Distribution and Variation of Arsenic in Wisconsin Surface Soils, With Data on Other Trace Elements

A total of 664 soil samples distributed among different geographic regions and soil types were collected across Wisconsin to describe the distribution of arsenic relative to parent material, soil texture, and drainage class. Soils from 6 inches in depth were composited, digested in aqua regia, and analyzed for 17 trace elements. Observed soil arsenic concentrations range from a high of 39 milligrams per kilogram (mg/kg) to less than the laboratory detection limit of 1 mg/kg. Ten samples with soil arsenic concentrations greater than 8.5 mg/kg were determined to be significantly separate from the main cluster of the dataset. With these outliers removed, overall soil arsenic concentrations in Wisconsin have a median value of 1.8 mg/kg, and the 95-percent upper confidence limit of the mean is 2.4 mg/kg. Soils with sandy glacial outwash as a parent material have a lower median arsenic concentration (1.0 mg/kg) than soils forming in other parent materials (1.5 to 3.0 mg/kg). Soil texture and drainage category also influence median arsenic concentration. Finer grained soils have a higher observed range of concentrations. For loamy and loess-dominated soil groups, drainage category influences the median arsenic concentration and observed range of values, but a consistent relationship within the data is not apparent. Statistical analysis of the 16 other elements are presented in this report, but the relationships of concentrations to soil properties or geographic areas were not examined. Introduction Arsenic, a trace element commonly found in surface soils, has both natural and anthropogenic sources. In uncontaminated soils, the most common natural source of arsenic is arsenic-bearing minerals in soil parent materials. Arsenicbearing minerals, such as glauconite, arsenopyrite, pyrite, and other sulfides, can weather in soils and release arsenic into the soil system. Arsenic-bearing minerals are commonly found in sedimentary rocks (Smith and others, 1998). Deposition from atmospheric sources of arsenic can contribute significantly to arsenic concentrations in soils. The ratio of atmospheric natural sources, such as volcanic activity, to atmospheric anthropogenic sources, such as fossil fuel combustion, varies regionally but on average their mutual contributions are about equal (O’Neill, 1995). Other non-atmospheric anthropogenic sources include industrial and mining waste products and agricultural applications (Smith and others, 1998). Elevated levels of arsenic in soils can compromise the health of the soil system as well as human health. The greatest risk to human health is through exposure that comes from arsenic that has leached from soils or bedrock into drinking-water sources (Das and others, 1996; Smith and others, 2000). Repeated exposure to arsenic at elevated concentrations can be toxic to the human system (Southworth, 1995; Das and others, 1996). If known, baseline arsenic concentrations can be used as a guideline for assessing contamination levels of soils and may be part of the process of regulating the cleanup of contaminated soils. An element’s baseline is defined as a range of concentrations around a mean, but the baseline does not strictly represent a natural background concentration because the range of values will also include contributions from nonpoint sources of anthropogenic arsenic (Gough, 1993; KabataPendis, 2001). States have used a variety of common statistics to characterize the baseline concentration in soils for cleanup purposes, including (1) the mean concentration (Illinois Environmental Protection Agency, 2007); (2) the 95th-percentile concentration (Vosnakis and others, 2009); and (3) the 95-percent upper confidence limit of the mean (Kentucky Energy and Environment Cabinet, 2004; Montana Department of Environmental Quality, 2005). However, existing data in most states are not abundant enough to properly estimate baseline arsenic concentrations. Baseline concentrations can have natural variations that are influenced by a soil’s physical and chemical properties. Arsenic availability in a soil system is controlled by chemical processes, mainly sorption, which is affected by specific physical and chemical soil properties. The presence of iron, aluminum, and manganese oxides increases the sorption of arsenic to soil surfaces (Elkhatib and others, 1984). Soil with fine-grained textures, such as clays, and soils with high organic matter content, such as wetland soils, have more exchange sites and thus have higher sorption capacity (Chen and others, 1999; Girouard and Zagury, 2009). Arsenic sorption on oxides, clays, and organic material can be influenced 2 Distribution and Variation of Arsenic in Wisconsin Surface Soils, With Data on Other Trace Elements by pH (Frost and Griffen, 1977; Thanabalasingam and Pickering, 1986; Goldberg and Glaubig, 1988) and redox conditions (Kabata-Pendias, 2001). Regional studies have quantified arsenic concentrations in soils with respect to parent material and soil physical and chemical properties. In New Jersey, soil samples were analyzed from three physiographic provinces: Valley and Ridge, Highlands, and Coastal Plain (Sanders, 2003). Mean arsenic concentrations of the Valley and Ridge and the Highlands provinces (4.98 and 6.04 mg/kg, respectively) were found to be significantly different from that of the Coastal Plain province (2.33 mg/kg). It is assumed, given the study design, that these differences were due to differences in soil parent materials among the three provinces. A California study developed a comprehensive database of trace elements in soils to determine baseline concentrations in benchmark soils (Bradford and others, 1996). In that study, positive correlation coefficients for some elements (for example, iron-vanadium, nickel-chromium, and copper-cobalt) indicated that chemical and physical properties of soils and parent materials influenced trace element distribution within a soil. Other states have also attempted to identify variations in baseline concentrations on the basis of changes in soil texture and parent material (Michigan Department of Environmental Quality, 2005) and soil taxonomic orders and chemical characteristics (Chen and others, 2002). For example, Chen and others (2002) found that wetland soils in Florida had high arsenic concentrations, which were attributed to immobilization of arsenic by high concentrations of organic matter, iron and aluminum oxides, and clays that accumulated in wetlands. A low-density geochemical soil survey of the United States (approximate spacing of one site per 6,000 km2 (2,300 mi2)) was completed by scientists from the U.S. Geological Survey (USGS). During this survey, soil samples were collected from a depth of about 20 cm (8 in.) across the conterminous United States from 1958 to 1976 (Boerngen and Shacklette, 1981; Shacklette and Boerngen, 1984). As part of the National Uranium Resource Evaluation Program (NURE), the Hydrogeochemical and Stream Sediment Reconnaissance (HSSR) component collected stream sediments across the conterminous United States between 1976 and 1984. The initial purpose of NURE was to identify potential uranium resources, but the scope was broadened to include analyses of additional elements, creating a second national-scale geochemical survey for the United States (Smith, 1997). Sampling density for the NURE program ranged from one sample per 10 km2 (3.9 mi2) to one sample per 23 km2 (8.9 mi2) (Information Systems Programs-Energy Resources Institute, 1985). Samples selected for a systematic reanalysis of NURE soil and stream sediment samples by the USGS to improve data quality have a sampling density of about one per 289 km2 (112 mi2). Data from these studies and more recent regional USGS geochemical baseline studies are available in the National Geochemical Survey database (U.S. Geological Survey, 2004). The National Geochemical Survey contains total elemental data from soil, stream sediment, and rock samples collected across the United States, though stream sediment samples make up the majority of the data. Soil analyses in the National Geochemical Survey database have been used to characterize regional soil composition. Grosz and others (2004) combined the U.S. data with an extensive Canadian geochemical database to create a map of arsenic concentrations. Their results show that typical arsenic concentrations in northern Wisconsin are less than 0.5 mg/kg and that the observed variability in surficial arsenic concentrations is controlled by bedrock characteristics. Cannon and others (2004) reanalyzed archived NURE stream sediment samples collected across northern Wisconsin. These data were compared to near-surface and subsurface soil samples collected from the same area to determine whether stream sediment samples could predict regional geochemical patterns in soils. Comparison of mean element concentrations of the three sample groups (stream sediments, surface soils, subsurface soils) showed that element concentrations among the groups were significantly different from one another. However, with some statistical manipulation that included multiple-sample averaging and smoothing, stream sediment data could be extrapolated to soil compositions for some elements (Cannon and others, 2004). Thus, in the absence of a high-density soil geochemical dataset, element concentration predictions based on stream sediment geochemistry could be used for regional background concentrations for some elements (Cannon and others, 2004). Despite these efforts to collect and compile a comprehensive geochemical database, sufficient data do not exist to determine baseline reference values for all of Wisconsin’s soils. The USGS, in cooperation with the U.S. Department of Agriculture, Natural Resources Conservation Service (NRCS), Wisconsin Department of Natural Resources (WDNR), and Wisconsin Department of Health Services, designed this study to collect trace element data from samples that are representative of all soils types and parent material sources in the State of Wisconsin. Data from this study will be used to characterize the baseline concentration of arsenic and any variation in this baseline value that may be a result of changing soil physical and chemical properties or parent material. Purpose and Scope The objective of this report is to describe the distribution of arsenic in Wisconsin surface soils and identify potential systematic variations of arsenic concentrations with respect to soil properties or parent materials. This information about the distribution and controls on arsenic concentrations in soils across Wisconsin could be helpful for the determination of regulatory levels of arsenic in soils. Sampling and analytic methods were designed to exclude obvious anthropogenic influences to achieve values that are as close as possible to a natural baseline while recognizing that few areas are pristine. Data from this study are not directly comparable to the geochemical data in the USGS National Geochemical