the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Mar 2025
Wintertime Evolution of Landfast Ice Stability in Alaska from InSAR
Abstract. Landfast ice in Alaska is experiencing rapid changes in extents and duration, impacting the safety and utility of the ice for Arctic coastal communities. Current datasets of landfast ice only distinguish landfast ice from mobile pack ice, omitting crucial information regarding the relative safety within landfast ice. InSAR (Interferometric Synthetic Aperture Radar) holds promise for identification of landfast ice and measurement of cm-scale deformation from a spaceborne sensor. We use two properties of interferometry: coherence to identify areas of landfast ice, and the interferometric phase gradient to approximate a new metric called apparent strain (εa) which acts as a proxy for estimating the relative stability of the landfast ice. Apparent strain is described as the horizontal gradient of interferometric phase in the line-of-sight displacement. We built on a previous study by Dammann et al. (2019) by assigning quantitative apparent strain values to identify 3 distinct stability classifications of landfast ice: Bottomfast (εa< 1.0x10-5), Stabilized (1.0x10-5 ≤ εa≤ 2.310-5), and Not stabilized (εa > 2.3x10-5). The monthly average apparent strain decreases as the season progresses, achieving the maximum stability in April or May depending on the region. This study introduces a novel approach to identify the relative stability for areas of landfast ice using InSAR. These findings have implications for enhancing the safety and planning of activities on landfast ice for Arctic coastal communities.
- Preprint
(2456 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-567', Anonymous Referee #1, 29 Apr 2025
General Comments
The authors present a study that proposes an InSAR coherence threshold to be used with Sentinel-1 image pairs to find the extent of landfast sea ice on the north coast of Alaska. They mention a loss of coherence in at the beginning and end of the winter season. Most of the study is an extension of Damman et al. (2019), with the current study finding ‘apparent strain’ values that separates landfast sea ice into three categories: bottom-fast ice, stabilized ice, and not-stabilized ice. The authors describe nuances related to the location of the category thresholds and the seasonal evolution of strain values. The efficacy of the methods is evaluated against a climatology for landfast sea ice extent.
The Introduction is rather brief. Please include material defining and relating stability, strain, and displacement, in the context of sea ice. Please also include a survey of similar InSAR techniques for sea ice, including the references listed below, and particularly a more-detailed description of the works on which this study is built (i.e., Dammann et al. 2019; Meyer et al. 2011; Pratt 2022).
- Dammert, P. B. G., Lepparanta, M., & Askne, J. (1998). SAR interferometry over Baltic Sea ice. International Journal of Remote Sensing, 19(16), 3019-3037.
- Li, S., Shapiro, L., McNutt, L., & Feffers, A. (1996). Application of satellite radar interferometry to the detection of sea ice deformation. Journal of the Remote Sensing Society of Japan, 16(2), 153-163.
- Morris, K., Li, S., & Jeffries, M. (1999). Meso-and microscale sea-ice motion in the East Siberian Sea as determined from ERS-1 SAR data. Journal of Glaciology, 45(150), 370-383.
- Wang, Z., Liu, J., Wang, J., Wang, L., Luo, M., Wang, Z., ... & Li, H. (2020). Resolving and analyzing landfast ice deformation by InSAR technology combined with Sentinel-1A ascending and descending orbits data. Sensors, 20(22), 6561.
The map figures in Figures 1, 2, 3, 5, 6 and 9 are inconsistent in the application of standard map elements. Use a consistent lat/lon grid, scalebar type, land colour, and shadow zone colour throughout.
The use of the 0.1 coherence threshold value found by Meyer et al. (2011) for L-band is not well-supported for C-band. Coherence values for C-band are usually significantly higher for sea ice. The assumption of an adequate trade-off between temporal baselines and wavelengths may be true, but this should be supported with evidence. A sensitivity study related to the Figure 4 results may help. Please also provide a representative example of a coherence image alongside a calibrated SAR image, showing landfast sea ice and open water or mobile sea ice. The coherence values should be presented in such a way that they are easy to discern.
The use of the 90th percentile (Figure 8) as the threshold between stabilized and not-stabilized is not convincing. It seems to be based on a rather vague notion that some of the not-stabilized is actually stabilized. This is presented without any supporting evidence. The 10th percentile threshold is also not well-supported. A more robust method for finding the thresholds should be applied. There are many statistical methods to choose from. Given the mix of distribution types, a non-parametric technique such as a decision tree is recommended.
However, it seems that the use of monthly averages is obscuring the abrupt change from stabilized to not stabilized strain values, as can be observed in the individual scenes in Figure 11. Perhaps a more meaningful method can be found to estimate the strain thresholds, based on individual images and the coast vectors.
Overall, the paper needs to substantiate the coherence and strain thresholds further. It may be better to localize the analysis to two or three sub-regions, and investigate these in more detail, as was done in Figure 11, in concert with air temperature data. This may lead to a better understanding of what is affecting the coherence and strain values, and lead to better estimates for the strain thresholds.
The analysis of the strain images to identify grounded ridges is a useful element of this study, and one that could be expanded upon by analyzing more image pairs of smaller regions.
Specific comments
Line 52: In Figure 1, the eleven smaller areas should be briefly mentioned in the text, or if they are unimportant, they can be removed from Figure 1. Indicate what the difference is between the cyan and blue vectors.
Line 66: If the coast-normal vectors are conceptual in Figure 1, please indicate that in the caption. Otherwise, indicate the time period that the vectors represent.
Line 88: Please describe the “other processes unrelated to motion that reduce coherence”, with references.
Line 93: The grammar needs to be improved for this sentence.
Line 146: How can the extent of the ‘not sheltered ice’ be known a priori, in order to create a mask for it. Also, it is not clear what these masks will be used for.
Line 160: The Figure 3 caption references Figure 1 regarding the ‘shadow’ zones. However, these zones are not indicated or obvious in Figure 1. Also, what do the shadow zones represent? Please indicate the Chuckchi-Beaufort border in Figure 3.
Line 163: In Figure 4, the x-axis text is too large, with the location names bleeding into one another. Also, these locations do not seem to align with the sub-regions in Figure 1. There are two Kotzebue locations on the x-axis.
Line 172: In Figure 4e, the high variability in extent just east of the Chuckchi-Beaufort border is not evident in Figure 3, which is also for April. Please explain the inconsistency between figures.
Line 191: Please explain how the percentage values (y-axis in Figures 7, 8, and 10) are calculated. What does percentage represent, especially in Figure 8? The percentages do not appear to add up when comparing Figures 7 and 8. Also indicate in the Figure 7 and 8 captions, the region the data represent.
Line 201: In Figure 8, please add an overall distribution so that the reader can see if the modes are present in the overall un-masked data.
Line 224: Is this single April 2017 comparison the only validation for the proposed thresholds for stability classes? If so, then the evidence is not convincing enough to say that the proposed threshold “…can be usefully applied…”. The large areas seemingly misclassified as bottom fast ice are adjacent to much of the not-stabilized ice. This juxtaposition does not support the statement that “the boundary between stabilized and non-stabilized landfast ice agrees between methods”. Furthermore, the outer extent of the not-stabilized ice is significantly greater in Dammann et al. (2019), which does not support the statement “both methods show good agreement on the position of the SLIE”. This comparative analysis should be redone with additional data. Given the delineations from Dammann et al. (2019) it is reasonable to include a quantitative comparison, e.g., a confusion matrix.
Line 230: Is the ‘particular time’ four years of April data? If so, then are the data for this region’s stabilized ice included in the distributions in Figure 8? If this is an anomalous area, then why not use a more representative area?
Line 237: Should this not be > 0.1?
Line 240: Provide a quantity instead of ‘slightly less’.
Line 250: Where is the Colville Delta?
Line 252: Please provide the May air temperatures to corroborate the attribution.
Line 253: It is not clear how or where the 12-day repeat prevents landfast ice identification.
Line 261: Refer to Figure 7 in the first sentence.
Line 265: In Figure 10, should not Figure 10l (whole study region) be the same as Figure 7? The values and monthly peaks are different. Also, is ‘interferometric phase gradient’ in the caption supposed to mean ‘apparent strain’?
Line 269: Is the extent not a function of coherence, which is indicated to be poor in May. Does this affect the pdf results shown for May?
Line 287: The statement that the apparent strain threshold ‘work well’ has yet to be shown. This may need to be revised in light of previous comments.
Line 301: In Figure 11’s caption, indicate that the thick red line is the boundary between the stabilized and not stabilized landfast ice. Indicate the location of the red line in panels c and d as well.
Line 304: Some of the text in Figure 11c and d is too small.
Line 302: The sentence needs to be reworded: “This boundary by another steep…”.
Line 320: The factors unrelated to motion causing a loss of coherence could be investigated in this study.
Line 323: Why would a shorter period improve coherence if the cause is temperature and snow moisture related? An analysis of the air temperature would likely provide some evidence towards a cause.
Line 343: The thickening of the ice is a reasonable assumption as to the cause of the seasonal decrease in apparent strain. This should be investigated in this study, using air temperature data and a simple freezing model.
Line 347: Why was a 2-D method not pursued?
Technical corrections:
Line 55: about => abut
Line 128: too => to.
Line 212: expect => suspect?
Line 238: immobile ice
Line 239: the average the
Line 268: ‘May in less’ => ‘May is less’
Line 306: ‘…of this feature elevated…’ => ‘…of this feature are elevated…’
Line 322: out => our
Citation: https://doi.org/10.5194/egusphere-2025-567-RC1 -
RC2: 'Comment on egusphere-2025-567', Anonymous Referee #2, 08 May 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-567/egusphere-2025-567-RC2-supplement.pdf
- AC1: 'Comment on egusphere-2025-567', Andrew Einhorn, 25 Jun 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 197 | 45 | 17 | 259 | 13 | 19 |
- HTML: 197
- PDF: 45
- XML: 17
- Total: 259
- BibTeX: 13
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1