Constraining Emissions of Methane in the Mountainous Uintah Basin with Ground-Based Observations and a Time-Reversed Lagrangian Transport Model

In the Uintah Basin of eastern Utah and western Colorado, extreme topography, complex mountain meteorology, and extensive oil and gas emissions combine to result in unusual air quality challenges. Our team at the University of Utah have been carrying out in situ observations of methane at three sites in the Uintah Basin since 2015: a baseline site (Fruitland), a site in the western portion of the Basin near oil wells (Roosevelt), and a site in the center of the Basin, near gas wells (Horsepool). In addition, a fourth site at Castlepeak in the central Basin near oil wells collected in situ data in 2016. To our knowledge, this study presents the first multi-year record of methane concentrations in the vicinity of major oil and natural gas extraction infrastructure in the United States. An analysis of the near-three-year long dataset in conjunction with meteorology provides insight into the influence of nearby methane emissions on concentrations observed at the sites. Overall, Horsepool and Castlepeak experience higher methane concentrations than Roosevelt, which is in part due to the higher density of oil and natural gas activity in those areas. Seasonal variations in meteorology impact methane concentrations observed at the sites. In the spring and fall, regional weather disturbances impact the normal diurnal cycle of methane more so than in the summer. In the winter, prolonged periods of atmospheric stability (cold-air pools) trap pollutants within the mountainous basin and methane concentrations gradually increase, while a diurnal cycle remains superimposed. We also carried out atmospheric modeling with the Stochastic Time-Inverted Lagrangian Transport (STILT) model driven with wind fields simulated by the Weather Research and Forecasting (WRF) model during a targeted period (19 April 2015 – 31 May 2015). At Horsepool and Castlepeak, the diurnal cycle of modeled methane concentrations was captured well by the WRF-STILT modeling framework when forced using recent National Oceanic and Atmospheric Administration aircraft-derived emission inventory estimates, but was underestimated using a bottom-up Environmental Protection Agency emission inventory.