Study On Recognition Algorithms Of Several Convective Disastrous Weather Based On Radar Data

Compared with other detection methods,radar has the advantage of being able to detect and extract information effectively and timely on convective weather that occurs and develops quickly.The data of radar is the main source of reference for Nowcasting.In our country,due to the vast area and the diversity of climate,catastrophic weather,such as hail,tornado,and windy weather,often causes damage to people's lives and loss of people's property.Therefore,how to use data of radar more efficiently,to warn about the upcoming severe weather,and track the severe weather,have become the important research of the severe weather nowcasting.This paper mainly focuses on three aspects of algorithm research:tornado hook echoes,narrow-band echoes of gust fronts,and hail area identification.The main contents are summarized as follows:?1?The hook echo of supercell storm indicates the high possibility of hail,tornado and other disastrous weather event.Image processing technology can be used to identify such a characteristic hook echo and is hereby beneficial to the reliability of the disastrous weather nowcast.According to the hook echo signature in the image of radar reflectivity factor,threshold segmentation is carried out to extract the strong echo regions.After binarization,Fourier descriptor and invariant moment are used to calculate the Contour feature of targets.The final recognition is determined according to similarity matching result.To test the ability of the recognition algorithm,25 dataset observed during the event of‘Funing-tornado'by 4Doppler weather radars are investigated.The analytical result shows that the Critical Successful index reaches 0.72 which proves that this method has a good recognition effect on the hook echo of supercell storm.?2?According to the morphological and intensity characteristics of the gust front,the convolution calculation of different grids is used to form the two-dimension fields based on the intensity classification.In order to obtain the more intuitive and clear recognition map of the gust front,mathematical morphology of digital image processing is adopted.Also,41dataset from 2 radar stations are used to evaluate the identification algorithm for gust front.The identification rate evaluated by ICS reaches 0.61 which proves that this method has a good recognition effect on the echo of the gust front.?3?For the strong convective hail weather that occurred in many places in Guangdong,we use the upgraded S-band dual polarization radar data to recognize the hail area.In this paper,hail recognition algorithm based on fuzzy logic,Fuzzy logic Hydrometeor Classification?FHC?,and index coefficient H_DR recognition algorithm are established.Three algorithms are compared and analyzed for different hail weather in Guangdong effect.The results show that the hail recognition algorithm based on fuzzy logic performs well,and the FHC and index coefficient HDR recognition algorithms which greatly affected by data quality have different degrees of misjudgment and underreporting.