Methane emissions are underestimated by the U.S. Environmental Protection Agency (EPA)

Institute of Environmental Physics, University of Bremen

Published: January 27, 2014

A recent paper by Miller et al. (2013), published in the Proceedings of the National Academy of Sciences, states that the emission of methane in the United States could be 50% higher than previously estimated. The results are based on a top-down method (see below), using an extensive set of atmospheric concentration measurements of methane for the years 2007 and 2008. The main findings of the study are:

The U.S. Environmental Protection Agency (EPA) and the Emissions Database for Global Atmospheric Research (EDGAR) underestimate U.S. methane emissions by a factor of 1.5 and 1.7, respectively.
Methane emissions from ruminants and manure are up to twice the magnitude of existing inventories.
Methane emissions due to fossil fuel extraction and processing could be 4.9 ± 2.6 times larger than in the global methane inventory EDGAR.

CH₄ emissions* can be estimated from two fundamentally different methods; the “bottom-up” and the “top-down” method. The bottom up method uses various sources of information and extrapolates this information to get an emission estimate for a larger region. For example, information on emissions from natural gas wells is obtained by taking measurements from a couple of wells, and these emissions are then assumed to be representative for all wells in the country, giving a countrywide emission estimate. The top-down method uses the measurement of atmospheric concentrations to estimate the emissions from certain regions. This method provides data on cumulative emissions for relatively large regions, but does not give any information regarding the source of said emissions.

The study by Miller et al. (2013) uses the top-down method. Allowing for stated uncertainties it provides reliable estimates on regional emissions for the years 2007 and 2008. In principle the top-down method only allows for the determination of cumulative emissions from various sources over a certain region. However, the high level of regional emissions over the South Central United States area (Texas, Oklahoma, Kansas), a key region for fossil-fuel extraction and refining, would strongly suggest that these industries are responsible. This theory is also supported by measurement of propane levels. Propane is a tracer of fossil hydrocarbons and over Texas and Oklahoma good correlations with methane are observed.

Miller´s article was published just two months after a paper in PNAS by Allen et al. 2013 found that methane emissions during the pre-production and production stage of hydraulically fractured shale gas wells emitted less methane than previously. These results are not in conflict for the following two reasons: firstly Allen et al. only investigated a small part of the well-to-consumer lifecycle. Secondly there is a fundamental difference between their approaches.

Miller et al. determine the cumulative total emissions by a top-down approach. In contrast, Allen et al. use a bottom-up approach and measurements at just a few wells are used to calculate the total emissions. The bottom-up method used by Allen et al. only provides realistic emission estimates if the investigated wells are representative. If, however, a few facilities were to be responsible for a large fraction of the total emissions, this would only be captured by the top-down method (Miller et al.).

* Correctly “flux” should be used instead of “emission”. Flux includes uptake. Since uptake is not relevant in the context of natural gas production and the term “emission” is widely known, “emission” will be used in the following.