Introduction
Loess-paleosol sequences (LPSs) are important terrestrial archives for reconstructing Quaternary climate and landscape history (e.g., Zöller and Faust, Reference Zöller and Faust2009; Frechen, Reference Frechen2011; Marković et al., Reference Marković, Catto, Smalley and Zöller2011; Stevens et al., Reference Stevens, Marković, Zech, Hambach and Sümegi2011; Antoine et al., Reference Antoine, Rousseau, Degeai, Moine, Lagroix, Kreutzer and Fuchs2013; Kadereit et al., Reference Kadereit, Kind and Wagner2013; Meszner et al., Reference Meszner, Kreutzer, Fuchs and Faust2013). Although biomarkers have been proved for decades to be of outstanding value as molecular fossils in marine and lacustrine sediments (Eglinton and Eglinton, Reference Eglinton and Eglinton2008), they have been much less investigated in LPSs (for a review see, e.g., Zech et al., Reference Zech, Zech, Buggle and Zöller2011b). This has several reasons, among others, but primarily is because of the generally very low concentration of organic matter (OM) in loess.
Only during the last decade, plant leaf wax–derived lipids, particularly long-chain n-alkanes, have come into focus for organic geochemists using LPSs as paleoenvironmental/climate archives (Xie et al., Reference Xie, Wang, Wang, Chen and An2002, Reference Xie, Chen, Wang, Wang, Gu and Huang2003; Zhang et al., Reference Zhang, Zhao, Eglinton, Lu and Huang2006; Bai et al., Reference Bai, Fang, Nie, Wang and Wu2009). The popularity of n-alkanes as biomarkers can be attributed to the relatively easy procedure to extract and analyze them in the laboratory, as well as their recalcitrance and good preservation in loess. This allows results to be obtained even from loess samples with low organic content. Furthermore, n-alkanes can be used to infer vegetation changes in terms of grasses (dominated by the n-alkane homologues nC31 and nC33) versus deciduous trees (dominated by the n-alkane homologues nC27 and nC29) (Kirkels et al., Reference Kirkels, Jansen and Kalbitz2013; Schäfer et al., 2016b). It may be noteworthy that soil microbial degradation can affect n-alkane patterns and may need to be considered in certain cases (Zech et al., Reference Zech, Buggle, Leiber, Marković, Glaser, Hambach and Huwe2009; Buggle et al., Reference Buggle, Wiesenberg and Glaser2010; Nguyen Tu et al., Reference Nguyen Tu, Egasse, Zeller, Bardoux, Biron, Ponge, David and Derenne2011; Zech et al., Reference Zech, Krause, Meszner and Faust2013a; Schäfer et al., 2016a). Last but not least, the compound-specific hydrogen isotope (δ2H) signature of n-alkanes and its coupling with the compound-specific oxygen isotope (δ18O) signature of hemicellulose-derived sugar biomarkers are emerging as highly innovative paleoclimate tools in loess research (Liu and Huang, Reference Liu and Huang2005; Zech et al., 2013b, Reference Zech, Tuthorn, Detsch, Rozanski, Zech, Zöller, Zech and Glaser2013c).
A major and crucial assumption of all paleoreconstructions based on n-alkanes, as well as on any other biomarker in loess, is that the n-alkanes are deposited synsedimentarily and are not “contaminated” significantly by postdepositional processes. This issue, however, has been highly disputed for several years (Wiesenberg, Reference Wiesenberg2012 vs. Zech et al., Reference Zech, Kreutzer, Goslar, Meszner, Krause, Faust and Fuchs2012a; Wiesenberg and Gocke, Reference Wiesenberg and Gocke2013 vs. Zech et al., Reference Zech, Rass, Buggle, Löscher and Zöller2012b; Gocke and Wiesenberg, Reference Gocke and Wiesenberg2013 vs. Häggi et al., Reference Häggi, Zech, McIntyre, Zech and Eglinton2014). Gocke et al. (Reference Gocke, Kuzyakov and Wiesenberg2010, Reference Gocke, Kuzyakov and Wiesenberg2013, Reference Gocke, Peth and Wiesenberg2014) claimed that a significant overprinting of loess OM, including n-alkanes, occurred during the Holocene and still occurs in modern times by deep subsoil rooting and rhizomicrobial deposition even in loess distant to former roots. Given that the interpretation of n-alkane biomarker patterns alone is not straightforward for reconciling this controversy, comparative 14C and optically stimulated luminescence (OSL) dating can serve as an elegant alternative approach. Postdepositional root/rhizomicrobial contamination would result in a rejuvenation of the synsedimentary leaf wax–derived n-alkanes in loess. OSL, by comparison, yields (synsedimentary) sedimentation ages. Although 14C dating of n-alkanes has already been proved to be a powerful tool in soil science and marine sediments (Lichtfouse and Eglinton, Reference Lichtfouse and Eglinton1995; Huang et al., Reference Huang, Li, Bryant, Bol and Eglinton1999; Kusch et al., Reference Kusch, Rethemeyer, Schefuß and Mollenhauer2010), this method has until now rarely been applied in loess research. Only recently, Häggi et al. (Reference Häggi, Zech, McIntyre, Zech and Eglinton2014) performed compound-specific 14C dating of individual n-alkane homologues after gas chromatography (GC) separation on a preparative GC. They found the n-alkanes to yield 14C ages that agree well with the luminescence ages for the investigated LPS Crvenka in Serbia. However, those results are still awaiting verification from other sites, and compound-specific 14C dating is very laborious. This latter disadvantage may be overcome by 14C dating of purified bulk n-alkane fractions (Haas et al., Reference Haas, Bliedtner, Borodynkin, Salazar, Szidat, Eglinton and Zech2016).
The aim of this study was therefore to evaluate the potential of bulk, rather than compound-specific, n-alkane 14C dating for answering the question of whether n-alkane biomarkers in loess are primarily synsedimentary or whether they are significantly contaminated by root/rhizomicrobial OM. Bulk n-alkanes were extracted, purified, and dated for the Middle to Late Weichselian LPS Gleina in Saxony, Germany. The sedimentation ages were assessed by OSL dating; therefore the methodological limits of this approach need to be kept in mind (i.e., possible age under- or overestimation because of postdepositional sediment mixing or insufficient bleaching). We propose a 14C isotope mass balance calculation to estimate the potential postdepositional root/rhizomicrobial contamination quantitatively.
Material and methods
The LPS Gleina
Loess and paleosol samples were collected from the LPS Gleina (51°13.83N, 13°14.55E), which is located in the Saxon loess region, eastern Germany (Lieberoth, Reference Lieberoth1963; Haase et al., Reference Haase, Lieberoth and Ruske1970; Meszner et al., Reference Meszner, Fuchs and Faust2011). A detailed description of the stratigraphy was provided by Meszner et al. (Reference Meszner, Fuchs and Faust2011). In brief, the uppermost meter is composed of decalcified material representing the Holocene pedogenetic impact. Major parts of the Holocene soil (Luvisol) are truncated, and a plough horizon (Ap horizon) nowadays forms the topsoil. According to the pan-European loess stratigraphic model suggested by Marković et al. (Reference Marković, Stevens, Kukla, Hambach, Fitzsimmons, Gibbard and Buggle2015), this unit refers to S0. Below, down to approximately 8 m depth and referring to L1LL1 according to Marković et al. (Reference Marković, Stevens, Kukla, Hambach, Fitzsimmons, Gibbard and Buggle2015), six paleosols, which can be described as fossil gelic gleysols and fossil brown tundra gleysols, alternate with loess layers (Fig. 1A). From approximately 8 m on, there is strong evidence for accumulation of partly reworked material. In situ pedogenesis contributed to the formation of a fossil gelic gleysol at approximately 8.5 m depth and to the formation of the decalcified so-called Gleina complex (defined by Lieberoth [Reference Lieberoth1963] and Meszner et al. [Reference Meszner, Kreutzer, Fuchs and Faust2013] and referring to L1SS1 according to Marković et al. [Reference Marković, Stevens, Kukla, Hambach, Fitzsimmons, Gibbard and Buggle2015]) at approximately 10 m depth. The underlying so-called Lommatzsch soil complex (Lieberoth, Reference Lieberoth1963) was sampled from a nearby profile (approximately 20 m away) and is separated in Figure 1 by an unconformity.
Figure 1 (color online) (A) Stratigraphy of the loess-paleosol sequence Gleina. Red numbers on the stratigraphic column and filled black squares, respectively, mark the position of the optically stimulated luminescence (OSL) and n-alkane samples. (B) Age-depth profiles of calibrated 14C bulk n-alkane ages (95.4% probability ranges) and OSL ages with their 2σ error bars.
Modern root contamination partly occurred in the upper 4.5 m of the sequence, whereas hardly any modern roots were found in the lower part of the sequence. The risk of modern lateral root contamination was minimized by setting back the exposure wall by at least half a meter over the whole height of the exposure. Few rhizoliths, presumably documenting a Holocene root contamination, were found in the fossil gelic gleysol at 2 m depth.
n-Alkane preparation, radiocarbon dating, and quantification of postdepositional n-alkane contamination by roots
For radiocarbon dating of the bulk n-alkanes, four samples were selected from the Gleina LPS, for which n-alkane concentrations and patterns were already investigated and published (Zech et al., Reference Zech, Krause, Meszner and Faust2013a; cf. fig. 1 therein). The samples were chosen according to their n-alkane concentrations (sum of long-chain n-alkanes >3 μg/g) and their stratigraphic position. While the samples 15 and 19 represent a weakly developed gleysol and a loess unit, respectively, from the upper part of the Gleina section (with no signs of reworking), samples 21 and 23 represent a weakly and a strongly developed gleysol, respectively, from the lower reworked part of the Gleina section (Fig. 1). For comparison, two approximately 160 ka old samples from LPS Tumara, northeast Siberia (Zech et al., Reference Zech, Andreev, Zech, Müller, Hambach, Frechen and Zech2010), and a modern litter sample were analyzed as radiocarbon dead and modern reference materials, respectively (Table 1).
Table 1 Results of radiocarbon analyses for the purified bulk n-alkane fractions of reference materials and four investigated samples from the loess-paleosol sequence Gleina.
a Corrected for a vacuum line combustion blank of 0.4±0.1 μg C with an F14C of 0.7±0.2.
b Error propagation after Shah and Pearson (Reference Shah and Pearson2007).
c Calibration done with OxCal 4.2 and the IntCal13 calibration curve (Bronk Ramsey, Reference Bronk Ramsey2009; Reimer et al., Reference Reimer, Bard, Bayliss, Beck, Blackwell, Bronk Ramsey and Buck2013).
d Relative error >50%. Therefore, according to convention, minimum ages were not calculated using the corrected F14C value, but twice the absolute error.
The n-alkane preparation followed the procedure described by Häggi et al. (Reference Häggi, Zech, McIntyre, Zech and Eglinton2014). In brief, free lipids were obtained with a Dionex ASE 200 accelerated solvent extractor using dichloromethane and methanol. n-Alkanes were purified and eluted over aminopropyl columns. Subsequent further purification steps involved AgNO3 and zeolite (Geokleen) pipette columns (McDuffee et al., Reference McDuffee, Timothy, Sessions, Sylva, Wagner and Hayes2004) and served to obtain as clean bulk n-alkane fractions as possible (chromatograms are shown in Fig. 2). Radiocarbon measurements were conducted on an accelerator mass spectrometer (MICADAS, Ionplus) system at the Laboratory for Ion Beam Physics, ETH Zurich. Radiocarbon ages were converted to calendar ages using OxCal 4.2 (Bronk Ramsey, Reference Bronk Ramsey2009) and the IntCal13 calibration curve (Reimer et al., Reference Reimer, Bard, Bayliss, Beck, Blackwell, Bronk Ramsey and Buck2013). Calibrated 14C ages (cal ka BP) are provided as 95.4% probability range in Table 1 and Figure 1B. All bulk n-alkane radiocarbon data were corrected for a vacuum line combustion blank of 0.4±0.1 μg C with a fraction of modern carbon (F14C) of 0.7±0.2 (Table 1) using an isotope mass balance according to Shah and Pearson (Reference Shah and Pearson2007).
$${\rm F}^{{{\rm 14}}} {\rm C}_{{{\rm corrected}}} {\,\equals\,}{{\left( {\rm F^{{14}} C_{{uncorrected}} {\times}mass_{{sample}} (in\,\:\rmu gC){\,\minus\,}0.7{\times}0.4} \right)} \over {\rm mass_{{sample}} (in\,\rmu \rm gC){\,\minus\,}0.4}}$$
Figure 2 (color online) GC-FID (gas chromatography - flame ionisation detector) n-alkane chromatograms for the four radiocarbon-dated bulk n-alkane samples. The x-axes represent the retention time in minutes; the y-axes represent the GC-FID signal intensity in picoamperes. The long-chain n-alkanes nC29, nC31, and nC33 originate from plant leaf waxes; the nC31 dominance indicates dominant grass orgin; potential sources of short- and midchained n-alkanes nC18, nC20, nC21, and nC22 are charred and soil microbial biomass.
OSL dating
A numerical chronostratigraphy of the Gleina section was established by OSL dating using fine-grain (4–11 µm) quartz separates. Eight luminescence samples, generally from loess below or above paleosols and covering the depth from 2.9 to 10.8 m, were taken during nighttime and prepared in the laboratory for equivalent dose (D e ) determination using standard methods for fine-grain quartz dating (e.g., Fuchs et al., Reference Fuchs, Straub and Zöller2005). The OSL measurements were carried out on Risø TL/OSL DA-15 luminescence readers using blue light stimulation (470 Δ 30 nm). Signals were detected in the UV band (340 Δ 80 nm). For D e determination on 12 aliquots per sample, a standard single aliquot regenerative (SAR) protocol according to Murray and Wintle (Reference Murray and Wintle2000) was applied using the first 0.2s after subtracting the background from the last 4 s of the shine-down curves. SAR protocol parameters were previously deduced from test measurements, and the purity of the quartz extracts was tested by infrared light stimulation. The mean D e and its standard error for each sample were used for age calculation. For dose rate calculation, the U, Th, and K concentrations were measured using inductively coupled plasma mass spectrometry (K) and thick source alpha counting (e.g., Zöller and Pernicka, Reference Zöller and Pernicka1989). The α-effectiveness (a-value) was measured for each sample following the procedure described in Mauz et al. (Reference Mauz, Packman and Lang2006) and Lai et al. (Reference Lai, Zöller, Fuchs and Brückner2008) and resulted in a mean a-value of 0.04 with a standard deviation of 0.01 (n = 8; cf. Supplementary Table 1). The cosmic dose rate was calculated according to Prescott and Hutton (Reference Prescott and Hutton1994). A water content of 20%±5% was used. Further methodological details on the luminescence dating are given in the Supplementary Materials.
Results and discussion
n-Alkane patterns
The n-alkane chromatograms of the four 14C-dated n-alkane samples reveal a bimodal pattern (Fig. 2). The long-chain n-alkanes (>nC25) are characterized by a high odd-over-even predominance; that is, particularly the n-alkanes nC27, nC29, nC31, and nC33 are by far more abundant than the n-alkanes nC28, nC30, and nC32. Such patterns are typical for leaf wax–derived n-alkanes (Eglinton and Hamilton, Reference Eglinton and Hamilton1967; Kolattukudy, Reference Kolattukudy1976). The dominance of nC31 indicates that grasses were the main source (Maffei, Reference Maffei1996; Zech et al., Reference Zech, Krause, Meszner and Faust2013a; Schäfer et al., 2016b). The compounds nC18 and nC20 also occur in relatively high amounts, and in sample 15, additionally nC21 and nC22 (Fig. 2). While these n-alkanes typically do not or rarely occur in higher plant leaf waxes (Kuhn et al., Reference Kuhn, Krull, Bowater, Grice and Gleixner2010), Wiesenberg et al. (Reference Wiesenberg, Lehndorff and Schwark2009) showed that charring of grass biomass at 400°C to 500°C produces such n-alkane patterns. Soil microorganisms are reported to produce short- and midchained n-alkanes as well (Grimalt and Albaigés, Reference Grimalt and Albaigés1987; Buggle et al., Reference Buggle, Wiesenberg and Glaser2010; Zech et al., Reference Zech, Pedentchouk, Buggle, Leiber, Kalbitz, Markovic and Glaser2011a). We hence conclude that apart from plant leaf waxes, charred biomass and soil microbial biomass also likely contributed to the investigated n-alkane fractions with short-chained n-alkanes serving as respective molecular markers.
14 C ages of the bulk n-alkane samples
Radiocarbon dating of the bulk n-alkane fractions of the two approximately 160 ka old samples from the LPS Tumara (Zech et al., Reference Zech, Andreev, Zech, Müller, Hambach, Frechen and Zech2010) yielded very low blank-corrected F14C values of 0.0072 and 0.0025 (Table 1). However, because of the error propagation after Shah and Pearson (Reference Shah and Pearson2007), the absolute errors are relatively large (0.0145 and 0.0269, respectively, resulting in relative errors >50%). According to convention, minimum ages were therefore not calculated using the corrected F14C values, but twice the absolute errors (Table 1), which yielded 14C n-alkane minimum ages of >28.4 and >23.5 ka BP. For comparison, the modern litter sample yielded a corrected F14C value of 1.0234, corresponding to a 14C n-alkane age of −185±75 yr BP (Table 1). Thus, the 14C n-alkane results of the radiocarbon dead and modern reference materials corroborate the reliability of the bulk n-alkane 14C analyses performed here.
The bulk n-alkanes of the four dated loess-paleosol samples from the LPS Gleina yielded blank-corrected F14C values ranging from 0.028 to 0.058. This corresponds to calibrated 14C ages of 32.0±4.5 calkaBP, 27.2±3.6 calkaBP, 38.7±11.0 calkaBP, and 33.1±7.2 calkaBP for the samples 15, 19, 21, and 23, respectively. Although the midvalues suggest the occurrence of chronostratigraphic inconsistencies or age inversions (Fig. 1B), the large 95.4% probability ranges, which are calculated according to the error propagation method of Shah and Pearson (Reference Shah and Pearson2007) and which have to be taken into account, actually do not support this first impression.
OSL ages and establishment of a chronostratigraphy for the Gleina LPS
OSL fine-grain quartz ages and their D e values are shown in Table 2 and Figure 1. Age results are given as mean and 2σ standard error. All age values consistently increase with profile depth, and within error no age inversion occurred. The quartz luminescence signals were bright, fast decaying, and highly reproducible for the determined preheat and dose recovery tests. For all samples, no significant feldspar contamination (infrared-stimulated luminescence/OSL <1%) was detected. Consequently, the derived D e distributions show narrow scatters (c υ ~5%). The dose rate (Supplementary Materials) varies between 3.1±0.2 Gy/ka (BT836) and 3.5±0.2Gy/ka (BT842) showing typical values for loess deposits from this region (e.g., Kreutzer et al., Reference Kreutzer, Fuchs, Meszner and Faust2012).
Table 2 Optically stimulated luminescence quartz fine-grain age estimates.
Note: Ages given as mean with 2σ uncertainty.
For the uppermost part of the profile (~3 m up to ~8 m), ages between 24.5±2.7 ka (BT835) and 26.6±2.9 ka (BT839) indicate rapid eolian deposition during marine isotope stage 2 (MIS2; Lisiecki and Raymo, Reference Lisiecki and Raymo2005). These results are in accordance with previous findings from the Saxonian loess region (Meszner et al., Reference Meszner, Fuchs and Faust2011, Reference Meszner, Kreutzer, Fuchs and Faust2013; Kreutzer et al., Reference Kreutzer, Fuchs, Meszner and Faust2012). The onset of loess sedimentation at ~10 m was attributed to the MIS3 (BT842: 45.6±5.3 ka; BT840: 39.0±4.4 ka). Below a hiatus at ~10.2 m, the sediment was dated to the early Weichselian (72.8±8.1 ka BT844, MIS5a to MIS4). This hiatus has been observed for all investigated loess profiles in the Saxonian loess region. In contrast to other investigated loess profiles in Saxony (Ostrau, Seilitz: Kreutzer et al., Reference Kreutzer, Fuchs, Meszner and Faust2012; Meszner et al., Reference Meszner, Kreutzer, Fuchs and Faust2013), where the onset of loess sedimentation has been dated to ca. 30 ka (transition MIS3 to MIS2), the Gleina loess OSL ages reveal an onset of loess deposition earlier in MIS3. These findings may indicate locally favored loess accumulation and preservation because of its special topographic position on the northernmost boundary of loess distribution in the Saxonian loess region. However, although unlikely in our case, an age overestimation as a result of an incomplete signal resetting during secondary translocation processes cannot be fully excluded, because it is not possible to detect any incomplete signal resetting using fine-grain quartz dating. Although great care was taken during sampling, an age underestimation because of postdepositional mixing by bioturbation might add some uncertainty with regard to interpreting the luminescence ages. The previously described processes have opposite effects and might cancel each other out, and we consider them as being covered by the provided standard errors. Further details on the luminescence dating results are provided in the Supplementary Materials.
Comparison—toward a quantification of postdepositional n-alkane contamination
The comparison of the 14C bulk n-alkane ages with the OSL ages (Fig. 1B) shows that within error the bulk n-alkanes of the three uppermost samples 15, 19, and 21 are well within the range or even slightly older than the OSL-inferred sedimentation ages for the LPS Gleina. This finding contradicts the thesis of a significant postdepositional n-alkane contamination by roots/rhizomicrobial OM in loess. Especially, for sample 15 where midchain length (i.e., C20–C25) compounds represent a dominant fraction, this finding has further implications. It may indicate either that the deposited charred biomass is of synsedimentary age or that microbial contribution of midchain n-alkanes was also synsedimentary and did not affect the radiocarbon age.
Only the lowermost n-alkane sample 23 is slightly younger than the OSL ages (Fig. 1B). As mentioned previously, an OSL age overestimation because of solifluction, cryo- or bioturbation, and incomplete signal resetting (i.e., partial bleaching) cannot be fully excluded for these sediments. Nevertheless, assuming that all OSL ages reflect sedimentation ages, it could be argued that sample 23 was contaminated postdepositionally by root/rhizomicrobial-derived n-alkanes and is therefore too young. In order to quantify the postdepositional contamination, we estimated for this sample the synsedimentary F14C value based on the (here interpolated) OSL age (42.3±7.1 ka; Table 3) according to the following equation:
$${\rm F}^{{{\rm 14}}} {\rm C}_{{{\rm synsedimentary}}} {\equals}\left( {{1 \over 2}} \right)^{{{t \over T{{{1 \over 2}}}}}} $$
where t is the estimated uncalibrated 14C age (assuming that OSL age equals calibrated 14C age and using the intercept method and the IntCal13 calibration curve of Reimer et al. [Reference Reimer, Bard, Bayliss, Beck, Blackwell, Bronk Ramsey and Buck2013]) and T½ is the half-life time of radiocarbon (5730 yr). In the same way, we estimated F14C for contaminating roots (Table 3). On the one hand, we hypothesize a more or less modern (F14C value of 1.0) or last-decadal n-alkane contamination by roots/rhizomicrobial deposition. Levin et al. (Reference Levin, Kromer, Schoch-Fischer, Bruns, Münnich, Berdau, Vogel and Münnich1985) observed atmospheric F14C values of more than 1.2 from 1962 to 1985 and up to almost 2.0 F14C in the year 1963 because of nuclear weapon tests in the Northern Hemisphere. An F14C value of 1.2 can be used to estimate the contamination that occurred during the last decades. On the other hand, we hypothesize a Holocene root/rhizomicrobial contamination and calculate according to Eq. 2 the theoretical F14C values for three Holocene time points (3 ka, 6 ka, and 9 ka; see Table 3). This is based on the finding of Gocke et al. (Reference Gocke, Kuzyakov and Wiesenberg2010) and Pustovoytov and Terhorst (Reference Pustovoytov and Terhorst2004) that rhizoliths in last-glacial loess are of Holocene age.
Table 3 14C isotope mass balance calculation for quantifying potential postdepositional contamination with root/rhizomicrobial-derived n-alkanes.
a Assuming that optically stimulated luminescence (OSL) age equals calibrated 14C age and using the intercept method and the IntCal13 calibration curve (Reimer et al., Reference Reimer, Bard, Bayliss, Beck, Blackwell, Bronk Ramsey and Buck2013).
b Calculated according to Eq. 2.
The percentage of postdepositional n-alkane contamination “x” was then calculated using a 14C isotope mass balance approach according to the following equation:
$$ x\,\%\,\rm{\,\equals\,}{{F^{{14}} C_{{corrected}} {\,\minus\,}F^{{14}} C_{{synsedimentary}} } \over {F^{{14}} C_{{postsedimentary}} {\,\minus\,}F^{{14}} C_{{synsedimentary}} }}$$
Confidence intervals were calculated based on the F14C and OSL errors (Table 3). Accordingly, a modern and last-decadal root/rhizomicrobial n-alkane contamination of 3% (confidence interval 1%–5%) or a Holocene contamination of up to 9% (confidence interval 2%–14%) cannot be fully excluded for this particular sample (sample 23).
It should be mentioned that all radiocarbon ages are relatively close to their methodological age limit, and that the error bars are therefore relatively large. Nevertheless, our results lend further support to suggest that long-chain n-alkanes are not significantly contaminated by postdepositional OM. This is in agreement with findings from Häggi et al. (Reference Häggi, Zech, McIntyre, Zech and Eglinton2014) based on compound-specific 14C dating of n-alkanes (and long-chain n-alkanoic acids) in the LPS Crvenka, Serbia. Haas et al. (Reference Haas, Bliedtner, Borodynkin, Salazar, Szidat, Eglinton and Zech2016) also reported good agreement of 14C ages from bulk n-alkanes with luminescence ages for the LPS Kurtak, central Siberia. It may thus not be necessary in all cases to do compound-specific radiocarbon dating, which is very laborious and requires specialized equipment and large amounts of samples. Dating bulk n-alkanes (i.e., at the compound-class level) probably has great potential for future loess-paleosol research.
Conclusions
For the investigated LPS Gleina, three out of four 14C bulk n-alkane ages are well within the range or even slightly older than the OSL-inferred sedimentation ages. Only the fourth and lowermost sample may contain in absolute worst-case scenarios up to 5% of modern/last-decadal root-derived n-alkanes or up to 14% of Holocene root-derived n-alkanes based on 14C isotope mass balance calculations. Our results show that (i) n-alkane biomarkers at our study site are not significantly contaminated by depositional OM, (ii) short- and midchain n-alkanes also have a synsedimentary age, and (iii) 14C dating of leaf wax–derived bulk n-alkanes in loess-paleosols has great potential as a complementary dating tool for Quaternary research. More comparative bulk 14C n-alkane and OSL dating studies should be performed at other study sites, ideally also covering younger loess-paleosol deposits (i.e., not so close to the radiocarbon dating limits).
The results of our study suggest that in suitable LPSs, the contamination of older deposits with younger carbon is not a significant contributor to the carbon budget of the sediments. However, such an observation should be determined empirically for each study site to exclude the possibility of postdepositional movement of young carbon into older sediments.
Acknowledgments
We are very grateful for constructive discussions and comments from B. Buggle, E. Lehndorff, J. Rethemeyer, and two anonymous reviewers on an earlier version of this manuscript. Part of these comments including our replies is available in an online discussion forum at http://www.biogeosciences-discuss.net/bg-2012-310/#discussion. Furthermore, we greatly acknowledge the constructive reviews and feedback provided by T. Reimann, the Associate Editor M. Lachniet, the Senior Editor N. Lancaster, and an anonymous reviewer, which helped to improve this manuscript. Prof. B. Huwe and Prof. L. Zöller generously provided laboratory facilities. We thank M. Sroka for support of the laboratory work and A. Kadereit from the Luminescence Laboratory of Heidelberg for her cooperation and making more machine time available.
This study was funded by the German Research Foundation (DFG; FA 239/13-2 and ZE 844/1-1). The work of S. Kreutzer and M. Fuchs was gratefully funded by the DFG (FU 417/7-2). M. Zech and R. Zech also gratefully acknowledge the support provided by the Alexander von Humboldt-Foundation and the Swiss National Foundation (SNF 131670 and SNF 150590), respectively.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/qua.2016.7
