Sensitivity of Multispectral Imager Liquid Water Cloud Microphysical Retrievals to the Index of Refraction

Round 1

Reviewer 1 Report

Review report on the paper titled “sensitivity of multispectral imager liquid water cloud microphysical retrievals to the index of refraction” in possible consideration of its publication in Remote sensing journal of MDPI.

1. Please define CER correctly to have consistency in the abstract and introduction
2. The motivation and significance of the work is not clearly presented. I encourage the authors to elaborate on the same.
3. Further, the objectives are not up to the mark and doesn’t found interesting to conduct the research. Please rewrite the objectives in view of importance of the work what is lagging and what can be done.
4. Also, the status of work that has already conducted globally is needed to be covered and discussed in the introduction.
5. Lines 170-180: I am not sure about the format followed by the journal.
6. Change section name for 3.2.
7. The results are nicely presented and at several instances the phrasing of sentence is not clearly. However, I recommend the authors to look in to the same before submitting the revision
8. Most of the times the authors are not following or inconsistent with the format or template given by the journal.
9. The conclusions is part too largely written and suggest to concise a little to avoid duplication of information with that mentioned in the results or discussion sections
10. Some amount of English grammar need to be refined to meet journal standards.

Author Response

Please see attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

BRIEF SUMMARY

The retrievals of cloud droplet effective radius CER rely on observations of reflectance in SWIR or MWIR channels. These reflectances depend mainly on cloud optical thickness (usually retrieved from visible channels) and the single-scattering co-albedo, which is roughly linearly proportional to the product of CER and the imaginary part of k of the refractive index. Therefore uncertainties in k give rise to uncertainties in CER. It is shown that systematic differences in retrieved CER related to different refractive index datasets can be surprisingly large, up to 1-2 µm, which is comparable to the total effect of other error sources in the retrieval of CER. A further complication is that the spectral channels of different satellite instruments are not identical; for example, in the near-2.2 µm window region, there is a MODIS band centred at 2.13 µm and a VIIRS band centred at 2.24 µm. Differring refractive index errors in the different channels translate into inconsistent values of CER. It is demonstrated in the context of the MODIS-VIIRS CLDPROP continuity cloud product, that the consistency between CER retrieved from the two
instruments can be improved by using the Kou et al. (1993) refractive index dataset for a temperature of 265 K.

BROAD COMMENTS

Overall, this is a good paper. The topic is highly very relevant for Remote Sensing. The paper represents important results that may be entirely new, and certainly not widely-known beforehand. The analysis represented in the paper gives a strong impression that the authors know what they are doing. I did not find any significant flaws in the paper. The few comments listed below are targeted mainly to improve the clarity of some points.

SPECIFIC COMMENTS

1. Some acronymes shoud be defined. "VNIR" on line 110, SNPP on line 142.

2. In Figures 1 and 2, the green and yellow shadings mentioned in the caption are actually not visible, so it is difficult to distinguish the bands used for retrievals.

3. lines 250-251: I think it would be worth showing explicitly a comparison of the Kou et al. curves for 265 K and 295 K, including their relative difference. While it is in principle well-known that the refractive index is a function of temperature, I don't think many people have recognized this properly. Furthermore, showing this difference explicitly would be helpful for interpreting some of the results shown later, e.g. the temperature
dependencies in Fig. 9.

4. line 350: "... channel-to-channel CER differences can be substantial". I think it would be pertinent to mention that these differences arise not only from "errors" in the retrieval (due to uncertainties in k, and other factors), but also because CER varies vertically, and the channels with stronger absorption give a larger weight to layers near the cloud top.

5. line 447: "Retrievals for northern polar latitudes ...". This sentence could/should be moved to line 412.

6. lines 448-450: Figs. 7 and 8: A plausible explanation for the strongly negative VIIRS - MODIS differences in the middle row of Fig. 7, and their reduction in the center panel of Fig. 8, would be the following. To the extent that the Kou et al. (265 K) data is accurate, the Palmer and Williams (300 K) refractive index data underestimates absorption
for the MODIS band centred at 2.13 µm but overestimates absorption for the VIIRS band centred at 2.25 µm (Fig. 3), which would lead to overestimated CER for MODIS and underestimated CER for VIIRS. (If you agree on this, I think that stating this possibility explicitly would help the reader to understand this somewhat complicated matter).

7. lines 519-520: "If the Kou et al. index of refraction dataset at 265 K chosen for use in the 2.x µm CER algorithm is physically correct, then it is reasonable to expect that the VIIRS minus MODIS CER difference would be dependent on cloud top temperature". I cannot fully follow the logic ("how B follows from A") here. I think that a sufficient
condition for temperature dependent CER differences here would be that the imaginary part of refractive index k depends differently on temperature for the MODIS and VIIRS bands. Please clarify. See also comment 3.

TYPOS ETC.

1. line 111: remove "the of" from the "use of the of a VIS/NIR".

2. line 140: something missing here. "Cloud mask and optical property retrievals are made at a nadir resolution of ..."?

3. line 303: "emissive radiative flux calculations"? Do you mean "longwave radiative flux calculations"?

4. Figures 7 and 8, titles above the top row: Assuming that "VIIRS 1.1" and "VIIRS v1.1" mean the same, the same notation should be used.

5. Caption of Table 3 is finished prematurely. "Values are consistent with".

6. line 591: "underly" should be "underlie".

Author Response

Please see attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript is dedicated to finding an optimal value (optimal set of values) of refraction index used in effective particle radius retrievals from the spectrally resolved infrared radiances measured by modern satellite instruments. As the authors show, this problem is multilateral and it is linked to availability of laboratory measurements, their capacity of reproducing real atmospheric conditions, the realization of the retrieval algorithms, and the differences between the ideal droplets and real-life ones. The research is topical and it will serve well to the scientists and engineers involved in the multispectral cloud property retrievals.

The article is well-structured and well-written, and I wouldn’t hesitate to recommend publishing it in the journal in its present form, but I’d like to ask the authors to consider the questions formulated below in the Major and Minor comments. Perhaps, they could add a few sentences to the article to clarify these points and to make stronger methodological statements where it is appropriate.   

Major comments:

The authors did an excellent job of studying the sensitivity of retrieved cloud parameters on the used dataset, and this is a necessary part of any methodological work of this kind. However, there is another side of this story. Besides solving a forward task a number of times, which corresponds to a number of datasets used, one can also consider the problem as an inverse one and introduce something like “an effective index of refraction”, which would comprise not only the property measured in the lab, but also the effects of asymmetry, particle composition, distribution, measurements errors, etc. This is a kind of technical solution, but it would serve a purpose of a better cloud microphysical retrieval, so it would be easy to justify. For this, one would need to collocate the MODIS measurements with some “reference” (in situ?) dataset and to retrieve the searched parameter (or choose from the available ones). Without such a comparison, the study is not complete, in my opinion. I understand that this is a separate work, and that it wouldn’t be fair to ask the authors to make it in the framework of the current study, but at least some bridge to a more general approach is needed here.

Minor comments

Lines 55 and 92: this might be a question of definitions, but it looks like “A” in SSA should stand for “Albedo” and not for an “Absorption”. Below, in line 317, the authors use a term “co-albedo” for (1-SSA), which confirms “Albedo” for SSA, but creates another contradiction – in general, the albedo is a ratio and the absorption is not and it would be good to keep this approach throughout the article. I would suggest to re-read the article having this in mind and to get rid of these ambiguities in the definitions.

Line 117: I believe, this has been discussed elsewhere, but to be on a safe side I’d ask if the SWIR channels are sensitive to middle/upper atmosphere contribution, and if yes then how the latter was calculated. The problem is that CO2 emits at these wavelengths, and the emission is not thermal (it does not depend on local kinetic temperature)

Line 374: Fig. 4 – the authors write that the asymmetry is not that significant for the retrievals, but here we see that there’s a double solution associated with this parameter, and for some cases (e.g. for g=0.79) the corresponding r_eff changes by 8um. Perhaps, this needs some clarifying/commenting.

Line 420 and elsewhere – it is quite logical that the 265K LUT should give better results for clouds than the room-temperature one, but in my opinion the temperature dependence of LUT and the necessity of using the individual temperature LUTs in the retrievals should be postulated from the very beginning. The microphysical cloud properties should not be separated from cloud retrievals themselves and the latter provide the information on cloud pressure and, therefore, temperature, so the question of using the pertinent data is purely technical. I would make a stronger bullet point for temperature dependence of LUTs in the conclusions.

Author Response

Please see attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have taken care and consideration of comments/suggestions given by the reviewers which in turn made them to improve the quality of present version of paper. Hence, I am satisfied and feel judged that the authors have implemented replies to all my queries. Now, I recommend the journal editor for its publication in remote sensing journal (MDPI) with the current version of manuscript.