the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Dec 2025
Capability of current observing systems to monitor CH4 emissions from the regional to the global scales
Abstract. Top-down methane CH4 flux estimates involve large uncertainties stemming from three main sources: (1) the coverage of the observing system, (2) systematic and random errors in the observation data and the priors and (3) errors in the atmospheric transport model. Quantifying these uncertainties is challenging, and methodological studies suggest they can be substantial. While global-scale uncertainties in total CH4 emissions are relatively small (±5%), they increase significantly at regional scales exceeding ±20% for high latitudes. Differences in satellite and in situ measurement uncertainties, as well as variations in data density, further influence the precision of CH4 flux estimates. Sectoral disaggregation of uncertainties improves error attribution, leading to more reliable regional flux assessments and trend detection. Its benefits are amplified in high-emission regions due to larger absolute uncertainties and more complex source mixtures.
- Preprint
(5370 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-5923', Anonymous Referee #1, 25 Mar 2026
-
RC2: 'Comment on egusphere-2025-5923', Anonymous Referee #2, 29 Mar 2026
The manuscript of Montenegro et al. provides a synthetic Observing System Simulation Experiments (OSSE) to evaluate the ability of different observing systems (satellites, in-situ networks) to constrain sub-continental surface fluxes of methane and sectorial information of the fluxes. The study includes different satellite instruments (GOSAT; TROPOMI and IASI) and in-situ network and combination of surface networks + Satellites and it is concluded that best results are obtained by combination of in-situ and satellites. Considering that the included satellite instruments are in orbit for several years now, I would have considered a comparison study using real instead of synthetic data as more beneficial or alternatively to study the impact of new and upcoming missions such as CO2m or GOSAT-GW. Nevertheless, the manuscript presents a comprehensive and valuable study that is of relevance and of interest and I suggest publication after addressing my comments given as a pdf attachment. My comments concern, besides a few spelling mistakes and inaccuracies, that the assumptions and setup of the OSSE is tailored towards in-situ inversions and thus this study assign little value to satellite observations. Therefore, I consider it necessary to discuss the impact of setup and assumptions in more detail and indeed it would have been valuable to study the sensitivity to setup and assumption of the OSSE experiment.
-
RC3: 'Review comment on egusphere-2025-5923 Montenegro et al', Anonymous Referee #3, 01 Apr 2026
General comments
The work presents a comprehensive evaluation of multiple methane observing systems combining surface and satellite observations on their ability to reduce uncertainty of the inverse model estimated emissions. The paper is well designed and covers substantial details. Suggestion to accept after minor revision, adding mostly technical corrections as introduced below.
Detailed comments
L2 (abstract) “systematic and random errors in the observation data and the priors” this formulation can be extended to include prior flux uncertainty (prior flux covariances)
L145 The motivation to use only one MetOp satellite out of 3 was not clearly explained.
L465 Would be useful to provide analytical expression for flux uncertainty reduction in case of matrix-based inversion, as it operates with observation sensitivities to surface fluxes and observation errors, making a solid base for later comparing quantitatively various observing systems differing in the sensitivity to fluxes, observation numbers, and observation errors.
L536 There are some works indicating the added value of GOSAT observations over South America (eg Tunnicliffe et al., 2020, Takagi et al. 2021), and later in the manuscript, which kind of contradict the conclusions here.
Minor comments and technical corrections
L4 GOSAT-1 can be reduced to GOSAT without ambiguity. (as used in Line 19)
L5 What is special about uncertainties high latitudes? Tropical lands also have quite sizable emissions and uncertainties. Need to provide a context in order to avoid confusing a reader.
L79 Can say “tropics in dry season”.
L133 According to GOSAT websites (https://www.gosat.nies.go.jp/en/about.html; https://earth.jaxa.jp/en/research/projects/gosat/index.html ) GOSAT mission is a joint effort by JAXA, MOE and NIES
L225 There are some stations outside of 3 cited networks (NOAA, ICOS, and CSIRO) as seen on Table A1
L500 Figure 6. Korean => Korea
Figure B3. In plot: Relatif => Relative
References:
Takagi et al., 2021: Meteorological control of subtropical South American methane emissions estimated from GOSAT observations. SOLA, 17, 213−219, doi:10.2151/sola.2021-037
Tunnicliffe et al., 2020: Quantifying sources of Brazil's CH4 emissions between 2010 and 2018 from satellite data, Atmos. Chem. Phys., 20, 13041–13067, https://doi.org/10.5194/acp-20-13041-2020.
Citation: https://doi.org/10.5194/egusphere-2025-5923-RC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 298 | 225 | 27 | 550 | 47 | 43 |
- HTML: 298
- PDF: 225
- XML: 27
- Total: 550
- BibTeX: 47
- EndNote: 43
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper uses a suite of Observing System Simulation Experiments (OSSEs) to explore the capability of the global surface flask monitoring network and various satellite instruments to constrain surface fluxes of methane (CH4). This technique is well-established and has been used previously in similar studies, but the authors here do a good job of updating such work and expanding the analysis in a nice way. The use of small Monte Carlo runs for each inversion is a good way to probe the posterior uncertainties of the results, although the ensembles are of fairly limited size.
The diagnostics are well-chosen and defined in an intuitive way, on the whole. The study’s findings imply that observation density alone is not sufficient to improve inversion performance, particularly if it comes at a trade-off with observation uncertainty. It also reinforces that a good surface flask network is of vital importance for top-down emission monitoring and must be maintained and, preferably, expanded significantly in order that we can better understand fluxes of species like CH4 in the future.
Overall, the paper is well-written, structured properly and has useful and well-designed figures. In my opinion it will serve as a useful reference for those attempting to do top-down estimations of CH4 fluxes in the future and should be published after the following minor suggestions have been incorporated.
Comments:
Figure 1: Please consider using a colour-blind friendly colour table for this figure (and others in the Appendix).The rainbow colourbar is not reader-friendly for such figures for a number of reasons. See https://blogs.egu.eu/divisions/gd/2017/08/23/the-rainbow-colour-map/
Line 438: How was the ensemble size of 10 members arrived at? Are you satisfied that it could represent the full uncertainty spectrum accurately? Please justify here.
Line 435: Did you consider randomly perturbing the synthetic observations within their precision to represent satellite ‘noise’? Wouldn’t this more accurately represent the satellite system in the OSSE and reduce the likelihood that you’re overestimating the error reduction? What is the impact of this decision on your results – please discuss at some point.
Line 453: Please briefly explain why the monthly averaging is done in log space before returning to real-space values.
Lines 461 – 463, and Section 4.1: The author’s discussion of the M metric can be confusing in places, in my opinion. For example, lines 462-463 state “…whereas positive values indicate that the posterior fluxes fail to improve upon the prior in reproducing the truth (darker blue in Figure 6).” However, darker blue colours relate to negative values and red colours relate to positive M values.
Elsewhere in Section 4.1 and beyond, the M value is referred to as positive when it is in fact negative (see line 509-510 “ M values above 20%”), whereas on line 513, similar M values are referred to as “below -20%”. Please check these sections and be consistent with your description of M values.
Line 510-514: I’m afraid I don’t quite follow your point on regions (such as China) with low sensitivity and high flux densities ‘appearing’ improved when this is not really the case. Can you clarify here?
Line 551: You mention representativeness errors here – am I correct in thinking that you have used a flat 10 ppb value for these errors on all satellite data, based on a 2014 paper relating to SCIAMACHY? Are you confident that this remains a useful assumption and would you expect a more complex model for representativeness error to change your results for the different satellite instruments included here?
Line 580 – 583: I don’t follow this logic. Please clarify? Why does the addition of GOSAT data degrading performance relative to the surface-only inversion having few sites there? Surely this is true in both cases? Apologies for missing the point!
Line 607: Rewrite this statement slightly – it currently implies that both M and UR are 2x GOSAT-NIES and 4x GOSAT-Leic, which I don’t think is true.
Line 842: In this section you mention the possibility of other combinations of satellites. Why did you choose not to include TROPOMI + surface, TROPOMI + GOSAT or IASI + other data in this manuscript? This would surely have strengthened the results. Although I understand that you might be reluctant to complete new model runs at this stage, I think that discussion of the other available combinations is warranted, at least.
Conclusions section: To what extent are your results limited by the fact that you use a single atmospheric model? Would you expect these conclusions to be upheld if an ensemble of models were used?
Technical corrections:
Line 13: “OSSEs are conducted over 18 regions…” is misleading. Change to “OSSE results are analysed over…”, or similar?
Line 17: Should “…reducing uncertainties of more than 30%...” be “…reducing uncertainties by more than 30%...”?
Line 25 and line 49: mixing ratio -> mixing ratios
Line 66-67: remove “the” in “… the Japan’s GOSAT mission…”
Line 471: “URr,t = …”. The ‘t’ should be subscript.
Line 494: “Figures 8” -> “Figure 8”
Line 608: Fix “Result reflecting…”
Figure 10 caption: Fix “10% was applied 10%”.
Line 733: Insert space after “…CO2).”