Past Exposure to Drinking Water From On-Base Wells 313 and 314
KELLY AIR FORCE BASE
SAN ANTONIO, BEXAR COUNTY, TEXAS
The purpose of this health consultation is to determine the likelihood of adverse health effects from drinking water obtained from Kelly Air Force Base on-base Wells 313 and 314. This consultation is Phase III of a three-phase process for evaluating the public health at Kelly Air Force Base. Phase III is divided into two parts: this health consultation and a public health assessment for East Kelly. The East Kelly assessment is being completed as a separate document. Phase I of this process was completed on August 20, 1999, with issuance of the public health assessment for Kelly Air Force Base (AFB). Phase II addresses air emissions and will be completed as a separate document.
The Restoration Advisory Board became concerned about water from on-base Wells 313 and 314 when a connection from the contaminated shallow aquifer to Well 313 was rediscovered in 1999. The connection was originally found in June 1989. Well 314 is located adjacent to Well 313, and was connected to Well 313 in the past. Details of the wells, the sampling, and discussion of the public health implications are provided in the next four sections. In summary, ATSDR concludes that past exposure from ingestion of water from Well 313 and Well 314 is not an apparent public health hazard because the concentrations of chemicals in the water combined with the length of estimated exposures were low.
In this health consultation, ATSDR presents its evaluation of exposure and potential health effects from consuming water from Wells 313 and 314. This consultation is organized into four sections: Background, Data Evaluation and Interpretation, Public Health Implications, and Public Health Action Plan. In the Background section, ATSDR discusses the pertinent environmental conditions and existing data. In the Data Evaluation and Interpretation section, the existing data is evaluated, and the concentrations people may have been exposed to is calculated. In the Public Health Implications section, ATSDR presents its final conclusions. Last, the Public Health Action Plan section discusses the activities Kelly AFB completed in the past and the activities the base is currently conducting to protect public health related to the drinking water. This last section also includes ATSDR's recommendations.
The following pertinent environmental conditions and existing data are presented and discussed in this section.
- The groundwater below Kelly AFB
- Groundwater contamination in the shallow aquifer
- On-base drinking water supply and distribution system
- Groundwater from the shallow aquifer flowing into Well 313
Groundwater below Kelly AFB
Groundwater below Kelly AFB is found in two main geologic zones (see Figure 1). These zones are called the shallow aquifer and the Edwards aquifer. The top of the shallow aquifer is found at depths 3 to 37 feet below ground surface across Kelly AFB with a thickness varying from 0 to 20 feet. The shallow aquifer generally exists as an unconfined, water table aquifer. This means that water may flow freely from the ground surface into the shallow aquifer. Below the shallow aquifer are seven different layers of clay and rock. These layers prevent the shallow aquifer from flowing deeper into the ground.
The second main water-bearing zone below Kelly AFB is the Edwards aquifer. The Edwards aquifer below Kelly AFB is approximately 1,500 feet below the ground surface and slopes to the surface northwest of San Antonio in an area known as the Balcones fault zone. Water enters (recharges) the Edwards aquifer in the Balcones fault zone via runoff from rain and from the bottom of rivers (CH2M Hill 1996). Under natural conditions, the Edwards aquifer is not connected to the shallow aquifer below Kelly AFB. Hence, surface water does not naturally flow into the Edwards aquifer below Kelly AFB. More information about the Edwards aquifer is provided in Appendix D.
The Edwards aquifer is the source of water for about 1.3 million people (as of 1997) in and near San Antonio and for ranches and farms in the region. The aquifer yields large quantities of water to wells and springs (United States Geological Survey 1997). The San Antonio metropolitan area is one of the largest cities in the United States to rely on groundwater (mainly from the Edwards aquifer) for its principal water source.
According to the Texas Water Development Board well database (data obtained from their website in February 2000), there are 1,116 wells and springs in Bexar County (Figure 2). Of these, 1,057 are used to withdraw water and 59 wells are used for oil or gas, mines, or groundwater observations. The 1,057 wells and springs are used for various purposes including commercial, domestic, irrigation, industrial, medicinal, public supply, aquaculture, stock, and institutions. Of these 1,057 wells, 943 tap into the Edwards aquifer. While this well inventory is not 100% accurate, it provides general information about the widespread use of the Edwards aquifer in Bexar County.
Groundwater contamination in the shallow aquifer
Water in the shallow aquifer below Kelly AFB is contaminated with chemicals from Kelly AFB and possibly other sources. The contamination from Kelly AFB is the result of past spills and releases which included solvents, fuels, and other organic chemicals. The area surrounding Kelly AFB is or was used for farming and industry, automotive repair shops, dry cleaners, and paint shops. ATSDR reviewed the shallow groundwater data because the shallow aquifer is the most likely source of contamination to Well 313. However, it is possible that contamination could enter the Edwards aquifer from the recharge area.
Kelly AFB investigated the groundwater contamination in the shallow aquifer by grouping the base into five areas called zones. Wells 313 and 314 are located in Zone 3. To investigate the groundwater contamination in Zone 3, Kelly AFB installed 214 monitoring and temporary wells under the Department of Defense's Installation Restoration Program (IRP). Figure 3 shows Wells 313 and 314 and the adjacent IRP monitoring wells. Two IRP monitoring wells are located within 50 feet of Wells 313 and 314. Data from these two IRP monitoring wells and others in the vicinity indicate that groundwater near Wells 313 and 314 contain tetrachloroethene at 100 micrograms per liter (µg/L) and trichloroethene at 10 µg/L (Figure 3). ATSDR used the data from the IRP monitoring wells in the following exposure assessment discussion as a worst case scenario. The highest concentrations of tetrachloroethene and trichloroethene in Wells 313 and 314 were 5.5 µg/L and 2.8 µg/L, respectively in Well 313 and <0.5 µg/L and <0.5 µg/L, respectively in Well 314.
On-base drinking water supply and distribution system
At Kelly AFB, records indicate that 21 wells tap into the Edwards aquifer currently or in the past (SAIC 1996, 1997, Griffith no date). Twelve of these wells, including Wells 313 and 314, are used or have been used for drinking water and industrial production; six are currently active. The other nine wells have mostly been used for irrigation. Wells 313 and 314 were permanently sealed in 1991 and 1998, respectively. A list of all these wells is provided in Tables 1 and 2.
The drinking water supply wells feed into a closed loop distribution system consisting of tanks, pumps, and pipes. The water is used exclusively for the base, both military and civilian personnel, and services on-base housing, industrial processes, operations, and office buildings.
The wells in the distribution system have been analyzed for volatile organic compounds since 1983. Table 3 provides a summary of the data. From 1983 to 1998, only Wells 313 and 1044 exceeded screening values and U.S. EPA maximum contaminant levels (MCLs) for drinking water and this occurred twice in Well 313 and once in Well 1044. The screening values are modified U.S. Environmental Protection Agency (EPA) Region 6 screening values (See Appendix A). Well 313 exceeded the screening value once in 1986 for 1,2-dichloroethane at 43 µg/L and chloroform at 12 µg/L and once for the MCL in 1988 for tetrachloroethene at 5.5 µg/L (Note: the MCL was not in effect until July 30, 1992 but is used here as a screening value). Tables 4, 5, and 6 provide individual sample results for Wells 313, 314, and 1044. Well 1044 exceeded the screening value and MCL once in 1986. For this health consultation, ATSDR only focused on Wells 313 and 314. Well 1044 is discussed further in Appendix B. The contamination in Wells 313 and 314 in 1986 may have been erroneous due to the same sampling quality control issues that occurred in sampling Well 1044 as discussed in Appendix B
Groundwater from the shallow aquifer flowing into Well 313
In June 1989, Kelly AFB discovered a horizontal pipe connected to Wells 313 and 314 and discovered water running into Well 313 from the horizontal pipe. Well 313 had been taken out of service in January 1989, 6 months prior to this discovery. Three days after discovery, Kelly AFB installed a temporary plug in Well 313 to seal off the horizontal pipe. Well 313 was permanently sealed in 1991. Well 314 was permanently sealed in 1998. A history of Wells 313 and 314 is provided in Table 7.
One day after the horizontal pipe was discovered, Kelly AFB sampled the water running from the horizontal pipe into well 313. The sample results found tetrachloroethene at 52 µg/L. The sample was analyzed on-base and the analysis was not calibrated to determine if other volatile organic compounds were present. To overcome this possible limitation, ATSDR reviewed the data from the shallow groundwater IRP monitoring program as presented in the previous section and in the Data Evaluation and Interpretation Section below.
Understanding the design of Wells 313 and 314 is important for understanding how the water from the surficial aquifer could drain into the well. The design of the wells prior to the 1970s is shown in Figure 4. The two wells, 30 to 50 feet apart, were connected by a horizontal pipe which in turn was connected to pumps. The artesian conditions of the Edwards aquifer brought the water from the Edwards aquifer into the horizontal pipe and to the pumps. The pumps pushed the water into the distribution system. Sometime in the 1970s, these pumps were removed and new pumps were installed in each well (Figure 5). At this same time, plugs were installed in the horizontal pipe. The new pumps then pumped water from the wells into the distribution system, bypassing the horizontal pipe.
Under natural conditions, the water from the shallow aquifer is not connected to the Edwards aquifer. However, if a leak occurred in the horizontal pipe and the plug, a connection could be established. Water could then flow between the well and the shallow aquifer. The direction of the flow would depend on the relative water levels of the shallow aquifer and the Edwards aquifer. If the well water level from the Edwards aquifer was below the water level of the top of the shallow aquifer, water from the shallow aquifer would leak into the horizontal pipe and the Edwards aquifer well (Figure 6). Conversely, if the well water level from the Edwards aquifer were above the water level of the top of the shallow aquifer, water from the Edwards aquifer well would leak into the shallow aquifer (Figure 7). Therefore, the relative water levels are important for determining how long the contaminated water could have been flowing into Well 313.
The water level of the Edwards aquifer at Kelly AFB rises and falls based on the amount of recharge and the amount withdrawn. Similarly, the water level in the shallow aquifer also rises and falls with time, based on rainfall. For the shallow aquifer, water level data is available from Kelly AFB from the two IRP monitoring wells located near Wells 313 and 314 (Figure 3). For the Edwards aquifer, water level readings are available from the San Antonio Water System (SAWS) Edwards Monitoring Well No. J-17 and other wells in the Texas Water Development database. ATSDR reviewed data from private well No. 6844214 because of its close proximity to Kelly AFB (Figure 2). The water levels from these wells are plotted in Figure 8. Water level readings from Wells 313 and 314 are not available. From this data, ATSDR found that the water levels in the SAWS well J-17 and well 6844214 do not correlate over time. Therefore, ATSDR cannot use either well for determining the water levels in Wells 313 and 314 and could not derive an exposure duration based on this data. Instead, ATSDR derived exposure durations based on the well sampling data discussed in the next section. The differences in water level readings are probably the result of the heterogenous nature of the Edwards aquifer.
In this section, data from the Background section is used to formulate exposure levels and durations. The results are then compared to health screening levels.
When did contamination first appear in Wells 313 and 314 and how long did the contamination last?
Volatile organic compounds were first detected in Wells 313 and 314 in January 1986 (Tables 4 and 5). These compounds were not detected in the previous sampling round conducted under the Safe Drinking Water Act (SDWA) in 1983. The samples collected in January 1986 were conducted under the IRP monitoring program and were reported by Kelly AFB to be erroneous. Kelly AFB sampled Wells 313 and 314 five-additional times in 1986 and found the last samples collected in May 1986 did not contain compounds above detection limits. Kelly AFB concluded that the errors were the result of improper cleaning of equipment after the equipment was used to sample the IRP monitoring wells. To be protective of public health, ATSDR assumed that the results in 1986 were not erroneous.
Another factor ATSDR considered in determining when contamination first appeared is the changing level of the Edwards aquifer as discussed previously and when the leak in the horizontal pipe occurred. Unfortunately, this data is not available. Therefore, ATSDR chose to be conservative and assumed that exposure started in 1983. ATSDR is assuming that the 1983 sample results are representative of previous water quality prior to 1983. For Well 313, ATSDR is assuming that the exposure duration was 6 years from just after the well was sampled in 1983 to January 1989 when the well was closed.
For Well 314, samples of the wells collected in February 1990 and thereafter did not detect volatile organic compounds. Therefore, the maximum duration of exposure from Well 314 would also be about 6 years, from just after the well was sampled in November 1983 to February 1990.
The exposure durations could underestimate actual exposure if the one sample in 1983 is not considered representative of past water quality. The misrepresentation is possible because of the design of Well 313 and the variable water levels in the aquifers discussed in the previous section. When the water level of the Edwards aquifer was below the upper level of the surficial aquifer, the potential existed for water to migrate into the well through any leak that may have developed in the horizontal pipe. Conversely, when the water level in the Edwards aquifer was above the shallow aquifer, the potential existed for the water to migrate from the well into the shallow aquifer.
It is not known when the leak began in the horizontal pipe. In addition, the historic water levels in the shallow aquifer and the Edwards Aquifer below Kelly AFB are not precisely known.
After considering all this information, ATSDR considered the 6-year exposure an adequate assumption based on the well sampling.
What contaminants and what concentrations were found in Wells 313 and 314 and the horizontal pipe?
Data for Wells 313 and 314 are presented in Tables 4 and 5, respectively. In Well 313, 1,2-dichloroethane, benzene, chloroform, methylene chloride, tetrachloroethene, and trichloroethene were detected (Table 4). Of these chemicals, 1,2-dichloroethane (43 µg/L) and chloroform (12 µg/L) were the only compounds detected above the screening levels. Tetrachloroethene was detected once in 1988 at 5.5 µg/L which is above EPA's MCL of 5 µg/L.
In Well 314, benzene, chloroform, ethylbenzene, methylene chloride, 1,1,1,2-tetrachloroethane, and 1,1,2-trichloroethane were detected. None of these chemicals exceeded the screening levels (Table 5) or MCLs.
As explained previously, some of these results were determined to be erroneous but were considered in ATSDR's evaluation anyway in order to be more protective of public health.
Of the contaminants that exceeded the screening levels, what concentrations were people exposed to?
Chloroform, 1,2-dichloroethane, and tetrachloroethene were the only compounds that exceeded the screening values or MCLs in the water from Wells 313 and 314. As a result, ATSDR then determined the concentrations of chloroform, 1,2-dichloroethane, and tetrachloroethene that were in the water as it reached the faucets.
Kelly AFB and the state do not sample the water from faucets in offices and on-base housing for volatile organic compounds (VOCs). Sampling for VOCs only occurs at the wells. Therefore, ATSDR took two approaches to calculate the VOC concentrations to which people were exposed. First, ATSDR used the VOC concentrations detected in 1986 and 1988 (12 µg/L chloroform, 43 µg/L 1,2-dichloroethane and 5.5 µg/L tetrachloroethene) in Well 313 and calculated the concentrations after mixing with water from Well 314. Water pumped from Well 313 and Well 314 was mixed together before distribution. Using the annual well production volumes for 1984 (1984 data were used because it resulted in the highest calculated concentrations for comparison, see Table 8 for the data) and assuming the concentration in Well 314 was at the detection level, the water would have been distributed at a concentration of 7.8 µg/L chloroform, 26.8 µg/L 1,2-dichloroethene, and 3.8 µg/L tetrachloroethene (See Appendix C-1). These concentrations would be further reduced as the water combined with water in the distribution system from other production wells. Based on this first dilution, chloroform and tetrachloroethene would be reduced below MCLs and screening values while 1,2-dichloroethane would be above screening values by a factor of four. Looking closer at the 1,2-dichloroethane results, 1,2-dichloroethane was detected twice in January 22, 1986 when Well 313 was sampled 5 times. In fact, two samples were collected on January 22, 1986. One sample contained 43 µg/L 1,2-dichloroethane but the second sample, 1,2-dichloroethane was not detected. In March 1986, 1,2-dichloroethane was detected at 3 µg/L and 2.1 µg/L. In April 1986, 1,2-dichloroethane was not detected. Therefore, the 43 µg/L is not representative of the water from Well 313. Nonetheless, if 26.8 µg/L 1,2-dichloroethane did reach people, it was for a relatively short duration, about 9 months. The cancer screening values are based on a lifetime of exposure, 30 to 70 years. The modified screening value we used for 1,2-dichloroethane, 6 µg/L for cancer and 17 µg/L for noncancer, is based on 6 years of exposure. For one year of exposure, the screening value for cancer would be 36 µg/L, indicating that cancer from exposure to 1,2-dichloroethane in the water is not likely.
The noncancer screening value used by Region 6 for 1,2-dichloroethane is based on a reference dose that has now been withdrawn (2.9 x 10-3 mg/kg/day). EPA Region 9's noncancer screening value for 1,2-dichloroethane has not been withdrawn and is based on a reference dose that is 10 times larger for ingestion (i.e., less toxic, 3 x 10-2 mg/kg/day). ATSDR's reference dose for intermediate exposure (less than one year, 0.2 mg/kg/day) is 100 times larger than EPA Region 6's reference dose. ATSDR's reference dose translates to concentrations of 2000 µg/L for children and 7000 µg/L for adults. Hence, if the water from Well 313 did contain 43 µg/L 1,2-dichloroethane, there would be no apparent public health hazard because the concentration was relatively low concentration and was for a short duration.
In the second approach, ATSDR calculated potential concentrations of the VOCs in tap water based on concentrations of VOCs leaking into Well 313 from the shallow aquifer. This approach requires information concerning the concentrations in the water leaking into the pipe, how much water was leaking into the pipe, the concentrations in the Edwards aquifer, and the amount of water coming from the Edwards aquifer. The concentration of the water from the horizontal well was measured at 52 µg/L. As discussed in the previous section, the concentration of tetrachloroethene and trichloroethene could have been as high as 100 µg/L and 10 µg/L respectively in the shallow aquifer. 1,2-Dichloroethene was not detected in the IRP monitoring wells adjacent to Wells 313 and 314 and was not analyzed in this approach. The highest concentration of chloroform detected in the IRP monitoring wells was 8 µg/L which is equal to the screening value and hence, not analyzed in this approach since dilution would reduce the concentration to levels below concern. ATSDR included trichloroethene in this analysis because the concentration in the shallow aquifer exceeds the MCL.
The amount of water leaking from the shallow aquifer has not been measured, but can be derived from the following information.
- Kelly AFB is pumping groundwater from the shallow aquifer for treatment. Pumping rates vary from 1.2 to 50 gallons per minute (gpm).
- The shallow groundwater is moving toward and discharging into Leon Creek. The rate at which this groundwater is moving provides information about how much water could flow into the horizontal pipe. The flow rates into Leon Creek range from 0.045 gpm to 1392 gpm (CH2M Hill 1999). These rates include the total amount of water over the entire 14,755 feet of Leon Creek as it passes through Kelly AFB. This equates to 0.047 gpm to 0.0000015 gpm per foot per side of Leon Creek.
From these numbers, the potential amount of water discharging into the pipe could be as much as 50 gpm (theoretical maximum) or as low as 0.0000015 gpm. If 50 gpm is used (based on the pumping rate), one must consider that this flow rate is the cumulative flow rate from the entire well screen. If a 10-foot well screen is assumed, the flow rate is about 5 gpm per foot of screen length. This value doesn't include the possibility of the screen being exposed to the unsaturated zone due to the cone of depression, which would increase the flow rate per foot of screen length. It also doesn't consider the total exposed surface area of the screen. For example, a one foot section of screen, 2 inches in diameter, would have an exposed surface of about 0.5-square feet. This is equivalent to a hole in the horizontal pipe 8.6 inches by 8.6 inches (not including the area occupied by the screen which could be more than one-half the exposed surface).
Assuming that the leak in the horizontal pipe is similar to one foot of screen length in the pipe, 10 gpm is flowing into the pipe. Using 10 gpm containing a concentration of 100 µg/L tetrachloroethene and 10 µg/L trichloroethene, combining with water from the Edwards aquifer with concentrations at the detection level of 1 µg/L, the concentration of water coming out of Well 313 would be 3.26 µg/L tetrachloroethene and 1.21 µg/L trichloroethene (See Appendix C, these calculations used the 1983 production data because it produced the highest concentrations for comparison).
If 50 gpm is used as the amount of water entering Well 313 from the horizontal pipe, which is unlikely, the concentration of water coming from Well 313 would be 12.3 µg/L tetrachloroethene and 2.03 µg/L trichloroethene. Mixing this water from Well 314, the concentrations would be 5.57 µg/L for tetrachloroethene and 1.42 µg/L for trichloroethene. These values are below the health screening values of 55 µg/L tetrachloroethene and 80 µg/L trichloroethene. The value of 5.57 µg/L tetrachloroethene is slightly above the MCL, but it is based on the 50 gpm of water entering the horizontal pipe, which is a high end estimate. When 50 gpm drops to 43.8 gpm, the concentration of tetrachloroethene would become 5.0 µg/L. This high end estimate of 50 gpm is not likely. Lower flow rates, lower than 43.8 gpm, are more likely.
Based on our calculated concentrations, flow rates from the horizontal pipe, and assumptions presented in the previous section, the water entering Well 313 from the horizontal pipe did not contaminate the well water at levels of public health concern. High end estimates indicate that the concentration could possibly have been above MCLs, but it is very unlikely. Concentrations measured in Well 313 were not reproducible and not consistent. If a chemical was present, the reported concentration was at relatively low levels and present for a short period of time. Hence, past exposures from potable use of water from Well 313 and 314 pose no apparent public health hazard. The exposure pathway is summarized in Table 9.
As part of this health consultation, ATSDR considered the unique vulnerabilities of infants and children exposed to environmental contamination and hazards. Children could be present in on-base housing and exposed to the water. Water from Well 313 and Well 314 is mixed together and put into the distribution system. The water in the distribution system is further mixed with water from other wells. ATSDR considered this mixing and levels of exposure and concluded that past use of water from Wells 313 and 314 does not pose an apparent public health hazard to children.
Based on the concentrations, lengths of exposure, and the toxicity of the contaminants in the drinking water, potable use of water from Well 313 and Well 314 pose no apparent public health hazard to children and adults.
Kelly AFB has sampled the production wells regularly since 1983 as part of the Safe Water Drinking Act. When the leak in the horizontal pipe was discovered, a temporary plug was placed in the well three days later to prevent the shallow aquifer water from leaking into the well. Kelly AFB plugged Well 313 in 1989 and permanently sealed it in 1991. Well 314 was permanently sealed in 1998.
Kelly AFB is required to monitor the production wells per the Safe Drinking Water Act. This requirement is sufficient to monitor water quality at Kelly AFB at the present time and in the future.
CH2M Hill 1996. 1995 Annual Report, Kelly AFB Basewide Remedial Assessment, July 1996.
CH2M Hill 1999. Semiannual Compliance Plan Report (7/98 - 12/98), Kelly AFB Basewide Remedial Assessment, January 1999.
Gargiulo 1998. Kelly AFB Safe Drinking Water Program, Presentation to the Restoration Advisory Board, October 1998. Major Robert Gargiulo, Bioenvironmental Engineering Flight.
Griffith no date. Map of the Kelly AFB Edwards Aquifer Wells.
Kelly AFB no date. Summary of sample results for drinking water wells.
SAIC 1996. Technical Report, Investigation and Abandonment of Edwards Wells, I-65, I-69, I-75, I-76, I-80, and I-99, Science Applications International Corporation, September 1996.
SAIC 1997. Technical Report, Investigation, Location, and Abandonment of Edwards Wells, I-72, I-74, I-80, Science Applications International Corporation, February 1997.
Texas Water Development Board 2000. Well Information http://www.twdb.state.tx.us/Newwell/well_info.html, accessed February 2000.
U.S. Geological Society 1997. Water Quality Assessment of South-Central Texas--Occurrence and Distribution of Volatile Organic Compounds in Surface Water and Ground Water, 1983-1994, and Implications for Future Monitoring, by Patricia B. Ging, Linda J. Judd, and Kirby H. Wynn. Water-Resources Investigations Report 97-4028.
Brian M. Kaplan, M.S.
Environmental Health Scientist
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation
John E. Abraham, Ph.D, M.P.H.
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation
David Fowler, Ph.D
Petitions Response Section
Exposure Investigations and Consultations Branch
Division of Health Assessments and Consultations
Office of Policy and External Affairs
Chief, Consultations Section
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation