The manuscript is well-written and presents the information in a clear and straightforward manner. I am concerned, though, that the new method for filtering DBS scans presented here is behind the state-of-the-art. For instance, Steinheuer 2022 (see below) introduces a method to filter DBS and VAD scans and adaptively reject low-quality measurements. An added benefit is that they use radial wind speed only, avoiding issues associated with fixing an SNR cut-off (which should vary with lidar settings). I would like to see more thought given to other methods such as the one they propose, or direct comparisons against results employing other methods. Or, if other methods are inappropriate for use on the lidar hardware itself given hardware limitations, that could be noted as well. That is, the proposed method might be preferred for running onboard the lidar given its simplicity. In any case, additional justification is needed for the current method compared to other methods.

J. Steinheuer, C. Detring, F. Beyrich, U. Löhnert, P. Friederichs, and S. Fiedler, “A new scanning scheme and flexible retrieval for mean winds and gusts from Doppler lidar measurements,” Atmospheric Measurement Techniques, vol. 15, no. 10, pp. 3243–3260, May 2022, doi: 10.5194/amt-15-3243-2022.

General comments:

The authors provide a concise summary of a comparison of two 10-min average wind retrieval algorithms specific to the BEAM 6x lidar from Lumibird. The newly proposed retrieval includes individual DBS retrievals from scans where only a subset of the six beams contains data with backscatter signals with signal-to-noise ratios above a critical level at the relevant range gates corresponding to a wind retrieval height. The article seems directed to justify the implementation of this algorithm for a commercial product, and some additional information could be included to make the article academically more relevant.

Specific comments:

The algorithm is specifically designed for 10-min mean winds. These average retrievals will ultimately include both, individual retrievals from a full, successfully retrieved set of six radial velocities, and a subset of now included DBS retrievals based on fewer radial velocities.  While the data shown (10-min averages) may suffice to market a machine using this updated algorithm, a reader may also be interested in a comparison of the retrievals from sets of a) the full six radial velocities, b) 5 valid velocities, c) 4 , etc. Adding this will allow the reader to evaluate the influence of these added members of the average, and the data shown in Fig 7 becomes more valuable from a more academic standpoint. Similarly, there may have been results published of experiments where subsets radial velocities were omitted from DBS and/or VAD retrievals. If so, these should be cited in this context, and differences in results should be discussed.

The description of the BEAM 6x lidar (Line100ff) should be moved to an earlier section as this information is needed prior to the “Methodology” section as the proposed retrieval is so some extend specific to that (or similar) lidar model. For example, you mention 10-min averages on line 71, but the reader does not know how many individual retrievals are contained in such an average until this information is given (i.e. 5.5-s scan repetition).

There are several aspects that could be clarified: The “sliding window” (Line 53), does it slide across the 2.6 m range gates of each of the six beams, or does this mean that winds are retrieved at vertical increments (i.e. heights) at these intervals?

It is stated that the addition of lidars increases the wind retrieval height (i.e. to 1000 m; line 97), but the range gates are listed as 2.6 m. How does adding lidars lead to extended ranges? Are they run with different gate lengths, and if so, which are chosen?

Several formulations could be more precise to adhere to scientific journal standards. For example, on line 27, you state: “doppler velocities”… fall “below”…”SNR threshold”. Isn’t it the backscatter that is below the SNR? See also line 51 states a bit colloquially:“beams fall below SNR”. In the paragraph starting with line 49, it is omitted that the described method is applied to each range gate. I’d like to encourage the authors to scan the manuscript for similar, slightly imprecise, formulations, and to improve them.

Technical corrections:

Line 124: add space:“analgorithm” -> “an algorithm”