Tajdarul Hassan Syed (Indian Institute of Technology Kanpur, India)

The Surface Water and Ocean Topography (SWOT) Mission provides high-resolution measurements of key parameters such as water surface elevation, channel width, and slope over inland rivers globally. The information provided through its node and reach products is crucial for estimating changes in river discharge, improving hydraulic and hydrological models, and providing reliable flood forecasts. Validating SWOT observations is a first step to ensure their reliability for water resource management, as it helps identify potential errors and quantify measurement uncertainties. Bitwise quality flags assess the quality of pixel cloud data used to generate SWOT’s reach and node products. Each bit flag corresponds to a source of error. This study evaluates the accuracy of SWOT Water Surface Elevation (WSE) measurements using bitwise quality flags, validated against in-situ gauge stations in five Indian river basins: Narmada, Mahanadi, Godavari, Krishna, and Cauvery. The performance of individual and combined bitwise filters on Node and Reach products is evaluated by minimizing median RMSE and maximizing the number of stations retained in each basin. This study demonstrates that excluding observations with Flag 19 (geolocation_quality_degraded) in the node product significantly improves the median RMSE across all basins while retaining approximately 70% of the stations. The accuracy improves further upon excluding the narrower channels (Width < 100 m) from the analysis. While the reach product yields a lower median RMSE than the node product in the Godavari, Krishna, and Cauvery basins upon excluding Flag 19 observations, it retains only ~30% of stations overall, making it less robust than the node product. This study provides improved optimization of SWOT data, enabling users to enhance accuracy while retaining the maximum number of viable observations.