AI and Met Office Partnerships
Artificial intelligence (AI) and machine learning (ML) is advancing rapidly, and the Met Office is at the forefront of harnessing the power of these technologies to enhance weather and climate science and services.
However, since a concerted and interdisciplinary effort is required, our ambition and vision for the application of data science can only be fully realised through partnership. By working across organisations, we can accelerate the application and adoption of emerging approaches and capabilities.
The Met Office have a number of existing multilateral relationships that support this endeavour, for example the Momentum Partnership with leading international operational weather forecast centres; the Met Office Academic Partnership, bringing together institutions which are among the primary universities in weather, climate and AI/ML; and the UK National Climate Science Partnership with the Natural Environment Research Council (NERC), that drives forward national capability and research in predicting weather and climate.
In addition, the Met Office have new programmes of collaborative AI/ML work with a number of world-leading organisations such as ECMWF – where we are one of several national meteorological services exploring the application of their open-source Anemoi framework for machine learning weather prediction, and participating in the WeatherGenerator project that aims to build a machine learning-based “foundation model” of the Earth System that can be used and adapted for a large range of specific tasks – alongside more traditional numerical weather prediction engagement.
Other relationships, more exclusively focussed on data science, include our formal partnership with The Alan Turing Institute – the UK’s national institute for data science and AI – focussed on the development of FastNet, an experimental AI-powered weather model built using graph neural networks.
Our working with technology companies such as Microsoft (with whom we joined forces in 2021 to build the world’s most advance supercomputer dedicated to weather and climate), alongside other key partners and suppliers, also presents considerable further opportunities to advance data science.
New institutional relationships are coordinated via our Science Partnerships team.