In their paper Can Li and co-authors present the NOAA-20/OMPS SO2 product. The paper is well written. It has a good introduction with a well chosen set of references. The reason for the work, linked to the creation of a multi-sensor long-term coherent NASA/NOAA SO2 data record, is motivated extensively (maybe even more than actually needed). The new dataset is complementing and extending the series from OMI and SNPP/OMPS, and it is important that this dataset is documented in the peer-reviewed literature. The added value of N20 compared to the other platforms is presented in a nice way. It was good to see the emission estimate comparisons, which are very convincing. I am in favour of publishing this work after my comments have been addressed.

General comments:

Validation (e.g. with MAXDOAS, PANDORA, IR satellite data) is not discussed in the paper. Could you add one or two paragraphs on how the (previous OMI/SNPP) retrieval compare with independent observations?

It seems that the product does not contain averaging kernels, like most recent satellite data products do. This would be useful, in particular for the PBL contributions and for data assimilation applications. Is there a chance that kernels will be added in a coming upgrade? 

The retrieval is called "version 1", but there are still a few elements missing, in particular the cloud retrieval, AMF calculation and the SAA filtering. Is a new update (v1.1) foreseen for the near future to include these aspects? The final summary paragraph mentions further extensions using machine learning, but this sounds like a more longer term development path (v2.0).

Detailed comments:

l 125: "none-SO2 PCs". Should this be "non-SO2"?

l 141: "Process each orbit separately". If an orbit contains mainly ocean, is there still enough information stored in the PCA to descibe pixels over land? Isn't it better to train the PCA with one or multiple days of data? 

l 183: (new volcanic screening): What is the impact? are there now more, or less pixels flagged?

l 194: "five subsectors". Are these five latitude bands? Why is this important? Please add a few lines to explain the subsectors.

l 202: "has not yet been implemented with N20/OMPS." Is there a special (technical) reason why this has not yet been implemented? The benefits seem to be big, as shown in Fig. 4.

l 203: "a fixed AMF (0.36) is used to convert all SCDs to SO2 VCDs," I was confused here, because line 169 mentioned "and a priori SO2 profiles based on a climatology from multi-year global model simulations". Please clarify. I assume this is linked to the missing cloud product. Maybe move this remark up to the beginning of the paragraph.

l 248: "the the"

Fig.2 left seems to indicate jumps in the NH. Is this within the noise, or linked somehow to the subsectors? 

The biases seem to be linked to changes in the algorithm, and especially the threshold for SO2 containing pixels. In Figs 6 and 7 the product is again bias corrected to check consistency with SNPP. This leaves the question why the procedure was changed in N20 after all. Why do you bias-correct N20 and not SNPP? It would be useful to extend the discussion on this. Does the old procedure lead to problems for N20? If so, where do the problems occur and how big are these problems?

Fig.8: It is nice to see the N20/SNPP ratio plots. Indeed, this seems to indicate that the increase in resolution has a benefit over the lower SNR for emission estimates.

What I found surprising is the good match with TROPOMI. The retrieval is very different, the AMF is different and the SNR and resolution are also very different. It would be useful to discuss this in more detail, maybe even add an extra figure, and an extra paragraph. Apparently the local contrasts in SO2 columns and absolute column values are very similar compared to the NASA algorithm.

l 448: A weighted average based on inverse variances is suitable for unbiased datasets. In the average of the four platforms algorithm and instrument differences play a role. Why did the authors choose this approach?

Sec 3.5: It would be interesting to see the TROPOMI results as well for the two eruptions.

Inviting one representative from the TROPOMI retrieval team as co-author could be considered.

The code and data availability section is missing.

Review of the manuscript “Version 1 NOAA-20/OMPS Nadir Mapper Total Column SO2 Product: Continuation of NASA Long-term Global Data Record” by Can Li et al.

The manuscript entitled “Version 1 NOAA-20/OMPS Nadir Mapper Total Column SO2 Product: Continuation of NASA Long-term Global Data Record“ describes the NOAA-20 (N20)/OMPS SO2 product, which aims at extending the long-term climate data record of OMI, SNPP/OMPS of SO2 column densities from both anthropogenic and volcanic activities.

The authors not only describe the new algorithm for N20/OMPS but also perform a comparison with the existing data record, showing the added value of this additional satellite product.

The manuscript is very well written and already in a very good state and require only minor revision, as detailed in the detailed comments hereafter:

Detailed comments:

Figure 1: Instead of showing arrows for each instrument that all end in 2024 it is perhaps better to show the planned mission timeline. Otherwise one gets the impression that all missions end in 2024 (except for JPSS3-4/OMPS). What is meant with “Direct readout only” for NOAA-21/OMPS?

Section 2.2.1 Line 134: Perhaps mention which RT code is used for the calculation of AMFs

Section 2.2.1 Line 162: You mention that the pixels are subdivided into 3 subgroups based on their latitude. What is the latitude range of each subgroup?

Section 2.2.2 Line 185: You are using two reference orbits to derive the PCs. Can you show or indicate the effect on your results when you use a different orbit/day, e.g. in 2024?

Section 2.3: I guess think this section should be moved to the end since it is out of context at this location and disturbs the readability.

Section 2.4 Line 248. Typo “the the”

Section 3.1, Figure 2 and 3: From the figure you see an offset between the mean SO2 VCD of SNPP/OMPS and N20/OMPS. Where is this bias of SNPP/OMPS coming from? Maybe you should address this as well. Does a rebinning of N2O/OMPS have an effect on the mean SCD and associated bias?

Section 3.2 Figure 4: Why do you see a stronger difference in mountain areas, especially in the South American Andes (negative) and Scandinavia (positive difference). Is this related to icy surfaces and related albedo effects?

Section 3.3 Third paragraph & Figure 6: The differences in the text and in the figure subtitles are slightly different, probably due to different rounding. E.g. for Norilsk a 8% difference is written in the text, but the figure title states 7% difference.

Section 3.3 Figure 6. It is really hard to distinguish the three colored lines from each other. Maybe it would help if you show only the timeframe with N20/OMPS results, i.e. show the plot with data from 2018 onwards?

Section 3.3 Figure 7 Same suggestion as the two above: Perhaps show only data for 2018+ and check numbers in text and figure title.

Section 3.4 Figure 8. The x axis label of d and e are missing a “/” character, i.e. N20/OMPS instead of N20 OMPS.

Section 3.4 Figure 8d-f. Perhaps you find a better y axis label, since “OMI and SNPP ratio” is a bit hard to understand when only looking at the figure. Perhaps use “OMI emission/uncertainty ratio” or so.

Section 3.5 Figure 10 Perhaps it would be useful to show the comparison with TROPOMI.

Section 4, This section should appear after the conclusions and then Section 2.3 should come after (see my comment above).