Environmental Drivers of Arctic Low-Level Clouds: Analysis of the Regional and Seasonal Dependencies Using Space-Based Lidar and Radar

Dziduch, Aymeric; Mioche, Guillaume; Coopman, Quentin; Bazantay, Clément; Delanoë, Julien; Jourdan, Olivier

doi:https://doi.org/10.5194/egusphere-2025-2698

Preprints

https://doi.org/10.5194/egusphere-2025-2698

Preprints

03 Jul 2025

| 03 Jul 2025

Environmental Drivers of Arctic Low-Level Clouds: Analysis of the Regional and Seasonal Dependencies Using Space-Based Lidar and Radar

Aymeric Dziduch, Guillaume Mioche, Quentin Coopman, Clément Bazantay, Julien Delanoë, and Olivier Jourdan

Abstract. Low-level clouds play a crucial role in the Arctic surface energy budget and hydrological cycle. However, their representation in climate models remains challenging due to limited observations and complex interactions between local processes and large-scale environmental conditions. This study analyzes eight years of active remote sensing observations from CALIPSO and CloudSat to investigate the regional and seasonal distribution of four types of low-level clouds: warm liquid, ice-only, mixed-phase clouds (MPCs), and unglaciated supercooled liquid clouds (USLCs). Our results show that 51 % of low-level clouds are located below 3 km. A reassessment of cloud-type frequencies reveals that MPCs occur 17 % of the time, ice-only clouds 20 %, and USLCs 12 %. Notably, this work provides the first satellite-based assessment of USLCs over the Arctic, showing that their occurrence can reach up to 25 % in oceanic regions during transition seasons. The Svalbard region and Bering Seas emerge as the cloudiest areas, where liquid-containing clouds are prevalent, while ice clouds dominate over Alaska and Siberia. Using multiple linear regression, we quantify the influence of key environmental drivers on cloud-phase occurrence. Surface temperature, lower tropospheric stability, mid-tropospheric humidity, and marine cold air outbreaks (MCAOs) are identified as dominant factors. MPCs are particularly linked to dynamic conditions such as MCAOs, especially over oceanic regions and during transition seasons, whereas USLCs are more associated with stable and drier mid-tropospheric environments. The results highlight the impact of air mass intrusions on the regional distribution of the cloud phase partitioning, offering guidance for improving their representation in climate models.

Received: 06 Jun 2025 – Discussion started: 03 Jul 2025

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 2453 KB)

Supplement (1156 KB)

Download & links

Aymeric Dziduch, Guillaume Mioche, Quentin Coopman, Clément Bazantay, Julien Delanoë, and Olivier Jourdan

Status: final response (author comments only)

RC1:
'Comment on egusphere-2025-2698', Anonymous Referee #1, 01 Aug 2025

Dziduch et al. use an existing spaceborne radar and lidar product developed for the Southern Ocean to map Arctic cloud distributions. They further assess the environmental drivers of the cloud distributions using reanalysis. I have three main concerns. First, the authors have not explained what new questions their study addresses, why those questions are important, and how their results add new understanding that answers those questions. This work is far from the first to assess the distribution of Arctic clouds and their environmental drivers using CloudSat and CALIPSO. While there many indeed be new and additional value added from this analysis, this value is not well motivated or explained by the manuscript text. Second, the authors use environmental factors and a cloud product that may be inappropriate for the Arctic. The reasons for the selection of a Southern Ocean cloud product and environmental factors selected and why they are appropriate for the Arctic is not explained. Finally, the scientific writing lack precision and organization. The scientific arguments also lack precision and at times are unconvincing. This paper was not written for the reader. It is dense in information that is not well placed. It is not easy for a reader to follow or understand. The text does not effectively communicate the main information in a way that flows naturally and answers obvious questions. If the authors chose to revise the text, I recommend guidance on writing a scientific paper (e.g., reading Schultz et al. 2022 https://journals.ametsoc.org/view/journals/mwre/150/11/MWR-D-22-0277.1.xml) for As such, I recommend this paper be rejected in its current form.

Specific comments
Line 13-19. The motivation for this study is not well explained. What science question does this study address and how does it build on what has already been done? How does this observational work connect to Arctic cloud process biases in models?
Line 18. “51% of low-level clouds are located below 3 km”. This text does not make sense. What is the definition of “low”? Did the authors mean all clouds?
Line 19. “reassessment”. Why is a reassessment needed? What is the new value of this work? First main concern.
Line 20. “first satellite-based assessment”. Why is this the first? Previous studies have mapped the seasonal distribution Arctic clouds by cloud phase (e.g., Cesana et al. 2012). What is meant by “unglaciated”? How do you know the spaceborne downward looking perspective didn’t miss ice near the surface? The radar cannot see close to the surface due to ground clutter. The lidar can be attenuated and therefore not detect the bottom of the cloud.
Line 34. The citation strategy here is strange. Why is the first study cited here in 2017? Studies on this topic range back well before 2017.
Line 35-79. This text lacks precision and organization. The text mixes topics from a wide range of previous studies without a coherent message. Instead of listing a wide range of previous studies, the introduction needs to review the specific key findings from previous work and then use that as motivation for why this additional work is needed in this study.
Line 78-80. The objectives for this work are not specific. It is unclear what new is done in this work that hasn’t been done by previous studies. Also, the placement of the two objectives at the end of two separate paragraphs is odd/confusing for the reader. Why not combine them in a single paragraph that discusses the objectives for this work.
Line 99-108. This list of what is contained in each section could be deleted. It is not specific enough to be useful.
Line 113. Why 2007 to 2016? I recommend either analyzing the whole mission, or analyzing the first five years (2006-2011) because the daylight operations only period for CloudSat begins in 2011.
Line 110-175. As this is far from the first study to analyze CALIPSO or use DARDAR – I think that this information can be more precisely and effectively communicated. An extensive description of CloudSat and CALIPSO can be found in the references cited and is not needed here. DARDAR could be introduced in a 1-2 paragraphs. Only include the information that is specifically relevant and important for this study. For example, Why is DARDAR an appropriate product to use to achieve the new objectives and advance new science in this study?
Line 150. Why is this study using a classification developed for the Southern Ocean over the Arctic? The Southern Ocean is dominated by mid-latitude clouds from the southern hemisphere storm-track. In contrast, the Arctic has polar clouds that are shallower, colder, and can be found in stable atmospheric conditions.
Line 178-204. While a discussion of the limitations and uncertainties is certainly important, this text does not effectively communicate that information. How are the instrument limitations going to affect the ability of this study to advance understanding and produce new insights about Arctic clouds?
Line 205-240. This information is very general. Why are the same environmental factors being used for Arctic clouds as for clouds at lower latitudes? That assumption is not well explained.
Line 240. Please use “cloud cover” instead of “OC”. OC is not a commonly used acronym and will make it hard for your reader to understand what you are saying.
Line 240-242. “(not shown here)” Why do you start with results that you do not show in the paper? Instead, start with the new results from this work.
Line 250. “vary significantly”. Be more specific and quantiatitve.
Line 256. What is “Arctic weather reanlsysi”? This sentence starting with “The spatial analysis…” does not make sense.
Line 268. The distribution of Arctic clouds has been presented in many other studies. See Cesana et al. 2012 Figure 1. What is this study adding that is new?
Line 273. “Weather systems have a direct influence on cloudiness in certain areas”. This text is not precise, and therefore does not add new information.
Line 276, 312. “This could be explained by the XXX”. “could be related to the weakening of the tropospheric vortex”, “could result from surface coupling”. This wild speculation about mechanisms without evidence is inappropriate. The results section should describe the new results obtained in this study. Citation to previous work in the results section is typically rare.
Line 460. “Previous results…”. This first sentence does not make sense. I am not convinced by what is written here because the authors have not been specific or precise about what previous studies have not included.
Line 465. The definition of cloud will also affect the comparison of this work to other studies. Have the authors considered that?
Line 483-485. While this study may be the first, it is unclear what the value added is for Arctic cloud studies because the authors have not justified the use of these cloud types and what new can be learned from them.
Line 506. I am not convinced that this study is innovative by looking at environmental drivers of clouds. Many cloud studies do so.
Line 523-525. I do not see any additional added value of this study in understanding cloud-sea ice interactions. The analysis completed here cannot separate the influence of the surface and the atmospheric circulation. See Morrison et al. 2018 (DOI:10.1002/2017JD027248) introduction for a description of why that separation is important.
Appendix B. Why is this key methodological information not presented in the main text? Reviewing linear regression equations is not needed. More justification for the selection of the selected environmental drivers is needed.

Citation: https://doi.org/10.5194/egusphere-2025-2698-RC1
- AC1:
  'Reply on RC1', Aymeric Dziduch, 10 Aug 2025
  Response to Reviewer #1
  We thank the reviewer for his detailed comments. Below, we address each of the main concerns raised by the reviewer : (1) needs to clarify the novelty of the work and the motivation (2) justification of the choice of datasets/products (3) improvements needed in the scientific writing.
  Novelty and added value compared to previous studies
  
  We are well aware that several studies have already analyzed the seasonal and geographical distribution of Arctic clouds and links with some specific (and limited set of) environmental conditions using CloudSat/CALIPSO data. However, we believe that our study yields several novel findings, which we will outline more explicitly in the revised version, should the editor/reviewer authorize it :
  
  First quantification of unglaciated supercooled low clouds (USLCs) in the Arctic: To our knowledge, the regional and seasonal occurrence of USLCs have never been investigated in the Arctic before. We find a median occurrence of 12%, with pronounced regional differences and a bimodal seasonal cycle (Sections 3.2, 3.3; Discussion, pp. 18–20). The aerosol-cloud interaction processes and larger scale environmental conditions responsible for the formation and maintenance of USLCs and MPCs may be different. For instance, we show that MPCs are particularly linked to marine cold air outbreaks (MCAO) over oceanic regions and during transition seasons, whereas USLCs are more associated with stable and drier mid-tropospheric conditions. Other studies do not investigate these two types of cloud separately, referring to them as 'liquid-containing clouds.
  
  New assessment of the frequency of occurrence of mixed-phase clouds (MPCs) and its distribution based on a “more relevant” definition of MPC: Unlike earlier analyses (Matus & L’Ecuyer, 2017), we define low-level MPCs as successive layers of single-phase or mixed-phase layers clouds. This definition is closer to in situ observations (e.g., Mioche et al., 2017) and leads to new estimates of the frequency of occurrence of MPCs which are more realistic (Section 2.2, pp. 5–6; Discussion, pp. 19–20).
  
  Arctic regionalization: We subdivided the Arctic into meteorologically coherent regions, allowing us to highlight processes specific to coastal, oceanic, and continental regimes. This was done rarely or differently in previous studies, and the segmentation of different cloud types using DARDAR-SOCP has never been carried out in association with an Arctic regionalization.
  
  Statistical study of the influence of environmental parameters on Arctic low-level clouds using multiple linear regression (MLR): We are aware that this method assumes linear relationships between environmental parameters and occurrence. Nonetheless, it allows us to identify the parameters (lower tropospheric stability – LTS, marine cold-air outbreaks – MCAO, specific humidity at 700 hPa, etc.) that most strongly influence cloud presence across different Arctic regions with an assessment of their relative importance by cloud type, region, and season (Sections 3.3, 4; Figures 3–4).
  
  We will revise the introduction to highlight the open questions related to the characterization and the distribution of the low-level cloud types in the Arctic and how our results address them. (The cloud phase partitioning is not accurately represented/simulated in the majority of the models).
  Relevance of the cloud product and environmental factors used
  
  The DARDAR-SOCP classification was first presented and developed for the Southern Ocean (Bazantay et al., GRL 2024). However, we originally developed and validated the classification using data from Arctic measurement campaigns (Mioche et al., 2015; Mioche & Jourdan, 2018). It was primarily designed for low-level clouds in the Arctic but applied to SO clouds in our GRL paper (Bazantay et al., 2024)
  
  We explicitly discuss the limitations and uncertainties of the product in the Arctic context (Section 2.3, pp. 6–7), citing validations with Arctic in situ datasets (Mioche & Jourdan, 2018; Achtert et al., 2020; Järvinen et al., 2023).
  
  We restricted the analysis to the 0.5–3 km altitude range to avoid near-surface biases and to focus on low clouds that have a direct impact on the surface energy budget.
  
  Regarding the environmental parameters, the nine selected variables (LTS, MCAO, specific humidity, temperature, etc.) are identified as main factors that are expected to influence clouds in the “Arctic” literature (Kay & Gettelman, 2009; Morrison et al., 2019; Taylor & Monroe, 2023;...). We implemented extensive statistical analysis to avoid multicollinearity of these parameters. We will strengthen the methods section to further justify their relevance for Arctic conditions.
  Quality of scientific writing, organization, and clarity
  
  We understand the concern regarding text density and flow. In the revised version, we will:
  
  Reorganize the introduction to clearly state the context, research questions, hypotheses, and novelty.
  
  Structure the results into three logical blocks (spatial variability, seasonal variability, role of environmental factors), with key messages at the end of each subsection.
  
  Streamline certain methodological descriptions by moving technical details to the appendices or supplementary material.
  
  Follow the recommendations of Schultz et al. (2022) to improve logical progression and guide the reader effectively.
  
  We thank the reviewer for his suggestions, which we hope will contribute to improving the manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2025-2698-AC1
RC2:
'Comment on egusphere-2025-2698', Anonymous Referee #2, 03 Aug 2025

The paper analyses an eight-year time series of active satellite measurements of low cloud frequency of occurrence over the Arctic, thus extending previous work to longer time series. An existing method is used to distinguish four different types of low clouds and to present their spatio-temporal distribution. The extended data record includes more daytime-only observations and thus analysis of differences between “only day” and “day and night” measurements (Appendix A) is valuable. In general, the data record is very interesting, as little information about Arctic clouds (in general and not just low clouds) and their variability exists, but some more information on the data itself and sampling limitations is needed. The authors might consider a data publication for this part of the study.
Ultimately, the authors want to identify the drivers explaining cloud variability. For this purpose, they use a multi-linear regression between cloud fraction and a set of environmental parameters to identify the strongest factors of influence. The motivation for the different parameters in this rather simple approach seems a bit arbitrary and requires a better motivation. Furthermore, the interpretation of the regression results needs to be more careful. For example, the abstract statement “The results highlight the impact of air mass intrusions on the regional distribution of the cloud phase partitioning (abstract)” needs to be backed up by further evidence. There is considerable work needed to improve this part.
The paper is written rather lengthily, and, in my opinion, much of the text can be removed without significant loss of information (see my suggestions below). In particular, this holds for the introduction (section 1), which needs to be better targeted to the objectives of the study, and the discussion (section 4). The paper only contains 4 figures in the main text, 4 in the Appendix, and 4 in the supplement (which most readers don’t look at). I suggest a reorganization to make the manuscript stand on its own.
Given the limitations mentioned above (and detailed below), substantial revisions are required.
Please see
General Comments:
1. Motivation/objectives/conclusion: The storyline of the paper needs strong improvement. The introduction emphasizes the role of clouds on the surface energy budget and provides an extensive review of the literature on that. However, the study does not perform any analysis with respect to the energy budget and only gets back to this in the last sentence of the paper as an outlook. Therefore, this part of the introduction should be significantly shortened, and instead, the two objectives (O1 - provide spatial distribution of low-cloud types; O2 – identify key parameters determining this distribution) need to be sharpened and motivated more clearly. This should lead to clear messages in the conclusions. Multiple times, the paper mentions the aim of improving climate models, but no clear path on how to use the results of the study for this purpose is given.
2. The analysis with respect to O1 is relatively straightforward, though some explanation and definitions (e.g., treatment of orography, reason for limitation to low clouds, vertical sampling with lidar attenuation) seem to be missing, even when looking at the appendix and supplement. Because many researchers have used the “short” Cloudsat/Calipso data record for various purposes in the past, it is certainly of interest for a wide community to understand how the longer data record (with changes in sampling) changes the “climatology”. See also my specific comments for details especially as I see problems in the treatment of the “blind zone”.
3. With respect to O1 the authors claim: “The second objective of our study is to identify a set of key thermodynamic parameters that control or affect the regional distribution of the cloud phase.” Why do the authors limit themselves (in their statement) to these thermodynamic parameters, but in the end also consider sea ice concentration (SIC) and aerosol (AOD)? In general, the interaction of synoptic forcing and local convection will control the regional distribution, and thus, parameters characterizing these phenomena should be considered. This would be in line with the “grand challenge” described by Bony et al., 2015. https://www.nature.com/articles/ngeo2398 .
More concretely, the motivations for the choice of parameters should provide physical reasoning, e.g., why consider both MCAO and LTS? How to treat MCAO over land? Why ignore surface type characteristics and orography? Or subsidence? In fact, one interesting analysis is hidden in the supplement; i.e.,the clouds coupled to the surface. Furthermore, as the conclusion mentions airmass intrusions being of major importance, why not include information on these, such as available by transport metrics, circulation weather types, atmospheric river/cyclone diagnostics, etc. to back this up. Somehow, I have the feeling that the data available (IFS analysis) plays a role, as likely many interesting parameters would be available from ERA5.
In general, I am skeptical that a multiple regression on a grid point level will be able to capture complex interactions that are certainly highly non-linear. Furthermore, it has been shown for SIC that time lags play a major role. I couldn’t identify any surprising new results from the correlations. Today, various machine learning methods might be more suitable to give new insights.
4. The division between the main text, the appendix, and the supplement is challenging for the reader. I suggest eliminating the appendix to provide a better flow and only have detailed information (only to be used by very few) within the supplement. Figures S2 (to be revised) and S4 provide interesting information that should be part of the main text. The lengthy (2 page) Table B1 should be part of the supplement.
Specific Comments:

Abstract
L18: “Our results show that 51% of low-level clouds are located below 3 km.” Misleading, I think you want to say that there are 51 % low-level (<3km) clouds.
L19:” Why “reassessment” and not “assessment”
L23/24 – why now talk on cloud-phase
L25: “MPCs are particularly linked to dynamic conditions such as MCAOs..” – what other dynamic conditions are you looking at?. Also, the second part of the sentences(oceanic regions” should be deleted as by definition “Marine CAO) are over ocean.
L27-30: This sentence overemphasizes the results as no “airmass intrusions” were studied; rather, a static analysis was performed.
Introduction
L36-57: can be summarized in one sentence that no consensus on the surface energy budget has been reached yet, and that this is a challenge.
L58: Now the focus is on low-level clouds. A clear motivation needs to be given explaining why the paper limits itself to low clouds and ignores the rest. The following text is very dense, packed with citations and could be shortened.
L76: In the introduction of O1 “the assessment of the cloud type” is mentioned. For the reader, it is not clear what is meant by type. It needs to be mentioned that “type” relates to the cloud phase, and this to the radiative effect
L87 on: It might be helpful to explain how clouds are treated in models, e.g., large-scale vs convective, to explain which “environmental” conditions are of interest
L97: not only thermodynamic
Methods and data
In general, I could not find the information on how profiles/columns with mid/high clouds are treated that cause a lidar attenuation. Are these profiles ignored?
L115: When you mention the DARDAR cloud mask product first, you should briefly mention what it contains, i.e., assignment of a cloud class (one out of 18) to each atmospheric pixel.
L127: delete “initial” or change to “true”
L133: Delanoe and Hogan
L154: Sometimes DARDAR-Mask is written with v2.23, sometimes not – any special reason? In line 161 DARDAR version 2 is mentioned – what is different here?
L155: The depth of Cloudsat’s blind zone depends on the surface type. It is lower over the ocean than sea ice and especially high over land, where it can reach up to 1.2 km (Maahn et al., https://doi.org/10.1002/2014JD022079), and even over the ocean, the effect is high (Schirmacher et al., 2023; https://doi.org/10.5194/amt-16-4081-2023). The authors only omit 0.5 km, which seems to be not sufficient. Does 0.5 km mean above ground level?
L167: ..ARE retrieved
L170+: in this paragraph of information is repeated, time frame, altitudes. Please check if the new information (leave out of years ) could be moved into the beginning of section 2.1 to have all satellite information together. This does not fit to the “cloud classification”
Maybe think about combining 2.1 and 2.2 and deleting repetitions and lengthy descriptions.
L172: Why actually 2 deg lat and 2 deg lon? At these latitudes one deg latitude is more than 3 longitudes…
L177 and following: In my opinion, section 2.3 is unnecessary and distracts the flow. The uncertainties should be mentioned when the respective parameters are introduced. It is also not appropriate to transfer uncertainties from the Southern Ocean to the Arctic, especially as clouds over land are also considered in the Arctic
L213: Why for a week? MCAOs have rather variable periods, and over land, we have (even in the Arctic) a daily cycle?
L203 and following: As mentioned before, a more physical explanation of the choice of parameters is needed. The lengthy discussion on the accuracy of SIC and AOD is not needed. For the latter, it is more important to discuss sampling as AOD can only be derived when no clouds are present.
L235: Here, a new paragraph discussing the processing for all parameters is needed. What is the temporal resolution? The description of the regression (currently) in the appendix should be moved into section 2. However, it needs to be discussed how the regression will handle non-existing data, e.g., SIC. This is later on stated in line 532 by the authors themselves, but would have been already clear from the approach.
L240 and following: Why not also show the “all clouds”? As the section is written now, it is difficult to identify significant new results. Therefore, I advise the authors to be less descriptive but more oriented towards answering scientific questions. This would avoid repetitive statements and trivia; e.g., with respect to warmer clouds over warmer regions (L485 “We show that warm and ice clouds exhibit opposite seasonal cycles,..”. In general, the issue of ocean vs land in the treatment of the data set would be good to discuss. Statements like “This shows the influence of oceanic regions on coastal zones” are unclear to me.
The separation between the discussion (section 4) and Sections 3.1 and 3.2 makes the paper very difficult to read, also because the use of different methods between all studies needs to be considered in the interpretation. Can the authors demonstrate the value of their classification into four classes compared to other approaches?
Conclusions:
The authors present four key results:
1) Provocatively summarized: warm liquid clouds occur mainly where it is warm, and ice clouds occur in cold conditions.
2) Description of the geographical distribution of MPC and USCL.
3) Results of the simple regression approach. Here the authors try to go to a physical explanation (whose intention is good) but I find the statement “MPCs, which are often associated with more pronounced air mass movements such as MCAOs” problematic as except for a simple MCAO index (averaged by one week and valid for a grid point) no “air mass movements” are considered.
4) The lack of a direct correlation with SIC is found to be minor, but this has been shown before (see also my comments above).
Afterwards, they state that “This study highlights important directions for future research aimed at refining our understanding and representation of Arctic low-level clouds”. But what are the new directions?

Figures
Figure S2: Box plot with min/max limitation does not make sense for a value between 0 and 1. I guess this is shown, though it is not explained in the caption. All regions have the same bar. I suggest to use 10 and 90 percentiles.
The plot would have been interesting in the main part also separately for low clouds.
The tables in Appendix B need to go into a supplement.

Citation: https://doi.org/10.5194/egusphere-2025-2698-RC2
- AC2:
  'Reply on RC2', Aymeric Dziduch, 10 Aug 2025
  Response to Reviewer #2
  We thank the reviewer for his comments and for recognizing the scientific value of our dataset and of the proposed analysis. The key issues raised by the reviewer are addressed below.
  1. Additional information on the dataset and sampling limitations are needed
  We understand that a better presentation of the dataset and its limitation is necessary. In the revised version of the manuscript, we plan to :
  In the methods and data section (section 2) : Add a paragraph describing more explicitly the satellite overpass frequency, the effective spatial coverage, the impact of “daytime-only” data after 2012, and its potential biases related to seasons and geographic regions.
  
  Insert the main results from Appendix A in the main text so that readers can better understand the impact of sampling on the statistics.
  
  2. Need of Motivating the choice of environmental parameters used in the regression
  We acknowledge that the rationale for selecting the nine environmental parameters may appear arbitrary. In the revised version, we will:
  Explicitly justify the choice of each variable based on the Arctic cloud literature (e.g., LTS, MCAO, specific humidity at 700 hPa, surface temperature, etc.), outlining the associated physical mechanisms. This justification will be supported by the statistical work carried out to select these parameters while minimizing collinearity within the model. The number of variables chosen for this study is sufficient, as adding further variables would introduce collinearity due to their high inter-correlation. Moreover, some variables are not necessarily available or computable at our resolution (ECMWF-AUX).
  
  Clarify that our multiple linear regression (MLR) approach is intended as an initial exploratory step to quantify broad regional and seasonal tendencies, rather than as a comprehensive predictive model.
  
  3. Interpretation of the regression results
  We agree that some statements in the abstract, particularly the sentence on the impact of air-mass intrusions, require stronger support. We will:
  Add in the results section clear examples (figures and maps) showing how the MCAO/WAI indices relate to the distribution of cloud phases.
  
  Revise the conclusions and abstract so that they only contain statements that are directly supported by our data.
  
  In the discussion, specify the limitations and actual scope of the MLR-derived results.
  
  4. Reduction and refocusing of the text
  We will follow the recommendation to make the manuscript more concise. In particular, we will modify :
  Introduction: removing overly general background on the impact of clouds on the arctic radiative budget and focusing on the specific objectives of the study (characterization of cloud types, geographical and seasonal distributions of the phase partitioning, large scale conditions that influence these distributions… )
  
  Discussion: refocusing on the main contributions and removing repetitions already covered in the results section
  
  Reorganizing sections to reduce overall length while improving the readability of the paper.
  
  5. Reorganization of figures
  We agree that the manuscript should be more visually self-contained. We will:
  Include some figures currently in the appendix/supplementary material into the main text.
  
  Reorganize figures so that key results (spatial variability, seasonal variability, MLR analysis) are visible without consulting supplementary material.
  
  We thank the reviewer once again for their suggestions, which will significantly contribute to improving the manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2025-2698-AC2

Aymeric Dziduch, Guillaume Mioche, Quentin Coopman, Clément Bazantay, Julien Delanoë, and Olivier Jourdan

Supplement

https://doi.org/10.5194/egusphere-2025-2698-supplement

Aymeric Dziduch, Guillaume Mioche, Quentin Coopman, Clément Bazantay, Julien Delanoë, and Olivier Jourdan

Viewed

Total article views: 320 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
280	26	14	320	22	11	16

HTML: 280
PDF: 26
XML: 14
Total: 320
Supplement: 22
BibTeX: 11
EndNote: 16

Views and downloads (calculated since 03 Jul 2025)

Month	HTML	PDF	XML	Total
Jul 2025	127	14	10	151
Aug 2025	143	12	4	159
Sep 2025	10	0	10

Cumulative views and downloads (calculated since 03 Jul 2025)

Month	HTML	PDF	XML	Total
Jul 2025	127	14	10	151
Aug 2025	143	12	4	159
Sep 2025	10	0	10

Viewed (geographical distribution)

Total article views: 318 (including HTML, PDF, and XML) Thereof 318 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 03 Sep 2025

Short summary

Low-level clouds play a central role in the Arctic climate, modulating the surface energy budget with competing warming and cooling effects. In this study, radar-lidar satellite products are used to investigate the geographical and seasonal variations of different cloud types observed over eight years. Oceanic regions are characterized by high occurrences of low-level mixed-phase clouds, which are influenced by atmospheric stability conditions, surface temperatures and cold air outbreaks.


Total:	0
HTML:	0
PDF:	0
XML:	0