June 18 -- Severe winter weather experienced in the UK over the last couple of years reduced the UK’s GDP by 0.5%, and resultant travel disruption cost the UK economy £280 million per day (link to report).  The weather has a huge impact on our lives, affecting transport, agriculture, energy use and leisure.

The Science and Technology Facilities Council (STFC), the Met Office and the Natural Environment Research Council (NERC), have together embarked on a project to design and build a next-generation weather forecasting model for the UK that will exploit advanced ultra-fast supercomputers and provide a boost to the effectiveness of forecasts that can not only save money, but also save lives.

By the end of the decade, scientists will be using supercomputers that are thousands of times faster than any of today’s systems.  Known as ‘exascale’ supercomputers they will contain millions of processors capable of performing a million trillion calculations per second.  Harnessing the power of these computers for weather and climate prediction could mean much more accurate forecasts that will help us to live more easily with episodes of severe weather and also to adapt to climate change, maintaining UK leadership in environmental prediction.

This research will be one of the first major projects to benefit from STFC’s new future software research facility at its Daresbury Laboratory in Cheshire. This centre was announced in March this year following £37.5m investment by the Department of Business Innovation and Skills (BIS) into High Performance Computing (HPC) at Daresbury as part of its UK e-infrastructure initiative.  It forms one of the world’s foremost centres in software development and is host the UK’s most powerful supercomputer, Blue Joule.

Minister for Universities and Science David Willetts said: "Supercomputers are fundamental to modern research, in particular very complex areas like weather forecasting. This project will harness the expertise of the UK's excellent research base to ensure we remain world-leading in climate science."

Andy Brown, Head of Foundation Science at the Met Office, said “The Met Office is at the forefront of scientific developments in weather forecasting and its forecasts are ranked in the top two national met services in the world.  This project between the Met Office, STFC and NERC will ensure that the UK continues to benefit from the best science and advice available.”

Associate Director of STFC’s Computational Science and Engineering Department, Dr Mike Ashworth said: “Ever more accurate prediction capabilities will help the UK to be more aware, and consequently more prepared, for severe weather impacts in the future. There are many challenges to overcome, the main issue being that the models used to simulate the atmosphere today would be unable to take advantage of the processing power of the ultra-fast computers available within the next few years. We are working together to design and develop a next-generation computer program that will do the key job of simulating the winds, temperature and pressure. This, when combined with other processes such as cloud formation, will allow us to simulate the changing weather conditions.”

It is anticipated that the new code will, in time, replace the dynamical core of the Met Office’s Unified Model (UM), the principal UK tool for weather and climate prediction, also used by national weather services around the world including Australia, South Korea, Norway, India, New Zealand and South Africa.

Professor Stephen Mobbs, Director of NERC’s National Centre for Atmospheric Science, said: “Tomorrow's "exascale" computers represent a huge opportunity and a huge challenge for the science of weather forecasting. The opportunity to produce forecast detail down to the scales which affect specific human activities are beckoning. For instance, details on the scales of transport infrastructure –roads, rail, etc – or individual towns will be resolvable. At the same time, the computer software challenges of effectively using millions of processors open up new areas of computer science. There are also a vast range of physical processes which affect the weather on these fine scales, stretching our understanding of the atmosphere itself and our ability to represent it within models.”

-----

Source: Science and Technology Facilities Council