Author(s): Siyoon Kwon; Il Won Seo
Linked Author(s): Il Won Seo, Siyoon Kwon
Keywords: Mixing; River confluence; Remote sensing; Hyperspectral imagery; SSC
Abstract: The mechanism of suspended sediment mixing at river confluences is challenging to interpret due to the complex flow conditions generated by the main river and tributary river. Additionally, the suspended sediment in the main and tributary rivers has different origins, leading to diverse mineral characteristics and particle distributions after confluence. Therefore, precise measurements of suspended sediment in confluences require high-resolution spatial data, but conventional measurement methods rely on labor-intensive and costly direct measurements, predominantly using numerical modeling for analysis. To overcome the limitations of these direct methods, recent advancements in remote sensing technologies have led to the development of suspended sediment concentration (SSC) measurement techniques. However, the accuracy of optical-based measurements has been hampered by the spectral variability caused by complex nature of water bodies and suspended sediments. The cluster-based machine learning regression with optical variability (CMR-OV) method proposed by Kwon et al. (2022) could account for this spectral variability, demonstrating high accuracy under various water and suspended sediment conditions. In this study, CMR-OV method was used for analyzing suspended sediment mixing in river confluences. Thanks to the detailed concentration map from CMR-OV, we could estimate the continuous variance of lateral concentration distributions along longitudinal distance, from which precise and continuous mixing patterns were extracted.
Year: 2024