The primary objective of this research was to investigate multicomponent transport of five major cations, Ca2+, Mg2+, NH4+, K+ and Na+, in laboratory soil columns. The soil columns were packed with soils from two different sites and were equilibrated with fresh groundwater from each respective site. Experimental data was obtained by flushing a simulated contaminant through the soil columns. The soil columns were then flushed with fresh groundwater to simulate decommissioning activities. The breakthrough data and soil exchange capacities obtained from both tests were used to identify key processes affecting the transport of the geochemical species. During the simulated contaminant flushing stage, NH4+ and K+ replaced Ca2+ and Mg2+ on the soil exchange sites. Breakthrough of NH4+ was attenuated by factors of 3.2 and 6 for Sites 1 and 2 soils, respectively. Breakthrough of K+ was attenuated by factors of 3.2 and 5.4 for Sites 1 and 2 soils, respectively. Generally, ions with higher valency will exchange for those of lower valency, but in this case the majority of the ions (NH4+ and K+) in the solution has a lower valency and will exchange with those of higher valency by mass action. Ca2+ was the first to be replaced, followed by Mg2+ once the ionic strength of the solution increased. The displacement of calcium and magnesium created a concentration pulse of these cations that coincides with the chloride breakthrough curve. Calcium and magnesium concentrations reached up to approximately 275% and 2000%, respectively, higher than the freshwater originally in the column. During the freshwater flushing stage, freshwater infiltrated the soil columns to assess the permanency of contaminant attenuation and to identify the geochemical mechanisms of contaminant release. Concentrations of NH4+ and K+ declined quickly. Ninety-five percent of attenuated NH4+ was released by the soil. Therefore, the attenuation of NH4+ is reversible but this occurs over several pore volumes at concentrations lower than those in the simulated contaminant and therefore would not result in a mass loading to the environment. Cation exchange was identified as the mechanism responsible for the release of the adsorbed ammonium and potassium into the soil pore water.

The primary objective of this research was to investigate multicomponent transport of five major cations, Ca2+, Mg2+, NH4+, K+ and Na+, in laboratory soil columns. The soil columns were packed with soils from two different sites and were equilibrated with fresh groundwater from each respective site. Experimental data was obtained by flushing a simulated contaminant through the soil columns. The soil columns were then flushed with fresh groundwater to simulate decommissioning activities. The breakthrough data and soil exchange capacities obtained from both tests were used to identify key processes affecting the transport of the geochemical species. During the simulated contaminant flushing stage, NH4+ and K+ replaced Ca2+ and Mg2+ on the soil exchange sites. Breakthrough of NH4+ was attenuated by factors of 3.2 and 6 for Sites 1 and 2 soils, respectively. Breakthrough of K+ was attenuated by factors of 3.2 and 5.4 for Sites 1 and 2 soils, respectively. Generally, ions with higher valency will exchange for those of lower valency, but in this case the majority of the ions (NH4+ and K+) in the solution has a lower valency and will exchange with those of higher valency by mass action. Ca2+ was the first to be replaced, followed by Mg2+ once the ionic strength of the solution increased. The displacement of calcium and magnesium created a concentration pulse of these cations that coincides with the chloride breakthrough curve. Calcium and magnesium concentrations reached up to approximately 275% and 2000%, respectively, higher than the freshwater originally in the column. During the freshwater flushing stage, freshwater infiltrated the soil columns to assess the permanency of contaminant attenuation and to identify the geochemical mechanisms of contaminant release. Concentrations of NH4+ and K+ declined quickly. Ninety-five percent of attenuated NH4+ was released by the soil. Therefore, the attenuation of NH4+ is reversible but this occurs over several pore volumes at concentrations lower than those in the simulated contaminant and therefore would not result in a mass loading to the environment. Cation exchange was identified as the mechanism responsible for the release of the adsorbed ammonium and potassium into the soil pore water.

There are many circumstances where lateral flows in the upper soil layers above the regional groundwater table are important for hillslope and catchment hydrology, and in particular for the transport of contaminants. Perched water tables frequently occur in Quaternary drift sequences, reducing rates of recharge to the underlying aquifers and altering contaminant migration pathways; recent experimental and modelling studies have demonstrated the potential importance of lateral flows in the unsaturated zone, even in homogeneous soils; and lateral interflow at the hillslope scale, and its role in generating storm runoff, is the subject of intense current debate amongst hydrologists. A numerical model for simulating transient three-dimensional variably-saturated flow in complex aquifer systems (the Variably-Saturated Subsurface flow, or VSS, model), capable of representing these conditions, is presented in this thesis. The VSS model is based on the extended Richards equation for saturated as well as unsaturated conditions, and also includes capabilities for modelling surface-subsurface interactions, stream-aquifer interactions, prescribed head and flow boundary conditions, plant and well abstractions, and spring discharges. A simple but novel approach is taken to solving the three-dimensional non-linear Richards equation on a flexible-geometry finite-difference mesh, using Newton-Raphson iteration and an adaptive convergence algorithm. The VSS model is implemented as a module of the catchment flow and transport modelling system, SHETRAN. The reliability of the full SHETRAN modelling system is demonstrated using verification and validation tests, including comparisons against analytical solutions for simple cases, and simulations of storm runoff in a small Mediterranean catchment. Simulations of flow and contaminant transport in complex sequences of Quaternary drift deposits demonstrate the full capabilities of the modelling system under real-world conditions.

Experiments exploring transport and bio-containment of contaminants in fractured rock were completed using fractured-limestone samples obtained in eastern Ontario, Canada. Three single-fracture samples, a fracture-intersection sample and a fracture-network sample were set into vertical flow systems. Three phases of experiments focused on the transport and hydraulic properties of each sample, the effects of biobarriers on diffusion processes in fracture rock, and methods of improving biobarrier stability and survivability. Hydraulic apertures were determined from constant-flow measurements and transport properties were interpreted from Lissamine and KBr tracer experiments with velocities of up to 8500 m/d for all five samples. At Re > 16, linear to non-linear transitions were observed in enlarged single fracture A and the fracture intersection samples. Reversible increases in aperture were observed at Reynolds numbers (Re) of 7, 4, and 3 for single fractures A and B, and the fracture-network, respectively. Non-linear effects were not observed in these samples over the range of velocities studied (up to Re = 20). Results from the 1-D analytical transport model overestimated values of matrix porosity, suggesting that diffusion from dead zones and slow-flowing regions are also contributing to observed breakthrough curves. Methods of improving biobarrier stability in fractured rock were studied in two single-fracture samples and the fracture-network sample by stimulating naturally-occurring groundwater bacteria. Survivability was improved with successive cycles of feeding and starving and stimulating growth at lower temperatures. Modeled values of matrix porosity decreased by up to 50%, indicating that diffusion processes are strongly influenced by biofilm development. Back diffusion of Lissamine was measured using one single-fracture sample and the fracture-intersection sample. Lissamine was allowed to diffuse into the matrix of each sample and, following a suitable loading period, the back-diffusion of residual Lissamine concentrations were measured from the outflow. This was done in the presence and absence of biofilm, and following the introduction of biofilm onto the fracture surfaces, diffusion was no longer a governing process affecting transport and only advective transport was observed. Experiments were interpreted using a 3-D finite difference model with a three-layer porosity approach, and indicated a decrease in aperture and porosity following biostimulation. / Thesis (Ph.D, Civil Engineering) -- Queen's University, 2012-11-22 11:23:24.065

The purpose of this project is to gain an improved understanding of the transformations that occur in the subsurface downstream of on-site wastewater disposal systems and septic systems. These systems are used widely throughout the United States to treat and discharge wastewater effluent. The approach involved the collection of samples from a septic research center in Cape Cod, MA, and analysis of these samples for nitrogen, phosphorus, dissolved oxygen, pH, alkalinity, suspended solids, metals, and other water quality parameters. Inverse modeling was used to compare samples collected upstream and downstream of subsurface “leaching” fields consisting of sand beds. This approach provided a basis to identify key reactions occurring in the subsurface downstream of the discharge. In addition, a reactive transport software package, based on the PHREEQC and Hydrus-1d models, was used to model the transport in these sand beds and identify possible reactions and changes in contaminant concentrations with depth. To understand the implications of the discharges, an additional field study was completed in an area where septic systems have impacts on surface waters. Samples collected from a stream provided an indication of the loads entering the stream as a result of septic system discharges. Combining the results from the modeling with the results of this field investigation provided an approach to estimate the transport of nutrients and other contaminants entering the surface waters from septic system discharges. The results provide a basis for understanding the impacts of septic systems on surface waters, and develop better approaches for reducing the impacts of these discharges.

Mobilization of colloidal particulate matter (most important, clay particles) from a soil matrix in the subsurface environment is an important environmental process. As many contaminants tend to adsorb onto various colloidal mineral particles, co-transport of contaminants in association with mobilized particles could contribute significantly to the migration of these contaminants in the environment. Numerous studies have observed the effects of pH on colloid mobilization but have overlooked the possible direct role of acidity. This study looked at the role of acidity with H⁺ as a chemical agent. Through cyclic elution of a natural sand column with a weak acid and base solution, there was an increase in mobilized clay colloids. It was found that low concentrations of organic acids could assist in detaching surface clays through lysing of labile Ca²⁺ and Mg²⁺ ions. The H⁺ ions sever the chemical bonds between the grain surface and the colloidal surface by being substituted for the interstitial Ca and Mg ions. This substitution has been found to release over 1 kg of surface clay per 1 mole of H⁺ consumed. It was postulated that pH oscillation addition to proton dynamics could play a major role in subsurface colloid transport. The results from this study could help improve predicting of subsurface contaminant fronts and aid in managing contaminant transport in the soil water environments.

Multicomponent nonaqueous phase liquids (NAPL) contaminating the subsurface can significantly inhibit remediation. One method of enhancing the rate of remediation of NAPL constituents, compared to pump-and-treat, involves source zone treatment with enhanced solubilization agents (ESAs) including cyclodextrins. Equilibrium cyclodextrin enhanced solubilization of simple 1, 2, and 3 component NAPL mixtures was examined to evaluate the applicability of Raoult's Law. The results suggest that Raoult's Law may be used to estimate equilibrium and early-time dynamic concentrations in contact with ideal NAPL mixtures, and Raoult's Law may be used to estimate cyclodextrin enhanced groundwater concentrations for ideal NAPL mixtures. Solubility enhancement of NAPL compounds was dependent on the cyclodextrin concentration and independent of NAPL composition. Column experiments and numerical modeling were used to evaluate the dissolution behavior of the NAPL mixtures in water and a cyclodextrin solution to estimate mass transfer rates. The aqueous multicomponent dissolution followed Raoult's Law, and the model-estimated lumped rate coefficients were independent of the NAPL composition. Addition of the cyclodextrin enhanced the dissolution and removal of compounds from residual NAPL due to an increase in the driving force (i.e. concentration gradient) and the mass transfer coefficient. The model results suggest that Raoult's Law is applicable for ideal NAPL mixture dissolution in water, but potential nonideality was observed and caused the model simulation to deviate from the dissolution behavior for NAPL mixture cyclodextrin experiments. The cyclodextrin dissolution experiments were less rate-limited than aqueous dissolution, and the mass transfer coefficients were quantified with the model. The results of the model suggest that NAPL mixture nonideality and intra-NAPL diffusion may also impact enhanced dissolution behavior. Additionally, the importance of NAPL mixture characterization was illustrated by evaluation of a mixture of PCE (tetrachloroethene) and diesel fuel collected from a site in Tucson, Arizona. A sample from the site was used to create mixtures with increasing PCE in the NAPL. Chemical evaluation of the complex NAPL was conducted, and physical property and phase partitioning testing was performed, which demonstrated the effect of NAPL composition on its distribution, interphase mass transfer, and potential mobilization.

Hydraulic conductivity (K) is a key parameter required for the accurate prediction of contaminant transport in an aquifer. Traditionally, pump tests, slug tests, grain size analysis and, to a lesser extent, tracer tests have been employed to estimate the K of an aquifer. These methods have disadvantages in respect to assessing the K of a contaminated aquifer, for example, pumping tests generate large quantities of potentially contaminated water, slug tests interrogate only a small portion of aquifer to generate K values, and tracer tests are costly to perform. The recirculating well pair (RWP) system, assessed in this study, attempts to minimise these disadvantages while producing accurate estimates of K. The RWP system uses two wells, each screened in two positions; one screen injects water and the other extracts water from the aquifer. One well extracts water from the lower screen and injects it into the aquifer via the upper screen, whereas the second well extracts water from the upper screen and injects it through the lower screen. When these two wells are pumped in tandem a recirculation system is created within the aquifer. No water is lost or gained from the aquifer in this system. Hydraulic conductivity can be estimated from a RWP system by either the multi dipole or the fractional flow methods. The multi dipole method estimates K by measuring steady state hydraulic heads, whereas the fractional flow method uses a tracer test to obtain steady state concentrations at the four screens to estimate K. Both methods utilise a 3D flow model to simulate the aquifer system. Inverse modelling in conjunction with a genetic algorithm simulate the hydraulic head values obtained from the multi dipole experiments or the tracer steady state values obtained from the fractional flow method. Hydraulic ii conductivity estimates are obtained by matching the simulated and observed steady state hydraulic head, or tracer steady state values. An investigation of the accuracy of the two RWP methods, when system parameters are varied, in estimating K values was undertaken. Five multi dipole experiments were undertaken with varying dipole flow rates to assess the effect of altering dipole flow rate on estimates of K. Two experiments were also undertaken to assess the effect of altering the pumping well incidence angle as compared to the regional flow on the accuracy of K estimates. Five fractional flow experiments were conducted, four to assess the effect of changing dipole pumping rates and one to assess the influence of altering the incidence angle of the pumping wells on estimation of K. All experiments were undertaken in an artificial aquifer that allowed control of hydraulic parameters and accurate measurement of aquifer K by independent methods. Experimental results were modelled with the two RWP methods. Results indicate that both the multi dipole and fractional flow methods provide accurate estimates of the K of the artificial aquifer (5 % to 57% greater than the actual K and -14% to 17% of the actual K, respectively). Altering the ratio between the pumping well and regional aquifer flow rates had no effect on the estimated K results in both methods. Although preliminary results were positive, further work needs to be undertaken to determine if changing the orientation of the well pairs affects the estimation of K.

Understanding the geology and the hydrology of the subsurface is important to model the fluid flow and the behavior of the contaminant. It is essential to have an accurate knowledge of the movement of the contaminants in the porous media in order to track them and later extract them from the aquifer. A two-dimensional flow model is studied and then applied on a linear contaminant transport model in the same porous medium. Because of possible different sources of uncertainties, the deterministic model by itself cannot give exact estimations for the future contaminant state. Incorporating observations in the model can guide it to the true state. This is usually done using the Kalman filter (KF) when the system is linear and the extended Kalman filter (EKF) when the system is nonlinear. To overcome the high computational cost required by the KF, we use the singular evolutive Kalman filter (SEKF) and the singular evolutive extended Kalman filter (SEEKF) approximations of the KF operating with low-rank covariance matrices. The SEKF can be implemented on large dimensional contaminant problems while the usage of the KF is not possible. Experimental results show that with perfect and imperfect models, the low rank filters can provide as much accurate estimates as the full KF but at much less computational cost. Localization can help the filter analysis as long as there are enough neighborhood data to the point being analyzed. Estimating the permeabilities of the aquifer is successfully tackled using both the EKF and the SEEKF.

<p>This thesis compares different methods for risk assessment of a creosote contaminated site in Vansbro, Dalarna. In a complementary study the importance of using representative data for calibration and validation of stationary groundwater models is discussed. The time dependency between surface water levels and groundwater levels on the site in Vansbro is evaluated by step response analysis. The results show that the groundwater levels reflect only long-term variation in the surface water levels. Hence the use of such long-term means for calibration and validation of the stationary groundwater models is the only theoretically defendable alternative.</p><p>As a first step in the comparison of risk assessment methods the degree of contamination is evaluated on the basis of generic guideline values for contaminated soils, developed by the Swedish Environmental Protection Agency. The guidelines specifically developed for contaminated petrol stations are also taken into consideration. Secondly, a site-specific evaluation is carried out. The transport of contaminants from the soil to the on-site groundwater and nearby surface water is calculated using the mathematical models behind the aforementioned guideline values. The calculated groundwater concentrations show a poor agreement with observed data and the thesis points out several possible explanations of this deviation. A sensitivity analysis is carried out and the influence of biodegradation during the transport in the aquifer is evaluated. The results indicate that the deviation between observed and calculated groundwater concentrations cannot be explained by the absence of biodegradation in the mathematical models, but rather by other simplifications and uncertainties in estimated model parameters.</p><p>Finally, a risk assessment is made on the basis of observed groundwater concentrations instead of observed soil concentrations as in the assessments above. Observed groundwater concentrations are compared with guideline values for groundwater on contaminated sites. The risk of negative environmental effects in the surface water is evaluated based on the estimated inflow of contaminated groundwater and its concentration of contaminants. The inflow is estimated from the interpolation of observed groundwater levels as well as with two numerical groundwater models of varying hydrogeological complexity: a three-dimensional flow model developed in Visual MODFLOW and a two-dimensional flow model developed in TWODAN. The models are compared with respect to the calculated inflows of contaminated groundwater into the surface water body as well as their ability to reproduce observed groundwater levels. The results indicate that the differences between the models, with respect to the calculated inflows, are of lesser importance to the risk assessment in comparison with the uncertainties associated with other parameters. The choice to base the assessment on observed contaminant concentrations in either the soil or the groundwater is shown to be of crucial importance.</p> / <p>I detta examensarbete jämförs olika metoder att bedöma de hälsomässiga och miljömässiga risker som föreligger med ett kreosotförorenat markområde i Vansbro, Dalarna. I en kompletterande undersökning belyses vikten av att stationära grundvattenmodeller kalibreras och valideras mot tidskompatibla indata. Tidsberoendet mellan yt- och grundvattennivåer på området i Vansbro studeras genom stegsvarsanalys. Resultaten visar att grundvattennivåerna på området anpassar sig mycket långsamt efter rådande ytvattennivåer. Grundvattennivåerna återspeglar således medelvärden av ytvattennivåerna över en mycket lång tidsperiod. Därför är användandet av sådana medelvärden för kalibrering och validering av modellerna det enda teoretiskt försvarbara alternativet.</p><p>Som ett första steg i jämförelsen mellan olika riskbedömningsmetoder utvärderas föroreningssituationen baserat på uppmätta föroreningskoncentrationer i jorden och med hjälp av Naturvårdsverkets generella samt branschspecifika riktvärden för förorenad mark. Därefter genomförs en platsspecifik riskbedömning. Föroreningstransporten från markområdet till grundvattnet samt till ett närbeläget ytvatten beräknas med de matematiska modeller som användes av Naturvårdsverket vid framtagandet av ovannämnda riktvärden. Beräknade grundvattenkoncentrationer jämförs med uppmätta koncentrationer och resultaten visar på stora avvikelser. En känslighetsanalys utförs och inverkan av biologisk nedbrytning i akviferen utvärderas. Resultaten tyder på att avvikelsen mellan uppmätta och beräknade grundvattenkoncentrationer ej förklaras av att biologisk nedbrytning inte beaktas i modellerna utan att avvikelsen härrör från andra förenklingar och från osäkerheter i indata.</p><p>Slutligen genomförs en platsspecifik riskbedömning, motsvarande den ovan, men denna gång utifrån uppmätta föroreningskoncentrationer i grundvattnet. De uppmätta grundvattenkoncentrationerna jämförs direkt mot Naturvårdsverkets riktvärden för grundvatten på förorenade områden. Risken för negativa ekologiska effekter i ytvattnet utvärderas utifrån det beräknade inflödet av förorenat grundvatten samt dess koncentration. Inflödet beräknas överslagsmässigt utifrån interpolerade grundvattennivåer samt med två numeriska grundvattenmodeller av olika hydrogeologisk komplexitet: en tredimensionell flödesmodell uppbyggd i Visual MODFLOW och en tvådimensionell modell uppbyggd i TWODAN. Grundvattenmodellerna jämförs med avseende på det beräknade inflödet och med avseende på deras förmåga att beskriva observerade grundvattenförhållanden.Resultaten tyder på att skillnaden i det, med de olika metoderna, beräknade inflödet är av liten betydelse för riskbedömningen i förhållande till osäkerheter förknippade med övriga parametrar. Störst betydelse för riskbedömningen får valet att utgå från uppmätta föroreningskoncentrationer i antingen jorden eller grundvattnet.</p>