Significant pedogenic and palaeoenvironmental changes during the early Middle Pleistocene in Central Europe

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology journal homepage: www.elsevier.com/locate/palaeo

Significant pedogenic and palaeoenvironmental changes during the early Middle Pleistocene in Central Europe

T

⁎

B. Bradáka, , Y. Setob, J. Nawrockic a

Department of Physics, University of Burgos, Av. de Cantabria, s/n, 09006 Burgos, Spain Department of Planetology, Kobe University, Nada, Kobe 657-8501, Japan c Maria Curie Skłodowska University, Faculty of Earth Sciences and Spatial Management, Kraśnicka 2cd, Lublin 20-718, Poland b

A R T I C LE I N FO

A B S T R A C T

Keywords: Loess Magnetic hysteresis Day plot European Loess Belt

Three well-developed paleosols from the Paks loess succession (Hungary), a key profile of the European Loess Belt, were sampled and studied by rock magnetic methods, such as low-field magnetic susceptibility and hysteresis measurements. The studied paleosols formed in MIS19, MIS15 and MIS11 and represent key periods in the Middle Pleistocene, namely, the Middle Pleistocene transition (MPT) (MIS19), and frame the mid-Brunhes transition (MIS15 and 11). The results of the low-field magnetic susceptibility measurements, a popular interglacial intensity proxy, showed different patterns and values. The different characteristics of the susceptibility curve indicate different degrees of pedogenesis and therefore different soil-forming palaeoenvironments. The comparison of the rock magnetic (hysteresis) parameters supports the results of the susceptibility measurements and revealed some components of two characteristic palaeoclimate types: i.) a humid and warm MIS19 (pre-)MPT interglacial environment with very intense weathering and pedogenesis and ii.) a moderate climate with seasonal precipitation in a post-MPT interglacial environment during MIS15 and MIS11 characterized by intense but different types of pedogenesis compared to MIS19. This result, compared with other loess successions in the Middle and Lower Danube Basin and the western part of the East European Plain, helps to improve the climate model regarding the changing of the characteristic palaeoenvironment from a sub-Mediterranean (in the south) and temperate and humid climate (with forests, in the north) towards a cooler grassland/forest steppe-ruled environment in the region.

1. Introduction The Middle Pleistocene transition (MPT) and mid-Brunhes event or transition (MBT) are important periods of the Pleistocene (Jansen et al., 1986; Head and Gibbard, 2005; Yin, 2013; Barth et al., 2018) characterized by significant changes in various parameters of the palaeoclimate, such as the periodicity of climate cycles and intensity of the interglacial. Although the studies about the suggested periods are increasing rapidly, there are many questions that remain unanswered, including the differences between the marine and terrestrial environments (Candy and McClymont, 2013) and between regional palaeoclimatic characteristics (Candy et al., 2010). The change in palaeoclimatic parameters is well reflected in various proxies from various records, such as the commonly used magnetic susceptibility. Low-field volumetric susceptibility (κlf) and mass susceptibility (Σf) have an important role in terrestrial climatostratigraphy and palaeoclimate

⁎

reconstructions (Forster et al., 1994). Magnetic susceptibility has been used as a pedogenic proxy and an important indicator of interglacial intensity (Hao et al., 2012; Past Interglacials Working Group of PAGES, 2016), especially but not exclusively limited in the case of transition periods (Heslop et al., 2002). In loess-paleosol sequences, loess horizons are indicated by low κ(Σ) lf and paleosols by high κ(Σ)lf. Increasing susceptibility in paleosols results from superparamagnetic (SP) mineral contributions compared to multidomain and stable single domain grains in loess. Magnetic enhancement of paleosols is due to biogenic processes strongly connected to pedogenesis (Maher and Taylor, 1988). This enhancement by pedogenic processes is called the pedogenic model (Evans, 2001). The application of the pedogenic model in the loess profile led to the generally accepted hypothesis about the relationship among the magnetic susceptibility, the degree of pedogenesis and the interglacial intensity. Based on the hypothesis, it can be assumed that the in situ formation of

Corresponding author. E-mail address: [email protected] (B. Bradák).

https://doi.org/10.1016/j.palaeo.2019.109335 Received 18 February 2019; Received in revised form 24 June 2019; Accepted 15 August 2019 Available online 20 August 2019 0031-0182/ © 2019 Elsevier B.V. All rights reserved.

Fig. 1. Location of the Paks profile in Hungary (a) and in the European Loess Belt (b) (the European loess map is based on Haase et al., 2007). AT-Austria; CZ-Czech Republic; HR-Croatia; RO-Romania; RS-Republic of Serbia; SI-Slovenia; SK-Slovak Republic and UA-Ukraine. Profiles: ČK - Červený Kopec (Red Hill), Czech Republic (Forster et al., 1996); Ki – Koriten, Bulgaria (Jordanova and Petersen, 1999); Ko - Korolevo, Ukraine (Nawrocki et al., 2016); Lu-Lubenovo (Jordanova et al., 2007, 2008); NE – Novaya Etuliya (45.51N, 28.41E), Ukraine (Gendler et al., 2006); Pa - Paks, Hungary (Sartori et al., 1999; Újvári et al., 2014); Ro – Roxolany, Ukraine (Gendler et al., 2006); SP – Skala Podilska (Boguckyj et al., 2009); St – Stalać (Kostić and Protić, 2000); SS - Stari Slankamen, Serbia (Marković et al., 2011); Vi – Viatovo, Bulgaria (Jordanova et al., 2008, Za Zahvizdja, Ukraine (Nawrocki et al., 2002); Zi – Zimnicea, Romania (Rădan, 2012). The lithostratigraphical subdivision of the studied section with the sampling points (c): L1 to L6-(aeolian) loess; SL1-sandy loess; S1 to S5: sand; LSi-laminated silt; MB-Mende Base paleosol; Mtp-forest soil horizon; and PD1 and PD2-Paks Double1 and 2 paleosol horizons. Asterisks indicate the sampling points of kT rock magnetic pilot samples. The figure is modified after Bradák et al. (2018a).

B. Bradák, et al.

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

2

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

area, where the red paleosol PK4 formed in MIS15 displayed the highest magnetic susceptibility (Tsatskin et al., 2008) (Σlf: ~140 × 10−8 m3/ kg), whereas the paleosol containing the MBB record displayed values of magnetic susceptibility almost two times lower, although they were evidently elevated with respect to the values noted in the loess. It should be noted that paleosols from this oldest part of the Brunhes chron at Novaya Etuliya were calcified during extensive post-burial alteration (Tsatskin et al., 2008). The character of paleosols in the northern part of MDB and Central and Eastern Europe seems to be more complex. Various types of paleosol horizons were identified in the early Middle Pleistocene as part of the Paks succession (e.g., Pécsi, 1995; Újvári et al., 2014). The paleosol horizons (as well as the formation environment) are characterized by older terms, namely, ‘Mediterranean-like forest soil’, ‘forest steppe soil’, ‘brown forest soil’ and ‘chernozem forest soil’ (Pécsi, 1995). Despite the differences in soil type, the geochemical study of Újvári et al. (2014) indicated a quasi-constant smectite content and disproved the appearance of progressive aridization in the northern part of the MDB, as expected based on the theory of Buggle et al. (2013, 2014). The loess profiles summarized above contributed to the conceptual pan-European loess model (Marković et al., 2015), but no profiles from the northwestern part of the EEP were included. To fill this gap, key sections from the region are summarized below in detail. The loess sections in the northwestern part of the EEP (West Ukraine) containing the Matuyama – Brunhes boundary (MBB), a key stratigraphic marker at the definition of MIS19, are not numerous. A high-resolution record of the older part of the Quaternary containing this boundary was disclosed in the sequences of the Skala Podilska (Boguckyj et al., 2009) and Zahvizdja (Nawrocki et al., 2002) sites. The MBB was also defined in the loess-paleosol key Palaeolithic section in Transcarpathia at Korolevo (Nawrocki et al., 2016). The stratigraphy of soils younger than the MBB and their correlation with marine isotope stages is based on the TL ages, correlation with the terrace chronology, and palaeopedologic and palaeobiologic criteria (Kusiak et al., 2013). Because of the discontinuous sedimentation and complex nature of paleosols, in some of the above-mentioned sections, this correlation is not unequivocal. The magnetic susceptibility measured for the profile in the Skala Podil'ska site displayed the highest value (κlf:848 × 10−6 SI) in the uppermost part of the forest paleosol formed in MIS19 (Boguckyj et al., 2009). It gradually decreased in the younger soils of the pedocomplex formed in MIS17 (κlf:200 × 10−6 SI). No remnants of genetic horizons of particular soils of this pedocomplex were distinguished. The κlf of the paleosol formed in MIS15 increases from the bottom part of the Bt horizon to its boundary with a peak at the accumulation horizon (κlf:563 × 10−6 SI). The upper MIS13 soil of this pedocomplex showed values of κlf not higher than 250 × 10−6 SI (i.e., comparable to those noted in the loess). The soil from the time of MIS11 was not preserved in the Skala Podil'ska section. It should be noted, however, that very high values of magnetic susceptibility (κlf:up to 739 × 10−6 SI) were also recorded here in the upper part of the Bt horizon of the MIS5e paleosol. The mean annual precipitation was almost two times higher in the pedocomplex formed in MIS19-17 than in the younger pedocomplex deposited in MIS15-13 (Boguckyj et al., 2009). The forest paleosols of the pedocomplex formed in MIS19 and MIS17 from another section, i.e., the Zahvizdja (Nawrocki et al., 2002), are strongly gleyed. During water stagnation, a portion of the magnetic oxides was most likely hydrated. Because of this, the magnetic susceptibility of the paleosols is only slightly higher than noted in the overlying loess (fluctuating here between κlf:100 and 206 × 10–6 SI only). Its maximum values were noted in the middle or upper parts of the Bt horizons of particular paleosols of this pedocomplex. The paleosol formed in MIS15 is rubificated red in colour, and it is assumed to relate it to a warm climate (Mediterranean-like). Similar red soil occurs in the central part of the Roxolany loess-paleosol sequence (soil PK4, Tsatskin et al., 1998). Elevated κlf no higher than 190 × 10–6 SI occur

a new magnetic fraction is increasing by a longer time span of pedogenesis and/or in an (very) intense pedogenic environment. In contrast, an irregular case was noted in the susceptibility curve from the Paks loess profile (Hungary) (Sartori et al., 1999), which was supported by recent measurements (Fig. 1) (see below). In the case of the Paks succession, four well-developed, thick palaosols can be identified as formed during the period from MIS19 to MIS11 (Pécsi, 1995; Sartori et al., 1999; Újvári et al., 2014). Based on its susceptibility value, the lowermost paleosol (PD2) (Pécsi, 1995), dated back to MIS19, is the most developed (Fig. 1c). There are two well-developed but different types of paleosols (MIS15 - Mtp and MIS11 - MB paleosols) (Pécsi, 1995). Although the well-developed MIS11 paleosol was formed during one of the most intense interglacials, its susceptibility is much lower than that of the MIS19 paleosol, which formed during a less intense interglacial among the interglacials in the studied time frame (from MIS19 to MIS11) (Past Interglacials Working Group of PAGES, 2016) (Fig. 1c). In addition, all the suggested paleosols are formed in ‘key’ interglacials, which represent an important period during the early Middle Pleistocene. Based on its magnetostratigraphic (Sartori et al., 1999) and chronostratigraphic position (Újvári et al., 2014), the MIS19 paleosol was formed during the MPT in the period when the 41-ky climate cycles were replaced by high-amplitude 100-ky cycles (~850 to 680 ka), as was identified in loess from the Chinese Loess Plateau (CLP) (Heslop et al., 2002). MIS15 and MIS11 paleosols framed another important transition: the MBT (Barth et al., 2018). Along with the studied Paks section (Fig. 1c and Material and methods chapter), there are other key sections in the region that date back to the early Middle Pleistocene and help in understanding the period of the suggested climatic transitions (Fig. 1b). The central and eastern part of the European Loess Belt (ELB) can be separated into various regions, including the Middle and Lower Danube Basin (MDB and LDB, respectively) (e.g., Buggle et al., 2013) and the Eastern European Plain (EEP). During the early Middle Pleistocene, a significant influence of the Mediterranean can be recognized on the soil development in the loess profiles from the MDB and LDB, including Batajnica (Marković et al., 2009), Stari Slankamen (Marković et al., 2011, 2015), Lubenovo, Viatovo, Koriten (Jordanova et al., 2007, 2008) and Zimnicea (Rădan, 2012). The characteristics of the environment forming the palaosols gradually change to steppe-like during development of the soils in the early Middle Pleistocene. This change is a possible indicator of the strengthening continental influence in the area (Buggle et al., 2013, 2014). The same phenomena can be observed in the Novaya Etuliya (Tsatskin et al., 2008) and Roxolany (Tsatskin et al., 1998; Gendler et al., 2006) loess profiles from the southwestern part of the EEP. The international importance of the Roxolany loess profile from the Black Sea region of Ukraine results from its abnormal thickness (> 50 m), numerous studies and interregional correlation within the European loess deposits, as well as the correlation of this section with the loess in China. It should be stressed, however, that the stratigraphy of the Roxolany profile is still a matter of controversy. Tsatskin et al. (1998) placed the MBB inside the loess deposit at approximately 33.2 m in depth, whereas Bakhmutov et al. (2017) defined this boundary in the interglacial paleosol at approximately 46.7 m in depth. In this profile, the highest values of magnetic susceptibility (up to 1000 × 10−6 SI) were noted in the topmost part of the paleosol PK4 formed in MIS3 (Bakhmutov et al., 2017) or MIS15 (Tsatskin et al., 1998). In the rest of the interglacial palaosols from Roxolany with elevated values of magnetic susceptibility, this parameter fluctuates between approximately 500 and 800 × 10−6 SI units. Referring to the stratigraphic scheme adopted by Bakhmutov et al. (2017), the magnetic susceptibility values of the paleosols formed in MIS19 and MIS15 are significantly lower than those of the paleosols formed in MIS11 and MIS3. The same pattern of magnetic susceptibility was observed in the loess/paleosol sequence at Novaya Etuliya in the western Black Sea 3

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

hysteresis measurements were conducted on a Micromag 2900 AGM alternating gradient field magnetometer (Princeton Measurements Co., United States). A maximum applied field of 1 T (the limitation of the instrument) was used during the hysteresis measurements. During the dia/paramagnetic correction of the hysteresis curve, the default (70%) setting of the dia/paramagnetic adjustment was applied (Suppl. Mat. 1). The shape of the hysteresis loop was characterized based on the studies of Tauxe et al. (1996, 2010). ΔM curves, the difference between the ascending and descending parts of the hysteresis loops for B > 0, were used to reveal the differences between similar hysteresis loops. The differences between the descending and ascending parts of the hysteresis loop for H > 0 were plotted and analysed based on the method of Tauxe et al. (1996). Slight differences between the wideness/ narrowness of the hysteresis loop, i.e., the distance between the descending and ascending parts of the curve, and its response to the changing applied field can be indicated by the comparison of the shape of the ΔM curves. ΔM is the difference between the M of the descending and ascending parts of the hysteresis loop at the same H with the premise that H > 0 (Tauxe et al., 1996). The coercivity of remanence (remanent coercive force) to coercivity (Hcr/Hc) ratio and the saturation remanence to saturation magnetisation (Mrs/Ms) plot, the so-called Day plot (Day et al., 1977; Dunlop, 2002), were used to reveal the multidomain, single-domain (SD), superparamagnetic (SP) and vortex state (Roberts et al., 2017) (SD + MD mixture or pseudo-single domain [PSD] in earlier studies) of the magnetic mineral components in the samples.

in the upper part of horizon Bt and in the accumulation horizon in the MIS15 paleosol. Significantly higher κlf (321 × 10−6 SI) was noted in one sample from the upper part of the Bt horizon of the MIS13 paleosol. The upper part of the Zahvizdja section seems to be more enhanced in magnetic oxides, at least in the loess parts. The loess underlying the paleosol formed in MIS11 displayed κlf up to 271 × 10−6 SI, whereas the magnetic susceptibility of this paleosol is below 200 × 10−6 SI. Most of the Bth and Btg horizons of paleosols distinguished in the Korolevo site are gleyed, and because of this, elevated κlf (exceeding 1000 × 10–6 SI) are dispersed in their profiles and are characteristic mainly of the Eet horizons of the paleosols formed in MIS19, MIS15 and MIS9 (Nawrocki et al., 2016). The paleosol formed in MIS11 did not record such κlf peaks (300 × 10−6 SI). However, isolated peaks of magnetic susceptibility are also characteristic of thin loess horizons of the Korolevo section. These peaks are probably due to the proximity of highly magnetic andesite lavas and river sediments containing heavy minerals that could be one of the primary sources of ferric oxides deposited in the Korolevo loess section. Because of this, the loess/paleosol section from Korolevo is not promising for further magneto-mineralogical studies dedicated to palaeoclimate reconstruction. No appearance of a warmer, (sub-)Mediterranean climate was reported in paleosols from the northern part of Central Europe and the EEP, except for one profile. Based on the characteristics of the MIS13 and MIS15 paleosols in the Zahvizdja loess succession (northwestern part of the EEP), the MIS13 and MIS15 interglacials are reported as warmer compared to MIS11 (Nawrocki et al., 2002). In other profiles, including Červený Kopec (Red Hill) (Forster et al., 1996) and Korolevo (northwestern part of the EEP) (Nawrocki et al., 2016), the change of forest soils to continental, steppe soils was observed between MIS15 and MIS11. For the reasons suggested above, systematic (re)sampling and rock magnetic (hysteresis) studies were performed on the pedogenic ‘subhorizons’ of the MIS19, MIS15 and MIS11 paleosols at Paks (Hungary) (Fig. 1c). The goal of this study was to determine the differences in pedogenic conditions among three similar-looking, well-developed paleosols from the Paks profiles by using magnetic susceptibility and hysteresis measurements of the soil subhorizons and the over- and/or underlying sediments. In addition, the detailed study may provide information about the terrestrial palaeoclimatic/environmental character of the interglacials around the important transition phases of the Middle Pleistocene.

2.2. Scanning electron microscope study Well-polished samples were studied by using a JSM-6480LAII scanning electron microscope (SEM; JEOL, Tokyo, Japan) equipped with energy-dispersive X-ray spectroscopy (EDS) and a JXA-8900 electron probe microanalyser (EPMA: JEOL) equipped with a wavelength-dispersive X-ray spectrometer (WDS). To obtain flat and smooth surfaces, we impregnated the samples with a low-viscosity resin (Petropoxy 154) and polished them using SiC and Al2O3 abrasives without a lubricant (to avoid altering the clay minerals). For SEM observations, we used back-scattered electron imaging. Chemical analyses using EDS were obtained at 15 kV and 0.4 nA. Data corrections were made by the ZAF method with well-established natural/synthesised materials as chemical standards. X-ray elemental maps were acquired on the EPMA at 15 kV and 20 nA with a pixel size of 1 μm and a counting time of 5 ms per pixel.

2. Materials and methods

3. Results

The Paks loess profile is located to the north of the town of Paks in the Pannonian Basin, Hungary, on the right bank of the Danube River. As a part of the ELB, the investigated succession plays an important role in pan-European loess stratigraphy and terrestrial palaeoclimate reconstruction (Újvári et al., 2014; Marković et al., 2015). As part of the recent sampling session, a 16-m thick loess/paleosol sequence was cleaned, high-resolution sampling was performed, and various rock magnetic and magnetic fabric studies were conducted (e.g., Bradák et al., 2018a, 2018b). A detailed description of the studied loess and paleosol units can be found in Bradák et al. (2018a, 2018b); therefore, only the studied profile with the stratigraphical position of the sampling points is presented here (Fig. 1c), and the main characteristics of the sampled layers are summarized in Supplementary Material 1.

3.1. Hysteresis loops and ΔM curves Based on the shape of the hysteresis loop, single vortex state (in previous works: PSD - Day et al., 1977; or mixed SD + MD - Dunlop, 2002) (Roberts et al., 2017) components are the characteristic rock magnetic contributors in the pilot samples (Tauxe et al., 1996; Tauxe et al., 2010) (Fig. 2). The loops of the loess samples closed to higher fields than the paleosol curves. This phenomenon can relate to SD or vortex (PSD) state magnetic grains, which are the most effective carriers of remanence (Deng et al., 2004) or can indicate a higher coercivity (antiferromagnetic) component (Necula and Panaiotu, 2012). In addition to this general statement, the differences between the ΔM curves reveal additional information (Table 1; Fig. 3; see below). In the MIS19 (PD2; Fig. 3a) paleosol (like the MIS15 paleosol), the overlying loess is characterized by wider hysteresis loops compared to those of the soil that are well separated from the soil curves, which reach 90% ΔM decay at approximately 200–250 mT (Fig. 3a). The paleosol samples have a narrow hysteresis loop with an exponential decay that reaches 90% ΔM decay at ~ 125 mT with a sharp drop between 0 and 40 mT.

2.1. Rock magnetic experiments An SM-100 ZH-Instrument portable magnetic susceptibility meter (Brno, Czech Republic) was used to determine the low-field mass susceptibility (500 Hz; xlf) characteristics of the samples. Six to 10 samples from every horizon were measured, and the averages of the measurements are presented (Suppl. Mat. 1) and were used during the analysis. To estimate the domain state of the magnetic components, 4

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Fig. 2. Results of hysteresis experiments of pilot samples from the studied paleosol horizons in the Paks loess succession. Hysteresis loops of the PD2 (MIS19) (a, b), Mtp (MIS15) (c, d) and MB (MIS11) (e, f) paleosols and loess (g, h). Corrected (adjusted) curves for the paramagnetic and diamagnetic contributions are in brighter tones in the diagrams.

vertical distribution in the body of the soil (Fig. 4c). In the MIS11 (MB; Fig. 4d) paleosol, the under- and overlying sediments contain < 10% (~5%) SD magnetic components (Fig. 4d; 1, 2 and 12). The content increases to SD ≥ ~10% in the transient horizons (Fig. 4d; 4 and 5) and reaches ~17% in the uppermost part of the soil (Fig. 4d; 6).

The shape and character of the soil curves are very similar to each other (Fig. 3a). In the MIS15 paleosol (Mtp; Fig. 3b), the under- and overlying sediments have significantly wider loops compared to those of the soils, and the curves reach the 90% ΔM decay above 300 mT. The curves from the paleosol horizons are well separated from the sediment samples (Fig. 3b). In the MIS11 paleosol (MB; Fig. 3c), the overlying sediment is characterized by a wider hysteresis loop, and a narrow one from the parent material of the soil. The curves from the sediment/paleosol transient horizons overlap with the pedogenic horizon. This indicates a relatively narrow hysteresis loop with an exponential-like decay reaching 90% ΔM decay between 100 and 150 mT with a sharp drop between 0 and 50 mT (Fig. 3c).

3.3. Relationship between hysteresis properties and xlf The relationship between Ms. and Σlf shows a weak linear relationship (R2 = 0.24) compared to the moderate relationship between Mrs. and Σlf (R2 = 0.52) (Table 1; Fig. 5). In contrast to the relationship among Ms., Mr. and Σlf, Hc and Hcr have a strong linear relationship throughout the samples (R2 = 0.88) (Fig. 5c). Sediment samples with low Σlf are characterized by high Hc and Hcr, and the paleosols are indicated by lower Hc and Hcr but higher Σlf. Hcr and Hc values decrease rapidly when Σlf is below 0.3–0.4 and decrease slowly when Σlf is above 0.3–0.4 (Fig. 5d and e).

3.2. Day plot analysis The coercivity of the remanence (remanent coercive force) to coercivity ratio (Hcr/Hc) ranged from ~3 to 4.6 for all samples and from 3.1 to 4.6 for MB, from 3.2 to 3.6 for Mtp and approximately 3.3 for the MIS19 (PD2) paleosols (Fig. 4). The saturation remanence to saturation magnetisation (Mrs/Ms) ranged from 0.032 to 0.077 for overall samples and from 0.066 to 0.072 for the MIS19 (PD2), from 0.049 to 0.054 for the MIS15 (Mtp) and from 0.059 to 0.074 for the MIS11 (MB) paleosols (Fig. 4b, c and d). The detailed Day plot analysis of samples related to individual paleosol horizons provided the following results. In the MIS19 (PD2) paleosol, as in MB and Mtp, the over- and underlying sediments contain 5–10% SD components (Fig. 4b; 1 to 4). The ratio of SD minerals in the transition horizon is approximately 12% (Fig. 4b; 5) and reaches ~15% in the lowermost part of PD2 (Fig. 4b; 10). In the MIS15 (Mtp) paleosol, the under- and overlying sediments contain < 10% (~5%) SD magnetic minerals, except for one sample with SD = ~10% (Fig. 4c; 2, 3, 4, 11 and 1). There are no large differences between the transient horizons and the soil horizon, which are characterized by an ~10–11% SD ratio with a relatively consistent

3.4. Scanning electron microscope analysis In comparison to the sporadic appearance of ferromagnetic minerals (magnetite, ilmenite), a high amount of micas (e.g., muscovite) and chlorite was observed in the loess. Other common mineral components were quartz and calcite in relatively greater quantities. The presence of apatite, rutile, K-feldspar and epidote was also noted (Fig. 6a and b). Along with the aforementioned features, the fragmentation of coarser grains was also observed (Fig. 6c and d). The microfabric of the paleosol samples was characterized by a more compact, clayey matrix and finer grains (Fig. 6c and d). Compared to ferromagnetic contributors such as ilmenite, magnetite/haematite and goethite, a smaller quantity of phyllosilicates was observed. Various marks of the higher stages of weathering were also observed, such as the degradation of micas or the faintness of the boundary between the minerals and the matrix (Stoops et al., 2010). The presence and quantity of other components, including quartz, epidote, K-feldspar, and rutile, was similar in loess. 5

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Pedogenic enhancement of magnetic components is the strongest compare to MB and Mtp. SD = ~12% SD max = ~15% (lowermost part)

SD ≥ ~10% SD max = ~17% (uppermost part of the soil) SD < 10% (~5%) except one sample with ~10% SD = ~10% SD max = ~10–11% (consistent vertical distribution). SD = 5–10% Overlaping of trans. hor. and paleosol; narrow; exponential like decay, reaching the 90% ΔM decay between 100 and 150 mT; sharp drop between 0 and 50 mT.

The shape of the hysteresis curves indicates a significant vortex state (PSD) and SD magnetic contributors in the studied samples. The differences between the ΔM curves reveal additional information about the character and vertical distribution of the magnetic contributors (Table 1; Fig. 3). In the MIS19 paleosol, the shape and character of the soil ΔM curves are very similar to each other (Fig. 3a), which possibly indicates uniform pedogenic processes in the entire paleosol horizon with characteristic mineral formation (Table 1; Fig. 3a). Similar to the MIS15 paleosol, the well-separated group of paleosol and sediment curves indicates in situ-formed pedogenic component(s). The well-separated group of paleosol and sediment ΔM curves of the MIS15 paleosol possibly indicates in situ-formed pedogenic component (s), which cannot be identified in the sediments but appear in the soil (Fig. 3b). The pedogenic characteristics of the curves increase towards the lower pedogenic horizons, which may indicate significant vertical pedogenic processes (e.g., higher precipitation-related water infiltration and material migration; see above). This phenomenon can also indicate increasing in situ mineral formation in the lower part of the soil (Table 1; Fig. 3b). The in situ mineral formation at the bottom of the paleosol unit is more likely a non-pedogenic process. Theoretically, the intense water infiltration and saturation of pores (above the compact CCa/Bk horizon) could provide various environments for magnetic mineral formation (Roberts, 2015). In the MIS11 paleosol (MB), a similar ΔM curve of the transient horizons and soil horizons possibly indicates limited vertical movements, such as the effect of water infiltration by continuous precipitation, which causes various weathering characteristics vertically in the body of the soil and the migration of the components downward. In aggradational systems (e.g., loess-paleosol systems), where the balance between the sediment input and pedogenesis has a significant role in soil formation, the transient horizon of a paleosol can be interpreted as a weak paleosol. In this horizon, due to the increasing sediment input, pedogenic processes, including a high degree of weathering or vertical migration, are limited (e.g., Kraus, 1999). These processes can be reflected by the significant differences between the ΔM curves of the transient horizon and soil horizon but cannot be recognized in the case of the MIS11 soil (Fig. 3c). The Day plot indicated that the data points settled into a field of vortex state (SD + MD mixture or pseudo-single domain [PSD] in earlier studies) with an approximately 85–95% MD component (Fig. 4). Although the samples from each (MIS11, MIS15 and MIS19) unit were yielded by similar Hcr/Hc and Mrs./Ms. ratios, the differences between the over- and underlying (parent material) sediments and pedogenic horizons (especially in the SD content) and the vertical change of the parameters may indicate different pedogenic environments, which are described in Table 1, Table 2 and Fig. 4. In the MIS19 (PD2) paleosol, the location of the maximum pedogenic-forming SD magnetic components in the lower part of the unit (Fig. 4b; 10) may indicate intense vertical processes, the vertical translocation of materials (Stoops et al., 2010), and possibly the highest precipitation among the three studied pedogenic periods. The upper and lower parts of PD2 are well separated by the distribution of the samples on the Day plot. This may indicate two different phases of the palaeoenvironment during MIS19 separated by abrupt climatic change, decreasing precipitation and possibly increasing sediment input. In the MIS15 paleosol (Mtp), the similar ratio of SD pedogenic mineral contents in the transient and soil horizons and the thin upper transient horizon (Fig. 4c; from 5 to 10 with the intermixing of some sediment samples, e.g., 1) indicate fast pedogenesis with low additional sedimentation during pedogenesis (interglacial) but fast burial of the

Parent mat./over. sed. Trans. hor. Pedogenic horizon

Parent mat./over. sed. Trans. hor. Pedogenic horizon

Parent mat./over. sed. Trans. hor. Pedogenic horizon

MIS15 (Mtp) paleosol

MIS19 (PD2)

Wide hysteresis loops (compare to soils); curves are reaching the 90% ΔM decay above 300 mT. Character of the curvesis in between the sediment and soil samples. Narrow hysteresis loop, with an exponential like decay; reaching the 90% ΔM decay between 125 and 175 mT; sharp drop between 0 and 50 mT Wider hysteresis loops compare to soil; and well-separated; reaching the 90% ΔM decay around 200–250 mT In between the sediment and soil. Narrow hysteresis loop; reaching the 90% ΔM decay at ~125 mT with a sharp drop between 0 and 40 mT; very similar soil curves;

SD < 10% (~5%) Overlaying sediments - wider loop; parent material - narrow loop.

Pedogenic enhancement of magnetic components slightly weaker compare to MB and weaker compare to PD2.

Pedogenic enhancement of the magnetic composition of the paleosol are stonger than Mtp, but weaker than PD2.

4.1. Palaeoenvironmental significance of the character and vertical distribution of magnetic components

MIS11 (MB paleosol)

ΔM

Table 1 The summarized results of the applied experiments.

Day plot

Pedogenic enhancement - xlf vs. Hc (and Hcr) plot

4. Discussion

6

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Fig. 3. The ΔM curves of the studied units. The ΔM curves reveal the differences between the shape and saturation character of the hysteresis loop from sediments (light grey colour), transient horizons (darker grey colour) and paleosols (black colour) in the a.) MIS19 (PD2), b.) MIS15 (Mtp) and c.) MIS11 (MB) paleosols.

Fig. 4. Day plot of the hysteresis measurement from the Paks loess-paleosol samples: a.) general overview, b.) the MIS19 (PD2) paleosol, c.) the MIS15 (Mtp) paleosol, and d.) the MIS11 (MB) paleosol. The increasing numbers at the marks indicate their vertical position in the paleosol body from the uppermost position (1) downward. The multidomain (MD), single domain (SD), and superparamagnetic (SP) areas are based on Day et al. (1977). The contribution of SD and MD components was determined by the SD–MD mixing curves (Dunlop, 2002). 7

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Fig. 5. Magnetic enhancement of the paleosols. Plots of low-field magnetic susceptibility (Σlf) vs. a.) saturation magnetization (Ms), b.) saturation remanence, c.) remanence coercivity (Hcr) versus coercivity (Hc), and d) Hc and Hcr versus Σlf.

clearly detected in the upper pedogenic horizon but could be possibly preserved in the lower part of the unit.

soil possibly due to the fast change of climate after the pedogenic period. That is, no long accretional period – development of the soil upward along with increasing sedimentation – can be recognized by the SD (gradually decreasing, instead of similar to the soil) and thickness (thick, instead of thin) of the transient horizon. The consistent vertical distribution of the SD ratio may indicate intensive pedogenesis in the entire body of the soil and/or vertical translocation of materials as suspension by soil water (precipitation) (Stoops et al., 2010). In the MIS11 (MB) paleosol, the high SD magnetic mineral content in the uppermost part of the soil indicates active nanoscale pedogenic mineral formation. No such ratio of pedogenic mineral component could be identified in the entire body of the paleosol unit (Fig. 4b; from 7 to 10) except in the lower transition, where the hysteresis parameters indicate a slight increase in SP content (Fig. 4d; 11). The appearance of the highest SD pedogenic mineral content may indicate i.) intense pedogenesis but ii.) less characteristic vertical mineral transportation downward, i.e., moderate precipitation (highest SD content located in the uppermost part of the paleosol). The appearance of superparamagnetic (e.g., bacterial) magnetic content in the lower transition may indicate i.) low evaporation during pedogenesis, i.e., SP (bacterial) minerals formed in situ in the constant capillary water or water film regime, which was preserved in the lower soil horizons due to the lack of strong evaporation, and/or ii.) nanosize, ‘ultra-sensitive’ grain sizedependent vertical transportation in the soil (‘larger’ SD magnetic grains stay in the upper horizon while smaller SP grains move downward). The formation of in situ ultrafine bacterial magnetic minerals in soils was suggested by Maher and Taylor (1988). The phenomenon of vertical translocation of fine materials (e.g., clay) suspended in infiltrating soil water is already known (Stoops et al., 2010). In the case of the MIS11 paleosol, the mark of in situ SP mineral formation cannot be

4.2. Theoretical pedogenic model Revealing the relationship between the hysteresis properties and Σlf indicated some irregularity compared to the results of Forster and Heller (1997), who performed similar study on some loess profiles (including the Paks loess) (Table 1; Fig. 5). The relationship between the increasing Ms. and Σlf only shows a weak linear relationship (R2 = 0.24) compared to the moderate relationship between Mrs. and Σlf (R2 = 0.52) and the ‘fairly linear’ results of Forster and Heller (1997) (Fig. 5a and b). Compared to the results of Forster and Heller (1997), who suggest that the susceptibility enhancement process is controlled by the increasing amount of ferromagnetic minerals with a constant Ms/Σlf and Mrs/Σlf during pedogenesis, the results show a weaker connection between Ms. and Σlf. The weakening relationship between Ms. and xlf could be the result of the scattering Ms. and Σlf of various paleosol horizons: MIS19 paleosol (PD2) is characterized by high Σlf but scattered Ms. The MIS15 paleosol (Mtp) showed significantly high Ms. (except for two samples, whose character was similar to that of the MB paleosol) and intermediate Σlf. The MIS11 paleosol (MB) was indicated by low Ms. and intermediate/low Σlf (Fig. 5a). These differences suggest different methods of magnetic enhancement during pedogenesis in the case of the studied paleosols. The linear relationship between the coercivity parameters Hc and Hcr (Fig. 5c) and the distribution of sediments and paleosols in the Hc, Hcr and xlf plots (Fig. 5d and e) indicated the changing ratio of sedimentary and pedogenic magnetic components during pedogenesis in the so-called two-component enhancement path (Forster and Heller, 8

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Fig. 6. Typical microfabrics of loess and paleosols from Paks and their characteristics. The mineral components in the fabric of sediments (a, b) with fragmentation, and submicron-scale nanofragments were observed around some coarser-grained magnetite (c, d) and paleosols (e, f). In the colour figures b and d, blue indicates quartz, red denotes minerals with a high Fe content (e.g., magnetite) and green indicates minerals with a lower Fe content (e.g., chlorite) (can be seen only in the online version). Abbreviations are as follows: Ap – apatite, Cal – calcite, Chl – chlorite, Dol – dolomite, Ep – epidote, Hem – haematite, Ilm – ilmenite, Ms. – muscovite, Kfs – potassium feldspar, Ky – kyanite, Qtz – quartz, Rt – rutile, Ttn – titanite. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

recognized weathering and clay mineral formation (Fig. 6c, d, e, and f). Unfortunately, because of the possible measurement errors, the appearance of the three-component model could not be clearly verified.

1997). That is, in loess, the magnetic enhancement of sediments and paleosols can be described by a (simplified) model of two different mixtures of magnetic minerals: the so-called original mixture (OM) appears in loess, which consists of coarser-grained multidomain magnetite and a ‘considerable proportion’ of higher-coercivity (e.g., haematite) magnetic components (Forster and Heller, 1997). The so-called enhanced mixture (EMp) represents the ‘final product of pedogenic enhancement process’ (Forster and Heller, 1997; 18 p). It is a mixture of magnetite and maghemite with a significant amount of ultra-fine-grained authigenic minerals developed by pedogenesis. The relationship between the coercivity parameters and xlf may reflect the appearance of the three-component enhancement path (Forster and Heller, 1997). In this enhancement model, the initial phase of pedogenesis (weathering, clay mineral formation) can also be recognized in sediments. In the case of the studied paleosols from Paks, both the twoand the three-component enhancement models were observed by Forster and Heller (1997), which may indicate various stages/types of pedogenesis in the paleosols. However, there are some marks of the appearance of the three-component enhancement path, including the

4.3. The character of pedogenesis and the interglacials Despite the differences in low-field magnetic susceptibility, the three studied paleosols are well developed and indicate an intense interglacial environment based on their pedogenic parameters observed in the field. The study of the parameters of pedogenic subhorizons provided deeper insight into pedogenesis and the paleosol-forming environment (Table 2). The pedogenic enhancement of the MIS11 soil indicates a stronger, more intense interglacial environment compared to that of MIS15 but a weaker environment than that of MIS19 (Fig. 5). The comparison of the magnetic parameters in the subhorizons of the soil indicates an active upper soil horizon. The observed increasing ultrafine (SP)-grain components in the lower transition horizon (Fig. 4a and b) may have multiple origins: 9

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Table 2 Some important components of the soil-forming environment and palaeoenvironment indicated by the studied magnetic parameters. Period

Paleoenvironmental interpretation ΔM

Day plot

Pedogenic enhancement - xlf vs. Hc (and Hcr) plot

MIS11 (MB paleosol)

Limited vertical movements; mixing of the pedogenic and transition horizon and/or fast burying (similar curve of transition and soil horizons).

Strong interglacial (but weaker than MIS19). Balance (seasonal?) between the precipitation, water infiltration and evaporation, moderately humid environment.

MIS15 (Mtp) paleosol

In situ formed pedogenic component (well separated curves of sediments and soils); significant vertical pedogenic processes (‘pedogenic’ characteristics of ΔM increasing downward) Similar enhancement processes of the magnetic components in various depth of the soil, intense (and fast) pedogenesis (very similar shape of the soil curves).

Active upper soil horizon; mineral forming; grain size dependent vertical transp. and/or in situ formed minerals in lower transition; moderate precipitation (increasing and highest SD in the upper subunits of the soil and slightly increase of SP in lower transition). Intense vertical material movement, precipitation (consistent SD content in the soil unit); and fast change of climate after the pedogenic period.

MIS19 (PD2)

Intense weathering in every pedogenic horizon, highest precipitation (most intense compare to Mtp and MB); indicate some abrupt climatic change, decreasing precipitation and possibly increasing sediment input (an upper, less-developed, and a lower, well-developed part)

Stronger interglacial, but weaker than MIS11 or MIS19

Strongest interglacial among the studied interglacials (MIS19, MIS15 and MIS11).

MIS19 can be characterized by the most intense paleosol-forming environment among the studied interglacials, as indicated by the highest amount of SD new-forming pedogenic mineral components (Fig. 4b). The paleoclimate of MIS19 supported continuous weathering during the formation of the paleosol, which resulted in similar rock magnetic characteristics in various subhorizons of the soil (Fig. 3a). Compared to the MIS15 and MIS11 paleosols, the upper horizon of the soil and the upper transition horizon indicate a continuous interglacial/ glacial transition. The interaction between sedimentary and pedogenic processes can be interpreted as a dynamic system in which the balance between accumulation and pedogenesis can determine the character of a paleosol (Kraus, 1999). Increasing sedimentation at the end of an interglacial, in balance with pedogenesis, can form a cumulic horizon (Retallack, 2001) (similar to cumulative paleosol, Kraus, 1999). Due to its mixed nature, the cumulic horizon is less compact and massive compared to the lower clayey pedogenic subhorizons and possibly (a part of it) can be eroded easily, e.g., by abrasion of stronger wind or water lain processes. Despite the appearance of potential erosion, the upper transient (cumulic) horizon may be used with some criterion as an indicator of interglacial/glacial transition. In the case of the studied paleosol, no marks of erosion, such as a sharp paleosol/sediment upper boundary and a redeposited, laminated sediment horizon overlying the paleosol, could be identified. Therefore, the character of the upper transient horizon of the MIS19, MIS15 and MIS11 paleosols may indicate the character of the interglacial/glacial transition. The continuous MIS19/MIS18 interglacial/glacial transition is more characteristic of the climate before the periodicity switch of MPT. Compared to the continuous transition, the MIS15/MIS14 and MIS11/MIS10 transitions indicate more abrupt changes. The differences are not limited to the character of the transition phases. MIS19 is characterized by a very intense pedogenic environment, possibly with a significant amount of precipitation and a warm environment, which supports very intense weathering. The palaeoclimate of MIS15 and MIS11 also favours strong pedogenesis, but with less (decreasing) precipitation and possibly more moderate temperature. The observation of some significant change in the pedogenic and palaeoenvironment agrees with the results of Buggle et al. (2014), who find a characteristic change in the Middle Pleistocene climate (from a Mediterranean-type climate towards a steppe-like climate) around the same time frame, and Heslop et al. (2002), who identified the periodicity switch between ~850 and 680 ka in the loess successions from the Chinese Loess Plateau.

i.) grain size-sensitive mineral translocation from the upper horizon (in contrast to the accumulation of SD components in the upper horizon) (e.g., Stoops et al., 2010); ii.) non-pedogenic in situ mineral formation; iii.) Mahowald et al. (2014) suggested a broad grain size range for atmospheric dust (< 1 nm up to 100 μm). The dust was transported from, e.g., the Sahara, and can appear in the paleosols of the Carpathian Basin during certain (palaeo)climatic conditions (Varga et al., 2016). iv.) the (physical) fragmentation of the coarser grains (Fig. 6c and d), e.g., due to some root activity from the soil. The described phenomena may indicate vertical migration that is not intense and is possibly more like seasonal than continuous migration. Water infiltration and vertical migration can change mineral fabrics and material migration in the soil body, as observed in welldeveloped MIS5e paleosols from the Cérna Valley loess profile (BradákHayashi et al., 2016), but no one mark of vertical movements could be identified in the studied soil (Bradák et al., 2018a). Relatively low precipitation and/or limited water infiltration did not trigger mineral reorientation and migration in the studied paleosols. In addition, enrichment of pedogenic magnetic components is found only in the upper part of the paleosol, which excludes intense and fast weathering in the entire body of the paleosol. Based on the results, seasonal increases in precipitation and moderate water infiltration and vertical pedogenic processes are expected during MIS11. The similarities between the ΔM curves of the paleosol and transient horizons (Fig. 3a) indicate mixing by biogenic activity (e.g., roots) and/or fast burying (no separable character of the transient horizon - no continuous soil aggradation and increasing of sedimentary character) of the soil after the interglacial and abrupt interglacial-to-glacial transition. The suggested abrupt, ‘saw-tooth’-like change is one of the characteristics of post-MPT interglacial/glacial transitions (Maslin and Brierly, 2015). Although the pedogenic enhancement of MIS15 turned out to be the weakest among the studied three paleosols, the character and the rock magnetic parameters indicate intense pedogenic processes and an intense interglacial. SD mineral formation during pedogenesis is indicated by the clear separation of paleosol and sediment curves in the ΔM plot (Fig. 3b) and the Day plot (increasing SD components in paleosols; Fig. 4a and c). The highest concentration of the pedogenic SD components was identified in the lower part of the soil, which possibly indicates intense vertical processes, possibly derived from a significant amount of precipitation and water infiltration. The transient and paleosol horizons cannot be separated well by the rock magnetic parameters (Fig. 3b), which possibly indicates fast and abrupt climate cooling for the same reason as it was described for the MIS11 paleosol. 10

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Table 3 Changes in the characteristics of the interglacial palaeoclimate recorded in the paleosols of loess profiles during the early Middle Pleistocene in Central and Eastern Europe. Region

Profile

Timeframe

Character of paleoclimate/environment change

References

Morava

Červený Kopec

MIS13 to MIS11

Kukla (1977); Forster et al. (1996);

NW East European Plain (EEP)

Korolevo Zahvizdja Paks Stari Slankamen

MIS15 to MIS13

Forest (moderate, humid) to grassland and forest (continental, steppe) Moderet, humid environment (MIS19-MIS15, forest) via warm (MIS13) towards MIS11 (sub-)Mediteranean to moderate, humid (decidous forest) sub-Mediterranean to continental

MDB

S MDB LDB

SW EEP

Batajnica Batajnica (2) Stalać Zimnicea Lubenovo Viatovo Koriten Novaya Etuliya Roxolany

MIS17 to MIS15 after MIS11 MIS13 to MIS11 after MIS11 MIS13 to MIS11 after MIS11 after MIS11

Moderate, humid (NW Europe-like) to continental sub-Mediterranean to continental Mediterranean to continental (sub Mediterranean) forest to (continental) steppe

MIS13 to MIS11

Mediterranean to continental (steppe)

Nawrocki et al. (2016) Nawrocki et al. (2002, 2016) In this study Marković et al. (2011); Marković et al. (2015) Kostić and Protić (2000) Marković et al. (2009) Kostić and Protić (2000) Rădan (2012) Jordanova et al. (2007, 2008)

Tsatskin et al. (2008) Tsatskin et al. (1998); Gendler et al. (2006)

series of (synchronous) forgoing events starting in MIS14 led to the MBT (MIS12/MIS11). These synchronous events may cause differences at the regional level, e.g., in the characteristics of an interglacial and thus in the way of pedogenesis. The pedogenic model of Finke et al. (2018) described that some components, such as the length of the interglacial, accumulative precipitation surplus and weaker dust deposition (also a key component in the model of Barth et al., 2018), play a significant role in pedogenesis, and well-developed soils can form even if the palaeotemperature (‘peak warm’) and the precipitation are less strong compared to other interglacials. Along with climatic components such as temperature and precipitation, these components could cause differences in the degree of pedogenesis, especially during the MBT when the regional differences were possibly strengthened due to synchronous events (suggested by Barth et al., 2018) and their global differences. The regional differences in the conditions of the palaeoclimate and pedogenesis may provide some explanation of the differences between, e.g., the character of paleosols and the interglacial intensities from the CLP (Guo et al., 2009; Lu et al., 2018) and the ELB (e.g., Paks and Batajnica). However, they cannot describe the significant differences between the referred Central and Eastern European loess sections, which are relatively close to each other and show different rhythms in the change of the pedogenic environment. Based on the identification of the changing points of (sub-) Mediterranean to continental climate in various profiles and their evolution in time (Table 3), a theoretical model can be built (Fig. 7). This theoretical model is based on the model of Buggle et al. (2013), and as some improvement, it proposes the integration of some sections from East-Central Europe (Červený Kopec) and the northwestern part of the East European Plain (EEP) (Korolevo and Zahvizdja). By integrating the suggested sections, the evolution of the changing pedogenic environment in Paks may be better understood. Based on the paleosols of the studied profiles (Table 3), MIS19 and MIS17 can be characterized by moderate, humid climate in the northern part of Central Europe and the northwestern part of the EEP. The Middle Danube Basin (MDB) (e.g., Paks, Stari Slankamen and Batajnica), Lower Danube Basin (LDB) (e.g., Korintien, Lubenovo and Viatovo) and the southwestern part of the EEP (Novaya Etuliya and Roxolany) climates were (sub)Mediterranean (Fig. 7a). The paleosols of MIS15 did not show significant change in the palaeoclimate setting, except in the Carpathian basin, where the strengthening of humid and cooler climate was recognizable (Fig. 7b). Following the MBT, significant changes occurred in the palaeoclimate of the region around the MIS11. The dominance of the Mediterranean and possibly Atlantic climate gradually changed to a

4.4. Proposal for a theoretical climate model in the Central/Eastern European region Similar to other loess profiles in Europe and East Asia, the potential marks of the MPT and MBT have also been found in the Paks profile. Buggle et al. (2008) and Marković et al. (2009) identified characteristic differences in pedogenesis before and after MIS11. The study of the Middle to Late Pleistocene paleosols of the Mircea Voda loess succession (southeastern Romania, Buggle et al., 2008) and Batajnica succession (Serbia, Marković et al., 2009) indicated the strengthening of rubification values in ‘older’ MIS15 to MIS11 paleosols. The strengthening of rubification indices was interpreted as a mark of sub-Mediterranean climate influence. The results indicate a characteristic palaeoclimate switch from sub-Mediterranean to steppe-like following the MBT in southeastern Central Europe. Buggle et al. (2013, 2014) (see above) agreed with Buggle et al. (2008) and Marković et al. (2009) and identified the change of palaeoclimate parameters from Mediterranean-type to more steppe-like during the (early) Middle Pleistocene in both the Middle and the Lower Danube Basin. Buggle et al. (2013) proposed the significance of the intensification of tectonic activity in the Alpine Orogenic Belt as a trigger of the gradual change of palaeoclimatic characteristics. At the Novaya Etuliya (Moldova) and Roxolany sections (Ukraine), a similar switch can be identified between the paleosols formed around the MBB and younger ones formed in the Brunhes chron (Gendler et al., 2006; Tsatskin et al., 1998, 2008). Compared to Stari Slankamen and Mircea Voda, the palaeoclimatic transition from subMediterranean to continental climate occurred after a gradual cooling around the MIS11 (following the MBT and the MIS11). The timing of the abrupt palaeoclimatic change in Novaya Etuliya and Roxolany was between MIS13 and MIS11. Based on the study of Chinese loess elaborated by Heslop et al. (2002), the switch between the 41- and 100-kyr periodicity of Pleistocene climate cycles, i.e., the actual transition of MPT, occurred during the period of MIS19 and MIS17 paleosol formation. Based on this timeframe, theoretically, the change of the pedogenic environment in Paks can link to the MPT, but more evidence that supports this hypothesis is needed. The climatic transition from sub-Mediterranean towards a cooler, steppe-like environment identified in, e.g., the Mircea Voda and Batajnica successions around MIS11, may relate to the end of the MBT. However, the description of the time gap between the palaeoclimatic change in regional neighbourhood sections such as Paks (around MIS17/MIS16), Novaya Etuliya and Roxolany (around MIS13/ MIS11) and Mircea Voda and Batajnica (after MIS11) seems more complex. The climate study and model of Barth et al. (2018) suggests that a 11

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Fig. 7. Theoretical model for the change in the characteristic climate influence between MIS19 and MIS11 in the Middle and Lower Danube Basin and the western part of the East European Plain. Profiles: ČK - Červený Kopec (Red Hill), Czech Republic (Forster et al., 1996); Ki – Koriten, Bulgaria (Jordanova and Petersen, 1999); Ko - Korolevo, Ukraine (Nawrocki et al., 2016); Lu - Lubenovo (Jordanova et al., 2007, 2008); NE – Novaya Etuliya (45.51N, 28.41E), Ukraine (Gendler et al., 2006); Pa - Paks, Hungary (Sartori et al., 1999; Újvári et al., 2014); Ro – Roxolany, Ukraine (Gendler et al., 2006); SP – Skala Podilska (Boguckyj et al., 2009); St – Stalać (Kostić and Protić, 2000); SS - Stari Slankamen, Serbia (Marković et al., 2011); Vi – Viatovo, Bulgaria (Jordanova et al., 2008, Za - Zahvizdja, Ukraine (Nawrocki et al., 2002); Zi – Zimnicea, Romania (Rădan, 2012). The lines indicate the expected areas of the humid and moderate (dark grey/blue line - online version), (sub) Mediterranean-like (black/red line - online version) and cold and drier (light grey/green line - online version) climate influence. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

very intense weathering and pedogenesis and ii.) a moderate temperature and humid post-MPT interglacial environment (MIS15 and MIS11) with significant but less intense pedogenesis compared to MIS19. Different characteristics of the interglacial/glacial transition were also observed: a continuous transition was identified during the MIS19/ MIS18 change. In contrast, the transition between MIS15/MIS14 and MIS11/MIS10 was abrupt and fast. This conclusion can be treated with reservations due to the possibility of erosion. The evolution of the changing point of a (sub-)Mediterranean to continental climate in time may indicate regional differences in the increasing influence of continental airmasses in Eastern and Central Europe, but more information is needed to reveal its nature. The suggested differences describe the differences between the magnetic susceptibility of the three similar-looking, well-developed paleosols and indicate characteristic changes in the palaeoclimate between MIS19 and MIS15. These differences possibly relate to the periodicity switch and global cooling during the MPT, but more detailed research of the profile is needed to verify the results of this theory. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.palaeo.2019.109335.

continental, steppe-like environment. The climate of the northern part of Central Europe and the western part of the EEP changed in the ‘first wave’ during the period from MIS13 to MIS11 (Fig. 7c).The ‘second wave’ of climate change was identified after MIS11 in the profiles of the MDB and LDB (Fig. 7d). The Paks and Zahvizdja profiles (Nawrocki et al., 2002) represent unusual cases. In the case of Paks, instead of two (such as humid/cool/Mediterranean and continental), three characteristic climate changes could be recognized: Mediterranean (MIS19 and MIS17), moderate and cooler climates with forest (MIS15 and MIS11) and continental (grassland/forest) environments (after MIS11). In the case of the Zahvizdja profile, there is a relatively warm period ‘wedged’ between the cooler humid and continental climate period.

5. Conclusions Various paleosol-forming environments have been revealed by the detailed rock magnetic (magnetic hysteresis) study of the well-developed paleosols of the Paks loess succession, one of the key profiles of the ELB. The study of various pedogenic subhorizons and the comparison of the vertical change in the parameters provided complex information about the palaeoclimate. Two characteristic pedogenic environments were identified: i.) warm and humid (pre-)MPT interglacial environment (MIS19) with 12

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Acknowledgements

Head, M.J., Gibbard, P.L., 2005. Early-Middle Pleistocene transitions: an overview and recommendation for the defining boundary. Geol. Soc. Lond., Spec. Publ. 247, 1–18. Heslop, D., Dekkers, M.J., Langeresi, C.G., 2002. Timing and structure of the midPleistocene transition: records from the loess deposits of northern China. Palaeogeography, Pataeoclimatology, Palaeoecology 185, 133–143. Jansen, J.H.F., Kuijpers, A., Troelstra, S.R., 1986. A mid-Brunhes climatic event: longterm changes in global atmosphere and ocean circulation. Science (4750), 619–622. Jordanova, D., Petersen, N., 1999. Palaeoclimatic record from a loess–soil profile in northeastern Bulgaria—I. Rock magnetic properties. Geophys. J. Int. 138, 520–532. Jordanova, D., Hus, J., Geeraerts, R., 2007. Palaeoclimatic implications of the magnetic record from loess/palaeosol sequence Viatovo (NE Bulgaria). Geophys. J. Int. 171, 1036–1047. Jordanova, D., Hus, J., Evgoliev, J., Geeraerts, R., 2008. Paleomagnetism of the loess/ paleosol sequence in Viatovo (NE Bulgaria) in the Danube Basin. Phys. Earth Planet Inter. 167, 71–83. Kostić, N., Protić, N. 2000. Pedology and mineralogy of loess profi les at Kapela-Batajnica and Stalać, Serbia: Catena, v. 41, p. 217–227, doi:10.1016/S0341-8162(00)00102-8. Kraus, M.J., 1999. Paleosols in clastic sedimentary rocks: their geologic applications. Earth-Science Reviews 47, 41–70. Kukla, G.J., 1977. Pleistocene land–sea correlations. Earth-Sci. Rev. 13, 307–374. Kusiak, J., Łanczont, M., Madeyska, T., Bogucki, A., 2013. Problems of TL dating of the Mesopleistocene loess deposits in the Podillya and Pokuttya regions (Ukraine). Geochronometria 40, 51–58. Lu, H., Jia, J., Wang, Y., Yin, Q., Xia, D., 2018. The cause of extremely high magnetic susceptibility of the S5S1 paleosol in the central Chinese Loess Plateau. Quat. Int. 493, 252–257. Maher, B.A., Taylor, R.M., 1988. Formation of ultrafine-grained magnetite in soils. Nature 336, 368–370. Mahowald, M., Albani, S., Kok, J.F., Engelstaeder, S., Scanza, R., Ward, D.S., Flanner, M.G., 2014. The size distribution of desert dust aerosols and its impact on the Earth system. Aeolian Res. 15, 53–71. Marković, S.B., Hambach, U., Catto, N., Jovanović, M., Buggle, B., Machalett, B., Zöller, L., Glaser, B., Frechen, M., 2009. The middle and late Pleistocene loess–paleosol sequences at Batajanica, Vojvodina. Serbia. Quat. Int. 198, 255–266. Marković, 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 loess–palaeosol sequence: revised chronostratigraphy and long-term environmental trends. Quat. Sci. Rev. 30, 1142–1154. Marković, S.B., Stevens, T., Kukla, G.J., Hambach, U., Fitzsimmons, K.E., Gibbard, P., Buggle, B., Zech, M., Guo, Z., Hao, Q., Wu, H., O'Hara, Dhand K., Smalley, I., Újvári, G., Sümegi, P., Timar-Gabor, A., Veres, D., Sirocko, F., Vasiljević, D.A., Jary, Z., Svensson, A., Jović, V., Lehmkuhl, F., Kovács, J., Svirčev, Z.., 2015. Danube loess stratigraphy — towards a pan-European loess stratigraphic model. Earth-Science Rewievs 148, 228–258. Maslin, M.A., Brierly, C.M., 2015. The role of orbital forcing in the Early Middle Pleistocene Transition. Quat. Int. 389, 47–55. Nawrocki, J., Bogucki, A., Łanczont, M., Nowaczyk, N.R., 2002. The Matuyama-Brunhes boundary and the nature of magnetic remenence acquisition in the loess-paleosol sequence from western part of the East European loess province. Palaeogeogr. Palaeoclimatol. Palaeoecol. 188, 39–50. Nawrocki, J., Łanczont, M., Rosowiecka, O., Bogucki, C., 2016. Magnetostratigraphy of the loess-palaeosol key Palaeolithic section at Korolevo (Transcarpathia, W Ukraine). Quaternary International 399, 72-85Necola, C., Panaiotu, C. 2012. Rock magnetic properties of a loess-paleosols complex from Mircea Voda (Romania). Romanian Reports in Physics 64, 516–527. Necula, C., Panaiotu, C., 2012. Rock magnetic properties of a loess-paleosols complex from Mircea Voda (Romania). Romanian Rep. Phys. 64/2, 516–527. Past Interglacials Working Group of PAGES, 2016. Interglacials of the last 800,000 years. Rev. Geophys. 54, 162–219. https://doi.org/10.1002/2015RG000482. Pécsi, M., 1995. The role of principles and methods in loess-paleosol investigations. GeoJournal 36, 117–131. Rădan, S.C., 2012. Towards a historical synopsis of dating the loess from the Romanian Plain and Dobrogea: authors and methods through time. GeoEcoMar 18, 153–172. Retallack, G.J., 2001. The Soils of the Past. Blackwell Science, An introduction to paleopedology (404 p). Roberts, A.P., 2015. Magnetic mineral diagenesis. Earth Sci. Rev. 151, 1–47. Roberts, A.P., Almeida, T.P., Church, N.S., Harrison, R.J., Heslop, D., Li, Y., Li, J., Muxworthy, A.R., Williams, W., Zhao, X., 2017. Resolving the origin of pseudo-single domain magnetic behaviour. Journal of Geophysical Research: Solid Earth 122, 9534–9558. https://doi.org/10.1002/2017JB014860. Sartori, M., Heller, F., Forster, T., Borkovec, M., Hammann, J., Vincent, E., 1999. Magnetic properties of loess grain size fractions from the section at Paks (Hungary). Phys. Earth Planet. Inter. 116, 53–64. Stoops, G., Marcelino, V., Mees, F., 2010. Interpretation of Micromorphological Features of Soils and Regoliths. Elsevier (720 p). Tauxe, L., Mullender, T.A.T., Pick, T., 1996. Potbellies, wasp-waists, and superparamagnetism in magnetic hysteresis. J. Geophys. Res. 101, 571–583. Tauxe, L., Butler, R. F.,Van der Voo, R., and Banerjee S. K. 2010. Essentials of Paleomagnetism. Berkeley: University of California Press. (512 p). Tsatskin, A., Heller, F., Hailwood, E.A., Gendler, T.S., Hus, J., Montgomery, P., Sartori, M., Virina, E.I., 1998. Pedosedimentary division, rock magnetism and chronology of the loess/palaeosol sequence at Roxolany (Ukraine). Palaeogeogr. Palaeoclimatol. Palaeoecol. 143, 111–133. Tsatskin, A., Gendler, T.S., Heller, F., 2008. Improved paleopedological reconstruction of vertic paleosols at Novaya Etuliya, Moldova via integration of soil micromorphology and environmental magnetism. In: Kapur, S., Stoops, G. (Eds.), New Trends in Soil

We thank Slobodan Marković and the anonymous reviewer for their advice and help to improve the manuscript. B. Bradák acknowledges the financial support of project BU235P18 (Junta de Castilla y Leon, Spain) and the European Regional Development Fund (ERD). A fellowship was awarded to B. Bradák at Kobe University, Japan, by the Japan Society for the Promotion of Science (JSPS; P15328) during the period of 2015.10 – 2017.10. We thank Professor Masayuki Hyodo (Kobe University, Japan) for his advice during the JSPS fellowship and Kochi University (Japan) for facilitating the hysteresis measurements. References Bakhmutov, V.G., Kazanskii, A.Yu., Mtasova, G.G., Glavatskii, D.V., 2017. Rock Magnetism and Magnetostratigraphy of the Loess-Sol Series of Ukraine (Roksolany, Boyanychi, and Korshev Sections). Izvestiya, Physics of the Solid Earth 53, 864–884. Barth, A.M., Clark, P.U., Bill, N.S., He, F., Pisias, N.G., 2018. Climate evolution across the Mid-Brunhes transition. Clim. Past 14, 2071–2087. https://doi.org/10.5194/cp-142071-2018. Boguckyj, A.B., Łanczont, M., Łącka, B., Madeyska, T., Nawrocki, J., 2009. Quaternary sediment sequence at Skala Podil'ska, Dniester River basin (Ukraine): preliminary results of multi-proxy analyses. Quat. Int. 198, 173–194. Bradák, B., Újvári, G., Seto, Y., Hyodo, M., Végh, T., 2018a. A conceptual magnetic fabric development model for the Paks loess in Hungary. Aeolian Res. 30, 20–31. Bradák, B., Seto, Y., Hyodo, M., Szeberényi, J., 2018b. Significance of ultrafine grains in the magnetic fabric of paleosols. Geoderma 330, 125–135. Bradák-Hayashi, B., Biró, T., Horváth, E., Végh, T., Csillag, G., 2016. New aspects of the interpretation of the loess magnetic fabric, Cérna Valley succession, Hungary. Quat. Res. 86, 348–358. Buggle, B., Hambach, U., Glaser, B., Marković, S.B., Glaser, I., Zöller, L., 2008. Long term paleoclimate records in SE-Europe — the loess paleosol sequences Batajnica/ StariSlankamen (Serbia) and Mircea Voda (Romania). In: Hauptversammlung der DEUQUA,Wien, 31 August – 6 September 2008, Abhandlungen der Geologischen Bundesanstalt, Band. 62. pp. 15–19. Buggle, B., Hambach, U., Kehl, M., Marković, S.B., Zöller, L., Glaser, B., 2013. The progressive evolution of a continental climate in southeast-central European lowlands during the Middle Pleistocene recorded in loess paleosol sequences. Geology 41, 771–774. Buggle, B., Hambach, U., Müller, K., Zöller, L., Marković, S.B., Glaser, B., 2014. Iron mineralogical proxies and Quaternary climate change in SE-European loess–paleosol sequences. Catena 117, 4–22. Candy, I., McClymont, E.L., 2013. Interglacial intensity in the North Atlantic over the last 800,000 years: investigating the complexity of the mid-Brunhes Event. J. Quat. Sci. 28, 343–348. Candy, I., Coope, G.R., Lee, J.R., Parfitt, S.A., Preece, R.C., Rose, J., Schreve, D.C., 2010. Pronounced warmth during early Middle Pleistocene interglacials: investigating the Mid-Brunhes Event in the British terrestrial sequence. Earth Sci. Rev. 103, 183–196. Day, R., Fuller, M., Schmidt, V.A., 1977. Hysteresis properties of titanomagnetites: grain size and composition dependence. Phys. Earth Planet. Inter. 13, 260–267. Deng, C., Zhu, R., Verosub, K.L., Singer, M.J., Vidic, N. J. 2004. Mineral magnetic properties of loess/paleosol couplets of the central loess plateau of China over the last 1.2 Myr, J. Geophys. Res., 109, B01103; doi:https://doi.org/10.1029/ 2003JB002532. Dunlop D. J. 2002. Theory and application of the Day plot (Mrs/Ms versus Hcr/Hc) 2. Application to data for rocks, sediments, and soils. J. Geophys. Res. 107:No. B3, 2057, https://doi.org/10.1029/2001JB000487, 2002. Evans, M.E., 2001. Magnetoclimatology of aeolian sediments. Geophys. J. Int. 144, 495–497. Finke, P.A., Yin, Q., Bernardini, N.J., Yu, Y., 2018. Climate-soil model reveals causes of differences between Marine Isotope Stage 5e and 13 paleosols. Geology 46, 99–102. Forster, T., Evans, M.E., Heller, F., 1994. The frequency dependence of low field susceptibility in loess sediments. Geophys. J. Int. 118, 636–642. Forster, T., Heller, F., Evans, M.E., Havlicek, P., 1996. Loess in the Czech Republic: magnetic properties and paleoclimate. Stud. Geophys. Geod. 40, 243–261. Forster, T.H., Heller, F., 1997. Magnetic enhancement path in loess sediments from Tajikistan, China and Hungary. Geophysical Research Letters 24/1, 17–20. Gendler, T.S., Heller, F., Tsatskin, A., Spassov, S., Pasquier, J.D., Faustov, S.S., 2006. Roxolany and Novaya Etuliya—key sections in the western Black Sea loess area: Magnetostratigraphy, rock magnetism, and paleopedology. Quat. Int. 152–153, 78–93. Guo, Z.T., Berger, A., Yin, Q.Z., Qin, L., 2009. Strong asymmetry of hemispheric climates during MIS-13 inferred from correlating China loess and Antarctica ice records. Clim. Past 5, 21–31. www.clim-past.net/5/21/2009/. Haase, D., Fink, J., Haase, G., Ruske, R., Pécsi, M., Richter, H., Altermann, M., Jäger, K.D., 2007. Loess in Europe—its spatial distribution based on a European Loess Map, scale 1:2,500,000. Quat. Sci. Rev. 26, 1301–1312. Hao, Q.Z., Wang, L., Oldfield, F., Peng, S.Z., Qin, L., Song, Y., Xu, B., Qiao, Y.S., Bloemendal, J., Guo, Z.T., 2012. Delayed build-up of Arctic ice sheets during 400,000-year minima in insolation variability. Nature 490 (7420), 393–396. https:// doi.org/10.1038/nature11493.

13

Palaeogeography, Palaeoclimatology, Palaeoecology 534 (2019) 109335

B. Bradák, et al.

Carpathian Basin and its possible effects on interglacial soil formation. Aeolian Res. 22, 1–12. Yin, Q., 2013. Insolation-induced mid-Brunhes transition in Southern Ocean ventilation and deep-ocean temperature. Nature 494, 222–225.

Micromorphology. Springer-Verlag, Berlin Heidelberg, pp. 91–110 (276p). Újvári, G., Varga, A., Raucsik, B., Kovács, J., 2014. The Paks loess-paleosol sequence: a record of chemical weathering and provenance for the last 800 ka in the midCarpathian Basin. Quat. Int. 319, 22–37. Varga, Gy., Cserháti, Cs, Kovács, J., Szalai, Z., 2016. Saharan dust deposition in the

14