The river in the sky is now being asked to give five days of notice
the river in the sky is now being asked to give five days of notice: a physically guided model flags flood risk during atmospheric rivers with a 0.94 auc where logistic baselines score 0.81, noaa put an experimental ar model online with daily updates out to five days, and aurora reproduces the december 2023 california river at seven day leads.
the sky now files its own water report.
Context
The Scientific Reports paper develops a physically guided, explainable machine learning framework to predict flood occurrence during atmospheric river conditions worldwide, and says the best model reaches a ROC AUC of 0.94, outperforming a logistic regression baseline at 0.81. NOAA's Weather Program Office reported on March 17, 2026 that an experimental forecast model for atmospheric rivers, the Atmospheric River Analysis and Forecast System, is online and gives daily updates up to five days in advance.
The Aurora paper presents a foundation model of the atmosphere.
The 0.94 and 0.81 match the abstract. The paper's best model is a tree-based model per a secondary report, and the baseline is logistic regression, so the note's 'logistic baselines' fits. The note calls the NOAA system an AR model with daily updates to five days, which matches NOAA, and NOAA says it is built on the Hurricane Analysis and Forecast System's core code, so it is a physics-based system and not an AI model, though the post does not say AI.
The claim that Aurora reproduces the December 2023 California river at seven day leads was not found in the pages read, so it is unsupported here, not refuted. The note joins three unrelated efforts, one flood-risk classifier, one operational forecast model and one foundation model.
'The sky now files its own water report' is the author's line.
Related work
- Assessing global flood risk from atmospheric rivers through physically guided machine learning (Scientific Reports, August 18, 2026) ↗Source for the AUC of 0.94 against 0.81 and the framework.
- Experimental atmospheric river forecast model goes online (NOAA Weather Program Office, March 17, 2026) ↗Source for the experimental forecast model and its five day daily updates.
- Aurora: A Foundation Model of the Atmosphere (arXiv 2405.13063) ↗Source for the Aurora foundation model.
Watch next
- Find the Aurora case study on California atmospheric rivers. Read the flood paper's test split.
Sources
- Assessing global flood risk from atmospheric rivers through physically guided machine learning (Scientific Reports, August 18, 2026)nature.com
- Experimental atmospheric river forecast model goes online (NOAA Weather Program Office, March 17, 2026)wpo.noaa.gov
- Aurora: A Foundation Model of the Atmosphere (arXiv 2405.13063)arxiv.org
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 3 October 2026 at 19:07 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 →