Reactive Transport in Soil and Groundwater

The unsaturated zone and the saturated/unsaturated interface region are important links between groundwater and the land surface. They provide storage capacity for both water and contaminants; a reactor medium for physical, chemical and biological processes; a delay time between the release of a contaminant into the unsaturated zone and its influx into the saturated zone and, a domain for the lateral and vertical transport of gases. We highlight some of the properties associated with these domains providing results of studies conducted in the phreatic Coastal Plain aquifer of Israel for the last 30 years.

Compound-specific isotope analysis is increasingly used to demonstrate contaminant degradation in groundwater. The method relies on the frequent occurrence of characteristic shifts in the isotope ratios of contaminants due to the faster degradation of molecules with light isotopes compared to those with heavy isotopes. The goal of the study was to evaluate if the method can also be used to assess biodegradation of petroleum hydrocarbons in the unsaturated zone. The study included laboratory experiments to determine isotopic enrichment for biodegradation of petroleum hydrocarbon compounds under unsaturated conditions and a field experiment. The field experiment consisted of burying an artificial fuel source in the unsaturated zone at a site in Denmark. Concentration and isotope ratios of individual compounds were monitored using a dense network of sampling points. Significant shifts of the isotope ratios of most of the compounds occurred. Initially, a depletion in ¹³C with distance was observed likely due to the faster diffusion of molecules with ¹²C only. Later, most of the compounds became in enriched in ¹³C compared to the source due to biodegradation. To evaluate the relative contribution of diffusion and biodegradation in more detail, the concentration and isotope ratio evolution was simulated using an analytical model. The calculations confirmed that diffusion can lead to significant isotope fractionation at early times. For hydrogen isotopes, the effect of diffusion relative to biodegradation is expected to be smaller due to the large hydrogen isotope fractionation frequently observed during biodegradation of organic compounds. In conclusion, the study demonstrates that especially at early times after spills, the effect of diffusion has to be taken into account in the interpretation of isotope data from the unsaturated zone. Furthermore, it demonstrates that for many compounds, hydrogen isotopes are likely the more sensitive indicator of biodegradation.

In this paper a model is proposed for describing persistent organic pollutants (POPs) bioavailability in soil. The model is written in Fortran 90 and describes POPs’ behaviour as resulting from four different processes: sorption-desorption equilibrium, slow diffusion (aged fraction), fast irreversible sorption (bound residues) and biodegradation of the bioavailable fraction.

The POP sorption to soil surfaces is described assuming a rapid rate of sorption-desorption to and from soil surfaces and a slower rate of diffusion into the internal matrix (aging). Biodegradation is described as resulting from bacterial growth using sigmoidal Monod kinetics for the contaminant dissolved in soil solution (for non-hydrophobic compounds) and first order kinetics for the degradation of the sorbed-available fraction. In the case of hydrophobic compounds, first-order kinetics is employed to describe also the degradation from the soil solution. Sorption and diffusion are approximated by first order kinetics. Finally, the formation of bound residues is described using an exponential saturation equation. The rate constants for the different processes are estimated using linear and non linear first order kinetics approaches. The rate constants of the irreversible processes are estimated from experimental data. Model evaluation was performed using data from previous experiments with phenanthrene as test compound.

Artificial groundwater recharge by percolation through the unsaturated zone is an important technique to enhance the water quality for drinking water supplies. Purification of the infiltration water results mainly from microbially mediated redox-reactions that involve the degradation of a wide range of organic substances. The aim of this study was to identify the spatial and temporal distribution of the redox zones that develop beneath an artificial recharge pond, a system that is characterised by regular hydraulic changes between the saturated and unsaturated condition within every operational cycle. The most significant hydraulic changes result from the formation and removal of a clogging layer at the pond’s bottom. Geochemical analyses of suction cup water as well as oxygen and hydraulic measurements showed that generally nitrate and manganese reducing conditions dominated below the pond as long as water saturated conditions prevailed. Iron and sulphate reduction occurred only in patchily distributed zones directly below the clogging layer and resulted from chemical and physical heterogeneity. When the sediment below the clogging layer became unsaturated, atmospheric oxygen penetrated from the pond fringes into this region, allowing re-oxidation of previously formed sulphide minerals.

The adsorption/ desorption processes in metallurgical wastes of Cr, Ni and Mn in single, binary and ternary solutions were investigated using batch tests and miscible displacement column experiments. For the experiments, mine wastes from Moa (Cuba) composed by oxi(hydroxi) iron minerals and heavy metals deposited in tailing dams were used. The adsorption isotherms of the three metals are strongly non-linear and are well described by the Freundlich isotherm. Desorption isotherms are linear. The adsorption/ desorption process of Cr, Ni and Mn shows hysteresis. The adsorption of metals can be attributed to the favourable electrostatic interaction due to mineralogical composition of the wastes. The shape of breakthrough curves from column experiments provide evidence that transport of Ni, Mn and Cr is non-ideal with asymmetry and great tailing, showing an important retardation relative to a tracer. When using a ternary solution (Cr, Ni, and Mn), retardation of Ni and Mn was reduced in comparison with the single solute experiment, consequence of the Ni and Mn competition for adsorption sites, while Cr did not show any change. In the Ni-Mn binary solution column transport experiment, Mn adsorption was more rapidly attained than Ni; and both metals show different adsorption behaviour, when adsorption of Ni increases, sorption of Mn decreases. One more aspect being investigated was the sorption capacity of the metallurgical waste. According to the mass-balance analysis of the results, the waste shows a very high effective sorption.

The effect of the ionic strength on adsorption of Cu on Ca-montmorillonite (SAz-1) was studied at concentrations ranging from 31 to 516 µM. An adsorption model was employed in the analysis of the data. When the background electrolyte was NaClO4 the ionic interchange was suppressed at 0.5 M, and Cu adsorption was taking place on edge sites, reaching a plateau at about 24 mmol/kg. A further increase in ionic strength did not make any effect on Cu adsorption, suggesting that the heavy metal was being adsorbed by inner sphere complexes on the edge sites. When the electrolyte used was NaCl the amounts of Cu adsorbed were reduced. The model predicted well the adsorption data by considering adsorption of CuCl⁺ species. Adsorption-desorption processes of Cu on Ca-montmorillonite in media of 0.01 and 0.1 M NaCl showed hysteresis. Model calculations predict the desorbed amounts fairly well. According to the model the hysteresis is mainly attributed to the heterogeneity of sites for the adsorption of Cu. The hysteresis arising from the planar sites is largely due to reduced competition for adsorption and enhancement in the magnitude of the surface potential. The presence of the cationic pesticide chlordimeform reduces strongly the sorption of the metal on the planar positions unlike the edge sites. However, Cu sorption increases on the clay treated previously with chlordimeform which was due to the opening of the clay platelets after some molecules of the pesticide are adsorbed, facilitating the subsequent penetration of the metal and its adsorption on planar positions. This cooperative adsorption was due to the fact that the loading of the pesticide on the clay was a very small amount of the CEC.

This paper reports studies from selected recent papers using NMR to investigate interactions between organic pollutants and soil matrices including SOM fractions, clays, whole soils or sediments, three main approaches are presented. First, an indirect approach consists to measure organic carbon normalized sorption coefficients (K_oc values) of organic pollutants with soil matrices and then to establish correlations between these values and structural elements of soil organic matter (carboxylic, carbonyl, aromatic, aliphatic groups⋯). In that case, NMR technique only plays a role in determining soil organic matter structure. Second are reported solid state NMR (¹³C and ¹⁵N CP-MAS, cross polarization magic angle spinning) experiments which are the main tools to investigate covalent bonds between organic pollutants and SOM. Finally are described experiments carried out on hydrated solid matrices using HR-MAS (high resolution magic angle spinning) NMR. This approach, which is the most recent and novel, allows characterizing molecular species at the solid-aqueous interface, these species are differentiated according to their degree of mobility. It should help to investigate the major question of bio-availability of organic pollutants, mainly for bacteria.

Reactive-transport models aim at a comprehensive, quantitative and, ultimately, predictive treatment of biogeochemical transformations and mass transfers in the subsurface. Not only do they provide environmental simulation tools, they can also be used to test new theoretical concepts or hypotheses. A major goal of the geochemistry group in Utrecht is to incorporate complex, microbially-driven reaction networks in reactive transport models, through a close collaboration between modelers and experimentalists. This paper gives an overview of some of the research activities we are carrying out in this area.

This study investigates the consequences of assuming different kinetic approaches for calculation of microbial degradation on plume development in a simple case of natural attenuation on field scale. If the required parameters are properly chosen, all approaches can simulate similar plumes for a particular given time step. The differences of contaminant concentrations in the plumes are small and would not attract attention in a natural aquifer. On long term prognoses the kinetics result in very different plumes: A complex Monod approach considering microbial growth prognoses a further spreading of the plume, compared to a first order rate law, which results in a short and early stationary plume. Other approaches show plumes between these two extremes. On the other hand, the forecasts for plumes assuming Monod kinetics are similar, even if different values for parameterization are chosen. The reason for this insensibility is, that degradation is not limited by microbial kinetics in the simulation, but by dispersive mixing. Simplifying approaches may have few and well determinable parameters, but they are not suited for proper prognoses if they neglect the prerequisite, that contaminant and electron acceptor have to be present for a reaction.

The unsaturated zone including the capillary fringe is a very dynamic and active environment for biogeochemical processes. Modeling of fate and transport of organic pollutants has to account for vapor phase and sewage water transport coupled to the biogeochemical processes occurring. This study presents results from numerical simulations validated with data of a well-controlled field experiment on volatilization of a multi-component organic mixture in the unsaturated zone. Sensitivity analyses show that the overall biodegradation rates depend mainly on properties of the organic pollutants such as Henry’s Law constant, the soil water content, and on the individual degradation rate constants or temperature. Low Henry’s law constants result in relatively high biodegradation rates whereas compounds with high vapor pressure und low water solubility are lost to the atmosphere. The contaminant transfer rates into groundwater are relatively small, but lead locally to concentrations above the legal limit in the capillary fringe region. Contaminant transport by seepage water is just minor compared to diffusive vapor phase fluxes of volatile compounds.

A process-based modelling approach of organic contaminant spreading is presented, which allows to quantify the increase of groundwater pollution risk due to preferential flow and particle-facilitated transport in the vadose zone. Both effects are illustrated by means of a parameter study based on realistic input data. When coupled with biodegradation, preferential flow and particle-facilitated transport lead to an increase in the long-term concentration to be expected at the groundwater table as both processes reduce the bioavailability of the contaminant due to either shorter residence times (preferential flow) or lower accessibility (particle-facilitated transport). Model predictions account for the amount and the velocity of preferential flow as well as the potentially kinetic interaction between organic contaminant and mobile particles.

An integrated approach of field and model investigations was implemented to an aquifer underlying urban infrastructure. The study focuses on the transport of Trichloroethylene (TCE) through the unsaturated-saturated zone. Simulations, subject to isothermal conditions, addressed a threedimensional continuum involving three interacting mobile phases: aqueous, NAPL (Non-Aqueous Phase Liquids) and gaseous. The mathematical model considers water, air and TCE as components in equilibrium partitioning between the three mobile fluids, while for the latter we account also for sorption on the solid matrix. Predictions of migration patterns were due to a continuous spill from a single NAPL source situated at the soil surface.

The specific flux of the gaseous phase proved to be directed to the surface at the vicinity of impervious segments, implying that TCE vapors may surround building foundations. We note a vertical, gravity driven, NAPL vertical displacement of the gaseous phase the velocity of which is also influenced by the gaseous pressure gradient. Fluids densities depend on TCE concentration, we thus note a density driven flow pattern. As TCE concentration decreases away from the NAPL stem, the gradient of the gas density yields its horizontal displacement, and vertical outward shift resulting from the vanishing capillary pressure. The gaseous velocity far away from the stem is associated only with the gas pressure gradient and is diminished at the unsaturated-saturated boundary, as the relative permeability of gaseous phase decreases near the capillary fringe.

A detailed field investigation was performed for studying groundwater recharge processes and solute downward migration mechanisms prevailing in the unsaturated zone overlying a chalk aquifer in Belgium. Various laboratory and field experiments were performed, among others tracer experiments in the unsaturated zone. From the experiments performed in the variably saturated chalk, it appears that the migration and retardation of solutes is strongly influenced by recharge conditions. Under intense injection conditions, solutes migrated at high speed along the partially saturated fissures, reaching the saturated zone in a few hours. At the same time, they were temporarily retarded in the almost immobile water located in the chalk matrix. Under normal recharge conditions, fissures were inactive and solutes migrated slowly through the chalk matrix, taking about one year to reach the water table. The tracer experiments performed in the chalk were modelled using the finite element flow and transport simulator SUFT3D. The mathematical representation of the unsaturated properties of the fissured chalk is based on a new modified van Genuchten — Mualem relationship that takes into account the dual-porosity characteristics of the rock. For modelling the transport of the tracers, an adapted first-order transfer, dual-porosity model is used, based on a dynamic partitioning of water between effective and immobile water porosity according to the water saturation of the chalk. Conclusions are drawn in terms of recharge mechanisms and vulnerability of the chalk to pollutions occurring at the land surface.

The emerging importance of large scale distributed-process modeling has generated a pressing need for detailed information on spatial distribution of input data, especially the parameters that describe the hydrologic behavior of soils. Traditional methods to determine soil hydraulic parameters at large scales are time-consuming and very costly. To carry out a soil hydraulic characterization in a cost-effective way, pedotransfer functions (PTFs) are profitable tools to estimate soil hydraulic parameters from easily measurable or already available soil data. Spatial distribution of soil properties can be explained, to a certain extent, in the light of the variability of landscape attributes. Digital terrain analysis can thus provide a quantitative basis for deriving topographic attributes and relating them to soil variables in order to improve the prediction of key soil hydraulic parameters. Topographical data have been used to improve the prediction of soil water retention by PTFs, whereas statistical indicators of goodness-of-fit helped in evaluating the effectiveness of the proposed procedure.

We model drainage flow by assuming gravitational regime. Central for the present paper is a fairly general analytical solution which is applicable to local (typically laboratory) scale. These results are then specified by adopting the Brooks et al (1964) model for the hydraulic conductivity curve. At field scale, we combine our analytical results with the stream tube model of Dagan et al. (1979) to derive the ensemble average of the effective saturation ‘S’. This is achieved by regarding the parameters of hydraulic properties as random space functions (RSFs) with given joint density distribution. For illustration purposes, we limit to consider the spatial variability of the saturated conductivity K _s , solely.

We demonstrate the potential of geophysical methods, in particular Electrical Resistance Tomography (ERT) in conjuction with tracer experiments, to derive hydraulic variables (e.g. velocity) and solute transport parameters (e.g. dispersivity) in subsurface environments. For this purpose, real tracer experiments were monitored using crosshole time-lapse ERT at reference planes at the Krauthausen test site. Numerical inversion of an ERT dataset obtained from a numerical tracer experiment in a heterogeneous 3D flow domain reveals that the 2D distribution of the bulk electrical conductivity field at selected reference planes can be recovered. In 2000, a first small-scale tracer experiment at the Krauthausen field site combined with ERT was performed to monitor the change in bulk electrical conductivity at a reference plane.

The imaged bulk electrical conductivity changes were transformed to local concentrations using an empirically derived relation with the tracer (bromide) concentration. The spatial variation of the stream tube local dispersion and stream tube velocity in the reference plane was then derived. This information can also be used to derive inversely the statistical properties (variance and correlation length) of the hydraulic conductivity field. In 2002 and 2003, a larger scale ‘positive’ and ‘negative’ tracer experiment was conducted with the tracer electrical conductivity larger and smaller, respectively, than the groundwater electrical conductivity. The negative tracer experiment was conducted to avoid downward movement of the plume such that it passes through the most sensitive part of the ERT-image plane.

To improve the imaging of solute concentrations, a new Magnetic-Electrical-Resitivity-Imaging-Technique (MERIT) is being developed which uses the additional information contained in the magnetic field strength. In a first step, this technique is being made available at column and lysimeter scale. In a further step, MERIT will also be developed for field scale applications.

In September 2001 a tracing experiment has been worked out at the groundwater test field “Wagna” (400 * 300 m in the Mur valley aquifer system) using 75 kg of sodium bromide as tracer and injected during 20 minutes at the groundwater table (pulse injection). Due to the density of the solute the tracer has been distributed over the whole aquifer depth in a very short time.

For estimating the longitudinal dispersivity an analytical model for calibrating Peclet’ Typecurves on measured tracer breakthrough curves has been used. Transversal dispersivity was estimated using 2D-analytical model calibration. The results of parameter estimation on the field scale are compared to the results of a 2D-numerical bromide transport model for the test field based on a transient groundwater flow model.

Within the PHREEQC framework, a dual Monod kinetics formulation has been included, which allows rate dependencies of both substrates, terminal electron acceptors and inhibitors. In this way, PHREEQC will simulate the redox processes under concern both with regard to kinetics and thermodynamics. Furthermore, PHREEQC allows one-dimensional reactive transport to be simulated.

The biogeochemical processes involved in transport and biodegradation of dissolved jet-fuel were simulated for two cases by PHREEQC with this Monod kinetics scheme. The column studies of Knudsen (2003) exploring dissolved jet-fuel transport and biodegradation within initially pristine aquifer sediments dominated by pyrite oxidation and calcite dissolution. Pyrite oxidation will compete with aerobic biodegradation, thereby reducing the efficiency of aerobic bioremediation. The 1D column simulations gave reasonable agreement with measured biodegradation, mineralization and pyrite oxidation rate, and reproduced the overall microbial processes well, but they failed to mimic the observed ferrous iron. A dual porosity approach should be included.

The second case with a jet-fuel contaminated plume under monitoring was simulated with a 1D PHREEQC column from a plume cross section along the flow direction. The biogeochemical reactions themselves were described reasonably well, but dispersional/diffusional transport effects could not be simulated sufficiently with a 1D column of PHREEQC alone. Here truly coupled models of 3D flow and biogeochemical reactions must be applied.

The aim of the present study is the assessment of natural attenuation potential at a TCE-contaminated site in Northern Italy. This site has a long history of contamination (approximately 50 years) mainly due to industrial activities.

Removal of contaminant sources or associated residual free phase pools often suffers from a combination of inefficiency, increased risk of contaminant spreading due to uncontrolled mobilization, and/or high treatment costs. The paper gives a brief overview of results from laboratory and field studies where chemical and biochemical in-situ source control actions are evaluated with respect to their efficiency on changes of source emission. The studies focused on the contaminant group of chlorinated ethenes. Chemical approaches aim to mobilize contaminant phases. By now these studies have been mainly executed in the laboratory and only few pilotscale field studies exist. The results indicate large emission rates at the beginning of the phase displacement but give less information on long-term emission rates. Biostimulation and bioaugmentation approaches revealed increased emission on short time scales and accumulation of cis- Dichloroethylene and Vinylchloride due to incomplete degradation of higher chlorinated solvents in the source zone. On long-term scales emission rates decreased and groundwater plumes were shrinking.