SWOT performance in monitoring water level of high-mountain lakes on the Tibetan Plateau

• SWOT data effectively capture water level dynamics of small lakes in high-mountain regions. • Gaussian KDE method improves SWOT-based water levels of small size lakes accuracy. • Geolocation quality and classification quality flags are key indicators of water level retrieval errors. • SWOT retrieved water levels for 2,900+ small lakes in the Himalaya with minimal layover effect. The Tibetan Plateau hosts over 60,000 lakes, which are critical to the regional hydrological cycle and potential downstream hazards. However, conventional satellite altimetry missions fail to capture many of these lakes, leaving major gaps in understanding their dynamics and impacts. The Surface Water and Ocean Topography (SWOT) mission addresses this limitation by providing wide-swath measurements of water surface elevation and area, especially in complex terrains. This study introduces a novel Gaussian kernel density estimation approach to retrieve water levels of high-mountain lakes from SWOT observations, improving retrieval accuracy and expanding the mission’s utility for hydrological and cryospheric studies. Using the SWOT Level-2 High-Rate Pixel Cloud product and in-situ validation data collected at nine lakes across the Qiangtang region and the southeast of the Tibetan Plateau during 2023‒2025, including real-time kinematic global navigation satellite system surveys, water-level gauge records, and uncrewed aerial vehicles measurements, we demonstrate that SWOT reliably captures variations in water level of high-mountain lakes (average r = 0.72, RMSE = 0.29 m; peak r = 0.94, p < 0.01, RMSE = 0.09 m). Further analysis indicates that water level errors are primarily associated with KaRIn retrieval degradations—including phase-unwrapping errors, low Signal-to-Noise Ratio, and a lack of coherent gain—rather than with terrain-induced layover effects. Across all PIXC pixels over the selected lakes, the retrieved elevations capture WSE accurately with a negligible bias (average CDF 50% signed error = 0.03 m). More importantly, SWOT’s swath altimetry demonstrates the observational potential to retrieve WSE for more than 2,924 glacial lakes across the Himalaya and for over 24,000 Tibetan Plateau lakes smaller than 1 km 2 . These findings provide practical guidance for future SWOT hydrologic data processing and applications, and underscore SWOT’s transformative potential for monitoring global small water bodies and advancing cryospheric research.