Geoarchaeology and paleoenvironmental context of the Beacon Island site, an Agate Basin...

lable at ScienceDirect

Quaternary International xxx (2014) 1e23

Contents lists avai

Quaternary International

journal homepage: www.elsevier .com/locate/quaint

Geoarchaeology and paleoenvironmental context of the Beacon Islandsite, an Agate Basin (Paleoindian) bison kill in northwestern NorthDakota, USA

Rolfe D. Mandel a, *, Laura R. Murphy a, Mark D. Mitchell b

a Kansas Geological Survey & Dept. of Anthropology, University of Kansas, 1930 Constant Avenue, Lawrence, KS 66047, USAb Paleocultural Research Group, P.O. Box 745309, Arvada, CO 80006, USA

a r t i c l e i n f o

Article history:Available online xxx

Keywords:Northern Great PlainsAgate Basin complexOahe FormationPaleopedologyStable carbon isotopesPhytoliths

* Corresponding author.E-mail address: [email protected] (R.D. Mandel).

http://dx.doi.org/10.1016/j.quaint.2014.06.0731040-6182/© 2014 Elsevier Ltd and INQUA. All rights

Please cite this article in press as: Mandel, R.(Paleoindian) bison kill in northwestern No

a b s t r a c t

A geoarchaeological investigation that included soil-stratigraphic and paleoecological analyses wasconducted at Beacon Island, located in the flood pool of Lake Sakakawea in northwestern North Dakota.Beacon Island is a multi-component archaeological site that includes an Agate Basin (Paleoindian)component consisting of the butchered remains of at least 29 Bison antiquus, along with projectile pointfragments and other artifacts. The bison bonebed is a product of a single kill at ca. 10,300 B.P. and isburied in a shallow kettle basin. The basin formed around 16,000e15,000 B.P. and initially trapped loesscomprising the Mallard Island Member of the Oahe Formation; other members of the Oahe Formationsubsequently filled the basin. Conditions probably were relatively dry when the Mallard Island Memberaggraded, but by ca. 10,300 B.P. paludal deposits comprising the Aggie Brown Member began to accu-mulate in the basin, indicating a shift to wetter conditions. Based on phytolith and stable carbon isotopedata, cool-season C3 prairie species dominated the site at the time of the bison kill, and the micro-mammal faunal assemblage suggests that this boreal grassland may have been punctuated by stands ofshrubby vegetation. Slight warming and drying occurred soon after ca. 10,300 B.P., indicated by relativelyhigher d13C values determined on soil organic matter and a significant increase in concentrations ofmicroscopic particulate charcoal and drought-resistant Stipa-type phytoliths, but paludal depositscontinued to aggrade, resulting in deep burial of the bonebed. Sedimentation was relatively slow in thebasin, allowing soil development to keep up with deposition. This cumulization process resulted in theformation of an overthickened A horizon in the Aggie Brown Member, typical of the Leonard Paleosol.Aggradation of the Aggie Brown Member ceased by ca. 8000 B.P., and loess comprising the Pick CityMember began accumulating soon after that time, likely marking the initiation of the warm, dry Alti-thermal climatic episode. Isotope and phytolith data also point to a warmer and probably drier climateafter 8000 B.P., indicated by higher d13C values and the appearance of warm-season C4 chloridoids at thesite. Hence the Agate Basin occupation coincided with the coolest and perhaps the wettest climaticepisode recorded in the sampled deposits. However, it is likely that the people associated with the AgateBasin culture at Beacon Island did not experience environmental conditions dramatically different fromthe modern conditions at the site.

© 2014 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction

The Agate Basin complex was first defined more than 50 yearsago in the Great Plains (Wheeler, 1954; Roberts, 1961), yet it re-mains today among the least well understood Paleoindian techno-complexes in North America because of a paucity of recorded Agate

reserved.

D., et al., Geoarchaeology andrth Dakota, USA, Quaternary

Basin components in stratified contexts. In addition, surface finds ofAgate Basin projectile points are comparatively rare, particularlyrelative to Folsom points, which date very nearly to the same period(Stanford, 1999). Prior to our investigation at the Beacon Island site,only six sites with Agate Basin components had been excavated, allof them scattered along the far western margin of the Great Plains(Fig. 1) (see Stanford, 1999; Kornfeld et al., 2010). Among the AgateBasin components at these sites, only the work in Area 2 at theAgate Basin type-site in eastern Wyoming has been described indetail (see Frison and Stanford, 1982). Interpretations of collections

paleoenvironmental context of the Beacon Island site, an Agate BasinInternational (2014), http://dx.doi.org/10.1016/j.quaint.2014.06.073

mailto:[email protected]

www.sciencedirect.com/science/journal/10406182

http://www.elsevier.com/locate/quaint

http://dx.doi.org/10.1016/j.quaint.2014.06.073



Fig. 1. Map showing the locations of sites with excavated Agate Basin components.

R.D. Mandel et al. / Quaternary International xxx (2014) 1e232

from several other sites are complicated by the presence of multiplePaleoindian occupations (Haynes and Agogino, 1966; Frison, 1984;Sellet, 2001). Agate Basin-like points have been recovered fromsubsurface contexts at sites in Alberta, Oklahoma, Texas, and else-where, but questions remain about the technological relationshipsbetween these specimens and better known western Plains as-semblages (Wyckoff, 1985; Knudson et al., 1998; Fedje et al., 1999;Stanford, 1999). Recent analyses of curated faunal and chipped-stone collections from Agate Basin components at the Frazier,

Please cite this article in press as: Mandel, R.D., et al., Geoarchaeology and(Paleoindian) bison kill in northwestern North Dakota, USA, Quaternary

Hell Gap, and Agate Basin sites provide valuable new insights(Sellet, 1999; Borresen, 2002; Slessman, 2004; Hill, 2008; Byers,2009; Lee et al., 2011), but investigation of the newly discoveredBeacon Island site (32MN234) offers a unique opportunity to un-derstand Agate Basin lifeways and the environmental conditionsthat prevailed before, during, and after the period of Paleoindianoccupation at the site.

Beacon Island is a multi-component archaeological site locatednear the confluence of the Missouri and Little Knife rivers in


Fig. 2. Map showing the location of the Beacon Island site in northwestern North Dakota.

R.D. Mandel et al. / Quaternary International xxx (2014) 1e23 3

northwestern North Dakota (Fig. 2). Covering approximately 0.9 ha,the site contains evidence of multiple Paleoindian through LatePrehistoric occupations. The site has been arbitrarily partitionedinto four non-contiguous sectors, designated A, B, F, and T (Fig. 3);artifacts and features also occur outside these defined areas. Area A(Fig. 4), the focus of this paper, contains the Agate Basin occupation,along with at least two ephemeral Holocene-age occupations.

Spread across the sloping sides of a shallow kettle basin (Fig. 5),the Agate Basin component at Beacon Island consists of thebutchered remains of at least 29 Bison antiquus, along with pro-jectile point fragments, other stone tools, and flaking debris (Leeet al., 2012; Mitchell and Johnston, 2012). The original rim of thebasin was not exposed during the excavation and may no longer bepreserved because of erosion associated with fluctuating waterlevels of Lake Sakakawea. However, the distribution of till to thenorth and east suggests that the basin was no more than about ameter deep at the time of the kill. Geomorphic and stratigraphicdata indicate that the floors of nearby basins were lower and mayhave held more water during the period of Agate Basin occupation(Timpson, 2003). The grade of the north and east slopes of thebasin, fromwhich the majority of faunal remains and artifacts wererecovered, varies from 5 to 10 percent. The floor of the basin is just50 or 60 cm below the highest preserved remnant of the bonebed.

This paper presents the results of geoarchaeological and pale-oenvironmental investigations in Area A. The geoarchaeologicalinvestigation focused on the stratigraphy, sedimentology, paleo-pedology, and geochronology of the site. The paleoenvironmentalinvestigation focused on obtaining multiple proxies (i.e. stablecarbon isotopes and microbotanicals) from the soil. Together, thegeoarchaeological and paleoenvironmental information gleanedfrom this study formed a cohesive picture of past environments,site context, and site formation processes.

Previous geoarchaeological investigations by Timpson andAhler, (2003) and Timpson (2003) were instrumental indescribing the geomorphology of the Beacon Island site and placingthe cultural deposits into a well-defined stratigraphic context.However, those studies did not provide detailed information aboutthe soils at the site. For example, descriptions of soil profiles (i.e.,soil horizons) were not presented, and little was said about soil-forming processes that may have affected the archaeological re-cord at the site. Understanding the paleopedology of the site is


especially important because the cultural deposits are associatedwith a buried paleosol. Also, recognizing the soil stratigraphyprovides a basis for identifying cycles of sedimentation and land-scape stability, which in turn is crucial for reconstructing the his-tory of basin filling at the site. Our study not only details thepaleopedology of Beacon Island, but it builds on the work of pre-vious investigations by providing additional stratigraphic andsedimentological information along with new radiocarbon ages.

This paper also presents the results of a combined isotopic andmicrobotanical analysis of soils at the Beacon Island site. Twocomponents of the soils were analyzed to infer changes in vege-tation through time and especially during human occupation at thesite: plant phytoliths, and stable carbon isotopes of soil organicmatter. In addition, microscopic particulate charcoal concentra-tions were determined for soils at the site. Particulate charcoalconcentrations from soils and sediments reveal past fire history(e.g., Patterson et al., 1987; Morrison, 1994; Scott, 2002), and whenpaired with phytolith data, aid interpretation of natural fire fre-quency versus anthropogenic burning (see Kealhofer, 1996; Boyd,2002; Piperno, 2006; Cordova et al., 2011). Understanding thepaleoenvironmental context of Beacon Island, including thecomposition of former plant communities, is crucial to under-standing the relationships between landscape evolution, climaticfluctuations, and human activities at the site.

Few published Holocene paleoenvironmental records exist forthe grasslands of the Northern Plains, and even fewer extend to thelate Pleistocene (e.g., Beaudoin, 1993; Cyr et al., 2011). However,previous studies in the region have gleaned paleoenvironmentalinformation from various proxies, including lithostratigraphy andpedostratigraphy (Clayton et al., 1976; Mason et al., 2008), pollen(Yansa, 2006), stable nitrogen and carbon isotopes of bison collagen(Leyden and Oetelaar, 2001), microscopic particulate charcoal(Boyd, 2002), opal phytoliths (Fredlund and Tieszen, 1997; Boyd,2002, 2005; Cyr et al., 2011), and stable carbon isotopes of soilorganic matter (Cyr et al., 2011). Valero- Garc�es et al. (1997)demonstrated the effectiveness of comprehensive multi-proxypaleoenvironmental reconstructions in the Northern Plains thatinclude stratigraphy, pollen, diatoms, and isotope geochemistry.

The episode of human occupation represented by the AgateBasin component at Beacon Island coincides with the YoungerDryas Chronozone (YDC) (ca. 11,000e10,000 B.P.), a period of cooler


Fig. 3. Plan and cross-section of Beacon Island, showing the locations of archaeological areas defined in May 2002.


climate compared to the preceding Bølling/Allerød interstadial(Broecker et al., 1989; Alley et al., 1993; Mayewski et al., 1993; Alley,2007). Although the YDC is often described as a cool, dry period(e.g., Alley, 2007), precipitation and run-off patterns varied, both ona continental and sub-continental scale (Meltzer and Holliday,2010). Several studies indicate that the Northern Plains


experienced cool, moist conditions during the time of the YDC. Forexample, Artz (1995, 2000) suggested that a cool, moist climatesupported a grassland community in the Northern Plains duringthe late Wisconsin and early Holocene, prior to the onset of theAltithermal at ca. 8000 B.P. A pollen record from Moon Lake, NorthDakota, shows that climate was cool and moist with high effective


Fig. 4. Aerial view of Beacon Island, looking south.


moisture between 11,700 and 9500 B.P. (Valero- Garc�es et al., 1997).Also according to Mason et al. (2008), the Leonard Paleosol, whichtypically dates to ca. 10,000e8000 B.P., formed at a time with higheffective moisture in the Northern Plains. At the Beacon Island site,Ahler (2003) suggested that during the time of Agate Basin occu-pation the plant community probably was a mesic C3 grasslandsuitable for bison herds. One of the primary objectives of ourpaleoenvironmental investigation was to confirm or refute Ahler'shypothesis.

2. History of archaeological research on Beacon Island

The archaeology of Beacon Island was first documented in 1974(Haberman and Schneider, 1975). By that time, Lake Sakakawea,one of the major reservoirs constructed by the U.S. Army Corps ofEngineers on the Missouri River, had inundated portions of the site.The surface elevation of the lake was 561.44m, just 2.4 m below themaximum “normal operations” pool elevation, and as a result theBeacon Island site was exposed on five small islands. The twolargest islands were surveyed, resulting in the discovery of a heartheroding from an active cutbank, and several more hearths werefound exposed on the wave-cut beach (Haberman and Schneider,1975:69e70, 75). Also, concentrations of bone, chipped stonetools, pottery, burned rock, and flaking debris were recorded. Pro-jectile points recovered during the survey included a Folsom base, apossible Late Paleoindian midsection, and a series of Archaic andLate Prehistoric fragments. Haberman and Schneider(1975:117e121) ranked Beacon Island among the most significantsites they documented in their shoreline survey.

In the years following this initial survey, local collectors pickedup projectile points from the site ranging in age from Clovis to LatePrehistoric (Ahler and Crawford, 2003). Among these items is acollection of 12 Folsom points and point preforms described andanalyzed by Ahler et al. (2002). In early 2000, owing to widespreaddrought, the level of Lake Sakakawea dropped, exposing a largerarea than was visible in 1974, and visitors to the site began findingnumerous Agate Basin points on the north side of the island. The


large number of recovered points, along with the presence ofnumerous bison bone fragments, suggested that an Agate Basin killor butchery locality was preserved in this part of the site. In 2001,the State Historical Society of North Dakota (SHSND) providedfunding for archaeological testing and evaluation work designed todocument the extent and condition of Agate Basin cultural depositson the island.

Initial reconnaissance confirmed the presence of intact archae-ological deposits exposed at the surfaceedeposits that appeared tocontain abundant bison remains and that covered an area at leasttens of square meters in size (Timpson and Ahler, 2003). Intensivefieldwork, directed by Stanley A. Ahler, began in May 2002 andinvolved systematic surveying, hand coring, and excavation of 121 m2 test units into and through the Agate Basin bonebed (Ahler,2003). A second round of testing, funded by the U.S. Army Corpsof Engineers (USACE), was carried out in September 2002 in orderto gather information necessary for a larger-scale mitigation anddata recovery project and to obtain additional data from the part ofArea A most vulnerable to shoreline erosion (Ahler et al., 2002).

In 2005, the National Park Service awarded a Save America'sTreasures grant to the SHSND for data recovery excavations in AreaA (Ahler et al., 2006a). The USACE and the SHSND also providedsubstantial funding for the data recovery effort (Ahler et al., 2006b).In summer 2006, the research team opened four excavation blocksin Area A in order to obtain data on Agate Basin hunting and bisoncarcass processing practices, site seasonality, and lithic technology(Fig. 6) (Mitchell, 2012a).

The archaeological collection from Area A includes 3473 piece-plotted bison bones, 240 kg of unplotted bone, 155 stone tools,and an estimated 3134 pieces of flaking debris (Lee et al., 2012;Mitchell and Johnston, 2012). Four of five radiocarbon ages deter-mined on materials other than soil organic matter from the site,including two determined on charcoal, one on tooth material, andone on bone, are statistically equivalent (Table 1). The weightedmean age of these four ages is 10,328 ± 28 B.P. A fifth assay oncharcoal produced a non-contemporaneous, slightly younger age of9911 ± 105 B.P.


Fig.

5.Geo

logicalcross-sectionof

AreaAof

theBe

acon

Island

site.T

hebe

achridg

esareassociated

withLake

Saka

kawea

.


Please cite this article in press as: Mandel, R.D., et al., Geoarchaeology and paleoenvironmental context of the Beacon Island site, an Agate Basin(Paleoindian) bison kill in northwestern North Dakota, USA, Quaternary International (2014), http://dx.doi.org/10.1016/j.quaint.2014.06.073

Fig. 6. Map showing the locations of Profiles 1, 2, and 3 in Area A of the Beacon Island site.

Table 1Results of AMS radiocarbon assays for samples from Area A of the Beacon Island site.a

Context Material assayed 14C age (yr B.P.) Cal. ageb (yr B.P.) d13C(‰) Laboratory No.

2Ak1b horizon, 0.36e0 46 cmbs, Profile 2 SOM 7980 ± 25 8981e8778 �24.2 ISGS-A17742Ak3b horizon, 0.52e0.62 cmbs, Profile 2 SOM 8740 ± 25 9765e9630 �24.0 ISGS-A1773Agate Basin bonebed, northwest block Bone 10,330 ± 45 12,375e12,049 NR CAMS-90966Agate Basin bonebed, north block Tooth 10,305 ± 45 12,370e11,989 NR CAMS-90966On the ABM-MIM contact, west block Charcoal 10,338 ± 82 12,384e12,046 �20.7 ETH-26779On the ABM-MIM contact, west block Charcoal 10,371 ± 80 12,390e12,095 �27.3 ETH-26780On the ABM-MIM contact, west block Charcoal 9911 ± 105 11,602e11,218 �24.5 ETH-26781

Abbreviations: SOM ¼ Soil organic matter, NR ¼ Not reported, ABM ¼ Aggie Brown Member, MIM ¼ Mallard Island Member.a Radiocarbon ages determined on the bone, tooth, and charcoal samples were reported and described by Ahler et al. (2012).b Calibration to calendar years at 1s (68.2% probability) was performed with OxCal v4.17 (Bronk Ramsey, 2010) using calibration dataset IntCal09 (Reimer et al., 2009).


Several lines of evidence indicate that the Agate Basin bonebedat Beacon Island represents a single, short-term occupation. Onaverage, the bonebed is just 5e10 cm thick. Statistical analysis in-dicates that the standard deviation of mean plotted bone elevationsis 0.0385 ± 0.0138 m, which yields a one-sigma bonebed thicknessof 7.7 ± 2.8 cm, or a median thickness of 7.1 cm. (Mitchell, 2012b).No internal stratigraphy was observed during excavationwithin thebonebed and no gaps or discontinuities exist in the vertical distri-bution of bones. The bonebed occurs on a single surface conformingto the original topography of the basin, as modeled by elevationdata on the contact between the Aggie Brown Member and the


Mallard Island Member of the Oahe Formation, located immedi-ately below the bonebed.

In addition, data on bone preservation and on the distributionsof faunal remains and lithic refits and conjoins suggest that discreteactivity areas are preserved at Beacon Island. Faunal remains areunevenly distributed across the site, but by far the densest con-centration of bison bone occurs in the northwest block (Fig. 7). Evenwithin the northwest block, bone is unevenly distributed, with thelargest mass in the northern part of the block located on the upperside-slope of the kettle basin. A total of 32 chipped stone artifactsrefit or conjoin to form 13 composite items, including nine


Fig. 7. Map of the Agate Basin component in Area A of the Beacon Island site showing the locations of plotted bones and unplotted bone weight.

Fig. 8. Profile 1 in Area A of the Beacon Island site showing three members of the Oahe Formation and the stratigraphic position of the Agate Basin cultural component.


Please cite this article in press as: Mandel, R.D., et al., Geoarchaeology and paleoenvironmental context of the Beacon Island site, an Agate Basin(Paleoindian) bison kill in northwestern North Dakota, USA, Quaternary International (2014), http://dx.doi.org/10.1016/j.quaint.2014.06.073


projectile points, one flake tool, one tested cobble, one core, andone unmodified flake. Themean distance between plottedmatchedspecimens is 53.2 cm, with a range of 3e126 cm (Mitchell andJohnston, 2012). These data suggest limited post-occupationdisturbance of the bonebed and associated activity areas. An un-lined basin hearth, measuring roughly 50 cm across, was docu-mented near the center of the kettle basin (Mitchell, 2012b). Thehearth originated in the lower portion of the Aggie BrownMember,approximately 10 cm above the undulating, patchy contact be-tween the Aggie Brown and the underlying Mallard IslandMember.Slopewash had transported a small number of bison bones over thefeature, but the primary post-occupation disturbance process wasanimal burrowing, again suggesting limited lateral movement ofartifacts and faunal remains.

The 2006 data recovery in Area A also involved a geo-archaeological field investigation and a paleoenvironmentalreconstruction derived from analyses of soil samples (Mandel,2012). The soil-stratigraphic context of the bonebed was deter-mined, and bioclimatic conditions were inferred from phytolithassemblages and stable carbon isotope values of soil organicmatter.The results of these studies are summarized in this paper.

3. Environmental setting

3.1. Physiography and geology

The Beacon Island site is in the Coteau du Missouri division ofthe Missouri Plateau. The Missouri Plateau is the largest sub-province of the Great Plains Physiographic Province (Fenneman,1931). The Coteau du Missouri, or “hills of the Missouri,” is aglacial landscape characterized by gently rolling end morainesinterspersed with kettle holes. This landscape formed during theMankato Substage of the Wisconsin glacial episode (Flint, 1955;Moran et al., 1976; Clayton and Moran, 1982; Mickelson et al.,1983), and runs roughly parallel to the Missouri River in a longdiagonal belt through North Dakota.

The Coteau du Missouri is separated from the main part of theMissouri Plateau by the Missouri River trench. The term “trench” ismore appropriate than “valley” because the cut made by the riverinto the plateau is deep, generally narrow, and very steep. Thebluffs range in height from 100 m to more than 150 m above thetrench floor. The Missouri River trench is considered a product ofglacial meltwater cutting into the soft bedrock (Flint, 1955;Bluemle, 2000).

Zones of closely spaced, steep-sided ravines descending into theMissouri River trench commonly are referred to as “breaks.” Theseravines or draws often extend several kilometers back into theadjacent plateau.

Beacon Island is a remnant of a high strath terrace of the Mis-souri River that was cut into bedrock (Ahler and Crawford,2003:20). This terrace forms the interfluve between the MissouriRiver and the Little Knife River. Following the initial period(s) offluvial erosion, Beacon Island was no longer part of the alluvialenvironment.

The Paleocene-age Sentinel Butte Formation, a bedrock unitgenerally consisting of poorly lithified siltstones, claystones, andmudstones with subordinate amounts of lignite and fine-grainedsandstones (Clayton et al., 1980), underlies Beacon Island. Out-crops of the Sentinel Butte Formation are common in the areaaround the island. Also, ice-thrust blocks of intact Sentinel Butterocks occur at multiple locations along the exposed face of the hillthat forms the highest ground in the southwestern part of the is-land (Ahler and Crawford, 2003:21).

At Beacon Island, late-Wisconsinan glacial deposits mantle theSentinel Butte Formation. According to Clayton (1972), the terminal


edge of the Wisconsinan ice front was in the area of Beacon Islandaround 16,000e15,000 B.P., and he named the glacial depositsassociated with this ice front the Lostwood Drift. Clayton suggestedthat an insulating blanket provided by supraglacial sedimentdelayed the final melt-out of the Lostwood ice until 9000e8000 B.P.However, based on observations made in previous studies (e.g.,Timpson and Ahler, 2003; Ahler and Crawford, 2003) and the cur-rent investigation, there is no geomorphological or sedimentolog-ical evidence indicating that ice was present at Beacon Islandduring the early Holocene.

The Lostwood Drift is comprised of pebbly to cobbly matrix-supported till that has been modified by soil development. Theprominent hill on the island mostly consists of a thick deposit ofglacial till that has not yet been removed by lake erosion. Elsewhereon the island a thinner deposit of till of the same age occurs. Erosionof the till by fluctuation of Lake Sakakawea has produced promi-nent boulder lag deposits, especially along the shore of the island.

Prior to the impoundment of Lake Sakakawea, it is likely thateverywhere on Beacon Island silty eolian deposits mantled theglacial till. These eolian deposits comprise the Oahe Formation, aformal lithostratigraphic unit named by Clayton (1972) anddescribed and defined by Bickley (1972) and Clayton et al. (1976).The Oahe Formation was originally defined to include only sedi-ment that contains a large proportion of silt and little or no sand orgravel. This material was considered to be largely eolian sedimentdeposited on uplands and high alluvial terraces during late-Wisconsinan and Holocene times. Clayton et al. (1976) subdividedthe Oahe Formation into four members: the Mallard Island, AggieBrown, Pick City, and Riverdale. Buried soils that developed in theAggie Brown and Riverdale members were informally named theLeonard and Thompson paleosols, respectively.

Although members of the Oahe Formation were defined on thebasis of color, Clayton et al. (1976) suggested that sedimentscomposing these deposits were emplaced on the landscape duringspecific periods. They proposed the following chronology for themembers: Mallard Island (ca. 13,000e10,000 B.P.), Aggie Brown (ca.10,000e8000 B.P.), Pick City (ca. 8000e5000 B.P.), and Riverdale(5000 B.P.-modern). However, this chronology is not based onactual radiocarbon ages determined onmaterials from the deposits.Instead, lacking adequate data for numerical dating, Clayton et al.(1976) proposed a simple model of hill-slope response to climaticchange in order to correlate Oahe Formation members withnumerically dated late-Quaternary climatic episodes. According tothis model, soils formed during relatively moist climatic episodes,when hillslopes and bluffs were stable because of adequate vege-tative cover. During intervening dry periods, vegetative cover wasreduced on uplands, resulting in landscape instability and hillslopeerosion. Dry periods also experienced frequent dust storms anddeposition of eolian sediments on the bluffs. Clayton et al. (1976)suggested that the Aggie Brown Member and associated LeonardPaleosol formed during the cool, moist climates of the terminalWisconsin and early Holocene, and that the Pick CityMember datedto the warm, dry middle Holocene. The Riverdale Member andassociated buried soils were considered products of relatively dryand moist late-Holocene episodes, respectively.

The chronology originally proposed by Clayton et al. (1976) forthe Oahe Formation is generally accurate. However, archaeologicalstudies such as those at Beacon Island continue to refine it. Also,some aspects of Clayton's climatic model have proven to be difficultto confirm (Artz, 1995, 2000). For example, Coogan (1983, 1987)determined that the entire package of sediments comprising theOahe Formation is rarely preserved in any one locality because theHolocene geomorphic record on uplands adjacent to the MissouriRiver Trench is primarily erosional, not depositional. The eolianrecord at many localities in the Dakotas appears to be entirely late



Holocene in age (Artz, 2000). Also, evidence confirming thepaleoclimatic implications suggested by Clayton and his coworkershas been elusive (Barnosky et al., 1987). Nevertheless, the OaheFormation model continues to be used to interpret the naturalstratigraphy of archaeological sites and to estimate their relativeages.

In 1979, the Oahe Formation was expanded to include all sedi-ments above the late-Wisconsinan Coleharbor Group (Clayton andMoran, 1979). As such, the formation is no longer restricted toeolian silt (loess) similar to the type section; it includes clay, silt,sand, and gravel. Clayton et al. (1980) recognized three lithogeneticsubdivisions of the Oahe Formation: (1) eolian, (2) paludal/lacus-trine, and (3) alluvial. Also, Bickley et al. (1970), Bickley and Clayton(1971) and Cvancara et al. (1971) presented an informal lithos-tratigraphic sequence for the paludal lithfacies of the Oahe For-mation. At Beacon Island, the Oahe Formation includes eolian,colluvial, and paludal lithofacies.

Archaeological and geoarchaeological investigations at BeaconIsland have focused on the Aggie Brown Member, which is repre-sented by the thick, organic-rich cumulic A horizon of the LeonardPaleosol. Most of the Paleoindian cultural deposits are in the lower10e15 cm of the Aggie Brown Member, though some extend into a“mixing zone” below the Aggie Brown and continue into the upper10 cm of the Mallard Island Member.

3.2. Climate

With the exception of the southwest corner of the state, NorthDakota is in Thornthwaite's (1948) Dry Subhumid (C1) climaticregion. The C1 climate is characterized by hot, dry summers andcold, dry winters. Most of the precipitation occurs in spring andearly summer. The mean annual precipitation at New Town, NorthDakota, located about 5 km southeast of Beacon Island, is 38.35 cm(High Plains Regional Climate Center, 2011). The average dailymaximum and minimum temperature in January is 8.5 �Cand �19.6 �C, respectively. The average daily maximum and mini-mum temperature in July is 28.3 �C and 13.3 �C, respectively (HighPlains Regional Climate Center, 2011).

3.3. Vegetation

The Great Plains is dominated by tall-grass, mixed-grass, andshort-grass prairies comprised of the three major Poaceae (grass)subfamilies: Chloridoideae, Panicoideae, and Pooideae. The climateof the Northern Plains, including northwestern North Dakota,largely supports C3 grasses of the Pooideae subfamily (Fredlund andTieszen, 1994; Rovner, 2001; Cyr et al., 2011). Native species of thePoaceae family found in the vicinity of Beacon Island today includePooideae varieties of wheatgrass (Agropyron), wild rye (Elymus),fescue (Festuca), needlegrass (Hesperostipa, e.g. Stipa), brome(Bromus), and bluegrass (Poa) (USDA, 2011). A smaller proportion ofthe plant community is made up of C4 Chloridoideae and Pan-icoideae, including varieties of bluestem (Andropogon), grama grass(Bouteloua), and three-awn (Aristida).

4. Methods

4.1. Field methods

The geoarchaeological investigation initially involved a recon-naissance of Beacon Island. At this early stage of the study, land-forms identified on topographic maps and aerial photographs werefield-checked for accuracy.

Following the reconnaissance, three profiles in Area A wereselected for study and sampling (Fig. 6). Detailed descriptions of the


three soil profiles were made in the field using standard proceduresand terminology outlined by Birkeland (1999) and Scheonebergeret al. (2002). However, in Profile 3, where more than one buriedsoil occurs, the buried soils are numbered consecutively from thetop of the profile downward, with the number following the suffix“b” (after Holliday, 2004:339). Stages of carbonate morphologywere defined according to the classification scheme of Birkeland(1999:Table A1e5).

For the purpose of determining grain-size distribution, organiccarbon content, and stable carbon isotope (d13C) values, 29 soilsamples from Profile 1 and 29 samples from Profile 2, eachapproximately 200 g, were collected at 5 cm intervals, exceptwhere stratigraphic breaks occur. Bulk soil samples spanning eachof the eight soil horizons in Profile 3 were collected, and eight soilsamples were collected from Profile 2 for phytolith analysis.

4.2. Laboratory methods

Bulk soil samples from the profiles were air-dried at the KansasGeological Survey's Geoarchaeology Laboratory and mechanicallysplit into equal halves. One split of samples was decalcified with0.5 N HCl and submitted to the Keck Paleoenvironmental andEnvironmental Stable Isotope Laboratory (KPESIL) at the Universityof Kansas to determine organic carbon (C) content. Those sampleswere analyzed on a Costech ECS 4010 Elemental Analyzer inconjunction with a series of atropine standards (Costech Code031042) of known percent C. From the analyzed standards, theCostech EAS32 software generates a calibration curve measuringarea (Vs) versus weight (mg C). Knowing the carbon content of thestandard and noting individual sample weights along withmeasured voltages, the software generates relative % C content foreach analyzed sample. Typical standard calibration r2 values arebetter than 0.9998.

The grain-size distribution of the samples was determined usinga slightly modified version of the pipette method (Gee and Bauder,1986). The samples were dispersed in a sodium hexametaphos-phate solution and shaken on a reciprocal shaker overnight. Silt andclay aliquots were drawn from the appropriate pipette depth basedon particle-size settling velocity, oven dried, and weighed to thenearest milligram. Wet sieving recovered the sand fraction. Theresults, presented as weight percentages, total to 100% of the<2 mm mineral fraction. Loess standards were used for inter-runcomparisons of grain-size data.

Two bulk soil samples from the paleosol developed in the AggieBrown Member were submitted to the Illinois State GeologicalSurvey Isotope Geochemistry Laboratory. The samples underwentstandard pretreatment for radiocarbon dating, including removal ofrootlets and calcium carbonate. Radiocarbon ages were determinedon the total decalcified soil carbon at the Keck Carbon Cycle AMSLaboratory, University of California-Irvine. These radiocarbon ages,as well as AMS ages determined on bone and charcoal, are reportedin uncalibrated years B.P. in the text, and in uncalibrated and cali-brated years (cal B.P.) in Table 1.

Soil samples were processed for stable carbon isotope analysisat the Kansas Geological Survey. The samples were dried in an ovenat 50 �C, and homogenized with a ceramic mortar and pestle.Samples were pretreated by adding 20 ml of 0.5 N HCl solution to1 g of soil to remove calcium carbonate. After the reaction wascomplete, 30 ml of distilled water were added to each sample andcentrifuged at 4000 RPMs for five minutes and decanted. Theprocess was repeated to ensure chlorine removal. Decalcifiedsamples were dried at 50 �C, pulverized using a synthetic rubymortar and pestle, and transferred to vials.

Prepared isotope samples from Profiles 1 and 3 in Area A wereanalyzed at KPESIL. However, samples from Profile 2 were sent to


Fig. 9. Diagram showing the stratigraphy, soil horizonation, and distribution of cultural deposits in Profile 1. The particle-size distribution, organic carbon content, and d13C valuesof soil samples collected from Profile 1 are also shown.


the Stable-Isotope Biogeochemistry Laboratory at McMaster Uni-versity. At both isotope laboratories, raw d13C values were obtainedusing a Costech element analyzer coupled with a Thermo-Finniganisotope ratio mass spectrometer. National Institute of Standardsand Technology (NIST) standards used to calibrate d13C values wereNIST USGS-24 (graphite) #8541, IAEA-600 (caffeine), and NIST ANU(sucrose) #8542. A pre-calibrated internal standard (DORM-2dogfish muscle; National Research Council of Canada) was used inthe d13C calibration curve. The precision of reported d13C values isbased on a linear correction of observed values versus expectedvalues of the standards. Typical standard calibration curves yield anr2 of 0.9994 or greater.

Eight soil samples representative of the Pick City, Aggie Brown,andMallard Island deposits from Profile 2 in Area Awere processedfor phytoliths and microscopic particulate charcoal using a modi-fied procedure described by Piperno (2006). This procedure isolatesthe phytoliths contained within the silt fraction by first removingthe clay and sand fractions, and then separating biogenic silica fromsilt particles through heavy liquid flotation. Each sample waspassed through a 2 mm sieve to remove sand and then 5 g wereweighed. Twenty-five ml of 10% HCl were added to each 5 g sampleto remove carbonates, samples were rinsed after the reaction wascomplete, and then 75ml of sodium pyrophosphate solution (71.4 gadded to 1600 ml of DI water) were added to each sample todisperse clay. The samples were then centrifuged at low speeds(1500e2000 RPMs) and decanted to remove clay while retainingphytoliths within the silt fraction; this process was repeated untilthe supernatant was clear and all clay was removed.

All phytolith samples from the Aggie Brown Member containedlarge quantities of very fine organic material and charcoal that


remained in the supernatant during clay removal centrifugations.The organic material was removed with a pipette after the first fivecentrifugations and transferred to 250 ml centrifuge bottles. Thebottles were centrifuged at 2500 RPMs for 15 min, decanted, andcentrifuged again at 2500 RPMs for 5 min and decanted to removeclay-sized organic particles in suspension. The remaining organicmaterials were then transferred back to the original samples.

After oxidation with hydrogen peroxide, 3 lycopodium sporetablets, each containing 18,585 spores, were added to each sampleto calculate phytolith concentrations (after Bozarth, 1992). Phyto-liths were then isolated by adding 20 ml of zinc bromide with aspecific gravity of 2.40 g/ml to each sample, and centrifuging at2500 RPM for one hour. Phytolith isolation and thin sectionmounting follow methods used for plant materials published byPiperno (2006). Samples were mounted on 25 � 75 � 1 mmmicroscope slides in Type A immersion oil. A standard petrographicmicroscope was used to analyze each slide at 40� and 100�magnification. Phytoliths were tallied by transect until at least 200cells were counted for each slide. Phytolith concentrations (phy-toliths per gram of treated sediment) were calculated with thefollowing formula: (# observed phytoliths � # lycopodium sporesintroduced/#observed lycopodium spores)/original sample weight.Recovered phytoliths with identifiable subfamilies and generawerecompared to a phytolith reference collection at the University ofKansas Geography Department. Charcoal particles were counted asindependent units within each phytolith slide, and then charcoalconcentrations were calculated with the same lycopodium-basedprocedure.

A digital pHmeterwas used to determine the pH of the eight soilsamples from Profile 2 that were selected for phytolith analysis.



Because biogenic silica generally goes into solution at pH 8.5, it wasimportant to determine the potential impact soil pH had on phy-tolith preservation.

5. Results

5.1. Stratigraphy and soils

5.1.1. Profiles 1 and 2Profiles 1 and 2 in the southwestern portion of the site are

representative of the soil stratigraphy observed near thedeepest part of the kettle basin. In both, the contact betweenthe Aggie Brown Member and the underlying Mallard IslandMember occurs at a grid elevation of 987.21 m. This is thelowest point in the kettle basin. To the north and east, close tothe original rim of the kettle basin, this contact occurs at about987.95 m. In Profile 3, located 9 m east and 2 m north of

Table 2Description of Profile 1 in Area A of the Beacon island site.

Depth (cm) Soilhorizon

Texture Structure Lowerboundary

Colo

Modern lakebed sand0e3 e e e e e

Pick City Member3e24 Bk1 SiL 1, f, pr~1, f, sbk G, S Ligh

comope

24e36 Bk2 SiL 2, m þ f, pr~1, f, sbk C, S Lighmanope

Aggie Brown Member36e47 2Ak1b SiL 1, f, pr~1, f þ m, gr G, S Gra

threman

47e54 2Ak2b SiL 1, f, pr~1, f, gr C, S Gramoiman

54e62 2Ak3b SiL 1, f, pr~1, f, gr G, S Darandman

62e67 2Ak4b SiL 1, f, gr C, S Darmoiopeand

67e72 2Ak5b SiL 1, f, pr~1, f, gr G, S Darandcomfew

72e77 2Ak6b SiL 1, f, gr C, S Darcombur

77e102 2Ak7b SiL 1, f, gr G, S Verstruchaman�20man

102e107 2ABkb SiL 1, f, pr~1, f, sbk G, S Darbrostroin twordiam

Mallard Island Member107e131 3Bk1b SiL 1, m þ f, pr~1, f, sbk G, S Ligh

finestroin tof C�20


Profile 2, the Aggie Brown-Mallard Island contact occurs atabout 987.45 m. Butchered bison bone is more abundant in thenortheast excavation block, represented by Profile 3, than it isin the west excavation block, represented by Profiles 1 and 2(Lee et al., 2012).

In Profile 1, an 8-cm-thick veneer of modern lakebed sedi-ment overlies the Pick City Member of the Oahe Formation(Figs. 8 and 9). The Pick City Member is 28 cm thick and consistsof loess with a fairly homogenous silt loam texture (Tables 2 and3). The surface soil is represented by a truncated light brownishgray (10YR 6/2-2.5Y 6/2, dry) Bk horizon with stage I carbonatemorphology (Table 2); the A horizon has been stripped off by thewave action of Lake Sakakawea. A clear, smooth boundary sep-arates the Pick City Member from the underlying Aggie BrownMember.

r and selected features

t brownish gray (10YR 6/2), dark grayish brown (10YR 4/2) moist;mon films and threads of CaCO3; common worm casts; commonn insect burrows; many fine and very fine pores.t brownish gray (2.5Y 6/2), dark grayish brown (2.5Y 4/2) moist,y films and threads of CaCO3; common worm casts; commonn insect burrows; many fine and very fine pores.

y (10YR 5/1), very dark gray (10YR 3/1) moist; many films andads of CaCO3; many worm casts; many open insect burrows;y fine and very fine pores.yish brown (10YR 5/2), very dark grayish brown (10YR 3/2)st; common films and threads of CaCO3; many worm casts;y open insect burrows; many fine and very fine pores.k gray (10YR 4/1), very dark gray (10YR 3/1) moist; common filmsthreads of CaCO3; many worm casts; many open insect burrows;y fine and very fine pores.k grayish brown (10YR 4/2), very dark grayish brown (10YR 3/2)st; common films and threads of CaCO3; many worm casts; manyn insect burrows; few krotovina �20 cm in diameter; many finevery fine pores.k gray (10YR 4/1), very dark gray (10YR 3/1) moist; common filmsthreads of CaCO3; many worm casts; common open insect burrows;mon krotovina �20 cm in diameter; common fine and very fine andmedium roots; many fine and very fine pores.k grayish brown (10YR 4/2), very dark grayish brown (10YR 3/2) moist;mon films and threads of CaCO3; many worm casts; many open insectrows; few krotovina �20 cm in diameter; many fine and very fine pores.y dark gray (10YR 3/1) black (10YR 2/1) moist; weak fine granularcture; slightly hard, friable; common bison bones and flecks ofrcoal throughout the horizon; few films and threads of CaCO3;y worm casts; many open insect burrows; common krotovinacm in diameter; common very fine and few fine and medium roots;y fine and very fine pores; gradual smooth boundary.k grayish brown (10YR 4/2) to brown (10YR 4/3), very dark grayishwn (10YR 3/2) to dark brown (10YR 3/3) moist; common fine prominentng brown (7.5YR 4/6) mottles; few bison bones and flecks of charcoalhe upper 4 cm of the horizon; few films and threads of CaCO3; manym casts; many open insect burrows; common krotovina �20 cm ineter; many fine and very fine pores.

t olive brown (2.5YR 5/3), olive brown (2.5Y 4/3) moist; manydistinct yellowish brown (10YR 5/6) and common fine prominentng brown (7.5YR 46) mottles; small concentration of bone fragmentshe upper 8 cm of horizon (krotovina?); few films and common threadsaCO3; common worm casts; common open insect burrows; few krotovinacm in diameter; common fine and many very fine pores.


Table 3Grain sizea and organic carbon data for Profile 1, Area A.

Soilhorizons

Depth(cm)

Sand(%) Silt(%) Clay(%) Texturalclassb

Organiccarbon(%)

Total C F Total Total

Bk1 8e12 8.8 44.7 23.1 67.8 23.4 SiL 0.7712e17 11.1 39.4 25.0 64.4 24.5 SiL 0.8017e24 11.8 37.3 25.6 62.9 25.3 SiL 0.72

Bk2 24e30 10.5 36.4 26.4 62.8 26.6 SiL 0.8330e36 11.7 35.5 26.1 61.6 26.8 SiL 0.79

2Ak1b 36e41 6.6 29.2 34.9 64.1 29.2 SiCL 1.6641e47 6.8 31.0 39.0 70.0 23.3 SiL 1.81

2Ak2b 47e54 7.4 35.2 39.1 74.3 18.3 SiL 1.412Ak3b 54e58 12.0 37.5 37.7 75.2 12.8 SiL 2.22

58e62 9.8 37.7 40.0 77.7 12.5 SiL 1.992Ak4b 62e67 10.9 39.1 35.9 75.0 14.2 SiL 1.872Ak5b 67e72 10.0 36.4 40.1 76.5 13.4 SiL 2.182Ak6b 72e77 5.6 31.0 48.5 79.5 14.9 SiL 2.182Ak7b 77e82 4.6 30.3 47.8 78.1 17.3 SiL 2.00

82e87 6.8 31.2 45.2 76.4 16.8 SiL 1.9487e92 13.0 35.5 38.2 73.7 13.4 SiL 2.4492e97 12.1 37.3 34.9 72.2 15.7 SiL 1.9697e102 9.3 41.0 32.7 73.7 17.0 SiL 1.42

2ABkb 102e107 10.0 42.4 24.9 67.3 22.7 SiL 0.893Bk1b 107e112 13.0 44.8 19.6 64.4 22.7 SiL 0.45

112e117 8.7 49.0 19.0 68.0 23.3 SiL 0.29117e122 11.2 48.5 18.1 66.6 22.2 SiL 0.37122e127 12.0 51.4 15.8 67.2 20.9 SiL 0.38127e131 13.1 46.7 17.6 64.3 22.6 SiL 0.43

3Bk2b 131e136 20.3 36.4 19.6 56.0 23.7 SiL 0.35136e141 19.7 40.0 17.8 57.8 22.6 SiL 0.35141e146 18.6 46.0 15.3 61.3 20.0 SiL 0.28146e150 16.0 48.4 16.0 64.4 19.6 SiL 0.29

a Particle-size limits (mm): Sand (Total) ¼ 2.0e0.05; Silt (Total) ¼ 0.05e0.002,Coarse (C) ¼ 0.05e0.005, Fine (F) ¼ 0.005e0.002; Clay (Total) ¼ <0.002.

b Textural classes: SiL ¼ silt loam; SiCL ¼ silty clay loam.

Table 2 (continued )

Depth (cm) Soilhorizon

Texture Structure Lowerboundary

Color and selected features

131e150 3Bk2b SiL 1, m, pr~1, f, sbk e Light olive brown (2.5YR 5/4), olive brown (2.5Y 4/4) moist; many fine andmedium distinct yellowish brown (10YR 5/6) and few fine prominent strongbrown (7.5YR 46) mottles; common gray (10YR 6/1) and light gray (10YR 7/1)reduction zones along macro-pores; few to common fine threads of CaCO3; commonworm casts; common open insect burrows; few krotovina �20 cm in diameter;common fine and many very fine pores.

Abbreviations and symbols: Texture SiL ¼ Silt loam; Structure 1 ¼ weak, 2 ¼ moderate, f ¼ fine, m ¼ medium, pr ¼ prismatic, sbk ¼ subangular blocky, gr ¼ granular,~ ¼ parting to; Boundary G ¼ gradual, C ¼ clear, S ¼ smooth.


The Aggie Brown Member is 71 cm thick in Profile 1 and hasbeen modified by pedogenesis. The Leonard Paleosol spans theentire thickness of the Aggie Brown Member and is characterizedby an overthickened, cumulic A horizon (2Ak1be2Ak7b) formed insilty paludal deposits. Cumulic soils receive influxes of parentmaterial while pedogenesis is occurring, but the rate of sedimen-tation is so slow that soil development keeps up with deposition(Nikiforoff, 1949; Birkeland, 1999:165; Mandel, 2008). In such soils,the A horizon builds up through time, and it is not unusual for the Ahorizon to look stratified, i.e., it may consist of alternating dark andlight zones. Because cumulic soils have parent material continu-ously added to their surfaces, their features are partly sedimento-logic and partly pedogenic (Birkeland, 1999:165).

Overthickened A horizons like the one developed in the AggieBrownMember are common in cumulic soils (Birkeland, 1999:166).Organic enrichment of the paleosol is due to deposition of organic-matter-rich material from upslope when the soil was at the surface,to in situ organic-matter accumulation at the site while sedimentwas accumulating, or to a combination of both processes. The


cumulative origin of the Leonard Paleosol is indicated by the highlyirregular organic-carbon content with depth in the soil profile(Fig. 9 and Table 3). In noncumulative soils, organic carbon steadilydecreases with depth.

With the exception of the upper 5 cm of the 2Ak1b horizon,which is silty clay loam, the cumulic A horizon formed in the AggieBrown Member has a silt loam texture (Table 3). Fine silt tends todominate over coarse silt through most of the Aggie Brown Mem-ber. This probably occurs because fine silt was more readily trans-ported into the kettle depression by slopewash.

Secondary calcium carbonate accumulation is apparentthroughout the A horizon of the Leonard Paleosol in Profile 1(Table 2), but carbonate morphology does not exceed stage I. Also,there are many biogenic features, especially worm casts and openinsect burrows. Krotovina up to about 20 cm in diameter occur inthe 2Ak4b and deeper soil horizons within the Aggie BrownMember.

Bison bones comprising the Agate Basin cultural componentoccur in the 2Ak7b horizon and upper 4 cm of the 2ABkb horizon.The 2Ak7b horizon is the deepest and most organically enrichedportion of the cumulic A horizon, with organic carbon contentsranging between 2.44 and 1.42% (Table 3 and Fig. 9). These highvalues may be due to the cultural deposits (i.e., bison remains,charcoal, etc.), but A-horizon material derived from upslope, anddetrital organic matter that accumulated on the floor of the kettlebasin also may have enriched the 2Ak7b horizon.

Organic carbon content declines to 0.89% in the 2ABkb, a tran-sitional horizon between the cumulic A horizon of the LeonardPaleosol and the 3Bk1b horizon developed in the Mallard IslandMember. Mandel (2012) referred to the 2ABkb horizon as a “mixingzone” because bioturbation and other pedogenic processes haveerased an abrupt or clear boundary that may have existed betweenthe Aggie Brown and Mallard Island members. Within the mixingzone, paludal deposits comprising the Aggie Brown Member gradeinto the underlying loess comprising the Mallard Island Member;there is no distinct boundary separating the two different parentmaterials. Also, bison bones associated with the Agate Basin cul-tural component occur in the mixing zone and may have beendisplaced downward from the 2Ak7b horizon through bioturbation,including trampling by bison. There are many biogenic features inthe 2ABkb horizon, including krotovina that are up to 20 cm indiameter.

In Profile 1 the Mallard Island Member is at a depth of107e150þ cm below the surface and consists of pedogenicallymodified loess. The Mallard Island Member has lower organiccarbon content and tends to be sandier compared to the otherstratigraphic units in Profile 1 (Table 3 and Fig. 9). The 3Bk1b and3Bk2b horizons developed in the Mallard Island consist of lightolive brown (2.5YR 5/3e5/4, dry) silt loam with stage I carbonatemorphology. These horizons appear to be genetically related to theLeonard Paleosol, so only one buried soil was identified in Profile 1,but other analytical methods, such as micromorphological analysis,



might be useful in determining whether a truncated paleosol ispresent at the top of the Mallard Island Member.

The soil stratigraphy recorded in Profile 2 is similar to the soilstratigraphy observed in Profile 1. A thin veneer of modern lakebedsediment overlies the Pick City Member, which in turn mantles theAggie Brown Member, and the Mallard Island Member comprisesthe lower 45 cm of the profile (Figs. 10 and 11). The cumulic Ahorizon of the Leonard Paleosol formed in the Aggie BrownMember in Profile 2 consists of five subhorizons (2Ak1be2Ak5b) inProfile 2 (Table 4), compared to seven subhorizons (2Ak1be2Ak7b)in Profile 1.


Depth (cm) Soil horizon Texture Structure Lower boun

Modern lakebed sand0e3 e e e e

Pick City Member3e23 Bk1 SiL 1, f pr~1, f, sbk G, S

23e36 Bk2 SiL 2, m þ f, pr~1, f, sbk C, S

Aggie Brown Member36e48 2Ak1b SiL 1, f, pr~1, f þ m, gr G, S

48e52 2Ak2b SiL 1, f, pr~1, f, gr C, S

52e69 2Ak3b SiL 1, f, pr~1, f, gr G, S

69e78 2Ak4b SiL 1, f, gr C, S

78e99 2Ak5b SiL 1, f, pr~1, f, gr G, S

99e105 2ABkb SiL 1, f, pr~1, f, sbk G, S

Mallard Island Member105e125 3Bk1b SiL 1, m þ f, pr~1, f, sbk G, S

125e150 3Bk2b SiL 1, m, pr~1, f, sbk e

Abbreviations and symbols: Texture SiL ¼ Sit loam; Structure 1 ¼ weak, 2 ¼ moderat~ ¼ parting to; Boundary G ¼ gradual, C ¼ clear, S ¼ smooth.


As was the case with Profile 1, bison bones were found at thebottom of the cumulic A horizon of the Leonard Paleosol (lower5 cm of the 2Ak5b) and in the 2ABkb horizon. Also, a small con-centration of bone fragments was recorded in the upper 8 cm of the3Bk1b horizon developed in the Mallard Island Member, but thefragments appeared to be in a krotovina.

The grain-size and organic-carbon data for Profile 2 areremarkably similar to the data for Profile 1 (Fig. 11 and Table 5). Forexample, the Pick City Member has a fairly homogeneous silt loamtexture, and the Mallard Island Member is sandier compared to theother stratigraphic units. The Aggie Brown Member has the highestproportion of silt in the profile, with fine silt dominating the silt

dary Color and selected features

e

Light brownish gray (10YR 6/2), dark grayish brown(10YR 4/2) moist; common films and threads of CaCO3;common worm casts; common open insect burrows; manyfine and very fine pores.Light brownish gray (2.5Y 6/2), dark grayish brown (2.5Y 4/2)moist, many films and threads of CaCO3; common worm casts;common open insect burrows; many fine and very fine pores.

Gray (10YR 5/1), very dark gray (10YR 3/1) moist; many filmsand threads of CaCO3; many worm casts; many open insectburrows; many fine and very fine pores.Grayish brown (10YR 5/2), very dark grayish brown (10YR 3/2)moist; common films and threads of CaCO3; many worm casts;many open insect burrows; many fine and very fine pores.Dark gray (10YR 4/1), very dark gray (10YR 3/1) moist; commonfilms and threads of CaCO3; many worm casts; many open insectburrows; many fine and very fine pores.Dark grayish brown (10YR 4/2), very dark grayish brown (10YR 3/2)moist; common films and threads of CaCO3; many worm casts;many open insect burrows; few krotovina �20 cm in diameter;many fine and very fine pores.Dark gray (10YR 4/1), very dark gray (10YR 3/1) moist; commonbison bones and flecks of charcoal in the lower 5 cm of horizon;common films and threads of CaCO3; many worm casts; commonopen insect burrows; common krotovina �20 cm in diameter;many fine and very fine pores.Dark grayish brown (10YR 4/2) to brown (10YR 4/3), very darkgrayish brown (10YR 3/2) to dark brown (10YR 3/3) moist;common fine prominent strong brown (7.5YR 4/6) mottles;common bison bones and flecks of charcoal; few films and threadsof CaCO3; many worm casts; many open insect burrows; commonkrotovina �20 cm in diameter; many fine and very fine pores.

Light olive brown (2.5YR 5/3), olive brown (2.5Y 4/3) moist; manyfine distinct yellowish brown (10YR 5/6) and common fine prominentstrong brown (7.5YR 46) mottles; small concentration of bonefragments in the upper 8 cm of horizon (krotovina?); few films andcommon threads of CaCO3; common worm casts; common openinsect burrows; few krotovina �20 cm in diameter; common fineand many very fine pores.Light olive brown (2.5YR 5/4), olive brown (2.5Y 4/4) moist;many fine and medium distinct yellowish brown (10YR 5/6) andfew fine prominent strong brown (7.5YR 46) mottles; commongray (10YR 6/1) and light gray (10YR 7/1) reduction zones alongmacro-pores; few to common fine threads of CaCO3; common wormcasts; common open insect burrows; few krotovina �20 cm indiameter; common fine and many very fine pores.

e, f ¼ fine, m ¼ medium, pr ¼ prismatic, sbk ¼ subangular blocky, gr ¼ granular,



fraction at most depths within this stratigraphic unit. Also, thecumulic A horizon of the Leonard Paleosol (Aggie Brown Member)is enriched with organic carbon, but the organic carbon content ishighly irregular with depth (Fig. 11).


Soil horizons Depth (cm) Sand (%) Silt (%) Clay (%) Textural classb Organic carbon (%)

Total C F Total Total

Bk1 8e18 9.2 43.6 22.4 66.0 24.9 SiL 0.718e23 8.2 42.3 24.6 66.9 25.0 SiL 0.8

Bk2 23e28 9.1 38.9 24.7 63.6 27.3 SiL 0.828e31 12.5 34.1 26.3 60.4 27.2 SiL 0.831e36 13.4 30.3 27.4 57.7 28.8 SiL 0.8

2Ak1b 36e41 6.1 24.3 38.3 62.6 31.3 SiL 1.841e44 5.9 32.5 35.8 68.3 25.9 SiL 2.144e48 7.8 34.6 38.0 72.6 19.6 SiL 1.8

2Ak2b 48e52 7.0 34.5 37.1 71.6 21.5 SiL 1.62Ak3b 52e55 6.9 38.8 39.5 78.3 14.9 SiL 2.2

55e60 8.6 37.7 39.1 76.8 14.6 SiL 2.260e65 7.2 36.8 41.8 78.6 14.3 SiL 2.465e69 9.4 53.1 23.3 76.4 14.2 SiL 2.6

2Ak4b 69e74 4.7 34.2 43.7 77.9 17.4 SiL 2.074e78 7.6 34.0 42.9 76.9 15.6 SiL 2.8

2Ak5b 78e83 6.0 33.8 43.0 76.8 17.2 SiL 2.283e88 8.9 36.1 38.7 74.8 16.3 SiL 2.388e93 9.6 37.9 37.0 74.9 15.6 SiL 2.793e99 11.7 32.6 47.0 79.6 8.6 SiL 2.2

2ABkb 99e105 16.5 37.2 31.9 69.1 14.5 SiL 1.33Bk1b 105e110 20.7 41.6 19.8 61.4 17.8 SiL 0.4

110e115 24.8 39.2 18.3 57.5 17.7 SiL 0.4115e120 27.7 38.0 16.5 54.5 17.7 SiL 0.4120e125 26.3 39.3 17.1 56.4 17.3 SiL 0.4

3Bk2b 125e130 26.9 37.6 18.8 56.4 16.7 SiL 0.4130e135 24.1 41.2 19.6 60.8 15.2 SiL 0.4135e140 22.7 42.9 20.6 63.5 13.8 SiL 0.4140e145 23.3 43.9 19.1 63.0 13.7 SiL 0.3145e150 21.7 45.3 19.9 65.2 13.0 SiL 0.3

a Particle-size limits (mm): Sand (Total) ¼ 2.0e0.05; Silt (Total) ¼ 0.05e0.002, Coarse (C) ¼ 0.05e0.005, Fine (F) ¼ 0.005e 0.002; Clay (Total) ¼ <0.002.b Textural class: SiL ¼ silt loam.

An effort was made to gain a better understanding of the nu-merical age of the sedimentary units at the Beacon Island site.Hence, bulk soil samples were collected from Profile 2 for radio-carbon dating. Soil organic matter from the upper 10 cm of the2Ak1b and 2Ak3b horizons yielded AMS radiocarbon ages of7980 ± 25 B.P. and 8740 ± 25 B.P., respectively (Table 1). The sig-nificance of these radiocarbon ages is discussed later.

5.1.2. Profile 3Profile 3 is located east of Profiles 1 and 2 (Fig. 6), mid-way

between the floor and rim of the kettle basin. All of the


Depth (cm) Soil horizon Texture Structure Lower boundary

Modern lakebed sand0e9 e e e e

Pick City Member9e24 Bk SiL 1, f, pr~1, f, sbk C, S

Aggie Brown Member24e34 2Ak1b SiL 1, f, pr~1, f, gr C, S

34e48 2Ak2b1 SiL 1, f, pr~1, f, gr C, S


stratigraphic units observed in Profiles 1 and 2 occur in Profile 3,plus the Lostwood Drift was exposed in the lower 10 cm of Profile3 (Table 6 and Fig. 12 and 13). However, the members of the OaheFormation, especially the Aggie Brown, are thinner in Profile 3

compared to the other profiles. For example, the Pick CityMember is only 15 cm thick, at least in part because of historicerosion, and the A horizon developed in the Aggie BrownMember (Leonard Paleosol) is only 38 cm thick. The absence ofan overthickened A horizon is due to landscape position.Whereas the deeper part of the kettle basin is a zone of netsediment accumulation (i.e., it is a sediment trap), the shallowerfringes of the basin did not receive much sediment. Also, themargins of the kettle probably experienced erosion during theearly Holocene, thus providing sediment to deeper parts of thebasin.

Color and selected features

e

Grayish brown (10YR 5/2), dark grayish brown (10YR 4/2) moist;few films and common fine threads of CaCO3; common worm casts;many open insect burrows; many fine and very fine pores.

Dark grayish brown (10YR 4/2), very dark grayish brown (10YR 3/2)moist; common films and many fine threads of CaCO3; common wormcasts; many open insect burrows; many fine and very fine pores.Dark gray (10YR 4/1) to dark grayish brown (10YR 4/2), very dark gray(10YR 3/1) to very dark grayish brown (10YR 3/2) moist; common films

(continued on next page)


Table 6 (continued )

Depth (cm) Soil horizon Texture Structure Lower boundary Color and selected features

and many fine threads of CaCO3; common worm casts; many open insectburrows; many fine and very fine.

48e62 2Ak3b1 SiL 1, f, gr G, S Very dark gray (10YR 3/1) to very dark grayish brown (10YR 3/2),black (10YR 2/1) to very dark brown (10YR 2/2) moist; weak finegranular structure; hard, friable; common bison bones in lower5 cm of horizon; few films and common fine threads of CaCO3;many worm casts; many open insect burrows; common krotovina2e5 cm in diameter filled with brown (10YR 5/3) to yellowishbrown (10YR 5/4) SiLt loam; many fine and very fine pores.

62e71 2ABkb1 SiL 1, f, sbk G, S Grayish brown (10YR 5/2) to brown (10YR 5/3, dark grayishbrown (10YR 4/2) to brown (10YR 4/3) moist; few fine prominentyellowish brown (10YR 5/6) and strong brown (7.5YR 4/6) mottles;common bison bones in upper 5 cm of horizon; few fine threads ofCaCO3; common worm casts; common open insect burrows;common krotovina 2e5 cm in diameter filled with brown (10YR5/3) to yellowish brown (10YR 5/4) SiLt loam; many fine and very fine pores.

Mallard Island Member71e88 3Bk1b1 SiL 1, m. pr~1, f sbk Brown (10YR 5/3) to yellowish brown (10YR 5/4), brown (10YR 4/3)

to dark yellowish brown (10YR 4/4) moist; common fine prominentyellowish brown (10YR 5/6) and strong brown (7.5YR 4/6) mottles;common to many encrusted threads of CaCO3; common worm casts;common open insect burrows; many fine and very fine pores; gradualsmooth boundary.

88e100 3Bk2b1 SiL 1, m. pr~1, f, sbk A, S Light olive brown (2.5Y 5/3 to 2.5Y 5/4), olive brown (2.5Y 4/3 to2.5Y 4/4) moist; common fine prominent yellowish brown (10YR 5/6)and strong brown (7.5YR 4/6) and few fine prominent yellowish red(5YR 4/6) mottles; common to few fine threads of CaCO3; few wormcasts; few open insect burrows; many fine and very fine pores.

Lostwood Glacial Drift100e110 4Bkb2 L w, f, sbk e Dark grayish brown (2.5Y 4/2), dark grayish brown (2.5Y 4/2) to very

dark grayish brown (2.5Y 3/2) moist; common fine faint yellowishbrown (10YR 5/4) and few fine distinct yellowish brown (10YR 5/6)and strong brown (7.5YR 4/6) mottles; weak common films andthreads of CaCO3; common pebbles and few cobbles scatteredthrough the matrix; few open insect burrows.

Abbreviations and symbols: Texture SiL¼ silt loam. L¼ loam; Structure 1¼weak, 2¼moderate, f¼ fine, m¼medium, pr¼ prismatic, sbk¼ subangular blocky, gr¼ granular,~ ¼ parting to; Boundary G ¼ gradual, A ¼ abrupt, C ¼ clear, S ¼ smooth.


Because Profile 3 is relatively thin, only a single bulk soil samplewas collected from each soil horizon for physical and chemicalanalyses. Consequently, the resolution of the grain-size andorganic-carbon data is not as high as the data for Profiles 1 and 2.Nevertheless, the general patterns observed in the data for Profiles1 and 2 are detectable in Profile 3 (Fig. 13) and Table 7. For example,organic carbon content dramatically increases down-profile goingfrom the Pick City Member to the Aggie Brown Member, and thereis great variability in the organic carbon content within the cumulicA horizon comprising most of the Aggie Brown Member (LeonardPaleosol). Also, the Mallard Island Member is sandier compared tothe other members of the Oahe Formation, though it is not as sandyas the Lostwood Drift.


Soil horizons Depth (cm) Sand (%) Silt (%)

Total C F

Bk 9e24 17.3 46.3 24.82Ak1b1 24e34 13.6 47.2 34.02Ak2b1 34e48 9.6 38.3 32.92Ak3b1 48e62 13.9 36.1 32.72ABkb1 62e71 19.3 39.5 24.63Bk1b1 71e88 22.9 39.1 23.03Bk2b1 88e100 19.9 45.1 23.54Bkb2 100e110 40.4 16.2 21.5

a Particle-size limits (mm): Sand (Total) ¼ 2.0e0.05; Silt (Total) ¼ 0.05e0.002, Coarseb Textural classes: SiL ¼ silt loam; L ¼ loam.


A dense concentration of bison bones occurs in the lower 5 cm ofthe 2Ak3b1horizon and in the upper 5 cm of the 2ABkb1 horizon(mixing zone). Immediately north of Profile 3, a few bones alsowere recorded in the upper 5e10 cm of the Mallard Island Member.The bones in the Mallard Island Member probably were displaceddownward, perhaps by trampling or other forms of bioturbation.

5.2. Paleoenvironmental reconstruction

5.2.1. Stable carbon isotope ratiosThe d13C values determined on soil organic matter at the Beacon

Island site range from �25.4 to �22.6‰, indicating a strong C3signature for the entire period of record (Figs. 9, 11 and 13). The

Clay (%) Textural classb Organic carbon (%)

Total Total

71.1 11.6 SiL 0.5981.2 5.2 SiL 1.1971.2 19.1 SiL 1.3968.8 17.3 SiL 1.7764.1 16.6 SiL 0.6862.0 15.0 SiL 0.5168.6 11.5 SiL 0.4137.7 21.9 L 0.87

(C) ¼ 0.05e0.005, Fine (F) ¼ 0.005e0.002; Clay (Total) ¼ <0.002.



lowest value, �25.4‰, occurs in the mixing zone between theMallard Island and Aggie Brown members, which is where theAgate Basin cultural component is located in the stratigraphicsequence. The highest value, �22.6‰, occurs in the Pick CityMember. The difference between themaximum andminimum d13Cvalue is �2.8‰.

Two distinct d13C excursions occur in Profiles 1 and 2: one at theboundary between the Mallard Island and Aggie Brown members,and the other at the boundary between the Aggie Brown and PickCity members (Figs. 9 and 11). In Profile 1, the d13C values deter-mined on samples from the Mallard Island Member remain fairlyconsistent, ranging from�24.1 to �24.0‰. However, the d13C valueabruptly shifts from �23.7‰ at the top of the Mallard IslandMember to �24.8‰ in the mixing zone at the bottom of the AggieBrown Member. Similarly, in Profile 2 there is little fluctuationwithin the Mallard Island Member, with d13C values rangingbetween �24.5 and �24.1‰, but a distinct shift from �24.2to �25.4‰ occurs across the boundary between the Mallard IslandMember and the mixing zone at the bottom of the Aggie BrownMember.

In the three profiles there is a general trend toward higher d13Cvalues up through the Aggie Brown Member until the top 40 cm ofthat stratigraphic unit is reached, where a slight shift to lower d13Cvalues occurs. Based on radiocarbon ages determined on soil

Fig. 10. Profile 2 in Area A of the Beacon Island site showing three members of theOahe Formation and the stratigraphic position of the Agate Basin Cultural component.


organic matter in Profile 2, this shift began around 8700 B.P. andended soon before ca. 8000 B.P. (Fig. 11). The subsequent shift tohigher d13C values that begins in the upper 20 cm of the AggieBrown Member and continues in the Pick City Member mayrepresent a slight warming event.

Identifying the point at which changes in the d13C values reflectactual changes in vegetation composition (C3 vs. C4) is difficult (Cyret al., 2011). According to Krull and Skjemstad (2003), changesbetween 1 and 3‰ are related to inherent soil processes, whereasdifferences exceeding 3‰ result from changes in the contributionof C3 and C4 vegetation. Ehleringer et al. (2000), however, notedthat changes as small as 1‰ may be caused by environmentalstress. A 1 to 3‰ shift in the d13C values may reflect increasedfractionation against 12C by C3 plants due to changes in respirationrates during drought, but may also represent small increases in C4plants within a C3-dominated community.

With only slight increases in d13C values, the stable carbonisotope record at Beacon Island indicates that C4 grasses were mi-nor components of the plant community and that the quantity of C4grasses fluctuated slightly during the terminal Pleistocene andearly- to middle-Holocene. Given that the lowest d13C values weredetermined on soil organic matter from the Agate Basin componentat the bottom of the Aggie Brown Member, it is likely that thecoolest and perhaps wettest period at the site occurred at ca.10,300 B.P. Such climatic conditions would have favored the accu-mulation of water in the kettle basin at that time, which in turnwould have attracted both bison and people. The distinct shift to-wards higher d13C values in the upper 20 cm of the Aggie BrownMember and continuing into the Pick City Member, and theaccompanying shift from paludal sedimentation (Aggie BrownMember) to loess deposition (Pick City Member), point to awarming and drying trend that began around 8000 B.P. andcontinued into the middle Holocene. Although the relatively smallchanges in the carbon isotope composition through the profile maybe related more to the differences in soil formation between themembers (i.e., eroded B horizons versus cumulic A horizons), thisinterpretation is supported by the phytolith data, which are pre-sented in the following section.

5.2.2. Phytoliths and microscopic particulate charcoalResults of opal phytolith analysis from Profile 2 are reported in

Fig. 14. C3 grasses of the Pooideae family dominated the phytolithassemblage. Short-cell phytolith morphologies were similar to thereference collection morphologies of Western wheatgrass (Agro-pyron smithii) and Canada wildrye (Elymus Canadensis), which aretypical C3 pooids, and to Green needlegrass (Stipa viridula), Porcu-pine grass (Stipa spartea), andwavy-sided (sinuate) Stipa, which aredrought-resistant pooids. Non-diagnostic grass long cells and par-ticulate charcoal were abundant, particularly in the Aggie BrownMember samples; no other siliceous morphotypes typical of treesor shrubs were present in any sample. One sponge spicule wascounted in the deepest Mallard Island Member sample, where noother siliceous material was preserved.

To detect aridity, the data are differentiated between the per-centages of general Pooid-type short cells, and the more drought-resistant Stipa-type short cells (Cordova, 2011). A significant shiftfrom pooids to drought-resistant species occurred across theboundary between the Agate Basin cultural horizon and overlyingAk5b horizon (Fig. 14). The shift in assemblage is from 96% pooidsand 4% drought-resistant Stipa-types in the Agate Basin horizon, to44% pooids and 56% drought-resistant Stipa-types in the Ak5bhorizon. Phytolith assemblages from the top and bottom of the2Ak1b and from the 2Ak4b horizons are typical of a stable plantcommunity, with each sample consisting of approximately 60%pooids and 40% drought-resistant Stipa-types. A small number of




phytoliths representing C4 grasses from the Chloridoideae familyappear in the two samples from the 2Ak1b horizon at the transitionfrom the Aggie Brown Member to the Pick City Member, indicatingslight warming and/or drying.

The concentration of microscopic particulate charcoal is highestin the 2Ak4b horizon in the Aggie Brown Member, with 124,000particles per gram of pre-treated soil (Fig. 14). Charcoal concen-trations in the 2Ak1b horizonwere 31,000 and 29,000 particles pergram of pre-treated soil. The 2ABkb horizon (mixing zone) had thelowest charcoal concentrations, with 12,000 and 15,000 particlesper gram of pre-treated soil. Charcoal was not observed in the PickCity or Mallard Island Members.

Burned (opaque) grass phytoliths were rare, and none wererecorded in the cultural horizon. The highest percentage of burnedphytoliths (8%) occurred with the highest charcoal concentration inthe 2Ak4b horizon, with only 2% burned phytoliths in the under-lying 2Ak5b horizon.

Soil samples from the Bk2 horizon in the Pick City Member andthe 3Bk1b and 3Bk2b horizons in the Mallard Island Member failedto yield phytoliths. The absence of phytoliths in some of thesesamples may be related to soil pH. The pH values in Profile 2 rangefrom 8.2 to 8.8, with the highest value, 8.8, in the Bk2 horizon (PickCity Member) (Fig. 14). High alkalinity may have caused silicacomprising the phytoliths to go into solution, which may explainthe absence of phytoliths in the Bk2 horizon. However, the soilsample from the 2Ak1b horizon (Aggie Brown Member) had goodphytolith preservation and high phytolith concentrations despite apH of 8.7, while soil samples from the 3Bk1b and 3Bk2b horizons,which lacked phytoliths, had relatively low pH values (8.3 and 8.4,


respectively). Hence, it is likely that factors other than pH, such asweathering or soil texture, have affected phytolith preservation atthe Beacon Island site. As the Pick City and Mallard Island samplesare from B horizons, phytoliths were probably already low in fre-quency compared to the A horizons of the Aggie Brown. However,considerably more solution pitting on the short and long grass cellsin the ABkb horizon at the contact between the Aggie Brown andMallard Island Members suggests soil pH played a role in at leastsome phytolith weathering.

6. Discussion

Based on the stratigraphic evidence and radiocarbon ages, thekettle basin at the Beacon Island site formed before ca. 10,300 B.P.,presumably around 16,000e15,000 B.P. when the terminal ice frontwas in the region. Initially, the kettle basin trapped eolian sedi-ment, the slightly sandy loess comprising the Mallard IslandMember. Conditions probably were relatively dry when theMallardIsland Member aggraded, which accounts for the presence of loessand the absence of paludal deposits immediately above the late-Wisconsinan glacial till that floors the kettle basin (Fig. 5). How-ever, shortly before 10,300 B.P., paludal deposits comprising theAggie Brown Member began to accumulate in the kettle basin,suggesting a shift to wetter climatic conditions. Ponds in smallkettle basins typically are marsh-like environments where thewater is shallow and grasses, rushes, reeds, typha, sedges, and otherherbaceous plants dominate the vegetation community.

According to Picha (2012), the gastropod assemblage at BeaconIsland reflects “an intermittent or semi-permanent, pond edge



habitat” during Agate Basin times and during aggradation of theAggie Brown Member. The gastropods are almost exclusivelyaquatic taxa, with only one terrestrial taxon, Discus cronkhitei, amesic environment indicator, represented in the assemblage.

The micromammal remains from the Agate Basin componentreflect the presence of a marsh or bog-like environment. Accordingto Falk and Semken (2012), the “Dark Faunule” micromammalassemblage recovered from the Aggie Brown Member in Area A, 98percent of which by specimen count are voles, indicates cool, mesicconditions with a high water table and a grassy substrated either ameadow or a bog. Two taxa in the Dark Faunule, the red-backedvole (Myodes sp.) and Ungava vole (Phenacomys ungava), alsorequire low shrubby vegetation; dwarf birch is preferred by bothtaxa. However, it is unlikely that shrubby vegetation was extensive.Otherwise one would expect the deer mouse (Peromyscus sp.) to berepresented by more than a single specimen in the faunal assem-blage (Falk and Semken, 2012).

The remains of two large rodents, the American beaver (Castorcanadensis) and common muskrat (Ondatra zibethicus), in thefaunal assemblage from the Agate Basin component indicate thepresence of standing water at or near the Beacon Island site (Falkand Semken, 2012). Beaver and muskrat are strongly linked toaquatic habitats and require permanent water (Jones et al., 1983).

Fig. 12. Profile 3 in Area A of the Beacon Island site showing three members of theOahe Formation, the Lostwood Drift, and the stratigraphic position of the Agate Basincultural component.


It seems more than a coincidence that the presence of bison andhumans at Beacon Island at about 10,300 B.P. coincides with theshift to mesic conditions. At about that time, sediment was deliv-ered to the kettle basin by sheetwash and airfall, but the contri-bution of loess to the Aggie Brown Member probably wasinsignificant. In low relief landscapes like the one at the BeaconIsland site, loess deposition forms a blanket of sediment that isuniformly thick across topographic highs and lows over a smallarea. This is one of the most distinct signatures of loess (see Masonet al., 2008). However, at Beacon Island, the Aggie BrownMember isabout 70 cm thick in the area of Profiles 1 and 2, but only 47 cmthick in the area of Profile 3, and even thinner along the rim of thekettle basin. The variation in the thickness of the Aggie BrownMember over such a small area is attributed to sheet erosion on therim and near-rim portion of the kettle basin and the accumulationof the eroded loess first as sheetwash, then as paludal sediment inthe bowl of the kettle basin. Ruhe (1969:147) documented similarspatial patterns of erosion and sedimentation at kettles on the DesMoines Lobe in north-central Iowa. At Beacon Island, data on thedistributions of faunal remains and chipped-stone artifacts indicatethat post-occupation sheetwash sediment quickly buried the AgateBasin bonebed, thereby isolating it from subsequent disturbanceprocesses.

Paludal deposits continued to aggrade at Beacon Island after10,300 B.P., resulting in deep burial of the Agate Basin culturaldeposits. However, sedimentation was at a relatively slow rate,allowing soil development to keep up with deposition. Thiscumulization process resulted in the formation of an overthickenedA horizon typical of the Leonard Paleosol and other Younger Dryas-age paleosols across much of the Great Plains (Mandel, 2008). Thecumulic A horizon of the Leonard Paleosol at Beacon Island is multi-storied, indicating that episodes of gradual sedimentation wereinterspersed with brief periods of landscape stability. This patternprobably reflects cycles of water accumulation (episodes of pondsedimentation) and drying (periods of landscape stability and soilformation) in the kettle basin.

Based on the radiocarbon ages determined on soil organicmatter from the 2Ak3b and 2Ak1b horizons of the Leonard Paleosol,up-building of the Aggie Brown Member slowed around 8700 B.P.and ceased by ca. 8000 B.P. The accumulation of the loess thatcomprises the Pick City Member was underway soon after 8000 B.P.and probably marks the beginning of the Altithermal climaticepisode in northwestern North Dakota. This interpretation is sup-ported by the stable carbon isotope and phytolith data.

The stable carbon isotope data reveal a slight shift to coolerconditions at ca. 10,300 B.P., which corresponds with the period ofAgate Basin occupation at the site. Also, the lowest concentrationsof microscopic particulate charcoal occur in the lower Aggie BrownMember (2Ak5b and 2ABkb), suggesting relatively low regional firefrequency, indicative of a cool, moist environment during the AgateBasin occupation. Typically, a high ratio of burned to unburnedphytoliths is suggestive of in situ burning (Piperno, 2006). Both theabsence of burned phytoliths and low concentrations of charcoalwithin the Agate Basin cultural component suggest anthropogenicactivity at the site did not include isolated fires or grasslandburning. Large-scale controlled grassland burning by NativeAmerican groups in the Northern Plains only seems to haveoccurred since 5000 B.P. (Boyd, 2002).

The results of phytolith and stable carbon isotope analysessupport Ahler's (2003) hypothesis that during the time of AgateBasin occupation at the Beacon Island site (ca. 10,300 B.P.) the plantcommunity probably was a mesic C3 grassland. However, it is likelythat people who made the kill at Beacon Island did not experienceclimatic conditions drastically different from modern conditions.Modern analogs of kettles and associated grass assemblages occur




in northwestern North Dakota. Our findings also support previouspaleoenvironmental studies indicating that a cool-season C3-dominated prairie existed in the Northern Plains during the Pleis-toceneeHolocene transition. Soon after ca. 10,300 B.P., gradual

Fig. 14. Diagram of Profile 2 in Area A of the Beacon Island site showing the phytolith a


warming and/or drying occurred at Beacon Island, as indicated byan increase in drought-resistant Stipa-type short cells immediatelyabove the Agate Basin component. During the latter part of theearly Holocene it is likely that Beacon Island resembled the

ssemblage, charcoal and phytolith concentrations, soil pH, and the soil stratigraphy.



seasonally dry playas of the High Plains of Texas (see Holliday et al.,1996, 2008). By ca. 8000 B.P., warm season C4 chloridoids appear atthe site and there is a trend towards higher d13C values from thebottom to the top of the Pick City Member. Hence, the climatebecame progressively warmer and probably drier with a decliningwater table and/or decreasing effective precipitation from ca. 8000B.P. through the middle Holocene.

Based on the combined isotope, phytolith, and charcoal data,gradual warming and drying occurred after ca. 10,300 B.P., thoughthe isotope data suggest that it was interrupted by a slight coolingepisode that began around 8700 B.P. and ended soon before ca.8000 B.P. Between 10,300 and 8000 B.P., is likely that the micro-scopic particulate charcoal was blown in fromnatural regional fires.Particulate charcoal from high plumes of smoke can be transportedby wind for long distances, settling into depressions such as kettlebasins, and can accumulate during long periods of soil formation(Boyd, 2002; Piperno, 2006). The increase in particulate charcoalafter the Agate Basin occupation suggests regional climate wasgetting drier, with an increase in fire frequency, followed by aleveling off, perhaps as woody fuel began to decrease. Despite thegeneral warming and drying trend, a relatively stable plant com-munity of C3 grasses existed at the site from the terminal Pleisto-cene through the middle Holocene. Also, no major turnover in thegrassland photosynthetic pathway was detected as might be ex-pected for the onset of the warm/dry middle Holocene Altithermal.

7. Summary and conclusions

Although a numerical chronology has not been established forthe entire sedimentary sequence at the Beacon Island site, four offive radiocarbon ages determined on materials from the bonebedare statistically equivalent, making it the most precisely datedAgate Basin component in the Americas. The suite of radiocarbonages for the site provides a numerical chronology for terminalPleistocene and early Holocene landscape evolution and the pale-oenvironmental record at Beacon Island.

During the period of Agate Basin occupation at the site (ca.10,300 B.P.), the landscape around the confluence of the Little Knifeand Missouri rivers featured a series of shallow, intermittent wet-lands in small kettle basins, or potholes, in theWisconsinan till. TheAgate Basin bonebed occurs in one such basin, and the presence of ahearth on the floor of the basin indicates that at least a portion ofthe depression was dry during the occupation.

The Agate Basin occupation occurred early in a period of relativelandscape stability marked by the Leonard Paleosol that formed inthe slowly aggrading Aggie Brown Member. Ahler et al. (2012)suggested that the Aggie Brown Member began to aggrade atBeacon Island nomore than 100 years before the mean radiocarbonage of the Agate Basin component, or between ca. 10,450 and 10,350 B.P. This provides an important timeline for the beginning of aclimatic event reflected by the formation of the Leonard Paleosol. AtBeacon Island, the inception of the Leonard Paleosol coincided witha shift from eolian deposition, represented by the slightly sandyloess of the Mallard Island Member of the Oahe Formation, to theprimarily paludal deposits of the Aggie Brown Member, indicativeof a wetter climate. Paludal deposition typically occurs in marshyenvironments, where water is shallow and grasses, rushes, andsedges dominate the vegetation community. Molluscan data indi-cate that the basin containing the bonebed intermittently heldwater following the occupation. The habitat requirements of theidentified gastropod taxa, which consist mostly of aquatic forms,point to the presence of locally dense, pond-edge vegetation. Also,the micromammal faunal assemblage is indicative of a marshy or


bog-like environment during and immediately after Agate Basintimes, and the presence of beaver and muskrat bones in the exca-vated assemblage indicates the presence of standing water nearby.

Outside the kettle basin, the local plant community at the timeof the occupationwas a mesic cool-season (C3) grassland.While theenvironment likely did not differ much from modern conditions,stable carbon isotope and phytolith data indicate that the AgateBasin occupation coincided with the coolest and perhaps thewettest climatic episode recorded in the sampled deposits. Data onlikely micromammal habitat suggest that this boreal grassland mayhave been punctuated by stands of shrubby vegetation, perhapsincluding dwarf birch.

Following the Agate Basin occupation, paludal depositscontinued to aggrade in the kettle basin, resulting in deep burial ofthe bonebed. However, gradual sedimentation allowed soil devel-opment to keep up with deposition, resulting in the formation of anoverthickened cumulic A horizon that comprises most of the Leo-nard Paleosol. The multi-storied A horizon of the Leonard Paleosolat Beacon Island formed as a result of brief episodes of slow sedi-mentation punctuated by periods of landscape stability, a patternreflecting alternating wet and dry climatic cycles, respectively.Overall, however, at Beacon Island the Leonard Paleosol is a productof mesic conditions during the Younger Dryas and first twomillennia of the early Holocene.

Aggradation of the Aggie Brown Member slowed around8700 B.P. and ceased around 8000 B.P. Loess comprising the PickCity Member of the Oahe Formation began accumulating soon afterthis time, likely marking the initiation of the Altithermal climaticepisode. Stable carbon isotope and phytolith data also point to awarmer and probably drier climate after 8000 B.P. indicated by theappearance of warm-season C4 chloridoids at the site and byheavier d13C values determined on soil organic matter.

In addition to reconstructing the history of landscape evolutionand environmental change at the Beacon Island site, our studyyielded information that is relevant to any investigation that usesphytoliths and microscopic charcoal as proxies for reconstructingpaleoenvironments. Specifically, our findings demonstrate thatmany complex factors control phytolith and charcoal preservation,including soil development, texture, and pH. Also gleaned from thestudy is the importance of separating cool-season grasses fromcool-season drought resistant Stipa species in phytolith assem-blages, a tool that has been largely ignored in the literature.Drought-resistant grasses provide a more precise indication ofincreasing aridity, and recognition of drought-resistant grass spe-cies in the phytolith record can help detect regional environmentalchanges. However, micro-environments, such as the area around akettle pond, can have highly variable vegetation communities, andphytolith assemblages may be drastically different only metersapart. Thus, at Beacon Island, multi-proxy analyses provide a meansfor comparing and more fully interpreting the paleoenvironmentalrecord.

The association of the Agate Basin bonebed with pond depositsunderscores the significance of kettle ponds, even shallow ones, aslocations for human activities through time. As focal points forwater, animals, and plant resources, kettles were attractive to hu-man groups in the Northern Plains. At Beacon Island, the age of thebison bonebed indicates that the kettle was a kill site about 12,000years ago, and the Leonard Paleosol, which contains the bonebed, isa prominent target for locating buried Paleoindian cultural depositsin other kettle basins in the region. Older sedimentary deposits arepresent within and adjacent to the kettle basin at Beacon Island,and they may contain earlier archaeological evidence. Futureresearch should focus on those deposits.



Acknowledgments

Stanley A. Ahler, Paleocultural Research Group's founder andResearch Director from 1996 to 2007, and Fern E. Swenson, Directorof the State Historical Society of North Dakota's Archaeology andHistoric Preservation Division, first recognized Beacon Island'ssignificance and worked tirelessly over many years to design andfund research at the site. The State Historical Society of NorthDakota provided initial funding for the project, and the U.S. ArmyCorps of Engineers funded portions of the field investigation.Additional fieldwork, along with subsequent laboratory analyses,were funded by a Save America's Treasures grant awarded to theState Historical Society of North Dakota by the National Park Ser-vice. Steven Bozarth, University of Kansas, aided in phytolith pro-cessing and identification. The Kansas Geological Survey providedadditional support. The authors thank Kathleen Nicoll and twoanonymous reviewers for their helpful comments.

References

Research Contribution No. 54. In: Ahler, S.A. (Ed.), Resurvey and Test Excavations atBeacon Island in Lake Sakakawea, Mountrail County, North Dakota. Paleocul-tural Research Group, Flagstaff, Arizona (Submitted to the State Historical So-ciety of North Dakota, Bismarck).

Ahler, S.A., Crawford, G.T., 2003. Area description, mapping, and surface studies.Research Contribution No. 54. In: Ahler, S.A. (Ed.), Resurvey and Test Excava-tions at Beacon Island in Lake Sakakawea, Mountrail County, North Dakota.Paleocultural Research Group, Flagstaff, Arizona, pp. 47e54 (Submitted to theState Historical Society of North Dakota, Bismarck).

Ahler, S.A., Feiler, E., Badorek, C., Smail, M., 2002. Stone tools and flaking debris.Research Contribution No. 40. In: Ahler, S.A. (Ed.), Prehistory on First Street NE:the Archaeology of Scattered Village in Mandan, North Dakota. PaleoculturalResearch Group, Flagstaff, Arizona, pp. 12.11e12.81 (Submitted to the City ofMandan, North Dakota, and the North Dakota Department of Transportation,Bismarck).

Ahler, S.A., Stafford, T.W., Haas, H., Zalucha, A., 2012. Appendix a: radiocarbondating. Research Contribution No. 86. In: Mitchell, M.D. (Ed.), Agate BasinArchaeology at Beacon Island. Prepared by the PaleoCultural Research Group,Arvada, Colorado, pp. 239e241 (Submitted to the State Historical Society ofNorth Dakota, Bismarck and the U.S. Department of the Interior, National ParkService, Washington D.C).

Ahler, S.A., Swenson, F.E., Sellet, F., Spurr, K., Madden, S., Mandel, R.D., Hurst, S.,2006a. Final Report on Pre-field Planning and Fieldwork under the SaveAmerica's Treasures Grant Program for Area A at the Beacon Island Site(32MN234), North Dakota. Research Contribution No. 76. PaleoculturalResearch Group, Flagstaff, Arizona (Submitted to the State Historical Society ofNorth Dakota, Bismarck).

Ahler, S.A., Swenson, F.E., Spurr, K., Mandel, R.D., Madden, S., Sellet, F., Hurst, S.,2006b. Field Report Regarding Excavations for Archaeological Mitigation duringJuly and August 2006 at the Agate Basin Component in Area A within theBeacon Island Site, 32MN234, Lake Sakakawea, North Dakota. ResearchContribution No. 75. Paleocultural Research Group, Flagstaff, Arizona (Submit-ted to the U. S. Army Corps of Engineers, Omaha District, Omaha, Nebraska).

Alley, R.B., 2007. Wally was right: predictive ability of the North Atlantic 'ConveyerBelt' hypothesis or abrupt climate change. Annual Review of Earth and Plane-tary Sciences 35, 241e272.

Alley, R.B., Meese, D.A., Shuman, C.A., Gow, A.J., Taylor, K.C., Grootes, P.M.,White, J.W.C., Ram, M., Waddington, E.D., Mayewski, P.A., Zielinski, G.A., 1993.Abrupt increase in Greenland snow accumulation at the end of the YoungerDryas event. Nature 362, 527e529.

Artz, J.A., 1995. Geological contexts of the early and middle Holocene archaeologicalrecord in North Dakota and adjoining areas of the Northern Plains. In:Bettis III, E.A. (Ed.), Archaeological Geology of the Archaic Period in NorthAmerica, pp. 67e86. GSA Special Paper 297, Boulder, Colorado.

Artz, J.A., 2000. Archaeology and earth sciences on the Northern Plains. In:Mandel, R.D. (Ed.), Geoarchaeology in the Great Plains. University of OklahomaPress, Norman, pp. 250e285.

Barnosky, C.W., Grimm, E.C., Wright Jr., H.E., 1987. Towards a postglacial history ofthe northern Great Plains: a review of the paleoecological problems. CarnegieMuseum Annals 56, 259e273.

Beaudoin, A.B., 1993. A compendium and evaluation of postglacial pollen records inAlberta. Canadian Journal of Archaeology 17, 92e112.

Bickley Jr., W.B., 1972. Stratigraphy and History of the Sakakawea Sequence, South-Central North Dakota (Unpublished Ph.D. dissertation). University of NorthDakota, Grand Forks, p. 366.

Bickley Jr., W.B., Clayton, L., 1971. Sedimentation in small ponds (sloughs) in themid-continent during late Quaternary time. North Dakota Academy of ScienceProceedings 25, 36e42.


Bickley Jr., W.B., Clayton, L., Cvancara, A.M., 1970. Seibold site: comparison withother late Quaternary fossil sites in North Dakota. North Dakota Academy ofScience Proceedings 24, 73e79.

Birkeland, P.W., 1999. Soils and Geomorphology, third ed. Oxford University Press,Oxford.

Bluemle, J.P., 2000. Face of North Dakota, third ed.. In: Education Series 26 NorthDakota Geological Survey, Bismarck.

Borresen, J.A., 2002. A Faunal Analysis of the Frazier Site, an Agate Basin-age BisonKill-butchery Site in Northeastern Colorado (Unpublished Master's thesis).University of Tennessee, Knoxville, p. 119.

Boyd, M., 2002. Identification of anthropogenic burning in the paleoecological re-cord of the northern prairies: a new approach. Annals of the Association ofAmerican Geographers 92, 471e487.

Boyd, M., 2005. Phytoliths as paleoenvironmental indicators in a dune field on thenorthern Great Plains. Journal of Arid Environments 61, 357e375.

Bozarth, S.R., 1992. Classification of opal phytoliths formed in selected Dicotyledonsnative to the Great Plains. In: Rapp Jr., G., Mulholland, S.C. (Eds.), PhytolithSystematics: Emerging Issues. Plenum Press, New York.

Broecker, W.S., Kennett, J.P., Flowers, B.P., Teller, J.T., Trumbore, S., Bonani, G.,Wolfli, W., 1989. Routing of meltwater from the Laurentide ice sheet during theYounger Dryas cold episode. Nature 341, 318e321.

Bronk Ramsey, C., 2010. OxCal Program, v. 4.1.7, Radiocarbon Accelerator Unit.University of Oxford, UK. http://c14.arch.ox.ac.uk/embed.php?File¼oxcal.html(last accessed August 2010).

Byers, D.A., 2009. The Agate Basin faunal remains at locality II. In: Larson, M.L.,Kornfeld, M., Frison, G.C. (Eds.), Hell Gap: a Stratified Paleoindian Campsite atthe Edge of the Rockies. The University of Utah Press, Salt Lake City,pp. 135e156.

Clayton, L., 1972. Geology of Mountrail County, North Dakota, Bulletin 55, Part IV.North Dakota Geological Survey, Grand Forks.

Clayton, L., Moran, S.R., 1982. Chronology of late-Wisconsinan glaciation in middleNorth America. Quaternary Science Reviews 1, 55e82.

Clayton, L., Moran, S.R., 1979. Appendix B: Oahe formation. Report of InvestigationsNo. 64. In: Groenewold, G.H., Hemish, L.A., Cherry, J.A., Rehm, B.W., Meyer, G.N.,Winczewski, L.M. (Eds.), Geology and Geohydrology of the Knife River Basin andAdjacent Areas of West-Central North Dakota. North Dakota Geological Survey,Bismarck, pp. 337e339.

Clayton, L., Moran, S.R., Bickley Jr., W.B., 1976. Stratigraphy, Origin, and ClimateImplications of Late Quaternary Upland Silt in North Dakota. MiscellaneousSeries No. 54. North Dakota Geological Survey, Grand Forks.

Clayton, L., Moran, S.R., Bluemle, J.P., 1980. Explanatory Text to Accompany theGeologic Map of North Dakota. Report of Investigations No. 69. North DakotaGeological Survey, Grand Forks.

Coogan, A.H., 1983. Geology and geological history of the site locality. ContributionNo. 193. In: Toom, D.L., Gregg, M.L. (Eds.), The Archaeology of the Mondrian TreeSite (32MZ58), McKenzie County, North Dakota. Department of Anthropologyand Archeology, University of North Dakota, Grand Forks, pp. 7.1e7.49.

Coogan, A.H., 1987. Holocene Geomorphic and Stratigraphic Framework ofArchaeological Sites Along the Missouri River, Central South Dakota. Report ofInvestigations No. DACW45-78-C-0136. U. S. Army Corps of Engineers, Omaha,Nebraska.

Cordova, C., 2011. The Stipa-type short cell: what does it mean taxonomically,climatically and ecologically?. In: 8th International Meeting on PhytolithResearch. Estes Park, Colorado.

Cordova, C., Johnson, W.C., Mandel, R.D., Palmer, M.W., 2011. Late Quaternary envi-ronmental change inferred from phytoliths and other soil-related proxies: casestudies from the central and southern Great Plains, USA. Catena 85, 87e108.

Cvancara, A.M., Clayton, L., Bickley Jr., W.B., Jacob, A.F., Ashworth, A.C., Brophy, J.A.,Shay, C.T., Delorme, L.D., lammers, G.E., 1971. Paleolimnology of late Quaternarydeposits: Seibold site, North Dakota. Science 171, 172e174.

Cyr, H., McNamee, C., Amundson, L., Freeman, A., 2011. Reconstructing landscapeand vegetation through multiple proxy indicators: a geoarchaeological exami-nation of the St. Louis site, Saskatchewan, Canada. Geoarchaeology: An Inter-national Journal 26, 165e188.

Ehleringer, J.R., Buchmann, N., Flanagan, L.B., 2000. Carbon isotope ratios in belowground carbon cycle processes. Ecological Applications 10, 412e422.

Falk, C.R., Semken Jr., H.A., 2012. Paleoenvironmental context: the local fauna re-cord. Research Contribution No. 86. In: Mitchell, M.D. (Ed.), Agate BasinArchaeology at Beacon Island. Prepared by the PaleoCultural Research Group,Arvada, Colorado, pp. 171e181 (Submitted to the State Historical Society ofNorth Dakota, Bismarck and the U.S. Department of the Interior, National ParkService, Washington D.C).

Fedje, D.W., White, J.M., Wilson, M.C., Nelson, D.E., Vogel, J.S., Southon, J.R., 1999.Vermillion lakes site: adaptations and environments in the Canadian rockiesduring the latest Pleistocene and early Holocene. American Antiquity 60,81e108.

Fenneman, N.M., 1931. Physiography of Western United States. McGraw Hill, NewYork.

Flint, R.F., 1955. Pleistocene Geology of Eastern South Dakota. U.S. Geological SurveyProfessional Paper No. 262. U.S. Government Printing Office, Washington, D.C.

Fredlund, G.G., Tieszen, L.L., 1994. Modern phytolith assemblages from the NorthAmerican Great Plains. Journal of Biogeography 21, 321e335.

Fredlund, G.G., Tieszen, L.L., 1997. Phytolith and carbon isotope evidence for lateQuaternary vegetation and climate change in the southern Black Hills, SouthDakota. Quaternary Research 47, 206e217.


http://refhub.elsevier.com/S1040-6182(14)00481-9/sref1





























































































http://c14.arch.ox.ac.uk/embed.php?File=oxcal.html

http://c14.arch.ox.ac.uk/embed.php?File=oxcal.html











































































Frison, G.C., 1984. The Carter/Kerr-McGee paleoindian site: cultural resource man-agement and archaeological research. American Antiquity 49, 288e314.

Frison, G.C., Stanford, D.J., 1982. The Agate Basin Site: a Record of the PaleoindianOccupation of the Northwestern High Plains. Academic Press, New York.

Gee, G.W., Bauder, J.W., 1986. Particle-size analysis. Agronomy Monograph No. 9. In:Klute, A. (Ed.), Methods of Soil Analysis, Part I: Physical and MineralogicalMethods, second ed. American Society of Agronomy/Soil Science Society ofAmerica, Madison, Wisconsin, pp. 383e411.

Haberman, T.W., Schneider, F., 1975. 1974 Archaeological Surveys of Portions of theGarrison Reservoir Shoreline, North Dakota (Submitted to the U. S. Departmentof the Interior, National Park Service, Lincoln, Nebraska). University of NorthDakota, Grand Forks.

Haynes Jr., V.C., Agogino, G.A., 1966. Prehistoric springs and geochronology of theClovis Site, New Mexico. American Antiquity 31, 812e821.

High Plains Regional Climate Center, 2011. Period of Record Monthly ClimateSummary for New Town 4W, North Dakota. University of Nebraska, Lincoln.http://www.hprcc.unl.edu/cgi-bin/cli_perl_lib/cliMAIN.pl?nd6383 (accessed31.05.11.).

Hill, M.G., 2008. Paleoindian Subsistence Dynamics on the Northwestern GreatPlains: Zooarchaeology of the Agate Basin and Clary Ranch Sites. Archaeopress,Oxford. British Archaeological Reports International Series No. 1756.

Holliday, V.T., Gustavson, T.C., Hovorka, S.D., 1996. Stratigraphy and geochronologyof playa fills on the Southern High Plains. Geological Society of America Bulletin108, 953e965.

Holliday, V.T., Mayer, J.H., Fredlund, G.G., 2008. Late Quaternary sedimentology andgeochronology of small playas on the Southern High Plains, Texas and NewMexico, U.S.A. Quaternary Research 70, 11e25.

Holliday, V.T., 2004. Soils in Archaeological Research. Oxford University Press,Oxford.

Jones Jr., J.K., Armstrong, D.M., Hoffmann, R.S., Jones, C., 1983. Mammals of theNorthern Great Plains. University of Nebraska Press, Lincoln, Nebraska.

Kealhofer, L., 1996. The human environment during the terminal Pleistocene andHolocene in northeastern Thailand: phytolith evidence from Lake Kumphawapi.Asian Perspectives 35, 229e254.

Knudson, R., Johnson, E., Holliday, V.T., 1998. The 10,000-year-old Lubbock artifactassemblage. Plains Anthropologist 43, 239e256.

Kornfeld, M., Frison, G.C., Larson, M.L., 2010. Prehistoric Hunters of the High Plainsand Rockies, third ed. Left Coast Press, Walnut Creek, California.

Krull, E.S., Skjemstad, J.O., 2003. d13C and d15N profiles in 14C-dated oxisol andvertisols as a function of soil chemistry and mineralogy. Geoderma 112, 1e29.

Lee, C.M., Lee, J.B., Turnbull, J.C., 2011. Refining the chronology of the Agate Basincomplex: radiocarbon dating the Frazier Site, northeastern Colorado. PlainsAnthropologist 56, 243e258.

Lee, J.B., Bennett, S.D., Crawford, G.T., 2012. Analysis of bison remains. ResearchContribution No. 86. In: Mitchell, M.D. (Ed.), Agate Basin Archaeology at BeaconIsland. Prepared by the PaleoCultural Research Group, Arvada, Colorado,pp. 85e112. Submitted to the State Historical Society of North Dakota, Bismarckand the U.S. Department of the Interior, National Park Service, Washington D.C.

Leyden, J.J., Oetelaar, G., 2001. Carbon and nitrogen isotopes in archeological bisonremains as indicators of paleoenvironmental change in southern Alberta. GreatPlains Research: A Journal of Natural and Social Sciences 11, 3e23.

Mandel, R.D., 2008. Buried paleoindian-age landscapes in stream valleys of thecentral Plains, USA. Geomorphology 101, 342e361.

Mandel, R.D., 2012. Geoarchaeology of area A. Research Contribution No. 86. In:Mitchell, M.D. (Ed.), Agate Basin Archaeology at Beacon Island. Prepared by thePaleoCultural Research Group, Arvada, Colorado, pp. 9e25 (Submitted to theState Historical Society of North Dakota, Bismarck and the U.S. Department ofthe Interior, National Park Service, Washington D.C.).

Mason, J.A., Miao, X., Hanson, P.R., Johnson, W.C., Jacobs, P.M., Goble, R.J., 2008.Loess record of the Pleistocene-Holocene transition on the northern and centralGreat Plains, U.S.A. Quaternary Science Reviews 27, 1772e1783.

Mayewski, P.A., Meeker, L.D., Whitlow, S., Twickler, M.S., Morrison, M.C., Alley, R.B.,Bloomfield, P., Taylor, K.C., 1993. The atmosphere during the Younger Dryas.Science 261, 195e197.

Meltzer, D.J., Holliday, V.T., 2010. Would North American Paleoindians have noticedYounger Dryas age climate changes? Journal of World Prehistory 23, 1e41.

Mickelson, D.M., Clayton, L., Fullerton, D.S., Borns Jr., H.W., 1983. The late Wisconsinglacial record of the Laurentide ice sheet in the United States. In: Porter, S.C.(Ed.), Late Quaternary Environments of the United States, The Late Pleistocene,vol. 2. University of Minnesota Press, Minneapolis, pp. 3e37.

Mitchell, M.D., 2012a. Project overview. Research Contribution No. 86. In:Mitchell, M.D. (Ed.), Agate Basin Archaeology at Beacon Island. Prepared by thePaleoCultural Research Group, Arvada, Colorado, pp. 1e8 (Submitted to theState Historical Society of North Dakota, Bismarck and the U.S. Department ofthe Interior, National Park Service, Washington D.C).

Mitchell, M.D., 2012b. Archaeological field investigations in Area A. ResearchContribution No. 86. In: Mitchell, M.D. (Ed.), Agate Basin Archaeology at BeaconIsland. Prepared by the PaleoCultural Research Group, Arvada, Colorado,pp. 27e72 (Submitted to the State Historical Society of North Dakota, Bismarckand the U.S. Department of the Interior, National Park Service, WashingtonD.C).

Mitchell, M.D., Johnston, C.M., 2012. Technological analysis of the modified stoneassemblage. Research Contribution No. 86. In: Mitchell, M.D. (Ed.), Agate Basin


Archaeology at Beacon Island. Prepared by the PaleoCultural Research Group,Arvada, Colorado, pp. 113e156 (Submitted to the State Historical Society ofNorth Dakota, Bismarck and the U.S. Department of the Interior, National ParkService, Washington D.C).

Moran, S.R., Arndt, M., Bluemle, J.P., Camara, M., Clayton, L., Fenton, M.M.,Harris, K.L., Hobbs, H.C., Keatinge, R., Sakreiter, P.K., Salomon, N.L., Teller, J., 1976.Quternary stratigraphy and history of North Dakota, southern Manitoba, andnorthwestern Minnesota. In: Mahaney, W.C. (Ed.), Quaternary Stratigraphy ofNorth America. Dowden, Hutchison, and Ross, Stroudsburg, Pennsylvania,pp. 133e158.

Morrison, K.D., 1994. Monitoring regional fire history through size-specific analysisof microscopic charcoal: the last 600 years in south India. Journal of Archaeo-logical Science 21, 675e685.

Nikiforoff, C.C., 1949. Weathering and soil evolution. Soil Science 67, 219e223.Patterson III, W.A., Edwards, K.J., Maguire, D.J., 1987. Microscopic charcoal as a fossil

indicator of fire. Quaternary Science Reviews 6, 3e23.Picha, P., 2012. Paleoenvironmental context: the shell record. Research Contribution

No. 86. In: Mitchell, M.D. (Ed.), Agate Basin Archaeology at Beacon Island.Prepared by the PaleoCultural Research Group, Arvada, Colorado, pp. 183e190(Submitted to the State Historical Society of North Dakota, Bismarck and theU.S. Department of the Interior, National Park Service, Washington D.C).

Piperno, D.R., 2006. Phytoliths: a Comprehensive Guide for Archaeologists andPaleoecologists. AltaMira Press, New York.

Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C., Blackwell, P.G.,Buck, C.E., Burr, G., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G.,Friedrich, M., Guilderson, T.P., Hughen, K.A., Kromer, B., McCormac, F.G.,Manning, S., Bronk Ramsey, C., Reimer, R.W., Remmele, S., Southon, J.R.,Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E., 2009.Intcal09 and Marine09 radiocarbon age calibration curves, 0-50,000 years calBP. Radiocarbon 51, 1111e1150.

Roberts, F.H.H., 1961. The Agate Basin Complex, Homenaje a Pablo Martinez del Rioen el XXV Aniversario de la Edicion de Los Origenes Americanos. InstitutoNacional de Antropologia e Historia, Mexico City, pp. 125e132.

Rovner, I., 2001. Cultural behavior and botanical history: phytolith analysis in smallplaces and narrow intervals. In: Meunier, J.D., Colin, F. (Eds.), Phytoliths: Ap-plications in Earth Sciences and Human History. A.A. Balkema Publishers, Lisse,pp. 119e127.

Ruhe, R.V., 1969. Quaternary Landscapes in Iowa. Iowa State University Press, Ames.Scheoneberger, P.J., Wysocki, D.A., Benham, E.C., Broderson, W.D., 2002. Field book

for describing and sampling soils. National Soil Survey Center, U.S. Departmentof Agriculture, Lincoln, Nebraska.

Scott, L., 2002. Microscopic charcoal in sediments: Quaternary fire history of thegrassland and savanna regions in South Africa. Journal of Quaternary Science 17,77e86.

Sellet, F., 1999. A Dynamic View of Paleoindian Assemblages at the Hell Gap Site,Wyoming: Reconstructing Lithic Technological Systems. Department of An-thropology, Unpublished Ph.D. dissertation. Southern Methodist University,Dallas, Texas.

Sellet, F., 2001. A changing perspective on Paleoindian technology and typology: aview from the northern Plains. Artic Anthropology 38, 48e63.

Slessman, S.A., 2004. The Frazier Site: an Agate Basin Occupation and LithicAssemblage on the Kersey Terrace, Northeastern Colorado. Department of An-thropology, Unpublished Master's thesis. Colorado State University, Fort Collins.

Stanford, D.J., 1999. Paleoindian archaeology and late Pleistocene environments inthe Plains and Southwestern United States. In: Bonnichsen, R., Turnmire, K.L.(Eds.), Ice Age Peoples of North America: Environments, Origins, and Adapta-tions. Center for the Study of the First Americans, Oregon State University Press,Corvallis, pp. 281e339.

Thornthwaite, C.W., 1948. An approach toward a rational classification of climate.Geographical Review 38, 55e94.

Timpson, M.E., 2003. Appendix B: natural stratigraphy and geomorphology in AreaA. Research Contribution No. 54. In: Ahler, S.A. (Ed.), Resurvey and Test Exca-vations at Beacon Island in Lake Sakakawea, Mountrail County, North Dakota.Paleocultural Research Group, Flagstaff, Arizona, pp. 136e138 (Submitted to theState Historical Society of North Dakota, Bismarck).

Timpson, M.E., Ahler, S.A., 2003. Geomorphic Analysis of Beacon Island, Site32MN234, Mountrail County, North Dakota. Research Contribution No. 47.Paleocultural Research Group, Flagstaff, Arizona (Submitted to the State His-torical Society of North Dakota, Bismarck).

USDA NRCS, 2011. The PLANTS Database. National Plant Data Center, Baton Rouge,LA 70874e4490 USA, 31 May 2011. http://plants.usda.gov.

Valero- Garc�es, B.L., Laird, K.R., Fritz, S.C., Kelts, K., Ito, E., Grimm, E.C., 1997. Holo-cene climate in the northern Great Plains inferred from sediment stratigraphy,stable isotopes, carbonate geochemistry, diatoms, and pollen at Moon Lake,North Dakota. Quaternary Research 48, 359e369.

Wheeler, R.P., 1954. Selected projectile point types of the United States. Bulletin ofthe Oklahoma Anthropological Society 2, 1e6.

Wyckoff, D.G., 1985. The Packard Complex: Early Archaic pre-Dalton occupations onthe prairie-woodland border. Southeastern Archaeology 4, 1e26.

Yansa, C.H., 2006. The timing and nature of Late Quaternary vegetation changes inthe northern Great Plains, USA and Canada: a re-assessment of the sprucephase. Quaternary Science Reviews, 263e281.



















http://www.hprcc.unl.edu/cgi-bin/cli_perl_lib/cliMAIN.pl?nd6383








































































































































































http://plants.usda.gov

















Geoarchaeology and paleoenvironmental context of the Beacon Island site, an Agate Basin...

Documents

Transcript of Geoarchaeology and paleoenvironmental context of the Beacon Island site, an Agate Basin...