Modeling Groundwater Vulnerability to Contamination from Unconventional Oil and Gas Development: Uncertainty Analysis Using Linear-based Methods

Activities related to unconventional oil and gas development (UD) can impact drinking water resources through various pathways such as surface spills and improperly constructed wellbores. Uncertainties about potential risks remain, especially among rural communities which simultaneously host UD and depend on shallow aquifers for domestic water supply. We develop a physically based groundwater vulnerability index to characterize these risks by integrating well capture probability and the geospatial characteristics of UD wells. The vulnerability assessment is predicated on a sufficient description of groundwater flow and solute transport through numerical modeling. We illustrate our approach in a field site in northeastern Pennsylvania, a region with a relatively high concentration of UD activity within the Marcellus Shale play. In this study, we explore the uncertainty in the delineated probabilistic well capture zones and the resulting vulnerability index through a suite of linear-based, first order, second moment techniques, also referred to as Bayes linear theory. Lastly, we explore the implications of these uncertainties on management decisions such as setback distances of UD activities from critical receptors.