General Comments

This paper analyzes spectrometer data from GOME-2, ERA5 reanalysis fields, and 10 km MAR meltwater output to evaluate the role of clouds atop the Greenland Ice Sheet (GrIS) in shaping the three most anomalous melt summers since 1979, which were 2012, 2019, and 2023 with the latter receiving the least attention to date. The authors find comparisons between these cases in terms of elevated two-metre air temperatures, melt area, and meltwater production over the ice sheet. They also find similarities between 2012 and 2023 in terms of July cloudiness linked to anomalous accumulation zone melt, though 2023 cloudiness was relatively more widespread across the ice sheet. The authors note that a colder-than-average June 2023 precluded that summer melt season from competing with 2012 for the largest melt summer on record.

Overall, the paper provides a timely comparison of monthly atmospheric circulation, clouds, and GrIS surface conditions through the progression of summers that experience notable melt. Pending some additional details about data products and considerations regarding methods (e.g., base period calculation), the paper could be a good fit for the journal and contribution to the literature on cloud-radiative effects on extreme monthly and seasonal GrIS melt. Comments are provided below by line number (L) of the submitted manuscript.

Specific Comments

L5 and L8: Negative anomalies in what?

L35: Is this role of clouds dependent on their type and occurrence within the summer season?

L50: “higher-level” data products and model results such as?

L78-80: What is the accuracy of GOME-2 retrievals? Have these measurements been validated by surface observations, especially over the GrIS?

L103-107: Why use ERA5 versus another global reanalysis? Specifically, how does ERA5 two-meter temperature compare versus other global atmospheric reanalyses over the ice sheet?

L137: Are 2012, 2019, and 2023 included in the 2007-2024 mean? If so, how do anomalies compare in Fig 2 if those years are removed? Are results impacted?

L143: The Geological Survey of Denmark and Greenland (GEUS), which manages the PROMICE weather station network, took over operations of GC-Net weather stations ~2020 (see https://promice.org/about/). It would be a good idea to confirm that more recent ERA5 datasets (2021-2024) do or do not assimilate data from one or both of these networks.

L237: “coming in second” since when and second to which year?

L284: What type(s) of clouds enhance meltwater run-off?

L295: A note here is warranted whereabouts on the GrIS this fixed elevation threshold may not accurately estimate the ELA.

Technical Corrections

L132: …and daily data forced by ERA5 reanalysis. Also, is this daily-averaged or sub-daily? Please clarify.

L272: “and go in opposite directions” – this portion of the sentence could be re-written to improve clarity.

L155: Should “all three months” be “all three weeks”?

Figure 9: The top portions of panels (a) and (b) are illegible due to the placement of the legends overlaid on the data.

L312: Would suggest adding “regional” before “model” for clarity.

This manuscript provided an informative comparison of the mechanisms that promoted melt of the Greenland Ice Sheet during the 2012, 2019, and 2023 anomalous melt seasons. The detailed examination of 2023 with respect to the other two record setting melt years, which highlights the importance of antecedent conditions in priming the ice sheet for high-volume melt later in the melt season represents a valuable contribution to the literature. It is my assessment that the analysis could be suitable for publication pending revisions to the manuscript.

General comments

The analysis would be more effectively communicated if the authors devoted some text earlier in the manuscript to briefly explain the significance of VIS-NIR reflectance to GrIS mass balance. In other words, what relevant physical processes does it inform on and how?

Related to my previous comment, the framing of the analysis in the introduction seems a bit inconsistent with the discussion. Understanding the contrasting mechanisms that promote melt of the ice sheet – namely, clear-sky vs cloud longwave forcing – is cited in the introduction as a primary motivation for the study, but the authors cite prior work in section 4.1 (L201) establishing that the primary measurement they utilize to conduct their analysis, , is not an effective indicator of cloud cover over the ice sheet. If, as suggested in section 4,  was not intended to be used to monitor cloud cover, that could be clarified when addressing my first comment above.

While the maps in Figs 2-4 show anomalies with respect to the long-term mean, many of the statistical relationships between  and other meteorological and SMB/SEB variables (Figs 5 and 6) appear to be derived using mean values of these fields. This means that the provided correlation coefficients may be strongly reflective of any shared seasonal cycles in these variables rather than a true physical link. To avoid this, the seasonal cycle should be removed from each variable before computing the correlation.

Specific comments

Abstract: Including a sentence briefly explaining the physical significance/relevance of to the ice sheet would make the abstract more accessible to a wider audience.

L9-10: Consider whether the use of parenthetical formatting is necessary here. I personally feel that the terms in parenthesis are clearer and would read well on their own; i.e., “… both clear-sky conditions, observed in 2019, and cloudy conditions, observed…”.

L20: replace “the melting” with “melt”

L27: delete “as a whole”

L41: Consider different, more descriptive phrasing to open this sentence; e.g., “Although it produced less melt than that of 2012, the summer of 2023…”

L48: The second half of this sentence, “…for the period of interest as well 2012 and 2019 for reference,” is not needed here.

L63: The parentheses are not needed here

L78: The phrasing here makes it sound like your results include band 3 imagery. But the next sentence suggests that band 3 was not utilized. If the latter is true, consider rephrasing to something like, “While band 3 (437-610 nm) includes part of the visible spectrum, band 4 better captures…”

Fig. 2 and 3: The authors describe this in the main text, but I would suggest making the difference between these two figures clearer up front in the captions. Perhaps with a short, descriptive title at the start of each caption that notes that Fig. 2 shows JJA values while Fig. 3. shows July alone.

L161: Please verify that the in-text references to the figures throughout this paragraph are correct. The opening line discusses anomalies, but references Fig 4a-c, which depicts cloud cover anomalies. I believe the next sentence should refer to Fig. 4, not Fig. 3. And the last sentence (L167) should refer to Fig. 4D-F, correct?

L169: I suggest using the more specific “JJA” instead of “summer-wide.” Also, the authors suggest that the fact that the anomalies in Fig. 4 are not as pronounced as those in Figs 2 and 3 implies that July conditions dominate the JJA average. But Fig. 3 shows select atmospheric fields for July alone. Considering this, and given the logic presented by the authors, shouldn’t the anomalies in Fig. 4 be consistent with Fig 2 but not Fig. 3?

L185: What is the basis for the statement that, “Like in 2012, a lot of melt produced in 2023 was either refrozen or deposited.” If Fig. 6c, please reference this at the conclusion of the sentence. Also, “a lot” is quite vague. Would it be possible to compute the fraction of meltwater production that was retained?

L200: Given the Lelli et al. (2023) results, was the original intent of this analysis to use the GOME2 data as a melt indicator or a cloud cover indicator?

L219: If GOME-2  data is used as a melt indicator, is there any advantage to using this product over passive microwave brightness temperatures?

L224: delete the “is” after “Greenland-wide”

L324: This seems like an important finding that may warrant being featured in the abstract.

L351 and 375: It may be a bit strong to say that blocking over Greenland is “expected” to become more frequent. Preece et al. (2023) present one line of evidence that Arctic amplification may promote anticyclonic conditions over Greenland. Modeling and observational analysis also point to a link between declining sea ice in Baffin Bay and increased Greenland blocking and should be acknowledged here:

Liu, J., Chen, Z., Francis, J., Song, M., Mote, T., and Hu, Y.: Has Arctic Sea Ice Loss Contributed to Increased Surface Melting of the Greenland Ice Sheet?, J. Climate, 29, 3373–3386, https://doi.org/10.1175/JCLI-D-15-0391.1, 2016.

Sellevold, R., Lenaerts, J. T. M., and Vizcaino, M.: Influence of Arctic sea-ice loss on the Greenland ice sheet climate, Clim Dyn, 58, 179–193, https://doi.org/10.1007/s00382-021-05897-4, 2022.

However, GCM simulations consistently suggest a decline in Greenland blocking frequency, so this is very-much still an open question:

Delhasse, A., Hanna, E., Kittel, C., and Fettweis, X.: Brief communication: CMIP6 does not suggest any atmospheric blocking increase in summer over Greenland by 2100, International Journal of Climatology, 41, 2589–2596, https://doi.org/10.1002/joc.6977, 2021.

Hanna, E., Fettweis, X., and Hall, R. J.: Brief communication: Recent changes in summer Greenland blocking captured by none of the CMIP5 models, The Cryosphere, 12, 3287–3292, https://doi.org/10.5194/tc-12-3287-2018, 2018.