Gold mining by artisinal (small-medium scale) miners causes an immense amount of damage to the environment (i.e. soil erosion, mobilization of heavy metals, etc.).1, 2 One of the most popular gold mining techniques employed by artisinal miners in Guyana is mercury amalgamation. During the amalgamation process approximately 300 metric tons/yr 11, 12 of mercury is used. Mercury once in the environment can be transported through the air, soil, and water column. It is estimated that 90-99% of total mercury (THg) is associated with the sediment. An understanding of the geochemical conditions that affect the fate of mercury in soils, which can act as potential sinks or sources for mercury, can provide solutions for reduced environmental impacts of mercury contamination. Local Guyanese agencies have become concerned with the quality of the water, soil, biota, and human impact in remote locations in the interior of Guyana. Therefore, soil samples were collected from two local mines in Guyana's Arakaka-Mathew's Ridge area. Two soil samples (Pakera Creek and Philip's Mine) and a commercially available iron-oxide sorbent, Kemiron, underwent CVAAS, BET surface area analysis, electron dispersion spectroscopy, and x-ray diffractometry. THg concentrations for recovered soil samples were approximately 300 ng/kg. In addition, samples were subjected to batch equilibrium sorption studies as a function of pH and mercury species/concentration added as Hg(NO3)2 and HgCl2. All samples showed significant amounts of sorption between pH 3-9 for 100-1,000 ppb Hg added as Hg(NO3)2. When HgCl2.was added to the batch reactor containing Kemiron, an iron-oxide surface, the adsorption behavior of Hg2+ decreased. Philip's Mine solids, characterized as silicon dioxide by BET, had the lowest surface area (4 m2/g) and sorption when added as Hg(NO3)2 and HgCl2. On the other hand, Kemiron and Pakera Creek displayed similar sorption behaviors with high sorption across all pH ranges. This may be due to similar chemistry and larger surface areas. Surface loadings were 200 mg/kg and 2,000 mg/kg for experiments with 100 ppb Hg and 1,000 ppb Hg, respectively. Further analysis is required to identify the binding mechanisms between mercury and samples as well as the role of organic matter content on samples.

The thesis deals with the numerical modeling of nonreactive and nonlinearly sorbing solutes in groundwater and analysis of the effect of heterogeneity resulting from spatial variation of physical and chemical parameters on the transport of solutes. For this purpose, a hy­brid flux corrected transport (FCT) and central difference method based on operator-split approach is developed for advection-dispersion solute transport equation. The advective transport is solved using the FCT technique, while the dispersive transport is solved using a conventional, fully implicit, finite difference scheme. Three FCT methods are developed and extension to multidimensional cases are discussed. The FCT models developed are anlaysed using test problems possessing analyt­ical solutions for one and two dimensional cases, while analysing advection and dispersion dominated transport situations. Different initial and boundary conditions, which mimic the laboratory and field situations are analysed in order to study numerical dispersion, peak cliping and grid orientation. The developed models are tested to study their relative merits and weaknesses for various grid Peclet and Courant numbers. It is observed from the one dimensional results that all the FCT models perform well for continuous solute sources under varying degrees of Courant number restriction. For sharp solute pulses FCT1 and FCT3 methods fail to simulate the fronts for advection dominated situations even for mod­erate Courant numbers. All the FCT models can be extended to multidimensions using a dimensional-split approach while FCT3 can be made fully multidimensional. It is observed that a dimensional-split approach allows use of higher Courant numbers while tracking the fronts accurately for the cases studied. The capability of the FCT2 model is demonstrated in handling situations where the flow is not aligned along the grid direction. It is observed that FCT2 method is devoid of grid orientation error, which is a common problem for many numerical methods based on Cartesian co-ordinate system. The hybrid FCT2 numerical model which is observed to perform better among the three FCT models is extended to model transport of sorbing solutes. The present study analyses the case of nonlinear sorption with a view to extend the model for any reactive transport situation wherein the chemical reactions are nonlinear in nature. A two step approach is adopted in the present study for coupling the partial differential equation gov­erning the transport and the nonlinear algebraic equation governing the equilibrium sorp­tion. The suitability of explicit-implicit (EI - form) formulation for obtaining accurate solution coupling the transport equation with the nonlinear algebraic equation solved using a Newton-Raphson method is demonstrated. The performance of the numerical model is tested for a range of Peclet numbers for modelling self-sharpening and self-smearing con­centration profiles resulting from nonlinear sorption. It is observed that FCT2 model based on this formulation simulates the fronts quite accurately for both advection and dispersion dominated situations. The delay in the solute mobility and additional dispersion are anal­ysed varying the statistical parameters characterising the heterogeneity namely, correlation scale and variance during the transport of solutes and comparisons are drawn with invariant, cases. The impact of dispersion during the heterogeneous transport is discussed.