BARRIER SYSTEMS for ENVIRONMENTAL CONTAMINANT CONTAINMENT and TREATMENT - PART 4 doc

Geophysical methods are needed to: 1 provide the spatial distribution ofcertain physical properties that are essential for site characterization; 2 map thedistribution of some contaminan

Trang 1

Geophysical Method Verification

is due to several different issues specific to barriers If one is trying to see achange in the properties of a barrier it is not as challenging as seeing absolutechanges If one is trying to characterize or find a leak in the barrier this may bejust as difficult as finding a contaminant The issue is particularly challengingbecause of the following:

* With contributions by Randolph J Cumbest, Westinghouse Savannah River Company, Aiken, South Carolina; Bruce Davis, National Aeronautics and Space Administration, Stennis Space Center, Mis- sissippi; William E Doll, Oak Ridge National Laboratory, Oak Ridge, Tennessee; Leland Estep, Lockheed Martin, Midland, Texas; Susan S Hubbard, Lawrence Berkeley National Laboratory, Berkeley, California; John D Koutsandreas, Florida State University, Tallahassee, Florida; David P Lesmes, Boston College, Chestnut Hill, Massachussetts; H Frank Morrison, University of California, Berkeley, California; Lee D Slater, University of Missouri at Kansas City, Kansas City, Missouri; Anderson L Ward, Battelle Pacific Northwest Laboratory, Richland, Washington; Chester Weiss, Sandia National Laboratories, Albuquerque, New Mexico

4040_C004.fm Page 209 Thursday, September 15, 2005 11:50 AM

Trang 2

210 Barrier Systems for Environmental Contaminant Containment & Treatment

In traditional oil and gas subsurface applications, a 50% recovery rate isconsidered a great success The great majority of geophysical and remote sensingmethods were developed with this level of sensitivity In remediation applications,this recovery rate is usually not sufficient

Although oil and gas applications are multi-phase, the variations in the erties are not as large as in near-surface, partially saturated systems encountered

prop-in the vadose zone or even prop-in saturated environments (i.e., groundwater inants can be particles, chemicals that dissolve in water, or liquids or gases thatare only partially soluble in water) Under certain conditions, some contaminantscan move through unsaturated soils and rocks as vapor Contaminants can alsointeract strongly with the minerals in the subsurface Clays can absorb somecontaminants while some may form chemical complexes with other groundwaterchemicals Immiscible dense liquids can settle vertically, while some may becomenutrients for microbes that are present naturally or have been introduced All ofthese interactions may or may not affect the geophysical signals

contam-A variety of methods exist that could be classified as geophysical techniques;however, this chapter focuses on geophysical methods that are used to infervolumetric (average over a volume of material rather than at a point) rather thanpoint properties, i.e., crosshole, surface, and surface to borehole methods ratherthan well-logging techniques which usually only measure a few centimeters to ameter away from the borehole The methods are assumed to be applied from thesurface and boreholes or by placing sensors and/or sources in or near the barriers,thus imaging the volume or planes between the surface and borehole, the volumefrom the surface to the borehole, or a volume from the surface to a reflector orother target in the subsurface Last but not least, two main applications areassumed with respect to barriers: (1) the initial and subsequent characterization

of the subsurface volume to be contained, and (2) the verification of the integrityand performance of the barriers These issues are linked and must be addressed

to validate overall system performance

4.1.1 C HARACTERIZATION AND G EOPHYSICS

A simple definition of characterization is mapping the distribution of contaminantsources and effluents as well as the physical, chemical, and biological properties

of the subsurface materials that control their distribution, concentration, andmovement Some of the physical properties required are lithology, fault/fractureproperties, porosity, permeability, grain size, and fluid type and saturation Rock

or soil types, mineralogy and distribution, and types of clay minerals are alsoneeded to model chemical processes Chemical state, temperature, fluid satura-tion, and other factors that affect the presence and amount of nutrients are alsoneeded to determine microbe behavior Characterization as defined here is theessential first step toward containment and/or remediation, but all too often theterm is used only to describe the extent of the contamination itself, usually over

Trang 3

Airborne and Surface Geophysical Method Verification 211

a small area or volume that is relatively small compared to the entire groundwatersystem in which it resides This concept of total system characterization is critical

in containment applications because, as will be seen, the application of ical methods for containment depends on detecting changes from background orinitial conditions As a result, characterization efforts are currently often limited

geophys-to determining the nature and extent of the geophys-toxic materials and not defining thewhole regime in which they are traveling, interacting, and evolving This limiteddefinition can be useful in small-scale sites where the solution is excavation, but

it is only half of the story at thousands of larger scale sites The additional conceptthat the distribution of properties and processes should also be characterized isjust now being incorporated into idealized or conceptual models of hypotheticalsites in anticipation of when actual site data permit contaminant fate and transportsimulation and eventual remediation Only in the last five years have geophysicalmethods been used to measure the spatial distribution of the properties at actualsites to provide constraints for ground water models in a quantitative sense(Hubbard et al., 2001, 2003; Grote et al., 2003) If the subsurface were uniform

or even uniformly layered, drilling on a loose grid of holes would probably suffice

to characterize the site Unfortunately, the subsurface is generally heterogeneous,and a program based on drill-hole samples and measurements would provideincomplete or, at worst, misleading information Thus, volumetric information(information connecting the actual points of measurements) is needed

Geophysical methods are needed to: (1) provide the spatial distribution ofcertain physical properties that are essential for site characterization; (2) map thedistribution of some contaminants; and, in some cases, (3) detect chemicalchanges associated with contaminant interaction with the subsurface and barriers.Indeed, a useful definition of applied geophysics is that it is the science of usingphysical measurements or experiments on the surface (or from boreholes drilledfrom the surface) to determine the physical properties and processes in thesubsurface Geophysics is ideally suited for extrapolating measurements obtainedfrom a borehole to the large-scale volume away from the borehole (Peterson et al.,1985; Parra, 1991; Krohn, 1992; Sheets and Hendrickx, 1995; Majer et al., 1997)

In this application, geophysical measurements obtained from the surface orbetween boreholes can be used to assess the continuity and homogeneity of theintervening material Geophysics can also serve to map the subsurface in theabsence of boreholes and can be used to detect the unexpected such as a change

in lithology, fractures, or fast paths (Leary and Henyey, 1985, Davis and Annan,

1989, Hendrickx et al., 2002, Hubbard et al., 2002, 2003) Failure to be aware ofsuch gross heterogeneity has a major impact on hydrologic flow models andcontaminant transport (Majer et al., 1997) Finally, geophysical methods could

be used to delineate contaminants if the waste was buried in containers becausethe waste containers produce a geophysical anomaly or the waste alters theproperties of the medium (Doll et al., 2000) Table 4.1 shows the different reso-lution of the seismic and electrical methods and their expected use and application

Trang 4

4.1.2 P ERFORMANCE M ONITORING AND G EOPHYSICS

An important need for geophysics is for monitoring the processes that are mented to remove, contain, or treat contaminants In the case of containments,the ability of geophysical methods to monitor the emplacement and performance

imple-of the barriers primarily depends on the geophysical contrasts imple-of the barrier andsubsurface However, in some cases, even though the barrier does not look anydifferent than the surrounding properties, geophysics could possibly monitorchanges in the barrier properties relative to the native materials, monitor flow

TABLE 4.1 Possible Surface Geophysical Methods for Verification

of Subsurface Barriers

Expected Resolution a

Surface Methods

characterization, caps and walls

Fair Use for structure

and lithology of interior

0.5–5 m

Electrical (electromagnetic, induced polarization, self potential, DC resistivity)

Host characterization, caps and walls

Good Fluid content and

conductivity

1.0–10 m

characterization, caps and walls

Good Water content

differences

0.25 m Radar tomographic

amplitudes

Barrier detection Excellent Processed for

differences

0.25 m Radar well-to-well

reflection

Barrier detection Poor Low

signal-to-noise ratio Electrical resistance

tomography

Electrical resistance tomography

Leak detection Excellent Differences

during salt water flood

a Estimated only for successful borehole methods.

Trang 5

paths within the contained zone, and/or detect processes occurring due to thepresence of contaminants Once a site has been characterized and modeled and

a remediation process designed and implemented, it is necessary to assess theeffectiveness of the remediation operation Geophysical methods are ideallysuited to this task, because it is often easier to monitor changes in some portion

of the subsurface than it is to uniquely determine the subsurface propertiesthemselves, i.e., time-lapse monitoring (Dailey and Ramirez, 2000) An example

of time-lapse data is given in Figure 4.1 This is a plan view of a site wheremoisture monitoring is performed by observing the changes in signals fromground-penetrating radar (GPR) (Grote et al., 2003) As seen in the differences

FIGURE 4.1 Comparison of volumetric water content estimates obtained from 900-MHz common off-set GPR ground wave data during two different times of the year over a natural field study site These images reveal a persistence of near-surface water content spatial distribution at the site, which was interpreted to be controlled by near-surface soil texture.

Volumetric water content

Trang 6

214 Barrier Systems for Environmental Contaminant Containment & Treatmentbetween the two plan views of the radar reflectivity, it is easy to determine wheremoisture changes occur.

Some of the information provided by geophysical methods is indirect, butthe parameters measured can be related to the rock/soil properties needed Forexample, the distribution of electrical conductivity is not a parameter that isdirectly useful in hydrological modeling, but when conductivity is used to obtaininformation on porosity, saturation, pore fluid salinity, and clay content then itbecomes a vital parameter needed for characterization The relationship betweenthe properties measured with geophysics and the hydrologic or mineralogic prop-erties is, in most cases, site-specific To be effective, site characterization requiresclose integration of the geologic, hydrologic, chemical, and geophysical data

4.1.3 G EOPHYSICAL M ETHODS FOR S ITE C HARACTERIZATION

The geophysical methods most directly applicable for characterizing and toring hazardous waste sites can be divided into the following general categories:seismic; electrical and electromagnetic; natural field and magnetic (e.g., gravity,tilt); and remote sensing methods These categories were chosen for the differentproperties that are fundamentally sensed

moni-Well-logging applications are considered here as point measurements and arenot included in the detailed discussions that follow This is not to imply that welllogging should not be included in a geophysical program The opposite is true.Well logging is fundamental to all databases and should be the rule, not theexception

4.1.3.1 Seismic

Seismic methods are used to measure the distribution of elastic wave velocity(compressional and shear) and the attenuation of the different seismic waves inthe ground Seismic velocity depends on many factors, but the primary factorsaffecting seismic measurements are porosity, mechanical compressibility, shearstrength, fracture content, density, fluid saturation, and clay content Some ofthese parameters are directly related to important hydrologic properties and othersare used to map the distribution of soil and rock types The most common use

of seismic methods is mapping interfaces between materials of different seismicvelocities to provide high-resolution images of the locations of lithologic prop-erties and thus infer main flow channels and soil types Cross-hole seismictomography is now used for petroleum reservoir characterization and will beequally important in hazardous waste site characterization

4.1.3.2 Electrical and Electromagnetic

Electrical and electromagnetic methods are used to measure the distribution ofelectrical conductivity and the dielectric constant of the ground Electrical con-ductivity of soils and rocks depends entirely on the conduction paths created by

Trang 7

fluids in the pore spaces and is determined by porosity, saturation, pore fluidsalinity, and clay content In certain cases where the contaminants are ionicsolutions, the electrical conductivity directly maps contaminant distribution(Endres et al., 2000) However, in most cases, the conductivity is used to extrap-olate hydrologic measurements obtained from boreholes The presence of claysthat is so important in fluid flow and chemical absorption models brings about adistinctive frequency-dependent conductivity — the induced polarization effect.This effect is of immense value in monitoring site remediation processes becausemany processes involve injecting materials that profoundly alter this effect (Slaterand Binley, 2003)

A separate electrical property of soils and rock is the streaming potentialeffect, which is but one aspect of a whole class of interactions called coupledflow phenomena Basically, driving forces of temperature gradients, hydraulicpressure gradients, chemical potentials, and voltage gradients produce flows ofheat, fluid, chemicals, and electric current (Slater and Binley, 2003) These flowsare coupled in the ground in the sense that not only does a pressure gradient produce

a fluid flow but it also produces an electrical current flow — the streaming potential.Similarly, temperature gradients drive currents to produce thermoelectriceffects Another cross-coupling term of immense potential in contaminant studies

is electro-osmosis, which is a flow of fluid produced by a voltage gradient Thisphenomenon has been used in geotechnical engineering applications to stabilizeembankments and assist in pile driving It could be used to alter subsurface flowpatterns by directing a particular contaminant plume to an extraction or treatmentregion Because electro-osmosis depends on fluid conductivity, rock permeability,and the configuration of the imposed voltage gradients, the site must be wellcharacterized in fluid conductivity and permeability before the design of a prac-tical system can be implemented

4.1.3.3 Natural Field and Magnetic

Natural field methods consist of gravity, magnetic, and tilt methods High racy measurements of gravity over the surface of the Earth (i.e., microgravitysurveys) yield a measure of the subsurface density distribution, which, in turn,depends on the distribution of porosity, water content, and rock type Boreholegravity measurements yield direct average volume values of density Similarly,high accuracy measurements of magnetic field can be used to infer the distribution

accu-of magnetic minerals, usually magnetite, which, in turn, is related to rock typeand certain sedimentary depositional environments where heavy minerals settleout of fluid flows Tilt measurements have recently been used to measure defor-mation associated with fluid withdrawal and injection By monitoring the rate oftilt or deformation, the rate of fluid movement can be inferred and an averagepermeability for the formation can be determined Tilt and strain methods arelow resolution, but for near-surface application they can be of some use in barriermonitoring If gross changes in the density or geometry of the barrier changes

on the order of a few percent, then these methods may be applicable The drawback

Trang 8

to achieving the necessary resolution is the installation of the gravity meters ortilt meters Careful attention to stability and repeatability of the data must bemaintained in addition to thermal stability and leveling General directions offluid movement, steam injections, or other density changes can also be monitored

In magnetic surveys, the distribution of the magnetization of the earth ismeasured from the surface, but these methods usually lack resolution for detailedsubsurface studies Borehole magnetometers are now being used to supplementmore conventional well-logging tools to search for lithologic changes and chem-icals/minerals that cause magnetization to change

4.1.3.4 Remote Sensing

Remote sensing is defined as the noninvasive observation of natural phenomena

It involves collecting information about an object by detecting differencesbetween the object and the surroundings without being in physical contact withthe object of observation The differences that can be detected between objects

of interest and their background involve shifts in various fields as observationmoves from the background to an object of interest Electromagnetic, acoustic,potential, and radiological are typical fields sampled by remote sensors for objectdetection These types of sensors mounted on spaced-based (satellite) or airborneplatforms can be used to rapidly and noninvasively characterize and monitorfeatures and events on the earth’s surface with broad coverage and high resolution.Space-based or airborne hyperspectral, thermal, radar, and/or radiation sensorscan provide a cost-effective alternative to traditional approaches The spatiallysynoptic look achieved by remote sensing methods can improve the accuracy ofarea interpolations generated by point-sampled data Ideally, the characterizationand monitoring of waste sites and their containment systems would includeremote sensing data, ground-based geophysical measurements, and point-sampleddata These data streams could then be integrated in a geographic informationsystem (GIS) database with ancillary data concerning the barrier construction,geology, watershed hydrology, and climatology of the site

4.2 SPECIFIC METHODS

Although each method has generally been described, there are subsets of eachmethod for specific applications For example, seismic methods can be catego-rized further into active and passive methods, and even further into surface andborehole methods or some combination Specific methods that are most applicable

to environmental remediation needs are described below

4.2.1 S EISMIC M ETHODS

Seismic methods can be divided into passive and active methods Passive methodsinvolve listening to seismic energy being created by stress changes or naturalseismicity such as micro-earthquakes or acoustic emissions near the boreholes

Trang 9

or underground openings Acoustic emissions for purposes described here are ofsecondary use When monitoring a barrier, however, the barrier may emit acousticemissions if it is brittle and possibly failing Monitoring would involve a simpleprocess of emplacing sensors in or near the barrier and monitoring for discreteevents above a certain threshold Active methods involve introducing energy intothe ground with either an impact or controlled swept frequency source andobserving how the seismic waveforms change due to heterogeneity or anisotropy

in the subsurface or barrier Both the direct and reflected arrivals of seismic waves(i.e., travel time and amplitude) can be used for this process More sophisticateduses can involve guided wave energy in the barrier either during emplacement

or for monitoring Seismic reflection methods are used extensively in the leum industry for structural delineation and lithologic definition New and sophis-ticated three-dimensional (3-D) surface and borehole methods have dramaticallyimproved imaging capabilities for the petroleum industry, and can potentially beapplied to remediation with proper instrumentation The utility of seismic tech-niques also depends on the resolution obtainable in a given soil or rock type Forthis reason, this discussion focuses on the seismic methods that have the highestresolution Figure 4.2 shows the typical field configuration of a seismic surfaceand a cross-borehole configuration of a seismic survey These configurations can

petro-be generalized to other techniques such as radar and electrical methods Knowingthe location of the source and receiver, the data can be inverted to derive theproperties of the earth

A typical set-up of a surface geophysical survey (top image of Figure 4.2)consists of a source and receiver on the surface and documentation of the differentarrivals from the source This example is typical for a radar or seismic survey(Hubbard et al., 2003) The bottom figure shows a typical example of a cross-borehole survey with different sources and receivers at different points so that atomographic and/or a reflection image between the boreholes is obtained.The goal of seismic surveys is to describe or map the velocity and attenuation

of seismic waves through the volume of interest In general, this process is referred

to as imaging, although the extent to which a complete or 3-D image can beformed depends on the availability of a suitable distribution of source–receivercombinations and the frequency content of the seismic waves When a crosssection of seismic parameters can be determined, the process is also referred to

as tomography Surface methods depend on sources and receivers distributed onthe surface Combined with sources and/or receivers in boreholes or the barriersthemselves, a true 3-D image can be formed Figure 4.3 shows typical imagesfrom a surface radar survey and a cross-borehole tomographic survey Shown arethe source and receiver pairs and the ray path coverage, very similar to seismicgeometry

Seismic imaging could play an important role in site characterization, formance confirmation, and monitoring tasks It could be used to estimate andextrapolate the extent and shape of soil property distributions that are measuredonly at discrete points with borehole methods It can also be effectively used todetect features not mapped in the exploratory or initial phase of remediation and

per-4040_C004.fm Page 217 Thursday, September 15, 2005 11:50 AM

Trang 10

to monitor changes in properties in the site area from measurements obtainedentirely outside the critical volume The transmission and attenuation of seismicwaves through the subsurface depends on the elastic parameters, which depend

on, among other things, the state of stress and strain, porosity, clay content, grainsize, and fluid saturation As recent research shows, high frequency seismic wavepropagation is sensitive to discontinuities (fractures or joints) in the media (Majer

et al., 1997) Seismic tomography can, therefore, be used to detect changes inthe soil column condition, locate major preexisting and new features, and measureoverall changes in the widths of these features The methods that can be used forthese studies use sources on the surface and detectors either in a borehole [referred

to as vertical seismic profiling (VSP)] or in cross-hole configurations with bothsources and receivers in boreholes VSP techniques are primarily used for eluci-dating subsurface structures and determining seismic velocities of the variousrock and/or soil horizons In addition to the more conventional uses of VSP, the

FIGURE 4.2 A typical set-up of a surface geophysical survey (top) where one places a source and a receiver and records the different arrivals from the source, this example is for a radar or seismic survey (Hubbard et al., 2003) The bottom figure shows a typical example of a cross-borehole survey with different sources and receivers at different points

so that one obtains a tomographic or reflection image between the boreholes.

Locations

of receivers

7 6

8 9

10

11

7 6

Ground wave

Air

Example of tomographic data acquisition geometry

Trang 11

Airborne and Sur

FIGURE 4.3 A typical example of data from a surface reflection survey showing the lithology (left-hand side) The figure on the

right shows typical results from a crosswell tomographic survey correlated with the lithology.

C02-C03aa 400-038

(49 ft N of 406) P4G3

400-207 (60 ft S of 407)

Sands

Gravels

Sands Silt

Trang 12

220 Barrier Systems for Environmental Contaminant Containment & Treatmentuse of three-component VSPs for detecting and characterizing 3-D features hasbecome routine in the oil industry.

In the last several years, the petroleum and gas industry have started to extendthe traditional uses of subsurface imaging from defining static properties tomapping changes in the reservoir conditions and monitoring production Toachieve accurate monitoring for the petroleum industry, new methods using multi-component (P- and S-wave) seismic surveys have been developed to map sub-surface anisotropy and heterogeneities from the surface and between boreholes(Daley et al., 1988a,b; Majer et al., 1988, 1997) The key to using the data,however, is the ability to relate the physical parameters measured using geophys-ical techniques to the parameters of interest to the hydrologist or reservoir engi-neer An example is the relationship of seismic velocity to permeability Frompast work in a variety of complex lithologies (Majer et al., 1988; Majer and Geller,1992; Tura et al., 1992; Tura and Johnson, 1993; Geller and Myer, 1995; Hubbard

et al., 2001; Geller et al., 2000), recent advances in wave propagation theory(e.g., shear wave splitting, fracture stiffness, guided waves, scattering, cross-wellseismic reflection, amplitude and frequency variation with azimuth) must beintegrated into the techniques employed in the petroleum industry and geotech-nical fields to fully utilize the potential of seismic techniques at any scale Theconventional field and analysis techniques [e.g., lower frequency VSP and surfacereflection less than 100 hertz (Hz)] do not detect thin features such as fractures

or steeply dipping or near vertical faults, low velocity zones, zones of small orhigh seismic velocity contrasts, not to mention resolution on the scale to charac-terize process behavior

To a large degree, the information contained in the cross-well/tomographictechniques offers promise of higher resolution, especially if more than first arrivalanalysis is performed, and the elastic solution as well as the acoustic case areincluded (i.e., S and P waves) Frequency effects must be investigated especiallywhen layered complex media exist Using seismic tomography/cross-well tech-niques as a tool for resolving heterogeneity within bedded and fractured structuresremains in development In terms of processing/inversion schemes for high-frequency seismic data, the following four main approaches are possible:

• Conventional and advanced ray and waveform tomography

• Guided/channel waves

• Scattered and reflected energy from voids/high contrast anomalies

• Cross-well/VSP/single well imaging employing azimuthal frequencyand time-varying effects

4.2.1.1 Conventional and Advanced Ray and

Waveform Tomography

Like most inverse problems, the quality of the solution depends directly on thecompleteness and accuracy of available solutions to the forward problem Con-ventional and advanced ray and waveform tomography include such simple

Trang 13

approaches as algebraic reconstruction tomography (ART), simultaneous iterativereconstruction tomography (SIRT), and singular value decomposition (SVD)using first arrival data Given sufficient data quality, these methods may be allthat are necessary Ray and waveform tomography approaches also consider moreadvanced analysis methods such as waveform tomography using exact and Fresnelapproximations and amplitude tomography as well as ensemble averaging tech-niques Conjugate gradient methods that can handle complicated structure andlow velocity/high contrast zones can also be considered

4.2.1.2 Guided/Channel Waves

Guided wave continuity logging is emerging as a new tool in the oil and gasindustry (Krohn, 1992; Nihei et al., 1999), and it likely will evolve into a powerfulmethod for shallow subsurface environmental characterization (Liu et al., 1991).For example, the complex geometry and fracturing of the basalts at the IdahoNational Engineering and Environmental Laboratory (INEEL), Idaho, may sup-port Rayleigh interface waves that propagate along horizontal fractures (Gu et al.,1996), and a new type of channel wave that propagates in the fluid-saturatedrubblized zones on the tops and bottoms of the flows Unlike body waves thatspread in three dimensions, channel waves are confined by the structure into twodimensions, resulting in less geometric spreading Recent results by Nihei et al.(1999) support that channel waves can play an important role in the attenuationmechanism of seismic energy, thus being a diagnostic of fracture properties.Therefore, these waves can be used to probe geologic structures between wellsspaced over substantial distances

4.2.1.3 Scattered and Reflected Energy

The third approach to consider is using scattered energy, particularly for detectingvoids and high-contrast heterogeneity As in the case of guided wave analysis,scattered wave field analysis needs full waveform data as opposed to only arrivaltimes and amplitudes This approach remains in the theoretical stage; practicalapplication, although very powerful, is still not routine The exact solution forscattering elastic waves by a homogeneous spherical obstacle is available andincludes a complete analytical treatment of the problem and the implementation

of the results in stable efficient computer codes (Korneev and Johnson, 1993a,b,c).The solutions were developed for incident P and S waves of arbitrary frequencyand for obstacles having arbitrary properties, including the cases of solid, fluid,and empty obstacles While a sphere may not be an accurate representation ofmany of the underground structures of interest, the solution to the problem ofscattering by a sphere has fairly general applicability A possible route of inves-tigation could be the numerous advances obtained for this type of problem, whichare an extension of the elastic inversion method found in Tura et al (1992) andTura and Johnson (1993) (This last paper contains a list of related work.) Aninvestigation of the reliability of solutions to the inverse scattering problem could

Trang 14

222 Barrier Systems for Environmental Contaminant Containment & Treatmentmake use of the developments in general inverse theory that are found in Vasco(1993) and Vasco et al (1993).

4.2.1.4 Cross-Well/VSP/Single Well Imaging

Last but not least, cross-well/VSP/single well reflection methods are fairly newapproaches, where reflection-processing methods developed for surface reflectiontechniques are used to image reflectors between and possibly below boreholes.These methods are applicable in layered sections with good impedance contrasts.However, if sufficient well coverage exists, a 3-D approach with varying azi-muthal coverage using three-component data can provide useful information onmedia complexity, especially in fractured media

In the cross-well method, the source is activated at various levels in one holeand the receiver is placed at similar levels in the other hole, creating a crossinggrid of ray paths for tomographic inversion Usually in the radar and seismiccases, the first arrival times for each source-receiver pair are used for a tomo-graphic image The two-dimensional cross section between wells is divided intosquare pixels and the velocity (in the seismic case) is estimated in each pixel.The resolution of each pixel is dependent on the ray density in the seismic orradar case (Peterson et al., 1985) and on the frequency content in the electro-magnetic or DC resistivity The data can also be inverted for attenuation In thisanalysis, the amplitude of the first arrival is computed for each trace with sufficientsignal-to-noise ratio The two-dimensional cross section between wells is thendivided into pixels, and each pixel is inverted for amplitude attenuation in decibelsper meter (dB/m) Cross-well seismic as well as radar surveys have been usedfor many years to tomographically image P-wave velocity between wells (e.g.,Mason, 1981; Peterson et al., 1985) More recently, cross-well S waves have alsobeen used to map S-wave velocity (Harris et al., 1995), and both P- and S-wavecross-well reflectivities have been analyzed for structural delineation Until aboutfive years ago, nearly all cross-well seismic tomography was performed in sed-imentary formations important to oil and gas exploitation A seismic source used

in the oil industry but not yet applied to environmental problems is the orbitalvibrator The operating principle of the orbital vibrator is rotation of an eccentricmass in the horizontal plane at increasing speeds, generating a swept frequencysignal of clockwise and counter-clockwise polarizations (Daley and Cox, 1999).The orbital vibrator has a high frequency (now up to 750 Hz) and high power,

is small (3.5 inches in diameter, 18 inches long), and puts out both P-wave andS-wave energy It is easy to deploy and can work in fluid-filled holes The orbitalvibrator propagated P waves up to 100 m in a fractured basalt aquifer (Daley

et al., 1999), significantly farther than a piezoelectric source used at the same site(albeit with a lower frequency band than the piezoelectric source) Tests usingthis source have been successful in acquiring P- and S-wave tomography datausing fluid-coupled hydrophone sensors In the case of direct current (DC) resis-tivity, multiple electrodes are placed in the subsurface either in uncased holes orwith electrodes on the exterior of polyvinyl chloride (PVC) or fiberglass holes

Trang 15

Each electrode acts as a source or receiver, making data collection very efficient.Electromagnetic sources and receivers have been developed and are in routineuse in the oil industry, but not efficiently downsized for the environmental field.The application of cross-well seismic methods to crystalline rock is often amore difficult problem than the application in sedimentary rock The advantage

of seismic imaging is the ability to detect or image features away from theborehole Cross-well seismic imaging in fractured crystalline rock has been used

to define the spatial distribution of velocity and attenuation that is related tofracture zones determined from other borehole techniques (e.g., Vasco et al., 1993;Cao and Greenhalgh, 1997) In fractured media, an important property definingthe rock is fracture anisotropy Anisotropy will also play an important role inimaging waste sites, as will heterogeneity in general Imaging using P and Swaves in borehole seismic studies is not a new idea (Stewart et al., 1981) It isbecoming increasingly apparent, however, that to utilize the full potential of theseismic methods for characterizing fractured media, three-component data should

be acquired In imaging barrier sites and contaminated sites this is rare Crampinnoted the importance of using three-component data in VSP work, particularlyfor fracture detection (Crampin, 1978, 1981, 1984a,b, 1985) These authors andothers have pointed out the phenomenon of shear wave splitting and the anisotropyeffects of horizontal and vertical shear component waves in addition to primaryand secondary wave anisotropy (Leary and Henyey, 1985) In addition to Crampin’stheoretical work on shear wave splitting (1978, 1985), laboratory (Pyrak-Nolte

et al., 1990a,b) and theoretical work (Schoenberg, 1980, 1983) explain shear waveanisotropy in terms of fracture stiffness.The fracture stiffness theory differs fromCrampin’s theory in that at a fracture or a nonwelded interface, the displacementacross the surface is not required to be continuous as a seismic wave passes Theonly solution boundary condition to the wave equation is that the stress mustremain continuous across an interface This displacement discontinuity is taken

to be linearly related to the stress through the stiffness of the discontinuity Theimplication of the fracture stiffness theory is that for very thin discontinuities(e.g., fractures), there can be significant effect on the propagation of a wave.Fracture sites, such as in basalt and other hard rocks, are of interest to barriersand their integrity This implication applies to voids or any feature that representdiscontinuity in the subsurface Usually one thinks of seismic resolution in terms

of wavelength as compared to the thickness and lateral extent of a bed or otherfeature In the stiffness theory, the lateral extent is still important, but if thestiffness of the feature is small enough (i.e., a sand-filled void), the thickness ofthe feature can be much less than the seismic wavelength

In the case of unconsolidated sediments, a coupled solution must be foundthat takes into account the pore fluid (or gas) and matrix interactions The situationbecomes even more complicated when clay content is introduced into the matrix

At this point, multi-phase models must be considered to account for the observedeffects One such approach was tried for acoustic velocities in shale (Minear,1982) where a two-phase model following Kuster and Tokoz (1974) was used.One phase was assumed to be the solid rock matrix and the other phase clay

Trang 16

224 Barrier Systems for Environmental Contaminant Containment & Treatmentinclusions This approach was attempted in order to explain deviations from whatconventional theory would predict Marion (1990) and Marion et al (1992) devel-oped relationships between seismic P-wave velocity and sand-clay mixtures usinglaboratory measurements In this work, it was possible to establish an empiricalrelationship between porosity and sand-clay content Klimentos and McCann(1990) developed empirical relationships between P-wave attenuation, porosity,and permeability in sandstone Partially saturated materials pose a further com-plication Anderson and Hampton (1980a,b) performed considerable work in boththeory and measurement to reach an understanding of seismic wave propagation

in gas-bearing sediments Bedford and Stern (1983) also developed models forwave propagation in sediments.Ito et al (1986) and Mochizuki (1982) alsodeveloped relationships between seismic velocity and attenuation for partiallysaturated material Parra (1991) analyzed elastic wave propagation in stratified fluid-filled media to examine the effect of porosity and permeability.He extended Biot’stheory to include a point force in fluid-filled porous media In a related study,Yamamoto et al (1994) used variations in seismic velocities at different frequencies

to map porosity variations These are just a few empirical and model studies thathave been conducted to relate seismic properties to physical parameters Theseapplications have been almost entirely for the petroleum industry

4.2.1.5 Summary

In summary, seismic methods historically have been used to image subsurfaceelastic properties Only in recent years have researchers focused on relatingseismic attributes to physical/chemical and microbial attributes at the scalesproposed for remediation (Hubbard et al., 1997, 1999; Chen et al., 1999) Seismicdata are well suited for extrapolating measurements obtained from a borehole tothe large-scale volume away from the hole In this application, measurementsobtained from the surface or between holes can be used to assess the continuityand homogeneity of the intervening material Therefore, field and modelingstudies have shown that such features as anisotropy, fluid content, and heteroge-neity have a measurable effect on the propagation of seismic waves It appearspossible to use shear wave anisotropy and 3-D tomography to map the orientation,density, and spacing of these features in the field and to give the hydrologist/res-ervoir engineer useful information on the fluid flow regime A few percent change

in properties produces effects that are easily detectable These seismic methodsare particularly informative if used in conjunction with the electrical methodsdiscussed below

4.2.2 E LECTRICAL AND E LECTROMAGNETIC M ETHODS

Electrical methods seem particularly promising in mapping and monitoring thegroundwater regime of a site because the electrical conductivity of the subsurfacedepends almost entirely on the fluid saturation, salinity (conductivity), and dis-tribution Electrical and electromagnetic methods traditionally have been used to

Trang 17

detect the presence of good electrical conductors (e.g., sulfide ore bodies) ordetermine the electrical layering in groundwater or petroleum exploration Quan-titative interpretation in terms of rock properties or even accurate mapping of thesubsurface distribution of electrical conductivity (imaging) is not as advanced asthat conducted seismologically Only recently have numerical and theoreticalstudies advanced to the point where quantitative imaging complementary toseismic imaging can be expected

The electrical conductivity of rocks and unconsolidated sediments in theupper part of the Earth’s crust is governed by the water content and the nature

of the water paths through the rock Electrical current is carried by ions in thewater; therefore, the bulk resistivity depends on the ionic concentration, ionicmobility, and the saturation and degree of connected pores Conductivity is alsotemperature and pressure dependent because as the temperature increases, ionmobility increases and the pressure affects the apertures of the conduction paths.Most studies on the electrical conductivity of rocks and soils have involvedsedimentary rocks because of their importance in petroleum and groundwaterexploration Archie (1942, 1947) established an empirical relationship betweenthe pore fluid resistivity, Rp (inverse of conductivity); the porosity, P; and theformation resistivity, Rf, that is now referred to as Archie’s Law:

where A and m are constants for a given rock type For a wide range of sedimentaryrocks and some volcanic and intrusive rocks as well, the constant, A, is close tounity and m is close to 2.0

Fluid saturation has a dramatic effect on the conductivity of porous materials(Telford et al., 1976) As water is withdrawn from a saturated rock, the largepores empty first; however, because small water passageways mainly control theresistivity, the bulk resistivity increases slowly The dependence is roughly pro-portional to one over saturation squared As desaturation progresses, criticalsaturation is reached when there is no longer any water to conduct along somepores This breaking of conduction paths leads to a much more rapid increase inresistivity, roughly proportional to one over saturation to the fourth power Thecritical saturation depends on the rock type (the nature of the porosity) and candepend strongly on the role of fast paths that are present Combined with seismicvelocity and attenuation, electrical measurements are valuable for monitoringresaturation progress at a site

An important and little studied aspect of rock and soil conductivity is therole of fast paths on the resultant bulk properties, particularly in barrier monitoringapplications Laboratory studies concentrate on small intact samples that, almost

by definition, do not include open voids or joints Field studies using surfaceresistivity measuring arrays are usually too strongly influenced by the inhomo-geneous nature of the near surface to allow any distinction between voids andpore porosity of a particular rock unit With the increased measurement accuracyand resolution provided by subsurface techniques and the interest in monitoring

Trang 18

226 Barrier Systems for Environmental Contaminant Containment & Treatmenttime changes in resistivity, it is now possible to investigate more closely the role

of porosity on the electrical conductivity of large masses It is well known thathydraulic conductivity is strongly influenced by the mean width, orientation, andspatial distribution of the fluid paths Also, as noted in the preceding section,seismic velocities are strongly affected by discontinuities However, expressionsfor the electrical conductivity of such material and taking advantage of thisvaluable physical property for characterizing and monitoring large subsurfacevolumes of soil remain to be developed

Channeling plays an important role in rock resistivity and is practicallydemonstrated in the work by Brace and Orange (1968a,b) Their work on theeffects of confining pressure on the resistivity of a water-saturated granite showedthat, at low pressures, the resistivity increased as the confining pressure increased.They attributed this effect to the closure of fracture porosity A resistivity increase

of a factor of 10 as the pressure increases could easily be explained by thedisappearance of only 0.1% fracture porosity in a granite of 1.0% pore porosity.The electrical conductivity of the ground can be measured in two ways Inthe first, referred to as the DC resistivity method, current is injected into theground through pairs of electrodes and the resulting voltage drops are measured

in the vicinity with other pairs of electrodes Any or all of the electrodes can beplaced in the subsurface, although traditionally surface arrays have been employed.Electrical resistivity tomography (ERT) uses electrodes in the subsurface to mea-sure resistivity between the boreholes (Daily and Ramirez, 2000) Measurements

of voltage and current for different electrode geometries are then used to inferthe subsurface distribution of conductivity These methods are indirect, but ideallysuited to measure the properties of a region for which it is impossible to gaindirect access The resulting interpretation of the conductivity distribution is notunique nor does it provide high resolution of subsurface features In many appli-cations, this latter property is an advantage because the measurements yield bulkaverage values of the conductivity that often include features that are not included

in hand samples or borehole logging measurements

The electrical conductivity can also be measured inductively Instead ofinjecting a DC current into the ground, currents can be induced to flow by achanging magnetic field The source of the changing magnetic field could be aloop of wire carrying alternating current or a long grounded wire carrying alter-nating current rather than direct current or the Earth’s natural electromagneticfield The currents induced in the ground are measured either by detecting themagnetic fields they produce or measuring the voltage drops in pairs of electrodes.Sources and receivers can be on the surface, below the ground, or a combination

of both In these inductive or electromagnetic methods, the interpretation dependsboth on transmitter-receiver geometry and frequency used In principle, the inter-pretation should be more definitive than with DC resistivity methods Rigorousconfirmation of this statement in heterogeneous media awaits the development

of generalized inversion techniques for electromagnetic methods

Electromagnetic methods offer some proven advantages over DC methods.Measurements can be obtained without contacting the ground; measurements are

Trang 19

insensitive to high resistivity zones; the investigation depth can be controlled by

the frequency of operation so that large transmitter-receiver spacings are not

required; and, because of the transmitter source field fall-off, the methods are not

sensitive to conductivity inhomogeneities far from the zone of interest The

resolution of subsurface features with electromagnetic methods is limited because

the frequencies that are low enough to penetrate to the desired depth cannot have

a wavelength short enough to define structural features The problem is

com-pounded by surface layers that are invariably conductive, highly variable in

thickness, and often act like shields to the subsurface To overcome these problems,

promising borehole electromagnetic methods exist Pulsed borehole radar is an

example of an electromagnetic technique that uses high frequencies (Hubbard

and Rubin, 2000) Radar is becoming prevalent in a variety of environmental

applications due to its ease of use and sometimes straightforward interpretation

(Grote et al., 2003) If the ground conductivity is sufficiently low, megahertz radar

waves can penetrate up to 100 m and can respond to dielectric contrasts within

the rock mass as well as conductivity anomalies Radar has been used successfully

at some toxic waste sites to map buried objects and determine fine-scale structural

features and map fluid flow in the vadose zone at a submeter scale (Hubbard and

Rubin, 2000) In typical soils the range of radar can be from a few meters (using

500 MHz) to tens of meters (using 50 to 100 Hz) (Hubbard and Rubin, 2000)

In more conductive rocks, the frequency of the electromagnetic fields must

be reduced to achieve significant penetration Then, the resolution decreases as

the fields become diffusive in nature The traditional low-frequency

implemen-tation of electromagnetic methods (less than a few kilohertz) for ore prospecting

relies on quasi-static magnetic induction theory and basically ignores the wave

propagation properties of the fields In subsurface applications, especially in

single- and cross-hole modes, there are exciting possibilities for electromagnetic

methods in the frequency band between the prospecting and radar frequencies

(i.e., the mid-frequency band)

4.2.3 N ATURAL F IELD AND M AGNETIC M ETHODS

Dramatic developments have occurred in natural electromagnetic field methods,

particularly magnetotellurics Although magnetotellurics may not have the resolution

for fine-scale studies, it is mentioned here for completeness In magnetotellurics,

the impedance of the ground is measured as a function of frequency This impedance

function is then interpreted in terms of a model of the earth Traditionally,

magnetotellurics has been plagued with problems in data quality and

interpreta-tion when the simple layered models used are inadequate The data quality

problem has been solved by using the remote reference method developed by

Goubau et al (1978) and improved instrumentation (sensors and high dynamic

range acquisition systems now permit high-accuracy surveys that were previously

not possible) Field evidence shows data errors of less than 1% in some frequency

bands (Nichols et al., 1985) Interpretation has been a problem because the

imped-ance was not sampled at adequate intervals on the surface The electric fields

Trang 20

change rapidly in response to near-surface resistivity variations and bias the

impedances that, in effect, mask the deep structure that is sought This bias can

be treated by conducting dense station sampling using larger lines for the electric

field measurements or, preferably, both Many of these issues are being overcome

with advanced computational methods and joint inversion of data (Gasperikova

et al., 2003) In principle, the electric fields could be measured over a grid on the

surface, with magnetic fields measured at the grid nodes and the conductivity

distribution recovered accurately and unambiguously Equipment is now available

for such surveys but has not yet been tested

4.2.4 A IRBORNE G EOPHYSICAL M ETHODS

Airborne geophysical methods hold a middle ground between the ground-based

geophysical methods described above and conventional remote sensing methods

spectral, hyperspectral, thermal, or radar systems, which are typically obtained

by satellites or aircraft at several hundred meters altitude Airborne geophysical

data include magnetic, electromagnetic, and GPR data, typically acquired at

sensor altitudes ranging from 50 m to about 1 m Conventional methods and

applications for airborne geophysics are described by the National Research

Council (1995) These airborne magnetic and electromagnetic systems have been

used to image United States Department of Energy (USDOE) waste areas and

caps (Doll et al., 2000) Recently, airborne magnetic and electromagnetic systems

have been developed in which the sensors are housed in booms that are mounted

directly to the helicopter This technique allows airborne data to be acquired at

altitudes as low as 1 to 2 m above ground level (AGL), where topography, cultural

features, and vegetation permit The boom-mounted systems have been used to

detect and map unexploded ordnance and other metallic objects and can

success-fully map these objects with < 0.2 m accuracy when the unexploded shells are

as small as 3 to 5 kg (i.e., the size of a small soup can)

The Oak Ridge Airborne Geophysical System-Arrowhead

(ORAGS-Arrowhead) is a production-level magnetometer system that is typically operated

at altitudes of 1.5 to 3.0 m AGL, depending on site conditions (Figure 4.4) The

sidebooms and foreboom house a total of eight cesium vapor magnetometers at

a nominal spacing of 1.7 m, with two magnetometers each at the ends of the side

booms and four spaced evenly across the V-shaped foreboom The sensor

posi-tioning is designed to minimize noise from the helicopter rotor and other sources

while maintaining a weight distribution that optimizes flight performance and,

above all, safety All data are recorded on a personal computer based console that

samples the magnetometers and keys analog inputs (e.g., fluxgate) at 1.2 kHz

and records laser-derived altitude and global positioning system (GPS) position

at the full output rates of those devices The magnetometer data are downsampled,

typically to 120 Hz, and the other data are interpolated to the same sample

frequency as the downsampled magnetometer data Navigation is directed by an

Agnav RT-DGPS system with Racal satellite real-time correction Aircraft position

described in Section 4.2.5 Remote sensing is generally used to refer to

Trang 21

multi-Airborne and Surface Geophysical Method Verification 229

is recorded on the system console and updated by post-processing with a DGPS

base station to provide an accuracy of 0.2 m or better Under optimal flight

conditions, the system acquires data over a 12 m swath with a 1.75-m sensor

spacing at a flight height of 1.5 m AGL An Ashtech ADU-2 GPS-based system

is used to monitor the altitude of the system to provide accurate sensor positioning

The ORAGS systems are typically operated at an air speed of 50 knots, allowing

full coverage acquisition of a rate of about 50 to 70 acres per hour under favorable

conditions

Figure 4.5 shows an analytic signal map for a site in Maryland where previous

ground-based geophysical surveys were conducted The airborne data set

delin-eated a spider web of underground pipes that was overlooked during

ground-based survey preparation and interpretation Such a network of conductors almost

certainly had a negative impact on the processing and interpretation of the ground

surveys The conductor network would not have been detected with a conventional

airborne survey at conventional altitudes

Another ORAGS system is the ORAGS-Hammerhead system, which is useful

for defining the boundaries and infrastructure of landfill sites that contain ferrous

waste materials or containers This system provides considerably more detail than

a conventional towed-bird system Oak Ridge National Laboratory (ORNL) and

partners have recently completed a successful demonstration of an airborne

time-domain electromagnetic prototype system, the ORAGS-TEM (Transient Electrical

Methods) system, as an electromagnetic complement to the ORAGS-Hammerhead

system A photograph of the system is shown in Figure 4.6, and data acquired

FIGURE 4.4 The ORAGS-Arrowhead total field magnetometer system in operation.

Trang 22

FIGURE 4.5 Analytic signal map for a site in Maryland showing anomalies associated

with a network of piping that had been overlooked in more localized ground-based surveys.

FIGURE 4.6 ORAGS-TEM system in transit near the Black Hills, South Dakota.

23.7 22.4 21.2 19.9 18.6 17.3 16.0 14.7 13.5 12.2 10.9 9.6 8.3 7.1 5.8 4.5 3.2 1.9 0.6 nT/m

Meters

200 300 N

Trang 23

with the system are compared with data from the ORAGS-Arrowhead magnetic

system in Figure 4.7 The system records the entire decay curve for each

trans-mission, as does the EM-643 ground-based system It is possible that the TEM

system can be adapted to provide resistivities through an appropriate calibration

procedure, but this possibility is only beginning to be investigated In its current

form, the TEM system responds to nonferrous metallic objects, ferrous materials,

and some nonmetallic objects Therefore, it is an appropriate tool for mapping

materials in waste sites If the TEM system can be successfully demonstrated as

a tool for measuring resistivities, it could be suitable for time-lapse measurements

of moisture or other resistivity-dependent effects that should be monitored at

landfills or similar areas

4.2.5 S TATE - OF - THE -P RACTICE R EMOTE S ENSING M ETHODS

Despite the fact that geophysics has been used successfully for many years in

mineral, petroleum, and geothermal exploration, it has not been used effectively

FIGURE 4.7 Comparison of (a) ORAGS-TEM measurements and (b) an analytic signal

map derived from ORAGS-Arrowhead magnetic measurements for a bombing target in

South Dakota TEM represent the first-time gate only, and data were acquired at 3 m

nominal flight line spacing and 1.5–2 m altitude Magnetometer results used the 8-sensor

magnetometer system at the same altitude and 12 m flight-line spacing The response of

both systems to an east-trending barbed wire fence is seen across the center of the diagrams.

The individual anomalies are associated with M-38 practice bombs, or their fragments.

These are sand- or cement-filled devices with a mass of 10–15 kg when intact Horizontal

scale is in meters (See color version insert of this figure.)

9.0 7.9 6.9 5.9 4.9 3.8 2.8 1.8 0.8

Trang 24

in remediation situations By and large, the examples of geophysics applied tothese problems that have appeared in the literature are very basic and display alevel of application that characterized geophysics 15 to 20 years ago Descriptions

of the state of the practice of remote sensing methods are described in thesubsections below

4.2.5.1 Aerial Photography

Aerial photography is a useful tool due to its well-understood technology andthe many historical records of sites that contain aerial photos Traditional aerialphotos provide high spatial resolution imagery using black and white, naturalcolor, or color infrared (CIR) film Black and white film can still be useful incases where high contrast differences enhance detection or location of a targetarea Natural color aerial photos are often used to overlay data layers when using

a GIS system CIR imagery captures a scene in the near infrared (NIR) bands ofthe electromagnetic spectrum The color imagery produced consists of false colorimages in which the colors serve to separate scene elements that reflect NIRradiant energy differently CIR photos have been used to detect vegetation stress,which can be important in identifying plants that have become damaged due toleachate exposure Moreover, CIR aerial photos can detect waterlogged areas andseparate out conifers from certain deciduous species (Jensen, 1968) For eitherfilm or digital camera technology, a current aerial photo allows the investigator

a synoptic view of site geographic/environmental features as well as its culturalaspects A significant aspect of aerial photos is the historical record that aerialphotos represent For example, many waste sites have had poor or little docu-mentation on their location or contents

Aerial photographs can also be used to construct digital elevation models(DEMs) Often stereo aerial photo capability is part of a standard collectionprocedure by many aerial photo firms Pairs of images are acquired with 60%overlap, which allows for stereo pair generation Standard photogrammetry isemployed to convert the information in the stereo pairs to digital contour mapsand/or DEMs For each picture element (or pixel) that comprises an image, anelevation datum can be assigned to it The DEM can be used to set up a baselinefor a waste site cap Later, DEMs generated can be used to determine relativesubsidence of the cap with respect to the baseline imagery, which could be anearly sign of cap compromise

4.2.5.2 Multi-Spectral Scanners

Multi-spectral scanners (MSS) are commonly used sensors for collecting imagery

in diverse application areas The National Aeronautics and Space Administration(NASA) Landsat satellite program established in the 1970s used filmless multi-spectral imagers As the Landsat program evolved, improved technologies wereused to enhance the quality of the imagery produced by the sensor In the 1980s

Trang 25

and 1990s, the number of spectral channels grew for both airborne and borne sensors while retaining good spatial resolution

satellite-MSS sensors have decided advantages over photographic systems For ple, film technology limits the spectral range that can be covered by film-basedphotographic systems, which are more difficult to calibrate to radiometric unitsthan is digital data The useable spectral range for film is about 300 nm to 900 nm,with wide spectral bandwidths Where photographic systems generally need touse separate optical systems to break out the different spectral bands, MSSsystems can use the same optical train to record data from each optical band.Finally, if the aerial photo film will be analyzed in an electronic computer, itmust be digitized, i.e., scanned by an aerial photo film scanner and saved asdigital number data The process of scanning not only degrades the spatial accu-racy inherent in the film, it is an extra step that is not needed with digital data(Lillesand and Kiefer, 1994)

exam-4.2.5.3 Thermal Scanners

Thermal scanners record radiant emissions that span a range of thermal infrared(TIR) wavelengths The TIR scanner integrates all of the emissions over thesewavelengths and composes an image of them using detectors specifically devel-oped for use in the TIR region of the spectrum Often, the wavelengths integratedover the span range from 8 µm to 12 or 14 µm due to the atmospheric transmissionwindow for these wavelengths The blue/green colors show cooler areas, whilethe orange/red show warmer areas The temperature regime of a landscape variesnaturally with the amount of solar insulation That is, solar input to a landscapedifferentially heats the constituent materials present (Elachi, 1987) Depending

on the application, the proper interpretation of thermal imagery must considerdiurnal heating effects Often pre-dawn imagery is requested because it tends tominimize the thermal shadow effects and differences in the slope of the landeffects in the imagery The optimum time for data collection depends on thespecific application and target characteristics

4.2.6 S TATE - OF - THE -A RT R EMOTE S ENSING T ECHNOLOGIES

Remote sensing systems, techniques, and practice are developing at a rapid pace.The use of hyperspectral sensors, light detection and ranging (LIDAR) topo-graphic systems, LIDAR fluorescence, satellite and airborne radar sensors, andsensor fusion approaches are rapidly moving from the research arena to applica-tions However, the exciting new developments in geophysics, especially newmethods of imaging the subsurface properties, have not been fully applied in wastestudies The subsections below reiterate the point that, in addition to monitoringthe emplacement, effectiveness, and performance of barriers, geophysics should

be used in a total system performance mode to monitor the total fluid matrixsystem that includes not only the barrier but also the zone being contained

Trang 26

4.2.6.1 Hyperspectral Imaging Sensors

A hyperspectral imaging spectrometer (HIS) or hyperspectral scanner acquires aseries of images of the same scene in a range of colors (i.e., wavelengths orspectral channels) similar to that of a MSS sensor The primary difference is inthe narrowness of the bandwidths of the spectral channels and their number Ahyperspectral scanner attempts to perform laboratory quality spectroscopy of alandscape from an aircraft Hence, the basic MSS technology has been enhancedand extended to handle and process the extracted spectral channels and concom-itant data load HIS imagery is often thought of as forming a cube of data(Lillesand and Kiefer, 1994) because of the many bands of data forming a stack

of images — one image of the same scene for each band For large areas imagedand, in some cases, 200+ bands of spectral data, the data load can become onerous.Nonetheless, the spectral information concerning the scene can be invaluable indetermining or classifying unknowns in the landscape, much as spectroscopy isused to determine unknown compounds in a chemical laboratory Figure 4.8shows a HIS cube

Currently, the premier HIS instrument is the NASA Jet Propulsion LaboratoryAdvanced Infra-Red Imaging Spectrometer (AVIRIS) This system is flown typ-ically on an ER-2 aircraft at an altitude of 20 km, producing a ground cell size

or pixel size of 20 m Alternatively, the AVIRIS can fly on a slower Twin Otterplatform at 2 km and produce 2- to 3-m pixels The AVIRIS possesses 224contiguous spectral channels that span 0.4 to 2.45 µm These spectral channelsare about 10 nm wide Figure 4.9 displays an AVIRIS image of Summitville,Colorado The Summitville Mine is shown in the picture as is the mapping of

FIGURE 4.8 HIS (AVIRIS) image cube of Moffett Field, California (See color version insert of this figure.)

Trang 27

iron-bearing minerals made possible by the detailed spectroscopic nature of theAVIRIS imagery

4.2.6.2 LIDAR Systems

LIDAR remote sensing systems are active sensors That is, LIDAR sensorsprovide their own illumination rather than relying on the sun for illumination.The basic principle of an airborne topographic LIDAR is time of flight of a round

FIGURE 4.9 AVIRIS HIS mapping of Summitville, CO, area (See color version insert

Alamosa River

Reynolds Tunnel

Sludge Fe–hydroxide

Ferrihydrite

Hematite not mapped Goethite

Bitter Creek

Wightman Fork

1 KM Crospy Mountain

Trang 28

236 Barrier Systems for Environmental Contaminant Containment & Treatmenttrip of a light pulse to the ground and its return back to a LIDAR receiver Becausethe airframe position can be known to centimeter-level accuracies, by extension,the point of the ground can be literally surveyed into the system Thus, as theairframe flies, a dense set of laser pulses is scanned perpendicular to the direction

of airframe motion and height and geographic position information for eachground return pulse are recorded (Measures, 1984) Another aspect of LIDARremote sensors is the recording of the intensity of the ground returning pulse aswell as its travel time to the target (time of flight) Because the reflectance ofvegetation and earth materials vary from one another at the wavelength of theLIDAR illumination, reflectance can be used to aid in separating out or classifying

a terrain

4.2.6.3 Laser-Induced Fluorescence (LIF)

Lasers can be used in remote sensing systems to invoke a fluorescence response

in different materials termed laser-induced fluorescence (LIF) The evolvedfluorescence is detected by a receiver and used to target or, in some cases, imagethe irradiated object Potential applications include pollutant/contaminant studiesand vegetation stress studies LIF involves the use of laser pulses at a specifiedwavelength to pump target molecules to excited states, followed by de-excitationand concomitant release of radiation or, in this case, fluorescence at longerwavelengths (Goulas et al., 1997).For example, uranium (in the form of uranyloxide) can be stimulated by a laser to produce a fluorescent spectrum Thus, LIFcan be used to detect uranium-bearing leachates or contamination hot spots on acap or in the surrounding cap environment

When laser light energy is absorbed by the chloroplast (i.e., the plant organellethat houses chlorophyll), the light energy excites an electron from a ground state

to a first excited state Plants that exhibit stress due to environmental factorsexhibit a decrease in the efficiency of photosynthesis (Bongi et al., 1994; Moya

et al., 1992).It has been shown that when photosynthesis is reduced, the amount

of heat energy increases by a factor of about five and the amount of chlorophyllfluorescence by a factor of six (Noonan, 1998).However, it is important to realizethat many factors can cause stress in plants Moreover, chlorophyll concentrationshave been known to alter because of shifts in lighting or season So, althoughstress can be indicated by LIF, the cause of the stress must be resolved bysupplementary information

Laser-induced fluorescence imaging (LIFI) is a project operated by theUSDOE that uses a camera to detect the fluorescence induced by a co-locatedlaser transmitter from selected targets (DiBenedetto et al., 1995).Contaminant-induced plant stress can be imaged and mapped by the LIFI instrument, as canuranyl-bearing soils or leachates It is possible to detect heavy metals and volatileorganic compounds that are often associated with landfills A handheld version

of the LIFI system was field tested at the ORNL K-25 site and was able to detecturanium during decontamination and decommissioning activities and on selected

Trang 29

surfaces Also, the LIFI was able to detect chromium-induced stress in plants atthe site A helicopter version of the LIFI is planned

4.2.6.4 Radar Systems

Radar systems are active sensors like LIDAR systems, but radar uses microwavesrather than light waves to probe areas of interest, which is an advantage becausemicrowaves penetrate clouds and rain Like LIDAR systems, radar systems usepulse transmissions of microwaves and record round-trip flight times from the radartransmitter to the target and back Generally, microwaves penetrate more deeplyinto vegetation than very near infrared (VNIR) wavelengths Penetration depthdepends on the actual microwave wavelength and the moisture content of thevegetation Radar returns are processed not only for their range information butalso for the intensity of their scatter and volumetric returns In VNIR wavelengths,scattering depends on the atomic/molecular makeup of the material irradiated.However, microwave scattering intensity depends on the following: (1) largerscale (on the order of centimeters) surface roughness features, (2) the dielectric

of the landscape material (which can be a function of moisture present), (3) thepolarization (horizontal or vertical electric field orientation) of the radar trans-mission, and (4) the angle the incident wave makes with the landscape element.Volumetric return refers to the total return from large-scale landscape elementslike a forest canopy Thus, total radar intensity return is the sum of the surface andvolumetric returns Hence, tonal values in a radar image are related to the intensity

of the radar return Specifically, the greater the backscatter values, the brighterthe tonal value of a landscape element (Toomay, 1982)

Two different radar technologies are often employed when collecting remotesensing data: side-looking airborne radar (SLAR) and synthetic aperture radar(SAR) SLAR represents the first imaging radar used SLARs are often referred

to as real aperture radars because the along-track resolution depends on the size

of the physical antenna of the radar system However, SLARs are inherentlylimited in their resolution by the antenna size that an airframe can support SARtechnology overcomes this limitation Moreover, a SAR system, due to the virtualantenna, can work at longer wavelengths than a SLAR system The greater range

of wavelengths available to a SAR system increases its flexibility and value forapplications

Typical satellite radar systems are the Japanese Earth Resources Satellite(JERS), European Resources Satellite (ERS-1 and ERS-2), and the CanadianRadarsat Table 4.2 includes a brief summary of system specifications of interest.Note the evolution in the spatial resolution capability of these sensors The 3-mresolution of Radarsat-2 means that the variety of applications for which the datacan be used is significantly increased

Radar imagery can contribute significantly to site monitoring Not only canDEMs be constructed from data, radar backscatter imagery can be used to lookthrough vegetation to reveal the ground surface beneath Texture and backscatter

Trang 30

changes indicate moisture shifts across the cap and surrounding areas, whichcould be the harbinger of the onset of closure cap compromise or a true breach.Moreover, DEM and geomorphology differences (slope analysis) from some ear-lier baseline data add an important information layer to the site assessment Radarimagery has been used to detect plant biomass and perform plant classifications(e.g., Ranson and Sun, 1994; Rignot et al., 1994; Dobson et al., 1998) Addition-ally, then, radar imagery can be used to detect changes in the composition of plantcommunities or plant biomass shifts that could be due to contaminant exposure

4.2.6.5 Fused Sensor Systems/Data Streams

Fused sensor approaches include sensor systems that are flown on the sameplatform over a target area, sensors on different platforms used simultaneouslyover a site, and sensors on different platforms used at different times over a site.The latter case often is the rule for GIS data layer accumulation for a given site It

is clear that data provided by multiple sensors, whether performing a simultaneousdata collect or not, are more valuable than data provided by a single sensor.Further, ancillary data concerning the construction, geology, watershed, and cli-matology of the site provide crucial data layers for input to the site GIS database.This is the systems approach to interrogating and monitoring waste sites Theconfluence of remote sensing/geophysical and ancillary data streams inform oneanother and end users of conditions present at a given target area

Examples of investigators using multiple data streams to successfully acterize a site include Vincent (1995), who used both aerial photos and MSS datafor assessing waste sites Van Eeckhout et al (1996) used aerial photos (somehistorical) and CIR photos to assess three landfill sites in New Mexico.Well et al.(1995) used both TIR and GPR with good results to investigate hazardous wastesites Smyre et al (1998) used aerial photos (some historical), CIR, TIR, MSS,magnetic, electromagnetic, and gamma ray data collection to assess an ORNL site

Trang 31

Airborne and Surface Geophysical Method Verification 239 4.3 PRB S

In this section, the potential application of geophysical methods for assisting inthe evaluation and monitoring of a PRB is discussed A PRB presents an enticinggeophysical target, although reported geophysical applications are currently few.The granular reactive iron media used to construct a PRB has unique geophysicalproperties Electronic conduction in iron results in high electrical conductivityrelative to near-surface geological formations The electrical conductivity (inverse

of resistivity) of iron is 1 × 107 siemens per meter (S/m) (Carmichael, 1989),whereas that of near-surface earth materials is typically less than 1 S/m.Iron hashigh magnetic susceptibility and will locally perturb the earth’s magnetic field.Endres et al (2000) investigated the electrical and magnetic properties of granularreactive iron mixed with sand.Laboratory measurements, reproduced in Figure4.10a and Figure 4.10b, illustrate the strong dependence of electrical conductivityand magnetic susceptibility on the volume of granular iron Note that the con-ductivity of the granular iron does not approach the reference value for theelectrical conductivity of iron, presumably due to the absence of continuouselectronic conduction paths in the granular media used in this study PRBemplacement in the subsurface also creates an interface between iron and thesediment at which charge transfer must switch between electronic and electrolyticconduction, making the PRB an intriguing target for the induced polarizationgeophysical method that is sensitive to the electrochemistry of a metal–fluidinterface Iron also has seismic properties distinctly different from most near-surface earth materials The acoustic velocity of iron in solid form is 5900 m/s(McIntire, 1991), whereas it is typically less than 1500 m/s in near-surfaceunconsolidated sediments in a diffuse form Seismic methods can also then assist

in PRB investigations through characterization if it does present a seismic aly or through general characterization structure

anom-4.3.1 R EQUIREMENTS , S ITE C HARACTERIZATION , D ESIGN

In this section, the utility of geophysical methods to PRBs is considered in respect

to the following four objectives: site characterization, PRB construction tion, short-term monitoring of PRB performance, and long-term monitoring ofPRB performance Case studies that illustrate the current state of the art ofgeophysical methods in PRB evaluation are subsequently presented Finally,future directions and recommendations for research are presented The potentialapplication of geophysical methods emerges at the design, installation/verifica-tion, and monitoring stages of the PRB life span PRB assessment effectiveness

verifica-is required over both the short term after PRB construction and over the term design life Short- and long-term monitoring issues are thus treated sepa-rately in the subsections below

long-4040_C004.fm Page 239 Thursday, September 15, 2005 11:50 AM

Trang 32

FIGURE 4.10 a (a) Electrical conductivity of sand-granulated iron mixtures with varying

iron content of solids Mixtures were saturated with 0.01 M KCl solution (b) Magnetic susceptibility of iron mixtures as a function of total volumetric content (Reproduced from

Endres, A.L et al., 2000 Proceedings of the Seventh International Symposium on Borehole

Geophysics for Minerals, Geotechnical and Groundwater Applications, Mineral and

Geo-technical Logging Society, Golden, CO, pp 1–8 With permission.) b (a) Schematic of

electrical charge transfer mechanisms in earth containing metal minerals (b) Simple circuit

model for this system: R nm represents the resistance exerted by the conduction path

associated with free electrolyte, R m represents the resistance exerted by conduction across

a metal–fluid pathway (electronic and ohmic), and W is a Warburg impedance that depends

Cation Anion Non-metallic path (nm)

Metallic path (m)

Interlayer (polarization)

Granular iron (electronic conduction) e

Rnm

Rm (iω X −c) = W 4040_C004.fm Page 240 Thursday, September 15, 2005 11:50 AM

Trang 33

are required to determine PRB design thickness Geophysical methods are tinely used to qualitatively characterize variability in subsurface lithology Low-permeability clay formations are distinguishable from hydraulically conductive

rou-sand units Although direct quantification of K from geophysical methods is

currently unachievable, recent advances illustrate the value of geophysical

imag-ing for providimag-ing spatially extensive representations of K variation (for review

see Hubbard and Rubin, 2000).This information can aid in the identification oflithologic variability at the immediate PRB installation site and define preferentialflow zones that could complicate contaminant plume transport upgradient of thebarrier Such techniques require ground-truth verification from whatever boreholerecords are available at the study site Effective PRB performance necessitatesaccurate barrier emplacement in the immediate path of the contaminant plumeunder remediation This implies that the plume geometry be well characterized.Direct geophysical detection of chlorinated solvents and heavy metals at typicalsite concentrations is unlikely Geophysical monitoring of the transport of tracerevolution injected upgradient of a proposed PRB installation can determinewhether the barrier is well placed to capture the plume Tracking electricallyconductive tracers using electrical resistivity, electromagnetic, and GPR methodshas been applied to characterize vadose zone transport (Daily et al., 1992, Hub-bard et al., 2002) and groundwater flow (White, 1988, 1994).These methods aredeployable using surface and/or borehole instrumentation Borehole methods areexpensive but enhance resolution of tracer transport at depth The results ofgeophysical tracer tests could assist in designing geochemical monitoring wellnetworks by identifying preferential flow paths and likely flow rates

4.3.1.2 PRB Construction Verification

The granular reactive iron used in PRB construction profoundly affects the trical and magnetic properties of the subsurface relative to the pre-installationcondition Thus, geophysical methods have a high potential for PRB constructionverification Geophysical imaging of subsurface conductivity structure usingresistivity, electromagnetic, or GPR techniques offers the possibility of definingthe continuity and uniformity (thickness) of the wall, as well as detecting thelocation of flaws in barrier construction Cross-borehole electrical resistivitytomography was used to examine the subsurface distribution of granular ironinstalled at the USDOE Kansas City facility in Missouri (Slater and Binley, 2003).Joesten et al (2001) used cross-borehole GPR to image differences in radar waveattenuation amplitude caused by PRB construction at the Massachusetts MilitaryReservation in Massachusetts.Endres et al (2000) showed that downhole elec-tromagnetic tools are sensitive to the presence of granular iron injected into aformation and offer a potential approach to PRB construction verification Tomo-graphic electromagnetic and seismic methods are also potentially valuable methods

Trang 34

242 Barrier Systems for Environmental Contaminant Containment & Treatmentwhen an image can be compared with an image from that of a pre-existingcondition In the case of the PRB, differences in the electrical structure caused

by the emplacement of reactive iron are of interest, necessitating geophysicaldata collection prior to PRB construction Ideally, boreholes for geophysical dataacquisition would be drilled prior to any subsurface disturbance to permit acqui-sition of a representative background data set The geometry of a typical PRB iswell suited to cross-borehole geophysical imaging Instrumentation can be placed

in boreholes drilled immediately upgradient and downgradient of the barrier,providing a two-dimensional cross-sectional image of the barrier wall (Joesten

et al., 2001; Slater and Binley, 2003).A closely spaced nest of boreholes permits3-D imaging of the barrier installation (Slater and Binley, 2003)

4.3.1.3 Short-Term Monitoring

Short-term PRB monitoring primarily focuses on the wall efficiency to degradeand remove contaminants The relatively low contaminant concentration typicallyencountered at a PRB installation site is unlikely to be detectable with geophysicalmethods Short-term monitoring is also concerned with possible disruption of thenatural flow regime due to PRB emplacement Most critical is that plume transportfollowing PRB construction is consistent with that predicted from the site char-acterization phase A geophysical tracer test could be an effective noninvasivemethod for assessing plume transport immediately after construction The use ofelectrical resistivity, electromagnetic, or GPR to track an electrically conductivetracer appears to be a promising technology and could be used in decision-makingregarding final placement of geochemical monitoring wells required for long-term performance evaluation

4.3.1.4 Long-Term Monitoring

The long-term performance of PRBs is highly uncertain but operational life, itsperiod of effectiveness, is expected by the user to exceed ten years Monitoringstrategies are required to determine deterioration in barrier performance as reac-tive iron is oxidized during hydrocarbon degradation The exact mechanism ofdegradation of chlorinated compounds is not fully understood, and a variety ofpathways are likely involved (Gavaskar et al., 1998).A fundamental aspect ofPRB performance is that degradation of chlorinated organics is a surface phe-nomenon and the available specific surface area of the reactive medium governsthe rate (Gavaskar et al., 1998) Clogging of the barrier and the resulting perme-ability reduction are also current concerns relating to reduction in PRB perfor-mance and can be significant issues Clogging at the influent end of the PRBcould potentially result from formation of iron precipitates under highly oxygenatedconditions (Gavaskar et al., 1998).In addition, deposition of inorganic suspendedsediments on the granular iron can also reduce permeability and performance.The presence of granular iron modifies the subsurface electrical propertiesdue to the following charge transfer mechanisms: electronic conduction in the

Trang 35

metal and polarization of charges at the interface between a metal and the filling electrolyte Figure 4.10a is a simple conceptual illustration of the keycharge transfer mechanisms in a medium containing metal particles Figure 4.10b

pore-is a circuit analogue of thpore-is system A frequency-dependent interfacial (Warburg)

impedance (W) is often used to simulate the electrical response of a metal–fluid

interface (e.g., Pelton et al., 1978).The magnitude of this interfacial impedance

is measured with the induced polarization method The frequency dependence ofthis impedance is also determined when spectral (multi-frequency) induced polar-ization measurements are made The chemistry of the metal–fluid interface exerts

a strong control on the induced polarization response (Olhoeft, 1985).Oxidation

of the granular iron surface as a result of continued chlorinated solvent tion might modify the induced polarization response of a PRB Precipitation ontothe granular iron does reduce the surface area of the metal–fluid interface andwill presumably modify its impedance It will also change the charge and thesurface complexation of the interface Induced polarization is thus considered apromising technology for long-term PRB monitoring

degrada-Self-potential is another geophysical method that is sensitive to interfacechemistry Small intrinsic voltages exist where ionic concentration gradientsoccur These gradients can result from physical movement of charge by fluid flow,charge diffusion at chemical interfaces, or thermal effects Changes in electro-chemistry at the iron–fluid interface can result in characteristic self-potentialsignals Extensive laboratory research is required to determine the induced polar-ization and self-potential signal as chlorinated solvent treatment by granular ironprogresses

4.3.2 C ASE H ISTORIES

Few published examples of the application of geophysical methods to PRBinvestigations exist The case studies that follow focus on the issue of constructionverification These examples illustrate the potential that geophysical imagingtechnologies offer with respect to noninvasive PRB construction evaluation.Applications of geophysical methods to site characterization and PRB monitoring(either short or long term) are currently unreported

4.3.2.1 Electrical Imaging of PRB Construction and

Installation (Kansas City, Missouri)

Slater and Binley (2003) report the results of cross-borehole resistivity andinduced polarization imaging on a PRB installation at the USDOE Kansas Cityplant.This PRB was constructed as a continuous 40 m long by 1.8-m-wide trench.The first 1.8 m of the trench immediately above bedrock was filled with 100%ZVI The remainder of the trench was filled with 0.6 m of zero-valent iron and1.2 m of sand Figure 4.11a shows the cross-sectional geometry of the barrierand site geology Superimposed is the position of electrodes and the finite elementmesh used to reconstruct the conductivity distribution between wells with electrical

Trang 36

imaging Figure 4.11b is a plan view of the site showing 16 boreholes drilled to

bedrock and installed with electrode arrays as per Figure 4.11a Each borehole

pair represents a two-dimensional panel for imaging the cross-sectional electrical

structure of the barrier Two sets of four boreholes were used to investigate barrier

integrity with 3-D electrical imaging (Figure 4.12)

Two-dimensional electrical images were obtained between boreholes 5 and 6

Both the resistivity and induced polarization parameters illustrate high contrasts

with background geology and accurately resolve PRB structure compared to the

design structure in Figure 4.12 and Figure 4.13 These images clearly illustrate

the capability of electrical resistivity and induced polarization imaging for in situ

PRB resolution The resistivity response results from electronic conduction in the

highly conductive granular iron The induced polarization response results from

the impedance at the metal–fluid interface Results of 3-D PRB visualization using

resistivity measurements are illustrated at two locations on the barrier in Figure

4.13 The images illustrate variability in the in situ PRB structure, particularly in

FIGURE 4.11 PRB installation at USDOE Kansas City plant (a) Cross-sectional

geom-etry showing electrical imaging mesh superimposed (b) Plan view showing location of

imaging boreholes along the barrier (After Slater, L and Binley, A., 2002 Geophysics,

Panels used to create 3-D dataset

Distance from midpoint (m)

Reactive iron Sand Clay backﬁll Silty clay Basal gravel Shale bedrock

Trang 37

the vicinity of BH 5, where the integrity of the upper thin section is compromised(Figure 4.12a)

4.3.2.2 Cross-Hole GPR Investigations (Massachusetts

Military Reservation, Massachusetts)

Joesten et al (2001) conducted cross-hole GPR imaging to monitor pilot-scaletesting of a hydraulic fracture method to install a PRB in unconsolidated sedi-ments at depth They also conducted numerical modeling of cross-hole radarpulses to assist in the interpretation of the barrier structure from the radar data.Design specifications called for the installation of two iron walls 5 m apart, 12 mlong, and at a depth of 24 to 37 m This installation depth precluded standardPRB installation procedures and emphasized the need for a noninvasive method

of emplacement evaluation The application of GPR was based on the largereduction in transmitted wave amplitude associated with emplacement of highlyconductive iron

Numerical finite difference modeling was used to predict the effects of holesand wall edges on the transmission amplitude of the radar pulse Figure 4.14

FIGURE 4.12 2-D PRB visualization using (a) electrical conductivity obtained from

resistivity imaging, and (b) normalized IP obtained from IP imaging Compare with ideal

ization (Modified from Slater, L and Binley, A., 2003 Geophysics, 68(3), 911–921 With

Trang 38

polar-246 Barrier Systems for Environmental Contaminant Containment & Treatment

compares the modeled response of a 3.1 m tall wall with the transmission tude data collected after PRB installation The model result is a close fit to thegeneral shape of the field data, suggesting that the top of the wall and wall heightare well defined by the geophysical data The results from cross-hole radaramplitude measurements between 14 boreholes were combined to define vari-ability in cross-sectional amplitude attenuation along the length of the two walls(Figure 4.15) Contour plots defined irregularly shaped walls about 8 m wide.Small-scale structure was tentatively interpreted as stringers of iron possiblyattributable to iron particles moving into higher permeability formations Thisstudy illustrates the potentially high spatial PRB resolution obtainable with radardata

ampli-4.4 VERTICAL BARRIERS

Constructed horizontal barriers are not included in this discussion It is assumedthat horizontal barriers are natural aquitards

The goal of vertical barriers is to prevent groundwater from either entering

or leaving a volume of interest, such that the contamination can be remediated orisolated Therefore, such issues as the location of the barrier, thickness, life

FIGURE 4.13 3-D conductivity images obtained at two locations along PRB: (a) between

BH 5 and 8, (b) between BH 9 and 12.

th from BH9 (m)

Conductivity (siemen) 1.0 0.8 0.6 0.4 0.2 0

Định dạng
Số trang	77
Dung lượng	4,91 MB