[A-CG45] Water and sediment dynamics from land to coastal zones

[ACG45-12] Spatially dense estimation of satellite-based river discharge along the mainstem of the Yellow River

*Yuki Ishikawa^1,2, Dai Yamazaki², Yuting Yang¹ (1.Department of Hydraulic Engineering, School of Civil Engineering, Tsinghua University, 2.Institute of Industrial Science, The University of Tokyo)

Keywords:Remote sensing, Satellite-based river discharge, Hydraulic geometry, Water resources monitoring

As 63% of human surface water use comes from rivers, monitoring changes in river discharge is crucial for stable and effective management of water. In particular, the identification of the specific spot of a river where water flows into and out of is useful for such management. It is difficult, however, to monitor discharge changes continuously along a river channel by in-situ stream gauges since they are scattered and distributed unevenly. Recently, the advancement in remote sensing technique and the development of the Mass-conserved Flow Law Inversion (McFLI) algorithms have enabled us to estimate discharge solely from satellite images. The validity of the algorithms was verified in previous researches, yet it remains to be elucidated whether and to what extent the McFLI method can describe the spatially continuous discharge changes along a river. In this study, we examined the applicability of the McFLI algorithm to the Yellow River, which is heavily consumed for agricultural activities and faces a serious water resource management problem, and attempted a spatially denser estimation of discharge along the mainstem of the river. By using in-situ discharge data and satellite-based river width at 16 gauging stations where no levees or diversion channels exist on both sides of the bank, log-linear regression analysis was conducted. We observed the at-many-stations hydraulic geometry (AMHG) at 9 stations in the upper reaches and 7 stations in the middle reaches, which indicates the applicability of the McFLI algorithm. Then, the Bayesian AMHG + Manning (BAM) algorithm, a novel McFLI algorithm combining empirical and physical flow laws, was employed to estimate river discharge along two river sections in the middle reaches. Each section has 101 and 140 sub-reaches respectively, and satellite-based discharge was obtained in each sub-reach. Although errors compared with in-situ data were more or less observed, the spatial changes in discharge along the river channel were successfully captured: the discharge increased at the confluence point and decreased in the sub-reaches where irrigated areas exist nearby. The approach in this study is applicable to areas with no or sparse stream gauges, and is expected to be used for monitoring rivers on a global scale.