HEALTH CONSULTATION
BARTON SPRINGS POOL
AUSTIN, TRAVIS COUNTY, TEXAS
Barton Springs Pool is a 1.9 acre pool, fed from underground springs which discharge from the Barton Springs segment of the Edwards Aquifer. The pool is located within the confines of Barton Creek; however, water from the creek only enters the pool during flood events. The pool is located in downtown Austin and is used year round for recreation. Barton Springs Pool also is one of the only known habitats of the Barton Springs salamander (Eurycea sosorum) an endangered species. The City of Austin has been collecting water and sediment samples from Barton Springs Pool since 1991.
Recent articles in the local daily newspaper have raised safety concerns regarding environmental contaminants found in the pool. In response to these concerns, the City Manager closed the pool pending an analysis of the perceived human health risks associated with chemical exposures occurring while swimming in the pool.
We reviewed the results from water and sediment samples collected by the City of Austin, the United States Geological Survey, the Lower Colorado River Authority, and the Texas Commission on Environmental Quality. We reviewed over 14,500 individual data points, involving approximately 441 analytes, collected over the past 12 years.
We screened the contaminants by comparing reported concentrations to health-based screening values and selected twenty-seven contaminants for further consideration. Of those 27 contaminants, 20 were polycyclic aromatic hydrocarbons (PAHs). The others included arsenic, boron, cadmium, bis(2-ethylhexyl)phthalate, total petroleum hydrocarbons (TPH), thallium, and lead.
The potential public health implications of the selected contaminants were assessed by considering the toxicological properties of the contaminants, the probable routes of exposure, and the types of exposures that might occur. Conservative assumptions (those that err on the side of protecting the public) were used in these assessments. In some instances we compared contaminant concentrations in the pool water to public drinking water standards assuming that if the reported concentrations were deemed appropriate for drinking, cooking, and bathing, then they were appropriate for recreational activities.
We did not find any information to support contention that swimming every day in Barton Springs would result in adverse health effects. Thus, we have concluded that swimming and playing in Barton Springs Pool does poses no apparent public health hazard. We recommend continued public health education to address any questions that the public may have concerning the risks associated with swimming in the pool.
During our investigation we did note elevated levels of certain contaminants in soil from a hill on the south side of the pool. On inspection of the sampling locations we discovered that the samples were collected adjacent to creosote-treated posts used to control erosion. We recommend that the city investigate the potential for these posts to serve as a source of contaminants that might migrate into the pool.
One of the contaminants we evaluated, TPH, is a common fuel-related urban pollutant. Benzene, toluene, ethylbenzene, and xylenes (BTEX) are commonly associated with fuel related contamination. While the levels of TPH - or TPH components - measured are not expected to result in adverse health outcomes, at least one citizen indicated that he periodically detected a gasoline type odor while swimming in the pool. We recommend that the city continue to monitor water and sediment from the pool for TPH and BTEX and that the samples be analyzed in a manner suitable for comparison to the TCEQ's protective contaminant levels.
The Texas Department of Health (TDH) and the Agency for Toxic Substances and Disease Registry (ATSDR) were asked by the City of Austin (COA) to determine the public health significance of the polycyclic aromatic hydrocarbons (PAHs) detected in sediment from Barton Springs Pool. Samples were collected by the COA, the Texas Commission on Environmental Quality (TCEQ), the Lower Colorado River Authority (LCRA), and the United States Geological Survey (USGS). Recent articles in the local daily newspaper have raised concerns regarding some of the chemicals found in the pool. In response to these concerns, the city manager closed the pool pending an analysis of the perceived human health risks associated with chemical exposures occurring while swimming and playing in the pool.
Barton Springs Pool is a 1.9 acre pool, fed from underground springs that discharge from the Barton Springs segment of the Edwards Aquifer. Water discharging from Barton Springs originates from the contributing watersheds: Barton Creek, Onion Creek, and their tributaries. The average flow of the combined Barton Springs is 53 cubic feet per second or about 34 million gallons per day. The pool is located in downtown Austin within the confines of Barton Creek, which begins northeast of Dripping Springs in northern Hays County. The creek flows east for forty miles to its mouth on the Colorado River at Town Lake in southwest Austin. It bypasses Barton Springs Pool through a culvert that runs beneath the sidewalk on the north side of the pool. Although water from Barton Springs Pool empties into Barton Creek at the east end of the pool, water from Barton Creek only enters Barton Springs Pools during flood events severe enough to exceed the capacity of the bypass culvert causing creek water to flow over the upstream dam. During these brief periods in which Barton Creek flows over the upstream dam, water and suspended solids from the Barton Creek watershed enter the pool, with much of it being discharged downstream of the pool. The City of Austin routinely closes the pool for cleaning after flood events.
The shallow end of the pool consists primarily of limestone rock covered with algae that accumulates on the rock and is removed weekly. The deep end of the pool consists of a mixture of gravel, sediment, and rock. The depth of the shallowest portion of the deep end ranges from approximately 4 to 5 ½ feet. This area is frequently referred to as the "beach" area and also has a bottom that consists of a mixture of gravel, sediment, and rock (Figure 1). The pool is closed when water turbidity is high; the vast majority of these high turbidity events are associated with heavy rain events. Barton Springs Pool is one of the only known habitats of the Barton Springs salamander (Eurycea sosorum) and on May 30, 1997 (62 FR 23377), the U.S. Fish and Wildlife Service listed the salamander as an endangered species. A second blind salamander (Eurycea waterlooensis) was recently discovered to inhabit the spring.
The City of Austin Watershed Protection and Development Review Department has been collecting water and sediment samples from the Barton Springs Pool since 1991 and the United States Geological Survey (USGS) has been sampling since 1978. The majority of the city's water samples were taken as close as possible to the spring outlets. Pool water data and analytes have varied over time but typically include field measurements and physical properties and then conventional pollutants (nutrients and ions) with toxics (metals, polychlorinated biphenyls (PCBs), pesticides, organics, and PAHs) on a less frequent schedule. The majority of the sediment samples were composited from the fissures and deeper portions of the pool which are potential Barton Springs Salamander habitat. These samples have been routinely analyzed for conventional pollutants, metals, pesticides, PCBs, and PAHs.
For this report, we reviewed the results from samples collected by the COA, the USGS, the LCRA, and the TCEQ. A complete list of the analytes included in the various sampling events are listed in Appendix A. Table 1 contains a list of the analytes for which at least one sample was reported above the detection limit in the sediment from the pool. Table 2 contains a list of the analytes for which at least one sample was reported above the detection limit in the water. Analytes not found above their respective detection limits in any of the samples have been omitted from further consideration. Many of the naturally occurring analytes found in the sediment were reported at concentrations equivalent to those normally reported for the Western United States (Tables 1 and 2). Analytes found within normal background ranges were omitted from further consideration; any potential risks associated with these contaminants, at Barton Springs Pool, would not be greater than those that would be experienced in any other natural water body used for recreational purposes.
Introduction
Exposure to, or contact with, chemical contaminants drives both the ATSDR public health assessment and health consultation processes. People may be adversely affected by chemicals only if exposure occurs; that is, they must come into contact with the chemicals and absorb them into their bodies. The presence of chemical contaminants in the environment does not always result in contact and contact does not always result in the chemical being absorbed into the body. The most common ways people come into contact with chemicals are by inhalation (breathing), ingestion (eating or drinking), or by dermal contact (absorption through skin) with a substance containing the contaminant.
Pathways Analysis
Generally, for chemicals found in sediment or water, absorption through the gastrointestinal (GI) tract by incidental ingestion or through the skin by direct contact are the exposure pathways of greatest concern. Whether adverse health effects occur depends on: 1) the toxicological properties of the chemicals; 2) the manner in which the person contacts the chemical; 3) the concentration of the chemical; 4) how often the exposure occurs; 5) how long the exposure occurs; and 6) how much of the chemical is absorbed into the body during each exposure event.
The lack of significant sediment in the shallow end where children play, the sporadic nature of the contaminants in sediment from the deep end, and the general nature of the types of contact that people might have with sediment in the pool eliminate absorption through the skin as a plausible exposure pathway for contaminants in the sediment. Thus, at Barton Springs pool, the only reasonable route of exposure to contaminants in the sediment is through incidental ingestion of sediment in the water column while swimming and playing in the pool. For contaminants in the water, exposure to contaminants both through dermal absorption and incidental ingestion are possible.
To help determine how much sediment a person might ingest while swimming, the City of Austin collected water samples from selected locations in the pool as well as upstream and downstream of the pool. These samples were analyzed for the amount of sediment suspended in the water column. Samples were collected mid-afternoon during peak swim periods over consecutive days of a hot summer weekend and the Labor Day holiday. The city chose a busy weekend to reflect conditions favorable for re-suspending sediment from the bottom of the pool. According to the City, hundreds of swimmers were in the pool during the collection of the samples. Suspended sediment measurements in Barton Springs Pool ranged from 0.4 (1) milligrams of sediment per liter of water (mg/L) to 4.95 mg/L with an arithmetic average concentration of 1.14 mg/L. Samples upstream and downstream of the pool were 0.8 and 2.55 mg/L, respectively [1].
Determining Contaminants of Concern
We screened contaminants for further consideration by comparing contaminant concentrations to media-specific health-based assessment comparison (HAC) values for non-cancer and cancer endpoints. Non-cancer screening values are based on ATSDR's minimal risk levels (MRLs) or EPA's reference doses (RfDs). MRLs and RfDs are based on the assumption that there is an identifiable threshold (both for the individual and for populations) below which there are no observable adverse effects. Thus, these values are estimates of a daily human exposure to a contaminant that is unlikely to cause adverse non-cancer health effects even if the exposure were to occur for a lifetime. For contaminants that are considered to be known human carcinogens, probable human carcinogens, or possible human carcinogens we calculated cancer risk evaluation guides (CREG) using EPA's chemical-specific cancer slope factors and an estimated excess lifetime cancer risk of one-in-one million persons exposed for a lifetime. For both non-cancer and cancer HAC values we used standard assumptions for body weight (70 kg adult and 15 kg child) and soil/sediment ingestion (100 mg/day for adults and 200 mg/day for a child) to calculate the HAC values.
The exposure assumptions used to establish these screening levels are chosen to be conservative with respect to protecting public health; thus, actual exposures are likely to be lower than those used to calculate the screening values. Thus, exceeding a screening value does not mean that a contaminant represents a public health threat, rather it suggests that the contaminant warrants further consideration. Assessing the public health significance of contaminants that exceed their respective screening levels involves reviewing and integrating relevant toxicological information with plausible exposures. We may estimate the magnitude of the public health significance by comparing the estimated exposures to reported "No Observed" and Lowest Observed" Adverse Effects Levels (NOAELs and LOAELs) in animals and to known effect levels in humans, when available.
In some instances, when MRL, RfD or CREG based HAC values are not available or when circumstances dictate that it is reasonable to do so, we use regulatory limits to screen the contaminants for further consideration. For instance, we may use EPA's maximum contaminant levels (MCLs) to screen contaminants in water, or the Texas Commission on Environmental Quality's (TCEQ) protective contaminant levels (PCLs) to screen contaminants in water, sediment, or soil. MCLs are chemical-specific maximum concentrations allowed in water delivered to the users of a public water system. Although MCLs are not solely health-based, they are considered protective of public health over a lifetime (70 years) of exposure at an ingestion rate of two liters per day. Although MCLs only apply to public water systems, we often use them to help assess the potential public health implications of contaminants found in water from other sources. Using MCLs to screen contaminants assumes that if the water is potable (safe for drinking, cooking, bathing), it also will be safe for recreation (swimming, boating, skiing, etc.). While PCLs are regulatory limits, they are generally calculated using conservative assumptions in order to protect public health. When applied appropriately - that is in a manner consistent with the scenarios for which they were developed - they can be used to screen contaminants for further consideration. As with other screening values, exceeding a regulatory limit does not mean that a contaminant represents a public health threat; rather, it suggests that the contaminant may warrant further consideration.
Table 3 contains a list of the contaminants selected for further consideration. This list consists of the PAHs, total petroleum hydrocarbons (TPH), and all contaminants for which at least one value exceeded the screening value. Thus, for some of the contaminants included in this table, only one value was above the screening value. The following sections discuss the potential public health implications of the contaminants selected for further consideration.
Polycyclic Aromatic Hydrocarbons (PAHs)
Polycyclic aromatic hydrocarbons (PAHs) are very common in the environment. They may occur naturally and also are formed during the incomplete burning of coal, oil, gas, wood, garbage, or other organic substances, such as tobacco and charbroiled meat. There are more than 100 different PAHs and they are generally found as mixtures, not as single compounds. While PAHs occur naturally, they also can be found in asphalt, crude oil, coal, coal tar pitch, creosote, and roofing tar. In general, PAHs do not dissolve well in water; rather, they tend to stick tightly to solid particles - such as soil/sediment - that can settle at the bottom of lakes, rivers, and creeks [2].
Because PAHs are so common in the environment, people are exposed to them everyday. The most common sources of exposure to PAHs are tobacco smoke, food, wood smoke, and ambient air. Exposure to PAHs via inhalation is estimated to range from 0.02 to 3 µg/day. Smoking one pack of unfiltered cigarettes per day increases this estimate by an additional 2 to 5 µg/day; chain smokers consuming three packs per day increase their exposure by an estimated 6 to 15 µg/day. The intake of carcinogenic PAHs from the average American diet has been estimated to range from 1 to 5 µg/day, mostly from the ingestion of unprocessed grains and cooked meats. This dietary estimate increases to 6 to 9 µg/day for individuals who eat large amounts of meat [2]. Estimated excess lifetime cancer risk estimates associated with common everyday exposures to PAHs are presented in Appendix D.
Under some circumstances, PAHs can be harmful. The harmful effects observed often depend on the type of exposure or the way that the contaminants enter the body. Absorption through the skin by direct contact, and through the gastrointestinal (GI) tract by incidental ingestion, are generally the exposure pathways of greatest concern when dealing with contaminated sediment. Adverse non-cancer effects have been noted in humans following intermediate-duration dermal exposure in patients with preexisting skin disorders. A solution of benzo[a]pyrene applied to the skin of patients with pemphigus vulgaris (a disease characterized by successive crops of blisters) caused local eruptions characteristic of the disease. The same type of solution applied to the skin of patients with xeroderma pigmentosum resulted in pigmentary changes and wart-like effects. When the solution was applied to patients with squamous cell cancer investigators observed a general improvement and/or retardation of the lesion. Adverse effects on the skin also have been observed in animals following dermal exposure to various PAHs [2].
While there is little evidence to indicate a relationship between ingestion of PAHs and adverse health effects in humans, animal studies have shown that ingestion of PAHs causes gastrointestinal (digestive system), hepatic (liver), reproductive, and developmental effects. The lowest doses associated with these effects have ranged from 40 milligrams per kilogram of body weight per day (40 mg/kg/day) to 700 mg/kg/day [2], levels that are orders of magnitude greater than those likely to occur at Barton Springs Pool.
Several of the PAHs also have been shown to cause tumors in laboratory animals when they breathed these substances in the air, when they ate them, or when they had prolonged skin contact with them. Benzo[a]pyrene (BaP) is perhaps the most toxicologically significant PAH and - along with several other PAHs - has been classified by the EPA as a "probable human carcinogen". This classification is based on animal data where repeated BaP administration in numerous strains of at least four species of rodents and several primates has been associated with increased incidences of total tumors and of tumors at the site of exposure. Human data specifically linking benzo[a]pyrene (BaP), or any of the other PAHs to a carcinogenic effect are lacking. Although lung cancer has been found in humans by exposure to various mixtures of polycyclic aromatic hydrocarbons known to contain BaP - including cigarette smoke, roofing tar, and coke oven emissions - it is not possible to conclude from this information that BaP or any other of the PAHs is the responsible agent. BaP also has produced positive results in numerous genotoxicity assays.
To assess the potential health risks associated with PAHs in sediment from the pool, we evaluated the toxicity of the contaminants with respect to exposures that might occur at the pool. We considered the incidental ingestion of sediment in the water while swimming as the primary pathway of exposure.
To assess this pathway, PAHs were converted to benzo(a)pyrene equivalents using established toxicity equivalency factors (Appendix B). The average benzo(a)pyrene equivalent of 1.46 mg/kg, for samples collected between 1991 and 2003 (Figure 2; Appendix C), was used in all subsequent calculations. To estimate a protective (with respect to public health) exposure scenario, we assumed that an individual would ingest 50 milliliters of pool water per hour, three hours per day, 7 days per week, 52 weeks per year, for 70 years. Using the average suspended sediment concentration (collected by the City of Austin during peak usage) we estimated the excess lifetime cancer risk associated with this pathway to be 2.6 x 10-8, or about one (1) in 38 million (Table 4). Qualitatively, we would interpret this to be an insignificant increased risk. Although the dermal exposure pathway is unlikely, we did evaluate the potential public health significance of possible dermal exposure for a child (Appendix D). Qualitatively, we would estimate the risks associated with this pathway to be insignificant.
Arsenic
Arsenic is a naturally occurring element in the earth's crust and is usually found in combination with other elements. Arsenic compounds can be classified into three main groups: 1) inorganic arsenic compounds, 2) organic arsenic compounds, and 3) arsine gas. In the environment, arsenic is most often found as inorganic arsenic, which is formed when arsenic combines with other elements such as oxygen, sulfur, and chlorine. Organic forms of arsenic, which result when arsenic combines with carbon and hydrogen, are generally considered less toxic than the inorganic forms. Background concentrations of arsenic in soil range from 0.1 to 97 mg/kg with an average value of about 7.0 mg/kg [3]. EPA Region 6 considers an arsenic concentration range of 1.1 to 16.7 mg/kg to be typical of background concentrations. TCEQ Risk Reduction Rules suggest 5.9 mg/kg to be the mean background concentration for Texas soils.
Analysis of the toxic effects of inorganic arsenic in soil and sediment can be complicated by the fact that arsenic can exist in different valence states. A number of studies have noted differences in the relative toxicity of these compounds, with trivalent arsenites tending to be somewhat more toxic than the pentavalent arsenates [4, 5, 6, 7]. For this consultation, we have not emphasized the difference in toxicity between the arsenates and the arsenites because in most instances, this difference is relatively small (about 2-3 fold) and is often within the bounds of uncertainty regarding the NOAELs or LOAELs. In addition, the different forms of arsenic may be interconverted, both in the environment and in the body, and as in most cases involving human exposure, the precise chemical speciation at this site is not known.
To assess the potential health risks associated with arsenic in soil and sediment from Barton Springs Pool, we compared the concentrations to HAC values for both non-cancer and cancer endpoints for the more toxic form of the compound. The non-cancer HAC values for arsenic in soil (20 mg/kg for children and 200 mg/kg for adults) are based on EPA's reference dose (RfD) for arsenic of 0.3 µg/kg/day [19]. For arsenic, the RfD was derived by dividing the identified NOAEL of 0.8 µg/kg/day, obtained from human epidemiologic studies, by an uncertainty factor of three. The LOAEL associated with these epidemiologic studies was 14 µg/kg/day, where exposure to arsenic above this level resulted in hyperpigmentation of the skin, keratosis (patches of hardened skin), and possible vascular complications [8-10].
Arsenic was detected in 43 of the 48 sediment samples; however, only two sediment samples (22 and 23 mg/kg) exceeded the non-cancer HAC value for children (20 mg/kg). The average concentration in the sediment (8.2 mg/kg) was approximately ½ of the HAC value. Arsenic also was detected in 19 of the 136 water samples, and although one sample exceeded the MCL, the average concentration of 1.8 micrograms of arsenic per liter of water (µg/L), is five (5) times lower than the MCL.
Using the suspended sediment data collected by the City of Austin, we considered dermal absorption, incidental ingestion of sediment, and incidental ingestion of water as possible pathways of exposure. (2) A child spending three hours per day in the pool, ingesting 50 ml of water per hour could theoretically receive a combined daily dose of arsenic, from all three pathways, 14 times lower than the RfD, 36 times lower than the NOAEL, and 636 times lower than the LOAEL. Adults could theoretically receive a combined daily dose of arsenic 55 times lower than the RfD, 148 times lower than the NOAEL, and 2,592 times lower than the LOAEL. On the basis of these data, adverse non-cancer health effects from exposure to arsenic in the sediment are not likely.
EPA classifies arsenic as a known human carcinogen on the basis of sufficient evidence from human data. An increase in lung cancer mortality was observed in multiple human populations exposed primarily through inhalation. Also, increased mortality from multiple internal organ cancers (liver, kidney, lung, and bladder) and an increased incidence of non-malignant skin cancers were observed in populations consuming water high in inorganic arsenic [9]. The carcinogenic HAC value for arsenic of 0.5 mg/kg is based on EPA's cancer slope factor (CSF) for skin cancer and an estimated excess lifetime cancer risk of one cancer in 1 million (1 X 10-6) people exposed for 70 years.
Although the average concentration of arsenic found in the sediment was similar to normal background concentrations, it was above the carcinogenic HAC value. To estimate a conservative (with respect to protecting public health) value for the potential excess lifetime cancer risk associated with exposure to arsenic in the sediment and arsenic in the water, we assumed that an individual would ingest 50 milliliters of water for each hour spent in the water and that they would swim, three hours per day, seven days per week, 52 weeks per year. Based on these conservative assumptions we estimate a theoretical excess lifetime cancer risk, from all three pathways of exposure, of 2.8 x 10-6, or about one in 350,000. Qualitatively, we would interpret this to represent an insignificant increased risk. Since the arsenic in the water would be deemed acceptable for drinking and bathing, we have concluded that the water is acceptable for recreational activities. Arsenic does occur naturally in the earth's crust and the reported concentrations are well within those normally found in the environment [9].
Boron
Boron is a naturally occurring element generally found combined with other elements throughout the environment. Boron is neither transformed nor degraded in the environment, but depending on environmental conditions the specific form of boron can change. Natural weathering is a significant source of environmental boron. Boron also can be found in several consumer products, including cosmetics, medicines, and insecticides. Populations residing in the western United States may be exposed to higher than average levels of boron due to natural boron-rich deposits. The most common routes of exposure to boron include ingestion of food (primarily fruits and vegetables) and water; however, occupational exposure to boron dusts also may be significant [11].
To assess the potential health risks associated with the boron in the water we compared the water concentrations to chronic- and intermediate-duration HAC values for non-cancer endpoints. The chronic duration, non-cancer HAC values for boron in water (900 µg/L for children and 3,000 µg/L for adults) are based on EPA's chronic RfD for boron of 0.09 mg/kg/day [12]. For boron, the chronic RfD was derived by dividing the identified NOAEL of 8.8 mg/kg/day, obtained from studies on dogs, by an uncertainty factor of 100. The intermediate duration, non-cancer HAC values for boron in water (100 µg/L for children and 400 µg/L for adults) are based on ATSDR's MRL for boron of 0.01 mg/kg/day [11]. For boron, the intermediate MRL was derived by dividing the identified LOAEL of 13.6 mg/kg/day (associated with reduced fetal body weight in rats) by an uncertainty factor of 1,000. Exposure to boron above this LOAEL was associated with reversible testicular atrophy.
Between 1987 and 1990, Boron was detected in 15 out of 15 water samples. Concentrations ranged from 60 µg/L to 560 µg/L with an average concentration of 149 µg/L. Neither the adult or child chronic HAC values were exceeded. Although the average concentration did exceed the intermediate HAC value for children, this HAC value is based on a child drinking one (1) liter of water per day. A child ingesting 150 ml of the water per day would receive a daily dose seven (7) times lower than the MRL and 9,100 times lower than the LOAEL. It also is important to note that the reported concentrations are well within the concentrations normally reported as background for the western United States. The average surface water concentration in the United States is about 100 µg/L; the concentration in the Western Gulf Basin has been reported to be 289 µg/L; the concentration of boron in seawater is reported to be about 4,500 µg/L. Groundwater concentrations in the western United States have ranged from 140 to 120,000 µg/L with a median concentration of 4,000 µg/L. Based on available information, exposure to the boron found in the water would not be expected to result in adverse health outcomes.
Cadmium
Cadmium is a naturally occurring metallic element that is present in small amounts in virtually all soils and rocks of the earth's crust. It is also present in coal and in both mineral and municipal sludge fertilizers. In its pure form, cadmium is a soft, silver-white metal that is easily cut with a knife. However, cadmium is not usually found in the environment as a pure metal, but instead is combined with other elements such as oxygen, chlorine, or sulfur. Cadmium concentrations in non-polluted soil are highly variable, depending upon sources of minerals and organic materials. The cadmium concentration of natural surface water and groundwater is usually less than 1.0 µg/L; however values of 405 µg/L have been reported [13]. The National Urban Runoff Program measured concentrations ranging from 0.1 to 14 µg/L in urban storm water runoff [13]. Most drinking water supplies in the United States probably do not contain more than 1.0 µg/L of cadmium, but concentrations may reach 10 µg/L.
People are exposed to cadmium everyday. Food is perhaps the major source of human exposure to cadmium in the general, non-smoking population. Average cadmium levels in U.S. food range from 1 to 42 parts of cadmium per billion parts of food (1-42 ppb). Adults consume approximately 30 g of cadmium from food sources each day, absorbing approximately 1 to 3 µg. Of all food items, vegetables generally contain the highest levels of cadmium, particularly potatoes and leafy vegetables (with levels of 42 ppb and 33 ppb respectively). Grain and cereal products also contain elevated levels of cadmium approaching 24 ppb. Meats, fish, and poultry generally contain relatively low levels of cadmium (less than 6 ppb), except for organ meats such as kidney and liver, which, of course, concentrate cadmium and may have levels 10 to 100 times higher [13].
On a national level, inhalation is another major route of exposure to cadmium. Average concentrations in air range from less than 1 ng/m3 in remote areas to 5 to 40 ng/m3 in U.S. urban areas, with isolated measurements of 7,000 ng/m3 in highly contaminated industrial areas [13]. Assuming an average air cadmium concentration of 10 ng/m3 for indoor and outdoor air combined and a daily inhalation rate of 16 m3 for an adult, the average cadmium intake by inhalation is 0.16 µg/day, of which about 25% or 0.04 µg/day will be absorbed. Another major source of inhalation exposure affects smokers, who absorb an additional 1 to 3 µg of cadmium per day for each pack of cigarettes smoked [13]. The toxic effects of chronic cadmium exposure occur primarily in the lungs and in the kidneys. Pulmonary effects are associated solely with inhalation exposures, while the kidney effects may occur after either oral or inhalation exposures. Long-term exposure to excessive cadmium can effect the kidneys, causing proximal tubular necrosis, lesions in the renal cortex, and kidney dysfunction [13, 14].
Cadmium was detected in seven (7) out of 146 water samples from the pool. Values ranged from non-detect to five (5) µg/L, with an average concentration of 1.08 µg/L. Although one sample was equivalent to EPA's MCL for cadmium, the average concentration was five (5) time lower than the MCL and well within the range normally found in surface and groundwater. ATSDR has established a chronic oral MRL for cadmium of 0.0002 mg/kg/day. Using the average concentration we considered both the incidental ingestion and dermal routes of exposure. A child using the pool three hours per day would receive a daily dose over 15 times lower than the MRL (3). An adult who swam three hours per day would receive a theoretical dose approximately 60 times lower than the MRL. Based on the frequency of detection and the reported concentrations, exposure to cadmium in Barton Springs Pool would not be likely to result in adverse health outcomes. Additionally, since the concentration of cadmium in the water would be deemed acceptable for drinking and bathing, we have concluded that the water is acceptable for the types of recreational activities common to Barton Springs Pool.
Bis(2-ethylhexyl)Phthalate
Bis(2-ethylhexyl)phthalate (BEHP) is a manufactured chemical that is commonly added to plastics to make them flexible. It is present in many plastics, especially vinyl materials, which may contain up to 40% BEHP. BEHP is present in plastic products such as wall coverings, tablecloths, floor tiles, furniture upholstery, shower curtains, garden hoses, swimming pool liners, rain wear, baby pants, dolls, some toys, shoes, automobile upholstery, packaging film, sheathing for wire and cable, medical tubing, and blood storage bags [15].
BEHP enters the environment predominantly through disposal of industrial and municipal wastes in landfills. It tends to adsorb strongly to soil and sediment and bioaccumulates in aquatic organisms. Biodegradation does occur under aerobic conditions. When BEHP is present in the environment, it is usually at very low levels; however, it often is difficult to measure these low levels accurately since BEHP is a very common laboratory contaminant, and laboratory contamination may cause false positives.
The principal route of exposure to BEHP is through ingestion. Recently, average total daily ambient exposures to BEHP of 3-30 µg/kg/day (70kg adult) have been proposed. These estimated general population exposures are 3-4 orders of magnitude lower than those observed to cause adverse health effects in animals. Information on the oral toxicity of BEHP in humans is limited to gastrointestinal symptoms (mild abdominal pain and diarrhea) in two individuals who ingested a single large dose of the compound. There is limited information available on the health effects of BEHP in humans or animals following dermal exposure. One dermal study found no indications of skin irritation or sensitization in humans or rabbits.
BEHP was detected in water from Barton Springs Pool in one (1) out of 11 samples. Concentrations ranged from non-detect to 5.6 µg/L with an average concentration of 3.9 µg/L. Although the concentration in this sample exceeded the CREG, the estimated excess lifetime cancer risk associated with long term exposure via both the dermal and incidental ingestion routes of exposure (5.8 x 10-7) would be deemed insignificant. (4) Additionally, the sample concentration was below the MCL which is 6.0 µg/L. Since the concentration of BEHP in the water would be deemed acceptable for drinking and bathing, we have concluded that the water is acceptable for recreational activities. Based on available information, the BEHP found in water from Barton Springs Pool would not be likely to cause adverse health outcomes.
Total Petroleum Hydrocarbons
Water and sediment from Barton Springs were periodically tested for Total Petroleum Hydrocarbons (TPH). TPH is a term used to describe a broad family of several hundred compounds that originate from crude oil. Thus, TPH is really a mixture of many chemicals all of which are made almost entirely from hydrogen and carbon. The exact nature of the chemicals within any measurement of TPH varies with the specific products from which they originate. Most products that contain TPH will burn and many have characteristic gasoline, kerosene, or oily odors. Because of the common use of petroleum based products, contamination of the environment with TPH is quite common [16]. TPH is a common pollutant associated with urban runoff and at one public meeting, a citizen expressed a concern about sporadically detecting a gasoline type smell in the pool.
While the measurement of TPH may be a good general indicator of petroleum contamination (i.e., used to estimate the effects of urban runoff), the large and varying number of hydrocarbons associated with any specific measurement of TPH makes it difficult to use this measurement to determine how the particular petroleum hydrocarbons in the sample may affect people. Although there is no one universally accepted method to make such an assessment, there are some general guidelines that can be applied, the application of which often may be affected by the specific nature of the samples.
One approach is to measure and assess individual compounds and while it is impractical to measure all the compounds potentially associated with TPH, it is reasonable to measure and assess surrogate compounds. Generally, some of the more common toxicologically significant compounds are chosen as surrogates. A number of the compounds that measured both in water and sediment from Barton Springs Pool are compounds often included in TPH. We were able to assess many of these compounds individually. These compounds include alkyl-benzenes (benzene, biphenyl, butylbenzene, ethylbenzene, styrene, toluene, xylene), naphtheno-benzenes (acenaphthene, acenaphthylene, benzo[b]fluoranthene, benzo[k]fluoranthene, fluoranthene, fluorene, indeno[1,2,3-cd]pyrene), alkyl-naphthalenes (naphthalene, 2,6-dimethylnaphthalene), and polynuclear aromatics (anthracene, benzo[a]anthracene, benzo[a]pyrene), benzo[e]pyrene, benzo[g,h,i]perylene, chrysene, perylene, phenanthrene, and pyrene).
In addition to the assessing individual compounds we also were able to compare the PAH concentrations measured in both water and sediment to TCEQ media specific PCLs. The PCLs consider both the dermal and ingestion routes of exposure and for TPH they are available for both the aliphatic and aromatic components of TPH and vary with the length of the carbon chain (Table 5).
Petroleum hydrocarbon measurements broken down by carbon chain length were not available for water; however, TPH was detected in eight (8) out of 31 water samples. Concentrations ranged from non-detect to 14,000 µg/L, with an average concentration of 1,093 µg/L; this is well within the range of PCLs available for TPH in groundwater (700 to 48,900 µg/L). For sediment, measurements were available both for TPH and for three fractions (C12-C28, C6-C12, and C6-C35). TPH was found in 16 of the 18 sediment samples for which this parameter was measured. Reported concentrations ranged from non-detect to 1,100 mg/kg, with an average concentration of 177 mg/kg. The concentrations reported for the three fractions ranged from non-detect to 70 mg/kg, with average concentrations of 53.5 mg/kg, 40.1 mg/kg, and 64.8 mg/kg for the three fractions, respectively. All reported concentrations were well below the sediment PCLs and would not be expected to result in adverse health outcomes.
Thallium
Thallium is a naturally occurring element that is widely distributed in trace amounts in the earth's crust. It can be found in pure form or combined with other substances such as bromine, chlorine, fluorine, and iodine to form salts [17]. Thallium is present in air, water, and soil.
Thallium is used mainly in the manufacture of electronic devices, switches, and closures. It also has limited use in the manufacture of special glasses and for medical procedures that evaluate heart disease. Up until 1972, thallium was used as a rat poison. Thallium is no longer produced in the United States [17].
Although thallium is present at low levels in air water and soil, the greatest exposure occurs through food, mostly home-grown fruits and green vegetables contaminated with thallium. It has been estimated that the average person eats approximately two parts thallium per billion parts of food. Cigarette smoking also is a source of thallium exposure [17].
Exposure to too much thallium can result in adverse health effects. Exposure to large amounts over a short time can affect the nervous system, lung, heart, liver, and kidney. Temporary hair loss (alopecia), vomiting, and diarrhea also can occur. There is little information on the effects in humans after long term exposure to smaller amounts of thallium; however, based on rodent studies, the EPA has established chronic oral RfDs for various thallium salts that range from 0.00008 mg/kg/day to 0.00009 mg/kg/day. These values are based on an identified NOAEL for increased liver enzymes of 0.26 mg/kg/day and an uncertainty factor of 3,000.
Thallium was detected in one (1) of six (6) samples at a concentration of 6.0 µg/L, a concentration above the MCL (2.0 µg/L). Using ½ the detection limit for the non-detect values, the average concentration of 3.9 µg/L also was above the MCL. Considering both the dermal and incidental ingestion routes of exposure we estimate that a child swimming three hours per day, ingesting 150 ml of pool water per day could receive a daily dose of approximately 0.00004 mg-thallium/kg/day (5), a value below the range of RfDs established for thallium compounds. Based on the low frequency of detection and the reported concentrations, exposure to thallium in the water at Barton Springs Pool would not be expected to result in adverse health effects.
Lead
Lead occurs naturally in the environment; however, most of the high levels found throughout the environment come from human activities. It is used in batteries, in the production of ammunition, in some kinds of metal products, and in ceramic glazes. In the past lead also was used in housepaint and gasoline; however, its use in these products has been phased out.
Preschool-age children and fetuses are usually the most vulnerable segments of the population for exposure to lead. This increased vulnerability results from a combination of factors which include the following: 1) the developing nervous system of fetuses and neonates are more susceptible to the neurotoxic effects of lead; 2) young children are more likely to play in dirt and to place their hands and other objects in their mouths thereby increasing the opportunity for soil ingestion; and, 3) the efficiency of lead absorption from the gastrointestinal tract is greater in children than in adults. The overall half-life of lead in blood is estimated to be 36 days ± 5 days [18]. In the United States, leaded paint continues to cause most of the severe lead poisoning in young children because it is the most widespread source and has the highest concentration of lead per unit of weight [18].
Chronic exposure to low lead levels has been shown to cause subtle effects on the central nervous system which manifest as deficits in intelligence, behavior, and school performance [18]. Available evidence is not sufficient to determine whether lead-associated deficits are irreversible. Although no threshold level for adverse health effects has been established, evidence suggests that adverse neurological and cognitive deficits occur at blood lead levels at least as low as 10 µg/dL. The Centers for Disease Control and Prevention (CDC) has determined that a blood lead level greater than or equal to 10 µg/dL in children indicates excessive lead absorption and constitutes the grounds for intervention. The 10 µg/dL level is based on observations of enzymatic abnormalities in the red blood cells at blood levels below 25 µg/dL and observations of neurologic and cognitive dysfunctions in children with blood lead levels between 10 and 15 µg/dL.
Lead is especially harmful to unborn children. Infants often are born with some lead in their bodies due to their mother's past exposure to lead. Exposure to lead during pregnancy has been correlated with premature births, low birth weight infants, and spontaneous abortions. While the impact of maternal and cord blood lead levels below 10 µg/dL have not been well-defined, reduced gestational age and reduced birth weight have been associated with blood lead levels of 10 to 15 µg/dL. In addition, lead has been found to lower intelligence quotient (I.Q.) scores, slow growth, and cause hearing problems in children. These adverse effects can persist and lead to decreased performance in school.
The most serious effect of acute high dose lead exposure is encephalopathy which is characterized initially by headache and drowsiness, and in more severe cases by coma, convulsions, and death. Virtually all children who recover from acute lead encephalopathy exhibit residual reduction in intelligence and behavioral dysfunction. Acute encephalopathy is usually associated with high blood lead levels (over 150 µg/dL). Another effect of acute high dose lead exposure is the Fanconi syndrome, an acute injury to the renal tubules, characterized by spillage of glucose, protein, amino acids, and phosphates into urine.
Anemia is the most serious effect of lead on the hematologic system. Lead-induced anemia occurs primarily by the lead-induced inhibition of several enzymes involved in the production of hemoglobin. Exposure to lead has been associated with hypertension, renal failure, and gout. Lead has not been shown to be carcinogenic in humans; however, high doses of lead have been found to produce kidney tumors in laboratory studies of rats and mice. The extremely high cumulative doses of lead used in animal studies are difficult to extrapolate to low-level exposure in humans, and do not provide a sufficient basis for quantitative risk assessment.
Lead was found in 34 of the 184 water samples taken from Barton Springs Pool at concentrations ranging from non-detect to 28.6 µg/L. It is important to note that in some instances the detection limits were higher than the highest reported value. The highest reported detection limit was 50 µg/L. Computing an average using the actual values and all non-detect represented as ½ the detection limit, the average concentration was 8.7 µg/L. EPA has established an action level for lead of 15 µg/L for public drinking water systems. This action level, which was established to reduce the lead level in drinking water at the consumers' tap, requires water system authorities to take specific actions when more than 10% of targeted tap samples exceed the action level. The estimated average concentration is below the action level for lead in drinking water. At Barton Springs Pool, only four (4) of the 184 samples actually exceeded the drinking water action level. Using EPA's Integrated Exposure Uptake Biokinetic Lead Model for children we estimate that chronic ingestion of drinking water containing 8.7 µg-lead/L would have an insignificant effect on a child's blood lead level. (6) The types of exposures possible at Barton Springs Pool would not be expected to result in adverse health outcomes.
We reviewed the results from water and sediment samples collected by: the City of Austin, the United States Geological Survey, the Lower Colorado River Authority, and the Texas Commission on Environmental Quality. This consisted of over 14,500 individual data points, involving about 441 analytes, collected over the past 12 years.
We screened the contaminants by comparing them to health-based screening values and selected twenty-seven (27) contaminants for further consideration. This screening process consisted of selecting all analytes for which any single value exceeded the screening level. Of the 27 contaminants, 20 were PAHs. The others included arsenic, boron, cadmium, bis(2-ethylhexyl)phthalate, total petroleum hydrocarbons, thallium, and lead.
The most likely route of exposure to the contaminants found in the pool is incidental ingestion of sediment while swimming. Based on our visit to the site, we would not expect absorption of contaminants through the skin by contact with sediment to be a major contributor to exposure. Even though children regularly play in the pool, the availability of sediment in the areas frequented by children is low. Additionally, the relative capacity of the contaminants, particularly the PAHs, to adhere to the sediment preferential to skin, further reduces this as a viable exposure pathway. For contaminants in the water we considered both the ingestion and dermal routes of exposure.
For most of the contaminants selected, the frequency of detection was low; however, in some instances this could have been an artifact of the detection limits used for some of the analyses. Thus, we did not omit contaminants from selection based on frequency of detection and we used ½ the detection limit for all reported non-detects in our calculation of the averages.
Uncertainties
General Uncertainties
In preparing this report, we relied on the information provided and assumed adequate quality assurance/quality control (QA/QC) procedures were followed with regard to data collection, chain-of-custody, laboratory procedures, and data reporting. The analysis and conclusions in this report are valid only if the referenced information is valid and complete.
There always will be uncertainties associated with any presentation of risk. These uncertainties include toxicologic uncertainties, exposure uncertainties, and data uncertainties. We tried to account for some of these uncertainties through the use of conservative assumptions. That is, we chose assumptions that err on the side of protecting the public. For instance, we assumed that exposures would occur 365 days per year for 70 years; actual exposures would be less frequent and for shorter duration. We also assumed that 100% of every contaminant that came into contact with the body was absorbed by the body. While in rare instances this may occur, in general, absorption rates would be much less.
The original concerns regarding this site focused on possible exposure to PAHs; however, other contaminants were measured during the course of the various investigations. To provide a complete analysis of the public health implications of the contaminants measured in the pool we reviewed all available data; this consisted of the examination of over 14,500 individual data points. Determining public health implications of possible exposure to multiple contaminants will always result in some uncertainty, particularly when simultaneous exposure to a variety of contaminants may occur. Such exposures could be additive, multiplicative, or even subtractive. To some limited extent we have addressed this issue in our determination of the risks associated with exposure to the PAHs. Although we can not address all the contaminants in this manner, many of the contaminants were found at concentrations several orders of magnitude below the levels at which adverse health effects have been observed. Even when we combined the cancer risk estimates from all pathways for all contaminants, the cancer risk estimates would still be considered to be insignificant. (7) In some instances we compared the concentrations of contaminants found in pool water to public drinking water standards. While there are many factors that are considered when establishing these standards, such a comparison assumes that if water is deemed safe for drinking, cooking, and bathing; it should be safe for recreational uses such as swimming.
There always will be some uncertainties associated with the contaminants for which we were not able to find HAC values or other toxicological information. Fortunately, this was limited to only four (4) contaminants; A, A'-dimethylphenyethylamine, hentriacosane, nonacosane, and butyl hexadecanoate. These uncertainties are somewhat reduced by the fact that each of these contaminants only was found in one sample and at low concentrations. Given the relative toxicity of other contaminants for which HAC values were available, the low frequency of detection of these contaminants, and the types of exposure scenarios that are likely to be encountered, we do not believe that these contaminants would pose a public health hazard.
Specific Uncertainties
There are specific uncertainties associated with some of the contaminants that we considered. For instance, there is considerable controversy associated with any estimate of risk (non-cancer or cancer) associated with exposure to arsenic. Both the RfD and the cancer slope factor (CSF) are based on human ecological studies that have recognized uncertainties with respect to exposure. In such studies it is difficult to avoid errors in assigning people to specific exposure groups. The studies upon which the RfD and the CSF are based also involved exposure to arsenic in drinking water. While we did consider arsenic in water, we also considered the risk associated with arsenic in sediment. The ability of the body to absorb arsenic in water is likely to be higher than the ability of the body to absorb arsenic in sediment. In our analysis we assumed that the arsenic in the sediment was 100% bioavailable. Studies conducted for EPA at various Superfund sites have found the bioavailability of the arsenic in the soil to be lower than 100 percent. Thus, assuming 100 % absorption is very conservative with respect to protecting public health and to some unknown degree overestimates the risk. We also did not consider the kinetics of arsenic in the body in our risk estimates. The RfD and the CSF are based on daily exposures over a lifetime. Since the half-life of arsenic in the body (the time it takes one-half of the arsenic to be excreted) is short (40-60 hours), the risk estimates for exposures that occur less frequently than everyday also may result in an overestimate of the risks.
With specific respect to the cancer risk estimates, the mechanisms through which arsenic causes cancer are not known; however, arsenic is not thought to act directly with DNA. Since the studies used to derive the CSF are based on exposure doses much higher than those likely to be encountered at this site, it is questionable whether it is appropriate to assume linearity for the dose-response assessment for arsenic at low doses. The actual dose-response curve at low doses may be sublinear which would mean that the above risk estimates overestimate the actual risks.
The use of toxicity equivalency factors (TEF) in our estimation of the risks associated with exposure to the PAHs also contributes a certain degree of uncertainty. While this is considered an acceptable practice, how well this procedure truly represents the toxicity is unknown.
Perhaps one of the biggest sources of uncertainty is associated with estimating risks associated with TPH. The shear number of possible contaminants that could be associated in any measurement of TPH makes trying to determine how the particular petroleum hydrocarbons in the sample may affect people very difficult. Where possible our evaluation individually considered some of the more toxic components of TPH. Additionally, the screening values that we used for TPH itself were, for the most part, based on the lowest ones available for surrogate compounds.
Children's Health Considerations
TDH and ATSDR recognize that the unique vulnerabilities of children demand special attention. Windows of vulnerability (critical periods) exist during development, particularly during early gestation, but also throughout pregnancy, infancy, childhood and adolescence -- periods when toxicants may permanently impair or alter structure and function [19]. Unique childhood vulnerabilities may be present because, at birth, many organs and body systems (including the lungs and the immune, endocrine, reproductive, and nervous systems) have not achieved structural or functional maturity. These organ systems continue to develop throughout childhood and adolescence. Children may exhibit differences in absorption, metabolism, storage, and excretion of toxicants, resulting in higher biologically-effective doses to target tissues. Depending on the affected media, they also may be more exposed than adults because of behavior patterns specific to children. In an effort to account for children's unique vulnerabilities, and in accordance with ATSDR's Child Health Initiative [20] and EPA's National Agenda to Protect Children's Health from Environmental Threats [21], TDH used the potential exposure of children as a guide in assessing the potential public health implications of the contaminants.
Actions Recommended
1. For statistical purposes samples below detection limits (DL) were assumed to be 0.25 mg/L (½ DL)
2. Dermal dose(water) (mg/kg/d) = 1.0 x 10-3 (cm/h) x 1.8 x 10-6 (mg/cm3) x 3 (h/d) x 10,000 (cm2)/15 (kg)
Ingested dose(water) (mg/kg/d) =( 0.0018 (mg/L) x 0.15 (L))/15 (kg);
Ingested dose(sed) (mg/kg/d) = (8.2 (mg/kg) x 1.71 x 10-7 (kg))/15 (kg); Dermal dose for adults was calculated using a body surface area of 20,000 cm2 and a body weight of 70 kg; Risk = ((Dermal dose + Ingested dose sed + Ingested dose water) x 1.5 (mg/kg/day)-1 ) x ((300 (d/y) x 30 (y))/(365 (d/y) x 70 (y))).
3. Dermal dose (mg/kg/day) = (1.0 x 10-3 (cm/h) x 1.08 x 10-6 (mg/cm3) x 3 (h/d) x 10,000 cm2)/15 (kg)
Ingested dose (mg/kg/day) = (0.00108 (mg/L) x 0.05 (L/h) x 3 (h))/15 (kg)
4. Estimated excess lifetime cancer risk of 5.8 x 10-7 based on incidental ingestion and dermal pathways. Dermal dose (mg/kg/day) = (0.033 (cm/h) x 3.9 x 10-6 (mg/cm3) x 3 (h/d) x 20,000 (cm2))/70 (kg);
Ingested Dose (mg/kg/d) = (0.0039 (mg/L) x 0.05 (L/h) x 3 (h))/70 (kg)
Risk = ((Dermal dose + Ingested Dose) x 0.014 (mg/kg/day)-1 ) x ((300 (d/y) x 30 (y))/(365 (d/y) x 70 (y)))
5. Dermal dose (mg/kg/day) = (1.0 x 10-3 (cm/h) x 3.9 x 10-6 (mg/cm3) x 3 (h/d) x 10,000 cm2)/15 (kg)
Ingested dose (mg/kg/day) = (0.00039 (mg/kg) x 0.15 (L))/ 15 (kg)
6. Using default parameters for all other sources of exposure the model predicts that chronic exposure to drinking water containing 8.7 µg-lead/L would result in an increase in a child's blood lead of less than 1.0 µg/dL.
7. Cancer risk estimates for all contaminants PAHs, arsenic, and BEHP combined, all pathways, ranged from 8.0 x 10-6 (300 days/year for 70 years) to 1.9 x 10-7 (50 days/year for 10 years).
Next Section Table
of Contents
