Lagrangian Modelling Techniques Simulating Wave and Sediment Dynamics ...
15
Thus, the cause of low mud content on the exposed banks in the Manukau
Harbour is associated with intermittent entrainment of the fines and losses from
shallow banks where wave orbital currents are largest. In general, the distribution of muds may therefore be in balance with (i) the input volumes and source
locations (river sources are often located up harbour), (ii) settlement time scales
as a function of grain size, (iii) the wave energy levels at each location in the estuary, (iv) tidal current strength and direction on the sand banks (including the
residual current) and (v) the flushing time of the estuary overall. Local wave energy determines the entrainment rate. Tidal currents and horizontal diffusion
move the fines into the channels while the flushing time determines the amount
of mud expelled through the estuary entrance when waves are stirring the bed
internally. The settlement time scales determine the excursion distance per
event for each grain size fraction. However, the key factor determining mud content on intertidal banks is wave energy.
To simulate the SSC, it was necessary to firstly generate accurate predictions
of water levels, currents and wave energy. Water levels and currents were effectively predicted at the RALPH site, notwithstanding input data limitations including (i) no boundary sea level measurements, (ii) coarse hydrographic survey data confounded by a spatially-varying datum shift and (iii) no representation in the 200 m model grid of the meso-scale sand waves observed at the site.
Improved predictions must therefore be possible with finer model grids and
more accurate measurements of sea level and bathymetry. While tidal modelling
is one of the most proven numerical simulations, the use of these models on intertidal sand flats in a storm is less well documented. The need for high-resolution modelling is highlighted by the findings of Bell et al. (1998) that bed slope
on the intertidal flat, which partly determines surf zone width, is also very important.
Wave conditions at the RALPH site were also effectively predicted, even
though the waves were considerably influenced by bathymetry during generation. The smaller heights as the tide drops were shown by the model to relate to
increased effect of bed friction, wave breaking around low tide and shortened
fetches as the banks emerge. The correctly modelled height oscillation at high
tide, for example, is due to an oscillation in the wind strength. Thus, pseudosteadiness, as assumed in the model, is evidently adequate at the 1-h time-scale
of the measured winds over the fetches of up to 20 km. The assumption is also
confirmed by the prevalence of single-peaked narrow spectra recorded at the
site. Other oscillations in the wave heights (e.g. 2184 and 2186) relate to changes
in bank submergence and these are being accurately predicted. The only significant deviations occurred at the start of the model period when wind speeds
were lower than expected for the measured orbital currents. The wave heights,
bed orbital motion and SSC prediction were all then affected by this anomalous
wind strength. The effective height predictions conversely indicate that deformation by tidal currents (which is not being modelled) is a secondary process.
The challenge to model a comprehensive dataset from a site within a complex
estuary was initially established as a test of present numerical modelling capac-
15
Thus, the cause of low mud content on the exposed banks in the Manukau
Harbour is associated with intermittent entrainment of the fines and losses from
shallow banks where wave orbital currents are largest. In general, the distribution of muds may therefore be in balance with (i) the input volumes and source
locations (river sources are often located up harbour), (ii) settlement time scales
as a function of grain size, (iii) the wave energy levels at each location in the estuary, (iv) tidal current strength and direction on the sand banks (including the
residual current) and (v) the flushing time of the estuary overall. Local wave energy determines the entrainment rate. Tidal currents and horizontal diffusion
move the fines into the channels while the flushing time determines the amount
of mud expelled through the estuary entrance when waves are stirring the bed
internally. The settlement time scales determine the excursion distance per
event for each grain size fraction. However, the key factor determining mud content on intertidal banks is wave energy.
To simulate the SSC, it was necessary to firstly generate accurate predictions
of water levels, currents and wave energy. Water levels and currents were effectively predicted at the RALPH site, notwithstanding input data limitations including (i) no boundary sea level measurements, (ii) coarse hydrographic survey data confounded by a spatially-varying datum shift and (iii) no representation in the 200 m model grid of the meso-scale sand waves observed at the site.
Improved predictions must therefore be possible with finer model grids and
more accurate measurements of sea level and bathymetry. While tidal modelling
is one of the most proven numerical simulations, the use of these models on intertidal sand flats in a storm is less well documented. The need for high-resolution modelling is highlighted by the findings of Bell et al. (1998) that bed slope
on the intertidal flat, which partly determines surf zone width, is also very important.
Wave conditions at the RALPH site were also effectively predicted, even
though the waves were considerably influenced by bathymetry during generation. The smaller heights as the tide drops were shown by the model to relate to
increased effect of bed friction, wave breaking around low tide and shortened
fetches as the banks emerge. The correctly modelled height oscillation at high
tide, for example, is due to an oscillation in the wind strength. Thus, pseudosteadiness, as assumed in the model, is evidently adequate at the 1-h time-scale
of the measured winds over the fetches of up to 20 km. The assumption is also
confirmed by the prevalence of single-peaked narrow spectra recorded at the
site. Other oscillations in the wave heights (e.g. 2184 and 2186) relate to changes
in bank submergence and these are being accurately predicted. The only significant deviations occurred at the start of the model period when wind speeds
were lower than expected for the measured orbital currents. The wave heights,
bed orbital motion and SSC prediction were all then affected by this anomalous
wind strength. The effective height predictions conversely indicate that deformation by tidal currents (which is not being modelled) is a secondary process.
The challenge to model a comprehensive dataset from a site within a complex
estuary was initially established as a test of present numerical modelling capac-