AMSR-E观测资料干扰对反演地表参数的影响

针对星载微波成像仪低频窗区通道观测数据中存在大范围无线电频率干扰(简称RFI)的情况,以微波扫描辐射计(AMSR-E)为例,首先用改进的主成分分析方法对RFI进行识别;进而探讨RFI存在对反演地表参数的影响.然后,利用经过线性拟合RFI订正后的AMSR-E观测资料,采用一维变分1D-Var方法进行地表参数反演,通过对美国地区陆地RFI订正前、后地表反演产品(地表温度及降水率)的比较,发现RFI干扰使得受影响区域反演的地表温度及降水率异常偏高,存在较大误差.因此,在使用星载微波成像仪低频窗区通道观测进行地表参数反演和资料同化前必须进行有效地RFI识别和订正,改进的主成分分析识别方法和线性拟合的RFI订正算法对陆地上观测是有效的.

Effect of AMSR-E data interference on the retrieval of land surface parameters

There are spread radio-frequency interference (RFI) problems of the low-frequency observations of the spaceborne microwave imagers, e.g. AMSR-E (Advanced Microwave Scanning Radiometer). A modified principal component analysis method was used to identify RFI and then its influence on the retrieval of surface parameters was studied. Next, the AMSR-E data corrected by linear fitting correction of RFI were used to retrieve the surface parameters through 1D-Var method. By comparing the retrieved products (e.g., land surface temperature and surface rainfall rate) before and after RFI correction over the United States, it was found that retrieved land surface temperature and surface rainfall rate which were interfered by RFI were abnormally high and with large deviations. Therefore, it was necessary to effectively identify and correct RFI prior to low-frequency observations with spaceborne microwave imagers to retrieve land surface parameters or assimilate these observations. And the modified principal component analyzing RFI identification method and linear fitting RFI correction algorithm were effective for the observations over land.