Flow dynamics in bedrock canyons

Bedrock rivers define base level for the watersheds they drain, and therefore incision of the largest bedrock rivers sets the rate of landscape evolution for entire mountain ranges. Previous investigations observed plunging flows in bedrock rivers that cause the fastest flows to submerge below the water surface, approaching the bed in deep pools. However, the frequency, statistical properties, and the impact on bedrock erosion of plunging flows are not well understood. This study explores the flow dynamics of a 375 km long reach of the Fraser River colloquially referred to as the Fraser Canyon, which drains a large portion of western Canada and is teeming with exposed bedrock. Plunging flows are shown to excavate sediment cover in scour pools during flood discharges exposing bedrock to the erosive particle impacts which carve the channel. Across the Fraser Canyon plunging flows occupy more than half of bedrock confined reaches. Clusters of plunging flows are found to dominate the flow dynamics of two steeper sections of the river where it flows through knickzones. Using particle-impact erosion modelling, plunging flows are shown to drive incision of the Fraser Canyon system, particularly in these knickzones. This prompts an exploration of the diversity of flow dynamics and morphology in bedrock reaches to understand how erosion is distributed at the reach scale. In longer bedrock confined reaches, many overlapping plunging flows can occur in sequence forming a complex, or even sustained submerged plunging flows where the maximum velocities are found submerged and following the bed topography. Detailed observations of plunging flows, sustained submerged flows, rapids, and overfalls are presented, and the reach-scale distributions of shear stresses and erosion potential of these various flow dynamics are compared. The flow dynamics, morphologies, and patterns of modelled erosion suggest a continuum of bedrock channel evolution over geologic timescales. This study provides new insight into the mechanical processes of bedrock erosion that are not represented in large-scale landscape evolution models, setting the stage for a re-evaluation of process representation in landscape evolution modelling.