlya few years during the history of SRS. This is particularly evidentin figure 4.4-3. Figures 4.4-5 and 4.4-6 give more detail about thearea where these releases occurred. Sampling line losses, a likely outcome given the long sampling linesconnecting the sampling probes and filter connectors, are discussedin the report as are possible impacts of various types of filterpaper and sampling velocities. The results in figure 3.2 accord well with those in figure 9 (p.82) of the phase 1, task 3 report except that the results for Pu-239,240 and Pu-238 both seem to be about 10 times higher in the phase1 draft than in figure 3.2. What is the reason for that? Sr-89 is missing from the list on page 3-6, line 2. Was that an oversightor intentional? If intentional, why? Tables 3-2 and 3-3 show the importance of the radionuclides of interestin air or water. It was instructive to tabulate the screening values(Sv/y) for all important radionuclides and all six pathways. Whenthat was done, the values for alpha-emitters in vegetables seemedhigher than that for Cs-137 in vegetables. In light of the majordifferences in gastrointestinal absorption between Cs-137, Pu, andAm, this result seems questionable. Is it related somehow to the“conservative values” used to produce the results? Or is it related to a misunderstandingof the screening process and how it should be used? The authors areencouraged to discuss their results more fully to eliminate doubtsin the readers' minds. In table 3A-2, the column headed “Percent total dose” adds up to 102.23%. That is curious in that each entry has threeto five significant figures, suggesting a precision that does notexist. Is there something wrong here and in the other percentagecolumns? Furthermore, summing the effective doses by different pathwaysdoes not usually give the same value as the total-dose entry. Whynot?
OCR for page 1
Sections 4.1, 4.2, and 4.3 each end with a brief summary of importantfindings. Both for consistency and as an aid to readers, a similarsummary should be added to Section 4.4. Because the factors of conservatism used in the NCRP screening mightnot be uniform, consideration should be given to a local best-estimateiteration of materials close to the cutoff. Only by doing that canone assume that the correct radionuclides are selected for use inphase III. Those to be reevaluated should be selected on the basisof the range of conservatism. For example, if some were 20 timesmore conservative than others, all those within a factor of 20 ofthe cutoff should be selected. That would entail some iteration ofthe phase II steps. In defense of its suggestion, the committee notesthe following: What is called “atmospheric concentration” appears to be stack exit with no dilution, and stack flows appearto be too low by factors of 10-50 (see page 3.5). For comparisononly at stack exit that might not matter, but why is it calculatedif the activities released will lead to the same ranking of the radionuclides?When the value called “atmospheric concentration” is used to calculate doses, the results areoverestimated by a factor of at least 1,000. It is misleading togive such absolute values in the report, for example, in tables 3A-2and 3A-3. Pu-238 is stated to represent 20% of the inhalation dose (page 3.7),and this seems high enough for Pu-238 to make it through first screening. Of the dose from meat, 73% is due to I-131, whereas 62% of the dosefrom milk is due to I-131. These estimates assume that animal thyroidsare consumed, and this might not be a common practice. Iodine charcoal filters are heated by NOx - stack source of NOx (pages 4.2-4.7). Was that considered in chemical releases? Pages 4.2-11 states that the iodine released from processing plantswas 300 times that from reactors, but pages 4.2-12 states at least50. Which is correct? The ratio of elemental iodine to organic iodine seems higher inreconstructed years than in measured years. Is there an explanationfor that? Almost all the iodine dose appears to be due to releases in 3 or4 years in a total of 36 years. Were those few reexamined for bestestimates? After 1968 or so, when carbon filters were used for sampling, alliodine species were collected. How was the elemental fraction versusthe organic fraction determined? To summarize the committee's reactions briefly, we believe that the selected radionuclides areappropriate. However, we view the list as incomplete. In particular,I-129 and Am-241, which rank higher than Pu-238 in the airborne releases,should have
OCR for page 1
been selected as well. That is also true for Cs-137 and other radionuclides,which rank higher than U-235, 238 in the airborne releases; withregard to waterborne discharges, Co-60, I-131, and Zn-65 could havebeen selected as well. Because of the limitations of the NCRP methodology, of the use ofthe initial release estimates in the screening process, and of thedesirability to take into consideration other criteria, the committeesuggests that RAC consider a value more stringent than 1% of thedose as is done in the report. RAC may already have done so, andif this is the case, it would be useful if the report described therationale for rejecting a more stringent criterion. We further recommendthat the relative importance of the naturally-occurring radionuclidesreleased by the coal-fired power plants be assessed. Were the methods used to estimate the source term for each radionuclideand chemical sufficient? RAC located and retrieved 50,000 boxes of archived materials. Analysisof that database is the basis of the delineation of chemical andradionuclide releases that they present in the Draft Final Report.The current accrual rate of boxes was estimated by RAC (at the December3, 1999, meeting) at about 8,000 per year, which suggests that atotal of 50,000 reflects substantial losses. It would be reassuringif RAC could devise an independent way of assessing the completenessof the data or, alternatively, if it could adduce from several internalsets of evidence a consistency that indicates that the records arefairly complete. Given the limitations of the data that were reviewed, the next questionis how adequately the files were searched and relevant data werefound and entered into the database. Quality control in this phaseof the operations included a second reviewer for recently entereddata, but additional attention to the data entered earlier couldincrease confidence in the data-retrieval operation. Chemical processing volumes are proportional to production output,so chemical inventories and releases are likely to have been greaterin the later years of plant operation, for which records are betterpreserved. Thus, the basic data on which source terms were estimatedcould be better for potential chemical releases than for radionuclidereleases, but this requires demonstration. It should be noted, however,that for the chemical source terms the results are scattered throughoutchapters 17 and 18, and it is difficult to form a unified pictureof the relative magnitude of releases of various chemicals. For communication purposes, it would be valuable to create a tablefor airreleases and a similar one for water releases in which thebest estimate of total releases over the 38-year period is givenfor each of the chemicals considered. Such tables should indicatethe plausible range of estimates (such as, the 5th to 95th percentile)when such estimates can be made.
OCR for page 1
Are the environmental monitoring data comprehensive and substantivefor uses in future phases of the dose reconstruction? The RAC effort to compile and assess all useful environmental monitoringdata is described in chapters 8-14 of the Draft Final Report. Thedata included both published summary reports and raw data (read fromhandwritten ledgers preserved on microfilm). The report considersmonitoring of specific nuclides in air, vegetation, agriculturalproducts, milk, drinking water, river water, and fish. SRS began environmental monitoring even before plant operations commencedand has continued throughout. Air and vegetation monitoring tookplace on site at numerous locations, at the plant periphery, andat a ring of locations about 25 miles off site. Monitoring of milkat farms and dairies in the off-site areas for I-131, Sr-90, andCs-137 began from the end of 1956 (for iodine) to 1960 and continuedthroughout the 1980s. Assessment of contamination of water on siteand downstream of SRS also began early in the operations of the facility. Monitoring data do not exist for all nuclides of interest becausemany of the samples contained concentrations below the minimum detectionlimit. The monitoring program tended to focus on gross alpha andbeta emissions, tritium, radioactive iodine, and a number of gamma-rayemitters (such as, Cs-137). The report notes a number of short comingsof the data, including changes in sampling locations over time withoutan explanation of why this occurred, frequency of sampling, the availabilityof only composites over time or space in many cases, changes in thelist of chemical contaminants monitored in the air, and changes inthe scheme of vegetation sampling. In general, there is a lack ofconsistency in the sampling method over time, but this was undoubtedlyinevitable given the rapidity with which monitoring methods and instrumentswere changing during the period of sampling (1950-present). In anumber of instances, estimates have been interpolated over time orspace; the details of how this was done are not always given or available. There are limitations in much of the collection of data. For example,I-131 data for the crucial year of 1956 are not available in rawform, and many of the data on microfilmed hand ledgers are presentlyunreadable. Most of the data collected for the first few decadesof operation were on punched cards or handwritten reports, scanned,and then interpreted by RAC; this was a tedious task for which itdeserves commendation. The limitations of the data notwithstanding,in sum, these data appear to be highly important. There is everyappearance that RAC made a substantial effort to assemble the existingdata, and the use of linked spreadsheets throughout the report isextremely helpful in reviewing the summaries that RAC has eitherreproduced from published data or compiled from the handwritten ledgers. In addition to describing the sources of published data and documentingthe availability of the original raw measurements, the Draft FinalReport presents a number of graphs and analyses of the environmentaldata. In particular, summaries of gross
OCR for page 1
activity and specific nuclides found in sampling by month or yearare given in each relevant report chapter, and these are very helpful. Chapters 8-14 discuss the suitability of the various monitoring programsfor dose reconstruction (or validation). In general, the committeeagrees with the RAC statements about usefulness, although in manycases off-site environmental levels show little direct correlationwith either on-site levels or releases. For air and vegetation samplingof radionuclides, the overall impression provided is that only tritiumlevels are found to be increased off site in a manner consistentwith contamination from SRS. However, original I-131 data for thecrucial year of 1956 (when the largest releases occurred) are evidentlynot available, in that monitoring of milk for I-131 did not beginuntil the end of that year. Often, the other radionuclides are reportedat below detection levels or the pattern of exposure is inconsistentwith that of SRS exposure. For example, many nuclides (such as, I-131and Cs-137) are found at similar levels at the periphery of the siteand 25 miles away, and the variation in time of the levels generallytracks the pattern of atmospheric fallout from atomic-bomb testing.Tritium levels in water vapor and vegetation are found to be higherat the periphery than at 25 miles from the site. However, there islittle temporal correlation between average monthly tritium levelat the periphery or 25 miles away and estimated production levels(figure 8-13); this suggests that it is difficult to use short-termpatterns in tritium measurements at individual measurement locationsto validate tritium dose reconstruction. Chapter 10 of the report discusses the extensive monitoring of off-sitemilk supply and distribution that took place. Indeed, figure 10-2indicates increased levels of I-131 in milk in a large number ofsamples taken shortly after increases in releases were detected inlate May 1961. The milk-monitoring data clearly can be importantin validating a crucial step (milk contamination) in the dose reconstructionof I-131 exposure of people in the region. As noted above, however,probably the most important iodine releases took place in 1956, beforethe milk-monitoring program began. Raw data by location are availableonly through 1973, after which only published annual summaries bysampling location were found. By that time, I-131 (but not Sr-90)was almost always found to be below the detection level of sampling. Chapter 11 describes the sampling of deer killed in large publichunts on site after 1965 as being potentially useful for dose reconstruction.That is stated despite the fact that Cs-137 levels are generallylower than those in deer taken from the South Carolina Coastal Plainhunting camp 65 miles away. Other than confirming that deer meatcould not be an important source of SRS-derived Cs-137, the usefulnessof these data appears minor. Other wild game was not regularly huntedon site, although some sampling did take place. To some degree, theseother animals might act as a gauge of on site contamination; buttheir contribution to personal dose reconstruction seems negligible. Chapter 12 minimizes the role of soil or sediment sampling in dosereconstruction other than possibly to validate the source term foruranium. However, the committee believes that soil sampling couldbe useful in assessing contributions to the overall source
OCR for page 1
term for some radionuclides, such as those of plutonium and uranium,and Sr-90, and Cs-137. Chapter 13 describes sampling for (initially) gross alpha and betaand (after 1960) tritium and Cs-137 in drinking-water supplies in14 communities near SRS. Surface-water sources for these communitiesare at least 40 miles from SRS, and it appears unlikely that SRShas made a significant contribution of tritium to the drinking water,although increased levels have occasionally been detected. In contrast,drinking water taken from the Savannah River downstream of SRS (atSavannah and Beaufort) has consistently shown higher levels of tritiumthan a control location (North Augusta) upstream. Because samplingof tritium began only in the middle 1960s and was sporadic in Savannah,the dose from drinking water is incompletely estimated by monitoringalone. Measurements of tritium would therefore serve primarily asvalidation data for dose reconstruction. Contamination of drinkingwater by other radionuclides at these locations requires extrapolationof Savannah River water measurements made adjacent to SRS. A source of measured contamination was found in the sediments fromsome water-treatment plants. The radionuclides were generally thoseof cobalt and cesium and at levels that were of no radiologic concern.The flocculation procedures used at some water-treatment plants depositedsediment that accumulated, and the longer-lived fission productswere adsorbed to this sediment. Over the years, the levels of tritiumin the Savannah River were detectable but appeared to be below thoseof reasonable concern at the time of measurement. With new chromatographicequipment, tritium was measured in the urine of persons living inthe Beaufort-Jasper area, and values were averaging about twice background;so doses to people were about 0.4 mrem/year, of no importance tohuman health compared with natural background. Radionuclides of plutoniumin urine samples of persons living in the same area around Augustayielded results in the femptocurie-per-liter range, of little importancefor the surrounding population. Chapter 14 describes the fish sampling in the Savannah River as beingpotentially useful for dose reconstruction for people who eat riverfish. Analyses for radionuclides in fish caught along the SavannahRiver were begun when the plant was initially operated. People inthe area eat a lot of fish, so this was a sensible measurement tomake in assessing the environment. Data from 1958-1969 from the SavannahRiver are compiled in table 14-8. In addition, data on fish in on-sitestreams were compiled. Generally, levels of radionuclides were higherin fish on site than in adjacent Savannah River locations, and adjacentSavannah River fish showed higher levels than fish taken upstreamof SRS. Although data before 1958 are unavailable the committee agreesthat the data from the fish-monitoring program will help in validationof possible dose reconstruction. Levels of Cs-137 in fish are describedas most relevant for humans because cesium accumulates in muscletissue. Levels of Cs-137 in both control and adjacent locations werehigher before 1971 than after. Chapter 14 states that the averageconcentration of radionuclides in fish had been “indistinguishableand near or below the LLD [lower limit of detection] since 1972.” That indicates to the committee that SRS contributed only a smallfraction of the Cs-137 found in fish even adjacent to the site.
OCR for page 1
In general, RAC's analysis for radionuclides in fish shows tritium and possibly Cs-137,but not other radionuclides, to be the major source of radiationoffsite. Most increases in environmental nontritium radioactivityhave been correlated with fallout patterns from global sources, butthe tritium is from SRS. Although increases in I-131 effluents weredetected over the years, sharp increases were generally correlatedwith atmospheric fallout from weapons testing. In summary, only for tritium (in air, vegetation and drinking water)and Cs-137 (in fish) do available environmental-monitoring data indicatethe likelihood of SRS contributions to human exposure to radionuclides.However, not all potentially relevant radionuclides were consistentlysampled (consider, for example, I-131 in 1956 or radionuclides ofplutonium at any time). While the environmental monitoring data arenot complete, and much of the data are present only in summary form,they do provide benchmarks for dose reconstruction. In some casesthese data virtually rule out the existence of exposure contributionsfrom the SRS, e.g., Cs-137 from either agricultural products or game.In other cases (tritium in drinking water taken from the SavannahRiver, contribution of radionuclides Cs-137 from fish, I-131 frommilk especially in 1956) detectable levels attributable to the SRShave been found. Determining their importance to human exposure andultimately disease risk could perhaps justify further analysis. To reiterate, no levels that would pose a practical problem for thehealth of the surrounding populations have been detected. However,as minimum detection limits decreased and EPA standards and NCRPexposure guidelines did likewise over the decades, the levels foundin the environment were deemed significant for almost any radionuclideat the 0.1% level. Of interest in this regard are the increases thatoccurred in relation to fallout from nuclear-weapons testing in someyears; most important, the gross alpha-particle levels in some reservoirswere attributed to naturally occurring isotopes of radium and thorium,not to effluent from SRS. Appendix F describes the collection of Geographic Information System(GIS) and demographic data. It is stated that GIS-data compilationwas curtailed for budgetary reasons, but there is no explanationof when or where this decision was made or by whom. Nor is thereany discussion of the possible consequences of the decision or anyindication of whether the collection of those data might be completedlater. To judge whether the data are comprehensive and substantive for usesin future phases of dose reconstruction, it is necessary to knowthe method for using the data. Although no dose-reconstruction modelis given in quantitative form, a diagrammatic version of a modelis contained in the diagrams on pages F-66 through F-75. The followingare proposed as potential uses for the monitoring data (appendixK): source-term verification, model validation, parameter development,and direct exposure assessment. There is little or no discussionas to how the data might be used for those purposes, so it is difficultto determine the sufficiency and quality of the data for them.
OCR for page 1
Do the Phase II results provide adequate information to establisha risk-based screening level for radionuclides? The investigators have done a commendable job in identifying thesites from which individual radionuclides have been released anddocumenting the release rates for each radionuclide; for example,tritium release rates in curies per year have been estimated andplotted. The estimates were derived from trial- and-error analysesbecause the procedures were new and the intricacies of the productionprocess, from which releases might have occurred, were not alwaysobvious. That was particularly true for tritium. Large differencesin stack-effluent measurements compared with the estimated sourceterm proved to be due to deposition in the stacks themselves, whichvaried as a function of humidity or moisture in the stacks. Nonetheless,there appears to be a good estimate of when accidental release ofradionuclides occurred and how the added effluent influenced theaverage tritium output (release). A scheme similar to that used for tritium was developed for I-131.Iodine was released in several forms but most abundantly as iodinegas from fuel reprocessing. Deposition on surfaces is relativelyhigh when iodine is in the vapor form. Other forms, such as HI andHC3I, were also common but not as prevalent as the vapor iodine. Iodinereleases with time are available and could be integrated to obtainan overall average dose to the population. If additional informationon meteorologic conditions or vegetation measurements is available,it would be better to adjust the estimated accumulation accordingly. Filter-paper collectors were used for estimation of the releasedalpha-emitters plutonium and uranium from the separation processesusing continuous stack sampling. About 70% of the effluent heavy-metalreleases (Pu-238, Pu-239, and U) was in the “respirable” range, that is, under3 mm in diameter. Releases to surface water might have much greater political importancethan releases to air because of the concern with the Savannah River.The soluble or possibly soluble radionuclides in groundwater areCs-137, Sr-90, I-131, Co-60, P-32, and Zn-65. The release of H-3to groundwater is reported to be about equal to all other solublenuclides combined that are released into groundwater of all othersoluble nuclides combined. On the basis of the methods of data retrieval used by RAC and theordering of the released nuclides according to their potential healthhazard, the results presented by RAC and CDC seem reasonable as astarting point for establishing risk-based screening. However, thecommittee wonders whether the environmental program was sufficientlyextensive historically—especially for atmospheric releases—to enable CDC and RAC to make extrapolationsto relevant populations with a solid measurement base, rather thanhaving to rely solely on meteorologic modeling. Although this questionmight lie beyond the scope of phase II, has there been an adequatecompilation of demographic, lifestyle (for example, fishing), anddietary data to perform a reasonable risk-based comparativescreeningfor radionuclides?
OCR for page 1
The Community Executive Summary prepared by CDC and RAC for publicinformational purposes CDC and RAC are to be commended for their credible work in translatinga highly technical and densely numerical document into a communityexecutive summary. The document, on the whole, is clear and shouldbe intelligible to lay readers at the college level, although itwill be difficult for readers with less education. However, someof the choices made by CDC and RAC are subject to second-guessing.For example, the figure related to I-131 release estimates in curiesper year uses a log-scale plot, instead of a linear plot which wouldhave made the distinction between elemental and organic iodine clearer. Obviously missing from the community executive summary is any discussionof how the tritium or iodine releases would pose a risk to humanor animal populations. Similarly, although cadmium, nickel, lead,and so on are identified as chemicals of concern, the document doesnot discuss how or why they represent a health concern for populationsin the Savannah River plain. It seems unlikely to the committee thatmuch attention will be paid to the summary unless more informationis included on the potential health risks; in this connection, itis extremely important that these risks be placed in a perspectiveintelligible to lay readers. It is our impression that much publicinterest focuses on the radioactive plume heading toward or enteringthe Savannah River, and we recommend therefore that the summary includemore information and explanation about the radioactive releases,particularly to water. Again, it is imperative that this informationbe put into perspective, preferably on a relative-risk scale. We have some questions: What criteria were used to establish “how harmful these materials might have been to humans”? In the section on releases of chemicals and heavy metals to airand water, the text refers to “available environmental monitoring”. Does that mean that environmental-monitoring information does notexist because it was not collected, or is it unavailable becauseit remains classified? Why does the community executive summary mentiononly the short-half-life radioisotopes (Ar-41 and I-131) and notthose with the more prodigious half-lives (Pu-239 and I-129)? We have some suggestions: Page 1 of the community executive summary has four steps in figure1 and five steps in the bulleted items. The items do not agree andcan confuse readers, particularly if they think that the items areabout the same things. That should be fixed. On page 1, the third paragraph toward the end refers to other radioactivematerials and chemicals but does not specifically identify how many.The expression “large number” is unlikely to be helpful to readers.
OCR for page 1
Why are the elements in table 1 possibly important to a person wholived near the site? How was it determined that they were important?How near the site would people have to live to be included? What criteria were used to select the chemicals to include, and whatrisks do these chemicals have for the population? Are the risks pertinentonly to those who live near the site? There is an inconsistency betweentables 1 and 2 concerning nearby site residents. The difference between elemental and organic iodine should be explainedon page 4 in addition to or instead of at the end of the communityexecutive summary. Effects of radioactive iodine on human populationsshould also be explained. On page 4, why are tritium, Cs-137, and Sr-90 of concern for waterreleases? Are they of concern only for those who live close to SRS?What about people living downstream? Page 6, last sentence before the “Conclusions”: Are the main points of the results of the environmental-monitoringdata review presented in the conclusions section that follows? Ifso, it should be stated. If not, summarize the main points here. The committee recommends that CDC and RAC reexamine the communityexecutive summary in the light of the recent study on stakeholders' risk perception at SRS, published in the December 1999 issue ofRisk Analysis (“Risk Perception in Context: The Savannah River SiteStakeholder Study”, by Bryan L. Williams, SylviaBrown, Michael Greenberg, and Mokbul A. Kahn, volume19, number 6, pages 1019-35). That study makes three points thatin the committee's view are not fully addressed in the summary. First, the study shows that being close to anddown river of SRS was associated with perceptions of risk by thoseinterviewed. Second, SRS residents appear to be cognizant of thepotential risk posed by living down river SRS. In fact, SRS residentsrated water contamination as the most “risky” source of human exposureto environmental contaminants. Third, the vast majority of the residentsin the vicinity of SRS do not believe that their community has beenadequately informed about the decisions at SRS. Despite the public-outreachinitiatives undertaken by site representatives, residents in thevicinity of SRS feel that they are inadequately informed. Given the number of uncertainties that are inherent in the informationgathered by RAC for this the Draft Final Report, particularly thoserelated to gathering data that are more than 40 years old, it isimperative that information about the uncertainties and potentialuncertainties in human doses be presented and discussed in much moredetail in the community executive summary than is now the case. Ease of use and utility of the electronic version of the Draft FinalReport The electronic version of the Draft Final Report is helpful and easyto use, especially in searching for data or references. The committeestrongly encourages CDC
OCR for page 1
to recommend this mode of presentation for future reports of thesame type. That recommendation rests on many considerations, includingthe following: It is easy for interested readers to identify a reference or definitionquickly by simply clicking the hypertext words— a handy way to checkitems without much disruption of the reading process. It is also possible to call up spreadsheet data analyses used inthe preparation and presentation of data in the text, tables, andfigures. This provides access to valuable data resources that forpractical editorial reasons could not be included in the report.Such displays of additional data make it possible for readers tostudy the underlying data in greater detail and draw their own conclusionsfor comparison with those in the main report. In addition to theobvious scientific value of such information, the availability ofmuch of the underlying data should provide further confidence tothe public that a thorough job of mining and analyzing the data wasdone by RAC and associates. Finally, we append to this report some editorial comments that shouldbe helpful in the preparation of the final report. I trust that you will not hesitate to contact me, Dr. Isaf Al-Nabulsi,or Dr. Evan Douple if you wish further elaboration on any of thecommittee's recommendations or suggestions. Sincerely, William J.Scull, Ph.D. Chairman Committee on the Assessment of CDC's Radiation Studies cc: Dr. Evan Douple Dr. Isaf Al-Nabulsi Members of the Committee
OCR for page 1
Editorial Comments: p. 1-11, 1. 33 “…can be used to evaluate…” p. 4.1-3, 1. 4-5 The federal fiscal year ran from 7/1-6/30 through 6/30/76. Aftera transitional period from 7/1-9/30/76, all future fiscal years haverun from 10/1-9/30. p. 4.1-19, 1. 24 “…for each…” p. 4.1-26, 1. 37 Should this be 481,954 Ci? p. 4.3-1, 1. 7 “…released from the reactor…” p. 4.3-5, 1. 23 Change half-live to half-life p. 4.3-18, 1. 14 Change RS to SRS p. 5-5, 1. 12 Isn't some unit missing in the statement 3 mo−1? p. 5-27, 1. 10 Delete “…as part of the…” sp. 5-35, 1. 15 “…and increased…” p. 5-37, 1. 14 Is this pCi/L or pCi L−1? p. 5-43, 1. 21 “…radioactivity being released to…? p. 5-46, 1. 6 “The bases for our…” p. 5-48, 1. 3 “…of 137Cs began, through…” Fig. 5-25 The two zeroes on the ordinate should actually be 0.1 and 0.01 p. 5-50, 1. 7 “…being released to…” p. 5-51, 1. 12-14 An opening quotation mark is missing from this quotation. p. 5-51, 1. 18 Change Cu to Ci. Figures 10-2, 10-3, and 10-6 have no labels on the y-axis to identifythe units. Finally, with regard to the CD-ROM, we note that sometimes the Excelspreadsheets did not work correctly. For example, the “CH 10-Alldata.xls” spreadsheet was supposed to contain data on I-131 (andpossibly strontium) in milk with a page for each year. However, thespreadsheet had no tabs to access other pages so only 1961 was available.