Silt and sand transport in a deep tidal channel of a large estuary (Manukau Harbour, New Zealand)

Time series of suspended-silt concentration (measured by an optical backsca tter sensor) and suspended-sand concentration (measured by an acoustic back scatter sensor) in a deep channel (14 m MSL) of a large New Zealand estuary (Manukau Harbour) are analysed. Suspended-sand concentration varied in pha se with the current speed and bed shear stress on a similar to 6.2-h cycle, but suspended-silt concentration varied on a semidiurnal (similar to 12.4- h) cycle such that the channel was clear of silt at high tide and maximum s ilt concentration occurred at low tide. Threshold for initiation of sand mo tion was found to be related to local skin friction, sand concentration pro files were consistent with settling flux balanced by gradient diffusion wit h a two-layer sediment diffusivity dependent on local friction velocity, su spended-sand reference concentration was explainable in terms of local skin friction, and changes in the channel-bed sediment were correlated with the suspended-sand load. Therefore, sand suspension in the channel is a "local " process and the channel bed is the source of the suspended-sand load. Sus pended-sand flux was found to be a highly nonlinear function of the tidal c urrent speed in the channel. Ln that case, tidal-current velocity asymmetry is the principal determinant of direction of net sand transport and the eb b/flood-dominance concept is valid. The relative phase of the M-2 and M-4 c onstituents of the tidal-current velocity, which quantifies tidal-current a symmetry, was calculated from a tide model of the estuary and results were mapped. The harbour was found to be broadly ebb dominant and therefore self -flushing of sand. Silt suspension in the channel was not related to the lo cal boundary-layer dynamics, but was explainable in terms of the movement u p and down the channel of a horizontal gradient in silt concentration, whic h implies the existence of a turbid water mass that is perched on the surro unding intertidal flats at high tide and that drains into the channel durin g ebb tide. Silt concentration in the channel increased during a short stor m, which corresponded to the passage through the measurement site of a "tur bid fringe" that was formed on the surrounding intertidal flats during the previous high tide under the action of breaking waves. Wave activity on the surrounding intertidal flats is therefore the principal determinant of sus pended-silt load in the channel and an approach that treats waves and sedim ent pathways between intertidal flats and channels is needed for modelling silt transport. (C) 2000 Elsevier Science B.V. All rights reserved.