High-resolution magnetic and palynological records of the last deglaciation and Holocene from Lake Xiarinur in the Hunshandake Sandy Land,Inner Mongolia

Abstract

High-resolution mineral magnetic and pollen records from overlapping piston cores from Lake Xiarinur (42°37′N, 115°28′E) document detailed changes in environment and vegetation since the last deglaciation in the Hunshandake Sandy Land, Inner Mongolia. The formation of Lake Xiarinur commenced during the Bølling–Allerød warming, as reflected by an abrupt shift in sedimentary facies from eolian sands to lacustrine sediments at a core depth of 3.5 m (~14.1 cal. kyr BP). The pollen records demonstrate that desert vegetation with sparse herbs occurred before 14.1 cal. kyr BP, which was succeeded by meadow grassland vegetation from ~14.1 –13.3 cal. kyr BP (the Bølling -Allerød). A dramatic decrease in pollen concentration occurred between ~13.3 and 11.7 cal. kyr BP corresponding to the Younger Dryas. The remarkable increases in pollen and Pediastrum concentrations at ~11.7 cal. kyr BP suggest that the environment began to ameliorate, and the most humid conditions prevailed until ~8.5 cal. kyr BP. The maxima of magnetic concentration–dependent and magnetic-grain-size-dependent parameters, together with high pollen concentrations, between ~10 and 9 cal. kyr BP are interpreted as a brief interval of high annual precipitation. Our results indicate a stepwise decrease in effective moisture in the mid-Holocene and an accelerated drying trend in the late Holocene, characteristics typical of an East Asian summer monsoon–influenced region. The broad similarities between our data and other lake records from central-eastern Inner Mongolia, well-dated speleothem δ¹⁸O records from southern and central China, and summer insolation at 45°N, support the idea that vegetation and climate changes in the Hunshandake Sandy Land were controlled mainly by fluctuations in the response of the East Asian monsoon to Northern Hemisphere summer insolation.

Keywords

Asian summer monsoon effective moisture Hunshandake Sandy Land mineral magnetism paleoclimate pollen

Introduction

Knowledge of climatic variability during the Holocene and the last deglacial transition is important for a better understanding of the natural background variability of paleoclimate and paleoenvironmental change (e.g. An et al., 2000; Stott et al., 2004; Wang et al., 2005; Yuan et al., 2004). Climatic transition zones are particularly sensitive to climate change, and thus have been extensively used for paleoclimatic reconstructions (e.g. Clarke and Rendell, 1998; Peck et al., 2002). The semi-arid and arid areas of the Mongolian Plateau and surrounding areas are of interest for paleoenvironmental and paleoclimatic studies because the environments in these regions are sensitive to climatic change (e.g. Feng et al., 2006a, 2006b; Sun and Ding, 1998; Sun et al., 1999; Yang et al., 2013; Zhang et al., 2011). In recent decades, a number of geological records from these regions, including eolian deposits, lake sediments, and soils, have been examined to reconstruct the environmental history since the last deglaciation, particularly during the Holocene (e.g. An et al., 2000; Chen et al., 2003; Feng et al., 2006a, 2006b; Jiang et al., 2010; Shi et al., 1993; Xiao et al., 2004, 2006, 2009; Yang et al., 2011, 2013). While some studies proposed that conditions were arid during the early Holocene in the western part of arid central Asia dominated by the Westerlies (Chen et al., 2008; Feng et al., 2005; Yang et al., 2011; Yang and Scuderi, 2010), the central-eastern part of arid central Asia, in the Westerlies–monsoon transition zone, was humid during the early Holocene (Jiang et al., 2010; Lee et al., 2013; Yang et al., 2013). Controversies over both the timing and duration of the Holocene climatic optimum in this region, in addition to spatial differences in climatic variations on centennial to decadal time-scales, have hampered detailed regional comparisons and caused difficulty for the prediction of the future effects of climate change on Mongolian ecosystems. These discrepancies may be partly because of the uncertainties in proxy interpretation and dating and by spatial differences in the Asian monsoon system. Thus, more independent multi-proxy data sets are needed to improve understanding of climatic change in the monsoon–Westerlies transition zone.

The Hunshandake Sandy Land in the semi-arid zone of eastern Inner Mongolia is located at the northern margin of the East Asian summer monsoon (EASM) and at the eastern edge of the Northern Hemisphere Westerlies (NHW; Figure 1). The dune landscape of the Sandy Land is therefore jointly influenced by the East Asian monsoon (EAM) and the NHW (Jiang et al., 2010; Yang et al., 2013). Earlier studies, primarily based on the ages and stratigraphic relationships between paleosols and eolian sand layers in the eolian sequences and on geochemical and pollen-derived moisture proxies in lake sediments, reported that significant environmental change has occurred in this region (e.g. Jiang et al., 2010; Xiao et al., 2008; Yang et al., 2013). It is therefore important to perform multi-proxy paleoenvironmental studies at different sites in order to provide a comprehensive understanding of climatic patterns in central-eastern Inner Mongolia.

Figure 1.

Topography and atmospheric circulation systems of China. Arrows indicate the wind directions of the East Asian summer monsoon (red), the East Asian winter monsoon (blue), the Indian monsoon (orange), and the mid-latitude Westerlies (lilac, dashed). Crosses mark the locations of Lake Xiarinur (red) and other sites referred to in the text (yellow): Lake Bayancagan (41°39′N, 115°13′E), Lake Angulinur (41°18′N, 114°20′E), Lake Dalinur (43°13–23′N, 116°29–45′E), Lake Daihai (40°35′N, 112°41′E), Lake Hulun (49°8′N, 117°30′E), Lake Ulaan (44°30′N, 103°39′E), Sanbao Cave (31°40′N, 110°26′E), and Dongge Cave (25°17′N, 108°5′E). The inset shows the bathymetry of Lake Xiarinur and the location of the coring sites (solid red circle).

In previous studies, a magnetic approach was seldom used to reconstruct short-term environmental fluctuations from different archives developed in the Hunshandake Sandy Land. By measuring variations in the composition, concentration, and grain size of magnetic minerals, mineral magnetism has been widely used to reconstruct paleoclimatic and paleoenvironmental processes recorded in marine sediments (Robinson, 1986; Rousse et al., 2006), lake sediments (Chen et al., 1999a, 1999b; Thouveny et al., 1994; Wang et al., 2008), and eolian sediments (Fang et al., 1999; Heller and Liu, 1984; Kukla et al., 1988). Of these natural archives, lake sediments have been widely used for studying environmental processes on a wide range of spatial and temporal scales (Li et al., 2006; Thompson and Oldfield, 1986; Verosub and Roberts, 1995). This is partly because lake sediments can provide data with higher temporal resolution than marine sediments, and they also offer abundant information on diverse aspects of environmental change, such as erosion history and soil-forming processes, in addition to the climatic history of continental regions and regional responses to large-scale climate change (Dearing et al., 2001; Oldfield et al., 1979; Rosenbaum and Reynolds, 2004).

Lake Xiarinur, located in the heart of the Hunshandake Sandy Land, is likely to be sensitive to the intensity and fine-scale shifts of the climatic drivers in eastern Inner Mongolia. Here, we present combined records of high-resolution magnetic, pollen, and particle size analyses of sediments from Lake Xiarinur to improve understanding of paleomonsoon variability since the last deglaciation of this region. The main aims of the study are as follows: (1) to assess the potential of mineral magnetic properties for reconstructing paleoenvironmental changes in the semi-arid Hunshandake Sandy Land, (2) to understand the processes of formation and evolution of the lake sediments using combined pollen and magneto-climatological analyses, and (3) to evaluate the variable contributions of the Holocene EAM and NHW in this climatically sensitive region.

Study site

Lake Xiarinur (42°37′N, 115°28′E) is a small closed saline lake, situated in the central area of the Hunshandake Sandy Land in central-eastern Inner Mongolia, ~260 km north of Beijing (Figure 1). The lake has a surface area of 3.1 km² and a maximum water depth of 1–2 m. The catchment is surrounded by semi-fixed dunes of ~2–20 m in height, on which the vegetation is dominated by meadow-steppe species with small forest patches surrounding the catchment area (Sun et al., 2007; Zheng et al., 2006). The mean annual temperature and precipitation in the area are ~1.5°C and ~365 mm, respectively. About 80–90% of the precipitation occurs between July and September, a distribution that is typical of regions influenced by the EAM. From the middle of November to early April, the ground is covered with ice. During winter, the Mongolian High generates the prevailing cold, dry, northwesterly airflow which brings strong winds and cold air to the region. During summer, warm and moist air masses are driven northwards by the pressure gradient between the Subtropical High and the Continental Low; they interact with cold air from the northwest, producing the majority of the annual precipitation.

Sampling and methods

Sampling

A suite of overlapping piston cores has been retrieved from Lake Xiarinur since 2007 (Figure 2). Core XN-2007 did not reach the lake basement, while core XRD-2012 reached eolian sands at the bottom of the lake sediment sequence, at depths of <4 m. In this study, we focused on the upper 5-m-long composite section of XRD-2012 that represents the complete sedimentary succession. The cores were split into half longitudinally; one half was sampled at 1-cm intervals for determination of water content and dry density and for pollen, and grain-size analyses, while the other half was used for mineral magnetic analysis. A total of 240 magnetic samples were extracted from the split face of the core sections using a continuous sequence of cubic plastic boxes (external dimensions 2 cm × 2 cm × 2 cm with an internal volume of 7 cm³). All samples were stored at 4°C. Radiocarbon ages were determined by accelerator mass spectrometry in the Poznań Radiocarbon Laboratory, Poland, and Peking University, China. All AMS ¹⁴C ages were calibrated using the program CALIB 4.3 (Stuiver et al., 1998).

Figure 2.

Lithological column and age–depth model for Lake Xiarinur. The black crosses and blue dots represent calibrated AMS ¹⁴C ages of cores XN-2007 and XRD-2012, respectively. The red dots denote ¹³⁷Cs-derived ages of core XRD-2012.

The upper 12 cm was sampled at a 0.5-cm interval for determination of ¹³⁷Cs and ²¹⁰Pb activity (Figure 3). Activity measurements of ¹³⁷Cs, ²¹⁰Pb, and ²²⁶Ra were carried out by gamma spectrometry using a low-background well-type germanium detector (EGPC 100P-15R). Each sample was packed in a 15-mm polyethylene tube for 3 weeks of storage in sealed containers to allow radioactive equilibration. Total ²¹⁰Pb activity was determined by gamma spectrometry via its energy at 46.5 keV. The short-lived ²²⁶Ra daughter nuclides ²¹⁴Pb (241.9, 295.2, and 351.9 keV) and ²¹⁴Bi (609.3 keV) were measured to determine the supported ²¹⁰Pb content in order to calculate the unsupported ²¹⁰Pb content. ¹³⁷Cs was measured by its emissions at 662 keV.

Figure 3.

¹³⁷Cs, ²¹⁰Pb, and ²²⁶Ra activities versus depth for core XRD-2012.

Mineral magnetic analyses

The magnetic susceptibility (χ, normalized by mass) was determined for all samples using an AGICO MFK1-FA Kappabridge magnetic susceptibility meter at frequencies of 976, 3904, and 15,616 Hz (with sensitivities of 2 × 10⁻⁸, 6 × 10⁻⁸, and 12 × 10⁻⁸ SI, respectively) in a peak magnetic field of 200 A/m. Here, χ_fd is defined as χ_976 Hz – χ_15,616 Hz. Natural Remanent Magnetization (NRM), Anhysteretic Remanent Magnetization (ARM), and Saturation Isothermal Remanent Magnetization (SIRM) of all samples were measured using a 2G Enterprises Model 755-4K superconducting rock magnetometer housed in a magnetically shielded room. ARM was imparted in a 2G alternating field demagnetizer in a DC field of 0.05 mT and a maximum AC field of 100 mT; it is expressed as an anhysteretic susceptibility (χ_ARM) by dividing by the strength of the DC field. SIRM was imparted in a DC field of 1 T using an ASC Model IM-10-30 pulse magnetizer. Back-field IRM was imparted in a DC field of 0.3 T to obtain the ratio S_−0.3 T (−IRM_−0.3 T/SIRM).

A total of 19 representative samples were selected for hysteresis and first-order reversal curve (FORC) measurements using a MicroMag 3900 AGM. The hysteresis loops were corrected for high field magnetic susceptibility, enabling the calculation of the ratio of saturation remanent magnetization to saturation magnetization (M_rs/M_s) and coercivity (B_c). DC demagnetization fields were used to determine the coercivity of remanence (B_cr). A total of 120 reversal curves were measured with an averaging time of 0.2 s per data point, and a smoothing factor of 5 was applied to suppress noise in the contours of the FORC diagram.

Pollen and particle size analyses

Pollen analysis was carried out on 129 samples taken from the core at intervals of 2–4 cm. Lycopodium spore tablets were added to samples to calculate pollen concentrations (Moore et al., 1991). Sediment samples (1–3 g) were treated with 10% HCl, 10% KOH, and 36% HF to remove carbonate, humid acid, and silicate, respectively, before mounting the fossil pollen samples on microscopic slides using water-free glycerol. The treated samples were examined using an Olympus BX51 transmitted light microscope. The remains of Pediastrum were also counted. Particle size analysis was performed on 165 samples taken from the core at 2-cm intervals using a Malvern Mastersizer 2000 laser particle size analyzer after the carbonate-free samples had been ultrasonically treated in a 10% (NaPO₃)₆ solution.

Results

Lithology and chronology

The cores XN-2007 and XRD-2012 are well correlated, except for obvious compression in the upper 90 cm of core XRD-2012 because of the freezing of the uppermost sediment (Figure 2). Freezing of the uppermost part of the sediment column is the result of shallowing of the lake (the water depth was only 0.2 m at the sampling site) caused by the extraction of ground water for agriculture in recent years. The lithological sequence can be divided into two parts: a lower interval (>3.5-m depth) consisting of homogeneous coarse-grained eolian sands, and an upper interval of lake sediments (0–3.5 m) consisting mainly of gray, silty marls intercalated with dark, clayey silts and/or brownish-red clay (Figure 2). The change from eolian sand to lake sediments at a depth of 3.5 m, in addition to a sudden increase in water content and a corresponding decrease in dry density, marks the formation of the lake at that level.

The ¹³⁷Cs, ²¹⁰Pb, and ²²⁶Ra activities versus depth in core XRD-2012 are shown in Figure 3. An abrupt increase in ¹³⁷Cs concentration at 5-cm depth is associated with the abrupt increase in nuclear testing in 1951/1952 (Crusius and Anderson, 1995), and provides satisfactory age control for the uppermost sediments. The absence of a prominent ¹³⁷Cs peak in the upper 5 cm may be attributed to disturbance of the surface sediments either during coring or by human activity. ²¹⁰Pb dating cannot be used because of the low levels of unsupported ²¹⁰Pb in the sediments, coarse-grained sediments, and large variations in accumulation rates, as suggested by Lamoureux and Gilbert (2004). The final age–depth model was constructed using linear interpolation of five AMS ¹⁴C dates from terrestrial plant remains (Table 1) together with the ¹³⁷Cs-derived ages for the uppermost sediments (Figure 2).

Table 1.

Radiocarbon ages from Lake Xiarinur.

Laboratory no.	Core	Material	Depth (cm)	Radiocarbon age (yr BP)	Calibrated age (median; cal. yr BP)
BA08846	XN-2007	Bulk	5.0	240 ± 45	294
Poz-50119	XRD-2012	Plant remains	20.0	1550 ± 80	1414
BA091134	XN-2007	Bulk	25.0	1370 ± 35	1290
BA07716	XN-2007	Bulk	28.5	1790 ± 90	1710
BA07717	XN-2007	Bulk	48.0	2580 ± 35	2743
Poz50120	XRD-2012	Plant remains	53.0	1170 ± 40	1063
BA07718	XN-2007	Bulk	82.5	3725 ± 50	4087
BA08845	XN-2007	Bulk	121.0	6235 ± 40	7169
BA08847	XN-2007	Bulk	159.0	7615 ± 45	7965
Poz-50121	XRD-2012	Grass	163.0	8110 ± 50	9025
Poz-42555	XN-2007	Bulk	169.0	7790 ± 35	8567
BA07719	XN-2007	Bulk	191.0	8490 ± 40	9509
Poz50122	XRD-2012	Branch	245.0	9390 ± 50	10,613
BA07720	XN-2007	Bulk	247.0	9320 ± 50	10,522
BA07721	XN-2007	Grass	247.0	9530 ± 55	10,982
Poz-50124	XRD-2012	Plant remains	365.0	12,600 ± 60	14,606
Poz-50125	XRD-2012	Plant remains	388.0	13,010 ± 60	15,643

All AMS ¹⁴C ages were calibrated using the atmospheric data set from the calibration program CALIB 4.3 (Stuiver et al., 1998). The AMS ¹⁴C age at the depth of 20 cm is too old and was not used in the age model.

Pollen assemblage zones

A total of 50 terrestrial fossil pollen types were identified, including 4 coniferous trees, 16 broadleaf trees, and 29 shrubs and herbs. In addition Pediastrum, one of the genera of green algae that are most abundant in very shallow lakes, was identified. Using the concentrations of the major pollen taxa and the affinity scores of meadow-steppe, broadleaf, and coniferous trees, the composite core can be divided into six pollen assemblage zones, as follows (Figure 4).

Figure 4.

Pollen concentration diagram for the Lake Xiarinur sediment core, including sediment dates. Six pollen assemblage zones are indicated to the right.

Zone 1 (3.9–3.5 m; ~15.3–14.1 cal. kyr BP)

The pollen concentration is relatively low in this zone (138–7,315 grains/g). The pollen assemblages are dominated by herbs, with Artemisia (55.2–72.6%), Ranunculaceae (1.28–13.04%), and Chenopodiaceae (2.2–14%) being the dominant taxa (Figure 4). The consistently low percentages and concentrations of pollen from coniferous trees, and from plants from aquatic and wetland environments, suggest relatively dry and cold conditions (Figure 4). Pediastrum was not detected, indicating that the lake did not exist at this time.

Zone 2 (3.5–3.2 m; ~14.1–13.3 cal. kyr BP)

The pollen concentrations are significantly higher in this zone (23,460–164,214 grains/g). Shrubs and herbs (Artemisia, Chenopodiaceae, Ranunculaceae, and Gramineae) are the dominant taxa, comprising 77.9–90% of total pollen, with Artemisia as the major contributor. Broadleaf trees comprise the majority of the remaining total pollen (8–20.3%), including Betula (3.4–14.1%), Carpinus (0.4–2.2%), Quercus (<1.7%), and Ulmus (<1%). Conifer pollen occurs at relatively low levels. Percentages of wetland herbs (Cyperaceae <4.7%) and aquatic Typha are significantly higher than in Zone 1. Pteridophyte spores are absent. Meadow grassland is the inferred vegetation type. Pediastrum is present in this zone at ~14.1 cal. kyr BP, indicating the formation of the lake.

Zone 3 (3.2–2.8 m; ~13.3–11.7 cal. kyr BP)

Zone 3 is characterized by a dramatic decrease in pollen concentrations (12,264–94,697 grains/g). Shrubs and herbs comprise 80–88.5% of the total pollen, with Artemisia as the dominant taxon (56.1–66.9%). Chenopodiaceae (4.4–9.3%), Ranunculaceae (4.7–9%), and Gramineae (2.0–8.2%) are significantly represented. Broadleaf trees comprise 8.7–18.5% of total pollen, with Betula (5.1–10.1%) the most significant component. Conifer pollen (e.g. Pinus, Abies, and Picea) is present at a consistently low level (0.7%) and probably represent long-distance transport. Percentages of aquatic and wetland pollen types are much lower (0.5–2.8%) than in Zone 2; Cyperaceae (0.5–2.5%) is the major component. Pediastrum occurs intermittently. Meadow grassland continues to dominate in this zone; however, the decreased pollen concentration indicates a trend toward sparser vegetation coverage.

Zone 4 (2.8–1.6 m; ~11.7–8.5 cal. kyr BP)

This zone is marked by a significant rise in total pollen concentration (76,846–486,550 grains/g). Although shrubs and herbs are dominant (76.7–90.6%), Artemisia percentages (31.2–60.3%) are much lower than in Zone 3. The representation of Chenopodiaceae (4.9–16.8%) and Gramineae (2.2–9.6%) increases steadily. Broadleaf trees (6.1–17.7%), coniferous trees (0.2–3.3%), and aquatic and wetland taxa (2.1–8.5%) are represented. Pediastrum reaches a peak at 2.3-m depth (~10.3 cal. kyr BP), indicating the highest lake level recorded during the interval of study. The high pollen concentrations and assemblages indicate a continuation of herb- and shrub-rich meadow grassland but with much denser vegetation coverage under an inferred relatively warm and moist climate.

Zone 5 (1.6–1.0 m; ~8.5–4.2 cal. kyr BP)

The pollen concentration in Zone 5 (72,630–513,375 grains/g) is generally lower than in Zone 4. Artemisia comprises 45.8–57.5% of total pollen, while Gramineae and Chenopodiaceae contribute 1.3–7.5% and 2.6–16%, respectively. The percentage of Betula increases slightly, while the percentages of both Ulmus (0.8–4.6%) and Carpinus (2.1–4.5%) decrease relative to Zone 4. The maximum total pollen and Pediastrum concentration occur at ~5.7 cal. kyr BP, reflecting a short interval of optimum climatic conditions.

Zone 6 (1–0 m; ~4.2 cal. kyr BP–present)

The most significant feature of Zone 6 is the sharp decrease in total pollen concentration (1,216–72,983 grains/g) from ~4.2 cal. kyr BP. Shrubs and herbs comprise a relatively low proportion of the total pollen (32.4–49.9%), while the percentages of conifers (10.2–70.4%) increased, and those of broadleaf trees (48–6,243 grains/g) declined. Pollen concentrations of wetland taxa, including Cyperaceae and Typha, are significantly lower than in Zone 5, although they still have relatively high percentages. Pediastrum exhibits a decreasing trend in this zone, indicating the continued shrinkage of the lake.

Mineral magnetic properties

The combination of the Day plot and the FORC diagram provides a powerful tool to identify both the coercivity and grain-size distribution of magnetic materials (Day et al., 1977; Roberts et al., 2000). The hysteresis loops are all closed below 500 mT, suggesting the predominance of low-coercivity ferrimagnetic minerals (Figure 5a and b). The average values (n = 19) of B_c and B_cr are 13.2 mT and 30.6 mT, respectively, both of which are close to the theoretical values of single-domain (SD) magnetite at 10 mT and 33 mT, respectively. In all FORC diagrams, the contours along the B_u axis represent the axis of the interaction field among magnetic grains between 5 and 10 mT (Figure 5c and d), indicating the dominance of SD magnetite in the samples (Muxworthy and Dunlop, 2002; Roberts et al., 2000). Compared with larger interacting grains, weakly or noninteracting uniaxial single-domain (UNISD) magnetite particles are characterized by relatively closed contours along the B_c axis and a narrow spread along the B_u axis (Egli et al., 2010; Muxworthy and Roberts, 2007). The FORC diagrams exhibit a narrow (central) ridge along the B_c axis, which is characteristic of UNISD particles (Figure 5c and d). Progressive acquisition of IRM in magnetic fields of <1 T shows that for both high- and low-χ samples, saturation is achieved at ~300 mT, consistent with the presence of a dominant ferrimagnetic mineral phase (Figure 5e). The hysteresis parameters (B_cr/B_c and M_rs/M_s), calculated after removal of the paramagnetic component of the magnetization, are plotted on a Day plot (Day et al., 1977) in Figure 5f. All samples fall in the pseudo single-domain (PSD) range regardless of the differences in the two ratios. Both S_−0.3 T and IRM_@100 mT/SIRM values reflect the relative proportion of anti-ferromagnetic minerals in the magnetic assemblage (Bloemendal et al., 1988; Hesse, 1997; Robinson, 1986). Values of S₋_0.3 T are close to 1 when virtually no high-coercivity minerals are present, while lower values indicate a higher content of high-coercivity minerals (Bloemendal et al., 1988). The S₋_0.3 T values are generally >0.85 throughout the core, implying the significant presence of low-coercivity minerals.

Figure 5.

Mineral magnetic parameters for the Lake Xiarinur sediment core. (a and b) Hysteresis loops, corrected for paramagnetism, and (c and d) FORC diagrams for two representative samples. (e) IRM acquisition curves, and (f) hysteresis ratios plotted on a Day diagram (Day et al., 1977).

Magnetic susceptibility (χ) commonly reflects the concentration of ferrimagnetic minerals in sediments (Thompson and Oldfield, 1986; Verosub and Roberts, 1995). ARM is notably sensitive to small SD particles, while SIRM indicates relatively coarse magnetic particles (Evans and Heller, 2003). The largely concordant variations among these magnetic concentration–dependent parameters throughout the core suggest that they are controlled mainly by variations in the concentration of magnetic minerals (Figure 6). The mass-specific χ_fd is a reliable proxy for the concentration of viscous SP particles (~20–25 nm for magnetite and maghemite; Liu et al., 2005; Zhou et al., 1990). χ_ARM/χ and χ_ARM/SIRM are widely employed to detect variations in magnetic mineral grain size (e.g. Evans and Heller, 2003; Maher and Thompson, 1991; Zhou et al., 1990). Figure 6 shows that higher χ_ARM/χ and χ_ARM/SIRM values generally correspond to higher χ, χ_fd, and ARM values, implying that the magnetic mineral grain sizes are generally finer in the intervals with higher concentrations of magnetic minerals. The ARM/SIRM ratio is also diagnostic of bacterial magnetosomes, and ARM/SIRM ratios between 0.15 and 0.25 are characteristic of intact magnetosomes (Moskovitz et al., 1993). ARM/SIRM values throughout the core are between 0.001 and 0.064, consistently lower than 0.15, suggesting a minimal contribution from bacterial magnetite.

Figure 6.

Plots of mineral magnetic parameters, water content, dry density, and grain-size distribution against age. (a) Mass-specified magnetic susceptibility (χ), (b) frequency-dependent susceptibility (χ_fd), (c) saturation isothermal remanent magnetization (SIRM), (d) anhysteretic remanent magnetization (ARM), (e) χ_ARM/χ, (f) χ_ARM/SIRM, (g) IRM_@100 mT/SIRM, (h) S₋_0.3 T (−IRM_−0.3 T/SIRM), (i) water content, (j) dry density, and (k) grain-size fraction > 35 µm. The blue bar denotes the Younger Dryas (YD) event, and the yellow bar indicates the most humid interval in the early Holocene. Here, χ_fd is defined as (χ_976 Hz - χ_15,616 Hz)/χ_976 Hz. IRM_@100 mT is the remanence after alternating field (AF) demagnetization of an IRM imparted in a 1-T field with a peak AF of 100 mT.

Discussion

Formation and evolution of Lake Xiarinur

Accurate AMS ¹⁴C dating of the formation and evolution of Lake Xiarinur is required for paleoclimatic reconstruction. In sandy land areas, it is difficult to obtain suitable material for macrofossil AMS ¹⁴C dating. In this study, macrofossils (e.g. the stems from shrubs) found in the sediment provide a reliable time control for the sedimentary sequence. Moreover, the core ages project to the sediment surface at an age that is very close to zero (Figure 2), also implying that the carbon reservoir effect is not significant in Lake Xiarinur. As shown in Figure 2, the sedimentary facies changes from eolian sands to lacustrine sediments at the depth of 3.5 m (~14.1 cal. kyr BP). A sharp increase in water content and pollen concentration at 3.5-m depth indicates that the lake formed at this time. Lake formation may have occurred in response to the Bølling–Allerød warming (Denton et al., 2010; Kienast et al., 2003), a period characterized by a rapid rise in temperature following the Last Glacial Maximum. Subsequently, Lake Xiarinur recorded the following additional climatic stages as inferred from the pollen assemblages, variations in lithology, sediment dry density, and mineral magnetic characteristics: Stage 1 (~13.3–11.7 cal. kyr BP), Stage 2 (~11.7–8.5 cal. kyr BP), Stage 3 (~8.5–4.2 cal. kyr BP), and Stage 4 (~4.2 cal. kyr BP–present), corresponding to Pollen Zones 3–6, respectively. These stages may correspond to the Younger Dryas (YD) and the early, middle, and late Holocene, respectively (Figures 4 and 5). During the YD, the vegetation was dominated by herbs and shrubs, including mainly Artemisia and Ranunculaceae, indicating relatively cold and dry conditions (Figure 4). This time interval is also characterized by a significant decrease in both pollen concentration and water content and by an increase in bulk and magnetic mineral grain sizes, as reflected by the grain-size fraction >35 µm (%) and the χ_ARM/SIRM ratio (Figure 6). We propose that this cold stage may represent a regional response to the YD event (Fairbanks, 1989; Lowe et al., 2008), during which Greenland Summit was 15°C colder, as inferred from thermally fractionated nitrogen and argon isotope ratios from the Greenland Ice Sheet Project Two (GISP2; Alley et al., 1993; Dansgaard et al., 1989). The period between ~11.7 and 8.5 cal. kyr BP was characterized by dramatic increases in both pollen concentration and water content, suggesting a dense vegetation cover (Figure 4). There was a remarkable increase in Pediastrum assemblages during the early Holocene, between 11.7 and 10.2 cal. kyr BP. Since Pediastrum, a genus of green algae, favors freshwater and has been regarded as a proxy for lake level change (Alhonen and Ristiluoma, 1973; Jiang et al., 2010; Komárek and Jankovská, 2001), the significant increase in its concentration clearly indicates a rapid increase in lake level at the onset of the Holocene. Total pollen concentrations began to decrease gradually between ~8.5 and 4.2 cal. kyr BP but still remained at relatively high levels. A significant increase in the concentrations of Pediastrum and various pollen taxa between 5.8 and 4.9 cal. kyr BP probably suggests a rise in lake level during the mid-Holocene (Figure 4). After ~4.2 cal. kyr BP, Pinus and other coniferous trees, which can survive in an arid environment, began to replace other broadleaf forest taxa. The significant reduction in the total pollen concentration during this period clearly indicates a decline in vegetation cover and general environmental degradation, although the concentrations of Pinus and other coniferous taxa were still relatively high (Figure 4). The ‘4.2 ka event’, which is characterized by cold and dry conditions, has been identified globally as a period of severe and abrupt climate change (Bond et al., 2001; Cullen et al., 2000; Stott et al., 2004; Thompson et al., 2002). At ~1 cal. kyr BP, a slight increase in the concentration of the pollen of Chenopodiaceae and aquatic plants implies a slight amelioration of the climate. In summary, the biostratigraphic data from Lake Xiarinur suggest humid conditions in the early and mid-Holocene, and a dry climate in the late Holocene. Our results provide a clear indication of the regional effects of global climate change on vegetation variations in the Hunshandake Sandy Land since the last deglaciation.

Precipitation-controlled magnetic records at Lake Xiarinur

Both wind-blown dust and a combination of sediment derived from surface runoff, rain splash, and remobilization of marginal lake sediments are potential sources of detrital material entering Lake Xiarinur. Fluctuations of the mineral magnetic properties of the lake sediments may, to some extent, reflect varying contributions from these two sources of sediment from within and beyond the catchment area (Thompson and Oldfield, 1986). Considering the location of Lake Xiarinur in relation to dust-transporting wind systems (Figure 1), both the East Asian winter monsoon and the Westerlies may have promoted inputs of eolian dust from the arid Gobi and other deserts in northwestern China when northwesterly/westerly surface winds predominated. These sandy materials may have also been transported to the lake either by strong winds during cold–dry periods, or by intensified rain splash/wave erosion under warm–humid conditions, particularly during the Holocene climatic optimum.

The vegetation changes at Lake Xiarinur since the last deglaciation indicate that maximum moisture levels occurred during the early Holocene, as indicated by the highest pollen concentrations, including abundant broadleaf trees, and of Pediastrum (Figure 4). During this period, the marked increase in annual precipitation resulted in enhanced vegetation coverage, which probably stabilized the dunes and enabled the formation of pedogenically produced soil-like material (Yang et al., 2013; Zhu et al., 1980). The occurrence of brownish-red clay in the depth interval 2.25–2.1 m (corresponding to ~10.2–9.9 cal. kyr BP) confirms that abundant and regular precipitation induced strong pedogenesis.

All magnetic concentration–dependent parameters (e.g. χ, ARM, and SIRM) show stepwise increases at the onset of the early Holocene, reaching peak values between ~10 and 9 cal. kyr BP (Figure 6). Higher χ_fd and χ_ARM/χ values during this period (Figure 6) indicate the presence of abundant fine-grained ferrimagnetic particles (Evans and Heller, 2003; Maher and Thompson, 1991; Zhou et al., 1990). This is consistent with the prevailing view that pedogenic processes result in the formation of fine-grained ferrimagnetic minerals and, therefore, significant magnetic enhancement (Banerjee et al., 1993; Maher and Thompson, 1991; Zhou et al., 1990). Previous studies have shown that ferrimagnetic minerals are usually relatively abundant in soils, particularly in the topsoil, compared with wind-blown dust and other materials originating from dunes (Oldfield et al., 1985; Peck et al., 2004). We propose that rain splash or wave erosion associated with increased rainfall and increasing lake levels eroded material surrounding the lake, causing increased detrital inputs. This eroded material probably contained higher concentrations of partly pedogenized loess-derived magnetic minerals with higher χ values, consistent with the magnetic properties of the sediment cores.

The period between ~8.5 and 4.2 cal. kyr BP is characterized by low χ, ARM, and SIRM values, implying low concentrations of ferrimagnetic minerals. Both χ_ARM/χ and χ_ARM/SIRM ratios show stepwise decreasing trends, indicative of a coarsening of the ferrimagnetic grains. A significant increase in IRM_@100 mT/SIRM and a corresponding decrease in S₋_0.3 T suggest decreasing contributions of low-coercivity ferrimagnetic grains. We attribute this to decreasing rates of rain splash/wave erosion and consequently declining surface-erosion-related input of ferrimagnetic minerals to the lake.

The period from ~4.2 cal. kyr BP to present is characterized by gradually decreasing trends of χ, ARM, and SIRM values, implying further climatic deterioration. Lower χ_ARM/χ and χ_fd values, compared with the early and middle Holocene, reflect further coarsening of the ferrimagnetic particles (Figure 6). Lower S₋_0.3 T (Figure 6h), but higher IRM_@100 mT/SIRM values (Figure 6g), suggest a further decrease in the percentages of ferrimagnetic minerals. χ_ARM/χ, χ_ARM/SIRM, and IRM_@100 mT/SIRM, together with pollen concentrations, show comparable high-amplitude variations between ~2 and 0.7 cal. kyr BP, reflecting a short time interval inferred to have been wet and warm (Figure 6).

In summary, pollen concentrations, the magnetic-grain-size-dependent parameter χ_ARM/χ, and the magnetic-composition-dependent parameter IRM_@100 mT/SIRM exhibit comparable trends of variation (Figure 7). We suggest that the enhanced magnetic properties of the Lake Xiarinur sediment cores are probably controlled by precipitation delivered by the Asian summer monsoon that resulted in detrital inputs dominated by pedogenic ferrimagnetic grains, and to a lesser extent by a varying influx of eolian dust carried by the northwesterly/westerly winds of the winter monsoon. Both the pollen and mineral magnetic data suggest high precipitation and surface erosion during the early Holocene, decreased effective moisture in the mid-Holocene, and a clear trend to arid conditions in the late Holocene.

Figure 7.

Comparison of sedimentary records from Lake Xiarinur with other lake records from Inner Mongolia, well-dated speleothem δ¹⁸O records from southern and central China, and summer insolation at 45°N. (a) Pollen concentrations from the Lake Xiarinur sediment core, (b) IRM_@100 mT/SIRM ratios of the Lake Xiarinur sediments, (c) χ_ARM/χ ratios of the Lake Xiarinur sediments, (d) annual precipitation (P_ann) at Lake Bayancagan (mm; Jiang et al., 2010), (e) total organic carbon (TOC; %) of Lake Ulaan sediments (Lee et al., 2013), (f) speleothem δ¹⁸O values (‰) from Sanbao Cave (red; Dong et al., 2010), (g) speleothem δ¹⁸O values (‰) from Dongge Cave (green; Dykoski et al., 2005; Wang et al., 2005), and (h) summer insolation at 45°N (W/m²; Berger, 1978). The dark gray, pink, blue, red, yellow, and light gray bars correspond to the Mystery Interval (Denton et al., 2006), the Bølling–Allerød warming, the Younger Dryas (YD) event, the early Holocene, the mid-Holocene, and the late Holocene, respectively.

Influence of the EASM on the climate of central-eastern Inner Mongolia

Effective moisture determines the areal extent of deserts in arid regions, the water level of lakes, and the density of vegetation cover, and therefore has a strong influence on the ecological environment (Lee et al., 2013). The effective moisture in arid Central Asia is dominated by the Westerlies and was lowest in the early Holocene and highest in the mid-Holocene (Feng et al., 2005; Xiao et al., 2004; Yang and Scuderi, 2010; Yang et al., 2011; Zhang et al., 2012; Zhao and Yu, 2012). However, in EASM-dominated regions, effective moisture was strongest in the early and mid-Holocene, and weakened after the mid-Holocene (Chen et al., 2008; Jiang et al., 2010; Lee et al., 2013; Yang et al., 2013; Zhang et al., 2011). Feng et al. (2006a) comprehensively summarized the evidence indicating that the Inner Mongolian Plateau was warm and dry during the early Holocene (from ~10.5–9.5 to ~8–7.5 cal. kyr BP), and was warm and wet from ~7.5 to ~3.5 cal. kyr BP. Using multi-proxy analyses of sediments from Lake Angulinur, Wang et al. (2010) proposed that the environment became more benign at ~10.9 cal. kyr BP, with relatively wet and warm conditions during ~10.9–8.9 cal. kyr BP, followed by environmental deterioration from ~7.4 cal. kyr BP; the driest and coolest conditions occurred during ~2.2–0.48 cal. kyr BP. Based on high-resolution, well-dated paleo-records (pollen, Pediastrum, and δ¹⁸O of authigenic carbonate) from Lake Bayancagan, Jiang et al. (2006, 2010) concluded that the most humid conditions occurred between ~10.5 and ~6.5 cal. kyr BP, and that the mid-Holocene cooling began at ~8 cal. kyr BP, with a drying phase commencing ~1.5 kyr later. At Lake Ulaan, to the east of the southeastern Gobi Altai Mountains, records of total organic carbon (TOC), C/N ratios, and weathering intensity also indicated that the effective moisture was highest during the early Holocene, and was followed by a decreasing trend, reaching the lowest values during the late Holocene (Lee et al., 2013). Lake Dalinur, located 120 km northeast of Lake Xiarinur, experienced the highest lake level during the early Holocene (~11.5–7.6 cal. kyr BP); dramatic fluctuations in lake level during the mid-Holocene (~7.6–3.4 cal. kyr BP) with three intervals of lower lake stands at ~6.6–5.8, ~5.1–4.8, and ~4.45–3.75 cal. kyr BP, and a general shrinking trend of the lake during the late Holocene (~3.45 cal. kyr BP to present; Xiao et al., 2009). The most recent study of eolian sequences in the Hunshandake Sandy Land clearly indicates the formation of paleosols at ~9.6–3 cal. kyr BP, confirming that the dune landscape was stabilized by vegetation to a large degree during the early and middle Holocene (Yang et al., 2013). Syntheses of moisture changes in monsoon-influenced regions also demonstrate that a humid climate dominated during the early and middle Holocene, and that a drier climate prevailed during the late Holocene, with an abrupt shift at ~4.5 cal. kyr BP (Herzschuh, 2006; Zhao et al., 2009).

The persistent deterioration of the climate at Lake Xiarinur since ~4.2 cal. kyr BP is also consistent with several climate records from central-eastern Inner Mongolia and adjacent regions. At Lake Dalinur, low organic carbon and high inorganic carbon concentrations during ~4.45–3.75 cal. kyr BP suggest a major interval of decreased monsoonal precipitation (Xiao et al., 2009). At Lake Hulun, a dramatic fall in tree and shrub pollen percentages indicates an extremely dry event at ~4.4–4.2 cal. kyr BP (Wen et al., 2010a, 2010b). This event is also recorded at Lake Daihai, as indicated by the decline in woody plants at ~4.45–2.9 cal. kyr BP (Xiao et al., 2004). The upper boundary of a dark paleosol layer at a well-exposed section at Saihanxili, in the western part of the Hunshandake Sandy Land, was dated to ~4.2 cal. kyr BP, indicating the onset of drier conditions in the late Holocene (Yang et al., 2013).

The location of Lake Xiarinur at the margin of the EASM and the mid-latitude Westerlies implies that both of these atmospheric circulation systems probably played key roles in determining the course of environmental change in this area. Figure 7 shows that the pollen and magnetic records from Lake Xiarinur, annual precipitation reconstructions at Lake Bayancagan, and TOC (%) records of Lake Ulaan, all exhibit very similar trends of variation during the Holocene, typical of a summer-monsoon-influenced domain and in broad agreement with the synthesized EAM strength (Wang and Feng, 2013). A striking resemblance between pollen concentrations from Lake Xiarinur and two absolute-dated stalagmite δ¹⁸O records from Sanbao Cave (Dong et al., 2010) and Dongge Cave (Dykoski et al., 2005; Hu et al., 2008; Wang et al., 2005), both of which are characterized by rapid decreases in δ¹⁸O values during the early Holocene that gradually increase toward the present, further confirms that the lake records from central-eastern Inner Mongolia provide a consistent pattern of Asian summer monsoon precipitation-controlled Holocene climate change (Figure 7). This conclusion can be reached regardless of the controversies over the contributions of EASM and Indian summer monsoon to speleothem oxygen isotope records from southern and central China (Dong et al., 2010; Dykoski et al., 2005; Wang et al., 2005). Comparison of the lake proxy records with summer insolation at 45°N (Figure 7) supports the idea that vegetation patterns and regional climate changes in central-eastern Inner Mongolia during the Holocene were controlled by fluctuations in the response of the EASM to Northern Hemisphere summer insolation, which acts as a major driver of global temperature change (Kutzbach and Gallimore, 1988).

Conclusion

A high-resolution, radiocarbon and ¹³⁷Cs dated, pollen and mineral magnetic record from Lake Xiarinur in the Hunshandake Sandy Land, central-eastern Inner Mongolia, reveals a detailed sequence of environmental changes during the last deglaciation and the Holocene. A change from eolian sand deposits to a lacustrine sequence at ~14.1 cal. kyr BP records the formation of Lake Xiarinur during the Bølling–Allerød warming, in response to rapid Northern Hemisphere deglaciation. The pollen data demonstrate that conditions began to ameliorate at ~11.7 cal. kyr BP, marking the end of the YD stadial, and maximum moisture levels were reached during the early Holocene. A gradual decrease in effective moisture occurred during the mid-Holocene (~8.5–4.2 cal. kyr BP), and an environmental deterioration toward increased aridity commenced at ~4.2 cal. kyr BP. Increases in the concentration of Pediastrum at ~11.7–10.2 and ~5.8–4.9 cal. kyr BP are inferred to represent two wet intervals with substantial rises in lake level. Mineral magnetic analysis reveals that the highest concentrations of magnetic minerals occurred at ~10–9 cal. kyr BP, which is interpreted as reflecting high detrital input of loess-derived magnetic minerals with a significant component of ferrimagnetic minerals of pedogenic origin by either rain splash or wave erosion associated with increased rainfall and/or a rise in lake level. The subsequent decline in the concentration of magnetic minerals is thus attributed to declining surface-erosion-related input of ferrimagnetic minerals into the lake. The stepwise decline in effective humidity through the Holocene, as inferred from parallel vegetation and magneto-climatological variations at Lake Xiarinur, is broadly consistent with other lake records from central-eastern Inner Mongolia, well-dated speleothem δ¹⁸O records from southern and central China, and summer insolation at 45°N. We therefore propose that regional climatic changes and associated variations in the density and composition of vegetation cover in the Hunshandake Sandy Land during the last deglaciation and throughout the Holocene were controlled mainly by fluctuations in the response of the EASM to Northern Hemisphere summer insolation.

Footnotes

Acknowledgements

We sincerely thank Professor Zhaoyan Gu, Dr Yongfu Chen, Bing Xu, and Zihua Tang for help in field sampling and in the laboratory. This manuscript benefits considerably from thorough and thoughtful reviews by J Bloemendal and JW Zhang.

Funding

This study was supported by the China Geological Survey (grant number 1212011120045) and the National Natural Science Foundation of China (grants 41274074 and 41371219).

References

Alhonen

Ristiluoma

(1973) On the occurrence of subfossil Pediastrum algae in a Flandrian core at Kirkkonummi, southern Finland. Bulletin of the Geological Society of Finland 45: 73–77.

Alley

Meese

Shuman

. (1993) Abrupt increase in Greenland snow accumulation at the end of the Younger Dryas event. Nature 362: 527–529.

Porter

Kutzbach

. (2000) Asynchronous Holocene optimum of the East Asian monsoon. Quaternary Science Reviews 19: 743–762.

Banerjee

Hunt

Liu

(1993) Separation of local signals from the regional paleomonsoon record of the Chinese Loess Plateau: A rock magnetic approach. Geophysical Research Letters 20: 843–846.

Berger

(1978) Long-term variations of daily insolation and Quaternary climatic changes. Journal of the Atmospheric Sciences 35: 2362–2367.

Bloemendal

Lamb

King

(1988) Paleoenvironmental implications of rock-magnetic properties of Late Quaternary sediment cores from the eastern equatorial Atlantic. Paleoceanography 3: 61–87.

Bond

Kromer

Beer

. (2001) Persistent solar influence on North Atlantic climate during the Holocene. Science 294: 2130–2136.

Chen

CTA

Lan

Lou

. (2003) The dry Holocene Megathermal in Inner Mongolia. Palaeogeography, Palaeoclimatology, Palaeoecology 193: 181–200.

Chen

Bloemendal

Zhang

. (1999a) An 800 ky proxy record of climate from lake sediments of the Zoige Basin, the Eastern Tibetan Plateau. Palaeogeography, Palaeoclimatology, Palaeoecology 151: 307–320.

10.

Chen

Shi

Wang

(1999b) Environmental changes documented by sedimentation of Lake Yiema in arid China since the late Glaciation. Journal of Paleolimnology 22: 159–169.

11.

Chen

Yang

. (2008) Holocene moisture evolution in arid central Asia and its out-of-phase relationship with Asian monsoon history. Quaternary Science Reviews 27: 351–364.

12.

Clarke

Rendell

(1998) Climate change impacts on sand supply and the formation of desert sand dunes in the southwest USA. Journal of Arid Environments 39: 517–531.

13.

Crusius

Anderson

(1995) Evaluating the mobility of ¹³⁷Cs, ^{239 + 240}Pu and ²¹⁰Pb from their distributions in laminated lake sediments. Journal of Paleolimnology 13: 119–141.

14.

Cullen

deMenocal

Hemming

. (2000) Climate change and the collapse of the Akkadian empire: Evidence from the deep sea. Geology 28: 379–382.

15.

Dansgaard

White

JWC

Johnsen

(1989) The abrupt termination of the Younger Dryas climate event. Nature 339: 532–533.

16.

Day

Fuller

Schmidt

(1977) Hysteresis properties of titanomagnetites: Grain size and composition dependence. Physics of the Earth and Planetary Interiors 13: 260–267.

17.

Dearing

Doody

. (2001) Preliminary reconstruction of sediment-source linkages for the past 6000 year at the Petit Lac d’Annecy, France, based on mineral magnetic data. Journal of Paleolimnology 25: 245–258.

18.

Denton

Anderson

Toggweiler

. (2010) The last glacial termination. Science 328: 1652–1656.

19.

Denton

Broecker

Alley

(2006) The mystery interval 17.5 to 14.5 kyrs ago. PAGES News 14: 14–16.

20.

Dong

Wang

Cheng

. (2010) A high-resolution stalagmite record of the Holocene East Asian Monsoon from Mt Shennongjia, central China. The Holocene 20: 257–264.

21.

Dykoski

Edwards

Cheng

. (2005) A high-resolution, absolute-dated Holocene and deglacial Asian monsoon records from Dongge cave, China. Earth and Planetary Science Letters 233: 71–86.

22.

Egli

Sartori

Mirabella

. (2010) The effects of exposure and climate on the weathering of late Pleistocene and Holocene Alpine soils. Geomorphology 114: 466–482.

23.

Evans

Heller

(2003) Environmental Magnetism: Principles and Applications of Enviromagnetics. San Diego, CA: Academic Press.

24.

Fairbanks

(1989) A 17,000-year glacio-eustatic sea level record: Influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature 342: 637–642.

25.

Fang

Ono

Fukusawa

. (1999) Asian summer monsoon instability during the past 60,000 years: Magnetic susceptibility and pedogenic evidence from the western Chinese Loess Plateau. Earth and Planetary Science Letters 168: 219–232.

26.

Feng

Wang

(2006a) Holocene climatic and environmental changes in the arid and semi-arid areas of China: A review. The Holocene 16: 119–130.

27.

Feng

Tang

Wang

. (2006b) Holocene vegetation variations and the associated environmental changes in the western part of the Chinese Loess Plateau. Palaeogeography, Palaeoclimatology, Palaeoecology 241: 440–456.

28.

Feng

Wang

Guo

. (2005) Lacustrine and eolian records of Holocene climate changes in the Mongolian Plateau: Preliminary results. Quaternary International 136: 25–32.

29.

Heller

Liu

(1984) Magnetism of Chinese loess deposits. Geophysical Journal International 77: 125–141.

30.

Herzschuh

(2006) Palaeo-moisture evolution in monsoonal central Asia during the last 50,000 years. Quaternary Science Reviews 25: 163–178.

31.

Hesse

(1997) Mineral magnetic ‘tracing’ of aeolian dust in southwest Pacific sediments. Palaeogeography, Palaeoclimatology, Palaeoecology 131: 327–353.

32.

Henderson

Huang

. (2008) Quantification of Holocene Asian monsoon rainfall from spatially separated cave records. Earth and Planetary Science Letters 226: 221–232.

33.

Jiang

Guiot

Chu

. (2010) An improved methodology of the modern analogues technique for palaeoclimate reconstruction in arid and semi-arid regions. Boreas 39: 145–153.

34.

Jiang

Guo

Sun

. (2006) Reconstruction of climate and vegetation changes of Lake Bayanchagan (Inner Mongolia): Holocene variability of the East Asian monsoon. Quaternary Research 65: 411–420.

35.

Kienast

Hanebuth

TJJ

Pelejero

. (2003) Synchroneity of meltwater pulse 1a and the Bølling warming: New evidence from the South China Sea. Geology 31: 67–70.

36.

Komárek

Jankovská

(2001) Review of the Green Algal Genus Pediastrum; Implication for Pollen-Analytical Research (Bibliotheca Phycologica, vol. 108). Berlin: J. Cramer, 127 pp.

37.

Kukla

Heller

Liu

. (1988) Pleistocene climates in China dated by magnetic susceptibility. Geology 16: 811–814.

38.

Kutzbach

Gallimore

(1988) Sensitivity of a coupled atmosphere/mixed layer ocean model to changes in orbital forcing at 9000 years B.P. Journal of Geophysical Research 93: 803–821.

39.

Lamoureux

Gilbert

(2004) A 750-yr record of autumn snowfall and temperature variability and winter storminess recorded in the varved sediments of Bear Lake, Devon Island, Arctic Canada. Quaternary Research 61: 134–147.

40.

Lee

Lim

. (2013) Late Pleistocene-Holocene records from Lake Ulaan, southern Mongolia: Implications for east Asian palaeomonsoonal climate changes. Journal of Quaternary Science 28: 370–378.

41.

Kodama

. (2006) A 14,000-year environmental change history revealed by mineral magnetic data from White Lake, New Jersey, USA. Earth and Planetary Science Letters 246: 27–40.

42.

Liu

Torrent

Maher

. (2005) Quantifying grain size distribution of pedogenic magnetic particles in Chinese loess and its significance for pedogenesis. Journal of Geophysical Research 110(11). DOI: 10.1029/2005JB003726.

43.

Lowe

Rasmussen

Björck

. (2008) Synchronisation of palaeoenvironmental events in the North Atlantic region during the Last Termination: A revised protocol recommended by the INTIMATE group. Quaternary Science Reviews 27: 6–17.

44.

Maher

Thompson

(1991) Mineral magnetic record of the Chinese loess and paleosols. Geology 19: 3–6.

45.

Moore

Web

Collinson

(1991) Pollen Analysis. 2nd Edition. Oxford: Blackwell Scientific Publication.

46.

Moskovitz

Frankel

Bazylinski

(1993) Rock magnetic criteria for the detection of biogenic magnetite. Earth and Planetary Science Letters 120: 283–300.

47.

Muxworthy

Dunlop

(2002) First-order reversal curve (FORC) diagrams for pseudo-single-domain magnetites at high temperature. Earth and Planetary Science Letters 203: 369–382.

48.

Muxworthy

Roberts

(2007) First-Order Reversal Curve (FORC) Diagrams, in Encyclopedia of Geomagnetism and Paleomagnetism. Dordrecht: Springer, pp. 266–272.

49.

Oldfield

Bloemendal

Thompson

(1979) Magnetic measurements used to access sediment influx at Llyn Goddionduon. Nature 280: 50–53.

50.

Oldfield

Maher

Donoghue

. (1985) Particle-size related, mineral magnetic source sediment linkages in the Rhode River catchment, Maryland, USA. Journal of the Geological Society 142: 1035–1046.

51.

Peck

Green

Shanahan

. (2004) A magnetic mineral record of Late Quaternary tropical climate variability from Lake Bosumtwi, Ghana. Palaeogeography, Palaeoclimatology, Palaeoecology 215: 37–57.

52.

Peck

Khosbayar

Fowell

. (2002) Mid to Late Holocene climate change in north central Mongolia as recorded in the sediments of Lake Telmen. Palaeogeography, Palaeoclimatology, Palaeoecology 183: 135–153.

53.

Roberts

Pike

Verosub

(2000) First-order reversal curve diagrams: A new tool for characterizing the magnetic properties of natural samples. Journal of Geophysical Research 105: 28461–28475.

54.

Robinson

(1986) The late Pleistocene palaeoclimatic record of North Atlantic deep-sea sediments revealed by mineral-magnetic measurements. Physics of the Earth and Planetary Interiors 42: 22–47.

55.

Rosenbaum

Reynolds

(2004) Basis for paleoenvironmental interpretation of magnetic properties of sediment from Upper Klamath Lake (Oregon): Effects of weathering and mineralogical sorting. Journal of Paleolimnology 31: 253–265.

56.

Rousse

Kissel

Laj

. (2006) Holocene centennial to millennial-scale climatic variability: Evidence from high-resolution magnetic analyses of the last 10 cal kyr off North Iceland (core MD99-2275). Earth and Planetary Science Letters 242: 390–405.

57.

Shi

Kong

Wang

. (1993) Mid-Holocene climates and environments in China. Global and Planetary Change 7: 219–233.

58.

Stott

Cannariato

Thunell

. (2004) Decline of surface temperature and salinity in the western tropical Pacific Ocean in the Holocene epoch. Science 431: 56–59.

59.

Stuiver

Reimer

Bard

. (1998) INTCAL 98 radiocarbon age calibration, 24,000 cal BP. Radiocarbon 40: 1041–1083.

60.

Sun

Ding

(1998) Deposits and soils of the past 130,000 years at the desert-loess transition in northern China. Quaternary Research 50: 148–156.

61.

Sun

Ding

Liu

. (1999) 580,000-year environmental reconstruction from aeolian deposits at the Mu Us Desert margin, China. Quaternary Science Reviews 18: 1351–1364.

62.

Sun

Chu

Liu

. (2007) Calibration of alkenone unsaturation index with growth temperature for a lacustrine species, Chrysotila lamellose (Haptophyceae). Organic Geochemistry 38: 1226–1234.

63.

Thompson

Mosley-Thompson

Davis

. (2002) Kilimanjaro ice core records: Evidence of Holocene climate change in tropical Africa. Science 298: 589–593.

64.

Thompson

Oldfield

(1986) Environmental Magnetism. London: Allen & Unwin, pp. 1–127.

65.

Thouveny

Beaulieu

Bonifay

. (1994) Climate variations in Europe over the past 140 kyr deduced from rock magnetism. Nature 371: 503–506.

66.

Verosub

Roberts

(1995) Environmental magnetism: Past, present and future. Journal of Geophysical Research 100: 2175–2192.

67.

Wang

Liu

Zhu

. (2010) Holocene environmental changes as recorded by mineral magnetism of sediments from Anguli-nuur Lake, southeastern Inner Mongolia Plateau, China. Palaeogeography, Palaeoclimatology, Palaeoecology 285: 30–49.

68.

Wang

Feng

(2013) Holocene moisture evolution across the Mongolian Plateau and its surrounding areas: A synthesis of climatic records. Earth-Science Reviews 122: 38–57.

69.

Wang

Løvlie

. (2008) Magnetic signature of environmental change reflected by Pleistocene lacustrine sediments from the Nihewan Basin, North China. Palaeogeography, Palaeoclimatology, Palaeoecology 260: 452–462.

70.

Wang

Cheng

Edwards

. (2005) The Holocene Asian monsoon: Links to solar changes and North Atlantic climate. Science 308: 854–857.

71.

Wen

Xiao

Chang

. (2010a) Holocene climate changes in the mid-high-latitude-monsoon margin reflected by the pollen record from Hulun Lake, northeastern Inner Mongolia. Quaternary Research 73: 293–303.

72.

Wen

Xiao

Chang

. (2010b) Holocene precipitation and temperature variations in the East Asian monsoonal margin from pollen data from Hulun Lake in northeastern Inner Mongolia, China. Boreas 39: 262–272.

73.

Xiao

Chang

Wen

. (2009) Holocene weak monsoon intervals indicated by low lake levels at Hulun Lake in the monsoonal margin region of northeastern Inner Mongolia, China. The Holocene 19: 899–908.

74.

Xiao

Zhai

. (2008) Hydrology of Dali Lake in central-eastern Inner Mongolia and Holocene East Asian monsoon variability. Journal of Paleolimnology 40: 519–528.

75.

Xiao

. (2006) Holocene climate changes in the monsoon/arid transition reflected by carbon concentration in Daihai Lake of Inner Mongolia. The Holocene 16: 551–560.

76.

Xiao

Nakamura

. (2004) Holocene vegetation variation in the Daihai Lake region of north-central China: A direct indication of the Asian monsoon climatic history. Quaternary Science Reviews 23: 1669–1679.

77.

Yang

Scuderi

(2010) Hydrological and climatic changes in deserts of China since the Late Pleistocene. Quaternary Research 73: 1–9.

78.

Yang

Scuderi

Paillou

. (2011) Quaternary environmental changes in the drylands of China – A critical review. Quaternary Science Reviews 30: 3219–3233.

79.

Yang

Wang

Liu

. (2013) Initiation and variation of the dune fields in semi-arid China – With a special reference to the Hunshandake Sandy Land, Inner Mongolia. Quaternary Science Reviews 78: 369–380.

80.

Yuan

Cheng

Edward

. (2004) Timing, duration, and transitions of the last interglacial Asian monsoon. Science 304: 575–578.

81.

Zhang

Feng

. (2012) Holocene hydrological and climatic change on the northern Mongolian Plateau based on multi-proxy records from Lake Gun Nuur. Palaeogeography, Palaeoclimatology, Palaeoecology 323–325: 75–86.

82.

Zhang

Chen

Holmes

. (2011) Holocene monsoon climate documented by oxygen and carbon isotopes from lake sediments and peat bogs in China: A review and synthesis. Quaternary Science Reviews 30: 1973–1987.

83.

Zhao

(2012) Vegetation response to Holocene climate change in East Asian monsoon-margin region. Earth-Science Reviews 113: 1–10.

84.

Zhao

Chen

. (2009) Vegetation response to Holocene climate change in monsoon-influenced region of China. Earth-Science Reviews 97(1–4): 242–256.

85.

Zheng

Xie

Robert . (2006) Did climate drive ecosystem change and induce desertification in Otindag sandy land, China over the past 40 years? Journal of Arid Environments 64: 523–541 (in Chinese).

86.

Zhou

Oldfield

Wintle

. (1990) Partly pedogenic origin of magnetic variations in Chinese loess. Nature 346: 737–739.

87.

Zhu

Liu

. (1980) An Outline of Chinese Deserts. Beijing: Science Press, 107 pp (in Chinese).