LAUSR

Step pools are a common bed morphology in boulder-rich gravel streams, but predicting how mountainous landscapes will respond to environmental perturbations such as climate-related hydrological changes requires a better understanding of channel morphodynamics and factors that influence bed stability. Flume experiments exploring bed stabilization demonstrate feedbacks among surface roughness, coarse grain clustering, and surface grain size. Clustering is quantified by using a novel normalization of Ripley's K statistic designed for use in power law functions. At 95% confidence, many but not all beds stabilized with coarse grains becoming more clustered than complete spatial randomness. The clustering statistic predicts hydraulic roughness better than D84 does (the diameter at which 84% of grains are smaller), suggesting that the spatial organization of the bed can be a stronger control than grain size on flow hydraulics. Initial conditions affect the degree of clustering at stability, indicating sensitivity to history.