This research has utilised high-resolution measurements of density and velocity to investigate the formation of a transient tidal intrusion front in a narrow, macrotidal estuary, and model the impact of such a frontal system on the accuracy of multibeam sonar surveys. The front was found to form during spring tides, when the barotropic inflow was sufficient to arrest the buoyant outflow from the estuary. This has been shown to be driven by changes in channel width and depth, creating a hydraulic control point. These changes in topography are demonstrated to interact with the flow m a similar manner to theoretical two-layer flow over the lee side of a sill. Enhanced shear at the density interface, provided by increasing barotropic and baroclinic flows during the flooding tide, eventually led to the decay of the frontal system. There was no surface manifestation of this front during neap tides. Further measurements have confirmed that the relatively weak barotropic flow at this time was not sufficient to overcome the stratification in a two layer regime. Frontal dynamics have been shown to conform to theoretical predictions, and an evaluation of the key frontal discrimmators has confirmed the validity of their use in such an environment. The development of a multibeam sonar refraction model has facilitated an assessment of the accuracy of hydrographic surveys conducted in the presence of a tidal intrusion fronts. Major reductions in swath width have been shown to be required when traversing a tidal intrusion front, with a flat sonar transducer array providing the most effective survey results. Undersampling the sound velocity field in the vicinity of a tidal intrusion front leads to major depth errors usmg all multibeam sonar transducer configurations; hence, accepted methods of sound velocity sampling in estuarine environments should be updated with immediate effect. Recommendations have been made that sampling in such an environment is undertaken at least hourly, at intervals of less than 50 m in order to maximise hydrographic survey efficiency.

One objective of this project is to set up a two-dimensional model for exploring the flushing process of trapped saltwater subject to upstream freshwater turbulent flow. The multiphase Eulerian model, a part of commercial code FLUENT6.2, has been applied for the first time to study this complex mixing interaction in estuary. The distinguishing characteristic of this model is to treat saltwater and freshwater as two single miscible phases instead of a mixture phase with density variation, and the advantage of using a multiphase approach over a single-phase model is that it can efficiently and accurately treat both the free water surface and relatively high density excess between two fluids simultaneously. The other objective of this project is to develop a three-dimensional model based on the FVCOM open source code, with the aim to better understand the estuarine hydrodynamics with or without the presence of typhoon. It is found that the original FVCOM code can not reproduce an accurate tidal hydrodynamics in estuary. An improved simulation of the bed friction has been incorporated into the existing code for estuarine tide. This model has also been developed by including air-pressure gradient term to study the hydrodynamic response to cyclonic typhoon. To include the effect of typhoon (wind stress and pressure deficit), a symmetrical cyclone model is adopted. However, the typhoon-induced wind field has been predicted poorly when the typhoon enters the near-shore region. This is because the typhoon quickly loses its symmetrical property in the near-shore region. To overcome this difficulty, an asymmetrical cyclone model is derived on the basis of characteristic isobar. The accuracy of open sea boundary for storm surge model has also been improved by using large scale model. Comparison results show a good agreement with numerical simulations and physical measurements.