the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Jul 2025
Assessment of Correlation between Sea Ice Fractures and Meteorological Conditions in Tuktoyaktuk
Abstract. Landfast sea ice is a crucial component of the Arctic ecosystem and holds significant cultural importance for indigenous communities in the Canadian Arctic. They rely on it for hunting, navigation, and maintaining connections between communities. As a climate indicator, landfast sea ice plays a key role in heat exchange between the ocean and the atmosphere, serves as a buffer against storms and coastal erosion, and impacts local environmental stability. However, the Arctic is experiencing rapid warming, which is leading to a reduction in the extent, thickness, and strength of sea ice. This trend increases the vulnerability of communities dependent on landfast ice to severe environmental events. This study focuses on the community of Tuktoyaktuk in the Northwest Territories, Canada, where landfast sea ice has been observed to fracture more frequently in the winter season. Using satellite-based Synthetic Aperture Radar (SAR) and meteorological data, this research identifies that changing wind directions, particularly strong western and north-western winds, are significant factors contributing to the displacement of ice from the shore. Observations reveal a growing frequency of ice fracture events from 3 to 13 fracture counts from 2016–2017 to 2022–2023, coinciding with an increase in the occurrence of strong winds along the Tuktoyaktuk coast. The study aims to assess the likelihood of landfast sea ice fracturing under specific wind conditions, providing insights into the impact of climatic changes on the stability of ice in this region.
- Preprint
(49188 KB) - Metadata XML
-
Supplement
(18993 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2025-2851', Anonymous Referee #1, 04 Aug 2025
Assessment of Correlation between Sea Ice Fractures and Meteorological Conditions in Tuktoyaktuk by Chaudhary and others
General Comments
This study uses SAR from Sentinel-1 to look at ice fractures off the coast of Tutoyaktuk from 2016-2023. They find wintertime ice fractures in the Beaufort Sea have increased over their time period and suggest they are associated with the decrease of the landfast extent off the coast of Tutoyaktuk. Community focused studies like this one are important but I have some major concerns.1. My main concern is the connection between the landfast ice and the polar pack processes of which the latter is in continuous motion. The mechanism suggested is “during winter, these winds push landfast sea ice away from the shore, resulting in weaker, thinner ice that is more susceptible to fracturing” but the evidence is rather weak and speculative. When the landfast edge sets up, it remains until break-up. Are the author’s suggesting this is not the case? If so, evidence for it changing during the winter needs to be presented. More ice fractures offshore is not really evidence. Landfast ice and open drift ice are completely different not to mention landfast ice is mechanically stronger. Fractures in the pack ice (Beaufort Sea) occur when the landfast ice remains intact so its not readily apparent how this relates to landfast ice break-up and the authors have not really demonstrated this linkage. For landfast breakup, you need warming, wind, etc. to drive the breakup. The author’s proposed process needs to be communicated better with more evidence as in its current form it is very vague with minimal supporting evidence.
2. A key result is an increased frequency in ice fractures over the 5-year time period. But this is a short time period so I’m not sure it is fair to say fractures are increasing. Looking at the actual values clearly shows a high in 2022 and a low in 2016 but here is considerable variability. Also, the fractures in Beaufort Sea are pretty far away from Tuktoyaktuk and that ice is constantly in motion so the ice conditions are not stable. So, there are two big problems associated with rather substantial statements, a short time series and continuous motion. I think more careful wording needs to be used here or a revised analysis.
Specific Comments
Line 18
Some quantitative values would be useful here.Lines 19-20
What local observations?
What are Northern Canadian Waters? When is the decrease? The reference is not formatted correctly. Revise.Line 22
No reference for landfast ice extent. Also, a definition of what landfast ice would be useful hereLine 31-32
It certainly deforms but only when mechanically weak. During the winter, landfast ice is very strong and represents a strong barrier not easily broken. Case in point, sea ice from the Beaufort Sea/Canadian Basin cannot penetrated into the Canadian Arctic Archipelago during the winter months. That is, the pack ice is the ice that fractures and deforms, not the landfast ice.Line 63-64
Statement seems redundant compared to previous sentences.Line 68
TUK should be Tuk or even Tuktoyaktuk. I’m not sure shortening a hamlet to Tuk is the best idea here. Place names should be spelled out in fullLine 75
Reference is cited strangely. Further, it would be better to cite a scientific study rather than the new media.Line 77
No date on the citation.
The word stable needs some context. I think it is still stable but the duration is reduced. What about the extent? Has that changed?Line 81
No date on the citation.Line 101
Does Sigma Nought need to be capitalized?Line 106-111
A lot of statements without references.Line 114
Why is TUKTOYAKTUK capitalized?Line 118
Mean Sea Level Pressure does not need capital letters.Line 122
There are better references for the CIS ice charts used in scientific studies. Tivy et al. 2011 is the baseline.Line 125
How does that increase precision? Don’t you mean accurate?Line 136
I’m not sure how Landsat helps here. You can see the fractures with SAR. I’m also not sure how computational burdens/storage are issues nowadays. The community has had the computing power to process and store large amounts of SAR data for many years.Line 161
Is the monthly scale suitable for all fracture events? Some can close in a manner of days. Has this been considered?Line 172
The SOM are useful spatially but there is a lot of information. What would be useful is a simple pressure gradient. This would help identify regions of divergence or convergence and more facilitate a time series plot which readers can more easily identify with. Numerous studies (especially by Kwok) have done this with great success.Line 181
You are looking at the relationship between atmospheric conditions and landfast ice but how do the fractures in the Beaufort Sea relate to this? This is not clear to me because the ice regimes are totally different. See General Comment.Line 213
Again, I don’t see how fractures in the winter outside of the landfast ice are related? They don’t influence landfast breakup. In fact, landfast is so strong that ice being forced against it will just ridge like an impenetrable barrier.Line 233
How do you known that the winds have influenced the boundary of the landfast edge? That’s an unsupported statement.Line 278
No date on the Canadian Ice Service reference.Line 319-325
I don’t see the evidence for this. I also don’t see the evidence for how this influences ice edge variability or now it has reduced its extent. Currently these statements are speculative and unsupported. If a physical link is going to be proposed it must be quantitatively supported which is currently not the case. The author’s have not made the case that winds in the winter translate to breakup in the spring/summer. See General Comment.Line 331-332
Winds do not push landfast ice away from the shore. The ice is landfast and remains immobile until break-up. The author’s have not shown how winds in the winter contribute to break-up the landfast ice. There is a disconnect between landfast processes and open drift processes here. See General Comment.Abstract
I think the abstract lacks quantitative information.Landsat data
I’m not certain the value of using optical imagery as validation. SAR alone is very good. Further the CIS ice charts also denote the landfast ice area/edge in your study region.Figure 5
This belongs in the Supplementary Material. It fairly well understood that ERA5 is representative.Figure 6
This is a lazy Figure that shows the regional analysis of the entire Western Canadian Arctic. There is way too much information and will also be impossible to read once copyedited. A better idea would be to make your own figures from the shapefiles focusing in on your study region. This what you can also construct a time series of ice conditions which would be useful and what the analysis is currently lacking.Figure 11
Are you sure 2023 isn’t just an anomaly?References
Tivy, A., Howell, S. E. L., Alt, B., McCourt, S., Chagnon, R., Crocker, G., et al. (2011). Trends and variability in summer sea ice cover in the Canadian Arctic based on the Canadian Ice Service Digital Archive, 1960–2008 and 1968–2008. Journal of Geophysical Research, 116,
C03007. https://doi.org/10.1029/2009JC005855Yu, Y., Stern, H., Fowler, C., Fetterer, F., and Maslanik, J.: Interannual Variability of Arctic Landfast Ice between 1976 and 2007, J. Climate, 27, 227–243, doi:10.1175/JCLI-D-13-00178.1, 2014.
Citation: https://doi.org/10.5194/egusphere-2025-2851-RC1 -
RC2: 'Comment on egusphere-2025-2851', Anonymous Referee #2, 03 Sep 2025
Coastal sea ice is an interesting area of research, particularly in the context of the complex coastal dynamics of the Tuktoyaktuk/Mackenzie delta region. The approach of using SAR to monitor landfast versus drifting sea ice and open water has been done before over larger areas. Lead formation events are interesting, but the approach presented in this manuscript does not make substantial contributions to this area.
Major comments:
- Throughout this paper, there is very little consideration of the processes at work in land fast ice formation, breakup, and the formation of flaw leads at the seaward land fast ice edge. Further, it is not clear what if what is being referred to as LFSI/land fast sea ice is actually land fast ice or if it is ice that is temporarily attached to /abutting actually-land fast ice. The mechanisms involved in both land fast ice formation (ridge building/grounding) and ice breakout events (fracture in the SLIE) may help make more sense of the wind patterns. Likewise, the paper does not distinguish between fractures that are a flaw lead opening (not destroying land fast sea ice) and fractures within otherwise land fast sea ice (which would be much more notable). Using the proper terminology based on the surrounding ice conditions will clarify that. Some papers that will be helpful in describing a more clear picture of the processes at work include:
Lange, K. A., Bradley, A. C., Duncan, K., and Farrell, S. L.: Grounded ridge detection and characterization along the Alaska Arctic coastline using ICESat-2 surface height retrievals, The Cryosphere, 19, 2045–2065, https://doi.org/10.5194/tc-19-2045-2025, 2025.
Mahoney, A. R., Eicken, H., Gaylord, A. G., and Gens, R.: Landfast sea ice extent in the Chukchi and Beaufort Seas: The annual cycle and decadal variability, Cold Reg. Sci. Technol., 103, 41–56, 2014.
Mahoney, A., Eicken, H., and Shapiro, L.: How fast is landfast sea ice? A study of the attachment and detachment of nearshore ice at Barrow, Alaska, Cold Reg. Sci. Technol., 47, 233–255, 2007.
The Mahoney papers talk a little bit about the shore (or flaw) lead that commonly forms between the SLIE and the drifting pack ice, but if you want to get a better intuitive sense for it, watch the UAF Utqiagvik sea ice radar during the winter: https://seaice.alaska.edu/coastal-ice-observations/utqiagvik/
In particular, the discussion around lines 315-325 suggests an incomplete understanding of the dynamics of land fast ice in the region.
2. SAR scenes from a number of dates are analyzed, but there is no description of how a SAR scene date is connected to a fracture event date: Given that there are only 267 SAR scenes over six winters of data, there clearly is not a scene available every day. The wind speed analysis depends on having dates for fracture opening. Line 135 seems to indicate a small subset was processed: does this refer spatially (in that case, what area?) or only certain images? What is “the subset” in figure 7? Is it just the two images? Or are those areas cut out of a larger image? Line 210 says that these images show the dynamic between the fast ice and pack ice, but without any description of what the images are actually showing.
3. Six winters of data over seven years is hardly adequate for the claims of trends made in this paper. Significance testing of the identified trends (e.g., lines 230, 288, 293) would likely indicate that these are high-variability signals over a short period of time.
4. A more robust analysis of wind/meteorological forcing would compare dates with fracture events (assuming you can be certain about the dates) to dates without fracture events. This might help pare down the number of figures you have (see comment 9). Section 3.2.1 would particularly benefit from reframing and considering the no-fracture "control" group.
5. The Tuktoyaktuk weather station is one of the (many) assimilated weather stations for the ERA5 product. The fact that the correlation for wind measurements and the reanalysis product is as weak as you show in Figure 5 suggests that maybe ERA5 is not a great estimate of wind in the area.
6. How are you using the Landsat data? It isn’t described in any detail in the Datasets section, but it shows up in Figure 2 and the introduction to the Datasets section says it is there for validation.
7. Be clear about where and how the land fast ice edge is detected: if I understand correctly, you are using the ECCC ice chart’s indicated land fast ice edge for the latest date of the winter? If that is correct, be consistent with interpretation of the ice fractures (lead opening) seaward of the land fast ice edge, and also reference their materials on how the land fast ice edge is determined/defined.
8. Since the temperature analysis showed nothing significant, you can abbreviate this section rather than spelling out temperatures (to an oddly high precision). The -15C threshold seems arbitrary, so if there is a reasoning behind it please explain.
9. Throughout the paper, the figures and tables are not necessarily appropriate for the messages being conveyed. Figures and tables are not in the order they are referred to in the text. There are a lot of figures with a lot of sub-panels, and very few of these are actually discussed in text. Limit the figures you show to what is relevant and consider alternate options for ways to show the relevant data in more concise ways.
- Figure 1: the inset requires a scale bar, and the fast ice edges are almost impossible to see without zooming in. Change the line weight and color to make it easier to read.
- Table 1: It is unclear why this is summarized in a table, when the same information would make more sense (and is) in the description of datasets in the text.
- Figure 2: The text reference to Figure 2 (line 137) has nothing to do with what is actually shown in this figure (examples of open flaw leads).
- Figure 3: The analytical approach flowchart is a nice idea, but it lacks so much detail as to be of little use. Adding detail and breaking it into each step (e.g., SAR preprocessing -> Adaptive thresholding/open water detection) would make it more useful.
- Figure 4: This looks like a textbook image showing how a SOM works, and it is not particularly useful in the current format in the paper. Using real world terms (what is the high-dimensional input space?) rather than the more generic terms would help make this clear.
- Figure 5: See comment 5 above. What are you trying to show with this table of plots? Do they need to be separated by year?
- Table 2: What do these dates refer to? Is there a particular geographic area that is ice covered?
- Figure 6: Why include all of these images? Add a map inset to show the extent of these images relative to the other north-up maps used throughout the paper. A summary figure would be better than having both figure 6 and table 2.
- Figure 7: Overlaying the images and the gridded open water detections would help evaluate the detection algorithm. This figure needs a more informative caption.
- Table 3: Is this the number of dates with any number of fracture events in the SAR imagery, the number of fractures, or the total number of fracture events in the full duration of the winter season? Be clear how many fracture events relative to how many SAR imagery dates you have.
- Table 4: Is the standard deviation the correct metric for variability in low pressure systems? Given the number of tables you have organized by year, they could be combined into one table.
- Figure 8: This table of wind roses is very large and totally illegible without really zooming in on the document (which you could not do in print). Consider what you are actually trying to show here and re-design the figure to focus on that. It might also help to overlay on the wind rose an outline of what directions are on- versus off-shore (though that is complicated by the coastline).
- Figure 9: This figure seems to replicate the data from Figure 8.
- Table 5: What is rate of change in temperature per month used for?
- Figure 10: This panel of figures is likewise illegible because they are too small. Consider what the message of the figure is actually meant to be and focus on presenting that (see also major comment 2 regarding dates).
- Figure 11: At first glance, this seems to show a lot of variability in the land fast ice edge rather than a decline/landward retreat over time. If you have more quantitative evidence of a retreat over time (that isn’t a change in definitions in the ice charts) then show that rather than the squiggly lines.
- Table 6: is this necessary? Would a figure be better? Perhaps a summary wind rose?
Addressing the major comments will make most of my smaller comments unnecessary, so these are just a few details that specifically need to be addressed:
63: Spell Tuktoyaktuk correctly. Throughout you use Tuk and TUK: Tuk is acceptable as a common abbreviation of the name, but you should indicate that it is popular usage for the Indigenous community.
162: The description of the color scale is backwards.
175: Clarify what the PCA did and what it means in real terms.
186: What do you mean by the “limited spatial resolution… of the weather station”?
232: The statement that increases in winds result in variability in boundaries of land fast ice is not necessarily true: more consideration of the mechanisms (and the relevant wind directions) of land fast ice formation would help make more sense of this argument.
247: Adding up these directions, it seems that you are trying to say “wind from WNW +/- 30° occurred on ~50% of the fracture dates”. Calling a direction “most frequent” at 23% of total dates seems to be overstating it.
269: What is high versus moderate variability in low versus high pressure occurrences?
271: Are these averages actually meaningful?
279: By “bay mouth” do you mean river outlet?
285: “clear predominance” overstates the statistics.
314: Do not use “correlates well” in this context if you are not actually measuring correlation.
328: The lack of acknowledgment of the role and value of local knowledge in this context is troublesome.
Citation: https://doi.org/10.5194/egusphere-2025-2851-RC2
Status: closed
-
RC1: 'Comment on egusphere-2025-2851', Anonymous Referee #1, 04 Aug 2025
Assessment of Correlation between Sea Ice Fractures and Meteorological Conditions in Tuktoyaktuk by Chaudhary and others
General Comments
This study uses SAR from Sentinel-1 to look at ice fractures off the coast of Tutoyaktuk from 2016-2023. They find wintertime ice fractures in the Beaufort Sea have increased over their time period and suggest they are associated with the decrease of the landfast extent off the coast of Tutoyaktuk. Community focused studies like this one are important but I have some major concerns.1. My main concern is the connection between the landfast ice and the polar pack processes of which the latter is in continuous motion. The mechanism suggested is “during winter, these winds push landfast sea ice away from the shore, resulting in weaker, thinner ice that is more susceptible to fracturing” but the evidence is rather weak and speculative. When the landfast edge sets up, it remains until break-up. Are the author’s suggesting this is not the case? If so, evidence for it changing during the winter needs to be presented. More ice fractures offshore is not really evidence. Landfast ice and open drift ice are completely different not to mention landfast ice is mechanically stronger. Fractures in the pack ice (Beaufort Sea) occur when the landfast ice remains intact so its not readily apparent how this relates to landfast ice break-up and the authors have not really demonstrated this linkage. For landfast breakup, you need warming, wind, etc. to drive the breakup. The author’s proposed process needs to be communicated better with more evidence as in its current form it is very vague with minimal supporting evidence.
2. A key result is an increased frequency in ice fractures over the 5-year time period. But this is a short time period so I’m not sure it is fair to say fractures are increasing. Looking at the actual values clearly shows a high in 2022 and a low in 2016 but here is considerable variability. Also, the fractures in Beaufort Sea are pretty far away from Tuktoyaktuk and that ice is constantly in motion so the ice conditions are not stable. So, there are two big problems associated with rather substantial statements, a short time series and continuous motion. I think more careful wording needs to be used here or a revised analysis.
Specific Comments
Line 18
Some quantitative values would be useful here.Lines 19-20
What local observations?
What are Northern Canadian Waters? When is the decrease? The reference is not formatted correctly. Revise.Line 22
No reference for landfast ice extent. Also, a definition of what landfast ice would be useful hereLine 31-32
It certainly deforms but only when mechanically weak. During the winter, landfast ice is very strong and represents a strong barrier not easily broken. Case in point, sea ice from the Beaufort Sea/Canadian Basin cannot penetrated into the Canadian Arctic Archipelago during the winter months. That is, the pack ice is the ice that fractures and deforms, not the landfast ice.Line 63-64
Statement seems redundant compared to previous sentences.Line 68
TUK should be Tuk or even Tuktoyaktuk. I’m not sure shortening a hamlet to Tuk is the best idea here. Place names should be spelled out in fullLine 75
Reference is cited strangely. Further, it would be better to cite a scientific study rather than the new media.Line 77
No date on the citation.
The word stable needs some context. I think it is still stable but the duration is reduced. What about the extent? Has that changed?Line 81
No date on the citation.Line 101
Does Sigma Nought need to be capitalized?Line 106-111
A lot of statements without references.Line 114
Why is TUKTOYAKTUK capitalized?Line 118
Mean Sea Level Pressure does not need capital letters.Line 122
There are better references for the CIS ice charts used in scientific studies. Tivy et al. 2011 is the baseline.Line 125
How does that increase precision? Don’t you mean accurate?Line 136
I’m not sure how Landsat helps here. You can see the fractures with SAR. I’m also not sure how computational burdens/storage are issues nowadays. The community has had the computing power to process and store large amounts of SAR data for many years.Line 161
Is the monthly scale suitable for all fracture events? Some can close in a manner of days. Has this been considered?Line 172
The SOM are useful spatially but there is a lot of information. What would be useful is a simple pressure gradient. This would help identify regions of divergence or convergence and more facilitate a time series plot which readers can more easily identify with. Numerous studies (especially by Kwok) have done this with great success.Line 181
You are looking at the relationship between atmospheric conditions and landfast ice but how do the fractures in the Beaufort Sea relate to this? This is not clear to me because the ice regimes are totally different. See General Comment.Line 213
Again, I don’t see how fractures in the winter outside of the landfast ice are related? They don’t influence landfast breakup. In fact, landfast is so strong that ice being forced against it will just ridge like an impenetrable barrier.Line 233
How do you known that the winds have influenced the boundary of the landfast edge? That’s an unsupported statement.Line 278
No date on the Canadian Ice Service reference.Line 319-325
I don’t see the evidence for this. I also don’t see the evidence for how this influences ice edge variability or now it has reduced its extent. Currently these statements are speculative and unsupported. If a physical link is going to be proposed it must be quantitatively supported which is currently not the case. The author’s have not made the case that winds in the winter translate to breakup in the spring/summer. See General Comment.Line 331-332
Winds do not push landfast ice away from the shore. The ice is landfast and remains immobile until break-up. The author’s have not shown how winds in the winter contribute to break-up the landfast ice. There is a disconnect between landfast processes and open drift processes here. See General Comment.Abstract
I think the abstract lacks quantitative information.Landsat data
I’m not certain the value of using optical imagery as validation. SAR alone is very good. Further the CIS ice charts also denote the landfast ice area/edge in your study region.Figure 5
This belongs in the Supplementary Material. It fairly well understood that ERA5 is representative.Figure 6
This is a lazy Figure that shows the regional analysis of the entire Western Canadian Arctic. There is way too much information and will also be impossible to read once copyedited. A better idea would be to make your own figures from the shapefiles focusing in on your study region. This what you can also construct a time series of ice conditions which would be useful and what the analysis is currently lacking.Figure 11
Are you sure 2023 isn’t just an anomaly?References
Tivy, A., Howell, S. E. L., Alt, B., McCourt, S., Chagnon, R., Crocker, G., et al. (2011). Trends and variability in summer sea ice cover in the Canadian Arctic based on the Canadian Ice Service Digital Archive, 1960–2008 and 1968–2008. Journal of Geophysical Research, 116,
C03007. https://doi.org/10.1029/2009JC005855Yu, Y., Stern, H., Fowler, C., Fetterer, F., and Maslanik, J.: Interannual Variability of Arctic Landfast Ice between 1976 and 2007, J. Climate, 27, 227–243, doi:10.1175/JCLI-D-13-00178.1, 2014.
Citation: https://doi.org/10.5194/egusphere-2025-2851-RC1 -
RC2: 'Comment on egusphere-2025-2851', Anonymous Referee #2, 03 Sep 2025
Coastal sea ice is an interesting area of research, particularly in the context of the complex coastal dynamics of the Tuktoyaktuk/Mackenzie delta region. The approach of using SAR to monitor landfast versus drifting sea ice and open water has been done before over larger areas. Lead formation events are interesting, but the approach presented in this manuscript does not make substantial contributions to this area.
Major comments:
- Throughout this paper, there is very little consideration of the processes at work in land fast ice formation, breakup, and the formation of flaw leads at the seaward land fast ice edge. Further, it is not clear what if what is being referred to as LFSI/land fast sea ice is actually land fast ice or if it is ice that is temporarily attached to /abutting actually-land fast ice. The mechanisms involved in both land fast ice formation (ridge building/grounding) and ice breakout events (fracture in the SLIE) may help make more sense of the wind patterns. Likewise, the paper does not distinguish between fractures that are a flaw lead opening (not destroying land fast sea ice) and fractures within otherwise land fast sea ice (which would be much more notable). Using the proper terminology based on the surrounding ice conditions will clarify that. Some papers that will be helpful in describing a more clear picture of the processes at work include:
Lange, K. A., Bradley, A. C., Duncan, K., and Farrell, S. L.: Grounded ridge detection and characterization along the Alaska Arctic coastline using ICESat-2 surface height retrievals, The Cryosphere, 19, 2045–2065, https://doi.org/10.5194/tc-19-2045-2025, 2025.
Mahoney, A. R., Eicken, H., Gaylord, A. G., and Gens, R.: Landfast sea ice extent in the Chukchi and Beaufort Seas: The annual cycle and decadal variability, Cold Reg. Sci. Technol., 103, 41–56, 2014.
Mahoney, A., Eicken, H., and Shapiro, L.: How fast is landfast sea ice? A study of the attachment and detachment of nearshore ice at Barrow, Alaska, Cold Reg. Sci. Technol., 47, 233–255, 2007.
The Mahoney papers talk a little bit about the shore (or flaw) lead that commonly forms between the SLIE and the drifting pack ice, but if you want to get a better intuitive sense for it, watch the UAF Utqiagvik sea ice radar during the winter: https://seaice.alaska.edu/coastal-ice-observations/utqiagvik/
In particular, the discussion around lines 315-325 suggests an incomplete understanding of the dynamics of land fast ice in the region.
2. SAR scenes from a number of dates are analyzed, but there is no description of how a SAR scene date is connected to a fracture event date: Given that there are only 267 SAR scenes over six winters of data, there clearly is not a scene available every day. The wind speed analysis depends on having dates for fracture opening. Line 135 seems to indicate a small subset was processed: does this refer spatially (in that case, what area?) or only certain images? What is “the subset” in figure 7? Is it just the two images? Or are those areas cut out of a larger image? Line 210 says that these images show the dynamic between the fast ice and pack ice, but without any description of what the images are actually showing.
3. Six winters of data over seven years is hardly adequate for the claims of trends made in this paper. Significance testing of the identified trends (e.g., lines 230, 288, 293) would likely indicate that these are high-variability signals over a short period of time.
4. A more robust analysis of wind/meteorological forcing would compare dates with fracture events (assuming you can be certain about the dates) to dates without fracture events. This might help pare down the number of figures you have (see comment 9). Section 3.2.1 would particularly benefit from reframing and considering the no-fracture "control" group.
5. The Tuktoyaktuk weather station is one of the (many) assimilated weather stations for the ERA5 product. The fact that the correlation for wind measurements and the reanalysis product is as weak as you show in Figure 5 suggests that maybe ERA5 is not a great estimate of wind in the area.
6. How are you using the Landsat data? It isn’t described in any detail in the Datasets section, but it shows up in Figure 2 and the introduction to the Datasets section says it is there for validation.
7. Be clear about where and how the land fast ice edge is detected: if I understand correctly, you are using the ECCC ice chart’s indicated land fast ice edge for the latest date of the winter? If that is correct, be consistent with interpretation of the ice fractures (lead opening) seaward of the land fast ice edge, and also reference their materials on how the land fast ice edge is determined/defined.
8. Since the temperature analysis showed nothing significant, you can abbreviate this section rather than spelling out temperatures (to an oddly high precision). The -15C threshold seems arbitrary, so if there is a reasoning behind it please explain.
9. Throughout the paper, the figures and tables are not necessarily appropriate for the messages being conveyed. Figures and tables are not in the order they are referred to in the text. There are a lot of figures with a lot of sub-panels, and very few of these are actually discussed in text. Limit the figures you show to what is relevant and consider alternate options for ways to show the relevant data in more concise ways.
- Figure 1: the inset requires a scale bar, and the fast ice edges are almost impossible to see without zooming in. Change the line weight and color to make it easier to read.
- Table 1: It is unclear why this is summarized in a table, when the same information would make more sense (and is) in the description of datasets in the text.
- Figure 2: The text reference to Figure 2 (line 137) has nothing to do with what is actually shown in this figure (examples of open flaw leads).
- Figure 3: The analytical approach flowchart is a nice idea, but it lacks so much detail as to be of little use. Adding detail and breaking it into each step (e.g., SAR preprocessing -> Adaptive thresholding/open water detection) would make it more useful.
- Figure 4: This looks like a textbook image showing how a SOM works, and it is not particularly useful in the current format in the paper. Using real world terms (what is the high-dimensional input space?) rather than the more generic terms would help make this clear.
- Figure 5: See comment 5 above. What are you trying to show with this table of plots? Do they need to be separated by year?
- Table 2: What do these dates refer to? Is there a particular geographic area that is ice covered?
- Figure 6: Why include all of these images? Add a map inset to show the extent of these images relative to the other north-up maps used throughout the paper. A summary figure would be better than having both figure 6 and table 2.
- Figure 7: Overlaying the images and the gridded open water detections would help evaluate the detection algorithm. This figure needs a more informative caption.
- Table 3: Is this the number of dates with any number of fracture events in the SAR imagery, the number of fractures, or the total number of fracture events in the full duration of the winter season? Be clear how many fracture events relative to how many SAR imagery dates you have.
- Table 4: Is the standard deviation the correct metric for variability in low pressure systems? Given the number of tables you have organized by year, they could be combined into one table.
- Figure 8: This table of wind roses is very large and totally illegible without really zooming in on the document (which you could not do in print). Consider what you are actually trying to show here and re-design the figure to focus on that. It might also help to overlay on the wind rose an outline of what directions are on- versus off-shore (though that is complicated by the coastline).
- Figure 9: This figure seems to replicate the data from Figure 8.
- Table 5: What is rate of change in temperature per month used for?
- Figure 10: This panel of figures is likewise illegible because they are too small. Consider what the message of the figure is actually meant to be and focus on presenting that (see also major comment 2 regarding dates).
- Figure 11: At first glance, this seems to show a lot of variability in the land fast ice edge rather than a decline/landward retreat over time. If you have more quantitative evidence of a retreat over time (that isn’t a change in definitions in the ice charts) then show that rather than the squiggly lines.
- Table 6: is this necessary? Would a figure be better? Perhaps a summary wind rose?
Addressing the major comments will make most of my smaller comments unnecessary, so these are just a few details that specifically need to be addressed:
63: Spell Tuktoyaktuk correctly. Throughout you use Tuk and TUK: Tuk is acceptable as a common abbreviation of the name, but you should indicate that it is popular usage for the Indigenous community.
162: The description of the color scale is backwards.
175: Clarify what the PCA did and what it means in real terms.
186: What do you mean by the “limited spatial resolution… of the weather station”?
232: The statement that increases in winds result in variability in boundaries of land fast ice is not necessarily true: more consideration of the mechanisms (and the relevant wind directions) of land fast ice formation would help make more sense of this argument.
247: Adding up these directions, it seems that you are trying to say “wind from WNW +/- 30° occurred on ~50% of the fracture dates”. Calling a direction “most frequent” at 23% of total dates seems to be overstating it.
269: What is high versus moderate variability in low versus high pressure occurrences?
271: Are these averages actually meaningful?
279: By “bay mouth” do you mean river outlet?
285: “clear predominance” overstates the statistics.
314: Do not use “correlates well” in this context if you are not actually measuring correlation.
328: The lack of acknowledgment of the role and value of local knowledge in this context is troublesome.
Citation: https://doi.org/10.5194/egusphere-2025-2851-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 617 | 35 | 14 | 666 | 15 | 50 | 41 |
- HTML: 617
- PDF: 35
- XML: 14
- Total: 666
- Supplement: 15
- BibTeX: 50
- EndNote: 41
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1