Computational Models of Flow, Sediment Transport and Morphodynamics in Rivers

Abstract

Numerical models have become fundamental tools to advance the scientific understanding of fluvial systems for a wide range of spatial and temporal scales. Despite recent advances in computational power and the development of advanced numerical methods, predicting the complex interactions among flow, sediment transport, and bed morphodynamics is still a significant challenge for models aimed at simulation and/or gaining new insights of physical processes in nature. The combination of advances in computing power and algorithms with the realization that complex dynamics can sometimes arise from surprisingly simple models has led to a wide spectrum of models from reduced‐complexity models that focus on self‐organization and formation of morphological patterns, to highly resolved direct numerical simulations that capture all the turbulent scales and the detailed dynamics of sediment particles. In between are morphodynamic models that, while drastically simplifying flow and sediment transport, remain anchored in classical PDE based modeling techniques, and models of fluid flow that represent the turbulence in some type of Reynolds‐averaged form. The rich range of approaches to modeling river dynamics raises the question of how to choose the best one. There is no recipe for this, but behaviors such as scale independence hint that the large‐scale pattern dynamics may not be sensitive to the details of the underlying dynamics; such cases, exemplified by dendritic erosional channel networks, should be amenable to reduced‐complexity modeling. At the other extreme are cases, including certain aspects of bedform behavior, which are mediated by large turbulent coherent structures that originate in small‐scale interactions near the bed and are thus hard to capture in simplified models. Finally, we suggest that there is a bright future in developing new ways to couple model approaches across this spectrum.