In the study “Influence of air mass origin on microphysical properties of low-level clouds in a subarctic environment” by Doulgeris et al. microphysical cloud properties measured during eight Pallas Cloud Experiments in the Finnish subarctic region are analyzed with respect to their air mass origin based on the Lagrangian particle dispersion model FLEXPART.

The scientific approach is valid and the manuscript is structured in a clear and concise way. 

However, two main deficits regarding the scientific relevance and thus the scientific quality of the study as described in the general comments would require major revisions.

General comments

1. Scientific relevance

The study is based on a large time series of measurements campaigns that have been conducted in a subarctic mostly pristine region adequate for the analysis of aerosol cloud-interactions (ACI). A clear statement is missing on how the presented results may advance the current state of the art. The dependence of cloud microphysics on the air mass origin (Twomey effect in continental air masses vs. marine air masses) and that cloud droplets are prone to grow in warmer air is known already from other studies (cited in the manuscript l.346). Also the introduction is not clearly leading to a research hypothesis or research question. Modifications in the Introduction and discussion of results as well as in the abstract and conclusions are required to specify the scientific relevance of the study in the context of existing literature. The identification and further interpretation of results that add new findings to the existing body of knowledge would be helpful for the ACI community and further studies. 

2. Scientific approach 

Cloud properties are analyzed according to their air mass origin in 5 predefined source regions. Cloud properties strongly depend on the air mass characteristics including humidity, wind speed, temperature etc. at different altitudes. Including air mass characteristics (e.g. from ERA5 reanalysis) in the analysis to understand differences in Nc, MVD, ED as started in Fig. 8 would make results more interpretable and scientifically relevant.

Also the approach of using predefined source regions is questioned as this classification may result in similar/mixed air mass characteristics as shown for the Eastern/Southern and Arctic/Western air masses in Fig. 8. More intuitive would be an automatic classification (grouping) based on the air-mass origins or pathways. 

Specific comments

l. 106: Do you have information on the cloud type, is this mainly fog or low stratus? This may imply different processes.

l. 123: Latitude and longitude is missing in Fig. 1.

l. 201: Delete "model" as this can be mixed-up with numerical models.

l. 207: it is not specified if the PES belongs to an aerosol type or Nc or which emission inventory is used to calculate the PES. If solely air mass trajectories are calculated backwards what is the PES referred to? Please provide more details on the FLEXPART model settings and assumptions here.

l. 244: Subtitle 3.1 should be bold as 3.2 and 3.3.

l. 245: The main message of the figure is not mentioned and should include something like: It shows the seasonal range of temperatures from on average XX°C in September to -XX°C in November and its interannual variability.  

Fig. 5 (also Fig. 7): Is there a reason to present each year separately? If not I suggest in accordance with the main message of the figure to present only one average line together with the standard deviation and include data gaps in the data section. This also applies to Fig. 7 and would increase clarity of the figures as 4 panels can even be summarized in one panel (4 lines - 4 regions). Data gaps and instrument specifications can be moved to the data section.

l. 271: anthropogenic aerosols: Is this an assumption, provide a reference?

l. 275: Fig. 6 a) What is the meaning of the cyan color? If not necessary please remove it. If it is representing a range please indicate it in the legend.

l. 275: Fig. 6 b) Symbols (stars and circles) representing different Nc measurements are difficult to distinguish. Would recommend either summarizing campaigns sorted by PES and CAS/FSSP (4 symbols per air mass) or summarizing it even further only by PES. If this is no option, increasing maker size and distance between campaigns would improve clarity.

l. 284 CAPS --> CAS? (as in the legend of 6b), please check usage throughout the manuscript.

l. 395-396 Fixed vertical position, but different layers? Something is missing: “…cases that we sampled WITH different layers”.

l. 409 Fig 10 a and b. I cannot see the difference between CAS and FSSP in the plot. If the difference is not important for conveying the message that MVD/ED is not dependent on the position of the probe I would skip the legend entry. This figure could be improved by using a scatter density plot (2-D histogram) and regression line. 

Further, If there is no dependence between MVD/ED and position of the probe, is this something still relevant for the main message of the paper and would it require a figure plus subsection? If the answer is no I recommend skipping it or putting it in the supplementary.

l. 428: to be representative or considered as representative

l. 429: Why are clouds more frequent when air masses originate from Southern and Eastern regions?

l. 440: What kind of measurements are needed?

In their study Doulgeris et al combine a long-term dataset of in-situ observed cloud microphysical properties at a sub-arctic location with simulations of air mass transport. While the general methodology, uniqueness of the dataset and presentation are reasonable for publication, I do share the concerns of the first referee regarding scientific relevance and quality. In my opinion the manuscript should be reconsidered after major revisions.

General comments

• The authors should point out more clearly where their work extends the current level of scientific knowledge. As the authors describe in the literature overview, the Twomey effect is well confirmed and no significant additions are provided in the manuscript. It should be considered to change the manuscript type and focus to a measurement report instead of a research article.
• It does not become clear how including the cloud base height information in Sec. 3.4 supports the manuscript. A distance of 4km of the ceilometer for cloud-base height retrieval seems quite far away. Also, it does not become clear if and how only stratiform cases are selected. The resulting Fig 10 looks more like a ‘point cloud’ without the chance to identify any physical relationship.
• The airmass source analysis raises some questions as well. The regions seem rather inconsistent. E.g., why is the Kola peninsula ‘Eastern’ and not ‘Arctic’ or why is Scotland an Ireland ‘Western’ while England is ‘Southern’. The simulation duration of 4 days is quite short. Was there any sensitivity analysis performed with 7 or 10 day simulations? How were contributions from outside the area of Fig. 3 treated?

Specific comments

• L31: The statement on larger droplets in warm clouds in the current form is not supported by the presented data. Fig 8 b, d shows a decrease of particle size for the ‘Arctic’ subsample in the FSSP data.
• L44: The issue of varying meteorological conditions is raised, but throughout the manuscript is does not become clear how different temperature and humidity within an airmass origin category are treated (or if they are uniform enough to be disregarded).
• L49-51: Consider rephrasing this sentence, it is hard to grasp what is reason of the limited knowledge and what is the consequence.
• L79-83: The sudden appearance of ice particle sizes confuses the reader. As the manuscript focuses only on cloud droplets, consider removing it.
• Fig 1: The information content of this map has to be increased. Include the elevation, ideally as shading or contour line, as you later argue based on the orography. A legend and lat-lon grid are lacking. Does the darker green color indicate forest? The labels are too small.
• L147 and 180: Please provide a histogram of wind direction/wind strength. Are certain airmass origin categories subsampled due to filtering periods when FSSP and CAPS did not look into the same direction.
• L152-157: At which of the sites were the meteorological observations conducted? Sammaltunturi? Given the amount of detail on hardware in this paragraph, it would be nice to have that the reader would not be forced to consult Doulgeris 2020/2022 for this piece of information.
• Fig 6 b: Please distinguish the years. Make clearer what data is from CAS and what from FSSP
• L304: Please provide the duration of the >80% periods also as fraction of the total in-cloud duration
• Fig 7: Without indicating the variability, the yearly averages are of limited use for the reader. How may hours of observations are available for each year and each cluster? Also, given the typical duration of the campaigns, shown is an autumn average.
• L350-352: The statement on shorter lifetime of warm Arctic clouds not well supported by the data presented. Please either extend the argumentation or omit that senstence.
• Fig 8: Please include the no of samples per bin, instead the vague statement in L354.
• L367: How did you account for different temperatures in different air masses for this conclusion?
• Fig 9 a: similarly to Fig 6b, please indicate the single years and make CAPS and FSSP more distinguishable
• L431-434: The conclusions on number concentration and diameter with respect to the height are not backed by the analysis presented in Sec 3.4 at all. Either remove this aspect or expand on the reasoning, including descriptive figures.