#GraphCast
Yash Diggikar, Milton Halem, Aishwarya Raman: An AI-Based OSSE Framework to Evaluate Learned Physical Consistency: A Multi-Model Study Using FourCastNet, Aurora, Prithvi WxC, and GraphCast https://arxiv.org/abs/2611.22024 https://arxiv.org/pdf/2611.22024 https://arxiv.org/html/2611.22024
November 18, 2026 at 6:42 AM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 16, 2026 at 9:10 PM
No, an AI that never learned physics just beat the world's best weather model. DeepMind's GraphCast outpredicted Europe's own physics-based model on 90% of forecast targets.

Like + repost if useful.
Joseph, geospatial scientist reading the planet from space. youtube.com/@Routero
November 12, 2026 at 5:02 PM
Ein guter Artikel von @cleanthinking.wsocial.eu zur KI in der Wettervorhersage, inklusive ihrer Stärken und Schwächen.

www.cleanthinking.de/ki-wettervor...
KI-Wettervorhersage: AIFS, AICON und ihre Grenzen
KI-Wettervorhersage erklärt: Wie AIFS, AICON und GraphCast rechnen, warum sie ERA5 brauchen und wo sie an Rekorden scheitern.
www.cleanthinking.de
November 8, 2026 at 3:50 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 18, 2026 at 7:30 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 17, 2026 at 7:58 PM
Global Weather Model Leaderboard: 전 세계 기상 예측 모델을 매일 같은 조건으로 채점하는 공개 리더보드 공개

소개

VIDRAFT가 Hugging Face에 기상 예측 관련 공개물 두 가지를 올렸습니다. 하나는 전 세계 주요 예보 모델의 정확도를 매일 채점하는 Global Weather Model Leaderboard이고, 다른 하나는 여러 전 지구 예보를 융합하는 예측 시스템 Darwin-Prediction-Weather입니다.

기상 예측 분야에서는 최근 몇 년 사이 GraphCast, Pangu-…
Global Weather Model Leaderboard: 전 세계 기상 예측 모델을 매일 같은 조건으로 채점하는 공개 리더보드 공개
소개 VIDRAFT가 Hugging Face에 기상 예측 관련 공개물 두 가지를 올렸습니다. 하나는 전 세계 주요 예보 모델의 정확도를 매일 채점하는 Global Weather Model Leaderboard이고, 다른 하나는 여러 전 지구 예보를 융합하는 예측 시스템 Darwin-Prediction-Weather입니다. 기상 예측 분야에서는 최근 몇 년 사이 GraphCast, Pangu-Weather, AIFS 같은 AI 모델이 등장해 물리 기반 수치예보(NWP) 모델과 함께 운영되고 있습니다. 모델 간 비교는 주로 WeatherBench 2처럼 과거 연도를 대상으로 한 벤치마크로 이루어졌습니다. 이 리더보드는 대상을 "올해, 매일"로 옮겼다는 점이 다릅니다. 현업에서 실제로 발표되는 예보를 날마다 기록하고, 관측값이 확정되면 채점해 누적 순위를 갱신합니다. 이 글에서는 리더보드의 평가 설계, 데이터 파이프라인, 융합...
discuss.pytorch.kr
November 13, 2026 at 12:45 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 15, 2026 at 5:46 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 13, 2026 at 6:35 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 11, 2026 at 6:39 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 12, 2026 at 6:58 PM
新しい記事を公開しました!

【2027年最新】AI×科学研究(AI for Science)10大革命!AlphaFold 3創薬・新素材発見・核融合制御・GraphCast気象予測までノーベル賞級の突破を解説
https://ama-industry.jp/future-2027-ai-science-nobel-prize-alphafold3-materials-weather-forecast-p254/
November 8, 2026 at 2:43 AM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 10, 2026 at 6:52 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 9, 2026 at 9:39 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 8, 2026 at 5:11 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 7, 2026 at 5:01 PM
No, Earth's best forecaster isn't a meteorologist. DeepMind's GraphCast beat Europe's top model on 89.3% of 2,760 forecast targets. Lam et al. 2023, Science. doi.org/10.1126/science.adi2336

Like + repost if useful.

Joseph, geospatial scientist reading the planet from space. youtube.com/@Routero
November 4, 2026 at 8:03 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 6, 2026 at 6:09 PM
This is literally my day job and it still blows my mind sometimes.

ECMWF now shows ML weather forecasts from FourCastNet, Pangu-Weather, and GraphCast alongside the operational IFS on our public charts. Same verification, same standards. Plus we open-sourced ai-models so anyone can run them.
This is literally my day job and it still blows my mind sometimes.
You can compare physics-based and ML forecasts yourself, right now.
amplt.de
November 3, 2026 at 10:58 AM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 5, 2026 at 6:53 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 4, 2026 at 6:45 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 3, 2026 at 6:45 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 2, 2026 at 8:59 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
November 1, 2026 at 5:12 PM
Best raw 2 m temperature forecast of the models with n≥30 at lead day 1 over the last 90 days: ECMWF AIFS Single (2.3 °F MAE, n=89)
Instantaneous values at 00/06/12/18 UTC vs the observation at the same instants. Raw 0.25° model output, no post-processing.
https://castcheck.zifanzhang.com
October 31, 2026 at 4:16 PM