This paper investigates the effect of wave-induced streaming on sediment transport by applying a newly developed numerical sediment transport model. The transport model solves the intra-wave oscillatory motion and sediment concentrations. The wave-induced streaming in the wave boundary layer, as occurs in propagating waves, is an integral part of the computed flow field. Comparison of the computed flow field with wave-current flume experiments gives good results. The model can also describe the horizontally uniform situation in an oscillating water tunnel, in which vertical orbital velocities and wave-induced streaming are absent. The model is successfully tested against measurements of sediment concentrations and net transport rates performed in a large-scale oscillating water tunnel. The importance of streaming is assessed by comparing model results for the case with wave-induced streaming and the case of the oscillating water tunnel. For a typical water tunnel condition, the effect of the absence of vertical velocities and streaming is found to be a 40% lower net transport rate than the corresponding situation under propagating waves. These differences result not only from the net transport by the streaming velocity, but are also due to an increased asymmetry in the model of near-bed velocity, bed shear-stress and resulting sediment concentrations. This result suggests that the differences in transport rates between water tunnel and propagating waves may not be insignificant.

A frequency-domain Boussinesq model with good linear shoaling, improved linear dispersion characteristics and a dissipation formulation to account for wave breaking is extended to include the computation of the vertical structure of the horizontal velocity. The extended model is used to predict bottom velocities and resulting velocity variance and skewness in (partially) breaking irregular waves. The comparison of measured and computed velocity moments indicates that for moderately long waves the spectral Boussinesq model can be successfully used for sediment transport purposes. For shorter waves the crest velocity values of the higher waves are significantly underestimated, and as a result the velocity skewness as well.

Existing Boussinesq models are extended to include the computation of the vertical structure of the horizontal velocity. A time-domain model is tested against laboratory measurements of the vertical profile of the horizontal velocity in regular waves; good results are obtained, especially in the near-bed zone. A spectral model, which includes a dissipation formulation to account for wave breaking, is tested against laboratory measurements of bottom velocities in (partially) breaking irregular waves. For moderately long waves, the comparison on velocity variance and skewness, which are relevant to sediment transport, yields good results.