The authors here present preliminary results from the MethaneSat satellite. The satellite is presented as bridging the gap between quantifying total regional fluxes and detection of some point sources, but the authors in this paper focus solely on the point source component of the technology. It is really interesting to see these results, and the satellite imagery displayed in this manuscript is quite compelling. I have several comments before being able to recommend for publication - in particular, more detail of their approach and contextualization of their results is needed. The authors also do not have a strong scientific question or hypothesis they are testing, rather they are mostly showing a demonstration and capability of a new technology. Given that focus, it's important that they do not draw too broad of conclusions based on these paper results alone. That said, I do look forward to what should ultimately be a deep set of scientific analyses that will come from the MethaneSat satellite record. 

1. Abstract - I am worried the authors are overstating a few things based on the evidence they've shown in the paper. In particular, Line 15. "Our results illustrate the potential of the MethaneSAT data archive for the discovery of new methane hotspot regions and super-emitters around the planet." This is not supported yet by the results in this paper. All the regions shown in this paper were previously known to be emitting methane super-emitters via other satellite detection. The authors need to reframe accordingly. 

2. How are the plumes actually quantified? Reading through the list of other cited papers (Chan Miller et al. 2024, Guanter et al., 2025) that use MethaneAir data, it appears that either a divergence integral method could have been applied, or an IME method. In the text the authors state "mass balance" but that could apply to variety of approaches. If it's an IME method (or really for either), how did the authors tune parameters to arrive at these emission rates? Through the comparison with IMEO-MARS data? Through simulation? This warrants deeper discussion given the presentation of emission rates in this paper.

3. Line 283. What makes these plumes unprecedented? That is an overstatement given that all the satellites listed in this paper (GHGSat, EnMAP, EMIT, etc) have seen plumes of the same magnitude or even at lower detection limits in Turkmenistan. 

4. Section 3.2.2. Can the authors provide more details about the emission rates distributions and the total point source emission total for all the Permian overpasses? In Figure A1 I only see the 22-May 2024 totals, but then this section is try to draw contrasts with other overpasses to say something about intermittency. Would be interesting to see how much the total point source rate changes across these observations.

5. In Figure A1, it looks like the total point source estimate is around 25 t/h, though still curious about those totals for other overpasses (comment #4). I crossed reference this against the MethaneAir paper the authors cited (Guanter et al., 2025) that also imaged the Permian and found ~35 t/h. Could you use that result or comparison (or maybe the distributions of MethaneAir vs MethaneSat) to say something more rigorous about MethaneSat's detection limit?

6. In the comparison with other satellite missions, the authors note that there can be grouping of multiple unique plumes into one plume complex, which complicates comparison. Then in section 3.2.2, the authors state that they see some plume emission rates that are exceptionally large relative to other studies - on Line 315 they state seeing 5 plumes greater than 5 tons/hr. Couldn't this really just be an issue of multiple plume aggregation? Given that repeated remote studies - the authors cite some (MethaneAir, Carbon Mapper, PRISMA/GF5) and miss some others (Insight M; Chen, Sherwin et al. 2021:https://doi.org/10.1021/acs.est.1c06458; Kunkel et al. 2023: https://doi.org/10.1021/acs.est.3c00229) - but none of those other studies see that frequency of exceptionally large emission sources, which calls into question whether the emission rates derived from MethaneSat truly represent individual sources, or whether the quantification is biased in some way. Can the authors further clarify what could be driving the exceptionally large emission rates? Why are you certain there isn't a quantification bias?

7. Can the authors provide the raw data from Figure A1 (and all plume detections + quantified emission rates) as part of this submission?

8. Authors state potential takeaway capacity limitations as driving the emissions they see in the Permian. Given the relatively small number of plume detections, I suggest the authors attribute the plumes to the most plausible source infrastructure types. That should be relatively easy and would be preferred to speculation. 

9. Figures 4b and 4d. Presumably Hassi Messaoud is going to be the optimal condition for plume detection (bright and homogeneous). As the authors point out, to the naked eye, the embedded synthetic methane plume in 4d, is marginally discernible at 1.6 m/s, and disappears at 3 m/s. Related, there is another enhancement in all figures around the index (x=420, y=90) that could potentially be a plume, or is at least as compelling as the LES 500 kg/h plume enhancement in Figure 4b. So this raises the question on how the plume detection process is taking place in the MethaneSat processing system. Can the authors comment on how the build confidence in a plume detection?

10. Figure 5. In the zoomed out view, both of these look like diffusive fields. I do appreciate seeing the whole image, but it's hard to reference the detections shown on the albedo map vs the actual concentration fields that are provided. Related - I think on the 14-June map, the "14-June-2024" label is covering one of the plumes

11. Line 325. Typo - "four" instead of "fout" 

12. The reported 20-30 ppb precision (~1-1.5% of background) is about a factor 2 higher than the single-sounding precision cited for TROPOMI, and is about on par with cited numbers from some point source imaging missions (Ramier et al, 2025: https://eartharxiv.org/repository/view/9307/; Duren et al. 2025:https://doi.org/10.5194/egusphere-2025-2275), while significantly better than precision numbers from other missions (PRISMA; Guanter et al., 2021). Given the emphasis on MethaneSat filling an area flux and point source imaging gap, and it is worthwhile for the authors to contextualize their precision against other missions. 

13. I understand that this paper's scope is not to map a realistic detection limit using detection probabilities. That said, this is really where the community is going, and minimum detection limit estimation is not really a super helpful metric for understanding the true detection capability of an instrument. I suggest the following that can at least provide more context about how detection limits vary in the areas MethaneSat observed: specifically, can you plot precision (and/or MDL via the MDL equation) either as a function of albedo, or for each of the scenes you summarize in Figure A1? It would be helpful to know how much the precision varies per geography and how that scales the MDL. 

The paper reports a global survey of methane point hot spots by EDF’s MethaneSAT’s high spectral resolution (0.25 nm), medium spatial sampling (110×400 m2 at nadir), and wide-area coverage (about 200 km2 at nadir) over its operational lifetime I year 3 months (March 24-June 25).  MethaneSAT’s ability to identify both point and diffuse large area super-emitting regions from O&G, agriculture and landfills is demonstrated with its L4 product (with citations for details). The paper partially mines MethaneSAT’s archive to identify and quantify 16 plumes, with emission rates ranging between roughly 800 and 7000 kg/h. WRF simulations are used to attribute plumes and infer emissions. Results are compared with available IMEO/TROPOMI data sets for rudimentary validation. Given that MethaneSAT was a high-quality NASA/ESA/JAXA-class satellite sensor funded by the non-profit and deployed at a much lower cost this is a critical paper that will help us develop the path forward on future low/medium cost global monitoring of natural gas leaks at a critical time. However, there are important gaps in its methodology and clarity that need to be addressed before I can recommend it for publication.

MeathneSAT’s plume detection capability is illustrated and compared with much coarser TROPOMI data using “adjustable arbitrary” parameters – notable “q” that measures the signal/noise at a pixel (# of standard deviations above noise) and np to measure the number of samples above noise threshold. While its ad hoc and qualitative nature are acknowledged no sensitivity studies to these parameters are presented and should be discussed. Their potential impacts on the matched filter plume reconstruction mentioned that use downscaled 30x30m WRF plume simulations that have uncertainties associated with winds and plume dispersion in GEOS-FP compared to the real world. It would also be valuable to MARK the point sources in Fig 1 (a) in addition to 1 (c) to illustrate the fidelity/challenges of plume separation/mixing (as is done later for less cluttered images).

Detection limits of 500 Kg/h under favorable winds are a key determination/claim that needs more careful caveats and/or evaluation. What are the favorable conditions? What is the limit under unfavorable conditions? What are the uncertainties? Also, standard deviation does not measure “systematic” biases and the sources of this should be mentioned and/or discussed.

There is a lot of discussion in the text on sensitivity/resolving power as a function of wind speed, direction and nonuniformity, surface albedo, cloud cover, multiple sources, matched filter low bias etc… that would be valuable to consolidate in a table for clarity.

The weakest part of the paper is proper uncertainty analysis and validation that demands much more discussion. For example, recent analysis using airborne-AVIRIS ng images of methane releases illustrates how do get reconstruct plume velocities from its shape and use this to show quantitative https://www.pnas.org/doi/10.1073/pnas.2507350122  – This should be cited and discussed. Are there similar methods that can be developed using WRF and/or MethaneSAT plume information? Can satellites be designed to take multiple snapshots of the plume that are closer in time to allow better velocity constraints for flux inversions?

Typos and question

Fig 8: Explain why the ‘source merged” – yellow L4 is much higher than IMEO

L28 tradeoff

L318 Delaware

L321 Also,

L325 four not fout, cannot

L386 widely used