基于南太平洋的AMSR2 L1R亮温数据质量评估

本文基于环境场较为稳定的南太平洋目标海区,以海洋大气微波辐射传输模型(Radiative Transfer Model,RTM)模拟亮温作为参考值,对2015年1月1日—2017年12月31日的高级扫描微波辐射计(Advanced Microwave Scanning Radiometer 2,AMSR2)L1R亮温数据产品进行了质量评估。结果表明AMSR2 L1R所有通道亮温数据总偏差和标准偏差的变化范围分别为1.466~6.352 K、0.270~1.693 K,其中标准偏差在水平极化通道较大的同时随着频率的增大而增大。相比同类遥感器如全球降水测量微波成像仪(Global Precipitation Measurement Microwave Imager,GMI)等的质量分析结果,AMSR2亮温数据的标准偏差较小,这表明AMSR2亮温数据精度较高。对AMSR2 L1R亮温数据3年长时间序列的变化趋势分析表明所有通道亮温偏差均在±0.5 K范围内波动但是存在微弱的季节性变化,标准偏差随时间的变化较小,这表明AMSR2 L1R亮温数据质量较为稳定。

Quality assessment of AMSR2 L1R brightness temperature data product base on South Pacific

The Advanced Microwave Scanning Radiometer 2 (AMSR2), which is equipped with the Global Change Observation Mission-Water 1 (GCOM-W1), and retrieves ocean atmospheric environmental parameters by observing the radiation brightness temperature of earth surface. In this paper, we set the simulation of brightness temperature of ocean-atmosphere radiative transfer model (RTM) as the reference value, and the south Pacific as a research area, then evaluate the quality of the L1R brightness temperature data which was produced from January 1, 2015 to December 31, 2017 via AMSR2. Compared toRTM brightness temperature data, the variation range of total deviation and standard deviation for the L1R data of all channels of AMSR2 are 1.466~6.352 K and 0.270~1.693 K respectively. Furthermore, compared to the vertical polarization, the standard deviation of L1R data is larger in the horizontal polarization channel, and will increases with the increasing of frequency. Compared to the quality analysis results of similar sensors, the standard deviation of the AMSR2 bright temperature data is smaller, which indicates that the AMSR2 bright temperature data has a higher accuracy. To carefully evaluate the stability of the AMSR2 L1R brightness temperature data, we also analyze the variation trend of deviation and standard deviation over time. The analysis results show that the brightness temperature deviations of all channels of AMSR2 L1R are within 0.5 K, but have slight changes with seasons, and the standard deviation changes slightly with time. In conclusion, theseresults indicate that the quality of AMSR2 L1R brightness temperature data is relatively stable.