The authors identified meteor echoes as those having altitude ranges of about 100 km and low spectral width, based on the results by Chisham and Freeman (2013). I have a concern about the possibility of the contamination of non-meteor echoes, such as E-region echoes, sporadic E echoes, and PMSEs. They have not fully discussed these possibilities, especially because they deal with echoes with Doppler velocities larger than 100 m/s. In addition to the echoes I already mentioned, there are other echoes such as HAIR (high aspect angle irregularity region, Milan et al., 2003) and FAIR (far aspect angle irregularity region, St.-Maurice and Nishitani, 2020), both located in the lower E-region ionosphere around 100 km altitude. They are not meteor echoes and do not always move with the ambient neutral wind. Some of them have similar characteristics as meter echoes (e.g., located around 100 km altitude and having low spectral width). Therefore, possible contamination of these echoes (E-region, sporadic E-region, PMSE, HAIR, and FAIR echoes) should be discussed in the text.

The following comment is a response to both reviewers:

The large majority of echoes measured by SuperDARN at near ranges (<400 km) are a result of scattering from meteor trails, which drift at the neutral wind speed. However, it is true that there is occasional contamination of these neutral wind measurements by E-region echoes, although these are only a significant problem in and near the auroral zone where field-aligned (and other) ionospheric irregularities are ubiquitous. Chisham and Freeman (2013) removed a large proportion of this E-region contamination from echoes measured by the Saskatoon SuperDARN radar, within the auroral zone, by understanding the differences in the probability distributions of the velocity error and spectral width error of the different echo types (as explained in detail in section 2.4 of that paper). Although not removing this E-region contamination completely, the filtering proposed in that paper reduced it significantly.

The echoes measured by the mid-latitude FIR SuperDARN radar and presented in this paper are very different to those observed at auroral latitudes. The near ranges are overwhelmingly dominated by meteor echoes. FIR is located at mid-latitudes (~55 degrees south geographic latitude; ~40 degrees south geomagnetic latitude), a significant distance from the auroral region, which only appears at far ranges in the FIR field-of-view during disturbed times. Hence, E-region echoes are very rare in the FIR data set, especially during the solar minimum interval studied here. At the near-ranges where meteor echoes are observed they are almost non-existent. Hence, contamination of meteor echoes by E-region echoes is not a problem with the FIR observations.

Visual inspection of daily plots of scatter throughout the analysis interval indicates that the only potential contamination of the meteor echoes within the FIR near-range data is from PMSE/MSE and sea scatter (following E-region reflection). This contamination is rarely a problem at the nearest ranges and occurs predominantly around noon/early afternoon in the summer months. This can be shown by a statistical investigation of echo characteristics at near ranges, which shows the domination of meteor echo characteristics at the nearest ranges. The potential for contamination of the data set from non-meteor echoes will now be discussed in more detail in the revised version of the paper, with a figure that demonstrates the changes in echo characteristics due to different scattering targets at near ranges.

The most promising way to solve the problem of distinguishing meteor echoes is to obtain the raw time series data, as reported by Yukimatu and Tsutsumi (2002) and Tsutsumi et al. (2009). Meter echoes should appear in the raw time series data as the echoes with a sudden increase of the echo power, followed by its exponential decay. Chisham and Freeman (2013) argue that the standard SuperDARN radars do not record raw time series data. However, the function of recording raw time series data has been implemented in the standard SuperDARN radar operation software, and several SuperDARN radars actually record raw time series data. I do not require the authors to analyze the raw time series data in the current study. However, it is not the right manner to ignore previous studies completely. Works related to raw time series data analysis should be introduced in the discussion.

The reviewer is right that the paper should discuss the previous methods that have been used to measure SuperDARN meteor echoes at high-time resolution. The method presented by Yukimatu and Tsutsumi (2002) is the only method that can presently be used to study individual SuperDARN meteor echoes in detail. We now intend to discuss this method in the introduction of the revised paper. However, our view is that this method is impractical for studying very long data sets to study phenomena such as atmospheric tides. Hence, we feel that the approach that we have taken with the SuperDARN data analysis is the most suitable for the phenomena that we wish to analyse.

Line 173 “persistant” should be “persistent”

This will be corrected in the revised paper.

In this paper, the authors presented high-resolution analysis of meridional winds obtained from meteor echoes observed by the superdarn radar.  High-resolution (1 min) continuous wind observations are very important for addressing the influence of the entire spectrum of waves on the  MLT region.   However, the data procedure is vaguely described. 

We don’t agree that the data procedure is vaguely described. The data procedure is standard as for many previous SuperDARN meteor scatter observations. Our paper includes a concise explanation of the meteor echo analysis and is well referenced regarding the measurements of meteor echoes with SuperDARN, and too much more detail in this paper would involve the repetition of text and information which is easily found in the referenced articles.

How many underdense meteors can be detected per minute or per hour?  Need to include a plot showing time variation of meteor counts. 

We already state in our paper `Figure 3 of Hibbins et al. (2011) presented the seasonal and diurnal variation of the mean number of meteor echoes observed during the first year of FIR operations.’ However, we can see that it would be instructive to show the number of meteor echoes typically observed per hour in the data set, so in the revised version of the paper we will add a figure showing this information. As the SuperDARN integration time is one minute, it is only possible to observe a single meteor echo during each one-minute interval. This is standard for SuperDARN meteor echo observations.

The authors need to show clearly how meteor echos and E-region echoes are differentiated, with a few examples. 

The following comment is a response to both reviewers:

The large majority of echoes measured by SuperDARN at near ranges (<400 km) are a result of scattering from meteor trails, which drift at the neutral wind speed. However, it is true that there is occasional contamination of these neutral wind measurements by E-region echoes, although these are only a significant problem in and near the auroral zone where field-aligned (and other) ionospheric irregularities are ubiquitous. Chisham and Freeman (2013) removed a large proportion of this E-region contamination from echoes measured by the Saskatoon SuperDARN radar, within the auroral zone, by understanding the differences in the probability distributions of the velocity error and spectral width error of the different echo types (as explained in detail in section 2.4 of that paper). Although not removing this E-region contamination completely, the filtering proposed in that paper reduced it significantly.

The echoes measured by the mid-latitude FIR SuperDARN radar and presented in this paper are very different to those observed at auroral latitudes. The near ranges are overwhelmingly dominated by meteor echoes. FIR is located at mid-latitudes (~55 degrees south geographic latitude; ~40 degrees south geomagnetic latitude), a significant distance from the auroral region, which only appears at far ranges in the FIR field-of-view during disturbed times. Hence, E-region echoes are very rare in the FIR data set, especially during the solar minimum interval studied here. At the near-ranges where meteor echoes are observed they are almost non-existent. Hence, contamination of meteor echoes by E-region echoes is not a problem with the FIR observations.

Visual inspection of daily plots of scatter throughout the analysis interval indicates that the only potential contamination of the meteor echoes within the FIR near-range data is from PMSE/MSE and sea scatter (following E-region reflection). This contamination is rarely a problem at the nearest ranges and occurs predominantly around noon/early afternoon in the summer months. This can be shown by a statistical investigation of echo characteristics at near ranges, which shows the domination of meteor echo characteristics at the nearest ranges. The potential for contamination of the data set from non-meteor echoes will now be discussed in more detail in the revised version of the paper, with a figure that demonstrates the changes in echo characteristics due to different scattering targets at near ranges.