Edward Salameh (Univ Rouen Normandie, Univ Caen Normandie, CNRS, M2C, UMR 6143, F-76000 Rouen, France, France); Frédéric Frappart (ISPA, UMR1391 INRAE Bordeaux Science Agro, F-33140, Villenave d'Ornon, France, France); Cassandra Normandin (ISPA, UMR1391 INRAE Bordeaux Science Agro, F-33140, Villenave d'Ornon, France, France); Valentine Sollier (ISPA, UMR1391 INRAE Bordeaux Science Agro, F-33140, Villenave d'Ornon, France, France)

River floodplains play a major role in water, carbon, and biogeochemical cycles. They serve as an interface between the catchment and the main river channel and are characterized by a complex water dynamic (e.g. multiple water sources and pathways, hydrological connectivity, interactions with groundwater). As such, they require accurate monitoring to better understand their seasonal variability, manage flood risks, and assess their ecological functioning in a changing climate. The Surface Water and Ocean Topography (SWOT) satellite, equipped with a Ka-band Radar Interferometer (KaRIn), offers high-resolution altimetric data with unprecedented potential for monitoring water surface elevation and extent. This study aims to assess the performance of SWOT in measuring water level and detecting water surfaces over floodplains of two river systems with contrasted characteristics: the Parana and the Ob Rivers, located respectively in temperate and boreal climates. The assessment is performed using the L2_HR_PIXC and the L2_HR_Raster products which provide geolocated water surface elevation, backscattering coefficients, pixel classification, and other relevant parameters. For the Paraná River Floodplains, in situ water level measurements were retrieved from the Argentine National Water Information System (SNIH) providing water level records. For water extents, the accuracy of SWOT water detection was validated using two datasets, the Global Land Analysis and Discovery (GLAD) product and water masks extracted from MODIS observations. SWOT Water levels were compated to in situ measurements from multiple stations along the Parana River and its tributaries. The mean absolute error (MAE) reached 20 cm while correlation coefficients (R) exceeded 0.97 for most stations, demonstrating SWOT consistency despite the complexity of the environment. Regarding water surfaces, a good agreement is generally found over inland water (i.e., rivers, lakes and floodplains) when compared to inundation maps from GLAD with overall accuracies ranging between 66% and 81%. Furthermore, water detected by SWOT occurred over agricultural areas in the Parana basin. This represents both a limitation due to misclassification generating an overestimation of water surfaces, but also an opportunity to monitor irrigation periods for agricultural areas. We also detected low backscatter signals classified as water over urban areas, which are likely due to strong reflections from buildings and other urban structures. This highlights the need for further refinement of SWOT algorithms to reduce false positives in urban environments. Parallel analyses are ongoing on the Ob River floodplains to evaluate SWOT’s performance across contrasting climatic and hydrological contexts.