A numerical study to investigate the roles of former hurricane Leslie, orography, and evaporative cooling in the 2018 Aude heavy precipitation event

In southeastern ::::::::::: south-eastern France, the Mediterranean coast is regularly affected by heavy precipitation events. On 14–15 October 2018, in the Aude department, a back-building quasi-stationary mesoscale convective system produced up to about 300 mm of rain in 11 h. The synoptic situation was perturbed by the former hurricane Leslie , :: At ::::::: synoptic ::::: scale, ::: the :::::: former :::::::: hurricane ::::: Leslie :::: was involved in the formation of a Mediterranean surface low that focused the convective activity ::::::::: channelled ::::::::::: conditionally ::::::: unstable :: air ::::::: towards ::: the :::: coast. At mesoscale, convective cells focused west of a quasi-stationary cold front ::::::: decaying 5 ::: cold :::::: front, ::: that ::::::: became :::::::::::::: quasi-stationary, : and downwind of the terrain. To investigate the roles of ::: the ::::::: moisture :::::::: provided ::: by Leslie, orography and evaporative cooling in the processes that led to the ::::: among ::: the ::::::: physical :::::::: processes :::: that ::: led :: to ::: the ::::::: location ::: and ::::::: intensity :: of ::: the observed rainfall, numerical simulations are run and evaluated with : at :::: 1 km :::: and ::::: 500 m ::::::::: horizontal ::::::::: resolutions ::: and :::::::: evaluated :::: with :::::::::: independent : near-surface analyses comprising standard and :::::::: including ::::: novel :::::::::::: crowd-sourced ::::::::::: observations :: of personal weather stations. Simulations show that, in a first part of the event, low-level conditionally unstable air parcels found 10 inside strong updrafts mainly originate from the Mediterranean Sea, east of 4.5 E, :::::::: updraughts ::::::: mainly :::::::: originated ::::: from ::::: areas ::: east :: of ::: the ::::::: Balearic ::::::: Islands, :::: over ::: the :::::::::::: Mediterranean :::: Sea, : whereas in a second part, an increasing number originates :::::::: originated from Leslie’s remnants. Air masses from east of 4.5 E appear :::: areas :::: east :: of ::: the ::::::: Balearic :::::: Islands :::::::: appeared as the first supplier of moisture over the entire event. Still, Leslie contributed to substantially moisten mid-levels over the Aude department, diminishing evaporation processes. Thus, the evaporative cooling over the Aude department does :: did : not play any substantial role 15 in the stationarity of the cold :::::::::::: quasi-stationary : front. Regarding lifting mechanisms, most of the air parcels found inside strong updrafts ::: the :::::::: advection :: of ::::::::::: conditionally ::::::: unstable ::: air ::: by : a :::::::: low-level ::: jet ::::::: towards ::: the ::::::::::::: quasi-stationary ::::: front, ::::::: confined :: to :::::::: altitudes ::::: below ::::: 2 km, ::::::::: reactivated ::::::::: convection ::::: along :::: and ::::::::: downwind :: of ::: the ::::: front. ::::: Most :: of ::: the ::: air :::::: parcels ::::: found :::::: inside ::::: strong :::::::::: updraughts near the location of the maximum rainfall are lifted above the cold front, attesting its key role in focusing convection :::: were ::::: lifted ::::: above ::: the ::::::::::::: quasi-stationary ::::: front. Downwind of the Albera Massif, mountains bordering the Mediterranean Sea, cells formed 20 by orographic lifting seem to be :::: were : maintained by low-level leeward convergence, mountain lee waves and a favourable directional wind shear; when terrain is flattened, rainfall is substantially reduced. The location of the exceptional precipitation appears to be driven primarily ::: was :::::::: primarily :::::: driven by the location of the quasi-stationary cold front and secondarily by the location of convective bands downwind of the orography.

coming from the Mediterranean Sea. Finally evaporative cooling did not seem to play a substantial role in the dynamics but only in the control of the total simulated amount of precip. The paper is well written and clearly structured however I feel it sometimes goes too long in the description of the dynamics of the event in a disproportionate way compared to the focus of the research. The analysis of Leslie's contribution and evaporative cooling is resolved in a short part compared to the long introduction on the analysis of the event and the methodology of 20 analysis which however I find accurate. So I do not have a strong recommendation if not to make the article more concise and to the point.
> As recommended, the article has been made more concise and thus substantially shortened.

Specific comments
I'm doubtful on the expression "personal weather station". What about "private weather stations" belonging to 25 citizen weather observing networks integrating the official Meteo-France Network. Or something similar. I would also put a reference if this initiative is coordinated by the National Met Service.
> Multiple names are given in the literature to personal weather stations: amateur weather stations (Bell et al., 2013;Muller et al., 2015;Chapman et al., 2017), citizen weather stations (Meier et al., 2017;Napoly et al., 2018;Nipen et al., 2020), personal weather stations (Muller et al., 2015;Chapman and Bell, 2018;McNicholas and Mass, 2018;de Vos et al., 2020;Hintz et al., 30 2021), private (automatic) weather stations (Waller, 2020). The initiative of using these networks is not coordinated for the moment but Eumetnet is working on it with the recent creation of a working group on crowd-sourcing. In the abstract, the expression has been detailed: [...] including crowd-sourced observations of personal weather stations. In the introduction, a reference has been added: [...] from observations of standard and personal weather stations (Mandement and Caumont, 2020). 35 The literature cited in the introduction is appropriate. I think that the synoptic description, as well the numerical simulations section can be shortened perhaps limiting to the essential level of details needed for the following dynamical assessment.
> The synoptic description and the numerical simulations section have been shortened.
I'm also suggesting to shorten the Conclusions eliminating unnecessary details like time of the day references of values of single variables to have a more compact way to present results. In the conclusions bullet points (line 590), I would add the part on the evaporative cooling reported earlier at line 585 > Conclusion has been shortened and modified as asked.
it is out of the scope of this particular study which focuses on what happened in a short time period at mesoscale but could be the topic of a future work. A short sentence has been added: This episode is part of a series of HPEs that occurred in October and November 2018 over the north-western Mediterraneannd particularly affected the Balearic Islands on 9 October (Lorenzo-Lacruz et al., 2019) and Italy on 27-30 October (Davolio et al., 2020). 5 5. The case description and the numerical simulations section could be shortened with major focus on the information relevant for the case analysis and answering of the questions initially raised.
> The case description (Sect. 2) and the numerical simulations section (Sect. 3) have been shortened as recommended.
6. Please rethink the sentence in L215 "... is run to realistically simulate the case".
> is run to realistically simulate the case has been replaced by is carried out to realistically simulate the rainfall observed during 10 the case.
7. It would be interesting providing an additional simulation in which the Pyrenees were removed since its position is upstream and one goal of this study is provide and analysis of the role of topography. If this is not possible, a hypothesis would be needed.  results of these additional simulations are explained in a small paragraph: Additional simulations (not shown) in which the Corbières Massif and the Eastern slopes of the Pyrenees are successively flattened show a substantial decrease in maximum accumulated precipitation downstream of these reliefs, associated with a spread of precipitation above 200 mm over a larger area along the quasi-stationary front.
8. Please provide information of why only near-surface information is analysed and validated.

30
> A comparison to high resolution soundings at the initial state has been added, as well as an explanation of why we focus on the validation of REF near the ground: Since the initial conditions of REF are provided by the AROME analyses in which all conventional observations are assimilated, there is little deviation from these observations at the initial time. At 12:00 UTC 14 October, comparison of REF fields on all vertical levels to high resolution soundings of Nîmes, Barcelona and Palma (not shown, see Fig. 1 for the locations) reveal 35 absolute bias (respectively root mean square error) of < 0.3 K (resp. < 0.6 K) in temperature, < 0.3 g kg -1 (resp. < 0.8 g kg -1 ) in water vapour mixing ratio, < 0.5 m s -1 (resp. < 1.9 m s -1 ) in wind speed and < 7 • (resp. < 16 • ) in wind direction.
Because the stationarity of precipitation is correlated to the quasi-stationarity of a MSLP trough and a virtual potential temperature (θ v ) gradient (Sect. 2.1), and because of the availability of near-surface observations that are not assimilated in the AROME model, this section focuses on validating the REF simulation near the surface.
9. There are too many figures in the article, and multiple panels in some of them. Please, restrict the number of figures to those strictly necessary and try to combine when possible several information in one figure.   Reply to anonymous referee #3

General comments
Paper would benefit from a more diverse collection of references, which place these heavy rainfall events into wider context and better draw upon literature from all across the world.
> More diverse references have been added to the paper as recommended.

5
A schematic diagram of typical synoptic setup during heavy precipitation event might be useful (you show a map in Figure 1, but as the reader, I want more information and detail). This kind of plot would really sell the paper to the reader.
> We agree, but given the number of figures this plot have not been added. A schematic diagram of main low-level mechanisms of HPEs in the western Mediterranean region is given by the Fig. 1 of Ducrocq et al. (2016), reference which has been added 10 to the article. Also, three composite analyses detailing the synoptic setup for HPEs occurring over Languedoc-Roussillon are given by the Fig. 11 of Ricard et al. (2012).
The detail in evaluating model performance against observations is impressive. The length of this paper could be reduced. The level of detail is commendable, but there are sections that are difficult to fully take in because of the consistent level of depth.

15
-The text on the model setup in Section 3.1 is an example. The level of detail is impressive, but do you need to go into this much detail? -Along the same line, you could lose a couple of figures (and accompanying text) without reducing the quality of your writing. As an example, you could remove either Figure 7 or 8, Figure 10, and Figure 13 without impacting the quality of your analysis.

20
> The numerical simulations section (Sect. 3) has been significantly shortened. Details of the Meso-NH configuration necessary to reproduce our experiments have been moved to an appendix. Figures 8, 10, 13 have been removed as recommended and the accompanying text has been shortened.
Careful to use the same tense throughout your writing. In places you switch between present and past, which is confusing.

25
> You're right, verbs tense has been homogenised throughout the article.
When you introduce diagnostics such as water vapour mixing ratio and precipitable water in Section 2, it would be useful to know how large these values are relative to climatology (how unusual?). in environments with integrated water vapour (IWV) ranging between 30 and 45 kg m -2 i.e. PW = IWV ρ ranging between 30 and 45 mm considering a water density ρ = 1000 kg m -3 .
Splitting section 2 into two sub-sections, one focusing on the synoptic evolution and the other on the mesoscale details of the rainfall, could help to make the writing more streamlined and easier to follow. At the moment, there is too much information crammed in; the section is too long.
> The section has been split in two sub-sections, which have both been shortened.
You have confused 'westwards' and 'eastwards' in places throughout the text. Make sure that your descriptions are accurate and consistent.
> You're absolutely right, four errors have been corrected.
-L85 westward movement eastward movement; 5 -L86 westwards eastwards; -L109 westwards eastwards; -L110 west east There are quite a few occasions where you describe details of the synoptic or mesoscale evolution without referring to figures. It is fine to do this on a few occasions (adding "not shown"), but not too much, or the reader will get 10 confused.
> Parts of the article which were not related to figures have been shortened or deleted, and "not shown" has been added in sentences not referring to figures.
Some of the more technical information could be included in an Appendices section at the end of the paper, rather than in the main text (would help to streamline the text). The first paragraph in Section 4 is a good example (on 15 REF_SP).
> Technical information about Meso-NH configuration have been moved to an appendix. The paragraph in Section 4 has been substantially shortened.
Generally, there is too much description of results, and not enough interpretation and putting your results into wider context. 20 > By shortening the descriptive parts of the article, adding more diverse and recent references to put this HPE in a wider context, and responding to the various referees' comments, we hope to have improved the quality of the article and addressed this point.
When explaining to the reader why this study is important (Introduction), you should discuss the topic of interactions between tropical cyclones and the midlatitude flow, and how they can impact upon predictability. Papers by 25 Christian Grams and Florian Pantillon provide good examples of this type of work. This type of discussion will help to link the specifics of this event that you're discussing with larger-scale issues of interest in the meteorology community.
> Thanks for the references. Sentences have been added or modified in the introduction and conclusion based on the information of articles written by Christian Grams and Florian Pantillon: 30 In the introduction: Transitioning hurricanes over the North Atlantic are known to disturb the midlatitude flow close or downstream of them, causing or modifying the location and intensity of high-impact weather such as HPEs (Grams and Blumer, 2015;Pantillon et al., 2015). As hurricanes can supply large amounts of moisture and because the moisture structure in the lower troposphere was shown to play a key role in the timing and location of precipitation of previous HPEs (Lee et al., 2018), it is of interest to 35 quantify the amount of moisture supplied by Leslie to the convective system.
In the conclusion: Future work could quantify Leslie's contribution in the cyclogenesis mechanisms of this Mediterranean low and more generally the role of ex-tropical cyclones in disturbing weather of the Mediterranean basin. Accurately track and represent the life cycle of Mediterranean lows seems crucial to better forecast HPEs.direct contribution in the formation of the Mediterranean 40 low and its associated cold front CF2 but also the remote impact of Leslie's extratropical transition in a similar way as Grams and Blumer (2015) or Pantillon et al. (2015): it could help to understand how accurately these systems need to be tracked to improve HPE forecasts.

Specific comments
Avoid vague language (a few replacements are suggested) convective activity -> convection synoptic situation seem to be maintained -> maintained 5 density departures thermal signature more reflectivities are observed precipitating activities -> precipitation > Vague language has been replaced in several parts of the manuscript: 10 L3: The synoptic situation was perturbed by the former hurricane Leslie, involved has been replaced by At synoptic scale, the former hurricane Leslie was involved L14: seem to be maintained has been replaced by are maintained L44: density departures between air masses that cause the most buoyant air parcels to ascend has been replaced by buoyancy differences between air masses 15 L109: Its precipitating activity has been replaced by Precipitation along CF1 L110: thermal signature has been replaced by thermal gradient L138: more reflectivities are observed has been replaced by an extended region of reflectivity >12 dBZ is observed Some occurrences of "synoptic situation" have been kept when its meaning of "meteorological situation at synoptic-scale" appeared useful. 20 Standard and personal weather stations -please elaborate > The sentence has been reformulated: with near-surface analyses comprising standard and including crowd-sourced observations of personal weather stations.
You introduce the concept of the stationarity of the cold front (L11) without any prior discussion -seems a bit rushed 25 > It referred to L4: "At mesoscale, convective cells focused west of a cold front". However, to be consistent with literature (e.g., Santurette and Joly, 2002), a cold front that become stationary is a quasi-stationary front. Thus, L4 has been modified: At mesoscale, convective cells focused west of a quasi-stationary cold frontdecaying cold front, that became quasi-stationary, L15 to 17: 'the location of the exceptional precipitation appears to be driven primarily by the location of the quasi-stationary cold front...' seems like an obvious statement to make. Would like more insight here.

30
> This front is, before the event, a decaying cold front along which precipitation decreased and almost stopped. Then during the event, it becomes active due to the increasing advection of conditionally unstable air by the low-level jet that strengthens at the beginning of the event. More details have been added in the abstract to express the specificity of this quasi-stationary front: At mesoscale, convective cells focused west of a decaying cold front [...] Regarding lifting mechanisms, the advection of conditionally unstable air by a low-level jet towards the quasi-stationary front, confined to altitudes below 2 km, reactivated 35 convection along and downwind of the front.
Which dataset did Ricard et al. (2012) use for their climatology of heavy precipitation events over the northwest Mediterranean?
'Mediterranean Sea supplied up to 60% of the total air parcels moisture (Duffourg and Ducrocq, 2013;Duffourg et al. 2018), modulating the intensity of convective precipitation.' How did the authors calculate this value (60%)? More information on the method is needed here, even if brief.

5
> This value was calculated by water budgets by Duffourg and Ducrocq (2013). Also, Duffourg et al. (2018) estimated with backward trajectories that 50 % of the moisture supply of the 14 October 2012 HPE originated from evaporation over the Mediterranean sea. Because the method varies between both articles, the last reference has been removed. The sentence L37 has been modified: The Mediterranean Sea supplies moisture -up to 60 % of the total air parcels moisture in previous HPEs according to water budgets of Duffourg and Ducrocq (2013) (Duffourg and Ducrocq., 2013;Duffourg et al., 2018). > You're right, it has been corrected. Other sentences L301-302 and L530-531 have also been corrected.
L52: need a sentence to tell the reader that you're introducing your case study here. Currently, you just start talking and the transition from the topic in the previous paragraph is not smooth enough.

15
> The following sentence has been added: Among these mechanisms, those at the origin of the HPE of 14 and 15 October 2018, on which this article focuses, are studied.
Paragraph on the event itself reads well (L52-62); you summarise the key points nicely.
> Thank you! L68-69: 'Because of the heavy rain observed in the area, evaporative cooling processes may have played a role 20 in the stationarity of the cold front.' Not sure I follow this argument. Evaporative cooling would be expected in conjunction with the rainfall, but why are you hypothesising that this cooling could play an important role in the movement of the front? You need to make this point more confidently here, and with more detail.
> This point has been detailed: Because of the heavy convective rain observed west of this front, evaporative cooling may have additionally cooled the west side of the front. This additional cold air may have caused a dynamic feedback that contributed to 25 the stationarity of the front. Similar dynamic feedback was described by Davolio et al. (2016) over north-eastern Italy: in cases of upstream events, a cold-air layer formation preceded the convection onset and evaporation and sublimation of precipitation beneath the convective system were able to additionally cool this cold-air layer, which influenced the propagation of this cold-air mass.
L69-71: link between evaporative cooling and conditionally unstable air? Are you hypothesising that evaporative 30 cooling occurred in the mid troposphere and destabilised the profile? Need to make these details clearer.
> You're right, the sentence L69-71 was unclear so it has been reworded. Establishing a link between evaporative cooling and conditionally unstable air was not the goal of the sentence. Consequently, the goal of the article is to address the questions raised by Caumont et al. (2021): what are the roles of (i) the moisture provided by Leslie, (ii) the Pyrenees relief and (iii) the evaporative cooling in the physical processes that led to the location and intensity of the observed rainfall? 35 L84-87: where is the evidence that the interaction between both lows seems to have strongly slowed the westward movement of the mid-level cut-off low? Or, do you mean eastward movement? This would make more sense. > Indeed, there was an error, we meant eastward. Because it is only speculative, this paragraph has been modified: -L85 westward movement eastward movement; -L86 westwards eastwards.
L88-92: which figure are you referring to in this discussion? Make the connection clearer and add detail to figure if necessary.
> There was no figure, (not shown) has been added. 5 L103-104: "A potential vorticity at upper levels is observed upstream of the low." Where is the evidence for this feature? Need to relate all statements to figures, or add '(not shown)'.
> There was indeed no figure. This part has been removed in order to shorten the manuscript as recommended. > This part has been removed in order to shorten the section as recommended.
Be as precise as possible in your discussion of figures. For example, in Figure 4 you plot the 925 hPa water vapour mixing ratio and geopotential height. In the accompanying text, make sure that you refer to the diagnostics in the figure ("925 hPa wind"), rather than using more vague descriptions ("near-surface wind").

15
> You're right, here because the 925 hPa wind was not shown in Figure 4, we have added (not shown) and have replaced near-surface by between the surface and 925 hPa.
Do the authors have any theory on why precipitation within the western band is overestimated by the model, or why the model overestimates precipitation over orography?
> Within the western band and over orography of the eastern Pyrenees, Fig. B shows large departures between quantitative 20 precipitation estimate (QPE) and simulation. However, during that case, because of a radar failure mentioned in the article but also the bad coverage of eastern Pyrenees by the French radar network, QPEs, even if they include rain gauge data, likely underestimate rainfall in these two mountainous areas. Thus, even if direct comparisons of the REF simulation with several rain gauges show overestimations, it is hard to give an explanation as the amplitude of the overestimations away of rain gauges is not known with certainty.  How do Leslie's remnants contribute to the formation of the cold front CF2, behind which a Mediterranean low deepened rapidly? > In the article, we indicate that thermodynamic anomalies associated to Leslie's remnants evolve in what is identified by meteorologists as a cold front on the analyses but the physical mechanisms at stake are not studied and the answer to this question has been mentioned in the conclusion as a future work.
Mark the Albera Massif on Figure 5. Also mark the Aude and Pyrénées-Orientales departments. You refer to these features, but the reader won't necessarily know where they are (without referring back to earlier figures). The convective lines that you mark on Figure 5 are not easy to see. The eastern line in Figs. 5a to c is fine, but the other two lines are much more difficult to pick out. Is there another way that you could annotate these figure panels? > In Figure 5, the color range has been modified to better highlight the lines. We have not found a better way than arrows 5 to make clear the direction of propagation of the cells. "A", "P", CF1 and CF2 have been added to respectively indicate the locations of the Albera Massif, the Eastern slopes of the Pyrenees, CF1 and CF2. We have not added the Aude and Pyrénées-Orientales departments as the figure is quite busy and as they are shown in Fig. 1. L229: do you mean downstream? Upstream suggests the windward (not lee) side of the ridges. > Our sentence was unclear: we meant that bands were generated upstream of the ridges, propagated, and then were enhanced 10 downstream. If we quote Cosma et al. (2002): "The bands are generated upstream of these ridges and enhanced on the lee side by convergence created by deflection around the obstacle and penetration of the flow into the valleys". Our sentence has been clarified, and the reference to the upstream generation has been removed: Similar convective band generation was observed in the south-eastern flank of the Massif Central by Miniscloux et al. (2001) and Cosma et al. (2002) upstream of small-scale topography ridges, with an enhancement of these bands : rainfall bands were 15 enhanced on the lee side of small-scale topography the ridges.
L243-255: when justifying your investigation of evaporative cooling, it's not immediately clear how evaporative cooling near the surface could modify the location of CF1 or change its stationarity. Could you describe how this could occur (maybe add a reference)?
> This part has been modified, and a reference has been added in the introduction (see the answer of L68-69).

20
L248-255: in the NOCOOL simulation, did you turn evaporation off completely in the child domain, or just set the temperature tendency from evaporation to zero? Assuming the second -it wouldn't make a difference to your results, but it's best to be as precise as possible when discussing the changes you made to the model output.
> You're right, a sentence is added in the paper: In NOCOOL, negative temperature tendency from evaporation of raindrops is set to zero.

25
L277-286: you discuss the evolution of precipitation in your model simulation relative to observations, but don't show any of the plots required to do so ( Figure 6 presents accumulated rainfall but gives no indication of how the structure of the rainfall evolves during the 24-h period).
> You're right, the paragraph has been significantly shortened and (not shown) has been added. Figure 8: why have you used virtual potential temperature rather than potential temperature to diagnose the 30 position of the front? Would be interesting to know what the pure temperature difference across the front looks like (potential temperature) as well as the moisture difference (virtual potential temperature). Don't change anything in the manuscript, it's just something to think about in future.
> We used virtual potential temperature because it takes into account the density effects of water vapour. In this case, virtual potential temperature and potential temperature gradients are quite similar because relative humidity is between 90 and 100 % 35 at screen level between 23:00 and 04:00 UTC on each side of the front. To answer, the following sentence has been added: It is based on the technique of Schär and Wernli (1993) in which three Eulerian passive tracers are initialised with the initial grid point position and are advected online by the resolved and subgrid-5 scale wind; a review of existing Lagrangian trajectory tools is given by Miltenberger et al. (2013). Figure 11 ( L355): make it clearer in the figure panels and in the text that the trajectories in (b) correspond to the transect (A -> D), etc. The reader will get confused otherwise. You have done this in your discussion of Figure 13 please apply the same method to the discussion of Figure 11. > The transects have been indicated in the text as recommended.

10
Nice illustration of lifting along the cold front in Fig. 11c. Overlay the position of the cold fronts (CF1 and CF2) in Figures 11 and 12, and on Figures 14 and 15.
> The position of the cold fronts has been overlaid.
L387-389: the claim that the trough over the Aude department may locally increase wind shear inland, based on Fig. 10a, is not based up strongly by the evidence.

15
> You're right, this claim has been removed.
Do you need Figure 13? The results are similar to those in Figure 11, and you could just include a sentence to tell the reader that trajectories ending at 0400 UTC 15th October are qualitatively similar to those ending at 0000 UTC 15th October.
> Figure 13 has been deleted, and a sentence has been added: Similar backward trajectories are simulated at the end of this first 20 part, e.g. at 04:00 UTC 15 October (not shown).
L417-418: be careful when describing the motion of cold air 'as a density current'; you don't have enough evidence to make that specific claim. Instead you say something like 'in a similar manner to a density current' to be less specific.
L418-420: not sure what you mean by equilibrium in this discussion. You need to be more specific here and 25 describe the important physical processes. Also, language like '...tends to show that...' is too vague and should be avoided.
> "As a density current" has been replaced by "in a similar manner to a density current". The expression "tends to show that" has been removed. Here equilibrium referred to the fact that the propagation of the cold air located west of the quasi-stationary front towards east appears countered by the low-level jet blowing perpendicularly to it. 30 This part has been rewritten: When wind speed decreased along CF1, the cold air west of CF1 started flowing eastwards, towards the Mediterranean Sea, as a density current, spreading out circularly over the sea. This cold air rapidly flowing as soon as the LLJ stops tends to show that an equilibrium maintaining CF1 quasi-stationary was reached between CF1 and the LLJ. Then, the equilibrium progressively broke from south to north by the advance of CF2. When wind speed abruptly decreases along CF1, CF1 stationarity breaks and the cold air west of CF1 immediately starts flowing eastwards in a similar manner to a 35 density current, and later spreads out circularly over the Mediterranean Sea (not shown). It indicates that, during the HPE, the propagation of the cold air located west of CF1 is countered by the LLJ blowing perpendicularly to it, also in a similar manner as the propagation of a cold pool can be countered by the environmental wind (Miglietta and Rotunno, 2014). L465: replace "probably" with "likely" > The modification has been made.
L486-490: I don't follow the argument. You say that backward trajectory analysis demonstrates that an increasing number of low-level moist air parcels that have not crossed the Albera Massif are found inside convective cells on the lee side of the mountain (i.e. originating on the lee side of the mountain, rather than further S-SE?). You then follow on from that point and say the supply of conditionally unstable air parcels (from the S-SE) along the line is 5 continuous. Have I misunderstood, or do these two sentences contradict each other?
> The sentences were unclear. Parcels that do not cross the Albera Massif (which is at the S-SE) are carried by the south-eastern LLJ and originate from the Mediterranean Sea, not from the lee side of the mountain. This part has been rewritten: Because of the slight directional vertical wind shear simulated in the lower part of the troposphere (see wind direction in Figs. 17a,c), the LLJ continuously supplies conditionally unstable air parcels to the convective cells formed from their 10 southeastern flank while these cells are advected by the south-southeastern mid-level wind. Backward trajectories starting from their updrafts (not shown) show that once convective cells are on the lee side of the mountain, as they are advected north-northwestwards, an increasing number of low-level moist air parcels that have not crossed the Albera Massif are found inside the cells. Thus, the supply of conditionally unstable air parcels along the line is continuous and possibly explains the maintenance of the convective cells long after they are formed. 15 once convective cells are on the lee side of the mountain, as they are advected by the south-south-eastern mid-level wind, the south-eastern LLJ supplies conditionally unstable air parcels that do not cross the Albera Massif to the cells from their southeastern flank. Backward trajectories starting from their updraughts (not shown) indicate that the number of low-level moist air parcels that do not cross the Albera Massif found inside the cells increases as they are advected. This supply mechanism possibly explains the maintenance of the convective cells long after they are formed.
I like the final paragraph of Section 5, in which you summarise the role of the Albera Massif. Follow-up question: is the topography of the Albera Massif only likely to play an important role when the wind direction is exactly as in this event? If the wind direction was slightly different, would you expect a different region of orography to play a more important role? > Thank you! 25 Considering a Mediterranean air mass having a conditional instability similar to this case, if the wind direction was slightly different and we are talking about the role on rainfall, we expect that the Albera Massif could still play an important role: Fig. 9 of Ducrocq et al. (2008) shows that the Albera Massif was a source of lifting in the 1999 Aude HPE (see the parcel n°1) while the 500 hPa wind direction (Fig. 18 of Nuissier et al., 2008) was slightly further south than in 2018 (with also differences in wind speed). Depending on the amplitude of wind direction changes, we indeed expect (i) a different region of orography to 30 play a more important role (ii) different interactions with other reliefs: e.g. if the mid-level wind had an easterly direction, convective cells would be advected from the Albera Massif towards the highest reliefs of the Pyrenees (iii) different wind shear between the ground and mid-levels, wind shear that may not be as favourable as in this 2018 case to the maintenance of convective cells downstream of the Albera Massif. > You're right, the orientation of the cross-section has been reversed in Fig. 19 and corresponding changes in the text have been made.
L530: replace "This probably" with "This difference possibly" 40 > The modification has been made. Figure 20: change the colour scale so that you can more easily distinguish between values around 60 to 80% and those nearer 90% and above. > The colour scale has been changed.
The summary paragraph from L588-595 is well-written and nicely set out. However, I don't think the statement that Leslie's remnants are involved in the formation of the cold front CF2 is backed up strongly enough by your analysis. Modify the text to include more evidence supporting this statement, or remove it from the manuscript.
> Thanks! You're right, the fact that Leslie's remnants are involved in the formation of the cold front CF2 has been moved to 5 future work.

Technical corrections
Model domain notation. Use parent + child rather than father / mother. > You're right, father L226 has been replaced by parent.
L88: replace with "a small jet branch circumvented the cut-off low to the south" 10 > It has been replaced.
L100: "participated"? Not sure of its meaning here.
> It has been replaced by "contributed".
L138: "more reflectivities are observed". Change to something like "an extended region of reflectivity > 12 dBZ is observed..." 15 > Done, more reflectivities are observed has been replaced by an extended region of reflectivity >12 dBZ is observed Figure 11: equivalent potential temperature above the surface is normally shown in K, not in°C.
> The modification has been made. L452: replace "couples" with "couplets" > The modification has been made. Brackets need editing in some of your time interval labels in Table 2 > Hyphens have been replaced by commas to match standard interval notations. The first seven intervals are voluntarily rightopen as well as the last one is closed. 25 L484-485: "...slightly directional vertical wind shear..." do you mean that there is only a slight change in wind direction with height? Replace "slightly" with "slight" if so.
> You're right, it has been corrected.