Plant It

Geology August **** Vol.** No.**: **** ****

doi: 10.1007/s11434-010-3174-8

SPECIAL TOPICS:

Pollen-inferred vegetation and environmental changes since

16.7 ka BP at Balikun Lake, Xinjiang

TAO ShiChen1, AN ChenBang1*, CHEN FaHu1, TANG LingYu2, WANG ZongLi1, L YanBin1,

LI ZhiFei1, ZHENG TongMing1 & ZHAO JiaJu1

1

MOR Key Laboratory of West China s Environmental System, Lanzhou University, Lanzhou 730000, China;

2

Nanjing Institute of Geology and Paleontology, Chinese Academy of Sciences, Nanjing 210008, China

Received June 18, 2009; accepted October 9, 2009

A high-resolution fossil pollen record from the sedimentary cores of Balikun Lake, northwestern China, combined with modern

surface pollen data, is used to reconstruct the history of vegetation and climatic change since 16.7 cal. ka BP. Fossil pollen assem-

blages and lithology indicate that the study area was dominated by desert. The desert had extremely arid climate and lower effec-

tive moisture during 16.7 7.9 cal. ka BP, especially from 16.7 to 8.9 cal. ka BP when the lake maybe dried up. During 8.9 7.9 cal.

ka BP, the environment gradually recovered in this area. It was then followed by the optimum period from 7.9 to 4.3 cal. Ka BP,

when the effective moisture obviously increased. It was characterized by the typical desert-steppe/steppe vegetation and was ac-

companied with several patch-birch woodlands around the lake. After that, a short but extremely arid climatic event occurred

during 4.3 3.8 cal. ka BP, and the vegetation quickly changed from desert-steppe/steppe to desert. It was a relatively optimum

period from 3.8 to 0.53 cal. ka BP showing typical desert-steppe/meadow-steppe landscape. Since 0.53 cal. ka BP, the climate has

shown signs of deteriorating again. Furthermore, regional comparison shows that the characteristics of climatic and environmental

evolution in this area were clearly different from East Asia monsoonal area during the last 16.7 cal. ka BP. It was characterized by

the arid climate during the late-glacial and early Holocene, and relatively wet during the mid-late Holocene.

Xinjiang, Balikun Lake, Holocene, pollen, vegetation, climate

Citation: Tao S C, An C B, Chen F H, et al. Pollen-inferred vegetation and environmental changes since 16.7 ka BP at Balikun Lake, Xinjiang. Chinese Sci Bull,

2010, 55: 2449 2457, doi: 10.1007/s11434-010-3174-8

A critical problem people face today is climatic change. of the climate change were relatively complicated during the

Studies of Holocene climate change in several sensitive Holocene. The oxygen isotope of carbonate and pollen data

regions have provided insights into future global warming. from Manas Lake indicated that Holocene climatic changes

Right now in East and South Asia, Asian summer monsoon were similar to those in east China [14 17]. Pollen records

variations have been well-documented by precisely dated from Lake Kendegelukol and Lake Tashkol in the Altay

speleothems [1 5]. The Asian monsoon was enhanced at the Mountains [18] also suggest that the climate was quite humid

onset of the Holocene. It was strongest in the early and from the early Holocene to middle Holocene. However, re-

mid-Holocene, but weakened after the mid-Holocene. cent studies [19 22] indicated that the climate was ex-

Similar changes in the strength of the Asian monsoons have tremely dry during the late-glacial and early Holocene pe-

been documented by other proxies records from peats [6,7], riod. The mid-Holocene was humid and relatively dry dur-

lake sediments [8 11], and marine sediments [12,13]. In ing the late Holocene, suggesting the climate was mainly con-

contrast, in arid and semiarid region of Xinjiang, the records trolled by westerly. All of these suggest that the patterns of the

climate change are still inconsistent in this region during the

Holocene. Thus, in order to fully understand the regional

*Corresponding author (email: ****@***.***.**)

Science China Press and Springer-Verlag Berlin Heidelberg 2010 csb.scichina.com www.springerlink.com

2450 TAO ShiChen, et al. Chinese Sci Bull August (2010) Vol.55 No.22

climate change and its mechanism, more fine-resolution cli- 1 Study area and site

matic records from the Xinjiang region are needed.

Balikun Lake is one of ideal study sites. Previous study

Balikun Lake (43 36 43 45 N, 92 42 92 54 E) is located

indicated that about 62 ka BP the stream outlets ended and 15 km west of Balikun County (Figure 1), eastern part of

formed a closed lake [23]. Therefore, the temperature, pre- the Tianshan Mountains, Xinjiang, China. The Lake Basin

cipitation and biology information of the regional climatic is between the Balikun Mountain and Meiqinwula Mountain,

environments could be well-preserved in the lake sediment north of the Tianshan Mountains (Figure 1). It covered a

since the lake was closed. Over the last decades, a series of maximum area of more than 800 km2, and decreased to 500

studies have been carried out on different geological time km2 since the early Quaternary [28]. By 1944, the area of

scales with different temporal-resolutions using multi-proxies water surface was 140 km2 with an elevation of 1585 m a.s.l.

records (such as pollen, stable isotopes, carbonates, geo- Then it shrank to 112.5 km2 with an elevation of 1581 m

chemistry) in the Lake [23 28], which provide valuable a.s.l. The average water depth changed from 2.5 m in the

insights into understanding the Holocene climate and vege- 1980s [29] to about 0.7 0.9 m with the elevation of 1575 m

tation history of this region. However, most of these studies a.s.l. at present. The maximum surface area is about 98.5

usually used relatively lower temporal-resolution as a result km2 with the watershed catchments of 4500 km2. The river

of the limitation of the experimental conditions at that time. rises from the Tianshan Mountains, and it flows from

In this paper, we present high-resolution pollen record com- southeast to northwest through Balikun grassland (Figure 1).

bined with the modern surface data, to reconstruct detailed Due to both dams built on the upstream of Dahe River and

evolutional history of climate and vegetation over the past agricultural activities, the lake level decreases sharply with

16.7 cal. ka BP, especially during the Holocene. Regional abundant Na2SO4 10H2O and saline deposits of halite in-

comparison is conducted to investigate the climatic change cluding minor amounts of gypsum and exposed mudflat, so

patterns of this region during the late-glacial and Holocene. the lake does not freeze in winter.

Figure 1 The location of study area and site. The black dot shows the location of the coring site (BLK06E). Insert map shows the location of the study site

in China [31].

2451

TAO ShiChen, et al. Chinese Sci Bull August (2010) Vol.55 No.22

This region is located in the northern temperate zone and

characterized by continental dry climate. The annual tem-

perature is 1.3 C with a July average of 17.1 C and a Janu-

ary average of 18.1 C. Mean annual precipitation is 210.9

mm, 54% of which falls in June-August with a peak of 42.2

mm in July. Mean annual evaporation reaches 1602.7 mm

[30]. The composition of vegetation is relatively simple

because of the dry climate in the Balikun Basin (Figure 1).

The regional vegetation changes from alpine-meadow

(mainly dominated by Kobresia capilliformis between 2900

and 2800 m a.s.l.) to coniferous forest (consisting of Larix

sibiria and Picea schrenkiana between 2900 and 2100 m

a.s.l.), then to desert-steppe (consisting of Stipa glareosa,

Festuca sulcata, Allium polyrrhizum with several desert

shrub) in the piedmont. Around the lake, the vegetation is

mainly dominated by Achnatherum splendens, Aneuro-

lepidium, Reaumuria soongorica, Salsola abrotenoides and Figure 2 The lithology, sediment rate and the age-depth model of core

Kalidium schrenkianum [32]. BLK06E at Balikun Lake, northwest China.

In March 2006, several consecutive cores were drilled

from Balikun Lake (Figure 1), including an 862.8 cm-long In the arid and semiarid region, the reservoir effect of

core (BLK06E). The upper 215 cm was collected by dig- lake is frequently the most important factor limiting accu-

ging an exploration trench and the low 647.8 cm was drilled rate carbon dating. Two different methods were used to

using a Kullenburg-type piston corer from the bottom of the evaluate the reservoir effect in Balikun Lake. Firstly, the

trench. In the laboratory, the cores were subsampled at 1 cm TPM usually is not influenced by the reservoir effect ac-

intervals and freeze-dried by the Thermo Savant s Lyopump cording to recent study [19]. Therefore, we use the 14C age

system. In this paper, we focus on the upper 446 cm. In of TPM to calibrate the carbon reservoir effect . The age

August 2008, we collected the modern samples as shown in of the TPM at 69 cm is 1945 50 a BP (Table 1); while the

Figure 1. Finally, we analyzed the relationship between the age of the organic matter at a depth of 69 cm interpolated

modern pollen assembleage and the vegetation types, in from the two adjacent dated samples, depths of 52 cm and

order to explain the climate meaning of the core sample. 71 cm (of age 2309 50 and 2785 48 a BP, respectively),

is 2735 years, which is 790 years older than the age of the

TPM. Secondly, in order to further calibrate the reservoir

2 Lithology, chronology and methods effect of the organic matter 14C age, we extracted pollen

for AMS 14C [33,34]. The AMS 14C age of the pollen at 232

2.1 Lithology

cm is 7020 40 a BP, and the age of the organic matter at

233 cm is 7810 50 a BP, which is also 790 years older

The stratigraphical descriptions of top 446 cm of the core

than the age of the pollen. Because of both methods (TPM

are as follows (Figure 2):

and pollen) are very similar, it is reasonable that 790 years

0 47 cm: White clay and silty clay including a few Mir-

were taken as the reservoir effect in this lake. Previous

abilite crystals (Na2SO4 10H2O); the sediment is relatively

study [35] showed a reservoir effect of 750 years, very

loose.

similar to our results. Similar results also appeared at other

47 72 cm: White-silt and silty clay, also with Mirabilite

study sites [19,20,22] in the Xinjiang area. Accordingly,

crystals.

after subtracting 790 years from organic matters dated 14C

72 244 cm: Blackish clay; the sediment emit effluvium.

ages, all reservoir corrected 14C ages were converted to cal-

244 336 cm: Darkish clay.

endar year ages using the CALIB 5.0 program [36], linearly

336 446 cm: Greenish-fine sand, with gravel.

interpolated and extrapolated by assuming a uniform sedi-

mentation rate between two paired ages. The linear sedi-

14

C dating and chronology

2.2

mentation rates thus obtained vary significantly within the

Radiocarbon dating was conducted using AMS14C and core, with high values of 1.4 mm per year between 233 cm

conventional 14C methods to date 13 samples, consisting of and 325 cm and relatively low values of 0.13 mm per year

from 142 to 166 cm (Figure 2).

organic matter, pollen and terrestrial plant macrofossil

(TPM) (Table 1). All samples were prepared with the stan-

dard pretreatment (alkali-acid-alkali). The 14C and calibrated 2.3 Pollen analysis method

calendar ages are shown in Table 1, while lithology and the

238 pollen samples were analyzed, including 198 core sam-

chronostratigraphy of the core are shown in Figure 2.

2452 TAO ShiChen, et al. Chinese Sci Bull August (2010) Vol.55 No.22

14

Table 1 C dates of BLK06E core

13C Calibrated 14C age (2 ) (cal. a BP)

14

Laboratory number Depth in core (cm) Dating material C age (a BP)

organic matter a)

LUG06-61 23-176*-**-***-***

a)

LUG06-62 52 organic matter 230*-**-**** 1521

plant macrofossils b)

LAMS07-008 69 37.98 194*-**-**** 1999

organic matter a)

LUG06-63 71-278*-**-**** 2062

organic matter a)

LUG06-64 76 15.92 304*-**-**** 2349

organic matter a)

LUG06-65 82-307*-**-**** 2364

organic matter a)

LUG06-66 142 455*-**-**** 4360

b)

LAMS07-005 166 organic matter 28.17 609*-**-**** 6208

organic matter b)

LAMS07-006 215 29.51 744*-**-**** 7594

pollen b)

LAMS09-11 232 27.97 702*-**-**** 7945

organic matter b)

LAMS07-002 233 25.89 781*-**-**** 7953

organic matter b)

LAMS07-003 325 25.85 852*-**-**** 8592

c)

OZL036 446 organic matter 21.40 147**-**-***** 17017

a) Chronology laboratory of Lanzhou University China; b) the Radiocarbon Dating Laboratory of Beijing University, China; c) ANSTO.

ples and 40 surface samples. The subsamples were processed on the total terrestrial pollen sum.

with the standard method including HCl, NaOH, HF, and

Acetolysis treatments. Before the chemical procedure, tablets

3 Pollen records, paleovegetation and paleocli-

containing a known quantity of Lycopodium spores (batch #

mate

938934) were added as exotic markers into the samples to

determine pollen concentration. Approximately 1g sample

3.1 Modern pollen assemblages

was digested with 10% HCl, 10% KOH, 40% HF, then fil-

tered with 7 m sieve mesh. Each pollen sample was The modern pollen assemblages are shown in Figure 3. Ac-

counted under a light microscope at 400 magnifications in cording to its origin, we divide the diagram into seven types:

regularly spaced traverses and each sample was identified Lake surface, desert, reed bush, steppe, meadow, alpine

more than two slices. Except for several samples, more than meadow and conifer forest.

300 and 500 terrestrial plant pollen grains (Figures 3 and 4) Chenopodiaceae (34.6% 42.7%) and Artemisia (35.7%

51.8%) pollen are the dominant components in the Lake s

were counted for the modern surface samples and core sam-

surface samples, which also includes the main pollen types

ples, respectively. All pollen and spores were identified to

of the Betula (0.7% 2.7%), Compositae (0.8% 2.6%), Gra-

genus or family level. Identifications of pollen followed

mineae (5.16% 5.98%), Cruciferae (0.2% 1.6%), Cyper-

Wang et al. [37], Xi and Ning [38] aided by modern refer-

aceae (0.6% 4.3%), Ephedra (3.4% 7.8%) and Tamari-

ence collection. Pollen percentages were calculated based

Figure 3 The vegetation types and pollen percentage diagram of selected pollen types of the modern samples in Balikun region.

2453

TAO ShiChen, et al. Chinese Sci Bull August (2010) Vol.55 No.22

caceae (0 1.4%). mation of climatic and environmental evolution.

The pollen assemblages of the desert samples are pre-

dominated by Chenopodiaceae (32.2% 95.6%), with an 3.2 Fossil pollen assemblages

average value of 74.3%. The percentage of the Gramineae

About 49 genera (families) of the pollen types are identified

pollen (39.6% 69.4%) is apparently higher than that of

in the samples of BLK06E core. The arboreal pollen mainly

other types, and the size of most Gramineae pollen are less

contains Pinus, Picea, Cupressaceae, Ulmus, Carpinus,

than 30 m in the samples collected from the local reed

Betula and Salix; the major shrub and semi-shrub pollen

bush. The pollen types of Artemisia (7.7% 43.4%), Com-

include Ephedra, Chenopodiaceae, Nitraria, Tamaricaceae,

positae (1.4% 22.8%), Gramineae (0.8% 16.8%), Cyper-

Umbelliferae and Zygophyllaceae. The major herb pollen

aceae (1.3% 53.6%) are the major elements in samples col-

types are Artemisia, Compositae, Gramineae, Leguminosae,

lected from the local steppe and meadow. Our results show

Cyperaceae, Ranunculaceae, Thalictrum and Caryophylla-

that the pollen assemblages have good consistence with its

ceae. There are also several aquatic plants pollen, such as

vegetation types. As a whole, the pollen types are predomi-

Sparganium, Typha and Myriophyllum. All the major types

nated by steppe, shrub and semi-shrub in the non-coniferous

are shown in pollen diagram (Figure 4). However, the per-

forest area, and the arboreal pollen types are sporadic.

centages of Chenopodiaceae (up to 94%), Artemisia (up to

The pollen types of the Pinus (5.8% 11.1%), Picea

45%), Gramineae (up to 23%) and Betula are predominated

(41.1% 56.7%), Abies (4.5% 7.7%) and Larix (1.2% 7.3%)

in whole pollen assemblages (Figure 4).

are the major elements in the samples of the conifer-forest.

The percentage pollen diagram was divided into five

It is worth noting that the Betula pollen is essentially spo- pollen assemblage zones, with subzones when necessary,

radic in the whole modern samples, and that the pollen per- based on stratigraphically constrained cluster analysis

haps is transported by wind from the other area. (CONISS) [40].

Figure 3 shows that the pollen assemblages mainly re- (1) Zone BLK-5 (446 244 cm; 16.7 7.9 cal. ka BP).

flect the local vegetation in the samples collected from the Chenopodiaceae pollen was the major element in this pollen

single vegetation types, and pollen assemblages from the zone (Figure 4). According to the composition of the pollen

surface of lake sediments are more uniform, indicating that percentage, we can divide this zone into two subzones.

the mixture of lake sediments is more evenly, and its pollen BLK-5a (446 336 cm; 16.7 8.9 cal. ka BP). In this sub-

assemblages should represent the comprehensive informa- zone, the pollen types were very poor, accompanied by al-

most the lowest pollen concentration ((0.04 14.7) 104

tion of the regional vegetation [39]. So the data of pollen

assemblages of the lake sediment could well reflect the grains/g). The main component was Chenopodiaceae pollen

composition of vegetation of the study region in the geo- whose percentages varied from 58% to 94%. The percent-

age of Artemisia pollen was relatively low (0.2% 12.6%).

logical period, and faithfully record the integrated infor-

Figure 4 Percentage pollen diagram from core BLK-06 of Balikun Lake, northwestern China. Only selected taxa are shown.

2454 TAO ShiChen, et al. Chinese Sci Bull August (2010) Vol.55 No.22

lected from the core sediment, we find that most of them are

Gramineae pollen declined from 18.3% to 0.3%. The Ar-

less than 35 m. However, in the modern samples collected

temisia/Chenopodiaceae (A/C) ratio was very low, with an

average value of

Contact this candidate