The Radio-Frequency Interference Identification Methods Based On Time Series Observations

With the development of microwave remote sensing technology,the measurements of spaceborne microwave radiometer have been widely used in meteorological research.However,Data from these sensors are likely to be contamined by Radio Frequency Interference(RFI),which reduce the utilization of microwave measurements(Especially for the low frequency observations).Radio frequency interference means that radiance received by the microwave radiometer not only includes natural emission from land surface but also include emission emitted by part of active microwave sensor of ground and emission reflected by ground,which result the surface can not be reflected by received information from spaceborne passive microwave sensors.In order to make a better reliable RFI sources distribution,AMSR-2(Advanced Microwave Scanning Radiometer-2)observations at 6.9 GHz and 7.3 GHz from January 1,2016 to December 31,2017 over global land are carefully analyzed by two methods(the means and standard deviation method,standard error of estimate)based on time series.Then,we use the mature spectral different method to verify results of these two methods.At the same time,the distribution and characteristics of long-term RFI signals were also analyzed and studied.It was found that these two methods based on long time series are effective for determining RFI over global land at 6.9 GHz and 7.3 GHz.What is more,the SE method is a more reliable method since it produces less false RFI pixels than other two methods.We also found that RFI sources are not always constant through time.The distribution of radio frequency interference(RFI)is also related to the polarization characteristic of channels.At 6.9 GHz,the total number of FOVs where determined RFI signals at horizontal polarization channel is greater than that at vertical polarization channel.However,at 7.3 GHz,the total number of land pixels where determined RFI signals at horizontal polarization channel is less than that at vertical polarization channel.The distribution of RFI signals at different passes is also different.Regardless of 6.9 GHz or 7.3 GHz,ascending-pass measurements are more contaminated than descending-pass measurements.In order to mitigate the RFI in microwave measurements at 6.9 GHz channel,AMSR-2 are added to two new C-band channels centered at 7.3 GHz based on AMSR-E.What is more,AMSR-2 also has a larger diameter antenna reflector(2.0 m compared to 1.6 m for AMSR-E)and narrower beam width.In order to make a evaluation about whether AMSR-2's hardware improvements and two new channels mitigate the RFI in microwave measurements,AMSR-2 and AMSR-E observations at 6.9 GHz and 7.3 GHz from August 2012 to July 2013?August 2016 to July 2017 and October 2010 to September 2011 over global land are also analyzed by the means and standard deviation method.Then we will make a detailed statistics and analysis for the total number of RFI signals in different instruments and time periods.Research shows,The number of RFI-contaminated pixels in AMSR-2 7.3 GHz observations is much less than that in AMSR-E 6.9 GHz observations,especially in the United States and Japan,which illustrates two new channels indeed mitigate the RFI in microwave measurements at 6.9 GHz.At the same time,AMSR-2's hardware improvements increases the main beam efficiency and improves the spatial resolution and the number of RFI in AMSR-2 6.9 GHz observations also less than that in AMSR-E 6.9 GHz observations.