Ocean warming drives rapid dynamic activation of marine-terminating glacier on the west Antarctic Peninsula

Ice dynamic change is the primary cause of mass loss from the Antarctic Ice Sheet, thus it is important to understand the processes driving ice-ocean interactions and the timescale on which major change can occur. Here we use satellite observations to measure a rapid increase in speed and collapse of the ice shelf fronting Cadman Glacier in the absence of surface meltwater ponding. Between November 2018 and December 2019 ice speed increased by 94 ± 4% (1.47 ± 0.6 km/yr), ice discharge increased by 0.52 ± 0.21 Gt/yr, and the calving front retreated by 8 km with dynamic thinning on grounded ice of 20.1 ± 2.6 m/yr. This change was concurrent with a positive temperature anomaly in the upper ocean, where a 400 m deep channel allowed warm water to reach Cadman Glacier driving the dynamic activation, while neighbouring Funk and Lever Glaciers were protected by bathymetric sills across their fjords. Our results show that forcing by warm ocean water can cause the rapid onset of dynamic imbalance and increased ice discharge from glaciers on the Antarctic Peninsula, highlighting the region’s sensitivity to future climate variability.

The study is well done and could be published after minor revisions and addressing a few notes and questions in the review.I will say that similar work has been reported numerous times for Greenland outlets by a host of researchers (Howat, Rignot, Joughin, Hamilton, Catania, Enderlin, Moon, Willis….) and this study details a very similar course of evolution for a fjord-bounded outlet glacier.However, this study is important in that the data sets are fairly comprehensive, span a significant period of time, and are temporally and spatially dense during the period of rapid evolution.Moreover, the breakup style of the relatively extensive ice tongue (relative to most Greenland outlets) appears to be unique --and more might be made of this fact: while hydrofracture processes and damage-related processes have dominated most literature on ice shelf break-up, the observation of a breakup with a significant component from basal melting (augmented by increased damage) may be new (not sure).
Several detailed comments are set as notes within the submitted .pdffile There are valuable Landsat 8 images of the Cadman showing details of the loss of the ice shelf in Dec-Jan 2020 and 2021 that might illustrate the break-up mechanism, and could be added to your S1 figure.This style of break-up could be unique in the area (possible exception of the Wordie) in that melt ponding or hydrofracture does not appear to be involved.This strengthens your case that the ice shelf broke apart by a combination of increased damage due to acceleration and loss of contact with bedrock ridges near the evolving grounding line.Check 17 Jan 2020 and 26 Jan 2021 (and there are likely more) -to me it looks as though sub-ice shelf channels weakened the shelf by creating ~polygonal blocks that then concentrated strain and led to rapid calving along the channel lines.More Could Be Made Of This in your paper as an example of a style of ice shelf break-up that has not been widely discussed to date.Make note that your study does not show a presence of surface ponding (if that's the case; I think it is), a harbinger of hydrofracturing processes on a shelf.
Reviewer #2 (Remarks to the Author): This paper investigates Cadman Glacier in the west Antarctic Peninsula (AP), focusing on its unique characteristics and recent changes.Unlike other glaciers in the region, Cadman Glacier velocity increased between 2018 and 2019 and experienced persistent thinning of its ice shelf.This thinning eventually led to retreat and rapid acceleration of the glacier.Warm ocean temperatures and bed geometry played significant roles in these changes.The study emphasizes the vulnerability of glaciers in the region to warm ocean waters and the need for further research to understand their future evolution.
The paper provides a comprehensive analysis of Cadman Glacier in the west Antarctic Peninsula, focusing on its unique characteristics and recent changes.The authors effectively present the background information and clearly articulate the research objectives.The methodology employed in the study, including the use of CryoSat-2 data and previous studies, appears to be partially appropriate for the research questions at hand.The analysis of surface elevation, elevation change, and ice velocity provides valuable insights into the glacier's behavior, but grounding zone measurements are missing which could hinder the interpretation of the results.Additionally, I was surprised the authors did not provide both an eulerian and lagrangian melt rate analysis.
Overall, the findings of the study are significant and contribute to the understanding of glacier dynamics in the region.The observed persistent thinning of Cadman Glacier's ice shelf and its subsequent retreat and acceleration are compelling evidence of the glacier's vulnerability to warm ocean waters.
The paper effectively discusses the role of bed geometry and bathymetric features in influencing the glacier's response to environmental forcing.The inclusion of comparisons with neighboring glaciers adds depth to the analysis and supports the conclusions drawn.The implications of the study are welldiscussed, particularly in terms of the impact on ocean circulation, productivity, and the ecosystem.The mention of potential regional-scale consequences further strengthens the significance of the research.The paper is well-written, with a clear and concise presentation of the research findings.The academic tone and language throughout the manuscript contribute to its overall quality.
However, one area that requires improvement is the discussion of limitations and future research directions.While the authors briefly mention the need for further studies and fieldwork measurements, more specific recommendations or suggestions for future investigations would enhance the paper's completeness specifically focused on the need of frequent grounding line measurements and melt rate measurements see (Shean et al 2019Milillo et al 2022) Additionally, it would be beneficial to provide a clearer connection between the findings of this study and their broader implications for climate change and sea-level rise.This would help contextualize the significance of the research within the larger scientific discourse.Overall, this paper presents valuable research on Cadman Glacier and its response to environmental changes.With minor revisions and additions to the discussion section, this study has the potential to make a significant contribution to the field of glaciology and climate science.

Few more comments follows:
Unless a reference is missing, lines 73-75 seem a bit out of context for the introduction section and could be more appropriate in the conclusions 79-81 when referencing the 4 areas of the Cadman glacier a reference to a figure should guide the reader and enable the spatial identification of these areas on a map.
Fig1 missing reference to the DEM used for shading.

Reviewer #3 (Remarks to the Author):
This manuscript presents a detailed set of observations of Cadman Glacier, a marine-terminating glacier on the west coast of the Antarctic Peninsular, spanning a period of rapid change in the glacier's ice shelf.The observations, which include newly derived data on ice thickness, velocity and terminus position, clearly document the thinning, ungrounding, acceleration, retreat and ultimate collapse of the ~5 km ice shelf over the period ~2010-2022.Through comparison with existing hydrographic and bathymetric data, the authors present a convincing case that this reflects the influence of the ocean, with gradual weakening of the shelf due to long term thinning followed by a final collapse triggered by the presence of particularly warm ocean waters on the shelf.Altogether, this makes a nice case study of a moderately sized marine terminating glacier experiencing ice shelf thinning and collapse in response to warming ocean waters.The detailed observations document this process well, and I believe the methodology, analysis and interpretation are sound.
While I am confident that the work is worthy of publication in a relevant journal, I am less confident that it represents the kind of major, high-impact advance that would typically be associated with Nature Communications.The finding that the retreat of glaciers on the western Antarctic Peninsular is predominantly ocean-driven was presented on a much larger scale by Cook et al (2016) -the current manuscript presents a detailed case study of one such example of this, but does not to me really advance our broader understanding of this topic beyond this particular glacier (which is not in itself of particular importance as a source of sea level rise etc).The sequence of processes observed -ice shelf thinning, acceleration and break up, grounding line retreat, dynamic thinning upstream -are quite familiar from the loss of other ice shelves and tongues in Antarctica and Greenland (e.g.Motyka et al (2011), to give just one example).The attribution to oceanic forcing supports the findings of Cook et al (2016), but the study lacks the kind of novel observations (such as in situ hydrographic measurements close to the glacier, or high resolution / spatially resolved measurements of shelf thinning) that might really advance our understanding of the underlying processes.
Putting this aspect aside, I do not have any major concerns with the methods, analysis or interpretation as currently presented.I found the manuscript to be largely well written, if perhaps overly descriptive at times -it might be possible to make better use of the figures and reduce the amount of descriptive text a little.A have a few more minor comments, outlined below.
L28.The 'cryosphere' is generally used to refer to ice and snow on a global scale -better to refer to the glaciers of this region or similar L37.Ambiguous phrasing -suggest something like 'increased by 400 % between the periods 1992-1997 and 2007-2012' L161.A figure reference would be useful L175-185.It would be useful to add a reminder of the years of the ASAID grounding line and DEMs L244-5.It's not easy to discern the grey dots in the figure, perhaps a different colour would work better L264-275.It wasn't clear to me why the addition of surface lowering merited such attention -is this making a methodological point (that other researchers need to make sure they factor this in), or that surface lowering has an important impact on the dynamics of the glacier?L303-306.Yes these glaciers have shallower sills, but its notable in Figures 6c-d  The study is well done and could be published after minor revisions and addressing a few notes and questions in the review.I will say that similar work has been reported numerous times for Greenland outlets by a host of researchers (Howat, Rignot, Joughin, Hamilton, Catania, Enderlin, Moon, Willis….) and this study details a very similar course of evolution for a fjord-bounded outlet glacier.
However, this study is important in that the data sets are fairly comprehensive, span a significant period of time, and are temporally and spatially dense during the period of rapid evolution.Moreover, the breakup style of the relatively extensive ice tongue (relative to most Greenland outlets) appears to be unique --and more might be made of this fact: while hydrofracture processes and damage-related processes have dominated most literature on ice shelf break-up, the observation of a breakup with a significant component from basal melting (augmented by increased damage) may be new (not sure).
Several detailed comments are set as notes within the submitted .pdffile We thank the reviewer for their comments regarding the quality and importance of the study, and for their insightful, informative and productive comments as a whole.In particular we thank the reviewer for taking the time to independently study satellite images of Cadman Glacier and provide helpful suggestions based on these.
We have replied to each comment in turn and made changes to the manuscript to address them.A major suggestion by the reviewer was to include more optical satellite images of Cadman Glacier in figure S1.We have done this and expanded the discussion of the failure mechanism for the Cadman Ice Shelf.We agree with the reviewer that the importance of this wasn't highlighted clearly enough in the previous version.
We believe that changes and clarifications presented in the revised manuscript have addressed the reviewer's concerns and have enriched the manuscript.We thank the reviewer for their time and effort.
1 We have changed the description of damage on the shelf to describe this: 120: 'Although advanced, during this period the seaward 3 km of glacier tongue shows a substantial increase in damage and crevassing on the ice shelf surface and a widening of the shear margins (Fig. S1), suggesting that the structural integrity of the ice was decreasing.Notably in the 17th January 2020 image (Fig. S1d) there is a distinctive Done -changed to 'we define' as this is a naming we have chosen.We chose to call this part lower as it is below the Cadman icefall which separates the upper slow flowing basin from this lower section.1.17 please spell this out a bit more -accelerated by 94%, a speed-up of 1.5 km/yr, reaching (something like) 2.9 km/yr by xx date.

93
Comment -This short paragraph is just to introduce the reader to the acceleration event being discussed, so that the following paragraphs are not out of context.We discuss the ice speed changes in detail in the results section so we would prefer not to duplicate such detail in the introduction.
1 We thank the reviewer for their comments and constructive feedback.In particular we were pleased to hear the reviewer found the study had the potential to make a significant contribution to the field of glaciology.
We have replied to the reviewers comments in turn and made changes to the manuscript to address them.Specifically, we have added an updated grounding line measurement, including details of the method used.We justify why we do not feel it is appropriate to include a Lagrangian basal melt rate analysis in the manuscript, through comparison to other studies and considerations relating to the other datasets required for this calculation.Finally, we have significantly expanded the discussion section of the manuscript to provide concrete recommendations for future work and to place the results in the greater context of the Antarctic Peninsula region's contribution to rising sealevels.
Overall, we believe that addressing the reviewer's comments has substantially enhanced the manuscript.We thank the reviewer for their time and consideration.

2.2
The Grounding line: Directly measuring the grounding line position of Cadman Glacier is difficult because there is no InSAR coherence on the glacier in available datasets.For example, a study which processed all available Sentinel-1 data in Antarctica for 2018 and automatically delineated grounding lines did not produce any measurements for Cadman Glacier (Mohajerani et al., 2021).We investigated measuring grounding line position using repeat track ICESat-2 laser altimetry (Fricker and Padman, 2006;Brunt et al., 2010), but found insufficient data coverage for the Cadman grounding zone, due to the large spacing between tracks at lower latitudes and an unfavorable across-glacier track orientation.
In the original manuscript, we discuss that the grounding line of Cadman glacier likely retreated from the ASAID dataset (1999)(2000)(2001)(2002)(2003) position to a position around 2km inland based on changes in ice flow speed and DEMs: Original MS L340-344: 'We interpret these changes to be indicators of a loss of basal traction and therefore the retreat of the grounding line, from the ASAID grounding line location to a position 2 km further inland which is co-located near the point (ii) 2015 surface bump and associated speed minimum (Fig. 3).'While a precise InSAR grounding line measurement is not possible, we have formalised our analysis of the pre-acceleration grounding line position by measuring the break in surface slope of the 2015 REMA DEM (Howat et al., 2019).We define this as the most seaward local peak in surface slope gradient (Hogg et al., 2018;Friedl et al., 2020) and find the grounding line position to be 1.25 km inland of the ASAID (1999 -2003) position.
Done -we have amended the text in several places to include this measurement.
In  3b).' In the discussion: 345: 'Between 1991 and 2010 there was a change from a local ice speed minimum feature to an area of spatially constant ice speed immediately behind the ASAID grounding line location, which was matched by the loss of an ice surface bump between the Cook AP (2000-2009) and the REMA DEMs ( 2015), together with the inland migration of the break in surface slope.
We interpret these changes to be indicators of a loss of basal traction and the retreat of the grounding line, from the ASAID grounding line location to a position 1.25 km further inland located near the point (ii) 2015 surface bump and associated speed minimum (Fig. 3).Such a loss of basal traction could also explain the increase in ice speed of 0.25 km/yr (15 %) between 1991 and 2010.' We have also updated Figure 1b and Figure 3b to show the 2015 grounding line position.

2.3
Additionally, I was surprised the authors did not provide both an eulerian and lagrangian melt rate analysis.

Basal melt rates:
The Furthermore, for any Lagrangian calculation individual altimetry results must be advected with ice flow and an elevation change rate calculated via a plane fit to an appropriate grid postings resolution.Grid resolutions for altimetry based basal melt data in the literature vary from 500 m (Gourmelen et al., 2017) to 30 km (Adusumilli et al., 2018), with the finer grid resolution calculations using high resolution swath mode CryoSat-2 altimetry which was not available on Cadman Ice Shelf.Additionally, the ice shelves to which Lagrangian calculations have been applied are many orders of magnitude larger than the Cadman ice shelf.
Overall, we believe these methodological limitations prevent the accurate calculation of a useful and robust basal melt rate value, which justifies our choice to use a Eulerian average ice shelf thinning rate which we can report with confidence.

2.4
Overall, the findings of the study are significant and contribute to the understanding of glacier dynamics in the region.The observed persistent thinning of Cadman Glacier's ice shelf No change required, thank you for the positive assessment.
and its subsequent retreat and acceleration are compelling evidence of the glacier's vulnerability to warm ocean waters.

2.5
The paper effectively discusses the role of bed geometry and bathymetric features in influencing the glacier's response to environmental forcing.The inclusion of comparisons with neighboring glaciers adds depth to the analysis and supports the conclusions drawn.The implications of the study are well-discussed, particularly in terms of the impact on ocean circulation, productivity, and the ecosystem.The mention of potential regional-scale consequences further strengthens the significance of the research.
The paper is well-written, with a clear and concise presentation of the research findings.The academic tone and language throughout the manuscript contribute to its overall quality.
No 2.10 79-81, Fig. 1 when referencing the 4 areas of the Cadman glacier a reference to a figure should guide the reader and enable the spatial identification of these areas on a map.

77-79
Comment -This description is intended to familiarize the reader with the glacier and the definitions here are only used qualitatively for description.Quantitative geometric definitions and those used for calculations are defined on a map in Figure 1b.We feel adding more labels to Figure 1b would reduce clarity, so have left it unchanged.2.11 Fig. 1 Fig1 missing reference to the DEM used for shading.

Fig. 1
No action required -Figure 1b,c,d all use the Landsat-8 image referenced in the caption as a background, no DEM was used.
2.12 152-153 Which grounding line has been used to measure these proportions ?
152 This relates to the sample areas defined in Figure 1b and Figure 3b.Comment -We thank the reviewer for their consideration of the manuscript and helpful comments.We are please that the reviewer found the work to be robust and well written.We thank the reviewer for their time.

Done -
We have addressed the reviewers comment point by point below.Specifically, the reviewer's comment about sill depth prompted us to expand the discussion of how fjord sills influence circulation and basal melt rates, which we feel notably improved this section of the manuscript.

3.2
While I am confident that the work is worthy of publication in a relevant journal, I am less confident that it represents the kind of major, highimpact advance that would typically be associated with Nature

Comment:
We feel the results presented in this study will be of great interest to scientists from a broad range of disciplines, which is why we chose to submit to Nature Communications.Our work combines high resolution remote sensing data with direct oceanographic measurements Communications.The finding that the retreat of glaciers on the western Antarctic Peninsular is predominantly ocean-driven was presented on a much larger scale by Cook et al (2016) -the current manuscript presents a detailed case study of one such example of this, but does not to me really advance our broader understanding of this topic beyond this particular glacier (which is not in itself of particular importance as a source of sea level rise etc).The sequence of processes observed -ice shelf thinning, acceleration and break up, grounding line retreat, dynamic thinning upstream -are quite familiar from the loss of other ice shelves and tongues in Antarctica and Greenland (e.g.Motyka et al (2011), to give just one example).The attribution to oceanic forcing supports the findings of Cook et al (2016), but the study lacks the kind of novel observations (such as in situ hydrographic measurements close to the glacier, or high resolution / spatially resolved measurements of shelf thinning) that might really advance our understanding of the underlying processes.
and reanalysis to significantly develop the process-based link between ice-ocean interactions in Antarctica, which is at the forefront of modern climate science and is essential for accurately projecting future sealevel rise.
In response to the comments of Reviewer 1, we have provided more detail about the novelty of the break-up of Cadman Ice Shelf without melt ponding and hydrofracture, a point which was not emphasized sufficiently in the original manuscript.See response to point 1.2.
Furthermore, through the responses to this point and those raised by other reviewers, we have substantially expanded the discussion to stress how the results of this study relate to broader research questions regarding ice-ocean interactions in Antarctica.See response to point 2.7.We also expanded our discussion of recommendations for future work and highlight the remaining challenges which must be overcome to better understand the glacier dynamics of this region.See response to point 2.6.
The reviewer is correct to cite the link between ocean warming and glacier retreat on the west AP found by Cook et al. 2016.We further develop this link to show that, as well as glacier retreat, ocean warming also causes increased ice discharge and mass loss, which directly contributes to sea-level rise.We note that since Cook et al.'s 2016, there have been few further papers exploring this link through observations.Compared to Greenland, the AP is a sparsely observed, but still important region making up 19% of Antarctica's sea-level contribution.For these reasons we believe our work will be of significant interest to the cryosphere, climate, and ocean science communities.
We hope the reviewer agrees that the changes made in response to all the reviewers' comment have improved the manuscript, further highlighted the important scientific results, and provide improved interest for the broad audience that Nature Communications commands.Of course, we respect the editor's final decision on the suitability of this manuscript for Nature Communications.

3.3
Putting this aspect aside, I do not have any major concerns with the methods, analysis or interpretation as currently presented.I found the manuscript to be largely well written, if perhaps overly descriptive at times -it might be possible to make better use of the figures and reduce the amount of descriptive text a little.A have a few more minor comments, outlined below.
No change required.

28
The 'cryosphere' is generally used to refer to ice and snow on a global scale 27 Done -changed to: -better to refer to the glaciers of this region or similar 27: 'major changes in the region's glaciers and ice shelves throughout the 20th and 21st centuries' 3.5 37 Ambiguous phrasing -suggest something like 'increased by 400 % between the periods 1992-1997 and 2007-2012' Done -we added a black border to these dots to make them clearer and removed the lower glacier sampling area (red dashed) which was not relevant to this plot 3.9 264-275 It wasn't clear to me why the addition of surface lowering merited such attention -is this making a methodological point (that other researchers need to make sure they factor this in), or that surface lowering has an important impact on the dynamics of the glacier?

378-382
The surface lowering term is significant because it reduces the ice discharge over time, even if surface velocities remain high.It is one of the ways that the glacier system can return to mass budget equilibrium.By accounting for this term, we can estimate the time taken for the glacier to return to equilibrium.This point is expanded upon in the discussion, lines 374-378.397-404 Comment -We thank the reviewer for raising this point and agree that the original discussion around the influence of the fjord sills could be expanded to be more comprehensive and nuanced.

Done
The oceanographic measurements we have used, while very interesting, are not direct measurements of the water properties in the individual glaciers' fjords, as they are taken from points on the west AP shelf and we are careful to note this clearly in the paper.
Recent observations in north Greenland and model studies show that fjord sills seaward of the grounding line of ice tongue glaciers exert a hydraulic control on buoyant meltwater plume driven fjord circulation (Jakobsson et al., 2020;Schaffer et al., 2020;Nilsson et al., 2022;Bao and Moffat, 2023).Conceptually, this can be understood as two regimes, one with low or insignificant sills which is melt-controlled and one hydraulically-controlled, where sills constrain exchange circulation.Importantly, modelling shows that sills do not need to be high enough to fully block warm deep water to substantially reduce melt rates at the grounding line.In this framework Cadman Glacier would be a melt-controlled system while neighboring Funk and Lever would be hydraulically controlled.Without direct oceanographic measurements it is not possible to definitively know the circulation regime of the Cadman, Funk and Lever Fjords, however these studies demonstrate that sills do not need to block all warm water to restrict melt rates.

REVIEWERS' COMMENTS
Reviewer #1 (Remarks to the Author): This is a re-review.
As I noted in my initial review, this is a well-done paper that assembles several data sets for a detailed analysis of a relatively important western Antarctic Peninsula glacier.
In my view, all of the reviewer comments were well-addressed.In particular, I appreciated the discussion of Reverwer2's suggestion of Eulerian and Langrangian melt tracking --without very dense altimetry data (in both time and space) this would not be easily possible, nor would such a result materially change the resuts presented.Reviewer 3 has only a few substantive comments.All three reviewers expressed appreciation of the work.
Reviewer 3 pointed out that this paper might not be apropos for Nature Communications, and there is some merit to that --Nature Geoscience might have been a better choice,, with essentially the same overall impact on the physical science community.However, similar works have appeared in Nature Communications in the past, and are higihly cited and I think appreciated by the wider earth and climate research community.
Reviewer #2 (Remarks to the Author): The Authors have addressed all my comments, I encourage the editor to accept this paper for publication.
Reviewer #3 (Remarks to the Author): Thank you to the authors for carefully and clearly responding to the comments I raised on the previous version of this manuscript.These were largely relatively minor queries and suggestions, which the authors have addressed satisfactorily.I am also pleased to see some increased discussion of the wider significance of the findings.
I have only a few comments on this revised manuscript, along with a number of suggestions relating to the clarity of the writing, which I will attach as an annotated PDF.
Comments: L266-7.The discharge here is described as being constant between 2016 and 2018, when Figure 5 shows it varies quite substantially in this time period.I think what is meant is that there is no trend of increasing ice discharge over this time period -if so, the phrasing should be amended accordingly.
L363-5.This indicates that one would assume that surface ponding was a necessary condition for ice shelf collapse, but this seems to be a bit of an Antarctic Peninsula-centric perspective.The laterally constrained ice shelf at Cadman may have more in common with the Greenland's relatively small ice shelves and tongues, where surface melt ponds are not generally deemed to have such an important part to play, a comparison that should probably be made here.
L396-425.A missing link here seems to be to make the connection between the divergent behaviour of Cadman, Lever and Funk Glaciers and the discussion around tipping points.Yes Lever and Funk Glaciers may be less sensitive to ocean conditions than Cadman due to the shallower sills, but this may be only part of the story.As discussed in this section, the rapid changes at Cadman Glacier may reflect the culmination of a lengthy period of mass imbalance in the terminal region, which is finally sufficient to trigger a dramatic dynamic response.Lever and Funk may also be slowly responding to ocean warming, but not presently in a position whereby this is sufficient to trigger a major response.A comparison can be made to Cadman Glacier around 1995 -if you only had the data for that period in Figure 2, you might argue that Cadman Glacier was insensitive to ocean warming, but then an ocean warming event of similar magnitude 25 years later caused the tongue to collapse.Similarly, Lever and Funk may seem insensitive now to ocean warming now, but this could dramatically change at some point down the line.
It would be interesting to see if there were any more gradual changes in ice thickness apparent on the tongues of these glaciers if they fall within the area covered by your data.At the least though, I think this is a point that merits mention in the Discussion.Reviewer 3 pointed out that this paper might not be apropos for Nature Communications, and there is some merit to that --Nature Geoscience might have been a better choice,, with essentially the same overall impact on the physical science community.However, similar works have appeared in Nature Communications in the past, and are higihly cited and I think appreciated by the wider earth and climate research community.

N/A
We thank the reviewer for their time reviewing this manuscript and for their kind and constructive comments.
No further changes are required to address these comments.
Reviewer 2: The Authors have addressed all my comments, I encourage the editor to accept this paper for publication.

N/A
We thank the reviewer for their time reviewing this paper and their helpful comments.
No further changes are required to address these comments.
Reviewer 3: I have only a few comments on this revised manuscript, along with a number of suggestions relating to the clarity of the writing, which I will attach as an annotated PDF.3.2 266-267 L266-7.The discharge here is described as being constant between 2016 and 2018, when Figure 5 shows it varies quite substantially in this time period.I think what is meant is that there is no trend of increasing ice discharge over this time period -if so, the phrasing should be amended accordingly.

266
Thank you for this comment, we agree that changing the phrasing to be more precise would be better.
Changed to: 'Throughout 2015 ice discharge increased until it reached a mean of 1.85 ± 0.13 Gt/yr between 2016 and 2018 with no positive or negative trend' 3.3 363-365 L363-5.This indicates that one would assume that surface ponding was a necessary condition for ice shelf collapse, but this seems to be a bit of an Antarctic Peninsula-centric perspective.The laterally constrained ice shelf at Cadman may have more in common with the Greenland's relatively small ice shelves and tongues, where surface melt ponds are not generally deemed to have such an important part to play, a comparison that should probably be made here.

371
Thank you, we agree that the precise meaning could be made clearer.

Changed to:
370: 'The absence of surface melt ponding on the Cadman Ice Shelf prior to its collapse demonstrates that surface ponding is not a necessary precondition for ice shelf collapse in the Antarctic Peninsula, which can occur through ocean-driven thinning and unpinning alone, similar to small ice shelf failures observed in Greenland' 3.4 396-425 L396-425.A missing link here seems to be to make the connection between the divergent behaviour of Cadman, Lever and Funk Glaciers and the discussion around tipping points.Yes Lever and Funk Glaciers may be less sensitive to ocean conditions than Cadman due to the shallower sills, but this may be only part of the story.As discussed in this section, the rapid changes at Cadman Glacier may reflect the culmination of a lengthy period of mass imbalance in the terminal region, which is finally sufficient to trigger a dramatic dynamic response.Lever and Funk may also be slowly responding to ocean warming, but not presently in a position whereby this is sufficient to trigger a major response.A comparison can be made to Cadman Glacier around 1995 -if you only had the data for that period in Figure 2, you might argue that Cadman Glacier was insensitive to ocean warming, but then an ocean warming event of similar magnitude 25 years later caused the tongue to collapse.We feel this sentence is adequately clear

37
Fig. 3b Done -the DEM years are already referenced in the legend for Fig 3b, but we have added the ASAID years to the figure caption.

Fig. 6b
Fig. 6b Done -Anomaly in Fig 6 is 2019.Added year to Fig 6b legend to make this clearer

Figure 4 .
Figure 4.The new markers in part a are better, but they are still the same colour as the topographic shading.Why not use a different colour, like green or pink?

Response to reviewers: Ocean warming drives rapid dynamic activation of a marine-terminating glacier on the west Antarctic Peninsula
that there is still substantial warming at depths shallower than 200-230 m, so they should still have experienced significantly warmer temperatures at this time.
reviewer raises a question about why both Eulerian and Lagrangian basal melt rate calculations were not included.Note that we did not calculate a Eulerian or Langrangian basal melt rate value for the Cadman Ice Shelf in the original manuscript, but instead calculated a total ice shelf thinning rate for the shelf area defined in Fig 1b and Fig 4a, which has an area of 9.3 km 2 .This was a deliberate choice we made because we did not feel that the Cadman Ice Shelf was adequately resolved in the additional datasets required to calculate a basal melt rate.Specifically, for surface mass balance (SMB), the highest spatial resolution regional climate model available to us is the 5.5 km RACMO 2.3p2 Antarctic Peninsula model(van Wessem et  al., 2016)and for firn air content (FAC), existing products, such as those produced by the IMAU firn densification model have a resolution of 27 km.The width of the Cadman ice shelf is approximately 3 km, therefore it would not be well resolved in either of these datasets.The shelf is surrounded by steep topography up to 2 km high, meaning the values for SMB and FAC in these models for the single grid cell containing the Cadman ice shelf will likely not be representative of conditions on the shelf.
-we have amended the discussion section to make this link clearer:

Response to second round of reviewers' comments: Ocean warming drives rapid dynamic activation of a marine-terminating glacier on the west Antarctic Peninsula
Similarly, Lever and Funk may seem insensitive now to ocean warming now, but this could dramatically change at some point down the line.It would be interesting to see if there were any more gradual changes in ice thickness apparent on the tongues of these glaciers if they fall within the area covered by your data.At the least though, I think this is a point that merits mention in the Discussion.3.5Figure4.The new markers in part a are better, but they are still the same colour as the topographic shading.Why not use a different colour, like green or pink?