Quantification of cohesive sediment transport is of fundamental importance in many engineering and ecological applications such as the quantification of yield of sediments from watersheds, reactivation of contaminated mud from industrial and other effluents, transport of nutrients for benthic organisms in aquatic ecosystem, etc. However, not much progress has yet been made in understanding cohesive sediment dynamics due to the inherent physical, chemical and biological / microbiological complexity of natural cohesive sediments. Both field (in-situ) and laboratory measurements of cohesive sediment erosion are often based on an experimental procedure in which constant levels of bed shear stress are applied step-wise over fixed time steps or intervals (e.g., 10 to 20 min, Maa et al. 1998; Houwing 1999; Ravens and Gschwend 1999; Aberle et al. 2003). A desired output from such measurements is the total surface erosion rate E defined as mass of sediment eroded per unit bed area per unit time. It is related to the time rate of change in bed elevation, , as , where z = the bed surface elevation with an arbitrary origin (positive upwards); t = time; and = the dry bulk density of bed material (Sanford and Maa 2001). In practice, E is usually calculated using measured fluxes of suspended sediment concentration (SSC) within the flume erosion channel, i.e., explicitly or implicitly assuming that the total erosion rate E is equivalent to the re-suspension rate. In general, however, the erosion rate (E) consists of two components, re-suspension rate (ER) and bed load component (Eb). Henceforth, throughout the paper the symbol E will indicate the total erosion rate, i.e., comprising both re-suspension and bed load. Re-suspension refers to sediments that go straight into suspension after erosion occurred. Some portion of the eroded sediment, in general, moves as bed load. This component is not taken into account in most cohesive sediment studies, although in principle it may contribute significantly to the total erosion (Mitchener and Torfs 1996; Aberle et al. 2004). The contribution of ER and Eb to E for flow through flumes can be seen in the equation of sediment conservation inside the erosion channel given by:
(1)
where C = SSC; H = height of the channel; q = specific water discharge; qb = specific bedload; and x = longitudinal co-ordinate along the flow direction (positive downstream) with an arbitrary origin. The data reported in the present study was obtained using NIWA in-situ flume II (Debnath et al. 2005) where E was measured with the Ultrasonic Ranging System (URS), ER was estimated from SSC measurements, and the bed load erosion component was calculated using Eq. (1). In this paper, we report new data obtained in several fresh and salt water environments in New Zealand using the recently developed NIWA in-situ flume II. The paper mainly focuses on (a) assessment of the total erosion rate, re-suspension rate, and bed load erosion component, based on Eq. (1); and (b) evaluation of dependency of ER / E on flow and sediment characteristics.
THE NIWA II IN-SITU SEDIMENT FLUME
NIWA in-situ flume II is a straight flow through flume designed to work under submersible conditions. It consists of an erosion channel through which water is sucked by rotation of an impeller driven by an air motor. The flume erosion channel is 0.6 m long, 0.16 m wide and 0.08 m high with a 0.14 m long entrance section. Water turbidity is monitored with two optical backscatter sensors (D&A, OBS-3). OBS-3 turbidity readings are calibrated against SSC for each in-situ experiment separately, where SSC is determined from at least six in-situ water samples taken with two small pumps near the OBS-3 sampling locations. A key improvement over existing in-situ flumes is that the NIWA flume II is equipped with seven SeaTek 5 MHz Ultrasonic Ranging System (URS) that allow non-invasive measurement of bed elevations inside the flume from which E can be directly estimated. Flow velocities are measured by an electromagnetic flow meter (EFM) located downstream of the erosion channel. Outputs from the OBS-3 probes, SeaTek sensors, and electromagnetic flow meter are sampled and logged into a data-logging unit with a frequency of 1 Hz. Water conductivity and water temperature are also recorded. To relate velocity readings to the bed shear stress, the flume was calibrated over three categories of bed surfaces (real cohesive bed, sandpaper, and flat wooden bed), using two Micro acoustic Doppler velocimeters (Micro-ADV). The flume and field experiments can be run in automatic or semi-automatic regimes when users can control most operations of the flume using a laptop and user-friendly Windows-based software.
INTRODUCTION
