Climate forecasters need to start saying what climate changes are most likely to occur, instead of just warning of all possible changes, says Myles R. Allen of Oxford University in this article.
Planners and engineers need practical information in order to prepare for the most likely events. At present, it is difficult for forecasters to say which scenarios can be ruled out as highly unlikely, since so many studies are based on the same limited set of observations.
Allen argues that realistic forecasts need to be based on many simulations working with numerous variable factors. The challenge is far beyond the scale of conventional supercomputers, so the users of personal computers are being invited to help through a new initiative to generate the world’s most comprehensive forecast of 21st century climate (see Scientists tap global computer power to probe climate).
Reference: Nature 425, 242 (2003)