the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Grounded Icebergs around Antarctica: A High-Resolution Dataset Derived from Deep Learning and Sentinel-1 Synthetic Aperture Radar
Abstract. Icebergs frequently run aground on shoals on the continental shelf around Antarctica. Once rendered immobile, they can anchor landfast sea ice (fast ice) and their supply of limiting trace nutrients contribute to driving coastal marine productivity. They are also associated with seabed scouring, with implications for benthic marine ecosystems. Despite their importance, there is currently a lack of accurate, continent-wide automated mapping of grounded iceberg distribution and size and as a consequence, no large-scale, complete map of grounded icebergs exists. To address these gaps, this study implements an automated grounded iceberg detection framework based on Sentinel-1 Synthetic Aperture Radar (SAR) imagery, integrating a proposed ResUNet deep-learning network with a multi-temporal identification algorithm, incorporating strict physical constraints derived from bathymetry and sea ice concentration to mitigate environmental false positives. The method demonstrates strong robustness against interference from complex coastal conditions, achieving a detection F1 score exceeding 0.91 and successfully reducing the minimum detectable iceberg size to 0.016 km². Since the presence of fast ice makes distinction between "truly grounded icebergs" and "those held motionless by fast ice", we capitalise on the unprecedented low fast ice extent in early 2025 (late February to early April) to construct the first high-resolution, continent-wide dataset of grounded icebergs around Antarctica. A total of 38,905 stationary icebergs were identified on the Antarctic continental shelf. We partition these stationary targets into "high-confidence grounded icebergs" and fast ice entrapped candidates. Analysis shows that 70.5 % of these targets are identified as high-confidence grounded. Across the entire dataset, tiny icebergs (< 1 km²) dominate numerically, accounting for 92.2 % of the total. These high-density grounded iceberg clusters are primarily concentrated on shallow continental shelves and west of (i.e., downstream of) the actively disintegrating fronts of ice shelves, forming complex and discontinuous "grounded iceberg chains". Our dataset reveals that these grounded icebergs cover a combined area of 13,719 km² and are widely distributed along 56.7 % of the Antarctic coastline, with just 14.2 % of the coastline containing 80 % of all grounded icebergs. Crucially, while typically overlooked, tiny icebergs (< 1 km²) contribute 52.6 % of this total grounded area. These dense clusters imply a potential "picket fence effect", providing a quantitative baseline for modelling fast ice stability, assessing nutrient fluxes, and mapping benthic habitats. The grounded iceberg dataset (Jiao et al., 2026) is available at https://doi.org/10.25959/54sx-pt47.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Earth System Science Data.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(31164 KB) - Metadata XML
-
Supplement
(53566 KB) - BibTeX
- EndNote
Status: open (until 09 Jul 2026)
-
RC1: 'Comment on essd-2026-179', Anonymous Referee #1, 30 Apr 2026
reply
-
AC1: 'Reply on RC1', Kaihong Jiao, 18 Jun 2026
reply
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2026-179/essd-2026-179-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Kaihong Jiao, 18 Jun 2026
reply
-
RC2: 'Comment on essd-2026-179', Anonymous Referee #2, 11 Jun 2026
reply
This study presents the first high-resolution circum-Antarctic grounded iceberg dataset by integrating Sentinel-1 SAR imagery with a ResUNet deep learning model, multi-temporal trajectory matching, and physical constraints derived from bathymetry and sea ice concentration. This dataset provides an important basis for future quantitative analyses of the impacts of grounded icebergs on the Southern Ocean environment, the growth and maintenance of landfast ice, and the stability of near-coastal ice shelves. Overall, this study has considerable scientific value and application potential, and I recommend publication after minor revision.
My specific comments are as follows:
Although the data publication page provides a general description of the main GeoPackage dataset, it still appears to lack detailed information on the individual attribute fields and their units. I suggest that the authors add a README file or a field-description table to the dataset, clearly listing the meaning, data type, and unit of each variable.
Page 5, Line 123: The expression “0.016 km2 (40 × 40 m)” is potentially ambiguous, because 40 × 40 m corresponds to 0.0016 km2 rather than 0.016 km2. According to the Methods section, connected components smaller than 10 pixels were removed as noise. Therefore, I suggest that the authors clarify here that the minimum retained iceberg area is 0.016 km2, corresponding to 10 pixels at a 40 m spatial resolution.
Page 6, Lines 139-140: The authors state that EW imagery was primarily used, while IW imagery was used to fill gaps in EW coverage. However, it remains unclear whether the algorithm directly performs cross-mode matching when two consecutive images over the same region are acquired in EW and IW modes, respectively, or whether IW imagery is only used to supplement spatial coverage and trajectories are continued only within the same acquisition mode.
I suggest adding a set of comparison figures to visualise cross-mode matching, including the EW and IW scenes, the predicted iceberg masks under both modes, and the differences in iceberg outlines.
Page 10, Line 198: The manuscript states that repeat-pass images with temporal baselines of 12 days or multiples thereof were used for grounded iceberg identification for each orbital track. It is unclear whether all Antarctic coastal regions have at least two SAR images available for matching. I suggest that the authors clarify whether any regions had only a single valid SAR acquisition. If such regions exist, the authors should explain whether they were excluded from grounded iceberg identification.
Page 16, Line 198: The authors use SIC and bathymetry data to remove iceberg candidates, but these auxiliary datasets have limitations in spatial resolution and accuracy. In particular, the AMSR2 SIC product has a spatial resolution of 10 km, which is much coarser than the scale of small icebergs. It may therefore fail to accurately represent the local sea ice conditions surrounding individual icebergs. As a result, some truly grounded icebergs may be incorrectly removed due to SIC uncertainty, leading to false negatives. I suggest that the authors further discuss this potential source of uncertainty.
Page 19-20, Lines 424-435: The authors state that the model shows strong cross-seasonal generalisation. However, the main validation area is a 50 × 50 km region in Adélie Land. Whether the validation results from this region are representative of the entire circum-Antarctic coastline requires further clarification.
Citation: https://doi.org/10.5194/essd-2026-179-RC2 -
AC2: 'Reply on RC2', Kaihong Jiao, 18 Jun 2026
reply
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2026-179/essd-2026-179-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Kaihong Jiao, 18 Jun 2026
reply
Data sets
Antarctic Grounded Iceberg Dataset Kaihong Jiao, Alexander D. Fraser, Johannes Lohse, Pat Wongpan, Caitlin Adams, and Alexander C. Bradley https://doi.org/10.25959/54sx-pt47
Grounded Iceberg Detection Dataset (Training, Test, and Validation) Kaihong Jiao https://doi.org/10.5281/zenodo.19228332
Model code and software
Antarctic Grounded Iceberg Detection Code Kaihong Jiao https://github.com/kj0703/Antarctic-Grounded-Iceberg-Detection
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 452 | 213 | 27 | 692 | 311 | 25 | 26 |
- HTML: 452
- PDF: 213
- XML: 27
- Total: 692
- Supplement: 311
- BibTeX: 25
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
See attached PDF-document. I did not check the data sets and the software code.