10
K. Black· M. Green· T. Healy· R. Bell· J. Oldman· T. Hume
sensor at high tide. This error is caused by accentuation of the high frequency
sensor noise in the spectrum by the linear theory transform for the deeper sensor (Hutt and Black 1997). The error occurs when depth is about 2 m for 0.4 m
waves with peak spectral periods of approximately 3 s and would increase for
smaller waves with shorter wavelengths or in deeper water.
Comparison of the measured sea surface wave heights and the model-predicted values indicates a close agreement (Fig. 2d). The deviation at the peak relates
to the three largest "measured" values which are over-estimated heights from the
lower pressure sensor. The model shows good agreement at this time with the
heights calculated from the upper sensor records. The only major deviation occurs around bursts 2156 and 2158. The wind speed was very low (2 ms- I ) and so
heights are under-predicted. Possibly, the total variance in the pressure records
used to calculate the "inferred" significant wave height may include variance
from other (non-wind-wave) processes, sensor noise or lower frequency seiching. One of the authors (Green) observed nonlinear solitary waves at the site of
unknown origin. Irrespective of the cause of the deviation, the wave model cannot generate the inferred wave height, given the very low wind speeds. Because
of the low wind speeds, near-bed orbital currents predicted by the model are
correspondingly low at the start of the period (Fig. 2e). Near-bed peak spectral
periods show good agreement with the measured values, except around burst
2156 when height predictions were low (Fig. 2f).
All processes (winds, sand bank emergence, fetch, shoaling, breaking, frictional dissipation, surface spectral shape and bed orbital transformation) are
calculated directly by the model, but the friction coefficient needs to be calibrated. Best fit to the measurements was obtained with a friction coefficient C f =
0.025. This value is physically acceptable for the low relief sandy beds experienced in the shallow parts of the estuary, compared with higher values of around
0.06-0.09 obtained on coastal rocky terrain (McComb et al. 1997; Hutt 1997).
The friction is highly important because of the long travel distances in shallow
water, particularly around low tide, and the model indicates that wave heights
are strongly influenced by friction.
4.3
Suspended Sediment Concentrations
In the next phase, the model predictions of currents and waves are used in the
sediment transport model POL3DD. While some deviations between models
and data have been noted, the model predictions are used here rather than actual measurements in order to use numerical models only to predict SSC at the
RALPH site. Pick-up function techniques [used by Black and Rosenberg (1991),
Black (1994), Black et al. (1995)] for the entrainment of sediment under waves
have been applied here without modification and the third velocity moment is
adopted as the "representative" orbital current, in accordance with the
recommendations in these publications. Similarly, the vertical eddy diffusivity
for sediment is taken as constant and equal to 0.0015 m 2 .S-I • A grain size popu-
K. Black· M. Green· T. Healy· R. Bell· J. Oldman· T. Hume
sensor at high tide. This error is caused by accentuation of the high frequency
sensor noise in the spectrum by the linear theory transform for the deeper sensor (Hutt and Black 1997). The error occurs when depth is about 2 m for 0.4 m
waves with peak spectral periods of approximately 3 s and would increase for
smaller waves with shorter wavelengths or in deeper water.
Comparison of the measured sea surface wave heights and the model-predicted values indicates a close agreement (Fig. 2d). The deviation at the peak relates
to the three largest "measured" values which are over-estimated heights from the
lower pressure sensor. The model shows good agreement at this time with the
heights calculated from the upper sensor records. The only major deviation occurs around bursts 2156 and 2158. The wind speed was very low (2 ms- I ) and so
heights are under-predicted. Possibly, the total variance in the pressure records
used to calculate the "inferred" significant wave height may include variance
from other (non-wind-wave) processes, sensor noise or lower frequency seiching. One of the authors (Green) observed nonlinear solitary waves at the site of
unknown origin. Irrespective of the cause of the deviation, the wave model cannot generate the inferred wave height, given the very low wind speeds. Because
of the low wind speeds, near-bed orbital currents predicted by the model are
correspondingly low at the start of the period (Fig. 2e). Near-bed peak spectral
periods show good agreement with the measured values, except around burst
2156 when height predictions were low (Fig. 2f).
All processes (winds, sand bank emergence, fetch, shoaling, breaking, frictional dissipation, surface spectral shape and bed orbital transformation) are
calculated directly by the model, but the friction coefficient needs to be calibrated. Best fit to the measurements was obtained with a friction coefficient C f =
0.025. This value is physically acceptable for the low relief sandy beds experienced in the shallow parts of the estuary, compared with higher values of around
0.06-0.09 obtained on coastal rocky terrain (McComb et al. 1997; Hutt 1997).
The friction is highly important because of the long travel distances in shallow
water, particularly around low tide, and the model indicates that wave heights
are strongly influenced by friction.
4.3
Suspended Sediment Concentrations
In the next phase, the model predictions of currents and waves are used in the
sediment transport model POL3DD. While some deviations between models
and data have been noted, the model predictions are used here rather than actual measurements in order to use numerical models only to predict SSC at the
RALPH site. Pick-up function techniques [used by Black and Rosenberg (1991),
Black (1994), Black et al. (1995)] for the entrainment of sediment under waves
have been applied here without modification and the third velocity moment is
adopted as the "representative" orbital current, in accordance with the
recommendations in these publications. Similarly, the vertical eddy diffusivity
for sediment is taken as constant and equal to 0.0015 m 2 .S-I • A grain size popu-