The study by Arosio et al. uses multiple satellite products of tropospheric column O3, derived from limb and nadir sounders, to investigate long-term trends in tropospheric ozone. Overall, this is a nice study and provides useful updates on tropospheric ozone trends, especially as part of TOAR-2. My main question is on Section 5 and Appendix A where the authors use a proxy of any OMI averaging kernel (AK) to investigate the impact of vertical sensitivity on comparisons to other datasets (i.e. HEGIFTOM) and long-term trends. I also have several minor comments listed below. Once these have been addressed, the manuscript is suitable for publication in AMT.

Major Comments:

The authors attempt to investigate the impact of satellite averaging kernels on tropospheric column ozone trends by applying an approximation of an OMI AK to the ozonesonde data. So, firstly, the function in Equation 3 is based on what? Just the approximate shape of an OMI AK or a peer-reviewed study? Depending on the satellite product used (i.e. UV-Vis vs. IR or DOAS vs. optimal estimation), the shape of the AK profile can change substantially. So, would it not be worth trying to simulate the impact of multiple types of AKs? Also, to support your choice of Eqn 3, could you plot some actual OMI ozone AKs?

Minor Comments:

Line 11: State which time-period the Southeast Asia trend is.

Line 17: Should the x in NOx be subscript?

Line 43-44: You state that ozone trends in Europe and North America has stabilised. This has been supported by several recent TOAR-2 studies (e.g. Pope et al., (2023 & 2024)), so might be worth mentioning them to support this statement.

Line 50-52: Gaudel et al. (2018) identified large scale discrepancies between satellite product tropospheric ozone trends and you state that more work is needed to try and reconcile these. However, some TOAR-2 studies have attempted this, would be good to site those (e.g. Gaudel et al., 2024 and Pope et al., 2024).

Figure 2: Could you add the global average trop O3 column average for each panel? While you can see the differences by eye, having a metric (e.g. average +/- standard deviation) above each map would add a useful overview for the reader.

Equation 1: It is not clear what tj is. Can this be defined?

Line 136: What does m represent? It is not overly clear.

Line 136-138: What are the offsets based on? Is this relative to the ozonesondes or the satellite ensemble average? Same for “after debiasing” in the Figure 3 caption.

Line 144: “in spring time”…is this the boreal or austral spring?

Line 148: To make this clearer, could you add an example. E.g. (e.g. “where m indicates the month (e.g. January) of the year and tm all months m (e.g. all Januarys) in the time series”).

Figure 6e: What is causing the large standard deviation at approx. 20-30N?

Line 213: What do you mean by “after the harmonization of their time series by the HEGIFTOM”?

Line 214: What data source is the thermal tropopause level based on?

Equation 3: Add a reference, if appropriate, for this choice of function to represent the AK.

Line 243: Excess space between “last” and “10 years”.

Lines 244-246: I did not follow this text. This point probably needs more explanation.

Lines 251-253: Would be useful to add the time-periods the authors derived trends for.

Line 284-285: “The positive trend from SCIA-OMPS in the Amazon is likely related to artefacts in the datasets at the very beginning and end of the time periods”. Can you expand on this. Not clear what you mean be artefacts and why this would drive an unrealistic trend.

Line 291: You discuss “insignificant trends at the 95% confidence level”. I might be wrong, but I believe TOAR-2 are trying to move away from such definitions.

Line 293: If the time-series are influenced by “positive drift”, since you can quantify this, could it not be removed to leave the real trends?

Line 343: Why chose the OMI-LIMB product for this example?

References:

 Gaudel, A., Bourgeois, I., Li, M., Chang, K.-L., Ziemke, J., Sauvage, B., Stauffer, R. M., Thompson, A. M., Kollonige, D. E., Smith, N., Hubert, D., Keppens, A., Cuesta, J., Heue, K.-P., Veefkind, P., Aikin, K., Peischl, J., Thompson, C. R., Ryerson, T. B., Frost, G. J., McDonald, B. C., and Cooper, O. R.: Tropical tropospheric ozone distribution and trends from in situ and satellite data, Atmos. Chem. Phys., 24, 9975–10000, https://doi.org/10.5194/acp-24-9975-2024, 2024.

Pope, R. J., Kerridge, B. J., Siddans, R., Latter, B. G., Chipperfield, M. P., Feng, W., Pimlott, M. A., Dhomse, S. S., Retscher, C., and Rigby, R.: Investigation of spatial and temporal variability in lower tropospheric ozone from RAL Space UV–Vis satellite products, Atmos. Chem. Phys., 23, 14933–14947, https://doi.org/10.5194/acp-23-14933-2023, 2023.

Pope, R. J., O'Connor, F. M., Dalvi, M., Kerridge, B. J., Siddans, R., Latter, B. G., Barret, B., Le Flochmoen, E., Boynard, A., Chipperfield, M. P., Feng, W., Pimlott, M. A., Dhomse, S. S., Retscher, C., Wespes, C., and Rigby, R.: Investigation of the impact of satellite vertical sensitivity on long-term retrieved lower-tropospheric ozone trends, Atmos. Chem. Phys., 24, 9177–9195, https://doi.org/10.5194/acp-24-9177-2024, 2024.

General comments: This work provides an intercomparison of tropospheric ozone column datasets from combined nadir and limb satellite observations. Although this intercomparison is of interest to the community and deserves publication into the TOAR-II Special Issue, its presentation in terms of scientific clarity and focus could be substantially improved. Providing a more consistent story that is less broad would increase readability and significance. “The overall goal … to assess the consistency between the datasets and explore possible strategies to reconcile the differences between them” as phrased in the abstract does not seem to be fully (quantitatively) addressed, or the information is too scattered to be properly captured.

Specific comments:

• Line 22: “due to the overlap in signals” is too vague.
• Line 26: It would be appropriate to refer at least to ESA’s operational TROPOMI nadir ozone product (possibly next to scientific products).
• Line 56: Does “bias” here refer to bias from different TPH definitions or in general?
• Line 65: Does “its profile” refer to sensitivity or to stratospheric ozone?
• Lines 95-96: Was this drift correction done by the authors, or by the data providers? Please elaborate, possibly with reference(s).
• Line 98: “WMO thermal definition from NCEP reanalysis” requires explanation and references.
• Figure 1: Ordering the plots north to south would seem more logical, but is not mandatory. What happened to the 20-40 °N band plot?
• Line 120 and following: Which global mean has been used as a reference for de-biasing?
• Note sure whether the more qualitative Figure 2 adds a lot to 1 and 3?
• Line 164: Explain and explicitly indicate the “Quasi Biennial Oscillation signature”
• Lines 168-169: “plays an important role in the biases between them” – How do you know, and doesn’t this at least partially contradict conclusions made later, e.g. on line 178?
• Line 183: “As reference TPH, the ERA5 dataset was selected.” This must refer to a different dataset than the monthly gridded ozone profiles mentioned in the previous sentence.
• Line 186: “We subtract the mean column gaps…” Is this the global mean, or determined for each latitude-longitude bin?
• Figure 7: More of interest than the trends themselves seem to be the differences in the trends due to different TPH definitions? The latter could be given more attention.
• Sections 5 and 6 refer to figures and tables that are distributed over the main text, appendices, and supplement, which hampers a fluent appreciation of the research and results by the reader.
• Line 220: Possibly briefly explain the difference between drift and trend studies? It doesn’t really help calling the drift plots in Figure 8 a “trend in DU per decade of the differences” and “trend values (drifts)”
• Equation (3): Where does this come from? Is this a fully arbitrary choice, or is this based on common approaches in the literature?
• Line 231: Does “collocated” mean containing the station location?
• Line 241: “with negative trends until around 2014 and positive trends in the last 10 years” Despite talking about drifts here (also see above), this is really not clear from Figure 8, especially as this figure only contains linear fitting to the full time series.
• Line 252: “trends of +(1-4)%/decade” over which period?
• Figure 9: Do the regions defined here correspond to those agreed upon within the TOAR-II initiative?
• Line 268: Can you provide info and references on the multivariate regression model that is being used?
• Despite the anticipated focus on geographical regions in Section 6, Figure 11 extends to the global anyways. Possibly, limit this work to one of both only?
• It would be very insightful to add an estimate (zonally) of the fraction of the trend that could be explained from the TPH trends in Figure 7?
• Line 298: “reconcile the discrepancies … rather than highlight the inherent differences” sounds conflicting. This requires some explanation of the distinction between both terms…
• Line 319-320: “Our analysis shows that the homogenization is a crucial step before using the datasets for global trends studies.” This is not what seems to come out of this work. Do you mean a global bias correction, a TPH correction, calculation of anomalies, or the ozonesonde homogenization for comparison? Both what this sentence is referring to as well as how this can be concluded from the analyses is very unclear, and makes this conclusion inappropriate.

Technical corrections:

• Often inconsistencies occur between singular versus plural nouns and verbs. Please verify throughout the text.
• Line 9: “morphology” does not seem to be the right term here, or should be explained.
• Line 91: Rephrase “a DU dataset”
• Line 105: Add monthly temporal resolution?
• Lines 108-117: “overall” is used quite often, sometimes inappropriately, i.e., conflicting with other parts of the sentence(s), e.g. “zonal averages” in line 108.
• Caption of Figure 4: “De-seasonalized anomaly time series…” ?
• The color scale of Figure 6 (b) is hardly visible in print.