AI & Environment Resource HubKent, Chris; Scaife, Adam A.; Dunstone, Nick J.; Smith, Doug; Hardiman, Steven C.; Dunstan, Tom; Watt-Meyer, Oliver · 2025-03-31
A machine learning weather model, ACE2, trained on reanalysis data, demonstrates skillful global seasonal forecasting, comparable to leading physics-based models—especially in predicting the North Atlantic Oscillation, suggesting ML's potential in near-term climate prediction.
Kent, Chris; Scaife, Adam A.; Dunstone, Nick J.; Smith, Doug; Hardiman, Steven C.; Dunstan, Tom; Watt-Meyer, Oliver. 2025-03-31. Skilful Global Seasonal Predictions From a Machine Learning Weather Model Trained on Reanalysis Data. https://arxiv.org/abs/2503.23953 (AI & Environment Resource Hub; record paper-195; collection snapshot 2026-09-15).
Included in the original Scientific Paper collection. Labels below are metadata based suggestions. They are not verified findings, claims of effectiveness, or endorsements.
Source sheet: Scientific Paper · Row 199 · Original ID: paper-195.
A paper listing is not a quality assessment. Peer review and findings require source-level confirmation. Source link reachable · checked 2026-09-15. Source identity and required metadata verified. The import date is not the original date added.
Suggest a correction for this record →Ordered by the number of shared provisional topic labels. This indicates a browsing connection, not agreement between sources.
Shared topics: Education and skills; Climate and greenhouse gases; Weather, hazards, and adaptation
The Artificial Intelligence Forecasting System (AIFS), developed by ECMWF, leverages graph neural networks and transformers to produce highly skilled medium-range weather forecasts, running alongside traditional numerical models and providing open-access predictions.
Shared topics: Education and skills; Climate and greenhouse gases; Weather, hazards, and adaptation
This paper introduces a proof-of-concept system that combines a curated knowledge graph with AI agents to overcome persistent barriers in climate data science, such as fragmented datasets and high technical skill requirements. By enabling natural-language interaction, automated data access, and cloud-native workflows, the system lowers entry barriers, improves reproducibility, and demonstrates how a knowledge-graph-driven approach can democratize climate data science and support scalable human–AI collaboration.
Shared topics: Education and skills; Climate and greenhouse gases; Weather, hazards, and adaptation
This study presents a machine learning framework for high-resolution regional weather forecasting using boundary-forced graph-based models, demonstrating strong predictive skill and lower computational costs compared to traditional methods.
Shared topics: Education and skills; Climate and greenhouse gases; Weather, hazards, and adaptation
Using natural language processing and interviews, this study finds that vague definitions and geographic inaccessibility of green jobs in Worcester, MA hinder the development of a climate-ready workforce and the implementation of local climate adaptation plans.