Archaeology: Geophysical Investigations: Additional Information
The Place of Geophysical Survey in Contemporary Fieldwork
R. Berle Clay
Cultural Resource Analysts, Inc.
Introduction
Near-surface geophysical survey techniques are being increasingly used in archaeology for a variety of reasons and there is every indication that the trend will continue as they gain greater acceptance from fieldworkers and the preservation world. The following addresses two points. First, given a variety of fieldwork demands, what sort of survey appears to be the most useful? Note, this question is not, which geophysical technique is the best? Rather, can we define meaningful, generic survey parameters, regardless of technique? Secondly, how do these relatively new techniques relate to traditional field techniques in archaeological research because geophysical survey techniques clearly cannot replace traditional techniques?
Geophysical Survey At The Phase I and II Level
Near-surface geophysical survey techniques are rapid, non-destructive techniques for identifying "structure" in the ground. They are ideally suited to Phase I and II fieldwork as a substitute for strip plowing, general plowing or disking, or dozer/pan "scrapes." Especially where survey and evaluation occur well in advance of mitigation and construction, these techniques provide an attractive alternative (from the standpoint of the landowner) to a badly damaged open field. In addition to this advantage, they are able to overcome some of the problems of sampling posed by systematic shovel testing coverage (from the standpoint of the preservation official), and are considerably faster--therefore cheaper (from the standpoint of the applicant and the CRM firm)--than shovel testing.
Ideally a geophysical survey should be able to place an archaeological "site" in a larger context of the "non-site" which surrounds it. This is important to remember because often archaeologists tend to think of "remote sensing" in terms of using metal detectors to "find" something (a task which metal detectors do very well, making them highly useful in certain fieldwork, for example "battlefield archaeology"). This means that a Phase I-II level geophysical survey should cover as large an area as possible, in effect placing the archaeological site in a local landscape. To do this they must be cost-effective. For example, British practice (where geophysical techniques have been widely integrated into CRM work) stresses survey areas of up to 5 acres as a minimum. Because of this the emphasis now is on high-speed, high-density survey techniques.
For this reason, at the Phase I-II level that old standby, resistivity survey, one of the earliest survey techniques to be developed, has become less cost effective because the technology must make electrical contact with the ground through probes. This is slow. The emphasis now is on magnetic and electromagnetic techniques that require no electrical contact with the ground and are continuous-reading (or nearly so). Where data recording is convenient and hand- or automatically-triggered, and where data can be downloaded for computer processing, cost can be reduced.
Data density is essentially the number of readings per unit area. High-density, overlapping geophysical readings, whatever the specific technology, tend to produce data that are more easily understood. If data recording is automatically triggered, density has become simply a product of how fast the technology can be triggered (some magnetic technologies can, for example, be triggered 10 times a second). Not so long ago data density was limited by computer capacity. But the days when you cranked up SURFER, then went off to lunch and looked for a grid file in the early afternoon, are passed with today's high-speed computer chips, ample memories, and commodious storage devices.
It used to be a rule of thumb that data density should be determined by the size of the expected target (for example, don't take readings two meters apart to detect a pit which is 50 cm in diameter). With automatic triggering this has become somewhat a non-issue, in other words it is now possible to take low interval recordings as a matter of course. They may not in fact be required by the nature of the Phase I-II survey, but if they can be done, then do them. However, there is at least one note of caution. Most geophysical survey technologies require that the instrumentation be held at a constant level above ground surface. If this is difficult in a given survey for whatever reason (high grass, fallen timber, trees, etc., anything which hinders walking), then a high density of readings may introduce considerable "noise" produced by the act of measurement as the sensor moves up and down despite the best intentions of the operator.
Putting these requirements together, Phase I-II geophysical survey requires a magnetic or electromagnetic technology capable of continuous readout and automatic triggering. The emphasis will be on high-density data sets, tempered by the need to maintain the quality of the data given varying survey conditions. Here at Cultural Resource Analysts, electromagnetic survey typically aims at data recording in 50 cm intervals along transects one-meter apart. Although this does produce an "asymmetrical" data set, the speed of ground coverage makes it attractive from a cost standpoint.
For a 20 meter data collecting square (in a sense a "standard" collecting unit from a phase in computer technology when a data file was limited in size), this gives a density of 861 readings per square. At the moment, daily data acquisition is limited by how fast the equipment operator can walk. This translates into survey time for a 20 meter square of between 13 and 15 minutes (partly dependent upon survey conditions, partly on assistance at moving marker lines). In theory it should be possible to cover 1 hectare in 6 to 7 hours. However, this would involve considerable wear and tear on the operator (over 10 km of walking in a day and intense concentration while doing it) and considerable boredom on the part of assistants moving the marker lines! In reality, something like 12-13, 20 meter squares a day is a better goal, perhaps one-half a hectare a day. Data acquisition rates are being expanded today by mounting multiple sensors on the operator, linking data acquisition with GPS technology freeing the operator of measuring ropes (and assistants), or doing away with the walking operator altogether by mounting an array on a sled or cart towed behind a vehicle. At present the cost of these alternatives goes up rapidly with their capacity.
Finally, and a question which is probably uppermost in the thinking of most archaeologists, which specific technique should be used in a Phase I-II survey? It goes without question that the different techniques are all "valid." For a host reasons, one technique may "work" in one context while another does not. There are even cases where no geophysical technique can produce meaningful results (from my experience a minority of cases, although this question is linked to that other question, what did you want to find out?). Different techniques may produce somewhat different results from the same geophysical context and this is related to the particular technology.
As a rule the archaeologist can count on vastly increasing his/her understanding of a particular survey area by using multiple survey techniques. These have the effect of multiplying the cost of data acquisition. However, if the techniques are high-speed/low cost to begin with, it is possible to build into fieldwork budgets multiple cycles of data collecting using different technologies. Here at Cultural Resource Analysts, using an electromagnetic technology, we routinely do both conductivity and magnetic susceptibility surveys. The results tend to be non-correlated, expressing different aspects of soil conditions. Putting the two results together increases the interpretive power of the whole exercise.
Geophysical Survey at the Phase II-III Level
While my comments on Phase I-II surveys also apply to Phase II-III, as an archaeological site becomes better understood the nature of the questions asked of it changes. With a more precise idea of the location of a site, and its nature, more cost intensive data collecting methods may be used to better advantage, simply because they are required to cover a more limited area. For example, while the speed of a resistivity survey may make it less attractive when a hectare must be covered, if only 20 square meters are to be covered the cost may be acceptable. Similarly, ground penetrating radar may be less cost effective over a large area (given the cost of the equipment coupled with the speed and nature of GPR data collecting) while in a smaller area the cost becomes more acceptable.
Geophysical Survey And Other Archaeological Field Techniques
Despite its ability to rapidly provide data on the landscape context of an archaeological site, geophysical survey is generally unable to provide high quality data on the nature of a site. That is, geophysical techniques may be able to locate a site, they are less able to describe it to the satisfaction of the archaeologist or the preservation agency. Much as we would wish it to do so, geophysical survey does not record archaeological sites with snapshot clarity. On the surface this tends to frustrate archaeologists who want more from the expensive technology and the popular press that expects more. There are classic exceptions to this--the buried earthwork which is revealed in its full extent, or the buried stone foundation--but the geophysical anomaly is only an analogue of the excavated feature with all that implies in interpretation.
This being the case, it goes without saying that geophysical survey is not a stand alone method for either locating or evaluating archaeological sites, it must be used its interaction with other traditional field methods. Realistically, geophysical survey of any sort should be one element in a multistage research design projected to accomplish certain things. In the CRM context this can be the location of archaeological sites, the determination of National Register eligibility, and--considerably more complex--fine detail recovery of site structure in the context of mitigation of adverse impact or more traditional, problem oriented research.
At its most informative, the results of a geophysical survey may be used to make decisions about how to best deploy more cost intensive field operations such as shovel testing, sample block excavation, strip plowing, dozer scrapes, etc. Under no circumstances, however, should the survey results be used to categorically "write off" a site or determine its eligibility. Again, in the non-CRM context, it is the unwary archaeologist who uses geophysical data alone to interpret site structure.
Some Examples of Geophysical Survey
All three examples below represent products of recent rapid, high density geophysical surveys conducted by the author.
1. Locust Grove Stock Barn Conductivity Survey
In this example, readings were taken with a Geonics EM38 every 50 cm along transects one meter apart. At the time of survey the barn had been removed for over ten years and the site was a grassed over pasture. The barn footprint is the darker rectangle in the upper half of the map. On the two long sides are the faint indications of the stone foundation walls running diagonally across the figure. Around the barn, but generally outside the floor area, are numerous metal signatures, probably scattered barn hardware. Although some purists would not use color as I have done, it is possible to bring out details that might otherwise be hidden. The barn was built in the early 19th century and included a combination of log and frame construction. The Locust Grove barn survey took about an hour to complete.
The purpose of this survey was straightforward, to locate the foundations of the vanished barn prior to its reconstruction. In this case the geophysics provide a solution to the question, to be tested by excavation.
2. Dillow Ridge Mississippian Site Conductivity Survey (data used with the permission of the Center for Archaeological Investigations, Southern Illinois University)
Again, conductivity readings were taken every 50 cm along transects one meter apart. Dillow Ridge is covered by open forest but it was possible to thread the measuring ropes between tree trunks. The high conductivity area for the most part seem to have identified house floors, in some cases structures which had been burned. They tended to coincide with circular depressions that marked the actual house locations. Generally, each structure locus was stratigraphically complex representing multiple rebuilding. Here, contour lines have been added to the color image, effectively I think, to emphasize the variation in conductivity. This survey took about a day to complete.
The survey was performed after a several years of limited fieldwork at the site by the Center for Archaeological Investigations. If additional fieldwork was planned, it might involve a sampling strategy to test what the "highs" as opposed to the "lows" represent.
3. Millstone Bluff Mississippian Site Magnetic Susceptibility (data used with the permission of the Center for Archaeological Investigations, Southern Illinois University)
A survey using the Inphase of the EM38. Readings were taken every meter along transects one meter apart. There was a fair amount of fallen timber on this site. The downed trees slowed walking speed, and about a day and a half were required for this coverage. Magnetic susceptibility is probably recording burning for the most part in this survey. The open plaza in this isolated Mississippian village is clearly indicated by the blue "sink" in the center. The diagonal alignment of somewhat higher susceptibility across the top of the figure (adjacent to the plaza) is the current Forest Service trail leading up to the site. In this case, the Inphase is recording something beside the effects of burning. The complicated patterning of readings around the plaza reflects a complex situation caused by multiple rebuilding of typical Mississippian structures. In the lower left-hand corner of the figure, a rectangular house floor is clearly outlined by a lower susceptibility bank. The house floor, which is probably burned, registered somewhat higher. The surrounding bank may represent earth which was literally "banked" against the wall of the structure for insulation. Further work is planned at this extraordinary site including conductivity and magnetometry surveys. Southern Illinois University has continuing excavations there in cooperation with the United States Forest Service.
Conventional fieldwork indicates that the core building area of the site is complex, with many structure rebuildings (also abundantly indicated by the geophysics). The clarity of the structure in the lower left hand corner suggests that further survey might profitably be used to identify less complex structures on the periphery of the site, in other words (and illustrating a nice point from my comments) place the site in the larger landscape context of the total bluff top.