David Bjerklie (U.S. Geological Survey, United States); Jack Eggleston (U.S. Geological Survey, United States); Dongmei Feng (University of Cincinnati, United States); Ryan Riggs (U.S. Geological Survey, United States); Cécile Cazals (CS Group France, France); Robert Dudley (U.S. Geological Survey, United States); Elisa Friedmann (University of Massachusetts Amherst, United States); Megan Garner (Environment and Climate Change Canada, Canada); Colin Gleason (University of Massachusetts Amherst, United States); Yuta Ishitsuka (University of Massachusetts Amherst, Japan); James Leach (Environment and Climate Change Canada, Canada); J. Toby Minear (Cooperative Institute for Research in Environmental Sciences, University of Colorado Boulder, United States); Hind Oubanas (INRAE, France); Tamlin Pavelsky (University of North Carolina Chapel Hill, United States); Taylor Rowley (University of Massachusetts Amherst, United States); Cassie Stuurman (Jet Propulsion Laboratory, California Institute of Technology, United States); Catalina Tagliatela (Jet Propulsion Laboratory, United States); Jacob Zwart (U.S. Geological Survey, United States)

The Surface Water and Ocean Topography (SWOT) mission enables transformative approaches to water resources management by offering high-resolution water surface elevation data and extents from novel satellite radar technology. This study presents the first large-scale assessment of SWOT water surface elevation data across the United States and Canada, evaluating elevation accuracy against in situ gage data. We compare SWOT Level 2 Hydrology data products (Version C) to in situ timeseries of water elevation using 421,677 paired observations across 2,143 river and 659 lake and reservoir gages. Data accuracy and temporal coverage of SWOT observations are explored as functions of various data quality filters. Paired measurements (median-normalized) of SWOT and in situ observations resulted in 68th percentile absolute elevation differences of 18.32 cm, 17.08 cm, and 7.79 cm for river reach, river node, and lake data, respectively, with an overall combined difference of 16.97 cm for filtered observations (79,909; 19% of total potential data volume). Here we emphasize the importance of applying quality filters (both self-reported and exogenous) to SWOT Version C data, with 68th percentile absolute elevation differences ranging from 61.23 cm to 75.93 cm for the full dataset. Filtering also improves the correlation statistics between SWOT and in situ water surface elevation data, increasing R2 values from 0.04 to 0.83. Overall, we find that SWOT river and lake elevations meet SWOT data accuracy requirements (elevation accuracies within 10 and 25 cm) with adherence to SWOT quality flags.