Summary

This study examines the influence of large-scale atmospheric processes on the formation of cyclones that lead to extreme surface winds in the central North Atlantic region (30° to 60°N, 10° to 50°W) during the extended winter period from 1950 to 2020, using the ERA5 reanalysis data. It identifies extreme surface wind events based on the 98th percentile of 10-m wind speed data. The study focuses on the top 1% most intense wind events (referred to as 'top extremes') associated with cyclones and compares them to wind events (also associated with cyclones) slightly above the 98th percentile (referred to as 'moderate extremes'). The analysis reveals that top extremes are characterized by the presence of pre-existing downstream cyclones, a strong jet, and positive upper-level potential vorticity anomalies to the north, whereas these features are weaker or absent in the case of moderate extremes. Additionally, the study finds that top extremes tend to have a greater diabatic contribution to surface pressure decrease compared to moderate extremes.

Major comments

Some parts of the paper are well structured and written, particularly the Results section, which is appreciated from a reviewer perspective. The figures are also very informative and effective, wich again is appreciated from a reviewer perspective. Still, I have some major concerns with the paper, which need to be addressed, prior to publication.

1.1 The abstract presents some inconsistencies with the rest of the paper. For instance, extratropical cyclones with tropical cyclone origins, which constitute approximately 10% of the top 1% extreme wind events, are excluded from the analysis. However, this exclusion is not mentioned in the abstract. Please ensure that the content presented in the paper is accurately reflected in the abstract whenever possible.

1.2 The authors should explicitly specify the large-scale atmospheric factors that influence the formation of cyclone extreme wind speeds, which they intend to investigate. While these factors are briefly mentioned in the introduction, they are not explicitly linked to the research questions. Furthermore, certain factors discussed in the paper, such as Rossby wave breaking, are not mentioned in the abstract (as highlighted in the previous point).

1.3 It is important to better emphasize the motivation behind the study. While the authors correctly mention that previous literature focused on extreme wind speeds over land due to concerns about loss of life and property associated with cyclone extremes, it is necessary to articulate why studying cyclone extreme wind speeds over the ocean is valuable. Aside from practical considerations, such as the absence of topographical features introducing noise in near-surface fields, what is the primary motivation for this aspect of the study? Clarifying this will enhance the paper's context.

1.4 I find the use of the destructiveness index somewhat an unnecessary complication, as it is typically derived for land-based scenarios. I appreciate the motivation behind its use is more leaning towards ranking the extremes rather than evaluating associated potential insurers’ losses. Therefore, it is crucial to provide a clearer explanation of the study's purpose, whether it serves scientific, practical, or both purposes. Perhaps consider using the term "severity index" instead to align with the study's objectives.

1.5 Another crucial element missing is a discussion of how the choice of cyclone tracking method used may impact the results. There may be significant variations in results, as vorticity-based methods often detect cyclones differently when compared to pressure-based methods (as demonstrated by Neu et al. 2013). Including such a discussion in Section 3 would strengthen the paper.

1.8 There are some concerns regarding the interpretation of the changes in the budget terms of the pressure tendency equation. Specifically, it is not entirely clear from the presented results whether the diabatic contributions are substantially larger in the top extreme cases relative to the moderate extremes compared to the baroclinic contribution group. To strengthen the robustness of the findings, it is advisable to incorporate non-parametric tests to evaluate the results rigorously.

1.9 An important aspect that appears to be missing is an exploration of how the identified signal varies within the historical period. This is particularly relevant considering the significant changes in our warming climate. It would be valuable to assess how the results compare between the periods 1950-1985 and 1985-2020, especially in terms of the magnitude of the difference between baroclinic and diabatic contributions to the pressure tendency equation. The latter period experiences a human-induced warmer climate signal, and investigating this aspect would greatly enhance the paper's relevance and comprehensiveness.

1.10 Finally, the literature is a bit lacking in some areas concerning storminess in Europe and in the North Atlantic region, motivation for studying extreme wind speeds over the ocean, and dependency of cyclone tracks on chosen methodology. Please add the following publications:

• Earl, N., Dorling, S., Starks, M. and Finch, R. (2017) Subsynoptic-scale features associated with extreme surface gusts in UK extratropical cyclone events. Geophysical Research Letters, 44, 3932–3940.
• Feser, F., Barcikowska, M., Krueger, O., Schenk, F., Weisse, R., & Xia, L. (2015): Storminess over the North Atlantic and Northwestern Europe - A Review. Q. J. R. Meteorol. Soc., 141, 350-382, January 2015 B.
• Gentile, E. & Gray, S. Attribution of observed extreme marine wind speeds and associated hazards to midlatitude cyclone conveyor belt jets near the British Isles.  J. Climatol.43, 2735–2753 (2023).
• Hart, N.C.G., Gray, S.L. and Clark, P.A. (2017) Sting-jet windstorms over the North Atlantic: climatology and contribution to extreme wind risk. Journal of Climate, 30, 5455–5471.
• Hewson, T.D. and Neu, U. (2015) Cyclones, windstorms and the IMILAST project. Tellus A, 67, 27–128.
• C. Leckebusch, D. Renggli, U. Ulbrich, Development and application of an objective storm severity measure for the Northeast Atlantic region, Meteorologische Zeitschrift, Vol. 17, No. 5, 2008
• Manning, C., Kendon, E.J., Fowler, H.J., Roberts, N.M., Berthou, S., Suri, D. and Roberts, M.J., 2022. Extreme windstorms and sting jets in convection-permitting climate simulations over Europe. Climate Dynamics, 58(9-10), pp.2387-2404.
• Messmer, M., I. Simmonds, 2021: Global analysis of cyclone-induced compound precipitation and wind extreme events. Weather and Climate Extremes.
• Ponce de León, S. and Bettencourt, J. (2021) Composite analysis of North Atlantic extra-tropical cyclone waves from satellite altimetry observations. Advances in Space Research, 68, 762–772.
• Ulbrich, U., G. C. Leckebusch, J. Grieger, M. Schuster, M. G. Akperov, N. Y. Bardin, Y. Feng, S. Gulev, M. Inatsu, K. Keay, S. F. Kew, M. L. R. Liberato, P. Lionello, I. I. Mokhov, U. Neu, J. G. Pinto, C. C. Raible, M. Reale, I. Rudeva, I. Simmonds, N. D. Tilinina, I. F. Trigo, S. Ulbrich, X. L. Wang, H. Wernli and the IMILAST team, 2013: Are Greenhouse Gas Signals of Northern Hemisphere winter extra-tropical cyclone activity dependent on the identification and tracking methodology? Meteorologische Zeitschrift, 22, 61-68.

Minor comments

1.11 Line 21: Could you provide more clarity regarding the specific features you are referring to? It may be beneficial to include a reference like Earl et al. (2017) to support your point.

1.12 Line 23: The phrase "concentrated on structure" lacks clarity. Please consider rephrasing it for better comprehension.

1.13 Line 26: Could you specify which "close relationship" you are referring to? Providing more specific details will enhance the clarity of your statement.

1.14 Line 36-37: The transition between these paragraphs seems abrupt. It would be helpful to create a smoother link between the two paragraphs for better flow and coherence.

1.15 Line 51-52: When stating "few studies," it is important to provide references for credibility. Consider citing relevant works such as Ponce de León and Bettencourt (2021) and Gentile and Gray (2023).

1.16 Line 59: Please elaborate on the large-scale factors that you intend to investigate, specifically those associated with the formation of cyclone extreme winds.

1.17 Line 70: It is worth noting that 10-meter winds are diagnostic rather than prognostic variables, influenced by surface characteristics and the surface layer. A brief discussion on how this might impact the interpretation of your results would be beneficial.

1.18 Line 84: Consider rephrasing "storm destructiveness index" to storm severity index as it may not be the most appropriate terminology, as discussed in major comment 1.4. Refer to Leckebush et al. (2008) for guidance. Additionally, better ehighlight why ranking extreme wind speeds by their cube is preferable over 10-meter wind speed, given that the focus is not on evaluating insured losses.

1.19 Line 113: While the scheme itself is robust, it's important to acknowledge that the choice of tracking scheme may still influence the results. Refer to Ulbrich et al. (2013) and Messmer and Simmonds (2021) for insights into how the tracking scheme chosen could affect your findings.

1.20 Line 136: The exclusion of extreme wind events from extratropical cyclones with tropical origins may lead to the omission of a significant number of events. Please provide additional details, including how large-scale features differ between these two types of extratropical cyclones, apart from the temporal lag between cyclogenesis and extreme events.

1.21 Line 145-147: The intended message in this section is unclear. Please rephrase for better clarity.

1.22 Line 152: Could you clarify what you mean by "another part of the analysis"? Please rephrase this section to provide better context.

1.23 Line 160: Consider incorporating the pressure tendency formula and discussing its relevance in relation to the content in Section 3.4.

1.24 Line 200: In addition to numerical values, including percentages would provide a more comprehensive understanding of the data.

1.25 Line 250 and beyond: It would be beneficial to discuss this paragraph in connection with the equation presented in Section 3.4 once it is included.

1.26 Line 260: To more robustly compare the shape of distributions between the moderate and top extremes group, consider incorporating a non-parametric test for a more robust analysis.

1.27 Line 280 and beyond: An essential aspect missing from the discussion is an exploration of historical variability within the selected time periods (1950-1985 and 1985-2020). Investigating changes in contributions from the pressure tendency equation terms within these periods is crucial, particularly in light of the non-negligible global warming signal within the analyzed time frame.

1.28 Line 322-324: It would be valuable to discuss this aspect in relation to Hart et al. (2017) and Manning et al. (2022) to provide a more comprehensive context.

1.29 Line 339: Your statement regarding the significant differences between moderate and top extremes relative to diabatic contributions to the pressure tendency equation, is strong. To strengthen your argument, consider conducting statistical tests to support your findings, as mentioned in previous comments

The authors present an analysis of drivers of extreme wind events in the central North Atlantic. This is an interesting study because, as the authors state, this is often less analysed compared to high impact events which cause damage over Europe. The scientific presentation is excellent and results very clear. I have some comments, although these are mostly minor with some clarification or further analysis requested. Overall, I think this will be an excellent paper for WCD once the revisions are addressed.

Main Points

1. I wonder how specific you are on differentiating your "footprints". You state that if they are not connected in your analysis region then they are treated as separate. However, what if the situation arises that they are connected outside of the analysis region as part of a larger wind anomaly, how would this work and is this considered?
2. Throughout you use ERA5 anomalies from the 1979-2020 climatology, yet your features are identified from 1950-2020. Why this choice? Surely your climatology should match the time period which your events are taken from?
3. When are your top 1% events found in the event data? Do these align with times of historically high EU windstorm activity (1990s) or are they regularly interspersed throughout the historical record. Some information on this would be of interest to the readers of this manuscript.
4. L290-291 - you talk about how the PV field is representative of Rossby wave breaking composites, does it not make sense to demonstrate this yourself with composites of the RWB field?
5. You only show upper-level PV in your composites. It would be good to also see composites of lower-level PV. As you state it is likely that your top1% cyclones are somewhat frontal in nature, and so the lower-level PV anomalies may help strengthen that argument and provide further distinction from your bottom 10% category.
6. L145-148 - In the bottom 10% category you have 117 events, which is very similar in number to those from the top 1%. How is this the case considering there should be 10x the number of events? Is this because most of them are not associated with ETCs or TCs? Please clarify this in the text
7. L136 - are all of your top 1% events associated with ETCs or TCs? Or did you have to discard some that did not

General Points

• Equation 1: You need a term to state that D is only calculated for gridpoints where v_i>v_98i
• L90 - is the climatology used to create the 98th percentile just Oct-Mar or is this annual?
• Section 3.2 - i'm a little confused as to what time you used for the tracking (6hour or 1hour). This section could do with some re-writing to clarify this.
• L111 - proxy to relative vorticity - please change
• L161 - you have not yet shown that the surface MSLP decreases in this timeframe
• L163 - Most cyclones (again you have not yet shown anything for your cyclone analysis set)
• L216-218 - i would move this paragraph up, or at least talk about the jet streaks first before you start discussing the structure of the PV anomalies