The paper provides a technical overview of the eVe lidar instrument with a clear and comprehensive description of the opto-mechanical components and blocks as well as a comprehensive description of the detection setup of the two WSU. One interesting aspect of the paper is the in depth description of the pre-processing and processing chains used to retrieve the lidar products – even if I have the feeling that this part could be improved, especially for the error calculation. The next section of the manuscript describes an intercomparison between synthetic signals and retrieved signals to assess the performance of the eVe inversion algorithm. The overall feeling is that this section needs to be revised since the purpose of the analysis is not clearly mentioned. On one hand, the section specifies that the synthetic profiles are used to test the inversion algorithm (this is clear) but on the other hand the synthetic coefficients are compared to so-called retrieved coefficients. The authors should specify why this comparison is relevant and what the retrieved coefficients are.

Next, the paper shows two case studies: one with low aerosol load and low depolarizing particles and a second with higher amount of depolarizing particles. These sections are well described but some details are missing and some uncertainty values must be explained (please see comments in the attached pdf).

The scientific significance makes the manuscript suited for publication but the content needs revision for the final submission.

Author’s reply:

We thank the reviewer for his/her constructive comments and suggestions. His/her comments helped us improve further the manuscript in terms of completement and also provide more details and/or clarifications where needed in order to be easily accessible for the readers.

Profiles of temperature, pressure, backscatter and extinction coefficients have been used as an input in a lidar signal simulator in order to produce synthetic signals corresponding to the signals that a lidar would measure under the given atmospheric conditions. These profiles are the simulated ones and they are provided in figure 2 in Pappalardo et al., (2004).

The simulated lidar signals have been used as input in the developed eVe software in order to retrieve the particle backscatter and extinction coefficients. The retrieved profiles of backscatter and extinction are compared with the simulated ones (that were used for the simulation of the lidar signals) for testing the performance (goodness of retrieval) of the developed software. The same algorithm intercomparison procedure has been also performed for other developed software (see Böckmann et al., 2004; Pappalardo et al., 2004)

Considering the fact that this section (Algorithm intercomparison) may be not clear for the other readers, too, further details have been added in the revised manuscript.

All the comments that are provided in the supplement of RC1 have been addressed and corresponding revisions and insertions have been made in the manuscript. The point-by-point responses are provided below each RC1’s comment along with the revised manuscript with track changes (attached pdf).

References

Böckmann, C., Wandinger, U., Ansmann, A., Bösenberg, J., Amiridis, V., Boselli, A., Delaval, A., Tomasi, F. De, Frioud, M., Grigorov, I. V., Hågård, A., Horvat, M., Iarlori, M., Komguem, L., Kreipl, S., Larchevêque, G., Matthias, V., Papayannis, A., Pappalardo, G., Rocadenbosch, F., Rodrigues, J. A., Schneider, J., Shcherbakov, V. and Wiegner, M.: Aerosol lidar intercomparison in the framework of the EARLINET project. 2.Aerosol backscatter algorithms, Appl. Opt., 43(4), 977–989, doi:10.1364/AO.43.000977, 2004.

Pappalardo, G., Amodeo, A., Apituley, A., Comeron, A., Freudenthaler, V., Linné, H., Ansmann, A., Bösenberg, J., D'Amico, G., Mattis, I., Mona, L., Wandinger, U., Amiridis, V., Alados-Arboledas, L., Nicolae, D. and Wiegner, M.: EARLINET: towards an advanced sustainable European aerosol lidar network, Atmos. Meas. Tech., 7(8), 2389–2409, doi:10.5194/amt-7-2389-2014, 2014.

Authors statement

We would like to thank the reviewer for his constructive comments. We acknowledge that addressing the comments and suggestions of the reviewer the manuscript has been considerably improved. We provide below the replies to the general comments of the reviewer.

All the comments that are provided in the RC2-supplement have been addressed and corresponding revisions and insertions have been made in the manuscript. The point-by-point responses are provided below each RC2’s comment along with the revised manuscript with track changes (attached pdf).

Reply on RC2 (Ronny Engelmann)

The manuscript "The eVe reference polarization lidar system for Cal/Val of Aeolus L2A product" by Peristera Paschou, et al. describes a new Raman Polarization lidar system that is capable of measuring atmospheric linear and circular depolarization, as well as backscatter and extinction at 355 nm wavelength.

The system was constructed by Raymetrics in cooperation with the University of Munich and the National Observatory of Athens.

The manuscript is well written, the outline is appropriate for such a technical description, the content is clear, the figures are of good quality, and the English language is of high quality except for a few minor points.

I recommend considering this manuscript for publication in ACP after some minor revisions have been considered. Most comments are given in the supplement pdf. However, the most important points that should be addressed are given next:

1.) As this is a descriptive paper for a new lidar system, it would be good to specify the most important optical elements as well as possible. If the policy of Raymetrics allows, it would be good to mention a bit more in detail manufacturers, and specs of e.g., telescope manufacturer & type, the collimation lens focal length, polarizing beam splitter types, sheet polarizer types, as well as interference filter manufacturer and specs, eyepieces, and so on. It has been seen from other papers in the past, that such a publication is the first point to look for such details during future work with the system.

Author’s reply:

Complying with the policy of Raymetrics, additional details about the manufacturers and the specifications of the optical elements that are deployed in the system have been provided in the revised manuscript. For more details, kindly read the author’s replies on the dedicated comments of RC2 in page 5 of the RC2-supplement.

2.) In Fig.5 a lot of optical elements are described by Müller matrices. This is nicely done and quite clear. But it should be mentioned, why those matrices are given and what they are used for. Do you know the values of these matrices? Can you specify them in the Appendix for future reference? Otherwise, it is not clear, why those matrices are given here.

Author’s reply:

The Müller matrices are used to model the changes in the polarisation state of the emitted light when the light interacts with the atmosphere or passes through the optical elements of the system. Thus, the expression of the polarisation lidar signals with the Stokes-Müller formalism facilitates the derivation of the equations that are applied in the polarisation calibration factor calculation techniques. More details have been added in the discussion about the use of the Müller-Stokes formalism in the revised manuscript.

We prefer to keep the definition of these matrices in the manuscript because it will be easier to reference the used matrices in future publications. More specifically, a follow up paper is planned where these matrices and their values will be extensively analysed. The follow up paper will provide an extended description on how each lidar setup is handled for calibration purposes, on the applied techniques for aligning the polarisation plane of the emission and the optical parts with respect to the reference plane as well as on diagnosing unwanted polarising effects in the system.

3.) If possible, also give real values of the GH parameters for future reference.

Author’s reply:

The values of the GHK parameters that were used for the retrieval of the volume linear and circular depolarisation ratios on the presented cases have been provided in section 5 of the revised manuscript.

4.) It should be mentioned, why a new algorithm for the retrievals is needed. After all, many algorithms already exist. But nevertheless, it is good to compare to a standard example like the EARLINET case. However, this case shows extremely high AOD and large extinction values. In most real-life cases the extinction can be smaller by as much as a factor of 10. Such cases would be really more challenging for a comparison. My suggestion is to at least mention this fact about the EARLINET case. Furthermore, I would suggest combining the graphs and tables to shorten this part of the Algorithm development, as

this is scientifically speaking nothing new.

Author’s reply:

We developed an algorithm for eVe lidar to ecompass all the retrievals needed, including the traditional ones (i.e., particle backscatter coefficient, particle extinction coefficient, linear depolarisation ratio), but also the new ones that are not included elsewhere (e.g. circular depolarisation ratio). There is not plan at the moment for the eVe system to be integrated in ACTRIS or EARLINET. Thus, eVe will not have an automated retrieval from the standard procedures of those networks/infrastructures (such as the Single Calculus Chain for example). As such, we have developed the eVe algorithm and we tested it against EARLINET synthetic data, that have been used for all the algorithms tested within the network and are considered necessary. In addition, the retrieval of the circular depolarisation ratio from lidar measurements is not documented elsewhere and the existed algorithms (e.g., the Polly^NET software or the SCC software and other algorithms by individual stations within EARLINET) can have different workflows or apply different approaches on the processing of the lidar signals and the retrieval of the lidar products. Hence, we included a section (section 4) in the manuscript in order to discuss the workflow of the developed eVe software on the processing of the lidar signals and the retrieval of the lidar products, and also to present the basic equations that are used for the retrieval of the traditional (particle backscatter and extinction coefficients, linear depolarisation ratio, etc.) and the new (circular depolarisation ratio) lidar products.

Concerning the high AOD in the EARLINET case and following the reviewer’s suggestion, a dedicated comment has been added in the manuscript.

5.) Error estimates: In the case study the error of the VLDR is below 0.0005. That seems to be very small, considering that signal noise plays a role, calibration uncertainties, and uncertainties of G&H are always present. I would suggest revising these calculations. Also, the PLDR has a very small uncertainty in regions with almost no backscatter. Please recheck, from experience, the errors should be much larger in Fig 10, right, above 5km height.

Furthermore, the uncertainties of the extinction coefficient seem to be very small as well. How is the extinction calculated? Usually, the derivative is calculated by a linear fit over a certain height range. Was the error of this fit propagated towards the extinction coefficient? Please denote the fit(smoothing) length. Can you also present the corresponding lidar ratio profile? Calculating the lidar ratio by eye leads to very small values between 10 and 20 in lower altitudes. Is this reasonable?

Author’s reply:

The uncertainties were erroneously provided as rounded up to the 3rd decimal place. Additionally, only the signals noise was taken into consideration for the estimation of the uncertainties in the "error estimation" module of the software resulting to lower uncertainties. During the revision phase the "error estimation" module has been revised (improving the parametrisation in the error simulation of the analogue signals) and additional error sources as indicated by the reviewer (polarisation calibration factor, GHK parameters) have been integrated in the software. The estimated uncertainties that are provided in the revised manuscript are larger and more reasonable than the ones provided in the initial submission.

In addition, during the revision phase, the retrieval of the optical properties was further optimised (better selection of the user defined parameters that are required in the Raman inversion method such as the scattering ratio in the Rayleigh reference height and the window length of the derivative for the extinction retrieval). The lidar ratio profile has been also provided along with the rest optical products. After the retrieval optimisation, the lidar ratio has more reasonable values (mean value of 37 sr on 24 September and 20 sr on 29 September).

Regarding the derivative in the extinction coefficient retrieval, it is calculated by a linear fit and the corresponding error of this fit is propagated towards the retrieval. In the initially submitted manuscript the extinction coefficient was retrieved using a monotonically increasing (with linear interpolation) window length for the derivative with increasing of the height range resulting to lower than expected statistical uncertainties as commented by the reviewer. In the revision phase, the derivative was calculated using different windows over certain signal height ranges (the approach of applying the same fixed window length in all height range is not preferred since it will result to a noisier extinction profile). The details about the applied fitting windows have been provided in section 5 in the revised manuscript. More specifically, in the first signal range node (up to 1.5 km) the derivative window is 200 m, in the second signal range node (from 1.5 to 4 km) the derivative window is 400 m, in the third signal range node (from 4 to 6 km) the derivative window is 600 m, and finally in the fourth signal range node (from 6 km to the end of signal) the derivative window is 800 m.

6.) Can you compare the measured Rayleigh volume depolarization (in particle-free air) to the theoretically determined depolarization (according to the filter bandwidth)?

Author’s reply:

For the dates of the selected cases, the temperature ranges from -10 oC to 20 oC within the altitude heights up to 5.5 km resulting to a mean molecular linear depolarisation ratio (mLDR) of 0.00586 ± 0.00004 with the minimum and maximum mLDR values within this height range to be 0.0058 and 0.00592 respectively. The mean molecular circular depolarisation ratio (mCDR) is 0.0119 ± 0.00009 with minimum and maximum value of 0.0118 and 0.012. For the comparison of the theoretically determined molecular depolarisation ratios with the retrieved volume depolarisation ratios in aerosol-free region, the mean mLDR and mCDR values have been indicated with a grey dashed line in the revised plots of the volume depolarisation ratio profiles. Additionally, related changes have been made in line 469 of the initially submitted manuscript.

7.) As the first case study shows a measurement with very low depolarization I would suspect the second case presents quite the opposite. While the second case has a slightly higher depolarization I wonder: is it not possible to find a real dust-dominated case? Then you could show the real performance of the system and how the Aeolus-like backscatter significantly differs from the "total" backscatter. I think such a case (instead of case #2) would improve the manuscript quite a lot.

Author’s reply:

This study is pending since eVe measurements are currently being collected on the tropical islands of Cape Verde where dust cases with high AOD values are observed. Our efforts will be finalized until summer 2022 when the ASKOS campaign will end. We aim to analyse the collected eVe dataset from Cape Verde and present a study related to the reviewer’s suggestions on an upcoming paper.

8.) Did you perform any of the EARLINET QA tests? Telecover, Rayleigh-fit, etc. Is it possible to include any of these results?

Author’s reply:

Yes, we perform QA tests proposed by the EARLINET network on a regular basis in order to monitor and evaluate the lidar performance. We have added Appendix B in the revised manuscript presenting a Rayleigh fit test and a Telecover test, both performed close to the presented cases.