In situ particle size instrumentation for improved parameterisation and validation of estuarine sediment transport models
In estuaries containing cohesive sediment, flocculation and break-up of the suspended particles during the tidal cycle has implications for the monitoring and modelling of sediment transport. Monitoring of suspended sediment concentration using in situ optical or acoustic instruments is problematic since the amount of light or sound scattered from the suspended sediment is proportional to both the suspended concentration and the size of the particles. Numerical sediment transport models are heavily reliant upon such concentration data. Particle size variation also directly affects model parameterisation by influencing the settling velocity. A critical review of current particle sizing techniques shows that in situ imaging offers the best option in terms of cost, accuracy and versatility. This thesis presents a new, low cost video-based instrument for measuring the in situ particle size distribution. The system uses two CCTV cameras to view a total size range of 4 to 3000 microm. Illumination is provided by miniature microsecond flash units. These allow blur-free images of particles to be obtained in current speeds of up to 1.4 ms-1, which are saved to hard disk at a frame rate of up to 10 s-1. The instrument package is designed for small-boat operation and deployment in profile mode. Calculation of size and shape parameters is accomplished in software using automated image-processing algorithms. An efficient and accurate edge coincidence technique is developed to detect in focus particles. Instrument performance is evaluated through a case study of the Blyth estuary (Suffolk, UK). Particle size data from a small reach of the estuary are presented for both a spring and neap tide. The process of flocculation is clearly shown, and a semi-empirical model of particle size variation is derived based on turbulent intensity and suspended sediment concentration. The modelled sizes are used to derive settling velocity data for a 2DH model of sediment transport using a simplified model of floe density. Model output is improved in comparison to using a fixed value of settling velocity. Two distinct particle size subpopulations are observed which affect both the settling velocity and the calibration of ADCP backscatter data for sediment concentration between flood and ebb. In addition, rapid resuspension of bed material at the beginning of the flood tide is successfully simulated using a two-layer bed model. It is concluded that the new instrument is a valuable aid to monitoring and modelling of sediment transport.