the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Mar 2024
Downward Coupling of Sudden Stratospheric Warmings: Important Role of Synoptic-Scale Waves Demonstrated by ERA5 Reanalysis
Abstract. Circulation anomalies accompanying Sudden Stratospheric Warmings (SSWs) can have a significant impact on the troposphere. This surface response is observed for some but not all SSWs, and their downward coupling is not fully understood. We use an existing classification method to separate downward- and non-propagating SSWs (d/nSSWs) in ERA5 reanalysis data for the years 1979–2019. The differences in SSW downward propagation in composites of spatial patterns clearly show that dSSWs dominate the surface regional impacts following SSWs. During dSSWs, the upper-tropospheric jet stream is significantly displaced equatorward. Wave activity analysis shows remarkable differences between d/nSSWs for planetary and synoptic-scale waves. Enhanced stratospheric planetary eddy kinetic energy (EKE) and heat fluxes around the central date of dSSWs are followed by increased synoptic-scale wave activity and even surface coupling for synoptic-scale EKE. An observed significant reduction in upper-tropospheric synoptic-scale momentum fluxes following dSSWs confirms the important role of tropospheric eddy feedbacks for coupling to the surface. Our findings emphasize the role of the lower stratosphere and synoptic-scale waves in coupling the SSW signal to the surface and agree with mechanisms suggested in earlier modeling studies.
- Preprint
(7453 KB) - Metadata XML
-
Supplement
(15689 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-667', Anonymous Referee #1, 29 Mar 2024
This study examines SSWs in ERA5 over 1979-2019 from the perspective of the ‘downward’ and ‘non-downward’ propagating (dSSW and nSSW) classification of Karpechko et al. 2017, which is based around the sign of the NAM after the SSW at the 1000 and 150 hPa levels. The study first looks at simple diagnostics (T, Z, MSLP) of the precursor and subsequent evolutions of the events in the stratosphere and troposphere, plus a jet latitude index. Then, EKE, heat flux and momentum flux are examined in latitude-pressure space for dSSW and nSWWs. The authors state the key finding is the role of synoptic-scale waves (zonal wavenumber >3) in the downward coupling of SSWs.
While I found the manuscript easy to follow, I have several significant concerns. My primary concern is one of a circular argument: the study classifies SSWs based on the lower-stratospheric and surface NAM. The first FIVE figures just show this in various different ways – i.e., we are informed that dSSWs have a larger signal in the lower stratosphere and at the surface and have an equatorward jet shift, but this is all by construction… we don’t learn anything new here, we are just shown various different ways in which the negative NAM manifests. Additionally, many other aspects of this (e.g., with regard to precursors of dSSW v. nSSW) are not novel (although the authors do make this clear). Thus, it is only at Figure 6 that anything potentially useful is shown. However, at this point, it is also not clear that the results are not arising just by construction: how do we know that these anomalies are not just arising by conditioning on the NAM? How can the causality be established? There is a lack of dynamical explanation for the differences seen. Furthermore, the figures appear to show very little difference between the two subsets, and it is not clear whether these are significant (a particular concern given the small sample).
The authors state five main findings in the Conclusions. Of these, (1) is by construction, (2) seems somewhat tangential and unrelated to the main findings and has been explored in more detail elsewhere, (3) is what happens during a negative NAM by definition, and has also been explored in more detail elsewhere, and (4) and (5) do not seem to be sufficiently explored in the text to make such claims.
Given these concerns I cannot recommend publication, but I would suggest the authors explore some of the analyses begun in Figures 6-8 in more detail, since this is a topic of interest. I was also surprised to see this paper submitted to ACP rather than WCD. I hope my comments can be of some use to the authors in how they may wish to proceed with this study.
Specific comments:
Title: ‘demonstrated by ERA5 reanalysis’ is superfluous in my opinion (most studies use reanalyses to some extent but don’t put it in the title)
L15: temperature increases locally exceed 50K, I believe this value usually refers to the increase over the polar cap
L19: the response to SSWs at the surface more closely resembles the negative NAO, rather than the NAM (e.g. Hitchcock and Simpson 2014, also Dai and Hitchcock 2022, Dai et al. 2023). This basin asymmetry is not explored or mentioned in this article and is a significant part of S-T coupling.
L20: the tropospheric response can last for two months, but it’s the lower-stratosphere where this signal persists and that which drives the persistence otherwise (e.g. Baldwin et al. 2003, White et al. 2020).
L21: weather *and climate*
L23: rather than ‘stratosphere-troposphere coupling’, I think what is meant here is the ‘downward coupling of stratospheric circulation anomalies’
L25: this description of downward control is a bit thin and doesn’t explain the downward part
L33: Mitchell et al. 2013 also looked at displacement vs split impacts
L30-37: what did these studies find? The text here describes what they did, but not what the key takeaways were or how they relate to the questions seeking to be answered here
L39: delete ‘successfully’ (if Karpechko hadn’t been successful, I don’t think it would have been published!)
Introduction in general misses a discussion of the importance of the lower stratosphere for downward impacts (e.g. Gerber et al. 2009, Maycock and Hitchcock 2015) and how this might relate to some of the other classifications (e.g. absorbing/reflecting, split/displacement)
L46: probably need to explain more about how this synoptic wave propagation can happen given usual filtering of stratosphere to k=1-3.
L57: ERA5 isn’t really ‘new’ anymore
L58: why does this study use a dataset which terminated five years ago? At the very least it should be updated to 2023, which would add at least two more SSWs or up to four if 2024 were included. The sample size (which is currently low) could also be increased by extending back to 1958, considered the start of more reliable stratospheric measurements, and the benefit of such a larger sample (e.g. Hitchcock 2019)
L69: so is the Charlton-Polvani definition used? I.e. 20 day separation?
L71: is the standard deviation daily or seasonal?
L80: does this ‘tropospheric’ jet latitude definition go a bit high? 100 hPa is in the stratosphere
General methods comment: consideration of effect of warming trends on T2m and geopotential height?
L88: not sure what is meant by ‘improve the detectability’
L90: for such composites I’d normally encourage non-parametric bootstrap tests. Further, we don’t see evidence that the composites are different from one another, only that they are different from climatology. Two subsets can be significantly different from climatology but not significantly different from each other, which is a key part of this study.
L96: see also the Butler and Gerber 2018 classification of SSWs re. March 2010
L98: ‘agree well’ – if there are differences, what are they?
L102: this is by construction!
L108: showing temperature as well as geopotential height seems redundant, since the main focus here is geopotential height/circulation and GPH closely follows T in the stratosphere anyway
Figures 1-4: the latitude cut-off seems a bit high – it’s more poleward than the latitude of the southern jet regime in the Atlantic which is seen after SSWs so potentially misses some interesting anomalies
Figure captions: the figures show stippling, not hatchingL121: by construction!
L123: the sentence beginning ‘Overall…’ just states the same thing in different ways – an SSW is a negative NAM which is a weak vortex.
L124: by construction!
L142: MSLP anomalies don’t propagate down
L148: unsubstantiated claim about sea ice
L151: and many others aside from Kretschmer et al. 2017 (which, in fairness, are mentioned later in this section)
L156: Aleutian Low precursor consistent with Garfinkel et al. 2012
L170: the lower stratosphere is baked into the definition of dSSW, so its importance cannot be known by what is shown here. One would have to composite just on the lower stratosphere to determine this.
L172: what is meant by ‘strong’ here?
L195: also Afargan-Gerstman and Domeisen (2020) looked at jet latitude shifts after SSWs
L199: but the response on the jet is in the Atlantic (hence why these studies have focused on it), the global focus will just be picking this same aspect up (and if not, then that would be interesting, but would be lost in the global perspective). In either case, it is not adding much beyond what we know from conditioning on the NAM
Figure 6: stippled and hatched might be clearer than grey and black. Colour scale oversaturated. The large tropospheric anomalies aren’t stippled? So the only panels that appear hugely different here are D15-30, but these aren’t significant?
L244: “After lag 15…. Significantly increased in the lower stratosphere” I do not see this, nor do I see evidence of a large difference between the two subsets. Lag15 to 30 in WN4-15 is almost identical.
L248: “the surface coupling… is through synoptic-scale waves” – this is a big claim and I do not see sufficient evidence to support it
Additional References:
Dai and Hitchcock 2021 J. Climate https://doi.org/10.1175/JCLI-D-21-0314.1
Dai et al. 2023 J. Climate https://doi.org/10.1175/JCLI-D-22-0300.1
Baldwin et al. 2003 Science https://doi.org/10.1126/science.1087143
Mitchell et al. 2013 J. Climate https://doi.org/10.1175/JCLI-D-12-00030.1
Gerber et al. 2009 GRL https://doi.org/10.1029/2009GL040913
Hitchcock 2019 ACP https://doi.org/10.5194/acp-19-2749-2019
Butler and Gerber 2018 J. Climate https://doi.org/10.1175/JCLI-D-17-0648.1
Garfinkel et al. 2010 J. Climate https://doi.org/10.1175/2010JCLI3010.1
Afargan-Gerstman and Domeisen 2020 GRL https://doi.org/10.1029/2019GL085007Citation: https://doi.org/10.5194/egusphere-2024-667-RC1 -
RC2: 'Comment on egusphere-2024-667', Anonymous Referee #2, 02 Apr 2024
Review of "Downward Coupling of Sudden StratosphericWarmings: Important
Role of Synoptic-ScaleWaves Demonstrated by ERA5 Reanalysis" by Rahm et alThis study uses reanalysis data to study the circulation anomalies before and after SSWs. In particular the authors compares the evolution of SSWs which impact the troposphere to those that do not. The authors find that the evolution noted in previous reanalysis studies which focused on all SSWs is particularly evident for downward propagating SSWs: changes in the upper-tropospheric jet stream, changes in wave activity (EKE, heat flux, momentum flux) for planetary and synoptic-scale waves, and regional changes in e.g. SLP. The main novelty the authors argue for this study is its use of reanalysis data instead of model study.
As currently constructed, I don’t think the novelty of this study is sufficient to warrant publication in a journal such as ACP. The findings emphasized by the author are very incremental, and to a large degree obvious and there by construction. Below are some comments that may help the authors focus on the novel elements of their results, and it could be that an essentially rewritten paper submitted to a more relevant journal would be publishable.
General comments:
The Karpechko et al 2017 definition of downward propagation will, by definition, discriminate between the NAM responses at 150hPa and 1000hPa in days 8 to 52 after SSW onset. Hence, it doesn’t make sense to me to focus much on panels that show the response in this period, other than a sanity check. Perhaps it is reasonable to show map views of the surface response in these later leads if the authors think there are interesting results, however to me it just looks like the typical -NAM pattern, and so not much to write about at least for a journal like ACP. (See also minor comment 1.)
To me, the more interesting aspect of this paper is to focus on days -15 to 8, i.e., the period in which anomalies are not trivial. While it would be reasonable to include on figures one column devoted to the response after day 8, the other columns should focus on the earlier period. The current column devoted to day 0 to 15 seems too broad to allow for interpretation – I strongly recommend showing anomalies from day 0 to day 8 only (I suspect the results will be very similar, but at least the results won’t be baked in by the methodology). The rest of my review will be devoted to the first column only which focuses on days -15 to 0.
There is an interesting difference between figure 3de and 4de with regards to the strength of the Ural high precursor. This seems to agree with White et al 2019: they also find that dSSWs have a stronger Ural high precursor. Further, they argue that around 25-30 dSSW and 25-30 nSSWs are needed before the difference in this Ural high precursor becomes statistically significant in the model they used. The authors here have a significantly smaller sample, yet based on the figure in the supplemental material there appears to be a statistically significant difference in this Ural high precursor despite this small number of events. This might suggest that the model of White et al underestimates the sensitivity of the polar vortex to tropospheric precursors.
I understand the authors want to focus on the post-1979 period only, however SLP and Z500 should be well captured by ERA5 since the 1950s. It would be very interesting to see whether this difference in the Ural low precursor is evident in the extra ~20 years of data and >10 extra SSWs.
There is also a difference in the strength of the North Pacific precursor between nSSW and dSSW in figure 3de and 4de. Perplexingly, the Aleutian low precursor is stronger for nSSW. In the model runs of White et al, the opposite was found: a stronger precursor for dSSW. However, the figure in the supplemental material indicates that this difference isn’t significant, so this doesn’t appear to be a robust result.
Minor comments
- Hitchcock et al 2013ab studies differences between PJO events and non-PJO events in reanalysis data. I suspect the events are very similar to the authors dSSW, and as best as I can tell the long-term impact in the lower stratosphere and troposphere seems very similar. Neither of these studies is cited at present.
- There are earlier studies that have focused on tropospheric precursors of SSWs and found anomalies in the Ural region, e.g. Garfinkel et al 2010 and Cohen and Jones 2011. Neither study is cited at present.
- The caption to figure 3 indicates these anomalies are pressure-weighted, but the figure is showing lat vs lon. I’m confused. More generally, there are large differences in the EKE between nSSW and dSSW before event onset. Is this just a reflection of Ural block in dSSW and deeper Aleutian low in nSSW?
- I recommend that figures showing nSSW minus dSSW be included in the main manuscript and not be in the supplemental material, at least for key diagnostics that are statistically significant and are discussed in the paper. This would better communicate the key point of the paper.
Cohen, Judah, and Justin Jones. "Tropospheric precursors and stratospheric warmings." Journal of climate 24, no. 24 (2011): 6562-6572.
Garfinkel, Chaim I., Dennis L. Hartmann, and Fabrizio Sassi. "Tropospheric precursors of anomalous Northern Hemisphere stratospheric polar vortices." Journal of Climate 23, no. 12 (2010): 3282-3299.
Hitchcock, Peter, and Theodore G. Shepherd. "Zonal-mean dynamics of extended recoveries from stratospheric sudden warmings." Journal of the atmospheric sciences 70, no. 2 (2013): 688-707.
Hitchcock, P., Shepherd, T. G., & Manney, G. L. (2013). Statistical characterization of Arctic polar-night jet oscillation events. Journal of Climate, 26(6), 2096-2116.
Citation: https://doi.org/10.5194/egusphere-2024-667-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 196 | 42 | 13 | 251 | 20 | 7 | 5 |
- HTML: 196
- PDF: 42
- XML: 13
- Total: 251
- Supplement: 20
- BibTeX: 7
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1