Letter Report to Review Identification and Prioritization of Radionuclide Releases from the Idaho National Engineering and Environmental Laboratory (2001)

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, Georgia 30341-3742

July 2, 2001

Dear Dr. Smith:

As you are aware, 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 critically review and comment on a draft report titled Identification and Prioritization of Radionuclide Releases from the Idaho National Engineering and Environmental Laboratory. The draft was submitted to CDC by the Risk Assessment Corporation of South Carolina (RAC) in partial fulfillment of the conditions set forth in a contract 200-95-0927, task order 5 concluded between RAC and the federal government, as represented by CDC. Among its obligations, RAC agreed to deliver “a prioritized list of radionuclides, with the key facilities identified that were the source of the priority radionuclides, the likely chemical forms of the priority radionuclides, and the important release paths for them, indicating the relative completeness of records relevant to a dose reconstruction for each one” (see page 8, contract 200-95-0927, task order 5).

The committee was asked to address the following four questions:

1. Were the methods and sources of information used in the draft report appropriate?

2. Are the methods and results clearly presented?

3. Where stack monitoring data were used, are the methods for correcting the data understandable and acceptable?

4. Do the results provide sufficient information for selection of the most significant release points and radionuclides for dose reconstruction purposes?

2101 Constitution Avenue, NW, Washington, DC 20418 USA

202-334-2232 (telephone) 202-334-1639 (fax) national-academies.org

At the initial meeting of the committee at the Emory Conference Center in Atlanta on January 30-31, 2001, representatives of RAC—John Till, Kathleen R. Meyer, and Justin Mohler—were present. They enlarged on their draft report and the methods they used in arriving at the conclusions set forth in it, and they responded to questions raised by the committee. Also present at the meeting were six people from the Radiation Studies Branch of CDC: James Smith, chief of the branch; Charles Miller, chief of the Environmental Dosimetry Section of the branch; Judy Qualters, acting chief of the risk analysis and communication section; Marie Spano, health physicist; C. M. Wood, health physicist; and Arthur Robinson, public-health educator.

In the pages to follow, the committee sets out its general reactions to the draft report and then turns to the specific questions it was asked to address.

General comments:

To establish the criteria that the committee should use in judging the draft report, we compared the CDC task order with the work set forth in draft. We note first that the task order does not seem wholly consistent with either the scope of the work or the budget. For example, in sentence 2 (page 2) of the task order, the description does not state “historical releases” only, nor does it state “offsite releases” only. It states “releases to the environment,” which we take to include radioactive materials buried in the ground. In contrast, the penultimate sentence of paragraph 2 (page 3) of the task order clearly limits the scope of the task, as does the last sentence of the first full paragraph on page 5. RAC’s proposed approach on page 5, item 2, does include offsite releases and also promises evaluation of environmental fate, atmospheric dispersion, water dispersion, and other appropriate dilution factors, but these evaluations do not appear to be provided in the draft report. The committee assumes that this full evaluation was not possible with the time and budget allowed. We note also that the last paragraph on page 6 promises to identify data-quality problems, such as sampling-line losses and sampler efficiencies, and sources of bias; such identifications do not appear to be explicitly provided.

The draft report does state that only historical releases were analyzed, although, as stated at the committee meeting, the evaluation is limited principally to historical exposures due to past releases.¹ This assertion should be clarified and, in our view, highlighted in a box or sidebar (explaining that exposures due to future offsite transport of material from historical releases were not estimated or predicted). It should also be clearly stated that only releases associated with historical offsite transport were analyzed. The committee suggests that the word potential be used sparingly to avoid confusion with the uncertain magnitude of future exposures, which are beyond the scope of the draft report by design.

The draft report includes a footnote explaining that the ranking accorded a radioisotope is not based on dose even if it is called such. This, too, should be set forth in a box or sidebar where it could be noted that actual exposures of individuals or their radiation doses were not calculated and that the ranking serves only to characterize or predict possible hazards and is not immediately relevant to possible future health risks.

RAC used the National Council on Radiation Protection and Measurements (NCRP) Report 123 titled Screening Models for Releases of Radionuclides to Atmosphere, Surface Water, and Ground issued January 22, 1996, for guidance in its work contained in the draft report. In the introduction to the NCRP report three suggestions, recommendations, or admonitions are given to readers: First, the methods were “designed primarily for facilities that handle small quantities of radioactive materials released as point-source emissions. However, these screening techniques can be employed by any facilities releasing radionuclides to the environment, and modifications can be made to the current screening methodology to account for releases from area sources.” Second, “conservative models and parameters are proposed. If compliance with the limits can be demonstrated using these screening models, then more sophisticated modeling techniques are not necessary.” Third, “the models should not be used to evaluate the results of accidental releases.” We recommend that the RAC draft report explicitly address those items with regard to their specific application. Each item should be discussed with emphasis on its effect on the estimation of the results of accidental releases (a type of episodic release) of radionuclides from the Idaho National Engineering and Environmental Laboratory (INEEL). The discussion should include the question of what to use if the NCRP method is not used and the precedent that has already been set on the basis of other release situations (such as the Hanford releases, and the DOE (1991a) report on the INEEL).

Many of the discussions of individual facilities and operations on the INEEL are confusing; the reader is assumed to know details that are not discussed or that are referred to with jargon. For example, on page 15, in a discussion of liquid radioactive wastes, the reader is assumed to know that “the Fuel Cycle Facility” and “the laboratory building” were components of the Argonne National Laboratory-West (ANL-W) and that the latter contained “shielded cave areas”. The expert will know what shielded cave areas means in this context, but the public at large might take it to refer to otherwise undisclosed underground facilities. Furthermore, are the “35,000-gallon underground seepage pit,” “a seepage pit,” and “this underground pit” one and the same? It is not possible to tell from the material in the draft report. Another example of confusing descriptions occurs in the discussion of the Aircraft Nuclear Propulsion (ANP) program (pages 69-87), where the discussion refers to Heat Transfer Reactor Experiment (HTRE) assemblies 1, 2, and 3, cores (A2, A-2, A3, and A-3), assemblies (D101A2 and D101A3), a reactor (A3), fuel cartridges, inserts (2B, 1-D, L2A-1, L2A1, L2C-1, and L2E-1), tubes, plates, and other terms. Again, some of these would be understood by experts, but for others (such as “inserts”) and for the precise relationships between and among them (tubes, cartridges, plates, inserts, and cores, for example), one has to know more detail about the HTRE assemblies than is given in the draft report (or even in the Department of Energy, DOE, report, 1991a). Much could be accomplished by simple schematics--for these two

examples, a map of major areas within each facility (presumably showing just one laboratory and one pit) and a schematic labeled diagram of the three HTRE--and by careful selection of the terms that are introduced. Additional comments are listed in Appendix A and B.

Specific issues that the committee was asked to address

1. Were the methods and sources of information used in the draft report appropriate?

Methods:

Explicit definitions of screening factor and screening units should be included. One should reserve use of screening as a technical term in this document. The more general use of the word to imply sorting or otherwise assessing the relative importance of variables should be avoided.

The draft report indicates that the documentation of the methods used is in the spreadsheets (on page 36), but the methods should all be documented in the draft report also and should be consistent between spreadsheets and between spreadsheets and the draft report. While the committee identified some inconsistencies in the methods applied none of these would have any significant effect on the results. Examples of the inconsistencies follow.

• For routine releases, in Offsite_Air_Screening.xls ¹³N releases are said to be reported only in 1972, and releases in all other years have been estimated as double that value. This approach was documented in the draft report on page 38. However, in Onsite_Air_Screening.xls that method was not applied: only the reported value for 1972 is included. The half-life of ¹³N is only 9.97 min, so this approach is more important for evaluating releases to the onsite environment.

• The manner in which ^239,240Pu releases were treated in Offsite_Air_Screening.xls, Onsite Air Screening.xls, and FacilityAirReleases.xls is not the same. In the first spreadsheet, there is a single row for ^239,240Pu + ²³⁹Pu. The second has three rows, for ²³⁹Pu, ^239,240Pu, and ^239,240Pu + ²³⁹Pu; the first two contain reconstructed releases from the Test Reactor Area (TRA) and the Idaho Chemical Processing Plant (ICPP), respectively. The third spreadsheet also has the three rows in the ICPP sheet, for ²³⁹Pu, ²⁴⁰Pu, and ^239,240Pu; but for early periods the ²³⁹Pu row has the ^239,240Pu value entered in it. Only by reference to DOE (1991a) can one reconstruct what is intended in the third spreadsheet, despite the notes in it, and it is not clear what happens in later periods without examination of the source documents which the committee has not done.

• The estimate of ¹⁴C emissions is inconsistent. The draft report (page 37) indicates that the lowest ⁴¹Ar/¹⁴C ratio reported for 1978-1992 was used and gives that minimum as 450. Table 4 and its footnote clearly indicate that the minimum was 420 (attained in

1978). Offsite_Air_Screening.xls implements this, using the value 420. However, Onsite_Air_Screening.xls uses instead an average value of the ⁴¹Ar/¹⁴C ratio for 19781992 of 6,680 (to confuse matters further, the explanatory note in the spreadsheet states that the value is 6,800, whereas the actual arithmetic average over that period is 6,646).

Some descriptions (in Onsite_Air_Screening.xls, Offsite_Air_Screening.xls, and the draft document) of the methods used for some isotopes appear odd, unjustified, or contradictory or to contain arbitrarily selected, variable, conservative factors. Examples follow (see also Appendix B).

• ⁵⁸Co. A notation attached to the row heading of the spreadsheets describes this as follows: “Activation product of stable ⁵⁵Mn; ⁵⁸Co releases reported only for 1973. For screening other years, we calculate annual releases of ⁵⁸Co based on the ratio of the predicted inventory of activation products, ⁵⁴Mn to ⁵⁸Co, of 0.5 (Till & Meyer 1983).” That reference is not in the reference list. The method appears plausible, but the prediction presumably depends on reactor or fuel cladding materials. What activation reaction with ⁵⁵Mn produces ⁵⁸Co? ⁵⁸Co is apparently produced through the reaction ⁵⁸Ni(n,p)⁵⁸Co, whereas ⁵⁴Mn is likely to be produced by ⁵⁴Fe(n,p)⁵⁴Mn. So the method is apt to depend on a stable relation between Ni and Fe in the claddings of fuels that were reprocessed. However, this discussion might be moot. For each individual year for ⁵⁸Co, there is another note in the spreadsheets specifying another method, which is the method that appears to have been used: “Activation product; releases only reported in 1973. For release estimates for other years we assumed ten times the ratio of ⁶⁰Co to ⁵⁸Co (ratio in 1973 was 1.73) and applied that to reported releases of ⁶⁰Co. For example, 0.21 Ci ⁶⁰Co reported in 1952; therefore, releases of ⁵⁸Co assumed to be 0.21 x 17.3 = 3.7 Ci ⁵⁸Co for 1952.” No rationale is provided for this method, and it appears that the result is the reciprocal of that intended (⁵⁸Co/⁶⁰Co = 1/1.73 = 0.58), so that an additional factor of 3 in addition to the intended factor of 10 has been inadvertently introduced.

• ⁵⁴Mn. This is described as follows in the spreadsheet: “Activation product of stable ⁵¹Cr; reported releases for 1965, 1973-1975; for other years we assume annual release based on the ratio of releases of ⁵¹Cr to ⁵⁴Mn of 3 (ratio ranged from 3 to 20).” See the proceeding comment. What activation reaction with ⁵¹Cr will produce ⁵⁴Mn? Moreover, the ratios for 1965 and 1973-1975 were 9.81, 5.88, 20.0, and 32.3—a range from 5.88-32.3 (not 3-20 – it appears that 32.3 was misidentified as 3.23).

• Furthermore, examining the values for ⁵¹Cr, we find a note, “Releases reported or [sic] 1952-1969 and 1986-1992; for other years, we assume an annual release equal to twice the average of other years. Average = 0.055 x 2 = 0.10.” So the ⁵⁴Mn estimates are themselves (except for 1965) based on ratios to other estimated values that have a conservative factor built in. And 0.055 x 2 = 0.11, not 0.10.

a 1995 press release from the INEEL (see Appendix D) is not included in the releases examined. Although the release was small, it exceeded the reported routine release in at least 1 year (for example, 1992). If the draft report intended to omit releases that were smaller than some threshold, that should be stated, and the threshold should be given and justified.

Pages 17-21 give some details of cross checks on releases from the radioactive lanthanum (RaLa) program, state that reported daily releases of ¹³¹I were compiled, and refer to FacilityAirReleases.xls for details. That file, however, does not contain any daily or weekly compilation of releases; it contains a monthly compilation, but gives references (HJM1999070733, HJM1999070739, HJM1999070757, HJM1999070761) that are not identifiable in the draft report.

Pages 22-26 discuss routine releases and give examples of the documents consulted (page 22). The section on “Confirmation of Reported Radionuclide Releases” (pages 23-24) discusses compilations of releases and cross checking between these compilations and “annual environmental and effluent release reports, from the Radioactive Waste Management Information System (RWMIS) data, and from HDE (DOE, 1991a) documents.” Several minor discrepancies are noted in ICPP data. However, there is no list of the documents consulted, and the committee could not locate, in the material provided to it, any spreadsheets or other documentation giving details of the compilations and cross checks.

The committee recommends that the documents used for the compilations of releases be listed in the report and that the cross checks be documented.

The source terms used in the draft report for routine releases are again discussed on pages 34-39. However, the only documentation of source documents in the draft report is on page 35, which indicates that “for annual releases to air for early years of operation, HDE (DOE, 1991a) data proved to be the most complete, because adjustments had been made to the RWMIS data that were the basis for the Historic Dose Evaluation source term data;” it is implied that the HDE values are used, but the period is not specified in the draft report. There does not appear to be any documentation in the draft report on the source documents for total routine releases for any period other than “early years.”

For some years, releases are given with no references in the draft report or the supporting spreadsheets. And for some years, no emissions are listed, and there is no reference to indicate documentation for zero emissions (ICPP in 1971-1973, and 1977; ANL-W in 1971-1973, 1975-1977, 1984-1987, and 1989). Low emissions for ANL-W after shutdown of the Fuel Cycle Facility in 1969 (page 15) would be plausible. However, no history of ANL-W after 1969 is provided in the draft document; in fact, the Fuel Cycle Facility was not shut down in 1969, so the source of emissions and the plausibility of years with no emissions are not addressed in the draft report.

Examination of the source citations indicates that in many instances they are not primary but secondary compilations (e.g. DOE, 1991a; Litteer et al. 1991; Litteer and Reagan 1989; Hoff et al. 1984, 1985, 1986, 1987, 1989, 1990, 1991, 1992, 1993). While DOE (1991a) and some of the corrections made therein are explicitly discussed in the draft report, it would be useful for the report to indicate the nature of each of the references. For example, what was the RWMIS, how was it compiled and maintained, and when did it cease operation? Were the data in it consulted independently of DOE (1991a)? Some mention that DOE (1991a) is based largely on the RWMIS is also warranted. There is no indication that a direct comparison was made between the data in DOE (1991a) and the RWMIS; was such a comparison made to check for errors in compilation of DOE (1991a)?

For total sitewide air emissions, as listed in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls, the only documentation the committee could find for the source of the data was in these spreadsheets. They indicate that for 1952-1989 the data came from DOE (1991a, Volume 2) and for 1990-1992 “From PDM1997080611; INEEL Site Environmental Report for CY 1990, 1991, and 1992” which the committee takes to indicate the references Hoff et al. (1991, 1992, 1993) in the draft report. The total emissions are not the sum of the facility emissions tabulated in FacilityAirReleases.xls, although the draft report does not make clear the relationships between the various spreadsheets—or indeed between the various sections of the draft report dealing with routine releases. The relationship between the entries in the spreadsheets FacilityAirReleases.xls and Onsite_Air_Screening.xls, and Offsite_Air_Releases.xls as well as the relationships between the sections of the draft report dealing with routine releases need to be clarified.

For routine releases, DOE (1991a) is itself a secondary source, in that it uses and modifies data from another source (the RWMIS). The total site emissions used in the draft report appear to have been obtained principally from Tables A-35 through A-40 of DOE (1991a). The exact sources of the data used in the draft report should be documented. Furthermore, the report needs to indicate any cross checks performed on these tables. For example, were Tables A-35 through A-40 taken directly from DOE (1991a), as indicated by the citation, or were they reconstructed from the RWMIS by using the modifications described in DOE (1991a)?

Despite their source reference, the tabulated values in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls are sometimes drastically different from those given in DOE (1991a) Tables A-35 through A-40, apparently because of transcription errors; the most common differences are factors of 10 or 100 or in location in the tables. The largest unexplained difference noted by the committee is a 100-fold difference for ^85mKr in 1956; this difference alters the estimated total emissions for that year by a factor of 2 and has more than a 10% effect on the onsite screening estimates for that year. Without consulting the original sources, the committee cannot be certain whether the error is in the draft report or in DOE (1991a), but examination of data on adjacent years suggests that the error is in the draft report. The 10-fold differences noted for ¹³⁷Cs in 1958 and

1963 could have a 20% effect on the offsite screening values for those years. Other differences noticed by the committee are listed in Appendix C.

The committee recommends that the report list the source documents used for the routine releases, both for individual facilities and for sitewide emissions. In addition, all transcriptions from original documents should be thoroughly proofread and checked for plausibility by comparing between years and between nuclides within years.

2. Are the methods and results clearly presented?

For the routine releases, many of the methods are documented only in the spreadsheets and not clearly presented elsewhere. It is impossible to discern the relationship, if any, between the various parts of the document dealing with individual facilities, the RaLa runs, and the overall emissions from the site as a whole.

The final report should provide:

• Better maps of the INEEL to indicate the location of all named places and identify which wells were used for observations and which were used for monitoring.

• Schematics of facilities or experiments to illustrate the location or function of all named facilities, buildings, and components.

• A complete chronologic list of all the releases described in the text and associated spreadsheets.

• A chronologic description of each of the facilities and programs that allows comparison with the emission lists. Separation of the listed emissions into emissions from the various individual facilities would also assist in understanding the relationships.

• More detailed description of the individual episodic releases. The descriptions given are generally less complete than those of DOE (1991a) although they draw heavily on the descriptions given there. In many places, the draft report is difficult to understand without simultaneous reference to DOE (1991a); where there are statements that the methods of DOE (1991a) are being followed, the methods should be described.

• Almost all quantitative material is relegated to the spreadsheets. The text for individual episodes should refer explicitly to the spreadsheet that contains relevant material, and the spreadsheets should clearly refer to the text. Otherwise, the reader has trouble in linking the documents effectively.

3. Where stack monitoring was used, are the methods for correcting the data understandable and acceptable?

The draft report does not include any calculations on stack-monitoring data, or on any methods used for their evaluation. Such data and evaluation clearly must have been used for the routine releases, and for several of the releases treated as episodic (such as the Initial Engine Test (IET) series). One possible reference within the draft report to correction methods was where it agreed with DOE (1991a) that estimates of ¹³²I emissions from the October 16, 1959 ICPP criticality had been miscalculated (draft report, page 97; DOE, 1991a, page A-174). However, there is no further discussion within the draft report, and the values used in the spreadsheet appear to correspond to the erroneously calculated values (see below under "spreadsheets"). The DOE report (1991a) mentions that small corrections were applied to the stack monitoring data from the IET series, but again the draft report does not discuss these corrections or indicate how those monitoring data were used in their evaluation of these events. The committee is thus not in a position to evaluate any "methods for correcting the data" from stack monitoring.

4. Do the results provide sufficient information for selection of the most significant release points and radionuclides?

The draft report does not yet provide sufficient information for the selection of the most significant release points. However, it seems unlikely that any radionuclide of dosimetric importance in historical discharges was missed but some radionuclides of less import that probably should not have been included were added. The use of different levels of conservatism for episodic and continued releases leads to the elevation in the rankings of short-lived radionuclides from episodic discharges. For the episodic releases, lack of sufficient documentation of discrepancies (such as in the evaluation of IET 10) leaves the committee unable to conclude that source terms have always been accurately evaluated. Moreover, the variable conservatism between episodic releases for both the source terms and the dispersion estimates introduces sufficient uncertainty that selection according to the ordering in the draft report cannot currently be justified. Accordingly, the committee believes that episodic releases and routine releases should be separated and the rankings reported separately because the ranking methods used are not equally applicable to the two cases.

There is a great deal of public concern about issues at INEEL and about the RAC draft report. As it presently exists, the latter report responds only minimally to these citizens’ concerns, particularly those relating to dose considerations. The committee is aware that such considerations were not the main objective of the RAC report, but, given the description by RAC staff of the time and effort spent with the INEEL Health Effects Subcommittee (HES), the report should have given more attention to citizens’ concerns.

The committee offers the following recommendations to make the report more understandable and meaningful for the public.

• There should be a public summary, written in easily understood, jargon-free language. It is difficult for a layperson to understand much of what is in the report. There has been considerable public concern about this report, and at least part of it should be understandable to citizens of the region.

• Major findings should be highlighted and explained in the public summary and in the conclusions of the report. As it stands, the report is difficult for even an educated layperson to distinguish the “wheat from the chaff” and determine what is most important. Did RAC recommend a detailed dose-reconstruction study on the basis of its findings? If a dose-reconstruction is contemplated, there will be a need to identify residential population shifts as a function of time and with respect to site locations. It will also be necessary to identify special groups brought in for short-term work on the site. The last sentence on page 136 of the report is especially weak. It needs to be strengthened to say clearly what RAC recommends—should further studies be done or not? In fact, the summary and conclusions section that starts on page 134 is far more summary than conclusions, and there are no recommendations. That is the substantial weakness of the report for the public and perhaps also for scientists and other interested persons.

• Public concerns about plutonium and other important issues should be addressed in the report in both the public summary and the main text. Even if there are no positive or negative findings related to these issues, that should be stated. Ignoring issues of public concern diminishes the value of the report.

• In both the public summary and the text of the report, two important issues should be discussed in more detail. First, the extensive use of the DOE (1991a) report as a basis for describing episodic releases should be explained clearly and justified. Given that there was public opposition to the DOE report, basing this report largely on the findings of the earlier one requires discussion - if for no other reason, to establish public credibility for the new report.

  Second, as recommended elsewhere in this report, there needs to be serious discussion of and reflection about the impact of the massive destruction of documents. Gathering and analyzing INEEL data were parts of task order 6, which was included in the RAC report. Avoiding discussion of the repercussions of the loss of hundreds of boxes of information means skirting an important question. There should be a section both in the public summary and in the main text that explains what happened to the boxes, what types of information the boxes were supposed to contain, how valuable such information would have been to the RAC study, and what could have changed in the RAC study’s results if the documents had been available. Clearly, the public is aware of the situation; it has been discussed by the Environment Defense Institute in its newsletters. Ignoring the issue will increase public suspicion about the credibility of this report.

• Another factor that should be explained in the public summary and in the introduction to the main report is the public input into the report. RAC staff members explained to the committee that many meetings and conversations were held with members of the INEEL HES. In addition, one of the RAC overheads shown at the committee meeting said “task order work is very important for establishing public participation.” Given those statements, it is important to document the degree of public input and participation and how they were used in and affected the report. Meetings between RAC staff and HES should be documented, and some record of what occurred at these meetings should appear in both the public summary and the main report. In addition, if HES committee minutes were related to the interactions, they could be appended to the report. The DOE (1991a) report was criticized for lack of public input, and the new report should indicate how RAC sought and used public input to reach its findings.

• There is a published history of the Idaho Nuclear Engineering and Environmental Laboratory at http://www.inel.gov/documents/ineel_history/ . That history discusses some of the episodic events, using eye-witness accounts, in a way that is difficult to reconcile, particularly for a layman, with the descriptions in the draft report. In particular, the descriptions of Boiling Water Reactor Experiment (BORAX-I) in Chapter 14, and SNAP 10A Transient (SNAPTRAN-3) in Chapter 17, might suggest release fractions considerably higher than used in the draft report. Some consideration should be given to a commentary on those published descriptions.

To summarize briefly, the committee compliments the RAC investigators on their efforts and many particulars in their draft report; however, we believe the report can be improved significantly. To this end, the following changes in the draft report are recommended: First, RAC should clearly state that its contractual charge was to analyze releases associated with historical offsite transport. The intent was to provide a ranking of the radionuclides released based on their potential importance and the results do not reflect actual doses to the public onsite and offsite. Furthermore, their charge did not include the estimation or projection of exposures due to future offsite transport of radioactive material from historical releases. Second, RAC should make a sharp distinction between episodic versus routine releases (they should clarify that they cannot add the episodic to the routine releases). Episodic releases and routine releases should be separated and the rankings reported separately because the ranking methods used are not equally applicable to the two cases. RAC should also include a discussion pertaining to the question of what to use if the NCRP method is not used and the precedent that has already been set on the basis of other release situations. Finally, in its report, RAC should give more attention to public concerns. While the committee recognizes that such considerations were not one of the major objectives of the report, nonetheless the acceptance and credibility of the findings will hinge on public understanding. This behooves RAC to make the report understandable and meaningful for the public.

If you desire further elaboration on the comments above or in the accompanying appendixes, please call or write Dr. Isaf Al-Nabulsi or me.

wjs:s

DOE . Idaho National Engineering Laboratory Historical Dose Evaluation: Volume 1 and Appendices: Volume 2 . DOE/ID-12119 . INEEL Task Order Database MC Number 51061 . August 1991a .

Hoff, D.L. , E.W. Chew , and R.L. Dickson . 1983 Environmental Monitoring Program Report for the Idaho National Engineering Laboratory Site . DOE/ID-12082 (83) . INEEL Task Order Database MC Number 50930 . May 1984 .

Hoff, D.L. , E.W. Chew , and R.L. Dickson . 1984 Environmental Monitoring Program Report for the Idaho National Engineering Site . DOE/ID-12082 (84) . INEEL Task Order Database MC Number 50949 . May 1985 .

Hoff, D.L. , E.W. Chew , and S.K. Rope . 1985 Environmental Monitoring Program Report for the Idaho National Engineering Site . DOE/ID-12082 (85) . INEEL Task Order Database MC Number 50946 . May 1986 .

Hoff, D.L. , E.W. Chew , and S.K. Rope . 1986 Environmental Monitoring Program Report for the Idaho National Engineering Laboratory Site . DOE/ID-12082 (86) . INEEL Task Order Database MC Number 50956 . May 1987 .

Hoff, D.L. , R.G. Mitchell , and R. Moore . The Idaho National Engineering Laboratory Site Environmental Report for Calendar Year 1988 . DOE/ID-12082 (88) . INEEL Task Order Database MC Number 50938 . June 1989 .

Hoff, D.L. , R.G. Mitchell , G.C. Bowman , and R. Moore . The Idaho National Engineering Laboratory Site Environmental Report for Calendar Year 1989 . DOE/ID-12082 (89) . INEEL Task Order Database MC Number 50936 . June 1990 .

Hoff, D.L. , R.G. Mitchell , R. Moore , and R.M. Shaw . The Idaho National Engineering Laboratory Site Environmental Report for Calendar Year 1990 . DOE/ID-12082 (90) . INEEL Task Order Database MC Number 50937 . June 1991 .

Hoff, D.L. , R.G. Mitchell , R. Moore , and L. Bingham . The Idaho National Laboratory Site Environmental Report for Calendar Year 1991 . DOE/ID-12082 (91) . INEEL Task Order Database MC Number 50951 . September 1992 .

Hoff, D.L. , R.G. Mitchell , R. Moore , and L. Bingham . The Idaho National Engineering Laboratory Site Environmental Report for Calendar Year 1992 . DOE/ID-12082 (92) . INEEL Task Order Database MC Number 50914 . June 1993 .

Litteer, D.L. , C.N. Peterson , and A.M. Sims . Radioactive Waste Management Information for 1990 and Record-To-Date . DOE/ID-10054 (90) . Idaho Falls, Idaho : EG&G Idaho, Inc. July 1991 .

Litteer, D.L. and B.D. Reagan . Radioactive Waste Management Information for 1988 and Record-To-Date . DOE/ID-10054 (88) . Idaho Falls, Idaho : EG&G Idaho, Inc. July 1989 .

1. There is no indication of whether the various references, such as the DOE study, were best estimates or contained substantial overestimates, so-called “conservatives”.

2. The introduction of the report should identify the time period encompassing all releases considered.

3. Page 3, last paragraph: What is the center of the “50-mile radius”, or is it 50 miles from the site boundary?

4. Page 3, last paragraph: “operating or operable” is a quote, but no citation is given.

5. Page 4, second paragraph: “monthly reports, particularly from the ICPP,” is then followed by three citations. The first is to Hayden (1957-1963) and the other two are individual monthly reports. Is this a complete list, or just a sample? If it is a complete list, it apparently covers only the period 1957-1963 and 1977. But the implication of the sentence is that there are monthly reports for the full period of operation.

6. Page 5: The map appears to be a badly reproduced copy of one in the HDE. It is impossible to discern the location of the INEEL from the inset (even at 4X magnification). A better map of the overall area should be used.

7. Page 6, Figure 2: What does “reported” mean? The citation to DOE (1991a) implies that these are the releases that are estimated in that document and so might be reconstructed from other records. Or is this a graph of the releases estimated in the draft report? Or are the two identical? Similarly for Figure 3 on page 7.

8. Page 7, Figure 3: These are not all legible regarding which facility is plotted. The reported source for this figure is DOE (1991a). But for ANL-W, DOE (1991a) has explicit entries for only 1960-1966 (in Table A-17). Furthermore, even the draft report has no entries (in the spreadsheet FacilityAirReleases.xls) for 1971-1973, 1975-1977, 1984-1987, and 1989. Where do the points for ANL-W for 1971, 1972, and 1973 come from? Similarly, the graph shows estimates for emissions from TRA in every year from 1952 to 1988. DOE (1991a) explicitly lists such emissions only for 1952-1968 (Tables A-7 through A-11; the spreadsheet notes incorrectly list all entries for 1952-1968 as being from Table A-7). Furthermore, the draft report (in the spreadsheet FacilityAirReleases.xls) has no entries for TRA emissions for 1971-1973, 1975-1977, 1984-1987, and 1989. Where do the points for TRA for dates before 1988 come from?

9. Page 8: What is (are) the source(s) for Figure 4? None is listed.

10. Page 9, first line: “the stack”: Three stacks are visible in Figure 5.

11. Page 9, first paragraph: References for this information on the MTR, ETR, and so on, would be helpful. Also, “the TRA had three 250-foot stacks for iodine collection”: when? For its entire period of operation?

12. Page 9, last sentence of first paragraph: The reference given (De Boisblanc 1958) is not to documentation, but to testimony. It is not clear whether this was written or oral testimony, but some indication needs to be given as to how this amounts to “fairly well documented” if it came from a single person’s testimony at a single time.

13. Page 9, last sentence: “the detection limits of the effluent monitor” ranged over more than 3 orders of magnitude with no explanation provided. Which monitors, and for what activities (the previous sentence discussed gross and alpha particulate, gamma, and iodine activities)? Why are these numbers relevant here?

14. Page 10, Figure 6 caption: Refers to 200-foot stack, but text on page 11 indicates a 250-foot stack.

15. Page 11, first paragraph: What does “steel-encased concrete binds” mean?

16. Page 11, Figure 7: See comments on Figures 2 through 4. Also, the caption here has interchanged 1.1 and 1.3. The discussion of the two peaks is not consistent with the graph.

17. Page 12, first paragraph, last sentence: Says 15 Ci per day for both 200-day cooled waste and 5-year cooled waste. Is that correct?

18. Page 12, second paragraph: What is a “particulate cleanup study”? There should be a reference.

19. Page 12, second paragraph: It is not clear whether this paragraph refers to what was done in this draft report or to what was done in some other evaluation (if another, a reference would be appropriate). Nor is it clear whether the stated assumptions have been accepted in this draft report or modified.

20. Page 12, third paragraph: There is a sudden and confusing jump in this paragraph from air to liquid emissions.

21. Page 14: Argonne West was established in 1950, not in the middle 1950s.

22. Page 15: The fuel cycle facility was not shut down in 1969; in fact, it is still operating and EBR II underground storage is not a burial ground.

23. Page 16, paragraph 6, lines 4-5: The meaning of this sentence is not clear with regard to the use of the word “screened”.

24. Page 18, second full paragraph: Where and what is cell L of the ICPP? What about the “Multicurie cell”? See the discussion of diagrams in the body of our text.

25. Page 18, second full paragraph: The statement that “in some months there are 3 to 4 orders of magnitude variation in ¹³¹I released to atmosphere” appears meaningless. Is this an attempt to quantify variation between instantaneous release rates, daily totals, weekly totals, monthly totals, or something else?

26. Page 19, second paragraph: The half-life of ¹³²I is given here as 2.5 h. The spreadsheets variously list it as 2.30 or 2.28 hours, and another source gives 2.295 h.

27. Page 19, line 24: What does it mean to say that “Neither expectation was seen”?

28. Page 19: The word “jetted” needs to be defined, as do “process off gas” and “vessel off gas”.

29. Page 19, third paragraph: “these reports contain results”: Only one report has been cited.

30. Pages 19 and 20: What are C slugs and J slugs?

31. Page 20: Table 1 appears to be pointless in the current context of the RaLa program if these are not associated with the RaLa program. The table might be appropriate in some other section on non-RaLa emissions.

32. Page 20, bottom: The source documents for these compilations are not identified and there are no references in the text.

33. Page 21: The switch from Ci to mCi in this paragraph appears unnecessary and could be confusing.

34. Page 22, “DOE (1991a) reports that RaLa run #15 occurred on June 2, 1958, while the daily discharge report stated that RaLa run #15 occurred on May 28, 1958 and no RaLa run was noted for June 2”: What, if anything, was done to resolve this discrepancy? DOE (1991a) also reports that run 14 occurred on May 28; if run 15 was on that date, when did run 14 occur?

35. Figure 12: The use of curies on the figure and millicuries in the description is confusing.

36. Page 22: Some RaLa runs were called routine and some were also evaluated as episodic events. Were they counted twice?

37. Page 23: Was the difference in plutonium release because RAC used the limit of detection when nothing was detected and DOE used zero? What are the references for monthly, quarterly, and annually reported releases? The “explanation” for the

61.2 mCi ¹³⁷Cs vs. the 63.2 mCi is not an explanation. The cited difference is 2.00 mCi, but the explanation refers to a correction of 4.50 mCi.

38. Page 24, lines 19-20: Citation/reference for this claim.

39. Page 29: The discussion of buried radioactive solids needs to be corrected. Substantial quantities of plutonium were buried from very early times.

40. Page 32: The section on NCRP methodology needs to be re-written.

41. Page 33: Documentation/details on “conservative and simple transport models”?

42. Page 36, end of second paragraph: The statement is made that special analyses were not done in many cases. That might be true because those analyses were not available at that time; they had not yet been developed. The wording is misleading.

43. Page 36: Basis for claiming that adjustments made to RWMIS data were “determined that they were technically defensible and soundly based”.

44. Page 36: Evidence for claim that “It appears that this instrumentation improved the analytical results”.

45. Page 36: Evidence for claim that “this type of improvement led to increasing confidence in the measurement techniques”. Is this a subjective or an objective conclusion?

46. Page 38, bottom: Is a factor of 3 extra added to the ¹²⁹I estimate?

47. Page 39, second paragraph: The ¹³⁴Cs estimate for clean years used twice the release as for dirtier years. This might be an overestimate by a factor of 3-10.

48. Page 39, fifth paragraph, “for our conservative screening, we assumed annual releases of ²⁴¹Am equal to 0.36 % of reported ^239,240Pu releases (the percentage of plutonium) that would be ²⁴¹Pu”: Examination of the spreadsheets shows that the estimate of ²⁴¹Am is 0.4% of the reported ^239,240Pu from ICPP; the ²³⁹Pu releases reported for TRA have been omitted from the calculation.

49. Page 43, Table 6: The ratio of air-screening value to percentage of total screening value seems confused for ¹³⁷Cs versus ¹³⁴Cs.

50. Page 45: What assumptions were made as to length of exposure of public driving by on the public road?

51. Page 49: Which “site environmental documents”, which “USGS reports”? Specific references/citations should be given.

52. Page 50: “Magnuson and Sanders, 1998” is not in the bibliography although cited on this page.

53. Page 50: “carefully studied and measured by the USGS”, citation/reference?

54. Page 50: “we have determined that the exposure pathway ...”, How?, Documentation, compuations, etc? (Section on Plutonium isotopes).

55. Pages 50-51: “complete offsite exposure pathway” needs to be clarified. The footnote at the bottom of page 51 needs to be clarified and justified.

56. Page 50, third paragraph, lines 1-4: Why was Pu not detectable as early as 1974?

57. Page 52: “the scientists looked at...”, which scientists? Documentation, evidence?

58. Page 52: “we believed ..”, vs “we concluded ...”?

59. Page 54, Figure 20: Not all the wells shown in the DOE report are shown here. Which figure is correct?

60. Page 55, footnote 9: Overexposure factor of 12 is used for tritium because it was found in only one month but was assumed to exist in every month.

61. Page 61: Equations/computations to go from Table 13 to Table 14?

62. Page 64: Typo in “(Ci-s m-³)”?

63. Page 115: For “ICPP Releases” in 1974, no reasons are given; were any found in the documentation? The total ⁸⁵Kr and ¹²⁵Sb releases in 1974 are 10 and 100 times larger, respectively, in 1974 than in adjacent years. The reported quantities in the releases described here amount to less than half the total excess, suggesting other similar releases that year. Is there any reason why these particular reports were located, but not others?

64. Page 122: “and generally believe that the work is thorough and scientifically defensible ..”, evidence/reasons for this claim?

65. Page 136: Is the last paragraph considered to be a significant conclusion in the report? If so, couldn’t this recommendation have been made prior to RAC’s work? What is different?

66. The material in this report is highly technical, and many terms will be unfamiliar to the lay reader. It would be helpful to include a glossary and to cite a single source to which a reader can be referred.

67. The mathematical equations contain several major errors. For example, see page 133, the equation for H.

68. The “References” section of the document would be more useful if it contained more information. Most readers will not have access to the INEEL task force database (how does one gain access to it?). A short description of what many of the documents are would be helpful. For example, “Hayden, R.E. 1957-1963. Activity discharged to Atmosphere. INEEL Task Order Database MC Number 60111” does not convey much information to a reader.

Although RAC acknowledges that its draft report draws heavily upon the earlier DOE report (1991a), the committee finds a number of troubling and unexplained inconsistencies between the two assessments. It is to these that the committee now turns.

The treatment of meteorologic conditions for the episodic releases (pages 6265) is relatively crude, particularly since a better treatment apparently had already been carried out by DOE (1991a) with substantially more contemporary data than evaluated in the draft report. DOE (1991a) states that it used the MESODIF model, and contemporary meteorologic records from the times of individual episodes. However, the main text of volumes 1 and 2 of DOE (1991a) do not provide details of meteorologic modeling for the individual events (the committee presumes that material is included in the microfiche addenda to volume 2). The draft report should evaluate the method used by DOE (1991a) and adopt it if it is supportable.

The simpler method used in the draft report results in variable, large factors of conservatism that could be significant. That can be demonstrated with the following table, taken from the draft report documents combined and Table B-2 of DOE (1991a). The events have been ordered by the draft report’s relative screening factors.

The four columns on the right are, respectively, the highest dispersion factor calculated by DOE with MESODIF, the corresponding dispersion factor in the draft report, the ratio of the two (draft report/DOE), and the offsite screening value obtained in the draft report, relative to the maximal offsite screening factor. Releases that DOE treats in multiple periods have been arbitrarily assigned the same draft report dispersion factor and screening value for all periods.

Examination of the table shows that if the dispersion factors obtained by DOE (1991a) are better estimates, the dispersion factors obtained in the draft report could be awry by factors ranging from 0.47 (FPFRT-8) to about 20,200 (SPERT-I, test No. 2). Such factors are large enough to change the ordering of the releases; for example, IET-3 and IET-10 would have similar offsite relative screening factors if the DOE dispersion estimates were correct. The observation of a ratio of 0.5 calls into question even the conservatism of the procedure adopted in the draft report. Examination of the information presented for this release (FPFRT-8) in the draft report (Table 15) and in DOE (1991a) (Figure B-22 and Figure B-23 and Table B-3) suggests that further investigation is in order. The draft report indicates a wind speed of more than 4 m/s for an hour after the release, with unstable conditions; the DOE report (1991a) indicates an average wind speed of 0.9 m/s for more than 10 hours, and the low dispersion suggests stable conditions. Similarly, the high ratio of 20,000 for SPERT-I, test No. 2 suggests that further investigation is required. Even without the two extreme values, the range of ratios (0.53-1270) is still large enough to call into question the ordering of releases obtained with the simple approach of the draft report.

The committee has examined some of the operational release data and a number of the episodic releases documented in the draft report, but not all of them in detail. It has found a number of errors in the calculations and the documentation (or at least a lack of agreement between the calculations and documentation), indicating the absence of adequate quality control.

As stated earlier, the operational releases of the draft report appear to have been copied directly from DOE (1991a). It is unclear how DOE (1991a) handled decay equilibria in its tables, and this confusion extends to the draft report. DOE (1991a) states that “when only the parent radionuclide appeared in the RWMIS data base, the progeny was also assumed to have been produced and to be present in the equilibrium ratio. For some pairs of radionuclides both the parent and the progeny appear in the RWMIS database. In such cases, the amount of the progeny expected from assuming equilibrium is included with the parent nuclide and indicated by the notation +D in Section A.3. If the reported activity of the progeny nuclide is less than that expected from equilibrium, the assumption that equilibrium exists results in a conservative overestimate of the

release. If the reported activity of the progeny exceeds that expected from equilibrium conditions, then the excess is listed explicitly in the table and included separately in subsequent dose calculations.”

That suggests that the progeny activities are added to the parent in Tables A-35 through A-41 in Section A-3. However, for episodic releases, in which the +D notation is used, the progeny activities have apparently not been added to the parent in Table A-41 (only the parent activity is reported, with the progeny activity implicit). For routine releases, the +D notation is not used in Tables A-35 through A-40, but the progeny are largely or entirely absent from the table, and it is unclear whether the progeny activities were included or are to be understood.

The draft report (page 34) accepts the DOE (1991a) rationale for using an assumption for decay-chain equilibria, but it is not clear whether the progeny activities have been included in the calculations of the draft report. The draft report largely uses the values in DOE (1991a) Table A-35 through A-40, with no additions for the progenies; but it does not appear to assume that the activities reported for the parents include contributions from the progenies. In contrast, in the episodic releases, the progenies are generally explicitly included in the tables of the draft report (but excluded in the similar tables in DOE, 1991a).

On page 58, the draft report treats the concept of release fractions for the first time. For the less technical reader, some explanation of what these are and why they might be selected for solids, noble gases, and halogens is in order. Some explanation of why they are not also separately selected for cesium and ruthenium (and it should be documented that the solids release fraction was used for the last two), is also in order inasmuch as all five release fractions have to be specified for the Radiological Safety Analysis Computer Program, Version 5 (RSAC-5) when this option is used. It might be worthwhile to reemphasize that decays introduce substantial quantities of nonhalogen and non-noble-gas radionuclides, even if these are the only nuclides released.

On page 65, for stability classes other than D, dispersion coefficients were obtained from Clawson et al. (1989), Figures VI-7 and VI-8. However, those figures are appropriate only for releases that extend over less than 15 min. Several of the episodic events involve releases over periods exceeding that time; for such cases, Figures VI-10 and VI-11 are specifically recommended for the INEEL by the cited reference. The values of σy are substantially larger for longer release periods under stable conditions because of the effects of plume meander.

On page 65, none of the entries in “equation 10” is dimensionless, so assuming them “unitless”, whatever that means, is obscure. It is conventional in writing physical equations for symbols to represent physical quantities (that is, for a scalar, a magnitude and a dimensionality, usually represented by a number in a specified system of units).

For numerical calculations, units must be compatible. Obtaining a dimensionless “relative screening value” requires division by some standard absolute value.

In BORAX-1 (pages 66-68), the description at the bottom of page 66 will mean little to the reader without maps that include the named locations (such as Buren-Adams junction) and facilities (such as Union Pacific Railway).

On page 66, the plume is said to have gone in a southwesterly direction; on page 68, the offsite location evaluated is Atomic City, which is east-southeast of BORAX-1. The discussion of “the location with the highest dispersion factor reported by DOE (1991a)” is confusing because the text of DOE (1991a) reports only a single “highest dispersion factor” for BORAX-1, and that is for Frenchman’s Cabin, to the southwest of BORAX-1.

Page 68 (top), needs to state which SPERT-I test is referred to. By implication, it is SPERT-I, test No. 1. A forward reference to page 100 is needed.

On page 68, the 30 elements are said to contain about 4200 g of ²³⁵U at 90% enrichment. For completeness, the assumed contents of ²³⁸U and ²³⁴U should also be documented, or the relationship between the three isotopes that is used throughout the draft report should be explained.

Page 38 states that the release from the NRF S1W Engineering Test extended over a period of days, so the straight-line-trajectory downwind assumption using sub-15min plume characteristics is practically meaningless. DOE (1991a) modeling included a total of 349.9 h, but that document does not describe how the wind varied over that period. “Modeling” a plume whose entire release traverses Atomic City is illogical, in that the prevailing wind was 90 degrees away.

The use of nondecayed estimates is questionable particularly in view of the relative ease of performing such calculations (for example, using RSAC-5). This conservatism reaches an extreme for some of the nuclides (a factor of 2 for ⁸⁸Kr, more than 10,000 for some short-lived Xe isotopes, and more than 1 million for ¹⁴²Ba, for a 3.7-h transit time)—large enough to have more than a 10% effect on some of the screening values (such as plume-immersion screening values, where ⁸⁸Kr contributes 30%

In NRF_S1W.xls, the columns for the relative screening value are mislabeled with incorrect distances.

For IET-3, the solids release fraction of 0.0054 implies a mass of 220 g of uranium released of the total of 40,800 g given on page 70. A note in IET-3.xls indicates that the activities used in the draft report are supposed to correspond to 93.4% enriched uranium, 37 fuel assemblies, 1100 g/assembly, and a release fraction of 0.0054. That yields a total core inventory of 30,700 g and a release of 219.8 g. However, the reported

activities (in IET-3.xls) correspond to about a 1-g release of 93.4% enriched uranium, not 220 (the method used in the draft report to calculate uranium activities is not reported in the text or in any of the spreadsheets).

The draft report defers to DOE (1991a) in the estimate of ⁴¹Ar release. Some description of this calculation would assist the reader.

The discussion of measured fission-product releases in the draft report is confusing, particularly compared with the more extensive discussion in DOE (1991a). Did the 2000-Ci release during a 4-h period include the 30-Ci h⁻¹ particulate activity, or are these two additive? Why should one assume these release rates over the entire 40 h of operation above 200 kW if no release was observed until February 11? Why should the reported total release of about 4000 10-min-decayed curies (DOE, 1991a) be considered inaccurate? Were measurement methods were so inaccurate that one should prefer an estimate 10 times higher based on guesses at release fractions?

For individual facilities, it appears that the only itemization of source documents for routine releases is in the spreadsheet FacilityAirReleases.xls, where we find

ICPP releases

TRA releases

ANL-W releases

Were the measurement methods applied so inaccurate that large fractions of the releases would have gone undetected, as implied by the many estimates used in the draft report? Were the reporting requirements so modest that detected releases would have gone unreported? Or are we in a position where contemporary documents do not include the majority of the reported releases? In this context, can we assume that emissions of ⁴¹Ar at 2.8 Ci/MW-h would not have been detected by the measurement methods in use, particularly during periods when there were no fission product emissions; or were they detected but not reported in any available contemporary documents?

IET-4 (page 71-72), DOE (1991a) did not report an⁴¹Ar release rate of 56 Ci h⁻¹ above 200 kW. It reported 2.8 Ci/MW-h, and production can be expected to be nearly proportional to total burnup. These two release rates coincide only at an average nuclear heat production of 20 MW. The draft report apparently unnecessarily overestimates ⁴¹Ar production by about 38%.

The draft-report release of uranium (reported in IET-4.xls) corresponds to about 1.75 g, not the 330 implied by a release fraction of 0.0081 and a total mass of 40,800 g.

The solids-release fraction appears nonconservative compared with the 1.27% estimated in DOE (1991a), on the basis of pictures of the fuel assemblies, although some of the fuel was deposited in the “engine and 76-in. diameter ducting to the stack”. The extra information that was used in the draft report to obtain the less conservative value should be documented.

Page 72 states that “the calculated relative screening value generally increases with increasing transit time.” If that were generally true, doses to populations farther from the site would be higher than doses to closer populations. The statement should be corrected, presumably to something like “for this particular location, the calculated relative screening value increases with increasing transit time (that is, because of lower wind speed).”

The DOE (1991a) modeling was for a period of 545 h, so selecting the highest dispersion-factor direction from that modeling and applying a straight-line-trajectory downwind assumption by using sub-15-min plume characteristics is practically meaningless.

For the description of IET-6, the draft report provides pointless details, for example, about how “the new A-3 core assembly differed from the A-2 core used during IET-3 and IET-4 in that the new insulation sleeve was designed to enhance the structural integrity of the sleeve against pressure collapse”; this is pointless in that the reader has no context (diagram, description of the core, or reasons that these affect damage) in which to place it. Indeed, it appears irrelevant to the analysis of release fractions. Later, the draft report gives negligible details about the damage that occurred, except that it was “less than that during IET-3”. DOE (1991a) did not report on an ⁴¹Ar release rate of 56 Ci h⁻¹

above 200 kW. It reported 2.8 Ci/MW-h, and production can be expected to be nearly proportional to total burnup. These two release rates coincide only at an average nuclear heat production of 20 MW. The draft report apparently unnecessarily overestimates ⁴¹Ar production by about 76%.

It is difficult to reconcile the descriptions of core damage reported in DOE (1991a) and in the draft report. DOE (1991a) indicates only that minor fuel damage was assumed at the time and that later photographic evidence suggested a loss of 5 g of fuel and associated fission products, whereas the draft report assumes the equivalent of substantial damage to one fuel cartridge and the release of 110 g of uranium. According to DOE (1991a), the measured release rates during IET-6 were less than 3% of those during IET-3, but the draft report suggests a total release of activity of about 50% of that in IET-6. The extra information on which the draft report relies to substantiate the increased release estimates should be noted.

The release of uranium given in IET-6.xls is substantially smaller than documented in the draft report or in the accompanying note in the spreadsheet (0.0027 of 37 assemblies of 1100 g would be 110 g, but the activities reported correspond to about 0.6 g).

For IET-8 (pages 73-74), it should be pointed out that what was found on the filters was (presumably) radioactive isotopes of Mo and Mn. The sentence suggests that Mo and Mn are necessarily radioactive.

For IET-10, the relationship of the spreadsheet pages labeled “IET-10a (total)” and “IET-10b (total)” and the corresponding RSAC-5 documentation to the discussion in the draft report is obscure and needs to be clarified. These might be the initial runs of RSAC-5 (or some of them), but the differences between them are difficult to discern. The page labeled “IET-10c (total)” seems to be what was ultimately used.

The description of IET-10 on pages 74-75 is not understandable with the amount of detail given. It is not clear how the election to “assume that the total reactor power was produced by the test insert” is “conservative”. It is physically unreasonable, but the method of calculation used does not appear to make this assumption. (Similarly, it is not clear what is meant by footnote 11 on page 75, in that again the methods of calculation used in the draft report do not necessarily make this assumption.)

It is not clear what is meant by the statement that using the DOE (1991a) estimated emission rates “resulted in a fission product release rate that was not conservative by comparison to release rates reported by Foster et al. (1958).” According to DOE (1991a), citing Foster et al. (1958), the values were selected “so that the total fission product release calculated using the RSAC-4 computer code equaled or slightly exceeded the total fission product release estimated for individual runs from the spot sampling data.” The discrepancy should be documented, and the source of any disagreement resolved.

The release fractions finally used—0.075, 0.1, and 1.0 for solids, halogens, and noble gases—suggest extensive core damage. However, the extent of core damage in IET-10 is described in DOE (1991a) with the statement that “during the disassembly and inspection of the insert following IET #10, some of the fuel tubes in the tenth stage were white in appearance, indicating that the UO₂ fuel had been released from the BeO ceramic matrix as a result of operations at temperatures of 3000 to 3200°F (Flagella, 1962, pp. 75-76).” That does not sound like damage equivalent to release of 7.5% of the total core solids, which is what appears to be suggested by the draft report.

For IET-11, the relationship of the spreadsheet pages labeled “IET-11a (total)” and “IET-11b (total)” and the corresponding RSAC-5 documentation to the discussion in the draft report is obscure and needs to be clarified. As in the instance of IET-10, these might be the initial runs of RSAC-5 (or some of them), but the differences between them are difficult to discern. The page labeled “IET-11c (total)” seems to be what was ultimately used.

The calculation of a core inventory for IET-11 and basing release estimates on it appear to be physically unreasonable. No fuel damage was noted in IET-11, and the fission-product and other emissions were apparently attributed to uranium deposited in the exhaust ducts from the previous IET-10 and still in the neutron flux of the core. The calculation of releases with the method used in the draft report might yield a reasonable estimate of emission, but it should be noted that the uranium emissions, at least, will be double-counted (with IET-10). However, because the uranium responsible for fission-product release was coated as a thin film, it might be more appropriate to model this release with a release fraction of 1 for all fission products (or perhaps 1 for all but solids and 0.5 for solids, to account for decay recoil either out of the uranium film or into it) and to adjust the burnup to match the total emission rate.

For SNAPTRAN-3, the header of Table 20 has been copied so faithfully from DOE (1991a) that it also includes the superscript (a). However, in the draft report, this superscript is incorrect: the superscript (a) is used elsewhere in the table for another purpose. No ⁴¹Ar is included in the release, although it was included by DOE (1991a); the rationale for that inclusion is not clear, inasmuch as the core apparently was immersed in water to generate the transient. A more detailed explanation of the generation of the transient would greatly enhance the utility of the draft report.

For SNAPTRAN-2, DOE (1991a) indicates that the core contained 4.75 kg of total uranium, and that is reflected in Table A-41, although SNAPTRAN-3 is stated to contain 4.75 kg of ²³⁵U. The draft report refers to reports 4.75 kg of ²³⁵U with 93% enrichment (and implicitly 1% ²³⁴U and 6% ²³⁸U). The core weight needs to be clarified, and the fractions of ²³⁴U and ²³⁸U explicitly documented. No ⁴¹Ar release is included for SNAPTRAN-2, although DOE (1991a) includes such a release. This test was apparently in air, so such a release is likely.

Neutral to light lapse meteorologic conditions were required for this test and it is difficult to see why stability class F was used for the modeling. Was that the observed

stability condition at the time of the test? The path of the plume was monitored by aircraft out to 30 km downwind. Did it pass over any potentially inhabited offsite area?

The release in connection with the blower failure at ICPP (page 113) is included in the wrong section of the draft report. The release was included in DOE (1991a, page A-180; it is omitted from the summary Table A-19).

Many of the episodic releases were included in the routine releases reported by DOE (1991a, Table A-19). It would be useful to indicate which of the episodic releases were included in the routine releases in the draft report and to state the fractional contribution of the episodic releases to the routine release estimates and screening-factor estimates.

For IET-12 (pages 76-77), it is not clear what is meant by “release fractions . . . in tube 6”. Is “tube 6” equivalent to a “fuel cartridge” that is mentioned. It does not seem conservative to assume only 50% release of noble gases from a fuel cartridge when there is such heavy damage (42% of the fuel unaccounted for).

For IET-14 (page 78), the object of this test is stated to be “to measure the fission product release rate”. But, these measurements appear to be disregarded later in favor of the draft report’s own estimate of release, higher by a factor of 5.7 than that reported. In other cases (such as IET-10, for which measurement of the fission-product release rate is not reported as an object of the test), it appears that there was some effort to calibrate the draft report’s estimates against measurements. The approaches appear inconsistent and would introduce arbitrary factors of conservatism in some estimates.

For IET-16 (page 79-80), the selection of Cerro Grande as the “nearest offsite location in the general direction of the prevailing winds” seems disingenuous. For “general direction”, Atomic City is somewhat closer to TAN. It appears more likely that Cerro Grande was selected because DOE (1991a) indicated that it was the location with the highest dispersion coefficient.

For IET-17, IET-18, IET-22, IET-24, and IET-25, the factors of conservatism between measured and estimated releases appear to be about 2, 18, 125, 181, and 8, respectively. They reinforce the appearance of arbitrary factors of conservatism.

For FEBT-A (page 87), a meteorologic lapse does not correspond to “increasing temperature with increasing elevation”, and the previous sentence states that the temperature decreased with elevation (although the sentence is correct in stating that the measured values correspond to a lapse condition).

For FPFRT (page 91), the disagreement in tellurium (Te) isotope inventories reported in Table 16 is curious and requires explanation. How was the “reported inventory” obtained? Did contemporary versions of programs like RSAC produce overestimates of Te? The use of the highest value in Table 16 is also curious. One or the other of the inventories should be best. Either RSAC-5 is the best tool for the job, or some factor (known to the original investigators) was not modeled by RSAC.

More detail on the October 1959 ICPP criticality would be desirable (as is true for most of the release descriptions). The reported energy release is sufficient to vaporize about 280 kg of water; at first sight this seems unlikely for a process designed to operate in critical-safe containers.

In Table 19 (page 100), the “stability parameter” needs some explanation, and the relation (if any) to the selected stability categories should be identified.

For CERTs (page 105), lists of the tests with the quantities released would improve the presentation. DOE (1991a) reports that CERT No.l was released north of Atomic City, rather than at the Experimental Dairy Farm, and cites Hawley (1966) as a source. The locations of release should be confirmed.

For LDTs (page 105), it would improve readers’ understanding to state that these releases were from the ICPP stack under controlled conditions. DOE (1991a) indicates that the quantities of ⁸⁵Kr were unknown; is this detail provided in the cited references?

For the MTR Fuel Melt Incident (page 110), was debris blocking flow in 40% of all fuel-assembly channels in the entire core, as implied here? DOE (1991) indicates two channels of one element.

For ETR Fuel Melt Incident (page 110), is De Boisblanc the author of that reference? DOE (1991a) indicates that Covington was the editor.

On Appendix A , the inclusion of cooling towers in the table requires explanation. Were there any releases from cooling towers? It is not clear that the list is complete; for example, the IET releases were from a stack that had an “80 ft level”, but the only stack listed for TAN is 40 ft high.

Some of the entries in the table might require attention. For example,

EBR-II: The circulation rate of 76,000 1/month and blowdown of 100-190 1/month, appear too low.

CFSGS: At what time is/was the “monitoring pending”?

Power Burst Facility: The listed cooling-tower capacity and maximal flow are not credible.

ATR: The listed evaporation rate appears extremely low.

ETR: The listed evaporation rate appears extremely low.

RSAC-5 was used for many calculations, and this appears to be appropriate. However, the use of RSAC-5 has disadvantages for a report that has to be transparent to the public. The user manual for RSAC-5.2 is freely available at http://ar.inel.gov , and a better copy is available through the Radiation Safety Information Computational Center (RSICC) at Oak Ridge at http://www-rsicc.ornl.gov/rsic.html . The executable code is available from RSICC, apparently at RSICC’s discretion. Obtaining it requires registration, the filling in of several forms, and the prepayment of $630; this is both intrusive and prohibitive for an individual member of the public.

The program has been “subjected to extensive independent verification and validation” (see http://www-rsicc.ornl.gov/codes/ccc/ccc1/ccc-125.html ) , or so states the preamble to the user manual obtained from RSICC.

1. In Offsite_Air_Screening.xls, an inappropriate comment related to ⁹⁹Tc appears in the 1952 column in the “All years – total” sheet.

2. In the “Sorted all years average” sheet of Offsite_Air_Screening.xls, many (perhaps all) of the comments are in inappropriate cells (they presumably did not get sorted with the rest of the entries).

3. The entry of 2.06E-04 for ²³⁸Pu in 1952 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with Table A-40 of DOE (1991a), which assigns that value to ²³⁹Pu.

4. The entry of 5.45E+00 for ⁹⁵Zr in 1955 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of 5.45E+01 in Table A-39 of DOE (1991a).

5. The entry of 4.02E+05 for ^85mKr in 1956 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of 4.02E+03 in Table A-39 of DOE (1991a).

6. The entry of 5.09E+00 for ¹³⁷Cs in 1958 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of 5.09E+01 in Table A-39 of DOE (1991a).

7. For ¹³²I, the entry for 1959 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls is 1.19E+04. An accompanying note states that “an annual total of 1330 Ci of !-132 (sic) was reported in HDE (DOE 1991a). Data from daily and monthly reports give total from ICPP alone of about 11,000 Ci for October 1959 due to a criticality on October 16, 1959. Although this criticality event was treated as an episodic release, too, we assumed that all releases in our assessment were routine for a conservative screening approach.” In fact, DOE (1991a) has an entry of 1.55E+03 Ci, not 1.33E+03, for routine releases (Table A-39 on page A-194). Furthermore, the explanation given in the note is not believable, given that the total inventory of ¹³²I calculated for this criticality according to RSAC-5 at 30-min decay is 9.17 Ci (ICCP_1959_Criticality.xls). DOE (1991a) estimated a release of 11.8 Ci at 0.52-h decay for this criticality (DOE, 1991a, Table A-41 on page A-199 and page A-174) on the basis of a release fraction the same as for ¹³¹I, which had a measured release of 3.73 Ci. The committee does not have the RSAC-4-calculated inventory of ¹³¹I used by DOE (1991a) to estimate a release fraction, but RSAC-5 apparently produced an inventory of about 4 Ci. As noted in the draft report, DOE (1991a) determined that the initial reports of release of ¹³²I to the environment were in error, and the draft report apparently agrees with that judgment (page 97). However, the

source of the 1000-fold discrepancy introduced by the note in the spreadsheets requires further elucidation.

8. The entry of 3.06E+00 for ¹³⁷Cs in 1963 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of 3.06E+01 in Table A-38 of DOE (1991a).

9. The entry of 3.41E+00 for 91mY in 1964 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of 3.73E+00 in Table A-38 of DOE (1991a). Indeed, the entries for ^91mY for the years 1964-1967 all differ from those in DOE (1991a), although they are identical with the entries in DOE (1991a) for 1965-1968.

10. The entries of 5.67E+00 for ¹³⁶Cs, 6.68E+02 for ¹³⁷Cs, and 0 for ¹³⁸Cs in 1968 in the “all-years” sheets of Onsite_Air_Screening.xls and Offsite_Air_Screening.xls do not agree with the values of 0 for ¹³⁶Cs, 5.67E+00 for ¹³⁷Cs, and 6.68E+02 for ¹³⁸Cs 3.06E+01 in Table A-38 of DOE (1991a). These values were propagated to the individual year (1968) in Onsite_Air_Screening.xls, but in Offsite_Air_Screening the values for the individual year agree with DOE (1991a), not with the entries in the “all-years” sheets.

11. The entry of 2.17E-01 for ¹⁴⁰Ba in 1968 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of 2.17E-02 in Table A-38 of DOE (1991a).

12. The entry of 6.0E+02 for ¹³³I in 1969 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of zero in Table A-37 of DOE (1991a). Similarly, the entry of zero for ¹³³Xe in 1969 does not agree with the value of 6.0E+02 in DOE (1991a).

13. The entry of 6.23E-03 for ²³⁹Pu in 1969 in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls does not agree with the value of zero in Table A-37 of DOE (1991a). It appears to have been transposed from the value for ²³⁸Pu, which has the value 6.2E-03 in DOE (1991a) but zero in Onsite_Air_Screening.xls and Offsite_Air_Screening.xls.

INEEL Press release at http://www.inel.gov/whats_new/press_releases/1995/hre2.html (at 3/17/2001).

FOR IMMEDIATE RELEASE

July 31, 1995

RADIATION EXPERIMENT RECORDS MADE AVAILABLE TO THE PUBLIC

Historical records found at the Idaho National Engineering Laboratory concerning past radiation experiments are available in the Department of Energy public reading room at University Place, 1776 Science Center Drive in Idaho Falls.

The records were found by a special team formed at the INEEL in early 1994 to conduct a thorough and comprehensive search of documents dating to the site’s birth in 1949. The team was formed to support Secretary of Energy Hazel O’Leary’s Openness Initiative, which requires disclosure of all past radiation experiments involving humans at DOE sites, and an Advisory Committee on Human Radiation Experiments established by President Clinton.

A summary report, Human Radiation Experiments: The DOE Roadmap to the Story and the Records, was released in February and also is available in the public reading room. This report describes human radiation experiments involving the release of radiation to the environment and records identified by DOE sites throughout the country, including the INEEL.

Last year, the team identified five actual or planned human radiation experiments which occurred at the INEEL. They were announced in news releases on Sept. 23, 1994, April 11, 1994 and June 27, 1994.

Another study was identified earlier this year involving intentional releases of iodine-129 into the environment. This experiment is being referred to as the Iodine-129 Technology Studies. The experiment involved no intentional exposure of humans to radiation and was not classified.

The experiment took place in August, 1964.

The Iodine-129 Technology Studies was conducted to examine the atmospheric mixing and dilution of gases and particles containing small amounts of iodine-129. There were a total of five tests: two with particles, one with gases, and two more with particles and gases combined. The first three tests were sampled to distances of about 10 miles over a densely instrumented grid located in the center of the INEEL site. The last two tests were sampled at distances of 25 to 35 miles in off-site area to the north-east of the point of release by use of an environment monitoring network. Less than 1 millicuries of iodine129 were released during the experiment. Although a formal dose calculation has not

been found, the amount of radioactivity involved was small when compared to current routine operational releases from INEEL facilities. For comparison, 98 millicuries of iodine-129 were released from INEEL facilities in 1993 that produced a hypothetical maximum exposure of an off-site individual to 0.3 percent of the current radiation protection standard established by the Environmental Protection Agency.

The studies were a collaborative effort of the U.S. Weather Bureau Research Station at the National Reactor Testing Station and the Nuclear Science and Engineering Corporation of Pittsburgh, Penn.

-- INEL --

Media contact: Isabel Valle (208) 526-9906

95-27

1. Page vii, regarding pages 125, 128, 129, 133, 134, 135, and 140: The correct numbers are 124, 127, 128, 132, 133, 134, and 139, respectively.

2. Page ix: Figure 21 is omitted from the list of figures.

3. Page x, regarding pages 124, 127, 128, 130, 132, 133, 138, and 139: The correct numbers are 123, 126, 127, 129, 131, 132, 137, and 138, respectively.

4. Page 6, first paragraph, line 2: Change the first “was” to “were”.

5. End of page 8 and beginning of page 9: "an ddicharged" should be "and discharged”.

6. Page 9, line 1: “maintined” should be “maintained”.

7. Page 9, line 1: "radation" should be "radiation".

8. Page 9, line 9: "monitorinf" should be "monitoring".

9. Page 12, third paragraph, line 5: “leel” should be “level”.

10. Page 23, last paragraph, line 5: “mullicuries’ should be “millicuries”.

11. Page 25, fifth paragraph, line 6: Change “was” to “were”.

12. Page 29, last paragraph, lines 2-3: Size of the RWMC increased from 13 acres in 1952 to 88 acres in 1988 and 144 acres by 1970? Something wrong?

13. Page 30, line 7: "50-300 feet wide, 250-100 feet long" should be "100-250 feet long or 50-100 feet long"?

14. Page 31, line 2: "as the site" should be "at the site".

15. Page 38, last paragraph, lines 1-2: "releases of ¹³N twice that reported" Should be "releases of ¹³N as twice that reported".

16. Page 49, second paragraph: Remove closing parenthesis after “⁶⁰Co”.

17. Page 49, line 4 from the bottom: “under which they lie” is awkward.

18. Page 56, first paragraph, line 1: “following is some information ..” is awkward.

19. Page 64, line 1: “equation used to address”, equations do not “address” anything!

20. Page 64,second paragraph, line 1: “concnetration” should be “concentration”.

21. Page 132, equation 15: Add "x" to the exponential term, that is. “e^−µx”.

22. Page 132, equation 17: "2" should be an exponent, that is “ln[(R² + h²)/h²]”.

23. Page 133, define “X”.

24. Page 133, formula and units: The use of “hr” and “h” for “hour”. Throughout the report, replace "hr" with "h". Change “/” to “(“ in the second line of the calculation.

25. Appendix A, Page 2: The last unit, “16 m s³ s⁻¹”, should be “16 m s⁻¹”.