Review Doensen et al.

This is an interesting paper, investigating the variability of Mediterranean cyclones in a prehistorical climate, using 3350 years of climate data from the CESM climate model. Given the length of the dataset, this could serve as a useful reference/climatology of the occurrence of Mediterranean cyclones. Although interesting, and generally clearly written with clear figures, I do have a few questions, mainly related to the choices made related to analyzing extreme precipitation and extreme wind cyclones.

Major comments:

1. Given a baroclinic growing cyclone, I would expect maximum precipitation (rates) on average around the time of maximum intensification (e.g. Papritz et al., 2021). This is also reflected in the precipitation composites in e.g. Figure 8, where the maximum precipitation for the extreme wind composites (Fig. b and c) is before the minimum pressure is reached. That the precipitation maximum is still at the minimum pressure for the extreme precipitation composites, is probably due to the sampling on the maximum precipitation, as the authors also remark themselves. That the cyclones grow on average baroclinically is visible in some of the composites a westward tilt of the minimum pressure with height. Can the authors argue why they did made this choice?
2. The minimum pressure is not the best way to assess cyclone strength, since it depends on both the latitude (since there is a equator to pole gradient of pressure) and the size, since a minimum pressure does not necessary indicate a strong gradient of pressure. Can the authors argue why this would be a sensible choice, and have they tested for example using the gradient of gpm per 1000 km, since given the cyclone detection algorithm they have this information available?
3. In several parts of the paper (e.g. line 50) the authors argue that the modes of variability influence the occurrence and strength of Mediterranean cyclones. This would suggest some causal link. However, I would argue that it is merely a correlation, as the authors also write for e.g. the link to the NAO (line 48). I would therefore suggest rewriting the text at these points a bit.

Minor comments:

Line 8: Does this variability refer to the variability of cyclone frequency?

Lines 47-49: This is an almost exact repetition of the sentence before, so I would suggest rewriting it and remove one the lines. I think this sentence is clearer than the previous, so I would suggest keeping this sentence.

Line 70: Does they refer to proxies?

Line 92: The focus on extremes could be motivated a bit more since it is a vital part of the manuscript.

Line 103: 1.9 x 2.5 degree longitude and latitude respectively?

Line 119: Given that you use the Z1000, you probably detect minima in the geopotential (height) and not pressure directly?

Line 121: Local minimum or averaged over the 1000 km?

Line 124-125: I do not understand it completely, since as far as I understand the authors track the cyclones on a 6-hourly resolution, so why is this criterion applied daily?

Line 143: Are these eruptions used, independently where they occurred on earth?

Line 144: Why is there a time frame of 5 years used before an eruption and only a time frame of 2 years after the eruption?

Line 154: This region could be indicated in one of the Figures.

Line 167: See above major remark, why not selecting on the time step of maximum intensification, since this is one would expect strongest precipitation rates?

Line 176: How many cyclones are detected in total, or in other words, what the is the fraction of selected ‘extreme’ cyclones?

Lines 200-205: The authors argue that the storm tracks are too zonal compared to ERA5 in the CESM model. However, if I would for example look at the DJF climatology, I would almost argue the opposite: there are relatively more cyclones detected at the northern side of the storm tracks, and less at the southern side. Can the authors explain why they argue that the storm tracks are too zonal in CESM?

Line 210: See previous remark, isn’t it more a northward shift of the storm tracks (away from the Mediterranean)?

Line 229: I might have missed this, but the 850 hPa temperature related to the cyclones is calculated in a certain area/radius around the cyclones? And the plotted temperature in Figure 4 is then the average over all cyclones occurring in a certain year?

Line 259: See remark above, I would be careful to describe these modes as ‘drivers of the circulation’.

Line 268: See previous remark

Line 287-288: I think this sentence could be moved to the methodology section, since it is not related to results shown.

Line 300: To what does the three different EXC types refer too? I assume it it the extreme precipitation, wind and compound composites? This could be further clarified.

Lines 320-321: I do not understand what the authors try to argue here, can the authors clarify their argument here?

Lines 337-338: Given this possible preselection bias, I would strongly suggest the authors to look at the sensitivity of this choice.

Line 352: Is Figure 8 then DJF?

Line 369: The authors write that the jet stream remains at certain position, which suggests that the jet stream remains at the certain position over a time period, but I don’t think that is what the authors mean.

Line 378: This already suggests that the (detected) cyclones grow baroclinically

Line 385: What is meant with an areawise more negative anomaly?

Lines 459-465: I would suggest to write the abbreviation EXC in full, since this probably would clarify the text.

Caption Figure 4c: I think it is a precipitation rate (in mm/6h), as also described in the label of the y-axis?

Figure 7 and elsewhere: I would suggest to make the text of the regions Central and Eastern Mediterranean bold, the first time I read the figure labels I was confused because I read them as ‘central Mediterranean longitude’

References:

Papritz, L., F. Aemisegger, and H. Wernli, 2021: Sources and Transport Pathways of Precipitating Waters in Cold-Season Deep North Atlantic Cyclones. J. Atmos. Sci., 78, 3349–3368, https://doi.org/10.1175/JAS-D-21-0105.1.

Review: “Characterization of the mean and extreme Mediterranean cyclones and their variability during the period 1500 BCE to 1850 CE” by Onno Doensen, Martina Messmer, Christoph C. Raible, and Woon Mi Kim.

Summary

The paper discusses characteristics of Mediterranean cyclones in a simulation with the Community Earth System Model (CESM) for a long period of 3350 years. The cyclone characteristics in CESM are first compared with the ERA5 atmospheric reanalysis for a short period of 30 years. Limitations are found in the representation of precipitation and wind speed and are attributed to the relatively coarse resolution of CESM. After a presentation of the internal variability, volcanic eruptions and circulation patterns are found to affect the North Atlantic more than the Mediterranean. Extreme cyclones are then compared between the central and eastern Mediterranean for rainy, windy and compound events. More intense cyclones are found in the central Mediterranean and the contrast is attributed to the different dynamical forcing.

The paper is generally well written and presents new results. It contains various ideas and approaches that will inspire the curious reader. However, the paper often looks like a portfolio of potentially interesting features rather than a consistent piece of research. A clearer focus should guide the reader throughout the paper, starting from key concepts in the introduction to more emphasis in the conclusions, possibly at the price of a restrained number of topics presented in the results. Also, inconsistencies in the methods and the lack of indications in the plethora of subfigures sometimes make the results challenging to appreciate. Comments are listed below to help improving the paper.

Major comments

1. The structure of the results section is imbalanced. On the one hand, the first half presents an overview of different aspects, where many results apply to the North Atlantic rather than the Mediterranean (Figs. 2, 5, 6). They may be relevant for the Mediterranean but this is not discussed. For instance, not much impact of volcanic eruptions is found, nor much link with atmospheric modes. On the other hand, the second half describes intense cases only. It is very detailed but contained in a single section. Some reorganization is required here, with important results highlighted and others streamlined.
2. Why are results compared between the western/central and eastern Mediterranean? Dynamical differences between these regions are not introduced, despite the large body of literature about Mediterranean cyclones. In contrast, Genoa lows, Vb cyclones, Sharav cyclones and Medicanes are introduced but not further discussed. Thus, the reader does not know what to learn from the comparison.
3. Why is CESM compared with ERA5 for the period 1980–2010 only and not for the full period 1940–present?
4. Methods are not always clear and definitions are sometimes repeated, which blurs the interpretation of results.
5. In the results section, please systematically refer to the figure and panel that is discussed. Currently, one has to guess where to look at. Also, figures tend to contain many panels, which are not all discussed and could likely be removed to focus on the main results.

Minor comments

l. 3 “their variability in the late Holocene is poorly understood”: more precisely?

l. 8 5% in what?

l. 9 the relation is described as “weak” in the conclusions

l. 24–25 What kind of variability and connection? This is the main motivation for the paper, thus requires (way) more details. Perhaps it is discussed below but it is unclear at that point

l. 40–41 References are expected here

l. 42 I don’t fully agree: see, e.g., the devastating cyclone Daniel of September 2023

l. 43–44 Feser et al discuss North Atlantic cyclones whereas Flaounas et al discuss Mediterranean cyclones

l. 46 Wintertime precipitation correlates negatively with NAO but positively with cyclone frequency

l. 65 What scales are referred to by short/long periods?

l. 84 What scale is referred to by low frequency?

l. 88 It would be worth citing and discussing the few studies, as the impact of volcanic eruptions on cyclones is investigated in the paper

l. 100 missing “in” Kim et al

l. 115 why not 1940–2024?

l. 125 I don’t fully understand the condition “across” 24 hours

l. 134 missing “in” Raible et al

l. 134–138 What is the meaning of fitting a Gaussian function to the geopotential field on a single grid point? Also, rather than describing first the “traditional” selection and then the adaptation, I recommend describing straight away what is actually used here.

l. 148 typo “significance”

l. 159 rendering issue

l. 161–165 not sure what “western Mediterranean” means here; actually the CMED region in Fig. 1 largely contains the region usually referred to as the West Med. Thus, for consistency with previous studies, and considering that only two regions are compared here, I recommend naming them West/East Med.

l. 166–177 The precipitation and wind speed metrics are repeated several times in the paragraph but in an inconsistent fashion (with/out max, with/out t0); I recommend defining them once for good and then simply using precipitation and wind speed (as in l. 177). Also, which radius is used along the track for precipitation and wind speed?

l. 173 what is the untransformed distribution?

l. 177 here and elsewhere: CMED and EMED to keep the same terminology

l. 180 t0 already defined

l. 185 missing “at” t0?

l. 186 variations in what? I don’t get the point here

l. 188 Why 30h? And missing “for” t0?

l. 191 are all fields averaged over the area of both regions for all cyclones?

l. 192 Cyclone tracks “that” did not appear (and no comma)

l. 195 Some details are expected about the RWP amplitude. And typo: Rossby wave “packet”

l. 200ff In the discussion on Fig. 2 please indicate which panel is referred to

l. 204 Fig. 2 suggests a northward shift rather than a zonal vs wavy storm track

l. 209–214 Any evidence for these assumptions? Otherwise they sound speculative

l. 216 “while” rather than “although”

l. 221 this is speculative but should be easy to verify

l. 224 where is it most often underestimated by 50%? it is unclear where this number comes from

Fig. 3 wind speed is the max value here, while it is defined as the value at t0 in Section 2.5

l. 234 the 30y running mean does not show similar patterns and is not appropriate for tendencies

l. 235 is this definition of cyclone frequency different from that used elsewhere in the paper?

l. 255–256 Repeats previous sentences

l. 260 “important” driver?

l. 263–265 Please refer to the corresponding panels

l. 266 Fig. 6 is already presented before

l. 287 this sentence is surprising, as sea ice anomalies are not discussed in the methods

l. 289 this should be mentioned earlier, and clarified that Fig. 6 shows DJF only

l. 303 the presence of the warm sector of the cyclone (I) could be verified and (II) does not dynamically explain the highest wind speeds (see, e.g., Raveh-Rubin and Wernli 2015, or papers for the North Atlantic)

l. 309–311, 320--321 Please refer to the corresponding panels in Fig. 7

l. 312–313 not only northward but also (obviously) westward! Western/central and eastern Mediterranean cyclones have different dynamics, which should be discussed in the introduction (see, e.g., Doiteau et al. 2024, or older papers)

l. 316 it is expected indeed; I don’t quite get the point at showing precipitation for windy cyclones and wind for rainy cyclones in Figs. 7–9

l. 327 Fig. 8 is already referred to on l. 308

l. 329 This is true for panels (a) and (d) but not (b) for instance

l. 330 I do not understand: there are triangles in the plots, indicating significant differences for EXC100 (also, it should be clarified that the symbols indicate statistical significance)

l. 335 This contradicts l. 320 (see comment above)

l. 337 This questions the relevance of the definition of EXCs (see above comments on methods

l. 349 typo: “cyclones”

l. 352–362 Why look at summer cyclones separately? This is quite a long description for a figure that is not shown in the paper

l. 365 See comment on l. 303

l. 380 why “often”? Why “again”?

l. 422ff references to specific figures are unexpected in the conclusion

l. 439 Please explicit, and clarify whether it is your result or arises from the cited study

l. 442 the wind and precipitation are also underestimated compared to ERA5

l. 451 an earlier study cannot confirm your current results: the other way round

l. 470 Flaounas et al. (2015b)

l. 471 and a very different time period!

l. 472 “region”: better “area” to avoid confusion with the east/west Med

l. 476 max???

l. 477 Homar et al. (2007)

l. 484 5% in frequency?