The paper addresses an interesting topic which fits WCD’s scope, and presents a robust set of analyses. While generally well-structured, the text is plagued by a large number of typos and shoddily formulated sentences, and a thorough proof-reading is needed. In a few cases, the authors are also excessively adamant in their statements. Once these issues are solved, I believe that this study will provide a welcome contribution to the literature on the drivers and maintenance mechanisms of European hot temperatures.

General Comments

1. The motivation of the paper is presented in a somewhat misleading fashion. The authors repeatedly highlight the impacts of heatwaves as motivation for their work, but then focus their analysis on moderate events (which they in fact term “hot spells” rather than heatwaves”. I do find the analysis of the hot spells interesting, but I would suggest a more nuanced framing of the introduction. In particular, the framing should recognise that the events the authors analyse are not necessarily high-impact ones from a societal viewpoint. If impacts were the only consideration, then the authors should revise their definition of hot spells.
2. While I understand that the focus of this study is on persistence rather than intensity, it would still be interesting to provide some information on the intensity of the events that you analyse in the main text, e.g. in Sect. 3.1. This would enable some understanding of what a “hot spell” entails in practice, and enable to relate the events being analysed to the typical intensity of recent European heatwaves.

See also my comment on ll. 262 – 264 below, which could be elucidated by comparing the severity of short versus long spells.

1. Throughout the paper, the authors equate persistence or stationarity of the large-scale flow with duration of the hot spells. However, this relationship is not as straight forward as the authors implicitly make it. For example, Holmberg et al. (2023) argued for a nuanced and variable relationship between the persistence of the large-scale flow and European temperature extremes.

Holmberg, E., Messori, G., Caballero, R., and Faranda, D. (2023) The link between European warm-temperature extremes and atmospheric persistence, Earth Syst. Dynam., 14, 737–765.

Specific Comments

Note that the list of typos and poorly phrased passages below is not comprehensive. I would recommend a thorough proof-reading of the text before a revised version is resubmitted.

l.6 This is poorly phrased. When you write “Temperatures are averaged across these regions for an analysis of hot spells in SW and W Europe.”, it sounds as though you use temperature information from all 6 regions for the analysis of 2 of them.

l.13 It is unclear why a zonal flow must necessarily mean a more transient flow. In fact, there are studies that argue that the zonal large-scale flow is the “default” state of the atmospheric circulation, with blocking representing an unstable fixed point (e.g. Faranda et al., 2016).

Faranda, D., Masato, G., Moloney, N., Sato, Y., Daviaud, F., Dubrulle, B., & Yiou, P. (2016). The switching between zonal and blocked mid-latitude atmospheric circulation: a dynamical system perspective. Clim. Dynam., 47, 1587-1599.

l.29 Do the authors mean excess deaths?

l.33 The citation of Tuel and Martius (2024) is misleading. The paper does not analyse impacts, and does not provide any evidence of a link between duration of hot spells and severity of impacts, beyond citing other literature on the topic.

l.49 “that” should be “the”

ll.64 – 65 This is an incomplete sentence.

l.100 Why start from 1959? This is neither the start of the available ERA5 data nor the start of satellite assimilations. Also, since the authors struggle with sample size, adding the summers of 2023 and 2024 could provide a few extra persistent hot spells.

l.114 This certainly works well for temperature and other smooth variables, but given that only 8 years are used, is this enough to give a smooth climatology also for noisy variables such as TP?

l.121 Why do the authors choose non-overlapping periods?

l.123 95^th percentile of what? Of the distribution of 3-week averaged temperatures? Please clarify in the text.

l.138 Verify that the figures are referred to consecutively in the text. Fig. 2 appears to be referred to before Fig. 1.

l.145 “and” should be “to”’

l.150 Why do the authors select a threshold in units of standard deviations here, but used a percentile threshold for the regionalisation? Also, the 1 sd threshold comes out of nowhere, and some context for it should be provided; for example, a sentence stating explicitly that a 1 sd threshold is used to define the hot spells.

l.158 The reference to Tuel and Martius (2023) is again potentially misleading here. The authors again use a paper that does not conduct any explicit analysis of impacts to support a sweeping statement on impacts of hot spells. It would be more instructive for the reader if the authors cited some of the papers that Tuel and Martius (2023) use to support the statements that they make on impacts.

l.185 Here the authors state that they “consider grid cells significant if they are identified as such in at least two-thirds of the 100 subsampled composites”. However, in some of the later figures they use two different types of stippling to highlight both the 1/3 and 2/3 threshold cases.

l.187 Here the authors could add a clarification that the stated confidence levels for the long spells should not be interpreted at face value, since the authors show cases where a given fraction of a sample satisfies a given confidence level. In other words, the stippling in the later figures for the long spells cannot be interpreted statistically in the same way as a 95% confidence level normally would be.

l.203 This makes the choice of regions that the authors focus on in the main text debatable. Why pick the two regions with the fewest long spells as focus regions if the authors repeatedly highlight sample size as a challenge for their analysis?

Fig. 3 I find the “Q” labels along the y-axis in panel (b) misleading. I would assume that these refer to “quantiles” or “quartiles”, while according to the caption they instead signify deciles.

l.240 What do the authors mean by “contrasting spread”?

l.241 “Bulk of values”. Do the authors mean bulk of events?

l.245 What does “significantly extreme” mean? The figure shows values significantly different from zero.

l.262 – 264 This is an interesting consideration. Does this suggest that long spells are long because they are weak while short spells may be short by virtue of their intensity, as the very high temperatures are a potential trigger for convection?

l.281 The authors should support their statement about persistence of the flow. Several of them have previously worked on this topic, and are certainly aware of appropriate references that could be used here.

l.285 “there is has”

l.286 Remove the extra full stop

l.298 Not everyone would agree with this statement, and in fact Lucarini and Gritsun (2020), identified blocking as not necessarily being a persistent configuration; see also my comment on l. 13. Holmberg et al. (2023) also argued against the notion that blocked large-scale atmospheric flows must be a priori anomalously persistent.

Lucarini, V., & Gritsun, A. (2020). A new mathematical framework for atmospheric blocking events. Clim. Dynam., 54(1), 575-598.

l. 304 Where can the reader find information on blocking size in the figure?

l.314 “VIPV”: define the acronym.

l.324 The statistically significant positive precipitation anomalies are also seen for the long spells, but the way the text is formulated presents this as a factor differentiating short hot spells from longer ones.

l.339 – 340 Why do the authors use a separate paragraph for these two lines?’

l.343 The sentence is missing an “a”

Fig. 7 It is very nice to see this summary figure, and it greatly helps the interpretability of the manuscript as a whole.

l.347 Both the text and the figure give prominence to RWB, but this is something that the authors never explicitly diagnose. I would recommend: clarifying that the RWB is assessed visually, or computing RWB with some existing algorithm, or toning down RWB in the summary text and figure presented here.

l.361 There is a missing space

Sect. 3.8 makes a number of interesting points and I enjoyed reading it.

l.367 “an summary”

l.393 Define the acronym “WN”

l.395 “strat”

l.410 “the factors their long duration”

l.411 An alternative to using climate models could be to leverage reforecasts, which may better reproduce the synoptic and large-scale features discussed by the authors.