Plucking phenomena in nonuniform flow

River channel erosion by plucking is poorly understood even though it is a dominant mechanism for bedrock river profile evolution. In an experimental flume with fractured slabs of plaster model bedrock, plucking from a bed lacking protrusion is produced by nonuniform flow, particularly in rapidly varied flow with hydraulic jumps and free-surface undulations. Model bedrock slides upstream toward water-surface depressions in regions lacking recirculation, groups of blocks bulge up when a trough of free-surface waves moves above, and bubbles and debris particles move in the bed crack network. The likelihood and completeness of plucking increases with average flow strength but relies on local flow properties for initiation of motion. Particle image velocimetry (PIV) analysis of flow during a plucking event suggests that flow structures smaller than the average size of the blocks may be important in the plucking process by generating velocity or pressure variations around the blocks. Because plucking typically occurred near free-surface undulations and we have observations consistent with crack network flow, we propose that the mechanism driving block lift starts in the static pressure gradients developed in the sub-bed crack network, which are locally and temporally enhanced by turbulent pressure fluctuations. Positive feedback occurs when plucked blocks alter flow character and allow other blocks to slide around the bed, promoting additional plucking. Negative feedback occurs where the deposition of plucked blocks downstream of nonuniform-flow reaches limits transport capacity by changing or damping the nonuniform flow upstream. Our experimental results are consistent with previous engineering studies of slab uplift under plunging jets and high-Froude-number hydraulic jumps in energy-dissipating spillways. Our results also point toward the ability of nonuniform flow in bedrock rivers with a low Froude number to generate lift of fractured bedrock below steps and constrictions, and suggest a need for further study of mechanisms that initiate block plucking in experimental and field settings.