Interpretation of Geophysical Survey Results

The results of geophysical surveys of archaeological sites are generally presented graphically. This is done because anomalies of cultural origin are generally recognized by their pattern, rather than by their numeric values alone. When rendered graphically, we can better recognize cultural and natural patterns and visualize the physical phenomena causing the detected anomalies.

Interpretation of survey data must be a cooperative process involving both archaeological geophysicists and archaeologists that are familiar with the specific cultural context of the site being studied. An understanding of the geological context of the survey area is also very important, and consultation with a geomorphologist can be very important in understanding survey results.

In areas that have been surveyed with more than one type of instrument, the results of the different surveys should be carefully compared. Correlations between data sets (or lack of correlation) can be as important as either data set by itself to our interpretation of the site.

Initial interpretations of the geophysical surveyor should be reviewed by archaeologists familiar with the cultural context of the site. Comparison of survey results with the range of expected feature types and intra-site patterning may result in different or elaborated interpretation. Initial interpretations should be considered hyptheses to be tested. Testing (or "ground truthing") will greatly inform further interpretation of survey data. Verification (or refutation) of preliminary interpretations and insights into feature composition and geology can allow us to revise or elaborate our interpretations, and to do so with greater confidence.

The results of geophysical surveys and testing should be used in conjunction with other available sources of information to understand the general site context, to locate features for excavation, and to understand the results of excavation within the greater site context.

Testing interpretations

In order to better understand the results of a geophysical investigation, ground truthing (subsurface testing) should be performed on selected geophysical anomalies. Limited invasive exploration should systematically address the following questions:

• Have preliminary interpretations correctly identified archaeological features?
• Can more ambiguous anomalies be identified (or dismissed) as archaeological features?
• What is the specific physical composition of the features?
• Can the cultural context of the features be better defined?
• What is the state of preservation or integrity of the features?

Ground truthing will allow more definitive interpretation of the geophysical data and provide data on a range of previously unexplored areas within the site. The results of ground truthing may (with caution) be interpolated and extrapolated to untested areas of the site.

Careful attention to spatial control during ground truthing will minmize search time and increase the chances of success. Failure to locate anomaly sources is very often due to imprecise placement of tests (more about spatial control here).

Careful testing of a meaningful sample within the interior and exterior of geophysical anomalies should identify the source of the anomalous signal. As ground truthing progresses, more information concerning the signal response is gained and interpretation by archaeologists familiar with the regional archaeology becomes increasingly detailed.

Resistance Survey

A hydraulic coring rig can sample more deeply, and penetrate very hard soils. Features are more easily identified in larger diameter cores.

For resistance data, testing should extend beyond the apparent location of the anomaly of interest in both the east-west and north-south directions. As a general guideline, testing should extend on either side of the anomaly for a distance of approximately two or three times the electrode probe separation used in the survey.

When the anomaly of interest is linear, testing should be oriented perpendicular to the edge of the anomaly, and should extend two or three times the probe separation on either side. Excavation to a depth of approximately three times the electrode probe separation distance or to the sterile soil horizon will generally reveal the subsurface components contributing to the measured resistance values. On occasion, resistance features may not be visually observed in the test trench. For example, a high salinity moist soil may be visually identical to a low salinity moist soil but the measured resistance could easily differ by a factor of 1,000. On these occasions careful attention must be paid to local variations in soluble ion concentration, physical soil particle size, and moisture variations.

Magnetic Survey

For magnetic field gradient data, testing should be centered over the anomaly of interest, and should extend approximately three times the diameter of the magnetic anomaly on either side in both the east-west and north-south directions. When the magnetic anomaly of interest is linear, testing should take place perpendicular to the anomaly and should extend approximately three times the diameter of the magnetic anomaly on either side. For magnetic data, maximum excavation depth is more difficult to specify. It is a complex function of the magnetic moment, angular orientation, soil susceptibility contrast, and physical size of the buried feature.

Ground Penetrating Radar Survey
The source of anomalous reflections in GPR data will generally be located near its apparent horizontal position in the data plot. Exceptions, however, are not uncommon. Reflector geometry and other factors may cause the apparent location of the anomaly source to be significantly displaced. Under typical survey conditions, the error in horizontal position of a feature of interest is not likely to be greater than its depth.

The estimated depth to an anomaly source is based on an estimate of the signal velocity through the soil (or other matrix). The velocity of the signal is not uniform through different materials, but in most instances the vertical error is not likely to be greater than 50% (up or down). Initial ground truthing results will enable the investigator to evaluate estimates of depth. It is possible to perform tests in the field to more accurately estimate the velocity of the GPR signal through the local soils, and therefore more accurately estimate the depth to anomaly sources without invasive testing.

Although they can usually be distinguished by their appearance, reflected air waves, "ringing," and other phenomena can create reflections that appear wildly displaced (vertically and horizontally) from their actual source, which may even be above the surface.

See also: Case Study - Odessa Yates Site, 1998 Ground Truthing