Brightness Temperature Restoration From Radio Frequency Interference And Its Impact On Practical Applications

Space-borne microwave imager data is the main source of surface observation,however,the low-frequency channel of the microwave imager has been interfered by a wide range of radio signals due to the lack of attention on channel protection,which seriously affects the assimilation of the imager data and the retrieval products from it.Based on the data in 2016,we use the normalized PCA method to identify the RFI-contaminated data of AMSR-2(Advanced Microwave Scanning Radiometer 2),and we attempt to establish a PCA-based restoration method to recover data with RFI.Both the ideal and real reconstruction experiments show that the new method can effectively repair abnormal observations interfered by RFI.Finally,the repaired microwave imager data were tested for practical application to evaluate its practical value by using CRTM simulation and retrieval of the surface soil moisture.The results show that the NPCA identification method can effectively identify the RFI interference of 6.925 GHz channel on land of America.It can be seen from the results that most large cities in the United States have obvious radio interference signals.In this paper,PCA iterative method and linear fitting method as well as the Cressman interpolation are respectively used to restore the abnormal signals affected by RFI.It is found that PCA iterative method can basically eliminate the influence of RFI,and the interpolation effect of iterative method is better than that of linear fitting method and the Cressman interpolation.This is because the PCA iterative method not only takes into account the typical spatial distribution of physical quantities,but also takes advantage of the high correlation of light temperature observed by each channel,so as to consider the effect of interpolation data on the whole.This method is universal to the terrestrial spaceborne microwave imager.The difference between the observed and restored brightness temperature and the simulated one by CRTM was further explored,and the surface soil moisture was retrieved with the observed and restored data based on LPRM algorithm.It can be seen that the new restoration method can improve the utilization rate of data thus it has a good application prospect for the research of climatic retrival and assimilation of microwave imager data.