HEALTH EFFECTS OF ARSENIC EXPOSURE IN BANGLADESH: PROGRESS AND PRELIMINARY FINDINGS FROM A COHORT STUDY.

Habibul Ahsan, M.D.

Department of Epidemiology and Herbert Irving Comprehensive Cancer Center, Columbia University, New York, NY 10032. An estimated 50 million people have been chronically exposed to arsenic exposure from drinking water. We are conducting a large epidemiologic cohort study of 11,500 men and women to comprehensively examine prospectively the health effects of arsenic exposure in Bangladesh with an initial emphasis on the full dose-response relationships of arsenic exposure with the incidence rates of skin lesions, skin cancers, and total and cancer-related mortalities. In addition, using cross-sectional and case-cohort designs within the main cohort, the interrelationships among urinary arsenic metabolites, a number of biomarkers and health outcomes are also being examined. To date, a complete survey and analysis of all 6,000 tube-wells in the study area are completed enumerating and characterizing the 60,000 users. Recruitment and collection of extensive interview and clinical data and biological samples have also been completed from the target 11,500 cohort members. Overall, data from this comprehensive study will provide information on the interrelationships between arsenic doses, intermediate biomarkers, and other host and lifestyle factors including the genetics and nutrition in the etiology of arsenic-induced cancers and other health outcomes. Design and progress of this comprehensive epidemiologic study, preliminary baseline data, and a preliminary analysis of cancer burden in Bangladesh based on the baseline data will be presented.



EFFECTS OF ARSENIC ON EXPRESSION OF DNA REPAIR GENES.

Angeline S. Andrew^1,2,3, Margaret R. Karagas^1,3, and Joshua W. Hamilton^2,3.

¹Department of Community and Family Medicine, ²Department of Pharmacology and Toxicology, ³Center for Environmental Health Sciences at Dartmouth, Dartmouth Medical School, Hanover, NH 03755-3835.

Exposure to high levels of arsenic in drinking water is associated with the occurrence of several types of cancers including lung, bladder and skin, as well as vascular disease and diabetes. However, the mechanism by which arsenic acts as a human carcinogen is not known. At low levels, arsenic is not directly genotoxic, but likely acts as a co-mutagen. We hypothesize that arsenic may act as a carcinogen, at least in part, through inhibition of DNA repair mechanisms, leading indirectly to increased mutations from other DNA damaging agents. In cell culture, low concentrations of arsenic inhibited nucleotide excision repair (NER) after UV irradiation, and specifically decreased incision frequency. Nucleotide excision repair is a major DNA repair pathway that removes DNA lesions including certain DNA crosslinks, UV photolesions, and bulky chemical adducts. The nucleotide excision repair system requires the cooperative function of many gene products for damage recognition, incision, excision, elongation, and ligation to restore DNA structure. The molecular mechanism underlying inhibition of nucleotide excision repair by arsenic is unknown, but could be due to decreased expression of critical genes involved in nucleotide excision repair. To test this hypothesis, we isolated mRNA from cryopreserved lymphocytes taken from a subset of individuals enrolled in a population based case-control study of bladder cancer in New Hampshire. Nucleotide excision repair gene expression was assessed by RT-PCR of nine individual genes normalized to GAPDH. Arsenic levels were determined in toenail clippings using instrumental neutron activation analysis and drinking water samples using high resolution ICP-MS with hydride generation. In a linear regression analysis, toenail arsenic levels were inversely correlated with expression of critical members of the nucleotide excision repair incision complex, ERCC1 and XPF, as well as the TFIIH DNA helicase XPB. Expression of these genes was significantly decreased in subjects whose drinking water arsenic levels were ³ 10 m g/L (p<0.05). In contrast, there was no correlation between arsenic exposure and expression of either XPG or XPA. These results indicate that intake of arsenic in drinking water at levels ³ 10 m g/L may decrease nucleotide excision repair gene expression, and support the hypothesis of a mechanistic role for nucleotide excision repair inhibition in arsenic-induced carcinogenesis (supported by NIEHS ES07373, NCI CA57494).



ARSENIC IN GROUND WATER IN EASTERN NEW ENGLAND: OCCURRENCE, CONTROLS AND IMPLICATIONS FOR HUMAN HEALTH.

Joseph D. Ayotte, Denise L. Montgomery, Sarah M. Flanagan, Keith W. Robinson, and Laura Hayes.

U.S. Geological Survey, 361 Commerce Way, Pembroke, NH 03275 .

In New England, low to moderate (1 to 50 micrograms per liter) concentrations of arsenic are known to occur in ground water, especially in the Eastern part of the region. There is increasing evidence that the source of the arsenic in New England is predominantly natural, originating from minerals within the rocks of the region. However, anthropogenic sources of arsenic (e.g. former pesticide use, treated lumber, manufacturing) may also contribute to ground-water contamination. Data from 88 wells in eastern New England showed that arsenic is more prevalent in water from private bedrock wells than in water from public wells in unconsolidated aquifers. Ground water from unconsolidated aquifers used for public supply accounts for about 33 percent of all drinking water use eastern New England. Water from these aquifers is less likely to be affected by contamination from arsenic (about 3 percent have water with arsenic greater than 10 micro-grams per liter) than privately supplied drinking water derived from bedrock aquifers. Water from private wells in bedrock aquifers accounts for about 14 percent of drinking water supplies in the area and is most likely to have arsenic concentrations at levels of concern to human health. Wells located in metasedimentary bedrock units that are described as variably calcareous are most likely to have elevated ground-water concentrations of arsenic compared to water from wells in other rock types. Nearly 30 percent of wells in these rock units had water with arsenic greater than 10 micrograms per liter. Arsenic concentrations were greatest where pH was greater than 7.5 and where dissolved oxygen concentrations were less than 1 milligram per liter. Arsenic was most commonly detected where the ground water was at or near saturation with respect to calcite. Fifty-eight of the 88 wells were sampled twice, 1 to 12 months apart. There was a strong correlation between concentrations for the two samples at each well (Spearman's rho = 0.86, 95% C.I. = 0.76-0.92) indicating that arsenic concentrations in bedrock wells did not change significantly during the interval. Although arsenic concentrations did not relate to any other metals, major ions, age of ground water, or hydraulic parameters, arsenic concentrations were related to major aquifer types (unconsolidated and bedrock) and to water supply types (public and private). From this information and associated water-use information, the total population potentially receiving drinking water with arsenic greater than 10 micrograms per liter was estimated at nearly 90,000 people on public supplies and about 114,000 people on private supplies. Whereas the estimate for the public-supply population will likely decrease because a lower standard will go into effect in 2006, the estimate for the population using unregulated private wells may not.



CELLULAR MECHANISMS FOR THE CARDIOVASCULAR EFFECTS OF ARSENIC.

Aaron Barchowsky¹, Nicole V. Soucy¹, Linda R. Klei ¹, Chandrashekhar D. Kamat² and Michael A. Ihnat² .

¹Dartmouth Medical School, Hanover, NH 03755, and ² The University of Oklahoma Medical School, Oklahoma City, OK 73104.

Chronic, low level exposure to arsenite increases the incidence of proliferative vascular diseases, arteriosclerosis, atherosclerosis, and ischemic heart disease. Arsenic-associated changes in blood vessels may also contribute to the vascular components of diabetes and tumor growth. Angiogenesis, the formation of new microvessels, is fundamental to most of these pathological changes and may underlie the vascular actions of arsenic. In support of this hypothesis, arsenite (0.033-1.0 m mol/L) significantly increased blood vessel density in an in vivo chicken chorioallantoic membrane (CAM) model. Above 1.0 m mol/L, arsenite inhibited blood vessel growth. Onset of angiogenesis correlated with increased expression of HIF-1a and plasminogen activator inhibitor-1 (PAI-1) in the CAMs. These increases were reproduced in cultured primary porcine aortic smooth muscle cells (SMC). In addition, arsenite stimulated SMC vascular endothelial cell growth factor (VEGF) expression in both a time- and dose-dependent manner. In contrast to endothelial cells, SMC were highly resistant to the toxic effects of arsenite and demonstrated proliferative responses at concentrations of up to 50 m mol/L. However, lower concentrations of arsenite, which that were compatible with endothelial cell proliferation, increased SMC mRNA levels for HIF-1 a , VEGF, and PAI-1 and caused sustained expression of HIF-l a and VEGF protein. In addition, microarray analysis in human airway epithelial cells, demonstrated that exposure to arsenite (5 or 50 microM) for 4 h increased HIF mRNA levels by 1.8-3 fold, suggesting that cells other than SMC can be the source of angiogenic factors. However, PAI-1 expression was not sustained at higher arsenite concentrations; indicating that PAI-1 may be a better predictor of the pro-angiogenic potential of arsenite. These data suggest that arsenite causes specific, dose-dependent effects on cell signaling that promote angiogenic and possibly anti-angiogenic responses that contribute to pathologic vascular changes (Supported by Superfund Basic Research Program Grant ES07373).



ARSENIC SOURCES TO GROUND WATER AND SIMULATION OF GEOCHEMCIAL EXPERIMENTS ON AQUIFER CORES AT A LANDFILL, SACO, MAINE: IMPLICATIONS FOR NATURAL REMEDIATION.

John A. Colman¹, Kenneth G. Stollenwerk², and Forest Lyford¹.

U.S. Geological Survey, ¹Northborough, MA 01532-1528, and ²Denver, CO 80225-0046.

Concentrations of arsenic (As) in leachate-plume ground water at the municipal landfill, Saco, Maine, varied with the character of geologic materials that underlie separate landfill piles. Comparisons were made between leachate plumes that pass through (1) glaciomarine fine-sand deposits not in contact with bedrock, and (2) sand and gravel upper- and till lower-layer deposits, over a fine-grained homfels bedrock. A low average As concentration (0.033 mg/L) and ratio of As-to-iron (As:Fe = 0.031 percent, by weight) were measured in plume water associated with the fine-sand deposits. A high average concentration (0.33 mg/L) and ratio (0.64 percent) were associated with the sand and gravel deposits. These differences corresponded with differences in average As concentrations and As:Fe values measured between the solid materials associated with the plumes, using both whole-rock digestions and hydroxylamine-hydrochloride leaches of aquifer solid materials. Speciation analysis of As in the leachate plumes indicated that high concentrations were primarily inorganic As (III), associated with anaerobic conditions and high concentrations of organic carbon and dissolved Fe and manganese. A core from the uncontaminated portion of the sand and gravel aquifer was eluted with uncontaminated ground water to which sucrose had been added to simulate the effect of reducing conditions in ground water from organic carbon. After reducing conditions were established in this core. As and Fe concentrations increased in the core effluent (after 50 pore volumes, average As = 0.2 mg/L, As:Fe = 0.5 percent). The results are compatible with a hypothesized source of As in the Fe-hydroxide minerals coating the aquifer solid materials. The coatings could be dissolved under reducing conditions in anaerobic leachate plumes releasing As and Fe to the ground water. An additional bedrock source is required to explain the highest As concentrations and As:Fe values, measured in the ground-water samples from deeper parts of the sand and gravel aquifer, and from bedrock wells. Cores from contaminated portions of the aquifer were eluted with uncontaminated ground water to simulate natural remediation conditions. Results indicated that substantial amounts of organic carbon have accumulated on the aquifer solids, causing continued biological oxygen demand. In laboratory leaching experiments of the most contaminated core, this pool of organic carbon caused complete consumption of the influent oxygen for 230 pore volumes. A geochemical model was developed to simulate the concentration changes of selected constituents in the natural remediation experiments. Concentrations of dissolved oxygen. As, and Fe in effluent from one core were used to calibrate the model. The model was then used successfully to simulate constituent concentrations in leachate from two other cores. The modeling indicates that reductive dissolution and sorption were the processes controlling As and Fe concentrations in the experiment effluents. Precipitation of As solid phases was not important. The data show that elimination of the source of landfill leachate and flushing with uncontaminated ground water may not return some constituents, including As, to pre landfill concentrations for decades (Funded by the U.S. Environmental Protection Agency, Region I, and the U.S. Geological Survey, National Research Program).



ARSENIC SOURCES AND PATHWAYS IN THE OVERBURDEN OF CENTRAL MASSACHUSETTS.

Rudolph Hon¹, Kevin Doherty², Thomas Davidson ¹, William C. Brandon³, Carol L. Stein⁴, and David F. McTigue⁴.

¹Department of Geology & Geophysics, Boston College, Chestnut Hill, MA 02467, ²Knoll Environmental, Inc., 69 Wexford Street, Needham, MA 02494, ³Office of Site Remediation and Restoration, USEPA Region I: New England Region, 1 Congress St, Boston, MA 02114, ⁴Gannett Fleming, Inc, 15 Willard Road, New Ipswich, NH 03071 .

Elevated levels of arsenic in unconsolidated layers within a zone that traverses N-S across Central Massachusetts had been at times noted, although without a specific reference to the source(s) of arsenic. Suspected sources included past applications of lead arsenate in orchards as a control for coddling moth, industrial applications in metal and leather processing facilities, and/or from natural sources. An accumulated set of data in the archives of state environmental agencies provides a confirmation of the widespread reports of arsenic levels that are well above the regulatory "background" levels (17 ppm) in overburden. We report data on arsenic that (1) were compiled for selected sites listed with the Massachusetts Bureau of Hazardous Waste within this region; and (2) data analyzed by this study on samples of overburden obtained from drilled profiles at randomly selected sites in Central Massachusetts. The compiled data include sites within a corridor along the NNE-SSW trending tract that passes through the geographic center of the state. Both data sets have similar arsenic frequency distribution curves (histograms) with two frequency subsets: 20 to 50 ppm and 50 to 800 ppm. Comparison with distribution curves for lead shows no correlation between lead and arsenic suggesting that lead arsenate is not the likely source for samples with elevated arsenic. Microprobe analysis of sulfides from bedrock cores confirms a presence of two different sulfide phases: pyrites (FeS₂) and cobaltites (CoAsS) in the underlying formations. Pyrites contain negligible amounts of arsenic, however, arsenic levels in cobaltites range from 30 to 50 % of As by weight. Elevated arsenic values in the overburden of Central Massachusetts are best explained by natural origin where the bulk of the overburden layer is derived from the local bedrock formations. An occasional presence of cobaltite clusters within the overburden is the likely explanation for the observed arsenic "hot" spots. Arsenic pathways have been recently documented in a water-supply aquifer located within the same geological zone. During installation of the monitoring wells, soil and groundwater samples were collected along vertical profiles between the top of the overburden aquifer and bedrock at three locations. Reducing conditions (ORP -50 to -200 mV) were encountered in the upper ~45 ft of the aquifer. Throughout this interval, total arsenic and iron yielded a strong correlation ranging up to a maximum of 189 m g/L and 21,900 m g/L, respectively. Below this redox boundary at ~45 ft bgs, dissolved arsenic and iron levels dropped below detection limits, and ORP increased correspondingly (0 to 100 mV). Soil analyses showed significant correlations between solid-phase iron and arsenic, aluminum, cobalt, copper, manganese, nickel, and zinc. These data and results obtained using PHREEQC to model arsenic adsorption are consistent with the reductive dissolution of iron oxides in the upper part of the aquifer and release of sorted arsenic.



ARSENIC EXPOSURES AND REPRODUCTIVE EFFECTS.

Claudia Hopenhavn.

Department of Preventive Medicine and Environmental Health University of Kentucky.

Inorganic arsenic is known to be associated with adverse human health outcomes such as skin, lung, and bladder cancers, vascular diseases, hypertension and diabetes. In the last two decades, numerous studies have been conducted to investigate the effects of arsenic on human health in several countries, but most of the research has focused on high exposure communities and on the dermatologic, vascular, and carcinogenic effects of arsenic. Conversely, limited attention has been directed towards understanding the potential reproductive health effects of arsenic exposure in human populations, or towards examining moderate or low exposure levels. Experimental studies support a role for arsenic as a developmental toxicant, and although limited, the findings from some human studies suggest that inorganic arsenic may be associated with several reproductive outcomes, including increased rates of spontaneous abortion, low birth weight, congenital malformations, pre-eclampsia and mortality. We have been conducting several studies in Chile, in an area with historically high exposures to arsenic (>500 m g/L) several decades ago, and which now has levels around 40 m g/L in the drinking water. We will present results of studies at both exposure levels in relation to reproductive and developmental outcomes, from past and current exposures. In particular, we will discuss infant mortality rates at high exposure levels, and decreases in birthweight at moderate or lower exposure levels. We will also discuss the potential for further analyses with data we have collected and for future work based on some of our results and experience.



AN INTERDISCIPLINARY STUDY OF ARSENIC EXPOSURE AND CANCER RISK IN NEW HAMPSHIRE.

Margaret R. Karagas.

Department of Community and Family Medicine, and Center for Environmental Health Sciences at Dart-mouth, Dartmouth Medical School, Hanover, NH 03755-3835.

Drinking water exposure to arsenic is clearly linked to cancer occurrence in highly exposed populations. However, the effects of environmental levels, typical of the USA are less clear. In particular, the health con-sequences of drinking water with 1 to 50 m g/L of arsenic remain controversial. Drinking water supplies throughout New Hampshire contain detectable levels of arsenic, and in four areas, several water systems exceed 50 m g/L. One of these "clusters" is located near a superfund site. We have detected arsenic primarily in private bedrock wells. Thus far, 35% of private wells in New Hampshire contain > 1 m g/L of arsenic and more than 10% contain >10 m g/L. Our ongoing epidemiologic investigation focuses on estimating the dose-response relationship between arsenic exposure and the risk of skin and bladder cancer in the New Hampshire population. Further, we are investigating whether specific subgroups of the population are greater risk of arsenic-induced malignancies due to co-carcinogen exposure, nutritional factors or genetic susceptibility. As part of our investigation, we are testing the reliability of various biological markers of arsenic exposure and are examining the molecular and genetic changes in lymphocytes and tumor tissues of individuals exposed to low levels of arsenic. New Hampshire has a unique population-based surveillance system for non-melanoma skin cancers, and a rapid reporting cancer registry from which to identify incident bladder cancers. We measure arsenic concentrations both in toenail clippings and in household tap water samples, and urinary arsenic on a subset. An especially challenging aspect of the study involves estimating lifetime drinking water exposure to arsenic by identifying and testing subjects' previous residences with private water systems. A large sample size is needed to detect effects at low levels of exposure and to identify possible interactions. Therefore, our study will include approximately 800 cases of squamous cell carcinoma, 1,100 cases of basal cell carcinoma, 850 cases of bladder cancer and over 1,200 controls. A summary of recent results will be presented.



ARSENIC: MOVING TOWARD A REGULATION.

Ira W. Leighton

U.S. Environmental Protection Agency, Region 1, 1 Congress Street, Boston MA 02114.

The U.S. Environmental Protection Agency (EPA) has spent much of 2001 reviewing the arsenic standard so that communities that need to reduce arsenic in drinking water can proceed with confidence that the new standard is based on sound science and accurate cost and benefit analyses. The Safe Drinking Water Act, as amended in 1996, required EPA to review current drinking-water standards for arsenic, propose a maximum contaminant level for arsenic by January 1, 2000, and issue a final regulation by January, 2001. EPA published a new standard for arsenic in drinking water on January 22, 2001 that would require public water sup-plies to reduce arsenic to 10 ppb by 2006. EPA withdrew the standard in March 2001 for review. Many small communities will be affected by the drinking water standard for arsenic, making it especially important to ensure that the Safe Drinking Water Act provision allowing balancing of cost, is based on accurate information. On May 22, 2001 EPA extended the previous delay of the rule's effective date to February 22, 2002 but did not change the compliance date (2006) for systems. On October 31, 2001, EPA affirmed the appropriateness of a maximum contaminant level (MCL) of 10 parts per billion for arsenic in drinking water. EPA believes setting the arsenic level at 10 parts per billion will provide additional protection to at least 13 million Americans.



ARSENIC IN GROUND WATER WELLS IN MAINE.

Marc C. Loiselle¹, Robert G. Marvinney¹, and Andrew E. Smith².

¹Maine Geological Survey, 22 State House Station, Augusta, ME 04333-0022, ²Environmental Toxicology Program, 11 State House Station, Augusta, ME 04333-00 11.

In the summer of 1993, residents of the towns of Buxton and Hollis, Maine, became concerned about the persistence of elevated arsenic concentrations (> 0.05 mg/L, the present EPA maximum contaminant level (MCL)) in the drinking water supply for a local school, although the elevated arsenic concentrations had been discovered several years earlier. This concern led to a town-wide survey of arsenic concentrations in over 1200 domestic water supplies. The survey found that over 13-percent of the tested samples had arsenic concentrations in excess of the MCL. Analysis of data on arsenic concentration in groundwater from several sets of random and pseudo-random bedrock wells indicates that 1- to 3-percent of wells in Maine have arsenic concentrations above the present maximum contaminant level (MCL) of 0.050 mg/1 and 12- to 13-percent above the proposed standard of 0.010 mg/1. These are comparable to levels found in New Hampshire (Peters and others, 1999) and New Brunswick (Brinsmead, personal communication, 2000), but significantly lower than values in the sample of private wells in the Buxton-Hollis area. Water samples from bedrock wells are much more likely to have elevated arsenic concentrations than dug wells or springs. The distribution of wells with elevated arsenic concentrations is not random. Statewide, the random and pseudo-random wells show a much higher occurrence of elevated concentrations in zones of biotite grade or higher metamorphism or adjacent to igneous intrusions. Several zones of elevated arsenic concentrations can be observed along an axis from northern York County (Buxton and Hollis) through central Kennebec County and in eastern coastal Maine. The arsenic concentration of groundwater is the most likely the result of both natural processes and human activities. A study of historical uses of arsenic in Maine (D’Angelo and others, 1996) showed widespread use of arsenic pesticides between 1920 and 1950, ending in the late 1960s. Conservative estimates indicate that 5 lbs/acre of As were applied to each year to apple orchards and blueberry fields, and 20 lbs/acre applied each year to potato fields. Locally high concentrations of arsenic in groundwater on the scale of kilometers (for example, Northport and Surry) appear to be associated with igneous activity within a belt of sulfidic pelites. An ongoing study of ambient ground water quality in bed-rock wells may provide additional information on the source and transport of arsenic in the bedrock flow system. D'Angelo and others, 1996, Historical uses and fate of arsenic in Maine, Final report to the Water Research Institute, University of Maine. Marvinney and others, 1994, NGWA Focus Conference on Eastern Regional Groundwater Issues, Burlington, VT, p. 701-714. Peters and others, 1999, Environmental Science and Technology, 33:1328-1333.



REMEDIATION STRATEGIES FOR PUBLIC AND PRIVATE WATER SUPPLIES—OVERVIEW.

Bernard Lucey, P.E.

NH Department of Environmental Services, PO Box 95, Concord, NH. 03302-0095.

Selection of an arsenic treatment system is a multifaceted process. Non treatment based alternatives include such options as interconnection with another water system free of arsenic, development of an additional well(s) and abandonment of the contaminated supply, and blending of supplies to dilute the arsenic to an acceptable level. Principle considerations are political practicality, loss of local control of the utility, and availability of a regional water supply alternative. Arsenic removal options must first consider system size (and thus economies of scale) and the level of operator expertise. The greatest impact of the arsenic rule will generally be on very small water systems. At a small-scale facility, process sophistication and instrumentation are normally minimal. Treatment processes must be accommodated within the existing floor plan or a costly expansion of the control building will be needed. Cost effective arsenic treatment methods vary with system size. For very small systems two methods look favorable: adsorptive media and anion exchange. These methods are easily automated, have a small footprint and require little operator knowledge. For much larger water systems, conventional coagulation with membrane filtration are favored. Also the element iron provides substantial arsenic removal when precipitated. As important as the above criteria are, the disposal of the arsenic waste by-product is a key element in selecting a treatment process. Most treatment processes produce a liquid waste, which is categorized as hazardous. Adsorptive media, on the other hand, hold the arsenic contaminant permanently attached on the adsorptive surface as documented by the Toxic Characteristics Leachate Procedure and thus the media can be disposed of at an ordinary land-fill without special considerations.



NEW TECHNOLOGIES FOR ARSENIC CONTAMINATED DRINKING WATER REMEDIATION.

Susan Murcott, Jessica Hurd, Tommy Ngai, and Barika Poole.

Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139.

The research described in this presentation involves the authors’ performance evaluation of four new and alternative remediation technologies, a discussion of several additional innovative technologies in terms of key criteria, and the lead author's work synthesizing the considerable literature on arsenic remediation technologies into an open-source database. The promise of and obstacles to the application of new and innovative arsenic remediation technologies will also be discussed. The four new and alternative technologies investigated were 1) iron filings, 2) iron oxide coated sand, 3) & 4) two different systems using modified activated alumina metal oxides. These systems were field tested at a total of eleven different sites in Massachusetts, New Hampshire and Nepal with source water total arsenic concentrations ranging from 100-1,000 m g/1. The iron filings media and the iron oxide coated sand gave treated water total arsenic concentrations respectively of < 5 m g/1-11 m g/1 and < 10-100 m g/l respectively. The two modified activated alumina metal oxide systems produced treatment water consistently < 5 m g/1 or < 10 m g/L. The costs of all these systems make them affordable, comparable to or less than conventional systems. This presentation describes several additional new technologies: granular ferric hydroxide, a ligand-based ceramic technology, biological filtration, new resin developments and nanofiltration and applies the criteria of performance, cost and environmental effects to come up with a preliminary evaluation of the viability of all these alternatives. The last part of the presentation "tours" two Web sites: http://web.mit.edu/murcott/www/arsenic and http://www.thinkcycle.org. The first Web site offers a database of over 50 specific arsenic remediation options. The thinkcycle Web site supports an open-source community engaged in design and engineering challenges and, under the arsenic treatment technology challenge, provides a virtual space in which to freely collect and exchange information on arsenic treatment so that this important public health issue can be swiftly and effectively addressed.



ARSENIC PROCESSES: EXAMPLES FROM NEW HAMPSHIRE

Stephen C. Peters¹, Joel D. Blum¹, Bjorn Klaue ¹, and Margaret R. Karagas².

¹Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109, ²Department of Community and Family Medicine, Dartmouth Medical School, Lebanon, NH 03756.

The geographic distribution of elevated arsenic concentrations in a fractured silicate bedrock aquifer in central New Hampshire correlates with the presence of pegmatites which border late Devonian granites and intrude metasedimentary rocks. Arsenic concentrations in the pegmatites average 9.6 mg/kg, which is much higher than the associated granites (0.24 mg/kg) and metasedimentary rocks (0.8 mg/kg). Arsenic is concentrated in these pegmatites by partial melting of calcareous metapelites and subsequent recrystallization as granites with low arsenic concentrations and pegmatites with high arsenic concentrations. Arsenic was observed to behave similar to Boron, both of which were concentrated into these late stage rock units. Arsenopyrite (FeAsS) with an oxidation reaction rim of scorodite (FeAsO₄.2H₂0) was observed in aquifer materials. Elevated arsenic concentrations observed in other New England locations occur in politic metasediments intruded by plutons (e.g. Marvinney, 1994). We propose that pegmatite formation from partial melting of pelitic metasediments may be a primary mechanism that concentrates arsenic in crystalline aquifer materials, which can then cause localized arsenic contamination of groundwaters. An alternate mechanism that can occur in the same location as pegmatite formation is hydrothermal circulation and vein formation. Movement of fluid along the temperature gradient from the intruding pluton into the surrounding rock units can mobilize arsenic from the intrusion zone, which would then precipitate in hydrothermal veins in the surrounding politic rocks. Both pegmatite formation and hydrothermal circulation may play an important role on a statewide basis, with the region around the Concord area being a case where pegmatite formation is the dominant mechanism. Groundwater arsenic concentrations in the region near the Concord Granite ranged from 0.001 m g/L to 400 m g/L with a median value (16 m g/L), more than thirty times higher than the median for groundwaters from all of NH (0.49 m g/L). High chloride concentrations (>1 mmol/L), resulting from road salt contamination of recharge waters, suggests that groundwaters are most likely very young (<50 years). All waters with highly elevated arsenic concentrations (>50 m g/L) have very low iron (<1 mg/L) and high pH (>7). Samples with low arsenic (<25 m g/L) had various concentrations of iron but occurred at lower pH values (<7). Sulfate is observed in excess of iron in the groundwaters and probably indicates the loss of iron as an oxyhydroxide precipitate, which then affects arsenic mobility via adsorption/ desorption reactions. At pH > 7, iron oxyhydroxides form rapidly and have a neutral or negative net surface charge that does not readily adsorb arsenic. At pH<7, iron oxyhydroxide formation is slow and depends on dissolved oxygen availability, however the resultant iron oxyhydroxides have a positive net surface charge, and readily adsorb arsenic. These data illustrate that reactions occurring after the initial dissolution of arsenic are as important as the spatial distribution of reactive arsenic source materials. This research high-lights the importance of characterizing not only the initial sources of arsenic, but also the geochemical processes occurring in the groundwater system.



ARSENIC GEOCHEMISTRY IN A BEDROCK AQUIFER, NORTHPORT, MAINE

Andrew Reeve¹, Michael Horesh, Robert Marvinney², and Robert Ayuso³.

¹Department of Geological Sciences, University of Maine, Orono, ME 04401, ²Maine Geological Survey, State House Station 22, Augusta, ME 04333, ³MS 954, National Center, U.S. Geological Survey, Reston, VA 12201.

Forty-four domestic water wells were sampled in Northport, Maine and surrounding areas and analyzed for metals, acid anions, alkalinity, pH and specific conductance. Arsenic concentrations ranged from less than 1 ppb to 1940 ppb, with concentrations as high as 5500 ppb previously reported in this area. Water wells providing water with the highest concentrations of arsenic are clustered around the Kellys Cove area. Wells in this cluster were re-sampled and arsenic was speciated in the field using ion exchange resin. The pe calculated from the As(3)/As(5) redox couple ranged from -1.5 to about 4, with most samples clustering near a pe of 0. Iron redox couple data, measured using a portable spectrophotometer, parallel the As data and have pe values about 2 pe units higher than those calculated from the As data. Two bedrock cores were collected within the high-As cluster for petrographic analysis. Arsenic-bearing sulfide minerals (cobaltite, arsenopyrite) were observed in thin sections of the bedrock core (Penobscot Formation), suggesting a natural source for the arsenic in the ground water. The close association of the As-bearing sulfides with epidote and tour-maline suggest a hydrothermal event may coincide with the production of these sulfide minerals. Ground-water chemistry data was evaluated using piper plots and principle component analysis. An arcuate pattern is present in the cation field on the piper plots, with water chemistry ranging primarily from calcium bicar-bonate to sodium bicarbonate type waters, with some samples enriched in chloride and sulfide. Arsenic-rich samples primarily plot in the middle and near the peak of the arcuate trend in the cation data. Bicarbonate and sulfide are the dominant anions in the As-rich samples and plot in a linear pattern in the piper plot anion field, trending from bicarbonate to sulfate-rich waters. The second PCA loadings (20.3% of variance) positively correlate Al, As, Fe, Mn, Pb and Si. Iron, Mn, and Al all occur in oxide-hydroxide coatings commonly produced during weathering, with stable mineral phases such as ferrihydrite and brucite precipitating during weathering. These mineral precipitates contain surface complexation sites that can sorb anions. The correlation of As and Si with these other metals suggests that the precipitation (or dissolution) of these coatings removes (or releases) As and Si oxyanions from the ground water within the bedrock aquifer underlying Northport. Our data suggests that sulfides are the ultimate source for As within the Northport As cluster. Sorption of As onto oxyhydroxide coatings may be an important control on As concentrations.



SPATIAL ASSOCIATIONS BETWEEN ARSENIC IN GROUNDWATER, SEDIMENTS, BEDROCK, AND AGRICULTURAL LANDUSE IN NEW ENGLAND.

Gilpin R. Robinson, Jr.¹, and Joseph D. Ayotte².

U.S. Geological Survey, ¹National Center, Reston, VA 20912 and ²361 Commerce Way, Pembroke, NH 03275.

Analytical data for arsenic concentrations in public-supply bedrock-groundwater wells, stream-sediment samples, and unmineralized bedrock samples from Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont provide a means to (1) map the occurrence and distribution of arsenic in the surface environment throughout New England and (2) measure the spatial associations between these data and other information on geologic units and land-use features in the region. Interpolation gridding of the bed-rock groundwater, stream sediment, and rock chemistry databases provides spatial coverage over 73, 71, and 31 percent of the New England area, respectively. Significant sources of arsenic to groundwaters and sediments include both natural weathering of rocks and anthropogenic applications of arsenical pesticides that were commonly used on apple, blueberry, and potato crops during the early twentieth century in New England. Agricultural census and landuse/land cover data were used to develop a spatial dataset portraying the intensity of cultivation of apple, blueberry, and potato crops in New England. Ten percent of the public-supply bedrock wells in our New England database exceed the new EPA arsenic standard of 10 m g/L. Twenty percent of the wells exceed an arsenic value of 5 m g/L; the distribution area of these high arsenic wells in the interpolation grid occupies 27% of the New England region where data are available. Equivalent statistics for stream sediments and rocks indicate that 20% of their arsenic values exceed 7.25 and 6.5 ppm and occupy 19 and 27 percent of the area of data coverage, respectively. Calcpelite, felsic volcanic, and sulfidic schist rock groups have higher background arsenic values than all other rock groups in New England. Kappa statistics (Bonham-Carter, 1994, chapter 8), measuring spatial agreement at the 80th percentile level, indicate that the distribution of arsenic in bedrock groundwater wells has a strong positive correlation with (1) the Coastal Maine and the Central Maine-New Hampshire geologic provinces, (2) calcpelite bedrock in the Central Maine-New Hampshire geologic province, (3) volcanic and sulfidic schist bedrock in the Coastal Maine geologic province, and (4) stream sediment and rock geochemistry. Bedrock wells have a weak positive correlation with past agricultural landuse. Stream sediments, which integrate both natural and anthropogenic sources, have a strong positive correlation with groundwater chemistry, geo-logic provinces, and rock chemistry and a weak positive correlation with past agricultural landuse. Although spatial correlation is not sufficient to demonstrate cause-and-effect, the spatial statistics favor rock-based arsenic as the dominant source of arsenic in stream sediments and groundwaters. The distribution of bedrock geology features at the geologic province and lithology group level closely match the areas of elevated arsenic in both groundwater and stream sediments. Bonham-Carter, G.R, 1994, Geographic Information Systems for Geoscientists: Modeling with GIS: Elsevier Science, Tarrytown, N.Y., 398 p.



PREDICTING ARSENIC IN WATER FROM AQUIFERS IN EASTERN NEW ENGLAND.

Sarah Ryker.

National Water-Quality Assessment, U.S. Geological Survey, PO Box 25046, MS-415, Denver CO 80225.

Several recent studies have identified eastern New England as a region with both high usage of ground water for drinking water, and widespread arsenic in ground water. Arsenic concentrations are especially high in ground water from eastern New England's bedrock aquifers. These aquifers are commonly used for private domestic wells; a small but increasing number of public supply wells tap bedrock aquifers. Precise assessment of population exposure is difficult as arsenic measurements are not available for every drinking-water well; information on private wells is particularly sparse. To fill these data gaps, the U.S. Geological Survey is currently developing a predictive model estimating likely concentrations of arsenic for wells in which arsenic has not been measured. Output from the arsenic concentration model will be used as input to exposure models for the six New England states, as part of the National Cancer Institute's New England Bladder Cancer Study. Concentrations of arsenic in ground water can be highly variable. Accounting for this variability in a predictive model requires information at a variety of scales. Many of the arsenic datasets and explanatory factors used in the model have been derived from regional studies covering large areas with relatively low data density. Several of these regional studies provide broad classifications of lithology and mineralogy that describe natural sources of arsenic in aquifer materials near the wells in question. These regional studies also provide water-quality parameters such as pH, dissolved oxygen, and sulfate that help to identify regional trends in ground-water geochemistry, which in turn help to predict mobilization of arsenic from the aquifer materials into ground water. Based on existing regional data, some initial predictions can be made of areas unlikely to have detectable arsenic in ground water vs. areas with the potential for arsenic in ground water. However, additional explanatory factors and more arsenic data are required to more finely distinguish areas likely to have moderate arsenic concentrations vs. high arsenic concentrations in ground water. Some additional explanatory factors may be derived through regional-scale exploration of the geochemical evolution of New England ground water. For example, the addition of geographic boundaries of historical salt-water contacts has refined the picture of coastal New England's ground-water geochemistry, with a commensurate refinement in model predictions. Local-scale explanatory factors are important in agricultural areas and near Superfund sites and mining operations, where human activities may exert field-scale influences over arsenic concentration. Areas under human influence are also likely to exhibit greater temporal variability in arsenic concentration than do areas where arsenic concentration is controlled primarily by natural factors; in areas where these practices influence ground water used as drinking water, quantifying temporal variability will help to refine the model. Above all. additional data on arsenic concentrations in bedrock wells will reduce the uncertainty and improve the spatial resolution of the model predictions. Some existing arsenic measurements from public supply wells in bedrock aquifers have been incorporated into the model. Additional sampling of private wells is planned during the New England Bladder Cancer Study; these measurements will be added to the model over the next several years.



"EVERYTHING BUT THE KITCHEN SINK": EXPOSURE TO ARSENIC FROM BATHING AND OTHER INDIRECT WATER PATHWAYS.

Andrew E. Smith.

Environmental Toxicology Program, Bureau of Health, Department of Human Services, State of Maine, Augusta, ME 04333.

An estimated 10 percent of Maine households have arsenic levels in well water above the recommended guideline of 10 m g/L. The most commonly selected technology to reduce exposure to arsenic from household well water is point-of-use treatment. These treatment systems are typically installed at the kitchen sink, leaving water elsewhere in the house untreated. This approach is predicated on the assumption that bathing and other indirect exposures are negligible. The purpose of this talk is to review the available studies that can be used to place bounds on the magnitude of exposures from bathing and other indirect pathways. In-vivo studies with monkeys and rats, and in-vitro studies with human and mouse skin have demonstrated that arsenic can be dermally absorbed. Dutkiewicz (1973) using an in-vivo rat tail exposure model and Rahman et. al. (1994) using an in-vitro mouse skin diffusional-cell model reported estimates of dermal absorption of arsenic from water that can be used to estimate dermal permeability coefficients ranging from 0.0015 to 0.0021 cm/hr. These values are similar to those reported for other inorganics. Use of these permeability coefficients with a USEPA (2001) model for estimating dermal uptake of inorganics from water predicts exposures ranging from 1 to 7 m g As for a child from a half-hour bath in water with 100 to 1000 ppb, respectively. Harrington et. al. (1978) reported data on urine arsenic levels from a survey of populations exposed to high and low concentrations of arsenic in well water. Average urinary arsenic levels were nearly identical for a group who had high arsenic water (mean of ca. 500 ppb) and used bottled water, as compared to a group who drank well water containing on average 11 ppb. Though limited, these data from studies of dermal absorption and community surveys indicate that indirect water exposures are small relative to direct ingestion from drinking and beverage preparation. These studies do not however, demonstrate that these secondary exposures are always negligible. Nor do these studies specifically evaluate children as a population that may have increased exposure due to play-related behavior during bathing. For these reasons, the Bureau of Health, in conjunction with the USCDC, has undertaken a study of to measure the extent of arsenic intake among adults and children who live in homes with high-arsenic well water, but use low-arsenic alternatives for drinking. This study is ongoing and in the middle of its data collection phase.



THE METABOLIC BASIS OF ARSENIC TOXICITY.

David J. Thomas.

Experimental Toxicology Division, National Health and Environmental Effects Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC.

Methylation of inorganic arsenic yields species that are more reactive and toxic than the parent compound; hence, the biomethylation of arsenic is an activation process. To understand the methylation process, we have purified a novel S-adenosyl-L-methionine: arsenic(III) methyltransferase from liver cytosol of adult male Fischer 344 rats that catalyzes transfer of a methyl group from S-adenosyl-L-methionine to trivalent arsenicals producing methylated and dimethylated arsenicals. The mRNA for this protein predicts a 369 amino acid-residue protein (molecular mass 41056 D) that contains common methyltransferase sequence motifs. Based on similarities in the sequence of the rat protein and other predicted protein sequences, this enzyme is designated as rat Cyt19. Rat Cyt19 mRNA is expressed in many rat tissues and human Cyt19 mRNA is expressed in HepG2 cells, a human hepatoma cell line that methylates arsenic. However, Cyt19 mRNA is not found in UROTsa cells, a human urothelial cell line that does not methylate arsenic. Recombinant rat Cyt19 is fully active as an arsenic methyltransferase. Human Cytl9 is a 375 amino acid-residue protein that is quite similar in sequence to rat Cyt19 and recombinant human Cyt19 is an arsenic methyltransferase. The catalytic functions of both rat and human Cyt19 have an absolute requirement for a dithiolcontaining molecule (This abstract does not necessarily reflect EPA policy).



NATIONAL TO LOCAL SCALE CYCLING OF ARSENIC IN GROUND WATER.

Alan H. Welch¹ and Joseph D. Ayotte².

U.S. Geological Survey, ¹333 W. Nye Lane, Carson City, NV 89706, ²361 Commerce Way, Pembroke, NH 03275.

Widespread high arsenic concentrations in potable ground water are most commonly caused by release from phyllosilicate, iron oxides, and sulfide minerals. Similarities between the geologic and geochemical characteristics of New England and other high-arsenic regions suggest that the latter two sources are important in the cycling of arsenic. A strong association between arsenic and weathered biotite in arsenic-rich ground water of Bangladesh has recently been demonstrated. Although weathering of biotite in a warm tropical delta is likely very different than weathering in the New England climate, the association between arsenic and biotite is worth consideration. Oxidation of pyrite and other less common sulfide minerals, such as arsenopyrite and cobaltite, can release arsenic to ground water. Pyrite commonly contains arsenic in at least trace amounts, with arsenic concentrations exceeding five percent in some cases. Molecular oxygen is quantitatively the most important oxidant in ground-water systems, although nitrate from agricultural activities also can oxidize sulfide minerals. Sulfide mineral oxidation is commonly limited by the amount of molecular oxygen contained in the water during recharge. Sulfate concentrations in most New England ground water are low (generally < 30 mg/L), suggesting that sulfide mineral oxidation is not much greater than could be attributed to oxygen in equilibrium with the atmosphere. In examples from other parts of the United States, exposing sulfide minerals to the atmosphere through lowering of ground-water tables can greatly increase oxidation. The resulting low pH ground water containing sulfate concentrations ranging greatly in excess of a few hundred mg/L, is not typical of ground water of New England. Arsenic can be released to ground water by desorption from, and dissolution of, HFO (hydrous ferric oxide) and other iron oxides. Desorption from iron oxide is an important process affecting arsenic concentrations in alkaline, oxic ground water because iron oxide commonly contains arsenic as an impurity. Desorption of arsenic can be promoted by an increase in either pH or the concentration of a competing ion, such as phosphorus. Sodium exchange for calcium can increase calcite dissolution (because of the lowered aqueous calcium concentration), thereby producing ground water with high pH and arsenic such as in the central Oklahoma aquifer. This scenario may be responsible for some of the high arsenic ground water of coastal New England where aquifer materials were affected by seawater either from their original depositional environment or from sea-level rise associated with Pleistocene deglaciation. As(III) is less readily sorbed onto HFO than As(V) within the pH range of most ground water. Because As(III) is present in moderate to high concentrations in some ground water from bedrock of New England, the lower adsorption may be an important factor affecting arsenic mobility. Dissolution of arsenic-bearing HFO, and other iron oxides, is an important source of arsenic in some ground water. High arsenic ground water in New England generally does not contain high iron concentrations, suggesting that oxide dissolution is not a major factor releasing arsenic; however, desorption may be a more important mechanism. While HFO is likely present in these aquifer materials, geochemical conditions generally are not favorable for dissolution of HFO and other iron oxides.



ARSENIC GEOCHEMICAL BEHAVIOR DURING GROUND WATER-SURFACE WATER INTERACTIONS AT A CONTAMINATED SITE.

Richard Wilkin¹, Robert Ford¹, Frank Beck¹ , Patrick Clark¹, Cynthia Paul², Joseph LeMay² , and Robert Puls¹.

¹U.S. Environmental Protection Agency, Office of Research and Development, National Risk Management Research Laboratory, Ada, OK 74820, ²U.S. Environmental Protection Agency, Region 1, Boston, MA 02114.

Arsenic mobility in groundwater at hazardous waste sites is often tied to redox reactions related to the geomicrobiological cycling of iron and sulfur. Important processes include adsorption or co-precipitation reactions of arsenate, arsenite, or thioarsenite species with poorly crystalline iron (oxy)hydroxides, iron monosulfides, and pyrite. Research results will be presented that address arsenic mobilization and cycling mechanisms at a Superfund site in eastern Massachusetts. The site is located in the headwaters of the Aberjona Watershed. In order to support assessments of the risk posed by off-site migration of arsenic to an adjacent downgradient wetland system, information is needed about the geochemical processes that control arsenic transport and fate on site. In addition, there is keen interest on the part of site stakeholders about the mechanisms and capacity of natural attenuation within the wetland area to prevent off-site migration of arsenic to the Aberjona River. In particular, data are needed to 1) assess the long-term assimilative capacity within the unconsolidated aquifer and the downgradient wetland, and 2) assess the potential for future mobilization of arsenic that is presently partitioned to soil/sediment solids. The study area encompasses a confined fluvial aquifer that locally discharges into a pond and wetland system. One of the challenges to assessing the impact of the discharge of arsenic contaminated ground water to surface water bodies is differentiating between the arsenic flux associated with ground-water discharge versus the arsenic flux due to dissolution of arsenic-bearing sediment components. An integrated sediment and ground water-surface water sampling strategy will be presented and monitoring results used to illustrate the method for capturing short-term temporal and spatial responses to storm events as a means to identify the sources of distinct arsenic fluxes. This presentation will also include discussion of a dynamic model of arsenic geochemical behavior that is closely tied to oxidation-reduction processes and the geobiochemical cycles of iron, sulfur, and carbon (This is an abstract of a proposed presentation and does not necessarily reflect EPA policy).