General Comments:

The work was prepared by a team of good specialists in the field of climatology and meteorology, statistical methods and atmospheric pollution. Measurements of the vertical structure of PM10 concentration and temperature in an atmosphere layer >200m in Krakow were taken during two winter seasons thanks to the use of sightseeing balloon. The authors also used all available meteorological and air pollution data from all measuring stations in Krakow and from meteorological models to explain the vertical structure of PM10 concentration. The results were discussed with other similar measurements around the world. The authors have demonstrated knowledge of the latest literature on this subject. This is evidenced by the citation of the latest works, even from 2020 and 2021, for example lines: 37, 44, 50, 60-62 71, 80-81, 84).

The most important result of the research is the finding of a significant influence of wind shear on the vertical distribution of PM10 concentration and the determination of three vertical zones of air pollution hazards. These results are representative of other parts of Europe with poor aero sanitary conditions.

In terms of content, I fully accept the text of the publication. If possible, I suggest making a few technical corrections, which are listed below.

Technical corrections:

Line 108: (PM10 emissions… 2020). In the reference is the year 2021, see line825.

Line 1003: change (Raport… 2020) to (Roczna… 2020). See line 834.

Line 122: (Kraków, 2019)? Is this citation?

Table 1 and 2: add “°” for Lat and Lon:  °N,  °E

Lines 398 and 399: 2-4 m·s^-1   is to be written together on one line.

Lines 552, 556, 566-567: citations must be in chronological order.

Line 568: Zangl  -  in reference list, line 892 is Zängl.

Lines 739-740: This position is not cited in the text (Drechsel and Mayr, 2008).

Lines 834-837: Roczna…       move to line 843.

• The authors state that they use a "subjective method of fitting the linear curve to each vertical profile" based on "R squared coefficient, the angle of the straight line and residual values classification". They further clarify that objective classification methodologies could not be used "due to differences in flight heights and the PM10 measurement altitudes". With all of the classification approaches available, I find this assertion difficult to accept, especially if the classification is being performed on quantitative metrics like the ones listed. Without a clearer, more objective approach I find it hard to consider this procedure sufficiently robust and reproducible.
• Following this subjective sorting, some comparison of local conditions for each profile is performed and described. However, the figures and metrics presented to support the final conclusions are not sufficiently clear and convincing, to my eye. Figure 6, in particular, would seem in the text to be a key figure in supporting final conclusions, but after several readings I still cannot figure out what it is supposed to be showing. Figure 7 highlights a particular day chosen as an example, but the information is scattered, unclear, and leaves many questions unanswered regarding the representativeness of the chosen day and the reasons for not somehow showing the characteristics of ALL observed days, rather than one or two. Throughout the final pages of text, I kept looking for the conclusive evidence that succinctly and unambiguously showed the connection (with uncertainties and other statistical metrics) between the patterns being studied, but I was not able to find this. If such evidence is in there, it should be made more clear. If not, it must be added.
• Finally, I would like to suggest that the order of operations being performed here (first manually sort vertical profiles, then try to analyze atmospheric conditions) is making the authors' (and reader's) task more difficult than it ought to be. If the main thesis is that strong wind shear drives key differences in surface and vertical profile PM10 patterns, why not first sort days based on this unambiguous metric (strength of wind shear) and then compare PM10 levels, boundary layer depth, vertical profiles, or any other dependent property based on that? Either way, the connection between pollution results and their potential drivers needs to be explored in a clearer, more defensible manner.

• References: Several references are in a notation unfamiliar to me, for example (Air… 2020). Is this a placeholder? I think these should be standardized and revised, unless I am missing something.
• Figure 1: The maps are useful, but their distance from corresponding tables makes it hard to interpret their labels. This may be better resolved at a later stage of typesetting, but some effort to line these up in some way would be helpful.
• Flight frequency details: What were the patterns of balloon flights over the course of each day? How much of the vertical profile patterns observed can be attributed to regular diurnal cycles? If the "frequency of flights depended on meteorological conditions" is it possible that some sampling bias has been introduced based on these limitations, and if so how is it being addressed?
• Figures 5 and 7: These panel plots appear to contain a great deal of information, but they are overpacked and underexplained. It's unclear in their current form what they are trying to show, on the whole, and some presentation details signifcantly impede clarity. For example, the humidity color bars run from 0 to 100%, but the data appear to be uniformly in one small sliver of that space, making the entire plot uniformly green. Both of these figures need rethinking.
• Figure 6: As mentioned, this figure is very confusing and it's unclear what the takeaway message is supposed to be. Furthermore, panel (b) appears to never be described in the caption.