Researchers at the European Centre for Medium-Range Weather Forecasts say that running simulations on inexpensive computer chips that create results tainted with random noise could improve climate models. Researchers trying to make climate models more realistic are limited by the processing power of the supercomputers running the models, says researcher Tim Palmer. However, adding a degree of randomness to a particular model and running it several times could be an inexpensive way to improve the realism of climate modeling, Palmer says. If multiple, slightly different simulations of the same model create the same result, it indicates the strength of that prediction. This technique has already been successfully demonstrated in the prediction of weather during a period of several weeks. Intentionally generating randomness for these models would consume a significant amount of computing power, but using cheap computer chips that naturally create random noise due to how electrons move through them could provide an inexpensive and processor-efficient solution. “It’s very speculative,” Palmer says. “But if it can be made to work, it would make much more efficient use of power.”
For more information please visit: http://www.cpccci.com
