(PDF) Resolving problematic luminescence

Quaternary Geochronology 45 (2018) 50e73 Contents lists available at ScienceDirect Quaternary Geochronology journal homepage: www.elsevier.com/locate/quageo Research paper Resolving problematic luminescence chronologies for carbonate- and evaporite-rich sediments spanning multiple humid periods in the Jubbah Basin, Saudi Arabia Laine Clark-Balzan a, *, Ash Parton b, c, Paul S. Breeze d, Huw S. Groucutt e, f, Michael D. Petraglia f a Institute of Earth and Environmental Sciences, University of Freiburg, Alberstr. 23-B, Freiburg im Breisgau 79104, Germany b Department of Social Sciences, Oxford Brookes University, Gibbs Building, Gipsy Lane, Oxford OX3 0BP, United Kingdom c Mansﬁeld College, University of Oxford, Oxford OX1 3TF, United Kingdom d Department of Geography, King's College London, Strand, London WC2R 2LS, United Kingdom e School of Archaeology, Research Laboratory for Archaeology and the History of Art, University of Oxford, Oxford OX1 2PG, United Kingdom f Max Planck Institute for the Science of Human History, Kahlaische Strasse 10, 07745 Jena, Germany a r t i c l e i n f o a b s t r a c t Article history: Most of the world's presently hyper-arid desert regions have experienced previous periods of signiﬁ- Received 21 January 2017 cantly higher humidity and milder environmental conditions. The timing of these 'greening events' is Received in revised form critical to research upon global climatic ﬂuctuations and for studies of hominin palaeodemography and 20 June 2017 range expansion, contraction, and extinction, but dating these climatic shifts via terrestrial sedimentary Accepted 20 June 2017 Available online 23 June 2017 records can be difﬁcult. Here, we outline the challenges inherent in the radiometric dating of carbonate- and evaporite-rich sediments preserved in the Jubbah basin (Nefud Desert, northern Saudi Arabia), a critical area for reconstructing the evolution of local hydrological regimes across long timescales. The Keywords: Luminescence dating Jubbah basin is surrounded by sandstone jebels (bedrock outcrops), which have prevented signiﬁcant Quartz leeward dune accumulation for at least 400,000 years. The sedimentary sequences in the basin indicate Feldspar repeated ﬂuctuations between arid and humid climatic conditions, and provide key hydroclimatic re- Post-IR IRSL cords for northern Arabia. Quartz OSL and feldspar pIRIR290 luminescence measurements and radio- Diagenetic alteration carbon dating efforts are reported from four palaeoenvironmental sections in the Jubbah basin. Dates Dose rate from sand-rich levels are relatively unproblematic, but signiﬁcant difﬁculties were encountered when calculating luminescence ages from carbonate and evaporite-rich sediments. Examination of the age- depth proﬁles, elemental composition, and sedimentological characteristics of these sections indicates that both secular disequilibrium and post-depositional alteration of the sediments has resulted in inaccurate dose rate assessment for multiple samples. In particular, we suggest that multiple ground- water pulses in the Jubbah basin have caused carbonate re-precipitation and concurrent uranium enrichment in subsurface deposits, whereas ‘perched’ sections (such as the carbonate-topped remnants reported elsewhere across the Nefud) seem to be free from such alteration. These difﬁculties highlight important considerations for the production of chronologies from comparable settings elsewhere. Careful evaluation of all results, however, yields a robust chronology indicating the presence of varying levels of groundwater from the Holocene, MIS 3, 5, and probably older sediments from MIS 7 through to 9 or 11. We therefore provide a detailed discussion of the production of a reliable chronological framework for the Jubbah basin as an exemplar of the challenges to be overcome in such settings, and the amount of information that can be derived in so doing. © 2017 Elsevier B.V. All rights reserved. 1. Introduction Numerous recent studies focusing on the Arabian Peninsula * Corresponding author. have provided chronologically secure data for key archaeological E-mail address:

[email protected]

(L. Clark-Balzan). (Armitage et al., 2011; Groucutt and Petraglia, 2012; Groucutt et al., http://dx.doi.org/10.1016/j.quageo.2017.06.002 1871-1014/© 2017 Elsevier B.V. All rights reserved. L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 51 2015b; Petraglia et al., 2015; Jennings et al., 2016), palaeontological comprising evaporite-rich and carbonate-rich, waterlain sedi- (Stimpson et al., 2016), and palaeoecological and palaeoclimatic ments, however, is not straightforward. Preservation of charcoals or applications (Rosenberg et al., 2011a, 2013; Parton et al., 2015a). plant macrofossils for radiocarbon dating is rare in this environ- From these data, it is clear that multiple regions in the Arabian ment, and in any case this technique cannot be used for samples Peninsula have been considerably more humid in the past, with with an age greater than approximately 50e55 ka (Pigati et al., signiﬁcant implications for biogeographical responses and popu- 2007). Uranium-series dating can be used to directly date lacus- lation expansions into and through this region. Yet the timing and trine carbonates if they comprise a suitable closed system magnitude of the precipitation events related to such humid pe- (Schramm et al., 2000). Such deposits are rare, though, with riods and their effects on the local and regional biogeography are contamination and open-system behavior common in palaeolake still known only in outline. Marine cores from the Arabian Sea (Des deposits (Debaene, 2003). Photostimulated luminescence dating Combes et al., 2005; Ziegler et al., 2010; Caley et al., 2011), and (Aitken, 1998), in which the burial age of sediment can be derived speleothem records from Oman and Yemen (e.g. Fleitmann et al., from the amount of absorbed energy in mineral grains, is currently 2003, 2011), have led to models of Late Pleistocene palaeoclimatic the most generally applicable technique to the creation of long- variability and environmental response at a regional scale (see range palaeoenvironmental frameworks in arid environments. Parton et al., 2015a,b for a full review). Alluvial, ﬂuvio-lacustrine This method utilizes ubiquitous quartz and feldspar grains as nat- and aeolian records (Blechschmidt et al., 2009; Atkinson et al., ural dosimeters and can potentially yield ages ranging from cen- 2011; Rosenberg et al., 2011a, 2013; Parton et al., 2013; Farrant turies to hundreds of thousands of years. et al., 2015; Parton et al., 2015a) have also provided insights into Dating long depositional sequences of waterlain deposits, while local hydroclimatic responses to arid-humid shifts, of signiﬁcance desirable from a palaeoenvironmental perspective, can cause for both human and animal range expansion and contraction, yet several practical difﬁculties. First, luminescence signal saturation these records are spatially and temporally heterogeneous. and age underestimation in quartz OSL dating have been encoun- Regions that may have been important biogeographic corridors tered in multiple environments around the world. In the Arabian during humid periods, such as northern Saudi Arabia (Breeze et al., Peninsula, such underestimation can begin at signal levels as low as 2016), lack published high-resolution, terrestrial hydroclimatic ar- ca. 100 Gy, or ca. 50e75 ka given the environmental dose rates in chives bridging multiple humid periods. Sabhkas, low-lying areas this region (Rosenberg et al., 2011b). Feldspar minerals offer a with deposits of evaporites, carbonates, and clays that form due to brighter signal and thus may extend this age range potentially to repeated, short ﬂooding events, may preserve useful proxies but several hundred thousand years (Huntley and Lamothe, 2001), but have not been comprehensively studied in this region. Sedimentary they are a complex mineral family, which has not yet been char- records from the Tayma sabhka on the Western edge of the Nefud acterized for accuracy and precision in this region. Second, lumi- (Fig. 1) have provided high resolution sedimentological data tracing nescence dating relies on accurate estimation of the average the development of hypersaline conditions during the early Holo- environmental dose rate during burial. In waterlain deposits, some cene humid period, however, there are no published pre-Holocene of the assumptions generally used for such calculations may be deposits at Tayma (Ginau et al., 2012). Extensive investigations in invalid due to chemical conditions at deposition (Krbetschek et al., both the Nefud and the Empty Quarter (Rub' al Khali) sand seas 1994). Even more concerning is recirculation of groundwater have demonstrated that remnant Pleistocene palaeolake sediments through carbonate and evaporite-rich deposits, which can lead to are typically preserved as scattered outcrops of consolidated, signiﬁcant post-depositional alterations (Dill, 2011). carbonate-rich material found within interdunal basins (e.g. In this study, we report investigations of the Jubbah basin as a Rosenberg et al., 2011a; Rosenberg et al., 2013). These small sec- high resolution palaeoenvironmental record spanning multiple tions, less than 0.5 m in depth to a rare maximum of nearly 3 m, humid periods. A chronometric framework including quartz and range in age from the Early Holocene to at least MIS 11, however, feldspar luminescence ages and radiocarbon dating results is examples of multiple stages of humidity in a single basin are developed for deposits from four palaeoenvironmental sequences, infrequent (Rosenberg et al., 2013). and we discuss the challenges entailed in this endeavour. Detailed In this context, the Jubbah basin seems to be unique in the Nefud sedimentological and palaeoenvironmental proxies will be dis- and possibly in Arabia as a whole (Whitney et al., 1983). Situated cussed in a related publication. approximately 50 km from the southern margin of the Nefud sand sea, this basin is bedrock-deﬁned and thus protected from sand 2. Site descriptions dune encroachment, with high potential for providing long- duration composite and/or continuous records that span multiple The Nefud sand sea in northern Saudi Arabia is one of three humid phases. Early investigations at Jubbah recorded a ca. 26 m connected sand seas, the others being the Rub’ al Khali and Ad- composite section (Garrard et al., 1981) including clays, carbonates Dahna, which together comprise approximately one-third of the and sands. More recently, both radiocarbon dating and lumines- land surface of the Arabian Peninsula. The Nefud itself is the second cence dating have provided data from disparate sections that largest and most intensely studied of these, consisting of indicate the presence of fresh water ecology during the Holocene ~57,000 km2 of predominantly barchanoid dunes in the south-west (Crassard et al., 2013; Hilbert et al., 2014), while minor lacustrine to western regions and linear dunes elsewhere (Edgell, 2006). As deposits of Late Pleistocene age have also been reported (Petraglia shown in Fig. 1A, it is bounded by Palaeozoic sandstones to the et al., 2012). Several lithic scatters in the Jubbah area ranging in age west, Cenozoic sandstones and limestones to the north, and from the Lower Palaeolithic (Shipton et al., 2014), through the Mesozoic formations to the east. To the south, the Nefud is bounded Middle Palaeolithic (Petraglia et al., 2011, 2012; Groucutt et al., primarily by outcrops of the Palaeozoic Saq Sandstone formation. 2015a), into the Epipalaeolithic (Hilbert et al., 2014) and Holocene The Saq aquifer is one of the most important aquifers in the Arabian (Crassard et al., 2013; Guagnin et al., 2017) also suggest that the Peninsula, having been intensively utilized since the 1980's for basin provided an important and recurring long term water source agricultural development in a number of nearby provinces (Zaidi in the region. et al., 2015). The southern edge of the Nefud directly to the south The sedimentary deposits in the Jubbah basin provide an of the Jubbah basin is bounded by the outcropping Precambrian important palaeoenvironmental record for the region. Providing a Arabian shield, as well as some younger granitic plutons such as the robust chronometric framework for these key sequences Jabal Aja or Aja Massif (Kellogg and Stoeser, 1985). 52 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 Fig. 1. Landscape and geological setting of studied region. (A) Geological map of the Nefud Desert and surroundings, redrawn from Whitney et al. (1983). (B) Location of the studied sections within the Jubbah basin. Roads, buildings, and centre pivot irrigation ﬁelds comprising the town of Jubbah now cover the dune-free area. L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 53 The Jubbah depression itself is a large endorheic basin located 2.3. JB2 near the southern edge of the sand sea (Fig. 1B). Pale carbonate-rich sands and silty carbonates are exposed to the east of a string of The JB2 section is located within the central Jubbah basin, outcropping Saq formation jebels running approximately N-S, approximately 5 km to the east of Jebel Umm Sanman (Fig. 1B, which have prevented sand accumulation over an area measuring Fig. S3). Quarry excavations were extended to reveal a total depth of approximately 20 km EeW and 6 km N-S. This primary basin occurs ca. 8.5 m of sediments, which were logged and sampled. The in the lee of Jebel Umm Sanman which rises ~200 m above the basin sedimentary sequence at JB2 is similar to JB1, having heavily iron- ﬂoor, but there are several smaller basins exposed at the base of stained greenish clays at the base, overlain by greenish sands, and other local jebels, such as in the lee of a line of jebels approximately an interstratiﬁed sequence of marls throughout the lower ca. 4 m. 10 km to the southwest. Currently, the region receives less than Unlike JB1, however, these progress into well-developed gypsum 90 mm of precipitation annually (Edgell, 2006), and the wider re- beds and interdigitated gypsiferous marls from ca. 4 m, and they do gion is classiﬁed as a desert (Bwh) according to the Ko €ppen and not display the extensive evidence for vegetation development Geiger climatic classiﬁcation. The town of Jubbah itself, however, present at JB1. Gypsum crystals, however, are similarly well sustains a population of several hundred people and several dozen developed, though the ﬁne marl/gypsum laminations present at JB1 center point irrigation ﬁelds due to the lack of sand-cover and are not seen in this section. outcropping Saq formation in this region (Hussein et al., 1992). The Saq aquifer yields the only near-surface water resource for at least 2.4. JB3 50 km to the SE/SW and several hundred kilometers in other directions. This section was identiﬁed as a consequence of road construc- Between 2010 and 2013, the mapping, excavation and sampling tion in the environs of the archaeological site of Jebel Umm Sanman of multiple palaeoenvironmental and archaeological sites within and 1 (JSM-1) (Petraglia et al., 2012) (Fig. 1B, Fig. S4). JB3 is located surrounding the Jubbah basin was undertaken. New chronometric approximately 100 m east of JSM-1 and the base of the jebel. Recent data is presented here for four palaeoenvironmental sections quarrying had revealed nearly 2 m of horizontally stratiﬁed car- sampled during the 2013 season (Fig. 1, Figs. S1eS4): three of these bonate rich sands, silts, and diatomites, underlain by 20e30 cm of are unpublished (JB1, 2, 3), and new data is presented from the grey-brown sand. Although clearly horizontally bedded, the previously published site of Al-Rabyah (ARY) (Hilbert et al., 2014). exposed sequence of sediments displays a distinct basinal structure indicative of a lake margin environment. The sequence at JB3 is 2.1. Al-Rabyah (ARY) underlain by a massive unit comprising well-sorted, ﬁne-medium red sands indicative of aeolian deposition. These are overlain by an This site comprises a surface lithic scatter (approx. interstratiﬁed sequence of silts, muds, gypsiferous marls, diatomite, 200 m 60 m) eroding from a low relief mound (Hilbert et al., marls and silt-sands, with high concentrations of mollusc shells and 2014). A trench (8 1 m) excavated through this mound yielded shell fragments throughout the upper ca. 0.7 m. a 2.5 m alternating sequence of discrete carbonate-rich and sand- rich sedimentary units, corresponding to lacustrine and palus- 3. Materials and methods trine deposition (Fig. 1B, Fig. S1). The trench was sampled for luminescence dating and sedimentological studies, and the recov- 3.1. Sedimentological analyses ered lithics were analysed (Hilbert et al., 2014). Published quartz multigrain ages for the sand-rich, carbonate-poor levels indicate All four sections were excavated and logged in the ﬁeld, with that excavated levels were deposited between 12.2 ± 1.1 ka and samples retrieved for further palaeoenvironmental analyses. A 6.6 ± 0.7 ka. Stratiﬁed Epipalaeolithic artefacts were discovered in a high-resolution study of these concerning lake development within unit with an age of 10.1 ± 0.6 ka. The capping carbonates were not the Jubbah basin is forthcoming. Analysis of carbonate content originally dated but were thought to have been deposited soon (LOIcarb) was conducted following the standard procedure after the underlying (ca. 6.6 ka) unit. Molluscan fauna were only described by Dean (1974) and Heiri et al. (2001). recoverable from these upper carbonates; these point to the pres- ence of a perennial, marshy water body. 3.2. Radiocarbon dating 2.2. JB1 Seven samples from JB1 were submitted to the Oxford Radio- carbon Accelerator Unit (ORAU) for dating (Table 1). These This section was exposed at the base of a quarry adjacent to Jebel comprised two charred plant macrofossils and ﬁve bulk sediment Ghawtar, one of the eastern jebels of the Jubbah basin (Fig. 1B, samples selected from layers found to be relatively rich in organic Fig. S2). Quarrying had exposed a ca. 3 m sequence of gypsiferous, carbon by initial sedimentological analysis. Sample pretreatment laminated sediments, and further excavations revealed a 9.5 m and AMS protocols are described in Bronk Ramsey et al. (2002) and section, which was logged and sampled. The sedimentary sequence Bronk Ramsey et al. (2004). Uncalibrated radiocarbon dates are at JB1 comprises thick, greenish, heavily iron-stained clays with calculated using a half-life value of 5568 years and reported in numerous root voids, overlain by coarse, poorly sorted sands and radiocarbon years BP (from 1950 CE). Dates have been calibrated gravels. These are overlain by a series of interstratiﬁed calcareous with the IntCal13 calibration curve (Reimer et al., 2013) via OxCal silts, sands and marls occasionally featuring well-developed diat- v4.2 (Bronk Ramsey, 2009). omite and gypsum lenses. The upper portion of the JB1 sequence displays increasing gypsum development, with the upper ca. 3 m 3.3. Luminescence dating comprising gypsiferous marls featuring well-developed, needle- like gypsum crystals. Marls and gypsum facies are ﬁnely interdig- 3.3.1. Sampling and extract preparation itated, with millimeter-scale laminations, particularly in the upper Samples were collected by hammering segments of plastic or 2.5 m. Numerous organic-rich humic layers are also conspicuous metal pipe into cleaned sections. Each tube was removed, and the throughout the upper ca. 3 m, with units featuring large quantities ends were capped. Sediment samples for water content determi- of plant and root impressions and locally calciﬁed reed beds. nation were also collected. Each section was sampled 54 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 Table 1 Radiocarbon dating samples (all collected from section JB1): depths and dating results. ORAU Sample Code Sample Type Depth (m) Results Uncalibrated Calibrated (14C years BP) (calBP, 95.4% range) OxA-30943 Plant fragment (charred) 2.65e2.68 7925 ± 45 8980e8609 P-37215 Plant fragment (charred) 2.58 Failed NA P-38476 Bulk Sed. 0.40e0.50 Failed NA P-38477 Bulk Sed. 0.85e0.90 Failed NA P-38478 Bulk Sed. 3.05e3.15 Failed NA P-38479 Bulk Sed. 3.40e3.50 Failed NA P-38480 Bulk Sed. 4.26e4.28 Failed NA approximately every 50 cm, where possible. Visibly sand-rich units the University of Freiburg; equipment details are identical to the were preferred, followed by carbonate-rich units; visibly gypsum- Oxford smart. rich units were avoided unless there were no alternate units to Quartz equivalent doses (Des) are derived from measurements be sampled nearby. Detailed photographs of sample locations and made for 15 to 20 multigrain aliquots (2 mm diameter, unless surrounding sediments are included in the supplementary mate- otherwise noted). Luminescence signals were measured via a blue rials (Figs. S1eS4). In total, we present ages or radioisotope con- light stimulation, single aliquot regenerative protocol (Murray and centrations for 10 samples from JB1, 11 from JB2, and 5 from JB3. Wintle, 2000, 2003), which included repeated (recycled) and zero From ARY, both quartz and feldspar results are presented for two dose steps, as well as an IR depletion step (Duller, 2003). Aliquots unpublished samples, and feldspar results are given for the samples were preheated at 260 C and 240 C for 10 s after regeneration and published in Hilbert et al. (2014). test dose irradiations, respectively. All samples were transported to the Research Laboratory for Ten multigrain feldspar aliquots (1 mm diameter) were also Archaeology and the History of Art (University of Oxford) for measured for each dating sample via the pIRIR290 protocol of Thiel preparation and measurement. Selected samples were opened and et al. (2011a) if sufﬁcient grains were present. In this protocol, the prepared in dim amber light (590 nm). Coarse grain quartz and ﬁrst IRSL bleach (50 C) is intended to deplete the portion of the feldspar were extracted according to the following methods. For signal that may have been affected by anomalous fading, while the samples collected from site ARY, 180e255 mm quartz was extracted second, elevated temperature IRSL stimulation (290 C) yields what according to the methods previously used at this site (Hilbert et al., is thought to be a more stable signal. Each SAR cycle also includes 2014); this procedure includes a hydroﬂuoric acid etch to remove repeated and zero dose regeneration steps for quality control pur- quartz grains' alpha irradiated exteriors. Feldspars (125e180 and poses. Des for the calculation of luminescence ages are derived from 180e255 mm) were concentrated by density separation (sodium the second stimulation (pIRIR290). A dose recovery experiment was polytungstate, r ¼ 2.58 g cm 3) after wet sieving, followed by hy- also conducted for sample JB1-OSL5. Sixteen aliquots (ca. 1e2 mm drochloric acid (10%) treatment until reaction ceased, drying, and diameter) were bleached under an Osram Ultra-Vitalux lamp for second sieving to remove any previously carbonate-coated 24 h. Twelve aliquots were then beta irradiated (259 Gy), and the grains < 180 mm in diameter. Sections JB1, JB2 and JB3 comprised given laboratory dose was measured via the above protocol, with much ﬁner sediment with a high proportion of gypsum, therefore acceptance of aliquots and data analysis as described below. quartz (125e180 mm) and feldspar (180e255 mm) extraction was Remaining residual dose was measured from the other four altered as described in Clark-Balzan (2016). Brieﬂy, this procedure aliquots. utilizes two density separations, r ¼ 2.35 g cm 3 and 2.58 g cm 3. Full pIRIR290 signal resetting at deposition was investigated in Three mineral extracts are obtained: r < 2.35 g cm 3 (primarily two ways. We tested for the existence of an unbleachable pIRIR290 gypsum, not discussed in this paper), 2.35 g cm 3< r < 2.58 g cm 3 signal by measuring a modern, aeolian-deposited sample from the (feldspar), and r > 2.58 g cm 3 (quartz). The latter two extracts may Arabian Peninsula. This sample was collected from the top centi- not be pure if gypsum-coated quartz or feldspar is present, how- meter of the surface sand layer at Mundafan al-Buhayrah (see ever, Clark-Balzan (2016) showed that luminescence signals from Groucutt et al., 2015b for site details). Coarse-grained feldspars each mineral can be discriminated. Therefore, we refer to these (180e255 mm) were extracted according to the ARY procedure extracts simply as quartz and feldspar in this study. Neither quartz described above, and twenty aliquots (1 mm diameter) were pre- nor feldspar grains collected from sites JB1, JB2, and JB3 were pared and measured. Aliquots were accepted for analysis as etched with hydroﬂuoric acid; this change in procedure both described below, but the zero ratio criterion was altered so that an maximized the number of available grains from problematic aliquot was accepted if the normalized post-zero measurement was gypsum-rich sediments, and removed concerns about heteroge- less than 0.4 Gy. Second, partial bleaching and the inheritance of a neous etching and etch depth uncertainty. residual geological signal was tested. No modern depositional an- alogues exist in the Nefud, therefore Des (IR50 and pIRIR290) were 3.3.2. Equivalent dose measurements calculated for feldspars from ARY and compared to quartz ages in an Due to the number of aliquots measured during this project, attempt to characterize possible partial bleaching in the Jubbah equivalent dose measurements for dating and residual estimates basin (see section 4.2.2 for further discussion). (ARY) were made on two luminescence readers, a Risø TL/OSL TL- Anomalous fading rates (IR50 and pIRIR290) were measured from DA-15 (Bøtter-Jensen et al., 2000, 2003) and a lexsyg smart the ARY feldspars via the pIRIR290 protocol according to the single (Richter et al., 2013) at the University of Oxford. Both machines aliquot regeneration method of Auclair et al. (2003). Three previ- incorporate automatic irradiation, optical stimulation, heating, and ously measured and accepted aliquots from each sample and each detection facilities. Full equipment details and measurement pa- grain size were repeatedly dosed, preheated, and stored before rameters, including illumination power, preheat parameters, and signal measurement. Storage times ranged from immediate mea- ﬁlters, are given in the Supplementary Material (Tables S1 and S2). surement to between 2 and 5 days for the longest storage period. The dose recovery experiment was performed with a lexsyg smart at For any given sample, no signiﬁcant differences were noted L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 55 between fading rates calculated for the two grain sizes, therefore all spectrometer. This spectrometer was calibrated using the doped aliquots from each sample were grouped together to calculate ﬁnal concrete blocks at the University of Oxford (Rhodes and g-values. Fading-corrected ages have been calculated according to Schwenninger, 2007), and dry gamma dose rates were calculated Huntley and Lamothe's (2001) procedure: sample-speciﬁc fading via the threshold technique (Mercier and Falgue res, 2007; Duval rates are used for samples from ARY, but an average pIRIR290 g- and Arnold, 2013), then corrected for on-site water content. A 5% value (m ± 1 s) from all ARY feldspar aliquots was used to correct RSE was used for all dry gamma dose rates. For sample ARY-OSL4, ages for samples collected from JB1, JB2, and JB3. Uncorrected ages the gamma dose rate was calculated from the measured are preferred; this will be discussed in section 4.2.2. elemental concentrations, with Gue rin et al. (2011) conversion Luminescence signals were analysed via Luminescence Analyst factors. Cosmic dose rates are calculated using the modern day v4.11. Due to the differing power outputs of the Risø and lexsyg LEDs burial depths and sediment overburden density of 1.8 ± 0.1 g cm 3. (both blue and IR), signal integration ratios were adjusted for each Present day water contents (%, masswater/masswet sediment) were machine so that approximately the same proportion of the total measured by weighing water content sediment samples before and signal was used for the initial signal measurement. Late background after oven-drying (at least 48 h, 60 C). Average water content subtraction was used for all analyses. Speciﬁc integration times are values during burial were then assigned in two groups: if the given in Table S2 (Supplementary Material). Aliquots were accepted present day water content was <5%, a 5 ± 3% average value was for analysis if they met the following criteria: used, if >5%, a 10 ± 3% value was used. These values were chosen to reﬂect higher water content during humid periods, but also to 1. A detectable net natural test signal greater than three sigma include heterogeneity in the moisture content of the sections due above the background signal, to varying porosity and the inﬂuence of the complex ground water 2. Test dose error less than 20% of the calculated test dose, system in the Jubbah region. Dose rate corrections were then 3. Recycling ratios less than >10% from unity or overlapping with calculated using the factors from Zimmerman (1971) and Aitken unity at an uncertainty level of 2s, and Xie (1990). 4. The calculated zero-ratio less than 5% of the magnitude of the normalized natural signal (LN/TN) [quartz and feldspar] or not 4. Results and discussion indistinguishable from 0 at an uncertainty level of 2s [quartz only]. This is discussed in Clark-Balzan (2016), 4.1. Radiocarbon dating 5. The IR depletion ratio greater than 90%, or indistinguishable from unity at an uncertainty level of 2s [quartz only]. Six of seven samples failed due to low radiocarbon yield (Table 1). The only successful sample (OxA-30943), collected from These criteria are sufﬁcient to reject aliquots with a signal 2.65 to 2.68 m depth in section JB1, yielded an uncalibrated age of dominated by gypsum luminescence, as shown in Clark-Balzan 7925 ± 45 14C years BP. When calibrated via the ‘IntCal13’ curve (2016). Sample Des and overdispersion values were calculated (Reimer et al., 2013) using OxCal v4.2 (Bronk Ramsey, 2009), this from accepted aliquots using the central age model (Galbraith et al., sample dates to 8980-8609 calBP (95.4% range). 1999). For a few samples (see results and discussion), the minimum Several factors may affect the accuracy of this radiocarbon date age model has also been applied (Galbraith et al., 1999). as an estimator for the depositional age of its sedimentary unit. First, it is possible for a sample to physically move within the 3.3.3. Dose rates sediment after deposition. The dated sample, a charred plant All dose rates have been calculated with DRAC, the ‘Dose Rate fragment, was extracted from a distinct, dark grey organic-rich and Age Calculator’ (Durcan et al., 2015; Durcan and King, 2017). layer, approximately 3 cm thick. This region of the section con- Measurement procedures and assumptions for all user-input values sists of ﬁnely interdigitated marls and organic silts, with preser- are described in this section. Portions of each sample, including vation of laminations on the scale of millimeters to centimeters and several that were not subsequently dated, were submitted for numerous plant remains. No visible evidence for bioturbation was fusion ICP-MS (Activation Laboratories, Ltd., Canada) in order to apparent. Secondly, macrofossils from woody plants may over- obtain uranium, thorium, and potassium concentrations. External estimate sediment age by more than 1000 years, both because they beta dose rates were calculated from these concentrations (5% may include much older material (up to several centuries) within relative standard error, ‘RSE’) using the dose rate conversion factors their inner rings whilst living and because they may remain in the of Guerin et al. (2011) and grain size attenuation factors for quartz landscape for signiﬁcant periods of time prior to burial (Oswald and feldspar from Gue rin et al. (2012). No external rubidium dose et al., 2005). Though the plant-type of the charred macrofossil rate was calculated. External alpha dose rates were also calculated could not be identiﬁed, it is highly improbable that such long- from the conversion factors in Gue rin et al. (2011), with grain size residence woods would have existed in the Nefud during the Ho- attenuation values from Brennan et al. (1991). Coarse-grain alpha- locene (Hoelzmann et al., 1998; Watts and Al-Naﬁe, 2013). Finally, efﬁciency factors (a-values) of 0.10 ± 0.02 and 0.15 ± 0.05 were used and most importantly in this environment, there may be a fresh- for quartz (Olley et al., 1998) and feldspar (Balescu and Lamothe, water reservoir effect due to the existence and uptake of ‘ancient’ 1994), respectively, following common practice in the lumines- carbon, depleted in 14C, by plants within the Jubbah basin. Both the cence community (Durcan et al., 2015). For etched quartz grains Saq sandstones and extensive Jubbah basin carbonate beds may (samples ARY-OSL4 and ARY-OSL5), the external alpha dose was function as sources of ancient carbon, and indeed the Saq Aquifer considered to be zero; the beta attenuation factors were adjusted pore water has yielded a radiocarbon age of 20,400 ± 500 14C years by assuming a 10e20 mm etching depth and calculated based on (Thatcher et al., 1961). The inﬂuence of this sedimentary carbon Brennan (2003). Potassium feldspars also include an internal dose depends upon whether the plant has ﬁxed carbon from the rate, arising from potassium and rubidium contained within the groundwater; fully submerged and emergent (submerged roots, crystal matrix. This proportion of the dose rate was calculated some vegetation above water surface) plants will uptake different assuming each grain contains by mass 12.5 ± 0.5% potassium proportions of this older carbon (Marty and Myrbo, 2014), while (Huntley and Baril, 1997) and 400 ± 100 ppm rubidium-87 (Huntley terrestrial plants ﬁx atmospheric carbon and will not be affected by and Hancock, 2001). Gamma dose rates were measured on site for such a reservoir (Grimm, 2011; Bronk Ramsey et al., 2012). As the all but one sample (ARY-OSL4), utilizing an Inspector 1000 gamma plant-type has not been identiﬁed, some age overestimation due to 56 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 a localized reservoir effect is possible. Finally, contamination by rejected for either recycling or zero ratio criteria. We conclude that modern carbon, which may cause an age underestimate, is highly the rejection criteria are appropriately excluding aliquots from unlikely given the young age of this sample and the pretreatment which the measured luminescence signal is not emitted by quartz used. Overall, in our estimation, this radiocarbon date is likely an grains, and that this protocol and the similar SAR OSL protocols accurate estimate of the age of sediment deposition at 2.65e2.68 m used in other studies in the region (Petraglia et al., 2011, 2012; depth in JB1, with some possibility that it may overestimate the Crassard et al., 2013) are appropriate for the extracted quartz. true age by up to a few centuries. Two dose regeneration curves are presented in Fig. 2, one of which saturates at a typical level for the Jubbah basin, and one that 4.2. Luminescence dating saturates at a much higher dose. For the samples presented here, at least one saturated aliquot is present in all samples with a popu- 4.2.1. Quartz characteristics lation CAM De > 200 Gy. Yet a number of individual aliquots have High proportions of aliquots (up to 13 of 20, 65%) were rejected been measured that yield De estimates in the range 400e500 Gy for insufﬁcient test dose signal or high test dose error in samples without saturation. The highest measured sample CAM De, with a high gypsiferous content (e.g JB2-OSL4) (Table 2A). A further, 288.62 ± 41.24 Gy (JB1-OSL4), yielded only three aliquots rejected typically low but sometimes signiﬁcant, number of aliquots has due to signal saturation. By contrast, multigrain and single grain been rejected by the IR depletion criteria and probably contained quartz OSL measurements from Mundafan Al-Buhayrah in the feldspars (e.g. JB3-OSL3). Relatively few aliquots, however, were Empty Quarter (Groucutt et al., 2015b), Jebel Faya (Armitage et al., Table 2 Rejection criteria results for measured aliquots (A) Quartz, (B) Feldspar (dating, residuals, and DRE reported). A) Quartz (2 mm aliquots, except where noted) Field Code Lab Code Measured (#) Signal Strength Tx err Zero Ratio Recycling Ratio IR depletion Saturated Accepted Aliquots (#) a ARY-OSL4 X6141 15 1 14 ARY-OSL5a X6142 15 1 2 12 JB1-OSL1 X6246 18 2 5 1 1 9 JB1-OSL2 X6247 18 3 1 14 JB1-OSL3 X6248 18 3 2 14 JB1-OSL4 X6249 18 2 3 2 3 9 JB1-OSL5 X6250 18 2 2 4 10 JB2-OSL1 X6216 18 6 12 JB2-OSL2 X6217 18 2 4 1 1 10 JB2-OSL3 X6218 18 2 5 11 JB2-OSL4 X6219 20 3 10 7 JB2-OSL5 X6220 18 1 1 1 2 1 5 7 JB2-OSL9 X6223 18 2 2 3 11 JB3-OSL1 X6231 18 3 1 14 JB3-OSL2 X6232 18 2 2 14 JB3-OSL3 X6233 18 8 10 JB3-OSL4 X6234 18 2 5 11 B) Feldspar (1 mm aliquots) Field Code Lab Code Grain Size Measured (#) Signal Strength Tx err Zero Ratio Recycling Ratio Saturated Accepted Aliquots (#) (mm) [IR50]b ARY-OSL2 X6139 125e180 10 1 [1] 9 [9] 180e255 6 1 [2] [2] [1] 5 [1] ARY-OSL3 X6140 125e180 10 1 1 [1] 8 [9] 180e255 10 8 [8]c 2 [2] ARY-OSL4 X6141 125e180 10 [1] 10 [9] 180e255 e e ARY-OSL5 X6142 125e180 10 1 [3] [2] 9 [7] 180e255 e e ARY-OSL7 X6144 125e180 10 10 [10] 180e255 24 14 [16]c 3 [1] 1 [1] 6 [6] JB1-OSL1 X6246 180e255 3 3 0 JB1-OSL2 X6247 180e255 10 1 1 5 1 2 JB1-OSL3 X6248 180e255 10 5 1 5 JB1-OSL4 X6249 180e255 7 1 6 JB1-OSL5 X6250 180e255 10 10 DRE 12 1 11 Residual 4 See text JB1-OSL6 X6251 180e255 7 1 1 5 0 JB1-OSL7 X6252 180e255 10 2 8 JB1-OSL8 X6253 180e255 10 1 1 8 JB1-OSL13 X6258 180e255 10 1 4 5 JB2-OSL14 X6228 180e255 8 1 1 6 JB3-OSL4 X6234 180e255 10 4 1 4 MUN13-BLEACH e 180e255 20 [3] 20 [17] a 4 mm aliquots. b Aliquots accepted for pIRIR290 De analysis and IR50 analysis were sometimes different populations. c Aliquots were ﬁrst screened for a measurable natural signal, then those with a signal were measured (in addition to the standard Tx signal strength criterion). This procedure should not affect the measured Des, as residual measurements on the modern surface sample indicated that Des as low as ca. 1 Gy could be measured with good signal strength. L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 57 Fig. 2. Quartz regeneration curves from sample JB1-OSL5, unsaturated (A) and saturated (B). Natural decay curves are shown inset. 2011), and even neighboring sites in the Jubbah basin such as JQ-1 distributions tend to by symmetric, with only three sample pop- and JKF-1 (Petraglia et al., 2011, 2012), yield grains and aliquots that ulations yielding positive unweighted skewness values greater than saturate at much lower levels. These differences may be due to one (Fig. 3). Both high overdispersion and skewness are present variable grain sources, but the majority of sands from both the only in aliquot populations ARY-OSL5 and JB3-OSL2 (Figs. 6 and 7), Nefud and Empty Quarter seem to be derived primarily from which have overdispersion values of 49 and 57%, respectively. This weathering of Palaeozoic sandstone formations as seen in Fig. 1 combination of characteristics is likely due to the inclusion of (Garzanti et al., 2013). partially bleached quartz grains within these samples (Olley et al., Overdispersion values for the Jubbah sites vary between ca. 10 2004), in which case the minimum age model can be used to and 60%, with a majority of values < 36% (Table 3). Overdispersion extract the De corresponding to the most fully-bleached grains values > 36% are measured only for the lower carbonate from Al- (Galbraith et al., 1999). The use of the minimum age model will be Rabyah (ARY-OSL5) and the three carbonate-rich samples from discussed further in section 4.2.3. JB3 (JB3-OSL1eJB3-OSL3). Comparative overdispersion values for quartz grains from the Arabian Peninsula are relatively scarce. 4.2.2. Feldspar characteristics Petraglia et al. (2011) measured values between 24 ± 3% and Feldspar aliquots measured from Jubbah basin sites were 29 ± 4% for small aliquots (ca. 50 grains) from both a calcrete and rejected primarily for low intensity luminescence signals (ARY) and palaeosols, and Rosenberg et al. (2013) calculated overdispersion zero ratio magnitude (JB1-3) (Table 2B). Half or more of the values between 12% and 28% for a suite of primarily aeolian quartz measured aliquots were rejected from samples ARY-OSL3, ARY- multigrain samples (unknown aliquot size) from the Nefud. De OSL7, JB1-OSL2, JB1-OSL3, JB3-OSL4, and JB1-OSL6 due to one of Table 3 Equivalent doses measured for all mineral fractions used to calculate luminescence ages. Field Code Depth (m) Quartz Feldspar Accepted De (Gy) Overdispersion (%) Accepted De (Gy) Overdispersion (%) ARY-OSL4 0.45 14 9.22 ± 0.50 19.21 ± 4.00 ARY-OSL5 0.56 12 17.94 ± 2.50a 57.32 ± 12.26 (37.19 ± 6.30) JB1-OSL1 2.50 9 24.94 ± 1.64 13.83 ± 6.09 JB1-OSL2 2.96 14 21.5 ± 1.07 16.13 ± 4.03 2 93.02 ± 6.12b NA JB1-OSL3 3.41 14 213.52 ± 16.43 21.38 ± 6.82 5 299.73 ± 43.24 31.37 ± 10.49 JB1-OSL4 4.16 9 288.62 ± 41.24 35.56 ± 11.74 6 376.02 ± 53.82 32.29 ± 10.85 JB1-OSL5 4.51 8 262.49 ± 28.55 26.46 ± 8.74 10 357.06 ± 28.46 19.43 ± 6.79 JB1-OSL7 6.35 8 698.34 ± 149.38 58.08 ± 15.68 JB1-OSL8 5.50 8 302.45 ± 48.79 43.49 ± 11.9 JB1-OSL13 9.00 5 889.16 ± 209.98 47.96 ± 18.18 JB2-OSL1 0.77 12 5.93 ± 0.32 14.43 ± 4.62 JB2-OSL2 1.57 10 8.52 ± 0.67 21.21 ± 6.39 JB2-OSL3 3.25 11 5.07 ± 0.26 10.67 ± 4.89 JB2-OSL4 3.94 7 9.78 ± 0.79 18.24 ± 6.62 JB2-OSL5 4.15 7 258.26 ± 38.91 35.73 ± 11.83 JB2-OSL9 5.95 11 236.03 ± 23.01 25.82 ± 8.47 JB2-OSL14 8.65 6 844.81 ± 189.89 54.11 ± 16.17 JB3-OSL1 1.20 14 61.63 ± 8.79 52.18 ± 10.31 JB3-OSL2 1.67 14 55.00 ± 6.32a 48.08 ± 9.90 (68.44 ± 9.20) JB3-OSL3 2.07 10 83.60 ± 16.75 62.22 ± 14.42 JB3-OSL4 2.50 11 94.98 ± 9.64 30.83 ± 7.77 4 224.75 ± 26.87 22.54 ± 8.95 a MAM De preferred; CAM De given in parentheses. b Calculated via weighted mean due to small number of accepted aliquots. 58 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 Fig. 3. Abanico plots (Dietze et al., 2016) for accepted quartz aliquots from representative Holocene (A, B) and Pleistocene (C, D) samples. These plots combine a radial plot (left) with a kernel density estimate curve (‘KDE’, right), so that distribution parameters and relationships between error and De can be analysed visually. Automatically determined bandwidth (‘bw’) values used to determine the smoothness of the KDE are given underneath each plot. The central age model De for each distribution is shown with a dotted line. these criteria. Feldspar aliquots measured from other sites in the The implication of low (<2%/decade) g-values for the pIRIR290 Nefud desert are rarely excluded by these criteria (e.g. 4 per 69 signal is still under debate. Several studies have reported mea- aliquots). Aliquots failing either criterion may comprise a signiﬁ- surement of ﬁeld saturated pIRIR290 signals, even though ﬁnite cant proportion of gypsum-coated grains; investigation of mineral fading values can be detected in the laboratory (Buylaert et al., extracts from several of these samples indicated signiﬁcant gypsum 2011; Thiel et al., 2011a; Thomsen et al., 2011). Similarly, quartz contamination, and the gypsum signal is both dim and character- grains, though apparently not suffering from signal fading in na- ized by signiﬁcant thermal transfer (Clark-Balzan, 2016). Feldspar ture, yield a 1%/decade g-value when measured via standard fading aliquots that passed both criteria were relatively bright with good tests (Buylaert et al., 2012). Buylaert et al. (2012) therefore propose recycling ratios. The natural pIRIR290 signal is in saturation for that fading rates of 1e1.5% are artifacts of the measurement pro- multiple measured aliquots, including nearly half of those cess. Fading-corrected pIRIR290 ages also seem to overestimate measured from sample JB1-OSL13 (Fig. 4). Dose recovery results quartz luminescence ages in some cases (Buylaert et al., 2011; Kars suggest that the pIRIR290 protocol is generally suitable for these et al., 2012), and Li et al. (2014) suggest that many uncorrected samples: a central age model De of 0.92 ± 0.05 (normalized by the pIRIR290 ages (70e80% of those compiled) agree with independent given dose) was recovered from 11 accepted aliquots. Residual age control within a two sigma range. Given that the average measurements were not considered to be reliable, as signal mag- pIRIR290 g-value measured in this study is less than 2%/decade, and nitudes for three of four aliquots were similar to the normalized that up to 44% of accepted aliquots from one of the oldest samples zero-dose response. Individual aliquots in the dose recovery measured are in saturation (JB1-OSL13), we suggest that pIRIR290 experiment are quite dim, and we think it is likely that feldspars in equivalent doses should not be corrected for anomalous fading. In this highly gypsiferous sample were measured at essentially a this, we follow the convention of a number of pIRIR290 feldspar single grain level. dating studies (Thiel et al., 2011b, 2012; Roskosch et al., 2012). Minimum fading rates measured from individual ARY aliquots Given the complexity of this issue (see for example Lowick et al., are indistinguishable from zero for both IR50 and pIRIR290 signals. 2012), however, we also provide fading-corrected ages in Table 6 IR50 g-values range up to a maximum of 20.13 ± 3.09%/decade, with for readers' information. an average value of 5.63 ± 4.46%/decade (m ± 1s) calculated from all We have attempted to estimate the magnitude of any pIRIR290 measured aliquots. The pIRIR290 signal yields much lower fading residual signal due to partial bleaching in these primarily low- rates, with a maximum individual aliquot g-value equal to energy, waterlain deposits. Such a residual may comprise both a 4.81 ± 0.49%/decade, and an average of 1.84 ± 1.32%/decade. G- fully unbleachable component (Buylaert et al., 2011), and an value distributions for all measured ARY aliquots can be seen for inherited geological signal due to the relative slowness of pIRIR290 both signals in Fig. S5. Such fading values are within the ranges signal bleaching (Buylaert et al., 2012; Murray et al., 2012). Mea- reported in other publications for both IR50 and pIRIR290 signals surements of modern, aeolian-deposited feldspars from the (see review by Li et al., 2014). Arabian Peninsula suggest a low magnitude for any unbleachable L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 59 Fig. 4. Feldspar regeneration curves (pIRIR290) from sample JB1-OSL13, unsaturated (A) and saturated (B). Natural decay curves are shown inset. signal: Des of only 0.39 ± 0.05 Gy and 1.07 ± 0.10 Gy were measured overdispersion values correlate with higher residual signals, sug- for IR50 and pIRIR290 signals, respectively (Table 2). These distri- gesting that high overdispersion values in feldspars cannot neces- butions have relatively high overdispersion values, with the former sarily be used to suggest the presence of partial bleaching (Fig. S6). yielding a value of 43.80 ± 10.72% and the latter 40.41 ± 6.94% In all cases, measured residual Des are below 50 Gy. (Fig. 5A). This estimate agrees well with the low residuals collated Poor bleaching of the feldspar pIRIR290 signal in waterlain sed- and published by Buylaert et al. (2012) from other coarse grained, iments has been noted in both coastal (Reimann et al., 2011) and modern samples. Kars et al. (2014) have also found that residual lacustrine environments (Buylaert et al., 2013), with inherited sig- doses continue to decrease even after 2e11 days of bleaching with a nals of hundreds of grays measured in laboratory experiments solar simulator, suggesting that any residual measured after two simulating such conditions (Lowick et al., 2012). The magnitude of days of laboratory bleaching (see for example Preusser et al., 2014) the residual signal measured in waterlain deposits for the Jubbah is not likely to accurately estimate the magnitude of a truly basin is quite variable between samples, however, the maximum unbleachable signal, and Thiel et al. (2011a,b) have also preferred measured value of ca. 50 Gy becomes negligible for older samples not to subtract laboratory-measured residuals from pIRIR290 Des. over the timescales considered. We do not subtract any residual No modern depositional analogies were available in the studied correction factor from our measured pIRIR290 De values, as the region, however, the quartz OSL chronology obtained by Hilbert dominant residual component is likely to be the inherited geolog- et al. (2014) allowed us to calculate feldspar pIRIR290 residuals ical signal. due to geological signal inheritance (Table 4). Expected Des (both Overdispersion values measured for 180e255 mm feldspars internal and external) were calculated for each sample based on the (pIRIR290 signal) range between approximately 20% and 60% for quartz OSL age obtained for the same sample. In the case of sample samples with four or more accepted aliquots (Table 3, Fig. 5). There ARY-OSL5, for which the luminescence age is likely under- is no clear relationship with either the samples' central equivalent estimating the depositional age of the sediment (see Section 4.2.3), dose or calculated age. Interpreting the meaning of these values is we have instead used the average quartz age of bracketing samples difﬁcult and beyond the scope of the current paper as, unlike ARY-OSL3 and ARY-OSL7 to calculate the expected feldspar dose. quartz, feldspar overdispersion values result from a convolution of Equivalent doses and ﬁnal ages calculated via the IR50 (fading variance in fading rate and internal dose rate variations, in addition corrected and uncorrected) and pIRIR290 signals (fading uncorrec- to shared sources of variability such as bioturbation, residual doses ted) are presented in Table 4. It is evident that for both IR50 and due to partial bleaching, and microdosimetric effects. External pIRIR290 signals, the 180e255 mm fraction yields lower equivalent microdosimetry has a small effect on coarse feldspars due to their doses than the 125e180 mm fraction. Several causes might explain relatively high internal dose rates, however, and fading rate varia- this systematic difference. First, it may be due to the differing tions should also be low, given the use of the pIRIR290 methodology. numbers of grains on the measured aliquots, with multigrain signal Therefore, we expect primary drivers of these overdispersion averaging of partially bleached single grain De populations pro- values to be related primarily to variations in partially bleached ducing systematically different results. Second, it is possible that residual doses, internal dose rate variations, and possibly bio- the size fractions comprise slightly differing mineralogies with turbation. We will discuss these possibilities further in section varying internal K-content, and these variable dose rates yield 4.2.3.1. different Des over time. Third, a real difference in the bleachability of the grains may be indicated. Several other studies have also re- ported better bleaching for coarser grains in both aeolian (Buylaert 4.2.3. Accuracy and reliability of luminescence ages et al., 2011) and ﬂuvial (see review by Rittenour, 2008) depositional Luminescence ages are presented for each of the four sections settings, which are likely due to differences in sedimentological (Table 6, Figs. 8 and 9). Two key observations are apparent: transport regimes. The calculated residuals for ARY range from 6 to 29 Gy for the 180e255 mm grains, and 7e45 Gy for the 125e180 mm 1. Age-depth reversals are present in all study locations, and for all grains. The magnitude of the residuals does not precisely correlate mineral types (where there is more than one sample measured with expected partial bleaching as inferred from the quartz De for comparison), distributions (Olley et al., 2004; Hilbert et al., 2014), but there 2. Independent age control via radiocarbon dating at 2.65e2.68 m seems to be some correlation. Nor do higher feldspar for section JB1 suggests that the bracketing luminescence dating 60 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 Fig. 5. Abanico plots with accepted aliquot distributions for pIRIR290 feldspar measurements: MUN-BLEACH13 residuals (A), dose recovery experiment results (normalized to the given dose) (B), and distributions with overdispersion > 40% (CeF). Central age model Des (AeF) and minimum age model Des (C-F, assigned overdispersion ¼ 40%) are indicated with dotted and solid black lines, respectively. samples (JB1-OSL1, JB1-OSL2) underestimate deposition age by signal (‘partial bleaching’) (Galbraith et al., 1999; Stokes et al., several thousand years. 2001), physical mixing of depositional units (Gliganic et al., 2015; Hanson et al., 2015), age underestimation due to signal saturation These observations may be explained by either sample/site (Duller, 2012), or anomalous fading (Wintle, 1973; Huntley and speciﬁc inaccuracies in the determination of equivalent doses or a Lamothe, 2001). We have already discussed the measurement pa- failure of the standard dose rate calculation assumptions. Each of rameters for both quartz and feldspar, and concluded that there is these possibilities will be discussed in turn. no evidence for a systematic inaccuracy in De estimation due to this cause. Anomalous fading of the pIRIR290 signal used for age calcu- 4.2.3.1. De estimation. The effects of equivalent dose determination lations has also been considered and assumed to be negligible (see upon age calculation are perhaps the most often discussed reasons section 4.2.2). The other issues will be discussed further here. for unexpected luminescence age results, either in the form of age 4.2.3.1.1. Partial bleaching. Partial bleaching refers to incom- reversals with depth or offsets from independent age controls. plete resetting of the OSL/IRSL signal during transport and depo- Population Des may yield inaccurate ages due to the use of inap- sition of mineral grains (Huntley et al., 1985). The absolute propriate measurement parameters (e.g. preheat temperatures, see magnitude (in gray) of a residual (post-bleaching) signal will Murray and Wintle, 2003), inheritance of a residual geological depend upon the magnitude of the grain's original equivalent dose, L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 61 Fig. 6. Accepted quartz aliquot distributions from ARY-OSL4 (A) and ARY-OSL5 (B), with CAM Des indicated by the dotted lines. bleaching stimulation intensity and wavelength, duration of shallow waters (Berger, 1990). Mineral characteristics will also have bleaching, and the mineral's characteristic crystal properties a strong effect on residual doses for a given depositional environ- (Aitken, 1998; Spooner, 1994). Many of these properties are ment. Experimental sunlight bleaching of quartz causes the photo- controlled by geomorphic transport processes, as variations in stimulated luminescence signal to decrease to a negligible level stimulation duration, intensity, and spectrum are determined by within seconds to minutes, whereas feldspar may retain a signiﬁ- the mode of grain transport. In general, aeolian transport should cant fraction of its original signal after multiple hours of bleaching result in the most complete bleaching (Stokes, 1992), though it is (Godfrey-Smith et al., 1988). Elevated temperature infrared stimu- possible to imagine scenarios such as dust/sand storms that might lation of feldspar, which allows de-trapped electrons to enter result in deposition of incompletely bleached grains. For waterlain higher energy band-tail states and recombine at luminescence sediment, water depth and turbidity strongly affect the amount and centers physically separated by larger distances within the crystal, wavelength of electromagnetic radiation incident upon the grains, allows measurement of a signal less prone to anomalous fading, with grains deposited by fast-ﬂowing, murky waters much less with the increase in temperature providing a concomitant increase likely to be fully bleached than grains transported through clear, in signal stability (Jain and Ankjærgaard, 2011; Jain et al., 2015). The Fig. 7. Accepted quartz aliquots from section JB3 samples, with CAM Des indicated by the dotted lines. 62 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 Table 4 Feldspar equivalent dose measurements and residuals for (A) Fading corrected IR50 and (B) Uncorrected pIRIR290 measurements; all samples were collected from site ARY. a) IR50 Field Code Depth (m) Fading Rate, g Grain Size Fraction: 125e180 mm Grain Size Fraction: 180e255 mm (% per decade) CAM De (Gy) Uncorrected Age (ka) Corrected Age (ka) CAM De (Gy) Uncorrected Age (ka) Corrected Age (ka) ARY-OSL2 0.87 2.46 ± 0.09 21.25 ± 0.78 10.4 ± 0.7 13.3 ± 0.9 17.00 ± 0.65 7.6 ± 0.6 9.6 ± 0.7 ARY-OSL3 1.07 1.55 ± 0.06 50.19 ± 4.50 15.8 ± 1.6 18.3 ± 1.8 40.41 ± 5.79 12.2 ± 1.9 14.1 ± 2.2 ARY-OSL4 0.45 2.07 ± 0.58 12.49 ± 0.92 5.5 ± 0.5 6.7 ± 0.7 ARY-OSL5 1.26a 5.40 ± 0.82 43.52 ± 8.52 9.2 ± 1.9 16.3 ± 3.9 ARY-OSL7 1.80a 3.85 ± 0.13 52.12 ± 2.50 13.5 ± 1.0 20.0 ± 1.5 45.85 ± 11.68 11.7 ± 3.0 17.3 ± 4.8 b) pIRIR290 Field Code Depth (m) Grain Size Fraction: 125e180 mm Grain Size Fraction: 180e255 mm CAM De (Gy) Overdispersion (%) Residual (Gy) CAM De (Gy) Overdispersion (%) Residual (Gy) ARY-OSL2 0.87 39.35 ± 2.35 17.8 ± 4.3 18.98 ± 2.86 28.02 ± 4.63 31.0 ± 12.6 5.64 ± 5.04 ARY-OSL3 1.07 80.79 ± 2.73 8.7 ± 2.6 44.87 ± 3.96 68.99 ± 7.23 12.4 ± 8.7 31.5 ± 7.87 ARY-OSL4 0.45 21.38 ± 1.30 18.7 ± 4.4 6.88 ± 1.82 ARY-OSL5 1.26b 83.95 ± 4.97 16.6 ± 4.4 28.97 ± 6.88 ARY-OSL7 1.80b 82.67 ± 1.89 5.9 ± 2.0 36.16 ± 5.16 76.42 ± 12.91 40.8 ± 12.1 29.09 ± 13.79 a Depths measured from stratigraphic zero, rather than from sloped surface directly above (see Fig. S1). b As note above. same physical process, however, results in slower bleaching of the closely bracketed to between 11.4 ± 0.8 ka and 12.2 ± 1.1 ka by OSL elevated-temperature IRSL signal, which may result in a high ages on sand-rich units above and below (Hilbert et al., 2014). The geological residual for feldspar crystals (see discussion and refer- quartz De population measured for ARY-OSL5 is both highly scat- ences in section 4.2.2). tered (s ¼ 57.3 ± 12.3%) and positively skewed, which suggests that The four study locations preserve a variety of depositional en- these grains may be partially bleached (Fig. 6). MAM-calculated vironments. Basal clayey silt-sands underlying JB1 and JB2 likely ages, however, underestimate the previously estimated deposi- represent undisturbed deposition in still water conditions and the tional age, if an average overdispersion for well-bleached samples dissolution of underlying bedrock material of the Saq sandstone. from the site is used (20%, see Hilbert et al., 2014): the De of ARY- These units are also free from large gravel clast inclusions, inter- OSL5 becomes 17.94 ± 2.50 Gy and the ﬁnal age 5.4 ± 0.8 ka (see bedding or signiﬁcant bioturbation. The overlying very poorly Fig. S7 for De dependence on assumed overdispersion). The CAM sorted, coarse sub-angular-subrounded gravelly and granulitic age (11.2 ± 2.0 ka) is stratigraphically consistent. There are two sands at both sites reﬂect the mobilization and deposition of possibilities that may explain the offset between MAM ages and weathered material from the adjacent jebels. As such, it is possible stratigraphy. The ﬁrst is that the CAM De more accurately repre- that samples from basal clayey or gravelly deposits may be affected sents the appropriate De for input into the age equation. This would by partial bleaching. At JB3, basal sands are very well sorted with no be the case if the skewed and overdispersed population were the signs of vegetation development, clast inclusions or bedding and result of microdosimetric variations in the sediment, and indeed are suggested to be of aeolian origin. Dated samples have been skewness and overdispersion are known consequences of hetero- collected primarily from overlying palustrine and lacustrine con- geneous beta radiation (Nathan et al., 2003; Mayya et al., 2006). texts. Higher gypsiferous and carbonate units are likely to represent This and other studies (Petraglia et al., 2011; Hilbert et al., 2014) of low energy waterbodies with ﬂuctuating extents and depths, analogous deposits in the Jubbah basin, however, show the typical of endorheic and seasonally variable water bodies; sedi- magnitude of the multigrain OSL overdispersion and skewness is mentary and palaeoecological data has been published from Al- highly unusual for similar Holocene deposits. Therefore, this does Rabyah (Hilbert et al., 2014) and analogous site JQ-200 (Crassard not seem to be the most likely cause. Second, we can observe that et al., 2013). Given the low energy setting and a lack of discern- the uranium concentration in this unit is nearly four times that of able ﬂuvial input into the water bodies, the overall probability of the upper carbonate, and thus it contributes ca. 86% of the beta dose partial bleaching should be low in these upper units, however, rate to quartz grains (as calculated via modern elemental concen- there is potential for the input of partially bleached mineral grains tration). We suggest that this sample does contain partially resulting from erosion of the local jebels or during mass move- bleached quartz grains, making the minimum age model a more ments of sandy layers in unusual storm events. The presence of appropriate choice, and that the stratigraphic reversal in the ages is wavy ﬁne laminations of marls and gypsum throughout the upper instead due to an overestimation of the dose rate. This will be ca. 2 m of JB1 are likely to be indicative of lacustrine wave action or discussed further in the next section. Second, as noted previously, subaerial aeolian scour. In addition, if seasonal rainfall variability all three samples collected from JB3 carbonates have unusually were to result in lake desiccation and large-scale deﬂation of sur- high overdispersion values in the context of the Jubbah basin face material, the subsequent reworking of sediments might result samples (Fig. 7). Assuming that partial bleaching can be identiﬁed in partial bleaching. It is noted, however, that there is no physical by a combination of positive skewness and high overdispersion, it evidence for substantial desiccation (i.e. curling of marl beds) was suggested that the luminescence age for sample JB3-OSL2 within any of the sedimentary sequences at Jubbah. It is likely that should be calculated from the MAM De. By analogy with the un- the presence of ﬁne to medium sand grains within lacustrine units derlying sand, and given the high overdispersions of quartz from reﬂect the transport and deposition of aeolian material into often the bracketing samples, we suggest an overdispersion value of 30%, shallow, low energy water bodies. which yields a MAM De of 55.00 ± 6.32 Gy with approx. 80% of Two samples share some luminescence features indicative of aliquots fully bleached. partial bleaching in measured quartz distributions. The ﬁrst of Considering the evidence for partial bleaching in the measured these is sample ARY-OSL5, the age of which has already been pIRIR290 feldspar distributions, it seems likely that some partial L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 63 Fig. 8. Stratigraphy, sediment composition, and geochronology for sections ARY (A) and JB3 (B). Unit 1 of JB3 continues below the presented section for 1 m, but it is unvarying and has been cropped for display. Feldspar dates shown for ARY are calculated from the pIRIR290 signal (125e180 mm). Note that the depths for site JB3 are in relationship to the undisturbed surface. bleaching may occur on a sample-speciﬁc basis (e.g. JB1-OSL2). identiﬁcation of poorly bleached samples via overdispersion is not Measurements on the modern, aeolian sample MUN-BLEACH13 straightforward. Nevertheless, we will consider the potential ef- indicated that a systematically unbleachable residual would be ca. fects of partial bleaching for feldspar samples with overdispersion 1 Gy, which is negligible. ARY quartz and feldspar comparisons, values > 40% in section 4.3. though, suggested that partial bleaching at deposition could lead to Generally low overdispersion and skewness values for quartz inherited signals of variable magnitude, up to ca. 50 Gy. The same and feldspar partial bleaching residuals conﬁrm that partial ARY data shows that distinguishing partially bleached samples via bleaching is unlikely to be a signiﬁcant, systematic issue in the distribution parameters is quite difﬁcult. Several samples with Jubbah basin. Most samples have low overdispersion values in the signiﬁcant residual signals (e.g. ARY-OSL3 and ARY-OSL7) yielded Holocene (around 20% or less) and acceptable values for the symmetric and well-grouped Des, with overdispersion values measured Pleistocene units (30e40%). Moreover, there is no cor- below 10% (Fig. S6). By contrast, overdispersion values of 20e40% relation between overdispersion magnitude and the age-depth are common for samples known to be well-bleached or fairly well- reversals apparent in sections JB1 and JB2 for either quartz or bleached (Fig. 5), including the modern aeolian sample, ARY-OSL2 feldspar; that is, if the reversals in the sediments are due to partial (180e255 mm fraction), and JB1-OSL3, JB1-OSL4, and JB1-OSL5 bleaching, the ‘overestimated’ upper ages should have higher (with similar feldspar and quartz Des). Therefore, we suggest that overdispersions. Finally, feldspar and quartz ages overlap at one 64 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 Fig. 9. Stratigraphy, sediment composition, and geochronology for sections JB1 (A) and JB2 (B). L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 65 sigma for two of four samples dated by both methods from section present in the Nefud by Garzanti et al. (2013), with 0e2% K-feldspar JB1, though some partial bleaching is clearly possible for the by petrographic counting and 0.1%e0.5% heavy minerals by vol- pIRIR290 signal based on the JB1-OSL2 feldspar age. ume, of which zircons comprise approximately one-third. Both 4.2.3.1.2. Physical mixing. Vertical sediment mixing through minerals have also been identiﬁed by the authors in preliminary desiccation cracks in marl beds, insect and/or animal burrowing SEM analysis of Jubbah basin samples. Zircons in particular may activity, or vegetative bioturbation may increase overdispersion have concentrations of tens to thousands of ppm U and Th (Hoskin values and cause age over- or underestimation. No visible evidence and Schaltegger, 2003), and therefore they have a signiﬁcant effect for substantial mixing of sedimentary units has been noted at any of on microdosimetric variation. ‘Coldspots’ can also occur, due to the four sections, though root voids are present in some units at JB1 shielding effects from inert minerals, such as carbonates (Nathan and ARY (Figs. 8 and 9). Horizontally bedded units are visible in et al., 2003). The combination of these microdosimetric effects each section, with clear and distinctive separation of carbonate-rich will cause some amount of scatter in measured De, the exact and sand rich-units (Figs. S1eS4). Millimeter-scale wavy beds of amount of which is highly variable and depends upon factors small gypsum crystals are also apparent in units 18 to 22 at JB1. including grain size, porosity, spatial distribution of radionuclides, These are typically composed of a few millimeters of ﬁne calcareous and total dose rate (Nathan et al., 2003; Mayya et al., 2006; silts, coupled with euhedral and prismatic to lenticular crystals that Cunningham et al., 2012; Gue rin et al., 2015). These can result in are laterally contiguous horizontally over the investigated area inaccurate luminescence ages if misinterpreted as evidence for (several meters wide), with no evidence for post-depositional physical mixing or partial bleaching. In the Jubbah basin, large disturbance. Extremely delicate carbonate casts of reed beds have differences in overdispersion values have been noted (e.g. JB3 vs. also been preserved in the upper 4 m of JB1. Mixing in basal sandy JB1) depending on the age and composition of the sampled unit. clay units is possible, due to root penetration or burrowing, Some of the samples with larger overdispersion values (e.g. JB3- particularly if signiﬁcantly younger lake deposits formed above an OSL1) are likely to be affected by microdosimetry, while others, older base. This is more difﬁcult to rule out visually, and must be having both large overdispersion values and skewed populations considered a possibility. (ARY-OSL5, JB3-OSL2) have been interpreted as partially bleached 4.2.3.1.3. Saturation effects. Due to the age of the sampled sed- sediments. We believe in this case comparison of all Jubbah basin iments, underestimation of absorbed dose due to saturation of al- samples suggests that partial bleaching and the MAM is justiﬁed in iquots must be seriously considered as a potential confounding these latter cases (see section 4.2.1 and 4.2.3.1). Larger-scale spatial factor for De measurements. In particular, Rosenberg et al. (2011a, b) heterogeneity, on the scale of centimeters to decimeters, has been and Groucutt et al. (2015b) have suggested that quartz multigrain accounted for by using a ﬁeld gamma spectrometer to measure the aliquots from the Arabian Peninsula measured with a standard SAR gamma dose rate for all samples except ARY-OSL4. OSL protocol may systematically underestimate Des greater than Waterlain sediments, such as those sampled in the Jubbah basin, ~100 Gy. Sections JB1 and JB2 already yield saturated quartz ali- are also known risks for signiﬁcant temporal variability in dose rate, quots by 4 m depth: samples JB2-OSL5 and JB1-OSL5 yield 5 and 4 which may cause inaccurate luminescence ages (Krbetschek et al., saturated aliquots, respectively, from 18 each measured. This was a 1994). This can occur via two routes: radioactive disequilibrium motivating factor for the measurement of both quartz and feldspar and post-depositional alteration of the sediment. The causes and grains from samples JB1-OSL1 through JB1-OSL5: because the consequences of each of these are discussed in detail below. feldspar signal saturates at a much higher level (in this study near Disequilibrium is deﬁned to be an imbalance in the secular 1000 Gy), feldspar Des should continue to grow normally even as equilibrium of a radioactive decay chain, i.e. that steady state in the quartz begins to saturate. Any underestimation by the quartz De which the decay rates of all daughters and parents in a radioactive measurement protocol would therefore be apparent as a systematic chain are equal. Disequilibrium arises due to the differential solu- offset between the quartz and feldspar ages. It is apparent from bility of various chemical species of parents and daughters in the Fig. 9 that quartz and feldspar ages are indeed congruent at one uranium and thorium decay series (Faure, 1986), as well as pro- sigma uncertainty for samples JB1-OSL2- JB1-OSL5 (JB1-OSL1 cesses such as alpha particle recoil and gaseous diffusion (Olley yielded no accepted feldspar aliquots). This suggests that the quartz et al., 1996). When a parent or daughter is lost or gained by any measured here, which saturates at an unusually high level, still process other than radioactive decay (i.e. the system is ‘open’), seems to yield accurate population Des even when up to a third of decay and production rates will no longer be balanced. Any un- aliquots are saturated. For samples JB2-OSL5 and JB2-OSL9, though, supported excess will decrease via decay until secular equilibrium where no direct comparison can be made to feldspar Des, quartz Des is regained, or conversely, a depleted daughter will increase. The must be considered a minimum value. It is also possible that time elapsed between initial disequilibrium and return to equilib- feldspar-derived Des are underestimated where saturated aliquots rium will depend on the concentration introduced/removed, and make up a signiﬁcant proportion of those measured, i.e. for JB1- the decay rates of the species involved. OSL7 and particularly JB1-OSL13. In both cases, we suggest that For the calculation of luminescence ages, the problems caused pIRIR290 ages should be considered, strictly speaking, as minimum by disequilibrium are twofold. First, dose rate conversion factors ages. assume a concentration to energy conversion based on the pres- ence of the entire decay chain for uranium and thorium, as well as 4.2.3.2. Dose rate estimation. Dose rates may lead to inaccurate an average ‘natural’ proportion between parent isotopes (e.g. 235U luminescence ages if the average dose rate during burial is mis- and 238U). When either parent or daughters are missing, this value calculated, or De distributions caused by variable dose rates are becomes inaccurate. Second, disequilibrium produces a time misattributed to another cause (e.g. partial bleaching). Given the dependent dose rate due to the ingrowth or unsupported decay of nature of the sediments at Jubbah and the De measurements ob- the daughter elements. The magnitude of the effect on the resulting tained, it seems likely that beta microdosimetry may cause un- luminescence ages depends on a number of factors: sample age, usually high overdispersion in some samples. Beta microdosimetry extent of disequilibrium in the decay chain, the proportion of the refers to variable beta doses to mineral grains within the same total dose contributed by the chain in disequilibrium, and the sample, and it arises from the short (several mm) range of beta method of dose rate calculation. Common sources of disequilibrium radiation. Minerals known to cause dose rate ‘hotspots,’ such as K- relevant for luminescence include: rich feldspars and U-rich zircons have been reported as sparse but 66 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 234 U/238U disequilibrium: 234U is a daughter of 238U, which is immediate consequences of this for the Jubbah basin system may more likely to escape from a sedimentary crystal lattice and be twofold, including both disequilibrium and open system effects form soluble uranyl complexes due to damage caused by alpha (diagenetic alteration). particles during 238U decay (Faure, 1986). 234U may be deﬁcient Open system behavior in carbonates is currently only directly in sediments that have been leached, and is usually in excess in detectable via disequilibrium measurements (i.e. ﬁnding disequi- water or waterlain sediments (Krbetschek et al., 1994). librium in sediments of a depositional age ca. 5x greater than the U/230Th disequilibrium: This is produced by the solubility of half-life of the daughter in question) or petrographic studies, which uranium in oxidizing conditions due to production of com- can detect the precipitation of secondary carbonates in pore spaces. pounds from the uranyl ion (UO2þ2), and relative insolubility of Unfortunately, neither analysis could be applied to these samples. thorium. However, we can examine closely the coherence of ages obtained Migration of radon. on carbonate-rich levels as they compare stratigraphically to bracketing sand-rich levels. And we may suggest that levels char- In an environment where disequilibrium is possible, samples acterized by particularly high uranium contents and which have are more likely to yield incorrect ages if uranium levels are high and been exposed to post-depositional water circulation are more likely not related to the detrital mineral component. to have been ‘overprinted’ by secondary precipitation/absorption. Post-diagenetic alteration of sedimentary characteristics such The potential effects of disequilibrium can also be gauged by as changing pore space availability or radioisotope concentrations calculating the Th/U ratio for samples; a typical crustal silicate Th/U may also cause inaccurate luminescence ages by altering the dose ratio is ca. 3.9 (O'Nions and McKenzie, 1993). rate to mineral grains. Standard calculation methods assume that Signiﬁcant variations in radioisotope concentrations are a dose rate calculated from modern sediment measurements apparent in Table 5 and Figs. 8 and 9, and uranium concentrations provides a good estimate for the average dose rate to the grains and thorium/uranium ratios for all measured samples have been during burial, or, in special cases, that corrections (such as plotted versus carbonate content in Fig. S8. Elevated uranium attenuation by moisture in pore spaces) can be assessed and levels (>15 ppm) occur in all sites but JB3, with JB1 samples applied. Waterlain, carbonate-rich sediments such as occur yielding the highest measured values at 2.5, 3.41, 4.16, and 6.35 m commonly in the Jubbah basin are particularly prone to temporal (JB1-OSL1, JB1-OSL3, JB1-OSL4, and JB1-OSL7). Three of these variation, via carbonate inﬁlling of pore spaces and radioisotopic samples (JB1-OSL1, JB1-OSL3, JB1-OSL4) also appear in the lowest deposition or leaching from a sedimentary unit. Post-depositional four Th/U ratios calculated, with samples JB2-OSL2, JB2-OSL5, and carbonate precipitation in pore spaces and consequent decreases JB2-OSL3 ﬁlling out the top six spots. It seems clear that in sediment water content will result in modern dose rate calcu- carbonate-rich layers may be prone to uranium enrichment, lations underestimating the average dose rate, as long as the though this is by no means uniform. The upper three samples carbonate inﬁll is inert (Nathan and Mauz, 2008). Conversely, collected from JB3 had some of the greatest recorded carbonate carbonate deposits are particularly prone to open system chemical contents, but also yielded uranium contents <2.5 ppm. This is behavior, as secondary carbonate inﬁlls in porous units may more in line with typical uranium concentrations measured in the coprecipitate with uranyl compounds (Faure, 1986), leading to Nefud, which tend to be 3 ppm (Rosenberg et al., 2013; Jennings overestimates of dose rate, or soluble uranium may be leached, et al., 2016; Stimpson et al., 2016). leading to dose rate underestimation. If signiﬁcant post- We suggest that there is a link between high modern-day depositional diagenetic alteration has occurred in the sediment, uranium contents, low Th/U ratios, and luminescence age under- leading to uptake or leaching of radioisotopes, then modern estimation due to dose rate overestimation, leading to the age measurements will not provide a good estimator of the average reversals and disagreement with independent chronological con- burial dose rate. This process may also lead to disequilibrium, but trol. It is quite difﬁcult to put constraints on the age offsets caused it is more difﬁcult to correct for, because there may be no con- by these processes. Given the overlying radiocarbon date as the straints on the timing of the alteration. Radioisotopic activity ra- minimum possible age for sample JB1-OSL2, we can calculate that tios of Quaternary lacustrine sediments cored from an endorheic the central dose rate has been overestimated by ca. 30%. Some of basin in Syria suggest that uranium leaching and re-precipitation this overestimation may be due to an average water content value in buried sediments is strongly linked to climatic conditions via higher than the 5% estimated here, but a signiﬁcant increase in humidity levels, and can result in high concentrations of uranium water content to 25e30% (water/wet sediment) would be required in subsurface levels as uranium is preferentially leached from near to correct the inverted age-depth relationship. It seems more surface deposits and reprecipitated in deeper levels (Ghaleb et al., likely that a signiﬁcant proportion of this overestimation is due to 1990; Li et al., 2008). Indeed, it has been suggested that important calculating alpha and beta dose rates using standard conversion sources of easily minable uranium may have been produced by the factors, which assume secular equilibrium. For sample JB1-OSL2, Pleistocene ﬂuvial systems of the Arabian Peninsula, which would 83% of the calculated combined alpha and beta dose are due to the have transported and concentrated uranium deposits in near- uranium concentration. The modern gamma dose rate assessment surface uranium-rich calcretes (Dill, 2011). should be less affected by current disequilibrium levels, as it has The Jubbah basin may be signiﬁcantly affected by high radio- been measured via a ﬁeld gamma spectrometer (Gue rin and isotope levels in groundwater that has chemically interacted with Mercier, 2011), but some error will be introduced due to tempo- either the Saq aquifer or the Ha'il granites to the south. Shabana and ral variance. If we assume that this age underestimation is entirely Kinsara (2014) have measured uranium and radium activities in due to disequilibrium, given that this sample is one of the youn- groundwater of the Ha'il region which are higher than the limits gest and has one of the highest dose contributions from uranium, imposed by national regulations. These authors link high uranium it seems plausible to suggest that a 30% dose rate overestimation is levels to leaching of granites and high radium concentrations to likely to be an upper limit for the age underestimation we might sandstone aquifers. The presence of 228-Radium, which has a half- expect via disequilibrium. If, however, elevated uranium levels life of only 5.75 years, indicates a strong and ongoing chemical (especially > 15 ppm) are due to post-diagenetic alteration, true exchange between Saq sandstone and regional groundwater. The ages become very difﬁcult to constrain. We do not believe that we L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 67 Table 5 Values measured for external dose rate estimation and calculated total dose rates via 'standard model.' Dose rates for feldspars include an internal component of 0.63 ± 0.10 Gy ka 1 or 0.89 ± 0.15 Gy ka 1 for grain sizes 125e180 mm and 180e255 mm, respectively. Samples ARY-OSL2,3,7 are also included as their dose rates have been recalculated via DRAC in order to calculate feldspar residual doses. Field Code K (%) Th (ppm) U (ppm) Water Content Gamma, Burial Depth Total Dose Rate (wet) Dry (Gy/ka) (m) Measured Burial Avg. Quartz Feldspar (% of wet) 125-180 mm 180-255 mm 125-180 mm 180-255 mm ARY-OSL2 0.32 2.85 2.32 0.36 5±3 0.53 0.87 1.26 ± 0.05b 2.03 0.12 2.23 ± 0.16 ARY-OSL3 1.08 4.9 4.6 0.79 5±3 0.72 1.07 2.26 ± 0.08b 3.18 0.16 3.32 ± 0.18 ARY-OSL4 0.29 0.46 4.09 0.49 5±3 –a 0.45 1.44 ± 0.05b 2.27 0.13 ARY-OSL5 0.38 1.97 16.04 0.42 5±3 0.96 0.56 3.32 ± 0.15b 4.71 0.32 ARY-OSL7 0.70 5.03 9.1 0.65 5±3 0.93 1.10 2.74 ± 0.11b 3.86 0.21 3.92 ± 0.21 JB1-OSL1 0.04 0.40 41.20 4.35 5±3 0.59 2.50 6.89 ± 0.45 JB1-OSL2 0.24 6.40 11.60 2.13 5±3 1.18 2.96 3.35 ± 0.15 4.21 ± 0.23 JB1-OSL3 0.25 1.00 45.40 10.88 10 ± 3 2.69 3.41 9.03 ± 0.48 9.82 ± 0.61 JB1-OSL4 0.12 0.60 39.80 1.75 5±3 3.23 4.16 9.2 ± 0.47 9.99 ± 0.60 JB1-OSL5 0.79 3.20 10.00 9.43 10 ± 3 2.11 4.51 4.01 ± 0.16 4.86 ± 0.23 JB1-OSL7 0.43 15.30 35.90 1.53 5±3 2.29 6.35 9.15 ± 0.54 JB1-OSL8 0.29 1.60 2.10 0.42 5±3 0.71 5.50 2.23 ± 0.16 JB1-OSL13 2.39 11.40 3.50 14.32 10 ± 3 1.12 9.00 4.30 ± 0.20 JB2-OSL1 0.03 0.50 2.40 1.80 5±3 0.09 0.77 0.69 ± 0.03 JB2-OSL2 0.05 <0.1 2.80 1.08 5±3 0.08 1.57 0.71 ± 0.04 JB2-OSL3 0.02 0.10 1.80 0.66 5±3 0.07 3.25 0.51 ± 0.03 JB2-OSL4 0.17 3.50 3.60 0.84 5±3 0.30 3.94 1.14 ± 0.05 JB2-OSL5 0.08 0.20 4.80 1.10 5±3 0.44 4.15 1.32 ± 0.06 JB2-OSL9 0.48 2.30 7.30 7.72 10 ± 3 1.19 5.95 2.56 ± 0.10 JB2-OSL14 0.80 3.50 1.30 14.32 10 ± 3 0.66 8.65 2.35 ± 0.16 JB3-OSL1 0.28 1.70 2.40 0.77 5±3 0.31 1.20 1.10 ± 0.04 JB3-OSL2 0.12 0.60 1.80 0.74 5±3 0.29 1.67 0.83 ± 0.03 JB3-OSL3 0.07 0.30 2.00 0.75 5±3 0.32 2.07 0.83 ± 0.03 JB3-OSL4 0.42 3.00 2.50 0.47 5±3 0.36 2.50 1.26 ± 0.05 2.13 ± 0.16 a On-site gamma dose rate not available. b Samples etched with hydroﬂuoric acid according to method in Hilbert et al. (2014), therefore no alpha dose rate included. have sufﬁcient evidence for which samples may have been subject Carb developed by Mauz and Hoffmann (2014). In order to provide to this process, or good constraints on the timing of the uranium some guidance, however, an attempt at subtraction dating deposition (e.g. early uptake or linear uptake) to successfully (Feathers, 2002) will be discussed in the next section for the propose ages based on model calculations such as those provided samples from JB1 which have both quartz and feldspar for open systems by Zander et al. (2007) or with a model such as measurements. Table 6 Ages measured for quartz and feldspar fractions, via standard and isochron methods. Both uncorrected and fading corrected (g-value ¼ 1.84 ± 1.32%/decade) feldspar pIRIR290 ages are included, but uncorrected ages are preferred (see text for discussion). Field Code Depths (m) Standard Ages (ka) Isochron Agesb Average D_ c (ka) (Gy ka 1) Quartz Feldspar Uncorrected Corrected ARY-OSL4 0.45 6.4 ± 0.4 ARY-OSL5 0.56 5.4 ± 0.8a (11.2 ± 2.0) JB1-OSL1 2.50 3.6 ± 0.3 JB1-OSL2 2.96 6.4 ± 0.4 22.1 ± 1.9 26.6 ± 4.5 81.9 ± 15.4 0.26 ± 0.05 JB1-OSL3 3.41 23.6 ± 2.2 30.5 ± 4.8 36.8 ± 7.7 111.6 ± 57.4 1.91 ± 0.99 JB1-OSL4 4.16 31.4 ± 4.8 37.6 ± 5.8 45.5 ± 9.8 116.5 ± 80.3 2.48 ± 1.74 JB1-OSL5 4.51 65.4 ± 7.6 73.4 ± 6.8 89.3 ± 16.1 117.1 ± 51.2 2.24 ± 1.01 JB1-OSL7 6.35 76.3 ± 16.9 92.8 ± 26.5 JB1-OSL8 5.50 135.8 ± 23.9 166.1 ± 38.9 JB1-OSL13 9.00 206.6 ± 49.7 253.6 ± 74.8 JB2-OSL1 0.77 8.6 ± 0.6 JB2-OSL2 1.57 12.0 ± 1.1 JB2-OSL3 3.25 10.0 ± 0.7 JB2-OSL4 3.94 8.6 ± 0.8 JB2-OSL5 4.15 195.2 ± 30.7 JB2-OSL9 5.95 92.0 ± 9.7 JB2-OSL14 8.65 359.4 ± 84.3 443.5 ± 135.3 JB3-OSL1 1.20 56.2 ± 8.3 JB3-OSL2 1.67 66.3 ± 8.0a (82.5 ± 11.6) JB3-OSL3 2.07 100.5 ± 20.5 JB3-OSL4 2.50 75.3 ± 8.1 105.3 ± 14.8 128.5 ± 27.3 a Minimum age model preferred, CAM age given in parentheses. b Calculated based on uncorrected feldspar ages. c Quartz grains, 125e180 mm. 68 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 4.3. Site chronologies regime. Instead, we propose that the older age found for the sand- rich sample JB1-OSL8 is more reliable than that calculated for the The issues raised above are discussed in detail for each of the younger, carbonate-rich, and uranium enriched sample JB1-OSL7. four palaeoenvironmental sections, and revised chronologies are The luminescence age calculated for JB1-OSL7 is likely to signiﬁ- presented. cantly underestimate the true burial age, and this is probably related to the OSL age underestimation higher in the sequence. As 4.3.1. ARY discussed above, comparison with independent chronological Two quartz ages have been added to those previously published control in this section also suggests that some OSL quartz ages may by Hilbert et al. (2014) (Fig. 8A, Table 6). Sample ARY-OSL4, be underestimating the true ‘burial age’. The single radiocarbon collected from the upper marls, dates the formation of the fresh- date obtained, 8980-8609 calBP (2.65e2.68 m), is signiﬁcantly water body to 6.4 ± 0.4 ka. These deposits, therefore, formed older than bracketing quartz OSL ages (3.6 ± 0.3 ka above and soon after deposition of the top of Unit 8, which was dated to 6.4 ± 0.4 ka below). It seems probable that the carbonate-rich levels 6.6 ± 0.7 ka (Hilbert et al., 2014). The age of the lower carbonate- in this section are signiﬁcantly affected by post-diagenetic uranium rich level (units 4e6) has already been closely bracketed to be- precipitation as well as disequilibrium. tween 11.4 ± 0.8 ka and 12.2 ± 1.1 ka by dating of sand-rich units In order to suggest an upper age limit for these carbonate-rich above and below (Hilbert et al., 2014). Direct dating of sample ARY- units, an equivalent dose subtraction method was used for sam- OSL5 from this deposit, however, has led to some unexpected re- ples JB1-OSL2, JB1-OSL3, JB1-OSL4, and JB1-OSL5. We are using a sults. In Section 4.2.1, it was noted that quartz from this sample simple calculation which leverages the intrinsic difference in dose yielded one of the highest overdispersion values measured in the rates between quartz (no internal dose rate), and feldspar (constant Jubbah basin (57.32 ± 12.26%), as well as a positively skewed De internal dose rate) (see Feathers, 2002), though a number of other distribution. Such characteristics suggest that the sample may be techniques have also been developed based on the same idea partially bleached. If a typical Holocene Jubbah quartz over- (Vogel et al., 1999; Zhao and Li, 2002; Li et al., 2008). Essentially, the dispersion value (20%) is assumed in the minimum age model, the quartz in each sample is used as a direct dosimeter (De,Q), which De of ARY-OSL5 becomes 17.94 ± 2.50 Gy and the ﬁnal age 5.4 ± 0.8 measures the total absorbed dose due to the external, environ- ka. This is signiﬁcantly younger than expected, as is the age mental dose rate. This value can be subtracted from the feldspar, calculated by using a less rigorous 31% overdispersion (see leaving only the effect of the internal dose rate. Age calculation then Figure S7), 9.2 ± 1.2 ka. The CAM age (11.2 ± 2.0 ka) is strati- follows naturally: graphically consistent. As discussed in Sections 4.2.3.1.1 and 4.2.3.2, we believe that this sample does contain partially bleached quartz Age ¼ De;F De;Q =D_ F;int grains, making the minimum age model a more appropriate choice, and that the stratigraphic reversal in the ages is instead due to an This technique may be powerful when the modern dose rate is overestimation of the dose rate. not expected to accurately reﬂect the average burial dose rate, i.e. in situations with extreme disequilibrium and/or secondary remobi- 4.3.2. JB1 lization of radioisotopes in the sediment. In this region, however, Five quartz and seven feldspar ages are available for this strat- the age results may be considered to be an upper age limit due to igraphic section (Fig. 9A, Table 6). Feldspar and quartz ages increase the potential for up to ca. 50 Gy residual doses caused by partial with depth as expected for most of the measured samples, from a bleaching of the pIRIR290 signal. top quartz age of 3.6 ± 0.3 ka (JB1-OSL1, 2.5 m) to a basal feldspar Central subtraction ages obtained for samples JB1-OSL2, JB1- age of >206.6 ± 49.7 ka (gravelly silt/sands, 9 m). This basal feldspar OSL3, JB1-OSL4, and JB1-OSL5 are in stratigraphic order, and age, calculated via the central age model, must strictly be consid- place these samples within MIS 5. The ages are unfortunately ered a minimum age due to the number of saturated aliquots imprecise, however, with propagated errors of up to 70%, and a one measured. The De distribution yields a relatively high over- sigma range extending from MIS 3 to 6. This is due primarily to a dispersion value (ca. 48%) and saturated aliquots. If we apply the combination of the De overdispersion and the error limits on the minimum age model (overdispersion value ¼ 40%) to this data, we internal dose rates calculated for the range of feldspar sizes used. calculate a De of 653.70 ± 151.67 Gy and an age of 151.9 ± 36.0 ka. Additionally, JB1-OSL2 clearly overestimates the depositional age, We suspect, however, that this minimum age is underestimating based on comparison with the radiocarbon date. If we derive the true age of the sediments, and suggest that the age of the de- average external dose rates to our measured quartz grains from the posits is more likely to be > 206.6 ± 49.7 ka. ages for samples JB1-OSL3eOSL5, we have signiﬁcant error. Even Further up in the sequence, an age-depth reversal occurs be- given this, it is apparent that these new dose rates are far more in tween 5.5 and 9 m, with the feldspar age of sample JB1-OSL7 being line with expected values based on the measurement of similar signiﬁcantly younger (76.3 ± 16.9 ka) than the age of the overlying sediments in the Nefud (Rosenberg et al., 2013; Jennings et al., sample JB1-OSL8 (135.8 ± 23.9 ka). The age reversal between these 2016; Stimpson et al., 2016). For samples JB1-OSL3 and JB1-OSL4, samples cannot be easily explained by the presence of partially the central subtraction-calculated dose rate is only ca. 20e27% of bleached feldspars in the upper sample. While both samples have the dose rate values derived from modern measurements. A much relatively high overdispersion values, applying the minimum age smaller change is calculated for sample JB1-OSL5. model with an overdispersion value of 40% (see discussion in sec- For this site, therefore, the balance of evidence suggests the tion 4.2.3.1) yields Des of 495.37 ± 97.57 Gy and 250.29 ± 40.08 Gy, following chronology. Underlying gravelly silt/sands (Unit 2) with a for samples JB1-OSL7 and JB1-OSL8, respectively. These values yield minimum age at least 151.9 ± 36.0 ka and probably 206.6 ± 49.7 ages of 54.1 ± 11.1 ka for the deeper sample and 112.4 ± 19.7 ka for ka, were deposited initially. Units 7e8 probably represent the initial the overlying sample, which does not resolve the age-depth stages of MIS 5e, with the top of unit 9 dated to 135.8 ± 23.9 ka. Unit discrepancy. If instead assuming that signiﬁcant partial bleaching 12 is most likely to have been deposited between 73.4 ± 6.8 ka has occurred in the upper but not lower sample, a feldspar residual (feldspar age) and 117.1 ± 51.2 ka (subtraction age), therefore it is of several hundred gray would be required. This is far higher than attributable to MIS 5. Given the subtraction ages obtained for any values calculated for the analogous site Al-Rabyah (Hilbert samples JB1-OSL3 and JB1-OSL4, we suggest that potentially units et al., 2014), and difﬁcult to explain given the depositional 10e14 all date to MIS 5a-c, though we cannot exclude the L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 69 possibility of MIS 3 deposition as at the nearby site ALM 3 (Jennings 4.3.4. JB3 et al., 2016). Units 16 and above must be Holocene, based on the Three samples (JB3-OSL1dJB3-OSL3) have been dated with magnitude of the quartz Des measured, and units 17 and above quartz OSL, while both quartz and feldspar ages have been obtained were deposited after 8980-8609 calBP. for underlying sample JB3-OSL4 (Fig. 8B, Table 6). Two anomalies are present in the calculated CAM ages: quartz ages for samples JB3-OSL3 and JB3-OSL4 are stratigraphically reversed, and there is a 4.3.3. JB2 signiﬁcant discrepancy between the quartz and feldspar ages for Six quartz ages and one feldspar age are available (Fig. 9B, sample JB3-OSL4. Table 6). The two quartz ages obtained for the upper 2 m, 8.6 ± 0.6 Some consideration has already been given to the potential for ka and 12.0 ± 1.1 ka, are in stratigraphic order, however, the next partial bleaching in this setting (Section 4.2.1). As noted previously, four quartz ages (2e6 m) do not yield a consistent age-depth all three samples collected from JB3 carbonates have unusually relationship. Based on the equivalent doses, it is apparent that high quartz De overdispersion values in the context of the Jubbah the upper 4 m of the section represent Holocene accumulation, basin samples (Fig. 7). Assuming that partial bleaching can be with a signiﬁcant increase in equivalent dose at approximately 4 m identiﬁed by a combination of positive skewness and high over- indicating a substantial gap in deposition. Units below 4 m are dispersion, it was suggested that the luminescence age for sample likely Pleistocene in age. Finally, the lowermost sample, measured JB3-OSL2 should be calculated from the MAM De. By analogy with with feldspar, yielded an age of 359.4 ± 84.3 ka. The De distribution the measurements from the underlying sand, and given the high for this sample is unusual, with one outlier aliquot (ca. 260 Gy) and overdispersions of quartz from the bracketing samples, we applied a cluster of ﬁve aliquots at approximately 1000 Gy. Interpreting an overdispersion value of 30%. A MAM De of 55.00 ± 6.32 Gy was such data is problematic, as it is quite difﬁcult to propose a calculated, with approximately 80% of aliquots fully bleached. mechanism by which the majority of the sediment would be Samples JB3-OSL1 and JB3-OSL3, with high overdispersion but partially bleached to such a degree. Physical sediment mixing via symmetric distributions, are likely to be better suited by CAM Des, bioturbation or another process may be possible, particularly if as microdosimetric variations in the sampled sediment seem to be younger sediments are building up over a much older basal sedi- a plausible cause for such distributions. These calcretes are far more ment. In this case, we prefer the central age model results, but note well-consolidated and have a lower porosity than any others this difﬁculty for future study. sampled in the Jubbah basin, which may indicate a differing Other anomalies in the age-depth relationship for this section geochemical and dosimetric environment. It seems plausible that indicate that multiple sample ages are problematic. Quartz De dis- some grains from these samples will have been completely over- tributions, however, appear suitable. All of the presumably Holo- grown by inert carbonates, perhaps unusually in the Jubbah basin, cene aliquot populations have low overdispersion values and therefore receive a negligible beta dose rate (Nathan et al., (maximum 21.21 ± 6.39% for sample JB2-OSL2) and acceptable 2003). Additionally, sample JB3-OSL1 was collected adjacent to a values were calculated for the sampled Pleistocene units well-developed shell bed and contained numerous shell fragments. (maximum 35.73 ± 11.83%). Moreover, there is no correlation be- Shell fragments may either comprise inert carbonate, and lower tween the magnitude of the overdispersion and the age reversals, dose rates to grains, or absorb uranium from the sediment and which would be expected if this were caused by partial bleaching. increase dose rates (Kaufman et al., 1996). In combination with the That is, if the reversal in the Holocene sediments is due to partial known presence of heavy minerals, such as zircons, signiﬁcant bleaching, the ‘overestimated’ upper ages should have higher microdosimetric variation seems likely, though a petrographic overdispersions. In the Pleistocene sediments, there are a number study would be necessary to prove this. Depending on the size of of saturated aliquots (up to 5 of 18 for sample JB2-OSL5), which the shell fragments, they may also cause systematic errors in suggests that the calculated quartz OSL age might underestimate calculated beta dose rates (Cunningham, 2016). Again, extensive the true burial age. However, this does not explain the dramatic age investigation of these sediments would be necessary to further reversal (more than 100,000 years) between samples JB2-OSL5 and reﬁne the calculated dose rates and test the effect of these various JB2-OSL9. confounding factors. Considering possible issues with dose rate determination, it has The single feldspar age obtained for this site yields an age of already been noted (section 4.2.3.2) that samples JB2-OSL2, JB2- 105.3 ± 14.8 ka for unit 1, an overestimate of ca. 30,000 years in OSL3, and JB2-OSL5 yield three of the six lowest Th/U ratios from comparison to the quartz age for the same sample. If this difference dated samples in this study (Fig. S8). Disequilibrium may therefore is attributed to partial bleaching, we can calculate a residual dose of cause these ages to be underestimated. Interestingly, however, approximately 55 Gy. This is certainly plausible, given the range signiﬁcant uranium enrichment is only noted at ca. 6.5 m depth, calculated for samples from section ARY, therefore we prefer the though the concentration is still less than half of that measured for quartz OSL age for sample JB3-OSL4. shallow samples from JB1. It can also be noted that the equivalent Age results based on a MAM De for JB3-OSL2 and CAM Des for all doses measured for samples JB2-OSL5 and JB2-OSL9 are within one other samples are shown in Fig. 8B. Given an examination of all the sigma uncertainty of each other, while their ages are 100,000 years data, we understand that the age calculations for this site are not offset. This also suggests potential difﬁculties with dose rate entirely straightforward. Nevertheless, we believe that the overall determination. weight of the evidence at the moment suggests that relatively slow Given the lack of sand-rich units within this section that can be deposition of waterlain sediments seems to have occurred after used as reliable age markers, it is very difﬁcult to suggest a likely 75.3 ± 8.1 ka, with the youngest sample dating to 56.2 ± 8.3 ka. chronology. All waterlain sediments must have been deposited Given the relative congruence between sample Des, the low ura- after the basal feldspar age of 359.4 ± 84.3 ka (clayey sands). A clear nium contents, and the low porosity of the upper three samples, it depositional hiatus is noted at around 4 m depth, between samples seems unlikely that circulating groundwater has caused diagenetic JB2-OSL4 and JB2-OSL5. Based on Th/U ratios and low U contents, alteration. we suggest that samples JB2-OSL1 and JB2-OSL4 are likely to be the most reliable Holocene samples. These ages are indistinguishable 5. Conclusions within one sigma error, suggesting the Holocene sediments were deposited around or slightly before 8.6 ka. The Jubbah basin preserves complex, heterogeneous 70 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 sedimentary records. Our work highlights that records from this post-depositional alteration of carbonate-rich sediments, and to basin can present some signiﬁcant problems for luminescence two anonymous reviewers for their helpful comments. dating, including: Appendix A. Supplementary data Signiﬁcant temporal gaps (as at JB2) that are not obvious in the ﬁeld. Supplementary data related to this article can be found at http:// Depositional ages for sediments potentially ranging from the dx.doi.org/10.1016/j.quageo.2017.06.002. Holocene through MIS 11 or even older. Mineral fractions containing signiﬁcant proportions of gypsum, References which may affect pIRIR290 signals without careful application of rejection criteria (Clark-Balzan, 2016). Aitken, M.J., 1998. An Introduction to Optical Dating: the Dating of Quaternary Sediments by the Use of Photon-stimulated Luminescence. Oxford University Carbonate-rich subsurface sediments subject to reprecipitation Press, Oxford. of carbonate leading to signiﬁcant post-depositional alteration Aitken, M.J., Xie, J., 1990. Moisture correction for annual gamma dose. Anc. TL 8, of radioisotope concentrations and disequilibrium (Ghaleb et al., 6e9. 1990; Li et al., 2008). Armitage, S.J., Jasim, S.A., Marks, A.E., Parker, A.G., Usik, V.I., Uerpmann, H.-P., 2011. The southern route “out of Africa”: evidence for an early expansion of modern humans into Arabia. Science 331, 453e456. http://dx.doi.org/10.1126/ It is likely that such difﬁculties are widespread in comparable science.1199113. settings, in which subsurface sediments have been exposed to Atkinson, O.A.C., Thomas, D.S.G., Goudie, A.S., Bailey, R.M., 2011. Late Quaternary chronology of major dune ridge development in the northeast Rub’ al-Khali, multiple cycles of groundwater percolation. It should be noted that United Arab Emirates. Quat. Res. 76, 93e105. http://dx.doi.org/10.1016/ it is difﬁcult a priori to suggest which units will have been signiﬁ- j.yqres.2011.04.003. cantly affected (e.g. signiﬁcant uranium enrichment throughout JB1 Auclair, M., Lamothe, M., Huot, S., 2003. Measurement of anomalous fading for feldspar IRSL using SAR. Radiat. Meas. 37, 487e492. http://dx.doi.org/10.1016/ and rare at JB2). S1350-4487(03)00018-0. Climatically relevant chronological frameworks can still be Balescu, S., Lamothe, M., 1994. Comparison of TL and IRSL age estimates of feldspar developed, however, via the approach used in this paper, i.e. coarse grains from waterlain sediments. Quat. Geochronol. 13, 437e444. Berger, G.W., 1990. Effectiveness of natural zeroing of the thermoluminescence in extensive sampling through the proﬁle, luminescence dating of sediments. J. Geophys. Res. 95, 12375e12397. http://dx.doi.org/10.1029/ both quartz and feldspar, and careful consideration of correlations JB095iB08p12375. between elemental concentrations, sedimentary characteristics, Blechschmidt, I., Matter, A., Preusser, F., Rieke-Zapp, D., 2009. Monsoon triggered formation of Quaternary alluvial megafans in the interior of Oman. Geo- Des, and dose rates. Even when ages cannot necessarily be esti- morphology 110, 128e139. http://dx.doi.org/10.1016/j.geomorph.2009.04.002. mated (e.g. sample JB2-OSL5 in JB2), equivalent dose versus depth Bøtter-Jensen, L., Andersen, C.E., Duller, G.A.T., Murray, A.S., 2003. Developments in plots may provide important palaeoenvironmental evidence. radiation, stimulation and observation facilities in luminescence measure- Studied sections preserve a remarkable record of recurrent hu- ments. Radiat. Meas. 37, 535e541. http://dx.doi.org/10.1016/S1350-4487(03) 00020-9. midity during the Holocene, MIS 3, MIS 5, and likely older basal Bøtter-Jensen, L., Bulur, E., Duller, G.A.T., Murray, A.S., 2000. Advances in lumines- sediments dating between MIS 7 and MIS 9 or MIS 11. Records of cence instrument systems. Radiat. Meas. 32, 523e528. http://dx.doi.org/ Holocene humidity from the interior of the Nefud are known only 10.1016/S1350-4487(00)00039-1. Breeze, P.S., Groucutt, H.S., Drake, N.A., White, T.S., Jennings, R.P., Petraglia, M.D., from Jubbah deposits, while MIS 3 deposits indicative of surface 2016. Palaeohydrological corridors for hominin dispersals in the Middle East humidity have only been reported from the nearby site ALM 3 on ~250e70,000 years ago. Quat. Sci. Rev. 144, 155e185. http://dx.doi.org/10.1016/ the southern edge of the Nefud (Jennings et al., 2016). The partic- j.quascirev.2016.05.012. Brennan, B.J., 2003. Beta doses to spherical grains. Radiat. Meas. 37, 299e303. ular environmental conditions supported by this humidity cannot http://dx.doi.org/10.1016/S1350-4487(03)00011-8. be inferred without signiﬁcant further analysis, which will be dis- Brennan, B.J., Lyons, R.G., Phillips, S.W., 1991. Attenuation of alpha particle track cussed in a subsequent publication. However, it seems likely that dose for spherical grains. Int. J. Radiat. Appl. Instrum. Part D. Nucl. Tracks Radiat. Meas. 18, 249e253. http://dx.doi.org/10.1016/1359-0189(91)90119-3. the groundwater in the Jubbah basin would have constituted a Bronk Ramsey, C., 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51, valuable resource, and the recurrent nature of this groundwater 337e360. indicated by the dates in this study may have led to the concurrent Bronk Ramsey, C., Higham, T., Leach, P., 2004. Towards high-precision AMS: prog- ress and limitations. Radiocarbon 46, 17e24. presence of archaeological sites of various periods in this generally Bronk Ramsey, C., Higham, T.F.G., Owen, D.C., Pike, A.W.G., Hedges, R.E.M., 2002. arid area. Radiocarbon dates from the Oxford AMS system: archaeometry datelist 31. Archaeometry 44, 1e149. Acknowledgements Bronk Ramsey, C., Staff, R.A., Bryant, C.L., Brock, F., Kitagawa, H., van der Plicht, J., Schlolaut, G., Marshall, M.H., Brauer, A., Lamb, H.F., Payne, R.L., Tarasov, P.E., Haraguchi, T., Gotanda, K., Yonenobu, H., Yokoyama, Y., Tada, R., Nakagawa, T., We thank His Royal Highness Prince Sultan bin Salman, Presi- 2012. A complete terrestrial radiocarbon record for 11.2 to 52.8 kyr B.P. Science dent of the Saudi Commission for Tourism and National Heritage 338 (6105), 370e374. Buylaert, J.-P., Thiel, C., Murray, A.S., Vandenberghe, D.A.G., Yi, S., Lu, H., 2011. IRSL (SCTH), and Prof. Ali Ghabban, Vice President for Antiquities and and post-IR IRSL residual doses recorded in modern dust samples from the Museums, for permission to carry out this research. We also thank Chinese Loess Plateau. Geochronometria 38, 432e440. http://dx.doi.org/ our Saudi colleagues from the SCTH, especially Jamal Omar, Sultan 10.2478/s13386-011-0047-0. Buylaert, J.-P., Jain, M., Murray, A.S., Thomsen, K.J., Thiel, C., Sohbati, R., 2012. Al-Fagir, and Abdulaziz al-Omari for their support and assistance A robust feldspar luminescence dating method for Middle and Late Pleistocene with the ﬁeld investigations. Dr. Abdullah Alsharekh, King Saud sediments. Boreas 41, 435e451. http://dx.doi.org/10.1111/j.1502- University, has also been a key supporter of our research in Saudi 3885.2012.00248.x. Buylaert, J.-P., Murray, A.S., Gebhardt, A.C., Sohbati, R., Ohlendorf, C., Thiel, C., Arabia. Financial support for the ﬁeldwork and project was pro- Wastegård, S., Zolitschka, B., The PASADO Science Team, 2013. Luminescence vided by the European Research Council (ERC) (grant number dating of the PASADO core 5022-1D from Laguna Potrok Aike (Argentina) using 295719, to MDP) and the SCTH. The senior author was supported by IRSL signals from feldspar. Quat. Sci. Rev. 71, 70e80. http://dx.doi.org/10.1016/ j.quascirev.2013.03.018. funding from the European Union's Framework Programme for , B., Zaragosi, S., Rossignol, L., Bourget, J., Eynaud, F., Martinez, P., Caley, T., Malaize Research and Innovation Horizon 2020 (2014e2020) under the Giraudeau, J., Charlier, K., Ellouz-Zimmermann, N., 2011. New Arabian Sea re- Marie Skłodowska-Curie Grant Agreement No. 658005 (Individual cords help decipher orbital timing of Indo-Asian monsoon. Earth Planet. Sci. Fellowship). Thanks also go to Mr. Christopher Doherty of the Lett. 308, 433e444. http://dx.doi.org/10.1016/j.epsl.2011.06.019. Clark-Balzan, L., 2016. Source and characteristics of blue, infrared (IR), and post-IR Research Laboratory for Archaeology and the History of Art (Uni- IR stimulated signals from gypsum-rich samples. Anc. TL 34, 6e13. versity of Oxford), for discussions concerning the deposition and Crassard, R., Petraglia, M.D., Parker, A.G., Parton, A., Roberts, R.G., Jacobs, Z., L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 71 Alsharekh, A., Al-Omari, A., Breeze, P., Drake, N.A., Groucutt, H.S., Jennings, R., 2015a. Late Pleistocene lakeshore settlement in northern Arabia: Middle Regagnon, E., Shipton, C., 2013. Beyond the levant: ﬁrst evidence of a pre- palaeolithic technology from jebel katefeh. Jubbah. Quat. Int. 382, 215e236. pottery neolithic incursion into the Nefud desert, Saudi Arabia. PLoS One 8, http://dx.doi.org/10.1016/j.quaint.2014.12.001. e68061. http://dx.doi.org/10.1371/journal.pone.0068061. Groucutt, H.S., White, T.S., Clark-Balzan, L., Parton, A., Crassard, R., Shipton, C., Cunningham, A.C., 2016. External beta dose rates to mineral grains in shell-rich Jennings, R.P., Parker, A.G., Breeze, P.S., Scerri, E.M.L., Alsharekh, A., sediment. Anc. TL 34, 1e5. Petraglia, M.D., 2015b. Human occupation of the arabian empty quarter during Cunningham, A.C., DeVries, D.J., Schaart, D.R., 2012. Experimental and computa- MIS 5: evidence from Mundafan Al-Buhayrah, Saudi Arabia. Quat. Sci. Rev. 119, tional simulation of beta-dose heterogeneity in sediment. Radiat. Meas. 47, 116e135. http://dx.doi.org/10.1016/j.quascirev.2015.04.020. 1060e1067. http://dx.doi.org/10.1016/j.radmeas.2012.08.009. Guagnin, M., Shipton, C., Martin, L., Petraglia, M., 2017. The Neolithic site of Jebel Dean, W.E., 1974. Determination of carbonate and organic matter in calcareous Oraf 2, northern Saudi Arabia: ﬁrst report of a directly dated site with faunal sediments and sedimentary rocks by loss on ignition; comparison with other remains. Archaeol. Res. Asia 9, 63e67. http://dx.doi.org/10.1016/ methods. J. Sediment. Res. 44, 242e248. j.ara.2017.02.001. Debaene, G., 2003. Uranium-series dating of marly sediments: application to Gue rin, G., Mercier, N., 2011. Determining gamma dose rates by ﬁeld gamma Jaroszow fossil lake (SW Poland). Geochronometria 22, 15e26. spectroscopy in sedimentary media: results of Monte Carlo simulations. Radiat. Des Combes, H.J., Caulet, J.P., Tribovillard, N., 2005. Monitoring the variations of the Meas. 46, 190e195. http://dx.doi.org/10.1016/j.radmeas.2010.10.003. Socotra upwelling system during the last 250 kyr: a biogenic and geochemical Gue rin, G., Mercier, N., Adamiec, G., 2011. Dose-rate conversion factors: update. Anc. approach. Palaeogeogr. Palaeoclimatol. Palaeoecol. 223, 243e259. http:// TL 29, 5e8. dx.doi.org/10.1016/j.palaeo.2005.04.007. Gue rin, G., Mercier, N., Nathan, R., Adamiec, G., Lefrais, Y., 2012. On the use of the Dietze, M., Kreutzer, S., Burow, C., Fuchs, M.C., Fischer, M., Schmidt, C., 2016. The inﬁnite matrix assumption and associated concepts: a critical review. Radiat. abanico plot: visualising chronometric data with individual standard errors. Meas. 47, 778e785. http://dx.doi.org/10.1016/j.radmeas.2012.04.004. Quat. Geochronol. 31, 12e18. http://dx.doi.org/10.1016/j.quageo.2015.09.003. Gue rin, G., Jain, M., Thomsen, K.J., Murray, A.S., Mercier, N., 2015. Modelling dose Dill, H.G., 2011. A comparative study of uranium-thorium accumulation at the rate to single grains of quartz in well-sorted sand samples: the dispersion western edge of the Arabian Peninsula and mineral deposits worldwide. Arab. J. arising from the presence of potassium feldspars and implications for single Geosci. 4, 123e146. http://dx.doi.org/10.1007/s12517-009-0107-4. grain OSL dating. Quat. Geochronol. 27, 52e65. http://dx.doi.org/10.1016/ Duller, G.A.T., 2003. Distinguishing quartz and feldspar in single grain luminescence j.quageo.2014.12.006. measurements. Radiat. Meas. 37, 161e165. http://dx.doi.org/10.1016/S1350- Hanson, P.R., Mason, J.A., Jacobs, P.M., Young, A.R., 2015. Evidence for bioturbation of 4487(02)00170-1. luminescence signals in eolian sand on upland ridgetops, southeastern Min- Duller, G.A.T., 2012. Improving the accuracy and precision of equivalent doses nesota, USA. Quat. Int. 362, 108e115. http://dx.doi.org/10.1016/ determined using the optically stimulated luminescence signal from single j.quaint.2014.06.039. grains of quartz. Radiat. Meas. 47, 770e777. http://dx.doi.org/10.1016/ Heiri, O., Lotter, A.F., Lemcke, G., 2001. Loss on ignition as a method for estimating j.radmeas.2012.01.006. organic and carbonate content in sediments: reproducibility and comparability Durcan, J.A., King, G.E., 2017. DRAC: Dose Rate and Age Calculator. of results. J. Paleolimnol. 25, 101e110. http://dx.doi.org/10.1023/A: Durcan, J.A., King, G.E., Duller, G.A.T., 2015. DRAC: dose rate and age calculator for 1008119611481. trapped charge dating. Quat. Geochronol. 28, 54e61. http://dx.doi.org/10.1016/ Hilbert, Y.H., White, T.S., Parton, A., Clark-Balzan, L., Crassard, R., Groucutt, H.S., j.quageo.2015.03.012. Jennings, R.P., Breeze, P., Parker, A., Shipton, C., Al-Omari, A., Alsharekh, A.M., Duval, M., Arnold, L.J., 2013. Field gamma dose-rate assessment in natural sedi- Petraglia, M.D., 2014. Epipalaeolithic occupation and palaeoenvironments of the mentary contexts using LaBr3(Ce) and NaI(Tl) probes: a comparison between southern Nefud desert, Saudi Arabia, during the terminal Pleistocene and early the “threshold” and “windows” techniques. Appl. Radiat. Isot. 74, 36e45. http:// Holocene. J. Archaeol. Sci. 50, 460e474. http://dx.doi.org/10.1016/ dx.doi.org/10.1016/j.apradiso.2012.12.006. j.jas.2014.07.023. Edgell, H.S., 2006. Arabian Deserts: Nature, Origin, Evolution. Springer-Verlag, Hoelzmann, P., Jolly, D., Harrison, S.P., Laarif, F., Bonneﬁlle, R., Pachur, H.-J., 1998. Heidelberg, Germany. Mid-Holocene land-surface conditions in northern Africa and the Arabian Farrant, A.R., Duller, G.A.T., Parker, A.G., Roberts, H.M., Parton, A., Knox, R.W.O., peninsula: a data set for the analysis of biogeophysical feedbacks in the climate Bide, T., 2015. Developing a framework of Quaternary dune accumulation in the system. Glob. Biogeochem. Cycles 12, 35e51. northern Rub' al-Khali, Arabia. Quat. Int. 382, 132e144. http://dx.doi.org/ Hoskin, P.W.O., Schaltegger, U., 2003. The Composition of zircon and igneous and 10.1016/j.quaint.2015.02.022. metamorphic petrogenesis. Rev. Mineral. Geochem 53, 27e62. http:// Faure, G., 1986. Principles of Isotope Geology. John Wiley & Sons, Inc, New York. dx.doi.org/10.2113/0530027. Feathers, J.K., 2002. Luminescence dating in less than ideal conditions: case studies Huntley, D.J., Baril, M.R., 1997. The K content of the K-feldspars being measured in from klasies river main site and duinefontein, South Africa. J. Archaeol. Sci. 29, optical dating or in thermoluminescence dating. Anc. TL 15, 11e13. 177e194. http://dx.doi.org/10.1006/jasc.2001.0685. Huntley, D.J., Hancock, R.G.V., 2001. The Rb contents of the K-feldspar grains being Fleitmann, D., Burns, S.J., Mudelsee, M., Neff, U., Kramers, J., Mangini, A., Matter, A., measured in optical dating. Anc. TL 19, 43e46. 2003. Holocene forcing of the Indian monsoon recorded in a stalagmite from Huntley, D.J., Lamothe, M., 2001. Ubiquity of anomalous fading in K-feldspars and southern Oman. Science 300, 1737e1739. http://dx.doi.org/10.1126/ the measurement and correction for it in optical dating. Can. J. Earth Sci. 38, science.1083130. 1093e1106. http://dx.doi.org/10.1139/e01-013. Fleitmann, D., Burns, S.J., Pekala, M., Mangini, A., Al-Subbary, A., Al-Aowah, M., Huntley, D.J., Godfrey-Smith, D.I., Thewalt, M.L.W., 1985. Optical dating of sedi- Kramers, J., Matter, A., 2011. Holocene and Pleistocene pluvial periods in Yemen, ments. Nature 313, 105e107. http://dx.doi.org/10.1038/313105a0. southern Arabia. Quat. Sci. Rev. 30, 783e787. http://dx.doi.org/10.1016/ Hussein, M.T., Bazuhair, A.G., Ageeb, A.E., 1992. Hydrogeology of the Saq formation j.quascirev.2011.01.004. east of hail, northern Saudi Arabia. Q. J. Eng. Geol. Hydrogeol. 25, 57e64. http:// Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., Olley, J.M., 1999. Optical dx.doi.org/10.1144/GSL.QJEG.1992.025.01.05. dating of single and multiple grains of quartz from Jinmium rock shelter, Jain, M., Ankjærgaard, C., 2011. Towards a non-fading signal in feldspar: insight into northern Australia: Part I, Experimental design and statistical models. charge transport and tunnelling from time-resolved optically stimulated Archaeometry 41, 339e364. luminescence. Radiat. Meas. 46, 292e309. http://dx.doi.org/10.1016/ Garrard, A.N., Harvey, C.P.D., Switsur, V.R., 1981. Environment and settlement during j.radmeas.2010.12.004. the upper Pleistocene and Holocene at Jubba in the great Nefud, northern Jain, M., Buylaert, J.P., Thomsen, K.J., Murray, A.S., 2015. Further investigations on Arabia. Atlal 5, 137e148. “non-fading” in K-feldspar. Quat. Int. 362, 3e7. http://dx.doi.org/10.1016/ Garzanti, E., Vermeesch, P., Ando , S., Vezzoli, G., Valagussa, M., Allen, K., Kadi, K.A., j.quaint.2014.11.018. Al-Juboury, A.I.A., 2013. Provenance and recycling of Arabian desert sand. Earth- Jennings, R.P., Parton, A., Clark-Balzan, L., White, T.S., Groucutt, H.S., Breeze, P.S., Sci. Rev. 120, 1e19. http://dx.doi.org/10.1016/j.earscirev.2013.01.005. Parker, A.G., Drake, N.A., Petraglia, M.D., 2016. Human occupation of the Ghaleb, B., Hillaire-Marcel, C., Causse, C., Gariepy, C., Vallieres, S., 1990. Fractionation northern arabian interior during early marine isotope stage, 3 (31), 953e966. and recycling of U and Th isotopes in a semi-arid endoreic depression of central http://dx.doi.org/10.1002/jqs.2920. Syria. Geochim. Cosmochim. Acta 54, 1025e1035. http://dx.doi.org/10.1016/ Kars, R.H., Busschers, F.S., Wallinga, J., 2012. Validating post IR-IRSL dating on K- 0016-7037(90)90436-O. feldspars through comparison with quartz OSL ages. Quat. Geochronol. 12, Ginau, A., Engel, M., Brückner, H., 2012. Holocene chemical precipitates in the 74e86. http://dx.doi.org/10.1016/j.quageo.2012.05.001. continental sabkha of Tayma (NW Saudi Arabia). J. Arid. Environ. 84, 26e37. Kars, R.H., Reimann, T., Ankjærgaard, C., Wallinga, J., 2014. Bleaching of the post-IR http://dx.doi.org/10.1016/j.jaridenv.2012.03.020. IRSL signal: new insights for feldspar luminescence dating. Boreas 43, 780e791. Gliganic, L.A., May, J.H., Cohen, T.J., 2015. All mixed up: using single-grain equivalent http://dx.doi.org/10.1111/bor.12082. dose distributions to identify phases of pedogenic mixing on a dryland alluvial Kaufman, A., Ghaleb, B., Wehmiller, J.F., Hillaire-Marcel, C., 1996. Uranium con- fan. Quat. Int. 362, 23e33. http://dx.doi.org/10.1016/j.quaint.2014.07.040. centration and isotope ratio proﬁles within Mercenaria shells: geochronological Godfrey-Smith, D.I., Huntley, D.J., Chen, W.H., 1988. Optical dating studies of quartz implications. Geochim. Cosmochim. Acta 60, 3735e3746. http://dx.doi.org/ and feldspar sediment extracts. Quat. Sci. Rev. 7, 373e380. 10.1016/0016-7037(96)00190-1. Grimm, E.C., 2011. High-resolution age model based on AMS radiocarbon ages for Kellogg, K.S., Stoeser, D.B., 1985. Reconnaissance Geology of the Ha’il Quadrangle, Kettle Lake, North Dakota, USA. Radiocarbon 53, 39e53. Sheet 27/41B, Kingdom of Saudi Arabia. USGS Open-File report 85e618. Groucutt, H.S., Petraglia, M.D., 2012. The prehistory of the Arabian peninsula: de- Krbetschek, M.R., Rieser, U., Zo € ller, L., Heinicke, J., 1994. Radioactive disequilibria in serts, dispersals, and demography. Evol. Anthropol. 21, 113e125. http:// palaeodosimetric dating of sediments. Radiat. Meas. 23, 485e489. dx.doi.org/10.1002/evan.21308. Li, B., Li, S.-H., Wintle, A.G., 2008. Overcoming environmental dose rate changes in Groucutt, H.S., Shipton, C., Alsharekh, A., Jennings, R., Scerri, E.M.L., Petraglia, M.D., luminescence dating of waterlain deposits. Geochronometria 30, 33e40. http:// 72 L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 dx.doi.org/10.2478/v10003-008-0003-z. K-rich feldspars for optical dating of young coastal sedimentsda test case from Li, B., Jacobs, Z., Roberts, R.G., Li, S.-H., 2014. Review and assessment of the potential Darss-Zingst peninsula (southern Baltic Sea coast). Quat. Geochronol. 6, of post-IR IRSL dating methods to circumvent the problem of anomalous fading 207e222. http://dx.doi.org/10.1016/j.quageo.2010.10.001. in feldspar luminescence. Geochronometria 41, 178e201. http://dx.doi.org/ Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey, C., 10.2478/s13386-013-0160-3. Buck, C.E., Cheng, H., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Lowick, S.E., Trauerstein, M., Preusser, F., 2012. Testing the application of post IR- Haﬂidason, H., Hajdas, I., Hatte , C., Heaton, T.J., Hoffmann, D.L., Hogg, A.G., IRSL dating to ﬁne grain waterlain sediments. Quat. Geochronol. 8, 33e40. Hughen, K.A., Kaiser, K.F., Kromer, B., Manning, S.W., Niu, M., Reimer, R.W., http://dx.doi.org/10.1016/j.quageo.2011.12.003. Richards, D.A., Scott, E.M., Southon, J.R., Staff, R.A., Turney, C.S.M., van der Marty, J., Myrbo, A., 2014. Radiocarbon dating suitability of aquatic plant macro- Plicht, J., 2013. IntCal13 and Marine13 radiocarbon age calibration curves fossils. J. Paleolimnol. 52, 435e443. http://dx.doi.org/10.1007/s10933-014- 0e50,000 Years cal BP. Radiocarbon 55, 1869e1887. http://dx.doi.org/10.2458/ 9796-0. azu_js_rc.55.16947. Mauz, B., Hoffmann, D., 2014. What to do when carbonate replaced water: Carb, the Rhodes, E.J., Schwenninger, J.L., 2007. Dose rates and radioisotope concentrations in model for estimating the dose rate of carbonate-rich samples. Anc. TL 32, the concrete calibration blocks at Oxford. Anc. TL 25, 5e8. 24e32. Richter, D., Richter, A., Dornich, K., 2013. Lexsyg d a new system for luminescence Mayya, Y.S., Morthekai, P., Murari, M.K., Singhvi, A.K., 2006. Towards quantifying research. Geochronometria 40, 220e228. http://dx.doi.org/10.2478/s13386- beta microdosimetric effects in single-grain quartz dose distribution. Radiat. 013-0110-0. Meas. 41, 1032e1039. http://dx.doi.org/10.1016/j.radmeas.2006.08.004. Rittenour, T.M., 2008. Luminescence dating of ﬂuvial deposits: applications to Mercier, N., Falgue res, C., 2007. Field gamma dose-rate measurement with a NaI(Tl) geomorphic, palaeoseismic and archaeological research. Boreas 37, 613e635. detector: Re-evaluation of the “threshold” technique. Anc. TL 25, 1e4. http://dx.doi.org/10.1111/j.1502-3885.2008.00056.x. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved Rosenberg, T.M., Preusser, F., Fleitmann, D., Schwalb, A., Penkman, K., Schmid, T.W., single-aliquot regenerative-dose protocol. Radiat. Meas. 32, 57e73. http:// Al-Shanti, M.A., Kadi, K., Matter, A., 2011a. Humid periods in southern Arabia: dx.doi.org/10.1016/S1350-4487(99)00253-X. windows of opportunity for modern human dispersal. Geology 39, 1115e1118. Murray, A.S., Wintle, A.G., 2003. The single aliquot regenerative dose protocol: http://dx.doi.org/10.1130/G32281.1. potential for improvements in reliability. Radiat. Meas. 37, 377e381. http:// Rosenberg, T.M., Preusser, F., Wintle, A.G., 2011b. A comparison of single and dx.doi.org/10.1016/S1350-4487(03)00053-2. multiple aliquot TT-OSL data sets for sand-sized quartz from the Arabian Murray, A.S., Thomsen, K.J., Masuda, N., Buylaert, J.P., Jain, M., 2012. Identifying Peninsula. Radiat. Meas. 46, 573e579. http://dx.doi.org/10.1016/j.radmeas.2011. well-bleached quartz using the different bleaching rates of quartz and feldspar 03.020. luminescence signals. Radiat. Meas. 47, 688e695. http://dx.doi.org/10.1016/ Rosenberg, T.M., Preusser, F., Risberg, J., Plikk, A., Kadi, K.A., Matter, A., Fleitmann, D., j.radmeas.2012.05.006. 2013. Middle and Late Pleistocene humid periods recorded in palaeolake de- Nathan, R.P., Mauz, B., 2008. On the dose-rate estimate of carbonate-rich sediments posits of the Nafud desert, Saudi Arabia. Quat. Sci. Rev. 70, 109e123. http:// for trapped charge dating. Radiat. Meas. 43, 14e25. http://dx.doi.org/10.1016/ dx.doi.org/10.1016/j.quascirev.2013.03.017. j.radmeas.2007.12.012. Roskosch, J., Tsukamoto, S., Meinsen, J., Frechen, M., Winsemann, J., 2012. Lumi- Nathan, R.P., Thomas, P.J., Jain, M., Murray, A.S., Rhodes, E.J., 2003. Environmental nescence dating of an Upper Pleistocene alluvial fan and aeolian sandsheet dose rate heterogeneity of beta radiation and its implications for luminescence complex: the Senne in the Münsterland Embayment, NW Germany. Quat. dating: monte Carlo modelling and experimental validation. Radiat. Meas. 37, Geochronol. 10, 94e101. http://dx.doi.org/10.1016/j.quageo.2012.02.012. 305e313. http://dx.doi.org/10.1016/S1350-4487(03)00008-8. Schramm, A., Stein, M., Goldstein, S.L., 2000. Calibration of the 14C time scale to >40 Olley, J.M., Murray, A., Roberts, R.G., 1996. The effects of disequilibria in the uranium ka by 234Ue230Th dating of Lake Lisan sediments (last glacial Dead Sea). Earth and thorium decay chains on burial dose rates in ﬂuvial sediments. Quat. Sci. Planet. Sci. Lett. 175, 27e40. http://dx.doi.org/10.1016/S0012-821X(99)00279-4. Rev. 15, 751e760. http://dx.doi.org/10.1016/0277-3791(96)00026-1. Shabana, E.I., Kinsara, A.A., 2014. Radioactivity in the groundwater of a high back- Olley, J., Caitcheon, G., Murray, A., 1998. The distribution of apparent dose as ground radiation area. J. Environ. Radioact. 137, 181e189. http://dx.doi.org/ determined by optically stimulated luminescence in small aliquots of ﬂuvial 10.1016/j.jenvrad.2014.07.013. quartz: implications for dating young sediments. Quat. Geochronol. 17, Shipton, C., Parton, A., Breeze, P., Jennings, R., Groucutt, H.S., White, T.S., Drake, N., 1033e1040. Crassard, R., Alsharekh, A., Petraglia, M.D., 2014. Large ﬂake acheulean in the Olley, J.M., Pietsch, T., Roberts, R.G., 2004. Optical dating of Holocene sediments Nefud desert of northern Arabia. PaleoAnthropology 446e462. http:// from a variety of geomorphic settings using single grains of quartz. Geo- dx.doi.org/10.4207/PA.2014.ART85. morphology 60, 337e358. http://dx.doi.org/10.1016/j.geomorph.2003.09.020. Spooner, N.A., 1994. On the optical dating signal from quartz. Radiat. Meas. 23, O'Nions, R.K., McKenzie, D., 1993. Estimates of mantle thorium/uranium ratios from 593e600. http://dx.doi.org/10.1016/1350-4487(94)90105-8. Th, U and Pb isotope abundances in basaltic melts. Philos. Trans. R. Soc. Lond. A Stimpson, C.M., Lister, A., Parton, A., Clark-Balzan, L., Breeze, P.S., Drake, N.A., 342, 65e77. Groucutt, H.S., Jennings, R., Scerri, E.M.L., White, T.S., Zahir, M., Duval, M., Oswald, W.W., Anderson, P.M., Brown, T.A., Brubaker, L.B., Hu, F.S., Lozhkin, A.V., Grün, R., Al-Omari, A., Al Murayyi, K.S.M., Zalmout, I.S., Mufarreh, Y.A., Tinner, W., Kaltenrieder, P., 2005. Effects of sample mass and macrofossil type Memesh, A.M., Petraglia, M.D., 2016. Middle Pleistocene vertebrate fossils from on radiocarbon dating of arctic and boreal lake sediments. Holocene 15, the Nefud desert, Saudi Arabia: implications for biogeography and palae- 758e767. oecology. Quat. Sci. Rev. 143, 13e36. http://dx.doi.org/10.1016/j.quascirev.2016. Parton, A., Farrant, A.R., Leng, M.J., Schwenninger, J.-L., Rose, J.I., Uerpmann, H.-P., 05.016. Parker, A.G., 2013. An early MIS 3 pluvial phase in Southeast Arabia: climatic Stokes, S., 1992. Optical dating of young (modern) sediments using quartz: results and archaeological implications. Quat. Int. 300, 62e74. http://dx.doi.org/ from a selection of depositional environments. Quat. Sci. Rev. 11, 153e159. 10.1016/j.quaint.2013.02.016. http://dx.doi.org/10.1016/0277-3791(92)90057-F. Parton, A., Farrant, A.R., Leng, M.J., Telfer, M.W., Groucutt, H.S., Petraglia, M.D., Stokes, S., Bray, H.E., Blum, M.D., 2001. Optical resetting in large drainage basins: Parker, A.G., 2015a. Alluvial fan records from southeast Arabia reveal multiple tests of zeroing assumptions using single-aliquot procedures. Quat. Sci. Rev. 20, windows for human dispersal. Geology 43, 295e298. http://dx.doi.org/10.1130/ 879e885. http://dx.doi.org/10.1016/S0277-3791(00)00045-7. G36401.1. Thatcher, L., Rubin, M., Brown, G.F., 1961. Dating desert ground water. Science 134 Parton, A., White, T.S., Parker, A.G., Breeze, P.S., Jennings, R., Groucutt, H.S., (3472), 105e106. Petraglia, M.D., 2015b. Orbital-scale climate variability in Arabia as a potential Thiel, C., Buylaert, J.P., Murray, A., Terhorst, B., Hofer, I., Tsukamoto, S., Frechen, M., motor for human dispersals. Quat. Int. 382, 82e97. http://dx.doi.org/10.1016/ 2011a. Luminescence dating of the Stratzing loess proﬁle (Austria)dtesting the j.quaint.2015.01.005. potential of an elevated temperature post-IR IRSL protocol. Quat. Int. 234, Petraglia, M.D., Alsharekh, A., Crassard, R., Drake, N.A., Groucutt, H., Parker, A.G., 23e31. http://dx.doi.org/10.1016/j.quaint.2010.05.018. Roberts, R.G., 2011. Middle paleolithic occupation on a marine isotope stage 5 Thiel, C., Buylaert, J.-P., Murray, A.S., Tsukamoto, S., 2011b. On the applicability of lakeshore in the Nefud desert, Saudi Arabia. Quat. Sci. Rev. 30, 1555e1559. post-Ir Irsl dating to Japanese loess. Geochronometria 38, 369e378. http:// http://dx.doi.org/10.1016/j.quascirev.2011.04.006. dx.doi.org/10.2478/s13386-011-0043-4. Petraglia, M.D., Alsharekh, A., Breeze, P., Clarkson, C., Crassard, R., Drake, N.A., Thiel, C., Buylaert, J.P., Murray, A.S., Elmejdoub, N., Jedoui, Y., 2012. A comparison of Groucutt, H.S., Jennings, R., Parker, A.G., Parton, A., Roberts, R.G., Shipton, C., TT-OSL and post-IR IRSL dating of coastal deposits on Cap Bon peninsula, north- Matheson, C., Al-Omari, A., Veall, M.-A., 2012. Hominin dispersal into the Nefud eastern Tunisia. Quat. Geochronol. 10, 209e217. http://dx.doi.org/10.1016/ desert and Middle palaeolithic settlement along the Jubbah palaeolake, j.quageo.2012.03.010. northern Arabia. PLoS One 7, e49840. Thomsen, K.J., Murray, A.S., Jain, M., 2011. Stability of IRSL signals from sedimentary Petraglia, M.D., Groucutt, H., Parton, A., Alsharekh, A., 2015. Green Arabia: human K-feldspar samples. Geochronometria 38, 1e13. http://dx.doi.org/10.2478/ prehistory at the crossroads of continents [special issue]. Quat. Int. 382, 1e302. s13386-011-0003-z. Pigati, J.S., Quade, J., Wilson, J., Jull, A.J.T., Lifton, N.A., 2007. Development of low- Vogel, J.C., Wintle, A.G., Woodborne, S.M., 1999. FOCUS: luminescence dating of background vacuum extraction and graphitization systems for 14C dating of coastal sands: overcoming changes in environmental dose rate. J. Archaeol. Sci. old (40-60 ka) samples. Quat. Int. 166, 4e14. http://dx.doi.org/10.1016/ 26, 729e733. http://dx.doi.org/10.1006/jasc.1999.0450. j.quaint.2006.12.006. Watts, D., Al-Naﬁe, A.H., 2013. Vegetation and Biogeography of the Sand Seas of Preusser, F., Muru, M., Rosentau, A., 2014. Comparing different post-IR IRSL ap- Saudi Arabia. Routledge, London. proaches for the dating of Holocene coastal foredunes from Ruhnu Island, Whitney, J.W., Faulkender, D., Rubin, M., 1983. The Environmental History and Estonia. Geochronometria 41, 342e351. http://dx.doi.org/10.2478/s13386-013- Present Condition of Saudi Arabia's Northern Sand Seas. USGS Open-File report 0169-7. 83e749. Reimann, T., Tsukamoto, S., Naumann, M., Frechen, M., 2011. The potential of using Wintle, A.G., 1973. Anomalous fading of thermo-luminescence in mineral samples. L. Clark-Balzan et al. / Quaternary Geochronology 45 (2018) 50e73 73 Nature 245, 143e144. http://dx.doi.org/10.1038/245143a0. Zhao, H., Li, S.H., 2002. Luminescence isochron dating: a new approach using Zaidi, F.K., Kassem, O.M.K., Al-Bassam, A.M., Al-Humidan, S., 2015. Factors governing different grain sizes. Radiat. Prot. Dosim. 101, 333e338. groundwater chemistry in Paleozoic sedimentary aquifers in an arid environ- Ziegler, M., Lourens, L.J., Tuenter, E., Reichart, G.-J., 2010. High Arabian Sea pro- ment: a case study from Hail Province in Saudi Arabia. Arab. J. Sci. Eng. 40, ductivity conditions during MIS 13dodd monsoon event or intensiﬁed over- 1977e1985. http://dx.doi.org/10.1007/s13369-014-1534-4. turning circulation at the end of the Mid-Pleistocene transition? Clim. Past. 6, Zander, A., Degering, D., Preusser, F., Kasper, H.U., Brückner, H., 2007. Optically 63e76. http://dx.doi.org/10.5194/cp-6-63-2010. stimulated luminescence dating of sublittoral and intertidal sediments from Zimmerman, D.W., 1971. Thermoluminescent dating using ﬁne grains from pottery. Dubai, UAE: radioactive disequilibria in the uranium decay series. Quat. Geo- Archaeometry 13, 29e52. chronol. 2, 123e128. http://dx.doi.org/10.1016/j.quageo.2006.04.003.

(PDF) Resolving problematic luminescence chronologies for carbonate-and evaporite-rich sediments spanning multiple humid periods