the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Jul 2026
Vertical structure and dynamical regimes of Mediterranean extreme warm events from surface to high troposphere
Abstract. Climate change is rapidly accelerating the frequency of Extreme Warm Events (EWEs) across the Mediterranean, yet their vertical thermodynamic structure and dynamical evolution remain poorly understood. This study investigates the three-dimensional patterns of EWEs over Rome during the 1991–2025 period using ERA5 temperature data at 1000 hPa (near-surface), 500 hPa (middle troposphere), and 300 hPa (high troposphere).
EWEs were identified through percentile- and persistence-based criteria and classified using a K-means clustering approach. Their evolution was further investigated through precursor analysis, synoptic composites, and HYSPLIT backward trajectories. Five Extreme Warm Regimes (EWRs) were identified, spanning from shallow boundary-layer warming to vertically coherent deep-column structures. These regimes naturally separate into two broader classes: vertically coupled events, characterized by warming throughout the tropospheric column, and vertically decoupled events, in which thermal anomalies remain confined to specific atmospheric layers.
Trend analysis reveals a marked increase in the frequency of vertically coupled regimes after 2015, particularly during summer and autumn. While the average duration and mean thermal intensity of episodes have remained remarkably stable over the 35-year period, the higher recurrence of deep-column anomalies is effectively extending summer-like atmospheric conditions into autumn months. Representative case studies show that the coupled regimes develop through different pathways, involving persistent blocking, subsidence, and/or subtropical warm-air advection, while the remaining regimes maintain persistent vertical decoupling throughout their evolution.
Overall, these findings indicate a progressive climatic transition toward more vertically organized and synoptically controlled EWEs, providing key information on the physical nature of Mediterranean heat extremes that cannot be inferred from near-surface temperatures alone. This three-dimensional framework offers a physically based approach for interpreting, classifying, and improving the prediction of EWEs in Mediterranean environments.
- Preprint
(1993 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 09 Sep 2026)
- RC1: 'Comment on egusphere-2026-4108', Anonymous Referee #1, 27 Aug 2026 reply
-
CC1: 'Comment on egusphere-2026-4108', Heini Wernli, 03 Sep 2026
reply
Dear Annalisa Di Bernardino and colleaguesI had a look at your recent submission to WCD and while I find the topic and approach promising, I would like to suggest some important improvements related to (i) the representativity of the results, (ii) the definition of extremes, and (iii) the consideration of recent relevant publications on the topic.With best regards,Heini Wernli
1) The title is potentially misleading (not many would call a situation an "an extreeme warm event" when the temperature anomaly is only in the upper troposphere); also since you only consider a single location (Rome) the title appears too general (the eastern or southern Mediterranean might behave differently).
2) It is not clear from the abstract how you identified extremes (absolute threshold, percentile criterion?); it would also be good if you explained already here how you defined "duration" and "thermal intensity".
3) You did not discuss your results in the context of recent studies about the vertical structure of heat extremes and processes leading to heat extremes, for instance:
Hotz et al. 2024. https://wcd.copernicus.org/articles/5/323/2024/
Hotz et al. 2026. https://egusphere.copernicus.org/preprints/2026/egusphere-2026-1522/
Röthlisberger, M., and L. Papritz, 2023. Quantifying the physical processes leading to atmospheric hot extremes at a global scale. Nature Geosci., 16, 210–216.
Fix et al. 2026a. https://wcd.copernicus.org/articles/7/17/2026/
Fix et al. 2026b. https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2350/
4) L26: not clear how your results can contribute to better predictions. This aspect is not considered in the paper and therefore should not appear in the abstract.
5) L45: the text implies that Arctic amplification leads to more blocking over Europe; I don't think that this is supported by the literature.
6) L60: I fully agree that such events exist (shallow warm layers in the upper troposphere are very common), but I am lacking the motivation why they should be discussed together with temperature extremes at the surface (that might have a complex vertical structure). They do not affect societies and ecosystem, so they clearly play a different role compared to surface extremes.
7) L69: I don't understand the distinction between EWE and EWR.
8) L75: this aspect is also discussed in Hotz et al. 2026.
9) L173/L304: it takes a long time until the reader understands how you defined EWEs. With the 90th percentile you use a very soft criterion and identify 1371 EWE days in 35 years, corresponding to on average about 40 EWEs per year! To me, this indicates that these are not really extreme events, as they occur about once a week. I suggest to avoid the term "extreme" when you investigate the 1371 events, or even better, you apply a much stronger percentile threshold to really focus on rare / extreme events.
10) I think the paper would be much stronger if the analyses were extended to a few more locations across the Mediterranean. A single location always bears the risk of not being representative.
Citation: https://doi.org/10.5194/egusphere-2026-4108-CC1 -
RC2: 'Comment on egusphere-2026-4108', Anonymous Referee #2, 04 Sep 2026
reply
This manuscript examines the vertical structure of persistent warm anomalies over Rome using ERA5 data and a clustering approach. I find the effort to look beyond surface temperature interesting, and the combination of climatological analysis, composites, precursor evolution, and case studies has potential. I have, however, several questions about the motivation for the framework, some methodological choices, and the interpretation of the results. Addressing these would make the paper clearer and more convincing.
Main comments
- Motivation for the vertical framework
I think the manuscript would benefit from a clearer explanation of why it is useful to include events confined to the middle or high troposphere that have no clear surface signature. For surface heatwaves, the relevance is immediate, but it is less clear to me why a persistent warm anomaly at 300 hPa should be considered an “Extreme Warm Event” in the same sense. What is the physical or practical motivation for focusing on these high-level events, and what do they add to our understanding of surface heat extremes?
The introduction should also discuss more fully the growing literature on the vertical structure of heat extremes. In particular, the recent study by Hotz et al. (2026), Global characterisation of the vertical temperature anomaly structure of heat extremes over land in ERA5, seems highly relevant.
- Choice of pressure levels and seasonality
The dataset contains 15 pressure levels, but the event definition and clustering use only 1000, 500, and 300 hPa. I would like to see a stronger justification for this choice. Why not use the full vertical profile for the classification, especially given the emphasis on three-dimensional structure and vertical coupling? With only three levels, it is difficult to know how well the identified categories represent the intervening atmosphere.
I also wonder how strongly the results depend on season. The vertical structure and physical mechanisms of warm anomalies are likely to differ substantially between winter, summer, and the transition seasons. This could affect the clustering and the composites. It would be helpful to show the seasonal distribution of events in each regime and discuss this issue more directly.
- Physical interpretation
Some of the dynamical interpretations currently feel more suggestive than demonstrated. For example, it is not clear to me how the authors distinguish a ridge from a blocking event in the composites. Similarly, (\Delta T_{850}) alone does not identify warm-air advection, and warm-air advection can also occur aloft. I think the paper would benefit from a more careful explanation of these interpretations, and possibly from additional diagnostics or anomaly fields.
More generally, I would like to see more discussion of the variability among events within each regime. The proposed mechanisms are plausible, but it would be useful to know how robust they are across events and seasons, rather than primarily relying on a single selected case study.
- Trends and recent changes
I was not fully convinced by the interpretation of the recent increase in coupled regimes. With a 35-year record, how can the authors distinguish a long-term trend from decadal variability? It would be helpful to quantify uncertainty in the changes in regime frequency and to discuss this limitation more explicitly.
Figure 6 also does not seem to show a clear systematic increase in CHI for the deep regimes, despite the text suggesting that it does. Similarly, the conclusion that blocking or synoptic control has become more important would need additional support. Do blocking events themselves become more frequent or persistent, or is this inferred only indirectly from the temperature structure?
Specific comments
- Abstract, lines 8–10: Why are only three levels used to describe the vertical structure? This choice should be justified already in the abstract or more clearly in the methods.
- Introduction, lines 54–60: The statement that most studies treat heatwaves as exclusively surface phenomena seems too broad. A number of recent studies examine the vertical structure of heat extremes; these should be discussed more fully.
- Lines 64–79: The distinction among heatwaves, EWEs, and EWRs is useful, but I still found the motivation for high- and mid-tropospheric EWEs unclear. Could the authors explain more directly why these events should be considered alongside surface heat extremes?
- Lines 108–110: Which season is being discussed here? The role of the sea breeze and urban heat island is likely to be quite season dependent.
- Lines 125–130: The paper uses the full 15-level profile for case-study cross-sections, but only three levels for identification and clustering. Why not use the full column for the classification itself?
- Lines 132–140: Please clarify the use of 1000 hPa over this region, given topography and the fact that it is not always equivalent to a near-surface or PBL level. Also, the phrase “advancing subtropical air masses” is not fully clear to me. Is this meant to refer specifically to lower-tropospheric air masses? Warm-air advection can occur aloft as well.
- Lines 150,164: A brief plain-language description of the Mann-Kendall test and Sen’s slope would help the reader understand what each diagnostic adds.
- Lines 186: Please explain the motivation for the four-day persistence criterion. Many heatwave definitions use three consecutive days; why was four days selected here?
- Table 1: I would be cautious about interpreting “Surface Only” events as “typical surface heatwaves.” Many conventional heatwaves do have a clear free-tropospheric and circulation signature. Likewise, the physical interpretations assigned to the Mid Troposphere and High Troposphere categories seem somewhat speculative and would benefit from further explanation.
- Lines 215–225: I did not fully follow the interpretation of (\Delta T_{850}). Why does a lower (\Delta T_{850}) in the presence of high (Z_{500}) imply that local vertical dynamics are more important? Please explain this reasoning more carefully.
- Section 3.2 / Figs. 2–3: The titles should be clearly distinguished from those in the methods, since this section presents results rather than the identification procedure. The captions for Figs. 2 and 3 are also too brief. Please define the colours and categories clearly and provide enough information for the figures to stand alone. I found it difficult to see directly from Fig. 3 the claimed seasonal expansion and transition from high-level precursor stages to full-column events.
- 4: Would it be more appropriate to show standardized anomalies, or at least discuss how differences in temperature variance with height affect the profiles and clustering? It would also be useful to discuss the fairly large spread within some regimes.
- 5 and associated text: Would geopotential-height anomalies be more informative than absolute (Z_{500})? As shown, it is difficult to assess whether the patterns represent anomalous ridging or blocking. The figure caption should also be expanded, and the basis for distinguishing a ridge from a block should be clarified.
- 6 and lines 419: I do not clearly see the stated increase in CHI in the recent period. Please revisit this interpretation. It would also be useful to discuss uncertainty, variability, and whether the apparent changes could reflect decadal variability rather than a robust trend.
- 7: The precursor analysis is interesting, but the physical explanation for the apparent cooling aloft in the Low Troposphere regime and at 500 hPa in the High Troposphere regime is not clear to me. Why do these layers cool as the event onset is approached? Showing the spread among events, rather than only the mean evolution, would also be helpful.
- Case studies: The case studies are useful illustrations, but it would help to explain more clearly why the selected cases are representative of each regime. For example, what is the variability among other events in the same regime? The language of “validation” may be too strong for a single case per cluster.
- 8: Please indicate (t_0) directly in each panel. It would also help to explain the contouring and make clearer how the inferred downward or upward development is identified.
- 9: Please identify which case each row represents. The time direction in the trajectory panels is somewhat confusing relative to the spatial tracks, and I found it difficult to see the claimed differences among the trajectories. A clearer presentation would help.
- Discussion and conclusions: The conclusions about increasingly synoptically controlled events, increased blocking, and reduced sea-breeze effectiveness seem stronger than what is directly shown. Could the authors either provide additional evidence for these interpretations or state them more cautiously?
Overall, I think the manuscript would be strengthened by a clearer motivation for the high-level events, a more fully justified treatment of vertical structure and seasonality, and a more cautious interpretation of the dynamical mechanisms and recent trends.
Citation: https://doi.org/10.5194/egusphere-2026-4108-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 172 | 61 | 20 | 253 | 22 | 14 |
- HTML: 172
- PDF: 61
- XML: 20
- Total: 253
- BibTeX: 22
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of the manuscript “Vertical structure and dynamical regimes of Mediterranean extreme warm events from surface to high troposphere”, written by Bernardino et al. for Weather and Climate Dynamics (WCD)
The manuscript explores the recently introduced approach to define temperature extremes as three-dimensional phenomena for the Mediterranean region (Rome, Italy) and analyses driving mechanisms of identified Extreme Warm Events (EWEs). In my opinion, the manuscript is well-suited for Weather and Climate Dynamics journal, as it extends current knowledge of vertical characteristics of temperature extremes from both statistical and dynamical perspective. The manuscript, however, contains several problematic parts, which should be clarified before consideration its publication.
1) Classification of Extreme Weather Events vs. Extreme Weather Regimes
Authors classified identified EWEs into 5 categories based on thresholds for 1000, 500, and 300 hPa levels. This is not an issue, we have used the analogous method, but the authors also introduce additional classification of Extreme Weather Regimes (EWRs) using K-means clustering. If I understand correctly, EWEs categories were used solely for Figure 2 and 3 (statistical perspective), while the remainder of the manuscript used EWRs. I am missing the link between these two approaches, for example which of the EWRs interprets Surface only EWE? Judging from Figure A2, this EWE falls into Low Troposphere, Standard Deep and partly to Upper Troposphere (!) clusters. This also points to a drawback a threshold-based methods. To sum up, in my opinion, classification of EWEs is redundant and potentially misleading, as EWE categories are not an input for EWR clustering. Authors may consider redrawing Figure 2 and 3 with EWR categories instead of EWE and omitting EWE classification from the manuscript. K-means would probably provide a more objective method to distinguish between vertical types of temperature extremes.
2) Number of EWR categories
Judging form Figure 5, the “Standard deep” and “Low Troposphere” EWR categories are very similar. They have also quite similar vertical profile (Figure 4) and precursor thermal animacies (to a lesser extent; Figure 7). I understand that from Figure A1b that the five clusters seems to be the most reasonable option but have authors tried using four or six EWRs?
3) Potential seasonal differences
Authors identify EWEs thorough the year (Figure 3), which is fine. But although they present temperature trends and EWEs statistical characteristics in different seasons, EWRs were determined from all events regardless a season. I am not an expert on the Mediterranean climate, but I can imagine that the stormier, moister and colder part of the year may affect development of EWE differently than the hot and dry part of the year. I am wondering if the number of EWR categories and their driving mechanisms would be different if EWEs would be partitioned between hot and cool part of the year.
4) Minor edits
Figure 1: In the “Detection of extreme warm events” panel, 90th percentile threshold, the red broken percentile line is drawn as horizontal. I believe that a proper representation of the threshold should be curved, as the threshold follows the annual temperature cycle.
Lines 179–180: “This moving-window approach reduces the influence of short-lived weather systems while providing a smooth representation of the seasonal cycle.” This sentence is confusing for me. If I understand the method correctly, the moving window was used only for a calculation of the thresholds. The short-lived EWEs are excluded due to a persistence criterion (four days; Line 185). Consider reformulating the sentence.
Table 1: I would omit the subjective meteorological interpretation, also because authors did objective analysis of EWRs below.
Figure 2 and 3: Please enlarge plot axis titles and texts. Also legends are hard to read, consider using a larger font.
Figure 6: In my opinion, there is not much added value in this figure. The upper panel (EWR frequency) could be integrated into revised Figure 2 see my first comment and the remaining panels could be omitted as the changes between time periods are relatively small. Also how interpretable are upper tropospheric positive temperature anomalies in 1996–2000 vs negative upper tropospheric temperature anomalies in 2001–2005?
Figure 7: Very hard to read, please enlarge labels and a legend.
Table 4: It reads “19/01/1997-22/01/1997”, while Figure 8a starts with 19/01/07, probably a typo.
Figure 8: Duration of x-axis varies between individual cases, which may mislead the reader about the case persistence. Consider aligning the x-axis between cases.
References: Line 268 (Draxel and Rolph, 2010), the name should be Draxler (typo) and the year (2010) do not match the year in the references (2026); Line 571 (Lhotka et al., 2018) is not in the reference list.