Reducing Uncertainty in Life Cycle Methane Emissions from Natural Gas using Atmospheric Inversions

Methane (CH₄₎ emissions associated with the production and use of natural gas (NG) are highly uncertain because of challenges to accurately measure fugitive CH₄ emissions from NG leaks and venting throughout a large and complex industry. Better understanding the CH₄ emissions from the NG life cycle is important for two reasons. First, the rising interest in NG use associated with the recent development of unconventional sources, such as shale gas, may cause a shift in the future energy system from coal towards more NG. Given its relatively high greenhouse gas potency, fugitive CH₄ emissions from the NG life cycle have the potential to outweigh lower CO₂ emissions compared to coal use in terms of their climate impacts over the next few decades. Second, worldwide NG related CH₄ emissions play a key role in understanding the global CH₄ budget. According to current atmospheric inversion studies, NG and oil production account for about 12% of global CH₄emissions. However, these results largely depend on prior emissions estimates whose uncertainties are poorly documented.The objective of this research is to analyze which ranges of global fugitive CH₄ emissions from the NG life cycle are reasonable given atmospheric observations as a constraint. We establish a prior global CH₄ inventory for NG, oil, and coal using emissions data from the life cycle assessment (LCA) literature. This inventory includes uncertainty estimates for different fuels, world regions, and time periods based on LCA literature, which existing inventories do not account for. Furthermore, global CH₄ inversion modeling will be used to test bottom-up hypotheses of high NG leakage and venting associated with the upper bound of the prior inventory. Given the use of detailed LCA emissions factors, we will test bottom-up scenarios regarding management and technology improvements over time. The emissions inventory will be established for the past decade, and inversion modeling will be carried out using NOAA’s CarbonTracker-CH₄ model over the same time period. Building prior emissions inventories using transparently documented LCA data allows us to directly link observation and model agreements to fugitive CH₄rate scenarios. Preliminary results suggest that on average, the upper bound of NG related CH₄ emissions factors in the LCA literature can be rejected with confidence using global inversion modeling. Considering a global average fugitive emissions rate of 6% translates into approximately 140 Tg CH₄/a in 2010 (80 Tg CH₄/a higher than current estimates), or 28% of the global CH₄ budget. Ongoing work will analyze in more detail the extent to which such high estimates disagree with observational data, e.g., the inter-hemispheric gradient given that 95% of NG is produced in the industrial north. The results also indicate that the LCA uncertainty estimates of NG systems combine epistemic uncertainty and spatial variability, which may significantly increase the perceived average life cycle CH₄ emissions of NG systems.