High Resolution Sea Ice Concentration Using a Sentinel-1 U-Net Ice-Water Classifier
Bibliographic record
Abstract
Accurate and high-resolution sea ice concentration (SIC) mapping is essential for polar navigation, environmental monitoring, and assimilation into forecast models. Traditional passive microwave sensors, such as AMSR2, provide reliable SIC estimates but are limited by coarse (5 km) resolution, particularly near coastlines and regions with mixed ice and open water, where finer spatial detail is critical. Synthetic aperture radar (SAR) imagery offers a high-resolution alternative. This study applies a U-Net convolutional neural network to Sentinel-1 SAR data, utilizing pixelwise ice-water labels to enhance SIC mapping. To address SAR noise challenges, we incorporate multilooking, adaptive noise correction, and overlapping patches at inference to improve SIC accuracy while preserving fine-scale features. We trained the U-Net across multilooking levels to balance resolution, noise reduction, and computational efficiency, allowing the model to handle noise artefacts effectively. Our results identify an optimal 7 × 7 multilooking level, achieving 280 m ice-water labels and a 2.5 km SIC field when an additional 9 × 9 SIC window is applied. This configuration enhances traditional SIC products by improving the representation of the ice edge, leads, and near-coastal features, which are critical for operational applications. SAR-derived SIC addresses the limitations of passive microwave products by providing superior spatial detail and ice edge resolution. Incorporating additional information from AMSR2 or wind features could strengthen SIC robustness and minimize the misclassification of open water, which is present in the results. These advancements would establish SAR-based SIC as a valuable tool for operational sea ice monitoring and integration into high-resolution models.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".