No document available.
Abstract :
[en] Conventional satellite-based altimetry, which relies on radar altimeters or microwave radiometers, achieves centimeter-level accuracy but is limited by spatial and temporal constrains, particularly in regions above and below approximately 80 degrees latitude in the polar regions. In contrast, L-band GNSS signals reflected from the Earth’s surface offer global coverage and can operate under all weather conditions. These signals can effectively fill the temporal and spatial gaps left by traditional satellite altimeters, addressing these limitations. While reflected signals, or multipath, have traditionally been a challenge in GNSS Positioning
Navigation and Timing (PNT) due to their potential to degrade measurement accuracy, they have proven to be a valuable resource for remote sensing. The study of these reflected signals, known as GNSS-Reflectometry (GNSS-R), demonstrates how an error source in PNT can be repurposed for a variety of applications, such as reconstruction of digital elevation models, altimetry, soil moisture estimation, and more.
This doctoral thesis explores the use of GNSS-R carrier phase measurements for altimetry retrieval across a variety of surface types worldwide, including oceans, sea ice, and ice sheets, with a particular focus on polar regions. The analysis used data from the Spire Global Inc. Grazing angle GNSS-R (GG-R) satellite constellation, providing a robust dataset to investigate the potential of dual-frequency GNSS-R altimetry at low elevation angles, ranging from 5 to 30 degrees.
Key findings of this dissertation include altimetric results for ocean, sea-ice, and ice-sheet, with decimeter and meter precision for water and land reflections, respectively. This work comprehensively spans model formulation, software development, parameter opti-mization, and validation against existing platforms and products. The development of the GRazing Angle ALTimetry (GRAALT) software enabled efficient processing and analysis of GNSS-R data, resulting in a refined Level-2 (L2) altimetry product derived from 113,765 GNSS-R tracks. This product achieved a SLA Root Mean Square Error (RMSE) of 16.8 cm with a standard deviation of 16.9 cm. Additionally, the L2 product includes parameters necessary for height recomputation, enhancing its flexibility and utility.
