Interpretation of Single-Pass Pol-InSAR Vegetation Signatures at X-Band: Dual- vs. Quad-Pol Case

Abstract

The analysis of the phase centre locations in early airborne high-resolution single-pass (Pol)-InSAR at X-Band indicated – in contrast to what was “believed” at this time – significant penetration depths not only into sparse but even in medium and dense vegetation. This new understanding triggered a number of case studies for the estimation of vegetation parameters from single- or dual-Pol-InSAR airborne measurements at X-band with surprisingly successful results indicating the potential of Pol-InSAR techniques at X-band [1, 2]. However, the lack of quad-pol and multi-baseline data sets and/or the presence of temporal decorrelation left many questions open as for example which mechanism causes volume decorrelation: Is it penetration meaning a continuous volume case or the presence of forest gaps referring to a discrete volume case associated to the high spatial resolution of the sensors? TanDEM-X is the first space-borne single-pass Pol-InSAR instrument operating in dual-pol mode. Since 2010, it acquires data primarily for generating a global DEM with high accuracy. Indeed TanDEM-X allows investigating the performance of forest height inversion by means of the Random Volume over Ground model (RVoG) in a large number of sites [3, 4] but did not allow a direct validity assessment of the used inversion models. The validation of the RVoG model in terms of the shape of the coherence region is possible only when quad-pol data are available. In the two-dimensional case the coherence region is always elliptic. DLR’s F-SAR is the first operational single-pass quad-Pol-InSAR airborne instrument for X-Band. It operates in a single-pass quad-Pol-InSAR mode with a spatial baseline of about 1.6 m at a bandwidth up to 800 MHz. F-SAR data allows for the first time to investigate the three-dimensional coherence region and to draw conclusions about the validity of the different Pol-InSAR inversion models. In this paper we investigate scattering and propagation characteristics at X-band using single-pass quad-Pol-InSAR data over different test sites. We interpret quad-pol coherence regions over vegetation scatterers and draw conclusions about the validity and constraints of different Pol-InSAR inversion models for a variety of vegetation conditions. Individual scattering mechanisms are interpreted in terms of the polarimetric scattering angle alpha and associated to their location within the vegetation i.e. the associated penetration depth. In a second step we compare the quad-Pol-InSAR (F-SAR) signatures against dual-Pol-InSAR (TanDEM-X) signatures and discuss the effect of the differences in acquisition geometry and system specifications (e.g. NESZ). Based on this we are able to understand the information content and potential of Tandem-X data in a better way. Finally, we generalise the quad- vs dual-Pol-InSAR discussion evaluating optimised dual-pol configurations and comparing them to the quad-pol case. [1] F. Garestier, P. C. Dubois-Fernandez, K. P. Papathanassiou, “Pine Forest Height Inversion Using Single-Pass X Band PolInSAR Data”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 46, NO.1, January 2008. [2] I. Hajnsek, F. Kugler, S. Lee, K. Papathanassiou, “Tropical Forest Parameter Estimation by means of Pol-InSAR: The INDREX II Campaign, IEEE Transactions on Geoscience and Remote Sensing, 2009. [3] F. Kugler, S. Sauer, S.-K. Lee, K. Papathanassiou & I. Hajnsek, “Potential of TanDEM-X for forest parameter estimation” Proceedings EUSAR 2010, Aachen, 2010. [4] F. Kugler, I. Hajnsek, “Forest characterisation by means of TerraSAR-X and TanDEM-X (polarimetric and) interferometric data”, Proceedings of IGARSS, Vancouver, July 2011.