Inner continenatl shlef grain-size sorted patterns: Instability and self-organization

Side-scan sonar observations in recent years have shown that a fascinating kind of large-scale pattern appears adorns many inner continental shelves around the world. Some mechanism sorts mixed-grainsize seabed sediment into domains of coarse material (coarse sand and/or gravel) separated by fine sand. The coarse domains commonly form swaths on the order of a hundred meters wide and extending up to kilometers from shore. These features have been interpreted as the imprint of hypothetical alongshore-localized, offshore-directed storm currents that locally winnow the fine material. We present a new hypothesis for this type of pattern that involves, rather than an imposed template in the forcing, feedback and subsequent self-organization. The seabed in coarse domains is typically sculpted into large wave-generated ripples. The large-scale, energetic turbulence generated as wave motions interact with the large ripples tend to suspend fine sediment. Any mean current can advect this fine material past the coarse domain. Starting from random variations in bed composition, the greater roughness in an area with slightly coarser sediment could winnow fine material locally, producing a feedback that would tend to produce a sorted pattern. To investigate whether these interactions operating over a spatially and temporally extended domain could produce large-scale sorted features with the realistic characteristics, we have developed a simple numerical model. In this model, rather than explicitly simulating relatively small-scale hydrodynamics, we parameterize the combined effects of wave motions and mean currents on sediment transport as a function of bed composition (as a proxy for ripple size). The model robustly generates sorted patterns. The size and appearance can match those of the natural features, although the details of the patterns depend on the wave and current characteristics. The size and appearance of the features that develop over an extended time in the model differ significantly from those of the pattern that forms initially. As perturbations grow to finite amplitude, they interact with each other in ways that lead to a larger-scale, better-organized pattern.