An investigation of braided river dynamics using a new numerical modelling approach
Cudden, Judith Rose
PublisherUniversity of Glasgow
MetadataShow full item record
Braided Cascade has been developed from Cascade (Braun and Sambridge, (1997)), a long-term (dt = 100 years) numerical model that simulates long-term landscape evolution. Herein it has been modified and applied to relatively short term process modelling of the evolution of complex river topography, discharge and sediment load of braided rivers. Braided Cascade is synthesist in spirit, there is no detailed hydrodynamic component to the model, a realistic simplification at the time scales considered. The major advantage of the model is the incorporation of an irregular time-varying grid using a triangulated irregular network (TIN) to represent a terrain surface. Advantages of using TINs include the ability to solve problems with non-rectangular geometeries and/or boundary conditions and the ability of river segments to form in all directions. The model routes water from node to node based on the local topographic slope. Sediment transport depends on the local stream power. Nodal elevation changes after each iteration according to the difference between the amount of sediment entering and leaving the node. Model output includes spatial and temporal (at one point) water discharge, bedload sediment transport, as well as maps of channel networks, erosion and deposition throughout the reach. Sensitivity analysis indicated that the most significant parameters for braiding are erosion length scale, splitting ratios and the allowance of the model to deposit sediment. Therefore an imbalance in the amount of sediment the river is carrying and the carrying capacity AND a reworking of the deposits is needed for a braided network to form. Sediment output from model runs indicate that the similarities between model data and other data sets are weak and all runs tended to reach static equilibrium. Braided Cascade therefore failed to adequately reproduce realistic data sets. It was found that the differences between model results and the flume data indicate that the model does not always match the physical systems as closely as physical systems match each other.