Exposure risk modeling and analysis: Contaminant migration in shallow waters

The prediction of the risk of contaminant migration in water environment requires understanding of various complex physical processes ranging from the dynamics of shallow waters, turbulent mixing, to the stochasticity of wind gust.; In this thesis, a Monte Carlo simulation model is presented that enables the prediction of contaminant concentration in shallow waters with quantitative measures of exposure risk. The shallow water environment is simulated as a stochastic system and essential processes are captured.; A new approach is introduced to simulate wind gusts over the water surface as a space-time random process. The so-called LASAR wind generation model is a combined local average subdivision and autoregressive random field model and is conceptualized based on the assumption of separability of the space-time correlation structure of the wind field. Various desired statistical characteristics of a wind field can be accounted for using the model. Shallow water currents are computed with full nonlinear features and the strong influence from wind shear and tidal forcing are given particular consideration. The model is developed based on the well-known wave equation formulation. The advection-dispersion process is modeled by solving the depth averaged form of the advection-dispersion equation using finite elements. A new finite element implementation is introduced. The model may be supported by the current model for velocity inputs.; The effects of variability in the wind field on contaminant spreading are of particular concern in evaluating exposure risk in this study. Some useful statistical quantities of level excursions of contaminant concentration, such as probability of excursion and mean excursion areas, are estimated.; Overall, the Monte Carlo simulation model is aimed at facilitating more realistic risk assessment of contaminant spread in shallow waters. The simulation model paves the way for a much more detailed study on contamination risk in shallow water environment. Additionally, each embedded model is believed to be of general applicability and can be readily applied to other engineering applications.; Examples and comparisons with analytical or simulation results are provided for each submodel whenever possible. The methodology is finally illustrated with a Monte Carlo simulation example.