A standard dynamic Reynolds stress model, with conventional coefficien
ts, is applied to oscillatory boundary layer flows. With a grid resolu
tion over the boundary layer thickness and wave period of the order of
100 and 600 respectively, well defined, grid-independent solutions ar
e obtained. The available data are predicted in great detail. However,
even with turbulence characteristics, the data from oscillatory flows
do not appear to be very model discriminant. A model based upon a sta
ndard (k-epsilon) closure also predicts them reasonably realistically.
With sediment entrainment, giving stably stratified flow, the Reynold
s stress model estimates that there is almost no turbulence above the
mean velocity maximum. This is probably a reason why a (k-epsilon) mod
el even predicts such flows accurately. Another reason is that the flo
w is strongly forced (by the oscillatory pressure gradient) and is not
, like for instance turbidity currents, decisively governed by the tur
bulence. An oscillatory flow with sediment entrainment on a slope is p
redicted to force a systematic turbidity current. At large enough slop
e angles, the waves are predicted to trigger self-accelerated turbidit
y currents.