How the Met Office uses artificial intelligence to predict the weather

The Met Office is one of the UK's biggest users of supercomputing - all geared towards forecasting weather in the UK and around the world

The Met Office is a world-leading organisation in weather and climate prediction. Its mission is to protect property and life through the operational delivery of products and services which rely on cutting edge science and technology.

In this video, Charles Ewen, CIO of the Met Office, explains how artificial intelligence and supercomputing is changing the face of weather prediction.

"Technology has always been at the heart of weather prediction and was one of the initial applications of ENIAC, the world's first computer," says Ewen. "In 1950, mathematician John von Neumann, in collaboration with an American team of Meteorologists led by the one-time Met Office employee, Lewis Fry-Richardson, produced the first computer-generated weather forecast.

"For many years, weather was one of the few domains utilising computers to make predictions; however, in recent years, computer simulation has become increasingly integral to modern life. The power and accessibility of computers have increased beyond recognition and the benefits of analytical, data-driven approaches have become more apparent in a range of fields.

"Though the power of computer simulation is now recognised by most organisations, the Met Office, and other early pioneers of the 'data revolution' have certainly not been standing still.

"The Met Office have been improving the quality, accuracy and usefulness of computer simulations at an accelerating rate. These improvements have been underpinned by world leading research and applied science as well as significant capacity and capability of the supporting information technology.

"Supercomputing is one of the purest representations of Moore's law, which predicts the doubling of the number of silicon gates that can be fit into a fixed area of silicon every 18 months.

"Whilst it could be argued that this specific measure of progress may be flattening, the capability and capacity of supercomputers continue to follow this power curve. Indeed, the progress of the Met Office's computational capacity and volume of data holdings matches Moore's law almost perfectly," he said.

Watch the video to see how the Met Office tackles the challenges in harvesting and analysing huge volumes of information quickly, and how it is building an SOA - API enabled architecture which can bring problems to data, as readily as the other way round.