Kai, Tyler, Randerson & Blake reply

Replying to Levin, I. et al. Nature 486, 10.1038/nature11175 (2012) Levin et al.1 report new methane data from Antarctica measured at the University of Heidelberg and combine these data with observations from two other networks to estimate long-term trends and the interhemispheric difference (IHD) in methane isotopes. These observations and the author’s analysis suggest that the δ13C IHD has been mostly uniform over the last three decades. This contrasts with the Niwot Ridge, Baring Head and Pacific Ocean time series that we analysed2 and suggests that both fossil fuel and microbial sources remained mostly constant during the late 1980s and 1990s, following a trajectory of emissions that is more similar to the control scenario in our analysis (Fig. 2 and Supplementary Fig. 2 of Kai et al.2). The estimate of δ13C inter-hemispheric difference is sensitive to different methods and selected observed sites. We note that the inter-hemispheric differences between Barrow and Cape Grim/Baring Head and between MLO and Cape Grim/Baring Head in Fig. 1 of Levin et al.1 were largest near the beginning of the record (circa 1990) and are broadly consistent with comparisons we made in the Supplementary Information of our paper (Supplementary Fig. 9)2. Important next steps, in our opinion, are (1) to use all available observations from both the Kai et al.2 and Levin et al.1 studies to assess trends in methane isotopes and their dependence on the inclusion or exclusion of individual time series, and (2) to quantify changes in microbial and fossil sources consistent with the expanded set of observations presented by Levin et al.1. Even with a levelling off of both fossil and microbial sources during 1990–2005, our forward biogeochemical modelling analysis of rice agriculture provides evidence that emissions from this sector of the budget declined as a consequence of changes in fertilizer application and water management.