Modeling the Probability of Arsenic in Groundwater in New England as a Tool for Exposure Assessment

Abstract

We developed a process-based model to predict the probability of arsenic exceeding 5 microg/L in drinking water wells in New England bedrock aquifers. The model is being used for exposure assessment in an epidemiologic study of bladder cancer. One important study hypothesis that may explain increased bladder cancer risk is elevated concentrations of inorganic arsenic in drinking water. In eastern New England, 20-30% of private wells exceed the arsenic drinking water standard of 10 micrograms per liter. Our predictive model significantly improves the understanding of factors associated with arsenic contamination in New England. Specific rock types, high arsenic concentrations in stream sediments, geochemical factors related to areas of Pleistocene marine inundation and proximity to intrusive granitic plutons, and hydrologic and landscape variables relating to groundwater residence time increase the probability of arsenic occurrence in groundwater. Previous studies suggest that arsenic in bedrock groundwater may be partly from past arsenical pesticide use. Variables representing historic agricultural inputs do not improve the model, indicating that this source does not significantly contribute to current arsenic concentrations. Due to the complexity of the fractured bedrock aquifers in the region, well depth and related variables also are not significant predictors.

An Integrated Approach to Identify Water Resources for Human Consumption in an Area Affected by High Natural Arsenic Content

“…As expected, the coexistence of rocks with slow kinetics of arsenic release and short residence time of groundwater generates relatively low arsenic concentrations, as reported in other regions of the world (e.g., [63]). The influence of residence time on arsenic content of sampled groundwater is clear from the comparison between the distribution in the basal aquifer and in the perched aquifers or dome-impounded groundwater (Figure 4): in the entire hydrogeological system, the lowest arsenic contents were found in the sources, mainly springs located at high elevation, where the short flow path from recharge to discharge area implies a reduced water-rock interaction (Figure 3).…”

Validity of spatial models of arsenic concentrations in private well water

“…These models resulted in a 75% agreement, NPV=0.85, PPV=0.53, sensitivity =0.62−0.63, and specificity=0.80. In comparison, an arsenic regression model built for the New England area resulted in lower sensitivity (0.37) and higher specificity (0.93) using 5 μg/L as the cut-off value (Ayotte et al, 2006). NPV and PPV were not reported in the New England study, but were estimated from the data provided as NPV=0.83, and PPV=0.60, similar to what we report in this paper using a higher cut-off value (10 μg/L).…”

“…Alternative approaches are available, such as land-use regression (Ayotte et al, 2006) and Classification and Regression Trees (CART) (Schroder, 2006), in which spatial variables are classified into distinct categories and used to predict a dependent variable (e.g., arsenic) in a-spatial analyses. These approaches fail to explicitly consider the spatial pattern or proximity of the dependent variable, but rather take advantage of regionally available factors, such as geology, land use/cover, and hydrology in making predictions (Ayotte et al, 2006). Factors that vary greatly from well to well, however, such as well depth, are often difficult to estimate at unsampled locations, and therefore challenging to include in these predictive models.…”

“…These factors include sediment mineralogy, well depth, microbial oxidation or reduction of arsenic, competing elemental species for sorption sites, groundwater recharge, groundwater flow path, and presence of fractures in bedrock formations (Ayotte et al, 2006;Ford et al, 2006;Harvey et al, 2006;Smedley and Kinniburgh, 2002;Van Geen et al, 2003;van Geen et al, 2006). These factors are often highly variable between wells and therefore have been difficult to incorporate into models for predicting arsenic concentrations in well water (Ayotte et al, 2006;Van Geen et al, 2006).…”

Predicting the risk of arsenic contaminated groundwater in Shanxi Province, Northern China

“…We used SAGA-GIS (SAGA Development Team, 2008) to calculate the following DEM derived parameters: (1) Topographic Wetness Index (TWI), (2) Catchment slope, (3) Topographic index, (4) Modified Catchment Area (MCA), (5) Slope and (6) Elevation. These variables can correlate with hydrologic processes of the aquifer (Ayotte et al, 2006). Studies in Cambodia and Bangladesh revealed that there is a strong correlation between topographic environmental variables and the content of As in groundwater (Rodríguez-Lado et al, 2008;Shamsudduha et al, 2009).…”