A pan-European analysis of large-scale drivers of severe convective outbreaks

Authors: Feldmann et al.

Severe convective outbreaks (SCO) are important drivers of weather-related damages in Europe. Hence, it is important to understand which large-scale drivers establish convectively favorable conditions to SCO in different regions, offering new opportunities in the applications of predictability and the understanding of climate processes since synoptic-scale patterns have much greater predictability than local convective precipitation. Although more work is required to establish a reliable synoptic proxy as an alternative to current CAPE-shear-based analysis, the scientific outlook is promising.

The paper nicely deals with this topic, identifying three macro-regions, affected in a different way by the large-scale patterns, and discussing land-surface conditions that are generally not considered in this kind of studies. The presentation is overall clear, and I like the concise and essential writing style. Thus, I think that the paper can be accepted after some relatively minor modifications, related both to improvements in the presentation and in the discussion of the results.

Major points:

- L212-217: What I really miss in the article is a quantification of your claims. I think adding a table (even in the Appendix) to quantify the anomalies in Figures 3 and 4 and the trends in Figure B1 of the different variables in each identified region could help support your analysis. For example (L279), you mention “a pronounced drying” in the humidity-limited regions, but I do not see a more pronounced drying than in the temperature-limited regions: can you quantify the trends in a table? This would make the summary and discussion in Section 4 more robust.

- L272: I think that here and in most of the paper there is an excessive emphasis on relative humidity rather than on specific humidity; however, RH is not a measure of humidity but rather of the proximity of the environment to saturation. So, why not adding specific humidity in Fig. 7?

- I do not understand the need to put Figure B1 (and in minor way Fig. A1) in the Appendix, as it is an important part of the study.

Minor points:

L11: delete “climate”

L48: which is the average duration of a persistent event?

Table 1: why is the average and not the maximum 2m temperature considered?

Figure 1: I do not think the selected names are representative of the regions considered: what about BA (Balkanic peninsula) or E-M (eastern Mediterranean) instead of HE, C-M (central Mediterranean) instead of AD, N-W (north-western region) instead of BE?

Figure 1 caption: panels b) and c) are reversed between the figure and the caption

L86: I think that “To calculate the anomalies, ...” should be put before “All descriptor variables”

L106: Also, squall lines frequently cross the Po valley from west to east (e.g., De Martin et al. (2024))

De Martin F., Davolio S., Miglietta M. M., and Levizzani V., A conceptual model for the development of tornadoes in the complex orography of the Po valley, Mon. Wea. Rev., 152, 2024, 1357-1377, https://doi.org/10.1175/MWR-D-23-0222.1

L108: does a high value of SCO necessarily mean that an outbreak occurs? Please shortly discuss this point

L119: “when absolute temperatures are high” is redundant

L124: “left column” instead of “top row”

L132: “climate average” instead of “baseline”

L132: please add “anomalies” after “much less temperature”

L133: forced lifting depends also on CIN, which is not considered here

L149-152: how do you explain the high variability in HE?

L157: really, only the positive temperature anomalies are mirrored in SST

L159: the sentence is ambiguous: in my opinion, all regions lie downstream of positive SST anomalies but only of the Mediterranean Sea. This makes some confusion with the following sentences.

L166: add “anomalies” after sea-surface temperatures

L206-208: although not necessary, it may be interesting to investigate the sea surface fluxes to support this hypothesis.

L237-249: all this part can be significantly reduced; there are some points repeated a few times.

L262: “Given the absolute thresholds required for convection, we refrain from removing long-term trends from the data.”: it is not clear to me what you mean here

L269: “or decreases only slightly (Alta-Italia, Central Europe)”: I would say it occurs in Slavic area rather than in Central Europe

L273: “upstream”: or downstream? Or both upstream and downstream?

L284-286: I think this is an important conclusion of the paper and deserved a bolded font

L314: remove “have”

Review of: A pan-European analysis of large-scale drivers of severe convective outbreaks

This study investigates the large-scale drivers of severe convective outbreaks (SCO) in Europe. SCO events cause significant damage to infrastructure, property, and human health and life, and they occur regularly across Europe. However, while a well-defined model of severe convection exists for the US, other extratropical regions like Europe still lack a comprehensive framework. This paper therefore represents an important step toward improving SCO predictability, and even understanding how these events may evolve under global warming.  

The authors detect convective events using a CAPE-shear threshold binary variable (CIX) and cluster them into regions where SCOs occur simultaneously. These regions are then grouped into three categories based on the dominant climatological perturbation driving SCO initiation during the extended summer. Using this framework, the paper examines the large-scale descriptors of convective outbreaks, differences between short-lived and persistent SCOs, and trends in these descriptors.  

The paper is well-organized, clearly written, and methodologically sound. I particularly commend the authors for their rigorous use of statistical inference, including the application of False Discovery Rate (FDR) correction for gridded testing—a theoretically necessary but often overlooked practice in climate studies.  

Overall, the manuscript requires only minor revisions. I have read the comments from Reviewer #1 and agree with their suggestions, particularly regarding the inclusion of Appendix figures (especially B1) in the main text. Below are my additional remarks:  

• Line 70: The use of CIX alone is one of the relatively weaker aspects of the study. As the authors mention comparing CIX-based clustering with lightning data, I recommend including this comparison in the Appendix for transparency.  

• Line 78: The authors note that they tested different CAPE and shear thresholds for defining CIX. It would be valuable to demonstrate the robustness of the results to these threshold choices, especially given the lack of validation for CIX as a convection proxy using observations.  

• Line 138: The statement “In each region, we find a maximum of convective precipitation on the day of the SCO” is technically correct but initially confusing when examining Fig. 3. At first glance, one might expect the precipitation maxima in panels (first two rows) to align strictly with the detected SCO regions. Instead, the key finding is that each region experiences its peak convective precipitation during its respective SCO events. I suggest clarifying this point and briefly discussing overlaps between regions (e.g., W-M/AL SCOs sometimes coincide with convection in BE/CE or SL/AD).  

Additional minor corrections:

• Line 193: 900 hPa → 800 hPa  

• Line 163 & Fig. 7 caption: Replace “insignificant” with “non-significant” or “not statistically significant” for precision. “Insignificant” can be misleading, especially when used in a text with no reference to the result of a statistical test as at Line 163.