WindSat Radio-Frequency Interference Signature and Its Identification Over Greenland and Antarctic

A detection of radio-frequency interference (RFI) in the space-borne microwave radiometer data is difficult under snow and sea ice-covered conditions. The existing methods such as a spectral difference technique or a principal component analysis (PCA) of RFI indices produce many false RFI signals near the boundary of Greenland and Antarctic ice sheets. In this paper, a double PCA (DPCA) method is developed for RFI detection over Greenland and Antarctic regions. It is shown that the new DPCA method is effective in detecting RFI signals in the C- and X-band radiometer channels of WindSat while removing the false RFI signals over Greenland and Antarctic. It also worked well in other snow-free or snow-rich regions such as winter data over the United States. The proposed DPCA can be applied to satellite radiometer data orbit-by-orbit or granule-by-granule and is thus applicable in an operational environment for fast processing and data dissemination.