Analysis of Propagation Effects in TerraSAR-X Images

Kurzfassung

TerraSAR-X, the first civil German synthetic aperture radar (SAR) satellite has been successfully launched in 2007, June 15th, 4:14 o`clock. After 4,5 days the first processed image has been obtained. The overall quality of the image was outstanding, however, suspicious features could be identified which showed precipitation related signatures. These rain-cell signatures motivated a further in-depth study of the physical background of the related propagation effects and in this contribution we focus on propagation effects in the case of TerraSAR-X. Synthetic aperture radars are often considered as day/night and all-weather imaging systems. Whereas the first argument is true, the second does not hold in every case as will be shown in more detail. From earlier missions such as the Shuttle imaging radar mission SIR-C/X, SRTM, ERS-1/2, ENVISAT-ASAR it was known that in some cases there might be a influence due to tropospheric effects, since the attenuation of rain may, in some cases, effect the signals of the sensors along the propagation path. In X-band, for instance, the specific attenuation of the signals may be around 1 dB/km assuming a rain-rate of 40 mm/hr, an occurrence, which may frequently take place for tropical areas over the rain-forest. As a first step, we provide an estimate on the potential influence of rain (attenuation) on SAR imaging for typical weather scenarios. During the commissioning phase, a total of 12000 scenes have been investigated for potential propagation effects and about 100 scenes have revealed atmospheric effects to a visible extent. Some of the particularly interesting events have been selected and will be shown and discussed in greater details. An interesting case of a data acquisition over New York will be presented which shows typical rain-cell signatures and the SAR image will be compared with weather-radar data acquired nearly simultaneously within the same minute. By comparing the images it can be clearly shown that reflectivities in the weather radar image of 50 dBz may cause visible artefacts in the SAR images. Furthermore, in this contribution we discuss the influence of the atmosphere (troposphere) on the external calibration (XCAL) of TerraSAR-X. By acquiring simultaneous weather-radar data over the test-site and the SAR-acquisition it was possible to flag effected SAR data and to exclude them from the procedure to derive the absolute calibration constant. Thus, it was possible to improve the absolute calibration constant by 0.15 dB. The conclusions in this contribution are manifold. First of all, propagation effects can be very important and they need to be considered in interpreting radar images. Attenuation caused by heavy rain events, has been identified as the main potential reason for image degradation and artefacts. The underlying analysis comprised more than 12000 data takes of TerraSAR-X. Maximum attenuation up to 20 dB through the precipitation volumes may occur in the cases of heavy precipitation, for instance, for data takes acquired over the Brazilian rainforest. Clouds with low liquid-water content or low rain rates and homogenous distribution will cause no or negligible distortions (visible artefacts). The flexible TerraSAR-X instrument offers new possibilities for investigation of propagation effects in SAR-imaging and the results presented herein may be also important for future systems operating with higher nominal frequencies, e.g. the Ku-band where propagation effects due to precipitation will become even more severe. For such cases the results presented in this contribution may act as a useful reference. In a test-case, it was shown for the first time that simultaneous precipitation measurements with weather radars are in good agreement with the observed rain-cell signature observed in the SAR-images and that such simultaneous measurements that are now available for further in-depth study of attenuation effects are an invaluable source of information for further investigations of attenuation effects and their quantification.