AUTHOR=Bazzi Henri , Baghdadi Nicolas , Ngo Yen-Nhi , Normandin Cassandra , Frappart Frédéric , Cazals Cecile 

TITLE=Assessing SWOT interferometric SAR altimetry for inland water monitoring: insights from Lake Léman

JOURNAL=Frontiers in Remote Sensing

VOLUME=Volume 6 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/remote-sensing/articles/10.3389/frsen.2025.1572114

DOI=10.3389/frsen.2025.1572114

ISSN=2673-6187

ABSTRACT=Monitoring water levels is crucial for managing water resources and addressing climate change challenges. The new Surface Water and Ocean Topography (SWOT) mission provides unprecedented spatial and temporal resolution estimates of water surface elevations (WSEs) globally. This study evaluates the accuracy of SWOT WSE estimates over Lake Léman, Switzerland. We evaluated the SWOT L2-HR-Raster product from the calibration and nominal phases using in situ measurements of water levels and compared its performance with other missions, including Sentinel-3A (S3A), Sentinel-3B (S3B), Sentinel-6 (S6), and Global Ecosystem Dynamics Investigation (GEDI) altimetry. From over 141 acquisitions, SWOT achieved a root mean squared error (RMSE) ranging from 13 cm to 21 cm compared to in situ water levels, depending on the measurement quality reported in the product. Data flagged as good quality had an RMSE of 19 cm and a correlation coefficient (R) of 0.8, although these represented only 42% of the total measurements. When considering WSE estimates of all quality levels and applying a median outlier filter, the RMSE reaches 21 cm, with a correlation coefficient of 0.79, while retaining approximately 83% of the dataset. A consistent bias of −10 cm was observed across the time-series. An analysis of SWOT accuracy relative to instrumental parameters revealed that nadir and near-nadir acquisitions (viewing angle near 0°) exhibited very high uncertainty, with mean absolute differences from in situ water levels potentially exceeding 5 m. To explore the sources of errors in SWOT WSE, a random forest analysis showed that atmospheric perturbations had the most significant impact on the SWOT WSE estimation accuracy. These perturbations were linked to dry tropospheric delays affecting interferometric height measurements and atmospheric effects on the Ka-band sigma0 values. Compared to other missions, SWOT demonstrated slightly better accuracy than S3A, S3B, and S6, with an RMSE of 11 cm on a daily scale, compared to 13 cm, 18 cm, and 20 cm for these three Sentinel missions, respectively. All radar-based missions (S3A, S3B, S6, and SWOT) exhibited correlation coefficients exceeding 0.95 with in situ water levels. In contrast, GEDI LiDAR data showed the highest RMSE (46 cm), a bias of 27 cm, and a correlation coefficient of 0.45.