Comments on Review of Radiological Monitoring at LBNL Preliminary Technical Report by Bernd Franke and Anthony Greenhouse of IFEU.
December 3, 2000
The main weakness of the IFEU report is that it does not focus closely enough on the question as to what extent do NTLF tritium emissions pose a threat to workers and visitors at the Lawrence Hall of Science. This important issue is obscured by several irrelevant side issues. Also, the report fails to consider two key lines of environmental evidence that indicate tritium levels have been much higher at the Lawrence Hall of Science than previously reported: the Menchaca organically bound tritium data and the Monheit rainfall data. I urge both the City and the authors to restructure the Scope of Services so that this evidence can be reviewed.
Another troublesome aspect of the report is that the authors always accept the LBNL position on controversial issues without considering alternative explanations. For example, they repeatedly state that there is no evidence that the 10 mrem limit has been exceeded even though they have not yet considered all the relevant evidence. A cynical interpretation would be that they are not truly independent of LBNL-DOE. However, I do not believe that this is the case. A more plausible explanation is that they were overwhelmed by a work plan that was impossibly broad in scope and therefore lost sight of the main issue.
Another general point I would like to make here is that LBNL-EHS has hindered my writing of this review by blocking my computer from the IFEU Report and SENES Comments sections of their tritium issues web site. At least one other Berkeley citizen concerned about the tritium issue has also been prevented access to these parts of the site. This selective blocking of access is particularly troubling and is further evidence that a cover-up is in place. On the same theme, I would like to know how LBNL received a copy of the report on June 30th whereas concerned community members did not receive it until July 5th one day before the LBNL Press Release was issued. Furthermore, why was the hard copy of the report eventually provided to community members incomplete in that a key graphic, Figure 5, lacked both labels and a caption? On the positive side, I must acknowledge that Bernd Franke has been most helpful in responding to my questions via email and in sending me data files.
LBNL employees who are responsible for communicating with the public on the NTLF tritium issue should be reminded that in the spring of 1997 Energy Secretary Federico Peña fired the Director of the Brookhaven National Laboratory and terminated the contract of Associated Universities Incorporated (AUI) in part because of poor communications with the public. Furthermore, the controversial issue there was a tritium leak that was trivial in comparison to the releases at the NTLF.
Finally, I would like to emphasize that even though some of the comments that follow are critical of the NTLF's proximity to the Lawrence Hall of Science, I have no doubt that the research done at the NTLF is important and that it should continue. However, I also feel that if the evidence shows that exposure levels at the Lawrence Hall of Science are even close to the NESHAPs 10 mrem standard, the facility should be moved to a more remote location.
A.1 Is the tritium inventory at NTLF adequately determined?
The authors report that the accuracy of the inventory data is very limited (+/-
30%) and therefore the data cannot be used as a check on the accuracy of reported releases. The question as to why the inventory data are so inaccurate is not clearly explained. In their final report the authors should include accuracy estimates for the various components of the inventory: (1) the amount of tritium purchased; (2) the amount shipped out as product; (3) the amount recycled; (4) the amount lost through radioactive decay; (5) the amount stored and/or shipped as radioactive/ mixed waste; and, most importantly, (6) the amount that goes up the stacks. Also, is any other nuclear facility within the DOE system allowed such a large range of uncertainty in its inventory reporting?
Community members have been particularly troubled by contradictory statements coming out of LBNL regarding the NTLF tritium inventory. The official version is that of the 10,000 Ci normal annual intake into the facility 80 percent is recycled, -20 percent is shipped as radioactive/ mixed waste, 1-2 percent is shipped as product, and 1 percent goes up the stack These figures are not supported by the evidence.
For example, it is claimed that since 1990 a secondary uranium bed recovers 8090 percent of the tritium used during labeling experiments, and that this recovered tritium is then shipped to other facilities for purification and recycling (letter from Ron Pauer to Nabil Al-Hadithy, February 18, 1994). In actual fact, during the years 1990 to 1996 less than 30 percent of the tritium coming into the facility was recycled (LBNL-N TLF Tritium Inventory dated 7/15/97). Furthermore, this 30 percent figure may be inflated because on October 28, 1998 a recycling shipment to LLNL labeled as containing 6,850 Ci was determined at Livermore to have only 3,300 Ci (Personal communication from Mark Mintz of LLNL to Pamela Sihvola, Co-Chair of the CMTW, February 3, 1999).
The radioactive/ mixed waste component is reportedly more difficult to quantify, but the question remains as to how much is generated each year. LBNL officials in trying to avoid having the new Hazardous Waste Handling Facility classified as a "nuclear facility" are on record stating that annual radioactive waste generation would be only 20 percent of the annual intake, i.e., - 2,000 Ci (letter from Klaus Berkner to Scott Samuelson, March 9, 1994). However, during the years 1990 to 1996 radioactive waste shipments totaled 38,500 Ci or - 50 percent of the incoming total for those seven years.
The inventory issue is not simply a question as to whether one can predict stack emissions from tritium inventory. The more important question is whether or not the public can trust the inventory data as reported by LBNL. If the inventory data cannot be trusted, why should the public trust any of the LBNL tritium data?
A.2 Are releases of airborne tritium accurately monitored?
This is the key question. Ever since the first reports of high tritium levels near the NTLF stack reached the public in 1995, there has been concern that the stack emissions were not being accurately monitored. Furthermore, LBNL's refusal to make available stack monitoring data has only increased these concerns. Finally, under pressure from the City of Berkeley, LBNL has provided data for the silica gel and realtime Overhoff stack monitors for 1998 and 1999 and summary data for 1995 through 1999 to the authors of the IFEU Report. Bernd Franke kindly provided me with these files. Since all of them take up less than 35 megabytes of disk space the question arises as to why LBNL asked the Committee to Minimize Toxic Waste to pay $42,000 when they asked for emissions and inventory data via a FOIA Request in 1998. A 650 megabyte CD, such as the one provided by LBNL to Franke, costs only a couple of dollars. I recommend that the authors request LBNL that all stack emissions data from 1995 to the present, Overhoff and silica gel, be made available on the LBNL-EHS web site in files that can be conveniently downloaded. CAP88 files and weather data should also be made available to the public.
This section of the report concludes on page 12 with the statement "there is no evidence to suggest at this time that offsite exposures resulted in radiation doses exceeding the 10 mrem/yr limit for any individual. " This conclusion is inappropriate because neither the environmental data nor the dispersion modeling have yet been adequately reviewed. The authors conclude that NTLF stack emissions are accurately monitored and reported although in doing so they overlook key evidence to the contrary
The silica gel stack monitors
It is not made clear in the text that the current silica gel sampling system consists of two columns; the first samples HTO, and the second HTO plus HT that has been converted to HTO. The latter does this because the air stream passes through an oxidizer before it enters the silica gel column. This system has been in place only since March 1999 (a typo at the top of page 9 indicates 1998). Prior to this date all measurements of HT were unreliable, an incredible revelation. The report states that the Overhoff data were used to estimate HT emissions in 1998. Was this also the case prior to 1998 or were the unreliable silica gel numbers used?
On page 8 it is reported that the silica gel system has a sampling and analysis error of -20%. Does this mean +/- 20 percent? On page 9 the statement is made that "the silica gel data for HTO appears to be reliable." How do we know this? Is the silica gel weighed at the beginning and at the end of the sampling period? Have samples been split and tested independently by different labs? Does the mass of water collected in a week correlate well with atmospheric humidity? Is the air pump accurately calibrated? Figure 4 shows clearly that there is no correlation between the silica gel weekly samples and the Overhoff real-time measurements. How can we be sure the silica gel results are reliable?
The Overhoff stack monitor
On page 9 of the Report it is stated that there are systematic problems with Overhoff real-time data: electrical spikes, temperature effects, high instrument background, etc. On the next page it is stated that the March 1998 HTO data presented in Figure 5 suggest significant fluctuation of instrument background due to changes in temperature and other factors. How do we know that this is the case? Perhaps the Overhoff system is accurately monitoring real variations in tritium emissions. Was the Overhoff Company contacted so that they could comment on how their system is being characterized by LBNL-EHS? Also, if the Overhoff system is so unreliable why was it not upgraded or replaced? One implication here is that LBNL prefers to have an unreliable system so that unwelcome high numbers can be conveniently discarded or factored down. This brings up the question of the 220,000 Bq/m3 "background."
The 220,000 Bq/m3 "background"
Perhaps the most troubling revelation in the Report is that LBNL-EHS subtracts 220,000 Bq (5,940,000 pCi) from the Overhoff real-time data for every cubic meter of air emitted from the main stack. In the caption to Figure 5, which is missing in copies of the report distributed by the City of Berkeley, the authors state "visual inspection shows the actual background may have been lower." This gentle criticism is as far as the authors take this issue. Why was LBNL-EHS not asked to explain in detail why such a high "background" is subtracted? How do they know there is an artificial background? The authors correctly point out that 220,000 Bq/m3 is equivalent to about 4 Ci a week, or 208 Ci a year. This is twice the 100 Ci total often cited as the annual average source term. One other interesting aspect of the Overhoff "background" is that LBNL-EHS subtracts it proportionately from the HTO and HT totals. This may be convenient for accounting purposes but is there any physical justification?
The stack airflow rate
In checking the data files made available to me by Bernd Franke it became apparent that different values were being assigned to the stack air flow rate. I asked him about this in an email note and he replied (email from Bernd Franke to Roger Byrne, 8/15/00) that he had noticed the same thing and asked LBNL-EHS why this was the case. Here is his question and their response:
<<1 noticed that flow rates in the spreadsheets with silica gel data is given to be 1.1 m3/s whereas the flow rates in Overhead (sic) data sheets are different, often larger. What is the reason for the difference and what is LANLs best estimate for NTLF flow rate.>>
LBL response: The stack flow rate varied throughout the year. For the period 1/1/98 -12/31/98, the average stack flow rate was 1.1 m3/s. 1.1 m3/s was used to determine the 1998 annual tritium emission. If weekly average flow rates were used to calculate the 1998 weekly releases (instead of the annual average flow rate of 1.1 m3/s), the calculated 1998 release from the NTLF increases from 115.2 to 117.5 Ci. This is approximately a 2% difference.
This answer does not indicate whether or not the silica gel sampler and the Overhoff realtime sampler are sampling at the same location on the trunk line. The NTLF Safety Analysis Document (SAD) indicates in its Figure 4-B that there are two EHS monitors on the stack side of the 4.66 m3/s blowers. If these represent the Overhoff monitor the airflow for the Overhoff should be 4.66 m3/s rather than 1.1 m3/s. This important matter obviously needs to be investigated.
Short term releases
On page 12 the authors conclude that the most significant conclusion is that realtime data show that tritium is often released in short events. This is echoed in the Executive Summary where it is stated that short term releases render the results of CAP88 to be inaccurate. This is not true. It is correct that CAP88 should not be used to model large "unplanned" releases because it is designed to use weather data that have been accumulated over a whole year. On the other hand, the short term releases that are shown in figures 5 and 7 are appropriately dealt with by CAP88. The 80+ tritiations carried out yearly at the NTLF, when it is operating normally, are more or less evenly spread out through the year. CAP88 is therefore an appropriate model to use if one has to calculate annual exposure levels.
LBNL-EHS's recent disenchantment with CAP88 most likely derives from my having pointed out that they are using it incorrectly by setting the stack height to 10 m and by using inappropriate weather data (see my report to NIH in 1999). If the source term is 100 Ci and the parameters for stack height and wind speed are set correctly, the dose estimates for the maximally exposed individual at the Lawrence Hall of Science exceed the EPA compliance level.
I agree with the authors that the short term nature of individual releases is significant. It is also unfortunate for the workers and visitors to the Lawrence Hall of Science that these releases tend to occur on weekday afternoons. Just a couple of weeks ago, I myself visited the Lawrence Hall of Science and encountered a young family with three children walking along the fence line 50 meters northwest of the NTLF stack. When I told them about the stack they quickly left the area. Had I not been there they may have sat down and had lunch on the slope just next to the fence. Coke bottles and beer bottles in the area indicate that this sometimes happens. My reason for bringing up this anecdote is to ask the authors to calculate what the air concentration would be in that area if we assume a release similar to the typical release shown in their figure 7.
Do the authors think it would be useful to have a radiation sign on the fence indicating to the public that tritium is being released? Also, would the authors comment on my suggestion that all the Eucalyptus trees on the slope between the NTLF stack and the Lawrence Hall of Science be cut down because: (1) they are a serious fire hazard; (2) they are all natural tritium stacks evapotranspiring unknown quantities of tritium; and (3) they reduce wind speed by a factor of 2 or 3 thereby increasing the dose to anyone in the plume at the Lawrence Hall of Science by that same amount?
The July 24, 1998 "unplanned" release
One surprising omission from the Report is the lack of any discussion of the "unplanned" release of July 24, 1998. At least thirty five Ci of tritium were released when a silica gel sample containing tritium was being heated in a kiln in room 107 at the NTLF. The sample was thought to contain only 5-10 mCi but apparently contained 35 Ci. The LBNL DOE notification report indicates that on July 28, the following Tuesday, it was determined that 24 Ci were released, but was later revised to 35 Ci. The room 107 fume hood is vented by a small stack on the roof of the NTLF building This fume hood has a silica gel monitor but does not detect HT. Would the authors care to comment on the appropriateness of this arrangement? For example, why is the 107 fume hood not vented through the main trunk stack? If it were, both HT and HTO releases would be monitored. The current arrangement allows for an unknown quantity of HT to be vented through the 107 stack.
According to the realtime Overhoff data, the amount released via the main stack from 11:30 A.M. to 2:00 P.M. was 0.2 Ci. This was presumably the tritium in room 107 that reached the stack via the building's ventilation system. The realtime HTO data for the main stack are shown here as Figure 1. One interesting aspect of the data is that emissions of HTO dropped to zero from 2:14 P.M. to 3:17 P.M. Also puzzling is the strange reversal in the Overhoff real-time data file of the HT v. HTO percentages. Anomalies such as this raise the question as to whether data values are being reported correctly.
On a more general point, perhaps the authors would care to comment on the fact that LBNL-EHS did not report the unplanned release to City of Berkeley and the public until a week after it happened? Fortunately, at the time of the release the wind direction was away from the Lawrence Hall of Science, but this could hardly have been known when the accident occurred. Neither was it known then how much tritium had been released. Would it not be prudent when accidental releases of this kind occur to warn the authorities at the Lawrence Hall of Science? In the 1980's a flask was broken and over 200 Ci of tritium were released. Other significant unplanned releases occurred in 1993 and 1994.
Figure 1. Overhoff realtime data showing the HTO fraction of the unplanned release of July 24, 1998.
The September 10,1997 TIWG field trip to the NTLF
On September 10, 1997 members of the Tritium Issues Work Group visited the NTLF. Dr. Philip Williams gave an oral presentation and led the group around the facility. In addition Dr. Henry Tran showed some members of the group the roof monitor and the silica gel and Overhoff systems in room 131. The tour was videoed by Mr L A Wood of Berkeley. I would like to suggest that copies of the video be made available to the authors of the report.
Two "incidents" during the field trip deserve special attention. The first is Dr. Williams' reluctance to let Mr Wood's video the oxidizer on the side of the building. My interpretation here was that the problem for Dr. Williams was that the oxidizer was not working. The second is Dr. Tran's nervousness while showing a small group the Overhoff monitor in room 131. While he was doing this it became evident from the computer display that the HT output was 29 percent of the total. Paradoxically, this happened at more or less the same time Dr. Tran was explaining how the oxidizer converted 90 percent of the HT coming out of the facility to HTO. When Dr. Iran noticed the high HT numbers he tried to screen the computer display from the people in the room. He also confused the situation by introducing a radioactive source into the ion chamber. The Overhoff display as recorded in the video before he introduced the source is presented in the following table. The data are in Bequerels/m3.
I assume that the cumulative total represents the integrated sum of the 100 second sampling intervals from midnight until the time the video was taken. If the video was taken at midday this would mean that there were 432 sampling intervals.
I also assume that to convert the cumulative numbers to a total flux from the stack (i.e., total Bequerels for the time period in question - midnight to midday) one would have to multiply by 100, i.e., the time of the sampling interval average in seconds, and by the average stack flow in m3/second. If the average stack flow rate was 1 m3/sec, this would mean the total emissions for the period midnight to midday was 1.090 E+11 Bequerels. If we extrapolate this rate for a 24 hour period the total would be 2.180 E+11 Bequerels, or 5.9 Ci. This is a surprisingly high total in view of the fact that no tritiations were being performed on the morning of the field trip.
To oxidize or not to oxidize...?
In the spring of 2000 a list of 10 CMTW priorities was presented to Bernd Franke. One of several items not covered in the Preliminary Report is a request for "An evaluation of the appropriateness of NTLF policy of converting HT to HTO on the grounds that it is easier to capture on silica gel, even though HTO is 25,000 times more rad toxic than HT." Just why this evaluation was not made in the preliminary report is unclear. The authors should include it in their second report.
Perhaps they can also find out why the oxidation policy was adopted in the first place. Dr. Williams was asked this question during the September 1997 NTLF Field Trip and responded that NTLF was told to oxidize by someone from the Lawrence Livermore National Laboratory. Another rumor is that EPA proposed that HT should be oxidized. Does any another tritium producing DOE facility follow this procedure? Also, in view of HT's 6-7 year half life in the atmosphere, does the policy of oxidation seem appropriate in anything other than a public relations context?
A.3 Is tritium in air measured at the right locations?
The authors state that their approach here was to "compare potentially affected locations with locations actually sampled." Unfortunately they do not appear to be aware of which areas have been affected so their findings are not very helpful.
Particularly disappointing is their unquestioning acceptance of LBNL's dose estimate as shown in Table 3. If the calculated dose to the MEl (maximally exposed individual, which by the way is a worker at the Lawrence Hall of Science), is only 0.28 mrem there is absolutely no need for an IFEU investigation in the first place. Why is it that the authors ignore the evidence that the actual dose is much higher than 0.28 mrem? The caption for table 3 is also misleading in that it implies that the purpose of the ambient air samplers at the various DOE facilities is simply to monitor tritium. That is not the case; all radionuclides are monitored.
The authors' recommendation that air monitoring at LBNL be expanded to cover 16 wind direction sectors further indicates that they are not yet aware that the primary concern is the Lawrence Hall of Science. There is abundant evidence that the environmental concentrations of tritium at LBNL are highest towards the northwest, that is towards the Lawrence Hall of Science. This was clearly pointed out in a 1996 report by Dr. Leticia Menchaca summarizing data on tritium in plant samples from several areas of the LBNL site. Dr. Menchaca explained the pattern as being a reflection of the prevailing south-easterly wind direction in winter. Winter is the rainy season and therefore the time when washout of the tritium plume occurs.
Another important point that the authors ignore is the topography of the site and its influence on tritium concentrations at the surface. The proposed location of a sampling station at the U.C. Botanical Garden illustrates this point. The contours shown in Figure 1 in the report indicate that the Botanical Garden is between 300 and 400 feet below the NTLF. This means that the effective stack height for this location is approximately 100 meters. Furthermore, if the scale on the map is correct the distance between the Botanical Garden and the NTLF stack is ca., 2,500 feet, or about 750 meters. All of this will mean that tritium concentrations at the U.C. Botanical Garden will be undetectably low. The construction of a sampling station at this location would be a waste of taxpayer's money.
In their discussion of current air sampling locations, the authors make no comments on the appropriateness of sampler placement. Apparently, they never visited the new sampler (ENV-75 EG) or the sampler at the Lawrence Hall of Science (LHS) because if they had they would surely have noticed how inappropriately they are located. The ENV-75 EG sampler is 21 meters from the stack and has its air intake at what appears to be 1-2 meters below the top of the stack. This plus the buoyancy factor of 5 meters per second makes it certain that the plume will not be meaningfully monitored. The sampler should be moved up the slope to an elevation where its air intake will be at a level 3 to 4 meters above the top of the stack. During the installation of the sampler a path was initially laid out up to a higher location but then inexplicably abandoned and the sampler installed at its present lower elevation.
The Lawrence Hall of Science monitor is also inappropriately located at a height ca., 3.5 meters above the ground surface. This is inappropriate for at least two reasons: (1) very few people 3.5 meters tall ever visit the Lawrence Hall of Science; and, more importantly, (2) the plume would have to have an average vertical thickness of 7 meters to be meaningfully sampled by this monitor. In reality the plume's thickness is probably rarely more than 2 meters at this location. The stack has a diameter of only 3 feet (0.9 meters) and the air in the Eucalyptus grove is typically stable. Plume behavior in a stable atmosphere involves minimal vertical dispersion.
The inappropriate height above the ground surface of the Lawrence Hall of Science Monitor is almost certainly the main reason for the low tritium concentrations in the much publicized LHS monthly silica gel samples. The behavior of the plume should be tested with smoke experiments.
I endorse the authors' recommendation that more air samplers be installed, but I disagree strongly that they be spread around the 16 wind directions. An array of realtime monitors should be set up between the stack and the Lawrence Hall of Science. A portable monitor should also be used to sample along the same array. Portable tritium air monitors are available now with a minimal resolution of 200,000 pCi per cubic meter so a release event such as the one shown in the report's Figure 7 should be detectable on the path between the stack and the Lawrence Hall of Science. The realtime data could be displayed in the Lawrence Hall of Science so that workers and visitors can decide whether they want to expose themselves and their children to the air concentrations indicated.
A.4 Is the sampling and analysis of tritium in air at a given location sufficiently accurate?
The data presented by the authors in this section clearly show that the laboratory analyses of the silica gel ambient air split samples are reliable The results reported by the two laboratories are similar although they show a systematic offset, the EPA sample counts always being higher than the LBNL counts. This joint LBNL-EHS/EPA sampling program was started in 1997 in response to public concern about the reliability of LBNL-EHS's air sampling data. I would only add that it is a great pity that the program did not also include the stack silica gel sampler. Insofar as the NESHAPS reporting is based on the stack silica gel data this is a puzzling omission.
The question as to whether the ambient air silica gel samples are effectively collecting all the water vapor in the air appears to be convincingly answered by the comparison of observed and expected water collected at the ENV-69 site shown graphically in Figure 8. However, in reviewing the data behind this graphic I have noticed that the Narel split samples have consistently more water in them than the LBNL samples. I show this here as Figure 2. The question therefore arises as to how EPA was able to extract more water from their silica gel samples than should have been there? Perhaps they have developed a new technology here that will be of great value in the more arid parts of the world? Or, perhaps the LBNL weather data are in some mysterious way unreliable?
Figure 2. Absolute humidity at the LBNL weather tower and the water content of LBNL and Narel silica gel samples from January 1999 to January 2000.
I would suggest that the authors obtain the LBNL weather data for 1999 and see if they can replicate the truly remarkable correlation shown in Figure 8. Weather data for the other years the split sampling program has been in operation should also be obtained so comparisons can be made for them as well.
B.1 Is LBNL's Draft Tritium Sampling and Analysis Plan sufficient to determine the extent and nature of legacy contamination at NTLF?
I have great difficulty understanding the logic of the Tritium Sampling and Analysis Plan. Superfund sites, as I understand them, are sites that have been contaminated in the past and now need to be cleaned up. In other words the source of contamination is no longer active. In this case, EPA is on record as stating that the NTLF should continue to operate in its present location. Why bother with a clean up if the facility is to remain active?
One difficulty for EPA is that its 50 pCi per cubic meter ambient air level cancer risk screening level is so low that it is often exceeded in the LHS silica gel sampler. For example, for the first 12 months of the split sampling program (September 1997 through August 1998) the monthly average was over 50 pCi per cubic meter. Not mentioned in the report is the EPA 600 pCi/liter ground water screening level for tritium. This also has been frequently exceeded in the area near the NTLF stack.
My interpretation is that the Tritium Sampling and Analysis Plan is designed to further confuse the public with an elaborate and expensive sampling program that will serve no useful purpose. The proposed sampling program in many ways replicates samples that have already been taken. It also extends into areas that are clearly not contaminated. The various reports prepared by Dr. Menchaca, the Monheit thesis, plus the voluminous data gathered by Dr. Javandel and the EHS Monitoring Program already document the extent of tritium contamination in the area around the stack.
On page 19 the authors indicate that integrated concentrations of tritium in ambient air at Building 3 (Calvin) are comparable to those measures at NTLF and could indicate a similar level of environmental contamination. I discuss this implausible conclusion in section C.2 which deals with past releases of tritium.
B.2 Which other factors need to be addressed in EPA's evaluation of the Superfund status for the NTLF site?
One concern here is whether or not NTLF operations will be typical during the sampling period. Earlier in the report (page 6) the authors reached the following ambivalent conclusion:
Product shipments may be a better indicator of the activities at NTLF that are associated with airborne releases of tritium. While shipments up until 1991 were typically between 100 and 300 Ci per year, reported product shipments dropped to a level around 10 Ci around the mid-1990s. According to LBNL, the values for 1998 and 1999 were 20 Ci and 10 Ci, respectively. This suggests less tritiation activity during the last few years. However, neither tritium shipments nor the tritium inventory appear to be a good indicator for the likelihood of potential releases (emphasis added).
I have emphasized the last sentence because it was used by Dr. Philip Williams, the NTLF Manager, in a handout to the Environmental Sampling Task Force on August 10, 2000. He used it to support his argument that the reduction in shipments of tritiated products (and therefore presumably the number of tritiations?) has nothing to do with the reduction in stack emissions.
Insofar as the Preliminary Report clearly documents that tritium emissions increase dramatically during a tritation, are we to believe that the number of product shipments is not an indirect index of tritium emissions? The data clearly indicate that it is. The implication is clear - a facility under close scrutiny by regulatory agencies and the public is hardly likely to be operating in a "normal" way.
One way in which the NTLF has reduced emissions of HTO is by reducing the use of the oxidizing system. The official position is that 90 percent of the tritium released from the facility is converted to HTO to facilitate capture and disposal as radioactive waste. However, according to the NTLF Release Summary files for 1997, 1998, and 1999 HTO has on average made up only 76 percent of the total release. The Overhoff Daily Cumulative files for 1998 and 1999 indicate even lower percentages at 69 and 60 percent respectively. The difficulties reported in measuring HT with the stack silica gel system further support the idea that HTO emissions are much less than 90 percent of the total.
C.1 Which exposures to neutron and gamma radiation resulted from LBNL operations?
This is a completely separate issue that should be dealt with in a separate investigation. Here it simply detracts from the tritium question.
C.2 Which exposures resulted from past releases of tritium?
This is the weakest part of the whole report insofar as it overlooks the Menchaca organically bound tritium data and the Monheit rainfall data, both of which are very relevant to the question of past exposures. Both of these issues were included in the Committee to Minimize Toxic Waste's List of Concerns that was given to Bernd Franke during his visit to Berkeley by Co-Chair Pamela Sihvola earlier this year (personal communication from Ms. Pamela Sihvola, as Co-Chair of the Committee to Minimize Toxic Waste).
The Menchaca/Monheit data
I have discussed the relevance of the Menchaca/Monheit data in my report to NIH and will not therefore review it again in detail here. I would simply emphasize that both lines of evidence show that tritium levels must have been significantly higher in the mid 1990s than reported by LBNL-EHS. The Monheit rainfall samples were collected in November and December of 1994 as part of a masters thesis research project. Several of the samples collected near the Lawrence Hall of Science had tritium levels of over 200,000 pCi's per liter. The average value for the 30 rainfall samples collected within 250 meter radius of the stack was 50,000 pCi per liter. These high numbers are in the same order of magnitude as the concentrations reported in the 1980's for the rain gauge on the roof of building 75. Ms. Monheit also collected rainfall samples at a site in Albany 4 km north west of the stack. These samples showed that a tritium plume was present in that area for 4 days (November 26 - November 29). In the fall of 1996 Dr. Menchaca collected vegetation samples from around the NTLF stack and had them analyzed for their organically bound tritium content. The levels she reported from the Lawrence Hall of Science area are several times higher than the levels reported by the U.S.G.S for leaf samples collected at DOE's Hanford and Savannah River facilities.
The high rainfall and organically bound tritium levels are not important in the sense that the rainwater or Eucalyptus leaves represented a direct threat to anyone. The significance of the data is that these numbers could be used to calculate the actual tritium concentration in the air at the time the rain was falling and in the leafwater when the leaf was being formed. The latter presumably came out of the ground water because in 1996 the NTLF was shut down for a large part of the year. I urge the authors of the report to assess the implications of the Menchaca/Monheit data. They also should be aware of the fact that both Dr. Menchaca and Ms. Monheit lost their positions at LBNL-EHS. Dr. Menchaca was even told that her organically bound tritium data would be shredded.
The Calvin diversion
On page 19 the authors state that integrated concentrations of tritium in ambient air at Building 3 (Calvin) are comparable to those measured at NTLF and could indicate a similar level of environmental contamination. This is an inaccurate and misleading statement. It is also unfortunate in that it diverts attention away from the real problem: the situation at the Lawrence Hall of Science.
First we should note the record high ambient air level of 3,000 pCi/m3 was not recorded at Calvin in 1985 but at Building 69. The authors misread the data presented in the Site Environment Report for 1985. Furthermore. Building 69 is about the same distance from the NTLF stack as is the Lawrence Hall of Science. In his telephone presentation to the Environmental Sampling Project Task Force Meeting of August 10, 2000 Bernd Franke included an overhead (#13) that shows the time integrated (1972-1999) tritium in air at the Lawrence Hall of Science to be about the same as that on the Building 3 (Calvin) roof, i.e., -10,000 pCi/yr/m3. It must be emphasized that the Calvin sampler is located only 6 feet from the roof stack and in this sense can not be compared with the Lawrence Hall of Science sampler 110 meters from the NTLF stack.
The Calvin sampler was also located on the roof of the building so the effective stack height as far as people on the ground is concerned is probably > 20 meters. I have no data to indicate what the tritium emissions were at Calvin in the late 1970's but whatever was released into the environment has decayed through two half lives since then. The chances of finding any tritium above background levels in the area around the Calvin Building will be infinitesimally small.
Unfortunately, local community members have already been alarmed by the suggestion that there could be tritium contamination in the area around the Calvin building. In large part this concern reflects the fact that a U.C. Day Care Center is located close to Calvin. The U.C. Department of Environmental Health and Safety is now developing a sampling program to ease community concerns.
Concerns not covered in the Preliminary Report
For some strange reason 5 of the 10 CMTW concerns presented to Bernd Franke are not covered in the Preliminary Report even though they involve subject areas that were supposed to be included. Perhaps the authors could explain why these concerns and suggestions were not considered, and also indicate to what extent they will be included in the final report?
I indicated at the beginning of this review, the primary issue is to what extent the tritium emissions from the NTLF stack pose a threat to workers and visitors to the Lawrence Hall of Science. Unfortunately, this issue is not given the attention it deserves in the Franke-Greenhouse preliminary report.
Particularly puzzling is the lack of any discussion of the Menchaca/Monheit tritium data. These data are, after all, the basic reason for community concern in this matter. Dr. Menchaca's organically-bound tritium leaf data from near the Lawrence Hall of Science are 3 to 7 times higher than those reported by the LBNL-EHS. Similarly, Ms. Monheit's rainfall data indicate that stack emissions in the fall of 1994 were very much higher than reported by LBNL-EHS. Also, her Albany rainfall data demonstrate that the NTLF tritium plume was still detectable 4 km away from the stack! These data must be evaluated and their implications explained to the public.
Another disappointing aspect of the report is the authors' willingness to accept the LBNL position on critical issues. LBNL's repeated claim that the exposure level to the maximally exposed individual at the Lawrence Hall of Science is well below the regulatory limit is based on two sets of data: the stack monitor data and the CAP88 dispersion model results. There is abundant evidence that both data sets are unreliable. The authors conclude that there are some uncertainties in the NTLF stack data for 1998 but go on to state that there is no evidence to suggest that the 10 mrem/yr limit has been exceeded. This contradictory position needs to be clarified. The incredible revelations that the Overhoff monitor is unreliable and the silica gel monitor has only recently been working correctly clearly indicate that the source term has not been accurately reported.
One positive outcome of the IFEU contract is that LBNL has finally been persuaded to release some of the NTLF stack monitor data. Having reviewed some of these data I can only say that they confirm my suspicions that the source term has been under-reported. Another piece of evidence here is the TIWG NTLF field trip video. The video coverage of the Overhoff Stack Monitor indicates that actual emissions may have exceed reported emissions by an order of magnitude. The authors of the report should be asked to view the video and to comment on this possibility.
According to the Scope of Services (Appendix A), the Preliminary Technical Report was not supposed to cover the appropriateness of the dispersion models used to estimate dose for NESHAPS reporting. However, the authors make several statements to the effect that CAP88 is not an appropriate model because of the short term nature of the NTLF tritium releases and the complex terrain. This is misleading. It is true that CAP88 is not designed to model large "unplanned releases" such as occurred three times in the 1990's at the NTLF. On the other hand, CAP88 does work well for routine emissions over a 12 month period, which is the NESHAPS reporting interval. At least it will work well if the parameters in the model are set correctly. The problem here is, as I indicated in my report to NIH, that LBNL-EHS is not correctly characterizing stack height, wind speed, or wind direction.
When these parameters are set correctly, and the source term is set at 100 Ci the EPA compliance level of 1,500 pCi/m3/yr is exceeded at the Lawrence Hall of Science. The complex terrain problem is easily resolved by setting the stack height to zero, the appropriate setting for the Lawrence Hall of Science. The appropriate settings for other locations around the NTLF are irrelevant because the maximally exposed individual is located at the Lawrence Hall of Science.
Finally, I would urge the authors to reconsider their statement that there is no evidence indicating a person near the fence could receive a radiation dose greater than 10 mrem/yr. Several lines of evidence point to the opposite conclusion: the Menchaca organically bound tritium data, the Monheit rainfall data, the subtraction of the Overhoff "background", the video of the Overhoff monitor during NTLF Field Trip, and the CAP88 model results when parameters are set correctly. I would also ask the authors to determine more precisely the stack flow at the monitoring locations, and to evaluate whether or not the ambient air monitors in the Eucalyptus grove and at the Lawrence Hall of Science have been improperly located.