## General comments

Schmid and coauthors examine the atmospheric circulation patterns linked to ice speed-up events at the Russell Glacier in southwest Greenland. They employ a mix of atmospheric and glaciological data, including weather station observations, atmospheric reanalysis data, regional climate model output, satellite records of supraglacial lakes, and in situ observations of ice velocity derived from GPS observations on the Russell Glacier during 2009–2012. The authors find that the majority of ice speed-up events are related to short-term increases in ice sheet surface melt that overwhelm the subglacial drainage system, with ice sheet melt contributing more than rainfall to runoff production during most of these events. The melt-induced speedups are linked to three distinct types of regional atmospheric circulation patterns, with the most intense melt and glacier speed-ups associated with strong moisture transport to southwest Greenland by atmospheric rivers. Less intense melt and speed-up events can also occur due to (a) anticyclonic blocking patterns over southwest Greenland and (b) downslope warming in southwest Greenland induced by a cyclone off the southern or southeast coast of Greenland.

In my opinion this is an excellent paper and I enjoyed reviewing it. There are a number of interesting results that will be of interest to the polar science community, and a particular strength of the paper is its detailed and novel synthesis of atmospheric and glaciological analyses that convincingly show the effect of specific atmospheric conditions (i.e. ARs) on glacier velocity speed-up events. I have a fair number of specific comments and technical corrections that are mostly aimed at refining the presentation and situating the work in the context of previous studies. Provided these comments are addressed I feel that this paper will be an excellent contribution to the literature.

## Specific comments

L9–10 (abstract): Are the 16 lake drainage events and 4 extreme rainfall events a *subset* of the 36 melt-induced speedup events? Or is there no overlap between these categories? I believe this is answered later in the paper e.g. in L300–304, but I found this information to be a little confusing as it is presented in the abstract.

L35–50: Nice explanation of the relationship between meltwater drainage and ice dynamics. This helps contextualize the "event-type" accelerations examined in this study.

L77–84: I think the authors should provide a more detailed and specific set of research objectives and questions in this last paragraph of the introduction. What specific questions about the current atmospheric drivers of speed-up events did the authors set out to answer? *Why* did the authors apply a Lagrangian trajectory model to analyse 5-day backward trajectories?

L84 (Fig. 1): Nice figure that does a good job of showing the regional and local setting for the study. I suggest the authors consider adding a white shaded area showing the coverage of the ice sheet in Greenland on the zoomed-out map.

L110–130: What is the difference between the melt values calculated from the IMAU vs PROMICE stations? Are only the station locations different, or is there also a different methodology applied to the station observations to calculate melt for each of the two networks?

L152–159: Can the the authors provide a reference for the blocking criteria and/or algorithm that was used to identify atmospheric blocking? Or is it an original methodology developed for this study (please state this if so)? Same for the cyclone identification and the object-tracking algorithms mentioned in this section – were these developed by the authors or adapted from prior studies?

L216–222: I find Figure 4 to be somewhat difficult to interpret... are there very few markers for lag days 2 and 3 because most of the cross-correlations for these days are < 0.5?

L305–309: I found it somewhat difficult to follow this discussion because of the differing temporal characteristics of MI and speed-up events, i.e. MI-days are only allowed to be a single day, while the ice speed-ups can extend over multiple days. For example, it took me a few reads to understand why the 8% (MI-day occurs 1 day before the *onset* of a multi-day speed-up event) and 47% (MI-day occurs 1 day before the *day of largest velocity increase* during a multi-day speed-up event) numbers are different. It may help to reiterate in this section that MI-days are single days, while ice speed-ups have durations of 1–8 days.

L334: Despite not being officially categorized in the C_L cluster, node 1 shares many of the same characteristics of nodes 14–15 and 19–20, namely a cyclone off the southern coast of Greenland and an IVT plume directed toward southeast Greenland. Its placement in the SOM space suggests that it's probably a hybrid or transition node between the C_L and the C_AR cluster, and the IVT directed toward southeast Greenland will likely lead to downsloping / foehn induced melt in southwest Greenland, as mentioned elsewhere in the paper.

L338–340 (Figure 7): I suggest the authors consider plotting mean 500 hPa wind vector arrows or barbs on the IVT / Z500 maps, in order to show the wind flow patterns in each cluster. This would help link this analysis more directly with the back trajectory results.

L356–359: MAR generally estimates lower rainfall rates because of the spatial mismatch between the MAR and DMI data, correct? (DMI station is located at low elevations below the ice sheet, while MAR grid cells are at higher elevations on the ice sheet.) It would be helpful to reiterate this here.

L363–365: Why did the authors choose to show the 750 hPa trajectories in the main paper and not the near-surface trajectories? It would be good to give a short explanation of this decision here. There are some interesting features in the near-surface trajectories in the supplement, such as the pronounced signal of downsloping-related strong warming and drying of the C_L cluster trajectories as they approach their end point.

L422 (Figure 9): It's interesting the SWnet is even higher in the C_L cluster than the C_H cluster. Is this evidence of strong foehn clearance as the air flows over southern Greenland? See next comment.

L429–432: The foehn-like mechanism in SW Greenland during the C_L cluster events is an interesting and novel result. I think foehn should be discussed earlier in the paper, i.e. in L384–390 where the authors describe the downsloping during C_L events but don't mention foehn. It would also be helpful to place this discussion in the context of previous work on foehn / downsloping in Greenland (e.g. Noël et al. 2019, Cullather et al. 2020, Hahn et al. 2020, Mattingly et al. 2020, Ward et al. 2020) and the Antarctic peninsula (e.g. Turton et al. 2018, Wille et al. 2019, Elvidge et al. 2020, Laffin et al. 2021). Given that southern Greenland is a relatively narrow plateau with steep topography descending to sea level on each side, there may be similar mechanisms at work here to what has been studied previously in the Antarctic peninsula.

L450–451: See also Neff et al. (2014).

L456–457: It is likely that many of the IVT bands that do *not* lead to extreme melting and ice speed-ups are weaker ARs associated with less intense moisture transport. See Mattingly et al. (2020) who showed a strong relationship between AR intensity and melt in Greenland.

## Technical corrections

L21: Need apostrophe in "Greenland's"

L67: One of the "most" well-studied regions?

L70: "K-transect" --> "The K-transect"

L70: "its" --> "their"

L78: No comma before "has"

L84 (Figure 1 caption): "on the GrIS" --> "in Greenland" (the overview map does not distinguish between the ice sheet and non-glaciated areas of Greenland)

L97 (and L102, L563; check elsewhere): "data is" --> "data are"

L142: "requires" --> "require"

L148: "uses hybrid" --> "uses a hybrid"

L149: "on 0.25°" --> "on a 0.25° grid"

L157: "a object tracking" --> "an object tracking"

L183: "consider" --> "considered"

L193: End sentence after "(L4–L6)" and start a new sentence with "The speed-up events..."

L206: "will" --> "with"

L236: All 3 words in the phrase "Self-organizing maps" are not capitalized here, but they are all capitalized elsewhere (e.g. L83). Be consistent with capitalization of this phrase. (I think it is generally not capitalized in other literature.)

L237: "SOMs is" --> "SOMs are"

L252 and elsewhere: No need to capitalize generic directional terms such as "southwest", "north", etc. (See also e.g. "South" and "East" in L366.)

L346: What does "shielding form the" mean in this sentence? Should this say "shielding the cyclones from arriving..."?

L385: I suggest "flowing over" or "traversing" instead of "overflowing" here.

L424: Remove comma after "Both"

L494: Reword "should not be over interpreted" - I suggest "are subject to large uncertainty"

L495: "data indicates" --> "data indicate"

L524-525: Switch the order of the first two clauses in this sentence - "Daily increases in rainfall are larger than in meltwater only during four ice speed-up events, despite..."

L529: Edit the first part of this sentence - "In addition, only a few speed-up events are not melt-induced and are linked to lake drainage events..."

L540: "from south" --> "from the south"

L540: "a blocking anticyclone"?

 

## References

Cullather, R. I., Andrews, L. C., Croteau, M. J., Digirolamo, N. E., Hall, D. K., Lim, Y., et al. (2020). Anomalous circulation in July 2019 resulting in mass loss on the Greenland Ice Sheet. *Geophysical Research Letters*, *47*(17), e2020GL087263. https://doi.org/10.1029/2020GL087263

Elvidge, A. D., Kuipers Munneke, P., King, J. C., Renfrew, I. A., & Gilbert, E. (2020). Atmospheric drivers of melt on Larsen C Ice Shelf: surface energy budget regimes and the impact of foehn. *Journal of Geophysical Research: Atmospheres*, *125*(17), e2020JD032463. https://doi.org/10.1029/2020JD032463

Hahn, L. C., Storelvmo, T., Hofer, S., Parfitt, R., & Ummenhofer, C. C. (2020). Importance of Orography for Greenland Cloud and Melt Response to Atmospheric Blocking. *Journal of Climate*, *33*(10), 4187–4206. https://doi.org/10.1175/JCLI-D-19-0527.1

Laffin, M. K., Zender, C. S., Singh, S., Van Wessem, J., Smeets, C. J. P. P., & Reijmer, C. H. (2021). Climatology and Evolution of the Antarctic Peninsula Föhn Wind‐induced Melt Regime from 1979‐2018. *Journal of Geophysical Research: Atmospheres*, *126*(4), e2020JD033682. https://doi.org/10.1029/2020JD033682

Neff, W., Compo, G. P., Martin Ralph, F., & Shupe, M. D. (2014). Continental heat anomalies and the extreme melting of the Greenland ice surface in 2012 and 1889. *Journal of Geophysical Research: Atmospheres*, *119*(11), 6520–6536. https://doi.org/10.1002/2014JD021470

Noël, B., van de Berg, W. J., Lhermitte, S., & van den Broeke, M. R. (2019). Rapid ablation zone expansion amplifies north Greenland mass loss. *Science Advances*, *5*(9), eaaw0123. https://doi.org/10.1126/sciadv.aaw0123

Turton, J. V., Kirchgaessner, A., Ross, A. N., & King, J. C. (2018). The spatial distribution and temporal variability of föhn winds over the Larsen C ice shelf, Antarctica. *Quarterly Journal of the Royal Meteorological Society*, *144*(713), 1169–1178. https://doi.org/10.1002/qj.3284

Ward, J. L., Flanner, M. G., & Dunn‐Sigouin, E. (2020). Impacts of Greenland Block Location on Clouds and Surface Energy Fluxes over the Greenland Ice Sheet. *Journal of Geophysical Research: Atmospheres*, *125*(22), e2020JD033172. https://doi.org/10.1029/2020JD033172

Wille, J. D., Favier, V., Dufour, A., Gorodetskaya, I. V., Turner, J., Agosta, C., & Codron, F. (2019). West Antarctic surface melt triggered by atmospheric rivers. *Nature Geoscience*, *12*(11), 911–916. https://doi.org/10.1038/s41561-019-0460-1

Overview:

The authors aim to fill a gap in understanding of the synoptic-scale atmospheric events driving melt-induced speed up events of the southwestern Greenland Ice Sheet. The authors focus their analysis on the Russell Glacier, a well-studied and monitored glacier in southwestern Greenland and use a combination of self-organizing maps and back trajectory analysis to characterize periods of increased ice velocity. Overall, this was a well written paper that probes at an interesting knowledge gap, and provides a nice process-driven study of atmosphere-ice interactions on Greenland. The manuscript was generally clearly presented, with informative and well-designed figures. I only have minor suggestions, mainly with respect to the discussion of the broader implications of this work.

My main comment is that the manuscript could benefit from more context on what the Lagrangian trajectory analysis provides, throughout the introduction and discussion. The air mass characteristics at the time of the melt speed up events could be assessed without knowing their history – does knowing the sources of these air masses and their trajectories provide information that would help predict increasing frequency of these melt events? Is there anything to be learned about the future frequency of these patterns and events given the dynamics revealed by the trajectory analysis? Can contextualization of these trajectories with respect to larger scale (i.e. hemispheric) circulation inform our understanding of the likelihood and predictability of melt events in the future?

It’s nice to see Table 1, and would be great to include more discussion on why the identified patterns do or do not co-occur with speed-up events. Based on this table, each of these patterns usually does not trigger a speed-up events. Is there an opportunity to learn something more from these conditional probabilities? For example, are speed-up events more likely to occur during these patterns given specific preconditions (e.g. elevated temperatures)?

A few specific minor comments:

-Line 52: ‘orographic forcing from North America’ – what does this mean?

-Lines 61-66: These are nice descriptions of possible future changes to regional circulation and a good explanation for the high-level motivation behind this work. The authors could revisit these ideas in the discussion to think through how their results contribute to this larger discussion.

-Line 132: Maybe clarify that the solid and liquid precipitation is the total precipitation referred to in the next sentence; the parenthetical sounds like the solid and liquid precip are available separately.

-Lines 200-208: These PCA results might fit better in the results section.

-Figure 5: With this x-axis, it’s hard to tell exactly how long different melt events are. Perhaps include a histogram of melt event duration?

-Some aspects included in the results would be more fitting in the discussion (e.g. the paragraph beginning line 275).

-Another brief point of discussion that could be useful to address is: how representative are the results from this study to other glaciers on (southwestern) Greenland?