Estimating acute impacts of unconventional oil and gas development on cause-specific hospitalization via satellite-based exposure assessment

Project summary Onshore expansion of Unconventional oil and gas development (UOGD) is considered an emerging environmental health issue due to its disproportional proliferation in rural U.S communities and the potential health impacts on 23 million nearby residents. However, two critical gaps in exposure assessment have

restricted previous studies from achieving rigorous and actionable evidence regarding the acute health effects of UOGD construction stages. First, current exposure assessments were too coarse to accurately capture the short-term variability in exposure. Due to a lack of detailed construction records, investigators had to impute

construction timelines for most wells based on questionable assumptions, which likely introduce misclassification. Second, none of prior large-scale observational studies has quantified exposure to specific UOGD-sourced pollutant. Investigators instead have developed various surrogates, which proxy residential

exposure with functions of auxiliary factors (e.g., distance to UOGD wells). These exposure surrogates do not account for the different transports and fates of UOGD-sourced pollutants in ambient environment, thus cannot indicate specific exposure pathway(s). Furthermore, causal inference, which provide a rigorous adjustment for

confounding, have not been widely used in estimating UOGD’s health effects. The objective of this proposal is to investigate the acute health effects of UOGD constructions based on improved satellite-based exposure assessment. In the K99 phase, I will take advantage of high-resolution images obtained by multiple recently

launched satellites, including Sentinel-1s, Sentinel-2s, and Dove satellites to automatically detect the daily construction stage of UOGD wells (Aim 1). Multiple state-of-art causal inferential methods will be used to estimate the acute effects of each UOGD construction stage on the risk of cause-specific hospitalization in

Medicare beneficiaries based on the new exposure assessment (Aim 2).In the R00 phase, I will develop and validate new exposure metrics based on the daily concentrations of air pollutants near UOGD wells, which are observed by Sentinel 5P (Aim3). Finally, I will explore the extent to which UOGD-related air pollutants mediate

any effect of UOGD construction stages (Aim4). The expected outcomes are 1) an open-source software to detect UOGD-related activities from remote sensing images; 2) exposure metrics for UOGD-related air pollutants; 3) acute causal health effects of UOGD construction. The exposure data will be made public

available to advance the field.