Next Article in Journal
Urban Air Quality in a Coastal City: Wollongong during the MUMBA Campaign
Next Article in Special Issue
A Geostationary Instrument Simulator for Aerosol Observing System Simulation Experiments
Previous Article in Journal
Temporal Patterns and Vertical Temperature Gradients in Micro-Scale Drainage Flow Observed Using Thermal Imaging
Previous Article in Special Issue
Response of Surface Ultraviolet and Visible Radiation to Stratospheric SO2 Injections
 
 
Review
Peer-Review Record

A Review of Ice Cloud Optical Property Models for Passive Satellite Remote Sensing

Atmosphere 2018, 9(12), 499; https://doi.org/10.3390/atmos9120499
by Ping Yang 1,*, Souichiro Hioki 1, Masanori Saito 1, Chia-Pang Kuo 1, Bryan A. Baum 2 and Kuo-Nan Liou 3
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Atmosphere 2018, 9(12), 499; https://doi.org/10.3390/atmos9120499
Submission received: 13 October 2018 / Revised: 5 December 2018 / Accepted: 11 December 2018 / Published: 17 December 2018
(This article belongs to the Special Issue Radiative Transfer in the Earth Atmosphere)

Round  1

Reviewer 1 Report

Reviewer’s comments on “A review of ice cloud optical property models for satellite remote sensing” by Ping Yang et al.


General comments: This is a review article on development of bulk ice scattering models for satellite remote sensing applications. The authors start from theoretical consideration, carefully laying down the foundation. The following section reviews various ice optical models that have been developed over the years for remote sensing purposes with a focus on satellite applications, in particular, those related to the MODIS and CERES teams.

 

The authors are with no doubt world-class experts in the area they reviewed. The review article is well written. I don’t have any major comments. Here are just a few minor suggestions to help improve the presentation of the paper.

 

(Line 63-65) Passive imager measurements go beyond IR and extends all the way to microwave, even for the purpose of cloud remote sensing. For example, microwave imagers have been routinely used to retrieve cloud property such as liquid water path (e.g., Greenwald et al. 1993; JGR). But if the authors wish to emphasize ice clouds and leave out mentioning of microwave, some clarification should be made here.

 

(Section 3) It will be a good idea to divide this big section into several subsections. In the current form, the authors put everything in Section 3 without a clear roadmap. A better structured section will help orient the readers. This is especially important for a review article.

 

(Summary) Given a long introduction and a sizable body part, it looks strange to me that the authors end the paper in such a rush. It will be nice if they can add more depth to the Summary section by, for example, including some discussion of future direction.

 

(Line 201, Line 323) Typo: quantify, not quantity.

 

Overall, I’d recommend the paper be accepted after minor revisions.


Author Response

We greatly appreciate the time and effort by the two reviewers, and thank them for their constructive and insightful comments and suggestions. We revised the manuscript to address the issues raised in the reviews. We hope the reviewers find that the revisions and the authors’ responses to the reviewers’ comments are appropriate and that the revised manuscript will be deemed acceptable for publication.


Response to reviewer 1


Comment 1

General comments: This is a review article on development of bulk ice scattering models for satellite remote sensing applications. The authors start from theoretical consideration, carefully laying down the foundation. The following section reviews various ice optical models that have been developed over the years for remote sensing purposes with a focus on satellite applications, in particular, those related to the MODIS and CERES teams.

The authors are with no doubt world-class experts in the area they reviewed. The review article is well written. I don’t have any major comments. Here are just a few minor suggestions to help improve the presentation of the paper.

 

Response

We thank the reviewer’s positive and constructive comments and suggestions.

 

Comment 2

(Line 63-65) Passive imager measurements go beyond IR and extends all the way to microwave, even for the purpose of cloud remote sensing. For example, microwave imagers have been routinely used to retrieve cloud property such as liquid water path (e.g., Greenwald et al. 1993; JGR). But if the authors wish to emphasize ice clouds and leave out mentioning of microwave, some clarification should be made here.

 

Response

This suggestion is well taken. Specifically, we have added descriptions regarding passive microwave measurements in the corresponding paragraph as Passive microwave measurements are useful to estimate not only liquid hydrometeor properties (e.g., Greenwald et al., 1993) but also ice counterparts (e.g., Wu and Dong, 2017).”.

 

Comment 3

(Section 3) It will be a good idea to divide this big section into several subsections. In the current form, the authors put everything in Section 3 without a clear roadmap. A better structured section will help orient the readers. This is especially important for a review article.


Response

Based on this suggestion, we divided Section 3 into three subsections.

 

Comment 4

(Summary) Given a long introduction and a sizable body part, it looks strange to me that the authors end the paper in such a rush. It will be nice if they can add more depth to the Summary section by, for example, including some discussion of future direction.

 

Response

The expanded summary provides more discussion, including potential future directions.

 

Comment 5

(Line 201, Line 323) Typo: quantify, not quantity.

 

Response

Corrected.


Reviewer 2 Report

As the authors note in the manuscript, different teams tend to apply different particle models in satellite retrievals of ice hydrometeor properties. Accordingly, a review article is motivated. However, the scope of the review is unclear.

First of all, is the manuscript inside the scope of the Atmosphere journal? The manuscript deals only with remote sensing aspects, as its title states. Implications for atmospheric physics or climate modelling are not considered at all (such as done in the similar review by Baran (2012) [3]). The manuscript presents particle models giving consistency between 0.9 and 2.3 um. This is promising, but does not imply that these particle models actually mimic reality fully and e.g. are valid at 10 um. Accordingly, the application outside remote sensing is unclear (and is not discussed). For this reason I suggest to redirect the manuscript to the Remote sensing journal.

Further, as mentioned, the manuscript deals only with a quite narrow spectral range. In addition, only passive observations are considered. Considering the high importance of the active observations by Calipso and CloudSat it is surprising that the particle models assumed for these instruments not are discussed. Either the manuscript has to be broadened, or the title must be changed to reflect actually content. As it is now, the title indicates a broader scope than the content covers.

While particle shape/habit is treated carefully, very little attention is given to the second part of a particle model, the particle size distribution (PSD). The PSD of each particle model should be reported and the impact of PSD should be clarified. Now all differences seem to be attributed to habit. Or PSD outside of the scope? If yes, then the title should reflect this.

Some further comments:

Section 3 should be divided into sub-sections.

I don't find the motivation for assuming random orientation convincing, or unclear. My understanding is that the difference between Fig. 6b and 6c reveals the fraction of particles having basically a perfect horizontal orientation. This fraction should be low, but this does not imply that totally random orientation dominates. The particles could still have preference towards horizontal orientation. That is, they could be "fluttering". The results by Gong and Wu (2017) strongly indicates that some degree of horizontal orientation is common, at least for larger ice particles.

The motivation in start of the Introduction refers to some relatively old papers. Reference [5] refers to the status of climate models before 2005. I assume the status is better today. Reference [12] is from 2009. Has not CloudSat improved the knowledge on IWP substantially?


References

Gong, Jie, and Dong L. Wu. "Microphysical properties of frozen particles inferred from Global Precipitation Measurement (GPM) Microwave Imager (GMI) polarimetric measurements." Atmospheric Chemistry and Physics 17.4 (2017): 2741-2757.

Author Response

We greatly appreciate the time and effort by the two reviewers, and thank them for their constructive and insightful comments and suggestions. We revised the manuscript to address the issues raised in the reviews. We hope the reviewers find that the revisions and the authors’ responses to the reviewers’ comments are appropriate and that the revised manuscript will be deemed acceptable for publication.


Response to reviewer 2


Comment 1

As the authors note in the manuscript, different teams tend to apply different particle models in satellite retrievals of ice hydrometeor properties. Accordingly, a review article is motivated. However, the scope of the review is unclear. First of all, is the manuscript inside the scope of the Atmosphere journal? The manuscript deals only with remote sensing aspects, as its title states. Implications for atmospheric physics or climate modelling are not considered at all (such as done in the similar review by Baran (2012) [3]). The manuscript presents particle models giving consistency between 0.9 and 2.3 um. This is promising, but does not imply that these particle models actually mimic reality fully and e.g. are valid at 10 um. Accordingly, the application outside remote sensing is unclear (and is not discussed). For this reason I suggest to redirect the manuscript to the Remote sensing journal.


Response

We agree that the manuscript is more suitable for a journal focusing on remote sensing. However, this manuscript is submitted to a special issue in Atmosphere, organized by Professor Irina Sokolik as a guest editor. To limit the scope of the manuscript, we do not discuss the applications of ice cloud radiative property models to climate study.

 

As the reviewer pointed out, we described the consistency of ice particle models between 0.9 and 2.3 um. Figure 17 also illustrate the consistency between visible-to-shortwave-infrared (VIS–SWIR) and thermal infrared (TIR) retrievals of optical thickness (i.e., 11 and 12 um). No ice particle model can reproduce perfect fits in optical thickness retrievals between VIS–SWIR and TIR measurements due to error sources other than ice particle model differences (e.g., cloud 3D effects appear differently in VIS–SWIR and TIR measurements). The main point illustrated by Figure 16 is that either MODIS C6 or THM can provide better consistency than other earlier models between VIS–SWIR and TIR based retrievals. In addition, we have added paragraphs regarding consistency with polarimetric observations in Section 3.

 

Comment 2

Further, as mentioned, the manuscript deals only with a quite narrow spectral range. In addition, only passive observations are considered. Considering the high importance of the active observations by Calipso and CloudSat it is surprising that the particle models assumed for these instruments not are discussed. Either the manuscript has to be broadened, or the title must be changed to reflect actually content. As it is now, the title indicates a broader scope than the content covers.


Response

This point is well taken. In particular, we explicitly state in the title that the review is “for passive satellite remote sensing”.

 

Comment 3

While particle shape/habit is treated carefully, very little attention is given to the second part of a particle model, the particle size distribution (PSD). The PSD of each particle model should be reported and the impact of PSD should be clarified. Now all differences seem to be attributed to habit. Or PSD outside of the scope? If yes, then the title should reflect this.

 

Response

The ice particle habit has a primary impact on optical properties of ice particles, and the review paper mainly focus on this aspect. We do note, however, that references were cited on in-situ measurements of ice clouds from numerous field campaigns (e.g. Heymsfield et al. 2013). We added an additional reference (Baran et al. 2018) that discusses the inference of bulk single-scattering properties at millimeter to sub-millimeter wavelengths, with further description of PSDs therein.

 

Comment 4

Some further comments:

Section 3 should be divided into sub-sections.

 

Response

In response to this suggestion, we have divided Section 3 into three subsections.

 

Comment 5

I don't find the motivation for assuming random orientation convincing, or unclear. My understanding is that the difference between Fig. 6b and 6c reveals the fraction of particles having basically a perfect horizontal orientation. This fraction should be low, but this does not imply that totally random orientation dominates. The particles could still have preference towards horizontal orientation. That is, they could be "fluttering". The results by Gong and Wu (2017) strongly indicates that some degree of horizontal orientation is common, at least for larger ice particles.

 

Response

Differences between Figs. 6b and 6c are due to quasi-horizontally oriented plates (not perfectly horizontally oriented plates). Gong and Wu (2017) (Ref. 26 in the revised manuscript) are based on polarimetric microwave passive measurements. We explicitly state Based on the current evidence, it is quite reasonable to assume ice particles to be randomly oriented for passive remote sensing applications that do not involve polarimetric measurements.. In the revised manuscript, we have added “Note that polarimetric microwave measurements are sensitive to hydrometeor orientations [26], and the assumption of random orientation may introduce errors in ice cloud retrievals based on the measurements.

 

Comment 6

The motivation in start of the Introduction refers to some relatively old papers. Reference [5] refers to the status of climate models before 2005. I assume the status is better today. Reference [12] is from 2009. Has not CloudSat improved the knowledge on IWP substantially?

 

Response

Thank you for the reviewer’s insightful comment. In the revised manuscript, we have added three references on IWP inferred with active sensor measurements (Matrosov 2015 and Deng et al. 2013, 2015)

Round  2

Reviewer 2 Report

In my review I found the particle size distribution (PSD) part poorly covered. The authors have made some changes to address this, but to what extent the influence of PSD is inside the scope or not is still not totally clear. The change of title could be interpreted as that less focus is given to PSD, but in the abstract habit and PSD are given the same weight. The reply to my comments seems to indicate that habit is more important than PSD. If this is true, this should be discussed in the manuscript. In fact, I think the authors could show this quite simply, e.g. by including CERES Edition 4 in Figure 15. This would at least show to what extent surface roughness matters.

I appreciate the longer summary section. This is a clear improvement, but I can not identify any concrete conclusions. It is clear that better spectral consistency has been achieved. But why? I get the feeling that inclusion of surface roughness is the main factor. Do the authors know?

One thing I missed to comment last time is Eq 21. The expression is said to be
"optimal estimation" (OE). The standard definition of OE implies that statistical regularisation, using a priori information, is performed. See
https://en.wikipedia.org/wiki/Optimal_estimation
Minimising J of Eq 21 corresponds to weighted least squares, see e.g. Menke 1989 and:
https://en.wikipedia.org/wiki/Weighted_least_squares

The new text around Figure 14 can be improved. I suggest to start with clearly describing the selection procedure. Now there are just a few comments at the end. Seems that is "ocean only" but this is not explained. It is said that "283594 pixels are selected". Does this mean that there is an additional "selection", or that the number of pixels found during the month of consideration?

Further comments:

Eq. 6 and 7: I suggest to merge these two equations, as the first step in Eq 7 is a repetition of Eq 6.

Line 480: What is "semi-maximum dimension"?

Figure 12: The panels look identical between top and bottom row. A mistake? If not, show results for only one wavelength and explain in text that the results are the same for the other wavelength.

Figure 13: Top and bottom row look identical also here. Shall "Effective radius is fixed at 30 um." in figure text be removed? If not, then the x-label of the figure must be wrong.

Figure 15: Seems unnecessary to have two panels here. Both lines can be put
into one panel.

References:

Gong, Jie, and Dong L. Wu. "Microphysical properties of frozen particles inferred from Global Precipitation Measurement (GPM) Microwave Imager (GMI) polarimetric measurements." Atmospheric Chemistry and Physics 17.4 (2017): 2741-2757.

Menke, W. "International Geophysics Series." Geophysical data analysis: Discrete inverse theory 45 (1989).


Author Response

We greatly appreciate the time, effort and constructive comments and suggestions. We revised the manuscript to address the issues raised in the review. We hope the reviewer find that the revisions and the authors’ responses to the reviewer’s comments are appropriate and that the revised manuscript will be deemed acceptable for publication.

 

Response to reviewer 2


Comment 1

In my review I found the particle size distribution (PSD) part poorly covered. The authors have made some changes to address this, but to what extent the influence of PSD is inside the scope or not is still not totally clear. The change of title could be interpreted as that less focus is given to PSD, but in the abstract habit and PSD are given the same weight. The reply to my comments seems to indicate that habit is more important than PSD. If this is true, this should be discussed in the manuscript. In fact, I think the authors could show this quite simply, e.g. by including CERES Edition 4 in Figure 15. This would at least show to what extent surface roughness matters.

 

Response

In response to the reviewer’s comments, we changed the title of the paper as “A review of ice cloud optical property models for passive satellite remote sensing” to deemphasize the focus on particle habit. The generation of bulk optical properties incorporates both particle size and habit distributions. Each team assumes different particle habit and size distributions, and this is clearly discussed in the manuscript. Also, as pointed out in Reference 27, modelers and active sensor teams may also employ different assumptions about ice particle habit and size distributions.

 

In particular, we call special attention to the paragraph just below Fig. 9, which emphasizes the importance of the particle size distribution. Specifically, we state that “The other important information necessary for the derivation of the bulk scattering properties is knowledge of the particle size distribution (PSD). Historically, the PSDs have been defined for this effort in two ways for ice cloud remote sensing by passive imagers: through the use of a well-defined gamma distribution, or through the use of in-situ measurements. The obvious benefit to use of a gamma distribution is that one has total control over the size limits and width of the PSD. The use of actual in-situ data is quite involved, but the relevant bulk single-scattering models can be assessed for different dynamical regimes (e.g., low updraft velocity such as synoptic cirrus versus tropical anvils formed in a convective dynamical environment). Different approaches may be adopted for active sensors, which can sense much larger particles and infer much higher values of IWP, and millimeter to sub-millimeter measurements [27].” An interested reader is referred to the paper by Baran et al. [27].


[27] Baran, A. J.; Ishimoto, H.; Sourdeval, O.; Hesse, E.; Harlow, C. The applicability of physical optics in the millimeter and sub-millimetre spectral region. Part II: Application to a three-component model of ice cloud and its evaluation against the bulk single-scattering properties of various other aggregate models. J. Quant. Spectrosc. Radiat. Transf. 2018, 206, 83-100, doi=10.1016/j.jqsrt.2017.10.027

 

We do not include CERES Edition 4 model-based result in Fig. 15 because it has been shown in a previous study (Yang, P., L. Bi, B. A. Baum, K. N. Liou, G. W. Kattawar, M.I. Mishchenko, and B. Cole, 2013: Spectrally consistent scattering, absorption, and polarization properties of atmospheric ice crystals at wavelengths from 02 to 100 µm. J. Atmos. Sci., 70, 330-347) that ice particle surface roughness leads to a better agreement between simulated and observed polarized reflectances.

 

Comment 2

I appreciate the longer summary section. This is a clear improvement, but I can not identify any concrete conclusions. It is clear that better spectral consistency has been achieved. But why? I get the feeling that inclusion of surface roughness is the main factor. Do the authors know?


Response

In response to the reviewer’s comment, we added a conclusion in lines 748–751 as “These two models incorporate severely roughened particle surfaces that lead to featureless bulk phase functions. At present, the MODIS Collection 6 and the CERES two-habit models are optimal ice particle models for passive remote sensing of global ice clouds.”


Comment 3

One thing I missed to comment last time is Eq 21. The expression is said to be
"optimal estimation" (OE). The standard definition of OE implies that statistical regularisation, using a priori information, is performed. See
https://en.wikipedia.org/wiki/Optimal_estimation
Minimising J of Eq 21 corresponds to weighted least squares, see e.g. Menke 1989 and:
https://en.wikipedia.org/wiki/Weighted_least_squares

 

Response

The optimal retrieval state is obtained based on a maximum likelihood method as described in the manuscript. We have cited Menke (1989) and revised the description accordingly.


Comment 4

The new text around Figure 14 can be improved. I suggest to start with clearly describing the selection procedure. Now there are just a few comments at the end. Seems that is "ocean only" but this is not explained. It is said that "283594 pixels are selected". Does this mean that there is an additional "selection", or that the number of pixels found during the month of consideration?

 

Response

We have revised the paragraph accordingly to enhance the clarity.

 

Comment 5

Eq. 6 and 7: I suggest to merge these two equations, as the first step in Eq 7 is a repetition of Eq 6.

Response

The suggestion has merit, but we wrote this review for people who are working on remote sensing but unfamiliar with the light scattering theory. Therefore, we prefer to keep Eqs. 6 and 7.

 

Comment 6

Line 480: What is "semi-maximum dimension"?


Response

In the revised manuscript, it is defined as “r indicates half of the maximum characteristic dimension (i.e., the radius in the case of a sphere) of an individual particle.”

 

Comment 7

Figure 12: The panels look identical between top and bottom row. A mistake? If not, show results for only one wavelength and explain in text that the results are the same for the other wavelength.


Response

The error has been corrected – good catch.

 

Comment 8

Figure 13: Top and bottom row look identical also here. Shall "Effective radius is fixed at 30 um." in figure text be removed? If not, then the x-label of the figure must be wrong.


Response

The error in Fig. 13 has been corrected. In addition, we have removed “Effective radius is fixed at 30 µm”.  

 

Comment 9

Figure 15: Seems unnecessary to have two panels here. Both lines can be put
into one panel.


Response

Following the reviewer’s suggestion, we have merged the two panels to one panel.  


Back to TopTop