Navigability at an unstable bifurcation: The Montaño - Murindó bifurcation of the Atrato River in Colombia

The Atrato River in Colombia should be navigable for bigger vessels when generating a corridor which should link the Atlantic Ocean with the Pacific Ocean. A likely segment of the river that forms a critical point for navigation is the point of bifurcation where the river is divided in the Montaño and Murindó branch. Both branches should be reasonably navigable, but bifurcation morphodynamics possibly render the water depth in one branch smaller than the draught of the vessel. It is noted that the Montaño branch is considered the main corridor, as this branch is shorter and generally wider. The main objective is to model and understand the morphology that evolves around the river bifurcation and discuss possible measures for creating a sustainable navigable channel. In order to create insights in these morphodynamics, a computational model has been set up. Using the software package Delft3D, a 2DH morphodynamic model of the area around the bifurcation has been made. Due to the lack of sufficient data available, a thorough morphodynamic calibration and verification is impossible. It was shown that the model predicts a development to a stable, but often highly asymmetrical discharge division. The model attains this equilibrium after 50 years of morphological modelling. This suggests the bifurcation is unstable. The evolution to an asymmetrical discharge division is mainly governed by the imposed downstream boundary water levels because a head difference over the branches determines the division of flow over the branches. The calibrated model assumes no backwater effects, which means that the shorter Montaño branch (83 km) has a larger water level gradient than the Murindó branch (91.5 km). It was shown that the unknown head difference over the branches greatly influences the global development of the branches (i.e. aggradation /degradation). Because only one bathymetrical survey was executed, erosion and sedimentation rates are unknown, which further decreases the possibility of accurately modelling the area. An important result is that when imposing the same water level gradients in both branches, the model still predicts a development to a highly asymmetrical division of flow discharge (i.e. 75% through the Montaño branch). This means that local geometrical phenomena influence the division of flow and sediments. The sharp angle of offtake of the Murindó branch results in flow separation which in turn causes a sediment trap at the entrance of this branch. As a result, the cross-section reduces significantly, which means less discharge is directed through this branch. Another phenomenon is caused by the present river bend just upstream of the point of bifurcation. This leads to a gravitational pull along the transverse slope which initially leads to an increasing bed load transport through the Murindó branch. Another local phenomenon is the Bulle effect which leads more bed load through the Murindó branch due to the river bend formed by the upstream bend and the offtaking channel. Discussing possible solutions is sometimes hypothetical as the global evolution is sensitive to the unknown downstream boundaries. The possibility of two different bend cut-offs are discussed. It was shown that the angle of offtake greatly influences the division of flow discharge and sediments. Decreasing the offtake-angle of the Murindó branch from 90° to 30° means the Murindó branch gets dominant (60% flow discharge through this branch instead of 20%). As the internal connection between the branches can result in significant backwater effects, it can be a simple solution to cut off this connection. In conclusion, the present available data prove to be insufficient. Measuring stage and discharges in both branches, river bed profiles and sediment characteristics can indicate which branch will be dominant and enable future decision making.River EngineeringHydraulic EngineeringCivil Engineering and Geoscience