• Betteridge, K. F. E., and Souza A. J. , 2008: Estimates of the turbulent kinetic energy (TKE) balance in shelf seas. Proc. 2008 PECS Conf., Liverpool, United Kingdom, Proudman Oceanographic Laboratory, 301–304.

    • Search Google Scholar
    • Export Citation
  • Binder, M. A., Mingham C. G. , Causon D. M. , Rahmati M. T. , Aggidis G. A. , and Chaplin R. V. , 2009: Numerical modelling of a surging point absorber wave energy converter. Proc. Eighth European Wave and Tidal Energy Conf., Uppsala, Sweden, EWTEC. [Available online at http://www.flow3d.com/pdfs/tp/mar_tp/FloSci-Bib32-09.pdf.]

    • Search Google Scholar
    • Export Citation
  • Bolaños, R., and Souza A. J. , 2010: Measuring hydrodynamics and sediment transport processes in the Dee Estuary. Earth. Syst. Sci. Data, 2, 157165.

    • Search Google Scholar
    • Export Citation
  • Bolaños, R., Souza A. J. , Bell P. S. , and Malarkey J. , 2009: Measuring tidal hydrodynamics and bedform evolution in the Dee estuary. River, Coastal and Estuarine Morphodynamics, C. A. Vionnet et al., Eds., CRC Press, 709–714.

    • Search Google Scholar
    • Export Citation
  • Cacchione, D. A., Sternberg R. W. , and Ogston A. S. , 2006: Bottom instrumented tripods: History, applications and impacts. Cont. Shelf Res., 26, 23192334.

    • Search Google Scholar
    • Export Citation
  • Chang, K.-A., Hsu T.-J. , and Liu P. L.-F. , 2001: Vortex generation and evolution in water waves propagating over a submerged rectangular obstacle Part I. Solitary waves. Coastal Eng., 44, 1336.

    • Search Google Scholar
    • Export Citation
  • Chang, K.-A., Hsu T.-J. , and Liu P. L.-F. , 2005: Vortex generation and evolution in water waves propagating over a submerged rectangular obstacle Part II. Cnoidal waves. Coastal Eng., 52, 257283.

    • Search Google Scholar
    • Export Citation
  • Chorin, A. J., 1968: Numerical solution of the Navier–Stokes equations. Math. Comput., 22, 745762.

  • Dolphin, T., and Vincent C. , 2009: The influence of bed forms on reference concentration and suspension under waves and currents. Cont. Shelf Res., 29, 424432.

    • Search Google Scholar
    • Export Citation
  • Goring, D. G., and Nikora V. I. , 2002: Despiking acoustic Doppler velocimeter data. J. Hydraul. Eng., 128, 117126.

  • Hirt, C. W., and Nichols B. D. , 1981: Volume of fluid (VOF) method for the dynamics of free boundaries. J. Comput. Phys., 39, 201225.

    • Search Google Scholar
    • Export Citation
  • Hsu, T.-J., Sakakiyama T. , and Liu P. L.-F. , 2002: A numerical model for wave motions and turbulence flows in front of a composite breakwater. Coastal Eng., 46, 2550.

    • Search Google Scholar
    • Export Citation
  • Humphery, J. D., Jones D. S. , and Doyle K. , 2007: The CFD-aided design of STABLE III, a new platform to study near-bed turbulent currents and associated sediment dynamics. Proc. OCEANS 2007, Aberdeen, Scotland, IEEE/OES, 1–6.

    • Search Google Scholar
    • Export Citation
  • Lara, J. L., Garcia N. , and Losada I. J. , 2006: RANS modelling applied to random wave interaction with submerged permeable structures. Coastal Eng., 53, 395417.

    • Search Google Scholar
    • Export Citation
  • Lee, G., Friedrichs C. T. , and Vincent C. E. , 2002: Examination of diffusion versus advection dominated sediment suspension on the inner shelf under storm and swell conditions, Duck, North Carolina. J. Geophys. Res., 107, 3084, doi:10.1029/2001JC000918.

    • Search Google Scholar
    • Export Citation
  • Lin, P., and Liu P. L.-F. , 1998a: A numerical study of breaking waves in the surf zone. J. Fluid Mech., 359, 239264.

  • Lin, P., and Liu P. L.-F. , 1998b: Turbulence transport, vorticity dynamics and solute mixing under plunging breaking waves in surf zone. J. Geophys. Res., 103, 15 67715 694.

    • Search Google Scholar
    • Export Citation
  • Lin, P., and Liu P. L.-F. , 1999: Internal wave-maker for Navier–Stokes equations models. J. Waterw. Port Coastal Ocean Eng., 125, 207215.

    • Search Google Scholar
    • Export Citation
  • Sherwood, C. R., Lacy J. R. , and Voulgaris G. , 2006: Shear velocity estimates on the inner shelf off Grays Harbour, Washington, USA. Cont. Shelf Res., 26, 19952018.

    • Search Google Scholar
    • Export Citation
  • Shih, T.-H., Zhu J. , and Lumley J. L. , 1996: Calculation of wall-bounded complex flows and free shear flows. Int. J. Numer. Methods Fluids, 23, 11331144.

    • Search Google Scholar
    • Export Citation
  • Torres-Freyermuth, A., Losada I. J. , and Lara J. L. , 2007: Modeling of surf zone processes on a natural beach using Reynolds-averaged Navier–Stokes equations. J. Geophys. Res., 112, C09014, doi:10.1029/2006JC004050.

    • Search Google Scholar
    • Export Citation
  • Traykovski, P., 2007: Observations of wave orbital scale ripples and a nonequilibrium time-dependent model. J. Geophys. Res., 112, C06026, doi:10.1029/2006JC003811.

    • Search Google Scholar
    • Export Citation
  • Vasquez, J. A., and Walsh B. W. , 2009: CFD simulation of local scour in complex piers under tidal flow. Proc. 33rd IAHR Congress, Vancouver, BC, Canada, International Association of Hydraulic Engineering & Research, 978-94-90365-01-1.

    • Search Google Scholar
    • Export Citation
  • Wahl, T. L., 2003: Discussion of “Despiking acoustic Doppler velocimeter data.” J. Hydraul. Eng., 129, 484486.

  • Williams, J. J., Bell P. S. , Coates L. E. , Metje N. , and Selwyn R. , 2003: Interactions between a benthic tripod and waves on a sandy bed. Cont. Shelf Res., 23, 355375.

    • Search Google Scholar
    • Export Citation
  • Williams, J. J., Bell P. S. , Thorne P. D. , Metje N. , and Coates L. E. , 2004: Measurement and prediction of wave-generated suborbital ripples. J. Geophys. Res., 109, C02004, doi:10.1029/2003JC001882.

    • Search Google Scholar
    • Export Citation
  • Williams, J. J., Bell P. S. , and Thorne P. D. , 2005: Unifying large and small wave-generated ripples. J. Geophys. Res., 110, C02008, doi:10.1029/2004JC002513.

    • Search Google Scholar
    • Export Citation
  • Wolf, J., Souza A. J. , Bell P. S. , Thorne P. D. , Cook R. D. , and Pan S. , 2008: Wave, currents and sediment transport observed during the LEACOAST2 experiment. Proc. PECS 2008, Liverpool, United Kingdom, Proudman Oceanographic Laboratory, 373–376.

    • Search Google Scholar
    • Export Citation
  • Yakhot, V., and Orszag S. A. , 1986: Renormalization group analysis of turbulence. Basic theory. J. Sci. Comput., 1, 351.

  • View in gallery

    Sketch of the two-dimensional wave model setup showing (top) the entire computational domain and (bottom) a close-up around the obstacle corresponding to the region delimited by the dashed lines in the top panel. In the bottom panel, the dashed lines represent the location at which vertical profiles are determined.

  • View in gallery

    Conceptual representation of STABLE III from CADdata used in Flow3D.

  • View in gallery

    Location of rigs in the (left) Dee estuary (west) and (right) Sea Palling (east). The Dee estuary data are from 2008 and the Sea Palling data are from deployments during 2006–07. The color bar represents water depth (m); negative values in the Dee estuary represent wet areas above mean sea level.

  • View in gallery

    Vertical profiles of the intrawave velocity components: (left) U and (right) W (m s−1) at the mid-obstacle location. (top to bottom) The free surface location (m), velocities from the reference case, differences in velocities for the solid obstacle, and difference in velocities for the porous obstacle (note differences in color scales).

  • View in gallery

    (left) Vertical profiles of the intrawave turbulent kinetic energy (TKE; k in m2 s−2) at the mid-obstacle location. (top to bottom) The free surface location (m), TKE for the reference case, the difference in TKE for the solid obstacle, and the difference in TKE for the porous obstacle. The white contour lines reperesent values of 10−6, 10−5, and 10−4 m2 s−2. (right) A zoom of the area near the cage.

  • View in gallery

    Wave-averaged friction velocity (m s−1) at five streamwise locations. Circles are for the reference results without obstacle, the plus signs for the porous obstacle, and the crosses for the solid obstacle.

  • View in gallery

    Profiles of the vertical and horizontal velocity components under and above the center of the cage for an upstream current of 1 m s−1 with constant vertical profile.

  • View in gallery

    Horizontal velocity profile for 3 positions: 4 m upstream of center of the frame, 1 m upstream (edge of the frame), and under and above the center of the cage for a logarithmic profile with a depth-averaged velocity of 0.7 m s−1.

  • View in gallery

    Vertical velocity vs horizontal velocity at the LEACOAST2 deployments. The line represents the linear fit to the data.

  • View in gallery

    Ripple profiler example showing (top) the bed scour under STABLE-III at the Hilbre Channel in the Dee estuary and (bottom) the scour at the LEACOAST2 data experiment 1 F1.

  • View in gallery

    Relationship between scour depth and current speed for the Dee and LEACOAST2 data.

  • View in gallery

    Vertical (positive = upward) and horizontal velocity component from a top-mounted ADV in Liverpool Bay.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 90 56 5
PDF Downloads 33 23 1

Effects of Instrumented Bottom Tripods on Process Measurements

View More View Less
  • 1 National Oceanography Centre, Liverpool, United Kingdom
  • | 2 CFD Solutions Ltd, Exeter, United Kingdom
© Get Permissions
Full access

Abstract

The measurement and assessment of ocean bottom processes are important sources of information for understanding bedform evolution and sediment entrainment and for improving numerical models. Instrumented tripods have been used to investigate bottom boundary layer and sediment dynamics processes for several decades. In this paper, the effects of instrumented tripods on hydrodynamics and on the sea bed are investigated via numerical modeling and field data collected under moderate to strong tidal currents and mild surface waves. Under high currents, streamlines are modified and structure-induced vertical velocities are produced. To minimize this effect, a rotation of the three-dimensional current measurement under the frame is recommended. Acceleration of the flow under the frame is also significant (on the order of 10%–20%), which leads to an increase in bottom stress and can produce a large scour pit in energetic currents. Wave–structure interactions mainly increase turbulence near the frame. No significant wave effect has been observed near the bed, and scouring thus mostly relates to tidal currents.

Corresponding author address: Rodolfo Bolaños, National Oceanography Centre, Joseph Proudman Building, 6 Brownlow Street, Liverpool L3 5DA, United Kingdom. E-mail: rbol@pol.ac.uk

Abstract

The measurement and assessment of ocean bottom processes are important sources of information for understanding bedform evolution and sediment entrainment and for improving numerical models. Instrumented tripods have been used to investigate bottom boundary layer and sediment dynamics processes for several decades. In this paper, the effects of instrumented tripods on hydrodynamics and on the sea bed are investigated via numerical modeling and field data collected under moderate to strong tidal currents and mild surface waves. Under high currents, streamlines are modified and structure-induced vertical velocities are produced. To minimize this effect, a rotation of the three-dimensional current measurement under the frame is recommended. Acceleration of the flow under the frame is also significant (on the order of 10%–20%), which leads to an increase in bottom stress and can produce a large scour pit in energetic currents. Wave–structure interactions mainly increase turbulence near the frame. No significant wave effect has been observed near the bed, and scouring thus mostly relates to tidal currents.

Corresponding author address: Rodolfo Bolaños, National Oceanography Centre, Joseph Proudman Building, 6 Brownlow Street, Liverpool L3 5DA, United Kingdom. E-mail: rbol@pol.ac.uk

1. Introduction

Instrumented bottom frames have been used since the 1960s to investigate bottom boundary layer processes and sediment dynamics. Data from these frames have led to a better understanding of near-bottom wave and current flows in the coastal ocean, and have been very important for the development of numerical models. These data have been used for calculations of bottom stress, bottom roughness, sediment flux, sediment resuspension, and near-bed and water column processes (e.g., Dolphin and Vincent 2009; Lee et al. 2002; Traykovski 2007; Sherwood et al. 2006; Williams et al. 2005). The frames on which the instruments are mounted have generally been three-legged (tripods), a structure that offers good stability without significant flow interference near the sea floor (Cacchione et al. 2006). Tripods are self-contained, fully submerged structures that can rest stably on the seafloor, and to which are attached various instruments, batteries, and dataloggers. Modern systems may contain numerous sensors; in particular, they include sensors that measure current speed and direction at multiple levels near the bed. For a more exhaustive review of tripod history and evolution the reader is referred to Cacchione et al. (2006).

Instrumentation on tripods has continued to evolve rapidly, providing a large variety of data and improved accuracy. Accurate measurements of vertical velocity are important for estimating turbulence and calculating bottom stresses. However, because of the relative magnitudes of vertical and horizontal velocities in the bottom boundary layer, measurements of vertical velocity are very sensitive to a small tilt of the instrument, to the effect of bed slope, or to a flow modification induced by the frame. Furthermore, as the temporal and spatial resolution of measurements is improved, the hydrodynamic features that can be observed are more detailed and thus the effect of the frame itself might need to be taken into account when analyzing and interpreting data.

Structures immersed in a fluid produce a modification of the ambient flow that depends on both the flow and the structure properties. For example, an acceleration of the flow and a change in pressure can be induced by the Bernoulli effect, eddies can be generated, and turbulence can be produced. These phenomena are widely used and studied in engineering. In particular, turbulence in laboratory studies is often generated by a grid structure.

In spite of the relatively large use of tripods, little work has been done to try to identify and understand the effects of the frame on the flow, on sediment transport, or on the bed. Williams et al. (2003) investigated the interactions between a tripod and waves in a full-scale laboratory experiment and found that the frame had surprisingly small effects on the surrounding hydrodynamics, bed morphology, and suspended sediment. However, they solely considered wave-only situations in a controlled laboratory setting. The main objective of the present work is thus to extend the investigation of the effect of a tripod on the surrounding flow and sediment bed to include wave and current situations and field measurements. Numerical modeling and real data from tripods are used to show the effects and implications when measuring boundary layer processes.

The different methodologies, specifically the models and data used, are described in section 2. Then, in section 3, the numerical and observational results are presented. In section 4, the implications for field measurements and data treatment are discussed. Finally, section 5 summarizes the study’s conclusions.

2. Methodology

a. Modeling wave–structure interaction

To numerically simulate the interactions between waves and a tripod, the Cornell Breaking Waves and Structures (COBRAS) model (Lin and Liu 1998a) was chosen. This model has been tested in numerous situations in the surf zone (e.g., Torres-Freyermuth et al. 2007), for breaking waves (e.g., Lin and Liu 1998a,b), and for wave–structure interactions (Hsu et al. 2002). More importantly, it has also been used to numerically study flows over submerged obstacles for solitary waves (Chang et al. 2001) and for cnoidal waves (Chang et al. 2005). In both cases, the numerical results compared well with experimental measurements, confirming the good performance of the model.

1) Model description

The numerical model solves the two-dimensional Reynolds-averaged Navier–Stokes (RANS) equations and uses a kɛ turbulence closure (Lin and Liu 1998b). The Reynolds stress is modeled following the nonlinear stress–strain relationship of Shih et al. (1996). The model therefore provides temporal and spatial variations of the velocity and turbulence fields. The free surface location is tracked following the volume of fluid (VOF) method (Hirt and Nichols 1981). Flow through a porous media can also be resolved and is calculated by solving the volume-averaged RANS equations (Hsu et al. 2002).

The two-dimensional RANS equations are solved numerically following the two-step projection method (Chorin 1968). Time derivatives are discretized using a forward time-differencing method. Pressure and stress gradient are discretized using a central difference method, while the convective terms employ a combination of central difference and upwind method. Similar algorithms are used to solve the balance equations for the turbulent kinetic energy k and the turbulence dissipation rate ɛ. On the free surface, boundary conditions enforce zero stress and zero normal gradient for k and ɛ. Near solid boundaries and at the bottom boundary, the logarithmic law of the wall is applied to obtain the wall shear stress and to provide estimates of the velocity gradient, turbulent kinetic energy, and turbulence dissipation rate.

2) Model setup

Although COBRAS is only two dimensional and cannot be used at a resolution sufficient to describe the precise structure of a tripod, it can be used for more idealized scenarios in which the instrument cage is represented by a solid or porous rectangular obstacle located some distance from the sea bed. In the present numerical simulations, monochromatic waves of height 1.5 m and period 5 s are generated in 10-m mean water depth using an internal wavemaker (Lin and Liu 1999). Sponge layers are used at both lateral boundaries to absorb possible reflected waves (e.g., Lara et al. 2006). The numerical setup is described in Fig. 1. Both sponge layers are taken to be 60 m thick. The source region is designed based on the recommendations of Lin and Liu (1999). The obstacle, which is 2 m long and 0.5 m high, is located 2 m above the bed and sufficiently far downstream of the source region to minimize errors introduced by the internal wave maker (Lin and Liu 1999). A constant 5-cm grid size is used in the vertical direction. In the horizontal direction, a variable grid size is employed, with a minimum grid size of 5 cm in the region 2 m upstream and downstream of the obstacle.

Fig. 1.
Fig. 1.

Sketch of the two-dimensional wave model setup showing (top) the entire computational domain and (bottom) a close-up around the obstacle corresponding to the region delimited by the dashed lines in the top panel. In the bottom panel, the dashed lines represent the location at which vertical profiles are determined.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

Three numerical simulations have been performed to investigate the impact of the idealized rectangular cage on the flow and the turbulence. A reference case is conducted with no obstacle and two other cases consider an idealized cage as either a solid or porous rectangular obstacle. In the porous case, the porosity of the obstacle is taken to be 0.8.

b. Modeling current–structure interaction

Although COBRAS is a validated model that performs well for many wave applications, it is limited for simulating currents. A commercial computational fluid dynamics model is used instead to investigate the modifications induced by the structure under tidal currents. The model employed is Flow-3D, which has been used in numerous maritime, riverine, and aerospace studies including tidal flows (Vasquez and Walsh 2009) and modeling energy converters (Binder et al. 2009). This model was also used as a tool during the original design of the Sediment Transport and Boundary Layer Equipment (STABLE) III tripod (Humphery et al. 2007).

1) Description of the model

The model used is part of a package developed by Flow Science in Santa Fe, New Mexico (www.flow3d.com). Despite using a rectangular grid, it can model complex geometries by applying the fractional area–volume method, which allows a computational cell to be partially blocked by an obstacle. Partial cell treatment is used to represent solid boundaries (bottom and obstacles). This method is similar to the VOF method and models the solid object as a special case of the two-phase flow with infinite density. Computational cells can then be partially blocked, which reduces the spurious reflections induced by a discrete rectangular representation of the boundary. The model solves a series of three-dimensional flow equations using a reduced set of the Navier–Stokes equations and includes schemes for the simulation of turbulence. The turbulence closure follows the renormalization group (RNG) approach of Yakhot and Orszag (1986), which results in a modified linear viscosity kɛ model. The RNG ɛ equation is derived from the Navier–Stokes equation, to which an extra ad hoc term is added. This modeling system was used to perform drag calculations to assess tripod stability in strong steady and wave-modulated flows (Humphery et al. 2007).

2) Model setup

The model was set up in a grid with approximately of 1.6 million nodes representing the solid geometry and surrounding seawater. A variable grid size was used with a minimum size of 3 cm close to the tripod and a maximum of 7 cm close to the computational domain boundaries. Two flow boundary conditions representative of a strong tidal current (0.7 and 1 m s−1) were applied to the model domain, each with either a logarithmic or a vertically constant profile. The model setup is the same as in Humphery et al. (2007) and Fig. 2 shows a conceptual model of the rig used to simulate the currents around the structure. The densely populated instrument platform, which was initially viewed as a squat cylinder with a drag coefficient (CD) of about 0.7, was shown to have a CD of about 1.6.

Fig. 2.
Fig. 2.

Conceptual representation of STABLE III from CADdata used in Flow3D.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

c. Measurements around the structures

The STABLE platform is used to measure and study near-bed hydrodynamics, near-bed turbulence, and the associated sediment dynamics. STABLE, which was originally developed in 1981, now includes several tripods designed and built at the Proudman Oceanographic Laboratory (POL; now the National Oceanography Centre). The biggest and most complex tripod is STABLE III, which stands 2.5 m high, weighs about 2500 kg, and (including feet) covers a circle of diameter ~3.5 m (Williams et al. 2003). Another version is MiniSTABLE, which stands about 1.8 m and has a triangular shape with each side of about 1.5 m.

Data from two sites are used to explore the effect of the frames on measurements. The first site is Sea Palling in the southeast of the United Kingdom (see Fig. 3), where several tripods were deployed during the Liverpool–East Anglia Coastal Study Phase 2 (LEACOAST2) project (Wolf et al. 2008). The second site is the Dee estuary (Fig. 3), located in the eastern Irish Sea. These two sites present different hydrodynamic and sediment properties. Sea Palling is a sandy beach with detached parallel offshore breakwaters. Tides are moderate with a tidal range of about 3 m and acoustic Doppler velocimeters (ADVs) have measured tidal currents of about 0.6 m s−1 in spring tides. Wave events are frequent with significant wave heights of the order of 2–3 m. The Dee estuary is a macrotidal estuary 20 km long and 8 km wide at the mouth, in which the flow is mainly confined to two channels at the mouth. The tidal amplitude can reach over 10 m and the tidal currents in the channels can exceed 1 m s−1, both at spring tides. The bed is primarily composed of mixed sediments. The estuary is only subject to moderate wave events that are strongly modulated by tidal levels and currents.

Fig. 3.
Fig. 3.

Location of rigs in the (left) Dee estuary (west) and (right) Sea Palling (east). The Dee estuary data are from 2008 and the Sea Palling data are from deployments during 2006–07. The color bar represents water depth (m); negative values in the Dee estuary represent wet areas above mean sea level.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

Data from the Dee estuary correspond to deployments during 2008 (Bolaños and Souza 2010). The LEACOAST2 experiments took place in two periods from March to May 2006 and from October 2006 to January 2007.

Data from these deployments are used to investigate the effect of the tripod frame on hydrodynamics and the sediment bed. The hydrodynamic effects are examined using data from SonTek ADVs that are fixed under the frames, while 2D and 3D ripple profilers are used to observe the sea bed response.

1) Acoustic Doppler velocimeter

The ADVs provide collocated measurements of the three components of flow at high temporal resolution at a single point by using the pulse-coherent technique. Detailed quality control including despiking of data was carried out for both sites (e.g., Bolaños and Souza 2010). The removal of spikes is necessary as they are known to erroneously modify the spectral analysis and estimation of bottom stress. The approach taken follows that of Goring and Nikora (2002) as modified by Wahl (2003). The method is based on a three-dimensional Poincaré map, in which each component of velocity and its first and second derivative are plotted against each other. The points located outside of the ellipsoid in the Poincaré map are excluded or replaced by a polynomial interpolation. The data used in this analysis have been despiked to remove the peaks of the time series while conserving the main original properties (e.g., Bolaños and Souza 2010).

2) Ripple profiler

Ripple profilers used on the deployments consisted of either a two-dimensional line scan ripple profiler or a three-dimensional scanner, which is a line scan ripple profiler that rotates to provide an area coverage. The transducer can determine the bed position with a spatial resolution of about 0.005 m (Williams et al. 2004, 2005).

Raw bed-level data from a rotating profiler located at STABLE-III in the Dee estuary have been treated for conversion into three-dimensional bed-level images of 1-cm resolution. There are some shadows and noise in the data due to the frame and instruments (particularly the ADVs). A method based on a local derivative threshold and a comparison with a polynomial curve fitted to the profile has been implemented to eliminate these sources of error and interpolate the bed elevation (Bolaños et al. 2009). This was not required for the fixed ripple profilers (2D) as they were carefully located to have a clear sweep without shadows. A summary of ADV and ripple profiler data together with a description of the hydrodynamic conditions at each deployment site is presented in Table 1. The magnitude of the tidal currents is defined as the 80th percentile of the velocity cumulative distribution (i.e., 80% of the data have lower velocity).

Table 1.

Summary of the ADVs, ripple profilers, and hydrodynamic conditions at the Dee estuary and Sea Palling (LEACOAST2) sites. The F number at the Sea Palling site indicates different rig deployments as shown in Fig. 3.

Table 1.

3. Results

a. Wave–structure interaction modeling

The presence of an obstacle is expected to induce a flow modification and to produce turbulence. When a wave propagates over a submerged obstacle, vortices are generated and shed within a wave cycle (Chang et al. 2005). The modification introduced by the idealized cage on the intrawave velocities is shown in Fig. 4. Very little influence of the obstacle on the wave height has been observed in the numerical simulations. An intensification of the horizontal flow of up to 20% is present below the idealized cage in both solid and porous cases. Above the cage, a reduction of the horizontal velocity occurs for the porous case, which is due to higher roughness and interaction with the vertical velocity component. Significant differences are also observed for the vertical velocity. In particular, the pattern for the solid case corresponds to almost no vertical flow under the cage and to vertical velocities vanishing at the solid obstacle top boundary. Comparatively, the difference is much smaller for the porous case; vertical flow is possible through the obstacle and still present below it. Similar behaviors are observed at the other two streamwise cage locations, while little flow disturbance is observed upstream and downstream of the idealized cage.

Fig. 4.
Fig. 4.

Vertical profiles of the intrawave velocity components: (left) U and (right) W (m s−1) at the mid-obstacle location. (top to bottom) The free surface location (m), velocities from the reference case, differences in velocities for the solid obstacle, and difference in velocities for the porous obstacle (note differences in color scales).

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

The effect of the idealized cage on the turbulent kinetic energy is shown in Fig. 5. Again, similar behavior is observed across the cage, while little change occurs at the upstream and downstream locations. There is strong flow turbulence generated by the obstacle in both solid and porous cases. However, the detailed behavior of this additional turbulence differs significantly depending on the obstacle. Turbulence around the solid obstacle is generated by two principal mechanisms: friction on the obstacle boundary and vortices generated by the wave flow over the obstacle (e.g., Chang et al. 2005). The first mechanism is limited to regions very close to the obstacle boundaries and is somewhat in phase with the wave height. The second mechanism occurs around flow reversal and is responsible for the elevated turbulence values (Fig. 5, third row) that are not directly adjacent to the boundaries but are of similar magnitude and impact the flow farther away from the obstacle boundaries. Turbulence around the porous obstacle is more intense but generally limited to regions closer to the obstacle. In particular, the strong turbulence that is associated with vortices for the solid case tends to vanish in the porous case. Also, the turbulence maximum coincides with flow reversals and originates within the porous obstacle.

Fig. 5.
Fig. 5.

(left) Vertical profiles of the intrawave turbulent kinetic energy (TKE; k in m2 s−2) at the mid-obstacle location. (top to bottom) The free surface location (m), TKE for the reference case, the difference in TKE for the solid obstacle, and the difference in TKE for the porous obstacle. The white contour lines reperesent values of 10−6, 10−5, and 10−4 m2 s−2. (right) A zoom of the area near the cage.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

The modification of the bed shear stress by the structure is an important effect that can result in scour. Figure 6 presents the wave-averaged friction velocity at the five streamwise locations specified in Fig. 1. While almost no difference is observed at the two locations upstream and downstream of the cage, a small increase in the bed shear stress up to about 5% is observed under the obstacle. The modeled case deals with monochromatic waves and the reported increase in bottom stress could be different in nonmonochromatic situations.

Fig. 6.
Fig. 6.

Wave-averaged friction velocity (m s−1) at five streamwise locations. Circles are for the reference results without obstacle, the plus signs for the porous obstacle, and the crosses for the solid obstacle.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

b. Current–structure interaction modeling

Modeling of horizontal and vertical velocity profiles shows a significant variation from the upstream profiles due to the presence of the rig (Fig. 7). The streamlines are distorted and vertical velocities are generated around the structure. Under the frame, negative (downward) vertical velocities can reach about 10% of the upstream horizontal velocity in the case simulated. This downward motion is present throughout the vertical section and decreases in magnitude toward the bottom. The horizontal velocity is also modified. For the first 40 cm under the cage, the velocity is reduced, the cage itself acts as a boundary, and a boundary layer is formed. Below that layer, the shape of the velocity profile is relatively similar to that of the upstream profile (for both monotonic and log profiles), but the magnitude is raised by about 10%–20% compared with the upstream value (Figs. 7 and 8). Turbulence is also expected to increase because of the presence of multiple boundary layers and the increase in the current vertical shear (Fig. 8). Given that the bottom stress is related to the velocity squared, this effect can produce an increase in the bottom stress of more than 30%. This can in turn be responsible for large erosion under environmental conditions near the threshold of motion of bed sediments.

Fig. 7.
Fig. 7.

Profiles of the vertical and horizontal velocity components under and above the center of the cage for an upstream current of 1 m s−1 with constant vertical profile.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

Fig. 8.
Fig. 8.

Horizontal velocity profile for 3 positions: 4 m upstream of center of the frame, 1 m upstream (edge of the frame), and under and above the center of the cage for a logarithmic profile with a depth-averaged velocity of 0.7 m s−1.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

c. Measurement evidence of the effect of the frame on the flow

Field data collected during tripods deployments are presented in this section. Results and analysis help identify frame effects on horizontal and vertical velocities, and scouring under the frame.

1) ADV velocities

An important and very sensitive variable in marine turbulence is the vertical component of the velocity. Its measurement can be used to estimate bottom stresses, turbulence production, and mixing. The vertical velocity component is more sensitive than the horizontal component to errors in measurements that result from heterogeneity of the field and tilting of the instrument. ADVs provide an accurate measurement of the three components of the velocity, and some contamination of the vertical component is expected since some tilting of the instrument is inevitable in the field. However, in a bidirectional flow such as a tidal system, such contamination is expected to change sign when changing flow direction. This is not the case for the data presented here, which present a constant negative (downward) velocity for most cases. Figure 9 shows a scatterplot of the burst-averaged vertical component versus the horizontal component from the LEACOAST2 data. A linear trend can be observed: larger vertical velocities correspond to larger horizontal components. The data from the Dee also present this pattern. This phenomenon can be attributed to the frame effect, which is persistent for all data under the cage as shown by the numerical modeling. Even though the figure presents some scatter, which is inevitable due to the different rig configurations, ADV heights, and environmental conditions, the linear regression shows a vertical velocity of about 10% of the horizontal velocity, in agreement with numerical results.

Fig. 9.
Fig. 9.

Vertical velocity vs horizontal velocity at the LEACOAST2 deployments. The line represents the linear fit to the data.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

Another indication of a frame-induced effect is the modification of the expected logarithmic profile near the bed. Even though this is not shown in the current-only numerical simulations, further evaluation with measured data is necessary to provide conclusive evidence on the matter. However, this requires measurements at several vertical locations, thus limiting the available datasets because, as shown in Table 1, the maximum number of measurements is three. The data do not show a logarithmic profile under the frame and the estimation of the bottom stresses with this method provides extremely large (unrealistic) stresses (not shown here). This might be a combination of the direct effect of the structure on the flow and of an indirect effect due to a scour pit under the frame (see next section). The estimation of bottom stresses by fitting a logarithmic profile to two or three ADV measurements can therefore not be reliable for our data.

2) Scour under the rigs

An evidence of a structure effect on the bed is the presence of a scour pit. The ripple profiler on STABLE-III in the Hilbre channel of the Dee estuary, located 1.34 m above the bed and providing a 3D image every hour, shows clear evidence of this: a very large scour pit about 0.8 m deep and with a diameter of 6 m was observed under the frame (Fig. 10, top). At this location, bedforms are dominated by the scour pit, which is stable with time, and ripples are not evident. Bolaños et al. (2009) studied this scour and observed that the mean bed location rises with time. This indicates that even though a scour pit is formed, the bed rises at the extremes of the ripple profiler image. A steepness parameter for the pit also increased with time, which was a direct consequence of the scour pit getting deeper. During spring tides, the oscillation of the scour steepness was clear. This location presents the largest scour, which is attributed to large near-bed velocities (close to 0.6 m s−1) compared to velocities of less than 0.43 m s−1 at the other locations.

Fig. 10.
Fig. 10.

Ripple profiler example showing (top) the bed scour under STABLE-III at the Hilbre Channel in the Dee estuary and (bottom) the scour at the LEACOAST2 data experiment 1 F1.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

The 2D ripple profiler deployed in the Welsh channel presented a relative lower scour (~15 cm) than the 3D ripple profiler in the Hilbre channel. Similar to the Welsh channel rig, the locations in Sea Palling presented smaller scour with features of the order of 10–30 cm (see Fig. 10, bottom), which is attributed to lower tidal velocities (<0.43 m s−1).

4. Discussion

Based on flume experiments, Williams et al. (2003) did not find any important wave–structure interaction in terms of either turbulence or the velocity field. However, the numerical results in the present work have shown that turbulence near the frame can be significantly increased under waves. This disagreement might be due to the previous measurements being out of the near-cage region that would be affected by structure–flow interaction. Williams et al. (2003) did not assess the effect of the frame under currents, which can be significant in tidal environments as shown in this paper.

The effect of the frame is clear, especially under high current conditions, and has been confirmed both numerically and experimentally. It produces an increase of the horizontal velocity under the frame, and it artificially creates negative (downward) vertical velocities. The cage acts as a boundary and induces boundary layers. The structure and bed boundary layers can overlap below the cage, which can cause a significant rise in turbulence as current shear is increased. The cage boundary layer is expected to be thinner as it does not have the length to completely develop. The generation of the downward vertical velocities under the frame suggests that an intraburst correction is needed. This is particularly important for estimating turbulence parameters and bottom stresses using approaches such as the turbulent kinetic energy method or the Reynolds stress method. A rotation of the intraburst data is suggested to minimize the vertical velocity and obtain more reliable estimates. It is not recommended to take measurements in the first 40 cm near the cage in order to avoid the rise of wave-induced turbulence, the current–structure boundary layer, and the large modification of the vertical velocities.

Even though the velocity measurements can be corrected in order to have a better estimation of the environmental conditions, the rise of the velocity and turbulence can have a very important local effect for studies of the interaction of flow with bedforms, ripple migration, and near-bed suspended sediment concentration. The presence of a scour pit under some of the frames is clear evidence of the importance of the frame effect. The scour depth should depend on a number of hydrodynamic, sediment, and frame parameters. The dependence on tidal velocity is evident both in theory and from the field data (Fig. 11). However, many other parameters will influence the scour depth and the data from the present field studies are not sufficient to completely isolate one parameter. The data summarized in Fig. 11 should therefore not be interpreted as precise quantitative evidence but rather as qualitative example of the dependence of the scour depth on tidal velocity. In particular, even though the pattern exhibited in Fig. 11 seems to suggest an exponential behavior, the amount of data is not sufficient to yield unambiguous results and uncertainty remains. Qualitatively, the observed scour may be separated in 3 groups: one with no scour that occurred for tidal velocities lower than about 10 cm s−1, one characterized by scour of 10–30 cm with tidal velocities of about 30–50 cm s−1 and a third one represented by the Hilbre channel, with a scour of about 80 cm for tidal currents of 58 cm s−1. The first group could suggest the existence of a threshold for the appearance of scour, the value of which should depend on a combination of current velocity and sediment erodability properties. From the wave conditions in the data presented and the wave modeling, waves do not seem to have a relevant effect on scour formation. A direct consequence of the formation of a significant scour pit below the instrumented frame is that data under a large tidal current (i.e., from the third group here) have to be treated with caution. In this situation, bed processes are dominated by the scour, the near-bed velocity does not follow a logarithmic profile, and sediment entrainment and flow near the bed may be complex processes highly influenced by the frame.

Fig. 11.
Fig. 11.

Relationship between scour depth and current speed for the Dee and LEACOAST2 data.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

A frame effect that has not been considered here is the possible eddy shedding due to the legs of the frame. However, this should be considered if the orientation of the rig cannot be controlled during deployment and if the ADV sampling volume may be located in the legs’ wakes. The body of the ADV (a cylinder about 10 cm in diameter and 40 cm long) itself might also create a significant disturbance of the flow and generate turbulence in the nearby area. If the ADV body is very close to the bottom, it might become another source of scour. It is therefore recommended to use the ADV in a vertical (down-looking) deployment for measurements very close to the bed.

While our results focused on velocity measurements conducted below the tripod frame, generation of vertical velocity by the frame can also contaminate results of top-mounted instruments. To illustrate this, a bed frame with a top-mounted ADV sampling at about 60 cm above the frame in the Liverpool bay is used (Betteridge and Souza 2008). Figure 12 shows the scatterplot of vertical and horizontal component. The vertical component is constantly positive (upward) and in linear correlation with the horizontal component because of a deflection of the horizontal component by the bed frame. The data present a shallower slope compared to the data under the rigs; the vertical velocity is about 5% of the horizontal component, which might be due to the configuration of the frame and the elevation of the ADV sampling volume.

Fig. 12.
Fig. 12.

Vertical (positive = upward) and horizontal velocity component from a top-mounted ADV in Liverpool Bay.

Citation: Journal of Atmospheric and Oceanic Technology 28, 6; 10.1175/2010JTECHO816.1

5. Conclusions

Depending on their configuration, tripods can produce significant alteration of the flow and turbulence and lead to scour pit generation. Wave–structure interactions increase turbulence near the cage, but no significant effect near the bed was evident. Current–structure interactions produce artificial vertical velocities, an increase of horizontal velocity under the frame, and enhanced bottom stress that, under large flows, produce considerable scour under the rig. The wave model did not show a significant increase in shear stress near the bed and therefore is not believed to lead to significant scour compared with the large current-induced effect. Environmental tidal currents of about 30 cm s−1 (defined as the 80th percentile of the distribution) produced a small scour pit, while a location with 58 cm s−1 showed a large scour under the frame. The generation of vertical velocities can be corrected by an intraburst rotation of the data, a procedure that is necessary for the extraction of turbulence time series and estimation of bottom stresses. The rigs’ data are expected to be a very valuable source of information in wave and moderate tidal current conditions but, in large tidal currents, the near-bed processes measured might be greatly affected by the scour and frame effects.

Acknowledgments

The authors thank the Research Vessel Prince Madog crew for their support during the deployment and recovery of the rigs at the Dee. Support from the engineering group at NOC is also acknowledged. Some of the data were provided by the LEACOAST2 project. We acknowledge the support of Dr P. Bell, at NOC, during the ripple profiler data analysis. The authors are grateful for the support of NERC through the FORMOST project (NE/E015026/1). Comments from two anonymous referees led to significant improvement of the manuscript.

REFERENCES

  • Betteridge, K. F. E., and Souza A. J. , 2008: Estimates of the turbulent kinetic energy (TKE) balance in shelf seas. Proc. 2008 PECS Conf., Liverpool, United Kingdom, Proudman Oceanographic Laboratory, 301–304.

    • Search Google Scholar
    • Export Citation
  • Binder, M. A., Mingham C. G. , Causon D. M. , Rahmati M. T. , Aggidis G. A. , and Chaplin R. V. , 2009: Numerical modelling of a surging point absorber wave energy converter. Proc. Eighth European Wave and Tidal Energy Conf., Uppsala, Sweden, EWTEC. [Available online at http://www.flow3d.com/pdfs/tp/mar_tp/FloSci-Bib32-09.pdf.]

    • Search Google Scholar
    • Export Citation
  • Bolaños, R., and Souza A. J. , 2010: Measuring hydrodynamics and sediment transport processes in the Dee Estuary. Earth. Syst. Sci. Data, 2, 157165.

    • Search Google Scholar
    • Export Citation
  • Bolaños, R., Souza A. J. , Bell P. S. , and Malarkey J. , 2009: Measuring tidal hydrodynamics and bedform evolution in the Dee estuary. River, Coastal and Estuarine Morphodynamics, C. A. Vionnet et al., Eds., CRC Press, 709–714.

    • Search Google Scholar
    • Export Citation
  • Cacchione, D. A., Sternberg R. W. , and Ogston A. S. , 2006: Bottom instrumented tripods: History, applications and impacts. Cont. Shelf Res., 26, 23192334.

    • Search Google Scholar
    • Export Citation
  • Chang, K.-A., Hsu T.-J. , and Liu P. L.-F. , 2001: Vortex generation and evolution in water waves propagating over a submerged rectangular obstacle Part I. Solitary waves. Coastal Eng., 44, 1336.

    • Search Google Scholar
    • Export Citation
  • Chang, K.-A., Hsu T.-J. , and Liu P. L.-F. , 2005: Vortex generation and evolution in water waves propagating over a submerged rectangular obstacle Part II. Cnoidal waves. Coastal Eng., 52, 257283.

    • Search Google Scholar
    • Export Citation
  • Chorin, A. J., 1968: Numerical solution of the Navier–Stokes equations. Math. Comput., 22, 745762.

  • Dolphin, T., and Vincent C. , 2009: The influence of bed forms on reference concentration and suspension under waves and currents. Cont. Shelf Res., 29, 424432.

    • Search Google Scholar
    • Export Citation
  • Goring, D. G., and Nikora V. I. , 2002: Despiking acoustic Doppler velocimeter data. J. Hydraul. Eng., 128, 117126.

  • Hirt, C. W., and Nichols B. D. , 1981: Volume of fluid (VOF) method for the dynamics of free boundaries. J. Comput. Phys., 39, 201225.

    • Search Google Scholar
    • Export Citation
  • Hsu, T.-J., Sakakiyama T. , and Liu P. L.-F. , 2002: A numerical model for wave motions and turbulence flows in front of a composite breakwater. Coastal Eng., 46, 2550.

    • Search Google Scholar
    • Export Citation
  • Humphery, J. D., Jones D. S. , and Doyle K. , 2007: The CFD-aided design of STABLE III, a new platform to study near-bed turbulent currents and associated sediment dynamics. Proc. OCEANS 2007, Aberdeen, Scotland, IEEE/OES, 1–6.

    • Search Google Scholar
    • Export Citation
  • Lara, J. L., Garcia N. , and Losada I. J. , 2006: RANS modelling applied to random wave interaction with submerged permeable structures. Coastal Eng., 53, 395417.

    • Search Google Scholar
    • Export Citation
  • Lee, G., Friedrichs C. T. , and Vincent C. E. , 2002: Examination of diffusion versus advection dominated sediment suspension on the inner shelf under storm and swell conditions, Duck, North Carolina. J. Geophys. Res., 107, 3084, doi:10.1029/2001JC000918.

    • Search Google Scholar
    • Export Citation
  • Lin, P., and Liu P. L.-F. , 1998a: A numerical study of breaking waves in the surf zone. J. Fluid Mech., 359, 239264.

  • Lin, P., and Liu P. L.-F. , 1998b: Turbulence transport, vorticity dynamics and solute mixing under plunging breaking waves in surf zone. J. Geophys. Res., 103, 15 67715 694.

    • Search Google Scholar
    • Export Citation
  • Lin, P., and Liu P. L.-F. , 1999: Internal wave-maker for Navier–Stokes equations models. J. Waterw. Port Coastal Ocean Eng., 125, 207215.

    • Search Google Scholar
    • Export Citation
  • Sherwood, C. R., Lacy J. R. , and Voulgaris G. , 2006: Shear velocity estimates on the inner shelf off Grays Harbour, Washington, USA. Cont. Shelf Res., 26, 19952018.

    • Search Google Scholar
    • Export Citation
  • Shih, T.-H., Zhu J. , and Lumley J. L. , 1996: Calculation of wall-bounded complex flows and free shear flows. Int. J. Numer. Methods Fluids, 23, 11331144.

    • Search Google Scholar
    • Export Citation
  • Torres-Freyermuth, A., Losada I. J. , and Lara J. L. , 2007: Modeling of surf zone processes on a natural beach using Reynolds-averaged Navier–Stokes equations. J. Geophys. Res., 112, C09014, doi:10.1029/2006JC004050.

    • Search Google Scholar
    • Export Citation
  • Traykovski, P., 2007: Observations of wave orbital scale ripples and a nonequilibrium time-dependent model. J. Geophys. Res., 112, C06026, doi:10.1029/2006JC003811.

    • Search Google Scholar
    • Export Citation
  • Vasquez, J. A., and Walsh B. W. , 2009: CFD simulation of local scour in complex piers under tidal flow. Proc. 33rd IAHR Congress, Vancouver, BC, Canada, International Association of Hydraulic Engineering & Research, 978-94-90365-01-1.

    • Search Google Scholar
    • Export Citation
  • Wahl, T. L., 2003: Discussion of “Despiking acoustic Doppler velocimeter data.” J. Hydraul. Eng., 129, 484486.

  • Williams, J. J., Bell P. S. , Coates L. E. , Metje N. , and Selwyn R. , 2003: Interactions between a benthic tripod and waves on a sandy bed. Cont. Shelf Res., 23, 355375.

    • Search Google Scholar
    • Export Citation
  • Williams, J. J., Bell P. S. , Thorne P. D. , Metje N. , and Coates L. E. , 2004: Measurement and prediction of wave-generated suborbital ripples. J. Geophys. Res., 109, C02004, doi:10.1029/2003JC001882.

    • Search Google Scholar
    • Export Citation
  • Williams, J. J., Bell P. S. , and Thorne P. D. , 2005: Unifying large and small wave-generated ripples. J. Geophys. Res., 110, C02008, doi:10.1029/2004JC002513.

    • Search Google Scholar
    • Export Citation
  • Wolf, J., Souza A. J. , Bell P. S. , Thorne P. D. , Cook R. D. , and Pan S. , 2008: Wave, currents and sediment transport observed during the LEACOAST2 experiment. Proc. PECS 2008, Liverpool, United Kingdom, Proudman Oceanographic Laboratory, 373–376.

    • Search Google Scholar
    • Export Citation
  • Yakhot, V., and Orszag S. A. , 1986: Renormalization group analysis of turbulence. Basic theory. J. Sci. Comput., 1, 351.

Save