Please refer to the pdf file in the supplement

Summary: The authors report on measurements of biomass burning aerosol (BBA) layers in the free troposphere and the lowermost stratosphere using a combination of multiparameter lidars, one of which was supplemented by an additional discrete fluorescence detection channel. The first part of the manuscript describes the experimental work and details the method used to calibrate the fluorescence signal. Individual measurement cases are then discussed, ranging from optically thick BBA layers in the middle troposphere to optically very thin BBA filaments in the UTLS region. The results presented are (1) that fluorescence is a better indicator for aerosol layers than elastic backscattering due to the lower noise floor, (2) that a distinction between aerosol and cloud particles is possible because the latter do not fluoresce, and (3) that fluorescence measurements can be used to prove that smoke can initiate the formation of cirrus clouds.

General comments: The study provides interesting results, even if some points have already been addressed in previous research. The manuscript is well written, notwithstanding the speculative nature of some of the discussion. After corrections, the paper can be published.

Specific comments:

l. 51: This is not correct and has to be changed. Since the 2010s, both the approach with a discrete channel and the spectrometric approach for measuring aerosol fluorescence with lidar have been pursued side by side. Actually, spectrometric fluorescence measurements predate those with discrete channels. So credit should be given here to Sugimoto et al. (2012) and Reichardt (2014). Both publications have established the methodology of fluorescence measurements and have already identified the most important applications of fluorescence measurements in aerosol and cloud research. For example, the spectral fluorescence capacity (or efficiency) was established as a measurement parameter. Furthermore, Reichardt (2014) pointed out that fluorescence measurements would enable aerosol typing and would be of great importance in the study of aerosol-cloud interaction.

l. 61: ‘practical’ is probably the wrong word, because fluorescence spectrometers are actually quite easy to implement and calibrate. In fact, calibration of discrete fluorescence channels is more tedious and error-prone. So, probably, ‘reduced the concept to a simpler approach’, or ‘cheaper’ and ‘more cost-effective’ might be better choices of words.

l. 180 ff.: This definition of fluorescence capacity is problematic because it does not account for the actual bandwidth of the fluorescence detection channel, and thus results from lidars with different optical setups cannot be compared effortlessly. The better way is to use the spectral fluorescence capacity instead. Further, why is the particle backscatter coefficient at 532 nm used and not the particle backscatter coefficient at 355 nm? Every lidar that measures fluorescence around 466 nm in the free troposphere deploys the THG wavelength of a Nd:YAG laser but not necessarily the SHG wavelength. The authors should therefore change the definition and rewrite the manuscript accordingly.

l. 229: The statement that the values corroborate the findings by Reichardt et al. (2018) is not correct: The latter research group reported spectral (!) fluorescence capacities of Saharan dust (!), so the results cannot be compared.

l. 245: One ‘already’ has to be discarded.

l. 271-278. This is highly speculative and cannot be discussed on the basis of a single measurement. Instead, a larger data base must be analyzed statistically.

l. 278 ff.: This is a mistake that is commonly made. One cannot argue in terms of sphericity alone, size does matter as well. Small particles do not depolarize even if they are irregular.

l. 280-288: This is all highly speculative. The reviewer observes different correlations with relative humidity from measurement to measurement. Do the authors have dependable humidity data at hand? Otherwise it is recommended to cut the discussion short.

l. 288: ‘above at around’?

l. 345 ff.: This argument is wrong. (Spectral) fluorescence capacity is a mixed parameter (inelastic and elastic) and is thus only a meaningful aerosol parameter when clouds are absent. Cloud elastic backscattering suppresses capacity, no matter what happens to the fluorescence signal itself. If one wants to see indications of quenching, he has to study the fluorescence backscatter coefficient alone.

l. 350: At the end of this sentence the reference to Reichardt et al. (2018) can be made, not in the preceding sentence.

l. 381-396: Again, this is all highly speculative. Reduced fluorescence in clouds can always be either quenching or transport-related. In fact, proving aerosol-cloud interaction on the particle scale is highly demanding and probably only possible with spectrometric fluorescence lidars. The reviewer strongly suggests to skip this discussion and to leave it to mentioning the two possibilities (quenching, transport).

Figs. 2d, 3d, 4d, 6d: Some axes are in grey and not in black color.