Investigating Construction History, Labour Investment and Social Change at Ocmulgee National Monument's Mound A, Georgia, USA, Using Ground-penetrating Radar

12 pages
of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Investigating Construction History, Labour Investment and Social Change at Ocmulgee National Monument's Mound A, Georgia, USA, Using Ground-penetrating Radar
  Investigating Construction History, LabourInvestment and Social Change at OcmulgeeNational Monument ’ s Mound A, Georgia, USA,Using Ground-penetrating Radar DANIEL P. BIGMAN 1 * AND  PETER M. LANZARONE 2 1 Department of Anthropology, Bigman - Georgia State University, 33 Gilmer St, Atlanta, GA, 30303, USA 2 Lanzarone - BP America, Inc., Houston, TX, USA  ABSTRACT Prehistoric societies from around the world constructed monumental mounded architecture (earthen pyramids) for avariety of functions, including the foundation for temples or a leader ’ s residence, community stages and cemeteries.Flat-topped earthen mounds often have complicated histories where the function, size, orientation and summit archi-tecture varied throughout time. This paper presents the results from a ground-penetrating radar (GPR) surveyconducted on the summit of Mound A, the largest prehistoric Native American mound at Ocmulgee NationalMonument located in central Georgia, USA. Our study indicates that performing depth-slice analyses of  󿬂 at-toppedmounds can effectively map successive construction stages over distinct periods of archaeological prehistory. TheGPR data show distinctive low-amplitude, discontinuous stratigraphic variations that we interpret to be related tomound  󿬁 ll, which are interrupted by high-amplitude coherent summit re 󿬂 ections. We also maintain suf 󿬁 cient verticaland horizontal resolution to identify summit architecture on earlier mound-use episodes, which is imaged by distinctre 󿬂 ection geometry as patterned linear, square and circular high-amplitude events in GPR depth-slices. The authorsrecorded four possible mound summits, the western expansion of the mound, and the shifting location and shapeof summit architecture; in addition to resolving a discrepancy regarding the location of early excavation units fromthe 1930s. Shallow geophysical data indicate a decline in the volume of material used over the course of moundconstruction, and by inference, a decline in the size of the labour force used to construct each stage. We concludethat the power and in 󿬂 uence of Ocmulgee ’ s early leadership subsided over the course of Mound A  ’ s use, and mayhave been contested by an emerging faction. Copyright © 2014 John Wiley & Sons, Ltd. Key words:  North American Mississippian Mound exploration; GPR; time-slice analysis; social capital; labourinvestment; DGPS Introduction This paper presents the results from a ground-penetratingradar(GPR)surveyconductedonthesummitofMoundA, the largest prehistoric Native American mound atOcmulgee National Monument located in centralGeorgia (Figure 1). The mound, which is situated on a bluff at the southern end of the site, is approximately15 m tall and measures roughly 91 m 2 (Figures 2 and3). Our project goals were to locate vintage excavationunits from the 1930s on Mound A ’ s summit, build anunderstanding of the mound ’ s construction historyand identify the presence or absence of summit archi-tecture during multiple use episodes. In this paper,we  󿬁 rst present some background information andthen detail previous investigations on Mound A tohighlight inconsistencies in the interpretations of dif-ferent archaeologists working at Ocmulgee NationalMonument. Next, we provide our data collection andprocessing methodology. We then present the resultsof our GPR survey and discuss how our  󿬁 ndings helpresolved discrepancies arising from previous work. Toconclude,webrie 󿬂 ydiscusssocialchangesatOcmulgee based on the results of our GPR survey. Ultimately, wehope our results expand the potentialutility for applyingthree-dimensional GPR modelling of monumentalmounded architecture; and provide additional insightinto the socio-political structure at Ocmulgee. * Correspondence to: D. P. Bigman, 340a Sparks Hall, 33 Gilmer St,Atlanta, GA, 30303, USA. E-mail: Copyright © 2014 John Wiley & Sons, Ltd.  Received 7 June 2013 Accepted 11 February 2014  Archaeological Prospection Archaeol. Prospect.  (2014)Published online in Wiley Online Library( DOI: 10.1002/arp.1483  Background Ocmulgee National Monument is an archaeologicallandscapewithvarieddegreesofpreservationandasig-ni 󿬁 cant occupational history from prehistoric to histori-caltimes.Thesiteislocatedapproximately400meastof theOcmulgeeRiver,wheretherivertransitionsfromtheGeorgia Piedmont physiographic terrain and enters theCoastal Plain (Hally and Williams, 1994, p. 86).There isa diverse soil setting encompassing the site. The parentmaterial for soils to the north of the site consists of sap-rolite of igneous and metamorphic srcin weatheredfrom biotite gneiss, granite gneiss and granites; whilethe parent material to the south of the site mainlyconsists of sedimentary rocks (Woods, 1979). Soils withinthe park boundaries consist of 1 – 2 m of red sandy soiloverlaying sandy clay material. The site ’ s topography isvaried with little to no accumulation of organic soil on bluff tops.This pedologic background facilitates GPR as anappropriate geophysical tool to explore the mostenduring features at Ocmulgee (the variety of earthenmounds constructed by Native Americans roughly1000 years ago), because construction techniques inthe area utilized mound-capping materials (includingsand, burnt charcoal, or hardened clay) with physicalproperties that contrast dramatically with those foundwithin the site ’ s pedogenic setting. Archaeologistshave often investigated  󿬂 at-topped earthen pyramidsthroughout the American Southeast, commonly calledmounds, using a variety of geophysical prospectiontechniques. Investigators have traditionally used verti-cal two-dimensional GPR pro 󿬁 les (Garrison, 1998;Welch  et al ., 2006; Pluckhahn  et al ., 2010; Thompson et al ., 2011), and more recently electrical resistivitypseudo-sections (Monaghan and Peebles, 2010;Kassabaum  et al ., 2014) and down-hole magnetic sus-ceptibility (McNutt  et al ., 2012; Rodning  et al ., 2013;Kassasbaum  et al ., 2014) to understand internal moundstructure and to interpret the sequence of mound con-struction episodes. Others have demonstrated the util-ity of magnetic gradiometry in locating subsurfacestructural remains on a mound ’ s  󿬁 nal summit (Butler et al ., 2011; King  et al ., 2011). Some investigators inother parts of the world have experimented withGPR time-slices to create horizontal maps of successivemound-use episodes, identify buried tombs and recon-struct mound manufacturing techniques, with varyingdegrees of success (Kamei  et al ., 2000; Campana  et al .,2009). Recent work in the American Southeast byMcNutt  et al . (2012), allowed the identi 󿬁 cation of pos-sible buildings on successive mound summits utilizingGPR time-slices. Flat-topped earthen mounds oftenhave complicated histories where the function, size,orientation and summit architecture varied throughtime. Shallow geophysical prospection has shown to be a viable tool used to elucidate these complexities,as demonstrated by the work of the aforementionedinvestigators.Although Native Americans reoccupied the Ocmul-gee landscape during every major period between ap-proximately 13 000 and 300 years BP, the mounds wereconstructed during the Early Mississippian period( AD  900 – 1200) (Hally and Williams, 1994). The Missis-sippian period in the American Southeast is character-ized by a variety of cultural phenomena that includes:population aggregation, the intensi 󿬁 cation of maizeagriculture, the integration of numerous settlementsinto larger decision-making bodies, the institutionali-zation of leadership, and in some areas, heavy compe-tition seen in site clustering and territorial boundaries(Smith, 1978; Hally, 1993; Milner, 2004; Anderson andSassaman, 2012). Mississippian political centres oftenadhered to an  ‘ architectural grammar ’  that consistedof mounds, plazas and palisades (Lewis and Stout,1998), but their arrangements varied, and eacharchitectural element was probably multifunctional Figure 1. Location of Ocmulgee National Monument. D. P. Bigman and P. M. Lanzarone Copyright © 2014 John Wiley & Sons, Ltd.  Archaeol. Prospect.  (2014)DOI: 10.1002/arp  (Bigman  et al ., 2011). Although the geotechnical and de-sign characteristics of mounds varied (Schilling, 2012),they probably represented monuments of memory(Pauketat and Alt, 2003) or the ascension of new leader-ship (Hally, 1996), and were the material representa-tions of social networks (Knight, 1998; King, 2004). Previous investigations Early informal investigations at Ocmulgee date backover a century, when Jones (1999 [1873]) produced ahand-drawn illustration of Mound A. In his depiction,the northern corners of the mound consisted of rampsof equal length and equal slope compared with theramp running down the northern face. However, morerecent investigators have questioned the accuracy of  Jones ’  drawings (e.g. Williams, 1992).The earliest excavations on Mound A ’ s summit bythe Civil Works Administration (CWA) began in1934, but details about these excavations were notpublished until 30 years later (Ingmanson, 1964).Ingmanson (1964, p. 9) reported that a 3.05 × 4.6 m testpit was excavated to approximately 9.15 m depth (allmeasurements have been converted to metric units).In 1937, excavations continued with two paralleltrenches cut approximately along the northern andsouthern edges of the summit. Sod located betweenthese trenches was stripped upon completion of the ex-cavations in hope of discovering post holes, pits or Figure 2. Aerial photograph of Ocmulgee National Monument with mounds mentioned in text labelled. Investigating Mound A at Ocmulgee National Monument Copyright © 2014 John Wiley & Sons, Ltd.  Archaeol. Prospect.  (2014)DOI: 10.1002/arp  other features related to architecture or summit occu-pation. The location, orientation and dimensions given by Ingmanson for these excavations were contradictedin later reporting (Stoutamire  et al ., 1983) (Figure 4).ThereisnoconclusiveevidencesuggestingMississippianinhabitants built structures on the  󿬁 nal Mound Asummit. Arthur Kelly, the  󿬁 eld director of the CWAinvestigations, encountered a clay cap at 13 – 18 cm below the surface (Ingmanson, 1964, pp. 10 – 11) and a5-cm hard band of clay approximately 0.6 m below thesurface that was interpreted as the remains of an earlierclaycap.Kellyalsoencounteredseveralhearthsapprox-imately 2.44 m below the surface.The National Park Service (NPS) excavated trenchesinto Mound A in 1967 and returned to Ocmulgee in1978 to carry out additional testing (Walker, 1969;Stoutamire  et al ., 1983, p. 12). Only limited reporting of these investigations are currently available and thelocationsoftheseunitscanonlybeinferred(Figure4).De-spite a lack of clarity of excavation techniques, Walkerpointedly reported his interpretation of the mound ’ sintegrity and structure. He concluded that  ‘ erosion wasall but imperceptible; there had been a stepped ramp-way down the northern and eastern sides of the moundduring the next to last construction stage ’  (Walker, 1994,p.33),butnonelocatedonthenortheasternornorthwest-ern corners (Stoutamire and Walker, , p. 3).In 2000, the NPS assessed the potential impact of constructing a stairway on the eastern face of MoundA (Halchin, 2000; Kidd  et al ., 2004). The investigationsconsisted of 78 shovel tests, 50 × 50 cm, on the easternedge of the mound (Figure 4). Fifty-two test unitsencountered prehistoric artefacts below the  󿬁 rst fewcentimetres of soil and indicated undisturbed stratigra-phy. Strata at intermediate depth contained a highdensity of prehistoric ceramic material and the deepestlayer showed evidence of a probable clay cap (Kidd et al ., 2004, p. 21).Our review of previous research has illuminated sev-eral important questions about Mound A ’ s form, func-tion and construction history. During Jones ’ investigation in 1873, he observed evidence of rampson the northern mound corners, but Walker ’ s work in1978 contradicts these  󿬁 ndings. The CWA and NPS ex-cavations identi 󿬁 ed little evidence for summit architec-ture on the  󿬁 nal summit, but it remains unclear if architecture was constructed on previous summitstages. The fragmentary evidence from the summit of Mound A may also be a product of destructiveploughing. All three projects observed a clay cap onthe mound, but only the CWA investigations encoun-tered multiple clay caps. Walker ’ s work on the easternface of the mound suggests that the stepped ramp ledtothetopofalower/earliermoundstageprobablypriortothe 󿬁 naluseepisode,butitisunclearifanyothermor-phological changes occurred during this time. Finally,investigators have yet to place the changing features of Mound A into a larger context that would help explainsocial change at Ocmulgee during the use history of Mound A. Figure 4. Topographic map with locations of excavation units andshovel tests from National Park Service investigations, and the variedinterpretations of CWA mound summit excavations.Figure 3. Three-dimensional elevation model of Mound A and sur-rounding area created using differential global positioning system(DGPS) elevation data corrected in real-time by Omni Star. We col-lected data using a Trimble ProXRT GPS unit. Data were  󿬁 ltered using ArcGIS Version 10 to remove all data points with accuracy worse than0.5 m in any direction (   x  ,  y  , or  z   ). From the viewer ’ s perspective, thescene is looking east-southeast. This  󿬁 gure is available in colouronline at D. P. Bigman and P. M. Lanzarone Copyright © 2014 John Wiley & Sons, Ltd.  Archaeol. Prospect.  (2014)DOI: 10.1002/arp  Survey methods The authors collected GPR data on the summit of Mound A to elucidate several of our project goals.Ground-penetrating radar was considered an appro-priate geophysical tool as we expected strong dielectriccontrasts in pedologic and sedimentary units typi 󿬁 edat Ocmulgee based on stratigraphic interpretationsfrom early excavation reports. We utilized a GSSISIR-2000 control unit and 500 MHz antenna duringthe GPR investigation. A time window of 100 ns wasutilized in order to capture the maximum depth of penetration at the mound for the antenna ’ s centre fre-quency. A 30 × 40 m grid was centred at the top of Mound A orientated NE – SW using 0.5 m spacing be-tween transects. This allowed the summit of MoundA to be adequately sampled and coincided with thesame geometry as other shallow geophysical data col-lected in previous investigations (Bigman, 2012a). Asurvey wheel was utilized with a 0.05 m horizontalsampling interval to ensure a high-degree of spatialaccuracy in tying into topographic maps surveyedwith a differential global positioning system (DGPS)(Figures 3 and 4).The GPR data were processed and visualized usingGPR SLICE Version 6.0. Minimal processing was doneto raw two-dimensional pro 󿬁 les in order to preservesrcinal  󿬁 eld data. Data were reversed and horizontalsampling was applied using  󿬁 eld markers collectedfrom the survey wheel. Raw two-dimensional pro 󿬁 leswere processed using background and band-pass  󿬁 l-tering to remove noise and enhance visualization of targets (Conyers, 2013) for the  󿬁 nal images. Time-to-depth conversion was performed using hyperbolamatching at various depths from shallow to deeper in-tervals throughout the pro 󿬁 les in order to provide anaverage velocity  󿬁 eld across the three-dimensional vol-ume (for a more in-depth discussion on depth conver-sion techniques see Conyers and Lucius (1996)). Wecalculated an average velocity of 19.49 ns m  1 , whichwe used to carry our interpretations in the depth do-main, and we could relate these to the descriptions of previous excavations and prehistoric mound construc-tion periods. Our total depth of penetration was ~5 m,which is re 󿬂 ected in the two-dimensional pro 󿬁 les wehighlight throughout this paper. It was observed thatthe radar signal becomes signi 󿬁 cantly attenuated past70 ns (~3.6 m), which probably re 󿬂 ects increasing claycontent and the limit of vertical resolution in this geo-logical environment. As interpretations of depth-slices beyond 3.6 m may be prone to data quality issues, cau-tion was used in identifying possible anthropogenicfeatures at these depths.Using inverse distance interpolation, GPR depth-slices were gridded in three dimensions. Data weresliced using a 6-ns (~0.3 m) time window. This windowwas considered an optimal value because it allowed fordiscrete sampling over a short depth interval, where wewould predict that variations due to clay caps or otheranthropological features would be identi 󿬁 ed withoutthe interference of overburden or underburden mate-rials. The authors experimented with other time win-dows with which to vary slice thickness, and foundthat the 6-ns (~0.3 m) time window selected here wasan optimal range to identify archaeological featuresnoted throughout this study. Finally, for visualizationthe depth-slice data were geo-referenced into ArcGISfor  󿬁 nal data interpretation. Results The shallowest section surveyed by GPR resolved adiscrepancy between two reconstructions of the 1930ssummit excavations. Documented reconstructions dif-fer in location, size and shape of the excavation units.According to archived  󿬁 eld notes, Kelly excavatedMound A twice during the 1930s. In 1934, a deep testunit was excavated, and in 1937, a spatially large butshallow excavation stripped much of the mound ’ stop layer. Stoutamire  et al . (1983) suggested that thedeep excavation in 1934 was square in shape and cen-trally located on the mound summit, but Ingmanson(1964) reported that the unit was rectangular and offsetto the northwest (Figure 4). Ingmanson (1964) also in-dicated that the shallow 1937 test unit was broad,and reached the western edge of the mound with a jag-ged northwestern corner, whereas Stoutamire  et al .(1983) indicated that the shallow unit did not reachthe western edge and was square (Figure 4). The GPRdata helped to delineate the extent of these excavations(Figures 5a, 6a, 6b, 7a, and 7b). A two-dimensionalpro 󿬁 le that spans NE – SW across the mound displaysa coherent, high-amplitude re 󿬂 ection signature from0.2 to 0.5 m (4 – 10 ns) depth in the northeastern portionof the pro 󿬁 le. This contrasts with the incoherent, loweramplitude response in the southeastern corner(Figure 5a). The change in this amplitude andcoherency response is visualized across the mound ’ spalaeo-anthropogenic surface by the depth-sliceshown in Figure 6a. The rapid change in the GPRsignature seen here probably results from variationsin sediment composition and compaction, and likelyrepresents the difference between unconsolidated,heterogeneous excavation  󿬁 ll and unexcavated, hard-ened clay material. The high-amplitude re 󿬂 ections Investigating Mound A at Ocmulgee National Monument Copyright © 2014 John Wiley & Sons, Ltd.  Archaeol. Prospect.  (2014)DOI: 10.1002/arp
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks