Anthropogenic organic compounds in source water of select community water systems in the United States, 2002-10

Abstract

Drinking water delivered by community water systems (CWSs) comes from one or both of two sources: surface water and groundwater. Source water is raw, untreated water used by CWSs and is usually treated before distribution to consumers. Beginning in 2002, the U.S. Geological Survey’s (USGS) National Water-Quality Assessment Program initiated Source Water-Quality Assessments (SWQAs) at select CWSs across the United States, primarily to characterize the occurrence of a large number of anthropogenic organic compounds that are predominantly unregulated by the U.S. Environmental Protection Agency. Source-water samples from CWSs were collected during 2002–10 from 20 surface-water sites (river intakes) and during 2002–09 from 448 groundwater sites (supply wells). River intakes were sampled approximately 16 times during a 1-year sampling period, and supply wells were sampled once. Samples were monitored for 265 anthropogenic organic compounds. An additional 3 herbicides and 16 herbicide degradates were monitored in samples collected from 8 river intakes and 118 supply wells in areas where these compounds likely have been used. Thirty-seven compounds have an established U.S. Environmental Protection Agency (EPA) Maximum Contaminant Level (MCL) for drinking water, 123 have USGS Health-Based Screening Levels (HBSLs), and 29 are included on the EPA Contaminant Candidate List 3. All compounds detected in source water were evaluated both with and without an assessment level and were grouped into 13 categories (hereafter termed as “use groups”) based on their primary use or source. The CWS sites were characterized in a national context using an extract of the EPA Safe Drinking Water Information System to develop spatially derived and system-specific ancillary data. Community water system information is contained in the EPA Public Supply Database, which includes 2,016 active river intakes and 112,099 active supply wells. Ancillary variables including population served, watershed size, land use, population density, and recharge were characterized for each of the watersheds for river intakes and contributing areas for supply wells. A total of 313 samples were collected from 20 river intakes. Between the years of 2002 through 2010, samples were collected approximately 16 times over the course of a year. Seventy-one compounds from 12 of the 13 use groups commonly occurred (detected in greater than or equal to 1 percent of samples using an assessment level of 0.05 microgram per liter or when a compound was detected in greater than or equal to 10 percent of samples without an assessment level) indicating a wide variety of sources and pathways to these rivers and highlighting the importance of source-water protection strategies. A total of 448 supply wells were sampled once during 2002–10 as part of 30 independent groundwater studies. About 15 CWS supply wells were sampled for each independent groundwater study. Twenty-eight compounds from 7 of the 13 use groups commonly occurred indicating a wide variety of sources and pathways exist for these compounds to reach these wells and highlighting the importance of wellhead protection strategies. About one-half the 265 compounds monitored (122) were detected in both surface water and groundwater samples. A more diverse suite of compounds were detected in surface water in comparison to groundwater. However, herbicides and herbicide degradates were the most frequent group of compounds detected in both surface water and groundwater. Sixty-five of the most commonly occurring compounds were detected in one or more samples from both surface water and groundwater. Human-health benchmarks (MCLs for regulated compounds and HBSLs for unregulated compounds) were available for more than one-half the compounds (160 of the 265) monitored in this study. Fifty-eight percent (41 of 71) of the commonly occurring compounds in surface water have a human-health benchmark to which concentrations can be compared; 19 have MCLs and 22 have HBSLs. Eighty-three percent (24 of 28) of the most commonly occurring compounds in groundwater have a human-health benchmark for which concentrations can be compared; 14 have MCLs and 10 have HBSLs. To put results from this study into context with the national distribution of river intakes and supply wells used by CWSs, sites were grouped into the respective national population of land-use quartiles. The increase in compound occurrence with increasing urban and agricultural land use in the watershed or contributing area was more evident for rivers than for supply wells. The increase in detection frequency of herbicides and herbicide degradates with increasing agricultural land use was more evident for rivers than for supply wells. The occurrence of solvents did not change substantially with increasing urban land use for rivers or supply wells. Basic co-occurrence analyses were completed with and without an assessment level. Considering all detections in surface water without an assessment level, approximately 86 percent of source-water samples contained 2 or more compounds, and 50 percent of samples contained at least 14 compounds. Considering all detections in groundwater without an assessment level, 50 percent of samples contained at least three compounds. For the most part, the compounds detected most frequently as individual compounds in the environment often composed the most frequent unique mixtures. Five of the 10 most frequently co-occurring unique mixtures in both surface water and groundwater were the same: atrazine and deethylatrazine; atrazine and chloroform; deethylatrazine and simazine; atrazine and simazine; and deethylatrazine, atrazine, and simazine. Because similar mixtures were identified in both surface water and groundwater without an assessment level, future studies could be directed toward better understanding the toxicological importance of these unique mixtures. Summed concentrations of herbicide degradates were compared to concentrations of the parent herbicides in surface-water and groundwater samples collected from 8 river intakes and 118 CWS wells, from which samples were analyzed for an additional 3 herbicides and 16 degradates. The toxicity to humans for many of these degradate products is largely unknown and thus points to the importance of monitoring these compounds (both the parent and degradate) in the environment. This study highlights the importance of anthropogenic organic compounds in source water of select CWSs in the United States by characterizing their occurrence in surface-water and groundwater samples. Compound concentrations and occurrence are summarized and evaluated in a human-health context, when possible. Additionally, compounds found to co-occur as mixtures for both surface water and groundwater highlight the significance of low-level compound co-occurrence.