Computer Requirements for Atmospheric Modelling

Abstract

It is a great pleasure for me to open this Workshop on ‘The usage of Multiprocessors to integrate atmospheric models’. As most of you know, the meteorological forecasting problem was identified already by von Neumann as an ideal application for computers. Under his guidance and supervision a special group was set up at the Institute for Advanced Studies in Princeton in 1984 under J. Charney to undertake an integration of the vorticity equation using real atmospheric observations. The result was successful and was one of the very first attempts to use electronic computers to solve non-linear equations by numerical methods. Since then, the meteorologists have stayed in the forefront in applying the fastest available computers to solve the forecasting problem using increasingly more realistic models of the atmosphere. The meteorological community is an active user and a very substantial user for supercomputers and discussions between the meteorological modellers and the experts within the computing industry are important and I am sure mutually beneficial.