A New View of Nearshore Sediment Sorting: the Self-Organization of Rippled Scour Depressions

On inner continental shelves, wave motions and mean currents interact with a bed that is commonly composed of mixed grain sizes. In this situation a simple feedback (described below) could produce spontaneous sediment sorting. 'Rippled scour depressions,' swaths of coarse sediment on the order of a hundred meters wide and extending up to kilometers from shore, are currently interpreted as the imprint of heterogeneous, offshore-directed storm currents, although such currents have not been observed and are not expected theoretically. We hypothesize, however, that the simple feedback could produce such features under the influence of homogeneous alongshelf currents. The coarse sediment of rippled scour depressions is typically arranged into large wave-generated ripples on the order of a meter in wavelength and a decimeter in height. Sharp boundaries separate these coarse domains from intervening areas with finer sediment and much smaller ripples. Divers have observed that fine sediment can be suspended an order of magnitude higher over the coarse domains than over the fine areas, because of the turbulence generated as wave motions interact with the large ripples. Any mean current can advect this sediment past the coarse domain. This interaction will tend to reinforce the sorting, because fine material entrained from the coarse domain will tend to be deposited preferentially in an adjacent fine domain. Starting from random variations in bed composition, the greater grain-size and bedform roughness and turbulence in an area with slightly coarser sediment could winnow fine material locally in the same way. This feedback will tend to produce a sorted pattern. However, whether these interactions operating over a spatially and temporally extended domain will produce large-scale sorted features with the characteristics of rippled scour depressions is not obvious. To investigate whether features such as rippled scour depressions could result from these basic interactions, we have developed a numerical model. In this simple model, we do not explicitly simulate hydrodynamics. Instead, the combined effects of wave motions and mean currents on sediment transport are parameterized as a function of bed composition (as a proxy for ripple size). For example, the friction factor that affects the magnitude of sediment transport depends on bed composition, represented by the proportions of the two grain sizes considered. Suspended-load transport of coarse and fine sediment is treated separately. A lag in the adjustment of a suspended-sediment concentration profile to changes in flow conditions, ignored in most sediment-transport modeling, plays a key role in this model. For example, sediment with a small settling velocity, suspended far from the bed, may not adjust rapidly enough to stay in equilibrium with the local conditions, such as when fine sediment is advected past the edge of a coarse domain. These non-local effects lead to characteristic spatial scales in the evolving pattern (which depend on wave and current parameters). The basic plan view and cross-sectional characteristics of the features that emerge in the model are consistent with those of natural rippled scour depressions, indicating that the simple interactions modeled provide a plausible explanation for such enigmatic sorting phenomena.