- Soil Science Society of America
Subsurface contamination plagues many USDOE sites and threatens groundwater supplies. This survey discusses research sponsored by the DOE Environmental Management Science Program (EMSP) for geophysical characterization of the vadose zone at the Idaho National Engineering and Environmental Laboratory (INEEL) and other contaminated sites. Various types of geophysical imaging techniques are used to characterize the shallow subsurface, including electromagnetic (EM), ground-penetrating radar (GPR), electrical, seismic, and nuclear magnetic resonance (NMR). Three common themes appear in the research surveyed in this article: (i) the development of high-resolution imaging capabilities to capture important details of the heterogeneous nature of subsurface properties and processes, (ii) the coupling of nonintrusive survey geophysical measurements (e.g., electrical surveys) with detailed quantitative precise point-sensor measurements (e.g., lysimeters and vapor-port systems) or borehole (e.g., NMR, neutron-based moisture, and geochemical tools) measurements to extend high-precision knowledge away from the borehole, and finally (iii) the application of multiple geophysical methods to constrain the uncertainty in determining critical subsurface physical properties. Laboratory, field, theoretical, and computational studies are necessary to develop our understanding of the manner in which contaminants travel through the vadose zone. Applications of geophysical methods to various contaminated areas at the INEEL are given.
- AVO, amplitude variation with offset
- CR, Complex Resistivity
- DNAPL, dense nonaqueous phase liquid
- EM, electromagnetic
- EMI, electromagnetic induction
- EMSP, Environmental Management Science Program
- ERT, electrical resistivity tomography
- FDFD, finite-difference, Fourier-domain code
- GPR, ground-penetrating radar
- INEEL, Idaho National Engineering and Environmental Laboratory
- INTEC, Idaho Nuclear Technology and Engineering Center
- IP, induced polarization
- MRI, magnetic resonance imaging
- NAPL, nonaqueous phase liquid
- RWMC, Radioactive Waste Management Complex
- SDA, Subsurface Disposal Area
- SIP, spectral induced polarization
- TAN, Test Area North
- TCE, trichloroethylene
- TDR, time-domain reflectometry
- VETEM, very early time electromagnetic
- XBGPR, cross-borehole ground-penetrating radar
Buried waste at the INEEL threatens the health of the aquifer and the people, environment, and economy that depend on it. The buried waste at the INEEL has resulted largely from past practices of disposing wastes into unlined pits and trenches, mainly at the Subsurface Disposal Area (SDA) at the Radioactive Waste Management Complex (RWMC). Contaminants of concern at the SDA include volatile organic compounds, fission and activation products (e.g., 14C, 99Tc, 90Sr), isotopes of U, 241Am, 129I, 94Nb, Cd, Pb, nitrates, and Pu isotopes. Other sources of subsurface contamination at the INEEL are underground pipes that have leaked at the Idaho Nuclear Technology and Engineering Center (INTEC). The principal radionuclide contaminants of concern at INTEC are 241Am, 90Sr, 137Cs, 154Eu, Pu isotopes, and 235U. There is risk that these contaminants will migrate into the Snake River Plain Aquifer, the sole aquifer supplying drinking water for tens of thousands residents in Idaho, Oregon, and Washington.
The 1995 Settlement Agreement between the USDOE, the U.S. Navy, and the State of Idaho states, “DOE shall ship all transuranic waste now located at INEL [INEEL], currently estimated at 65,000 m3 in volume, to the Waste Isolation Pilot Plant or other such facility designated by DOE, by a target date of December 31, 2015, and in no event later than December 31, 2018.” In April 2003, an Idaho court ruled that the phrase “all transuranic waste” includes buried waste.
Geophysical imaging will be necessary to locate buried waste and residual contamination in the subsurface. Advanced geophysical characterization techniques developed for the subsurface are emerging from the research world ripe for implementation at waste sites. This article focuses on geophysical imaging research for vadose zone characterization developed under auspices of the DOE EMSP, which funds basic science research targeted at DOE's most pressing cleanup challenges. These basic science research projects will improve our knowledge of subsurface features, properties, processes, and contaminant distributions, which in turn leads to better decision-making regarding remediation and long-term stewardship strategies.
The goal of these research projects is to develop high-resolution measurement and interpretation packages that provide accurate, timely information needed to characterize the vadose zone and the contaminants it contains. Accurate characterization of the subsurface beneath DOE waste sites, such as at the INEEL, is essential to
Assess the inventory, distribution, and movement of contaminants
Develop improved predictive methods for liquid flow and contaminant transport
Provide a foundation for the selection of remediation alternatives
Enable effective performance monitoring of corrective actions
Figure 1 shows that basic research provides tools and techniques to yield information used in the decision-making process to protect groundwater supplies for future generations. High-resolution imaging of the subsurface can be used to map subsurface structures and contamination and investigate whether contaminants are leaking into the lower geological horizons via transport pathways. As a geologic system, the subsurface beneath the INEEL comprises highly heterogeneous fractured basalts interlayered with sedimentary deposits. Inadequate descriptions of geologic heterogeneity have resulted in uncertain predictions of contaminant behavior in the vadose zone. An accurate description of transport pathways on the gross scale, characterization of heterogeneity, and the location of contamination within the vadose zone are vital for proper treatment, confinement, and stabilization of subsurface contamination at DOE waste sites.
Our understanding of the subsurface structure and properties of a site is derived from field observations. The technical foundation for making decisions concerning the vadose zone requires detailed knowledge of subsurface processes and properties coupled with numerical and experimental capability to perform simulations to assess the effectiveness of remediation strategies. Engineers and scientists must be able to represent and visualize the subsurface structure and properties as part of a whole-earth system to integrate computational predictions of relevant processes with subsurface data. A combination of multidisciplinary laboratory, computational, and field-scale studies are necessary to test specific hypotheses and obtain a comprehensive understanding of vadose zone behavior. Therefore, characterization of the subsurface environment is a necessary precursor to developing isolation, stabilization, and treatment strategies, since the subsurface properties coupled with physicochemical and biological processes affect contaminant fate and transport. The overarching purpose of such analysis is to predict the potential for infiltration to cause movement of contaminants in the vadose zone before they migrate into the groundwater, where they are difficult and costly to remediate. Knowledge of contaminant levels and distribution throughout the subsurface is thus necessary to develop remediation and long-term stewardship monitoring strategies.
Stewardship plans for protecting groundwater supplies rely on an accurate predictive capability for contaminant transport in the subsurface. Current abilities to predict and optimally manage fluid flow and contaminant transport processes in the subsurface are limited by inadequate descriptions of subsurface physical, hydrological, geochemical, and biological heterogeneities (i.e., those occurring naturally and those created by waste interaction with the porous medium) that will influence the distribution of contaminants. Consequently, predictive models have often failed to provide accurate results and have underestimated transport rates. Radioactive and hazardous contaminants have been found in groundwater monitoring wells at the INEEL. At the SDA, the presence of 241Am has been detected at least twice in one well (Olson et al., 2003). These detections were not expected nor predicted by models developed from conventional flow and transport theory for contaminant migration through the vadose zone. Remediation of contaminants in the shallow vadose zone is preferable to waiting until they have migrated to greater depths or into the aquifer.
ADVANCED GEOPHYSICAL CHARACTERIZATION
We discuss a variety of complementary geophysical methods providing subsurface characterization data, including EM, GPR, electrical, seismic, and NMR. These geophysical imaging techniques enable characterization of subsurface conditions and processes from a limited number of boreholes and surface measurements.
Noninvasive, in combination with minimally invasive, borehole techniques are becoming increasingly important for shallow subsurface characterization of waste sites. Such techniques can provide continuous coverage of features and properties to supplement critical quantitative information obtained from more invasive methods, such as vertical survey profiles with borehole tools, point sampling and measurements, and permanent sensors (e.g., lysimeters). Information from a limited number of test wells can thus be extrapolated to a larger area, reducing the risks and cost of obtaining information needed to manage contaminated waste sites. Besides the obvious expense of drilling boreholes for characterization and monitoring, drilling poses the hazard of spreading contaminants and entails disposal of secondary contaminated waste from drilling operations. Geophysical high-resolution imaging can potentially be used as a site-screening tool to detect and locate specific buried waste (e.g., containers of radioactive wastes) and areas of contamination (e.g., leaking containers and gas or liquid plumes) for subsequent drilling and sampling if necessary. Since individual geophysical characterization tools all have specific limitations, multiple techniques are being explored to provide more widespread applicability across a range of hydrogeologic settings. This article describes how combined techniques can provide complementary information for more complete understanding of subsurface fate and transport.
INTERPRETATION AND APPLICATIONS OF GEOPHYSICAL DATA
Geologic mapping provides critical information on subsurface features and the distribution of hydrologic properties. In the unsaturated zone, soil properties change dramatically with moisture content. Subsurface hydrology is a particularly important aspect of site characterization because subsurface water and vapor transport are the primary natural pathways for the movement of contaminants to the accessible environment (Tindall et al., 1999). To convert between geophysical and hydrological properties requires application of petrophysical models, which relate the physics of geophysical properties, such as electrical conductivity and dielectric properties, to subsurface state properties, such as water or moisture content and permeability. Improved relationships between geophysical measurements, hydrological properties, and soil composition are needed. A more reliable mapping between geophysical properties and hydrological properties would provide information that could ultimately be used to reduce cleanup costs, accelerate schedules, and reduce risk.
Improved methods for interpreting geophysical properties from the raw measurement data collected in the field are also needed. Inversion algorithms are frequently used to “invert” or transform certain field measurements (such as radar, EM, and remote sensing data) into high-resolution images that provide more readily usable information. Two- and three-dimensional, forward modeling and inversion algorithms play an important role in environmental site characterization.
SURVEY OF EMSP-SPONSORED GEOPHYSICAL RESEARCH
Geophysical imaging techniques, including EM, GPR, electrical, seismic, and NMR, investigated by EMSP researchers at institutions across the country are being developed to characterize the contents of the INEEL waste pits and trenches. Since individual geophysical characterization tools have specific limitations, multiple techniques are being explored to provide broad applicability across a range of hydrogeologic settings. Combining techniques offers the potential to exploit complementary capabilities of each technique. These projects couple field techniques with data acquisition, data processing algorithms, and software to interpret the data. Modeling, including numerical methods, inverse theory, data analysis, and scientific visualization, is a necessary counterpart to experimental investigation. An example of a combined experimental and numerical approach is shown in Fig. 2 , where GPR antennas are used to acquire field data and numerical simulations are used to calculate the temporal and spatial behavior for radiation into dispersive and lossy earth for a variety of antenna geometries and earth conditions.
Each EMSP geophysics research project outlined in this survey serves to further the state-of-the-art for shallow waste site characterization. Recent research results and applications at the INEEL and other sites with similar geology and contaminants are discussed.
Electromagnetic induction (EMI) techniques provide a measure of the electrical conductivity variations in the subsurface. The source is a magnetic field generated by currents in wire coils, whereby sources and receiver coils are either located in boreholes or with a coil array on the surface. Electromagnetic induction methods are sensitive to the amount and distribution of fluid present in porous media. Monitoring soil moisture and electrical conductivity changes can provide indications of fluid-conveyed contaminant transport. Electromagnetic induction methods have been shown to be effective in environmental site characterization, but there is a need for increased spatial resolution for waste form characterization, verification, and monitoring activities.
Detection of Buried Metal Objects
A high-frequency impedance methodology that uses a window in the EM spectrum (1.0–100 MHz) between low-frequency conventional EMI and GPR has been investigated to yield high-resolution mapping of electrical conductivity, as well as the permittivity of near-surface formations. This technique, called the very early time electromagnetic (VETEM) system, can resolve small buried metal objects and determine both electrical conductivity and dielectric permittivity, which are related to shallow subsurface geochemistry and geohydrology. Very early time electromagnetic is designed to produce enhanced high-resolution EM images in shallow (0–5 m) subsurface regions where the electrical conductivity of the earth is too high for GPR to be effective, such as in the clay capping material covering the waste pits. The VETEM system was deployed at Pits 4, 9, and 10 in the Radioactive Waste Management Complex SDA at the INEEL. Data from Pits 4 and 10 were obtained using a VETEM system towed by an all-terrain vehicle equipped with a global positioning system that provides decimeter lateral positional accuracy. Inversion of the VETEM data provides depth estimates. Very early time electromagnetic geophysical images of waste drums buried in Pit 9 show high detail, and this information is being used to help determine the location and estimated depths of the buried waste containers and support remediation operations. Software and hardware enhancements to VETEM prototype instruments have enabled physical modeling experiments, numerical forward and inverse modeling, and field demonstrations (Cui et al., 2003).
A full, three-dimensional, nonlinear inversion algorithm, which uses a scattered-field solution approach, is being developed for imaging electrical impedance data. Inversion algorithms are typically very complex and computationally intensive. The INEEL is proposing to implement this algorithm on an INEEL cluster computing resource and perform inversion calculations on relevant data sets. Limitations of the EM impedance method include not precisely knowing when one is in the far-field of the transmitter, since this is geology dependent, and the scarcity of rigorous two- and three-dimensional algorithms to properly invert the data and bound the range of applicability of approximate methods. To address these limitations, the research team is currently (i) implementing full nonlinear two- and three-dimensional inverse solutions that incorporate source coordinates and polarization characteristics, (ii) using these solutions to study improvements in image resolution that can be obtained by making measurements in the near- and mid-field regimes using multiple source fields, (iii) collecting field data with recently developed earth impedance measurement systems, and (iv) interpreting the field data with the newly developed inversion capability, as well as with additional, independent information, such as well logs from boreholes. The goal of this research is a combined measurement–interpretation package for noninvasive, high-resolution characterization of larger transport pathways and the location of certain types of contamination. The application of rigorous two- and three-dimensional inversion codes will extend the range of complex heterogeneous imaging capabilities of these measurements (EMSP Project 70220, information available online at the EMSP Project Database http://emsp.em.doe.gov [verified 17 Dec. 2003]).
Surface and Borehole Electromagnetic Imaging
This research has produced a forward modeling capability based upon a three-dimensional finite-difference, Fourier-domain code (FDFD) (Champagne et al., 2001) to produce forward predictions of electrical and magnetic fields for a given source configuration and to compute the adjoint operator to be used in the inversion method. An adjoint method of data inversion was developed to reduce computational efforts of the inversion to allow more rapid processing of large geophysical data sets than previously possible. Data from cross-well and surface-to-borehole field tests show that extended conductivity information between boreholes results in much higher resolution than surface techniques and much greater penetration than radar. Reasonable agreement was obtained between experimental data and the FDFD code, especially when the receiver coil is located not too far from the transmitter well (Buettner and Berryman, 1999). Before this research, the state of the art in EM data inversion was based on the Born approximation. However, a nonlinear analysis is necessary since electrical conductivity variations in the subsurface exhibit a range of several orders of magnitude.
High-Frequency Electromagnetic Impedance Measurements
This research team is seeking to interpret the EM impedance data by simultaneously inverting electrical conductivity and dielectric permittivity using a single data set (Song et al., 2002; Frangos, 2001). Upon completion of the project, the system may be field deployed to monitor the moisture content of clay caps, detect buried objects, and monitor and predict vadose zone contaminant movement (related to electrical conductivity) at INEEL and other DOE sites. Proof of concept of the methodology has been successfully demonstrated using a prototype 0.1- to 30-MHz system with off-the-shelf components, including a magnetic dipole transmitter and electric and magnetic antennae. Electronics of these instruments have been carefully redesigned and repackaged for improved performance and field deployment. The electronic components have been miniaturized, packed together, and positioned at the center of the stub antenna; wiring was replaced with optical fibers; and the lock-in amplifier was replaced with a network analyzer to extend the operating bandwidth beyond 100 MHz (Tseng et al., 2003).
Ground-penetrating radar and time-domain reflectometry (TDR) techniques are based on the propagation and reflection of EM energy in the subsurface. Electromagnetic wave propagation and scattering are employed to image, locate, and quantitatively identify changes in electrical and magnetic properties of the ground. Ground-penetrating radar uses EM waves to detect buried objects and provide information about the variation in dielectric properties of the subsurface. For waste site characterization, GPR is typically performed from the surface of the earth, in a borehole, or between boreholes. Ground-penetrating radar has one of the highest resolutions in subsurface imaging of any noninvasive geophysical method, approaching a few centimeters, under the right conditions. Information such as depth, orientation, size, and shape of buried objects and the density and water content of soils can be derived from GPR data by quantitative interpretation through modeling. Detectability of a subsurface feature or object depends on depth, size, shape, and orientation relative to the antenna; contrast in electrical and magnetic properties with the host medium; complexity of the host medium; and radiofrequency noise and interferences (Olhoeft, 2000). Existing GPR systems have performed very poorly in highly conductive subsurface environments, such as the clay present at the Cold Test Pit at the SDA, which attenuates high frequency EM energy. The Cold Test Pit is a simulated waste disposal area used to test and demonstrate characterization, retrieval, and treatment technologies. It provides known targets and waste forms for accurate evaluation, calibration, and testing of procedures, technologies, and equipment. Areas at the INEEL where GPR is expected to perform well include the Idaho Nuclear Technology and Engineering Center, where improvements to GPR that would extend the depth of investigation for locating utility lines from about 2.4 to 3.7 m (8–12 feet) would be of immediate use (Schneider, 2002).
Spatial Variability of Subsurface Moisture Content
A focus of this research has been how surface and crosshole radar data can be used together to quantify the distribution of water content in the vadose zone. Three-dimensional maps of subsurface moisture content are being constructed from GPR data (Knight, 2001). One approach involves obtaining estimates of water content for a radar-sampled volume of the subsurface. A critical part of this research is quantifying the uncertainty in water content estimates due to heterogeneity below the scale of the sampled volume. A second approach involves working directly with the radar image to quantify the spatial variability in water content (Irving and Knight, 2000, 2003a, 2003b). Researchers are using a modified version of GSLib for geostatistical analysis of GPR images. Field tests at the Sisson and Lu site at Hanford have yielded encouraging results when the analysis of the radar data is compared with the information about water content obtained from neutron probe data (Knight et al., 2003).
A suite of methodologies for direct detection of dense nonaqueous phase liquid (DNAPL), specifically chlorinated solvents, is being developed via material electrical property estimation from surface GPR data. Most organic liquids have lower dielectric permittivity and electrical conductivity than water, so a contrast in properties is induced when DNAPL displaces water. To identify shallow DNAPL source zones, the research team is examining three aspects of reflected wave behavior—propagation velocity, frequency-dependent attenuation, and amplitude variation with offset (AVO). Velocity analysis provides direct estimate of electric permittivity, attenuation analysis provides a measure of electrical conductivity, and AVO behavior provides an estimate of the permittivity ratio at a reflecting boundary (Bradford, 2003). Two- and three-dimensional multioffset, multipolarization data sets were acquired at nine research sites in seven field areas at two DOE and five Department of Defense facilities. The researchers believe they have obtained the first reported case of GPR AVO and migration velocity analysis being used for direct detection of NAPL in an uncontrolled field setting over an existing plume. Quantitative analysis of multioffset radar data to identify electric property anomalies that may otherwise have gone unnoticed in qualitative interpretation of conventional radar profiles enabled the detection of NAPL in a location previously thought to lie outside of the NAPL plume (Deeds and Bradford, 2002).
Subsurface Flow and Transport Imaging
Cross-borehole GPR (XBGPR) velocity and attenuation imaging has been employed to better understand contaminant movement in the vadose zone. During the first part of the research, the team analyzed the ability of XBGPR imaging to recover in situ estimates of moisture content and to monitor unsaturated water flow through the near surface. More recent research is focusing on the inclusion of attenuation data to better delineate low-permeability clay-bearing zones, and to image unsaturated contaminant transport processes (Alumbaugh et al., 2002).
Improvements for Challenging Field Conditions
Ground-penetrating radar is limited by losses in materials, such as many types of mineralogical clay, with a high electrical conductivity. For common earth soils with gravel-, sand-, and silt-sized particles, the scattering losses are high when there are subsurface property variations with random orientations that are close in size to the dominant wavelength of the propagating energy. This research seeks to extend the limits of performance of GPR, specifically the depth of investigation and the interpretability of images from highly variable, attenuable, and dispersive earth via improvements in hardware and in numerical computations. Key features include (i) greater dynamic range through real-time digitizing, signal averaging, and receiver gain improvements; (ii) modified, fully characterized antennas with current sensors to allow dynamic measurement of the changing radiated waveform; (iii) modified deconvolution and depth migration algorithms exploiting the new antenna output information; and (iv) development of automatic, full waveform inversion made possible by the known radiated pulse shape (EMSP Project 86992, information available online at http://emsp.em.doe.gov [verified 17 Dec. 2003]).
Biostimulation Monitoring Using Radar Techniques
Remediation approaches, such as in situ chemical oxidation or bioremediation, can induce dynamic transformations in subsurface systems that impact remediation efficacy. On the basis of successful use of geophysical data to estimate hydrological properties at both laboratory and field scales (Hubbard and Rubin, 2000), the research team is investigating the geophysical methods as a noninvasive means of providing information on subsurface-coupled processes (Hubbard et al., 2002). The researchers demonstrated the ability to monitor the evolution of biogenic gas using TDR during a controlled column-scale biostimulation experiment. Hydraulic conductivity decreased by 55% during the experiment as a result of generated gas bubbles clogging the pore spaces in an originally water-saturated sand column. Radar velocity data suggest that 24.6% of the available pore space was occupied by gas, which agrees favorably with the estimate of 23.3% obtained from column weight loss measurements (Hubbard and Williams, 2004).
Electrical resistivity tomography (ERT) is a direct-current resistivity technique that measures electric potentials generated by a current source either on the earth's surface or in the subsurface. These potentials are sensitive to the bulk electrical properties, indicative of porosity; the amount and connectivity of pore fluid; and the pore fluid chemistry. Electrical methods have the potential to identify fluids, and time-lapse measurements can assess fluid movement, saturation changes, and deduce permeability (Wilt et al., 1995). However, these measurements can be affected by buried pipelines and cables, topographical variations, and geochemical heterogeneity of the subsurface. Electrical resistivity tomography data require an inversion algorithm to produce a subsurface image. Mathematical inversion of the resistivity measurements can be used to derive a three-dimensional image of the possible water and contaminant distributions in the subsurface by taking advantage of water and contaminants' ability to conduct electricity much more readily than rock. Induced polarization (IP) is a current-stimulated electrical method that provides information on the subsurface chemical reactions taking place by using an electrode setup identical to the resistivity method and measuring the time-domain decay of voltage in the ground from an induced electrical signal (National Research Council, 2000). In the frequency domain, the phase and amplitude of the induced voltage are measured at several frequencies. Spectral IP (SIP) or complex resistivity (CR) techniques typically employ an extended number or spectrum of frequencies ranging from several millihertz to as high as a megahertz.
Spectral Induced Polarization
Electromagnetic coupling noise presents a critical limitation for field implementation of SIP, and conventional correction methods are inadequate. Research to resolve the EM coupling problem with SIP includes explicitly combining EM induction physics in three-dimensional IP modeling inversion codes. This project treats EM coupling signals as data, rather than as useless noise, and extends the usable frequency range of the data to >1 KHz. In 2001, the team performed electrical surveys using electrical-impedance tomography to detect DNAPL contamination at the Savannah River Site (Wang, 2002). Electrical-impedance tomography was chosen since it is capable of mapping reactions involving inorganic compounds. The field study shows that potential bearing electrodes should be carefully chosen to avoid capacitive coupling problems, which become more important for borehole sampling (EMSP Projects 55300 and 73836, information available online at http://emsp.em.doe.gov [verified 17 Dec. 2003]).
Four-Electrode Complex Resistivity
Complex electrical resistivity measurements are being refined for monitoring DNAPL contamination in the subsurface. Strong electrical signatures are characteristic of certain organic solvents, notably toluene, perchloroethylene, and trichloroethylene (TCE), in clay-bearing soils (Slater and Lesmes, 2002b).This technique is potentially useful to detect the presence of DNAPLs along interfaces between basalt and clay-bearing interbeds, which constitute a large portion (depending on location) of the subsurface at the INEEL (Nace et al., 1956, 1975; Hackett and Smith, 1992). The work is based on a four-electrode electrical resistivity measurement, where two electrodes are used to impose a sinusoidal current, and the remaining two electrodes sense the response voltage of the sample. A laboratory system has been designed and built, including a sample holder, electrodes, electronics, and data analysis software. Sample holders fit directly into an agricultural soil-sampling auger, allowing samples to be collected and their electrical properties measured with minimal disturbance to the soil microstructure. The team has successfully reproduced published results for the amplitude and phase response of brine-saturated glass-bead packs with specified volume fractions and grain diameters of iron pyrite, and also for glass beads mixed with calcium montmorillonite clay (Slater and Lesmes, 2002a). The team is currently trying to reproduce published results on the complex resistivity response of toluene-contaminated clay-rich samples.
Seismic methods can resolve certain physical characteristics of the subsurface geology by measuring the travel times of acoustic waves. Fluid-gas boundaries and gas production from biomass activity can be detected using the amplitude of the reflected seismic signal, such as the “bright-spot” signature in seismograms used to identify gas reserves in the oil and gas industry. Seismic methods have the potential to fully depict the contours and connectivity of subsurface aquitards and other geologic structures for design of groundwater remediation programs (Mathisen et al., 1995). Use of seismic imaging methods can potentially reduce uncertainties associated with contaminant pathways due to the scarcity of test wells and the highly heterogeneous nature of the subsurface. A wide range of source–receiver configurations can be used, including surface to surface (seismic reflection profiling), well to surface (vertical seismic profiling), and well to well (cross-well transmission tomography) (Berryman et al., 2002). Images of shallow structures can be obtained from the periphery of the waste site and do not require access to the surface above the structure.
Integrated Suite of Imaging and Inverse Techniques
The objective of this research is to develop structural maps of the shallow subsurface with sufficient detail to be useful for site cleanup. The research team conducted an extensive three-dimensional seismic experiment at a groundwater contamination site in Utah, where DNAPLs are confined at the base of a shallow aquifer (Levander et al., 2001a, 2001b; Dana et al., 2001). Images characterizing the test site were produced at a number of different scales, from tens of meters to submeters. Three-dimensional seismic results show images of a shallow paleochannel incised in a clay acuiclude that acts as a structural trap for DNAPLs. An integrated suite of imaging and inverse techniques for two-dimensional and three-dimensional seismic data in the near-vertical to wide-angle propagation regime have been developed by extending Kirchhoff inversion, depth focusing, and full waveform inversion methods to three-dimensional imaging of the shallow subsurface (Zelt et al., 1999).
Surface-Wave Group-Velocity Tomography
An analysis technique for surface-wave group-velocity tomography is being developed to map the primary fluid pathways in shallow soils. Measurements of perturbations in the shear wave velocity of the soil are made by measuring changes in seismic response induced by fluid flow. The observed changes in surface-wave velocity are then used to map areas where fluids have modified the fluid pressure and, hence, the shear-wave velocity. Areas experiencing the greatest pressure and material properties changes would indicate primary flow paths and zones that are most likely amenable to remediation by extraction or flushing. The technique could be used before remediation to detect flow paths and help design an optimal remediation process, or during remediation to detect and determine which zones are reached by the remediation. The sensitivity of shear waves to fluid content could eventually allow surface-wave tomography to track fluid (i.e., water or organic contaminants in soils) movement with time. Information such as the time variations in the depth of the water table would be provided without drilling wells to obtain single-point values (Long, 2002; Long and Kocaoglu, 2001).
Automated, Ultrashallow Seismic Imaging
The research team has demonstrated the use of an automated geophone-planting device to plant large numbers of geophones to acquire data, which is expected to make shallow seismic surveying more efficient and less expensive. The automated approach has been demonstrated to provide imaging results comparable with those obtained with conventionally planted geophones. Experiments were performed to show that good seismic data can be recorded when interconnected geophones are mounted on steel bars (Tian et al., 2003a, 2003b). Their progress can be attributed to refinements in the design of the geophone-planting device and to an improved ability to measure the near-source wavefield. Traditional shallow seismic-reflection methods are capable of imaging the subsurface from about 2 to 30 m, but are ineffective in the very near-surface regime. By modifying the field layout of the geophones and using an alternative seismic source (a single shot from a 0.22-caliber rifle), it has been possible to image subsurface depths of 0.6 to 2.1 m. Data were collected using single 100-Hz geophones spaced at intervals of 5 cm, rather than 1 m or more, as in typical shallow seismic surveys. Increasing the spatial density of the geophones by a factor of 20 or more improves the ability to delineate the coherence of ultrashallow seismic reflections over other interfering phases. Also, it was found that larger, more powerful sources generated near-field nonlinear soil deformation strong enough to prevent detection of ultrashallow seismic reflections (Baker et al., 2000). Complementary site characterization capabilities of three-component shallow (compressional) P-wave seismic reflection techniques and GPR methods at ultra-shallow depths are being explored (Baker et al., 2001).
Crosswell Seismic Imaging for Bioremediation
Crosswell seismic tomography was applied at the INEEL to image the fractured basalt aquifer at the Test Area North (TAN) site where wastes were disposed of directly into an injection well. Such noninvasive imaging complements more invasive techniques, such as core sampling, and offers the potential to estimate the distribution of the biologic properties needed for detailed site characterization and assessment of the effectiveness of bioremediation. An orbital vibrator seismic source was used to simultaneously acquire P- and S-wave tomography data using fluid coupled hydrophone sensors (Daley and Cox, 1999). Regions of slow seismic velocity were found to correspond with zones of contaminant transport (and by inference higher permeability and increased fracturing); regions of low amplitude (high attenuation) appeared to correspond with zones of permeability. High attenuation of seismic energy occurred at a known zone of high concentration of contaminants, which may be due to either high concentrations of DNAPL in the fractures or trace amounts of gas from biodegrading contaminants. This work was performed before nutrient injection (lactate addition) for accelerated bioremediation. These results suggest that geophysical methods may be sensitive to detecting by-products of biogeochemical processes, such as N2 gas (from the sludge injected at TAN); biogenic iron, carbonate, and sulfide minerals; and biogenic methane gas (Daley et al., 1998). The seismic study identified primary flowpaths for the chlorinated contaminants near the “hot spot” at TAN, which gave a separate EMSP research team the ability to reconcile samples of the microbial community taken near these the biologically active zones (Tobin et al., 2000; Lehman et al., 1999).
Nuclear Magnetic Resonance
Nuclear magnetic resonance and magnetic resonance imaging (MRI) can be used to observe and quantify the location and size distributions of pores with proton-containing liquid or vapor, such as nonaqueous phase liquids (NAPLs) and water. Nuclear magnetic resonance field logging tools have been successfully used for more than a decade by the oil and gas industry to characterize fluid types (hydrocarbons and water), distinguish mobile fluids from bound fluids, and determine formation permeability. Surface NMR allows direct, noninvasive sounding of groundwater distribution vs. depth (Shushakov et al., 2004). Magnetic resonance imaging provides three-dimensional images of the phases within the individual pores as a function of space and time. Ultimately, miniaturized magnets on sensor platforms deployed in existing test wells may provide in situ species identification and quantification of various subterranean chemical compounds. Depending on the type of MRI unit used, these methods can provide quite good spatial characterization (on the order of centimeters), but at present cannot detect trace quantities of NAPLs. Chlorinated solvents are present in the subsurface soils at the SDA.
Magnetic Resonance Image of Soil Columns
Magnetic resonance imaging has been used to observe and quantify the location and size of individual pores containing DNAPL, water, and vapor flow in soil columns. Laboratory column experiments with decane have shown that MRI signal intensity is linearly correlated to the amount of decane trapped in the columns. In conjunction with MRI, constitutive relations are being developed for use in modeling techniques to describe the transient distribution of phases inside a column experiment. These constitutive relations will be incorporated into a site-scale transport model to evaluate how processes, such as slow interphase mass transfer, retarded vapor phase transport, and diffusion from unswept zones of low permeability, affect the performance of soil vapor extraction. A unique aspect of their model is the effect of water saturation on NAPL volatilization. One-dimensional simulations at both high and low water saturation reveal that temperature and water content changes resulting from water evaporation are only marginally important. The effect of vapor sorption can be neglected except under extremely dry conditions, which indicates that simplified models that neglect temperature change and water evaporation are valid under most field conditions (Yoon et al., 2003).
Magnetic Resonance Imaging of Water Content Distribution
Field measurements were performed using surface NMR to determine water content distribution in the subsurface. As opposed to techniques that infer water content from other measurements, NMR offers a direct, noninvasive measurement through its sensitivity to H nuclei in free (mobile) water. Most other geophysical methods (e.g., electrical resistivity, dielectric constant, seismic velocity) do not directly measure water content. For example, electrical measurements require a geophysical inversion to determine the conductivity of the subsurface, which is followed by a geophysical interpretation to infer the water content from the geophysical property, conductivity. From the results of the field measurements, it was concluded that (i) water content distribution could be determined in coarse-grained, but not fine-grained, soils; (ii) the signal to be measured is low amplitude (tens to hundred of nanovolts) and is susceptible to interference from power lines; (iii) the presence of ferromagnetic minerals in the soil and paramagnetic ions in the water interferes with reliable measurements; (iv) the field strength of the NMR signal is concentrated under the loop and therefore little lateral dissipation occurs; and (v) accurate inversion requires knowledge of the geoelectric section (dielectric and conductivity structure) provided by auxiliary measurements. Several advances need to be made to make surface NMR a practical geophysical characterization tool. These include improvements to noise suppression to enable use in EM environments, shortening the instrument delay time to enable detection of water in small pores and in soils with magnetic impurities, and advances in modeling (EMSP Project 54857, information available online at http://emsp.em.doe.gov [verified 17 Dec. 2003]).
Measurements of Adsorbed Organic Contaminants
Experiments were performed to measure the NMR response of clay-free water-saturated sands with various amounts of adsorbed oil. It was shown that NMR measurements can be a useful way of detecting trace amounts of adsorbed organic contaminants in geological porous media (Daughney et al., 2000). Oil sorbed on the surfaces of sand grains affected NMR relaxation time by changing both the area and the effective relaxivity of the grain surfaces (Bryar and Knight, 2003b; Bryar et al., 2001a). This research also obtained results regarding the paramagnetic effects of Fe(III) on NMR data, which must be taken into account to accurately interpret NMR measurements in the near-surface environment (Bryar et al., 2000). In addition, changes in relaxation time observed when Fe(II) was oxidized to Fe(III) have important implications for the potential use of NMR field instruments to monitor changing redox conditions in near-surface environments using the Fe(III)–Fe(II) redox couple as an indicator (Bryar and Knight, 2002).
Nuclear Magnetic Resonance Downhole Logging Tool
The capability and limitations of low-field NMR relaxation decay-rate measurements for determining environmental properties affecting DNAPL flow in the subsurface are being investigated. This research leverages the work on NAPLs described in the paragraph above (Bryar et al., 2001b; Bryar and Knight, 2003a). The oil and gas industry uses NMR measurements to determine porosity and hydrocarbon content and to estimate formation permeability in deep subsurface formations. These determinations rely on the ability of NMR to distinguish between water and hydrocarbons in the pore space and to obtain the distribution of pore sizes from relaxation decay-rate distributions. The potential of NMR decay-rate distributions for characterizing DNAPL fluids in the subsurface and understanding their flow mechanisms has not been exploited, but is being explored in this project, as shown in Fig. 3 . Near-surface, unsaturated environments provide unique challenges for using NMR. These challenges are being addressed through systematic laboratory experiments at the INEEL and through a program of research to extend and adapt current low-field NMR measurements to near-surface environmental problems. The product of this research will be a downhole logging tool that is sensitive to microbiological influences on TCE fate and distribution. Recent experiments have started to explore the potential of combining complex resistivity and NMR to characterize large volumes of the subsurface. Both measurement modes have complementary sensitivity to subsurface properties such as subsurface chemistry, microgeometry, and pore-size distributions. The research focuses on the effects of sand–clay mixtures with fluids comprising water and DNAPL, and biological impact on these fluids. Understanding the CR and NMR responses to these pore-fluid interactions will lead to a technique than can simultaneously extend NMR measurements away from the borehole and validate CR interpretations (EMSP Project 86804, information available online at http://emsp.em.doe.gov [verified 17 Dec. 2003]).
Multigeophysical Mode–Data Fusion Methods
Hybrid Hydrological–Geophysical Electrical Inversion Techniques
Hydrologic data are being combined with ERT images in a hybrid hydrological–geophysical inverse technique to improve our ability to characterize the vadose zone and to monitor and predict contaminant movement in the vadose zone. During the first part of the research, the team obtained hydrologic and geophysical data at the Sandia-Tech Vadose Zone Facility infiltration test site in New Mexico. This data set was used in both forward and inverse hydrologic algorithms to describe the relationship between in situ moisture content and matric potential, as well as the saturated hydraulic conductivity (LaBrecque et al., 2002). Subsequent research has focused on the use of geophysical imaging to estimate transport processes in unsaturated conditions. The researchers have developed a new anisotropic ERT inversion code for commercial applications and a new aquifer test method, called hydraulic tomography, for cost-effective characterization of aquifer heterogeneity at high resolutions that were previously not possible (Yeh and Liu, 2000; Liu et al., 2000). Furthermore, the research has developed a stochastic information fusion approach to synthesize the hydrological and geophysical data sets. The approach aims at maximizing the usefulness of available information to obtain the best unbiased estimates of the subsurface properties and, meanwhile, quantify their uncertainties (Yeh and Simunek, 2002). Other objectives of the research include to thoroughly evaluate how well ERT and XBGPR can resolve subsurface hydrological features and processes within the vadose zone, to provide synthetic data sets for other researchers to use for testing and assessment, and to assess an interactive method for determining subsurface flow and transport properties within the vadose zone (LaBrecque et al., 2002).
Seismic–Electrical–Hydrological Data Interpretation
Research is being performed to relate measured geophysical properties, such as seismic velocity and electrical conductivity, to hydrogeology parameters of interest, such as porosity, saturation, and soil composition. The approach combines laboratory research, available field data, rock physics, and modeling to find relationships between geophysical measurements, hydrogeological parameters, and soil composition. Controlled laboratory measurements were conducted to determine ultrasonic velocities and complex impedances in artificially constructed soils of known compositions under known saturation and pressure conditions to simulate prescribed depths. Researchers compared their measurements to theoretical models, other laboratory measurements from literature, and available field data and investigated the role of microstructure, fluid and clay distribution, and chemical effects on measured geophysical properties. The researchers have developed and tested algorithms to relate measured geophysical properties to porosity, soil composition, and fluid distribution (Bertete-Aguirre et al., 2002; Berryman et al., 2000).
DISCUSSION AND SUMMARY
This article has surveyed EMSP-sponsored geophysical research being developed at the INEEL and EMSP research by others performed at the INEEL or applicable to INEEL problems and subsurface soils. These research results are broadly applicable to other waste sites with similar geology and subsurface contamination. Additional information on the EMSP research described herein, including annual and final reports, can be found online at http://emsp.em.doe.gov.
We identified three common themes in the EMSP-sponsored research surveyed here: (i) the development of high-resolution imaging capabilities to capture important details of subsurface properties and processes, (ii) the coupling of nonintrusive geophysical surveys (e.g., electrical surveys) with detailed quantitative precise point-sensor measurements (e.g., lysimeters and vapor-port systems) or borehole (e.g., NMR, neutron-based moisture and geochemical tools) measurements to extend high-precision knowledge away from the borehole, and finally (iii) the application of multiple geophysical methods to constrain the uncertainty in determining critical subsurface physical properties.
Examples of the first theme include the electrical forward and inverse modeling techniques for complex hydrogeological structures and the development of high-resolution GPR systems through integrated system developments using detailed sensor physics (antenna–soil characterization), full waveform inversions, and enhanced instrument capabilities. In the final analysis, high-resolution geophysical imaging is critical to measuring and monitoring transport properties and coupled biogeochemical and hydrological processes in the subsurface. Preferred flow paths must be identified, and these typically result from small-scale features such as fractures and fissures. Also, the ability of geophysical methods, including seismic, radar, and electrical, to monitor dynamic subsurface processes that occur during remediation could reduce the uncertainty concerning the efficacy of remediation (Hubbard and Williams, 2004).
An example of the second theme is the recent combined EMSP-sponsored borehole NMR research with electrical CR capabilities to extend the spatial range of characterizing DNAPL pore-fluid interactions away from the borehole. Huisman et al. (2003) discussed the use of GPR to measure soil water content at an intermediate scale between remote sensing and TDR data. Nuclear magnetic resonance and MRI techniques can help determine constitutive relations for fluid movement (e.g., DNAPL sliding over water-wet pore surfaces), which will lead to improved computational models for site-scale transport of contaminants. Combining GPR with neutron-probe moisture measurements to extend high-precision moisture characterization between boreholes (Knight, 2001) is another example of this theme. Neutron-probe measurements are sensitive to the subsurface H content and can be quantitatively calibrated to the water content. A GPR survey is sensitive to the dielectric constant of the subsurface, which is dominated by the water content, and can be used as a proxy survey to map the water content in the subsurface of the large regions between the boreholes.
Examples of the third theme include data fusion with hydrological measurements using ERT (Yeh and Liu, 2000; Yeh and Simunek, 2002) and GPR in combination with ERT (LaBrecque et al., 2002). These approaches enhance the determination of key vadose zone subsurface properties such as moisture content and capillary pressures, as well as saturated hydraulic conductivity. Integrating seismic measurements with electrical data and hydrological properties may also improve the capability to determine soil porosity, composition, and fluid distributions (Bertete-Aguirre et al., 2002; Berryman et al., 2000).
These themes highlight the need for higher-resolution noninvasive imaging capability and the fact that subsurface characterization goes much further than simply acquiring raw geophysical data and then inverting the data to obtain a physical property. All modes of geophysical measurements are subject to the impact of the sensor-environmental physics, and, as such, they depend on a full characterization of the environmental influence on the geophysical parameter. A variety of geophysical techniques with complementary capabilities (e.g., EM, radar, electrical, seismic, and NMR) are being pursued to provide a more complete picture of vadose zone properties and processes. Since individual geophysical characterization tools all have specific limitations, a suite of different types of measurements is sought to provide better understanding of contaminant fate and transport in the vadose zone.
There is no “magic bullet” in geophysical characterization. Uncertainties in interpreting geophysical properties and determining the key subsurface parameters critical for shallow waste site characterization can be reduced through multiple modes of physical measurements, sensors, and data fusion. The synergy of noninvasive, in combination with minimally invasive, borehole measurements can increase our capability to image the shallow subsurface at the INEEL. This knowledge will help scientists understand vadose zone issues at other sites with similar subsurface structures. Research transfer of the latest geophysics techniques, methodology improvements, and interpretation algorithms to practicing environmental geophysicists and site engineers is essential to support cleanup operations at contaminated sites.
This work is sponsored by the U.S. Department of Energy under DOE Idaho Operations Office Contract DE-AC07-99ID13727. The research cited herein was supported by the U.S. Department of Energy Environmental Management Science Program.
- Received December 4, 2003.