全 文 ：Received 20 Aug. 2003 Accepted 12 Dec. 2003
Supported by the Knowledge Innovation Program of The Chinese Academy of Sciences (KZCX1-10-05), the National Natural Science
Foundation of China (90102009), the State Key Basic Research and Development Plan of China (G1999043502) and the National Postdoctor
Science Foundation of China (2003033253).
* Author for correspondence. Tel: +86 (0)10 62591431 ext. 6273; E-mail: .
http://www.chineseplantscience.com
Acta Botanica Sinica
植 物 学 报 2004, 46 (6): 655-667
Vegetation Changes and Environmental Evolution in the Urumqi River
Head, Central Tianshan Mountains Since 3.6 ka BP:
a Case Study of Daxigou Profile
ZHANG Yun1, KONG Zhao-Chen1, YANG Zhen-Jing1, 2, YAN Shun3, NI Jian1, 4*
(1. Laboratory of Quantitative Vegetation Ecology, Institute of Botany, The Chinese Academy of Sciences, Beijing 100093, China;
2. Institute of Hydrologic and Environmental Geology, The Chinese Academy of Geological Sciences, Zhengding 050803, China;
3. Xinjiang Institute of Ecology and Geography, The Chinese of Academy of Sciences, Urumqi 830011, China;
4. Max Planck Institute for Biogeochemistry, Jena 07701, Germany)
Abstract: A relatively high resolution pollen record and data of loss of ignition (LOI), grain size and
susceptibility of the Daxigou profile in the head area of the Urumqi River, central Tianshan Mountains,
revealed new information about vegetation changes and environmental evolution since 3.6 ka BP. Results
showed that from 3.6 ka BP to present, climate was unstable with multi-changes of warming-cooling and
wetting-drying. From ca. 3.6 to 3.2 ka BP, climate was warmer and more humid than today. Climate changed
to cooler and drier between ca. 3.2 and 2.0 ka BP, coinciding with a glacier advance in the head area of the
Urumqi River. From ca. 2.0 to 1.4 ka BP, climate became warmer and more humid again. From ca. 1.4 to
0.5 ka BP temperature and humidity went on increasing and a period of Climatic Optimum since 3.6 ka BP
might occur. A few limnetic hydrophytes pollen are counted for all zones, indicating a freshwater habitat
since 3.6 ka BP in this region. Based on synthetically analysis of ecological characteristics and dispersal of
spruce pollen, the abundance of Picea is influenced by treeline moving upward, valley wind and glacier
ablation. Statistics of charcoal concentration and susceptibility further suggest that fires may have
occurred in this region since 0.5 ka BP and the peak value of charcoal might be related to human
activities.
Key words: Tianshan Mountains; vegetation changes; climatic change; pollen record; charcoal; multiproxy
data
Reconstruction of past environment and vegetation is
an important task for studying past global and regional
changes in climate, environment and vegetation. Pollen data
often provide detailed information concerning regional veg-
etation and climate as well as their dynamics and changes.
Studies of fossil pollen can be palaeoecologically informa-
tive and are used worldwide to reconstruct Holocene envi-
ronments (Elenga et al., 2000; Muller et al., 2003). Recently,
researchers emphasized the importance of higher time reso-
lution of pollen data (Popescu, 2001; Pokomy, 2002). Re-
sults showed that multiproxy records including pollen data
are very important to research of environmental evolution
(Mann, 2002; Schmidt et al., 2002).
The semi-arid and arid areas are sensitive belts of global
change in the world, and the palaeoclimatic importance of
pollen research has been long recognized in such regions
(Newsome, 1990; Zhang et al., 2000). For example, vegeta-
tion types in two regions of semi-arid southwestern Aus-
tralia were studied using pollen traps (Newsome, 1990).
Pollen data from sediment stratums in the arid northwest-
ern China recorded abundant palaeoclimatic information.
Periglacial and mountain tundra in the head area of the
Urumqi River in the central Tianshan Mountains have a
profound effect on historical and modern environments,
hence it is complicated to study palaeoenvironment and
vegetation changes by pollen data from surface soils, air,
stratum and archaeological sites. However, many Quater-
nary pollen works in this region since the 1980’s provided
us with more information to accelerate such kind of research
(Zhou et al., 1981; Yan and Ye, 1983; Pan, 1985; Yan and Xu,
1989, 1995; Yan, 1991; Yan et al., 1991; Xu and Yan, 1996).
But it is still lack of studies based on periglacial geomorphy,
glacier change, and vegetation and environmental evolu-
tion in the Holocene through pollen records, especially the
pollen data in high temporal resolution. Furthermore, study
on the relationship between vegetation change and human
activity is still needed in this region.
Spruce (Picea spp.) is one of the most widespread tree
Acta Botanica Sinica 植物学报 Vol.46 No.6 2004656
species, and it is an important component of boreal conifer
forests in the study area. Several studies have indicated
that the shift of spruce is influenced by climatic change
and human activity (Markgraf, 1970; Tsukada, 1983;
Kullmann, 1986). On the other hand, local-scale pollen dia-
grams often give the most detailed information concerning
local vegetation history and human interference on
vegetation. Interpretations of anthropogenic activity in
pollen diagrams have often been revealed by arboreal (AP)/
non-arboreal (NAP) relations and also by the occurrence
of charcoal (Hjelle, 1997). Recently, change of charcoal con-
tent in pollen diagram has significance in research of natu-
ral fire and human activity (Swain, 1973; Singh, 1981; Clark,
1988; Chen, 1990; Figueiral and Mosbrugger, 2000; Sun et
al., 2000).
In this study, we choose the Daxigou profile in the Urumqi
River Head area of central Tianshan Mountains, Xinjiang
Autonomous Region, as the target area. The aim is to dis-
cuss the dynamic changes of vegetation and environmen-
tal evolution since 3.6 ka BP using pollen records and some
auxiliary environmental indices such as loss of ignition
(LOI), grain size and susceptibility. Relationship of the char-
coal content with environmental change and human activ-
ity is also discussed in this study.
1 Data and Methods
1.1 The study area
The source region of the Urumqi River is located on the
north slope of Tianger Peak (43°04¢-43°08¢ N, 86°48¢-
87°00¢ E) in the central Tianshan Mountains (Fig.1). The
Urumqi River originates at an elevation of 4 486 m on Tianger
II peak, the highest point in the eastern Tianshan
Mountains, and flows northward to the city of Urumqi, the
capital of Xinjiang Uygur Autonomous Region, China. The
altitudes of main ridges are 4.1-4.3 km, the modern snowline
at 4.0-4.1 km, the bottom of glacier tongue 3 650-3 700
mm, and the lower limit of permafrost 3.2-3.3 km, respec-
tively (Zhu and Cui, 1992) and the upper limit of treeline is
2.6-2.8 km (Editorial Committee of Xinjiang Forest, 1989).
This mountain is mainly composed of Paleozoic granites
and metamorphic rocks. With regard to geotectonic unit it
is the Tianshan folding belt belongs to the middle part of
the Tianshan geosyncline folding system. Incised river val-
leys developed greatly for the geotectonic movement. The
present climate is continental mountainous climate with
obvious vertical zonality. It is very dry at the lower altitudes,
but climate becomes cold and humid upwards. The tem-
perature reduces quickly with increasing altitude. The
Daxigou weather station is located at an elevation of 3 588
m. The mean annual air temperature is -5.4 °C, with mean
temperature in January of -15.9 °C and in July of 4.7 °C.
Mean annual precipitation is 430.2 mm. But at the elevation
of 4.0-4.5 km mean annual temperature is -8 to 12 °C (Zhu
and Cui, 1992). Snow and ice exist at above 3.8 km. The
natural vegetation changes with increasing altitudes. Al-
pine meadow and subalpine meadow dominate between
Fig.1. Location and stratigraphy of the Daxigou profile.
ZHANG Yun et al.: Vegetation Changes and Environmental Evolution in the Urumqi River Head, Central Tianshan Mountains
Since 3.6 ka BP: a Case Study of Daxigou Profile 657
2.8 and 3.8 km in elevations. Between 1.6 km and 2.8 km
there are forests, and steppe and desert are in lowland be-
low 1.6 km (Yan, 2002).
1.2 The Daxigou profile
The Daxigou section (43°07¢ N, 86°51¢ E), a manual exca-
vated profile in July 2001 with 110 cm thick, is situated in
the head area of the Urumqi River, central Tianshan Moun-
tains at an elevation of 3 450 m (Fig.1). Main vegetation in
this area is alpine meadow. The profile is composed of clay
and barren peat (it is also called turf) with continuous depo-
sition and better stratification.
Two samples at the depth of 44-48 cm and 108-110 cm
were collected for 14C dating (14C dating was analyzed by
the 14C Laboratory of Institute of Geology, China Seismo-
logical Bureau, but it was not calibrated by tree rings). They
are (890±60) a BP and (3 640±60) a BP, respectively.
Age calculations are based on a 14C half-life of 5 568 a. The
age of top section of the profile is set to 0 a BP, because the
characteristics of pollen assemblage at top section and in
surface soil are similar and then the top section is consid-
ered to be not denuded. Ages of the remaining samples
were interpolated by assuming that the sediment rate is
constant between the two dated samples. Then the chro-
nology was established on the basis of these two dating. A
total of 52 samples, at a interval of 2 cm each (time resolu-
tion is about 38-88 a) were obtained for pollen analysis. At
least 300 pollen and spore grains of terrestrial plants (the
maximum is 1 284) in each sample were counted, except the
deep sample at 102-74 cm with poor pollen (but > 150
grains). At the same time charcoal was analyzed by its size.
Pollen diagram was drawn using the TILIA and TILIA-
GRAPH. Pollen percentage for tree, shrubs, and herbs were
calculated upon a pollen sum including aquatic pollen and
spores. Mass concentration (grains per gram) was calcu-
lated by the method of direct concentration. No Lycopo-
dium spore was added.
Proxy environmental indices, such as grain size
(determined by Mastersizer 2 000 laser particle distribution
analyzer), loss on ignition and susceptibility (determined
by MS2 susceptibility analyzer) were also measured for the
same samples as pollen identification (by the State Key
Laboratory of Western China’s Environmental System,
Lanzhou University).
2 Results
Thirty-nine genera and families were identified in 52
samples. Among them, main arboreal pollen was Picea
schrenkiana, together with a fewer pollen of Pinus,
Cupressaceae, Larix, Betula and Salix. Main meso-xero-
phytic shrub and herb pol len was Ar temis ia ,
Chenopodiaceae, Ephedra, Leguminosae, Rosaceae,
Nitraria and Tamarix. Main meso-hydrophytes herb pol-
len was Thalictrum, Polygonum, Gramineae, Plantago,
Caryophyllaceae, Umbelliferae, Cyperaceae and Cruciferae.
Mesic and aquatic vascular pollen types were Potamogeton,
Sparganium and Typha. Freshwater phycophyta spore in-
cluded Spirogyra, Zygnema, Chlamydomonas and Diatom.
Fern spore types were Polypodiaceae and Batrychium. Li-
chen and moss spore included Athalamia, Asterella and
Sphagnum.
Based on variations in pollen concentration and main
pollen percentage, pollen diagram was divided into five
zones (Fig.2). Artemisia / Chenopodiaceae (A/C) ratio was
usually used to indicate the humidity or aridity condition in
arid area (Ann, 1990; Huang, 1993). Here Artemisia / Ephe-
dra (A/E) ratio was also introduced to research of environ-
mental and vegetation change. The high A/E ratio reflects
the more humid.
Zone Ⅰ ca. 3.6-3.2 ka BP (110-102 cm). The sediment
consists of brown-black peat with high pollen
concentration. The pollen assemblage is dominated by
Picea schrenkiana and Artemisia. Pollen percentage and
concentration of Picea are high (37% at the depth of 108
cm). A/C and A/E ratios are significantly high.
Zone Ⅱ ca. 3.2-2.0 ka BP (102-74 cm). Sediment is
brown-black peat with clay. More than 150 pollen grains
were counted, which is the minimum of the whole profile.
The pollen assemblage is still dominated by P. schrenkiana
(25%-35%), Artemisia and Ephedra. This zone did not
have freshwater Chlamydomonas and Diatom.
Zone Ⅲ ca. 2.0-1.4 ka BP (74-60 cm). Sediment com-
position is the same as zone Ⅱ, but pollen concentration is
higher than Zone Ⅱ though pollen percentage decreased
relatively. The pollen assemble is characterized by Ephe-
dra and Artemisia. The A/C and A/E ratios are very high.
Zone Ⅳ ca. 1.4-0.5 ka BP (60-32 cm). The lowest li-
thology consists of brown-black peat with clay, the middle
is light brown peat, but the upper is light brown clay with
peat. Pollen concentration went on increasing and reached
a maximum value. Pollen assemblage is dominated by shrub
and herb pollen (60%), but Picea percentage rose up to
15%, whereas a few exceeded 20%. At the same time Ephe-
dra percentage decreased. Concentration of aquatic pollen
is the highest and A/C and A/E ratios are slightly high.
Zone Ⅴ ca. 0.5-0 ka BP (32-0 cm). The lowest sedi-
ment is light sandy clay, but the upper is brown clay. Pollen
concentration decreased, but average concentration is still
high. Picea percentage decreased gradually, but it is not
Acta Botanica Sinica 植物学报 Vol.46 No.6 2004658
Fig.2. Pollen diagram of the Daxigou profile.
ZHANG Yun et al.: Vegetation Changes and Environmental Evolution in the Urumqi River Head, Central Tianshan Mountains
Since 3.6 ka BP: a Case Study of Daxigou Profile 659
less than 10%, while Ephedra increased. Ephedra and Ar-
temisia are two important components of this zone. A/C
and A/E ratios decreased gradually and aquatic vegetation
decreased, too.
Accordingly, pollen percentage curve of Picea contents
changed from peak value at the bottom of Zone Ⅰ, to low
value in Zone Ⅰ, peak value in Zone Ⅱ, low value in Zone
Ⅱ and Ⅲ, and peak value in Zone Ⅳ and decreased in
Zone Ⅴ (Fig.2). By contrast, Ephedra and Chenopodiaceae
had the opposite trends with the change of Picea, while
aquatic vegetations and A/C and A/E ratios are the same as
Picea.
A few hygrophytes and aquatic pollen were counted
from bottom to top of the profile. Among these plants, con-
tents of Cyperaceae, Spirogyra, Zygnema, Diatom and
Chlamydomonas are very high. It is remarkable that the
minimum value of total pollen concentration of the Daxigou
profile occurs in Zone Ⅱ, whereas the peak of Picea oc-
curs in ZonesⅠ and Ⅳ.
3 Discussion
3.1 Palaeoclimate and palaeoenvironment reconstruc-
tion since 3.6 ka BP
From ca. 3.6 to 3.2 ka BP, slightly higher percentage and
concentration of Picea indicated that its biomass was high
at that time. High A/C and A/E ratios, low percentage of
Artemisia and high pollen concentration of aquatic plants
(Fig.3) reflected that climate was warmer and more humid
than present and Picea treeline shifted upwards. In addition,
Fig.3 shows that average granularity and LOI value are
slightly high. This might indicate that climate was more
humid and vegetation coverage was high, but coarser par-
ticle could be easily transported because of high precipita-
tion and runoff.
From ca. 3.2 to 2.0 ka BP, relatively higher Picea per-
centage but lower pollen concentration (Fig.3) could not
suggest that Picea forest appeared there or treeline moved
upwards. Low LOI value, A/C and A/E ratios and aquatic
vegetation reflected that at that time vegetation coverage
was low and the climate became relative drier and cooler
than before.
From ca. 2.0 to 1.4 ka BP, Picea percentage and pollen
concentration were very low, whereas total concentration
and aquatic pollen began to increase (Fig.3), indicating that
climate became warmer and more humid and vegetation
coverage increased.
From ca. 1.4 to 0.5 ka BP (550-1 350 a AD), Picea per-
centage and concentration rose rapidly. It is also worth
notice that LOI value reached a maximum and average
granularity increased (Fig.3), which indicated that coarser
particle could be easily transported due to high precipita-
tion and much melt water from glacier and snow. In particu-
lar the obvious peak of concentrations of total pollen and
arboreal, shrub, aquatic and herb pollen in this zone could
represent the Climatic Optimum stage in the Daxigou sec-
tion during 1.4-0.5 ka BP. At that time Picea forest-line
moved upwards again. Susceptibility decreased greatly from
Zone Ⅳ because oxidized ferromagnetic particles were
changed into hydrate by stronger soil gleization in the
long-term reductive condition (Sun et al., 1995).
From ca. 0.5 ka BP to the present, the percentage and
concentration of Picea decreased obviously (Fig.3). This
implied that vegetation coverage was still high and the cli-
mate was warmer and more humid than present but less
than Zone Ⅳ, because concentration of total pollen,
Artemisia, Chenopodiaceae, and Ephedra were high.
3.2 Picea’s indicative significance of environments
Another profile of a thaw channel was obtained with
low time resolution at the elevation of 3.5 km in the same
study area several years ago (provided by Liu Gennian in
Peking University). Two dating data were measured at the
depth of 110 cm ((2 015 ± 80) a BP) and 134 cm ((3 640 ± 70)
a BP), respectively. Pollen identification (by Kong Zhaochen
and Du Naiqiu in Institute of Botany, The Chinese Acad-
emy of Sciences) and pollen diagram (Fig.4) showed that
Picea content also varied obviously through this profile
with three remarkable peaks of Picea since 3.6 ka BP. It is
not incidental that there is the same fluctuation of Picea
curve at the same period for those two similar profiles. To
give a rational explanation it is essential to synthetically
analyze the ecological characteristics and pollen transport
of genus Picea.
3.2.1 Transport of Picea pollen in North Xinjiang Ex-
perimental studies on Picea pollen rain showed that the
transport capacity of Picea pollen is lower than Pinus
(Janssen, 1966; Li, 1991). However, Picea amount changes
obviously when it is influenced by strong updraft and val-
ley wind. For example, the Chaiwopu Basin of North Xinjiang
is the windy belt and wind speed is up to 5 m/s. Around the
period of Picea florescence from April to June it is the sea-
son with the highest wind speed in a year. Contents of
Picea pollen in surface soils of this basin increase gener-
ally during this season, but Picea pollen in desert soil away
from the spruce forest reduces to 2%-5%. In the alpine
meadow above the Picea forest belt of the Nanshan
Mountain, however, Picea pollen content of the surface
soil is still 7.1% although here the altitude is 0.5-0.8 km
higher than the upper limit of Picea forest. In the alpine
Acta Botanica Sinica 植物学报 Vol.46 No.6 2004660
Fig.3. Pollen concentration (grain/g), grain size, susceptibility and loss of iqnition (LOI) of the Daxigou profile.
ZHANG Yun et al.: Vegetation Changes and Environmental Evolution in the Urumqi River Head, Central Tianshan Mountains
Since 3.6 ka BP: a Case Study of Daxigou Profile 661
Fig.4. Pollen percentage diagram of a thaw channel profile at the Daxigou weather station in the Head Area of the Urumqi River, central
Tianshan Mountains.
Acta Botanica Sinica 植物学报 Vol.46 No.6 2004662
meadow pollen is obviously transported by updraft, which
is much more than those in desert transported by downdraft.
Such phenomenon that updraft moves a great deal of pol-
len exists in common in the mountainous areas of West
China (Li, 1988). Hence it is an universal phenomenon of
pollen transport. Statistics of Picea pollen contents in sur-
face soils from the Tianshan Mountains, Altay Mountains,
Kunlun Mountains, Tarim Basin and Junggar Basin also
showed that those samples that Picea contents exceed 30%
were collected at 1.3-2.8 km, and those samples that Picea
contents exceed 20% were collected at 1.3-3.3 km. The
former altitude range is basically the same as the distribu-
tion range of Picea forest, but the latter obviously exceeds
its upper limit. Therefore, Picea pollen in surface soils at
0.5 km above upper limit of forest is more than 20%, which
is likely to be influenced by valley wind (Yan et al.,
submitted).
In addition, other 28 surface pollen samples (Fig.5) col-
lected from the same study area but with altitudes (3 880 m)
down to the Chaiwopu Basin (1.1 km) indicated that, in the
alpine rocky vegetation belt (3 880 m) spruce pollen occu-
pied 9.6% of total pollen content and 15.9% in the alpine
meadow (3.5 km), but it was only 2.2% in Chaiwopu (1.1 km)
under the lower limit of spruce forest (Fig.5). This might be
due to that valley wind at the Daxigou area exists all the
year and the wind speed here is 1 m/s higher than the vici-
nal hillside (Editorial Committee of Land Resource in
Xinjiang, 1993).
3.2.2 Treeline movement of Picea forest and its response
to climatic change Fossil pollen analysis also indicated
that cold temperate forests consisting of Abies and Picea
were distributed widely in mountains and plains in main-
land of southwestern, northwestern, northern and eastern
China and Taiwan during the Quaternary Glacial Stage (Xu
et al., 1980; Chen, 1988; Fang, 1996; Editorial Committee of
Forests in China, 1997), which suggested that coniferous
forest of cold temperate zone had larger distribution area
than today. Vegetation zone has horizontal and vertical
shifts along with the frontier and retreat of the glacier. Gen-
erally speaking, snowline and treeline shifted down when
glacier expanded, while cold temperate coniferous forest
retreated gradually along with the mountain glacier melting
in the interglacial periods when temperature increased.
For this profile, from ca. 3.6 to 3.2 ka BP, high pollen
concentration and percentage of Picea might suggest that
at that time the climate was warmer and more humid than
today. Picea forest zone expanded and treeline moved
upwards. A great deal of Picea pollen was transported from
the forest area to the place of profile. The stage looked like
a “small interglacial stage”. From ca. 3.2 to 2.0 ka BP, Picea
percentage was still very high, but pollen concentration,
LOI and susceptibility were low, indicating a drier and colder
“ice age” climate. Furthermore, pollen percentages of
samples from the fourth to eleventh excluding the eighth at
Zone Ⅱ were very low. In contrast, a previous study indi-
cated that there was high pollen content of Picea in
Fig.5. Relationship between spruce pollen percentage and elevation in the northern slope of the Tianshan Mountains.
ZHANG Yun et al.: Vegetation Changes and Environmental Evolution in the Urumqi River Head, Central Tianshan Mountains
Since 3.6 ka BP: a Case Study of Daxigou Profile 663
surface soils of the First Glacier in the source area of the
Urumqi River (Zhou et al., 1981), which might be due to the
transportation of valley updraft wind. On the other hand,
when identifying pollen samples of Zone Ⅱ, we found that
pollen grains in this zone were not only small but also
fragmentized mostly. It is possible that after they were
blown into the glacier by valley updraft Picea pollen grains
were crashed by the friction of glacier. Then they depos-
ited to the lowland by glacial melt water. A glacier frontier in
the head area of the Urumqi River occurred in 2.8 ka BP
(lichen dating) since the New Ice Age and the moraine ridge
of Shanbei group Ⅲ formed. This “S” type of moraine ridge
is located 250-300 m east away from the Second Glacier in
the head area of the Urumqi River and is mainly composed
of eyeball gneiss, schist, granitic diorite and dolerite (Chen,
1988). So there is an obvious relationship between Picea
content and the advance and retreat of glacier. The Zone
Ⅳ also suggested that from ca. 550 to 1 350 a AD not only
the percentage and concentration of Picea pollen increased
again, but also the pollen concentration of shrubs, herbs
and aquatics increased. At the same time there were high
LOI and grain size and low susceptibility, which reflected
that climate became warmer and more humid again, as coin-
cided with Zhao et al. (1983).
Besides the influence of valley wind and glacier
movement, Picea percentage could reflect the historical
treeline changes. Treeline is more sensitive to climate
change, which has become one of the hotspots of global
change study (Liu, 2002). At present the lower limit of Picea
distribution is 1.7 km and the upper limit is 2.7 km, with a
maximum of 1.0 km Picea forest belt. The snowline is at
4.2-4.5 km. The altitude difference between the upper limit
of forest and the snowline is 1.5-1.8 km (Editorial Commit-
tee of Xinjiang Forest, 1989). Altitude of the treeline de-
pends mainly on temperature, but the altitude of snowline
relies mainly on precipitation. Studies suggested that aver-
aged air temperature of the warmest month is the most im-
portant thermal control of the development of zonal dark
coniferous forests (Li and Zhou, 1979; Wu, 1983; Liu et al.,
2002). The mean temperature of the warmest month at the
upper limit of this forest is generally about 10-14 °C. Pollen
analysis of the Daxigou profile also indicated that the up-
per limit of spruce forest is mainly controlled by tempera-
ture because precipitation is enough to the survival of
spruce trees. The impact of moisture condition on the dis-
tribution of dark coniferous forests is also presented to
some extent by the impact of moisture condition on thermal
condition.
Based on long-term (1950-2000) meteorological records
of Daxigou (3 539 m), Xiaoquzi (2 160 m), Tianchi (1 938 m),
Jimsar (735 m) and Fukang (547 m) weather stations and
using a linear regression, the mean annual temperature and
mean annual precipitation along elevations were roughly
interpolated (Fig.6). The mean annual temperature at the
upper limit of spruce forest is –1.4 °C and the mean annual
precipitation is 490 mm. At the lower limit of spruce forest
Fig.6. Changes of mean annual temperature and mean annual precipitation along with the increasing elevation in Tianshan Mountains.
Acta Botanica Sinica 植物学报 Vol.46 No.6 2004664
the temperature and precipitation are 3.3 °C and 450 mm,
respectively (Fig.6). According to pollen analysis at ca.
3.6-3.3 ka BP, spruce pollen occupied 40% of total pollen
(excluding aquatic pollen), indicating that the upper limit of
spruce forest moved upward by ca. 0.6 km. Then, the mean
annual temperature at that time is ca. 3 °C and annual pre-
cipitation is ca. 50 mm higher than today.
In a nutshell, we could conclude that the Picea abun-
dance might be related to climatic change, valley wind, the
advance and retreat of glacier and treeline moving upward.
However to further understand which factor is the most
important, more pollen and multi-proxy data are needed in
future study.
3.3 Charcoal distribution, vegetation change and human
activity
Fire is one of the major disturbances in forest and grass-
land ecosystems. Charcoal, often called “fossil of fire”, is
the product by plants after incomplete burning. Small size
of charcoals can be transported by air and water from the
origin place to deposit site, but airflow only transports about
2% of total charcoals and water can transport most of
charcoals. However, large charcoals bigger than 100 mm
cannot be transported easily and they deposit mostly in
their origin site. Therefore, the number of charcoal particles
and their existing section in deposits can be taken as indi-
cators of regional fire occurrence and frequency (Swain,
1973; Singh et al., 1981; Patterson et al., 1987; Clark, 1988;
Chen, 1990; Sander and Gee, 1990; MacDonald et al., 1991;
Huang et al., 1996; Zhang et al., 1997; Sun et al., 2000).
Figure 3 shows that a small peak of charcoal content
existed at depth of ca. 28 cm and the charcoal particles at
different sizes ≥100 mm, 50-100 mm and ≤50 mm had the
same peaks at the same time. We could not distinguish
whether they were woody charcoal or herb charcoal and
whether the fire was natural or man-made, but very low
pollen contents of Picea, Ephedra, Cyperaceae, Spirogyra
and Zygnema and a small peak value (26%) of Picea pollen
before charcoal peak (Figs.2, 3), indicated that at that time
climate was relatively dry and forest fire might occur
frequently. The peak Picea pollen before charcoal peak
also showed that Picea forest grew well at that time and
forest biomass, shrubs and herbs as well as litters increased
that contributed to more occurrences and higher frequency
of fires. Therefore charcoal content was abruptly increased.
It is worth mentioning that many peak values of char-
coal occurred in Zone Ⅴ. Compared it with susceptibility
curve, it can be found that abundance of charcoal particles
corresponded very well with high value of susceptibility.
Generally, more magnetic oxides lead to combustion within
the soil; accordingly the susceptibility increases
(Thompson and Oldfield, 1995). Therefore to some extent
high charcoal numbers can be correlated with fire.
Additionally, biomass should be high if natural fire
occurred. But Fig.3 shows that LOI is very lower, indicat-
ing biomass was not high. So the likelihood of natural fire
can be eliminated. It might be related to the increase of use
of fire by increased human activity in recent hundred years,
but dry climate increased the possibility of occurrence of
fire.
4 Conclusions
Based on pollen concentration, pollen percentage and
the analysis of average granularity, susceptibility and LOI,
we concluded that the climatic conditions since 3.6 ka BP
can be divided into five stages: there was a warm-humid
period between ca. 3.6 to 3.3 ka BP, with temperature was
higher than today. Then climate deteriorated between ca.
3.2 to 2.0 ka BP. From ca. 2.0 to 1.4 ka BP climate became
warmer and more humid again, and the period of Climatic
Optimum since 3.6 ka BP prevailed between ca. 1.4 and 0.5
ka BP (550-1 350 a AD). Since ca. 0.5 ka BP, climate became
drier.
Based on the synthetically analysis of ecological and
transport characteristics, Picea pollen abundance is related
to valley wind, treeline moving upward and glacier retreating,
which reflects the correlation between vegetation dynamic
change and environment evolution.
Many peaks of total number of charcoal particles were
recorded in Zone Ⅴ. Furthermore, abundance of charcoal
concentration corresponded very well with high suscepti-
bility by contrasting it with susceptibility and LOI curve.
Since 0.5 ka BP increased human population and human
activities as well as dry climate caused the possibility of
fires.
Acknowledgements We thank Prof. LIU Gen-Nian of Pe-
king University and XU Qing-Hai of Hebei Normal Univer-
sity for their assistance with sampling, Dr. ZHU Yan etc. of
Lanzhou University for chemical analysis, and YIN Jin-Hui
of China Seismological Bureau of Geology Institute for 14C
dating. Thank also specially to Prof. CUI Zhi-Jiu for his
valuable comments on the early version of this manuscript.
References:
Ann P E. 1990. Ecological significance of common nonarboreal
pollen: examples from drylands of the Middle East. Rev
Palaeobot Palyno, 64: 343-350.
Chen J-Y. 1988. A study on lichen dating and other issues of
Holocene glacial change in the head Area of Urumqi River of
ZHANG Yun et al.: Vegetation Changes and Environmental Evolution in the Urumqi River Head, Central Tianshan Mountains
Since 3.6 ka BP: a Case Study of Daxigou Profile 665
Tianshan Mountains. Sci China (Ser D), 1: 95-104. (in Chi-
nese with English Abstract)
Chen Y-S. 1990. Forest fire in early Holocene forest changes at
lake Barrine, Australia. Acta Bot Sin ,32: 69-75. (in Chinese
with English Abstract)
Clark J S. 1988. Particle motion and the theory of charcoal analysis:
source area, transport, deposition, and sampling. Quaternary
Res, 30: 81-91.
Editorial Committee of Forest in China . 1997. Forest in China.
Vol. 1. Beijing: China Forestry Press. 513-574. (in Chinese)
Editorial Committee of Land Resource in Ürümqi . 1993. Land
Resource in Xinjiang. Ürümqi: Xinjiang People Press. 52. (in
Chinese)
Editorial Committee of Xinjiang Forest. 1989. Forest in Xinjiang.
Ürümqi: Xinjiang People Press. 10-100. (in Chinese)
Elenga H, Namur C, Vincens A, Roux M, Schwartz D. 2000. Use
of plots to define pollen-vegetation relationships in densely
forested ecosystems of Tropical Africa. Rev Palaeobot Palyno,
112: 79-96.
Fang J-Y . 1996. The distribution pattern of Chinese natural veg-
etation and its climatologic and topographic interpretations.
Wang R-S, Fang J-Y , Gao L , Feng Z-W . Researches on
Hotspots of Modem Ecology. Beijing: China Science and Tech-
nology Press. 369-380. (in Chinese)
Figueiral I, Mosbrugger V. 2000. A review of charcoal analysis as
a tool for assessing Quaternary and Tertiary environments:
achievements and limits. Palaeogeogr Palaeocl, 164: 397-
407.
Hjelle K L. 1997. Relationships between pollen and plants in
human-influenced vegetation types using presence-absence
data in western Norway. Rev Palaeobot Palyno, 99: 1-16.
Huang C-X. 1993. A study on pollen in surface soil from western
Xizang. Arid Land Geogr, 16(4): 75-83. (in Chinese with
English Abstract)
Huang C-Y , Kong Z-C , Pu Q-Y . 1996. Research on sediment of
Kunming Lake since 3500 a BP. Beijing: China Ocean Press.
91-142. (in Chinese)
Janssen C R. 1966. Recent pollen spectra from the deciduous and
coniferous forests of Northeastern Minnesota: a study in
pollen dispersal. Ecology, 47: 804-824.
Kullmann L. 1986. Recent tree-limit history of Picea abies in the
southern Swedish Scandes. Can J Forest, 16: 761-770.
Li W-H , Zhou P-C . 1979. Study on distributional general rules
and mathematical models of dark coniferous forests in Asian-
European Continent. Nat Res, 1: 21-34. (in Chinese)
Li W-Y . 1991. Spreading efficiency of Picea pollen. Acta Bot Sin,
33: 792-800. (in Chinese with English Abstract)
Li X. 1988. Holocene vegetation and environment evolution of
Luoji Mountain in Xichang City, Sichuan. Acta Geogr Sin ,
43: 44-51. (in Chinese with English Abstract)
Liu H-Y . 2002. Quaternary Ecology and Global Change. Beijing:
Science Press. 91-97. (in Chinese)
Liu Z-L, Fang J-Y, Piao S-L . 2002. Geographical Distribution of
Species in Genera Abies, Picea and Larix in China. Acta Geogr
Sin , 57: 77-58. (in Chinese with English Abstract)
MacDonald G M, Larsen C P S, Szeicz J M, Moser K A. 1991.
The reconstruction of boreal forest fire history from lake
sediments: a comparison of charcoal, pollen, sedimentological,
and geochemical indices. Quaternary Sci Rev, 10: 53-71.
Mann M E. 2002. The value of multiple proxies. Science, 297:
1481-1482.
Markgraf V. 1970. Palaeohistory of the spruce in Switzerland.
Nature, 228: 249-251.
Muller U C, Pross J, Bibus E. 2003. Vegetation response to rapid
climate change in Central Europe during the past 140 000 yr
based on evidence from the Furamoos pollen record. Quater-
nary Res, 59: 235-245.
Newsome J C. 1990. Pollen-vegetation relationships in semi-arid
southwestern Australia. Rev Palaeobot Palyno, 106: 103-
119.
Pan A-D. 1985. Late quaternary pollen assemblages and their
signification of Junggar Basin, Xinjiang Province. J Glacio
Geocryo, 7: 257-264. (in Chinese with English Abstract)
Patterson W A. Edwards K J, Maguire D J. 1987. Microscopic
charcoal as a fossil indicator of fire. Quaternary Sci Rev, 6: 3-
23.
Pokorny P. 2002. A high-resolution record of late-glacial and early-
Holocene climatic and climatic and environmental change in
the Czech Republic. Quaternary Int, 91: 101-122.
Popescu S M. 2001. Repetitive changes in Early Pliocene vegeta-
tion revealed by high-resolution pollen analysis: revised
cyclostratigraphy of southwestern Romania. Rev Palaeobot
Palyno, 120: 181-202.
Sander P M, Gee C T. 1990. Fossil charcoal: techniques and
applications. Rev Palaeobot Palyno, 63: 269-279.
Schmidt R, Koinig K A, Thompson R, Kamenik C. 2002. A multi
proxy core study of the last 7000 years of climate and alpine
land-use impacts on an Austrian mountain lake (Unterer
Landschitzsee, Niedere Tauern). Palaeogeogr Palaeocl, 187:
101-120.
Acta Botanica Sinica 植物学报 Vol.46 No.6 2004666
Singh G, Kershaw A P, Clark R L. 1981. Quaternary vegetation
and fire history in Australia. Gill A M, Groves R H, Noble I R.
Fire and the Australia Biota. Canberra: Australia Academy of
Science. 23-54.
Sun J-M, Ding Z-L , Liu D-S . 1995. Environment evolution of
desert-loess border belt since Last Interglacial. Quaternary
Sci, 15: 117-122. (in Chinese with English Abstract)
Sun X-J, Li X, Chen H-C. 2000. Evidence for natural fire and
climate history since 37 k a BP in the northern part of the
South China Sea. Sci China (Ser D), 30: 163-168. (in Chinese
with English Abstract)
Swain A M. 1973. A history of fire and vegetation in northeast
Minnesota as recorded in lake sediments. Quaternary Res, 3:
383-395.
Thompson R, Oldfield F. 1995. Yan R-J , Wu B-C , Chen D ,
Zhang F-L Trans. Environmental Magnetism. Beijing: China
Geological Press. 57-69. (in Chinese)
Tsukada M. 1983. Late-Quaternary spruce decline and rise in
Japan and Sakhalin. Bot Magaz Tokyo, 96: 127-133.
Wu X-H . 1983. Study on temperature of dark coniferous forests.
Chin Sci Bull, 23: 1451-1457. (in Chinese)
Xu R , Kong Z-C , Du N-Q . 1980. Pleistocene Picea and Abies
floras and their Quaternary significance in China. Quaternary
Sci China , 5: 48-56. (in Chinese)
Xu Y-Q , Yan S, Jia B-Q , Yang Y-L . 1996. Numerical relationship
between the surface pore-pollen and surrounding vegetation
on the southern slope of Tianshan Mountains. Arid Land
Geogr, 19: 24-30. (in Chinese with English Abstract)
Yan F-H , Ye Y-Y. 1983. Pollen assemblages and their significance
of Luosijin profile of Lop Nur region, Xinjiang Province.
Seismol Geol , 5: 75-80. (in Chinese with English Abstract)
Yan S , Li W-Y , Liang Y-L, Xu Y-Q . 1991. Pleistocene pollen
assemblages and environment of Chaiwopu Basin, Xinjiang.
Geogr Symp Arid Zone , 2: 1-14. (in Chinese with English
Abstract)
Yan S , Xu Y-Q . 1989. Spore-pollen association in surface-soil in
Altay, Xinjiang. Arid Land Res , 1: 26-33. (in Chinese with
English Abstract)
Yan S , Xu Y-Q. 1995. Pollen assemblage in the moraine and the
environment in glacial epoch in Tianshan Mountains. Arid
Land Geogr , 18: 21-26. (in Chinese with English Abstract)
Yan S. 1991. Quaternary pollen assemblages and vegetation
succession. Arid Land Geogr , 14: 3-8. (in Chinese with
English Abstract)
Yan S. 2002. The information of environmental evolvement of the
northern piedmonts of the Tianshan Mts. in the history. Acta
Phytoecol Sin , 26: 82-87. (in Chinese with English Abstract)
Yan S , Kong Z-C , Yang Z-J , Zhang Y, Ni J. Seeking relationships
between spruce (Picea spp.) pollen in surface soils and veg-
etation in Xinjiang, northwestern China. Acta Ecol Sin , 24. (in
Chinese with English abstract)
Zhang H C, Ma Y Z, Wunnemann B, Pachur H J. 2000. A Ho-
locene climatic record from arid northwestern China.
Palaeogeogr Palaeocl, 162: 389-401.
Zhang J-H, Kong Z-C , Du N-Q . 1997. Charcoal analysis and fire
changes at Dongganchi of Fangshan in Beijing since 15 000
years BP. Acta Phytoecol Sin, 21: 161-168. (in Chinese with
English Abstract)
Zhao L, Qiu G-Q. 1983. The climate fluctuation and the perma-
frost formation since the Last Glaciation, Tianshan, China. J
Glaciol Geocryol , 15: 103-109. (in Chinese with English
Abstract)
Zhou K-S, Liang X-L, Liu R-L . 1981. Primary studies on pa-
lynology of quaternary sedimentary and glacier-ice in the Head
Area of Urumqi River of Tianshan Mountains. J Glaciol
Geocryol , 3: 97-105. (in Chinese with English Abstract)
Zhu C , Cui Z-J. 1992. The distribution and evolution of periglacial
Landforms in the source region of Urumqi River on the
Tianshan Mountain. Acta Geogr Sin, 47: 526-535. (in Chi-
nese with English Abstract)
(Managing editor: HAN Ya-Qin)