Sediment mixing in aeolian sandsheets identified and quantified using single-grain optically stimulated luminescence

Quaternary Geochronology 32 (2016) 53e66

Contents lists available at ScienceDirect

Quaternary Geochronology journal homepage: www.elsevier.com/locate/quageo

Research paper

Sediment mixing in aeolian sandsheets identified and quantified using single-grain optically stimulated luminescence Luke Andrew Gliganic a, b, *, Tim J. Cohen b, Michael Slack c, d, James K. Feathers e a

Institute for Geology, University of Innsbruck, Innrain 52, 6020 Innsbruck, Austria GeoQuest Research Centre, School of Earth and Environmental Sciences, University of Wollongong, Northfields Avenue, Wollongong, NSW 2500, Australia c Scarp Archaeology, PO Box 191, Terrey Hills, NSW 2084, Australia d College of Arts, Society and Education, James Cook University, PO Box 6811, Cairns, QLD 4870, Australia e Department of Anthropology, University of Washington, Seattle, WA 98195-3100, USA b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 September 2015 Received in revised form 22 December 2015 Accepted 26 December 2015 Available online 29 December 2015

Post-depositional mixing processes are extremely common and often obscure a record of deposition in dune and sand sheet deposits. We show that the upper half metre of a dune in southeastern Australia is currently being turned over through bioturbation, but that single-grain OSL dating and contextual knowledge can be used to identify and model these modern mixing processes. In the sandy deposits investigated, mixing processes were observed to be acting to a predicable depth of ~50e60 cm. This observation was used to develop a conceptual framework that can be applied to buried deposits and used to temporally constrain the evolution of the landform and quantify rates of mixing. When our mixing zone conceptual framework was combined with the MAM we show that phases of significant dune aggradation occurred at ~29.9, ~18.3, ~10.3 ka, and continued through the Holocene. We also present an approach using single-grain OSL data to estimate downward mixing rates, which show a strong depth dependency and are coherent with previously reported mixing rates. Modern downward mixing rates indicate that the upper ~50 cm (Zone 1) will be completely turned over on millennial time scales. While caution needs to be used when interpreting archaeological and OSL data from bioturbated sandy environments, our results demonstrate that contextual knowledge and single-grain OSL data can resolve mixing processes and contribute to an understanding of landscape evolution. © 2015 Elsevier B.V. All rights reserved.

Keywords: Aeolian dunes OSL dating Bioturbation Sandsheet Mixing rate

1. Introduction Aeolian dunes and sandsheets are common landforms around the world and such features have the potential to inform us about how landscapes respond to changing climate and environmental conditions. The activation and de-activation of dune-building has been previously used to infer changes in the climate and surrounding environment, such as wind behaviour, precipitation, the presence or absence of vegetation, and sediment availability (Hesse et al., 2003; Hesse, 2014; Thomas and Shaw, 2002; Fitzsimmons et al., 2007; Telfer and Thomas, 2007). In addition, sandsheet and dune settings are often rich archaeological archives (e.g., Feathers et al., 2006; Hughes et al., 2014), and analyses of the surrounding

* Corresponding author. Institute for Geology, University of Innsbruck, Innrain 52, 6020 Innsbruck, Austria. E-mail address: [email protected] (L.A. Gliganic). http://dx.doi.org/10.1016/j.quageo.2015.12.006 1871-1014/© 2015 Elsevier B.V. All rights reserved.

soils can yield insights into how humans used the landscape in the past (Walkington, 2010). Thus, understanding depositional processes and post-depositional modification provides an important constraint when using such records for archaeological or palaeoenvironmental reconstruction. However, accurately dating and interpreting these landforms can be difficult due to their susceptibility to post-depositional mixing and turnover processes. Bioturbation by flora and fauna are processes that laterally and vertically displace objects (e.g., grains, gravels, artefacts) from one position to another (Schaetzl and Anderson, 2005), thereby obscuring a record of depositional history and complicating the relationship between buried materials (e.g., archaeological, palaeontological, ecological) with the surrounding substrate (Leigh, 1998; Balek, 2002; Peackock and Fant, 2002). The development of optically stimulated luminescence (OSL) dating over the past three decades has allowed unprecedented insights into the evolution of dunes and sandsheets which often lack material suitable for radiocarbon dating. OSL dating is used to

54

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

estimate the last time a grain of quartz or feldspar was exposed to sunlight and was subsequently buried. Aeolian-transported sediments are ideal for this method, as they are unlikely to suffer from a common problem of insufficient exposure of the grain to sunlight to completely remove its residual signal. OSL dating has allowed the development of Pleistocene dune building chronologies across the world (e.g., Thomas et al., 2003; Fitzsimmons et al., 2013; Hesse, 2014) and has potential to explore how soil mixing is manifested in grain movement. The OSL signal from single grains of quartz has been used to quantify soil movement and turnover rates (Heimsath et al., 2002; Wilkinson and Humphreys, 2005; Stockmann et al., 2013; Johnson et al., 2014; Kristensen et al., 2015) and identify phases of enhanced mixing processes in the past (Gliganic et al., 2015). Bateman et al. (2007a,b) used the OSL signal from single quartz grains to investigate a range of independently dated sandy archaeological deposits that have experienced post-depositional mixing to varying degrees. At one extreme, an unmixed stratified site with homogeneous OSL equivalent dose (De) distributions yielded ages that are consistent with independent age controls (Bateman et al., 2007a). By contrast, highly spread (i.e., overdispersed) and skewed De data from a structureless sandy site indicated sub-surface movement of sand grains and the exhumation of underlying sandstone bedrock (Bateman et al., 2007a). While the later site yielded a stratigraphically coherent OSL chronology, it was inconsistent with robust independent age controls. This supports the conceptual bioturbation schema described by Bateman et al. (2003) that predicts that various pedoturbation and sedimentation rates may result in a coherent OSL-age-depth relationship. Importantly, this indicates that stratigraphic coherence is not necessarily indicative of an accurate OSL chronology. In these cases, the calculated OSL ages will reflect mixing-intensity, not deposition or sedimentation. Bateman et al. (2007b) suggest that bioturbational processes may explain large De scatter observed for structureless linear dune deposits from the Kalahari, questioning the robustness of the existing OSL framework for dune activity in the region (Thomas et al., 2003). Rink et al. (2013) empirically demonstrated that over a seven month period a colony of Florida harvester ants could build a 130 cm deep nest and move ~54,000 sand grains upward, thus significantly affecting De distributions and OSL age estimates for the site. They suggest that an OSL age derived from the grains with the lowest De values, identified using the minimum age model (MAM), yields the most accurate estimates of deposition for the sediments. Indeed, statistical modelling and contextual information can be used to resolve finite burial ages for some mixed sediments (Rodnight et al., 2005; Duller, 2006; David et al., 2007; Jacobs et al., 2008; Araujo et al., 2008; Lombard et al., 2010; Lomax et al., 2011; Cohen et al., 2012; Bueno et al., 2013; Gliganic et al., 2014, 2015; Hanson et al., 2015) and OSL data has been used to quantify rates of soil mixing for in situ weathered soil profiles (Wilkinson and Humphreys, 2005; Stockman et al., 2013). When mixing is significant, however, the depositional signal can be completely obscured, impeding any attempt to estimate a depositional age (Feathers et al., 2006; Bateman et al., 2007a,b; Rhodes et al., 2010; Tribolo et al., 2010; Chazan et al., 2013; Gliganic et al., 2012). In these cases, a new approach is needed to develop a chronology for the evolution of the landform in question and to ascertain the degree and rates of soil/sediment turnover. In this study we investigate significantly mixed dune deposits in a sandsheet complex in the lower Hunter River Valley in southeastern Australia. Despite decades of intensive ecological and archaeological consulting work associated with coal mining in the region, very little is published or known about the age and development of these sandsheet deposits which often contain buried archaeological remains (Hughes et al., 2014). Here, we describe

novel approaches for estimating phases of dune aggradation and calculating downward mixing rates using a conceptual model that is supported by empirical observations of our optical data. Our approach utilises the identification of zero-dose grains mixed into deposits when they are within ~60 cm of the surface (the “mixing zone”) and thus focuses on the downward movement of grains (i.e., “downward mixing”). We discuss the probable mechanics of mixing in this landscape, posit realistic downward mixing rate estimates, and discuss the chronology for the evolution of the dune in the regional context. 2. Context 2.1. Regional setting Dunes and sandsheets cover approximately one third of the Australian continental land surface (Wasson et al., 1988). While linear dunes comprise the bulk of Australian desert dune fields, dunes and sandsheets are also present in the more temperate regions of eastern New South Wales such as the Blue Mountains and the Hunter valley (Story et al., 1963; Hesse et al., 2003). Dune building is not a process that is currently active in non-coastal vegetated locations in temperate southeastern Australia where relict (now vegetated) aeolian features are found. Hesse et al. (2003) showed that aeolian dunes in the Blue Mountains, New South Wales (NSW) formed during the Last Glacial Maximum (LGM) and are composed of locally derived sand from reworked soils formed from the underlying Triassic sandstone bedrock. They suggest that tree cover was vastly reduced during the LGM due to lower atmospheric carbon dioxide levels, modestly lower precipitation and colder temperatures that induced water stress in local vegetation. Such an interpretation of LGM conditions is supported by pollen records from Lake George (Singh and Geissler, 1985) and Barrington Tops (Sweller and Martin, 2001) which show that grasslands and herbfields dominated the vegetation of southeastern Australia during the LGM. This would suggest that conditions in the temperate regions of eastern Australia were more favourable for dune construction and sand sheet development during the Late Pleistocene in contrast to the modern climate. Many rivers in temperate southeastern Australia that drain west of the Great Dividing Range and some that drain east display evidence of source-bordering dunes. Early work by Bowler (1967) reported extensive source bordering dune deposits formed on or adjacent to the Goulburn River system after the LGM (between 20 and 13 ka). Such evidence was further substantiated for other inland draining rivers which have extensive source-bordering dunes associated with large Quaternary palaeochannels (Page et al., 1991, 2009). Aeolian dunes adjacent to rivers but also on high elevation plateaus have been dated using thermoluminescence (TL) by Nott and Price (1991) as being Late Pleistocene in age. All such investigations on source-bordering dune deposits have invoked late Pleistocene climatic conditions suitable for downwind sand accumulation (e.g. increased seasonality and wind strength and/or reduced vegetation). However none of the existing investigations have directly assessed the potential role of bioturbation or mixing on the derived chronologies, representing a potential issue for the derived depositional ages and the associated palaeoclimatic interpretation. The distribution, age and palaeoenvironmental significance of aeolian deposits in other temperate locations is less clear. Hughes et al. (2014) report the results of archaeological and geochronological investigations of bioturbated sand deposits nearby those discussed in this study in the Hunter River Valley in southeastern Australia. They report OSL ages derived from multi-grain aliquots and central age model analyses from single quartz grains of ~88-83

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

ka, 61 ± 3 ka, 53 ± 3 ka, and 11.9 ± 0.6 ka, which they argue reflect the timing of dune growth and possible Pleistocene occupation of the site. These estimates along with the work by Hesse et al. (2003) may suggest that aeolian activity in these temperate regions has occurred intermittently throughout the last glacial cycle. TL-based studies have been used to constrain the history of coastaldraining rivers in NSW, which exhibit broadly synchronous activity during the Late Pleistocene with alluvial deposits dated to MIS 5 (terminating ~75 ka), ~60e40 ka, 30e25 ka, and then resuming fluvial activity 20e15 ka (Nanson et al., 2003). 2.2. Study setting On the high river terraces of Wollombi Brook and the Hunter River (Fig. 1) extensive aeolian sand deposits exist in the form of sandsheets and NWeSE aligned 1e6 m high discontinuous linear dunes (Story et al., 1963; Kovac and Lawrie, 1990). The sand was most likely sourced from the Hunter River and Wollombi Brook (Galloway, 1963; Page, 1971, Fig. 1), the latter of which drains a catchment lithologically dominated by Triassic Sandstones (Rasmus et al., 1969). The bedrock underlying the sandsheets is the Singleton Coal Measures, which contains Permian quartz-lithic sandstone, shale, carbonaceous mudstone, and coal seams (Rasmus et al., 1969). The sandy soils that occur on the fluvial terraces and the lower slopes have been previously grouped by Storey et al. (1963) into two broad soil groups: sandy aeolian regosols and duplex/texture contrast soils (solonetzic or podzolic). They are generally composed of a homogeneous massive sandy A horizon overlying an abrupt transition with a dark coloured underlying and mottled fine-grained B horizon with blocky structure. The deposit investigated in this study is a low relief dune in the lee of a bedrock ridge, where the sandsheet occurs 40e50 m above

55

the valley floor of the Wollombi Brook and Hunter River (Fig. 1). The site is ~4 km downwind (based on the dominant wind direction) and east of Wollombi Brook (Figs. 1 and 2). Lockwood (2009) used ground penetrating radar and soil auguring to determine the depth of unconsolidated sand (up to 3 m deep) overlying the more compact underlying fine-grained unit (Fig. 2). The ecosystem of the surrounding area is defined as the Warkworth Sands Woodland, which is dominated by trees such as Angophora floribunda (Rough-barked Apple) and Banksia integrifolia (Coast Banksia) in addition to Callitris endlicheri and Callitris glaucophylla (Peake, 2006; Bell, 2010). Vegetation on the site is characterised by regrowth of an unknown age but suggests that the site was cleared in the recent past. Local fauna is similar to that for the Sydney basin, including a range of common endemic and introduced vertebrates (e.g., goannas, snakes, rabbits, kangaroos, wallabies, possums, wombats, echidnas) and invertebrates (e.g., termites and ants) (Attenbrow, 2002). Biomixing and mounding rates in other sandstone landscapes in the Sydney region have been shown to be 10e35 times dune accumulation rates (Humphreys and Mitchell, 1983; Humphreys and Field, 1998) and have removed all primary aeolian sedimentary structures in LGM-age dunes in the Blue Mountains (Hesse et al., 2003). 2.3. Methods Seven trenches of 5 m
1 m were excavated in 10 cm spits using an excavator to depths of between 50 and 250 cm. Previous ground penetrating radar (GPR), soil auguring studies (Lockwood, 2009), and LiDar data were used to determine sampling locations. Six trenches were excavated across an east-west transect incorporating a bedrock ridge and a low-relief, vegetated, lee-side dune (Fig. 2). The seventh trench was located 120 m north of the transect in an

Fig. 1. Regional setting. Extent of sandy soils (siliceous sands, soloths, yellow solodic soils, and yellow podzolic soils; Kovac and Lawrie, 1990) and alluvial sediments mapped over Shuttle research topography mission (SRTM) data for the Hunter region, NSW. Note also the location of the study site.

56

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

Fig. 2. Plan view of study site. Transect, trench locations (numbered), and the depth-to-sand results of Lockwood (2009) are shown on a digital elevation model (LiDar data). Inset is a wind rose diagram showing the prevailing wind directions for Williamtown, NSW (79 km southeast of the study site) at 9:00 am (data from the Australian Bureau of Meteorology: http://www.bom.gov.au/climate/averages/wind/selection_map.shtml). Note that the lines point in the direction from which the wind came.

area with no sand cover. Each trench was excavated to the contact between loose sand (A Horizon) and the more compact, underlying sandyeclay unit (B Horizon). All trenches were logged, photographed, and sampled for grain size analysis (using a Malvern Laser Mastersizer) and OSL dating. OSL sample names are composites of the trench number (e.g., TR1-30) and the sample depth in cm (e.g., TR1-30). OSL is a dating method that can be used to provide an estimate of the time elapsed since luminescent minerals, such as quartz, were last exposed to sunlight (Huntley et al., 1985; Rhodes, 2011). If a quartz grain is exposed to sunlight, thereby zeroing its latent OSL signal, and is subsequently buried, charge will accumulate in the crystal lattice of the grain at a rate that is proportional to the flux of cosmic rays and ionising radiation from the surrounding environment (i.e., the dose rate). When the grain is stimulated with light in the laboratory, the stored energy is released and photons (i.e., OSL) are emitted, the amount of which is proportional to the charge that has accumulated in the grain during burial. This signal can be compared to the OSL signals measured following known laboratory-induced irradiations to estimate the equivalent dose (De) absorbed during burial. Optical ages are calculated by dividing the measured De by the corresponding dose rate. Samples for single-grain OSL dating (n ¼ 20) were collected from each trench comprising the transect by hammering 5 cmdiameter opaque PVC tubes into cleaned and logged sections. Quartz grains of 180e212 mm diameter were extracted from the sediment samples in the laboratory under dim red illumination using standard procedures (Wintle, 1997). Hydrochloric acid (32%) and hydrogen peroxide (50%) were used to remove carbonates and organics, respectively. A sodium polytungstate solution with a density of 2.70 g/cm3 was used to isolate quartz and feldspar grains from heavy minerals. A hydrofluoric acid (40% for 40 min) etch was used to remove the external, alpha-dosed layer (Aitken, 1998) and feldspars. Finally, grains were rinsed in Hydrochloric acid and sieved again to retain the 180e212 mm diameter fraction. Grains

were loaded into a Risø DA20 TL/OSL reader and were optically stimulated by a green (532 nm) laser light (Bøtter-Jensen et al., 2003) and the ultraviolet OSL emissions were measured using an Electron Tubes Ltd 9635Q photomultiplier tube fitted with 7.5 mm of Hoya U-340 filter. Optical stimulations were performed for 2 s (at 90% laser power) at 125  C and equivalent doses were estimated by summing first 0.17 s of signal and using the final 0.3 s as background. Laboratory irradiations were given using a calibrated 90 Sr/90Y beta source mounted Risø DA20 TL/OSL reader. The OSL signal from individual quartz grains was measured to eliminate those grains with unsuitable luminescence properties and to calculate single-grain De values, which enables the identification of incomplete bleaching and post-depositional sediment mixing prior to age calculation. For each sample, between 300 and 500 grains were measured using the single-aliquot regenerativedose (SAR) procedure (Murray and Wintle, 2000). Preheats of 200  C for 10 s and 160  C for 5 s were applied prior to the optical measurement of the natural/regenerative dose and test dose, respectively. Satisfactory single-grain dose recovery experiments (Roberts et al., 1999; Murray and Wintle, 2003) indicate the appropriateness of the SAR procedure for dose estimation using these preheats. Standard tests were applied to ensure the appropriateness of the SAR procedure. These include a zero-regenerative dose check for recuperation (5%), a repeat-regenerative dose for calculation of a recycling ratio to ensure that sensitivity changes are being corrected for (recycling ratio within 2s of unity) (Murray and Wintle, 2000), and an OSL-IR depletion ratio test (OSL-IR depletion ratio within 2s of unity) (Duller, 2003) to identify and eliminate contaminant feldspar grains. The statistical models that were used to analyse De data were fitted using the Luminescence R package (Kreutzer et al., 2012) and include the central age model (CAM; Galbraith et al., 1999), the 3-parameter minimum age model (MAM; Galbraith et al., 1999), the unlogged MAM (ul-MAM; Arnold et al., 2009), and the finite mixture model (FMM; Roberts et al., 2000). The MAM and ul-MAM were both applied with a 10%

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

overdispersion. The ul-MAM was used for samples with large numbers of negative De values. For samples with few (<4) negative De values or where the ul-MAM could not be fitted, De datasets were transformed (e.g., “De þ 3”) and the MAM was calculated, then transformed back (e.g., “calculated MAM e 3”). This same transformation approach was used to calculate CAM De estimates for all datasets that contain negative De values. Additional parameters used to describe De distributions include the overdispersion (calculated using the CAM), skewness, and the percentage of zero-dose grains (i.e., grains yielding a De consistent with zero, following terminology of Bateman et al., 2007a). Two criteria were used to identify zero-dose grains. Grains were defined as zero-dose grains if their De values are (i) consistent with 0 at 2s and (ii) less than 5 Gy; the latter criterion removes the influence of high-dose De values with very low precisions. The dose rate for each sample is derived from cosmic rays and the alpha, beta, and gamma dose rates due to 238U, 235U, 232Th (and their decay products) and 4 K. The alpha-affected portion of each grain was removed by HF-etching. The cosmic-ray dose rate was calculated following Prescott and Hutton (1994). The concentrations of U, Th, and K in bulk sediment samples were measured using inductively-coupled plasma mass spectrometry (ICP-MS; e.g., Bailey et al., 2003) at ALS Geochemistray, Reno, USA. Dose rates were calculated using the conversion factors of Guerin et al. (2011) and an allowance was made for beta-dose attenuation (Mejdahl, 1979) and sample water content (Aitken, 1985). The latter value represents the average water content over the duration of burial and is assumed to be 5 ± 2%. The uncertainty on this value is sufficient to accommodate the likely range of water contents experienced by these deposits. An effective internal alpha dose rate of 0.03 ± 0.01 Gy/ka was assumed, based on previous measurements made on quartz from Australia (Bowler et al., 2003). 3. Results 3.1. Sedimentology and stratigraphy Fig. 2 shows the locations of all trenches overlaying the results of Lockwood (2009) showing the projected depth of sand. Fig. 3 shows the stratigraphy for Trenches 1e6. The grain-size characteristics of sediment samples from Trenches 1e7 are presented in Fig. 4. Trenches 1e6 have the same general stratigraphy: an upper unit composed of well-sorted massive, structureless sands (mean grain

57

size of 289 mm) between 20 (Trench 6) and 250 (Trench 2) cm thick, overlying a compact sandy unit with a high proportion of silt and clay (60% sand with a mean grain size of 213 mm). The underlying silty-clayey-sand has been interpreted as the B-horizon of a duplex soil (Storey et al., 1963). However sandstone blocks were observed in the bases of most trenches, suggesting that this unit may be associated with saprolitic decay of the underlying Permian quartzlithic sandstone. Grain-size distributions for samples from Trenches 4e7 (shallow or no sand cover), and the deepest samples from Trenches 1e3 (deep sand cover) show a higher proportion of silt and clay in samples nearby or from this underlying unit (Fig. 4). The overlying sandy unit is very well sorted (87% sand, Fig. 4). The depth of the sands indicates that the dune is oriented downwind (based on the modern yearly prevailing winds) and on the leeward side of a bedrock ridge (Fig. 2). The grain-size data, dune shape and orientation, prevailing wind direction, and position of a suitable sediment source (Wollombi Brook) ~4 km upwind of our study site strongly suggest that the Warkworth sandsheet and dunes are aeolian forms with material sourced from Wollombi Brook. Unlike some of the nearby stratified source-bordering dune deposits identified by Page (1971) no depositional structures were observed in the massive sand deposits in any of the trenches. Extensive mixing of the sandy sediments is indicated by the lack-ofdepositional structures, and the direct observation of roots (<2 cm thick) to a depth of 200 cm below the surface (Trench 2) and tree stump holes (~20 cm wide and ~60 cm deep) in several trenches (Trench 1 and 3, Fig. 5). 3.2. Optically stimulated luminescence Single-grain dose recovery experiments were performed for samples TR1-60 and TR3-170. Grains were bleached for 60 min using a Sol2 solar simulator positioned 100 cm above the samples before being given a surrogate natural dose of 20 Gy in the Risø TL/ OSL reader. The measured/given dose ratios for both TR1-60 (n ¼ 106, 0.99 ± 0.01, 0% overdispersion) and TR3-170 (n ¼ 44, 1.00 ± 0.02, 0% overdispersion) are consistent with unity at 1s. These results indicate that the SAR procedure is appropriate for estimating known radiation doses. 3.2.1. Dose rates and de distributions Dose rate data for all samples are shown in Table 1. The De data

Fig. 3. Cross sectional view of the transect (Fig. 2) with the stratigraphy of each trench and the locations of OSL samples shown as circles in stratigraphy (colours correspond to mixing zones e see text and Fig. 7). The topography (thick black line) is derived from LiDar data, and the dashed line shows the contact between massive sands and sandy-silts. Note that the left “bump” is a bedrock hill and the right “bump” is a cross section across the dune (see Fig. 2 for reference).

58

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

Fig. 4. Grain-size distributions for all OSL samples from Trenches 1e7.

for all samples are presented in Table 2. De data from three typical samples are plotted as radial plots in Fig. 6 (De data for all samples are shown in Supp. Fig. 1). Between 21% and 50% of the measured grains have OSL signals suitable for dating using the SAR procedure. All but one (TR2-190) of the MAM and CAM age estimates are in the correct stratigraphic order, but most samples have positively skewed De data (Table 2, Fig. 7b) and all samples have highly dispersed De data (Fig. 6, Fig. S1). The De datasets from all samples contain wide ranges of De values, as quantified by overdispersion values between 36% and 90% (Table 2). This suggests that these deposits do not represent a single depositional event in which quartz grains were well bleached and subsequently unmixed. Given

the depositional setting and the likely mode of sediment transport (i.e., aeolian), it is unlikely that grains were incompletely bleached prior to deposition. Spatial heterogeneity in the dose rate cannot be ruled out, but is unlikely to be significant given the homogeneous nature of the sediments. Given the vegetated nature of the study site, geomorphic field observations (Fig. 5), and similar observations from similar massive-sandy deposits (Bateman et al., 2003, 2007a,b; Feathers, 2003; Chazan et al., 2013), post-depositional mixing is the most likely cause of the large spread in De data. In this case, the deposits of interest are composed of sand grains that were deposited at different times, and were subsequently moved, vertically upward, vertically downward, and laterally, after burial.

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

59

Fig. 5. Geomorphic observations of mixing. a) Photo of Trench 1. The percentage of modern grains and the MAM ages for each OSL sample are shown at their collection position. Note that the sample in the stump hole (root infill) is from the same depth below the surface as the middle sample in the profile to the left. b) Photo of Trench 3 showing two tree stump holes (root infill). Note 25 cm trowel in bottom-right for scale. c) Photo of top ~60 cm of Trench 2 showing woody and stringy roots in upper 30 cm. Note ~25 cm hammer for scale. d) Photo of woody root 100 cm below the surface in Trench 3. e) Photo showing many roots in upper 30 cm of Trench 2.

Field observations of modern roots and tree stump holes (Fig. 5) at a depth greater than 50 cm indicate that these mixing processes are still occurring. A common approach to resolving depositional ages in mixed sediments is to use the FMM to identify the De population associated with burial (Bateman et al., 2007a; Rhodes et al., 2010; Tribolo et al., 2010; Gliganic et al., 2015). However, in extremely bioturbated deposits, sediments can be mixed to such an extent that the De population related to burial cannot be identified using this approach (Bateman et al., 2007a; Tribolo et al., 2010; Chazan et al., 2013). We applied the FMM to our De data using two approaches to fitting. In the first approach, s (overdispersion) was assumed to be 10% and the optimal k (number of components) was determined using two statistical measures: the maximum log likelihood (llik) and the Bayes Information Criterion (BIC) (Roberts et al., 2000; Jacobs et al., 2008) (Table S1). Most samples' De distributions were best fitted by k ¼ 4 (n ¼ 7) and k ¼ 6 (n ¼ 6), though other samples were best fitted by 3, 5, 7, or 8 components (Fig. S2a). A population that comprised the majority (>50%) of grains in a given De distribution was only identified for eight samples. The large range of De

populations identified suggests that both upward and downward mixing processes are occurring. In the second approach, the optimal combination of k and s were determined for each distribution using the llik and BIC (Table S2). Most samples were best fit by 1 component (n ¼ 9; identical to the CAM) or two components (n ¼ 10), with high s-values (>40%; Fig. S2b). The high s-values suggest that the identified components do not represent discrete populations of grains that were deposited and subsequently unmixed. In the absence of independent age controls, the ambiguity in FMM results, the large spread in De values from all samples, the highly skewed De data and large numbers of zero-dose grains in the upper samples suggest that identification of a De population related to deposition using the FMM is not appropriate. 3.2.2. Elucidating modern mixing processes The De data can be used to explore how modern bioturbation processes affect these deposits, such as the depth to which mineral grains can penetrate into the substrate. Grains that were very recently on the surface would have had their latent OSL signal bleached, and thus yield De values close to zero. These zero-dose

60

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

grains can be used as markers for how deeply modern mixing processes are occurring. The proportions of zero-dose grains from all samples are plotted against depth (Fig. 7a). As expected, it is clear that most zero-dose grains are present near the surface and the number of zero-dose grains decreases with depth. Specifically, three depth zones are apparent in the data: Zone 1 (0e50 cm), Zone 2 (50e60 cm), and Zone 3 (>60 cm). De distributions from samples from Zone 1 are composed of between 16% and 79% zero-dose grains. Samples from Zone 2 have between 2% and 15% zero-dose grains. Samples from Zone 3 have <3% zero-dose grains (with the exception of the outlier sample e TR2-190 e highlighted in Fig. 7). While most De datasets are positively skewed, samples from Zones 1 and 2 have more skewed De distributions than samples from Zone 3 (Fig. 7b). These data suggest that Zone 1 (the upper ~50 cm) is actively receiving bleached grains presently. Zone 2 is the threshold zone into which zero-dose grains can be introduced, though not in all trenches. Zone 3 is deep enough that processes acting presently may only insignificantly contribute grains to these sediments except in extreme cases (e.g., TR2-190). The identified mixing zones can be further identified in our De data using statistical age models to temporally constrain the evolution of the investigated landform. When the MAM age for each sample is plotted against depth it is clear that each zone has a unique MAM age signature (Fig. 7c). All Zone 1 (0e50 cm depth) ages from Trenches 1e3 (dune) are less than 710 years old and two of the samples yield a modern age (i.e., an age that is consistent with 0 e the day of measurement e at 2s). Zone 1 samples from Trenches 4 (base of a sand-mantled bedrock ridge) and 5 (top of a sandmantled bedrock ridge) yield MAM ages between 171 years and 3.66 ka. Zone 2 (50e60 cm depth) samples yielded a spectrum of MAM ages between modern and 7.91 ± 0.65 ka, where the modernaged sample (TR1-60Bio) was collected from a tree stump hole and the oldest sample (Tr1-60) was collected in homogeneous sands next to the stump hole (Fig. 5). For Zone 3 (>60 cm depth), all MAM ages, with the exception of sample TR2-190, are older than 9 ka. 4. Discussion 4.1. Aggradational chronology using the mixing zone conceptual framework In light of our results it is imprudent to identify finite

depositional ages for samples using a single age model. However, these results allow a conceptual framework to be constructed and used to build an aggradational chronology for the site based on the concept that mixing is most likely to occur in the upper ~60 cm of a sediment body (i.e., Zones 1 and 2 e “the mixing zone”; Fig. 7d) and assuming that modern mixing processes are representative of past mixing processes. In a given sedimentary deposit, Zone 1 (0e50 cm depth) is receiving zero-dose grains through currently-acting mixing processes and the MAM is sensitive to this, yielding modern or near-modern ages for samples collected from the top 50 cm of deposit (Fig. 7c). Zone 2 (50e60 cm) represents a transitional depth between Zone 1 sediments that are currently being disturbed and Zone 3 (>60 cm depth) sediments that are deep enough that they are not significantly affected by modern mixing processes. Therefore, if modern mixing processes are representative of past mixing processes and the MAM ages for samples from Zone 1 are close to 0, then the MAM ages for samples from the present-day Zone 3 will indicate when these sediments were last within the mixing zone. To put another way, the MAM age for a sample from 150 cm below the surface in 2012 will reflect when these sediments were within the top 60 cm of the past surface e i.e., when surface was 90e150 cm lower than in 2012. The MAM ages vs. depth plot demonstrates the relationship described above for a given sedimentary body e showing that most Zone 1 and some Zone 2 MAM ages from the dune are effectively consistent with 0 and, therefore, that the MAM age reflects the last time a sample was within the top ~60 cm of the surface (i.e., the mixing zone). Fig. 8 shows a schematic of this model through time. At time T1 a sand body (A) is transported and deposited. At time T2, deposition has ceased and bioturbational processes mix grains of T2 age into the sand body A. At time T3, a new sand body (B) is deposited, thus burying sand body A below the mixing zone and ceasing the introduction of zerodose grains into the deposit. When the OSL signal is measured, samples from sand body A will yield MAM age estimate of T3 (Fig. 8). Two points should be clarified now. Firstly, for the purposes of this study, an aggradational chronology is different from a depositional chronology even though both deal with the timing of sand accumulation in a given place. A depositional chronology (i.e., what is usually constructed using OSL dating) reports the timing of the deposition and accumulation of specific sediments (i.e., the grains that were dated). By contrast, an aggradational chronology deals

Table 1 Dose rate data. Trench

Depth (cm)

Radionuclide concentrations K (%)

1 1 1 1 2 2 2 2 2 2 2 3 3 3 3 3 3 4 4 5

30 60 60-Bio 100 30 60 100 130 160 190 240 25 55 80 110 145 170 20 40 25

0.27 0.28 0.33 0.28 0.25 0.27 0.28 0.28 0.27 0.30 0.35 0.21 0.23 0.30 0.27 0.28 0.42 0.58 0.61 2.09

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

Th (ppm) 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.04 0.02 0.02 0.03 0.03 0.03 0.04 0.06 0.06 0.21

2.60 2.90 2.80 3.00 3.20 4.00 4.60 3.80 4.80 4.00 3.40 2.50 2.80 3.20 3.20 3.00 6.00 3.30 3.70 7.00

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.26 0.29 0.28 0.30 0.32 0.40 0.46 0.38 0.48 0.40 0.34 0.25 0.28 0.32 0.32 0.30 0.60 0.33 0.37 0.70

Dose rates U (ppm) 0.50 0.70 0.60 0.60 0.60 0.70 0.70 0.60 0.70 0.70 0.70 0.50 0.60 0.70 0.60 0.70 1.30 0.80 0.80 1.60

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.05 0.07 0.06 0.06 0.06 0.07 0.07 0.06 0.07 0.07 0.07 0.05 0.06 0.07 0.06 0.07 0.13 0.08 0.08 0.16

Gamma (Gy/ka) 0.23 0.27 0.27 0.27 0.27 0.32 0.35 0.30 0.36 0.33 0.31 0.22 0.24 0.29 0.27 0.28 0.51 0.37 0.40 0.98

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.02 0.03 0.02 0.02 0.06

Beta (Gy/ka) 0.30 0.34 0.36 0.33 0.31 0.35 0.37 0.34 0.37 0.37 0.40 0.26 0.29 0.36 0.33 0.34 0.57 0.57 0.60 1.79

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.02 0.02 0.03 0.02 0.02 0.02 0.03 0.02 0.03 0.03 0.03 0.02 0.02 0.03 0.02 0.02 0.04 0.05 0.05 0.16

Cosmic (Gy/ka) 0.19 0.19 0.19 0.18 0.20 0.19 0.18 0.17 0.17 0.16 0.15 0.20 0.19 0.18 0.18 0.17 0.16 0.20 0.19 0.20

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02

Total (Gy/ka) 0.76 0.83 0.85 0.80 0.81 0.89 0.93 0.85 0.92 0.89 0.89 0.70 0.75 0.86 0.81 0.82 1.27 1.17 1.22 3.00

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.05 0.05 0.05 0.05 0.05 0.06 0.06 0.05 0.06 0.06 0.06 0.04 0.05 0.05 0.05 0.05 0.08 0.08 0.08 0.24

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

61

Table 2 Equivalent dose and OSL age data. Trench

Depth (cm)

N¼

Zero-dose grains (n ¼ /%)

Skewness

1 1 1 1 2 2 2 2 2 2 2 3 3 3 3 3 3 4 4 5

30 60 60-Bio 100 30 60 100 130 160 190 240 25 55 80 110 145 170 20 40 25

215 252 225 160 110 82 78 116 133 117 50 62 178 110 151 171 62 131 156 62

37/17 4/2 25/11 0/0 64/58 12/15 2/3 4/3 3/2 24/21 1/2 44/71 18/10 0/0 1/1 0/0 0/0 104/79 25/16 16/26

4.9 3.8 4.1 1.5 3.5 3.1 2.1 1.6 1.6 0.5 1.3 4.3 6.0 2.0 0.9 1.4 0.0 5.5 3.3 1.0

Over-dispersion 63 59 90 44 49 75 44 44 44 82 36 71 74 62 69 56 49 63 61 37

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

4 3 5 3 6 7 5 4 3 7 5 11 4 5 4 4 5 9 4 5

CAM de (Gy) 4.2 14.6 8.9 33.5 3.7 14.5 23.0 39.5 49.9 42.6 104.6 1.6 7.8 20.1 41.4 93.7 126.9 1.3 8.3 7.8

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.2 0.6 0.6 1.2 0.3 1.3 1.5 1.9 2.1 3.7 6.5 0.2 0.5 1.2 2.5 4.5 8.9 0.1 0.5 0.5

CAM age (ka) 5.5 17.6 10.5 41.7 4.5 16.3 24.7 46.4 54.0 47.7 117.5 2.3 10.3 23.3 51.3 114.7 99.6 1.1 6.8 2.6

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.4 1.3 1.0 3.2 0.4 1.8 2.3 3.8 4.3 5.2 10.8 0.3 0.9 2.1 4.6 9.3 9.7 0.1 0.6 0.3

MAM

MAM de (Gy)

ul-MAM MAM ul-MAM MAM ul-MAM MAM MAM MAM MAM MAM MAM ul-MAM MAM MAM MAM MAM MAM ul-MAM MAM ul-MAM

0.05 6.56 0.12 14.80 0.57 5.10 10.73 15.46 28.24 5.95 58.71 0.05 1.92 8.05 8.64 23.80 50.97 0.20 4.45 3.09

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.13 0.32 0.24 0.80 0.08 0.52 0.97 1.14 2.07 0.74 5.95 0.04 0.16 0.61 0.60 2.33 6.24 0.04 0.28 0.61

MAM age (ka) 0.07 7.91 0.14 18.40 0.71 5.72 11.51 18.18 30.59 6.67 65.94 0.07 2.54 9.33 10.72 29.13 40.00 0.17 3.66 1.03

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.17 0.65 0.28 1.57 0.11 0.69 1.29 1.81 3.04 0.95 8.03 0.06 0.27 0.94 1.03 3.44 5.59 0.04 0.35 0.22

Fig. 6. Single-grain De distributions, shown as radial plots, for typical samples from each mixing zone (all De distributions are shown as radial plots in Fig. S1). Grey band shows grains consistent with 0 at 2s (zero-dose grains) and the solid line shows the MAM estimate.

with accumulation of sediment on a landform scale. In Fig. 8, the depositional age for sand body A is T1, while the aggradational age derived from sand body A is T3. In this study, sediments are too mixed to allow depositional ages for samples to be obtained. However an aggradational chronology for the dune can be

developed by using grains from lower deposits to infer the accumulation of overlying sediments, as described below and shown in Fig. 8. Secondly, it should be noted that in addition to mixing vertically downward, upward vertical mixing and laterally mixing are almost certainly contributing to the spread in De data for each

Fig. 7. Proportion of accepted grains that are zero-dose grains (a.), skewness (b.), and MAM ages (c.) plotted against depth below surface (cm) of sample. Data is plotted as three groups based on depth downprofile: Zone 1 (0e50 cm: red), Zone 2 (50e60 cm: green), and Zone 3 (>60 cm: blue). Note that the outlier sample (TR2-190) has been circled. d. Mixing zone schematic showing Zone 1 is currently receiving grains from the surface while Zone 3 is not, and Zone 2 is transitional between the two. Zones 1 and 2 are the “mixing zone” (see text for more details). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

62

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

Fig. 8. Schematic of the mixing zone conceptual framework through time.

sample and the overall mixing at the site. However, because the mixing zone conceptual framework (Fig. 8) relies on the MAM ages for each sample, vertical upward mixing and lateral mixing will have no effect on the resulting aggradational chronology. For a given sample, underlying older grains that are vertically mixed upwards and coincident grains mixed laterally into the sample will not affect the MAM age estimate for that sample. Consequently, since only vertical downward mixing is being investigated in this study, the term “downward mixing” is used. Using the mixing-zone framework and assuming that mixing processes have remained constant through time, a chronology for dune aggradation can be developed. The MAM age for a sample represents the last time that sample was within ~60 cm of the surface and, therefore, when sediment accumulated. In this way, a chronostratigraphy can be constructed that represents dune aggradation. MAM ages for each sample are plotted on their

Fig. 9. Cross sectional view of dune portion of transect with MAM OSL ages plotted in their stratigraphic position (a.). Schematic chronostratigraphic boundaries are interpreted based on OSL chronology and shown in (b.).

stratigraphic position (Fig. 9a). Where multiple samples yield consistent ages, the weighted mean has been calculated and plotted to show the chronostratigraphy for the Warkworth dune (Fig. 9b). Each MAM age represents dune aggradation at the sampling location. Our chronostratigraphy indicates phases of aggradation at 65.9 ± 8.0 ka (n ¼ 1), 29.9 ± 2.3 ka (n ¼ 2), 18.3 ± 1.2 ka (n ¼ 2), and 10.3 ± 0.6 ka (n ¼ 3). Within the modern mixing zone, samples of middle and late Holocene age indicate the most recent phase of dune aggradation. Most of the aggradational phases identified are represented by two samples from across the dune surface. The coincidence of ages from different trenches supports the presence of significant aggradational phases at ~29.9, ~18.3, ~10.3 ka, and continued aggradation though middle and late Holocene times. In particular, the presence of overlying samples in Trench 3 and a sample from the adjacent Trench 2 with coincident ages suggests that aggradation at 10.3 ka was significant. Interpreting the aggradational chronology for temperate source bordering sand bodies such as the Warkworth sand sheet without additional data is difficult. A host of competing factors influence dune growth including sediment supply, windiness, precipitation, and vegetation type and cover (Telfer and Hesse, 2013; Hesse, 2014), complicating the interpretation of environmental information from dune records. For example, while phases of increased precipitation/stream runoff may increase the sediment supply along a river channel, more riparian and in-channel vegetation cover may inhibit the active sediment supply. Likewise, dry and windy conditions may increase aeolian activity, but result in decreased river activity reducing the available sediment supply for dune building. Reduced vegetation cover in southeastern Australia during the LGM (Singh and Geissler, 1985; Sweller and Martin, 2001) likely allowed temperate dunes to form in the Blue Mountains (Hesse et al., 2003). Similarly, our record indicates dune aggradation at 18.3 ± 1.2 ka, however this phase is not more pronounced than other late Pleistocene and Holocene aggradational phases. While available records for Wollombi Brook do not extend beyond the middle Holocene (Erskine and Melville, 2008), fluvial activity on the Nepean, Bellinger, Nambucca, and Shoalhaven rivers (Nott et al., 2002; Nanson et al., 2003) show marked activity at ~80, ~60e40, ~30e25, and then ~20e15 ka, largely mirroring three phases of dune aggradation in our record (~65.9, ~29.9, and ~18.3 ka). Ages between 20 and 13 ka for source bordering dunes along the Goulburn river system (Bowler, 1967) are consistent with our identified dune aggradation at 18.3 ± 1.2 ka, but the largest phase of aggradation in the Warkworth sand sheet was at 10.3 ± 0.6 ka. These differences reflect the difficulty in making inferences about Quaternary environments from aeolian archives.

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

4.2. Mechanisms and rates of downward mixing Our observation that mixing processes are strongest at the surface and become more diffuse with depth conforms to other published data (Roering et al., 2002; Bateman et al., 2003, 2007a; Wilkinson et al., 2009; Stockmann et al., 2013). Stockman et al. (2013) observed many modern grains in the top 35 cm of clayand silt-rich soils in southeastern Australia and conclude that grains in this zone are likely to move to and from the surface. Roering et al. (2002) showed that as the depth of a tephra marker bed on a hillslope in New Zealand declined from 80 cm to 40e50 cm, the layer transitioned from a thin concentrated tephra layer (>50 cm depth) to a less concentrated, uniformly distributed tephra in the upper 40 cm of soil. They linked the even distribution of tephra in the upper 40e50 cm of soil and the coincident rooting depth of two local tree species and conclude that biological disturbance (notably floral characteristics) can generate a wellmixed mobile soil layer. Phillips and Marion (2006) investigated rocky soils in the Ouachita Mountains in Arkansas, USA to assess the biomechanical effects of trees on soil mixing. They used rock fragments as tracers and show that tree uprooting, physical displacement of soil by root growth, and infilling of tree stump rot depressions (tree stump holes) are significant processes that maintain a continuously mixed surface biomantle, with complete forest soil turnover potential over Holocene time scales. Brimhall et al. (1991) used a pseudoroot (surgical tubing) buried vertically in a sandy matrix to experimentally demonstrate vertical mixing of soil. They repeatedly inflated and deflated the pseudoroot, which caused material on the surface of the matrix to become mixed through the soil column, with the depth of mixing increasing with the number of inflation and deflation cycles. Our field observations of multiple tree stump holes to a depth of 60 cm and roots in our deposits conform to these published observations of floralturbation. Consequently, we infer that floralturbation by tree roots (displacement by root growth, uprooting/treethrow, and infilling of tree stump holes) is a significant factor that transports grains and rocks from the surface to the subsurface. In addition to developing an aggradational chronology, our De data and the mixing zone conceptual model can be used to quantify bioturbational processes in this environment. The proportion of zero-dose grains and MAM ages plotted against depth (Fig. 7) shows that the magnitude of downward mixing is most strongly observed in the upper 50 cm, and is insignificant below 60 cm in our study area, demonstrating the depth to which the bulk of mixing agents are acting. In addition to knowing the depth to which these processes act, it is useful to know how long it takes for mixing processes to affect these deposits. Since the MAM age of a sample represents the timing of accumulation of the overlying ~60 cm of sand, the MAM age estimates for samples from the upper 60 cm can be used to estimate i, how much time has passed since the overlying sands were deposited and have since been affected by mixing processes and ii, a downward mixing rate for these deposits. The youngest packet of non-modern sand in the dune is represented by a sample from a depth of 55 cm (TR3-55) that yields a MAM age of 2.5 ± 0.3 ka. Two other samples from a depth of ~60 cm (TR1-60 and TR2-60) yield MAM ages of 5.7 ± 0.7 ka and 7.9 ± 0.7 ka. While these estimates are variable, they indicate that mixing processes are capable of significantly influencing the composition of the upper 60 cm of soil and the OSL age estimates derived from these sediments over millennial time scales (e.g., Phillips and Marion, 2006). Others have used OSL data to quantify soil mixing rates for in situ weathered soil profiles (Stockman et al., 2013; Wilkinson and Humphreys, 2005), but these methods cannot work in our deposits given the aggradational and geomorphic contexts. Conceptually, all grains in an in situ weathered soil profile have an infinite

63

age to start with and any grain with a finite age must have been bleached on the surface and mixed into the soil from above, allowing the calculation of mixing rates (grain depth/age). In our study area, the landform is composed of grains with various transport and burial histories, complicating the use of previously described mixing rate calculation methods based on OSL data. Given the geomorphic context of our study site, the term “downward mixing rate” can be used to denote the rate at which grains from the surface penetrate the subsurface (i.e. the time it takes to replace some proportion of grains at a given sediment depth with grains from the surface). Here, we suggest an approach to quantify downward mixing rates in the aeolian Warkworth sands using single-grain OSL data. For each sample from the mixing zone, the zero-dose grains were investigated. Firstly, we estimated how long it has taken for the proportion of zero-dose grains (P) to reach the sampling depth (D) for each sample. To do this, the weighted mean De value of the zero-dose grains in each sample was calculated using the CAM as described for modern samples in Gliganic et al. (2015) and converted to an age in years (T) (Table 3). This value provides an approximation of the time it has taken for a given proportion of grains in a sample (i.e., P) to be replaced with grains from the surface and represents the intensity of the currently ongoing mixing processes. Thus, for each sample, T/P provides an estimate of the time it takes for 1% of the grains at depth D to be replaced with grains introduced from the surface, and T/P*100 (i.e., the “replacement time”) quantifies the time it takes to replace all grains with surficial grains (Table 3, Fig. 10a). These data indicate that, as expected, the replacement time is depth dependent (Fig. 10a). Zone 1 samples have replacement times ranging from 0.015 ± 0.002 ka (i.e., 15 ± 2 years) to 8.3 ± 1.9 ka, while Zone 2 samples have replacement time ranging from 16.4 ± 4.8ka to 266.5 ± 90.1 ka. This indicates that complete overturning of the upper 50 cm of these deposits will occur on Holocene timescale if mixing rates remain constant. Finally, the replacement times for each sample can be used to calculate downward mixing rates at a given depth, as commonly used in soil science (Schaetzl and Anderson, 2005; Wilkinson et al., 2009). Each sample was treated as the centre of a 5 cm thick sediment package corresponding to the diameter of the OSL sampling tubes. Based on a density estimate for these loose sand deposits of 1.50 ± 0.15 g/cm3, the mass estimate per sample unit/volume was then divided by the replacement time to calculate a downward mixing rate (g/m2/year) at each depth (Table 3, Fig. 10b). Downward mixing rates are depth dependent; the downward mixing rates for Zone 1 samples (9.1e4856 g/m2/year) are orders of magnitude larger than for Zone 2 samples (0.3e4.5 g/ m2/year). The observed depth dependency and apparent non-linear decrease with depth of replacement time and downward mixing rates supports results presented in Fig. 7 and corresponds to previously reported data (Wilkinson et al., 2009). The estimated downward mixing rates are realistic and within the range of rates presented by Schaetzl and Anderson (2005, and references therein). The congruence of our single-grain OSL-based downward mixing rate results with field observations, OSL age data, and previous studies suggests that this approach to downward mixing rate calculation is robust. 5. Conclusions The development of OSL dating over the past three decades has allowed unprecedented insights into the evolution of aeolian dunes and sandsheets. However developing chronologies for these and many other landforms can be difficult due to their susceptibility to post-depositional mixing processes. We investigated significantly mixed dune deposits in an archaeologically rich sandsheet complex

64

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

Table 3 Replacement time and downward mixing rate data. Trench

1 1 1 2 2 3 3 4 4 5 a b c d

Depth (cm)

Percent zero-dose grainsa

CAM age (years)

D

P

T

30 60 60 30 60 25 55 20 40 25

17 2 11 58 15 71 10 79 16 26

790 4230 2072 621 2404 16 2314 12 1322 266

bio

T/Pb

Replacement time (years)c

Downward mixing ratesd g/m2/yr

± ± ± ± ± ± ± ± ± ±

109 1430 364 70 702 2 443 1 299 52

45.9 2664.8 186.5 10.7 164.3 0.2 228.9 0.2 82.5 10.3

4593 266481 18649 1067 16425 23 22885 15 8251 1031

± ± ± ± ± ± ± ± ± ±

636 90081 3273 121 4800 3 4384 2 1867 200

16.3 0.3 4.0 70.3 4.6 3262.2 3.3 4856.7 9.1 72.8

± ± ± ± ± ± ± ± ± ±

t/Ha/yr 2.8 0.1 0.8 10.6 1.4 578.5 0.7 678.9 2.2 15.9

0.180 0.003 0.044 0.775 0.050 35.960 0.036 53.536 0.100 0.802

± ± ± ± ± ± ± ± ± ±

0.031 0.001 0.009 0.117 0.016 6.377 0.008 7.483 0.025 0.175

Percent of grains measured that are zero-dose grains. Time it takes for 1% of sampled sediment body to be replaced by grains from the surface. Time it takes for 100% of sampled sediment body to be replaced by grains from the surface. Calculated assuming average density of loose sand deposits at 1.50 ± 0.15 g/cm3.

Fig. 10. Replacement time (a.) and downward mixing rate (b.) data plotted against depth below surface (cm) of sample. Values were calculated as described in text and are shown in Table 3. Note that replacement time and downward mixing rates are shown on a log x-axis.

in the lower Hunter River Valley, southeastern Australia. We collected 20 OSL samples from trenches excavated as a transect across the dune and adjacent sand-mantled bedrock ridge. The stratigraphy of the trenches showed 20e200 cm of deep loose, massive, structureless, well-sorted sands with prolific tree stump holes, root infills and modern roots overlying compact, clay-rich sands. OSL samples from these mixed deposits yield single-grain De datasets from which depositional ages could not be determined. Plots of the proportion of modern grains per sample and MAM ages against depth are used to identify the present day “mixing zone” (the upper 50e60 cm) in which modern mixing processes are acting. This observation was used to develop a conceptual framework that is applied to buried deposits and used to estimate phases of aggradation of the aeolian deposits and quantify rates of downward mixing. The mixing zone conceptual framework was combined with the MAM to show that phases of significant sand aggradation likely occurred at ~29.9, ~18.3, ~10.3 ka, and continued through the Holocene. We also present a new approach utilising single-grain De data to estimate downward mixing rates. The calculated downward mixing rates show a strong depth dependency and are in line with previously reported mixing rates, indicating that the upper ~50 cm of these deposits will be completely turned over on millennial time scales. While caution is still required when interpreting archaeological and OSL data from bioturbated deposits, our results demonstrate that single-grain OSL data and contextual knowledge can be used to better understand mixing processes and the evolution of sandy landscapes.

Acknowledgements The project was funded by Scarp Archaeology and Rio Tinto Coal and Allied under a NSW Planning and Assessment Commission permit. We thank Dr Dave Cameron from RTCA for facilitating this project. The authors acknowledge the participation of the Aboriginal stakeholders of the upper Hunter Valley in the fieldwork component of this project. We also thank several anonymous reviewers for their helpful comments, which have improved this manuscript. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.quageo.2015.12.006. References Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press, London. Aitken, M.J., 1998. An Introduction to Optical Dating: the Dating of Quaternary Sediments by the Use of Photon-stimulated Luminescence. Oxford University Press, Oxford. Araujo, A.G.M., Feathers, J.K., Arroyo-Kalin, M., Tizuka, M.M., 2008. Lapa das boleiras rockshelter: stratigraphy and formation processes at a paleoamerican site in Central Brazil. J. Archaeol. Sci. 35, 3186e3202. Arnold, L.J., Roberts, R.G., Galbraith, R.F., DeLong, S.B., 2009. A revised burial dose estimation procedure for optical dating of young and modern-age sediments. Quat. Geochronol. 4 (4), 306e325. http://dx.doi.org/10.1016/j.quageo.2009.02.017. Attenbrow, V., 2002. Sydney's Aboriginal Past. UNSW Press, University of New

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66 South Wales. Bailey, R.M., Stokes, S., Bray, H., 2003. Inductively-Coupled Plasma Mass Spectrometry (ICP-MS) for dose rate determinations: some guidelines for sample preparation and analysis. Anc. TL 21, 11e15. Balek, C.L., 2002. Buried artifacts in stable upland sites and the role of bioturbation: a review. Geoarchaeology 17 (1), 41e51. Bateman, M.D., Frederick, C.D., Jaiswal, M.K., Singhvi, A.K., 2003. Investigations into the potential effects of pedoturbation on luminescence dating. Quat. Sci. Rev. 22 (10e13), 1169e1176. http://dx.doi.org/10.1016/S0277-3791(03)00019-2. Bateman, M.D., Boulter, C.H., Carr, A.S., Frederick, C.D., Peter, D., Wilder, M., 2007a. Detecting post-depositional sediment disturbance in sandy deposits using optical luminescence. Quat. Geochronol. 2 (1e4), 57e64. http://dx.doi.org/ 10.1016/j.quageo.2006.05.004. Bateman, M.D., Boulter, C.H., Carr, A.S., Frederick, C.D., Peter, D., Wilder, M., 2007b. Preserving the palaeoenvironmental record in Drylands: bioturbation and its significance for luminescence-derived chronologies. Sediment. Geol. 195, 5e19. Bell, S.A.J., 2010. Review of Flora Investigations: Proposed Expansion to Warkworth Mine Singleton LGA. Report to Bulga-Milbrodale Progress Association Inc. C/- H. Davis 245 Inlet Road Bulga NSW 2330 June 2010. Bøtter-Jensen, L., Andersen, C.E., Duller, G.A.T., Murray, A.S., 2003. Developments in radiation, stimulation and observation facilities in luminescence measurements. Radiat. Meas. 37 (4e5), 535e541. http://dx.doi.org/10.1016/S13504487(03)00020-9. Bowler, J.M., Johnston, H., Olley, J.M., Prescottk, J.R., Roberts, R.G., Shawcross, W., Spooner, N.A., 2003. New ages for human occupation and climatic change at Lake Mungo, Australia. Nature 421, 20e23. http://dx.doi.org/10.1038/ nature01391.1. Bowler, J.M., 1967. Quaternary chronology of Goulburn Valley sediments and their correlation in Southeastern Australia. J. Geol. Soc. Aust. 14, 287e292. Brimhall, G.H., Chadwick, O.A., Lewis, C.J., Compston, W., Williams, I.S., Danti, K.J., Dietrich, W.E., et al., 1991. Deformational mass transport and invasive processes in soil evolution. Science 255, 695e702. Bueno, L., Feathers, J.K., De Blasis, P., 2013. The formation process of a paleoindian open-air site in Central Brazil: integrating lithic analysis, radiocarbon and luminescence dating. J. Archaeol. Sci. 40, 190e203. Chazan, M., Porat, N., Sumner, T.A., Horwitz, L.K., 2013. The use of OSL dating in unstructured sands: the archaeology and chronology of the Hutton Sands at Canteen Kopje (Northern Cape Province, South Africa). Archaeol. Anthropol. Sci. 5, 351e363. http://dx.doi.org/10.1007/s12520-013-0118-7. Cohen, T.J., Nanson, G.C., Jansen, J.D., Gliganic, L.A., May, J.-H., Larsen, J.R., Goodwin, I.D., et al., 2012. A pluvial episode identified in arid Australia during the medieval climatic Anomaly. Quat. Sci. Rev. 56, 167e171. http://dx.doi.org/ 10.1016/j.quascirev.2012.09.021. David, B., Roberts, R.G., Magee, J., Mialanes, J., Turney, C., Bird, M., White, C., et al., 2007. Sediment mixing at Nonda Rock: investigations of stratigraphic integrity at an early archaeological site in northern Australia and implications for the human colonisation of the continent. J. Quat. Sci. 22, 449e479. Duller, G.A.T., 2003. Distinguishing quartz and feldspar in single grain luminescence measurements. Radiat. Meas. 37 (2), 161e165. http://dx.doi.org/10.1016/S13504487(02)00170-1. Duller, G.A.T., 2006. Single grain optical dating of glacigenic deposits. Quat. Geochronol. 1 (4), 296e304. http://dx.doi.org/10.1016/j.quageo.2006.05.018. Erskine, W.D., Melville, M.D., 2008. Geomorphic and stratigraphic complexity: holocene alluvial history of upper Wollombi Brook, Australia. Geogr. Ann. Ser. A, Phys. Geogr. 90, 19e35. Feathers, J.K., 2003. Single-grain OSL dating of sediments from the Southern high Plains, USA. Quat. Sci. Rev. 22 (10e13), 1035e1042. http://dx.doi.org/10.1016/ S0277-3791(03)00049-0. Feathers, J., Rhodes, E., Huot, S., Mcavoy, J., 2006. Luminescence dating of sand deposits related to late Pleistocene human occupation at the Cactus Hill Site, Virginia, USA. Quat. Geochronol. 1 (3), 167e187. http://dx.doi.org/10.1016/ j.quageo.2006.05.011. Fitzsimmons, K.E., Rhodes, E.J., Magee, J.W., Barrows, T.T., 2007. The timing of linear dune activity in the Strzelecki and Tirari Deserts, Australia. Quat. Sci. Rev. 26, 2598e2616. Fitzsimmons, K.E., Cohen, T.J., Hesse, P.P., Jansen, J., Nanson, G.C., May, J.-H., Barrows, T.T., et al., 2013. Late quaternary palaeoenvironmental change in the Australian drylands. Quat. Sci. Rev. 74, 78e96. http://dx.doi.org/10.1016/ j.quascirev.2012.09.007. Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., Olley, J.M., 1999. Optical dating of single and multiple grains of quartz from Jinmium rock shelter, northern Australia: part I, experimental design and statistical models. Archaeometry 41, 339e364. Galloway, R.W., 1963. Geomorphology of the Hunter Valley. In: Story, R., Galloway, R.W., van de Graaff, R.H.M., Tweedie, A.D. (Eds.), General Report on the Land of the Hunter Valley. CSIRO, Melbourne, pp. 90e102. Land Research Series 8. Gliganic, L.A., Cohen, T.J., May, J.-H., Jansen, J.D., Nanson, G.C., Dosseto, A., Larsen, J.R., et al., 2014. Late-Holocene climatic variability indicated by three natural archives in arid southern Australia. Holocene 24 (1), 104e117. http:// dx.doi.org/10.1177/0959683613515732. Gliganic, L.A., Jacobs, Z., Roberts, R.G., 2012. Luminescence characteristics and dose distributions for quartz and feldspar grains from Mumba rockshelter, Tanzania. Archaeol. Anthropol. Sci. 4, 115e135. http://dx.doi.org/10.1007/s12520-0110085-9.

65

Gliganic, L.A., May, J.-H., Cohen, T.J., 2015. All mixed up: using single-grain equivalent dose distributions to identify phases of pedogenic mixing on a dryland alluvial fan. Quat. Int. 362, 23e33. http://dx.doi.org/10.1016/ j.quaint.2014.07.040. Guerin, G., Mercier, N., Adamiec, G., 2011. Dose-rate conversion factors: update. Ancient TL. 29, 5e8. Hanson, P.R., Mason, J.A., Jacobs, P.M., Young, A.R., 2015. Evidence for bioturbation of luminescence signals in eolian sand on upland ridgetops, southeastern Minnesota, USA. Quat. Int. 362, 108e115. http://dx.doi.org/10.1016/ j.quaint.2014.06.039. Heimsath, A.M., Chappell, J., Spooner, N.A., Questiaux, D.G., 2002. Creeping soil. Geology 30, 111e114. http://dx.doi.org/10.1130/0091-7613(2002)030<0111. Hesse, P.P., 2014. How do longitudinal dunes respond to climate forcing? Insights from 25 years of luminescence dating of the Australian desert dunefields. Quat. Int. http://dx.doi.org/10.1016/j.quaint.2014.02.020 (in press). Hesse, P.P., Humphreys, G.S., Selkirk, P.M., Adamson, D.A., Gore, D.B., Nobes, D.C., 2003. Late quaternary aeolian dunes on the presently humid Blue Mountains, Eastern Australia. Quat. Int. 108, 13e32. Hughes, P., Spooner, N., Questiaux, D., 2014. The central lowlands of the Hunter Valley, NSW: why so few early sites have been found in this archaeologicallyrich landscape. Aust. Archaeol. 79, 34e44. Humphreys, G.S., Field, R., 1998. Mixing, mounding and other aspects of bioturbation: implications for pedogenesis. In: 16th World Congress of Soil Science. International Society of Soil Science, Montpellier. Registered paper no.18. Humphreys, G.S., Mitchell, P.B., 1983. A preliminary assessment of the role of bioturbation and rainwash on sandstone hillslopes in the Sydney Basin. In: Young, R.W., Nanson, G.C. (Eds.), Aspects of Australian Sandstone, Landscapes Special Publication, vol. 1. Australian and New Zealand Geomorphology Group, Wollongong, pp. 66e80. Huntley, D.J., Godfrey-Smith, D.I., Thewalt, M.L.W., 1985. Optical dating of sediments. Nature 313, 105e107. http://dx.doi.org/10.1038/313105a0. Jacobs, Z., Wintle, A.G., Duller, G.A.T., Roberts, R.G., Wadley, L., 2008. New ages for the post-Howiesons Poort, late and final middle stone age at Sibudu, South Africa. J. Archaeol. Sci. 35, 1790e1807. http://dx.doi.org/10.1016/ j.jas.2007.11.028. Johnson, M.O., Mudd, S.M., Pillans, B., Spooner, N.A., Fifield, K.L., Kirkby, M.J., Gloor, M., 2014. Quantifying the rate and depth dependence of bioturbation based on optically-stimulated luminescence (OSL) dates and meteoric 10 Be. Earth Surf. Process. Landforms 39 (9), 1188e1196. http://dx.doi.org/10.1002/ esp.3520. Kreutzer, S., Schmidt, C., Fuchs, M.C., Dietze, M., Fischer, M., Fuchs, M., 2012. Introducing an R package for luminescence dating analysis. Anc. TL 30 (1), 1e8. Kristensen, J.A., Thomsen, K.J., Murray, A.S., Buylaert, J.-P., Jain, M., BreuningMadsen, H., 2015. Quantification of termite bioturbation in a savannah ecosystem: application of OSL dating. Quat. Geochronol. 30, 334e341. Kovac, M., Lawrie, J.M., 1990. Soil Landscapes of the Singleton 1:250,000 Sheet. Soil Conservation Service of NSW, Sydney. Leigh, D.S., 1998. Evaluating Artifact burial by eolian versus bioturbation processes, South Carolina Sandhills, USA. Geoarchaeology 13 (3), 309e330. Lockwood, P., 2009. Mount Thorley-warkworth Extension Area Soil Survey (Unpublished report to Coal and Allied). Lomax, J., Hilgers, A., Radtke, U., 2011. Palaeoenvironmental change recorded in the palaeodune fields of the western Murray Basin, South Australia e new data from single grain OSL-dating. Quat. Sci. Rev. 30 (5e6), 723e736. http:// dx.doi.org/10.1016/j.quascirev.2010.12.015. Lombard, M., Wadley, L., Jacobs, Z., Mohapi, M., Roberts, R.G., 2010. Still Bay and serrated points from Umhlatuzana rock shelter, Kwazulu-Natal, South Africa. J. Archaeol. Sci. 37 (7), 1773e1784. http://dx.doi.org/10.1016/j.jas.2010.02.015. Mejdahl, V., 1979. Thermoluminescence dating: beta-dose attenuation in quartz grains. Archaeometry 21, 61e72. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single- aliquot regenerative-dose protocol. Radiat. Meas. 32, 57e73. Murray, A.S., Wintle, A.G., 2003. The single aliquot regenerative dose protocol : potential for improvements in reliability. Radiat. Meas. 37, 377e381. http:// dx.doi.org/10.1016/S1350-4487(03)00053-2. Nanson, G.C., Cohen, T.J., Doyle, C.J., Price, D.M., 2003. Alluvial evidence of major Late-Quaternary climate and flow-regime changes on the coastal rivers of New South Wales, Australia. In: Gregory, K.J., Benito, G. (Eds.), Palaeohydrology: Understanding Global Change. John Wiley and Sons, Chichester, pp. 233e258. Nott, J.F., Price, D.M., 1991. Late Pleistocene to early holocene aeolian activity in the upper and middle Shoalhaven catchment, New South Wales. Aust. Geogr. 22, 168e177. Nott, J., Price, D., Nanson, G., 2002. Stream response to quaternary climate change: evidence from the Shoalhaven River catchment, southeastern highlands, temperate Australia. Quat. Sci. Rev. 21, 965e974. Page, K.J., Nanson, G.C., Price, D.M., 1991. Thermoluminescence chronology of late quaternary deposition on the Riverine Plain of southeastern Australia. Aust. Geogr. 22, 14e23. Page, K.J., 1971. Riverine source bordering sand dune. Aust. Geogr. 11 (6), 603e605. Page, K.J., Kemp, J., Nanson, G.C., 2009. Late quaternary evolution of Riverine Plain paleochannels, southeastern Australia. Aust. J. Earth Sci. 56, S1. S19-S33. Peacock, E., Fant, D.W., 2002. Biomantle formation and Artifact Translocation in Upland Sandy Soils : an example from the Holly springs national Forest, NorthCentral Mississippi, U. S. a. Geoarchaeology 17, 91e114. Peake, T.C., 2006. The Vegetation of the Central Hunter Valley, New South Wales. A

66

L.A. Gliganic et al. / Quaternary Geochronology 32 (2016) 53e66

Report on the Findings of the Hunter Remnant Vegetation Project. Hunter Central Rivers Catchment Authority, Paterson. Phillips, J.D., Marion, D.A., 2006. Biomechanical effects of trees on soil and regolith: beyond treethrow. Ann. Assoc. Am. Geogr. 96, 233e247. Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiat. Meas. 23 (1), 497e500. Rasmus, P.L., Rose, D.M., Rose, G., 1969. Singleton 1:250,000 Geological Sheet SI/56-01, first ed. Geological Survey of New South Wales, Sydney. Rhodes, E.J., 2011. Optically stimulated luminescence dating of sediments over the past 200,000 years. Annu. Rev. Earth Planet. Sci. 39, 461e488. http://dx.doi.org/ 10.1146/annurev-earth-040610-133425. Rhodes, E.J., Fanning, P.C., Holdaway, S.J., 2010. Developments in optically stimulated luminescence age control for geoarchaeological sediments and hearths in western New South Wales, Australia. Quat. Geochronol. 5 (2e3), 348e352. http://dx.doi.org/10.1016/j.quageo.2009.04.001. Rink, W.J., Dunbar, J.S., Tschinkel, W.R., Kwapich, C., Repp, A., Stanton, W., Thulman, D.K., 2013. Subterranean transport and deposition of quartz by ants in sandy sites relevant to age overestimation in optical luminescence dating. J. Archaeol. Sci. 40 (4), 2217e2226. http://dx.doi.org/10.1016/j.jas.2012.11.006. Roberts, R.G., Galbraith, R.F., Olley, J.M., Yoshida, H., Laslett, G.M., 1999. Optical dating of single and multiple grains of quartz from Jinmium rock shelter, northern Australia: part II, results and implications. Archaeometry 41 (2), 365e395. Roberts, R.G., Galbraith, R.F., Yoshida, H., Laslett, G.M., Olley, J.M., 2000. Distinguishing dose populations in sediment mixtures: a test of single-grain optical dating procedures using mixtures of laboratory-dosed quartz. Radiat. Measur. 32, 459e465. Rodnight, H., Duller, G.A.T., Tooth, S., Wintle, A.G., 2005. Optical dating of a scrollbar sequence on the Klip River, South Africa, to derive the lateral migration rate of a meander bend. Holocene 15, 802e811. Roering, J.J., Sciences, G., Almond, P., Tonkin, P., 2002. Soil transport driven by biological processes over millennial time scales. Geology 30, 1115e1118. Schaetzl, R., Anderson, S., 2005. Soils: Genesis and Geomorphology. Cambridge University Press. Singh, G., Geissler, E.A., 1985. Late cainozoic history of vegetation, fire, lake levels and climate, at Lake George, New South Wales, Australia. Philos. Trans. R. Soc. Lond. B 311, 379e447. Stockmann, U., Minasny, B., Pietsch, T.J., McBratney, A.B., 2013. Quantifying

processes of pedogenesis using optically stimulated luminescence. Eur. J. Soil Sci. 64, 145e160. http://dx.doi.org/10.1111/ejss.12012. Story, R., Galloway, R.W., van de Graaff, R.H.M., Tweedie, A.D., 1963. General Report on the Land of the Hunter Valley. In: Land Research Series, vol. 8. CSIRO, Melbourne. Sweller, S., Martin, H.A., 2001. A 40,000 year vegetation history and climatic interpretations of Burraga Swamp, Barrington Tops, New South Wales. Quat. Int. 83e85, 233e244. Telfer, M.W., Thomas, D.S.G., 2007. Late quaternary linear dune accumulation and chronostratigraphy of the southwestern Kalahari: implications for aeolian palaeoclimatic reconstructions and predictions of future dynamics. Quat. Sci. Rev. 26, 2617e2630. Telfer, M.W., Hesse, P.P., 2013. Palaeoenvironmental reconstructions from linear dunefields: recent progress, current challenges, and future directions. Quat. Sci. Rev. 78, 1e21. Thomas, D.S.G., Shaw, P.A., 2002. Late quaternary environmental change in central southern Africa: new data, synthesis, issues and prospects. Quat. Sci. Rev. 21, 783e797. Thomas, D.S.G., Brook, G., Shaw, P., Bateman, M., Haberyan, K., Appleton, C., Nash, D., et al., 2003. Late Pleistocene wetting and drying in the NW Kalahari: an integrated study from the Tsodilo Hills, Botswana. Quat. Int. 104, 53e67. Tribolo, C., Mercier, N., Rasse, M., Soriano, S., Huysecom, E., 2010. Kobo 1 and L'Abri aux Vaches (Mali, West Africa): two case studies for the optical dating of bioturbated sediments. Quat. Geochronol. 5 (2e3), 317e323. http://dx.doi.org/ 10.1016/j.quageo.2009.03.002. Walkington, H., 2010. Soil science applications in archaeological contexts: a review of key challenges. Earth-Sci. Rev. 103 (3e4), 122e134. http://dx.doi.org/10.1016/ j.earscirev.2010.09.002. Wasson, R.J., Fitchett, K., Mackey, B., Hyde, R., 1988. Large-scale patterns of dune type, spacing and orientation in the Australian continental dune field. Aust. Geogr. 19, 89e104. Wilkinson, M.T., Humphreys, G.S., 2005. Exploring pedogenesis via nuclide-based soil production rates and OSL-based bioturbation rates. Aust. J. Soil Res. 43, 767e779. Wilkinson, M.T., Richards, P.J., Humphreys, G.S., 2009. Breaking ground: pedological, geological, and ecological implications of soil bioturbation. Earth-Sci. Rev. 97 (1e4), 257e272. http://dx.doi.org/10.1016/j.earscirev.2009.09.005. Wintle, A.G., 1997. Luminescence dating: laboratory procedures and protocols. Radiat. Meas. 27 (5), 760e817.