Fusing Ice Surface Temperature with the AI4Arctic Dataset for Enhanced Sea Ice Concentration Estimation: A Preliminary Assessment
Bibliographic record
Abstract
While Arctic sea ice mapping supports several key applications (e.g., navigation, climate monitoring), its accuracy is impacted by remote sensing uncertainties and data limitations. The recent AI4Arctic dataset combines Sentinel-1 Synthetic Aperture Radar (SAR) imagery, AMSR2 brightness temperature (TB) measurements, ERA-5 reanalysis data, and ice charts to improve deep learning-based mapping approaches. Nevertheless, AI4Arctic excludes thermal infrared data and it is critical to explore the use of these products, which may improve predictions where SAR and passive microwave measurements are challenging to interpret. This study investigates the use of VIIRS ice surface temperature (IST) for improving SIC predictions. Our work builds on a competitive U-Net architecture, which estimates three parameters for automated sea ice mapping: SIC, stage of development, and floe size. A 30-scene subset of the AI4Arctic dataset is selected based on established criteria, and co-registered with VIIRS IST data. The impacts of fusing IST with other remote sensing data at the input- and feature-levels are explored using two fusion architectures. Experimental trials are conducted using these models, spanning six input channel combinations. Predictions are compared using evaluation metrics and SIC maps. When using IST measurements in combination with the original input channels, both the input- and feature-level approaches outperform the baseline model. This preliminary study suggests that IST data, in combination with TB measurements, improves predictions where ambiguous textures are present in SAR imagery or PM data is not able to contribute.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".