The Airborne EMIRAD L-band Radiometer System Used in the CoSMOS Campaigns
Skou, Niels; Soebjaerg, Sten; Kristensen, Steen; Balling, Jan
Technical University of Demark, DENMARK

Introduction. An airborne, fully polarimetric L-band radiometer system, primarily intended for sea salinity and soil moisture campaigns in support of SMOS, the so-called CoSMOS campaigns, is described. The radiometer is of the digital kind: the L-band signals are directly fed into fast A to D converters using sub-harmonic sampling. All Stokes parameters are calculated digitally in a fast FPGA. Special attention is paid to detection and mitigation of radio frequency interference (RFI): the digital radiometer principle with Nyquist sampling provides unique possibilities for RFI detection and mitigation before final integration [1]. The antenna system consists of 2 large Potter horns and bulky waveguide OMTs (Ortho Mode Transducers) for almost ideal antenna patterns and low loss. One antenna having a 38° half power beam width (HPBW) is pointed nadir, the other (31° HPBW) 40° aft. Thus each point on the ground is covered twice with different incidence angle and very short time interval. The radiometer features input multiplex switches to handle the 2 antennas. The antennas are of considerable size: the nadir looking horn has a diameter of 41 cm and a length of 151 cm incl. the OMT, while the aft looking horn has a diameter of 51 cm and a total length of 174 cm. The radiation patterns are from a practical point of view rotational symmetric. The performance of the Potter horns is very satisfactory with side- and backlobes being suppressed more than 40 dB while the cross polar level is even smaller. The receiver is a fully polarimetric correlation radiometer, operating in the protected 1400 - 1427 MHz band, featuring direct sampling. Two MAXim (MAX106) A/D converters directly sample the L-band signals. The sampling frequency is 139.4 MHz, and the input signal level to the converters, corresponding to a 300 K radiometer input signal, is set to -13 dBm in order to minimize nonlinearities in the sampled signal due to saturation. The calibration is carried out using internal loads and a noise diode, adding around 125 K to the radiometer input. This setup ensures frequent internal calibration capability, and it allows the radiometer to calibrate the possible phase difference between the two input channels in front of the digital correlator. The digital back-end: The required digital signal processing is implemented in an FPGA. On start-up, the FPGA will configure itself from the built-in PROM holding already uploaded firmware. The FPGA calculates any bias on the incoming sampled signals, and via two digital to analog converters, the DC offset from the ADCs is neutralized. A clock signal feeds the FPGA, and via controllable delay circuits also the ADCs. This way, minor phase imbalances can be fine tuned to a minimum under computer control, which is especially important for polarimetric performance. The FPGA program: Basically, squaring and correlations are carried out to find polarimetric signatures, and following the idea outlined in [2], second and fourth order central moments are calculated for later off-line RFI detection. The quadrature channels are formed in IQ units following the ADCs, and a switch selects which to use. The calculation of the Stokes vector and the kurtosis elements takes place in a multiplication unit, resulting in a 6-element output vector. The signals are integrated in an accumulator unit to yield results with proper standard deviation for estimation, and to reduce data rate. After splitting the signal in several sub-bands the same procedure is carried out for each individual band. Finally, data is stored for later processing during which kurtosis is calculated, RFI is detected, and mitigation followed by proper integration in time and frequency takes place. Campaigns and aircraft installations: The CoSMOS campaigns have covered a range of targets at very different locations. In the fall of 2005 agricultural areas were covered in Australia. In the spring of 2006, the North Sea west of Norway was the target. Finland was the target area in 2007 with campaigns over the ice-covered sea in winter, and over open sea in summer. In 2008 the so-called Rehearsal Campaign saw the coverage of agricultural and urban areas in Germany and Spain, as well as transits including France and the Mediterranean Sea. The Cal/Val Campaign was carried out in 2010, covering agricultural targets in Denmark and Germany. Finally, the Dome-C in Antarctica was investigated for surface homogeneity in the DOMECair campaign in 2013. The radiometer system has been installed on several aircraft for these campaigns. A Rockwell Aero Commander was used in Australia in 2005, while a Short SC-7 Skyvan was used for the campaigns in Europe, 2006 - 2010. For the Antarctic DOMECair campaign the system was mounted on a Basler BT-67. RFI was experienced right from day one during the campaigns, despite operating in the protected band, as reported in [3] and [4]. The percentage of data corrupted by RFI was low in Australia and over open ocean, while it varied considerably over Europe - from a few percent in the north to 10 - 20 percent in the south (higher percents typically over urban areas). As expected, the kurtosis method is a good RFI detector in many cases, but also the polarimetric method (looking for non-zero 3rd and 4th Stokes parameters) is a powerful tool [5]. Typically, kurtosis and polarimetry detects about the same percentage of RFI, but to a large extent the methods complement each other: some samples are detected by both methods while more samples are detected by one or the other method. References: [1] N. Skou, S. S. Sobaejrg, J. Balling, and S. S. Kristensen: "A Second Generation L-band Digital Radiometer for Sea Salinity Campaigns", Proc. IEEE Geosc. Remote Sens. Symposium, Denver, pp. 3984-3987, July 2006. [2] C. S. Ruf, S. M. Gross, and S. Misra, ''RFI detection and mitigation for microwave radiometry with an agile digital detector'', IEEE Trans. Geosc. Remote Sens., vol. 44, no. 3, pp. 694-706, Mar. 2006. [3] N. Skou, S. Misra, J. E. Balling, S. S. Kristensen, and S. S. Sobjaerg: "L-band RFI as Experienced During Airborne Campaigns in Preparation for SMOS", IEEE Trans. on Geosc. Remote Sens., Vol. 48, No. 3, pp. 1398-1407, Mar. 2010. [4] J. E. Balling, S. S. Kristensen, S. S. Sobjaerg, and N. Skou: "Surveys and Analysis of RFI in Preparation for SMOS: Results from Airborne Campaigns and First Impressions from Satellite Data", IEEE Trans. on Geoscience and Remote Sensing, Vol. 49, No. 12, pp 4821-4831, December 2011. [5] J. Balling, S. S. Sobjaerg, S. S. Kristensen, and N. Skou: ''RFI detected by kurtosis and polarimetry'', Proceedings of Microrad 2012, March 2012.