Hiwassee Island: Partnering with Tribes to Ground Truth Geophysical Studies

30 Days of Tennessee Archaeology, Day 28

Erin Pritchard
Tennessee Valley Authority

Hiwassee Island, located at the confluence of the Tennessee and Hiwassee Rivers in Meigs County, Tennessee contains a rich history dating back many thousands of years.   The island is now owned in fee by the U.S. Government under the stewardship of the Tennessee Valley Authority (TVA) and is under easement with the Tennessee Wildlife Resource Agency (TWRA) for the management of wildlife (Figure 1).


Figure 1. Hiwassee Island

Prior to the inundation of Chickamauga Reservoir, numerous archaeological excavations occurred on Hiwassee Island and the true extent of intact deposits remaining was unknown. In 2016 (Previous Blog ) we reported on TVA’s efforts to document these remaining deposits through geophysical survey (Figure 2). This survey produced outstanding results indicating that the Mississippian village located on the island still retained significant integrity with as many as seven palisade features. Only one palisade feature had initially been identified on the island from previous excavations. The island was recently determined eligible for listing in the National Register of Historic Places using data collected from this as well as the many other studies conducted on the island.


Figure 2. Shawn Patch (New South Associates) working with Gano Perez (Muscogee Creek Nation of Oklahoma) during the geophysical field school on Hiwassee Island.

In 2017 TVA continued its research efforts on the island by conducting limited test excavations to ground truth the results of this previous study and to collect radiocarbon dates from the newly identified palisade features. The following results highlight some of the more interesting findings for this project.

The scope of work for the test excavation was limited as the agency did not wish to damage more features than was necessary to collect sufficient carbon samples. Fieldwork was limited to one week and eight test units (measuring 1m by 50cm, 1m by 1m, or 1m by 2m depending on the anomaly being tested). The research design for the project sought to identify a sequence for the palisade construction through the radiocarbon dates as well to examine differences in construction methodology for each of the five palisades that were investigated.

TVA partnered with federally recognized tribes to provide an archaeological field school opportunity for non-archaeological staff and to provide training for tribal monitors. Eight tribal participants from five tribes (Chickasaw Nation, Eastern Band of Cherokee Indians, Muscogee (Creek) Nation of Oklahoma, Sac and Fox Nation in Oklahoma, and the United Keetoowah Band) participated in the excavations along with TVA staff and managers, Tennessee Wildlife Resources Agency, and volunteers from TVA’s Thousand Eyes Archaeological Site Stewardship Program (Figures 3a-c).


Figure 3a-c: a: (upper left) Shawn Patch excavates a unit while Corain Lowe-Zepeda (Muscogee Creek Nation of Oklahoma) and Jason Jackson (Tennessee Wildlife Resource Agency) screen for artifacts; b (bottom left): Matt Girty (United Keetoowah Band) excavates as Karen Loven (Thousand Eyes Volunteer), Jamie McCabe (New South Associates, Inc.) and Michaelyn Harle (TVA) screen for artifacts; c (right)  Corain Lowe-Zepeda excavates as Benny Wallace and Catie Hamilton (both from the Chickasaw Nation) observe her findings.

Results of the excavations (Table 1) confirmed our initial interpretation of feature type in all but one of the units. Radiocarbon dates collected suggest that the village expanded over time reaching its peak during the later Mississippian Dallas Phase. While additional radiocarbon dates would be needed to fully support this hypothesis, these initial results confirm that the information potential of this island is still extensive further supporting its eligibility for inclusion in the National Register of Historic Places.


Table 1.  Radiocarbon results from each of the excavated units.

The one surprising result was found in Test Unit 1 where the thick linear anomaly was initially hypnotized to be a late Mississippian wall surrounding what was once a large platform mound excavated during the 1930s Works Progress Administration work which is believed to have been the center of the village.   Excavation of the feature now suggests that this feature is actually a ditch of unknown origin. Based on the content of the feature, we believe it was filled during the later portion of the Hiwassee Island Phase (Figure 4).


Figure 4.  Test Unit 1 profile showing fill placed in the feature.

While not all features ended up being what we initially hypothesized, it was clear the technology is very effective for identifying archaeological features and TVA was able to obtain radiocarbon dates to further hypothesize on the expansion and/or contraction of the village. Results of these limited excavations confirmed that this non-invasive technology can be an effective approach to collecting data from known archaeological sites in order to evaluate potential significance and provide the agency with sufficient data to properly manage archaeological sites on its federal lands.

More importantly, the experience TVA had with inclusion of Federally recognized tribes in the field work was extremely rewarding. What started out as a field school for the tribal reps ended up being a learning opportunity for all of the participants in the project (Figure 5). Similar projects are already planned for the future and staff is very excited for the opportunity to work with tribes to learn more about their rich history in the Tennessee Valley.


Figure 5. Participants in the Hiwassee Island field school.


Archaeological Geophysics at Runion, a Protohistoric Site on the Middle Nolichucky River

30 Days of Tennessee Archaeology, Day 12

Eileen G. Ernenwein and Jay D. Franklin
East Tennessee State University

In last year’s blog post, we reported on the results of geophysical survey and excavation at the Cane Notch site, a mid to late sixteenth century Cherokee Town on the Nolichucky River in Upper East Tennessee. This work is also featured in the documentary film, Secrets of the Nolichucky River. A host of other pre-contact and contact-era sites are known along this stretch of the Middle Nolichucky, a region we believe represents an early Cherokee culture area where interactions with Spanish Explorers almost certainly took place. We continue to explore this idea with a focus on the Runion site (40WG20), some 10 miles downriver from Cane Notch.

In January 2017, we surveyed six hectares of the site with magnetic gradiometry (MG) using a SENSYS MXPDA push-cart system with five fluxgate sensors. The results (Figure 1) reveal a small village consisting of square houses almost identical (in the data) to those at Cane Notch. Twenty-one probable structures are highlighted in Figure 1b, while dozens of amorphous anomalies not highlighted could also be structures or associated features. One structure in the south is likely a townhouse given that it is roughly four times the size of the others (twice the length and width). We will refer to this structure as a townhouse hereafter. In addition, there is evidence of fortifications surrounding the houses on the north, west, and south, with the river along the east forming a natural defense. It is also possible that these curvilinear anomalies are related to river channel geomorphology, however, and are entirely natural.


Figure 1a and 1b: Preliminary magnetic gradiometry (MG) results: (left) full survey data showing location of the 30 x 30 m test grid; (right) nineteen roughly square structures (red) measuring 5-7 meters on a side are identified in the central area. The larger structure in the south, likely a townhouse, measures about 13 x 13 meters

Magnetic gradiometry (MG) is a well-known method for rapidly surveying large areas as a means of reconnaissance. The houses, possible fortifications, and other features are visible because they were burned, are filled with accumulated topsoil, or they contain rich cultural deposits (often with burned material such as charcoal). As with our results at Cane Notch, however, we know that there is much to gain by applying more than one geophysical method. Ground penetrating radar (GPR) and electromagnetic induction (EMI) often reveal more about known features aid discovery other features and their function, depth, composition, and level of preservation. GPR uses the elapsed time and amplitude of radio wave reflections to detect the depth and nature of archaeological and other subsurface layers and features. EMI, a less commonly used geophysical method, transmits much lower frequency electromagnetic waves to measure electrical conductivity (EC) and magnetic susceptibility (MS) at multiple depths.

We tested the effectiveness of GPR and EMI at Runion by collecting data in a 30 x 30 m grid over the townhouse and surrounding area (highlighted in green in Figure 1). Figure 2a shows a close-up of the townhouse as previously revealed by MG, illustrating the clear outer limits of the structure, a central hearth, and evidence that deep plowing has cut into it as revealed by the light colored lines trending north to south. MS data show portions of the townhouse quite clearly (Figure 2b-c), while EC data can only be interpreted within the context of the known structure (Figure 2d-e). GPR depth slices (Figure 3) show that this structure is situated approximately 28-56 cmbs and that there is a concentration of material that produces strong reflections in the northeast corner. Most interesting is that the GPR also shows a structure situated about 6m due north of the townhouse (circled in the fourth slice), a feature not clearly shown in any other dataset.


Figure 2. Magnetometry and EMI results for the 30 x 30 m test area. (a) MG, (b) shallow MS (~0-25 cmbs), (c) deeper MS (~0-50 cmbs), (d) shallow EC (~0-75 cmbs), (e) deeper EC (~0-150 cmbs).


Figure 3. GPR slices representing the following depth intervals from left to right, top to bottom in cm: 28-32, 32-35, 35-39, 39-42, 42-46, 46-49, 49-53, 53-56. The smaller structure located north of the townhouse is circled in the fourth slice.

All data collected in the 30 x 30 m area were combined in GIS to see how they complement one another. Figure 4 maps anomalous areas of positive MG, the inverse of the average EC (labeled ER to indicate these are high electrical resistivity areas), average MS, and the mean of all strong GPR reflections. This representation makes it clear that the combination of data from all three instruments provides a detailed view of the townhouse and a better depiction of the nearby smaller structure. The fill inside the northeast portion of this structure is more electrically resistive (and therefore causes strong GPR reflections), while it is more magnetic (likely a higher concentration of burned materials) in the southwestern portion. Given these results, it is clear that we have much to gain by expanding EMI and GPR survey over as much of the site as possible this fall and winter. Excavations in our 2018 field school will target features revealed by geophysical survey with the aim of understanding what types of fill are associated with each anomaly type.


Figure 4. Vector Fusion of all geophysical data in the 30 x 30 m test area situated over the townhouse: (blue) strong GPR reflections from multiple depths; (crosshatch) low EC values; (red) Strong positive MG values; and (red horizontal lines) strong positive MS.

Acknowledgements. We thank Wolfgang Suess, Gorden Konieczek, and Michaela Schulze at SENSYS Sensorik & Systemtechnologie GmbH (http://www.sensysmagnetometer.com) for loan of the MXPDA magnetometer system and technical support.  Thanks to the Runion family for allowing the research to take place on their property. Thanks to Nate Shreve and Claiborne Sea for help with site logistics and geophysical survey.