| 01 Sep 2026
Extreme hazard cyclones – concept and ERA5 climatology
Abstract. Extratropical cyclones shape much of the weather variability outside the tropics and can be associated with extreme precipitation and damaging winds with potentially substantial socioeconomic impacts. Consequently, identifying and understanding the cyclone characteristics associated with widespread weather extremes along their tracks is essential. This study uses 75 years of ERA5 reanalyses to characterize “cyclone extremeness” by quantifying the occurrence of extreme precipitation and surface winds along cyclone tracks. We do this by linking precipitation and 10 m wind speed extremes (values exceeding the local climatological 99.9th percentile) to individual cyclones and quantify the number of extremes associated with each cyclone along its track on a global scale. Extremes are considered at two time scales, 1 hour and 3 days, reflecting that the impact of an extreme surface weather event may vary with its duration. An “extreme hazard cyclone” is then defined as a cyclone that accumulates a large number of precipitation and/or wind extremes along its track. Compared to all cyclones on average, extreme hazard cyclones form at lower latitudes and are deeper, larger in cyclone area, longer-lived, and more rapidly intensifying. Most of the top 100 extreme hazard cyclones globally form in the tropics and undergo extratropical transition. When restricting the ranking to cyclones producing extremes in the North Atlantic storm track, the top 100 extreme hazard cyclones predominantly occur in summer and autumn for precipitation extremes (including extratropical transition events) and mainly in winter for wind extremes. Extreme wind hazard cyclones typically form in strongly baroclinic environments at the left exit of intense upper-level jets. In contrast, extreme precipitation hazard cyclones are slightly less deep and form in weaker baroclinic environments, but they develop a prominent vertically coherent potential vorticity tower indicative of strong diabatic processes. At the time a precipitation or wind extreme occurs, extreme hazard cyclones are generally classified as ‘symmetric’ in the classical cyclone phase space and those linked to many wind extremes exhibit a ‘shallow warm core’. Long-duration extremes are associated either with more stationary cyclones or with cyclone clustering, both of which keep the affected region under cyclonic influence for an extended period, whereas short-duration extremes are linked to more intense and rapidly moving cyclones. The concept of extreme hazard cyclones, as introduced in this study, reveals key characteristics of cyclones associated with many precipitation or wind extremes along their track, and it provides a framework to investigate how climate change influences cyclone-driven surface weather extremes. The concept can easily be applied to specific regions of interest and extended to study cyclone impacts by including particular fields of exposure.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Weather and Climate Dynamics.
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.
The study on "Extreme hazard cyclones" by Voigt et al. mainly addresses the question, in which way the most intense or hazardous cyclones differ in typical properties like wind and precipitation intensity, extent, lifetime, (thermo-)dynamical structure and others compared to "normal" cyclones which they refer to as climatology. Basis to approach this is the analysis of the global ERA5 reanalysis with some deeper investigations over the North Atlantic. The paper is clearly structured, and the results are supported by the used figures and tables. The topic is highly relevant for various applications and for a better theoretical understanding what makes a cyclone more hazardous as others, for example to improve forecasts, and fells within the scope of WCD.
I therefore suggest a publication after what I consider as minor revisions, as most of my comments related to a better wording and description. Please see my detailed comments below.
General comments:
1. At several points within the text, I highly encourage the authors to sharpen the text (see also details below) regarding method description or what is actually meant. This would generally help the reader which is probably not that deep into this topic to better understand what is done and what do the authors want to say, which will also improve the readability. Currently, the reader needs more than usual comprehensive background knowledge regarding methods and cyclones.
2. Abstract and Introduction suggest that there will be a profound comparison between the top hazard storms and the "normal" storms, referred to as climatology. Beside the analyzed properties in Fig. 5, I do not see this kind of comparison. Most of the time, the authors just present the results of the top storms. You could add such analysis to a supplementary, for example.
3. I have doubts that mixing tropical and extratropical cyclones in this sense is helpful. In my opinion, the tropical ones will always dominate the heavy-tail of the statistics of wind and precipitation. Maybe it would be helpful to either focus on extratropicals, only, as partly done in the focus section on the North Atlantic, or to do the analysis separately for tropicals and extratropicals.
Detailed comments:
L.12: what do you exactly mean by "North Atlantic storm tracks"?
L.28: "over continents and within storm tracks" This sounds like the tracks are not over the continents. Generally, you should once define what you mean by "storm tracks".
L.31: how do you define "intensity" of a cyclone? Is it minimum SLP?
L.35: again, what do you mean by "storm tracks". Is it the center line of the pressure minimum, or a certain area around, or . . . .
L.72: "number of surface extremes associated with each cyclone along its track" I only understand this after I had read the entire paper, maybe already state here how you count extremes?
L.73: "large number", also here. Intuitively one cyclone has extreme precipitation and this would be one extreme, but you count the grid points exceeding a threshold, so probably rephrase this or define already here.
L.103: Why March 1950 to March 2025? ERA5 start in 1940 and depending on when you started the analysis, why not full 2025?
L.104: Did you do the interpolation? Can you provide some more details on the method you used? Furthermore, why did you choose a 0.5° grid. The original resolution is roughly 0.3°
L.110ff: The cyclone area from this approach can be quite large, right? Which can also be seen in Fig.1. Are there any further restrictions to limit the extent?
L.124: the 99.9the percentile is quite high, which limits the data amount in an only 75-year-long record. Can you provide some words as justification why you used this high percentile and not a lower one (95, 99, 99.5)?
L.127: do you use a running 72h window on the hourly data or is it blockwise meaning e.g. day 1-3, day 4-6, etc.?
L.127: So you first do the averaging or accumulation and from this new values you calculate the 99.9th percentile as basis for your analysis?
L.146: Again, what do you mean by "intensity"?
Sect. 3.1. It would be nice to see an example (case study) for a 72h case (e.g. in a supplement) to get a better insight to this part of the analysis
Fig.2: Especially in g-i it is hard to distinguish between the land contour and the storm track as both a black solid outlines. Could you try, e.g. gray land contours or something else that better separates the two things?
L.218ff: So did I understand this correctly: you have 4 categories with each 100 storms, which is a total of 400. Removing duplicates, 274 individual storms remain, and out of these 188 are then only attributed to a single category?
L.221ff: The given TCs: Do they remain TCs throughout their entire lifetime or do some of them undergo an ET?
L.225: Do you mean, there is only one cyclone assigned to the combination of P1 and W72 or do you mean one cyclone for each category?
L.271: Only one time step seem rather low for me to take a storm into account for the NA region. How often does this low threshold really occur, and wouldn't it be more robust to define a combination of a certain storm area for at least one time step for consideration?
L.290: "larger deepening rates than climatology" - do you mean mslp? Should be stated or defined somewhere. It would also be nice to have some exemplary numbers of the rates, so what actually is larger than climatology?
L.305-308: Description of an additional metric - isn't this better placed in the method section?
L.312/313: "the majority of extremes occur during intensification phase" - where can I see that/from which result stems this?
L.325: "at least parts if the cyclone" - can you specify this more? Is it a certain percentage?
L.330: "95th percentile" - so you define extremes by using 99.9%, and now you look at the "ensemble statistics" and analyze the 95%, right? So these are the extremes of the extremes?
L.340: 22m/s - is it also the 95th percentile?
Fig.6: I think, parts of the caption are better placed within the text, especially the parts linked to the method of selection. Furthermore, you use here now a threshold of 10%, which is fairly low in my opinion. In the main text you discuss the 95th percentile, so it gets a bit confusing now.
Fig.7: "zonal vertical cross-section" - is this through the storm center?
Fig.9: From the caption, I don't really get what is shown in (a, b) and what is then the difference to (c, d). Can this be improved?
L.488: In my opinion, it is not necessary to repeat the research questions again. Maybe just say something like "with respect to the initially given RQs, the following main conclusions can be drawn from the results"
L.496: With respect to one of my general comments: You conclude here on differences to the climatology, but this was never shown.