Archaeological geophysical instruments (conductivity, groundpenetrating

Transcription

1 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page Archaeological Geophysics: Seeing Deeper with Technology michael rogers Archaeological geophysical instruments (conductivity, groundpenetrating radar, magnetometry, magnetic susceptibility, and resistivity) record the near- surface geophysical properties of the soils and objects within the soils. Each of these noninvasive methods records different physical parameters, and each method s effectiveness depends upon the properties of the materials being studied (Clark 1990; Gaffney and Gater 2003; Scollar et al. 1990). The goal of using these methods is to obtain a view into the subsurface prior to or without excavating, and the use of multiple instruments can enhance one s view into the subsurface (Clay 2001*; Kvamme 2003b; Maillol et al. 2004; Murdie et al. 2003; Rogers 2001; Simpson et al. 2009; Watters 2008). Archaeological geophysical surveys record information about geophysical parameters, but more importantly they record geophysical signals related to human interaction with the landscape. Interpretation of archaeological geophysical data attempts to identify human activities on the horizontal landscape by identifying house locations, activity areas, burials, and midden deposits, to name a few relevant cultural features (Conyers and Goodman 1997; Johnson 2006; Rogers 2001; Rogers et al. 2005). A highly successful archaeological geophysical survey can map layouts of entire villages, providing guidance to archaeological investigations (Conyers 2008; Gaffney 2008b; Kvamme 2003a). Archaeological geophysical instruments have been noninvasively probing the very near surface to several meters deep since the 1940s. Recent advances in portable computer (specifically laptop) processing speeds, memory, and 112

2 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 113 Quality archaeological geophysical surveys use a high- precision survey instrument, such as a total station, to set- out the corners of the archaeologiarchaeological geophysics hard- drive capacity have aided archaeological geophysics in making the transition from low- resolution to high- resolution surveys. Instead of taking readings from every meter to a few meters, modern instruments can gather readings every millimeter to every few centimeters at casual walking speeds. These data can be downloaded to a field computer and initially processed in the field, quite often over lunch and again near the end of the day. This near- immediate feedback facilitates the making of adjustments to the sampling strategy and research design to maximize the quality of the data gathered and the time spent on- site. This high- resolution sampling allows for the geophysical results to not only aid in the excavation strategy, but directly address anthropological and project questions independent of excavation. Archaeological geophysical methods are routinely used in Europe, but less frequently in the United States (Watters and Ball 2001). The U.S. has fallen behind Europe, in spite of serious advances in these methods during the past two decades. Along with the hardware improvements mentioned above, there have been improvements in data- processing software (Gaffney 2008a), as well as a growing body of literature to aid practitioners in interpreting archaeological features (The English Heritage Database, last visited July 22, 2008; The North American Database of Archaeological Geophysics, last visited July 22, 2008). The infrequent use of these methods in the U.S. and, in particular, the Intermountain West may be related to the expense of equipment, which leads to a lack of training and a tendency to rely on more familiar methods (Gaffney 2001). A panel session at the 2001 meeting of the Society for American Archaeology discussed these issues at length (Watters and Ball 2001), but several years later the U.S. has yet to see significant gains in archaeological geophysical usage, which is unfortunate because archaeological geophysical surveys can cover larger areas at higher resolution and in less time than traditional methods. Additionally, landscape- scale archaeological questions can now be addressed in meaningful ways by combining archaeological geophysics with targeted archaeological excavation (Buteux et al. 2000; Conyers 2008; Kvamme 2003a). Not only can we dig deeper, we can dig broader. Archaeological Geophysical Field Methods 113

3 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 114 michael rogers cal geophysical survey grid. When possible it is useful for the archaeological survey to use the same grid system as the archaeological excavation plan. In an ideal situation, the archaeological geophysical surveys will precede archaeological testing, facilitating the establishment of a single project grid. Highprecision instruments are needed to establish the survey grid because archaeological geophysical surveys often cover much larger areas then traditional excavations and require higher precision than pedestrian surveys using shovel testing. It is common archaeological practice to avoid assigning grid coordinate 0 m, 0 m to the primary site datum (PSD) and instead to use large positive numbers in order to avoid dealing with negative values. Negative coordinates have the potential to introduce unnecessary confusion due to misspoken or incorrectly written numbers. It is easy to forget to write or read the minus sign. Abandoning the use of 0 m, 0 m for the PSD, many archaeologists choose 100 m, 100 m or 500 m, 500 m or some larger paired values for the PSD, based on the size of the study area. This method is only clear if one always attaches an E (for easting) or N (for northing) to every coordinate. This problem is compounded by the fact that in the U.S. it is convention to report latitude (the y- coordinate) and then longitude (the x- coordinate). This is precisely backwards from graphing convention. Most scientists outside of the U.S. report their easting (longitude or x- coordinate) before their northing. In the author s system, the PSD is assigned local coordinates 1,000 m, 5,000 m, and the northing is always the larger of the two numbers. This convention adds additional information to minimize communication errors. To facilitate use of this survey grid, all measurements should be referenced to the nearest southwest grid marker. By selecting the southwest marker, all measurements become additions and stay consistent with scientific graphing practice. A common practice when conducting archaeological geophysical surveys is to divide the site into 20-x-20-m subunits. Every corner of these subunits is marked by driving a plastic or wooden stake (both of which are nonmagnetic) into the ground. Nonmagnetic and nonconducting fiberglass survey tapes are stretched between the corner pegs to facilitate the laying out survey of transect lines. Blaze- orange, plastic, weed- trimmer line is very useful for marking survey transects every meter within the 20-x-20-m subunit. This material can be slightly stretched to facilitate getting straight lines and is nonmagnetic and nonconductive, easy to see in a wide range of settings, and inexpensive. 114

4 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 115 archaeological geophysics Figure 8.1. Diagram of a highly efficient bi-directional survey with two cesium magnetometers mounted horizontally on a cart. Transect lines are spaced every meter, but magnetic data are collected with a transect spacing of.25 m. Archaeological geophysical instruments can be programmed to gather data in a unidirectional or bidirectional method. During a unidirectional survey the instrument operator walks along the first transect from south to north as an example while gathering data. They then return to the beginning of the second transect while not gather data and then survey the second transect traveling from south to north, again gathering data. During a bidirectional survey the operator walks from south to north along the first transect and then from north to south, while gathering data in both directions (Figure 8.1). A bidirectional survey reduces survey time, but has the potential to introduce larger positional errors that need to be corrected for postacquisition. Most archaeological geophysical survey instruments collect data along the survey transect at a higher resolution than the transect spacing. Transect spacings are often.125 m,.25 m, or.5 m to keep resolutions sufficiently high, while keeping the survey time reasonable, and sampling along the transect ranges from every millimeter to a maximum of 50 cm. When the survey of the first subunit is completed, the data file is saved, the transect lines are moved to the next survey subunit, and that area is surveyed. Five to 10 of these subunits can be surveyed each day, depending on the number of obstacles, the difficulty in putting in the transect lines, and the number of instruments employed. With enough survey equipment, a second survey unit can be set up by the crew while the first unit is being surveyed with the first instrument (usually one of the magnetometers, to avoid magnetic contamination from the other instruments). This method allows for the most efficient use of crew time. When it becomes necessary to move transect lines from the first unit into a new survey unit, it is advantageous to 115

5 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 116 michael rogers have the next unit positioned above or below the first unit (i.e., if the first survey unit is oriented on grid north, the new survey unit should be immediately to the north or south of it). When the new unit is above or below the first unit, the transect lines can be easily pulled forward into the new unit, instead of their having to be rolled up and moved laterally. Magnetometry Human interaction with the landscape often involves digging, burning, and creating localized activity areas such as middens and cooking, butchering, living, and other activity areas that tend to leave long- lasting magnetic signatures. If any of these activities alter the distribution of iron in the soils or enhance the magnetic properties of the soils, the Earth s local magnetic field will be slightly altered. The same is true for any object that contains iron or has magnetic properties. Burning tends to enhance the magnetic properties of the soils and rocks surrounding the burning area, leading to the appearance of a localized magnetic anomaly. The moving of soils during the creation of a pithouse, storage pit, or firepit may disturb natural stratigraphic layers and create a new, mixed soil that has a different iron concentration than the original layers (Rogers et al. 2010). Magnetometers are sensitive enough to record these small changes in iron concentration (Figure 8.2). Magnetic surveys are commonly conducted using fluxgate or optically pumped magnetometers. Modern magnetometers generally record the magnitude of the Earth s local magnetic field to a precision of.1 nanotesla (nt). To put this number in context, the Earth s magnetic field has an approximate value of 53,000 nt at 45 degrees North Latitude. Subsurface features that have their own magnetic field or magnetic properties will alter the magnitude of the Earth s local magnetic field. Magnetometers are sufficiently sensitive and are able to record many subsurface geological features and human- made features (Figure 8.3). The instrument is so sensitive that before beginning a survey, the magnetometer operator removes all magnetic materials from their person to minimize contamination of the data. Zippers on pants, shoe eyelets, and even metal eye glasses will contaminate the signal (plastic sports glasses can be used to eliminate contamination from metal eye glasses). Magnetometer data can be plotted in profile or plan view, with plan- view image plots or shaded- relief plots being the most commonly used. The data are post- acquisition processed to remove any unusually high or low signals 116

6 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 117 archaeological geophysics Figure 8.2. Example of how subsurface objects alter the Earth s local magnetic field. Objects containing iron will have an induced magnetic field due to being in the Earth s magnetic field. This induced field changes the Earth s local magnetic field. (called despiking), correct for positional errors (destriping and destaggering), and gridded using an interpolation algorithm such as Krigging to enhance the resolution. Additional processing such as high- pass, low- pass, and additional filters are applied when appropriate. Fluxgate Gradiometry The fluxgate gradiometer is a hand- held device that is carried by a single operator (Figure 8.2). The most commonly used instruments are manufactured by Geoscan Research and Bartington Instruments. Inside the plastic housing are two fluxgate magnetometers that are carefully oriented to each 117

7 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 118 other and hang vertically with respect to the ground. The signal from one sensor is subtracted from the other to obtain a magnetic gradient. This is done to eliminate errors encountered due to not having the magnetometer perfectly aligned with the Earth s magnetic field, which would be impossible to achieve during an archaeological geophysical survey. The gradient also removes background effects due to the Sun and geological scale changes, leading to an enhancement of the local, near- surface signals. A fluxgate magnetometer is basically an electromagnet with an iron core wrapped with two current- carrying wires. Running current through one of the coiled wires creates a magnetic field in the core. Local variations in the Earth s magnetic field cause variations in the iron core that create a current in the second wire coil that is proportional to the local magnetic field. Rapid reading of the current in the second coil allows for rapid recording of changes in the Earth s local magnetic field as the instrument is carried along the transect at a normal walking pace (Scollar et al. 1990). Advantages of this approach and instrumentation include fast readings, high precision, light weight, moderate cost. Disadvantages are long calibramichael rogers Figure 8.3. Plan view shaded-relief plot of the magnitude of the Earth s local magnetic field acquired using a Geometrics cesium magnetometer. Excavation inset photograph shows firepits and post molds being excavated that show up as black dots in the magnetic data (just a few magnetic signals are highlighted using arrows). The region immediately outside of the excavation unit shows similar magnetic signals that are probably associated with more post molds and firepits. 118

8 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 119 archaeological geophysics Figure 8.4: Side view of the Geophysical Software Systems Inc. G858G Cesium Magnetometer with two magnetometers mounted horizontally on a cart. tion time, limited configuration modes, limited interface, and inability to use local grid coordinates. Optically Pumped Magnetometry/Gradiometry Optically pumped magnetometers can be hand- held or mounted to a cart as individual magnetometers, in vertical gradient mode (similar to a fluxgate magnetometer) or in horizontal gradient mode (Figure 8.4). The most commonly used instruments are manufactured by Geometrics Inc., Gem Systems, and Scintrex. An optically pumped magnetometer uses the fact that the energy levels of an electron in an atom (cesium is often used) become split when exposed to an external magnetic field. This splitting is proportional to the external magnetic field. Light (thus the optical portion of the name) and radio- frequency waves (the pumping part) are used to record electrons that are making the transition between these energy levels, resulting in a measurement of the magnetic field that is done quickly and with very high precision. Advantages of this approach include fast readings, high precision, light weight, an easy- to- use interface, and the ability to assign local grid coordinates with some instrument. As for disadvantages, the instrument is expensive and heavy in hand- held mode. 119

9 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 120 michael rogers Figure 8.5. Image plot of resistivity data showing the location of a subsurface stone wall, indicated by the horizontal white band. Resistivity The method of electrical resistivity was one of the earliest used for archaeological geophysical surveying (Gaffney and Gater 2003). In its simplest form, a resistivity meter is similar to a voltage meter that an electrician might use when testing electrical circuits. Two probes are pushed into the ground and the voltage between them is recorded. This voltage is related to the current and resistance via Ohm s Law, V = IR where V is the voltage, I is the current, and R is the resistance. The resistance of the material between the movable voltage probes can be calculated because the current is set by the instrument and a voltage reading is taken between the probes. Differences in the physical properties of subsurface materials can lead to differences in resistance (Figure 8.5), or how difficult it is for the electrical current to flow through the material (Clark 1990; Gaffney and Gater 2003; Scollar et al. 1990). There are two general types of resistivity meters used in archaeology, with one requiring the insertion of probes into the soil and the other using a capacitively coupled dipole system in which the probes are pulled along the ground surface. 120

10 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 121 archaeological geophysics Twin Probe Resistivity The popular Geoscan RM15 Multiplexer instrument uses probes directly inserted into the ground and has a main control unit that is used to set the grid parameters, the sampling intervals, and the fixed current. Below the control unit is a multiplexer that allows for the use of multiple acquisition probes in a wide range of configurations (e.g., Twin, Pole- Pole, Double- Diploe, Wenner, Schlumberger, and Gradient). The control unit and multiplexer are attached to a frame that also holds the current probe and one or more voltage probes. A long cable runs from the control unit out to remote probes that measure the current and a reference voltage. The operator inserts the movable probes into the soils, and the control unit gives an audible signal to inform the operator that a reading has been recorded. The RM15 can store 30,000 data points using a built in data logger. Figure 8.6 highlights the components of the Geoscan RM15 and the method of resistivity. The resistivity technique has the advantages of actively sending current into the ground, supporting multiple configurations, providing audible feedback to let the operator know that a reading is recorded, low cost, being able to use two or more instruments simultaneously to reduce survey time, and portability. Its disadvantages include having a hard- to- use interface, needing additional grid markers to move along each transect precisely, and a data- acquisition time that can be several seconds in dry, sandy soils. Dipole- Dipole Capacitively Coupled Resistivity The Geometrics Ohm- Mapper is designed for use by a single operator. Readings are taken every half- second and stored in the control console. Up to five sensors can be coupled to gather a wider range of data corresponding to different depths without additional survey time (Figure 8.7). The Geometrics Ohm- Mapper can gather data more rapidly than older systems because it uses sensor to ground contact instead of individual sensor stakes inserted into the ground. Changes in voltages are recorded as the resistivity meter is pulled along a survey transect. The voltages are then converted into resistivity readings using a modified form of Ohm s law to account for the instrument parameters. Post- acquisition processing of the data is done through Geometrics DataMap software, which allows for exporting of the data into file formats recognized by mapping programs such as Golden Software s Surfer or Geosoft s Oasis Montaj. Although this device has not seen 121

11 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 122 michael rogers Figure 8.6: The Geoscan RM15 in parallel twin probe configuration that allows for two voltage readings to be acquired at once. Probe B sends current into the ground that is recorded by remote probe M (shown as solid lines). The voltage between remote probes M and N is recorded to establish a baseline reference resistance. The voltage between probes A and B will be higher than between B and C due to the subsurface wall having a higher resistance to current flow. widespread field use, it has potential to facilitate large- scale surveys in short time periods. The device has several advantages: the operator can walk at a casual pace, multiple depth readings are possible with a multi- sensor array, the device uses the same console as the Geometrics magnetometer, and the control unit employs local grid coordinates. Its disadvantages include its cost, the fact that the multi- sensor array is over 10 m long requiring the use of longer transects, and the lack of example surveys due to it being a relatively new instrument. 122

12 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 123 archaeological geophysics Figure 8.7. The Geometrics OhmMapper with one transmitter and one receiver. Ground- Penetrating Radar Increased computer processing and more efficient power usage have made field use of Ground- penetrating Radar (GPR) more viable for archaeological projects. Large amounts of data can be gathered, digitally stored, initially processed while still in the field, and transported to the laboratory for indepth analysis. Even with this processing power, the success of a GPR survey in identifying subsurface features is very dependent upon the material through which the radar signal is travelling. When the soils at a site are conducive to propagating a radar wave, a GPR survey can yield information about the three- dimensional spatial position of stratigraphic layers and subsurface features. A GPR survey records the reflection of the radar wave from a subsurface object and, because many different objects can have the same reflection, it is often difficult to know precisely what object created an observed reflection. For correctly interpreting the data, all of the following are helpful: results from other GPR surveys, computer modeling, knowledge of the site, knowledge of possible features, and experience. GPR is sensitive to a range of physical and chemical properties of materials. Electrical conductivity, magnetic permeability, and the relative dielectric 123

13 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 124 michael rogers Figure 8.8. Profile of a possible pithouse at the Springs Preserve, Las Vegas highlighted by the white box. There is a distinct 3-4 m long dip in the profile that appears consistent with the expected signal of a pithouse floor. There is a large bush in this area, and the dip may be caused by dirt piling around the base of the bush. permittivity of materials determine the speed of the radio wave in the soil and the intensity of the wave reflected from and transmitted through an interface between two materials (Conyers and Goodman 1997). The movement of soil during the creation of houses and pithouses creates new soil interfaces and stratigraphy. If the relevant properties of the fill soils and undisturbed soils are different from those of the materials on the house or pithouse living surface, the radar wave will reflect from these interfaces (Figure 8.8). The amplitude of the reflected wave will be larger if the difference between the relevant properties is larger. The depth into the soil and the size of the interfaces that can be detected are dependent upon the frequency of the antenna being used; the higher the frequency, the shallower the penetration, but the smaller the objects that can be detected. The primary components of a standard GPR device are the control unit, the transmitting antenna, the receiving antenna, the antenna cable, the display, and the power supply. Figure 8.9 shows a Geophysical Survey Systems, Inc. SIR-3000 GPR with a 400 MHz antenna (Sensors and Software, Inc. also manufactures radar units commonly used in archaeology). A typical GPR antenna transmits 25,000 to 50,000 radar pulses per second into the ground (Conyers 2004a). Due to the enormous number of pulses being transmitted in such a short time, the control unit must sample the reflected wave data. The SIR-3000 r control unit allows the user to set the samples per scan to 128, 256, 512, 1024, or The higher the number, the more data points per pulse (often called scan or trace), but the larger the data file. 124

14 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 125 archaeological geophysics Figure 8.9. Geophysical Survey Systems, Inc. SIR-3000 GPR with a 400 MHz antenna and distance wheel. The solid line shows how the radar wave reflects multiple times when it encounters a feature such as a firepit. The dashed line shows radar wave reflection and transmission when encountering interfaces. For archaeological applications where the features of interest lie 2-3 m beneath the surface, 512 samples per scan is often sufficient (Conyers 2004a). Another adjustable parameter that is linked to the samples- per- scan setting is the range (or time window). By increasing the range, radar signals from deeper into the ground will be recorded, but the trade- off is less resolution due to the samples- per- scan setting. For example, if the samples per scan is set to 512 and the range is 64 ns, then eight data points will be recorded every nanosecond. If the range is increased from 64 to 128 ns and 125

15 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 126 michael rogers the samples per scan remains at 512, only four data points are recorded for every nanosecond. It is advantageous to set the range to an appropriate value for a particular study area to maximize the amount of data recorded per nanosecond and keep the data files a manageable size. There are two more adjustable parameters that are essential to keeping data files manageable. The bits per sample can be set to 8 or 16. The benefit of selecting 16 bits per sample is better resolution of smaller objects, but this setting doubles the file size. When each 20 m long transect can produce files 1 2 Megabytes in size using a setting of 8 bits per sample, careful consideration must be given to the appropriate bits per sample setting. With the recent development of large and inexpensive hard drives, it is now easier to record 16 bits per sample in the field. The user can also set the scans per second to 16, 24, 32, 48, or 64. This setting is used when the GPR survey is conducted in continuous mode. Again, because of the enormous number of pulses per second, the control unit must also sample the number of scans recorded. The scans per unit is a horizontal sampling of the data, whereas the samples per scan is a vertical sampling of the data. The scans per unit is based on the number of turns of the distance wheel. Recording more horizontal points per unit will require the operator to walk more slowly. This parameter is set based on the expected horizontal size of the features and the amount of time available to the project. A final, important setting is the gain. As the radar wave travels deeper into the ground, the amplitude of the wave decreases. This decrease is due to reflections from interfaces and attenuation due to travelling through different media. Under ideal conditions, identical objects at the same depth in the same material generate reflected signals with the same amplitude. Now move one of the identical objects deeper into the ground, and that object will have a smaller reflection amplitude due to the radar wave travelling though more soil. If this object is moved even deeper into the ground, the reflection amplitude may get so small that it cannot be identified. By applying gains to the signal, the amplitude of the reflection from the deeper object is boosted to facilitate better identification. Gain adjustments can be made at varying depths and with varying intensity. Additional filters are available that can be applied during or after the survey. Low Pass and High Pass Vertical Frequency filters, Horizontal Smoothing, and Horizontal Background Filters assist in removing unwanted noise from the data. 126

16 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 127 archaeological geophysics Figure Two types of conductivity meters: (left) Geonics EM31 and (right) GSSI EMP-400. The advantages of GPR include the facts that 3-dimensional images are possible, depth information is provided, the resolution is high, and the interface is easy to use. Disadvantages are that the approach is expensive, the data files are large, and much time can be spent on post- acquisition processing. Electrical Conductivity Conductivity meters use an alternating magnetic field to induce a current in the subsurface materials that leads to a secondary electromagnetic wave, which is recorded by a receiver. The magnitude of the secondary magnetic field is related to the conductivity of the materials. Conductivity is the inverse of resistivity and a measure of how easy it is to pass a current through a material. Due to the fact the method of measuring conductivity (or inverse resistivity) differs from resistivity methods, the results are different and complementary to resistivity surveys (Clark 1990; Scollar et al. 1990; Sharma 1996). Geonics, Inc. has been a longtime manufacturer of conductivity meters for archaeological applications. Geophysical Survey Systems Inc. (well known for their GPR systems) has recently produced a conductivity meter (Figure 8.10). A conductivity meter is operated by a single operator who carries the instrument, batteries that provide 20 hours of continuous operation, and a control console with data logger. Several readings are taken per second as the operator walks continuously along a transect. Conductivity meters do not need contact with the ground to take readings, which simplifies operation, but the overall size of the Geonics EM-31 meter can be problematic in areas with many vertical obstacles (such as trees). The Geonics EM-38 and the GSSI EMP-400 are more portable. The length of the instrument is related to the separation between the transmission and receiving coils. The larger the spacing the greater the depth being measured (Clay 2006*). 127

17 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 128 michael rogers Figure Magnetic susceptibility survey conducted at the Clonmacnoise Bridge Site, Ireland by the author and Kevin Barton (National University of Ireland-Galway). The magnetic susceptibility survey identified easy-to-magnetize material along a Norman castle s defensive earthen embankment. Iron slag used during construction appears to be the source of the signal. The conductivity meter has the advantages of being generally easy to maneuver along a transect, of measuring both conductivity and magnetic susceptibility at the same time, and of working well in a range of environments (wet and dry). As for disadvantages, it is sensitive to near surface and subject to inference from power lines and other electrical sources; also, depth is difficult to distinguish, and some instruments are hard to use at sites with vertical obstacles. Magnetic Susceptibility The most common stand- alone magnetic susceptibility meter is manufactured by Bartington Inc. The Bartington magnetic susceptibility meter (Figure 8.11) allows for the field- gathering of data on the magnetic susceptibility of near- surface soils (first few centimeters) and for laboratory analysis of soils, rocks, and other artifacts to better understand their magnetic properties. Magnetic susceptibility meters use an alternating current in a wire coil to create an alternating magnetic field. The frequency of the field is altered 128

18 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 129 archaeological geophysics due to its interaction with the soils. The change in frequency is related to the magnetic susceptibility of the materials (Clark 1996*; Gaffney and Gater 2003; Scollar et al. 1990). A single operator carries the Bartington magnetic susceptibility meter with a single reading being recorded in a few seconds which is slower than data collection with instruments such as GPR and resistivity meters. The instrument is therefore used most effectively when instruments with more rapid data acquisition have identified small areas of interest. The Bartington instrument measures the magnetic susceptibility of only the first 10 cm or so of the soil. Even with the shallow penetration, the active method facilitates data collection of the near surface where passive instruments like a magnetometer work poorly. Another strength of the Bartington magnetic susceptibility meter is the ability to convert the instrument from a field sensor to a laboratory sensor to make measurements on rocks, soils, and artifacts. The magnetic susceptibility meter has several advantages: it can be used in both the field and the laboratory for understanding magnetic properties of materials, and it can be used to measure exposed excavation profiles and features. The instruments disadvantages include its slow data acquisition, outdated interface, and shallow (only 10 cm) depth of penetration. Variables that Influence a Successful Archaeological Geophysical Survey Ultimately, and regardless of instrument being used, archaeological geophysical surveys measure contrasts between the geophysical properties of the materials that are present. If there is little to no contrast in those properties, archaeological geophysical surveying will not be successful in producing usable images. It is unlikely there will be no contrast in all properties, and thus the advantage to using multiple instruments or selecting the appropriate instrument or instruments based on a firm understanding of the study site (Table 8.1). Having stratigraphic layers with differing soil types and archaeological features made of materials that differ from the surrounding soil matrix creates the necessary contrast for these techniques to work. Creating a firepit involves digging that can cut through soil layers. Rocks may be used to line the pit. Over time the pit is filled with ash, charcoal, debris, and windblown soil that will generally be more organic than the subsurface soils. All of these 129

19 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 130 michael rogers Table 8.1. Summary of the Geophysical Properties Measured by Each Instrument. Instrument Geophysical Property Archaeological Implications Conductivity Ease of sending electrical current Good at detecting landscape-scale (also measures how easy it is to changes in soil types and human magnetize) modification of these soils. Works when wet or dry. GPR Ability to transmit radio wave Effectively maps stratigraphy and human modification of the stratigraphy. Only instrument to give a direct measure of feature depth. Magnetometry Magnetic field strength Humans tend to leave long-lasting magnetic signatures due to burning, digging through soil layers, and use of iron Magnetic How easy it is to magnetize Essentially measures iron Susceptibility concentration, which is useful at historic sites Resistivity Difficulty in sending electrical Tends to work well when soils are wet current and GPR may not work. Able to image a wide range of features. elements lead to contrasting geophysical properties. If a firepit that was cut into only one soil type becomes filled with that same soil, the contrast between the firepit and the surrounding soils will be low. The sediment filling the firepit will generally be less compact then the surrounding soils, but archaeological geophysical methods do not directly measure density. Instead, the less compact or less dense materials will preferentially retain or release water relative to the adjacent compact materials/soils. This preferential retention of water has been shown to influence the outcome of archaeological geophysical surveys (Conyers 2004b). Additionally, wet clay tends to attenuate ground- penetrating radar signals in just a few centimeters, but is also highly conductive of electricity. Each instrument has a maximum depth of effectiveness, with signal loss generally occurring with increasing depth. Some instruments have components (such as different GPR antennas) that allow the operator to select the appropriate depth for the project. Certain parameters of each instrument (Table 8.2) determine maximum effective depth, but soil types and conditions also play a strong role. It is helpful to have some sense ahead of time of the soil conditions and depth of features to facilitate selecting the proper instrument and configuration. 130

20 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 131 archaeological geophysics Table 8.2. Effective Depth of Investigation for each Type of Instrument Instrument Key Instrument Variable Depth Conductivity Distance between 1 to 2 times the transmitter-receiver transmitter and receiver spacing GPR Frequency of antenna 900 MHz 1 m, 400 MHz 4 m, 200 MHz 9 m Magnetometry Distance from ground Magnetometers passively measure the magnitude of the Earth s local magnetic field at the location of the sensor. Deep objects with large magnetic fields can be recorded, whereas very near surface objects with small magnetic fields may not appear in the data. Magnetic Size of the coil / From just a few millimeters to Susceptibility surface area approximately 10 cm Resistivity Distance between current Maximum of 1 to 1.5 times the and voltage probe probe separation Sampling interval is another key factor that influences successful surveys. Both the distance between transects and the number of readings taken along each transect must be appropriate to sufficiently resolve the geophysical signal of features of interest. Keep in mind that it is the size of the geophysical signal and not the size of the feature that is important. Early in the development of archaeological geophysics, sampling intervals of 1 m or more were not uncommon. Sampling intervals have decreased to.50 m and quite often smaller as instruments and computing power have improved. There is always a tension between selecting a sampling interval to produce the highest resolution and completing the survey in a timely fashion. Experience and knowledge of the site can help in finding the appropriate middle ground. The sampling interval along the transect is often dictated by how fast the instrument can collect data and how quickly one wants to complete each transect. The Geometrics cesium magnetometer can take 10 readings every second in continuous mode. At a casual walking pace, the end of 20- m- long transect is reached in seconds, yielding a reading every 5 10 cm. Ground- penetrating radar instruments collect data rapidly, allowing for a reading every millimeter or so along the transect. Resistivity meters requiring the insertion of probes and the Bartington magnetic susceptibility meter can take several seconds for each reading. It is not uncommon to use a spacing of 50 cm along the transect with instruments that take several seconds to record a single reading. 131

21 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 132 michael rogers The distance between transects should also be small, but is restricted by the need to have some way for the operator to keep straight on the transect. As mentioned in the general field methods section, some form of string, survey tape, or weed- trimmer line can be used to mark transects. It is not practical to have a transect spacing of 5 cm, to mark every transect on the ground, and to complete the survey in a timely fashion. If a project does in fact demand this degree of resolution, one must commit the time for such an intensive grid setup. More commonly, one sees transect lines stretched at intervals of.5 m or 1.0 m, with the operator heading north along the first line and then south between that line and the next, which marks the location of the third transect. Figure 8.1 shows how two cesium magnetometers mounted horizontally on a cart can sample two transects at once, allowing for wider spacing between the lines that mark the transects. Procedures that combine widely spaced lines and more narrowly spaced transects are desirable due to the time- intensive setup, movement, and take- down of the lines. Current research is examining the use of differentially corrected GPS in order to reduce the need for lines marking the locations of the transects (Barratt et al. 2000). A final key influence on the success of archaeological geophysical surveys is careful coordination between geophysical survey techniques and archaeological research questions. A useful starting point is to have a research design that uses the same grid system for all work at a site. This allows for the production of plots of the geophysical data that are easily read by anyone on the project team. Interpreting geophysical data is challenging because the actual object cannot be directly examined. Plots can be created while still in the field, but post- acquisition processing will often be necessary to obtain the best view of the data. Some surveys result in planview plots that are quite easy to interpret, whereas others may be difficult or impossible to interpret. As the project continues, it is important to consider the archaeological geophysical survey as part of the overall research effort. Additional information obtained from the site after an initial interpretation of the geophysical data may warrant reexamination of the geophysical data. A better understanding of how to interpret geophysical data will emerge from this cyclic process. Past Successes and Future Prospects for the Intermountain West Several dozen successful archaeological geophysical surveys have been conducted in the Intermountain West. Many of these surveys were done just 132

22 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 133 archaeological geophysics Table 8.3. Summary Description of Archaeological Geophysical Surveys Site* Name Investigator Instruments Features Imaged 1 Escondido Pueblo site Ernenwein (2006); GPR, Mag, Sus, Pithouse, adobe walls, garden plots Ernenwein and Kvamme (2008) Cond 2 La Gila Encantada site Rogers Mag Pithouses, hearths 3 Terry Canyon Village site Rogers GPR Pithouses 4 Cobble Pueblo site Abbott, Roxlau, Butler GPR, Mag, Res Not conclusive 5 Cloudblower site Abbott, Roxlau, Butler GPR, Mag, Res Not conclusive 6 Adobe Pueblo site Abbott, Roxlau, Butler GPR, Mag, Res Not conclusive 7 Archaic site Abbott, Roxlau, Butler GPR, Mag, Res Not conclusive 8 Camp Lewis (Pecos National Historic Park) DeVore Mag Santa Fe Trail 9 San Marcos Pueblo historic mission Archaeo-Physics, LLC Mag, Res Church and convent ruins 10 Chaco Canyon (Chetro Ketl) Loose, Lyons Mag Garden plots 11 Aztec Ruins National Monument Johnson Mag, Res Six room structures 12 Mesa Verde Lancaster, Watson Pithouses, post holes 13 Shield site Conyers and Cameron (1998) GPR Not conclusive 14 Bluff sites Conyers and Cameron (1998) GPR Pithouses, kiva 15 Vaughn site Conyers and Cameron (1998) GPR Not conclusive 16 Cottonwood Falls site Conyers and Cameron (1998) GPR Not conclusive 17 Fort Crawford Cemetery Charles Mag, Res Burials, ditches, roads 18 Sieber Flats Kvamme Mag Hearths 19 Black Canyon Butler Sus Not conclusive 20 Site 5MT2848 Weymouth Mag Pithouse 21 Kaplan-Hoover (5LR3953) Kvamme GPR, Res Arroyo 22 Fort Laramie National Historic Site Somers Mag, Res Fort 23 Seminoe s Fort DeVore Mag Trading post, trash dump 24 Fort Caspar (48NA209) DeVore Mag, Cond Fort 25 Fish Creek Ranch Butler, Cannon, DeVore Mag, Res, Sus Pithouses 26 Bighorn Canyon National Recreation Area DeVore Res, Cond, Res Burials 27 Sand Draw Dump site (48FR3123) DeVore Mag Hearths 28 Aspen site (Craters of the Moon National Monument) DeVore Res Roads 29 Sites 42TO1009, 42BO681, and 42BO683 (Hill AFB) DeVore Mag, Cond Not conclusive 30 Springs Preserve Rogers GPR, Mag, Res Pithouses *see Figure

23 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 134 michael rogers Figure Locations of archaeological geophysical surveys conducted in the Intermountain West (Table 8.3). This is a representation of the range of surveys already conducted and not meant to be an exhaustive list. west of the Rocky Mountains and, as a group, involved the use of a wide range of instruments. Table 8.3 and Figure 8.12 highlight a few surveys that were identified using the North American Database of Archaeological Geophysics ( and a basic literature search using references cited in publications and found on Google Scholar. The North American Database of Archaeological Geophysics provides an excellent resource for examining representative data plots of features such as pithouses, hearths, firepits, burials, and historic structures. The projects presented in table and figure should not be considered an exhaustive list, but rather sufficient evidence that archaeological geophysical surveys can successfully image features of interest in the Intermountain West. It is important to remember that sometimes surface survey, surface scraping, shovel testing, backhoe testing, excavation, and geophysical survey either do not work in a given research context or produce results that are confusing to interpret. There are no magic methods, and the tricorders depicted in the TV show Star Trek really are the stuff of science fiction. Archaeological geophysical methods have, however, seen significant advances 134

24 *ch08 Rogers_Bradford+Linn 1/3/11 8:39 AM Page 135 archaeological geophysics in the past decade that make the instruments far more portable and effective (Jordan 2009). The future prospects for the Intermountain West are encouraging, based on the range of existing successful surveys. As more surveys are done especially when intentionally and thoughtfully integrated with other archaeological methods the quality and success rate of the integrated archaeological methods to answer the questions posed will increase. 135