Archaeogeophysical Investigations at the Late Mississippian Parkin Site in Arkansas

By Jami J. Lockhart, Jeffrey M. Mitchem, and Timothy S. Mulvihill
Arkansas Archeological Survey


During four days in late May of 2001, Jami Lockhart, Jeff Mitchem, and Tim Mulvihill of the Arkansas Archeological Survey carried out preliminary archeogeophysical surveys at Parkin Archeological State Park. The goal was to test the efficacy of five different geophysical techniques of near-surface prospection for archaeological features at the Parkin site (3CS29). The five technologies used were electrical resistance, magnetometry, electromagnetic conductivity, magnetic susceptibility, and ground penetrating radar. This paper summarizes some of the early findings of this brief preliminary work.

The Parkin site is a fortified 17-acre Native American village located on the St. Francis River in the Mississippi Alluvial Valley of northeastern Arkansas. It was occupied as early as A.D. 1000 until at least 1550, and is one of the best-preserved village sites of this time period in the region. The Parkin site is also important because many scholars believe it was the Native American town of Casqui visited by Hernando de Soto's expedition in the summer of 1541 and described in the written accounts of the surviving Spaniards.

Parkin, like all archaeological sites, is the product of formation processes occurring over a period of time. These processes are both cultural and non-cultural. Sites may be transformed by human action during and long after their original use. Non-cultural formation processes determine what decays and what is preserved. Soils are physically and chemically altered over time, and the archaeological record is transformed — spatially, quantitatively, and in content — from its original state. These processes result in combinations of physical properties defined by soil texture, stratigraphy, artifact type and distribution, burning, biogenic and biochemical components, compaction, and differential moisture retention. Geophysical technologies can measure the variable strength and location of these properties, giving clues to what lies beneath the surface.

But archeogeophysics is not just about finding things. Through an iterative process — by matching the geophysics to the archaeology and vice versa — we hope to build a "library" of geophysical signatures that will allow us to better interpret the patterns of cultural organization, lifeways, and even world view that are left to us in the ground.

Figure 1. Images of the Parkin site: a 1939 map and a modern-day digital photograph from the air. CLICK ON IMAGE TO ENLARGE
Figure 1. Images of the Parkin site: a 1939 map and a modern-day digital photograph from the air

Figure 1 displays a 1939 map of Parkin, made by archeologist Philip Phillips, alongside a modern-day digital photograph of Parkin taken from the air. The St. Francis River can be seen on the left in both images, and formed the western boundary of the site. The large extant platform mound is located close to the river. The rectangular defensive moat is clearly seen in the modern image, and is also still visible on the ground.

Figures 2a and 2b. Views of the large flat-topped mound at Parkin: left, as it appears today; right, artist Dan Kerlin's conceptualization of the mound as it would have appeared in 1541. CLICK ON IMAGES TO ENLARGE
Figure 2a. The large flat-topped mound at Parkin as it appears today Figure 2b. Dan Kerlin's conceptualization of the mound as it would have appeared in 1541

The appearance of the large mound today is illustrated in the photograph in Figure 2a. Dan Kerlin's artistic representation (Figure 2b), shows the mound from approximately the same vantage point, as it may have appeared in 1541. The two platforms of the mound have become less well defined over time as a result of erosion, but are still quite evident.

Figure 3. May 2001 survey grids. Click image to enlarge
Figure 3. Topographic map of the Parkin site, showing the May 2001 geophysical survey grids in solid black

We used a Total Station transit to shoot a large number of discrete x,y,z coordinates across the site. These three-dimensional measurements can be interpolated to produce a topographic map, as shown in Figure 3. This figure also shows the grids used for geophysical survey (in solid black). An L-shaped block of grids in the southern portion of the site was laid out over an area where we expected to find evidence of palisade fortification near the moat. Two grid blocks in the Locus 4 area in the northern portion of the site were placed close to previous excavations. In the central area, a rectangular block of three contiguous 20 x 20 m grids was placed in front of the large mound. Smaller grid units were located on the high and intermediate platforms of the mound. We will discuss the results for these survey grids, moving roughly from north to south.

Figure 4. Linear anomaly in Locus 4. Click image to enlarge
Figure 4. This linear anomaly in Locus 4 proved, upon excavation, to be a filled ditch of recent origin

A very striking linear anomaly shown in the electrical resistance imagery from Locus 4 was the first thing that got our attention (Figure 4). There was no accompanying surface expression whatsoever. Parkin site archeologists Mitchem and Mulvihill searched old records and spoke with long-time residents of the area, but could find no evidence of a buried water line or other utility. Subsequent excavation of a 2 x 2 m test unit over the anomaly in October, 2001, turned up substantial amounts of aboriginal material, along with numerous historic period artifacts. The profile of this feature revealed it to be a filled historic ditch — probably used for drainage — that had cut through one or more buried prehistoric house floors. This portion of the project highlights the value of geophysics in locating underground features for which there are no surface indications. It also underscores the efficiency of this technology, when accompanied by pinpointed excavation units for "ground-truthing," to maximize the amount of information that can be recovered with only limited disturbance to the site.

Figure 5. House floors revealed during excavation. Click image to enlarge
Figure 5. House floors revealed during excavation

Much more extensive excavations had been conducted by Mitchem and others between 1991 and 1999 in the same part of the site (Locus 4) as the linear anomaly. These excavations uncovered a number of prehistoric features, including house remains, post molds, trash pits, and burial pits, as well as modern looters' holes. All of the house floors had been disturbed, either in prehistoric times by continued use of the area for new house construction, refuse disposal, or burial, or more recently by pothunting and farming. All the houses were about 4 m square with prepared floors consisting of packed-down yellow or olive-colored clay. Figure 5 shows one corner of Structure 13 (in the right hand portion of the image) during excavation.

Figure 6. Locus 4. Click to enlarge
Figure 6. Electrical resistance imagery in Locus 4, with anomalies that may prove to be prehistoric house floors, burials, pits, and other features

These 1990s excavation units were located in the 10 x 20 m blue area of Figure 6. (Since this blue area had been excavated prior to the geophysical work, there is no remote sensing data for this portion of the grid.) The geophysical anomalies, outlined in red on the same figure, are in areas that have not been excavated. We think that some of the larger discrete anomalies may be additional prehistoric houses, and that some of the smaller ones are burials or other pits — even looter holes. Ground-truthing these hypotheses will be important in building up a library of geophysical signatures that will help us to more accurately interpret the site.


Figure 8 (left) - Enlarged portion of the images shown in Figure 7. Note the very faint linear anomaly cutting across the upper left-hand corner. Figure 9 (right) - the same linear anomaly is much more strongly expressed in the magnetometry image, illustrating the value of using multiple technologies. CLICK ON IMAGES TO ENLARGE
Figure 8. Enlarged portion of the images shown in Figure 7. Note the very faint linear anomaly cutting across the upper left-hand corner. Figure 9. The same linear anomaly more strongly expressed in the magnetometry image

Figure 7. Electrical resistance. Click image to enlarge
Figure 7. Electrical resistance results from the smaller survey grid in Locus 4

Electrical resistance results for the smaller survey block in Locus 4 are shown at the bottom of Figure 7. Again, there are roughly rectangular anomalies of about 4 m square — the same size as the previously excavated houses. This same grid unit contained another interesting anomaly, shown enlarged in Figure 8, consisting of a very faint line cutting across the northwest (upper left) quadrant of the image. Figure 9, the magnetometry image, gives a much stronger result for the same anomaly. This feature presents a different signature in both technologies than the linear anomaly discussed above that proved to be a historic ditch. There is thus the possibility that it is related to the Native American occupation. The other discrete magnetic anomalies in Figure 9 are located within those 4 x 4 m rectangular anomalies shown in the electrical resistance imagery. We suspect that some of them represent prehistoric hearths associated with the hypothesized houses. The hearths, if excavated, might look like the example in Figure 10, which shows the profile of a previously excavated hearth not 20 m away. Other anomalies might be concentrations of burned daub, such as the excavated feature seen in Figure 11, also located in the near vicinity.

Figure 10 (left) - Smaller magnetic anomalies may prove to be hearths, such as this one, excavated in Locus 4 in 1996. Figure 11 (right) - A concentration of burned daub, also in Locus 4, excavated in 1997. Some of the smaller geophysical anomalies may represent similar features. CLICK ON IMAGES TO ENLARGE
Figure 10. Smaller magnetic anomalies may prove to be hearths, such as this one, excavated in Locus 4 in 1996 Figure 11. A concentration of burned daub, also in Locus 4, excavated in 1997

Figure 12. Electrical resistance images from in front of the mound. Click image to enlarge
Figure 12. Electrical resistance images from the area in front of the large mound, with anomalies possibly representing structures

We move now to the survey grids located on and alongside the large mound at Parkin. Time constraints allowed only electrical resistance survey in the larger block. The image shows several rectangular anomalies, perhaps representing houses or other structures and related features (Figure 12). This result is very interesting, as it is tempting to interpret this flat area in front of the mound as an open plaza, which is often seen at Mississippian sites. Contiguous coverage using multiple technologies should provide us with a basis to interpret the function of this area in the heart of the site.

Figure 13. Ground-penetrating radar on the mound. Click image to enlarge
Figure 13. Tim Mulvihill and Jeff Mitchem operate the ground-penetrating radar equipment in a depression on the high platform of the mound.

We were able to obtain only preliminary results for the smaller survey grids on top of the mound. Still, the work in this part of the site was very useful. We used ground penetrating radar (GPR) — with which we were less familiar at the time — to get some comparative results among the various technologies (Figure 13). The GPR time slice was collected using a 400 MHz antenna capable of 3 meter penetration. The images in Figure 14 are all from the lower platform of the mound. These experiments produced good correspondence between the GPR data, magnetometry, and electrical resistance — all showing similar anomalies in the same areas.

Figure 14. GPR contour map. Click to enlarge
Figure 14. GPR contour map

All of the work we have done at Parkin has been carefully georeferenced using a Total Station. This allows us to compare results from archaeological excavation with those from remote sensing with great precision. For example, Figure 2b showed an artist's conception of a historical event at Parkin/Casqui — the raising of a large wooden cross on top of the platform mound by de Soto's men (an act performed in order to impress the Indians). Excavations on top of the mound in the 1960s recovered pieces of an unusually large burned post. Could it be remnants of the cross? We do not know. But the coordinates of the archaeological feature, when georeferenced to our grid, plot on top of one of the GPR anomalies seen in the southwest quadrant of the images in Figure 15. We hope that more extensive surveys later on will help us interpret this intriguing correspondence, as well as other features of the large mound.

Continuing use of these techniques at Parkin will help us determine whether the area in front of the platform mound was indeed an unoccupied plaza, or just part of the crowded settlement. Comprehensive data will also help us locate houses and other features upon which to focus future excavation and research effort. By surveying large portions of the site and carrying out limited excavation on specific anomalies, we may eventually be able to recognize specific signatures of various types of archaeological features. Such an archeogeophysical image index would help guide future research not only at Parkin, but at other Mississippian sites in northeast Arkansas as well, for decades to come.

Aside from pure research, there are practical applications for the archeogeophysical data. For instance, if we can define a specific signature for human burials, we will be able to plan excavations to avoid disturbing them. Since burials are considered sacred sites by Native Americans, and are also protected by federal and state laws, it is proper to respect such remains by leaving them undisturbed. Most archaeological work is carried out under tight constraints of budget and labor. The potential savings in time, money, and energy would be a considerable benefit whenever excavation strategies can be informed by geophysical results.

Figure 15. GPR slices from the high mound platform. Four levels show anomalies at varying depths. Remnants of a large burnt post excavated in the 1960s plot on one of the anomalies in the lower left quadrant. CLICK ON IMAGE TO ENLARGE
Figure 15. GPR slices from the high mound platform

The 2001 archeogeophysical work at Parkin demonstrates that these techniques are of use in locating subsurface features at the site. This study serves as a pilot project that highlights the need for additional geophysical survey to guide future research at Parkin. The well-preserved nature of the Parkin site makes it a prime testing ground for this technology and a useful "laboratory" for identifying geophysical signatures of specific archaeological features that can be applied as well to other Mississippian sites in the region.

Back Return to the News and Current Research page.

Home | News | About | State Archeologist | Education | Publications | SRP | Related

Petroglyphs

Copyright 2002, Arkansas Archeological Survey, Revised - December 13, 2002
http://www.uark.edu/campus-resources/archinfo/parkin_rs.html
We welcome your comments.
Mail