Cloud Top Thermodynamic Phase from Synergistic Lidar-Radar Cloud Products from Polar Orbiting Satellites: Implications for Observations from Geostationary Satellites

Round 1

Reviewer 1 Report

BRIEF SUMMARY

This manuscript employs vertically-resolved radar-lidar DARDAR data to investigate issues related to cloud top thermodynamic phase, to support the goal of cloud top phase retrievals from passive instruments on geostationary satellites.  The cloud top phase is classified into four categories (ice, mixed-phase, supercooled liquid and warm liquid) and it is evaluated based on the phase information in the DARDAR data for the uppermost 240 meters of the
cloud, to roughly mimic what would be seen by passive satellite instruments.  Firstly, the impact of spatial aggregation of DARDAR data (1.1 x 1.7 km resolution) to the lower-resolution SEVIRI pixels (3 x 3 km at nadir) is considered. The impacts are relatively minor, except for a more frequent occurrence of mixed-phase cloud tops when evaluated at the coarser resolution. Second, climatological features (i.e., spatial and seasonal dependence) of cloud top phase are analyzed using the spatially aggregated data. Third, the spatially aggregated DARDAR data is employed to evaluate the CiPS product which retrieves cloud top phase (ice, non-ice) from SEVIRI data on the MSG2 dataset.

GENERAL COMMENTS

The topic is relevant for Remote Sensing. The paper is generally well-written and of good technical quality, although rather long as it addresses three separate but related subtopics (as noted in the broad comments). Obviously, the issue of cloud top phase has been addressed in many previous studies, but the idea of using lidar/radar data for providing the "ground truth" of cloud top phase for SEVIRI might well be new. One issue that probably deserves more discussion is the definiton of "SEVIRI-like" cloud top phase. It ís not fully clear how well the selected 240 m depth for analyzing cloud top phase represents SEVIRI observations. Even if/when this cannot be addressed comprehensively in this study, some sensitivity tests would be warranted. See specific comment 3.

SPECIFIC COMMENTS

1. The title of this paper only mentions geostationary satellites. I found this slightly misleading because the paper mostly employs active remote sensing DARDAR data (geostationary satellite data is actually only used in Section 4, for evaluation of the CiPS algorithm). So I recommend to also mention DARDAR in the title. One possibility: "Cloud-top thermodynamic phase from radar-lidar DARDAR data: implications for retrievals from geostationary satellites".

2. The introduction  is quite long (125 lines). Consider splitting it into subsections (e.g., 1.1. Background, 1.2. Scope of the present work), where the latter part could start from around line 111.

3. lines 270, 289: The choice of considering four DARDAR gates (i.e., the uppermost 240 m) may be a reasonable compromise, but it is still arbitrary because (as acknowledged on lines 285-296), the thickness of the layer dominating the signal seen by a passive instrument depends on the cloud extinction coefficient, and also on the spectral bands considered.  The sensitivity of the results to this choice should be tested.  One option for that: show the zonal annual-mean phase occurrence frequencies for IC, MP, SC and LQ for the default 240 m sampling and some alternative values (120, 360 m?).  
4. lines 279-280: "for multilayered clouds ... we only consider the uppermost cloud gate (even if its geometrical thickness is thinner than 240 m)". This is
a bit confusing. Do you mean that for multilayer clouds you only consider the uppermost 60 m (if so, state this explicitly)? Or that you consider only the gates within the uppermost cloud (240 meters or less, depending on the cloud geometric thickness)?

5. line 334: I suggest replacing "different spatial resolutions" with "different  horizontal resolutions". The vertical sampling is still the same for SEVIRI-like CTP and DARDAR CTP.

6. lines 357-361: "Next ... (Section 3.3-4.4). Finally ... (Section 3.2)". Please correct the order of these sentences.

7. line 379 and Fig. 2. It is stated that the percentage of omitted pixels is small in general (at most 0.2%). This sounds extremely small. Should it actually be 20%? Also, if it is only 0.2%, it is hard to see how the issue of omitted DARDAR LQ values could have any significant effect on the differences between DARDAR and SEVIRI-like CTP discussed on lines 454-472.

8. lines 493-494: Cloud ice content (or cloud ice water content) usually refers to the amount of ice in units of g m^-3. It would be more appropriate to say "... with the fraction of cloud tops increasing ...".

9. lines 534-535: "In some rare cases, MP can be observed for very high CTHs up to 17 km". While these cases are indeed rare, have you got any idea what causes them? Is it a failure of the DARDAR algorithm?  In the tropics (where these cases occur), the temperature at 17 km is far too cold (ca. -70 C or lower) for liquid droplets to survive. Also the CTH variations should not be the reason, as you screen out cases with CTH variations of more than 1 km (lines 310-311).

10. Figure 6 is a bit cluttered. The clarity could be improved by showing only JJA and DJF (the shoulder seasons MAM and SON seem not to be discussed in the text, anyway).

11. line 651: As a follow-up to comment 9, the large False Alarm Rate for very cold cloud top temperatures (ca. 220 K) in Fig. 7a (where DARDAR indicates mixed phase but CiPS ice) could be an artifact of DARDAR data. On physical grounds, mixed-phase clouds should not occur in these temperatures.

12. Lines 652-646: In this long sentence, the parts in parentheses require
some grammatic correction. Suggestion: "We find that the FAR of CiPS for clear-sky (i.e., pixels classified as clear-sky in the SEVIRI-like data set and as ice by CiPS) depends mainly on the latitude, while the FAR of CiPS for non-ice cloudy pixels (i.e., pixels classified as LQ, SC or MP in the SEVIRI-like data set and as ice by CiPS) depends mainly on CTT."

13. lines 781-1184: Please correct the technical issues with the reference
list (it is now given twice and starts from reference 81!).

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The authors did a good job to show variation of cloud top phase at the SEVIRI resolution. The active remote sensing cloud phase data were used to simulate what cloud top phase will be at the SEVIRI resolution. It is a very interesting research that suggest a great potential detecting cloud top phase more clearly from geostationary satellite imagers, therefore, I suggest to accept this submission after a minor revision.

1. I fell the title is not appropriate. My first impression is that the authors would detect cloud top phase by using geostationary satellite measurements.

2. Line 95-96, it is not very easy for understand this point, what's the relationship between the first part and the second part?  

3. Line 123, a modeled, modeled here is not appropriate although I have no idea which word is more appropriate,

4. Line 139-143, actually, it seems not clear whether the mixed and super cooled cloud top can be detected by imagers onboard geostationary satellites.

5. I'm not sure why the reference list begins with 100, not 1.

Author Response

Please see the attachment.

Author Response File: Author Response.docx