In 1922, first the mathematician Lewis Frei Richardson had the idea to describe physical processes that take place in the atmosphere by using the hydrodynamics of fluid equations. In his work “Weather prediction with numerical methods” described the mathematical methods which one could use in order to predict the weather evolution with the time. His idea required a great number of people who would solve the mathematical equations that describe the atmosphere motions and so they would predict the weather evolution. In particular, he estimated that 64000 people were needed in order to make a prediction with the same speed as the weather develops in real time. He also claimed that in distant future people would compute faster than the weather evolution. These words were prophetic although “the distant future” was only 40 years later
In the followings years until 1960 scientific world paid no regard to prediction and did not trust computers. Weather forecast was an empirical work and practical meteorologists predicted the weather for the next day by using mostly their own experience. That season Lorenz one of the biggest pioneers in numerical forecast worked at M.I.T (Technological Institute of Massachusetts) and he had a totally different opinion:
"The equations that describe the procedures which take place in the atmosphere, are already known from a long time. If numerical theory draws precise roots for spaceships and rockets and predicts planets trajectories why it cannot be used in weather forecast?"
Lorenz run into a lot of difficulties especially he had to face off the scientific establishment which considered his opinions extremes. But after continual tries he managed with his primitive computer to create a simple model of the atmosphere that constituted the ancestor of current atmospheric models.
In the next years the scientific world had a lot of hopes for the capability of weather prediction. There were several tries until the numerical weather forecast had been an undeniable reality. At this period of time the bigger successes for the numerical weather prediction have happened in Reading, a small town outside London where sheltered and continue to shelter E.C.M.W.F ( European Center for Medium Range Weather Forecasts). The Europeans showed their confidence to the numerical weather prediction by using the supercomputer CRAY for the integrations. They also engaged the best scientific personnel of the ECMWF states of members. In the other end of ocean efforts overwhelmed also the Americans. Their supercomputer one Cyber 205 by Control Data executed millions of calculations in one second and solved in short time model’s systems with 500.000 equations. The results were very satisfactory and the messages were continuously more optimistic.
From then up to today they have passed 20 years of continuous efforts and has been achieved enormous progress. In this it has contributes the development of supercomputers which today allocates a big number of processors of enormous possibilities. The velocity of one only processor measured in Megaflop (One Megaflop is equal with one million calculations in one second). The atmosphere simulated continuously with more realistic way after the numerical methods are improved permanently. The simulation of the atmosphere is continuously more realistic as the numerical methods are improved, the grid becomes denser, and the under study atmospheric layers are increased. Also detail information about orography ,vegetation and type of the ground are introduced in the models. Thus the results that are received are more reliable and more analytic.