Reconciling divergent estimates of oil and gas methane emissions

Published estimates of methane emissions from atmospheric data (top-down approaches) exceed those from source-based inventories (bottom-up approaches), leading to conflicting claims about the climate implications of fuel switching from coal or petroleum to natural gas. Based on data from a coordinated campaign in the Barnett Shale oil and gas-producing region of Texas, we find that top-down and bottom-up estimates of both total and fossil methane emissions agree within statistical confidence intervals (relative differences are 10% for fossil methane and 0.1% for total methane). We reduced uncertainty in top-down estimates by using repeated mass balance measurements, as well as ethane as a fingerprint for source attribution. Similarly, our bottom-up estimate incorporates a more complete count of facilities than past inventories, which omitted a significant number of major sources, and more effectively accounts for the influence of large emission sources using a statistical estimator that integrates observations from multiple ground-based measurement datasets. Two percent of oil and gas facilities in the Barnett accounts for half of methane emissions at any given time, and high-emitting facilities appear to be spatiotemporally variable. Measured oil and gas methane emissions are 90% larger than estimates based on the US Environmental Protection Agency's Greenhouse Gas Inventory and correspond to 1.5% of natural gas production. This rate of methane loss increases the 20-y climate impacts of natural gas consumed in the region by roughly 50%.