Research and Test on Distinguish Daytime Fog Using Geostationary Satellite in Zhejiang Province

Li Wenjuan; Peng Xiayun; Li Minjie

doi:10.16765/j.cnki.1673-7148.2017.01.013

Abstract：According to the characteristic differences of fog reflection and radiation that visible light,long-wave infrared bands showed,the research was carried out about fog recognition based on the high resolution of FY2E geostationary satellites during the daytime.Statistical analyzing the satellite features on daytime fog,extracting multi-channel dynamic threshold,and combining with the automatic station observations,the threshold superposition can identify fog automatically.The study found that daytime fog reflectivity typically is between 20％ and 50％,the infrared brightness temperature is generally concentrated in the range 270-285 K,and through a combination of dual-channel threshold generally can roughly identify fog by filtering oceans,land surface and thick clouds.In addition,because of the positive correlation between surface temperature and infrared brightness temperature,seasonal setting threshold can be appropriate to improve the recognition accuracy.Furthermore,the albedo and infrared brightness temperature of fog showed pixel uniformity characteristics in the same time data,therefore,dynamically extract threshold at real-time,reduce the threshold range,help to reduce the rate of false.Fine judgment of fog requires a combination of relative humidity data to filter out low humidity area that can make TS score increased by more than 20％.Through 16 days fog sample test showed that the detection rate reached 92％,missing report rate was 8％,but a higher rate of false,reaching 32％,TS score was 62.It is showed that the threshold superposition method can achieve high accuracy for recognizing the fog automatically,but fog recognition which low cloud covered showed high false,because of the fine separation of low clouds,mist and fog remains difficult.