Simultane Schätzung von Topographie und Dynamik polarer Gletscher aus multi-temporalen SAR Interferogrammen

摘要

This thesis presents a new technique to simultaneously estimate topography and motion of polar glaciers and ice sheets from multi-temporal SAR interferograms. In this approach the unknown parameters, i.e. topography and surface displacement, are estimated from a redundant number of observations. The resulting redundancy is systematically used to attach realistic accuracy and robustness measures to all estimated unknowns and to reveal possible gross errors in the observations or errors in the model. Additionally, the redundancy facilitates replacing very restrictive model components by more flexible models with a higher amount of free parameters. This leads to a better description of the real behavior of the object. The approach combines several SAR interferograms into a least-squares adjustment based on the Gauss-Markov model. In order to connect the multi-temporal data sets, a spatio-temporal model is proposed that describes the properties of the surface and its temporal evolution. The parameterization of the model can be adapted considering the properties of the object under investigation. Tests with simulated data show that, with this technique, both topography and motion of polar ice caps can be estimated with an accuracy of a few meters or a few centimeters per day, depending on the configuration of the adjustment. The quality of the estimated parameters depends on the accuracy of the phase observations, the observation geometry, and the number of involved data sets. Analyses of the robustness of the approach additionally indicate a high robustness of the estimated unknowns with respect to gross errors in the observations. The influence of the atmospheric path delay on the unknowns is analyzed and quantified for various observation geometries and different atmospheric conditions. It is shown that errors in the assumed deformation model significantly falsify the estimated parameters. Thus, the configuration of the spatio-temporal model must be chosen carefully. In addition to the estimation of surface topography and motion, the method and its inherent analysis tools allow to determine the observation geometry that is necessary to reach a pre-defined accuracy level in advance. Therefore, the method is a very useful tool for project planing. By means of a case study aimed at the observation of a huge ice dome in the Russian arctic the feasibility of the approach in real-life situations is demonstrated.%Die vorliegende Arbeit präsentiert eine neue Methode zur simultanen Schätzung von Topographie und Bewegung polarer Gletscher und Eisflächen aus multi-temporalen SAR Interferogrammen. Die unbekannten Parameter Topographie und Bewegung werden dabei aus einer redundanten Anzahl an Beobachtungen geschätzt. Die Redundanz wird systematisch zur Ableitung realistischer Genauigkeitsaussagen sowie zur Aufdeckung möglicher grober Fehler bzw. Modellfehler genutzt. Die Überbestimmung eröffnet ebenfalls die Möglichkeit, restriktive Modellkomponenten durch flexiblere Modellkomponenten mit einer größeren Anzahl an freien Parametern zu ersetzen und gewährleistet dadurch eine bessere Beschreibung der tatsächlichen Verhältnisse. Der Ansatz basiert auf der Kombination einer Anzahl von SAR Interferogrammen in einer Kleinste-Quadrate-Ausgleichung nach dem Gauss-Markov Modell. Zur Verknüpfung der multi-temporalen Datensätze wird ein räumlich-zeitliches Modell vorgeschlagen, das die Eigenschaften der beobachteten Oberfläche und deren zeitliche Veränderung beschreibt.