aceModerate evidence
Published recordClimateGood Space
More
Atmospheric Predictability beyond 30 Days with Machine Learning
- Truvace record
- TRV-2026-2623raw data (JSON)
- Version
- v1Published
- First published
- Evidence type
- Benchmark or lab evaluation
- Sources
- 1 sourcePeer-reviewed
- sha256
- 503d47686204…check it
Abstract Atmospheric predictability research has long held that rapid error growth at small spatial scales imposes an intrinsic limit of roughly 2 weeks on deterministic weather forecast skill. We challenge this limit using GraphCast, a machine learning weather model, by optimizing initial conditions for twice-daily forecasts spanning 2020. This approach yields an average error reduction of 86% at 10 days relative to control forecasts from reanalysis initial conditions, with skill lasting beyond 30 days. Mean op…
The gain
In a study of 732 twice-daily hindcasts for 2020, optimizing starting conditions in the GraphCast AI weather model cut 10-day error by 86% against control forecasts and kept significant skill to 33 days, more than double the usual two-week limit.
What this doesn’t fix
The optimization uses knowledge of the verifying trajectory, so the authors say it does not yet identify such initial conditions in real time, which would be required before any operational forecasting gain. The 1.0° model also does not resolve mesoscale processes, so the findings apply only to synoptic and planetary scales.
Sources
1 on file- Artificial Intelligence for the Earth Systemsdoi:10.1175/aies-d-26-0009.1
Cite this record
v1Truvace Impact Record TRV-2026-2623, v1: “Atmospheric Predictability beyond 30 Days with Machine Learning.” Truvace, 2026-11-12. /record/TRV-2026-2623 (accessed on the day you copy it). sha256 503d476862041b84…
Change history
1 versionEvery change made to this record since it was first published, and why. Nothing has changed since it went up.
Certified into the record
Check this record hasn’t been altered
How to check it yourself, without trusting this page
Download any version from /api/record/TRV-2026-2623 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.