Summary

In this study, Karalis and coauthors study the transformation of an air mass entering the Atlantic sector of the Arctic during March 2022. They use observations from the HALO-(AC)3 field campaign along with a single-column model to dissect the physical processes occurring within the air mass and validate the model simulations. They find that different physical processes influence the air mass cooling along its path, with near-surface radiative and turbulent cooling dominating over the ocean, and cloud processes and adiabatic cooling becoming more important as the air mass progresses into the marginal ice zone and sea ice areas. They also find that the single-column model generally simulates the air mass transformation realistically, but struggles to reproduce the stable boundary layer and is highly dependent on the vertical motion prescribed by the ERA5 reanalysis data used to force the model.

This study provides a unique perspective on Arctic air mass transformation, a process that is still not fully understood but is critically important to understanding the causes of Arctic-amplified warming. The paper is generally well-written and scientifically robust. I have a number of minor comments and technical corrections listed below. Once these comments are addressed, in my evaluation this will be a valuable addition to the literature on Arctic air mass transformation.

Minor comments

- General comment: Is this air mass considered to be "fully transformed" at the end of the 12–14 March 2022 study period, or did it continue cooling after the HALO-(AC)3 sampling ended on 14 March? At the end of the study period, was the air temperature characteristic of a cold Arctic air mass, or was its thermal state more characteristic of an air mass still in transition from mid-latitude to Arctic conditions? If it continued cooling, do the authors expect that the dominant cooling processes at the end of the study period continued to be most important for air mass cooling as the air continued to reside in the Arctic? From Fig. 4 it appears the air mass was still cooling, albeit at possibly a cooler rate, at the end of the study period. I understand that further simulations outside the study period are likely outside of the scope of this study, but it would be useful to provide some discussion about these aspects for context.

- General comment: The authors provide qualitative descriptions of which physical processes were most important for air mass cooling at different stages of its life cycle. Is it possible to integrate these contributions over time to provide a comparison of which processes contributed the most to cooling throughout the entire study period?

- L6: The meaning of "undistorted" air column isn't quite clear here and doesn't become apparent until later in the paper (e.g. L99–101, L112–118, L180–191). I suggest using the word "cohesive" in the abstract (as in L182) to be more clear.

- L150–152: What type of adjustment is needed for the model to be able to produce realistic skin temperature values?

- L145–154: I'm not entirely clear on the mixture of data sources here. So ERA5 is used for SIC, then CMEMS is used to quantify snow on top of sea ice and sea ice thickness? So both the snow on sea ice and the sea ice thickness are taken into account by the AOSCM? This is also unclear in L364–366.

- Fig. 2 and Fig. 3: Are these maps showing instantaneous snapshots of IVT, IWV, LWP, etc.? Or are these quantities integrated over time? Is the (Eulerian) ERA5 regular grid field of these values plotted, or are the values interpolated to the Lagrangian trajectories? I assume the cloud fields (LWP, IWP) and SEB values (SHF, LHF, etc.) are taken from ERA5, is this correct?

- L206–208: This is an interesting hypothesis about the quality controls in the assimilation scheme filtering the profiles out – is there any way to check this?

- L214–219: This paragraph is describing the cloud radiative effect – is it possible to directly calculate the cloud radiative effect and plot it on the maps?

- Fig. 4: I don't quite understand how cloud liquid and ice are represented in Fig. 4. Does the shaded area represent the additional atmospheric water in ice or liquid phase, in addition to the vapor-phase water (IWV)?

- Fig. 4: It is difficult to distinguish between the faded perpendicular lines for AOSCM/ERA5/IFS. Perhaps some could be plotted as dotted or dashed lines to make them easier to tell apart? Does each of these lines represent a timestep, such that the wider spacing of the lines over sea ice can be interpreted as faster air mass cooling and drying? Is appears that the uncertainty range is greater for the AOSCM than the other two products, is that correct?

- L326–327: How was Bulk Richardson number = 0.25 chosen as the threshold for the boundary layer? Is this threshold based on previous studies?

- Fig. 5: Is this figure created by averaging all the trajectories? Also, the uncertainty contours are difficult to see on the figure panels – perhaps they could be plotted with a darker color and/or thicker line.

- L366–368: So are the ERA5 and IFS-OF representation of the boundary considered more reliable than the AOSCM?

- L372: I think the reference to Fig. 5h is actually referring to Fig. 5k,h here? Please also check the other figure references in this paragraph (e.g. reference to Fig. 5i on L374).

- L376: To my eye, it looks like the IFS-OF mostly shows a single-layer cloud structure for about 75% of the MIZ and early sea-ice leg.

- Fig. 6: I don't see several features on this figure that are described in the text. For example, where does AOSCM simulate a drop in temperature below freezing levels (L394–395)? L395 states that dropsondes released over full sea-ice cover show minor surface cooling, but it looks the dropsonde observations are within the envelope of the other temperature profiles in panel (k)?

- Fig. 6: Unless I am missing something, I don't see where the cloud liquid comparisons (right column) are addressed in the text.

- L414–416: It sounds like it would be more accurate to call it the "liquid cloud layer" rather than the "cloud layer".

- Fig. 7: The caption does not describe panels e–g. Please check that all figure captions describe the figures in sufficient detail.

Technical corrections

- L3: "is" --> "are"

- L9: I think "simulate" or "reproduce" would be a better word choice than "emulate" here

- L26: Remove comma after "As"

- L42: "Airborn" --> "Airborne"

- L58: "imporant" --> "important"

- L68: "on" --> "to"

- L134: Add the word "are" after "tracks"

- L197: "dropping" --> "decreasing"

- L211: "air mass" --> "airmass" (to be consistent with the use of this word throughout the manuscript, I would argue that "air mass" is more commonly used in the literature but will leave it up to the authors whether they wish to change it throughout the manuscript)

- L216: Space needed in "ofthe"

- L293: "big" --> "large"

- L303: "uncertainty range ERA5 and IFS-OF curves" --> "uncertainty range of the ERA5 and IFS-OF curves" (?)

- L305: "and" --> "an"

- L307: "heat-to-moisture" --> "heat-to-moisture ratio" (?)

- L345: "while" --> "with" (?)

- L369: "dropping" --> "decreasing"

- L374: "bares" --> "bears"

- L391: "profiles" --> "profiles are" (?)

The paper investigates air mass transformations (radiative, turbulent, clouds, precipitation) associated with Arctic warm air intrusions. The paper is very well written, with a clear and concise introduction highlighting the existing knowledge and gaps in understanding cloud processes and air mass transformations, and presenting important results which advance our understanding of processes associated with warm air intrusions strongly affecting the Arctic climate. My major recommendations are to strengthen the abstract including key conclusions and slightly modify the results section structure to bring forward the air mass transformation processes and drivers, shifting the focus from model intercomparison. Also, the methodology section requires more details about the three models used in the study including relevant parameterizations. Below I provide more details. These are relatively minor revisions to clarify certain interpretations and to strengthen the presentation of the paper. I recommend the paper publication after they are addressed.

Abstract:

The abstract includes detailed methodology description and however lacking somewhat the main results. It will be beneficial for the paper if the readers could learn from the abstract what are the key conclusions regarding the air mass transformation.

Data and Methods section: overall it is very clear and well written however some key details regarding observations and models used in the study are missing. In particular:

2.1 Observations: it will be helpful to know more details about the dropsondes (which  type, parameters measured directly, vertical resolution, accuracy, etc)

2.4 Model description: key details are missing and will be helpful to include in the model description: resolution (ERA5 and IFS), cloud parameterization schemes, convection parameterization, and the snow pack model – in particular, details on how snow on sea ice is represented in the AOSCM.

Figure 1 caption: “Isobars between 940 hPa and 1080 hPa are plotted with thin(thick) white lines with a 5(10) hPa step” – while it is obvious from the values, it has to be noted that this is mean sea level pressure. Also, including selected makers on the plots will help

Marginal sea ice zone: typically SIC of 80% is used as the upper limit to define MIZ, while the authors used here 90%. Could the authors justify their choice?

Lines 93-94: “On March 13, at 12 UTC we launch 24-hour long trajectories, 600 in total, half of which were computed backward and half forward in time. “ – it will be helpful to show on the figure from where the trajectories launched on Fig. 2a (eg, can highlight in bold the 81ºN line portion near the appropriate meridian not to clutter the figure)

Fig 2 c) c) temporal evolution and spatial variability of integrated water vapor transport (IVT).

• Could the authors explain in more detail how IVT temporal evolution is calculated – is it the value for each specific trajectory (which would be difficult given the number of trajectories), along latitudinal line across the trajectory ensemble? Also for trajectories at which level?  

Lines 106-107: “These are identified using an integrated vapor transport (IVT) threshold of 100 kg m−1 s−1, generally preferred for Arctic WAI and AR detection “ – I suggest adding also an Arctic- focused paper, eg Viceto et al (2022), and a polar-focused reference by Zhang et al (2024) where specific thresholds are mentioned:

Viceto et al: Atmospheric rivers and associated precipitation patterns during the ACLOUD and PASCAL campaigns near Svalbard (May–June 2017): case studies using observations, reanalyses, and a regional climate model, Atmos. Chem. Phys., 22, 441–463, https://doi.org/10.5194/acp-22-441-2022, 2022.

Zhang et al: Extending the Center for Western Weather and Water Extremes (CW3E) atmospheric river scale to the polar regions, The Cryosphere, 18, 5239–5258, https://doi.org/10.5194/tc-18-5239-2024, 2024.

Lines 145-150: “The presence of snow on ice, not allowed in OpenIFS, has also been shown to mitigate surface energy and near-surface air-temperature biases (Pithan et al., 2016). “ – it is not clear how the presence of snow on ice is treated in the AOSCM – please include more details as this is an important parameter influencing surface fluxes (especially the surface albedo and netSW). Is it from observations or parameterized?

Fig 3: For the flux plots, it should be indicated in the caption that the flux is positive towards the surface. For SW and LW fluxes – please specify in the caption that these are net fluxes.

Fig 3 caption: “in terms of integrated specific water content “ – suggestion to add “integrated”

Line 218: “On the western flank of the airmass, where the LWP is larger, less solar radiation reaches the surface.. “ – the statement is not clear. As the plot is showing netSW radiation at the surface, a large impact over the perennial sea ice and MIZ is most probably explained by the high surface albedo and reflection of a large portion of the incoming SW flux. My earlier question – how the snow on sea ice is treated – is an important factor to consider also over the sea ice zones. However, it is not clear why the netSW flux sharply decreases from rather large values south of 70ºN to almost zero north of it and then stays around zero over the open ocean not changing much over sea ice. It will be useful to include also a map of the surface albedo together with downwelling SW and investigate processes controlling netSW in more detail (in the later section using AOSCM). Part of this can be probably explained by changing solar zenith angle however the differences across the 70ºN are too sharp.

Lines 220-225: “The spatial variability in skin temperature over the ocean also appears to be controlling the exchange of latent heat at the surface (Fig. 3h). Over the warm ocean, the strongly negative (upward) fluxes indicate the ongoing moisture uptake by the airmass. “ : can you please clarify your interpretation. The upward LH flux indicates surface evaporation, which indeed seems to be related to the skin T according to the plots, while it is also strongly controlled by the near-surface winds and the boundary layer RH. To be sure that this evaporated moisture is taken by the air mass needs verification if the trajectory was within the boundary layer. Was this the case over the region with surface evaporation? It is anticipated that these questions are considered further when using AOSCM. Then the limitations of using ERA5 shall be stated clearly also highlighting the added value of modeling investigations. Boundary layer height is later shown in the AOSCM (Fig. 5) however the two sections (3.2 and 3.3.4) are somewhat disconnected.

See for example:

Sodemann, H.: The Lagrangian moisture source and transport diagnostic WaterSip V3.2, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2025-574, 2025.

Sodemann, H., & Stohl, A. (2009). Asymmetries in the moisture origin of Antarctic precipitation. Geophysical Research Letters, 36(22). https://doi.org/10.1029/2009GL040242

Section 3.3.3: I suggest including a reference to Fig. 4 to make it clear the results are based on this figure

Lines 300-305: “The uncertainty range ERA5 and IFS-OF curves grows larger due to the slight divergence of the trajectory ensemble. “ – I am not sure to follow this interpretation. My understanding from reading the methodology is that the trajectories are the same, while thermodynamics state is represented by 3 different models (ERA5, IFS-OF and AOSCM). Thus, this is not the divergence of the trajectory ensemble but shall be explained by the differences in the model physics and processes representation. Could you please clarify and rephrase the statement.

Fig 5: Please indicate the time 0 (2/03/2022, 00UTC) in the caption

Line 355: “Over the MIZ, the subsidence spikes abruptly and over the sea-ice leg the vertical motion is predominantly upward, with ω increasing the deeper the airmass intrudes into the Arctic. “: is this updraft driven by cloud top radiative cooling (as described in Morrison et al 2012)? This can be seen in Fig. 7a discussed later in section 3.3.6.

Morrison, H., de Boer, G., Feingold, G. et al. Resilience of persistent Arctic mixed-phase clouds. Nature Geosci 5, 11–17 (2012). https://doi.org/10.1038/ngeo1332

To make more emphasis on the process understanding I suggest to move section 3.3.5 “Comparison with observed transformation“ before section 3.3.4 – this will show how each model represents each parameter before investigating the evolution in these parameters. Further, it will be beneficial to combine sections 3.3.4 “Vertical structure” with section 3.3.6 “Physical and dynamical drivers” explaining the drivers (Fig. 7) right away when presenting the vertical structure transformations (Fig. 5).

Section 3.3.5: As cloud ice and liquid content are key drivers of the radiative fluxes and updrafts, can the authors also include cloud evaluation, eg with cloud LWP from HAMP onboard HALO? I understand that this can be beyond the scope of the paper but if the data are already available this will be beneficial to see how AOSCM represents cloud properties.

Minor edits:

Line 168 : “of a strong cyclone”  - add ‘a’

Throughout the text – spaces missing at multiple places

Summary

Karalis et al. aim to enhance the understanding of airmass transformation occurring during warm air intrusions (WAI) in the Arctic. They propose a new single-column Lagrangian framework for simulating realistic WAI events and justify this by their findings that the WAI behaves as a column in the atmosphere. A case study of a WAI from 12-14 March is used to evaluate the performance of their framework, in which they do a comparison against dropsonde measurements, ERA5 and IFS forecast data. By looking at heat, moisture content and the vertical thermodynamic and cloud structure they conclude that the model adequately reproduces the transformation. They state the value of the model as a source for identifying common features between airmass transformations and for identifying model biases. 

The paper presents an interesting study aiming to enhance the understanding of warm air intrusions in the Arctic. My main recommendations revolve around highlighting the novelty and use of your work and clarifying the description of the methodology for this new approach. After these comments have been addressed I think the paper will be a valuable contribution for further work in understanding airmass transformation in the Arctic. Below I attach major and minor comments that hopefully will be useful in preparing a revised version of the manuscript.

Major comments

1. It was not very obvious from the Introduction how this paper innovates from the plain Lagrangian framework applied in Svensson et al., 2023, even though this becomes somewhat clearer later on in the methods section. The novelty and use of the single-column framework should be better highlighted in the text upfront in the introduction, throughout the results, and in the discussion/conclusions (see comment 2). 
2. The airmass trajectories for this case are almost entirely north-south oriented. It seems thus that for such straight-line trajectories one could get a lot of the same information from a simple cross-section without running the single-column model in addition. Thus, it is important to highlight what the additional value of this approach is. 
3. The authors make a strong statement in the conclusion about the novelty of this study, but it is difficult to distinguish the contribution here from the works of others, which is discussed  both in the conclusion and also to some extent in the results part. In order to make it easier for the reader to follow the argumentation of how these results are novel, a dedicated discussion section would be useful. This would also allow to focus more on the conclusions from this work in the final section. 
4. The method is not sufficiently clear, in particular when it is being referred to an ensemble. Some additional details on how the trajectories and the ensemble are obtained would be helpful for the interpretation. I suggest to illustrate this with a conceptual figure instead of using a result figure in section 2.2/2.3. 
5. Another step in the method that needs further justification is the meridional search for threshold values from trajectory points. It seems odd to go from a Lagrangian framework to a search for threshold values in an Eulerian perspective. Why not use a threshold in a Eulerian map of TCWV directly? When ‘stepping outside’ the trajectories, there is no more guarantee for that the airflow aligns and goes into the same direction.
6. L180: This appears to be a fundamental conclusion to move ahead, but the vertical alignment is not clear from the results. Maybe it could be quantified with a dispersion metric to underline how the trajectories move together? Additionally, a figure showing this result would be helpful, for example showing the vertical position of traced air parcels over time.
7. The title is not sufficiently connected to what is shown in the manuscript, which is a novel model framework illustrated by a single WAI case from a campaign. For that it is not necessary to promote the HALO-(AC)3 campaign in the title. I suggest a title along the lines of: “Lagrangian single-column modeling of Arctic airmass transformation during a major warm air intrusion”

General comments

Several sentences and parts of the manuscript are hard to read and it is not easy to grasp the flow in a paragraph. This is probably due to interruption of the sentences by references and by a reversed order of the old and new information in sentences (see for example Gopen and Swan, https://www.americanscientist.org/blog/the-long-view/the-science-of-scientific-writing). See technical comment L214 and minor comment L155 and L156 for examples. 

Figures

Figure 1 is too small, consider using 2 rows and 2 columns instead. The features are hard to distinguish, especially the green on top of the dark blue. The purple dots are nearly invisible. 

Figure 2: this figure is also a bit small. The figure caption does not sufficiently make it clear how to understand this figure. This is also connected with the uncertainty of how the trajectories are obtained (see major comments). 

Figure 7: the caption lacks information on panel e, f and g. 

Minor comments

L21: Could be useful to give an indication of the typical timeline referred to here. 

L69: The sentence is long and unclear, please rephrase. “The suite of Lagrangian observations available …”

L76: This section seems to be more of a weather description based on the observational data. Consider using a more descriptive section title. 

L147: Clarify the connection between the two sentences 

L155: Rephrase as “The modeled profiles at the final timestep of the previous simulation are used as initial conditions for the following simulation at each transition point between surface regimes.” 

L156: Rephrase as “Two additional preparatory simulations are performed over each sea-ice leg. The first one using… “

L165: Consider whether the information on climatological perspective could be better placed somewhere else. 

L189: Rephrase  “To the degree that this feature is common along WAI, …. “

It is unclear whether this is a statement or a question on whether they are common. 

L240. Section 3.3 contains mostly method material and should be moved to the methods section

L259: The reference to Fig. 4 is too early, the reader doesn't know what to look for yet. 

L266: Expand on this first description of Fig. 4, guide the reader through the details of the figure. 

L317: Rephrase as “At the end of the simulation, uncertainty in the heat content grows as well, due to slight variations in the forcing among the trajectory.” 

L318: Unclear sentence: “The same behavior is exhibited by the airmass in …”

L325: It is currently unclear now whether the warm and moist airmass is confined within the boundary layer or if the boundary layer depth is additional information. Rephrase sentence. 

L362: Add a reference here

L391: Unclear what this statement means: “The ensemble mean AOSCM, ERA5 and IFS-OF profiles in the center of each dropsonde cluster.” 

Technical comments 

L43: Remove the double “and”

L50: sampling →  sampled 

L53: Connect to a narrative instead of “them”  —> “…  reveal the time-scales and processes that drive them”

L95: Connect to a narrative instead of “Their” 

L143: Remove the last 6 words of the sentence: “for this part of the simulation” 

L163: develops →  developed 

L190: Replace “while” with “then/and”

L214: Move the reference to Fig. 3d to the back of the sentence. 

L216: Missing space 

L250: Add: “stems from two reasons”, or drop the point markings. 

L259: Connect to narrative instead of “Their”

L327: Drop the additional “the” before “threshold” 

L387: Drop the additional “observed” 

L416: therefore  → replace with “thus”

L421: moistest →  “most moist/humid”

L429: insert at → to zero at the top 

L458: misplaced dot

L469: mid-April → mid-March