GNSS-R as a source of opportunity for remote sensing of the cryosphere
- Autores: Francisco José Fabra Cervellera
- Directores de la Tesis: Adriano José Camps Carmona (dir. tes.), Estel Cardellach (dir. tes.)
- Lectura: En la Universitat Politècnica de Catalunya (UPC) ( España ) en 2013
- Idioma: inglés
- Tribunal Calificador de la Tesis: Manuel Martín Neira (presid.), Jordi Mallorqui (secret.), Ismael Colomina Fosch (voc.)
- Materias:
- Enlaces
- Tesis en acceso abierto en: TDX
- Resumen
This work evaluates the potential use of signals from the Global Navigation Satellite Systems (GNSS) that scatter off the Earth surface for the retrieval of geophysical information from the cryosphere. For this purpose, the present study is based on data collected with a dedicated reflectometry GNSS receiver during two field campaigns, which were focused on two types of characteristic surfaces of the cryosphere: thin sea ice covers and thick dry snow accumulations. During the first experiment, the complete process of formation, evolution and melting of sea ice was monitorized for more than seven months in a bay located in Greenland. This type of ice is typically characterized by its thickness, concentration and roughness. Different observables from GNSS reflections are analyzed to try to infer these properties. The ice thickness is linked to the free-board level, defined as the height of the sea ice surface. Accurate phase altimetry is achieved, showing good agreement with an Arctic tide model. In addition, the long term results of ellipsoidal height retrievals are consistent with the evolution of the ice surface temperature product given by MODIS, which is a key parameter in the rate of growth of sea ice. On the other hand, the presence of salinity in the sea ice modifies its dielectric properties, resulting in different amplitude and phase for the co- and cross-polar components of the complex Fresnel coefficients. The polarimetric measurements obtained show good agreement with visual inspections of ice concentration from an Arctic weather station. Finally, the shape of the reflected signals and its phase dispersion are tested as potential signatures of surface roughness. For comparison, ice charts of the experimental area are employed. In particular, maximums in roughness given by the GNSS observables coincide with fast ice events. Fast ice is defined as ice anchored to the coast, where the tidal movements contribute to the development of strange patterns, cracks, and fissures on its surface, thus consistent with the GNSS-R roughness retrievals. The second experiment took place on Antarctica, monitoring a pristine snow area which is well-known for the calibration of remote sensing instruments. Due to the relative stability of the snow layers, the data acquisition was limited to ten continuous days. Interferometric beats were found after a first analysis of the amplitude from the collected signals, which were consistent with a multipath model where the reflector lies below the surface level. Motivated by these results, a forward model is developed that reconstructs the complex received signal as a sum of a finite number of reflections, coming from different snow layers (a snow density profile obtained from in-situ measurements). The interferometric information is then retrieved from the spectral analysis applied to time series from both real and modeled signals (lag-holograms). We find that the frequency bands predicted by the model are in general consistent with the data and the lag-holograms show repeatability for different days. Then, we attempt a proper inversion of the collected data to determine the dominant layers of the dry snow profile that contribute to L-band reflections, which are related to significant gradients of snow density/permittivity.
