Review of Statistical validation of Aeolus L2A particle backscatter coefficient retrievals over ACTRIS/EARLINET stations in the Iberian Peninsula by Abril-Gago et al.
The paper describes the comparison/validation of Aeolus aerosol and cloud spin-off products with 3 ground-based sites on the Iberian Peninsula. The paper is well written and in a good shape. It is a valuable contribution to the scientific community, even though submitting it to AMT would have been more appropriate. Nevertheless, many issues from the open discussion has been addressed in the revised version and after considering my comments below, in my opinion, the paper can be published at Copernicus.
As I haven’t been involved in the first round of the review process (but both reviewers have done a great job), I focussed on the author’s replies to the reviewers from the public discussion but also have some own comments to the paper.
General remarks:
• There are many ESA internal references which seem to be not publicly available. All reference must be available somehow. You might discuss with ESA to submit these documents as supplement or ask if they can make it publicly available. I just googled some of them, an many documents can be downloaded, like for example the "Aeolus-Sensor-and-Product-Description ", which is freely available here: https://earth.esa.int/eogateway/documents/20142/37627/Aeolus-Sensor-and-Product-Description.pdf
Please add links to all ESA documents and or make them available to the reader in another way.
• With respect to the topic above, would like to mention that a description of the L2A retrievals is currently in discussion in the same special issue. Even though not yet accepted, it might be worth to cite: https://amt.copernicus.org/preprints/amt-2021-181/
• I am still struggling with the quantity you named BIAS. For me, a bias is a systematic error. But what you calculate is the Difference between ground-based lidar and Aeolus for some cases. As the reason of these differences could be also due to atmospheric nature (horizontal heterogeneity etc), I would prefer not to use this wording. Thus, I recommend to rename this parameter simply to (absolute) difference or something similar, but not bias, please.
• In my opinion, Figure 15 is not useful. It is known, that Aeolus uncertainty estimates (even though not perfect yet) handle only statistical uncertainties. Thus, it is not meaningful to compare them to the absolute differences (BIAS) derived from the comparison.
Not satisfying responses to the previous reviewers:
• Reviewer 1, comment 2:
R1: “…It is not very clear why the authors prefer to use the literature values instead of spectral conversion factor retrieved from measurements.”
Authors: “…Considering all these arguments, we prefered to use the spectral conversion factor derived from the literature in order to base our results on measurements of different stations (i.e. different environmental conditions and aerosol types), which themselves have been reviewed through peer review process.”
Me: I understand your arguments and can agree on your conclusions, but why to show “your own” conversion factors then, if they are not used?
This is confusing and has nothing to do with the topic! Please always have in mind: “Illustrations should only be shown if they are necessary for the understanding of the paper, not because they have been created. “
Therefore, I recommend to remove Fig. 2a and b, and only show Fig. 2c. The comparison of different depolarization ratio at different wavelength is not the topic of your paper but could be followed in another study.
• Reviewer 1, comment 4:
“In section 4.3 Case studies, the authors stated “Sun -photometer measurements are taken into account for the sake of completeness aerosol typing, through the study of the aerosol optical depth at 675 nm (AOD 675 )”. Why was the 657 nm wavelength chosen? Why not choose the AOD values in the UV region as 340 or 380 nm, instead?”
Me: In my opinion your response is not sufficient. As reviewer 1 noted, you could have used a wavelength much closer to Aeolus (440 nm). And I don’t think that 675 nm is most used in the community and is not the closed one to the wavelengths of your lidars. So I also do not see the point why to use that wavelength.
…anyhow, as the AOD is only of minor importance for your work, it is not crucial, but you might consider to change this.
• Reviewer 1, comments 5:
Authors:” Regarding Barcelona station location just in the coastline, as the satellite overpasses the station at a close distance we can assume that both instruments (ground-based and space-borne lidars) detect the same air masses. Consequently, both instruments will register the same effects that the geographical layout might produce, so no special considerations have to be taken into account in the statistical analysis.”
Again I think your response is not sufficient. In my opinion you cannot assume that “both instruments detect the same airmasses” given the complex location of Barcelona and the very coarse resolution of Aeolus with 8 km in the horizontal. Rather, I would assume, that the coarse horizontal resolution of Aeolus and the horizontal heterogeneity in this area is one of the main reasons for the discrepancies found. Therefore, I think, you should discuss this issue in the paper with some sentences. Then, your newly added sentences in the conclusion would also fit much better, as it holds not only for the vertical but also for the horizontal domain.
• Reviewer 1, comment 9:
“9) Page 15 - lines 474-475 - “With the implementation of the quality flags (Figure 12c and 12 d), all of the sets range from 0 Mm-1sr -1 .”. Please, consider correct this sentence. All the sets range from 0 to which value?
Authors: The quoted sentence has been restated and completed as: With the implementation of the quality flags (Figure 13c and 13d), all of the sets range from 0 Mm-1 sr-1 onwards. Actually, the maximum values mentioned are still flagged as valid, 86 Mm-1 sr-1 and 68 Mm-1 sr-1 in the case of the SCA and SCAmb, respectively.”
Me: the sentence is still awkward. I guess what you simply mean is that with the implementation of the QA/QC flags, no negative values are existent anymore?
• Reviewer 1, comment 10:
Authors: “Aeolus backscatter coefficient uncertainties (known as Aeolus error estimates) are addressed through the biases between satellite and ground-based measurements (as presented in Section 3.3).”
It is known, that the Aeolus uncertainties (even though not full developed) handle only statistical errors. So how can this be addressed through “your” BIAS?--> makes Fig. 15 obsolete.
Specific comments:
• Line 428: Why are values up to 85 Mm^-1 sr^-1 unrealistic? Please explain or give reference!
• Line558: I would use “Nevertheless” or “Even though” instead of “Thus”
• Line 935: Is the WMO report available somewhere? DOI? Internet page?
• Table 2 caption: Please write at least in the caption the full name of delta_par_linear. For example: “The linear particle depolarization ratio delta_par_linear at 355 nm….”
And please indicate the UNIT, so that one at least knows that the unit is not [%].
• Fig. 1, Caption: in spring 2021, the orbit has changed, can you indicate that here: E.g. “Distribution of Aeolus overpasses during the studying period from .. too...”
• Figure 2b Caption: Please indicate were the measurements were taken.
• Figure 3. Caption: Please remove “Daily”
• Fig 13 caption: Caption does not describe what is seen. Please write: Frequency disruption plot of ... derived from....
Also, please make the caption more concise.
• Fig. 14, caption: What is meant with “combined database”? Please write more exactly.
• Figure 15: See comment on Bias and Aeolus uncertainty above. In my opinion these plots are not useful.
• Caption Fig. 16: Replace measures by measurements. |
