Response to reviews for Hedetoft, Bang Brinck, Mottram et al.. Mélange or landfast ice: What controls seasonal calving at Greenland outlet glaciers?

Reviewer 1:

The authors combine in-situ tracking buoys, remote observations, CTD, and climate model data to explore the effects of melange and landfast sea ice on calving dynamics at three Greenland outlet glaciers: Tracy, Melville, and Farquhar glaciers. The data is novel and important for future investigations of Greenland calving dynamics. However, there are major concerns that need to be addressed before the manuscript to be considered for publication.

Thanks for your nice remarks, we’re happy you see the importance of this study.

Major comments:

1) Melange is defined as "a granular mixture of calved icebergs and sea ice that forms in front of glaciers". Therefore, landfast sea ice is part of melange and directly modulates melange properties. For instance, in Robel 2017 Nat Commun paper, landfast sea ice is modeled as bonding material (cohesion) in between iceberg particles within the melange. In Meng 2025 Nat Commun paper, they stated that "On the other hand, cooler ocean and air temperatures in winter enhance mélange rigidity [due to the formation of landfast sea ice], making it easier to pile up thick mélange at the terminus to provide buttressing. How warmer oceans and atmospheric influence the mélange strength is the subject of future work." This manuscript could be a great contribution on quantifying how landfast sea ice modulates melange rigidity/thickness/buttressing force. But the current title/conclusion is confusing: landfast ice and melange should not be listed as two independent factors. With landfast sea ice, melange is easier to pile up to be thicker (due to the enhanced cohesion), which provides larger buttressing force. If melange does not affect calving dynamics, it's possible that it is just not thick enough. Theoretical derivation on how melange buttressing depends on melange thickness has been conducted in the following papers:

Amundson, Jason M., and J. C. Burton. "Quasi‐static granular flow of ice mélange." Journal of Geophysical Research: Earth Surface 123, no. 9 (2018): 2243-2257.

Kavinda Nissanka, Nandish Vora, Joshua Méndez Harper, et al. Experimental Investigations of Ice Mélange and the Flow of Floating Granular Materials. ESS Open Archive . August 24, 2024.

Meng, Y., Lai, C. Y., Culberg, R., Shahin, M. G., Stearns, L. A., Burton, J. C., & Nissanka, K. (2025). Seasonal changes of mélange thickness coincide with Greenland calving dynamics. Nature Communications, 16(1), 573.

Amundson, J. M., Robel, A. A., Burton, J. C., & Nissanka, K. (2025). A quasi-one-dimensional ice mélange flow model based on continuum descriptions of granular materials. The Cryosphere, 19(1), 19-35.

Thanks for your comment, it is indeed conceptually challenging to separate out landfast sea ice from mélange, and as we note in the abstract we consider the calving front outlet glaciers as part of a continuum that in fact starts before the glacier terminus and ends well into the landfast sea ice outside of the mélange zone. As landfast sea ice also extends well beyond the region of mélange we do think it helpful to distinguish between the two categories of land fast sea ice and unbonded mélange. Our observations show that the unfrozen or unbonded mélange at these outlet glaciers has very little influence on glacier velocity or calving rates and it is only when the separate ice blocks in the mélange zone are frozen into a matrix of landfast sea ice, losing the granular material properties that are characteristic of mélange and have the material properties more associated with rigid multi-year sea ice, that they start to exert any kind of (limited) influence on the calving rates. We have therefore clarified this in the definition of mélange and land fast sea ice in subsection 1.1 and added some additional sentences to the discussion where we focus also on the differences in mélange behaviour when bonded in landfast sea ice and when free floating on modulating calving processes.

We propose slightly adjusting the title to make this clearer too. Our new proposed title is: “Mélange needs landfast sea ice to modulate seasonal calving behaviour at Greenland outlet glaciers.”

2) Figure 5 implies that melange velocity is always faster than glacier, over the 1.5yr data acquisition period. This is surprising and not consistent with the finding in Amundson2018. Can the authors add some explanation of why this is the case for the 3 reported glaciers?

We were also surprised by this feature in our analysis. We also checked against the ITS_LIVE dataset and found a very similar result. In fact, careful examination of Figure 5 (which we have replotted to make the result clearer) shows the inner mélange zone velocity is usually very close to the velocity of the lower glacier point, particularly on Melville and Tracy glaciers but with some large deviations over some periods which likely represent calving events. The buoy data shows that the average daily velocities in the mélange zone are generally low but with abrupt jumps which we interpret to be associated with calving events and the addition of these abrupt jumps likely explains the above average velocity when smoothed in the time series. It is one of the reasons that we suggest that the mélange does not exert much back stress on the glacier fronts even when held in place by land fast sea ice. Ocean measurements indicate substantial tidally driven mixing close to the mid point between Tracy and Farquhar which drives some of the movements we detect at these locations and the driving stress of the glaciers is clearly sufficient to keep the glaciers moving through the winter.

As the glacier front advances into the mélange zone the mélange is also pushed forward and has less resistance than grounded glacier ice. This also leads to the crumple zone of ice ridges, shown in figure 16 at a location around 4km in front of Melville Glacier. The landfast sea ice further out in the fjord is in general stationary, except for some tidally driven movement, so this velocity away from the glaciers does decay with distance. We have added some further sentences to the discussion of this result, to clarify this result and have also plotted up 2D velocity maps (see reply to Reviewer 2 comment also).

3) Figure 9 could be a nice place to showcase landfast sea ice is important for interpreting calving dynamics. What will happen if the author superimpose air temperature profile (below 0C as blue, above 0C as red, as done in Fig10)? Will we see terminus advance coincide with below C (more sea ice, thicker melange, larger buttressing); terminus retreat with above C (less sea ice, thinner melange, less buttressing)?

This is a nice suggestion and we have modified Figure 9 with the blue/red colours suggested and lines showing the sea ice break-up, but note that figures 6, 7 and 8 also show calving front position by month, in general all three glaciers have their maximum advanced position in April and then calve back through the spring and summer, even before the fast ice has broken up in June to reach their minimum frontal position in October.

4) Figure 11, can the authors provide summer CTD too? Are there any seasonal changes, and how do this correlate with landfast sea ice thickness? Also, any explanation on why mix layer is 100m at the mouth of the fjord (dotted red line, line 320 in text)? Can the author report more data on seasonal sea ice thickness (Line325 briefly mentions sea ice in between was consistently measured to be around 1m thick)?

Unfortunately, we do not have a year round programme of CTD observations in front of the glacier termini, though summer observations from other years in other parts of the fjord do not show a significant change compared to the profile shown here. The mixed layer depth of 100m at the mouth of the fjord is very consistent with other locations in NW Greenland and represents the Atlantic water entering the fjord. Closer to the glaciers, there is more turbulent mixing, probably driven by upwelling from the bed of the glaciers which changes the mixed layer depth. The sea ice thickness was measured systematically in a few locations during this study and was remarkably consistent both inside the mélange zone and in the fast ice area outside the mélange zone, as it has been in other years. Ongoing work in a data rescue project is compiling a large database of sea ice thickness and snow depth on sea ice observations from this region over at least the last 20 years and will be published as a dataset separately. We will add some sentences clarifying both these points in the results and discussion.

5) Figure 13, are there tidal/diurnal melange velocity signal in fall? Are the data presented here winter data, when melange is strong due to the cohesion induced by landfast sea ice, and thus less vulnerable to tidal forcing?

The buoy measurements are only possible when there is thick sea ice to travel on, and that we can use as a platform to place the instruments. We therefore do not have any autumn or early winter measurements when the ice would be thinner. The fast ice reaches maximum thickness in March/April and starts to thin from May, so the observations we have include also very thin sea ice in June where we do not however see either a diurnal or tidal signal either. We note in addition that when the landfast sea ice breaks up, both the GNSS buoys and many of the icebergs are very quickly (within a few weeks) ejected from the fjord, indicating rapid circulation through the fjord system that may also be related to tidal circulation. The movement of icebergs and mélange out of the fjord is visualised in the short timelapse videos of Sentinel-2 satellite observations linked to in the Appendix. This is a good question though so we have clarified the field deployments in the methods section.

6) After drawing conclusion that "melange does not matter for these three glaciers", can the author look into melange thickness data (i.e., ArcticDEM) and check whether melange is just too thin to do anything? The 3 fjord here are all very short. If there is not enough lateral friction from the fjord, it will be very hard to pile up thick melange, which might explain the observation here. 

We have pointed out in the discussion section that while we think these glaciers are very representative of many if not most Greenland glaciers, there are other very large outlet glaciers like Sermeq Kujalleq/Jakobshavn Isbræ and Helheim glacier, where large masses of mélange pile up within their long fjords and that this may have different properties and be able to exert more back stress, although modelling studies are not convincing on this point as the packing density is also not well observed. Mélange thickness is a very difficult concept to apply because it is very clear when standing in the mélange zone that it is a highly heterogenous material and trying to reduce to a single number or even field of “thickness” is not very helpful, at least at these glaciers, given uncertainties in fjord bathymetry and iceberg sizes. Even high resolution datasets like ArcticDEM or satellites like TerraSAR-X and Sentinel 1/2 do not resolve this heterogeneity sufficiently at these glaciers, which is very clear when in the field directly. Future work will try to address the question of mélange thickness in more detail with photogrammetry based on UAV mapping that was carried out concurrently with the buoy deployments, from which we will derive a size distribution of icebergs and ice floes that can be applied in a modelling framework. We will be sure to make all data openly available for all groups to use when it becomes available. We have added some more text on this point to the discussion.

Minor comments:

• Line 245 "there is a break point where the ... rather than the properties of the melange itself". If landfast sea ice modulates melange cohesion and in turn thickness, then it should affect the properties of the melange itself. 

Thanks that is a good catch, we have adjusted the sentence slightly to read:

“There is a break point where the ice in the mélange zone shifts between these two regimes in mid-June in both years, suggesting that the onset of enhanced calving is related to the break up of landfast sea ice in the melange zone.”

• Figure 3(b), is it better to use log scale for velocity so that we can see the baseline velocity more clearer?  Also, the buoy data is very valuable. Can you archive some of the data with full temporal resolution (10~30mins, instead of the 1-day averaged shown here) in SI or repository? It's especially useful for future study on melange dynamics during calving, where iceberg velocity can vary by 3 order of magnitude (5~5000m/day).

We have replaced the figure 3b and 4b with log-scale plots as we agree it is clearer to see the baseline velocity.

And yes we are very much in favour of open data publishing whenever possible. We are currently preparing a full data paper which will include raw data as well as the processed analysis here as well as an additional two years of similar observations. We hope to keep adding data to it as part of a monitoring programme in the fjord for as long as we can keep the programme going. This will be published in an open access repository in 2026. We are happy to share the raw data before then with others who would like to use it and have added a line to this affect in the data availability section and in the future work section.

• Line 270, extensional melange flow does not mean no buttressing. As long as the melange is thick enough, there can be enough buttressing. For instance, Meng2025 model and Amundson2018 remote observation data both show melange extensional flow during winter seasons;

We agree it does not mean no buttressing, we have slightly reformulated the sentence to read:

 “Interestingly, our results confirm that the sea ice in the melange zone and the outer fast ice zone is moving away from the glacier fronts slightly faster than the outlet glaciers are advancing, even when the landfast ice is at its seasonal maximum thickness. This extensional flow has also been observed by Meng 25 and Amundson 2015 in remote sensing observations. We interpret this to indicate that the buttressing effect of mélange is limited in winter, though it likely still exerts some resistive force on the glacier termini that suppresses, though does not entirely eliminate calving activity.”

• Figure 5(b) legend of glacier position, "upper/lower" is confusing, how about "upstream/downstream"? 

We have changed this to upstream and downstream in the figure.

5) Line490: "while several models incorporating iceberg melange exist, few of them have been tested against a range of datasets or for general application in Greenland". The author should double check the references here, at least Meng2025 looks into 32 Greenland termini with 108 ArcticDEMs. 

Firstly, thank you for the reference to Meng 2025, we were not aware of this when writing our paper. We have added Meng et al 2025 and a couple of other references to the manuscript on this point. However, we are focused here on in situ datasets to resolve high time resolution calving events that satellite data cannot capture. We have updated the sentence to read:

While several models incorporating iceberg melange exist (Krug et al., 2015; Brough et al., 2023; Everett et al., 2021; Schlemm et al., 2022; Meng et al., 2024)  few of them have been tested against a range of in situ observations. Satellite data has proved immensely useful for testing these models (e.g. Meng et al., 2025) but our results also show that for detailed process understanding, the time resolution of Earth Observation data is insufficient for resolving the rapid movements of the mélange, though Meng et al (2025) demonstrate how useful it can be for generalising across termini. In situ observations of ice melange are rare and difficult and expensive to obtain and we are therefore committed to open data access to datasets such as this one, we also urge the mélange community to use such in situ observations where available.

Response to reviews for Hedetoft, Bang Brinck, Mottram et al.. Mélange or landfast ice: What controls seasonal calving at Greenland outlet glaciers?

Reviewer 1:

The authors combine in-situ tracking buoys, remote observations, CTD, and climate model data to explore the effects of melange and landfast sea ice on calving dynamics at three Greenland outlet glaciers: Tracy, Melville, and Farquhar glaciers. The data is novel and important for future investigations of Greenland calving dynamics. However, there are major concerns that need to be addressed before the manuscript to be considered for publication.

Thanks for your nice remarks, we’re happy you see the importance of this study.

Major comments:

1) Melange is defined as "a granular mixture of calved icebergs and sea ice that forms in front of glaciers". Therefore, landfast sea ice is part of melange and directly modulates melange properties. For instance, in Robel 2017 Nat Commun paper, landfast sea ice is modeled as bonding material (cohesion) in between iceberg particles within the melange. In Meng 2025 Nat Commun paper, they stated that "On the other hand, cooler ocean and air temperatures in winter enhance mélange rigidity [due to the formation of landfast sea ice], making it easier to pile up thick mélange at the terminus to provide buttressing. How warmer oceans and atmospheric influence the mélange strength is the subject of future work." This manuscript could be a great contribution on quantifying how landfast sea ice modulates melange rigidity/thickness/buttressing force. But the current title/conclusion is confusing: landfast ice and melange should not be listed as two independent factors. With landfast sea ice, melange is easier to pile up to be thicker (due to the enhanced cohesion), which provides larger buttressing force. If melange does not affect calving dynamics, it's possible that it is just not thick enough. Theoretical derivation on how melange buttressing depends on melange thickness has been conducted in the following papers:

Amundson, Jason M., and J. C. Burton. "Quasi‐static granular flow of ice mélange." Journal of Geophysical Research: Earth Surface 123, no. 9 (2018): 2243-2257.

Kavinda Nissanka, Nandish Vora, Joshua Méndez Harper, et al. Experimental Investigations of Ice Mélange and the Flow of Floating Granular Materials. ESS Open Archive . August 24, 2024.

Meng, Y., Lai, C. Y., Culberg, R., Shahin, M. G., Stearns, L. A., Burton, J. C., & Nissanka, K. (2025). Seasonal changes of mélange thickness coincide with Greenland calving dynamics. Nature Communications, 16(1), 573.

Amundson, J. M., Robel, A. A., Burton, J. C., & Nissanka, K. (2025). A quasi-one-dimensional ice mélange flow model based on continuum descriptions of granular materials. The Cryosphere, 19(1), 19-35.

Thanks for your comment, it is indeed conceptually challenging to separate out landfast sea ice from mélange, and as we note in the abstract we consider the calving front outlet glaciers as part of a continuum that in fact starts before the glacier terminus and ends well into the landfast sea ice outside of the mélange zone. As landfast sea ice also extends well beyond the region of mélange we do think it helpful to distinguish between the two categories of land fast sea ice and unbonded mélange. Our observations show that the unfrozen or unbonded mélange at these outlet glaciers has very little influence on glacier velocity or calving rates and it is only when the separate ice blocks in the mélange zone are frozen into a matrix of landfast sea ice, losing the granular material properties that are characteristic of mélange and have the material properties more associated with rigid multi-year sea ice, that they start to exert any kind of (limited) influence on the calving rates. We have therefore clarified this in the definition of mélange and land fast sea ice in subsection 1.1 and added some additional sentences to the discussion where we focus also on the differences in mélange behaviour when bonded in landfast sea ice and when free floating on modulating calving processes.

We propose slightly adjusting the title to make this clearer too. Our new proposed title is: “Mélange needs landfast sea ice to modulate seasonal calving behaviour at Greenland outlet glaciers.”

2) Figure 5 implies that melange velocity is always faster than glacier, over the 1.5yr data acquisition period. This is surprising and not consistent with the finding in Amundson2018. Can the authors add some explanation of why this is the case for the 3 reported glaciers?

We were also surprised by this feature in our analysis. We also checked against the ITS_LIVE dataset and found a very similar result. In fact, careful examination of Figure 5 (which we have replotted to make the result clearer) shows the inner mélange zone velocity is usually very close to the velocity of the lower glacier point, particularly on Melville and Tracy glaciers but with some large deviations over some periods which likely represent calving events. The buoy data shows that the average daily velocities in the mélange zone are generally low but with abrupt jumps which we interpret to be associated with calving events and the addition of these abrupt jumps likely explains the above average velocity when smoothed in the time series. It is one of the reasons that we suggest that the mélange does not exert much back stress on the glacier fronts even when held in place by land fast sea ice. Ocean measurements indicate substantial tidally driven mixing close to the mid point between Tracy and Farquhar which drives some of the movements we detect at these locations and the driving stress of the glaciers is clearly sufficient to keep the glaciers moving through the winter.

As the glacier front advances into the mélange zone the mélange is also pushed forward and has less resistance than grounded glacier ice. This also leads to the crumple zone of ice ridges, shown in figure 16 at a location around 4km in front of Melville Glacier. The landfast sea ice further out in the fjord is in general stationary, except for some tidally driven movement, so this velocity away from the glaciers does decay with distance. We have added some further sentences to the discussion of this result, to clarify this result and have also plotted up 2D velocity maps (see reply to Reviewer 2 comment also).

3) Figure 9 could be a nice place to showcase landfast sea ice is important for interpreting calving dynamics. What will happen if the author superimpose air temperature profile (below 0C as blue, above 0C as red, as done in Fig10)? Will we see terminus advance coincide with below C (more sea ice, thicker melange, larger buttressing); terminus retreat with above C (less sea ice, thinner melange, less buttressing)?

This is a nice suggestion and we have modified Figure 9 with the blue/red colours suggested and lines showing the sea ice break-up, but note that figures 6, 7 and 8 also show calving front position by month, in general all three glaciers have their maximum advanced position in April and then calve back through the spring and summer, even before the fast ice has broken up in June to reach their minimum frontal position in October.

4) Figure 11, can the authors provide summer CTD too? Are there any seasonal changes, and how do this correlate with landfast sea ice thickness? Also, any explanation on why mix layer is 100m at the mouth of the fjord (dotted red line, line 320 in text)? Can the author report more data on seasonal sea ice thickness (Line325 briefly mentions sea ice in between was consistently measured to be around 1m thick)?

Unfortunately, we do not have a year round programme of CTD observations in front of the glacier termini, though summer observations from other years in other parts of the fjord do not show a significant change compared to the profile shown here. The mixed layer depth of 100m at the mouth of the fjord is very consistent with other locations in NW Greenland and represents the Atlantic water entering the fjord. Closer to the glaciers, there is more turbulent mixing, probably driven by upwelling from the bed of the glaciers which changes the mixed layer depth. The sea ice thickness was measured systematically in a few locations during this study and was remarkably consistent both inside the mélange zone and in the fast ice area outside the mélange zone, as it has been in other years. Ongoing work in a data rescue project is compiling a large database of sea ice thickness and snow depth on sea ice observations from this region over at least the last 20 years and will be published as a dataset separately. We will add some sentences clarifying both these points in the results and discussion.

5) Figure 13, are there tidal/diurnal melange velocity signal in fall? Are the data presented here winter data, when melange is strong due to the cohesion induced by landfast sea ice, and thus less vulnerable to tidal forcing?

The buoy measurements are only possible when there is thick sea ice to travel on, and that we can use as a platform to place the instruments. We therefore do not have any autumn or early winter measurements when the ice would be thinner. The fast ice reaches maximum thickness in March/April and starts to thin from May, so the observations we have include also very thin sea ice in June where we do not however see either a diurnal or tidal signal either. We note in addition that when the landfast sea ice breaks up, both the GNSS buoys and many of the icebergs are very quickly (within a few weeks) ejected from the fjord, indicating rapid circulation through the fjord system that may also be related to tidal circulation. The movement of icebergs and mélange out of the fjord is visualised in the short timelapse videos of Sentinel-2 satellite observations linked to in the Appendix. This is a good question though so we have clarified the field deployments in the methods section.

6) After drawing conclusion that "melange does not matter for these three glaciers", can the author look into melange thickness data (i.e., ArcticDEM) and check whether melange is just too thin to do anything? The 3 fjord here are all very short. If there is not enough lateral friction from the fjord, it will be very hard to pile up thick melange, which might explain the observation here. 

We have pointed out in the discussion section that while we think these glaciers are very representative of many if not most Greenland glaciers, there are other very large outlet glaciers like Sermeq Kujalleq/Jakobshavn Isbræ and Helheim glacier, where large masses of mélange pile up within their long fjords and that this may have different properties and be able to exert more back stress, although modelling studies are not convincing on this point as the packing density is also not well observed. Mélange thickness is a very difficult concept to apply because it is very clear when standing in the mélange zone that it is a highly heterogenous material and trying to reduce to a single number or even field of “thickness” is not very helpful, at least at these glaciers, given uncertainties in fjord bathymetry and iceberg sizes. Even high resolution datasets like ArcticDEM or satellites like TerraSAR-X and Sentinel 1/2 do not resolve this heterogeneity sufficiently at these glaciers, which is very clear when in the field directly. Future work will try to address the question of mélange thickness in more detail with photogrammetry based on UAV mapping that was carried out concurrently with the buoy deployments, from which we will derive a size distribution of icebergs and ice floes that can be applied in a modelling framework. We will be sure to make all data openly available for all groups to use when it becomes available. We have added some more text on this point to the discussion.

Minor comments:

• Line 245 "there is a break point where the ... rather than the properties of the melange itself". If landfast sea ice modulates melange cohesion and in turn thickness, then it should affect the properties of the melange itself. 

Thanks that is a good catch, we have adjusted the sentence slightly to read:

“There is a break point where the ice in the mélange zone shifts between these two regimes in mid-June in both years, suggesting that the onset of enhanced calving is related to the break up of landfast sea ice in the melange zone.”

• Figure 3(b), is it better to use log scale for velocity so that we can see the baseline velocity more clearer?  Also, the buoy data is very valuable. Can you archive some of the data with full temporal resolution (10~30mins, instead of the 1-day averaged shown here) in SI or repository? It's especially useful for future study on melange dynamics during calving, where iceberg velocity can vary by 3 order of magnitude (5~5000m/day).

We have replaced the figure 3b and 4b with log-scale plots as we agree it is clearer to see the baseline velocity.

And yes we are very much in favour of open data publishing whenever possible. We are currently preparing a full data paper which will include raw data as well as the processed analysis here as well as an additional two years of similar observations. We hope to keep adding data to it as part of a monitoring programme in the fjord for as long as we can keep the programme going. This will be published in an open access repository in 2026. We are happy to share the raw data before then with others who would like to use it and have added a line to this affect in the data availability section and in the future work section.

• Line 270, extensional melange flow does not mean no buttressing. As long as the melange is thick enough, there can be enough buttressing. For instance, Meng2025 model and Amundson2018 remote observation data both show melange extensional flow during winter seasons;

We agree it does not mean no buttressing, we have slightly reformulated the sentence to read:

 “Interestingly, our results confirm that the sea ice in the melange zone and the outer fast ice zone is moving away from the glacier fronts slightly faster than the outlet glaciers are advancing, even when the landfast ice is at its seasonal maximum thickness. This extensional flow has also been observed by Meng 25 and Amundson 2015 in remote sensing observations. We interpret this to indicate that the buttressing effect of mélange is limited in winter, though it likely still exerts some resistive force on the glacier termini that suppresses, though does not entirely eliminate calving activity.”

• Figure 5(b) legend of glacier position, "upper/lower" is confusing, how about "upstream/downstream"? 

We have changed this to upstream and downstream in the figure.

5) Line490: "while several models incorporating iceberg melange exist, few of them have been tested against a range of datasets or for general application in Greenland". The author should double check the references here, at least Meng2025 looks into 32 Greenland termini with 108 ArcticDEMs. 

Firstly, thank you for the reference to Meng 2025, we were not aware of this when writing our paper. We have added Meng et al 2025 and a couple of other references to the manuscript on this point. However, we are focused here on in situ datasets to resolve high time resolution calving events that satellite data cannot capture. We have updated the sentence to read:

While several models incorporating iceberg melange exist (Krug et al., 2015; Brough et al., 2023; Everett et al., 2021; Schlemm et al., 2022; Meng et al., 2024)  few of them have been tested against a range of in situ observations. Satellite data has proved immensely useful for testing these models (e.g. Meng et al., 2025) but our results also show that for detailed process understanding, the time resolution of Earth Observation data is insufficient for resolving the rapid movements of the mélange, though Meng et al (2025) demonstrate how useful it can be for generalising across termini. In situ observations of ice melange are rare and difficult and expensive to obtain and we are therefore committed to open data access to datasets such as this one, we also urge the mélange community to use such in situ observations where available.

Please see attached file

Response to reviews for Hedetoft, Bang Brinck, Mottram et al.. Mélange or landfast ice: What controls seasonal calving at Greenland outlet glaciers?

Reviewer 1:

The authors combine in-situ tracking buoys, remote observations, CTD, and climate model data to explore the effects of melange and landfast sea ice on calving dynamics at three Greenland outlet glaciers: Tracy, Melville, and Farquhar glaciers. The data is novel and important for future investigations of Greenland calving dynamics. However, there are major concerns that need to be addressed before the manuscript to be considered for publication.

Thanks for your nice remarks, we’re happy you see the importance of this study.

Major comments:

1) Melange is defined as "a granular mixture of calved icebergs and sea ice that forms in front of glaciers". Therefore, landfast sea ice is part of melange and directly modulates melange properties. For instance, in Robel 2017 Nat Commun paper, landfast sea ice is modeled as bonding material (cohesion) in between iceberg particles within the melange. In Meng 2025 Nat Commun paper, they stated that "On the other hand, cooler ocean and air temperatures in winter enhance mélange rigidity [due to the formation of landfast sea ice], making it easier to pile up thick mélange at the terminus to provide buttressing. How warmer oceans and atmospheric influence the mélange strength is the subject of future work." This manuscript could be a great contribution on quantifying how landfast sea ice modulates melange rigidity/thickness/buttressing force. But the current title/conclusion is confusing: landfast ice and melange should not be listed as two independent factors. With landfast sea ice, melange is easier to pile up to be thicker (due to the enhanced cohesion), which provides larger buttressing force. If melange does not affect calving dynamics, it's possible that it is just not thick enough. Theoretical derivation on how melange buttressing depends on melange thickness has been conducted in the following papers:

Amundson, Jason M., and J. C. Burton. "Quasi‐static granular flow of ice mélange." Journal of Geophysical Research: Earth Surface 123, no. 9 (2018): 2243-2257.

Kavinda Nissanka, Nandish Vora, Joshua Méndez Harper, et al. Experimental Investigations of Ice Mélange and the Flow of Floating Granular Materials. ESS Open Archive . August 24, 2024.

Meng, Y., Lai, C. Y., Culberg, R., Shahin, M. G., Stearns, L. A., Burton, J. C., & Nissanka, K. (2025). Seasonal changes of mélange thickness coincide with Greenland calving dynamics. Nature Communications, 16(1), 573.

Amundson, J. M., Robel, A. A., Burton, J. C., & Nissanka, K. (2025). A quasi-one-dimensional ice mélange flow model based on continuum descriptions of granular materials. The Cryosphere, 19(1), 19-35.

Thanks for your comment, it is indeed conceptually challenging to separate out landfast sea ice from mélange, and as we note in the abstract we consider the calving front outlet glaciers as part of a continuum that in fact starts before the glacier terminus and ends well into the landfast sea ice outside of the mélange zone. As landfast sea ice also extends well beyond the region of mélange we do think it helpful to distinguish between the two categories of land fast sea ice and unbonded mélange. Our observations show that the unfrozen or unbonded mélange at these outlet glaciers has very little influence on glacier velocity or calving rates and it is only when the separate ice blocks in the mélange zone are frozen into a matrix of landfast sea ice, losing the granular material properties that are characteristic of mélange and have the material properties more associated with rigid multi-year sea ice, that they start to exert any kind of (limited) influence on the calving rates. We have therefore clarified this in the definition of mélange and land fast sea ice in subsection 1.1 and added some additional sentences to the discussion where we focus also on the differences in mélange behaviour when bonded in landfast sea ice and when free floating on modulating calving processes.

We propose slightly adjusting the title to make this clearer too. Our new proposed title is: “Mélange needs landfast sea ice to modulate seasonal calving behaviour at Greenland outlet glaciers.”

2) Figure 5 implies that melange velocity is always faster than glacier, over the 1.5yr data acquisition period. This is surprising and not consistent with the finding in Amundson2018. Can the authors add some explanation of why this is the case for the 3 reported glaciers?

We were also surprised by this feature in our analysis. We also checked against the ITS_LIVE dataset and found a very similar result. In fact, careful examination of Figure 5 (which we have replotted to make the result clearer) shows the inner mélange zone velocity is usually very close to the velocity of the lower glacier point, particularly on Melville and Tracy glaciers but with some large deviations over some periods which likely represent calving events. The buoy data shows that the average daily velocities in the mélange zone are generally low but with abrupt jumps which we interpret to be associated with calving events and the addition of these abrupt jumps likely explains the above average velocity when smoothed in the time series. It is one of the reasons that we suggest that the mélange does not exert much back stress on the glacier fronts even when held in place by land fast sea ice. Ocean measurements indicate substantial tidally driven mixing close to the mid point between Tracy and Farquhar which drives some of the movements we detect at these locations and the driving stress of the glaciers is clearly sufficient to keep the glaciers moving through the winter.

As the glacier front advances into the mélange zone the mélange is also pushed forward and has less resistance than grounded glacier ice. This also leads to the crumple zone of ice ridges, shown in figure 16 at a location around 4km in front of Melville Glacier. The landfast sea ice further out in the fjord is in general stationary, except for some tidally driven movement, so this velocity away from the glaciers does decay with distance. We have added some further sentences to the discussion of this result, to clarify this result and have also plotted up 2D velocity maps (see reply to Reviewer 2 comment also).

3) Figure 9 could be a nice place to showcase landfast sea ice is important for interpreting calving dynamics. What will happen if the author superimpose air temperature profile (below 0C as blue, above 0C as red, as done in Fig10)? Will we see terminus advance coincide with below C (more sea ice, thicker melange, larger buttressing); terminus retreat with above C (less sea ice, thinner melange, less buttressing)?

This is a nice suggestion and we have modified Figure 9 with the blue/red colours suggested and lines showing the sea ice break-up, but note that figures 6, 7 and 8 also show calving front position by month, in general all three glaciers have their maximum advanced position in April and then calve back through the spring and summer, even before the fast ice has broken up in June to reach their minimum frontal position in October.

4) Figure 11, can the authors provide summer CTD too? Are there any seasonal changes, and how do this correlate with landfast sea ice thickness? Also, any explanation on why mix layer is 100m at the mouth of the fjord (dotted red line, line 320 in text)? Can the author report more data on seasonal sea ice thickness (Line325 briefly mentions sea ice in between was consistently measured to be around 1m thick)?

Unfortunately, we do not have a year round programme of CTD observations in front of the glacier termini, though summer observations from other years in other parts of the fjord do not show a significant change compared to the profile shown here. The mixed layer depth of 100m at the mouth of the fjord is very consistent with other locations in NW Greenland and represents the Atlantic water entering the fjord. Closer to the glaciers, there is more turbulent mixing, probably driven by upwelling from the bed of the glaciers which changes the mixed layer depth. The sea ice thickness was measured systematically in a few locations during this study and was remarkably consistent both inside the mélange zone and in the fast ice area outside the mélange zone, as it has been in other years. Ongoing work in a data rescue project is compiling a large database of sea ice thickness and snow depth on sea ice observations from this region over at least the last 20 years and will be published as a dataset separately. We will add some sentences clarifying both these points in the results and discussion.

5) Figure 13, are there tidal/diurnal melange velocity signal in fall? Are the data presented here winter data, when melange is strong due to the cohesion induced by landfast sea ice, and thus less vulnerable to tidal forcing?

The buoy measurements are only possible when there is thick sea ice to travel on, and that we can use as a platform to place the instruments. We therefore do not have any autumn or early winter measurements when the ice would be thinner. The fast ice reaches maximum thickness in March/April and starts to thin from May, so the observations we have include also very thin sea ice in June where we do not however see either a diurnal or tidal signal either. We note in addition that when the landfast sea ice breaks up, both the GNSS buoys and many of the icebergs are very quickly (within a few weeks) ejected from the fjord, indicating rapid circulation through the fjord system that may also be related to tidal circulation. The movement of icebergs and mélange out of the fjord is visualised in the short timelapse videos of Sentinel-2 satellite observations linked to in the Appendix. This is a good question though so we have clarified the field deployments in the methods section.

6) After drawing conclusion that "melange does not matter for these three glaciers", can the author look into melange thickness data (i.e., ArcticDEM) and check whether melange is just too thin to do anything? The 3 fjord here are all very short. If there is not enough lateral friction from the fjord, it will be very hard to pile up thick melange, which might explain the observation here. 

We have pointed out in the discussion section that while we think these glaciers are very representative of many if not most Greenland glaciers, there are other very large outlet glaciers like Sermeq Kujalleq/Jakobshavn Isbræ and Helheim glacier, where large masses of mélange pile up within their long fjords and that this may have different properties and be able to exert more back stress, although modelling studies are not convincing on this point as the packing density is also not well observed. Mélange thickness is a very difficult concept to apply because it is very clear when standing in the mélange zone that it is a highly heterogenous material and trying to reduce to a single number or even field of “thickness” is not very helpful, at least at these glaciers, given uncertainties in fjord bathymetry and iceberg sizes. Even high resolution datasets like ArcticDEM or satellites like TerraSAR-X and Sentinel 1/2 do not resolve this heterogeneity sufficiently at these glaciers, which is very clear when in the field directly. Future work will try to address the question of mélange thickness in more detail with photogrammetry based on UAV mapping that was carried out concurrently with the buoy deployments, from which we will derive a size distribution of icebergs and ice floes that can be applied in a modelling framework. We will be sure to make all data openly available for all groups to use when it becomes available. We have added some more text on this point to the discussion.

Minor comments:

• Line 245 "there is a break point where the ... rather than the properties of the melange itself". If landfast sea ice modulates melange cohesion and in turn thickness, then it should affect the properties of the melange itself. 

Thanks that is a good catch, we have adjusted the sentence slightly to read:

“There is a break point where the ice in the mélange zone shifts between these two regimes in mid-June in both years, suggesting that the onset of enhanced calving is related to the break up of landfast sea ice in the melange zone.”

• Figure 3(b), is it better to use log scale for velocity so that we can see the baseline velocity more clearer?  Also, the buoy data is very valuable. Can you archive some of the data with full temporal resolution (10~30mins, instead of the 1-day averaged shown here) in SI or repository? It's especially useful for future study on melange dynamics during calving, where iceberg velocity can vary by 3 order of magnitude (5~5000m/day).

We have replaced the figure 3b and 4b with log-scale plots as we agree it is clearer to see the baseline velocity.

And yes we are very much in favour of open data publishing whenever possible. We are currently preparing a full data paper which will include raw data as well as the processed analysis here as well as an additional two years of similar observations. We hope to keep adding data to it as part of a monitoring programme in the fjord for as long as we can keep the programme going. This will be published in an open access repository in 2026. We are happy to share the raw data before then with others who would like to use it and have added a line to this affect in the data availability section and in the future work section.

• Line 270, extensional melange flow does not mean no buttressing. As long as the melange is thick enough, there can be enough buttressing. For instance, Meng2025 model and Amundson2018 remote observation data both show melange extensional flow during winter seasons;

We agree it does not mean no buttressing, we have slightly reformulated the sentence to read:

 “Interestingly, our results confirm that the sea ice in the melange zone and the outer fast ice zone is moving away from the glacier fronts slightly faster than the outlet glaciers are advancing, even when the landfast ice is at its seasonal maximum thickness. This extensional flow has also been observed by Meng 25 and Amundson 2015 in remote sensing observations. We interpret this to indicate that the buttressing effect of mélange is limited in winter, though it likely still exerts some resistive force on the glacier termini that suppresses, though does not entirely eliminate calving activity.”

• Figure 5(b) legend of glacier position, "upper/lower" is confusing, how about "upstream/downstream"? 

We have changed this to upstream and downstream in the figure.

5) Line490: "while several models incorporating iceberg melange exist, few of them have been tested against a range of datasets or for general application in Greenland". The author should double check the references here, at least Meng2025 looks into 32 Greenland termini with 108 ArcticDEMs. 

Firstly, thank you for the reference to Meng 2025, we were not aware of this when writing our paper. We have added Meng et al 2025 and a couple of other references to the manuscript on this point. However, we are focused here on in situ datasets to resolve high time resolution calving events that satellite data cannot capture. We have updated the sentence to read:

While several models incorporating iceberg melange exist (Krug et al., 2015; Brough et al., 2023; Everett et al., 2021; Schlemm et al., 2022; Meng et al., 2024)  few of them have been tested against a range of in situ observations. Satellite data has proved immensely useful for testing these models (e.g. Meng et al., 2025) but our results also show that for detailed process understanding, the time resolution of Earth Observation data is insufficient for resolving the rapid movements of the mélange, though Meng et al (2025) demonstrate how useful it can be for generalising across termini. In situ observations of ice melange are rare and difficult and expensive to obtain and we are therefore committed to open data access to datasets such as this one, we also urge the mélange community to use such in situ observations where available.