AI and Met Office capabilities

It requires a vast amount of observational data, an advanced understanding of complex physics, models and their outputs, as well as access to significant computational power.  

Complicated workflows are involved, with the process stretching from observations through numerical simulation, analysis and interpretation, and translation into products and services.

The revolution in artificial intelligence (AI) and machine learning (ML) offers an exciting opportunity to develop and deploy new approaches across the entirety of the Met Office’s work, complementing existing techniques by harnessing the power of data science in conjunction with conventional physics-based modelling and expert knowledge.

As pioneers of weather prediction and climate projection, we continue to demonstrate global skill, technological transformation and realisation of wider impact and benefits (both in the UK and internationally) through our development and application of AI and ML, and the opportunities it affords in agility and innovation, purposeful data and intelligence, and the realisation of growth and value for our customers.

Key areas of Met Office focus for the use of AI and ML capability include:

  • Discovery and attribution – e.g. data mining to analyse, extract and understand information; and test, quantify and discover causal relationships.
  • Fusing simulation with data science – e.g. substituting components of numerical weather prediction and climate models to reduce model run-times and/or improve accuracy.
  • Uncertainty and trust – explainable, interpretable and trustworthy artificial intelligence to understand predictions (and uncertainty) of data-driven methods.
  • Data to decisions – e.g. improving output from numerical weather prediction and climate models, combining this with user-specific metrics, and delivering insights through new interfaces to improve decision-making.

Further details of these advances will be periodically communicated in the Met Office blog and newsletter, as well as in external journal articles and presentations. You can find out more about AI in weather and climate science using the links below: