UBC Theses and Dissertations
Episodic processes in fluvial landscape evolution Dadson, Simon
The fractal dimension provides a succinct summary of topographic intricacy and a measure of landscape roughness. The work presented in this thesis is an attempt to examine the spatial variation of fractal properties of topography. It is hypothesized that distinct zones, or domains, may be identified in the landscape, within which the fractal dimension is constant, owing to the dominance of a characteristic geomorphic process. The variogram method (structure-function analysis) is used to measure the fractal properties of digital elevation models of the Queen Charlotte Islands and small tributaries of the Capilano River (in the Southern Coast Mountains). Process domains are identified using upslope-area criteria, and are shown to have distinct fractal dimensions. In the Capilano basin, the fractal dimension is found to be higher on upper slopes; whereas on the Queen Charlotte Islands, it is higher on lower slopes. This difference is explained with reference to the details of specific surficial processes operating in each of the areas, and to their different glacial histories. A multiple-process landscape evolution model is presented to explore the link between geomorphic process and fractal form. The model is carefully calibrated using field data from several recent surveys within the study areas. Non-linear diffusion is used to represent surficial slope failure; fluvial processes are represented using a simplified Bagnold equation; and deep-seated rock slides are introduced at random on the steepest, highest slopes. The model results confirm the smoothing effect of diffusion, which lowers the fractal dimension of the resulting topography. Structures introduced by rock-slope failures introduce visible short-term complexity on hillslopes; however, this is quickly incorporated into the channel network so does not appear to increase the fractal dimension of areas defined as hillslopes using the threshold slopearea criterion. Further work is suggested to improve modelling of fluvial processes in steep, headwater channels, and to investigate the details of hillslope-channel coupling at higher spatial resolution.
Item Citations and Data