Environ. Sci. Technol. 2007, 41, 5902-5907
Lead Bioaccessibility in Food Web Intermediates and the Influence on Ecological Risk Characterization CLAIRE A. KAUFMAN, JOSEPH R. BENNETT, IRIS KOCH, AND KENNETH J. REIMER* Environmental Sciences Group, Department of Chemistry & Chemical Engineering, PO Box 17000 Stn Forces, Royal Military College of Canada, Kingston, Ontario K7K 7B4, Canada
Models simulating gastric conditions of mammalian (eastern cottontail, Sylvilagus floridanus; short-tailed shrew, Blarina brevicauda) and avian (American robin, Turdus migratorius) receptors were used to investigate the proportion of lead (Pb) mobilized into the digestive juices (the bioaccessible fraction) from soil, earthworms, and vegetation collected at a rifle and pistol (RP) range in eastern Ontario, Canada. Pb concentrations averaged 5044 mg kg-1 in RP range surface soils, 727 mg kg-1 in earthworm tissue, and 2945 mg kg-1 in unwashed vegetation. For mammalian gastric models, the bioaccessible fraction of Pb in soils was 66 ( 22%, in earthworm tissue was 77 ( 14%, and in unwashed vegetation was 50 ( 37%. For the avian gastric model, the bioaccessible fraction of Pb in soil was 53 ( 43% and in earthworm tissue was 73 ( 13%. The incorporation of soil and food web intermediate bioaccessibility data into standard risk calculations resulted in predicted risk being reduced for all receptors. The inclusion of bioaccessibility during ecological risk assessment affords a more realistic estimate of contaminant exposure, and is a valuable tool for use in contaminated sites management.
Introduction The threat posed to ecological receptors by soil contaminants is frequently assessed using total concentrations of these contaminants. Risk is frequently evaluated by comparison of these values to soil quality guidelines (1), or through the calculation of hazard quotients for ecological risk assessment (2). This practice is inherently conservative because it assumes that during exposure, receptors will absorb 100% of the soil contaminant. However, the use of total metal concentrations may result in overestimates of exposure to soil contaminants, potentially leading to site remediation that is costly, time-consuming, and largely unnecessary (3, 4). Organisms are generally exposed to inorganic soil contaminants via three pathways: dermal, oral, and inhalation, although oral ingestion of contaminated soil, food, and water is recognized as the dominant pathway of contaminant uptake (5). To pose a risk, ingested contaminants must become “bioavailable,” meaning they must dissolve in the gastrointestinal tract of the organism and enter the blood * Corresponding author phone: (613) 541-6000 ext. 6161; fax: (613) 541-6596; e-mail:
[email protected]. 5902
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 16, 2007
stream (6). Several soil characteristics such as pH, presence of crystalline or amorphous iron oxides, organic carbon content, and cation exchange capacity can influence metal bioavailability, often reducing it to less than 100% (4). Matters are further complicated when trying to estimate the bioaccessibility of metals in food products (e.g., plants, invertebrates, etc.) ingested by organisms occupying higher trophic levels. Therefore, understanding both contaminant movement through food webs and contaminant bioavailability is crucial to estimate exposure accurately. Tests to determine bioavailability, whereby contaminant concentrations are measured in the blood and body tissue of live animals dosed with contaminated soil, have been developed (6). However, these in vivo studies are expensive, time-consuming, and pose an ethical dilemma. For those reasons, tests measuring bioaccessibility (the fraction of a contaminant mobilized into the digestive juices of a receptor from ingested soil and food) have been developed as an alternative. Bioaccessibility can be estimated using these simple and inexpensive in vitro models simulating gastrointestinal conditions (7). Models estimating the bioaccessibility of lead (Pb) in soil have received considerable attention and recent efforts have focused on validation of these models through comparison with in vivo studies (8). However, the bioaccessibility of contaminants present in food web intermediates has remained largely unexplored. This study focused on the investigation of Pb bioaccessibility in soil, earthworms, and vegetation collected at a rifle and pistol (RP) range in Eastern Ontario, Canada. It is estimated that there are more than 12 000 active outdoor small arms firing ranges in the United States alone, and over a million kilograms of Pb are discharged at these ranges annually (9, 10). Previous studies have focused on contaminant characterization at small arms firing ranges (e.g., 11, 12), risk posed to receptors (13), and the movement of Pb from firing ranges soils into the tissues of mammalian and amphibian receptors (e.g., 14, 15). However, the proportion of the Pb body burden due to the ingestion of either soil or contaminated food products remains unclear. For this study, existing bioaccessibility methods were modified to simulate the gastric conditions of three receptors found throughout North America and commonly explored during ecological risk assessment: the American robin (Turdus migratorius), an avian omnivore, the eastern cottontail (Sylvilagus floridanus), a mammalian herbivore, and the short-tailed shrew (Blarina brevicauda), a burrowing mammalian insectivore. The diet of robins and shrews is comprised of a significant proportion of earthworms (16, 17) and both have been shown to bioaccumulate Pb due to dietary preference (18, 19). Herbaceous and woody plants dominate the diet of the eastern cottontail, and Pb poisoning via food chain transfer of Pb has previously been observed (20). The objective of the study was to determine how Pb bioaccessibility might influence contaminant exposure to the three receptors via direct ingestion of soil and food (earthworms and vegetation). Exposures were then evaluated by calculating hazard quotients for the receptors using soil Pb values typically found at small-arms firing ranges.
Materials and Methods Study Site. The rifle and pistol (RP) range selected for this study is located near Kingston, Ontario. The range is approximately 100 m long by 20 m wide, is built to standards outlined by the Canadian Firearms Centre (21; Figure 1), and has been in use for 5 years by approximately 325 individuals annually. Ammunition is discharged from a series 10.1021/es062443u CCC: $37.00
2007 American Chemical Society Published on Web 07/21/2007
FIGURE 1. Layout of study site, including soil, earthworm, and vegetation sampling locations. Two samples representing background conditions were collected outside the range boundaries and are not shown on this map. of firing lines toward stationary targets. Weapons discharged at the site are chiefly 5.56 mm rifles, 9 mm pistols, 0.38 caliber pistols, and shotguns. The RP range is located in the mixed wood plains ecozone of the St. Lawrence Basin (22). The range floor consists of manicured grass; the earthen berms that surround the range are vegetated with various grasses, forbs, and shrubs; the bullet pockets are devoid of vegetation. The RP range is surrounded by forest to the north and an abandoned field to the east, which is being colonized by trees. Bedrock in the area consists of paragneiss overlain with fine-grained silt and clay (23, 24). Native soils at the site are moderately to poorly drained albic luvisols (25). The face of the impact berm is lined with approximately 30 cm of imported wellsorted, medium-grained sand. Average pH for soils at the site is 8.06 ( 0.46 (n ) 14). Bullets (or buckshot in the case of shotgun cartridges) used contain approximately 97% Pb (26). The majority of the Pb mass accumulates behind targets in bullet pockets within a 6-m high impact berm. High impact speeds cause a certain amount of bullet fragmentation, and Pb in microscopic bullet fragments may agglomerate or become smeared onto soil particles. Although a small quantity of Pb is emitted from the muzzle of the gun as airborne particulate matter, a larger amount becomes airborne when bullets strike the impact berm (27, 28). This airborne dust settles over time, becoming adhered to vegetation and incorporated into the surface media, as was apparent at the study site. Soil, Vegetation, and Earthworm Sampling. Site assessments conducted at RP ranges located near the study site have shown that Pb concentrations in surface soils (0-10
cm depth) exceed those at depths greater than 10 cm (see Table 1 in Supporting Information). For that reason, our sampling program was limited to the collection of surface soil samples from 20 locations on the face, top, and backside of the impact berm (Figure 1). Reference samples were collected from two areas outside of the firing range boundaries in order to obtain representative background concentrations of Pb. Vegetation was collected within a 50-cm radius of all soil sampling locations. Grazing patterns of the eastern cottontail were simulated by collecting only the above ground portion of plant species. An effort was made to collect earthworms from all 20 established soil sampling locations; however, specimens were recovered from only 16 of the locations. Earthworms were collected by hand-sorting soil from depths between 0 and 30 cm in order to obtain adequate sample sizes, although the majority were found concentrated in the top 10 cm of the soil profile near the litter layer. Soil, Vegetation, and Earthworm Sample Preparation. Soil sample heterogeneity is an issue at firing ranges (29), therefore the following method was designed to reduce analytical uncertainty. Each soil sample was sieved and homogenized, then a 30-g subsample of the
