Lata et al. present in their manuscript "Measurement report: Role of Organic Coating and Chemical Composition on Ice Nucleation Potential of Atmospheric Particles in European Arctic" results of STXM/NEXAFS, SEM-EDX, and INP measurements of samples collected on Svalbard during the winter season of 2020/2021. They found indications that the properties of the organic coating influence the ice nucleation activity of their samples. While this study only features five samples, which of course limits the statistical significance of the results, one must recognize the labor that went into the analyis of these samples. STXM/NEXAFS measurements of so many individual particles are no small feat. Furthermore, STXM/NEXAFS is rarely used in studies of ambient aerosol samples in conjunction with ice-nucleating particle measurements, which makes these results valuable. Considering the limited scope of the study due to the small sample size and the value of the results, I think the publication as a measurement report is appropriate after making some revisions.

Please see my comments and suggestions below. Some comments are marked with "***"; these are the ones I deem more important and that need to be addressed. There is also a list of technical comments (typos, etc.) at the end.   

Comments and suggestions

L51

Immersion freezing is mentioned for the first time without further explanation. I suggest adding one or two sentences to the previous paragraph to explain that there are different freezing modes and describe immersion freezing, stating that it is the most relevant mode for MPCs.

L62

Since the em dash is now often seen as a sign of AI usage, I recommend removing it and rephrasing the sentence if AI was not actually used. Using AI is generally fine, but according to the ACP submission guidelines, if AI was used to prepare the manuscript, its use must be described in the Methods section or the Acknowledgments.

L91 - L94

Also specify the distance between the Gruvebadet and the AWIPEV station. 

L95

Introduce the disdrometer properly (see the 2nd technical comment for details)

L96

What is "Pluvio total"? Does it simply mean that the precipitation volume is derived from the OTT Pluvio² L regardless of the phase determined by the OTT Parsivel²?

L98 - L99

Do I understand correctly that the sampling for the INP analysis is continuous, but co-located samples for the single particle analysis were collected only on five occasions? Or were co-located samples always taken, but only five were analyzed? In either case, the text should mention the criteria used to select these five occasions.

*** Particle sizes

I am a bit surprised by the large number of supermicron particles you found on the TEM grids (Fig. 3). The TEM grids should mainly contain particles with an aerodynamic diameter between 250 and 500 nm, because you only used the last impactor stage with d₅₀ = 250 nm and the previous stage has d₅₀ = 500 nm. Of course, AED and aerodynamic diameter are not identical, and it is known that the AED is on average larger than the geometric diameter and also that the aerodynamic diameter is typically larger than the geometric diameter (Huang et al., 2021). But knowing this and considering typical shape factors and density of SSA, I find it hard to imagine a scenario where many particles with a significantly larger AED than aerodynamic diameter occur.

Do you have an explanation for the large number of supermicron particles? Was e.g., the impactor operated with the correct flow rate?

L108

I assume these filters are made of polycarbonate 

L110

The citation in parentheses is not needed if the same papers are cited within the sentence

L111 - L112

Does that mean that some of the samples are measured in regular intervals in Ny-Ålesund, but most samples are analyzed at the authors' home institutes? Why are not all samples measured on-site? Is it the lack of manpower or is on-site only a rudimentary freezing array setup to ensure sample integrity?  Even though you cite Li et al. (2023), an additional sentence describing these on-site verification measurements would not hurt.

Section 2.2.

Later, you present results from the FLEXPART back trajectory analysis. Why did you choose to use HYSPLIT to characterize the air masses and identify potential source regions, given that one could argue FLEXPART is better suited for studying aerosol particles? 

L116

While it is known that back trajectories become significantly less accurate the farther back in time they go, and every trajectory going back more than five days should be viewed critically, 48 hours is rather short and might not reveal the full picture.

I suggest reproducing Fig. 2 with five-day back trajectories for the supplement.

Additionally, I suggest reproducing Fig. 2 with color-coded heights so readers can assess which regions sources close to the surface might be important in.

L135

How does the computer identify the particles? Is it by a brightness threshold of the BSE image? If so, do you think biological particles (i.e., C-rich particles), which appear less bright in the BSE image, could be undercounted? Is there also a lower size limit for what the computer recognizes as a particle? Please mention these things in the manuscript.

L139

Please provide more information about the EDX measurement, such as how long the spectrum was collected and whether it was measured at only one central spot or if the beam scanned across the detected particle.

***L151

You write that STXM/NEXAFS was performed on a subset of particles that were already analysed with CCSEM/EDX. If you measure a particle twice, isn't it possible that a particle received beam damage, which might alter the STXM/NEXAFS results?

L165

I suppose that also at the ALS the STXM/NEXAFS measurements do not provide the convenience of computer controlled measurements such as those of the SEM. Therefore, it is valid to mention (somewhere in this paragraph) that the nearly 2000 particles were measured manually to emphasize the effort involved in these measurements.

L170

What is the upper detection limit of WT-CRAFT?

L177

Are the "< 3%" calculated in the f_ice or N_INP space? I would also recommend showing the field blank and, if available, pure MilliQ water measurements in the supplement.

L181

I would not talk about "airborne samples" as this as this gives the impression of samples collected on aircrafts, balloons, drones etc.

Figure 1

A cyclic colormap would be better suited for color-coding the wind direction.

Please also consider making as "zoomed in" version of this figure for the supplement, showing only the sampling periods

Figure 2

It would be great if the sea ice concentration or at least the sea ice edge is shown in these maps.

L212

Where does the information about turbulence come from? If the statement is based on measurements, they should be also be part of the manuscript.

L217

Can you elaborate on the possible enhanced atmospheric processing for this event? Is the cause for that primarily the farther transport (hence more time for the aging to take place), the presence of reactive species  that do not exist at higher latitudes, or the availability of solar radiation?

Figure 3 and 4

Instead of a separate figure for CCSEM and STXM results, I think it would be better to have CCSEM and STXM sample wise side by side, eliminating the need to scroll through the document when a reader wants to asses the chemical composition of a specific sample.

Section 3.3 first paragraph

It is of course good to provide context and show what others have found, but I am missing the connection to your specific samples. Since you only have five samples from four distinct meteorological scenarios, you can directly compare your findings during, for example, the warm air intrusion with studies that also measured during a warm air intrusion.

L291

This statement should be removed. The 0.2 INP/L is simply the upper detection limit of the achievable with the used droplet freezing array. During the same expedition, also a CFDC (operated in immersion mode) was deployed and at -28°C, a max. INP concentrations of around 100 L-1 were measured, which can be seen in the cited overview publication by Welti et al. (2020), but also in more detail in the expedition related publication by Hartmann et al. (2021) (which you also already cite). If you look into the latter study more carefully, you will also notice, that the name of the campaign was PASCAL, Polarstern is the name of the research vessel and PS106 is the the consecutive expedition identifier asssigned by the operator of the vessel. Hence you might want to change "[...] from the Arctic expedition (Polarstern) - PS 106) [...]" accordingly so that a reader can more easily find information on that expedition online.

L295 - L297

I suggest removing all sentences referring to the dynamic filter processing chamber measurements in the Rinaldi study. You measured immersion freezing, not condensation freezing; therefore, these results are not relevant here. Also, the discrepancy between the dynamic filter processing chamber and the droplet freezing assay is a separate topic, and mentioning it here only distracts from your work.

L298

To which cold air outbreak are you referring? Was it during the Rinaldi study or yours? From the phrasing, this is unclear.

L301

It is not clear what the difference between the preparation techniques is. If you wash off the particle, you also end up with a particle suspension.

***L307

How do you come to the conclusion that the SML contributed? This has to be explained.

Figure S1

While potential readers are likely aware of the location of Svalbard, they may not know where Gruvebadet is located within the archipelago. Therefore, I suggest marking the study site on a map of Svalbard instead of the current map (or adding it to a separate panel, for example).

Table S3

How do you define open ice?

*** Heat sensitivity

1. Somewhere, the calculation of the percentage of heat-labile INPs should be explained.

2. I doubt that a percentage accurately captures the logarithmic nature of INPs.

3. Your calculation method is skewing the results since it is based on an unsound assumption. You assume that the concentration of heat-stable INPs is zero when it is simply below the detection limit of your setup. It is not only very unlikely that the concentration of heat-stable INPs goes immediately to zero when below the detection limit, but it is also not good scientific practice to make this assumption without mentioning it in the manuscript.

Unfortunately, there is no established, scientifically sound way to report the heat-labile fraction. One uncritical approach would be to report only values where the INP spectra of the heated and unheated samples overlap (see Gong et al., 2022). The disadvantage is that, especially at higher temperatures, it is not possible to report any values. 

I suggest calculating the heat-labile fraction in a conservative, robust way by replacing values below the detection limit with the lowest detectable value. This method provides a conservative estimate of the heat-labile fraction that represents the case with the least amount of heat-labile INPs. Along with your  existing values, which represent the maximum amount of heat-labile INPs, you can report a range for the heat-labile fraction of a sample. In my view, this is a better representation of the sample since it is impossible to determine the true value.

L307

What are you basing the statement that SA2 and SA4 exhibit high INA on? At -25°C, these two samples appear to have higher N_INP. However, at higher temperatures (>-15°C), this does not seem to be the case. Therefore, labeling them as high INA seems inaccurate.

***L309 - L311

The phrasing is this sentence gives the impression that the cold conditions are one of the indicators pointing towards INPs that are internal mixtures of organic/biological and Na-rich particles. If that is what you meant, then this needs explanation, as I do not understand how cold conditions alone are an indicator for this. If this is not what you meant, then the sentence needs rephrasing.

L315 - L316

It is certainly possible that the SML contributes biological INPs, but can you exclude other sources that are often discussed in this context, like  low-INA dust acting as carrier for high-INA biological INPs? If you can, then this is also worth mentioning in the text.

L324

Do you expect the biogenic part to be from local sources or from long-range transport as well?

L332

Please rephrase this statement. Polysaccharidic INP certainly can be enriched in the SML, but are rather heat-resistent, so the heat treatment does not indicate their presence (see Hartmann et al., 2025).

L343

Tab. S1 contains the precipitation at the measurement site. However, for your statement, the precipitation along the airmass trajectory is more important. Since precipitation is one of the HYSPLIT output variables, I recommend reporting the average accumulated precipitation along the airmass trajectory for each sample in the supplement to emphasize your point.

L351

At least a reference should be added to explain how the OVF is determined. Even better would be a few sentences describing it.

L371 - L372

Are these "overall OVF values" the mean of all the individual OVF values of all the particles in a sample? If yes, would not the median be a more appropriate measure? I would also recommend adding these numbers to Fig. 7.

L372 - L373

In general I would recommend the use of the Spearman rank coefficient instead of the Pearson correlation coefficient, because it does not assume linearity and is not easily affected by outliers. Regardless of Pearson or Spearman, five  data points are of course not a very solid basis to determine correlation. For that reason I would recommend to remove figure 8 from the main manuscript, because it highlights too much results with weak statistical basis. It can still be shown in the supplement and the values can be reported in the text, but the reader should be reminded of the small sample size and hence assess these values carefully. Also a scatter plot matrix could be shown in the supplement, as this allows the reader to assess correlation themselves.

L373

While I am not aware of fixed rules for the descriptive words used for the strength of a correlation, I never seen anyone to refer to an R of 0.5 as strong. Generally, this would be considered a moderate positive correlation. 

L394

I do not see how "promoting heterogeneous freezing via organic-induced deliquescence or restructuring" can be directly inferred from your measurements. You can of course mention this as a possible mechanism that could take place, but then the respective literature has to be cited and sentences that explain the mechanism need to be added.

L402

I have not seen where you show and discuss potential influence of the morphology. You should either add this or remove "morphology" from this sentence.

L402-403

Can this statement be supported quantitatively? Could you derive a mean organic layer thickness for the particles in a sample? Is there an existing metric for "patchiness" that could be applied to your measurements?

L413 - 421

In the first sentence you basically only state that there was a case of many Na-rich particles with a thin organic coating and that they had only moderate INA. I do not understand why this finding underscores the critical role of biological INPs, nor is the connection to mild, humid conditions clear.

Then, you discuss the particles from cold and dry cases, which I assume are SA2 and SA5. You state that these particles have enhanced INA. However, this is not entirely consistent with what you wrote in the Results section. In L282-283, you wrote that SA1 and SA5 are less ice-active than SA2, SA3, and SA4.

Fig S3

The airmass sampled on SA5 seems to have experienced the highest BC concentrations of all samples. BC emissions are also typically accompanied by the emission of sulphate species, but SA5 hardly contains sulphate in the CCSEM/EDX measurements. Do you have an explanation for this?

Summary & Conclusion

With only five samples it might be a viable idea to have summary table, which contains for each sample the central results from your analysis like:  meteo. conditions, dominant particles classes (EDX), mixing state, OVF, organic coating thickness/patchiness, N_INP at a high and low temperature, N_INP of heat stable INPs also at a high and low temperature.

Technical comments

L93

missing spaces:  "[...] GVB(Mazzola et al., 2016) [...]" and "[...] (Pluvio2 L 400 RH)(Ebell et al., 2025) [...]"

L93

"OTT Pluvio (Pluvio2 L 400 RH)"

It is not clear what is the company name and what is the instruments name. Please follow common conventions of citing instruments. E.g., OTT Pluvio² L (OTT HydroMet GmbH, Kempten, Germany). After that, please refer to the instrument by its correct name or introduce an abbreviation properly: "OTT Pluvio² L (OTT HydroMet GmbH, Kempten, Germany, hereafter referred to as Pluvio2, ...)"

L94

Typo "Ny-Aalesund" should be "Ny-Ålesund"

L94

missing article "OTT Pluvio2L weighing gauge [...]" should be "The OTT Pluvio2L weighing gauge [...]"

L96

remove "So,"

L99

Plural: "[...] 5 pairs of filter samples was collected [...]" should be "[...] 5 pairs of filter samples were collected [...]"

Table 1

The headers are aligned differently. One is centered and the other is left-aligned. Please use consistent formatting. Additionally, I recommend that the authors take a look at the "APA Style manuals" and consider revising their tables. These manuals provide good guidelines for creating appealing, readable scientific tables.

L135

"figures out" sounds colloquial. Suggestion: "determines"

L138

missing space  "[...] behavior(Lata et al., 2021, 2023)." should be  "[...] behavior (Lata et al., 2021, 2023)."

L246

Write "back trajectory" consitently with or without hyphen throughout the document.

L286

too many parentheses "[...] range (e.g., (Hill et al. [...]", should be "[...] range (e.g., Hill et al. [...]"

L287

missing space "[...] N_INPmeasured [...]" should be  "[...] N_INP measured [...]"

L295

citation in parentheses is unnecessary since the study is already cited within the sentence

L309

Plural. Suggestion: "Though SA2 and SA4 contain 16.1 and 12% OC-rich particles, respectively."

L331

There seems to be a space at the beginning of this line.

Use consistent panel identifier throughout the manuscript (Fig. 1, 2, 3 used "(a), (b), etc."; Fig. 4 used "a), b), etc."; Fig. 5 used "a., b., etc.")

Use consistent parameter abbreviations: Fig. 5 denotes the INP concentration with a lower case "n" (n_INP), while the text uses the upper case letter (N_INP)

References

Huang, Y., Adebiyi, A. A., Formenti, P., & Kok, J. F. (2021). Linking the different diameter types of aspherical desert dust indicates that models underestimate coarse dust emission. Geophysical Research Letters, 48, e2020GL092054. https://doi.org/10.1029/2020GL092054

This manuscript presented a case study on the impacts of aerosol physiochemical characteristics on the ice nucleation activity. They conducted micro-spectroscopic single-particle analysis and derived the ice nucleation concentrations from droplet freezing experiments. Although with limited number of samples analyzed, the manuscript shows the complexity of interaction between particle mixing state and ice nucleation activity. The topic of this study fits the scope of this journal. There are several issues need to be addressed before it can be considered for publication.

Major issues:

1. Missing information on the methods for (1) INP concentration calculation, (2) fraction of heat liable stable INP, (3) FLEXPART back-trajectory analysis. Brief descriptions are important to readers and reviewers for judgement on the methodology.
2. The focus of this manuscript is connecting the particle physiochemical properties to ice nucleation activity, however, this a mismatch in particle size and sampling duration between the samples used for particle characterization and ice nucleation experiments. In addition, some organic or salt components are soluble, it is very tricky when discuss the impacts of organics what like contain soluble components on the immersion freezing ability of particles. Please comment on these and make this clear or add a few sentences to discuss these potential artifacts in the manuscript when possible.
3. Ice nucleation ability (INA) and ice nucleation concentration are different concepts. Ice nucleation concentration is not only depended on the ice nucleation ability of particles but also on the total particle surface area in the sample. To compare the INA of different particle samples, it would be better to use normalized parameters, for example, ice nucleation active site density (over particle surface area or mass). Is there particle surface or mass measurements or estimation that can be used for this purpose? This is also related to previous comment. In addition, frozen fraction and freezing temperature (e.g., median freezing temperature) are usually presented in ice nucleation references. These are missing here.
4. The descriptions and discussion in the differences for five samples are inconsistent, for example, in the Line 277-283, L305, it claims they are similar; however, in Line 307, 321-326, the discussion assumes that five samples have different INA. Please provide quantitative comparison with uncertainties on this comparison.

Technical comments:

1. L119, for the HYSPLIT analysis, using 50 m maybe too low,  100 or 200, 500m which are usually used. Please comment on this.
2. L280, the mean is 0.3 but the range is between 0.1-0.3? these numbers don’t match.
3. Figure 5, how do you calculate the percentage of heat liable INPs? In this figure, are these averaged over different temperatures?
4. Figure 8, how many data points were used for this analysis? How about the significance level?