the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Aug 2025
Northern Hemisphere Stratospheric Polar Vortex Morphology under Localized Gravity Wave Forcing: A Shape-Based Classification
Abstract. The Northern Hemisphere stratospheric polar vortex (SPV) response to localized gravity wave (GW) forcing remains poorly understood, particularly in terms of its detailed morphology. Here, we investigated geometry-specific impacts of enhanced orographic GW drag in three hotspot regions, the Himalayas, Northwest America, and East Asia, using ensemble simulations with the high-top UA-ICON global circulation model. By classifying daily SPV geometries into ten distinct clusters with a novel unsupervised, shape-based hierarchical clustering framework, we isolated geometry-specific responses using the class contribution method. Our results showed that all hotspot forcings consistently reduce planetary wave 1 (PW1) amplitude and induce a PW1-like displacement of the SPV core, though spatial patterns vary with hotspot location. This response manifested as negative geopotential height (GPH) anomalies within the forced region and positive anomalies to the north, indicating localized SPV edge mixing. The response was also sensitive to the forcing’s latitudinal position: the Himalayas, as the southernmost hotspot, produced a deepened vortex, while the more poleward Northwest America and East Asia forcings showed similar patterns with greater intrusion of positive GPH anomalies into the vortex core. The forcing reduced PW1 amplitude both by shifting the frequency of specific clusters and altering the mean structure of the most frequent classes. Our results demonstrate that shape-based clustering combined with the class contribution framework can reveal robust, spatially coherent signals that might otherwise be masked by internal variability, providing a new perspective for understanding SPV variability and its predictability.
- Preprint
(45219 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 02 Oct 2025)
- RC1: 'Comment on egusphere-2025-3612', Anonymous Referee #1, 17 Sep 2025 reply
-
RC2: 'Comment on egusphere-2025-3612', Anonymous Referee #2, 18 Sep 2025
reply
General Comments:
This paper utilizes an unsupervised, hierarchical clustering technique to define geometric clusters representing the behavior of the stratospheric polar vortex as defined by boundary objects and features from a FFT of 10-hPa GPH fields. Thereafter, this work examines the stratospheric polar vortex response to gravity wave forcings in climate model scenarios, with a focus on three hotspot regions: East Asia, the Himalayas, and North America. Their results conclusively show that the hotspot forcings reduce amplitudes of planetary wave size 1 features, with the largest of those changes occurring in the Himalayan hotspot. Additionally, the observed gravity wave forcings hold implications for geometric responses in the stratospheric polar vortex. The results from this paper are compelling and provide interesting contributions to the discussion of stratospheric polar vortex morphology. Outcomes from this paper, specifically regarding the clustering of the vortex, hold applicability for examinations of subseasonal stratospheric-tropospheric teleconnections.
The paper itself is technically dense and difficult to follow at times. It reads like a chapter of a dissertation. The paper would improve greatly with some changes to the wording and extrapolation of specific methodologies to help the content stand as an individual publication. Additionally, the author should explain the underlying motivation for using Fourier descriptors to define the vortex boundary in place of pre-existing, similar methodologies.
I would recommend publication following the completion of these major and minor revisions.
Major Comments:
- There is a lot of “hand waving” done with respect to explaining the specifics of this experiment and its methodology to the point that a lot of backtracking was required to make sense of what was being discussed in Section 3. This is particularly an issue for anything surrounding the model simulations since the author assumes the reader has read Mehrdad et al. 2025a. It is not necessary to include all clarifying information about the specifics of the original experiment in Mehrdad et al. 2025a, but enough detail is needed so that an individual can read this paper on their own and know what is happening. I would suggest that you include more specifics about the methodology surrounding model simulations and data in Section 2. Clarified methodological information will improve this section of the paper substantially.
- This paper would benefit from an explanation and motivation for the choice to use Fourier descriptors for diagnosing the vortex boundary as opposed to using pre-existing vortex diagnostic methods, like those from Seviour et al. 2013, or k-means clustering. The FFT method presented here is very compelling, but what is the underlying motivation for using it? Seviour, W. J. M., D. M. Mitchell, and L. J. Gray (2013), A practical method to identify displaced and split stratospheric polar vortex events, Geophys. Res. Lett., 40, 5268-5273 doi:10.1002/grl.50927.
- While there is a lot of useful information in the discussion of the sensitivity experiments, sections 3.3-3.5 of this paper suffer from an oversaturation of material. I found it difficult at times to maintain focus on the key takeaways from the figures. This quality of writing is ideal for a dissertation but not necessary for a publication. I would suggest thinning out the text from this area, focusing on the primary results. I may also suggest modifying the figures here to show only the most relevant clusters. For example, the panels in Figures 10-18 for C1, C7, and C10 are generally empty.
Minor Comments:
Line 1-2 – I would personally include a reference to what months specifically define the “winter pole” in the Northern Hemisphere.
Line 30 – You may also want to cite this paper as well: Butchart, N. 2022: The Stratosphere: A Review of the Dynamics and Variability. J. of Weather and Climate Dynamics. 3(4), 1237-1272. https://doi.org/10.5194/wcd-3-1237-2022
Line 47 – This could be a technical comment, but I would be inclined to put the initial reference of “compensation mechanisms” in quotes, since it is a reference to terminology.
Line 48 – “These” referring to what? The compensation mechanisms?
Line 53 – What is meant by “inter-model comparisons of stratospheric dynamics”? Would it be possible to provide an example?
Line 59 – “They” referring to what?
Line 60-62 - “hotspot-induced changes” … to GWs? There is a lack of specific clarifying information in this sentence.
Line 65-72 – As stated in Major Comment 1, some readers will find this paper on its own and may not read Mehrdad et al. 2025a. Silly as it may sound, it is necessary to provide a small amount of context to the reader in the Introduction, one or two sentences, by giving a general overview of Mehrdad et al. 2025a’s general goals and key takeaways. The author does a good job of explaining how this specific paper has different motivations, continuing in this paragraph, but does not fully address the previous paper.
Line 65 – Since this is the first mention of UA-ICON, I would reference it in full: “high-top UA-ICON global circulation model experiments.”
Line 67-69 – Something about point (ii), starting with a verb, while points (i) and (iii) start with “we” disrupts the flow of this paragraph and makes it difficult to digest.
Line 71 – I might replace “Arctic” with “Northern Hemisphere” or similar. Or reflect on future nomenclature used in the paper and try to maintain consistency.
Line 93 – What was the period of record for this winter season data? 30-year simulations are indicated, but what is the start date? Again, this information may be included in the previous paper from Mehrdad et al. 2025a, but it is still important to include clarifying details about the specifics of this experiment as it is its own paper.
Line 110 – Explain why you retain data northward of 25°N. Similarly, why 18%?
Line 135 – Perhaps be a little more direct/descriptive beyond saying “this feature space” to provide further clarity to the reader.
Line 168-170 – C1 is distinguished for unstable-vortex clusters, in this case meaning more than two or no SPV boundaries. C7-C10 are split vortex events. Would C2-C6 contain a cluster specified for displaced vortex events? What about strong vortex events? You may mention this in detail later, but it is good to give some preliminary indication of what these clusters might contain.
Figure 4 – The cyan lines are difficult to read in this figure. I may suggest using a different colored contour. Additionally, I may separate Panels d and h into their own figures. Included here, the panels are a little small, and their inclusion makes this figure a bit too busy.
Figure 6 – Similar comments to Figure 4. The coloring is difficult to see, and the subplots on this figure are small and difficult to read. Breaking up or structuring this figure differently may make it easier to read.
Line 234 – This sentence may not be necessary.
Line 262 – What exactly is an “occurrence frequency”/how is it calculated?
Line 375, 377 – “gravity wave” may be replaced with GW.
Line 391 – Which configuration? Be a little more specific.
Line 393 – At this point within the discussion section, I may remind the reader what each of the clusters (C5, C6, etc.) physically represents.
Technical Comments:
In general, I noticed many tense inconsistencies (e.g., present vs. past tense within the same sentence, etc.) Choose one tense to use throughout the paper.
All of the figures would benefit from being increased in resolution. Figure formatting within the text should also be changed to keep the discussion collocated with the image as close as possible.
Line 45 – “exert” should be “exerts”
Line 133 – “features 4 to 23” may be replaced with “features four to twenty-three” for consistency
Figure 5 – Increase figure size.
Section titles 3.3, 3.4, 3.5 – Use the unabbreviated versions of the sensitivity experiments for the titles.
Line 318-320 – “consistent” is used three times in some form here. I may suggest rewording.
Line 431, 448 – “zonal mean” should be “zonal-mean” here
Line 461 – “help” should be “helps”
Citation: https://doi.org/10.5194/egusphere-2025-3612-RC2 -
RC3: 'Comment on egusphere-2025-3612', Anonymous Referee #3, 18 Sep 2025
reply
The study addresses the impact of long-term intensified GW forcing on SPV. The authors classified SPV into ten groups based on morphology and analyzed the impact of GW forcing on three major orographic GW hotspots (HI, EA, and NA) separately. Using the UA-ICON global circulation model, the authors conducted multi-ensemble simulations to account for the internal variability of the signal. The paper showed that all hotspots exhibited a consistent decrease in the amplitude of planetary wave number 1 and the corresponding change in circulation. Furthermore, the paper showed that these anomalies are due to changes in the mean SPV structure and frequencies of several clusters, to varying extents depending on the hotspot.
While Mehrdad et al. address highly relevant scientific questions within the scope of ACP, the paper requires minor revisions to be published. The paper introduces a novel methodology for SPV classification and carefully addresses the impact of GW forcing on different clusters. However, the discussions of statistical significance and the mechanism of the signals are lacking and could be improved.
General comments
- Are there any differences in the vertical profiles of GW forcing by the hotspot? Perhaps there are no significant differences, but showing the general picture of how the forcings are applied would be helpful. If differences exist, they could suggest potential reasons for different responses. The example figure in Mehrdad et al. (2025a) helps with understanding, but a more general distribution of the forcing by hotspot would be helpful.
- The methodology in the paper is convincing and suggests a novel approach. However, a discussion on why this method was chosen over well-known clustering methods such as EOF, k-clustering, and self-organizing maps would be beneficial. What are the advantages?
- Most of the statistical discussion here relies on the consistency of the signals. However, the significance of these signals (e.g., GPH, zonal wind, and frequency) is questionable. For instance, the authors repeatedly emphasize the strong signal in zonal mean zonal wind, yet the magnitude is less than 1 m/s in most figures. Considering the strong and highly variable winds in the stratosphere, can this signal be significant enough to lead to a meaningful change in circulation?
- Figure 6 shows that GW forcing seems to explain the zonal wind response over the hotspot and to some extent northward (L230-231). In higher latitudes, however, especially in the HI experiment, EP flux divergence does. What do the authors think led to this difference? L440-441 argues that the delayed adjustment is a potential cause of this inconsistent response, but how can this explain the latitudinal dependence of the balance between GW and resolved waves?
- Overall, the mechanistic discussion is lacking. Why do we see certain anomalies in WCVC? What causes frequency changes? Why do they respond differently depending on the hotspot? While this may be outside the scope of this study, it would be better to express the authors' opinion and leave it as an open question.
Minor comments
L53: The first sentence of this paragraph does not seem related to the rest of it.
L145-146: Is rescaling not necessary for the boundary size and time feature?
Figure 2. What do different colors mean?
L219-221: divergence is reduced -> convergence is reduced?
Figure 7. The GPH differences in C7-C10 are difficult to discern. Therefore, it is difficult to determine if the boundary aligns with the GPH. Is it possible to show the discrepancy more clearly?
L259-262: The first sentence states that C3 has a relatively circular vortex, but a later sentence says that C3 is strongly deformed. This is confusing and needs clarification.
L349: How is SSW linked to C4 and C8? There is no earlier discussion about this. It would be better to either remove the statement or show the relationship in the supplementary material.
L383-392: First, it is argued that positive GPH anomalies north of the hotspot lead to the weakening and mixing of the SPV edge. However, for the HI experiment, a similar signal leads to the sharpening of the SPV. This needs to be clarified, as a similar signal led to a different response, and the two arguments conflict with each other.
L405-407: Same here. The SSW linkage to C4 and C8 must be shown explicitly first.
L426-428: Could you clarify what it means to "follow mean circulation"?
L428-429: Again, the strengthening of the SPV edge conflicts with the sentence before. Please clarify this.
L442: C9 -> C8
L444-446: Positive anomalies seem to exist at different longitudes.
Citation: https://doi.org/10.5194/egusphere-2025-3612-RC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,798 | 39 | 14 | 1,851 | 26 | 24 |
- HTML: 1,798
- PDF: 39
- XML: 14
- Total: 1,851
- BibTeX: 26
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
This study assesses the Northern Hemisphere stratospheric polar vortex response to localized gravity wave forcing above three hotspot regions, the Himalayas, Northwest America, and East Asia, using UA-ICON GCM. The results highlight that all hotspot forcings consistently reduce planetary wave 1 amplitude, which is discussed in detail. I find the study highly relevant, especially due to the classification framework developed and its application to transient climate simulations and reanalyses. I recommend publication once the minor comments below are addressed.
Specific comments
I noticed that in Fig. 6 in Mehrdad (2025a), the zonal-mean climatology of the tendencies induced by the OGW parameterization scheme shows a secondary maximum in the lower stratosphere over midlatitudes, within the so-called valve layer (Kruse et al., 2016). However, this maximum is located around and below 100 hPa. This contrasts with the breaking of freely propagating OGWs above the center of the UTLS jet starting rather above in CMIP6 AMIP simulations (Hajkova and Sacha, 2024). Can you comment on this deficiency or model tuning with respect to vertical profiles of OGWD in the sensitivity simulations above NA and HI, where the breaking is maximized below 100 hPa (see Fig. 3 in Mehrdad (2025a))? This can consequently affect the polar vortex response simulated by UA-ICON.
I miss the motivation why such a methodology has been applied to classify the SPV geometry compared to either standard clustering techniques (e.g. k-means in Kretschmer et al (2018)) and/or standard techniques for split and displacement identification (e.g. Seviour et al , 2013)
Can you include the value of the 18% threshold mentioned in Section 2.3.1 and Fig. 1?
Due to the extensiveness and unique methodology of the study, I think the whole community would appreciate an adoption of Open Science approaches to allow reproducing the extensive analysis in this study (e.g. Laken, 2016). In particular, I would recommend any kind of willingness of the authors to publish the code or a series of functions allowing to reproduce the figures in the paper. There are multiple ways to proceed, either to allow access upon request or via portals that allow assigning Digital Object Identifier (DOI) to the research outputs, e.g. ZENODO. I think it could enhance the quality and reliability of this publication.
As shown in Mitchell et al (2011), splits are also accompanied by equatorward shift of the vortex (diagnosed by centroid latitude), i.e. a PW1-like pattern. In this view, I would suggest discussing results in Sections 3 and 4. Some studies find little (e.g. Maycock, A. C., and P. Hitchcock, 2015) or strong (e.g. Mitchell e al, 2013) differences between the surface impacts of split and displacement events. Have authors found any surface signatures in the sensitivity experiments?
I had trouble seeing dotted regions and contours in Figs. 3,4,5. I encourage authors to enhance their clarity. The choice of colours (cyan and green) in Fig. 10 could also be improved.
I would replace abbreviations (EA, HI, NA) with their full length in subsection titles.
I would move Fig. 5 to the appendix/supplement.
Have you considered decomposing EPFD into leading zonal planetary wave modes? As shown in Sacha et al (2021), diverse dynamical responses to OGWD hotspots, particularly given the different wavenumbers. This has been in details discussed in Kuchar et al (2022), highlighting that strong and intermittent OGW drag events above the Himalayas in the lower stratosphere are associated with anomalously increased upward RW propagation in the stratosphere. This is somewhat different to the conclusion of this study. Overall, I suggest discusses differences in findings from previous studies in the manuscript.
Technical comments
L315 (Figure 9. -> (Figure 9).
References
Hajková, D., Sacha, P. Parameterized orographic gravity wave drag and dynamical effects in CMIP6 models. Clim Dyn 62, 2259–2284 (2024). https://doi.org/10.1007/s00382-023-07021-0
Kretschmer, M., D. Coumou, L. Agel, M. Barlow, E. Tziperman, and J. Cohen, 2018: More-Persistent Weak Stratospheric Polar Vortex States Linked to Cold Extremes. Bull. Amer. Meteor. Soc., 99, 49–60, https://doi.org/10.1175/BAMS-D-16-0259.1.
Kruse, C. G., Smith, R. B., and Eckermann, S. D.: The midlatitude lower-stratospheric mountain wave “valve layer”, Journal of the Atmospheric Sciences, 73, 5081–5100, https://doi.org/10.1175/JAS-D-16-0173.1, 2016.
Kuchar, A., Sacha, P., Eichinger, R., Jacobi, C., Pisoft, P., and Rieder, H.: On the impact of Himalaya-induced gravity waves on the polar vortex, Rossby wave activity and ozone, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2022-474, 2022.
Laken, B. A. (2016). Can Open Science save us from a solar-driven monsoon? Journal of Space Weather and Space Climate, 6, A11. http://doi.org/10.1051/swsc/2016005020.
Maycock, A. C., and P. Hitchcock (2015), Do split and displacement sudden stratospheric warmings have different annular mode signatures?, Geophys. Res. Lett., 42, 10,943–10,951, doi:10.1002/2015GL066754.
Mehrdad, S., Marjani, S., Handorf, D., and Jacobi, C.: Non-zonal gravity wave forcing of the Northern Hemisphere winter circulation and effects on middle atmosphere dynamics, EGUsphere, 2025, 1–35, https://doi.org/10.5194/egusphere-2025-3005, 2025a.
Mitchell, D. M., A. J. Charlton-Perez, and L. J. Gray, 2011: Characterizing the Variability and Extremes of the Stratospheric Polar Vortices Using 2D Moment Analysis. J. Atmos. Sci., 68, 1194–1213, https://doi.org/10.1175/2010JAS3555.1.
Mitchell, D. M., L. J. Gray, J. Anstey, M. P. Baldwin, and A. J. Charlton-Perez, 2013: The Influence of Stratospheric Vortex Displacements and Splits on Surface Climate. J. Climate, 26, 2668–2682, https://doi.org/10.1175/JCLI-D-12-00030.1.
Sacha, P., Kuchar, A., Eichinger, R., Pisoft, P., Jacobi, C., & Rieder, H. E. (2021). Diverse dynamical response to orographic gravity wave drag hotspots—a zonal mean perspective. Geophysical Research Letters, 48, e2021GL093305. https://doi.org/10.1029/2021GL093305
Seviour, W. J. M., D. M. Mitchell, and L. J. Gray (2013), A practical method to identify displaced and split stratospheric polar vortex events, Geophys. Res. Lett., 40, 5268-5273 doi:10.1002/grl.50927.