Characterization of the land surface water dynamics using in-situ and spaceborne GNSS reflectometry ; Étude des dynamiques du réservoir hydrologique continental de surface par réflectométrie GNSS : de l'in-situ au spatial

Spatial and temporal variations of the water cycle are still subject to large uncertainties. Yet, freshwater is essential for life on Earth and human populations, for health, industry and agriculture. Natural ecosystems also depend on freshwater availability. The interactions between the energy, water and biogeochemical cycles directly impact the climate. Within the frame of climate change mitigation, public and environmental policies need reliable estimates of water stocks and their dynamics. The knowledge of water levels and flooded extent are required, and the latter is currently poorly constrained. It is particularly the case for the large Amazon and Congo basins, where most remote sensing techniques are limited by dense tropical forests. The only global maps of surface water extent (GIEMS) were derived from passive microwave sensors, and their spatial resolution of 0.25° is too low for many hydrological applications. For this, the contribution of GNSS Reflectometry (GNSS-R) can be very useful. GNSS-R is a remote sensing technique that uses L-band GNSS signals after their reflection on the Earth's surface. These signals dedicated to positioning penetrate well the vegetation layers. The reflected signals are particularly sensitive to the presence of water bodies. Hydrological applications include both the monitoring of water levels at local scale, and the detection of large flooded areas in tropical forests. The latter was made possible by the launch of a spaceborne GNSS-R mission named CYGNSS in 2016, which collects numerous observations over the intertropical band. The different hydrological applications and spatial scales have been the focus of my thesis work. First, I studied water level on rivers monitored by in-situ GNSS-R. Existing methods were developed for oceanic case studies, and limited on rivers due to noisier measurements, a smaller number of GNSS satellites in visibility, as well as faster water level variations. All this generated a lot of noise that I was able to filter out by adding ...