Kyle Hilburn gave a talk at MICRORAD, the 13th Specialist Meeting on Microwave Radiometery and Remote Sensing of the Environment in Pasadena, CA, March 2014. His talk presented a method for Radio Frequency Interference (RFI) detection in AMSR-E data. RFI is a rapidly increasing source of error in microwave radiometer sea surface temperatures and is highly variable in time. The primary source of RFI over the ocean is television broadcasts from geostationary satellites and RFI is therefore located mostly in coastal areas of Europe, African, North American and South America. There can also be direct emission from surface-based sources that produce strong RFI over a localized area, such as the Azores.
At RSS, we identify RFI by calculating the difference in AMSR-E SST obtained from two RSS algorithms, one algorithm using all microwave channels (6.9, 10.7, 18.7, 23.8, and 26.5 GHz) and the other using all channels except the 10.7 GHz (we call this the no-10 SST). We use a threshold of 0.5K to distinguish RFI from random small differences between the two SSTs. RFI always has an additive effect on brightness temperature, but the effect on SSTs can be positive or negative.
The no-10 algorithm can be used to not only detect RFI, but also to mitigate the effects of RFI. The figure to the left shows the exceedance frequency, or the number of times the 0.5K SST difference threshold was exceeded in the 2003 to 2011 time period, divided by the total number of valid observations for that pixel. The red region depicts the RFI that exists for more than half the months during AMSR-E operation. A time series of fractional area of RFI shows an increase of 19% per decade.
More details and other examples are in the MICRORAD Meeting Extended Abstract: Hilburn et al, 2014, "RFI Detection and Mitigation for AMSR-E Ocean Retrievals",