Conferences, Videos & Testimony

Report of March 3, 1997 Workshop

by Kevin L. O'Neill

March 3, 1997

report author: Kevin L. O'Neill with review and concurrence of David Albright, Chair of the Source Term Subcommittee Final draft submitted: May 19, 1997


On March 3, 1997 the Source Term Subcommittee of the Health Advisory Panel (HAP) sponsored a workshop to provide an opportunity for more discussion with concerned stakeholders about airborne releases of plutonium from the Rocky Flats Plant that occurred as a result of the fires in 1957, 1969 and through routine operations. The workshop was held because of the apparent need to bring the source term investigation to a close in the near future, despite remaining uncertainties about these events. These releases have been investigated by Radiological Assessments Corporation (RAC), who are preparing reports for the HAP. Documents and records that have helped RAC to estimate the quantities of plutonium released from the two fires and routine releases have been found during the investigation. However, it is increasingly unlikely that further original documents and records will be found. Consequently, many questions about the fires in 1957, 1969 and routine releases may never be answered definitively. Nevertheless, a large amount of information and many assessments have been generated about these plutonium releases, resulting in a better understanding about these releases than ever before. The workshop was designed to give the public an opportunity to obtained detailed information about RAC's conclusions and to question the principal RAC investigator, Paul Voillequé, about his findings. The workshop sought to provide more detailed and thorough comments than is possible at HAP meetings or other public meetings sponsored by HAP or the Colorado Department of Public Health and the Environment. The workshop's primary objective was to identify the agreements and disagreements that remain about the source terms, and to identify those issues that need to be addressed during the remainder of the investigation. It also proved to be a useful public education tool, notably by giving the public an opportunity to thoroughly understand RAC's assessments and to constructively challenge these assessments. While the workshop intended to provide an opportunity to discuss all three of these source terms, there was insufficient time for RAC to give a presentation on the 1969 fire. This was primarily a result of the detailed presentations given on routine releases and the 1957 fire, and the accompanying questions, discussions and comments from workshop attendees. The purpose of this report is to summarize the comments at this workshop, to identify those areas where agreements and disagreements exist, and, based on the comments received from stakeholders who attended the workshop, to recommend further issues for investigation.

"Routine Releases" from Building 771 and 776

Paul Voillequé described the effort to estimate plutonium releases from routine operations. This effort has concentrated on the main exhaust duct downstream from all inputs to the ventilation system and downstream from the filters. The design of the main exhaust plenum and sampling systems, notably a large bank of HEPA filters and a single sampling point downstream from the filter bank, were very similar in the both Building 771 and 776. The assessment used concentration ratios developed from the data from Building 771, because the concentrations were higher. The presentation concerned primarily a discussion of the Building 771 data and the methodology used by RAC. Voillequé explained the methodology that was adopted to estimate these releases. He noted that for the buildings' early years of operations, air samples collected by Dow Chemical to measure releases were collected at a single point in the main exhaust duct. Since single point sampling data are insufficient to estimate releases, RAC's principal task has been to estimate air concentrations in unsampled areas of the exhaust duct. RAC made these estimates by determining concentration ratios in other areas of the duct with respect to areas where samples were taken in later years. Three-point sampling along a center line in the exhaust duct, adopted by Dow Chemical in 1963, enabled RAC to develop concentration ratios to the central sampling point, which had been employed since operations began. These ratios, plus additional duct design information, were then used to estimate concentration ratios for other unsampled areas of the duct, both before and after the additional two sampling points were installed. Monte Carlo techniques were then employed, based upon different centerline concentration ranges at the sampled areas, to estimate average monthly releases through the duct.

Comments on RAC's Routine Release Estimates

Comments on RAC's estimates of routine releases were generally favorable. Most participants agreed that RAC's methodology was sound. The exhaust systems of Building 771 and Building 776 are the designed airborne pathways for plutonium to escape to the environment during the course of routine operations. Samples upon which RAC based its estimates were taken downstream from all inputs into the system and from the main filter banks. Therefore, the methodology assumes that all significant airborne releases were accounted for in the sampling data (ie., exited through the exhaust duct and out the stack). One workshop participant said after the presentation that he accepted RAC's estimates of plutonium releases from routine operations. In a letter to Voillequé, this participant noted further that "it appears to the credit of the operators that, despite increased production [over the years], routine losses decreased." Participants also agreed that records and physical evidence needed to check RAC's estimates no longer existed. Daily air sampling records from the early years of the plant have not been found and are presumed to have been thrown away. RAC has noted that in the absence of daily records, weekly or monthly reports have been used. Additionally, fan operating records, which would help inform estimates of effluent mixing in the exhaust duct, thereby checking estimated concentration ratios relative to the centerline sampling point, no longer exist. Finally, the original ductwork and fans were replaced long ago, making it impossible to test the mixing in the duct. Many participants who attended the presentation had no comments. This silence is difficult to interpret, since it is assumed that all who attended the workshop did so because they were skeptical of RAC's findings. To be sure, Voillequé's presentation did not convince every one present. Although most participants agreed that the sampling data used by RAC is as complete as possible, a few challenged the veracity of this data. Plant workers may have improperly recorded data, falsified data, or never recorded sample data at all. Voillequé said that he had no indication that plant workers would be motivated to falsify data in this way, particularly in the early years of the plant's operations when releases were the greatest. A few participants disagreed with the methodology, noting that numerous fires, filter inefficiencies and leakages in and around the main filter plenum suggested that there were higher releases. For example, one participant noted that for lower concentrations, the range of concentration ratios with respect to the centerline concentration was highly variable, suggesting that concentrations on the outside part of the duct may, in some cases, be higher than those estimated by RAC. This participant suggested that these higher concentrations could be produced by leakage around the filters, and would not be picked up by the center-point sample collector, as the effluent in the duct would not be well mixed if the fans were not operating at optimal speed. In response, Voillequé pointed out that leakages through and around the main filter bank occurred at all times, and that the whole range of concentration ratios were considered by RAC in estimating releases. In any case, these participants offered no other methodology as an alternative to the one used by RAC.

Further Work

The presentation and comments made at the workshop indicated that further work had to be accomplished before the routine release estimate was complete:
  1. Voillequé noted during his presentation that flow rate uncertainties, issues related to sample collection and factors affecting sample analysis were addressed by ChemRisk during Phase I of the Dose Reconstruction Project. During its investigation, ChemRisk calculated an uncertainty bias to account for these factors, which RAC accepted. However, Voillequé agreed that further investigations into alpha particle absorption by the sample filters will be completed before the Phase II report is finished. These changes may increase estimated releases from the Buildings.
  2. One participant noted during the presentation that RAC has to include all of daily release records that it had collected as part of the investigation. It was clear from the presentation that many gaps in these records existed. If these gaps coincided with operations that might have led to significant releases, then the total release to the environment could be higher than RAC's estimate. RAC needs to determine when these gaps occurred and check if events coincident with these gaps could have led to higher releases.
  3. RAC should take note of anecdotal evidence indicating that sampling data was improperly recorded by plant employees, falsified after the fact, or never calculated. However, if such evidence is to be incorporated into RACs's estimates in a meaningful way, then stakeholders who are aware of former plant workers who possess such evidence have an obligation to cooperate with RAC or HAP to identify these workers or at least to obtain their statements.

Releases from the 1957 Fire in Room 180, Building 771

Voillequé next detailed RAC's attempt to estimate releases from the 1957 fire in Room 180 of Building 771. He described this analysis as an "inside-out approach." RAC's analysis attempts to reconstruct the amounts of plutonium available for release from different sources, identify release pathways, and apply various empirical airborne release fraction estimates for plutonium fires. Key issues raised by this investigation include: the amount of plutonium in room 180 at the time of the fire; the amount of plutonium deposited on the glovebox outlet filters, booster system and main filter plenum at the time of the fire; how and when the fire spread from the glovebox system to the booster system and main filter plenum; fan operations during the fire; the severity of the fire, and of the oxidation of the metal in Room 180; and the nature and location of the reported explosion in the main filter plenum.

Comments on RAC's 1957 Fire Release Estimates

RAC's estimates of releases from the 1957 fire were more controversial than its estimates of routine releases. Principally, this was due to the inability to conclusively reconstruct the events of the fire. Some comments were quite favorable, despite the uncertainties that remain. Many participants agreed that RAC's investigation into the 1957 fire was detailed and informative, with one key participant declaring the presentation to be "thorough and complete." As in the case of the routine release presentation, many participants simply had no comments at all, despite their obvious interest in RAC's findings. On the negative side, some participants refused to accept RAC's findings. Many of these reactions were highly emotional. One workshop participant even accused Voillequé of holding a preconceived bias that releases from the fire were small. Most comments were thoughtful and constructive. One participant noted in a letter to Voillequé following the workshop that the source term reconstruction had been "carried as far as it can go." The "inside-out approach was a noble attempt but not too convincing," this participant wrote, "too many 'guesses' had to be made."

Further Work

The presentation and comments made at the workshop indicated that further work had to be accomplished before the estimate was complete:
  1. The key uncertainty of the 1957 fire release estimate concerns the probability that the scenario adopted by RAC reflects the actual events of the fire. Most disagreements over RAC's conclusions stem from disagreements over what actually happened. Since the actual events can not be conclusively reconstructed, RAC should present a range of possible scenarios, some more probable than others, that can bound the releases from the event. For example, RAC might consider a worst-case scenario where a sufficient area of the filter plenum, approximately 2 square feet, was displaced early in the fire such that the effluent from the fire essentially vented up the stack unfiltered. By developing such worst case scenarios, and explaining why these scenarios are plausible or implausible, RAC's estimates would become more credible.
  2. One participant noted that RAC's assessment did not include an estimate of plutonium released to the environment after the fans went off and before water was used to put out the fire in the main filter plenum. The participant noted that the "chimney" effect during the later stages of the fire would carry smoke, ash, and plutonium up the stack. Voillequé agreed to include this analysis in the next draft of the report.
  3. RAC must defend more conclusively its assessment that the explosion took place on the upstream side of the main filter bank since, as several participants noted, in prior assessments RAC had reached the opposite conclusion. Alternatively, since the true location of the explosion may never be known, RAC might accept that both scenarios are possible, and incorporate both scenarios into its uncertainty analysis.


The sponsors of the workshop have concluded that the workshop was extremely useful and satisfied many of its objectives. Voillequé's detailed and thoughtful presentations gave attendees a deeper understanding of plutonium releases from routine operations and from the 1957 fire. Stakeholders who attended this workshop responded with a wide range of comments about the presentations given by RAC. Several attendees raised technical issues which need to be addressed. RAC also needs to provide the public with more "worst-case" analyses to provide reasonable bounds on the source terms, especially where disagreements over fundamental issues remain unresolved. It is clear from the presentations and the comments that RAC's source term estimates are extremely detailed. These estimates are based upon a large body of information that, until recently, was unknown and inaccessible to the public. However, the workshop revealed that RAC's assessments remain controversial; incomplete data and information remains a significant problem, and many questions that could affect estimated releases from Rocky Flats remain unanswered. Significantly, however, the extent to which this data is incomplete, lost or destroyed is better understood now than ever before. The workshop has shown that, by itself, source term reconstruction has its limits in terms of public acceptance and credibility. Consequently, a reassessment of off-site environmental monitoring data, including plutonium isotopic analysis, may become more important. Notwithstanding, RAC should press ahead with the source term reconstruction. These further investigations should take into account the comments of workshop participants and the specific recommendations made in this report. Workshop sponsors also recommend that additional workshops be conducted over the next several months in parallel with scheduled HAP meetings. If nothing else, the workshop revealed the need to carefully educate the public about these releases. One participant who had heard some of Voillequé's presentations and was familiar with the final analyses commented that it "took a few times to sort it all out." He said that the release estimates were very complicated, and that it was difficult for anyone to understand these estimates after hearing the analysis only one time. Workshop participants told the sponsors that they felt the public education value of this event was high. Additional events would help to build public credibility in the HAP's efforts to understand the plutonium releases from Rocky Flats and the associated risk to the public. The next workshop should take place on a weekend so that stakeholders have fewer conflicts with their occupations and are better able to attend. Topics to be discussed at this next workshop or workshops include the releases from the 1969 fire and the 903 pad. Further investigations into routine releases and the 1957 fire might also be discussed. Technical reviews of RAC's analysis should concentrate on specific issues, such as how different possible locations of the explosion in the 1957 fire affect releases. However, it is probably unnecessary to subject RAC's assessments to full-scale technical reviews by outside investigators. As one workshop participant noted, the Phase II source term reconstruction is already a review of the Phase I estimate of releases from the plant. In any case, there is a need to close on these estimates in a reasonable amount of time, even if all of the outstanding issues are not resolved to everyone's satisfaction.

email us twitter share on facebook