Assessing and Enhancing SWOT Hydraulic Visibility and Slopes in Narrow Rivers Using High-Resolution Models

Evaluating the accuracy of SWOT water surface elevation (WSE) observations, including, slopes, and hydrodynamic signals, is crucial to determine their usefulness for river network monitoring and modeling.  The spatially distributed nature of SWOT measurements enables the characterization of local hydraulic signatures, which carry unprecedented information for improving hydraulic-hydrological models, enabling finer inferences of bathymetry and monitoring of river morphological evolution, which could support infrastructure management.

Ledauphin et al. (2025) demonstrated, using a comprehensive in situ dataset over the Franco-German Rhine (130–350 m wide), that SWOT elevation accuracy can exceed expectations for large rivers. Analyses of products from PIXC to reach-averaged scale confirm SWOT’s ability to detect fine-scale hydraulic variations driven by longitudinal hydraulic controls and dynamic phenomena such as flood wave propagation and associated water surface slope.

Building on these results, this study evaluates SWOT’s ability to capture hydraulic signatures over narrower rivers (20–80 m wide) located in France’s Grand Est region, with a focus on the Moselle River (25–80 m) This river, which includes diverse channel morphologies (e.g., step-pool sequences, meanders) as well as hydraulic structures (weirs, dams), benefits from long-term in situ gauge records complemented by field data, such as LiDAR bathymetric surveys and WSE profiles measurements. This rich dataset enabled to build and calibrate a high-resolution 1D HEC-RAS hydraulic model (Piasny G. 2023), providing simulated water surface elevation profiles for a range of discharges used as an independent reference for satellite-based validation. Complementary analysis were also performed on narrower rivers such as the Meurthe (20–40 m) and the Sarre (20–30 m, including a major flood event in May 2024).

Using data from the nominal science orbit, this study investigates SWOT performance close to the limits of its design specifications for narrow rivers. In this context, the use of official SWOT river products becomes challenging, as WSE profiles can be noisy, and multi-pass acquisitions introduce temporal variability in data quality that is difficult to filter with conventional methods, requiring advanced techniques. To overcome these limitations, hydraulic-preserving filtering methods specifically designed for SWOT data are applied to improve local slope estimation (Montazem et al., 2025; Larnier et al., 2025).  In the absence of full RiverSP coverage, the analysis here relies on PIXC pixel-cloud classes water-near-land and open-water, spatially filtered using a narrow riverbed polygon and existing flags, then projected onto the river centerline to produce a 1D product.

The impact of these processing and filtering methods is evaluated at fine scale through comparison with in situ measurements taken during the SWOT acquisitions and with WSE profiles from 1D hydraulic models at equivalent discharges. The use of high-resolution hydraulic model profiles enables a robust spatio-temporal validation of swot derived river altimetry and slope profiles at the node scale.

Variations in WSE due to discharge, bathymetry, and exceptional floods are well depicted with filtered SWOT data and validated against independent datasets. SWOT observations therefore demonstrate high accuracy across various hydrological conditions and river morphologies, even for narrow rivers, with a 1‑sigma error below 18 cm and a standard deviation below 30 cm compared to models and in situ measurements.