In this paper, we analyze the performance of several noncohesive bank erosion algorithms to be embedded into two‐dimensional models for river hydromorphodynamics on nonmoving meshes. To avoid the complexity of analyzing two‐dimensional model results arising from the nonlinear interaction between flow, fluvial transport, and bank erosion, we develop a simplified framework. In detail, we reduce the two‐dimensional morphodynamic model to a cross‐sectional model under the assumption of longitudinal morphodynamic equilibrium, and apply bank erosion algorithms therein. To build candidate bank erosion models, we break down bank erosion algorithms into three modeling steps: identification of the bank, computation of sediment fluxes due to bank erosion, and bank updating. Different potential models are created by choosing different options for each step. We assess model performance against surveyed bank erosion over a flood event in the transitional Selwyn River, New Zealand. This study is preliminary to implementation of bank erosion in a fully two‐dimensional setting, to model braided planform dynamics.
10 October 2020 in Earth Surface Processes and Landforms
In this paper, we utilize a numerical morphological simulation approach to study braiding channel dynamics under steady flow. We conduct 11 different runs in a laboratory configuration, with each initial condition featuring a unique small‐amplitude perturbation pattern of the bed topography. From these infinitesimal initial differences, braided channels evolve into patterns that are macroscopically different and unique to each run, thus showing sensitive dependence on the initial condition and hence a chaotic behavior. Leveraging the analogy of braiding and fluid turbulence put forward by Paola (1996), by introducing a Reynolds‐type decomposition of bed elevations into a reference and a fluctuating component we characterize braided channel dynamics through statistics of bed fluctuations. We observe that, over the simulation time frame, braiding is not statistically homogeneous, but is stationary in a spatially averaged sense. We prove that braiding is anisotropic and extract two distinct length scales associated with the correlations of bed fluctuations in the streamwise and cross‐channel direction. We observe that, under the time averaging window that yields stationary braiding, braiding shows hints of ergodic behavior, as the time statistics of one run converge to the ensemble statistics computed across all the other runs.
We use two‐dimensional physics‐based numerical modeling to study multi‐decadal evolution of vegetation and morphology under different flow regimes in real‐world gravel‐bed braided rivers. To assess model realism, we focus on two rivers in Canterbury (New Zealand) that, despite having been subjected to the introduction of similar invasive vegetation species in the last ∼100 years, show very different trajectories of vegetation presence due to their different flow regimes: the Lower Waitaki River and the Waimakariri River. The former, featuring a naturally damped flow regime—and having experienced further artificial flow damping due to hydropower generation from the 1930s, experienced vegetation encroachment; while the latter, featuring a flashy flow regime, retains an unvegetated braided planform. We propose an innovative calibration and validation procedure to determine an optimal setup of vegetation parameters that allows the model to robustly reproduce the trajectories of both rivers, thereby proving that the model responds sensibly to different hydrological conditions. Then, we isolate the impact of hydropower‐related flow regime modifications on vegetation encroachment in the Lower Waitaki by running the calibrated model with a natural flow regime that does not feature the effect of hydropower generation, and find that vegetation encroachment would have happened even without flow alteration, albeit to a milder degree. Finally, we apply to simulation results a conceptual framework based on a synthetic parameter that compares the relative strength of hydrological and vegetation controls, and discuss the use of this parameter as a predictor of vegetation presence across flow regimes.