Quaternary International 296 (2013) 82e93
Contents lists available at SciVerse ScienceDirect
Quaternary International journal homepage: www.elsevier.com/locate/quaint
An environmental magnetic fingerprint of periglacial loess: Records of Late Pleistocene loessepalaeosol sequences from Eastern Germany Philipp Baumgart a, *, Ulrich Hambach b, Sascha Meszner a, Dominik Faust a a b
Chair of Physical Geography, Dresden University of Technology, D-01062 Dresden, Germany Chair of Geomorphology, University of Bayreuth, D-95440 Bayreuth, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Available online 25 December 2012
A detailed rock magnetic analysis of the last glacial/interglacial loess–palaeosol sequences was carried out in the framework of a comprehensive stratigraphic study of loess in Saxony (Eastern Germany). Magnetic susceptibility and laboratory-induced remanences have been determined to compare individual sections and to identify the specific rock magnetic characteristics of the Saxonian Loess Province. According to the model of pedogenic magnetic enhancement, an increasing neoformation of ferrimagnetic minerals in the course of pedogenesis was observed only in the uppermost Late Weichselian lithological units consisting of pure aeolian loess and indicating dryer climatic conditions. In contrast, the rock magnetic characteristics of the lower Middle and Early Weichselian units exhibit a significant destruction of primary magnetic minerals caused by such secondary processes as climatically controlled waterlogging and reworking. This change in magnetic composition with stratigraphic depth was additionally proved by factor analysis in which the main observation, an increasing cfd with decreasing c, argues for a general magnetic depletion in conjunction with decreasing magnetic grain sizes caused by weathering of larger primary particles. The magnetic fingerprint of the Saxonian loess is characterised by prevailing magnetic depletion processes, which effectively rules out the application of the wind vigour model. Moreover, the magnetic characteristics differ significantly from that of other loess regions and support a new and independent model, which explains the magnetic behaviour of strongly reworked and waterlogged (gleyed) loessepalaeosol sequences in the relatively humid Central European periglacial areas. Ó 2013 Elsevier Ltd and INQUA. All rights reserved.
1. Introduction Loess and loess-related sediments are sensitive recorders of the Pleistocene palaeoclimate and provide – almost globally distributed – detailed palaeoclimatic information. Especially in the relatively humid periglacial loess provinces, the last glacial/interglacial palaeosols and soil sediments provide evidence for high frequency climatic fluctuations on supra-millennial time scales. In addition to the established pedological and sedimentological analyses, studies of rock magnetic properties provide important contributions to palaeo-environmental and stratigraphical investigations in such archives. As a common element in the Earth’s crust and as an essential trace element in living organisms and various life processes, iron
* Corresponding author. Dresden University of Technology, Institute of Geography, Helmholtzstr. 10, 01069 Dresden, Germany. E-mail address:
[email protected] (P. Baumgart). 1040-6182/$ e see front matter Ó 2013 Elsevier Ltd and INQUA. All rights reserved. http://dx.doi.org/10.1016/j.quaint.2012.12.021
constitutes a major component of the environment. Moreover, the types and concentrations of iron compounds are indicative of pedogenesis due to their characteristic change in colour during redox processes. Natural iron oxides, hydroxides and sulphides exhibit remarkable properties in ambient magnetic fields, ranging from weak paramagnetic to strong ferromagnetic (sensu lato) effects, of which the latter is characterised by a memory effect known as magnetic remanence. Rock magnetic parameters are sensitive to the fate of iron-bearing minerals in the environment (Maher, 1986, 1998); they provide information on the occurrence and concentration of magnetic minerals and are indicative of magnetic grain size distributions. In addition to the recording of palaeoclimatic conditions, in a wider sense, they may reflect geomorphologic formation processes and landscape evolution from outcrop to landscape scale. Two established models can explain most of the rock magnetic characteristics of the loessepalaeosol sequences found around the world. The neoformation and enhancement of ferrimagnetic minerals in palaeosols in the course of pedogenesis was exemplarily demonstrated for loessepalaeosol sequences in the Chinese
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
Loess Plateau (CLP) and is valid in a wide range of environments, from the CLP in the east of the Eurasian continent via the Central Asia loess provinces to the Danube Basin in the west (Evans and Heller, 1994, 2003; Fitzsimmons et al., 2012; Markovic et al., 2012). In certain regions, however, the wind vigour model, primarily defined for Alaskan (Begét and Hawkins, 1989) and later for Siberian loessepalaeosol sequences (Chlachula et al., 1998), describes a higher amount of ferrimagnetic minerals in unaltered loess than in palaeosols. This phenomenon is explained by higher inputs of dense magnetic particles that were transported by stronger winds during glacial periods. Evans (2001) compared both models in detail and demonstrated their specific characteristics and possible interactions. Selected other studies exhibit distinct magnetic features not considered within those models. For instance, Matasova et al. (2001) compared two different Siberian loessepalaeosol sequences: the section at Bachat consisting of pure unweathered loess, and in contrast, the section at Kurtak built up from locally derived sediments and influenced by seasonal flooding, colluvial and solifluction processes. According to Zhu et al. (2003), these processes lead to degradation and destruction of primary magnetic minerals, resulting in magnetically depleted palaeosols. The present study demonstrates the specific magnetic record of the last glacial/interglacial Saxonian loessepalaeosol sequences and discusses a broader comprehension of the influence of secondary processes on the environmental magnetic signal, pointing out significant differences to the existing models. The specific aims of the current study are: (1) To compare the investigated sections based on the achieved magnetic data; (2) To contribute to the already completed stratigraphical investigations through the interpretation of the magnetic parameters linked to stratigraphical units (according to Meszner et al., 2011); (3) To identify specific magnetic characteristics of the Saxonian Loess Province.
83
1.1. Location and settings The study area is located in the northern part of the Saxonian Loess Province (Eastern Germany) at relatively low latitudes (52 N) west of the city of Dresden and the Elbe River (Fig. 1). Situated between the area of the Scandinavian and Alpine glaciations, the Weichselian loess (L1 loess) was extensively deposited in the foreland of the southern Erzgebirge Mountains, forming a plateau-like topography 150e250 m above sea level (a.s.l.). With a maximum thickness of 18 m at the northern loess boundary, the aeolian deposits of this province mainly consist of loess derivatives shaped by periglacial processes. Northward, relocation processes of sandy material dominated, wherefore the deposits pass over into the sandy and cryoturbated loessic sediments of the North German Plain. On its southern edge at 250e 300 m a.s.l, the loess is interfingered with periglacial slope deposits and shows only a thin and incomplete cover (Eissmann, 2002). Relatively low precipitation favoured the accumulation of loess in the Saxonian Loess Province whereas higher precipitation hindered the preservation of aeolian deposits in the Erzgebirge Mountains. In addition to this sedimentological northesouth gradient, the influence of a westerly wind component is also evidenced by grain size differences between the left and right bank deposits of the Elbe River. The modern-day landscape of gently rolling hills was shaped by small rivers that incised the Weichselian loess cover down to the bedrock at the end of the Weichselian glacial period. During the Holocene, the present-day flood plains were formed by the deposition of fine-grained alluvial sediments. This region has been the subject of intense scientific research since 1950. A first stratigraphic effort was proposed by Lieberoth (1959) and supplemented by Haase (1963) and Richter et al. (1970). Recent investigations by Meszner et al. (in press) present a new composite stratigraphy, adding previously unrecognised palaeosols and new stratigraphic marker horizons. The sites of Gleina, Ostrau and Seilitz presented in this paper have been described in detail by Meszner et al. (2011), who recognised and
Fig. 1. Location of the investigated loessepalaeosol sequences of Gleina, Ostrau and Seilitz in the study area and the general distribution of loess (>3 m thickness) in Saxony. The Zeuchfeld (Saxony-Anhalt) section is situated southwest of the city of Leipzig.
84
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
defined five stratigraphical units. Short intervals of the section at Zeuchfeld (Saxony-Anhalt, 40 km southwest of Leipzig, GK: R 4486052; H 5677220) were additionally investigated, and the results have been used for comparison. A brief lithological and palaeopedological description is given below and in Table 1. For details, please refer to Meszner et al. (in press).
2. Methods 2.1. Sampling and laboratory analyses Before sampling, all sections were carefully cleaned. In total, 483 non-oriented specimens (Gleina: 197, Ostrau: 191, Zeuchfeld: 55,
Table 1 Brief lithological and palaeopedological description of the Saxonian loess deposits based on the composite profile from Meszner et al. (2011). Stratigraphical units IeV, main loess layers and palaeosols are listed. Stratigraphical unit
Main loess layers and palaeosols
I
S0
IIa IIb III
IV
Ostrau
LIIa NB LIIb fBvc LIII fG
Btv LFZ Bv Loess NB Loess br. NB (fBvc) Loess fG
Bt LFZ Bv Loess NB Loess br. NB(fBv) Loess e
LIV HZ
ABS Loess HZ
e Loess HZ
S1
fS(e)w fBtSd
fS(e)w fSd/Bt
IVw
IVs V
Gleina
Lithological and palaeopedological description
Holocene soil, decalcified, developed as a Luvisol (“Parabraunerde”) Horizon with a lenticular structure (“Lamellenfleckenzone”) Light brown horizon, decalcified Loess layer, calcified, homogeneous structure, includes 2 palaeosols with weak Gelic Gleysol features Strong greyish Gelic Gleysol (“Nassboden”), increased humic content Loess layer consisting of loess derivatives, calcified, laminated, weakly modified, small frost cracks Weak double Cambisol (“Braunerde”), congeneric to NB Loess layer consisting of loess derivatives, calcified, strongly reworked Grey strong hydromorphic solifluction layer, Stagnic Gleysol, decalcified, upper part of “Gleinaer Boden-Komplex” (GBK) Arctic Brown Soil sediment, formed by reworked material from S1, lower part of GBK Loess layer consisting of loess derivatives, slightly calcified, strongly reworked Humic horizon, enriched humic content (“Humuszone”), typical for several European loess sections (e.g., Rhine catchment area), upper part of “Lommatzscher Boden-Komplex” (LBK) Grey bleached soil sediment, earliest Weichselian deposit, part of LBK Relict of Eemian interglacial soil, decalcified, developed as a Luvisol with strong hydromorphic features (“Pseudogley”) on Late Saalian sediments, lower part of LBK Glaciofluvial sandy sediments, Late Saalian
In all sections, unit I represents the decalcified Holocene soil (S0), which is developed as a Luvisol (“Parabraunerde”, Bt-horizon) and is followed downwards by a horizon revealing a lenticular structure (“Lamellenfleckenzone”, LFZ) and a brownish pedohorizon (Bv). Unit II extends over two loess layers (LIIa, LIIb) in which four palaeosols are intercalated. The upper two palaeosols show only weak Gelic Gleysol features in contrast to the preserved soil at the base of unit IIa, which is developed as a strong and greyish Gelic Gleysol (“Nassboden”, NB) with an increased humic content. A weak interstadial double Cambisol (“Braunerde”, fBvchorizon) follows within the LIIb loess layer. The base of unit III is formed by a grey and strong hydromorphic solifluction layer (fG), which is interpreted as the upper part of an interstadial soil complex, the so-called “Gleinaer Boden-Komplex” (GBK) according to Lieberoth (1963). It is assumed that this formation took place under cool and moist climatic conditions. A sediment derived from an Arctic Brown Soil beneath (ABS, unit IV) represents the lower part of the GBK. This sediment is formed by reworked material derived from the Eemian interglacial soil. At the Zeuchfeld section, the local formation of GBK can be parallelised to the “Kösener Verlehmungszone” (Ruske and Wünsche, 1961; Wansa, 2009), which is developed in a gravel matrix. A horizon revealing an enriched humic content (“Humuszone”, HZ) has been found in unit IV as well. Such a horizon is frequently observed in several Early Weichselian European loess sections, e.g., in the Rhine catchment area (Semmel, 1997). The HZ is followed by a grey and bleached soil sediment (fS(e)w, fossil stagnic horizon) that is regarded as the earliest Weichselian deposit in these investigated sections. Below the Weichselian sediments, remnants of the Eemian interglacial soil (MIS 5e, S1) are preserved in unit IV as an in-situ Bt-horizon with strong hydromorphic features (“Pseudogley”, fBtSd horizon). The sequence of the humic horizon, the soil sediment and the remnants of the Eemian interglacial soil has been previously defined by Lieberoth (1963) as the “Lommatzscher Boden-Komplex” (LBK). In most cases, unit V represents glaciofluvial sandy sediments of Saalian age (MIS 6).
Seilitz: 40) were taken for rock magnetic analyses during the years 2009 and 2010. The high-resolution sampling of the loesse palaeosol sequences in the main profiles of Gleina (6 cm spacing) and Ostrau (5 cm spacing) was carried out to obtain a detailed view and a precise separation of the loess layers and intercalated palaeosols. The underlying remnants of the Eemian palaeosol were sampled with a resolution of 20 cm, as well as the entire section at Seilitz. Discrete intervals of the Zeuchfeld section were sampled with a resolution of 5 cm. The rock magnetic measurements were performed at the Laboratory for Palaeo- and Environmental Magnetism, Chair of Geomorphology at the University of Bayreuth. For laboratory analyses, the dried material was carefully ground to a homogeneous structure and compressed into plastic boxes (6.4 cm3). The parameter of the initial low field magnetic susceptibility (c) provides information on the amount of magnetically effective material in a sample as well as the magnetic grain size distribution (GSD), if measured at different frequencies. The value of c was determined using the AGICO KLY-3 Kappa-Bridge (sensitivity greater than 5 107 SI) operating at an AC field of 300 Am1 and at a frequency of 0.92 kHz. At the Ostrau section, the initial low field magnetic susceptibility was additionally determined in-situ in 10 cm steps by means of a portable Bartington MS2 susceptibility meter operating at 0.58 kHz (MS2F-sensor). Four repetitive measurements were taken at each level and subsequently averaged. For measurement of the frequency dependent magnetic susceptibility (cfd) each specimen was measured twice at two different frequencies (0.3 and 3 kHz) in a magnetic AC field of 300 Am1 using a MAGNON VFSM susceptibility bridge (sensitivity greater than 5 106 SI). For imprinting of the anhysteretic remanent magnetisation (ARM), the samples were exposed to a magnetic alternating field (AF) of 100 mT and a bias DC field of 50 mT. The acquired remanence was measured by employing a JR-6A AGICO spinner magnetometer. A MAGNON AF Demagnetiser 300 was used for demagnetisation. The isothermal remanent magnetisation (IRM) was imprinted in fields of 2000 mT and 350 mT (back field) using a MAGNON Pulse
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
Magnetiser II. During all remanence analyses, a constant temporal offset between inducement and measurement was maintained to avoid differences in the acquisition of viscous remanent magnetisations (VRM). 2.2. Methodological background and data processing The relevant ferrimagnetic minerals in loesses and palaeosols are iron oxides and sulphides such as magnetite, titanomagnetite, maghemite and greigite (e.g., Stanjek et al., 1994; Maher, 1998), which show high c-values. Others, like haematite, goethite, ilmenite and pyrite, show only a weak contribution to the susceptibility. In addition to the concentration of ferrimagnetic minerals, the magnetic GSD also plays an important role for susceptibility. Maher (1986, 1988) found for fine-grained superparamagnetic magnetite (SP < w0.03 mm) 2e3 times higher c-values as for single domains, pseudo-single domains (SD, PSD w0.03e10 mm) and multi domains (MD > w10 mm). With knowledge of the mass of each specimen, a mass specific susceptibility was calculated (c in m3 kg1) in addition to the volume specific susceptibility (k in SI). Measurement of frequency dependent magnetic susceptibility (cfd in %) is sensitive to the presence and relative contribution of the SP particles in a sample (Maher, 1986). The effect of blocking SP grains in the SP/SD transition zone and shifting that boundary towards smaller domain sizes at higher frequencies results in lower c-values compared with low-frequency measurements (Heller et al., 1991). For that reason and because of the weak influence of MD particles on cfd, this parameter is often used as a proxy for pedogenetic formed SP particles (Liu et al., 2007). The inter-parametric ratios of laboratory-induced remanences, including their normalisation by susceptibility, are highly sensitive to magnetic grain size variations (e.g., Evans and Heller, 2003). Therefore, laboratory-induced magnetisations have been determined to obtain more detailed information on the occurrence and relative abundance of the various magnetic mineral types in these samples. Anhysteretic remanent magnetisation (ARM mass normalised in Am2 kg1) is an indicator for magnetic GSD, especially for SD and small MD grains, as well as a parameter for the concentration of soft ferrimagnetic minerals (Maher, 1986). Remanent magnetisations induced within a short time and under strong magnetic fields at room temperature are known as isothermal remanent magnetisations (IRM mass normalised in Am2 kg1). The IRM experiments are suitable for discovering further characteristics of the remanence-carrying magnetic phases. Thus, the IRM reflects the concentration of ferromagnetic SD and MD particles. Furthermore, the IRM can be used as a parameter for the relative amount of such high-coercive (hard) magnetic minerals as the antiferromagnetic haematite as well as for such low-coercive (soft) magnetic minerals as the ferrimagnetic magnetite, whose remanences saturate at very different fields. This correlation is shown by the so-called s-ratio (e.g., Maher, 1986). The following formulas were required to describe the cfd and the s-ratio:
cfd% ¼ ððc0:3 kHz c3 kHz Þ=c0:3 kHz Þ 100½%
(1)
s-ratio ¼ ððIRM350 mT =IRM2000 mT Þ þ 1Þ=2
(2)
For further investigations, combinations of previously presented parameters were employed (particularly ARM/c, IRM2000/c and ARM/IRM2000) that do not depend on concentration changes at all but are indicative for magnetic GSD. The observed high correlation between these parameters was derived from factor analysis as well. Factor analysis was performed with the PASW Statistics 18 software
85
using the extraction method of principal component analysis. This mathematical procedure enables the user to reduce the complexity of measured parameters to a set of linear and assessable principal components (factors), which allows further interpretations of the measured data. The ARM/c-ratio provides information on the composition and GSD of the magnetic mineralogies dominated by magnetically soft minerals (e.g., King et al., 1982; Hambach, 2010); it is sensitive to variations in the SD fraction to the SP fraction. The IRM2000/c mainly reflects a ratio of the SDeMD fraction to the SP fraction (Zhou et al., 1990; Evans and Heller, 2003) and also indicates the composition of the magnetic minerals (e.g., a low ratio indicates a higher presence of paramagnetic minerals). The ARM/ IRM2000 detects variations of the SD fraction to the SDeMD fraction (Oldfield, 1991). An increasing concentration of SD particles during pedogenesis results in a higher value of the ARM/IRM2000 ratio (van Velzen and Dekkers, 1999; Liu et al., 2007). 3. Results For each profile, the individual magnetic parameters were plotted as a function of depth. Significant peak values and specific changes in the magnetic parameters observed in all sections in the same lithostratigraphical units demonstrate the remarkable potential of rock magnetic measurements for identification of stratigraphical units and palaeosols. 3.1. Gleina section The section at Gleina (Fig. 2, Table 1) represents one of the thickest and most complete Weichselian loessepalaeosol sequences in the Saxonian Loess Province. The previously described interstadial soil complex “Gleinaer Boden-Komplex” (GBK) was named after that locality. High susceptibilities and remanences can be observed in the Holocene soil S0, reaching maximum values of c (472 108 m3 kg1) and ARM (439 106 Am2 kg1) in the uppermost two samples. This observation can be interpreted as a near-surface anthropogenic influence; however, the reason for this influence could not yet be clarified. Moreover, a significant enhancement of magnetically effective material in the S0 is obvious and is underlined by the relatively high values of cfd (6%) and the low values of IRM2000/c (6 kAm1), indicating a higher amount of SP particles. Furthermore, the s-ratio (0.94) shows the dominance of lowcoercive (soft) magnetic minerals (magnetite, maghemite), which are characterised by high susceptibilities and remanences. A similar but notably weaker effect can still be observed in the palaeosols NB (depth of 3 m) and fBvc (6 m) in which the local peaks are weak, and in contrast to S0, the cfd-values are low. A remarkable change is clearly visible beneath the NB (3 m) by an increasing and more heterogeneous course of cfd, as well as by a decreasing c. Unit III shows significant reductions of susceptibility and remanences (cmin ¼ 10 108 m3 kg1; ARMmin ¼ 4 106 Am2 kg1) apart from high cfd-values (up to 7.6%) and an increasing ARM/IRM2000 ratio, indicating a higher amount of SD particles in the SDeMD fraction. The solifluction layer fG (10 m) at the base of that unit offers the most extreme values, whereby the HZ of unit IV displays nearly the same features. Compared with the Holocene S0, the remnants of the Eemian interglacial soil S1 are characterised by low c-values m3 kg1), s-ratios (0.84), remanences (15 108 6 2 1 (ARM ¼ 9 10 Am kg ) and a distinct minimum in cfd (0.9%). 3.2. Ostrau section Similar to the Gleina section, the S0 in the Ostrau profile (Fig. 3, Table 1) reveals high values of susceptibility and laboratory-induced
Fig. 2. Rock magnetic results from the Gleina section. The following parameters are given: initial low field magnetic susceptibility c (108 m3 kg1); frequency dependent magnetic susceptibility cfd (%); s-ratio (dimensionless); anhysteretic remanent magnetisation ARM (106 Am2 kg1); ratio of isothermal remanent magnetisation IRM2000 to c (kAm1); ratio of ARM to IRM2000 (dimensionless). Stratigraphical units IeV are highlighted. Note that the high anthropogenic-influenced values of the uppermost two samples are not presented because of unfavourable displacement of the graphs (e.g., cmax ¼ 472 108 m3 kg1, ARMmax ¼ 439 106 Am2 kg1).
Fig. 3. Profile of Ostrau with rock magnetic results and highlighted stratigraphical units.
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
remanences caused by the enhancement of ferrimagnetic minerals. In addition to the typical change in magnetic behaviour below the NB (5 m), weak local maxima in the palaeosols NB and fBv (7 m) can be observed, similar to those of the Gleina section. However, in contrast to Gleina, the cfd in unit IIb generally decreases, and unit III exhibits lower values of susceptibility and remanences, especially in the lower part. Note that the fG could not be recognised in the Ostrau section in which no significant correlation to that horizon at the Gleina section is possible. However, the HZ (11 m) of unit IV offers identical magnetic properties as those of Gleina with typical low c values (down to 9 108 m3 kg1), high cfd-values (6.7%) and a high ARM/IRM2000 ratio. The interglacial pedo-horizon S1 (12 m) displays more extreme features than in the Gleina section: low c values (8 108 m3 kg1), low cfd-values (2.4%), low s-ratios (0.84) and low remanences (ARM ¼ 4 106 Am2 kg1) but evidently high values of ARM/IRM2000-ratio. Fig. 4 exhibits a correlation between the field- and laboratorymeasured initial low field magnetic susceptibility at the Ostrau section. The graphs correlate quite well and despite the use of two independent measuring systems with different sensitivities, both results are able to detect nearly the same variations in c necessary for the recognition of stratigraphical units. 3.3. Seilitz and Zeuchfeld sections Investigations at the Seilitz and Zeuchfeld sections were carried out to verify the magnetic trends identified in the main profiles of the Gleina and Ostrau sections. Therefore, the sections were sampled at lower resolution (20 cm), and the obtained results are presented in Fig. 5. At the Seilitz section, the same magnetic features can be observed: high c, cfd, s-ratio and remanences in the Holocene S0, and in addition to the characteristic change beneath
Fig. 4. Two independent measurements of magnetic susceptibility k (dimensionless, 106 SI) at the Ostrau section as a function of stratigraphy (stratigraphic units IeIV are highlighted). The field data (Bartington MS2 susceptibility meter) show a good correlation with the data measured in the laboratory (AGICO KLY-3 Kappa-Bridge). Trends in k, specific for each stratigraphical unit, can be reconstructed notably well with both methods.
87
unit IIa, weak local maxima in the interstadial palaeosols NB and fBvc. Thus, the c decreases in unit IIb whereas the cfd increases at the same level as in the Gleina section. Furthermore, unit III exhibits the previously noted lower susceptibilities and remanences, whereas the fG-horizon of unit III, unit IV and the underlying unit V are not developed at this site. At the Zeuchfeld (Saxony-Anhalt) section, the remarkable GBK and the underlying Early Weichselian loess sediments of unit IV were sampled only to correlate these outstanding layers with the Saxonian profiles. The upper part of the GBK, unit III, shows the same significant minima in c, s-ratio and ARM as in the fG of the Gleina section as well as high cfd values. A moderate amount of SD particles can be observed in the SDeMD fraction, as indicated by a lower ARM/IRM2000-ratio. The underlying part of the GBK at the top of unit IV offers the same magnetic features as the Arctic Brown Soil sediment beneath the fG in the Gleina section. 3.4. Combination of magnetic parameters and factor analysis Fig. 5 compares the summarised magnetic properties as a function of stratigraphical position. Several combinations of the magnetic parameters are displayed in Fig. 6. First, each single stratigraphical unit is grouped in all presented sites. Moreover, the data cloud of each unit shows the identical graphical positions for all sections and documents their remarkable similarity and comparability. Hence, the unit-dependent differences in the magnetic features are obvious. For example, units I and IIa indicate a correlation of cfd/c: the higher the cfd, the higher the c. With increasing depth, in addition to a wider scattering of the values, this correlation changes to the opposite trend: the higher the cfd, the lower the c. Furthermore, a relationship between the magnetic GSD and the type of magnetic mineral is demonstrated in the cross-plot of the IRM2000/c to s-ratio. Consequently, the Holocene soil and the loess of unit IIa are characterised by the dominance of low-coercive minerals of the magnetiteemaghemite group in addition to a higher amount of particles in the SDeMD fraction. The lower units display significantly high values of cfd and lower values of IRM2000/c, which indicate an increasing amount of SP particles and the dominance of hard magnetic minerals, most likely haematite and goethite. In this figure, horizons fG (unit III), HZ and S1 (unit IV) offer the most extreme values. The selected combinations of parameters have been derived supplementarily from factor analysis using the obvious high correlations among them (compare Fig. 7). The extraction method of the principal component analysis, and the rotation method varimax with Kaiser Normalisation, were determined for the statistical evaluation. The rotation converged in three iterations. The results of the factor analysis are presented in Fig. 8 in which the highly correlated parameters of c, ARM, IRM2000 and a weakly correlated s-ratio are represented by factor 1. Factor 2 mainly reflects cfd. Together, both factors explain more than 86% of the observed variance (factor 1: 63.1%; factor 2: 23.1%). Positive values indicate a strong correlation and negative values indicate a weak correlation of the measured properties with the respective factor. Consequently, units I and IIa are well represented by factor 1, but with increasing depth, factor 2 plays a more important role. Several parts of units IIb and III are not explained at all, neither by factor 1 nor by factor 2 due to negative values in both correlations. In addition to the significant separation of units through the derived factors, the previously described change in magnetic behaviour beneath unit IIa is also clearly demonstrated by the factor analysis. Unit III, especially the fG and several parts of unit IV and the HZ, indicate the highest correlations with factor 2. The remnants of the Eemian interglacial soil (unit IV) are described by a weak correlation with factor 2 only.
Fig. 5. Comparison of rock magnetic parameters of all investigated sections classified according to stratigraphical units with main loess layers and palaeosols (arithmetic means are given). S0 ¼ Holocene Bt-horizon; LIIa ¼ loess layer unit IIa; NB ¼ interstadial strong Gelic Gleysol (“Nassboden”); LIIb ¼ loess layer unit IIb; fBvc ¼ interstadial Cambisol; LIII ¼ loess layer unit III; fG ¼ strong hydromorphic solifluction layer of interstadial soil complex GBK; LIV ¼ loess layer unit IV; HZ ¼ interstadial humic horizon; S1 ¼ relict of Eemian interglacial soil (fBtSd-horizon, MIS 5e).
Fig. 6. Cross-plots of rock magnetic parameters showing clear separations of units and demonstrating the remarkable similarity of all investigated sections. Furthermore, a change in magnetic behaviour beneath unit IIa (c w30 108 m3 kg1; s-ratio w0.92) is clearly visible. High values of cfd indicate a growing relative amount of SP particles in a sample and are a proxy for increasing soil moisture. The IRM2000/c reflects the ratio of the SDeMD fraction to the SP fraction, whereby a high s-ratio indicates a dominance of low-coercive minerals of the magnetiteemaghemite group.
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
89
Fig. 8. Summarised results of factor analysis from all investigated sections. Stratigraphical units IeV are indicated. Factor 1 represents the highly correlated parameters of c, ARM, and IRM2000, including a weak correlated s-ratio. Factor 2 mainly reflects cfd. Positive values indicate a strong correlation and negative values indicated a weak correlation of the measured properties with the respective factor. Consequently, units I and IIa are well represented by factor 1, but factor 2 plays a more important role with increasing depth.
Fig. 7. Strong correlations of the measured initial low field magnetic susceptibilities with the laboratory-induced ARMs are obvious in all sections proving the extreme low absolute amount of SP particles, which are in turn an inevitable element of the classic magnetic enhancement model.
3.5. Correlation of Fet and Fed with c For the Weichselian loessepalaeosol sequence at the Gleina section, there is a weak correlation (R2 ¼ 0.149) between the total iron content (Fet) and the magnetic susceptibility (Fig. 9), similar to the results of investigations from a loessepalaeosol sequence in Luochuan (CLP) described by Evans and Heller (2003). According to them, the paramagnetic minerals and weak magnetic phases formed during oxidation processes mask a clear correlation between these parameters. The present study did not detect increasing correlations (R2 ¼ 0.028) by individual consideration of dithionite soluble iron (Fed). Other authors also describe only weak Fet/c e correlations, e.g., for Chinese loess and palaeosol samples (Fine et al., 1995) and for recent soils in England (Dearing et al., 1996). The interglacial soils S0 (Holocene) and S1 (Eemian) contain maximum amounts of Fet but on rather different susceptibility levels. The interstadial GBK, consisting of the solifluction layer fG and the Arctic Brown Soil sediment ABS, shows higher Fet-values and low susceptibilities as well, whereas the HZ and the strong hydromorphic bleached fS(e)w-horizon above the remnants
Fig. 9. Gleina section: Correlation of total iron Fet and dithionite soluble iron content Fed (both in g kg1) to c (108 m3 kg1) with highlighted palaeosols. No significant correlation is observed.
90
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
of the Eeman interglacial soil display the lowest Fet contents. Similar to the interglacials with their maxima in Fet and Fed, the ABS horizon also shows comparable Fed and Fet contents, which indicates an intense pedogenesis. Contrary to this observation, the fG displays the lowest Fed amounts in this section and emphasises the differences in development of the fG and the ABS. 4. Discussion and interpretation The presented results show almost identical magnetic characteristics in all investigated sections, demonstrating the general suitability of rock magnetic investigations for correlation of different profiles over short distances. Moreover, the established environmental magnetic record proves the existing pedo- and litho-stratigraphy by retracing the previously defined stratigraphical units with an environmental magnetic approach. Thus, the changes in the rock magnetic-dependent parameters can be correlated with the changes in the pedological and sedimentological parameters. Consequently, the pedostratigraphical units IeV can also be defined as environmental magnetic units. The magnetic properties of all investigated sections are stratigraphically summarised in Fig. 10. Because the Saxonian loess deposits and loess derivatives are stratigraphically non-uniform, they deserve a differentiated view. Units I and IIa primarily consist of pure aeolian loess with a homogeneous structure. This characterisation is indicated by the cfd/c-cross-plots (Fig. 6) in which the uppermost two units show the typical magnetic behaviour (the higher the cfd, the higher the c) of unaltered loess-deposits with intercalated palaeosols. Thus, the record displays a significant enhancement of the magnetically active material in the Holocene soil following the classical magnetic enhancement model. On the one hand, a high concentration of superparamagnetic particles causes this high susceptibility signal. On the other hand, the s-ratio and the ARM indicate the presence of low-coercive magnetic minerals of the magnetiteemaghemite group, which own high susceptibilities as well and are detritic or pedogenic in origin (Maher, 1998). In other words, the dominant process in the Holocene soil is the magnetic enhancement as a result of pedogenesis in a pre-existing substrate, as documented by the higher c-values in palaeosols compared with unaltered parent loess. The intercalated palaeosol NB (unit IIa) shows similar magnetic enhancements but on a distinctly lower level compared with the Holocene S0.
However, the lower units, which demonstrate the opposite trend, are characterised by partly reworked and hydromorphic loess layers and palaeosols. The loess in unit IIb becomes more laminated and reworked. This feature might be caused by shortrange dislocation processes that occurred after the deposition of loess. Rousseau and Kukla (1994) describe the same characteristics for the Eustis Loess Section in Nebraska in which they assume a niveoaeolian deposition or frequent oscillations in the local microclimate as the reason for this lamination. Following these observations, unit II was separated into two subdivisions at the lower edge of the NB. Furthermore, the results of the factor analysis also showed this change in magnetic behaviour by an increasing importance of factor 2 with increasing depth. The cfd, however, can be used as a proxy for hydromorphy and reworking in the investigation area. Thus, the formation of the magnetic characteristics of the reworked and hydromorphic lower units cannot be explained by magnetic parameters alone, in contrast to the upper units I and IIa that are dominated by primary loess. Unit III, which consists of loess derivatives, is marked by strong reworking processes culminating in clear features of solifluction at the base (fG). This solifluction layer represents a distinctive discordance in the Saxonian loess and is tentatively correlated to the Denekamp Interstadial Complex. The fG is characterised by significantly low values of susceptibility and laboratory-induced remanences but exhibits distinct maxima in the SP and SD fraction. This evidence suggests an intense pedogenesis, while, caused by strong gleying conditions, the production of ferrous minerals dominates. Furthermore, the decreasing s-ratio points to a drastic depletion of the highly susceptible and remanence-carrying minerals of the magnetiteemaghemite group and to the neoformation of highly coercive minerals of the haematiteegoethite group. Unit IV is similarly characterised by reworking of soils, most notably in the upper part. In contrast to the fG, the ABS horizon shows clearly increasing values of susceptibility and remanences in addition to high contents of dithionite soluble iron. This high Fed content reaches values comparable to those of the interglacial soils and remarkably even on the same weak susceptibility level as the interglacial S1, which may be indicative for an intense pedogenesis. However, it is hypothesised that this high Fed content is possibly a heritage of the reworked Eemian soil (S1) as proposed by Meszner et al. (2011). The HZ shows magnetic properties similar to those of the fG and indicates an outstanding pedogenesis temporally
Fig. 10. Summarised magnetic parameters of all investigated sections classified according to stratigraphical units with main loess layers and palaeosols (arithmetic means and standard deviations are given).
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
followed by waterlogging, resulting in strong magnetic depletion. Directly beneath the HZ-horizon, a grey and intense bleached (fS(e) w)-horizon reveals the lowest susceptibility values. The remnants of the Eemian interglacial soil display a moderate increase in the c and s-ratio but are generally influenced by gleyification (“Pseudogley”) and carry a much weaker magnetic signal compared with the Holocene soil. Only the ARM/IRM2000 ratio indicates a comparable relative amount of stable SD particles and proves the typical intense Eemian pedogenesis. This characterisation could be observed in the field and is corroborated by the high clay contents and clear sub-angular structures. In a nutshell, the rock magnetic characteristics of the Saxonian Loess Province differ significantly from that of other neighbouring European loess provinces. Antoine et al. (2001) found interstadial tundra gley layers in the Nussloch profile in the Upper Rhine Area developed during the Weichselian Upper Pleniglacial (end of MIS 3 and MIS 2). These layers were influenced by periodical waterlogging and display similarly weak magnetic susceptibilities but not on such a notably low level as the solifluction layers and gleyed palaeosols in the Saxonian Loess Province. In the Late Würmian loess from the area of Krems (Lower Austria), Hambach (2010) found k and ARM/k values that were mostly increased in pale greyish horizons occurring in generally primary aeolian loess showing the typical weakly brownish colours. He interpreted these horizons as the result of partial gleying during permafrost conditions, indicating at the same time a slightly more humid climate. These slightly more humid conditions may have only led to low concentration changes of magnetic minerals but are clearly expressed in the relative amount of ultra-fine SP particles formed during slightly more intense pedogenesis. Thus, the greyish horizons may be the product of a stronger seasonality with more humid summers but cold winters. Other studies from southeastern/ eastern European loessepalaeosol sequences, e.g., Czech Republic (Forster et al., 1996), Poland (Nawrocki et al., 1999) and Serbia (Buggle et al., 2009; Markovic et al., 2011) present magneticenhanced last interglacial and Early Weichselian palaeosols, reaching comparable high susceptibility values as older interglacial soils due to the complete absence of waterlogging processes. The magnetic behaviour of the Saxonian loess cannot be explained by one of the established magnetic models only, because of its specific palaeoclimatically driven environmental conditions. The model of magnetic enhancement in palaeosols was exemplarily demonstrated for loessepalaeosol sequences from the CLP (e.g., Evans and Heller, 1994) and is also applicable for most European loess sections. After the deposition of the parent material (loess), which holds an inherent province-dependent composition of magnetic particles, the pedogenesis and weathering processes led to an in-situ neoformation of ferrimagnetic minerals, particularly ultra-fine-grained SP magnetite, during warmer climatic conditions. As a result, the magnetically enhanced (palaeo) soils exhibit higher susceptibility and remanence values in contrast to the unweathered parent loess material. Three major pathways for the natural origin of SP minerals can be specified: (1) weathering of larger magnetic mineral grains, which produces high amounts of ultra-fine SP in addition to low temperature oxidation of magnetite (van Velzen and Dekkers, 1999); (2) bacterial production of extracellular SP iron minerals, especially in soils (Fassbinder et al., 1990); and (3) production of SP hydromaghemite during the transformation processes of ferrihydrite (Liu et al., 2008a). While c reaches values up to 300 108 m3 kg1 and cfd up to 12% in Chinese palaeosols (e.g., Heller and Evans, 1995), the values in Saxony are generally lower (compare with Fig. 5). Only a weak magnetic enhancement was found in the palaeosols of the upper two loess layers. Indeed, significant pedogenesis is also obvious in the lower units: however, their primary magnetic enhancement is
91
not reflected in the observed data. Surprisingly, they exhibit completely opposite features: a lower c in the palaeosols than in the loess layers. Within a first view, this phenomenon seems to be explained by the so-called wind vigour model, which was previously described for Alaskan loess (Begét and Hawkins, 1989; Begét et al., 1990) and later for Siberian loess sections (Chlachula et al., 1998). This model demands higher values for c in unweathered parent loess than in intercalated palaeosols caused by higher inputs of dense iron oxide particles (especially ferrimagnetic minerals) carried by stronger winds in the glacial periods (Evans, 2001). In contrast, the results of this study show a leading formation of highcoercive (hard) magnetic minerals (particularly haematite) in the palaeosols upon a strong depletion of primary MD and SD grains relative to SP particles. These observations provide evidence for an intense degradation and destruction of pre-existing remanencecarrying minerals of the magnetiteemaghemite group. Consequently, the wind vigour model cannot be applied in Saxony because of the dominance of processes that initiated a magnetic depletion; in particular, these include intense waterlogging and highly dynamic reworking processes. According to Maher (1986), permanently wet and anoxic soils do not display a neoformation of magnetite and may lead to a magnetic depletion. Such depletions of ferrimagnetic minerals have been previously discussed in the literature (e.g., Evans and Heller, 2003; Maher, 2011) and were exemplarily observed in waterlogged/gleyed palaeosols in Siberia (Matasova et al., 2001; Zhu et al., 2003) but also in gleyed soils in Poland and Western Ukraine (Nawrocki et al., 1996, 1999). Matasova et al. (2001) took into account that the magnetic properties of the Siberian Kurtak section most probably depend on concentration changes of ferrimagnetic minerals, which cannot be explained by the wind vigour model only. Furthermore, this section shows clear external influences of seasonal flooding, colluvial and solifluction processes that generate its specific magnetic signal. Furthermore, a thorough study by Liu et al. (2008b) demonstrates the contrasting palaeo-environmental conditions in southern Siberia and in the CLP and their effect on the rock magnetic properties. Based on investigations from Maruszczak and Nawrocki (1991) and Nawrocki (1992), Heller and Evans (1995) discussed the possibility that certain Polish palaeosols were generated under more humid and cooler periglacial climatic conditions, thus showing a drastically depleted and degraded magnetic phase. According to them, these climatic circumstances are mainly responsible for gleyification (waterlogging) and leaching of the soils, preferably during interstadial and interglacial times. Such secondary processes as waterlogging and reworking resulted in decreasing magnetic grain size, as well as in general depletion of detritic and pedogenetic magnetic particles. These processes are particularly dominant for the genesis of the environmental magnetic record of the Middle and Early Weichselian loess units in Saxony. Therefore, the strong correlation of the concentration dependent magnetic parameters with pedologically proven magnetic depletion processes is evident. This observation indicates that the classic magnetic enhancement model is not applicable here. Furthermore, the observed increasing cfd with decreasing c gives rise to the strong weathering of primary MD and SD grains according to the model of van Velzen and Dekkers (1999). The magnetic fingerprint of the Saxonian loess is due to climatically driven processes that reflect the destruction of magnetic particles in palaeosols rather than its neoformation. According to Eissmann (2002), the Early Weichselian Saxonian loess is often associated with mud, peat, insect and faunal remains from a rich glacial flora, which indicate relatively humid conditions. During the last glacial maximum, the Saxonian Loess Province was located in the periglacial area with a minimum distance to the former ice sheet of approximately 90 km (Eissmann, 2002).
92
P. Baumgart et al. / Quaternary International 296 (2013) 82e93
However, in addition to the accumulation of loess, strong solifluction and waterlogging due to permafrost dominated the surface under relatively humid conditions, generating magnetically depleted palaeosols and reworked loess derivatives. Those phenomena occurred in all Saxonian profiles as well as in the Saxony-Anhalt Loess Province located to the northwest of the study area. Similar features occur in other Eurasian loess provinces as well, especially in the middle and higher latitudes. The uppermost layers, however, do not exhibit any discordances or reworking features and represent a semi-continuous Late Weichselian loess blanket. This archive reflects drier climatic conditions, which reveal the neoformation and weak enhancement of magnetic particles in the course of pedogenesis in their palaeosols. Hence, the magnetic fingerprint of the Early and Middle Weichselian loess in Saxony is mainly due to cooler and more humid climatic conditions, which are in turn responsible for waterlogging. Maher (2011) demands a site-specific interpretation of palaeosol magnetic properties with consideration not only of pedogenic enhancement but also of pedogenic dilution or depletion and allochthonous inputs of magnetic minerals. Therefore, a new magnetic facies model is proposed for more humid (Central European) loess provinces dominated by typical periglacial conditions, including widespread permafrost, which control the intense reworking and waterlogging (gleyification) processes. 5. Conclusions Detailed rock magnetic studies of four Weichselian loesse palaeosol sequences from Eastern Germany yield data to establish a stratigraphical model of Saxonian Loess Province based on rock magnetic characteristics. The results of this study lead to the following conclusions: (1) All investigated sections provide almost identical magnetic properties and could be mutually correlated. Furthermore, the previously defined pedostratigraphical units are well reflected by the established environmental magnetic record in which they can also be defined as rock magnetic units. (2) The Early and Middle Weichselian units IIbeIV exhibit a leading formation of hard magnetic minerals in addition to a strong reduction in the relative abundance of primary detritic or pedogenic MD and SD grains. This observation gives evidence for a radical destruction of pre-existing magnetic minerals (particularly magnetite) resulting in magnetically depleted palaeosols. These findings result from such secondary processes as reworking (solifluction) and waterlogging (gleyification) due to permafrost conditions. This indicates cool and permanently humid climatic conditions during the Early and Middle Weichselian in the Saxonian Loess Province. However, because of the complex composition of the loess derivatives, precise palaeoclimatic interpretations are not currently possible. (3) The upper Late Weichselian units I and IIa, which consist of an almost unweathered loess blanket, display a weak neoformation of ferrimagnetic minerals in palaeosols (particularly SP-magnetite) according to the classical model of magnetic enhancement. This observation and the prevailing homogeneous structure indicate drier climatic conditions during the accumulation of these youngest loess deposits. (4) The wind vigour model cannot be applied because of the prevailing magnetic depletion in all stratigraphical units, especially in the interstadial palaeosols and in the remnants of the Eemian interglacial soil. (5) Because of its specific properties, the magnetic fingerprint of the Saxonian loess cannot be explained by one of the
established environmental magnetic models alone. Moreover, the presented results support the need for a new and independent model that explains the magnetic characteristics of loessepalaeosol sequences dominated by sediment reworking and waterlogging. Acknowledgements We are grateful to the reviewers for the helpful and constructive comments that significantly improved the quality of the manuscript. References Antoine, P., Rousseau, D.D., Zöller, L., Lang, A., Munaut, A.V., Hatté, C., Fontugne, M., 2001. High-resolution record of the last Interglacialeglacial cycle in the Nussloch loessepalaeosol sequences, Upper Rhine Area, Germany. Quaternary International 76/77, 211e229. Begét, J.E., Hawkins, D.B., 1989. Influence of orbital parameters on Pleistocene loess deposition in central Alaska. Nature 337 (6203), 151e153. Begét, J.E., Stone, D.B., Hawkins, D.B., 1990. Paleoclimatic forcing of magnetic susceptibility variations in Alaskan loess during the late Quaternary. Geology 18 (1), 40e43. Buggle, B., Hambach, U., Glaser, B., Gerasimenko, N., Markovi c, S., Glaser, I., Zöller, L., 2009. Stratigraphy, and spatial and temporal paleoclimatic trends in Southeastern/Eastern European loessepaleosol sequences. Quaternary International 196 (1e2), 86e106. Chlachula, J., Evans, M.E., Rutter, N.W., 1998. A magnetic investigation of a Late Quaternary loess/palaeosol record in Siberia. Geophysical Journal International 132 (1), 128e132. Dearing, J.A., Hay, K.L., Baban, S.M.J., Huddleston, A.S., Wellington, E.M.H., Loveland, P.J., 1996. Magnetic susceptibility of soil: an evaluation of conflicting theories using a national data set. Geophysical Journal International 127 (3), 728e734. Eissmann, L., 2002. Quaternary geology of eastern Germany (Saxony, SaxoneAnhalt, South Brandenburg, Thüringia), type area of the Elsterian and Saalian Stages in Europe. Quaternary Science Reviews 21 (11), 1275e1346. Evans, M.E., Heller, F., 1994. Magnetic enhancement and palaeoclimate: study of a loess/palaeosol couplet across the Loess Plateau of China. Geophysical Journal International 117 (1), 257e264. Evans, M.E., Heller, F., 2003. Environmental Magnetism e Principles and Applications of Enviromagnetics. Academic Press, Amsterdam. Evans, M.E., 2001. Magnetoclimatology of aeolian sediments. Geophysical Journal International 144 (2), 495e497. Fassbinder, J.W.E., Stanjek, H., Vali, H., 1990. Occurrence of magnetic bacteria in soil. Nature 343 (6254), 161e163. Fine, P., Verosub, K.L., Singer, M.J., 1995. Pedogenic and lithogenic contributions to the magnetic susceptibility record of the Chinese loess/palaeosol sequence. Geophysical Journal International 122 (1), 97e107. Fitzsimmons, K.E., Markovi c, S.B., Hambach, U., 2012. Pleistocene environmental dynamics recorded in the loess of the middle and lower Danube basin. Quaternary Science Reviews 41, 104e118. Forster, T., Heller, F., Evans, M.E., Havlícek, P., 1996. Loess in the Czech Republic: magnetic properties and Paleoclimate. Studia geophysica et geodaetica 40 (3), 243e261. Haase, G., 1963. Stand und Probleme der Lössforschung in Europa. Geographische Berichte 27 (2), 97e129. Hambach, U., 2010. Palaeoclimatic and stratigraphic implications of high resolution magnetic susceptibility logging of Würmian Loess at the Upper Palaeolithic KremseWachtberg Site. In: Neugebauer-Maresch, C., Owen, L.R. (Eds.), New Aspects of the Central and Eastern European Upper Palaeolithic e Methods, Chronology, Technology and Subsistence. Proceedings of the Prehistoric Commission of the Austrian Academy of Sciences. Vienna, pp. 295e304. Heller, F., Evans, M.E., 1995. Loess magnetism. Reviews of Geophysics 33, 211e240. Heller, F., Liu, X.M., Liu, T.S., Xu, T., 1991. Magnetic susceptibility of loess in China. Earth and Planetary Science Letters 103, 301e310. King, J.W., Banerjee, S.K., Marvin, J., Özdemir, Ö., 1982. A comparison of different magnetic methods for determining the relative grain size of magnetite in natural materials: some results from lake sediments. Earth and Planetary Science Letters 59, 404e419. Lieberoth, I., 1959. Beobachtungen im nordsächsischen Lössgebiet. Zeitschrift für Pflanzenernährung, Düngung, Bodenkunde 86 (2), 141e155. Lieberoth, I., 1963. Lösssedimentation und Bodenbildung während des Pleistozäns in Sachsen. Geologie 12 (2), 149e187. Liu, Q., Deng, C., Torrent, J., Zhu, R., 2007. Review of recent developments in mineral magnetism of the Chinese loess. Quaternary Science Reviews 26 (3e4), 368e 385. Liu, Q., Barrón, V., Torrent, J., Eeckhout, S.G., Deng, C., 2008a. Magnetism of intermediate hydromaghemite in the transformation of 2-line ferrihydrite into hematite and its paleoenvironmental implications. Journal of Geophysical Research 113 (1), B01103. http://dx.doi.org/10.1029/2007JB005207.
P. Baumgart et al. / Quaternary International 296 (2013) 82e93 Liu, X.M., Liu, T.S., Hesse, P., Xia, D.S., Chlachula, J., Wang, G., 2008b. Two pedogenic models for paleoclimatic records of magnetic susceptibility from Chinese and Siberian loess. Science in China (Series D: Earth Sciences) 51 (2), 284e293. Maher, B.A., 1986. Characterisation of soils by mineral magnetic measurements. Physics of the Earth and Planetary Interiors 42, 76e92. Maher, B.A., 1988. Magnetic properties of some synthetic sub-micron magnetites. Geophysical Journal 94, 83e96. Maher, B.A., 1998. Magnetic properties of modern soils and Quaternary loessic paleosols: paleoclimatic implications. Palaeogeography, Palaeoclimatology, Palaeoecology 137, 25e54. Maher, B.A., 2011. The magnetic properties of Quaternary aeolian dusts and sediments, and their palaeoclimatic significance. Aeolian Research 3, 87e144. Markovi c, S.B., Hambach, U., Stevens, T., Kukla, G.J., Heller, F., McCoy, W.D., Oches, E.A., Buggle, B., Zöller, L., 2011. The last million years recorded at the Stari Slankamen (Northern Serbia) loessepalaeosol sequence: revised chronostratigraphy and long-term environmental trends. Quaternary Science Reviews 30, 1142e1154. Markovi c, S.B., Hambach, U., Stevens, T., Jovanovi c, M., O’Hara-Dhand, K., Basarin, B., Lu, H., Smalley, I., Buggle, B., Zech, M., Svircev, Z., Sümegi, P., Milojkovi c, N., Zöller, L., 2012. Loess in the Vojvodina region (Northern Serbia): an essential link between European and Asian Pleistocene environments. Netherlands Journal of Geosciences e Geologie en Mijnbouw 91, 173e188. Maruszczak, H., Nawrocki, J., 1991. Stratigraphicepaleogeographic interpretation of the results of magnetic susceptibility investigations of loesses at Niedelew (SE Poland). Annales e Universitatis Mariae Curie Sklodowska, Section B 46, 173e 185. Matasova, G., Petrovský, E., Jordanova, N., Zykina, V., Kapicka, A., 2001. Magnetic study of Late Pleistocene loess/palaeosol sections from Siberia: palaeoenvironmental implications. Geophysical Journal International 147 (2), 367e 380. Meszner, S., Fuchs, M., Faust, D., 2011. LoessePalaeosol-sequences from the loess area of Saxony (Germany). Quaternary Science Journal 60 (1), 75e93. Meszner, S., Kreutzer, S., Fuchs, M., Faust, D. Late Pleistocene landscape dynamics in Saxony, Germany: paleoenvironmental reconstruction using loess–paleosol sequences. Quaternary International, in press.
93
Nawrocki, J., Wojcik, A., Bogucki, A., 1996. The magnetic susceptibility record in the Polish and western Ukrainian loessepalaeosol sequences conditioned by palaeoclimate. Boreas 25 (3), 161e169. Nawrocki, J., Bakhmutov, V., Bogucki, A., Dolecki, L., 1999. The Paleo- and Petromagnetic record in the Polish and Ukrainian LoessePalaeosol Sequences. Physics and Chemistry of the Earth (A: Solid Earth and Geodesy) 24 (9), 773e777. Nawrocki, J., 1992. Magnetic susceptibility of Polish loesses and loess-like sediments. Geol. Zb. Geologica Carpathica 43, 179e180. Oldfield, F., 1991. Environmental magnetism e a personal perspective. Quaternary Science Reviews 10 (1), 73e85. Richter, H., Haase, G., Lieberoth, I., Ruske, R., 1970. Periglazial-Löß-Paläolithikum im Jungpleistozän der Deutschen Demokratischen Republik. In: Petermanns Geographische Mitteilungen, Ergänzungsheft, vol. 274 (Berlin). Rousseau, D.D., Kukla, G., 1994. Late Pleistocene climate record in the Eustis loess section, Nebraska, based on land snail assemblages and magnetic susceptibility. Quaternary Research 42 (2), 176e187. Ruske, R., Wünsche, M., 1961. Lösse und fossile Böden im mittleren Saale- und unteren Unstruttal. Geologie 10, 9e29. Semmel, A., 1997. Referenzprofile des Würmlösses im Rhein-Main-Gebiet. In: Jahresberichte der Wetterauischen Gesellschaft für die Gesamte Naturkunde zu Hanau, vol. 148, pp. 37e47. Stanjek, H., Fassbinder, J.W.E., Vali, H., Wagele, H., Graf, W., 1994. Evidence of biogenic greigite (ferrimagnetic Fe3S4) in soil. European Journal of Soil Science 45, 97e104. van Velzen, A.J., Dekkers, M.J., 1999. Low-temperature oxidation of magnetite in loessepaleosol sequences: a correction of rock magnetic parameters. Studia geophysica et geodaetica 43 (4), 357e375. Wansa, S., 2009. Kösener Verlehmungszone. In: LithoLex. BGR, Hannover. Last updated 09.07.2009. Cited 03.10.2011. Record No. 1006026. Available from: http://www.bgr.bund.de/litholex. Zhou, L.P., Oldfield, F., Wintle, A.G., Robinson, S.G., Wang, J.T., 1990. Partly pedogenic origin of magnetic variations in Chinese loess. Nature 346 (6286), 737e739. Zhu, R.X., Matasova, G., Kazansky, A., Zykina, V., Sun, J.M., 2003. Rock magnetic record of the last glacialeinterglacial cycle from the Kurtak loess section, southern Siberia. Geophysical Journal International 152, 335e343.
