Climatology and Long-Term Trends in the Stratospheric Temperature and Wind Using ERA5

Round 1

Reviewer 1 Report

Kozubek et al., (2021) presented the results on climatology and long-term trends in the stratosphere temperature and winds using ERA-5. Their study focused on the altitude region of 100-1 hPa. They compared the impact of NAO, ENSO, QBO, solar flux, and SSW on observed trends. And they have concluded that the observed trends are due to the occurrence of SSW. This article reads nicely I recommend it for publication with the changes by following comments/questions. I have a few questions and comments: (1) Their analysis is on ERA-5, It may be important to state in the introduction that why do they have chosen the ERA-5 datasets over others. (2) The high altitude balloons and even regular balloons could reach ~35 km in a given time; Can they comment why they did not use the temperature and winds from those GPS radiosonde balloons of a few locations for comparison with ERA-5 trends and climatology. (3) It is better to make a comparison (plots) of these trends with GPS radiosonde balloons reported trends since ERA-5 may not represent the short vertical scale waves. (4) On page 4 & line 4 and thereafter, the wording lower pressure levels seem to represent the 30 hPa and 100 hPa but there 1, 5, and 10 hPa in lower in magnitude wise; it is confusing; Can you rephrase these words? (5) Why in Figure 10, Y-axis does not have magnitude for their individual variables Can your make it. (6) Can you make a comparison by including the SSW (Intensity of SSW and/or occurrence of SSW) intensity in Figure 10. (7) Can you make a time-series plot (at 10 hPa & 100 hPa) of temperature-residuals at the hot spots (using the intense contours from Figure 1:K/year) with trend line for (a) Canada/Northern Pacific, (b) Northern Atlantic, (c) Euroasian continent, (d) tropics (land and ocean) along with the vertically stacked plots of NAO, ENSO, QBO, Solar flux, and SSW.

Author Response

Kozubek et al., (2021) presented the results on climatology and long-term trends in the stratosphere temperature and winds using ERA-5. Their study focused on the altitude region of 100-1 hPa. They compared the impact of NAO, ENSO, QBO, solar flux, and SSW on observed trends. And they have concluded that the observed trends are due to the occurrence of SSW. This article reads nicely I recommend it for publication with the changes by following comments/questions. I have a few questions and comments:

1. Their analysis is on ERA-5, It may be important to state in the introduction that why do they have chosen the ERA-5 datasets over others.

We choose ERA5 reanalysis because this reanalysis has not been included into last S-RIP report and it is one of the most used reanalysis for atmospheric analysis not only in the troposphere but in the stratosphere as well. It is the newest released reanalysis. Moreover, it covers whole analyzed period.

2. The high altitude balloons and even regular balloons could reach ~35 km in a given time. Can they comment why they did not use the temperature and winds from those GPS radiosonde balloons of a few locations for comparison with ERA-5 trends and climatology.

Of course that we can use balloon measurements but problem with these measurements is that comparison with regular gridded reanalysis is not ideal. Balloons provide information at a given time and grid point but when it reaches higher altitude, comparison with specific grid point is not accurate, because balloon may be moved by winds substantially away. More information can be found in Kozubek et at. (2014), where we used balloon measurements.

Kozubek, Michal ; Laštovička, Jan ; Križan, Peter, 2014: Differences in mid-latitude stratospheric winds between reanalysis data and versus radiosonde observations at Prague, Annales Geophysicae, 32, -, 353-366

Moreover, GPS RO can provide much more information that balloon measurements.

3. It is better to make a comparison (plots) of these trends with GPS radiosonde balloons reported trends since ERA-5 may not represent the short vertical scale waves

GPS RO observations have also high vertical resolution and they provide much more data than balloons.

4. On page 4 & line 4 and thereafter, the wording lower pressure levels seem to represent the 30 hPa and 100 hPa but there 1, 5, and 10 hPa in lower in magnitude wise; it is confusing; Can you rephrase these words?

Thank you; this was corrected.

5. (5) Why in Figure 10, Y-axis does not have magnitude for their individual variables. Can you make it?

Because there are five different variables with five different scales. For the purpose of the paper, only relative variations, not absolute values, are important.

6. Can you make a comparison by including the SSW (Intensity of SSW and/or occurrence of SSW) intensity in Figure 10.

First, the five variables shown are continuous variables, whereas SSWs are sporadic events, typically 0-2 per winter. Second, Figure 10 is busy enough even without SSWs. Therefore SSW is not included there.

7. (7) Can you make a time-series plot (at 10 hPa & 100 hPa) of temperature-residuals at the hot spots (using the intense contours from Figure 1:K/year) with trend line for (a) Canada/Northern Pacific, (b) Northern Atlantic, (c) Euroasian continent, (d) tropics (land and ocean) along with the vertically stacked plots of NAO, ENSO, QBO, Solar flux, and SSW.

Thank you for your suggestion. We know that it is possible to do more detailed analysis for specific area (hotspot) for temperature and winds. In this study we focus mainly on the overview of situation and conditions in the middle and higher latitudes. This study will show us the direction for the next studies where we plan to focus on the hotspots and if possible to compare this with more reanalysis or satellite measurements.

Reviewer 2 Report

1^st Review of “Climatology and long term trends in the stratospheric temperature and wind using ERA-5” submitted by Michal Kozubek, Jan Lastovicka and Radek Zajicek to Remote Sensing

The current manuscript described the long-term characteristics of stratospheric temperature and zonal wind by using ERA-5, reanalysis data. The previous study found that the zonal and vertical asymmetric structure of stratospheric temperature and wind field, however there is no explanation of the mechanisms. Overall the manuscript, the descriptions are less and the quantitative explanation/discussion of figures are less. Totally, the quality of the manuscript is poor, the current manuscript is rejected. Further this manuscript is not fit to the aim of the journal of remote sensing, should submit to other journal, like atmospheric science.

Major comments:

• The latest scientific knowledge on remote sensing is not found in this study.
• The figure is too small to see, and there is not enough explanation. Why is the low latitude white in Figs2-9? There is no explanation that it focuses only on the winter season in both hemispheres.
• The mechanism of the asymmetric (cell) structure was absent, that is the key point of this study. The authors tried to explain the SSW influence on temperature at the high latitudes of NH, but how influence? And how much to explain the variability by major SSW?

Minor comments:

Data set : There is absent of data accuracy of GPS and the explanation of MERRA2.

Fig1: The white line (0 degree longitude) should be removed or modify to use smoothing and so on. Why was the 50hPa pressure level selected? In the bottom panels, the line of continent and explanation of contour interval was absent. The anomaly or scatter plot would be effective to discuss the difference, like ERA5-MERRA2 or ERA5-GPS. The average period of GPS, 2010-2020, was different from ERA5 1980-2020, it is big impact because the LS temperature had negative trend before 2000.

Figures : (Same comments for Fig 2-9) the contour color is not seen clearly the positive/negative values, the blue and red color gradation should be fine. Further the label of latitude was absent. The description about 10 and 30hPa was less, the removal There is almost no description on the 10 and 30hPa so you can delete it and enlarge the figures. In Figure4, what does the hatch mean?

Figure 10 : The time serise of temperature and wind were absent, and the y-axis was also absent. Further, each index explanation were absent.

Author Response

1st review of “Climatology and long term trends in the stratospheric temperature and wind using ERA-5” submitted by Michal Kozubek, Jan Lastovicka and Radek Zajicek to Remote Sensing
The current manuscript described the long-term characteristics of stratospheric temperature and zonal wind by using ERA-5, reanalysis data. The previous study found that the zonal and vertical asymmetric structure of stratospheric temperature and wind field, however there is no explanation of the mechanisms. Overall the manuscript, the descriptions are less and the quantitative explanation/discussion of figures are less. Totally, the quality of the manuscript is poor, the current manuscript is rejected. Further this manuscript is not fit to the aim of the journal of remote sensing, should submit to other journal, like atmospheric science.

A: This paper has been invited into this special issue. I sent abstract in advance to editor and editor confirmed acceptability of this paper. I understand that main content of the manuscript is focused on model results but we use reanalysis. And reanalysis is mainly based on satellite observation which could match the scope of Atmospheric remote sensing.

Major comments:

• 
     The latest scientific knowledge on remote sensing is not found in this study.

A: We think that the latest knowledge related to the topic of this paper is essentially included in the paper. But we included SPARC report with more relevant information.

• 
     The figure is too small to see, and there is not enough explanation. Why is the low latitude white in Figs2-9? There is no explanation that it focuses only on the winter season in both hemispheres.

A: We have tried to improve all figures. We add explanation why we used only middle and higher latitudes in winter. Lower latitudes in winter is not so important for stratospheric dynamics.

• 
     The mechanism of the asymmetric (cell) structure was absent, that is the key point of this study.

 A: We added reference into the paper for this mechanism. It is due to stationary planetary wave with zonal wavenumber 1.

The authors tried to explain the SSW influence on temperature at the high latitudes of NH, but how influence? And how much to explain the variability by major SSW?

A: We agree that there is no quantitative analysis for this conclusion so we can delete this statement. On the other hand, more major SSW definitely influences climatology of temperature and zonal wind because this phenomenon affects the whole stratosphere for more than 14 days.

Minor comments:

Data set: There is absent of data accuracy of GPS and the explanation of MERRA2.

A: The neutral atmospheric profiles retrieved from COSMIC-1 radio occultation (RO) data have been demonstrated to be very useful for studying atmospheric processes (Shao et al., 2021 and 15 references herein). GPS (GNSS) RO data are highly accurate; they were used among others to calibrate AMSU instruments onboard satellites (Shao et al., 2021 and references herein). A note about that is now included in our paper.   Shao,X., Ho, S.-P., Zhang B., Cao, C., Chen, Y., Consistency and stability of SNPP ATMS microwave observations and COSMIC-2 radio occultation over ocean, Remote Sens. 2021, 13, 3754. https://doi.org/10.3390/rs13183754.

Fig1: The white line (0 degree longitude) should be removed or modify to use smoothing and so on. Why was the 50hPa pressure level selected? In the bottom panels, the line of continent and explanation of contour interval was absent. The anomaly or scatter plot would be effective to discuss the difference, like ERA5-MERRA2 or ERA5-GPS. The average period of GPS, 2010-2020, was different from ERA5 1980-2020, it is big impact because the LS temperature had negative trend before 2000.

A: This figure is taken form the Kozubek et al. (2020), where we compare climatology and trends in MERRA2, ERA5 and GPS RO. This figure is shown here as an example.

Kozubek, Michal, Križan, Peter, Laštovička, Jan, 2020: Homogeneity of the Temperature Data Series from ERA5 and MERRA2 and Temperature Trends, Atmosphere, 11(3): 235, doi: 10.3390/atmos11030235.

Figures : (Same comments for Fig 2-9) the contour color is not seen clearly the positive/negative values, the blue and red color gradation should be fine. Further the label of latitude was absent. The description about 10 and 30hPa was less, the removal There is almost no description on the 10 and 30hPa so you can delete it and enlarge the figures. In Figure4, what does the hatch mean?

A: Thank you for your suggestion. We have improved our figures and tried to rearrange all panels. But we prefer to use panels as they are because we think that the better way is to show temperature and its differences and after that wind analysis. 10 and 30 hPa is very important as a representative of middle stratosphere even if they are not used much in the description.

Figure 10 : The time series of temperature and wind were absent, and the y-axis was also absent. Further, each index explanation were absent.

A: The meteorological and solar indices used are basic, well-known and regularly/often used indices; therefore their description is not included. If necessary, it could be included. Description of y-axis is absent because each of five indices plotted has its own scale (y-axis) and because only relative changes/variations, not absolute values, are of interest for the sake of the paper. Time series of temperature and wind is absent – which time series? We are working with individual grid points, so for which grid point to take temperature or wind?

Reviewer 3 Report

Please see attached file

Comments for author File: Comments.docx

Author Response

Major:

1. The comparison of ERA5 to GPS RO and MERRA2 is meant to show that ERA5 is suitable for this analysis (as stated on P 3, end of 1^st paragraph of Results), but the comparison is much too superficial to draw any conclusion on the matter. There is only one figure and it is difficult to interpret.  The maps images are much too small to see any detail, the longitude labels are unreadable and they obscure much to the data since they are place on each map, and part of the latitude labels also run into all of the maps.  For a more accurate picture of how these three datasets compare, there needs to be another figure with differences shown between ERA5 and the other two datasets.  Maps with actual differences are much more quantitative than trying to compare colors and patterns by eye. Also, this comparison only covers 2010-2020 (as stated in the text at the top of p. 3), or 2008-2020 (as stated in the caption). 

A: This figure 1 is taken from Kozubek et al. (2020), where we compare climatology and trends through MERRA2, ERA5 and GPS RO. This figure is shown here as an example. We found differences in amplitude, especially at 100 hPa in warm areas, where ERA5 reanalysis and GPS RO were about 3 K warmer than MERRA2. Comparison is only made for 2008-2020 due to availability of GPS RO data.

Kozubek, Michal, Križan, Peter, Laštovička, Jan, 2020: Homogeneity of the temperature data series from ERA5 and MERRA2 and temperature trends, Atmosphere, 11(3): 235, doi: 10.3390/atmos11030235.

2. Either way, since the paper looks at ERA5 back to 1980, it is important to present a comparison for an earlier period. GPS is likely not available but there definitely should be something for MERRA2. In this respect, the authors could benefit greatly by examining the SPARC Reanlaysis Intercomparison Project S-RIP report (Masatomo Fujiwara, Gloria L. Manney, Lesley J. Gray, and Jonathon S. Wright (Eds.), SPARC Report No. 10, WCRP-6/2021, doi: 10.17874/800dee57d13, sparc-climate.org/publications/sparc-reports), which should also be cited in this paper.  In fact there is a lot of information in this report and the associated papers in the ACP special issue (https://acp.copernicus.org/articles/special_issue829.html), which is highly relevant to the topic of this paper, and all of these aspects should be discussed and compared with the results in this paper.

A: S-RIP report is now cited in our paper but it does not include ERA5. We have added some information about difference between reanalysis and GPS RO.

3. Analysis of the impact of SSW, NAO, ENSO, and QBO on the trends is lacking in depth. First, the data sources for all four of the time series for these phenomena are not described, along with any of the limitations for these datasets, and whether or not they are also suitable for trend analyses.

A: We have added some new information about datasets and all sources are cited. Particularly ENSO and QBO have been used in many multi-parametric regression studies of stratospheric trends.

4. Second, the reader is asked to compare the time series in Figure 10 with the differences and trends in temperatures and winds, which is practically impossible. A more quantitative analysis would compare these time series with suitably developed time series from the ERA5 dataset, and then conduct a principal component analysis or a multivariable least squares fit in order to compare the relative contributions from each forcing agent.

A: Thank you for your suggestion. Unfortunately it is not technically possible to compare mentioned datasets of phenomena directly with ERA5 in each grid point. Some zonal or other averaging would be necessary for such comparison. We agree that using multivariable least squares analysis is more proper way how to test connection between these datasets. But this study should show an overview of the situation using ERA5 and only basic analysis and identification of specific areas for future detailed studies to be done.

5. Note that the axis labels for years in Fig 10 are too crowded to read – perhaps labeling every other year or every five years might fix this.

We have fixed this problem. Thank you.

6. Third, the SSW table in Figure 11 needs much more explanation in order for it to be useful. Major versus minor versus final needs to be defined,

A: We have used WMO definition of major SSW.

The data sources identified (again with uncertainties and limitations). 

A: We have used ERA 5 reanalysis data for identification of SSW.

7. Why is the final warming noted in only a few cases, doesn’t a final warming occur every year? Why include events in March when that month is not even included in the ERA5 analysis for the NH (DJF)?  The same issue arises for having September and October in Figure 11 but these months are not used for the SH analysis (JJA).

A: This analysis and figure 11 has been done for other study and we show it only as an example. We understand that it includes more month than we use in this study. We can delete this figure and because it generally known the occurrence of SSW during last four decades we can only describe this information. But we prefer to use this figure for readers.

8. There is not a quantitative analysis to support conclusion (4) “The possible reason for this behavior is the occurrence of major SSWs in the stratosphere”. This unsubstantiated point is also made in the abstract.

A: We agree that there is no quantitative analysis for this conclusion so we can delete this statement. On the other hand, more major SSW definitely influences climatology of temperature and zonal wind because this phenomenon affects the whole stratosphere for more than 14 days. We reformulate this point of conclusion.

Minor:  .

1. 1 end of 1^st paragraph: “radiation forcing” should be “radiative forcing”.

Done.  

2. 2 2^nd paragraph: the discussion of SSWs is lacking in completeness and missing a number of important references: see [Baldwin, M. P., Ayarzagüena, B., Birner, T., Butchart, N., Butler, A. H., Charlton-Perez, A. J., Domeisen, D. I. V., Garfinkel, C. I., Garny, H., Gerber, E. P., Hegglin, M. I., Langematz, U., and Pedatella, N. M.: Sudden Stratospheric Warmings, Rev. Geophys., 59, e2020RG000708, https://doi.org/10.1029/2020RG000708, 2021] and many of the references therein.

This reference is now included and discussed. 

3. P 3, top: The text says the analysis is done between 40-90 N/S, but conclusion (2) says 20-70 degrees latitude.

This is corrected.   

4. Figures 3, 5, 7, and 9 would be greatly improved for interpretation if there was some kind of delimiter to separate the panels according to the difference. There are three sets of differences shown in each figure. For example, all of the differences with respect to 80-90 should be grouped together, and the same for differences with respect to 90-00, and the one difference with 00-10. Perhaps solid horizontal lines for separation into these groupings might be considered, or some other way to designate sets of differences.

Thank you for your suggestion. We have improved our figures and tried to rearrange all panels. But we prefer to use panels as they are because we think that the better way is to show temperature and its differences and after that wind analysis.

5. Reference to “first” “second” and “third” decades should be avoided especially in the abstract where there is no basis for context. It would be an improvement to refer to 1980-90, 1990-2000, etc. throughout the manuscript.

We have corrected this.

6. The last row for Figure 2 (2010-2020) is clipped at the bottom in my copy.

This problem does not occur in our docx copy.

7. P 9, middle, the statement that there is evidence of tropospheric warming based on positive trends at 100 hPa is unsubstantiated because the 100-hPa level corresponds to the lower stratosphere between 40 and 90 degree latitudes in winter.

We have corrected this.

8. The title uses “ERA-5”, whereas the text uses ERA5. These should be consistent.

We have corrected this.

Reviewer 4 Report

If I understand it right, this manuscript was submitted to the section "GNSS Atmospheric Modeling” and “Atmosphere Remote Sensing”. Even though the manuscript includes the COSMIC RO temperature observations, which belong to GNSS remote sensing, however, the main contents of the manuscript focus on the atmospheric model results. So I do not think the manuscript matches the scope of this journal and section. Regarding the manuscript itself,  it describes the long-term climatology temperature from ECMWF model and observation. This is a very interesting and important topic, especially it is about the troposphere where most people neglect it. However, the manuscript does not present the results well. For the 4 decades of ECMWF data, there might be better ways of showing the long-term variations, rather than decade averaging. The discussions about the temperature/zonal wind and their trends are too general. The readers would expect something new, rather than results agreeing with other older studies. Only winter data is included and the data over the equatorial region is not included in the results. The authors need to better explain the reasons the methods. Some minor issues: The map view should choose the same projection method, for Figure 1.

Author Response

If I understand it right, this manuscript was submitted to the section "GNSS Atmospheric Modeling” and “Atmosphere Remote Sensing”. Even though the manuscript includes the COSMIC RO temperature observations, which belong to GNSS remote sensing, however, the main contents of the manuscript focus on the atmospheric model results. So I do not think the manuscript matches the scope of this journal and section.

A: This paper has been invited into this special issue. I sent abstract in advance to editor and editor confirmed acceptability of this paper. I understand that main content of the manuscript is focused on model results but we use reanalysis. And reanalysis is mainly based on satellite observation which could match the scope of Atmospheric remote sensing.

Regarding the manuscript itself, it describes the long-term climatology temperature from ECMWF model and observation. This is a very interesting and important topic, especially it is about the troposphere where most people neglect it. However, the manuscript does not present the results well. For the 4 decades of ECMWF data, there might be better ways of showing the long-term variations, rather than decade averaging. The discussions about the temperature/zonal wind and their trends are too general. The readers would expect something new, rather than results agreeing with other older studies. Only winter data is included and the data over the equatorial region is not included in the results. The authors need to better explain the reasons the methods.

A: Thank you for your suggestion. Analyzing of all grid points instead of zonal averaging will bring us an overview which regions should be analyzed in more details. We have added more discussion about results and comparison with SPARC report. We also added more information about used datasets (SSW, NAO, ENSO or MERRA2). This study should show an overview of the situation using ERA5 and only basic analysis and identification of specific areas for future detailed studies to be done. We choose ERA5 reanalysis because this reanalysis has not been included into last S-RIP report and it is one of the most used reanalysis for atmospheric analysis not only in the troposphere but in the stratosphere as well. It is the newest released reanalysis. Moreover, it covers whole analyzed period. We have improved our figures and tried to rearrange all panels. But we prefer to use panels as they are because we think that the better way is to show temperature and its differences and after that wind analysis.

Masatomo Fujiwara, Gloria L. Manney, Lesley J. Gray, and Jonathan S. Wright (Eds.), SPARC Report No. 10, WCRP-6/2021, doi: 10.17874/800dee57d13

Some minor issues:

The map view should choose the same projection method, for Figure 1.

A: This figure is taken from the Kozubek et al. (2000), where we compare climatology and trends through MERRA2, ERA5 and GPS RO. This figure is shown here as an example.

Kozubek, Michal, Križan, Peter, Laštovička, Jan, 2020: Homogeneity of the temperature data series from ERA5 and MERRA2 and temperature trends, Atmosphere, 11(3): 235, doi: 10.3390/atmos11030235.

Round 2

Reviewer 2 Report

Major revisions on figures

Figures in revised manuscript were not improved yet.

Figure 1 should be unified to the Mercator projection like the other figures. The longitude label is complicated.

Remove the diagonal lines in the figures, such as Figs 4,5 and 8.

In Figs 2-9, because the contour color is not seen clearly the positive/negative values, the contour with zero value should be drawn.

Author Response

Figures in revised manuscript were not improved yet.

Figure 1 should be unified to the Mercator projection like the other figures. The longitude label is complicated.

A: The figure 1 has been removed and replaced ba another figure.

Remove the diagonal lines in the figures, such as Figs 4,5 and 8.

A: Diagonal lines have been deleted.

In Figs 2-9, because the contour color is not seen clearly the positive/negative values, the contour with zero value should be drawn.

A: All figures has been redrawn to see positive and negative values.

Reviewer 3 Report

please see attached review

Comments for author File: Comments.pdf

Author Response

Major:

1. The comparison of ERA5 to GPS RO and MERRA2 has not been improved in any way. Even though this figure has been published previously, it does not change the fact the figure is too small, it has text that covers much of the data, and it is not quantitative because it does not show actual differences. Atmosphere should not have allowed such a poor figure to appear in their publication. The authors pointed out in response that GPS RO is not available for the earlier period, but completely ignored this reviewer’s suggestion to show earlier periods in comparison with MERRA2.

A: Thank you for your suggestion. We have changed figure 1 and now it shows only comparison of ERA5 and MERRA-2 (climatology and trends) for period 1980-2020. This comparison shows that at least down from 10 hPa the main features are very similar. The comparison with GPS RO are only discussed and reference has been added.

2. The data sources for the QBO, F10.7, and NAO time series shown in figure 10 are only listed in the end of the paper, within a “Data Availability Statement”, and they are only a laundry list of websites. It is well known that there is an abundance of faulty data, misinformation, and outright falsehoods on the internet. If this paper is intended to be a contribution to the body of scientific knowledge, then any data that is presented should have a clear connection to the peer-reviewed literature, and the paper should not require that the audience verify the appropriateness of the data shown in the paper. Instead, it must include a thorough discussion and summary of previous papers that have examined the soundness and/or any known limitations of such data.

A: Thank you for your suggestion. We have added information and discussion about main datasets like solar flux, QBO or NAO. We understand that it is possible that some errors or problems can occur in the dataset websites but on the other hand NOAA, NASA and berlin university website could be considered as a verified source of data for scientific community. One reference concerning F10.7 and QBO data series homogeneity is added.

3. The assertion that “The possible reason for behavior of climatology is irregular occurrence of major SSWs in the stratosphere.” is still not supported by the results shown in this paper. The reader is asked to compare Figure 11, which is a table of symbols, with maps that show trends in temperatures and zonal winds in Figures 4 and 8. Not only is it non-quantitative, but it is nearly impossible to do even qualitatively by eye,.

A: We understand that comparison is not possible directly. That´s why we added some information about several major SSW behaviour during analysed period which show that changes of temperature and wind during major SSWs are enormous and last sometimes longer than 3 weeks so the probability that it affects at least climatology is very high.

Minor:

1. There is new text about the QBO, but it is completely wrong in saying “On the other hand, QBO is generally very stable parameter without any big variation.” There have been some major disruptions to the QBO recently, see

Osprey, S. M. et al. An unexpected disruption of the atmospheric quasi- biennial oscillation. Science 353, 1424–1427 (2016).

Newman, P. A., Coy, L., Pawson, S. & Lait, L. R. The anomalous change in the QBO in 2015–2016. Geophys. Res. Lett. 43, 8791–8797 (2016).

           A: We have deleted this statement.

After the new text discussion on NAO, there is a non-English sentence. Also after the caption for figure 3.

A: we have deleted these sentences.

English of the paper has been checked and improved.

Reviewer 4 Report

I would like the authors could go through the manuscript to check some English wording issues, one instance is about the use of article words.  There are some sentences with non-English words, please correct them.

Author Response

We went through whole text and corrected English spelling, article etc. Non english sentences have been removed.