An Ionospheric Calibration Scheme for the BIOMASS Pol-InSAR Data Space

Kurzfassung

Long wavelength SAR observations are being high-lightened because of their excellent performances on the global forest and the glacier observations. Several missions including TERRASAR-L in L-band and BIOMASS in P-band are now under planning phase. One of the biggest obstacles of long wavelength SAR observation is the effect of the Earth’s ionosphere. Interactions among the free ions, the Earth’s magnetic field and the electromagnetic waves delays its propagation group velocity, rotate its polarization plane, and advances its phase, in L- and P-band spectrum. These effects degrades the polarimetric information in the final SAR product, cause miscorregistration along both range and azimuth direction, and introduces unwanted interferometric phase. In L- and P-band observations, the higher order effects of the ionosphere are negligible, and the only one parameter, total electron content (TEC), can represent the ionospheric state. TEC can be estimated from Faraday rotation (FR), when the quad-pol data are available and the observation region is not magnetic equator. In the presentation, in addition to FR, the relative azimuth direction shift of two azimuth subband data are fused into the FR derived TEC in order to suppress the TEC variation. A discussion about the enhancement of the TEC estimation using these two data will be followed. The next point of the presentation will be about the estimation of differential TEC. The first adversary effect of the ionosphere on InSAR is the decorrelation owing to the shift of focusing along the azimuth direction. Azimuth shifts can be 1) directly measured from the comparison of two scenes, 2) indirectly inferred from differential TEC estimation, or 3) predictable from sub-band interferograms. The presentation compares several shift estimation-correction schemes and then discusses their performances. The second adversary effect is the ionospheric fringe in InSAR. This is very sensitive to the small change of TEC and, therefore, can tell keenly which kind of differential TEC estimation scheme can give the best answer. The required accuracy depends on the type of the application. And each TEC or differential TEC estimation methods have different characteristics on their estimation result and their applicability, depending on the sensor specification and the geolocation. These points will be discussed in the discussion section. In addition, the treatment of the spatial or spectral filtering and the directionality of the ionosphere on the SAR data will be shortly discussed before showing the results.