This is an excellent overview of the CALIPSO CALIOP lidar and very useful information on the global distribution of high clouds.

The authors make a laudable contribution to global ice cloud measurements through a useful and well-documented data product with valuable comparisons to peer data products. Especially useful for users are formulae and examples of how to process the product's vertically-resolved histograms into more commonly used quantities. Below, I highlight several points where further clarification could be useful.

Minor comments:

• Lines 221-223: The authors write "Comparison of panels (e) and (f) shows the most significant impact of applying quality filters is the exclusion of bins deep within opaque cloud layers where the overlying optical depth exceeds 2, such as near latitudes 19.0◦ N and 11.0◦ N." This sentence is confusing since it refers to overall product behavior anecdotally via the example of a single granule -- the "exclusion of bins deep within opaque layers" is a general statement about the product while "near latitudes 19°N and 11°N" is a particular statement about the example.
• Lines 263-265: The authors write "In L2-CPro, the AVD is reported at 60 m vertical resolution between 8.2 km and 20.2 km but reported at 30 m vertical resolution below 8.2 km, while the vertical resolution of L3-Ice is 60 m at all altitudes." However, the reported resolution of L3-Ice is 120 m as documented elsewhere in the text. Perhaps the authors could clarify how this 60 m resolution appears to be an intermediate aggregating resolution rather than a final output resolution.
• Figures 7 & 8: I am curious about the small but noticeable elevated IWC and extinction at 4 km visible at most latitudes, appearing as a horizontal band, which is also noticeable in the percent of ice samples rejected. It looks as if it might be an effect of either the CPro retrieval or QC filtering, but I did not see it discussed in the text.
• Line 403-407: I found the impact of cloud-clearing on scattering/extinction fields somewhat vague. Are the authors stating that, for intermittently cloudy layers at 5 km resolution with <4 km top height, the resulting cloudy extinction/IWC used by L3-ICE ignores scattering from single-shot-detectable clouds?
• Section 6: In the comparison between L3-ICE and the combined radar-lidar products, many differences are stated to be "significant," but I do not see any statistical significance testing in this section.
• Radar-lidar vs. lidar-only comparison: The authors provide an invaluable comparison to similar products but provide relatively little discussion of large differences in ice cloud frequency in the region where one would expect the greatest agreement (>15 km/lidar-only). One product agrees well with L3-ICE, while another (DARDAR) shows far fewer ice clouds (Fig. 20a). Do the authors expect L3-ICE ice cloud occurrence to be more accurate for high clouds than the combined products, or do the authors think definitional differences could explain such a large discrepancy?

Text/figure corrections:

• Figure 5: caption reads "Level 3Tropospheric" (missing a space)
• Figure 7: units of IWC and ice extinction are not specified in either the figure or the caption.
• Figure 11: Meaning of grey pixels in both panels is not stated.
• Lines 287-298: "all-sky IWC (Figure 7 (h) and (j) shows a rainbow-shaped maximum" -- missing ")"
• Line 386: "when averaged using a 10◦×10◦ (red))" -- extra ")"
• Line 408: "impacts L2 CPro ice cloud extinction" -- elsewhere it's spelled as "L2-CPro"
• Line 504: "L3-ICE reports very little IWC greater than 0.1 gm−3 while DARDAR and 2C-ICE report a significant number of larger values contribute to an IWC which is 3 to 5 times higher than from L3-ICE." -- grammatically confusing
• Line 562: the CloudSat data center is at CIRA (CSU/NOAA), not CIRES (CU Boulder/NOAA)