015
THE GRID
SECTION 13
ISSUE 001
OBSERVEDsafetynowconfidence / high
Weather Distribution-Shift Alarms
An AI forecast can drift because the world changed—or because its upstream analysis system did. Weather services need alarms for both: observing-network shifts, climate regimes, data pipelines, and physics-model upgrades. When distributions move, widen uncertainty, compare alternatives, and document the retraining decision before yesterday’s score becomes today’s false assurance.
Why this idea is here
What the evidence establishes.
ECMWF documented downstream effects from an analysis-system upgrade and stopped some external real-time models after verification; AIFS itself was adapted.
Source ledger
Read the sources.
- S01Farewell to the External AI Models
official verification report / published 2026-05-11 / retrieved 2026-07-09
- S02AIFS Machine Learning Data
official operational dataset / published 2026-05-12 / retrieved 2026-07-09