| Satellite-based and ground-based lightning detection technology provides important methods for monitoring and studying the characteristics of lightning activity.Lightning detection based on geostationary satellite orbit,because of its unique platform advantage,is a new technology that countries are competing to develop.LMI(Lightning Mapping Imager),boarded on Feng Yun4 A meteorological satellite,is the geostationary satellite orbit lightning optical detection payload,which is first in China,and the first batch in the world.There are some urgent issues of LMI,including data acquisition and service、quality control、data comparison and verification、fusion application of multi-source lightning data.According to the logical relationship of information flow in LMI ground application system,the effect of space single particle bombardment on LMI output data is analyzed,the methods of ground autonomous control repair and multil-evel data calibration are proposed,the data service platform is designed in this paper.According to the feature of LMI multilevel data,the quality control methods based on artificial intelligence technology and threshold recognition technology are studied,machine learning models and deep learning models are built,which improved the quality of LMI data;based on satellite-based and ground-based lightning data,characteristics of lightning activity in severe convective weather are studied from the perspectives of individual case analysis and statistical feature analysis,and the different characteristics between LMI data and traditional ground-based lightning data are also studied,a collaborative analysis method of satellite-based and ground-based lightning data is proposed.The specific research contents and conclusions are as follows:⑴ In order to solve the problem of LMI output data error caused by single particle bombardment in the universe,an autonomous control repair method is proposed,the accuracy of instructions has reached 100% after testing,MTTR(Mean Time To Repair)is less than 1 minute,which minimize the impact of single particle bombardment on LMI detection data.The method of multi-level data calibration is studied and a data service platform is designed,the statistical characteristic relationship between lightning activities and deep convective precipitation in customized space-time interval is studied,realizing the reliable acquisition of LMI data,providing standard,accurate,efficient data service.⑵ The organization structure of LMI Level-0(L0),Level-2(L2)data is analyzed,quality control methods are studied for different levels of LMI data.Training data sets and validation data sets(L0)are generated,machine learning models based on CART(Classification and Regression Tree)、NB(Naive Bayes)and SVM(Support Vector Machine)and deep learning models of convolutional neural network are built,a data classification and recognition method based on artificial intelligence is proposed,experiments show that L0 lightning data can be effectively identified by the method.Based on the coupling relationship between lightning activities and severe convective weather,the selection method of recognition threshold is studied,a quality control method for L2 group data based on multi-source meteorological data is proposed,the effectiveness of the method is proved by the comparison based on the ground lightning data,which improved the quality of LMI data.⑶ The relationship between satellite-based and ground-based lightning data is analyzed,a comprehensive analysis method combining the two source data is proposed.Case study shows that LMI can detect the cloud flash more quickly than the groundbased lightning location system,which is helpful for early warning of severe convection process;LMI usually detects 5 to 10 times as much lightning as ground-based lightning system;LMI has better detection and positioning accuracy during the daytime;the lightning data detected by LMI and ground-based ADTD system shows good consistency in different scales such as interannual,season and time,but the matching ratio varies greatly between the east and west of China,and it is not related to the lightning signal strength,which reflects the different advantages of satellite-based and ground-based lightning detection systems.⑷ The different characteristics of LMI data and traditional ground-based lightning detection data are researched,and the fusion application of multi-source data in severe convection monitoring is studied,the intracloud-cloud-to-ground(CG)lightning ratio(termed Z)in China and the characteristics of lightning activity in China’s coastal waters and typhoons are also analyzed based on the new data contributed by LMI.The results show that the average value of Z in China is 2.82,the average standard deviation is 0.31,and the value of Z is not highly correlated with the latitude and the proportion of positive and negative ground lightning;under the influence of different topographical and climatic characteristics,the lightning activities in the four main offshore seas vary greatly in season,time,day/night,and peak current intensity;lightning mainly occurred in the stage of super strong typhoon;the distribution of lightning activities on both sides of the typhoon path is uneven;there are also dense lightning activities before the typhoon made landfall twice due to the influence of topographical uplift;the intense lightning activities,mainly in the low cloud top bright temperature region(TBB 210K);the frequency/density of LMI lightning data is closely related to the distance from the eye of the typhoon center,there are basically a strong-weak-strong three-cycle periodic distribution,which is related to the changes of precipitation structure and typhoon intensity.In particular,the degree of correlation between lightning activities and typhoon intensity in the eye-wall region is higher than that in the outer rain zone.It indicates that LMI can detect lightning with different intensities in typhoons in a more comprehensive way,which is helpful to deepen the understanding of lightning activity characteristics in typhoons.The work of this paper has improved the reliability of LMI data acquisition,enhanced the data service ability,promoted the research on data quality control,the comprehensive application of satellite-based and ground-based lightning observation data has realized.It also has good theoretical guidance and important practical significance for the exploration of satellite-based lightning detection technology and the optimization of LMI. |
PDF Full Text Request