Review of the paper

Real-time UV-Index retrieval in Europe using Earth Observation based techniques and validation against ground-based measurements by Panagiotis G. Kosmopoulos et al.

The paper deals with the description of UV indices evaluation using the Earth observation system in Europe. This, so called UV-Index Operating System, or UVIOS exploits both radiative transfer models and the data available from Meteosat Second Generation and Meteorological Operational Satellite-B as well as the information available from Tropospheric Emission Monitoring Internet Service, Copernicus Atmosphere Monitoring Service and the Global Land Service.

The simulations include the account of main factors affecting UV radiation: ozone, clouds and aerosols as well as ground elevation and surface albedo with resolution of 5 km and 15 minutes.

This work is highly important “for the provision of operational early warning systems that will help raise awareness among European Union citizens of the health implications of high UVI doses” as the authors wrote.

I like the idea of this approach and its technical solution and can recommend paper for publishing. However, I have several comments, which are presented below.

Natalia Chubarova

Comments:

1. row 57. I would recommend to add the references to the numerous EEAP reports (for example, ENVIRONMENTAL EFFECTS OF OZONE DEPLETION AND ITS INTERACTIONS WITH CLIMATE CHANGE: 2014 ASSESSMENT).
2. row 73. The only mentioning UV-A as a spectral region, where the NO2 play important role seems to be misleading. See, for example, Table 7-1 from the Ozone Assessment 2006 (Chapter 7) concerning the role of NO2 and SO2 effects on erythemal irradiance. Should be clarified in the text.
3. Organic gases like formaldehyde can be also important in both UV-B and UV-A regions. I would recommend to re-write this part taking this into account.
4. row 103. The areas with extremely high positive UV trends over Northern Eurasia over the 1979-2015 period were shown recently in (Chubarova et al., 2020). (https://www.mdpi.com/2073-4433/11/1/59).
5. row 108. The reference should be given concerning the turnout point in UV trend in 2007.
6. row 145. This is not exactly so, since the method proposed by Jean Verdebout used geostationary Meteosat instruments data. This should be accounted for in the text. (Verdebout, J., A method to generate surface UV radiation maps over Europe using GOME, METEOSAT, and ancillary geophysical data, J. Geophys. Res., [Atmos.] 105, 5049–5058, 2000. )
7. row 160. The authors should begin this part mentioning that using their approach they could combine information on input parameters from different satellite sources to provide the better quality UV estimates. I would recommend to re-write the text.
8. row 188. I do not see the information on factor of asymmetry of aerosol phase function in the list, which is one of the important aerosol parameters, necessary for model simulations. Also I do not see the cloud amount parameter in the list. I understand the difficulties with its application but this should be discussed here in the text.
9. row 198-199. The references should be provided to the internet link at least.
10. row 231. The title should be changed. Like ”The description of the geophysical parameters”, for example.
11. 248. “However, since such measurements are associated with very low UV Index (<1). Depending on different parameters ( ozone, cloud amount).” Should be proved by simulations.
12. row 261. The reference should be given to the Albedo product. The parenthesis is missed.
13. row 268. I would recommend also to add the reference to Table1 here.
14. 271. What is the range of overestimation?
15. row 279. Misprint ( I Note)
16. row 316. Previously you mentioned the threshold of 75 degrees for MSG COT retrievals ( row 246)? Should be clarified.
17. row after 333. I do not find the information on altitude correction. Since all other factors are analyzed here it should be also discussed here even you have the detailed analysis after.
18. row 338. I would propose to replace “while” on “and”
19. row 340-348. I would propose to re-write this part in a more compact way. This is obvious.
20. row 357. Remove the extra dot, please.
21. row 370-371. The values should be given.
22. 452. The reference should be given or it should be clarified that this has been obtained using model simulations provided by the authors.
23. row 458. “conditions”
24. row 468. – “conditions”.
25. row 475 – change to: and “in case”
26. row 500. Concerning the changes with altitude: there are other factors in addition to TOC, which may influence the altitude dependence like aerosol and surface albedo. Please, look for details in our paper (Chubarova et al., 2016, ACP, https://acp.copernicus.org/articles/16/11867/2016/)
27. row 507. The estimates in term of UVI should be made here to be consistent with other sections.
28. row 543. SSA? Seems to be misprint. SSA is usually used as single scattering albedo abbreviation. Here you describe the albedo effects. If you are speaking about real SSA, this should be made closer to aerosol effect discussion.
29. row 926. “result in”
30. row 1036. It would be nice to see in this Table also the RMSE and R statistics, like in Table 5.
31. Figure 6. The name of Y-axis should be changed.
32. 1275. (a) – is not clear. To my understanding the angular dependence due to 3D geometry should be taken into account. Please, clarify.

General Comments

The manuscript evaluates the UV index generated from the UV- Index Operating System (UVIOS) and also conducted sensitivity tests to quantify the uncertainty caused by the different input parameters. The system is important for providing early warning, which will benefit the general public. The evaluation of the UV estimation against ground observation is thorough. However, the presentation of the manuscript can be improved, and I have provided some detailed comments as follows. Lastly, English needs to be improved. There are grammar errors in some sentences.

• The manuscript mentions that the UVIOS system provides both now-casting and forecasting UV index products. It is not clear how different they are and it would be helpful to provide some information on these two products.
• Line 319-322 talks about the different stations. It can be moved to the next paragraph where it focuses on describing the geographic and climate information about each station.
• I suggest changing title 3.1 to Overall performance of the UVIOS system.
• Line 368: how did you decide the criteria for low, moderate and high UVI differences?
• For Figure 4, it would be nice to show a map of the annual UV index values from the UVIOS system overlaid with ground observational data.
• Besides Figure 5, it would be helpful to have a scatter plot here to better show the overall performance for all the stations together.
• Table 3 is hard to read. Figure 11 can be used as a complement. I would suggest showing a histogram distribution of the errors for all the stations together rather than showing each station on one plot with different lines. Histogram for individual station can also be shown as supplementary plots.
• Figure 9 and Figure 15 can be put together into one figure.

Corrections:

Line 98: what is “0 to 8/8”?

Line 99-102: This sentence has some grammar errors, please check and reorganize the sentence.

Line 105: It is not necessary. Or what is the main conclusion of Calbo et al. (2005)?

Line 117: what does it mean by in contrast to ozone”? Is ozone column not important for the trend?

Line 170: change next section to Section xx.

Line 292: change Aeronet to AERONET.

Line 344: which section is it for the cloud effect section?