A synoptic-dynamic interpretation of the Southern Annular Mode

Barnes, Michael A.; Risbey, James S.; Parker, Teresa J.; Earl-Jones, Nicholas; Tozer, Carly; Monselesan, Didier P.

doi:10.5194/egusphere-2026-3141

Preprints

https://doi.org/10.5194/egusphere-2026-3141

Preprints

17 Jun 2026

| 17 Jun 2026

A synoptic-dynamic interpretation of the Southern Annular Mode

Michael A. Barnes, James S. Risbey, Teresa J. Parker, Nicholas Earl-Jones, Carly Tozer, and Didier P. Monselesan

Abstract. In this study, we describe the Southern Annular Mode (SAM) as a function of regional extratropical variability in order to understand the synoptic-dynamic features associated with it. The SAM is often considered to be the leading mode of variability in the extratropics and is used to interpret changes and impacts in the large-scale mid-latitude circulation and surface weather in the Southern Hemisphere. This interpretation of the SAM has however been questioned in recent years, given the SAM's lack of annularity and weak correlation to surface weather and synoptic features. From this regional perspective, we once again highlight the non-annularity of the SAM and show explicitly how these asymmetries can result in the misinterpretation and misattribution of regional surface weather impacts. Through analysing a set of weather features including Rossby wave breaking (RWB) zones, cyclones and jets, we show the SAM can be interpreted as a function of the RWB characteristics across the hemisphere, in line with similar perspectives of the North Atlantic Oscillation. In fact, changes in the storm track and jets occur locally and non-annularly across the hemisphere in response to regional RWB, in contrast to the annular view assumed by the SAM. This interpretation provides opportunity for further understanding of present and future Southern Hemispheric extratropical variability and its associated surface weather impacts.

Received: 01 Jun 2026 – Discussion started: 17 Jun 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Michael A. Barnes, James S. Risbey, Teresa J. Parker, Nicholas Earl-Jones, Carly Tozer, and Didier P. Monselesan

Status: final response (author comments only)

RC1: 'Comment on egusphere-2026-3141', Anonymous Referee #1, 08 Jul 2026

The study investigates the link between the commonly used and hemispheric-wide definition of the Southern Annular Mode (hSAM), and a very local definition of the Southern Annular Mode (rSAM). Both definitions include a zonal mean pressure difference between two latitudes 40°S and 65°S. hSAM is defined by a zonal mean over all longitudes while rSAM is defined by a zonal mean over a small sector of 10° longitude. One main result is to show that hSAM is dependent on a multi linear combinations of all rSAMs. The most intense hSAM phases correspond to rSAM phases of the same sign in most regions but moderate hSAM phases are less dominated by the equivalent phase in rSAM. This result is not surprising but is well quantified. As expected from the link between spatial and temporal time scales in the atmosphere, the study also shows that the relation between hSAM and the rSAM is closer at longer time scales (monthly, seasonal) than at daily time scales. Then the link between Rossby wave breaking, jets, cyclones and rSAM is further detailed in the rest of the paper. An illustration of the weak link between rSAMs in some regions and moderate hSAM phases is done by showing two examples at the end of the paper. The figures have high quality, the text is clear in most part of the paper, and the sections well organized. I am not convinced the paper clearly brings very new results, it rather illustrates with clean figures and diagnostics the relation between synoptic-scale variability and the hemispheric-wide SAM. The most important concern I have is that the whole paper is based on an instantaneous view of the relationship between rSAM and hSAM and there is no attempt to illustrate potential time-lags links. For instance, it is well-known that Rossby wave breakings are associated with a transfer of energy from synoptic to planetary scales and this transfer may take some days as illustrated in Benedict et al or Franzke et al papers mentioned in the study. Also I do not think rSAM is a well chosen acronym. It is confusing because A in SAM means annular but the diagnostic is local and not annular at all. I therefore recommend publication of the paper after a major revision.
Main comments:
1) About time-lags links.

a) In the composites of rSAM and RWBs of Figs. 6 and 8, sophisticated diagnostics based on the identification of jet streaks and cyclones are shown, but it would be important to simply investigate time lag composites of the zonal wind. Indeed one can expect that after synoptic events like rSAM or RWB, some important zonal wind changes may appear at larger scales. For instance, Benedict et al. (2004) show that following RWB, the anomalies get larger and larger and finally project onto NAO anomalies. One could think that RWB events in the Southern Hemisphere could lead to pronounced zonal wind anomalies at large scales. This question seems important when one wants to link rSAM, RWB and hSAM. I am not saying that a single RWB could affect the whole hemisphere of course but at least this could affect a sector of the Southern Hemisphere larger than a synoptic-scale event. So my suggestion is to analyze time-lag composites of zonal wind anomalies in Figs.3, 6 and 8

b) In the same vein, have the authors looked at time-lag correlations between rSAMs in different sectors ? Same question for RWB. In Riviere and Drouard (2015, JAS), the NAM was shown to be first characterized by anomalies in RWBs and zonal wind in the North Pacific before propagating in the North Atlantic. In other words, several RWB events of the same nature may happen in different regions but with a time lag of several days.
2) Figure 6. Figures 6b and d show that CWB and AWB generally occur at the same place in terms of longitude/latitude/isentropic level during negative coherent rSAM events. Since AWB and CWB cannot occur simultaneously at the same place, do you think the composites of Figures 6b and d reflect the occurrence of different events ? To check that, I would encourage the authors to look at the pdf of CWB minus AWB densities for the negative coherent SAM events. Maybe the distribution is not centered over 0.
3) do the authors think that "rSAM" is the most appropriate wording for their regional diagnostic of pressure difference ? the letter "r" and the letter "A" in SAM are not compatible. Maybe "rDeltap" or "rzi" for regional zonal index are better suited. See for instance Feldstein and Lee (1996) for the zonal index definition.
Minor comments:
1) line 45: in Codron (2007, JAS), the JJA SAM is characterized by jet pulsing in the Pacific and jet shifting in the Indian Ocean. I would suggest to mention it.
2) Line 50: Please refer also to Ambaum et al (2001) for the discussion about NAM vs NAO
3) Lines 67-70: The NAM / RWB relationship has been investigated by Drouard et al (2015, J Clim) and Riviere and Drouard (2015, JAS). In particular, the latter paper shows that RWB signal in the North Pacific precedes the one in the North Atlantic (see my main comment above on the time lag)
4) Lines 76-80 please mention that the study considers all the seasons. Since the SAM is quite different between JJA and DJF, I am surprised there is no difference noted in the present study.
5) Line 97: 10° in longitude is a very small distance in mid-latitudes and even smaller than the synoptic scales. Why did the authors choose such a limited sector ? Please mention how the results would change by increasing the size of that sector.
6) line 103: "where" to be suppressed
7) Line 107 "predictor"
8) What do you mean by "generalized least squares solutions" ? Is it the same thing as multiple least squares regression models ?
9) Line 111-114: Where is such a framework of autocorrelation coefficient used in the paper ? Could the authors be more precise in this paragraph.
10) Lines 123-125: it would be nice to visualise the RWB zone in a schematic because the RWB zone varies from one RWB detection algorithm to another, and it is not clear how the present one works.
11) Line 156: additional figures are usually put in a supplementary file in EGUsphere journals and not in appendixes. Appendixes are more used to provide details on the methods. So please shift the current additional figures in a supplement information file.
12) Line 199: "highlights the asymmetry and lack of annularity of the SAM". Even though I understand what the authors want to say I am not sure this sentence is adequate. Figure 5 shows that the more rSAM of the same sign we have the stronger the hSAM phase is. So in my mind, hSAM has an annular property and extreme hSAMs occur when similar rSAMs events occur in very different sectors.
13) Line 232. The results of table 1 are consistent with the usual correspondence between spatial and temporal scales. The longer the time scale the larger the spatial scale.
14) Line 239: "consecutive" --> "contiguous" ? I think "contiguous" is more appropriate since we are talking about neighbouring regions. Also please specify that the coherence is spatial and not temporal
15) Figure 6: as previously said, please composite the zonal wind anomalies and look at different time lags. Also how do the composite fluctuate by looking at summer and winter separately ? Color curves in e and f are not described in the caption
16) Line 255-261. Some of the observations can be commented at the light of previous studies. The fact that cyclonic RWB occurs on the poleward side and anticyclonic RWB on the equatorward side is logical with usual baroclinic life cycles (Thorncroft et al. 1993, QJRMS). The fact that cyclonic RWB occurs at a lower vertical levels than anticyclonic RWB has been explained by Riviere (2009) in terms of refractive index properties.
17) Figure 7b. Do negative rSAMs systematically occur with both CWB and AWB at the same time, or do the statistics hide different types of events, some of them being dominated by CWB some others by AWB ?
18) Lines 320-327: this paragraph was a bit difficult to read. Are we talking about positive coherent rSAM only in the whole paragraph ?
19) Figure 8, please provide composite of the zonal wind anomalies at different time lags
20) Line 331: are you sure this is Figure 4 you want to refer to ?
21) Line 346: replace "SAM-like" by "rSAM"
22) Lines 377-378: in my opinion, it is less important to discuss jet streaks or cyclones densities in relation with rSAM than zonal wind anomalies at different lags that can be more easily compared with hSAM properties.
23) Lines 401-402. I am not convinced by that statement. Following the present study, we should now consider that there are different ways to talk about SAM (hSAM/rSAM) so saying "this interpretation is inconsistent with the SAM" is misleading because we do not know which SAM we are talking about. If we talk about hSAM, maybe the relation between the different orientations of RWB and the hSAM phases exist. As mentioned above, the NAM was shown to be closely related to the nature of wave breaking in Drouard et al (2015) and Riviere and Drouard (2015). So we cannot avoid the possibility that hSAM is also related to the nature of wave breaking. This could be even a test to be made computing CWB/AWB frequencies for positive and negative hSAM.
24) Conclusion. I think the text should be clearer. In many places, the text should mention "rSAM" (or "rzi" if the acronym is changed) or "hSAM" rather than simply "SAM".

Citation: https://doi.org/10.5194/egusphere-2026-3141-RC1
RC2: 'Comment on egusphere-2026-3141', Anonymous Referee #2, 11 Aug 2026

This study presents a sectoral view of dipolar MSLP anomaly structures between the southern mid- and high latitudes. Their integrated variability over the entire hemisphere corresponds to the Southern Annular Mode (SAM). The authors refer to the sectoral/regional MSLP dipole as regional SAM (rSAM). Statistical analyses of the frequency and structure of Rossby wave breaking (RWB) events, cyclonic storms and jet features are used to study the underlying dynamics of positive and negative phases of rSAM.

The analyses of different types of RWB (cyclonic, anticyclonic), jets and cyclones are well done and illustrated. I very much like the quality and information content of the corresponding plots (especially Figs. 6 & 8). The manuscript reads well. However, I struggle to follow the setup and interpretations of this study as described below. Because of this I think the manuscript needs major revisions, especially surrounding the motivation, main interpretations and conclusions.

Major comments

1) Misleading use of term "annularity"

To be completely honest I sense a bit of a strawman argument, as follows. By demanding that annularity is only given if the (hemispheric) SAM translates into zonally symmetric/coherent anomaly structures, which is a too narrow view of what the SAM is, the authors claim to show that this (zonally coherent anomalies) is in fact not the case. Why the annular structure of the dominant mode of hemispheric variability should automatically translate into zonal coherence of circulation anomalies is never explained. This underlying question has in fact been addressed in the literature. For example, Gerber & Thompson (2017, DOI: 10.1175/JAS-D-16-0191.1) distinguish zonally coherent fluctuations ("dynamical annularity") from mere zonally symmetric statistics ("statistical annularity"). The cited Spensberger et al. (2020) also refer to this distinction. It seems that the authors equate "annularity" with "dynamical annularity", but they never clarify or justify this. The interpretation of the results changes quite a bit if both kinds of annularity are considered (even more so given that daily fluctuations are considered, see next comment).

2) Time scales

The most common usage of the SAM as an index describing the hemispheric-scale circulation anomalies appears to be based on monthly or longer time scale variability (e.g., the recent review article by Purich et al. uses monthly-mean data throughout, as does the work by Thompson & Wallace, which is used to introduce the SAM in the present paper). It is clear, essentially by definition, that at daily time scales, as focused on here, a) synoptic dynamics will play a dominant role, b) these synoptic dynamics will be much more regional, c) they will therefore in many cases project only weakly onto the (h)SAM. This is why other studies first remove this synoptic variability, e.g., by taking monthly means or applying a low-pass filter, before defining the SAM. I think the fact that one can still define a (hemispheric) SAM at daily time scales should not be confused with the fact that this index is more appropriate for describing low-frequency fluctuations. In this sense, even in the context of "dynamical annularity" I find the presented arguments and interpretations misleading.

3) Index definition

I'm not sure I follow the choice of index definition (sect. 2.1). The simple difference of MSLP between 40°S and 65°S is certainly appealing from a practical perspective and is consistent with past work for the hSAM (although I wonder whether the long-term climate trend would call for some modifications of a ~30-year old index; an EOF-based index using detrended data is more immune to such issues). In any case, at regional scales it's not a priori clear whether the same latitudes will still serve well as a proxy of dominant modes of variability. So some of the shown differences between sectors may simply arise from different projections of the index used onto the actual dominant mode of variability. To me this issue is quite apparent in Fig. 2a/b: the two latitude circles are representative of the centers of action is some regions (crossing through regions of large correlation), while not in others (e.g., 65°S barely straddles the region of negative correlation, especially for JJA).

Further, by choosing a sector width of 10° a preferred spatial scale is effectively built into the analysis - about 700 km, i.e., that of synoptic scale dynamics (this doesn't change much in section 5 when requiring 3 coherent neighboring sectors, so the scale becomes ~2000 km). So in some sense it's obvious that synoptic scale dynamics are then found to explain the corresponding variability (at daily time scales). Note that when defining an EOF-based NAO pattern (the NH analogy of a regional index), a much larger spatial scale is typically used (often roughly a third of the hemisphere, i.e., a spatial scale of 10,000 km). So even the regional pattern is of planetary scale by definition, which makes studying its underlying synoptic drivers a less obvious problem.

Minor comments

line 15: I wasn't quite sure what "longest-lived" refers to - does it refer to the SAM's autocorrelation time scale, or the fact that it exists at very different time scales (daily, monthly etc.), or that it has existed as a mode of variability the longest?

line 30: I can't find any statements in Gerber & Vallis (which is primarily about the NAM) that question the physical meaning and interpretability of the SAM

line 32: I think this is a misrepresentation of the findings of Simpson & Polvani, who studied the jet response to climate change and made no comment about the physical meaning of the SAM itself

line 41: this is a good example where stating the time scale of variability is crucial (explained variance strongly depends on it)

line 144/145: on what basis are these correlation coefficient small? In the extreme case of uncorrelated sectors the fraction of explained variance per sector would only be 1/36 ~ 0.028, whereas the obtained fractions are a factor of 3-5 larger than this base line. So, perhaps they are not huge, but I wouldn't consider them small either.

line 178/179: where in the literature is it claimed that local weather should be fully captured by the hSAM (which, if anything, could only be used to obtain likelihoods for local weather patterns)? This statement as is seems trivial to me.

line 194: I find "impact" in this context misleading: all of these measures are statistical (hSAM, linear coefficient, correlation coefficient); the results of Fig. 4b/c really just mean that the Indian ocean has less overall variability (Fig. 4b should simply be equal to Fig. 4c multiplied by each sectors standard deviation)

line 198/199: see major comment - yes these highlight the zonal asymmetry, but that doesn't mean there is a lack of annularity: as you show in Fig. 2a/b yourself, the SAM is indeed quite annular; and the fact that all sector show statistically significant correlations could also be read as indication of robust annularity...

line 201/202: again see major comment (related to "dynamical" versus "statistical" annularity)

line 236: not sure about the usefulness of the analogy to the NAO (here and at other places throughout the manuscript): first, if anything the NAM (a more recently identified phenomenon) is a hemispheric version of the NAO (a much longer established phenomenon); second, the regional nature of the NAO is already given by the regional scale anomalies of the climatological circulation (which includes the Icelandic low and Azores high), so there's much less of a need to motivate a regional scope (obviously due to different land distribution in the NH)

line 275/276: not sure I follow - without information about the climatological position of maximum jet and cyclone frequency I can't tell whether the fields shown in Fig. 6e represent a shift or local increase. BTW, if I understand correctly, the increased jet frequency roughly midway between the chosen latitude circles to define rSAM is essentially by construction: demanding a strong pressure gradient between these latitudes gives increased zonal wind by geostrophy.

line 297: this is a kinematic statement; I suggest to rewrite as "jet streams are associated with strong PV gradients" (similar in line 300: I suggest to write "RWB is associated with a split jet ...")

Fig. 8: I suggest to swap left and right so that the panels correspond better to Fig. 6 (i.e., put anticylonic RWB on the left because it's dominant for positive rSAM)

line 330: see major comment - "misdiagnosis" seems to refer to the expectation that the hSAM phase should show up concurrently over all sectors ("dynamical annularity") ...

line 332: do you mean "examples in Figure 2"?

line 359 and rest of this and following paragraph: again, whether correlations of 0.39-0.52 are considered weak is a matter of perspective (i.e., you need to clarify in what sense they are weak); based on our understanding of the SAM it could be expected that at daily time scales its annularity is mostly statistical and in this sense your work confirms the annularity by showing that _all_ sectors are needed to produce the full SAM.

line 384/385: another instance of the flawed portrayal of what the SAM really is

line 395: I don't see this as something "counter-intuitive", in fact the last sentence of the abstract of Ding et al. states "Although the SAM is an intrinsic pattern of high-latitude variability, the SAM index reflects the superposition of both high-latitude and tropically forced variability", so is not at odds with the previous statement at all

Citation: https://doi.org/10.5194/egusphere-2026-3141-RC2

Michael A. Barnes, James S. Risbey, Teresa J. Parker, Nicholas Earl-Jones, Carly Tozer, and Didier P. Monselesan

Viewed

Total article views: 90 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
48	38	4	90	3	3

HTML: 48
PDF: 38
XML: 4
Total: 90
BibTeX: 3
EndNote: 3

Views and downloads (calculated since 17 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	38	27	4	69
Aug 2026	10	11	0	21

Cumulative views and downloads (calculated since 17 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	38	27	4	69
Aug 2026	10	11	0	21

Viewed (geographical distribution)

Total article views: 69 (including HTML, PDF, and XML) Thereof 69 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 24 Aug 2026

Short summary

The physical interpretation of the Southern Annular Mode (SAM) has been questioned in recent years. Here we show that the SAM can be interpreted through the diagnosis of synoptic-scale weather systems. In particular, the SAM is described here by the sum of the characteristics of Rossby wave breaking zones across the Southern Ocean. These differences produce local differences in the frequency of mid-latitudes cyclones and jets which the SAM is assumed to alter hemispherically.


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