Recent Advances in Groundwater Hydrology Posters
Presiding: J J Butler, Kansas Geological Survey, University of Kansas; T Scheibe, Pacific Northwest National Laboratory
H51A-01 0830h
Sandbox Experiments of Anomalous Dispersion
We performed a series of tracer experiments in a sandbox. We changed the size of grain (small, medium and large) for each sand layer. We monitored the concentration of sodium chloride at 15 points to obtain the breakthrough curves. We found that the breakthrough behavior of the three cases is different from the results of the clasical advection-dispersion equation. We discuss the tailing phenomena of the experimental result together with a stochastic model of the flow.
H51A-02 0830h
Identification of Physical and Chemical Mass Transfer Processes by a Tracer Flush Experiment
A small-scale field tracer test with a long pulse injection (48 hours) was performed in a highly heterogeneous aquifer. Diffusive/dispersive mass transfer between a highly conductive gravel layer and adjacent saprolitic materials was hypothesized as a significant solute transport process. Some solutes of interest (e.g., uranium) are also strongly impacted by sorption to aquifer solids. In addition to observations of injected tracer breakthrough, we also measured concentration histories of several background groundwater constituents at monitoring wells, including sorbing (uranium) and non-sorbing (chloride, nitrate) solutes. As expected, the concentrations of these constituents decreased as the tracer pulse flushed the accessible aquifer pore space. However, the form of the inverse relationship between tracer and ambient solute concentrations was indicative of chemical and physical non-equilibrium between groundwater in advection-dominated zones (gravel) and that in relatively immobile pore space (saprolites). The observations were compared to field-scale transport simulations with and without mass transfer mechanisms included. These comparisons demonstrate that the response of background solutes to a tracer flush can be used to quantify field-scale mass transfer rates. This information is being used to design and evaluate experiments aimed at development of a distributed biologically-active microbarrier between advection- and diffusion-dominated pore regions.
H51A-03 0830h
An ELLAM approximation for advective-dispersive transport with combined, nonlinear equilibrium and nonequilibrium sorption
We consider an Eulerian-Lagrangian localized adjoint method (ELLAM) applied to nonlinear model equations governing solute transport and sorption in porous media. Solute transport in the aqueous phase is modeled by standard advection and hydrodynamic dispersion, while two types of solid phase are distinguished --- a fraction which achieves equilibrium with the aqueous phase quickly, and another which does not. The rapidly sorbing fraction is modeled using a local equilibrium assumption, while a first-order rate expression is used for the slowly sorbing fraction. In both cases, the sorption isotherms are assumed to be nonlinear. This can be a difficult problem to model numerically for several reasons. Depending on the choice of isotherms, solutions may exhibit self-sharpening fronts and finite speed of propagation. The presence of both equilibrium and non-equilibrium sorption can be challenging for Eulerian-Lagrangian methods, since information may propagate along different characteristic directions in the space-time domain. Here, we present an implementation of a finite element ELLAM discretization in both fully coupled and operator-split frameworks for the reactive transport model. We then evaluate our method for several test problems spanning a range of auxiliary and physical conditions and compare its performance to more standard approaches.
H51A-04 0830h
Modeling the Effects of Mass Transfer on Contaminant Biodegradation in Porous Media
The effect of mass transfer on the rate of contaminant biodegradation in porous media has been evaluated using a mathematical model that accounts for the mechanisms of mass transport in the bulk liquid by advection and dispersion, mass transfer from the bulk liquid to the biofilm (BF) by diffusion, molecular diffusion within the BF, and biological reaction within the BF. The resulting set of equations has been solved using a finite difference method with operator splitting and an iterative solution procedure. The mathematical model was applied to a hypothetical column of homogeneous grains. The main result is that mass transfer may be a limiting factor in the rate of biodegradation. Mass transfer is strongly related to characteristics of the porous medium, namely porosity, hydraulic conductivity, water flow velocity, grain size distribution, contaminant characteristics, and biofilm characteristics. Therefore, the results presented here represent a first step to improve our understanding of biodegradation in biological reactors or during groundwater bioremediation, and, consequently, to enable the design of more effective treatment technologies.
H51A-05 0830h
Estimation of groundwater flow distribution in structured media from non-reactive tracer results under unsaturated condition
Subsurface contaminant transport is a complex process especially within structured media, i.e. saprolite, in which fracture flow dominates matrix flow. The media is made up of components of different pore size ranges and hydraulic conductivity rendering an uneven distribution of water flow and contaminant concentrations. An ability to estimate the distribution of hydraulic conductivity within the media is beneficial in several aspects including providing a better understanding of the mechanisms of contaminant transport and a proper allocation of remedial effort to critically contaminated fractions. In this study pore water velocity and diffusion coefficients of relevant fractions of three intact saprolite samples are estimated from breakthrough curves of non-reactive tracer using a one dimensional transport model. Mariot devices are used to deliver solution at specific tensions and a unit gradient to a fraction of samples that remained saturated within an otherwise partially drained sample. Hydraulic conductivity is then calculated from the model-fitted pore-water velocity and estimated porosity. At a tension of around -10 cm, the hydraulic conductivity of the saturated fraction contributing to flow ranges from 0.01-0.07 cm/hr compared to 0.1-0.5 cm/hr of whole samples under fully saturated condition assuming 40% porosity. The conductivity contrast allows the estimation of 1-6% of saturated flow through the fine pore fraction with tension of -10 cm and higher. The approach is simple yet is a helpful tool in understanding complex contaminant transport in subsurface media.
H51A-06 0830h
Immobilization of uranium (VI) in structured saprolite with microbial U(VI) reduction
Lab-scale biostimulation experiments were conducted to investigate microbial reduction of uranium (VI) in intact saprolite columns. Three intact saprolite columns derived from fractured, interbedded shale, limestone, sandstone sequences were collected and transported to the laboratory. The columns were first flushed with a solution of 10ΜM U(VI) in 50mM NaCl. Once the effluent U/U0 reached approximately 0.5, each column was treated differently. The injection of U(VI) was continued for Column 1 to serve as a control. U(VI), as well as ethanol and nutrients were supplied to Column 2 to stimulate indigenous metal reducing activity. In addition to electron donor and nutrients, the soluble electron shuttling compound anthraquinone-2, 6-disulfonate (AQDS) was supplied to Column 3. AQDS has the ability to diffuse into micropore regions that remain inaccessible to bacterial cells potentially increasing the region of influence of bacterial metal-reducing activity. Successful immobilization of U(VI) will be indicated by incomplete mass recovery in the column effluent relative to the control column (no ethanol addition). The experiment provides lab-scale evaluation of hypothesis that U (VI) in low-permeability porous regions (micropores) of saprolite can be immobilized and isolated from mobile groundwater by stimulating localized microbial U (VI) reduction in hydrologically accessible fractured zones (meso and macropores).
H51A-07 0830h
Flow Dimension Analysis of Pumping Tests in Indicator Fields
The generalized radial flow approach is rarely used in hydraulic test analysis. The flow dimension of a hydraulic test can be described as the change in cross-sectional area of flow with respect to radial distance from the borehole of a hydraulic test. Using this relationship, we have developed a simple algorithm that generates geometries, or conduits, that correspond to arbitrary flow dimensions and have validated this approach through numerical pumping test simulations. Flow dimension can be estimated from the second derivative of pumping-test drawdown versus log time. To investigate geometry effects on flow dimension, we look at transmissivity conduits and indicator fields of varying complexity. We believe constant non-radial flow dimensions, commonly observed in field data, may result from variations in the transmissivity of the conduit and bounding media. Analysis has shown that flow dimension varies with indicator field complexity and/or transmissivity contrast. We investigate the hypothesis that flow dimensions in complex indicator fields may be affected not only by transmissivity contrasts, but also by the interconnectivity of the field.
H51A-08 0830h
Application of a new Model for Ground-Water age Distributions
Modeled ground-water age distributions can aid in understanding ground-water flow systems because they provide a continuous indication of contributions from different source waters that span a range of ages. The ground-water age equation of Ginn (Ginn, T.R., 1999, On the distribution of multi-component mixtures over generalized exposure time in groundwater flow and reactive transport: Foundations; formulations for groundwater age, geochemical heterogeneity, and biodegradation, Water Resources Research, 35(5):1395-1408) was solved numerically to obtain ground-water age distributions for a part of a real-world ground-water flow system. The results of the ground-water age model were compared with results from a particle-tracking model that accounts for dispersion. The means of the ground-water age distributions also were compared with isotopic (tritium and carbon-14) ages of ground-water at selected locations. The ground-water age model consists of two physical dimensions (axial and vertical) and the exposure-time dimension. The two-dimensional velocity field for the physical dimensions in the ground-water age model was extracted from a three-dimensional ground-water flow model of the Rialto-Colton Basin, California. The velocity in the exposure-time dimension is unity. The numerical approximation to the initial conditions is zero everywhere in the ground-water age model domain. Source concentrations are located in the same cells where recharge occurs in the flow model. Dispersive mixing occurs in the physical dimensions in the ground-water age model whereas transport in the exposure-time dimension is through advection only. The resulting simulated ground-water age distributions showed the expected trend of a greater distribution of mass in older exposure-time cells with distance from the recharge cells and with depth in the ground-water age model domain. As a consequence of this trend, the calculated mean ages increased with distance downgradient from the recharge sources and with depth. Comparison of the ground-water age distributions and mean ages with the results of the particle-tracking model showed that the ground-water age distributions were similar to the frequency of particles of different ages at selected locations but the mean ages were younger for the ground-water age model. The computed mean ages from the ground-water age model compared favorably to the isotopic ages near the recharge sources; however, the difference between the isotopic ages and the computed mean ages increased with distance from the recharge sources.
H51A-09 0830h
Empirical analysis of electromagnetic profiles for groundwater prospecting in rural areas of Ibadan, southwestern Nigeria
The Slingram electromagnetic (EM) survey using a coil separation of 60 and 100 meters was carried out in 10 villages in Akinyele area of Ibadan, southwestern Nigeria to aid in the development of groundwater. Five main rock types including an undifferentiated gneiss complex (Su), biotite-garnet schist/gneiss (Bs), quartzite and quartz schist (Q), migmatised undifferentiated biotite/hornblende gneiss (M) and pegmatite/quartz vein (P) underlie the study area. A total of 31 EM profiles was made to accurately locate prospective borehole sites in the field. Four main groups with different behavioural pattern were categorized from the EM profiles. Group 1 is characterized by high density of positive (HDP) or high density of negative (HDN) real and imaginary curves, Group 2 by parallel real and imaginary curves intersecting with negligible amplitude (PNA), Group 3 by frequent intersection of high density of negative minima (FHN) real and imaginary curves, and Group 4 by separate and approximately parallel (SAP) real and imaginary curves. Qualitative pictures of the overburden thickness and the extent of fracturing have been proposed from these behavioural patterns. A comparison of the borehole yield with the overburden thickness and the level of fracturing show that borehole yield depends more on the fracture density than on the overburden thickness. Asymmetry of the anomaly was also found useful in the determination of the inclination of the conductor/fracture.