the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Jul 2026
Data-driven Tide Correction for SAR-derived Ice Velocity Measurements on Antarctic Ice Shelves
Abstract. Synthetic Aperture Radar (SAR) offset tracking is widely used to measure ice velocity on floating ice shelves, however, vertical displacement caused by ocean tides projects into the radar-range direction, introducing apparent horizontal motion unrelated to true ice flow. Here, we develop an observation-driven tide-correction for SAR-derived ice speed measurements using differential range offset tracking (DROT) applied to Sentinel-1 SAR data. The method estimates spatially and temporally varying tidal response coefficients directly from observed radar-range displacement variability and an ocean tide model. Application of the method to SAR-derived ice velocity measurements on Antarctic Ice Shelves shows spatial variability in the magnitude of the tide-induced ice-speed correction, with an Antarctic-wide mean tide-correction range of 106 m/yr, a standard deviation of 71 m/yr and maximum values exceeding 300 m/yr. The largest tide correction range predominantly occurs on freely floating ice shelves in regions of strong tidal forcing, particularly in the Weddell Sea and Antarctic Peninsula. Smaller tide corrections are observed near grounding zones and pinned regions where tidal motion is mechanically constrained. Time-series analyses show that the tide correction systematically reduces non-physical temporal variability in SAR-derived ice velocity products while preserving the large-scale spatial structure of ice flow. Independent validation using GNSS observations on Brunt Ice Shelf shows improved agreement between tide-corrected satellite-based ice speed and in situ measurements. Our results demonstrate that tide-induced apparent motion is a source of variability in SAR-derived ice velocity products on floating ice shelves, highlighting the importance of applying a correction when tide-induced signals are not required.
- Preprint
(7838 KB) - Metadata XML
-
Supplement
(825 KB) - BibTeX
- EndNote
Status: open (until 11 Oct 2026)
-
RC1: 'Comment on egusphere-2026-3795', Anonymous Referee #1, 27 Aug 2026
reply
SummaryIn Zhu et al., the authors develop a new empirical tide correction for Synthetic Aperture Radar (SAR) data using differential range tracking and outputs from both a regional tide model (CATS2008) and an inverse-barometer correction. The empirical correction they compute appears to be technically sound, but I do not believe they properly place it in context. And so while the work presented by the authors falls within the scope of The Cryosphere, I would not recommend it for publication in its current form.General Comments
- My main complaint is that applying a tide correction to SAR data is standard operating procedure for floating ice shelves and ice tongues. There are a few different methods used depending on the data producer, but the reasoning for doing this correction is known (e.g. Rignot and MacAyeal, 1998; Rignot et al., 2000). I believe the authors should be comparing their data against the present-day best estimates and not against raw SAR ranges. In other words, the debate should be "How is this better than existing tide corrections?", and not "How is this better than nothing?" The authors do include a comparison late in the paper against a tide correction computed using a flexure model, but they do not compare this other method against the GPS data. There's also a third possible method that is simply outputs from an "off-the-shelf" model converted to horizontal displacements using equation 1 and the time-variable incidence angles. This third model would likely be quite bad in areas of ice flexure, but it would be nice to see how the DROT method compares to it in hydrostatic areas (typically a few ice thicknesses downstream of the grounding zone). For evaluation, you might want to do something akin to a "hold-out" method to compare your DROT method against the flexure method and the simple method (i.e. use newer Sentinel data that haven't been folded into your routine as an independent evaluator).
- In equation 2, ⍵ should be some convolution of cos(θ) and flexure (I'll call it ξ), correct? If so, what I might want to see is maps of this modulation factor (ξ) and not maps of the apparent horizontal tide correction (S).
- Only comparing against GPS data at Brunt is an interesting choice. That ice shelf is pretty heavily crevassed and weakly connected. Are there other GPS sites that might provide better evaluation statistics? Maybe even multiple sites?
- How could this work help improve modern-day global and regional tide models?
Line-by-Line CommentsLine 179: This sounds similar to what was done in Rignot et al. (1996) to calculate the tidal signature and flexure at Petermann Gletscher.Figure 3: These figures might be way too zoomed out. You might want to put full versions in the supplement. The tidal amplitude plot is so tiny it is almost worthless.Figure 4: Ice speed difference (typo)Figure 5: Number of raw observations (typo) and Number of tide-corrected observations (typo)Figure 6: You might want this to be 4 sub-plots. To me it is confusing to have the differences on the same plot as my brain wants it to follow the 1-to-1 line. Again I would probably have the existing methods in this comparison as well.Line 492 (and many other places): I might be misunderstanding, but I do not know where the idea came from that SAR data is not commonly corrected for tides. It has been standard operating procedure to correct slant ranges for tides for decades. Is it complicated to correct for tides in areas of ice flexure, ephemeral grounding or grounding line retreat? Yes, of course, but attempts have been made for a long time.Lines 530-534: This is a good idea! You should look into making that product.Line 537: NISAR has launched so maybe change "forthcoming" to "new and forthcoming"Figure S1: I assume these are estimates of ⍵. Correct? Similar to Figure 3, these are very small and hard to make out any features.References- Rignot, E., "Tidal motion, ice velocity and melt rate of Petermann Gletscher, Greenland, measured from radar interferometry," Journal of Glaciology, 42(142), 476-485 (1996). https://doi.org/10.3189/s0022143000003464
- Rignot, E. & MacAyeal, D. R., "Ice-shelf dynamics near the front of the Filchner-Ronne Ice Shelf, Antarctica, revealed by SAR interferometry," Journal of Glaciology, 44(147), 405-418 (1998). https://doi.org/10.3189/s0022143000002732
- Rignot, E., Padman, L., MacAyeal, D. R., & Schmeltz, M., "Observation of ocean tides below the Filchner and Ronne Ice Shelves, Antarctica, using synthetic aperture radar interferometry: Comparison with tide model predictions," Journal of Geophysical Research: Oceans, 105(C8), 19615-19630 (2000). https://doi.org/10.1029/1999jc00001
ReplyCitation: https://doi.org/10.5194/egusphere-2026-3795-RC1
Data sets
Code and data associated with the manuscript "Data-driven tide correction for SAR-derived ice velocity measurements on Antarctic ice shelves Yikai Zhu, Anna E. Hogg, Benjamin J. Wallis, Ross A. W. Slater, and Oliver J. Marsh https://doi.org/10.5281/zenodo.20815930
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 185 | 82 | 27 | 294 | 49 | 27 | 17 |
- HTML: 185
- PDF: 82
- XML: 27
- Total: 294
- Supplement: 49
- BibTeX: 27
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
