Dear Fernando Chouza et al.,



Thank you for this well written and clear paper. It was easy to read, logically laid out, and showed clear results.  The ozone comparisons look very good. It is nice to see relatively low cost tropospheric ozone lidars produce such high quality data sets.  I only have a few minor criticisms:

i) The technical information on the system and data processing could be more detailed. The paper is submitted to AMT so technical details are in the scope of the journal and of interest to likely readers.  Summary tables for technical specs of the transmitter, receiver, and counting electronics in an Appendix could be helpful for the reader.

ii) It could also benefit from a bit more discussion as to why you have chosen various hardware/optical components and some system efficiency calculations. 

iii) Introduction of scientific motivations could be stronger and more explicit.  Why make these lidar measurements of tropospheric ozone above 250m?  Is the objective that these instruments provide real time data to a weather forecast or reanalysis (i.e. ECMWF) or are these instruments only useful for CalVal of satellites/aircraft? Explicitly citing the work of a few other small network deployable lidars for water vapour, wind and aerosol could also be good.



Comments/Questions (please address, answer, or ignore as you see fit.  These points are possibly important but shouldn't hold up publication if they take a long time to address.):

Section 2.1 Transmitter.

What are the other laser parameters? A measured histogram of laser pulse frequency over 30 minutes would give you an idea of what the stability of your laser is during a single ozone profile. 

Does your emission frequency drift over the day due to day/night heating of the instrument?  Or due to the air conditioner turning on and off?



How broad is your emission line before and after the gas cells?  How does this compare to the target lines for O3 and the offline?  Coupled with the histogram above, you can estimate how efficiently your laser pulse is illuminating the target O3 line(s) compared to the offlines.   



How is the collimation ensured? How well is this known?  Is the beam shape OK after the gas cells?  Is this why a wide fov is used in the receiver section?



typo L94:  "transmitted through..."



L97: Nominal Ocular Hazard Distance not defined



Section 2.2 Receiver.

Is there a benefit to choosing a smaller NA on the fibres to reduce the field of view further and reduce the influence of the solar background even more, or is the lidar already sufficiently solar blind?   Some diagnostic plots could be helpful. Later in the paper, comments are made about complications with background estimations.  

What is your estimated beam spot size inside your field of view?  How much can the beam jitter and drift inside the fov?   Do you have stabilization for day/night observations to account for thermal expansion of the lidar unit and slow diurnal misalignments?  What about daytime turbulence corrections?  Local convection can jitter your beam a lot inside the fov on the timescales of 1 second to minutes, reducing signal stability.   Where you place the SMOL could also be important for the signal stability (grassy field vs. concrete building roof, ground level, vs. elevated platform etc)



L110:  Is FWHM of 32 nm a typo?  



If you are throwing away half your signal with the intensity beam splitters, why not just use narrower filters?  Even if the transmission efficiency of a cheaper deep UV filter is low (~30%), isn't it better to have a 30% intensity signal at a well known, narrow line than a 50% signal over a broad wavelength range?



Saturation problems.  It would be good to have more details.  This is a bit vague and the reader can't tell how impactful the problem is in the current design.  A signal plot could be helpful.  Is there a benefit to increasing the rep rate of your laser from 20Hz and increasing the speed of acquisitions?  

What are the technical specification of the MCS and counters?  Dark count rate? Quantum efficiency? Counter clock rate?  Data write rate?



Section 2.3 Automation



If the alignment is at a set time, how do you account for variable sky conditions?  For example, a patchy cirrus moving overhead could lead to a variable maximum back scatter due to cloud structure and result in a poor alignment.



Running the piezios every second could give you real-time beam stabilization to compensate for daytime turbulence.  Disabled during cloud detection.



Section 3: Data processing



Non-linear backgrounds are an unfortunate reality for some historical lidar datasets.  But this should be corrected in hardware for newly designed systems.  Showing some plots of lidar returns at different signal intensities, could show the linear range for SMOL1 and 2.



Is vertical resolution in GLASS achieved through integration or filtering.  If filtering, what filter?



4.1.1 Comparison with airborne measurements

It seemed to me that there were more technical specifications and error tolerances given for the HSRL than the SMOL system.  This comes back to my main point that more details of the new instrument would be appreciated.



Figure 6. could include the 2 sigma limit on the mean. Seeing the measured ozone profiles might also be nice, perhaps a 2-panel plot.



Comparison with a reanalysis or forecast (ECMWF) seems relatively straight forward to include, and could provide more motivation for the importance of high resolution observations to capture small-scale features at a local level.  

General comments

A well-structured and well written paper about an inspiring new implementation of a tropospheric ozone differential absorption lidar system. The methodology is well known from previous implementations of the technique in other instruments, but the noteworthy achievement is in the small and affordable package.

The system itself is well described, concise but clear, and several interesting results are shown from field deployment of three lidar systems, among which two are the newly built instruments. Furthermore, the lidar observations, potentially having to deal with similar methodological issues, are also favourably compared to independent airborne and ground based observations of well known characteristics.

The conclusions are valid and corroborated by the presented material.

Specific comments

• Abstract line 16: change ‘physical’ to ‘physical and chemical’.
• Abstract line 23: it says three ozone DIAL pairs, but the listing is ambiguous. Please clarify. One DIAL pair is applied twice, but perhaps this is better left to the descriptions later
• Introduction line 60: what is the expected bias? Is that specified in TOLNET, before and after corrections for e.g. aerosol interference? Please add some text to explain this.
• Instrument description line 90: How is the output energy for 266 nm optimised and stabilised and what is the (long term) stability without adjustment of the fourth harmonic crystal? Is the laser output power monitored and can it be optimised remotely?
• Instrument description line 97: what is NOHD?
• Instrument description line 117: potential future updates might be moved to a summarising section, or the conclusions.
• Instrument description line 128: which parameters are recorded needed for the retrieval?
• Instrument description Figure 2: in the figure the wavelengths at output and detection are missing. Please add.
• Instrument description Figure 2: a second figure will be helpful to indicate various control sub-system units as described in Sec.2.3
• Data processing Line 183: use ‘… specified optical thickness’.
• References: perhaps a few older references can be selected to indicate earlier attempts for routine monitoring of tropospheric ozone profiles, in particular from the European Eurotrac TESLAS and TOR programmes. E.g.
  • https://doi.org/10.5194/amt-13-6357-2020
  • TESLAS: Tropospheric Environmental Studies by Laser Sounding (TESLAS), in: Transport and Chemical Transformation of Pollutants in the Troposphere, Vol. 8, Instrument Development for Atmospheric Research and Monitoring, edited by: Bösenberg, J.,Brassington, D., and Simon, P. C., Springer (Berlin, Heidelberg, New York), ISBN 3-540-62516-X, 1–203, 1997.