Inversion of river discharge from remotely sensed river widths: A critical assessment at three-thousand global river gauges

Peirong Lin, Dongmei Feng, Colin J. Gleason, Ming Pan, Craig B. Brinkerhoff, Xiao Yang, Hylke Beck, Renato Prata de Moraes Frasson

Research output: Contribution to journalArticlepeer-review

5 Scopus citations


Accurately estimating river discharge from satellite-derived river hydraulic variables (e.g., width, height, and slope) is the overarching goal of the remote sensing of discharge (RSQ) community. Numerous past studies have developed and intercompared different RSQ algorithms to demonstrate their feasibility, yet relatively few have focused on assessing how the RSQ algorithms are adapted to a wide range of rivers globally. As the community is now ready to expand to the global scale given advances in computing power, sensors, and the launch of the Surface Water and Ocean Topography (SWOT) satellite mission, a much broader geographic view of RSQ accuracy should be prioritized toward “better generalizability” instead of “higher accuracy at limited places”. To help close this gap, we extracted multi-temporal river widths from >350 K Landsat scenes at >3 K river reaches globally, and used them to estimate discharge using the Bayesian AMHG-Manning (BAM) algorithm and the geomorphologically-enhanced variant (geoBAM). We use this framework to demonstrate how to apply an ‘off the shelf’ RSQ algorithm and test it globally without methodological intervention and answer: does it live up to its promise? Our daily discharge inversions (1984–2019) showed positive Kling-Gupta Efficiency (KGE) at 27% of the gauges for BAM and 39% for geoBAM, and this percentage increased to 46–65% after feeding richer priors on flow seasonality and monthly variability, amounting to 1400–2000 successful inversions. Exploratory analyses showed that the inversion is the most sensitive to a channel geomorphological parameter b and climate aridity, where the optimal conditions are high-b, humid environments, as well as moderate width variability, leaf area index (LAI), and river width. Although specific to BAM/geoBAM, constraining the factors to their optimal ranges led to a median KGE of 0.33 for >600 gauges, which highlights the promising potential for global RSQ. We further discussed the optimal configuration of the RS/priors by analyzing results derived from different information content in priors. Overall, our critical assessment of BAM/geoBAM reveals a successful global implementation of an existing algorithm that SWOT will improve. We suggest similar large-scale assessments for other RSQs be prioritized to identify the emerging challenges as we move into a new era with global river monitoring capability from space.
Original languageEnglish (US)
Pages (from-to)113489
StatePublished - Feb 4 2023

Bibliographical note

KAUST Repository Item: Exported on 2023-03-21
Acknowledgements: This study is supported by the Open Research Program of the International Research Center of Big Data for Sustainable Development Goals, Grant No. CBAS2022ORP05, the National Key Research and Development Program (2022YFF0801303), and the startup fund from Peking University on “Numerical modeling and remote sensing of global river discharge” (#7100603564). We thank Prof. Eric F. Wood (deceased in November 2021) who offered valuable input to this work at its early stage. Renato Frasson is supported by NASA SWOT Science Team funding (80NSSC20K1339). Colin Gleason is supported by NASA SWOT Science Team funding (80NSSC20K1141) and NSF CAREER funding (VGJHK59NMPK9). Dongmei Feng is supported by NASA Terrestrial Hydrology Program funding (80NSSC22K0987). Craig Brinkerhoff is supported by a NASA FINESST Fellowship (80NSSC21K1591). We would like to thank the Global Runoff Data Centre (GRDC) for providing river discharge data. A portion of this work was performed at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration (80NM0018D0004). Three anonymous reviewers and the editors are thanked for offering constructive comments to this manuscript. Author contribution: PL conceived the study, conducted the main analyses, produced the figures, and wrote the initial draft; DF assisted with RivWidthCloud data collection and processing; CJG assisted with the BAM/geoBAM experimental setup; CJG, MP and CB assisted with part of the analyses; XY, HEB, RF provided discussions and technical support; all authors contributed to the editing and revisions of the manuscript.


Dive into the research topics of 'Inversion of river discharge from remotely sensed river widths: A critical assessment at three-thousand global river gauges'. Together they form a unique fingerprint.

Cite this