Dating fossil root cast (Black Sea coast, Turkey) using
thermoluminescence: Implications for windblown drift of shelf
carbonates during MIS 2
George S. Polymeris
a,b,*, George Kitis
c, Na
fiye G. Kiyak
d, Eleni Theodosoglou
e,
Nestor C. Tsirliganis
a, Ahmet Ertek
f, Ahmet E. Erginal
gaLaboratory of Archaeometry and Physicochemical Measurements, R.C.‘Athena’ PO BOX 159, GR-67100, Xanthi, Greece bInstitute of Nuclear Sciences, Ankara University, TR-06100, Bes¸evler-Ankara, Turkey
cAristotle University of Thessaloniki, Nuclear Physics Laboratory, GR-54124, Thessaloniki, Greece dIs¸ık University, Faculty of Science and Arts, Physics Department, TR-34980, S¸ile-Istanbul, Turkey
eAristotle University of Thessaloniki, School of Geology, Department of MineralogyePetrologyeEconomic Geology, GR-54124, Thessaloniki, Greece fIstanbul University, Faculty of Letters, Department of Geography, TR-34459, Laleli, Istanbul, Turkey
gArdahan University, Department of Geography, TR-75000, Ardahan, Turkey
a r t i c l e i n f o
Article history:
Available online 13 June 2015 Keywords:
Rhizolith TL dating Black Sea
a b s t r a c t
Rhizoliths are mostly sub-aerially exposed root prints which appear through removal of the rock units that cap them. A horizontal-lying residual of a rhizolith, consisting purely of soft inner core material of white color was recovered 10 km west of S¸ile, Istanbul, in northwest Turkey within laminated oolithic massive aeolianite. The inner part, consisting purely of calcium carbonate, was dated by applying ther-moluminescence, while for the outer shelves optically stimulated luminescence of quartz was used for age assessment. The age of the CaCO3infill occupying the original place of the decayed plant roots was
found to be 26.8 (±5.0) ka, corresponding to MIS 2. When compared with the ages of the middle (105.2± 15.6 ka) and outer (127 ± 9 ka) layers, corresponding to the later stage of MIS 5e or early stage of MIS 5d, the inner core coincides with the last glacial period when the sea-level was lower than the present, promoting transportation of ooids by offshore winds in conjunction with the exposed shelf carbonates. Based on the results yielded, rhizolith is much younger than the host rock aeolianite and witnesses to last glacial sea level lowstand when removal of shelf carbonates by offshore winds was promoted from the exposed shallow shelf plain. The results provide strong evidence that rhizoliths may not be coeval with the aeolianites within which they are embedded.
© 2015 Elsevier Ltd and INQUA. All rights reserved.
1. Introduction
Contributing to illumination of paleo-pedogenesis conditions in sedimentary rocks of various origins and ages, rhizoliths are foot-prints of decayed roots. These occurrences are preferably retained within Pleistocene carbonate dunes or aeolianites and are simply defined as organo-sedimentary formations produced by roots, providing evidence of higher plant colonization of sub-aerially exposed sediments and rocks (Klappa, 1980). The generic term
rhizolith, derived from the Greek rhizo¼ root þ lith ¼ stone, was adopted in 1980 by Klappa to include accumulation and/or cementation around, cementation within, or replacement of, higher plant roots by minerals.
Rhizoliths are mostly sub-aerially exposed root prints which appear through removal of the rock units that cap them. Their size is usually on a centimetric scale with the exception of those having a length of 1 m or more, recently defined as megarhizoliths (Alonso-Zarza et al., 2008). In theirfield appearance, these occur-rences have some constitutive properties, such as several down-ward bifurcations similar to those in living tree roots with decreasing diameters of second, third and higher order branches, distinguishing thus rhizoliths from animal burrows. In general, they can be categorized into five different morphological types: root
* Corresponding author. Institute of Nuclear Sciences, Ankara University, TR-06100, Bes¸evler-Ankara, Turkey.
E-mail address:gspolymeris@ankara.edu.tr(G.S. Polymeris).
Contents lists available atScienceDirect
Quaternary International
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m/ l o ca t e / q u a i n t
http://dx.doi.org/10.1016/j.quaint.2015.05.060
tubules, root casts, root molds, rhizocretions and root petrifactions (Klappa, 1980).
Rhizoliths retain a record of paleo-environmental conditions (Retallack, 2001). Therefore, their characteristics as facies, age and isotopic attributes could provide extremely valuable information. Despite the abundance of preserved records, fossil root marks and casts have received little attention from geologists, and their paleo-ecological potential has scarcely begun to be examined. In previous studies, carried out on both continental and coastal aeolianites all around the world, many authors were solicitous about the signi fi-cance of rhizolith formation as it accompanies paleo-pedodiagenesis as an organic component (Abegg et al., 2001;Kraus and Hasiotis, 2006; Frebourg et al., 2008). However, most of these citations go no further than merely listing their presence in thefield. For further evidence supporting the aforementioned statement, the authors could refer toGocke et al. (2010; 2011; 2014)and references therein. A limited number of studies were carried out concerning the morphological and petrological features and the mineralogical composition of the aeolianites and the presence of rhizoliths as paleo-environmental indicators. However, the majority of these are focused on the morphological characteristics of the rhizoliths. Among these, only a scarce number of studies are focused on the structure and composition of the rhizoliths and even fewer use X-ray diffraction techniques to determine their mineralogical composition (Gocke et al., 2010, 2011, 2014). Moreover, there is no literature data concerning direct luminescence dating of the cal-cium carbonate based rhizolith inner part, with only limited reports of rhizolith dating by14C (Gocke et al., 2011).
In this paper, the microfacies, compositional characteristics as well as the thermoluminescence (TL) age are presented for a rhi-zolith ascertained within coastal aeolianites 10 km west of S¸ile on the western Black Sea coast of Turkey. The formation of these aeolianites was recently dated by both optically stimulated lumi-nescence (OSL,Erginal et al., 2013) and TL (Polymeris et al., 2012). Besides some recent publications (Erginal et al., 2010, 2013; Kiyak and Erginal, 2010; Polymeris et al., 2012), the knowledge for the existence of both aeolianites and rhizoliths on either Turkey's long (8483 km) coastline or the Black Sea is greatly restricted. For instance, inBrooke's (2001)review, listing the global distribution of aeolianites pinpointing 89 different locations, no record was included from either the Turkish coastline or the Black Sea coast.
The present study is based mainly on direct observations ranging fromfield to microscopic. The morphological and miner-alogical features of the rhizolith, as well as its chemical composi-tion, are examined in this paper using optical microscopy, X-ray powder diffraction (XRPD) and energy dispersive X-ray spectros-copy/scanning electron microscopy (EDX/SEM) techniques, respectively. For the age of the formation, two independent lumi-nescence approaches were applied, depending on the composition. To the best of the authors' knowledge, this is thefirst study dealing with the simultaneous compositional and dating examination of a rhizolith sample around the Mediterranean Sea.
2. Regional settingse site description
The studied rhizolith sample was collected within highly indu-rated fossil terrestrial carbonate dunes (aeolianites) 10 km west of S¸ile (Fig. 1), Istanbul, in northwest Turkey (411105300N, 292604600E). The host rock has a maximum of 6 m thickness and backs onto a sandy beach containing abundant bioclastic material such as Venus gallina and Ostrea edulis. The aeolianite overlies disconformably Pliocene-aged clay-rich terrestrial deposits, consisting of clayey sands with lignite intercalations (Okay, 1948). From a climatological point of view, the area under study falls within a semi-humid re-gion based on data from S¸ile meteorological station (41100N
29360E). The area receives an annual precipitation of 749 mm. The average annual temperature is 13.6C and the predominant wind activity is from the northeast (Erginal et al., 2013).
3. Materials and methods 3.1. Micro-analytical examination
The samples of rhizolith were thin sectioned and subsequently examined under a CHEBIOS microscope. Energy dispersive X-ray spectroscopy (EDX) was used to determine the elemental compo-sition of the samples. The micro-morphological characteristics of the rhizolith were studied using a scanning electron microscope (Phillips XLe30 S FEG). The XRPD patterns of the different layers (inner core, middle, and outer layers) observed in the sample were acquired on a Phillips PW 1820/00 X-ray diffractometer, equipped with a PW 1710 microprocessor and using PC-APD software. Operating conditions for the samples were 35 kV and 25 mA using Ni-filtered CuKaaveradiation. The 2theta (2
q
) scanning range wasbetween 3 and 63and the scanning speed was 1.2/min. Identi-fication of the samples was made using the JCPDS-ICCD 2003 database.
3.2. Luminescence apparatus and protocols
All luminescence measurements were acquired using a RISØ TL/ OSL reader (model TL/OSL-DA-15) equipped with a 0.075 Gy/s
90Sr/90Y beta particle source (Bøtter-Jensen et al., 2000). The reader
wasfitted with a 9635QA photomultiplier tube. In the case of OSL, the stimulation wavelength was 470 (±20) nm, emitting at the sample position a maximum of 40 mW cm2, while the detection optics consisted of a 7.5 mm Hoya U-340 (
l
p ~ 340 nm, FWHM 80 nm)filter. However, all TL measurements were performed usinga combination of a Pilkington HA-3 heat absorbing filter and a Corning 7e59 (320e440 nm) blue filter. All heatings were per-formed in a nitrogen atmosphere with a constant heating rate of 1 C/s in order to avoid significant temperature lag and conse-quences from thermal quenching effect (Subedi et al., 2011; Kalita and Wary, 2015); all TL measurements were performed up to a maximum heat temperature of 500C; all blue OSL measurements were performed at 110C in the continuous wave mode (CW-OSL). For the age of the formation, two independent luminescence approaches were applied, depending on the composition. The multigrain, multi-aliquot, TL-based additive dose, total optical bleaching procedure suggested bySinghvi et al. (1982)was applied to the CaCO3of the rhizolith, while the Single Aliquot Regenerative
OSL (SAR-OSL,Murray and Wintle, 2000, 2003) was applied to the quartz extracted from the middle layer and outer, aeolianite surface.
4. Results
4.1. Sample descriptione composition and origin
Based on field observations, rhizoliths are mostly found as horizontal-lying root structures with maximum 80 cm length and 8 cm diameter. With few exceptions, the typical diameter is not more than 3e4 cm, comprising discrete fragments of curved shape.
Fig. 2shows a picture of the collected sample. In agreement with the literature, the examined rhizolith sample collected for the present study was almost 10 cm long and 4 cm in diameter, without any fragments of curved shape. As a typical horizontal rhizolith, it has a soft inner core material of white color, in contrast to those preserved vertically within aeolianite sands. The exterior of the rhizolith, typically consisting of aeolianite, is marked by a rough and yellowish surface; it will be termed the“outer surface” here-after. Between the core and the outer surface, a light grayish, thin layer-material ring could be easily identified; this will be termed as “middle part” hereafter.
Vertically, these occurrences are only implicit in the upper (~1 m below surface) level of the aeolianite, which, in that section, shows a thin lamination (Fig. 3a, b). Thin section analyses of the sample (Fig. 3c) verified the presence of three different rhizocreations, which implies cementation during the life or after death of the
plant roots (Klappa, 1980; Flugel, 2004) comprised roughly of inner, middle and outer layers. The inner core is entirely composed of 3 cm-thick, powdery,fine grained calcium carbonate (Figs. 2 and 3d), which can be easily engraved by afingernail, and is possibly the original place of the decayed roots. Based on EDX analyses, this part contains predominantly Ca (40.78%), O (39.25%), while to a lesser extent C (19.77%). This observation is in excellent agreement with the XRPD pattern of the inner core, where only the phase of calcite occurs asFig. 4reveals. This centripetal infill is believed to be formed as a result of the death and decay of plant roots and its subsequent calcification.
In addition, observation of the thin sections demonstrates the abundance of circular and ellipsoidal-shaped ooids (Fig. 3e), sug-gesting concurrent wind-blownfine sediment drift from a shallow marine environment onto the decayed roots during rhizolith for-mation. The ooids observed are composed of several concentrically-arranged laminae. Such materials with high intragranular porosity are, in terms of ooid microfacies, known as indicators of high-energy shallow marine environments (Aalto and Dill, 1996). In this study, it can be stated that the ooids, dispersed within abun-dant carbonate precipitates, might have been deposited within the voids, left by decayed roots, during a lowstand sea-level, a well-known model for carbonate supply to glacial aeolianites (Abegg et al., 2001; Brooke, 2001; Kraus and Hasiotis, 2006; Frebourg et al., 2008).
The middle layer of the studied rhizolith is a thin (<1 cm), well-cemented micritized ring, consisting of a mixture of calcite and quartz grains with almost equal dimensions; the amount of the latter increases towards the surrounding outer aeolianite sands. The detected elements within that layer were Ca (45.23%), O (35.93%), C (16.58%), Na (0.70%), Si (0.43%), K (0.28%), and Al (0.24%), suggesting the predominance of calcite and quartz verified by the XRPD analysis (Fig. 4) and probably possible traces of the feldspar group. The fact that this layer constitutes the hardest part of the rhizolith is likely associated with the microcrystallic size of the minerals and an excessive amount of calcite microcrystal and low porosity ratio.
The outer part of the rhizolith corresponds to carbonated-cemented aeolianite itself. This section is characterized by high porosity but strong induration (Fig. 3f). From petrographical point of view, it contains quartz and carbonate particles as well as abundant ooids (Fig. 3). EDX analysis of the cement material demonstrated the existence of Ca (39.51%), O (37%), and C (18.9%) as the predominant components. The other components are repre-sented by Si (1.35%), Na (0.93%), Fe (0.78%), Mg (0.52%), K (0.52%), and Al (0.51%). Of great significance here is the abundance of ooids, confirming the transportation from a low sea-level environment. Even though accumulation of ooids within aeolianite sands has been tied to high wave energy conditions (Aalto and Dill, 1996), this can also be attributed to deflation by offshore winds when a carbonate-rich shelf emerged during low sea-level stands. A tabular form of the outline of the materials geochemistry is presented in
Table 1.
4.2. Luminescence dating; the rationale
The examined coastal dunes are covered by salt-tolerant plant canopies. During a low stand of one of the glacial periods, the sea-level regressed and an extensive shelf, rich in carbonate, emerged. This carbonate of marine origin was shifted onto the dune sands by off-shore winds and then enwrapped the plant bodies. During short-term rainfalls, dune sands, abundant in quartz were cemen-ted. Plants died in the course of time, and their organic parts were replaced by CaCO3, while voids left by the decayed plants were also
filled by the carbonate. This is a short summary of how both the
Fig. 2. Rhizolith sample subjected to present study. Three distinctive layers are easily identified; an unconsolidated inner core of white color full of powdery carbonate, the tightly lithified outer layer consisting of yellowish aeolianite, and middle thin grayish ring as transition unit between them.
aeolianite and the so-called rhizolith were formed, based on the classification ofKlappa (1980).
The soft carbonate whichfills the inner core is derived from the exposed shelf. This is geologically evidenced by the presence of abundant ooids, which are formed as result of carbonate binding around nuclei, which in turns, is made of a mineral fragment as small fragment of rock or grain. The ooids were deposited concurrently with the carbonate infills into the rhizolith. CaCO3was
transported, together with the ooids, which are geologically typical of deposition in an agitated shallow marine environment. Their interesting presence within the rhizolith suggests transportation from the exposed shelf to the dunes and rhizoliths. There is no crystallization in the inner core; this is further supported by the soft form of calcium carbonate. The event involves dating layers of the carbonate shelf, which relies on the sensitive electron traps responsible for luminescence, having been previously bleached by sunlight, either by another layer or by being incorporated into material of the same layer. Thus the target event, namely the date of accumulation, is contemporaneous with the optically-dated event, i.e. the time since the material's grains were last exposed to light.
The exposure time of each mineral's grains to sunlight depends on the time taken for the sedimentation/formation procedure. Sunlight exposure may also happen before transportation and sedimentation. Nevertheless, these carbonate materials, as well as the quartz grains, are expected to have been exposed to the sun for a prolonged time, thus enabling adequate bleaching of the material during the formation of the rhizolith. Aeolianites are multiple-stacked, indurate dune forms comprising skeletal carbonate sand with interbedded paleosols. Consequently, cementation guaran-teesfirm contact between the grains of the formation. The light attenuation in rocks is up to several microns (Laskaris and Liritzis, 2011). Therefore, the cemented layers exclude light penetration from the overlying layers, protecting the latter from further solar bleaching. From the moment that any grain is no longer exposed to sunlight and put infirm contact with other grains, the optically-sensitive electron traps are filled by electrons produced by the ionization caused by nuclear radiation of natural uranium, thorium, potassium, rubidium and cosmic radiation. These iso-topes are present in both the surrounding minerals as well as the mineral itself.
Regarding age determination, two different physical quantities are required; the total accumulated dose during the past, termed as paleodose or equivalent dose, as well as the rate at which this energy-dose accumulated, termed as dose rate. Thus, the lumi-nescence age estimate equals the paleodose (De or ‘equivalent
dose’, in units of Gy, derived from luminescence measurements) divided by the annual dose rate. The latter (usually in mGrays per year) denotes the radiation dose accumulated in a year. It is comprised of the three radiation dose components (alpha, beta and gamma radiation) derived from the natural radioisotopes of ura-nium (238U,235U), thorium (232Th), potassium (40K), and rubidium (87Rb), of the sample itself, the surrounding environment and the soil surrounding the sampling point, including cosmic-rays (Aitken, 1985; Liritzis, 1994; Liritzis et al., 2013a,b).
4.3. Luminescence characteristics of rhizolith (inner core)
According to both the XRD and SEM-EDX results, as well as
Table 1, the inner part of the rhizolith consists solely of calcium carbonate. The latter has been shown to yield a TL signal that grows with increasing radiation dose (McDougal, 1968; Engin and Güven, 2000; Stirling et al., 2014). The temperature-dependence of the luminescence emission from calcite appears as a set of three main peaks (Wintle, 1977; Debenham, 1983; Debenham and Aitken, 1984; Down et al., 1985; Liritzis, 1989; Kalita and Wary, 2014; Stirling et al., 2014). Nevertheless, in geological samples of calcite, differences in the behavior of these peaks have been observed. The reader may refer to McDougal (1968) for a review of early achievements on this related topic. Unfortunately, even though the TL signal of calcium carbonate is capable of providing relative dating information, a notable exception is the optically stimulated
luminescence (OSL) signal from CaCO3, as the latter material does
not yield a natural OSL signal (Liritzis, 1989, 2000, 2011; Galloway, 2002). From the dating point of view, the studies so far were limited to TL of mostly biogenic calcite, with both limited (Ninagawa, 1987; Ninagawa et al., 1988, 1992) as well as good success (Duller et al., 2009; Liritzis, 2010, 2011). While in the former case the problems arose from the organic nature of the samples, in the latter case the cornerstone of the successful applications was the event being dated, being the crystallization of the sample itself as well as the last exposure to light for the surface material of the curved stone, respectively.
Fig. 5A presents the natural TL (NTL) signal glow curve of the calcite-based inner layer. The NTL glow curve consists mainly of one peak at around 350e375 C. Nevertheless, the possibility of the presence of another TL peak corresponding to Very Deep Trap (VDT), similar to the case of ordinary chalk (Polymeris et al., 2013) could not be excluded, due to the presence of a rising part to the glow curve after background subtraction (curve III) for tempera-tures greater than 450C. Note the intense TL background signal (curve II).Fig. 5B shows characteristic glow curves of the additive-dose procedure for the sample. Each TL glow curve presented in
Fig. 5 stands as the mean value of four independently-acquired glow curves. A TL peak at low temperatures, around 95C, corre-sponding to a shallow trap, is obvious. The additive doses applied were 10, 20, and 30 Gys.
The TL signal is never totally zeroed due to bleaching. Instead, an un-bleachable residual component is always left. Whatever the level of the unknown un-bleachable residual TL, it serves as the ‘non-zero clock’ level upon which subsequent radiation builds up. Calcite, unlike quartz and feldspar, is not an easily bleached min-eral, and in most cases the un-bleachable residual is reached after a prolonged period of some dozens of hours (Liritzis and Bakopoulos, 1997; Habermann et al., 2000; Kim and Hong, 2014). Therefore, there is always the risk that the exposure to sunlight prior to deposition was not sufficient to reach this level. In order to check the effective zeroing of the luminescence under solar light and establish an adequate duration of solar bleaching, several fresh aliquots were kept aside in order to estimate the residual bleaching levels of TL. These aliquots were divided into groups consisting of four aliquots each. Samples from thefirst group were measured in order to obtain the NTL. The rest were exposed to sunlight bleaching in the summer sun in Xanthi\Northern Greece, over variable lengths of time, ranging from 30 min to 10 days, hereafter noted as Nblj.Fig. 5C shows a selection of glow curves after various
bleaching times for the calcium carbonate grains resulting from the inner layer. Bleaching does not seem to take place rapidly, espe-cially in thefirst two hours of solar exposure. This feature is further supported by curve (a) ofFig. 6, where the residual curve of solar bleaching indicates that exposure to sunlight for 3 days is adequate for the TL signal to reach its equilibrium level.
Each bleached TL curve was subtracted from the corresponding natural TL curve, after which a corresponding dose-temperature plateau was constructed. The method involves the estimation of, and allowance for, the residual level of TL, assuming that its level
Fig. 4. XRPD patterns of inner layer (a) of rhizolith, consisting of pure calcite crystals and middle layer (b) consisting of calcite and quartz (qz) crystals. Upper XRD pattern belongs to pure calcite mineral and is projected for sake of comparison.
Table 1
Geochemical content of rhizolith. The detection limit for the (%) concentration is 0.01%.
Layer Ca (%) O (%) C (%) Si (%) K (%) Na (%) Others U (ppm) Th (ppm) DR (Gy/ka) De(Gy) Age (ka)
Inner 40.78 39.25 19.77 e e e e 0.25 (0.14) 0.13 (0.08) 0.668 (0.103) 17.81 (2.09) 26.80 (5.05) Middle 45.23 35.93 16.58 0.43 0.28 0.70 Al (0.14) 2.30 (0.15) 3.51 (0.25) 0.795 (0.076) 83.52 (4.98) 105.21 (15.62) Outer 39.51 37.01 18.89 1.35 0.38 0.93 Fe (0.78) Mg (0.52) Al (0.51) 2.80 (0.23) 3.36 (0.28) 0.927 (0.083) 117.61 (5.69) 126.88 (8.69)
was reduced to its minimum at the time of deposition. Equivalent dose was estimated according to the equation:
De¼ ½ðNTL NbljÞ=ððNTL þ biÞ NTLÞÞ bi;
where NTL is the natural TL, NTLþ bithe natural TL curves after the
i-th added beta dose, bithe i-th administered beta dose in Gy, and
Nbljthe bleached TL for each one of the j-duration (Liritzis, 2010; Polymeris et al., 2012). The application of this equation implies that linearfittings were performed to the dose response curves. As the plateaus obtained for different residuals are expected to be of variable lengths, the starting non-zero residual TL is determined as the residual TL level which, when subtracted from the additive dose growth curve, produces the longest plateau in the temperature-dose plateau test. All residual TL signals from 3 days of exposure to the sun, and longer, provide adequately wide plateaus, over 65e70C in length, ranging from 350 to 410
C. Although it was experimentally established by the residual curve of solar bleaching that sun bleaching for 3 days is adequate to leave only the un-bleachable component, the existence of the plateau in conjunc-tion with its length are both prerequisite criteria for the estimaconjunc-tion of an equivalent dose (Liritzis and Vafiadou, 2005). These plateaus, being longest for the prolonged bleaching times, namely, for 3 days or more, provide important information regarding the bleaching level prior to deposition; a short plateau is assumed to be a sign of incomplete bleaching at deposition (Wintle, 1997; Polymeris et al., 2009, 2012).
The best equivalent dose plateau (Mejdahl, 1988) was calculated in the temperature range between 345 and 420C for residual after 5 days of bleaching and is plotted against glow curve temperature inFig. 5D based on the corresponding linear built-up growth curve (inset ofFig. 5D). The equivalent dose of 17.81 (±2.09) Gy was ob-tained as the mean value of the best plateau for the sample. Errors derived mainly from the uncertainties in curvefitting are 1
s
and were calculated by standard error propagation analysis (Knoll, 1999).Fig. 5. (A) Raw NTL (without bgk subtraction, curve I), background (curve II) and NTL after background subtraction of the inner core of soft CaCO3(curve III). (B) NTL (curve a) along
with three additive glow curves (10, 20 and 30 Gy, curves b, c and d respectively). (C) NTL (black) along with a selection of bleached glow curves (green¼ 1day, blue ¼ 2days, red¼ 5days). (D) Plot of Deversus temperature; solid line indicates best Deplateau; inset presents a typical TL built-up growth curve (For interpretation of the references to colour in
thisfigure legend, the reader is referred to the web version of this article.).
Fig. 6. Residual TL levels after bleaching. Curve (a) corresponds to main Deestimation
The aforementioned procedure for Dedetermination was further
cross-checked as well as strengthened with a simulation dose recovery experiment, administering (a) a prolonged bleaching in the summer sun in Xanthi\Northern Greece over 5 days, and (b) a recovery dose of 18 Gy employing the beta source housed in the Risø OSL/TL system. This specific dose, which was administered since it is equal to the Devalue yielded by TL, was treated as
un-known during the next stages of the dose recovery procedure. The same steps and protocol were applied, in order to recover the estimated value of 18 Gy. Curve b ofFig. 6represents the residual curve of solar bleaching after applying the same bleaching intervals with the original Deestimation approach. As is clear, the residual
level is (a) reached after 5 days instead of 3 in the original study and (b) almost 5e8% higher than the original curve. The recovered dose was estimated to be 16.78± 1.4 Gy. This minor discrepancy could be attributed to possible TL sensitivity changes in calcite at the 330C peak (Galloway, 2003). Nevertheless, this recovered dose yields a ratio of estimated over attributed dose of 0.91, indicating the suit-ability of the specified TL procedure to recover successfully the given dose delivered to the CaCO3sample.
4.4. Luminescence characteristics of middle and outer layers According to both the XRD and SEM-EDX results, as well as
Table 1, the middle and outer layers are of bi-mineral nature, including mostly quartz and calcite. Therefore, the quartz in both cases was used as a chronometer. After sieving, grains of di-mensions 140e180
m
m were obtained. In the case of the outer layer, typical chemical treatment was applied, including treatment with HCl (10%), H2O2(10%), HF (40%) and afinal treatment with HCl(10%) in order to obtain a pure quartz extract. However, asFigs. 2 and 3d show, the dimensions of the second layer (being within the dotted lines) are somewhat limited. Therefore, no chemical treatment was applied to that layer, mostly due to the limited quantity of material extracted. However, as already mentioned, CaCO3does not yield an OSL signal. Aliquots with mass of 5 mg each
were prepared by mounting the material on stainless-steel disks. The preparation of samples was formed under dim red light conditions.
The De accumulated in quartz grains was estimated using the
conventional single-aliquot regenerative-dose protocol (OSL-SAR), based on a comparison of the natural OSL signal with regenerative OSL signals produced by known laboratory doses (Murray and Wintle, 2000, 2003). After applying seven different regenerative doses (25e175 Gy in steps of 25 Gy), the OSL signals were corrected for sensitization using a steady test dose (15 Gy) and then a growth curve was formed. The corrected experimental points on these growth curves werefitted using an exponential function, and the equivalent dose, measured as the natural signal, was interpolated onto the growth curve.Fig. 7shows the effect of varying preheating temperature on the equivalent dose, the recuperation (Rec) and the recycling ratio (RR) for the outer layer. Equivalent doses almost form a plateau for high preheat temperatures, ranging between 240 and 280 C, while for lower temperatures there is a substantial decrease of the measured De, with high scattering. Therefore, the
temperature of 260C was chosen as the preheat temperature for the rest of the measurements of both middle and outer layers. It should be noted that some OSL signals acquired for the middle layer are faint; nevertheless, a trustworthy Dewas obtained.Fig. 8
pre-sents in a radial form the Devalues yielded for each layer.
Corre-sponding mean values as well as standard deviations are included inTable 1. A total of 25 aliquots were measured for the outer layer, while the lack of adequate quantity in the middle layer restricted the number of measured aliquots to 14.
4.5. Dose rate estimation
The annual dose of the radiation environment was estimated using the concentrations of major radioactive isotopes of the ura-nium and thorium series, as well as potassium. Uraura-nium and thorium concentrations were estimated by applying thick source alpha counting while the potassium content was estimated by EDX analysis. Thick source alpha counting measurements were per-formed both in the integral and in the pair counting mode, for discrimination between Th and U (Hossain et al., 2002; Liritzis and Vafiadou, 2012). Dose-rate calculations were made using the con-version factors ofLiritzis et al. (2013b).
In the case offine-grained CaCO3, all three components were
calculated. In fact, the dose rate for the Ca-based inner core com-prises the alpha component arising from the CaCO3itself, the beta
component arising from the CaCO3 itself and the surrounding
layers, as well as the gamma component arising mostly from the outer layer as well as cosmic rays (Prescott and Hutton, 1994). Based on the geochemistry of all layers (Table 1), the dose rate for the calcium carbonate part of the rhizolith was estimated to be 0.668± 0.103 Gy/ka, after corrections applied for moisture of the powder.
In the case of quartz originated from the other two layers, namely the middle and the outer part, only beta and gamma dose rates were taken into account for this study. Alpha radiation was ignored due to the short penetration of alpha particles and very low internal radioactivity of quartz grains. Over time, carbonate-cemented deposits experience accumulation as well as dissolu-tion of carbonate material as a pore-filling substance.Nathan and Mauz (2008) suggested dose rate corrections for carbonate-rich sediments similar to those for water and organic components,
Fig. 7. Dependence of equivalent dose (De), recuperation (Rec) and recycling ratio (RR)
on preheat temperature for aliquots collected from outer layer. Each data point rep-resents mean value for each corresponding preheating temperature. Error bars indicate 1sdeviation.
under the assumption that interstitial material is inert. Thefinal burial dose rates obtained for the middle and outer layers were estimated to be 0.795 ± 0.076 Gy/ka and 0.864 ± 0.083 Gy/ka respectively; these values are presented inTable 1. These dose rates are comparable to the dose rates already reported byErginal et al. (2013) for the same sampling site, which were independently estimated.
5. Luminescence ages and interpretation
Luminescent ages are listed inTable 1. The TL age of the CaCO3
filling the original place of the decayed plant roots was estimated to be 26.8 ka with associated errors up to 19%. Optical ages for the middle and outer layers vary between 105 and 127 ka and exhibit good agreement with independent age estimates from the area, based on OSL (Erginal et al., 2013) and TL (Polymeris et al., 2012) dating of the aeolianite samples. Uncertainties in optical age of the outer layer are less than 7%; however, in the case of the inner layer, the low intensity of the signal results in a relatively large error. These ages indicate that rhizoliths and enclosing aeolian sands that were both preserved at the near surface (1 m) of the aeolianite could provide reliable clues indicating the manner in which late Pleistocene coastal morphodynamics acted on the Black Sea coast of Turkey.
As is clear from the ages of the different materials/layers, the outer and middle layers of aeolianite, dated at 127 ka and 105 ka, respectively, were formedfirstly. The exact deposition period of these aeolian sands, when age accuracy values are taken into consideration, could correspond to the later stage of OIS 5e or early stage of MIS 5d. Considering the Black Sea during the so-called Eemian (Mikulinian) interglacial period, spanning from 125 ka to 65 ka (Panin and Popescu, 2007), the sea-level was approximately 8 m higher than today (Federov, 1978; Chepalyga, 1984; Svitoch et al., 2000), allowing precipitation of abundant carbonates on the shelf plains when the Mediterranean and Black Sea were con-nected. Landward drift of the connective carbonates from subtidal to supratidal, as well as ooids brought by offshore winds, might have contributed to deposition of these components and to cementation of aeolian sands (Abegg et al., 2001; Brooke, 2001). Similar examples are quite common not only in the western Mediterranean coast (Fumanal, 1995), but also in Bahamas (Hearty
and Kindler, 1995; Kindler and Hearty, 1995) and the shoreline of Australia (Murray-Wallace et al., 2001).
The inner, soft carbonate core of the rhizolith was, on the other hand, formed much later, either as a form of the chemical trans-formation process driven by water passing through the aeolianite sediment, allowing deposition of the dissolved organic and inor-ganic constituents as a post-depositional chemical cementation process, or, most likely, by a wind-blown drift from the emerged shelf plain during the last glacial period (MIS 2). This result bears a resemblance to the rhizoliths (dated to 16.18± 1.70 ka using OSL) embedded within aeolianites on the south coast of the island of Bozcaada on the north Aegean coast of Turkey (Kiyak and Erginal, 2010). During that regressive stage, ooid-rich carbonates of the studied rhizolith might have drifted from an emerged marine shelf. The presence of ooids with fresh appearance in conjunction with lack of any evidence for reworking sediments within the carbonates confirms the possibility of deposition during the last glacial low-stand. Thus, luminescence ages obtained from the rhizolith and enclosing aeolianite sands are indicative of deposition during the last glacial and interglacial period, respectively. Deposition of rhi-zolith carbonates and ooids might have occurred after decay of the roots of dune plants which formed canopies on the near surface layers of the aeolianite.
6. Conclusions
A horizontal-lying residual of a rhizolith was recovered 10 km west of S¸ile, Istanbul, in northwest Turkey within laminated oolithic massive aeolianite. It is composed of three distinctive layers or rhizocreations; i.e. an unconsolidated inner core full of pure powdery carbonate, the tightly lithified outer layer consisting of yellowish aeolianite, and a thin grayish ring as transition unit between them. The carbonate infill deposited after the death and decay of the dune plant's roots contains plenty of circular and ellipsoidal-shaped ooids. Comprised of concentrically arranged laminae, the ooids are suggestive of wind-blown drift from an exposed shelf plain into voids left by decayed coastal dune plants. For thefirst time in the literature, the inner core, consisting purely of calcium carbonate, was dated by applying TL. The TL age of the CaCO3infill occupying the original place of the decayed plant roots
was found to be 26.8 (±5) ka. When compared with OSL ages of the middle (105 ka) and outer (127 ka) layers, evidence of the last
Fig. 8. Radial plots for OSL equivalent doses measured for middle (triangles, green zone) and outer (dots, blue zone) layers. Each data point corresponds to an individual, single-aliquot equivalent dose estimation. Measured Devalues (in Gy) for each aliquot can be read by tracing a line from y-axis origin through the point, until line intersects radial axis on
right. Corresponding standard error for each estimate can be read by extending a line vertically to intersect x-axis. Values with smallest relative errors and highest precision are plotted closest to radial axis. All data points lie in 2sregions (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.).
interglacial highstand, the inner core coincides with the last glacial period when the sea-level was lower than the present, promoting transportation of ooids by offshore winds in conjunction with the exposed shelf carbonates. This explains how rhizoliths commonly found within coastal aeolianites cannot be contemporaneous with the host rock aeolianite, which also clarifies why the rhizoliths are exceptionally found embedded within the upper level of the studied sequence. Further work is required in order to perform similar studies at other sampling sites around the coastline of Turkey and the Mediterranean.
Author contributions
G.S.P. designed the methodology of the TL measurements, per-formed the TL measurements, all data analysis, and wrote the pa-per. G.K. helped with interpretation of the luminescence results. N.G.K. designed the methodology for the OSL measurements and performed the OSL measurements. E.T. performed the XRD mea-surements as well as their analysis and their interpretation. N.C.T. performed the dose rate measurements and calculations. A.E. hel-ped with the geographical interpretation of the dating results. Finally, A.E.E. collected the sample, performed the micro-analytical examination and SEM measurements, and wrote the part of the paper dealing with interpretation of the results.
Acknowledgements
A.E.E. thanks the Scientific and Technological Research Council of Turkey (TÜB_ITAK; project code: 113Y418) for their financial support as well as the Turkish Academy of Sciences (TÜBA) as part of the Young Scientist Award Program.
References
Aalto, K.R., Dill, R.F., 1996. Late Pleistocene stratigraphy of a carbonate platform margin (Exumas, Bahamas). Sedimentary Geology 103, 129e143.
Abegg, F.E., Loope, D., Harris, P.M., 2001. Carbonate aeolianites: depositional models and diagenesis. In: Abegg, F.E., Harris, P.M., Loope, D.B. (Eds.), Modern and Ancient Carbonate Aeolianites: Sedimentology, Sequence Stratigraphy, and Diagenesis. SEPM Special Publication 71, pp. 17e30.
Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press, London.
Alonso-Zarza, A.M., Genise, J.F., Cabrera, M.C., Mangas, J., Martin-Perez, A., Valdeolmillos, A., Dorado-Valino, M., 2008. Megarhizoliths in Pleistocene aeolian deposits from Gran Canaria (Spain): Ichnological and palae-oenvironmental significance. Palaeogeography Palaeoclimatology Palae-oecology 265, 39e51.
Bøtter-Jensen, L., Bulur, E., Duller, G.A.T., Murray, A.S., 2000. Advances in lumines-cence instrument systems. Radiation Measurements 32, 523e528.
Brooke, B., 2001. The distribution of carbonate eolianite. Earth-Science Reviews 55, 135e164.
Chepalyga, A.L., 1984. Inland sea basins. In: Velichko, A.A., Wright, H.E., Barnowsky, C.W. (Eds.), Late Quaternary Environments of the Soviet Union. University of Minnesota Press, Minneapolis, pp. 229e247.
Debenham, N.C., 1983. Reliability of thermoluminescence dating of stalagmitic calcite. Nature 304, 154e156.
Debenham, N.C., Aitken, M.J., 1984. Thermoluminescence dating of stalagmitic calcite. Archaeometry 26, 2e9.
Down, J.S., Flower, R., Strain, J.A., Townsend, P.D., 1985. Thermoluminescence emission spectra of calcite and Iceland spar. Nuclear Tracks and Radiation Measurements 10, 581e589.
Duller, G.A.T., Penkman, K.E.H., Wintle, A., 2009. Assessing the potential for using biogenic carbonates as dosemeters for luminescence dating. Radiation Mea-surements 44, 429e433.
Engin, B., Güven, O., 2000. The effect of heat treatment on the thermoluminescence of naturally-occurring calcites and their use as a gamma-ray dosimeter. Radi-ation Measurements 32, 253e272.
Erginal, A.E., Kiyak, N.G., Ekinci, Y.L., Demirci, A., Ertek, A., Canel, T., 2013. Age, composition and paleoenvironmental significance of a Late Pleistocene aeo-lianite from the western Black Sea coast of Turkey. Quaternary International 296, 168e175.
Erginal, A.E., Kiyak, N.G., Ertek, T.A., 2010. Late Pleistocene carbonate dunes and rhizoliths on Bozcaaada Island (Northwestern Turkey) and S¸ile area (Western Black Sea) paleoenvironmental implications for MIS 2 and MIS 3 lowstands. In:
IGCP 521-INQUA 0501 Sixth Plenary Meeting and Field Trip. Rhodes Island, Greece. 27 Sept.-5 Oct. 2010, pp. 67e68.
Federov, P.V., 1978. Pleistocene Ponto-Kaspiia (The Pleistocene of the Ponto-Caspian region). In: Trudy Geologicheskogo Instituta Akademii Nauk SSSR, 168. Nauka, Moscow.
Flugel, E., 2004. Microfacies of Carbonate Rocks. Analysis, Interpretation and Application. Springer-Verlag, Berlin, New York, p. 976.
Frebourg, G., Hasler, C.-A., Guern, P.L., Davaud, E., 2008. Facies characteristics and diversity in carbonate aeolianites. Facies 54 (2), 175e191.
Fumanal, M.P., 1995. Pleistocene dune systems in the Valencian Betic cliffs (Spain). INQUA Subcommission on Mediterranean and Black Sea Shorelines Newsletter 17, 32e38.
Galloway, R.B., 2002. Does limestone show useful optically stimulated lumines-cence? Ancient TL 20, 1e7.
Galloway, R.B., 2003. Limestone: some observations on luminescence in the region of 360 nm. Radiation Measurements 37, 177e185.
Gocke, M., Pustovoytov, K., Kühn, P., Wiesenberg, G.L.B., L€oscher, M., Kuzyakov, Y., 2011. Carbonate rhizoliths in loess and their implications for paleoenvir-onmental reconstruction revealed by isotopic composition:d13C,14C. Chemical Geology 283, 251e260.
Gocke, M., Pustovoytov, K., Wiesenberg, G.L.B., 2010. Rhizoliths in loesse evidence for post-sedimentary incorporation of root-derived organic matter in terrestrial sediments as assessed from molecular proxies. Organic Geochemistry 41, 1198e1206.
Gocke, M., Gulyas, S., Hambach, U., Jovanovic, M., Kovacs, G., Markovic, S.B., Wiesenberg, G.L.B., 2014. Biopores and root features as new tools for improving paleoecological understanding of terrestrial sediment-paleosol sequences. Palaeogeography, Palaeoclimatology, Palaeoecology 394, 42e58.
Habermann, J., Schilles, T., Kalchfruber, R., Wagner, G.A., 2000. Steps towards surface dating using luminescence. Radiation Measurements 32 (5e6), 847e851.
Hearty, P.J., Kindler, P., 1995. Sea-level highstand chronology from stable carbonate platforms (Bermuda and the Bahamas). Journal of Coastal Research 11, 675e689.
Hossain, S.M., De Corte, F., Vandenberghe, D., Van den haute, P., 2002. A comparison of methods for the annual radiation dose determination in the luminescence dating of loess sediment. Nuclear Instruments and Methods in Physics Research A 490, 598e613.
Kalita, J.M., Wary, G., 2014. Thermoluminescence study of X-ray and UV irradiated natural calcite and analysis of its trap and recombination level. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 125, 99e103.
Kalita, J.M., Wary, G., 2015. Thermal quenching in calcite and evaluation of quenching parameters from composite glow curve by a computerized resolved peak technique. Journal of Luminescence 160, 134e137.
Kim, K.B., Hong, D.B., 2014. Kinetic parameters, bleaching and radiation response of thermoluminescence glow peaks separated by deconvolution on Korean calcite. Radiation Physics and Chemistry 103, 16e21.
Kindler, P., Hearty, P.J., 1995. Pre-Sangamonian eolianites in the Bahamas? New evidence from Eleuthera Island. Marine Geology 127, 73e86.
Kiyak, N.G., Erginal, A.E., 2010. Optical stimulated luminescence dating study of eolianite on the island of Bozcaada, Turkey: preliminary results. Journal of Coastal Research 26, 673e680.
Klappa, C.F., 1980. Rhizoliths in terrestrial carbonates: classification, recognition, genesis, and significance. Sedimentology 26, 613e629.
Knoll, F.G., 1999. Radiation Detection and Measurements. J. Wiley& Sons.
Kraus, M.J., Hasiotis, S.T., 2006. Significance of different modes of rhizolith preser-vation to interpreting paleoenvironmental and paleohydrologic settings: ex-amples from Paleogene paleosols, Bighorn Basin, Wyoming, USA. Journal of Sedimentary Research 76, 633e646.
Laskaris, N., Liritzis, I., 2011. A new mathematical approximation of sunlight attenuation in rocks for surface luminescence dating. Journal of Luminescence 131, 1874e1884.
Liritzis, I., 1994. A new dating method by thermoluminescence of carved megalithic stone building. Comptes Rendus de l' Academie des Sciences, Paris (Geo-materials) 319, 603e610.
Liritzis, I., 2000. Advances in thermo- and opto- luminescence dating of environ-mental materials (sedimentary de-posits): part II: applications. Global Nest 2 (1), 29e49.
Liritzis, I., 2010. Strofilas (Andros Island, Greece): new evidence for the cycladic final Neolithic period through novel dating methods using luminescence and obsidian hydration. Journal of Archaeological Science 37, 1367e1377.
Liritzis, I., 2011. Surface dating by luminescence; an overview. Geochronometria 38, 292e302.
Liritzis, I., Bakopoulos, Y., 1997. Functional behaviour of solar bleached thermolu-minescence in calcites. Nuclear Instruments and Methods B 132, 87e92.
Liritzis, I., Singhvi, A.K., Feathers, J.K., Wagner, G.A., Kadereit, A., Zacharias, N., Li, S.-H., 2013a. Luminescence Dating in Archaeology, Anthropology and Geoarchaeology: an Overview. Springer Briefs in Earth System Sciences, Springer Verlag.
Liritzis, I., Vafiadou, A., 2005. Dating by luminescence of ancient megalithic ma-sonry. Mediterranean Archaeology& Archaeometry 5 (1), 25e38.
Liritzis, I., Vafiadou, A., 2012. Calibration aspects of thick source alpha counter ZnS system. Radiation Measurements 45, 1966e1980.
Liritzis, Y., 1989. Dating of calcites: some aspects of radiation survey in caves and dose-rates. Bulletin Geologique des Pays Helleniques 34 (1), 123e136.
Liritzis, I., Stamoulis, K., Papachristodoulou, C., Ioannides, K., 2013b. A re-evaluation of radiation dose rate conversion factors. Mediterranean Archaeology and Archaeometry 13 (3), 1e15.
McDougal, D.J., 1968. Thermoluminescence of Geological Materials. London Uni-versity Press.
Mejdahl, V., 1988. The plateau method for dating partially bleached sediments by thermoluminescence. Quaternary Science Reviews 7, 347e348.
Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved singlee aliquot regenerative e dose protocol. Radiation Measurements 32, 57e73.
Murray, A.S., Wintle, A.G., 2003. The single aliquot regenerative dose protocol: po-tentials for improvements in reliability. Radiation Measurements 37, 377e381.
Murray-Wallace, C.V., Brooke, B.P., Cann, J.H., Belperio, A.P., Bourman, R.P., 2001. Whole rock aminostratigraphy of the Coorong Coastal Plain, South Australia: towards a 1 million year record of sea-level highstands. Journal of the Geological Society of London 158, 111e124.
Nathan, R.P., Mauz, B., 2008. On the dose rate estimate of carbonate-rich sediments for trapped charge dating. Radiation Measurements 43, 14e25.
Ninagawa, K., 1987. Thermoluminescence dating of fossil calcite shells. Journal of Physics D: Applied Physics 26, 2127e2133.
Ninagawa, K., Adachi, K., Uchimura, N., Yamamoto, I., Wada, T., Yamashita, Y., Takashima, I., Sekimoto, K., Hasegawa, H., 1992. Thermoluminescence dating of calcite shells in the Pectinidae family. Quaternary Science Reviews 11, 121e126.
Ninagawa, K., Takahashi, N., Wada, T., Yamamoto, I., Yamashita, N., Yamashita, Y., 1988. Thermoluminescence measurements of a calcite shell for dating. Qua-ternary Science Reviews 7, 367e371.
Okay, 1948. A.C. Istanbul University Review of Faculty of Science (Series B) 13, pp. 311e335.
Panin, N., Popescu, I., 2007. The Northwestern Black Sea: climatic and sea-level changes in the Upper Quaternary. In: Yanko-Homback, V., Gilbert, A.S., Panin, N., Dolukhanov, P.M. (Eds.), The Black Sea Flood Question: Changes in Coastline, Climate, and Human Settlement. Springer, New York, pp. 387e404.
Polymeris, G.S., Erginal, A.E., Kiyak, N.G., 2012. A comparative morphology, compositional and TL study of Tenedos (Bozcaada) and S¸ile aeolianites, Turkey. Mediterranean Archaeology and Archaeometry 12 (No 2), 117e131.
Polymeris, G.S., Kitis, G., Afouxenidis, D., Sfampa, I.K., Tsirliganis, N.C., Rousaki, A., Kouloumpi, E., Paraskevopoulos, K.M., 2013. On the feasibility of dating portable paintings: preliminary luminescence measurements on ground layer materials. Mediterranean Archaeology and Archaeometry 13 (3), 93e103.
Polymeris, G.S., Kitis, G., Liolios, A.K., Sakalis, A., Zioutas, K., Anassontzis, E.G., Tsirliganis, N.C., 2009. Luminescence dating of the top of a deep water core from the NESTOR site near the Hellenic Trench, east Mediterranean Sea. Quaternary Geochronology 4, 68e81.
Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for lumi-nescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, 497e500.
Retallack, G.J., 2001. Soils of the Past: an Introduction to Paleopedology, second ed. Blackwell Science, Oxford, U.K. 404 pp.
Singhvi, A.K., Sharma, Y.P., Agrawal, D.P., 1982. Thermoluminescence dating of sand dunes in Rajasthan, India. Nature 295, 313e315.
Stirling, R.J., Duller, G.A.T., Roberts, H.M., 2014. Developing a single-aliquot protocol for measuring equivalent dose in biogenic carbonates. Radiation Measurements 47, 725e731.
Subedi, B., Oniya, E., Polymeris, G.S., Afouxenidis, D., Tsirliganis, N.C., Kitis, G., 2011. Thermal quenching of thermoluminescence in quartz samples of various origin. Nuclear Instruments and Methods in Physics Research B 269, 572e581.
Svitoch, A.A., Selivanov, A.O., Yanina, T.A., 2000. Paleohydrology of the Black Sea Pleistocene basins. Water Resources 27 (6), 594e603.
Wintle, A.G., 1977. Thermoluminescence dating of minerals - traps for the unwary. Journal of Electrostatics 3, 281e288.
Wintle, A.G., 1997. Luminescence dating: laboratory procedures and protocols. Radiation Measurements 27, 769e817.
