Integrating Oil and Gas Measurement Data to Estimate Spatially-Gridded Methane Emissions in the Barnett Shale

Abstract:

In October 2013, a dozen research teams measured methane emissions from oil and gas (O&G) and other sources in the Barnett Shale region of Texas at multiple scales ranging from bottom-up component measurements to top-down regional emission measurements. This work integrates ground- and aircraft-based measurements of site-level emissions from the campaign and a recent national study of gathering and processing facilities to construct a spatially resolved emission inventory for the Barnett Shale. Spatially referenced activity data including O&G site locations were obtained from multiple databases. O&G site emission factors were estimated with two-step Monte Carlo simulations that integrated emission rates from unbiased datasets with higher measurements obtained with targeted sampling. Emissions from other fossil and biogenic sources were estimated from reported emissions data or published emission factors. We constructed a 4 km x 4 km gridded emission inventory to estimate emissions by source category in the 25-county Barnett region. Total methane emissions in October 2013 were estimated to be 72.3 (+10.1/-8.9) Mg CH₄ h^-1 with 46.2 (+7.9/-6.2) from O&G sources. Fat-tail sites, which were defined as emission rates above the unbiased sampling distributions, accounted for 19% of O&G emissions but less than 2% of sites. In comparison to alternative estimates of O&G emissions based on the United States Environmental Protection Agency Greenhouse Gas Inventory, EPA Greenhouse Gas Reporting Program, and Emissions Database for Global Atmospheric Research, our custom inventory was higher by factors of 1.5, 2.7, and 4.3, respectively, similar to published ratios of top-down and bottom up estimates. Our custom inventory was higher than alternatives primarily due to more complete activity data and the inclusion of fat-tail site emissions. Gathering facilities, which accounted for 40% of our O&G emission estimate, had the largest difference from alternative inventories.