Interannual variability of winds in the Antarctic mesosphere and lower thermosphere over Rothera (67° S, 68° W) in radar observations and WACCMX
 ^{1}Centre for Space, Atmospheric and Oceanic Sciences, Department of Electronic Engineering, University of Bath, Bath, UK
 ^{2}Atmosphere, Ice and Climate Team, British Antarctic Survey, Cambridge, UK
 ^{3}Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, CO, USA
 ^{4}Virginia Polytechnic Institute and State University, Blacksburg, VA, USA
 ^{5}High Altitude Observatory, National Center for Atmospheric Research, Boulder, CO, USA
 ^{1}Centre for Space, Atmospheric and Oceanic Sciences, Department of Electronic Engineering, University of Bath, Bath, UK
 ^{2}Atmosphere, Ice and Climate Team, British Antarctic Survey, Cambridge, UK
 ^{3}Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, CO, USA
 ^{4}Virginia Polytechnic Institute and State University, Blacksburg, VA, USA
 ^{5}High Altitude Observatory, National Center for Atmospheric Research, Boulder, CO, USA
Abstract. The mesosphere and lower thermosphere (MLT), at heights of 80–100 km, is critical in the coupling of the middle and upper atmosphere and controls the momentum and energy transfer between these two regions. However, despite its importance, many General Circulation Models (GCMs) do not extend upwards into the MLT and those that do remain poorly constrained. In this study, we use a longterm meteor radar wind dataset from Rothera (67° S, 68° W) on the Antarctic Peninsula to test the Whole Atmosphere Community Climate Model with thermosphereionosphere eXtension (WACCMX). This radar has an interferometer to determine meteor heights and has been running since 2005. This unique combination yields a dataset ideally suited to investigate interannual variability. We find that although some characteristic features in monthly median winds are represented well in WACCMX, the model exhibits significant biases. In particular, the observations reveal a ∼10 ms^{1} eastward wind at heights of 85–100 km in Antarctic winter, whereas the model predicts winds of the same magnitude but of opposite direction. We propose that this bias exists because WACCMX is missing eastward momentum forcing in the MLT from the breaking of secondary gravity waves.
Both the model and observations reveal significant interannual variability in monthly median winds. We investigate the role of particular key external phenomena in driving the winds in this region. These phenomena are; i) variations in Solar activity, ii) the El Nino Southern Oscillation (ENSO), iii) the QuasiBiennial Oscillation (QBO) and iv) the Southern Annular Mode (SAM). We use a linear regression method to investigate how the observed and modelled winds, and modelled gravity wave tendencies in the Antarctic MLT vary in relation to the indices that quantify these phenomena.
We find that there are some times of year and some height ranges at which there are significant correlations between the indices and the observed/modelled winds. In particular, in summer, there is a strong positive correlation in the modelled and observed zonal winds with the 11year Solar cycle of magnitude up to 9 ms^{1} per 70 Solar flux units. However, there appears to be little significant influence of the ENSO on the winds observed by the radar although WACCMX zonal winds display a negative correlation throughout January–February and a positive correlation during March–May. Results from the QBO indices are varied and we find differing correlations in the model and observations. Finally, we find a positive correlation between observed summertime zonal winds and the SAM which has a magnitude of 9 ms^{1} per 2.5 hPa change in the SAM index. However, in WACCMX zonal winds the summertime response is negative and around 10 ms^{1} per 2.5 hPa. The significance of this work lies in our quantifying the biases in a leading GCM and demonstrating there is significant interannual variability in both modelled and observed winds, some of which are consistent with the proposal of external forcing.
Phoebe Noble et al.
Status: closed

RC1: 'Comment on acp2022150', Anonymous Referee #1, 03 Jun 2022
Comments on the ACP manuscript 'Interannual variability of winds in the Antarctic mesosphere and lower thermosphere over Rothera (67°S, 68°W) in radar observations and WACCMX' by Noble et al.
Journal driven questions:
Does the paper address relevant scientific questions within the scope of ACP? Yes
Does the paper present novel concepts, ideas, tools, or data? Yes
Are substantial conclusions reached? Yes
Are the scientific methods and assumptions valid and clearly outlined? Mostly  see below
Are the results sufficient to support the interpretations and conclusions? Mostly  see below
Is the description of experiments and calculations sufficiently complete and precise to allow their reproduction by fellow scientists (traceability of results)? Mostly  see below
Do the authors give proper credit to related work and clearly indicate their own new/original contribution? Mostly  see below
Does the title clearly reflect the contents of the paper? The title does not reference consideration of linkages to climatological indices. But then, the title is already long enough.
Does the abstract provide a concise and complete summary? Yes
Is the overall presentation well structured and clear? Yes
Is the language fluent and precise? Mostly.
Are mathematical formulae, symbols, abbreviations, and units correctly defined and used? Mostly
Should any parts of the paper (text, formulae, figures, tables) be clarified, reduced, combined, or eliminated? No
Are the number and quality of references appropriate? Yes.
Is the amount and quality of supplementary material appropriate? None is supplied. That is appropriate.
This paper describes the annual cycle of winds above Rothera station as measured by a meteor radar and as modelled by WACCMX. The results of a regression analysis with a number of climatological indices are also described. Comparisons are made to other analyses and conclusions are drawn. The paper represents a useful contribution to our understanding of the polar MLT region and our ability to model it and warrants publication after the consideration of the points described below.
Some aspects of the analysis method are novel but their rationale are not fully described. The authors often use the median to describe the data. At line 102, they say this removes tides and planetary waves but do not say why this is so. Such wave phenomena can lead to velocity distributions that have strange shapes. The authors’ use of the median and the interdecile range to describe the data gets around this. In their comparisons with other results, though, they need to make sure the reader is cognizant of the use of different statistical measures.
In the abstract and the conclusions, the authors propose that a bias exists in WACCM zonal winds due to missing eastward momentum forcing. This conclusion can be supported on the time scale of a single time step (because on that scale, the model is using the forcing to define a grid point’s next velocity value). But on longer time scales, the dynamical equations include the influence of other forcings such as the Coriolis effect. If balanced flow was assumed, the zonal wind bias would be due to incorrect meridional forcing. The authors need to consider their proposal here more carefully.
In the introduction, the authors introduce WACCMX. A short description needs to be included here (Line 63).
The authors note that QBO10 and QBO30 are ‘orthogonal’ near line 134. They should provide an explanation of what they mean by this.
Near line 136, the authors say that the influence of the SAM is important for Antarctic winds. How do they know this is true or that the SAM does not influence nonAntarctic winds. Please rephrase.
In the discussion of multicolinearity in section 3.1.1, VIFs are introduced without a reference. It is also not clear what R is (it too needs a reference). An example on which dependent and independent variables contributed to the calculation of a VIF would perhaps clarify this section.
The authors should note the presence of the following two Antarctic observation and analysis papers in the context of both trends and linkages to climatological indices:
French W.J.R., Mulligan F.J., Klekociuk A.R.(2020) Analysis of 24 years of mesopause region OH rotational temperature observations at Davis, Antarctica – Part 1: longterm trends, Atmospheric Chemistry and Physics 63796394; doi:10.5194/acp2063792020
French W.J.R., Klekociuk A.R., Mulligan F.J. (2020) Analysis of 24 years of mesopause region OH rotational temperature observations at Davis, Antarctica – Part 2: Evidence of a quasiquadrennial oscillation (QQO) in the polar mesosphere, Atmospheric Chemistry and Physics 8691–8708; doi:10.5194/acp2086912020, 2020.;
The discussion of gravity wave tendencies in section 6 and near line 434 should include consideration of the relationship between GW tendency and the background wind. Wave breaking is affected by background wind conditions so a discussion of GW tendency needs to consider the existing relationships between climatological factors and the winds discussed earlier in the paper. The plot title on Fig 8a should include a word or symbol that denotes tendency. (It is misleading when read by itself as it is.)
Specific comments:
Near line 292 (first paragraph of section 5.2), the zonal 95100km ENSO significant zone (fig 6c) seems to extend into June.
Near line 349 – The authors should note that the 10 deg latitude difference between Scott Base and Rothera could affect comparisons.
Sentence starting L 443. At what height does the effect on GW forcing seen by Li et al (2016) occur?
L35 ‘constrain’ to ‘constraining’
L64 move parenthesis after ‘1992’ to after ‘(2009)’
L94 Suggest delete ‘In this section we discuss’
L130 Suggest small ‘s’ on ‘Solar’ and follow by a comma
L 182 replace ‘to test’ with ‘which’
L212 Start a new sentence after ‘Figure 3’
L282 Delete last ‘s’ in ‘westwards’
L293 delete ‘a’ before ‘south’
L294 ‘… WACCMX in either component’? This is now what I see. Please check wording.
L333 Do you mean QBO30 in this line?
L391 insert ‘by’ after ‘paper’
L406 Suggest replace ‘Although’ with ‘However’
L466 replace ‘affects’ with ‘effects’
L473 End sentence after ‘all heights’

RC2: 'Comment on acp2022150', Anonymous Referee #2, 07 Jun 2022
Interannual variability of winds in the Antarctic mesosphere and lower thermosphere over Rothera (67S, 68W) in radar observations and WACCMX
by Noble et al.
This paper presents results of longterm meteor wind radar observations at Rothera (67S, 68W), Antarctic, and also WACCMX modelling results for corresponding area and period, focusing on the interannual variability at MLT region. Monthly median zonal and meridional winds are displayed as timeheight crosssection between 2005 and 2020 and difference between observed and modelled wind is discussed. The authors also applied linear regression analysis for the period between 20052015 on five indices of Solar cycle, ENSO, (two heights of ) QBO, and SAM. The analysis is extended to gravity wave tendency (zonal drag).
The data set and analysis of longterm observation over the Antarctic station and corresponding WACCMX data are rare and valuable in understanding MLT dynamics and interannual variability. Results shown in ‘4. Results: The winds in radar observations and WACCMX’ is overall fine. However, the linear regression analysis results shown in ‘5. Results: Linear regression analysis’ need more careful consideration and evaluation.
My main concern is the number of samples in linear regression analysis. The sample for linear regression analysis is threemonth window times 11 years, i.e. 33 samples (data points) for each time and height bin. The estimated coefficient (β beta) is as many as six, as shown in Equation (1). I suppose this number of coefficient is too many for 11 years of data. The variance in these 33 samples is not only due to the interannual (or yeartoyear) variability, but also intraseasonal (or monthtomonth) variability, because variation in threemonth window is also included. Statistical significance of each regression coefficient must be overestimated because the interannual variability is not included in three sample in the same year window (of three month). Data number of 33 are not number of independent samples in the sense of interannual variation, and therefore freedom of 27 is overestimated. Multicollinearity (3.1.1) check by VIF should be carefully evaluated because under the presence of both intraseasonal and interannual variability VIF should be underestimated. Also, if the autocorrelation in 3.1.2 is calculated for all 11year timeseries, this will smooth out different autocorrelation function at different season (or threemonth window) so the value should be underestimated. Effect of such seasonal smoothing (or averaging) should affect DW test, and this should be checked. Therefore, my recommendation is that authors clarify the above questions before publishing the results of Chapter 5, 6 and 7.2.
Specific comments:
L 1213
‘These zonal gravitywave tendencies inWACCMX were found to be noisy when examined over the meteor collecting region. We therefore calculated tendencies as zonalmeans in a band of 300 km latitudinal width, centred over the latitude of Rothera.’
This assumes that the gravity wave drag is uniform at all longitude if monthly averaged. Is this an appropriate assumption? It is known that Andes and Antarctic Peninsula is the region of strong gravity wave generation. I do not think the zonally uniform assumption is correct.
L 255
‘In (a), it can be seen that the interdecile range of the zonal wind maximises over the summer, when the zonal wind reversal occurs, due to the considerable variability in the
strength and timing of this reversal.’
Please check whether this difference of the interdecile range could be due to the difference of zonal wind magnitude between summer and winter. If the fluctuation is a certain % of the amplitude, this also cause the variation of interdecile range measured by m/s.
L 2689
‘Hatched regions show where the relationship is statistically significant at the 90% level, using the Student’s ttest.’
I am suspicious about this, considering the point described as ‘main concern’ above.

CC1: 'Comment on acp2022150', Karanam Ramesh, 14 Jun 2022
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp2022150/acp2022150CC1supplement.pdf
 AC1: 'Authors response to reviewers and community comment', Phoebe Noble, 05 Aug 2022
Status: closed

RC1: 'Comment on acp2022150', Anonymous Referee #1, 03 Jun 2022
Comments on the ACP manuscript 'Interannual variability of winds in the Antarctic mesosphere and lower thermosphere over Rothera (67°S, 68°W) in radar observations and WACCMX' by Noble et al.
Journal driven questions:
Does the paper address relevant scientific questions within the scope of ACP? Yes
Does the paper present novel concepts, ideas, tools, or data? Yes
Are substantial conclusions reached? Yes
Are the scientific methods and assumptions valid and clearly outlined? Mostly  see below
Are the results sufficient to support the interpretations and conclusions? Mostly  see below
Is the description of experiments and calculations sufficiently complete and precise to allow their reproduction by fellow scientists (traceability of results)? Mostly  see below
Do the authors give proper credit to related work and clearly indicate their own new/original contribution? Mostly  see below
Does the title clearly reflect the contents of the paper? The title does not reference consideration of linkages to climatological indices. But then, the title is already long enough.
Does the abstract provide a concise and complete summary? Yes
Is the overall presentation well structured and clear? Yes
Is the language fluent and precise? Mostly.
Are mathematical formulae, symbols, abbreviations, and units correctly defined and used? Mostly
Should any parts of the paper (text, formulae, figures, tables) be clarified, reduced, combined, or eliminated? No
Are the number and quality of references appropriate? Yes.
Is the amount and quality of supplementary material appropriate? None is supplied. That is appropriate.
This paper describes the annual cycle of winds above Rothera station as measured by a meteor radar and as modelled by WACCMX. The results of a regression analysis with a number of climatological indices are also described. Comparisons are made to other analyses and conclusions are drawn. The paper represents a useful contribution to our understanding of the polar MLT region and our ability to model it and warrants publication after the consideration of the points described below.
Some aspects of the analysis method are novel but their rationale are not fully described. The authors often use the median to describe the data. At line 102, they say this removes tides and planetary waves but do not say why this is so. Such wave phenomena can lead to velocity distributions that have strange shapes. The authors’ use of the median and the interdecile range to describe the data gets around this. In their comparisons with other results, though, they need to make sure the reader is cognizant of the use of different statistical measures.
In the abstract and the conclusions, the authors propose that a bias exists in WACCM zonal winds due to missing eastward momentum forcing. This conclusion can be supported on the time scale of a single time step (because on that scale, the model is using the forcing to define a grid point’s next velocity value). But on longer time scales, the dynamical equations include the influence of other forcings such as the Coriolis effect. If balanced flow was assumed, the zonal wind bias would be due to incorrect meridional forcing. The authors need to consider their proposal here more carefully.
In the introduction, the authors introduce WACCMX. A short description needs to be included here (Line 63).
The authors note that QBO10 and QBO30 are ‘orthogonal’ near line 134. They should provide an explanation of what they mean by this.
Near line 136, the authors say that the influence of the SAM is important for Antarctic winds. How do they know this is true or that the SAM does not influence nonAntarctic winds. Please rephrase.
In the discussion of multicolinearity in section 3.1.1, VIFs are introduced without a reference. It is also not clear what R is (it too needs a reference). An example on which dependent and independent variables contributed to the calculation of a VIF would perhaps clarify this section.
The authors should note the presence of the following two Antarctic observation and analysis papers in the context of both trends and linkages to climatological indices:
French W.J.R., Mulligan F.J., Klekociuk A.R.(2020) Analysis of 24 years of mesopause region OH rotational temperature observations at Davis, Antarctica – Part 1: longterm trends, Atmospheric Chemistry and Physics 63796394; doi:10.5194/acp2063792020
French W.J.R., Klekociuk A.R., Mulligan F.J. (2020) Analysis of 24 years of mesopause region OH rotational temperature observations at Davis, Antarctica – Part 2: Evidence of a quasiquadrennial oscillation (QQO) in the polar mesosphere, Atmospheric Chemistry and Physics 8691–8708; doi:10.5194/acp2086912020, 2020.;
The discussion of gravity wave tendencies in section 6 and near line 434 should include consideration of the relationship between GW tendency and the background wind. Wave breaking is affected by background wind conditions so a discussion of GW tendency needs to consider the existing relationships between climatological factors and the winds discussed earlier in the paper. The plot title on Fig 8a should include a word or symbol that denotes tendency. (It is misleading when read by itself as it is.)
Specific comments:
Near line 292 (first paragraph of section 5.2), the zonal 95100km ENSO significant zone (fig 6c) seems to extend into June.
Near line 349 – The authors should note that the 10 deg latitude difference between Scott Base and Rothera could affect comparisons.
Sentence starting L 443. At what height does the effect on GW forcing seen by Li et al (2016) occur?
L35 ‘constrain’ to ‘constraining’
L64 move parenthesis after ‘1992’ to after ‘(2009)’
L94 Suggest delete ‘In this section we discuss’
L130 Suggest small ‘s’ on ‘Solar’ and follow by a comma
L 182 replace ‘to test’ with ‘which’
L212 Start a new sentence after ‘Figure 3’
L282 Delete last ‘s’ in ‘westwards’
L293 delete ‘a’ before ‘south’
L294 ‘… WACCMX in either component’? This is now what I see. Please check wording.
L333 Do you mean QBO30 in this line?
L391 insert ‘by’ after ‘paper’
L406 Suggest replace ‘Although’ with ‘However’
L466 replace ‘affects’ with ‘effects’
L473 End sentence after ‘all heights’

RC2: 'Comment on acp2022150', Anonymous Referee #2, 07 Jun 2022
Interannual variability of winds in the Antarctic mesosphere and lower thermosphere over Rothera (67S, 68W) in radar observations and WACCMX
by Noble et al.
This paper presents results of longterm meteor wind radar observations at Rothera (67S, 68W), Antarctic, and also WACCMX modelling results for corresponding area and period, focusing on the interannual variability at MLT region. Monthly median zonal and meridional winds are displayed as timeheight crosssection between 2005 and 2020 and difference between observed and modelled wind is discussed. The authors also applied linear regression analysis for the period between 20052015 on five indices of Solar cycle, ENSO, (two heights of ) QBO, and SAM. The analysis is extended to gravity wave tendency (zonal drag).
The data set and analysis of longterm observation over the Antarctic station and corresponding WACCMX data are rare and valuable in understanding MLT dynamics and interannual variability. Results shown in ‘4. Results: The winds in radar observations and WACCMX’ is overall fine. However, the linear regression analysis results shown in ‘5. Results: Linear regression analysis’ need more careful consideration and evaluation.
My main concern is the number of samples in linear regression analysis. The sample for linear regression analysis is threemonth window times 11 years, i.e. 33 samples (data points) for each time and height bin. The estimated coefficient (β beta) is as many as six, as shown in Equation (1). I suppose this number of coefficient is too many for 11 years of data. The variance in these 33 samples is not only due to the interannual (or yeartoyear) variability, but also intraseasonal (or monthtomonth) variability, because variation in threemonth window is also included. Statistical significance of each regression coefficient must be overestimated because the interannual variability is not included in three sample in the same year window (of three month). Data number of 33 are not number of independent samples in the sense of interannual variation, and therefore freedom of 27 is overestimated. Multicollinearity (3.1.1) check by VIF should be carefully evaluated because under the presence of both intraseasonal and interannual variability VIF should be underestimated. Also, if the autocorrelation in 3.1.2 is calculated for all 11year timeseries, this will smooth out different autocorrelation function at different season (or threemonth window) so the value should be underestimated. Effect of such seasonal smoothing (or averaging) should affect DW test, and this should be checked. Therefore, my recommendation is that authors clarify the above questions before publishing the results of Chapter 5, 6 and 7.2.
Specific comments:
L 1213
‘These zonal gravitywave tendencies inWACCMX were found to be noisy when examined over the meteor collecting region. We therefore calculated tendencies as zonalmeans in a band of 300 km latitudinal width, centred over the latitude of Rothera.’
This assumes that the gravity wave drag is uniform at all longitude if monthly averaged. Is this an appropriate assumption? It is known that Andes and Antarctic Peninsula is the region of strong gravity wave generation. I do not think the zonally uniform assumption is correct.
L 255
‘In (a), it can be seen that the interdecile range of the zonal wind maximises over the summer, when the zonal wind reversal occurs, due to the considerable variability in the
strength and timing of this reversal.’
Please check whether this difference of the interdecile range could be due to the difference of zonal wind magnitude between summer and winter. If the fluctuation is a certain % of the amplitude, this also cause the variation of interdecile range measured by m/s.
L 2689
‘Hatched regions show where the relationship is statistically significant at the 90% level, using the Student’s ttest.’
I am suspicious about this, considering the point described as ‘main concern’ above.

CC1: 'Comment on acp2022150', Karanam Ramesh, 14 Jun 2022
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp2022150/acp2022150CC1supplement.pdf
 AC1: 'Authors response to reviewers and community comment', Phoebe Noble, 05 Aug 2022
Phoebe Noble et al.
Phoebe Noble et al.
Viewed
HTML  XML  Total  BibTeX  EndNote  

333  84  14  431  4  3 
 HTML: 333
 PDF: 84
 XML: 14
 Total: 431
 BibTeX: 4
 EndNote: 3
Viewed (geographical distribution)
Country  #  Views  % 

Total:  0 
HTML:  0 
PDF:  0 
XML:  0 
 1