Degree Name

Doctor of Philosophy


School of Geosciences


Alluvial channel geometry has been studied quantitatively for over one hundred years and it is now recognised to be a multi-variable problem. However, as yet there is no widespread agreement on what variables are the most important, nor is there agreement on the relationships between variables or on the form of general model. This study tests the applicability to both stable canals and natural alluvial channels of an experimental flume relationship between channel shape and boundary shear distribution. It establishes a multivariate model of downstream hydraulic geometry, showing clearly that channel geometry is generally controlled by four factors: flow discharge, channel roughness, slope and bank strength.

This study makes five important contributions. Firstly, it presents a physical explanation for each of the major factors controlling channel geometry. Secondly, because it has been developed from a wide range of field observations accumulated over many years and studied in combination here, it provides a high level of prediction of channel form. Thirdly, it presents a partially quantitative description of the influence of bank strength on channel form. Vegetated and highly cohesive banks result in narrower and deeper sections while the non-vegetated and noncohesive sandy banks produce wider and shallower sections. Fourthly, it demonstrates the role that sediment transport plays in influencing channel form, finding that velocity and slope are highly dependent on sediment concentration while channel cross-section is slightly dependent. An increase in bed-material load causes marked increase in width only when the adjustment of channel slope is severely restrained. Finally, this study finds that regime theory, threshold channel condition, channel shape relations and bivariate hydraulic geometry models provide appropriate explanation only for specific conditions and do not provide a general explanation or predictive model for channel geometry.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.