Summary

In this paper, Zhang and coauthors present a modified Atmospheric River Scale tailored to analyze AR events in the polar regions. They first describe the justification for a polar AR scale by comparing the integrated water vapor transport (IVT) climatology in polar regions to the mid-latitudes, finding that most ARs impacting Greenland and Antarctica would go undetected by the original AR scale due to the colder and drier conditions in the polar regions. They introduce a modified version of the AR scale that includes three new "polar" categories with lower IVT thresholds, then use this scale to analyze the frequency, seasonality, and interannual trends in polar ARs. They assess the precipitation and melt impacts of the ARs identified by the new scale, finding that weak and moderate ARs account for most AR-related precipitation in Greenland and Antarctica, while stronger ARs are infrequent but cause extreme precipitation when they occur. Finally, they describe a web product that provides polar AR forecasts in real time.

The paper is well-organized and well-written overall. Most figures are clear and the references are extensive and appropriate. In my assessment the polar AR scale will be highly useful to a broad range of users, and I am impressed by how much information about ARs in Greenland and Antarctica can be captured by this simple but well-designed scale. However, I think there may be a serious error in the ice sheet melt analysis that should be addressed before the paper can be published, as described in my major comment below. I also have a number of minor comments and technical corrections, as described below.

Major comments

(1) I strongly suspect there is some error in the analysis of surface melt in Fig. 13. Numerous studies have documented an increase in Greenland Ice Sheet melt during the 21st century, but Fig. 13b shows more summer days with surface melt even in the 1980s compared with the 2000s. This is almost certainly incorrect. For a quick check, compare the years 1993 and 2012 using the Greenland Surface Melt Extent Interactive Chart at the National Snow and Ice Data Center Ice Sheets Today page (https://nsidc.org/ice-sheets-today/melt-data-tools). The NSIDC chart shows a much more extensive Greenland melt area throughout virtually all of JJA in 2012 compared to 1993, but Fig. 11b taken literally shows that 1993 was the most extensive melt year along the Greenland coast and had over 3 times the melt days of 2012. This figure also doesn't line up with Fig. 11b which shows increasing AR frequency along the Greenland coastline.

I am less familiar with melt trends in Antarctica so I am not sure if Figs. 13a and 13c contain obvious errors.

The authors should check their analysis and the underlying microwave melt dataset in both Greenland and Antarctica for potential errors in processing.

Minor comments

(1) General comment: Have the authors thought about adjustments to the polar AR scale that may need needed in future warming scenarios? Will the AR scale remain constant in climate change scenarios, despite the projected increases in IVT in the polar regions? Any new analysis on this topic is likely outside the scope of this manuscript, but it could be a nice addition to the paper to discuss this as a topic for future research in the conclusions section.

(2) Title: I suggest removing "CW3E" from the title, or at least stating the full name of the Center for Western Weather and Water Extremes in the title. The abbreviation "CW3E" will not be familiar to many in the cryospheric science readership of this journal. I also note that the Polar AR Scale is described as a collaborative effort between CW3E and the Byrd Center at Ohio State (L523–526), and the paper introducing the original scale (Ralph et al., 2019) does not describe it as the "CW3E Atmospheric River Scale".

(3) L44: I suggest framing this sentence along the lines of "Our results show that that the polar AR scale better characterizes the strength and impacts of ARs in the Antarctic and Arctic regions than the original AR scale, and has the potential..."

(4) L170–171: Why were 1-degree ERA5 data used instead of the finer native resolution of ERA5? Do the authors expect that this has any influence on their results? I note that the original AR Scale in Ralph et al. (2019) used 0.5-degree gridded data.

(5) L218–228: This is a nice analysis of the climatology of IVT in the parts of Greenland and Antarctica that extend outside of the polar latitudes.

(6) L286–287, 319–321: Out of curiosity, do the authors know how many AR4 events there are in the historical record in Antarctica? I see in L456–457 that no AR5 events have ever been recorded in Antarctica, but it would be nice to state the number of AR4 events here to provide historical context for the March 2022 event. Would it be straightforward for the authors to include a map of the maximum AR category ever reached in the historical record at the Antarctic coastline points shown in Fig. 6d?

(7) Figure 6d and elsewhere: How / why were the locations of the these points along the Antarctic coastline chosen to calculate AR scale data? Are they selected to be useful for particular communities, such as Antarctic research stations?

(8) Fig. 7: To help interpret these maps, it would help to add a few solid contours with contour labels. Perhaps the contours of 1, 5, and 10 average annual ARs could be labeled.

(9) Fig. 8: Why are there more AR 2 events (panel e) in this "Atlantic Arctic gateway" region than ARs in the weaker AR P1 through AR 1 categories (panels a–d)? Is this correct?

(10) L385–389: Nice analysis of the seasonality of Greenland ARs. This is an interesting result and Fig. 10 is an interesting figure.

(11) L445–456: How / why was this 12-hour window chosen to define AR-associated precipitation? Is there precedent for this method in the literature? I have not performed an extensive literature review but I note that Maclennan et al. (2022) defined AR-associated precipitation in Antarctica using precipitation from the time of the AR + the following 24 hours.

(12) L484–485: This delay of 18–24 hours found by Mattingly et al. (2023) applies specifically to the delay between AR landfall in northwest Greenland and melt in northeast Greenland due to the foehn effect, not generally to all Greenland ARs.

(13) L546–548: Are there any plans to extend the CW3E polar AR scale forecasts to the Arctic, and to Greenland in particular? I could envision it being highly useful to the scientific and public communities in Greenland.

Technical corrections

- L35: application --> applications

- L36: "the intensity"... of what? IVT?

- L38 and elsewhere (e.g. L568): Find a better word than "insufficient" to describe the unsuitability of the standard AR scale. I suggest "unsuitable". "Insufficient" implies that the scale does not reach high enough IVT values to characterize polar ARs, but the opposite is actually the case.

- L43: Antarctic --> Antarctica

- L46: "observation, research, and forecasts" – this list is a grammatically incorrect mixture of singular and plural verbs. Please revise.

- L71: "the diabatic process" --> "diabatic heating"?

- L75: "the polar ice" --> "the polar cryosphere"

- L116: starts --> start

- Fig. 1 caption: Labels b and c don't match the figure panels. They refer to panels c and b in the figure.

- L158: its --> their

- L158, 525: The abbreviation "CW3E" is defined in multiple places in the manuscript.

- L170 and elsewhere (e.g. L177, L180): "data was" --> "data were". (The word "data" is a plural noun. Please check this throughout the manuscript.)

- L174: The abbreviation "EA" is not defined anywhere in the manuscript.

- L191, L196: The phrases "southern hemisphere" and "northern hemisphere" are not capitalized in this paragraph, but "Southern Hemisphere" and "Northern Hemisphere" are capitalized elsewhere in the manuscript (e.g. L204–205, L397). Please be consist with capitalization.

- L199: A space is needed before the opening parenthesis in "(Fig. 3b)".

- L204: The caption states that the maps show the Southern and Northern Hemisphere, but technically the maps only show the mid- and high-latitude areas of each hemisphere.

- L212: percentages --> percentiles

- L253: What are "variant" meteorological conditions? Please rephrase.

- L330: Rather than "the gap between Greenland and Northern Europe", a more specific term that is often used to describe this region in the atmospheric and marine science literature is the "Atlantic gateway to the Arctic", or it could also be described as the "Nordic Seas".

- L426: increase --> increasing

- L555: An open parenthesis is missing before the word "colored"

- L603: was --> were

- L608: illustrating --> illustrative

SUMMARY

This paper presents a novel tool for identifying polar atmospheric river (AR) events and their intensities. The methodology builds on Ralph et al. (2019)’s mid-latitude AR scale by including three new categories for polar ARs, which account for lower amounts of atmospheric moisture at higher latitudes. The authors present the methodology and results from the new scale, with analysis of the frequency of ARs of the different rankings in each polar region, and then discuss the precipitation and surface melt in Greenland and Antarctica attributed to ARs detected by this scale. Furthermore, the authors present a new tool for forecasting ARs and AR intensity in Antarctica. This represents a significant step forward in the development of a wider range of tools to detect and analyze polar ARs, and because this tool is Eulerian (i.e., the AR ranking is on a point-by-point basis), it allows for novel interpretations of the frequency and likelihood (risk) of ARs of different intensities.

The paper is generally well-organized and written in a clear and effective manner. However, I have several major comments below, and overall, my primary concern is that the results and analyses presented in this paper are not well grounded in the results from previous studies. There is little comparison of the AR frequency and precipitation and melt impacts found here to prior studies of these features (using different AR detection algorithms) in both Greenland and Antarctica. I think this paper could be significantly improved by providing more context and comparison to the results from previous studies in both polar regions, regarding both the method for detection and the impacts attribution.

MAJOR COMMENTS

(1) The use of integrated vapor transport (IVT) versus the meridional component of integrated vapor transport (vIVT) for AR detection in the polar regions – Figure 1 in Shields et al., 2022 (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2022GL099577) shows a comparison between the Wille et al., 2021 AR detection algorithm, which is based on vIVT, and a number of standard global AR detection algorithms, most of which use IVT for AR detection. The figure highlights that IVT-based detection methods struggle to capture Antarctic ARs, particularly in the interior of the ice sheet. I think it would be incredibly important to assess the limitations of using one or the other (IVT or vIVT) in detecting ARs in the polar regions in this paper, given the results from previous studies like Wille et al. 2021. For example, in the context of this result presented in Shields et al. 2022, what does a frequency map of ARs detected by this new AR scale look like compared to the Wille et al. 2021 algorithm (as in, a frequency difference map)? Or alternatively, how well does the AR scale capture known AR events, for example those presented in Gorodetskaya et al. (2014)?

(2) Description of how to determine the rank of an AR event – regarding the ranking of AR events in the polar regions (as described on P11 L246 to L268), did the authors consider adjusting the duration requirements for ARs as well? In the Wille et al. (2021) AR detection algorithm, there are frequently ARs that appear to make landfall for only a few hours, less than the amount of time required to meet the 24-hour qualification presented here. I found the discussion on the IVT climatology analysis and the choice of IVT thresholds for the polar scale quite interesting, but I am wondering if you also examined the sensitivity of the AR scale detection method to the time period requirements to meet certain rankings?

(3) Also regarding the AR scale description (now P12 L276 to 290), I’m not sure that the March 2022 AR-heatwave event is optimal in showcasing the capabilities of the new polar AR scale, since it ranks as an AR4 on the midlatitude scale. While this was a standout event, and it’s interesting to know how it ranks, I think it could be helpful to provide greater detail on known events that rank between AR P1 and ARP3, given that this will be the most relevant application for this scale. I am aware that an example of this was mentioned earlier in the text in Figure 2, as well as a brief description from L291 to L294 on P13. I would strongly encourage the authors to provide details on ranking these types of events in the level of detail presented for the March 2022 event.  

(4) Seasonality of polar ARs – I would ask that the authors please include the statistical significance of the results on AR seasonality on P17 – P18. Regarding the comparison of the AR scale results with the Wille et al. (2021) findings on AR seasonality, the authors suggest that the Wille algorithm has a higher frequency of ARs in winter months because the vIVT thresholding method (instead of IVT) is more closely tied to extratropical cyclones. My interpretation of the seasonality difference is that the Wille algorithm uses a threshold for vIVT that accounts for the seasonality of vIVT, where the vIVT threshold is higher in summer, due to higher atmospheric temperatures and increased moisture in the atmosphere, and lower in the winter, when conditions are drier. Comparatively, the polar AR scale uses an absolute threshold for ARs regardless of the season. Because of this absolute threshold, I would certainty expect the polar AR scale to detect more ARs in summer than in winter, just based on seasonal differences in the amount of atmospheric moisture / IVT / vIVT. To me, this doesn’t suggest that there can be a conclusion formed about which detection method is more or less affected by seasonality in extratropical cyclones. I would be interested to hear what the authors think about this and how it relates to the seasonality analysis. (and as a sidenote, can the authors please also provide a citation for “is closely related to the occurrence of extratropical cyclones, which are more active during JJA in the Southern Hemisphere.”)

(5) Precipitation analysis for polar ARs – could the authors please describe what the basis is for using the 12 hour before and after window for AR-attributed precipitation (mentioned on P20 L444)? I would be interested to know if the authors used a method to determine when AR precipitation tends to fall with respect to the timing of landfall (how long before, during, and after), given that this really impacts how much precipitation we attribute to ARs and their relative importance in contributing to the surface mass balance of the Greenland and Antarctic ice sheets. Or alternatively, is this window based on a previous study of AR precipitation in polar regions?

(6) Surface melting time series in Figure 13 – I am surprised to see a relatively low number of melt days in 2007 in Greenland compared to the 1990s, when we know 2007 was a record melt year for Greenland (Mote et al. 2007 - https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2007GL031976). Similarly, I would expect anomalously high melt in Antarctica for the 2019/2020 austral summer. Could the authors double-check the melt analysis presented in this figure, or compare the melt observations used with another melt dataset? I am not an expert on surface melting, but the Greenland melt time series especially does not necessarily look how I would expect it to. Also, when describing the satellite-observed melt in the methods, both papers cited (Picard and Fily, 2006, and Torinesi et al., 2003) are Antarctic – is there a study you can cite that applies these observations to Greenland as well?

(7) Role of IVT intensity in relation to AR impacts (as described on P27 L593) – “but including an objective description of AR strength can improve the understanding of ARs’ impacts on polar regions”. Baiman et al., 2023 (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2022JD037859) recently showed that the strength of the AR is not necessarily correlated with precipitation impacts, focusing on the region of Dronning Maud Land in Antarctica. The study found that one of the more important factors in determining AR precipitation intensity is having a mechanism for lift to produce the precipitation. I think it would be valuable to mention this in the discussion section and how it relates to the AR ranking – precipitation impacts you found in this study.

MINOR COMMENTS

Title – I’m not sure how many cryospheric scientists are familiar with CW3E (the acronym), so I would recommend spelling it out in the title or removing it

P3 L75 – “notable influence on the polar ice” is not the clearest word choice/descriptor for ARs impacts on surface mass balance and ice shelf stability

P3 L76 to L81 – I find it misleading that this section on AR impacts begins with “hot spells and heatwaves” and only includes “as well as intense snow accumulation” at the end. Many of the studies cited in this section have shown that by far the dominant impact of ARs is snowfall, especially in Antarctica. I would recommend that the authors frame the impacts description to reflect the relative importance of each AR impact in the present – though of course with the caveat that this might change in a warming climate (as mentioned on P4 L109).

P4 L89 – “ARs can also interact with other weather systems” – by interact, do you mean compound? Or as AR-extratropical cyclone systems?

P4 L90 to L107 – this reads more like a detailed summary of the heatwave than an introductory paragraph – consider condensing or tie more directly to the motivation for this study

P4 L107 – “Under a warming climate, the extreme ARs are expected to increase in both frequency and intensity” – I think the studies cited here all refer to the midlatitudes, is that correct? If so, I would recommend mentioning that in the sentence, or alternatively looking for polar/Arctic-specific studies to cite (I don’t know that this has been done for ARs in the Antarctic….).

P5 L126 – “using flexible thresholds” – it would be interesting and helpful here to have one or two sentences that elaborate on what a flexible threshold means (as in, a percentile of vIVT or IVT relative to the climatology, etc.), as well as why a different (lower) threshold is needed for polar ARs.

P5 L129 – “on-the-ground applications and communications”. I’m not sure what “communications” refers to here – as in, communicating to meteorologists and fieldworkers at weather stations to collect observations during the AR period?

Figure 2 – I would strongly recommend labelling the Ross Ice Shelf on the Antarctic maps, since scientists less familiar with the geography of the Antarctic continent may not know where it is. Similarly, it may be helpful to label East Greenland as well.

P7 L170 – what was the motivation for decreasing the resolution of the ERA5 data from 0.25 x 0.25 deg at hourly resolution to 1 x 1 deg at 6-hourly resolution? (is there a reason not to use the data with higher resolution?)

P9 L208 – I would specify that you are including Antarctic ice shelves in the coastline

P9 L212 – “percentages of IVT” as in percentages of IVT “values”?

P9 L218 – I find it slightly surprising that the authors decided to exclude southern Greenland from the analysis, given that surface melting, an impact highlighted by in the introduction, is prominent in this region. From the ice sheet and sea level rise perspective, it seems valuable to include the whole of Greenland in this study, especially since the new AR scale still includes the midlatitude scale for AR intensity.

P14 L306 to L312 – the distance covered by the Antarctic coastline is huge, so I would recommend being more specific in the locations described here (i.e., instead of “most of the East Antarctic coast”, mention the names of specific regions).

P14 L311 – I’m confused by this sentence: “… there are more AR P1 events over the inland area close to the coast compared to a similar area in East Antarctica.”

P14 L314 to L319 – the analysis jumps from AR P1 to AR1 rankings – since AR P2 and P3 are the other two new categories, could you list the statistics for these events here too?

Figures 7 and 8: the red-white color map used in this figure makes it extremely difficult to discern any difference between AR frequencies from AR P1 to AR P3, and the colors are very washed out for AR3 and AR4. Can you try using a different color map for these figures that better highlights spatial differences among the panels? Also, please write out the full figure caption for Figure 8 instead of referring to Figure 7.

P16 L354 – “0.011 events” and “0.001” are numbers that I find slightly difficult to interpret in a physically meaningful way – would it be possible to list the number of AR3 and AR4 events that occurred in parentheses?

P19 L408 – “along Greenland” --> “along the Greenland coast?”

P19 L424 – “coast of East Greenland” – this does not include southern Greenland, right? If so, would recommend saying “central-north East Greenland coast”

P21 L447 – what is the standard deviation in annual AR precipitation?

P21 L466 to L469 – nice summary!

P24 L522 – is there a citation you can include for YOPP-SH?

P25 L527 – I am missing a methods/data description with respect to GEFS, which is introduced here. It sounds like this might be described more in Bromwich et al. 2024 (is this in review?), but I think it would be highly relevant to include more information on the reliability of the GEFS in capturing the intensity, extent, and duration of Antarctic ARs.

P26 L565 – you introduced the AR acronym quite a bit earlier in the paper

P27 L587 – “related to” as in “associated with”?

P28 L613 – “… enhance situational awareness, contributing to timely preparedness and effective decision-making for high impact events…” I’m not sure who this is referring to – citizens, fieldworkers, meteorologists who can launch weather balloons during the events? I’m not aware of structures outside of the YOPP-SH campaign (and maybe the research stations that need to keep fieldworkers safe?) that employ decision making strategies for polar AR events, so I would welcome more specificity/clarity here on what this sentence means.