The fog is now being asked to give its own visibility log
the fog is now being asked to give its own visibility log: a generative network nowcasts dense marine fog under 400 meters on the grand banks with 95 percent intervals, a physics informed forecaster predicts the four cloud species that matter for aviation seven days out, and a multi satellite hub detects marine fog across ocean regions.
the haze now files its own clearance forecast.
Context
The Grand Banks study (arXiv, NeurIPS 2023 climate workshop) uses data from the FATIMA campaign in July 2022 near Sable Island, with a forward scatter visibility sensor and a weather transmitter on a research vessel. A conditional GAN and an XGBoost baseline nowcast visibility 30 and 60 minutes ahead. At 30 minutes the cGAN had the best RMSE for visibility under 1 km (0.151 km); at 60 minutes XGBoost was best (0.167 km) and the cGAN was similar.
AviaSafe (CVPR 2026) is a hierarchical, physics-informed forecaster for four cloud hydrometeor species, global, six-hourly, up to 7 days, with the Icing Condition index as a physics constraint. Trained on ERA5, it reports lower RMSE than baseline and better results than operational numerical models on certain key variables at 7 days.
MFogHub (CVPR 2025) is a dataset of annotated marine fog from 15 coastal regions and six geostationary satellites with over 68,000 samples, built for marine fog detection and forecasting benchmarks.
The note's 'under 400 meters' and '95 percent intervals' are not in the Grand Banks abstract, which reports 1 km and 10 km visibility thresholds, 30 and 60 minute leads and RMSE values, so those two details are unsupported here, not refuted. The paper also says XGBoost won at 60 minutes, and that nowcasting at 30 minutes was 'quite difficult'. AviaSafe predicts four species of cloud water and ice for icing risk, which the note calls 'cloud species that matter for aviation', and says it beats operational models only on certain variables. MFogHub is a benchmark dataset, so the note's 'multi satellite hub detects marine fog' describes data and baselines, not a deployed detector.
Related work
- Generative Nowcasting of Marine Fog Visibility in the Grand Banks area and Sable Island (arXiv) ↗Source for the cGAN and XGBoost nowcast results.
- AviaSafe: A Physics-Informed Data-Driven Model for Aviation Safety-Critical Cloud Forecasts (arXiv) ↗Source for the four species, 7 day cloud forecasts.
- MFogHub: Bridging Multi-Regional and Multi-Satellite Data for Global Marine Fog Detection and Forecasting (arXiv) ↗Source for the 15 region, six satellite fog dataset.
Watch next
- Look for a fog nowcast paper that reports a 400 metre threshold and 95% intervals. Check the Grand Banks result on campaigns beyond one month of observations.
Sources
- Generative Nowcasting of Marine Fog Visibility in the Grand Banks area and Sable Island (arXiv)arxiv.org
- AviaSafe: A Physics-Informed Data-Driven Model for Aviation Safety-Critical Cloud Forecasts (arXiv)arxiv.org
- MFogHub: Bridging Multi-Regional and Multi-Satellite Data for Global Marine Fog Detection and Forecasting (arXiv)arxiv.org
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 02:03 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 →