Research On Integration Technology Of Medium-short Range Rolling Weather Forecast Results

The integration weather forecast synthesizes each kind of numerical forecast model's forecasting results into a reasonable result. It well embodied integrated member models' forecast level in different time and place respectively, and then provided the relatively accurate forecast reference for forecasters. In a word, it enhanced the forecast accuracy.This project is a cooperation project together with Meteorological Institute of Tianjin. This thesis mainly discussed the research on integration technology of medium-short range rolling weather forecast results and the process of Integrated Weather System realization. This thesis adopted the forecast method of optimal positive weight integration in time-space domain. Using good global optimization characteristics of real-coded genetic algorithm, I designed integrated forcast strategy of numerical forecast products and deal with integration of serveral different models' forcast products. The thesis did research on this method. Integrated forecast strategy was implemented through evolution of weight coefficients for every model.Meteorological workers in china have done a lot of work in the integrated weather forecast, but not much on the medium-short term weather forecast. Genetic algorithms is a adaptive global optimization search learning from Genetic and Evolutionary of nature Species. It has stronger superiority on some problems which need global optimization and are resolved difficultly.The thesis researched on principles and main methods of the integrated forecasting techniques and Optimization Problems. It introduced the basic principles and the steps of genetic algorithm, and then discussed the key technical issues, such as the fitness structure function of genetic algorithm chromosome encoding, genetic operators and genetic parameters. In addition, the thesis analysised and researched on that genetic algorithm was applied to multi-mode integration of weather forecast.At the same time,an integrated weather forecasting system was designed and the process of system implementation was introduced in detail. This System adopted a modular design method. Data pre-processing module implemented function of data processing. Integrated module generation module implemented function of integrated model generating. Integrated results release module implemented function of releasing integrated results. System is strong on reliability, and high in operating efficiency. The performance of the algorithm was analyzed and validated by real data provided by meteorological institute of Tianjin. Finally, the factors impact on the accuracy of integrated forecast was analyzed.