Research on bifurcations in rivers

Abstract

One of the unsolved problems in the water resource engineering is the morphological behaviour of bifurcations in rivers. Bifurcations can be found in deltas, in estuaries and in braided rivers. When modelling river reaches which contain a bifurcation, a nodal point relation is needed. The problem is to determine this nodal point relation. The distribution of sediment over the two branches is determined by local three-dimensional phenomena and it has to be specified explicitly at every bifurcation that is modelled. Nowadays nodal point relations are used, which are not based on thorough (experimental) research. In this report the search on the distribution of the sediment as a function of the discharge distribution is described. An experimental model of a bifurcation in a river is designed and constructed. This experimental model is used to do experiments which lead to a better insight in the behaviour of the bifurcation. Before starting the experiments, all parts of the test rig were tested. Then several experiments have been designed and carried out with the test rig. The measurement errors, made during the experiments, are described, which gives a good view of the quality of the experiments. Since it was time-consuming to obtain data from the experiments, a thorough statistical analysis of the found data is carried out. Several statistical techniques were used to obtain as much information as possible from the data. With the use of this statistical analysis, nodal point relations were found for the specific types of bifurcations in the test rig, for three different upstream discharges and two different shapes of the bifurcation. It appeared that the general nodal-point relation proposed by Wang et al (1993) was appropriate. The unknown parameters of this relation were found for the circumstances of the experiments; Three relations are found for the first shape of the bifurcation, with a respective upstream discharge of 20 1/s, 30 1/s and 40 1/s. Two relations are found for the second shape of the bifurcation, with a respective upstream discharge of 30 1/s and 40 1/s. It is statisticaly proven that some of the coefficients in these relations are comparable for the different circumstances, but others are not. In order to see whether it is possible to carry out numerical simulations of a bifurcating river, the configuration of the experimental model and the found nodal point relations were used as input for simulations with WENDY. Thus, simulations were carried out of some of the experiments. It appeared that the results of these simulations were comparable to the measured data from the experiments. Therefore it is proven that if a good nodal-point relation is known, a good simulation of a bifurcating river can be carried out. Recommendations are given how to continue this research project in the future. The relations that are found here are only valueable for the given circumstances of the experiments. The final goal, however, of this research is to get a relation of the distribution of the sediment over the two downstream branches which can be used in all circumstances. More research has to be done to obtain this.