Ratio of River Channel Bar to Bank Height Sets Bank Erosion Rate

River channel changes during floods, including bank erosion and bar deposition, are difficult to predict but important for infrastructure, river management, and riparian zone health. Upstream dams can also drive disequilibrium changes to channel morphology, potentially changing bar and bank susceptibility to flooding. To understand feedback between bar morphology, flood flow, bank erosion and channel cross-sectional evolution, we (a) evaluate how bank erosion during natural floods in six gravel-bed rivers with alternate and/or mid-channel bars varied with bar and bank height, and (b) conduct laboratory experiments scaled to the Chubetsu River in Japan, a pseudo-meandering gravel-bed river that has experienced bank erosion influenced by alternate bars. The field data show a strong correlation between bank erosion during floods and the bar to bank height ratio measured before the respective floods. We also find that rivers with upstream dams tend to have higher bar to bank height ratios. In the experiments, we systematically vary flood discharge, initial bar height, and initial bank height, and measure bank erosion and bar evolution. We experimentally find that bank erosion initially increases with bar height and flood discharge. As bar height evolves, bank erosion rates tend to converge under different initial conditions. Bank erosion rate increases with decreasing bank height. Because bars in natural rivers tend to not be adjusted to the largest floods, our experimental results and field data suggest that surveyed ratios of bar height to bank height could be used to infer relative bank erosion risks prior to large floods.