Seasonal evolution of the supraglacial drainage network at Humboldt Glacier, North Greenland, between 2016 and 2020&nbsp;

Rawlins, Lauren D.; Rippin, David M.; Sole, Andrew J.; Livingstone, Stephen J.; Yang, Kang

doi:https://doi.org/10.5194/tc-2023-23

Preprints

https://doi.org/10.5194/tc-2023-23

Preprints

22 Feb 2023

| 22 Feb 2023

Status: a revised version of this preprint was accepted for the journal TC and is expected to appear here in due course.

Seasonal evolution of the supraglacial drainage network at Humboldt Glacier, North Greenland, between 2016 and 2020

Lauren D. Rawlins, David M. Rippin, Andrew J. Sole, Stephen J. Livingstone, and Kang Yang

Abstract. Supraglacial rivers and lakes are important for the routing and storage of surface meltwater during the summer melt season across the Greenland Ice Sheet (GrIS), yet remain poorly mapped and quantified across the northern part of the ice sheet, which is rapidly losing mass. Here we produce, for the first time, a high-resolution record of the supraglacial drainage network (including both rivers and lakes) and its seasonal behaviour at Humboldt Glacier, a wide-outlet glacier draining a large hydrologic catchment (13,488 km²), spanning the period 2016 to 2020 using 10 m spatial resolution Sentinel-2 imagery. Our results reveal a perennially extensive yet interannually-variable supraglacial network extending from an elevation of 200 m a.s.l to a maximum of ~1440 m a.s.l recorded in 2020, with limited development of the network observed in the low melt years of 2017 and 2018. The supraglacial drainage network is shown to cover an area ranging between 965.7 km² (2018) and 1566.3 km² (2019) at its maximum seasonal extent, with spatial coverage of up to 2685 km² recorded during the early phases of the melt season when a slush zone is most prominent. Up-glacier expansion and the development of an efficient supraglacial drainage network as surface runoff increases and the snowline retreats is clearly visible. Preconditioning of the ice surface following a high melt year is also observed, with the earlier widespread exposure of the supraglacial drainage network in 2020 compared to other years; a finding that may become representative with persistent warmer years into the future. Overall, this study provides evidence of a persistent, yet dynamic, supraglacial drainage network at this prominent northern GrIS outlet glacier and advances our understanding of such hydrologic processes, particularly under ongoing climatic warming and enhanced runoff.

Received: 10 Feb 2023 – Discussion started: 22 Feb 2023

Download & links

Preprint (PDF, 3071 KB)

Supplement (3472 KB)

Download & links

Lauren D. Rawlins et al.

Status: closed

RC1:
'Comment on tc-2023-23', Anonymous Referee #1, 22 Mar 2023

Rawlins et al TC 23
This is a super well written, clear, thoroughly referenced (rare!), and interesting paper. I don’t get many of these to read, so this was a delight. I do have some suggestions below for how the paper could be improved, but they are all easily done and I think will make the paper stronger. I am very familiar with the subject matter and literature here, so my eyes glaze over on the methods sections and I took them for granted.
Improvements needed:
Slush definition- Slush zones are a key part of these results, but the authors have not described how they were defined/mapped/identified. I’d like to see their definition of a slush zone and how they uniquely identified them in images from other supraglacial features.
MAR uncertainty- I do not think the authors need an ensemble of models, but I would expect to see some discussion of the fact that MAR (or any coupled or uncoupled atmos-ice model) is our least bad representation of reality. There are known issues in this sort of modelling, and errors in the model will only strengthen your conclusions by potentially reducing scatter. Figure 5, for instance, should have a MAR uncertainty plotted on it, perhaps as a confidence interval. Discussion should also be added.
MF uncertainty- as above, please discuss uncertainties in your MF. You have spatial errors resulting from pixel size, spatial errors from undetected sub-S2 channels, DEM resolution errors, and errors of classification that might omit/comit water area. This seems a larger omission from the paper- I (and readers) want to know where the method is good and where it is not. Since you aren’t about to map a scene manually to provide true validation (although I wouldn’t object to that!), I think you can propagate the variance from each of those terms based on the literature surrounding your classification methods. I think this information is essential to this paper.
Drainage density and other stats- you have the data to create a vector river network and determine drainage density, stream orders, and other stats for comparison (see last point). Author Yang has published many papers on this topic and therefore I believe this should be a straightforward task that would add needed richness to this paper in terms of comparison.
Split MF into river and lake areas- I am quite interested in this divide. This is figure S3, but for me this belongs in the main text as a very interesting expression of the supraglacial hydrology here. Some discussion should also occur.
Section 5.4- I am ok with this section as it is fairly hedged and well referenced, but other reviewers may not like to see such speculative conjecture.
Missing comparisons- the 2^nd major omission I see (beyond uncertainty discussion noted above) is a lack of comparison to the rich literature of the SW GrIS. This is the right paper to use the discussion to first outline this bit of ice sheet (as you have done) and then explicitly compare to the SW to see what is the same and what is different- are the fractions of rivers and lakes the same? Elevations of highest melt features? Density of persistent channels? Channel lengths? Width distributions of these channels? Prevalence of slush zones/bare ice and their interaction with the network? I think you have all the data to answer those questions (and more) and I think this paper really needs it to move this beyond an interesting and well written observational study into a richer contextual understanding of this unique bit of ice. I’d like to see these differences quantified where possible (e.g. from a vector network) or discussed qualitatively and referenced where not possible (as you have nicely done throughout this paper!!)

Citation: https://doi.org/10.5194/tc-2023-23-RC1
- AC1: 'Reply on RC1', Lauren Rawlins, 24 Aug 2023
  
  We would like to thank the reviewer for their positive and insightful comments. Please see attached for our responses.
  
  Citation: https://doi.org/10.5194/tc-2023-23-AC1
RC2:
'Review on tc-2023-23', Anonymous Referee #2, 04 Apr 2023

Review for "Seasonal evolution of the supraglacial drainage network at Humboldt Glacier, North Greenland, between 2016 and 2020" by Rawlins et al.
General comments
This paper was a pleasure to read as it is written so well. I only have some minor comments (below), followed by minor specific line by line comments.
The incorporation of slush mapping is very interesting and novel, as I am not aware of prior studies that have mapped slush in detail Greenland(?), though this has been done by at least one study on Antarctic ice shelves (e.g. Dell et al 2022). So it would be great to see more details regarding how exactly slush was mapped in this study. Additionally, I think slush (e.g. as a component of total meltwater area fraction (MF)), could feature more heavily as a discussion point in the Results/Discussion.
I do not know Humboldt Glacier well, so I wondering how much of this marine terminating glacier is floating? Unless I missed it, the only mention of floating ice is on line 208, where a ‘7 km floating section’ is mentioned. Depending on the area of this floating area of ice, I’d also be particularly interested to know whether the authors see a difference between lake/other meltwater feature characteristics on the floating versus grounded portions of the glacier? (For example, comparisons of meltwater features on floating vs. grounded ice have previously been made on Peterman Glacier (Macdonald et al 218) and well as Paakitsoq (SW Greenland) vs. Larsen B Ice shelf, Antarctica (Banwell et al 2014).
I am wondering why the authors choose to base the NDWI on the Green and NIR bands rather than the blue and red bands, e.g. as used by Bell et al (2017) and Williamson et al (2018). I am sure there were/are good reasons, but perhaps a sentence or two about this could be added to the paper. Also related to the use of the NDWI, I am wondering how/why the authors decided to use a threshold of 0.4, i.e. which they call a ‘high-value global NDWI threshold’? Again, I am not suggesting this threshold is not appropriate, but perhaps some more detail and/or reference(s) be added about this choice of value?
My final general comment is that I think the authors should add a short paragraph about uncertainty quantification, particularly regarding their MF analysis.
Specific line by line comments
Abstract lines 16 and 17: Not necessary to state areas to 1 d.p. Round up as is done for other area/elevation values in the abstract.
Abstract line 19: I suggest adding an few extra words to explain what you mean by ‘preconditioning’ here.
56 – 57: Two relevant studies focusing on the surface hydrology of Petermann Glacier in NW Greenland could also be referenced here: Macdonald et al (2018) and Boghosian et al (2021).
133: Gledhill and Williamson just have a 2017 paper, but Williamson et al. have a 2018 paper (both are already in the reference list).
241 – 247: I find this paragraph confusing regarding the maximum extent of meltwater mapped, versus the maximum extent of the study region. For example, the first sentence says that rivers and lakes were mapped up to a “maximum melt extent of 1500 m a.s.l.”. Is 1500 m the highest elevation of the study area analyzed, or is this the maximum elevation of observed melt? I assume the latter(?), but if so, then the following sentence is repetitive (i.e. this states “The mapped supraglacial drainage network across HG is shown extend up to 1500 m a.s.l,”). Also, later in the paragraph (line 246), it says rivers and lakes from up to a “maximum of 1440 m” (as opposed to 1500 m). So these sentences need to be re-written for clarity.
253 - 255: For the sentence “In Figure 3b, we also see some evidence of a potential main-river reconfigurations, with the north-westward advection of a river channel that runs transverse to ice flow”; maybe this river in Fig 3b could be labelled? As I see various rivers/streams that are transverse to ice flow. Also, it looks to me as though similar examples may also be seen in panels c and e?
310 – 314: Can the authors suggest a possible explanation for why these two parallel lines that track across glacier exist? Could they be fractures?
497 - 499: For the sentence:”… this study also notes that many well-established rivers that are longitudinal to ice flow, including many with canyonised features, also reoccupy locations.”, studies focused on Petermann Glacier could also be mentioned here, which found similar findings I believe (Macdonald et al. 2018, Boghosian et al. 2021).
622/623: Mention Summer 2019 somewhere in this sentence to remind the reader which melt season is being described.
Figures
Fig 2: it would be interesting to know the locations of these figure panels, so perhaps an extra panel could be added to show this (e.g. as is done in Fig 3a), or perhaps the locations should be shown somewhere in Fig 1? Also, the ‘off edge river termination’ feature in panel 2c) is interesting, and I’m wondering how comparable this feature could be to the large river/waterfall described in Bell et al (2017)?
References (those in bold are not referenced in the current paper)
Banwell, A.F., Cabellero, M., Arnold, N., Glasser, N., Cathles, L.M., MacAyeal, D. 2014. Supraglacial lakes on the Larsen B Ice Shelf, Antarctica, and Paakitsoq Region, Greenland: a comparative study. Annals of Glaciology. 55(66), doi:10.3189/2014AoG66A049.
Bell, R. E., Chu, W., Kingslake, J., Das, I., Tedesco, M., Tinto, K. J., Zappa, C. J., Frezzotti, M., Boghosian, A., and Lee, W. S.: Antarctic ice shelf potentially stabilized by export of meltwater in surface river, Nature, 544, 344–348, 2017.
Boghosian, A.L., Pitcher, L.H., Smith, L.C. et al. Development of ice-shelf estuaries promotes fractures and calving. Nature Geoscience, 14, 899–905 (2021). https://doi.org/10.1038/s41561-021-00837-7
Dell, R., Banwell. A.F., Willis, I., Arnold, N., Halberstadt, A.R.W., Chudley, T.R., Pritchard, H. 2022, Supervised classification of slush and ponded water on Antarctic ice shelves using Landsat 8 imagery, Journal of Glaciology, 1–14. https://doi.org/10.1017/ jog.2021.114.
Gledhill, L. A. and Williamson, A. G.: Inland advance of supraglacial lakes in north-west Greenland under recent climatic warming, Annals of Glaciology, 59, 66-82, https://doi.org/10.1017/aog.2017.31, 2018.
Macdonald, G.J., Banwell, A.F., MacAyeal, D.R. 2018, Seasonal evolution of supraglacial lakes on a floating ice tongue, Petermann Glacier, Greenland. Annals of Glaciology, doi:10.1017/aog.2018.9
Williamson, A. G., Banwell, A. F., Willis, I. C., and Arnold, N. S.: Dual-satellite (Sentinel-2 and Landsat 8) remote sensing of supraglacial lakes in Greenland, The Cryosphere, 12, 3045-3065, https://doi.org/10.5194/tc-12-3045-2018, 2018.

Citation: https://doi.org/10.5194/tc-2023-23-RC2
- AC2: 'Reply on RC2', Lauren Rawlins, 24 Aug 2023
  
  We would like to thank the reviewer for their positive and insightful comments. Please see attached for our responses
  
  Citation: https://doi.org/10.5194/tc-2023-23-AC2

Status: closed

RC1:
'Comment on tc-2023-23', Anonymous Referee #1, 22 Mar 2023

Rawlins et al TC 23
This is a super well written, clear, thoroughly referenced (rare!), and interesting paper. I don’t get many of these to read, so this was a delight. I do have some suggestions below for how the paper could be improved, but they are all easily done and I think will make the paper stronger. I am very familiar with the subject matter and literature here, so my eyes glaze over on the methods sections and I took them for granted.
Improvements needed:
Slush definition- Slush zones are a key part of these results, but the authors have not described how they were defined/mapped/identified. I’d like to see their definition of a slush zone and how they uniquely identified them in images from other supraglacial features.
MAR uncertainty- I do not think the authors need an ensemble of models, but I would expect to see some discussion of the fact that MAR (or any coupled or uncoupled atmos-ice model) is our least bad representation of reality. There are known issues in this sort of modelling, and errors in the model will only strengthen your conclusions by potentially reducing scatter. Figure 5, for instance, should have a MAR uncertainty plotted on it, perhaps as a confidence interval. Discussion should also be added.
MF uncertainty- as above, please discuss uncertainties in your MF. You have spatial errors resulting from pixel size, spatial errors from undetected sub-S2 channels, DEM resolution errors, and errors of classification that might omit/comit water area. This seems a larger omission from the paper- I (and readers) want to know where the method is good and where it is not. Since you aren’t about to map a scene manually to provide true validation (although I wouldn’t object to that!), I think you can propagate the variance from each of those terms based on the literature surrounding your classification methods. I think this information is essential to this paper.
Drainage density and other stats- you have the data to create a vector river network and determine drainage density, stream orders, and other stats for comparison (see last point). Author Yang has published many papers on this topic and therefore I believe this should be a straightforward task that would add needed richness to this paper in terms of comparison.
Split MF into river and lake areas- I am quite interested in this divide. This is figure S3, but for me this belongs in the main text as a very interesting expression of the supraglacial hydrology here. Some discussion should also occur.
Section 5.4- I am ok with this section as it is fairly hedged and well referenced, but other reviewers may not like to see such speculative conjecture.
Missing comparisons- the 2^nd major omission I see (beyond uncertainty discussion noted above) is a lack of comparison to the rich literature of the SW GrIS. This is the right paper to use the discussion to first outline this bit of ice sheet (as you have done) and then explicitly compare to the SW to see what is the same and what is different- are the fractions of rivers and lakes the same? Elevations of highest melt features? Density of persistent channels? Channel lengths? Width distributions of these channels? Prevalence of slush zones/bare ice and their interaction with the network? I think you have all the data to answer those questions (and more) and I think this paper really needs it to move this beyond an interesting and well written observational study into a richer contextual understanding of this unique bit of ice. I’d like to see these differences quantified where possible (e.g. from a vector network) or discussed qualitatively and referenced where not possible (as you have nicely done throughout this paper!!)

Citation: https://doi.org/10.5194/tc-2023-23-RC1
- AC1: 'Reply on RC1', Lauren Rawlins, 24 Aug 2023
  
  We would like to thank the reviewer for their positive and insightful comments. Please see attached for our responses.
  
  Citation: https://doi.org/10.5194/tc-2023-23-AC1
RC2:
'Review on tc-2023-23', Anonymous Referee #2, 04 Apr 2023

Review for "Seasonal evolution of the supraglacial drainage network at Humboldt Glacier, North Greenland, between 2016 and 2020" by Rawlins et al.
General comments
This paper was a pleasure to read as it is written so well. I only have some minor comments (below), followed by minor specific line by line comments.
The incorporation of slush mapping is very interesting and novel, as I am not aware of prior studies that have mapped slush in detail Greenland(?), though this has been done by at least one study on Antarctic ice shelves (e.g. Dell et al 2022). So it would be great to see more details regarding how exactly slush was mapped in this study. Additionally, I think slush (e.g. as a component of total meltwater area fraction (MF)), could feature more heavily as a discussion point in the Results/Discussion.
I do not know Humboldt Glacier well, so I wondering how much of this marine terminating glacier is floating? Unless I missed it, the only mention of floating ice is on line 208, where a ‘7 km floating section’ is mentioned. Depending on the area of this floating area of ice, I’d also be particularly interested to know whether the authors see a difference between lake/other meltwater feature characteristics on the floating versus grounded portions of the glacier? (For example, comparisons of meltwater features on floating vs. grounded ice have previously been made on Peterman Glacier (Macdonald et al 218) and well as Paakitsoq (SW Greenland) vs. Larsen B Ice shelf, Antarctica (Banwell et al 2014).
I am wondering why the authors choose to base the NDWI on the Green and NIR bands rather than the blue and red bands, e.g. as used by Bell et al (2017) and Williamson et al (2018). I am sure there were/are good reasons, but perhaps a sentence or two about this could be added to the paper. Also related to the use of the NDWI, I am wondering how/why the authors decided to use a threshold of 0.4, i.e. which they call a ‘high-value global NDWI threshold’? Again, I am not suggesting this threshold is not appropriate, but perhaps some more detail and/or reference(s) be added about this choice of value?
My final general comment is that I think the authors should add a short paragraph about uncertainty quantification, particularly regarding their MF analysis.
Specific line by line comments
Abstract lines 16 and 17: Not necessary to state areas to 1 d.p. Round up as is done for other area/elevation values in the abstract.
Abstract line 19: I suggest adding an few extra words to explain what you mean by ‘preconditioning’ here.
56 – 57: Two relevant studies focusing on the surface hydrology of Petermann Glacier in NW Greenland could also be referenced here: Macdonald et al (2018) and Boghosian et al (2021).
133: Gledhill and Williamson just have a 2017 paper, but Williamson et al. have a 2018 paper (both are already in the reference list).
241 – 247: I find this paragraph confusing regarding the maximum extent of meltwater mapped, versus the maximum extent of the study region. For example, the first sentence says that rivers and lakes were mapped up to a “maximum melt extent of 1500 m a.s.l.”. Is 1500 m the highest elevation of the study area analyzed, or is this the maximum elevation of observed melt? I assume the latter(?), but if so, then the following sentence is repetitive (i.e. this states “The mapped supraglacial drainage network across HG is shown extend up to 1500 m a.s.l,”). Also, later in the paragraph (line 246), it says rivers and lakes from up to a “maximum of 1440 m” (as opposed to 1500 m). So these sentences need to be re-written for clarity.
253 - 255: For the sentence “In Figure 3b, we also see some evidence of a potential main-river reconfigurations, with the north-westward advection of a river channel that runs transverse to ice flow”; maybe this river in Fig 3b could be labelled? As I see various rivers/streams that are transverse to ice flow. Also, it looks to me as though similar examples may also be seen in panels c and e?
310 – 314: Can the authors suggest a possible explanation for why these two parallel lines that track across glacier exist? Could they be fractures?
497 - 499: For the sentence:”… this study also notes that many well-established rivers that are longitudinal to ice flow, including many with canyonised features, also reoccupy locations.”, studies focused on Petermann Glacier could also be mentioned here, which found similar findings I believe (Macdonald et al. 2018, Boghosian et al. 2021).
622/623: Mention Summer 2019 somewhere in this sentence to remind the reader which melt season is being described.
Figures
Fig 2: it would be interesting to know the locations of these figure panels, so perhaps an extra panel could be added to show this (e.g. as is done in Fig 3a), or perhaps the locations should be shown somewhere in Fig 1? Also, the ‘off edge river termination’ feature in panel 2c) is interesting, and I’m wondering how comparable this feature could be to the large river/waterfall described in Bell et al (2017)?
References (those in bold are not referenced in the current paper)
Banwell, A.F., Cabellero, M., Arnold, N., Glasser, N., Cathles, L.M., MacAyeal, D. 2014. Supraglacial lakes on the Larsen B Ice Shelf, Antarctica, and Paakitsoq Region, Greenland: a comparative study. Annals of Glaciology. 55(66), doi:10.3189/2014AoG66A049.
Bell, R. E., Chu, W., Kingslake, J., Das, I., Tedesco, M., Tinto, K. J., Zappa, C. J., Frezzotti, M., Boghosian, A., and Lee, W. S.: Antarctic ice shelf potentially stabilized by export of meltwater in surface river, Nature, 544, 344–348, 2017.
Boghosian, A.L., Pitcher, L.H., Smith, L.C. et al. Development of ice-shelf estuaries promotes fractures and calving. Nature Geoscience, 14, 899–905 (2021). https://doi.org/10.1038/s41561-021-00837-7
Dell, R., Banwell. A.F., Willis, I., Arnold, N., Halberstadt, A.R.W., Chudley, T.R., Pritchard, H. 2022, Supervised classification of slush and ponded water on Antarctic ice shelves using Landsat 8 imagery, Journal of Glaciology, 1–14. https://doi.org/10.1017/ jog.2021.114.
Gledhill, L. A. and Williamson, A. G.: Inland advance of supraglacial lakes in north-west Greenland under recent climatic warming, Annals of Glaciology, 59, 66-82, https://doi.org/10.1017/aog.2017.31, 2018.
Macdonald, G.J., Banwell, A.F., MacAyeal, D.R. 2018, Seasonal evolution of supraglacial lakes on a floating ice tongue, Petermann Glacier, Greenland. Annals of Glaciology, doi:10.1017/aog.2018.9
Williamson, A. G., Banwell, A. F., Willis, I. C., and Arnold, N. S.: Dual-satellite (Sentinel-2 and Landsat 8) remote sensing of supraglacial lakes in Greenland, The Cryosphere, 12, 3045-3065, https://doi.org/10.5194/tc-12-3045-2018, 2018.

Citation: https://doi.org/10.5194/tc-2023-23-RC2
- AC2: 'Reply on RC2', Lauren Rawlins, 24 Aug 2023
  
  We would like to thank the reviewer for their positive and insightful comments. Please see attached for our responses
  
  Citation: https://doi.org/10.5194/tc-2023-23-AC2

Lauren D. Rawlins et al.

Supplement

https://doi.org/10.5194/tc-2023-23-supplement

Lauren D. Rawlins et al.

Viewed

Total article views: 725 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
440	265	20	725	47	8	9

HTML: 440
PDF: 265
XML: 20
Total: 725
Supplement: 47
BibTeX: 8
EndNote: 9

Views and downloads (calculated since 22 Feb 2023)

Month	HTML	PDF	XML	Total
Feb 2023	182	105	3	290
Mar 2023	84	60	3	147
Apr 2023	76	39	3	118
May 2023	25	14	2	41
Jun 2023	16	15	1	32
Jul 2023	20	14	3	37
Aug 2023	27	13	5	45
Sep 2023	10	5	0	15

Cumulative views and downloads (calculated since 22 Feb 2023)

Month	HTML	PDF	XML	Total
Feb 2023	182	105	3	290
Mar 2023	84	60	3	147
Apr 2023	76	39	3	118
May 2023	25	14	2	41
Jun 2023	16	15	1	32
Jul 2023	20	14	3	37
Aug 2023	27	13	5	45
Sep 2023	10	5	0	15

Viewed (geographical distribution)

Total article views: 720 (including HTML, PDF, and XML) Thereof 720 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 19 Sep 2023

Short summary

We map and quantify surface rivers and lakes at Humboldt Glacier to examine seasonal evolution and provide new insights of network configuration and behaviour. A widespread supraglacial drainage network exists, expanding up-glacier as seasonal runoff increases. Large interannual variability affects the areal extent of this network, controlled by high vs low melt years, with late-summer network persistence likely preconditioning the surface for earlier drainage activity the following year.


Total:	0
HTML:	0
PDF:	0
XML:	0