the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Feb 2026
Interannual and Seasonal Evolution of Supraglacial Channel Networks on Nivlisen Ice Shelf, East Antarctica
Abstract. Supraglacial channel networks (SCNs) represent critical components of the surface drainage system on Antarctic ice shelves, facilitating the redistribution of meltwater, including into and out of supraglacial lakes, potentially influencing ice-shelf stability. However, the seasonal and interannual variability in their spatial extent, structural evolution, and climatic sensitivity remain poorly quantified, particularly compared to observations of supraglacial lakes. In this study, we investigate the interannual evolution of SCNs on Nivlisen Ice Shelf in East Antarctica between 2017 and 2024 using Sentinel-2 and Landsat-8 satellite imagery, with a detailed seasonal analysis during the 2018–2019 melt season. Key metrics including surface water area, total channel length, number of SCNs, and the number of individual channels were derived to quantify the drainage evolution and system connectivity and complexity. We find significant interannual variability in SCN count (from 34 to 186) and total channel length (from 345 to 1275 km), with peak length of channels observed in 2018. During the 2018–2019 melt season SCNs progressively integrated and connected to form an efficient drainage system, which then fragmented towards the end of the melt season. We find that SCNs are strongly influenced by the ice shelf surface structures and ice surface undulations at a range of scales, with larger channels reforming in the same location. Correlation analysis shows that annual SCN evolution is tightly coupled to regional air temperature, with surface water extent and network connectivity strongly linked to reanalysis-based temperature outputs (r > 0.9, p < 0.05). Snow had variable effects, depending on its timing and how readily it could be melted. Our findings underscore the important dynamic nature of supraglacial channels in redistributing meltwater on Antarctic ice shelves.
- Preprint
(4900 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-5562', Amber Leeson, 16 Apr 2026
-
RC2: 'Comment on egusphere-2025-5562', Anonymous Referee #2, 21 Apr 2026
This paper uses remotes sensing and the output of a regional climate model and a reanalysis product to examine the surface drainage system that forms on the Nivlisen Ice Shelf. The authors map the channel network over eight melt seasons, derived metrics related to development of the drainage system, and compare these to weather conditions. They also examine the evolution of the drainage system over one melt season (2018-2019).
The paper could be a useful extension to the growing literature on similar systems and different aspects of the same system. This work is novel in its detailed focus on surface channel networks on Nivlisen; other papers have worked on Nivlisen Ice Shelf, but have either performed only a preliminary analysis of the drainage system or have focused on the lakes rather than the whole drainage network, while others have performed detailed analysis of drainage systems on other ice shelves. As such, I consider the current work a relatively minor advance over those previous efforts. It has the potential to be a useful addition nonetheless.
A couple of main comments are below that the authors should consider before the paper can be published in my opinion, followed by more minor commentsMain comments
Efficiency is mentioned in the abstract and in several other places in the manuscript. However, I do not think the paper presents any direct observations of efficiency. In fact, I am not sure what efficient means in this context. It seems like the way it is being used in this paper could be close to ‘well-connected’. Is that correct? If so, maybe use a term like that which is more closely tied to the observations, rather than efficient, which is something that is vaguely associated with how fast/easily water moves through a system, which of course we do not know from these remote sensing observations alone. Alternatively, a clearer argument should be made linking the observations to efficiency specifically, but if the route is taken, I would like to see this link introduced, proposed and discussed in a critical way, i.e. in a way that considers the arguments for and against linking the observations to efficiency in this way.There is little discussion of limitations. A major limitation that has the potential to undermine many of the inferences made from derived statistical relationships is the very short times series analyzed. Only eight melt annual snapshots of the drainage systems are considered and only eight snapshots of the system are considered during the analysis of seasonal evolution during the 2018-19 season. This must be critically discussed.
Line-by-line comments
Line 38: An additional relevant citation is Spergel et al 2021
L49: I am not sure that an increase in hydrostatic pressure is a good summary of the stress changes involved in the mechanism proposed in Banwell et al. (2013). I suggest editing this sentence.
L50: Point 3 is not an example of a primary mechanism of hydrofracturing, as is suggested by its presence in this list.
L56: Lai et al. (2020) is a relevant citation here. Particularly given that they talk about spatial variations in ice-shelf vulnerability to meltwater and SCNs move water around.
L104: ‘during the past 20 years’, Kingslake et al. (2015) only mentioned the 2002-2009 in their paper
L111: Is this the citation for the speed, and coastline and GL products? If not, add all required citations.
L122: Delete ‘very much’.
L136: Delete ‘by’.
L142: Would it be simpler to just use ‘lake’ instead of SGL and ‘networks’ instead of SCNs?
L152-153: “Furthermore, while observations suggest that these networks often spatially merge into a small number of large channels that may terminate in a large lake” If this is something that the previous work highlights, I suggest noting that above when you discuss the individual papers. This will strengthen the case for focusing on this question specifically in this paper.
L155: The last sentence of this paragraph takes on an urgent tone that seems out of place and is unsupported by the preceding literature review. It is unclear from reading the preceding text what is urgent about the need to quantify channels on Nivlisen specifically. Is there an indication that collapse of the Nivlisen is imminently if meltwater delivers water to specific areas, for example?
L190: imagery —> image
L190: There is a change in tense here.
L203: Rephrase “Red band on optical imagery”
L239: delete ‘the’
L241: Do not capitalize “Density”
Table 1: replace ’n’ in the units column with ‘-‘ or something similar to indicate no units.
L277: ‘stands for’ ‘is’
L293-294: Is there a typo here or is it the ‘accumulated melt’ This seems a bit confusing to combine ‘melt’ and ‘accumulation’.
L371-372: Is there a missing word in this sentence?
L375-377: I thought a decrease in the number of networks would be consistent with less fragmented and more connected networks. This sentence seems to say the opposite. Please elaborate.
L456: This reads as saying that the mean depth became more uniform. I am not sure I understand this. Are you saying that the depth became more uniform spatially? If so, rephrase this and point to a different figure which quantifies this.
L469-470: This phrase implies channels connect the GL and the ice-shelf front. This is not the case. Rephrase.
Figure 12: The key is very confusing. I suggest using a normal colorbar. Typo in the caption.
L490: “along the northwest direction” Awkward phrasing.
Figure 13: Do the colors match between the large panel, the key and the other panels? The thick green lines that extend the furthest downstream to the location of the large terminal lake do not appear the same as the lines in (e) and (i), which I think are intended to be the same color.
L507: ‘Particularly the narrow tributaries’: I am not sure what this is referring to.
Figure 14b: The horizontal axis label should be latitude not longitude.
L574-576: I am not sure that I understand/agree with this statement. Comparing Figures 17a, b and d, suggests quite a bit of disagreement between the melt time series and the PDD time series.
For example, while all three peak then drop to lower levels, the PDD time series drops from near its peak significantly later than both the melt time series.
Can you expand upon this point, adjust the figures to make the comparison easier to see, and discuss the similarities and differences in more detail.
L622-624: Or could these backscatter values suggest a refrozen unit that acts as specular reflectors that reflect away the radar energy?
L639: Is it an ice rise? Or an ice rumple? It seems very small for an ice rise, which require a flow center independent of the ice-shelf flow. In fact, looking at figure 1, it seems unlikely that the flow speed goes to zero here, which is what is required to make this an ice rise.
L657: “Although individual channels do not evolve at the same rate”: Does this mean they do not evolve at the same rate throughout the season, or as each other? In addition to this confusion, I am not sure how this phrase relates to the rest of the sentence.
L666: I do not think these are the most appropriate references for meltwater patterns on Larsen C Ice Shelf?
L726-727: I am not sure how useful it is to be speculating on physical processes to explain a weak correlation which is not statistically significant.
L731: “less energy is required to melt snow than it is to melt ice” This is true for a given depth of snow vs the same depth of ice, but all quantities are in mm w.e. so I am not sure I understand this statement. In fact, fresh snow increases albedo, reducing melt compared to the same weather conditions over ice. This needs more discussion or removal.L760: I do not think this is the best citation for this sentence about how channel formation should precede lake formation. The paper cited here is focused on the opposite - lakes feeding channels.
L767-768: This sentence seems to imply that channels operate all the time. In contrast, they must either empty or their water must freeze during the winter months.
L772: I think this statement is based on an argument made earlier that I do not understand that snow is “less energy is required to melt snow than it is to melt ice”. This needs revisiting in my opinion. The argument should be elaborated on or removed. Given that the correlations this agreement is trying to explain are found not to be statistically significant I think this can mostly be removed without losing much of value.
L793: “(>100%)” it is unclear what this means precisely.
L794: should interannual be seasonal?
L798: “near the ice front”: This is qualitative. Personally, I wouldn’t consider any of the terminal lakes to be ‘near the ice front’ but it is subjective, so quantify this or remove it.
Data availability section: I strongly think these data should be archived and freely available, rather than available upon request.
Lai, C.-Y., Kingslake, J., Wearing, M. G., Chen, P.-H. C., Gentine, P., Li, H., Spergel, J. J., and Van Wessem, J. M.: Vulnerability of Antarctica’s ice shelves to meltwater-driven fracture, Nature, 584, 574–578, https://doi.org/10.1038/s41586-020-2627-8, 2020.
Spergel, J. J., Kingslake, J., Creyts, T., Van Wessem, M., and Fricker, H. A.: Surface meltwater drainage and ponding on Amery Ice Shelf, East Antarctica, 1973–2019, J. Glaciol., 67, 985–998, https://doi.org/10.1017/jog.2021.46, 2021.Citation: https://doi.org/10.5194/egusphere-2025-5562-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 156 | 86 | 14 | 256 | 10 | 29 |
- HTML: 156
- PDF: 86
- XML: 14
- Total: 256
- BibTeX: 10
- EndNote: 29
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript quantifies the interannual (2017–2024) and seasonal (2018–2019 melt season) evolution of supraglacial channel networks (SCNs) on the Nivlisen Ice Shelf, East Antarctica, using Sentinel-2 and Landsat-8 imagery. Specifically, the authors derive key network metrics including surface water area, total channel length, number of SCNs and number of individual channels. They report strong interannual variability, and demonstrate progressive seasonal integration and late-season fragmentation during 2018–2019. Correlation analyses indicate strong coupling between SCN metrics and broad scale climate.
The study directly addresses a knowledge gap: while supraglacial lakes (SGLs) on Nivlisen (and indeed in Antarctica more broadly) have been extensively studied, SCNs have not been systematically quantified. This is important because SCN can efficiently route surface meltwater laterally across an ice shelf and, in some cases, off the ice front, thereby reducing the residence time of water on the surface (e.g. Leeson et al., 2020). By limiting widespread ponding, SCNs may reduce the potential for surface water to infiltrate and densify the firn layer, which would otherwise enhance future meltwater retention and decrease pore space. At the same time, the spatial focusing of meltwater within channels and connected lakes could locally increase hydrostatic loading and fracture propagation, leading to instability.
More work is required to understand SCNs and their role in the potential fate of Antarctica’s ice shelves. The manuscript therefore makes a clear and valuable contribution to Antarctic ice-shelf hydrology and I recommend publication following revisions. Specific comments can be found below.
1. Methodological clarity
I find the use of use of ‘network’ and ‘channel’ a bit confusing. The manuscript defines SCNs as ‘drainage networks including both lakes and channels’ however, it is missing a definition of network. Fig 3 suggests that a single channel is classed as a network, if it is not connected to another, but this seems counterintuitive and at odds with the definition of SCNs.
I’m also not clear why NDWI is used since it seems that this is augmented with manual delineation. Why not exclusively manual delineation given the small scale of data?
I also suggest that you give water depth as relative rather than absolute given high uncertainty in depth detection methods (e.g. Melling et al., 2024).
2. Topographic dependence
The distribution of features in space is explained with reference to local topography and firn conditions. I agree topography is likely the dominant control, and would like to see all your elevation maps given on a log scale so that small scale variations on the ice shelf can be seen in addition to broad scale grounded-to-floating gradients.
3. Interannual variability.
I understand that S2 was used for the interannual assessment and L8 was used for the seasonal comparison intentionally so that you are not mixing sensors, however I wonder if a bit of mixing might be useful. The interannual variability you see in your dataset is striking, and similar to signals that we see elsewhere in Antarctica (e.g. GVIIS, Barnes et al., 2020). It would be interesting to know how far you can go back, and what you might find, if you brought Landsat in for the inter-annual comparison. 8 years is not so very many given the high degree of variability you see. Even if this is qualitative, it would be interesting to know if the patterns you see in 2019 and 2020 are unique to that epoch, or have occurred regularly in the past. On GVIIS for example, extensive meltponding has been observed as far back as the 1970s.
Figure 8a seems to show frozen lake area downstream of the open-water lake. Could this be a remnant from the previous year?
4. Seasonal evolution.
It would also be nice to see a comparison of contemporaneous images from both L8 and S2 to get a feel for what might be missing from the seasonal comparison. Since channels narrower than 20-30 m may be unresolved or only partially resolved in Landsat, what uncertainty might this introduce into your findings, especially in lower melt years?
The discussion around seasonal evolution is interesting, although I would like to see greater clarity (and possibly stronger caveats) applied to the more speculative arguments including the ‘unstable drainage’ paragraph. I note that we see a transfer of water downstream on LBIS in the months before collapse (Leeson et al., 2020, Figure 5) so this seems plausible. I don’t understand however why lakes would discharge laterally, and then refill later given that we would probably assume the overflow channel to persist throughout the remainder of the met season.
5. Role of climate
A lot of weight is put on the correlation analysis between hydrological parameters and climate variables. I think this is perhaps unjustified given the short (8 year) time series and that as a consequence this could be streamlined. It is more likely that broader scale climatic patterns as indicated by ERA5 temp would show a statistically significant correlation on these timescales than specific small scale signals e.g. MAR snowfall. It is also probable that because of the short time series the correlation analysis is sensitive to single year outliers (e.g. 2019). I’d be interested to see you include the melt-over-accumulation ratio in your analysis (van Wessem et al., 2023) and a map of snow/firn depth somewhere (perhaps in fig 1).
6. Implications?
The study frames the importance of SCNs as a mechanism for redistribution of meltwater on Antarctic ice shelves and, as a corollary, stabilising or destabilising them. It doesn’t however then go on to comment on what the implications of their specific findings are for the future stability of the Nivlisen ice shelf, which is what I would expect from such framing. The authors may feel their contribution is more useful in terms of new process level understanding than impacts of stability. This is of course fine, and very useful, but means that the framing needs to be adjusted accordingly.
7. Data Availability and Reproducibility
Data availability statement indicates mapped SCNs are available upon request, I suggest that these are instead deposited in a public repository (e.g., Zenodo) upon publication.
New References:
Barnes, T. J., Leeson, A. A., McMillan, M., Verjans, V., Carter, J., and Kittel, C.: Changes in Supraglacial Lakes on George VI Ice Shelf, Antarctic Peninsula: 1973–2020, The Cryosphere Discuss. [preprint], https://doi.org/10.5194/tc-2021-214, 2021.
Leeson, A. A., Forster, E., Rice, A., Gourmelen, N., & vanWessem, J. M. (2020). Evolution of supraglacial lakes on the Larsen B ice shelf in the decades before it collapsed. Geophysical Research Letters, 47, e2019GL085591. https://doi.org/10.1029/2019GL085591
Melling, L., Leeson, A., McMillan, M., Maddalena, J., Bowling, J., Glen, E., Sandberg Sørensen, L., Winstrup, M., and Lørup Arildsen, R.: Evaluation of satellite methods for estimating supraglacial lake depth in southwest Greenland, The Cryosphere, 18, 543–558, https://doi.org/10.5194/tc-18-543-2024, 2024.
van Wessem, J.M., van den Broeke, M.R., Wouters, B. t al. Variable temperature thresholds of melt pond formation on Antarctic ice shelves. Nat. Clim. Chang. 13, 161–166 (2023). https://doi.org/10.1038/s41558-022-01577-1