Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Methods of Characterization The principal methods for determining subsurface properties are reviewed in this chapter. The methods are reviewed briefly (scientific and technical details can be found in the referenced literature), followed by their range of application and limitations, and the prospects for their improvement. The major noninvasive characterization tools involve geophysical sensing of potential and propagating fields. In addition, a limited number of noninvasive geochemical and geobio- logical measurements can be made. Measurements for characterizing the subsurface may be performed from laboratory to planetary scales; from instrument platforms in boreholes, on the surface, and in vehicles (trucks, boats and airplanes); and from satellites in orbit. Some methods work only from certain platforms (e.g., seismic measurements cannot be made from satellites or aircraft), and a few can be done from all (e.g., electromagnetic observations). In general, the closer the instrument is to the material being measured, the higher is the resolution. Measurement techniques "at a distance" (usually from aircraft or satellites) are remote-sensing methods with meter to tens of meter resolution. Measurement techniques requiring bore- holes (single-hole well logging; geophysical sensing from hole to hole, hole to surface, surface to hole, hole to tunnel, etc.) are invasive, requiring the drilling of a hole. However, they can often provide greatly improved resolution compared to surface measurements. Many of these invasive well-logging techniques are thor- oughly reviewed by Ellis (1987~. In noninvasive characterization, the depth of investigation is highly dependent on technique, logistical constraints, and other factors discussed below, ranging from no surface penetration (surface photoimaging) to hundreds of kilometers in depth (seismic and electromagnetic induction). 52
METHODS OF CHARACTERIZATION 53 Independent of sensor type, all of these methods of characterization also can produce anomaly maps. Such maps yield information about the location of places or regions that are somehow different (or anomalous) from other places. Even if only anomaly information is available, it at least guides later invasive investiga- tion (drilling) to sample these differences. To go beyond simple anomaly maps requires knowledge of the sensor func- tion, logistics of deployment, sensor location and orientation, sources of noise and interference, and so forth. Such information allows computer processing to correct for biases introduced in the measurement process, for example, because of limitations of the instrument (the instrument transfer function), logistical con- straints, or sources of noise. Further, such detailed information can allow the modeling of the measurements and prediction of success in problem application as well as interpretation of derived quantities. For example, if fluid flow is of interest, because the techniques only directly measure changes in some physical field (such as electric or elastic fields), the fluid flow parameters have to be derived through modeling. In all of the methods of characterization, there are certain common problems. Historically, the single largest error often has been precise knowledge of the location and orientation of the measurement sensor. It does not help to have a good measurement but be unable to relocate the measurement site to guide a drill rig to penetrate a contaminant plume. This location sensitivity is especially true of moving sensors in vehicles and satellites, but also of fixed sensors (such as seismic geophones) where later processing and modeling brings out features in the data that must be located. Inadequate locational information has been ru- mored to be the reason for the failure of more than one site characterization or exploration survey. Adequate location surveying may also take longer and cost more than the geophysical survey, although the growing use of GPS (Global Positioning System) technology is ameliorating this problem. Another common problem is lack of property contrast. In comparison to lack of optical contrast, which makes it difficult to find a black cat in a dark coal bin, it is easy to find a furry cat against hard coal by touch. Thus, it is important to consider the available contrast in properties between the target and the host background materials. In a practical sense, many environmental and many engineering geophysical surveys are conducted under less than ideal conditions; for example, often a site is disturbed by human activities including prior excavation of soil or delivery of fill material. Other problems include the presence of either buried or surface utilities such as tunnels, gas lines, sewer drains, and water mains. The mere presence of these more or less passive anthropogenic features disturbs the signals that would otherwise be obtained. Other noise sources include active field disturbances caused by human ac- tivities such as interferences (electromagnetic methods pick up all nearby good conductors, e.g., metallic pipes, wires, and fences), and sources of noise (seismic
54 SEEING INTO THE EARTH noise from wind or nearby traffic; electromagnetic noise from radio stations, cellular phones, and so forth). Seismometers can be susceptible to 60 Hz noise from power lines, as well as higher modes of 60 Hz (such as 120 Hz, 240 Hz, etc.~. In addition to noise problems, there are often logistical constraints (e.g., denial of access to secure or hazardous areas) and physical requirements (e.g., seismic methods require ground contact and are not often effective through con- crete) that are difficult to meet. Each of these is discussed in further detail for the individual method. There are two major types of geophysical measurements. One is measure- ment of potential fields that result from forces decaying away from a source of stored energy. The most common potential-field techniques measure gravita- tional and magnetic fields; less commonly used are thermal and stress fields, which exhibit a quasi-static, nearly time-invariant (or slowly varying) depen- dence on a force generated by a gradient in a field. For all of these methods, the depth of investigation and the resolution are controlled by the measurement sampling interval. Closely spaced measurements give higher resolution for nearby changes in properties, but the resolution decays exponentially with increasing depth. In general, a discrete object with a high contrast against its background is detectable at a depth ten times the size of the object. Measurements of small perturbations in the large source field are made with part-per-million precision and accuracy. The other major type of geophysical measurements, uses propagating fields. Propagating fields result from a disturbance in a field within a material medium that has the capability to store energy. Principal techniques include various adap- tations of seismology and ground penetrating radar (GPR). Resolution is con- trolled by the frequency and the velocity of the propagating wave and is generally comparable to a wavelength. Resolution is also related to the geometry of the sensors and may be much better than one wavelength for arrays of sensors. The depth of penetration is linear with the inverse of frequency (period) and con- trolled by the losses that cause the eventual decay of the propagating wave. Measurements are made of scattered waves in the absence of the source field. POTENTIAL-FIELD METHODS Gravity Measurements Gravity (a potential field) methods measure changes in the earth's natural gravitational field caused by internal variations in bulk density. Density is a basic property of all materials describing the volumetric packing of mass in space. Gravity describes not only the density of minerals but also the packaging of minerals, including fluids and voids in the interparticle spaces (porosity) between mineral grains. The gravity field is a vector quantity pointed toward the center of the earth, with a minor horizontal component near extremes of topography (moun
METHODS OF CHARACTERIZATION 55 tains and canyons). Commercially available sensors are quite simple in prin- ciple measuring the vertical field strength, but they are delicate, expensive, and sophisticated in practice owing to the required precision of measurement (parts per billion) and necessary corrections for location on the earth (altitude and latitude) and for environment (temperature, barometric pressure, tides). Funda- mental principles are described in Blakely (1996) and Hinze (1994~. The subsurface condition that leads to a surface anomaly in the gravitational field is a density variation that changes with horizontal location (lateral density contrast) or depth. A variety of geological conditions cause lateral density con- trasts (e.g., lithologic changes, cavities, faults, folds) as do buried human-made features (e.g., trenches, tunnels, disposal containers). For example, Roberts et al. (199Oa) detected density differences within landfill material in a glaciated area in the U.S. midcontinent. Applications and Limitations Applications of gravity address engineering, environmental, groundwater, and archaeological requirements, such as detection of cavities and tunnels, map- ping of density variations in landfills or aquifer materials, location of under- ground storage tanks, location of buried river channels, detection of faults and fracture zones, and infrastructure assessments. Butler (1984) discusses the use of gravity gradients for near-surface investigations. Because gravity measurements can be taken virtually anywhere, surveys are possible on, inside, and immediately adjacent to structures; on pavement and concrete slabs; and under conditions where other noninvasive methods are not always applicable (Yule et al., 1998~. However, certain frequencies of mechanical vibrations can make attaining preci- sion measurements difficult. Future Prospects Because most technical and theoretical aspects of gravity measurements are quite mature, future improvements will probably be evolutionary in nature. New possibilities are starting to be realized by the application of airborne gravity surveys, which combine gravity determinations with accurate land and sensor positioning using the GPS in a differential mode (NRC, 1994~. At present, the resolution of airborne gravity systems are on the order of a few milligals (1 gal = 1 cm/s2), which is about a thousand times less accurate than microgravity surveys on land. However, the resolution will probably improve with increased attention to this relatively new approach, particularly since it can cover large areas at a smaller cost than land surveys. For certain types of applications, gravity measure- ments from satellites will be possible (NRC, 1997~. A Department of Defense program in the 1980s helped develop a viable gravity "radiometer system (Jakeli, 1993~. The system allows determination of
56 SEEING INTO THE EARTH all independent components of the gradient tensor from moving platforms. Di- verse applications of the gravity "radiometer measurements are a rapidly evolv- ing area of research (Bell, 1997~. Magnetic Measurements Magnetic methods measure changes in the earth's natural magnetic (poten- tial) field caused by variations in magnetic susceptibility and remanence. Mag- netic susceptibility is the property of some minerals (mostly iron bearing) that describes their ability to be magnetized by an external magnetic field. Magnetic remanence is the property that describes the ability of a material to retain mag- netic field strength and direction in the absence of an external magnetic field. Magnetic fields are static vector fields with three-dimensional variation in direc- tion over the surface of the earth with a small superimposed time-varying compo- nent. It is sometimes important to measure both field strength and direction. Modern commercial sensors are simple in principle, measuring either the total field strength (a scalar) or the three-components of the directional field (a vector). Gradient measurements (derivatives of the field) are less often measured. Mea- surements are performed easily and routinely at the part-per-million level. The techniques are quite mature. Fundamental principles are described in Blakely (1996) and Hinze (1994~. Magnetic interpretation is similar to gravity interpretation because both are based on potential-field theory, except that magnetic anomalies are almost al- ways asymmetrical. It is important to realize that anomalies express the net effect of two bipolar vector magnetic fields (induced and remanent) that usually have different intensities and directions of magnetization. Wavelength filtering can be used to better separate the effects of shallow versus deep-seated sources. Using a high-pass filter brings out anomalies at greater depths. Derivative methods accen- tuate the boundaries of anomalies, both shallow and deep. As with all potential-field techniques, it is impossible to calculate the anomaly's depth unambiguously without knowing the shape and magnetic prop- erties of the source of a magnetic anomaly. However, with a prior knowledge about the source, it is often possible to estimate depths to within a factor of 10 to 20 percent depending on the complexity of the anomaly and site noise conditions. Magnetic "radiometry involves simultaneous measurement by two magnetom- eters close to each other (about 0.5 m). The interval gradient is the difference in magnetic intensity readings divided by the distance between sensors. Commonly, two total field instruments are placed on a vertical staff and the vertical gradient is determined. Two key advantages of gradient surveys are (1) they tend to resolve complex anomalies into their component parts (higher resolution than the magnetic field alone), and (2) because the readings are taken simultaneously, it is not neces- sary to correct for diurnal variations and magnetic storms. The orientation of the line between the two sensors must be kept constant, or at least monitored, because
METHODS OF CHARACTERIZATION 57 the gradient will vary with the orientation of this line. In addition, the magnetic cleanliness of the operator (belt buckles, watches, etc.) and magnetic cleanliness of the surface of the area surveyed become even more important than for simple total field measurements. The "radiometer technique is extremely sensitive to surface debris such as nails, cans, wire, and other metallic objects. Applications, Limitations, and Prospects In site characterization, magnetic methods commonly are used for finding buried objects such as drums and abandoned underground fuel storage tanks. Often, the analysis can be very simple. A survey is carried out on a grid or profile line, the results are contoured or plotted, and anomaly locations are noted. Buried metallic objects usually show up as dipolar anomalies (magnetic highs with an adjacent low on their north sides in the Northern Hemisphere). However, sophis- ticated filtering and analysis techniques for separation of superimposed anoma- lies and depth determinations can make processing and interpreting magnetic surveys more complicated (e.g., Telford et al., 1990; Burger, 1992~. Where there are localized changes or contrasts in magnetic properties, the earth's field will induce a secondary or anomalous magnetic field. For buried ferrous metal objects, the magnetic permeability is large relative to surrounding soil and rock and results in a large induced magnetic field. Many magnetic objects, particularly ferromagnetic objects, also have a large remanent or perma- nent component. Accurate interpretation of magnetic data depends on being able to distinguish between the induced and remanent components. Gravity and magnetic methods can be used in a complementary fashion to more tightly constrain geological interpretations. Roberts et al. (199Ob) give an example in which magnetic data recorded over a landfill was enhanced by digital processing. Hinze et al. (1990) show how the gravity and magnetic data from the landfill can be combined to assist in the interpretation of its extent and of the material within it. Future prospects include several innovations related to increased use of GPS, high-temperature superconductivity, and cheaper electronics. Increased use of GPS could lead to robotic control of magnetometers, including unmanned air- craft. High-temperature superconductors may lead to additional sensitivity for portable magnetometers. Cheaper electronics and computing could lead to real- time contouring of data in the field and to increased use of magnetic "radiometry in which two or more magnetometers are read simultaneously at slightly different locations. ELECTRICAL AND ELECTROMAGNETIC METHODS "Electrical methods" refer to measurements of natural or impressed electri- cal fields (potential fields) at low-frequency alternating current (ac) or direct
58 SEEING INTO THE EARTH Electncai Resist~vity Method C, P1 ~, ~ . P2 ; C '~..~.,~ , ^., ~ .r, at, .~ . ~ ~ ,\ ,., ,~. ,~, _' b ': g '. r2 ~ X 2~' ~ i'; =dii~ dX ~ } ~ ~ ~ ~ r -; ;~ ~ ~ ~ ~5 ·~ ~ I, ., 2 ,~ ,5 5 Electromagnetic Induction | Methods I ( ~ Ti HE ~. ~i ~ ;~ Hs:RX ` ... . . I' ~ ~ ~ at,, . ., -. ~ ~. ~ >~N FIGURE 4.1 Simple comparison of electrical resistivity methods and electromag- netic induction (EM) methods. EM methods are generally noncontact, whereas resis- tivity methods require driving metal electrodes into the ground. current (do) using electrodes attached to the ground. By contrast, electromagnetic (EM) methods measure magnetic fields associated with time-varying subsurface currents induced by a natural or artificial electromagnetic source (propagating fields). A schematic comparison of the two is shown in Figure 4.1. GPR is based on high-frequency electromagnetic wave propagation.
METHODS OF CHARACTERIZATION 59 Field Electrical Measurements Electrical field methods measure changes in the earth's natural and induced electrical fields caused by changes in the source origins of the fields and in the electrical properties of the earth. Electrical field methods include do resistivity, complex resistivity, and self potential. Electrical properties of interest are (1) the electrical conductivity, which describes the ability of a material to transport electrical charge, and (2) certain electrochemical and coupled processes. Sources of the electrical fields are the natural fields in the earth caused by the natural magnetic field, solar-wind interaction with the earth, lightning from storms, elec- trochemistry (e.g., the battery-like corrosion of naturally occurring sulfide miner- als in water), and coupled processes (e.g., a voltage called the streaming potential is generated by fluids flowing through pores). Human-made sources also exist from grounding of power grids, corrosion of buried metallic objects, and inten- tional artificial sources connected to the ground (do resistivity sounding). Electrical fields are time-varying vector fields with three-dimensional varia- tion in direction over the surface of the earth. Commercial electric field sensors are simple in principle, consisting of a porous container filled with nonpolarizing electrolyte and electrode. Measurements are made by connecting voltmeter ter- minals to electrodes in the ground at two locations. Measurements are easily and routinely performed at the microvolt or millivolt level. Fundamental principles are described in Keller and Frischknecht (1970), for example. dc Resistivity The do resistivity method is a widely used, inexpensive technique for near- surface investigations. Electrical resistivity methods measure the bulk electrical resistivity of the subsurface directly by measuring the voltage generated by trans- mission of current between electrodes implanted at the ground surface (Figure 4.2~. Resistivity data are collected using single or multiple pairs of current and voltage electrodes (dipoles) with known relative positions. They are interpreted by matching them to theoretical models having a subsurface structure of varying conductivity. In the past, resistivity measurements were usually taken in a straight line on the surface, and the interpretation was done in terms of one-dimensional or two- dimensional models. In a sounding, the measurement array can be expanded about a central position and the data interpreted with a vertical one-dimensional model. In profiling, the relative array geometry and electrode spacing are fixed, but the entire array is moved laterally. Variations indicate lateral or two-dimen- sional changes in subsurface resistivity. This work can be reviewed in texts such as Keller and Frischknecht (1970) and Koefoed (1979~. Sounding provides a resistivity map as a function of depth, comparable to drilling a well and logging it for this information. Resistivity measurements are
60 SEEING INTO THE EARTH current sounce ~ measured 22~ potential . ~ . ~ ~ . ~< FIGURE 4.2 Schematic of a resistivity survey. made at a variety of electrode separations, the depth of the investigation increas- ing with larger separations. Electrode geometry can differ among various appli- cations of this method. Profiling is performed using a constant spacing between electrodes (two outside current electrodes and two inside voltage probes), in which case the arrangement is known as a Wenner configuration (Figure 4.2), and sounding using the Schlumberger configuration where the potential elec- trodes are located in the center of two widely spaced current electrodes. Modern resistivity systems use a multicore cable, multiple electrodes, and computer- controlled switches in a noise-reducing and field-efficient procedure that speeds the data collection process. Depth and resistivity estimates are made with one- or two-dimensional inversion programs. Complex resistivity or induced polarization measurements refer to nonlinear or frequency-dependent resistivity measurements and are treated later in this chapter. It was clear even from early studies (Schlichter, 1933; Pekeris, 1940) that results of the resistivity method must account carefully for both nonuniqueness and resolution issues. More recently, there have been systematic studies of the uniqueness of one-dimensional resistivity sounding results (Parker, 1984; Zohdy, 1989; Simms and Morgan, 1992~. A number of easy-to-use one-dimensional inversion programs are commercially available. Recently, a number of approximate imaging schemes have been developed (e.g., Niwas and Israil, 1987; Zohdy, 1989) that give a better representation of a spatially varying geoelectric section than simple layered-earth models. Two-dimensional resistivity models of the subsurface avoid some of the limitations in one-dimensional models. Clearly, they also entail the collection of much more data. The first step in data processing is usually to display the data graphically in a pseudosection, which is constructed by assigning a resistivity
METHODS OF CHARACTERIZATION 61 value, measured with a specific geometry, to an approximate position at depth. The pseudosection is a data display, not a geoelectric section. Unfortunately, some practitioners mistakenly contour this pseudosection and use it as the end product for interpretation. Resistivity data may be inverted on a computer using algorithms (e.g., Tripp et al., 1984~. Today, the state of the science is to collect extensive current and electrical potential data not in one direction but rather in two surface directions. These data are then interpreted by computer inversion in terms of two-dimensional or three- dimensional subsurface models. Hundreds to thousands of data points must be collected. The generation and inversion of a full three-dimensional subsurface model requires complex computer codes (Ellis and Oldenburg, 1994; Li and Oldenburg, 1994; Zhang et al., 1995~. However, the general uniqueness and resolution of three-dimensional resistivity inversion have not been investigated sufficiently thus far. Self Potential Self potentials (SP) (sometimes called spontaneous potential) are natural do voltages that exist in the earth. They are measured with a high-input impedance voltmeter using nonpolarizing electrodes, often as a by-product of a do resistivity measurement. Natural voltages rarely exceed 100 mV over several hundred meters, and they usually average to zero over distances that are a few times larger than whatever size anomalies may be present. These electrical fields are caused by fluid flow, subsurface chemical reactions, and temperature differences. Depth of placement of the electrodes can have an effect on the reliability of the readings, as can roots and nearby vegetation. Through the fluid flow streaming-potential mechanism, the SP method rep- resents the only known noninvasive passive method directly related to subsurface fluid flow (the seismoelectric method also can measure fluid flow). Small fluid flows associated with cracks in contaminant containment barriers are probably too small to be observed. However, significant fluid movement associated with remediation, such as pump-and-treat and sparging, should produce measurable anomalies. Furthermore, significant fluid flow from leaking dams can be moni- tored and modeled (e.g., Wurmstich et al., 1991; Wurmstich and Morgan, 1994~. Underground chemical pollution, by definition, produces chemical concen- tration or diffusion potentials. However, a number of factors must be favorable for surface anomalies to be detectable. Large chemical concentration differences, shallow depth, and a high electrical resistivity background all contribute to en- hancing the effect. Furthermore, the specific chemistry involved in setting up the diffusion potentials will determine the level of sustainable electric current avail- able from such an electrochemical battery. The SP method is one of the oldest geophysical methods and shows signifi- cant correlation with subsurface processes. Field data are also relatively easy and
62 SEEING INTO THE EARTH inexpensive to obtain. However, detailed interpretation is relatively difficult. Because the voltages are low, they are subject to noise from power lines, pipe- lines, electrical storms, and other environmental sources. Care must be taken with the data acquisition field procedures to ensure that the data are repeatable. Induced Polarization (IP) Using an electrode setup identical to that of the resistivity method, the re- sponse of the ground to the removal of an induced electrical signal can be inves- tigated. The IP method involves measurement of the decay of voltage in the ground following the cessation of an excitation current pulse (time-domain method) or low-frequency (less than 100 Hz) variations of earth impedance (fre- quency-domain method). Most of the stored energy involved is chemical, involv- ing variations in the mobility of ions and variations due to the change from ionic to electronic conduction where metallic minerals are present, and can be likened to a capacitive discharge. Various electrode configurations can be used, com- monly dipole-dipole arrays. In the resistivity method, the passage of electric current through the pores of rocks and soils is dominated by the movement of ions in the pore solution. The earth behaves capacitively at low frequency. Induced polarization measures the low-frequency or capacitive behavior. As ions progress through the pore fluid of rocks they also accumulate along and across surface boundaries. It is this induced accumulation of charge that produces the capacitive effect. Induced polarization is present in varying degrees in all earth materials. However, it manifests itself strongly in two situations. When electrically con- ducting metallic minerals are present, charges accumulate at surface boundaries as charge flow changes from ionic in solution to electronic through the mineral. IP effects are also significant in earth materials such as clays with high internal surface areas. Here the charge accumulation or capacitance is associated with the ubiquitous electrochemical boundary layer. Application, Limitations, and Improvement of IP Methods Traditionally, IP was assessed in the field by measuring the resistivity at two frequencies or by monitoring the change in decay in response to a current pulse. Modern instruments can also measure the phase difference between the real and imaginary parts (complex conductivity) over a wide range of frequency. Such instrumentation opens a new domain because it allows a broad frequency or spectral response to be recorded in the field. The idea is that the spectral response will have behaviors characteristic of the specific chemical reactions taking place. Historically, IP has been used mainly to locate metallic minerals in the near surface (Wait, 1959; Madden and Cantwell, 1967; Bertin and Loeb, 1976; Sumner, 1976~. There were a few early attempts to use the method in groundwa
METHODS OF CHARACTERIZATION 63 ter studies (Vacquier et al., 1957~. Currently, with the widespread emphasis on environmental problems, there has been renewed interest in IP. The idea is that pollutants may alter or influence the surface chemistry and attendant chemical reactions in such a manner that the IP response will be anomalous relative to unpolluted areas. How successful this will be is still a matter of debate, but IP represents one of the few means of possibly performing noninvasive chemistry. As a parallel to the above, because IP is sensitive to clays at depth, it is often of tremendous use in mapping low-permeability clay zones that impede pollutant movement. The negative side of this sensitivity is that it is not possible to uniquely determine if an IP anomaly is due to the actual contamination or to the confining clay zone. The current status of practice is to perform single-frequency, time-domain or phase IP and to plot this as a pseudosection at an approximate depth. Layered-earth IP inversion is the current state of the art, but is not widely practiced. In addition, techniques for two-dimensional IP inversion have also been developed (e.g., Pelton et al., 1978), and attempts are being made to perform three-dimensional IP inver- sion (Oldenburg and Li, 1994~. The main limitations appear to be lack of consistent high-quality, high-volume data and dissemination of computer codes. The IP method has some unique features and possibilities in terms of non- invasive chemical characterization. Good instrumentation is available for embark- ing on the more interesting spectral IP and three-dimensional interpretation meth- ods, and the subject is moving mainly in this direction. Recommendations for needed research in IP are given by Ward et al. (1995), and condensed below: . Opportunities exist in the areas of controlled laboratory and in situ mea- surement to better understand IP signatures of various chemical contaminant situations, especially rock-fluid interactions at a wide range of frequencies. · Further development of digital signal processing and both forward and inverse modeling techniques for IP methods could enhance the extraction of relevant geophysical parameters from IP data. . For environmental work, research in IP data acquisition using the order of 100 data recording channels is needed, along with systems that would quickly, efficiently, and safely control a large array of electrodes with minimal human intervention. Low-Frequency Electromagnetic Field Measurements Electromagnetic induction techniques operate at frequencies less than 1 MHz and are based on inducing eddy currents at the surface. Eddy currents diffuse into the earth at a rate that depends on the electrical conductivity and, to lesser extent, the magnetic susceptibility of the earth. At induction frequencies, the attenuation of electromagnetic waves is proportional to the square root of conductivity and frequency.
64 SEEING INTO THE EARTH At high frequencies (generally greater than 1 MHz), electromagnetic fields propagate like seismic waves, responding mostly to the complex dielectric per- mittivity and, to a lesser extent, to the electrical conductivity and complex mag- netic susceptibility. Electromagnetic measurements above 1 MHz are generally referred to as GPR, which is discussed later in this chapter. Electromagnetic waves are three-dimensional, time-varying, complex vector fields, propagating with directional and polarization properties. Electromagnetic waves may be of natural or induced origin, such as power grids, electric subways, and communica- tions broadcasts. At lower frequencies, the commercial sensors are coils of wire (magnetic sensors or induction coils), and at high frequencies, the commercial sensors are electric field antennas. Measurements are made of the strength (magnitude and phase) and orientation (direction and polarization) of the complex vector fields. Frequency-Domain Electromagnetics This active (as opposed to passive) induction technique uses a transmitting coil that emits a fixed- or swept-frequency EM oscillation and a receiving coil that measures changes in amplitude and phase of the secondary magnetic field associated with eddy currents induced in the ground. These eddy currents and their associated secondary magnetic fields are directly proportional to the electri- cal properties of the shallow subsurface sediments and fluids beneath and be- tween the two coils. The simplest frequency-domain EM instruments, known as terrain conductivity meters, yield depth-integrated measurements of soil conduc- tivity from a depth of a meter to more than 30 m. Conductivity data can be interpreted qualitatively or quantitatively, often in conjunction with other proce- dures designed to directly measure conductivity as a function of depth, such as resistivity sounding. The depth of investigation from frequency-domain EM pro- cedures is a function of coil separation, transmitted frequency, and transmitter power. The end product is a map showing conductivity (millisiemens or millimhos per meter) as a function of lateral position and is used for reconnaissance of a site's electrical properties from the surface down to some depth of interest. Re- sults from a high-resolution frequency-domain EM system at the University of Arizona are shown in Plate 2. Recently, there has been interest in high-frequency EM surveys (Sternberg and Poulton, 1997~. At frequencies of 1 to 30 MHz, it is possible to measure both conductivity and dielectric constant. At these frequencies, the depth of penetration of the EM energy is much greater in conductive soils, compared with standard GPR, which typically uses frequencies of 30 MHz to 1 GHz. The measurement of both conductivity and dielectric constant provides greatly enhanced capability to infer more about the earth's properties (e.g., presence of organic contaminants, engineered structures, and buried nonmetallic ob- jects).
METHODS OF CHARACTERIZATION Time-Domain Electromagnetics 65 Time-domain EM techniques are fundamentally similar to frequency-do- main EM methods, except the transmitted signal is in the form of discrete pulses and the secondary magnetic field is measured during the interval between pulses. The rate of decay of the secondary magnetic field depends on the electrical conductivity structure in the earth. In the presence of highly conductive bodies, the decay is slower than in a less conductive earth. The decay signal can be interpreted in terms of lateral and depth variations in conductivity. The depth of investigation increases with sample time and decreases with ground conductivity, but it can penetrate more than 100 m in some cases. For near-surface investiga- tions, very early time systems have been developed with small portable transmit- ter loops suitable for rapid profiling. An application of time-domain EM is illus- trated in Plate 3. Very Low Frequency Electromagnetics This technique measures the magnetic (and sometimes electric) components of the electromagnetic field generated by long-distance radio transmitters in the very low frequency (VLF) band. These transmitters are used for long-distance naval communication with submarines and operate in the 10- to -30-kHz fre- quency range. Conductive structures on the surface or underground, even when covered with thick overburden, locally affect the direction and strength of the field generated by the transmitted radio signal. The method can locate structures where quantities of groundwater may be held in rock fractures or cavities, and it is sensitive to geological features with long strike length. Large anomalies are associated with electrical cables and buried metallic pipes in urban areas. Com- mercial adaptations display the in-phase and quadrature magnetic field tilt-angle components from which interpretations of lateral changes in conductivity are made. If topsoil is electrically conductive, it is difficult to obtain information from deeper structures. Some VLF equipment also measures the electric field, allowing calculation of average ground conductivity. Applicability of Electrical and Electromagnetic Methods Electrical and electromagnetic methods have tremendous potential for sig- nificant advancements in the field of near-surface investigations. They are cur- rently among the most used techniques for environmental and engineering site investigations, however their potential is far greater than is currently being real- ized. Applications include site stratigraphy, depth to the groundwater table or electrically conductive contaminant plumes, and buried wastes. Time-lapse mea- surements can help detect leaks in engineered contaminant barriers or in tracking the movement of contaminant plumes.
66 SEEING INTO THE EARTH Electromagnetic methods (in contrast with seismic or GPR) have relatively low resolution. Nevertheless, theoretical studies show that EM techniques can have much higher resolution than is achieved currently in normal field surveys. For example, Fullager (1984) showed that these methods "are, in principle, im- bued with unlimited resolving power ... provided no noise is present." A small amount of noise, which is always present to some degree, can have significant degrading effects on the resolving power. Electromagnetic methods can be particularly sensitive to the parameters of greatest interest in near-surface investigations. These methods include direct de- tection of contaminants in the subsurface, sensitivity to geological formation changes, and a correlation with parameters of interest in geotechnical studies. Although there have been controlled demonstrations indicating sensitivity of the EM fields to some of these parameters, much more development of this technol- ogy is needed to apply this in routine field surveys. Electromagnetic systems are in many respects relatively crude in compari- son, for example, to standard three-dimensional seismic survey systems used in the petroleum industry. Some of the pressure to use simpler and relatively unso- phisticated instrumentation comes from the desire to emphasize low cost, easy to-use, and easy-to-understand techniques. Unfortunately, this has limited the usefulness of the techniques and has resulted in much greater expense during drilling and excavation phases in some site investigations. Electromagnetic measurements can use an almost endless variety of sources, instruments (e.g., receivers, array types, recording techniques) and techniques (e.g., those discussed above). On the one hand, this is a great advantage because of the wide diversity of measurements and the opportunity for novel techniques. On the other hand, it is also a disadvantage because much of the past effort in this field has been diffused over a great many different, and incompatible, techniques. Controlled tests are needed to help define the best approaches for each problem of interest in near-surface investigations. Potential Improvements of Electrical and EM Capabilities There are a great many potential research and technical improvements in capabilities. Among those that can be undertaken are the following: r Irequencles. . areal coverage. More sophisticated arrays of sensors and sources. Further development of broadband measurements from do to gigahertz More rapid data collection to allow essentially continuous profiling and · Greatly enhanced capabilities to handle cultural interference, in particular grounding-line interference, not just electrical noise. · Sophisticated systems for critical applications where the alternative would
METHODS OF CHARACTERIZATION 67 be expensive excavation, as well as more economical, easier-to-use systems that contain a subset of new capabilities for EM systems that could be operated in smaller-scale surveys by skilled technicians. Interpretations that more often include complex resistivity at low fre- quencies and combined use of conductivity and dielectric constant at higher ~ . frequencies. · Published case histories are essential for showing applications of im- proved EM techniques for mapping properties of interest in near-surface investi- gations, including more studies of contaminant mapping, permeability determi- nation, formation type, rock strength, and water chemistry. There has been a number of case histories studying some of these properties, but, few have used the full capabilities of EM, including novel arrays, wide bandwidths, complex resistivity and dielectric constants, and high data density measurements. · New field acquisition methods will require greatly improved interpreta- tion techniques that allow handling of complex geometries and widely varying background responses. These techniques include analytical, numerical, and physi- cal modeling as well as novel methods of transforming the raw data into a mean- ingful image of the subsurface. · Easy-to-use interpretation techniques that allow some of the interpreta- tion to be done in near real time in the field. · More laboratory electrical property measurements are needed to deter- mine what can be interpreted reliably from surface electrical and electromagnetic measurements. For example, are there distinctive electrical property changes due to contaminants, what is the relationship between engineering properties such as rock strength and electrical properties, and how well can hydraulic permeability be predicted from electrical properties? Another crucial area for laboratory elec- trical property studies is to find ways to better relate laboratory-scale measure- ments to field-scale measurements. . GROUND PENETRATING RADAR GPR is similar to the seismic reflection method in the basic wave propaga- tion physics, but uses high-frequency electromagnetic waves in the tens of mega- hertz to gigahertz range. Details of the acquisition process differ markedly from the seismic method, most notably because only one channel is acquired. The contrasts being measured with GPR are differences in dielectric permittivity across earth boundaries. The dielectric permittivity is a measure of the ability of a material to store electrical charge (like a capacitor or battery) and principally determines the velocity of propagation of the electromagnetic wave. The product of the dielectric permittivity and the magnetic permeability is analogous to seis- mic impedance. The real part of this product (complex modulus) usually de- scribes how the material stores energy and the imaginary part describes how the material loses (or dissipates) energy.
68 SEEING INTO THE EARTH The EM wave propagates in the earth at the speed of light divided by the square root of the dielectric constant of the geological material. The depth of investigation is inversely proportional to the near-surface conductivity of soils and pore fluids. Due to the smaller wavelengths used in the GPR method, resolu- tion is commonly as much as one order of magnitude better than current seismic reflection techniques. The quality of GPR data and its usefulness in site characterization are deter- mined by (1) the electrical properties of the site, (2) the equipment used, (3) data acquisition procedures and parameters, (4) data processing, and (5) methodolo- gies for interpretation and visualization. The greatest limitation to the widespread use of GPR is the electrical conductivity at a site, which determines the depth of penetration. As a rough guide, GPR is considered to be most useful when the conductivity is less than 10 maim (Davis and Annan, 1989~; this generally pre- vents effective applications of GPR in clay-rich environments. Applications of Ground Penetrating Radar GPR is used to delineate near-surface site stratigraphy, map the extent of buried waste, locate the water table, and find buried utilities. Recent develop- ments in GPR allow direct detection of organic contaminants by observing changes in scattering properties (the texture of the radar record) or dielectric contrast (e.g., oil floating on water). GPR can contribute to site characterization in three ways. The most common use of GPR is in obtaining information about the large-scale (meters to tens of meters) geological structure at a site. A second common use is to detect anoma- lous regions superimposed on the natural geological background; this includes the possible detection of liquid contaminants and the detection of buried objects. The third, and most challenging, potential use of GPR is to obtain information at the meter scale (or less) about the specific physical or chemical properties of the subsurface. The first two applications emphasize the use of GPR as a means of imaging the subsurface; in the third, information about dielectric properties is extracted from GPR data and then related to physical and chemical properties. Each of these is expanded upon below. Large-Scale Imaging The first step in site characterization often involves determining the geologi- cal setting and locating key geological boundaries. Given a site with suitable electrical conductivity, GPR can obtain excellent images of the subsurface that can be used for this purpose. With detailed horizontal and vertical sampling, it is possible to obtain high-resolution (tens of centimeters to meters) images of the subsurface to average depths of 10 m or more. To extract information about the geological structure, the approach usually taken is to identify within the GPR
METHODS OF CHARACTERIZATION 69 section reflectors with a distinct geometry or onentation, or packages of reflec- tors with a characteristic appearance. There are a number of examples in the literature in which GPR data have been used to reconstruct the geological setting by relating the GPR image to the subsurface stratigraphy and sedimentary facies. In such studies there is always prior knowledge from surface outcrop or wells of the lithologies likely to be present and of the depositional environment. Published examples include the use of GPR images to determine the geometry of Pleistocene gravel deposits (Huggenberger et al., 1994), the orientation of major sedimentary structures, and facies thickness and depths in deltas environments (Jol and Smith, 1991~. An example of the GPR image of a deltas deposit is shown in Figure 4.3. The distinct appearance of the radar reflections makes it relatively easy to locate some sedimentary units in the subsurface. In addition, GPR data can be used to help target the anisotropy expected in hydraulic properties in this sedimentary pack- age. The structure seen in a radar section contains information about the spatial heterogeneity of the subsurface and provides a basis for mapping the geological units in the subsurface. DISTANCE (m) (a) 0 20 40 60 1 80 2G It11111111~111111111~111114111~11111 ~11111~1111tlll~llllll~llllllllil~ll 1111 1~111111 ~1~11111111 1~111111111~1 11111111~111 1 01~111111111861 1111111~111111 111~16111~111 ~1 80 1 OG 120 140 160 O 1 00 200 (b) - 1 ~ ~_, ~_~ ~,~ -an, ~n.,,.~,,,~,~,, ~. ~~.~"" ~_ ~ _ aim FIGURE 4.3 GPR profile along the escarpment of the Slave River Valley, Fort Smith, North West Territories, Canada. Early Holocene wave-influenced deltaic de- posits with possible postdepositional slumping. (From Jot and Smith, 1992.)
70 aft : . . . : ,~, Iran t.Rl to, ~1 ye I.; 1 HER R 3 3 ~ COO W. a SO EM ·_. Ct o ·_. · _. Ct So So ~ Ct O ~ ·0 S'-d Id . So ·0 Ct o Ct Cal ;^ ~ Ad Ct ~ Cal Cal Cal o ·_. Cal ~0 Cal By Cal Cal ~ . _. Cal Cal ;^
WHATIS CHARACTERIZED? 7 An important aspect of characterizing the geological setting is to locate the boundaries that can affect the physical, chemical, and biological behavior of regions of the subsurface. One of the key geological boundaries of interest in a number of different applications is the top of the bedrock. This often can be imaged with GPR due to the contrast in dielectric properties between the bedrock and the overlying material. In an example of a GPR image of the bedrock topog- raphy under a fine sand overburden (Davis and Annan, 1989), the contrast in dielectric constant between the overlying sand and the granodiorite bedrock and the lateral continuity of the feature made this a relatively easy target for GPR imaging. Further processing of GPR data can improve the presentation of the information (see Figure 4.4~. Determining the depth to the water table is a characterization objective for which GPR is well suited if the electrical conductivity at the site is not high. The water table can be identified as a flat-lying, high-amplitude reflector in a GPR section (Knoll et al., 1991; Sutinen et al., 1992~. A dominant reflector is seen due to the contrast between the dielectric constant of the unsaturated and the fully saturated materials; therefore the "water table" reflector seen in GPR sections may actually be the top of the capillary fringe. The clearest images of the top of the saturated zone are obtained in coarse-grained materials where the capillary fringe does not "smear" the dielectric contrast. Significant progress has been made in the collection and display of GPR data. With current technologies, it is possible to collect and display three-dimen- sional data in a way that makes it relatively easy for the nonexpert to visualize useful information. This user-friendly aspect of GPR is likely to contribute sig- nificantly to the increased use of GPR in site characterization. In all of the above applications the objective is to obtain a representation of the subsurface in which geological units and boundaries are located. As a useful caveat, the vertical positioning of any feature seen in a GPR record is only as accurate as the velocity determination of the radar signal at that site. Detection of Organic Contaminants Using GPR There have been a number of examples in which GPR has been used to image the presence of organic contaminants in the subsurface. The contrast be- tween the low dielectric constant of most organic contaminants and the high dielectric constant of water, and the availability of pre-spill radar data are what make detection possible. A recent example is the direct monitoring of a sinking organic liquid (tetrachloroethylene) during a controlled spill (Annan et al., 1992; Greenhouse et al.,1993; Brewster et al., 1994~. This investigation was conducted under the most ideal of conditions: the background geology was a homogeneous sand, and a GPR profile was available from the site before the spill. The collec- tion of GPR data as a function of time during this experiment greatly simplified the interpretation by making it possible to relate the time-dependent changes in
72 SEEING INTO THE EARTH the data to the movement of the contaminant. Monitoring such a process is an application for which GPRis well suited. In a more typical situation, GPRis used after the spill of a contaminant, and time-dependent data are not collected. In some case studies the presence of an organic contaminant has been associated with the region in the GPR record where there is a "washed-out" appearance (Olhoeft, 1986~. This change in character of the radar reflectors is by no means a conclusive way of determining the pres- ence or lack of a contaminant. This change in GPR signals can lead to a high degree of uncertainty when GPRis used for contaminant detection without addi- tional information from other types of data. Detection of Buried Objects Using GPR GPR has been found to be a useful technique for the detection of subsurface voids, buried drums, bodies, storage tanks, and utilities. In some cases, an object can be located using the changes in the dielectric properties in the surrounding zone disturbed during the digging and burial of the object. The main limitations to the use of GPR for these purposes have been the background electrical conduc- tivity of the site, the resolution of the GPR data, and cultural interference. In many of the GPR searches for buried objects, the procedure is simply to use unprocessed data and look for anomalous regions in terms of the appearance of the GPR reflectors. Looking for anomalous regions is usually what is done in archeological and forensic studies, where there are many examples of the suc- cessful use of a GPR image to locate an object in the subsurface. Undoubtedly, there have also been numerous times that regions identified as "anomalous" have not corresponded to the target of interest; unfortunately, it is more difficult to find published examples of these failures. A description of various case studies in which GPR was used both successfully and unsuccessfully to find buried bodies is given by Mellett (1996), with a discussion of the various reasons a GPR anomaly can be associated with the burial. If digital signal processing capabilities are available, the ability to resolve the presence of a buried object can be improved dramatically. Examples are given in Bradford et al. (1996), where advanced processing methods were used to improve the resolution of GPR data for the purpose of locating metal and polyvinyl chlo- ride (PVC) pipes. Clear images were obtained of pipes with diameters near the limits of resolution (2 inches in this case) for the antennas used in the survey. Characterizing Small-Scale Properties by GPR In the above applications, the GPR was used to obtain information about the geological structure of the subsurface or the presence of anomalous fluids or solids. It is the geometry and character of the reflectors in the GPR data that are used in a predominantly qualitative way to characterize the subsurface. It is for
METHODS OF CHARACTERIZATION 73 imaging the subsurface in this way that GPR is currently most widely used and, given the current technology, most ideally suited for. There is, however, addi- tional information contained in GPR data that can, ideally, be extracted for the purposes of site characterization. GPR image obtained at a site is one representation of the recorded changes in Dielectric properties of the subsurface. Given that dielectric properties are related to the physical and chemical properties of the subsurface, it should be possible to extract information about these properties from GPR data. Determination of the dielectric properties is not commonly done in practice and represents one of the current limits (or forefronts) in applying GPR to site characterization problems. The two main challenges are in collecting sufficient data to allow inversion for dielectric information and in relating dielectric properties to the physical and chemical properties of interest. A recent example (Greaves et al., 1996) in which GPR data were used to obtain estimates of water saturation at a site illustrates both the problems with and the enormous potential for using GPR data in this way. Currently, GPR can provide excellent images; the future is to provide de- tailed information about physical, chemical, and biological properties that can be used in characterizing the subsurface. Opportunities for Improvement of GPR GPR is a relatively young observational technique. Research needs in GPR, to some degree, resemble those of reflection seismology about 40 years ago. Research is needed in data acquisition, data processing, and inversion and inter- pretation of the data. Some specific examples in these four areas are given below. · The use of multichannel receiving antennas would allow much faster recording of data with different distances between source antenna and receiving antennae. · Multichannel receiving antennae would also allow the use of true three- dimensional recording in an efficient manner. . New strategies for introducing GPR source energy into the ground, in- cluding pulse coding and swept frequency techniques, should improve the pen- etration depth and image resolution. · Collection of cross-polarized data would make it possible to characterize the full vector nature of the electromagnetic wave field. This would lead to new ways of discriminating among subsurface targets. · Digital signal processing of GPR data using reflection seismology data processing software has been on the increase, but algorithm development is needed that accounts for the aspects of GPR data that are not common to seismic methods. For example, processing is required to account for dispersion due to frequency-dependent attenuation and scattering, both of which are much more dominant in GPR data than in seismic data.
74 SEEING INTO THE EARTH . A better understanding of factors that affect the source waveform (e.g., antenna radiation patterns, antenna-ground coupling) would lead to improved deconvolution techniques, which would enhance the temporal resolution. Char- acterizing the source waveform is also a critical part of developing full waveform inversion techniques. . Inversion of the GPR data to obtain a dielectric model is a critical step in using GPR data to describe the structure and properties of the subsurface. Inver- sion methods are needed that account for the complex nature of EM wave propa- gation. . An understanding of the link between the dielectric properties of the subsurface, as imaged in GPR data, and material properties (water content, poros- ity, permeability) is fundamental if we are to use GPR data to describe the magnitude and spatial variation of material properties in the subsurface. SEISMIC METHODS Sound waves propagate through air or water as waves (like the ripples around a rock thrown into a pond). In the earth at lower frequencies, such waves are called seismic waves. In fluids (air or water), the mode of propagation is as a pressure wave with particle motion in the direction of wave propagation (called a compressional wave). In solids, there are both compressional waves and shear waves (where particle motion is perpendicular to the direction of propagation, like the motion of a rope laid on the ground and wiggled sideways). At interfaces between two different materials there are a variety of surface wave modes of propagation. The property to which the seismic wave responds is the complex elastic modulus of the material (density dependent), which determines the veloc- ity of propagation and the rate of decay of the propagating signal. The real part of the complex modulus describes how the material stores energy, and the imagi- nary part describes how the material loses (or dissipates) energy. Seismic waves are three-dimensional, time-varying, complex vector fields, propagating with directional and polarization properties. Seismic waves may be of natural or an- thropogenic origin. Seismic waves are generated naturally by earthquakes (the breaking of rocks under stress), landslides, and events in the ocean and atmosphere (like thunder from lightning). Seismic waves are also generated from anthropogenic sources such as explosions, hammer strikes, and vehicular traffic. Seismic methods are concerned with the production, propagation, and mea- surement of elastic waves that travel within earth materials. The variety of seismic sources commonly used for shallow environmental and engineering investigations includes sledgehammers, weight-drop devices, and explosive sources often in the form of large-gauge shotgun shells fired by percussive or electrical means. Two commonly measured elastic body waves that propagate in the earth are compressional (P) and shear (S) waves. P- and S-waves have veloci
METHODS OF CHARACTERIZATION / /Flood Plain ............................................................................... .................................................. .. ~ me. V A , 1 . i...,, .,,,,, ., I .' , . . . .. .....i .. .. ' .................. ,:.:.: :.:,:-:,:, : 1 ...................... ·:-:-:-:~:-:-:-:~:~:: 1 . ..... . . ................... . 2,,: ,.' . ................... .; ;.;.;.-.; ; .. ; ; ;.; ;. . ...................... ':',2-.-,.: :,:.:- .~2 2 ~ 2 2 3 i:,: :,:,:,:,:,:,.,:- L ... , , .,,,.,,, ~::: ::: :: :: :: : : : .: ~ :: I::: : : -'"-'.-.2; ~ ~ ~ :,2,.,2,:,:~:.:, -,:,:,:,:~:~:,:,:.:,: -.:,:'':'-:':'':'':':':':':" :-:-:':':':':':''':'- -: ::::::: :~ }'''''-'''3 ~t ~. :-:22:2,":.:':-:-:',2:22-:2:-:2.,:,:2: :.':: ::::::::: it: 3 ~: 2"-2.2.~2--''* ''.2- ' ~.-.'2 -2-''"'-'--'"'-""" "' '-:"',, '':'-:' ·.' - - ,,,, ;;~ 75 V' ... ~ ................ ' .............. , ,,,, . ..... .;.; ;.;.; ;.;.;.;.;.;.;.;} .;.;.;.; ;.;.~. .......... . ~, ,~,., . r:::.: .~::::::::::::::.~:::::::. 3 . [.~ ~ ~ 2'' ~'. ~- 'd:.' '.2 '.2.' '~ 2,2,-~ : t::::2. ::~:::::::::::::::::::::::::'::~::- :::::::::::::::::- :::: at, ,~,.. :,:,: :,: am:: :.:,:.:',:,:,:,: I............ ~ :: ::: .~. ::: : :~:~:~:-~:i 7 I.:: ..... I,.,: :-:-.:,:,,:,:: :,: ',:,:.:-,:,:-:-: .,, i ,., : 22 .,.,,i,,,.,j /, . -i i.,i ~- i.,i,. :jj.:2:2-2: ::: :2: :-: :~.':,: :j,:. . ,j - ~- ;., :: ;.- ............ . ~............ ::.~:::::~:: ::::::: ::::::::::::: -:. :.:-: :,: :,:- .:-:: :.: ,.22:~.. ]2:.:.:-,:2:2:,:.:-l2:.:.:.:,:, ~:-l,---j: ~ 'l-2-2, ~;~"2.-2-22.2 -2-i\-- 2-2 I:: :: ::::::::~:::::: ::::::.::: ~:x: ::::::::::::::: 1~ 2~ 2 23.2 '.-.- - ~ ~'7 ~2.~ ;, . :: :~::: -: i:::: :~:: ::: :~::: :~:~.~:~:~:] :~::: :.~:~:~:~: '~'~:~:~:~ h:e'.:`'',:: ::::::::: ._':-: . ~ .. ..... .~ .,. .Y : : _.:.-. , :: :~:-:-:~.~:~:~:.2':-::: .::: :~:2-:-:-: :-:-:-:-: ~.. ·"""':' ... ~ .:~:~''.`'~:: :~: :~::::: :~:: :~: =.'a'*'. : .-~-.2b~ ill 2-2~ ~k ~ ~ ~...1 . .~: . 2" . -.:,:-,:,:,:-: .,. ... - . ·:,-2: :2:;,:,":-:,:,:'., : 2 ~ :,:2:-,:-: ,:.:'.:2:,:2",'-: :2: :2:2 me- 2. - . 2 - 2 2 2: :.2 ~ "2-2-~i "-"2-3 ..-2~ 2.2) 1 ..... : :.::: :- ~.,.j.2,. ~2) ~i , (,j ~ 1 ; - . my. :,.',:: :i,:,-,:,:-: :.:2: Is 2~ 2~ :~:-:,:',:,:-:,:,:2:,:,:.:2:', 2:,:,:.:~:,: :j.:,:2:~:-:-'; :':':-:-:jp':- ' ':-:':-:':-:':':':- ':~:i':' - --:- ' ' ' ' ' '3 ':' ':' ' ' ' ' - '' ' ' --.~ '5. ''I:: .' :: : X::::: :.:.:: :.:: :.:,.~.: :',: :2:',:,:,:,:.: ,:,:.:{ ~ ~ .:~.2....2.2...."-.,.',i .:2:.: :. aft:.:: :-,:.:: :,:: :2.,:,.: :-,:,..: :2:.-.... - :. ..... ~ .: . r : : it:. FIGURE 4.5 Simplified cartoon showing the "mapping" of bedrock and other fea- tures by seismic waves traced to individual geophone receivers. ties related to the physical properties of the material in which they travel.1 The wave velocities are inversely proportional to the square root of density and di- rectly proportional to the square root of the shear modulus for both types of elastic waves and, in addition, bulk modulus for compressional waves. Detectors (geophones) are implanted in the ground and arrayed at known distances and locations from the controlled energy source (see, for example, Figure 4.5~. Precise times of the arrival of the initial seismic waves and subse- quent vibrations are recorded at each geophone; also recorded are the amplitude and period of the waves. The receivers are digital, multichannel detectors that respond to particle-velocity changes associated with the passage of the elastic wave. Seismic methods have also been applied in cross-borehole environments, where the geophones are deployed down boreholes to a known depth. In addition to the physical parameters affecting their velocities, seismic waves are reflected, refracted, and variably attenuated (absorbed) as they pass through media with different elastic properties. These properties allow their use in the interpretation 1 The P-wave velocity is P ~ .._ ,, _, _ , _.~_.~,, . ~ 1 is the bulk modulus, ,u is shear my d 4~) K+ | 3 p , and the S-wave velocity is Vs = plus, and ~/ is the density of the material. ::, where K
76 SEEING INTO THE EARTH SOURCE Hi, RECEIVER FIGURE 4.6 Simple reflection from bedrock. Either a seismic wave-velocity con- trast or a mass density contrast is required for seismic waves to be reflected from the geological interface. of geological layering and waste-zone geometry based on analysis of signal travel time and frequency content. The seismic reflection method is an image-based technique that produces a cross section of the volume of earth under investigation and shows acoustic- impedance contrasts. The cross section (image of the actual data) has traverse distance as the abscissa and reflected-wave travel time as the ordinate. These acoustic-impedance contrasts can be associated with both fluid and rock bound- aries within the earth. The returning signals that constitute the image are stored as traces associated with a particular ground position along the traverse. Each trace is a mathematical vector of particle velocity as a function of time for that position (see Figure 4.6~. As these traces are lined up side by side and corrected for geometrical aspects of their acquisition, the individual responses make an echo mosaic that has the appearance of an image of the shallow-earth cross section. The seismic refraction technique uses a series of geophones arrayed on the surface to analyze the refraction of seismic waves along subsurface interfaces of differing materials, as indicated in Figure 4.7. The technique records the time of the first response of each geophone. Plotting these responses as a function of
METHODS OF CHARACTERIZATION 77 location from the source and processing the information produces a cross section of seismic wave velocity, which reflects geological layering of the subsurface. Many seismic methods were first developed in the petroleum industry as a way of interpreting the geological structure of sedimentary basins. The signals can be processed by a computer to produce an image a seismic reflection pro- file of the subsurface to depths of several kilometers. An uninterpreted profile is not a true geological cross section, although the gross geometry of the bedrock can be determined from it. Normally seismic methods do not provide any infor- mation about the chemical makeup of pore fluids. Recent developments include the adaptation of reflection seismology for uses as shallow as a few meters and the civil engineering adaptation of spectral analysis of surface waves (SASW) used in determining shear wave velocity profiles and soil stiffness for ground response analyses. The SASW method has a variety of earthquake, environmental, and other geotechnical engineering appli- cations; a recent review is given by Stokoe et al. (1994~. SEISMIC REFRACTION S1 R1 R2 R3 R4 As Velocity = V1 s i n i = V0 /V1 , I, , FIGURE 4.7 Seismic refraction. Seismic energy produced by source (S) is detected at a series of geophone receivers (R) after traversing the surface layer and refracting along the interface between layers having two different seismic wave velocities. The velocity in the lower layer must be higher than that in the surface layer for the method to work properly.
78 SEEING INTO THE EARTH Applications of Near-Surface Seismology There are relative advantages and disadvantages of both refraction and re flection seismic techniques (Table 4.1~. The reflection technique can be more powerful in terms of generating interpretable observations over complex geologi- cal structures. This power, however, comes at a cost, because reflection surveys are more expensive than refraction surveys and more computationally intensive. Also, usable reflections are often not obtained in shallow surveys. As a result, many engineering and environmental concerns generally opt for refraction sur- veys when possible. On the other hand, the petroleum industry uses reflection seismic methods almost exclusively. As more channels become available, the increased use of three-dimensional engineering surveys can be expected along with additional applications. The successful use of near-surface seismology spans the spectrum of applications- from those that are well understood and routine to those that are beyond present understanding and technical capabilities. It is important to distinguish where these limits are because vendors and consultants sometimes make unrealistic claims about the capabilities (particularly their capabilities) of near-surface seis- mic techniques. Seismic Refraction Historically, the use of seismic refraction techniques in geoscience and civil engineering investigations has been widespread (Stem, 1962; Redpath, 1973; Mooney, 1977~. The method has advanced throughout the past half-century, which parallels the use of portable, multichannel seismographs. Improvements in both acquisition and processing of data have allowed geophysicists to account for layer dip and spatial velocity variations of both the target refractor and the over- burden soil velocities. Resolution of the geometry of the target refracting surface has been another source of improvement. Development of the method reached maturity in 1980 with the publication of the generalized reciprocal method (GRM) of seismic refraction interpretation (Palmer, 1980~. Since then, most papers on seismic refraction have described refinements of the GRM technique or explained the method to a larger audience through clarification and example (Lankston and Lankston, 1986; Lankston, 1988~. A developing alternative to GRM analysis is refraction travel time tomogra- phy (e.g., Zhang and Toksoz,1998~. Tomography tends to work better than GRM when the near-surface seismic velocity structure is not discrete, continuous, gen- tly dipping homogeneous layers. The method is analogous to medical computer- ized axial tomography (CAT) scans, except that the measurements are made along the earth's surface rather than around a three-dimensional volume. One
METHODS OF CHARACTERIZATION 79 TABLE 4.1 Advantages and Disadvantage of Seismic Refraction and Seismic Reflection Methods. Refraction Method Reflection Method Advantage Disadvantage Advantage Disadvantage Observations generally use fewer source and receiver locations; relatively cheap to acquire Little processing is needed except for trace scaling or filtering to help pick arrival times of the initial ground motion Modeling and interpretations fairly straightforward Observations require relatively large source- receiver offsets Only works if the speed at which motions propagate increases with depth Observations generally interpreted in layers that can have dip and topography; produces simplified models Observations are collected at small source-receiver offsets Method can work no matter how the propagation speed varies with depth Reflection observations can be more readily interpreted in terms of complex geology; subsurface directly imaged from observations Many source and receiver locations must be used to produce meaningful images; expensive to acquire Processing can be . . . expensive as it Is very computer intensive, needing sophisticated hardware and high-level of expertise. Interpretations require more sophistication and knowledge of the reflection process approach employs a two-point ray tracing technique to calculate forward travel times for a model, followed by a least-squares inversion to fit the data to a model that is iteratively adjusted to reduce the misfit between the data and the modeled traveltimes (White, 1989~. Applications in Which Shallow Refraction Usually Works. A common use of the GRM technique is determining the thickness of the soil column (depth to bed- rock) and thereby producing an image of a layer (soil) over a half-space (bed- rock). Locating channels in the bedrock surface and the fill material in these channels that differentially control the flow of fluids in the subsurface is another important use of the GRM refraction technique. These channels represent a natu- ral "French drain" that needs to be known and charted in the subsurface in order to install the proper remediation system at the site (Young et al., 1995~. If infor- mation about the subsurface is obtained on the basis of drill hole information alone, the phenomenon of "spatial aliasing" of these crucial channels could create a distorted view of the subsurface (Henson and Sexton, 1991~.
80 SEEING INTO THE EARTH Applications in Which Shallow Refraction Sometimes Works. One current use of GRM seismic refraction is finding zones of increased fracture density within areas of bedrock where flow and transport of groundwater might occur. Several investigators have succeeded in seismically finding fracture zones by noting decreased target refractor velocity along a segment of the bedrock. Applications in Which Shallow Refraction Does Not Work. One basic theoretical assumption with seismic refraction is that seismic velocity increases with depth. If this assumption is not true at a given site, refraction methods will give incorrect depths or thicknesses of one or more layers. Lankston (1988) discusses ways to detect these errors and to estimate how large such errors might be. Seismic Reflection Three conditions must exist for shallow seismic reflection to work. First, the frequency must be high enough for the reflection to be separable from other arrivals in the early part of the seismogram. In genera!, 3 to 5 cycles of dominant period of the data in time must pass after the onset of the first arrival before the reflection can be easily separated from the direct waves, refractions, and air blast. In some excep- tional data sets, this might be reduced to 1.5 to 2 cycles, but investigators who claim such early reflection arrival times must be able to defend such claims with scientific rigor. In a practical sense, for most data sets with dominant frequencies of less than 150 Hz, reflections at times smaller than 50 milliseconds must be demonstrated as valid by the use of phase identification on unstacked data with the phases traceable through the intermediate processing stages. Second, acoustic impedance contrasts between layers must be large enough to give rise to detectable reflections. These contrasts require a variation in either seismic velocity or material density, or both. Third, the seismic system (energy source, receivers, and seismograph) must work together with sufficient seismic energy and signal sensitivity to register the desired information coming from the ground motion. Three-dimensional seismic reflection has been widely adopted in the petro- leum industry since the mid 1980s. The use of three-dimensional seismic reflec- tion in near-surface work has not been widespread because of the high costs involved. For example, Buker et al. (1998) reported requiring 85 days of field work with a crew of 5 to 7 people to perform a shallow three-dimensional seismic reflection survey of an area 357 m wide by 432 m long. Applications in Which Shallow Reflection Usually Works. Although one cannot tell in advance of field testing whether shallow seismic reflection will work at a particular site or for a particular objective, it often works in the applications discussed below. Data quality is commonly better where the water table is at a depth of only a few meters and where the near-surface materials have not been
METHODS OF CHARACTERIZATION 0.05 0.1 - 19- 15- E 12 ,, u, cat a) ' 3 81 ground surfac7 -datum FIGURE 4.8 Seismic reflection cross section and interpreted cross section. The purpose of the characterization was to determine the placement of a monitoring well, which was designed to be placed at the deepest bedrock-alluvium contact. From D. Steeples. disturbed by construction fill materials or by recent mass wasting such as earthflows or landslides. Working on top of paved surfaces is difficult. Determining gross geological structure is one of the classic uses of seismic reflection, and the technique works well in near-surface applications if the im- pedance contrast and frequency are sufficiently large. Examples of this applica- tion include determining depth to bedrock (see Figures 4.8 and 4.9) and produc- ing a contour map of bedrock beneath alluvium or till. Fault detection is another major use of shallow reflection, primarily for earthquake hazard studies, detection of near-surface pathways of high permeabil- ity, and geological mapping. Offset detection limits under favorable conditions may be as small as one-tenth of the wavelength of the dominant wave frequency. Because of the reflection time uncertainty introduced by static corrections (e.g.,
82 SEEING INTO THE EARTH C - ~ 81. 61. 61. ~ 50 ! 150 200 2SO 300 H~O9 3 n~.ure~ to ~ ~ - . ! ~ ~;~ ._ 5 - U ~ . ~ - 1 - ·'s da,. 47~ days 5 - - I~)~ ~'1 a ~ ~)~f-~,,~ _ FIGURE 4.9 Seismic reflection profiling for geological variability. O lo. . Shallow seismic reflection profiles can provide a picture of geometric complexity and variability of contacts between different types of unconsolidated sediments and the sediment-bedrock interface. Based on a seismic study of the sediments overlying bedrock (depth of about 700 feet) at the Aberdeen Proving Ground in Maryland (Miller et al., 1996), the detail and horizontal interpretation confidence provided by shallow seismic profiles are not possible from drillhole data alone. Extrapolation of drill data from borehole to borehole required significant speculation and assump- tions about lithologic correlations. The seismic investigation was able to describe and detect subtle features of po- tential local hydrologic and geological significance, such as scour and infill patterns (horizontal expanse of less than 200 feet and vertical extent of less than 20 feet). The figure illustrates one of the seismic sections, which is correlated with litholo- gies determined from well logs. To even detect these features (and by no means to image them) with drilling methods would require closely spaced holes and signifi- cant expense. Seismic profiling proved to be a cost-effective method for interpola- tion between boreholes.
METHODS OF CHARACTERIZATION 83 near-surface velocity anomalies), auxiliary use of the presence of diffractions from broken layers is sometimes needed to detect small offsets. Stratigraphic studies can also be done successfully with shallow reflection, although the limits of resolution are debatable. Vertical resolution limit (seeing both top and bottom of a bed) is commonly described as a quarter-wavelength of the dominant frequency (Widess, 1973~. In a practical sense, however, Miller et al. (1994) have shown that half-wavelength is sometimes abetter (or at least more conservative) estimate of vertical resolution limit. The water table usually presents a contrast in seismic wave velocity and a smaller contrast in density, both of which are likely to produce seismic wave reflections. Consequently, the detection of unconfined and perched water tables is often successful. Indeed, in some cases the water table reflection can be strong enough to make detection of slightly deeper reflectors difficult. Applications in Which Shallow Reflection Sometimes Works. Shallow reflection seismology sometimes works in applications that require very high resolution, which necessitates both broad bandwidth and high frequencies. Because high frequencies usually fade rapidly or attenuate with increasing depth, the time window during which these applications work in a satisfactory manner is often quite small. Confidence in the validity of reflections usually increases at times >3 to 5 dominant-frequency cycles after the first break. In contrast, high frequencies are often lost at times greater than perhaps 150 to 200 ms. Consequently, the most commonly successful time window for the applications listed below is from 50 to 200 ms. Detecting voids and tunnels is difficult, although a few occurrences are noted in the literature (e.g., Branham and Steeples, 1988; Miller and Steeples, 1991~. There do not appear to be any examples of direct detection of voids using seismic reflection at depths exceeding 20 m. Robinson and Coruh (1988, p. 215) de- scribed a case of indirect detection of underground coal mining where reflections from times exceeding 120 ms are masked or attenuated by the presence of voids in coal seams. Shallow reflection techniques can sometimes detect and delineate facies changes in the shallow subsurface. The detection of facies changes requires a high signal-to-noise ratio and expert interpretation skills. The facies changes often manifest themselves as subtle changes in amplitude or other seismic at- tributes loosely referred to as "seismic character." Occasionally, stratigraphic detail such as foreset beds on the scale of a few meters can be seen in deltaic deposits and other favorable environments. Intra-alluvial reflections can some- times be seen on the scales of a few meters in thickness. Delineation of beds thinner than a quarter-wavelength based on the shape of the reflection wavelet requires a higher-than-normal signal-to-noise ratio and substantial experience on the part of the interpreter (Widess, 1973~.
84 SEEING INTO THE EARTH Applications in Which Shallow Re;fiection Does Not Work. Currently, shallow seismic reflection techniques appear unable to discriminate the interface between two liquids in near-surface materials, such as between water and dense, nonaque- ous-phase liquids (DNAPLs) or other chemicals. Modeling suggests that the velocity contrasts at the interfaces may be too small to detect with current tech- nology. Furthermore, frequencies at least an order of magnitude higher than those available in shallow seismic reflection are needed to detect such chemical satura- tion lenses at the thicknesses commonly encountered in real-world pollution situations. Direct detection of tunnels or other voids at depths of 100 m or more with surface-seismic reflection appears to be unlikely at this point. Cross-borehole seismic methods, with their substantially higher frequencies, may be able to detect voids a few meters across at these depths under favorable circumstances. Improving Near-Surface Seismic Methods Though relatively well established in petroleum exploration, the use of seis- mology for near-surface applications is still in an emerging state. Areas of poten- tially fruitful research and new applications follow. · The combination of GRM refraction of the compressional wave (P-wave) with the second body wave (S-wave) opens many new possibilities (Hasbrouck, 1987~. For instance, various soil and rock mechanical parameters (e.g., Poisson's ratio, Young's modulus, and shear modulus) can be determined from the combi- nation of compressional wave velocity (Vp), shear wave velocity (Vs), and den- sity (possibly derived from a gravity survey). These elastic constants can help identify rock type and possibly fluid content of pore space (Domenico and Danbom, 1987~. · The combination of GRM refraction of the compressional wave (P-wave) with the second body wave (S-wave) also allows differentiation of "true" geo- metrical relative minima for the surface of a target refractor from artifacts of overburden (soil) velocity variations. Confirmation of existence and correct posi- tion of relative minima is important in potentially locating DNAPL pools in the subsurface (Brewster et al., 1995~. · Research is needed to compare VSP (vertical seismic profile) surveys with those of shallow three-dimensional surveys. Multioffset, multiazimuth VSP may have some advantages in resolution at many locations where boreholes are available. Hole-filling pressurized bladders, which would allow the use of hy drophones at shallow depths above the water table, require further develop- ment. Hydrophones have some advantages over geophones because they are less sensitive to the passage of non-P-wave modes and to distortional surface waves. · Three-component seismology at small-interval (<1 m) offsets is virtually
METHODS OF CHARACTERIZATION 85 nonexistent in the literature. Research in this area is necessary to examine the unaliased high-frequency components of the seismic wavefield, which could lead to improved use of shallow S-wave reflection and to simultaneous use of surface wave (both Love and Rayleigh) inversions to help constrain the near-surface velocity models. Such research could lead to a better understanding of anisotropy of the near-surface materials. · Shallow (1- to 15-m depth) S-wave reflection seismology (e.g., Goforth and Hayward, 1992; Hasbrouck, 1993) is far from routine, but improvements could be of great assistance in engineering seismology, particularly for predicting amplification during earthquakes (Miller et al., 1986~. · When seismic P-wave reflection surveys are conducted, a large portion of other seismic information is unanalyzed. There is a need for collection and analy- sis of whole three-component seismograms that would also allow analysis of S- waves, mode converted waves, and Love waves. · In the petroleum industry, time-varying reflection surveys are now being used to monitor reservoir conditions during hydrocarbon production, including following velocity variations within the reservoir induced by enhanced produc- tion procedures such as steam injection. Time-varying near-surface surveys could possibly be used to good advantage in a number of research applications. Birkelo et al. (1987) have shown that the top of the saturated zone can be monitored during a pumping test. Bachrach and Nur (1998) monitored tide-induced varia- tions in near-surface velocity on an ocean beach in California. Jefferson and Steeples (1995) noted amplitude changes of 12 dB or more in reflection signals as soil moisture varies from about 18 to 36 percent by volume. Time-varying appli- cations of near-surface seismic surveys in the future might include pro- and posttunnel construction to examine the effects of a tunnel's presence. Some possible improvements involve seismic equipment and associated tech- nologies including the following. · There will always be a need for improved seismic sources. With the use of explosives becoming more difficult for various social reasons, the need for im- proved vibratory and impact sources will increase. · One way to reduce the cost of data acquisition is to improve the speed of acquisition. Fifteen years ago the cycle time between shotpoints was about 20 seconds; today it is down to about 5 seconds. However, as the cycle time between shotpoints decreases, an attendant increase in the rate of geophone emplacement must occur. Consequently, there is a need to develop a way to rapidly or auto- matically plant geophones. One way to rapidly deploy geophones might be a draggable or automatically movable set of geophones, similar in concept to a hydroplane streamer used in marine applications but adapted for land applica- tions. Such a low-frequency set of sensors has been used for several years by C. B. Reynolds Associates (Foster et al., 1992), but the primary challenge with their
86 SEEING INTO THE EARTH use is effective coupling to the ground to obtain the broader bandwidth and higher frequencies necessary for high-resolution near-surface applications. · For many years, the seismic receivers of choice in reflection seismology have been velocity geophones. However, manufacturers' specifications some- times do not reach to the high frequencies used in shallow reflection surveys. Consequently, an unbiased and independent research evaluation of receiver at- tributes of high-frequency geophones, following the work of Duff and Lepper (1980), could be useful. Tests should include amplitude, phase, spurious response analysis over a broad bandwidth, at least from 10 to 2 kHz. For shallow high- resolution purposes, accelerometers have become a possible alternative. Other motion-sensitive technologies may be applicable in the future. REMOTE SENSING Remote sensing offers unique observations of the earth's surface and shal- low subsurface that complement conventional mapping and exploration methods. When employed in timely conjunction with field observations, remote sensing can be used to extrapolate local observations over extensive regional areas. A report summarizing remote sensing from satellite and aircraft (Watson and Knepper, 1994) provides a comprehensive evaluation of the state of the art for geological mapping, mineral and energy resources, and environmental studies. It recognizes the evolution from aerial photography to multispectral systems that record solar reflected, thermal emitted, and radar illuminated radiation, and the emergence of imaging spectrometers, which acquire data with spectral resolution comparable to laboratory instruments. There are also several texts on remote sensing. A good source for explaining the physical basis is Elachi (1987~; a report that summarizes many of the opportunities for remote sensing was issued by the National Research Council (NRC, 1995~. An annual conference with published proceedings is sponsored by the Environmental Research Institute of Michigan and is a good source for current application focus. Technical instrument work- shops on instruments are sponsored by the Jet Propulsion Laboratory, and sub- stantial information and illustrative material are available on the Internet. Aerial Photography An ideal environmental remote-sensing system requires high spatial resolu- tion, high sensitivity to changes in baseline characteristics, proven and accessible technologies, and low cost. Airborne photography, which is familiar and rela- tively inexpensive, is still ubiquitous in environmental studies despite obvious limitations awkward archiving, lack of spectral resolution and sensitivity, and difficult integration with correlative geospatial data and digital technologies. Historical photos may provide evidence of waste sites and facilities that are now abandoned. Recent photography facilitates the analysis of current waste disposal
METHODS OF CHARACTERIZATION 87 practices and locations, drainage patterns, geological conditions, signs of vegeta- tion stress, and other factors relevant to contamination site assessment. Addition- ally, aerial photograph fracture trace analysis is used at sites where bedrock contamination is a concern. Overlapping photo pairs can be used to model topog- raphy. Photography is gradually being replaced by digital image data, a trend that will be hastened as commercial satellites with spatial resolution in the 1- to 5-m range are launched in the next few years. Multispectral Scanners Multispectral scanners digitally record several images simultaneously at dif- ferent wavelength bands. The bands are selected to exploit the greatest sensitivity to features of interest and allow significantly more definitive characterization of surface composition and state than does photography. The data are processed using computer image analysis algorithms based on physical or statistical models and knowledge of laboratory-measured physical properties. The most familiar system is the 30-m-resolution TM (Thematic Mapper) satellite instrument that has six reflectance channels (and a 120-m-resolution thermal channel). A number of aircraft systems are available to acquire additional spectral channels with comparable spectral resolution and somewhat higher ground resolution. Imaging radar, acquired as part of a national program is archived (along with photography from a similar program) at the U.S. Geological Survey's EROS Data Center. These data and their derivative images provide uniform spatial coverage, avail- ability at different resolutions, and the digital format that are important for geo- graphic information systems (GIS) analysis. Reflectance data have been success- fully used to distinguish among geological units, to find hydrothermally altered rocks, to infer tectonic setting and local fold and fault structures, to map linear features that may indicate fracture controls, and to indirectly infer lithologic and structural information in heavily vegetated areas based on empirical correlations between vegetation type, density, distribution, and local geological conditions. Thermal infrared data can be used to map silicification and igneous lithologies, fractures, heat (due to near-surface exothermic reactions or underground coal fires), and changes in near-surface thermal properties and to examine surface water changes and groundwater discharge and seepage. Airborne and satellite radar provides all-weather weather capability to define terrain units, to map topo- graphically expressed features that reflect local and regional geological struc- tures, and in hyperarid terrain, to penetrate the upper meter or two. Imaging Spectroscopy Imaging spectrometry can be used to map minerals at the surface for a wide variety of environmental studies. An excellent example (see Plate 7) is the map- ping by aircraft of acid-generating minerals at the Superfund site in Leadville,
88 SEEING INTO THE EARTH Colorado (Swayze et al., 1996~. Mine waste material is dispersed over a 30-km2 area in which oxidation of sulfides releases heavy metals that are carried into the Arkansas River, a major source of water for urban centers and agricultural com- munities along the Rocky Mountain Front Range. The spectroscopy identified areas with higher acid-generating capacity based on the identification and map- ping of distinctive zones of iron-bearing minerals. Research Instruments There are also a number of remote sensing instruments that have consider- able promise for surface characterization but are not yet well established. Passive Microwave Radiometry Natural surfaces radiate mainly in the thermal infrared region; however, radiation at lower intensities extends throughout the electromagnetic spectrum into the submillimeter and microwave region. The radiant power emitted is a function of the surface temperature and its emissivity, which in turn are functions of surface composition and roughness. The large emissivity difference between ice and open water makes mapping polar ice cover and its change one of the most useful applications of microwave radiometry. The high dielectric constant (low emissivity) of water relative to most natural surfaces leads to applications involv- ing mapping of soil moisture variation. However, because large variations can also result from differences in surface roughness or composition, repeat measure- ments are required to resolve ambiguities in interpretation. Microwave data can also be used to infer snow extent, onset of snowmelt, and water equivalent of snow. Limitations are the availability of data and the low spatial resolution. (For passive electromagnetic radiation, resolution is proportional to the ratio of re- ceiver diameter to wavelength; thus, to preserve spatial resolution, very large receivers are required at longer wavelengths.) Radar Interferometry Radar interferometry from satellites can be used to detect minute changes in land surface geometry by comparing the phase difference between observations at two different times. Because the method is sensitive to differences as small as a few centimeters, it is sensitive to active faulting, subsidence caused by fluid withdrawal, pre-eruption volcanic swell, erosion, or tectonic creep. Atmospheric differences between the two observation times can cause substantial errors, and it is necessary that the surface has not been too greatly disrupted. This technique appears to have substantial potential for worldwide study of geological hazards once a good database and case history experience have been established.
METHODS OF CHARACTERIZATION Lasers 89 A number of experimental laser systems have been used from aircraft to illuminate the ground in order to measure surface conditions including surface texture, composition, elevation (decimeter accuracy), and water quality and depth (using fluorescence). GEOCHEMICAL METHODS Assessment of the subsurface geochemistry involves describing the chemi- cal composition of solids, liquids, or gases. That is, the geochemistry of the subsurface may be defined as the chemical composition of bedrock and soil, groundwater and its dissolved or suspended load, and the atmosphere in the unsaturated zone. It is unlikely that all aspects of subsurface geochemistry can be determined remotely. In fact, relatively few chemical parameters can be readily detected without direct sampling and analysis. However, remote methods of chemical sensing for some constituents of interest in contaminated aquifer sys- tems show promise. Volatile Gas Emission Wide use of organic solvents in the industrial and commercial sectors and of refined petroleum as fuels in numerous applications has led to nearly ubiquitous contamination of the environment with volatile organic compounds (VOCs) of a variety of compositions. Volatilization at the surface of the water table and diffusion through the air- filled pore spaces in the vadose zone cause VOCs to be present at the surface overlying a contaminated site (see, for example, Figure 2.1~. Soil-gas analysis became a popular screening tool for detecting VOCs during the 1980s. Soil-gas surveys can generate extensive chemical distribution data quickly at a fraction of the cost of conventional invasive methods and offer the benefits of real-time data. There are two types of soil-gas sampling. Grab sampling typically involves the insertion of a hand-held probe to depths of only tens of centimeters, with the volatiles pumped directly into a portable gas chromatograph. Passive sampling provides a measure of VOCs over time. It uses a sorbent material, such as acti- vated carbon, that is placed below ground and later retrieved for analysis. VOCs and gases can also be important as indicators of biological degrada- tion reactions proceeding at depth. Isotopic information on these gases, obtained through mass spectroscopic methods in the laboratory, may yield even more information about the nature and extent of biodegradation reactions occurring within an aquifer.
9o SEEING INTO THE EARTH Water Composition Ground water moves into, through, and out of a given portion of the shallow subsurface. In doing so, reactions occur among the components of the aquifer system (water, minerals, atmosphere, and associated biota) that can lead to a change in the composition of the groundwater. Sampling the groundwater in wells or springs downgradient of the site may allow inferences to be made about a portion of the subsurface that we cannot sample directly. That is, the composi- tion of the dissolved or suspended load in the groundwater may be used as an indicator of the composition of the solids, liquids, and gases in the study area, as well as of the reactions they are undergoing. (Many of the principles are similar to the geochemical water sampling developed in ore deposit exploration.) Tracers may be passive or natural products of the environment, or they may be introduced purposefully for the purposes of sampling. Natural or artificial tracers may be introduced and sampled without disturbing the physical integrity of the study site. Analysis of the outcome is by traditional chemical methods in the field or the laboratory. Most solutes in natural and contaminated groundwater are ionic; that is, they are present as charged cations and anions in solution. Dissolved ions can carry an applied electrical current; if they are present in high enough concentrations in groundwater, noninvasive electrical geophysical methods can detect their pres- ence and location. An example of a useful and successful application is in the mapping of saltwater intrusion fronts in coastal water supply aquifers. Fresh groundwater has highly contrasting electrical properties to the intruding seawa- ter, and because of density differences and poor mixing in a porous medium, the contact between the two types of water can be fairly sharp and, in these cases, relatively easily detected and mapped. Another widespread problem is the presence of plumes of landfill leachate within an otherwise clean groundwater system. Electrical methods can map such plumes as well as their migration because the leachates are typically high in dissolved salts and metals and contain a variety of organic compounds. Similarly, acid mine drainage can also be mapped because of high concentrations of dis- solved solids and metals. Significant challenges remain in detecting nonionic contaminants, including many dissolved organic compounds such as pesticides. Composition of the Solid Phase Remote assessment of the chemical composition of the subsurface's solid portions (soil and bedrock) is problematic. Relatively few material properties that can be remotely measured yield information about the chemical composition of solid materials, although the presence of some minerals can be modeled. One approach to the composition would be to use a combination of the knowledge of site geology with geophysical determinations of density or porosity contrasts to support an interpretation of rock type, but this does not go much beyond what a
METHODS OF CHARACTERIZATION 91 geologist can do without noninvasive technologies. A metallic object can be detected from the surface through the contrast of its electrical or magnetic prop- erties with the enclosing silicate, carbonate, or oxide rock, but little specific knowledge can be gained about chemical composition. Use of self potentials and induced polarization methods potentially could be applied to such chemical de- terminations. Radioactive Methods Detection of natural radioactivity (or that resulting from disposal of radioac- tive materials) can be of use in characterizing the shallow subsurface. Applica- tions include, for example, regional mapping, prospecting for some minerals, and detection of leaking storage facilities containing radionuclides. The same proper- ties that make radionuclides dangerous also make them easy to track in the environment. "The Multi-Agency Radiation Survey and Site Investigation Manual (MARSSIM) provides a nationally consistent consensus approach to conducting radiation surveys and investigations at potentially contaminated sites" (Environ- mental Protection Agency, 1997~. The manual describes well-tested methods and details the specific methodology and analysis that should be used. Several other aspects of radioactivity that can be valuable in site characterization involve inva- sive (e.g., borehole logging [Ellis, 19871) or direct sampling (e.g., tritium or bomb-pulsed chlorine tracers in subsurface water). GEOBIOLOGICAL METHODS Properties of the biota of most interest to site characterization biologists may be the most difficult to determine noninvasively. The identity, abundance, diver- sity, and ecology of the resident organisms, as well as their overall physiological status, are the most important general properties to assess. Biological processes in the near surface ultimately depend on the genetic makeup of the near-surface biota, which in turn depends on physical and chemi- cal environmental factors that select the biota at a given site. Generic properties of the biota (identity, abundance, diversity, and ecology and their overall physi- ological status and activity) will be important in most site characterizations. However, given the spatial variability and heterogeneity of geological settings, large variations in metabolic activities may occur across a given site. For ex- ample, a process such as aerobic respiration of a pollutant chemical (or biomineralization) depends upon the availability of oxygen, which itself can be controlled by water content, inorganic oxidation-reduction reactions, and content of exchangeable organic compounds. The pollutant chemical itself may be more or less available for respiration depending on its solubility in water, its octanol- water partition coefficient, its organic matter content, or competition with soil particle surface absorbers. Finally, the total abundance of aerobic heterotrophic
92 SEEING INTO THE EARTH organisms controlled by oxidizable organ~cs will greatly influence the oxygen available for respiration of the pollutant. Presently there are no noninvasive methods for direct measurement of bio- logical presence or metabolic activity in the near surface. However, in some geological settings, subsurface biological activity can be inferred indirectly from near-surface biogeochem~cal activity, which might be measurable using noninvasive methods. For example, near-surface biogeochem~cal activity in the vicinity of oil reservoirs has been mapped by electrical resistivity methods (Sternberg, 1991), and airborne imaging spectroscopy has been used to detect and map biogenic minerals in acid and neutral drainage areas of acidified watersheds (see Plate 7~. Minimally invasive methods such as soil-gas analysis by gas chro- matography, chemical assays of bioaccumulating plants, and bacterial indicator culturing of surface soil have been used to a limited extent for petroleum and mineral exploration as well as environmental pollution studies. Noninvasive tech- nologies show some promise in biological assessments, but until more research is done to develop other methods, the characterization of site biology will still depend to a large degree on analysis of samples obtained by invasive methods. Development of coordinated noninvasive and minimally invasive methods for geobiological site characterization remains a challenge (e.g., Ghiorse, 19971. REFERENCES Annan, A. P., M. L. Brewster, J. P. Greenhouse, J. D. Redman, G. W. Schneider, G. R. Olhoeft, and K. A. Sander, 1992. Geophysical monitoring of DNAPL migration in a sandy aquifer, 62nd Annual International Meeting, Society of Exploration Geophysicists, Expanded Abstracts 62, 344-347. Bachrach, R., and A. Nur, 1998. High-resolution shallow-seismic experiments in sand, Part I: Water table, fluid flow, and saturation, Geophysics 63, 1225 1233. Bell, R. E., 1997. Gravity "radiometry resurfaces, The Leading Edge 16(1), 55-59. Bertin, J., and J. Loeb, 1976. Experimental and Theoretical Aspects of Induced Polarization, Vols. I and II, Gebruder Borntraeger, Berlin-Stuttgart. Birkelo, B. A., D. W. Steeples, R. D. Miller, and M. A. Sophocleous, 1987. Seismic-reflection study of a shallow aquifer during a pumping test, Ground Water 25, 703-709. Blakely, R. J., 1996. Potential Theory in Gravity and Magnetic Applications, Cambridge University Press, Cambridge, 441 pp. Bradford, J., M. Ramaswamy, and C. Peddy, 1996. Imaging PVC gas pipes using 3-D GPR, Pro- ceedings of the Symposium on the Application of Geophysics to Engineering and Environmen- tal Problems, Environmental and Engineering Geophysical Society, Wheat Ridge, Colorado, pp. 519-524. Branham, K. L., and D. W. Steeples, 1988. Cavity detection using high-resolution seismic reflection methods, Mining Engineering 40, 115-119. Brewster, M. L., A. P. Annan, J. P. Greenhouse, B. H. Kueper, G. R. Olhoeft, J. D. Redman, and K. A. Sander, 1995. Observed migration of a controlled DNAPL release by geophysical methods, Groundwater 33, 977-987. Buker, F. A., G. Green, and H. Horstmeyer, 1998. Shallow 3-D seismic reflection surveying: Data acquisition and preliminary processing strategies, Geophysics 63, 1434-1450.
METHODS OF CHARACTERIZATION 93 Burger, H. R., 1992. Exploration Geophysics of the Shallow Subsurface, Prentice-Hall, Englewood Cliffs, N.J. Butler, D. K., 1984. Interval gravity-gradient determination concepts, Geophysics 49(6), 828-832. Davis, J. L., and A. P. Annan, 1989. Ground-penetrating radar for high-resolution mapping of soil and rock stratigraphy, Geophysical Prospecting 37(5), 531-552. Domenico, S. N., and S. H. Danbom, 1987. Shear-wave technology in petroleum exploration Past, current, and future, in Shear-Wave Exploration, S. H. Danbom and S. N. Domenico, eds., Society of Exploration Geophysicists Special Publication. Duff, B. M., and C. M. Lepper, 1980. A high-resolution controlled-source audiomagnetotelluric system for mining applications, 50th Annual International Meeting, Society of Exploration Geophysicists, Reprints, 80, Session:E.8. Elachi, C., 1987. Introduction to the Physics and Techniques of Remote Sensing, John Wiley & Sons, New York, 413 pp. Ellis E. V., 1987. Well Logging for Earth Scientists, Elsevier, New York, 532 pp. Ellis, R. G., and S. W. Oldenburg, 1994. The pole-pole 3-D dc-resistivity inverse problem: A conjugate-gradient approach, Geophysical Journal International ll9, 187-194. Environmental Protection Agency (EPA), 1997. Multi-Agency Radiation Survey and Site Investiga- tion Manual, EPA-402-R-97-016 (www.epa.gov/rpdwebO0/marssim/filestoc.htm). Fisher, E., F. A. McMechan, A. P. Annan, and S. W. Cosway, 1992. Examples of reverse-time migration of single-channel ground-penetrating radar profiles, Geophysics 57, 577-586. Foster, M. S., K. R. Nunn, S. A. Lewis, and D. J. Reynolds, 1992. Zero phasing seismic data without wells offshore W. Africa: Reducing uncertainty and variability of the wavelet, 62nd Annual International Meeting, Society of Exploration Geophysicists, Expanded Abstracts 92, 241-243. Fullager, P. K., 1984. A uniqueness theorem for horizontal loop electromagnetic frequency sound- ings, Geophysical Journal, Royal Astronomical Society, 559-566. Ghiorse, W. C., 1997. Subterranean life, Science 275, 789-790. Goforth, T., and C. Hayward, 1992. Seismic reflection investigations of a bedrock surface buried under alluvium, Geophysics 57(9), 1217-1227. Greaves, R. J., D. P. Lesmes, M. J. Lee, and M. N. Toksoz, 1996. Velocity variations and water content estimated from multi-offset, ground-penetrating radar, Geophysics 61(3), 683-695. Greenhouse, J., M. Brewster, G. Schneider, D. Redman, P. Annan, G. Olhoeft, J. Lucius, K. Sander, and A. Mazzella, 1993. Geophysics and solvents: The Borden experiment, The Leading Edge 12, 261-267. Hasbrouck, W. P., 1987. Hammer-impact, shear-wave studies, in Shear-Wave Exploration, S. H. Danbom and S. N. Domenico, eds., Society of Exploration Geophysicists Special Publication. Hasbrouck, J. C., 1993. An Integrated Geophysics Program for Non-Intrusive Characterization of Mixed-Waste Landfill Sites, U. S. Department of Energy, DOE Contract No. DE-AC04- 86ID 12584. Henson, H., and J. L. Sexton, 1991. Premine study of shallow coal seams using high- resolution seismic reflection methods, Geophysics 56, 1494-1503. Hinze, W., 1994. Engineering and environmental applications of gravity and magnetic methods, in Introduction to Applied Geophysics: Short Course, Environmental and Engineering Geophysi- cal Society. Hinze W. J., R. L. Roberts, and D. I. Leap, 1990. Combined analysis of gravity and magnetic anomaly data in landfill investigations, in Geotechnical and Environmental Geophysics, Vol. II: Environmental and Groundwater, S. H. Ward,, ea., Society of Exploration Geophysicists, Investigations in Geophysics #5, pp. 267-272. Huggenberger, P., E. Meier, and A. Pugin, 1994. Ground-probing radar as a tool for heterogeneity in gravel deposits: Advances in data processing and facies interpretation, Journal of Applied Geophysics 31, 171-184.
94 SEEING INTO THE EARTH Jakeli, C., 1993. A review of gravity "radiometer survey system analyses, Geophysics 58(4), 508- 514. Jefferson, R. D., and D. W. Steeples, 1995. Effects of short-term variations in near-surface moisture content on shallow seismic data, 65th Annual International Meeting, Society of Exploration Geophysicists, Expanded Abstracts 95, 419-421. Jol, H. M., and D. G. Smith, 1991. Ground penetrating radar of northern lacustrine deltas, Canadian Journal of Earth Sciences 28, 1939-1947. Keller, G. V., and F. C. Frischknecht, 1970. Electrical Methods in Geophysical Prospecting, Pergamon Press, Oxford. Killham, K. 1994. Soil Ecology, Cambridge University Press, New York. Knoll, M. D., F. P. Haeni, and R. J. Knight, 1991. Characterization of a sand and gravel aquifer using ground-penetrating radar, Cape Cod, Massachusetts, in U.S. Geological Survey Toxic Substances Hydrology Program, Water Resources Investigations (USGS-WRI-91 -4034), Reston, Virginia. Koefoed, O., 1979. Geosounding Principles, I: Resistivity Sounding Measurements, Elsevier, New York Lankston, R. W., 1988. High resolution refraction seismic data acquisition and interpretation, in Environmental Geophysics, Society of Exploration Geophysicists Special Publication. Lankston, R. W., and M. M. Lankston, 1986. Obtaining multilayer reciprocal times through phantoming, Geophysics 51, 45-49. Li, Y., and D. W. Oldenburg, 1994. Inversion of 3D dc-resistivity data using an approximate inverse mapping, Geophysical Journal International 116, 527-537. Madden, T. R., and Cantwell, 1967. Induced polarization: A review in mining geophysics, BE Soc. Exploration Geophysics. Mellett, J. S., 1996. Location of human remains with ground-penetrating radar, in Fourth Interna- tional Conference on Ground-Penetrating Radar, Special Paper 16, Geological Survey of Finland, pp. 359-365. Miller, R. D., and D. W. Steeples, 1991. Detecting voids in a 0.6-m coal seam, 7 m deep, using seismic reflection, in Geoexploration 28, 109-119. Miller, R. D., S. E. Pullan, J. S. Waldner, and F. P. Haeni, 1986. Field comparison of shallow seismic sources, Geophysics 51, 2067-2092. Miller, R. D. S. E. Pullan, D. W. Steeples, and, J. A. Hunter, 1994. Field comparison of shallow P- Wave seismic sources near Houston, Texas, Geophysics 59, 1713-1728. Miller, R. D., J. Xia, S. Swartzel, J. Llopis, and P. Miller, 1996. High-resolution seismic reflection profiling at Aberdeen Proving Grounds, Maryland, in SAGEEP '96, Environmental and Engi- neering Geophysical Society, Wheat Ridge, Colorado, pp. 189-201. Mooney, H. M., 1977. Handbook of Engineering Geophysics, Bison Instruments. Niwas, S., and M. Israil, 1987. A simple method for interpretation of dipole resistivity soundings, Geophysics 52, 1412-1417. National Research Council (NRC), 1994. Airborne Geophysics and Precise Positioning: Scientific Issues and Future Directions, Board on Earth Sciences and Resources, National Academy Press, Washington, D.C., 111 pp. NRC, 1995. Earth Observations from Space: History, Promise, and Reality, Space Studies Board, National Academy Press, Washington, D.C., 310 pp. NRC, 1997. Satellite Gravity and the Geosphere: Contributions to the Study of the Solid Earth and Its Fluid Envelope, Board on Earth Sciences and Resources, National Academy Press, Wash- ington, D.C. Oldenburg, D. W., and Y. Li, 1994. Inversion of induced polarization data, Geophysics 59, 1327- 1341.
METHODS OF CHARACTERIZATION 95 Olhoeft, G. R., 1986, Direct detection of hydrocarbon and organic chemicals with ground penetrating radar and complex resistivity, in Proceedings of the NWWA/API Conference on Petroleum Hydrocarbons and Organic Chemicals in Ground Water-Prevention, Detection and Restora- tion, National Water Well Association (NWWA), Dublin, Ohio, pp. 284-305. Palmer, D., 1980. The Generalized Reciprocal Method of Seismic Refraction Interpretation, Society of Exploration Geophysicists Monograph. Parker, R. L., 1984. The inverse problem of resistivity sounding, Geophysics 49, 2143-2158. Pekeris, C. L., 1940. Direct method of interpretation in resistivity prospecting, Geophysics 5, 31-46. Pelton, W. H., L. Rijo, and C. M. Swift, Jr., 1978. Inversion of two-dimensional resistivity and induced polarization data, Geophysics 63(4), 788-803. Redpath, B. B., 1973. Seismic Refraction Exploration for Engineering Site Investigation, National Technical Information Service. Roberts, R. L., W. J. Hinze, and D. I. Leap, 1990a. Application of the gravity method to the investigation of a landfill in glaciated midcontinent, U. S. A., in Geotechnical and Environmen- tal Geophysics, Vol. II: Environmental and Groundwater, S. H. Ward,, ea., Society of Explo- ration Geophysicists, Investigations in Geophysics #5, pp. 253-260. Roberts, R. L., W. J. Hinze, and D. I. Leap, 1990b. Data enhancement procedures on magnetic data from landfill investigations, in Geotechnical and Environmental Geophysics, Vol. II: Environ- mental and Groundwater, S. H. Ward,, ea., Society of Exploration Geophysicists, Investiga- tions in Geophysics #5, pp. 261-266. Robinson, E. S., and C. Coruh, eds., 1988. Basic Exploration Geophysics, John Wiley & Sons, New York, 562 pp. Shi, W., and F. D. Morgan, 1996. Non-uniqueness in self-potential inversion, 66th Annual Interna- tional Meeting, Society of Exploration Geophysicists, Expanded Abstracts 66, 950-953. Simms, J. E., and F. D. Morgan, 1992. Comparison of four least-squares inversion schemes for studying equivalence in one-dimensional resistivity inversion, Geophysics 57, 1982-1293. Slichter, L. B., 1933. The interpretation of resistivity prospecting method for horizontal structures, Physics 4, 307-322. Stam, J. C., 1962. Modern developments in shallow seismic refraction techniques, Geophysics 27(2), 198-212. Sternberg, B. K., 1991. A review of some experience with the induced-polarization/resistivity method for hydrocarbon surveys: Successes and limitations, Geophysics 56, 1522-1532. Sternberg, B. K., 1997. The LASI high-frequency ellipticity system, in Proceedings of the High- Resolution Geophysics Workshop, Tucson, Arizona. Sternberg, B. K., and M. M. Poulton, 1997. High Resolution Subsurface Imaging and Neural Net- work Recognition: Non-Intrusive Buried Substance Location, DOE Report on Contract DE- AC21- MC 29 101 A001. Stokoe, K. H., S. G. Wright, S. A. Barg, and J. M. Roesact, 1994. Characterization of geotechnical sites by SASW method, in Geophysical Characterization of Sites, R. D. Woods, ea., Oxford & 1BH Publishing Co., pp. 15-25. Sumner, J. S., 1976. Principles of Induced Polarization for Geophysical Exploration, Elsevier Scientific Publishing Co. Sutinen, R., P. Hanninen, R. Cromwell, and E. Hyvonen, 1992. GPR and dielectric classification of glacial materials, in Fourth International Conference on Ground-Penetrating Radar, Special Paper 16, Geological Survey of Finland, pp. 133-138. Swayze, G. A., R. N. Clark, R. M. Pearson, and K. E. Livo, 1996. Mapping acid-generating minerals at the California Gulch Superfund site in Leadville, Colorado using imaging spectroscopy, Summaries of the 6th Annual JPL Airborne Earth Science Workshop. Telford, W. M., L. P. Geldart, and R. E. Sheriff, 1990. Applied Geophysics, Cambridge University Press, New York.
96 SEEING INTO THE EARTH Tripp, A. C., G. W. Hohmann, and C. M. Swift, Jr., 1984. Two-dimension al resistivity inversion, Geophysics 49, 1708- 1717. Vacquier, V., C. R. Holmes, P. R. Kintzing, and M. Laverone, M., 1957. Prospecting for ground water by induced electrical polarization, Geophysics 22, 660-687. Wait, J. R., ea., 1959. Overvoltage Research and Geophysical Applications, Pergamon Press, Lon don. Ward, S. H., B. K. Sternberg, D. J. LaBrecque, and M. M. Poulton, 1995. Recommendations for IF research, The Leading Edge 14, 243-247. Watson, K., and D. H. Knepper, eds., 1994. Airborne Remote Sensing for Geology and Environ- ment-Present and Future, U.S. Geological Survey Bulletin 1926, 43 pp. White, D. J., 1989. Two-dimensional seismic refraction tomography, Geophys. J. 97, 223-245. Widess, M. B., 1973. How thin is a thin bed? Geophysics 38, 1176-1180. Wurmstich, B., and F. D. Morgan, 1994. Similarities in modeling groundwater flow and do resistiv- ity, 64th Annual International Meeting, Society of Exploration Geophysicists, Expanded Ab- stracts 94, 578-579. Wurmstich, B., D. F. Morgan, G. P. Merkler, and R. L. Lytton, 1991. Finite element modeling of streaming potentials due to seepage: Study of a dam, 61st Annual International Meeting, Soci- ety of Exploration Geophysicists, Expanded Abstracts 91, 542-544. Young, R. A., J. M. Forgotson, L. White, D. O'Meara, Z. Deng, Z. Liu, C. L. Liner, R. Weindel, and S. H. Danbom, 1995. The OCAST project: Integrated geophysical characterization assisting flow simulation, 65th Annual International Meeting, Society of Exploration Geophysicists, extended abstract. Yule, D. E., M. K. Sharp, and D. K. Butler, 1998. Microgravity investigations of foundation condi- tions, Geophysics 63(1), 95-103. Zhang, J., and M. N. Toksoz, 1998, Nonlinear refraction traveltime tomography, Geophysics 63, 1726-1737. Zhang, J., R. L. Mackie, and T. Madden, 1995. 3-D resistivity forward modelling and inversion using conjugate gradients, Geophysics 60, 1313-1325. Zohdy, A. A. R., 1989. A new method for the automatic interpretation of Schlumberger and Wenner sounding curves, Geophysics 54(2), 245-253.
