From avalanche predictions to storm surge warnings, the chances of rain on a national holiday or the arrival of perfect picnic conditions, we would all like to know what the weather is going to do tomorrow.
Luckily we’ve come a long way since basing our predictions on whether the cows are lying down or the color of the sky at night. Now weather prediction, alongside climate analysis, is done by intensely complex computer models running on some of the world’s most powerful supercomputers.
Supercomputers for Predicting Snowflakes
At Météo-France, 460 teraflops of big data computing are now available for radically improved weather and climate modeling. France’s national meteorological service relies on two bullx supercomputers powered by the Intel Xeon processor E5 family to help its dozens of physicists, atmospheric chemists, hydrologists and oceanographers make sense of data gathered from millions of points around the world.
The French government has requirements that the resolution of the weather forecasting model be increased from 2.5 kilometers to 1.3 kilometers in order to pinpoint predicted weather patterns at a more granular level and deliver more accurate, highly localized, forecasts. The adoption of the Bull supercomputers has enabled Météo-France to meet these requirements. Internal tests at Météo-France show that the new supercomputers have already increased useful performance by a factor of 12 when compared to the previously installed configuration. Electricity consumption has also been reduced by 20 percent overall.
The new supercomputers cement the organization’s position at the heart of international weather and climate research.
To continue this conversation on Twitter, please follow us at @IntelITCenter or use #ITCenter.
Jane Williams
Online Sales Development Manager
Intel Corporation
The post Météo-France Improves Forecast Accuracy Using Intel Xeon Processors appeared first on Blogs@Intel.
