Please see attached for my review comments.

The authors have presented a clear, well-written discussion of the characteristics of scaled vertical velocity profiles for different environmental conditions as measured by Doppler lidars at the Lindenberg site. However, their analysis depends on some assumptions that have not been fully tested and their classification method possibly obscures scientifically useful results.  Because of these issues, I believe that the paper needs major revisions before acceptance by ACP.

Major revisions:

The first issue pertains to the use of two different instruments for the analysis. The authors note that there is substantial agreement between the two Doppler lidars when they were co-deployed for three days, as quantified by the strong correlation of the observed vertical velocity at a single range gate. However, observations of the individual vertical velocities are less important than the derived values of w* and ML depth. With the different instrument subject to different SNR filtering, different ML calculations, different near-surface cutoffs, and different w* calculations, correlation between the derived quantities is not guaranteed.  The authors’ implicit claim that the observations between these two systems can be interchanged would be stronger if a comparison were done between the actual observations being used for the bulk of the analysis. While it is likely that there is a causal relationship between the different Bowen ratios and the different profile behaviors in the two analysis years,, the fact that different instruments with different processing pipelines were used in each of the two years cannot be overlooked as a potential source of the observed changes. A new Fig. 3 that showed w* and ML values instead of w values at a single height would help validate the claim that the two systems produce interchangeable observations.

The second significant issue is the classification of analysis days into clear, cloud-topped, and rainy. This sorting can obscure the actual processes at work: rain can come from convectively-driven clouds forced by PBL processes, or it can come from stratiform clouds with no real connection to the PBL; convective boundary layers might or might not form clouds depending on conditions both within and outside of the PBL. One reason some of the results show little to no dependence of the scaled w variance profiles on different environment types are occurring could be because different processes are being aggregated into the same bins while the same processes are being distributed into different ones. The authors should go into some detail about why these specific categories were chosen over more robust measures of PBL turbulence that could be quantified using other instruments at the Lindenberg site.

Minor issues:

Line 77:  How, specifically, were clouds identified? Is it a backscatter threshold from the Doppler lidar? What role did the ceilometer have in this?

Line 80:  Given that the two instruments were both Halo Streamline XR Doppler lidars, why were they treated differently in processing, with different noise filtering and other changes?

Line 98:  Related, why different ML calculations?

Line 105:  What is a “day” for the purpose of this analysis?  Is it a 24 h period? Sunrise to sunset (meaning it’s different for each day in the analysis)? A fixed period of time as shown in Figs. 4 and 5?

Line 107:  Can you explain the value of finding a characteristic scaled variance profile for cloudy days regardless of cloud type when certain cloud types are going to be closely coupled to PBL processes?

Line 131/Fig. 8:  Seeing as there are rather large overlaps between the latent heat fluxes in the three categories (implying that absolute moisture fluxes have little to no impact), why is there such a substantial difference by relative humidity? It seems that rather than relative humidity having an impact on the mean profile, the relative humidity is more an indicator that an environment is clear/cloudy/rainy. Is there a causal relationship between RH and the scaled profile? Is RH actually the best measure when these processes would seeminly depend on absolute moisture quantities?

Lines 169-170: Can you explain why this is happening?

Line 221:  This sentence seems to downplay the importance of the cloud-fraction analysis performed in this paper. Since soil moisture is important to these processes, it seems like a robust analysis of the scaled profiles would include soil moisture characteristics. Are there soil moisture observations at Lindenberg that could be used to illustrate the dependency of the profiles on soil moisture?

Line 223: SGP, not GPS