Comment on acp-2021-610

This study addresses how wintertime warm and moist air intrusions into the Arctic affect the surface/boundary-layer energy budget. The analysis is very thorough and highlights intriguing differences between fully ice-covered sectors and sectors which instead have open waters to the south. The analysis and conclusions provide an interesting addition to the literature and I believe that they should eventually be published. Nonetheless, I would recommend a number of revisions before the paper is accepted. I have some concerns on the methodology and on how this may affect the results, and I find the paper to be overall poorly and with a large number of careless which no proof-reading of the text been carried out prior to submission.

2. The paper is poorly written and with a number of careless errors/typos, which suggest that no proof-reading has been performed prior to submission. The text needs a thorough review before it may be published. Below are a few examples, but this is by no means a comprehensive list: Several "the" missing from the abstract.
l. 331 "is warmed" --> "are warmed" ll. 331-332 "before advected" --> "before being advected"  Was there a reason to exclude the ~20 degrees wide ocean sector immediately to the east of Greenland? Sect. 2.1 Doesn't the use of coarse-grained ERA5 data (at 1/3 of the spatial resolution available from ECMWF) significantly reduce the accuracy of the airmass trajectory calculations? l. 73-80 I agree with the authors that we need to make use of the data we have, even though it has known limitations. However, the discussion on this point should be qualified with references to the relevant literature. There are several different reanalysis products available for the Arctic region, and several studies have provided intercomparisons of how they perform. To clarify, I am not suggesting the use of additional datasets, but rather a more robust argument grounded in the existing literature as to: (i) why ERA5 is a sensible choice and how it compares to other data options that the authors could potentially have used; and (ii) what sort of uncertainty we may expect from the use of reanalysis data (are we talking about an uncertainty of the same order of magnitude as the actual values?). ll. 88-90 This is a key part of the methods but the details are hard to follow. Please revise the phrasing, specify what the 95th percentile is calculated for (all sensitive regions, each region in turn or other? One should not need to second-guess it from Fig.  2), specify that WaMAIs are >>continuous<< periods when f_bar_w exceeds 0, explain that a single WaMAI always contains at least one EMI, but may contain several, show the 95 th percentile as a horizontal line in Figs. 2c, d etc. Also, the highlighted WaMAI in Fig. 2c does not seem to match the definition, as the portion of the line immediately preceding the marked WaMAI is still above zero, yet coloured in black. l. 93 Do the authors mean that they select the single point along the latitude circle within each sector with the highest T850 on the day when a WaMAI is selected and initialise the trajectories only from there? This needs to be rephrased to clarify what the procedure actually is. l. 95 Isn't this a very restrictive criterion? Airmass trajectories penetrating the Arctic basin may have a strong zonal component and in some cases even have short sectors of their tracks with a southward component. Could the authors quantify how many events they exclude which have a predominantly northward component but track southwards for 1 or 2 timesteps during the 4 days considered? l. 97 ERA5 data is typically used on pressure levels (with some variables also being available on sigma levels), yet here the authors seem to imply that they initialise their trajectories at fixed geometric heights. Does this mean that interpolation is not only used to retrieve the vertical profiles of the trajectories but also to determine the starting points of the trajectories? If so, does this not significantly degrade the performance of the tracking algorithm? l. 106 Why use a 0.5 interpolation from data at 0.75 degrees when higher-resolution ERA5 data is available directly from ECMWF? Sect. 3.3 I am not sure that the description of these four WaMAI boundary-layer energybudget classes allows for full reproducibility of the results. Could the authors add a short section in the Metods or an Appendix where they describe in detail how each class is defined, how border-line cases at the cross-over between different classes are treated etc.?
4. The authors mention on l. 84 that the Kara sea, like the Barents sea, also experiences some wintertime sea-ice variability. However, this aspect is never picked up again in the analysis, and the Kara Sea seems to be missing altogether from Table 2 (at least  according to the table caption). Could the authors include a discussion of the Kara sea in their analysis and comment on whether it is a special case (intermediate between the Barents Sea and the other sectors) or follows one of the two patterns already discussed (open ocean vs. land-locked ice)? 5. Sect. 3.2 There are some very interesting results in this section, but I struggled to read it. Right now it reads more as a point-by-point description of the figures rather than as a description which highligths the key results from the figures. I would encourage the authors to try to distil the relevant information provided by the figures, and communicate it more effectively in the text. This also applies to a few other passages in Sect. 3.3, but is most evident here.

Minor Comments
7. l. 227-228 Would composites over all events of a given category in a given sector show some features similar to those shown for these individual events, or are the boundarylayer characteristics so variable as to average out in the composite? This is a point that may be of interest to readers. I would be grateful if the authors could show a composite figure in their reply (even if they decide not to include it in the manuscript) to get a feel for what the variability of these events actually is. 8. Fig. 2b, d The blue dot is very hard to view. Perhaps use a vertical line to mark the 95th percentile? 9. Fig. 6 To ease interpretation, you may want to specify that the range of the colourbars is different from Fig. 5. 10. Fig. 9 "Note that this is not necessarily the distance travelled, since WaMAIs need to travel due northward." I though the reason for this not being the distance travelled is that WaMAIs need to track nortwhards but can have also a zonal component to their path? 11. Figs. 9, 10 and 16 Mixing red and green is not a good idea as colour-impaired readers will not be able to distinguish the different curves.
12. Fig. 10 Specify in the caption that panels (a), (b) refer to the Barents Sea.