See supplement.

• l. 44: "significantly higher signal to noise ratio in the TTL" When such statements are made comparing the SNRs from BeCOOL and CALIPSO, it would be nice to clarify each time that the improvements are due to the accumulation time and proximity to the clouds. In absence of such precisions, a hasty reader could assume that BeCOOL has significantly higher SNR than CALIOP *in similar operating conditions*, which is probably not the case. These precisions are provided in the conclusion (l. 327) but they should also be added to the introduction.
• l. 56: daytime observations are not mentioned, which is surprising given the attention to the improved SNR. Could you clarify the reasons behind this?
• l. 56: "fiels" > "field"
• l. 57: What happens to the microlidars at the end of the flights? Are they lost?
• l. 82:"15% of the cloud optical depth lay above the new top altitude": Here I have failed to understand what has been done, and why. Could you clarify?
• table 2: "wavelenght"
• table 2: "CALIOP’s 1064nm channel has a low SNR…" I was not aware of that, could you provide a reference? (Also line 193)
• l. 128: "as been degraded"
• l. 127-129: why degrade CALIOP’s resolution horizontally, but not vertically? Averaging from 30 to 60m below 8.2km would get rid of the vertical step visible on Fig. 2. Also, if the goal is to display lidar curtains from both instruments at a comparable resolution, why not regrid BeCOOL's profiles vertically on a 60m resolution? I guess I don't understand very well the re-averaging choices made here.
• Fig. 2 and others: the vertical black bars make the BeCOOL results hard to decipher. They make the CALIOP results appear of higher signal quality, which should definitely not be the case. Low-signal areas are particularly difficult to evaluate, which is a pity. Could the black bars be made thinner here, as in Fig. 1? Or maybe set to white/grey?
• Sect. 3.1: As in this case the BeCOOL/CALIPSO colocation is perfect, I would be very curious to see a superposition of the profiles measured by both instruments at the same time and location. The scene being sampled (a high optically thin cirrus + more opaque clouds below + clear sky) is quite rich and would give a good idea of the signal performance that can be reached from both instruments in the same conditions. This is such a unique situation and exceptional measurements that in my opinion it is worth spending a little more time on it.
• l. 147: "thicken" > "thickens"
• l. 155: "above above"
• l. 155: why don’t you show the total column optical depth, instead of showing only the optical depth above 10km?
• l. 158: "both retrieval"
• Sect. 3.2: In the second case study, optical depths from both instruments agree very well (4 10-3 vs 3.8 10-3), in presence of an extremely optically thin cloud. How do you explain that the agreement appears much worse in the first case study (0.6 vs 0.4), which benefits from a much better colocation/coincidence and a more opaque, easier to detect cloud? Do you think you would get a much better agreement if you reprocessed the CALIOP data in the second case study, as you did with the first?
• l. 186: "We can attempt to estimate the horizontal extension of this UTTC assuming an horizontal extension…" could you please rephrase this? It reads as if you estimate the extension by assuming an extension. Also, what supports the assumption that the cirrus expands a few hundred kms in any direction?
• Table 4 and the following paragraph: These results make me wonder if the cirrus cover is only limited by the instrument’s sensitivity, i.e. if there is an optical depth continuum. Are there reasons to believe that an instrument with an even finer sensitivity to optically thin clouds wouldn’t report even higher number of cirrus cover than BeCOOL, and detect super-super-thin cirrus everywhere? If you plot the cirrus cloud cover as a function of the instrument’s optical depth sensitivity, would the cloud cover be 100% at the origin?
• I know this is quite tricky, but could you discuss a little bit if you think the optically very thin cirrus that are so frequent in the microlidar dataset are a generic/persistent feature of the regions you sample? Of other regions/periods? Previous studies such as Balmes and Fu 2018 suggest that similar optically-very-thin cirrus are found even in the extratropics and over land. Could you contrast your results with those? Doing so could help generalize your results to longer time periods and/or larger spatial scales.