GMS实时资料识别冰雹研究

Study of Distinguishing Hail with GMS Real-time Data

利用１９９７～１９９８年３～４月ＧＭＳ－５的红外亮温和水汽亮温资料，根据云南春季冰雹和非冰雹的红外、水汽亮温值二维概率分布，建立数学判别式。结果表明，冰雹和非冰雹的识别率为９４％，冰雹判别准确率为３０．８％。

By using GMS-5 infrared and water vapour brightness temperature data from Mar. toApr. in 1997 - 1998, the methods of extracting hail weather parameters were developed anddigital remote sensing pictures were obtained in spring in Yunnan. Based on the bivariate probability distribution of infrared and water vapour brightness temperature of hail and no- hail, a criterion distinguishing hail from no-hail was set up. The results show that a precisiondistinguishing hail and no-hail is 94%,the accuracy of hail distinguishing is 30. 8%.