Identifying Future Drinking Water Contaminants (1999)

Timothy L. Miller and William G. Wilber

Several programs and initiatives of the U.S. Geological Survey (USGS) contribute information useful in identifying emerging drinking water contaminants. These programs include the National Water-Quality Assessment (NAWQA) program, the Toxic Substances Hydrology Program, and the Drinking Water Initiative. The goals of the NAWQA Program, the largest of these efforts, are to assess the status of, and trends in, the quality of the nation's surface- and groundwater resources and to link the status and trends with an understanding of the natural and human factors that affect water quality (Hirsch et al., 1988; Leahy et al., 1990; Gilliom et al., 1995). The purpose of this paper is to provide an overview of the design of the NAWQA Pro ram, especially those aspects of the program that contribute information relevant to identifying emerging contaminants of concern to drinking water. Selected findings from the program are also highlighted, along with suggested enhancements to existing monitoring programs that would greatly increase the utility of these data for identifying emerging drinking water contaminants.

The NAWQA program attempts to balance the unique assessment requirements of individual hydrological systems with a nationally consistent design structure that incorporates a multiscale interdisciplinary approach. The building blocks of the national assessment are investigations in major hydrological basins of the nation, referred to as "study units." The goal for the first phase of investigation in each study unit is to characterize, in a nationally consistent manner, the broad-scale geographic and seasonal distributions of water quality conditions in relation to the most significant contaminant sources and background conditions.

The NAWQA study units cover about 40 percent of the conterminous United States, encompass 60 to 70 percent of both national water use and the population served by public water supplies, and include diverse hydrological systems that differ widely in the natural and human factors that affect water quality. The study units are divided into three groups, which are studied on a

rotational schedule of three-year periods of intensive data collection. About one-third of the study units are in the intensive data collection phase at any given time, and the nine-year cycle is designed to be repeated perennially. The first complete cycle of intensive data collection in the study units began during 1992 and is scheduled to be completed in 2002.

The national assessment goals of NAWQA are being accomplished in two main ways. First, the accumulation of consistent and comparable water quality assessments for some of the largest and most important hydrological systems of the nation will stand alone as a major contribution to our knowledge of regional and national water quality conditions. Second, the NAWQA national synthesis builds on and expands the findings from individual study units by combining and interpreting results from multiple study units together with historical information reported by the USGS and other agencies and researchers. National synthesis analyses produce regional and national assessments for priority water quality topics by comparative analysis of study unit findings. National synthesis efforts focused on pesticides and nutrients began in 1991; synthesis of data on volatile organic compounds began in 1994. Draft plans for synthesis of data on trace elements and stream ecology were developed in 1998 and are currently in peer review.

External coordination at all levels is an integral component of the NAWQA Program. Information exchange and coordination occur through several mechanisms. First, study unit liaison committees help ensure that the water quality information produced by the program is relevant to regional and local interests. The liaison committees are comprised of non-USGS members who represent a balance of technical and management interests. Represented organizations include, as appropriate, federal, state, interstate, and local agencies; Indian nations, and universities. At the national level, the NAWQA Advisory Committee, a subcommittee of the National Water-Quality Monitoring Council, helps ensure that federal and nonfederal interests and needs at the regional and national levels are met. Finally, the USGS and NAWQA have established liaison positions with the U.S. Environmental Protection Agency's (EPA) Offices of Pesticide Programs and Ground Water and Drinking Water to provide a stronger and more timely linkage with the information needs of these two groups.

Study unit investigations are designed to meet national synthesis requirements for consistent and integrated information and study unit requirements for assessing water quality with sufficient flexibility to adapt to local conditions. Each study unit investigation consists of four interrelated components: (1) retrospective analysis; (2) occurrence and distribution assessment; (3) trend and change assessment; and (4) case studies of sources, transport, fate, and effects. The retrospective analysis consists of a review and analysis of existing water quality data. It forms the basis for addressing what is already known and what needs to be further investigated with respect to current water quality conditions, trends and changes, and understanding causes and effects. The occurrence and distribution assessment builds on findings of the retrospective analysis to complete a broad-scale geographic and seasonal

characterization of the distribution of current water quality Conditions in relation to major point and nonpoint sources and natural and background conditions. A key objective is to fill gaps in existing data for each study unit. Design features, such as physical, chemical, and biological measurements, media sampled, and spatial and temporal resolution, are consistent among the study units.

The occurrence and distribution assessment is the largest and most important component of the first three-year intensive study phase in each study unit because it provides the foundation of data and information needed for other components. Periodic repetition of selected parts of the occurrence and distribution assessment during future intensive study phases is a key part of the trend and change assessment. Results of the occurrence and distribution assessment also are used to identify the most important questions about sources, transport, fate, and effects to be addressed by the case studies. The trend and change assessment focuses on documenting long-term trends and changes, results in new questions about causes and effects, and identifies changes that need to be made in the periodic intensive study phases. Case studies are used to improve understanding of selected questions about sources, transport, fate, and effects that arise from all aspects of NAWQA investigations and often lead to changes in assessment approaches over time.

The national study design for surface water and groundwater has been described by Gilliom et al. (1995, 1998), and only a brief outline is provided here. The surface water design focuses on water quality conditions in rivers and streams using the following interrelated components:

• Water-column studies assess physical and chemical characteristics, which include suspended sediment, major ions, nutrients, organic carbon, and dissolved pesticides.
• Bed sediment and tissue studies assess trace elements and organic contaminants that are hydrophobic (tend to associate with particles and accumulate in biological tissues rather than be dissolved in water).
• Ecological studies evaluate characteristics of benthic algae, macroinvertebrate, and fish communities and physical habitat in streams.

Sampling designs for all three components rely on coordinated sampling of varying intensity and scope at integrator sites, which are chosen to represent water quality conditions of rivers in large basins that are commonly affected by complex combinations of land-use settings, and at indicator sites, which are chosen to represent water quality conditions of streams and smaller rivers associated with specific environmental settings. The most complete data collection for the three components is at a selected core of three to five integrator sites and four to eight indicator sites in each study unit, which constitute the

fixed-site monitoring network for regular collection of samples over time. A subset of two to five sites in each study unit, usually one integrator site and two to four indicator sites, is sampled more intensively than the rest, and these are the only sites for which water samples are routinely analyzed for pesticides. Samples at these sites are analyzed for 76 pesticides and 7 degradation products, accounting for about 75 percent of agricultural pesticide use in the United States and substantial commercial, garden, and home use. Low-level analyses of 87 volatile organic compounds (VOCs) are completed for samples collected at urban indicator sites only.

The 226 surface water sampling sites in the first 20 study units include a wide range of stream sizes, types, and land-use settings in major regions of the nation, but the sites were not selected to be a statistically representative sample of the nation's streams. NAWQA sites with relatively small basins (17 to 1,243 km²), include a greater prevalence of basins with large proportions of agricultural and urban land compared with all similarly sized basins in the United States, particularly the subset of sites sampled for pesticides. NAWQA sites with large basins (1,244 to 221,497 km²), which are mainly integrator sites, also have a greater prevalence of agricultural land compared with similarly sized basins in the United States, although the pattern is less clear than for the smaller basins. This bias toward agricultural and urban land is the expected consequence of the NAWQA design. Most of the streams sampled are not directly used as sources of drinking water. The agricultural and urban indicator sites can generally be viewed as extreme examples of what drinking water sources would be like in highly developed watersheds in the region. Some of the sites sampled on major rivers are close enough to water-supply withdrawals that they are reasonably representative of the source water used for drinking water, while others are in regions where groundwater or remote surface water sources are used for drinking water.

The national study design for groundwater focuses on water quality conditions in major aquifers and in recently recharged shallow groundwater associated with current and recent land uses:

• Aquifer or "study unit" surveys assess the quality of water in the major aquifer systems of each study unit. All samples collected for the aquifer surveys are analyzed for the same field characteristics and dissolved constituents that are analyzed samples from surface water sites plus VOCs. Depending on study unit priorities, trace elements were determined in samples from some of the aquifer surveys in the first 20 study unit investigations.
• Land-use studies assess the quality of recently recharged shallow ground water (generally less than 10 years old) associated with specific combinations of land uses and hydrogeological conditions. In general, the same national target constituents are analyzed in samples from the land-use studies as for the aquifer surveys. The addition of other constituents varies among land-use studies as in the aquifer surveys.

Generally, each aquifer survey and land-use study consists of sampling about 30 randomly selected sites (wells or springs) in the geographic area and aquifer zone targeted for the specific study. One sample was collected from most of the sites. Thirty-six aquifer surveys and 56 land-use studies were completed in the first 20 study units. The 36 aquifer surveys have mixed land-use influences. Of the 56 land-use studies, 41 targeted agricultural settings, 14 targeted urban settings, and 1 targeted forested setting. Although suitable geographic delineation of aquifer boundaries throughout the United States is not available to enable comparison of NAWQA groundwater study areas to groundwater resources of the entire United States (as it was for stream drainage basins), the focus of the NAWQA groundwater design on agricultural and urban settings is similar to surface water. For purposes of evaluating water quality characteristics and comparing constituent concentrations to established water quality criteria, aquifer surveys and land-use studies were reclassified according to two categories of groundwater resources:

• Drinking-water aquifers, which are currently used as sources of drinking water (though wells sampled by NAWQA were not necessarily drinking water supply wells).
• Shallow groundwater, which is recently recharged groundwater, that may or may not be currently used as a source of drinking water. Because of the nature of their design, some land-use studies were reclassified as both drinking water aquifers and shallow groundwater.

The NAWQA program has selected a wide range of physical, chemical, and biological measurements to monitor in a nationally consistent manner. Selection of these target analytes was based on their relevance to important water quality issues and on the existence of appropriate analytical methodologies. Water measurements include field measurements of stream flow, temperature, pH, dissolved oxygen, and specific conductance, and laboratory analyses of major ions, nutrients, trace elements, organic carbon, pesticides, and VOCs. Lists of analytes and, where appropriate, associated reporting limits are presented in Gilliom et al. (1995, 1998) and for pesticides and VOCs through the World Wide Web at: http://water.wr.usgs.gov/pnsp.anstrat and http://wwwsd.cr.usgs.gov/nawqa/vocns/targets91.html.

For some water quality issues the choice of target analytes is a relatively simple task. For example, there are a small number of analytes relevant to the issues of nutrient enrichment, acidification, salinity, and sedimentation, and their chemical analysis is relatively inexpensive. In contrast, for the issue of chemical contamination, selection of target analytes is much more difficult because of the large number of substances to consider and their relatively high cost of measurement.

The process used to select target VOC analytes for the groundwater component of the NAWQA Program serves as an example of how this design

element has been addressed. Fifty-five of the 87 VOCs routinely analyzed in groundwater samples by the NAWQA Program are targeted for emphasis. These compounds were selected from an initial list of 130 compounds using an eight-step screening process that considered available information about the compounds: (1) physical properties; (2) human cancer rating; (3) noncancer human-health risk; (4) toxicity to freshwater aquatic organisms; (5) occurrence data for VOCs in groundwater, surface water, and drinking water; (6) potential for atmospheric ozone depletion and bioconcentration in aquatic organisms; (7) use or potential use as a fuel oxygenate, and (8) feasibility of analysis using purge-and-trap gas chromatography/mass spectrometry (Zogorski et al., in press).

Of the initial 130 compounds, 124 were from one or more regulatory lists associated with the Safe Drinking Water Act and Clean Water Act. Six additional compounds were included on the basis of: (1) listing as a carcinogen in the EPA's Toxics Release Inventory (chloromethyl methyl ether and bis-2-chloroethyl sulfide); (2) potential for ozone depletion (1,1,2-trichloroethane, 1,2,2-trifluoroethane); and (3) use or potential use as a gasoline oxygenate (diisopropyl ether, ethyl tert-butyl ether, and tert-amyl methyl ether). Of the 55 VOC target analytes, 29 have a national enforceable drinking water regulation; 28 are classified as known, probable, or possible human carcinogens; 35 have noncancer human health effects; and 33 are known to impart taste and odor to water. For the protection of freshwater biota, 33 of the 55 target analytes have water quality guidelines established by the EPA, and 17 have water quality guidelines established by Environment Canada.

NAWQA samples for pesticides and VOCs are routinely analyzed at the USGS National Water-Quality Laboratory by several broad-spectrum methods usually involving gas chromatography/mass spectrometry. Each method is capable of detecting a number of target pesticides and/or VOC analytes. In an ideal world, modifications to target analyte lists would consider changes in chemical-use patterns and many of the other characteristics described above for screening candidate VOCs. In some situations, however, the above information may not exist. Furthermore, in addition to parent compounds, there is growing interest and concern over contaminant degradates—particularly for pesticides. In the NAWQA Program, nontarget analytes are tentatively identified using computerized searches of libraries of mass spectra. For those nontarget compounds that are frequently detected, efforts are made to obtain standards for the compound and, where appropriate, include it as a target analyte. in 1991 methyl tert-butyl ether (MTBE), a fuel oxygenate, was added to the USGS's VOC analyte list using just such an approach.

Major findings for individual study units and national synthesis teams are published in reports, many of which are available on the Internet via the World Wide Web. The NAWQA Program's home page is http://water.usgs.gov/lookup/get?nawqa.

At the end of the initial assessment, each study unit published a report summarizing major findings for the period 1992 to 1995. Included in each of these reports is a series of tables summarizing concentrations of individual contaminants detected in groundwater, surface water, bed sediment, and tissue for the study unit in relation to the national range observed across all 20 NAWQA study units and in relation to existing environmental standards and guidelines.

Results from the first 20 study units during 1992 to 1996 include analyses of 76 pesticides and seven pesticide degradates in more than 8,000 samples of water from rivers, streams, and wells. The 76 pesticides account for about 75 percent of national agricultural use Coy weight) and a substantial portion of urban and suburban use. More than 95 percent of all samples collected from streams and rivers contained at least one pesticide compared to about 50 percent for groundwater. Seventy-four of the 83 pesticide compounds were detected at least once in streams and ground water. Most detections in streams were greater than 0.01 µ/L and more than haft were greater than 0.05 µ/L. Agricultural and urban streams, as well as major rivers, had relatively similar high frequencies of detection, although the highest concentrations generally occurred in the smaller streams. Detection frequencies in ground water were highest in shallow groundwater in agricultural areas, somewhat lower in shallow groundwater in urban areas, and lowest in major aquifers. The shallow groundwater is generally derived from relatively recent recharge within the land-use area of interest. The major aquifers are generally deeper, have variable land use influences, and were sampled using existing production wells.

Concentrations of the 21 most commonly detected pesticides exceeded 0.05 µ/L in more than 10 percent of the stream samples or in more than 1 percent of groundwater samples at least one of the land-use categories. The most frequently detected pesticides in agricultural areas were the major herbicides: atrazine and its degradation product deethylatrazine, metolachlor, cyanazine, and alachlor, which rank first, second, fourth, and fifth, respectively, in national herbicide use for agriculture. These most heavily used herbicides also account for most of the detections in larger rivers and major aquifers and many detections in urban streams and groundwater. The herbicides that were generally found most often in urban streams are simazine, prometon, 2,4-D, diuron, and tebuthiuron, with simazine and prometon accounting for most detections in

streams and shallow groundwater. 2,4-D and prometon rank first and fourteenth among herbicides in frequency of home and garden use, and 2,4-D, simazine, and diuron rank third, eighteenth, and twenty-third, respectively, in national herbicide use for agriculture. Prometon and tebuthiuron have no reported agricultural use. Insecticides were detected much more frequently in urban streams than in agricultural streams and were seldom detected in groundwater in any setting. Most detections were accounted for by diazinon, carbaryl, malathion, and chlorpyrifos, which nationally rank first, eighth, thirteenth, and fourth, respectively, among insecticides in frequency of home use.

Low-level mixtures are the most common form of pesticide occurrence in streams and groundwater. More than 50 percent of all stream samples contained five or more pesticides, and about 25 percent of groundwater samples had two or more pesticides. In accordance with use patterns, the composition of the most common mixtures differs between urban and agricultural areas and among agricultural areas with different crops and pests. For example, simazine and prometon were present in the most commonly occurring mixtures of two or more compounds in urban areas, whereas atrazine, deethylatrazine (DEA), and metolachlor were the most common compounds in mixtures found in agricultural areas. The most common pair of compounds found in groundwater in agricultural areas was atrazine and DEA. A distinctive feature of urban streams was the common occurrence of mixtures with both herbicides and insecticides. More than 10 percent of urban stream samples contained a mixture of at least. four herbicides plus diazinon and chlorpyrifos.

Drinking water standards for individual pesticides were rarely exceeded in streams or groundwater, but aquatic life criteria were commonly exceeded in some streams. Most of the major aquifers and about half of the shallow groundwater zones sampled are sources of drinking water. Most concentrations are substantially below EPA drinking water standards, which were exceeded in less than 1 percent of the wells sampled. In streams, peak concentrations of several herbicides frequently occurred above EPA drinking water standards in some agricultural areas, but annual average concentrations, which are used for regulation, rarely exceeded standards. For drinking water, NAWQA results are generally good news regarding individual pesticides and current regulations. This conclusion is tempered, however, by the fact that criteria are not established for many pesticides and mixtures and degradation products are not considered. Furthermore, a limited range of potential effects have been assessed.

Preliminary results from the first cycle of 20 NAWQA study units are summarized and available on the Internet via the World Wide Web at: http://wwwsd.cr.usgs.gov/nawqa/vocns/datasum91.html. The summary represents analyses of VOCs in 1,600 groundwater samples collected in 1993 to 1995. For most VOCs, the minimum reporting level was 0.2 µ/L. Forty-six VOCs were detected in groundwater. VOCs were most frequently detected in shallow groundwater in urban areas, in comparison to shallow groundwater in agricultural areas and major aquifers. About 54 percent of the samples of shallow ground water in urban areas contained one or more VOCs. Frequently detected VOCs

included trichloromethane (chloroform), MTBE, tetrachloroethene (perchloroethylene), and trichloroethene (trichloroethylene). While VOCs were frequently detected in shallow groundwater, similar to pesticides, their concentrations were low and almost always below maximum contaminant levels or health advisories for drinking water.

To augment data collected by the NAWQA Program for assessment of the "occurrence" of VOCs in groundwater nationwide, the USGS has compiled a national dataset of VOC analyses of samples of groundwater collected between 1985 to 1995 by 43 federal, state, and local nonpoint source monitoring programs (Lapham et al., 1997, 1998). The dataset includes analyses representing 5,320 wells in 20 states. Eight attributes of this dataset were evaluated to determine the suitability of the data to augment NAWQA data in answering occurrence questions of varying complexity. These eight attributes are: (1) VOC analytes, (2) associated reporting levels, (3) water use, (4) well-casing materials, (5) well depth, (6) depth to water level in the well, (7) aquifer lithology, and (8) distribution of sampling points.

Twelve of the 13 most frequently analyzed VOCs are in the use category of solvents, industrial reagents, and refrigerants. Other compounds of current interest were not frequently analyzed for. MTBE was analyzed for in only about 6 percent of the samples. Three other oxygenates (diisopropyl ether, ethyl tertiary-butyl ether, and tertiary-amyl methyl ether) were not included in any of the analyses in the dataset. About 70 percent of the sampled wells have the associated water use documented in the dataset. However, the dataset generally lacks documentation of the other characteristics of interest. Three modifications to these monitoring programs that would enhance the suitability of the resulting data for determining the occurrence of VOCs are: (1) expand the VOC analyte list beyond currently regulated compounds, (2) consistently record the reporting level for each analyte for every analysis, and (3) consistently record key ancillary information about each well.

The NAWQA program provides information useful in identifying emerging contaminants in the nation's streams, rivers, and groundwater used for drinking water in several ways. First, NAWQA study units cover almost half of the conterminous United States and represent some of the largest and most important hydrological systems. These hydrological systems differ widely in the natural and human factors that affect water quality and thus provide a monitoring framework that is useful for identifying existing and emerging contaminants of regional and national concern.

Second, the first major component of a study unit investigation is focused on determining the broad-scale geographic and seasonal distribution of water quality conditions in relation to major point and nonpoint contaminant sources. As part of this initial component, bed sediment, tissue, and ground and surface water samples are analyzed for a large number of potentially toxic trace elements, pesticides, and other synthetic organic contaminants of concern.

Analytical reporting limits are typically much lower than many other routine monitoring efforts. In addition, nontarget analytes are tentatively identified using computerized searches of libraries of mass spectra. This combined approach is designed to fill gaps in existing data and to help identify the most important existing and emerging contaminants of concern.

Third, the NAWQA program is perennial and places a high emphasis on repetition of measurements with time and on documentation of both the methods of data collection and analysis and the locations and characteristics of data collection sites. This type of approach is essential for identifying emerging contaminants and changes in concentrations of contaminants in relation to changes in land- and water-use activities.

Fourth, in addition to investigations of individual hydrological systems, national synthesis currently focused on pesticides, VOCs, nutrients, and trace elements helps to identify regional and national patterns in the occurrence and distribution of contaminants of concern and their relation to contaminant sources.

Finally, communication and coordination with scientists and managers in other federal, state, and local agencies are critical components of the NAWQA Program. The ongoing exchange of information about emerging water quality issues and new contaminants of concern, improved methods of measurement, and new sources of data provides some of the additional insight necessary for the program to evolve and to be relevant to future water quality information needs.

Gilliom, R. J., W. M. Alley, and M. E. Gurtz. 1995. Design of the National Water-Quality Assessment Program: Occurrence and distribution of Water-Quality Conditions. USGS Circular 1112. Reston, Va.: U.S. Geological Survey.

Gilliom, R. J., D. K. Mueller, and L. H. Newell. 1998. Methods for Comparing Water-Quality Conditions Among National Water-Quality Assessment Study Units, 1992-1995. USGS Open-File Report 97-589. Reston, Va.: U.S. Geological Survey.

Hirsch, R. M., W. M. Alley, and W. G. Wilber. 1988. Concepts for a National Water Quality Assessment Program. USGS Circular 1021. Reston, Va.: U.S. Geological Survey.

Lapham, W. W., K. M. Neitzert, M. J. Moran, and J. S. Zogorski. 1997. USGS compiles dataset for national assessment of VOCs in ground water. Ground Water Monitoring and Remediation (fall issue):147-157.

Lapham, W. W., M. J. Moran, and J. S. Zogorski. 1998. Enhancements of nonpoint source monitoring programs to assess volatile organic compounds in the nation's ground water. Pp. 371-381 in Proceedings of the National Water Quality Monitoring Council, National Conference on Monitoring—Critical Foundations to Protect Our Waters, Reno, Nev., July 7-9, 1998.

Leahy, P. P., J. S. Rosenshein, and D. S. Knopman. 1990. Implementation Plan for the National Water-Quality Assessment Program. USGS Open-File Report 90-174. Reston, Va.: U.S. Geological Survey.

Zogorski, J. S., D. A. Bender, M. J. Halde, and B. L. Rowe. In press. Selection Procedure and Salient Information for Volatile Organic Compounds Emphasized in the National Water-Quality Assessment Program . USGS Open-File Report. Reston, Va.: U.S. Geological Survey.