Dear Authors, 

please refer to the attached PDF for my review of your manuscript.

Review of „Atmospheric Deserts: Detection and Consequences“ by Fix et al. submitted to Weather and Climate Dynamics

Summary and general assessment

Fix et al. introduce a new concept of atmospheric deserts (AD), a phenomena that advects hot and dry air from desert or semi-arid areas towards the midlatitudes. ADs are a generalization of the more known elevated mixed layers (EML). The authors analyse the impact of an atmospheric desert event from mid-June 2022 to Europe, which was partly effected by high temperatures and thunderstorms. For that, they computed forward trajectories from North Africa and clustered the trajectories with a k-means clustering approach. These clusters were analysed according to the evolution of different meteorological variables traced along the trajectories (e.g., height, location, specific humidity etc.). Overall, they found four different clusters, one resembled an EML, the other one a warm conveyor belt and the third one slightly descend. The last one was rather ambiguous. In addition, they show how the AD influenced the formation of heat wave and thunderstorms in parts of Europe.

Overall, the study is well-written, nicely analysed and figures are of high quality. However, I have some major concerns that should be addressed prior to publication.

Major comments:

1. In my view, the title of the study promises a little bit too much. When I first read the title, I imagined that the authors were introducing a new concept and showing how this new concept affects weather in the extratropics in general. In reality, the authors „only“ performed a case study, which is not negative in itself, but not enough for this title. This brings me to my main critique of this study. Although the case study is certainly very nicely analysed and interesting, the main findings of the study are not too significant. The authors argue above all that the atmospheric desert influences two phenomena in particular, namely thunderstorms and heat waves. For the thunderstorms, they argue that the dry air influenced the formation of the thunderstorm on the edge of the heat wave, which was affected by a cold front. However, this process appears to be very case-sensitive, so that it cannot be generalised that atmospheric deserts have an influence on the formation of thunderstorms. In addition, the authors argue in section 3.4.1 that atmospheric deserts co-occur with heat wave, but they not really show whether the atmospheric desert has an influence on the heat wave in this case study. In L264-270, backward trajectories started from the heat wave area showed that they mainly originated from the Atlantic and only a small part from North Africa, which is the origin area of the atmospheric desert. So to what extent is then the phenomena „atmospheric desert“ important for the heat wave itself? Since they only analysed this case study, the co-occurrence of heat waves and atmospheric deserts may just be a coincidence. This could be better quantified by a climatological analysis.

2. Climatological analysis: I know that this needs very much computing time. But I would suggest that the study would extremely benefit from that. For a climatological analysis a calculation of trajectories on a much coarser resolution would be sufficient (I think this would be feasible in terms of computation time). The authors already stated in L368 that they plan to do a climatological analysis in the future.

3. Maybe you can discuss the atmospheric desert with the opposite „atmospheric river“? Both can lead to extreme events, one to extreme precipitation and the other one to extreme temperatures? Maybe you can elaborate on this (also this discussion would need a climatological analysis)

4. I find the definition of an atmospheric desert in L60-61 a bit too weak. Is only one trajectory really enough to significantly affect a grid box? Maybe you could perform a sensitivity analysis and elaborate on that.

Minor comments: 

L20-21: Is your postulation corroborated by your findings?

L21: Can atmospheric deserts also transport dust towards the mid-latitudes?

L30: which properties?

L34-35: The effects of EMLs and ADs are similar. Is there a process in your case study which is new in ADs and not yet found in EMLs?

L36-39: For the special case of EMLs … → add something like „and thereby contributing to potential instability when lifting mechanism is available“

L52-55: this was already mentioned very similarly in the introduction

Section 2.1: How long are the forward trajectories and at which pressure/model levels have you initiated them?

Section 3.1: This is better suited in the data and methods section 2.

L114: Is there a meteorological reason why you use exactly this region as source region?

L149: where do the other 80% of the trajectories are going to?

L162: I don’t see a warm sector because the low is already occluded.

L164: Which heights are encompassed in the column?

L168: At which height is the majority of ADs?

L186: does not explain this warming → insert „diabatic“ warming

L200: Does it make sense to regard Cluster C2 still as an atmospheric desert? Because it leads to precipitation and is not dry anymore?

L203-204: evaporative cooling as precipitation falling … → is this the precipitation from Cluster C2?

L207-209: You lost me at this point. C2 is ascending, heated diabatically due to latent heat release, therefore increase in cwc-variables and decrease of q due to precipitation. But why do cwc-variables of C3 act very similar to C2, although C3 is descending and cooled diabatically?

L245: above the cold air mass at the surface → do you mean at around 12°W?

L247: repetitive to L241 and following lines

Section 3.4.1 → this section is a bit disappointing because it does not provide new insights into the formation of heat waves. How long was this heat waves? Typically, a heat wave should at least last three days to be defined as such.

L261-262: perhaps you should look for another period, in which high temperatures persisted at least 3 consecutive days in order to fulfil the criterion of a heat wave? Then you would maybe see that AD form a lid.

L264: the explanation for high surface temperatures of advection and subsidence heating is not really new in the literature … (please review papers on heat wave formation, in particular from a lagrangian point of view)

L269-L270: Hence, the analysis of the AD event … → yes, this is correct but it still can be a coincidence, especially when you don’t compare this with climatology (either your own or from existing literature on this subject)

Section 3.4.2: I don’t get the message of this section. Is penetration of air into the boundary layer not just a normal process when the boundary layer grows during the day? What is now the special with ADs?

L284: … while thunderstorm tend to erupt violently along its edges → is this always the case or only under certain circumstances, e.g. a cold front at its edge?

L289: … the warm AD air still suppresses thunderstorm formation in most parts. → okay, but I assume that the major reason for suppressing the convection is the large-scale subsidence in the anticyclone

L310: … while the equivalent potential temperature decreases with height → very hard to see in the figure

Conclusion: a critical assessment of the approach used in this study is lacking

L369-371: please review e.g. the following paper on the formation of heat waves: https://www.nature.com/articles/ngeo2141#citeas, https://rmets.onlinelibrary.wiley.com/doi/full/10.1002/qj.3599, https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.2339 

Figures:

Figure 1: Display of the situation during the second half … → what is the first half, since you analyse 15-19 June 2022?

Figure 1: 800 hPa fronts → how were fronts identified?

Figure 2: are the initial values at (b), (c), (d) relevant? Because you don’t inserted the initial values at (a), (e) and (f).

Figure 3: 00 UTC instead of 06 UTC

Figure 3: What do the red lines show?

Figure 4: Why 00 UTC instead of 12 UTC? 12 UTC perhaps better due to the BLH topography?

Figure 5 doesn’t provide many new insights → perhaps a cross-section in the area from Fig. 6 would be better to further the insights on the penetration of AD air into the local boundary layer?

Figure 6: Back-trajectories started at which level?