Retrieval of Precipitation by Using Himawari-8 Infrared Images

SUN Shaohui1,2, LI Wanbiao1,†, HUANG Yipeng1

1. Department of Atmospheric and Oceanic Sciences, School of Physics, Peking University, Beijing 100871; 2. 93313 Troop, Changchun 130111; † Corresponding author, E-mail: lwb@pku.edu.cn

Abstract Using the matched data between infrared images of the geostationary meteorological satellite Himawari-8 and the product 2AGPROFGMI of Global Precipitation Measurement (GPM), the matched infrared brightness temperature (BT), brightness temperature difference (BTD) and surface precipitation are connected by pixel to pixel. Further more, because of the advantage of the more channels of geostationary meteorological satellite, two-dimensional and three-dimensional lookup tables of rain rate (RR) are established with the matched infrared brightness temperature, brightness temperature difference and surface precipitation. The lookup tables which can identify the rain rate of different grades are found by the retrieval tests. The three-dimensional lookup tables (BT10.4, BTD12.4-10.4, BTD6.2-7.3) show higher probability of detection (POD=0.8817) and lower false alarm rate (FAR=0.4042) when detecting the rain, so it can identify the areas of rainfall. Key words satellite remote sensing; brightness temperature; brightness temperature difference; rain rate; retrieval

降雨的准确测量对工农业生产、水资源利用以及洪涝和干旱等自然灾害的预报有重要意义[1]。由于地面雨量计和地基测雨雷达观测网的密度不足,很难监测和跟踪暴雨及其他灾害性天气系统。利用卫星遥感数据反演降雨, 覆盖区域广, 时空分辨率高, 可以弥补地面常规观测的不足。卫星遥感可分为主动微波遥感、被动微波遥感和可见光/红外遥感。主动、被动微波遥感可以分别探测降雨的垂直结构和云内部的结构信息, 但目

前微波遥感器搭载在极轨气象卫星上, 时间分辨率不高。静止气象卫星可见光/红外遥感可以提供高时间分辨率的降水产品, 对降雨起到很好的监测作

[2–3]用 。因此,加强利用静止气象卫星多光谱遥感测量信息反演降雨的研究很有必要。

Ba等[4]采用温度梯度、坡度参数[5]和云顶粒子有效半径[6], 利用地球静止业务环境卫星(Geostationary Operational Environmental Satellite, GOES) 0.65, 3.9, 6.7, 11和12 μm 5个通道建立GOES多光