Dear authors,
Thank you for careful responding to the reviewer comments. However, the new added text is of lower quality than the rest of the manuscript. Please find below some minor comments to be taken into account.
* Page 2, line 8: The Meisei iMS-100 radiosonde … --> Why are you giving this information if this type of radiosonde is not used in the manuscript? Or is the Meisei RS-11G radiosonde replaced by this type of radiosonde since September 2017.
* Page 2, line 30-31: Replace with: “GRUAN sites take over some of the duties of the Lead Centre …’
* Page 3, line9: a Vaisala DigiCora III Sounding System, I assume
* Page 4, lines 3-5: The minima-pass filter is applied to the temperature measurements and only picks up minimum values within a certain time window (Kizu et al., 2018).
* Page 4, line 11: add “are” before difficult to quantify.
* Page 5, line 13: add “(optional)” after a silicon pressure sensor because not all RS92 radiosondes have this pressure sensor implemented!
* Page 5, lines 22-23 vs. Page 5, lines 27-28: these two statements seem contradictory. First, you state that the raw temperature data of RS92 are corrected for heat spike errors as with RS-11G, so that the reader might conclude that you apply the same minima-pass filtering. But at the end of the paragraph, you write that heat spike errors are removed by applying a low-pass digital filter with a cut-off frequency of 0.1 Hz. Please remove this ambiguity.
* Page 6, lines 17-19: It is completely not obvious what is meant here. If the quality control procedures have not been established and are still under consideration, why do you use these 82 flight data after all? From this statement, it seems that you are using data that have not been quality controlled at all. Is this the case? Please clarify and in particular, give more details about the reason why those 5 RS-11G flight data have been removed.
* Page 6, lines 20- 30: this (new) part is poorly written. Here is a suggestion: “These quality control procedures for the RS92 GDP (Dirksen et al., 2014) consist of two steps: first, the results of the ground check procedure are verified, after applying the GRUAN corrections to the raw RS92 measurements. In a second step, it is checked that the profile data have estimated uncertainties of GDP that are within the uncertainties provided by the manufacturer. For instance, for one of the 25 rejected RS92 data, there was more than 1.5%RH difference between the RH sensors at the ground check. However, most of the excluded 25 RS92 flights failed the second step of the quality control procedures . Among these, two thirds were daytime observations. Furthermore, of the remaining 57 dual soundings, 5 dual soundings were additional blacklisted due to spurious RS-11G RH measurements or based on outlying temperature differences. So we end up with 52 sets of dual flight data for the data analysis”. Please note that I dropped the part of the sentence in line 28, as it is not clear to which sample those 8 dual sounding belong (25 RS92 profiles that have been excluded? 5 dual soundings with a lot of outliers?).
* Page 7, lines 7-19: the described results of the analysis of the temperature differences are not so obvious from the figures to my opinion. For instance, you write “The temperature difference and standard deviation for the paper cardboard rod tend to be somewhat larger than those for the bamboo and plastic rods in the lower troposphere and the lower stratosphere.” But, from the figure, could you also not conclude that paper cardboard and plastic rod having similar temperature differences? The main question from this analysis :has the rod material an impact on the temperature differences? is not convincingly answered in this analysis. Are the differences and standard deviations between the temperatures for the different rod materials significant or not? Also the last sentence of this paragraph (by the way, I think that you mean here the standard deviations of differences) is not obvious to me from the figures. You might also include to which panels of Fig 7 you are alluding to when making these statements in this paragraph.
* Page 9, lines 1-3: does this statement also hold for the differences, next to the standard deviations? Also, you might repeat the remaining spring daytime sample size here after removing the exceptional case.
* Page 12, lines 14-22: another added piece of text of lower quality. First of all, you should refer to the figures that you provided in the Supplement here. Also, you do not explicitly mention the possible reasons (brought up in the short comment SC1) to which you refer in the first sentence and that you try to investigate [a) the ensemble average difference is not zero, b) the distribution is not Gaussian, c) the number of samples is small]. You should first mention those. Also replace “The authors” by “We”, “is supposed to be available” with “will”, and “Further works” with “Further work”.
* Page 13, line 1: please specify the pressure range you are referring to for the middle troposphere.
* Page 13, lines 4-7: Change to “In addition, we noted that the humidity sensor of RS-11G has a dry bias in the lower troposphere and a wet bias in the upper troposphere when compared to chilled-mirror hygrometer measurements (see Section 6). In the future, we will try to improve…”
* Page 13, lines 27-28. Change to “These comparisons confirm that the RS-11G GDP has a wet bias between 400 and 200 hPa, as was also common in the results shown in Fig. 11”. Please repeat the origin for this effect here: “We ascribed this to …”
* Page 14, lines 5-7. Change to “We compared the GDPs using a general statistical approach based on 13 allocated pressure layers, but an analysis using the functional regression approach (e.g. Fasso et al., 2014) might be applied in the future as well.
* Page 14, lines 13-16. Please add the reason for this important finding.
* Fig 6: please add “0.9 or” 1 m for the rod length on this figure.
* Caption of Fig 7: change to “bamboo rod with radiosondes hanging freely”
* Caption of Fig 16: Replace with “Seasonal profiles for wind speed and direction (as for Fig. 9).
* Caption of Fig 17: Replace with “Profiles of differences and standard deviation, as for Fig. 10, but for wind speed for all soundings (00 and 12 UTC combined)”.