The numerical weather forecast model was born.

At that time, the weather map analysis and data reading were all done manually, and the time spent on these tasks was about 10 times more than the calculation time (experiment) on the electronic computer at that time. In order to improve work efficiency and reduce human error, people have put forward some methods of objective analysis and automatic data processing with electronic computers since 1954. Soon, from the report to the analysis and prediction, automation began. In the late 1950s, it was found that the prediction made by quasi-geostrophic model had great limitations, and the system strength of the prediction changed little. In the future, although the equilibrium model is used to improve the effect, it may bring difficulties to the calculation if the physical factors are further considered because of the high order of the equation used. In 1956, A. Eliasen put forward a scheme of using primitive equation model to predict gravity waves. 1959 K. Hinkelmann made a prediction with the original equation model, which succeeded, and its effect was not lower than that of the quasi-geostrophic model. However, his practice is different from that of Sun Guohao, who learned the essence of fast gravity waves and took new measures (such as shortening the time step and filtering gravity waves). ). 1960, the United States successfully launched the Terez Meteorological Satellite and found a new way to provide meteorological data in desert and marine areas. With the development of dynamic meteorology and computing technology, the effect of primitive equation model forecast gradually exceeds that of quasi-geostrophic model forecast, and it has been widely used as operational forecast by the early 1970s (see atmospheric model).