Conceptual Model of Vadose-Zone Transport in Fractured Weathered Shales
The mobility of water and solutes in the vadose zone occurs through a complex continuum of pores that vary in size and shape. Fractured subsurface media that have weathered often depict the extreme case of pore-class heterogeneity because highly conductive voids surround low-permeability, high-porosity matrix blocks. In this chapter, the physical and chemical processes controlling water and solute transport in fractured, weathered shales are discussed. At the Oak Ridge National Laboratory (ORNL), the weathered shales are commonplace and are characterized by a highly interconnected fracture network with densities of 100-200 fractures per meter. The media are conducive to extreme preferential flow that results in physical, hydraulic, and geochemical nonequilibrium conditions between fractures and the surrounding soil matrix. This scenario is of significance with regard to contaminant fate and transport issues at ORNL, where the subsurface burial of toxic metals and radionuclides has occurred. A multiregion flow and transport concept has been adopted to describe the movement of water and solutes through the weathered shale. A variety of multiscale experimental and numerical endeavors have been undertaken to justify the multiregion concept. Experimental manipulations at the laboratory, intermediate, and field scales are designed to quantify the rates and mechanisms of intra- and inter-region mass transfer. These experimental techniques
Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee.
Southern Nevada Science Center, Desert Research Institute, Las Vegas, Nevada.
Center for Computational Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee.
have significantly improved our conceptual understanding of time-dependent solute migration in fractured subsurface media and have also provided the necessary experimental constraints needed for accurate numerical quantification of the physical or geochemical nonequilibrium processes that control solute migration.
Inceptisols, common to the southeastern United States, are weakly developed soils that have weathered from interbedded shale-limestone parent material. At the Oak Ridge National Laboratory (ORNL) in eastern Tennessee, the Department of Energy (DOE) has historically used these soils for the disposal of low-level radioactive waste in shallow-land burial trenches. Physical and chemical barriers were seldom used to impede waste migration since it was thought that the high cation exchange capacity (CEC) of the media (CEC ~20 cmol/kg) would significantly retard radionuclide mobility from the primary trench sources. Although the geochemistry of the media is sufficient for slowing radionuclide migration, the physical structure of the inceptisols coupled with large annual rainfall inputs (~1400 mm/y) have resulted in the formation of a large secondary contaminant source where radionuclides have been disseminated across a vast subsurface environment. The basic problem is that the soils are highly structured and conducive to rapid preferential flow coupled with significant matrix storage. This circumstance causes large physical and geochemical gradients between the various flow regimes of the media and typically results in nonequilibrium conditions during solute transport.
Because transport processes contributing to the formation of secondary contaminant sources are ill-defined, it is difficult to accurately assess the risk associated with the off-site migration of the contaminants and thus the need for remediation. To help resolve this dilemma, investigations were conducted to provide an improved conceptual understanding and predictive capability of solute transport processes in highly structured, heterogeneous subsurface environments that are complicated by fracture flow and matrix diffusion. The investigation involves multiscale experimental and numerical approaches to address coupled hydrological and geochemical processes controlling the fate and transport of contaminants in fractured vadose zone saprolites. Novel tracer techniques and experimental manipulation strategies using laboratory-, intermediate-, and field-scale experiments helped unravel how coupled transport processes affect the nature and extent of secondary contaminant sources. In so doing, a multiregion flow and transport concept is developed and numerically implemented.
SELECT CHARACTERISTICS OF WEATHERED, FRACTURED SHALE
The subsurface materials used in the disposal of low-level radioactive waste at ORNL are acidic inceptisols that have been weathered from interbedded shale-limestone
sequences within the Conasauga Formation. The limestone has weathered to massive clay lenses devoid of carbonate, and the more resistant shale has weathered to an extensively fractured saprolite. Fractures are highly interconnected with densities in the range of 200 fractures per meter (Dreier et al., 1987). Fold-related fractures in the saprolite are observable in the field and consist of (a) fractures along bedding planes, (b) two sets of orthogonal extensional fractures that are perpendicular to bedding planes, and (c) shear fractures. The extensional fractures are either parallel or perpendicular to the strike of the bedding planes and form an orthogonal fracture network with the bedding-plane fractures. Bedding-plane fractures dominate the fracture network of the media and have aperture spacing of <0.05 mm. Cross-cutting extensional fractures are less numerous but have larger apertures ranging from 0.2-0.5 mm (G. R. Moline, pers. comm.). Many of the fracture surfaces contain secondary deposits (e.g., clays, Mn and Fe oxides) suggesting that they are hydrologically active. Fracture orientation and connectivity can give rise to extensive preferential flow within the media (Solomon et al., 1992). Fractures surround low-permeability, high-porosity matrix blocks that have water contents ranging from 30 to 50 percent. Functional relationships between water content (θ), pressure head (h), and hydraulic conductivity (K) suggest that the soil behaves as a three-region media consisting of macropores, mesopores, and micropores (Wilson et al., 1992). Macropores and mesopores are conceptualized as primary and secondary fractures, and micropores as the soil matrix. The three-region conceptualization of the fractured, weathered shales is supported by field measurements of Wilson and Luxmoore (1988) that showed mesopore convergent flow into macropores with subsequent bypass of the soil matrix. The occurrence of preferential flow in these media is common due to large annual rainfall inputs—as much as 50 percent of the infiltrating precipitation results in groundwater and surface water recharge (10 and 40 percent, respectively). The condition of variably saturated preferential flow suggests that hydraulic nonequilibrium (i.e., hydraulic gradient between flow regions) and physical nonequilibrium (i.e., concentration gradient) exist during storm-driven solute transport.
The fractured, weathered shales are slightly acidic (e.g., pH = 4.5 to 6.0) with typical cation exchange capacities of 15-20 cmol/kg. The <2 µm clay fraction is predominately illite with lesser quantities of 2:1 interstratified material and vermiculite. Carbonates are completely weathered from the upper several meters of soil, and most of the fracture pathways and matrix blocks are coated with amorphous Fe-and Mn-oxides. The diverse mineralogy results in a highly reactive solid phase that can significantly alter the geochemical behavior and transport of solutes. In an effort to enhance our conceptual understanding of and predictive capability for solute transport processes in these unsaturated fractured soils, a multiscale experimental and numerical approach was used to quantify the rates and mechanisms of pore class interactions.
LABORATORY-SCALE ASSESSMENT OF TRANSPORT IN FRACTURED WEATHERED SHALES
Undisturbed columns (typically 15 cm diameter × 40 cm length) are used in tracer transport experiments to assess the interaction of hydrology, geochemistry, and microbiology on the fate and transport of nonreactive and reactive solutes. The primary purpose of research at this scale was to quantify transport mechanisms that are operative at the field scale, but difficult to quantify at these larger scales. At the column scale, several techniques were employed to assess non-equilibrium processes that result from the large difference in hydraulic conductivity of fractures versus matrix blocks. The techniques include (1) controlling flow-path dynamics with manipulations of pore-water flux and soil-water tension, (2) isolating diffusion and slow geochemical processes with flow interruption, (3) using multiple tracers with different diffusion coefficients, and (4) using multiple tracers with grossly different sizes.
Controlling Flow-Path Dynamics
Variations in Pore-Water Flux
A relatively simple technique for confirming and quantifying physical non-equilibrium in soil systems involves displacement experiments performed at a variety of experimental fluxes using a single representative tracer. Alteration of the experimental flux or specific discharge through a soil system perturbs the rate of approach toward equilibrium by changing the hydraulic or concentration gradient. In heterogeneous systems that exhibit a large distribution of pore sizes, an increase in the overall pore-water flux should result in greater system non-equilibrium due to a decrease in solute residence time within the porous media. This is usually the case, as solute movement into the matrix is a combination of advective and diffusive processes, and is typically the rate-limiting step as the system approaches equilibrium. This condition can be observed in Figure 3-1, which shows the breakthrough of a nonreactive Br− tracer at several different pore-water fluxes through an undisturbed column of weathered, fractured shale. As is typical of heterogeneous media, tracer displacement was characterized by an initial rapid solute breakthrough followed by extended tailing to longer times. In this system, the fracture network of the weathered shale controlled the advective transport of solutes, which was coupled with diffusion into the surrounding matrix blocks. The largest and smallest flux experiments were conducted over periods of 0.25 days and 94 days, respectively, with the relative amount of tracer mass remaining in the column at the end of each pulse ranging from 22 percent (fast flux) to 38 percent (slow flux). These results indicate that the system became increasingly removed from equilibrium as the pore-water flux was increased. At
the conditions of faster flux, the tracer residence times in the mobile fracture regions were significantly decreased, and thus not as much mass was lost to the matrix. However, mass loss to the matrix as a function of pore-water flux was highly nonlinear, suggesting that the average rate of mass transfer from fractures to the matrix is greater at larger fluxes. The velocity dependence of mass-transfer processes in various porous media and soils has been shown by others (Akratanakul et al., 1983; Nkedi-Kizza et al., 1983: Jensen, 1984; Schulin et al., 1987; Anamosa et al., 1990; Kookana et al., 1993; Reedy et al., 1996).
Controlling flow-path dynamics through variations in pore-water flux is a relatively simple technique for assessing the significance of physical nonequilibrium in soil systems. However, when used by itself the technique is semiquantitative as it is difficult to know how system variables change in response to flux variations (e.g., are the proportions of advective flow paths constant with changes in flux?). When this technique is combined with the other manipulative experimental strategies (e.g. multiple tracers, flow interruption) described below, it can become a powerful means of quantifying physical nonequilibrium processes in structured media (see, for example, Hu and Brusseau, 1995; Reedy et al., 1996).
Variations in Pressure Head
Controlling flow-path dynamics by manipulation of the soil water content with pressure-head variations is an excellent technique to assess nonequilibrium processes (Seyfried and Rao, 1987; Jardine et al., 1993a). The basic concept of the technique is to collect water and solutes from select sets of pore classes in order to determine how each set contributes to the bulk flow and transport processes that are observed for the whole system. In heterogeneous systems, a decrease in pressure head (more negative) will cause larger pores, such as fractures, to drain and become nonconductive during solute transport. Because advective flow processes tend to dominate in large pore regimes, a decrease in pressure head, which will restrict flow and transport to smaller pores, will limit the disparity of solute concentrations among pore groups. By minimizing the concentration gradient in the system, the extent of physical nonequilibrium is decreased. Figure 3-2 conveys this concept by showing the breakthrough curves of a nonreactive Br− tracer at three different pressure heads in an undisturbed column of the weathered, fractured saprolite from the Oak Ridge Reservation (ORR). The increasing asymmetry of the breakthrough curves with increasing saturation (less negative pressure head) is indicative of enhanced preferential flow coupled with mass loss into the matrix. As the soil becomes increasingly unsaturated, breakthrough curve tailing becomes less significant because of a decrease in the participation of larger pores (fractures) involved in the transport process. These findings suggest that mass-transfer limitations (nonequilibrium conditions) become less significant for these unsaturated conditions because fracture flow has been eliminated. An interesting finding of these studies was that the application
of −10- and −15-cm pressure heads resulted in 5- and 40-fold decreases in the mean pore-water flux, respectively, with relatively little change in soil water content relative to saturated conditions (0.55 cm3/cm3 at h = 0 to 0.51 cm3/cm3 at h = −15 cm). This suggests that most of the water flux may be channeled through pores that hold water with tensions <10 cm (primary fractures, macropores) even though their surface area and contribution to the total system porosity is very small (Wilson and Luxmoore, 1988; Wilson et al., 1989). These results differ from the findings of Krupp and Elrick (1968), who investigated Cl− breakthrough in variably saturated glass bead media. In this study, breakthrough asymmetry was more pronounced during unsaturated conditions relative to saturated flow. This observation resulted from the uniformity of the media (i.e., single pore class) and the use of a constant volumetric flow rate that caused a wide pore-water velocity distribution to develop during unsaturated flow (Wilson and Gelhar, 1974).
Reactive solute transport is also dramatically influenced by water content changes and pore regime connectivity. With enhanced preferential flow, important geochemical reactions such as sorption, oxidation/reduction, complexation/dissociation, and precipitation/dissolution become increasingly limited due to a decrease in residence time of pore water with the soil matrix. Jardine et al. (1988,
1993a, 1993b) found that the reactivity of divalent contaminants and chelated radionuclides increased dramatically with a slight decrease in pressure head from 0 to −10 cm. This was the result of a decrease in preferential bypass when fractures were empty and a greater contact time between the high surface area matrix and secondary fracture network.
Another useful technique for isolating diffusion or slow time-dependent geochemical reactions involves flow interruption during a portion of a tracer displacement experiment. The technique involves inhibiting the flow process during an experiment for a designated period of time and allowing a new physical or chemical equilibrium state to be approached. When physical nonequilibrium processes are significant in a soil system, the flow-interruption method will cause an observed concentration perturbation for a conservative tracer when flow is resumed. Interrupting flow during tracer injection will result in a decrease in tracer concentration when flow is resumed, whereas interrupting flow during tracer displacement (washout) will result in an increase in tracer concentration when flow is resumed. The concentration perturbations that are observed after flow interrupts are indicative of solute diffusion between pore regions of heterogeneous media. Conditions of preferential flow create concentration gradients between pore domains (physical nonequilibrium), resulting in diffusive mass transfer between the regions. Therefore, during injection, tracer concentrations within advection-dominated flow paths (i.e., fractures, macropores) are higher than those within the matrix. Upon flow interruption, the relative concentration decrease that is observed indicates that solute diffusion is occurring from larger, more conductive pores into the smaller pores. During tracer displacement or washout, the concentrations within the preferred flow paths are lower than those within the matrix. Thus, solute diffusion is occurring from smaller pores into larger pores, and a concentration increase is observed when flow interruption has been imposed.
The utility of the flow-interrupt method for confirming and quantifying physical nonequilibrium can be observed in Figure 3-3, which shows Br− breakthrough curves at two fluxes in an undisturbed column of fractured, weathered shale from the ORR. The observed concentration perturbations on the ascending and descending limbs of the breakthrough curves are the result of prolonged flow interrupt and the system approaching a new state of physical equilibrium. The concentration perturbations that are induced by flow interruption are significantly more pronounced at larger fluxes (Figure 3-3b). This is because the system is further removed from equilibrium at the larger fluxes, as a greater concentration gradient exists between advection-dominated flow paths and the soil matrix.
The flow-interrupt method has also been used during reactive tracer studies that focus on the determination of rate-limiting geochemical reactions (i.e., geo-
chemical nonequilibrium) such as sorption, precipitation, transformation, and complexation (Murali and Aylmore, 1980; Hutzler et al., 1986; Brusseau et al., 1989; Ma and Selim, 1994; Mayes et al., 2000). Mayes et al. (2000) provided the first application of the flow-interruption technique for quantifying geochemical nonequilibrium of reactive contaminant and chelated radionuclides in fractured, weathered shales. The flow-interruption process was very effective at decoupling rate-limiting redox and dissociation reactions versus time-dependent solid phase desorption reactions.
Multiple Tracers with Different Diffusion Coefficients
A powerful technique for quantifying physical nonequilibrium in structured subsurface media is the simultaneous use of multiple tracers with different diffusion coefficients. In general, when the technique is used to quantify physical nonequilibrium processes in soils and rock, two or more conservative tracers that have different diffusion coefficients are simultaneously displaced through the porous media. Tracers such as Br−, Cl−, fluorobenzoates, 3H2O, and, for watersaturated conditions, dissolved gases such as He, Ne, SF6, and Kr, are frequently suitable for assessing physical nonequilibrium processes in structured media using the multiple-tracer technique (Carter et al., 1959; Raven et al., 1988; Bowman and Gibbens, 1992; Wilson and MacKay, 1993; Gupta et al., 1994; Jaynes, 1994; Linderfelt and Wilson, 1994; Clark et al., 1996; Sanford et al., 1996; Sanford and Solomon, 1998; Jardine et al., 1999). Colloidal tracers that are size-excluded from the matrix porosity of the media are not included in this type of experimental technique. When physical nonequilibrium processes are significant in porous media, tracers with larger molecular diffusion coefficients will be preferentially lost from advective flow paths (i.e., fractures) due to more rapid diffusion into the surrounding solid phase matrix. Likewise, tracers with smaller molecular diffusion coefficients (e.g., larger molecules) will remain in the advective flow paths for longer times due to slower diffusion into the matrix porosity. When advective processes are dominant in a system and matrix diffusion is negligible, multiple tracer breakthrough profiles will not differ considerably (see Brusseau, 1993, for an example).
The utility of multiple tracers for quantifying physical nonequilibrium processes in structured media can be observed in Figure 3-4, where the simultaneous transport of two nonreactive tracers, Br − and pentafluorobenzoic acid (PFBA), was investigated in large undisturbed columns of fractured, weathered shale at two different pore-water fluxes. The molecular diffusion coefficient for PFBA is 40 percent smaller than the diffusion coefficient for Br− (Bowman, 1984). Differences in the breakthrough curves for these solutes can be attributed to differences in the rates of tracer diffusion into the soil matrix. Because the PFBA diffused more slowly into the weathered shale matrix, its breakthrough at the column exit was initially more rapid than Br− but required longer times to approach equilib-
rium (C/C0 = 1) (Figure 3-4 a, Figure 3-4 b). Thus, Br− had a larger mass loss to the matrix at any given time and exhibited a more retarded breakthrough relative to PFBA. However, Br− will reach equilibrium more rapidly than PFBA and the tracer breakthrough curves will eventually cross at longer times. In contrast, the mobility of these two nonreactive tracers would be identical in a column of unstructured media because pore class heterogeneity would be minimal, thus limiting the significance of physical nonequilibrium during transport (see Brusseau, 1993, for an example).
Shropshire (1995) investigated the transport of an inorganic anion (Br−) and a dissolved gas tracer (He) in saturated columns of fractured weathered shale from the ORR. The breakthrough and displacement (washout) of He was significantly more sluggish relative to Br−. This resulted from the fact that the molecular diffusion coefficient for He was about three times greater than that for Br−. Thus, the He tracer was preferentially lost to the matrix relative to Br−. Using a discrete fracture flow model, Shropshire (1995) found that the multiple-tracer technique was an excellent method for approximating fracture aperture and network geometry in the weathered shale media.
Multiple Tracers with Grossly Different Sizes
Another sensitive technique for confirming and quantifying physical nonequilibrium in heterogeneous soil and rock systems is the use of multiple tracers with distinctly different sizes. Specifically, this technique uses both dissolved solutes and colloidal tracers so that flow-path accessibility can be controlled. Viruses, bacteria, fluorescent microspheres, DNA-labeled microspheres, radiolabeled Fe oxide particles, and synthetic polymers have all been used as colloidal tracers in various subsurface media (Barraclough and Nye, 1979; Gerba et al., 1981; Smith et al., 1985; Bales et al., 1989; Harvey et al., 1989, 1993, 1995; Toran and Palumbo, 1991; McKay et al., 1993a, 1993b; Hinsby et al., 1996; Reimus, 1996; Yang et al., 1996). Colloidal particles are typically large enough to be excluded from the matrix porosity of soils and geologic material. If they are not severely retarded by the porous media, colloidal particles serve as excellent tracers for quantifying advective flow velocities in systems conducive to preferential flow. When colloidal tracers are coupled with dissolved solutes that can interact with the matrix porosity, a unique technique emerges for assessing physical nonequilibrium processes in subsurface media.
The utility of using multiple tracers of different sizes can be seen in Figure 3-5, which shows the simultaneous injection of two strains of bacteriophage (PRD1 and MS-2) and two dissolved solutes (Br− and PFBA) into a column of fractured, weathered shale in order to assess physical nonequilibrium processes. The bacteriophage travel times were much more rapid than those of the dissolved solutes, and the bacteriophage strains exhibited significantly less total dispersion in their transport, as evidenced by steeper breakthrough characteristics relative to
PFBA and Br−. The larger bacteriophagies were preferentially transported through a smaller range of flow paths and were minimally affected by diffusion into the matrix porosity. The dissolved tracers, on the other hand, were influenced by diffusive mass-transfer processes between fractures and the matrix (Figure 3-5). Besides providing visual evidence of physical nonequilibrium processes in structured media, the experiments provided advective flow velocities that were used to parameterize numerical models designed to simulate the observed data.
Significant research regarding colloidal tracer transport in the fractured, weathered shales has been conducted by the research group led by Dr. Larry McKay at the University of Tennessee. This team of researchers has investigated the influence of coupled hydrologic and geochemical conditions on the fate and transport of bacteriophage and microspheres in undisturbed columns of the fractured saprolites. Studies by Harton (1996) have shown that the retention of bacteriophage to the solid phase is extremely sensitive to flow rate, where essentially
no retention was observed at rapid pore-water fluxes and a gradient of 1.0, versus near-complete retention at slower fluxes and a gradient of 0.01. Cumbie and McKay (1999) further showed that for any given pore-water flux, the transport characteristics of microspheres in the fractured saprolites were optimized in the size range of 0.5-1.0 µm. Smaller particles down to 0.05 µm tended to adhere to the fracture walls and enter the soil matrix, and larger particles up to 4.25 µm tended to clog advective flow paths due to straining and gravitational settling. These results were corroborated by Haun (1998), who also found that microsphere retention mechanisms were enhanced by increasing aqueous phase ionic strength and the valence state of cotransported cations. Studies by G. R. Moline (pers. comm.) at ORNL have also used fluorescent microspheres to interrogate the fracture network and flow-path heterogeneity of the weathered saprolites. These results were coupled with computed tomography imaging of the undisturbed core specimens, and the fracture network patterns were used to simulate tracer displacement experiments conducted on the undisturbed cores.
INTERMEDIATE-SCALE ASSESSMENT OF TRANSPORT IN FRACTURED WEATHERED, SHALES
A logical progression from laboratory-scale undisturbed columns is the use of intermediate-scale in situ pedons for assessing the interaction of coupled processes on the fate and transport of solutes in the fractured, weathered shales (Figure 3-6). This research scale, unlike the column scale, encompasses more macroscopic structural features common to the field (e.g., dip of bedding planes, more continuous fracture network, convergent flow processes), yet allows for a certain degree of experimental control as the pedon can be hydrologically isolated from the surrounding environment. The pedon is an undisturbed block of soil (2m × 2m × 3m deep) with three excavated sides refilled with compacted soil and a concrete wall with access ports placed in good contact with the front soil face. The pedon was instrumented with a variety of solution samplers designed to monitor water and solutes through various pore regimes as a function of depth. Fritted glass plate lysimeters of varying porosity were held under different tensions to derive solutions from various pore regimes. Coarse fritted glass samplers were held at zero tension and collected free-flowing advective pore water (primary fractures); medium porosity frits were held at 20 cm tension and collected pore water from mesopores or secondary fractures; and fine frits were held at 250 cm tension and collected pore water from the soil matrix. Numerous tracer experiments have been conducted at this facility and others on the ORR using both nonreactive and reactive tracers with various steady-state infiltration rates or transient flow conditions driven by storm events (Jardine et al., 1989, 1990; Wilson, unpublished data). The purpose of the experiments was to determine transport properties and mass transfer rates for the various pore regimes. An example of tracer mobility (Br−) through the soil for two different infiltration
rates can be seen in Figure 3-7. At an infiltration rate of 30 cm/d, Br− is transported exclusively through medium and small pore regimes indicative of secondary fractures and the soil matrix, respectively. Flow through the large pore regimes, indicative of primary fractures, is essentially excluded since the imposed
infiltration velocity is not sufficient to accommodate their conductivity. Larger infiltration rates (e.g., 300 cm/d), representative of larger-scale convergent flow processes, did produce flow through the primary fractures, and thus three distinct breakthrough curves can be observed for tracer movement through primary and secondary fractures, as well as the soil matrix. Large infiltration rates, however, create local-scale perched water tables within the soil, allowing small pore regime samplers to extract a portion of the larger pore water. This is why the ascending limb of the three breakthrough curves at 300 cm/d exhibit diminutive differences. Nevertheless, the distinction of tracer migration through the different pore regions allows for semi-quantitative estimates of pore-flow velocities, dispersion coefficients, and mass transfer rates between the pore classes at a scale one step closer to the realities of the field scale.
When the primary fracture network is nonconductive (e.g., 30 cm/d) solute residence times are sufficient to allow significant interaction of reactive solutes with the secondary fracture walls and the soil matrix. Near-equilibrium conditions were observed for Sr2+ transport through the pedon at an infiltration rate of 30 cm/d. In contrast, when primary fractures are conductive, reactive tracers can bypass the soil matrix and migrate in the same way as a nonreactive tracer (e.g., Jardine et al., 1988, 1993a). This scenario can be observed in undisturbed laboratory columns as well, where it was found that batch adsorption isotherms were appropriate for describing the transport of reactive solutes during unsaturated conditions (primary fractures nonconductive and preferential flow restricted to secondary fractures) but were inappropriate for describing solute transport during saturated conditions (Jardine et al., 1988, 1993a). During unsaturated conditions in the undisturbed columns, solute residence times (~10 d) were sufficient to establish geochemical equilibrium between the secondary fracture network and the soil matrix. These results are consistent with those of Reedy et al. (1996), who used a flow-interruption technique to show that ~4 days was sufficient to establish equilibrium conditions between nonreactive solutes in the fractures and nonreactive solutes in the soil matrix. Both the column and pedon studies have shown that the rate of mass transfer between fractures and the soil matrix is relatively rapid, and thus it can be inferred that the soil matrix remains an integral part of the entire system during storm events and is not necessarily excluded from transport processes when preferential flow occurs in fractures.
FIELD-SCALE ASSESSMENT OF TRANSPORT IN FRACTURED, WEATHERED SHALES
Waste migration issues in fractured, weathered shales are field-scale problems that are complicated by large-scale media heterogeneities that cannot be replicated at the laboratory or intermediate scale. In order to validate our conceptual understanding of vadose-zone transport that was derived from laboratory and pedon-scale observations, field-scale solute fate and transport experiments must
be performed. At ORNL, a field facility has been constructed for assessing stormdriven solute mobility in the vadose zone of the fractured saprolites (Luxmoore and Abner, 1987). The experimental subwatershed drains an area of 0.63 hectares from an elevation of 275-258 m. The landscape is forested with hardwoods, and the soil profile is 0.5-3 m thick depending on elevation. The facility contains a buried line source for tracer release to simulate leakage of trench waste from contaminated areas and is equipped with an elaborate array of water and solute monitoring devices. Lateral subsurface flow is intercepted by a 2.5 m deep by 16 m long trench that has been excavated across the outflow region of the subwatershed (Figure 3-8). Six massive stainless steel pans with 10-cm lips have been pressed back into the soil face to capture subsurface drainage from different portions of the landscape. Subsurface drainage from the pans, as well as overland storm flow and drainage under the pans, is routed into tipping bucket rain gauges situated in two H-flumes that are equipped with ultrasonic sensors for measuring water levels. The tipping buckets and ultrasonic sensors are equipped with computer data acquisition, allowing for real-time monitoring of tracer fluxes during storm events. Besides the ability to capture subsurface drainage, the field facility is equipped with numerous multilevel solution samplers, tensiometers, and piezometers, used to assess perched water table dynamics during storm events. The field facility has been well characterized with respect to spatial variability of surface and subsurface hydraulic conductivities, basic geochemical properties, and mineralogical analyses (Watson and Luxmoore, 1986; Jardine et al., 1988, 1993a, b, 1998; Wilson and Luxmoore, 1988; Wilson et al., 1989, 1992, 1993, 1998; Luxmoore et al., 1990; Kooner et al., 1995; Reedy et al., 1996).
In an effort to address field-scale fate and transport processes during transient flow in the fractured weathered shale media, Wilson et al. (1993) released a Br− tracer from the ridgetop buried-line source during a storm event and monitored its mobility through the subsurface for nearly 8 months. During the release they observed that a small portion of the total Br− mass (~5 percent) migrated very rapidly through the hillslope via lateral storm flow with subsequent export through the weirs that were located 70 m from the line source. The first arrival of Br− at the weirs was 3 h after the initiation of the release, which illustrates the incredibly large hydraulic conductivity of the primary fracture network during hillslope convergent flow. These results are consistent with the measurements of Wilson and Luxmoore (1988) that showed 85 percent of ponded water flux went through primary fractures that constitute only 0.1 percent of the total soil porosity. Subsequent storm events over the 8-month period resulted in the export of ~20 percent of the injected Br− mass (e.g., Figure 3-9a). While the actual tracer release revealed a rapid transport through the fracture network of the soil, the mass transfer into the low-permeability matrix was significant since >50 percent of the applied tracer was found to reside within matrix porosity of the soil primarily at a depth of 1-1.5 m (Figure 3-9b), which is synonymous with where lateral storm flow occurs through the hillslope. Large-scale structural heterogeneities of the subsurface
media controlled solute mobility, since tracer was preferentially transported along bedding planes that dipped in an opposing direction to the hillslope topography. The interconnected nature of bedding plane parallel fractures with cross-cutting extensional fractures results in continuous preferential flow paths that can accommodate large storm-flow inputs. The sheer magnitude of bedding plane parallel fractures versus strike parallel extensional fractures tends to drive solutes along the dip of the bedding planes (G. R. Moline, pers. comm.).
Strong evidence indicating matrix-fracture interactions during transient storm flow can be inferred from tracer breakthrough patterns at the subsurface weirs. Storm events that followed the release of tracer resulted in delayed tracer breakthrough pulses relative to the subsurface flow hydrograph (Figure 3-9a). Stable isotope and solute chemistry analysis revealed that subsurface flow was predominately new water at peak flow and almost exclusively old water during the descending limb of the subsurface hydrograph. These results suggested that the storm-driven export of Br− through the weirs was the result of tracer mass transfer from the soil matrix into the fracture network with subsequent mobility through the hillslope. The field-scale endeavors provided an improved conceptual understanding of how transient hydrodynamics and media structure controlled the migration and storage of contaminants in the subsurface. The field-scale findings were consistent with the multiregion conceptual framework established with laboratory- and intermediate-scale observations of flow and transport through these heterogeneous systems.
MULTIREGION CONCEPTUAL FRAMEWORK OF VADOSE ZONE TRANSPORT IN FRACTURED MEDIA
The multiscale experimental endeavors have provided a conceptual framework for a multiregion flow and transport mechanism that controls solute mobility in the fractured, weathered shales. Experimentally we are able to distinguish three pore-size classes (primary fractures, secondary fractures, soil matrix); thus a representative elemental volume (REV) at any point in the soil consists of three regions, each with its own flow and transport parameters (Figure 3-10). Intraregion mass transfer is described by flow from a physical point to a neighboring one and interregion mass transfer between the various pore regimes.
This transfer is controlled by both advective and diffusive processes, where hydraulic gradients, caused by differences in fluid velocity in different-sized pores, drive advective mass transfer and concentration gradients drive diffusive processes. This concept is illustrated by the following example.
Clean rainwater infiltrating into secondary fractures and converging into primary fractures would bypass an initially contaminant-rich soil matrix. Thus, physical nonequilibrium—that is, concentration gradients—would exist between regions and cause diffusive transfer from the matrix to the mobile regions. However, pressure gradients would also exist initially from the secondary and primary
fracture regions toward the soil matrix. Thus, diffusion of contaminants from the matrix would initially be counter to the interregion advective transfer (Figure 3-10b). The resulting export of contaminants out of the profile would be slow, followed by a gradual increase as advective mass transfer between regions subsides. In this example, primary fractures would serve to provide relatively clean water to aquifers and streams. In the absence of primary fractures, the initiation of outflow may require a longer period; however, the eventual export of contaminants would be much greater due to the increased displacement of this contaminant-rich matrix water. From this example, advection as well as diffusion of solutes between pore regions are relevant processes in solute transport under field conditions.
Another process inherent in multiregion flow and transport in structured systems is the time-dependent nature of both the advective and diffusive mass transfer rates between the various pore domains. Time-dependent mass transfer rates take into account changes in concentration gradients as solute mass is transferred between pore regimes. Time-dependent rate coefficients also indirectly account for variabilities in matrix block sizes.
The concept of multiregion flow and transport in structured media has been numerically implemented by Gwo et al. (1991), with the mathematical formulation of the code described in Gwo et al. (1994, 1995a, b, 1996) and the simulation of experimental data illustrated in Gwo et al. (1998, 1999). A two-region formulation of the model has been used to describe multiple tracer transport through undisturbed soil columns of fractured, weathered shale. The structure of the porous media was embedded into the two-region model in which the interregion mass flux was characterized by mesoscale mass-transfer coefficients. Predictions were significantly more accurate than those obtained using a simple single-fracture conceptual model that represents a least-information scenario. Modeling results suggested that mesoscale spreading of tracer in structured porous media may be largely attributed to interregion mass transfer. The multiregion model has also been used to numerically simulate storm-driven solute transport experiments conducted at the watershed scale within the fractured saprolites. Comparison of numerical results with the field data indicated that multiregion, preferential flow occurs under partially saturated conditions that can be confirmed theoretically, and that mass transfer between pore regions is an important process influencing contaminant movement in the subsurface.
In a manner similar to that of Solomon et al. (1992), the basic concepts of groundwater flow and transport in the vadose zone of fractured, weathered shales on the ORR are summarized below.
Flow and transport through the fractured saprolites is a multiporosity scenario where an interconnected network of highly conductive primary and secondary fractures surrounds low-permeability, high-porosity matrix blocks.
Flow and transport are driven by storm events with significant water flow occurring through a 1- to 2-m-thick storm-flow zone that develops as a result of perched water tables. Rapidly infiltrating storm water in the upper soil horizons becomes perched on lower conductivity, high bulk density saprolites, promoting subsurface lateral flow through hillslopes. The flow is transient and preferentially follows the dip of bedding planes, with most storm flow being discharged into local streams.
The majority of storm flow occurs in primary and secondary fractures, and most water storage occurs in the soil matrix. Fractures can accommodate large rainfall inputs (85 percent of convergent storm flow) but only constitute 0.1 percent of the total soil porosity. The soil matrix, on the other hand, has low permeabilities but porosities as high as 50 percent.
The fractured, weathered shales are conducive to extreme preferential flow, which results in hydraulic, physical, and geochemical nonequilibrium between fractures and the surrounding soil matrix. Differences in fluid velocities in different-sized pores create hydraulic gradients that drive time-dependent interregion advective mass transfer. Solute concentration differences between the various sized pores creates concentration gradients that drive time-dependent inter-region diffusive mass transfer. Storm-enhanced preferential flow in these weathered shales disrupts geochemical equilibrium between the solid, liquid, and gas phases.
The rapid transport of small amounts of mass occurs through large pores (fractures), but because of the prevalence of nonequilibrium conditions, the majority of mass resides in small pores (soil matrix), which greatly retards bulk mass migration rates.
This research was supported by the Environmental Technology Partnership program of the Office of Biological and Environmental Research, U.S. Department of Energy. The authors thank Mr. Paul Bayer, contract officer for DOE's ETP program, for financially supporting this research. Oak Ridge National Laboratory is managed by University of Tennessee-Battelle, LLC, for the U.S. Department of Energy, under contract DE-AC05-00OR22725. Publication No. 5057, Environmental Sciences Division, ORNL.
Akratanakul, S., L. Boersma, and G. O. Klock, 1983. Sorption processes in soils as influenced by pore water velocity. 2. Experimental results. Soil Science 135: 331-341.
Anamosa, P. R., P. Nkedi-Kizza, W. G. Blue, and J. B. Sartain, 1990. Water movement through an aggregated, gravelly oxisol from Cameroon . Geoderma 46: 263-281.
Bales, R. C., C. P. Gerba, G. H. Grondin, and S. L. Jensen, 1989. Bacteriophage transport in sandy soil and fractured tuff. Appl. Environ. Microbiol. 55: 2061-2067.
Barraclough, D., and P. H. Nye, 1979. The effect of molecular size on diffusion characteristics in soil . Journal of Soil Science 30: 29-42.
Bowman, R., 1984. Evaluation of some new tracers for soil water studies. Soil Sci. Soc. Am. J. 48: 987-993.
Bowman, R. S., and J. F. Gibbens, 1992. Difluorobenzoates as nonreactive tracers in soil and groundwater. Ground Water 30: 8-14.
Brusseau, M. L., 1993. The influence of solute size, pore water velocity, and intraparticle porosity on solute dispersion and transport in soil. Water Resources Research 29: 1071-1080.
Brusseau, M. L., P. S. C. Rao, R. E. Jessup, and J. M. Davidson, 1989. Flow interruption: A method for investigating sorption nonequilibrium . Journal of Contaminant Hydrology 4: 223-240.
Carter, R. C., W. J. Kaufman, G. T. Orlob, and D. K. Todd, 1959. Helium as a ground-water tracer. Journal of Geophysical Research 64: 2433-2439.
Clark, J. F., P. Schlosser, M. Stute, and H. J. Simpson, 1996. SF6-3He tracer release experiment: A new method of determining longitudinal dispersion coefficients in large rivers. Environmental Science and Technology 30: 1527-1532.
Cumbie, D. H., and L. D. McKay, 1999. Influence of diameter on particle transport in a fractured shale saprolite. Journal of Contaminant Hydrology 37(1-2): 139-157.
Dreier, R. B., D. K. Solomon, and C. M. Beaudoin, 1987. Fracture characterization in the unsaturated zone of a shallow land burial facility. In: Flow and Transport Through Unsaturated Rock. D. D. Evans and T. J. Nicholson, eds. Geophysical Monograph 42, Washington, D.C.: American Geophysical Union , pp. 51-59.
Gerba, C. P., S. M. Goyal, I. Cech, and G. F. Bogdan, 1981. Quantitative assessment of the adsorptive behavior of viruses to soils. Environmental Science and Technology 15: 940-944.
Gupta, S. K., L. S. Lau, and P. S. Moravcik, 1994. Ground-water tracing with injected helium. Ground Water 32: 96-102.
Gwo, J. P., G. T. Yeh, and G. V. Wilson, 1991. Proceedings of the International Conference on Transport and Mass Exchange Processes in Sand and Gravel Aquifers. 2. Field and Modeling Studies. Ottawa, Canada, pp. 578-589.
Gwo, J. P., P. M. Jardine, G. T. Yeh, and G. V. Wilson, 1994. MURF user's guide: A finite element model of multiple-pore-region flow through variably saturated subsurface media. Oak Ridge National Laboratory , ORNL/GWPO-011.
Gwo, J. P., P. M. Jardine, G. V. Wilson, and G. T. Yeh, 1995a. A multiple-pore-region concept to modeling mass transfer in subsurface media. Journal of Hydrology 164: 217-237.
Gwo, J. P., P. M. Jardine, G. T. Yeh, and G. V. Wilson, 1995b. MURT user's guide: A finite element model of multiple-pore-region transport through variably saturated subsurface media. Oak Ridge National Laboratory , ORNL/GWPO-015.
Gwo, J. P., P. M. Jardine, G. V. Wilson, and G. T. Yeh, 1996. Using a multiregion model to study the effects of advective and diffusive mass transfer on local physical nonequilibrium and solute mobility in a structured soil. Water Resources Research 32: 561-570.
Gwo, J. P., R. O'Brien, and P. M. Jardine, 1998. Mass transfer in structured porous media: Embedding mesoscale structure and microscale hydrodynamics in a two-region model. Journal of Hydrology 208: 204-222.
Gwo, J. P., G. V. Wilson, P. M. Jardine, and E. F. D'Azevedo, 1999. Modeling subsurface contaminant reactions and transport at the watershed scale. In: Assessment of Non-Point Source Pollution in the Vadose Zone, D. L. Corwin, K. Loague, and T. R. Ellsworth, eds. Geophysical Monograph Series 108: 31-43.
Harton, A. D., 1996. Influence of Flow Rate on Transport of Bacteriophage in a Column of Highly Weathered and Fractured Shale. M.S. thesis, University of Tennessee.
Harvey, R. W., L. H. George, R. L. Smith, and D. L. LeBlanc, 1989. Transport of microspheres and indigenous bacteria through a sandy aquifer: Results of natural- and forced-gradient tracer experiments . Sci. Technol. 23: 51-56.
Harvey, R. W., N. E. Kinner, D. MacDonald, D. W. Metge, and A. Bunn, 1993. Role of physical heterogeneity in the interpretation of small-scale laboratory and field observations of bacteria, microbial-sized microsphere, and bromide transport through aquifer sediments. Water Resources Research 29: 2713-2721.
Harvey, R. W., N. E. Kinner, D. MacDonald, D. W. Metge, and A. Bunn, 1995. Transport behavior of groundwater protozoa and protozoan-sized microspheres in sandy aquifer sediments. Applied and Environmental Microbiology, January: 209-217.
Haun, D. D., 1998. Influence of Ionic Strength and Cation Valence on Transport of Colloid-Sized Microspheres in Fractured Shale Saprolite. M.S. thesis, University of Tennessee.
Hinsby, K., L. D. McKay, P. Jørgensen, M. Lenczewski, and C. P. Gerba, 1996. Fracture aperture measurements and migration of solutes, viruses and immiscible creosote in a column of clay till. Ground Water 34: 1065-1075.
Hu, Q., and M. L. Brusseau, 1995. Effect of solute size on transport in structured porous media. Water Resources Research 31: 1637-1646.
Hutzler, N. J., J. C. Crittenden, J. S. Gierke, and A. S. Johnson, 1986. Transport of organic compounds with saturated groundwater flow: Experimental results. Water Resources Research 22: 285-295.
Jardine, P. M., G. V. Wilson, and R. J. Luxmoore, 1988. Modeling the transport of inorganic ions through undisturbed soil columns from two contrasting watersheds. Soil Sci. Soc. Am. J. 52: 1252-1259.
Jardine, P. M., G. V. Wilson, R. J. Luxmoore, and J. F. McCarthy, 1989. Transport of inorganic and natural organic tracers through an isolated pedon in a forested watershed. Soil Sci. Soc. Am. J. 53: 317-323.
Jardine, P. M., G. V. Wilson, and R. J. Luxmoore, 1990. Unsaturated solute transport through a forest soil during rain storm events. Geoderma 46: 103-118.
Jardine, P. M., G. K. Jacobs, and G. V. Wilson, 1993a. Unsaturated transport processes in undisturbed heterogeneous porous media. I. Inorganic contaminants. Soil Sci. Soc. Am. J. 57: 945-953.
Jardine, P. M., G. K. Jacobs, and J. D. O'Dell, 1993b. Unsaturated transport processes in undisturbed heterogeneous porous media. II. Co-contaminants. Soil Sci. Soc. Am. J. 57: 954-962.
Jardine, P. M., R. O'Brien, G. V. Wilson, and J. P. Gwo, 1998. Experimental techniques for confirming and quantifying physical nonequilibrium processes in soils. In: Physical Nonequilibrium in Soils: Modeling and Applications. H. M. Selim and L. Ma, eds. Chelsea, Michigan: Ann Arbor Press, Inc. , pp. 243-271.
Jardine, P. M., W. E. Sanford, J. P. Gwo, O. C. Reedy, D. S. Hicks, R. J. Riggs, and W. B. Bailey, 1999. Quantifying diffusive mass transfer in fractured shale bedrock. Water Resources Research 35: 2015-2030.
Jaynes, D. B., 1994. Evaluation of fluorobenzoate tracers in surface soils. Ground Water 32: 532-538.
Jensen, J. R., 1984. Potassium dynamics in soil during steady flow. Soil Science 138: 285-293.
Kookana, R. S., R. D. Schuller, and L. A. G. Aylmore, 1993. Simulation of simazine transport through soil columns using time-dependent sorption data measured under flow conditions. Journal of Contaminant Hydrology 14: 93-115.
Kooner, Z. S., P. M. Jardine, S. Feldman, 1995. Competitive surface complexation reactions of SO42− and natural organic carbon on soil. Journal of Environmental Quality 24: 656-662.
Krupp, H. K., and D. E. Elrick, 1968. Miscible displacement in an unsaturated glass bead medium. Water Resourc. Res. 4: 809-815.
Linderfelt, W. R., and J. L. Wilson, 1994. Field study of capture zones in a shallow sand aquifer. In: Transport and Reactive Processes in Aquifers. Dracos and Stauffer, eds. Rotterdam: Balkema.
Luxmoore, R. J., and C. H. Abner, 1987. Field facilities for subsurface transport research. DOE/ER0329. Washington, D.C.: U.S. Department of Energy. 32 pp.
Luxmoore, R. J., P. M. Jardine, G. V. Wilson, J. R. Jones, and L. W. Zelazny, 1990. Physical and chemical controls of preferred path flow through a forested hillslope. Geoderma 46: 139-154.
Ma, L., and H. M. Selim, 1994. Predicting the transport of atrazine in soils: Second-order and multireaction approaches. Water Resources Research 30: 3489-3498.
Mayes, M. A., P. M. Jardine, I. L. Larsen, S. C. Brooks, and S. E. Fendorf, 2000. Multispecies transport of metal-EDTA complexes and chromate through undisturbed columns of weathered fractured saprolite. Journal of Contaminant Hydrology 45: 243-265.
McKay, L. D., J. A. Cherry, R. C. Bales, M. T. Yahya, and C. P. Gerba, 1993a. A field example of bacteriophage as tracers of fracture flow. Environmental Science and Technology 27: 1075-1079.
McKay, L. D., R. W. Gillham, and J. A. Cherry, 1993b. Field experiments in a fractured clay till. 2. Solute and colloid transport. Water Resources Research 29: 3879-3890.
Murali, V., and L. A. G. Aylmore, 1980. No-flow equilibration and adsorption dynamics during ionic transport in soils. Nature 283: 467-469.
Nkedi-Kizza, P., J. W. Biggar, M. Th. van Genuchten, P. J. Wierenga, H. M. Selim, J. D. Davidson, and D. R. Nielsen, 1983. Modeling tritium and chloride 36 transport through an aggregated oxisol. Water Resources Research 19: 691-700.
Raven, K. G., K. S. Novakowski, and P. A. Lapcevic, 1988. Interpretation of field tracer tests of a single fracture using a transient solute storage model. Water Resources Research 24: 2019-2032.
Reedy, O. C., P. M. Jardine, G. V. Wilson, and H. M. Selim, 1996. Quantifying the diffusive mass transfer of nonreactive solutes in columns of fractured saprolite using flow interruption. Soil Sci. Soc. Am. J. 60: 1376-1384.
Reimus, P. W., 1996. The Use of Synthetic Colloids in Tracer Transport Experiments in Saturated Rock Fractures. Ph.D. dissertation, University of New Mexico, LA-13004-T.
Sanford, W. E., R. G. Shropshire, and D. K. Solomon, 1996. Dissolved gas tracers in groundwater: Simplified injection, sampling, and analysis. Water Resources Research 32: 1635-1642.
Sanford, W. E., and D. K. Solomon, 1998. Site characterization and containment assessment with dissolved gases . Journal of Environmental Engineering 124: 572-574.
Schulin, R., P. J. Wierenga, H. Flühler, and J. Leuenberger, 1987. Solute transport through a stony soil. Soil Sci. Soc. Am. J. 51: 36-42.
Seyfried, M. S., and P. S. C. Rao, 1987. Solute transport in undisturbed columns of an aggregated tropical soil: Preferential flow effects. Soil Sci. Am. J. 51: 1434-1444.
Shropshire, R. G., 1995. Dual-Tracers: A Tool for Studying Matrix Diffusion and Fracture Parameters . M.S. thesis, University of Waterloo, Ontario, Canada.
Smith, M. S., G. W. Thomas, R. E. White, and D. Ritonga, 1985. Transport of Escherichia coli through intact and disturbed soil columns . Journal of Environmental Quality 14: 87-91.
Solomon, D. K., G. K. Moore, L. E. Toran, R. B. Dreier, and W. M. McMaster, 1992. A hydrologic framework for the Oak Ridge Reservation. ORNL/TM-12026.
Toran, L., and A. V. Palumbo, 1991. Colloid transport through fractured and unfractured laboratory sand columns. Journal of Contaminant Hydrology 9: 289-303.
Watson, K. W., and R. J. Luxmoore, 1986. Estimating macroporosity in a forest watershed by use of a tension infiltrometer. Soil Sci. Soc. Am. J 50: 578-582.
Wilson, G. V., and R. J. Luxmoore, 1988. Infiltration, macroporosity, and mesoporosity distributions on two forested watersheds. Soil Sci. Soc. Am. J. 52: 329-335.
Wilson, G. V., J. M. Alfonsi, and P. M. Jardine, 1989. Spatial variability of saturated hydraulic conductivity of the subsoil of two forested watersheds. Soil Sci. Soc. Am. J. 53: 679-685.
Wilson, G. V., P. M. Jardine, and J. P. Gwo, 1992. Modeling the hydraulic properties of a multiregion soil. Soil Sci. Soc. Am. J. 56: 1731-1737.
Wilson, G. V., P. M. Jardine, J. D. O'Dell, and M. Collineau, 1993. Field-scale transport from a buried line source in variable saturated soil. Journal of Hydrology 145: 83-109.
Wilson, G. V., J. P. Gwo, P. M. Jardine, and R. J. Luxmoore, 1998. Hydraulic and physical non-equilibrium effects on multi-region flow and transport. In: Physical Nonequilibrium in Soils: Modeling and Application. H. M. Selim and L. Ma, eds. Chelsea, Michigan: Ann Arbor Press, Inc., pp. 37-61.
Wilson, J. L., and L. W. Gelhar, 1974. Dispersive mixing in a partially saturated porous medium. Water Resources and Hydrodynamics, Department of Civil Engineering, Massachusetts Institute of Technology , Boston, Massachusetts, Report No. 191.
Wilson, R. D., and D. M. MacKay, 1993. The use of sulfur hexafluoride as a conservative tracer in saturated sandy media. Ground Water 31: 719-724.
Yang, Z., R. S. Burlage, W. E. Sanford, and G. R. Moline, 1996. DNA-labeled silica microspheres for groundwater tracing and colloid transport studies. In: Proceedings, 212th National Meeting, American Chemical Society, Orlando, Fla., Aug. 25-30.