[A-CG37] Water and Sediment Dynamics from Land to Oceans [En]

[ACG37-06] Enhancing Flow Simulations in Mega River Deltas by Upgrading the Bifurcation Scheme in the CaMa-Flood Model

*Dung Trung Vu¹, Dai Yamazaki¹ (1.Institute of Industrial Science, The University of Tokyo)

Keywords:Large-scale hydrodynamic models, River bifurcation, Mega river deltas

Mega river deltas, located at the most downstream of major rivers globally (e. g., Amazon and Mekong) are the habitat of about 500 million people; yet they are highly vulnerable to flood hazards due to their low-lying and flat terrain. Additionally, the frequency and severity of floods in these regions are forecast to increase under the impacts of climate change and sea level rise. Consequently, developing and improving hydrodynamic models are crucial for effective flood forecast and risk assessment in mega river deltas. However, modeling flow dynamics in these deltas is a challenging task due to their complex river network, often characterized by channel bifurcations—where rivers split into multiple channels. This phenomenon is rarely captured in large-scale hydrodynamic models because, for several decades, the river network maps used in such models have relied on the assumption that each river pixel has only one downstream flow directory. Previous studies have attempted to overcome the challenge by introducing approaches for representing bifurcation channels in hydrodynamic models; but the approaches are applicable to single deltas only or still require manual processing, that hinders them from being applied to large-scare simulations for multiple basins and deltas. Based on the previous development of the representation of bifurcation channels in the CaMa-Flood model (a global hydrodynamic model), in this study, we further improve its computational scheme. Specifically, we use an empirical equation to estimate the parameters of bifurcation channels from their average annual discharge. Doing so, we introduce an automatic approach for representing bifurcation channels in the CaMa-Flood model and estimating the channel parameters. The results in this study show that our improved bifurcation scheme helps to enhance the simulated flows in both bifurcation and mainstem channels. Our results also reveal an interesting interaction between simulated flows in mainstem and bifurcation channels when river parameters get improved after each simulation iteration. The impact of our improved bifurcation scheme on simulated flows varies depending on the size of the deltas and the complexity of their river network. Finally, our improved flow dynamic simulations provide a promising basis for more accurate flood simulations and forecast in mega river deltas.