How to predict the temperature

Numerical weather forecast is the main method of modern weather forecast.

Since 1980s, with the wide application of remote sensing, computing technology and meteorological satellite data, the world weather forecast has made a new leap, and the traditional weather map has been replaced by numerical weather forecast. The so-called numerical weather forecast is to describe the laws of atmospheric motion with seven differential agendas of fluid mechanics and thermodynamics. The seven agendas contain seven unknowns-maximum temperature, minimum temperature, precipitation, humidity, air pressure, wind direction and wind speed. By solving the equations with a large-scale high-speed computer, we can get the space-time analysis of seven unknowns in the future, that is, the future weather distribution. More than 30 countries in the world carry out numerical forecasting business, and developed countries have established global and limited regional forecasting models, with the highest running speed of computers reaching 4 billion times per second. The numerical prediction model of total stress has more than 30 vertical layers, and the horizontal resolution is 60 kilometers. The prediction can be effective for 10 days in the middle and high latitudes and for 5 days in the low latitudes. The horizontal resolution of the limited area prediction model is 15km, which can reach 2km in the United States and Germany.