Research And Analysis Of Precipitation Based On Radar Echo Data

The analysis and research of precipitation in meteorology has always been a very popular field.In the research and analysis of precipitation,the traditional analysis methods mainly have the problems that the model is too complicated,the calculation is too large,and the existing data cannot be fully tapped and utilized.With the explosive growth of meteorological data in recent years,these problems have become more serious.This article studies and analyzes the precipitation based on the actual business requirements of the Liaoning Meteorological Disaster Early Warning Center,which is different from the traditional method.First,it starts with the data and pre-processes the collected radar echo data accordingly.Thus,the multi-layer radar echo reflectance data is obtained,and then the problem is researched and analyzed based on the obtained multi-layer radar echo reflectance data using a machine learning method.In terms of data processing,this paper makes full use of existing radar echo data and information mining.According to the working mode of Doppler meteorological radar,the radar echo data collected by it is decomposed and the clutter is removed to obtain the corresponding multilayer radar echo reflectance data.Through the data collection time and latitude and longitude coordinates and other information,the historical precipitation data and the decomposed radar echo reflectance data are integrated,and then the precipitation level is divided according to the amount of precipitation,and finally the multilayer radar echo is obtained Reflectivity and precipitation level dataset.This paper uses LightGBM integrated learning algorithm model to complete the research and analysis of precipitation.Firstly,the multi-layer radar echo and precipitation level data sets are standardized,and the multi-layer radar echo reflectivity value is used as the input of the algorithm model,and the precipitation intensity level is used as the output to train the LightGBM model.In this process,by comparing the experimental results obtained with multiple sets of different model parameter values,the appropriate parameters are finally selected for the model.In order to facilitate the storage,management,search and statistics of massive radar echoes and precipitation data.After completing the above research,according to the actualoperation and use requirements,a precipitation analysis system based on radar echo data was designed and completed.This system is developed based on the B / S architecture,and users can directly complete the required operations through the graphical interface,thereby avoiding cumbersome and complicated operations.