The sun is now being asked to file its own flare diary
the sun is now being asked to file its own flare diary: nasa's coffies model predicts storm causing active regions up to 12 hours before they appear, the surya foundation model trained on 9 years of sdo observations forecasts flares winds and irradiance, and sepnet jointly predicts flares, coronal mass ejections and energetic particles with lstm and transformer blocks.
the star now files its own storm watch.
Context
NASA’s August 14, 2026 COFFIES account describes a sliding-window transformer that detects subtle acoustic and magnetic precursors of active-region emergence up to 12 hours before the region appears. This is an emergence forecast, not a direct forecast of a particular storm.
NASA-IMPACT’s Surya repository describes a 366-million-parameter solar foundation model using SDO AIA and HMI observations. Its training section names 2011 through 2019, and downstream tasks include flare forecasting, solar-wind prediction and EUV spectra modeling.
The SEPNET paper jointly predicts flare and coronal-mass-ejection counts and solar energetic particle probability. The base model uses feed-forward layers; the SEPNET-TS variant adds an LSTM and transformer encoder. NASA CCMC describes an operational forecast of particle events exceeding 10 pfu at energies of at least 10 MeV in the next 24 hours.
An emerging active region does not necessarily produce a damaging eruption. Surya is a reusable representation model with downstream tasks, while SEPNET is a multi-task particle-forecasting system; they are not interchangeable storm forecasts. LSTM and transformer blocks belong to SEPNET-TS rather than every SEPNET version. The SEPNET paper acknowledges relatively high false-alarm ratios and shows a limited real-time November 2025 example, not proof of universally reliable warnings.
Watch next
- Prospective false-alarm and missed-event rates during a full solar cycle. Whether active-region emergence forecasts improve operational warning decisions beyond existing inputs.
Sources
- NASA’s COFFIES Uses AI to Predict Storm-Causing Active Regionsscience.nasa.gov
- NASA-IMPACT/Suryagithub.com
- Solar Energetic Particle Forecasting with Multi-Task Deep Learning: SEPNETarxiv.org
- SEPNET 1 (NASA CCMC)ccmc.gsfc.nasa.gov
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 4 October 2026 at 07:18 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 →