Continental Shelf Research 22 (2002) 305–325
Sediment resuspension across a microtidal, low-energy inner shelf J. Guille! na,*, J.A. Jime! nezb, A. Palanquesa, V. Graciab, P. Puiga, A. Sa! nchez-Arcillab b
a " Institut de Ciencias del Mar, CSIC, Passeig Marı´tim de la Barceloneta, 37–49, 08003 Barcelona, Spain ! Laboratori d’Enginyeria Mar!ıtima, ETSECCPB, Universitat Politecnica de Catalunya, C/Jordi Girona, 1–3, Campus Nord, ed. D1, 08034 Barcelona, Spain
Received 21 January 2000; received in revised form 1 February 2001; accepted 6 February 2001
Abstract Simultaneous field measurements were carried out across the Ebro delta inner shelf using two instrumented bottom boundary layer tripods deployed at 8.5 and 12.5 m water depth. The period analysed corresponds to the transition from fair-weather to wave-storm conditions. The recorded data show that the sediment resuspension was associated with the increases in wave activity (‘‘skin’’ wave shear stress) and started simultaneously at both study sites. Wave and current related parameters and the concentration of suspended sediment reflect an increasing trend towards the shallower location. The measured profile of suspended sediment concentration agrees well with exponential and power deterministic models. The estimated resuspension coefficient (g0 ) shows temporal and spatial variability, although average values (4–6 103) are near the ‘‘typical’’ resuspension parameter previously defined for rippled bottom conditions (2 103). r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Inner shelf; Sediment resuspension; Gradients across Ebro shelf
1. Introduction Sediment resuspension is one of the main processes that control the magnitude of suspended sediment transport because it determines the sediment concentration in the water column to be advected by currents. The intensity of the process is usually formalised in terms of the Shields parameter, (y), which measures the ratio between stirring and restoring forces. On the inner *Corresponding author. Tel.: +34-93-239-9500; fax: +3493-230-9555. E-mail address:
[email protected] (J. Guill!en).
shelf, the wave-induced bottom shear stress is generally taken as the main stirring factor, which is counteracted by the immersed weight of the particles. In natural coasts, this process can be altered (enhanced or prevented) by factors such as the presence of bedforms (e.g. Li et al., 1996; Li and Amos, 1998), bed armouring (e.g. Wiberg et al., 1994; Reed et al., 1999), cohesive sediments (e.g. Drake and Cacchione, 1989) or biological control (e.g. Wright et al., 1997). In tidal areas, the tidal flow can also stir the bottom sediment and this current is responsible for the cycling of particulate material through resuspension and deposition (Jago et al., 1993; Jones et al., 1998).
0278-4343/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 8 - 4 3 4 3 ( 0 1 ) 0 0 0 5 9 - 0
306
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Table 1 Reported values of the resuspension coefficient, g0 ; and corresponding reference elevation, z; and bottom characteristics Authors
g0
Z
Observations
Smith and McLean (1977) Wiberg and Smith (1983) Hill et al. (1988) Drake and Cacchione (1989) Vincent and Green (1990) Vincent et al. (1991) Madsen et al. (1993) Bedford and Lee (1994) Wiberg et al. (1994) Vincent and Downing (1994) Green et al. (1995) Li et al. (1996)
2.4 103 1.6 105 1.3 104 1.5 105–3 104 1–1.7 104 1.7 104–9.7 103 4 104 (SD 1.6 104) 2 103 (SD 8.6 104) 5.5 105–2.2 103 3.5 105–3 102 2.5 103 1.5 105–1 102
z0 z0 (0.2–0.5 cm) 2 cm z0 (0.1–0.6 cm) 2 cm 2 cm 7 d50 0.1–0.3 cm 0.02 cm 2 cm 2 cm z0 (0.1–0.6)
River, unidirectional flow Sandy smooth bottom Laboratory, unidirectional flow Silty, flat bottom Sandy, rippled bottom Sandy, rippled bottom Sandy, smooth bottom Sand-silty, rippled bottom Sandy rippled bottom Sandy rippled bottom Measured Sandy rippled bottom
The magnitude of the resuspension is quantified in terms of a reference concentration that is usually estimated as some function of the excess of the near-bottom shear stress above a critical value (e.g. Smith, 1977; Zyserman and Fredse, 1994). Within this deterministic context, the calculated reference concentration will largely depend on the proportionality constants included in the respective formulations, which in Smith’s model are called the resuspension coefficient, g0 : This coefficient is in fact only a calibration parameter that, according to the reported values, is poorly constrained and can vary up to three orders of magnitude (see Table 1). This extremely large range of variation could be explained in different ways: (i) calibrations were made following different procedures, e.g. using different reference levels; (ii) external factors affect the process, e.g. armouring or sediment cohesion; or (iii) the predictive formula does not account for all the relevant processes, and therefore those that are not resolved are implicitly included in the g0 value. The main implication of the scatter found in g0 values is that when the reference concentration has to be predicted, the proper value of g0 must be determined locally unless only ‘‘indicative’’ concentrations are desired. Although resuspension processes have been investigated during the last two decades (e.g. Smith and McLean, 1977), most of the studies correspond to oceanic coasts where wave conditions can cover the full range of the theoretical
‘‘wave-spectrum’’ (from sea waves to well-developed swells). In low-energy areas such as Mediterranean coasts, long period waves, i.e. swell, are hardly found, which can be illustrated by the fact that the yearly averaged mean wave period on the Ebro delta shelf is 4 s (e.g. Jime! nez et al. 1997). Thus, it is reasonable to assume that in these coasts, under the more frequent conditions, the near-bottom wave-induced velocity field will be weak, and that significant activity on the inner shelf will only take place under storms when wave periods increase significantly. As an example, wave peak period can increase up to a maximum of 12 s with mean periods of 8–9 s under the most energetic storms on the Ebro shelf (Jime! nez et al., 1997). In addition to this, the Ebro delta area is a microtidal environment (astronomical tidal range of 0.25 m) with very weak tidal currents, and therefore currents on the inner shelf are probably insufficient to resuspend sediment by themselves during usual conditions. Due to these factors (limited wave period and weak tidal currents), the main question to be solved for this kind of inner shelf is how intense the sediment resuspension can be. To adequately analyse the inner shelf functioning in terms of sediment dynamics, it is important to resolve the across-shelf variations of resuspension processes (Wright et al., 1999). These variations can be caused by an across-shelf gradient in near-bottom hydrodynamics, by across-shelf variations in bottom sediment, or by a combination of
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
the two. In general, near-bottom wave-induced shear stress increases shoreward, whereas sediment grain size usually shows a coarsening trend in the same direction. The combination of variations in near bottom hydrodynamics and grain-size can cause time lags in the sediment response across the inner shelf. These gradients will also control the across-shore distribution of suspended sediment concentration, which is extremely important when the inner shelf morphodynamic evolution is to be assessed (see e.g. Niedoroda et al. 1995). For instance, a strong gradient of the sediment grain size from medium sand to mud has been described across the Ebro inner shelf (Guille! n and Palanques, 1997). In this environment, a 6-h lag between sediment concentration at 1 m. a. b has been reported between two locations with depths of 8.5 and 12.5 m and a separation of 2.5 km (Jime! nez et al., 1999). This lag was associated with the seaward advection of fine sediments from the inner site, which did not necessarily correspond to sediment resuspended from the bottom. In this paper, the sediment resuspension across the Ebro delta inner shelf is analysed using in-situ measurements. The relevance of the study lies in the characteristics of the study area and the hydrodynamic conditions during the experiment. Thus, the experiment covers a period in which the hydrodynamic conditions range from very weak near-bottom flows with no sediment entrainment to a full-developed storm with a high sediment resuspension. Moreover, there is also an acrossshelf gradation in the grain size of the sediment at the bottom, which can induce differentiated responses of the inner shelf for a given set of hydrodynamic conditions.
2. Methods 2.1. Field campaign Data used in this paper were acquired during a field campaign carried out in front of the Trabucador Bar, the Ebro Delta (Fig. 1). Two instrumented tripods were deployed at 8.5 and 12.5 m water depth on the inner shelf during April 1997. Flow velocity and suspended sediment
307
concentration were measured at three different levels of the water column (approximately 0.1, 0.5 and 0.9 m above the bottom) using three bi-axial electromagnetic currentmeters (Delft Hydraulics P-EMS, Delft Hydraulics, 1993) and three optical backscattering sensors (OBS-3, D&A Instruments, 1991) respectively. The OBS sensors were calibrated in the laboratory with sediment of the bottom taken at each site. The superficial bottom sediment was sampled at both sites at the beginning of the deployment, the d50 of the sediment being 135 and 105 mm at the inner and outer locations respectively. Subsurface pressure was measured by means of an absolute pressure sensor (Druck PDCR-1830) in each tripod placed 1.75 m above the bottom. A KVH Industries C100 compass monitored the tripod orientation during the experiment. Sensor elevations with respect to the sea bottom were measured by divers at the beginning of the deployment. Data obtained using electromagnetic currentmeters and optical backscatter sensors represent average flow and sediment concentration conditions at a certain distance from the sensor (about 10–20 cm). Therefore, the small-scale spatial differences in the hydrodynamic structure and the suspended sediment concentration due to small-scale (o10– 20 cm) horizontal bottom features (such as the differential response on the crest and trough of ripples) cannot be detected with these sensors. Both tripods worked in burst-mode by recording data at a frequency of 2 Hz during a burst’s duration of 20 min every 3 h. The data analysed here correspond to the measurements acquired from 7 April 1997 at 12:00 (hour 0) to 8 April at 15:00 (hour 27), i.e. the initial period of the deployment. Recorded velocity components were projected onto a reference system aligned with the shoreline in such a way that the so projected u-component was perpendicular to the shoreline (positive seaward) and the v-component was alongshore directed (positive towards the northeast). The Ebro River water discharge was lower than 250 m3/s during the study period. Under these conditions the suspended sediment concentration in the lower river course is lower than 15 mg/l (Guille! n and Palanques, 1992) and the potential
308
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Fig. 1. Study area (A) and tripod locations across the Ebro inner shelf (B).
influence of the river sediment supplies into the analysed data can be neglected.
2.2. Wave parameters Wave conditions during the monitored period were calculated from subsurface pressure and velocity time series. The wave direction was estimated from the velocity time series by using the method proposed by Madsen et al. (1993). The
representative orbital velocity, ubm ; was calculated from the demeaned velocity time series after projection onto the mean wave direction, and it was calculated as the root-mean-square velocity, i.e. pffiffiffi ubm ¼ 2 * su ; ð1Þ where su is the standard deviation of the projected orbital velocities. The ubm was estimated from time series recorded at the uppermost currentmeter in each tripod. The representative wave period was
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
computed as the centroid of the velocity spectrum (after projection onto the mean wave direction), T; T ¼ m0 =m1 ;
ð2Þ
where m0 and m1 are the 0 and 1 order moments of the velocity spectrum. The root-mean-square wave height, Hrms ; was estimated from subsurface pressure time series, assuming the linear theory to be valid (see details in Jime! nez et al., 1999). 2.3. Friction velocity and bottom stress The relationship between current velocity, u; at a level z above the bottom and the shear velocity u * c in the current dominated portion of the near-bed bottom boundary layer is typically given by u ¼ ðu * c =kÞlnðz=z0a Þ;
309
tained. Output parameters of the model were the shear wave velocity, u * w ; the total shear velocity, u * wc and the bottom boundary layer thickness, dwc : The required kb necessary to obtain the ‘‘measured’’ u * c and z0a values can be also considered as an output of the model. Estimates of bottom roughness calculated by using the bottom boundary layer model and those obtained from ripple predictions were compared. The bottom roughness (kb ) was characterised as kb ¼ 4 * Z (Wikramanayake and Madsen, 1994), Z being the ripple height predicted by using the model of Wiberg and Harris (1994). Differences in kb were assumed to be due to additional roughness sources such as biological and suspended sediment stratification effects. 2.4. The suspended sediment concentration profile
ð3Þ
where k is the Von Karman constant (0.4) and z0a is the apparent roughness, which is the increased roughness scale felt by the current above the waveboundary layer caused by the wave-enhanced dissipation in the wave boundary layer (cf. Glenn and Grant, 1987). Values of u * c and z0a were obtained by fitting the measured velocities to the logarithmic model. Since only three measurement in the vertical were obtained, the consistency test proposed by Madsen et al. (1993) was applied to all the data, and only data fulfilling this test were retained. In addition to this, only profile fits with a coefficient of determination (r2 ) larger than 0.95 were accepted as velocity log-profiles. To fully characterise the structure of the nearbottom flow, the wave-current interaction bottom boundary layer model of Grant and Madsen (1979, 1982, 1986) was employed. Since measurements were taking in a natural environment with random waves, the modifications introduced in the model by Madsen (1993) were used. Values of ubm and T; the measured median grain size of the bottom sediment (d50 ), the current velocity and the angle between waves and currents were employed for running the model. The model was run iteratively until the required bottom roughness, kb ; necessary to get the fitted u * c and z0a values from the logarithmic velocity profile, was ob-
The time-averaged vertical distribution of the sediment concentration profile can be described by ws C þ es dc=dz ¼ 0;
ð4Þ
where ws is the settling velocity of sediment grains, C the mean suspended sediment concentration and es the turbulent eddy diffusivity or sediment mixing coefficient. The first term of the equation represents the downward transport of sediment by gravity, whereas the second term represents the upward transport by turbulence, which due to the adopting of the diffusion model is proportional to the vertical concentration gradient. The solution of Eq. (4) depends on the selected form of the vertical distribution of the mixing coefficient (see e.g. Van Rijn, 1993; Soulsby, 1997). Although many different vertical profiles of the mixing coefficient have been proposed (Soulsby et al., 1993; Williams et al., 1999), in this paper we used only two different methods, which in terms of the resulting mathematical expression are called the exponential and the power profiles. 2.3.1. Exponential concentration profile The exponential profile results from the assumption of a constant-in-vertical mixing coefficient in Eq. (4). It implicitly means that the water column is assumed to be well-mixed and the resulting
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
310
profile is given by C ¼ C0 expðzes =ws Þ;
ð5Þ
where C0 is the reference concentration at the bottom. Nielsen (1992) proposes the use of this profile but assuming that the dominant process is convective instead of diffusive. This means that if this approach is accepted, coherent structures in the fluid motion must exist to produce the upward transport of sediment particles. Under non-breaking waves, he assumes that the environment is mainly dominated by wave action and that the mixing length scale, L; depends on the ripple dimensions. Thus, Eq. (5) can be re-written as (Nielsen, 1992) C ¼ C0 expðz=LÞ;
ð6Þ
where L for rippled bottoms is given according to Nielsen (1992) by L ¼ k1 ubm Z=ws L ¼ k2 Z
for ubm =ws o18;
for ubm =ws > 18;
ð7aÞ ð7bÞ
k1 and k2 being proportionality constants with recommended values of k1 ¼ 0:075 and k2 ¼ 1:4 (Nielsen, 1992). By equating (5) and (6), we will obtain the sediment mixing coefficient by es ¼ Lws and then it can be written as es Dk1 ubm Z
for ubm =ws o18;
es Dk2 ws Z for ubm =ws > 18;
ð7a0 Þ ð7b0 Þ
2.3.2. Power concentration profile This profile results from the assumption of a sediment diffusivity varying linearly with elevation above the bottom, and it is usually employed in environments with combined waves and currents contributing both to the sediment mixing (see e.g. Glenn and Grant, 1987; Soulsby, 1997). In this kind of environment the sediment is suspended within the wave boundary layer and diffused further up into the flow by the turbulence associated with the current. The sediment diffusivity is defined as (Glenn and Grant, 1987) es ¼ gku * wc z es ¼ gku * c z
for z=dw o1 for z=dw > 1
ð8aÞ ð8bÞ
where g is a constant accounting for the differences in turbulent diffusion of water and sediment particles (here taken as 1). If we substitute (8) in (4) and after integration, the mean suspended sediment concentration C at a height z above the bottom is given by C ¼ Cðdw Þ ðz=dw Þgðws=ku * c Þ
for z > dw
ð9Þ
for zodw ;
ð10Þ
and C ¼ Cðz0 Þ ðz=z0 Þgðws=ku * wc Þ
z0 being the height of the bottom roughness, Cðdw Þ and Cðz0 Þ the reference concentration at the top of the wave boundary layer and at z0 respectively. 2.3.3. Reference concentration and resuspension coefficient The magnitude of the near-bottom reference concentration Cðz0 Þ is highly dependent on the reference level at which it is evaluated. For a matter of standardisation a common elevation has been selected to be used in the application of (6) and (10). Thus, in this study, the elevation for estimating the reference concentration was taken as 7 * d50 ; following Wikramanayake and Madsen (1994), who assumed this distance to be the thickness of the bedload layer. To calculate the reference concentration according to the exponential model, field data were fitted to (6), with C0 being obtained from the fit at z0 ¼ 7 * d50 : The following procedure was applied to calculate the reference concentration according to the power law: (i) since the three measurement points were above the wave boundary layer, they were fitted to (9) with known values of ws ; u * c and dw to obtain the concentration at the top of the boundary layer Cðdw Þ ; (ii) the thus fitted Cðdw Þ was introduced in Eq. (10) and with known values of u * wc and z0 ; the Cðz0 Þ value was calculated. Finally, once the reference concentration had been determined, the resuspension coefficient, g0; was calculated following the approach of Smith and McLean (1977). They propose that the reference concentration at the bottom, Cðz0 Þ ; can be expressed as a function of the excess of shear
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
stress as 0
Cðz0 Þ ¼ Cb
g0 S ; 1 þ g0 S 0
ð11Þ
where Cb is the maximum volumetric concentration at the bottom (0.65), S 0 is the normalised excess of shear stress above the critical one, and S0 ¼ ðt0 tcr Þ=tcr ; is tcr the critical shear stress to mobilise the sediment. If it is assumed that g0 S 0 51; then Eq. (11) can be re-written as Cðz0 Þ ¼ Cb g0 S0 :
ð12Þ
3. Results 3.1. Meteorological and wave conditions Meteorological conditions during the studied period were typical of the development of a northeast storm (Gregal). At the initial stage, weak winds were blowing from the southwest with a mean velocity of 2 m/s to rapidly veer anticlockwise towards the east, reaching a stable situation after 12 h with a wind speed of 16 m/s and with the direction stabilised at ENE, developing a Gregal storm (Fig. 2). The atmospheric
311
pressure displayed an increasing trend during the period from 1026 to 1035 mbar (Fig. 2). Easterly offshore winds caused a fast generation of wind waves that propagated in the wind direction. The orbital velocity near the bottom, ubm ; increased at both field sites from less than 0.08 m/s at the beginning of the experiment to maximum values of 0.47 m/s (Fig. 3). Wave heights ranged from 0.2 to 1.6 m and the wave period ranged from 5.9 to 6.7 s. The direction of the wave approach became progressively more perpendicular to the shoreline and the angle between the wave direction and the shoreline was smaller at the shallower field site than at the outer site (Fig. 3). Wave-related parameters (orbital velocity or shear stress) increased shoreward and, as expected, they showed the same temporal behaviour at both sites (Figs. 3 and 4). 3.2. Current Mean 0.22 m/s 0.32 m/s directed half-day
current speed ranged between 0.05 at the outer site and between 0.02 at the inner site (Fig. 5). The flow towards the northeast during the and progressively turned towards
Fig. 2. Temporal evolution of wind speed, wind direction and atmospheric pressure during the study period.
and and was first the
312
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Fig. 3. Temporal evolution of orbital velocity and the wave-current angle during the study period.
Fig. 4. Temporal evolution of the ‘‘skin’’ wave stress during the study period and critical values for sediment mobilisation for different grain sizes.
As the wind continued blowing, a clear increase in near-bottom currents was detected and a welldeveloped log-velocity profile was observed from hour 6 onwards. From hour 6 to hour 9 the current shear velocity u * c rapidly increased from initial values smaller than 0.010 m/s to 0.019 and to 0.015 m/s at the inner and outer field sites respectively (Fig. 5). Maximum u * c values were simultaneously recorded at both sites at hour 21 (0.039 and 0.030 m/s respectively) under a steady strong blowing wind. Comparing current shear velocities at the two sites, a consistent larger value was found at the inner location in all the times (about 27% larger). This increase must mainly reflect the differences in current intensity and water depth at the two sites. 3.3. Bottom sediment and roughness
southeast as a consequence of the change in wind direction. The alongshore component of the mean current was always dominant and it was mainly directed southwards, whereas the cross-shore component was always directed offshore. During the first six hours of the period analysed no well-defined vertical log-profile was developed, since current velocities were very weak and the wind-induced circulation was not yet developed.
The median grain size of the bottom sediment was 135 and 105 mm at the shallower and deeper field sites respectively at the beginning of the deployment. The mud content ranged between 7% and 20% of the total sediment. No direct observations of bottom roughness were available during the monitored period, although visual observations and photographs made at the begin-
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
313
Fig. 5. Temporal evolution of the mean current speed and shear velocity at both tripod locations during the study period.
ning of the deployment (4 April) indicate the presence of ripples with a height of about 1 cm and a wavelength of about 10 cm. From 4 April (the instant of visual observations) to the beginning of the study period (7 April) waves and current were of very low intensity, so no suspended sediment activity was detected. Therefore, we can assume that bottom roughness conditions observed on 4 April are representative of the bed conditions at the beginning of the study period. To assess bottom roughness evolution during the study, the model of Wiberg and Harris (1994)
was applied to recorded wave conditions and the predicted ripple dimensions are shown in Fig. 6a. Ripple height would increase from hour 0 (situation of ‘‘no movement’’ of sediment) to hour 15, with maximum heights of about 1.6 and 1.3 cm at the inner and outer sites respectively. These heights are equivalent to that of the observed ripples at the beginning of the deployment (about 1 cm) and the differences in height at the two locations reflect the differences in sediment size and near-bottom waveinduced velocities. From hour 15 to the end of measurements, the predicted ripple height decreased progressively as wave-induced velocity
314
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Fig. 6. (A) Temporal evolution of the ripple height estimated using the Wiberg and Harris (1994) model at both tripod locations during the study period. (B) Comparison between ripple height estimated using Wiberg and Harris (1984) and the bed roughness calculated from Grant and Madsen (1986) (unfilled symbols indicate the bed roughness when a stratification correction is introduced in the model).
increased and ripples evolved from orbital-type to anorbital ones. 3.4. Wave-current interaction Values of u * c ; u * w ; u * wc (and their skin components) were computed with the bottom boundary layer wave-current model of Grant and Madsen (1979) and Madsen (1993). The value of wave-related parameters tends to increase shoreward, showing a similar temporal variation at the two field sites, as expected. For instance, the skin wave shear stress showed a sharp increase simultaneously at both sites at hour 3 of the record, and afterwards values at the inner site were nearly twice those at the deeper site (Fig. 4), reflecting the difference in near-bottom wave induced velocity. The Grant and Madsen model was run inversely to estimate the required kb to produce the fitted u * c values (from the log-profile); the kb values obtained were similar to that predicted with the model of Wiberg and Harris (1994) (Fig. 6b). However, in the two last bursts, when the largest suspended sediment concentration were recorded, an unrealistic large roughness was required to reproduce the large fitted u * c values. The ripple heights needed to reproduce the u * c values were about 10 cm. These values must indicate that there existed something in the bottom boundary layer
that increased the roughness artificially. To analyse this, these two bursts were fitted to a velocity profile corrected for the presence of stratification due to gradients in sediment concentration (Jime! nez et al., in preparation). The inclusion of this correction resulted in much lower bottom roughness values, in agreement with those obtained from the ripple formation model (Fig. 6b). 3.5. Suspended sediment The suspended sediment concentration was below the resolution limit of OBS sensors in both deployments at the beginning of the experiment (hour 0). From hour 3 onwards, the suspended sediment concentration in the water column presented an increasing trend (Fig. 7). The maximum recorded suspended sediment concentration, at 10 cm above the bottom, was about 7 g/l at the 8.5 m water depth site at hour 27. Simultaneously, the concentration measured at the outer site, at the same height above the seabed, was 3.3 g/l. The initiation of movement of sediment observed in the suspended sediment concentration was compared to that predicted using the Shields criterion according to Madsen and Grant (1976). Thus, using as an input the near-bottom wave induced velocity, the skin wave stress t0w exceeds
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
315
alongshore towards the south and offshore. However, the content of fine sediment in the bottom was small (less than 5% at 8.5 m depth) and the sediment in suspension discharged from the Ebro river was undetectable (o15 mg/l) during this period. Therefore, it can be considered that advection of fine sediment was not relevant in the measured concentration profile. 3.6. Exponential profile
Fig. 7. Temporal evolution of the suspended sediment concentration in both tripods at three different heights above the bottom during the study period.
the threshold criterion, tcri ; just after the first burst at both sites, in agreement with the observed sediment response (Fig. 4). The first maximum in sediment concentration was reached at hour 6 and 9 at the inner and outer locations respectively, followed by a decrease in concentration and finally a continuous increase in concentration during the last part of the study period. On the Ebro inner shelf, the distribution of the suspended sediment can be affected by advective transport of fine sediment (Jime! nez et al., 1999). This advection could be suggested by the continuous increasing of the concentration at both sites during the second part of the observation period, despite the fact that the skin wave shear stress does not significantly increase at the deeper site (Fig. 4). The advection could occur from the inner to the outer site because the current was directed
Measured vertical profiles of suspended sediment concentration were fitted to the exponential profile (6) and in all the bursts a coefficient of determination, r2 ; greater than 0.8 was obtained in the fit (Fig. 8). These results suggest that the use of a simple exponential law to characterise the shape of the near-bottom vertical profile of the suspended sediment concentration is accurate enough from the empirical standpoint (based on the goodness of the fit). Sediment diffusivity coefficients obtained in the fit vary between 2 and 7 103 m2/s with higher values at the inner location (Table 2). These values show some relation to the gradient of suspended sediment concentration (Fig. 9). If the sediment diffusivity is predicted using Eq. (7b0 ) (since ubm =ws > 18) and the ripple height calculated from the Wiberg and Harris model, es -values range between 6 105 and 3 104 with an average value of about 1 104. This value is one order of magnitude lower than the es calculated from field measurements, suggesting that the constant k2 in Eq. (7b0 ) must be higher than the value considered. Thus, differences between observed and calculated es would suggest that the assumption in Eq. (6) that the concentration profile is only controlled by the interaction of the oscillatory flow with ripples is not suitable for this data set. 3.7. Power profile The second distribution law considered for the vertical profile of suspended sediment was the Glenn and Grant (1987) model (Eq. (9)). The recorded data were fitted to the model by assuming a constant and known settling velocity of the sediment at each site that corresponds to the d50 of
316
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Fig. 8. Measured (dots) and fit of exponential (dashed line) and power (discontinuous line) of the suspended sediment concentration profiles for both tripod locations: (a) shallower site; (b) deeper site.
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Fig. 8. (continued)
317
318
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Table 2 Gradients of selected parameters across the Ebro inner shelf considering average values during the study period Parameter
12.5 m depth
8.5 m depth
Difference
d50 bottom sediment (mm) d50 suspended sediment (mm) Orbital velocity (m/s) Skin wave shear stress (N/m2) Excess wave shear stress Friction velocity (m/s) Reference concentration (g/l) Resuspension coefficient Ripple height (m) Apparent roughness (z0a ) (m) Sediment diffusivity (m2/s)
105 111 0.302 0.335 1.39 0.022 2.00 3.8 103 0.0076 0.032 3 103
135 117 0.386 0.552 2.45 0.028 6.91 5.4 103 0.0106 0.023 6 103
+28% +5% +28% +65% +76% +27% +245% +42% +28% 28% +100%
Fig. 9. Sediment mixing coefficient obtained from fitting Eq. (7) with the measured suspended sediment concentration and gradients of suspended sediment concentrations between the lowest and highest sensors for both tripod locations (dots: shallower site; circles: deeper site).
the bottom sediment (135 and 105 m at the inner and outer sites respectively). The shear velocities introduced in Eq. (9) were the ones estimated from the logarithmic velocity profiles (from 9 to 27 h), so the only free variable in Eq. (9) was Cðdw Þ : Fig. 8 shows this fit, where the coefficient of determination, r2 ; was always larger than 0.87. Therefore,
adjustment of field data to the model is correct in terms of goodness of the fit and no significant improvement was obtained by considering different values for constants (g ¼ 0:74 in Eq. (9)) or by introducing stratification effects in computations. It is clear from Eqs. (9) and (10) that ws is an important parameter for controlling the value of
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
319
Fig. 10. Particle size of suspended sediment estimated from the power profile of suspended sediment concentration.
the Rouse number (=ws =ku * ), although it was not directly measured during the experiment. To account for its importance, the settling velocity was considered as a second free variable in Eq. (9), together with the reference concentration, Cðdw Þ : This also made it possible to analyse the expected variability of the suspended sediment grain size during the survey. Including ws as an additional variable, the fit improved as expected because we increased the degrees of freedom of the fit (r2 > 0:99 for all the cases). The theoretical values of ws varied following the same pattern at both sites, with the sediment coarser at the end of the deployment (Fig. 10). This apparent change in the suspended sediment grain size must be related to the increasing bottom shear velocity during the study. If a simple criterion for sediment suspension by currents is considered (ws ou * c ; Bagnold, 1966), it is apparent from Fig. 5 that friction velocities at the beginning of the deployment were insufficient to maintain sand particles in suspension and only silt particles fulfil the suspension criterion. However, wave-induced shear stress may resuspend both silt and sand particles from the second burst to the end of measurements (Fig. 4). Consequently, wave activity supplied silt and sand fractions to the water column, but only silt could be maintained in suspension by currents, whereas the sand fraction was probably deposited on the bottom during the first part of the study period. The current friction velocity increased significantly
Fig. 11. Reference concentration C0 at 7 * d50 calculated from: (a) considering an exponential profile (Eq. (6)), with constant L (=0.45) and C0 ¼ 0:005y3r ; yr being the effective Shields parameter defined by Nielsen (1986); (b) Eq. (12), considering a constant g0 (=3 103); (c) fitting the exponential profile to field data; and (d) fitting the power profile to field data.
6–9 h after the beginning of the storm and maintained most of the sediment grain size in suspension. 3.7.1. Reference concentration and resuspension coefficient The reference concentration, C0, was estimated by fitting the data of suspended sediment concentration to the exponential and power profiles (Fig. 11). The C0 displays high spatial and temporal variability, although it is higher at the inner site and tends to increases during the study period. Typical values of C0 range from 7 to 35 g/l and from 3 to 20 g/l at the inner and outer site
320
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Fig. 12. (a) Temporal evolution of the resuspension parameter at both tripod locations during the study period; and (b) relations between the resuspension parameter and the excess ‘‘skin’’ wave shear stress.
respectively. The value of g0 ranges between 2.1 103 and 7.2 103, with mean values of 3.8 103 and 5.4 103 for the outer and inner sites respectively (Fig. 12a).
4. Discussion 4.1. Mechanisms for resuspension on the Ebro inner shelf The hydrodynamic conditions required to entrain the sediment across the inner shelf of the Ebro delta were calculated by applying the Shields criterion according to Madsen and Grant (1976). The results show that at the outer location (12.5 m depth) the critical shear stress, tcr ; was about 0.14 N/m2 (d50 ¼ 105 mm), whereas that associated with the inner location (8.5 m depth) was about 0.16 N/m2 (d50 ¼ 135 mm) (Fig. 4). By assuming that such stresses will only be wave-induced and converting them to wave characteristics for typical wave conditions in the Ebro shelf (mean Hrms ¼ 0:5; T ¼ 4 s), we can estimate theoretical minimum waves to stir the bottom sediment. Considering wave periods ranging from 4 to 6 s, the minimum Hrms to entrain the bottom sediment varied between 0.55–1 and 0.75–2.5 m at the shallower and deeper sites respectively. Therefore, the inner shelf would be inactive as regards resuspension processes under relatively fair-weather conditions.
This indicates that during a large part of the climatic year the inner shelf is inactive as regards sediment resuspension, which can be considered as the regular situation for inner shelves in areas where short period waves are dominant. This agrees with previous theoretical studies that noted that near-bottom wave-induced sediment transport on the Ebro inner shelf was only significant during storm conditions (10% of the incoming waves during the year induce about 95% of the potential transport) (Jime! nez et al., 1997). Another possibility for entraining the sediment would be currents inducing the required threshold shear stress. In this case, the current skin stress exceeding the previously estimated critical shear stress must correspond to currents at 1 m above the bottom of 0.36 and 0.33 m/s for the inner and outer locations respectively. According to currents previously measured at this elevation in the study area (Jime! nez et al., 1999; Palanques et al., 2002), they are not likely to occur on this inner shelf region. On the other hand, if actual wave and bottom roughness effects are incorporated into the computations and the current skin friction is calculated for the measured conditions, it is below the critical stress for sediment resuspension at both field sites. This is the other typical situation for this inner shelf, and in general for microtidal shelves, where currents are usually not strong enough to entrain the bottom sediment by themselves for most of the time.
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
4.2. Gradients across the shelf Sediment dynamics are controlled by the interaction of different processes that change their intensity and characteristics across the shelf. Table 2 shows gradients of different parameters measured across the Ebro inner shelf, considering their average value for the whole observation period. Most of the parameters increase in the onshore direction, although with different intensities. For instance, the skin wave stress, the excess shear stress, the reference concentration of suspended sediment, the resuspension coefficient and the sediment diffusivity almost double their value between 12.5 and 8.5 m water depth. The wave orbital velocity, the current friction velocity and the grain size of the bottom sediment show a more moderate increase onshore (about 25%). Finally, some parameters such as the apparent roughness (28%) decay towards the shore, as was expected, due to the increase in near-bottom wave induced velocities. The spatial variability of different parameters can be considered representative of the Ebro inner shelf for the measured conditions (from calm weather to mid-storm conditions). In general, the observed variability corresponds to the increasing wave and wind-induced current intensity and the coarsening of the sediment towards the coast. The increase in these parameters (about 25% between 12.5 and 8.5 m depth) causes a major increase (about 100%) in the skin wave stress and the parameters related to the sediment resuspension. Consequently, moderate changes in wave characteristics can lead to major changes in the concentration of suspended sediment and fluxes of sediment across the shelf. 4.3. Concentration profile and resuspension coefficient The vertical profile of suspended sediment is governed by a number of interacting variables, including wave and current parameters and sediment and bottom properties. The knowledge of this profile is required for the estimation of the sediment transport near the bottom. Since only limited field information is usually available,
321
several models have been developed to predict the concentration profile using a number of parameters (Soulsby et al., 1993; van Rijn, 1993; Soulsby, 1997). The reference concentration at a fixed distance above the bottom is one of the most critical parameters in these models. The reference concentration is not usually measured directly but estimated by extrapolation of the concentrations measured close to the bottom. Due to this, its value largely depends on the number and quality of the field measurements and the law used to make the extrapolation. To illustrate this, Fig. 11 shows the reference concentration estimated by fitting the field data to the two models employed in this work at the same elevation. In statistical terms, they were empirically equally good to explain the vertical distribution of suspended sediment concentration. However, the differences in the calculated C0 are relatively large, especially at the inner site (Fig. 11). These differences are due to the different shape of the profile very close to the bottom, which in the case of the power profile predicts an infinite concentration at z ¼ 0 m, resulting in a larger concentration close to the bottom. This sensitivity of C0 to the model used is directly transferred to differences in g0 (Eq. (12)), and it can be considered as one of the sources of g0 variability (Fig. 12a). The estimated values are in the range of other values reported in the literature (Table 1). The mean g0 during the study period is close to the ‘‘practical’’ boundary suggested by Madsen et al. (1994) for rippled (g0 ¼ 2 103 ) beds. The validity of the resuspension coefficient estimated for wave-dominated conditions following Smith and McLean (1977) has been questioned (Webb and Vincent, 1999). Several studies have established that the resuspension coefficient is dependent on the excess shear stress (S) (e.g. g0 =kS n ) (Drake and Cacchione, 1989; Vincent et al., 1991; Li et al., 1996; Green and Black, 1999). Including this type of relation in Eq. (12) would suggest that the reference concentration would be proportional to the excess skin friction (C0 Dconstant * S n ), as suggested by Nielsen (1992). Webb and Vincent (1999) pointed out that g0 decreases and C0 increases with increasing S:
322
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
However, the Smith and McLean model only makes sense for a fixed value of g0 : Consequently, these authors consider that this model is not appropriate for predicting the reference concentration over bedforms in a wave-dominated environment. The exponential model of Nielsen (1986) is theoretically more suitable for these conditions, although the comparison between measured and predicted concentrations can display very significant differences (Webb and Vincent, 1999). On the Ebro shelf, the resuspension coefficient seems to decrease with increasing S; although a significant relation between the two parameters is not detected. In addition to this, the resuspension coefficient is lower at the inner site, where the excess wave shear stress is higher (Fig. 12a). This means that in the relation g0 ¼ kS n ; the ‘‘constant’’ k is lower at the inner site or, in other words, the constant k has spatial variation and it is characteristic of each site. This is probably related to the control of the physical roughness (ripples) in the resuspension coefficient, which has been extensively used in this kind of approach (see Table 1). Fig. 12b shows the increase in the reference concentration value with the excess shear stress when the mean g0 is considered. The mean values of g0 obtained here can be useful in studies where only an approximation to the value of the reference concentration is desired. Therefore, very detailed measurements of the vertical profile of the sediment concentration are still required to obtain realistic estimations of the resuspension coefficient or the sediment diffusivity coefficient. Fig. 13 displays the comparison between the measured concentration at 0.98 m above the bottom and the predicted concentration using the exponential and power model, considering the average values of es and g0 estimated at each site. The concentration predicted for both models is of the same order of magnitude as measured values during the entire study period. In general, models underpredict and overpredict the concentration at the inner and outer sites respectively. In spite of these differences, the comparison indicates that we can obtain a semi-quantitative approach to the suspended sediment concentration from both models using the estimated average values of es and g0 and typical wave/current parameters.
Fig. 13. Suspended sediment concentration at 0.98 m above the bottom at both tripod locations: (a) measured; (b) estimated from exponential profile considering the mean sediment diffusivity at each site; and (c) estimated from the power profile considering the mean resuspension parameter at each site.
However, the variation of the es and g0 at each site with changing wave-current conditions and variations in the bottom morphology remains unsolved for this Mediterranean inner shelf.
5. Conclusions Sediment dynamics on the bottom boundary layer display a fast and quasi-simultaneous response across the Ebro inner shelf to changes in hydro-meteorological conditions. The wave shear stress affecting the bottom sediment exceeded threshold conditions for resuspension a few hours after an easterly wind began to blow and caused the waves to develop. The resuspension of bottom
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
sediment started simultaneously at 8.5 and 12.5 m water depth, although the suspended sediment concentration and energetic wave conditions were higher at the shallower site. Most parameters showed significant changes between 12.5 and 8.5 m water depth. The ‘‘skin’’ wave shear stress, the reference concentration of suspended sediment and the resuspension coefficient displayed maximum positive gradient (increasing onshore), whereas the apparent roughness decreased in an onshore direction. These data indicate that the inner shelf is an area of strong gradients, which should be taken into account for understanding the sediment dynamics across the shelf. The estimation of the reference concentration of suspension profiles from field measurements is still an unsolved problem, which largely depends on the data characteristics and the method employed. We obtained good fits (r2 > 0:9) for our suspension profiles using both exponential and power profiles. Significant spatial and temporal variability of the sediment diffusivity and resuspension coefficient were observed across the inner shelf. These variations reflect spatial changes in the ‘‘skin’’ friction shear stress, the grain size and the morphology of the bottom sediment, and they should be considered for modelling sediment transport across the shelf. The average resuspension coefficient estimated for the Ebro inner shelf is about 5.4 103 and 3.8 103 for the inner and outer site respectively, and they fall near the typical range of values measured in different areas. Average values of the sediment diffusivity coefficient are 6 103 and 3 103 at the inner and outer sites respectively. Using these values, an adequate approximation to the concentration profile is obtained for a semi-quantitative analysis. This would imply that each location across the inner shelf could be characterised by a ‘‘typical’’ or mean sediment diffusivity and resuspension parameter. However, the temporal variability observed in these parameters associated with both bottom morphology and excess shear stress changes has not been clearly explained during the study period and should be considered for a more accurate modelling of the concentration profile.
323
As a final conclusion it can be stated that sediment dynamics in microtidal, low-energy (seadominated) continental shelves is largely controlled by wave-storm events. Sediment resuspension is caused by the wave-storm activity and it mostly affects the shallow part of the shelf, whereas the deeper areas of the shelf are hardly influenced by wave action. This is a significant difference from shelves affected by strong tidal currents and/or ‘‘swell’’ waves, where high-intensity sediment resuspension processes can occur across the entire shelf and are caused by both unidirectional and oscillatory flows.
Acknowledgements This work was carried out in the framework of the FANS and TRASEDVE projects funded by the EU (MAS3-CT95-0037) and CICYT (MAR98-0691-C02-01) respectively. Additional support was also given by CICYT (MAR962268-C03-01-CE). We appreciate the critical review of the manuscript from three anonymous referees. This is a contribution to the IGCP project No. 396: Continental Shelves in the Quaternary.
References Bagnold, R.A., 1966. An approach to the sediment transport problem from general physics. Geological Survey Prof. Paper 422-I, Washington, USA, 37p. Bedford, K.W., Lee, J., 1994. Near-bottom sediment response to combined wave-current conditions, mobile bay, gulf of Mexico. Journal of Geophysical Research 99 (C8), 16 161– 16 177. D&A Instruments, 1991. OBS 1 & 3. Suspended Solids & Turbidity Monitor. Instruction Manual, 41pp. Delft Hydraulics, 1993. P-EMS. Programmable Electronic Liquid Velocity Meter. User’s Manual. Drake, D.E., Cacchione, D.A., 1989. Estimates of the suspended sediment reference concentration (Ca) and resuspension coefficient (g0) from near-bottom observations on the california shelf. Continental Shelf Research 9, 51–64. Glenn, S.M., Grant, W.D., 1987. A suspended sediment stratification correction for combined wave and current flows. Journal of Geophysical Research 92, 8244–8264. Grant, W.D., Madsen, O.S., 1979. Combined wave and current interaction with a rough bottom. Journal of Geophysical Research 84 (C4), 1797–1808.
324
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen
Grant, W.D., Madsen, O.S., 1982. Moveable bed roughness in unsteady oscillatory flow. Journal of Geophysical Research 87, 469–481. Grant, W.D., Madsen, O.S., 1986. The continental-shelf bottom boundary layer. Annual Review of Fluid Mechanics 18, 265–305. Green, M.O., Black, K.P., 1999. Suspended-sediment reference concentration under waves: field observations and critical analysis of two predictive models. Coastal Engineering 38, 115–141. Green, M.O., Vincent, C.E., McCave, I.N., Dickson, R.R., Rees, J.M., Perason, N.D., 1995. Storm sediment transport: observations from the British North Sea shelf. Continental Shelf Research 15 (8), 889–912. Guill!en, J., Palanques, A., 1992. Sediment dynamics and hydrodynamics in the lower course of a river highly regulated by dams: the Ebro river. Sedimentology 39, 567–579. Guill!en, J., Palanques, A., 1997. A shoreface zonation in the Ebro Delta based on grain size distribution. Journal of Coastal Research 13 (3), 867–878. Hill, P.S., Nowell, A.R.M., Jumars, P.A., 1988. Flume evaluation of the relationship between suspended sediment concentration and excess boundary shear stress. Journal of Geophysical Research 94, 12 499–12 509. Jago, C.F., Bale, A.J., Green, M.O., Howarth, M.J., Jones, S.E., McCace, I.N., Millward, G.E., Morris, A.W., Rowden, A.A., Williams, J.J., 1993. Resuspension processes and seston dynamics, southern North Sea. Philosophical Transactions of the Royal Society of London A343, 475–491. Jim!enez, J.A., S!anchez-Arcilla, A., Valdemoro, H.I., Gracia, V., Nieto, F., 1997. Processes reshaping the Ebro delta. Marine Geology 144, 59–79. Jim!enez, J., Guill!en, J., Gracia, V., Palanques, A., Garc!ıa, M., S!anchez-Arcilla, A., Puig, P., Puigdefabregas, J., Rodr!ıguez, G., 1999. Water and sediment fluxes on the Ebro delta shoreface. On the role of low frequency currents. Marine Geology 157, 219–239. Jones, S.E., Jago, C.F., Bale, A.J., Chapman, D., Howland, R.M.J., Jackson, J.J., 1998. Aggregation and resuspension of suspended particulate matter at a seasonally stratified site in the southern North Sea: physical and biological controls. Continental Shelf Research 18, 1283–1309. Li, M.Z., Amos, C.L., 1998. Predicting ripple geometry and bed roughness under combined waves and currents in a continental shelf environment. Continental Shelf Research 18, 941–970. Li, M.Z., Wright, L.D., Amos, C.L., 1996. Predicting ripple roughness and sand resuspension under combined flows in a shoreface environment. Marine Geology 130, 139–161. Madsen, O.S., 1993. Sediment Transport on the Shelf. Lecture Notes, Ralph M. Parsons Laboratory for Water Resources and Hydrodynamics, Department of Civil Engineering, Massachusetts Institute of Technology, Cambridge, USA, 147pp. Madsen, O.S., Grant, W.D., 1976. Sediment transport in the coastal environment. Report No. 209, Ralph M. Parsons
Laboratory for Water Resources and Hydrodynamics, Department of Civil Engineering, Massachusetts Institute of Technology, Cambridge, USA, 105pp. Madsen, O.S., Wright, L.D., Boon, J.D., Chisholm, T.A., 1993. Wind stress, bed roughness and sediment suspension on the inner shelf during an extreme storm event. Continental Shelf Research 13, 1303–1324. Madsen, O.S., Chisholm, T.A., Wright, L.D., 1994. Suspended sediment transport in inner shelf waters during extreme storms. Proceedings of the 24th Coastal Engineering Conference, ASCE, pp. 1849–1864. Niedoroda, A.W., Reed, C.W., Swift, D.J.P., Arato, H., Hoyanagi, K., 1995. Modeling shore-normal large-scale coastal evolution. Marine Geology 126 (1/4), 181–200. Nielsen, P., 1986. Suspended sediment concentrations under waves. Coastal Engineering 10, 23–31. Nielsen, P., 1992. Advanced Series on Ocean Engineering. Vol. 4, World Scientific, Singapore, 324pp. Palanques, A., Puig, P., Guill!en, J., Jim!enez, J., Gracia, V.S!anchez-Arcilla, A., Madsen, O., 2002. Near bottom suspended sediment fluxes on the microtidal low energy Ebro continental shelf (NW Mediterranean) Continental Shelf Research 22, 285–303. Reed, C.W., Niedoroda, A.W., Swift, D.J.P., 1999. Modeling sediment entrainment and transport processes limited by bed armoring. Marine Geology 154, 143–154. Smith, J.D., 1977. Modelling of sediment transport in continental shelves. In: Goldberg, E.D., et al. (Ed.), The Sea, Vol. 6. Wiley, New York, pp. 539–577. Smith, J.D., McLean, S.R., 1977. Boundary layer adjustments to bottom topography and suspended sediment. In: Nihoul, J.C.J (Ed.), Bottom Turbulence. Elsevier, New York, pp. 123–151. Soulsby, R., 1997. Dynamics of Marine Sands. A Manual for Practical Applications. Thomas Telford, London, 249pp. Soulsby, R., Hamm, L., Klopman, G., Myrhaug, D., Simons, R.R., Thomas, G.P., 1993. Wave-current interaction within and outside the bottom boundary layer. Coastal Engineering 21, 41–69. Van Rijn, L.C., 1993. Principles of Sediment Transport in Rivers, Estuaries and Coastal Seas. Aqua Publications, Amsterdam, 638pp. Vincent, C.E., Downing, A., 1994. Variability of suspended sand concentrations, transport and eddy diffusivity under non-breaking waves on the shoreface. Continental Shelf Research 14, 223–250. Vincent, C.E., Green, M.O., 1990. Field measurements of the suspended sand concentration profiles and fluxes, and of the resuspension coefficient over a rippled bed. Journal of Geophysical Research 95, 15 591–15 601. Vincent, C.E., Hanes, D.M., Bowen, A.J., 1991. Acoustic measurements of suspended sand on the shoreface and the control of concentration by bed roughness. Marine Geology 96, 1–18. Webb, M.P., Vincent, C.E., 1999. Comparison of time-averaged acoustic backscatter concentration profile measurements with existing models. Marine Geology 162, 71–90.
! et al. / Continental Shelf Research 22 (2002) 305–325 J. Guillen Wiberg, P.L., Harris, C.K., 1994. Ripples geometry in wavedominated environments. Journal of Geophysical Research 99, 775–789. Wiberg, P.L., Smith, J.D., 1983. A comparison of field data and theoretical models for wave-current interaction at the bed on the continental shelf. Continental Shelf Research 2, 147–162. Wiberg, P.L., Drake, D.E., Cacchione, D.A., 1994. Sediment resuspension and bed armoring during high bottom stress events on the northern California inner continental shelf: measurements and predictions. Continental Shelf Research 14 (10/11), 1191–1219. Wikramanayake, P.N., Madsen, O.S., 1994. Calculation of suspended sediment transport by combined wavecurrent flows. Contract Report DRP-94-7, US Army Engineers Waterways Experiment Station, Vicksburg, MS.
325
Williams, J.J., Rose, C.P., Thorne, P.D., O’Connor, B.A., Humphery, J.D., Hardcastle, P.J., Moores, S.P., Cooke, J.A., Wilson, D.J., 1999. Field observations and predictions of bed shear stresses and vertical suspended sediment concentration profiles in wave current conditions. Continental Shelf Research 19, 507–536. Wright, L.D., Schaffner, L.C., Maa, J.P.-Y., 1997. Biological mediation of bottom boundary layer processes and sediment suspension in the lower Chesapeake bay. Marine Geology 141, 27–50. Wright, L.D., Kim, S.-C., Friedrichs, C.T., 1999. Across-shelf variations in bed roughness, bed stress and sediment suspension on the northern california shelf. Marine Geology 154, 99–115. Zyserman, J., Fredse, J., 1994. Data analysis of bed concentration of suspended sediment. Journal of Hydraulic Engineering 120 (9), 1021–1042.
