East Asian monsoon variability over the last ... - Wiley Online Library

Geochemistry Geophysics Geosystems

3

G

Article Volume 7, Number 12 29 December 2006 Q12Q02, doi:10.1029/2006GC001287

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society

ISSN: 1525-2027

Click Here for

Full Article

East Asian monsoon variability over the last seven glacial cycles recorded by a loess sequence from the northwestern Chinese Loess Plateau Youbin Sun Department of Earth and Planetary Sciences, University of Tokyo, Hongo 7-3-1, Tokyo 113-0033, Japan Institute of Surfacial Geochemistry, Department of Earth Science, Nanjing University, Nanjing 210093, China Now at State Key Laboratory of Loess and Quarternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi’an 710075, China ([email protected])

Jun Chen Institute of Surficial Geochemistry, Department of Earth Science, Nanjing University, Nanjing 210093, China

Steven C. Clemens Department of Geology, Brown University, Providence, Rhode Island 02912-1846, USA

Qingsong Liu National Oceanography Centre, University of Southampton, European Way, Southampton, SO14 3ZH, UK

Junfeng Ji Institute of Surficial Geochemistry, Department of Earth Science, Nanjing University, Nanjing 210093, China

Ruiji Tada Department of Earth and Planetary Sciences, University of Tokyo, Hongo 7-3-1, Tokyo 113-0033, Japan

[1] A 180-m-thick loess-paleosol sequence from the northwestern Chinese Loess Plateau was investigated to construct a high-resolution record of the East Asian monsoon variability over the last seven glacialinterglacial cycles. The low-field magnetic susceptibility (c, mass-specific) and the mean grain size are used as proxies for changes in the intensity of the East Asian summer and winter monsoon, respectively. Because of the weaker pedogenesis at the northwestern Chinese Loess Plateau compared to the central Chinese Loess Plateau, our c and mean grain size records show a muted glacial-interglacial contrast for the Asian summer monsoon but an enhanced contrast for the Asian winter monsoon. Although better resolved, most orbital-scale East Asian monsoon variations captured by our c and grain size records are similar to those reported from the central Chinese Loess Plateau. Nevertheless, variations in c exhibit clear precessional cycles in three paleosol layers (i.e., S1, S2, and S3), corresponding with solar insolation maxima. Furthermore, unlike c records at the central Chinese Loess Plateau where c is dramatically enhanced at paleosol S5SS1 (corresponding to marine isotope stage 13), our new c record shows a major enhancement at paleosol S4 (corresponding to marine isotope stage 11), which indicates geographic differences in the timing of local monsoon precipitation in the two regions.

Copyright 2006 by the American Geophysical Union

1 of 16


3

G

sun et al.: east asian monsoon variability

10.1029/2006GC001287

Components: 9541 words, 9 figures, 1 table. Keywords: Chinese Loess Plateau; loess-paleosol sequence; magnetic susceptibility; mean grain size; East Asian monsoon. Index Terms: 1616 Global Change: Climate variability (1635, 3305, 3309, 4215, 4513); 1637 Global Change: Regional climate change; 3344 Atmospheric Processes: Paleoclimatology (0473, 4900). Received 23 February 2006; Revised 28 September 2006; Accepted 18 October 2006; Published 29 December 2006. Sun, Y., J. Chen, S. C. Clemens, Q. Liu, J. Ji, and R. Tada (2006), East Asian monsoon variability over the last seven glacial cycles recorded by a loess sequence from the northwestern Chinese Loess Plateau, Geochem. Geophys. Geosyst., 7, Q12Q02, doi:10.1029/2006GC001287. ————————————

Theme: Eolian Dust as a Player and Recorder of Environmental Change Guest Editors: Jan-Berend Stuut, Maarten Prins, and Denis-Didier Rousseau

1. Introduction [2] The East Asian monsoon (EAM) is generated by the heating contrast between the Pacific Ocean and the Asian landmass, and is further enhanced by the thermal and dynamic effects of the Tibetan Plateau [Webster et al., 1998]. During warm and wet phases, the Asian summer monsoon transports heat and moisture from the equatorial Pacific Ocean to the Asian continental interior [L. Chen et al., 1991; Zhang and Lin, 1992; Emile-Geay et al., 2003]. In contrast, during cold and dry phases, the Asian winter monsoon and northern hemisphere westerly jet flow across eastern Asia and transport dust from Asian deserts to eastern Asia and north pacific regions [e.g., An et al., 1991a; Rea, 1994; An, 2000; Irino and Tada, 2003]. Those living within monsoon-influenced regions are strongly impacted by changes in these circulation regimes. Studies of past changes in monsoon circulation at different timescales would aid in defining the underlying mechanisms and hence in predicting future climate change scenarios with regard to the EAM. [3] Loess deposits in northern China provide direct and quasi-continuous records of changes in the EAM at different timescales [e.g., Liu and Ding, 1998; An, 2000; An et al., 2001]. Over the last several million years, EAM variability was dominated by changes in solar insolation and Northern Hemisphere ice volume [e.g., An et al., 1990; Ding et al., 1995; Liu and Ding, 1998; An, 2000], as well as processes operating at much shorter and more abrupt timescales [Porter and An, 1995; Guo et al., 1996; Chen et al., 1997; Ding et al., 1999a; Wang et al., 2001]. Previous long-term reconstructions of

EAM evolution focused mainly on the loesspaleosol sequences in the central Chinese Loess Plateau (CLP) where linkages among paleoclimatic proxies and monsoonal fluctuations are relatively simple; both the Asian summer and winter monsoons are strong in this region, resulting in distinct glacial-interglacial contrasts [e.g., Liu and Ding, 1998; An, 2000; Ding et al., 2002; Sun et al., 2006a; Deng et al., 2005, 2006]. [4] Recently, attention has turned to the western CLP (west of Liupan Mountains) to better assess geographic patterns of the EAM variation over the whole CLP [e.g., Chen et al., 1997; Ding et al., 1999a; Guo et al., 2002; Wu et al., 2005]. Most of previous studies in the western CLP focused on the last two glacial cycles [e.g., Chen et al., 1999; Fang et al., 1999; Ding et al., 1999a; Parker and Bloemandal, 2005]. To extend our understanding of the paleoclimate history prior to the last two glacial cycles, we investigate a 180-m-thick loesspaleosol sequence at the Jingyuan section from the northwestern CLP. This site is characterized by a high sedimentation rate (10–100 cm/kyr) [Yue et al., 1991; Lei, 1995], which permits a highresolution study of high-frequency EAM variability. Moreover, compared to loess/paleosol sequences at the central CLP, the relatively weaker effects of the summer monsoon at the western CLP minimize post-depositional pedogenic alterations, which improve the fidelity of the monsoon signals recorded by both grain size and magnetic susceptibility variations [e.g., Chen et al., 1997; Fang et al., 1999; Ding et al., 1999a]. [5] On the basis of the correlation among the monsoon records (i.e., magnetic susceptibility and grain size), the stacked benthic d 18 O record 2 of 16


3

G


10.1029/2006GC001287

Figure 1. (a) Map showing the location of the Jingyuan (JY), Lingtai (LT), and Zhaojiachuan (ZJC) loess sections in the Chinese Loess Plateau (CLP). (b) Regional map showing that the sampling site is located in the south of Jingyuan town. Deserts adjacent to the CLP and mountains within the CLP are also shown.

[Lisiecki and Raymo, 2005], and the summer insolation at 65°N [Laskar, 1990], we construct an astronomical timescale for the loess-paleosol records. By applying the low-field magnetic susceptibility (c, mass-specific) and the mean grain size (MGS) as indicators of the summer and winter monsoons intensity, respectively, we then evaluate temporal variations of the EAM over the past seven climatic cycles. Finally, we address geographic variability of the EAM by comparing the monsoon indices generated from the loess sequences in the northwest and central CLP.

2. Sampling and Methods [6] The Jingyuan (JY) section (104.6°E, 36.35°N, 2120 m above sea level) is situated in the northwest-

ern part of the CLP at a distance of 80 km from the modern Tengger Desert margin and 28 km south of Jingyuan town (Figure 1). In this region, the mean annual temperature and precipitation are 8.3°C and 275 mm, respectively. The field outcrop is developed on the sixth terrace of the Yellow River (Figure 2a) and has a thickness of 500 m with a basal age of 1.4 Ma [Yue et al., 1991]. The Holocene soil (S0) and last interglacial paleosol (S1) at this site are well documented by Heslop et al. [1999] and Chen et al. [1999]. Here we extend the work to greater depth covering the upper 180-m of the JY section. [7] To a first order, the major loess-paleosol cycles at this section can be identified visually by changes in soil color and structure. The Holocene soil (S0) is dark in color (rich in organic matter), and contains

Figure 2. (a) A profile shows that the outcrop is well exposed on the sixth terrace of Yellow River (modified after Li and Kang [1989]). Two pictures show the working sections: (b) the uppermost part (Holocene soil) and (c) the underlying part (last glacial loess). 3 of 16


3

G


10.1029/2006GC001287

Figure 3. (a) Pedostratigraphy of the JY section. Depth plots of variations in (b) c, (c) sand content, and (d) MGS records of the JY section. (e) The stacked marine oxygen isotope record [Lisiecki and Raymo, 2005]. (f) The La90(1,1) summer insolation at 65°N [Laskar, 1990]. Gray bars indicate palaeosol units and interglacial stages. Solid lines indicate the correlation between the loess-paleosol cycles and marine isotope stages (MIS). Dashed lines are tie points linking the monsoon proxies to the insolation cycles (IC). MIS and IC numbers as well as the loess-paleosol units are labeled.

abundant root or earthworm holes (Figure 2b). The loess layers have a yellowish color and a massive microstructure. Fourteen paleosol units are readily identified by their light reddish color and weak biological microstructures. Pedogenic differentiation of those paleosols is weak and they can thus be regarded as A horizon (topsoil). [8] The loess-paleosol units and the exact stratigraphic boundaries are further determined by variations in the paleoclimatic proxies (Figures 3a–3d). Among paleosol units, four paleosol complexes (S1, S2, S3 and S5) are composed of three sub-paleosol layers interbedded with two thin loess horizons.

This stratigraphic pattern correlates well with the loess-paleosol sequences above L7 in the central CLP, but with a much higher resolution. The depth of each paleosol layer is summarized in Table 1. The thickness of loess deposits since the upper L7 (720 ka) is about 180 m, which is almost threefold thicker than previous reported loess sections from the central CLP [e.g., Ding et al., 2002; Sun et al., 2006a]. This permits a much higher resolution evaluation of EAM variability than previously was possible. [9] Since loess outcrops are not well exposed in this region, to get fresh samples, the surface 4–5 m 4 of 16


3

G


10.1029/2006GC001287

Table 1. Thickness and Estimated Ages of Paleosol Units at the JY Section, and Their Comparison With the Ages of Those Paleosol Layers Derived From Recently Published Age Modelsa Units

Depth

Age (JY)

S0 top S0 base S1 top S1 base S2 top S2 base S3 top S3 base S4 top S4 base S5SS1 top S5SS1 base S5SS2 top S5SS3 base S6 top S6 base Reference

0 1.7 37.5 45.6 66.6 76.7 89.4 96.5 116.5 119.9 139.6 142.4 147.8 153.1 173.5 175.5 this study

0 11 73 129 189 244 280 335 396 424 473 506 567 624 682 710 this study

Age (LC)

71 129 188 254 279 334 385 428 471 576 658 670 Lu et al. [1999]

Age (BJ)

Age (5 Sections)

Age (LT/ZJC)

79 129 196 250 290 342 386 417 503 556 568 625 693 713 Heslop et al. [2000]

0 11 73 128 190 245 307 336 360 412 479 531 549 622 684 710 Ding et al. [2002]

0 11 72 128 190 245 279 336 394 426 473 506 558 622 682 712 Sun et al. [2006a]

a Thickness is in meters; estimated ages are in ka. Note that although all those age models are generated by orbital tuning method, different proxy indicators and tuning targets are employed in the chronological construction. Detailed descriptions of those age models are given by Lu et al. [1999], Heslop et al. [2000], Ding et al. [2002], and Sun et al. [2006a]. JY, Jingyuan; LC, Luochuan; BJ, Baoji; ZJC, Zhaojiachuan; LT, Lingtai. Five sections of Ding et al. [2002] are located in Jingchuan, Lingtai, Puxian, Baoji, and Pingling.

of sediment was removed to eliminate possible surface weathering effects and disturbance due to vegetation (Figure 2c). A total of 1800 samples were taken at 10 cm intervals for both c and grain size analyses. c was measured using a HKB-3 Digital Kappabridge at the Department of Earth Science, Nanjing University. [10] To evaluate the influence of organic matter and carbonate on the grain size composition, 50 samples were pretreated using three different methods: (1) removal of organic matter with H2O2, (2) removal of carbonate with HCl, and (3) no removal of either organic matter or carbonate. Similar to previous study of Ding et al. [1999a], mean grain size shows only a slight decrease after the removal of organic matter, but exhibits undetectable differences before and after the removal of carbonate. Thus the correction for the removal of organic matter and carbonate was not necessary for the loess deposits at the JY section. [11] Prior to grain size measurement, ultrasonic pretreatment with 10 mL 10% (NaPO3)6 solution was used to disperse samples. Grain size composition of all samples was determined using a Marvin 2000 laser instrument at the Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences. Replicate analyses (n = 30) on two samples (one loess at 5 m and a paleosol sample at 45 m) indicate that the mean grain size has an analytical error of less than 2%, corresponding to

0.4 mm and 0.8 mm for the paleosol and loess samples, respectively.

3. Monsoon Proxies and Chronology 3.1. c as a Proxy Indicator of the East Asian Summer Monsoon Intensity [12] It is generally accepted that during warm periods, intensified summer monsoon circulation has a significant impact on the formation of paleosols through pedogenesis [e.g., Zhou et al., 1990; An et al., 1991b; Maher and Thompson, 1992; Liu et al., 2004]. During pedogenic processes, finegrained magnetic minerals (dominantly maghemite in superparamagnetic and single domain grain size regions) are formed, and further enhance the bulk c. Although arguments still exist regarding the origins of c enhancements in Chinese loess, there is a consensus that c can be used to quantify the amount of paleorainfall, an indicator of East Asian summer monsoon intensity [e.g., Kukla et al., 1988; Kukla and An, 1989; An et al., 1990, 1991b; Liu et al., 1995; Bloemendal et al., 1995; Han et al., 1996; Chen et al., 1997; Liu et al., 2004; Bloemendal and Liu, 2005]. [13] On the basis of low-temperature techniques, Liu et al. [2005a] quantified a continuous grain size distribution of pedogenic fine-grained maghemite particles, showing that the dominant grain size is just above the threshold (about 22.5 nm for maghe5 of 16


3

G


mite) of the superparamagnetic and single domain. This grain size distribution is independent of the degree of pedogenesis. For loess profiles at the central CLP where pedogenic effects are strong, fluctuations in the absolute bulk c dominantly reflect changes in the concentration of superparamagnetic + single domain maghemite particles with a fixed grain size distribution. The lithogenic magnetic components with a c of 0.5 10 7 m3 kg 1, which include non-magnetic silicate matrix and the large coarse-grained magnetite, also contribute to the bulk contribute c, but have lower amplitude than that of the pedogenic components [Deng et al., 2005]. In contrast, since effects of pedogenesis are negligible at the northwest CLP [Chen et al., 1997; Liu et al., 2005b], variations in the bulk c during the glacial periods will dominantly reflect the lithogenic components, which are related more to the eolianderived material; whereas the c enhancement in the paleosol layers (S0 –S6) can be confidently used as an index for changes in amount of the paleoprecipitation and, in turn, the intensity of the Asian summer monsoon (Figure 3b).

3.2. Grain Size as a Proxy Indicator of East Asian Winter Monsoon Intensity [14] The Chinese loess reflects the deposition of dust transported to the CLP by the northwesterly winds [e.g., An et al., 1990, 1991a, 1991b; An, 2000]. A number of sedimentological and geochemical studies reveal that decrease in the loess grain size from northwest CLP to southeast CLP is consistent with the spatial pattern of the northwesterly winter monsoon winds [e.g., Ding et al., 1999b; Nugteren and Vandenberghe, 2004; Chen et al., 2006]. Therefore it is generally accepted that loess grain size mainly reflects the energy of winds that transport dust particles from modern deserts in northern and northwestern China [e.g., An et al., 1991a; Xiao et al., 1992, 1995; Liu and Ding, 1998]. Thus the grain size indices, such as the mean and median grain size, coarse fraction content and grain size ratio, have been widely employed as proxies for the intensity of the winter monsoon [e.g., Porter and An, 1995; Xiao et al., 1995; Vandenberghe et al., 1997; Liu and Ding, 1998; An, 2000]. [15] Recently, Ding et al. [1999b, 2005] suggested that the location of the desert margin has a potential influence on the grain size, especially on the sand content (>63 mm) of loess deposits. Due to gravity effect, sand particles can be transported only a limited distance. Therefore sand particles accumu-

10.1029/2006GC001287

lating in the central CLP probably originate primarily from a nearby source (e.g., the Mu Us desert). On the basis of the systematic decrease in sand content with increasing distance between the loess sections and the south margin of the Mu Us desert, a semiquantitative sand-content model was proposed to estimate the source-sink distance [Ding et al., 1999b]. On the basis of this model, the sand particle contents of 30% and 15% (by volume) within the loess unit (L1LL1) correspond to a source-sink distance of 100 and 200 km, respectively. [16] At the JY section, the sand content varies between about 6% and 10% in loess units and between about 0.5% and 3% in paleosols (Figure 3c). By applying the sand-content model of Ding et al. [1999b] to our site, the lower sand content during both glacial and interglacial periods suggests that the distance of the JY section to the desert margin is always >200 km over the last seven glacial/interglacial cycles. This is inconsistence with the modern 80 km distance from the JY section to the south margin of Tengger desert. [17] There are two possible mechanisms for such a contradiction. The first involves a different geomorphology between the western CLP and central CLP. The sand-content model is based on loess sections in the central CLP, in particular the Jinbian section which is adjacent to the modern Mu Us desert without mountain barriers [Ding et al., 2005]. The JY section in this study, however, is isolated from the Tenger desert by the eastern branch of the Qilan Mountains, namely Hasi Mountain with an average altitude of 2500 – 3000 m. This mountain is 1000–1500 m higher than the altitude of Tengger desert and therefore may block the southward expansion of the Tengger desert. If this is the case, the grain size variation at the JY section (Figure 3d) may be less sensitive to changes in the source-sink distance, mainly reflecting changes in the intensity of East Asian winter monsoon. The other possible mechanism is that the provenance of the loess deposits in particular the coarse fraction at the JY section does not originate from the Tengger desert, but from a nearby source (e.g., Yellow River beach sands). Further provenance study is necessary to clarify the latter mechanism, but this does not impact our final conclusion especially with regard to the glacial-interglacial contrast.

3.3. Calibration of the Orbital Age Model [18] The close relationships between the Earth’s orbital forcing and East Asian monsoon changes 6 of 16


3

G


have been documented by numerous studies on Chinese loess deposits [e.g., An et al., 1990; Ding et al., 1995; An, 2000]. Orbital tuning has been widely employed to reconstruct the chronology of the loess-paleosol sequences [Ding et al., 1994; Lu et al., 1999; Heslop et al., 2000; Ding et al., 2002; Sun et al., 2006a]. In addition to orbital tuning, magnetic susceptibility model [Kukla et al., 1988] and grain size model [Porter and An, 1995; Vandenberghe et al., 1997; Nugteren et al., 2004] have also been employed in construction of loess chronology. [19] In this study we employed the orbital tuning method proposed by Yu and Ding [1998] to reconstruct the chronology of the JY loess-paleosol sequence. The method involves 4 steps as follows: (1) selection of the tuning targets, (2) selection of the time controls and construction of an initial timescale, (3) orbital tuning by adjusting the age of each time control, and (4) fine adjustment of the age of each data point using a dynamical optimization method. The tuning targets were generated from the orbital solution of Laskar [1990], employing the SPECMAP-defined lag of 5-kyr and 8-kyr for obliquity and precession [Imbrie et al., 1984], respectively. [20] During the orbital tuning processes, selection of the time controls is critical for generating an accurate age model. An effective means of determining the tie points is to directly compare the monsoon proxies (e.g., c and grain size) to the marine oxygen isotope record and summer insolation at 65°N [e.g., Heslop et al., 2000]. On the basis of the visual correlation of monsoon proxies (i.e., c and MGS) to the stacked benthic d 18O [Lisiecki and Raymo, 2005] (Figure 3e), we correlate each paleosol layer to an interglacial marine isotope stage (MIS), for example, S0 to MIS 1, S1 to MIS 5, etc., which is identical to previous correlations between Chinese loess and marine oxygen isotope records [e.g., Kukla and An, 1989; Bloemendal et al., 1995; Liu et al., 1999; Heslop et al., 2000; Ding et al., 2002]. The top and base of each paleosol layer are used as tie points; the age of each point is derived from the stacked benthic d 18O curve (see solid lines in Figure 3). All together, sixteen initial time controls were selected; their depth and ages are listed in Table 1. [21] Meanwhile, we make a detailed correlation of monsoon proxies to the summer insolation at 65°N [Laskar, 1990] (Figure 3f), which is widely employed as a reliable target curve in the orbital tuning of terrestrial and marine sedimentary

10.1029/2006GC001287

sequences [e.g., Lourens et al., 1996; Heslop et al., 2000; Wehausen and Brumsack, 2002]. It is obvious that c variations in S1, S2 and S3 and MGS variations in L1, L2, L4, L5 and L6 can be correlated to precessional insolation cycles (IC). For example, three c cycles in S1 are correlated to IC 7, 9 and 11, respectively, while two MGS cycles within L2 are correlated to the IC 13 and IC 15. [22] A close check of the correlation between the oxygen isotope record and the summer insolation reveals that abrupt decreases in the benthic d18O record during glacial terminations correspond well with the insolation maxima, whereas rapid d18O increases during the interglacial/glacial transitions are closely coupled with the insolation minima. More recently, the absolute-dated speleothem d18O records [Wang et al., 2001; Yuan et al., 2004] reveal that abrupt monsoon shifts during the last climatic cycle match well with the most rapid changes in the summer insolation (i.e., insolation maxima and minima) [Ji et al., 2006]. Therefore, by assigning rapid c increases or MGS decreases to the insolation maxima and rapid c decreases or MGS increases to the insolation minima, respectively, we choose twenty-four new time controls.

3.4. Evaluating the Orbital Age Model [23] Comparison of the filtered components of MGS with orbital parameters is a reliable means for assessing the orbital age model; although temporal fits are expected as a result of the tuning process, similar amplitude modulation is not. The correlation coefficient between the filtered components of MGS and orbital curves is 0.93 at the obliquity band and 0.88 at the precession band, respectively, which indicates good phase and amplitude matches between orbital targets and monsoon variance. The filtered 41-kyr and 21-kyr components of the MGS match well with the obliquity and precession target curves both in amplitude and in phase (Figures 4a and 4b), with a minor amplitude mismatch at the precession band during the time interval 470–620 ka. Since this interval (corresponding to S5) is the prolonged and strongest interglacial in the CLP [Han et al., 1998], the MGS record only exhibits three longer cycles, whereas the precession cycles have been damped to a larger degree (Figure 4c). Such an amplitude mismatch probably indicates that the winter monsoon intensity is less sensitive to the precessional forcing during this interval than during other periods. The exact mechanism for this mismatch requires further study. 7 of 16


3

G


10.1029/2006GC001287

Figure 4. Comparison of the filtered (a) 41-kyr and (b) 21-kyr components of JY MGS (solid lines) with the orbital obliquity and precession curves (dashed lines) [Laskar, 1990], (c) temporal changes in the MGS, and (d) estimated sedimentation rate of the JY section. Both the filtered signals and orbital records are normalized. Band-pass filters with central frequencies of 0.02439 and 0.04762 kyr 1 and bandwidths of 0.004 and 0.015 kyr 1 are used to isolate the 41-kyr and 21-kyr components of the grain size time series. Gray bar shows the amplitude mismatch at the precession band during the interval 470– 620 ka.

[24] Sedimentation rate (Figure 4d), another parameter useful in assessing the age model, is higher in loess layers than paleosol units, consistent with those reported in the central CLP [Sun et al., 2006a]. Sedimentation rate is almost consistent within interglacial times (6–10 cm/kyr), but varies significantly during glacial periods with values of 20–100 cm/kyr. During the last glacial, sedimentation rate varies from 30 cm/kyr to 100 cm/kyr, yielding a sampling resolution ranging from 330-yr to 100-yr. [25] Comparison of our age model to previous orbital age models can provide additional insights into the chronology of the loess-paleosol sequences (Figure 5). As indicated by the estimated ages for the top and base of each paleosol in Table 1, discrepancies among different age models are evident over the past 720 kyr. Two evident discrepancies between our age model and the age model generated by Lu et al. [1999] are the ages for the bases of S4 and S5 (Figure 5a), likely resulting from the use of inconsistent geomagnetic reversal boundaries and different initial correlations to tuning targets.

[26] The differences between the age model generated by Heslop et al. [2000] and our age model are evident in the top of all paleosol layers (S1 –S6) and the bottom of S4 (Figure 5b). Ages of the top of those paleosol layers are 6 to 30 kyr younger in our age model. Compared with the age model published by Ding et al. [2002], age for the top of S3 is 27-kyr younger, while age for the base of S4 is 12-kyr older in this study (Figure 5c). The age model in this study is consistent with a recently published astronomical timescale (Figure 5d), in which a number of adjustments have been made in order to generate an improved loess timescale for better evaluation of the EAM evolution [Sun et al., 2006a].

4. Results and Discussion 4.1. Temporal Variations in the c and MGS [27] Temporal variations in the c and MGS of the JY section and their comparison with marine oxygen isotope record and solar insolation are shown in Figure 6. Generally, loess units are characterized by lower c and coarser MGS, where8 of 16


3

G


10.1029/2006GC001287

Figure 5. Comparison of our age model to recently published age models. Solid lines denote the MGS record of JY section. Dashed curves from top to bottom are as follows: (a) Luochuan c on the age model of Lu et al. [1999], (b) Luochuan c on the age model of Heslop et al. [2000]; (c) stacked median grain size on the age model of Ding et al. [2002]; and (d) stacked mean grain size of quartz on the age model of Sun et al. [2006a]. Gray bars indicate differences between our age model and published age models.

as paleosol layers have higher c and finer MGS values. Within the paleosol complexes (e.g., S1, S2 and S3), alternations between the paleosol and interbedded loess also exhibit distinct c and MGS fluctuations. Temporal variations of the two proxies are apparently different in two aspects: (1) the c variation is more significant during interglacials than glacials whereas the MGS value is less variable during interglacial periods than glacial periods; (2) c is dramatically enhanced at paleosol S4 whereas the MGS displays relatively uniform amplitude variations at glacial-interglacial scales over the last 720 kyr. [28] Since the pedogenic processes within the loess layers at the site is insignificant, the c change of these loess units may be attributable mainly to change in the lithogenic magnetic components. A slight up-section c increase from L6 to L1 is consistent with changes in lithogenic magnetic components of two loess-paleosol sequences in the central CLP [Deng et al., 2005, 2006], reflecting an increased dust input over the CLP, which presumably results from a gradual drying of the Asian interior [Sun and An, 2005; Deng et al.,

2006]. As discussed above, the high c values of the paleosol layers provide more insights into the relative amount of local paleoprecipitation. The c values of S5 and S6 are significantly lower than in the overlying paleosols (i.e., S0 –S4), implying that the amount of the local precipitation during the interglacials has increased significantly since 425 ka. The c values of three paleosol complexes (S 1 , S 2 and S 3 ) exhibit three distinct peaks corresponding well with insolation maxima. Moreover, the c values are higher in the S1SS3 compared with the S1SS1 and S1SS2, suggesting a heavier monsoon precipitation in the early/full interglacial (MIS 5e) than the MIS 5a and 5c. [29] Unlike the c variation, MGS values are less variable during interglacial periods, but show evident precession cycles in most glacial times. For example, two precessional MGS cycles can be readily identified within the loess units L1, L2, L4 and L6. These MGS peaks (finer MGS) roughly correspond to the insolation maxima, however similar precessional cycles are not evident in the stack benthic d 18O record [Lisiecki and Raymo, 2005] (Figure 6). The distinct precessional varia9 of 16


3

G


10.1029/2006GC001287

Figure 6. (a) The stacked benthic d 18O record [Lisiecki and Raymo, 2005]; age plots of (b) MGS and (c) c over the past 720 ka from the JY section; and (d) the summer insolation at 65°N [Laskar, 1990]. Gray bars denote the precessional cycles captured by the monsoon indices, which roughly correspond to the ice volume minimum and insolation maximum. Dashed line draws attention to the sudden c increase around the MIS 12/11 transition. Bold solid lines indicate the relatively weak winter monsoon during two moderate glaciations (MIS 8 and MIS 14).

tions in monsoon proxies relative to other proxy indicators suggests that the EAM is more sensitive to the precessional forcing than climate systems in high-latitude regions, where the damped precessional signals may be due to the 180° out-of-phase nature of southern- and northern-hemisphere heating/cooling at the precessional band [Clemens, 1999]. [30] Moreover, the lower MGS values of loess units L3 and S5LL1 relative to other loess layers suggest that the winter monsoon intensity is not so strong during the two periods, which are consistent with moderate glaciations revealed by benthic oxygen isotope records [Lisiecki and Raymo, 2005] and weakened changes in summer insolation during MIS 8 and MIS 14 [Laskar, 1990]. An abnormal coarsening of the MGS is observed within the paleosol S2, which is about 10 mm coarser than the MGS value of loess unit L3. This MGS coarsening is correlated to the ice volume maximum during MIS 7d and insolation minimum around 225 ka, suggesting a dynamical linage between the winter monsoon variation and the solar/glacial forcing. [31] In summary, the difference between c and MGS is due to different sedimentary processes

associated with the two proxies. Grain size reflects surface processes, which are largely related to wind intensity; whereas the c record contains both eolianinput and post-depositional signals associated with the amount of paleorainfall. As discussed above, the c record is relatively muted and stable within the loess units compared with paleosol layers, which presumably indicates that the summer monsoon did not affect this region during the glacials [Hao and Guo, 2005]. Whereas higher and more variable c values during interglacials reveal that the intensified summer monsoon regime has penetrated into this region. Grain size of the loess deposits in this region, however, is sensitive to changes in the wind intensity because of negligible pedogenic alteration and high accumulation rate. Thus MGS can record distinct fluctuations at both orbital and millennial timescales.

4.2. Spatial Differences in the c and MGS Variations [32] Comparing the temporal patterns of the c and MGS variations in the central and northwestern CLP can provide valuable insights into the spatial variability of the East Asian summer and winter monsoons over glacial-interglacial timescales 10 of 16


3

G


10.1029/2006GC001287

Figure 7. Comparison of (a) c and (b) MGS variations of the LT and ZJC section in the central CLP [Sun et al., 2006a] and the JY section in the northwest CLP. MGSQ indicates that the mean grain size data of the LT and ZJC sections are generated from the monomineral quartz. Gray bars denote major paleosol units. The dashed lines indicate the amplitude of the winter and summer monsoon intensities during interglacials.

(Figure 7). It is obvious that most orbital-scale features captured by our MGS (winter monsoon proxy) are similar to those reported from the Lingtai (LT) and Zhaojiachuan (ZJC) sections in the central CLP [e.g., Sun et al., 2006a]. However, the c record of the JY section is different from those in the central CLP (Figure 7a), having three distinct precessional cycles within three paleosol complexes (S1, S2 and S3). The lack of precessional-scale summer monsoon variations in the central CLP may be attributed to the relatively low sedimentation rate and the combination of sub-paleosol units during these interglacials [Feng et al., 2004]. [33] Different amplitude of the c and MGS variation among the three sites provides insight into the spatial gradient and glacial-interglacial contrast between the summer and winter monsoons (Figure 8). The c difference between the central and northwestern CLP is larger in the interglacials relative to the glacials, implying that the southeastnorthwest gradient of the summer monsoon intensity/precipitation is larger in interglacials compared

with glacials. This finding is consistent with previous studies on the spatial pattern of the c variation of the loess-paleosol sequences in the central CLP [e.g., Hao and Guo, 2005; Bloemendal and Liu, 2005]. The c difference between glacials and interglacials is larger in the central CLP than the northwest CLP, suggesting that glacialinterglacial contrast in the summer monsoon intensity/precipitation is stronger in the southern site (i.e., LT section) than in the northern site (i.e., JY section). [34] Similarly, the MGS difference can be used to determine the relative intensity and gradient of the East Asian winter monsoon across the CLP. Although the mean grains size of the LT and ZJC section are generated from mono-mineral quartz, the glacial-interglacial variation of the mean grain size of quartz (MGSQ) is similar to the MGS of bulk samples [Sun et al., 2006b]. The glacialinterglacial difference in the MGS variation is higher at the JY section (30 mm) than the LT (14 mm) and ZJC (18 mm) sections. Moreover, the average MGS value shows a twofold increase 11 of 16


3

G


Figure 8. c versus MGS plot showing the relative amplitude of glacial-interglacial changes in the summer and winter monsoon intensities at the JY, ZJC, and LT. Larger c and MGS values are associated with stronger summer and winter monsoons, respectively.

from the central CLP to the northwest CLP during glacials (14 mm) relative to interglacials (7 mm). These spatial patterns are consistent with earlier investigations on the spatial variations of grain size for the last two glacial cycles [Yang and Ding, 2004; Nugteren and Vandenberghe, 2004], indicating that (1) glacial-interglacial contrast in the winter monsoon fluctuations is higher in the northwest CLP than in the central CLP and (2) southeast-northwest gradient of the winter monsoon variation is larger in glacials compared with interglacials.

4.3. Sudden c Increase During the Mid-Brunhes Epoch [35] Inspection of the c records from the central CLP and northwestern CLP suggests a sudden increase in the summer monsoon intensity during the mid-Brunhes epoch (Figure 7a), but with different timing at these two regions. c records from the central CLP reveal a sudden strengthening of the interglacial summer monsoon intensity around MIS 13, corresponding to the development of the paleosol S5SS1 [Kukla and An, 1989; An et al., 1990; Xiao and An, 1999; Hao and Guo, 2005; Sun et al., 2006a]. The c records of the Lanzhou and Huining loess sections [F. Chen et al., 1991; Hao and Guo, 2005], as well as our JY c record, however, suggest a remarkable increase of the interglacial summer monsoon intensity since the paleosol S4 (corresponding to MIS 11) in the northwest CLP.

10.1029/2006GC001287

[36] Since variation in the c is mainly dependent on the monsoon precipitation [e.g., Zhou et al., 1990; An et al., 1991b; Maher and Thompson, 1995; Liu et al., 1995], the c difference among different loess sections should reflect the relative amount of local precipitation [e.g., Liu et al., 2005b; Hao and Guo, 2005; Bloemendal and Liu, 2005]. The asynchronous shift of the c records between the central CLP and the northwest CLP indicates a spatial difference in the local amount of monsoon precipitation. Figure 9 illustrates a schematic map showing the spatial evolution of the summer monsoon during the mid-Brunhes epoch. Prior to MIS 13, the monsoon precipitation was relatively weak and had a limited impact on the whole CLP (Figure 9a). During MIS 13 (S5SS1), the sharply increased monsoon precipitation strongly affected the central CLP [Balsam et al., 2004]. However, the relatively lower c value at the JY section suggests that intensified precipitation did not penetrate into the northwestern CLP (Figure 9b). Since MIS 11 (S4), the sudden c increase at the JY section indicates that the strong influence of summer monsoon precipitation extended northward spanning the whole CLP during the subsequent interglacial intervals (Figure 9c). [37] As for the origin of the mid-Brunhes shift in summer monsoon intensity, it has been speculated that tectonic uplift of the Tibetan Plateau may have played a part in changing the East Asian monsoon circulation around this time interval [An et al., 1990; Xiao and An, 1999]. On the basis of climate-model simulations, continued uplift of the Tibetan Plateau during the Quaternary period should enhance both summer and winter monsoons in the region of the Loess Plateau [An et al., 2001]. However, the winter monsoon intensity (as revealed by the MGS variations) does not indicate a corresponding change either during MIS 13 in the central CLP or during MIS 11 in the northwestern CLP (Figure 7b). Evidently, tectonic uplift forcing is not sufficient to explain the monsoon shift during the mid-Brunhes epoch. [38] Another possible mechanism triggering the mid-Brunhes summer monsoon shift may reside in the tropic region, because the tropical ocean may amplify the eccentricity forcing through carbon reservoir disturbances with consequent impact on global climate systems [Wang et al., 2003]. Carbon isotope records (d 13C) of ODP site 1143 in the southern South China Sea document a d 13C maximum during MIS 13, which is consistent with heavy precipitation in monsoon-effected regions 12 of 16


3

G


10.1029/2006GC001287

Figure 9. A schematic map showing the spatial difference in the local monsoon precipitation over the CLP during (a) MIS 17, (b) MIS 13, and (c) MIS 11, respectively. The gradient gray areas denote that the affecting areas are apparently different during the three interglacial stages, while the different lightness of the three gray areas suggests that the relative intensity of the monsoon precipitation increases significantly from MIS 17 to MIS 11.

during this time interval [see Wang et al., 2003, Figure 3]. [39] Clearly, climate shift during the mid-Brunhes epoch is a global phenomenon, although the timing is inconsistent among land (Chinese loess), lake [Prokopenko et al., 2002], ocean [Droxler and Farrell, 2000; Wang et al., 2003] and ice core records [EPICA Community Members, 2004]. Our monsoon indices generated from the loess deposits add to the global inconsistencies in timing and suggest even regional inconsistencies relative to the mid-Brunhes climate shift.

5. Conclusions [40] The c and MGS records generated from a loess sequence in the northwestern CLP exhibit distinct glacial-interglacial variations over the past seven climatic cycles. Generally, the winter and summer monsoon intensities vary in a compensation mode, characterized by intensified winter monsoons during glacials and enhanced summer monsoons during interglacials. During two moderate glaciations (MIS 8 and MIS 14), however, the MGS of loess units L3 and S5LL1 suggests a relatively weak winter monsoon compared with previous and subsequent glaciations. Moreover, spatial differences in amplitude of the c and MGS variations suggest the glacial-interglacial contrast in the northwest CLP is stronger for the winter monsoon but weaker for the summer monsoon than in the central CLP. [41] Although the monsoon indices (c and MGS) are better resolved at the JY section, orbital-scale features captured by the MGS are similar to those reported from the central CLP, indicating that MGS is a reliable proxy and can be used for highresolution inter-profile correlations. The JY c is

different from the c records in the central CLP in two aspects. First, due to the high sedimentation rate and weak pedogenic alterations in the northwest CLP, the JY c record resolves precessional cycles during MIS 5, 7 and 9. Second, although the c record from both the central and northwestern CLP reveal a marked strengthening of the c during the mid-Brunhes epoch, the timing of such a summer monsoon strengthening is different between the central CLP and the northwest CLP.

Notation CLP EAM c MGS MGSQ MIS IC JY LT ZJC

Chinese Loess Plateau. East Asian monsoon. magnetic susceptibility. mean grain size of bulk samples. mean grain size of quartz. marine isotope stage. insolation cycle. Jingyuan. Lingtai. Zhaojiachuan.

Acknowledgments [42] We thank L. Zhao, T. Wu, and H. Wang for assistance in field sampling and lab analysis. D. W. Oppo and R. X. Huang are gratefully acknowledged for their comments on an early version of the manuscript. We are especially grateful to D. D. Rousseau, G. Kukla, and A. Berger for their thoughtful comments, which significantly improved the manuscript. This work was supported by the Natural Science Foundation of China (40331001), the National Basic Research Program of China (2004CB720200), and a fellowship to Y. B. Sun from the Japan Society for the Promotion of Science (P04293). Q. S. Liu was supported by a European Commission Marie-Curie

13 of 16


3

G


Fellowship (proposal 7555). This is a contribution to IGCP476 project.

References An, Z. S. (2000), The history and variability of the East Asian paleomonsoon climate, Quat. Sci. Rev., 19, 171 – 187. An, Z. S., T. S. Liu, Y. C. Lu, S. C. Porter, G. Kukla, X. H. Wu, and Y. M. Hua (1990), The long-term paleomonsoon variation recorded by the loess-paleosol sequence in central China, Quat. Int., 7/8, 91 – 95. An, Z. S., G. Kukla, S. C. Porter, and J. L. Xiao (1991a), Late Quaternary dust flow on the Chinese Loess Plateau, Catena, 18, 125 – 132. An, Z. S., G. Kukla, S. C. Porter, and J. L. Xiao (1991b), Magnetic susceptibility evidence of monsoon variation on the loess plateau of central China during the last 130,000 years, Quat. Res., 36, 29 – 36. An, Z. S., J. E. Kutzbach, W. L. Prell, and S. C. Porter (2001), Evolution of Asian monsoons and phased uplift of the Himalaya-Tibetan plateau since late Miocene times, Nature, 411, 62 – 66. Balsam, W., J. Chen, and J. F. Ji (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, 335 – 348. Bloemendal, J., and X. M. Liu (2005), Rock magnetism and geochemistry of two Plio-Pleistocene Chinese loesspalaeosol sequences: Implications for quantitative palaeoprecipitation reconstruction, Palaeogeogr. Palaeoclimatol. Palaeoecol., 226, 149 – 166. Bloemendal, J., X. M. Liu, and T. C. Rolph (1995), Correlation of the magnetic susceptibility stratigraphy of Chinese loess and the marine oxygen isotope record: Chronological and palaeoclimatic implications, Earth Planet. Sci. Lett., 131, 371 – 380. Chen, F., J. Li, and W. Zhang (1991), Lanzhou loess profile and its comparison with deep sea sediment and Antarctic ice core, GeoJournal, 24, 201 – 209. Chen, F. H., J. Bloemendal, J. M. Wang, J. J. Li, and F. Oldfield (1997), High-resolution multi-proxy climate records from Chinese loess: Evidence for rapid climatic changes over the last 75 Kyr B. P., Palaeogeogr. Palaeoclimatol. Palaeoecol., 130, 323 – 335. Chen, F. H., J. Bloemendal, Z. D. Feng, J. M. Wang, E. Parker, and Z. T. Guo (1999), East Asian monsoon variations during Oxygen Isotope Stage 5: Evidence from the northwestern margin of the Chinese Loess Plateau, Quat. Sci. Rev., 18, 1127 – 1135. Chen, J., Y. Chen, L. W. Liu, J. F. Ji, W. Balsam, Y. B. Sun, and H. Y. Lu (2006), Zr/Rb ratio in the Chinese loess sequences and its implication for changes in the East Asian winter monsoon strength, Geochim. Cosmochim. Acta, 70, 1471 – 1482. Chen, L. X., Q. G. Zhu, H. B. Luo, J. H. He, M. Dong, and Z. Q. Feng (1991), The East Asian Monsoon (in Chinese), pp. 304 – 346, Meteorol. Press, Beijing. Clemens, S. C. (1999), An astronomical tuning strategy for Plio-Pleistocene sections: Implications for global-scale correlation and phase relationships, Philos. Trans. R. Soc. London, Ser. A, 357, 1949 – 1974. Deng, C., N. J. Vidic, K. L. Verosub, M. J. Singer, Q. Liu, J. Shaw, and R. Zhu (2005), Mineral magnetic variation of the Jiaodao Chinese loess/paleosol sequence and its bearing

10.1029/2006GC001287

on long-term climatic variability, J. Geophys. Res., 110, B03103, doi:10.1029/2004JB003451. Deng, C. L., J. Shaw, Q. S. Liu, Y. X. Pan, and R. X. Zhu (2006), Mineral magnetic variation of the Jingbian loess/ paleosol sequence in the Northern Loess Plateau of China: Implications for Quaternary development of Asian aridification and cooling, Earth Planet. Sci. Lett., 241, 248 – 259. Ding, Z. L., Z. W. Yu, N. W. Rutter, and T. S. Liu (1994), Towards an orbital time scale for Chinese loess deposits, Quat. Sci. Rev., 13, 39 – 70. Ding, Z. L., T. S. Liu, R. W. Rutter, Z. W. Yu, Z. T. Guo, and R. X. Zhu (1995), Ice-volume forcing of East Asian winter monsoon variations in the past 800,000 years, Quat. Res., 44, 149 – 159. Ding, Z. L., J. Z. Ren, S. L. Yang, and T. S. Liu (1999a), Climate instability during the penultimate glaciation: Evidence from two high-resolution loess records, China, J. Geophys. Res., 104, 20,123 – 20,132. Ding, Z. L., J. M. Sun, N. W. Rutter, D. Rokosh, and T. S. Liu (1999b), Changes in the sand content of loess deposits along a north to south transect of the Chinese Loess Plateau and the implications for desert variations, Quat. Res., 52, 56 – 62. Ding, Z. L., E. Derbyshire, S. L. Yang, Z. W. Yu, S. F. Xiong, and T. S. Liu (2002), Stacked 2.6-Ma grain size record from the Chinese loess based on five sections and correlation with the deep-sea d 18O record, Paleoceanography, 17(3), 1033, doi:10.1029/2001PA000725. Ding, Z. L., E. Derbyshire, S. L. Yang, J. M. Sun, and T. S. Liu (2005), Stepwise expansion of desert environment across northern China in the past 3.5 Ma and implications for monsoon evolution, Earth Planet. Sci. Lett., 237, 45 – 55. Droxler, A. W., and J. W. Farrell (2000), Marine isotope stage 11 (MIS 11), New insights for a warm future, Global Planet. Change, 24, 1 – 5. Emile-Geay, J., M. A. Cane, N. Naik, R. Seager, A. C. Clement, and A. van Geen (2003), Warren revisited: Atmospheric freshwater fluxes and ‘‘Why is no deep water formed in the North Pacific’’, J. Geophys. Res., 108(C6), 3178, doi:10.1029/2001JC001058. EPICA Community Members (2004), Eight glacial cycles from an Antarctic ice core, Nature, 429, 623 – 628. Fang, X. M., Y. Ono, H. Fukusawa, B. T. Pan, J. J. Li, D. H. Guan, K. Oi, S. Tsukamoto, and T. Torii (1999), Asian summer monsoon instability during the past 60,000 years: Magnetic susceptibility and pedogenic evidence from the western Chinese Loess Plateau, Earth Planet. Sci. Lett., 168, 219 – 232. Feng, Z. D., H. B. Wang, C. Olson, G. A. Pope, F. H. Chen, J. W. Zhang, and C. B. An (2004), Chronological discord between the last interglacial paleosol (S1) and its parent material in the Chinese Loess Plateau, Quat. Int., 117, 17 – 26. Guo, Z. T., T. S. Liu, J. Guiot, N. Q. Wu, H. Y. Lu, J. M. Han, J. Q. Liu, and Z. Y. Gu (1996), High frequency pulses of East Asia monsoon climate in the last two glaciations: Link with the North Atlantic, Clim. Dyn., 12, 701 – 709. Guo, Z. T., W. F. Ruddiman, Q. Z. Hao, H. B. Wu, Y. S. Qiao, R. X. Zhu, S. Z. Peng, J. J. Wei, B. Y. Yuan, and T. S. Liu (2002), Onset of Asian desertification by 22 Myr ago inferred from loess deposits in China, Nature, 416, 159 – 163. Han, J. M., H. Y. Lu, N. Q. Wu, and Z. T. Guo (1996), The magnetic susceptibility of modern soils in China and its paleoclimate reconstruction, Stud. Geophys. Geod., 40, 262 – 275. Han, J. T., W. S. Fyfe, and F. J. Longstaffe (1998), Climatic implications of the S5 paleosol complex on the southernmost Chinese Loess Plateau, Quat. Res., 50, 21 – 33. 14 of 16


3

G


Hao, Q., and Z. Guo (2005), Spatial variations of magnetic susceptibility of Chinese loess for the last 600 kyr: Implications for monsoon evolution, J. Geophys. Res., 110, B12101, doi:10.1029/2005JB003765. Heslop, D., J. Shaw, J. Bloemendal, F. H. Chen, J. Wang, and E. Parker (1999), Sub-millennial scale variations in East Asian monsoon systems recorded by dust deposits from the north-western Chinese Loess Plateau, Phys. Chem. Earth, Part A, 24, 785 – 792. Heslop, D., C. G. Langereis, and M. J. Dekkers (2000), A new astronomical timescale for the loess deposits of Northern China, Earth Planet. Sci. Lett., 184, 125 – 139. Imbrie, J., J. D. Hays, D. B. Martinson, A. McIntyre, A. C. Mix, J. J. Morlet, N. G. Pisias, W. L. Prell, and N. J. Shackleton (1984), The orbital theory of Pleistocene climate: Support from a revised chronology of the marine d 18O record, in Milankovitch and Climate, Part 1, edited by A. Berger et al., pp. 269 – 305, Springer, New York. Irino, T., and R. Tada (2003), High-resolution reconstruction of variation in aeolian dust (Kosa) deposition at ODP site 797, the Japan Sea, during the last 200 ka, Global Planet. Change, 35, 143 – 156. Ji, J., W. Balsam, X. Chen, J. Chen, Y. Chen, and H. Wang (2006), Rate of solar insolation change and the glacial/interglacial transition, Geophys. Res. Lett., 33, L04706, doi:10.1029/2005GL025401. Kukla, G., and Z. S. An (1989), Loess stratigraphy in central China, Palaeogeogr. Palaeoclimatol. Palaeoecol., 72, 203 – 225. Kukla, G., F. Heller, X. M. Liu, T. C. Xu, T. S. Liu, and Z. S. An (1988), Pleistocene climates in China dated by magnetic susceptibility, Geology, 16, 811 – 814. Laskar, J. (1990), The chaotic motion of the solar system: A numerical estimate of the size of the chaotic zones, Icarus, 88, 266 – 291. Lei, X. Y. (1995), Formation age and microfabric characteristics of loess in Caoxian, Jingyuan, Gansu, China (in Chinese with English abstract), Acta Geogr. Sin., 50, 521 – 533. Li, J. J., and J. C. Kang (1989), Quaternary glaciations, physiographic stages and loess records in China (in Chinese with English abstract), Quat. Sci., 3, 269 – 278. Lisiecki, L. E., and M. E. Raymo (2005), A PliocenePleistocene stack of 57 globally distributed benthic d 18O records, Paleoceanography, 20, PA1003, doi:10.1029/ 2004PA001071. Liu, Q., M. J. Jackson, S. K. Banerjee, B. A. Maher, C. Deng, Y. Pan, and R. Zhu (2004), Mechanism of the magnetic susceptibility enhancements of the Chinese loess, J. Geophys. Res., 109, B12107, doi:10.1029/2004JB003249. Liu, Q., J. Torrent, B. A. Maher, Y. Yu, C. Deng, R. Zhu, and X. Zhao (2005a), Quantifying grain size distribution of pedogenic magnetic particles in Chinese loess and its significance for pedogenesis, J. Geophys. Res., 110, B11102, doi:10.1029/2005JB003726. Liu, Q. S., S. K. Banerjee, M. J. Jackson, C. L. Deng, Y. X. Pan, and R. X. Zhu (2005b), Inter-profile correlation of the Chinese loess/paleosol sequences during Marine Oxygen isotope Stage 5 and indications of pedogenesis, Quat. Sci. Rev., 24, 195 – 210. Liu, T. S., and Z. L. Ding (1998), Chinese loess and the paleomonsoon, Annu. Rev. Earth Planet. Sci., 26, 111 – 145. Liu, T. S., Z. L. Ding, and N. Rutter (1999), Comparison of Milankovitch periods between continental loess and deep sea records over the last 2.5 Ma, Quat. Sci. Rev., 18, 1205 – 1212. Liu, X. M., T. Rolph, J. Bloemendal, J. Shaw, and T. S. Liu (1995), Quantitative estimates of palaeoprecipitation at Xi-

10.1029/2006GC001287

feng, in the loess plateau of China, Palaeogeogr. Palaeoclimatol. Palaeoecol., 113, 243 – 248. Lourens, L. J., A. Antonarakou, F. J. Hilgen, A. A. M. van Hoof, C. Vergnaud-Grazzini, and W. J. Zachariasse (1996), Evaluation of the Plio-Pleistocene astronomical timescale, Paleoceanography, 11, 391 – 413. Lu, H. Y., X. D. Liu, F. Q. Zhang, Z. S. An, and J. Dodson (1999), Astronomical calibration of loess-paleosol deposits at Luochuan, Central Chinese loess plateau, Palaeogeogr. Palaeoclimatol. Palaeoecol., 154, 237 – 246. Maher, B. A., and R. Thompson (1992), Paleoclimatic significance of the mineral magnetic record of the Chinese loess and paleosol, Quat. Res., 37, 155 – 170. Maher, B. A., and R. Thompson (1995), Paleorainfall reconstructions from pedogenic magnetic susceptibility variations in the Chinese loess and paleosols, Quat. Res., 44, 383 – 391. Nugteren, G., and J. Vandenberghe (2004), Spatial climatic variability on the Central Loess Plateau (China) as recorded by grain size for the last 250 kyr, Global Planet. Change, 41, 185 – 206. Nugteren, G., J. Vandenberghe, C. van Huisseden, and Z. S. An (2004), A Quaternary climate record based on grain size analysis from the Luochuan loess section on the Central Loess Plateau, China, Global Planet. Change, 41, 167 – 183. Parker, E. J., and J. Bloemandal (2005), Aeolian process and pedogenesis under the influence of the East Asian monsoon: A statistical approach to particle-size distribution variability, Sediment. Geol., 181, 195 – 206. Porter, S. C., and Z. S. An (1995), Correlation between climate events in the North Atlantic and China during the last Glaciation, Nature, 375, 305 – 308. Prokopenko, A. A., D. F. Williams, M. I. Kuzmin, E. B. Karabanov, G. K. Khursevich, and J. A. Peck (2002), Muted climate variations in continental Siberia during the midPleistocene epoch, Nature, 418, 65 – 68. Rea, D. K. (1994), The paleoclimatic record provided by eolian deposition in the deep-sea: The geologic history of wind, Rev. Geophys., 32, 159 – 195. Sun, Y., and Z. An (2005), Late Pliocene-Pleistocene changes in mass accumulation rates of eolian deposits on the central Chinese Loess Plateau, J. Geophys. Res., 110, D23101, doi:10.1029/2005JD006064. Sun, Y. B., S. C. Clemens, Z. S. An, and Z. W. Yu (2006a), Astronomical timescale and palaeoclimatic implication of stacked 3.6-Myr monsoon records from the Chinese Loess Plateau, Quat. Sci. Rev., 25, 33 – 48. Sun, Y. B., H. Y. Lu, and Z. S. An (2006b), Grain size of loess, palaeosol and Red Clay deposits on the Chinese Loess Plateau: Significance for understanding pedogenic alteration and palaeomonsoon evolution, Palaeogeogr. Palaeoclimatol. Palaeoecol., 241, 129 – 138. Vandenberghe, J., Z. S. An, G. Nugteren, H. Y. Lu, and C. van Huisseden (1997), New absolute time scale for the Quaternary climate in the Chinese loess region by grain-size analysis, Geology, 25, 35 – 38. Wang, P. X., J. Tian, X. R. Cheng, C. L. Liu, and J. Xu (2003), Carbon reservoir changes preceded major ice-sheet expansion at the mid-Brunhes event, Geology, 31, 239 – 242. Wang, Y. J., H. Cheng, R. L. Edwards, Z. S. An, J. Y. Wu, C. C. Shen, and J. A. Dorale (2001), A high-resolution absolute-dated late Pleistocene monsoon record from Hulu Cave, China, Science, 294, 2345 – 2348. Webster, P. J., V. O. Magan˜a, T. N. Palmer, J. Shukla, R. A. Tomas, M. Tania, and T. Yasunari (1998), Monsoons: Processes, predictability, and the prospects for prediction, J. Geophys. Res., 103, 14,451 – 14,510. 15 of 16


3

G


Wehausen, R., and H. J. Brumsack (2002), Astronomical forcing of the East Asian monsoon mirrored by the composition of Pliocene South China Sea sediments, Earth Planet. Sci. Lett., 201, 621 – 636. Wu, G. J., B. T. Pan, Q. Y. Guan, and D. S. Xiao (2005), Terminations and their correlation with solar insolation in the Northern Hemisphere: A record from a loess section in Northwest China, Palaeogeogr. Palaeoclimatol. Palaeoecol., 216, 267 – 277. Xiao, J. L., and Z. S. An (1999), Three large shifts in East Asian monsoon circulation indicated by loess-paleosol sequences in China and late Cenozoic deposits in Japan, Palaeogeogr. Palaeoclimatol. Palaeoecol., 154, 179 – 189. Xiao, J. L., H. B. Zheng, and H. Zhao (1992), Variation of winter monsoon intensity on the Loess Plateau, central China during the last 130,000 years: Evidences from grain size distribution, Quat. Res., 31, 13 – 19. Xiao, J. L., S. C. Porter, Z. S. An, H. Kumai, and S. Yoshikawa (1995), Grain size of quartz as an indicator of winter mon-

10.1029/2006GC001287

soon strength on the Chinese Loess Plateau of central China during the last 130,000 yr, Quat. Res., 43, 22 – 29. Yang, S. L., and Z. L. Ding (2004), Comparison of particle size characteristics of the Tertiary red clay and Pleistocene loess in the Chinese Loess Plateau: Implications for origin and sources of the red clay, Sedimentology, 51, 77 – 93. Yu, Z. W., and Z. L. Ding (1998), An automatic orbital tuning method for the Paleoclimate records, Geophys. Res. Lett., 25, 4525 – 4528. Yuan, D. X., et al. (2004), Timing, duration, and transitions of the last interglacial monsoon, Science, 304, 575 – 578. Yue, L. P., X. Y. Lei, and H. J. Qu (1991), A magnetostratigraphic study on the Jingyuan loess section, Gansu, China (in Chinese with English abstract), Quat. Sci., 4, 349 – 353. Zhang, J. C., and Z. G. Lin (1992), Climate of China, 376 pp., John Wiley, Hoboken, N. J. Zhou, L. P., F. Oldfield, A. G. Wintle, S. G. Robinson, and J. T. Wang (1990), Partly pedogenic origin of magnetic variations in Chinese loess, Nature, 346, 737 – 739.

16 of 16