The physics model just lost its own scoreboard. ecmwf reports its ai ensemble beats its…
the physics model just lost its own scoreboard. ecmwf reports its ai ensemble beats its physics-based system by up to 20 percent on surface temperature, per oct 8 numbers, since going operational in july 2025.
equations are now the second-best forecast.
Context
A tuput roundup dated Oct 8 reports that ECMWF put an AI ensemble into operation on 1 July 2025 and reports gains of up to 20 percent on surface temperature. ECMWF's 2025 annual report says AIFS ENS became operational as an ensemble of 51 forecasts, faster and more energy efficient while maintaining high skill.
A 2026 arXiv comparative study looks at AI and physics-based forecasting at ECMWF, and a Science Advances study cited in the roundup found three other AI models lose to a physics model on record-breaking heat, cold and wind.
The 20 percent surface temperature gain and the July 2025 operational date match the Oct 8 roundup. The annual report confirms AIFS ENS went operational but the excerpt read gives no percentage.
'Equations are now the second-best forecast' is the author's opinion. The extremes studies suggest physics models still hold an edge on record events, which the post does not claim either way.
Related work
- arXiv: AI and physics-based weather forecasting, a comparative study ↗ECMWF AI versus physics.
- Forecasting extremes: AIFS and km-scale IFS ↗Where AI still lags on extremes.
- ECMWF annual report 2025 ↗AIFS ENS operational.
Watch next
- Read the arXiv comparison for variable by variable scores.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 10 October 2026 at 11:06 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 →