The glacier is now being asked to keep its own retreat schedule
the glacier is now being asked to keep its own retreat schedule: a model couples landsat imagery with climate reanalysis to forecast ice boundaries across five glaciers in four climate regimes, an ml emulator maps daily meltwater over greenland from 1979 to 2024, and a unet on sar watches high risk himalayan lakes for outburst floods.
the ice now files its own melt calendar.
Context
GlacierCastAI (July 5, 2026) combines Landsat time sequences, ERA5 climate reanalysis and terrain features across five glaciers in four climate regimes. Adding climate inputs improves image-only intersection-over-union from 0.326 to 0.337 in its reported ablation; the paper distinguishes forecasting future boundaries from mapping current ice.
Learning to melt (The Cryosphere, June 3, 2026) emulates daily Greenland surface melt from HIRHAM5 and DMIHH model output using atmospheric drivers and climate memory. Its reported mean absolute error is below 0.21 millimetres water equivalent per day relative to the model it emulates, across six drainage basins. A separate Himalayan lake paper uses Sentinel-1 SAR and a U-Net with an EfficientNet-B3 encoder on four named lakes, reporting validation IoU of 0.9130.
The melt error is measured against a physical model’s output, not direct observations. The original note’s 1979 to 2024 range was not established in the sources read and remains unsupported here, not refuted. GlacierCastAI reports single runs and notes the need for multi-seed evaluation; Gangotri has no test sequences under its temporal split. The Himalayan pipeline maps lake boundaries and proposes an early-warning architecture; segmentation skill alone is not evidence of successful outburst prediction.
Watch next
- Temporal coverage of the melt emulator. Out-of-region and multi-seed glacier forecasts. Prospective lake-growth alerts validated against observed flood events.
Sources
- GlacierCastAI: Predicting Glacier Retreat from Multi-Modal Satellite Imagery and Climate Signalsarxiv.org
- Learning to melt: Emulating Greenland surface melt from a polar RCM with machine learningtc.copernicus.org
- Targeted Semantic Segmentation of Himalayan Glacial Lakes Using Time-Series SAR: Towards Automated GLOF Early Warningarxiv.org
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 07:48 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →