The Chacalapan Geophysical Survey, Veracruz, México

Methodology

Our geophysical work in and around Chacalapan was conducted over two periods. The first period consisted of approximately three and one half weeks during the month of March 2000. During this period we conducted extensive magnetic survey. This work concentrated on covering a maximum amount of area in an effort to locate as many features as possible. Using maps generated in the field from survey results, some suspected features were excavated, but we also decided to wait for more intensive data processing before testing all interpreted anomalies. During the second period, two weeks during May 2000, we used the processed data to generate detailed maps to locate suspected features that were not immediately visible in the raw data. Certain possible archaeological deposits were then tested with electrical resistivity in an effort to further characterize them. In addition, resistivity testing was performed on the walls of an open excavation unit in order to better refine the understanding of the area’s stratigraphic record. This section details the reasons for our choosing particular methods and equipment. A discussion of the techniques of magnetic gradiometry and electrical resistivity is presented, followed by a discussion of our field methods and computer processing methods.

Magnetic Gradiometry

After a preliminary study of the region’s archaeological remains and underlying geology during the Proyecto Arqueológico Hueyapan in 1998, we selected magnetometry and electrical resistivity as the two methods best suited to meet our goals. Magnetometry surveys are used to measure minute variations, measured in nanoTeslas (nT), in the Earth’s magnetic properties across an area. Archaeological features are detected through the contrast between their magnetic properties and the magnetic properties of the surrounding soil. Ferrous metals, dense deposits of fired ceramics, and burned areas (kilns, hearths) are readily detectable through magnetics (Breiner, 1973; Weymouth, 1986). Furthermore, areas of topsoil disturbance or organic input (whether ancient or modern) are also detectable. Depending on the strength of the source of magnetic interference the features as deep as 2.5-3 meters may be detected. Very slight changes in ground chemistry may also influence magnetic results. Areas of a site may have a higher concentration of ferrous minerals in the soil that could affect readings over a large area and need to be identified and dealt with during data processing and interpretation.

A major problem with magnetometry is that the Earth’s magnetic field is not constant throughout the day. A magnetometer records what is commonly known as the total field. The total field is the magnetic susceptibility of the soil, buried features, normal Earth magnetism, and influences from the Sun. Magnetic storms and sunspot activity can cause short term changes in the strength of the magnetic field that are orders of magnitude greater than changes caused by archaeological features. The total field can be in the range of 30,000 - 50,000 nT, while archaeological features may only show anomalies on the order of several nT or fractions of nT.  The timing and intensity of these storms are wholly unpredictable. Furthermore, the Earth’s normal magnetic field changes in intensity throughout the day. This is called the diurnal effect, and can severely hamper data quality. In order to remove these influences, archaeological geophysicists employ the technique of magnetic gradiometry. Gradiometry employs a pair of linked magnetometers set to record readings at exactly the same time. One magnetometer is positioned above the other so as to read only the Earth’s field and any stellar interference. The lower magnetometer reads as the upper one but also reads the influences from the ground. One dataset may then be subtracted from the other, thus filtering out most of the background noise and leaving only the influence of the ground.

A Geometrics G-858 Cesium Vapor Magnetometer configured in gradient mode was rented for a total field time of 30 days. The Geometrics G-858 represents the current state of the art in high-resolution magnetics and is capable of extremely fast, accurate data collection and has been employed successfully in archaeological applications (Watters, personal communication; Hervanger, 1996). Gradiometry was our primary technique and was employed as a prospecting tool for large area coverage. When possible, areas were divided into 40 x 20-m grids. It was preferable to survey as large, regularly shaped area as possible so that newly surveyed grids could be easily fitted to previous grid. However, with real world field conditions in mind we selected an instrument which has the capability to survey irregularly sized and shaped areas if we were constricted as a result of topography or ground cover (Geometrics, 1995). The ground cover over the bulk of the surveyed area was low pasture grass and thus was no impediment. Wherever bushes or trees were present, their location was marked on maps so as to identify them in the data. The majority of the terrain was relatively flat. In instances where a slope was surveyed, transects were run in one direction only. This was to keep the sensors at a constant distance from the ground surface to eliminate the substantial error that occurs when sensor distance to ground varies. Flat areas were surveyed in a zigzag pattern; with the end of one transect being the beginning of the adjacent one. The G-858 internal software is specifically designed for, and anticipates this pattern and surveying in this manner is extremely rapid. The survey time for each grid was approximately 1 hour. Within each grid the magnetic field was sampled every 10-cm along traverses spaced 50 cm apart. These survey parameters represent the very high level of resolution needed to locate prehistoric features. Data was downloaded into Geometrics’ MagMap96 software for preliminary processing and then into Golden Software’s Surfer Version 6.0 GIS for intensive processing, analysis, and display of results.

Data processing was necessary to get the magnetic information into an interpretable format. The G-858 has an internal memory system that records readings for both sensors (the total fields) as well as the physical position of the reading in the survey area. The information was downloaded into laptop computers using Geometrics MagMap96 software. MagMap is instrument-specific software for downloading and analyzing magnetic data from Geometrics instruments. It provides a schematic representation of the survey area so that the operator can make sure all readings are in their correct spatial orientation, but the software has limited processing and interpretation capabilities. Once corrections to the spatial orientation of data points were made, the X and Y position of the survey points and the difference between the total fields as measured by the two sensors (the gradient) were exported to Surfer. In addition to performing the statistical computations necessary to process information, Surfer has the capability to deal with truly huge datasets. This was significant to our work because we wanted to concatenate (or fit together) all survey grids in a given area prior to processing. This would insure that mathematical functions would operate equally on the whole dataset. To illustrate the type of processing capability we needed, the larger of the two areas from the site of El Tecolote was made up of nearly one half million individual data points.

Once in Surfer, descriptive statistics were computed for each dataset and data were clipped outside of two standard deviations from the mean reading. This eliminated the influence of data "spikes." Spikes are readings that are found at the extremes of the dataset and are likely caused by instrument error or metallic surface debris such as bottle caps or cans. Their presence can obscure the more ephemeral signals that may be caused by archaeological features. It is still useful to examine the data carefully with the spikes present to discern if they may be archaeologically significant. Data were then reexamined with descriptive statistics to check if any drastic alteration of the mean occurred which could signal error. The clipped dataset was then gridded by Surfer with no interpolation. From this, contour and color image maps were produced. The displayed data range was manipulated to try to zero in on features. We produced final maps in Surfer for use in planning excavations. These maps were grayscale depictions and because magnetic readings have a positive and negative numerical component, extremes were identified by either red or blue. This allowed the computer to "stretch" its 256 shades of gray across a finer range and bring out more subtle detail. Anomalies were identified based on their shape. Regular shapes, or patterned geometries, rarely occur in the natural world and usually signal manmade features. Furthermore, discrete, highly magnetically susceptible features tend to cause a paired high/low reading called a dipole. Dipoles were also noted as possible archaeological features.

Electrical Resistivity

Resistance surveys measure variations, expressed in ohms (Ω), in the electrical resistance of soil across a site or down through a column of earth. An electric current is introduced into the ground by a series of probes. The current travels through the ground to another series of probes that measure the potential difference. Electricity always seeks the path of least resistance so that if it encounters a highly resistant feature such as a stone wall, it will travel around it. A longer travel path will result in greater resistance. Electrical properties of a soil are primarily dependent on matrix type, compaction, moisture content, and objects buried within (Clark, 1996; DeVore and Heimmer, 1995). The ohm measurement is later calculated with respect to depth to normalize it so that readings taken as different depths can be compared, this is expressed as resistivity (ρ). This technique was used at Chacalapan primarily to test areas of magnetic interest, to take resistance "soundings" at different depths over a single location, and to test the walls of open excavations. In general, humic rich midden areas will tend to produce lower resistance values due to a higher concentration of organic ions and the presence of larger, moisture filled voids in coarse midden fill. Resistivity measurements vary tremendously with differing soil moisture, so there is much more interpretive guesswork involved with this technique. Since the same types of features produce anomalies in both techniques, resistivity correlates well with magnetics for certain types of features. Resistance soundings also enabled us to get an idea of the degree and type (sand, silt, clay) of the fill at varying depths. Resistivity was not used for large area survey, but only to "spot check" certain areas. By altering spacing between the soil probes, the instrument can be configured to read at differing depths. We used a Gossen Geohm 3 Soil Resistance Meter lent by the Boston University Department of Archaeology at no cost. The Geohm 3, while an older instrument, was well suited to our needs. It is small, has a compact power source, and features a manual probe configuration that allows easy surveying at multiple depths.

Over an area of magnetic interest, readings were taken at a consistent interval at several depths. For the majority of features, readings were spaced every 50 cm along a transect located to straddle the magnetic anomaly. Readings were then over a point taken at depths of .5, 1, and 1.5 meters. Readings were then converted into resistivity (ρ) measurements so as to be able to compare the composition of the soil through depth. No additional processing was needed.

Previous Page  |  Table of Contents  |  Next Page

Return to top of page