Thesis abstract:

«p»Water flow and contaminant migration in soils are governed by the Richards Equation and the diffusion-advection equation jointly. This project aims to develop new probabilistic solutions of coupled governing equations and apply them to the risk assessment of groundwater contamination by PFAS in surface sources, including landfills or other waste repositories. Firstly, statistical characterization of relevant hydraulic and diffusive parameters are determined by Bayesian inference, based on experimental data on PFAS migration through, and sorption on, geosynthetic clay liners. Secondly, for a given model of the liner-subsoil system, contaminant migration simulations, whose essence is solving equations, can be realized with existing numerical software. Meanwhile, we will further explore the application of advanced techniques, such as neural networks, to provide a superior scheme for equation-solving. Thirdly, the probabilistic analysis of PFAS migration through base liners and underlying soil will be carried out. That is, by establishing random fields of property parameters, Monte Carlo simulations will be performed, leading to statistical results that characterize the contamination risk of PFAS.«/p»