Modeling the basics of river-bar formation, and model evaluation

A. B. Murray and C. Paola

Researchers have modeled river bars by solving equations for flow and sediment transport in as much detail, and at as small a scale, as possible. However, a cellular model that treats only simple, large-scale interactions between water and sediment also produces bars. This result suggests that the detailed processes may not be essential in explaining the basic phenomenon of bar growth. In the model, the most essential requirement for bar formation is a non-linear relationship between flow strength and sediment flux, with an exponent > 1. In this case, flow divergence causes deposition, and with most combinations of the parameters involved, leads to bar formation. Similarly, flow convergence leads to erosion. We have not attempted to produce bars with completely realistic shapes. The model is designed to simulate the behavior and large-scale features of braided rivers. We are evaluating the model by comparing typical sequences of flow convergence and divergence, i.e. areas of erosion and bar formation, respectively, in the model with those in natural real braided rivers. To do this, we have developed a method for measuring the difference between state-space plots representing these sequences.