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

PUBLIC HEALTH ASSESSMENT

PADUCAH GASEOUS DIFFUSION PLANT (U.S. DOE)
PADUCAH, MCCRACKEN COUNTY, KENTUCKY


APPENDIX C: HEALTH GUIDELINES, COMPARISON VALUES, AND EXPOSURE FACTORS

When a hazardous substance is released to the environment, people are not always exposed to it.Exposure happens when people breathe, eat, drink, or make skin contact with a contaminant. Peoplecan also be exposed to radioactive contaminants by irradiation--if they get close to the radioactivematerial and if the contaminants are present at high concentrations.

Several factors determine the type and severity of health effects associated with exposure tocontaminants. Such factors include exposure concentration, frequency and duration of exposure,route of exposure, and multiplicity of exposure (i.e., the combination of contaminants and routes).Once exposure takes place, individual characteristics--such as age, sex, nutritional status, genetics,lifestyle, and health status--influence how that person absorbs, distributes, metabolizes, and excretesthe contaminant. These characteristics, together with the exposure factors discussed above and thespecific toxicological effects of the substance, determine the health effects that may result.

ATSDR considers these physical and biological characteristics when developing health guidelines.Health guidelines provide a basis for evaluating exposures estimated from concentrations ofcontaminants in different environmental media (soil, air, water, and food) depending on thecharacteristics of the people who may be exposed and the length of exposure.

ATSDR reviews health and chemical information in documents called toxicological profiles. Eachtoxicological profile covers a particular substance; it summarizes toxicological and adverse healtheffects information about that substance and includes health guidelines such as ATSDR's minimalrisk level (MRL), EPA's reference dose (RfD) and reference concentration (RfC), and EPA's cancerslope factor (CSF). ATSDR public health professionals use these guidelines to determine a person'spotential for developing adverse non-cancer health effects and/or cancer from exposure to ahazardous substance. ATSDR does not have guidelines for exposure to radioactive materials.Instead, the agency uses existing regulatory values and national or international recommendations.

An MRL is an estimate of daily human exposure to a contaminant that is likely to be without anappreciable risk of adverse non-cancer health effects over a specified duration of exposure (acute,less than 15 days; intermediate, 15 to 364 days; chronic, 365 days or more). Oral MRLs areexpressed in units of milligrams per kilogram per day (mg/kg/day); inhalation MRLs are expressedin micrograms per cubic meter (g/m3). MRLs are not derived for dermal exposure.

RfDs and RfCs are estimates of daily human exposure, including exposure to sensitivesubpopulations, that are likely to be without appreciable risk of adverse non-cancer health effectsduring a lifetime (70 years). These guidelines are derived from experimental data and lowest-observed-adverse-effect levels (or no-observed-adverse-effect levels), adjusted downward usinguncertainty factors. The uncertainty factors are used to make the guidelines adequately protective ofpublic health. RfDs and RfCs should not be viewed as strict scientific boundaries between what istoxic and what is nontoxic.

For cancer-causing substances, EPA established the CSF [1]. A CSF is used to determine thenumber of excess cancers expected from maximal exposure for a lifetime.

Comparison values are estimated contaminant concentrations that are unlikely to cause detectableadverse health outcomes when these concentrations occur in specific media. Comparison values areused to select site contaminants for further evaluation. They are based on health guidelines.Comparison values are calculated using conservative assumptions about daily intake rates by anindividual of standard body weight. Because of the conservatism of the assumptions and safetyfactors, contaminant concentrations that exceed comparison values for an environmental medium donot necessarily indicate a health hazard.

For nonradioactive chemicals, ATSDR uses comparison values like environmental media evaluationguides (EMEGs), cancer risk evaluation guides (CREGs), reference dose (or concentration) mediaevaluation guides (RMEGs), and others. EMEGs, since they are derived from MRLs, apply only tospecific durations of exposure. Also, they depend on the amount of a contaminant ingested orinhaled. Thus, EMEGs are determined separately for children and adults, and also separately forvarious durations of exposure. A CREG is an estimated concentration of a contaminant that wouldlikely cause, at most, one excess cancer in a million people exposed over a lifetime. CREGs arecalculated from CSFs. Reference dose (or concentration) media evaluation guides (RMEGs) aremedia guides based on EPA's RfDs and RfCs.

EPA's maximum contaminant levels (MCLs) are maximum contaminant concentrations ofchemicals allowed in public drinking water systems. MCLs are regulatory standards set as close tohealth goals as feasible and are based on treatment technologies, costs, and other factors.

For radiological contaminants, ATSDR uses information on radiation exposure and its effects prepared by federal agencies, including EPA, DOE, and the US Nuclear Regulatory Commission.The agency also uses other publicly available data sources and recommendations on radiation doselimits. The National Council on Radiation Protection and Measurements (NCRP), the InternationalCommission on Radiological Protection (ICRP), and the United Nations Scientific Committee onthe Effects of Atomic Radiation are a few of the sources.

ATSDR uses standard or site specific intake rates for inhalation of air and ingestion of water, soil,and biota. Table C-1 presents the intake rates for groundwater, surface water, soil, and sediment thatwe used in estimating doses for PGDP. (The dose calculation equations, and our assumptions aboutexposure factors, are derived from the ATSDR Public Health Assessment Guidance Manual [2].)For screening purposes, ATSDR often uses the maximum contaminant concentration detected in aspecific medium at a site to identify contaminants requiring specific exposure evaluations; using themaximum concentration results in a more protective evaluation. When unknown, the biologicalabsorption of a substance within the human body is assumed to be 100%.

After estimating the potential exposure at a site, ATSDR identifies the site's "contaminants ofconcern" by comparing the exposures of interest with health guidelines, or contaminantconcentrations with comparison values. As a general rule, if the guideline or value is exceeded,ATSDR evaluates exposure to determine whether it is of potential health concern. Sometimesadditional medical and toxicological information may indicate that these exposures are not of healthconcern. In other instances, exposures below the guidelines or values could be of health concernbecause of interactive effects with other chemicals or because of the increased sensitivity of certainindividuals. Thus additional analysis is necessary to determine whether health effects are likely tooccur.

Exposure doses via ingestion are calculated on the basis of the following equation:

Dose (Ingestion) = (Chemical Conc. x IR x EF x ED) / (BW x AT)

where:
Chemical Conc. = concentration of each contaminant (in mg/g,µg/g, mg/L, or µg/L)
IR = ingestion rate (in grams/day or liters/day)
EF = exposure frequency in days per year
ED = exposure duration in years
BW = body weight in kilograms
AT = averaging time in days

For soil and sediment doses, we take an additional step to determine exposure via dermal absorption, with the total dose being the sum of the ingestion dose and the dermal dose.

Dose (Dermal) = (Chemical Conc. x ABS x TSA x EF x ED) / (BW x AT)

where all factors are as above except:
ABS = a chemical-specific absorption or bioavailability factor (unitless)
TSA = total soil adhered in milligrams (skin surface area x soil adherence value)

Once we have calculated the dose (in mg/kg/day) for a contaminant, we evaluate that contaminant's non-cancer and cancer health effects. For the former, we compare the dose with studies that have investigated the health effects of exposure to the contaminant. For the latter, we multiply the dose by the pathway-specific CSFs which are expressed in units of inverse dose--that is, (mg/kg/day)-1.

Excess Cancer Risk = Dose (mg/kg/day) x Cancer Slope Factor (mg/kg/day)-1

The excess cancer risk is the expected increase in cancer risk due to contaminant exposure. All of the uncertainties and health-protective exposure assumptions associated with the dose calcuations are included in the risk estimation, as well as the uncertainty in deriving the CSF. Excess cancer risks are described by the following categories [3]:

No increased risk less than 1 per 100,000 < 0.00001
No apparent increased risk 1 per 100,000 0.00001
Low increased risk 1 per 10,000 0.0001
Moderate increased risk 1 per 1,000 0.001
High increased risk 1 per 100 0.01
Very high increased risk more than 1 per 100 > 0.01

None of the excess cancer risk estimates necessarily indicate that exposure to carcinogenic contaminants will result in cancer in the exposed population.

References

  1. US Environmental Protection Agency. Integrated Risk Information System, 2000. http://www.epa.gov/iris/index.html

  2. Agency for Toxic Substances and Disease Registry. Public Health Assessment GuidanceManual. Atlanta: US Department of Health and Human Services; 1992.

  3. Agency for Toxic Substances and Disease Registry. Public Health Decision StatementTOX.14. Draft QAA-27. Atlanta (GA): US Department of Health and Human Services;1991 Oct 21.

Table C-1.

Dose equations and factors used in calculating exposure doses at PGDP
Dose Parameters Groundwater Surface Water Soil Sediment
Ingestion rate
WKWMA workers
Adults
Children
Pica children
2 liters/day
2 liters/day
1 liter/day
1 liter/day
0.5 liters/day
0.5 liters/day
0.5 liters/day
None
200 mg/day
50 mg/day
200 mg/day
2,000 mg/day
100 mg/day
None
100 mg/day
None
Total soil adhered (for dermal contact)
WKWMA workers
Adults
Children
Pica children
NA
NA
NA
NA
NA
NA
NA
NA
37,600 mg
9,400 mg
5,250 mg
3,000 mg
37,600 mg
9,400 mg
5,250 mg
3,000 mg
Exposure frequency
WKWMA workers
Residents (adults and children)
All groundwater exposures based on residential scenario 12 days/year
12 days/year
1.5 days/week
5.6 days/week
0.75 days/week
12 days/year
Exposure area
(location of stations used to determine 67th percentile concentration)
All groundwater exposures based on specific well data All surface-water stations outside of security fence WKWMA workers: buffer zone stations

Residents: stations outside buffer zone

All sediment stations outside security fence
Exposure duration ~14 years
(1974-1988)
30 years (adult)
6 years (child)
30 years (adult)
6 years (child)
3 years (pica child)
30 years (adult)
6 years (child)
Body weight
Adults
Children
Pica children
70 kg
13 kg
10 kg

Averaging time
Non-cancer (exposure duration x 365 days)


Cancer


Adult: 30 years x 365 days/year
Child: 6 years x 365 days/year
Pica child: 3 years x 365 days/year

70 years x 365 days/year

This table does not include information for the food and biota pathway; see the food and biota section of the public health assessment.
Key: kg = kilograms; mg = milligrams; mg/day = milligrams per day; WKWMA = Western Kentucky Wildlife Management Area


APPENDIX D: ESTIMATION OF EXPOSURE DURATION FPR GROUNDWATER PATHWAY

Four residences near the northwest boundary of PGDP were exposed to trichloroethylene (TCE) andtechnetium 99, and possibly to lead, pentachlorophenol, and vinyl chloride via contaminatedgroundwater. The exposure occurred via ingestion of and dermal contact with groundwater, andinhalation of vapors from contaminated groundwater. Residents were provided with an alternatewater source upon discovery of the contaminants in August 1988.

Very little groundwater monitoring took place before 1988, so monitoring data cannot be used todetermine the duration of contaminant exposure. The rate of contaminant transport after 1988 hasbeen used to estimate the annual rate of contaminant migration. The locations of the 100-microgram-per-liter TCE isocontours were qualitatively interpreted from monitoring data for 1988,1991, and 1995. (This concentration was chosen not for health reasons but for better reliability inthe data.) These contours were interpolated using maximum annual concentrations from residentialand monitoring wells. The contouring procedure locates the line of equal concentration (100 g/L)based on point values and the distances between adjacent values.

Figure D-1 shows isocontours for 1991 and 1995, which were drawn using ArcView overlaid on asite map. ArcView's map measurement tool was used to measure the plume progressions from 1988to 1991 and from 1991 to 1995. These distance measurements divided by the number of years ofplume progression (3 years and 4 years, respectively) yield a plume progression rate between 125and 330 meters per year, depending on the time interval (1988-1991, 330 meters per year;1988-1995, 207 meters per year; 1991-1995, 125 meters per year); see Table D-1.

Table D-1.

Estimated plume migration rates based on plume locations for different time periods
Time Period Plume Progression Annual Migration Rate
1988-1991 (3 years)960 meters330 meters/year
1988-1995 (7 years)1,450 meters207 meters/year
1991-1995 (4 years)500 meters125 meters/year

The largest uncertainty associated with estimating exposure duration is in interpreting the TCEisocontours. The 1988 contour is based on 21 annual data points (maximums at each well), whichare irregularly distributed. Because data are limited, the resulting isocontour is a conservativeestimate. The 1988 isocontour is approximately 1,200 meters (0.75 miles) downgradient of theresidential well closest to the site boundary. The small number of data points used to interpret theplume location suggests that it had progressed at least that far but probably further. Also, the timebetween measured TCE concentrations was rounded to annual values. The data used to generate thecontours are annual maximums, which occurred at approximately the same time each year.

Site personnel estimate plume migration at about 1 foot (30 centimeters) per day, which adds up to110 meters per year. (This information came from a January 22, 1998, communication with BradMontgomery of Bechtel Jacobs Company and a February 2, 1998, communication with Ross Millerof Geo Consultants, LLC.) This estimate is based on extensive flow modeling and the measuredmigration rate of tracers injected into the Regional Gravel Aquifer (RGA). Given theabovementioned uncertainty about the 1988 isocontour's location, we used a contaminant migrationrate of 110 to 125 meters per year to evaluate the duration of contaminant exposure.

Table D-2 indicates the distance that the contaminant plumes have migrated beyond the affectedresidential wells. The 1991 and 1995 isocontours show that the plume has moved 2,200 meters (asof 1991) and 2,640 meters (as of 1995) downgradient of the wells. Dividing these distances by theannual migration rate of 125 meters per year provides an estimate of the total duration of plumemigration. Subtracting the years of post-1988 migration from this total provides an estimate of thepre-1988 exposure duration. Using the 1991 and 1995 plume locations, the estimated pre-1988exposure durations are 14.6 and 14.1 years (Table D-2). This estimate is for those wells closest tothe site boundary (RW-002 and RW-113). Exposures for wells further downgradient would be ofshorter duration. Also, this exposure duration is for TCE concentrations greater than 100 g/L.Exposures at lower concentrations probably had a longer duration.

Table D-2.

Estimated exposure durations, based on 125 meters per year migration rate and distance of plume migration downgradient of residential wells
Plume Distance Beyond Residential Wells Years Migration Past 1988 Exposure Estimated Duration of Exposure (using 125 meters/year migration rate)
1988: 1,200 meters0 years9.6 years
1991: 2,200 meters3 years17.6 years - 3 years = 14.6 years
1995: 2,640 meters7 years21.1 years - 7 years = 14.1 years

TCE concentrations in the affected wells probably varied considerably over the exposure period.While there are no data for this period, well concentrations in the years after 1988 indicatesignificant variation in concentrations. (See Figure 4 in the main body of this report.) Thesevariations are probably due to the changes of seasonal water levels in the Ohio River (i.e., riverstages). Changes in the river stage directly affect both flow rate and direction in the RGA. In the caseof well RW-017, high TCE concentrations correspond with times of lower river stages. Although weused maximum annual concentrations to calculate exposure doses (Table 5 in the main body of thisreport), ingested concentrations probably varied by a factor of two.

TCE Isocontours (1991-95) and Contaminated Off-Site Well Locations
Figure D-1. TCE Isocontours (1991-95) and Contaminated Off-Site Well Locations


APPENDIX E: EXPOSURE TO AIRBORNE RADIONUCLIDES

Exposure doses to airborne radionuclides were estimated using the Clean Air Act AssessmentPackage--1988 (CAP88), a system developed by EPA [1,2]. CAP88 uses a modified Gaussianplume equation to estimate the average dispersion of radionuclides released from up to six sources.The sources can be either elevated stacks, such as a smokestack, or uniform area sources, such as apile of uranium mill tailings. Plume rise can be calculated assuming either a momentum orbuoyancy-driven plume. Assessments are done for a circular grid of distances and directions with aradius of 80 kilometers (50 miles) around the facility.

The program computes radionuclide concentrations in air, rates of deposition on ground surfaces,concentrations in food and intake rates to people ingesting food produced in the assessment area.Estimates of radionuclide concentrations in produce, leafy vegetables, milk, and meat consumed byhumans are made by coupling the output of atmospheric transport models with the U.S. NuclearRegulatory Commission Regulatory Guide 1.109 terrestrial food chain models.

Dose and risk estimates from CAP88 are applicable only to low-level chronic exposures, since thehealth effects and dosimetric data are based on low-level radionuclide intakes. The populationestimates used in this evaluation are the 1980 Census data provided with the CAP88 model. Inaddition to population estimates, the model requires information on radionuclide emission rates,meteorological data, and agricultural data on consumption of locally grown food and dairy products.Radionuclide emission data were obtained from the annual site environmental monitoring reports.

The two meteorological data sets that were used in the evaluations are provided with the CAPP88model. The 1950s emission years used 1960-1964 meteorological data; the 1996 emission yearevaluation used 1989-1993 meteorological data. Agricultural input data, stack parameters, andsource partitioning were based on information provided in the 1996 National Emission Standardsfor Hazardous Air Pollutants (NESHAP) report [3]. Four sources account for most PGDPradionuclide emissions: the C-310 stack, C-400 combined sources, the seal and wet air exhausts,and the C-710 laboratory. CAP88 places all sources at the center of the facility with respect to thesurrounding population and varies only the height of the release. This evaluation used a zero plumerise factor based on emission temperature and velocity information in the NESHAP report.

The radionuclides evaluated include technetium 99, uranium 234, uranium 235, and uranium 238.The results reported in Table E-1 are for 1956 through 1959 (the years with the largest releases) andfor 1996 (a recent year for which there is complete information). The uranium isotope releases werepartitioned between the sources in the following proportions for all years:

  • C-310 stack: 6%

  • C-400 group: 8%

  • Seal/wet air exhaust: 65%

  • C-710 laboratory: 21%

Although these proportions may have changed with process and control operations, any variations inthe sources had minimal effect on the estimated dispersion concentrations, because CAP88 locatesall emissions at the same geographic point and because a zero plume rise was used.

Table E-1.

Annual radionuclide emissions for selected isotopes and years [4,5,6,7]
Year Technetium 99
in curies (gigabecquerels)
Uranium 234
in curies (gigabecquerels)
Uranium 235
in curies (gigabecquerels)
Uranium 238
in curies (gigabecquerels)
19562.6 (96.2)1.62 (59.94)0.08 (2.96)3.50 (129.5)
19574.8 (177.6)1.10 (40.7)0.05 (1.85)1.20 (44.4)
19586.3 (233.1)1.09 (40.33)0.05 (1.85)1.16 (42.92)
19595.1 (188.7)0.93 (34.41)0.04 (1.48)1.10 (40.7)
19960.04 (1.48)0.003 (0.111)0.0001 (0.004)0.001 (0.037

In addition to the chronic or long-term process releases, accidental releases of UF6 have occurredthroughout the operating history of the PGDP facility [8,9]. The largest reported accidental releaseoccurred in 1960, when a cylinder ruptured releasing about 11,000 pounds (approximately 5,000kilograms) of UF6 . This accident occurred in Building C-333 on November 17, 1960, at about 4:00a.m. Another accidental release occurred during a fire at Building C-337 in December 1962. About5,062 pounds (2,278 kilograms) of UF6 were released during the fire.

Acute airborne uranium hexafluoride (UF6) concentrations near PGDP from the 1960 and 1962accidents were estimated using the RASCAL 3.0 air dispersion and dose model [10] and weatherobservations from the Paducah/Barkley Airport [11]. The RASCAL model (beta test version)provides a general assessment of potential uranium air concentrations following accidental releases.Due to the confluence of the water vapor from PGDP cooling towers with any airborne releases,atmospheric humidity is assumed to be similar to conditions of light precipitation.

Our data on weather conditions at the time of the 1960 release indicate a stable to very stableatmosphere (stability class F), very low wind speed from the northwest, and a temperature of 39oF(dry bulb) [11]. Under these release conditions and according to our modeling of this accident, anestimated uranium inhaled radiation dose of 1.5 rem (0.015 sieverts) and an estimated uraniuminhaled chemical dose of 20 milligrams (mg) could have been received by the maximally exposedresident southeast of the site. The U.S. Nuclear Regulatory Commission's action level for intake ofsoluble uranium is 10 mg. (At this action level, residents may be instructed to evacuate or to stayindoors with windows closed.) A report assessing PGDP accidents [10] indicates that a 5-mguranium dose can produce detectable, non-permanent kidney damage. The 1960 cylinder rupturecould have resulted in inhaled exposure doses of 5 mg to 20 mg to people who lived approximately2.5 miles (4 kilometers) from the release site. That includes off-site areas to the southeast of the site.

According to accident records, this release occurred on November 17, 1960, at approximately 4:00a.m. At that time of day and year, it is unlikely that nearby residents would be outside, whereexposure to the maximum concentrations would occur. Air temperatures were in the 30s, sowindows and doors would have been shut--very little exposure to residents inside their housesprobably occurred. Additionally, this exposure scenario assumes that 62% of the UF6 cylindercontent was vented from the building over a 1-hour period and became airborne. Notes fromaccident summaries suggest that a considerable portion of the UF6 remained in the liquid phase andwas recovered [9].

Estimated uranium air concentrations and doses from the 1962 fire are much lower than from the1960 cylinder accident. The explosion and fire that caused this release resulted in much greateratmospheric dispersion and much lower air concentrations and doses. Off-site uranium airconcentrations from this accident probably did not present a health hazard to the surroundingcommunity.

In addition to the documented 1960 and 1962 accidents, the community had concerns about twoother potential incidents: a 3-day UF6 release on March 15 through 17, 1970; and a large accidental release sometime in 1969 or 1970. A Union Carbide memorandum contained reference toa 3-day UF6 (March 15-17, 1970) that was detected via on-site air monitoring inside the building[8]. This memorandum also indicated that the average gross alpha air monitoring results for theperimeter east location from October 1969 to May of 1970 were higher than normal. However, theindividual weekly air monitoring results indicated that this average was elevated for a different timeperiod than March 15-17, 1970 [12]. Also, the plant's original report for this incident indicated thata total of 15 grams of uranium was released inside the building and eventually released through thebuilding ventilation system. This amount of uranium would not have an adverse impact off site. Forthe second concern (an accident that occurred in 1969 or 1970 when houses to the southeast turnedblack and trees died), the site accident records for the 1969/70 time frame do not report any eventscapable of producing significant off-site uranium or hydrogen fluoride concentrations.

However, an extensive review of the weekly air monitoring data indicates that there were severalperiods of elevated gross alpha and gross beta (presumably, uranium and technetium 99)concentrations at perimeter air monitors during the 1969 and 1970 timeframe [12]. There is someindication that the site investigated elevated gross beta levels to the north of the plant during thistime, but there was no explanation of the cause. Due to the limited information available on thesespecific events, ATSDR cannot evaluate potential exposure doses off site. However, monitoring datado indicate that some type of release event(s) occurred that are not reflected in the accident reportsreviewed.

At this time, it is not possible to determine if nearby residents were actually exposed to hazardousconcentrations of uranium from any of these accidental releases. However, this analysis doesestimate that potentially hazardous releases have occurred and that rupture of a UF6 cylinderrepresents potentially hazardous conditions for residents living adjacent to PGDP. In addition, theair dispersion models suggest that significant concentrations of uranium may have been deposited inoff-site areas. Currently, we have no reports of health effects related to these accidents; however, ifdata become available suggesting that health effects did occur, we will re-evaluate the need forfollowup activities.

References

  1. US Environmental Protection Agency. AIRDOS-EPA: A Computerized Methodology forEstimating Environmental Concentrations and Dose to Man From Airborne Releases ofRadionuclides. Washington (DC): US Environmental Protection Agency; 1979 Dec.Document No. EPA 520-1-70-009.

  2. US Environmental Protection Agency. User's Guide for CAP88-PC, Version 1.0.Washington (DC): US Environmental Protection Agency; 1992 Mar. Document No. EPA402-B-92-001.

  3. US Enrichment Corporation. United States Department of Energy Air Emissions AnnualReport (40 CFR 61, Subpart H), Calendar Year 1996, Paducah Gaseous Diffusion Plant.Paducah (KY): US Enrichment Corporation; 1997 May 23.

  4. US Department of Energy. Historical Radionuclide Releases From Current DOE Oak RidgeOperations Office Facilities. Oak Ridge (TN): US Department of Energy; 1988 May.Document No. 707576.

  5. Baker RC, Brown EG. Environmental Monitoring Summary for the Paducah Plant for 1958.Paducah (KY): US Atomic Energy Commission; 1959 May 22. Document No. KY-273.

  6. Brown, EG, Mitchell, KK. Environmental Monitoring Summary for the Paducah Plant for1959. Paducah (KY): US Atomic Energy Commission; 1960 May 31. Document No. KY-332.

  7. Lockheed Martin Energy Systems, Inc. Paducah Site Annual Environmental Report for1996. Kevil (KY): US Department of Energy; 1997 Dec. Document No. KY/EM-206.

  8. Letter from RF Smith, Union Carbide Nuclear Division, to VG Katzel. Subject: airborneuranium contamination. June 5, 1970.

  9. Mayo T. Draft UF6 Releases at Cylinder Handling Facilities. Paducah (KY): Union CarbideNuclear Division; Date Redacted. Document No. KY-L-863 (draft).

  10. US Nuclear Regulatory Commission. RASCAL 3.0 Beta 2, Rev. 08-18-2000. [Note: thisversion is for review and testing only, not for operational use.] Washington (DC): USNuclear Regulatory Commission; 2000.

  11. National Climatic Data Center. Surface Weather Observations for Paducah/Barkley Airport,November 17, 1960. Asheville (NC): US Department of Commerce; 1960.

  12. Unsigned. Paducah Gaseous Diffusion Plant Environmental Monitoring Worksheets -Environmental Air Sampling (1969-1974).

APPENDIX F: EXPOSURE TO AIRBORNE HYDROGEN FLOURIDE

During the uranium enrichment processes at PGDP, uranium hexafluoride (UF6) is released into theair. The UF6 reacts rapidly with water in the air to form particulate uranium and fluorides, and alsohydrogen fluoride gas (HF) [1]. HF is the most abundant form of atmospheric fluoride and reactswith atmospheric water to form hydrofluoric acid aerosols. Airborne particulate fluorides have lowsolubility and are removed from the atmosphere through dry and wet deposition.

Releases of UF6 (with atmospheric conversion to HF) occurred both as long-term releases due toprocess operations and as short-term releases due to accidents. Long-term (chronic) exposure to HFis evaluated based on correlation of annual UF6 releases with measured site perimeter HFconcentrations. Short-term (acute) HF exposures are evaluated using accident records and airdispersion modeling.

Estimated uranium releases and ambient air monitoring results have been reported consistentlythroughout PGDP's operational history; fluoride releases and HF ambient air concentrations havenot. Evaluation of potential HF exposures to nearby residents presents several problems: noreporting of HF release quantities or ambient air monitoring during the period of highest potentialfluoride and HF emissions (1956), changes in sampling locations, and changes in the data reported(e.g., annual medians vs. means). Consequently, evaluation of chronic HF exposures during theperiod of highest potential emissions requires estimation or modeling of HF emissions from periodsof consistent data reporting.

Uranium emissions are a good proxy for prediction of chronic or long-term HF ambient airconcentrations. Ambient airborne HF concentrations were measured at several locations for theyears 1961 to 1970. Comparison of mean ambient airborne HF concentrations from these locationswith uranium emission estimates for the same years provides a correlation coefficient of 0.8863,which indicates a strong positive relationship between uranium emissions and measured HFconcentrations at the perimeter north monitoring site (HF concentrations increase proportionatelywith increases in uranium emissions). That relationship is plotted in Figure F-1.

The strong correlation of uranium emissions and HF concentrations at the perimeter north station inthe years where both data sets are available allows for the prediction of HF concentrations fromuranium emission data in the years for which no HF monitoring data are available. These HFconcentrations are predicted using the linear regression forecasting function in the computerprogram EXCEL (version 7.0a). Figure F-2 shows the relationships between uranium emissions (incuries per year), estimated and monitored HF concentrations at the perimeter north and one milenorth stations, and measured fluoride concentrations in grass samples near the perimeter northstation.

The perimeter north station consistently had the highest concentrations of both particulate uraniumand HF. The perimeter north station was closer to the fluoride processing facility than other stations[2], and was downwind of the processing facility with respect to the prevailing south-southwestwinds [3]. Therefore, it was assumed that this station would have had the highest concentrations ofHF during the year of highest release.

All of the measured parameters show a strong relationship to uranium emissions and to the estimatedHF concentrations. Figure F-2 shows measured and estimated HF ambient concentrations in relationto the Kentucky ambient air standard for average annual HF exposure (500 parts per billion, or ppb)[4] and the ATSDR provisional long-term guidance value of 12 ppb. None of the measured orestimated HF concentrations exceed the Kentucky ambient air standard.

ATSDR has established a provisional guidance value of 10 micrograms per cubic meter (12 ppb) forannual average air concentrations of HF [5,6]. HF concentrations below 12 ppb (annual averagevalue) are not likely to cause adverse health effects. This guidance value is more than 100 timeslower than an exposure concentration that caused mild irritation to the eyes and noses of humanvolunteers exposed for 10 days [1]. None of the measured or estimated HF concentrations at the onenorth sampling station exceeded this guidance value (Figure F-2).

Some of the estimated HF concentrations at the perimeter north station did exceed the ATSDRguidance value; the maximum value was 28 ppb (Figure F-2). The maximum annual HF emissionoccurred in 1956, which is the period of maximum uranium emissions. Because HF concentrationsat the perimeter north station are consistently higher than at other locations, this station represents aworst-case exposure scenario. It is important to point out that no off-site residents live at theperimeter security fence. The nearest houses are closer to the one mile north and east stations than tothe perimeter stations. Consequently, the concentrations at the nearest house would have been closerto the concentrations estimated by the one north station (Figure F-2) than to the concentrations at theperimeter north station. The estimated annual average HF concentrations at these points of exposureare below levels of health concern.

There is some uncertainty associated with deriving HF concentrations from uranium emissions. Onemeasure of this uncertainty is the standard error, which is represented by error bars on the predictedHF concentrations in Figure F-2. The error bars, which show the predicted maximum and minimumHF values, do not significantly change the predicted HF concentrations with respect to the ATSDRand Kentucky health guidance values. Note that the largest standard errors occur between 1965 and1968, the period with the highest variability and lowest uranium emissions.

In addition to the chronic or long-term process releases, accidental releases of UF6 and HF(estimated from reported UF6 releases) have occurred throughout the operating history of the PGDPfacility [7,8]. The largest reported accidental release occurred in 1960, when a cylinder rupturedreleasing about 11,000 pounds (approximately 5,000 kilograms) of UF6. This accident occurred inBuilding C-333 on November 17, 1960, at about 4:00 a.m. Another accidental UF6 release occurredduring a fire at Building C-337 in December 1962. About 5,062 pounds (2,278 kilograms) of UF6were released during this fire. Many other smaller releases have occurred, but these were at least anorder of magnitude smaller than the 1960 release and less than 30% of the size of the 1962 release.

Airborne UF6 and HF concentrations surrounding PGDP from the 1960 and 1962 accidents wereestimated using the RASCAL 3.0 air dispersion and dose model [9] and weather observations fromthe Paducah/Barkley Airport. The RASCAL model (beta test version) is used to provide a generalassessment of potential HF and uranium air concentrations following accidental releases. Due to theconfluence of the water vapor from PGDP cooling towers with any airborne releases, atmospherichumidity is assumed to be similar to conditions of light precipitation.

Our data on weather conditions at the time of the 1960 release indicate a stable to very stableatmosphere (stability class F), very low wind speed from the northwest, and a temperature of 39ºF (dry bulb) [10]. Under these release conditions, short-term hazardous HF concentrations (6 parts permillion, or ppm; 15-minute Short Term Exposure Limit) could have extended more than 1 kilometer(0.6 miles) from the release site (Building C-333) toward the southeast. This means that theestimated HF concentrations could have been at hazardous levels immediately off site. Estimatedconcentrations of more than 30 ppm, which is considered immediately dangerous to life/health,extended more than 500 meters (1,640 feet) from Building C-333 and would not have reached theoff-site community. Table F-1 summarizes the air dispersion analysis.

According to accident records, the 1960 release occurred on November 17, 1960, at approximately4:00 a.m. At that time of day and year, it is unlikely that nearby residents would be outside, whereexposure to the maximum concentrations would occur. Air temperatures were in the 30s, sowindows and doors would be shut--very little exposure to residents inside their houses probablyoccurred.

The explosion and fire that caused the 1962 release resulted in much greater atmospheric dispersionand much lower air concentrations and doses. This release was modeled using a fire scenario inRASCAL 3.0, which did not analyze HF dispersion. However, HF in a fire is atmosphericallyunstable and very unlikely to undergo significant atmospheric dispersion. Off-site HF airconcentrations from this accident probably did not present a health hazard to the surroundingcommunity.

In addition to the documented 1960 and 1962 accidents, there were community concerns about twoother potential incidents: a 3-day UF6 release on March 15 through 17, 1970; and a large accidental release sometime in 1969 or 1970. A Union Carbide memorandum contained reference toa 3-day UF6 (March 15-17, 1970) that was detected via on-site air monitoring inside one of thebuildings [7]. This memorandum also indicated that the average gross alpha air monitoring resultsfor the perimeter east location for the period from October 1969 to May of 1970 were higher thannormal. However, the individual weekly gross alpha air monitoring results indicated that thisaverage was elevated for a different time period than March 15-17, 1970 [11]. The airborne fluorideresults from the perimeter east location for March 1970 were not elevated. Also, the plant's originalreport for this incident indicated that a total of 15 grams of uranium was released inside the buildingand eventually released through the building ventilation system. This amount of uraniumhexafluoride would not have an adverse impact off site. For the second concern (an accident thatoccurred in 1969 or 1970 when houses to the southeast turned black and trees died), the siteaccident records for the 1969/70 time frame do not report any events capable of producingsignificant off-site uranium or hydrogen fluoride concentrations.

However, an extensive review of the weekly air monitoring data indicate that there were severalperiods of elevated fluoride concentrations at perimeter air monitors during the 1969 and 1970timeframe [11]. There is some indication that the site investigated elevated hydrogen fluoride levelsto the east of the plant later in 1970, but there was no explanation of the cause. Due to the limitedinformation available on these specific events, ATSDR cannot evaluate potential exposure doses offsite. However, monitoring data do indicate that some type of release event(s) occurred that are notreflected in the accident reports reviewed.

At this time, it is not possible to determine if nearby residents were actually exposed to hazardousconcentrations of uranium from any of these accidental releases. However, this analysis doesindicate that potentially hazardous releases have occurred and that rupture of a UF6 cylinderrepresents potentially hazardous conditions for residents living adjacent to PGDP. Currently, wehave no reports of health effects related to the reported accidents; however, if data become availablesuggesting that health effects did occur, we will re-evaluate the need for followup activities.

Table F-1.

Summary of RASCAL 3.0 model for assessing accidental releases of HF from November 17, 1960, UF6 cylinder rupture
SITE DATA INFORMATION:
  Location: PADUCAH, KENTUCKY
  Building C-333
  Time: November 17, 1960 0400 hours CST (user specified)
CHEMICAL INFORMATION:
  Chemical Name: HYDROGEN FLUORIDE
  Molecular Weight: 20.01 kg/kmol
  IDLH: 30 ppm
  Footprint Level of Concern: 6 ppm
  Boiling Point: 19.52ºC
  Vapor Pressure at Ambient Temperature: 0.56 atm
  Ambient Saturation Concentration: 568, 348 ppm or 56.8%
ATMOSPHERIC INFORMATION: (MANUAL INPUT OF DATA)
  Wind: 1.95 knots from 303º true at 10 meters
  No Inversion Height
  Stability Class: F Air Temperature: 39º F
  Relative Humidity: 100%
  Ground Roughness: urban or forest
  Cloud Cover: 0 tenths
SOURCE STRENGTH INFORMATION:
  Direct Source: 17800 pounds, 62% released from building
  Source Height: 83 feet
  Release Duration: 65 minutes
  Release Rate: 2.9 pounds/sec
  Total Amount Released: 11,036 pounds
FOOTPRINT INFORMATION:
  Dispersion Module: Gaussian
  User-Specified LOC: 6 ppm
  Max Threat Zone for LOC: 1.5 kilometers
  Max Threat Zone for IDLH: 0.5 kilometers

References

  1. Agency for Toxic Substances and Disease Registry. Toxicological Profile for Fluorides,Hydrogen Fluoride, and Fluorine. Atlanta (GA): US Department of Health and HumanServices; 1993.

  2. Baker RC. Environmental Monitoring Summary for the Paducah Plant for 1962 and 1963.Paducah (KY): US Atomic Energy Commission; 1964 Jul. Document No. KY-458.

  3. Lockheed Martin Energy Systems, Inc. Paducah Site Annual Environmental Report for1996. Kevil (KY): US Department of Energy; 1997 Dec. Document No. KY/EM-206.

  4. Commonwealth of Kentucky. Ambient Air Quality Standards, 401 KAR 53:010 (1988).

  5. Agency for Toxic Substances and Disease Registry. ATSDR Record of Activity forTelephone Communication With S Chou, ATSDR Division of Toxicology. Atlanta,Georgia. October 21, 1993.

  6. Agency for Toxic Substances and Disease Registry. Health Assessment for US DOEPortsmouth Gaseous Diffusion Plant, Piketon County, Ohio. Atlanta (GA): US Departmentof Health and Human Services; 1996.

  7. Letter from RF Smith, Union Carbide Nuclear Division, to VG Katzel. Subject: airborneuranium contamination. June 5, 1970.

  8. Mayo T. Draft UF6 Releases at Cylinder Handling Facilities. Paducah (KY): Union CarbideNuclear Division; Date Redacted. Document No. KY-L-863 (draft).

  9. US Nuclear Regulatory Commission. RASCAL 3.0 Beta 2, Rev. 08-18-2000. [Note: thisversion is for review and testing only, not for operational use.] Washington (DC): USNuclear Regulatory Commission; 2000.

  10. National Climatic Data Center. Surface Weather Observations for Paducah/Barkley Airport,November 17, 160. Ashville (NC): US Department of Commerce; 1960.

  11. Unsigned. Paducah Gaseous Diffusion Plant Environmental Monitoring Worksheets -Environmental Air Sampling (1969-1974).

Correlation coefficient of PGDP uranium emissions and hydrogen fluoride measurements at the perimeter north station
Figure F-1. Correlation coefficient of PGDP uranium emissions and hydrogen fluoride measurements at the perimeter north station

Measured and predicted HF concentrations at the perimeter north and one mile north sampling stations
Figure F-2. Measured and predicted HF concentrations at the perimeter north and one mile north sampling stations


APPENDIX G: AIRBORNE TRICHLOROETHYLENE DISTRIBUTION AND POTENTIAL OFF-SITE EXPOSURE

Past operations at PGDP involved large quantities of trichloroethylene (TCE) as an organic solventand degreaser. Although significant amounts of TCE were released into the groundwater system,most TCE from operational processes volatilized into the atmosphere [1]. To determine if thoseairborne releases present a potential for inhalation exposure to nearby residents, we conducted aTCE air dispersion analysis using the Industrial Source Complex (ISC3) model [2].

The ISC3 model uses meteorological data to generate air concentration averages, for periods from 1hour to 1 year, for any location surrounding an air emission source. ATSDR's analysis used 1989meteorological data from the nearby Paducah Municipal Airport (Barkley Airfield). According todata available to us, the largest annual release of TCE to the atmosphere (62,826 kilograms, or138,845 pounds) occurred in 1986 [3]. The dispersion from this release was modeled as a singlesource from Building C-400--a vent 5 meters (16 feet) off the ground with a diameter of 25centimeters (10 inches)-- and the annual release proportioned over the entire year.

For this analysis, we assumed conservative dispersion with no chemical degradation orphotochemical breakdown of TCE. (Typically, TCE is estimated to degrade in the atmosphere witha chemical half-life of 3 to 7 days [4].) Under these conservative assumptions, the maximumairborne TCE concentration is 112 micrograms per cubic meter (g/m3) for a 1-hour averagingperiod, and 3 g/m3 for a 1-year averaging period, at a location 1 kilometer (0.6 miles) north ofBuilding C-400 (that is, off site). Some animal studies have shown carcinogenic effects from TCE;however, ATSDR and EPA are re-evaluating TCE's carcinogenic effects on humans. Until TCE'scarcinogenicity for humans is determined, minimal risk levels (MRLs) for non-cancerous effects areused to screen for contaminants of concern. The MRL for TCE is 10,920 g/m3 for acute exposures(1 to 14 days) and 546 g/m3 for intermediate exposures (15 to 365 days) [4,5]. The estimated TCEair concentrations during the highest TCE release year are two orders of magnitude lower than theMRLs and below levels of health concern. Consequently, we did not choose TCE as a contaminantof concern for airborne releases at PGDP.

References

  1. Martin Marietta Energy Systems, Inc. Paducah Gaseous Diffusion Plant EnvironmentalReport for 1992. Paducah (KY): US Department of Energy; 1993 Sep. Document No.ES/ESH-36, KY/E-164.

  2. Trinity Consultants, Inc. Breeze Air Suite Industrial Source Complex (ISC3) DispersionModels Software Package and Users Guide, Version 1.07. Dallas (TX): Trinity Consultants,Inc.; 1996.

  3. Martin Marietta Energy Systems, Inc. Environmental Surveillance of the U.S. Department ofEnergy Paducah Reservation and Surrounding Environs During 1986. Paducah (KY): USDepartment of Energy; 1987 Apr. Document No. ES/ESH-1/V3.

  4. Agency for Toxic Substances and Disease Registry. Toxicological Profile forTrichloroethylene. Atlanta (GA): US Department of Health and Human Services; 1997 Sep.

  5. Agency for Toxic Substances and Disease Registry. Public Health Assessment GuidanceManual. Atlanta: US Department of Health and Human Services; 1992.

APPENDIX H: AIRBORNE HEXAVALENT CHROMIUM DISTRIBUTION FROM THE PGDP WATER COOLING TOWERS AND POTENTIAL OFF-SITE EXPOSURE

Isotopic diffusion operations at PGDP generate excess heat, which is released to the environmentthrough four cooling systems. In these systems, heat exchangers transfer heat to cooling waters,which in turn release the heat to the atmosphere through 14 water cooling towers (located in fourdiscrete areas). Until 1993, a chromium solution was added to the cooling waters to preventcorrosion [1]. This caused hexavalent chromium to be released to the atmosphere at the watercooling towers.

Although annual chromium emissions have been calculated based on the quantities of chromiumcompounds added to the cooling system, the airborne chromium concentration has never beenmeasured at on-site or off-site locations. DOE has measured and modeled chromium deposition insurrounding soils and plants, and found that chromium concentrations are at background levels forlocations more than 1,500 meters from the cooling towers [2,3]. However, the security fence to theeast and north of the easternmost cooling systems is less than 500 meters from the towers, and theclosest residence is about 1,000 meters from the towers.

Because inhalation of hexavalent chromium can be toxic, we estimated potential exposures toairborne concentrations of chromium using the ISC3 air dispersion model [4]. This model uses site-specific meteorological data (in this case, from Barkley/Paducah Municipal Airport) and a Gaussianair dispersion equation to estimate contaminant concentrations.

The 1992 chromium release of 2,015 kilograms per year (0.064 grams per second, or g/sec) wasused as the emission rate for the dispersion calculations [5]. The 1992 chromium release was thehighest annual emission on record, and thus represents the most conservative source term forevaluation of public health affects. To characterize local weather, we used the most recent completemeteorological data set (a 1990 hourly data set) from the EPA SCRAM Web site(http://www.epa.gov/ttn/scram/) for the Barkley/Paducah Municipal Airport weather station.

The chromium emissions were modeled as four sources, based on information from the study ofcooling tower drift at PGDP [2]. Relative locations of the cooling towers were derived from Figure 1of that study. Chromium concentrations at breathing height were estimated for a 5,000-meter polargrid, with potential receptors located at 500-meter intervals along 16 transects (every 22.5 degrees).

The source-specific release rates and source dimensions were based on a total annual emission rateof 0.064 g/sec, allocated between the four sources [2]. The four sources were modeled as volumesources, with release heights of 8 meters and lateral dimensions of 25 meters (towers 1 and 2) and75 meters (towers 3 and 4). Release rates were estimated as:

  • Tower 1: 0.011 g/sec

  • Tower 2: 0.011 g/sec

  • Tower 3: 0.021 g/sec

  • Tower 4: 0.021 g/sec

The model used regulatory default options and dry settling/deposition to estimate all chromiumconcentrations [6]. Chromium particles in the cooling tower drift have diameters of 5 to 50 microns.However, these particles are contained within water droplets that have diameters of 100 to 1,300microns [3]. A particle size distribution of 100 to 999 microns (with a 700-micron mean) was usedin calculating settling velocities.

Maximum chromium concentrations were calculated for each potential receptor for 1-hour, 8-hour,24-hour, and 1-year averaging periods. Because wind directions and speeds change so drasticallyover a year, these maximum concentrations represent the highest estimated concentrations for eachtime period for each location. Averages for 1-hour periods are significantly higher than the 8-hour,24-hour, or 1-year averages.

The results of this modeling indicate that dispersed hexavalent chromium air concentrations onsite and off site were lower than the health guidelines for intermediate and chronic exposures[7]. The intermediate minimal risk level (MRL) for inhalation of particulate hexavalent chromium is0.5 micrograms per cubic meter (g/m3). The intermediate and chronic MRL for inhalation ofdissolved hexavalent chromium as an aerosol is 0.1 g/m3. The highest estimated 1-hour, 8-hour,24-hour, and 1-year average air concentrations were on site, between the four cooling towers; theyare listed in Table H-1. The maximum estimated off-site concentrations were about 500 metersnorth-northeast of the cooling towers, outside the security fence, and are listed below. The closestresidence is approximately 1,000 meters east of the easternmost cooling system, and the estimatedmaximum concentrations for this location are listed below. The results of this air dispersion modelare in agreement with chromium distribution studies that found no air-dispersed chromium in soil orvegetation samples beyond 1,500 meters of the cooling towers [2,3].

Table H-1.

Maximum estimated airborne hexavalent chromium concentrations
Exposure Time Maximum On Site Maximum Off Site Maximum at Closest Residence
1-hour maximum0.0215 g/m30.005 g/m30.0011 g/m3
8-hour maximum0.0108 g/m30.0025 g/m30.0005 g/m3
24-hour maximum0.0067 g/m30.0007 g/m30.0005 g/m3
1-year maximum0.0009 g/m30.0003 g/m30.0004 g/m3

The results of this modeling study, which uses conservative assumptions for settling, dispersion, andemission rates, did not find any areas where exposure to airborne chromium exceeded healthguidelines. The distribution of these airborne concentrations is supported by measurements ofdeposited chromium in soil and vegetation samples. Therefore, airborne hexavalent chromium wasnot selected as a contaminant of concern at PGDP.

References

  1. Martin Marietta Energy Systems, Inc. Paducah Gaseous Diffusion Plant Annual SiteEnvironmental Report for 1993. Paducah (KY): US Department of Energy; 1994 Oct.Document No. ES/ESH-53, KY/ERWM-18.

  2. Taylor FG, Hanna SR, Parr PD. Cooling Tower Drift Studies at the Paducah, Kentucky,Gaseous Diffusion Plant. Oak Ridge (TN): Oak Ridge National Laboratory, EnvironmentalSciences Division; 1978. p. 32. Document No. 1275 (ORNL/TM-6131).

  3. Taylor FG Jr. Chromated Cooling Tower Drift and the Terrestrial Environment: A Review.Nuclear Safety 1980;21(4):495-508.

  4. Trinity Consultants, Inc. Breeze Air Suite Industrial Source Complex (ISC3) DispersionModels Software Package and Users Guide, Version 1.07. Dallas (TX): Trinity Consultants,Inc.; 1996.

  5. Martin Marietta Energy Systems, Inc. Paducah Gaseous Diffusion Plant EnvironmentalReport for 1992. Paducah (KY): US Department of Energy; 1993 Sep. Document No.ES/ESH-36, KY/E-164.

  6. US Environmental Protection Agency. User's Guide for the Industrial Source Complex(ISC3) Dispersion Models, Volume I. Research Triangle Park (NC): US EnvironmentalProtection Agency; 1995.

  7. Agency for Toxic Substances and Disease Registry. Toxicological Profile for Chromium.Atlanta (GA): US Department of Health and Human Services; 1998 Aug.

Next Section     Table of Contents

  
 
USA.gov: 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. #