the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Apr 2026
The evolution of cloud microphysical properties during Cold Air Outbreaks – a composite approach to in situ measurements
Abstract. When cold, dry polar air is advected over a warmer ocean, a rapid development of clouds is observed. Several airborne field campaigns have been dedicated to these marine Cold Air Outbreaks (mCAOs). However, their properties, as well as their impact on the energy budget and water cycle, remains poorly understood. This study investigates the evolution of cloud microphysics during mCAOs through use of airborne in situ observations of ice crystals and water droplets from a recent Spring campaign in the Norwegian Sea (ISLAS2022). As individual flights only offer snapshots into certain parts of the mCAO evolution, a composite approach has been developed that integrates the in situ microphysical observations from multiple flights in order to capture the entire mCAO development. Thin, low ice clouds were observed over sea ice, reaching cloud top altitudes around 1200 m high. A rapidly developing "stratiform" region has been observed, where liquid-topped mixed-phase clouds increase in vertical extent, as the boundary layer deepens with increasing distance from the sea ice edge. At around 600 km fetch, rapid glaciation in a subsequent "convective" region leads to deep (up to 4500 m cloud top), almost completely glaciated, precipitating clouds that are reaching the end of their lifetime. While the observed microphysical properties are in agreement with earlier studies, the study highlights the potential of the composite approach in moving away from individual case studies to a more holistic microphysical picture of mCAOs, especially when statistics are improved by including additional campaign datasets in the future.
- Preprint
(29783 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-1650', Anonymous Referee #1, 12 May 2026
-
RC2: 'Comment on egusphere-2026-1650', Anonymous Referee #2, 16 Jun 2026
General comments:
The manuscript investigates the microphysical properties of clouds associated with marine cold-air outbreaks (mCAOs) during the ISLAS aircraft campaign in spring 2022. A Cloud Droplet Probe (CDP) and a Cloud Imaging Probe (CIP) were mounted on a research aircraft and used to perform in situ measurements of cloud microphysical properties during several flights over the Fram Strait under mCAO conditions.
The novelty of the manuscript is the presentation of cloud microphysical measurements obtained across multiple research flights. For this purpose, the authors introduce a framework in which all measurements are presented as a function of altitude and distance from the sea-ice edge. The authors further divide this representation into shallow, stratiform, and convective parts. Based on this categorization, several previously reported microphysical characteristics of Arctic clouds are identified and compared with results from other studies.
At this point, however, I must highlight what I consider to be the main weaknesses of the manuscript: The processing and description of the in situ measurements are not fully consistent. For the processing of the CIP data, the “all-in” method is applied, which results in the rejection of a substantial fraction of the measurements. The implications of this method, particularly the associated uncertainties in the sampling volume, are not discussed. This gives the impression that the consequences of applying this method have not been sufficiently considered. Other processing steps, such as the estimation of ice water content using a mass–dimension relationship, are also not applied consistently.
The division of the observed clouds into shallow, stratiform, and convective parts appears to be based largely on the authors’ subjective assessment. I recommend analyzing the individual flights separately and supporting this categorization with additional observations, for example satellite imagery.
The scientific conclusions presented in the Results section are based primarily on the findings shown in Fig. 6. These findings are compared with previous studies and generally show good agreement. However, the manuscript proposes several causal relationships that cannot be demonstrated on the basis of the current analysis. Additional data would need to be included to support these interpretations, particularly meteorological measurements from the aircraft.Overall, I consider the collected in situ cloud dataset to be valuable and to have considerable scientific potential. However, the issues in the following need to be addressed before the manuscript can be considered for publication. I therefore recommend major revisions.
Specific comments:
Sect. 2.2: The in situ instruments and the corresponding data-processing procedures are described in considerable and, in some parts, overly fundamental detail. These instruments and processing methods are well established and have already been described extensively in previous studies. This applies, for example, to Eq. (2), the calculation of SV(t), and Eqs. (3), (5), and (6). These relationships do not need to be derived or explained again in detail; references to the relevant previous publications would be sufficient.
The section should instead focus on the instrument-specific processing choices applied in this study and on any deviations from established procedures. For the CIP data, for example, it would be sufficient to state which settings were used in SODA and to describe any differences from the standard processing approach. At the same time, several important details are currently missing, including the binning applied to the CDP data, the calibration procedures for CIP and CDP, the treatment of dead-time corrections for CIP, and a discussion of the associated measurement uncertainties for all instruments.The authors average the in situ measurements over 5 s. Please provide a justification for the choice of this averaging interval. What is the basis for selecting 5 s, and why were the data not analyzed at the 1 s resolution?
27: This sentence appears to imply a direct causal relationship between cold-air outbreaks and extreme weather events. However, CAOs are generally part of the broader meteorological setting associated with such events rather than their direct cause. I suggest rephrasing the sentence to avoid implying a direct causal relationship.
31: I recommend moving the explanation of the mCAO index to the Methods section. It is also unclear why the mCAO index is calculated, as it does not appear to be used in the subsequent scientific analysis.
59: This statement seems somewhat far-reaching. Should the role of these clouds in the Arctic climate system not first be understood before assessing how they may change in response to Arctic amplification?
72 and 81: The cited paper by Sodemann et al. (2026, in preparation) is not publicly accessible. Consequently, important background information about the measurement campaign is unavailable to the reviewer. Please provide further information on the scientific objectives of the campaign, the measurement strategy used during the research flights, and the additional instruments operated on board that could potentially be included in the analysis.
105: Please clarify what is meant by the “standard suite of instruments.” Are there references that describe these instruments and their specifications? This information is important for the reader to assess potential errors and measurement uncertainties in the meteorological data.
Sect. 2.2.1: The manuscript states that the CDP measurements are divided into 30 size bins. However, it is not explained how these bin boundaries were defined. Were they selected arbitrarily, or were they based on a specific criterion? In particular, were Mie ambiguities taken into account when defining the size bins? Was the CDP calibrated during the campaign? The manuscript states that the smallest size bins are particularly affected by shattering effects, electronic noise, and coincidence. Were these effects investigated in laboratory experiments, or on what basis do the authors conclude that the smallest bins are especially affected? Figure 2 suggests that aerosol particles were measured in addition to cloud droplets. In particular, the retrieved particle size strongly depends on the assumed Mie curve. Which assumptions were made for the particle-sizing procedure, and how were aerosol particles distinguished from cloud particles?
The dark shaded green area labelled “CIP high conf.” cannot extend to 1600 µm when the “all-in” method is applied.119: What is the source of the value of 0.269 mm²? Is this a manufacturer specification or was it determined from an instrument specific calibration?
136: The assumption that the smaller particles measured by the CDP are liquid droplets, whereas the larger particles measured by the CIP are ice particles, is reasonable here. However, the manuscript should provide a discussion, supported by previous studies, explaining why this assumption is justified in this study. Numerous studies have shown that, in stratiform mixed-phase clouds, the smaller particles is typically dominated by liquid droplets, while the larger particles are predominantly ice crystals. The authors should discuss this evidence before applying the assumption.
150: Could the authors please explain why the “all-in” method was applied? I recommend reprocessing the data using the “center-in” method, as the “all-in” method results in the rejection of a substantial fraction of the particle images.
The authors further state that the maximum measurable particle size is 1600 µm. However, this is not possible when applying the “all-in” method, because at least one illuminated pixel is required at each edge of the array. The technically achievable maximum particle size is therefore 1550 µm.
Furthermore, for large particles, the sampling volume becomes very small when the “all-in” method is used (because the effective array width gets smaller), resulting in large measurement uncertainties. This could potentially explain why the cloud measurements indicate a substantially larger ice fraction than liquid-water fraction, as the concentration of larger ice particles may be overestimated because of the large uncertainty in the sampling volume. This could be assessed by examining the particle size distributions. However, the underlying data are not available to the reviewer, and this issue therefore cannot be evaluated.156: Were the data corrected for dead time, or were any analyses performed to quantify its effect? Optical array probes manufactured by DMT are particularly susceptible to dead-time effects, which cannot be corrected automatically by SODA. One possible approach would be to compare the 1D and 2D data to estimate the dead time. Since this is a mono-scale probe, the effect is likely to be negligible; nevertheless, it should be discussed.
162: The Brown and Francis (1995) mass–dimension relationship is based on a dataset in which particle size was defined as the mean of the particle dimensions in the (x) and (y) directions, where (x) is along the flight direction and (y) is along the photodiode array. This differs from the circle-fit method applied in the present study. Could the authors therefore explain why the Brown and Francis (1995) relationship was selected? More recent mass-dimension relationships are available that are better suited to modern cloud probes and to cloud types other than mid-latitude cirrus.
176: Could the authors please explain in more detail why the MVD was used here instead of applying the effective diameter to both thermodynamic phases?
201: The threshold of 0.01 g m-3 used by Korolev and Milbrandt (2022) is based on measurements from a Nevzorov probe. It would be more robust to apply a cloud threshold derived from a study using similar cloud microphysical instruments and based on observations of Arctic clouds.
233–241: The authors divide the evolution of the mCAO into three parts: shallow, stratiform, and convective. According to the authors, this classification is based on the microphysical properties of the clouds. However, the relevant microphysical characteristics, on which the classification is based on, are not described in sufficient detail. At present, the division into these three regions appears to be based on a subjective assessment rather than on a scientifically defined classification method.
I recommend that the classification should not be based solely on changes in cloud microphysics, as the variability of these properties is too large. Additional parameters should be considered. Satellite imagery, for example, could be used to determine whether a measurement region is located over the sea ice, within the cloud streets, or in the convective region. Further potentially relevant parameters include horizontal and vertical wind speed, temperature, relative humidity, and boundary-layer height.300: Could the authors explain why mean values are used in the analysis? Cloud microphysical quantities are often not normally distributed; therefore, the median would generally be a more appropriate measure of central tendency.
316: Where is an adiabatic LWC profile visible in Fig. 6a? This needs to be explained in more detail so that the reader can understand why the observed profile is interpreted as adiabatic.
328: this could be shown with vertical wind data from the meteorological data from the aircraft.
337–345: The authors suggest that the higher ice fraction in the shallow region is related to the number concentration of INPs and CCN. The current wording gives the impression that the number concentration of INPs is higher relative to that of CCN. However, the partitioning between LWC and IWC is most likely controlled by dynamical processes. I strongly recommend including relative humidity and boundary-layer dynamics in this analysis. In this case, the air may be supersaturated with respect to ice while remaining subsaturated with respect to liquid water.
352–354: The authors suggest that the lower ice-particle concentration near cloud top allows the liquid droplets to persist because of a limited Wegener–Bergeron–Findeisen process. However, I suspect that the enhanced LWC near cloud top is not caused by the vertical distribution of ice particles. (Such an interpretation would also imply a local depletion of INPs.) Instead, dynamical processes within the boundary layer are more likely responsible for the observed increase in LWC near cloud top. This could be shown by including other data from the aircraft.482: I ask to publishing the in situ cloud-probe data in a suitable repository. Without access to the underlying data, the results presented in the manuscript cannot be independently reproduced.
Technical corrections:
88: Please avoid using “(a)” and “(b)” here, as these labels may be interpreted as referring to panels (a) and (b) of Fig. 1.
260: Where can the main wind direction be identified in Fig. 4a?
317: Strictly speaking, this should be referred to as the “maximum mean LWC,” since individual 5 s measurements exceed 0.23 g m-3.
326: WBF is not explained in the manuscript.
685: The provided link is not accessible.
Citation: https://doi.org/10.5194/egusphere-2026-1650-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 253 | 167 | 22 | 442 | 14 | 19 |
- HTML: 253
- PDF: 167
- XML: 22
- Total: 442
- BibTeX: 14
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper introduces a new methodological framework for studying marine Cold Air Outbreaks (mCAOs). By integrating in situ airborne data collected from multiple flights during the ISLAS2022 campaign, the authors go beyond isolated "snapshot" case studies, offering a comprehensive perspective on cloud development from sea ice to open ocean. The manuscript is well-prepared and suitable for publication, although I have a few minor suggestions for improvement.
135-140: The authors assume that all particles with a measurement below 0 degrees Celsius are classified as ice by the CIP. It would be beneficial to address the presence of large supercooled droplets, which are often found in mCAOs, as the CIP is unable to distinguish between these large supercooled droplets and ice crystals. This potential misclassification of supercooled droplets could lead to significant errors in SLF values later on.
430-440: Does this field campaign include INPs or CCN observations? It would be beneficial to incorporate INPs and CCN measurements analysis to support your interpretations.
440: This hypothesis would benefit from a more detailed examination of LWF vs Temperature, rather than the isolated contours currently depicted in Figure 5.
Some mechanisms, physical-process interpretations, and literature comparisons may be better suited to the Discussion section. For example, lines 265-270: This indicates that the crystals ...
The discussion section could benefit from a more schematic conceptual model, with the physical mechanisms (arrows and process names) labeled directly on the diagram, rather than repeating them in both the results and discussion sections.