Comparing spatial exposure surrogates with detections and concentrations of unconventional oil and gas-related chemicals in residential drinking water

Approximately 30 epidemiologic studies of unconventional oil and gas development (UOG) to date have identified associations with adverse health outcomes, such as birth outcomes, asthma, and cancer. These studies used spatial surrogates of exposure, which are feasible approaches for assessing aggregate exposures in large-scale studies. However, exposures and pathways captured by these surrogates (e.g., water contamination, air pollution) are poorly understood. We used logistic and linear regression to assess whether commonly-used spatial surrogates were associated with UOG-related chemicals in residential drinking water, an exposure pathway of public health concern. We compared detections and concentrations of 64 organic and inorganic chemicals from 94 private water wells in Bradford County, Pennsylvania with count of UOG wells within a buffer around the home, distance to nearest UOG well, and inverse distance weighted (IDW) and inverse distance-squared weighted well counts at buffer sizes of 2, 5, and 10km. We also compared chemical concentrations to health-based drinking water standards to understand human health risks from exposure. A total of 28 chemicals were detected in ≥20% of water samples, generally at low concentrations, though seven chemicals exceeded health-based drinking water standards at one or more homes. Of the organics, only benzene and toluene were associated with spatial surrogates. The odds of detecting benzene or toluene decreased by 55% and 50%, respectively, with each increasing kilometer between the residence and nearest UOG well (95%CI: 0.24-0.87 and 0.27-0.94); there were no clear relationships between other chemicals and metrics. Concentrations of iron, manganese, and sulfate were consistently inversely associated with IDW metrics, but the other 14 inorganics were poorly associated with metrics. Preliminary results indicate that spatial surrogates are weakly associated with detections and concentrations of UOG-related chemicals. More complex models incorporating groundwater flow and contaminant fate/transport may be needed to accurately capture drinking water exposure for health studies.