ADP, EDP and Computers (continued)

However in 1922 Richardson's solution would have required a huge team of human computers for the technique to be implemented. His single-handed laborious computation to produce a numerical six-hour forecast for Europe was spectacularly unsuccessful. It was not until a sufficiently sophisticated electronic computer became available in the early 1950s that von Neumann was able to bring Richardson's dream to reality. It took a further 20 years before computer-aided NWP became operationally useful.

For those wishing to find romance in meteorology I recommend a reading of the brief summary of the Richardson story in my 1972 ANZAAS address or better still the stories of Platzman (1967) and Ashford (1985).

In retrospect it was wise for those of us in the Bureau in the Dwyer years to proceed cautiously in planning the acquisition of a computer for NWP. Firstly the network of observations in the southern hemisphere was insufficient to satisfy the requirements of item 1 of the Bjerknes statement of 1904, secondly techniques to satisfy the requirements of Bjerknes item 2 required much refinement and thirdly computer hardware and software were developing so rapidly that it was desirable to make a detailed evaluation of the state of the art in overseas countries before recommending which computer should be procured. Further developments are discussed in Chapter 6.

Bureau of Meteorology