The (appropriately titled) paper, “Tethered balloon-borne observations of thermal-infrared irradiance and cooling rate profiles in the Arctic atmospheric boundary layer” by M. Lonardi et al. provides an interesting analysis of a valuable new dataset, tethered balloon profiles of radiative fluxes and associated parameters obtained during the MOSAiC expedition. The combination of analysis of the tethered balloon profiles with simulations provides a valuable description of the radiative characteristics of the arctic environment. I appreciated the modeling sensitivity tests, which provided some physical understanding of the clouds beyond their radiative properties.

For the most part, I thought that the flow of the paper was logical and appropriately cited previous work. It appears that sufficient information is also provided that would allow someone to attempt to reproduce these calculations. I would have appreciated more of a summary of some of the measurements. I know that details are provided in Lonardi et al. (2022), but the CloudNet retrievals play a central part in the discussion so it would be useful to have at hand a list of the measurements used for those retrievals. Similarly, while the focus of this paper is on radiative profiles, there are places in the paper where information about other measurements would have been useful (e.g. surface temperatures or temperature profiles).

I generally found the discussion of the net radiative fluxes and cooling rates to be clear. However, there were locations (see specific examples in the line-by-line comments below) where I feel that statements are made about the implications of these net fluxes without regard for other components of the heating budget. I particularly noticed this in section four, which focuses on the transition between atmospheric states. There seemed to be an argument made that the radiative cooling rates were responsible, or partially responsible, for the transition between states. But there are very likely other processes at play (e.g. advection, solar heating, turbulent heat fluxes). From my perspective, I don’t think this section adds significantly to the overall analysis and could be removed – unless there is an intent to make a more direct causal relationship between the cooling profiles and the changes in atmospheric state. If that is case, I believe that a stronger case needs to be made.

With the exception of the caveat noted for section four, I found the overall analysis to be useful and the conclusions to be an appropriate summary of the results.

Specific comments tied to line numbers in the document are provided below.

Comments

7-8: The magnitude of radiative cooling is given for cloudless and low-level liquid-bearing clouds – but the numbers have different meanings. For the cloudless case, I believe the value is a layer or column average, whereas the value for the low-cloud case is the peak at cloud top. That should be explained or corrected for consistency.

16-17: I am uncomfortable with the statement: “The enhanced warming in this region is a result of different feedback mechanisms known as Arctic amplification”. The references listed given provide good discussions of this topic. I think it would be more correct to say that the enhanced warming in the arctic relative to the global average is known as arctic amplification – and the cause for this phenomena, though not yet fully understood, includes several feedback mechanisms and other processes.

18: What is meant by “In this framework”? What framework is being referred to? I think it is fair to say that representation of clouds in most environments includes uncertainties.

14: I think the text would be more correct if modified to read: “demonstrating the (greater) radiative significance of the liquid clouds (relative to ice clouds)”. Or something that conveys the fact that liquid and ice cases are being compared here.

76: The text indicates the MOSAiC camp was “placed” on an ice flow with melting snow. It was located in such conditions in the summer – but it was originally placed in moderately thick ice in the late fall and allowed to freeze into the floe over the winter.

95: The solar component is noted to be of secondary importance and this is likely true; however, in some later conclusions, I am wondering how true this is – especially during summer – and whether some of the residual effects could be due in part to the neglected solar component. Has this been explored?

100: Is the longwave uncertainty really 7 W/m^-2 regardless of the actual irradiance? I would have thought it would be a percentage of the absolute value.

115: I believe plan-parallel should be plane-parallel

132-134: The text states that the summer atmosphere over ice is “typically covered by fog and/or liquid-bearing low clouds” but the text then goes on to say that the summer of 2020 was unusually warm resulting a “lower share of ice-bearing clouds in favor of purely liquid-water clouds”. This seems to suggest that the absence of ice is atypical – in contradiction to the earlier description.

138: The text states that autumn clouds are typically centered at about 1km height. I’m not clear what this means. Does this mean that cloud bases are typically around 1km? Or that the mid-point between cloud base and cloud top is typically around 1 km? Does this only refer to single-layer boundary layer clouds?

139-140: I don’t think it is appropriate to refer to 13 cases as “a vast presence”

140: The text states that low-level clouds were “inhibited by local topography” but no explanation is offered by how low clouds are inhibited. I have often seen clouds form in valleys – so it is not clear what is being referred to here. I believe further explanation is needed.

141: It may be a question of semantics, but I think it would be more correct to say that cloudless cases were the dominant state sampled in spring rather than “observed” There appear to be plenty of cloudy cases in the spring, but they didn’t occur at a time that BELUGA was flying (e.g. the middle part of April)

171: Could the difference between Ny-Alesund and summer (MOSAiC) be due in part to differences in the surface (land vs. ocean/sea ice)?

175-176: I would expect that variations in in the profile top Fnet would be due to the superposition of the cloud cover variation and variation in the vertical thermodynamic profile.

176-177: The text states that MOSAiC “mostly featured only low-level cloud”. To me, the top-of-profile distribution of Fnet looks more flat across the range of Fnet. There is a peak around -60 W/m^2 but there are a significant number at or above -30 W/m^2 as well.

189-190: I am wondering whether gradients in emissivity due to gradients in water vapor also play a role in the Fnet profile.

198: I don’t think this is a correct usage of the term “radiative equilibrium”. I believe that radiative equilibrium refers to the state where the radiative flux absorbed by a volume is equal to the radiative flux emitted by the volume. In the case here – a downward flux is stated as being equal to the upward flux. I don’t think that guarantees radiative equilibrium.

205-206: Would the assertion about cloud top cooling and its variation with the height of cloud top still be true as the cloud top was varied across a temperature inversion?

216: The text states that the “(cooling) signal becomes more variable due to the surrounding topography.” What is the basis for concluding that topography is responsible for cooling rate variability? It seems that there could be other explanations such as variability in the thermodynamic profile.

234: Are the cloud boundaries assumed to be fixed at those observed by CloudNet?

244: I am not following the meaning of the statement: “an increase in the LWP offsets almost homogeneously the Fnet in the layer between surface and cloud base”

249: The statement is made that “Nd plays a role in offsetting the net irradiance in and below cloud”. But doesn’t that depend on how Nd is varying relative to the LWP?

254-261: I have a few comments regarding this section on maximum and integrated cooling rates. To begin with, I am not clear why the figure is an appendix when the discussion is in the body of the article. I would suggest either fully integrating this discussion into section 3.4 and moving the figure there – or moving the discussion to the appendix with more explanation on how it amplifies the article. I am also unclear on the concept of integrated cooling rate. This is not a quantity that I have encountered before. I looked at the cited reference (Williams and Igel, 2021) and it is still not clear there how the quantity adds to the understanding of the cloud – but I would also note that Williams and Igel use different units for integrated cooling vs. the local (cloud top) cooling. I suggest that if you are going to present the integrated cooling, you provide some discussion of the importance of the quantity and verify that the units are appropriate.

277-78: The statement is made that “The net irradiance became strongly negative due to a reduced downward component in the cloudless atmosphere, resulting in an enhanced temperature decrease at the surface on the order of -5.3 K d^-1”. I’m not sure I believe this cause and effect. There may be other factors impacting the temperature trend including shortwave fluxes, turbulent heat fluxes, or advection. There are statements made at other parts of the manuscript (e.g. line 297 or 301) noting the possible role in some of these heating components – so I am surprised by this cause and effect statement here.

287: Is the quoted radiative cooling rate of -7 K d^-1 an average for the below-cloud layer? What is the “lowermost layer”?

309: Another cause-effect question – it seems that it is being asserted that the elevated cloud is causing the low cloud to dissipate – is that true? If so, what is the driver? The additional heating from the upper cloud? There must be other factors at play since multi-layer clouds are not unusual.

330: What is the basis for this hypothesis? This is another cause-effect situation. As stated – the impact of advection is not known. Is there any indication of what the advective tendencies might be?

338: In Figure 11, there seem to be more lines than I can find definitions for in the text.

347: The text mentions “three modes”. I would think of modes as three observed states. Two of the modes would seem to be better described as limiting states.

361: It appears that the ice amount in the upper level cloud has no impact on the net irradiance within the cloud. Is that true?

388: The statement is made that the radiative cooling ratees “(have) the capacity to modify the entire temperature profile”. I am not clear on the point of this statement. It is typical for there to be non-zero heating/cooling rates through the atmospheric column. This can result in modifying the temperature profile – but it can also have other effects (as is noted in places in the article).

404: I believe “car” should read “it can” – but I’m not certain.

418: There appears to be another cause/effect statement being made here about the transition between states. It appears (but I am not certain) that an argument is being made that the cooling profiles are responsible (or partly responsible) for the transition between states – but I don’t think that is necessarily the case. There could be (and likely are) larger scale processes at work.

The authors utilize upward and downward-looking irradiance measurements that were obtained from instruments tethered to balloons to infer radiative heating and cooling rate profiles. The measurements cover multiple Arctic locations and different seasons. They find characteristic differences for clear and cloudy scenes, largely in line with radiative transfer calculations, and highlight the role of multi-layer clouds to suppress cloud-top cooling in the lower layers. The authors explore the role of cloud micro- and macro-physical properties for this suppression.

The paper is well written, and the figures complement the text adequately. I suggest one major and several minor modifications before publication.

Major concerns

The paper lacks a discussion section. There are a few points the authors touch on and additional ones that come to mind. I suggest writing a discussion section preceding the summary that includes, for example, the below points:

• I like the brief discussion (ll. 326-331) on the dynamical impacts of cloud-top cooling (or the lack thereof). Perhaps the authors could expand on it.
• Are there other constituents the authors haven’t explore in this paper? I could image lofted aerosol as well as lofted moist air could impact boundary layer heating and cooling rates.
• Based on the profiles used here, can the authors claim full understanding of Arctic radiative transfer or are there remaining deficiencies in RTM simulations?

Minor concerns

ll. 38-40 This sentence needs additional information. Is free-tropospheric cooling or warming meant here?

Section 2: Please explain how aerosol was included in radiative transfer calculations and which aerosol properties were assumed.

ll. 124ff The authors should explain here how they define “cloudy”, perhaps foreshadowing the use of ground-based net radiation that is currently mentioned much later. It would be important to translate the net radiation threshold into an optical thickness threshold for reference.

Fig. 3 Please show the -30 W m^-2 as vertical (thin, gray) line in each panel.

Fig. 5 Please include the interquartile range of respective profiles, perhaps as shading.

Fig. 6 Please increase the vertical extent of this figure to better display fine changes.

ll. 233 ff Please explain whether the adiabatic assumption is realistic for Arctic clouds.

Section 4: Satellite imagery of these cases could be helpful.

Fig. 7 and 9: Please add a line marking the altitude of 0 and also -30 degree Celsius to illustrate where there are supercooled and homogeneous freezing conditions.

l. 273 The sentence suggests that the timeline shows the same airmass. Is that really the case or was there advection?

ll. 301-303 Would the inclusion of a high-altitude ice cloud in radiative transfer simulations produce plausible heating rate profiles? Perhaps include a simulation with the best guess in ice cloud properties. With respect to Section 5, the authors could use it an example of a relatively thin cloud where moderate boundary layer cloud-top cooling is permitted.

ll. 326-331 Perhaps best reserve this for a discussion (see major point). Similar to an above concern, was there advection happening? If so, please rephrase to omit the misleading impression of having sampled the same airmass twice.