Dispersion controlled natural attenuation : The role of conservative plume characteristics in reactive mixing processes

The aim of this project was to gain a fundamental understanding into competing mixing processes involved in natural/enhanced attenuation; processes which occur chiefly in the transition zone of a contaminant plume. Analytical and numerical methods, in combination with laboratory and field data, were employed to quantify these mixing processes. In particular, the focus of this work was on transverse dispersion, believed to be the most important mechanism resulting in the mixing of electron acceptors (EA) and electron donors (ED). A new two-dimensional analytical solution was derived, proving conclusively that transversal dispersion is indeed the most important dispersion mechanism leading to the attenuation of steady-state contaminant plumes where species (e.g. EA and ED) mix and instantaneously react together. This model was extended generically in three-dimensions and also to further include anaerobic core degradation processes, which have been demonstrated to play an important role in some remediation cases. This occurs when core degradation rates are comparable to fringe degradation, which is often the the case when EA availability is limited. It must be noted that although three-dimensional solutions can be formulated, in practice most problems can be considered in two-dimensions. In another study, a numerical model was developed to access the competing mechanisms of transverse mixing and kinetic-controlled biodegradation where two species react together. The base model consisted of three non-linear governing partial differential equations (PDE), describing EA and ED migration, and moreover microbial growth, in time. These can be reduced to two PDE by the introduction of a new parameter relating concentrations of EA and ED to each other. The PDE describing this variable can be solved using an appropriate "intermediate" analytical solution. In the steady-state, the equations reduce to just one PDE, in terms of one variable, requiring a numerical solution. The new model provides an elegant solution to the problem and agrees excellently with the base model. Furthermore, the new model allows an easy assessment of the effects of interacting dispersion and reaction parameters on plume development. The value of longitudinal dispersivity is paramount to provide an accurate solution, whilst transversal dispersivity and microbial growth control fringe width and the attenuation capacity of contaminant plumes. The combination of these effects can be seen as `effective' dispersion. In another study, a model-based interpretation of laboratory scale experimental data was undertaken. It was demonstrated that the classical form of the dispersion coefficient does not hold at the laboratory scale, when plumes of reacting species are considered. Using a new empirical formula combined with derived diffusion coefficients, the laboratory data could be accurately simulated using pure forward modelling techniques. In the last exercise, the analytical model derived in a previous study was applied to field data for a well documented petroleum hydrocarbon plume. The plume consists of a non-reactive benzene plume and a steady-state, reactive toluene plume. Extracting parameters from data profiles for conservative species, it was possible to accurately predict plume lengths of the reactive hydrocarbon plume. The results of these studies provide useful modelling tools, particulary in the assessment of studies in monitored natural attenuation