Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe

Quaternary International xxx (2016) 1e17

Contents lists available at ScienceDirect

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

Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe Teresa Wriston a, *, Gary Haynes b a b

Desert Research Institute, Reno, NV 89512, USA University of Nevada, Reno, NV 89557, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 13 December 2015 Received in revised form 5 November 2016 Accepted 9 November 2016 Available online xxx

The transition from foraging to food production in interior southern Africa is still not well understood, despite being central to hypothesized migration and diffusion routes connecting the earliest agropastoralist sites in central Africa with the southern African subcontinent. Hwange National Park (14,650 sq. km), Zimbabwe is located along these proposed routes and has both Late Stone Age and Early Iron Age sites that can address this transition. To better understand this change in lifeways and any environmental and/or topographic factors that influenced it, geoarchaeological investigations characterized the soils, sediments, and landforms in six different areas of the park. Forty-eight stratigraphic sections varying from two to four m thick were described, six of which are detailed here. These stratigraphic sections include dated material that allows correlation between basins of different substrates and comparisons of carbon isotopic signatures to reveal whether conditions were locally favorable to grasses (hot/dry) or woodlands (cool/wet) through time. Examined river cuts and augered samples show that some river basins have sediment textures, soil types, and topographic relief that should have been attractive to farmers. However, radiocarbon ages reveal variable potential for finding remains of any early farmers from basin-to-basin, with some bedrock-constrained valleys having no preserved deposits of appropriate age. The repeated scouring of these river basins testifies to sometimes volatile conditions that would have challenged sustained food production. In fact, in the study area landscape instability and variable environmental conditions were commonplace during the past three millennia. Drought between ca. 2.8 and 2.6 ka may have lessened the risk of disease vectors for the earliest agropastoralists that arrived during more mesic conditions between 1.9 and 1.65 ka. At this time basins began to accumulate fill and soil development suggests landscape stability even in headland basins. Changes in soil carbon isotope values imply relatively hot/ dry conditions prevailed around 0.73 ka and 0.41 ka and cool/wet conditions around 0.53 ka, but much information has been lost due to the numerous erosional events that wiped away swaths of sediments during intervening droughts. The landscape in Hwange National Park continues to change rapidly as reflected by several meters of river downcutting in response to recent man-made disturbance and/or drought. This downcutting both reveals and endangers the few remaining archaeological sites that are critical to understanding the first food producers in the region. © 2016 Elsevier Ltd and INQUA. All rights reserved.

Keywords: Agropastoralism Late Stone Age Early Iron Age Geoarchaeological land use modeling Carbon isotopes African archaeology

1. Introduction

* Corresponding author. E-mail address: [email protected] (T. Wriston).

Hwange National Park is within or near several hypothesized southward expansion routes of agropastoralists from core areas in central Africa (e.g., see Bousman, 1998; Elphick, 1977; Huffman, 2007; Mitchell, 2002; Sadr, 2008). Farming and pastoralism

http://dx.doi.org/10.1016/j.quaint.2016.11.010 1040-6182/© 2016 Elsevier Ltd and INQUA. All rights reserved.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

2

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

(known together as agropastoralism), are considered hallmarks of the Early Iron Age and were relatively late adaptations in southern Africa when compared to the northern extent of the continent, where they occurred as early as 10 ka (Van Zinderen Bakker, 1976). As Gifford-Gonzalez (2000) suggests, this delay was partly due to the disease vectors found in the wooded area, including tsetse fly and mosquitoes that respectively carry sleeping sickness (nagana in animals) and malaria. These vectors often foiled European colonists' plans for expansion (cf. Chapman and Tabler, 1971; Holub, 1975), and continue to be a threat to the health of both man and beast when, after a drought, wet conditions and woodland expansion occur (cf. Endfield et al., 2009). Therefore, the initial adoption of agropastoralism in our study area, situated near the recent boundary of tsetse fly habitat (cf. Cecchi et al., 2008), would have had to take place within an environmental window that had sufficient precipitation for crops to survive, but was sufficiently dry to minimize woodland cover and suppress the risk from disease vectors. This study’s aim was to provide baseline environmental and landscape information for northwestern Zimbabwe between ca. 3.5 and 0.5 ka, a period that encompasses the earliest agropastoralism in the region. How and why plant cultivation and/or pastoralism spread is complex and varies by locality, but minimally, the environmental conditions had to be able to support agriculture and domestic animals. Stratigraphic analyses can reveal environmental conditions through soils (Catt, 1990), which signal landscape stability during mesic conditions, and unconformities, which signal instability and drought (cf. Eze and Meadows, 2015; Heinrich and Moldenhauer, 2002). Any landscape level changes can be correlated between basins using these diagnostic strata to establish a soil chronosequence (cf. Botha and Porat, 2007). This process also reveals the areas of Hwange National Park that are most likely to contain sediments and soils that preserve the remains of early agropastoralist settlements. 2. Regional setting 2.1. Environmental background 2.1.1. Geology and soils Hwange National Park straddles the watershed between the Zambezi Basin to the north and the Makgadikgadi pans of Botswana's Kalahari Desert to the south. The northern portion of the park has undulating topography with drainages cutting into bedrock ridges of basalt, sandstone, granite, gneisses, and various other materials (Sithole, 1994; Fig. 1). The central and southern portion of the park is covered with red Kalahari sand sheets separated by long linear dunes. These dunes were formed during the Pleistocene and have been sporadically reworked during Holocene dry periods (Stokes et al., 1998; O'Connor and Thomas, 1999). This parent material influences the success or failure of agricultural crops. Basic volcanic rocks, such as basalt, contain the full range of the earth's crustal elements and weather to clay, allowing good soil development and high plant productivity due to the abundance of available nutrients (Bell, 1982; Limbrey, 1975); but within Hwange National Park, basalt areas have developed only thin lithosols according to Nyamapfene (1991). Conversely, granite rocks and quartz sand, comprised of relatively coarse, weatherresistant minerals, have less variety and abundance of essential minerals and nutrients required by plants (Bell, 1982; Limbrey, 1975; Norman et al., 1984), as in the regosols of the Kalahari sands (Nyamapfene, 1991). However, fersiallitic soils (Nyamapfene, 1991; a Zimbabwe-specific soil categorization similar to alfisols) have been identified in the paragneisses and granites of the

northeastern study area and are good for agriculture due to their well-developed B-horizon/soil structure, high field capacity, and good drainage. In addition to the importance of the type of bedrock from which sediment originates, erosion redistributes it and associated nutrients through the drainage system, enhancing the fertility of valley floors and river floodplains while depleting the uplands (Limbrey, 1975). Given the bedrock lithology and soil types within the Park, we expect that prehistoric farmers would be attracted to basins with fersiallitic/alfisol soils and to floodplains that drain areas with basic volcanic rocks. 2.1.2. Precipitation patterns Northwestern Zimbabwe is near the southern extent of the Intertropical Convergence Zone (ITCZ) that impacts rainfall amount and distribution. Precipitation falls November to March while the remainder of the year is dry. Any surface waters accumulated during the wet period dwindle during the dry period, eventually leaving few, if any, sources of water until rains come again. This wet-and-dry season distribution of rainfall creates annual short term drought signatures within the geomorphic record (cf. Thomas and Burrough, 2012). Long droughts are frequent, and when drought extends over several years, significant geomorphic response, such as deflation and scour of slopes and alluvial basins, is likely. The mean annual rainfall in the park is around 620 mm/ year (Rogers, 1993), but annual values and distribution are erratic and unpredictable (Nyamapfene, 1991). For instance, between 1918 and 1990, annual rainfall at Main Camp ranged from 335.6 mm to 1159.8 mm (Rogers, 1993). The unimodal distribution of annual precipitation within southern Africa creates challenging conditions for farmers, who have learned to use sediment texture and topographic position to their advantage. For example, Scudder (1971, 1976) reports that the Gwembe Tonga people of the middle Zambezi River Basin plant wet season crops that grow during the rains in well-drained, sandy soils high on the landscape, while water-tolerant crops are planted along river ways. With the onset of the dry season, loamy soils on alluvial terraces are again planted in hopes that the water absorbed by the clay portion of the loam slowly releases to the crops, allowing their maturation under dry conditions. Gardens on alluvial terraces are therefore most valued because they could be cropped twice per year while local topographic relief that allows flexible field placement above the river is also desired. These strategies of midtwentieth century farmers allow us to speculate that similar substrates and topographic locations were preferred by prehistoric farmers within this volatile environment. Similar patterns have also been recognized by archaeologists working in Botswana. Denbow (1984) has shown that Early Iron Age archaeological sites with small kraals are preferentially situated on boundaries of clay and sand substrates. 2.1.3. Vegetation and carbon isotopes Trees and shrubs in Africa use the C3 photosynthetic pathway, whereas grasses predominantly use the C4 photosynthetic pathway (Wang et al., 2009). C4 plants use water more efficiently than C3 plants, particularly during the dry season in arid environments (Wang et al., 2010), and they also have higher photosynthetic rates at high temperatures and under high light conditions (Farquhar et al., 1989a,b). Thus, C4 plants are more successful than C3 plants when growing season temperatures are high and light levels are moderate to high (Sage et al., 1999; Still and Powell, 2010), such as in grasslands and savannas with limited tree cover (Long, 1999). Goodfriend (1999:503) and others (Ambrose and Sikes, 1991; Ambrose and DeNiro, 1989; Bird et al., 2004; Farquhar et al., 1989b; Levin et al., 2006; Wang et al., 2010; Wynn and Bird,

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

3

Fig. 1. Geology of Hwange National Park (adapted from Sithole, 1994). t - Jurassic-Triassic (251e145.5 Ma) Karoo grits, sandstones, and siltstones; Bj e Jurrassic-Triassic (251e145.5 Ma); p - Permian (299-251 Ma) glacial beds, coal measures, mudstones, and sandstones; Gn - gneisses of various age; pG - various ages Proterozoic (2500-542 Ma) intrusive granites; w - mid-Precambrian (~2500 Ma) Piriwiri and minor quartzes; n - mid-Precambrian (~2599 Ma) Malaputese and Kahire paragneiss, other metasediments, and amphibolite; s - Late Precambrian (~1500-542 Ma) Sijarira red grits, sandstones, shales, and conglomerates; r e Pleistocene Kalahari aeolian sand; gray lines e linear dunes.

2008) have also shown that C4 plant frequency increases with drier and/or hotter conditions. The d13C isotope ratio is significantly different for plants that use C3 (
26.5 ‰ in southern Africa) and C4 (
12.5‰ in southern Africa) pathways, allowing calculation of the relative contributions of each to a d13C isotope value of soils and sediments. Increases in mean annual precipitation (MAP) are strongly correlated to decreases in vegetation d13C due to an increase in tree (C3) cover. Conversely, decreasing the precipitation increases the amount of C4 cover and enriches d13C (Still and Powell, 2010). It is for these reasons that d13C can be used as a paleoclimatic indicator. 2.1.4. Paleoenvironment During the Holocene, local conditions become increasingly important in paleoecological reconstructions because no dramatic worldwide events (e.g., glacial-to-interglacial periods) masks the regional response to atmospheric and global driving forces (Ashley et al., 2011; Burrough et al., 2009; Jones et al., 2009; O'Brien et al., 1995; Thomas, 2008). However, the few paleoenvironmental studies in Zimbabwe concern periods either older (e.g., Brain, 1969; Cruz-Uribe, 1983; Stokes et al., 1998) or younger (e.g., Mazvimavi, 2010; Sithole and Murewi, 2009) than our time span of interest, so it is difficult to draw any cohesive, well-supported conclusions about local past environmental change. Summarizing regional trends drawn from elsewhere in the southern Africa summer rainfall zone provides contextual complexity.

Overall, the strongest agreement amongst various types of proxy data within the summer rainfall zone suggests that c. 8.0 to c. 5.0 ka was wet and warm (Beaumont et al., 1984; Butzer, 1984b; Chase, 2009; Chase et al., 2010; Finch and Hill, 2008; Scott et al., 2003; Shaw and Cooke, 1986; although see Wadley, 1986; Deacon, 1974; Walker, 1995; Rust et al., 1984), c. 5.0 to c. 3.0 ka was dry and cold (cf., Beaumont, et al., 1984; Butzer, 1984a,b; Finch and Hill, 2008; Lancaster, 1989, 2002; O'Connor and Thomas, 1999; Rust et al., 1984; Scott et al., 2003; Stager, 1988; Van Zinderen Bakker, 1982, 1995; although see Chase, 2009; Chase et al., 2010; Robbins et al., 1998; and Shaw and Cooke, 1986 for evidence of a humid interval early in this period), and c. 3.0 to c. 2.0 ka was wet and warm (cf., Butzer, 1984b; Cooke, 1984; Deacon, 1984; Deacon and Thackeray, 1984; Haynes and Klimowicz, 2005; Robbins et al., 2005; Shaw, 1985a,b; Shaw and Cooke, 1986; Talma and Vogel, 1992). During the past c. 2.0 ka, regional variation and detailed study of various types of proxy data produce contradictory results. For instance, speleothems from Cold Air Cave in South Africa (Holmgren et al., 1999) suggest cool and dry conditions the past c. 1.6 ka while lake levels, soil development (Butzer, 1984b), and ongoing peat development (Norstrom et al., 2009) suggest mesic conditions between c. 1.5 ka and 1.3 ka. Arid conditions between 1.2 and 1.1 ka are evinced by dune reactivation in Namibia (Blümel et al., 1998; Thomas et al., 1997) and likely also by fluvial aggradation (Lancaster, 2002) and lunette development at Koopan Suid in South Africa (Lancaster, 1989).

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

4

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

Beginning around c. 1.2 ka and continuing to c. 0.8e0.5 ka, relatively wet and sometimes volatile conditions are demonstrated by various proxy data, including peat development (Norstrom et al., 2009), microfaunal remains (Brain and Brain, 1977), stalagmite growth at two separate caves (Cooke, 1984; Holmgren et al., 1999), and sediments (Butzer, 1984b). Even so, on the western margin of the Kalahari and into the Namib, arid conditions persist through this time period, based largely on deflation (Butzer, 1984a) and isotope data from hyrax middens (Chase, 2009; Chase et al., 2010). Denbow (1986) noted that there is a marked decrease in archaeological sites dating to c. 0.75 ka, implying abandonment of the Makgadikgadi basin because of drought. Other records suggest the past c. 0.7 to 0.5 ka have generally been relatively dry, such as speleothems from Cold Air Cave (Holmgren et al., 1999), dune reactivation (Blümel et al., 1998), fluvial aggradation (Lancaster, 2002), and lake levels (Shaw, 1985a,b). Even so, at the Kathu Pan and Vlei (Beaumont et al., 1984; Butzer, 1984b), and the Okavango Delta, wetter conditions are supported. As the proxy data for the late Holocene make clear, any broadly interpreted trends across southern Africa may over-simplify the record and mask variability in the surface water availability and distributions of plants and animals that effected human populations. The contradictory regional environmental data shows the importance of establishing local baseline paleoenvironmental data to provide context for any archaeological remains. 2.2. Archaeological background The earliest evidence of nomadic hunter-gatherers in Hwange National Park dates to the Early Stone Age (Haynes and Klimowicz, 1998), and some semblance of this hunting-and-gathering way of life (albeit with major technological and evolutionary advances), continued to be practiced there into the historic period. Impala and Ngabaa Rockshelters, which are situated atop a sandstone ridge overlooking the Deka River and along the northern boundary of the study area, were first occupied before ca. 6 ka (Wriston and Haynes, unpublished results) when relatively large forager groups began to use the area, leaving characteristic microlithic tools and manufacturing debris behind along with a generalized tool kit and ostrich eggshell beads. Human use of the rockshelters intensified between 4 ka and 3 ka, and unique, finely-engraved animal spoor (hoofprint) carvings that cover much of the rockshelters' sandstone walls were likely created during this time (Haynes et al., 2011; Wriston and Haynes, 2009). However, beginning around 1.8 ka, hunter-gatherer occupation of the study area had shifted from large aggregate groups to small family or task groups (Wriston and Haynes, unpublished results). The use of these rockshelters by smaller groups of hunter-gatherers continued even as a new cultural adaptation in the area took hold in the lowlands, namely agropastoralism. Robinson (1966) was the first to describe and date the only early agropastoralist site yet known in the Hwange National Park at Kapula Vlei to 1.0 ka. This site was revisited during this study and different loci were identified and dated to ca. 1.9 ka, 1.65 ka, and 1.1 ka (Wriston and Haynes, unpublished results). As such, Kapula Vlei contains the earliest evidence of agropastoralists so far reported in the region, which makes our understanding of this basin and the variables that drew agriculturalists here especially important. 3. Methods 3.1. Fieldwork Geoarchaeological reconnaissance covered over 50 linear km along river cuts and around 20 linear km of ridgelines, pans, and

alluvial plains. These investigations focused on identifying the basic geology within the study area and characterizing sediments and soils of Holocene age, particularly near known archaeological sites. This documentation was accomplished by examining the cut banks of incised seasonal tributaries and rivers and hand augering across landforms that lacked any subsurface exposure. Selected stratigraphic profiles were photographed, point-plotted using a GPS, drawn and described. Descriptions included: topographic position, vegetation cover, any disturbances, strata identification and the nature of their upper and lower boundaries, Munsell color (dry and moist), effervescence with 10% HCl, sediment texture, likely depositional environment, the relative amount of organic material, gravel abundance, clast lithology, roundness, and size, and a summary description (following Schoeneberger et al., 2002). Block surveys were not possible given the park's wildlife abundance and associated safety concerns. At Josivanini, two units were hand-excavated on the upper portion of a large linear stabilized dune. One of these units was dug to 2.5 m below surface, and the other to 4 m below surface. Charcoal for radiocarbon analysis was retrieved from these units by sieving excavated materials through a 3 mm mesh screen or collecting directly from the wall profile. In total, 20 natural exposures, 26 auger samples, and two handdug units were described and/or sampled. Samples of buried soil horizons, other strata of interest, and charcoal were collected for laboratory analyses and AMS radiocarbon dating to establish chronostratigraphic units that could be correlated between basins. 3.2. Laboratory procedures Materials were preferentially selected for further analysis based on their potential to provide important temporal control of stratigraphic layers traceable across different landscapes (e.g., soils) and any association with archaeological materials. The samples were analyzed by the University of Arizona Radiocarbon laboratory and subjected to standard pretreatment, combustion, reduction, and AMS measurement procedures. AMS radiocarbon analysis of the sediments included both bulk organic carbon and humate fractions. These paired dates were compared, and whichever provided the oldest age (generally the humate fraction) was accepted as the most reliable given that any contamination is likely by younger carbon (Holliday, 2004; Martin and Johnson, 1995; Walkington, 2010). The d13C isotope ratio determined during the radiocarbon dating procedures was used to calculate vegetation types that contributed to each sample's soil carbon. As mentioned above, this ratio is significantly different for C3 (trees and shrubs) and C4 (grasses) photosynthetic pathway plants, allowing us to calculate the relative contributions of each to the soil sample (Cerling and Hay, 1986; Wynn, 2000). Using the average African d13C isotope ratio in C3 plants as
26.5 /oo, and C4 plants as
12.5 /oo (Bird et al., 2004), we calculated the contribution of each plant type, given that C3 ¼ X and C4 ¼ 1-X in the equation: d13C ¼ X (
26.5) þ (1-X)(-12.5) (Von Schirnding et al., 1982; Johnson et al., 1998; Segalen et al., 2002). 4. Results 4.1. Geochronology Twenty-three radiocarbon dates are reported in Table 1. As these radiocarbon ages show, most of the sediments within the study area date to within the past one and a half millennia. Unfortunately, these are too recent to address the earliest evidence of agropastoralism and settled village life, given the earlier dates of 1.9 ka and 1.65 ka from the Kapula Vlei Early Iron Age site (Wriston and Haynes, unpublished results). This means that many of the

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

5

Table 1 Radiocarbon dated material from stratigraphic profiles. Area

Lab #

Locale

Depth (cmbs)

Material

d13C

%C3

%C4

14

Kapula river

AA95105h AA95105 AA95106 AA95106h AA95094 AA95089 AA95102 AA95102h AA95103 AA95103h AA95104 AA95104h AA95100 AA95100h AA95101 AA95101h AA95096 AA95107 AA95107h AA95108 AA95108h AA95097 AA95098

KR08-3 KR08-3 KR08-3 KR08-3 KRL1 KRl4 Br08-3 Br08-3 Br08-3 Br08-3 Br08-3 Br08-3 10-A8 10-A8 10-A8 10-A8 Br09-1 Dtma09-1 Dtma09-1 Sin08-L3 Sin08-L3 JosSu2 JosSu2

95 95 185 185 100 85e87 50e70 50e70 145e150 145e150 300 300 57e64 57e64 170e183 170e183 205 335e345 335e345 140e145 140e145 110 200e205

humates sediment sediment humates bovine tooth charcoal sediment humates sediment humates sediment humates sediment humates sediment humates charcoal sediment humates sediment humates charcoal charcoal


14.9
14.3
18.1
15.5
6.6/-17.1b
24
18.3
16
15.9
19.4
18.4
18.1
19.1
17.4
15
15.4
23.9
15.6
17.4
20
17.2
23.7
27.3

17 13 40 21 33

83 87 60 79 67

41 25 24 49 42 40 47 35 18 21

59 75 76 51 58 60 53 65 82 79

22 35 54 34

78 65 46 66

292 380 825 1009 1253 1790 284 408 487 549 1347 1466 278 398 812 863 573 2206 2133 3580 2902 2574 415

Bumbusi river

Deteema Sinamatella Josivanini a b

C age BP

±

Cal yrs BPa

34 35 35 34 43 37 34 34 35 34 35 35 34 34 35 34 36 39 36 38 37 38 35

311 396 705 859 1128 1654 302 431 504 529 1228 1324 295 407 700 728 539 2176 2068 3819 2978 2606 445

Mean probability 2-sigma (95%) calibration; calibrated using Calib 7.0 SHCal13 (Stuiver et al., 2013). Bovine tooth corrected by 10.5‰ for fractionation/trophic level (DeNiro and Epstein, 1978).

sedimentary basin fills exposed within the natural river cuts have accumulated only within the relatively recent past. The paucity of older sediment packages is due to periodic scour, particularly in bedrock constrained basins. Samples in dune fields that accumulate deposits while other points of the landscape erode may help fill these data gaps. 4.2. Sample locations and descriptions Of the 20 natural exposures, 26 auger samples, and two handdug units described and/or sampled, six (Fig. 2) are detailed here as representative of each general area. 4.2.1. Bumbusi River area The Bumbusi River empties into the Deka River on the northern boundary of the study area. The Deka holds water most of the year within large pools cut into the basalt bedrock, creating habitat for fish, crocodile, and hippopotamus. Near-modern sand bars and terrace remnants were discovered along the Deka, but most deposits have been scoured from the channel. The Deka's tributaries, such as the Bumbusi River, have better potential to provide stratigraphic evidence of past environments due to the presence of buried soils and pockets of deposition. During the extended archaeological excavations at nearby Impala and Ngabaa Rockshelters, numerous samples were obtained from Bumbusi River terraces, slopes between the river and Bumbusi Ridge, and promising sediment catchments further to the west (Fig. 3). Bumbusi Ridge is comprised of sandstone, and the Bumbusi River has cut a gorge through the ridge and the basalts that outcrop upstream of its confluence with the Deka River (Fig. 3). Most the deposits along the Bumbusi River are cumulative, and within the river's lower section they form a sequence of organic-rich standing water deposits with upper weathered zones commonly separated by channel bed sands and gravels (Fig. 3; Table 2). Inspection of the area for channel meandering both on the ground and from satellite images revealed no relic scrolls. Rather, these deposits represent wet-and-dry cycles, when seasonal ponding and vegetation growth

accumulated organic material in fine sediments. After the water receded, these deposits weathered subaerially. The alluvial terraces are up to 4 m above the channel bottom, but within the surrounding Bumbusi slope deposits, bedrock is generally less than 60 cmbs, with localized deposits of up to 200 cmbs before impenetrable material was reached using a hand auger. The deepest deposits encountered within this basin are near the geologic contact of the dipping basalt and sandstone bedrock exposed in the Bumbusi River. Within an augured sample west of the Bumbusi River, a grassy swale was tested and found to have loamy sand in the upper 55 cmbs, loamy clay and clay below 55 cmbs, CaCO3 nodules and calcrete below 130 cmbs, and was too indurated to auger at 180 cmbs (presumably just above bedrock). East of the Bumbusi River on the flank of Bumbusi Ridge, the slopes are underlain by loose, quartz-dominated alluvial red sand with little to no soil development. Impala and Ngabaa Rockshelters are situated on the rocky upper slope of the ridge, and their fill is comprised of anthropogenic charcoal ash, artifacts, and rockfall. The rockshelters had been scoured to bedrock in the early-to-mid Holocene, after which human occupation led to an accumulation of 80e97 cm of anthropogenic fill (Wriston and Haynes, 2009). Deposits within the Bumbusi River terraces show some evidence of human impacts, including an upper leached zone possibly the result of historic irrigation or waterway modification, and a lithic flake discovered in an auger sample from 180 cmbs, suggesting prehistoric use of the area. Before the Bumbusi River became entrenched, its alluvial terraces had potential for agricultural use. The availability of surface water most of the year, the loamy sediments, and varied topography and sediment types over a short distance offers flexibility in field placement across various elevations and substrates to increase the odds of success in this volatile environmental regime. However, most the area has shallow soils inundated during the wet season, or is sandy, and therefore not as nutrient rich as other basins. These factors could have lessened the Bumbusi River basin's overall appeal to early agriculturalists.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

6

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

Fig. 2. Sampled Locales (black circles) within Hwange National Park (black line) on Landsat imagery.

Fig. 3. Landsat image of Bumbusi area left, BR08-3a's upper profile right (250 cm in height).

4.2.2. Kapula Vlei Kapula Vlei is the grassy and seasonally-inundated land adjacent to the Kapula River channel (Fig. 4), which is itself usually dry most of the year. The river is a headwater tributary that drains into the Lukosi River just before the Lukosi cuts through a sandstone

ridge, forming a relatively deep gorge downstream. Within the Kapula Vlei basin, two seasonal headwaters drain its southern portion. One emerges from granite kopjes (South African term for a small hill) located to the south, and the other (the current trunk of the Kapula River) from the ancient mudstones and glacial deposits

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

7

Table 2 Lower Bumbusi river stratigraphic profile description (Br08-3a). Stratum

Depth (cm)

Soil horizon

Field texture

Color

Description

Dep/Landform

1

0e25

AB

sandy clay

10YR5/3d; 10YR4/2m

alluvial terrace

2

25e60 60e88

ABqb Bwb

loamy clay sandy clay

10YR4/1d; 10YR3/1m 10YR5/3d; 10YR4/2m

88e100

C

sand & gravels

7.5YR5/5d&m

100e115 115e135

Ab Aqb

sandy clay clay

10YR5/3d; 7.5YR4/2m 10YR4/2d; 7.5YR3/2m

135e160

C2

sand & gravels

7.5YR5/2d&m

160e175

Ab2

clay

10YR5/2d; 10YR3/1m

175e190

Ab3

clay

10YR4/1d; 10YR3/1m

190e200

C3

loamy sand

10YR3/2d&m

5

200e235

Atqb

clay

10YR3/2d&m

6

235e270

Atkb

loamy clay

10YR4/1d&m

270e295

Btkb

loamy clay

10YR4/2-7/2d; 10YR3/2-4/3m

295e305 305e400

C4 Btkb2

loamy sand sandy loam

10YR5/3d; 10YR3/3m 10YR5/3d; 10YR3/2m

400e415

C5

loamy sand

10YR5/2d; 10YR3/2m

415e430

C6

sandy loam

7.5YR3/2m

430

water

NE w/10% HCl; sparse shrub veg; wk sab; gradual lb VSE w/10% HCl; wk sab; abrupt lb; org-rich SLE w/10% HCl; wk sab; some ox mottles; abrupt lb NE w/10% HCl; loose med-coarse sand; ox mottles; distinct bedding planes; channel deposit; undulating, abrupt lb NE w/10% HCl; mod sab; gradual lb NE w/10% HCl; mod sab; mod silica; abrupt lb; some charcoal flecks NE w/10% HCL; loose; abrupt lb; wk silica cemetation; interbedded sands/small gravels/ thin mud beds; many charcoal flecks NE w/10% HCl; wk columnar; wk gley; thin clay films; some charcoal flecks; gradual lb NE w/10% HCl; mod sab; abrupt lb; org-rich; many charcoal flecks NE w/10% HCl; loose; abrupt lb; coarse sand & pebbles NE w/10% HCl; wk sab; mod wilica; abrupt lb; many charcoal flecks 10YR7/2d&m nodules & stringers; STE w/10% HCL; st sab; wk silica; gradual lb nodules & stringers; VE w/10% HCL; mod sab; wk silica; bands created by precipitation zones STE w/10% HCl; medium sand; abrupt lb STE w/10% HCL; massive; hard; wk silica; gradual lb VSE w/10% HCL; massive coarse sand & pebbles; gradual lb saturated; VSE w/10% HCl; oxidized zone above water table; loose when moist groundwater table, basalt bedrock nearby

3

4

7

alluvial terrace alluvial terrace channel bed

standing water standing water channel bed

standing water standing water channel bed standing water standing water alluvium w/groundwater channel bed alluvium w/groundwater channel bed alluvium w/groundwater

d-dry, m-moist, HCl e hydrochloric acid, NE e noneffervescent; VSE e very slightly effervescent; SLE e slightly effervescent, STE e strongly effervescent, VE e violently effervescent, veg e vegetation, wk e weak, sab e sub-angular blocky, lb e lower boundary, org e organic, mod e moderate, ox e oxidation, vs e very strong, vvs e very very strong, s e strong, vw e very weak.

Fig. 4. Landsat image of Kapula Vlei area left, KR08-3 profile right (280 cm in height).

to the southwest. Evidence of repeated channel cut-offs, and cutand-fill sequences, show that the location of these two drainages’ confluence has changed many times during the history of the basin.

Presently, they join below the recorded archaeological site of Kapula Vlei, but granitic material 2 km upstream hints at earlier confluences there, on the opposite side of the valley.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

8

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

Meander scrolls that cut tightly through the south-central portion of the basin also (Fig. 4) show that a change in the amount of water, its velocity, and/or sediment load has occurred given that the modern channel is relatively straight and entrenched. This change may have resulted from tectonic activity or an increase in climate volatility, with punctuated flood events rather than gentle precipitation throughout the rainy season. However, channel alteration could also have resulted from prehistoric agricultural practices and/or historic activities in the area. Landscape alterations since the park's creation include dambuilding, emplacement of a water pump to feed nearby Masuma Dam, and creation of two different campgroundsdone of which was abandoned during the 1980s when the drainage repeatedly downcut, making access impossible (Anon. National Park staff member, personal communication, 2008). The artificial concentration of wildlife at a pumped water source during the dry season may also be affecting erosion rates and sediment load by lowering the groundwater levels, encouraging concentrated overgrazing and vegetation destruction, and increasing trampling and bioturbation, particularly by elephants. The soils and sediments exposed in Kapula River cutbanks are variable due to the history discussed above, but we focused on documenting the dark, organic-rich soils near the Early Iron Age village exposed at the Kapula Vlei Site. Within the site itself, pots were found eroding out of prehistorically excavated pits beneath a strongly-developed Ap horizon of clay loam with discreet dagalined floors exposed around 20 cmbs. At approximately 45 cmbs, the Ap horizon transitions into a Bt horizon that, albeit obviously formed on a medium sand substrate, still forms a ribbon more than 5 cm long due to the clay and sesquioxide accumulation. Below 82 cmbs is a channel bed deposit of an ancient loamy subangular and subrounded coarse sand, which by 100 cmbs has granular structure and many CaCO3 nodules and stringers. Of note, a significant amount of time is missing between the Bt horizon and the much older relic channel sands. Upstream of the Kapula Vlei site, soils are also organic-rich sandy-to-clay loams in the cut's upper portion (Fig. 4; Table 3). The deep A-horizon here may be due to human tilling and organicenrichment; however, grassland soils also commonly exhibit a thickened A-horizon (Martin et al., 1976), and given the dense modern grass cover, this may be a natural phenomenon. Several cumulative horizons evident in the area merge into the upper profile layers due to pedogenic alteration, and sand lenses deposited by migrating prehistoric rills and channels occasionally can be discovered within the otherwise loamy sediments. KR08-3's profile section has a distinct shelf formed at the contact above the same ancient sands described for the Kapula Vlei site. Due to differential cohesion along the stratigraphic boundary, this shelf hangs over the undercut sands below, a process accentuated by elephant-digging along this transition, presumably for minerals that precipitate due to the change in sediment texture (Fig. 4). Alluvial terraces within Kapula Vlei are generally around 280 cm above the channel bottom. Historic channel incision of over 1 m implies that recent land use changes, such as groundwater pumping for Masuma Dam, may be altering the local base level/ groundwater table. However, more work needs to be done to determine whether land use, changing climatic conditions, or tectonic activity, is causing the downcutting (Fig. 5). Although the nutrient rich sandy-to-clay loams, thriving vegetation, and local topographic and substrate variety suggest that agriculture would be successful at Kapula Vlei (as it was in the past), the lack of surface water during the dry season would make permanent villages here difficult to sustain under modern conditions. In fact, there has been some difficulty finding enough groundwater to extract for the Masuma Dam reservoir (Anon.

National Park staff, personal communication, 2009). Salvage archaeological excavations here recovered ceramics, charcoal, animal teeth, and animal bone. A bovine tooth recovered from locus 1 was dated to 1253 ± 40 14C yr BP (AA95094), or 1179 to 1069 cal yr BP (Calib 7.0; Stuiver et al., 2013). Ceramic temper and charcoal from nearby locus 4 dated to 1990 ± 40 14C yr BP (Beta275280; 1931 to 1864 cal yr BP; Calib 7.0; Stuiver et al., 2013) and 1790 ± 40 14C yr BP (AA95089; 1705 to 1611 cal yr BP; Calib 7.0; Stuiver et al., 2013) respectively. The locus 4 dates are some of the earliest in the region for settled village life, making our understanding of this basin, and the variables that drew agriculturalists here, especially important. 4.2.3. Deteema The Deteema area (Fig. 6) was not extensively sampled, but is included here because it provides a unique example of CaCO3-rich deposits associated with seep or spring activity. The Deteema basin is within Permian-age mudstones, shales, glacial deposits, and sandstones bordered by dolomite dikes and gneisses of various ages (Sithole, 1994). It is characteristically different from other basins we visited because it has a possible spring head less than several hundred meters upstream. Surrounding the drainage, dense grasslands are well populated by various ungulates and a large baboon troop that uses the seep for drinking water. Unfortunately, a bull elephant chased us away before a detailed profile description could be completed, but Fig. 6 shows the multiple bands of light colored, deep sediments with compact structure, zones of dense CaCO3 accumulation, and some manganese stains on peds within the lower strata. Luckily, a few sediment samples were described before interruption by the elephant and are detailed in Table 4. The terrace cutbank is up to 4 m above the channel, which is locally very steep-sided and narrow due to the fine-grained cutbanks. Near the base of the sampled area, groundwater was seeping from a stratigraphic break. Although the deposits exposed in the examined Deteema stratigraphic profile would not be conducive to productive agricultural development, the presence of an active seep during the dry season, and thriving grasslands at higher topographic locations nearby, imply that plant cultivation could have been successful on the surrounding slopes and nearby plains. In fact, agricultural activities may have been secondary, with successful pastoralist activities likely only limited by the past distribution of disease vectors. 4.2.4. Sinamatella The Sinamatella River winds its way across a wide alluvial plain below a sandstone mesa where a camp has been established for tourist lodging and game viewing (Fig. 7). In addition to this sandstone, which locally contributes a significant amount of alluvium and colluvium, the geologic substrate includes Permian-age mudstones, sandstones, shales, and coal seams (Sithole, 1994). These old, relatively soft deposits are eroding, leaving more resistant remnants of Jurassic-Triassic age sandstone high above the gradually lowering plain. Vegetation here is rather sparse woodland with an understory of thin grasses. Along Sinamatella River, three profiles were examined that reveal three buried soils (Fig. 7; Table 5). A relatively recent alluvial overbank deposit in the upper 10 cm overlies the uppermost soil. All the soils are formed on overbank alluvial terrace deposits separated by channel bed sands and gravels. The soil between 10 and 60 cmbs is relatively strongly developed whereas the two lower soils have been partially truncated by channel scour. At the base of the profile, large rocks are found above a calcrete layer. Near the river, vegetation is relatively dense, and roots, rootlets, and termite activity is common.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

9

Table 3 Kapula River (Kr08-3) stratigraphic profile description. Stratum

Depth (cm)

Soil horizon

Field texture

Color

Description

Dep/Landform

1

0e35

A

loamy sand

10YR5/3d; 10YR3/2m

alluvial terrace

2

35e55

Ab

sandy loam

10YR3/2d; 10YR3/1m

3

55e69 69e72 72e105

Btkb C ABkb

clay loam coarse sand clay loam

10YR5/2d; 10YR3/2m 10YR4/3d&m 10YR3/1d; 10YR2/1m

4

105e115 115e127

BCb ABb

clay loam sandy loam

10YR4/2d; 10YR3/2m 10YR3/3-4/2d; 10YR3/2m

5

127e145 145e160

Ab2 C2

sandy loam loamy sand

10YR3/2; 10YR3/2m 10YR3/3d; 10YR3/2m

6

160e180

Ab3

loam

10YR3/2d; 10YR3/2m

180e190

Bqb

clay loam

10YR4/2d; 10YR3/2m

190e225

Bqb2

clay loam

10YR4/2d; 10YR3/2m

7

225e295

C3

loamy sand

7.5YR3/2d; 7.5YR3/2m

STE w/10% HCl; wk sab; grassy surface; abrupt lb STE w/10% HCl; wk sab; organic-rich; poorly sorted; gradual lb STE w/10% HCl; compact; s sab; abrupt lb NE w/10% HCl; clean quartz sand; abrupt lb VSE w/10% HCl; wk sab; slightly gleyed; gradual lb SLE w/10% HCl; abrupt lb VSE w/10% HCl; shallow soil on top of loamy sand overbank deposit; org rich on top grading to sand below; abrupt lb VSE w/10% HCl; shallow soil; org rich; abrupt lb SLE w/10% HCl; abrupt lb; some termite concretions SLE w/10% HCl; wk ab w/thin silica coatings; gradual lb SLE w/10% HCl; ab peds w/silica coatings; gradual lb; targeted elephant diggings SLE w/10% HCl; very hard wedge to platey peds w/silica coatings; abrupt lb; targeted elephant diggings NE w/10% HCl; termite-affected; ab structure; wk silica films; oxidized

8

295þ

C

sand & gravels

alluvial terrace alluvial terrace channel bed alluvial terrace alluvial terrace alluvial terrace

alluvial terrace channel bed alluvial terrace alluvial terrace alluvial terrace

channel bed channel bed

d-dry, m-moist, HCl e hydrochloric acid, NE e noneffervescent; VSE e very slightly effervescent; SLE e slightly effervescent, STE e strongly effervescent, VE e violently effervescent, veg e vegetation, wk e weak, sab e sub-angular blocky, lb e lower boundary, org e organic, mod e moderate, ox e oxidation, vs e very strong, vvs e very very strong, s e strong, vw e very weak.

Fig. 5. Recent downcutting/channel movement evident at the base of a living tree with ranger for scale (approximately 180 cm in height).

It is unlikely that the Sinamatella area would have been targeted by early agriculturalists. The steep slopes of the sandstone scarp and relative flat plain below do not provide flexible topographic options for agricultural field placement and the deposits tend to be sandy, nutrient-poor, and porous. No evidence of prehistoric farming has been reported from the locale. However, during our early dry season fieldwork we did see occasional pools of water above the calcrete layer within the river bottom, which could have been used by prehistoric pastoralists for livestock-watering. 4.2.5. Shumba Shumba is a sink for much of the central portion of the study

area, and is where waters begin to drain southward towards the Kalahari Desert. Standing water is common in natural pans (Fig. 8), but pumping of underground water is necessary to assure that resident hippopotamus have plenty of water. The substrate at Shumba is loose aeolian Kalahari sand (Sithole, 1994). Grasslands dominate the vegetation with sparse clusters of trees concentrated on local high points. A profile sample augered in the sand sheet above the Big Shumba pan (Fig. 8; Table 6) revealed a thick, well-developed, organic-rich soil throughout the upper 178 cm. Organic and aeolian accumulation on this persistently developing grassland has thickened the soil. Pedogenesis has apparently kept pace with the

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

10

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

Fig. 6. Landsat image of Deteema area left, stratigraphic profile Dtma09-1 right (325 cm in height above trowel at base).

Table 4 Deteema (Dtma09-1) stratigraphic sample description. Stratum

Depth (cm)

Soil horizon

Field texture

Color

Description

Dep/Landform

e

105e115

e

clay loam

10YR3/1d; 10YR3/1m

spring mound?

e

160e165

e

sandy loam

10YR5/1d; 10YR3/1m

e

185e195

e

loamy clay

10YR4/1d; 10YR3/2m

e

195e205

e

fine sandy loam

10YR5/2d; 10YR3/2m

e

215e225

e

loamy sand

10YR5/3d; 10YR4/3m

e

240e255

e

loamy clay

10YR5/2d; 10YR3/1m

e

280e290

e

loamy clay

10YR5/2d; 10YR3/1m

e

305e310

e

loamy fine sand

10YR8/1d; 10YR6/2m

VE w/10% HCL; mod sab; abundant CaCO3 stringers; many small burrows and vesicles; some mica VE w/10% HCL; mod sab; CaCO3 stringers & nodules; abundant coarse sand and small gravels VE w/10% HCL; s sab; dark organic decomposing rootlets; mottled; many small vesicles; CaCO3 stringers; mica-rich VE w/10% HCL; mod sab; CaCO3 stringers;; many small vesicles; some mica VE w/10% HCL; loose med-coarse sand w/occasional small gravel; CaCO3 stringers; occasional gastropod shell fragment VE w/10% HCL; vs sab; CaCO3 stringers; many rootlets VE w/10% HCL; vs sab; many CaCO3 stringers; many rootlet cast and small vesicles; some mica VE w/10% HCL; wk sab; abundant precipitant CaCO3; occasional gastropod shell fragment; abundant small vesicles (rootlets)

channel shallow water/marsh

shallow water channel

standing water standing water shallow water

d-dry, m-moist, HCl e hydrochloric acid, NE e noneffervescent; VSE e very slightly effervescent; SLE e slightly effervescent, STE e strongly effervescent, VE e violently effervescent, veg e vegetation, wk e weak, sab e sub-angular blocky, lb e lower boundary, org e organic, mod e moderate, ox e oxidation, vs e very strong, vvs e very very strong, s e strong, vw e very weak.

aeolian deposition, suggesting no severe droughts have decreased the vegetation, or increased the depositional rate of sand accumulation, beyond a maintainable threshold. Below 178 cmbs, the sediments transition to what appears to be clean white quartz sand. However, silica and calcium carbonate coat the well-rounded quartz grains and are forming nodules in this seasonallysaturated sediment. The relatively loose sand at Shumba would not be able to sustain early plant cultivation due to limited nutrient availability, little topographic relief, and seasonal inundation. However, livestock may have thrived foraging on the abundant grasses and surface waters. No evidence of agropastoralism has been found at the locale, but Later Stone Age lithics are not uncommon.

4.2.6. Josivanini dune Two units were hand-excavated into the upper slope of one of

the long, linear, stabilized dunes shown in Fig. 9. One unit was excavated down to 4 m below surface, and the other to 2 m. Charcoal and bulk samples were collected from the profiles. Little soil development was revealed within the units (Fig. 9; Table 7), but several horizons were noted based on color changes, differences in compaction, and/or charcoal accumulation. Based on these slight differences, relic surfaces or erosional unconformities were provisionally noted at 90 cmbs, 160 cmbs, and 193 cmbs. These transitions reflect periods of dune reactivation often associated with fire, drought, and high winds (Lancaster, 1988; O'Conner and Thomas, 1999; Stokes et al., 1998; Haynes and Klimowicz, 2005). Wild fires left relatively large chunks of charred wood now buried in the dune, with a small sample identified as coming from the same tree taxa found there today. These pieces of charcoal are especially common in discontinuous lenses at around 110 cmbs. Agriculture could not be supported at Josivanini due to the lack

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

11

Fig. 7. Landsat image of Sinamatella area left, Sin08-L3 profile right (200 cm measuring stick).

Table 5 Sinamatella River cut stratigraphic profile description (Sin08-L3). Stratum

Depth (cm)

Soil horizon

Field texture

Color

Description

Dep/Landform

1 2

0e10 10e50

C ABb

sandy loam loam

7.5YR4/2d; 7.5YR3/2m 7.5YR45/2d; 7.5YR3/2m

alluvium alluvial terrace

50e57 57e85

BCb C2

loam loamy sand & gravels

7.5YR6/2d; 7.5YR3/2m 7.5YR6/2d&m

3 4

85e115 115e135 135e155

C3 Ab Bqb

sandy loam sandy loam loamy sand

7.5YR5/2d; 7.5YR5/2m 7.5YR5/2d; 7.5YR5/2m 7.5YR6/2d; 7.5YR4/4m

5 6

155e160 160e180

C4 ABb2

sand & gravel loamy sand

7.5YR5/2d&m 7.5YR5/2d; 7.5YR4/4m

180e205

C5

sand & gravel

7.5YR5/2d&m

205þ

C6

calcrete

calcrete

NE w/10% HCl; surface duff NE w/10% HCl; org rich; compact; st sab; many animal burrows; gradual lb NE w/10% HCL; compact sab; abrupt lb NE w/10% HCL; bedded gravels & coarse sands; abrupt lb NE w/10% HCL; ox mottling; abrupt lb NE w/10% HCl; org rich; gradual lb NE w/10% HCl; wk sab w/silica films; mediumsand substrate; abrupt lb NE w/10% HCl; loose; abrupt lb NE w/10% HCl; compact irregular peds w/silica films; some redox mottling; medium sand substrate; abrupt lb NE w/10% HCl; bedded with subangular rocks up to 8 cm; abrupt lb VE w/10% HCl

7

alluvial terrace channel bed alluvial terrace alluvial terrace channel bed channel bed channel bed

channel bed

d-dry, m-moist, HCl e hydrochloric acid, NE e noneffervescent; VSE e very slightly effervescent; SLE e slightly effervescent, STE e strongly effervescent, VE e violently effervescent, veg e vegetation, wk e weak, sab e sub-angular blocky, lb e lower boundary, org e organic, mod e moderate, ox e oxidation, vs e very strong, vvs e very very strong, s e strong, vw e very weak.

of nutrients in this deep, unconsolidated clean quartz sand with limited and seasonal surface water. No archaeological evidence from any period has been found on dune crests similar to the Josivanini locality, but Stone Age lithics do occur within many of the interdunal troughs in the region. In addition, some of the seasonal pans have sparse, very Late Iron Age/recent potsherds around them, which may indicate brief occupations by livestock herders or foragers using ceramics in the last two centuries (Haynes and Klimowicz, 1998). 5. Discussion 5.1. Soil stratigraphy and landscape evolution Although sediments of the appropriate age to address the earliest days of agropastoralism have not been preserved in much

of the study area, some important chronostratigraphic trends did emerge from the sampling project. As Fig. 10 shows, representative samples from the six different localities in Hwange National Park can be related by comparing common stratigraphic markers (e.g., soils, unconformities) and radiocarbon ages. Within the Bumbusi River and Kapula River basins, the alluvial sequences correlate well, with periods of stability marked by weak, but identifiable soil development at ca. 400 14C yrs BP (408 cal yrs BP or 0.41 ka), ca. 550 14C yrs BP (532 cal yrs BP or 0.53 ka), and between ca. 860 and ca. 1000 14C yrs BP (725 and 857 cal yrs BP or 0.73 ka and 0.86 ka; Fig. 10). These soils are all present within the representative stratigraphic section Br08-3a (Fig. 3), although dates were also obtained from augered materials in 10-A8 not detailed here. Within this region, soil development often indicates greater than average rainfall received in incremental amounts during the

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

12

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

Fig. 8. Landsat image of Shumba area left, augered 10-A23 sediments on right (surface lower left, reverse sequence on middle bag, 392 cmbs).

Table 6 Shumba 10-A23 stratigraphic profile description. Stratum

Depth (cm)

Soil horizon

Field texture

Color

Description

Dep/Landform

1

0e69 69e125

A Bw

sand clayey sand

10YR4/2d; 10YR3/2m 10YR4/3d; 10YR3/2m

aeolian aeolian

125e178

BC

sandy loam

10YR4/3d; 10YR3/2m

2

178e213

C

sandy loam

10YR6/3d; 10YR4/2m

3

213e392

C2

loamy sand

10YR8/2d; 10YR5/3m

NE w/10% HCl; poorly sorted; loose NE w/10% HCl; wk irregular peds w/silica; oxidized NE w/10% HCl; wk irregular peds w/silica; clear round quartz sand NE w/10% HCl; redox mottling; clear round quartz sand; gley STE w/10% HCl; silica nodules w/CaCO3 coatings; clear round quartz sand

aeolian aeolian aeolian

d-dry, m-moist, HCl e hydrochloric acid, NE e noneffervescent; VSE e very slightly effervescent; SLE e slightly effervescent, STE e strongly effervescent, VE e violently effervescent, veg e vegetation, wk e weak, sab e sub-angular blocky, lb e lower boundary, org e organic, mod e moderate, ox e oxidation, vs e very strong, vvs e very very strong, s e strong, vw e very weak.

growing season. Therefore, these periods of soil development are interpreted as relatively mesic intervals when the landscape stabilized, although soil carbon isotopes can also reflect relative hot and dry periods that encourage the growth of grasses. Soils are often separated by flood and channel deposits that indicate climatic variability and possible drought.

Within the upper Bumbusi River, a piece of charcoal dating 573 C yrs BP (539 cal yrs BP or 0.54 ka) was found on top of a wellcemented bedded gravel that contained a secondarily-deposited levallois-like (Middle Stone Age) lithic flake. Although this profile is not typical and was not detailed herein, these temporal indicators suggest that sediments from a span of time of perhaps 30,000 years 14

Fig. 9. Landsat image of Josivanini area left, JosSu2 profile right (250 cm deep).

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

13

Table 7 Josivanini Dune (Jos-SU2) stratigraphic profile description. Stratum

Depth (cm)

Soil horizon

Field texture

Color

Description

Dep/Landform

1

0e10

A

loamy sand

10YR4/4d; 10YR3/2m

aeolian dune

10e90

C

loamy sand

5YR5/4d&m

2

90e160

C2

loamy sand

5YR5/6d; 5YR4/4m

3

160e193

C3

loamy sand

5YR5/8d; 2.5YR4/4m

4

193e200

C4

loamy sand

5YR5/8d; 2.5YR4/4m

NE w/10% HCl; loose med rounded quartz sand; many charcoal flecks; gradual lb loose; NE w/10% HCl; fine to medium round quartz sand w/sesquioxides & clay films; many termite runs and roots; gradual lb NE w/10% HCl; relatively compact; dense charcoal lense; many termite runs and root casts; abrupt lb NE w/10% HCl; loose fine to medium round quartz grains with sesquioxides & clay films; gradual lb NE w/10% HCl; charcoal throughout; loose fine to medium round quartz grains with sesquioxides & clay films

aeolian dune

aeolian dune aeolian dune aeolian dune

d-dry, m-moist, HCl e hydrochloric acid, NE e noneffervescent; VSE e very slightly effervescent; SLE e slightly effervescent, STE e strongly effervescent, VE e violently effervescent, veg e vegetation, wk e weak, sab e sub-angular blocky, lb e lower boundary, org e organic, mod e moderate, ox e oxidation, vs e very strong, vvs e very very strong, s e strong, vw e very weak.

are missing at that location. Another illustration of this is an archaeological site upstream from Br08-3's sampled location, where Late Iron Age ceramics are intermingled with a Middle Stone Age lithic assemblage. At Br08-3a (Fig. 3), although early and midHolocene sediments are also missing, sediments began

accumulating on top of Pleistocene-age sediments (undated, but red in color with strong cementation) around 1500 14C yrs BP (1352 cal yrs BP or 1.35 ka) based on radiocarbon dates obtained from the lower profile. The youngest dated soil in the Bumbusi River Basin is ca. 400 14C

Dtma09-1 10-A23

cmbs

cmbs

Br08-3a cmbs

Kr08-3 50

cmbs

50

50

408+/-34 BP 100

50

100

100

380+/-35 BP 100

150

150

Josivanini Dune JosSu2

150

549+/-34 BP 200

150

200

Sinamatella River Floodplain Sin08-L3

cmbs 200

1009+/-34 BP

cmbs

250

250

200

50 250 300

50

300 250 100 300

1466+/-35 BP

2206+/-39 BP 300

350

Kapula River Floodplain

2574+/-38 BP

350

100

350

Deteema River Floodplain

150

3580+/-38 BP 150 200

415+/-35 BP

Shumba Pans

400

Bumbusi River Floodplain

200 250

Fig. 10. Stratigraphic correlations across the study area. All dates are uncalibrated radiocarbon ages.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

14

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

yrs BP (408 cal yrs BP or 0.4 ka) in age. This soil is overlain by another, even more recent soil that is likely several hundred years old and is comprised of fine-grained overbank alluvium. The current channel is 4 m below this ca. 200 year old overbank deposit. What this suggests is that at least 2 m of downcutting has occurred during the past two centuries within the Bumbusi River. Based on historic accounts of Kapula Vlei, we know this to be true there: 1e2 m of downcutting has occurred since the 1960s (Anon. National Park staff, personal communication 2008; Fig. 5). At Deteema, abundant material has accumulated over the past 2200 14C yrs (2164 cal yrs BP or 2.2 ka); including standing water deposits and channel fill as represented at Dtma09-1 (Fig. 6). Deposits at Sinamatella are more conservative, and a date of 3580 14C yrs BP (3819 cal yrs BP or 3.8 ka) from Sin08-L3 (Fig. 7) on a soil around 140 cmbs suggests relatively low sedimentation rates and little potential for buried preservation of agropastoralists' cultural deposits, particularly within the channel gravels between 57 and 85 cmbs. Within Josivanini Dune, horizonation is slight but detectable. Radiocarbon ages of 2574 14C yrs BP (2606 cal yrs BP or 2.6 ka; Table 1) on chunks of charcoal from JosSu2 (Fig. 9) and 2730 14C yrs BP (2805 cal yrs BP or 2.8 ka; Haynes, unpublished results) from nearby JosSu1 (not detailed here) testify to past wildfires. These dates, combined with re-activation of the dune and deposition of around 100 cm of aeolian sand, suggest that drought and volatility characterize this time. Another date of charcoal was stratigraphically reversed in JosSu2 (Fig. 9), but given that it is directly dated to 415 14C yrs BP (445 cal yrs BP or 0.45 ka), we know that a wildfire also burned here during that time, even while soils were forming within the loamy sediments of Bumbusi River and Kapula Vlei. This may reflect increased fuel loads coupled with normal dry season aridity and ignition. Within the Shumba deposits represented by augered sample 10-A23 (Fig. 8), no indicators of environmental change are evident, and soil development has kept pace with the aeolian accumulation above the water-logged sands within this sink. However, no dates were obtained from these sampled sediments. Characterization of changing vegetation patterns due to shifting environmental conditions can be calculated using the d13C values from the radiocarbon analyses. Within the Bumbusi River area, conditions around 1466 14C yrs BP (1324 cal yrs BP or 1.3 ka) are relatively mesic (60% C4 plants; Br08-3a; Fig. 3; Table 1) compared to 863 14C yrs BP (728 cal yrs BP or 0.73 ka; Auger 10-A8; Table 1), when the percent of C4 grasses increases to 79% (Table 1). However, the 863 14C yrs BP (728 cal yrs BP or 0.73 ka) layer also had a lithic flake in the auger bucket, and the increase in the percentage of C4 plants here may reflect farming of C4 crops rather than a climatic signature. By 549 14C yrs BP (529 cal yrs BP or 0.53 ka), conditions again became more mesic, with C4 plants dropping down to 51% of the vegetation cover (Br08-3a, Fig. 3; Table 1). Ca. 400 14C yrs BP (408 cal yrs BP or 0.41 ka) the Br08-3a (Fig. 3; Table 1) soil is 65e75% C4 plants, suggesting warmer conditions than at 549 14C yrs BP (529 cal yrs BP or 0.53 ka), but more mesic conditions than at 863 14C yrs BP (728 cal yrs BP or 0.73 ka). At Kapula Vlei, C4 plants dominate throughout (as they do today), making up 79% of the plant cover at 1009 14C yrs BP (859 cal yrs BP or 0.86 ka) and 83% of plant cover at 380 14C yrs BP (396 cal yrs BP or 0.4 ka) at Kr08-3 (Fig. 4). At Deteema and Sinamatella, we unfortunately do not have a sequence to compare change over time, but at Dtma09-1 locale (Fig. 6) the 2206 14C yrs BP (2176 cal yrs BP or 2.2 ka) sediments are comprised of 78% C4 plants (Table 1), while in the Sin08-L3 (Fig. 7) 3580 14C yrs BP (3819 cal yrs BP or 3.8 ka) soil, C4 plants made up only 46% of the vegetation (Table 1). Given the characteristic vegetation of these areas today, these values reflect the different substrates, topography, and vegetation communities of their respective systems and

so cannot be compared to analyze broad scale climatic changes. 5.2. Archaeological implications Geomorphological investigations in Hwange National Park have revealed several trends important in our search for the earliest evidence of agropastoralists in the area. Although pockets of preserved sediments and soils of this age exist, such as at Kapula Vlei, they are not widespread, and may be missing completely from many basins. Based on Scudder's (1971, 1976) study of the Gwembe Tonga people, key environmental variables that determine agropastoralist success include: the amount of precipitation, variance in sediment textures and topographic positions for fields in relatively close proximity to each other, and, most importantly, loam alluvial terraces. Given these criteria, Shumba, Sinamatella, and Josivanini would not have been attractive locations for the cultivation of plants by early farmers. Insufficient information has been collected at Deteema to judge whether or not this area would have drawn early agropastoralists, but given that reliable surface water seems to be available year round here, and that it has deposits of appropriate age, it should be investigated further. The Bumbusi River drainage basin may have drawn early agropastoralists to the area given the relatively reliable waters, loamy sediments, varied textures, and elevational gradients; however, most of the uplands are too sandy for successful farming. Also, before around 1500 14C yrs BP (1352 cal yrs BP or 1.35 ka), sediments were scoured from the basin, taking any evidence with them. In addition, excavations in nearby Impala and Ngabaa Rockshelters did not find any evidence of Early Iron Age cultural markers, such as combstamped ceramics, within the hunter-gatherer assemblage dating to this time (Wriston and Haynes, unpublished results). Given the evidence presented, investigations seeking archaeological materials created by the earliest agropastoralists in the region should focus in the Kapula Vlei area and other localities with similar characteristics elsewhere in northwestern Zimbabwe. Furthermore, we speculate that sites similar to Deteema also have the potential to reveal as yet undiscovered Early Iron Age archaeological deposits. 5.3. Overview Within Kapula Vlei, prehistoric pits dating to between 1990 and 1790 14C yrs BP (1901 and 1654 cal yrs BP or 1.9 ka and 1.65 ka; Wriston and Haynes, unpublished results) were apparently excavated into the same soil as those dated ca. 1200 14C yrs BP (1061 cal yrs BP or 1.1 ka), suggesting surficial stability over this time. However, preceding the earliest farmer's occupation, the wildfire and dune activity at Josivanini beginning around 2730 14C yrs BP (2805 cal yrs BP or 2.8 ka; Haynes, unpublished results) and extending later than 2574 14C yrs BP (2606 cal yrs BP or 2.6 ka) show that the climate was volatile. Volatility is also evident in the scouring of upper, bedrock-constrained tributaries such as the Bumbusi River. Fortunately, Kapula Vlei is within a broad basin with abundant fill, making degradation and erosion less likely. During modern times, the lack of surface water during the dry season (outside of the pumped Masuma Dam) would make year-round occupation of Kapula Vlei difficult, even if crops grew well here during the wet season and a second cropping from the clay loam alluvial terrace was possible. It follows that at ca. 1200e1100 and 1990e1790 14C yrs BP (ca. 1061e963 and 1901-1654 cal yrs BP, or ca. 1.1 ka and ca. 1.8 ka), less distinct seasonality in rainfall distribution and more mesic conditions would have been necessary to successfully farm the area and maintain villages. We have not yet discovered any direct archaeological evidence of early livestock herding (pastoralism) associated with these

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

settlements. Plug (1997) suggested that agricultural villages in northeastern Zimbabwe did not keep domestic animals due to disease vector prevalence, and instead relied on hunting, most particularly of buffalo. This adaptation may have been preferred by early agriculturalists in what is now Hwange National Park, where game is plentiful. In 1928, when the precursor to today's Hwange National Park, the Wankie Game Reserve, was created, both agropastoralists and hunter-gatherers were still using the area. This suggests that people with very different resource procurement systems may have minimized risks and benefited by participating in interdependent trade and social interactions. This is an efficient strategy in a drought-prone area with numerous factors largely out of anthropogenic control (cf. Ellis and Swift, 1988; Homewood, 1994; Dublin, 1995; Sullivan, 1999; Laris, 2003), such as disease outbreaks, drought, and famine. Perhaps for these reasons, as Smith (2000) and others have discussed, there was never a complete abandonment of hunting-and-gathering or foraging activities even as new people or new ideas were introduced into the area. Hybrid ways of life with pastoro-foragers (Elphick, 1977; Wilmsen, 1989, 1991; Kinahan, 1991), pastoro-fishermen, pastoro-miners, agropastoralists, and specialized pastoralists and agriculturalists all coexisted and often traded foodstuff and other services (e.g., Goula in Clark, 1976). 6. Conclusions As this study shows, understanding the geomorphic system in Hwange National Park is essential to focus the search for the earliest food producers in the area. Tremendous landscape evolution has occurred over the past three millennia due to drought that likely began around 2730 14C yrs BP (2805 cal yrs BP or 2.8 ka) and continued for several hundred years. This drought and others may have depleted disease vectors and woodlands within the study area and caused downcutting and scour of drainage basin fill. By the time the earliest agropastoralists arrived ca. 1990e1790 14C yrs BP (1901-1654 cal yrs BP or 1.9e1.65 ka), conditions had become more mesic, with sediment fill beginning to accumulate and soil development beginning as early as 1466 14C yrs BP (1324 cal yrs BP or 1.3 ka), even in bedrock-constrained upper tributaries such as the Bumbusi River. Periodic droughts after this time are reflected by stratigraphic unconformities and by soil carbon isotopic signatures that indicate relatively warm/dry conditions at around 0.73 ka and 0.41 ka separated by a mesic interval around 0.53 ka. Although early agropastoralists' land use in this volatile climatic regime would have required flexibility and the probable seasonal supplementation by hunting and gathering, the relatively abundant Later Iron Age settlements in the study area testify to this lifeway's resilience. However, evidence of the earliest farmers in the region will not be found in many areas they undoubtedly inhabited due to regular basin scour and erosion. Based on our findings, researchers looking for the earliest agropastoralists in the region should target areas with broad basin fill, meandering drainages, loamy sediments, predictable surface waters, and topographic relief. Acknowledgements The authors wish to thank two anonymous reviewers for their constructive comments that improved this manuscript. Numerous individuals and organizations contributed to the success of this project, including: Kathryn Krasinski, Jarod Hutson, Simon and Violah Makuvaza, Mike Scott, Bernard Sibanda, the late Shadreck Muleya, Verla Jackson, Rob Burrett, and Paul Hubbard; the National Museums and Monuments of Zimbabwe, especially the Executive Director, Dr. G. Mahachi, the former Director of Research, Mr. P.

15

Taruvinga, and the former Western Regional Director Dr. Darlington Munyikwa; the Directorate and staff of the Parks and Wildlife Management Authority of Zimbabwe, especially Park Rangers Anderson Munkuli and Lucky Lengu, and other Park employees too numerous to name; and the many staff at the Zimbabwe Natural History Museum in Bulawayo. Funding was provided by The National Science Foundation (Grant BCS-0741877), The Wenner-Gren Foundation (Grant 7789), UNR Graduate Student Association Research Grants, The Geological Society of America Research Grant, The Douglas C. Kellogg Fellowship of the Society for American Archaeology, The William Self Fellowship of the Anthropology department at the University of Nevada, Reno, and the Am-Arcs of Nevada. References Ambrose, S., Sikes, N.E., 1991. Soil carbon isotope evidence for Holocene habitat change in the Kenya Rift Valley. Science 253 (5026), 1402. Ambrose, S.H., DeNiro, M.J., 1989. Climate and habitat reconstruction using stable carbon and nitrogen isotope ratios of collagen in prehistoric herbivore teeth from Kenya. Quat. Res. 31 (3), 407e422. Ashley, G.M., Ndiema, E.K., Spencer, J.Q.G., Harris, J.W.K., Kiura, P.W., 2011. Paleoenvironmental context of archaeological sites, implications for subsistence strategies under Holocene climate change, Northern Kenya. Geoarch. Int. J. 26 (6), 810e837. Bell, R.H.V., 1982. The effect of soil nutrient availability on community structure in African ecosystems. Ecol. Stud. 42, 193e216. Beaumont, P.B., Van Zinderen Bakker, E.M., Vogel, J.C., 1984. Environmental changes since 32 000 BP at Kathu Pan, northern cape. In: Vogel, J.C. (Ed.), Late Cainozoic Palaeoclimates of the Southern Hemisphere. A.A. Balkema, Rotterdam, Netherlands, pp. 329e338. Botha, G., Porat, N., 2007. Soil chronosequence development in dunes on the southeast African coastal plain, Maputaland, South Africa. Quat. Int. 162e163, 111e132. Bird, M.I., Veenendaal, E.M., Lloyd, J.J., 2004. Soil carbon inventories and d13C along a moisture gradient in Botswana. Glob. Change Biol. 10 (3), 342e349. Blümel, W.D., Eitel, B., Lang, A., 1998. Dunes in southeastern Namibia: evidence for Holocene environmental changes in the southwestern Kalahari based on thermoluminescence data. Palaeogeogr. Palaeoclimatol. Palaeoecol. 138 (1e4), 139e149. Bousman, C.B., 1998. The chronological evidence for the introduction of domestic stock into southern Africa. Afr. Archaeol. Rev. 15 (2), 133e150. Brain, C.K., 1969. New evidence for climatic change during middle and late Stone age times in rhodesia. South Afr. Archaeol. Bull. 24 (95/96), 127e143. Brain, C.K., Brain, V., 1977. Microfaunal remains from Mirabib: some evidence of palaeocological changes in the Namib. Madoqua 10, 285e293. Burrough, S.L., Thomas, D.S.G., Singarayer, J.S., 2009. Late quaternary hydrological dynamics in the middle Kalahari: forcing and feedbacks. Earth Sci. Rev. 96 (4), 313e326. Butzer, K.W., 1984a. Archeogeology and Quaternary environment in the interior of southern Africa. In: Klein, R.G. (Ed.), Southern African Prehistory and Paleoenvironments. A.A. Balkema, Rotterdam/Boston, pp. 1e64. Butzer, K.W., 1984b. Late quaternary environments in South Africa. In: Vogel, J.C. (Ed.), Late Cainozoic Palaeoclimates of the Southern Hemisphere. A.A. Balkema, Rotterdam, Netherlands, pp. 235e264. Catt, J.A., 1990. Field Recognition, description and spatial relationships of paleosols. Quat. Int. 6, 2e20. Cecchi, G., Mattioli, R.C., Slingenbergh, J., De La Rocque, S., 2008. Land cover and tsetse fly distributions in sub-Saharan Africa. Med. Veterinary Entomol. 22, 364e373. Cerling, T.E., Hay, R.L., 1986. An isotopic study of paleosol carbonates from Olduvai Gorge. Quat. Res. 25 (1), 63e78. Chapman, J., Tabler, E.C., 1971. Travels in the interior of South Africa, 1849-1863: hunting and trading journeys from natal to walvis bay and visits to lake ngami and victoria falls. In: South African Biographical and Historical Studies, vol. 2. A.A. Balkema, Cape Town. Chase, B., 2009. A record of rapid Holocene climate change preserved in hyrax middens from southwestern Africa. Geology 37, 703e706. Chase, B.M., Meadows, M.E., Carr, A.S., Reimer, P.J., 2010. Evidence for progressive Holocene aridification in southern Africa recorded in namibian hyrax middens: implications for African monsoon dynamics and the “African humid period”. Quat. Res. 74 (1), 36e45. Clark, J.D., 1976. Prehistoric populations and pressures favoring plant domestication in Africa. In: Harlan, J.R., de Wet, J.M.J., Stemler, A.B.L. (Eds.), Origins of African Plant Domestication, The Hague, pp. 67e105. Cooke, H.J., 1984. The evidence from northern Botswana of climate change. In: Vogel, J.C. (Ed.), Late Caenozoic Palaeoclimates of the Southern Hemisphere. Balkema, Rotterdam, pp. 265e278. Cruz-Uribe, K., 1983. The mammalian fauna from Redcliff Cave, Zimbabwe. South Afr. Archaeol. Bull. 38 (137), 7e16.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

16

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17

Deacon, J., 1974. Patterning in the radiocarbon dates for the wilton/smithfield complex in southern Africa. South Afr. Archaeol. Bull. 29 (113/114), 3e18. Deacon, J., 1984. The Later Stone age of Southernmost Africa. Cambridge Monographs in African Archaeology 12. B.A.R. International Series 213, Oxford. Deacon, J., Thackeray, J.F., 1984. Late Pleistocene environmental changes and implications for the archaeological record in southern Africa. In: Vogel, J.C. (Ed.), Late Cainozoic Palaeoclimates of the Southern Hemisphere. Balkema, Rotterdam. Denbow, J., 1984. Cows and Kings: a spatial and economic analysis of a hierarchical Early Iron Age settlement system in eastern Botswana. In: Hall, M., Avery, G., Avery, D.M., Wilson, M.L., Humphreys, A.J.B. (Eds.), Frontiers: Southern African Archaeology Today. British Archaeological Research Monographs, Cambridge, pp. 24e39. Denbow, J., 1986. A new look at the later prehistory of the Kalahari. J. Afr. Hist. 27 (1), 3e28. DeNiro, M.J., Epstein, S., 1978. Influence of diet on the carbon isotopes in animals. Geochim. Cosmochim. Acta 42, 495e506. Dublin, H., 1995. Vegetation dynamics in the Serengeti-Mara ecosystem: the role of elephants, fire, and other factors. In: Sinclair, A., Arcese, P. (Eds.), Serengetic II: Dynamics, Management and Conservation of an Ecosystem. Chicago University Press, Chicago. Ellis, J., Swift, D., 1988. Stability of African pastoralist ecosystems: alternate paradigms and implications for development. J. Range Manag. 41, 450e459. Elphick, R., 1977. Kraal and Castle: Khoikhoi and the Founding of White South Africa. Yale University Press, New Haven and London. Endfield, G.H., Ryves, D.B., Mills, K., Berrang-Ford, L.E.A., 2009. 'The gloomy forebodings of this dread disease', climate, famine and sleeping sickness in East Africa. Geogr. J. 175 (3), 181e195. Eze, P.N., Meadows, M.E., 2015. Geochemistry and palaeoclimatic reconstruction of a palaeosol sequence at Langebaanweg, South Africa. Quat. Int. 376, 75e83. Farquhar, G.D., Ehleringer, J.R., Hubick, K.T., 1989a. Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol. 40 (1), 503e537. Farquhar, G.D., Hubick, K.T., Condon, A.G., Richards, R.A., 1989b. Carbon isotope fractionation and plant water-use efficiency. In: Rundel, P.W., Ehleringer, J.R., Nagy, K.A. (Eds.), Stable Isotopes in Ecological Research. Springer-Verlag, New York, pp. 21e40. Finch, J.M., Hill, T.R., 2008. A late Quaternary pollen sequence from Mfabeni Peatland, South Africa: reconstructing forest history in Maputaland. Quat. Res. 70 (3), 442e450. Gifford-Gonzalez, D., 2000. Animal disease challenges to the emergence of pastoralism in sub-Saharan Africa. Afr. Archaeol. Rev. 17, 95e139. Goodfriend, G.A., 1999. Terrestrial stable isotope records of Late Quaternary paleoclimates in the eastern Mediterranean region. Quat. Sci. Rev. 18 (4e5), 501e513. Haynes, G., Klimowicz, J., 1998. Archaeological sites. In: Hwange National Park Management Plan 1999-2003, Pp. Annex 7, Section 7.2. Zimbabwe Department of National Parks and Wild Life Management, Harare. Haynes, G., Klimowicz, J., 2005. Possible evidence for early woodland burning by agropastoralists in northwestern Zimbabwe at 2,000 years ago. Nyame Akuma 64, 72e77. Haynes, G., Makuvaza, S., Wriston, T., 2011. The Bumbusi engravings and paintings in Hwange National Park, Zimbabwe: preliminary results of recording and rockshelter excavations. Zimbabwean Prehistory 29, 12e25. Heinrich, J., Moldenhauer, K.M., 2002. Climatic and anthropogenic induced landscape degradations of West African dry Savanna environments during the later Holocene. Quat. Int. 93e94, 127e137. Holliday, V.T., 2004. Soils in Archaeological Research. Oxford University Press, Oxford. Holmgren, K., Karlen, W., Lauritzen, S.E., Lee-Thorp, J.A., Partridge, T.C., Piketh, S., Repinski, P., Stevenson, C., Svanered, O., Tyson, P.D., 1999. A 3000-year highresolution stalagmite based record of palaeoclimate for northeastern South Africa. Holocene 9 (3), 295e309. Holub, E., 1975. Emil Holub's travels north of the Zambezi 1885-6. In: Johns, C. (Ed.), The Institute for African Studies University of Zambia, 2 vol. s. Manchester University Press, Lusaka. Translated by. Homewood, K., 1994. Pastoralism, environment and development in east African rangelands. In: Zaba, B., Clarke, J. (Eds.), Environment and Population Change, vol. ORDINA Editions. Cerouaux, Liege. Huffman, T., 2007. Handbook to the Iron Age: the Archaeology of Pre-colonial Farming Societies in Southern Africa. University of KwaZulu-Natal Press, Scottsville, South Africa. Johnson, B.J., Fogel, M.L., Miller, G.H., 1998. Stable isotopes in modern ostrich eggshell: a calibration for paleoenvironmental applications in semi-arid regions of southern Africa. Geochim. Cosmochim. Acta 62 (14), 2451e2461. Jones, P.D., Briffa, K.R., Osborn, T.J., Lough, J.M., van Ommen, T.D., Vinther, B.M., Luterbacher, J., Wahl, E.R., Zwiers, F.W., Mann, M.E., Schmidt, G.A., Ammann, C.M., Buckley, B.M., Cobb, K.M., Esper, J., Goosse, H., Graham, N., Jansen, E., Kiefer, T., Kull, C., 2009. High-resolution palaeoclimatology of the last millennium: a review of current status and future prospects. Holocene 19 (1), 3e49. Kinahan, J., 1991. Pastoral Nomads of the Central Namib Desert: the People History Forgot. Namibia Archaeological Trust. New Namibia Books, Windhoek. Lancaster, N., 1988. Development of linear dunes in the southwestern Kalahari, southern Africa. J. Arid Environ. 14, 233e244. Lancaster, N., 1989. Late quaternary palaeoenvironments in the southwestern

Kalahari. Palaeogeogr. Palaeoclimatol. Palaeoecol. 70, 367e376. Lancaster, N., 2002. How dry was dry? Late Pleistocene palaeoclimates in the Namib Desert. Quat. Sci. Rev. 21 (7), 769. Laris, P., 2003. Grounding environmental narratives. In: Moseley, W., Logan, B. (Eds.), African Environment and Development: Rhetoric, Programs, Realities. Ashgate Publishing Ltd., Aldershot. Levin, N.E., Cerling, T.E., Passey, B.H., Harris, J.M., Ehleringer, J.R., 2006. A stable isotope aridity index for terrestrial environments. Proc. Natl. Acad. Sci. U. S. A. 103 (30), 11201e11205. Limbrey, S., 1975. Soil Science and Archaeology. Studies in Archaeological Science. Academic Press, London, New York, and San Francisco. Long, S.P., 1999. Environmental responses. In: Sage, R.F., Monson, R.K. (Eds.), C4 Plant Biology. Academic Press, New York, pp. 215e249. Martin, C.W., Johnson, W.C., 1995. Variation in radiocarbon ages of soil organic matter fractions from late Quaternary buried soils. Quat. Res. 43 (2), 232e237. Martin, J.H., Leonard, W.H., Stamp, D.L., 1976. Principles of Field Crop Production, third ed. Macmillan Publishing Co., Inc., New York. Mazvimavi, D., 2010. Investigating changes over time of annual rainfall in Zimbabwe. Hydrol. Earth Syst. Sci. 14, 2672e2679. Mitchell, P., 2002. The Archaeology of Southern Africa. Cambridge World Archaeology. Cambridge University Press, Cambridge. Norman, M.J.T., Pearson, C.J., Searle, P.G.E., 1984. The Ecology of Tropical Food Crops. Cambridge University Press, Cambridge. Norstrom, E., Scott, L., Partridge, T.C., Risberg, J., Holmgren, K., 2009. Reconstruction of environmental and climate changes at Braamhoek wetland, eastern escarpment South Africa, during the last 16,000 years with emphasis on the Pleistocene-Holocene transition. Palaeogeogr. Palaeoclimatol. Palaeoecol. 271 (3/4), 240e258. Nyamapfene, K., 1991. Soils of Zimbabwe. Nehanda Publishers, Harare. O'Brien, S.R., Mayewski, P.A., Meeker, L.D., Meese, D.A., Twickler, M.S., Whitlow, S.I., 1995. Complexity of Holocene climate as reconstructed from a Greenland ice core. Science 270 (5244), 1962e1964. O'Connor, P.W., Thomas, D.S.G., 1999. The timing and environmental significance of late Quaternary linear dune development in western Zambia. Quat. Res. 52 (1), 44e55. Plug, I., 1997. Early Iron age buffalo hunters on the kadzi river, Zimbabwe. Afr. Archaeol. Rev. 14 (2), 85e105. Robbins, L.H., Murphy, M.L., Campbell, A.C., Brook, G.A., Reid, D.M., Haberyan, K.A., Downey, W.S., 1998. Test excavations and reconnaissance palaeoenvironmental work at Toteng, Botswana. South Afr. Archaeol. Bull. 53 (168), 125e132. Robbins, L.H., Campbell, A.C., Murphy, M.L., Brook, G.A., Srivastava, P., Badenhorst, S., 2005. The advent of herding in southern Africa: early AMS dates on domestic livestock from the Kalahari Desert. Curr. Anthropol. 46 (4), 671e677. Robinson, K.R., 1966. The iron age site in kapula vlei, near the Masuma dam. Wankie game reserve. Rhodesia. Arnoldia 39 (2), 1e7. Rogers, C.M.L., 1993. A Woody Vegetation Survey of Hwange National Park. Department of National Parks and Wild Life Management, Harare, Zimbabwe. Rust, U., Schmidt, H.H., Dietz, K.R., 1984. Palaeoenvironments of the present day arid south western Africa 30,000-5,000 BP; results and problems. Palaeoecol. Afr. Surround. Isl. Antarct. 16, 109e148. Sadr, K., 2008. An ageless view of first millennium AD southern African ceramics. J. Afr. Archaeol. 6 (1), 103e129. Sage, R.F., Wedin, D.A., Li, M., 1999. The biogeography of C4 photosynthesis: patterns and controlling factors. In: Sage, R.F., Monson, R.K. (Eds.), C4 Plant Biology. Academic Press, New York, pp. 313e373. Schoeneberger, P.J., Wysocki, D.A., Benham, E.C., Broderson, W.D. (Eds.), 2002. Field Book for Describing and Sampling Soils, Version 2.0. Natural Resources Conservation Service, National Soil Survey Center, U.S. Department of Agriculture, Lincoln, NE. Scott, L., Holmgren, K., Talma, A.S., Woodborne, S., Vogel, J.C., 2003. Age interpretation of the Wonderkrater spring sediments and vegetation change in the Savanna Biome, Limpopo province, South Africa. South Afr. J. Sci. 99, 484e488. Scudder, T., 1971. Gathering Among African Woodland Savannah Cultivators, a Case Study: the Gwembe Tonga. Zambian Papers. University of Zambia and Manchester University Press, Manchester. Scudder, T., 1976. Social anthropology and the reconstruction of prehistoric land use systems in tropical Africa: a cautionary case study from Zambia. In: Harlan, J.R., De Wet, J.M.J., Stemler, A.B.L. (Eds.), Origins of African Plant Domestication. Mouton Publishers, Paris, pp. 357e381. Segalen, L., Renard, M., Pickford, M., Senut, B., Cojan, I., Le Callonnec, L., Rognon, P., 2002. Environmental and climatic evolution of the Namib Desert since the middle Miocene: the contribution of carbon isotope ratios in ratite eggshells. Comptes Rendus Geosci. 334 (12), 917. Shaw, P., 1985a. The desiccation of Lake Ngami: an historical perspective. Geogr. J. 151, 318e326. Shaw, P., 1985b. Late quaternary landforms and environmental change in northwest Botswana: the evidence of lake ngami and the mababwe depression. Trans. Inst. Br. Geogr. 10, 333e346. Shaw, P.A., Cooke, H.J., 1986. Geomorphic evidence for the late Quaternary palaeoclimates of the middle Kalahari of northern Botswana. CATENA 13 (4), 349e359. Sithole, R.T., 1994. Geological Map of Zimbabwe 1:1,000,000. Zimbabwe Geological Survey, Harare. Sithole, A., Murewi, C.T.F., 2009. Climate variability and change over southern Africa: impacts and challenges. Afr. J. Ecol. 47, 17e20.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010

T. Wriston, G. Haynes / Quaternary International xxx (2016) 1e17 Smith, A.B., 2000. The origins of domesticated animals of southern Africa. In: Blench, R., MacDonald, K. (Eds.), The Origins of Domesticated Animals of Southern Africa. UCL Press, London, pp. 222e238. Stager, J.C., 1988. Environmental changes at lake cheshi, Zambia since 40,000 yrs BP. Quat. Res. 29, 54e65. Still, C.J., Powell, R.L., 2010. Continental-scale distributions of vegetation stable carbon isotope ratios. In: West, J.B., Bowen, G.J., Dawson, T.E., Tu, K.P. (Eds.), Isoscapes: Understanding Movement, Pattern, and Process on Earth through Isotope Mapping. Springer, New York, pp. 179e194. Stokes, S., Haynes, G., Thomas, D.S.G., Horrocks, J.L., Higginson, M., Malifa, M., 1998. Punctuated aridity in southern Africa during the last glacial cycle: the chronology of linear dune construction in the northeastern Kalahari. Palaeogeogr. Palaeoclimatol. Palaeoecol. 137 (3e4), 305e322. Stuiver, M., Reimer, P.J., Reimer, R., 2013. CALIB 7.0. http://calib.qub.ac.uk/calib/. Sullivan, S., 1999. The impacts of people and livestock on topographically diverse open wood and shrub-lands in arid north-west Namibia. Cimbebasia 15, 1e28. Talma, A.S., Vogel, J.C., 1992. Late quaternary paleotemperatures derived from a speleothem from cango caves, cape province, South Africa. Quat. Res. 37 (2), 203e213. Thomas, D.S.G., Burrough, S.L., 2012. Interpreting geoproxies of late Quaternary climate change in African drylands: implications for understanding environmental change and early human behaviour. Quat. Int. 253 (0), 5e17. Thomas, D.S.G., Stokes, S., Shaw, P.A., 1997. Holocene aeolian activity in the southwestern Kalahari Desert, southern Africa: significance and relationships to latepleistocene dune-building. Holocene 7 (3), 273e281. Thomas, M.F., 2008. Understanding the impacts of Late Quaternary climate change in tropical and sub-tropical regions. Geomorphology 101 (1e2), 146e158. Van Zinderen Bakker, E.M., 1976. Paleoecological background in connection with the origin of agriculture in Africa. In: Harlan, J.R., De Wet, J.M.J., Stemler, A.B.L. (Eds.), Origins of African Plant Domestication. Mouton Publishers, Paris, pp. 43e63. Van Zinderen Bakker, E.M., 1982. Pollen analytical studies of the wonderwerk cave,

17

South Africa. Pollen Spores 24, 235e250. Van Zinderen Bakker, E.M., 1995. Archaeology and palynology. South Afr. Archaeol. Bull. 50 (162), 98e105. Von Schirnding, Y., Van der Merwe, N.J., Vogel, J.C., 1982. Influence of diet and age on carbon isotope ratios in ostrich eggshell. Archaeometry 24 (1), 3e20. Wadley, L., 1986. Segments of time: a mid-holocene wilton site in the transvaal. South Afr. Archaeol. Bull. 41 (144), 54e62. Walker, N.J., 1995. Late Pleistocene and Holocene hunter-gatherers of the matopos. Stud. Afr. Archaeol. 10, 284. Walkington, H., 2010. Soil science applications in archaeological contexts: a review of key challenges. Earth Sci. Rev. 103 (3e4), 122e134. Wang, L.X., Okin, G.S., Caylor, K.K., Macko, S.A., 2009. Spatial heterogeneity and sources of soil carbon in southern African savannas. Geoderma 149 (3e4), 402e408. Wang, L., D'Odorico, P., Ries, L., Macko, S.A., 2010. Patterns and implications of plant-soil d13C and d15N values in African savanna ecosystems. Quat. Res. 73 (1), 77e83. Wilmsen, E., 1989. Land Filled with Flies. University of Chicago Press, Chicago and London. Wilmsen, E., 1991. Pastoro-foragers to “bushmen”. In: Galaty, J.G., Bonte, P. (Eds.), Herders, Warriors and Traders: Pastoralism in Africa. Westview Press, Boulder, CO. Wriston, T.A., Haynes, G., 2009. Preliminary results of archaeological excavations in Impala and Ngabaa rockshelters, Hwange national park, Zimbabwe. Nyame Akuma 72, 61e66. Wynn, J.G., 2000. Paleosols, stable carbon isotopes, and paleoenvironmental interpretation of Kanapoi, northern Kenya. J. Hum. Evol. 39 (4), 411e432. Wynn, J.G., Bird, M.I., 2008. Environmental controls on the stable carbon isotopic composition of soil organic carbon: implications for modelling the distribution of C3 and C4 plants, Australia. Tellus Ser. B Chemical Phys. Meteorol. 60 (4), 604e621.

Please cite this article in press as: Wriston, T., Haynes, G., Sediments, soils, and the expansion of farmers into a forager's world: A geoarchaeological study of the mid-to-late Holocene in Hwange National Park, Zimbabwe, Quaternary International (2016), http:// dx.doi.org/10.1016/j.quaint.2016.11.010