Design And Implementation Of Meteorologocal Data Management And Forecast System

Meteorological enterprise is public welfare undertakings in our country, and it provides guarantee to national security, disaster prevention and reduction, so meteorological enterprise is paid high attention to the government. With the continuous usage of weather observation tools in meteorological departments, a large number of data are accumulated, covering weather, climate, lightning, agriculture ecology, atmospheric composition and so on. It has important practical significance to analyze and manage the data in order to make weather forecast more automatic and intelligent.Meteorological data management and forecast system (MDMFS) implements effective storage of meteorological data, and accurate weather forecast. The MDMFS includes three major parts:meteorological data management module, weather forecast module and evaluation module. Meanwhile, the system uses Microsoft Visual Studio.NET as development platform and C# as development tool. Meteorological data management module uses data warehouse and ETL technology, and achieves meteorological data storage by the theme, which improves the searching speed of the system and provides the data for weather forecast and analysis. The weather forecast module includes simple forecast models and statistics forecast models. In the statistics forecast model, SVM algorithm is introduced, with the parameters optimized by PSO algorithm, which is applied to conventional weather forcast, weather-related disasters forecast and weather conditions forecast. The evaluation module includes evaluations of the station, the region and the guidance forecast.First, in this thesis, the background and related technology are introduced, Which includes data warehouse, ETL and support vector machines. And then, the architecture of system and function module are given by analysis of function and performance.The thesis focuses on the theme analysis in data warehouse, multi-dimensional table design, algorithm of weather forecast and evaluation table design. Also the implementation process of the data extraction module, cleaning module, conversion module, loading module, weather forecast module and evaluation module are given in detail. Finally, the MDMFS is tested with the point of view of the interface testing, unit testing, integration testing and system testing.