Exploring the Effect of River Spatial Variability on SWOT Discharge Accuracy
Authors:
- Yushan ZhangYZYushan Zhang
- Michael DurandMDMichael Durand
- Stephen Paul CossSCStephen Paul Coss
- Renato Prata de Moraes FrassonRFRenato Prata de Moraes Frasson
- Elizabeth M. PriorEPElizabeth M. Prior
- Colin J. GleasonCGColin J. Gleason
Preprint
- 7 February 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
The Surface Water and Ocean Topography (SWOT) mission measures river width, water surface elevation (WSE), and slope for discharge estimation. In this study, we explore the impact on remotely sensed discharge of within-reach river spatial variability, defined as the coefficient of variation (εp = σ/μ). After quality control, we analyze 484 representative reaches across the United States with paired SWOT observations and United States Geological Survey (USGS) discharge records from January 2023 to October 2024. Discharge is calibrated using modified Manning equations and evaluated using the Nash-Sutcliffe efficiency (NSE), absolute normalized residuals and correlation coefficient. Reaches with high spatial variability exhibit substantially lower calibration skill than reaches with low variability, with the median NSE reduced by 0.09 (0.78 vs. 0.69). Two possible reasons for this phenomenon are examined: SWOT observation errors and structural limitations of modified Manning’s equations. Regarding observation errors, we apply a river hypsometry constraint to correct width and WSE. Overall calibration accuracy improved, and the median NSE gap between low- and high-variability reaches reduced to 0.05 (0.86 vs. 0.81), emphasizing the significance of observation errors in this degradation. Conversely, we find no direct evidence that degradation arises from structural errors. We confirm a key prediction of spatial averaging theory—that effective resistance changes more for reaches with higher spatial variability—but adding an explicit slope variability term to flow laws does not lead to systematic improvement. This research has important implications for better estimating discharge from SWOT data.Keywords
This publication has 38 references indexed in Scilit:
- Discharge Estimation in Ungauged Basins Through Variational Data Assimilation: The Potential of the SWOT MissionWater Resources Research, 2018
- BAM: Bayesian AMHG‐Manning Inference of Discharge Using Remotely Sensed Stream Width, Slope, and HeightWater Resources Research, 2017
- Skin-friction drag reduction in a channel flow with streamwise-aligned plasma actuatorsInternational Journal of Heat and Fluid Flow, 2017
- The SWOT Mission and Its Capabilities for Land Hydrology2016
- Improved error estimates of a discharge algorithm for remotely sensed river measurements: Test cases onSacramento andGaronneRiversWater Resources Research, 2015
- Spatiotemporal interpolation of discharge across a river network by using synthetic SWOT satellite dataWater Resources Research, 2014
- Data assimilation methods in the Earth sciencesAdvances in Water Resources, 2008
- Variability and uncertainty in reach bankfull hydraulic geometryJournal of Hydrology, 2008
- Estimating regression models with unknown break‐pointsStatistics in Medicine, 2003
- Changing ideas in hydrology — The case of physically-based modelsJournal of Hydrology, 1989