Previous pageTable of ContentsNext Page

Interface--

Archaeology and Technology

Geophysics, Ground-Penetrating
Radar, and Archaeology

Lawrence B. Conyers

 

Archaeology has traditionally been a multidisciplinary endeavor. Most of the major advancements of the last 50 years, such as radiocarbon and other dating methods, palynology, bone chemistry and trace element analysis, to name just a few, were incorporated into archaeology directly from other disciplines. Before they became accepted all were tried, tested and then retested at many different sites before gaining acceptance by the archaeological community. Geophysical methods are being subjected to this trial-and-error period. Total acceptance by the archaeological community is still pending.

Geophysical techniques for subsurface mapping first started to be applied to archaeological projects in the 1950s. Today many field projects are incorporating geophysics as not only an exploration tool but also for the creation of predictive models prior to excavation and for mapping the extent and spatial distribution of buried features that may never be excavated due to funding limitations and preservation concerns. For the most part, however, geophysical techniques are not commonly employed in North American archaeology because of their expense, perceived limitations, lack of well-published successes, or a reluctance by some archaeologists to use "complicated high-tech equipment" they don't understand. Geophysical archaeology must overcome these and other hurdles if it is to be taken seriously and become more widely applied.

Common Geophysical Methods

The most commonly used geophysical methods used in archaeology are electromagnetic methods (EM), soil resistivity, and magnetometry. The EM and resistivity methods induce a field (electromagnetic or electrical) into the ground and then measure how the medium responds. As the sensors are moved across the ground in transects, changes in the induced field can be related to the presence or absence of buried archaeological features or geological changes. When gridded in map form, many buried features are visible. Magnetometry is a passive method which measures changes in the ambient magnetic field of the earth over an area due to physical and chemical changes in near-surface material, which can sometimes relate to buried materials or very near-surface archaeological features that are invisible to the human eye. These geophysical tools collect data quite rapidly and can quickly produce useful maps, but cannot accurately measure the depth of buried features. In contrast, ground-penetrating radar (GPR), which measures the reflection of radar pulses transmitted into the ground, can measure both the depth and spatial extent of buried features. For archaeological mapping in some environments, GPR is rapidly gaining acceptance as its successes are demonstrated and data collection and interpretation techniques improve.

The GPR Method

The GPR method uses surface antennas to transmit pulses of radar energy into the ground, which are then reflected off stratigraphic interfaces and buried archaeological features. Reflected waves are received back at the surface and recorded digitally. Many reflections from different depths can be recorded at any one location on the ground, and as antennas are moved along transects in a grid, many thousands can be recorded. Ground-penetrating radar surveys record massive amounts of digital data, making this type of data set one of the more difficult to collect and interpret. Its effectiveness lies in the ability to collect data in three dimensions, producing images of buried archaeological features that are very precise both spatially and in depth.

In archaeological GPR mapping, a 50-x-50-m survey (consisting of about 100 individual transects) can be collected in about three to four hours. A survey of this type can yield more than 400 megabytes of digital data. Prior to about five years ago, these data would have been processed and interpreted transect-by-transect, and then interpreted manually. This process could take months per grid, making it infeasible or extremely costly for most projects and budgets. Today, processing and interpretation of these types of data sets can be accomplished in a few hours at most, and usable images of the subsurface can sometimes be created while still in the field. More interpretation time may of course be necessary if subsurface conditions are very complex. When renting equipment and paying consultants by the day or hour, this time saving has now put GPR methods within the range of many archaeological projects.

A Recent Successful GPR Survey

A GPR survey was recently conducted at the Bluff Greathouse site in southeastern Utah. The survey was centered over a subtle surface depression that was considered a possible great kiva. It was done to produce maps that could help in planning future excavations by the University of Colorado-Boulder field school. Individual reflection profiles showed the kiva feature as a bowl-shaped depression incised into bedrock and filled with nonreflective sediment, but little could be visualized within the kiva in the reflection profiles (Figure 1). In an attempt to create images of features within the kiva, the reflection data were processed in
Ground Surface
Interior of kiva
Bedrock reflections

Figure 1 -- Great Kiva reflection profile

amplitude time-slices, which creates a series of horizontal images of the ground similar to arbitrary levels in a standard excavation. Subtle changes in reflection amplitudes at given depths were then plotted and mapped. When this was done, one of the upper slice showed the outer, partially collapsed wall of the kiva and a deeper slice revealed a "squarish" feature within the kiva (Figure 2). Two test trenches were then excavated that confirmed the presence of both the shallow outer wall and the deeper inner feature, whose origin is still being debated.

 

Reflections from .5­.75 m
Outer wall of Great Kiva

Reflections from 1.25­1.5 m
Inner wall of Great Kiva

Figure 2

A Potential GPR Failure

The potential to produce a glaring failure with GPR surveying was recently made clear to me at another archaeological site in southeastern Utah. By visually interpreting GPR reflection profiles, I discovered two pit-structures, buried by wind-blown sand and alluvium, near the San Juan River. The initial survey was conducted after four very dry months. This was a good time to do the survey because water, which is differentially retained in sediments, can sometimes severely disrupt GPR surveys because moisture boundaries can cause anomalous reflections that have little relationship to the buried geological or archaeological features of interest. Excellent reflection data from the pit house floors were recorded in my initial survey because what little moisture remained in the ground was probably ponded on the pit house floors, creating an excellent reflection surface (Figure 3). One of the pit structures discovered by GPR was later confirmed by augering.

A few months later, the survey was repeated, but this time a few days after about 2 inches of rain had fallen. The reflection data from this survey showed little indications of the pit structure floors, but abundant anomalous pockets of water that had been differentially retained in sediment layers (Figure 3). If my initial survey had been conducted after the rain in this case, it would have created images of nothing but water pockets. Instead of conducting the geophysical survey after the rain storm, it would probably have been prudent to postpone it for at least a few days to give the water time to percolate through the ground.

The Future of GPR and Other Geophysical Surveying Methods

In the last 10 to 15 years, there has been a small group of geophsicists and technically trained archaeologists that have been applying GPR techniques (and other methods) to archaeological sites around the world. Some surveys have been successful in revealing archaeological features, and others less so. The GPR surveys that most hear about, and are published, appear to be almost
Dry soil Soil after rain

Figure 3 -- Floor of pit-structure

uniformly successful and they yield important information about archaeological sites that could not be acquired in any other fashion. But there are probably an equal number of surveys that are at least initially judged to be failures by those involved due to problems similar to mine with rain water at the buried pit-structure.

One potential problem in employing any new technology is a perception by new enthusiasts that "high-tech" methods can do almost everything an archaeologist desires. This misperception can be perpetuated because only geophysical successes are published or widely discussed and many experienced geophysicists are reluctant to say their techniques are not appropriate in some areas, especially if they do not completely understand the site conditions in advance. Fortunately, geophysical successes are becoming common as we learn more about how to apply different techniques to archaeological problems. The successes need to be critically evaluated, studied, and published. Unfortunately, information about unproductive surveys, which is rarely published, is transmitted by a different and potentially more destructive method: word of mouth. If the person doing the survey has not attempted to address why a particular method yielded poor results, or if reasons for failure are known but have not been adequately communicated to all concerned, the method receives a "black eye" that can be hard to recover from, as more people hear the story.

Why So Few Geophysical Surveys?

Although geophysical surveys are becoming more common, there is still a general reluctance to use these methods in archaeology for a number of reasons. One may be due to the way most archaeologists are trained. Many of our college instructors were traditional field archaeologists who have a perception of what are "real" archaeological data (artifacts and architectural features) and what are "suspect" data (information or maps produced from processed digital data). Typical anthropology programs offer little in the way of computer and other technical field methods classes, and little encouragement is given to students wishing to take geophysics courses in other departments. As a result, geophysical techniques (especially GPR) are quite intimidating to many archaeologists, and are often looked at as "black box" methods where strange equipment is moved around on the ground and mysterious images, which are difficult to interpret, are produced.

In today's archaeology, the practitioners of what appear to be obscure geophysical techniques are usually not archaeologists, but consultants with nonarchaeological backgrounds who are brought in to do a quick survey, produce some results (that may or may not be useful), and then leave. There is often poor communication between the anthropologically trained field archaeologists and the technically trained geophysicists, which can often result in a "clash of cultures" (or at least a moderate chasm in understanding) between the two. The geophysicist often is unsure what the archaeologist really wants to map and may have a limited understanding about site conditions. The archaeologist often lacks the geophysical background to appreciate the nature and limitations of the method. Thus the geophysical data output produced in the field can be confusing, at best, for both.

Even worse if a consulting geophysicist tells an archaeologist to "dig here" because "I found an anomaly" and then after excavating "nothing is found" by the archaeologist, what little intellectual capital the geophysicist might have earned can be squandered. (I am always puzzled when those conducting excavations tell me that "nothing was found". . . surely "something" was uncovered-at least soil changes, moisture differences, stratigraphic boundaries, or other features that might be useful to the geophysicist in the interpretation process).

One Approach to Advance a Geophysical Archaeological Understanding

In an attempt to make geophysics more "real," or at least more understandable, a geophysical archaeology test facility was constructed in Champaign, Illinois. Built by the U.S. Army Corps of Engineers at their Construction Engineering Research Laboratory, with funds from the National Center for Preservation Technology and Training, a model archaeological site was constructed and then buried, plowed, and replanted. In a 50-x-50-m test area house floors of different construction materials, storage pits, hearths, ditch and wall features, and even pig carcasses (to simulate human burials) were put in the ground according to a precisely mapped plan. Archaeological geophysicists can use this site as a test of different geophysical methods under varying soil conditions and field procedures.

I conducted a GPR survey over the entire test site in October 1998. My initial perception of the data, as they were being acquired, indicated the survey would likely be a failure. Recent rains had saturated the ground and the high clay content of the soil appeared to be severely attenuating the radar energy before it could be transmitted to any significant depth. In a normal setting, I would have been tempted to give up and go home. Extremely complex reflections appeared to be all but worthless using traditional interpretation techniques because the human brain is not capable of processing huge amounts of random reflections within a three-dimensional volume.

In an attempt to create images of what I knew was there, these reflection data were processed using the amplitude slice-map method, in the same way as the data from the great kiva in Utah. Subtle changes in reflection amplitudes, which were not visible in individual profiles were then mapped, which could be directly correlated to what was known in the subsurface. The amplitude map of radar reflections collected at the depth of four house floors is particularly interesting (Figure 4). Although none of this information was visible in individual transects, the slice-map showed each individual floor, some of the standing walls, hearths, storage pits, and a number of other features. And they were mapped at precisely the correct burial depth. This survey vividly illustrates the power of collecting and processing three-dimensional data, even if it "looks bad" in the field. Sometimes the computer is capable of doing what the human brain cannot by combining information from adjacent lines and processing subtle changes that are present in digital data, but not immediately visible to the human eye. Hopefully, this test site will be more widely used for research and training of archaeologists in the future.

The Future of Geophysical Archaeology

New acquisition, processing, and interpretation techniques in GPR, as well as those of other geophysical methods, have the potential to change the way field archaeology is done. As geophysical field techniques are refined and processing and interpretation methods are experimented with and perfected,



Figure 4

archaeologically useful data sets will become the norm. Time must always be spent, however, with the data from perceived failures so that they can be understood, or maybe reprocessed and interpreted in different ways.

Geophysics is slowly being incorporated into research designs as a new generation of archaeologists (many of whom were computer literate almost from grade school) are entering the field. They are still, however, obtaining most of their archaeological training in traditional anthropology departments that rarely incorporate geophysical methods into field methods courses or as part of a field school experience.

For geophysics to become more widely accepted, more surveys must be attempted, the results published and methodological approaches repeated and modified for different conditions. Geophysical results should be tested by excavation, and what was uncovered (whether it was what was predicted by geophysics or not) must be studied and compared to the processed data. What appear to be marginally successful surveys should be evaluated not so much as failures, but as part of a learning experience that can hopefully be amplified in the future. Using this approach geophysics will gain in stature as a legitimate way to collect archaeological data. I hope that someday soon it will become not just a common field technique, but one that is universally accepted as standard practice by the archaeological community.

Much the GPR research discussed in this paper was funded by the National Center for Preservation Technology and Training. ·

Lawrence B. Conyers is an assistant professor in the Department of Anthropology at the University of Denver.

References:

Conyers, L. B., and D. Goodman
1997 Ground-Penetrating Radar: An Introduction for Archaeologists.
AltaMira Press, Walnut Creek, CA.

Conyers, L. B., and C. M. Cameron
1998 Ground-Penetrating Radar Techniques and Three-dimensional Mapping in the American Southwest.
Journal of Field Archaeology
25( 4): 417­430.

back to top of page


Previous pageTable of ContentsNext Page