Integrating Suspended Sediment Flux in Large Alluvial River Channels: Application of a Synoptic Rouse‐Based Model to the Irrawaddy and Salween Rivers. Issue 9 (18th September 2020)

• Record Type: Journal Article
• Authors: Baronas, J. Jotautas; Stevenson, Emily I.; Hackney, Christopher R.; Darby, Stephen E.; Bickle, Michael J.; Hilton, Robert G.; Larkin, Christina S.; Parsons, Daniel R.; Myo Khaing, Aung; Tipper, Edward T.
• Is Part Of: Journal of geophysical research. Volume 125:Issue 9(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-09-18
• Subjects: sediment transport -- hydrodynamic sorting -- megarivers -- sediment flux -- particulate organic carbon -- POC flux; Geomorphology -- Periodicals; 551.3
• DOI: 10.1029/2020JF005554
• Languages: English
• ISSNs: 2169-9003

Abstract

Abstract: A large portion of freshwater and sediment is exported to the ocean by a small number of major rivers. Many of these megarivers are subject to substantial anthropogenic pressures, which are having a major impact on water and sediment delivery to deltaic ecosystems. Due to hydrodynamic sorting, sediment grain size and composition vary strongly with depth and across the channel in large rivers, complicating flux quantification. To account for this, we modified a semi‐empirical Rouse model, synoptically predicting sediment concentration, grain‐size distribution, and organic carbon (%OC) concentration with depth and across the river channel. Using suspended sediment depth samples and flow velocity data, we applied this model to calculate sediment fluxes of the Irrawaddy (Ayeyarwady) and the Salween (Thanlwin), the last two free‐flowing megarivers in Southeast Asia. Deriving sediment‐discharge rating curves, we calculated an annual sediment flux of 32 6 − 70 + 91 Mt/year for the Irrawaddy and 15 9 − 51 + 78 Mt/year for the Salween, together exporting 46% as much sediment as the Ganges‐Brahmaputra system. The mean flux‐weighted sediment exported by the Irrawaddy is significantly coarser ( D 84 = 193 ± 13 μm) and OC‐poorer (0.29 ± 0.08 wt%) compared to the Salween (112 ± 27 μm and 0.59 ± 0.16 wt%, respectively). Both rivers export similar amounts of particulate organic carbon, with a total of 1 . 9 − 0 . 9 + 1 . 4 Mt C/year, 53% as much as the Ganges‐Brahmaputra. These results underline the global significance of the Irrawaddy and Salween rivers and warrant continued monitoring of their sediment flux, given the increasing anthropogenic pressures on these river basins. Plain Language Summary: The sediment (clay, silt, and sand) carried by rivers is a crucial but dwindling resource, sustaining agriculture in fertile deltas, while huge amounts of sand particularly are used to produce concrete, glass, and electronics. The amount of sediment that rivers carry globally is, however, not well known. It is especially difficult to measure in large rivers because most sand is carried near the channel bottom, tens of meters beneath the surface. In this study, we present an improved approach to measure the amount of sediment carried by large rivers. It combines sediment samples collected at various depths in the river with measurements of river flow via acoustic sensors. We apply this method to some of the world's largest rivers ‐ the Irrawaddy (Ayeyarwady) and the Salween (Thanlwin) in Myanmar, which have been understudied for decades. Our results show that they both currently discharge immense quantities of sediment to the ocean. However, this is likely to decrease drastically in the coming decades, given the projected industrialization and future damming of these two basins. The results presented in this study thus provide an important baseline against which to measure future changes in sediment discharge by these rivers. Key Points: An updated empirical Rouse modeling framework to calculate sediment flux and composition of large alluvial rivers is presented Model was applied to compute annual sediment flux of Irrawaddy and Salween rivers as 326 − 70 + 91 and 159 − 51 + 78 Mt/year, respectively Fluxes calculated using simple means of depth point samples result in errors of up to 50% relative to Rouse‐based model