An industrial activity model approach to differentiating exposure categories for oil and gas epidemiological studies

摘要

Oil and gas (O&G) production in the United States has increased in the last 15 years and these operations release pollutants to air. Health studies have relied on proximity to O&G wells as an exposure metric, typically using an inverse distance weighting (IDW) approach. Since O&G emissions are dependent on multiple factors, such as development phase, production volume, and scale, a dynamic model is needed to describe the variability in air pollution emissions over space and time. We used information on Colorado O&G activities, production volumes, and air pollution emission rates to create a spatiotemporal industrial activity model. The Spearman correlation coefficient between our modeled intensities and measured emissions was 0.56. We used our monthly model output to adjust the IDW well count metric for intensity and applied this metric to households in Greeley, Colorado, which is in the middle of the densely developed Denver-Julesburg basin. Our intensity-adjusted metric increased the dynamic range of "exposure" by 30-fold compared to the unadjusted IDW metric. Our approach distinguishes between high intensity events such as locations undergoing drilling, hydraulic fracturing/flowback, or large-scale production pads and lower intensity ones such as single well pads with small production volumes. This industrial activity model improves upon the proximity model approach and accounts for both inter- and intra-phase variability. As operators continue to move toward larger multi-well pads that reduce the overall geographic footprint of O&G development, individuals that live near those pads may be subject to much longer periods of high intensity and it will become more important to account not just for the density of wells, but the intensity of activity.