PETITIONED PUBLIC HEALTH ASSESSMENT
ST. THOMAS, U.S. VIRGIN ISLANDS
The Agency for Toxic Substances and Disease Registry (ATSDR) evaluates contaminants detected in environmental media at the site and determines whether exposure to the contaminants has public health significance. ATSDR uses the following factors to select contaminants for discussion:
- concentrations of contaminants on and off site,
- community health concerns, and
- comparison of on- and off-site concentrations with ATSDR health comparison values for (1) noncarcinogenic endpoints and (2) carcinogenic endpoints.
ATSDR health comparison values are concentrations of contaminants that are media specific (e.g. specific to water, air, or soil). The comparison values are considered to be safe under default conditions of exposure and are used as screening values in the preliminary identification of site-specific contaminants of concern. Contaminants of concern are those contaminants detected above the screening comparison values and contaminants without comparison values. The fact that a contaminant is mentioned in this section does not mean that site-specific exposure to the substance will result in adverse health effects. It means, rather, that the contaminant will be evaluated in subsequent discussions in the document. Please refer to Appendix C for additional clarification and a description of the comparison values used in this public health assessment.
Following the preliminary identification of site-specific contaminants of concern which are described in this section, ATSDR staff determine whether workers and nearby residents are exposed to contamination migrating from the site. Discussions of migration appear in the Pathways Analyses Section of this public health assessment. If exposure to contamination is identified, the Toxicological Evaluation Section of this public health assessment contains discussion of the significance of this exposure and its relation to adverse health effects. ATSDR staff also address specific community concerns in the Community Health Concerns Evaluation section. Finally, based on the evaluations from all preceding sections of the public health assessment, ATSDR staff determine conclusions and prepare recommendations.
The remainder of this section will discuss the sampling of each environmental medium separately. Appendix B contains tables that list the contaminants detected in each environmental medium. Contaminants listed in the tables will not necessarily cause adverse health effects at the levels detected. Instead, the list indicates which contaminants an ATSDR toxicologist will evaluate in subsequent parts of this document.
Six wells were sampled during on-site groundwater monitoring at the landfill from July 1982 through March 1983 . Arsenic, cadmium, chromium, lead, and silver exceeded ATSDR comparison values. Please refer to Table Three, Appendix B, for the results of this groundwater monitoring.
Two monitoring wells were installed and sampled in September 1996 as part of the field investigation at the landfill. Please refer to Figure Four, Appendix A, for a location map of the monitoring wells. Two additional wells were planned, but attempts to sample the planned locations yielded fractured and weathered rock and no groundwater . The samples were analyzed for volatile organic compounds, semivolatile organic compounds, organochlorine pesticides, polychlorinated biphenyls, cyanide, phenolics, and metals. All concentrations were below ATSDR comparison values. Please refer to Table Three, Appendix B, for the results of this groundwater monitoring.
The Army Corps of Engineers removed debris from the Bovoni Landfill after Hurricane Marilyn from October 1995 through April 1996. Air monitoring at designated worker areas at the landfill in November 1995--not during an above-ground fire event--assessed worker exposures.
Figure Three, Appendix A, contains a location map of the air monitoring areas for the November 1995 sampling event. The Upwind monitoring area was upwind of the landfill, approximately 500 feet east of the road bordering the east side of the working landfill. This location provided background concentrations of compounds. The Downwind monitoring area was located relatively downwind of the landfill, approximately 8 feet west of the dirt path bordering the west edge of the landfill. This location was approximately 20 feet above the level area, across the top of the landfill between the burnable debris pile and the daily trash area. The Smolder Pit monitoring area was approximately 8 feet west of the Smolder Pit and was directly in the line of escaping smoke from several fissures in the pit. Monitoring at this location determined the types and quantities of pollutants being emitted from the open fissures. The Burn Pit monitoring area was approximately 50 feet west of the operating burn pit and was approximately 20 feet lower than the burn pit because of the slope of the terrain. This location was in the direct line of smoke from the burn pit and of dust emissions from burn pit and debris-sorting operations . Please refer to Table Four, Appendix B, for the 1995 air monitoring results.
Air monitoring was also performed during the field investigation in 1996. Please refer to Figure Four, Appendix A, for a location map of the air monitoring areas. Grab samples (for volatile organics, carbon monoxide, phosgene, mercury and hydrogen sulfide) and short-term samples (for particulates) were taken both upwind and near vents at the landfill . Upwind samples were collected to determine background concentrations. Vent samples were collected directly from smoking vents to determine landfill emissions. A vent measurement at the landfill yielded a maximum concentration of 6% of the lower explosive limit for methane . The sampling was not conducted during an above-ground fire event. Please refer to Table Five, Appendix B, for the 1996 air monitoring results.
Investigators sampled sediment from the Mangrove Lagoon and Benner Bay in the 1970s . The samples consisted mainly of bulk surface samples and a few core samples. Total phosphorus exceeded relevant ATSDR comparison values. Please refer to Table Six, Appendix B, for the results of this sampling.
In preparing public health assessments, ATSDR relies on environmental data in referenced documents. ATSDR staff assumes that adequate quality assurance and quality control measures were followed regarding chain-of-custody, laboratory procedures, and data reporting. The analyses, conclusions, and recommendations in public health assessments are valid only if the referenced documents are complete and reliable.
For this public health assessment, ATSDR staff obtained and evaluated environmental data for groundwater, air and sediment. However, it is important to note that ATSDR staff were not able to obtain the complete reports to review for Reference 7, Reference 9, Reference 10 and Reference 11. Therefore, the information and data provided on groundwater collected in 1982/83 (obtained from Reference 7) and on sediment collected in the 1970s (obtained from Reference 9) are questionable. ATSDR staff have included the information in this report even though the reliability of this information is questionable because of the lack of environmental data for this site. ATSDR staff were informed that many records on the landfill were destroyed during hurricanes. Because of the lack of environmental data, ATSDR staff considered it necessary to include all data for public health significance, regardless of concerns about the quality of the data.
Also of note, phosgene was detected at a maximum of 0.19 parts per million (ppm) 15 feet downwind from the Smolder Pit area and near several fissures, and at 0.75 ppm 6 to 12 inches away from one of the fissures in the smoke. The referenced report noted that the analyzer utilized has several known interferents, including trimethylamine, a gas associated with decomposing plant and animal materials . Mercury vapor was detected at 0.078 milligrams per cubic meter (mg/m3) at the burnable debris pile and at 0.121 mg/m3 at the Burn Pit area. The referenced report noted that the analyzer utilized has several interferences including hydrogen sulfide, chloride compounds, and temperature changes .
Access to the site is restricted by fence and natural barriers; however, trespassing has been reported. One trespasser accidentally started the March 1996 tire fire while using a blowtorch to try to remove a fender from a junked car. Trespassers have also been caught taking food and beverages from the food rubbish area . Other hazards include speeding trucks, unstable ground created by voids beneath the landfill surface, and explosions . ATSDR staff noted during the August 1996 site visit that several refrigerators still had the doors attached; this could pose a physical hazard if children trespass on-site and are trapped inside.
To determine whether nearby residents are exposed to contamination migrating from the site, the Agency for Toxic Substances and Disease Registry (ATSDR) evaluates the environmental and human components that lead to human exposure. An exposure pathway contains the following five elements: a source of contamination, transport through an environmental medium, a point of exposure, a route of human exposure, and an exposed population.
ATSDR categorizes an exposure pathway as a completed, potential, or eliminated exposure pathway. Completed pathways require that the five elements exist and indicate that exposure to the contaminant has occurred, is occurring, or will occur. Potential pathways are those in which at least one of the five elements is missing but could exist. Potential pathways indicate that exposure to a contaminant could have occurred, could be occurring, or could occur in the future. In an eliminated exposure pathway, at least one of the five elements is missing and will never be present. The discussion that follows describes only those pathways that are relevant to the site. Table One and Table Two, Appendix B, present the completed and potential pathways and their elements.
Contamination of the ambient air creates a past, current, and future completed exposure pathway for on-site workers. The underground fire at the landfill emits combustion products to the ambient air through fissures in the landfill surface. Thirty individuals, including approximately 10 Department of Public Works employees and 20 contractors, work at the landfill . Air monitoring data indicate 1,2,3-trichloropropane; 1,1,2-trichloroethane; benzene; acetaldehyde; acrolein; mercury vapor; arsenic; nickel; ethylbenzene; and m,p-xylene exceeded relevant comparison values. Please refer to Table Four and Table Five, Appendix B.
Residents living near the Bovoni Landfill might have been exposed to contamination during the above-ground fire events and the tire fire. No data exist off-site to confirm this exposure; however, there are many reports of black smoke covering the island during these episodic events. Discussions with local authorities indicate that approximately 1,600 people live in the locality of the landfill. Because of the nature of prevailing winds, it is primarily the 100 to 300 people in the Bolongo Bay area west of the landfill that have experienced smoke and fumes from the periodic eruptions of the above-ground fire at the landfill.
The primary route of exposure is inhalation. Sensitive populations in the area are children, residents with chronic medical problems, the elderly, and pregnant and nursing mothers. Other groups of concern include the landfill workers and the firefighters who have fought the various eruptions of the fire.
Residents can obtain groundwater for domestic purposes through groundwater wells pumped by private water haulers that fill household cisterns or private wells. No private water-hauling companies are located in the vicinity of the landfill. Residents using private wells for drinking water and other domestic purposes could experience exposure to landfill contaminants. Ingestion, inhalation, and direct skin contact with private well water could be routes of exposure. No private wells are known to exist in the locality of the landfill; however, ATSDR staff conducting the investigation were unable to confirm this lack of private wells. Although sampling data from on-site monitoring wells in the past (1982/83) indicated that metals exceeded ATSDR comparison values, 1996 sampling data from on-site wells indicate concentrations below comparison values. Because of the difficulty securing groundwater in the faulted and fractured rock in the Bovoni Landfill area and the determination that the flow of groundwater at the landfill is toward the east, it is unlikely that landfill contamination could affect private water wells to the north/northeast. Based on available data, ATSDR staff determined that human exposure to groundwater in the Bovoni Landfill area is unlikely and that no current contaminant concentrations are at levels of health concern.
Most households in the vicinity of the landfill obtain fresh water from rooftop rain catchments that fill household cisterns. Exposure to contaminants could occur for residents using water from household cisterns for drinking and other domestic purposes. If contamination were present, routes of exposure would include ingestion, inhalation, and direct skin contact with cistern water. No cistern water sampling data exist; however, the impact of combustion products from fires at the landfill on cisterns in the area is expected to be minimal, and no significant exposure would be expected from utilizing this water . Please refer to the Community Health Concerns Evaluation Section of this public health assessment for further information on cistern water.
Sampling data from the 1970s exist for the Mangrove Lagoon and Benner Bay . Individuals use the outer Mangrove Lagoon for recreational activities, including boating and fishing. Exposures to contaminants in sediment could occur during these recreational activities. Direct skin contact and incidental ingestion could be routes of exposure. Because such exposure is intermittent, and it is unlikely that recreational users would have ingested sediments in quantities of concern, the contaminants detected in sediment are not considered to be at levels of health concern.
This section addresses the likelihood that exposure to on-site contaminants at the maximum concentrations detected would result in adverse health effects. While the relative toxicity of a chemical is important, the response of the human body to a chemical exposure is determined by several additional factors: the concentration (how much); the duration of exposure (how long); and the route of exposure (breathing, eating, drinking, or skin contact). Lifestyle factors (i.e., occupation and personal habits) have a major impact on the likelihood, magnitude, and duration of exposure. Individual characteristics such as age, sex, nutritional status, overall health, and genetic constitution affect how a human body absorbs, distributes, metabolizes, and eliminates a contaminant. A unique combination of all these factors will determine the individual's physiological response to a chemical contaminant and any adverse health effects the individual may suffer as a result of the chemical exposure.
The Agency for Toxic Substances and Disease Registry (ATSDR) has determined levels of chemicals that can reasonably (and conservatively) be regarded as harmless, based on the scientific data the agency has collected in its toxicological profiles. The resulting comparison values and health guidelines, which include ample safety factors to ensure protection of sensitive populations, are used to screen contaminant concentrations at a site and to select substances that warrant closer scrutiny by agency health assessors and toxicologists. Please refer to Appendix C for a more complete description of ATSDR's comparison values, health guidelines, and other values ATSDR uses to screen site contaminants. It is important to understand that ATSDR's comparison values and health guidelines do not represent thresholds of toxicity. Thus, although concentrations at or below ATSDR's comparison values may reasonably be considered safe, it does not automatically follow that any concentration above a comparison value will produce toxic effects. To the contrary, ATSDR's comparison values are intentionally designed to be lower, usually orders of magnitude lower, than the corresponding no-effect levels (or lowest-effect levels) determined in laboratory studies.
When screening individual contaminants, ATSDR typically compares the lowest comparison value available, typically the cancer risk evaluation guide (CREG) or chronic environmental media evaluation guide (EMEG) for the most sensitive, potentially exposed individuals (usually children or pica children) with the highest single concentration of a contaminant detected at the site. This high degree of conservatism results in the selection of many contaminants as chemicals of concern that will not, upon closer scrutiny, be judged to pose any hazard to human health. However, ATSDR judges it prudent to use a screen that lets through many harmless contaminants rather than one that overlooks even a single potential hazard to public health. The reader should keep in mind the conservativeness of this approach when considering the potential health implications of ATSDR's toxicological evaluations.
Since a contaminant must first enter the body before it can produce any effect, adverse or otherwise, on the body, this evaluation will focus solely on those contaminants associated with completed pathways of exposure. The only completed pathway documented at the Bovoni Landfill is air. All currently available site-specific air data are from on-site monitoring performed at worst-case locations (i.e., at vents, burn pits, etc.), and at times when no major landfill fire was raging at the surface. The general absence of air data for the breathing zone hampers efforts to estimate exposure doses onsite. Furthermore, no data at all were available for airborne contamination offsite. The only safe assumption is that the concentrations of contaminants in air will be substantially lower offsite (i.e., 500 ft to 1 mile away from the source) than they are on-site.
ATSDR predicts that all exposures to airborne contaminants at the Bovoni Landfill are likely to be intermittent and of intermediate duration (or less), since they will depend, from moment-to-moment, on (1) the activity of the landfill itself in relation to surface fires, (2) the type and concentration of waste fueling the fires both above and below the surface, (3) The direction of the wind and how hard it is blowing, and (4) the specific location and duration of worker activity on-site. Therefore, the data tables list only acute and intermediate comparison values, where they are available, and chronic comparison values where they are not. Given that exposure scenario, ATSDR's CREGs would not be applicable. CREGs are based on numerical risk estimates, such as the Environmental Protection Agency's (EPA) cancer slope factors which, in turn, are based on certain assumptions (e.g., chronic, lifetime exposure and the absence of a threshold) that are not always relevant to real world situations. Although ATSDR recognizes the utility of numerical risk estimates, agency staff consider these estimates in the context of the variables and assumptions involved in their derivation and in the broader context of biomedical opinion, host factors, and actual exposure conditions . For those compounds with no ATSDR comparison value other than the CREG, ATSDR staff used an appropriate (i.e., noncancer) health benchmark from another agency, such as an EPA Region III risk-based concentration (RBC). Finally, staff used the National Institute for Occupational Safety and Health and Occupational Safety and Health Association (OSHA) limits for occupational exposure when there was no RBC.
Thus, while staff initially screened all contaminants at the Bovoni Landfill by comparing their maximum detected concentrations to the lowest available comparison values, as described earlier, only comparison values with some relevance to site-specific exposure scenarios are listed in Tables 3-6 in Appendix B. This approach is intended to provide the reader with a meaningful perspective on potential health hazards at this site. Most on-site contaminants were detected at levels that were below comparison values that are based on acute and intermediate effects.
CONTAMINANTS IN AIR AT BOVONI LANDFILL, 1995 AND 1996
Apart from the obvious physical hazards (e.g., speeding trucks, unstable earth, potential fires and explosions) and biological hazards (e.g., medical waste disposed improperly, tires that can serve as breeding areas for disease-carrying insects, and access to the food refuse pile), the primary health hazards at Bovoni Landfill are the respiratory irritants (both chemicals and smoke) generated by the subterranean fire that occasionally breaks through to the surface. Air monitoring data from 1995 and 1996 show concentrations of individual chemicals below their respective thresholds for respiratory irritation during an 8- to 12-hour (hr) workday. However, the irritant effects of these chemicals (e.g., phosgene and various aldehydes) may be additive, and more sensitive individuals will be at greater risk than less sensitive individuals. Nearby residents, as well as on-site workers, will be at greatest risk during a surface landfill fire, i.e., when concentrations of chemical respiratory irritants may be further elevated and smoke inhalation will exacerbate their effect. Some workers, especially those with prolonged exposure, may also experience respiratory effects between fires. However, contaminants at Bovoni Landfill are not likely to produce adverse health effects in nearby residents except possible effects during major surface fires. At such times, the increased risk will be due primarily to the potential for smoke inhalation augmented by the (potential) generation of increased concentrations of phosgene, aldehydes, and other chemical respiratory irritants. The following discussions describe contaminants detected at maximum concentrations exceeding one of the more site-relevant comparison values listed in Appendix B.
The maximum detected concentration of acetaldehyde (10 micrograms [ug]/cubic meter [m3] in the Burn Pit area) exceeded ATSDR's only noncancer comparison value for this compound (EPA's response dose concentration [RfC] of 9 ug/m3, which contains a safety factor of 1,000) only marginally (11%). The RfC is based on the assumption of chronic exposure over an entire lifetime, and on-site exposures at Bovoni Landfill will be intermittent and of relatively short duration. Therefore, ATSDR does not consider the concentrations of acetaldehyde measured at Bovoni Landfill to represent any hazard to public health.
Acrolein vapor is a direct irritant with excellent warning properties. It has a sharp, disagreeable odor and causes irritation of the eyes and upper respiratory tract. The maximum concentration of acrolein detected on-site (6 ug/m3 or 2.6 parts per billion [ppb] at the Burn Pit) was well below the odor threshold (50-500 ug/m3), and one two-hundredth of the levels that produce irritation (1,250 ug/m3 ) [15, 16]. ATSDR's Toxicological Profile for acrolein puts the odor threshold at 160 ppb (369 ug/m3) and levels that are irritating to the eyes, nose, and throat at 170 ppb (392 ug/m3), 260 ppb (600 ug/m3), and 430 ppb (992 ug/m3), respectively . The maximum level was also one thirty-eighth of the 8-hr threshold limit value (TLV) of 230 ug/m3 . ATSDR's intermediate EMEG of 0.009 ppb (0.02 ug/m3) is identical to EPA's RfC and incorporates a safety factor (or cumulative uncertainty factor) of 1,000 [17, 19]. Thus, the maximum level detected (so far) on-site would not, by itself, be expected to produce any adverse health effects in workers intermittently exposed on-site, and should present even less hazard to residents off-site. (See comment on additive effects in the summary at the end of this section).
However, it is not clear just how representative the two measurements made at the Burn Pit (Nov. 9 and 13, 1995) are of the concentration of acrolein in ambient air at other times and places at Bovoni Landfill. If concentrations were much higher on occasion (e.g., during landfill fires), then acrolein (and such other aldehydes as formaldehyde) may well have contributed to the instances of respiratory irritation reported by workers and nearby residents. The effects of chemical respiratory irritants are very similar to those of smoke inhalation, and the two may act in concert to produce a greater effect than either would by itself [20, 21].
Chronic inhalation exposure to arsenic trioxide at copper smelters and arsenate at chemical plants is associated with an increased risk of lung cancer in workers . However, when smoking is taken into account, the association between lung cancer and arsenic is seen only at relatively high, occupational exposure doses . High concentrations of arsenic dusts in air can also produce various symptoms of respiratory irritation, but the latter are unlikely to occur below concentrations of 100-1,000 ug/m3 . The maximum level of arsenic in air detected at Bovoni Landfill in 1995 was 7.3 ug/m3 at a burn pit. The latter concentration is lower than the American Council of Governmental and Industrial Hygienists' time-weighted average (TWA)-TLV of 10 ug/m3 but exceeds the EPA Region III RBCs for noncancer effects (1.1 ug/m3) [18, 24]. (ATSDR has no noncancer comparison value for arsenic in air.) However, RBCs are ultimately based on EPA's RfDs or reference doses or RfCs, which are levels considered "safe" for a lifetime of exposure, and are generally much lower than minimal risk levels (MRLs) based on acute or intermediate exposure . (The non-cancer RBC for arsenic in air was based on the oral RfD because no RfC was available.) Because neither workers nor residents will be chronically exposed to levels as high as 7.3 ug/m3, the RBC is not strictly applicable. The fact that the maximum concentration detected at the landfill (7.3 ug/m3) exceeds the RBC does not imply that adverse health effects will necessarily ensue under site-specific conditions of exposure. Since both workers and residents will presumably be exposed only intermittently to lower levels for shorter periods of time, arsenic in air at the landfill does not represent a likely public health hazard.
Although the maximum detected concentration of benzene in air in 1995, i.e., 210 ppb or 670 ug/m3 in an on-site smolder pit (Table 4, Appendix B), exceeded ATSDR's current acute EMEG/MRL of 50 ppb (160 ug/m3) in air, the acute toxicity of benzene is actually relatively low [25, 26]. ATSDR's conservative acute MRL is based on a lowest observed adverse effect level (LOAEL) of 10 parts per million (ppm) (10,000 ppb) for lymphocyte depression in mice . Humans appear to be less sensitive to immunotoxic effects than are laboratory animals, especially mice . Nevertheless, ATSDR's conservative acute EMEG includes a safety factor of 300 . ATSDR's conservative acute MRL is one two-hundredth of the TWA-TLV of 10,000 ppb set by the ACGIH . The maximum detected on-site concentration of benzene in 1995 is only 2% of the TLV and would have been much lower off-site, well below levels of health concern.
The maximum concentration of benzene in air detected in 1996 (2,500 ppb or 8,000 ug/m3) was measured at a "vent" or open fissure on-site and was not, therefore, representative of concentrations in the breathing zone. It substantially exceeded the acute EMEG of 50 ppb but was still only 25% of the 8-hr TLV of 10,000 ppb. Given the location of the measurement (i.e., at a vent), the limited duration of potential worker exposure, and the inevitable dilution of any air contaminants migrating off-site, ATSDR considers it unlikely that the measured concentration (2,500 ppb) of benzene, or rather the lower, unmeasured concentration in the breathing zone, would result in any adverse health effects in workers or nearby residents.
An isolated reading of 5 ppm benzene (5,000 ppb) was taken at a smoldering vent with a colorimetric tube (data not shown). However, because they have many interferences and do not provide accurate data, these devices are usually used only for screening. If such a reading were accurate, it would still be only half the 8-hr TWA-TLV and would not represent levels in the breathing zone. Because of open air movements at the landfill, levels in the breathing zone would be much lower and transitory, and would not be expected to produce any adverse health effects.
Carbon monoxide (CO) binds to hemoglobin about 250 times more tightly than does oxygen. CO displaces the latter, forming carboxyhemoglobin (COHb), and reduces the amount of oxygen delivered to the tissues of the body, particularly the heart and brain [28, 29]. At Bovoni Landfill, a concentration of 110 ppm (0.011%) CO was detected at a vent by use of a colorimetric tube (data not shown). The tubes are used primarily for screening purposes and do not yield precise or reliable results. However, this reading is the only data ATSDR has regarding CO levels on-site. Therefore, it will be assumed, for the purposes of this toxicological evaluation, that this single reading accurately reflects CO concentrations on-site.
ATSDR has no comparison value for CO. Practical information is available elsewhere on the effects of CO on humans. Exposure to 25 ppm CO (the current TLV) under normal working conditions (8 hr/day, 5 day/week [wk]) corresponds roughly to a carboxyhemoglobin level of 3.5% . Some studies suggest that such low levels of COHb may significantly (in the statistical sense) reduce performance on psychomotor tasks of long duration (e.g., "careful driving skills" and "video game performance"), but conflicting results have been reported . Most effects of low-level carbon monoxide exposure are entirely reversible upon removal of the victim from the source of exposure and resumption of proper ventilation. No symptoms or shortness of breath usually occur during vigorous muscular exercise at COHb levels of 0-10% . A mild headache and breathlessness on moderate exercise may occur at COHb levels of 10-20%. At equilibrium, CO levels of 100 ppm (i.e., close to those measured at a vent at Bovoni Landfill) produce an average COHb level of 16%, which may be associated with perceptible clinical effects. Throbbing headache, irritability, emotional instability, impaired judgement, defective memory, and rapid fatigue occur at 20-30% COHb. Carbon monoxide levels at fires may reach 100,000 ppm (10%), which can raise COHb levels in active firefighters without respiratory protection to 75% within 1 minute, leading to unconsciousness, convulsions, and, if exposure is prolonged, death .
In the context of the human effects levels described, intermittent exposures of acute or intermediate duration to ambient air concentrations significantly less than 110 ppm or 0.011% CO are not likely to produce any appreciable adverse health effects. The maximum reading for CO on-site was taken at a vent and was not representative of probable (i.e., lower) levels in the breathing zone.
Ethylbenzene concentrations at Bovoni Landfill ranged from 1.31-391 ppb (5.7-1,700 ug/m3). The maximum concentration, which only marginally exceeded (by 30%) ATSDR's intermediate EMEG/MRL of 300 ppb (1,300 ug/m3), was detected in 1996 at a vent on-site (Table 5, Appendix B). Elsewhere on-site, concentrations were generally less than 2.3 ppb (10 ug/m3).
The measurements taken at a vent on-site are not representative of levels in the breathing zone, which should be much lower. In addition, the intermediate EMEG/MRL is based on the lowest animal NOAEL available (100 ppm for skeletal abnormalities in newborn rats) and a safety factor of 100 . Clearly, an exposure limit based on developmental effects in baby rats is of limited relevance to the workers at Bovoni Landfill. EPA's RfC, based on lifetime exposure in adults, is 230 ppb (1000 ug/m3), and the TLV, based on repeated exposure (8-hr/day, 40 hr/wk) in adult workers, is 100,000 ppb (434,000 ug/m3) [32, 18]. Therefore, even the maximum concentration of ethyl benzene detected at Bovoni Landfill (391 ppb or 1,700 ug/m3) should not be associated with any adverse health effects in workers intermittently exposed for relatively short durations. It is reasonable to expect that levels off-site will be substantially lower and will also represent no hazard to public health.
Formaldehyde is ubiquitous in the environment, occurring as a common contaminant of smoke and as a component of photochemical smog. At room temperature, formaldehyde is a colorless gas with a pungent, irritating odor detectable at 500 ppb or about 610 ug/m3 . The odor threshold ranges from 50 to 1,000 ppb (62-1,200 ug/m3), depending on the individual and the degree of acclimatization. At exposure levels of 1-4 ppm (1.000-4,000 ppb or 1,200-4,900 ug/m3), it is a strong mucous membrane irritant, producing burning and lacrimation . However, in a controlled study, formaldehyde gas at levels of 3 ppm (3,700 ug/m3) or less did not aggravate asthma symptoms in patients with presumptive formaldehyde-induced asthma. The short-term exposure limit (TLV-STEL) is the concentration to which workers can be exposed for a short time (up to 15 minutes, 4 times a day) without suffering from irritation, chronic or irreversible tissue damage, or narcosis sufficient to increase the likelihood of injury . The formaldehyde TLV-STEL is 300 ug/m3 or about 245 ppb . ATSDR has no noncancer comparison value for formaldehyde in air.
Based on the above numbers, the maximum concentration of formaldehyde detected on-site in 1995 (70 ug/m3 or 57 ppb) in the smoke plume at the Burn Pit was below the odor threshold for most people, was only 5-6% of levels normally expected to produce significant irritation, and was less than one-fourth of the TLV-STEL. Therefore, no adverse health effects would be anticipated in workers exposed only intermittently to generally lower concentrations on-site, and any potential hazard to residents off-site should be even lower. However, there are reports in the literature of people experiencing effects below 1 ppm, including conjunctival irritation (10-50 ppb or 12-62 ug/m3) and pulmonary irritation (30-3,000 ppb or 37-3,700 ug/m3) . As a highly water-soluble gas, formaldehyde in sufficient concentrations, can cause immediate upper airway injury. (Poorly soluble gases such as phosgene produce delayed effects on the terminal bronchioles and alveoli.) Therefore, if concentrations were much higher than the recorded maximum on occasion (e.g., during landfill fires), formaldehyde (and other aldehydes such as acrolein) might have contributed to the respiratory irritation reported by nearby residents. However, smoke inhalation and respiratory irritants produce similar physiological effects, and the effects of the two may not be readily distinguishable during a fire [20, 21].
Airborne mercury (Hg) is primarily elemental. Elemental Hg is well absorbed by inhalation. High levels of mercury vapor are extremely irritating to the lungs. Acute exposure to high concentrations may lead to metal fume fever with pneumonitis, bronchitis, and bronchiolitis . The signs and symptoms of chronic intoxication by mercury vapor, which is more frequent than acute toxicity because of the cumulative nature of mercury exposure, begin with psychic and emotional disturbances, loss of concentration, depression, headaches, fatigue, weakness, loss of memory, drowsiness, and insomnia, and progress to fine muscular tremors. Kidney and brain tissue are the major depots after Hg vapor exposure .
The maximum on-site concentration of mercury vapor (121 ug/m3) was detected November 12, 1995, in the Burn Pit area (Table 4, Appendix B). This concentration was 6,000 times higher than ATSDR's acute inhalation EMEG/MRL for colloidal mercury (0.02 ug/m3). However, this conservative acute MRL, which is based on a LOAEL of 50 ug/m3 for impaired spatial learning in rats and a safety factor of 100, is about one seven-hundredth of the published chronic effects levels in humans. It is almost identical to the average concentration of mercury reported in ambient air in non-industrialized areas during the 1970s (0.01-0.02 ug/m3) . Of more relevance to human health is the fact that the maximum concentration of mercury vapor detected at Bovoni Landfill was almost 5 times the TLV of 25 ug/m3  and marginally exceeded the range of LOAELs for less serious effects (14-106 ug/m3) in chronically exposed humans .
It is not clear what produced such high readings for mercury vapor in the Burn Pit area on November 12, 1995. It is possible that the fire was fed partially by waste mercury or mercury-containing medical wastes (e.g., pharmaceuticals and dental amalgams), agricultural wastes (e.g., pesticides), and/or electrical products (mercury switches, fluorescent lamps, and dry-cell batteries). Also, the instrument used to take the reading has a number of interferences, including hydrogen sulfide, chloride compounds, and temperature changes. These interferences may have accounted for some unknown portion of the high reading. However, the concentration of hydrogen sulfide was recorded as 0.000 ppm at the same time and place that the 121 ug Hg/m3 reading was taken.
Whatever the cause of the abnormally high reading, the maximum concentrations detected were not representative of concentrations of mercury vapor in the breathing zone at other times and places on-site, and no one is likely to have been exposed to such levels for very long. Of the 15 readings of mercury vapor taken at the Burn Pit area from November 8 to November 13, 1995, (range less than 1 through 121 ug/m3, average 10.5 ug/m3), only 4 were greater than 1 ug/m3, and average concentrations of Hg vapor were lower everywhere else on-site except at the smolder pit (range 0-31 ug/m3, average 15 ug/m3 based on only two samples). Thus, the average concentration of mercury vapor at the Burn Pit was, at most, 40% of the LTV (i.e., ignoring the possibility that interferences accounted for at least some part of the high readings). It was also below the range of LOAELs for less serious effects in chronically exposed humans mentioned earlier. Therefore, assuming that on-site exposures will generally be intermittent and of acute or intermediate duration, the average level of mercury vapor detected in the Burn Pit area (and on-site in general) during the monitoring period would not be expected to produce any adverse health effects.
The maximum level of nickel measured in air at Bovoni Landfill was 14.6 ug/m3, which is 146 times higher than ATSDR's intermediate EMEG of 0.1 ug/m3 [35, 36]. However, it was only one-fifth of the EPA Region III RBC of 73 ug/m3 for "Nickel and Compounds"  and between one-seventh and one sixty-eighth of the TLVs for soluble (100 ug/m3) and insoluble (1,000 ug/m3) nickel compounds . (RBCs are based on lifetime exposure rather than on acute or intermediate exposure.) ATSDR's conservative intermediate EMEG is based on a NOAEL of 20 ug/m3 in rats, and a safety factor of 100 [35, 36]. Therefore, the maximum detected level of nickel on-site is not expected to pose any health hazard to workers intermittently exposed at the landfill nor, considering the effect of dilution during off-site migration of on-site contaminants, to nearby residents.
Phosgene is a heavier-than-air, colorless gas (COCl2) that was used as a chemical warfare agent in World War I . Toxic levels are reached before its suffocating odor (resembling musty hay or decaying fruit) becomes detectable. It is used in the manufacture of isocyanates, dyestuffs, insecticides, and pharmaceuticals. Phosgene is also produced by the decomposition of heated organochlorine compounds. Thus, phosgene concentrations could dramatically increase during a landfill fire if that fire were partially fueled by synthetic materials such as polyurethane and/or polyvinyl chloride.
The maximum concentration of phosgene (750 ppb or 3,000 ug/m3) was detected on November 9, 1995, approximately 6-12 inches away from one of the smoking fissures in the smolder pit (Table 4, Appendix B). The maximum reading in ambient air (190 ppb or 760 ug/m3) was taken the previous day at the smolder pit ambient air station. Both of these readings exceeded the 8-hr TLV of 100 ppb or 400 ug/m3 . (ATSDR has no comparison value for phosgene.) At a concentration of 1 ppm in air (1,000 ppb), phosgene causes little or no immediate irritation; however, it may, after a latent period of some hours, cause severe pulmonary edema . Thus, it is entirely possible that phosgene (in combination with various aldehydes and smoke) contributed to the respiratory complaints of on-site workers in 1996 and late 1995. Because of dilution in air as it migrates off-site, phosgene gas emanating from a smolder pit on-site would not necessarily have resulted in any adverse health effects in nearby residents. However, the evolution of even higher concentrations of phosgene and other respiratory irritants during previous, major landfill fires, especially in combination with smoke inhalation, would be consistent with the health effects data (i.e., respiratory complaints) obtained from local residents at those times. Between the times of surface fires, the apparent concentration of phosgene may be expected to vary widely, depending on the level of interferences and the nature of the materials fueling the underground fire. For example, during the 1996 air monitoring event, no phosgene was detected, even near vents.
ATSDR and EPA do not have noncancer comparison values (or RfCs) for 1,1,2-trichloroethane. However, the maximum concentration detected at Bovoni Landfill (260 ug/m3) was less then one two-hundredth of the current TLV of 55 mg/m3 (55,000 ug/m3), a concentration to which nearly all workers may be repeatedly exposed for a normal 8-hr workday and a 40-hr workweek without adverse effects  and one seventeenth of Kentucky's acceptable ambient air concentration of 4.5 mg/m3 (8-hr average) . (Note that state and federal standards designed to protect sensitive members of the general population are typically much more conservative than are occupational exposure limits designed to protect healthy workers.) Therefore, the maximum concentration detected at Bovoni Landfill (260 ug/m3) would not be expected to be of health concern to workers on-site. The concentrations to which residents may be exposed will have been further diluted as air migrates off-site and, based on the limited data available, are not expected to pose any health hazard. Upwind levels of 1,1,2-trichloroethane (260 ug/m3) were higher than downwind levels (180 ug/m3) at Bovoni Landfill in 1995, the opposite of what would usually be expected if the landfill were the source. Concentrations were below detection limits in 1996.
As Table 4, Appendix B, indicates, the maximum concentration of 1,2,3-trichloropropane (110 ug/m3 or about 18 ppb) exceeded ATSDR's acute EMEG/MRL of 0.3 ppb (1.8 ug/m3) in air . However, this particular comparison value is unusually conservative, especially for an acute MRL. The 8-hr TWA-TLV, a concentration to which nearly all workers can be repeatedly exposed 8 hours a day, 40 hours a week without adverse effect, is more than 30,000 times higher (i.e., 10,000 ppb) . ATSDR's 0.3 ppb acute MRL was based on a NOAEL of 1,000 ppb for nasal irritation (i.e., "increased thickness of the olfactory epithelium") in rats and a cumulative uncertainty factor (i.e., "safety factor") of 100 . However, rats are actually more sensitive to nasal irritation than are people because the former are obligate nose breathers, i.e., rats cannot breath through their mouths as humans can to relieve irritation of sensitive nasal membranes. Based on human experience with 1,2,3-trichloropropane, it is highly unlikely that the 18 ppb (110 ug/m3) measured at an on-site burn pit would have any adverse effects on workers intermittently exposed at the landfill. Concentrations off-site (i.e., 500 feet to a mile from the source) would routinely be considerably lower than those on-site. Any risks that might be associated with exposure to 1,2,3-trichloropropane in air would be correspondingly reduced as well.
In the 1996 sampling, the maximum detected level of xylene in on-site air was 102 ppb (Table 5, Appendix B). This value does not exceed ATSDR's intermediate EMEG/MRL of 700 ppb for total xylenes and only marginally exceeds ATSDR's chronic EMEG/MRL (100 ppb), which is one one-thousandth of OSHA's TWA-TLV [41, 18]. Thus, exposure to xylene at the levels detected on-site are not expected to be associated with any adverse health effects in on-site workers or nearby residents.
Information on general health-related trends in the Virgin Islands is available through the Department of Health, Bureau of Vital Statistics, Cancer and Birth Registries. No relevant health outcome data are available for the residential population living near the site or for the worker population; therefore, ATSDR staff have not evaluated health outcome data specifically for the Bovoni Landfill.
To address health concerns related to the Bovoni Landfill, a health clinic was opened for the residents. The clinic was originally operated on a weekly basis by the Virgin Islands Department of Health in March of 1996. Currently, the health clinic receives patients at scheduled appointments. Dr. Audria Thomas has been directing this clinic. She reported that there was an initial reluctance to come to the clinic, despite the fact that it provided free physical examinations, complete blood counts, serum chemistries, urinalyses, chest X rays, and arterial blood gas reports. Dr. Thomas indicated that approximately 85 individuals have been examined at the clinic. Of these individuals, 55 have been seen only once, and the remaining 30 have had 2 or more visits. Most of these individuals have been from the residential areas west and north of the landfill .
As part of the petition process, Agency for Toxic Substances and Disease Registry (ATSDR) staff have gathered health concerns from several sources. These sources include: the health concerns noted by the petitioning environmental group, the health complaints expressed by residents attending the public meeting in August 1996, the health concerns noted by the Army Corps of Engineers personnel assisting in the debris reduction mission at the landfill in late 1995 and early 1996, and health concerns described in letters ATSDR received from concerned local residents. All of these health concerns are described below:
- concerns that health effects such as cough, congestion, irritated throats, irritated lungs, light-headedness, headaches, nausea, chest tightness, and shortness of breath are attributable to landfill smoke
The primary symptoms reported to the health clinic were upper respiratory and nasal irritation. Respiratory health concerns were also expressed during the public meeting and in the letters ATSDR received. Local residents complained of a symptom complex of upper respiratory irritation, congestion, cough, headache, conjunctival irritation, chest tightness, and shortness of breath. These symptoms were common in residents to the north and west of the landfill, particularly when a fire was actively burning. These symptoms were also noted some nights when no fire was visible but small traces of smoke were reported by the residents. Although no imminent, life-threatening health effects were noted, there is a good deal of respiratory symptomatology, primarily during times of active smoke from the landfill. There is a plausible relationship between the respiratory symptoms that have been described and smoke from the landfill fire with its accompanying contaminants or particulates. Please refer to the Toxicological Evaluation Section for information on specific contaminants.
- concerns that musculatory ailments, degenerative arthritis, knee problems, joint pains, stiffness, and skin rashes are related to landfill contaminants
Dr. Thomas reported for the health clinic that a few complaints relating to joint and musculoskeletal concerns have been voiced. Most of the symptoms have involved the aggravation of pre-existing chronic medical problems. Gastrointestinal symptoms and complaints have been fairly rare .
During the public meeting in August 1996, concerns were expressed regarding the firefighters who have dealt with the landfill. Reports to ATSDR indicated that there had been a 28% increase in sick leave from March to August 1996. Much of this sick leave was related to arthritis symptoms.
The sporadic and intermittent nature of this particular landfill fire makes it difficult to determine possible exposures, doses, and health effects. As previously noted, there is a plausible relationship between the respiratory symptoms reported and the smoke, particulates, and chemical contaminants generated by the landfill fire. However, the other health complaints that residents relate to the landfill and the fire, including musculoskeletal, dermatological, and gastrointestinal symptoms, seem much less plausible based on the chemicals involved and the probable magnitude and duration of exposure.
- concerns over drinking and bathing with cistern water
The underground fire at the landfill emits combustion products to the ambient air through surface fissures and through above-ground fires that occasionally break through the surface. The primary contaminants detected during 1995 and 1996 air sampling events included several volatile organic compounds (VOCs) and polycyclic aromatic hydrocarbons (PAHs).
During rainstorms, contaminants in the air could be picked up by falling raindrops. These raindrops would then be trapped by the rooftop rain catchments and collect in household cisterns. No air monitoring was conducted in residential areas or during above-ground fire events; however, air contaminant levels would be significantly reduced by dilution and diffusion as contamination migrated from the landfill. In addition, the amount of VOCs that would be absorbed by falling rain is expected to be relatively small, as most VOCs have high vapor pressures and relatively low water solubilities . Because VOC concentrations in ambient air are reduced during migration from the landfill and the amount of VOCs absorbed by falling rain is small, no significant exposure to VOCs would be expected from drinking cistern water .
There was a tire fire at the landfill in March 1996. The dense, black smoke emitted during a tire fire is composed largely of carbon black particulates that may absorb other combustion products, such as PAHs . Smoke particulates from the burning tire debris could have been deposited on the roofs of area homes. This particulate matter could then have been washed into the cisterns during subsequent rainstorms. However, carbon black is insoluble and would not dissolve in the water column, and PAHs are relatively insoluble in water and would tend to concentrate in sludge on the bottom of the cistern . Therefore, the concentration of PAHs in the water column would also be expected to be low .
Although ATSDR does not consider it likely that cistern water in the Bovoni Landfill area has been significantly contaminated as a result of the landfill fires, this water has not actually been sampled or analyzed. In the absence of such cistern water sampling data, ATSDR suggests that concerned residents might use diversion valves or roof washers in conjunction with their cisterns. These kinds of devices can prevent foreign matter on the roof, such as bird droppings, insects, stray dirt, and particulates, from entering household cisterns . Residents with cisterns should also periodically flush out sludge that collects in the bottom of the cisterns. Residents following these good hygienic practice recommendations will further reduce their exposure to potential contamination from the landfill as well as that from the more significant general sources mentioned previously (bird droppings, insects, etc.).
- concerns related to community education
Residents feel strongly that issues concerning Bovoni should be made more public and information should be disseminated frequently. Community education was a key issue, especially education related to learning safeguards that could immediately prevent potential exposures to contaminants and education in preparation for the next occurrence of an above-ground fire. ATSDR's Division of Health Education and Promotion is available to provide health education to community members to help them understand the risk in living and working near this site.
- concerns over improper disposition of medical waste
Many reports of improper disposition of medical waste have surfaced over the years. Potential risks may have been present for landfill workers in the past because of the improper disposal of medical waste in the landfill. Landfill workers may have come in contact with items of concern, including bagged medical waste and medical "sharps" (syringes, needles, scalpel blades, etc.), that were dumped indiscriminately along with other wastes. Since nearby residents did not come in contact with the medical waste, no public health concerns are evident.
ATSDR staff toured the landfill site and did not notice any visible medical waste on the surface of the landfill. The concern over improper disposition of medical waste is no longer an issue, because current practice involves the incineration of medical waste at the Roy Schneider Hospital and not burial at the landfill.
- concerns over the landfill damaging the Mangrove Lagoon
The petitioning community group expressed concern about the overall effect of the landfill on the environment, particularly on the Mangrove Lagoon. The Mangrove Lagoon, one of the largest remaining mangrove areas in the U.S. Virgin Islands, provides a refuge for a diverse wildlife population . The tides bring ocean water past barrier islands through channels in the lagoon. During a visit to the landfill in August of 1996, ATSDR staff noticed that the lagoon area directly adjacent to the landfill did not have much vegetation with leaves while most other vegetation on the island was green and leafy.
ATSDR received reports commenting on conditions at the lagoon in the 1970s and 1980s [10, 11]. While there were several reports on the lagoon, none contained environmental data, such as surface water sampling results, for ATSDR to review and evaluate. The reports described a variety of potential sources of contamination, including discharge from Turpentine Run, effluent from a sewage treatment plant, and the landfill. One report stated that effluent from sewage treatment plants is probably the main factor leading to habitat degradation, decrease in dissolved oxygen, and increase in suspended solids . The report also stated that the middle and outer Mangrove Lagoon areas demonstrated a trend toward cleaner water . A different report indicated that the lagoon water contained alarming concentrations of heavy metals, VOCs and total petroleum hydrocarbons . Actual contaminants and concentration levels were not reported.
Trapping results from the inner Mangrove Lagoon area in the 1980s indicated that while most fish caught appeared healthy, occasional individual fish had surface growths or deformities . No fish sampling data were available for ATSDR to review and evaluate; therefore, it is impossible to make a public health statement concerning the ingestion of fish from the inner Mangrove Lagoon area.
Because of a lack of environmental data, conflicting reports, and the influence of other sources such as the discharge from Turpentine Run and the sewage treatment plant effluent, it was impossible for ATSDR to discern the overall effect of the landfill on the lagoon. ATSDR has recommended sediment sampling in an area of the Mangrove Lagoon adjacent to the landfill in this public health assessment. If environmental data concerning the Mangrove Lagoon are provided to ATSDR in the future, ATSDR will evaluate the data for public health significance.