Global medium-range NWP★ Highlight
AIFS -- ECMWF's data-driven forecasting system
Why it matters. ECMWF's operational data-driven AIFS, a GNN-transformer model running alongside IFS with open data; a key reference for operational ML weather forecasting.
In one sentence. AIFS is ECMWF's operational GNN-transformer forecast system, trained on ERA5 and operational analyses, with skilled upper-air, surface and tropical cyclone forecasts run publicly.
Abstract
Machine learning-based weather forecasting models have quickly emerged as a promising methodology for accurate medium-range global weather forecasting. Here, we introduce the Artificial Intelligence Forecasting System (AIFS), a data driven forecast model developed by the European Centre for Medium-Range Weather Forecasts (ECMWF). AIFS is based on a graph neural network (GNN) encoder and decoder, and a sliding window transformer processor, and is trained on ECMWF's ERA5 re-analysis and ECMWF's operational numerical weather prediction (NWP) analyses. It has a flexible and modular design and supports several levels of parallelism to enable training on high-resolution input data. AIFS forecast skill is assessed by comparing its forecasts to NWP analyses and direct observational data. We show that AIFS produces highly skilled forecasts for upper-air variables, surface weather parameters and tropical cyclone tracks. AIFS is run four times daily alongside ECMWF's physics-based NWP model and forecasts are available to the public under ECMWF's open data policy.
Labels
- Topic
- Global medium-range NWP
- Type
- New model or method
- Method
- Graph neural network, Transformer / ViT
- Time scale
- 3–15 days
- Domain
- Global
- Data
- ERA5 / reanalysis, NWP analysis or forecasts (IFS, GFS, ...)
- Status
- Operational
- Identifiers
- arXiv:2406.01465 OpenAlex:W4399401448
- Added
- 2026-10-22
Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.