Earth & environment · OECD FoS 1.5 · China MoE 0706 大气科学
Meteorology & climate science
气象与气候科学
L3Systemic transformation ↑ Rising Score 3.35 · Last reviewed 2026-10-21 · Next review 2026-12-20
BIGGEST BREAKTHROUGH 2023-07-05 AI leapField leap
Pangu-Weather (Nature) ↗
First AI model to beat ECMWF’s operational NWP on medium-range skill
Why this level
AI weather models are operational: ECMWF's AIFS went operational in Feb 2025 and v2 on 12 May 2026; NOAA runs AI global models (AIGFS etc.). GraphCast, GenCast and Pangu laid the peer-reviewed groundwork.
S = 0.30·method + 0.35·discovery + 0.25·adoption + 0.10·education (education scores 0 when there is no evidence; if empty, the other weights are renormalised)
Sources
- ecmwf.int/en/about/media-centre/news/2025/ecmwfs-ai-forecasts-become-operational
- ecmwf.int/en/newsletter/187/news/implementation-aifs-v2
- noaa.gov/news-release/noaa-deploys-new-generation-of-ai-driven-global-weather-models
Composite 2.5–3.49 (or ≥3.5 with any of method/discovery/adoption below 3): AI is standard infrastructure with multiple independent breakthroughs. All levels →
Research directions
Data-driven medium-range forecastingProbabilistic ensemblesHybrid physics-AI climate modelsExtreme weather
Papers
- Advancing subseasonal forecasting with machine learning
AI: a light model learns any system’s biases and won ECMWF’s AI Weather Quest; Field: improves ECMWF subseasonal forecasts for 98% of precipitation targets
2026-10-14 · Nature
- A foundation model for the Earth system
Earth-system foundation model
2025-05-21 · Nature
- End-to-end data-driven weather prediction
Bypasses data assimilation, forecasting straight from observations
2025-03-20 · Nature
-
Field leap
Probabilistic weather forecasting with machine learningBeats ECMWF ENS on 97% of targets
2024-12-04 · Nature
- Neural general circulation models for weather and climate
Hybrid ML–physics GCM usable for climate
2024-07-22 · Nature
-
Field leap
Learning skillful medium-range global weather forecastingBeats HRES on 90% of targets, 10-day forecast in a minute
2023-12-22 · Science
-
AI leapField leap
Accurate medium-range global weather forecasting with 3D neural networksFirst AI model to beat ECMWF’s operational NWP on medium-range skill
2023-07-05 · Nature
- Space weather dependence of decameter-scale ionospheric irregularities spatial distribution
Data-driven self-learning pattern recognition for nowcasting and forecasting high-latitude ionospheric irregularities.
2026-10-21 · arXiv · Oleg I. Berngardt, Ivan A. Lavygin
- Disentangling forced and internal components in observed climate evolution using spatiotemporal deep learning
A ConvLSTM separates forced response from internal variability, finding internal variability dominated the La Niña-like Pacific pattern.
2026-10-20 · arXiv · Parvathi Kooloth, Wenyu Zhou, L. Ruby Leung, Shang-Ping Xie
- AirMPA: A Meteorology-to-Pollution Adapter for Global Air Quality Forecasting
AirMPA, a decoupled meteorology-to-pollution adapter for global air-quality forecasting.
2026-10-19 · arXiv · Yiheng Wang, Kai Zheng, Yuetan Lin, Fanglu Fan
- Spherical Physics-informed Neural Operator with Multi-scale Coupling for Meteorological Downscaling
A spherical physics-informed neural operator with multi-scale coupling for meteorological downscaling.
2026-10-17 · arXiv · Yiqiang Ye, Yichi Wang, Haodi Zhang, Jiawei Wen
- Energy Flow Shapes the Decomposition Dynamics of a Reactive Atmospheric Intermediate
Neural-network potential MD simulations of Criegee intermediate decomposition under different excitations.
2026-10-15 · arXiv · Mitradip Das, Cangtao Yin, M. Meuwly
- Ranked Right, Sized Wrong: Ensemble Spread Fails Silently at Streamflow Extremes
LSTM streamflow ensemble spread under-represents errors at extreme flows.
2026-10-14 · arXiv · Justin Gallant, Tyler Wilson
- Improved Precipitation Variability in the Community Atmosphere Model through Implementation of a Neural-Network Deep Convection Scheme
A DNN trained on storm-resolving output replaces CAM6’s deep convection scheme, improving precipitation variability.
2026-10-14 · arXiv · Xavier J. Levine, Griffin Mooers, Judith Berner, Jack W. Atkinson
- Generative modelling for multivariate spatio-temporal downscaling of monthly climate fields
Generative modelling for joint spatio-temporal downscaling of monthly climate fields.
2026-10-14 · arXiv · Yu Huang, Shangshang Yang, Sebastian Bathiany, Michael Aich
- From arXiv, bioRxiv, medRxiv and OpenAlex (last ~30 days); summaries written by AIMap from the abstract.
Milestones 12· 4 leaps
Big glowing nodes are leaps: AI leap cool · Field leap warm · double ring = both
-
Deployment
CMA’s Fengqing AI global ensemble forecast system V1.0 passes operational review and goes live (11-day skill)AI: a fully domestic AI ensemble enters national operations; Field: 11.0-day skill at 500 hPa, some scores above ECMWF systems (developer evaluation)
中国新闻网 -
Paper
Advancing subseasonal forecasting with machine learning — Probabilistic Bias Correction (Nature)AI: a light model learns any system’s biases and won ECMWF’s AI Weather Quest; Field: improves ECMWF subseasonal forecasts for 98% of precipitation targets
Nature -
Deployment
ECMWF IFS 50r1 and AIFS v2 go live; AIFS ENS adds waves and snow coverSecond operational generation of AI forecasts at ECMWF
ECMWF -
Deployment
NOAA deploys new AI-driven global weather models (AIGFS etc.)NOAA deploys AI global models
NOAA -
Paper
Aardvark Weather: end-to-end data-driven forecasting from observations (Nature)Bypasses data assimilation, forecasting straight from observations
Nature -
DeploymentField leap
ECMWF AIFS becomes operationalFirst operational AI model at a top forecasting centre
ECMWF -
PaperField leap
GenCast probabilistic ensemble forecasting (Nature)Beats ECMWF ENS on 97% of targets
Nature -
PaperField leap
GraphCast medium-range forecasting (Science)Beats HRES on 90% of targets, 10-day forecast in a minute
Science -
PaperAI leapField leap
Pangu-Weather (Nature)First AI model to beat ECMWF’s operational NWP on medium-range skill
Nature
No milestones match this filter.
More papers & repos 11
| Type | Title | Date |
|---|---|---|
| Paper | Advancing subseasonal forecasting with machine learning | 2026-10-14 |
| Paper | A foundation model for the Earth system | 2025-05-21 |
| Paper | End-to-end data-driven weather prediction | 2025-03-20 |
| Paper | Probabilistic weather forecasting with machine learning | 2024-12-04 |
| Paper | Neural general circulation models for weather and climate | 2024-07-22 |
| Paper | Learning skillful medium-range global weather forecasting | 2023-12-22 |
| Paper | Accurate medium-range global weather forecasting with 3D neural networks | 2023-07-05 |
| Repo | google-deepmind/graphcast To be verified | — |
| Repo | ecmwf/anemoi-core To be verified | — |
| Repo | google-research/neuralgcm To be verified | — |
| Repo | microsoft/aurora To be verified | — |
Latest news 11
- AI reveals how a warming climate will intensify NZ’s most extreme downpoursThe Conversationen
- Interpretability of AI weather models via intermediate decodingarXiv physics.ao-phenPaperPreprint
- AI and climate: Hero or villain?Latitude Mediaen
- Forecasting space weather risks on power gridsMicrosoft Researchen
- Using AI to plan for climate variability in renewable energy systemsPhysics World（IOP）en
- Introducing WeatherNext 3, our most advanced and accurate global weather AI modelGoogle DeepMind Blogen
- WeatherNext: AI model achieves breakthrough in forecasting cyclonesGoogle DeepMind Blogen