Modeling the Concentration Distribution of Benzene from Flowback using AERMOD

Oil and gas well operations are often located near populated areas in communities including schools and parks.  This raises concerns about the exposure of people in these areas to compounds emitted from activities associated with oil and gas production. Compounds emitted include a class of compounds known as BTEX (benzene, toluene, ethylbenzene and xylenes). Some of which are known carcinogens and can affect the central nervous system. Understanding the increases in the ambient concentration of these species is important for at-risk populations and the general public living in areas close to oil and gas operations.  In this study, AERMOD was used to model the dispersion of benzene during flowback from oil and gas wells in northern Colorado.  We focus on comparing concentrations at 500 feet, the current well operation setback distance, with 2000 ft, the proposed setback distance in the State of Colorado.  In our simulations (with no terrain), we used an average benzene emission rate during flowback of 0.094 g/s from a recent study by Colorado State University. There was a factor of about ten difference in the concentration of benzene at 500 ft and 2,000 ft. Meteorology from two different sites in the Front Range of Colorado from different land use areas to represent the varied conditions in the region was used.  The first site, Rawhide, was located in Larimer County, north of Fort Collins in a natural landscape, while the second site, St. Vrain, was in Weld County in an agricultural setting.  The highest concentrations over the course of a year were observed at 500 ft for the meteorology data from St. Vrain but were observed closer to the source at Rawhide.  Simulations were also run with multiple wells, which indicated that the meteorology data may be an important factor to consider as the results between sites varied greatly. For St. Vrain, there was an average factor of 6 difference in concentration between the single and four source cases and a factor of 4 difference using the Rawhide meteorology data.