Polarimetric Radiometers and their Applications

Research output: Book/ReportPh.D. thesis

396 Downloads (Pure)


A new space borne L-band radiometer, known as SMOS, is currently under phase-B studies at the European Space Agency. With high resolution the instrument is intended for measurements of soil moisture and ocean salinity. The influence from other effect must be known, however, and studies of Earth signatures at L-band are carried out prior to the mission, thus involving assembly of new L-band radiometers as well as field campaigns.

This thesis describes the design and integration of a new airborne radiometer, measuring the full Stokes vector. Initial considerations include evaluation of the influence from the aircraft motion on the measured parameters. It is shown how knowledge of the aircraft attitude to a 0.05° level is necessary, and how corrections can be applied to the measured data. The instrument design involves concept studies, and traditional implementations of polarimetric radiometers are studied. A new radiometer type may be attractive, however, enabled by the availability of new analog to digital converters: the digital subharmonic sampling radiometer. The radiometer has a very simple analog front-end, only with an RF-filter and amplifiers, while all further data processing is moved to the digital domain. The RF is sampled at a lower frequency without down conversion, and it is shown, how a proper selection of the sampling frequency can lead to a combined digitizing and conversion to base band using the well known frequency band aliasing.

The radiometer study is a technology development, opening future perspectives for signal processing, increasing the total radiometer performance, and a digital radiometer has a unique flexibility for reconfiguration. Adaptive filtering of the input signal will provide an opportunity to filter out unidentified RF-interference in the analog pass band, and a band extension beyond the protected 27 MHz bandwidth is thus possible, as well as an active protection within the band. The other perspective is the opportunity to achieve smaller and less power consuming instruments with the digital technique, along with the potential reduction of the front-end drifts, typically the limiting factor for receiver stability.

Based on the study, a radiometer of the digital subharmonic sampling type is constructed, and evaluation of the complete system shows the expected performance in instrument sensitivity. In stability the design goal of 100 mK drift within 15 minutes of operation is reached, and the instrument is thus useable for field campaigns aiming at determination of potential Earth signatures of importance for the SMOS mission.

Field experiments over a land site at Institut National de la Recherche Agronomique in Avignon, France as well as an airborne experiment over Danish ocean target sites are carried out. The Avignon experiment demonstrates the calibration of the EMIRAD radiometer over a period of four month, and it is demonstrated that a calibration within 1 K is achievable. The experiment also proves the presence of polarimetric signatures under special soil conditions, when the instrument is scanned over a range of azimuth look angles. Bare soil, plowed and sowed in a wave pattern of 8 cm height, is shown to give 10 K of peak-to-peak variation in the measured signals, and more than 15 K variation with the look angle is identified in a cornfield with 80 cm crop. The experiments identify both a 1st and a 2nd harmonic component, and thus it is shown, that some polarimetric signatures, known from measurements of wind driven patterns on the sea surface at higher frequencies, are also present at land sites. The experiment concludes, that a potential risk exist, that results from SMOS may be influenced by these effects, when large homogenous areas are observed.

The other field campaign aimed at identification of eventual wind driven sea surface signatures at L-band. At higher frequencies it is well known, that some Kelvin of variation may result from a 360° azimuth scan, and the effect is important for the SMOS mission in order to evaluate the necessity for a wind direction dependent data correction. As part of the experiment four flights have been carried out; one technical test flight and three scientific flights at wind speeds of 3.6 m/sec., 5.1 m/sec. and 11.0 m/sec, respectively. Methods for correction of the measured signals in order to remove the influence from the aircraft attitude are tested, and it is demonstrated, how a large-scale platform movement, e.g. a 45° roll variation, may provide a good model for the correction.

For all flights it was noticed, that large noise was still present in the signals after the corrections, and peak-to-peak variations of several Kelvin were noticed. From the first two scientific flights, no azimuth signatures were identified, and the data analysis indicated, that derived azimuth directions resulted from a random distribution. A conclusion based on these data is not possible, however, as the statistical background is too small compared to the amount of noise. For the high-wind case a larger number of circles were measured, and the flight altitude was varied in order to investigate, if a larger antenna footprint, i.e. a larger spatial integration, would decrease the noise. The data shows, however, that no decrease can be proven, and instead the circles are treated as equal samples of the same sea pattern. Averaging shows, that the 2nd, 3rd and 4th Stokes parameters have a random behavior, and from the statistical analysis it is seen, that less then 30 mK of harmonic magnitude can be identified in the parameters. Due to the limited number of signatures, however, the statistical uncertainty of this value is large, and additional experiments are needed to improve the precision.

The 1st Stokes parameter has a clear 490 mK 2nd harmonic component. It is shown, however, that the signature has a significant correlation with the downwelling galactic background radiation as a function of the azimuth look angle, and a calculation of the background level for one of the circles shows, that a significant fraction of the signature can be removed. The effect must be subject for further investigation, however, and individual corrections for all measured data must be modeled in order to correct the SMOS data.
Original languageEnglish
Place of PublicationKgs. Lyngby, Denmark
PublisherTechnical University of Denmark
Number of pages144
Publication statusPublished - Apr 2003

Cite this

Søbjærg, S. S. (2003). Polarimetric Radiometers and their Applications. Technical University of Denmark.