This paper seeks to organize all measurements of methane clumped isotope measurements published to date. The introduction, explanations, and discussions are complete and it was a pleasure to read. 

I have only two larger suggestions for the authors. I leave it to the authors to decide if they want to follow these suggestions.

First, I would suggest providing plots of dD and d13C vs the 13CD and D2 measurements as well as dD vs. d13C – I would provide in the background prior databases for bulk so the data here can be compared to that. This is a way of showing whether the data set assembled here compares well to prior data measured or if there are gaps. Second, I think it is important to consider the clumped data in general in the context of prior bulk isotope measurements such that interpretations of clumping are considered in conjunction with standard measurements.

Second, in the database, I strongly suggest providing the additional 'metadata' such as, for surficial samples, the dD of waters, d13C of CO2 etc. For experiments, such would also be extremely helpful and, where known, the isotopic composition of the organic molecules provided (where relevant). For the thermogenic samples, I would suggest providing the gas compositions and isotopic composition of other molecules (where know). I know this will be annoying to do, but this is the kind of information that makes the database become extremely useful as, any study using this data, will likely need that as well. And so future authors will be stuck compiling this other information over and over again. I note, maybe this is provided, but I only saw a cell indicating what metadata exists.

****

A few minor comments:

Line 101-102: clumping is only independent of bulk composition for an equilibrated system. It is a strong function of bulk composition for many non-equilibrium processes (mixing, chemical kinetics, distillation, etc.).

Line 113: In terms of history — there are older attempts to do methane clumping for 12CD4 and claims of exceptional values. Eiler 2007 summarizes this. Ma et al (2008 https://doi.org/10.1016/j.gca.2008.08.014) discussed the idea of the measurement. Tsuji et al. (2012 https://doi.org/10.1016/j.saa.2012.08.028) also attempted this and developed a method, but I wasn’t applied to measurements of environmental samples as far as I recall.

Line 379-383. I understood that there is an emerging understanding that for thermogenic methane, it likely forms out of clumped equilibrium but, at high enough temperatures, rapidly reequilibrates so it reflects peak formation temperature prior to expulsion. This is discussed as far as I remember it the cited papers from Dong, Eldridge, and Xie et al. This is a nuance, but is different from methane representing formation temperatures and formation in equilibrium but rather represents rapid kinetics of H exchange post methane formation and then quenching of the reaction. I recommend checking those papers to verify what they said (or asking them as two are on this paper).

For figure 6, what are the catalytic equilibration samples that are +30 to+40‰? Are those labeled experiments where 13CD was added to a sample then removed during equilibration to verify the catalyst was working? If so, I might not include as they are spiked experiments. I would in general avoid including anything in which labels were added.

The Defratyka et al. manuscript describes the distribution of doubly substituted isotopologues in exospheric reservoirs and their relevance in constraining the sources of atmospheric methane. The paper is important in that describes how methane clumped isotopes can potentially help in the quantification of atmospheric sources which are of key importance to atmospheric scientists working to model the source and sink of key greenhouse gases, including methane.

Two challenges to this approach (as described in the introduction) have been the incorrect assignment of kinetic isotope effects in sink reactions, and also assumptions regarding the source signatures (Line 122). In order to correct this, the authors have compiled a database of 1500 previously published  clumped isotope results. Because the equilibrium distribution of isotopologues is temperature dependent, the sources can be inferred based upon expected source temperatures, although this also has some caveats (Line 97).

Of the field-sampled sources, thermogenically derived methane is the most common, as has been verified through traditional gas measurements of cold seeps in recent decades (e.g. stable carbon isotopes of methane, methane/ethane ratios, etc.). Similar approaches have been taken to determine biogenic end-member distribution in cold seeps. As Defratyka et al. concede in this manuscript (line 331), abiotic methane is quite poorly constrained. The sources and distribution of abiotic methane are adequately described in this paper (line 405). Nonetheless, the difficulties in ascertaining the abiotic component are not that straightforward, given that the methane concentrations are quite low in high-temperature gases such as those in submarine hydrothermal vents, geothermal wells, and fumoroles. In these cases, CO2 gas and, in some cases nitrogen gas, are major constituents while methane may only be a couple percent of the total gas content (as opposed to cold seeps where methane often comprises over 90% of the total gas content). As such, abiotic methane derived at high temperatures can easily be contaminated by other sources, such as thermogenic methane gas released during pyrolysis of organic matter in surrounding sediment and, as reviewer, it is of my personal opinion that methane ascribed as being of purely abiotic origin be viewed with some caution.

As described by the authors, the net sink for methane in wetlands and marine systems is either AOM or AeOM. The kinetic isotope effects associated with each process are unique and help distinguish between the two. On the other hand, tropospheric methane removal through OH reduction seems to have only a minimal effect on the remaining CH4.

This is the most complete compilation of doubly substituted methane isotopologues and the global map of  sample distribution covers an impressive global range. Nonetheless, the sample density is patchy (fig. 7) and there are vast areas around the globe that remain unsampled, including some areas of very high methane emission. Given that data is lacking, and sampling is skewed towards some areas more than others, it is possible that further refinements in global sample distribution will change the distribution of clumped isotope results (fig. 5).

All in all, this data description paper is well written and describes the potential for clumped isotope studies to be applied to methane source determination in future studies. As such, I recommend this paper for publication, and also look forward to ongoing work which provides a more complete data set in the future.