The manuscript presents a newly developed payload, “PUFIN,” for vertically resolved sampling of ice-nucleating particles (INPs) using tethered balloon systems. The PUFIN system samples aerosol particles with a powerful scroll pump onto polycarbonate filters in three separate filter holders, each selected for sampling at a different altitude, using remote-controlled magnetic valves. INP abundance and temperature spectra are derived from offline filter analysis using an INP detection system (here, the Colorado State University INS). This is a very valuable approach to extending INP sampling into the vertical with tethered balloon systems, particularly because multiple filters allow contrasting altitudes, such as the free troposphere from the boundary layer. The paper is very well written and provides a sound description of the PUFIN system, its deployments across the US, and the INP analysis of the sampled filters. The INP community may also benefit from the author’s open-source approach, which provides technical drawings and details about PUFIN in a public repository, as well as freely available INP data from their previous balloon deployments. I recommend accepting the manuscript for publication in AMT.

Some recommendations for improvements are listed below:

1. The manuscript could benefit from showing meteorological data from the iMet-XQ2, particularly given that the INP samples exhibit varying abundance at different altitudes, rather than simply assuming atmospheric layering as the sole explanation (e.g., lines 303-305, 320). This could further emphasize the need for vertically resolved INP sampling.
2. Some more background on the sampling strategy would be nice. For instance, why were the specific sampling altitudes chosen? Which atmospheric layers are of interest?
3. Is it possible to derive a recommendation for the minimum sampling time in differing environments or seasons from the provided results in Section 4.2? A brief “lessons-learned” could also foster PUFIN rebuilds and deployments by other groups.
4. Figure 5 is difficult to interpret due to overlapping dots. Is each data point shown, or could one point be entirely covered by another? A separation into seasons might be helpful, as the temperature spectra are not only sampling time-dependent.
5. Some more interpretation and discussion of the derived temperature spectra in section 4.2 would be interesting, e.g., line 309: Which INP types are representative of the winter spectra?

General:

Creamean et al., 2025, describes the development of an aerosols sampling system (PUFIN) with focus on deployment on tethered balloon systems. The authors successfully built a system capable of sampling ice nucleating particles (INPs) in vertical profiles. Ice nucleation measurements were conducted downstream after extracting INPs from the filters of PUFIN. The technology is of great value for the scientific community, and the paper is very well suited for publication in AMT. Especially the accessibility to the data and possibility to request PUFIN for field measurements is highly appreciated. Nevertheless, the paper could be significantly improved by comparing the results and the sampling technique with literature and explain the example dataset in more detail. A discussion of the advantages and disadvantages of the sampling technique in comparison with other UAS and balloon systems and sampling techniques (e.g. impinger or impactor) could further improve the study. Therefore, I suggest the publication of the study after minor revisions.

Key Considerations:

Introduction

The introduction is well written and summarizes important knowledge and state of the art about aerosol and INP measurement at different heights in the troposphere. The authors also discuss the advantages of tethered balloon systems (TBS) over uncrewed aerial systems (UAS), such as longer sampling times. However, I was wondering how the flight path and position of sampling is different between TBS and UAS. For example, using fixed-wing drones¹ or rotary wing drones^2–4 could allow for a more precise flight path and location than a balloon system which also relies on the wind conditions and ground accessibility of the sampling locations. Furthermore, I was wondering what the key advantages of the PUFIN system are over the SHARK kit.⁵ I wish the authors could address the key differences in the introduction or in the discussion of the paper. A table comparing different unoccupied sampling tools for INPs^3,5,6 could be helpful in understanding the need for PUFIN.

Methods

PUFIN samples INPs on filters which can be further extracted for ice nucleation experiments. Their workflow is clever and straightforward, the cleaning procedures and the immediate cooling of the samples to -80°C prior INP measurements with the CSU INS is thorough. However, there are some remaining questions which I think deserve attention prior publication: What is the size cutoff and sampling efficiency for aerosol sampling on the filter? How do the authors make sure that all INPs are washed of the filter? What is the advantage of sampling INPs on filters over impingers⁷ or impactors⁵? Does the developed sampling procedure limit downstream analysis of INPs, for example cultivation of ice nucleating microbes⁷? What is the total weight of the setup? Could it in theory also be applied for other balloons or UAS?

Results

Figure 5 could be discussed with more detail. What is the main message of this figure? That PUFIN has the advantage over IcePuck with the sampling time? This is not obvious to me as a reader.

The resulting INP data (Figure 6 and 7) show a wide range of onset temperatures and INP concentrations. How does this range compare to other studies that measured vertical profiles of INPs?

The February vs July comparison in Figure 6 is particularly interesting. Could that correlate with the agricultural activity the authors were discussing earlier? Where any additional test performed on the sample (e.g. heat test) to test for the composition and origin of the INPs?

I wish the authors could spend more time discussing the data evaluation of the results from their freezing experiments: It might be worth (i) showing frozen fractions for some samples, at least in the SI, (ii) discussing how the blank background nucleated ice and why background subtraction was applied, and (iii) discussing the choice of calculating confidence intervals using Agresti & Coull⁸ over using the approach of Fahy et al.⁹ or others.

It seems that the authors use the two-sample t-test for determining statistical differences between that samples at different altitudes. Is that test suitable for testing cumulative nucleation spectra? If yes, what values were used for the test? Alternatively, the authors could just discuss the overlapping confidence intervals or also use a bootstrapping algorithm⁹ to back calculate the variations of the ice nucleation measurements.

Summary

What are future improvements that could be made for PUFIN?

Minor suggestions:

• Line 20: It might be worth dropping two or three concentration levels in the abstract to support the boundary layer stratification and the local vs. transported aerosol argument.
• Line 34: Change “freezing supercooled liquid …” to “freezing supercooled liquid water …”
• Line 46: Could you perhaps give an example of INP measurements at “surface” and how high these measurements normally extend? It is worth noting that INP measurements from towers can reach sampling heights of 60 m above ground.¹⁰
• Line 112: It might be worth adding the list and schematic design to the SI of the paper.
• Line 117: It might be worth defining standard liter in “sL min^-1”.
• Figure 1: Very minor edit, but it seems the mass flow meter symbol does not connect to the airflow arrows correctly.
• Table 1: The authors give the altitude of sample in meters above mean sea level (AMSL). Although sampling from 0 to 1000 m AMSL is theoretically possible, it would mean flying from sea level (0 m) to 1000 m. Their sampling locations in Colorado, Alabama and Maryland seem to be already a couple of meters above sea level. The authors should consider adding more accurate sampling heights into Table 1 and also consider adding the GPS coordinates of the sampling locations.
• Line 217: Consider changing “proportion” to “fraction”
• Line 220: “INP spectra are corrected using DI negative controls and subsequently blank-subtracted.” Please specify what was subtracted, the DI blank or the sample blank (filter from the PUFIN setup).
• Line 269: Worth citing the flow rate values of IcePuck for comparison with PUFIN.
• Figure 5: Check the left-hand side of the label, it seems to be cut off. I also got confused with the caption saying “Temperatures at which INPs were detected”. Does that mean onset temperatures, mean freezing temperatures or all temperatures at which droplets froze in the INS experiment?
• Figure 6: A statement to why and when sampling over an altitude range versus sampling at a set altitude was applied might be useful.
• Line 341: Is the “concentrations as low as 10-2 INP L-1 in as little as 28 minutes” also tied to a minimum onset temperature? I guess with the current workflow, 10-2 INP L-1 active below -30°C would not be detectable. Consider attaching a minimal onset temperature to this statement.
• This is very picky, but the authors use “warmer temperature” quite frequently. Technically, an object like a droplet can be warm, however, the temperature is a property of the object and therefore either high or low, not warm or cold.
• It might be worth considering including PUFIN in the title for further papers referencing the technique. Suggestion: “Reaching new heights: Profiling Upper altitudes For Ice Nucleation (PUFIN) on the Atmospheric Radiation Measurement (ARM) tethered balloon systems”
• Line 350: Again, great job and big step to help the ice nucleation community with access to INP measurements.

References:

(1)         Schmale Iii, D. G.; Dingus, B. R.; Reinholtz, C. Development and Application of an Autonomous Unmanned Aerial Vehicle for Precise Aerobiological Sampling above Agricultural Fields. J. Field Robot. 2008, 25 (3), 133–147. https://doi.org/10.1002/rob.20232.

(2)         Borchers, C.; Moormann, L.; Geil, B.; Karbach, N.; Wasserzier, D.; Hoffmann, T. Development and Use of a Lightweight Sampling System for Height-Selective UAV-Based Measurements of Organic Aerosol Particles. Atmospheric Meas. Tech. 2025, 18 (23), 7231–7242. https://doi.org/10.5194/amt-18-7231-2025.

(3)         Bieber, P.; Seifried, T. M.; Burkart, J.; Gratzl, J.; Kasper-Giebl, A.; Schmale, D. G.; Grothe, H. A Drone-Based Bioaerosol Sampling System to Monitor Ice Nucleation Particles in the Lower Atmosphere. Remote Sens. 2020, 12 (3), 552. https://doi.org/10.3390/rs12030552.

(4)         Seifried, T. M.; Bieber, P.; Kunert, A. T.; Schmale, D. G.; Whitmore, K.; Fröhlich-Nowoisky, J.; Grothe, H. Ice Nucleation Activity of Alpine Bioaerosol Emitted in Vicinity of a Birch Forest. Atmosphere 2021, 12 (6), 779. https://doi.org/10.3390/atmos12060779.

(5)         Porter, G. C. E.; Sikora, S. N. F.; Adams, M. P.; Proske, U.; Harrison, A. D.; Tarn, M. D.; Brooks, I. M.; Murray, B. J. Resolving the Size of Ice-Nucleating Particles with a Balloon Deployable Aerosol Sampler: The SHARK. Atmospheric Meas. Tech. 2020, 13 (6), 2905–2921. https://doi.org/10.5194/amt-13-2905-2020.

(6)         Böhmländer, A. J.; Lacher, L.; Brus, D.; Doulgeris, K.-M.; Brasseur, Z.; Boyer, M.; Kuula, J.; Leisner, T.; Möhler, O. A Novel Aerosol Filter Sampler for Measuring the Vertical Distribution of Ice-Nucleating Particles via Fixed-Wing Uncrewed Aerial Vehicles. September 4, 2024. https://doi.org/10.5194/amt-2024-120.

(7)         Šantl-Temkiv, T.; Amato, P.; Gosewinkel, U.; Thyrhaug, R.; Charton, A.; Chicot, B.; Finster, K.; Bratbak, G.; Löndahl, J. High-Flow-Rate Impinger for the Study of Concentration, Viability, Metabolic Activity, and Ice-Nucleation Activity of Airborne Bacteria. Environ. Sci. Technol. 2017, 51 (19), 11224–11234. https://doi.org/10.1021/acs.est.7b01480.

(8)         Agresti, A.; Coull, B. A. Approximate Is Better than “Exact” for Interval Estimation of Binomial Proportions. Am. Stat. 1998, 52 (2), 119–126. https://doi.org/10.1080/00031305.1998.10480550.

(9)         Fahy, W. D.; Shalizi, C. R.; Sullivan, R. C. A Universally Applicable Method of Calculating Confidence Bands for Ice Nucleation Spectra Derived from Droplet Freezing Experiments. Atmospheric Meas. Tech. 2022, 15 (22), 6819–6836. https://doi.org/10.5194/amt-15-6819-2022.

(10) Schrod, J.; Thomson, E. S.; Weber, D.; Kossmann, J.; Pöhlker, C.; Saturno, J.; Ditas, F.; Artaxo, P.; Clouard, V.; Saurel, J.-M.; Ebert, M.; Curtius, J.; Bingemer, H. G. Long-Term Deposition and Condensation Ice-Nucleating Particle Measurements from Four Stations across the Globe. Atmospheric Chem. Phys. 2020, 20 (24), 15983–16006. https://doi.org/10.5194/acp-20-15983-2020.