The Research Of Rainfall Prediction Algorithm Based On Radar Data Cloud Extrapolation

In Meteorology,short-time strong convective weather is not only sudden and destruction,but also usually accompanied by thunderstorms winds,tornadoes,heavy rain and other disastrous weather,it threats people's lives and property safety and causes serious damage to the normal economic activities.Therefore,it is of great significance to study the prediction of strong convective weather,especially heavy rainfall.Cloud extrapolation refers to two reflectivity factor graphs formed by the same cloud at the adjacent time,and we calculate the change between them,and then get the short-term temporary precipitation based on these changes.In this paper,research the cloud extrapolation algorithm based on the reflectivity factor formed by Doppler weather radar.The main innovations are as follows:(1)Analysis the radar data and extract the cloud reflectance data,establish the corresponding data model,and then based on the existing cloud structure and morphology extrapolation algorithm,combining image processing technology,the cloud extrapolation algorithm is improved.(2)The simulation experiment is designed based on the test and evaluation standard of literature for the improved algorithm is tested and evaluated.The evaluation results show that the proposed algorithm improves the accuracy of clouds extrapolation.(3)Based on the improved cloud extrapolation algorithm,a short-term and small-range rainfall forecasting system is designed and implemented for short-term temporary small area rainfall forecast.The system uses Spring + SpringMVC + MyBatis system architecture,based on JAVA language.The foreground displays the precipitation forecast information in many ways,the background implements that the timing dynamic acquisition and analysis of radar data,the extraction of the reflectivity factor,the extrapolation of the reflectivity factor graphs and the data interface.