General Comments:

In this new study, the authors evaluate the boreal summer thermodynamic response (temperature and precipitation) to a sudden reduction in Arctic sea ice. In particular, they run targeted experiments using two fully-coupled climate models and two different horizontal resolutions for evaluating changes in climate extremes over nearby high latitude regions. Sea-ice loss is imposed in these experiments using a modified sea ice albedo scheme, which is consistent with previous studies. However, the temperature and precipitation responses are scaled by the amount of sea-ice loss, which helps to more easily compare the two climate models that show different sea-ice forcing (both regionally and in overall magnitude). In summary, the authors find an increase in frequency and persistence of maximum surface air temperature in the high latitude regions of the Arctic, which is especially large near Svalbard. In contrast, the extreme precipitation response is less robust compared to atmospheric internal climate variability.

Overall, this is a useful study for understanding the local response to rapid changes in Arctic sea-ice loss during the boreal summer. While there is an abundance of literature for assessing Arctic-midlatitude climate linkages (especially due to sea ice), there is less work on understanding the response in summer (relative to remote linkages in winter). For the most part, the methods here are logical, and the text is well-written. However, the paper is generally too long and can be substantially shortened to focus on the novel results of these climate model experiments. The introduction is quite long and confusing as it conflates winter and summer Arctic-midlatitude linkages, which are quite different in their respective dynamical tropospheric/stratospheric mechanisms and pathways. Instead, it would be helpful to compare this study’s results with those that have focused on summertime changes within the high latitude regions. Again, focusing on ‘what is new here.’ I have some more detailed comments, suggestions, and references below. After some shortening of the main text, this paper should be acceptable for publication in Weather and Climate Dynamics.

Recommendation: 

Major Revisions.

Specific Comments:

1.    L2-3; This sentence is a bit confusing to me. It makes it sound like this is the first study to conduct experiments with a large reduction in Arctic sea ice, which is not the case.

2.    L18-82; This introduction is quite long. I think it can be more concise by focusing on the direct connections with this work (e.g., L37-62), rather than restating Arctic climate change, which is already well documented in plenty of studies.

3.    L28; Latest data from observations/reanalysis reveal that warming is now at least “three” times as fast as the global average

4.    L29; Ballinger et al. (2020) can be updated to the newest Arctic Report Card 2021 (referenced below)

5.    L35; Review papers, such as Cohen et al. (2018) and Overland et al. (2021), are more appropriate studies to cite here

6.    L44-46; This might be the case for historical forcing, but some studies have shown that the remote response to future Arctic amplification is more robust than sea-ice loss alone later in the 21st century (see Labe et al. 2020)

7.    L52-53; In my view, this is a key/novel point for this study and should further be distinguished in the introduction (rather than conflating studies that have only focused on the peripheral response during winter).

8.    L52-53; Coumou et al. (2018) should be cited here. See references within for summertime studies.

9.    L70-71; Are the perturbations here realistic compared to CMIP projections of sudden/rapid ice loss events (on this timescale)?

10.    L79-82; This outline paragraph is extra text and unnecessary.

11.    L88; This table is very helpful!

12.    L107-108 & L127; Do the models include an internally generated QBO? This has been shown to modulate the response to Arctic sea-ice loss in some climate models.

13.    L139-140; Sun et al. (2021) compared methods between modifying ice albedo and nudging

14.    L142-143; Although this may still be an underestimate (Peings et al. 2021) for some of the regions outlined in Figure 2 (e.g., “ER”)

15.    L154; This is just a suggestion, but it might be helpful putting these climate indices in a table. It can be a bit cumbersome to read in the paragraph.

16.    L188; This study is not relevant here.

17.    L194-196; How does the effective sea-ice thickness actually change between PERT and CTRL in these experiments? Is the mean state realistic, such as compared to CryoSat-2 or PIOMAS? Changes in sea-ice thickness can influence surface turbulent fluxes and thus the local thermodynamic response.

18.    L209-210; Again, could this also be due to difference in the sea-ice thickness mean state? 

19.    L224-235; Is there any role for changes in ocean heat transport that result in the temperature response? This is one advantage in using the fully-coupled experiments here.

20.    L228; This is not necessarily the case during summer, where sea ice is mostly confined to the Fram Strait and northern Greenland.

21.    L228-229; And major differences in topography/elevation

22.    L258-264; This transition paragraph is a bit confusing to me. I suggest rewriting to improve clarity for the readers, especially when discussing the effect of temperature extremes over the continental regions. While there is a quick investigation of the NAO response, is it possible there is another dynamical contributor to the changes in temperature extremes (i.e., not just a turbulent heat flux response)?

23.    L324; Some studies have considered the response to very rapid ice loss events (e.g., Semmler et al. 2016)

24.    L325-328; Sorry, but I am not sure I understand what you mean here.

25.    L334-335; It may be helpful to remove “cold days” and “warm days” to improve interpretation of this result

26.    L348-350; But this could be a product of assessing responses in summer versus winter

27.    How do these horizontal resolution results compare to Streffing et al. (2021)?

Technical Comments:

1.    L19; “more pronounced in [late] summer”

2.    L274; change “models” to “experiments”

Figures/Tables:

1.    Figures 6/10; Is there any way that statistical significance could be denoted here? For example, comparing the CTRL and PERT PDFs in each respective region and adding a star for statistical significance.

Appendix:

1.    Figure A1; Are any of these changes in the NAO statistically significant? If so, could they be indicated on the graph?

2.    Could the authors include a data availability statement for the climate model experiments?

References: 

Ballinger and Coauthors, 2021: Surface Air Temperature. Arctic Report Card 2021, T. A. Moon, M. L. Druckenmiller, and R. L. Thoman, Eds., https://doi.org/10.25923/53xd-9k68

Cohen, J., Zhang, X., Francis, J., Jung, T., Kwok, R., Overland, J., ... & Yoon, J. (2018). Arctic change and possible influence on mid-latitude climate and weather: a US CLIVAR White Paper. US CLIVAR reports.

Coumou, D., Di Capua, G., Vavrus, S., Wang, L., & Wang, S. (2018). The influence of Arctic amplification on mid-latitude summer circulation. Nature Communications, 9(1), 1-12.

Labe, Z., Peings, Y., & Magnusdottir, G. (2020). Warm Arctic, cold Siberia pattern: role of full Arctic amplification versus sea ice loss alone. Geophysical Research Letters, 47(17), e2020GL088583.

Overland, J. E., Ballinger, T. J., Cohen, J., Francis, J. A., Hanna, E., Jaiser, R., ... & Zhang, X. (2021). How do intermittency and simultaneous processes obfuscate the Arctic influence on midlatitude winter extreme weather events?. Environmental Research Letters, 16(4), 043002.

Semmler, T., Jung, T., & Serrar, S. (2016). Fast atmospheric response to a sudden thinning of Arctic sea ice. Climate Dynamics, 46(3-4), 1015-1025.

Streffing, J., Semmler, T., Zampieri, L., & Jung, T. (2021). Response of Northern Hemisphere Weather and Climate to Arctic Sea Ice Decline: Resolution Independence in Polar Amplification Model Intercomparison Project (PAMIP) Simulations. Journal of Climate, 34(20), 8445-8457.

Sun, L., Deser, C., Tomas, R. A., & Alexander, M. (2020). Global coupled climate response to polar sea ice loss: Evaluating the effectiveness of different ice‐constraining approaches. Geophysical Research Letters, 47(3), e2019GL085788.

General comments

The authors examine changes in climate extremes in temperature and precipitation in the Arctic summer using 40-member climate model ensembles where the sea ice albedo is replaced by the open ocean value. They include results from two different climate models ad two different resolutions, allowing some sampling of model uncertainty. The authors focus on impacts in the peripheral Arctic regions, where changes will likely have societal relevance. This is a welcome addition to the literature, which has often focused more on Arctic-midlatitude linkages, rather than impacts in the peripheral Arctic regions. The methods are generally appropriate, although some aspects require clarification. Some of the presentation and discussion should be improved.

As noted by Sun et al. (2020, https://doi.org/10.1029/2019GL08578) and others, the use of albedo reduction to investigate the response to sea ice loss mainly impacts summer sea ice, with little to no impact during the winter. This can be seen in the present manuscript in Fig. 3, panels c and d. However, satellite observations show a decrease in winter sea ice area over 1979-present, and CMIP6 models project this to increase over the 21^st century (see e.g. Fig. 2a in SIMIP Community 2021, https://doi.org/10.1029/2019GL086749). This manuscript focuses on summer, but the authors should add some discussion on the seasonality of sea ice loss imposed in these experiments, and how that may affect the interpretation of the results. The existence of other techniques for considering the impact of sea ice loss and the seasonality is hinted at in the Conclusions at L357, but I think it deserves more attention in the manuscript, perhaps in the Discussion.

I think the Introduction should be revisited – some of the logical flow is a bit unclear.

Importantly, there is no data availability statement.

Specific comments

L44: I don’t follow the logic of the sentence beginning ‘Furthermore’

L55: I think this sentence is too strong. Suggest rephrasing to ‘An increase of climate extremes (…) can have substantial impacts on human activities’ or similar.

L58: Suggest adding a reference to Landrum & Holland (2020, http://dx.doi.org/10.1038/s41558-020-0892-z) somewhere in this paragraph.

L159: Can you explain further why these eight variables were chosen? I don’t understand what is meant by ‘because they can show extreme changes in frequency or in persistence’

L165: I understand that the naming convention of ETCCDI, but the term ‘ice days’ is confusing here, since it doesn’t refer to the presence of sea ice, but rather than the occurrence of freezing conditions.

L181: Are land and ocean grid cells treated equally in these regions? At L255, the text says that Screen et al. 2015 takes oceanic areas into account – does this manuscript not do that? It might also be helpful to add the fraction of each region that is land/ocean in Table 2.

L181: May be worth stating that there is no longitudinal variation in grid cell area.

Figure 3: Please add the satellite observations to the top panels.

L193: ‘in the Eastern Arctic’ – I would highlight the Barents Sea here.

L194: Suggest softening this sentence from ‘The main reason for these discrepancies’ to ‘These discrepancies may arise due to’, as it could be due to other factors.

L196: Can you comment on which is more realistic? It would be helpful to show sea ice thickness maps in the Appendix.

L230: Which kind of ice? Sea ice?

L147 and Figure 5 caption (and elsewhere) – when referring to ‘amount of sea ice loss’, is this total Northern Hemisphere sea ice area loss or sea ice volume?

L244: Please clarify how the regional average metric is computed. Regional average of temperature first, then compute e.g. number of days below freezing?

L263: Why not show this? This seems like an interesting point.

Technical comments

L11: Remove ‘statistically speaking’ – how else would you define robustness?

L19: Replace ‘Actually’ with ‘In particular’, or similar

L30: Replace ‘associated to’ with ‘associated with’

L40: Replace ‘the same confounding factor’ with ‘a common factor’?

L52: Replace ‘less attention was paid’ to ‘less attention has been paid’

L60: Replace ‘The projected Arctic sea ice loss can be responsible’ with ‘The projected Arctic sea ice loss could be responsible’

L150: I found this wording a bit confusing. How about ‘which limits spurious local test rejections’, or similar?

L191: Replace ‘even if the same protocol has been applied’ with ‘although the experimental set up is the same’?

L196: Replace ‘could lead to more sea ice in this model’ with ‘could lead to more sea ice being retained in this model’?

L207: Replace ‘continents’ with ‘landmasses’ – the places named are not continents

L226: Replace ‘a further south island’ with ‘an island further south’

L227: Replace ‘is the second region in terms of warming’ with ‘is the region with the second-strongest warming’

Figure 5 caption: Replace ‘associated to’ with ‘associated with’

L261: Replace ‘the importance to study’ with ‘the importance of studying’

L263: Replace ‘extreme frequency’ with ‘frequency of extremes’ and similarly for ‘extreme persistence’

L289: Missing ‘m’ in symmetrical

L294: Replace ‘in the northern Canada’ with ‘in northern Canada’, ‘generates’ with ‘generate’

L350: Avoid using the word ‘probably’ -it’s vague.

Figure 6 – Is there a better way of showing the PDFs for the different regions besides adding 0.1 to the y-axis for each region?