Factors controlling floc settling velocity along a longitudinal estuarine transect

Abstract

A 147 km longitudinal transect of flocculated cohesive sediment properties in San Francisco Bay (SFB) was conducted on June 17th, 2008. Our aim was to determine the factors that control floc settling velocity along the longitudinal axis of the estuary. The INSSEV-LF video system was used to measure floc diameters and settling velocities at 30 stations at a distance of 0.7 m above the estuary bed. Floc sizes (D) ranged from 22 μm to 639 μm and settling velocities (Ws) ranged between 0.04 mm·s − 1 and 15.8 mm·s − 1 during the longitudinal transect. Nearbed turbulent shear stresses throughout the transect duration were within the 0.2–0.5 Pa range which typically stimulates flocculation growth. The individual D–Ws–floc density plots suggest the suspended sediments encountered throughout SFB were composed of both muddy cohesive sediment and mixed sediments flocs. Mass-weighted population mean settling velocity (Ws mass ) ranged from 0.5 mm·s − 1 to 10 mm·s − 1 . The macrofloc and microfloc (demarcation at 160 μm) sub-populations demonstrated parameterised settling velocities which spanned nearly double the range of the sample mean settling velocities (Ws mean ). The macroflocs tended to dominate the suspended mass (up to 77% of the ambient suspended solid concentration; SSC) from San Pablo Bay to Carquinez Strait (the vicinity of the turbidity maximum zone). Microfloc mass was particularly significant (typically 60–100% of the SSC) in the northern section of South Bay and most of Central Bay. The transect took eleven hours to complete and was not fully synoptic. During slack tide, larger and faster settling flocs deposited, accounting for most of the longitudinal variability. The best single predictor of settling velocity was water velocity 39 min prior to sampling, not suspended-sediment concentration or salinity. Resuspension and settling lags are likely responsible for the lagged response of settling velocity to water velocity. The distribution of individual floc diameters and settling velocities indicates that floc density for a given floc diameter varies greatly. A small portion (a few percent) of suspended sediment mass in SFB is sand-sized and inclusion of sand in flocs appears likely. Fractal theory for cohesive sediment assumes that there is a single primary particle size that flocculates, which is not the case for these types of mixed sediment flocs. The wide variability in the physical, biological and chemical processes which contribute to flocculation within SFB means that spatial floc data is required in order to accurately represent the diverse floc dynamics present in the Bay system. The importance in determining accurate estimates of floc density has been highlighted by the SFB data, as these provide the basis for realistic distributions of floc dry mass and the mass settling flux across a floc population. However, although video floc sampling devices can produce the various floc property trends observed in SFB, good survey practice is still paramount. One can see that if the sampling coverage (i.e. data collection frequency) is poor, this could lead to potential mis-interpretations of the data and only limited conclusions may be drawn from such a restricted survey. For example, a limited survey (i.e. only 3 stations, compared to the 10 stations in the full survey) in South Bay produces an under-estimate in both the macrofloc SSC macro distribution by a factor of four and the Ws macro by a factor of two. To develop sediment transport numerical models for SFB, high quality floc size and settling data are needed to understand and simulate the depositional qualities of both suspended cohesive sediment and mixed sediments in San Francisco Bay. This study has shown that the most pragmatic solution is a physically-based approach, whereby the detailed flocs D vs. Ws spectra are parameterised in terms of their macrofloc and microfloc properties. This aids in model calibration, whilst retaining more of the dynamical aspects of the floc populations. All forms of flocculation are dynamically active processes, therefore it is important to also include both SSC and turbulence functions together with the floc data.