2.6. Summary of Unremediated Site Risks
- Human Health Risks
- Contaminant Identification
- Media of Concern
- Contaminants of Concern
- Concentrations of Chemicals of Concern Used in the Risk Assessment
- Exposure Assessment
- Exposure Pathways
- Potentially Exposed Population
- Exposure Point Concentration Estimates
- Exposure Frequency and Duration
- Toxicity Assessment
- Cancer Potency Factors
- Reference Doses for Noncarcinogens
- Risk Characterization
- Carcinogenic Risks
- Potential for Noncarcinogenic Effects
- Combined Carcinogenic Risks and Hazard Indices
- Sources of Uncertainty
- Environmental Risks
- Risk Assessment Conclusions
As part of the RI report (Thorpe et al., 1990), the BPHA (Layton et al., 1990)
was conducted to estimate the potential future health risks if contaminants in
ground water and sediments originating from LLNL were not remediated.
Evaluation of a no-action scenario is a requirement of the NCP, 40 CFR section
300.430(e)(6), to represent a baseline condition. In addition, a risk
assessment was conducted as part of the FS (Isherwood et al., 1990) to estimate
the potential public health risks if the concentrations of VOCs in ground water
were reduced to their respective MCLs. These and other assessments of
potential risks are summarized in the PRAP (Dresen et al., 1991) and below.
Details of the risk assessments are contained in the RI and FS.
The LLNL risk assessment consisted of several steps:
- Identifying the contaminants of concern (see Section 2.5 of this ROD).
- Identifying the media through which exposure may occur.
- Assessing the exposure.
- Assessing the toxicity of each contaminant.
- Quantifying the risk.
Each of these is discussed below.
The primary medium through which public exposure to LLNL contaminants may occur
is ground water. Air is also a medium of concern for contaminants that may
volatilize from contaminated soil or ground water. The public is not directly
exposed to contaminated soils because no offsite surficial soils contain
significant concentrations of contaminants originating from LLNL. Contaminated
onsite surficial soils were evaluated as a potential medium of concern.
However, a screening analysis of the risks resulting from potential onsite
exposure to contaminated soils has shown these risks are insignificant (Layton
et al., 1990; Hoffman, 1991b; Macdonald et al., 1991). Therefore, surficial
soils are not a medium of concern for the LLNL site.
A screening analysis was conducted to determine which substances and exposure
pathways are potentially important from the perspective of potential adverse
health effects. A statistical analysis of thousands of water and soil samples
estimated the relative abundance of particular contaminants in the study area
(Layton et al., 1990). TCE, PCE, and chloroform account for an estimated 91%
of the total amount of VOCs dissolved in the LLNL-area ground water. Of the
remaining VOCs, the most hazardous are carbon tetrachloride and 1,1-DCE, which
were used to represent the potential adverse effects of the remaining 9% of the
VOCs. Nearly 60% of the mass of the remaining 9% of VOCs is 1,1-DCE. These
compounds were used to estimate the public health risks resulting from the
offsite migration and domestic use of contaminated ground water. According to
the U.S. EPA, PCE, TCE, chloroform, and carbon tetrachloride are classified as
B2 carcinogens, which are described as "probable human carcinogens indicated by
sufficient evidence in animals and inadequate or no evidence in humans" (U.S.
EPA, 1989a). 1,1-DCE is classified as a Class C carcinogen by the U.S. EPA
(possible human carcinogen).
Other contaminants in soil and ground water include benzene at the Gasoline
Spill Area, tritium, and inorganic substances, such as chromium, lead, nitrate,
sulfate, and manganese. A screening analysis of the transport and fate of
benzene indicates that benzene or other gasoline-related contaminants (toluene,
xylene isomers, and ethylbenzene) are not likely to reach detectable
concentrations west of LLNL. Similarly, tritium continues to undergo
radioactive decay with a 12.3-year half-life such that by the time ground water
containing elevated levels of tritium would migrate to the western LLNL
boundary in the absence of remediation, concentrations would be within
background levels. As stated in Section 2.5.4, LLNL plans to monitor tritium
in ground water over the life of the cleanup.
As discussed in a letter to the regulatory agencies (Hoffman, 1992), there is
strong evidence that the lead in LLNL ground water is naturally occurring.
Furthermore, as described in Section 2.5.3, it appears that the migration
potential for lead is very low, and its occurence above the remediation
standard is very limited. Several inorganic substances, including chromium,
nitrate, sulfate, and manganese, occur in ground water in concentrations
exceeding regulatory limits in various monitor wells, sporadically located
onsite and offsite. Except perhaps for chromium, which has been used in LLNL
cooling towers, the observed concentrations appear to reflect background levels
of these constituents in ground waters in the Livermore Valley.
To assess the ground water exposure pathway, migration of the five VOCs of
concern (PCE, TCE, chloroform, carbon tetrachloride, and 1,1-DCE) was simulated
using the January-September 1988 concentrations as initial conditions. These
concentrations range from the various detection limits up to a maximum of 6 ppm
for TCE in the Building 518 Area.
The only potential exposure pathway for present and future offsite populations
is use of contaminated well waters. For domestic water uses, the potential
exposure pathways are ingestion of drinking water, inhalation of volatile
substances, and entry through the skin. For irrigation uses, the potential
exposure pathways are inhalation of volatilized chemicals from sprinklers, and
ingestion of foods from crops or home gardens irrigated with water containing
the chemicals of concern. Exposure from contact with surface water runoff or
sediment in local arroyos that receive drainage waters from the LLNL site is
not a pathway of concern, because no chemicals of concern have been detected in
downstream drainage channels near LLNL, and ground water does not discharge to
streams near LLNL. The most important offsite exposure pathways with regard to
health risk are those that result from domestic well water use from offsite
wells (Thorpe et al., 1990).
As described in the BPHA and in Section 2.6.1.1.1 above, there are no
significant onsite exposure pathways for LLNL site contaminants. Prior to any
soil excavation at LLNL, the existing soil cleanup data are reviewed and maps
of known or suspected contamination are consulted to determine whether
additional sampling needs to be conducted prior to excavation. If no samples
have been previously collected in a given area, preconstruction sampling is
performed before excavation begins. If contamination is found, appropriate
safety and disposal practices are overseen by the LLNL Hazards Control
Department.
The only potentially exposed offsite population consists of residents who use
ground water that has migrated from LLNL. In the assessments of risk for the
LLNL site, a future residential-use scenario was not considered because it is
unlikely that transfer of ownership of the site from DOE would occur in the
foreseeable future. No change in ownership of the LLNL Main Site or any
portion thereof, or notice pursuant to Section 120 of CERCLA, will relieve DOE
of its obligation to clean up contamination resulting from DOE activities, or
any future contamination resulting from DOE activities at LLNL. In addition,
no change of ownership of the site or any portion thereof will be consumated by
DOE without provision for continued maintenance of any containment system,
treatment system, monitoring system, or other response action(s) installed or
implemented under terms of the LLNL FFA.
To assess the potential future health risks of the known contaminants in ground
water, the movement of VOCs from their current distribution was simulated with
a model. A semianalytical model of contaminant transport and fate in ground
water was used that considers advection, dispersion, retardation, and
degradation. The BPHA contains details on the assumptions and the parameters
used in the model.
To address uncertainty inherent in all contaminant migration calculations, two
scenarios were investigated, one called "best-estimate" and the other
"health-conservative." The health-conservative scenario uses parameter values
and assumptions that yield exposures that are very unlikely to be exceeded.
U.S. EPA prefers using the most conservative of the health-conservative
scenarios (footnote "b," Table 4, Section 2.6.1.4.3) as their estimate of the
potential health risk from the LLNL site. The best-estimate simulations use
parameter values that are considered to be the most likely or the most
representative, based on existing knowledge of the LLNL ground water system and
contaminant properties. Best-estimate simulation assumes no human exposure to
the ground water until it reaches the currently used municipal supply wells in
downtown Livermore because no private wells are currently contaminated and
administrative control limits the potential for domestic well installation into
a contaminated zone. The administrative control consists of notification by
Zone 7, the local water agency, that a proposed new well is in or near the
contaminant plume.
The exposure period for the offsite public for any exposure pathway of concern
was assumed to be a 70-year lifetime. For offsite exposures to contaminated
ground water, the fate and transport model was used to calculate maximum
70-year average concentrations in ground water at existing and potential
offsite wells. It was assumed that the exposed population uses ground water as
its sole source of domestic water for this continuous 70-year period. These
and other assumptions were used to estimate the total daily uptake of each
chemical of concern in milligrams of chemical per kilogram body mass per day
(mg/kg-day).
Cancer potency factors (CPFs) have been developed by U.S. EPA to estimate
excess lifetime cancer risks associated with exposure to potentially
carcinogenic chemicals. CPFs, expressed in units of (mg/kg-day)-1, are
multiplied by the estimated intake of a potential carcinogen, in mg/kg-day, to
provide an upper-bound estimate of the excess lifetime cancer risk associated
with exposure at that intake level. The term "upper bound" reflects the
conservative estimate of the risks calculated from the CPF. Use of this
approach makes underestimation of the actual cancer risks highly unlikely.
CPFs are derived from the results of human epidemiological studies or chronic
animal bioassays to which animal-to-human extrapolation and uncertainty factors
have been applied (e.g., to account for the use of animal data to predict the
effects on humans).
CPFs for the LLNL chemicals of concern are listed in Table 2. In conformance
with EPA methodology, cancer potencies are based on applied, rather than
metabolized, doses.
Reference doses (RfDs) have been developed by EPA for indicating the potential
for adverse health effects from exposure to chemicals exhibiting
noncarcinogenic effects. RfDs, which are expressed in units of mg/kg-day, are
estimates of lifetime daily exposure levels for humans, including sensitive
individuals. Estimated intakes of chemicals from environmental media (e.g.,
the amount of a chemical ingested from contaminated drinking water) can be
compared to the RfD. RfDs are derived from human epidemiological studies or
animal studies to which uncertainty factors have been applied (e.g., to account
for the use of animal data to predict the effects on humans). These
uncertainty factors help ensure that the RfDs will not underestimate the
potential for adverse noncarcinogenic effects to occur.
Reference doses for the LLNL chemicals of concern are listed in Table 3.
The information from the preceeding steps was combined to determine if an
excess health risk would exist if the site were not remediated. Excess
lifetime cancer risks are determined by multiplying the intake level with the
CPF. These risks are probabilities that are generally expressed in scientific
notation (e.g., 1 x 10-6 or 1E-6). An excess lifetime cancer risk of 1 x 10-6
indicates that, as a plausible upper bound, an individual has a one in one
million chance of developing cancer as a result of site-related exposure to a
carcinogen over a 70-year lifetime under the specific exposure conditions at a
site.
Tables A-1 and A-2 in Appendix A summarize the estimated cancer risks for
offsite exposure to ground water for both the best-estimate and
health-conservative exposure scenarios for PCE, TCE, 1,1-DCE, chloroform, and
carbon tetrachloride. Under the best-estimate exposure scenario (Table A-1),
the greatest incremental cancer risk is seven in ten million (7 x 10-7), which
is associated with a well 2 miles west of the LLNL site that is in the path of
the plume containing the highest concentrations of 1,1-DCE. Under the
health-conservative exposure scenario (Table A-2), the incremental cancer risks
are on the order of one in one thousand (10-3) to one in one million (10-6) for
all wells. The highest predicted risk, two in one thousand (2 x 10-3), is for
a hypothetical well about 250 feet west of the LLNL site. However, no such
wells have been constructed to date or are planned for installation prior to
cleanup. The most conservative of the health-conservative scenarios (i.e., the
one with the 2 x 10-3 incremental risk) is the scenario prescribed by EPA for
the LLNL site.
Potential noncarcinogenic effects of a single contaminant in a single medium is
expressed as the hazard quotient (HQ) (or the ratio of the estimated intake
derived from the contaminant concentration in a given medium to the
contaminant's reference dose). By adding the HQs for all contaminants within a
medium or across all media to which a given population may be reasonably
exposed, the hazard index (HI) can be estimated. If only one compound is
involved, then the HQ is equivalent to the HI. If the HI value is greater than
1.0, exposure could result in adverse health effects. The HI provides a useful
reference for gauging the potential significance of multiple contaminant
exposures within a single medium or across media.
Tables A-3 and A-4 in Appendix A summarize the estimated HQ's for offsite
exposure to ground water for both the best-estimate and health-conservative
exposure scenarios for the chemicals of concern at LLNL. Under the
best-estimate exposure scenario (Table A-3), the greatest HQ is 1.4 x 10-3,
which is for a hypothetical well 2 miles west of the LLNL site in the path of
the plume containing the highest concentrations of carbon tetrachloride. Under
the health-conservative exposure scenario (Table A-4), the HQ's are on the
order of 10-2 to 10-1 for all wells. The highest predicted HQ (0.8) is for a
hypothetical well that is 250 feet west of the LLNL site.
The maximum theoretical excess cancer risks for a hypothetical, no-remediation
scenario, based on the assumption that an individual will use well water for a
70-year (lifetime) period, are presented in Table 4. The maximum additional
cancer risk associated with the best-estimate scenario in Table 4 means that
the cancer risk from a lifetime exposure to VOCs (PCE, TCE, chloroform, and
carbon tetrachloride) in well water derived from a downtown Livermore municipal
supply well could be as high as 7 in 10 million (7 x 10-7), using EPA
assessment methods. This means that each individual that consumes 2 liters
(about 2 quarts) of this water each day for 70 years would increase his or her
risk of developing cancer by 7 in 10 million above the normal 1 in 4 cancer
risk for Americans (U.S. EPA, 1989a). The HI associated with the best-estimate
scenario is far below 1.0, indicating exposure at the predicted concentrations
would not produce any adverse health effects from noncarcinogens (see the RI,
Thorpe et al., 1990, for details).
Under the health-conservative no-remediation scenario, the maximum additional
cancer risk is two in one thousand (2 x 10-3) for a lifetime exposure to
contaminants in water from a potential monitor well drilled 250 feet west of
LLNL. The HI calculated for this scenario is 1. Because no drinking water
wells are likely to be drilled in the area 250 feet west of LLNL, we also
calculated the risk based on a lifetime exposure to well water derived from
downtown Livermore using the health conservative assumptions. This unlikely
scenario results in a maximum additional cancer risk of one in one thousand (1
x 10-3) and an HI of 1. The HI of 1 for the health-conservative scenario
indicates that there is some potential for noncarcinogenic health effects if
the very conservative assumptions of the health conservative scenario were ever
realized, and if there was an additive effect of all the individual compounds.
Both health-conservative risks in Table 4 exceed EPA's one in ten thousand to
one in ten million (1 x 10-4 to 1 x 10-7) acceptable risk range for Superfund
sites.
Uncertainties are associated with all estimates of cancer and noncancer health
hazards. These uncertainties result from incomplete knowledge of many physical
and biological processes, such as carcinogenesis. Where specific information
is not available, it is necessary to make assumptions and/or use predictive
models to compensate for lack of information. The assumptions, models, and
calculations are chosen such that the resulting risk and hazard estimates are
health-conservative. The specific sources of uncertainty in the risk and
hazard estimates presented here are further discussed in the BPHA.
Currently, there is no potential risk of ecological impacts related to
environmental exposure to ground water because no ground water containing
contaminants is present at the surface, either onsite or offsite. No perennial
streams exist at or near the site and no streams receive flow from ground
water. No critical habitats are affected by the ground water and soil
contamination. No endangered species or habitats of endangered species are
affected by the site contaminants, as described in the FS (Isherwood et al.,
1990).
In summary, the identified compounds of concern, if not addressed by
implementing the response actions selected in this ROD, may present a potential
risk to public health.
2.7. Description of Remedial Alternatives
UCRL-AR-109105
