General comments

The authors propose a mechanism involving the interaction of the mesospheric semiannual oscillation (SAO) with the quasi-biennial oscillation (QBO) to explain the occurrence of SH weak vortex events. Overall I thought the authors’ mechanism certainly warrants consideration and they provide some thoughtful, though qualitative, evidence for the important of this interaction.  However, my main criticism is that there was little quantitative or statistical evidence provided to show that these processes are linked.  Additionally, the authors make some comments (e.g. line 10-11, 339) that argue the SAO/QBO configuration is the primary ingredient for these somewhat rare SH events to occur, without really discussing the fact that tropospheric wave driving was extraordinarily large, for example, in August 2019 (Rao et al. 2020).  For these reasons and the other major comments below, I suggest a major revision.

Major Comments:

• A general recommendation for the authors is to be careful or more exacting with their wording regarding this event- for example about calling it “the second SSW” (line 3, line 27-28, line 325). This event was not a major SSW (by technical criteria) like 2002, and if you mean to refer to minor events, there have been other minor SH events before too as mentioned on line 57 (though none as large as this one in terms of temperature).  Maybe instead for line 3, you could state something like “while this was only the second time stratospheric temperatures have risen this rapidly in the Southern Hemisphere (SH) winter/spring”.  Also the number of 6 events per decade in the NH (or one almost every other year) is mentioned a few times but again that’s referring to events that meet the major SSW criteria so it’s not really comparable.  The authors do try to clarify this point on line 55 but it would improve the text to be careful with wording throughout (or move the statement on line 55 to earlier to try to be clear from the beginning). I would disagree with the statement on line 61 as well, as 2019 saw no wind reversal at 60S and 10 mb; additionally, there is evidence to suggest that dynamics were quite different between the 2002 and 2019 warmings, so I’d rephrase.

• There needs to be some more analysis to connect the changes in the critical line/wave propagation in the subtropics with the momentum changes happening in the polar regions. The authors have a couple nice figures (Figure 3,4, and 6) qualitatively showing a potential connection.  But what’s not clear is if the momentum changes over the polar region are just happening simultaneously as the changes in the critical line, or if these things are truly related. Perhaps there is some way to examine the correlation between the momentum flux convergence near 60S/1 hPa with some measure of the critical line at 1 hPa (latitudinal location for example), for all years in the record during July, and see if 2019 and 2002 correspond to exceptional conditions on that scatterplot.  This might also provide better quantification for how other years are different (e.g. Figure 9). Similarly, in section 3.3.2 there is a lot of qualitative discussion of the SAO-QBO interaction (e.g. lines 243, 250) but it would greatly improve the arguments if this relationship and its link to the polar vortex could be better quantified.  For example, is the latitude of the SAO-QBO interaction the most important feature? The vertical extent? The magnitude of the SAO/QBO winds? These factors are brought up on lines 260-261 and 340 but can these be quantified in a more robust statistical manner?

• It wasn’t clear to me why it was necessary to break the averaging into MERRA-2 streams. Is there justification for doing this? Maybe there are large jumps introduced (e.g., Long et al. 2017), but I wouldn’t think it would have such a substantial effect on weekly averages. This made for a lot of unnecessary figures in my opinion, and it is not clear that a lot of the differences between, e.g. Figures 9, 10, and 11, aren’t just due to sampling (only a few cases in each composite)- thus this was not particularly meaningful for me to be able to draw any conclusions. Why not just composite all the non-weak vortex years together?  Or just pick a couple of years of interest and just show those individual years? (the averaging could be introducing differences just from differences in the timing of SAO-QBO interaction between years, for example). Similarly, the statements on lines 257-260 and lines 281-285 are implying some decadal differences based on arbitrary averaging over streams, but I’m not sure these statements are justified (what would be a physical explanation for why these relationships would be different over time?).

• The authors make an interesting case for the role of mesosphere-stratosphere interactions, but I think they ignore too much the role of tropospheric forcing and/or internal stratospheric processes. For example I think Figure 11 shows cases in which the SAO-QBO interaction is very similar to 2019 and 2002 yet no SSWs happened- why? Particularly during 2019 when tropospheric wave forcing was at record high levels, I think this is a mistake to not mention this or explore the combined effect of these two features- perhaps the mesospheric wind interaction is a necessary but not sufficient condition for a SSW to occur (or likewise for the tropospheric wave driving; perhaps you need both factors). For this reason I think statements such as 338-339 are overstated, and should be reworded unless stronger evidence is provided that this interaction can actually “trigger” the event.

Specific comments

Line 2, also line 30: specify the level/latitudes that this wind reversal and temperature rise occurred at (or at least something general like “polar cap” and “mid-stratosphere”).

Line 4, line 62: This is one important aspect that is thought to drive SSWs but one difficulty is that the only known SSW in the SH in 2002 was likely not due to tropospheric wave driving but resonant amplification internal to the stratosphere (see Albers and Birner 2014, Esler papers). A number of NH events are also thought to be caused by this, with the link to tropospheric wave forcing not always prevalent (de la Camara et al. 2019). So either also mention this mechanism or caveat these statements so that it’s clear this is not the only possible mechanism.

Line 33-35: This is true but the other key dynamical difference between a final warming and a sudden warming is that for the SSW the vortex recovers back to westerlies afterwards.

Line 31, 34: Here, and throughout, there are a lot of textbook references for things that are more or less “facts”- I don’t think every statement that is considered general knowledge needs a reference (or at least, there might be more appropriate references to cite; for example on line 34 & 39 I suggest instead the review on SSWs, Baldwin et al. 2020 in Reviews of Geophysics).

Line 47-49: Rephrase; they don’t have to be easterlies to descend to the troposphere; stratosphere-troposphere coupling occurs when the winds are anomalously strong, e.g. positive NAM, as well.

Line 54, and line 160: From Figure 2d, I’d argue that the vortex is not centered over the pole once the SSW occurs (even at 40 hPa it’s fairly displaced).

Line 63: it’s not really the North pole but the middle latitudes where the mountain ranges matter for wave forcing.

Line 76, and line 101: easterly QBO at what level? Note that a lot of the Holton-Tan relationships are looking at QBO at 40-50 hPa, not 10 hPa as done in this study.

Line 82: westerlies maximise at what level?

Line 140: It’s not clear if “all experienced mesospheric wind reversals in October” is meant to refer to 1988 and 2017 or to the other years identified- also why October, isn’t this when the final warming or seasonal transition begins to occur? Seems late compared to 2002 and 2019.

Lines 143-145: For a list of years when the SH polar vortex was weak, see Lim et al. 2019 (https://www.nature.com/articles/s41561-019-0456-x)

Line 168-169: “as may happen with wQBO”- can you clarify what is meant here?

Figure 5, figure 7: 15-20S is a narrow latitude band- which makes it difficult to see the easterly SAO winds in the tropics.  What about 0-30S or 0-20S instead? Perhaps that would make the descent between the SAO and the QBO also clearer. The “downward connection” between the SAO and QBO in these figures was not particularly compelling for that reason; particularly for 2002, there only seemed to be some interaction for a week in mid-June, when the SSW didn’t happen until late September (the connection is easier to see for this reason in Figure 6).

Line 225, 276: Figure 8d- there’s not a wind reversal near 60S except above 1 hPa; and isn’t this just the final warming/season transition? It’s not clear to me the utility of showing late October for this reason since it seems likely there are different processes at play (radiative changes driving the seasonal cycle).  Why not show what Sept looked like instead (as in Figure 3 and 6) to show what was different about 2017 so that it didn’t result in a warming, given otherwise strong similarities with the critical line/QBO/SAO in June?

Line 257: smaller in what way? Can this be quantified? (see major comments)

Line 265: I don’t understand why there are parentheses around polar and low-latitude, but I’m also not sure this sentence is clear.

Line 283-4: “We didn’t find evidence of easterly momentum being deposited throughout the winter as we did for 2002 and 2019”; this is a bit circular, but it also shows the issue with causality in this study.  The other years didn’t have build up of easterly momentum and there was no SSW; but is this a reflection of differences in the SAO-QBO relationship or the fact that there was no SSW due to other processes? This lack of causality is reflected in the statement on line 286, where you mention the years you did have an SAO-QBO interaction but there was no easterly momentum build up or wind reversal.

Minor discussion comment/line 303-304: Some attempts have been made to look at the upper stratosphere in the SH winter to make predictions for the following spring which should be acknowledged here:

Lim, E.-P., Hendon, H.H. & Thompson, D.W.J., 2018. Seasonal Evolution of Stratosphere-Troposphere Coupling in the Southern Hemisphere and Implications for the Predictability of Surface Climate. Journal of Geophysical Research: Atmospheres, doi: 10.1029/2018JD029321.

Byrne, N.J. & Shepherd, T.G., 2018. Seasonal persistence of circulation anomalies in the Southern Hemisphere stratosphere, and its implications for the troposphere. Journal of Climate, doi: 10.1175/JCLI-D-17-0557.1.

Line 305, 344: 20-30 days is not typical for SSW predictability (maybe a handful of events in the record have been predictable on timescales that long).  Refer to Domeisen et al. 2020 (part I)- not the one that is cited in the next sentence- typical predictability limits are 10-15 days in the NH.  However, it might turn out that SH SSWs are predictable at longer leads, if they are more influenced by the mesosphere/stratosphere interaction than the NH.

Line 316: One question I had was why was the focus of this study on the SH rather than the NH (which would greatly increase your sample size?). Is this something presented already in the Gray et al. (2020) paper? Perhaps it should be stated somewhere why the focus here is on the SH (maybe it is and I missed it). In terms of the QBO, ENSO, and MJO influence it should be mentioned that the Rao et al. 2020 paper considers several of these factors on the 2019 SH warming.

Line 320: what is meant by the “MJO index was positive”? (the amplitudes of the EOF1/2 in the MJO phase space were positive?) What was the phase?

Line 348: suggest non-gendered associations such as “their boreal counterpart”.

Line 347-348: Are their anti-symmetrical aspects of the QBO/SAO across hemispheres that could explain a difference between NH and SH frequency of SSWs? To me there seems an obvious difference in tropospheric wave driving between the two, but if there are clear differences in the upper atmosphere that could explain frequency differences it would be interesting to describe or speculate on here.

Technical corrections

Line 22: “Annual” should be “Annular”

Line 42: should be “forming a ‘comma shape’”

Line 46: replace “intrusion” (which has a very specific meaning tied to stratosphere-troposphere exchange) with “mixing”

Line 50: correct misspelling of “split”

Line 57: remove “of”

Line 63: replace “geography” with “topography”or “orography” perhaps

Line 85: change to “roughly in June”

Line 117: change “were” to “where”

Line 139: change “left out of these groups” to “considered separately”

Line 146, 170, 175, 294, 329: change “deposit” to “deposition”

Line 182: change “form” to “from”

Line 187: change to “western side of Antarctica”

Line 192: change to “before it connects”

Line 223: not sure what is meant by “subsides” in this context- “moves down”?

Line 253: change “on” to “of”

Line 261: change “there” to “their”

Line 321: change “difference” to “differences”

Line 331: change “influencing” to  “influences”

Review of “Does the coupling of the mesospheric semiannual oscillation with the quasi-biennial oscillation provide predictability of Antarctic sudden stratospheric warmings?” by Nordström and Seppälä

The goal of this study is to show that zonal winds in the tropical stratosphere affected the development of the Southern Hemisphere minor sudden stratospheric warming (SSW) in 2019. The results are emphasized by comparisons of this winter with dynamical developments during two other disturbed SH winters and during quiet winters. This is an interesting topic and is timely since it investigates developments during the recent 2019 winter, which showed an enhanced level of disturbance that is rarely seen in the SH.

Overall, the paper lays out the authors’ argument in a straightforward manner. However, the language is often lacking in precision so some steps along the way are over-simplified or not carefully described. The net effect comes across as a very qualitative description in which complexities are ignored or glossed over; at times the discussion is not well-grounded in dynamical theory. This could be remedied with an extensive revision focused on more careful description of the links between cause and effect, better definition of terms, and overall effort to avoid generalizations that do not hold up. It would also be very helpful if the figures were revised to better illustrate the points being made in the text. The comments below elaborate on these concerns. The aspects that are of most concern to this reviewer are those given in major comments 1, 4, and 6.

Major comments

1. The paper is focused on the way that variations in tropical winds might contribute to the deceleration of the zonal winds in midlatitudes. The potential impact of the disturbed dynamics on the tropical wind is not discussed. The accepted understanding is that the stratopause SAO easterly winds in solstice periods develop in response to the global Brewer-Dobson circulation that is driven by planetary wave dissipation in the winter hemisphere. In other words, the anomalous EP flux divergence in particular southern winters could contribute to stronger SAO easterlies during those winters. That is not to say that the tropical winds cannot also affect the midlatitude dynamical disturbance but it is not appropriate to treat it as a one-way influence. The direction of influence is likely to be two-way; cause and effect cannot be easily separated.
2. Section 2.1: The paper by Kawatani et al (2020), which was cited in the manuscript, shows that the equatorial winds in the upper stratosphere in MERRA-2 have significant deficiencies. The winds in the lower mesosphere have even poorer agreement with observations. Some mention of these discrepancies is necessary.
3. It is often hard to see the features being discussed in the latitude x pressure figures (3, 4, 6, etc.). Can you make them wider in the latitude dimension by, for example, stacking two over two and/or cutting off the northern latitudes?
4. Besides being difficult to discern in the plots, the text describing the easterly winds is difficult to follow. One problem is labeling all easterly winds to be part of the QBO or SAO. For example (l. 218) “eQBO comes to lay between 50-2 hPa extending to about 40°S”. The QBO is an oscillation in the tropics; the term is also sometimes applied to periodic signals elsewhere that could be affected by the tropical oscillation. What evidence do you have that this individual instances of easterly winds in mid-latitudes is part of the QBO rather than a response to other dynamical activity? It is good to keep in mind that the QBO and SAO are oscillations that are defined by a timeseries of winds at a given location. Take care when referring to winds with the same sign seen on a single occasion elsewhere in the atmosphere.
5. Figure 4 and discussion: The EP flux and its impact on the zonal wind are derived using zonal means and waves defined as perturbations from that mean. They cannot be applied to an arbitrary longitude sector. To look at the regional wave-mean flow process, you should apply analysis tools specifically developed for this purpose, such as the formulation given by Plumb (1985). I suggest to remove this figure and delete the discussion.
6. Please define what you mean by “coupling of the SAO and QBO wind signals” used in reference to Figure 5. Do you mean to indicate that some dynamical interaction is “coupling” these winds? If so, please describe what it is. The terms “interact” and “interaction” are also used in the discussion of several other figures and in the discussion section but the nature of this interaction and the evidence that it is occurring are never presented.
7. The speculation in the paragraph beginning at line 316 has several problems. First, it ignores the role that dynamics in the winter hemisphere have in driving the SAO and affecting its variability. Second, to support the speculation about the contrast with the NH, it is necessary to show or cite evidence that the development of the SAO is different there. This is not done.

Minor comments:

1. The reason for analyzing the results from quiet years in streams is not clear. Do we learn anything from this that would not be equally apparent from treating all the quiet winters together? If so, please describe what it is that we learn and what about the different streams accounts for the differences. If not, it would be useful to eliminate this as an unnecessary complication.
2. (l. 36-37) “a reversal of the meridional temperature gradient, creating an easterly zonal wind” The meridional temperature gradient and vertical wind gradients are consistent. Be careful about saying that one is causing the other unless you have evidence that only one is being forced.
3. (l. 46-47) “These clouds are where ozone depleting reactions occur (Solomon, 1999)” Not really. Reactions within clouds produce reactive species that then destroy ozone in the presence of sunlight.
4. (l. 64-67) “The North Pole is ringed by mountain ranges, perfect for producing atmospheric waves (Duck et al., 2001). An enhancement of wave activity over winter causes disruption to the vortex, as the waves deposit their momentum at higher altitudes (Brasseur, 2005). However, Antarctica is enclosed by flat oceans, which don’t excite waves as effectively as mountains (Holton, 2012).” It is not necessarily the topography near the pole that is associated with planetary wave generation. It would be better just to say that the NH has higher winter planetary wave activity and variability than the SH.
5. (l. 86-87) “Westerlies maximise close to the equinoxes, whilst the easterlies maximise near the solstices (Brasseur, 2005)” Please be clear that this description applies for the SAO near the stratopause, not above.
6. (l. 138-139) “if the 10 hPa equatorial flow is easterly during the austral winter months of June/July” Do you mean if the monthly mean values are easterly for these two months or if the values are easterly for every day?
7. (l. 156-157, l. 181, and elsewhere) “As planetary scale waves can only propagate where the zonal flow is westerly (eastward)” This should refer specifically to stationary waves. To be precise, insert “stationary” and also verify that the wave or waves that are driving the SSW under investigation are themselves quasi-stationary.
8. 1 and eq. 2 are the quasi-geostrophic version of EP flux. The full values derived from the primitive equations would be preferable.
9. (l. 189) “the zero wind line formed by the QBO and SAO subsides” What do you mean by subsides?

reference:

Plumb, RA (1985) On the three-dimensional propagation of stationary waves, J. Atmos. Sci., v 42, p 217-229.

Review on:

“Does the coupling of the mesospheric semiannual oscillation with the quasi-biennial

oscillation provide predictability of Antarctic sudden stratospheric warmings?”,

ACPD, 2021, by Nordström, V. J. and Seppälä, A..

In their paper, Nordström and Seppälä present a coupling of SAO and QBO phases that has

occured both during the 2002 and during the 2019 Antarctic SSW events. The authors describe

some characteristics of the two wind phenomena that, in combination, were special in these two

years and hence may be crucial for the generation of the SSW events. In other years of similar

QBO and SAO phase combinations, some differences can always be pointed out, but the method

of doing so is not convincing. The conclusions of the paper are all in all pretty vague, and I think

they could, with little effort, be concretised such that a clear statement stands at the end of the

paper. I think the study presents an interesting mechanism that has the potential to add a fair bit

of understanding to the SSW phenomenon and I recommend publication in ACP. However, there

are a couple of major and numerous minor points that I would like to see addressed beforehand

and also the presentation quality requires some improvement, especially because the language

is often too unprecise. Please consider the list of major, minor and technical issues below for that.

Major issues:

• I do not understand how it can be fair in this study to compare in Figs. 9-12 an average

over several years with particular years. This way you can average out particular features

of certain years or even create artefacts in the averages. The dates that are chosen appear

rather arbitrary and not well defined. If they are specific for each year (which seems to

be the case for the two SSW years), how can they fit to the averages over several years? I

also did not really undestand how it makes sense to put these particular years into one pot

(the use of the ’streams’). I guess this requires some good explaining or serious revision

of the method.

• In some of the additional years, the SAO-QBO interaction takes place, but then the signal

is not moving down all the way. Too often in the paper, it is stated that eQBO and eSAO

come together, but then this does not result in change in deposition of wave forcing/wind

reversal/... (see 301-302 as example) but then, the reason for that is not discussed. This

happens throughout the paper, it always leaves the reader behind puzzled with nothing

in his hands. Some more analysis or discussion to at least point out the one or the other

mechanism that could be responsible for this would be very helpful.

• In the analysis, gravity wave (GW) drag is completely missing. GWs can also largely

contribute to the overall wind forcing in these areas. These data should be available from

MERRA-2 as well. That should be added to the resolved wave drag and the contours

should be plotted more nuanced. As is, the caption does not even provide information

on the values of the drag and there is only one (or no) contour line. Possibly, clearer

conclusions can be drwan on the wave activity statements with this done.

• The paper fails to make any quantitative statements about the events. In consideration of

the few examples, it is clear that this cannot be made in detail. However, some more effort

could be made to achieve something in that direction. One could maybe make statements

about when it is too late in winter for the two signals to fall together for generating an SSW

as seen from example years where this happened. Or the magnitude that is reached could

be described better. For example, in L269-270 a ’large’ eQBO and SAO is mentioned but

not even stating what ’large’ means. Is it the strength or the expansion of the easterly winds

that matters? With statements that are a bit clearer, the conclusions of the paper could step

up a little from what is now stamped by a conditional that reflects in the word “may” too

often (see e.g. L316-). The mechanism that is described in the paper is quite clear and

interesting, so it is a little disappointing that at the end the conclusions are so vague. I

think this could be improved fairly easily for example by some statistical analyses.

Minor issues:

• L10: “these features“. Please name them.

• L28: When did record keeping begin?

• L31-32: This is a bit oversimplified for my taste, please don’t forget about PSC processes

here. (below you mention it then) Moreover, use the common wording ’ozone depleting

substances’. These substances are there, no matter if ”trapped” or not. Please remove the

word “trapped”.

• L55: ’at higher altitudes’ and ’at lower altitudes’ in one sentence confuses me. Please

rephrase.

• L61: Rephrase to: Hence, the dynamic situation in 2010 was unlike the situations in 2002

and 2019, because in the latter two years, rapid warmings and wind reversals occured.

• L96: What does ’well’ mean here? Realistically or at all or ...?

• L104-109: I do not think this kind of summary belongs here, rather state how the paper is

structured at this point.

• For my taste, the introduction (and also the methods section) is too long. I do appreciate

the detailed information, but I think the authors could try to somewhat compress it. Some-

times it is a bit of a back and forth and moreover, I am not sure if all the information is

really needed for the paper.

• L136: Remove: ’in zonal winds’

• Fig. 3 caption: Denote the magnitude of the EPfc line!

• L160: cite here: Middle atmosphere dynamics, Andrews, David G and Holton, James R

and Leovy, Conway B, 40, doi:10.1002/qj.49711548612, 1987, Academic press

• L195-196: “reverse further down than the mesosphere“ I do not understand what you want

to say here. Can you rephrase that please.

• L200-2005: Rephrase to: In Fig. 4, we average Fig. 3d according to the two geopotential

anomaly patterns from Fig. 1, namely clockwise from 40..... This way, we find ....

But actually, do you conclude anything from these figures? Or will you refer to this

lateron? If so, state it. If not, you can remove this bit, because it shows simply what you

would expect.

• L206-209 and associated figure: That is a nice depiction too, but not needed for under-

standing. If no conclusions are drawn from this here, move it to where it is needed, or

remove it. (but see below)

• L218-219: rephrase ’comes to lay’ and please the entire sentence, it is very confusing

• L221: Rephrase to: Fig. 6d shows that the polar winds revers between ...

• L223-225 If the purpose is to contrast Fig. 5 and Fig. 7, show the two figures together

here and describe differences. Wouldn’t it be interesting to extend these time series to

when the SSWs happen to see the entire evolution? Or is not so much happening in these

latitudes then?

• Remove subsection headers 3.3.1 and 3.3.2

• L235: Fig 8 is as.... remove the sentence

• Fig. 9-12: I cannot see any EP flux convergence contours in these plots (exception is 12b).

I can hardly imagine that there is none anywhere there. So how come?

• L249: ’as before’ What are you referring to exactly? Be precise.

• L256: How can you expect the timing to be so similar to 2002 or 2019, when even the

timings of 2002 and 2019 are not really similar. How similar does it have to be?

• L273-275: ’smaller’ and ’larger’. In what sense? Expansion, or strength, both? Be

specific.

• L273: Rephrase to: ...QBO and SAO and the two patterns only occasionally interacted....

But actually, I do not like this statement at all. How can you summarise over whole

decades, when each year is so specific and has its very own dynamics and each feature

can be of importance. This refers to my main point, why I think the averaging over several

years is not feasible in this study.

• L280: Do you mean: In both years, the SAO is in its easterly phase with winds extending

into the SH above 1 hPa... ? (also L349)

• L287: The ’However’ here confuses me. Do you want to express ’in contrast’? But I don’t

see how these two situations stand in contrast. Can you resolve this?

• L288: What does ’links across latitudes’ mean?

• L308: I guess you mean stratospheric extreme events? As is it appears like referring to

troposphere. Better simply write SSWs.

• L317-320: The reason that SSWs are less common in the SH is not explained here! I

reckon the eQBO and eSAO co-occur about as often in both hemispheres, and then the

wave guide is changed similarly in both hemispheres. But what is different is the amount

waves that then uses that waveguide and makes its way up to the mesosphere. In the

SH there are generally much less/weaker waves and hence the perturbation in the upper

atmosphere is weaker.

Technical issues:

• L30: What do you mean by ’descend’ here? I guess a different word could be more

suitable.

• L21: Change ’aftermath’ to ’influence’

• L59: Change ’had’ to ’included’

• L66: Change ’However’ to ’In contrast’

• L67: remove ’flat’ and change ’don’t’ to ’do not’

• L83: change ’a switching’ to ’ and ’alternation’

• L96: change ’found’ to ’showed’

• L99: Do you mean forced QBO and SAO-like or forced QBO- and SAO-like or...?

• L101: change to: ... we are here following suggestions that the upper atmosphere may be

key to understand the...

• L102: draw

• L104: In the present study, we analyze the interactions of the ...

• L114: change ’vertical levels’ to ’levels in the vertical’

• L122: change ’were’ to ’where’

• L124: remove one ’average’

• L133: Analogously

• L134: Rao et al. (2020)

• L136: ...by contrasting the zonal winds to sonde radiosonde measurements from Singa-

pore...

• L154: MERRA-2

• L154: deposition

• L160: Change ’*’ to ’·’

• L166: ...correspont to the ...

• caption Fig. 1: ... at 2 hPa over ... and remove last sentence

• L183: In the austral winter of 2019, the QBO was in ...

• L189: ’comes to sit’ ⇒ ’remains’

• L198: form ⇒ from

• L218: between 50 and 2 hPa

• L238 and L241: (Figure 8b)

• L245: ... for these years.

• L272: reversals

• L287: In 2002, the zonal mean....

• L305 ...was it necessary...

• L321 heed?

• L333-: Please move these links to the references or so.

Comments on “Does the coupling of the mesospheric semiannual oscillation with the quasi-biennial oscillation provide predictability of Antarctic sudden stratospheric warmings?” by Nordström and Seppälä (2021).

Summary

Nordström and Seppälä (2021) used the MERRA-2 reanalysis to study the possible impact of the Quasi Biennial Oscillation and the Semiannual Oscillation on the SH SSWs in 2002 and 2019. They proposed that the interaction between QBO and SAO can improve the predictability of the SH SSW. However, such a link between the QBO-SAO combined interaction and the predictability of the SH SSW is not stable and not applicable to other year with similar QBO-SAO configuration in the upper stratosphere and mesosphere. Most of the results are fairly descriptive and different SH stratospheric responses to the QBO-SAO in different years are not well explained. Therefore, I would possibly recommend publication of the paper after a major revision is performed with the following comments considered.

Major comments

1. The introduction has a large bias toward the studies of QBO in literature. The authors should present a comprehensive review on the recent studies of the QBO in literature. Can the authors also give a complete review on the modelling of the QBO and possibly the SAO in literature? In addition, the QBO simulations from CMIP5/6 models have made a big progress recently, and the authors did not review in this paper (several JC, GRL, QJ, JGR papers report the simulation of the QBO in CMIP5/6 models, please refer to them for details). The Holton-Tan relation has also been assessed for the CMIP5/6 models in literature. I suggest the authors to read those new publications.

2. L68-73: Factors influencing the occurrence of SSW in SH and NH have been reported in some papers, and summarized in Baldwin et al. (2021). Those factors include the QBO, ENSO, MJO, Solar cycle, and extratropical blockings. The authors did not read those original papers but see the summary in Baldwin et al. (2021). It is an efficient way to get an overall image on the latest study progress of SSW by reading the review paper Baldwin et al. (2021), but more details should be traced to the original paper. See the original paper (https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2019JD030826) Tables 1-3 for more details.

3. The structure of the paper should be well trimmed and organized. The paper tries to emphasize the impact of the QBO and SAO, but the timeseries of the QBO and SAO are not shown in the paper. Show the QBO index at 10 hPa and the SAO index at 1 hPa at least. From the display of the paper, the maximum SAO westerlies are much stronger than the maximum SAO easterlies in some years. Readers do not know the SAO zonal wind asymmetry in other years. Further, the authors display some redundant figures (9-12). Is it necessary to composite the SSW group in different decades (or data streams in MERRA-2, L118-121)? Please condense the paper with your largest endeavor.

4. Some papers might not be well understood by the authors. Gray et al. (2020) indeed tried to improve the prediction (rather than simulation or modelling, L304, L353) of the extreme polar vortex by considering the high-level impact.

Other comments

1. L28: has occurred => have occurred

2. L33: The strong polar vortex in March 2020 and the Arctic ozone loss have also been reported (https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020JD033524).

3. L44: The impact of SSW is wide, from polar region to midlatitudes. Please well read the reference and use the correct word.

4. L49: The relationship between SSW and NAM in the NH is discussed for S2S model, and the uncertainty is also present (https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2019JD031919).

5. L51: Here a reference should be added, like Allen et al. (2003) https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2003GL017117

6. L68-73, L81: Please see the discussion in Rao et al. (2020) for the possible impact of ENSO, QBO, MJO, Solar cycle on the SSW.

7. L74: east and westward => eastward and westward

8. L124: average average => (delete one)

9. L135: is discussed => are …

10. L146: The author might use a composite analysis. Please write more specific.

11. L151: Only two SSWs are commonly reported for the SH in record. There was no SH SSW in 1988. (also see your L340)

12. L159, 161: there are two “theta” in the formulas, the authors should change the second one to “phi” (latitude). Please well read the book or reference and avoid those unnecessary errors.

13. L163-164: What mean did the authors use? Zonal mean? Or time mean? Please be more specific.

14. L166: correspond => correspond to

15. L167: What is the “upcoming EP flux”?

16. L190: What latitude band do you describe?

17. L198: form => from??

18. L207-209: Could you provide the timeseries of the SAO from 1979 – nowadays?

19. L210-225: This paragraph is too descriptive.

20. L228-230: How regular is the SAO? See the major comment and provide the timesires of the SAO.

21. L238: Figure 8b should be put in a pair of parentheses.

22. L240: grammar error “is does”

23. L242-L277: This subsection is too descriptive and should be condensed.

24. L317-319, L364: I don’t think the SAO can help to explain why the SSW in SH is much less than the NH. This is mainly caused by the land-sea distribution responsible for the forced planetary waves.

25. L320: It is nearly impossible to forecast an SSW with 2-3 months in advance.

26. L326-327: The QBO is still a challenge for forecast systems. Not to mention the SAO. Please refer to https://journals.ametsoc.org/view/journals/clim/33/20/jcliD200024.xml

27. L330: read the original paper mentioned by Baldwin et al. (2021) for details.

28. L334-335: What do the authors mean “the MJO index was positive”? There are two MJO modes, and their combination presents a phase space. We use phases 1-8 to describe the MJO rather than “positive/negative”.

29. L344: into the SH => into the SH extratropics.

30. L347-348: Incomplete sentence. It is a long phrase.