Using reanalysis data and MOSAiC in situ observations, the authors present a detailed and comprehensive analysis of two Arctic atmospheric rivers (ARs) that occurred within one week in April 2020. They show that both events were associated with a cyclone–anticyclone couplet and had substantial impacts on the Arctic environment, including intense warming, sea-ice loss, and extreme precipitation. Using a Lagrangian back-trajectory approach, the manuscript further characterizes the thermodynamic evolution of these two ARs and identifies the principal moisture source regions contributing to AR-induced precipitation. Overall, the manuscript is well written and logically organized. The combined Eulerian–Lagrangian framework provides valuable mechanistic insight into Arctic AR processes. I therefore believe this manuscript is suitable for publication after minor revisions, which I outline below.

Specific comments:

Line 26: The increases in AR frequency over the Arctic are not spatially uniform, with faster increases over the Atlantic sector (Ma et al., 2024a).

Line 211: ‘Coinciding’ is not very accurate here. It seems that Q2m begins to decline sharply only after precipitation returns to ~0. Please explain why the decline in Q2m lags behind changes in precipitation.

Fig. 2a: The results shown here are consistent with the findings of Ma et al. (2024b) (see their Fig. 10). Arctic ARs are usually driven by a cyclone–anticyclone couplet. This couplet can be further divided into a high-pressure-dominant regime (associated with the Siberian AR), with a relatively weak low-pressure system to the left, and a low-pressure-dominant regime (associated with the Atlantic AR), with a relatively weak high-pressure system to the right. I think the authors should provide additional discussion in the Introduction on the flow regimes that drive Arctic AR formation. This would give readers more context for interpreting the results presented in this study.

Fig. 5b: Sea-ice concentration decreases of comparable magnitude can also be observed in the marginal sea-ice regions of the Pacific sector. However, ARs were not observed in those regions. Is it possible that the sea-ice reduction shown in Fig. 5b is simply driven by the seasonality of sea-ice melt? Please also examine the sea-ice concentration differences between 12 April and 22 April in years when no AR event influenced these regions.

Fig. 6c: What drives the precipitation peaks prior to the occurrence of ARs? Are they also driven by extreme moisture-transport events that are not captured by the Guan and Waliser (2024) AR detection algorithm? I recommend that the authors also plot IVT evolution in Fig. 1 using shaded contours and use line contours to indicate SLP anomalies. This would allow readers to assess whether these precipitation peaks are also associated with enhanced moisture transport.

Lines 267-268: Based on Fig. 6d, precipitation starts to increase and peaks after the passage of the AR.

Fig. 9 caption: Should be “moisture uptake minus the absolute magnitude of moisture loss”?

Line 346: I can see the strong moisture loss in the figure. However, I don’t see how strong upward motion is reflected in the figures. Additional explanations might be helpful here.

Fig. 5: I suggest that the authors also plot the spatial distribution of precipitation intensity. This would help readers better understand the magnitude of extreme precipitation associated with Arctic ARs.

Fig. 7: Please explain how you calculated this spatial density distribution for parcel trajectories. This can help readers better interpret the results.

Fig. 10: Since all these parcels are released in central Arctic, shouldn’t central Arctic have the highest density distribution?

Fig. 12: Please explain why there is a lack of moisture loss over the central Arctic where extreme AR precipitation occurs, and parcels are released.

Lines 364-365: If that is the case, why not track the parcels backward in time for more than 7 days?

Line 366: Should be decreasing pressure.

Lines 398–400: This is consistent with the Arctic AR trajectory-clustering results shown in Fig. 7d of Ma et al. (2025). They show that ARs that travel along the U.S. East Coast or the western Atlantic before entering the Arctic through the Atlantic pathway tend to be associated with enhanced moisture uptake from the warm waters of the Gulf Stream.

References:

Ma, W., Wang, H., Chen, G. et al. The role of interdecadal climate oscillations in driving Arctic atmospheric river trends. Nat Commun 15, 2135 (2024). https://doi.org/10.1038/s41467-024-45159-5

Ma, W., Wang, H., Chen, G., Qian, Y., Baxter, I., Huo, Y., and Seefeldt, M. W.: Wintertime extreme warming events in the high Arctic: characteristics, drivers, trends, and the role of atmospheric rivers, Atmos. Chem. Phys., 24, 4451–4472, https://doi.org/10.5194/acp-24-4451-2024, 2024.

Ma, W., Wang, H., Zhang, S., Singh, B., Qian, Y., Huo, Y., et al. (2025). Quantifying moisture sources of Arctic atmospheric rivers during the recent historical period. Journal of Geophysical Research: Atmospheres, 130, e2025JD043918. https://doi.org/10.1029/2025JD043918

General remarks:

In their manuscript title “Impact, drivers and pathways of two Arctic atmospheric rivers in April 2020”, the authors, in details, introduced two ARs penetrating into the Arctic Circle within several days in April 2020. Evidenced by reliable observations by MOSAiC and supported by extremeness analyses using ERA5 data, these two ARs were found to bear great impacts on the Arctic surface, including near surface temperature, precipitation, and sea ice. Although these two intense ARs have been investigated before, the authors complemented further in this study with revealing their thermodynamic evolution by means of backward trajectory analysis, the findings of which are novel and interesting. The manuscript is clearly written, and all the figures are well polished, but there are some of my concerns below that I would like the authors to elaborate on.

1. For the drivers of the two ARs: the authors emphasized the steering impacts of a cyclone-anticyclone couplet. Although the two high/low MSLP centers in between in Fig. 2a have also been analyzed, but it seems that their roles are secondary compared to those two highlighted in Fig. 2b-c. I was wondering whether the cyclone-anticyclone couplet is a must for intense ARs to penetrate into the Arctic. Also, why the anomalous cyclone is important for the Atlantic AR and the anomalous anticyclone for the Siberian AR? Are they random case-dependent or more like common features?
2. According to the distinct pathways, the Atlantic AR have strong impacts on Greenland and central Arctic, whereas the Siberian AR on Eurasia and also central Arctic. These are evident results that have been revealed by previous studies. I would like to see some further discussion about comparing the strength of Siberian and Atlantic AR impacts. For example, which of them could exert greater surface impacts on the central Arctic (e.g., the Atlantic AR seems to bring about more precipitation over the central Arctic in Fig. 3b)? In this case, it is difficult to isolate the impacts of the two ARs because they ended up merged together. But the simultaneous occurrence of the Atlantic and Siberian ARs would not always take place, very likely I supposed, thus the difference between them is worth being discussed.
3. I suggest, if there is any, the authors could also provide some simple characteristics of similar/comparable Atlantic or Siberian ARs in the global AR database, which could help support the findings. This is also a potential way to scope with the first two concerns.

Specific remarks:

L31-32: I don’t understand why the “observed increase …”  is in line with “projections”? Do the projections here mean future predictions or historical simulations?

L54: Since the strong near-surface winds have been mentioned here, I suggest the effect of the surface wind should also be discussed, besides the results in Fig. 6.

L87: The Atlantic AR has not shown up on 15 April within the Arctic Circle.

L99-105: Why the calculations of the reference distribution are different for different variables (MSLP, T2m, and precipitation)? It sounds a bit complicated and subjective (how about changing the length of days to include more or fewer days?). It would be better if some of them could be unified.

Figure 4b: It would be more meaningful also providing the PDF of absolute T2m, like that in Fig. 3a. Or just provide readers the climatology mean of all 7-day mean T2m, helping contextualize the anomaly magnitude.

L280: Differences “relative to their trajectory endpoints” more sound like the initial status minus the endpoint, but the authors actually mean the reverse.

L302: The orographic precipitation augmentation by the steep topography over the southeastern Greenland is a key factor for the in-situ extreme precipitation in Fig. 5a. I suggest this should be mentioned more and in advance.

Review of “Impact, drivers and pathways of two Arctic atmospheric rivers in April 2020”

This manuscript examines two intense and well-documented Arctic atmospheric river (AR) events that occurred during the MOSAiC campaign in April 2020, with a focus on their thermodynamic characteristics and regional surface impacts using ERA5 and MOSAiC observations. While these AR events have been widely studied, the authors apply a Lagrangian parcel tracking approach to further investigate their moisture sources and thermodynamic evolution along the transport pathways. Overall, the paper is well written and provides an interesting and valuable perspective on diagnosing Arctic AR events and their surface impacts. I have only a few major/minor comments and suggestions that may help improve clarity and strengthen the manuscript.

Major/clarifying comments

1. The AR identification is based on Guan and Waliser (2024), who also employ a Lagrangian tracking framework. According to their tracking methodology, the Siberian AR event and the Atlantic AR evet appear to be merged into a single AR event (if my understanding is correct). However, throughout the manuscript the two events are described as separate ARs. It would be helpful to include a brief clarification explicitly stating that, under the Guan–Waliser AR algorithm, these two ARs are identified as a single merged event, while they are analyzed separately in this study.
2. Regarding Figure 3: In April, decreases in net shortwave radiation associated with ARs can also contribute to the net surface energy budget (SEB), in addition to enhanced net longwave radiation (Zhang et al., 2025). It would be helpful to discuss the contribution of net shortwave radiation to net SEB, as net shortwave radiative appears to decrease during the two AR events shown in Figure 3e, which may help explain the slight decrease on net SEB on April 16.

In addition, prior to the arrival of the Siberian AR, net longwave radiation, 2-m temperature (T2m), precipitation, 2-m specific humidity (Q2m), and net SEB already show increases, while during the Siberian AR itself the net longwave radiation remains relatively steady. Could the authors please elaborate on the atmospheric conditions preceding the arrival of the Siberian AR that may explain these features?

I also suggest computing T2m, precipitation, Q2m, and surface energy budget components (including longwave, shortwave, sensible, and latent heat fluxes) from ERA5 and comparing them with the MOSAiC in situ observations shown in Figure 3. This comparison could help readers to appreciate a broader spatial context and allow for an assessment of the consistency between the reanalysis and the MOSAiC observations.

Minor comments:

• Lines 19–22: I recommend citing the canonical AR definition paper by Ralph et al. (2018).
• Figure 1: Please consider using two distinct colors to clearly differentiate the Siberian AR from the Atlantic AR.
• Figure 3 (title): Please specify that the in situ observations refer to measurements fromRV Polarstern.
• Line 145 (Section 2.5): Please provide additional details on the Lagrangian Analysis Tool (LAGRANTO v2.0) used in the parcel tracking analysis.
• Lines 343–345: Figure 12 shows

References:

Zhang, C., Cassano, J. J., Seefeldt, M. W., Wang, H., Ma, W., and Tung, W.: Quantifying the impacts of atmospheric rivers on the surface energy budget of the Arctic based on reanalysis, The Cryosphere, 19, 4671–4699, https://doi.org/10.5194/tc-19-4671-2025, 2025.

Guan, B. and Waliser, D. E.: A regionally refined quarter-degree global atmospheric rivers database based on ERA5, Scientific Data, 11, https://doi.org/10.1038/s41597-024-03258-4, 2024.

Ralph, F. M., Dettinger, M. C. L. D., Cairns, M. M., Galarneau, T. J., and Eylander, J.: Defining “Atmospheric river”: How the glossary of meteorology helped resolve a debate, B. Am. Meteorol. Soc., 99, 837–839, https://doi.org/10.1175/BAMS-D-17-0157.1, 2018. 

Summary

In this manuscript, the authors study two atmospheric rivers (ARs) that entered the Arctic in April 2020 during the MOSAiC field campaign. They analyze the large-scale circulation features that steered the two ARs from the Atlantic and Eurasia, respectively, into the Arctic, use ERA5 and MOSAiC data to study the impacts of the ARs on near-surface meteorology and sea ice, and model Lagrangian air parcel trajectories to study the evolution of air masses associated with the ARs. They find that diverse moisture sources contributed to the ARs and that a subset of the modeled air parcels exhibit AR characteristics, with decreasing temperature and increasing potential temperature as they ascend during their arrival at the locations of maximum precipitation.

The paper is generally well written, with sound methodology and thorough references. The trajectory analysis in particular provides some novel insights into the moisture sources and dynamics of Arctic ARs. The figures are mostly of high quality and straightforward to interpret, though I suggest some tweaks to improve the figures. I have a number of other minor comments related to the moisture source analysis, climatological context for the AR events, and various other topics, as detailed in the comments below. Provided these comments are addressed, I feel this paper will provide valuable advances to the scientific understanding of AR drivers and impacts in the Arctic.

Minor overview comments

- Some refinement and clarification about the climatological context for these AR events is needed. It appears that the claims of an "unprecedented event" (L84) and "unprecedented transport of heat and moisture into the region" (L384) are based on the previous analysis of Rinke et al. (2021), is that correct? If so, this should be stated clearly, and some further context for what was unprecedented about this event should be provided - i.e. it appears that Rinke et al. (2021) found that there were daily record highs of temperature, TCWV, and longwave radiation at the MOSAiC site relative to the ERA5 1979–2019 climatology on some of these dates? If so, while these daily record highs do confirm that this was an extreme event, I think that a wider net (e.g. monthly or seasonal climatology) would need to be cast to claim that this event was without precedent in the historical record.

- On a similar point, the different climatological reference periods for the statistical calculations in Section 2.1 are a bit hard to follow. For example, why are there two sets of T2m anomalies (daily anomalies from the April mean climatology, and 7-day centered mean anomalies)?

- The moisture sources for the "Siberian" AR need to be better explained and contextualized. The authors suggest in L410–419 that the 2020 Siberian heat wave provided a significant moisture reservoir for the Siberian AR. However, Fig. 12 shows that the majority of moisture uptake occurred further west over southern and central Europe. Further, the two references cited in L413 do not seem to directly support a Siberian moisture source for this AR – Kwon et al. (2021) is about the impacts of this heat wave on CO2 uptake, while Fig. 3 in Gloege et al. (2021) does show some positive soil moisture anomalies over Siberia, but these appear to be spatially displaced from the area of moisture uptake in Fig. 12 of this paper. Do the authors contend that evapotranspiration from land areas can constitute a significant moisture source for ARs during the spring, and is there support for this in the literature? Or is there some other process in the atmosphere that results in the moisture uptake over Eurasia in Fig. 12? Relatedly, it's not clear to me how the conclusion is reached that "a portion of the moisture reaching the central Arctic is not newly acquired over Eurasia, but carried from more distant regions over a period exceeding seven days". Could the authors spell out in more detail how this conclusion follows from the observations in the previous sentences?

- It appears that rainfall is quantified using ERA5 data in this study (L93). Although the partitioning of rain vs snow is not a central focus of the work, the uncertainties in quantifying precipitation amount and phase using reanalysis data should be acknowledged. For example, L257 states that "a mix of rain and snow is observed" and L267 states that "intermittent rainfall events are measured" - since this is reanalysis data, the text should be clear that rainfall is simulated rather than observed or measured. Can the authors provide any references to show that ERA5 rainfall data are reliable in the Arctic?

- Title and abstract: I wonder if it's worth mentioning in the title and abstract that the two ARs being studied took place during the MOSAiC field campaign (i.e. maybe the title could be something like "Impact, drivers, and pathways of two Arctic atmospheric rivers during MOSAiC in April 2020"). Upon first reading the paper, I was a bit confused on why these two ARs in April 2020 were being studied, and the connection to the MOSAiC campaign with its rich observational dataset would make the context and motivation of the study more clear to readers. However, this isn't a strong opinion and I will leave it to the authors' discretion as to whether they accept this suggestion.

- Fig. 1 and elsewhere: Maybe nitpicking, but I'm not sure if "Siberian" is the best description of the second AR. From Fig. 1 it appears that the AR spent much of its life cycle over central and eastern Europe before entering the Arctic from western Siberia. Perhaps "Eurasian" would be a more accurate label?

Minor specific comments

- L12: The way the abstract reads right now, "one group of parcels" is described without any mention of the other parcel groups. I understand after reading the rest of the paper that this is the group of "AR-like" air parcels that is singled out for more detailed analysis, but it would be helpful to try and make this a little more clear in the abstract. I suggest a rewrite to something like "During both AR events, there is a group of parcels that experiences overall cooling and increases in potential temperature associated with classic AR characteristics: ..."

- L19: Atmospheric rivers are typically defined according to strong water vapor *transport*, not just enhanced presence of water vapor. See the definition in the American Meteorological Society Glossary of Meteorology: https://glossary.ametsoc.org/wiki/Atmospheric_river

- L36–46: Another useful reference on the physical mechanisms that induce Arctic warming (in this case, specific to Greenland) associated with AR-like air streams, and analyzed from a Lagrangian perspective, is Hermann et al. (2020).

- L65: This is a bit of an abrupt transition to discussing the two April 2020 AR events. Some transition language about why these ARs are being studied would be helpful. Maybe the introductory text from the succeeding paragraph could be pulled forward to introduce this case study, i.e. something like "Previous studies have examined a sequence of two strong ARs that impacted Arctic sea ice during the MOSAiC campaign during April 2020..."

- L173: "within a few days"... of what? Within a few days of entering the Arctic, or a few days after their origin as AR features?

- L184: How are "residual AR airmasses" defined? I do not see anywhere these residual AR airmasses are represented in the figures.

- Fig. 3: This figure would be easier to interpret if there were minor ticks on the x-axis at the start of each day. I also suggest locating the major x-axis ticks and labels every 5 days, i.e. at 5 Apr, 10 Apr, etc.

- L210–211: It looks to me like the spike in precipitation is simultaneous, or even slightly leads, the peak in Q2m...?

- L218: From Fig. 3b, it looks like the peak in precipitation is more like 1.5 times greater for the second AR? (not 2.5 times)

- L220–221: I agree that turbulent fluxes appear to contribute only marginally during the positive-SEB part of the second AR event, although it is a bit hard to say for sure because of the different shapes of panels (d) and (e) in Fig. 3 and because of the missing SEB data for part of the event in Fig. 3d. But what about the strong positive SHF during the latter stages of this event? Do the authors have any hypothesis or explanation about what causes this (multi-day?) spike in SHF?

- Fig. 4: Why is the color scale for the T2m anomaly map (panel a) not centered on 0C?

- Fig. 5: The precipitation percentile color scale seems somewhat odd. The linearly increasing color scale does not match the somewhat arbitrarily-chosen percentile levels. I suggest finding a color scale to more effectively display this map data.

- L265: It appears to me (from Fig. 6b) that the temperature over the BKS doesn't actually reach 0C by 18 April, more like -1C or -2C?

- L267–268, Fig. 6: "During the ARs, precipitation increases substantially" - it appears to me that the increase in precipitation over BKS occurs after the gray-shaded AR period in Fig. 6d?

- L268–270: It's not quite clear to me from Fig. 6 how there is a strong correlation between SIC and rainfall & T2m. Are these correlations calculated at the daily scale, such that the change in SIC is correlated with the daily rainfall and temperature? Could the authors provide more details about how these correlations are calculated?

- L322–323: How are the air parcels along these "pT..." trajectories interacting with the surface if they are mostly at pressures < 650 hPa (and not located over elevated terrain)? Is there some other process in the free troposphere that can explain the specific humidity changes in these parcels?

- L326–327: I like the decision to focus on a subset of "AR" air parcels chosen using simple physical criteria, rather than all air parcels as a composite. This is an interesting and effective methodological choice.

- L339–341: Is there anything that can be concluded from the fact that the percentage breakdown of the air parcel grouping is so different for the Arctic AR comparing to the Greenland-landfalling AR?

- L343–344: It's really neat to see the distinct pathways of the two ARs in the trajectory density map!

- L434–435: The Komatsu et al. (2018) paper cited here does not show that Siberian ARs are becoming more frequently occurring features of the Arctic climate system. The only mention of increasing trends in this paper are references to other papers that have analyzed trends in water vapor and precipitation over the Arctic and Eurasia.

Technical corrections

- L14: Suggest commas around "however"

- L70: The word "the" is duplicated in this sentence

- L71: "from ship" --> "from the ship"

- L80 and Fig. 2 caption: Find a more grammatically appropriate description than "extremeness", such as "extreme nature" or "statistical context"

- L142: "its life cycle" --> "their life cycles"

- L157: "5a" --> "Fig. 5a"

- L177–178: "along the Atlantic Ocean" is awkward phrasing, please rephrase

- L226: "After having" --> "Having"

- Fig. 5 caption: The word "gridded" is unnecessary (it is assumed that ERA5 reanalysis data are on a regular lat/lon grid)

References

- Hermann, M., Papritz, L., & Wernli, H. (2020). A Lagrangian analysis of the dynamical and thermodynamic drivers of Greenland melt events during 1979–2017. _Weather and Climate Dynamics_, _1_(2), 497–518. https://doi.org/10.5194/wcd-1-497-2020