Tectonic uplift-influenced monsoonal changes promoted hominin occupation of the Luonan Basin: Insights from a loess-paleosol sequence, eastern Qinling Mountains, central China

Quaternary Science Reviews 169 (2017) 312e329

Contents lists available at ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Tectonic uplift-influenced monsoonal changes promoted hominin occupation of the Luonan Basin: Insights from a loess-paleosol sequence, eastern Qinling Mountains, central China Qian Fang a, Hanlie Hong a, *, Lulu Zhao a, Harald Furnes b, Huayu Lu c, Wen Han d, Yao Liu a, Zhuoyue Jia a, Chaowen Wang a, e, Ke Yin a, Thomas J. Algeo a, f, g, * a

State Key Laboratory of Biogeology and Environmental Geology & School of Earth Sciences, China University of Geosciences, Wuhan, 430074, China Department of Earth Science & Centre for Geobiology, University of Bergen, Allegt. 41, 5007 Bergen, Norway School of Geographic and Oceanographic Sciences, Nanjing University, Jiangsu Collaborative Innovation Center for Climate Change, Nanjing, 210023, China d National Gems & Jewelry Technology Administrative Center, Beijing, 100013, China e Gemological Institute, China University of Geosciences, Wuhan, 430074, China f State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Wuhan, 430074, China g Department of Geology, University of Cincinnati, Cincinnati, OH 45221-0013, USA b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 January 2017 Received in revised form 19 May 2017 Accepted 30 May 2017

Quaternary soil deposits from northern and southern China are distinctly different, reflecting variability of the East Asian monsoon north and south of the Qinling Mountains. Coeval sediments from the transitional climatic zone of central China, which are little studied to date, have the potential to improve our understanding of Quaternary monsoon changes and associated influences on hominin occupation of this region. Here, we investigate in detail a well-preserved and continuous Quaternary loess-paleosol sequence (Shangbaichuan) from the Luonan Basin, using a variety of weathering indices including major and trace element ratios, clay mineralogy, and Fe-oxide mineralogy. The whole-rock samples display similar rare earth element patterns characterized by upper continental crustal ratios: (La/Yb)N z 9.5 and Eu/Eu* z 0.65. Elemental data such as (La/Yb)N, La/Th and Eu/Eu* ratios show a high degree of homogeneity, suggesting that dust in the source region may have been thoroughly mixed and recycled, resulting in all samples having a uniform initial composition. Indices for pedogenic weathering such as Na/K, Ba/Sr, Rb/Sr, CIA, CIW, CPA, PIA, kaolinite/illite, (kaolinite þ smectite)/illite, and hematite/ (hematite þ goethite) exhibit similar secular trends and reveal a four-stage accumulation history. The indices also indicate that the climate was warmer and wetter during the most recent interglacial stage, compared with coeval environments of the Chinese Loess Plateau. Secular changes in weathering intensity can be related to stepwise uplift of the Qinling Mountains and variation in East Asian monsoon intensity, both of which played significant roles in controlling climate evolution in the Luonan Basin. Furthermore, intensified aridity and winter monsoon strength in dust source areas, as evidenced by mineralogic and geochemical changes, may have been due to the mid-Pleistocene climate transition. Based on temporal correlation of warmer and wetter climatic conditions with more frequent hominin occupation, we infer that the paleoclimate in the eastern Qinling Mountains remained mild and favorable during glacial stages of the Late Quaternary, thus promoting early human settlement. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Pedogenic weathering Pleistocene East Asian monsoon Paleoclimatology Sediment mineralogy Rare earth elements

1. Introduction

* Corresponding authors. State Key Laboratory of Biogeology and Environmental Geology & School of Earth Sciences, China University of Geosciences, Wuhan, 430074, China. E-mail addresses: [email protected] (H. Hong), [email protected] (T.J. Algeo). http://dx.doi.org/10.1016/j.quascirev.2017.05.025 0277-3791/© 2017 Elsevier Ltd. All rights reserved.

The East Asian monsoon (EAM), which is an important component of the global climate system, is a major control on hydrological and environmental changes in East Asia (An, 2000; An et al., 2015). The EAM is characterized by long-term (~104e105 yr) alternations in dominance between the warm/humid summer

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

monsoon and the cold/dry winter monsoon, which are coupled with variations in solar radiation and continental ice volume (An et al., 2001, 2015; Wang et al., 2005). The EAM circulation influences water supply and agricultural food production in East Asia, and can also create natural hazards such as flooding and drought, affecting 30% of the world's population (Sun et al., 2015). Because knowledge of monsoon variability during the geological past is essential to a better understanding of the present climate system and prediction of future climatic changes, interest in the history of the EAM has been increasing (e.g., Chen et al., 2006; Clift et al., 2014; An et al., 2015; Lu, 2015; Sun et al., 2011, 2015). The climate evolution of northern and central China during the Quaternary has been largely controlled by variations in intensity of the East Asian summer monsoon (EASM) and East Asian winter monsoon (EAWM). In northern China, the Chinese Loess Plateau (CLP) contains abundant loess-paleosol sequences that represent valuable records of multi-cycle changes of continental climatic conditions (Kukla and An, 1989; Liu and Ding, 1998; Balsam et al., 2004; Torrent et al., 2007). In southern China, Quaternary red earth sediments are accretionary in nature, recording intense syndepositional weathering in a subtropical to tropical monsoonal climate (Qiao et al., 2011). The relatively sharp transition between Quaternary sediment types in northern and southern China is due to the presence of the Qinling Mountain Belt (QMB) in central China (Fig. 1; Cai et al., 2010; Zhang et al., 2012). The QMB partially blocks the circulation of monsoon air and marks a significant climatic boundary: the northern and southern sides of the QMB are characterized by distinct climates and vegetation cover (Lu et al., 2007, 2011). The Luonan Basin to the east of the QMB, which contains many surface and in situ lithic artifacts, is viewed as a vital region for understanding the history of hominin settlement in East Asia (Lu et al., 2007, 2011). This basin is also recognized by archaeologists as the transitional region between the northern China flake-core tool industries and the southern China pebble-core tool industries (Wang, 2005; Sun et al., 2013, 2014). Studies of sediment magnetic stratigraphy and optically stimulated luminescence (OSL) have shown that early humans occupied the Luonan Basin intermittently during the early to middle Pleistocene (Lu et al., 2007, 2011). Existing research has proposed that the Luonan Basin was a suitable place for hominin settlement based on investigations of environmental magnetism (Wang, 2005; Wang-X et al., 2016). Nevertheless, the stimulus for multiple stages of hominin occupation of the Luonan Basin during the Late Quaternary remains unclear. Also unresolved is the driving force of monsoonal changes in the Luonan Basin, which is important for an improved understanding of the EAM generally. Several well-preserved and continuous loess-paleosol sequences have been found recently in the Luonan Basin, offering an excellent opportunity to better understand monsoonal changes in the transitional climate zone of central China, as well as their specific influences on hominin occupation. The well-preserved Shangbaichuan (SBC) loess-paleosol sequence is representative of loess deposits in the Luonan Basin and, thus, an ideal archive for paleoclimate reconstructions (Lu et al., 2007, 2012; Sun et al., 2014). Mineralogic and geochemical analyses of loess deposits enable correlations with likely source areas, elucidate secular variation within the loess sequence, and shed light on pedogenic weathering processes and paleoclimate changes (Yang et al., 2007a; Muhs et al., 2008; Jiang et al., 2016). In this study, we examined variations in clay and Fe-oxide mineralogy, elemental ratios, and a variety of element-based indices for pedogenic weathering in the SBC loesspaleosol sequence using X-ray diffraction (XRD), diffuse reflectance spectrophotometry (DRS), X-ray fluorescence (XRF), inductively coupled plasma-mass spectrometer (ICP-MS), and scanning

313

electron microscopy (SEM). The objectives of this study were: 1) to determine which mineralogic and geochemical indices most effectively record variations in intensity of the EASM and EAWM; 2) to reconstruct variations in the EASM and EAWM since 870 ka within the transitional climate zone of central China; 3) to compare characteristics of the SBC sequence with coeval loess deposits on the CLP and in southern China in order to illuminate the climatic role of the Qinling Mountains; and 4) to elucidate possible relationships in timing and causality between tectonic uplift, changes in the EAM, and early human occupation of the Luonan Basin. 2. Geographical and geochronological background The SBC loess-paleosol sequence (34 040 0300 N, 110 030 0600 E) is situated on the second terrace of the Shimenhe River, which is a tributary to the South Luo River, ~7 km west of Luonan County, Shangluo City, Shaanxi Province, central China (Fig. 1; Lu et al., 2007). The South Luo River flows eastward through the Luonan basin, before finally joining the Yellow River. The Luonan Basin is located in a region with a generally warm and semi-humid climate. This basin is an intermontane depression with a surface elevation of ~990 m, a mean annual temperature (MAT) of ~11  C, a mean annual precipitation (MAP) of ~705 mm (Lu et al., 2007). Loess-paleosol deposits in the Luonan Basin are distributed mainly on tablelands and river terraces, with total thicknesses ranging from 1 to 30 m, composed of several distinct loess-paleosol alternations (Lu et al., 2007, 2012; Zhang et al., 2012). The Luonan loess is harder (i.e., more firmly indurated) than that of the CLP, containing numerous black manganese oxides nodules and brown ferric oxide horizons (Zhang et al., 2012). Many important and famous early to middle Pleistocene archeological sites (e.g., Dali and Lantian) are present in the vicinity of the Luonan Basin (Xiao et al., 2002; Bae, 2010; Yin et al., 2011). More than 90,000 stone artifacts, as well as one hominin tooth thought to belong to Homo erectus, have been reported from these sites (Xue, 1987; Wang et al., 2008; Bae, 2010). The QMB, which is located dominantly in Shaanxi Province, central China, is situated on a tectonically active belt located between the Yangtze Craton and North China Craton. The QMB was formed by the collision of these continental plates during the Mesozoic (Meng and Zhang, 2000). The belt stretches for ~2200 km in a roughly east-west direction at an average elevation of ~2000 m. Field observations indicate that the SBC loess-paleosol sequence in the Luonan Basin is more intensively weathered than that of the CLP. The paleosols have a dark reddish-brown (5YR 3/6) to light reddish-brown hue (5YR 4/6), and the loess units show a light reddish-brown (7.5YR 5/8) to brown hue (7.5YR 5/6) (Table 1). The loess sequence consists mainly of wind-blown silt, with abundant pedogenic Fe and Mn concretions in the middle and lower parts (10.6e19.7 m) of the SBC study section (Fig. 2; Lu et al., 2007). The age framework of the SBC sequence was established in earlier studies through magnetostratigraphic analysis, regional correlation of loess-paleosol layers, and OSL dating (Fig. 2; Lu et al., 2007, 2011). The Brunhes/Matuyama (B/M) boundary, which is dated to ~780 ka (Cande and Kent, 1995), has been identified at ~17.2 m below the top of the SBC profile (Fig. 2; Lu et al., 2007, 2011). Assuming a constant sedimentation rate above and below the B/M boundary, an age of ~870 ka for the base of the section was obtained by linear extrapolation (Lu et al., 2011). The SBC sequence has been approximately correlated with the Luochuan loesspaleosol sequence: (1) the uppermost four paleosols at SBC correspond to paleosols S1-S4 at Luochuan, and (2) the paleosol unit just below the B/M boundary at 17.2 m is equivalent to paleosol S8 at Luochuan (Lu et al., 2011). However, the SBC loess and paleosol units do not exactly match those of the Luochuan sequence on a

314

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Fig. 1. (A) Monsoonal circulation of East Asia (modified from Ge et al., 2013). The study area in central China is influenced by the East Asian summer monsoon (EASM), the East Asian winter monsoon (EAWM), and the Indian summer monsoon (ISM). (B) The study section at Shangbaichuan (SBC) is located in the Luonan Basin, south of the Chinese Loess Plateau and east of the Qinling Mountains. It is located on a terrace of the South Luo River, which flows eastward into the Yellow River. Other sections discussed in the text (e.g., Luochuan and Lingtai) are also shown. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table 1 Loess stratigraphy and sample descriptions of the SBC sequence. Sample number

Stratigraphy

Depth (m)

Munsell color

SBC001-019 SBC020-026 SBC027-036 SBC037-046 SBC047-064 SBC065-076 SBC077-082 SBC083-105 SBC106-115 SBC116-120 SBC121-128 SBC129-150 SBC151-157 SBC158-170 SBC171-178 SBC179-185 SBC186-197

Paleosol Loess Paleosol Loess Paleosol Loess Paleosol Loess Paleosol Loess Paleosol Loess Paleosol Loess Paleosol Loess Paleosol

0.1e1.9 1.9e2.6 2.6e3.6 3.6e4.6 4.6e6.4 6.4e7.6 7.6e8.2 8.2e10.5 10.5e11.5 11.5e12.0 12.0e12.8 12.8e15.0 15.0e15.7 15.7e17.0 17.0e17.8 17.8e18.5 18.5e19.7

2.5YR 4/6 7.5YR 5/8 5YR 3/6 10YR 6/8 7.5YR 5/6 10YR 6/8 7.5YR 5/8 7.5YR 5/6 2.5YR 4/4 7.5YR 5/6 5YR 4/6 7.5YR 5/6 7.5YR 5/8 10YR 6/6 2.5YR 4/6 7.5YR 5/8 5YR 4/6

one-to-one basis; for example, the prominent unit S5 of the Luochuan sequence corresponds to at least three units of the SBC (Fig. 2). We have labeled the loess and paleosol layers in the SBC sequence using “L*” and “S*”, in order to distinguish them from the standard numbering system for loess (“L”) and paleosol (“S”) layers, as at Luochuan (Fig. 2). 3. Sampling and analytical methods 3.1. Sampling Among dozens of loess deposits in the Luonan Basin, the SBC loess-paleosol sequence was chosen for the present study, because it is well-preserved without any significant depositional gaps (Lu et al., 2007, 2012; Zhang et al., 2012). The SBC sequence has a total thickness of ~20 m and contains 17 loess-paleosol alternations (Fig. 2). A total of 197 bulk-rock samples, weighing ~1 kg each, were

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

315

Fig. 2. Field photos of SBC section: (A) upper half (0e10.6 m), and (B) lower half (10.6e19.7 m). The section consists mainly of wind-blown silt, with abundant pedogenic Fe and Mn concretions in the lower half. (C) Age framework of the SBC section, as established by magnetostratigraphy, OSL dating, and regional correlation of loess-paleosol layers (after Lu et al., 2004, 2007, 2011). The Brunhes/Matuyama (B/M) boundary (0.78 Ma) is located at ~17.2 m; the red stars represent horizons containing hominin artifacts (after Lu et al., 2011). Loess and paleosol units at SBC are marked with asterisks (*) to indicate a lack of exact correspondence to similarly numbered units at Luochuan. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

collected at 10-cm intervals in a steep-sided gully from the present land surface to the bottom of the exposure. Each sample was airdried and then ground to <200 mesh using a mortal and pestle, in preparation for bulk-rock mineralogic, clay mineralogic, Fe-oxide mineralogic, elemental geochemistry analyses. Loess stratigraphy and sample descriptions are given in Table 1. For elemental analyses, we collected 18 representative samples from the 17 alternating loess and paleosol layers, with one sample taken from each layer except for S*1 (the youngest paleosol), from which two samples (i.e., SBC002 and SBC016) were collected because this layer exhibits a light-colored upper part and a darkcolored middle-lower part (Table 1). The geochemical data yielded by these samples provide a low-resolution record of average compositional variation between loess and paleosol layers, serving as a useful complement to the higher-resolution mineralogic and geochemical indices generated for the larger (n ¼ 197) sample set.

and Reynolds, 1989; Hong et al., 2015, 2016). The ADM was prepared by pipetting clay paste onto a glass slide. The EGM was obtained by saturating the ADM with ethylene-glycol vapor at 70  C for 8 h, for determination of the expandable clays. The 15.6-, 14.2-, and 10.0 Å reflections were used to identify smectite, chlorite, and illite, respectively; 3.57 and 7.15 Å represented the basal (002) and (001) reflections of kaolinite. The ADM was heated to 600  C for 2 h to obtain the HM, which distinguished chlorite from kaolinite; the two minerals can also be distinguished from the 3.52 Å peak of chlorite versus the 3.57 Å peak of kaolinite. The XRD operating conditions for clay-mineral analysis were identical to those for bulk-rock analyses. The clay mineral abundances were semiquantitatively determined following Kahle et al. (2002), which used weighting factors to convert the peak area of a given clay mineral into its relative abundance. The illite crystallinity (IC) was determined by the full width at half maximum of the illite 10 Å reflection.

3.2. X-ray diffraction For bulk-rock mineralogy, organic matter in the powder was removed with H2O2 prior to analysis. The powder was side-loaded into a sample holder to ensure random orientation (Eberl and Smith, 2009; Fang et al., 2017b). The X-ray diffraction (XRD) pattern was recorded using a Panalytical diffractometer with CuKa radiation (40 kV, 40 mA). The clay fraction (<2 mm) was isolated from suspension by centrifugation after dispersion. Clay-mineral species were determined on the basis of air-dried oriented mounts (ADM), ethyleneglycol saturated mounts (EGM), and heated mounts (HM) (Moore

3.3. Elemental geochemistry analyses and related weathering proxies Major-element concentrations were determined by X-ray fluorescence (XRF) spectrometry. The loss-on-ignition (LOI) value was calculated from the difference in mass after the sample was heated to 105  C and 950  C. The analytical precision for major-element concentrations was <1% and the detection limit was 0.01 wt%. Elemental concentrations are given in the form of weight percentages of oxide in the Results section and in Table 3. These values were converted to molar proportions in making calculations of

316

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Table 2 Geochemical proxies for paleoclimate variations used in this study. Weathering indices

Formula

Reference

Chemical index of alteration (CIA) Chemical index of weathering (CIW, CIA - K) Chemical proxy of alteration (CPA) Weathering index of Parker (WIP) Plagioclase index of alteration (PIA)

Al2O3/(CaO* þ Na2O þ Al2O3 þ K2O)  100 Al2O3/(CaO* þ Na2O þ Al2O3)  100

Nesbitt and Young (1982) Harnois (1988)

Al2O3/(Na2O þ Al2O3)  100 [2(CaO*/0.70) þ 2(Na2O/0.35) þ 2(K2O/0.25) þ (MgO/0.90)]  100 [(Al2O3 e K2O)/(CaO* þ Na2O þ Al2O3 - K2O)]  100

Buggle et al. (2011) Parker (1970) Fedo et al. (1995)

Table 3 Major element data for the 18 SBC loess-paleosol samples. Sample name

Depth(m)

SiO2 (%)

TiO2 (%)

Al2O3 (%)

Fe2O3 (%)

MnO (%)

MgO (%)

CaO (%)

Na2O (%)

K2O (%)

P2O5 (%)

LOI (%)

SBC002 SBC016 SBC023 SBC031 SBC041 SBC055 SBC070 SBC079 SBC094 SBC110 SBC118 SBC124 SBC139 SBC154 SBC165 SBC175 SBC182 SBC196 UCCa

0.2 1.6 2.3 3.1 4.1 5.5 7.0 7.9 9.4 11.0 11.8 12.4 13.9 15.4 16.5 17.5 18.2 19.6

62.52 63.53 65.17 63.27 62.03 65.68 64.35 63.23 64.43 64.48 67.10 63.84 64.52 65.01 66.30 60.74 66.19 64.45 66.62

0.78 0.71 0.84 0.79 0.83 0.84 0.78 0.76 0.77 0.79 0.89 0.82 0.76 0.82 0.86 0.79 0.85 0.85 0.64

15.82 15.29 14.96 16.07 17.01 15.01 15.54 15.34 15.15 14.94 13.99 15.55 15.28 15.00 14.48 17.07 14.62 15.03 15.40

6.30 6.26 5.93 6.35 6.51 5.81 6.08 6.31 5.94 5.87 5.42 6.10 5.94 5.82 5.69 6.71 5.66 6.24 5.04

0.11 0.09 0.14 0.09 0.09 0.14 0.09 0.09 0.10 0.13 0.28 0.11 0.11 0.11 0.14 0.11 0.13 0.20 0.10

1.62 1.67 1.54 1.61 1.39 1.49 1.66 1.72 1.72 1.69 1.32 1.65 1.75 1.57 1.35 1.64 1.35 1.37 2.80

0.98 0.99 0.95 0.91 0.77 0.76 0.91 1.04 1.09 1.05 0.81 0.79 1.08 0.89 0.82 0.96 0.80 0.79 3.59

1.09 1.23 1.27 1.03 0.64 1.24 1.18 1.28 1.37 1.43 1.16 1.13 1.35 1.27 0.99 0.71 1.04 0.91 3.27

2.39 2.52 2.44 2.48 2.20 2.50 2.56 2.55 2.63 2.62 2.25 2.48 2.71 2.45 2.28 2.36 2.28 2.37 2.80

0.07 0.15 0.09 0.08 0.09 0.08 0.09 0.12 0.14 0.13 0.10 0.07 0.15 0.09 0.09 0.09 0.09 0.11 0.15

7.91 6.63 6.58 7.54 8.24 6.36 6.86 6.75 6.41 6.24 5.93 6.89 6.42 6.38 6.47 8.68 6.77 7.52

a

UCC data are from Taylor and McLennan (1985).

paleoweathering proxies. Simple elemental ratios in bulk-rock samples have been widely employed as paleo-weathering proxies (Chen et al., 2006; Hosek et al., 2015; Lü et al., 2016). The principle underlying their use is that selective removal of more mobile elements leads to a comparative enrichment of more immobile elements (Price and Velbel, 2003; Buggle et al., 2011). For example, elemental ratios such as K/Al, Na/Al, and Mg/Al are controlled mainly by depletion of K, Na, and Mg and passive enrichment of Al. Moreover, Ca (which is present in modest amounts in silicate minerals and in larger quantities in carbonates) and Na are more mobile in the weathering environment than K (both Na and K reside mainly in silicate minerals such as feldspar and illite). Thus, a high K/Ca ratio represents loss of carbonate and silicate Ca through weathering (Varga et al., 2011, and references therein). Multi-element weathering proxies are calculated using a combination of immobile and mobile elements, providing quantitative/ semiquantitative measures of mineral alteration and weathering (Hosek et al., 2015; Schatz et al., 2015). A variety of chemical weathering proxies have been developed (Table 2). The most commonly used proxy is the chemical index of alteration (CIA; Nesbitt and Young, 1982), which is based on molar oxide ratios: CIA ¼ Al2O3/(CaO* þ Na2O þ Al2O3 þ K2O)  100). CaO* stands for the CaO content of silicate minerals, so total CaO must be corrected for carbonate and phosphate CaO contents (we used the procedure of McLennan, 1993). Higher CIA values indicate more intense chemical weathering. Other weathering proxies include: (1) the chemical index of weathering (CIW, also known as CIAeK), an index without potassium (Harnois, 1988); (2) the chemical proxy of alteration (CPA), an index based on Na and Al (Buggle et al., 2011); (3) the weathering index of Parker (WIP, Parker, 1970); and (4) the

plagioclase index of alteration (PIA), a modification of CIW and CIA designed to emphasize plagioclase weathering (Fedo et al., 1995). Trace-element and rare earth element (REE) analyses were measured by inductively coupled plasma-mass spectrometry (ICPMS) performed on an Agilent ICP-MS, with an analytical precision of <4% for Y and REE, and 4e10% for other trace elements. Some of these elements have also been widely used in chemical weathering studies. For example, Sr is usually depleted during intense pedogenesis, whereas Ba shows more stable values in paleosol layers compared with loess layers (Varga et al., 2011; Hosek et al., 2015). Thus, the Ba/Sr ratio can serve as an indicator for pedogenic weathering and leaching (Price and Velbel, 2003; Fitzsimmon et al., 2012). Similarly, the Rb/Sr, Sr/Ca, Mg/Sr and Zr/Rb ratios are also used as indicators of paleoenvironmental conditions, although they may be influenced by other factors (Chen et al., 2006; Buggle et al., 2011; Qiao et al., 2011). 3.4. Diffuse reflectance spectrophotometry We used diffuse reflectance spectrophotometry (DRS) as a sensitive and accurate approach to determining Fe-oxide mineralogy. In order to quantitatively determine goethite and hematite content, the sample was prepared and analyzed for DRS following the protocol of Ji et al. (2001) and Balsam et al. (2004). Sample powder was pressed into a plastic holder at a pressure of ~200 kPa to reduce the influence of grain-size variation, and then was analyzed using a Persee TU-1901 spectrophotometer equipped with a diffuse reflectance instrument, between 380 and 750 nm (visible to infrared spectrum). This is the most sensitive EM region for iron oxides responsible for color variations in the loess-paleosol sequences (Balsam et al., 2004; Torrent et al., 2006; Buggle et al.,

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

317

2014). Reflectance intensity data were measured with respect to a white BaSO4 standard, and were recorded at 2-cm intervals. Significant correlations exist between the second derivative of the reflectance spectrum and the abundances of goethite and hematite, which can be expressed by the following equations (Scheinost et al., 1998): Goethite (g$kg1) ¼ 0.06 þ 268X1 (R2 ¼ 0.86, p(a) < 0.001)

(1)

Hematite (g$kg1) ¼ 0.09 þ 402X2 (R2 ¼ 0.85, p(a) < 0.001) (2) where X1 represents the second-derivative amplitude between the 415 and 445-nm bands of goethite, while X2 represents the secondderivative amplitude between the 535-nm and 580-nm reflections of hematite. 3.5. Scanning electron microscopy Based on the XRD results, four representative samples were selected for scanning electron microscopic (SEM) observations. A fresh surface on each slabbed sample was Pt-coated and observed with an EDS-equipped HITACHI FEG-SEM in secondary-electron mode. The microscope was operated at a voltage of 10e15 kV and a current of 3e5 nA. 4. Results 4.1. Major element geochemistry Major-element data for 18 SBC loess-paleosol samples are summarized in Table 3. The SBC loess is characterized by high concentrations of SiO2 (60.7e67.1%, mean 64.3%), Fe2O3 (5.42e6.71%, mean 6.04%) and Al2O3 (14.0e17.1%, mean 15.34%), and low contents of TiO2 (0.76e0.89%, mean 0.81%), MnO (0.09e0.28%, mean 0.13%), MgO (1.32e1.75%, mean 1.56%), CaO (0.76e1.09%, mean 0.91%), Na2O (0.64e1.43%, mean 1.12%), K2O (2.20e2.71%, mean 2.44%), and P2O5 (0.07e0.15%, mean 0.10%). Major-element contents for these samples vary within narrow ranges (Table 3). The loss on ignition (LOI) also varies in a narrow range, from 5.93 to 8.68%, and shows a positive correlation with Al2O3 content (R2 ¼ 0.77) but no clear correlation with CaO content (R2 ¼ 0.01). These relationships indicate that samples components lost during LOI were closely associated with clay minerals rather carbonate minerals. Most SBC samples contain little carbonate, as shown by wt%(CaO) < wt%(Na2O) and by lack of calcite peaks in bulk-rock XRD patterns (see Section 4.3). The major-element abundances of the whole-rock samples exhibit similar upper continental crust (UCC)-normalized patterns (Fig. 3). Si, Al and K contents are similar to those of UCC; Ti, Fe and Mn are enriched and Mg, Ca, Na and P depleted relative to UCC (Fig. 3). The loss of Ca and Mg was probably due to dissolution of carbonates, and the loss of Na and K due to chemical weathering of feldspars and subsequent formation of clays (Gallet et al., 1996). The partial dissolution of organic matter and phosphate nodules may be responsible for the mobility of P. 4.2. Trace element and rare earth element geochemistry Similar to the major element data, trace element concentrations also vary within relatively narrow ranges (Table S1). The distribution patterns of all samples are similar: the Co, Cr, Ni, and Cu contents are close to UCC; Li, Zn, Ga, Sn, La and Pb are enriched whereas Sr and Ba are depleted relative to UCC (Fig. 3B). The Sr and Ba depletions may be related to high mobility of large-ionlithophile elements during pedogenic weathering (Jahn et al.,

Fig. 3. UCC-normalized distributions of (A) major elements and (B) trace elements, and (C) chondrite-normalized REEs in the SBC section. Note that blue lines represent loess samples and orange lines represent paleosol samples. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2001; Hong et al., 2016). The chondrite-normalized REE patterns (designated by a subscripted “N”) of the SBC samples are presented in Fig. 3C. REE distributions show highly similar features, including distinct negative

318

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Eu anomalies, enriched light-REEs (LREEs), and relatively flat heavy-REE (HREE) distributions (Fig. 3C). Similar REE distributions have been reported from the CLP. The (La/Yb)N ratio, which is a measure of light-to-heavy REE fractionation, exhibits narrow variation (9.2e10.1, mean 9.5), revealing no evident distinction between loess and paleosol units throughout the sequence. Eu anomalies Eu/Eu* are stable (0.62e0.67, mean 0.65) and similar to that of the UCC. Unlike samples from the Luochuan sequence (Gallet et al., 1996), the SBC samples show no positive Ce anomalies (Fig. 3C). Relatively narrow ranges are shown by total REEs (SREE) (183e227 ppm, mean 202 ppm), SLREE (168e205 ppm, mean 180 ppm), and SHREE (19.4e24.3 ppm, mean 22.0 ppm). The SLREE/SHREE ratio ranges from 7.3 to 9.4, with a mean of 8.2. This may be attributed to the characteristics of the parent material and the fact that LREEs are preferentially enriched in fine-grained sediments (Taylor and McLennan, 1985; Yang et al., 2007a). 4.3. Mineralogic composition Representative XRD spectra of bulk samples, ADMs and EGMs are shown in Fig. 4. The mineral components of the SBC loesspaleosols are quartz, clay minerals and minor feldspars. Clay species in the SBC sequence are primarily smectite, kaolinite and illite (Fig. 4). Trace amounts of fine-grained quartz are also commonly present in the clay fractions, whereas there is no obvious evidence for the presence of clay-sized feldspars. No chlorite is observed in the SBC samples. After ethylene-glycol saturation, the ~14.5 Å reflection in the ADM disappears and a ~16.8 Å peak appears in the EGM (Fig. 4). In the 400  C-heated mount, the ~14.5 Å reflection completely collapses with an increase in the intensity of the ~10 Å reflection (Fig. 4). These changes indicate similar clay components including smectite and kaolinite. The ~10, ~5, and ~3.3 Å reflections occur in the whole-rock samples, implying that illite is ubiquitous throughout the SBC sequence. The semi-quantitative analyses of the clay fraction suggest that clay-mineral assemblages consist predominantly of illite (39e77%, mean 59%) and smectite (18e59%, mean 34%). Kaolinite is present only in trace amounts, ranging from 2 to 15%, with a mean value of 7%. Illite crystallinity (IC) varies between 0.25 and 0.50 D2q, with a mean value of 0.37 D2q, implying relatively good illite crystallinity (Wang and Yang, 2013). Variation trends in clay mineral abundances reveal four distinct mineralogical intervals within the sequence (Units 1 to 4, from top to bottom), with transitions at 10.6 m (~480 ka), 4.8 m (~330 ka), and 2.6 m (~180 ka) (Fig. 5). In order to better reflect climatic evolution, we describe the results in chronological order. Unit 4 (19.7e10.6 m, S*9 to S*5, ~870e480 ka), the lower part of the sequence, is characterized by distinct compositional cycles linked to loess-paleosol alternations. The paleosol units contain more smectite and kaolinite and less illite than the loess units (Fig. 5). The Kao/I and (Kao þ Sm)/I ratios covary positively with kaolinite and smectite concentrations, and negatively with illite (Fig. 5). Variation ranges for clay mineral contents in this unit are relatively wide, with illite content ranging from 34 to 73%, smectite from 21 to 59%, and kaolinite from 2 to 11%. IC values fluctuate but show a variation pattern similar to that of illite (Fig. 5). Unit 3 (10.6e4.8 m, L*5 to S*3, ~480e330 ka), the lower middle part of the sequence, has a relatively uniform clay mineralogy (Fig. 5). Illite abundance remains high and stable in both loess and paleosol horizons, whereas smectite and kaolinite abundances remain relatively low (Fig. 5). Despite generally stable clay mineralogy, kaolinite content and Kao/I ratio in this unit tend to increase upsection. Unit 2 (4.8e2.6 m, L*3 to S*2, ~330e180 ka), the upper middle part of the sequence, exhibits higher illite content in paleosol than

in loess horizons, whereas kaolinite and smectite abundances, as well as Kao/I and (Kao þ Sm)/I ratios, exhibit the opposite pattern of variation (Fig. 5). Unit 1 (2.6e0 m, L*2 to S*1, ~180e10 ka), the uppermost part of the sequence, is also marked by relatively stable mineralogy. Its illite content is slightly lower than that of Unit 2, whereas smectite and kaolinite contents are generally higher (Fig. 5). 4.4. Fe-oxide mineralogy The DRS patterns were used to identify and quantify goethite and hematite at the SBC sequence. For each sample, we calculated goethite and hematite concentrations using Equations (1) and (2) (see Section 3.4) and then calculated the hematite/ (hematite þ goethite) ratio, i.e., Hm/(Hm þ Gt). Representative first-derivative DRS patterns for loess and paleosol samples from the SBC sequence show goethite peaks at ~430 and ~500 nm and hematite peaks at 550e580 nm (Fig. 6). Because the goethite peak at ~500 nm is usually masked due to slight overlap with hematite peaks, we regard the peak at ~430 nm as a reliable indicator of goethite. Both goethite and hematite concentrations are generally low throughout the sequence; goethite concentration ranges from 0.33 to 4.04 g kg1, with a mean of 1.38 g kg1, whereas hematite concentration has limited variation, ranging from 0.05 to 2.07 g kg1, with a mean of 0.94 g kg1. Variations of hematite content and Hm/(Hm þ Gt) display somewhat similar trends with those of some clay indices, e.g., kaolinite and smectite contents and Kao/I and (Kao þ Sm)/I ratios, but are more fluctuating (Fig. 5). In order to obtain smoother curves, we replotted the data using five-point moving averages (Fig. 5). In Unit 4, hematite content and Hm/(Hm þ Gt) ratio show higher values in paleosol than in loess horizons (Fig. 5). Unlike clay minerals, variations of hematite-related indices are irregular in Unit 3 (Fig. 5). In Unit 2, hematite-related indices covary positively with illite and IC, opposite to other clay indices (Fig. 5). Hematite content and Hm/(Hm þ Gt) ratio show positive covariation with, but are more variable than, clay mineralogy in Unit 1 (Fig. 5). 5. Discussion 5.1. Provenance of clay minerals Clay mineral assemblages in soils are dependent not only on weathering and transport processes but also on the provenance of the sediments (Chamley, 1989; Sheldon and Tabor, 2009). Thus, interpretations of the significance of mineralogical and geochemical proxies for chemical weathering and monsoonal intensity are more robust if the sediment source can be constrained (e.g., WangC et al., 2016). With regard to the origin of the SBC loess, Zhang et al. (2012) demonstrated that the deposits are eolian, with mixed dust sources derived from (i) the Gobi Desert and drylands in northern/ northwestern China, and (ii) weathering of local sedimentary rocks derived from bedrock of the proximal orogen belt and adjacent alluvial-fluvial deposits. Here, we briefly summarize several lines of evidence from their work and our study: (1) Modern synoptic data from China Meteorological Administration clearly show the common direction and process of the dust storm, confirm the eolian origin of the SBC loess, and strengthen the argument for a dust source from northern/northwestern China; (2) both major and trace-element (including REE) data show similar distribution patterns and conform well to those of typical loess-paleosol samples, suggest that the SBC loess is well-mixed; (3) Sr-Nd isotopic compositions indicate that the SBC loess has mixed sources, i.e., in addition to the sources of materials to the CLP loess, material has

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

319

Fig. 4. XRD patterns for whole-rock and clay-fraction mounts of a representative sample (SBC-41) using various preparation techniques (e.g., ethylene-glycol saturation, ion saturation, and heat treatment). Mineralogic legend for XRD peaks shown in upper right.

also been derived from nearby bedrock of the Qinling Mountains, with source initially derived from the old crust. SEM analysis allows us to observe textures and morphologies of clay minerals in the loess sediments that can provide insights into their origins (Hong et al., 2007; Fang et al., 2017a, b). Clays occur as aggregates filling the space between detrital minerals with no preferred orientation, showing various morphologies (Fig. 7). Generally, clay particles are present as flakes with irregular/round and undulating outlines and discrete boundaries with other grains, having formed as replacements of detrital minerals such as quartz and feldspars. The clay particles vary in size from 0.1 to 5.0 mm and in thickness from 0.1 to 0.4 mm, with much of this variation due to breakage and mechanical erosion during transport (Hong et al., 2007; Fig. 7). Irregular/embayed edges and varying thicknesses of clay minerals are consistent with intense dissolution during leaching and weathering processes (Fig. 7). A clay-mineral ternary diagram can also be employed to better illuminate the provenance of SBC samples by comparison with CLP loess deposits (Fig. 8A). Clay mineral assemblages of the SBC differ from those of the CLP (e.g., the Luochuan and Lingtai loess

deposits). Specifically, the SBC samples show greater variation in smectite, lesser variation in kaolinite, and a similar dominance of illite relative to CLP loess (Fig. 8A; Gylesjo and Arnold, 2006; Won et al., 2017). These data reinforce the mixed-source hypothesis for SBC loess (Zhang et al., 2012) and suggest that the proximal weathered bedrocks of the QMB may have served as parent materials for a non-trivial fraction of smectite clays in the SBC samples. The clay-mineral assemblage of eolian deposits derived from bedrock is typically a mixture of chlorite, illite, kaolinite, smectite, and mixed-layer illite-smectite in varying proportions (Jeong et al., 2008, 2011; Wang and Yang, 2013). Any changes in clay-mineral assemblages from those inherited from the parent materials provide information about the types of pedogenic processes and climatic conditions that operated on a sediment (Chadwick and Chorover, 2001; Liu et al., 2005; Jeong et al., 2008). Almost all of the SBC samples are devoid of chlorite and mixedlayer illite-smectite clays (Fig. 8A). Chlorite is commonly a physical erosion product of feldspars, micas and/or greenschist-grade metamorphic rocks exposed to weak chemical weathering (Chamley, 1989). Thus, the absence of chlorite indicates intense

320

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Fig. 5. Profiles of clay-mineral concentrations and ratios, illite crystallinity, and Fe-oxide phases in the SBC section. Dotted lines show the horizons sampled for geochemical analysis. For the Fe proxies, five-point moving averages were calculated. Variation trends in clay mineral assemblages reveal four distinct intervals within the section (Units 1 to 4, from top to bottom).

The absence of mixed-layer illite-smectite in the SBC sequence indicates a lack of transformation of smectite to illite or illite to smectite in the soil environment (Varga et al., 2011). The full SBC sequence exhibits a pronounced inverse relationship between illite and smectite (Fig. 5), which is sometimes an indication of illite-tosmectite weathering conversions within a soil profile (Varga et al., 2011; Hong et al., 2016). However, illite is also a primary mineral derived from rock erosion under cool and/or arid conditions (Chamley, 1989), and the inverse trends in illite and smectite in the SBC sequence may therefore reflect variations in source-rock composition (Miao et al., 2016). Such variations commonly reflect changes in transport vectors, which may be linked to monsoonal changes (Ji et al., 2001; Gylesjo and Arnold, 2006). Based on comparison of SEM observations with the smectite concentration profile (Figs. 5 and 7), we infer that smectite in the SBC sequence had a predominantly detrital origin, and that the contribution of in-situ pedogenic production was minor. Secular trends in smectite concentration are similar to those for kaolinite, Kao/I and (Kao þ Sm)/I (Fig. 5), which suggests that smectite may be an appropriate proxy for paleoclimatic reconstruction. In this regard, it is important to emphasize that both smectite and kaolinite form under warmer, wetter conditions favoring more intense chemical weathering, and that illite forms mainly under cooler, drier conditions favoring physical weathering (Chamley, 1989). Thus, the opposite trends between smectite and kaolinite on the one hand and illite on the other reflect broad variation in weathering styles at the SBC site during the late Quaternary. Fig. 6. First-derivative (FD) curves of representative samples of (A) loess and (B) paleosol from the SBC section. Gray bars denote peak positions of the FD values at 420e440 and 500e520 nm, corresponding to goethite, and at 550e580 nm, corresponding to hematite.

chemical weathering of the source area under relatively warm and/ or wet conditions. Mixed-layer illite-smectite clays may be a product of diagenetic illitization of smectite, or the end-product of rock degradation through moderate chemical weathering in poorly drained areas (Chamley, 1989; Hong et al., 2010b; Fang et al., 2017a).

5.2. Quasi-uniform primary composition of SBC sequence The main factors controlling chemical weathering intensity include sediment lithology, climate, and geomorphology (Chadwick and Chorover, 2001; Sheldon and Tabor, 2009; Wang-C et al., 2016). Eolian materials potentially undergo weathering and chemical differentiation in their source regions, and such weathering ideally should be distinguished from weathering that takes place following eolian transport and loess deposition. Thus, understanding variations in the primary composition of eolian deposits is important as a prelude to analysis of mineralogical and

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

321

Fig. 7. Morphological variation of clay minerals in SEM images. (AeB) Detrital smectite shows curved and/or undulating outlines (sample SBC091). (C) Tabular detrital kaolinite shows nearly hexagon outlines and ragged edges (sample SBC136). (D) Detrital illite shows poorly developed plates and embayed margins (sample SBC073).

geochemical changes due to in-situ weathering (Xiong et al., 2010). Paleoclimatic reconstructions based on loess deposits are most robust when the entire loess sequence has a quasi-uniform primary composition (Schatz et al., 2015). This crucial condition must be verified by evidence based on immobile elements that do not change much due to in-situ weathering (Jahn et al., 2001; Lan et al., 2016). Major- and trace-element concentrations are relatively homogenous for both loess and paleosol samples (Fig. 3), indicating that SBC loess deposits may have been well mixed prior to eolian transport and deposition. Also, the similarity of trace-element distributions for the SBC loess-paleosol units and UCC further validates our inference that eolian transport and deposition have yielded a representative sample of crustal material at the SBC site (Taylor and McLennan, 1985; Jahn et al., 2001). Moreover, as REEs are especially resistant to remobilization during weathering, REE characteristics such as Eu/Eu* and (La/Yb)N should be inherited from the source. The striking similarity of REE distributions between analyzed samples at SBC indicates little to no change in eolian source composition throughout the sequence (e.g., Ding et al., 2001; Yang et al., 2007a, b). Comparing the La/Th ratios of the SBC samples (2.6e2.9) with the referenced value of 2.8 ± 0.2 for UCC (Taylor and McLennan, 1985; Fig. 9), we conclude that insoluble Th has also been little influenced by transport and deposition, and that materials constituting the loess deposits were well mixed in the source regions. Paleosols of the Luochuan sequence exhibit a distinct negative Ce anomaly, which was interpreted to have been due to intense

chemical weathering during pedogenesis (Gallet et al., 1996). However, none of the SBC samples displays this Ce anomaly (Fig. 3C), which is consistent with the results from the Xining, Jixian and Xifeng loess deposits of the CLP (Jahn et al., 2001). Rather, the general homogeneity of REE distributions, including Eu and Ce anomalies and (La/Yb)N and La/Th ratios, all support our inference that the SBC loess-paleosol sequence had a relatively uniform primary composition. 5.3. Chemical weathering and climate evolution since 870 ka Pedogenic weathering occurs due to the thermodynamic instability of primary materials exposed to meteoric conditions (Chadwick and Chorover, 2001), and it can greatly influence the mineralogy and geochemistry of loess deposits (Lu et al., 2012; Schatz et al., 2015). Various quantitative/semi-quantitative geochemical and mineralogical indices have been developed to reflect chemical weathering intensity and to reconstruct paleoenvironments at the time of soil formation (Ahmad and Chandra, 2013; Schatz et al., 2015; Li et al., 2016; Sun et al., 2016). Clay mineralogy has been widely used as a proxy for paleoclimate and pedogenesis as it is strongly influenced by temperature and precipitation, along with source material composition, duration of weathering, water chemistry, and water-rock ratio (Chamley, 1989; Weaver, 1989). The main clay-mineral components found in soils include kaolinite, smectite, illite, chlorite, and mixedlayer clays (Chamley, 1989; Varga et al., 2011; Churchman and Lowe, 2012; Hong et al., 2007, 2015). Kaolinite and smectite form

322

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Fig. 9. La-Th covariation for SBC section. UCC (red square; Taylor and McLennan, 1985) is plotted for comparison. The nearly constant La-Th concentrations indicate that these fine-grained sedimentary rocks are derived from a nearly homogenous source. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 8. (A) Clay mineralogy and (B) major cation chemistry of the SBC, Luochuan (LC) and Lingtai (LT) sections shown on ternary plots. Panel B is an A-CN-K-type diagram (A: Al2O3; CN: CaO þ Na2O; K: K2O) in which CIA is shown on the vertical axis. Note weathering trends subparallel to the A-CN boundary for both SBC and LC. Clay-mineral data for Lingtai and Luochuan are from Gylesjo and Arnold (2006) and Won et al. (2017), respectively, and geochemical data for Luochuan are from Gallet et al. (1996).

during advanced stages of chemical weathering, corresponding to intense weathering; kaolinite forms under hot and humid (tropical) conditions, and smectite forms in seasonally wet (temperate) climates (Chamley, 1989; Gylesjo and Arnold, 2006; Hong et al., 2010a). Generally, chlorite and illite form in early stages of weathering, dominating in immature soils that have undergone relatively weak chemical weathering; they are commonly abundant in drier and cooler pedoclimates (Hong et al., 2007). Because these two minerals are regarded as less susceptible to the effects of chemical weathering, they can be used as “internal standards”, allowing proxies such as smectite/illite, kaolinite/illite, smectite/ (illite þ chlorite), and (smectite þ kaolinite)/(illite þ chlorite) to be useful for characterization of chemical weathering intensity and climate regime (Varga et al., 2011). Given negligible variations in eolian source composition and weathering, pedogenic weathering indices such as clay-mineral assemblages, Fe-oxide phases, and geochemical ratios can be

used to reconstruct pedogenic weathering and climatic conditions in the Luonan Basin over the past ~870 kyr. CIA values for the SBC sequence range from 66 to 79, with a mean value of 71, suggesting an intermediate weathering condition (Fig. 8B), similar to that at Luochuan on the CLP (Gallet et al., 1996). The SBC samples are characterized by absence of both carbonates and a Ce anomaly (Figs. 3 and 4), in contrast to the Luochuan loess (Jahn et al., 2001). Negative Ce anomalies may be linked to formation of carbonate concretions under oxidizing soil water conditions (Udvardi et al., 2016). The preservation or dissolution of primary and secondary carbonate minerals depends strongly on soil pH, which can vary considerably at a local scale due to variations in plant cover (e.g., Jeong et al., 2011; Udvardi et al., 2016). That the SBC sequence yields a similar CIA to Luochuan but lacks any carbonate may imply more acidic porewater conditions, and lack of soil carbonate may account for the absence of a Ce anomaly (since the detrital fraction of the sediment typically shows none). Distinguishing pre-transport (source) from post-transport (insitu) weathering effects is generally difficult for loess deposits. The CIA cannot distinguish between source and in-situ weathering (Yang et al., 2006). The A-CN-K [Al2O3-(CaO* þ Na2O)-K2O] ternary diagram (Fig. 8B), which was developed to show weathering trends and the initial composition of parent materials (Nesbitt and Young, 1989; Buggle et al., 2011), shows an overall trend for the SBC samples subparallel to the A-CN boundary. This implies that Naand Ca-bearing minerals were preferentially weathered over Kbearing minerals, and that the deposits were not subjected to Kmetasomatism (Ahmad and Chandra, 2013). A single trend for weathering in this diagram is an indication that the source material had a relatively invariant primary composition (Fig. 8B). Fe-oxide phases display notable changes through the SBC sequence (Fig. 5). Iron-oxides in soils have been frequently employed as pedogenic weathering indicators as their assemblages are sensitive to variations in temperature and in the amount and seasonal distribution of precipitation (Chen et al., 2005, 2010; Hu et al., 2013; Zhao et al., 2017). Relatively high temperatures together with intermittent or seasonal rainfall favor the formation of hematite (Long et al., 2016, and references therein). In contrast, continuously wet conditions promote the formation of goethite as it precipitates directly from solution (Ji et al., 2001). Among the Feoxide indices, hematite content and Hm/(Hm þ Gt) exhibit secular

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

variation trends that are generally similar to those of clay content and geochemical weathering proxies. Hematite and goethite can have diverse origins that are hard to discriminate (Chen et al., 2010; Long et al., 2016). For example, three different sources of hematite in CLP deposits have been inferred: 1) weathering of Fe-bearing minerals (e.g., pyroxene, hornblende and chlorite), 2) magnetite oxidation in surface soils, and 3) eolian hematite (e.g., from basement rock or desert soils; Chen et al., 2010). Thus, the Fe mineralogy of the SBC sequence is influenced not just by the pedogenic weathering regime. Despite these complexities, it is useful to integrate various approaches including Fe indices to obtain a more comprehensive interpretation of paleoclimatic variation at a given locale (Buggle et al., 2014; Lyons et al., 2014). Proxies for chemical weathering reveal four distinct intervals (Units 1e4) within the SBC sequence (see Sections 4.3 and 4.4), on the basis of which we identify four stages in the climate evolution of the Luonan Basin during the middle-late Quaternary: 5.3.1. Unit 4 Unit 4 corresponds to the interval from ~870 to 480 ka (MIS 22MIS 13). It comprises four distinct cycles of mineralogic variation with a mean periodicity of ~100 kyr that closely reflects loesspaleosol alternations (Fig. 5). Illite content is higher in the loess than in the paleosol units, whereas kaolinite, smectite, Kao/I, and (Kao þ Sm)/I are higher in the paleosols. The clay-mineral indices reveal strengthened pedogenic weathering and weakened accumulation during interglacial stages versus weaker pedogenesis and higher depositional rates during glacial stages, a pattern that is typical of climatic conditions for loess-paleosol formation (Jahn et al., 2001; Ji et al., 2001; Udvardi et al., 2007). However, the pattern of variation in clay-mineral indices within loess-paleosol couplets at SBC is opposite to that documented in the Lingtai and Baoji loess sections of the CLP. In the latter sections, the loess layers contain more kaolinite and less illite than the paleosols, an unusual pattern that was attributed to source-controlled mineralogy rather than pedogenesis (Kalm et al., 1996; Gylesjo and Arnold, 2006). Most geochemical indices (e.g., CIA, CIW, CPA, PIA, Mg/Al, Ba/Sr and Rb/Sr) in the SBC section exhibit a similar symmetrical shape (Fig. 10), suggesting that they are approximately equally well-suited to assessment of variations in weathering intensity. Evidence of Kmetasomatism, which can lead to differences between K-based and K-free proxies, is not observed in the study section (Fig. 10), probably owing to the low content of K-feldspar (Fig. 8B; Schatz et al., 2015). Geochemical proxies for Unit 4 show two rather than four cycles of variation, perhaps due to the lower resolution of the geochemical profiles, but differences between loess and paleosol horizons are nonetheless manifest (Fig. 10). The end of Unit 4 deposition at ~0.6e0.5 Ma was marked by strong variation in clay mineralogical indices, reflecting intense climate fluctuations. On the whole, chemical weathering intensity weakened from ~870 to 480 ka, as reflected in both geochemical and mineralogic indices (Figs. 5, 10 and 11). Thus, a stepwise cooling and drying of the Luonan Basin during this period can be inferred. 5.3.2. Unit 3 Unit 3 corresponds to the interval from ~480 to 330 ka (MIS 12MIS 10). It is characterized by relatively constant mineralogy and geochemistry regardless of loess-paleosol alternations, along with an overall increase in chemical weathering upsection that is especially evident in geochemical indices (Figs. 5, 10 and 11). We infer a rather stable environment with a weak trend toward warmer and wetter climatic conditions during this interval. Among mineralogic indices, kaolinite and Kao/I show more distinct trends than illite and smectite, possibly because the latter minerals are dominant and small variations in their abundances are not obvious. Loess-

323

paleosol alternations reflect changing balances between dust accumulation and pedogenesis, as controlled by climatic conditions, and higher smectite and kaolinite contents are expected in the paleosols (Liu and Ding, 1998; Huang et al., 2012). In this context, the nearly identical clay-mineral composition of loess and paleosol layers in Unit 3 is unusual (cf. Zhao et al., 2017; Varga et al., 2011; Udvardi et al., 2016) and may reflect a combination of source and pedogenic influences (Gylesjo and Arnold, 2006; Huang et al., 2012), as opposed to dominantly pedogenic influences in the other units at SBC. 5.3.3. Unit 2 Unit 2 corresponds to the interval from ~330 to 180 ka (MIS 9MIS 7). This unit exhibits pronounced cyclic variation in both mineralogic and geochemical proxies (Figs. 5, 10 and 11), suggesting a climatic alternation between warm/humid and cool/dry conditions. The distribution of kaolinite and illite in loess and paleosol layers is opposite to that seen in Unit 4 but matches that in correlative loess deposits (i.e., Unit 2-equivalent) at Baoji (Kalm et al., 1996), Lingtai (Gylesjo and Arnold, 2006), and Luochuan (Won et al., 2017). Clay-mineral and geochemical weathering proxies in Unit 2 show consistent patterns (Fig. 11), indicating that the clay-mineral assemblage reflects climatic and pedogenetic controls rather than the composition of source materials (Miao et al., 2016). 5.3.4. Unit 1 Unit 1 corresponds to the youngest interval at SBC, dating to ~180 to 10 ka (MIS 6 to MIS 1). Both mineralogic and geochemical proxies suggest relatively stable climatic conditions since 180 ka (Figs. 10 and 11). Although major climate fluctuations are known throughout this interval (Oh and Shin, 2016), especially between the last glacial maximum (LGM) at ~20 ka and the beginning of the Holocene at ~10 ka, the sampling resolution of the present study is not sufficient to resolve these high-frequency climate events. 5.3.5. Climate trends since 870 ka The profiles of various weathering proxies (especially clay mineralogical indices) do not show an obvious overall climate trend but, rather, changes between highly variable conditions (Units 4 and 2) and more stable conditions (Units 3 and 1) (Fig. 11). In addition, the mean value of (Kao þ Sm)/I changes from 0.84 (Unit 4, 19.7e10.6 m, 870e480 ka) to 0.65 (Unit 1e3, 10.6e0.1 m, 480e10 ka), suggesting that the overall climate trend was toward milder with less precipitation. The trend is in good agreement with field observations of abundant pedogenic Fe and Mn concretions in the lower part (19.7e10.6 m) of the study section, which are characteristic of a soil gleying process induced by long-term water-saturated conditions related to flooding (Zhang et al., 2012; Wang-X et al., 2016). The inferred trend is generally consistent with rock magnetic data of the Liuwan section in the Luonan Basin (Wang-X et al., 2016). Different from the climatic conditions in northern China, which became cooler and drier with more severe oscillations during the late Quaternary, the long-term climate of the Luonan Basin changed overall from highly variable conditions with frequent flooding to more stable and milder environments with less influence from the EAWM due to uplift of the Qinling Mountains, providing a more suitable place for hominin occupation (Zhang et al., 2012; Wang-X et al., 2016). 5.4. Tectonic uplift of the Qinling Mountains Climate ranks among the most important factors that control weathering processes and products (Chadwick and Chorover, 2001; Torrent et al., 2007; Sheldon and Tabor, 2009). Weathering

324

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Fig. 10. Profiles of geochemical weathering proxies for the SBC section. The weathering indices CIA, CIW, CPA, PIA, Na/K, Ba/Sr and Rb/Sr covary positively among themselves and negatively with Mg/Al, Na/Al and (Ca þ Na þ Mg)/Ti ratios. The winter-monsoon index Si/Ti (Jahn et al., 2001) is also shown. Different types of weathering indices are indicated with different colors.

Fig. 11. Summary of weathering indices showing climate evolution in the Luonan Basin since 870 ka. Global benthic d18O data are from Lisiecki and Raymo (2005), artifact horizons from Lu et al. (2011), and tectonic uplift events from Sun (2005). These records demonstrate temporal relationships among tectonic uplift, climatic conditions, and hominin settlement in the Luonan Basin.

intensity increases with higher precipitation levels and mean annual temperatures (Yang et al., 2006). Beside the monsoonal climate, tectonism may also exert influences on physical erosion and chemical weathering through accelerating mechanical denudation in the source areas, and hence increasing the exposure of fresh rock surfaces, as well as shifting atmospheric circulation

patterns and climatic conditions (An, 2000; Udvardi et al., 2007). Also, differential tectonic uplift can produce steep local relief (Sun, 2005). Clay-mineral assemblages are particularly sensitive to chemical weathering linked to tectonic movements (Hong et al., 2007, 2015; Varga et al., 2011; Wang-C et al., 2016). The Qinling Mountains have been actively rising during the

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Quaternary (Wang et al., 2002; Xue et al., 2004; Sun and Xu, 2007). Wang et al. (2002) calculated a mean rate of uplift since 0.4 Ma of 0.3e0.5 mm yr1 using 230Th dating, equivalent to a total uplift of ~400 m over 1 million years. The effects of tectonic uplift are recorded in fluvial terraces of the Fen Wei Graben, adjacent to the Qinling Mountains (Sun, 2005). This uplift is regarded as a far-field effect of the India-Asia collision, which caused movement on a series of major strike-slip faults in eastern and central China that are linked to the ‘escape tectonics’ extrusion of large crustal blocks in this region (Liu, 2004; Sun, 2005). One of the most important of these faults is a sinistral strike-slip fault along the base of the Qinling Mountains that marks the boundary between the South China and North China blocks (Peltzer et al., 1985; Meng and Zhang, 2000). Displacements between these blocks are absorbed not only by the Qinling Mountains but also by NW-SE crustal extension across the study area in the form of normal-fault-bounded pullapart basins (Sun, 2005; Sun and Xu, 2007). Uplift of the eastern Qinling Mountains has occurred in four stages since ~1.2 Ma, with episodes dated to ca. 1.2, 0.8, 0.6, and 0.15 Ma (Liu, 2004). These ages agree well with those of river terraces in the Luonan Basin (ca. 1.2e1.1, 0.9e0.8, 0.7e0.6, and 0.2e0.1 Ma) based on a combination of paleomagnetic and climatostratigraphic techniques (Sun, 2005). They also correspond to four intervals of accelerated northward movement of India towards Asia and uplift of the Tibetan Plateau since ~1.2 Ma (An et al., 2001). In this study, proxies for pedogenic weathering and summer monsoon intensity reach high values at ca. 0.8e0.7, 0.6e0.5, and 0.2e0.1 Ma (Fig. 11), approximately matching the timing of the three younger uplift episodes inferred by Liu (2004) and Sun (2005). We propose that this temporal association reflects a causal link between stepwise uplift of the Qinling Mountains and shifts toward warmer/wetter climate conditions within the Luonan Basin. The proximity of the Luonan Basin to the Qinling Mountains makes it likely that sediments of the former will record monsoonal variations induced by uplift of the latter (Zhang et al., 2012). The Qinling Mountains of central China act as a barrier to atmospheric circulation: they obstruct humid, warm air masses of the southeasterly summer monsoon, and dry, cool air masses of the northwesterly winter monsoon. The effect of the latter obstruction is to significantly reduce the amount of loess deposited in the southern and eastern QMB (Cai et al., 2010; Lu et al., 2011). The climatic importance of the Qinling Mountains also lies in the fact that they act as heat sinks in winter and as thermal sources in summer (Cai et al., 2010). Unlike the uplifts of the Qinghai-Tibetan Plateau, which commonly result in climate cooling and drying due to blockage of tropical moisture (Zhang et al., 2012; Zhao et al., 2017), uplifts of the Qinling Mountains are associated with warmer and wetter conditions in the Luonan Basin and, thus, enhanced chemical weathering. Consequently, stepwise uplift of the Qinling Mountains during the Pleistocene has played a key role in triggering climate variations in this region, i.e., weakening the intensity of the EAWM and strengthening the EASM. These changes were recorded as intervals of enhanced pedogenesis in the SBC sequence of the Luonan Basin of central China. 5.5. Implications for hominin settlement and global changes 5.5.1. Paleoenvironmental implications for early humans A wealth of research has shown that the Luonan Basin was a vital area for early human migration and settlement in central China during the Pleistocene (Wang, 2005; Lu et al., 2007, 2011, 2012; Bae, 2010; Zhang et al., 2012; Sun et al., 2013). Numerous in-situ flakes, cores, and stone artifacts discovered from this region, together with magnetic stratigraphy, OSL and carbon-isotope studies, have revealed that early humans repeatedly occupied this area (Lu et al.,

325

2007, 2011). River incision during the Pleistocene generated suitable low-relief surfaces for early human settlements, and evidence of hominin activities has been found in both open fields and cave sites (Sun et al., 2013, 2014). The Luonan Basin was more frequently and densely settled by early humans than the CLP, as indicated by more abundant finds of Paleolithic stone artifacts and other evidence of hominin occupation. The Luonan Basin was more favorable for human settlement owing to a denser vegetative cover and less extreme climate fluctuations, factors that led to more abundant wildlife that may have been an important influence on early humans (Lu et al., 2011, 2012). Dating of hominin artifacts has identified several intervals during which human occupancy of the Luonan Basin intensified, e.g., at 0.8e0.6, 0.4e0.3, and 0.2e0.1 Ma (Fig. 11; Lu et al., 2007, 2011). Geochemical proxies for pedogenic intensity (e.g., CIA, CIW, CPA, CIW, Ba/Sr and Rb/Sr) exhibit two maxima corresponding to pronounced climatic optima at ~0.8 and ~0.3 Ma (Figs. 10 and 11), supporting the results of higher-resolution clay-mineral indices showing evidence of warm and wet conditions at ca. 0.8e0.7 Ma and 0.4e0.2 Ma (Figs. 5 and 11). These intervals are linked to stages of greater hominin occupation and activity within the Luonan Basin. This association implies that paleoclimatic conditions influenced early human settlement patterns in central China, and that increased hominin migration and settlement occurred during warmer, wetter intervals. These relationships among tectonic uplift, climate change, and early human occupation are summarized in Fig. 12. 5.5.2. Comparisons with Quaternary deposits in southern China Unconsolidated Quaternary deposits are also widely distributed in southern China, especially in the middle to lower reaches of the Yangtze River, although their origins remain controversial (Hu et al., 2009). These soils consist dominantly of red earth that has undergone intense chemical weathering (Hu et al., 2009; Hong et al., 2015). They differ in appearance from the SBC deposits of the Luonan Basin, which have a well-defined alternation of loess and paleosol layers exhibiting different structures and colors. In contrast, the red earth profiles in southern China are nearly massive in structure and whereas the latter contain a plenty of exhibit a distinctive net-like pattern of veins within the red matrix (Hong et al., 2016; Zhao et al., 2017). The clay-mineral assemblages of the southern China red-earth deposits differ from that of the SBC in containing abundant vermiculite and mixed-layer clays (e.g., illitesmectite, kaolinite-smectite, and illite-vermiculite; Hong et al., 2007, 2015), which is interpreted as the result of more intense pedogenic weathering in southern China. Greater weathering intensity is reflected in the notably higher CIA values (85e92) of the red earth sediments (Hong et al., 2010a). These differences reflect spatial variation in monsoon intensity, i.e., an extremely strong EASM has dominated the climate of southern China since the midPleistocene (Hu et al., 2009; Qiao et al., 2011), whereas northern and central China have been influenced by competing changes in strength between the EASM and EAWM (Liu and Ding, 1998; An, 2000). 5.5.3. Relationship to global climate changes Representative climatic proxies for the Luonan Basin are generally well correlated with the global benthic d18O record (Fig. 11; Lisiecki and Raymo, 2005). For example, weathering intensity peaks at ca. 850, 780, 700, and 620 ka correspond to high absolute d18O values in the LR04 stack at MIS 21, 19, 17, and 15, respectively. Unit 3 is an exception in that weathering intensity is relatively low and nearly invariant, not lending itself to an obvious correlation with the d18O stack (Figs. 5 and 11). The general correlation between climatic proxies and the d18O stack for the other

326

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

Fig. 12. Generalized relationships among uplift of the Qinling Mountains, monsoon intensity, and early human occupation of the Luonan Basin. Note that the three stages shown in the figure correspond to a single phase of uplift and associated climate change, a sequence that was repeated multiple times during the middle to late Quaternary. (A) Stage 1: EASM and EAWM circulation is unimpeded when the Qinling Mountains are lowlying. (B) Stage 2: Uplift of the Qinling Mountains obstructs monsoonal circulation, weakening the EAWM and strengthening the EASM in the Luonan Basin. (C) Stage 3: Further uplift of the Qinling Mountains triggers warmer and wetter conditions linked to orographic precipitation in the Luonan Basin, promoting early human occupation. The same uplift event results in climate cooling/drying and vegetation reduction to the north of the Qinling Mountains.

units of the SBC sequence suggests that the climate of the Luonan Basin has responded relatively sensitively to global climate

variations since ~780 ka. As is the case for most monsoon systems, the EAM denotes seasonal climate change and is driven by the

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

annual cycle of insolation resulting from ocean-land thermal contrast and ocean/atmosphere interplay (An, 2000; An et al., 2015). The thicknesses of loess-paleosol layers above S*5 (i.e., L*5-S*3) are significantly greater than those of S*6-S*5, implying higher dust fluxes and relative greater aridity in source regions since ~480 ka (Yang et al., 2006). Aridification of the source areas would have enhanced physical erosion and yielded more primary minerals such as illite. This may account for the relatively invariant and strongly illite-enriched mineralogy of Unit 3 (Figs. 5 and 11), as suggested by Gylesjo and Arnold (2006). In addition, cooler and drier depositional environments and higher dust accumulation rates limited pedogenic weathering intensity in the Luonan Basin (Shao et al., 2012). Finally, we used the molar Si/Ti ratio as a proxy to reconstruct variation in the intensity of the EAWM. This proxy is useful for analysis of winter monsoon intensity because 1) the two elements are nearly insoluble and thus largely uninfluenced by weathering processes, and 2) a more intense winter monsoon delivers coarser eolian particles that have higher Si/Ti concentrations (Jahn et al., 2001, and references therein). The profile for Si/Ti is complementary to the profiles of summer monsoon proxies to a certain extent, with intervals of intensified winter monsoons (high Si/Ti) corresponding to weak summer monsoons, and vice-versa (Fig. 10). Intense winter monsoons tend to promote aridity in loess source areas, generating greater amounts of eolian dust (Liu and Ding, 1998). Winter monsoon intensity was high at ~480e330 ka, reflecting a trend toward increasing source-area aridification. This aridification is in good agreement with records from the North Pacific since the mid-Pleistocene (Rea et al., 1998; Yang et al., 2006). The mid-Pleistocene climate transition event, which occurred at ~0.9e0.6 Ma, is thought to have been related not only to solar insolation and global ice volume changes, but also to intensified aridification and expansion of deserts in central Asia (Song et al., 2005, 2014). We therefore suggest that aridification and an intensified winter monsoon in eolian dust source areas, as reflected in unique mineralogic and geochemical features of Unit 3 of the SBC sequence, may be attributed to the mid-Pleistocene climate shift.

6. Conclusions (1) The Shangbaichuan (SBC) loess deposits from the Luonan Basin, eastern Qinling Mountains, central China have a mixed provenance, being partly sourced from arid regions of central Asia through EAWM transport and partly through weathering of local bedrock (the latter being an important source of smectite clays). (2) Major, trace, and rare-earth element compositions indicate that the parent materials for the SBC deposits were wellmixed in the source areas prior to transport and deposition in the Luonan Basin; they had a nearly uniform starting composition throughout the SBC sequence (i.e., prior to local pedogenic weathering). (3) Proxies for chemical weathering suggest an intermediate weathering condition for the SBC sequence that represents more intensive post-depositional weathering than for the CLP loess deposits to the north; this is consistent with differences in climate and vegetation cover between the Luonan Basin and the CLP. (4) The clay-mineral assemblages, Fe-oxide species abundances and ratios, and chemical weathering indices collectively reveal four periods of monsoon evolution in the Luonan Basin, i.e., ~870 to 480 ka (Unit 4), ~480 to 330 ka (Unit 3), ~330 to 180 ka (Unit 2), and ~180 to 10 ka (Unit 1).

327

(5) Based on temporal correlations of weathering intensity variation with patterns of regional tectonic and global climate change, we propose that stepwise tectonic uplift of the Qinling Mountains modulated the strength of the EAM in the Luonan Basin, exerting significant influences on local climate and weathering conditions. (6) The temporal relationship between warm/humid climate stages and early human occupation of the Luonan Basin implies that climatic conditions were an important factor affecting hominin migration and settlement patterns. Acknowledgments This work was supported by the NSFC (Project Number 41272053 and 41472041), NSFC for Young Scholars (41402036 and 41602037), NSF of Hubei for Young Scholars (2016CFB183), Postdoctoral Science Foundation of China (2015M582301), and Fundamental Research Funds for Central Universities (Wuhan, CUG160848). Q. Fang acknowledges the China Scholarship Council (CSC) for financial support (201706410017). We thank Feng Cheng, Hongyu Zhao, Yao Xiao and Na Yue for sample preparations, and Zhao Liu, Shuling Chen, Hao Yang, Yuheng Fang, Tingwang Yin, Haihong Chen and Jishun Yu for their helps with experimental analyses. We also thank Prof. Xiaoping Yang, and two anonymous reviewers for thoughtful and constructive reviews of the manuscript. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.quascirev.2017.05.025 References Ahmad, I., Chandra, R., 2013. Geochemistry of loess-paleosol sediments of Kashmir valley, India: provenance and weathering. J. Asian Earth Sci. 66, 73e89. An, Z., 2000. The history and variability of the East Asian paleomonsoon climate. Quat. Sci. Rev. 19, 171e187. An, Z., Kutzbach, J.E., Prell, W.L., Porter, S.C., 2001. Evolution of Asian monsoons and phased uplift of the Himalaya-Tibetan plateau since Late Miocene times. Nature 411, 62e66. An, Z., Wu, G., Li, J., Sun, Y., Lu, Y., Zhou, W., Cai, Y., Duan, A., Li, L., Mao, J., Cheng, H., Shi, Z., Tan, L., Yan, H., Ao, H., Chang, H., Feng, J., 2015. Global monsoon dynamics and climate change. Annu. Rev. Earth Planet. Sci. 43, 29e77. Bae, C.J., 2010. The late Middle Pleistocene hominin fossil record of eastern Asia: synthesis and review. Am. J. Phys. Anthropol. 143, 75e93. Balsam, W., Ji, J., Chen, J., 2004. Climatic interpretation of the Luochuan and Lingtai loess sections, China, based on changing iron oxide mineralogy and magnetic susceptibility. Earth Planet. Sci. Lett. 223, 335e348. Buggle, B., Glaser, B., Hambach, U., Gerasimenko, N., Markovic, S., 2011. An evaluation of geochemical weathering indices in loess-paleosol studies. Quat. Int. 240, 12e21. €ller, L., Markovic, S.B., Glaser, B., 2014. Iron Buggle, B., Hambach, U., Muller, K., Zo mineralogical proxies and Quaternary climate change in SE-European loesspaleosol sequences. Catena 117, 4e22. Cai, Y., Tan, L., Cheng, H., An, Z., Edwards, R.L., Kelly, M.J., Kong, X., Wang, X., 2010. The variation of summer monsoon precipitation in central China since the last deglaciation. Earth Planet. Sci. Lett. 291, 21e31. Cande, S.C., Kent, D.V., 1995. Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic. J. Geophys. Res. 100, 6093e6095. Chadwick, O.A., Chorover, J., 2001. The chemistry of pedogenic thresholds. Geoderma 100, 321e353. Chamley, H., 1989. Clay Sedimentology. Springer, Berlin, 623 pp. Chen, J., Chen, Y., Liu, L., Ji, J., Balsam, W., Sun, Y., Lu, H., 2006. Zr/Rb ratio in the Chinese loess sequences and its implication for changes in the East Asian winter monsoon strength. Geochimica Cosmochimica Acta 70, 1471e1482. Chen, T., Xu, H., Xie, Q., Chen, J., Ji, J., Lu, H., 2005. Characteristics and genesis of maghemite in Chinese loess and paleosols: mechanism for magnetic susceptibility enhancement in paleosols. Earth Planet. Sci. Lett. 240, 790e802. Chen, T., Xie, Q., Xu, H., Chen, J., Ji, J., Lu, H., Balsam, W., 2010. Characteristics and formation mechanism of pedogenic hematite in Quaternary Chinese loess and paleosols. Catena 81, 217e225. Churchman, J., Lowe, D.J., 2012. Alteration, formation, and occurrence of minerals in

328

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329

soils. In: Huang, P.M., Li, Y., Sumner, M.E. (Eds.), Handbook of Soil Sciences, second ed. CRC Press, Boca Raton, Florida, pp. 20.1e20.72. Clift, P.D., Wan, S., Blusztajn, J., 2014. Reconstructing chemical weathering, physical erosion and monsoon intensity since 25 Ma in the northern South China Sea: a review of competing proxies. Earth-Sci. Rev. 130, 86e102. Ding, Z.L., Sun, J.M., Yang, S.L., Liu, T.S., 2001. Geochemistry of the Pliocene red clay formation in the Chinese Loess Plateau and implications for its origin, source provenance and paleoclimate change. Geochimica Cosmochimica Acta 65, 901e913. Eberl, D.D., Smith, D.B., 2009. Mineralogy of soils from two continental-scale transects across the United States and Canada and its relation to soil geochemistry and climate. Appl. Geochem. 24, 1394e1404. Fang, Q., Churchman, G.J., Hong, H., Chen, Z.-Q., Liu, J., Yu, J., Han, W., Wang, C., Zhao, L., Furnes, H., 2017a. New insights into microbial smectite illitization in the Permo-Triassic boundary K-bentonites. Appl. Clay Sci. 140, 96e111. Fang, Q., Hong, H., Chen, Z., Yu, J., Wang, C., Yin, K., Zhao, L., Liu, Z., Cheng, F., Gong, N., Furnes, H., 2017b. Microbial proliferation coinciding with volcanism during the Permian-Triassic transition: new, direct evidence from volcanic ashes, South China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 474, 164e186. Fedo, C.M., Nesbitt, H.W., Young, G.M., 1995. Unraveling the effects of potassium metasomatism in sedimentary rocks and paleosols, with implications for paleoweathering conditions and provenance. Geology 23, 921e924. Fitzsimmons, K.E., Markovic, S.B., Hambach, U., 2012. Pleistocene environmental dynamics recorded in the loess of the middle and lower Danube basin. Quat. Sci. Rev. 41, 104e118. Gallet, S., Jahn, B., Torii, M., 1996. Geochemical characterization of the Luochuan loess-paleosol sequence, China, and paleoclimatic implications. Chem. Geol. 133, 67e88. Ge, J., Dai, Y., Zhang, Z., Zhao, D., Li, Q., Zhang, Y., Yi, L., Wu, H., Oldfield, F., Guo, Z., 2013. Major changes in East Asian climate in the mid-Pliocene: triggered by the uplift of the Tibetan Plateau or global cooling? J. Asian Earth Sci. 69, 48e59. Gylesjo, S., Arnold, E., 2006. Clay mineralogy of a red clay-loess sequence from Lingtai, the Chinese Loess Plateau. Glob. Planet. Change 51, 181e194. Harnois, L., 1988. The CIW index: a new chemical index of weathering. Sediment. Geol. 55, 319e322. Hong, H., Li, Z., Xue, H., Zhu, Y., Zhang, K., Xiang, S.Y., 2007. Oligocene clay mineralogy of the Linxia Basin: evidence of palaeoclimatic evolution subsequent to the initial-stage uplift of the Tibetan Plateau. Clays Clay Minerals 55, 492e505. Hong, H., Gu, Y., Li, R., Zhang, K., Li, Z., 2010a. Clay mineralogy and geochemistry and their palaeoclimatic interpretation of the Pleistocene deposits in the Xuancheng section, South China. J. Quat. Sci. 25, 662e674. Hong, H., Zhang, K., Li, Z., 2010b. Climatic and tectonic uplift evolution since ~7 Ma in Gyirong basin, southwestern Tibet plateau: clay mineral evidence. Int. J. Earth Sci. 99, 1305e1315. Hong, H., Cheng, F., Yin, K., Churchman, G.J., Wang, C., 2015. Three-component mixed-layer illite/smectite/kaolinite (I/S/K) minerals in hydromorphic soils, south China. Am. Mineralogist 100, 1883e1891. Hong, H., Fang, Q., Cheng, L., Wang, C., Churchman, G.J., 2016. Microorganisminduced weathering of clay minerals in a hydromorphic soil. Geochimica Cosmochimica Acta 184, 272e288. Hosek, J., Hambach, U., Lisa, L., Grygar, T.M., Horacek, I., Meszner, S., Knesl, I., 2015. An integrated rock-magnetic and geochemical approach to loess/paleosol sequences from Bohemia and Moravia (Czech Republic): implications for the Upper Pleistocene paleoenvironment in central Europe. Palaeogeogr. Palaeoclimatol. Palaeoecol. 418, 344e358. Hu, P., Liu, Q., Torrent, J., Barron, V., Jin, C., 2013. Characterizing and quantifying iron oxides in Chinese loess/paleosols: implications for pedogenesis. Earth Planet. Sci. Lett. 369e370, 271e283. Hu, X., Wei, J., Xu, L., Zhang, G., Zhang, W., 2009. Magnetic susceptibility of the Quaternary Red Clay in subtropical China and its paleoenvironmental implications. Palaeogeogr. Palaeoclimatol. Palaeoecol. 279, 216e232. Huang, C., Zhao, W., Li, F., Tan, W., Wang, M., 2012. Mineralogical and pedogenetic evidence for palaeoenvironmental variations during the Holocene on the Loess Plateau, China. Catena 96, 49e56. Jahn, B., Gallet, S., Han, J., 2001. Geochemistry of the Xining, Xifeng and Jixian sections, Loess Plateau of China: eolian dust provenance and paleosol evolution during the last 140 ka. Chem. Geol. 178, 71e94. Jeong, G.Y., Hillier, S., Kemp, R.A., 2008. Quantitative bulk and single-particle mineralogy of a thick Chinese loess-paleosol section: implications for loess provenance and weathering. Quat. Sci. Rev. 27, 1271e1287. Jeong, G.Y., Hillier, S., Kemp, R.A., 2011. Changes in mineralogy of loess-paleosol sections across the Chinese Loess Plateau. Quat. Res. 75, 245e255. Ji, J., Balsam, W., Chen, J., 2001. Mineralogic and climatic interpretations of the Luochuan loess section (China) based on diffuse reflectance spectrophotometry. Quat. Res. 56, 23e30. Jiang, H., Guo, G., Cai, X., Thompson, J.A., Xu, H., Zhong, N., 2016. Geochemical evidence of windblown origin of the Late Cenozoic lacustrine sediments in Beijing and implications for weathering and climate change. Palaeogeogr. Palaeoclimatol. Palaeoecol. 446, 32e43. Kahle, M., Kleber, M., Jahn, R., 2002. Review of XRD-based quantitative analyses of clay minerals in soils: the suitability of mineral intensity factors. Geoderma 109, 191e205. Kalm, V.E., Rutter, N.W., Rokosh, C.D., 1996. Clay minerals and their paleoenvironmental interpretation in the Baoji loess section, Southern Loess Plateau, China. Catena 27, 49e61.

Kukla, G., An, Z.S., 1989. Loess stratigraphy in central China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 72, 203e225. Lan, X., Li, R., Mi, B., Zhang, Z., Guo, X., Huang, L., 2016. Distribution characteristics of rare earth elements in surface sediment and their provenance discrimination in the eastern Bohai and northern Yellow Seas. Earth Sci. 41, 463e474 (in Chinese with English abstract). Li, Y., Song, Y., Chen, X., Li, J., Mamadjanov, Y., Aminov, J., 2016. Geochemical composition of Tajikistan loess and its provenance implications. Palaeogeography, Palaeoclimatology. Palaeoecology 446, 186e194. Lisiecki, L., Raymo, M., 2005. A Pliocene-Pleistocene stack of 57 globally distributed benthic d18O records. Paleoceanography 20, PA1003. Liu, H., 2004. Formation and Evolution of the Wei He River Basin and Uplift of the Eastern Qinling Mountains. Doctoral dissertation. Northwest University, Xi’an, China, 70 pp. (in Chinese with English abstract). Liu, T.S., Ding, Z.L., 1998. Chinese loess and the paleomonsoon. Annu. Rev. Earth Planet. Sci. 26, 111e145. Liu, Z., Colin, C., Trentesaux, A., Siani, G., Frank, N., Blamart, D., Farid, S., 2005. Late quaternary climatic control on erosion and weathering in the eastern Tibetan plateau and the Mekong Basin. Quat. Res. 63, 316e328. Long, X., Ji, J., Barron, V., Torrent, J., 2016. Climatic thresholds for pedogenic iron oxides under aerobic conditions: processes and their significance in paleoclimate reconstruction. Quat. Sci. Rev. 150, 264e277. Lu, H.Y., 2015. Driving force behind global cooling in the Cenozoic: an ongoing mystery. Sci. Bull. 60, 2091e2095. Lu, H.Y., Zhang, F., Liu, X., Duce, R.A., 2004. Periodicities of palaeoclimatic variations recorded by loess-paleosol sequences in China. Quat. Sci. Rev. 23, 1891e1900. Lu, H.Y., Zhang, H., Wang, S., Richard, C., Zhao, C.F., Thomas, S., Zhao, J., 2007. A preliminary survey on loess deposit in Eastern Qinling Mountains (central China) and its implication for estimating age of the Pleistocene lithic artifacts. Quat. Sci. 27, 559e567 (in Chinese with English abstract). Lu, H.Y., Zhang, H., Wang, S., Cosgrove, R., Sun, X., Zhao, J., Sun, D., Zhao, C., Shen, C., Wei, M., 2011. Multiphase timing of hominin occupations and the paleoenvironment in Luonan Basin, Central China. Quat. Res. 76, 142e147. Lu, H.Y., Zhang, H.Y., Sun, X.F., Wang, S.J., Cosgrove, R., Shen, C., Zhang, W.C., Zhang, X.B., Wang, X.Y., Yi, S.W., Ma, X.L., Wei, M., 2012. Landform, loess deposit and palaeoenvironmental changes in the South Luohe River (central China) during the hominin occupations. Quat. Sci. 32, 167e177 (in Chinese with English abstract). Lü, X.X., Versteegh, G.J.M., Song, J.M., Li, X.G., Yuan, H.M., Li, N., 2016. Geochemistry of middle Holocene sediments from south Yellow Sea: implications to provenance and climate change. J. Earth Sci. 277, 51e762. Lyons, R., Tooth, S., Duller, G.A., 2014. Late Quaternary climatic changes revealed by luminescence dating, mineral magnetism and diffuse reflectance spectroscopy of river terrace palaeosols: a new form of geoproxy data for the southern African interior. Quat. Sci. Rev. 95, 43e59. McLennan, S.M., 1993. Weathering and global denudation. J. Geol. 101, 295e303. Meng, Q.R., Zhang, G.W., 2000. Geologic framework and tectonic evolution of the Qinling orogen, central China. Tectonophysics 323, 183e196. Miao, W., Fan, Q., Wei, H., Zhang, X., Ma, H., 2016. Clay mineralogical and geochemical constraints on late Pleistocene weathering processes of the Qaidam Basin, northern Tibetan Plateau. J. Asian Earth Sci. 127, 267e280. Moore, D.M., Reynolds, R.C., 1989. X-ray Diffraction and the Identification and Analysis of Clay Minerals. Oxford University Press, Oxford, 332 pp. Muhs, D.R., Bettis, E.A., Aleinikoff, J.N., McGeehin, J.P., Beann, J., Skipp, G., Marshall, B.D., 2008. Origin and paleoclimatic significance of Quaternary loess in Nebraska: evidence from stratigraphy, chronology, sedimentology, and geochemistry. Geol. Soc. Am. Bull. 120, 1378e1407. Nesbitt, H.W., Young, G.M., 1982. Early Proterozoic climates and plate motions inferred from major element chemistry of lutites. Nature 299, 715e717. Nesbitt, H.W., Young, G.M., 1989. Formation and diagenesis of weathering profiles. J. Geol. 97, 129e147. Oh, H., Shin, H.J., 2016. Climatic classification over Asia during the middle Holocene climatic optimum based on PMIP models. J. Earth Sci. 27, 123e129. Parker, A., 1970. An index of weathering for silicate rocks. Geol. Mag. 107, 501e504. Peltzer, G., Tapponnier, P., Zhang, Z.T., Xu, Z.Q., 1985. Neogene and quaternary faulting in and along the Qinling Shan. Nature 317, 500e505. Price, J.R., Velbel, M.A., 2003. Chemical weathering indices applied to weathering profiles developed on heterogeneous felsic metamorphic parent rocks. Chem. Geol. 202, 397e416. Qiao, Y., Hao, Q., Peng, S., Wang, Y., Li, J., Liu, Z., 2011. Geochemical characteristics of the eolian deposits in southern China, and their implications for provenance and weathering intensity. Palaeogeogr. Palaeoclimatol. Palaeoecol. 308, 513e523. Rea, D.K., Snoeckx, H., Joseph, L.H., 1998. Late Cenozoic eolian deposition in the North Pacific: Asian drying, Tibetan uplift, and cooling of the northern hemisphere. Paleoceanography 13, 215e224. Schatz, A.K., Scholten, T., Kuhn, P., 2015. Paleoclimate and weathering of the Tokaj (Hungary) loess-paleosol sequence. Palaeogeogr. Palaeoclimatol. Palaeoecol. 426, 170e182. Scheinost, A.C., Chavernas, A., Barron, V., Torrent, J., 1998. Use and limitations of second derivative diffuse reflectance spectroscopy in the visible to nearinfrared range to identify and quantity Fe oxide minerals in soils. Clays Clay Minerals 46, 528e536. Shao, J., Yang, S., Li, C., 2012. Chemical indices (CIA and WIP) as proxies for integrated chemical weathering in China: inferences from analysis of fluvial

Q. Fang et al. / Quaternary Science Reviews 169 (2017) 312e329 sediments. Sediment. Geol. 265e266, 110e120. Sheldon, N.D., Tabor, N.J., 2009. Quantitative paleoenvironmental and paleoclimatic reconstruction using paleosols. Earth Sci. Rev. 95, 1e52. Song, Y., Fang, X., Naoto, I., Masayuki, T., Li, J., Shi, Z., Qiang, X., Fu, C., 2005. Rock magnetic evidences of middle Pleistocene climate transition recorded in Chinese loess-paleosol sequence. Mar. Geol. Quat. Geol. 25, 43e51 (in Chinese with English abstract). Song, Y., Fang, X., King, J.W., Li, J., Naoto, I., An, Z., 2014. Magnetic parameter variations in the Chaona loess/paleosol sequences in the central Chinese Loess Plateau, and their significance for the middle Pleistocene climate transition. Quat. Res. 8, 433e444. Sun, J., 2005. Long-term fluvial archives in the Fen Wei Graben, central China, and their bearing on the tectonic history of the India? Asia collision system during the Quaternary. Quat. Sci. Rev. 24, 1279e1286. Sun, J., Xu, L., 2007. River terraces in the Fen Wei Graben, central China, and the relation with the tectonic history of the India-Asia collision system during the Quaternary. Quat. Sci. 27, 20e26 (in Chinese with English abstract). Sun, X., Lu, H., Wang, S., Yi, S., Shen, C., Zhang, W., 2013. TT-OSL dating of Longyadong Middle Paleolithic site and paleoenvironmental implications for hominin occupation in Luonan Basin (central China). Quat. Res. 79, 168e174. Sun, X., Lu, H., Wang, S., Cosgrove, R., Zhang, W., Yu, K., Zhuo, H., 2014. Age of newly discovered paleolithic assemblages at liuwan site Luonan Basin, central China. Quat. Int. 347, 193e199. Sun, Y., Clemens, S.C., Morrill, C., Lin, X., Wang, X., An, Z., 2011. Influence of Atlantic meridional overturning circulation on the East Asian winter monsoon. Nat. Geosci. 5, 46e49. Sun, Y., Kutzbach, J., An, Z., Clemens, S., Liu, Z., Liu, W., Liu, X., Shi, Z., Zheng, W., Liang, L., Yan, Y., Li, Y., 2015. Astronomical and glacial forcing of East Asian summer monsoon variability. Quat. Sci. Rev. 115, 132e142. Sun, Z., Owens, P.R., Han, C., Chen, H., Wang, X., Wang, Q., 2016. A quantitative reconstruction of a loess-paleosol sequence focused on paleosol genesis: an example from a section at Chaoyang, China. Geoderma 266, 25e39. Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: its Composition and Evolution. Blackwell Scientific Publications, Oxford, 312 pp.  n, V., Liu, Q., 2006. Magnetic enhancement is linked to and precedes Torrent, J., Barro hematite formation in aerobic soil. Geophys. Res. Lett. 33, 1e4. Torrent, J., Liu, Q., Bloemendal, J., Barron, V., 2007. Magnetic enhancement and iron oxides in the upper Luochuan loess-paleosol sequence, Chinese Loess Plateau. Soil Sci. Soc. Am. J. 71, 1570e1578. Udvardi, B., Kovacs, I.J., Szabo, C., Falus, G., Ujvari, G., Besnyi, A., Bertalan, E., Budai, F., Horvath, Z., 2016. Origin and weathering of landslide material in a loess area: a geochemical study of the Kulcs landslide, Hungary. Environ. Earth Sci. 75 (1299), 19 pp. Varga, A., Ujvari, G., Raucsik, B., 2011. Tectonic versus climatic control on the evolution of a loess-paleosol sequence at Beremend, Hungary: an integrated approach based on paleoecological, clay mineralogical, and geochemical data. Quat. Int. 240, 71e86. Wang, F., Li, H., Zhu, R., Hu, Y., 2002. Incision speed and tectonic uplift of the middle Qinling Mountains since the late Quaternary. Sci. Bull. 47, 1032e1036 (in Chinese with English abstract). Wang, P., Clemens, S., Beaufort, L., Braconnot, P., Ganssen, G., Jian, Z., Kershaw, P., Sarnthein, M., 2005. Evolution and variability of the Asian monsoon system: state of the art and outstanding issues. Quat. Sci. Rev. 24, 595e629. Wang, Q., Yang, S., 2013. Clay mineralogy indicates the Holocene monsoon climate

329

in the Changjiang (Yangtze River) catchment, China. Appl. Clay Sci. 74, 28e36. Wang, S.J., 2005. Perspectives on Hominid Behaviour and Settlement Patterns: a Study of the Lower Palaeolithic Sites in the Luonan Basin, China. BAR International Series, 1406. Archaeopress, Oxford, 248 pp. Wang, S.J., Cosgrove, R., Lu, H.Y., Shen, C., Wei, M., Zhang, X.B., 2008. New progress on paleolithic archaeological studies in the Luonan Basin, China. In: Matsufuji, K. (Ed.), Loess-paleosol and Paleolithic Chronology in East Asia. Yuzankaku, Tokyo, pp. 145e161. Wang, X., Lu, H., Zhang, W., Hu, P., Zhang, H., Han, Z., Wang, S., Li, B., 2016. Rock magnetic investigation of loess deposits in the Eastern Qingling Mountains (central China) and its implications for the environment of early humans. Geophys. J. Int. 207, 889e900. Wang, C., Hong, H., Abels, H.A., Li, Z., Cao, K., Yin, K., Song, B., Xu, Y., Ji, J., Zhang, K., 2016. Early middle Miocene tectonic uplift of the northwestern part of the Qinghai-Tibetan Plateau evidenced by geochemical and mineralogical records in the western Tarim Basin. Int. J. Earth Sci. 105, 1021e1037. Weaver, C.E., 1989. Clays, Muds, and Shales. Developments in Sedimentology, vol. 44. Elsevier, Amsterdam, 818 pp. Won, C., Hong, H., Cheng, F., Fang, Q., Wang, C., Zhao, L., Churchman, G.J., 2017. Clay mineralogy and its palaeoclimatic significance of the Luochuan loess-palaeosols since ~1.3 Ma, Shaanxi, northwestern China. Front. Earth Sci. http://dx.doi.org/ 10.1007/s11707-017-0625-4 (in press). Xiao, J.L., Jin, C.Z., Zhu, Y.Z., 2002. Age of the fossil Dali Man in north-central China deduced from chronostratigraphy of the loess-paleosol sequence. Quat. Sci. Rev. 21, 2191e2198. Xiong, S., Ding, Z., Zhu, Y., Zhou, R., Lu, H., 2010. A ~6 Ma chemical weathering history, the grain size dependence of chemical weathering intensity, and its implications for provenance change of the Chinese loess-red clay deposit. Quat. Sci. Rev. 29, 1911e1922. Xue, X., 1987. Human fossil tooth from Luonan, Shaanxi and its geological age. Acta Anthropol. Sin. 6, 284e288 (in Chinese with English abstract). Xue, X., Li, H., Li, Y., Liu, H., 2004. The new data of the uplifting of Qinling Mountains since the middle Pleistocene. Quat. Sci. 24, 82e87 (in Chinese with English abstract). Yang, S., Ding, F., Ding, Z., 2006. Pleistocene chemical weathering history of Asian arid and semi-arid regions recorded in loess deposits of China and Tajikistan. Geochimica Cosmochimica Acta 70, 1695e1709. Yang, X., Liu, Y., Li, C., Song, Y., Zhu, H., Jin, X., 2007a. Rare earth elements of aeolian deposits in Northern China and their implications for determining the provenance of dust storms in Beijing. Geomorphology 87, 365e377. Yang, X., Zhu, B., White, P.D., 2007b. Provenance of aeolian sediment in the Taklamakan Desert of western China, inferred from REE and major-elemental data. Quat. Int. 175, 71e85. Yin, G.M., Bahain, J.J., Shen, G.J., Tissoux, H., Falgueres, C., Dolo, J.M., Han, F., Shao, Q.F., 2011. ESR/U-series study of teeth recovered from the palaeoanthropological stratum of the Dali Man site (Shaanxi Province, China). Quat. Geochronol. 6, 98e105. Zhang, H., Lu, H., Jiang, S., Vandenberghe, J., Wang, S., Cosgrove, R., 2012. Provenance of loess deposits in the Eastern Qinling Mountains (central China) and their implications for the paleoenvironment. Quat. Sci. Rev. 43, 94e102. Zhao, L., Hong, H., Fang, Q., Yin, K., Wang, C., Li, Z., Torrent, J., Cheng, F., Algeo, T.J., 2017. Monsoonal climate evolution in southern China since 1.2 Ma: new constraints from Fe-oxide records in red earth sediments from the Shengli section, Chengdu Basin. Palaeogeogr. Palaeoclimatol. Palaeoecol. 473, 1e15.