Acceptance Speech – Rome, 19.11.1996

Norway

Arnt Eliassen

1996 Balzan Prize for Meteorology

For his fundamental contributions to dynamic meteorology that have influenced and stimulated progress in this science during the past fifty years.

Mr. President,
Members of the Balzan Foundation,
Ladies and gentlemen,

To be recipient of the Balzan Prize is a great honour and a wonderful experience of recognition for an old scientist. The Prize has for the first time been given for research in the field of meteorology; this is a sign that meteorology has now reached a stage of maturity, being firmly based on the laws of physics. This progress is the result of cooperation of scientists in many countries around the world. They should all share the honour with me.

At the beginning of the century the Norwegian physicist Vilhelm Bjerknes advocated that weather prediction should be based on numerical integration of the differential equations of hydrodynamics and thermodynamics. But that was more easily said than done. These non-linear equations did not easily reveal their secrets in the form of information of practical value to the world’s weather services. A heroic attempt to calculate tomorrow’s weather from the theory was published in 1922 by the great British physicist Lewis Richardson. The result was discouraging, and showed in particular that the amount of computation necessary was so enormous, so that, with the primitive calculators of the time, the computation of a weather change would take much longer time than the weather change itself. It seemed that Vilhelm Bjerknes’ dream of weather prediction by numerical integration of the governing differential equations had to be dropped. The traditional method of graphical extrapolation of weather systems from one weather map to the next remained the principal basis of weather forecasting.
In the 1930’s Vilhelm Bjerknes was still active at the University of Oslo, studying atmospheric wave motions and vortices in co-operation with his co-workers Halvor Solberg and Einar Høiland. I happened to attend some of their seminars in 1938, and as a result, I decided to choose meteorology as my graduate subject. I was extremely fortunate to end up in this stimulating milieu.

With the advent of electronic computers after World War II, the situation for meteorology changed radically. Computations which before had been utterly impossible now seemed to be within reach. At the Institute for Advanced Study in Princeton, the Hungarian-American mathematician John von Neumann planned an attack on the numerical weather prediction problem by means of an electronic computer which was being built there. A small group of theoretical meteorologists was set up, headed by the renowned American meteorologist Jule Charney. He visited Oslo in 1947, and I was invited to join the Meteorology Group in Princeton. I had a wonderful year there working together with Jule.

The philosophy in Princeton during the first years was to start out with the simplest possible model, namely the “flat” atmosphere designed by the Swedish meteorologist Carl-Gustaf Rossby. The first successful 24 hour prediction was made with this model by Charney, Fjørtoft and von Neumann, and published in 1950. The accuracy of the result was not too impressive, but the work still represented an important break-through and gave a strong impetus to meteorological research throughout the World. A long-lasting development process began, where advances in computer technology, in meteorological theory and in numerical integration techniques went hand in hand. The research was not limited to the problem of prediction, however. A long list of special motion phenomena were studied and analysed both theoretically and empirically. One example is various kinds of wave motion, their excitation mechanisms and their ability to transfer energy and momentum through the atmosphere. Another problem concerns atmospheric fronts, how they are formed, how they produce clouds and precipitation, and their connection with extra-tropical cyclones.
By means of global numerical models of the atmosphere and modem supercomputers, we now get weather forecasts for about one week ahead. We all know that these forecasts are not perfect, but they are very much better than the one day forecasts of fifty years ago before the era of numerical weather prediction. A legitimate question then is this: how far can the forecast period be stretched as a result of future development? To this question there is a definite answer. The American meteorologist Edward Lorenz has shown that the atmosphere, although it is a deterministic system, is also chaotic, in the sense that it is oversensitive to errors in the initial state. Inevitable imperfections in the observed initial state of the atmosphere will magnify during the integration process, and after a time of no more than about two weeks render the prediction worthless. The possible extention of the forecast period is thus quite limited.

On the other hand, I believe the prospects are good for important progress in our understanding of the factors that determine the global climate. This is particularly important for our possibility to control harmful climate changes of anthropogenic origin.
Atmospheric scientists from all over the world have contributed to the progress in meteorology during the last decades. In the 1950’s the number was very modest. We all knew each other; it was like belonging to a large family. Now, new and more numerous generations of scientists have taken over, and I wish them success in their important research. I consider myself highly privileged to have had the opportunity to participate in the exciting development of meteorology, and I feel deep gratitude to find that my work has been appreciated.