This is a most comprehwensive, well written paper that carefully assesses the data and enumerates limitations of various assumptions made in the data assessment.  It presents the best assessment/prediction to date of retrieved aerosol lidar ratios and AODs obtained from CALIPSO observations, made possible using the independent CALIPSO-SODA lidar ratio retrievals for comparison.  Unceertainties in the CALIPSO Version 4.2 and the CALIPSO-SODA retrievals, and biases they contribute, are carefully considered.  Discussions about agreements/disagreements in the comparison results are well reasoned and extensive. 

The submitted paper titled “Assessment of tropospheric CALIPSO Version 4.2 aerosol types over the ocean using independent CALIPSO-SODA lidar ratios” retrieved aerosol extinctions and lidar ratios from a synergetic use of CALIOP L1 profiles and CALIPSO-SODA AODs then assessed lidar ratios used in the CALIOP version 4 aerosol algorithm. The results of this study are very meaningful and may be very helpful to develop further version of the CALIOP aerosol retrieval algorithm. The manuscript is well organized but needs minor revision before publication.

Some minor comments are as follows.

line 55-56: better to move to Section 2 (method). If you declare that you used data for only over ocean, you don't need to mention about plluted dust over desert land in line 65. 

line 81: So, this is your motivation of this study which is very simple. please describe limitations of previous studies in more detail to emphasize this study. You mentioned the spatiotemporal coverage is extremely limited. Is there any other related studies covering larger regions (global) or long period of time? Your study show any spatiotemporal vatiations which are not discussed in previous studies? Considering a long introduction, your motivation is too simple.

line 89: You have introduced only level 2 product. What about the CALIOP level 1 product used in your study? You had used VFM for cloud screen only for the range of 1 - 8.2 km? what about clouds out of that range? The title of subsection is "CALIPSO V4 and SODA data". But there is nothing about SODA.

line 117: You mean that any aerosol layers below BL are considered as clean marine, even if the CALIOP algorithm classify as other aerosols such as dusty marine, dust, or polluted continental? It could be acceptable for remote ocean but may wrong for coastal regions. How can you justify this?

line 130-175:  Why did you select VFM_max+2km as a top height for the retrieval? Is it only because Painemal et al. (2019) showed best agreement with HSRL for that criteria? What did they compare here in Painemal et al. (2019)? Coulmn AOD from CALIPSO-SODA and HSRL AODs from surface to VFM_max+2km? It should be very careful here because you retrieve only below VFM_max+2km, but CALIPSO-SODA is a coulmn integrated AOD from surface to TOA. You ignored aerosols above the VFM_max+2km (line 157). It could be negligible but you showed the discrepancy up to ~70% (line 154) which is quite large. Based on this results, it looks better to select higher altitude as an upper limit for the retrieval. Authors should expain more acceptable reasons for selecting VFM_max+2km as an upper limit for the retrieval and specify resulting unceartainty in the retrived lidar ratio. Authours mentioned uncertainty of the retrived lidar ratios due to Stratospheric AOD (line 173), but tropospheric aerosols above VFM_max+2km is much more important to discuss.

line 209: 2L assumption for polluted dust and elevated smoke is acceptable. however, dust aerosols may frequently exist near the sea surface, especially near the continents. Have you checked this?

line 236: better not to conclude like this. present uncertainties of each data here.

line 258-260: "misclassification of ~ tenuous aerosol layers that are not detected by the CALIPSO algorithm" This may be a reason why the authors should select higher upper boundary for the retrieval instead of VFM_max+2km.

The authors assess the CALIPSO Version 4.2 aerosol typing and assigned lidar ratios over ocean. To do that, they use (i) the aerosol optical depth (AOD) retrievals from the Synergized Optical Depth of Aerosols (SODA) algorithm and (ii) retrieved columnar lidar ratio estimated by combining SODA AOD and CALIPSO attenuated backscatter (CALIPSO-SODA). This paper is of good quality, well written and structured. It will be worthy of publication once the issues below are addressed.

Major comments

We recommend that the authors:

. first compare CALIOP-SODA and CALIOP V4 AOD as well as lidar ratios for all CALIOP V4 aerosol types before they classify their analysis by CALIOP V4 aerosol types. As they know, CALIOP V4 is likely to miss tenuous aerosols or misclassify aerosols. Starting by classifying CALIOP-SODA lidar ratios per (likely misclassified) CALIOP V4 aerosol type is confusing and slightly circular.

. clarify why CALIOP does not consider the possibility of polluted dust within the PBL

. clarify their filtering method and technique as a bullet list or a table (includes 1L vs 2L techniques, cloud masking, altitude selection etc)

. clearly suggest which CALIOP product is accurate, which one is not. And proposes fixes to the algorithm or additional filters to be applied by the users moving forward

Detailed comments:

. line 14 – “This implies that the CALIPSO classification scheme generally categorizes aerosols correctly” is too strong of a statement. Please consider watering it down.

. line 21 – “value”

. abstract – the authors point out issues with CALIOP V4 polluted continental/ smoke, dust, dusty marine and clean marine aerosol types. They should suggest some future fixes and possible filtering

. line 105 – why not talk about HSRL instead of saying “airborne lidar observations”?

. Figure 1 – can you briefly specify how reliable is GEOS-5 ML?

. line 226 – I don’t understand the sentence. Please rephrase.

. line 240 or eq. (1) – explain i=1 to N

. line 249 – largest mean AOD

. line 221 – consider replacing “modest” by low and then “particularly low” instead of “negligible”

. Figure 5 & 6 – consider saying “density plot” instead of histograms

. Figure 7 – add (a), (b) and (c) on graphs

. These Figures and Tables are redundant, consider consolidating/ simplifying – Fig 9 and Table 4 as well as Fig. 7-8 and Table 3

. Line 275 to 277 – aerosol variability depends on the environment. Please refer to e.g., Shinozuka and Redemann (2011).

. Line 305 – there is a repeat in the sentence. Please rephrase

. Line 303 to 307 – can we make sure to say that the lidar ratio for a specific aerosol type should be independent of the AOD? The fact that it varies with AOD points to some issues in the CALIOP algorithm and I suggest describing them.

. Line 432 – assumed by the CALIPSO

. first bullet in the conclusion – consider explaining the reasons and implications

. Line 443 – within +-10sr of those

. Line 460 – is a repeat of the sentence above

Shinozuka, Y. and Redemann, J.: Horizontal variability of aerosol optical depth observed during the ARCTAS airborne experiment, Atmos. Chem. Phys., 11, 8489–8495, https://doi.org/10.5194/acp-11-8489-2011, 2011.