Rock magnetism and pedogenetic processes in Luvisol profiles: Examples from Central Russia and Central Mexico

ARTICLE IN PRESS

Quaternary International 156–157 (2006) 212–223

Rock magnetism and pedogenetic processes in Luvisol profiles: Examples from Central Russia and Central Mexico Jorge Rivasa,, Beatriz Ortegab, Sergey Sedovc, Elizabeth Solleiroc, Svetlana Sycherad a

Universidad Nacional Autonoma de Mexico, Posgrado en Ciencias de la Tierra, Ciudad Universitaria, 04510 Mexico D.F., Mexico b Universidad Nacional Autonoma de Mexico, Instituto de Geofisica, Ciudad Universitaria, 04510 Mexico D.F., Mexico c Universidad Nacional Autonoma de Mexico, Instituto de Geologia, Ciudad Universitaria, 04510 Mexico D.F., Mexico d Russian Academy of Sciences, Institute of Geography. Staromonetny per. 29, 119017 Moscow, Russia Available online 30 June 2006

Abstract Despite a vast literature concerning magnetic properties of loess sequences, we still do not fully understand how magnetic components and properties are related to particular soil-forming processes that can vary with each type of genetic horizon. In order to establish the role of lithogenic factors in the link between magnetic properties and soil-forming processes, we carried out a study of two complete profiles of buried interglacial Luvisols, one formed in loess in Russia (Alexandrovsky quarry) and the other in volcaniclastics in Mexico (Barranca Tlalpan). In both profiles, soil genetic horizons have contrasting differences of their magnetic properties. In the Alexandrovsky profile, the magnetic susceptibility (w) is enhanced in the paleosol compared to parent material. In the Barranca Tlalpan sequence, w enhancement is absent in the soil profile. Increase of fine-grained magnetic components in the soil is attributed to neoformed minerals. However, this process cannot compensate for the loss of lithogenic magnetic minerals in any of the genetic horizons, and the resulting trend is w depletion in the whole soil profile. The pedogenic environment of eluvial horizons in both Luvisols is destructive to all magnetic components, both primary and secondary. Higher concentrations of antiferromagnetic components (hematite and goethite) found in E horizons are related to redoximorphic processes. r 2006 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction Since the discovery of differences in magnetic susceptibility values between paleosols and loess layers in Chinese loess sequences (Heller and Liu, 1982, 1986; Kukla and An, 1989; Maher et al., 2003; Tang et al., 2003), a surge of interest has resulted in the emergence of a vast literature on this topic. Similar patterns of susceptibility depth variation, characterized by strong maxima in paleosol horizons, were found in geographically distant loess–paleosol sequences in Central Asia, Europe, and North Africa (e.g. Dearing et al., 1996; Forster and Heller, 1997; Rousseau and Puisse´gur, 1998; Dodonov et al., 1999). These data were successfully used as a correlation tool, in particular to compare loess and marine records (including marine oxygen isotope curve) (Kukla and An, 1989; Heller and Evans, 1995; Evans et al., 2003). Corresponding author. Tel.: +52 55 56224226; fax: +52 55 55509395.

E-mail address: jorger@geofisica.unam.mx (J. Rivas).

At the same time, the phenomenon of susceptibility enhancement in loess-derived paleosols drew attention as a possible paleoenvironment indicator. It was widely accepted that the major part of enhancement is due to postdepositional pedogenic accumulation of fine-grained magnetic minerals (Zhou et al., 1990; Zheng et al., 1991; Maher and Thompson, 1992), and thus could depend upon paleoclimatic conditions of interglacials. Various approaches were developed to evaluate the proportion of pedogenic versus lithogenic magnetic components (Banerjee et al., 1993; Fine et al., 1995; Grimley et al., 1998). Another branch of research was focused on searching for dependence functions between susceptibility values and modern climatic conditions (in particular precipitation) in modern surface soil climosequences as a tool for comparison with susceptibility patterns in paleosols (Maher et al., 1994; Liu et al., 1995; Maher and Thompson, 1995). Although the pedogenic origin of enhancement was widely accepted, little effort was made to understand the

1040-6182/$ - see front matter r 2006 Elsevier Ltd and INQUA. All rights reserved. doi:10.1016/j.quaint.2006.05.007

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

interrelation between magnetic characteristics and paleosol pedogenesis in the early studies of magnetism in loess sequences. In most cases, researchers did not discuss sufficiently the kind of pedogenic horizons/features that were present in the buried soil profile, and in many studies, the sampling was accomplished mostly on equal depth basis without regard to soil horizonation. Furthermore, many studies failed to use a soil classification system, to fully characterize of the degree of development, or to note if the profile was truncated by erosion before burial or transformed by diagenesis after burial, although these phenomena are known to be frequent in loess sequences (Bronger et al., 1998). More interest towards the link between specific soilforming processes and magnetic properties of paleosols arose when a number of studies documented an inverse pattern of the magnetic susceptibility curve in some loess–paleosol sequences of Alaska (Bege´t et al., 1990; Vlag et al., 1999), Siberia (Chlachula et al., 1998; Evans et al., 2003), and South America (Orgeira et al., 1998; Nabel et al., 1999). Similarly, a curve with susceptibility minima in paleosols was established in the Pleistocene volcanic sequences of Mexico (Ortega-Guerrero et al., 2004). The reasons for this deviation were synthesized in two main viewpoints: (i) the pattern is controlled by variations of depositional processes such as change of wind intensity, source area, and presence of volcanic material (e.g. Bege´t et al., 1990; Evans et al., 2003; Schellenberger et al., 2003), or (ii) it is induced by gleization, a post-depositional pedogenic process occurring in a chemically reducing environment typical of water-logged soils. Chemical processes in a reducing environment are destructive to magnetic minerals and are responsible for susceptibility minima in paleosols (Liu et al., 1999, 2001), as well as in poorly drained modern soils (de Jong et al., 2000; Grimley and Vepraskas, 2000; Grimley et al., 2004). More extensive information about the interrelation between soil magnetism and type of pedogenesis was accumulated through detailed study of modern surface soils, which started earlier and independently from magnetic research of loess sequences (Le Borgne, 1955, 1960; Vadyunina and Babanin, 1972; Tite and Linington, 1975; Maher, 1998). Various works present datasets collected to establish the connection between the behavior of magnetic parameters and the variability of soil environments (e.g. Singer and Fine, 1989; Geiss et al., 2004) and soil age (Singer et al., 1992). Singer et al. (1996) proposed a conceptual model for susceptibility enhancement in soils that considers the mechanisms of pedogenic magnetic mineral neoformation and links it to the soil-forming factors climate and time. Maher (1998) developed a broader concept of interaction between pedogenesis and the magnetic mineral system, which includes possibilities of both susceptibility enhancement and ‘‘depletion’’. Despite these studies, there has been relatively little discussion about how magnetic components and properties

213

are related to particular soil-forming processes such as carbonate leaching/precipitation, type of weathering, clay eluviation/illuviation, humus accumulation, and redoximorphic processes in surface versus groundwater saturation conditions. These processes vary not only from soil to soil, but within soil profiles, generating different conditions for magnetic minerals. Another limitation for understanding the link between pedogenesis and soil magnetism is the lack of knowledge about rock magnetic properties of paleosols and modern soils formed on parent materials other than loess. With the aim to characterize Pleistocene paleosols developed in different environmental conditions and with variations in their parent materials, we initiated a multidisciplinary study of buried interglacial Luvisols (soils, known to be unequivocally indicative of subhumid forest ecosystems) in central Mexico and western Russia. The studied profiles were chosen on different parent materials (loess and volcanic ash) with the goal of establishing the role of lithogenic factors as well as soil-forming processes on magnetic properties. In this paper, we present a study of soil magnetic minerals as a function of particular pedogenetic environments within genetic horizons of Luvisol profiles. 2. Site descriptions 2.1. Barranca Tlalpan In central Mexico, a paleosol sequence called the Barranca Tlalpan (BT) profile is located in the Transmexican Volcanic Belt (191270 41.300 N; 981180 52.500 W, 2580 m a.s.l.) (Fig. 1a), and is comprised by a set of paleosols developed on Pliocene to Quaternary volcaniclastic deposits, locally known as ‘‘tepetates’’. Previous work on the Barranca Tlalpan described in detail the pedogenic and mineral magnetic characteristics of three major units of Pleistocene–Holocene volcaniclastics–paleosol sequences, classified as Luvisols (Ortega-Guerrero et al., 2004). In one of the youngest paleosols, the humus of A horizon was dated by 14C as 38,160 yr BP. In the oldest paleosol of this sequence, the ‘‘Red Unit’’ pedocomplex (formerly labeled P6 and P7), the only horizons found were Bt, BC, and C. The Bt horizon was subdivided into four subhorizons (Bt1, Bt2, Bt3, and Bt4) due to differences in soil properties. Bt1 and Bt2 exhibit in situ and thick clay cutans while Bt3 and Bt4 have reworked pedofeatures. In later field reconnaissance, the upper horizons Ah and EBtg were found in another exposure few hundred meters away. The Red Unit sequence was resampled from the two exposed profiles, labeled in this work as URU (at the upper part of pedocomplex profile, Ah, EBtg, and Btg horizons) and RU (at middle and bottom of the pedocomplex profile, Bt, BCt, and C horizons), at vertical intervals of nearly 30 cm (Fig. 1b and c). In URU, the unit shows a complete profile. The Ah horizon is dark gray, enriched in humus, and has

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

214

19° 30' Tlaxcala City 30

115

O ce a

00

fic

Barranca Tlalpan

I n

105

95

P

30

25

20

Gulf of Mexico

3500

MEXICO ci

3 000

Pa

25

M

00

10 km

19°00' 98° 00'

98° 40'

(a)

00

Horizons Section 0 1

Depth [m]

2

Ah EBtg

URU

URU

Bt1

Btg Bt1

Ah

Bt2

3 4

Bt3 Bt4

Btg

Bt3 Bt4

C

Bt1

6 (b)

Bt2

EBtg RU

BCt

5

RU

BCt C

(c)

Fig. 1. (a) Geographic location of the Barranca Tlalpan profile, near Tlaxcala city, in central Mexico. Popocatepetl (P), Iztaccihuatl (I), and Malinche (M) are the major stratovolcanoes in the area. (b) Pedocomplex of Upper Red Unit (URU) and Red Unit (RU) profiles. (c) Photos of the studied paleosol.

well-developed granular and subangular blocky structure. EBtg is more silty, lighter in color, with a weak subangular blocky structure, and redoximorphic features (pale-brown ferruginous mottles). It also exhibits thin clay cutans in voids and pores. The boundary with the underlying Bt horizon is tonguing, which is typical for Albic Luvisols. The Btg horizon also has abundant redoximorphic features such as gray and reddish-brown mottles, Fe–Mn concretions, and black Mn coatings on ped surfaces. No dates are available for the paleosol presented in this work, but according to its stratigraphic position, it was formed during the Middle Pleistocene. 2.2. Alexandrovsky The Alexandrovsky quarry site (AQ) is located in the Central Russian Highlands some 10 km south of Kursk (Fig. 2a). This elevated (200–260 m a.s.l.) part of the East European Plain was never covered by Pleistocene ice sheets. During glaciations, it formed part of an extensive periglacial tundra-steppe megazone, characterized by intensive cryogenesis and loess accumulation. A temperate environment during interglacials promoted soil formation. Currently loessic sediments with paleosols form a rather continuous mantle (Velichko, 1990), overlying pre-Qua-

ternary rocks. The contemporary environment is temperate forest-steppe, under which thick Chernozems are formed. The Alexandrovsky profile exposure is located in the upper near-watershed divide part of a gentle slope towards the valley of the Seim River and was extensively studied from a geomorphological and paleopedological standpoint by Sycheva (1998, 2004). In the cut of a brick quarry, we observed a buried fluvial geoform, a ‘‘balka’’ (a small valley) cut into loess of the Dnepr (Riss) glaciation (Fig. 2b and c). Both balka slopes are outlined by the buried interglacial paleosol of the Mikulino (Riss-Wu¨rm, Eemian) interglacial. This paleosol corresponds to the lower member of the Mezin pedocomplex, but in this paper, we will use the name Mikulino paleosol or soil to refer to it informally. In one place on the western side of the balka, an older Middle Pleistocene paleosol, consisting of AB and BCt horizons strongly deformed by cryogenic processes, was observed below the Dnepr loess. In contrast to the Mikulino soil, this earlier paleosol does not conform to the profile of the buried balka. It marks an older land surface that was buried by Dnepr loess and then locally cut by the balka slope. The depression was afterwards filled by Early Valday (Wu¨rm) colluvium that includes three weakly expressed

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

215

Fig. 2. (a) Geographic and stratigraphic position of the Alexandrovsky quarry (AQ), near Kursk, Russia; (b) photos of the studied paleosol; (c) schematic profile of the buried fluvial geoform balka, developed in the Dnepr loess.

paleosols formed during interstadials. Finally, this colluvial sequence is overlain by Middle-Late Valday strata including the Bryansk Paleosol, then loess that accumulated during the last glacial maximum; the latter is the parent material of the Holocene Chernozem. In the upper part of both slopes of the buried balka, both the AQ Mikulino paleosol and Early Valday colluvium pinch out. Consequently, the Bryansk Paleosol rests directly on Dnepr loess. The AQ Mikulino Interglacial Paleosol, which correlates to MIS 5, demonstrates a complete well-preserved sequence of genetic horizons, typical for an Albic Luvisol. For the rock magnetic study, we analyzed a pedocomplex profile with Ah-E-BEt-Bt(1-3)-BCt-Cg horizons, and additionally the AB and BCt horizons of the middle Pleistocene paleosol. 3. Methods Samples of 200 g were collected at roughly 20 cm intervals, homogenized, and placed in 8 cm3 acrylic boxes for magnetic measurements. One or two samples for each horizon were collected, 19 samples from URU and RU profiles and 17 in AQ profiles. All magnetic measurements were carried out with bulk samples. Magnetic susceptibility versus high-temperature experiments were used to determine the magnetic mineralogy by its Curie temperature, which is the temperature at which minerals loose their magnetization. Curie temperature estimations were performed on a Bartington MS2WF furnace system, in which changes in w were measured from

20 to 650 1C, during both heating and cooling in an air atmosphere. Samples with very low magnetic susceptibility were indurated with Omega CC high-temperature cement. Stepwise thermal demagnetization of remanences was conducted with a Thermal Specimen Demagnetizer set Model TD-1, between room temperature and 700 1C. Antiferromagnetic goethite has a Curie temperature at 80–120 1C. Pure, Ti-free magnetite has a Curie temperature close to 580 1C, while the content of Ti in titanomagnetites (Fe3
xTixO4, 0pxp1, represented as TM0–TM100) decreases the Curie temperature (Dunlop and O¨zdemir, 1997; McElhinny and McFadden, 2000). A Curie temperature of 150–200 1C is typical of TM60, which is the primary Timagnetite in rapid-cooled basaltic lavas, and 300 1C for TM45. Hematite, which has a Curie temperature of 675 1C, and goethite are difficult to observe in thermal demagnetization and susceptibility curves due to their low magnetic signal that is overshadowed by the stronger ferromagnetic Ti-magnetites. Magnetic susceptibility (w), which is a measure of the concentration of magnetic minerals, was measured in all samples at low (0.47 kHz) and high frequencies (4.7 kHz) with a Bartington MS2B dual sensor. We calculated frequency dependence of susceptibility wfd% as [(wlf–whf)/ wlf]100, to approximate possible ultrafine (o 0.05 mm) superparamagnetic (SP) contribution. When initial susceptibility was o30  10
5 SI units, samples were measured 10 times in a 0.1 scale, and average values were plotted. The natural remanent magnetization (NRM), anhysteric remanent magnetization (ARM), and isothermal remanent

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

Susceptibility 10-5 [S.I.]

magnetization (IRM) were measured with a Molspin Minispin magnetometer. ARM was imparted in a 50 mT bias field, superimposed on a peak alternating field of 100 mT in a Molspin AF demagnetizer. IRM was imparted with a pulse magnetizer at a forward field of 1 T and at backward fields of 100 and 300 mT. The magnetization acquired at 1 T was considered the saturation magnetization (SIRM). Several magnetic ratios were calculated, as indicators of magnetic hardness and stability. ‘‘Hard’’ IRM (HIRM) is obtained by imparting a back field at 300 mT (IRM
300) on a sample previously given SIRM. It is a tool to estimate the concentration of antiferromagnetic minerals with higher coercivity, or very fine-grained ferrimagnetic grains, and was calculated as [SIRM+IRM
300]/2 (Opdyke and Channell, 1999). The S-ratios are useful to estimate the presence of minerals with high coercivity, commonly hematite or goetithe, or high stability minerals such as fine (40.05 mm) single domain (SD) magnetite. They were calculated as Sx ¼ IRMx/SIRM, with x as the field applied, in this case we used 100 and 300 mT in a backward field. S100 display grain size variations in low coercivity components among coarse-grained ferromagnetic particles, although it cannot be distinguished from mixtures of a high-coercivity and a fine-grained low-coercivity fraction (Robinson, 1986). S300 show the proportion of lowcoercivity minerals in a sample (Opdyke and Channell, 1999), where values close to zero indicate antiferromagnetic concentration.

120

BCt horizon

60

0 0

100

(a)

200

300

400

500

600

To estimate grain size, we used ARM/SIRM to estimate the concentration of SD grains contribution parameter (Geiss, 1999). wfd% is a sensitive parameter of the presence of SP grains (Jordanova et al., 1997), but it is too dependent on titanium present in Ti-magnetites (Wall and Worm, 2000). Saturation magnetization (Ms) and coercivity (Hc) were obtained from hysteresis loops measured with a Molspin Vibrating Sample Magnetometer (VSM), at room temperature in all samples. Coercivity of remanence, Hcr, was estimated by demagnetization of SIRM applying IRM backfields. Both Hc and Hcr are useful to estimate harder grains. As the para-, dia-, and ferrimagnetic fractions account for w, the ferrimagnetic susceptibility, wf, was calculated by subtracting the paramagnetic contribution estimated from the high field slope in the hysteresis loops from the bulk susceptibility. 4. Results 4.1. Barranca Tlalpan sequence Magnetic susceptibility (w) versus temperature curves from BCt horizons presents two clear decays, at around 160–300 and 580 1C, and a persistence of susceptibility above 580 1C (Fig. 3a). In samples of Btg and C (not shown) and most Bt horizons, w versus temperature curves show a larger drop near 580 1C, and a weak decay above this temperature (Fig. 3b). All of these curves are almost reversible. The mineralogical phases suggested by these

Susceptibility 10-5 [S.I.]

216

Bt1 horizon

50

0 0

100

(b)

Temperature [°C]

200

300

400

500

600

700

Temperature [°C]

1

0.5 EBtg horizon

Normalized Intensity

Normalized Intensity

100

700

1

0

Ah horizon

0.5

0 0

(c)

150

100

200

300

400

500

Temperature [°C]

600

700

0

(d)

100

200

300

400

500

600

700

Temperature [°C]

Fig. 3. Curie temperature measurements for selected samples from Red Unit, Barranca Tlalpan paleosol: (a) w versus tempertarure BCt horizon Curie temperatures suggest Ti-magnetite/Ti-maghemite, almost pure magnetite plus hematite as the main magnetic mineral components. Bt4 horizon, not shown, displays the same behavior. (b) Curie temperature of 580 1C in Bt1 horizon sample shows pure magnetite, and the weak decay in w between 580 and 670 1C suggest the presence of hematite. (c) and (d) Thermal demagnetization of SIRM curves for EBtg and Ah horizons, respectively. A weak decay in remanence is observed at 120 1C in both samples, followed by a large decay around 300 1C, and a final loss of remanence at 575 1C. This behavior suggests the possible presence of antiferrimagnetic goethite, additionally to Ti-magnetite (Ti-maghemite), and pure magnetite.

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

Normalized IRM

1

Horizons Ah EBtg Btg BCt

0.5

0 0

200

400

600

800

1000

Field (mT )

Fig. 4. IRM acquisition curves of Red Unit paleosol samples. Btg and BCt horizons show a quick acquisition of remanence, and reach saturation in fields close to 300 mT, which suggest dominance of ferrimagnetic grains. Ah and EBtg samples curves, in contrast, have a more gradual increment in remanence and are not fully saturated at 500 mT fields, which suggest the significant presence of antiferrimagnetic phases, such as hematite or goethite.

(a)

(b)

(c)

(d)

217

curves correspond to Ti-magnetite/Ti-maghemite with variable Ti content, along with pure magnetite and hematite. Due to the low w in Ah and EBtg horizons, samples were prepared for thermal demagnetization of SIRM. Curves display a smooth decay in remanence. In the EBtg curve, subtle drops in remanence are found around 80–100 1C, a second inflection at 300 1C, and a final decay close to 580 1C (Fig. 3c). The sample from Ah horizon shows decays at 200, 300, and 580 1C (Fig. 3d). In addition to Timagnetite/Ti-maghemite and pure magnetite phases inferred for these horizons, goethite may also be present in the EBtg horizon. Acquisition curves of IRM show two different paths (Fig. 4). The Btg and Bt horizons have similar behavior, so we only present Btg and BCt horizons, which show steeper curves in low fields, and saturation is reached in fields close to 300 mT, this suggests dominance of relatively soft ferrimagnetic grains. Curves of Ah and EBtg samples are

(e)

(f)

(g)

(h)

Fig. 5. Down profile magnetic properties for Red Unit paleosol, Barranca Tlalpan. (a) Bulk magnetic susceptibility (w). (b) Percentages of ferrimagnetic (wf) and paramagnetic (wp) contribution to total bulk susceptibility. (c) Saturation isothermal remanent magnetization (SIRM). w and SIRM are proxies of abundance of ferrimagnetic minerals. (d) Frequency-dependent susceptibility, wfd%, is a proxy of the abundance of ultrafine (o 0.05 mm) (SP) grains. (e) ARM/SIRM ratio is an indicator for the abundance of fine (0.01–0.05 mm) SD grains. (f) HIRM300 reflects the concentration of hard magnetic minerals, such as very-fine-grained (SD) magnetite, or antiferrimagnetic minerals (hematite or goethite). (g) S-ratios estimate the variations in magnetic mineralogy or grain sizes: S100 (diamonds) is sensitive to both variables, whereas S300 (triangles) is sensitive to hard, high-coercivity antiferrimagnetic minerals such as hematite or goethite. (h) Coercive force Hc is sensitive to grain sizes (MD); coercivity of remanence Hcr is a useful guide to magnetic mineralogy, as it is higher for antiferrimagnetic minerals.

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223 1 Normalized Intensity

harder to magnetize, the increment in remanence is more gentle, and the curves are not fully saturated at 500 mT fields, all suggesting a relative increase of antiferrimagnetic phases such as hematite or goethite. Magnetic ‘‘concentration’’ parameters w and SIRM indicate higher concentrations of magnetic grains in the lower horizons and a general up-profile decay in concentration. The BCt has w46 mm3/kg, while Bt horizons present w values between 4 and 6 mm3/kg (Fig. 5a and c). The Btg, EBtg, and Ah horizons are the lowest with wo1 mm3/kg. Ferrimagnetic contribution to susceptibility (wf) in BCt and Bt horizons is 470%, while in the upper horizons (Btg, EBtg and Ah), wf is lower than the paramagnetic contribution (wp) (Fig. 5b). wfd% is fairly constant at around 5% in most of the profile, suggesting moderate abundance of SP grains, except in EBtg, where wfd% is close to 1% suggesting that SP grains are absent. In contrast, the Ah horizon is 13%, which indicates that SP particles are abundant (Fig. 5d). The abundance of fine grains (SD), according to ARM/ SIRM, is higher in the upper samples of Bt4 and Bt2 horizons, and in Bt1 and Btg horizons (Fig. 5e). HIRM300 measurements (Fig. 5f) suggest that the higher abundance of high-coercivity minerals is in the Ah, EBtg, and C horizons. S300 ratios and Hcr have constant values, except in EBtg, where S300 is close to 80% and H cr 450 mT. These values suggest that antiferrimagnetic minerals such as goethite or hematite are present in this horizon (Fig. 5g and h). This interpretation agrees with the magnetically harder phases detected in these horizons by IRM acquisition.

A

0.5

horizon

0 0

100

200

(a)

300

400

500

600

700

Temperature [°C] 1

Normalized Intensity

218

Cg

0.5

horizon

0 0

100

200

(b)

300

400

500

600

700

Temperature [°C]

Fig. 6. Curie temperature curves for Alexandrovsky quarry paleosol section for A (a) and Cg (b) horizons. All curves are similar, as they drop near 300 1C, which indicates that Ti-magnetite/Ti maghemite are present. Weaker drops at 120 and near 580 1C suggest that goethite and pure magnetite may also be present. The remanence between 580 and 670 1C suggests the content of hematite. 1

Low initial magnetic susceptibility in most of these samples (o 0.20 mm3/kg) prevented our obtaining reliable warming curves for Curie temperature determinations, except for the A horizon. Curie temperatures were thus estimated from thermal demagnetization in cemented samples. Most curves from all horizons showed inflections at 300 and 580 1C, and a weak remanence until 675 1C. A Curie temperature close to 120 1C is suggested in samples from the A and Cg horizons (Fig. 6a and b). Acquisition curves of IRM point to magnetically hard components, as remanence is not fully saturated at 500 mT fields in all samples (Fig. 7). Only the A horizon sample presents a steeper acquisition curve, which suggests relatively softer magnetic grains. Paramagnetic contribution (wp) to w is apparently higher than the ferrimagnetic contribution (wf) to w in most of the samples with wo0.2 mm3/kg. This might be misleading, as w and wp are obtained from two different instruments, and differences in sensitivity may be significant in these low initial w samples. A more reliable proxy for ferrimagnetic concentration may be the SIRM. Higher concentration of ferrimagnetic minerals are found in the BEt, E, and A horizons (Fig. 8c). wfd% was measured repeatedly, and gave

Normalized IRM

4.2. Alexandrovsky quarry sequence Horizons A 0.5

E BEt BCt

0 0

200

400

600

800

1000

Field (mT)

Fig. 7. IRM acquisition curves of Alexandrovsky quarry paleosol samples. The increase in remanence in fields above 300 mT in all curves indicates that magnetically hard minerals are important fraction of magnetic mineralogy. The two extreme paths in these curves are represented by the A and E horizons, in which the E horizon apparently has the higher content of antiferrimagnetic phases, such as hematite or goethite.

consistent values. This parameter does not give a clear pattern of high concentration of SP grains in most horizons (wfd%47%). However, one E horizon sample resulted in comparable values than those to Bt (Fig. 8d). ARM/SIRM indicates that the fine SD grains are more abundant in the A and E horizons, and less in the Cg and

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

(a)

(b)

(c)

(d)

(e)

(f)

219

(g)

(h)

Fig. 8. Down profile magnetic properties for Alexandrovsky quarry paleosol. Captions as in Fig. 5.

BCt horizons (Fig. 8e). HIRM300 is higher in the Bt1, BEt, and E Alexandrovsky profile horizons (Fig. 8f). S-ratios do not indicate significant differences in magnetic hardness, except in one sample of E horizon (Fig. 8g). In contrast, the highest coercivity Hcr is found in Cg and lower BCt Alexandrovsky samples (Fig. 8h). Two samples of the Middle Pleistocene paleosol are plotted in order to compare the magnetic behavior of the upper AB horizon and minimally weathered BCt horizon, to the respective horizons of the AQ paleosol (Fig. 8). Both paleosols have in common higher w and SIRM in the AB/A upper horizons than in the minimally affected horizons, a relationship that is much higher in the AQ Mikulino paleosol (Fig. 8a and b). 5. Discussion The Red Unit of the Barranca Tlalpan and the Alexandrovsky quarry paleosols have different magnetic mineral characteristics. The Curie points indicate that the main ferrimagnetic carriers are Ti-magnetite or Ti-maghemite with variable Ti content, for most of the RU profile. The dominance of relatively soft ferromagnetic grains is also suggested in IRM acquisition curves of samples from Btg, BCt, and C horizons. For these samples, the presence

of hematite is also suggested. In addition to Ti-magnetite/ Ti-maghemite, in the EBtg horizon, the Curie point around 100 1C suggests the presence of antiferrimagnetic goethite, which is supported by the IRM acquisition curve. In the URU and RU sections, parent material and soil horizons weakly affected by pedogenesis are characterized by a higher concentration of magnetic minerals, dominated by MD Ti-magnetites or Ti-maghemites. Pedogenic processes seem to be responsible for the formation of fine, SD magnetite, as observed in Bt4 (Fig. 5). However, a progressive upward destruction of magnetic minerals results in the decrease in magnetic concentration and coarsening of the magnetic fraction. Although the concentration of ferrimagnetic grains is very low in the EBtg and Ah horizons, antiferrimagnetic goethite accounts significantly to the magnetic signal. Ultrafine SP minerals, magnetite or goethite, are present in the Ah horizon. If the SP fraction in the upper horizons corresponds to goethite or hematite, this could be due to either the neoformation by pedogenesis, or the oxidation of ferrimagnetic minerals, such as Ti-magnetite or Ti-maghemite. In previous work in the Red Unit, direct observation of magnetic minerals in the P6 and P7 pedocomplex found Timagnetites and ilmenite grains weakly affected by corrosion and with several degrees of hematization in Bt and C

ARTICLE IN PRESS 220

J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

horizons (Ortega-Guerrero et al., 2004). Ti-magnetites, hematized Ti-magnetites, pure magnetites, and scarce hematite were identified by rock magnetism methods. In the AQ sequence, the magnetic mineralogy inferred by Curie temperatures consist of ferrimagnetic minerals such as Ti-magnetites and pure magnetites, and antiferrimagnetic phases such as hematite and goethite are suggested. However, the IRM acquisition curves, the S300 ratios, and HIRM300 strongly suggest that the antiferrimagnetic phases are the main constituents of the AQ sequence. The Mikulino paleosol profile at the Alexandrovsky site portrays only minor differences in magnetic concentration between the different pedogenic horizons along the profile. Here, only the upper sample of E horizon has lower magnetic concentration. The major difference in magnetic properties is in the distribution of small, SD grains, which are more abundant in the horizons where pedogenesis has been more intense (horizons Bt3 up to A). One of the remarkable characteristics of this sequence is the magnetic hardness of its components. Although there is low concentration of magnetic minerals, the magnetic parameters indicate that antiferrimagnetic phases, both goethite and hematite, are present throughout the profile. An apparent magnetic enhancement in susceptibility in the A horizon is not mirrored in the proxies of ferrimagnetic concentration as SIRM, or in the ultrafine fine fraction wfd%. This enhancement is only coincident with the increase in the abundance of small SD grains. A possible explanation of these characteristics is that paramagnetic ironbearing minerals are being formed in the upper horizon, and only a small fraction of them are ferrimagnetic minerals, such as SD magnetite. As shown above, within both Albic Luvisol profiles, there are layers having contrasting differences of their magnetic properties, which coincide with soil genetic horizons. This provides evidence that these properties are linked to the specific set of soil-forming processes operating in each horizon. In the Kursk profile of the buried Albic Luvisol developed on loess, the well-known phenomenon of magnetic susceptibility enhancement in the paleosol compared to parent material is expressed. The high values of magnetic susceptibility in this profile are probably mostly due to accumulation of fine-grained magnetic components, which we assume to be neoformed in the soil environment. However, this phenomenon is observed only in some parts of the paleosol profile. It is strongest in the topsoil, in the Ah of the AQ paleosol, and then to a lesser extent in the Middle Pleistocene AB horizon at the base of the studied sequence (Fig. 8). This corresponds to the zone of highest past biological activity and accumulation of organic material, a zone that is rather thin in Albic Luvisols. Less-pronounced w enhancement was detected in the uppermost horizons of the illuvial part of the profile, namely the EBt and partly the Bt1. The values decrease downwards in the Bt2, Bt3, and BCt horizons, although they are still somewhat higher than in the Cg horizon. In

general, this zone is characterized by carbonate leaching, weathering, and clay illuviation (associated with precipitation of iron oxides). However, we refrain from directly linking of the second susceptibility maximum just with these processes. It should be taken into account that this maximum is located not in the middle part of illuvial zone (Bt2 and Bt3) where clay accumulation due to illuviation is greatest, but in the upper horizon, where illuviation is still moderate and eluvial pedofeatures (concentrations of bleached skeletal material) are frequent. The distribution of different kinds of clay illuvial pedofeatures could be helpful to explain the susceptibility curve. Targulian and Bronnikova (2002) report in their recent study of an Albic Luvisol cutan complex that EBt and Bt1 horizons are characterized by a high concentration of specific iron–clay and iron–manganese cutans, which quickly diminish downwards. These pedofeatures indicate intensive precipitation of iron oxides together with manganese and/or clay (which still persists in this horizon in minor quantities, despite eluviation). Finally, somewhat higher susceptibility values in the Bt2, Bt3, and BCt horizons compared to the Cg horizon should be partly controlled by carbonate leaching. The parent material of Mikulino paleosol at AQ, the Dnepr loess, is known to contain 20% and more of calcium carbonate, which ‘‘dilutes’’ magnetic minerals present in this sediment. The decarbonatization (leaching) front is located at the upper Cg horizon boundary, where an increase of mineral magnetic components is apparent. Two major maxima of the susceptibility curve for the Ah and EBt horizons are separated by a strong minimum in the eluvial E horizon. We believe that this minimum is linked to a specific pedogenic environment of an E horizon, in which simultaneously operate strong acid weathering, leaching, and surface redoximorphic (stagnic) processes that are likely aggressive towards magnetic minerals, both inherited (lithogenic) and neoformed. The specific role of stagnic conditions, caused by periodic water saturation above low permeable clay-illuvial horizons, consists of mobilization of iron in the form Fe2+. This iron, mobilized in eluvial horizons and capable of migration with soil solutions, is believed to be the source for iron oxides that precipitate in the specific illuvial pedofeatures of the subadjacent EBt and Bt1 horizons (considered above as a possible reason for susceptibility enhancement). 6. Conclusions In general, the samples of a Luvisol profile formed in ‘‘tepetates’’ from Barranca Tlalpan show much higher values of magnetic susceptibity than the AQ Albic Luvisol formed in loess. A conspicuous feature of the BT profile is the complete absence of enhancement in the soil profile. The values in the parent material are higher than in any soil horizon. At the same time, we observe considerable increase of fine-grained magnetic components in the soil, compared to the C horizon where these components are

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

lower. We conclude that the high susceptibility values in the parent material are controlled by abundant coarse volcanogenic magnetic minerals (mainly titanomagnetite) and the susceptibility decrease in the soil results from the partial loss of these minerals due to intra-soil weathering. At the same time, the neoformation of magnetic components in soil takes place, which accounts for accumulation of fine-grained magnetic minerals. However, this process cannot compensate for the loss of lithogenic magnetic minerals in any genetic horizon, and the resulting trend is magnetic susceptibility depletion throughout the soil profile. This model was already proposed earlier for volcanic paleosols of Mexico (Ortega-Guerrero et al., 2004) and supported by microscopic observations of weathering features. A significant point of this finding is that it is critical for researchers to fully characterize a paleosol profile including a detailed morphological description and also analysis of both genetic soil horizons and the parent material (C horizon). This should also include quantitatively establishing that the described C horizon is the parent material for the genetic soil horizons and is not a different lithologic unit. Within this hypothesis, the divergence between Albic Luvisols on tephra (depletion case) and loess (enhancement case) can be explained by the differences in original quantity of the lithogenic coarse magnetic components. Since coarse lithogenic magnetic components are few in loess, their loss due to weathering is of minor importance compared to neoformation in A, BEt, and Bt1 horizons, which results in susceptibility enhancement, as was shown with the help of CBD test by TenPas et al. (1999). Behavior of magnetic properties within the soil profile shows their remarkable correlation with the type of genetic horizon. Susceptibility is higher in Bt horizons, decreases in the BEt, reaches a minimum in the E, and increases a little bit in the A horizon. It is interesting also to trace the differences in fine-grained magnetic components accumulation, which have two maxima—a major one in the Bt horizons (especially in the lowest Bt3 horizon) and minor peak in the topsoil A horizon, with a sharp decrease in the E horizon. We speculate that the neoformation of finegrained magnetic minerals is associated with two pedogenetic environments: (1) topsoil area of humus accumulation and highest past biological activity, and (2) area of clay illuviation and weathering in the middle part of the profile. A question arises concerning the accumulation of fine magnetic components in the Bt horizons: Since the fine components are well expressed in the paleosol on tephra, why are they weak or absent in the loess-derived paleosol (except the uppermost illuvial horizons)? We attribute this divergence to the differences of parent material composition. Pyroclastic sediments provide abundant (up to 25% of coarse fractions) primary minerals containing iron (crystalline Fe–Ti oxides, Fe–Mg silicates) which generate finegrained iron oxides in the course of weathering. On the contrary, Dnepr loess has rather ‘‘poor’’ mineralogical

221

composition of primary minerals, dominated by quartz and K–Na feldspars, and low concentrations of iron-bearing minerals. Magnetic susceptibility curves of both studied profiles coincide in demonstrating strong minima in eluvial horizons, which shows also low concentrations of finegrained components. We conclude that the pedogenic environment of eluvial horizons, which experience strong weathering, leaching, surface water redoximorphic processes, and low biological activity, is destructive for all magnetic components (both primary and secondary) and does not support neoformation. Our data and interpretation contradict to some extent the results of Singer and Fine (1989), who found greater susceptibility in the eluvial part of the profile than in the illuvial part or subsoil in a number of soils of California. Probably the detected phenomenon of E-horizon minimum is specific for Albic Luvisols and is related to the combination of eluvial processes and surface gleying (stagnic conditions) responsible for bleaching in this layer. Terhorst et al. (2001), who also observed this phenomenon in the E horizon of a buried Eemian soil in southern Germany, report minima of magnetic susceptibility in the bleached zones of the Bt horizon of the same soil, also affected by eluviation and gleying. Finally, the higher concentrations of antiferromagnetic components (hematite and goethite), which were found in E horizons of both studied soils, also seem to be related to redoximorphic processes. Such processes involve both dissolution of iron oxides and also their local fast precipitation as ferruginous mottles and nodules that are observable in thin sections.

Acknowledgments This work was funded by ICSU (International Council of Science) project ‘‘Polygenetic models for Pleistocene paleosols’’. Partial support also was given by UNAM DGAPA IN107902, IX102104, and CONACyT G-28528T. D. Hernandez collaborated in technical support. The authors thank Dr. D.A. Grimley and an anonymous reviewer for their helpful comments on the paper.

References Banerjee, S.K., Hunt, C.P., Liu, X.M., 1993. Separation of local signal from the regional paleomonsoon record of the Chinese loess plateau: a rock-magnetic approach. Geophysical Research Letters 20, 843–846. Bege´t, J.E., Stone, D.B., Hawkins, D.B., 1990. Paleoclimatic forcing of magnetic susceptibility variations in Alaskan loess during the late quaternary. Geology 18, 40–43. Bronger, A., Winter, R., Heinkele, T., 1998. Pleistocene climatic history of east and central Asia based on paleopedological indicators in loess–paleosol sequences. Catena 34, 1–17. 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, 128–132.

ARTICLE IN PRESS 222

J. Rivas et al. / Quaternary International 156–157 (2006) 212–223

Dearing, J.A., Livingstone, I., Zhou, L.P., 1996. A late quaternary magnetic record of Tunisian loess and its palaeoclimatic significance. Geophysical Research Letters 23, 189–192. de Jong, E., Pennock, D.J., Nestor, P.A., 2000. Magnetic susceptibility of soils in different slope positions in Saskatchewan Canada. Catena 40, 201–305. Dodonov, A.E., Shackleton, N.J., Zhou, L.P., Lomov, S.P., Finaev, A.F., 1999. Quaternary loess–paleosol stratigraphy of Central Asia: geochronology, correlation, and evolution of paleoenvironments. Stratigraphy and Geological Correlation 7 (6), 581–593. Dunlop, D.J., O¨zdemir, O¨., 1997. Rock Magnetism, Fundamentals and Frontiers. Cambridge University Press, New York, 573pp. Evans, M.E., Rutter, N.W., Catto, N., Chlachula, J., Nyvlt, D., 2003. Magnetoclimatology: teleconnection between the Siberian loess record and North Atlantic Heinrich events. Geology 31, 537–540. 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, 97–107. Forster, T., Heller, F., 1997. Magnetic enhancement paths in loess sediments from Tajikistan, China and Hungary. Geophysical Research Letters 24, 17–20. Geiss, C.E., 1999. The development of rock magnetic proxies for paleoclimate reconstruction. Ph.D. Thesis, University of Minnesota, 274pp. Geiss, C.E., Zanner, C.W., Banerjee, S.K., Minott, J., 2004. Signature of magnetic enhancement in a loessic soil in Nebraska, United States of America. Earth and Planetary Science Letters 228 (3–4), 355–367. Grimley, D.A., Vepraskas, M.J., 2000. Magnetic susceptibility for use in delineating hydric soils. Soil Science Society of America 64 (6), 217–235. Grimley, D.A., Follmer, L.R., Mc Kay, E.D., 1998. Magnetic susceptibility and mineral zonations controlled by provenance in loess of unglaciated southwestern Illinois and central Mississippi Valleys. Quaternary Research 49 (1), 24–36. Grimley, D.A., Arruda, N.K., Bramstedt, M.W., 2004. Using magnetic susceptibility to facilite more rapid, reproducible and precise delineation of hydric soils in the middle western USA. Catena 58, 183–213. Heller, F., Evans, M.E., 1995. Loess magnetism. Reviews of Geophysics 33, 211–240. Heller, F., Liu, T.S., 1982. Magnetostratigraphic dating of loess deposits in China. Nature 300, 431–433. Heller, F., Liu, T.S., 1986. Palaeoclimate and sedimentary history from magnetic susceptibility of loess in China. Geophysical Research Letters 13, 1169–1172. Jordanova, D., Petrovsky, E., Jordanova, N., Evlogiev, J., Butchvarova, V., 1997. Rock magnetic properties of recent soils from northeastern Bulgaria. Geophysical Journal International 128, 474–488. Kukla, G., An, Z.S., 1989. Loess stratigraphy in central China. Palaeogeography, Palaeoclimatology, and Palaeoecology 72, 203–225. Le Borgne, E., 1955. Susceptibilite magnetique anomale du sol superficial. Annales de Geophysique 11, 399–419. Le Borgne, E., 1960. Etude experimentale du trainage magnetiques dans le Cas d’un ensemble de grains magnetiques TRES anormale du sol superficiel. Annales de Geophysique 16, 445–494. Liu, X.M., Rolph, T., Bloemendal, J., Shaw, J., Liu, T.S., 1995. Quantitative estimates of palaeoprecipitation at Xifeng, in the Loess Plateau of China. Palaeography, Palaeoclimatology, Palaeoecology 113, 243–248. Liu, X.M., Hesse, P., Rolph, T., Bege´t, J.E., 1999. Properties of magnetic mineralogy of Alaskan loess: evidence for pedogenesis. Quaternary International 62, 93–102. Liu, X.M., Hesse, P., Bege´t, J.E., Rolph, T., 2001. Pedogenic destruction of ferrimagnetics in Alaskan loess deposits. Australian Journal of Soil Research 39, 99–115. Maher, B.A., 1998. Magnetic properties of modern soils and loessic paleosols: implications for paleoclimate. Palaeogeography, Palaeoclimatology, Palaeoecology 137, 25–54.

Maher, B.A., Thompson, R., 1992. Paleoclimatic significance of the mineral magnetic record of the Chinese loess and paleosols. Quaternary Research 37, 155–170. Maher, B.A., Thompson, R., 1995. Paleorainfall reconstructions from pedogenic magnetic susceptibility variations in the Chinese loess and paleosols. Quaternary Research 44, 383–391. Maher, B.A., Thompson, R., Zhou, L.P., 1994. Spatial and temporal reconstructions of changes in the Asian palaeomonsoon: a new mineral magnetic approach. Earth and Planetary Science Letters 125, 461–471. Maher, B.A., Alekseev, A., Alekseeva, T., 2003. Magnetic mineralogy of soils across the Russian Steppe: climatic dependence of pedogenic magnetite formation. Paleogeography, Paleoclimatology, Paleoecology 201, 321–341. McElhinny, M.W., McFadden, P.L., 2000. Paleomagnetism of Continents and Ocean. Academic Press, San Diego, CA, 386pp. Nabel, P.E., Morra´s, H.J.M., Petersen, N., Zech, W., 1999. Correlation of magnetic and lithologic features of soils and Quaternary sediments from the Undulating Pampa, Argentina. Journal of South American Earth Sciences 12, 311–323. Opdyke, N.D., Channell, J.E.T., 1999. Magnetic Stratigraphy. Academic Press, San Diego, CA, 346pp. Orgeira, M.J., Walther, A.M., Va´squez, C.A., Di Tommaso, I., Alonso, S., Sherwood, G., Yuguan, H., Vilas, J.F.A., 1998. Mineral magnetic record of paleoclimate variation in loess and paleosol from the Buenos Aires formation (Buenos Aires, Argentina). Journal of South American Earth Sciences 11, 561–570. Ortega-Guerrero, B., Sedov, S., Solleiro-Rebolledo, E., Soler, A., 2004. Magnetic mineralogy in Barranca Tlalpan exposure paleosols, Tlaxcala, Mexico. Revista Mexicana de Ciencias Geolo´gicas 21 (1), 120–132. Robinson, S.G., 1986. The late Pleistocene paleoclimatic record of North Atlantic deep-sea sediments revealed by mineral-magnetic measures. Physics of the Earth and Planetary Interiors 42, 22–47. Rousseau, D.D., Puisse´gur, A., 1998. A 350,000-year climatic record from the loess sequence of Achenheim Alsace, France. Boreas 19, 203–216. Schellenberger, A., Heller, F., Veit, H., 2003. Magnetostratigraphy and magnetic susceptibility of the Las Carreras loess–paleosol sequence in Valle de Tafı´ , Tucuma´n, NW-Argentina. Quaternary International 106–107, 159–167. Singer, M.J., Fine, P., 1989. Pedogenic factors affecting magnetic susceptibility of northern California soils. Soil Science Society of America Journal 53, 1119–1127. Singer, M.J., Fine, P., Verosub, K.L., Chadwick, O.A., 1992. Time dependence of magnetic susceptibility of soil chronosequences on the California coast. Quaternary Research 37, 323–332. Singer, M.J., Verosub, K.L., Fine, P.T., TenPas, J., 1996. A conceptual model for the enhancement of magnetic susceptibility in soils. Quaternary International 34–36, 243–248. Sycheva, S.A., 1998. New data on composition and evolution of the mezin loess–paleosol complex in the Russian plain. Eurasian Soil Science 31 (10), 1062–1074. Sycheva, S., 2004. Aleksandrov quarry late pleistocene-holocene. Cover complex: soils, loesses, buried balka. In: Proceedings of Conference: geological heritage concept, conservation and protection policy in Central Europe. Polish Geological Institute Special Papers, 13. Warszawa, pp. 175–181. Tang, Y., Jiab, J., Xie Xiande, 2003. Records of magnetic properties in quaternary loess and its paleoclimatic significance: a brief review. Quaternary International 108, 33–50. Targulian, V.O., Bronnikova, M., 2002. Sequence of illuviation cutans formation in Albeluvisols. In: International Union of Soil Science (Eds.), Paper published at the 17th World Congress Soil Science, Thailand Symposium 36 (942), pp. 942-1–942-13. TenPas, J.D., Vidic, N.J., Singer, M.J., Verosub, K.L., 1999. Mineral magnetic and pedogenic studies of the paleoclimatic record of the upper part of the loess/paleosol sequence at Jiaodao. Chinese Science Bulletin 44 (1), 259–263.

ARTICLE IN PRESS J. Rivas et al. / Quaternary International 156–157 (2006) 212–223 Terhorst, B., Appel, E., Werner, A., 2001. Palaeopedology and magnetic susceptibility of a loess–palaeosol sequence in southwest Germany. Quaternary International 76–77, 231–240. Tite, M.S., Linington, R.E., 1975. Effect of climate on the magnetic susceptibility of soils. Nature 256, 565–566. Vadyunina, A.F., Babanin, V.F., 1972. Magnetic susceptibility of some soils in the U.S.S.R. Soviet Soil Science 4, 588–599. Velichko, A.A., 1990. Loess–paleosol formation on the Russian Plain. Quaternary International (7–8), 103–114. Vlag, P.A., Solheid, P.A., Oches, E.A., Banerjee, S.K., 1999. The paleoenvironmental-magnetic record of the gold hill steps loess section

223

in central Alaska. Physics and Chemistry of the Earth Part A 24, 779–783. Wall, H., Worm, H., 2000. A cautionary note on interpreting frequencydependence of susceptibility solely in terms of superparamagnetism. Institute for Rock Magnetism, Quarterly 10(4), 1–6. Zheng, H., Oldfield, F., Yu, L., Shaw, J., An, Z., 1991. The magnetic properties of particle-sized samples from the Luo Chuan loess section: evidence for pedogenesis. Physics of the Earth and Planetary Interior 68, 250–258. 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, 737–739.