Content Data

SCIENCE CHINA

Earth Sciences

RESEARCH PAPER October 2012 Vol.55 No.10: 1656 1668

doi: 10.1007/s11430-012-4447-7

Grain size records reveal variability of the East Asian Winter

Monsoon since the Middle Holocene in the Central Yellow Sea

mud area, China

HU BangQi1,2,3, YANG ZuoSheng3*, ZHAO MeiXun4, Yoshiki SAITO5,

FAN DeJiang3 & WANG LiBo1

1

Key Laboratory of Marine Hydrocarbon Resources and Environmental Geology, Qingdao Institute of Marine Geology,

Ministry of Land and Resources, Qingdao 266071, China;

2

State Key Laboratory of Loess and Quaternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi an 710075, China;

3

College of Geo-Marine Sciences, Ocean University of China, Qingdao 266100, China;

4

Key Laboratory of Marine Chemistry Theory and Technology, Ministry of Education, Ocean University of China, Qingdao 266100, China;

5

Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305-8567, Japan

Received September 19, 2011; accepted January 13, 2012; published online July 2, 2012

Three cores (ZY-1, ZY-2 and ZY-3) retrieved from the Central Yellow Sea mud (CYSM) were analyzed in sensitive grain size

and AMS 14C dating to reconstruct the history of the East Asian Winter Monsoon (EAWM) since the Middle Holocene in the

study area. The results show that these data provide a continuous history of the EAWM over the past 7.2 ka and that the

EAWM can be divided into three periods: strong and highly fluctuating during 7.2 4.2 ka BP; moderate and relatively stable

during 4.2 1.8 ka BP; and weakened during 1.8 0 ka BP. Compared with the East Asian Summer Monsoon (EASM) recorded

in the previous studies, the evolutionary history of the EAWM broadly follows the orbital-derived winter insolation with a

similar long-term step-decreased trend as the EASM. At the centennial scale, however, the EAWM intensified events correlate

well with the EASM weakened events and the North Atlantic climatic variations (Bond events 0 to 5) within the dating error,

most likely forced by the reduction of solar irradiance through changes in the oceanic-atmospheric circulation patterns.

East Asian Winter Monsoon, Central Yellow Sea mud area, sensitive grain size, solar activity, Middle Holocene

Citation: Hu B Q, Yang Z S, Zhao M X, et al. Grain size records reveal variability of the East Asian Winter Monsoon since the Middle Holocene in the Cen-

tral Yellow Sea mud area, China. Sci China Earth Sci, 2012, 55: 1656 1668, doi: 10.1007/s11430-012-4447-7

The East Asian Monsoon (EAM) is strongly influenced by global climate change in recent decades [1 3]. The East

the temperature gradient of the Eurasian continent and the Asian Summer Monsoon (EASM), due to its significant

Indo-Pacific Ocean as well as the uplift of the Tibet Plateau. hydrological character, has been well reconstructed by multi-

As an important component of the global climate system, proxies archived in lakes [4, 5], peat [6, 7], stalagmites

the EAM is responsible for transferring enormous amounts [8 11], ice cores [12, 13], and loess [14, 15].

of heat and moisture between the Eurasian continent and the Compared with the EASM, the East Asian Winter Mon-

adjacent ocean [1], and thus, knowledge of the past changes soon (EAWM) has proved to be difficult to reconstruct [16].

in EAM and its mechanism will help further explain the The grain sizes of loess-paleosol sequences have been used

as a proxy for winter monsoons in studying the large-scale

evolution of the EAWM [17 20]. However, depicting the

*Corresponding author (email: *******@***.***.**) monsoonal variability at the millennial or centennial time-

Science China Press and Springer-Verlag Berlin Heidelberg 2012 earth.scichina.com www.springerlink.com

1657

Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

scales is complicated by low sedimentation rates and dating southern Bohai Sea (BS) and along the Shandong Peninsula

difficulties in the loess-paleosol sequences. By analyzing can be resuspended under strong ocean dynamics in winter

the magnetic susceptibility (MS), S-ratio and Ti content in and then transported by the EAWM-forced coastal current

the Lake Huguang-Maar sediment in South China, Yancheva eastward via the southern Bohai Strait to the Yellow Sea

et al. [16] found an anti-relationship between the EAWM (YS) [36 40], forming a unique omega-shaped distal

and EASM during the past 16 ka and explained it as the subaqueous deltaic lobe deposited around the eastern tip of

result of north-south shifts in the intertropical convergence the Shandong Peninsula in the YS [41] (Figure 1). Field

zone (ITCZ). Subsequently, Zhang and Lu [21] argued that observations [42 44] and satellite imagery [45 47] have

the strengths of EAWM and EASM were not anti-correlated shown that the winter suspended sediment concentration

during AD 700 900 based on historical records. Further- (SSC) of the BS and YS is significantly larger than that in

more, Zhou et al. [22] and Han et al. [23] indicated that the summer. Likewise, the high SSC zones are concentrated on

Lake Huguang-Maar sediments were mainly supplied by the coastal areas, whereas they are diffused to offshore areas

local pyroclastic rocks instead of originating from remote in winter, as summarized by the phrase store in summer,

loess. Thus, the reliability of the MS, S-ratio and Ti content transport in winter [48]. Most recently, Yang et al. [49]

in the Lake Huguang-Maar sediments as a proxy of EAWM studied the sediment transport regimes off the Huanghe

should be re-evaluated. delta and in the adjacent BS in winter and summer. They

Recently, Tian et al. [24] reported the history of EAWM indicated that although the river water and sediment dis-

during the last 23 ka using the latitudinal sea surface tem- charges to the sea are much greater in summer, the intensity

perature gradient ( SST) of the north-south in the southern of sediment transport in winter is much stronger than that in

South China Sea (SCS). Their results indicated that the summer due to the powerful effect of winter storms. Based

EASM and EAWM were probably not anti-correlated dur- on field observations and MODIS imagery, Bi et al. [50]

ing the last glacial-interglacial cycle, and the EAM varia- have shown that the sediment flux through the southern

tions cannot be explained simply as the migration of the Bohai Strait mainly occurs in winter and is approximately

ITCZ. Reconstruction of the upper water column dynamics four times greater than that in summer. Strong resuspension

in the SCS also suggested that the intensity of the EAWM and enhanced coastal currents due to EAWM activity are

and EASM shows an inverse behavior during the last glacial responsible for this strong seasonal variability [50]. More-

over, the spatial distribution pattern of the 210Pb sedimenta-

and deglaciation but covaried during the middle to late

tion rates in the YS [51] is broadly similar to the Holocene

Holocene [25, 26], consistent with model simulation [27].

mud thickness [41], which represents a late Holocene High-

By compiling a variety of alkenone-based SST estimates in

stand System Tract. This scenario suggests that the sedi-

the SCS, Huang et al. [28] reconstructed the variations of

ment transport and deposition regimes of the central YS

the EAWM over the past 26 ka using the west-east SST in

mud area are most likely to remain relatively unchanged

the SCS. The results suggested that the EAWM was signifi-

over the past 7 ka due mainly to being supplied by the sus-

cantly related to the Atlantic meridional overturning circu-

pended sediments delivered with the EAWM-forced coastal

lation (AMOC), solar insolation and ice-sheet dynamics

currents in winter.

operating over different time scales [28]. In summary, these

In the current study, the grain size, sedimentary sequence,

studies in the SCS provided insights into the variations of

and AMS 14C dating were analyzed for three cores (ZY-1,

the EAWM in different time periods; however, their rela-

ZY-2, and ZY-3) located in the southern tip of the central

tively long records with low temporal resolution are insuffi-

YS mud area. We used the sensitive grain size groups as a

cient to decipher the centennial-scale history of EAWM

proxy of the EAWM to produce a high-resolution recon-

variability. Consequently, higher temporal resolution

struction of the history of the EAWM during the last 7.2 ka.

EAWM records are required to study the phases between

The results show that the evolutionary history of the

the EAWM and EASM and their influencing factors.

EAWM broadly followed the orbital-derived winter insolation

In recent years, progress has been made in using

with a similar long-term step-decreased trend as the EASM.

high-resolution grain size records as a proxy of EAWM,

However, the EAWM intensified events correlate well with

archived in the mud areas of the East China Sea (ECS) (e.g.,

the EASM weakening events within the dating error, most

the mud belt of the inner shelf of the ECS [29 32], the mud

likely due to the reduction of solar irradiance through changes

patch in the southwest of Cheju Island [33, 34]). However,

in the oceanic-atmospheric circulation patterns.

there is also debate regarding the reliability of this proxy.

Wang et al. [35] suggested that both the EAWM and EASM

would influence the grain size of the ECS inner shelf sedi-

1 Material and methods

ments. Therefore, one should first adequately determine the

sediment provenance and its associated transport mecha-

nism in the mud areas of the ECS when using sediment T hree gravity cores (ZY-1, ZY-2, and ZY-3) were

grain size as a proxy of the EAWM. recovered from the central YS mud area (Figure 1) by R/V

Previous studies have confirmed that the sediments in the Dong-Fang-Hong 2 in 2006. The water depths of these

1658 Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

Figure 1 Map of the study area showing the locations of the cores. YSWC: Yellow Sea Warm Current; LBCC: Lubei Coastal Current; SBCC: Subei

Coastal Current; KCC: Korean Coastal Current. The dashed lines indicate the Holocene mud thickness [41]. 0 40 are isopach (m).

cores are 70.3, 68.5 and 56 m, and core lengths are 395, 342 to dark gray clayey silt (Figures 2 4). The compositions of

and 378 cm, respectively. In the laboratory, these cores sand in these three cores are below 1%, although several

were split, described and subsampled in 1 cm intervals for layers exceed 2.5%. The mean sedimentation rates of cores

grain size analysis. ZY-1, ZY-2, and ZY-3 are 67, 56, and 51 cm/ka (Figure 5),

The grain size analyses were conducted using a Master- respectively. The grain size parameters, such as the mean

sizer-2000 laser particle size analyzer at the Key Laboratory (Mz), sorting (QD), kurtosis (Kg), and skewness (Sk), are

of Marine Sedimentology and Environmental Geology, First shown in Table 2.

Institute of Oceanography, State Oceanic Administration, (1) ZY-1. The compositions of silt and clay range from

Qingdao, with a measurement range of 0.02 2000 m and a 55% to 68% and from 31% to 44% with average values of

size resolution of 0.01 . The measuring error was within 60% and 40%, respectively. The Mz are coarser in the upper

3%. Before the grain size analyses, the samples were pre- (0 80 cm) and bottom (315 395 cm) parts, but finer in the

treated with 10% H2O2 and 1 mol/L HCl for 24 h to remove middle (80 15 cm) interval. The variability of QD, Sk and

organic matter and biogenic carbonate, respectively. Kg is small, with worse sorting, symmetrical distribution

An age model of the three cores was obtained by Accel- and high kurtosis (Figure 2).

erator Mass Spectrometry (AMS) 14C dating of mixed ben- (2) ZY-2. The silt content ranges from 60% to 72%

thic foraminifers at the AMS Laboratory of Peking Univer- with a mean of 65%. The clay content varies from 27% to

sity, China, and the Beta Analysis Co., USA (Table 1). The 39% with a mean of 35%. Two intervals can be distin-

radiocarbon ages were converted to calibrated calendar ages guished: the Mz of the upper part (0 262 cm) is finer than

following Hughen et al. [52] using the CALIB 5.0.2 pro- that in the bottom part (262 342 cm), and the variability of

gram with a marine reservoir age correction of 400 a. In this the grain size parameters in the bottom is also larger than

paper, all ages are expressed as calendar ages (Cal ka BP, that in the upper (Figure 3).

Before 1950 AD). (3) ZY-3. The silt content ranges from 57% to 78%

with a mean of 71%. The clay content varies from 22% to

43% with a mean of 29%. The Mz displays an upward fin-

2 Core lithology and sedimentation rate ing trend and a mud layer (15 cm thickness) occurs at 25 40

cm, where the QD abruptly increases and Sk and Kg de-

Three cores (ZY-1, ZY-2, and ZY-3) are composed of gray crease dramatically (Figure 4).

1659

Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

AMS 14C ages and calendar ages of Cores ZY-1, ZY-2 and ZY-3 (mixed benthic foraminifera) a)

Table 1

Depth (cm) Conventional 14C age error (a BP) 2 age range (cal a BP)

Cores Sample Calendar age (cal a BP) Code No.

ZY-1-1 45-150*-**-****-*** 115*-******

ZY-1-2 155 275*-**-**** 2337 261*-******

ZY-1

ZY-1-3 275 382*-**-**** 3643 388*-******

ZY-1-4 375 5270 -50-563*-****-**** 271485

ZY-2-1 68 1825 -30-136*-****-**** BA09250

ZY-2-2 129 234*-**-**** 1861 2089 BA081342

ZY-2 ZY-2-3 238 395*-**-**** 3831 4072 BA081343

ZY-2-4 273 445*-**-**** 4499 4792 BA081344

ZY-2-5 330 555*-**-**** 5857 6082 BA09252

ZY-3-1 26-267*-**-**** 2270 250*-******

ZY-3-2 65-348*-**-**** 3253 345*-******

ZY-3

ZY-3-3 168 486*-**-**** 5029 528*-******

ZY-3-4 359 650*-**-**** 6875 714*-******

a) The ZY-1 and ZY-3 samples were analyzed at Beta Analytic, USA, and the ZY-2 samples were analyzed at the State Key Laboratory of Nuclear

Physics and Technology, Peking University.

Figure 2 The lithology and vertical variations of the grain size parameters of ZY-1.

3 Sedimentary environment of the South Yel- ment of modern-type circulation in the South Yellow Sea

(SYS) [53]. The benthic foraminiferal, UK-37 SST, oxygen

low Sea

and carbon isotopic ( 18O and 13C) data from the adjacent

Core YE-2 also suggest three stages of paleoenvironmental

Previous studies show that a major faunal transition oc- development: a low salinity estuarine environment at

curred between 8.47 and 6.63 ka, indicating the establish- 8.4 6.9 ka, a low salinity shallow sea environment at 6.9 to

1660 Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

Figure 3 The lithology and vertical variations of the grain size parameters of ZY-2.

Figure 4 The lithology and vertical variations of the grain size parameters of ZY-3.

1661

Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

Figure 5 Comparison of the sedimentation rates of ZY-3 (a), ZY-2(b), ZY-1(c) and the averaged sedimentation rate (d).

The grain size parameters of ZY-1, ZY-2 and ZY-3 a)

Table 2 However, there is some debate on the sediment provenances,

formation times, and dynamic mechanisms of these mud

Mz Core Sorting Skewness Kurtosis

deposits in the SYS. Liu et al. [56, 58] suggested the top

ZY-1 7.27 7.85 1.30 1.69 0.02 0.12 1.04 1.12

fine sediments of Cores YA131 and YSDP102 were the

(7.66) (1.45) (0.07) (1.08)

ZY-2 7.03 7.67 1.33 1.77 0.01 0.13 0.98 1.15

deposited products after 6 7 ka, when the sea level ap-

(7.42) (1.48) (0.06) (1.10)

proximated the Holocene Highstand, whereas Zhuang et al.

ZY-3 6.84 7.76 1.35 1.59 0.04 0.15 1.02 1.12

[57] indicated that the top fine sediments of Core EY02-2

(7.21) (1.48) (0.10) (1.07)

have developed since 9.7 ka.

a) The value in brackets is the averaged value.

The lithology of Cores ZY-1, ZY-2 and ZY-3 is clayey

silt, a typical fine-grained deposit. The high-resolution

approximately 6 5 ka, and a modern marine environment shallow seismic profile [41] shows that these cores do not

after approximately 6 5 ka [54, 55]. During the course of penetrate into the acoustic reflective surface (maximum

paleoenvironmental development in the SYS, mud deposi- flooding surface, MFS) (Figure 6), indicating that the mud

tion widely occurred in this shallow sea, such as in the up- deposits of the SYS began to accumulate before at least 7 ka.

per part of Cores YA131 [56] and CC02 [53] in the northern As shown by the AMS 14C dating data, the average sedi-

SYS, of Cores EY02-2 [57] and YE-2 [55] in the central mentation rates of these three cores (50 70 cm/ka) slightly

SYS, and of Core YSDP102 [58] in the northeastern SYS. decreased from the west to east with the maximum sedi-

These mud deposits are considered to have formed under mentation rate up to 100 cm/ka (Figure 5). The average

conditions of small-amplitude change in sea level, a stable sedimentation rates are approximated to that of the adjacent

sedimentary environment and continued sediment supply. Core YE-2 [55], but they are much higher than those of

1662 Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

Cores CC02 (16 cm/ka, since 5.72 ka) [53] and EY02-2 could mostly represent the influences of suspended sedi-

(13.3 cm/ka, since 7.9 ka) [57]. In addition, all the grain size ments delivered by the coastal current in winter. PC2 likely

frequency curves of these three cores are unimodal (Figures reflects the sedimentary processes in summer, when the

2 4). These results suggest a sufficient sediment supply and study area is characterized by weaker hydrodynamic condi-

stable sedimentary environment in the study area during the tions and sediment supply starvation. In the SYS, the ocean

last 7 ka. hydrodynamics are dominated by the EAWM in winter, and

Spatially, although Cores ZY-1, ZY-2 and ZY-3 are lo- thus, the stronger EAWM would lead to more coarse sedi-

cated at the southern tip of the omega-shaped mud ments being resuspended and an increase for the SSC.

deposits of YS (Figure 1), Cores ZY-1 and ZY-2 are on the Meanwhile, the stronger EAWM would also strengthen the

bottomset, whereas Core ZY-3 is on the foreset (Figure 6). coastal current, resulting in the transport of coarser sedi-

This difference led to a great discrepancy between the sedi- ment fractions to the study area. Thus, the PC1 loading can

mentation rates of the top part of these three cores. The top be used to indicate the variations of the EAWM. In Cores

part of Core ZY-3 shows a minimum sedimentation rate of ZY-2 and ZY-3, the coarse fractions are negative loadings

only 11 cm/ka. The mean grain size of these three cores in PC1, and therefore, the high (low) values denote a weak

decreased eastward likely associated with the sample sta- (strong) EAWM. The situation is reversed in ZY-1, where

tions gradually becoming closer to the center of the cyclonic its coarse fractions are a positive loading in PC1. We also

circulation of the SYS [59]. However, the temporal varia- extracted the environmentally sensitive grain size compo-

tions of grain size parameters of these three cores are basi- nent of these three cores following the method of Xiang et

cally identical to each other, suggesting that their sedimen- al. [33]. The results gave a similar variation trend to that of

tary environment may be similar over the past 7 ka. Con- PCA. In this study, the PC1 loadings of PCA were used as a

sidering the relatively stable sea level since 7 ka and that the proxy of the EAWM.

sample stations are away from the river mouths (e.g., Yel- As shown in Figure 7, the PC1 loadings display a nearly

low River and Yangtze River), we inferred that the temporal identical variation trend; however, the fluctuation points of

variations of grain size of these three cores are controlled ZY-1 are always prior to the other two cores by ca. 200

years, likely caused by the AMS 14C dating error. Generally,

mainly by the suspended sediments delivered by the coastal

the EAWM displayed fluctuation changes during the last

current in winter.

7.2 ka, and it can be divided into three periods: strong and

4 Proxy of EAWM

Table 3 The loading of PC1 and PC2 of 45 grain size groups in Core

ZY-2

Principal component analysis (PCA), a multivariate analyti-

cal tool, is widely used to extract important information Grain size ( m) PC1 PC2 Grain size ( m) PC1 PC2

from a large data matrix to represent it as several new or- 0.11 0.04

0.34 0.50 5.71 0.92

thogonal variables called principal components (PC). PCA 0.17 0.19

0.39 0.93 6.45 0.94

with varimax rotation was performed on the grain-size data 0.35

0.44 0.04 0.98 7.29 0.92

of Cores ZY-1, ZY-2, and ZY-3 using the program SPSS 0.51

0.50 0.05 0.98 8.23 0.84

13.0 for Windows (SPSS Inc., Chicago, Illinois). Taking 0.65

0.56 0.09 0.98 9.30 0.68

Core ZY-2 as an example, the grain-size of ZY-2, varying 0.73

0.63 0.13 0.97 10.51 0.45

from 0.3 74.1 m, was divided into 45 groups by logarith- 0.74

0.72 0.15 0.97 11.87 0.18

mic distribution. After performing PCA on the grain size 0.08 0.70

0.81 0.18 0.96 13.42

datasets, the scree plot of eigenvalues indicated that 81% of 0.29 0.63

0.91 0.21 0.95 15.16

the data variance could be explained by the first two com- 0.47 0.56

1.03 0.25 0.93 17.12

ponents (42% of PC1 and 39% of PC2) (Table 3). The PC1 0.61 0.48

1.17 0.29 0.90 19.35

consisted of two groups (2.7 9.3 and 21.9 74.1 m), of 0.72 0.42

1.32 0.33 0.87 21.86

which the coarse part (21.9 74.1 m) gave a high negative 0.80 0.36

1.49 0.38 0.83 24.70

loading on the PC1 and the fine part (2.7 9.3 m) was the 0.87 0.31

1.68 0.42 0.79 27.90

positive loading. The PC2 is similar to PC1, being consti- 0.92 0.26

1.90 0.47 0.74 31.53

tuted by two groups of 0.3 2.4 m (positive loading) and 0.95 0.23

2.15 0.52 0.69 35.62

10.5 19.4 m (negative loading). The PCA results of ZY-1 0.97 0.19

2.43 0.57 0.63 40.24

and ZY-3 are shown in Tables 4 and 5, respectively. 0.98 0.17

2.74 0.62 0.56 45.47

By comparing the components of PC1 and PC2, it was 0.99 0.15

3.10 0.67 0.49 51.37

found that the grain sizes of the PC1 positive and negative 0.98 0.13

3.50 0.73 0.41 58.04

loading were coarser than those of PC2. Given the seasonal 0.97 0.12

3.96 0.78 0.31 65.58

differences in the sedimentary environment of the SYS 0.94 0.12

4.47 0.83 0.21 74.09

(e.g., sediment supplies and ocean hydrodynamics), PC1 5.05 0.88 0.09

1663

Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

Figure 6 The high-resolution shallow seismic profile crossed Cores ZY-1, ZY-2, and ZY-3.

Table 4 The loading of PC1 and PC2 of 45 grain size groups in Core Table 5 The loading of PC1 and PC2 of 45 grain size groups in Core

ZY-1 ZY-3

Grain size Grain size ( m) PC1 PC2 Grain size ( m) PC1 PC2

Grain size ( m) PC1 PC2 PC1 PC2

( m)

0.61

0.21 0.00 0.78 3.50 0.75

0.18 0.95 0.29

0.34 0.76 5.71

0.86

0.24 0.15 0.97 3.96 0.43

0.18 0.88 0.46

0.39 0.77 6.45

0.96

0.27 0.14 0.97 4.47 0.03

0.18 0.76 0.63

0.44 0.84 7.29

0.30 0.93

0.30 0.15 0.97 5.05

0.16 0.59 0.78

0.50 0.91 8.23

0.52 0.83

0.34 0.18 0.97 5.71

0.13 0.36 0.90

0.56 0.95 9.30

0.66 0.74

0.39 0.20 0.97 6.45

0.10 0.11 0.94

0.63 0.97 10.51

0.75 0.65

0.44 0.23 0.96 7.29

0.06 0.93

0.72 0.99 11.87 0.13

0.81 0.58

0.50 0.27 0.96 8.23

0.03 0.88

0.81 0.99 13.42 0.33

0.86 0.51

0.56 0.30 0.95 9.30

0.01 0.81

0.91 0.98 15.16 0.49

0.88 0.46

0.63 0.33 0.94 10.51

0.01 0.74

1.03 0.98 17.12 0.61

0.90 0.41

0.72 0.37 0.93 11.87

0.02 0.66

1.17 0.97 19.35 0.71

0.91 0.37

0.81 0.42 0.91 13.42

0.06 0.58

1.32 0.97 21.86 0.79

0.91 0.33

0.91 0.46 0.88 15.16

0.12 0.50

1.49 0.97 24.70 0.85

0.89 0.29

1.03 0.52 0.85 17.12

0.21 0.42

1.68 0.95 27.90 0.89 0.85 0.24

1.17 0.58 0.81 19.35

0.33 0.35

1.90 0.92 31.53 0.93 0.75 0.14

1.32 0.65 0.75 21.86

0.46 0.28 0.52

2.15 0.87 35.62 0.95 1.49 0.72 0.68 24.70 0.01

0.30

1.68 0.80 0.59 27.90 0.09

0.61 0.21

2.43 0.77 40.24 0.96

0.15

1.90 0.87 0.47 31.53 0.05

0.74 0.14

2.74 0.64 45.47 0.96

0.07

2.15 0.94 0.32 35.62 0.00

0.84 0.07

3.10 0.48 51.37 0.95

0.08 0.02

2.43 0.98 0.15 40.24

0.92 0.01

3.50 0.31 58.04 0.92

0.06 0.10 0.04

2.74 0.98 45.47

0.96

3.96 0.16 65.58 0.87 0.06

0.32

3.10 0.93

0.98

4.47 0.01 74.09 0.80 0.12

0.97 0.14

5.05

pressure gradients between the two air masses seasonally

highly fluctuating during 7.2 4.2 ka, moderate and rela- changing its direction (Figure 8(a)). Previous studies have

tively stable during 4.2 1.8 ka, and weakened after 1.8 ka. demonstrated that the intensity of the EASM predominantly

Centennial-scale EAWM intensified events are superim- oscillates at the precession cycle (23 ka) regulated by boreal

posed on the long-term EAWM variation trend, especially summer insolation, whereas the 100 ka rhythm (eccentricity

during the periods 7.2 4.2 ka (e.g., 7.2 6.9, 5.9 5.5, cycle) dominates the EASM intensity only when the solar

5.0 4.7 and 4.4 4.2 ka). After 4.2 ka, only small EAWM forcing reaches a low-level threshold [8, 61]. As the

intensified events were recorded at 3.1 3.2 and 2.4 2.6 ka. Lau-rentide ice sheet melted, the ice volumes in North

The weakest EAWM period of 1.1 1.4 ka agrees well with America stabilized and reached to its modern-level since the

the Sui-Tang Warming Period (600 900 a AD) [60]. Middle Holocene (7 8 ka) [63]. Thus, the influence of the

northern hemisphere ice sheet on the EAM in this period

can be neglected.

5 Comparison with the EASM since the Middle The boreal summer (winter) solar insolation (30 N) has

Holocene: Millennial-scale trend vs. centennial- decreased (increased) gradually since the Middle Holocene

scale events [62] (Figure 8(b)), resulting in the summer (winter) tempera-

tures of the Eurasian continent and the adjacent ocean de-

Essentially, the seasonal variability of the EAM is the result creasing (increasing) together. However, the temperature

of thermal differences between the Eurasian continent and variability is greater on the continent than in the oceans due

the Pacific Ocean, which leads to the temperature and air to the larger thermal capacity of the ocean (Figure 8(a)).

1664 Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

Figure 7 The evolutionary history of the East Asian Winter Monsoon recorded by the PC1 loadings of ZY-1, ZY-2, and ZY-3 during the last 7.2 ka.

previous studies [9]. Here, our grain size records demon-

strate that the intensity of the EAWM also displays a

long-term step-decreased trend like that of the EASM dur-

ing the last 7.2 ka (Figure 9(a) and (b)), which is consistent

with the theory discussed above.

Seasonally opposing insolation changes since the Middle

Holocene, as mentioned above, played a dominant role in

the temporal evolution of land-ocean surface temperatures,

which have been discussed using climate simulations [27,

66]. A new pollen-based reconstruction of latitudinal tem-

perature gradient (LTG) covaried with the latitudinal insola-

tion gradient (LIG) during the Middle Holocene, as both

exhibited a gradually decreasing trend in summer and win-

ter [67]. Coupled records of Sr/Ca and 18O of coral skele-

tons from the subtropical northwestern Pacific also indicat-

ed that the EASM and EAWM were more intense in the

Middle Holocene than the present day [68]. Our results

show that both the EASM and the EAWM display a similar

long-term step decrease since the Middle Holocene, result-

ing from the orbital forcing seasonal variation of solar inso-

lation. This in-phase variation of EASM and EAWM during

the last 7.2 ka supports the views of Zhou and Zhao [27]

and Steinke et al. [25, 26].

However, the EAWM intensified events are correlated

well with the EASM weakened events [9], the North

Atlantic climatic variations (Bond events 0 to 5) [64] and

Figure 8 Seasonal atmospheric circulations of the East Asian Monsoon reduced solar irradiance [65] at the centennial scale (Figure

and its influencing factors. (a) Measured annual temperature profile from

9). Although the changes in solar irradiance are small

inland Eurasia and the Pacific [61]; (b) seasonal variations of the 30 N

(0.1% 0.3%), the Earth s climate system is highly sensitive

solar insolation [62]. Tw and Ts are winter and summer temperature gradi-

to the slight changes in the Sun's energy output, which may

ent between the Eurasian continent and the Pacific Ocean, respectively.

be amplified via the atmosphere and/or oceans [69, 70]. The

affinity between the solar irradiance and climate change has

Consequently, the summer (winter) temperature and air been demonstrated by previous studies [9, 11, 64, 71 73].

pressure gradients between the two should both decrease Xiao et al. [32] indicated that a close relationship existed

during this period. The intensity of the Holocene EASM between the solar irradiance and the EAWM during the last

broadly following summer insolation has been confirmed by 7.6 ka. They speculated that the reduction of solar irradi-

1665

Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

Figure 9 The variation trend of the East Asian Winter Monsoon (a), stalagmite 18O records from the Dongge Cave [9] (b), the Holocene record of drift ice

(hematite-stained grains, HSG) in the North Atlantic [64] (c) and variations of the total solar irradiance [65] (d). The numbers in (c) indicate the ice-rafted

debris (IRD) events of the North Atlantic.

ance leads to changes in atmospheric or oceanic circulation, differences, and the decreased intensity of the AMOC led to

which could cause an increasing land-ocean temperature an enhanced EAWM in winter but a weakened EASM in

difference during winter, resulting in a strengthening of the summer [83, 84]. Hong et al. [7] suggested that the

EAWM [32]. The TOC and grain-size records from the melt-water outburst and reduced solar irradiance appeared

Kusai Lake, northern Qinghai-Tibetan Plateau, also suggest to modulate the Earth system changes in the same direction

three anti-variations of EASM and EAWM events during because both would have triggered changes in the mean

the last 3.8 ka, which are well correlated with reduced solar state of the equatorial Pacific Ocean and ENSO may have

irradiance [74]. Our results indicate that these EAWM in- acted as a mediator between the Sun and the Earth s climate

tensified events are likely forced by the centennial-scale [73, 85]. Combined with our results, we hypothesize that the

changes in solar irradiance since the Middle Holocene, as inverse behavior of EAWM and EASM at the centennial

discussed below. scale is related to the reduced solar irradiance, which caused

The Atlantic Meridional Overturning Circulation the intensity of AMOC to decrease due to changes in the

(AMOC) is a switch to the global thermohaline circula- oceanic-atmospheric circulation patterns. The underlying

tion (THC) and thus plays an important role in abrupt cli- mechanism of these intensified EAWM events may resem-

mate changes [75, 76]. At the millennial scale, the stopping ble that of millennial abrupt cooling events, but with smaller

or slackening of AMOC is likely responsible for the abrupt amplitude.

global cooling events (e.g., Heinrich, Younger Dryas, 8.2

ka); it also significantly influenced the EAM, resulting in a

6 Conclusions

declined EASM [77, 78] but intensified EAWM [17, 18, 25,

28, 79] during these cooling events. Model simulations in-

Three cores (ZY-1, ZY-2, and ZY-3) obtained from the

dicated that changes in the AMOC can also have important

central mud area of the SYS consisted of clayey silt with

impacts on the EAM at decadal [80, 81] and centennial

average sedimentation rates of 50 70 cm/ka. These fine

scales [82]. Furthermore, the influence of the AMOC on the

sediment deposits began to develop when the sea level

Asian monsoonal climate displayed significant seasonal

1666 Hu B Q, et al. Sci China Earth Sci October (2012) Vol.55 No.10

reached the Holocene highstand. The vertical variations of 11 Zhang P Z, Cheng H, Eewards R L, et al. A test of climate, sun, and

culture relationships from an 1810-year Chinese cave record. Science,

grain-size parameters of these three cores are basically

2008, 322: 940 942

identical to each other, mainly controlled by the suspended 12 Thompson L G, Yao S T, Davis M E, et al. Tropical climate

sediments delivered by the coastal current in winter, which instability: The Last Glacial Cycle from a Qinghai-Tibetan ice core.

Science, 1997, 276: 1821 1825

is dominated by the intensity of the EAWM. PCA was used

Contact this candidate