To maintain a navigable channel and improve high-flow conveyance, engineers have recently proposed constructing longitudinal training walls as an alternative to the traditional transverse groynes. However, previous work has shown that the system of parallel channels created by a longitudinal training wall might be unstable in rivers with alternate bars. Many questions remain unanswered, particularly whether a stable system can be obtained by carefully designing the bifurcation point. This work analyses the stability of the bifurcating system created by a thin longitudinal wall in sand-bed rivers with alternate bars or point bars. The methodology includes performing 102 numerical tests using the Delft3D code to reproduce an idealized low-land river, either straight or meandering. The results show that the system of parallel channels separated by a training wall may indeed become unstable. An important factor is found to be the location of the bifurcation point with respect to a neighboring bar or point bar. The same trends are observed for both constant and variable discharge, in straight and meandering channels. The results suggest that cyclic growth and decline of the bifurcating channels may arise as inherent system behavior, without the need of any additional external forcing. We explain this from changes in the relationship between sediment transport ratio and discharge ratio as the bifurcation evolves. This cyclic behavior can be regarded as a form of system stability and can be obtained by carefully placing the starting point of the longitudinal training wall, and thus the bifurcation point, near the top of a bar.@en

Rivers have been trained for centuries by series of transverse groynes. This generally results in damages to their ecosystems as well as in undesirable longterm morphological developments. We analyze here the possibility to train rivers in a new way by subdividing their channel in parallel channels with specific functions with longitudinal training walls. In most cases, the goal is that of obtaining one deep, regular navigation channel and one shallower channel that is able to preserve some ecological functions of the river and to contribute to convey high flow discharges. The effectiveness of longitudinal training walls in achieving this goal and their longterm effects on the river morphology have not been thoroughly investigated yet. In particular, studies that assess the stability of the parallel channels separated by the training wall are still lacking. This work studies the long-term morphological developments of river channels subdivided by one or two longitudinal walls, focusing on low-land rivers. These rivers are normally characterized by the presence of steady alternate bars, or point bars inside their bends. For this reason, the presence of these large deposits is taken into account. This is the first study dealing with the combined effects of bars and longitudinal walls. The methodology comprises both laboratory experiments and numerical simulations...@en

The Dutch River Waal, a branch of the Rhine, has been trained for centuries to mitigate the effects of ice-jams and improve navigation. The works, started in 1850, involved river straightening and narrowing by a series of transverse groynes. Besides fulfilling their goal, the groynes also created the need to raise flood protection works and caused undesirable channel incision. This study assesses the effectiveness of training the river with a longitudinal wall instead of with groynes. The investigation analyzes the long-term response of the historical river with a two-dimensional depth-averaged (2DH) morphodynamic model. The results show that the wall would create two parallel channels, one becoming deeper and the other one shallower. The former would be as suitable for navigation as an equally-wide channel obtained with groynes. The latter would contribute in conveying water during high flow events and improve the river ecology. Training the river with a wall would also lessen channel incision. The best performance is obtained if the wall is built on the channel centerline, starting just upstream of a point bar top.@en

Rivers have been trained for centuries by channel narrowing and straightening. This caused important damages to their ecosystems, particularly around the bank areas. We analyse here the possibility to train rivers in a new way by subdividing their channel in main and ecological channel with a longitudinal training wall. The effectiveness of longitudinal training walls in achieving this goal and their long-term effects on the river morphology have not been thoroughly investigated yet. In particular, studies that assess the stability of the two parallel channels separated by the training wall are still lacking. This work studies the long-term morphological developments of river channels subdivided by a longitudinal training wall in the presence of steady alternate bars. This type of bars, common in alluvial rivers, alters the flow field and the sediment transport direction and might affect the stability of the bifurcating system. The work comprises both laboratory experiments and numerical simulations (Delft3D). The results show that a system of parallel channels divided by a longitudinal training wall has the tendency to become unstable. An important factor is found to be the location of the upstream termination of the longitudinal wall with respect to a neighboring steady bar. The relative widths of the two parallel channels separated by the wall and variable discharge do not substantially change the final evolution of the system.@en

Recent years have seen a marked increase in the availability of morphodynamic models and a proliferation of new morphodynamic codes. As a consequence, morphodynamic models are increasingly developed, used and evaluated by non-experts, possibly leading to mistakes. This paper draws attention to five types of common mistakes. First, new morphodynamic codes are developed as extensions of existing hydrodynamic codes without including all essential physical processes. Second, model inputs are specified in a way that imposes morphodynamic patterns beforehand rather than letting them evolve freely. Third, detailed processes are parameterized inadequately for application to larger spatial and temporal scales. Fourth, physical and numerical phenomena are confused when interpreting model results. Fifth, the selection of modeling approaches is driven by the belief that complete data are a prerequisite for modeling and that the application of 2D and 3D models requires more data than the application of 1D models. Examples from fluvial morphodynamics are presented to illustrate these mistakes.@en

Recent developments in river training include the replacement of series of transverse groynes by a longitudinal training wall. The idea is to obtain higher flood conveyance capacity during peak discharges and maintain the navigation channel at low flows. This new type of interventions creates a (near-bank) side channel parallel to the main (navigation) channel, splitting water discharge and sediment transport between the two channels. However, it is not clear yet whether the two-channel system is stable, i.e. both channels tend to remain open, and whether the intervention will meet the objectives, particularly considering the long term morphological adaptations. This study examines the stability of two-channel systems created by longitudinal training walls in low-land rivers with a bed topography characterized by alternate bars. Based on the results of preliminary numerical simulations, showing high risk of system instability with the closure of one of the two channels, this study concentrates on experimental tests. The work is on-going and the preliminary results suggest that the position of a training wall related to a near-bank bar plays an important role on the stability of the system. Further work will be carried out with more extensive experimental investigations to carefully verify the long-term effects of a longitudinal training wall in a river with alternate bars, starting at different locations related to a bar.@en

Parallel channels created by a longitudinal training wall in a river with bars tend to be unstable. The evolution of the bifurcation can be predicted based on the location of the starting point of the wall with respect to the nearest bar.@en

Parallel channels created by a longitudinal training wall in a river with bars tend to be unstable. The evolution of the bifurcation can be predicted based on the location of the starting point of the wall with respect to the nearest bar.@en