the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Apr 2026
Aerosol-cloud interactions in Saharan mineral dust over the Eastern Mediterranean
Abstract. Mineral dust is a major component of atmospheric aerosol mass, yet its role in cloud processes remains poorly understood. While laboratory studies have explored its ability to act as cloud condensation nuclei (CCN) and ice nucleating particles (INPs), in situ atmospheric measurements are scarce.
This study presents unique in situ and remote sensing observations of dust-embedded clouds from the A-LIFE campaign in the Eastern Mediterranean. In situ observations on 20 April 2017 including shadow images and size distributions of aerosol and cloud particles combined with Doppler wind lidar (DWL) measurements were used to characterize dust and cloud microphysical properties. A novel size distribution retrieval for the second-generation Cloud, Aerosol, and Precipitation Spectrometer (CAPS) was developed – accounting for instrumental uncertainties, particle non-sphericity, and refractive index via a Monte Carlo method – and a machine learning-based calibration of the CAPS hotwire sensor for LWC measurements was introduced.
Using the newly developed analysis tools, this study presents a quantitative assessment of two different aerosol-cloud interaction processes in dust-embedded clouds: Measurements at the cloud-top indicated heterogeneous ice nucleation, in accordance with laboratory experiments and INP parameterizations. Furthermore, small-scale CCN activation of mineral dust particles into liquid droplets (~10 µm) was observed in the middle part of the cloud, with activation diameters as small as 0.13–0.23 µm, consistent with laboratory findings. DWL observations revealed vertical lifting as the driver of increased water supersaturation, enabling CCN activation. These findings provide new insights into the microphysical processes of dust-embedded clouds and their interactions with the atmosphere.
- Preprint
(2564 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-531', Anonymous Referee #1, 17 Apr 2026
-
AC2: 'Reply on RC1', Bernadett Weinzierl, 23 Oct 2026
We want to thank the two reviewers for their time and for their valuable comments, which helped us to improve the manuscript. We address all points below; our responses are given in bold.
RC1:
This manuscript presents a unique and comprehensive set of airborne in situ and remote sensing observations to characterize aerosol-cloud interactions within Saharan mineral dust layers over the Eastern Mediterranean. The study is technically innovative, introducing a novel Monte Carlo-based size distribution retrieval and a machine learning-based calibration for liquid water content measurements. The scientific conclusions regarding both heterogeneous ice nucleation and CCN activation of mineral dust appear robust and provide valuable insights into atmospheric microphysical processes. Overall, the paper is very well written and properly organized. I recommend accepting the paper for publication after a few, relatively minor revisions.
Detailed Comments
- Abstract, Line 23: The acronym LWC should be explicitly defined at its first mention to ensure clarity for all readers.
Thank you very much for this comment. We updated the text accordingly. - Introduction, Lines 24-25: The authors state that heterogeneous ice nucleation is the primary mechanism for ice formation at temperatures larger than 235 K. Given that this is a foundational physical threshold in atmospheric science, the manuscript should include a formal reference to support this specific value.
We included the reference. - Section 2.2.1 (Instrument description): While the spatial resolution and size ranges of the CAS and CIP are well-defined, the authors should include the temporal resolution (sampling frequency) of these instruments.
We deliberately do not specify a single temporal resolution for the CAS and CIP in the brief instrument description. The UNIVIE-CAPS data analysis is based on particle-by-particle data, which are inherently non-equidistant in time. Consequently, the effective temporal resolution depends on the particle arrival rate and on the subsequent data processing and aggregation procedure. A detailed discussion of these aspects, including the instrument data processing, is provided in Dollner et al. (2024). - Section 2.2.4 (Novel machine learning method for hotwire calibration): The application of a neural network to determine the convection component of the LWC sensor is a modern approach. However, to ensure scientific reproducibility and robustness, the authors must specify the architecture characteristics of the network (e.g., architecture type, number of hidden layers, activation functions) and provide details about the training process, including dataset split sizes and training strategies used to predict the convective loss.
We updated the text accordingly. - Section 3.2 (Investigation of ice crystals in the dust-embedded clouds): The manuscript utilizes the DeMott et al. (2010 and 2015) parameterizations (D10 and D15) to predict INP concentrations. However, these theoretical frameworks are introduced directly in the Results section without prior mention in the Methods. I recommend including a brief description of these parameterizations in Section 2.
We included a dedicated paragraph 2.5. - Section 4.2 (Mineral dust as CCN): The authors report that CCN activation was observed at supersaturations (SS) between 2% and 4%. This is significantly higher than the 0.2-0.3% SS typically reported in laboratory studies for similar dust samples. While the authors suggest that instrumental resolution may play a role, the manuscript would be strengthened by a deeper discussion of the physical implications.
Thank you very much for this comment. A deeper discussion of the physical implications of the observed CCN activation in relation with the observed supersaturations would have been ideal. However, with the mentioned uncertainties of the relative humidity measurements we do not want to overstretch the discussion and interpretation of our data. We believe, that for the discussion of the physical implications a different instrumentation with respect to relative humidity needs to be deployed, which would be a suggestion for future investigations.
Citation: https://doi.org/10.5194/egusphere-2026-531-AC2 - Abstract, Line 23: The acronym LWC should be explicitly defined at its first mention to ensure clarity for all readers.
-
AC2: 'Reply on RC1', Bernadett Weinzierl, 23 Oct 2026
-
RC2: 'Comment on egusphere-2026-531', Anonymous Referee #2, 28 Apr 2026
-
AC1: 'Reply on RC2', Bernadett Weinzierl, 23 Oct 2026
We want to thank the two reviewers for their time and for their valuable comments, which helped us to improve the manuscript. We address all points below; our responses are given in bold.
RC2:
This study captures a Saharan mineral dust transport event to the Mediterranean on 20 April 2017 through airborne observations, investigating the effects of mineral dust on ice-nucleating particles and cloud condensation nuclei. The findings make a meaningful contribution to advancing the current state of knowledge. However, the methodological descriptions lack sufficient clarity, and both the figure presentation and the writing of the results section are difficult to follow. I would recommend this manuscript for publication in Atmospheric Chemistry and Physics after the authors address the following comments accordingly.
Major comments:
- The authors have not included continuous line numbers in this version of the manuscript, which makes it difficult to pinpoint specific locations precisely. In the comments below, I will try to indicate the relevant section and sentence. In the Methods section "The Cloud Indicator Algorithm", the authors should briefly describe the computational approach underlying this algorithm, or alternatively, provide a detailed description in the supplement. Regarding the model inputs, it should be clarified whether the "size distribution" refers to the aerosol size distribution or the cloud droplet size distribution. A brief explanation of the distinction between liquid clouds, clouds in the Mixed-Phase Temperature Regime (MPTR), and cirrus clouds should also be provided. Furthermore, I believe that the calculation equation of the cloud-aerosol volume factor (fca) should be presented.
Thank you very much for this comment. We have changed the line numbers to a continuous format.
The Cloud Indicator Algorithm is extensively described in a dedicated publication (Dollner et al., 2024). We indicated more clearly, that the details of the algorithm can be found in this publication. - In Section 2.2.3 UNIVIE size distribution retrieval, I am not sure how to distinguish between hydrated sea salt (SS) and dry SS.
For the UNIVIE size distribution retrieval, the distinction between hydrated sea salt (SS) and dry SS is done using the relative humidity, where dry is considered up to a relative humidity of 65%. We updated the section accordingly. - In Section 3.2, I suggest that the authors should first introduce the significance of using the ratio between ND>30 μm and predicted INPs before presenting the results. Furthermore, it remains unclear why two schemes are employed for INP calculation, as no discussion of the differences between the two approaches is provided in the manuscript.
We included a dedicated paragraph 2.5. - In Section 3.3, the two CCN activation events reported by the authors are separated by only approximately 30 seconds, with each event being of very short duration. Given that these measurements were obtained during flight, it is considering whether the observed differences may reflect spatial variability in the sampled air mass rather than two distinct CCN activation events.
We thank the reviewer for raising this important point. The fact that aircraft was continuously ascending during the time period of the two CCN activation events, together with the observation of the growth of the aerosol particles to small droplets in both sequences, lead to our conclusion that these possible had been two CCN activation events.
Specific comments
- In Section 2.1, it is questionable whether citing a manuscript currently listed as "Weinzierl and Coauthors, in prep." is appropriate, as unpublished work in preparation may not be accessible to readers.
We have removed the citation. - I think the term "cloud indicator" does not appear to require italicization and should be formatted in regular text throughout the manuscript.
We changed the text accordingly. - In Section 2.2.3, it is suggested that the authors use numbered subsection headings (e.g., 2.2.3.1, 2.2.3.2, etc.) rather than bold and italicized text to denote sub-sections.
We changed the text accordingly. - Abbreviations for certain terms, such as sulfate and organics, should be introduced at their first occurrence in the manuscript. For instance, these terms appear earlier in the beginning of Section 2.2.3, yet their abbreviations are only introduced later in the manuscript.
We changed the text accordingly. - Abbreviations such as D and N should be consistently formatted in italics throughout the manuscript. While italics are applied correctly in Section 2.2.3, they are not maintained in the Results section.
We changed the text accordingly. - In Figure 4, the legend for each panel should be placed within or immediately adjacent to the corresponding panel, as the current layout is difficult to follow.
We changed the figure accordingly. - In Figures 1 and 5, it is recommended that the colors used to represent sulfate, nitrate, and sea salt be revised to align with the conventions commonly adopted in other literatures.’
We changed the figure accordingly. - In Figure 6, the particle number concentration appears to reach values as low as 10-4 cm-3. The authors should clarify the detection limit of the instrument or bias of algorithm used and discuss whether measurement/calculation uncertainties may be significant at concentration in this range.
We thank the reviewer for this important comment. Figure 6 shows vertically averaged particle number concentration profiles using altitude bins of 100m. Because the aircraft did not descend at a constant rate, the time spent within each 100m altitude bin varied. Consequently, the number of measurement samples contributing to each bin, and therefore the effective detection limit of the averaged concentration, differs between altitude bins. At very low particle number concentrations, the derived values are increasingly affected by the instrument detection limit and by counting statistics. We have indicated the concentration range in which the CIP detection limit begins to influence the displayed data by adding a grey-shaded area to Figure 6. - The ice crystal images presented in Figure 7 do not appear to be discussed in the main text, and the authors should either provide a corresponding discussion or consider removing them from the manuscript.
We updated the relevance of the images for the discussion of the heterogeneous ice nucleation section (4.1). - Figure 8: The same comment as Figure 4.
We changed the figure accordingly. - The tables in the manuscript do not appear to follow standard academic formatting conventions, and the authors are encouraged to revise them accordingly.
To the best of our understanding, the final formatting and layout of the tables will be revised by the journal’s typesetting team during the production stage. - I think language quality of the manuscript should be improved. For instance, the sentence on line 15 of page 19 appears to contain a grammatical error, and its intended meaning is unclear.
We have revised the sentence identified by the reviewer and carefully screened the manuscript for potentially similar cases.
References:
Dollner, M., Gasteiger, J., Schöberl, M., Gattringer, A., Beres, N. D., Bui, T. P., Diskin, G., and Weinzierl, B.: The Cloud Indicator: A novel algorithm for automatic detection and classification of clouds using airborne in situ observations, Atmosph. Res., 107504, https://doi.org/10.1016/j.atmosres.2024.107504, 2024.
Citation: https://doi.org/10.5194/egusphere-2026-531-AC1 - The authors have not included continuous line numbers in this version of the manuscript, which makes it difficult to pinpoint specific locations precisely. In the comments below, I will try to indicate the relevant section and sentence. In the Methods section "The Cloud Indicator Algorithm", the authors should briefly describe the computational approach underlying this algorithm, or alternatively, provide a detailed description in the supplement. Regarding the model inputs, it should be clarified whether the "size distribution" refers to the aerosol size distribution or the cloud droplet size distribution. A brief explanation of the distinction between liquid clouds, clouds in the Mixed-Phase Temperature Regime (MPTR), and cirrus clouds should also be provided. Furthermore, I believe that the calculation equation of the cloud-aerosol volume factor (fca) should be presented.
-
AC1: 'Reply on RC2', Bernadett Weinzierl, 23 Oct 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 629 | 231 | 56 | 916 | 58 | 56 |
- HTML: 629
- PDF: 231
- XML: 56
- Total: 916
- BibTeX: 58
- EndNote: 56
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents a unique and comprehensive set of airborne in situ and remote sensing observations to characterize aerosol-cloud interactions within Saharan mineral dust layers over the Eastern Mediterranean. The study is technically innovative, introducing a novel Monte Carlo-based size distribution retrieval and a machine learning-based calibration for liquid water content measurements. The scientific conclusions regarding both heterogeneous ice nucleation and CCN activation of mineral dust appear robust and provide valuable insights into atmospheric microphysical processes. Overall, the paper is very well written and properly organized. I recommend accepting the paper for publication after a few, relatively minor revisions.
Detailed Comments