GoKawiilGoogle DeepMind details machine-learning system for subseasonal weather forecasting
Researchers describe a probabilistic forecasting method for predicting weather two to six weeks ahead, built on ERA5 reanalysis data from 1979-2026 and evaluated using ECMWF's AI Weather Quest competition framework. The approach forecasts temperature, precipitation and sea-level pressure as probability distributions across climatological quantile bins, scored using the Ranked Probability Skill Score.
GoKawiil's interpretation of the reporting above, not reported fact.
Subseasonal forecasting, the gap between short-term weather prediction and long-term seasonal outlooks, has historically been a weak point for meteorological models, so improvements here could help sectors like agriculture and energy plan weeks in advance. The framing around ECMWF's competition suggests the work is positioned to be benchmarked against established forecasting centers rather than evaluated in isolation.
- The method predicts probability distributions over five climatological quantile bins rather than single-point forecasts
- It targets 2-meter temperature, precipitation and mean sea-level pressure over week-long periods
- Performance is measured using the Ranked Probability Skill Score against ERA5 reanalysis ground truth
Source: nature.com — Guan, 2026-10-14
Published there as: “Advancing subseasonal forecasting with machine learning”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.