USGS Home
Contact USGS
Search USGS

 Go to:      NAWQA Home     National-Synthesis Assessments

Trace Elements National Synthesis Project

1998 conference paper--
Arsenic in Ground Water Supplies of the United States

Citation: Welch, A.H., Helsel, D.R., Focazio, M.J., and Watkins, S.A., 1999, Arsenic in ground water supplies of the United States, in: Arsenic exposure and health effects, W.R. Chappell, C.O. Abernathy and R.L. Calderon, Eds., Elsevier Science, New York, pp. 9-17.

• Abstract
• Introduction
• Arsenic in ground water
• Estimating arsenic in regulated water supplies
• Comparison of public and non-public ground water samples
• Estimates of the frequency of exceedance
• Discussion
• References

Abstract

High arsenic concentrations in ground water have been documented in many areas of the United States. Within the last decade, parts of Maine, Michigan, Minnesota, South Dakota, Oklahoma, and Wisconsin have been found to have widespread arsenic concentrations exceeding 10 mg/L. These high concentrations most commonly result from: (1.) upflow of geothermal water, (2.) dissolution of, or desorption from, iron-oxide, and (3.) dissolution of sulfide minerals.

Because the MCL for arsenic is currently being evaluated, estimating the exceedance frequency for different arsenic concentrations in regulated water supplies is particularly timely. Estimates of the frequency of exceedance, which are based on analyses of about 17,000 ground water samples, suggests that about 40% of both large and small regulated water supplies have arsenic concentration greater than 1 mg/L. The frequency of exceedance decreases for greater arsenic concentrations -- about five percent of systems are estimated to have arsenic concentration greater than 20 mg/L. Comparison of these estimates with previously published work, based on 275 samples collected from regulated water supplies, shows very good agreement for the United States as a whole, although the two approaches yield somewhat different results for some parts of the nation.

Introduction

Arsenic can impact human health through the ingestion of ground water used for water supply. An understanding of arsenic concentrations in ground water can: (1) assist water managers and users in overcoming adverse health effects through avoidance or treatment, (2) assist epidemiologists interested in evaluating the intake of arsenic from drinking water, which can contribute much of the human exposure to inorganic arsenic (Borum and Abernathy, 1994), and (3) provide a basis for evaluating the costs of adopting a particular value for a drinking-water standard (or MCL - Maximum Contaminant Level). Estimating the frequency of exceedance for arsenic in regulated ground water supplies is the focus of this manuscript. The estimates were made for values in the range being considered for a revised MCL. These estimates are timely because the U.S. Environmental Protection Agency (EPA) must issue a proposed and a final regulation for arsenic in drinking water by 2000 and 2001, respectively.

Arsenic in Ground Water

Within the conterminous United States, widespread high arsenic concentrations in ground water most commonly result from: (1.) Upflow of geothermal water, (2.) dissolution of, or desorption from, iron-oxide, (3.) dissolution of sulfide minerals, and (4.) evaporative concentration. Figure 1 and Table 1 indicate areas where these processes appear to be important. Concentrations of naturally occurring arsenic in ground water vary regionally due to a combination of climate and geology. At a broad regional scale, arsenic concentrations exceeding 10 mg/L appear to be more frequently observed in the western U.S. than in the east (Welch et al., 2000). Investigations of ground water in Maine, Michigan, Minnesota, South Dakota, Oklahoma, and Wisconsin within the last decade suggest that arsenic concentrations exceeding 10 mg/L are more widespread and common than previously recognized (Table 1).

Arsenic release from iron oxide appears to be the most common cause of widespread arsenic concentrations exceeding 10 mg/L in ground water. This can occur in response to different geochemical conditions, including release of arsenic to ground water through reaction of iron oxide with organic carbon. Iron oxide also can release arsenic to alkaline ground water, such as that found in some felsic volcanic rocks, including areas 8, 14, and 16 (Figure 1) in the western U.S. Geothermal water and high evaporation rates also are associated with arsenic concentrations greater than 10 mg/L in ground water, particularly in the west. Geothermal systems occur throughout much of the western United States, although most affect relatively small regions. A notable exception is the Yellowstone geothermal system, which causes arsenic concentrations as high as 360 mg/L in the Madison River at the park boundary, and as high as 19 mg/L arsenic in the Missouri River at a point 470 km downstream (Nimick, 1994; Nimick et al., 1998). These studies clearly demonstrate the existence of ground water with arsenic concentrations that exceed both the current and possible new MCL for arsenic.

(Larger source-areas map is an 18Kb GIF.)
Figure 1. Areas where high arsenic concentrations have been documented in ground-water systems. Physiographic provinces modified from Fenneman (1931).

Table 1. Areas with high arsenic concentrations derived from natural sources

¹ Known or inferred. For areas with Fe-oxide as a source of arsenic, dissolution of the oxide and desorption are important processes that can release arsenic to ground water. The letters `D' and `P' in parentheses refer to the processes of dissolution and pH-influenced desorption of arsenic, respectively.

² Arsenic concentrations in ground water are generally higher in bedrock aquifers compared with overlying glacial aquifers.

³ The sandstone contains arsenic rich pyrite, which may be a source of the arsenic in the overlying glacial aquifer. Pyrite has not been identified in the glacial deposits.

⁴ May include a contribution of arsenic from underlying coal-bearing units. Arsenic-rich ground water may extend into the upper Kankakee River basin within Indiana, as suggested by high arsenic in surface water, sediment and biota (Fitzpatrick et al., 1998; Schmidt and Blanchard, 1997).

⁵ Arsenopyrite has been mentioned as a possible source of arsenic. However, high pH (the median pH of 11 samples with arsenic >50 µg/L is 8.25) and generally low sulfate concentrations (<25 µg/L; Ficklin et al., 1989) imply that sulfide mineral oxidation is limited, suggesting that the arsenic may be from Fe-oxide that was formed from the oxidation of arsenopyrite. The ground water with the highest arsenic concentration (15,000 µg/L) also had the highest pH (9.23).

Estimating Arsenic in Regulated Water Supplies

The arsenic content in ground water used to supply water for regulated systems has been estimated from arsenic data from across the United States. The basic approach consisted of combining arsenic data for ground water with the locations of water supplies that use ground water (Figure 2). The arsenic data were retrieved from the U.S. Geological Survey's National Water Information System (NWIS). These data, along with other water-quality measurements and ancillary data are collectively referred to as the USGS Arsenic Database. Within this database, geothermal water (water with a temperature > 50^o C) and slightly saline water -- (dissolved solids >3,000 mg/L or specific conductance >4,000 mS/cm) are considered to be unlikely to be a source for a regulated water supply. Accordingly, analyses of this ground water were excluded from the evaluation.

Figure 2. Approach used to estimate arsenic exceedance in regulated supplies using ground water.

The USGS integrated the arsenic occurrence data with information on public suppliers and population served. This was accomplished by relating arsenic concentrations on a county level from the USGS data base with public water supplier data from US EPA's Safe Drinking Water Information System (SDWIS) database. Data were retrieved from SDWIS for all (surface water, ground water, and purchased water) community public suppliers and their sources during late summer of 1997. Exceedance frequencies were estimated for public water supplies for two system sizes, based on population served. The size classification corresponds to that used by Frey and Edwards (1997) to allow direct comparison with their estimates. The size classes are large systems -- those that serve populations greater than 10,000 -- and small system, which serve 1,000 to 10,000 people.

Comparison of Public and Non-public Ground Water Supplies

Ancillary data within the USGS Arsenic Database includes information on the use of the water. These data allow a comparison of arsenic concentrations in water that is used for public supply with water that is not used for public supply. The classification as public supply was based on primary use of water listed in NWIS as bottling, commercial, medicinal, public supply, or institutional; all other primary use categories were considered non-public supply. It is worth emphasizing that the use of the term 'public water supply' is not synonymous with 'regulated water supply,' where the latter refers to systems that are subject to the Safe Drinking Water Act. The term 'public water supply' is an informal term used here and refers to the water uses listed above.

The arsenic analyses for public and non-public supply sources were compared using nonparametric tests employing the procedure NPAR1WAY in the SAS/STAT® software system (SAS Institute Inc., 1990, p. 1195-1210). The statistical comparisons indicated that arsenic concentrations in non-public supply samples tended to be higher than concentrations in public supply samples. However, the magnitude of the differences is not very large -- the results are so significant due to the large amount of data used. The medians of both the non-public and public supply samples are < 1 µg/L. The test results are summarized in the following table.

When the data are split into Physiograhic Provinces, only the Atlantic Plain had a significant difference in arsenic concentration between public supplies and non-public ground water sources (Table 2). Yet, the medians for both groups are < 1 mg/L.

Table 2. Comparison of arsenic concentrations in public and non-public water supplies, by physiographic province. A total of 18,468 sites were used.

These data suggest that the magnitude of difference in arsenic concentrations between public supply and other data represented in the USGS Arsenic Database is very small. Although the USGS data were not collected with the specific intent of developing national occurrence estimates in drinking water, these data appear to be appropriate for characterizing arsenic concentrations in ground water used for public supply.

Estimates of the Frequency of Exceedance

A common way to summarize national occurrence data for use in the regulatory process is to present the data in terms of the frequency that systems exceed specified concentrations of a contaminant. The systems are categorized by size classes and the concentrations are often selected to correspond with ranges of potential drinking water standards that are being assessed as part of the regulatory process. Estimates of arsenic exceedance in regulated water supplies are based on 17,496 arsenic analyses for 595 counties that have five or more sites represented in the USGS Arsenic Database (Figure 3). About 47% of large systems are located in these counties. About 32% of all small systems are located in those same 595 counties.

(2Kb, GIF)
Figure 3. Counties with five or more sites with ground-water samples analyzed for arsenic.

Estimates based on the USGS Arsenic Database suggest that an MCL of 20 mg/L would increase the frequency of exceedance over that estimated for the current MCL of 50 mg/L, although less than five percent of both the small and large systems would be affected at either level. Decreasing the MCL below 10 would result in a substantial increase in the frequency of exceedance, with about forty percent of the systems exceeding the standard if an MCL of 1 mg/L were adopted. Generally, the frequency of exceedance for the large and small systems at the various levels shown on Figure 4 are similar.

Figure 4. Exceedance frequency of arsenic concentrations in small and large regulated water supply systems.

Estimates of exceedance based on the USGS Arsenic Database are broadly similar to previous estimates (Figure 4) using the NAOS (National Arsenic Occurrence Survey) by Frey and Edwards (1997). (Comparisons were not possible at concentrations of 20 and 50 µg/L because they were not published for the NAOS data.) This broad agreement is somewhat surprising considering the differences in the two databases. NAOS includes 275 filtered samples of ground water used by regulated water supplies, whereas over 17,000 analyses were used in the approach described above. The exceedance estimates based on the NAOS database also were adjusted for treatment, which resulted in lower arsenic values. This adjustment is one factor that could account for the generally lower frequency of exceedance compared with the estimates based on the USGS Database. The adjustment was not large, however, ranging from zero to ten percent (Frey and Edwards, 1997).

The agreement between the approach discussed above with that of Frey and Edwards (1997) ranges from very good to poor at 5 µg/L for regions of the United States (Figure 5). The exceedance estimates for the Western, Midwest, and New England regions show very good agreement. Although the estimates for the South East region differ by more than a factor of two, both suggest that this region has the lowest frequency of exceedance at the at 5 µg/L level. The poor agreement for the North Central, South Central, and Mid-Atlantic regions may, in part, be due to combination of sparse data and the approach described above. For instance, only a few counties are represented in the USGS estimates for several of the states in the Mid-Atlantic region (Figure 3). One weakness of the current approach is that counties are relatively small in the central and eastern United States compared to the west. An undesirable artifact of the approach is that the data density for the east must be much greater to include a county in the frequency estimates. An alternative approach is to include data within a search radius greater than the size of most counties in the central and eastern United States.

 
Figure 5. Exceedance frequencies of arsenic at 5 µg/L for seven regions. Regions and exceedance frequency data for NAOS are from Frey and Edwards, 1997. The western region includes Alaska and Hawaii, although the USGS Arsenic database has less than 5 samples for the state.

Discussion

Estimates of the frequency of exceedance based on the USGS Arsenic Database suggest that about 40% of both large and small water supplies have arsenic concentration greater than 1 µg/L. The frequency of exceedance decreases for greater arsenic concentrations -- about five percent of systems are estimated to have arsenic concentration greater than 20 µg/L. Comparison of these with the estimates based on the NAOS Database shows very good agreement for the United States as a whole, although the two approaches yield somewhat different results for some parts of the nation. This agreement suggests that the USGS Arsenic Database may be used to estimate of the frequency of exceedance for the large number of systems that supply small populations.

References

Ayotte, J. D., M. G. Nielson, and G. R. Robinson. 1998. Relation of arsenic concentrations in ground water to bedrock lithology in eastern New England. 1998 Geol. Soc. of Am. Annual meeting abstracts with programs. pp. A-58.

Ball, J. W., D.K. Nordstrom, E.A. Jenne, and D.V. Vivit. 1998. Chemical Analyses of Hot Springs, Pools, Geysers, and Surface Waters from Yellowstone National Park, Wyoming, and Vicinity, 1974-1975.U.S. Geol. Surv. Open-File Rep. 98-182. p. 45.

Borum, D.R. and C.O. Abernathy. 1994. Human oral exposure to inorganic arsenic. In: W.R. Chappell, C.O. Abernathy, and C.R. Cothern (eds.), Arsenic exposure and health. Sci. and Tech. Letters, Northwood. v. 16, pp. 31-30.

Boudette, E. L., F. C. Canney, J. E. Cotton, R. I. Davis, W.H. Ficklin, and J.M. Motooka. 1985. High levels of arsenic in the groundwater of southeastern New Hampshire. A Geochemical Reconnaissance. U. S. Geol. Surv. Open-File Rep. pp. 25.

Carter, J.M., S.K. Sando, T.S. Hayes, and R.H. Hammond. 1998. Source Occurrence, and Extent of Arsenic Contamination in the Grass Mountain Area of the Rosebud Indian Reservation, South Dakota. U.S. Geol. Surv. Wat. Res. Inv. Rep. 97-4286. p. 90.

Davies, J., R. Davis, D. Frank, F. Frost, D. Garland, S. Milham, R.S. Pierson, R.S. Raasina, S. Safioles, and L. Woodruff. 1991. Seasonal study of arsenic in ground water: Snohomish County, Washington. Snohomish Health District and Washington State Dept. of Health unpublished report. p. 18.

Eccles, L.A. 1976. Sources of arsenic in streams tributary to Lake Crowley California. U. S. Geol. Surv. Wat. Res. Inv. Rep. 76-36. pp. 76-86.

Fenneman, N.M. 1931. Physiography of western United States. McGraw-Hill, New York. p. 534

Ficklin, W.H., D.G. Frank, P.K. Briggs, and R.E. Tucker. 1989. Analytical results for water, soil, and rocks collected near the vicinity of Granite Falls Washington as part of an arsenic-in-groundwater study. U. S. Geol. Surv. Open-File Rep. 89-148. p. 9.

Fitzpatrick, F., T.L. Arnold, and J.A. Colman. 1998. Surface-Water-Quality Assessment of the Upper Illinois River Basin in Illinois, Indiana, and Wisconsin-Spatial distribution of geochemicals in the fine fraction of streambed sediment, 1987. U.S. Geol. Surv. Wat. Res. Inv. Rep. 98-4109. p. 89.

Frey, M.M. and M.A. Edwards. 1997. Surveying arsenic occurrence. J. Am. Water Works Assoc. v. 89, pp. 105-117.

Fujii, R. and W.C. Swain. 1995. Areal distribution of trace elements, salinity, and major ions in shallow ground water, Tulare basin, Southern San Joaquin Valley, California. U.S. Geol. Surv. Wat. Res. Inv. Rep. 95-4048. p. 67.

Goldblatt, E.L., S.A. Van Denburgh, and R.A. Marsland. 1963. The unusual and widespread occurrence of arsenic in well waters of Lane Country, Oregon. Lane County Health Dept. Rept. p. 24.

Goldstein, L. 1988. A review of arsenic in ground water with an emphasis on Washington state. Unpublished M.S. thesis, Evergreen State College, Olympia, Washington. p. 102.

Hinkle, S.R. 1997. Quality of shallow ground water in alluvial aquifers of the Willamette Basin, Oregon, 1993-95. U.S. Geol. Surv. Wat. Res. Inv. Rep. 97-4082-B. p. 48.

Holm, T.R. 1995. Ground-water quality in the Mahomet Aquifer, McLean, Logan, and Tazewell Counties. Illinois State Water Survey Contract Report 579. p. 42.

Holm, T.R. and C.D. Curtiss. 1988. Arsenic contamination in east-central Illinois ground waters. Illinois Dept. of Energy and Natural Res., Energy and Environmental Affairs Division. p. 63.

Kanivetsky, R. in press. Arsenic in ground water of Minnesota: Hydrogeochemical modeling and characterization. Minnesota Geol. Surv. Rep. p. zz.

Kolker, A., W.F. Cannon, D.B. Westjohn, and L.G. Woodruff.1998. Arsenic-rich pyrite in the Mississippian Marshall Sandstone: Source of anomalous arsenic in southeastern Michigan ground water. 1998 Geol. Soc. of Am. Annual meeting abstracts with programs. pp. A-59.

Korte, N. E. 1991. Naturally Occurring Arsenic in Groundwaters of the Midwestern United States. Oak Ridge National Laboratory Environmental Tech. Section Informal Rep. p. 20.

Korte, N.E. 1995. Naturally Occurring Arsenic in the Groundwater at Chanute Air Force Base, Rantoul, Illinois. Environmental Sciences Division Publication no. 3501. p. 54.

Mariner, R.H. and L.M. Willey. 1976. Geochemistry of Thermal Waters in Long Valley, Mono County, California. J. of Geophys. Res. v. 81, no. 5, pp. 792-800.

Marvinney, R.G., M.C. Loiselle, J.T. Hopeck, D. Braley, and J.A. Krueger. 1994. Arsenic in Maine Groundwater: An Example From Buxton, Maine. 1994 Focus Conference on Eastern Regional Ground Water Issues. pp. 701-714.

Matisoff, G., C.J. Khourey, J.F. Hall, A.W. Varnes, and W.H. Strain. 1982. The nature and source of arsenic in northeastern Ohio Ground Water. Ground Water. v. 20, no. 4, pp. 446-456.

Nadakavukaren, J.J., R.L. Ingermann, and G. Jeddeloh. 1984. Seasonal Variation of Arsenic Concentration in Well Water in Lane County, Oregon. Bull. of Environ. Contamination and Toxicology. v. 33, no. 3, pp. 264-269.

Nimick, D.A. 1994. Arsenic transport in surface and ground water in the Madison and upper Missouri River Valleys, Montana. EOS. pp. 247.

Nimick, D.A., Moore, J.N., C.E. Dalby, and M.W. Savka. 1998. The fate of arsenic in the Madison and Missouri Rivers, Montana and Wyoming. Water Resour. Res. v. 34, no. 11, pp. 3051-3067.

Norvell, J.L.S. 1995. Distribution of, sources of, and processes mobilizing arsenic, chromium, selenium, and uranium in the central Oklahoma aquifer. Unpublished M.S. thesis, Colorado School of Mines, Golden, Colorado. p. 169.

Owen-Joyce, S.J. and C.K. Bell. 1983, Appraisal of water resources in the Upper Verde River area, Yavapai and Coconino Counties, Arizona. Arizona Dept. of Water Res. Bull. 2. p. 219

Owen-Joyce, S.J. 1984. Hydrology of a Stream-Aquifer System in the Camp Verde Area, Yavapai County, Arizona. Arizona Dept. of Water Res. Bull. 3. p. 60.

Panno, S. V., K.C. Hackley, K. Cartwright, and C.L. Liu. 1994. Hydrochemistry of the Mahomet Bedrock Valley Aquifer, East-Central Illinois: Indicators of Recharge and Ground-Water Flow. Ground Water. v. 32, no. 4, pp. 591-604.

Peters, S.C., J.D. Blum, B. Klaue, and M.R. Karagas. 1998. Arsenic Occurrrence in New Hampshire Ground Water. 1998 Geol. Soc. of Am. Annual meeting abstracts with programs. pp. A-58.

Roberts, K., B. Stearns, and R. L. Francis. 1985. Investigation of arsenic in southeastern North Dakota ground water, A Superfund Remedial Inv. Rep. North Dakota State Dept. of Health. p. 225

Robertson, F.N. 1989. Arsenic in ground-water under oxidizing conditions, south-west United States. Environ. Geochemistry and Health. v. 11, no. 3/4, pp. 171-186.

SAS Institute Inc., 1990, SAS/STAT® User's Guide, Version 6, Fourth Edition, Volume 2: Cary, North Carolina, SAS Institute Inc., p. 1686.

Schlottmann, J.L. and G.N. Breit. 1992. Mobilization of As and U in the central Oklahoma aquifer. In: Y.K. Kharaka and A.S. Maest (eds.), Water-Rock interaction. Balkema, Rotterdam. pp. 835-838.

Schmidt, A.R. and S.F. Blanchard. 1997. Surface-water-quality assessment of the upper Illinois River Basin in Illinois, Indiana, and Wisconsin-Results of Investigations through April 1992. U.S. Geol. Surv. Wat. Res. Inv. Rep. 96-4223. p. 63.

Simo, J.A., P.G. Freiberg, and K.S. Freiburg. 1996. Geologic constraints on arsenic in groundwater with applications to groundwater modeling: Groundwater Research Rept. WRC GRR 96-01, University of Wisconsin. p. 60

Stauffer, R.E. and J.M. Thompson. 1984. Arsenic and antimony in geothermal waters of Yellowstone National Park, Wyoming, USA. Geochim. et Cosmochim. Acta. v. 48, no. 11, pp. 2547-2561.

Swartz, R.J. 1995. A study of the occurrence of arsenic on the Kern Fan element of the Kern Water Bank, southern San Joaquin Valley, California. Unpublished M.S. thesis, California State University, Bakersfield, California, p. 138.

Swartz, R.J., Thyne, G.D., and Gillespie, J.M. 1996. Dissolved arsenic in the Kern Fan San Joaquin Valley, California: Naturally occurring or anthropogenic. Environmental Geosciences. v. 3, no. 3, pp. 143-153.

Voelker, D.C. 1986. Observation-well network in Illinois, 1984. U.S. Geol. Surv. Open-File Rep. 86-416. p. 108.

Welch, A.H. and Lico M.S. 1998. Factors controlling As and U in shallow ground water, southern Carson Desert, Nevada. Appl. Geochem. v. 13, no. 4, pp. 521-539.

Welch, A.H., D.B. Westjohn, D.R. Helsel, and R.B. Wanty. 2000. Arsenic in ground water of the United States: Occurrence and geochemistry: Ground Water, v.38, no.4, p.589.

Westjohn, D.B., A. Kolker, W.F. Cannon, and D.F. Sibley. 1998. Arsenic in ground water in the "Thumb Area" of Michigan. The Mississippian Marshall Sandstone Revisited, Michigan: Its Geology and Geologic Resources, 5th symposium. pp. 24-5.

Wilkie, J.A., and J.G. Hering. 1998. Rapied oxidation of geothermal arsenic(III) in steamwaters of the eastern Sierra Nevada. Environ. Sci. Technol. v. 32, pp. 657-662.

Yarling, M. 1992. Anomalous concentrations of arsenic in the groundwater at Wakarusa, Indiana: A byproduct of chemical weathering of shales. Indiana Dept. of Environmental Management. p. 38.

Ziegler, A.C., W.C. Wallace, D.W. Blevins, and Maley, R.D. 1993. Occurrence of Pesticides, Nitrite Plus Nitrate, Arsenic, and Iron in Water From Two Reaches of the Missouri River alluvium, northwestern Missouri--July 1988 and June-July 1989. U.S. Geol. Surv. Open-File Rep. 93-101. p. 30.

Zuena, A.J. and P.E. Keane. 1985. Arsenic Contamination of Private Potable Wells. EPA National Conference on Environmental Engineering Proceedings, Northeastern University Boston, MA. pp. 717-725.