The Rosa-Keystone Dunes Field: The geoarchaeology and paleoecology of a late Quaternary stabilized dune field in Eastern Beringia

Abstract

Stabilized sand sheets and dunes hold a remarkable amount of information on paleoenvironmental conditions under which late Quaternary landscapes evolved in northern subarctic regions. We provide the results of a project focused on understanding the development of lowland environments and ecosystems, including dunes and sand sheets, which were critical habitat for early human occupations in subarctic regions. Our study area is the Rosa-Keystone Dunes Field in the Shaw Creek Flats of the middle Tanana River basin, interior Alaska, one of the oldest continuously occupied areas in North America (14,000 cal. BP to present). The disturbance regimes of reactivated dunes and associated forest fire cycles between 12,500 and 8800 cal. BP fostered a unique early to mid-successional mixed vegetation community including herbaceous tundra, shrubs, and deciduous trees. This environment provided key habitats for large grazers and browsers, significant resources for early hunter-gatherer populations in central Alaska. After 8000 cal. BP, the expansion of black spruce and peatlands heightened landscape stability but decreased the range of local habitat for large grazers. Hunter-gatherer economic change during these periods is consistent with human responses to local and regional landscape disturbance and restructuring.

Keywords

Eastern Beringia hunter-gatherers paleoecology sand dunes Shaw Creek Flats subarctic interior Alaska

Introduction

Eolian deposits cover a wide expanse of eastern Beringian (Alaska, US; and Yukon, Canada) landscapes. Sand dunes and low-lying sand sheets developed along major drainage systems and coastal regions, and are estimated to cover over 32,500 km² across eastern Beringia (Figure 1a; >30,000 km² in Alaska (Black, 1951; Hopkins, 1982)) and >2500 km² in the Yukon (Wolfe et al., 2011)).

Figure 1.

(a) Hillshade model of Alaska showing the extent of eolian sand sheets and dune deposits (adapted from Wolfe et al. (2009, 2011)). (b) Hillshade model showing the extent of sand dunes and locations of archaeological sites (mentioned in the text) in the Shaw Creek Flats, middle Tanana Valley. (c) Hillshade model showing transverse dune complex in Rosa-Keystone Dune Field and sampling locales (Dunes 23 and 25 were not sampled and not mention in the text). (d) LIDAR of selected portion of the Rosa-Keystone Dune Field near Rosa Creek showing transverse dune complex with reversing crests and superimposed parabolic and star-like forms and blowouts.

Despite their vast extent in eastern Beringia, little research has focused on the paleoenvironmental and paleoecological aspects of dune and sand sheet histories (notable exceptions include Collins, 1985; Galloway et al., 1990; Lea, 1996; Lea and Waythomas, 1990; Mann et al., 2002). An understanding of the history of dune stability and activation can indicate periods of local vegetative loss and aridity, shifts in paleo-wind directions and intensity, and changes within the composition of local habitat and biotic communities (Matthews and Seppälä, 2014; Wolfe and Nickling, 1997; Wolfe et al., 2007). The reconstruction of long-term changes in local ecological contexts informs us about the mechanisms driving shifts in faunal diversity and in human ecology and foraging behavior in eastern Beringia (Bigelow and Powers, 2001; Graf and Bigelow, 2011; Mason and Bigelow, 2008; Mason et al., 2001; Potter, 2008a, 2008b, 2008c, 2011; Potter et al., 2013; Yesner, 2007).

Cultural systems confront many challenges in the midst of changing climates and environments and variability across these systems is reflected in the technological and social adaptations to local ecological contexts (Butzer, 1982). Our goal is to provide long-term windows into local landscape evolution in order to understand how humans may have responded to ecological and environmental change in subarctic settings, in particular eastern Beringia. This paper provides the results of our studies on an eastern Beringian dune field, the Rosa-Keystone Dunes Field (RKD Field; Figure 1b) in the Shaw Creek Flats (SCF) in the middle Tanana Valley (mTV), which focuses on the relationships of past stable and active eolian landscapes and the late Glacial and early Holocene environments under which they formed in interior Alaska.

The formation history of the RKD Field provides an important 14,500-year window into the ecology of a region containing a rich and diverse archaeological record, including the earliest evidence for human colonization of northern North America (Holmes, 2011; Potter et al., 2013). Understanding environmental change that influenced prehistoric human land use can help address changes in technology and subsistence and adaptive strategies through time.

General setting

mTV

The Tanana River is predominately a glacially fed, braided river system that drains the Yukon-Tanana Uplands and the foothills and mountainous areas of the central and eastern Alaska Range (Wahrhaftig, 1965). This project is focused on the mTV basin, defined here as the region between Healy Lake to the east and Fairbanks to the west (Figure 1a). Loess and eolian sand deposits mantle much of the lowlands and hills of the region, and currently provide excellent substrates for deciduous-dominated and mixed coniferous and deciduous forest growth and soil development (Dilley, 1998). The areal extent of these eolian sand deposits in subarctic Alaska is unknown, but dunes cover large tracts of land in the interior regions along most, if not all, of the large rivers including the Tanana River. In the Tanana Lowlands, dunes and sand sheets are estimated to cover over 2000 km² (Collins, 1985; Péwé, 1965 [1977]).

Two late Pleistocene regional glacial periods, the Delta and Donnelly glaciations, are pertinent to the discussion of loess and dune and sand sheet development in the mTV (Hamilton, 1973; Péwé and Reger, 1983). The Delta and Donnelly glaciations (70–40,000 and 25–16,600 years ago, respectively), are recognized by moraines and glacial drift situated on the north side of the Alaska Range (Matmon et al., 2010; Reger et al., 2008; Ten Brink and Waythomas, 1985). Katabatic winds derived from the Delta and Donnelly massive ice sheets created increasing wind frequency and intensity of eolian erosion and deposition across the Tanana basin (Muhs and Budhan, 2006; Péwé and Reger, 1983; Thorson and Bender, 1985). Péwé and others (Péwé, 1975; Reger et al., 2008) have hypothesized that an immense amount of fluvial deposition within the Tanana Valley occurred during glacial periods and provided sediment sources for subsequent wind deposition and the growth of dune fields and accumulation of sand sheets and loess covering the hills.

The SCF

Geology

The SCF is a low-lying alluvial plain or terrace formed by the aggradation of the Tanana River during the height of the Donnelly glaciation (Last Glacial Maximum (LGM); Reger et al., 2008). Shaw Creek and its tributaries drain the Yukon-Tanana Uplands into the Tanana River in this region.

A Delta-aged terminal moraine, less than 16 km south of the SCF, was dissected during the subsequent interglacial period, creating a gap among till deposits (Reger et al., 2008). Donnelly-aged terminal moraines are situated in the Delta River Valley, approximately 32–48 km south of the SCF, and fan deposits associated with this period fill gaps in the Delta-aged terminal moraine. A period of massive alluviation during the Donnelly glaciation forced the Tanana River northward toward the southern bases of the Uplands (Reger et al., 2008). Gravel, sand, and silt from the Donnelly-aged outwash fan underlie the modern surface of the SCF (Reger et al., 2008). Reger and Péwé (2002) and Reger et al. (2008: 14) suggested that the upper gravels from the SCF were deposited around 23–21,000 cal. BP). Broad, unvegetated braided channels of the Tanana River and the newly formed surface of the SCF provided ample sediment sources for subsequent eolian remobilization of sand and silt. Loess blankets the hills of the Uplands adjacent to the SCF (Dilley, 1998; Péwé, 1965 [1977]; Péwé and Reger, 1983; Weber, 1971; Weber et al., 1978).

Sand dunes and sheets cover a vast majority of the northern and western portions of the SCF (Figure 1), and the adjoining lower valleys of the Uplands up to an elevation of 487 m a.s.l. (Dilley, 1998; Jordan and Yarborough, n.d.; Péwé, 1965 [1977]; Weber et al., 1978). The formation histories of these Shaw Creek deposits have yet to be detailed. Péwé (1965 [1977]: 47; Péwé and Reger, 1983: 30) estimated dunes in the SCF as covering 100 km², many of which are dissected by several drainages including Keystone, Rosa, and Shaw Creeks.

The larger expanse of sand dunes covering the SCF and adjacent uplands has been informally referred to as Shaw Creek or SCF Dunes (Dilley, 1998; Hopkins, 1982; Jordan and Yarborough, n.d.). The dunes that lie at the far western edge of the SCF area between Rosa and Keystone Creeks have been informally referred to as Rosa Creek Dune Field or the Rosa-Keystone Dunes (Alaska Boreal Forest Council, 2002; Reger and Péwé, 2002); they will be referred to here as the ‘Rosa-Keystone Dunes’ Field (RKD Field).

Péwé (1965 [1977], 1975) and Weber et al. (1978) suggested that the dunes in the SCF formed during the Delta glaciation based on the attribution of overlying loess deposits to the more recent Donnelly glaciation. However, several researchers have attributed the formation of dunes in the lowlands and eolian sand and silt deposits on hills in the mTV to the Donnelly and late Glacial periods (25–11,600 cal. BP; Dilley, 1998; Reger et al., 2008: 20). Kreig and Reger (1982), Lea and Waythomas (1990), and Dilley (1998) suggested that the RKD Field, and likely the greater expanse of dunes in the SCF, experienced local reactivation during the late Glacial and Holocene, although absolute ages on dune and sand sheet development and reactivation have been lacking until now.

The surface deposits of the SCF contain approximately 1 m of organic-rich silt, overlying 1–3 m of silt and sand, overlying sand and gravels to at least 10 m in depth of the surface (Kreig and Reger, 1982: 35). Holocene aggradation in the SCF has been extremely minor compared with that occurring during glacial periods (Dilley, 1998). Surficial soils have been mapped around the lower reaches of Keystone, Rosa, and Shaw Creeks (Schoephorster, 1973). Poorly drained silt loams and peats cover the lower lying areas of the SCF (Kreig and Reger, 1982; Schoephorster, 1973). Well to moderately drained, shallow silt loams have developed on steep slopes of higher ridges. Silt loams that are thicker and well drained are present on south-facing slopes of ridges and hills. Silt loams with increasing sandy substrates are present within dune fields (Schoephorster, 1973).

Vegetation, animals and humans

The geographic setting of the SCF against the southwestern foothills of the Yukon-Tanana Uplands contributes to the diverse topography and biota within this region. Modern vegetation over the mTV consists of muskeg and black spruce-larch forest in the lower terrain, closed canopy forests on lower slopes of the hills, and open mixed forests and tundra at higher elevations. Sedge tussocks, sedges, grasses, and shrub willow and dwarf birch dominate the muskeg. In areas of disturbance, alder and willow shrubs, grasses, sedges, and horsetail (Equisetum) tend to be early colonizers of newly exposed ground. On the edges of bedrock cliffs and terraces, the southern slope faces foster the presence of more xeric plant communities, such as Artemisia, sedges, and grasses that have been referred to as relicts of the Pleistocene steppe vegetation (Guthrie, 1990a, 2001).

The SCF interface with the Uplands creates a unique array of ecological niches for small and large mammals, birds, and fish, resources used by human populations in the mTV for over 14,000 years (Potter et al., 2013; Yesner, 2007). Moose are relatively abundant in the Flats, and the Fortymile caribou herd historically ranged in the area during the winter months (Durtsche and Hobgood, 1990). Black and brown bears, beaver, coyotes, fox, lynx, martin, and weasels are among the other furbearers that inhabit this region. The wetlands offer excellent habitat for birds that include waterfowl (swan, geese, ducks), eagles, gulls, hawks, ravens, and ruffed grouse. Salmon species (coho, Chinook, and grayling) spawn in the SCF drainage system, and pike are present in some lakes.

Since the early 1990s, the SCF has been a major focus of archaeological research in interior Alaska. Several important archaeological sites are located in the area within 10 km of each other, including the Broken Mammoth, Mead, Swan Point, and Bachner sites, and are situated on the edges of bluffs at the interface of the uplands and lowlands, on bedrock knobs and dunes at the edge of the flats, and along raised landforms of lakes (Figures 1b and 2(1); Holmes, 2011; Potter et al., 2013; Reuther, 2013; Yesner, 2007). These sites are well-stratified, multi-component records of prehistoric occupation dating back nearly 14,500 cal. BP, and provide undisputed evidence of some of the earliest human occupations in the Western Hemisphere. A wide range of site types are represented throughout the late Glacial and Holocene, including residential base camps, short-term camps, stations, caches, and lithic workshops (Holmes, 2001; Potter et al., 2013). Faunal remains recovered from these occupations provide a rich Paleoindian subsistence record, indicating that a broad spectrum of resources was exploited throughout the area during the late Glacial and early Holocene (Holmes, 2011; Potter et al., 2013; Yesner, 1996, 2007). This fauna include grazers (bison, wapiti, horse, mammoth), generalized browsers (caribou and moose), a variety of waterfowl and terrestrial birds, and other mammals (e.g. bear, arctic fox, hare, marmot, and ground squirrel).

Figure 2.

(1) Calibrated radiocarbon ages on archaeological components from the Broken Mammoth, Mead, Swan Point, and Bachner sites; calibrated radiocarbon and IRSL ages, from soils and sands, respectively, from the Rosa-Keystone sections. Calibrated radiocarbon ages (IntCal13) quoted as 2-sigma ranges, and IRSL as 2 sigma ranges for comparison with radiocarbon ages. (2) Archaeofaunal assemblage changes in large mammal composition in the middle Tanana Valley (Potter, 2011): 2a = no archaeological component data before 14,500 cal. BP; 2b = Swan Point CZ4b archaeofaunal assemblage with horse, mammoth, and caribou dominating the large mammals (Lanoë and Holmes, unpublished data). (3) Generalized pollen zones for the middle Tanana Valley (adapted from Bigelow (1997) and Tinner et al. (2006)): Herb Zone = herbaceous taxa (Poaceae, Cyperaceae, and Artemisia) with Salix sp. as dominant shrub taxa; Betula Rise = Betula sp. increases (3a = short-term increase in xerophytic species including Juniperus, Salix, Artemisia, and Poaceae); Populus Rise = Populus sp. increases amongst Betula and Salix shrubs; Picea Rise = 3b – Picea glauca and Alnus viridis increased; 3c – Picea mariana increased. (4) Generalized late Pleistocene and Holocene vegetation records for central Alaska (adapted from Hu et al. 2006 and Lynch et al. 2004).

The presence of waterfowl indicates that wetlands were present in the SCF area between 14,500 and 13,000 cal. BP (Yesner, 1996, 2007). The wide variety of grazers may indicate an important late Glacial refugium of herbaceous tundra communities during this period (Figure 2(2); Yesner, 1996, 2007).

Several of these taxa are extinct or locally absent within the lowlands of the SCF and mTV by the early Holocene, including mammoth, horse, bison, wapiti, and ground squirrel. It is clear that mammalian diversity was higher in this region, and throughout much of interior Alaska, prior to 11,000 cal. BP (Guthrie, 1990a). Our study is geared toward providing environmental and ecological contexts to understand the shift in the lessening of mammalian diversity and human resource foci throughout the Holocene in this region. While there is debate as to the cultural relationships between early cultural groupings (East Beringian tradition, Nenana or Chindadn, and Denali complexes (Goebel et al., 1991; Hamilton and Goebel, 1999; Holmes, 2001; Potter et al., 2013), we are focused on the ecological relationships between these populations and the regional landscape.

Study area

The RKD field

An expansive field of dunes comprises the majority of the terrain between Rosa and Keystone Creeks in the SCF, covering roughly a 33.4 km² area. The large majority of the dunes in the RKD Field abut the bases of bedrock ridges at the northern and western extent of the SCF (Figure 1b). The surrounding hills range in elevation between 366 and 762 m a.s.l., while the dune field ranges between 290 and 467 m a.s.l.

Very little fieldwork has been conducted in the RKD field to understand the specifics of their morphology and depositional history. Péwé (1965 [1977]) initially described the SCF area dune fields, and Weber (1971) and Weber et al. (1978) provided more detailed mapping. The RKD Field was mapped as a part of the larger expanse of dunes within this area. Kreig and Reger (1982) provided the initial detailed field description of the RKD Field based on aerial photograph analysis and geotechnical coring results, with subsequent field descriptions by Dilley (1998) and Jordan and Yarborough (n.d.).

The RKD Field is a complex dune field dominated by large crescentic morphologies, including barchanoid ridges and transverse dunes with reversing crests (Figure 1d; McKee, 1979), rising up to 20 m above the surrounding terrain (Kreig and Reger, 1982; Reuther, 2013). Interdunal areas between crescentic dunes are currently vegetated, with the exceptional few that contain very small ponds. The morphologies of dune forms increase in complexity toward the field’s southern extent, likely a result of episodic dune reactivation, as discussed below. As noted above, absolute dating on the timing of the initial dune formation and potential reactivation of the RKD Field was absent prior to our current study.

Field and laboratory methods

The areal extent of dune locales and morphologies was mapped through the aide of aerial photography and Light Detection and Ranging (LIDAR) intensity imagery (Hubbard et al., 2011), and supplemented by field observations. These geomorphic data assisted in the selection of the geologic sections sampled, along with the interpretation of wind patterns, environmental conditions (e.g. presence of vegetation and moisture) within which these dunes formed, and reactivation and modification of landforms after their initial formation. Dune cuts in the RKD Field and along road corridors were trenched to (1) describe the lithostratigraphy and pedostratigraphy (sediment and soil stratigraphy, respectively) to understand periods of eolian aggradation and landscape stability and (2) recover geochronological, paleoecological, and sedimentological samples for use in providing a more detailed reconstruction of the geologic and environmental characteristics of these settings through time (Figure 1c).

Radiocarbon dating of organic materials was used to establish periods of landform stabilization and soil formation and provide bracketing ages on periods of sand accumulation, while infrared stimulated luminescence (IRSL) dating of buried feldspar sand grains allowed us to directly date the timing of initial dune construction (Olley et al., 2004; Wintle, 2008). Radiocarbon ages are presented as calibrated 2-sigma age ranges and are presented in Table 1. The IRSL age estimates, shown in Table 2, are also presented as 2-sigma age ranges.

Table 1.

Radiocarbon dates from the Keystone Dune site and RKD Dune 2 sections. Conventional radiocarbon dates were calibrated and median probabilities calculated using Calib v7.1 software (Stuiver et al., 2015) and the IntCal13 terrestrial calibration model (Reimer et al., 2013). Pooled averages were calculated following Ward and Wilson (1978).

Depth (cmBS)	Unit (soil horizon)	Lab no.	Material	δ¹³C	Conventional age (BP)	Cal. BP (median probability) 2σ	Remarks
Keystone Dune Site
40	2 (Bwb)	AA96878	Charcoal (unidentified angiosperm (shrub?))	−24.20	7980 ± 50	9000 (8850) 8650	Recovered from the boundary of Bwb and C/B horizons
314–315	1 (Pedocomplex 2)	Beta-313858	Charcoal (Populus cf. tremuloides)	−25.00	9720 ± 50	11,240 (11,160) 10,870	Recovered from uppermost Ab horizon in Pedocomplex 1 in Trench 11-01
326	1 (Pedocomplex 2)	AA88626	Charcoal (possible Alnus sp.)	−29.00	10,370 ± 60	12,520 (12,260) 12,040	Recovered from lowest Ab horizon in Pedocomplex 2 in Trench 1a
336	1 (Ab2)	AA88625	Charcoal (possible Alnus sp.)	−28.50	9670 ± 140	11,340 (10,990) 10,580	Recovered from Ab2 horizon in Trench 1a. (REJECTED as stratigraphic marker: stratigraphically inconsistent)
397	1 (Pedocomplex 3)	AA96879	Charcoal	−24.90	10,310 ± 60	12,400 (12,125) 11,830	Sample recovered from the lowest Ab horizon in Pedocomplex 3 in Trench 11-01
420	1 (Ab4)	Beta-311495	Charcoal (Populus cf. tremuloides)	−26.00	10,590 ± 50	12,690 (12,580) 12,420	Recovered from Ab4 horizon in Trench 11-01
420	1 (Ab4)	UGAMS#22800	Charcoal	−30.0	10,590 ± 30	12,670 (12,590) 12,430	Recovered from Ab4 horizon during the 2015 excavations.
420	1 (Ab4)	–	Charcoal	–	10,590 ± 20	12,660 (12,590) 12,540	Average of UGAMS#22800 and Beta-311495
460–480	1 (Pedocomplex 5)	UGAMS#18143	Wapiti (elk) collagen	−19.4	10,990 ± 30	12,970 (12,830) 12,730	Recovered from Pedocomplex 5 from near a hearth in excavation in 2014. (REJECTED as stratigraphic marker: stratigraphically inconsistent)
460–480	1 (Pedocomplex 5)	UGAMS#18143	Wapiti (elk) hydroxyproline (amino acids specific)	−25.6	11,470 ± 40	13,420 (13,320) 13,220	Recovered from Pedocomplex 5 from near a hearth in excavation in 2014
460–480	1 (Pedocomplex 5)	AA105234	Charcoal (Betula sp.)	−25.8	11,450 ± 40	13,410 (13,300) 13,190	Recovered from hearth in Pedocomplex 5 in excavation in 2014
460–480	1 (Pedocomplex 5)	–	Bone – hydroxyproline (amino acids specific) and charcoal	–	11,460 ± 30	13,400 (13,310) 13,220	Average of UGAMS#18143 (hydroxyproline) and AA105234
RK Dune 2
660	1 (Pedocomplex 1)	AA88746	sediment (alkali-soluble)	−23.90	9840 ± 60	11,390 (11,240) 11,170	Sample recovered from the lowest Ab horizon in Pedocomplex 1.
660	1 (Pedocomplex 1)	AA88745	sediment (alkali-insoluble)	−23.20	17,770 ± 180	21,960 (21,500) 20,970	Sample recovered from the lowest Ab horizon in Pedocomplex 1.

Table 2.

Infrared stimulated luminescence (IRSL) ages from the Rosa-Keystone Dune Field sampled sections.

Lab no.	Field sample	Section	Depth (cmBS)	Central age (ka) 1σ	Central age (ka) 2σ
UW1956	Keystone-OSL-SAMPLE-01	Keystone Dune Site	420–460	11.3 ± 0.8	11.3 ± 1.6
UW2655	KD2-OSL-01	RK Dune 2	700	14.9 ± 1.3	14.9 ± 2.6
UW2656	KD13-OSL-01	RK Dune 13	90–100	15.9 ± 1.1	15.9 ± 2.2
UW2657	KD26-OSL-01	RK Dune 26	155	12.5 ± 0.8	12.5 ± 1.6
UW2658	KD26-OSL-02	RK Dune 26	750	8.9 ± 0.5	8.9 ± 1.0
Lab no.	Field sample	Section	Proportion (%)	Mixture model age 1σ Component 1	Mixture model age 2σ Component 2
UW2655	KD2-OSL-01	RK Dune 2	72.9	11.4 ± 1.0	–
UW2655	KD2-OSL-01	RK Dune 2	27.1	–	31.7 ± 5.6

A selection of charcoal and woody fragments and seeds, shown in Table 3, from buried soils were taxonomically identified to determine shrub and tree taxa locally present within the SCF and RKD Field over their developmental history. Particle size analysis and inorganic and organic carbon contents (%CaCO₃ and %OC, respectively) were used to supplement and complement field stratigraphic observations to understand changes in the composition of particle sizes and the development of soils. The Supplemental Materials (available online) contain detailed information on the field and laboratory methods and data, including stratigraphic descriptions, sedimentological analyses, and radiocarbon and IRSL dating.

Table 3.

Wood identifications from the Keystone Dune site.

Stratigraphic section	Unit (soil horizon)	Depth (cmBS)	Description	Species ID (no. of fragments identified)	Conventional radiocarbon date (BP)	δ¹³C (‰)	Lab no.	Cal. BP (median probability) 2σ	Indirect stratigraphic date (cal. BP)
Keystone Dune Site	2 (Bwb)	40	Charcoal	Unidentified angiosperm (shrub?) (n = 1)	7977 ± 48	−24.20	AA96878	9000 (8850) 8650	–
Keystone Dune Site	1 (Pedocomplex 2; upper Ab horizon)	314–315	Charcoal	Gymnosperm (n = 1)	–	–	–	–	11,240 (11,160) 10,870
Keystone Dune Site	1 (Pedocomplex 2; upper Ab horizon)	314–315	Charcoal	Populus cf. tremuloides (n = 2)	9720 ± 50	−25.00	Beta-313858	11,240 (11,160) 10,870	–
Keystone Dune Site	1 (Ab2)	336	Charcoal	Alnus sp. (n = 2)	9670 ± 140	−28.50	AA88625	11,340 (10,990) 10,580	–
Keystone Dune Site	1 (Ab2)	336	Charcoal	Picea sp. (n = 1)	–	–	–	–	11,340 (10,990) 10,580
Keystone Dune Site	1 (Pedocomplex 3; lowest Ab horizon)	397	Charcoal	Populus cf. tremuloides (n = 1)	–	–	–	–	12,400 (12,125) 11,830
Keystone Dune Site	1 (Ab4)	420	Charcoal	Populus cf. tremuloides (n = 1)	10,590 ± 50	−26.00	Beta-311495	12,690 (12,580) 12,420	–
Keystone Dune Site	1 (Ab4)	420	Charcoal	Unidentified angiosperm (shrub?) (n = 1)	–	–	–	–	12,690 (12,580) 12,420
Keystone Dune Site	1 (Pedocomplex 5) (hearth matrix)	460–480	Charcoal	Betula sp. (n = 18)	–	–	–	–	13,410 (13,300) 13,190
Keystone Dune Site	1 (Pedocomplex 5) (hearth matrix)	460–480	Charcoal	Populus sp. (n = 3)	–	–	–	–	13,410 (13,300) 13,190

Results

Section stratigraphy and dating

Keystone Dune Site

The Keystone Dune Site (KDS) is situated at the northeastern horn of a parabolic dune that overlooks the lower reaches of Keystone Creek and its valley floor to the west and south (Figures 1c and 3). Several parabolic dunes, including the KDS, are superimposed over and integrated with crescentic dune forms at the northeastern extent of the RKD Field. The northeastern horn of the KDS trends 150°SSE to 330°NNW toward the dune’s slipface, and rises between 20 and 30 m above the surrounding valley floor.

Figure 3.

The Keystone Dune site examined and sampled in Rosa-Keystone Dunes Field.

A total of 700 cm of vertical stratigraphy was described from five 1-m-wide trenches placed along a 200–300 m road cut through the KDS (Figures 3 and 4). A 100 × 50-cm pit was dug at the top of the horn to describe the upper 1 m stratigraphy.

Figure 4.

Generalized profiles of the Keystone Dune site and RKD Dune 2 sections stratigraphy. Italicized ages are IRSL ages (2σ); **averaged radiocarbon ages.

The cut reaches 15 m high at its peak and is between 3 and 6 m high at the northern and southern ends. Dune sands (Unit 1) consist of over 650 cm of coarse to fine sands that generally fine upward. The lower 550 cm of Unit 1 sands are intercalated coarse to fine sands. Unit 1 beds and laminations are horizontal to sub-horizontal (<5° slope) lying and range from 0.5 to 5 cm thick. Sands become finer grained and more massive in the upper 100 cm of Unit 1 as sediments transition to the sandy and silt loams of Unit 2 (80 cm thick).

Pedocomplexes 4 and 5 represent the oldest periods of soil development currently recognized at the KDS. These pedocomplexes consist of dark brown beds of higher silt content (fine sandy loams) that are more oxidized than the sand deposits above and below, and circular oxidized stains and thin laminations of organics that represent ancient root casts. A 10- to 15-cm-thick bed of fine sand with a thin (<1 cm thick) weakly developed black Entisol (Ab4 horizon) that contains charred plant materials separates these pedocomplexes.

Radiocarbon dating on charcoal from archaeological features (i.e. hearths) and associated animal bone fragments provide age estimates on the formation of Pedocomplex 5. A charred Betula fragment from the hearth was radiocarbon dated to 13,410–13,190 cal. BP (11,450 ± 40 BP; AA105234). A long bone fragment of a large ungulate was also radiocarbon dated to 12,970–12,730 cal. BP (10,990 ± 30 BP; UGAMS#18143 collagen) and 13,400–13,220 cal. BP (11,470 ± 40 BP UGAMS#18143 amino acids (hydroxyproline)). We suspect the later age accurately reflects the true age of the bone (see Supplemental Materials, available online). The weighted average of the accepted ages for the human occupation at the KDS is 13,430–13,230 cal. BP (11,460 ± 30 BP); this average age also indicates that Pedocomplex 5 was in development by this time.

Feldspar within Unit 1 medium-to-fine sands around 420–460 cmBS, between Pedocomplex 5 and the Ab4 horizon, were luminescence (IRSL) dated and provided a central age of 11,300 ± 1600 years ago (UW1956), a range between 12,900 and 9700 years ago (Figure 4). This IRSL date was taken from sands below the Ab4 horizon and above Pedocomplex 5. The older end of UW1956’s age range is around 330 years younger than the weighted average radiocarbon age on archaeology in Pedocomplex 5.

A series of at least 16 buried soils (Ab4, Ab3, and Ab2 horizons and Pedocomplexes 2 and 3) developed between 12,690 and 10,870 cal. BP, a period of 1820 years, and are preserved between 314 and 445 cmBS in Unit 1. These soils are weakly developed black Entisols (Ab horizons), are 0.5–2 cm thick, and display evidence of burning, interpreted here as forest fires (Figure 4). Each soil is separated by 2–23 cm of sands that show a progression of fining upward; coarser sands cover buried soils with sediments becoming finer sandy loams that soils have developed on. Populus cf. tremuloides (quaking aspen), Picea sp. (spruce), and angiosperm (likely Alnus sp.) were among the charcoal fragments identified throughout these soils. Similar charcoal-rich Entisols dating to 12,880–12,610 cal. BP (10,850 ± 40 BP; Beta-311361) have been identified in dune deposits in the upper Shaw Creek watershed (personal communication, Aubrey Morrison and Michael Yarborough, Cultural Resources, Consultants; see Reuther, 2013).

A third series of at least three weakly developed buried soils (Pedocomplex 1) are present at the interface of Units 1 and 2 between 65 and 80 cmBS. These soils are dark reddish brown to black sandy loams (Ab horizons; 1–2 cm thick Entisols) that are similar to the deeper buried soils observed in Unit 1 (Figure 4).

In the upper 50 cm of Unit 2, buried soils (Inceptisols) are more defined and thicken to between 3 and 10 cm thick. A discontinuous light brown sandy loam (Ab1 horizon; 3–10 cm thick) developed 20 cm above Pedocomplex 1 and a dark reddish brown sandy loam (Bwb horizon; 7–9 cm thick) partially weld onto the Ab1 horizon. Charcoal (unidentified angiosperm) recovered from the lower boundary of the Bwb horizon dated to 9000–8650 cal. BP (7980 ± 50 BP; AA96878). A surficial incipient podzolic forest soil (O-A/E-B horizons; 15–30 cm thick) is separated from the Bwb horizon by a yellowish gray loam (C1 horizon).

Particle sizes in the KDS stratigraphic column show a trend toward fining upward, with coarse and medium sands dominating Unit 1 sediments, while finer sands and silts compose the majority of Unit 2 sediments. Clay is present in minor amounts in Unit 1, but increases in Unit 2 likely because of increases in soil development. The highest presences of organic carbon and matter are generally associated with soils; while CaCO₃ content is much higher in the upper portion of the column as silt and clay particles become more dominant, possibly lending to their more calcareous nature than the sands.

Rosa-Keystone Dune 2

The Rosa-Keystone Dune 2 (RK Dune 2) section is located at the eastern horn of a long parabolic dune that is part of a set of compound dune forms at the eastern extent of the RKD Field (Figure 1c). Keystone Creek flows at the base of the eastern horn, and the RK Dune 2 section is situated on the right bank of the creek opposite the KDS section. The eastern horn trends roughly 160°SSE to 340°NNW toward the dune’s slipface; the horn’s slipface edge has a 50–60° slope to the northeast toward and approximately 15 m high above Keystone Creek’s floor. The windward slope of the horn has a gentler slope (~25–30°) to the southwest. Dune reactivation may have modified the southwestern edge of the horn. Wind direction during the period of the development of RK Dune 2 was from the south to the southwest.

A 1 m wide and 7 m deep trench was excavated at an 8 m high exposure, approximately 15 m south of the horn’s slipface edge, and a 50 cm² pit was dug at the top of the horn to describe the upper 1 m stratigraphy (Figures 4 and 5). Bedded and laminated coarse to fine sands (Unit 1) compose the lower 700 cm of stratigraphy. The slopes of beds and laminations are oriented toward the northeastern slipface of the horn, and become steeper in the upper portion of the column, denoting reworking of the dune. Slopes of the beds between 200 and 800 cmBS are 5–10°, while the slope between 100 and 200 cmBS increases to 30–35°. Beds and laminations disappear in the upper 27–100 cm of Unit 1 sands. The particle sizes of sediments at the RK Dune 2 section generally show fining upward throughout the column. Silt begins to increase around 27 cmBS as sandy and silty loams (Unit 2) become more prominent.

Figure 5.

RKD Dune 2, RKD Dune 13, and RKD Dune 26 examined and sampled in Rosa-Keystone Dunes Field.

Two thin soils (Entisols; Pedocomplex 1) are present between 650 and 660 cmBS. The upper Ab(2) horizon at 650 cmBS is a 0.5 cm thick discontinuous black organic stringer, while lower Ab(3) horizon at 660 cmBS is a continuous 1 cm thick dark reddish brown sandy loam (Figure 4). Charcoal was not recovered from either Pedocomplex 1 horizon. Sediment and soil organic fractions from the lower Ab horizon were assayed by radiocarbon with the alkali-soluble (i.e. humic acids) dating between 11,400 and 11,170 cal. BP (9840 ± 60 BP; AA88746), and the alkali-insoluble (i.e. soil residue) dating between 21,960 and 20,970 cal. BP (17,770 ± 180 BP; AA88745). Ideally, the age of the alkali-soluble fraction represents the average age for soil organic matter to develop and break down (mean residence time of organic carbon compounds) in the lower Ab horizon between the time of its formation and its burial (Haas et al., 1986; Martin and Johnson, 1995). For this reason, the alkali-soluble fraction date is considered as a minimum date for the formation of the soil, and a maximum age on its burial by the upper sand deposits (above 650 cmBS) at RK Dune 2. The 20,000-year-old radiocarbon soil residue age is regarded here as too old to date the soil development at this section (see Supplemental Materials, available online).

An IRSL sample (UW2655) was taken at 700 cmBS, and from 40 cm below the Ab(3) horizon that we are inferring dates to at least 11,400–11,170 cal. BP. The central age model estimate for the IRSL sample was 14,900 ± 2600 years ago. A finite mixture model on this sample showed a bimodal distribution of feldspar luminescence ages: 11,400 ± 2000 years ago (72.9% of sample) and 31,700 ± 11,200 years ago (27.1% of sample). The ages are stratigraphically coherent with the radiocarbon ages, given that IRSL ages from the central age model and largest component of the finite mixture model date to the late Glacial between 11,400 and 14,900 years ago. This range is consistent with IRSL ages on the lower sands at the nearby KDS. The older component of the finite mixture model likely represents older and partially bleached grains mixed in during this period of late Glacial activation of the dunes.

Soils in the upper portion of the RK Dune 2 stratigraphic column consist of incipient soil (Inceptisol) horizons developed within Unit 2. A dark reddish brown sandy loam (Bwb horizon; 7 cm thick) is present around 20–27 cmBS with an overlying light brown sandy loam (Ab1 horizon; 12 cm thick) between 9 and 21 cmBS. The Ab1 horizon is welded upon by more recent surficial soils (O-A/B horizons; 9 cm thick). These upper soil horizons in Unit 2 began to develop in finer grained and more massive sediments. Their welding indicates that the crest was highly stable throughout the Holocene, allowing for a series of soils to development one after another, similar to the KDS.

Rosa-Keystone Dune 13

Rosa-Keystone Dune 13 (RK Dune 13) consists of the southwestern edge of a low-lying crescentic dune or an irregular-shaped parabolic dune that developed at the southwestern limit of the RKD Field (Figures 1c and 5). This dune is part of a set of compound dunes and older crescentic dunes that had younger dune forms transposed over them. These dunes rise 4–5 m above the surrounding terrain, and their southwestern edges are dissected by Rosa Creek and its tributaries.

The axis of RK Dune 13 is oriented 340°NNW to 150°SSE. A 1 m wide trench was excavated to 160 cmBS at a 3 m high exposure, with the upper 60 cm of the stratigraphy disturbed. The lower 100 cm of the trench is fine and medium bedded sands (1–6 cm thick beds) that have a horizontal orientation. Buried soils were not observed at this section. Clay, organic carbon and matter, and CaCO₃ contents are relatively negligible throughout the column. The lack of visible soil development and clay, organic carbon, and CaCO₃ contents likely reflects a rapid development of this dune than at the KDS. An IRSL age estimate (UW2656) of 15,900 ± 2200 years ago was obtained on sands from 90 to 100 cmBS.

Rosa-Keystone Dune 26

Rosa-Keystone Dune 26 (RK Dune 26) is a star dune that is part of the complex dunes at the western edge of the RKD Field (Figures 1c and 5). Star-like dunes are superimposed over crescentic and parabolic forms, and the product of multiple dune forms fusing into each other as several wind directions persisted. This star-like dune has three arms that are between 160 and 240 m in length with orientations 40–50°NE, 180°S, and 310–320°NW, and winds from the east, south, and southwest influenced its shape.

Two trenches were placed within the upper 8 m of a 14-m high cut at the northeastern arm of RK Dune 26. Slump from the upper deposits and gravels from road construction between 200 and 600 cmBS and below 800 cmBS hampered our trenching here. Medium sands dominate dune deposits (Unit 1) between 90 and 800 cmBS. Fine and medium bedded sands (1–10 cm thick) are present between 600 and 800 cmBS and are horizontal lying (2–5° slopes) to the northwest. The upper 190 cm of Unit 1 consists of massive gray medium sands that grade into a 10 cm thick silty sand (Unit 2) deposit which caps dune sediments.

IRSL samples were taken at 155 and 750 cmBS from the Unit 1 sands, with their central age estimates showing a confusing stratigraphic dating reversal. The lowest sample (UW2658) at 750 cmBS dated to 8000 ± 1000 years ago, and the upper sample (UW2657) estimated at 12,500 ± 1600 years ago. While these estimates show stratigraphic dating incoherencies, they do however follow a pattern of late Glacial and early Holocene dune activation that was widespread throughout the RKD Field.

A pedocomplex, consisting of at least two 1–2 cm thick continuous black to dark reddish brown sandy loams (Entisols; Ab horizons), developed in Unit 1 sands between 700 and 714 cmBS. Patches of organics make up the sandy loams, and are relatively similar in coloration and general depth to the weak soils at the KDS and RK Dune 2 sections. However, unlike the KDS soils, macrofossils of charcoal were not found within the RK Dune 26 buried soils, and organic carbon, CaCO₃, and clay contents are relatively negligible throughout the stratigraphic column. This may reflect that the RK Dune 26 vegetation had less woody shrubs or trees growing on it at the time of the pedocomplex development, and/or a more continual aggradation of this landform and less stability than at the KDS and RK Dune 2. A recent immature forest soil (Inceptisol; O-A horizons; 10 cm thick) developed in sandy silts (loess; Unit 2) at the surface of the dune.

Archaeology at the KDS

Archaeology is present in Pedocomplex 5 at the KDS, and consists of large ungulate remains (wapiti (Cervus elaphus)) and waste flakes from stone tool manufacture and maintenance that were recovered from within and surrounding a hearth. The preservation of the bones is variable across the excavated area, with some fragments being structurally intact and recoverable, while others incredibly friable and non-recoverable. This difference in bone preservation is likely because of two factors: (1) cultural practices in cooking and (2) the attack of soil acids and bacteria. While much of the site has been lost because of road construction, the types of excavated artifacts and features and their sparse quantity and spatial distribution appear to represent a short-term camp at the KDS that dates between 13,430 and 13,230 cal. BP (11,460 ± 30 BP), a similar time frame for occupations at the Mead and Bachner sites (Potter et al., 2013; Wooller et al., 2012).

Discussion

Initial formation and activation history of the RKD field

The mTV landscape during the Donnelly glaciation and the beginning of the late Glacial was in a state of significant transformation involving a dichotomy between an arid environment and a substantial amount of glacio-fluvial aggradation (Guthrie, 1990a; Péwé and Reger, 1983). Our studies complement the research of Reger and Péwé (2002) and Reger et al. (2008) in understanding the inception of the SCF landscape as a result of Donnelly glaciation outwash infilling the lower valleys of the Uplands around 25,000–20,000 years ago. Outwash and succeeding alluvial deposition created the SCF plain and provided an immense quantity of sediment available for subsequent eolian mobilization.

The environment in the mTV and interior eastern Beringia was arid and a thin xerophytic (grasses, sedges, and tundra species) vegetation regime covered the landscape (Blinnikov et al., 2011; Gaglioti et al., 2011; Guthrie, 1990a; Zazula et al., 2006). The lack of expansive vegetation coverage and paucity of moisture (Abbott et al., 2000; Finkenbinder et al., 2014) would have maintained a substantial local sediment source shortly after the initial fluvial deposition into the SCF. IRSL ages on lower sand deposits in the southern portion of the RKD Field (RK Dune 13) indicate that at least low-lying crescentic dunes were forming by 18,000–14,000 years ago (Figure 4).

The prominent large crescentic ridges that dominate the topography across the entire RKD Field formed under a predominant southern wind regime and required a large supply of sand and strong winds to develop (Figure 1c and d; McKee, 1979). Crescentic ridge slipfaces are oriented toward 310°NW to 360°N.

Subsequent reworking of the RKD Field at the end of the late Glacial and into the early Holocene is evident by the series of radiocarbon ages on deeply buried soils and IRSL ages on dune sands, and played a major role in the expression of the topography and landscape. Dates from above and below deeply buried soils recognized at the far western and eastern edges of the field (at KDS and RK Dunes 2 and 26) show that over 6 m of sand accumulated after 12,900 and 11,400 cal. BP. Sand accumulation continued in several areas of the dune field until around 8800 cal. BP when silt became the dominant sediment accumulating by wind transport.

Parabolic and star-like forms are superimposed over the stoss slopes of larger crescentic ridges and blowouts are present in several areas (Figure 1d). Parabolic dunes are also present within interdunal areas, suggesting that these areas were viable sand sources during the periods of reactivation, and the morphologies of dune forms increase in complexity toward the field’s southern extent, likely a result of episodic reactivation. A unique dune form is present near the center of the field in the shape of relatively large and flat, oval-shaped dunes that have slipfaces 360° around the landforms. However, their steepest slipfaces appear along the northern and western edges with more gradual slopes at the southern and eastern edges. Kreig and Reger (1982) referred to these RKD Field forms as ‘rosette’ dunes, similar to forms described by Collins (1985) that developed at the crests of transverse dunes in the Tanana Lowlands.

Parabolic dunes show a wider variance in orientation between 300°NNW and 40°NE. Reversing crests along crescentic ridges and rosette dunes slipfaces are oriented toward 180°S to 130°SE. The slipface orientations of these landforms demonstrate the presence of multidirectional seasonal winds with shifts from southern to northern wind regimes that affected the dune field. Star-like and rosette dune forms also attest to a multidirectional wind regime during periods of reactivation of the southern portion of the field.

Our radiocarbon and IRSL age estimates that support late Glacial and early Holocene reactivation in this field come from parabolic and star-like forms. Given the superimposition of parabolic, star-like, and rosette forms over larger crescentic ridges, we suggest that expansive crescentic ridges accumulated during the Donnelly glaciation. Furthermore, we suggest that the reversing crests of crescentic ridges and parabolic, star-like, and rosette forms were the product of dune reactivation during periods of shifting wind patterns in the late Glacial and early Holocene once air circulation patterns began to fluctuate regionally and seasonally as the heights of the large ice sheets were reduced. Coastal air masses were no longer blocked and bifurcated by the glaciers (Anderson and Brubaker, 1994; Bartlein, 1997).

The predominant southern wind regime during the formation of crescentic ridges likely corresponds to katabatic winds originating from Donnelly-aged glaciers in the Alaska Range (Thorson and Bender, 1985), and agrees with paleo-wind models based on loess thickness and mineral composition (Muhs and Budhan, 2006; Péwé, 1965 [1977]). However, it contrasts to earlier suggestions based on interior Alaska dune orientations and the magnetic mineral fabric of loess deposits that northeast paleo-winds were dominant during Donnelly glaciation (Hopkins, 1982; Lagroix and Banerjee, 2002; Lea and Waythomas, 1990). Regional topography in interior Alaska may play an important role in the expression of dune morphology and simply reflect local wind directions (Dijkmans and Koster, 1990).

The shift in eolian sediment sizes in the early Holocene (~8000 cal. yr BP) marks a decrease in wind intensity as the glaciers receded further into the Alaska Range (Matmon et al., 2010; Young et al., 2009), and a change in sediment sources occurred as vegetation and soil coverage became thicker and wider spread (Bigelow, 1997; Dilley, 1998) and the more distant sources of the Tanana and Delta Rivers supplied eolian silt deposits. These shifts in sediment size and vegetative coverage appear to correspond with rises in regional lake levels that reached near modern heights, interpreted as reflecting increased effective moisture (Abbott et al., 2000; Barber and Finney, 2000; Finkenbinder et al., 2014).

Late Glacial and early Holocene local ecology in the SCF and mTV

Weakly developed soils (Entisols) formed shortly after loess began to accumulate, and became widespread across the region by 13,000 cal. BP (Dilley, 1998; Reuther, 2013). Pollen and macrofossil records from lacustrine cores reflect a transition from herbaceous tundra communities toward shrub tundra dominated vegetation around 14,000 cal. BP (Figure 2; Ager, 1975, 1982; Bigelow and Powers, 2001; Tinner et al., 2006). Birch (Betula sp.) and willow (Salix sp.) increase in the pollen and macrofossils records around 14,000–13,000 cal. BP from lacustrine settings (Bigelow, 1997; Tinner et al., 2006). Charcoal remains identified from soils in loessic deposits at the Mead and Swan Point sites indicate the presence of shrub birch and willow, and possibly Populus sp. in the SCF area around 14,000 and 13,000 cal. BP (Holmes, 2011; Potter et al., 2013).

Shrubs and deciduous trees were locally present by 12,000 cal. BP, as evident in the identification of charcoal in buried soils at Mead, Quartz Lake, and the RKD Field, including shrub birch, quaking aspen (Populus cf. tremuloides), and possibly alder (Alnus sp.; Reuther, 2013). The presence of Populus charcoal in the SCF is at least 1000 years earlier than the rise in Populus pollen in lake cores in the region (Bigelow and Powers, 2001; Tinner et al., 2006). A similar scenario of early Populus colonization occurred at Dune Lake in the Tanana Valley, west of the RKD Field (Bigelow, 1997). Populus likely colonized pockets of well-drained soils and south-facing slopes that were favorable for this taxa prior to its spread across the landscape into less favorable soil and topographic conditions (Ager, 1975; Bigelow and Powers, 2001).

Sand dune mobilization, episodic stability and weak soil development, and likely wild fires were widespread throughout the Shaw Creek Valley during this time, and possibly throughout the mTV (Tinner et al., 2006). At the KDS section, a series of buried very weakly developed and burnt soils date between 12,970 and 10,380 cal. BP. These soils are separated by coarser sands overlain by finer loamy sands and sandy loams that the soils have developed within. Remains of angiosperms (likely alder (Alnus sp.)), quaking aspen, and spruce (Picea sp.) were identified among seeds and charcoal fragments from these soils. Picea charcoal from the KDS section at 11,350–10,580 cal. BP marks a relatively early presence in interior Alaska, at the early range of their expansion exhibited by pollen records (Brubaker et al., 2005). Thus, the vegetation on these landforms appears diverse and within a general mid-successional cycle for northern forest cycles. Charcoal-lain soils from this time period are also evident in dunes in the upper Shaw Creek and at the Mead and Bachner sites (Gilbert, 2011; Reuther, 2013).

These events most likely signify a cycle of natural fires that burned vegetation and root mats that would have resulted in the loss of the vegetative cover and subsequent eolian deposition and periodic, localized reactivation of the RKD Field. Early-to-middle successional species (herbs, grasses, shrubs, deciduous trees) took hold in the finer eolian sediments and Entisols began to develop before the next fire.

Blowouts and parabolic dune forms transposed over crescentic ridges most likely occurred during late Glacial to early Holocene periods of reactivation. Blowout and parabolic dune forms are controlled by the presence of vegetation and/or moisture to anchor part of the sediment source from wind mobilization (Dijkmans and Koster, 1990; Pye and Tsoar, 2009). The portions of the sediment source that are not anchored down by vegetation and moisture are free to move. The increased frequency of these forms during this period logically coincides with an expansion of shrub tundra vegetation, a possible early deciduous forest, periodic soil development, and natural forest fires.

As noted above, late Glacial faunal remains include waterfowl (geese, ducks, swan), ptarmigan and grouse, shrews, pika, hare, marmot, ground squirrel, arctic fox, river otter, wolf, caribou, horse, mammoth, moose, wapiti, bison, sheep, and Salmonidae species (Holmes, 2011; Potter et al., 2013; Yesner, 2007). The ecological overlap of these species indicates a uniquely heterogeneous environment that supported diverse faunal communities during the late Glacial. These communities became less diverse into the Holocene as several species became regionally extinct (bison, horse, mammoth, wolf, and wapiti), other species’ ranges were fragmented and reduced (caribou, sheep, ground squirrel), and a few species’ ranges expanded (moose; Guthrie, 1990a, 1990b, 2006; Potter, 2008a, 2008b; Yesner, 1996).

The most recent radiocarbon ages for horses and mammoth in interior Alaska are from Cultural Zone 4 at Swan Point, the earliest evidence of human occupation in eastern Beringia (Guthrie, 2006; Holmes, 2011). The Swan Point horse remains date to 14,030–13,490 cal. BP and mammoth remains date between 14,100 and 13,490 cal. BP. The SCF area may have been a refugium of xerophytic vegetation (grasses and sagebrush) that could support the grazing of horse and mammoth within an increasingly expanding shrubby landscape. The co-existence of large mammal grazers (bison, wapiti, and sheep) and browsers (caribou, moose) suggests a unique combination of vegetative communities of xerophytic species and deciduous shrubs and trees in the SCF region into the early Holocene (Figure 2). The presence of ground squirrels in the mTV during this period also indicates the wider presence of herbaceous vegetation (graminoids, forbs, and sage) in this region; Zazula et al., 2006), and relatively year-round ice-free eolian deposits that allowed for burrowing and caching behaviors, a situation not seen evident today because of the presence of thick seasonal freezing active layers.

Boreal forest, paludification, and extreme landscape stability

The stability of the RKD Field heightened as forest soils became established, and regional landscape stability increased. Dark reddish brown sandy loams (Inceptisols; Bwb horizons) are evident at several locales throughout the mTV by 9000–8650 cal. BP, marking increased soil development in the Holocene as forests expanded (Dilley, 1998; Potter, 2005; Reuther, 2013). The stabilization of dune fields in the SCF, and likely much of the mTV region, occurred after 8000 cal. BP characterized by a transition to silt deposition and increased soil development. This period coincides with region-wide landscape stability likely a response to increased effective moisture and expansion of the boreal forest and paludification, which expanded around 6000 cal. BP (Ager, 1975, 1982; Anderson et al., 2004; Bigelow, 1997). The ranges of herbaceous tundra and deciduous forests and shrubs became more fragmented and restricted to well-drained sediments, such as on south-facing slopes and terrace edges, and areas of relatively high disturbance frequencies (e.g. active floodplains). In contrast, some northern Alaska sand dune fields remained active throughout the middle to late-Holocene as a response to shifts in local variability in effective moisture and vegetative coverage extent, such as the Kobuk Dunes (Mann et al., 2002).

Population declines of several small and large mammals, range restrictions, and local extinctions appear to coincide with the extreme landscape stabilization that occurred with the expansion of the boreal forest and peatlands and decline in the ranges of herbaceous tundra and deciduous forests and shrubs (Figure 2; Bigelow, 1997; Guthrie, 1990a; Jones and Yu, 2010). Bison and wapiti ranges overlapped in this region well into the Holocene until their extirpation (Potter, 2005). Wapiti presence in the faunal record begins to decline around 9000–8000 cal. BP, but is locally present at least until 2000 cal. BP (Guthrie, 2006; Potter, 2008a, 2008c). Bison and sheep occur in lower numbers in the middle Holocene faunal assemblages, likely as their ranges began to fractionate in interior Alaska (Guthrie, 1990a, 2006; Holmes and Bacon, 1982; Potter, 2008a, 2008c). Bison and wapiti habitat in the middle to late-Holocene was relegated to windswept areas and floodplains, where early successional plant species (grasses, sagebrush, and sedges) would have been prevalent and snow depth minimal (Guthrie, 1990a). The distributions of small mammals, such as ground squirrels, became more restricted into the early Holocene. Ground squirrels disappeared from the lowlands of the mTV, in the early to middle Holocene, probably in response to the development of thicker boreal forest mats and wider spread seasonally frozen ground in loess and eolian sand deposits.

The transition from deciduous shrub tundra and tree communities toward a coniferous forest in the mTV appears to have reduced faunal diversity in the Holocene (Potter, 2008b, 2008c). Frequent disturbances, such as natural fires and eolian aggradation, may have been a key for sustaining deciduous vegetation communities into the early Holocene in the mTV, and thus the region’s attractiveness to the diverse animals whose remains we find in archaeological sites (Reuther, 2013).

Human response to late Glacial and Holocene environmental change in the mTV

While some archaeologists have argued for a succession of possibly unrelated cultural groupings in the region (e.g. Nenana/Chindadn and Denali complexes) (Hamilton and Goebel, 1999), recent analyses of human adaptive strategies (i.e. integrated subsistence economy, settlement, and technological strategies) in the region suggest responses to some of the ecological changes identified in this study (Mason et al., 2001; Potter, 2008a, 2008b; Potter et al., 2013; Yesner, 2007). Technological trends suggest broad similarity in artifact classes (e.g. microblade-composite points) throughout the period under study (14,000–6000 cal. BP), but changes in artifact types (e.g. Chindadn points, biconvex bifacial knives) around 12,000 cal. BP, at the end of the Young Dryas period (Holmes, 2001; Potter, 2011; Powers and Hoffecker, 1989). Habitat use suggests a broader range of resources and seasons of habitation (spring through fall) in lowland areas, including SCF, with narrower diet breadth evident in upland areas, such as the northern foothills of the Alaska Range (Guthrie, 1983), although taphonomy may be a factor in small mammal, bird, and fish preservation (Potter, 2008a). The main changes in overall adaptive strategies during this period are the incorporation of lower ranked prey (small mammals, birds, fish) during the later part of the Younger Dryas (~12,200–11,500 cal. BP), and a possible reduction in regional populations during the earlier part of the Younger Dryas (~12,800–12,200 cal. BP) (Bever, 2006; Potter, 2008c). The hunting of higher ranked prey species (bison, wapiti) continued throughout the period 14,000–6000 cal. BP, suggesting a resilient system that could accommodate considerable local and regional climate and vegetation regime changes.

More substantial cultural change is evident at ~6000 cal. BP, when earlier cultures, identified as Denali complex or American Paleoarctic tradition, are replaced by a new cultural construct, the Northern Archaic tradition, characterized by new artifact types and classes (side-notched bifaces, side-notched cobbles, the latter perhaps associated with fishing) (Anderson, 1968; Dixon, 1985; Mason and Bigelow, 2008). Bison and wapiti are replaced by increased focus on caribou as the primary large ungulate prey (Potter, 2008a). Settlement systems also change after 6000 cal. BP, with increasing use of upland areas (e.g. Upper Nenana and Upper Susitna basins, Tangle Lakes), perhaps associated with increased caribou exploitation (Potter, 2008b).

Given the relatively limited resources available to early Alaskan hunter-gatherers, the cultural changes at ~12,000 and ~6000 cal. BP appear closely linked with the environmental and climate change described above, specifically the heightened landscape stability that correlates to the proliferation of coniferous forests and paludification throughout the region (Bigelow, 1997), thought to be a response to warming temperatures and increases in effective moisture. Habitat reduction, decreases in below-ground biomass, introduction of toxic plants, and increases in moisture and paludification have been put forward as some of the primary processes leading to changes in interior Alaskan mammalian diversity from the LGM to the Holocene (Guthrie, 1982, 1984, 1990a, 2001). In turn, the shifts in human economy (primary focuses from wapiti and bison, to caribou in the uplands) appear to be responses to the changes in ranges and diversity of mammals in the mTV in the middle Holocene.

Conclusion

These studies of the RKD Field have allowed us to refine our understanding of the environmental history and evolution of landscapes in the SCF, and the broader mTV, an area significant to the human colonization and occupation of eastern Beringia and North America. The expressions of the regional lowland topography generally reflect the climate of the Donnelly glaciation and the late Glacial and early Holocene periods (25,000–6000 years ago). The initial formation of sand sheets and dunes in this area occurred around 18,000–15,000 years ago shortly after the formation of the SCF plain, but extensive dune reactivation in the RKD Field occurred between 12,900 and 8800 cal. BP. This coincides with the reactivation and development of other dunes and sand sheets throughout the lowlands of the Tanana Valley (Fernald, 1965; Forgacs et al., 2013; Reuther, 2013), and loess deposition on higher terraces and ridges across the region that were farther away from sediment sources (Dilley, 1998; Péwé and Reger, 1983; Reger and Péwé, 2002; Reuther, 2013).

The dynamic nature of the mTV landscape between 14,500 and 8000 cal. BP with high rates of eolian deposition, periodic soil development, and a rise in fire regimes fostered a diverse ecosystem of early to mid-successional vegetation and faunal communities. Many of the large mammal species, such as bison and wapiti, which were important to subsistence systems of late Glacial and early Holocene human occupations, went regionally extinct or their ranges shifted later in the Holocene. As the middle Holocene (6000–1000 cal. BP) approached, landscapes became increasingly stable with the expansion of the boreal forest and eolian deposition drastically decreased throughout the mTV (Bigelow, 1997; Reuther, 2013). The highly productive late Glacial and early Holocene early-to-middle successional vegetative communities became progressively partitioned and primarily relegated to active floodplains during the middle Holocene, as the expanse and impact of eolian disturbance regimes on vegetation cycles lessened and broader ranges of mTV landscapes become more stable (Guthrie, 1990a, 2001).

When viewing the overall expanse of the RKD Field, our sampling strategy was relatively small and relegated to easy access road cuts. Much more effort needs to be expended in sampling to refine models for the evolution of landscapes, such as the RKD Field and SCF, and the climatic conditions and local environmental histories under which they develop, especially given that larger dune fields and sand sheets can experience localized variability in their activation and stabilization histories (Matthews and Seppälä, 2014). This research on the RKD Field reminds us that prior to the boreal forest stronghold and extreme stability over landscapes across the northern subarctic regions, vast areas were dynamically changing with ecosystems and flora and fauna patch overlap that are not evident today in northern subarctic environments. This situation was likely critical to the colonization of and sustained hunter-gatherer occupation in the region. Human adaptive strategies in the subarctic, such as those described above, appear intrinsically linked to local ecological systems that may shift in response to changes in environment and regional climate.

Footnotes

Acknowledgements

The authors would like to thank Jason Rogers for editing an earlier version of this manuscript and providing insights into his work at the KDS; Mike Yarborough, Sarah Meitl, and Aubrey Morrison for Cultural Resources Consultants for providing information and discussions on their current work in the SCF; Richard Vanderhoek of the Office of History and Archaeology for discussions on the work he conducted in the 1990s in the SCF. Claire Alix, of the Université Paris 1 Pathéon-Sorbonne, and Owen Davis of the University of Arizona for wood identifications, and Aren Gunderson, Mammals Department at the University of Alaska Museum of the North, for facilitating large mammal skeletal comparative specimens; Vance Holliday, University of Arizona, for providing laboratory support for the sediment analyses and discussions on the manuscript. They thank two anonymous reviewers whose comments helped to improve the manuscript.

Funding

Funding was provided through a Dissertation Improvement Grant (Award#1107631) to JDR through the National Science Foundation’s Arctic Social Sciences Program, Office of Polar Programs. Northern Land Use Research supplied funding for some of the radiocarbon dates and field supplies.

References

Abbott

Finney

Edwards

. (2000) Lake-level reconstructions and paleohydrology of Birch Lake, central Alaska, based on seismic reflection profiles and core transects. Quaternary Research 53: 154–166.

Ager

(1975) Late Quaternary Environmental History of the Tanana Valley (Alaska. Institute of Polar Studies Report No. 84). Columbus, OH: Ohio State University Research Foundation.

Ager

(1982) Pollen studies of Quaternary-age lake sediments in the Tanana Valley, Alaska. In: Coonrad

(ed.) The United States Geological Survey in Alaska: Accomplishments During 1980 (Geological Survey Circular 844). Alexandria, VA: Department of the Interior, U.S. Geological Survey, pp. 64–67.

Alaska Boreal Forest Council (2002) Assessments of Shaw Creek watershed in the Tanana River Basin, Alaska. Available at: http://www.tanana-watershed.org/shaw_creek/notes1/Open-House-Material/ShawCK-Flyer.pdf (accessed 28 October 2014).

Anderson

(1968) Early notched point and related assemblages in the Western American Arctic. In: Proceedings of the paper presented at the 67th annual meeting of the American Anthropological Association, Seattle, WA, November 21–24, 1968.

Anderson

Brubaker

(1994) Vegetation history of Northcentral Alaska: A mapped summary of late-Quaternary pollen data. Quaternary Science Reviews 13: 71–92.

Anderson

Edwards

Brubaker

(2004) Results and paleoclimate implications of 35 years of paleoecological research in Alaska. In: Gillespie

Porter

Atwater

(eds) The Quaternary Period in the United States: Developments in Quaternary Science, vol. 1. Amsterdam: Elsevier, pp. 427–440.

Barber

Finney

(2000) Late Quaternary paleoclimatic reconstructions for interior Alaska based on paleolake-level data and hydrologic models. Journal of Paleolimnology 24: 29–41.

Bartlein

(1997) Paleoclimatic variations in Beringia: Large-scale controls and regional responses in general circulation model simulations. In: Edwards

Sher

Guthrie

(eds) Terrestrial paleoenvironmental studies in Beringia (Proceedings of a joint Russian-American workshop, Fairbanks, Alaska, June 1991). Fairbanks: Alaska Quaternary Center, pp. 43–47.

10.

Bever

(2006) Too little, too late? The radiocarbon chronology of Alaska and the peopling of the New World. American Antiquity 71(4): 595–620.

11.

Bigelow

(1997) Late Quaternary vegetation and lake level changes in central Alaska. PhD Thesis, University of Alaska Fairbanks.

12.

Bigelow

Powers

(2001) Climate, vegetation, and archaeology 14,000–9000 cal yr B.P. in central Alaska. Arctic Anthropology 38(2): 171–195.

13.

Black

(1951) Eolian deposits of Alaska. Arctic 4(2): 89–111.

14.

Blinnikov

Gaglioti

Walker

. (2011) Pleistocene graminoid-dominated ecosystems in the Arctic. Quaternary Science Reviews 30: 2906–2929.

15.

Brubaker

Anderson

Edwards

. (2005) Beringia as a glacial refugium for boreal trees and shrubs: New perspectives from mapped pollen data. Journal of Biogeography 32: 833–848.

16.

Butzer

(1982) Archaeology as Human Ecology. Cambridge: Cambridge University Press.

17.

Collins

(1985) Map Showing a Vegetated Dune Field in Central Alaska (U.S. Geological Survey Map MF-1708). Washington, DC: Department of the Interior, U.S. Geological Survey, U.S. Government Printing Office.

18.

Dijkmans

JWA

Koster

(1990) Morphological development of dunes in a subarctic environment, central Kobuk Valley, Northwestern Alaska. Geografiska Annaler: Series A, Physical Geography 72(1): 93–109.

19.

Dilley

(1998) Late Quaternary loess stratigraphy, soils, and environments of the Shaw Creek Flats Paleoindian sites, Tanana Valley, Alaska. PhD Thesis, University of Arizona.

20.

Dixon

(1985) Cultural chronology of central interior Alaska. Arctic Anthropology 22(1): 47–66.

21.

Durtsche

Hobgood

(1990) Distribution, Movements and Seasonal Use Areas of Caribou in the White Mountains National Recreation Area, Alaska, 1982–1988 (BLM-Alaska Open File Report 29). Anchorage, AK: Bureau of Land Management, U.S. Department of the Interior.

22.

Fernald

(1965) Recent history of the Upper Tanana River Lowland, Alaska. In: Chase

(ed.) Geological Survey Research 1965. Geological Survey Professional Paper 525-C. Washington, DC: United States Government Printing Office, pp. C124–C127.

23.

Finkenbinder

Abbott

Edwards

. (2014) A 31,000 year record of paleoenvironmental and lake-level change from Harding Lake, Alaska, USA. Quaternary Science Reviews 87: 98–113.

24.

Forgacs

Johnson

Rittenour

. (2013) IRSL dating of a vegetated dune field in central Alaska. Geological Society of America Abstracts with Programs 47(7): 579.

25.

Gaglioti

Barnes

Zazula

. (2011) Late Pleistocene paleoecology of arctic ground squirrel (Urocitellus parryii) caches from interior Alaska’s Mammoth Steppe ecosystem, USA. Quaternary Research 76: 373–382.

26.

Galloway

Huebner

Lipkin

. (1990) Early Holocene calcretes from the subarctic active Nogahabara sand dune field, Northern Alaska. In: Bradley

Ford

(eds) Geologic Studies in Alaska by the U.S. Geological Survey, 1990. Washington, DC: United States Government Printing Office, pp. 100–111.

27.

Gilbert

(2011) Micromorphology, site spatial variation and patterning, and climate change at the Mead Site (XBD-071): A multi-component archaeological site in interior Alaska. MA thesis, University of Alaska Fairbanks.

28.

Goebel

Powers

Bigelow

(1991) The Nenana complex of Alaska and Clovis Origins. In: Bonnichsen

Turnmire

(eds) Clovis: Origins and Adaptations. Corvallis, OR: Center for the Study of the First Americans, Oregon State University, pp. 49–79.

29.

Graf

Bigelow

(2011) Human responses to climate during the Younger Dryas Chronozone in central Alaska. Quaternary International 242: 434–451.

30.

Guthrie

(1982) Mammals of the Mammoth Steppe as paleoenvironmental indicators. In: Hopkins

Matthews

Jr Schweger

. (eds) Paleoecology of Beringia. New York: Academic Press, pp. 307–326.

31.

Guthrie

(1983) Paleoecology of the site and its implications for early hunters. In: Powers

Guthrie

Hoffecker

(eds) Archeology and Paleoecology of a Late Pleistocene Alaskan Hunting Camp (Unpublished report submitted to the National Park Service, Contract CX-9000-7-0047 under the Federal Antiquities Act Permit 77-AK-098). Fairbanks, AK: University of Alaska Fairbanks, pp. 209–287.

32.

Guthrie

(1984) Mosaics, allelochemics, and nutrients: An ecological theory of late Pleistocene megafaunal extinctions. In: Martin

Klein

(eds) Quaternary Extinctions: A Prehistoric Revolution. Tucson, AZ: University of Arizona Press, pp. 259–298.

33.

Guthrie

(1990a) Frozen Fauna of the Mammoth Steppe. Chicago, IL: University of Chicago Press.

34.

Guthrie

(1990b) New dates on Alaska Quaternary moose, Cervalces-Alces – Archaeological, evolutionary, and ecological implications. Current Research in the Pleistocene 7: 111–112.

35.

Guthrie

(2001) Origin and causes of the Mammoth Steppe: A story of cloud cover, woolly mammoth tooth pits, buckles, and inside-out Beringia. Quaternary Science Reviews 20: 549–574.

36.

Guthrie

(2006) New radiocarbon dates link climatic change with human colonization and Pleistocene extinctions. Nature 441: 207–209.

37.

Haas

Holliday

Stuckenrath

(1986) Dating of Holocene stratigraphy with soluble and insoluble organic fractions at the Lubbock Lake Archaeological Site, Texas: An ideal case study. Radiocarbon 28: 473–485.

38.

Hamilton

(1973) Late Quaternary Glacial History, Delta-Johnson Rivers Region, Northeastern Alaska Range (Unpublished manuscript). Fairbanks, AK: Department of Geology, University of Alaska Fairbanks.

39.

Hamilton

Goebel

(1999) Late Pleistocene peopling of Alaska. In: Bonnichsen

Turnmire

(eds) Ice Age Peoples of North America: Environments, Origins, and Adaptations. College Station, TX: Texas A&M University Press, pp. 156–199.

40.

Holmes

(2001) Tanana River Valley archaeology circa 14,000 to 9,000 B.P. Arctic Anthropology 38(2): 154–170.

41.

Holmes

(2011) The Beringian and transitional periods in Alaska: Technology of the East Beringian tradition as viewed from Swan Point. In: Goebel

Buvit

(eds) From the Yensei to the Yukon: Interpreting Lithic Assemblage Variability in Late Pleistocene/Early Holocene Beringia. College Station, TX: Texas A&M University Press, pp. 179–191.

42.

Holmes

Bacon

(1982) Holocene bison in central Alaska: A possible explanation for technological conservatism. In: Paper presented at the 9th annual meeting of the Alaska Anthropological Association, Fairbanks, AK, April 2–3, 1982.

43.

Hopkins

(1982) Aspects of the paleogeography of Beringia during the Late Pleistocene. In: Hopkins

Matthews

Jr Schweger

. (eds) Paleoecology of Beringia. New York: Academic Press, pp. 3–28.

44.

Hubbard

Koehler

Combellick

(2011) High-Resolution Lidar Data for Alaska Infrastructure Corridors (Alaska Division of Geological and Geophysical Surveys Raw Data File 2011-3). Fairbanks, AK: Alaska Division of Geological and Geophysical Surveys, Department of Natural Resources.

45.

Jones

(2010) Rapid deglacial and early Holocene expansion of peatlands in Alaska. Proceedings of the National Academy of Sciences 107(16): 7347–7352.

46.

Jordan

Yarborough

(n.d.) Geomorphology and Archaeological Site Potential of the Shaw Creek Hillside Access and Power Line Corridors (Report). Anchorage, AK: Cultural Resource Consultants.

47.

Kreig

Reger

(1982) Air-Photo Analysis and Summary of Landform Soil Properties along the Route of the Trans-Alaska Pipeline System. Fairbanks, AK: Alaska Division of Geologic and Geophysical Surveys.

48.

Lagroix

Banerjee

(2002) Paleowind directions from the magnetic fabric of loess profiles in central Alaska. Earth and Planetary Science Letters 195: 99–112.

49.

Lea

(1996) Vertebrate tracks in Pleistocene eolian sand-sheet deposits of Alaska. Quaternary Research 45: 226–240.

50.

Lea

Waythomas

(1990) Late-Pleistocene eolian sand sheets in Alaska. Quaternary Research 34: 269–281.

51.

Lynch

Hollis

(2004) Climatic and landscape controls of the boreal forest fire regime: Holocene records from Alaska. Journal of Ecology 92(3): 477–489.

52.

McKee

(1979) Introduction to a study of global sand seas. In: McKee

(ed.) A Study of Global Sand Seas (Geological Survey Professional Paper 1052). Washington, DC: U.S. Government Printing Office, pp. 1–19.

53.

Mann

Heiser

Finney

(2002) Holocene history of the Great Kobuk Sand Dunes, northwestern Alaska. Quaternary Science Reviews 21: 709–731.

54.

Martin

Johnson

(1995) Variation in radiocarbon ages of soil organic matter from late Quaternary buried soils. Quaternary Research 43: 232–237.

55.

Mason

Bigelow

(2008) The crucible of early to mid-Holocene climate in northern Alaska: Does Northern Archaic represent the people of the spreading forest? Arctic Anthropology 45(2): 39–70.

56.

Mason

Bowers

Hopkins

(2001) The early Holocene Milankovitch Thermal Maximum and humans: Adverse conditions for the Denali Complex of Eastern Beringia. Quaternary Science Reviews 20: 525–548.

57.

Matmon

Briner

Carver

. (2010) Moraine chronosequence of the Donnelly Dome region, Alaska. Quaternary Research 74: 63–72.

58.

Matthews

Seppälä

(2014) Holocene environmental change in subarctic aeolian dune fields: The chronology of sand dune re-activation events in relation to forest fires, palaeosol development and climatic variations in Finnish Lapland. Holocene 24(2): 149–164.

59.

Muhs

Budhan

(2006) Geochemical evidence for the origin of late Quaternary loess in central Alaska. Canadian Journal of Earth Sciences 43: 323–337.

60.

Olley

Pietsch

Roberts

(2004) Optical dating of Holocene sediments from a variety of geomorphic settings using single grains of quartz. Geomorphology 60: 337–358.

61.

Péwé

1965 [1977]) Middle Tanana River Valley. In: Péwé

Ferrains

Nichols

. (eds) Guide to the Quaternary Geology: Central and South-Central Alaska. College, AK: Division of Geological and Geophysical Surveys, Department of Natural Resources, pp. 36–54.

62.

Péwé

(1975) Quaternary Geology of Alaska (U.S. Geological Survey Professional Paper 835). Washington, DC: United States Government Printing Office.

63.

Péwé

Reger

(1983) Richardson and Glenn Highways, Alaska: Guidebook to Permafrost and Quaternary Geology. Fairbanks, AK: Division of Geological and Geophysical Surveys, State of Alaska Department of Natural Resources.

64.

Potter

(2005) Site structure and organization in Central Alaska: Archaeological investigations at Gerstle River. PhD Thesis, University of Alaska Fairbanks.

65.

Potter

(2008a) A first approximation of Holocene inter-assemblage variability in central Alaska. Arctic Anthropology 45(2): 89–113.

66.

Potter

(2008b) Exploratory models of intersite variability in mid to late-Holocene central Alaska. Arctic 61(4): 407–425.

67.

Potter

(2008c) Radiocarbon chronology of central Alaska: Technological continuity and economic change. Radiocarbon 50(2): 181–204.

68.

Potter

(2011) Late Pleistocene and early Holocene assemblage variability in central Alaska. In: Goebel

Buvit

(eds) From the Yensei to the Yukon: Interpreting lithic assemblages variability in late Pleistocene/early Holocene Beringia. College Station, TX: Texas A&M University Press, pp. 215–233.

69.

Potter

Holmes

Yesner

(2013) Technology and economy among the earliest prehistoric foragers in interior eastern Beringia. In: Graf

Ketron

Waters

(eds) Paleoamerican Odyssey College Station: Center for the Study of the First Americans. College Station, TX: Texas A&M University, pp. 81–104.

70.

Powers

Hoffecker

(1989) Late Pleistocene settlement in the Nenana Valley, Central Alaska. American Antiquity 54(2): 263–287.

71.

Pye

Tsoar

(2009) Aeolian Sand and Sand Dunes. Berlin: Springer.

72.

Reger

Péwé

(2002) Geologic map of the Big Delta A-4 quadrangle, Alaska (Alaska Division of Geological & Geophysical Surveys Report of Investigation 2002-2, 1 sheet, scale 1:63,360). Fairbanks, AK: Division of Geological and Geophysical Surveys, State of Alaska Department of Natural Resources.

73.

Reger

Stevens

DSP

Solie

(2008) Surficial Geology of the Alaska Highway Corridor, Delta Junction to Dot Lake, Alaska. Fairbanks, AK: Division of Geological and Geophysical Surveys.

74.

Reimer

Bard

Bayliss

. (2013) IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 55(4): 1869–1887.

75.

Reuther

(2013) Late Glacial and early Holocene geoarchaeology and terrestrial paleoecology in the lowlands of the middle Tanana Valley, subarctic Alaska. PhD Thesis, University of Arizona.

76.

Schoephorster

(1973) Soil Survey of Salcha-Big Delta Area, Alaska. Washington, DC: Soil Conservation Service, United States Department of Agriculture.

77.

Stuiver

Reimer

(2015) CALIB v7.1 (WWW program and documentation). Available at: http://calib.qub.ac.uk/calib/ (accessed November 2015).

78.

Ten Brink

Waythomas

(1985) Late Wisconsin glacial chronology of the North-central Alaska Range: A regional synthesis and its implications for early human settlements. National Geographic Research Reports 1978(19): 15–32.

79.

Thorson

Bender

(1985) Eolian deflation by ancient katabatic winds: A late Quaternary example from the North Alaska Range. GSA Bulletin 96: 702–709.

80.

Tinner

Beer

. (2006) Postglacial vegetational and fire history: Pollen, plant macrofossil, and charcoal records from two Alaskan lakes. Vegetation History and Archaeobotany 15: 279–293.

81.

Wahrhaftig

(1965) Physiographic Divisions of Alaska. Washington, DC: United States Government Printing Office.

82.

Ward

Wilson

(1978) Procedures for comparing and combining radiocarbon age determinations: A critique. Archaeometry 20: 19–31.

83.

Weber

(1971) Preliminary Engineering Geologic Maps of the Proposed Trans-Alaska Pipeline Route, Fairbanks and Big Delta Quadrangles (U.S. Geological Survey Open-File Report 71-317). Washington, DC: U.S. Geological Survey, Department of the Interior, U.S. Government Printing Office, 2 sheets, scale. 1: 250,000.

84.

Weber

Foster

Keith

TEC

. (1978) Preliminary Geologic Map of the Big Delta Quadrangle, Alaska (U.S. Geological Survey Open-File Report 78-529A). Washington, DC: U.S. Geological Survey, Department of the Interior, U.S. Government Printing Office, 1 sheet, scale 1:250,000.

85.

Wintle

(2008) Luminescence dating: Where it has been and where it is going. Boreas 37: 471–482.

86.

Wolfe

Nickling

(1997) Sensitivity of Eolian Processes to Climate Change in Canada: Impact of Global Change on Geological Processes in Canada (Geological Survey of Canada Bulletin 421). Ottawa: Geological Survey of Canada.

87.

Wolfe

Bond

Lamothe

(2011) Dune stabilization in central and southern Yukon in relation to early environmental change, northwestern North America. Quaternary Science Reviews 20: 324–334.

88.

Wolfe

Gillis

Robertson

(2009) Late Quaternary Eolian Deposits of Northern North America: Age and Extent. Geological Survey of Canada. Open File 6006. Ottawa, ON: Geological Survey of Canada.

89.

Wolfe

Hugenholtz

Evans

. (2007) Potential aboriginal occupation-induced dune activity, Elbow Sand Hills, northern Great Plains, Canada. Great Plains Research 17: 173–192.

90.

Wooller

Kurek

Gaglioti

. (2012) An ~11,200 cal yr BP paleolimnological record from Quartz Lake, Alaska. Journal of Paleolimnology 48: 83–99.

91.

Yesner

(1996) Human adaptation at the Pleistocene-Holocene boundary (circa 13,000 to 8,000 BP) in eastern Beringia. In: Straus

Eriksen

Erlandson

. (eds) Humans at the End of the Ice Age: The Archaeology of the Pleistocene-Holocene Transition. New York: Plenum Press, pp. 255–276.

92.

Yesner

(2007) Faunal extinction, hunter-gatherer foraging strategies, and subsistence diversity among eastern Beringian Paleoindians. In: Walker

Driskell

(eds) Foragers of the Terminal Pleistocene in North America. Lincoln, NE: University of Nebraska Press, pp. 15–31.

93.

Young

Briner

Kaufman

(2009) Late Pleistocene and Holocene glaciation of the Fish Lake Valley, northeastern Alaska Range, Alaska. Journal of Quaternary Science 24(7): 677–689.

94.

Zazula

Schweger

Beaudoin

. (2006) Macrofossil and pollen evidence for full-glacial steppe with an ecological mosaic along the Bluefish River, eastern Beringia. Quaternary International 142–143: 2–19.