| Summary: | Radar backscatter represents the portion of a transmitted electromagnetic signal that is redirected back toward the antenna from a target on ground. Its properties change depending on the radar wave frequency and polarization, acquisition geometry, ground cover type, and soil conditions. Backscatter information is of paramount importance for the design of SAR missions and is widely used for the development of scientific models in the fields of, e.g., the biosphere and cryosphere. The main goal of this work is to exploit the global TanDEM-X SAR data set to model radar backscatter at X band, considering different acquisition parameters and land cover types and to provide then the scientific community with an up-to-date set of backscatter models at a global scale. A novel approach for statistically model the backscatter information, which takes into account the quality of the input measurements, has been developed. The results are weighted polynomial models for different land cover types, taken from the ESA GlobCover map. A dedicated validation approach is presented as well, together with additional considerations on backscatter seasonality and a dedicated analysis of backscatter behavior over tropical rainforests. The attention is then focused on the Greenland Ice Sheet, which is characterized by the presence of different kinds of snow cover, from dry to wet snow. Here, the insufficient level of detail that is provided by the GlobCover map over Greenland (one single class for the entire ice cap) does not allow for a reliable modeling of backscatter. This obstacle set the motivation for developing a new approach for analyzing the information content of interferometric TanDEM-X data, aimed at locating different snow facies by means of the c-means fuzzy clustering algorithm. A set of four different snow facies has been derived, and their properties interpreted with the help of external reference data. The obtained map has then been used to generate an incidence angle dependent backscatter model for each snow facies, ...
|
|---|