FLOW IN FRACTURED BEDROCK CRACK NETWORK AND ITS ROLE IN PLUCKING

The erosional mechanism of plucking is poorly understood, but it plays a large role in the erosion and transport of bedrock in rivers. Understanding the mechanism for plucking is necessary to model river incision going forward. We propose three hypotheses for how plucking is initiated in rapidly varied flow. The first two hypotheses: change in head and coherent flow structures along the bed, are driven by the hydraulic jump. The third hypothesis: pressure pulse transmission, is aided by upstream steps but is not studied.

We used a 2.44 m X 14.3 cm x 33 cm Plexiglas flume with plaster blocks as experimental, fractured bedrock. We instrumented the flume with dye ports at the downstream end of the fractured plaster to observe water flow underneath and between the cracked bedrock. In some experiments we measured pressure under the plaster blocks at the channel centerline. We varied the location and shape of the hydraulic jump by changing flume slope, presence and height of upstream bedrock steps, and outlet velocity. In all experiments, dyed water flowed upstream beneath the bedrock in the crack network, becoming more diffuse upstream. When an upstream bedrock step was present, dye collected behind it upstream. Dye, sometimes as high- concentration events, emitted from the crack network into the overlying flow, and at times, they occurred just prior to the initiation of plucking. Dye “tornadoes” appeared in areas of low head at the channel margin at the toe of the jump, suggesting lateral and vertical flow. We coordinate the observations of dye and flow behavior with the sub-bed pressure measurements. The head of the jump causes a high pressure zone where water is forced into bedrock cracks and flows upstream until it reaches the low pressure zone at the toe of the jump. The results support previously observed initiation of plucking in rapidly-varied flow. Flow in the sub-bed crack network is present, it varies response to the flow conditions, and is highly nonuniform and nonsteady in response to the location and changing character of a jump. Our observations of water flow and pulsed behavior under and above the crack network suggest that plucking initiation in rapidly-varied flow is caused by a complex combination of drivers.