Suspended cohesive sediments commonly occur in flocculated form. The flocculation process significantly increases particle size, settling velocity, and settling flux in comparison to unflocculated particles. To better understand suspended particles, a new instrument for in situ imaging using digital inline holography is presented and evaluated. With a resolution of 7.4 µm per pixel and a field of view of 7.4 × 7.4 mm, the instrument generates sharply focused images of particles from approximately 20 µm to 7 mm in diameter at up to 25 frames per second. A significant advantage of holography over current imaging systems is that in-focus images are obtained over a substantial depth of field. Contained in small-diameter cylindrical housings, the instrument presents minimal flow disruption and is easily deployable. Digitally recorded holograms are reconstructed and analyzed for number and size of particles in a fully automated manner. To assess the systems' potential, particle size distributions for two grades of quartz sand are compared with those from a laser diffraction particle sizer and are found to exhibit good similarity. To simultaneously estimate particle size and settling velocity, a modification to the system is demonstrated in which settling particles are automatically tracked and sized. The resulting relationships between settling velocity and effective density as functions of particle size are characteristic of suspended cohesive sediments and empirical power-law descriptors of these relationships show excellent agreement with previously published studies.

No abstract

Seven sets of 2D particle image velocimetry data obtained in the bottom boundary layer of the coastal ocean along the South Carolina and Georgia coast [at the South Atlantic Bight Synoptic Offshore Observational Network (SABSOON) site] are examined, covering the accelerating and decelerating phases of a single tidal cycle at several heights above the seabed. Additional datasets from a previous deployment are also included in the analysis. The mean velocity profiles are logarithmic, and the vertical distribution of Reynolds stresses normalized by the square of the free stream velocity collapse well for data obtained at the same elevation but at different phases of the tidal cycle. The magnitudes of ͗uЈuЈ͘, ͗wЈwЈ͘, and Ϫ͗uЈwЈ͘ decrease with height above bottom in the 25-160-cm elevation range and are consistent with the magnitudes and trends observed in laboratory turbulent boundary layers. If a constant stress layer exists, it is located below 25-cm elevation. Two methods for estimating dissipation rate are compared. The first, a direct estimate, is based on the measured in-plane instantaneous velocity gradients. The second method is based on fitting the resolved part of the dissipation spectrum to the universal dissipation spectrum available in Gargett et al. Being undervalued, the direct estimates are a factor of 2-2.5 smaller than the spectrum-based estimates. Taylor microscale Reynolds numbers for the present analysis range from 24 to 665. Anisotropy is present at all resolved scales. At the transition between inertial and dissipation range the longitudinal spectra exhibit a flatter than Ϫ5/3 slope and form spectral bumps. Second-order statistics of the velocity gradients show a tendency toward isotropy with increasing Reynolds number. Dissipation exceeds production at all measurement heights, but the difference varies with elevation. Close to the bottom, the production is 40%-70% of the dissipation, but it decreases to 10%-30% for elevations greater than 80 cm.

Six sets of particle image velocimetry (PIV) data from the bottom boundary layer of the coastal ocean are examined. The data represent periods when the mean currents are higher, of the same order, and much weaker than the wave-induced motions. The Reynolds numbers based on the Taylor microscale (Reλ) are 300–440 for the high, 68–83 for the moderate, and 14–37 for the weak mean currents. The moderate–weak turbulence levels are typical of the calm weather conditions at the LEO-15 site because of the low velocities and limited range of length scales. The energy spectra display substantial anisotropy at moderate to high wavenumbers and have large bumps at the transition from the inertial to the dissipation range. These bumps have been observed in previous laboratory and atmospheric studies and have been attributed to a bottleneck effect. Spatial bandpass-filtered vorticity distributions demonstrate that this anisotropy is associated with formation of small-scale, horizontal vortical layers. Methods for estimating the dissipation rates are compared, including direct estimates based on all of the gradients available from 2D data, estimates based on gradients of one velocity component, and those obtained from curve fitting to the energy spectrum. The estimates based on vertical gradients of horizontal velocity are higher and show better agreement with the direct results than do those based on horizontal gradients of vertical velocity. Because of the anisotropy and low turbulence levels, a −5/3 line-fit to the energy spectrum leads to mixed results and is especially inadequate at moderate to weak turbulence levels. The 2D velocity and vorticity distributions reveal that the flow in the boundary layer at moderate speeds consists of periods of “gusts” dominated by large vortical structures separated by periods of more quiescent flows. The frequency of these gusts increases with Reλ, and they disappear when the currents are weak. Conditional sampling of the data based on vorticity magnitude shows that the anisotropy at small scales persists regardless of vorticity and that most of the variability associated with the gusts occurs at the low-wave-number ends of the spectra. The dissipation rates, being associated with small-scale structures, do not vary substantially with vorticity magnitude. In stark contrast, almost all the contributions to the Reynolds shear stresses, estimated using structure functions, are made by the high- and intermediate-vorticity-magnitude events. During low vorticity periods the shear stresses are essentially zero. Thus, in times with weak mean flow but with wave orbital motion, the Reynolds stresses are very low. Conditional sampling based on phase in the wave orbital cycle does not show any significant trends.

Turbulence in shelf seas strongly affects the spread of pollution (such as oil spills 1 ) as well as the distribution of sediment 2 and phytoplankton blooms 3 . Turbulence is known to be generated intermittently close to the sea bed 4 , but little is known of its evolution through the water column, or to what extent it affects the surface. Here we present observations of the surface effects of bottom-generated turbulence in a tidally in¯uenced and well mixed region of the North Sea, as derived from acoustic and visual images. Although the sea bed in the area is¯at, we ®nd that at any one time, 20±30% of the water surface is affected by boilsÐ circular regions of local upwellingÐof diameter 0.960.2 times the water depth. The signature of individual boils persists for at least 7 minutes and, in accordance with laboratory 5,6 and numerical 7 studies, shows the appearance of eddies. The boils contribute to the replacement of surface waters from depth in unstrati®ed waters, and may therefore enhance the¯uxes of gases between atmosphere and ocean.There are no reported observations of the surface signature of boils in well mixed, open sea. It is, however, clear from dynamical measurements made using current meters that tidal¯ows generate turbulence through the action of shear stress at the sea bed in much the same manner as in channels in laboratory experiments 4,8 . Fluid is intermittently ejected away from the bottom in turbulent bursts that may reach vertical speeds of 25% of the forcing current 4 . Numerical 7 and laboratory 5,6 studies show that the upwardmoving water produces a boil as it impinges on the water surface.Upward-pointing side-scan sonar has been used to observe a wide variety of processes in the upper ocean 9±13 . In experiments designed to study the processes leading to dispersion of an oil plume in the tidally well mixedÐand consequently unstrati®ed 14 Ðsouthern North Sea, a two-beam side-scan sonar system was mounted on a frame set on the sea bed at a depth of 45 m. The sonars operate at 80 kHz and 90 kHz, are set at 908 apart in the horizontal, and produce vertical fan-like beams with axes aligned upwards, 208 from the horizontal 15 . The site is 56 km from the shore and the local sea bed is¯at, with no sand banks or other notable bed forms. The principal acoustic scatterers are bubbles, of typical diameter 20± 200 mm, produced in clouds by wind waves as they break 16 . The clouds are detectable to ranges of ,150 m along the sea surface from the sonar, bubbles accumulating in regions of surface convergence and downwelling 17,18 . Oil reduces wave breaking and acoustic scatter from the sea surface 10 . Acoustic observations were made, however, at least 0.5 km from the oil plume, and in a variety of wind speeds from near calm to 14 m s -1 and in tidal currents reaching 1 m s -1 . letters to nature NATURE | VOL 400 | 15 JULY 1999 | www.nature.com 251 Figure 1 Sonar and video images of boils at the sea surface. The scale is the same in both images (and along both axes), and the tidal current is from ...