Skip directly to search Skip directly to A to Z list Skip directly to site content


Community Exposures to the 1965 and 1970 Accidental Tritium Releases



The Lawrence Livermore National Laboratory (Livermore site, hereafter referred to as LLNL) is a multi-program research facility owned by the U.S. Department of Energy (DOE) and operated by the University of California. LLNL was placed on the Superfund National Priority List (NPL) in 1987. The Agency for Toxic Substances and Disease Registry (ATSDR) is required to conduct a public health assessment of all facilities proposed for listing on the NPL. During the LLNL public health assessment process, potential off-site exposure to tritium released by LLNL has been identified as a specific community concern (CDHS 2003). In response to this concern, ATSDR convened an expert panel to assess tritium monitoring and dosimetry issues at the Lawrence Livermore National Laboratory (LLNL) and Savannah River Site (SRS) facilities. Although the expert panel determined that approximately 80% of the total radiologic releases from the LLNL facility occurred during two accidents in 1965 and 1970, they did not explicitly evaluate potential short-term tritium doses from those accidental releases. This public health assessment will specifically evaluate potential short-term tritium doses from the accidental tritium releases to determine whether these releases presented a public health hazard to members of the Livermore community.

There are insufficient historic environmental sample data available to adequately evaluate the total tritium doses from these releases. Consequently, this evaluation will use modeled data combined with available measured data to estimate past exposure concentrations and doses. This evaluation focuses on exposure doses to maximally exposed individuals. Available meteorological data indicate that for the 1965 release winds were blowing to the north-northeast at about 3 meters per second (m/s). The maximally exposed residence is more than 1 mile from the tritium facility for the 1965 release (January 20, 1965) with an estimated maximum of 18 people living in the plume area to a distance of 2 miles from the tritium facility. During the 1970 release, winds were blowing to the north-northeast at about 1.5 m/s and the closest residence was also more than 1 mile from the tritium facility. An estimated maximum of 55 people were living in the area of the 1970 plume to a distance of 2 miles from the tritium facility.

Tritium is a radioactive isotope of hydrogen. As hydrogen, tritium might be present in the environment as any chemical form or compound of hydrogen, including hydrogen gas (HT), tritiated water (HTO), or as various organic compounds (known generically as organically bound tritium; OBT). The specific absorbed dose from tritium exposure depends on the chemical form of the tritium that is ingested or inhaled. The radiologic dose is determined by how many tritium decays occur in the body after ingestion or inhalation. As hydrogen gas, very little tritium is absorbed and retained in the body following exposure. Consequently, very few tritium decays occur in the body from HT inhalation. Conversely, most of the tritium taken in as water or HTO (including a lesser OBT contribution) is absorbed and retained in the body with an effective half-life that varies from 1 to about 40 days.

Both of the accidental tritium releases from LLNL occurred in the HT form. Consequently, there is very little radiologic dose from direct inhalation of the HT plumes. However, HT is converted by soil microbes into the HTO form of tritium. Subsequent exposure to HTO creates the potential for much more significant radiologic doses. This health assessment is based on potential exposures to each of the significant tritium forms as it moves through the environment. As there are insufficient environmental measurements of each of the tritium forms in air following the accidental releases, this health assessment relies on air dispersion and exposure models to evaluate potential historic short-term tritium exposures. Specifically, this assessment uses the RASCAL air dispersion model to determine concentrations of airborne HT in areas of potential exposure. The Industrial Source Complex air model is used to estimate concentrations of HTO in areas of potential exposure due to emission of HTO from the soil. To accommodate the uncertainty inherent in each of these modeling steps, a Monte Carlo simulation is conducted to determine the most likely tritium doses from each type of exposure.

The estimated total tritium doses include direct inhalation of the HT plumes, inhalation of HTO following emission from soil, direct absorption of HTO through skin, ingestion of foods containing HTO and OBT, and also sum potential chronic exposures from ongoing (past) LLNL tritium releases. The estimated maximum doses (to a child; 95th percentile) are less than 149 millirem/year (mrem/year) for both the 1965 and 1970 releases. The more likely average doses are about 42 mrem/year. On the basis of current peer-reviewed scientific literature, the one-time exposure to tritium resulting in a committed effective adult dose of 42 mrem (0.42 mSv) or a child dose of 149 mrem (1.49 mSv) from the LLNL accidental HT releases is not expected to be a public health hazard.

While some public exposure to tritium probably did occur as a result of the accidental releases of tritium gas (HT), estimated maximum exposures are below levels of public health concern and no adverse health effects would be expected. This conclusion is based on tritium doses developed from analytical models and is supported by human biological samples that showed no detectable tritium from either LLNL workers or affected community members. The above doses represent the 95th percentile doses on the basis of health protective exposure and dosimetry assumptions. It is unlikely that actual doses approached these conservatively estimated values.

All of the adverse health effects from exposures to tritium (or low-energy external gamma radiation or x-rays) that we found in the medical literature occurred at levels higher than the exposure levels we estimated for people living near the LLNL facility at the time of the accidental releases. Therefore, we conclude that inhalation and ingestion of tritium from the acute releases that occurred in 1965 and 1970, plus the annual contribution from chronic or long-term exposures, were never a public health hazard. Because these historic accidental releases are below levels of public health concern, no specific recommendations are warranted.


The Lawrence Livermore National Laboratory (Livermore site; hereafter referred to as LLNL, is a multi-program research facility owned by the U.S. Department of Energy (DOE) and operated by the University of California. The LLNL is a science, technology, and engineering facility with a special focus on nuclear weapons research and development. Other areas of research include arms control and treaty verification control technology, energy, the environment, biomedicine, the economy, and education (DOE 1992).

LLNL was placed on the Superfund National Priority List (NPL) in 1987 on the basis of volatile organic compounds (VOCs; trichloroethylene, tetrachloroethylene, chloroform, 1,1-dichloroethylene, and others) in monitor wells and nearby drinking water wells (LLNL 1990). The Agency for Toxic Substances and Disease Registry (ATSDR) is required to conduct a public health assessment of all facilities proposed for the NPL.

During the course of the LLNL public health assessment process, potential off-site exposure to tritium released by LLNL was identified as a specific community concern (CDHS 2003). In response to this concern, ATSDR convened an expert panel to assess tritium monitoring and dosimetry issues at the Lawrence Livermore National Laboratory (LLNL) and Savannah River Site (SRS) facilities. ATSDR's summary of the panel's conclusions, which was released as a public health consultation, found that potential community exposures to chronic or long-term tritium releases from LLNL (and SRS) are not likely to produce adverse health effects in the surrounding communities and consequently are below levels of public health concern (ATSDR 2002A). The expert panel report included evaluations of the tritium monitoring programs at each site, an assessment of current and historic tritium releases at each site, an evaluation of the uncertainties related to tritium dosimetry procedures, and an assessment of potential chronic tritium doses to the off-site community for each facility.

Although the specific results and discussion of those evaluations will not be repeated in this exposure assessment, the conclusions and recommendations regarding tritium dosimetry parameters will be included by reference. The expert panel also determined that approximately 80% of the total radiologic releases from the LLNL facility occurred during two accidents in 1965 and 1970 (ATSDR 2002A). However, the expert panel dose assessment did not evaluate potential short-term exposures from those accidental releases. Consequently, this public health assessment will specifically evaluate the potential for adverse health effects in the Livermore community from the accidental tritium releases.

The two releases are of similar magnitude (1965--350 kCi; 1970--300 kCi)(1). However, because of differing meteorological conditions, the 1970 release resulted in larger potential doses to the offsite community. Consequently, this public health assessment will focus on the evaluation of tritium dispersion and potential exposures from the 1970 release. Further, this evaluation will emphasize the evaluation of those people who were potentially exposed to the highest doses. If doses to the maximally exposed residents are below levels of health concern, it will be assumed that all other, lower, doses are also below levels of health concern.

Both accidental releases are included in a list of site accidental releases (DOE 1992) and LLNL has recently declassified the 1965 accident report (Peterson et al. 2002). Available documentation confirming the magnitude and chemical form of the releases is limited to the referenced list of accidental releases, newspaper reports, the Myers et al. (1971) study, and the Peterson et al., report (2002). As these references all report similar release magnitudes and conditions, the referenced information will be used in this report. ATSDR will re-evaluate the following results and conclusions if additional information becomes available.

Myers and others (1971) conducted a study of the 1970 release that included real time monitoring and environmental sampling of the release plume and provides general meteorological data. That study also included human dose estimates based on sampling of tritium in air, vegetation, cows milk, and human urine. The results from the Myers et al. (1971) study will be compared with the modeled dispersion and estimated doses from this evaluation, which implicitly includes the dose contribution from OBT.

ATSDR also released a public health consultation that updated and reevaluated the Myers et al. study (ATSDR 2001). That health consultation used the same assumption (that 1% of the released HT will occur or be transformed to HTO) as the Myers (et al. 1971) study and used updated dosimetry factors in an attempt to verify the results and conclusions of the Myers analysis. As documented in the following section on "Community Health Concerns…", the assumptions and approach of that consultation do not adequately evaluate the dispersion and transformation of the tritium released into the environment. Consequently, in response to comments received on the 2001 health consultation, this public health assessment presents a more comprehensive approach to estimating potential tritium exposures and doses from the 1965 and 1970 accidental releases. As the approach and assumptions underlying this PHA and the earlier 2001 health consultation are very different, we will not attempt to compare the results of the different approaches. All assumptions underlying this PHA are appropriately referenced in the following sections and appendices.

Site Description and History

The LLNL site is within the southeast end of the Livermore Valley in southern Alameda County, California, and approximately 40 miles east of San Francisco (Figure 1). The LLNL is about 3 miles east of the central business district of the City of Livermore but directly abutted by residential properties to the west, commercial and industrial properties to the north, agricultural and residential land to the east, and the Sandia National Laboratory to the south. LLNL also operates the LLNL Site 300 near Tracy, California (about 12 miles east of the main site). Operations and potential contaminant releases of the Site 300 will be addressed in a separate public health assessment.

The LLNL main site, including a buffer zone acquired in 1989, covers an area of approximately 821 acres in the southeastern portion of the Livermore Valley. In 1942, the U.S. Department of the Navy acquired 681 acres of agricultural and ranch land to establish the Livermore Naval Air Station. Although the original use of the Naval Air Station was for flight training, by October 1944, aircraft assembly, repair, and overhaul was conducted at the Livermore NAS. From 1945 until the Livermore NAS was deactivated in 1946, extensive aircraft repair and assembly occurred at the site. In 1950, the Atomic Energy Commission occupied the site, with formal transfer of the property in 1951. The AEC, its successor agencies, and ancillary entities have occupied the site for defense-related research. In 1952, the site was established as a separate part of the University of California Radiation Laboratory. In 1971, the Livermore site became the Lawrence Livermore Laboratory, and in 1979 was renamed by Congress as the Lawrence Livermore National Laboratory. The University of California operates LLNL under contract with the U.S. Department of Energy.

In 1992, DOE published the "Final Environmental Impact Statement and Environmental Impact Report for Continued Operation of the Lawrence Livermore National Laboratory and Sandia National Laboratories, Livermore." This document includes a detailed statement of LLNL operations and facilities. The information from that report outlining LLNL operations and facilities will not be reproduced here, but will be referenced as appropriate to define environmental releases and potential community exposures to chemical and radiological materials.

Exposure Pathways

A release of a chemical or radioactive material into the environment does not always result in human exposure. For an exposure to occur, a completed exposure pathway must exist. A completed exposure pathway exists when all of the following five elements are present: (1) a source of contamination, (2) an environmental medium through which the contaminant might be transported to (3) a point or area of human exposure, (4) a route or process of human uptake (ingestion, inhalation, etc.), and (5) an exposed population. A potentially completed pathway exists when one or more of the above elements are missing or unknown, but available information indicates that exposure is likely. An incomplete exposure pathway exists when one or more of the five elements are missing and available data indicate that human exposure is unlikely.

A conceptual model of the exposure pathway from the accidental tritium releases is presented in Figure 2. From the pathway perspective, exposures to the 1965 and 1970 LLNL tritium releases are considered to be complete because known residences were located within the potentially affected areas (as described below and shown in Figure 1). The remainder of this assessment will use modeled and measured data to estimate whether the tritium exposures occurred at doses likely to cause sickness or other adverse health effects.

Map of the LLNL facility showing the area of potential exposures from the 1965 and 1970 tritium releases
Figure 1. Map of the LLNL facility showing the area of potential exposures from the 1965 and 1970 tritium releases. The inset shows the estimated number of people living within 2 miles of the LLNL tritium facility for each release. Population estimates are derived from counting buildings from historic aerial photographs, assuming all buildings are home-sites, and that an average number of residents live in each home. Note that the city of Livermore boundary is based on 1995 Census Bureau data (TIGER/Line files).

Estimated Population of Affected Areas

LLNL is just east of the city of Livermore in eastern Alameda County. Although the site is currently adjacent to the city limits, in the 1960s and 1970s the areas to the north and east of the site (in the direction of the winds during both releases) were sparsely populated rangeland. The populations that were potentially exposed to the highest tritium concentrations lived within wedge-shaped areas to the northeast of the on-site tritium facility (see Figure 1). These areas of potential exposure are based on the footprints of the dispersed tritium plumes as described below. Note that the footprints of the two plumes do not overlap, such that residents living in one plume area would not likely have any exposure to tritium from the other plume.

An estimate of the number of people exposed to the tritium plumes has been determined by counting buildings within the plume area on a photo-revised 1968 U.S. Geological Survey topographic map (USGS 1968; and related aerial photographs dated 1966 and 1977) and multiplying the number of buildings times the average number of persons per household. According to the 1970 Census of Population, there were 3.24 persons per household in census tract 4511, which includes the entire modeled area.

For the 1965 release, 6 buildings are visible in aerial photographs. This translates to about 18 people within the maximally exposed area (assuming that all visible buildings are single-family homes; Figure 1). For the 1970 release, there are 16 visible buildings within the plume area (from 1 to 2 miles) in the 1968 photo-revised topographic map. The closest off-site buildings to the tritium facility (within the plume areas) are more than 1 mile from the tritium facility. Assuming that all visible buildings are single-family homes provides a conservative(2) estimate of approximately 52 people residing in the area of maximum exposure in 1970. Figure 1 also includes a population breakdown of the maximally exposed residents for age, gender, and other demographic categories based on average proportions for census tract 4511. There were too few residents living in the area of the 1965 plume for an accurate breakdown.

The estimated populations within the areas of maximum exposure are probably overestimates of the potentially affected population. An area-proportional estimate of population indicates that the potentially affected areas had total populations of about six people for the 1965 plume and six people for the 1970 plume. The area-proportional method assumes that the population of the census tract is evenly distributed throughout the tract and uses the ratio of the potentially affected area relative to the total census tract area as a multiplier to correct the total census tract population (affected area/census tract area x census tract population = affected area population). The lower total population estimate produced by the area-proportion method indicates that the home-site counting method produces a very conservative estimate of the total population within the potentially affected plume area.

Conceptual exposure pathway and exposure factors related to short-term (12 day) human inhalation and ingestion of tritium from the 1965 and 1970 LLNL HT releases
Figure 2. Conceptual exposure pathway and exposure factors related to short-term (12 day) human inhalation and ingestion of tritium from the 1965 and 1970 LLNL HT releases. Ranges of uncertainty and assumptions used in this evaluation for each of the exposure variables are discussed in following sections and in Appendices 2-4. Total tritium intake and dose also includes a term to account for direct dermal absorption of HTO through exposed skin.


As previously stated, this public health assessment has been developed in response to specific community concerns related to potential health effects from the accidental tritium releases. An initial evaluation of the potential exposures from these releases was published by ATSDR as a health consultation (ATSDR 2001). The results of that health consultation were presented to the Livermore community at a public meeting on November 8, 2001. ATSDR received a number of comments relating to the process of exposure assessment used in that consultation as well as specific concerns related to the number of people potentially exposed during the accidental releases.

Specifically, the comments regarding the earlier health consultation are as follows:

  • Community members have requested funds to hire an independent technical consultant to review the health consultation (exposure assessment).
  • The health consultation did not adequately address the environmental fate and cumulative effects of the tritium released during the accidents.
  • The health consultation did not address any potential increased vulnerability of women and children to tritium exposures.
  • The health consultation was too brief and/or too technical.
  • The distance to the site boundary (and to the maximum off-site exposures) was too short.
  • Some of the wording regarding tritium regulatory standards was misleading.
  • Potential tritium ingestion doses should be evaluated.
  • Errors in labeling of some tables and attachments.
  • The precautionary principle should be incorporated in all health consultations.

In response to those comments, ATSDR has revised and expanded the health consultation into this public health assessment that addresses all of the above comments(3). As presented in the following section on Methods of Exposure Assessment and Dose Estimation, this evaluation specifically addresses the environmental fate and dispersion of tritium released to air, deposited on soil, and taken up by vegetation. This assessment also includes specific sections on the number of people potentially exposed to the tritium releases, potential tritium ingestion doses, and the special vulnerabilities of women and children. Although the content of this assessment is inherently technical due to the modeling approach required to address historic exposures, we have attempted to increase the readability by including a number of illustrations and summary tables and by moving some of the technical detail to expanded appendices.

Also, because this assessment evaluates tritium doses differently than the earlier health consultation, it was re-released for additional public and independent peer review comment. The specific comments received on the public comment version of this PHA (ATSDR 2002b) and the ATSDR responses to those comments are listed in Appendix 1. In general, these comments warranted numerous revisions to the public comment version of this PHA. The most significant of these revisions is a re-calculation of the initial 1970 HT plume dispersion based on a class "E" atmospheric stability (rather than class "B"). Although the rationale for this change in atmospheric stability is discussed in the following section on "Dispersion of the HT Releases", the primary implications of this change is an increase in the centerline concentration of the HT plume, a decrease in the lateral spread of the dispersed plume, and a translation of the area of maximum concentration from an on-site to an off-site location.

Another comment addressed the potential for tritium uptake from direct HTO absorption through exposed skin (dermal absorption). Tritium dose calculations have been modified to account for this type of human uptake with appropriate changes to the dose estimates. Another important comment concerned the description and rate of HTO emission from soil. Although the HTO loss rate of 1% per hour that was assumed in the prior PHA adequately represents the loss rate over a 10-12 day period, the actual HTO emission from the soil is more variable and goes to zero during night-time. As further explained in following sections, the revised HTO air concentrations are based on a variable HTO loss rate and use the maximum 12 hour concentration averages, rather than a maximum one hour average concentration.

In addition to the comments received from the independent peer reviewers and the various stakeholders, the TriValley CAREs (Communities Against a Radioactive Environment), the San Francisco Bay Area Physicians for Social Responsibility (PSR), and the Western States Legal Foundation (WSLF) commissioned two independent reviews of this public health assessment. These reviews resulted in the development of two documents (Russ and Goble 2003; and Wing 2003). The findings and conclusions of those documents that represent specific comments on this public health assessment are included verbatim in Appendix 1 and addressed as appropriate.

Following release of the public comment version of this PHA (May 2002), LLNL published a review of the 1965 release that included previously unavailable meteorological and accident-specific data (Peterson et al. 2002). This newly available information has been incorporated into the dispersion and dose calculations within this document. Specific changes to the dispersion and dose estimations are referenced in the following sections.


The estimation of tritium doses from the accidental releases is complicated by the dynamics of tritium dispersion and transformation in the environment. Tritium, which is a radioactive isotope of hydrogen, occurs in several chemical and physical forms. A diagram of the path of tritium from the release stack to a point of off-site exposure is illustrated in Figure 3. HT, which is tritiated hydrogen gas, is the form released at the stack (Myers et al. 1971; DOE 1992). As hydrogen gas, there is very little radiological dose from HT inhalation (ICRP 1996). The HT gas, if inhaled, is biologically inert and is almost completely exhaled with very little uptake in the lungs. However, the HT gas is also deposited onto the ground where it is rapidly transformed by soil microbes into HTO (tritiated water), which becomes associated with soil moisture (Ogram et al. 1988; Brown et al. 1990). Following transformation into HTO, a significant portion of the tritiated soil water may be re-emitted into air as atmospheric moisture, or the soil moisture might be taken up by vegetation with subsequent transpiration to the atmosphere or incorporation into the plant as OBT.

Inhalation and dermal absorption of the HTO as air moisture and ingestion of HTO or OBT in vegetation present the potential for radiological exposures. In areas of significant groundwater recharge, tritiated soil moisture might also enter into the groundwater flow system and underlying aquifers. In dry, sedimentary environments, such as the Livermore Valley, transfer of soil moisture back to the atmosphere and to plants is much more rapid than aquifer recharge. Although some portion of the accidental tritium releases might enter the groundwater flow system, due to the slow rate of groundwater flow there is unlikely to be a measurable, short-term spike in tritium concentrations that could be attributed to the short-term releases. Assuming that all soil moisture is re-emitted to the air maximizes potential inhalation exposure and is a conservative, or health-protective, assumption. Consequently, this exposure analysis will not attempt to include a short-term groundwater ingestion component. Potential long-term groundwater ingestion of tritium has been evaluated (ATSDR 2002A).

Calculation of short-term tritium doses from inhalation and ingestion is based on the flow path illustrated in Figure 3. Important factors in the diagram are (1) the dispersion of the HT plume from the release source (the tritium facility stack); (2) the HT deposition velocity onto surface soils; (3) the rate of emission as HTO; (4) the rate of decline in the initial soil HTO concentration as it is emitted to air; (5) the area and duration of potential exposure (which determine the tritium concentrations and exposure time frame); (6) dispersion of the atmospheric HTO from the soil (both horizontally and vertically); (7) the rate of inhalation and body weight; and (8) the amounts and types of tritium-containing food ingested. All of these factors represent variables in the following dose calculations(4).

All of the doses calculated or described in this assessment are "effective dose equivalents" as recommended by the ICRP (1991).

Diagram of tritium dispersion and transformation based on airborne release of tritium gas (HT) and subsequent microbial transformation into tritiated atmospheric and soil moisture (HTO)
Figure 3. Diagram of tritium dispersion and transformation based on airborne release of tritium gas (HT) and subsequent microbial transformation into tritiated atmospheric and soil moisture (HTO). The quantity of tritium released, the processes of dispersion, and the environmental transfer rates are discussed in the text and appendices.

This evaluation focuses on the potential tritium exposure to a hypothetical maximally exposed individual. While we know that people lived within the area of potential maximum exposure, calculation of total tritium doses requires assumptions about specific exposure factors, such as activity and breathing rates, garden grown food consumption rates, body weights, and other information. As such information about actual Livermore residents is not available, we must make assumptions about those exposure factors based on population averages (for body weights, breathing rates, activity durations) or health protective values (for garden grown consumption fruit, vegetable, and milk ingestion rates).

The resulting doses are considered to be maximum doses because the individuals are assumed to reside at the area of maximum HT and HTO air concentrations (1.0 mile from the LLNL tritium facility) . The maximally exposed residences are also assumed to be on the centerlines of each plume and the residents present in those areas for 24 hours per day for 12 days, which creates maximum potential concentrations. The assumption underlying the evaluation of a maximally exposed person is if exposures to that hypothetical person are below levels of health concern, then all other lesser exposures are also below levels of health concern.

The calculation of overall tritium doses in this assessment is comprised of several components. The first component is the atmospheric model used to determine the HT plume dispersion and off-site HT concentrations. The second component is another atmospheric model used to estimate the emission of HTO from the soil surface and dispersion in the atmosphere. The third component is a Monte Carlo analysis used to estimate tritium uptake and dose from the dispersed HT and HTO concentrations. The final step in the analysis is an evaluation of the potential health effects to maximally exposed people from the estimated total tritium doses. Discussion of the specific models, underlying assumptions, and variables inherent to each component of the dose calculation are included in the following sections and appendices.

Dispersion of the HT Releases

Dispersion of the HT plumes was modeled using the RASCAL 3.0 codes developed by the Nuclear Regulatory Commission (NRC 2000). Input assumptions and case summary information for the RASCAL model are presented in Appendix 2 (Tables A-1 and A-2). Output from the RASCAL model includes both instantaneous and cumulative concentrations of HT deposited on the soil. Instantaneous and cumulative air concentrations are calculated from those surface concentrations and the RASCAL deposition velocity. The instantaneous HT air concentration is a measure of the HT present at breathing height (1 m) at any single time and location after the release. The cumulative air concentration is a measure of the instantaneous air concentration integrated over the entire duration of the release and subsequent dispersion.

Although RASCAL can compute deterministic tritium doses, those doses are based on HTO exposures rather than HT as released from LLNL. In order to assess the complex environmental fate of tritium and use the tritium dosimetry parameters recommended by the expert panel, tritium doses from the 1970 accidental HT release were calculated using a probability-based Monte Carlo procedure. The deterministic doses estimated by RASCAL are compared with the probabilistic doses estimated using the Monte Carlo model in the following section on "Total Tritium Doses and Uncertainties in the Modeling Process."

Tritium from both accidents was released as hydrogen gas (HT) from a 30 m stack(5).All of the available accident reports and descriptions state that the releases were comprised of HT. As this was the form supposedly contained within the vessels that leaked, there is no basis for assuming that any other form of tritium comprised a significant proportion of the releases.

The 1970 release occurred at 5:49 am on August 6, 1970. During the time of the accident, the wind was from the southwest (~200º) at a speed of 1-2 m/sec (2.2-4.5 mph) with stable atmospheric conditions (Pasquill-Turner Stability Class E or F). The stability classification is derived from tables in "Air-EIA: Air quality technical data: atmospheric stability" ( using known wind speeds and assuming no incoming solar radiation based on the time of day (Sunrise occurred at 6:15 AM on 08/06/1970; U.S. Naval Observatory, Predicted meteorological conditions for the 1965 release are described in a later section.

RASCAL model runs were conducted using both "E" and "F" atmospheric stability classifications. As the "E" classification resulted in slightly greater estimated HT air concentrations in areas of potential exposure, subsequent analyses will be based on results from the "E" analysis. Two sets of results are used from the RASCAL dispersion model(6). Inhalation of HT during the 30-minutes of direct exposure to the plume is based on maximum instantaneous tritium concentrations (Table A-3). The deposition of HT onto soil is based on the cumulative HT concentrations (Tables A-4 and A-5). Estimated HT concentrations for various downwind locations are presented in each table.

The HTO soil concentrations are obtained by multiplying the cumulative air HT concentrations times an empirical HT deposition velocity derived from several sources including, Sweet and Murphy (1981), Dunstall et al. (1985), Ogram et al. 1988, Spencer and Vereecken-Sheehan (1994), and Paillard et al. (1988). These studies identified a range of deposition velocities, which varied from 0.00001 to 0.0013 m/sec. Rather than use a single point estimate of the HT deposition velocity, this analysis used a probability distribution using the a low, commonly referenced rate (the lowest values were for frozen soil) as the 10th percentile value (0.0003 m/sec) and a similar higher value (0.0012 cm/sec) as the 90th percentile value of a normal distribution. Consequently, the resulting soil HTO concentrations are also represented by probability distributions.

After the dispersed HT plume has diffused onto the soil, microbes transform (oxidize) the HT into the HTO form of tritium. This evaluation assumes that 100% of the deposited HT is oxidized into HTO. Following oxidation, the HTO becomes associated with soil moisture and is emitted to the atmosphere via evaporation and plant transpiration. A number of studies, such as, Russell and Ogram (1988), Foerstel et al. (1988), Brown et al. 1990, and a review by Brown et al. (1996) have identified a range of loss rates for different soils, seasons, vegetation types, and other factors. These HTO loss rates have upper values of 10-20% per hour and lower values of less than 0.25 % per hour. Although these studies identified a range of HTO loss rates, there was common agreement that the initial rates were in the range of ~1% to 8% per hour for the first 48 hours after the HT release and then declined to rates of 0.25% to 1.5% per hour after 48 hours. These studies also indicate that HTO emissions from both plants and soil are zero during night-time hours when fog or dew are present. Consequently, the HTO loss rate for this assessment is based on an exponential decay probability distribution that varies from 0 to 8% per hour with an average loss rate of 1% per hour (Appendix 4).

The HT concentrations used in the Monte Carlo analysis are normal distributions around the maximum instantaneous and cumulative concentrations (at the locations of maximum exposure) derived from the RASCAL output. The highest dispersed tritium concentrations occurred at the maximally exposed off-site location (1 mile from source). HT air concentrations from the RASCAL analysis are multiplied by the HT deposition velocity range to obtain a range of cumulative HTO soil concentrations (for each area or portion of the dispersed HT plume). These HTO concentrations are then multiplied by the range of HTO loss rates to obtain a range of HTO emission concentrations in units of Curies per second per square meter (Ci/sec-m2).

The RASCAL output provides an estimate of the instantaneous maximum HT deposition rate (in Ci/m2/sec)(7) that is used in calculating inhalation exposure to the HT plume (Appendix 2). The range of instantaneous concentrations used for estimating inhalation doses is a normal distribution around the point estimate at 1 mile and 20E from the source (1970 release). The normal distribution, which looks like a bell-shaped curve, assumes that the mean value is the most frequent estimate of the true value and that estimated values are just as likely to under-estimate the true value as they are to over-estimate the true value.

Estimation of the atmospheric HTO concentration that is re-emitted from microbially transformed HT requires an additional set of calculations on the basis of the rate of HT deposition onto the soil surface and the rate of HTO re-emission. Ogram et al. (1988) found that there is little direct atmospheric conversion of HT to HTO (compared with the conversion rate from soil microbes). The HT deposition velocity used in this analysis is assumed to be a normal distribution using the rate of 4.0e-4 as the 10th percentile value and the measured rate of 1.2e-3 as the 90th percentile value. This HT deposition velocity range encompasses the commonly referenced rates identified by a variety of studies (Sweet and Murphy 1981; Dunstall et al. 1985; Spencer and Vereecken-Sheehan 1994; and Paillard et al. 1988).

Based on the previously referenced studies the HTO loss rate from soil to air varies from zero to >20% per hour and decreases with time after HT release and initial deposition to soil. It is important to note that emission of the HTO as soil moisture to the air is effectively zero during night-time if evaporation and plant transpiration cease and fog or dew are present (Forestel et al. 1988). A daily HTO loss rate must account for periods when emission is zero or greatly reduced from daytime values. This also indicates that the resulting HTO air concentrations will vary greatly over a 24 hour period. The HTO loss rate used in this analysis is an exponential decay distribution with an average value of 1 % per hour. All of the assumptions underlying the Monte Carlo analysis are included in Appendix 4.

The tritium emitted from the soil as HTO is also subject to dispersion. The Industrial Source Complex Short-term (ISC) air dispersion model developed by the Environmental Protection Agency (Lakes Environmental 2002) is used to estimate HTO concentrations in the breathing zone (one meter above ground surface). The ISC model uses hourly site-specific meteorological data and soil HTO emission rates to estimate the HTO air concentrations(8). Because HTO emissions and air concentrations are expected to change greatly over a daily cycle, this analysis uses the maximum 12 hour average to estimate doses for the initial daily concentration. Daily HTO air concentrations for subsequent days are based on an exponential decrease (averaging 1% per hour) of the day one air concentration.

Hourly surface weather data were not available for the years 1965 or 1970(9). Five years of hourly August weather data from the LLNL on-site weather station for the weather years 1990 to 1994 (Gouviea 2001) were used in model runs to determine the maximum 12 hour average HTO concentrations. Each annual analysis used two weeks of hourly data centered on the release dates (the 1970 analysis included hourly weather data from July 31 to August 13). Although the various weather years produced very similar hourly concentrations, the highest maximum concentrations were from weather year 1992, which was used in subsequent dose calculations for the 1970 exposure calculations. Details of the dispersion of the 1965 release are presented in the next section.

Soil HTO emissions for the ISC model are based on a series of eight rectangles oriented along the footprint of the dispersed tritium plume. The soils within these rectangles represent secondary tritium sources. For the 1970 release, the rectangles cover the area between 10º to 30º and extend approximately 4 miles from the LLNL tritium facility source. HTO emission rates for each rectangle are assumed equal within each rectangle. Rectangle 1 covers the area from 0 to 0.25 miles from the source (between the 10º to 30º radii; see Figure 1). Rectangles 2-8 are bounded at distances of 0.5, 0.75, 1.0, 1.5, 2, 3, and 4 miles, respectively. HTO emissions from the soil or plants were assumed to only occur during daylight hours. From 7:00 PM to 6:00 AM HTO emissions were assumed to be zero.

Figure 4A shows the probability distribution of HTO emission rates at one mile from the tritium facility stack (on the plume centerline). The mean and 95th percentile HTO emission rates were used for separate ISC model runs to estimate the mean and 95th percentile HTO air concentrations. HTO emission rates for each rectangular source area were adjusted for the different analyses. The resulting mean and 95th percentile HTO air concentrations were used to create a lognormal probability distribution which was used in the subsequent dose calculations (Figure 4B).

Only the initial or day 1 HTO air concentrations are based on the ISC model output. HTO air concentrations for days 2-12 are based on a constant decrease in the HTO soil source. The HTO emission rate equals the decrease in the HTO soil concentration. The use of only initial or day one dispersed concentrations from the secondary soil sources assumes that dispersion from the soil will not change significantly from day 1 to day 12.

This evaluation assumes that HTO concentrations in plant moisture are proportional with concentrations in soil moisture and that the plants are a conduit for the emission of soil HTO into the surrounding atmosphere. This assumption is supported by the field measurements of Spencer et al. (1988) who found that the HTO concentrations in soil, atmospheric, and plant moisture were in proportional equilibrium after a lag phase of 12-24 hours. Assuming that all deposited HT is re-emitted as HTO may overestimate the decline of the HTO soil source. However, this assumption will also result in a conservative estimation of the air HTO concentration, and double count any plant moisture tritium removed from the soil/air system and ingested as food (because it assumes the tritium is present in each compartment at the same concentration). The net result of this assumption is to cycle the tritium into the atmosphere and vegetation where human exposure may occur.

Estimated HTO emission rate at 1 mile from the LLNL tritium facility on the plume centerline
Figure 4A. Estimated HTO emission rate at 1 mile from the LLNL tritium facility on the plume centerline. Units are in Curies HTO per second per square meter. The mean and 95th percentile emission rates were used as inputs for the ISC model to estimate mean and 95th percentile HTO air concentrations.

Estimated lognormal probability distribution of HTO air concentrations at 1 mile from tritium facility
Figure 4B. Estimated lognormal probability distribution of HTO air concentrations at 1 mile from tritium facility. This distribution was generated by using the mean and 95th percentile HTO emission rates (Figure 4A) in separate ISC model runs to generate mean and 95th percentile HTO air concentrations. This distribution has a mean value of 9.8e-7 Ci/m3 (geometric mean of 6.5e-7 Ci/m3) and a 95th percentile value of 2.9e-6 Ci/m3. This distribution of HTO concentrations was used to calculate the 1 mile HTO inhalation doses.

Evaluation of the 1965 Accidental Release

The 1965 release has been estimated to be somewhat larger than the 1970 release--350 vs. 300 kCi (DOE 1992; Tate et al. 1999). The 1965 release occurred at approximately 3:30 PM on January 20, 1965. Release and site-specific weather data for the 1965 release are available in a recently released report (Peterson et al. 2002) and from daily weather observations by Quarterman (1965).

These data indicate that at the time of release, the wind was from 230º to 250º (or towards ~60º) at 7 knots (~3.6 m/sec). The maximum temperature for January 20th was 53ºF and the minimum temperature was 42ºF. Based on these weather observations, the atmospheric stability classification is most likely to be "B" or "C" (

Using the above meteorological observations in the RASCAL air dispersion model (with other release conditions as presented in Appendix 2 (Table A-2), the predicted HT dispersion and HTO emissions from soil are summarized in Table 1 as the average of the estimated maximum HTO air emission rates (Ci/second-m2). These are the average values of the initial release rates, which are at maximum rates because initial HTO soil concentrations are at their highest levels. These values are calculated by multiplying the cumulative HT air concentrations from Tables A-4 and A-5 by the HT deposition velocity (3.0e-4 to 1.2e-3 m/sec) times the HTO loss rate in percent per second (~0% to 8% per hour/3,600 seconds per hour). Specifically, these emission values represent the maximum rate at which HTO is emitted from the soil following initial HT deposition and conversion to HTO.

In Table 1, at distances of 1 to 2 miles from the tritium source, which corresponds to the maximally exposed locations, the HTO soil emission rates are higher for the 1970 release than for the 1965 release. Even though the amount of HT released in 1965 was greater than that of 1970, the meteorological conditions during winter in 1965 resulted in higher rates of HT deposition and HTO re-emission (per unit area) at close-in locations and lower HT deposition and HTO emission at off-site areas of potential exposure (1 mile or more from tritium facility). Conversely, the increased dispersion of the 1970 release results in a larger area of lower concentration tritium deposition and re-emission (the more dispersed plume covers a larger area). However, as stated in the introduction, this analysis is focusing on potential doses and health effects to maximally exposed individuals.

As the HT and HTO air concentrations for the 1965 release were less than those for the 1970 release, all subsequent dose calculations are based on the higher 1970 HT and HTO tritium concentrations. Because HTO emission rates at areas of maximum exposure were lower for the 1965 release, ISC dispersion analysis was not repeated for the 1965 release. The highest HTO emission rates for the 1965 release are for areas less than 0.5 miles from the tritium facility stack and are therefore at on-site locations. It is also significant to note that HTO emissions for the 1970 release (and the resulting air concentrations) are highest at the 1 to 1.5 mile locations. Consequently, these locations represent the area of maximum exposure and tritium doses at those locations will be commensurately higher than any other area.

Release 0.25 mile 0.5 mile 0.75 mile 1.0 mile 1.5 mile 2 miles 3 miles 4 miles
1965 1.9e-8 6.6e-9 3.1e-9 1.8e-9 8.0e-10 4.5e-10 2.0e-10 1.2e-10
1970 1.9e-11 5.2e-9 1.6e-8 2.2e-8 2.2e-8 1.6e-8 7.0e-9 2.7e-9
Table 1. Estimated average soil HTO emission rates (Ci/second-m2) for the 1965 and 1970 releases at the initial maximum rates. All estimates are on plume centerlines. The 1970 values are used in ISC-ST3 air dispersion model. Note that due to differences in meteorological conditions, the 1970 HTO emission values are higher at the maximally exposed offsite locations (1 to 2 miles from tritium source). These values are obtained by multiplying the cumulative air HT concentrations from Tables A-4 and A-5 by the HT deposition velocity (3.0e-4 to 1.2e-3 m/sec) times the HTO loss rate in percent per second (~0%to 8% per hour/3,600 seconds per hour).

Inhalation Exposure

Calculation of tritium inhalation doses requires an estimation of the tritium concentration at the point and time of exposure. Tritium inhalation doses are calculated for two time frames. Inhalation during the 30-minute passage of the HT plume presents the potential for direct inhalation of HT. Deposition of HT onto soil and subsequent emission as HTO presents the potential for HTO inhalation over a 10-12-day period (after 12 days the atmospheric HTO concentration is negligible (10)). The dose from HT exposure is ~10,000 times less effective than the dose from HTO exposure due to the very limited human uptake of hydrogen gas or tritiated hydrogen gas (HT) (ICRP 1996).

Potential 30-minute HT doses were calculated for a maximally exposed person at the 1 mile location (area of highest cumulative and instantaneous HT concentrations) using four different inhalation rates (resting, light activity, moderate activity, and strenuous activity). HT doses were calculated for ½ hour of exposure with concentrations based on a normal distribution of the instantaneous HT air concentration. Because the aerial photo analysis indicates that the closest residence is more than 1 mile from the tritium release point, potential 12-day HTO doses were also calculated for a maximally exposed person at the same 1 mile area downwind of the source and for the 2 mile location (from the source).

The 30-minute HT inhalation dose is calculated from a normal distribution of the instantaneous HT concentration. The 12-day HTO inhalation dose is based on a declining HTO concentration. Atmospheric HTO concentrations, and the resulting inhalation dose, decline over time as the HTO emitted from the soil surface is dispersed through the environment. The 12-day doses are calculated assuming an initial day one tritium concentrations as illustrated in Figure 4B, and assuming inhalation rates based on 7.5 hours per day resting, 14 hours per day of light activity, 1.7 hours per day of moderately strenuous activity, and 0.8 hours per day of strenuous activity.

The probability distributions, assumptions, and parameter statistics for each variable are presented in Appendix 4. Variables used in the Monte Carlo analysis are:

1) the energy of the tritium beta decay (ATSDR 2002A)
2) the effective half-life of tritium in the body (ATSDR 2002A)
3) body mass using a standard human model (EPA 1999)
4) tritium weighting factor (ATSDR 2002A)
5) the tritium dose and dose rate reduction factor (ATSDR 2002A)
6) the instantaneous HT concentration at each exposure location (Table A-3)
7) the cumulative HT concentration, or total HT deposition (Table A-4)
8) the HT deposition velocity based on measured data (see previous section for references)
9) HTO loss rates based on measured data (reviewed by Brown et al. 1996)
10) breathing rates (adult and child) for rest, light, moderate, and strenuous activity (EPA 1999)
11) exposure factor (days/day)

The mean and 95th percentile inhalation doses for several locations (relative to the 1970 release) are listed in Table 2. The 30-minute HT doses in Table 2 are based on the strenuous activity inhalation, which results in the maximum HT dose. The 95th percentile dose at the maximum exposure location is 0.7 mrem for children and 0.2 mrem for adults(11). The HT dose, in comparison with the 12-day HTO dose, is insignificant.

The average 12-day HTO dose to a maximally exposed adult (Table 2) is 5 mrem and the 95th percentile dose (same location) is 20 mrem. Doses for a maximally exposed child are 19 and 74 mrem, respectively. Doses for children are higher due to differences in body mass and breathing rates. Doses for more distant locations are significantly lower than the 1-mile location. Doses at 2 miles from the source are about 60% of the 1-mile dose (Table 2). The potential health effects of those tritium inhalation doses (plus the other dose components) and estimates of the potentially affected populations are discussed in the following section on "Public Health Implications of Estimated Tritium Doses."

The average daily HTO inhalation doses for a child (17 kg body weight) are presented in Figure 5. This figure shows the decline in dose for each day. The day one dose is ~3 mrem, the day 12 dose is ~ 0.2 mrem. Although theoretical doses could be calculated for additional days, the additional doses would not substantially change the total dose. Also, it is not clear that those theoretical doses would occur given the dispersion and cycling of tritium in the environment. The exposure that occurs over a 24-hour period is not to the daily maximum concentration, but rather to an average of all values over the 24-hour period. In order to be protective of public health, the doses in this evaluation are calculated from the highest 12 hour average concentrations, rather than 24 hour averages.

Distance from source 30-minute HT Dose (mrem) 12-day HTO Dose (mrem)
30 min. HT


30 min. HT

95th percentile

12 day HTO


12 day HTO

95th percentile

1 mile--adult 0.1 0.2 5 20
1 mile--child 0.3 0.7 19 74
2 miles--adult Same as above- all HT inhalation doses are calculated from concentrations at the 1 mile location 3 12
2 miles--child 12 43
Table 2. 30-minute HT inhalation doses and 12-day HTO inhalation doses (in mrem) at various distances from the tritium source for the 1970 release. All locations are along centerline (maximum concentration) of plume. The 1-mile location represents the maximally exposed individuals. Note that the estimated 12-day inhalation child doses are about 4 times larger than the adult doses due to differences in breathing rates and body weights relative to adults. The HT inhalation doses are based on strenuous exertion breathing rates at the location of the maximum HT concentration (1 mile from tritium facility stack).

Estimated average daily HTO child inhalation dose at 1 mile from the LLNL tritium facility (1970)
Figure 5. Estimated average daily HTO child inhalation dose at 1 mile from the LLNL tritium facility (1970). The cumulative 12-day inhalation dose is about 13 mrem. The dose for the 12th day is about 0.2 mrem. Although theoretical doses could be calculated for days after day 12, such additional doses do not constitute a significant addition and it is also unclear that such theoretical doses would actually occur given the environmental dynamics of tritium transformation and dispersion. These daily HTO inhalation doses are doubled to account for HTO absorption through the skin.

Dermal Absorption Exposure

HTO is absorbed into the human body through exposed skin (Osborne 1966; ICRP 1979). According to R. Osborne (personal communication 2003) and K. Eckerman (personal communication 2003) dermal absorption can be conservatively evaluated by doubling the estimated HTO inhalation dose. This means that HTO intake through the skin is equal to HTO absorption by the lungs via inhalation. Consequently, the total tritium doses estimated in the following section include a dermal absorption component that is equal to the HTO inhalation component.

Ingestion Exposure

ATSDR estimated the radiological dose resulting from ingestion of food crops in the Livermore area following the accidental tritium releases. Using data derived from the EPA National Center for Environmental Assessment, Exposure Factors Handbook (EPA 1999), ATSDR evaluated the water content for several types of fruits and vegetables and the estimated food intake of fruits and vegetables for adult males and females. The analysis indicated that the average water content of fruits was 85% water, whereas the average water content of vegetables was 88%. Our analysis also indicated that the average intake of fruits (in terms of water) in males of all ages was 124 grams per day and in females 125 grams per day. Although child ingestion rates are slightly smaller, they are within the same range; therefore, the higher adult values will be used for both children and adults. The water intake values for vegetables were 162 g/d and 135 g/d for males and females, respectively. Collectively, these intake rates result in water ingestion rates of about 0.3 L/day. Ingestion of dairy products from local herds was conservatively assumed to be 1 L/day for both children and adults.

Tritium concentrations in food items are based on measured concentrations in vegetation following the 1970 release (Myers et al. 1971). Measured vegetation tritium concentrations for this assessment are based on measurements from their stations VS-8 and VS-11, which were located along the approximate centerline of the 1970 plume (VS-8 was just north of the site boundary). Maximum measured values at these locations, which are at least 10 times larger than any other locations, varied from the detection limit of 5,000 pCi/L to a maximum value of 680,000 pCi/L (Myers et al. 1971; their Table 4). This range of values was used in the Monte Carlo analysis by assuming a normal probability distribution with a 10th percentile value of 5,000 pCi/L and a 90th percentile value of 680,000 pCi/L. The probability distribution of food HTO concentration had an average value of 343,000 pCi/L(12).

Tritium concentrations in milk are also based on measured concentrations. Of the 13 milk analyses conducted, only three had detectable concentrations ranging from 6,000 to 8,000 pCi/L (Myers et al. 1971; their Table 7). Dose estimates for this assessment assumed a normal probability distribution with an average value of 7,000 pCi/L and a standard deviation of 700 pCi/L. The above tritium distributions for both milk and vegetation are used as initial, or day one concentrations. Tritium concentrations in both milk and vegetation are assumed to decline in similar proportion to the declining concentrations in soil and dispersed air. The assumption of decreasing vegetation tritium concentrations is supported by measurements of tritium in vegetation as reported by Spencer et al. (1988). Their measurements of tritium in vegetation following a test release found that tritium concentrations in a variety of plants decreased by rates of ~0.5% to 8% per hour. As with air and soil tritium concentrations, we assumed that vegetation and milk concentrations declined by the HTO loss rate (0 to 8% per hour; average value of 1% per hour).

Using the above distributions of vegetation and milk tritium concentration, and using the water content and intake of the foods, we then estimated the intake of HTO into the body and the resulting dose. For the estimation of radiological dose, we used the values discussed in the ATSDR Expert Panel Report for the radiation weight factor and the DDREF (see ATSDR [2002A]and Appendix 4 for details of the dosimetry assumptions and probability simulations). The calculations also assumed that the ingestion of potentially contaminated foods occurred over 12 days.

The OBT component of the ingestion dose is estimated by multiplying the HTO dose times a "percentage increase in dose from tritium as OBT above that from tritium in the form of water (HTO)…" (from Table 2.3 of the Expert Panel Report; ATSDR 2002A). The percentage increase values in that table range from 16% to 400% and the panel concludes that the most likely range of increase is less than 32%. Consequently, we have assumed that the ingestion dose due to OBT is a triangular distribution ranging from 16% to 400% of the HTO dose with a most likely value of 32%.

The results of these calculations are illustrated in Figure 6. This figure shows the probability distribution of the estimated child tritium dose, in mrem, due to ingestion of vegetables, fruit, and milk for the 12-day period following the 1970 release. The average dose is about 0.4 mrem and the 95th percentile dose is about 1.5 mrem. These potential tritium ingestion doses will be added to inhalation doses and potential long-term tritium exposures doses to estimate cumulative doses during 1970.

Probability distribution of 12-day child ingestion doses following the 1970 LLNL accidental tritium releases
Figure 6. Probability distribution of 12-day child ingestion doses following the 1970 LLNL accidental tritium releases. The average ingestion dose is 0.4 mrem, and the 95th percentile ingestion dose is about 1.5 mrem. Child doses are approximately 3 times larger than adult doses due to the smaller body mass of children (relative to adults). Approximately 2/3 of the ingestion dose is due to HTO with the remaining 1/3 due to OBT ingestion. Assumptions underlying the dose estimation are discussed in the text.

Total Tritium Doses and Uncertainties in Dose Modeling Procedures

A summary of total or cumulative tritium doses is presented in Table 3 and Figure 7 (child doses). The cumulative doses include the 30-minute HT inhalation dose, the 12-day HTO inhalation dose, the 12-day HTO dermal absorption dose, the 12-day ingestion dose, and potential contributions from long-term or chronic environmental tritium exposure (ATSDR 2002A). The 12-day HTO inhalation and dermal absorption doses are several orders of magnitude larger than any of the other potential dose components.

The tritium doses presented in Figures 6 and 7 are presented in terms of probability distributions. These figures show the relative probability or frequency of occurrence for different doses from the Crystal Ball simulations (in mrem). Both figures show the estimated frequency of doses for a child from the location with the highest HT and HTO concentrations (1 mile from the tritium facility). Both figures show that the most frequent or probable doses are at the lower ends of the dose ranges and that high doses are relatively rare. Note that the mean or average doses in both figures over-estimate the most frequent or probable doses. The cumulative doses presented in Table 3 for child and adult doses at the 1 and 2 mile locations are presented in terms of the mean and 95th percentiles of the estimated probability distributions.

At the point of maximum exposure (1 mile from the tritium source), the estimated mean and 95th percentile cumulative doses (1970 annual dose) ranged from 41 to 149 mrem for children, and from 11 to 42 mrem for adults. Figure 7 shows the distribution of estimated total tritium doses to a child at one mile from the tritium facility stack. As a comparison, exposure to background radiation in the United States is on the order of 360 mrem/year (3.6 mSv/year) including atmospheric radon. Therefore, the average doses (to children) represent about 11% of the typical annual exposure one receives from background radiation. Doses were also calculated for 30-minute exposure to the HT plume and for ingestion of food items containing HTO and OBT. Estimated doses for those exposures are only fractions of a mrem and do not constitute a significant route of tritium exposure relative to the inhalation and dermal absorption of HTO released from soil.

The estimated inhalation doses in Tables 2 and 3 do not include an explicit contribution from OBT. However, the dosimetry parameters included in the Monte Carlo analysis do include an implicit acknowledgment of OBT contributions. For example, the effective half-life (biological retention) term varies from 1 to 40 days with a most likely value of 10 and an average value of 17 days. This distribution reflects the potential longer retention of OBT relative to HTO. Similar acknowledgment of potential OBT contributions is also reflected in the tritium-weighting factor and the dose and dose rate reduction factor (Appendix 3). Food ingestion doses do include an explicit OBT contribution, however, the total ingestion dose is not significant relative to the much larger 12-day HTO inhalation dose.

It should be noted that one of the peer reviewers suggested that the estimated food ingestion doses are significantly lower than doses estimated from a simple model of soil moisture HTO concentration (Appendix 1, comment 76).

The potential average tritium doses (to the maximally exposed individuals) presented in Table 3 range from about 11 to 41 mrem for adults and children (respectively; 95th percentile doses from 40 to 149 mrem). Because these doses are based on the health protective assumptions concerning potential exposure to outside air and food ingestion, it is unlikely that any member of the public was actually exposed at those concentrations. Myers et al. (1971) provided similar total dose estimates of potential exposure (3 mrem inhalation plus 70 mrem milk ingestions) at the site boundary), albeit with distinctly different dose component contributions. It should be noted that the Myers et al. (1971) estimate of dose from milk was based on a hypothetical cow consuming pasture grasses contaminated at the maximum measured HTO concentration and an assumed moisture content of 4 L/m2. The Myers report states that "a) no other locations reaching this level of contamination were found and b) it is unlikely that any such locations were missed because of the extensiveness of the survey." Our dose estimates are based on milk measurements from the dairy cows that were present in the plume area.

Estimated total tritium dose to a child from the August 6, 1970 tritium release
Figure 7. Estimated total tritium dose to a child from the August 6, 1970 tritium release. This dose is at a distance of one mile from the LLNL tritium facility and represents the location of maximum exposure. The dose includes a 30 minute HT inhalation dose, a 12 day HTO inhalation and skin absorption dose, a 12 day HTO and OBT food ingestion dose, and an annual HTO and OBT dose (inhalation and ingestion). The average dose is ~41 mrem and the 95th percentile dose is 149 mrem. Note that due to the shape of the probability function, the most likely doses are considerably lower than the average dose.

In addition to the tritium measurements in milk, the Myers et al. (1971) study also analyzed urine samples from potentially exposed people, including both LLNL workers present at the release source and potentially affected community members. Those measurements did not detect any elevated tritium body burdens. The detection limit for those human samples corresponded with a potential tritium exposure of 0.025 mrem. The results of those human biological samples indicate that the modeled exposures in both the Myers study and this analysis over-estimate the actual human exposures.

Examples of the health protective assumptions used in this evaluation include use of daily 12 hour maximum (rather than 24 hour average) HTO concentrations to estimate total daily exposure. The use of daily maximum HTO air concentrations instead of daily average concentrations is significant. Because the source concentrations are continually declining and HTO soil emissions are negligible during night-time actual exposure will be to the integrated daily average, rather than the 12 hour maximum. This analysis also assumes that in-house HTO concentrations are the same as outdoor concentrations. As soil and plants are the pathway for soil moisture to enter the atmosphere, indoor air, most likely, has lower HTO concentrations.

Doses in mrem (mSv) 30-minute HT inhalation 12-day HTO inhalation 12-day HTO dermal absorption 12-day ingestion Annual or chronic doses Total tritium Dose mrem/yr
~41 (0.41)
~149 (1.49)
95th %
11 (0.11)
42 (0.42)
Table 3. Short term and cumulative tritium doses to maximally exposed individuals following the 1970 accidental tritium release from the LLNL tritium facility. Doses are in mrem with total doses also in milliSievert (mSv). The derivation of the annual doses is described in ATSDR's 2002 health consultation on LLNL, which did not estimate specific annual doses for children. Total child tritium doses assume that annual child doses are 3 times the annual adult dose. As the expert panel report indicated that the estimated annual doses represent maximum values, the stated annual doses are used for both mean and 95th % values. Values are rounded so component doses may not precisely sum to total dose.

The exposure assessment also assumes that all deposited HT is re-emitted as atmospheric HTO with no losses to other media or retention in soil or groundwater. The result of this assumption is to maximize air concentrations for inhalation exposure. Transpiration from plants and evaporation from soil are the primary conduits for transferring tritium from soil to air. The exposure evaluation also assumes that the potentially exposed individuals spent 24 hours per day for 12 days in the plume area and that each individual spent 1.7 hours per day of moderate exertion, and 0.8 hours per day of strenuous exertion. The 30-minute HT inhalation exposures are based on potential exposure at the 1 mile location (maximum concentrations) using strenuous activity (maximum) inhalation rates.

Residents living within the plume area beyond the 1 to 2 mile locations were also exposed to tritium from the accidental releases. However, those doses significantly decrease with distance from the LLNL tritium facility. Based on dispersed air concentrations, doses at 2 miles are about 60% of those at the 1 mile location and are 31% and 12% at 3 and 4 miles, respectively. Also, in 1965 and 1970, the areas northeast and east of the LLNL facility were very sparsely populated, so there were very few people exposed at low doses in areas beyond the 1 to 2 mile area of maximum exposure.

The dose estimates in Table 3 assume that all of the released tritium was in an HT form. There would be an additional dose component if some percentage of the tritium release occurred in an HTO form. The RASCAL model estimates a maximum dose of 103 mrem (at 1 mile) based on the assumption that 100% of the tritium plume was present as HTO. Noguchi et al. (1989) and Noguchi (1995) conclude that the conversion of HT to HTO in air will be less than 0.5% over the time period of a few hours (according to the RASCAL model the HT plume has completely dispersed in 30 minutes). Using the RASCAL dose estimates, if 1% of the HT plume was present in an HTO form, it would add about 1 mrem to the total tritium dose at the point of maximum exposure (1% of 103 mrem is ~1 mrem). Alternatively, if 1% of the 30 minute HT inhalation dose from Table 2 was present in the HTO form, the increase in dose would be about 30 mrem (1% of 0.3 mrem x 10,000= 30 mrem). A potential, albeit unlikely, additional dose component of between 1 and 30 mrem is within the probability distribution of Figure 7 and Table 3.

The estimates of tritium ingestion also rely on conservative assumptions that will overestimate potential doses. The food intake rates assume that all produce consumed is locally derived from a home garden. Also, according to the California Agricultural Extension Service, fruits and vegetables harvested in August would include only specific items, such as tomatoes, squash, certain apples, and pears (J. Schweggman, University of California Agricultural Extension Service, personal communication from M. Evans, Agency for Toxic Substances and Disease Registry, April 2, 2002). Cold weather crops such as lettuce, broccoli, and other items would not be harvested during August. Conversely, the opposite would occur during the January release with only harvesting of cold weather crops, but no summer vegetables or fruit. Additionally, it is very unlikely that any August garden would produce significant yields without extensive irrigation. Such irrigation would dilute the tritium from residual soil moisture and result in much lower tritium concentrations in vegetation (Spencer et al. 1988). Similarly, there is no pasture grazing for dairy or beef cattle during the August timeframe (as indicated by the very low tritium measurements in milk and livestock; Myers et al. 1971). The ingestion dose also assumes that 1 liter per day of locally produced milk or milk products are consumed.

Appendix 1, comment 76 contains an alternate estimate of the ingestion dose. This estimate, based on theoretical assumptions of tritium distribution and partitioning within soil and plant moisture produces ingestion estimates of ~20 mrem. This dose estimate is considerably larger than that calculated from tritium concentrations in plant moisture (0.4 mrem). However, this estimate is based on an assumed distribution of 10% soil moisture which likely overestimates tritium concentrations in plants during the dry conditions of August. We believe actual plant tritium concentrations from the plume area are a better measure on which to base dose estimates.

Similar health protective assumptions were used by the expert panel to estimate long-term tritium doses (ATSDR 2002A). The air dispersion and tritium dosimetry values used in this exposure analysis are estimates for the parameters. Ranges and most likely values of dosimetry parameters were obtained from an expert panel review of tritium issues (ATSDR 2002A). The net result of using these conservative or health protective assumptions is the likely over-estimation of potential rates of tritium exposure.

1 Myers et al. (1971) indicate that 289,000 Ci (289 kCi) of tritium were released in the 1970 accident. This assessment conservatively rounds this value up to 300 kCi.
2 Any non-residential buildings such as barns, sheds, or commercial buildings are counted as houses with a commensurate increase in the estimated residential population.
3 Although ATSDR does not have a contractual process for directly providing funds to community members, we have submitted this public health assessment for independent technical peer review.
4 In addition to the above variables related to HT plume dispersion and exposure, there are also a number of variables related to tritium dosimetry. These dosimetry variables, along with their sources of uncertainty and most likely values, are discussed in detail in the ATSDR health consultation on "Tritium Releases and Potential Off-Site Exposures" (ATSDR 2002A). Those discussions are not repeated in this document.
5 Myers et al. (1971) indicate that, due to plume rise, the effective stack height was 45 m. RASCAL computes plume rise from known discharge and stack parameters. The specific parameters used for the plume rise calculations are presented in Appendix 2.
6 The RASCAL concentrations (Appendix 2) are specifically output as cumulative surface concentrations in Ci/m2, based on a default deposition rate of 0.3 cm/sec. These values are divided by the default deposition rate (0.3 cm/sec) to obtain the cumulative air concentrations. In subsequent dose calculations, these point value air concentrations are used as the average value of a normal distribution (see Appendices 2 and 4).
7 The instantaneous HT deposition rate (Ci/m2/sec) is divided by the RASCAL default deposition velocity (cm/sec) to obtain an instantaneous HT air concentration (Ci/m3).
8 The HTO loss rate and HTO emission rate are both measures of the rate of HTO transfer from soil to atmospheric moisture. The loss rate is defined in terms of percent per hour and is used to measure the decreasing HTO concentration in soil moisture. The HTO emission rate is used as input to the ISC model and is an area-based unit of Ci/sec-m2.
9 A request for the hourly surface weather observations for the Livermore Airport (National Weather Service ID 044997) from the National Climatic Data Center indicated that data for the years 1965 and 1970 are not available. However, daily weather records from the National Weather Service volunteer observer network (M. Quarterman, 2284 Evans St., Livermore CA) are available and have been incorporated in the dispersion calculations.
10 After 12 days, more than 95% of the soil HTO has been emitted to the atmosphere assuming there has been no additional uptake by plants or dilution by rainfall or dew. Doses for days past day 12 are assumed to be accounted for in the uncertainties in the concentration and exposure parameters and in the chronic doses which are added to the short term doses from the accidental releases.
11 The maximum dose for a 30 minute exposure is taken at the 1 mile location, which is the area of highest HT concentration and the location of the closest residence to the LLNL tritium facility.
12 The Gaussian air models on which this evaluation is based have an underlying assumption of normal distribution of concentrations in the plume. Consequently, this assessment similarly assumes that HT deposition to soil and HTO concentrations in plants will be normally distributed.

Next Section     Table of Contents The U.S. Government's Official Web PortalDepartment of Health and Human Services
Agency for Toxic Substances and Disease Registry, 4770 Buford Hwy NE, Atlanta, GA 30341
Contact CDC: 800-232-4636 / TTY: 888-232-6348

A-Z Index

  1. A
  2. B
  3. C
  4. D
  5. E
  6. F
  7. G
  8. H
  9. I
  10. J
  11. K
  12. L
  13. M
  14. N
  15. O
  16. P
  17. Q
  18. R
  19. S
  20. T
  21. U
  22. V
  23. W
  24. X
  25. Y
  26. Z
  27. #