This manuscript presents a work of developing a deep-learning-based model for plume segmentation of XCO2 over cities or power plants in European countries. They evaluated the model for model generalization on new data from the same region and model extrapolation on unseen data from another region. The results indicate the proposed segmentation model outperforms the usual segmentation technique based on thresholding.

In general, the presentation of the paper is clear, and the potential of this technique is well-suggested. However, further explanation is needed on how this technique can be applied to estimate emissions from satellite imagery.

Detailed comments:

1)In the introduction section,

- The additional reference is needed for that NO2 can be a proxy to CO2 and with NO2, the plume detection capabilities are significantly improving. 

- Since CO2M is a satellite mission, the author is considering applying this technique; a more detailed explanation of CO2M is needed, such as the spatial resolution, channel information, etc.

2) In the 2.2 section, page 5.

- The data for Paris are selected for Jan., Mar., and Aug. Is there any specific reason to use these three month?

- How much has the results performance improved using data augmentation techniques?

The following paper introduced the data augmentation technique for weather applications considering major wind direction. Like this, have you considered the domain characteristics in data augmentation methods?

"Seo, Minseok, et al. "Domain Generalization Strategy to Train Classifiers Robust to Spatial-Temporal Shift." arXiv preprint arXiv:2212.02968 (2022)." 

 3) In the 3.4 section,

- The results showed when the concentration is low or signal-to-noise is small, the performance is significantly degraded. The author mentioned NO2 is helpful for that in the introduction section. Then, why is NO2 data not used as an additional input to solve this problem? 

- In the deep-learning approaches, the data split is important. Generally, the training and validation dataset are randomly split, while the test is separated from the training and validation. It would be best if you used separate datasets, not days in the middle of the same month used in the training dataset. And please indicate how many datasets are in each training, validation, and test dataset.



4) In the results,

- Most plume smoke shapes are long-tailed, and when the smoke does not spread and gathers in the middle, the segmentation results are not as good as those from long-tailed shapes. There has been a bias towards the plume shape. It seems necessary to analyze whether the result of having a higher wbce score was influenced by the shape of the plume.

 - How you get the emission amount in the Figure 13.

“Segmentation of XCO₂ images with deep learning: application to synthetic plumes from cities and power plants” by Dumont Le Brazidec et al. describes a new method for segmentation of CO₂ plumes in satellite imagery that could improve methods for quantifying CO₂ emissions from anthropogenic sources.  The authors develop and train a convolutional neural network (CNN) for XCO₂ plume segmentation from satellite observations that outperforms the more common thresholding approaches, when demonstrated on modelled plumes with random noise and non-uniform backgrounds.

The manuscript applies atmospheric models to address a relevant scientific question that has implications for the Copernicus CO2M mission and monitoring, verification and support (MVS) efforts to support climate policy. The CNN approach developed is novel and represents a substantial advance for the field. In general, the method is described well and the manuscript overall is well-written, well-structured and well-presented.

Specific Points

Line 1: It would be best if expanded text matched acronym: “CO₂ Monitoring Verification and Support” system and “CO2MVS” system.

Line 2: “Amount of CO₂” should be replaced with “Distribution of CO₂”

Line 35: It is my understanding that 2 of 3 satellites CO2M are scheduled for launch in 2026, with the 3^rd to follow by a year or more. (The authors should confirm whether this is correct and if not, revise with the most up to date information).

Line 113: including only January, March and August barely the minimum for representing “seasonal variability” with 3 of 4 seasons. A more thorough effort could have been made here.

Line 120: Some basic information about CO2M observing characteristics would be useful for the reader, at minimum, the nominal image pixel size (2x2 km²) and swath width warrant mentioning. Furthermore, it is unclear whether the 0.7 ppm Gaussian random noise is applied at the model resolution of 1.1x1.1 km² or the CO2M imagining pixel size, which was never mentioned. 0.7 ppm noise at 1.1x1.1 km² is equivalent to a lower noise level at 2x2 km².

For improved readability, capitalization of “WBCE”, “NWBCE” and “DDEQ” acronyms is strongly recommended throughout the entire manuscript.

It might be difficult to reproduce these results from only the brief description about how the plumes were modified to generate the training dataset.

Figure 13 seems to be the only mention of the magnitude of emission sources in the manuscript. For some perspective the authors should mention either the annual emissions for the sources in the study (Paris, Berlin and power plants). As a suggestion for additional perspective, the authors can cite the recent real world example of quantifying CO₂ emissions from Europe’s largest power plant using satellite observations (https://doi.org/10.3389/frsen.2022.1028240), consistent with the high end of the blue scale in Figure 13.