Threshold of motion of bivalve and gastropod shells under oscillatory flow in flume experiments

The threshold of motion of non-fragmented mollusc shells was studied for the first time under oscillatory flow. In this regard, flume experiments were used to investigate the threshold of motion of three bivalve and three gastropod species, two typical mollusc classes of coastal coquina deposits. The sieve diameters ranged from 2·0 to 15·9 mm. These experiments were performed on a flat-bottom setup under regular non-breaking waves (swell) produced by a flap-type wave generator. The critical Shields values for each species of mollusc were plotted against the sieve and nominal diameter. Moreover, the dimensionless Corey shape factor of the shells was evaluated in order to investigate the effect of mollusc shell shapes on the threshold of motion. According to their critical Shields parameter, the mollusc threshold data under oscillatory flow present smaller values than the siliciclastic sediments when considering their sieve diameter. In addition, the mollusc datasets are below the empirical curves built from siliciclastic grain data under current and waves. When considering the nominal diameter, the critical Shields parameter increases and the mollusc data are closer to siliciclastic sediments. Bivalves, which have a flat-concave shape (form factor: 0·27 to 0·37), have a higher critical Shields parameter for smaller particles and more uniform datasets than the gastropod scattered data, which have a rounded shape (form factor: 0·58 to 0·62) and have varied morphologies (ellipsoidal, conical and cubic). The comparison between previous current-driven threshold data of bioclastic sediment motion and the data of mollusc whole shells under oscillatory flow shows a fair correlation on the Shields diagram, in which all datasets are below the mean empirical curves for siliciclastic sediments. These findings indicate that the shape effect on the transport initiation is predominant for smaller shells. The use of the nominal diameter is satisfactory to improve the bioclastic and siliciclastic data correlation.