JOURNAL OF QUATERNARY SCIENCE (2014) 29(2) 189–199
ISSN 0267-8179. DOI: 10.1002/jqs.2693
Environmental dynamics and luminescence chronology from the Orlovat loess–palaeosol sequence (Vojvodina, northern Serbia) ´ ,1* ALIDA TIMAR-GABOR,2 THOMAS STEVENS,1,3 ULRICH HAMBACH,4 DRAGAN POPOV,1 SLOBODAN B. MARKOVIC 1 ´ ´6 NEMANJA TOMIC, IGOR OBREHT,1,† MLADJEN JOVANOVIC´,1 FRANK LEHMKUHL,5 HOLGER KELS,5 RASTKO MARKOVIC and MILIVOJ B. GAVRILOV1 1 Chair of Physical Geography, Faculty of Sciences, University of Novi Sad, 21000, Novi Sad, Serbia 2 Faculty of Environmental Science, Babes¸-Bolyai University, Faˆntaˆnele 30, 400294 Cluj-Napoca, Romania, and Interdisciplinary Research Institute on Bio-Nano-Science of Babes¸-Bolyai University, Treboniu Laurean 42 400271, Cluj-Napoca, Romania 3 Centre for Quaternary Research, Department of Geography, Royal Holloway University of London, Egham, Surrey TW20 0EX, UK 4 Chair of Geomorphology, University of Bayreuth, 95440, Bayreuth, Germany 5 Department of Geography, RWTH Aachen University, Wu¨llnertsr. 5b D-52056, Aachen, Germany 6 High school “Isidora Sekulic´”, Vladike Platona 2, 21000 Novi Sad, Serbia Received 5 July 2013; Revised 13 January 2014; Accepted 18 January 2014
ABSTRACT: The Carpathian Basin contains some of the best preserved loess deposits in Europe, including some of the continent’s longest and best resolved climate records. Large areas of the basin have been intensively investigated in recent years, although deposits in the east remain largely unstudied, despite considerable regional variation in climate records. Here we discuss the sedimentary record exposed in the Orlovat brickyard using detailed litho- and pedo-stratigraphic, enviromagnetic parameters and luminescence dating. The results show an atypical Late Pleistocene succession for the Carpathian Basin. Notably, the normally widespread pedocomplex V-L1S1 is missing. This contrasts with other parts of the sequence, which appear highly resolved, such as the thicker pedocomplex V-S1 and the detailed transitions between modern pedocomplexes V-S0 and the last glacial loess unit V-L1. The luminescence chronology demonstrates a lack of intensive pedogenesis during the Early Holocene and raises an important general question about the beginning of Holocene soil formation in the region. The later Holocene soil formation adds to a growing body of evidence that suggests more complex terrestrial responses of climate to global climate change. This evidence weakens the validity of previously generalized direct stratigraphic correlations between regional terrestrial environmental archives, and global marine and ice core records. Copyright # 2014 John Wiley & Sons, Ltd. KEYWORDS: enviromagnetism; loess; luminescence dating; Serbia; Orlovat.
Introduction Loess deposits in the Carpathian Basin provide evidence for highly variable climate, dust deposition and soil formation over the region during the Quaternary (Fuchs et al., 2008; Markovic´ et al., 2008; Schmidt et al., 2010; Stevens et al., 2011). It is not clear whether this implies strong regional gradients in environment or local differences in loess sedimentation. While certain areas of the Carpathian Basin loess region have been well studied (notably the north, central and west of the basin), deposits in the east have received little attention and the late Quaternary environmental and dust accumulation record from that region is poorly known. This represents a major gap in understanding climate and dust deposition variability over the region. The northern Serbian province of Vojvodina is a lowland part of the southern Carpathian Basin encompassing the confluence of the Danube, Sava, Tisa (Tisza), Drava (Drau), Morava and Tamis (Temes, Timis¸) rivers (Markovic´ et al., 2008). These rivers separate three Vojvodinian sub-regions: Srem, Backa and Banat, which partly extend to the neighbouring countries of Croatia, Hungary and Romania (Fig. 1). Loess sediments in the Vojvodina region are among the oldest and most complete loess–paleosol formations in Europe. These thick sequences contain a detailed palaeoclimatic record since the Early Pleistocene (Markovic´ et al., 2011).
Correspondence: S. B. Markovic´, as above. E-mail:
[email protected] † Present address: 5Department of Geography, as above.
Copyright # 2014 John Wiley & Sons, Ltd.
Recently, Vojvodina loess–palaeosol sequences have been intensively investigated, especially at the Srem, Titel and Backa loess plateaus (Markovic´ et al., 2006, 2007; Antoine et al., 2009; Bokhorst et al., 2009, 2011; Schmidt et al., 2010; Stevens et al., 2011). The continuous presence of apparently much drier conditions in the region and persistence of stable ‘plateau’ accumulation (Markovic´ et al., 2012) has resulted in more detailed and long-term sedimentary record preservation than in other European areas. However, the Tamis loess plateau of the eastern Carpathian Basin has not previously been investigated. Lying on the Tamis plateau, the Orlovat loess site is located further to the east than similar previously investigated Serbian loess sections (Fig. 1). Its geographical position provides a unique opportunity to reconstruct climatic and environmental evolution in the transitional area between the south-eastern limit of the Carpathian Basin and the Western Carpathian slopes and allows us to extend our understanding of regional climate and dust deposition into the poorly understood east of the basin. The Carpathians are located at the boundary between central and eastern European lowlands, with apparently sharp transitions in morphology and climatic conditions. During the winter, conditions are dominated by inflow of polar-continental air masses arriving from the east and northeast, while during other seasons oceanic air masses from the west predominate. The morphology of the Carpathian Mountains also controls the area’s precipitation distribution. Most of the mountain chain is characterized by humid environmental conditions, contrasting with continental conditions in the intermontane depressions and on the lower parts of the
190
JOURNAL OF QUATERNARY SCIENCE
Figure 1. The study area. (A) Map of the Vojvodina region with the geographical positions of the main loess sections (Markovic´ et al., 2004, modified). (B) A geomorphological map of the Tamis loess plateau surrounded by the Tamis and Begej river valleys (Popov et al., 2012a, modified). This figure is available in colour online at wileyonlinelibrary.com.
Copyright # 2014 John Wiley & Sons, Ltd.
J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
THE ORLOVAT LOESS–PALAEOSOL SEQUENCE, NORTHERN SERBIA
191
Figure 2. Topographic profile A–A0 (Fig. 1) from the investigated Orlovat section to the Tamis river. This figure is available in colour online at wileyonlinelibrary.com.
southern mountain slopes leading to the Valahian plain. The Carpathians thus represent a humid biogeographical island and biodiversity hotspot. Considering the temporal dynamics of Pleistocene biogeographical change, the transitional area between the Carpathian Basin and the south-western Carpathian slopes can be regarded as a zone of significant biogeographical and biodiversity variation (e.g. Su¨megi and Krolopp, 2002), potentially accounting for the variation noted from previous research into Quaternary loess profiles. This study presents the results of the first detailed multi-proxy and chronometric investigation of a loess–palaeosol sequence from this region: the Serbian part of the Banat region in the south-east of the Carpathian Basin. The environmental conditions and dust depositional record at the site are compared with previously investigated loess sequences elsewhere in the Vojvodina region to gauge variability in climate and loess-dust accumulation and preservation across the Carpathian Basin.
Geomorphological setting and methods The Tamis loess plateau is the smallest loess plateau in the Vojvodina region, extending between the settlements of Ecka, Botos, Orlovat and Farkazdin, as a slightly elevated geomorphological unit between the floodplain of the Tamis River (on the NE–ESE), the palaeo channel Petra (west) and the palaeo channel Sozov (north). These palaeochannels are relicts of ancient fluvial activity of, most probably, the palaeo Tisa River (Figs 1 and 2). Hydromorphometric parameters of the Petra channel and recent Tisa meanders support this assumption and it is plausible that the Begej or even Tamis rivers used this riverbed after the Tisa had shifted westwards (Popov et al., 2008, 2012a, 2012b). Thus, the current distribution of the Tamis loess plateau is a relic of a previously much larger loess belt reduced in extent by intensive fluvial erosion. The investigated loess–palaeosol sequence is exposed in a brickyard at the village of Orlovat (45˚150 N, 20˚350 E, 88 m a. s.l; 24 km SE of the city of Zrenjanin), in the central part of the Tamis loess plateau in contact with the Tamis river valley (Fig. 1). Recent and palaeorelief conditions are very important to consider in interpreting this site. Figure 2 shows a topographic profile from the Orlovat brickyard, over the Tamis loess plateau to the Tamis River alluvial plain. To illustrate the palaeorelief conditions we used spatial data about the extent of the V-S1 palaeosol obtained at other existing exposures. Copyright # 2014 John Wiley & Sons, Ltd.
During April and May 2012, the profile near Orlovat was carefully cleaned and sampled for rock magnetic properties and sediment colour proxies as well for grain size and geochemical analyses. About 500 g sample material was collected for magnetic analysis in plastic bags at 5-cm intervals. Additionally, eight samples were collected for luminescence dating. Here we report the field observations (litho- and pedostratigraphy), magnetic measurements and the results from absolute luminescence chronology.
Litho- and pedostratigraphy The profile was cleaned and described in detail on a 10-mhigh vertical loess cliff. Sediment colour was determined on dried samples in the laboratory, using the Munsell Soil Colour Chart. Based on investigations at various loess exposures in Vojvodina, Markovic´ et al. (2008) developed a stratigraphic labelling scheme following the Chinese loess stratigraphic model (Kukla, 1987; Kukla and An, 1989). The loess and palaeosol stratigraphic units were designated as L and S and were numbered in order of increasing age. The prefix ‘V’ is used to refer to the Pleistocene loess–palaeosol stratigraphy in Vojvodina (Markovic´ et al., 2012). This scheme is applied to the Orlovat section for ease of comparison.
Enviromagnetic properties Environmental magnetic analyses were carried out on bulk samples. The dried sediment was packed into plastic boxes, and subsequently compressed and fixed with cotton wool to prevent movement of sediment particles during measurement. The individual density of the specimens served to normalize the results. The initial low field susceptibility [x (108 m3 kg1)] was measured in an AC-field of 300 A m1 at 920 Hz using the KLY-3-Spinner-Kappa-Bridge (AGICO, Brno, Czech Republic). The frequency dependence of susceptibility (xfd, %) was determined with a MAGNON Susceptibility Bridge (MAGNON, Dassel, Germany) at AC-fields of 300 A m1 at 0.3 and 3 kHz at the Palaeomagnetic Laboratory of the Chair of Geomorphology, University of Bayreuth.
Luminescence dating Sampling and sample preparation The eight samples for luminescence dating were collected by hammering opaque PVC or stainless steel tubes into the J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
192
JOURNAL OF QUATERNARY SCIENCE
freshly cleaned loess profiles. Sample preparation was performed under low-intensity red light conditions in the luminescence laboratory of Babes¸-Bolyai University, ClujNapoca, Romania. Three days of HCl (concentration 10%) treatment was employed for removal of carbonated material followed by another 3 days of H2O2 (30%) treatment for organic matter removal. The 63–90-mm fraction was separated through wet sieving. A two-step density separation using heavy liquids and centrifuging was performed to isolate quartz from other minerals. The heavy liquid solution contained sodium metatungstate, Na6[H2W12O40].xH2O, and distilled water. First, the quartz (density 2.65–2.66–g cm3) and plagioclase grains were separated in a –2.63 g cm3 heavy liquid solution from the lighter minerals, such as potassium and sodium feldspars. The second step employed a – 2.75-g cm3 heavy liquid solution, in which the quartz and plagioclase feldspars were isolated from the heavy minerals, such as zircons and apatite. Quartz grains were isolated from this fraction through an HF acid (40% concentration) treatment for 40 min. The etching also removed the outer surface of the quartz grains, reducing to negligible the internal ionizing alpha radiation contribution to the grains. Precipitated fluorides after the HF treatment were removed via a 60min HCl (10%) immersion. Instrumentation Quartz grains were mounted as large (8 mm) aliquots in a monolayer using silicone oil on a 10-mm-diameter stainless steel disc. Luminescence investigations were carried out using an automated TL/OSL-DA-20 Risø reader equipped with blue (470 30 nm) light-emitting diodes. Stimulation with infrared (IR) light was through IR light-emitting diodes (875 80 nm). All luminescence emissions were detected through a 7.5-mmthick Hoya U-340 UV filter. Further details regarding the equipment can be found in Thomsen et al. (2008). The radioactive 90Sr-Y beta source was calibrated using gammairradiated calibration quartz supplied by Risø National Laboratory and had a dose rate to sandy grains (63–90 mm) on stainless steel discs of 0.15 Gy s1. The heating rate was 5 ˚C s1. All thermal treatments and stimulations at temperatures higher than 200 ˚C were carried out in a nitrogen atmosphere. The specific activities of radionuclides of interest for dose rate determination were obtained through high-resolution gamma spectrometry using an ORTEC hyperpure germanium detector having the following characteristics: active volume of 181 cm3, 0.878 keV FWHM at 5.9 keV, 1.92 keV FWHM and 34.2% relative efficiency at 1332.5 keV, calibrated in efficiency using International Atomic Energy Agency standards.
and possible feldspar contamination, the IR depletion ratio was employed. This represents the ratio of two repeated dose measurements, the second one having an IR stimulation before the (blue) optically stimulated luminescence (OSL). The sensitivity to IR stimulation was defined as significant if the IR depletion ratio deviated by more than 10% from unity, with aliquots not satisfying this criterion being rejected (Duller, 2003).
Results Litho- and pedo-stratigraphy The stratigraphic framework of loess in the Vojvodina region is generally relatively simple (Markovic´ et al., 2008) because the sequences were formed from relatively continuous deposition of aeolian dust on near-horizontal platforms similar to the Chinese loess plateau (e.g. Liu, 1985; Kukla, 1987; Kukla and An, 1989). Here we test whether this is also the case at the Orlovat site in the south-east of the basin and present the first chronostratigraphic model for loess in this area. The profile has a total thickness of approximately 10 m. The loess unit V-L2 is exposed for only about 100 cm at the base of profile. The lower part of this loess unit has not been excavated during previous raw material exploitation. The unit displays light yellowish brown (10YR 6/4) sediment scattered with many root channels and carbonate concretions. The thickness of the overlying strongly developed soil complex VS1 is approximately 400 cm. A major change in sedimentology occurs at the pedocomplex V-S1 boundaries, with loess layers V-L1 above and V-L2 below it. The basal part of pedocomplex V-S1 is a strongly developed olive brown (2.5Y 4/3) blocky AB horizon that gradually transforms from light olive brown (2.5Y 5/4 and 5/6) to three Ah granular horizons. The uppermost part of the pedocomplex includes a light yellowish brown loessic layer, a granular Ah3 horizon and a weakly developed initial A horizon at the top (Fig. 3). In contrast to other loess sections in the Vojvodina region, a weakly developed palaeosol complex V-L1S1 is not detected at Orlovat. V-L1S1 normally corresponds to Marine Isotope Stage (MIS) 3 (Stevens et al., 2011). At Orlovat, only an undifferentiated V-L1 is represented, reaching 405 cm thickness. It is a typical porous loess (10YR 6/3) with one initial pedogenetic horizon in the middle and two similar weakly developed palaeosols below the modern soil V-S0. The Holocene soil S0 spans the upper 75 cm of the section with an upper steppic Ah horizon and a lower transitional AC horizon with small soft spherical carbonate concretions (from 1 to 2 cm in diameter). The whole of the uppermost part of the section is scattered with krotovinas (Fig. 3).
Enviromagnetic record Measurement protocol The luminescence characteristics of the samples were investigated using a single-aliquot regenerative-dose (SAR) protocol (Murray and Wintle, 2000). Stimulation with the blue lightemitting diodes was for 40 s at 125 ˚C. The initial 0.30 s of the decay curve was used, less a background evaluated from the interval 2.31–3.08 s. A constant test dose of 15 Gy was used in all experiments. Unless stated otherwise, a preheat of 10 s at 220 ˚C preceded the measurement of natural and regenerative signals. A cutheat to 180 ˚C (ramp heating held for 0 s at maximum temperature) preceded the stimulation of each test dose signal. Following measurement of each test dose signal, a high-temperature bleach was performed by stimulating with the blue light-emitting diodes for 40 s at 280 ˚C (Murray and Wintle, 2003). To define the sensitivity to infrared stimulation Copyright # 2014 John Wiley & Sons, Ltd.
Mass-specific magnetic susceptibility (x) variations are generally similar to those seen in previously investigated loess– palaeosol sequences along the Danube River, at the eastern and northern shores of the Black Sea, and in Central Asia and China (e.g. Heller and Evans, 1995; Evans and Heller, 2001; Buggle et al., 2009; Markovic´ et al., 2009). The absolute x values vary between 19 and 96 108 m3 kg1. These values are similar to previously measured low field magnetic susceptibility records from other loess sites in the Vojvodina region (Fuchs et al., 2008; Markovic´ et al., 2008, 2011, 2012; Bokhorst et al., 2009; Stevens et al., 2011). The two major soil complexes V-S1 and V-S0 have significantly higher x values (from 35 to 60 108 m3 kg1, and from 55 to 96 108 m3 kg1, respectively) than the loess unit V-L1 (small amplitude of variations around 20– 108 m3 kg1). J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
THE ORLOVAT LOESS–PALAEOSOL SEQUENCE, NORTHERN SERBIA
193
Figure 3. Sediment colour, x and xfd records related to pedostratigraphy of the Orovat loess–palaeosol sequence. Ages shown next to the sequence represent the results of luminescence dating. This figure is available in colour online at wileyonlinelibrary.com.
However, contrary to the typical pattern of the magnetic signal recorded at other previously investigated sections in Vojvodina, the typically gentle increase of x values related to the pedocomplex V-L1S1 is not observed at Orlovat (Fig. 3). The frequency-dependent susceptibility (xfd) varies from approximately 0.2 to 10% with significantly higher values in pedocomplex V-S1 and peaks in the recent soil V-S0, as compared with loess unit V-L1. Values of xfd gradually decrease upwards in pedocomplex V-S1 from almost 9 to 2%. In loess unit V-L1, xfd values range from 0.23 to 8.86% with a clearly gradually increasing trend. Modern soil V-S0 has high values of xfd from 8.73 to 9.71% (Fig. 3) and, again, there is no sign of V-L1S1 in the xfd record.
Luminescence dating Luminescence characteristics Figure 4(a) presents a typical dose–response curve for the sample. The growth of the signal with dose passes very close to the origin (i.e. recuperation is negligible) and the correction for sensitivity change is apparently properly performed (recycling ratio is close to unity). The OSL signals exhibit a rapid decay during optical stimulation and the natural and regenerated signals have the same shape and appear indistinguishable (Fig. 4a, inset). The parameters indicating the suitability of the measurement protocol for all the samples analysed are also presented in Table 1. For high doses (Fig. 4b), the growth of the signal with dose is best represented by the sum of two single saturating exponential functions of the form:
IðDÞ ¼ I0 þ Að1 expðD=D01 ÞÞ þ Bð1 expðD=D02 ÞÞ Copyright # 2014 John Wiley & Sons, Ltd.
This has been as reported by other recent studies on quartz (Lowick et al., 2010; Chapot et al., 2012; Timar-Gabor et al., 2012). It can also be seen that the artificial dose–response is not dependent on sample age (Fig. 4b) and there is very little scatter between aliquots and samples. The average characteristic doses (D01 and D02) of 33 and 300 Gy have been obtained on 20 aliquots of samples ORL1, ORL2, ORL3 and ORL4, for which dose–response curves have been constructed for doses up to 1000 Gy. Contamination with a thermally less stable slow component is unlikely to be significant considering the background subtraction and the shape of the decay curves. However, we have performed a preheat plateau test to test this further. The preheat plateau is specifically intended to isolate a thermally stable signal; we observed a lack of dependence of the equivalent dose on preheat temperature (Fig. 5). Thus, it can be concluded that: (i) the dosimetric signal investigated comes from traps that are thermally stable; (ii) any unwanted thermal transfer is not significant; and (iii) sensitivity changes are properly corrected for (as different thermal treatments may lead to a different degree of sensitivity change). The applicability of the SAR protocol to date the fine sandsized (63–90 mm) quartz grains was also tested through a dose–recovery test. Natural aliquots were bleached twice for 250 s at room temperature using blue light-emitting diodes; the two bleaching treatments were separated by a 10-ks pause. Aliquots were then given a known dose chosen to be approximately equal to the estimated equivalent dose, and measured using the SAR protocol. The results are shown in Fig. 6. The given dose can be recovered reasonably accurately over the entire dose range (i.e. from 30 to 350 Gy). The results from the dose–recovery test indicate that the SAR J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
Copyright # 2014 John Wiley & Sons, Ltd.
9.2 7.9 7.9 6.7 6.7 6.7 6.6 6.6 11 10.5 5.8 5.6 5.6 5.3 5.0 3.9 159 23 92 12 81 8 70 6 67 6 43 4 13 1 10 1 2.18 0.02 2.49 0.02 2.76 0.02 2.57 0.02 2.70 0.02 3.10 0.02 2.78 0.02 2.77 0.02 20 15 10 10 10 10 10 10 406 4 456 7 471 7 444 7 462 6 546 7 481 7 476 6 35.6 0.3 38.6 0.4 38.2 0.3 36.0 0.4 37.1 0.3 42.6 0.3 34.9 0.4 35.7 0.2 32.4 0.9 31.9 0.6 37.0 0.6 32.1 0.4 34.4 0.4 37.4 0.5 35.7 0.6 34.8 0.7 0.19 0.03 0.20 0.08 0.09 0.02 0.15 0.06 0.05 0.02 0.07 0.01 0.09 0.01 0.11 0.01 0.91 0.01 0.90 0.01 0.94 0.01 0.94 0.01 0.96 0.01 0.96 0.01 0.96 0.01 0.95 0.01 965 700 600 500 400 300 180 150 ORL1 ORL2 ORL3 ORL4 ORL5 ORL6 ORL7 ORL8
347 38 (8/12) 229 24 (5/8) 225 13 (10/16) 181 10 (11/16) 180 10 (10/14) 133 7 (13/13) 36.3 1.8 (13/13) 28.8 1.1 (13/13)
0.93 0.01 0.91 0.01 0.94 0.01 0.94 0.01 0.96 0.01 0.95 0.01 0.97 0.01 0.97 0.01
Relative random error sr (%) Age (ka) Dose rate (Gy ka1) Water content (%) K (Bq kg1)
40
Th (Bq kg1)
232
Ra (Bq kg1)
226
Recuperation (%) IR depletion ratio De (Gy)
Table 1 summarizes the information relevant to the age calculation. Dose rates were calculated from radionuclide activities using the conversion factors tabulated by Adamiec and Aitken (1998). A factor of 0.94 ( 5% relative uncertainty) was adopted to correct the external beta dose rates for the effects of attenuation and etching (Mejdahl, 1979) and an internal dose rate of 0.010 0.002 Gy ka1 was assumed for the coarser fraction, based on the recommendation of Vandenberghe et al. (2008). The dose rate contribution from cosmic rays was calculated based on the formula given by Prescott and Hutton (1994). The as-found moisture content was used for samples collected from and beneath the palaeosol horizon. For samples collected from L1, the measured moisture content ranged between 6 and 12%, and an average moisture content of 10% with an assumed relative error of 25% was used in calculations. Uncertainties on the ages were calculated following the error assessment system formalized by Aitken and Alldred (1972) and Aitken (1976). It can
Recycling ratio
Luminescence ages
Depth (cm)
protocol is able to measure laboratory doses given before any heat treatment both accurately and reasonably precisely.
Lab. code
Figure 4. (A) Representative SAR growth curve for a single, representative aliquot of quartz extracted from sample ORL6. Recycling and IR depletion ratio points are shown as open triangles. The recuperation point is represented as an open square, and the natural signal is depicted as a star. The inset shows the natural and a regenerated OSL decay curve. (B) Comparison of average SAR–OSL growth curves for four samples of different ages; the number of aliquots (n) used to obtain the average is specified in the legend. The average dose response fitting function for all aliquots is given. This figure is available in colour online at wileyonlinelibrary.com.
Relative systematic error ss (%)
JOURNAL OF QUATERNARY SCIENCE Table 1. Summary of equivalent doses (De), recycling ratios, recuperation values, IR depletion ratios, calculated dose rates, optical ages, and random (sr) and systematic (ss) uncertainties. The uncertainties in the luminescence and dosimetry data are random; the uncertainties quoted with the optical ages are the overall uncertainty. All uncertainties represent 1s. The number of replicate De measurements is given in parentheses. The total dose rate includes the contribution from cosmic rays, and allowance was made for the effect of moisture.
194
J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
THE ORLOVAT LOESS–PALAEOSOL SEQUENCE, NORTHERN SERBIA
195
Figure 5. Preheat plateau test for sample ORL6. For all investigated preheat treatments the cutheat was fixed to 180 ˚C. With the exception of the 220 ˚C treatment (13 aliquots), for each temperature four aliquots were used and the data plotted present the average value.
be seen that the systematic uncertainties are usually larger than the random uncertainties. Random uncertainties increase with the age of the sample (from 4% for ORL8 to 11% relative error for samples ORL1 and ORL2) and mainly reflect the scatter in the obtained equivalent doses. This scatter can be attributed to the shape of the growth curve in the region where the natural signals are interpolated. Nevertheless, this is not unusual, as other recent studies also report significant scatter for equivalent doses obtained using coarse quartz extracted from loess (Lomax et al., 2013).
Discussion Figure 7 compares the stratigraphy and luminescence chronology of four loess sections in the Vojvodina region: Crvenka (Stevens et al., 2011), Surduk (Fuchs et al., 2008), Stari Slankamen (Schmidt et al., 2010) and Orlovat. According to proposed multi-millennial chronostratigraphy for Serbian loess sections (Martinson et al., 1987; vanKreveld et al.,
Figure 6. Summary of dose recovery data for all samples using a preheat of 10 s at 220 ˚C. The given dose was chosen to approximate the equivalent dose. The solid line (eye guide) represents the 1: 1 relation; the dotted lines (eye guide) bracket a 10% deviation from unity. Copyright # 2014 John Wiley & Sons, Ltd.
Figure 7. Comparison between luminescence chronologies of Crvenka (Stevens et al., 2011), Surduk (Fuchs et al., 2008), Stari Slankamen (Schmidt et al., 2010) and Orlovat loess sections and MIS chronostratigraphy (Martinson et al., 1987; vanKreveld et al., 2000; Thompson and Goldstein, 2006). This figure is available in colour online at wileyonlinelibrary.com.
2000; Thompson and Goldstein, 2006; Markovic´ et al., 2008, 2012) the V-S0 unit should correspond with MIS 1 (approximately 0–12 ka), V-L1L1 covers MIS 2 (12–25 ka), V-L1S1 is related to MIS 3 (25–56 ka), V-L1L2 is an equivalent of MIS 4 (56–80 ka), V-S1 covers MIS 5 (80–130 ka) and V-L2 should represent MIS 6 (>130 ka). However, the results presented indicate only very general agreement between the luminescence dating and the expected ages based on a correlation to the MIS stratigraphy-based age model (Martinson et al., 1987; Markovic´ et al., 2008). While the models broadly match to about 40 ka, below this (i.e. lower part of V-L1S1, V-L1L2, VS1 and V-L2) all the luminescence ages are younger than those expected from MIS stratigraphy. However, some sedimentological properties of the Orlovat loess–palaeosol sequence demonstrate several important differences when compared with other previously investigated loess sites in the Vojvodina region (e.g. Markovic´ et al., 2008, 2009, 2011; Antoine et al., 2009; Bokhorst et al., 2009; J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
196
JOURNAL OF QUATERNARY SCIENCE
Figure 8. Correlations between Mosorin, Orlovat and Ruma xfd records. Potential depositional hiatuses are indicated with a ‘?’. This figure is available in colour online at wileyonlinelibrary.com.
Stevens et al., 2011). The most important include the lack of the Middle Pleniglacial pedocomplex V-L1S1, and the unusually thick pedocomplex V-S1 (405 cm), apparently characterized by higher accumulation rates during the final part of MIS 5 and MIS 4 based on luminescence dates, magnetic properties, and litho- and pedo-stratigraphic features. There is also evidence of at least two phases of initial pedogenesis during the Early Holocene and for a relatively late start of the formation of the strongly developed chernozem V-S0, as indicated by the luminescence dating. We suggest that the reason for this is related to the site’s palaeogeomorphological conditions. Almost all previously investigated loess sections in the region are associated with plateau-like dust deposition, characterized by stable and quasi-continuous accumulation and sediment preservation, at least over multi-millennial timescales (e.g. Markovic´ et al., 2012). The lower part of the Orlovat loess–palaeosol sequence indicates sedimentary conditions characterized by the influence of occasional slope processes at the southern limit of the Tamis loess plateau. Figure 2 shows that reconstruction of the palaeotopographic surface at the base of pedocomplex V-S1 at Orlovat was about 4 m lower than at approximately 1-km distant exposures on the Tamis River bank. However, at the end of formation of V-S1 this difference in relative altitude was reduced to 2.5 m, indicating a significantly thicker last interglacial–early glacial palaeosol at Orlovat, probably due to basin infilling and slope deposition. This observation is clearly supported by luminescence ages of samples ORL2 (92.0 12.0 ka) and ORL3 (81.0 8.0 ka), confirming high deposition rates during the latest part of MIS 5 (0.09 mm a1) despite intense soil formation. The middle Pleniglacial interval is represented in the Vojvodina region by a weakly developed soil complex L1S1, which appears either as a single, complete pedohorizon (Ruma site), or as double (Batajnica, Irig, Miseluk, Susek and Copyright # 2014 John Wiley & Sons, Ltd.
Petrovaradin) or multiple (Stari Slankamen, Titel and Crvenka) palaeosols (Markovic´ et al., 2008). Loess sub-layers intercalated into V-L1S1 have preserved evidence of sudden changes in climatic and environmental conditions. In contrast to other European Late Pleistocene loess records (Vandenberghe et al., 1998; Antoine et al., 2001; Rousseau et al., 2001; Schatz et al., 2011), the middle Pleniglacial palaeosol V-L1S1 is weakly developed in the Vojvodina region. At the Orlovat loess section, evidence for this middle Pleniglacial pedogenesis is not observed at all. In contrast to other sections in the Vojvodina region (Markovic´ et al., 2005, 2006, 2007, 2009, 2011; Fuchs et al., 2008; Bokhorst et al., 2009; Stevens et al., 2011), the Orlovat sequence does not indicate a typical pattern of gently increased x values associated with the appearance of the V-L1S1 pedocomplex. xfd, which is an even more sensitive indicator of pedogenetic intensity, also indicates just two main periods of pedogenesis in the sequence, related to interglacial pedocomplexes V-S1 and V-S0, without any significant pedogenetic overprint on the loess V-L1 unit. Figure 8 shows a comparison between xfd records of the Orlovat, Ruma and Mosorin sections in the Vojvodina region (Fig. 1). Arrows represent the generalized xfd trends. Dashed arrows indicate intervals of more highly resolved deposition at the Orlovat section. A missing sedimentary interval at Orlovat is indicated by a lack of a typical pattern of xfd records seen in Carpathian Basin loess that is related to slightly higher values associated with the appearance of the V-L1S1 complex, as observed in both the Ruma and the Batajnica sites further west. This interval is indicated with a ‘?’ in Fig. 8. This anomaly in the Orlovat environmental magnetic record indicates a potential depositional hiatus between approximately 2.5 and 2 m depth in the Orlovat profile. It is partly supported by the luminescence dates. After a relatively smooth decrease in luminescence age from the top of palaeosol V-S1 (ORL2, 92 12 ka) to the middle part J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
THE ORLOVAT LOESS–PALAEOSOL SEQUENCE, NORTHERN SERBIA
of the last glacial loess unit V-L1 (ORL5, 67 6 ka) a chronological shift occurs below and above sample ORL6 (43 4 ka) over the interval between 4 and 3 m depth, indicating a significant reduction in accumulation rates, or the likely occurrence of a hiatus in sedimentation. To identify more precisely these depositional gaps we plan to apply more detailed sampling for luminescence dating, similar to that presented in the recent studies of Stevens et al. (2006, 2007, 2008) on the Chinese Loess Plateau. Thus, two independent lines of evidence, the magnetic record and luminescence chronology, confirm a discontinuity in deposition as suggested by an incomplete stratigraphy when compared with sections further west. However, on the basis of the existing results it is very difficult to indicate precisely the age of the hiatuses in the Orlovat loess–palaeosol sequence, especially given the large variability in the sedimentological and pedological characteristics of pedocomplex V-L1S1 in the region. These sedimentary gaps can also be linked to a reduction in the Tamis loess plateau surface area by intensive fluvial erosion. Dry valleys on the Tamis loess plateau, and former river beds and large palaeomeanders indicated from the fluvial terraces near the investigated section at Orlovat are relicts of ancient fluvial activity (Fig. 1). For example, Popov et al. (2012b) reported quite intensive fluvial activity at the nearby sequence at Muzlja sand pit (Tisa River valley) during the final period of terrace formation between 15 and 18 ka. Finally, the two uppermost luminescence dates clearly demonstrate an Early Holocene age almost 0.7 m below the Holocene soil V-S0. This raises important questions regarding the onset of soil formation and the cessation of loess deposition in the region: When do soil-forming conditions regain dominance in the region during the Holocene? What does this imply for the timing of glacial–interglacial transitions in the region? According to the results presented here, the beginning of chernozem formation should be younger than the most recent estimated age (ORL8, 10.1 0.8 ka), implying that loess deposition continued well into the Holocene and that chernozem soil-forming conditions did not fully take hold until later in the Early Holocene. This interpretation is supported by very young IRSL ages (7.6 0.5 ka) from the lower part of modern soil V-S0 at Stari Slankamen (Schmidt et al., 2010) and ages of 10.0 1.1 ka from the last glacial loess V-L1 at 1.6 m below the lower boundary of the modern soil in the Rogulic´ gully on the Titel loess plateau (Bokhorst et al., 2011). It is also supported by a very young 14 C age (7.3 0.38 cal ka BP) for the uppermost part of the last glacial loess V-L1 at Surduk (Hatte et al., 2013). However, at Crvenka, quartz OSL dates from the Holocene soil and the boundary with last glacial loess unit V-L1 suggest that soil formation began at the onset of the Holocene (Stevens et al., 2011). It may be likely that many terrestrial records are characterized by a similar lag between Holocene global climate shifts and local or even regional environmental responses. For example, Lehmkuhl et al. (2103) have also reported an absence of soil formation during the Early Holocene at the Suohuduo loess–palaeosol located on the eastern margin of the Tibetan Plateau. Although variable between sites, based on the dates presented here, previous statements that the boundary between the Holocene soil and last glacial loess below it represents the exact time of the boundary between the Holocene and Late Pleistocene in globally integrated records such as global ice volume (e.g. Antoine et al., 2001; Rousseau et al., 2001; Markovic´ et al., 2005) may be problematic. The precise timing of Termination 1 does not necessarily occur coevally between loess and palaeosol units Copyright # 2014 John Wiley & Sons, Ltd.
197
as compared with marine oxygen-isotope stratigraphic transitions. This has major implications for understanding the terrestrial response to global climate changes. It implies a significant lag between ice volume or insolation forcing and environmental response on land, perhaps via delayed landscape stabilization or onset of more humid conditions. Potentially, sediment availability remained high during the earliest Holocene, leading to high dust accumulation rates and retarded soil formation. This may have been accentuated through Alpine glacial meltwater transport of material. Alternatively, atmospheric circulation did not change sufficiently to allow more humid conditions to prevail over the basin until later in the Holocene. Such delayed onset of full Holocene humid westerly flow is seen in arid zone records further east in Central Asia (Chen et al., 2008). High sampleresolution OSL dating has also shown this kind of asynchronous regional response on the Chinese Loess Plateau and it may be a feature of a large number of continental settings where the interaction of multiple atmospheric systems precludes linear responses to global climate forcing (Stevens et al., 2008). Regardless, these results call into question many basic assumptions underpinning the correlation of records worldwide and imply that the terrestrial environment in the large Eurasian loess belt shows complex responses to changing boundary conditions. More focused studies on the Holocene–Pleistocene boundary are needed to fully consider the timing of the transition to full interglacial conditions in this region as most of the studies showing this discrepancy have not dated the boundary at sufficiently high resolution. This is especially important because the dominant modes of air circulation are associated with different geographical latitudes which do not experience completely synchronous insolation forcing during the summer (e.g. Ruddiman, 2008). If a major global climatic shift such as Termination 1 does not dictate a uniform response of the terrestrial ecosystem, it is hard to imagine that climatic fluctuations of smaller magnitude would be characterized by worldwide synchronicity in environmental change. This situation may be typical for the investigated region, which was drier than ‘classical’ European loess provinces (Markovic´ et al., 2007, 2008; Bokhorst et al., 2009, 2011; Stevens et al., 2011; Hatte et al., 2013; Zech et al., 2013) where environmental responses to the last glacial climate variations have been close to the threshold between loess deposition and the initiation of pedogenesis. This explains why it is hard to distinguish differences between loess and initial pedogenetic layers in these sections and why over short distances the same stratigraphic subunits have quite different expressions.
Conclusions Multi-proxy investigations of the loess–palaeosol sequence at Orlovat demonstrate the importance of this site as a record of Late Pleistocene palaeoclimate and palaeoenvironment in Serbia. The Orlovat section provides an opportunity to reconstruct local and regional environmental and climatic conditions during the Late Pleistocene in a poorly studied region of the Carpathian Basin. Due to the slope depositional conditions at Orlovat, the preserved loess–palaeosol sequence also reveals a pattern of sedimentation that contrasts with other previously investigated northern Serbian loess sections, characterized by plateaulike deposition. Correlation with environmental magnetic records from other key loess sites in the Vojvodina region confirms the absence of at least parts of pedocomplex VL1S1. This is also supported by independent evidence from the luminescence chronology. Contrary to many previous J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
198
JOURNAL OF QUATERNARY SCIENCE
interpretations, the presented chronological results indicate the absence of significant pedogenesis during the earliest part of the Holocene in the Vojvodina region. Thus, the formation of the modern soil (V-S0) started with a delay of several thousands of years in comparison with previous expectations derived from analysis of globally integrated records. This interpretation raises an important question mark over the validity of generalized direct stratigraphic correlations between European loess records and global marine records and suggests a more complex response of loess terrestrial systems to global forcing. Finally, this study again confirms the importance of the simultaneous application of detailed exposure description, multi-proxy approaches and a detailed independently derived chronology for accurate temporal and spatial reconstruction of the evolution of environmental and depositional changes in complex terrestrial environments (2007, 2008; Vandenberghe, 2012; Vandenberghe et al., 2014) and that inadequate one-dimensional investigations can suggest misleading associations between global and regional processes. Acknowledgements. This research was financially supported by Project 176020 of the Serbian Ministry of Education and Science, as well as a grant (114-451-2262/2011) of the Provincial Secretariat for science and technological development of the Vojvodina Government, A.T-G. acknowledges the financial support from a grant of the Romanian National Authority for Scientific Research (CNCS-UEFISCDI, PN-II-RU-TE-2011-3-0062, no. 73/05.10.2011). F.L. wishes to thank the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) for funding in the frame of the Collaborative Research Center ‘Our way to Europe’ (CRC 806). Abbreviations. SAR, single-aliquot regenerative-dose; IR, infrared; OSL, optically stimulated luminescence; MIS, Marine Isotope Stage.
References Adamiec G, Aitken M. 1998. Dose-rate conversion factors: update. Ancient TL 16: 37–0. Aitken MJ. 1976. Thermoluminescent age evaluation and assessment of error limits: revised system. Archaeometry 18: 233–238. Aitken MJ, Alldred JC. 1972. The assessment of error limits in thermoluminescent dating. Archaeometry 14: 257–267. Antoine P, Rousseau DD, Fuchs M, et al. 2009. High resolution record of the last climatic cycle in the Southern Carpathian basin (Surduk, Vojvodina, Serbia). Quaternary International 198: 19–36. Antoine P, Rousseau DD, Zoller L, et al. 2001. High resolution record of the last interglacialglacial cycle in loess palaeosol sequences of Nussloch (Rhine Valley–Germany). Quaternary International 76/77: 211–229. Bokhorst MP, Beets CJ, Markovic´ SB, et al. 2009. Pedo-chemical climate proxies in Late Pleistocene Serbian-Ukrainian loess sequences. Quaternary International 198: 23–33. Bokhorst MP, Vandenberghe J, Su¨megi P, et al. 2011. Atmospheric circulation patterns in central and eastern Europe during the Weichselian Pleniglacial inferred from loess grain size records. Quaternary International 234: 62–74. Buggle B, Hambach U, Glaser B, et al. 2009. Magnetic susceptibility stratigraphy and spatial and temporal paleoclimatic trends in East European loess paleosol sequences. Quaternary International 196: 86–106. Chapot MS, Roberts HM, Duller GAT, et al. 2012. A comparison of natural and laboratory generated dose response curves for quartz optically stimulated luminescence signals from Chinese Loess. Radiation Measurements 47: 145–142. Chen FH, Yu ZC, Yang ML, et al. 2008. Holocene moisture evolution in arid central Asia and its out-of-phase relationship with Asian monsoon history. Quaternary Science Reviews 27: 351–364. Duller GAT. 2003. Distinguishing quartz and feldspar in single grain luminescence measurements. Radiation Measurements 37: 161– 165. Copyright # 2014 John Wiley & Sons, Ltd.
Evans ME, Heller F. 2001. Magnetism of loess/paleosol sequences: recent developments. Earth Science Reviews 54: 129–144. Fuchs M, Rousseau DD, Antoine P, et al. 2008. High resolution chronology of the upper Pleistocene loess/paleosol sequence at Surduk, Vojvodina, Serbia. Boreas 37: 66–73. Hatte C, Gauthier C, Rousseau DD, et al. 2013. Excursions to C4 vegetation recorded in the Upper Pleistocene loess of Surduk (Northern Serbia): an organic isotope geochemistry study. Climate of the Past 9: 1001–1014. Heller F, Evans ME. 1995. Loess magnetism. Reviews of Geophysics 33: 211–240. Kukla GJ. 1987. Loess stratigraphy in Central China. Quaternary Science Reviews 6: 191–219. Kukla GJ, An ZS. 1989. Loess stratigraphy in Central China. Palaeogeography, Palaeoclimatology, Palaeoecology 72: 203–225. Lehmkuhl F, Schulte P, Zhao H, et al. 2013. Timing and spatial distribution of loess and loess-like sediments in the mountain areas of the northeastern Tibetan Plateau. Catena. (In press. doi: 10.1016/j.catena.2013.06.008). Liu TS, et al. (unnamed) (eds). 1985. Loess and the Environment. China Ocean Press: Beijing. Lomax J, Fuchs M, Preusser F, et al. 2013. Luminescence based loess cronostratigraphy of the upper Paleolithic site in Krems-Wachtberg, Austria. Quaternary International. (In press. dx.doi.org/10.1016/j. quaint.2012.10.037). Lowick SE, Preusser F, Wintle AG. 2010. Investigating quartz optically stimulated luminescence dose–response curves at high doses. Radiation Measurements 45: 975–984. Markovic´ SB, Bokhorst MP, Vandenberghe J, et al. 2008. Late Pleistocene loess–paleosol sequences in the Vojvodina region, north Serbia. Journal of Quaternary Science 23: 73–84. Markovic´ SB, Hambach U, Catto N, et al. 2009. Middle and Late Pleistocene loess sequences at Batajnica, Vojvodina, Serbia. Quaternary International 198: 255–266. Markovic´ SB, Hambach U, Jovanovic´ M, et al. 2012. Loess in Vojvodina region (Northern Serbia): the missing link between European and Asian Pleistocene environments. Netherlands Journal of Geosciences 91: 173–188. Markovic´ SB, Hambach U, Stevens T, et al. 2011. The last million years recorded at the Stari Slankamen loess–palaeosol sequence: revised chronostratigraphy and long-term environmental trends. Quaternary Science Reviews 30: 1142–1154. Markovic´ SB, Kostic´ N, Oches EA. 2004. Paleosols in the Ruma loess section. Revista Mexicana de Ciencias Geologicas 21: 79–87. Markovic´ SB, McCoy WD, Oches EA, et al. 2005. Paleoclimate record in the Upper Pleistocene loess–paleosol sequence at Petrovaradin brickyard (Vojvodina, Serbia). Geologica Carpathica 56: 545–552. Markovic´ SB, Oches E, Su¨megi P, et al. 2006. An introduction to the Upper and Middle Pleistocene loess–paleosol sequence at Ruma brickyard, Vojvodina, Serbia. Quaternary International 149: 80–86. Markovic´ SB, Oches EA, McCoy WD, et al. 2007. Malacological and sedimentological evidence for ‘warm’ glacial climate from the Irig loess sequence, Vojvodina, Serbia. Geophysics, Geochemistry and Geosystems 8: Q09008. Martinson DG, Pisias NG, Hays JD, et al. 1987. Age dating and the orbital theory of the ice ages: development of a high resolution 0 to 300,000-year chronostratigraphy. Quaternary Research 27: 1–29. Mejdahl V. 1979. Thermoluminescence dating: beta-dose attenuation in quartz grains. Archaeometry 21: 61–72. Murray AS, Wintle AG. 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements 32: 57–73. Murray AS, Wintle AG. 2003. The single aliquot regenerative dose protocol: potential for improvements in reliability. Radiation Measurements 37: 377–381. Popov D, Markovic´ SB, Jovanovic´ M, et al. 2012a. Geomorphological investigations and GIS approach of the Tamis loess plateau, Banat region (Northern Serbia). Geographica Pannonica 16: 1–9. Popov D, Markovic´ SB, Strbac D. 2008. Generations of meanders in Serbian part of Tisa valley. Zbornik radova Geografskog instituta ‘Jovan Cvijic´’ SANU 58: 31–42. J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
THE ORLOVAT LOESS–PALAEOSOL SEQUENCE, NORTHERN SERBIA Popov D, Vandenberghe DAG, Markovic´ SB. 2012b. Luminescence dating of fluvial deposits in Vojvodina, N. Serbia: first results. Quaternary Geochronology 13: 42–51. Prescott JR, Hutton JT. 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23: 497–500. Rousseau DD, Gerasimaenko N, Matvviishina Z, et al. 2001. Late Pleistocene environments of the Central Ukraine. Quaternary Research 56: 349–356. Ruddiman WF. 2008. Earth’s Climate – Past and Future, 2nd edn. W. H. Freeman New York Schatz AK, Zech M, Buggle B, et al. 2011. The late Quaternary loess record of Tokaj, Hungary: Reconstructing palaeoenvironment, vegetation and climate using stable C and N isotopes and biomarkers. Quaternary International 240: 52–61. Schmidt E, Machalett B, Markovic´ SB, et al. 2010. Luminescence chronology of the upper part of the Stari Slankamen loess sequence (Vojvodina, Serbia). Quaternary Geochronology 5: 137–142. Stevens T, Armitrage SJ, Lu H, et al. 2006. Sedimentation and diagenesis of Chinese loess: implications for the preservation of continuous, high-resolution climate records. Geology 34: 849–852. Stevens T, Lu H, Thomas DSG, et al. 2008. Optical dating of abrupt shifted in the late Pleistocene East Asian monsoon. Geology 36: 415–418. Stevens T, Markovic´ SB, Zech M, et al. 2011. Dust deposition and climate in the Carpathian Basin over an independently dated last glacial–interglacial cycle. Quaternary Science Reviews 30: 662–681. Stevens T, Thomas DSG, Armitrage SJ, et al. 2007. Reinterpreting climate proxy records from the late Quaternary Chinese loess: A detailed OSL investigation. Earth Science Reviews 80: 111–136. Su¨megi P, Krolopp E. 2002. Quatermalacological analysis for modeling the Upper Weichselian palaeoenvironmental changes in the Carpathian Basin. Quaternary International 91: 53–63.
Copyright # 2014 John Wiley & Sons, Ltd.
199
Thompson LG, Goldstein SL. 2006. A radiometric calibration of the SPECMAP timescale. Quaternary Science Reviews 25: 3207– 3215. Thomsen KJ, Bøtter-Jensen L, Jain M, et al. 2008. Recent instrumental developments for trapped electron dosimetry. Radiation Measurements 43: 414–421. Timar-Gabor A, Vasiliniuc S¸, Vandenberghe DAG, et al. 2012. Investigations on the reliability of SAR-OSL equivalent doses obtained for quartz samples displaying dose response curves with more than one component. Radiation Measurements 47: 470– 475. Vandenberghe D, De Corte F, Buylaert JP, et al. 2008. On the internal radioactivity in quartz. Radiation Measurements 43: 771– 775. Vandenberghe J. 2012. Multi-proxy analysis: a reflection on essence and potential pitfalls. Netherlands Journal of Geosciences 91: 263– 269. Vandenberghe J, Hujizer B, Mucher H, et al. 1998. Short climatic oscillations in a western European loess sequence (Kesselt, Belgium). Journal of Quaternary Science 13: 35–38. Vandenberghe J, Markovic´ SB, Jovanovic´ M, et al. 2014. Site specific variations of loess and paleosols (Ruma, Vojvodina, northern Serbia). Quaternary International, doi: 10.1016/j.quaint. 2013.10.036. vanKreveld S, Sarnthein M, Erlenkeuser H, et al. 2000. Potential links between surging ice sheets, circulation change, and the Dansgaard–Oeschger cycles in the Irminger Sea, 60-18 kyr. Palaeoceanography 15: 425–442. Zech R, Zech M, Markovic´ SB, et al. 2013. Humid glacials, arid interglacials? Critical thoughts on pedogenesis and paleoclimate based on multi-proxy analyses of the loess–paleosol sequence Crvenka, Northern Serbia. Palaeogeography, Palaeoclimatology, Palaeoecology 387: 165–175.
J. Quaternary Sci., Vol. 29(2) 189–199 (2014)
