This paper deals with a statistical analysis of aerosol higroscopicity determined with the combination of multiwavelength measurements and radiosondes. The topic is of great interest to advance in the water vapor uptake by atmospheric aerosols particles and thus advance in understanding the role of aerosols as potentials cloud condensation nuclei. Of particular interest is the analysis of aerosol hygroscopicity below clouds and for clean skies. Therefore, the study topic is of interest for Atmospheric Chemistry and Physics goals.

To my knowledge, this is one of the first papers that use a long-term database for statistical analyses which implicitly is a novelty. The database of multiwavelength Raman and micropulse lidars plus radiosondes is unique. The paper is well structured and references are updated. Results are unique.  The paper has the potential of becoming a reference study for future statistical analyses that involved more ACTRIS stations and similar infrastructures worldwide. However, because of the expectative, I have some concerns that must be addressed before its final publication in Atmospheric Chemistry and Physics.

• Although there are no needs for given specific details about your lidar systems, I would like to know if your Raman lidar measurements used are also for daytime. If so, how does it affect the temporal resolution? Also, it is needed to specify which type of radiosondes  used and if data are publicly available.
• The methodology proposed for determining aerosol higroscopicity parameters from lidars measurements is not new. The authors give some recent references but I would go the originals papers that explain why the specific thermodynamic conditions must be fulfilled and give a short overview in the paper. But what is not clear to me is the novelty introduced by the authors in using RHref. Moreover, I miss an error analysis in f(RH) and γ because determining their uncertainty is essential for a discussion of similitudes and differences between seasons and between different sites.
• The authors make a great discussion comparing their results with other stations. But they refer many times to chemical properties of aerosols in Barcelona. To me, that should be accompanied with data. I understand that most of your chemical data are from ground-measurements and lidar measurements refer to different altitudes. But having a study case with well-mixed conditions and chemical properties from the surface would help the discussion.
• I get lost with the statistical comparison. The databases for spectral analyses and climatological analyses seem different. That needs a clarification

MINOR COMMENTS

• The time period when data were acquired is not mentioned. That must be clarified
• Line 46: The statement that aerosol higroscopicity depends on chemical composition need reference. The same for the discussion of deliquescence and cristalization.
• Line 60: Reference needed for measurements of higroscopicity with tandem nephelometers
• Line 76: I agreed that multispectral lidar measurements are fundamental for aerosol hygroscopic growth study. But this technique has limitations. Please clarify.
• Line 97: Details or reference are needed for your algorithm for MPLNET data
• Lines 140:144: Did you carry out backward-trajectories analyses for stdudy each case?
• Line 167: I do not understand the statement ‘ it is the first time a conversion fmin to fref is proposed
• Figure 3: Are the data those of Figure 1?
• Line 229: I do not understand the expression ‘en passant’. Generally I would recommend double-checking English grammar. 
• Line 277: Do you expect marine aerosols above 2 km? Please clarify
• Line 313: Please clarify the statement ‘our findings suggest that hygroscopic layers near the top or slightly above the PBL’.
• Table 3: Why the analyses are only done at 532 nm?
• Lines 355-356: Why as the aerosol size grows, its potential to keep growing are reduced compared to a drier aerosol?

Manuscript analyzes hygroscopic growth parameter based on collocated radiosonde and multiwavelength lidar measurements. Hygroscopic growth parameter and enhancements factors are derived at three laser wavelengths for different types of aerosols and seasons. Paper is clearly and well written and authors provide very complete review of previous studies in this field. The results presented are new and original and manuscript is appropriate for publishing.

I have just several technical comments.

            Authors consider two parameters: hygroscopic growth parameter and enhancements factor, which are recalculated to each other via a simplified scheme. May be it is worth to mention, that these parameters are not really independent.

Simplified parameterization may not work for deliquescent aerosols and sea salt particles are often occur in cases presented. Probably, corresponding comment id needed.

Are backscattering coefficients calculated by Raman or Klett method? For Klett method lidar ratio may vary with RH.

It is a pity, that depolarization measurements are not presented. These could help in particle identification. Besides dependence of depolarization on RH is also interesting. Hope the authors will include it in their future studies.

Ln  46.  “Atmospheric aerosols can be classified as hydrophobic (e.g. dust) or hydrophilic with monotonic (smoothly varying, e.g. volcanic) or deliquescent (step change, e.g. marine) growth (Carrico et al., 2003)”. I think “growth” should be changed for something else. “Types”?

Ln 310. “Hygroscopic layers”. I am confused, because probably authors mean layers where hygroscopic growth was calculated? Because most part of aerosol is hygroscopic everywhere…

Fig.8a. Hygroscopic growth is calculated for low RH (below 55%) and then extrapolated to RH=85%. Uncertainty can be high.

Ln.439. “This behavior could 440 be attributed to the aerosol size: the smaller the aerosol, the more hygroscopic”. Can it be explained? For spectral dependence, it is explained by the presence of the coarse mode particles. But if we have only fine particles, how hygroscopicity depends on size?