We have constructed an atmospheric inversion framework based on TM5-4DVAR to jointly assimilate methane and δ13C methane measurements to estimate their emission sources. This article presents global emission estimates for the period 1999-2016. A newly constructed multi-agency database of CH4 and δ13C measurements has been assimilated. It is considered unlikely that traditional atmospheric inversions of methane alone estimate emissions consistent with δ13C atmospheric data. Therefore, δ13C data assimilation is necessary to derive emissions consistent with both measurements. Our approach attributes approximately 85% of the post-2007 atmospheric methane growth to microbial sources, with about half originating from the tropics between 23.5° N and 23.5° S. This contradicts the attribution of recent methane budget growth by the Global Carbon Project (GCP). It is noted that GCP attribution aligns with the top-down estimate obtained in this article conducted without δ13C data.

It has been found that globally and continentally, δ13C data can better distinguish microbial from fossil methane emissions than methane data alone, and on smaller scales, this capability is only limited by the current coverage of δ13C measurements. Finally, it is noted that the greatest uncertainty in using δ13C data to separate different methane source types stems from our limited knowledge of atmospheric chemistry, particularly the tropospheric chlorine distribution and isotopic discrimination of methane absorption (so-called sink).