Fluvial channel geometry is controlled by the interaction of a number of geologic and hydraulic variables. The width of mixed alluvial-bedrock channels generally is a function of discharge, with variations due to local conditions. The aspect ratio (width/depth) of channels is heavily influenced by substrate size and erodibility. How channel width and aspect ratio vary as a function of other variables, such as structure, lithology, slope, large scale valley topography, and rock uplift, has not been fully quantified. The New River is ideal for examining these relationships because it shows considerable variability in width and aspect ratio and flows through three structurally and lithologically distinct geologic provinces. Through these provinces, the New River does not follow the expected trends of channel widening with increasing drainage area. Topographic maps show that channel width of the New River has a significant variation that far outscores an overall widening downstream. Aspect ratios for the New River are also large, approaching 500. We collected a field data set of 29 sites of the river's channel geometry, along with characteristics of bedrock, sediment, and confinement. Fifteen of the 29 sites are bedrock reaches. The data set allows empirical analysis of how width and aspect ratio of the New River are related to different variables, including slope, discharge, flow velocity, curvature, trend, bedrock type, and structure. Sediment characteristics and confinement of the channel do not affect channel morphology. Bedrock is shown to affect channel width directly through the percent of bedrock exposed in the channel and indirectly through the modified rock mass strength, rock hardness, obliquity to regional strike, dip orientation, and degree of joint intersection. / Master of Science

Des études de terrain, expérimentales et numériques ont été menées pour rendre compte de l'incision des rivières à fond rocheux par le transport sédimentaire. Toutefois, le caractère fluctuant du système étudié rend difficile l'établissement de lois générales. La thèse s'articule autour de l'analyse des résultats fournis par un programme informatique qui simule la mise en mouvement de sédiments par un écoulement turbulent. La modélisation et l'intégration des interactions entre les sédiments et le fluide ainsi que les contacts entre sédiments se basent sur la description mécanistique du système. La validation du modèle numérique a été effectuée par la comparaison de paramètres dont les valeurs pour des rivières naturelles ou expérimentales sont abondantes dans la littérature. Le nombre de Shields critique de mise en mouvement, la relation entre la vitesse du fluide et le débit de sédiment sont ces grandeurs de références. Le travail échangé entre les sédiments et le fond est comptabilisé et ce flux d'énergie est rattaché au taux d'érosion avec des paramètres mécaniques de la roche. Nous avons établi une relation entre le taux d'érosion et le nombre de Shields. Les variations du taux d'incision avec masse de sédiment présente dans le lit de la rivière montrent que nous reproduisons le « tool-effect » et le « cover-effect ». Nous prédisons l'existence d'un nombre de Shields critique propre à l'incision qui est supérieur au nombre de Shields critique de transport. Les travaux échangés entre la couverture sédimentaire et les parois du canal permettent de quantifie le taux d'érosion latérale. Les travaux échangés au sein de la couverture sédimentaire sont utilisés pour évaluer le taux d'usure des sédiments. À partir du bilan de puissance du sous-système couverture sédimentaire, nous proposons un scénario pour la croissance du chenal d'une rivière à fond rocheux. Plusieurs études paramétriques rendent compte de l'influence de la rugosité du fond, du coefficient de restitution et de la gravité. / Field measurements, empirical studies and numerical approaches had been led in order to explain bedrock river incision due to transport of sediments. However, the studied system is roughly fluctuating, what make difficult to establish general laws. The thesis is based on the analysis of data extracted from a numerical code that simulate the motion of sediment by a turbulent flow. The interactions between the sediments and the fluid and the contact between sediments are modeled on physical basis and integrated following the principles of mechanics. The numerical process has been validated by comparison between two parameters extracted from natural rivers and experimental setup. The critical Shields number for sediment transport and the relationship between flow velocity and flux of sediments has been compared to values taken from literature. The work of dissipative forces between sediments and bedrock is computed. We propose a quantification of vertical incision rate based on rock resistance parameters. We have established a relationship between incision rate and Shields number that fit with a power law. We reproduce the two antagonistic effects that rules incision process: tool-effect and cover-effect. We predict the existence of a critical Shields number for incision that is higher than sediment transport threshold. We quantify the lateral incision rate with the work of dissipative forces between channel benches and sediments. We also quantify the attrition rate of the sediments with the value of energy dissipated within the sediment cover. The power balance of the sub-system sediment cover gives basis that lead to a scenario for bedrock river channel growth. Other results about influence of bedrock roughness, coefficient of restitution and surface gravity on incision rate are exposed.