Florin Zăinescu (Aix-Marseille University, University of Bucharest, Romania); Florin Tătui (University of Bucharest, Romania); Edward J Anthony (Aix-Marseille University, University of Bucharest, France)

The launch of the Surface Water and Ocean Topography (SWOT) satellite mission offers unprecedented opportunities for surface hydrology in deltas, particularly those containing large flooded vegetation surfaces. SWOT reveals delta plain-river connectivity, crucial for sea level rise adaptation through sedimentation and is promising as a calibration/validation data source in these data sparse regions. However, the raster product often suffers from degraded and unusable data from various errors including spectral ringing effects that hinders the immediate use of the raster product (L2_HR_RASTER). This work presents the development and application of a comprehensive methodological workflow for processing SWOT Level 2 Pixel Cloud (L2_HR_PIXC) data to characterize in detail the hydrological dynamics of several deltas and wetlands (e.g., Rhône, Danube, and Mississippi deltas). The workflow is implemented entirely in Python, and addresses the full processing chain from data acquisition to error removal and final analytical products and is available on GitHub. A classification algorithm is applied to the point data to refine the distinction between water and land, improving the standard raster product classification particularly in thin ridge areas such as levees, beach ridges and coastal barriers, often misclassified as water. Water-surface elevation points are first filtered for errors through local outlier removal and the application of weighted flags. These filtered points are then interpolated into a regular 30 m grid using a hybrid k-nearest neighbors (KNN) approach and aggregated into a spatio-temporal data cube, forming the basis for multi-temporal analyses. Key final products derived from this data cube include maps of mean water level; temporal variability highlighting areas with the greatest fluctuations; a spatially weighted water-frequency map; anomaly maps; and minimum and maximum water height and extent maps, which, when integrated with wind speed and direction, river discharge, and wave conditions, allow us to better understand water dynamics and ongoing processes. The final accuracy of the water-surface elevation estimates is quantitatively assessed through validation against water-level data measured at in situ stations. Initial validation of SWOT data against these in situ measurements revealed a high level of accuracy (R² > 0.8). Finally, we show the validation of a numerical model using SWOT data for the Danube Delta. This demonstrates SWOT's crucial role in creating reliable models for poorly monitored areas, which is essential for understanding these systems and predicting their response to climate change.