Tidal sand ridges on the shelf: A numerical study of their natural morphodynamic evolution and response to interventions

Abstract

Tidal sand ridges are large-scale bottom features that occur on continental shelves with strong tidal currents and abundance of sand. These ridges have longshore spacings of several kilometres, heights of tens of metres, they evolve on centennial time scales and their crests are cyclonically rotated with respect to the direction of the dominant tidal currents. The coupled Delft3D-SWAN numerical model is used to study the natural morphodynamic evolution of tidal sand ridges and their response to interventions (sand extraction and construction of islands) in a setting that resembles the Belgian shelf. New aspects in this study are that a sophisticated wave model (SWAN), a shelf with a sloping bottom and state-of-the-art formulations for sediment transport (including suspended load) are considered. Starting from an initially flat bed, model results show that sand ridges develop on the shelf that have similar characteristics as those of observed tidal sand ridges. Results further show that ridges recover after sand extraction, i.e., the original sand volume of the ridge crests is recovered on decadal time scales. Deepening of the troughs provides sand for this recovery. The recovery of the ridges is weaker when pits are located further offshore and/or are deeper. Ridges recover faster when waves are included. If a sand pit is too deep (>4 m), the ridge loses sand in the first tens of years prior to its recovery on longer time scales. This initial sand loss seems to agree with the observed loss of sand volume from a tidal sand ridge on the Belgium shelf (Kwinte Bank), which has been subject to intense extractions. Model results further show that islands lose sand on decadal time scales until they eventually disappear. This sand loss is deposited on the crests and in the troughs. The presence of islands causes the surrounding ridges to break into smaller ridges that have smaller orientation angles than those of the natural case. Using different island configurations (location, geometry and number of islands, longshore distance between islands) does not lead to qualitative changes of the model results.