General comments

The authors propose an interesting method of filtering away the adverse effects of rain on velocity estimation by sampling the Doppler spectra faster than the raindrops’ beam transit time. The method is validated with two continuous-wave Doppler lidar measurements. The paper is well-written. I recommend acceptance after minor revisions.

1.  Lidar #1 has a relatively bigger elevation angle of 57.9◦ compared to Lidar #3 of 15.3◦. Generally, the velocity difference between aerosols and raindrops appears in the vertical direction. Therefore, large elevation angles should suffer more influence from rain signals. While figure 17 exhibits the opposite results (Raw data with the red circle). The authors explain that the short probe length may contribute to it. I think adding a comparison experiment or detailed analysis will be better.
2. The proposed method is verified by continuous-wave Doppler lidar measurements. I’m also interested in whether it’s also suitable for a pulsed Doppler lidar which often uses a collimated beam. The author is advised to add related discussions.

Dear authors,

Please consider attached review comments. 

This paper proposed a method for filtering away the adverse influence of rain on wind velocity estimation for continuous wave lidar. It’s interesting to set the high sampling period (0.33 ms) shorter than the rain drops’ beam transit time (0.35 ms), which makes it possible to see the performance of rain on the lidar spectrum. The method is validated with sonic anemometers, which shows good agreement. The weak point is that all the samples listed in this paper were collected on the same day that containing limited rain intensities or rain events. I would recommend major revision by the suggestions and comments below.

Specific comments:

• Line 5 and Line 139: If this paper could provide more cases or results on several days with various rain intensities (containing light rain, moderate rain, and heavy rain) would make the conclusion more convincing.
• Line 100: How do the 0.35 ms of the raindrops’ beam transit time calculate? Please clarify.
• Line 158: “… where the line-of-sight speed is away from zero.” Please clarify and explain the reason for this processing.
• Line 169: This paper mentioned the rain spectrum with a high value of PSD and a narrow peak. However, considering the strong attenuation of laser energy caused by the raindrops, sometimes the PSD of rain spectra gets weak and has the nearly same magnitude as the aerosol spectrum. How to distinguish the wind and rain in these cases?
• Is this method proposed in this paper also suitable for pulsed Doppler lidar?
• This paper evaluates the performance of this method under several rain intensities. How about the influence of horizontal velocity on the results? Because a big raindrop will break up more small raindrops with high wind speed.

General comments

The authors investigate a method to improve the accuracy of wind speed measurements of continuous-wave Doppler lidars during rain. They show a reduction of the difference between wind measurements of lidar and sonic anemometer by processing the raw lidar data with high frequency in a different manner than usual. The proposed processing routine is very interesting as it can extend the application range of lidars.   

The manuscript is easy to read. However, there are some major points that should be addressed in the manuscript before it can be published.    

Major comments

The main critical point of this manuscript is that the conclusion was based only on the values of three minutes with (no) rain during one event. However, more minutes with rain are existing (Figure 16), but only a visual comparison was provided. It is essential to validate the proposed data processing procedure by considering more rain minutes (if possible also from other rain events on different dates with different intensities) to conclude the applicability of the method in various rain conditions.   

From the text, it is not easy to distinguish between the steps in the standard procedure of WindScanner data processing and the new proposed procedure. As this is the key point in this manuscript, a sketch showing the steps with and without rain-signal exclusion would be really helpful to understand the differences in the data processing.

The authors provide 0.35 ms as transit time of a raindrop through the lidar beam. Which assumptions were made to calculate this time? It would be interesting to have a range of potential transit times, because different raindrop sizes exist. Depending on the size and other factors, the fall velocity of a raindrop varies as shown already in Figure 9 (b).

The PDFs of the no-rain minutes were higher in case of the new procedure compared to the old one. Can the authors elaborate on the reasons for that and possible consequences? This example shows that a validation with more data is necessary to see how the data processing procedure behaves in non-rainy periods as well. This is important, because it raises the question if the proposed procedure can only be applied for measurements taken during rain or if it can be applied during dry and wet conditions.

Minor comments and technical corrections

 L10: The authors write significant reduction. Was the reduction analysed with a statistical test that supports the assumption of a significant reduction? If yes, please include this result in the manuscript. If not, please consider to remove the word ‘significant’.

L11:  It is not clear what should be understood by ‘the measurement distance of the lidar’. Distance to what?

L16-22: When starting with meteorology, examples of this application area should be mentioned first.  

L22-L23: Accurate wind speed measurements are required as well for example for data assimilation of numerical weather prediction models. Hence, accurate measurements are also relevant for authorities.

L42: I assume the measurements of Doppler lidars are influenced, not the instrument itself? Maybe the authors can clarify that.

L43-L44: Please remove the brackets around the reference of Träumner et al.

L46: Please remove the brackets around the reference Wei et al.

L54: The acronym ‘cw’ is not defined. Please add the information.

L107: What does Risø in the brackets mean? Is this the type/manufacturer of the cup anemometers? Please clarify.

L108-L109: What is the manufacturer of the wind vane and the air temperature sensor? In this connection, the wording ‘absolute temperature’ sounds strange. Maybe ‘air temperature’ is more appropriate?

L138-L139: It is not clear how the wake influence of the turbine was determined and why this was important for the experiment. Please clarify.

L144: Can the authors provide a reference to the Met Office’s definition?

L185: Please remove the brackets around the reference of Angelou et al.

L186: Why was the number 2.5 used for the analysis?

L195-L197: It is not clear what the authors want to express with the sentence starting with ‘Consequently, the projection of …’. Maybe a sketch could help?

L204: ‘… in detail’ instead of ‘in details’.

L208-L210: How much do raindrops influence the sonic measurements? The authors should provide some information about that in the sensor description in Chapter 2. Furthermore, what interpolation method was used?

Figures in general: It would be easier to read the caption if (a), (b), … are written before the actual description.

Figure 1: Can the authors add the information about the location of the disdrometer?

Figure 2: Do the red arrows indicate the location of the common focus point on the met mast? Please add some explanation about the arrows in the figure caption.

Figure 6: The disdrometer shown on this photo is not a Thies LPM, but a Ott Parsivel2. Please check the manufacturer of the disdrometer which was used in this experiment.

Figure 9: Is the plotted rain intensity taken from the automatic output of the disdrometer or calculated based on a quality-controlled rain-drop-size distribution?

Figure 10: It is a bit confusing using the same colours in (a), (b) and (c), although the colours in (c) describe not the same as in (a) and (b). The authors should consider using other colours or adding a legend to (c). Furthermore, the acronym ‘PSD’ is not described in the paper. This information should be added.

Figure 11: Strictly speaking, the Doppler signal is caused by aerosols not by wind.

Figure 12 and Figure 13: To the last sentence the information ‘in the scatter plot’ should be added to make the description clearer.

Figure 15: The figure could be simplified by plotting the bars in the same direction and the two different methods (SonicToRaw and SonicToNorm) are visualised by different colours (e.g. bright and dark). This would allow an easier comparison of the values.

Table 4: For ‘Light-rain minute 16:36’ in two cases three digits are given. Depending on the possible accuracy, please provide two or three digits for all numbers.

Table 4 & 5: Are the values calculated for the same time period plotted in Figure 12 and Figure 13? The figures represent the values for a bit more than exact one minute. The authors are asked to state exactly which time period (including seconds) was used for the values provided in the tables.

The authors are not consistent by using ‘rain drops’ and ‘rain droplets’. Please harmonize the description to ‘rain drops’.

The authors should check whether ‘filter out’ is more appropriate than ‘filter away’.

Sometimes the description of the lidars is ‘lidar #1/#2/#3’, sometimes ‘WindScanner #1/#2/#3’ and sometimes ‘WindScanner lidar #1/#2/#3’. To improve the reading, I suggest using the same description throughout the manuscript.