UK supercomputer to weather climate change
Scientists from the Walker Institute at the University of Reading will study the problem of climate change using a UK academic supercomputer called HECToR, which was launched on 14 January.
HECToR (High End Computing Terascale Resources) is a national high-performance computing service for the UK academic community. Initially, HECToR is expected to be about six times more powerful than the previous UK academic supercomputer and 25 times more powerful by 2009.
HECToR has been set up to allow a wide range of research from physics to economics.
At Reading, scientists will be using the new supercomputer to run some of the most detailed global climate models in the world. Models have already been developed on the Earth Simulator computer in Japan for the last three years by the Reading-based researchers and it will now be possible to run these in the UK.
Len Shaffrey, a senior scientist developing the new high resolution models, said: ‘Current climate models struggle to give us details of how climate might change at regional and local levels and in particular how high impact weather, like storms, will change. We’ll be using the power of HECToR to run much finer resolution global climate models than has been possible before in the UK. Early results are showing that our high resolution global models can simulate regional and local climate more realistically.’
Shaffrey added: ‘Details of European weather, like storms and summer heat waves, are much improved in the new high resolution models. The number and intensity of hurricanes are also more realistic.
Another major project at Reading to make use of HECToR will investigate tropical weather systems.
Dr Steve Woolnough, who is leading the tropical weather project, said: ‘With the increased computing power of HECToR we will be able to tackle some of the most fundamental questions about our weather and climate. Using very high resolution models we can look at the way complex tropical weather systems develop and interact. The tropics act as the powerhouse for all our weather, even affecting us over Europe. Improving our ability to model the tropics will give us better forecasts of weather and climate from days to decades ahead.’
Current weather forecasting and climate models have serious problems in modelling the cloud systems that make up tropical weather, largely because they range from the scale of a few clouds to large scale circulations like the Indian monsoon. All these scales can interact in ways that are not captured by current weather forecasting and climate models. These problems limit, for example, the ability to forecast European weather beyond about 10 days and the ability to forecast accurately the onset and break down of El Niño events in the tropical Pacific.