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




Appendix A

Health Hazard Categories


This category is used for sites where short-term exposures (< 1 yr) to hazardous substances or conditions could result in adverse health effects that require rapid intervention.

This determination represents a professional judgement based on critical data which ATSDR has judged sufficient to support a decision. This does not necessarily imply that the available data are complete; in some cases additional data may be required to confirm or further support the decision made.


Evaluation of available relevant information* indicates that site-specific conditions or likely exposures have had, are having, or are likely to have in the future, an adverse impact on human health that requires immediate action or intervention. Such site-specific conditions or exposures may include the presence of serious physical or safety hazards, such as open mine shafts, poorly stored or maintained flammable/explosive substances, or medical devices which, upon rupture, could release radioactive materials.

* Such as environmental and demographic data; health outcome data; exposure data; community health concerns information; toxicologic, medical, and epidemiologic data.


This category is used for sites that pose a public health hazard due to the existence of long-term exposures (> 1 yr) to hazardous substance or conditions that could result in adverse health effects.

This determination represents a professional judgement based on critical data which ATSDR has judged sufficient to support a decision. This does not necessarily imply that the available data are complete; in some cases additional data may be required to confirm or further support the decision made.


Evaluation of available relevant information* suggests that, under site-specific conditions of exposure, long-term exposures to site-specific contaminants (including Radio nuclides) have had, are having, or are likely to have in the future, an adverse impact on human health that requires one or more public health interventions. Such site-specific exposures may include the presence of serious physical hazards, such as open mine shafts, poorly stored or maintained flammable/ explosive substances, or medical devices which, upon rupture, could release radioactive materials.

*Such as environmental and demographic data; health outcome data; exposure data; community health concerns information; toxicologic, medical, and epidemiologic data.


This category is used for sites when a professional judgement on the level of health hazard cannot be made because information critical to such a decision is lacking.


This category is used for sites in which "critical" data are insufficient with regard to extent of exposure and/or toxicologic properties at estimated exposure levels. The health assessor must determine, using professional judgement, the "criticality" of such data and the likelihood that the data can be obtained and will be obtained in a timely manner. Where some data are available, even limited data, the health assessor is encouraged to the extent possible to select other hazard categories and to support their decision with clear narrative that explains the limits of the data and the rationale for the decision.


This category is used for sites where human exposure to contaminated media may be occurring, may have occurred in the past, and/or may occur in the future, but the exposure is not expected to cause any adverse health effects.

This determination represents a professional judgement based on critical data which ATSDR considers sufficient to support a decision. This does not necessarily imply that the available data are complete, in some cases additional data may be required to confirm or further support the decision made.


Evaluation of available relevant information* indicates that, under site-specific conditions of exposure, exposures to site-specific contaminants in the past, present, or future are not likely to result in any adverse impact on human health.

*Such as environmental and demographic data; health outcome data; exposure data; community health concerns information; toxicologic, medical, and epidemiologic data; monitoring and management plans.


This category is used for sites that, because of the absence of exposure, do NOT pose a public health hazard.


Sufficient evidence indicates that no human exposures to contaminated media have occurred, none are now occurring, and none are likely to occur in the future.


Appendix B

Demographic Data


A. Comparison of key demographic variables by zip codes.
  Zip Codes
  78211 78237
Total population 38,998 38,998
Number of households 8,097 10,227
Number of persons per household 3.78 3.77
Percentage of population with Hispanic ethnicity 95.3 94.2
Percentage of persons aged <10 years 19.6 19.1
Percentage of persons aged <65 8.8 9.7
Percentage of housing units occupied by owners 69.8 66.2
Median value ($) of owner-occupied housing units 31,300 30,700

Source: 1990 Census of Population and Housing, Summary Tape File 3 (Texas).

B. County and State Demographics
  Bexar County Percentage Texas Percentage
White 74.1 75.2
Black 7.1 11.9
American Indian, Eskimo, or Aleut 0.4 0.4
Other race 1.3 1.9
Hispanic origin 17.1 10.6
Under age 10 16.8 16.4
Age 65 and older 9.9 10.1
Households occupied by owners 57.8 60.9

C. Community Housing and Age Distribution of Population
  Percentage, by Zip Code of Residence
  78211 78237
Age in years    
0-14 29.3 28.8
15-29 26.4 26.5
30-44 18.2 18.7
45-64 17.3 16.4
<65 8.8 9.7
Total population 30,610 38,998

Year structure was built    
1989-1990 1.6 0.8
1985-1988 5.0 4.3
1980-1984 6.1 4.3
1970-1979 13.9 13.5
1960-1969 22.8 26.2
1950-1959 28.2 32.3
1940-1949 15.7 14.4
Before 1940 6.9 4.2
Total number of housing units 8,936 11,013

Year resident moved into unit    
1980-1990 15.7 17.2
1985-1988 17.8 17.6
1980-1984 11.8 8.5
1970-1979 20.4 16.8
1960-1969 13.8 18.9
Before 1960 20.4 21.0
Total number of households 7,999 10,136


Appendix C

Evaluation Methodology

Evaluation Methodology

In performing a health assessment, ATSDR uses the weight-of-evidence approach, which considers the strength of all of the evidence to evaluate potential health effects (50). In some instances, ATSDR uses quantitative risk assessment to estimate cancer risk levels. These estimated risk levels are one of the factors considered by ATSDR in its professional judgment to define exposures that present a potentially significant human health hazard (51,52). ATSDR recognizes the utility of quantitative risk assessment conclusions, but such estimates must be considered in the context of the variables and assumptions used to derive them (53). Risk assessments conducted by ATSDR at hazardous wastes sites include these four steps: hazard identification, exposure assessment, dose response assessment, and risk characterization (54).

Hazardous waste sites typically contain a mixture of hazardous substances, and some combinations of these substances may be much more hazardous than any of the individual chemicals (53). ATSDR believes that no single approach is appropriate for all risk assessments of multiple chemical exposures. This is a complex issue, and there is a paucity of empirical data. In the absence of information regarding the interaction of these substances, ATSDR assumes that the effects of these combinations are additive. Such assessments should also be accompanied by a weight-of-evidence statement on the potential for interactive effects (54).

ATSDR generally know two things about health effects associated with a particular contaminant: 1) the level of exposure at which ATSDR would consider most individuals to be "safe" (comparison values like ATSDR MRLs and CREGs), and 2) the level of exposure at which adverse health effects have been reported in the scientific literature, generally in occupational workers. ATSDR's comparison values may be hundreds or thousands of times below the levels for which adverse health effects have been reported. Comparison values include a margin of safety to protect sensitive populations. Therefore, ATSDR knows a concentration for which health effects might be expected and a concentration for which health effects would not be expected. ATSDR is uncertain what may happen when individuals are exposed to concentrations that are in between these values. ATSDR performs a health assessment to determine whether health effects would be likely or unlikely at a given concentration in a site-specific exposure scenario. Other risk factors, which are unique to each individual, also determine whether an individual will actually become sick. These are risk factors such as smoking, alcohol consumption, diet, nutrition, and exposure to other chemicals at home or at work. ATSDR does not know the risk factors for each individual in a community, so ATSDR's health assessments focus on the community as a whole, realizing that the risk of an individual becoming sick depends on that individual's risk factors.

ATSDR performs health assessments in two phases. The first phase is a screening phase which may consist of several steps, and the second phase is based on professional judgment. In the first step of the screening phase, ATSDR compares health comparison values with concentrations of contaminants to which the community could be exposed. Exposures at levels less than or equal to these comparison values are not expected to make people sick and are thus considered to be "safe" levels, even under the conditions of maximum exposure. For example, comparison values for chronic-duration (long-term) exposures are based on lifelong, continuous exposure. While it is unlikely that anyone would be exposed this frequently or for this duration, it represents a worst-case condition. If a contaminant concentration does not exceed a comparison value, ATSDR does not need to evaluate the contaminant any further, because even under the maximum conditions of exposure, an individual would not be expected to get sick.

If a contaminant concentration exceeds a comparison value, but is below levels where adverse health effects have been reported, ATSDR performs a more in-depth evaluation using realistic exposure scenarios. This is a way to describe the exposure, relative to other exposures. It does not predict whether an individual will get sick. For example, if an individual were exposed to a contaminant half as much as another individual, his risk would be half as great. This risk does not mean that the individual would get sick. ATSDR does not know if either individual would get sick, but ATSDR uses this way of describing an individual's risk of getting sick from a specific exposure to a specific contaminant as compared to another individual's risk of getting sick from a different exposure to the same contaminant or to a different contaminant.

If the risk of getting sick from exposure to a contaminant is low (no increased risk or no apparent increased risk), ATSDR does not need to evaluate the contaminant any further. If the risk from exposure to a contaminant is high, ATSDR further evaluates the contaminant and the exposure during the second phase of the health assessment. ATSDR uses professional judgment based on the principles and knowledge of toxicology, physiology, biochemistry, anatomy, epidemiology, and other scientific disciplines to describe the likelihood of adverse health effects occurring in an exposed population. Risk assessment is one of the tools that a health assessor uses in considering all evidence. Because of the conservative nature of risk assessments, the health assessor is reasonably assured that no one is likely to become sick if the risk assessment does not indicate a very high level of risk.

Appendix D

Air Exposure Pathway

Air Dispersion Model

To estimate the concentrations of contaminants in the air in the communities surrounding Kelly Air Force Base, ATSDR conducted a dispersion air model. This model is a way of predicting how much of a contaminant would be present in the community by describing how other factors (such as gravity, weather, wind, and chemical reactions) would affect the contaminants as they are released from the base. Modeling is a good way of predicting how much of a contaminant would be present at any location between the release point and distant locations in the community. Modeling can estimate the concentrations 24 hours per day for years. Sampling could not realistically provide the same information in a timely fashion because it would require many samplers operating 24 hours per day for years to get the same information.

The ATSDR model used emission data from the Kelly AFB emissions data records. This data contained information on more than 1400 emission sources, many of which emitted multiple chemicals, resulting in more than 7000 separate modeling data inputs. In addition, ATSDR included emission data calculated from landings and takeoffs and "grandfathered" emission sources. Grandfathered emissions sources are those sources which were in operation when regulations were developed requiring new air quality criteria. The base was allowed to keep these sources without having to change them to meet the new air quality criteria.

Emission data, including emission rates in pounds per hour, hours of operation, and location of source, was integrated with emission stack height and meteorological data. Meteorological data consisted of the 5-year average conditions of wind velocity and direction, temperature, and humidity. The model calculates the predicted concentrations for each chemical at any point as it leaves the source and is dispersed by wind and diffusion. Therefore, ATSDR was able to estimate the 24-hour average annual concentrations and 1-hour maximum concentrations. The closest point to the fence line at the nearest residence downwind of Kelly AFB was selected as the point of reference. This point of reference represents the location of the maximum concentration to which a resident of the neighborhood would be exposed. Estimated concentrations of contaminants at this location were screened with comparison values. Risk assessments were performed for the contaminants that exceeded screening concentrations.

The air dispersion model used at Kelly AFB was developed and validated by the Environmental Protection Agency (EPA). A model is developed by measuring emissions and meteorological conditions and monitoring air quality at various points downwind of the source. The data is then mathematically treated to find a model that "fits" the various data. Models are validated by sampling to determine if the model successfully predicts the downwind concentrations for known emissions and meteorological data. The model accounts for the dispersion or spreading out of a chemical as it is carried away from the source by wind and diffusion. The model can also account for deposition due to particulate matter settling to the ground and depletion due to reactions with sunlight and other materials. As a chemical spreads out as it leaves the source, the concentration of the chemical will generally decrease as it get farther from the source.

An advantage of using the modeling approach is that short-term fluctuations in emissions and meteorological conditions are averaged out to provide a good long-term average. If sampling events were attempted, they would be subject to limitations in their ability to account for fluctuations in emissions from a number of contributors, including fluctuations in meteorological conditions. Of course, a model is subject to limitations and uncertainty as well, but ATSDR believes that it is the best tool to predict concentrations over a long period of time in the absence of long-term sampling data. The model becomes less certain when ATSDR tries to determine short-term concentrations. "The accuracy of chronic risk modeling is quite high, perhaps +/- 20% or better, because annual average concentrations are computed. The accuracy of unit risk factors--the risk of a person contracting cancer if continuously exposed for a 70-year lifetime at the site--is subject to more error because of the difficulties of relating human health effects to low concentration levels of a substance. Thus, the major source of error in chronic risk assessment lies in the risk factors, not in the modeling" (55).

General Assumptions for the Air Dispersion Model

Modeling provides ATSDR with a means to evaluate air concentrations, which are hard to measure because they are very variable over time. The means include historical information that correlates the hard-to-measure parameters with ones that can be measured effectively, like wind speed, wind direction, and source release rate.

Other parameters such as building height, source temperature, particle size, and decay rate, have an influence on the final concentration. These parameters are considered in refined modeling runs in which they are measured reliably so that their contribution is not offset by uncertainty.

Often, there are insufficient data to discern the effects of one parameter or another. So there is a balance used between the two terms or one term is used to consider the effects of both. This will be discussed further in the section on refined modeling parameters.

The latest version of the Industrial Source Complex Short Term (ISC3) Model was used to estimate the ambient air concentrations of contaminants from the operations at the Kelly Air Force Base (56). The ISC3 model is a computer model that ATSDR used to estimate present exposures to contaminants. The model includes wind, temperature, and dispersion factors that allows ATSDR to estimate the amount of a contaminant to which one could be exposed.

Following is a list of assumptions, other than default, used to set up the model for evaluating the dispersion of emissions from Kelly AFB (57).

  • Emissions reported by Kelly AFB
  • Specifications of buildings at Kelly AFB

  • Reports of surface weather observations collected hourly at San Antonio Airport for the years 1987-1991
  • Reports of upper air observations collected twice daily at Del Rio Airport for the years 1987-1991.

1. For data that was reported by Kelly Air Force Base to the Texas Natural Resources and Conservation Commission, ATSDR conducted the following activities:

  • Reviewed and compared the data against other Air Force, navy, and army bases and activities. The values for emissions were consistent with other similar operations.
  • Requested operation practices, source-specific parameters, and building characteristics from the Air Force.

  • Used information about throughputs from similar Air Force or naval operations at the facilities when data was lacking. These do not include emission rates, but do include flow rates in cubic feet per minute (cfm).

  • For those sources that did not have sufficient stack height or where flow rate data were missing, a stack height of 20 feet (ft) and a flow rate of 0.1 cfm was chosen.

2. ATSDR ran the model using data from the Kelly AFB emission inventory, consisting of 7,016 modeling inputs.

  • The sources that emitted contaminants that were carcinogenic or had a Hazard Index associated with them were screened as concerned sources for modeling.

  • Those contaminants that had a risk that was significant (which ATSDR designated as greater than 1E-10) were contoured and plotted on a map of the facility and surrounding neighborhoods. Risk was determined by multiplying the concentration in micrograms per cubic meter (mg/m3) of each contaminant by the inhalation slope factor for unit risk (per mg/m3).

  • For individual contaminant runs, some compounds (such as metals) were emitted as particulates and volatiles. Some were subject to decay or reactions with other compounds. For these compounds, there was insufficient data to determine the individual decay products. The following approximations were used.

    • Many source emissions were handled as volatile organic compounds (VOCs) in the model.

    • Others were handled as VOCs with decay factors (in the form of half lives) to account for other depleting mechanisms in the atmosphere.
  • Hexavalent chrome was handled differently. Information indicates that approximately 85%-95% of the total chrome emissions originate as hexavalent chrome. It then reduces to trivalent chrome over time. Laboratory results indicate that the half life varies from 16 seconds to years. In soil, chrome may undergo several changes from hexavalent to trivalent and back over time. Realistically, the half life in the ambient air is closer to the low end of the seconds-to-years range. A value of 16 minutes was considered to be reasonable. This is an important issue because hexavalent chrome is a compound that has the potential to significantly increase the total predicted carcinogenic risks of the facility. The refined modeling considerations used for this evaluation are discussed in the following sections.

3. Refined Modeling Parameters Considered

As mentioned previously, there are often insufficient data to determine the effects of one parameter from the effects of another. For example, particle size concentration must be collected downwind to determine the contribution of the deposition versus decay. When studies are designed to measure concentrations at different distances downwind, they seldom evaluate particulate as well as vapor species. If a certain compound undergoes both mechanisms of plume depletion, there is no way to determine the individual contributions of each. A solution may be either to develop some balance between the two terms or to use one term to consider the effects of both.

Downwash, decay, and deposition are three major issues considered during this modeling exercise. Neither downwash nor deposition was incorporated in the design of the Kelly AFB evaluation, because insufficient data was available and the outcome of their incorporation was considered to be more refined than the available data would allow. Additionally, the impact of the refined solution was considered to be negligible when compared to the uncertainty of the input parameters. These individual considerations for refined modeling are discussed in the next section.

  • Building Downwash

    Building downwash was considered, but not incorporated into the model because detailed stack and building information was not available for all of the sources. Because the distance between the source and receptors is extensive, there would be sufficient time and distance for the plume to spread before the plume reaches the receptors. It is agreed that near field receptors are greatly impacted by a building's effect on the flow of air. An illustrative example was calculated using Huber-Sneider wake effects of dispersion from a 6.6 meter stack that sits on a 6-meter high building (58). Figure B-1 is a graphic illustrating the mathematical dilution that occurs 50 meters downwind of this stack, with and without downwash. To simplify the problem, the solution was only obtained for a single meteorological event (unstable conditions with 2.5 meters/second winds).

    Figure D-1 reveals as much as a 25% reduction due to the initial dilution caused by the eddies which are developed by wind moving around the building. Different meteorological combinations can vary the size of the eddies and therefore vary the amount of dilution. The effect of the initial dilution caused by the building works to spread out the concentrations initially, but as distance increases, the regular Gaussian dilution occurs for distances farther downwind. Figure D-2 is an illustration of the same scenario's effect at a distance of 1500 meters (about the distance from the center of the neighborhoods to the north of Kelly AFB). For the receptors that are 1500 meters away, there is virtually no difference in the solution to the dispersion equation, with and without downwash.

    The principle of conservation of mass is applied to the solving of the dispersion equation and the concentration is spread out over a larger distance (vertically and horizontally). The variability in the horizontal component of the spreading is somewhat offset by the variability of wind direction (which also serves to spread the plume) when the solution is applied to many hours of data.

  • Deposition

    Deposition reduces the mass of particulates in the air causing a decrease in the total concentration in the plume far downwind.

    The deposition algorithm is a constant rate depletion term which accounts for particles falling out of the plume and onto the ground. The effects of the use of this term are variable depending on the distance downwind because more particles are available close to the source and less are available farther from the source. Heavy particles drop out immediately while finer particles get carried farther. Generally, concentrations are greater near the source and decrease as the distance from the source increases.
  • Decay

    Decay coefficients are provided in the model to deplete the plume concentrations by applying a standard logarithmic decay. Compound-specific half lives may be input into the model so that there is a reduction of total mass when viewing each cross section of the plume as distance increases. Hexavalent chromium is known to decay to trivalent chrome at different rates depending, most critically, on medium. The rates vary in air. In conversations with representatives of the California Air Pollution Officers Association and EPA Research Triangle Park during the development of risk assessments, values of half life ranging from 16 seconds (in a laboratory) to several years were discussed (59-62). The geometric mean of the available data was more than 15 minutes; the groups involved suggested that a half-life of one day would be reasonable and still conservative to health. For the assessment at Kelly, a more reasonable (and less conservative) value of 16 minutes was used.

4. Effect of the Refined Assumptions

Maintaining some consistency in setting up models for evaluating similar types of operations is important when comparing the impact of one facility to another. However, some overly conservative values need to be adjusted when evaluating actual impact. Steps were taken to refine the modeling of dispersion from the Kelly AFB operations so that they more realistically represented the actual conditions.

It is difficult to discuss the effects of collectively varying each of the parameters on the net ambient air concentrations predicted by the model. ATSDR has performed a sensitivity analysis to convey the effect of varying individual parameters. A summary of the results obtained for individual effects is provided in Table D-1.

Table D-1.

Effects on predicted downwind breathing-zone concentrations.
Parameter Percentage of Change in Parameter
Nearby Receptors (less than 800 meters) Distant Receptors (greater than 800 meters)
Emission Rate 100 100
Downwash - 57 - 9
Temperature -11 -1
Dry Deposition +5 -20*
Wet Deposition +8 -23*
Decay -5 -48
Stack Height -52 -8
Stack Diameter +12 -1
Exit Velocity -66 -2

* Particle size dependent

Integration of Air Dispersion Modeling Results & Geographic Information Systems (GIS)

Air dispersion models produce contaminant concentration estimates along with the geographic coordinates at which those estimated values occur. Geographic Information Systems can be used to query the estimated value at all points generated by an air dispersion model. Therefore, the model estimate for a point in any desired vicinity within the geographic coverage of the air model may be determined.

Risk Assessment for Air Emissions

Current Air Emissions

Kelly AFB provided ATSDR with air emissions data from 1996, the most recent available data for chemicals and air emissions. Specific emission rates, hours of operation, sources, and meteorological data were used with the general assumptions, previously described, in the model as a screening activity to estimate contaminant concentrations in the community. ATSDR terms this as a screening activity because default values are used for many assumptions for which ATSDR does not have specific data and because conservative emission rates are used when grouping emissions from common locations. Those contaminants with the highest estimated risks are presented in Table D-2. Current emissions that failed to pass an initial screening (which compared the estimated concentration in the community with ATSDR comparison values) were included for further risk assessment. Methylene chloride, benzene, PCE, cadmium, and formaldehyde, although not failing the initial screening, were included because of their potential importance in past air emissions. Initial screening with ATSDR cancer risk evaluation guides (CREGs) are based on a continuous air exposure for a lifetime (estimated at 70 years). Risk assessment involves a more realistic exposure scenario using site-specific conditions, if known. Assumptions used in the risk assessment of current air emissions are as follows:

  • Exposure duration was assumed to be 30 years total. Surveys indicate that 95% of the population in the United States stays in a location for less than 30 years. Because the contact rates may be different for children and adults, carcinogenic risks during the 30 years were calculated using age-adjusted factors. These factors approximate the integrated exposure from birth until age 30 by combining contact rates, body weights, and exposure durations for two age groups--small children and adults. Small children were assumed to weigh 15 kilograms (kg) and breathe 10 cubic meters of air per day for 6 years. Adults were assumed to weigh 70 kg and breathe 20 cubic meters of air per day for 24 years.

  • The frequency of exposure was assumed to be 365 days per year to compare with predicted concentrations which were based on a continuous exposure of 365 days per year.

  • Cancer Slope Factors for inhalation were obtained from EPA Region 6 Human Health Media-Specific Screening Levels, April 1998.

Table D-2 presents the results of air dispersion modeling and risk analysis. Air dispersion modeling is presented as a concentration estimated from modeling the emissions (listed as tons per year or TPY), compared to a cancer comparison value (ATSDR CREG), and the results of risk assessment presented as a potential increase in cancer of 1 case per 100,000 people. These results indicate that chromium and cadmium may represent the most current risk for health effects in the community from air emissions, although the increase in risk may not be detectable as health effects. ATSDR has conservatively estimated that hexavalent chromium represented 85% of the total chromium emissions reported. ATSDR does not know that this is an accurate estimation. In conclusion, exposure to current air emissions is unlikely to result in detectable health effects in the community.

Table D-2.

Present Air Dispersion
Chemical Emissions
(TPY) a
Estimated Increase in Cancer Risk
Hexavalent Chromium 0.38 0.001 0.00008 4/100,000
1,3-Butadiene 0.7 0.014 0.004 0.004/100,000
Arsenic 0.017 0.0003 0.0002 0.06/100,000
Formaldehyde 6.06 0.035 0.08 0.02/100,000
Cadmium 0.003 0.0006 0.0006 0.1/100,000


3.4 0.005 3.0 0.006/100,000
PCE 9.7 0.2 2.0 0.005/100,000
Benzene 1.04 0.012 0.1 0.004/100,000
Cumulative 4/100,000

Shaded areas indicate that concentrations exceed ATSDR comparison values for cancer.
     a TPY: Tons per year. All emissions data are from Kelly AFB 1996 Air Emissions.
     b mg/m3: micrograms per cubic meter (1000 liters). Concentrations are estimated from the Air Dispersion Model.
     c CREG: Cancer risk evaluation guide


Table D-3.

Category Definitions* Used by ATSDR
Category Fraction Decimal Exponential
No Increased Risk <1/100,000 <0.00001 <1E-05
No Apparent Increased Risk 1/100,000 0.00001 1E-05
Low Increased Risk 1/10,000 0.0001 1E-04
Moderate Increased Risk 1/1,000 0.001 1E-03
High Increased Risk 1/100 0.01 1E-02
Very High Increased Risk >1/100 >0.01 >1E-02

* ATSDR Decision Statement TOX.14. Draft QAA-27. Revised October 21, 1991.

Past Air Emissions

Past air emissions will be evaluated and the results will be presented in Phase II of the public health assessment.


Air Exposure Scenario

In this section, ATSDR will discuss how variability and uncertainty may affect the numbers associated with the inputs to some of the relevant sections of the health assessment, particularly the air model and risk assessment, and the conclusions presented in these sections. The main areas of interest are (1) how concentrations were determined in the community, (2) the selection of the exposure scenario describing who came into contact with this concentration of contaminant and for what period of time, and (3) how this relates to potential health effects. Because most uncertainty cannot be accurately quantitated, ATSDR has attempted a qualitative explanation.

1. Uncertainty involving estimation of the concentration of contaminant in the community involves uncertainty of the emission factors (such as the rate of emission, physical location of emission, or the physical form of the chemical in emission) and factors determining the dispersion of chemical from the source to the community (such as meteorological data, decay rates, deposition rates, or obstructions). Emission factors were supplied by Kelly AFB and were assumed to be complete and accurate. However, errors can occur in gathering and calculating information supplied by base personnel, as for information from any source. Estimation of past emissions may contain error because it is not known how representative the selected values were. In addition, because only 4 out of more than 200 chemicals were used in estimates of past emissions, the uncertainty is evident, even though these 4 emissions were determined to be the ones most likely to result in the highest toxicity (relative to the other chemicals).

Uncertainty involving the use of a model to predict community concentrations of chemicals is inherent in the modeling process. Many assumptions must be made. The predictions of long-term modeling tend to be more accurate than short-term modeling. The use of meteorological data in a long-term model tends to buffer the highs and lows, achieving a long-term annual average concentration that is useful in risk analysis, but may not accurately predict the concentration reaching the community at a given time and location. Modeling may be more useful in predicting long-term exposure concentrations than actual sampling and analysis, due to the many sources of error inherent in air sampling and analysis, and particularly considering low detection levels and variability of conditions.

2. Uncertainty associated with the exposure scenario can be reduced by gathering more community-specific information. ATSDR has recommended gathering additional information to reduce this uncertainty. Nevertheless, the scenario is selected to be protective of the public health based on conversations with community members and personal observations. The model averages the chemical concentrations over 24-hour periods for 365 days per year for a 5-year period. While actual concentrations are likely to be higher for shorter periods if Kelly AFB operated on a 8-hour per day, 5 days per week schedule, averaging allows comparison to ATSDR comparison values, which are based on continuous exposure. Residents who do not fit the exposure scenario would be identified in subsequent investigations of exposure history. For example, the concentrations in the community are outdoor concentrations, and it is highly unlikely that any resident would be outdoors at all times. Indoor air concentrations would likely be less than outdoor air concentrations, so exposure concentrations and risk would be less than that predicted for a worst-case scenario. In addition, some residents may not be in the community during exposure periods during the day or throughout the year, which would lessen exposure and risk. Residents may not have lived in the community for a long time, thereby reducing their exposure and risk.

3. Perhaps the greatest area of uncertainty is in assessing the likely health effects from exposure to predicted concentrations of contaminants. The range of human variability, exposures to multiple chemicals in the home, environment, and occupational situations, medications, life style, and other risk factors (such as smoking, alcohol consumption, or nutritional status) may preclude an accurate prediction. Comparison values based on animal studies may not be relevant in some cases, no matter the precautions taken in the use of safety factors. Safety factors attempt to account for the possible differences in responses among humans, between humans and animal subjects, among animals, and between different time frames. Scientists have very little knowledge of the effects of multiple chemical exposure. Nevertheless, ATSDR would be remiss in its duty to protect the public health if ATSDR ignored the plausible links. Health effects have been documented in communities surrounding Kelly AFB which have plausible links to emissions from Kelly AFB. While ATSDR has attempted to quantitate the risk from the concentration predicted by the air model, ATSDR acknowledges that the values that it has used to estimate potential past emissions may not be representative. Those emissions could have been higher or lower. In addition, ATSDR does not have information on the emissions of the vast majority of contaminants potentially emitted in the past by Kelly AFB. It is possible that mixtures of these contaminants may have contributed to potential risk, but ATSDR has no basis on which to quantitate past emissions.

In summary, exposure of communities to air contamination from Kelly AFB and other sources is obvious. The level of exposure to contaminants from Kelly AFB remains uncertain and will remain so, due to the unavailability of past emissions data. Plausible links exist but health effects at predicted levels of contamination are uncertain, both because of the uncertainty in the level of contamination and the uncertainty in predicting health effects from low levels of multiple chemical exposures. Health effects exist in the communities with plausible links to contaminants from Kelly AFB and from other sources. The uncertainty in community exposure scenarios must be clarified to reduce this uncertainty. ATSDR is investigating the feasibility of activities to reduce uncertainty in the following areas:

  • The population with the highest risk for health effects is the on-base population of workers. If this population could be identified and potential health effects investigated, the information obtained would be a valuable indicator for further investigations.

  • Exposure histories of residents could clarify the numbers and extent of exposures, as well as identify confounding variables and risk factors.
  • Significance of elevated numbers of cases of liver cancers could be clarified by further investigation to determine length of residence, stage of disease at diagnosis, whether the liver cancer is primary or secondary, and the existence of contributing factors such as hepatitis infection. Although considering the widespread incidence and mortality of liver cancer in Bexar County and the state of Texas, it is unlikely that an association with emissions from Kelly AFB could be determined.
  • The plausibility of the leukemia outcomes associated with environmental exposure needs to be further investigated. Questions need to be answered to shed additional light on the potential association. Some of these questions involve the proportion of adult/child leukemia; whether the adult leukemia population consists of on-base workers; whether the types of leukemia reported match the types reported to be associated with environmental exposure; and whether exposure history could clarify uncertainty.
  • A method of determining potential past emissions of contaminants from Kelly AFB should be identified.
  • A determination should be made about the feasibility of using biomarkers as an index of exposure for potential past exposures.

Appendix E

Leon Creek


Contaminants of potential concern were selected based on their persistence in the environment, their toxicity, and their ability to bioconcentrate in fish. In addition, some contaminants were selected because of their detection in other environmental media. Any contaminant which had at least one value exceeding its cancer comparison value was selected for further evaluation. ATSDR calculated estimated exposure doses and estimated cancer risk values using the assumptions described under each environmental media. Although the calculations which follow indicate a certain amount of precision, they are estimates using a range of values that include several safety factors. When there is uncertainty, ATSDR overestimates rather than underestimate risk by factors ranging from 10 to 1000. Therefore, our recommendations are highly protective of public health.

ATSDR's conclusion categories describing any increased cancer risk are defined in the following table:
Category* Fraction Decimal Exponential
No Increased Risk <1/100,000 <0.00001 <1E-05
No Apparent Increased Risk 1/100,000 0.00001 1E-05
Low Increased Risk 1/10,000 0.0001 1E-04
Increased Risk
1/1,000 0.001 1E-03
High Increased Risk 1/100 0.01 1E-02
Very High Increased Risk >1/100 >0.01 >1E-02

* ATSDR Decision Statement TOX.14. Draft QAA-27. Revised October 21, 1991.

Estimated Exposure Dose and Cancer Risk from Incidental Ingestion of Surface Water

EPA's Region III Risk-Based Concentration (RBC) Tables for tap water were used as comparison values. Risk-based screening levels for carcinogens were based on combined childhood and adult exposure.

Maximum concentrations of each contaminant were used in lieu of averages or means to reflect worse case conditions.

  • Standard body weights of 70 kg for adults and 15 kg for children were used in dose calculations.
  • Using the scenario describing incidental ingestion while swimming or wading, an incidental ingestion rate of 50 mL/day was assumed with a frequency of 50 days/year for a duration of 30 years. These are highly conservative values.

  • The carcinogenic potency slope for benzo(b)fluoranthene (7.3E-01 risk per mg/kg/day) was used for the total benzo(b+k)fluoranthene since analysis reported the combined concentration. This carcinogenic potency slope for benzo(b)fluoranthene is more conservative than using the carcinogenic potency slope for benzo(k)fluoranthene (7.3E-02 risk per mg/kg/day).

  • Other carcinogenic potency slopes were as follows:
              Phenanthrene: 7.3E+00 risk per mg/kg/day
              Tetrachloroethylene: 5.2E-02 risk per mg/kg/day
              Trichloroethylene: 1.1E-02 risk per mg/kg/day
              Vinyl chloride: 1.9E+00 risk per mg/kg/day.

The results are presented in Table 5.

Estimated Exposure Dose and Cancer Risk from Sediment in Leon Creek

  • ATSDR selected an exposure scenario of incidental ingestion of suspended sediment in Leon Creek as representing the greatest potential exposure to sediment. Dermal exposure was assumed to be minimal because the chemicals in sediment tend to be strongly absorbed to sediment and contact with sediment would be infrequent, with a minimal skin surface area exposed. Organic material accumulates in pockets on the stream bed of Leon Creek; thereby limiting contact to specific areas of the stream bed. Although ingestion of suspended sediment may be unlikely, it represents the greatest potential exposure to chemicals in sediment.
  • ATSDR screened the maximum contaminant concentrations by comparison to the EPA's Region III Risk-Based Concentration (RBC) Tables for residential soil, and selected any contaminant with any value above the comparison value for further evaluation. Maximum concentrations were also used to calculate an exposure dose. Risk-based screening levels for carcinogens were based on combined childhood and adult exposure.
  • ATSDR assumed the same ingestion rates as recommended for incidental soil ingestion of 200 mg/day for children and 100 mg/day for adolescents and adults. ATSDR used standard body weights for children (15 kg), adolescents (42.5 kg), and adults (70 kg).

  • The exposure duration and frequency were assumed to be the same as the scenario described in the investigation of the incidental ingestion of surface water, except the exposure durations were segmented to account for differences in ingestion/body weight ratios. Durations were assumed as follows: children, 0-6 years, adolescents, 7-8 years, and adults, 19-30 years.

  • The oral slope factor that was previously described was used for benzo(b+k)fluoranthene. The polychlorinated biphenyl (PCB) Aroclor 1242 does not have an individual carcinogenic potency slope, so ATSDR used the carcinogenic potency slope for Aroclor 1254 (2.2/mg/kg/day) for all PCB assessments.

  • Other oral slope factors were as follows:
              Benz(a)anthracene: 7.3E-01 risk per mg/kg/day
              DDT: 3.4E-01 risk per mg/kg/day.

The results are presented in Table 6.

Estimated Exposure Dose and Cancer Risk from Consumption of Fish from Leon Creek

  • ATSDR used the Texas Natural Resources and Conservation Commission's estimated ingestion rate for fish of 10 g/day, a rate that reflects local knowledge of cultural preferences and environmental conditions. This would be the most appropriate value to use because the potential population fishing in these waters would be recreational anglers, not subsistence anglers.

  • A frequency of exposure of 365 days/year and a duration of 30 years was assumed.

  • Maximum concentrations were used to reflect an assumed maximum exposure.

  • A standard adult body weight of 70 kg was used for dose calculations.

  • EPA's Region III Risk-Based Concentration (RBC) Table was used for comparison values for fish tissue concentrations, even though these values are based on a greater ingestion rate (54grams/day). All RBCs were based on adult exposure. Any contaminant which had a maximum detection above the comparison value was selected for further evaluation. The selected contaminants, comparison values, and estimated risks are presented in Table 7.

  • The PCB Aroclor 1260 does not have an individual carcinogenic potency slope, and ATSDR used the carcinogenic potency slope for Aroclor 1254 (2.2/mg/kg/day) for all PCB assessments, as previously described. Total PCBs did not exceed the FDA action level of 2000 parts per billion.

  • For polycyclic aromatic hydrocarbons (PAHs) not having an individual carcinogenic potency slope, ATSDR used the total toxic equivalency factor approach, which utilizes the carcinogenic potency slope of 7.3E+00 mg/kg/day derived for benzo(a)pyrene.

  • Carcinogenic potency slopes used were as follows:
              Benzo(a)pyrene: 7.3E+00 risk per mg/kg/day
              DDD: 2.4E-01 risk per mg/kg/day
              DDE: 3.4E-01 risk per mg/kg/day
              DDT: 3.4E-01 risk per mg/kg/day.
  • Pesticides DDD, DDE, and DDT were all included in the risk assessment, even though their total maximum concentrations did not exceed the FDA action level of 5000 parts per billion.

Results are presented in Table 7.

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. #