Résumé
The water issue is currently a major challenge for our societies. Even if water is a renewable resource, its natural cycle supports great stresses, both human activity and climate change. River discharge is a key variable of the water cycle, whose quantification requires heavy field measurements. Therefore the global monitoring of river discharge remains problematic and satellite remote sensing techniques could be a major asset. Earth Observation radar techniques are currently limited to the measurement of surface variables and cannot measure river bottom hydraulic parameters. The current study proposes a method to estimate these parameters from surface variables, in order to estimate the river discharge. This method has been validated on exact simulated data and its robustness to measurement noise has been studied on noisy simulated data and on real data. The second part of this work explores the abilities of radar along-track interferometry to measure river surface velocity. Velocity measurements carried out on the Rhône river during an airborne campaign (ONERA-Cemagref) show a strong consistency with ADCP field measurements. M4S backscattering model, tested in the framework of this study, is poorly adapted to river scenes : it appears to be extremely sensitive to wind intensity which could be explained by a poor modeling of surface roughness. As part of the SWOT mission program (NASA-CNES) a method allowing in situ characterization of river surface roughness has been developed. It was validated under laboratory controlled conditions. It was implemented on the Rhône river to characterize water surface and quantify the influence of wind intensity on the water surface roughness parameters.