ECMWF is at the forefront of research in numerical weather prediction.

We use the scientific advances made in areas such as data assimilation, Earth system modelling, predictability and reanalysis to improve our forecasts.

We work collaboratively with scientists across the world on all aspects of prediction systems as well as the many areas that support forecast production.

Data assimilation

We optimise methods that combine observations with the forecast model to produce an initial state of the Earth system.


We develop models for different components of the Earth system and their interactions.


We assess predictability at all forecast ranges and investigate methods to represent forecast uncertainty.

Real-time forecast data

We create global datasets of the recent history of the atmosphere, land surface, and oceans.


ECMWF is part of a vibrant meteorological community, leading and taking part in projects, making resources available to others, and sharing knowledge.

Key results are documented in ECMWF publications and in many cases in peer-reviewed journals.

Projects and collaborations

We are funded by our Member and Co-operating States but receive some extra funding from various bodies to carry out cutting-edge research, often in collaboration with other organisations.


We produce a range of publications documenting our latest research.

Member State Special Projects

Member States can use ECMWF’s supercomputer for projects likely to be of interest to the general scientific community.