Machine Learning Models of Arsenic in Private Wells Throughout the Conterminous United States As a Tool for Exposure Assessment in Human Health Studies

Abstract

Arsenic from geologic sources is widespread in groundwater within the United States (U.S.). In several areas, groundwater arsenic concentrations exceed the U.S. Environmental Protection Agency maximum contaminant level of 10 μg per liter (μg/L). However, this standard applies only to public-supply drinking water and not to private-supply, which is not federally regulated and is rarely monitored. As a result, arsenic exposure from private wells is a potentially substantial, but largely hidden, public health concern. Machine learning models using boosted regression trees (BRT) and random forest classification (RFC) techniques were developed to estimate probabilities and concentration ranges of arsenic in private wells throughout the conterminous U.S. Three BRT models were fit separately to estimate the probability of private well arsenic concentrations exceeding 1, 5, or 10 μg/L whereas the RFC model estimates the most probable category (≤5, >5 to ≤10, or >10 μg/L). Overall, the models perform best at identifying areas with low concentrations of arsenic in private wells. The BRT 10 μg/L model estimates for testing data have an overall accuracy of 91.2%, sensitivity of 33.9%, and specificity of 98.2%. Influential variables identified across all models included average annual precipitation and soil geochemistry. Models were developed in collaboration with public health experts to support U.S.-based studies focused on health effects from arsenic exposure.

Drivers of Spatiotemporal Variability in Drinking Water Quality in the United States

“…22 Annual precipitation and groundwater recharge were found to be important drivers of elevated arsenic levels in the United States based on the application of machine learning approaches to private wells. 23 Another study showed that arsenic MCL (10 μg/L) violations were found mostly in the Southwest United States, primarily in groundwater systems (95%), serving small populations (mean ∼1,100 people). 24 Dominant drivers of nitrate MCL (10 mg/L) violations were found to be land attributes (e.g., cropland, irrigation), N surplus inputs, and surplus precipitation based on random forest modeling.…”

Mapped Predictions of Manganese and Arsenic in an Alluvial Aquifer Using Boosted Regression Trees

“…Studies in Louisiana and Arkansas found that As was mobilized via dissimilatory iron (Fe) reduction from Fe and Mn oxide coatings on sediment (Sharif et al 2008; Borrok et al 2018). At the continental scale, predictions of As focused on domestic drinking‐water wells and found low As in the MRVA (Ayotte et al 2017; Lombard et al 2021). At the global scale, As has been studied due to its significant health effects; low probabilities of high As were predicted in the MRVA (Amini et al 2008; Podgorski and Berg 2020).…”

A Statistical Approach for Identifying Private Wells Susceptible to Perfluoroalkyl Substances (PFAS) Contamination

“…Modeling analyses can help prioritize testing in regions that are most likely to contain detectable concentrations of PFAS. For example, spatial modeling approaches have been successfully used to predict inorganic contaminant concentrations (especially arsenic and nitrate) in well water at the local, regional and national scales. − A Bayesian network model was applied to predict the occurrence of a novel PFAS, GenX, in private wells around a fluorochemical manufacturing facility . These studies have identified potentially important predictors based on understanding of the sources and transport of chemical contaminants in groundwater.…”