The Contribution of Groundwater Arsenic to Public Water Arsenic Concentrations in the United States, 2006-2011

Abstract

Background: Although many aquifers in the US contain elevated concentrations of naturally occurring arsenic and groundwater is a major water source for many community water systems (CWSs), the contribution of groundwater to CWS arsenic concentrations has not been systematically studied. We evaluated the contribution of domestic well arsenic concentrations to CWS arsenic concentrations over time and by region for conterminous US at the county-level from 2006-2011. We hypothesized domestic well arsenic would be more strongly associated with CWS arsenic concentrations in the Southwestern US and during 2006-2008 versus 2009-2011, reflecting the 2006 implementation of EPA’s 10μg/L MCL. Methods: County-level 90th percentile probability of domestic well water arsenic exceeding 10μg/L was estimated by US Geological Survey (USGS) using boosted regression tree models. County-level CWS arsenic was estimated using EPA’s 2006-2011 Six Year Review of Contaminant Occurrence Database. We used linear regression to assess associations between domestic well arsenic, a proxy for groundwater arsenic, and CWS arsenic, stratified by CWS monitoring period (2006-2008 vs. 2009-2011) and region for 2,231 counties across conterminous U.S. Results: From 2006-2011 nationwide, the relationship between the 90th percentile probability of domestic well arsenic exceeding 10μg/L and CWS arsenic decreased by 1.14 (95% CI: -1.54, -0.73)μg/L between 2006-2008 and 2009-2011. A declining relationship was observed in all regions, except the Central Midwest and Pacific Northwest. Discussion: The probability of high arsenic in domestic well water, a proxy for groundwater, is relevant contributor to CWS arsenic nationwide. This contribution was weaker in 2009-2011, possibly reflecting implementation of the 2006 MCL and uncertainty in modeled probability estimates. Declines in contribution of domestic well arsenic to CWS arsenic varied regionally, suggesting that interventions to implement the MCL were more successful in some regions (Southeast, Southwest) than others (Central Midwest, Pacific Northwest), or that aquifers used for domestic wells and CWSs differ regionally. Keywords: Water, water quality, biomarkers, arsenic, environmental epidemiology