Design And Implementation Of Weather Forecast Service System Based On Distributed Technology

With the progress and development of the society,the weather forecast has more and more influence on people's production and life.The demand for the weather forecast is getting higher and higher.In the past,the acquisition of meteorological data quantity is less and the data accuracy is not high,the weather forecast system using a single database and server.With no guidance forecast data,weather forecasters mainly according to their subjective experience to make a small range forecast,the weather forecast accuracy rate and the production efficiency is low.With the rapid development of meteorological science,radar and satellite business,the theoretical research of weather forecast become more theoretical and standardized,and the data acquisition equipment become more precise and diversified.In order to make weather forecast become more theoretical and normalized,improve the storage and management of meteorological data,and improve the efficiency of forecast production and simplify the production process,developing the meteorological forecast business system.Domestic and foreign meteorological forecast method produced a lot.the most popular method is the grid full coverage forecast.This system applies this method to realize the fixed point,fixed time prediction in the region.The system of weather forecast process is as follows:Firstly,acquire automatic station data and the original message site guidance forecast data through the FTP client;Secondly,change the sation data into two-dimensional lattice data by using the interpolation method,and make grid data displayed on the map through Gis technology;Thirdly,using site data as reference,the forecaster use grid making tool to correct forecast guidance data,making weather forecast grid production.Smplify production process in the production process by means of grid cooperative scheme;Finally,making the lattice data as the foundation,according to the service demand and service production rules,forecast create different services.The system is designed and developed according to the above weather forecast process.According to the functional requirements,the system is divided into four major modules,comprehensive analysis,grid point production,service production,and forecast inspection.The main work of this thesis is as follows:1)Studying the technologies related to the development of meteorological website,including the establishment of distributed system,java thread pool,Modjk middleware,Memcache technology,Mycat technology,weather forecast theory and so on.2)Using the above technologies,building a distributed system framework.Building the distributed cluster needed by website development with Mod_jk and Mycat middleware,and optimize the system performance with Memcache.3)Designing the overall function module of the system according to the requirement of the system,which includes four modules:comprehensive analysis,forecast making,service production and forecast inspection.Design and develop each module according to the specific needs of the business,which includes a comprehensive analysis of the observational data and numerical forecast,forecast making made according to the production process of development,using the space and time coordination algorithm to improve the production efficiency.service production services based on data in grid production,according to the rules of weather phenomena to make service production,forecast inspection according to the area,elements,and inspection rules to classify history forecast data.4)Designing and implementing the system of meteorological data management module,including meteorological data collection(a packet data from FTP request,CIMISS data from HTTP request)and the normalization(conversing file data and intermediate data format to specific meteorological.nc grid file or.mic file),the design and implementation of weather.