Data Process
Resume:
A RTICLES
Chinese Science Bulletin **** Vol. ** NO.9 953-960 Ngola Shan is mostly deposited above 3500-4500 m
altitude. It has been growingly attracted how these loesses
Dust storms and loess form and where they come from (Fig. 1). Current geologic
evidence shows that the Tibetan loess is "cold loess" or
accum ulation on the Tibetan periglacial loess, has coarser grain size than that of the
Plateau: A case study of dust loess on the Loess Plateau to the northeastern Tibetan
Plateau, and most possibly comes from the Tibetan Pla-
event on 4 March 2003 in teau itself. The middle and high levels of the westerly jet
and the Tibetan monsoon are likely to the major generator
Lhasa and carrier of dusts[I-5]. Comparison of ESR signals of
quartz grains of Japanese Kosa (means yellow sand or
FANG Xiaomin 1,2, HAN Yongxian~ 1,3, MA Jinghui 1, dusts) suggests that much of the Kosa is transported by the
3
SONG Lianchun, YANG Shengli & ZHANG Xiaoye westerly jet most likely from the Tibetan Plateau[6]. The
study of dust records of deep-sea sediments also reveals
1. Key Laboratory of Western China's Environmental Systems (Lanzhou
that the Asian arid land between 25 and 40 0 N is the major
University), Ministry of Education & College of Resources and En-
dust source area and the westerlies is the carrier conveying
vironment, Lanzhou University, Lanzhou 730000, China;
2. State Key Laboratory of Loess and Quaternary Geology, Institute of dusts[7,8]. Current studies demonstrate that loess in the Ti-
Earth Environment, Chinese Academy of Sciences, Xi'an 710054,
betan Plateau and its marginal areas are mostly formed at
China;
about 0.8-1.15 MaBP[2,4,5,9-12], mostly due to a combina-
3. Lanzhou Arid Meteorological Institute of China Meteorological Bu-
reau, Lanzhou 730000, China tion of the Tibetan Plateau being raised into a key tlrresh-
Correspondence should be addressed to Fang Xiaomin (e-mail:
old height of atmospheric kinetics and physiographic land-
******@***.***.**)
scape where the Plateau rising into the cryosphere and the
atmosphere circulation being greatly adjusted with the
Abstract Whether the Tibetan Plateau is a significant
dust source area is of great importance, because this is re- lower and middle levels of the westerlies being forced to
lated to the understanding of sources, accumulation and diverged into two jets to flow around the Plateau and the
environmental effects of dusts on the Tibetan Plateau and in stable occurrence of the Tibetan monsoon[5,l1,12]. Further-
the Far East-Pacific Ocean regions as well as to the evolution more, this event may have caused a strong desertification
of coupling of the Tibetan Plateau and atmosphere-ocean-
of Asian inland giving rise to formation of great deserts in
continent exchange. Synoptic dynamics and remote sensing
NW China and a great expansion of loess southeastwards
tracing of a dust storm on 3 to 5 March, 2003 in Lhasa on
from the Loess Plateau to East China[5,1l,12] as well as have
South Tibet demonstrate that the Tibetan Plateau possesses
triggered the global cooling at tlris time (so-called MPR =
all factors and conditions of generating dust storms. Accom-
Mid-Pleistocene Revolution) [1 3] .
panied with this dust storm is a strong ascending stream on
the Plateau which has raised various sizes of dust particles lOO E
80 90
into different levels. The lifted coarse particles were largely
fallen down and accumulated as loess on the eastern Tibetan
Plateau, and the fine particles were translated by the west-
erly jet and subsided in the northern Pacific Ocean. The spa-
tial-temporal distribution of dust-storms between years 1961
and 2000 on the Plateau shows that dust-storms mainly occur
in winter and early spring with high frequency, and the path
of dust storm moves gradually from south to north, which is
closely coupled with the northward moving of the westerly
jet from winter to spring over the Tibetan Plateau. Com-
pared with other twelve dust source areas in China, the Ti-
betan Plateau is one of the key dust source areas for the
long-distance transport because its high occurring frequency
and elevation cause fine particles easily to be lifted into the
zone of the westerly jet.
o
Keywords: Tibetan Plateau, dust storm, dust source areas, Tibetan
1 4 Dune or sandland of modem DuneofLGM
loess.
Imilia
E:;J Dust days from I% I to 2000 Loess
DOl: 10.1360/03wd0180
Fig. 1. The distribution map of modern and last glacial sand dunes,
desertified sandy land and loess and the spatial mean annual days plot of
Loess is widely distributed on the eastern part of the
dust storm from 1961 to 2000 on the Tibetan Plateau. Aeolian sand and
Tibetan Plateau and Plateau marginal areas surrounding loess distribution was drawn according to our field investigation and refs.
the central and western Tibetan Plateau. Particularly the [1-5], [9-12] and [19-21]. Dust storm data are from the Climate
Plateau loess to the south of the eastern Kunlun Shan-
953
Chinese Science Bulletin May
Vol. 49 NO.9 2004
A RTICLES
Center of China Meteorological Administration. When the westerly jet adjusts, the Tibetan Plateau will
Zhang Xiaoye and his co-authors used the chemical appear westerly winds from the Yarlung Zangbo valley
element tracer system to analyze the aerosol characteris- and the Qiangtang Plateau as well as northwesterly wind
tics and flux of dust storms occurred between September from the Qaidam Basin. These winds meet respectively in
and October 1993 and between April and March 1994 at the source areas of the Yangtze and Yellow Rivers and in
an altitude of 4800 mat Wudaoliang (35.2 N, 93.1 E) on the Lhasa region and form strong convergent upwelling
the central Tibetan Plateau. They concluded that the Ti-
air flows there, providing power for dusts to be lifted. In
betan Plateau is not a dust source area[l4-16 J. However, addition, there are widely distributed flowing sand dunes
because the observation data they used are short and only on the Tibetan Plateau. For example, an area of about
come from one locality, their data are hard to secure a 2000 km2 of desertified surface (with moving sand dunes
whole pattern of dust fluxes on the Tibetan Plateau both in in winter) occurs in the valleys of Yarlung Zangbo River
time and in space. Therefore, whether the Tibetan Plateau
and its tributaries with an average altitude of about 3600
is or not a significant dust source area remains still uncer-
m (Fig. 1). At the rim of this desertified surface is loess
tain. The area of the Tibetan Plateau is account for about
with thicknesses of up to 30 m and ages old back to 0.8
one fourth of total area of Chain, and is mostly over alti-
Ma[9,19J. Between the Gangdese and Kunlun Shan are a
tudes of 3500-4500 m, much higher than that of the sur-
series of near west-east glaciated mountains and a vast
rounding desert areas (e.g., about 840-1200 m above the
area of planation surface (Plateau Surface) on which there
Taklimakan Desert to the north, 1300-1800 m above the are distributed many varied sizes of lakes and wind-blown
geomorphology and moving sand dunes[20 J. The wind-
Badain Jaran Desert and 400-1600 m above the Tengger
Desert to the northeasti 17J . As early as in early 1980s, blown and desertified surfaces occupy about an area of
about 400 km2, on which the most are moving sand dunes.
Duce and others recognized that fine dusts derived from
These sand dunes are greatly activated in winded winterI 21J .
Asia in spring are mostly transported by the westerly jet at
500 hPa (~5500 m) and fallen down in the northern Pa- For example, vast patched moving sand dunes are found
cific Ocean[l8J. Therefore, fine dust particles on the Ti- in the source areas of the Huang He (the Yellow River)
betan Plateau are much easier lifed into high level of the and Chang Jiang (the Yangtze River) (e.g., source area of
westerly jet than those in other dust source areas. It de-
Tongtian He (River), the vast area to the west of the road
serves being soon clarified if the Tibetan Plateau is a sig-
from the Kunlun Shan pass to Lhasa, and areas at Xingx-
nificant dust source area, because it is not only a regional
ing Hai, Rigecuo and Ghoring Lake beach). Moreover,
scientific question related to understanding the loess ori-
desertified meadow and grass soils occupy area much
gin and formation on the Tibetan Plateau, but also is sig-
wider than that of sand dunes and include almost all
nificantly related to the uplift of the Tibetan Plateau and
source areas of the Huang He and Chang Jiang except
its coupling with global ocean-continent-air system.
sand dunes and mountain bedrocks. The parent materials
On 4 to 5 March, 2003, CCTV (China Central Tele-
of these soils are relatively homogeneous and consist
vision) successively reported an event of strong blow-
mostly of silts and fine sands[20 J. At present, the desertifi-
ing-dust from 3 to 5 March in Lhasa. This event forced all
cation land is rapidly increasing. The investigation shows
the scheduled flights to Lhasa to be cancelled, stopping
that the area of desertification land only in the Tibetan
over 2000 passengers at the Shuangliu airport in Chengdu.
Autonomous Region reaches about 2047.41 X 10 4 hm2,
This unusual synoptic process of strong blowing-dust on
distributed in every county and city, accounting for
the Tibetan Plateau provides a fair opportunity to study the
17.03% of the total area of the Tibetan Autonomous Re-
above-mentioned scientific question. Taking this case, we
gion[21 J. Moreover, thick connected sand dunes widely
analyzed the path and vertical ascending movement of the
appear in those desertification areas. These sand dunes
dust-storm. Furtermore, on the basis of statistics of about
40-year data of dust-storms from 91 weather stations on were mostly formed at 14-24 kaBP in the last glacial
the Tibetan Plateau and integrating dust storm literatures maximum (LGMpoJ, indicating much stronger desertifi-
in Japan and Korea, we analyzed the spatial-temporal dis- cation occurring in the LGM (Fig. 1). These widely dis-
tribution and long-distance translation of dust storms on tributed sand dunes and desertification land provide plen-
the Tibetan Plateau, providing our information on the teous materials for dust storm generation. The average
source of Tibetan dust storm. annual days spatial figure of dust storm occurring between
1961 and 2000 shows that the Qiangtang Plateau and Tak-
1 Geography and source materials of dust storms in
limakan Desert to the north are the two high frequency
the studied region
areas of dust storm occurrence. The frequency is over 15
The studied region includes nearly the whole Tibetan days a year in the Qiangtang Plateau and over 5 days a
Plateau having topography inclining from the northwest at year in the other parts of the main Tibetan Plateau, with a
average altitudes 5000 m to the southeast at about 3500 m. decreasing trend southeastward from the center of the
Chinese Science Bulletin
954 Vol. 49 NO.9 May 2004
negative allobaric center along the west path, the blow-
Qiangtang Plateau (Fig. 1).
ing-dust appeared successively at Shiquanhe, Gaize, Din-
2 Synoptic dynamic diagnosis of the strong blow-
gri, Rikaze, Gongga, Lhasa, Dangxun, and Anduo, reach-
ing-dust event in Lhasa
ing its maximum at about -8 hPa at 14:00 to 20:00 on 4
The gale and strong blowing-dust in the Tibetan Pla- March. At the same time, blowing-dust occurred in the rear
teau on 3-6 March, 2003 was formed in the process of of the north center successively at Dehaling and Maduo.
the adjustment of the circulation of mid-low latitudes in 80 90 100 E
Eurasia and the southward moving of the strong westerly
jet. Before blowing-dust occurred, the pattern of the cir-
4O N
culation on the Eurasia continent presents as two troughs
and one ridge in the weather chart of 300 hPa, in which
the area from the Tibetan Plateau to West China was con-
trolled by the high pressure ridge, resulting in a 50-60
mls strong westerly jet locating at 33 N near the Plateau
30
and a weak south branch trough. With the Europe trough
being deeper and moving eastwards, the upper air front
moved southwards and the south branch trough gradually
intensified and developed. Up to 4-5 March, the two low
pressure systems in the mid and low latitudes phased to-
gether, greatly deepening the south branch trough in the
Fig. 2. Sketch map summarizing the weather of the blowing-dust event
western side of the Plateau and leading the strong westerly on 4-6 March, 2003. Black solid line: Boundary and time of the ap-
jet nearby the Plateau move southwards to 30 0 N with a pearance of the blowing-dust event; black circle: Time and location of
wind speed of 60-65 mls. This combined system led the the appearance of the 24-hour allobaric center; dash arrow: direction of
movement of the allobaric center; 1: 11 :00 am on March 3; 2, 3 and 4:
momentum transfer downwards to low level, providing a
02:00 am, 08:00 am and 20:00 pm on March 4, respectively; 5: 20:00 pm
basic atmospheric circulation necessary for the dust event on March 5; 6: 08:00 am on March 6.
in this study. This weather system moved out the main
part of the Tibetan Plateau at 08:00 am on 6 March, 2003. We used continuous integral equation to calculate
Dust storm is a middle scale weather process. Its forma- vertical speed OJ modified by the O'Brien nonlinear equa-
tion requires three conditions, i.e., sand source, gale and tion[23]. Fig. 3 shows the OJ at 500 hPa (5.5 km) and from
vertical ascending air stream. Because gale and dust storm 500-100 hPa along 70 oE-100oE at 30 0 N on the Tibetan
frequently appear in the heel of negative allobaric center,
Plateau. We can clearly see the ascending movement pre-
we use chiefly the movement of 24-hour allobaric center
vailing on the Tibetan Plateau with three rising centers.
to trace the moving track of the vertical ascending air
The first center is the strongest one with OJ at -1.0 hPaJS
stream[22,23], thus derterming the trajectory of dust storm.
locating roughly in the source region of the Yellow and
Subsequently, we calculated and plotted the spatial distri-
Yangtze Rivers; the second has OJ at -0.6 hPaJS located in
bution of the vertical speed OJ of the dust storm at its big-
the regions of Lhasa-Namucuo Lake-Dangxun; the last
gest on the Tibetan Plateau and the altitude of the vertical
has OJ at -0.4 hPaJS located in the southern Kunlun
ascending air stream in the Lhasa region using methods of
Mountains and northern Quangtang Plateau (Fig. 3(a)).
dynamics diagnosis.
The vertical transection along 86 E-93 E indicates that
When the dust storm occurred, the wind speeds had
the vertical ascending movement prevailing the whole
achieved the critical wind speed over 6 mls in every direct
levels from 500 (5.5 km) to 100 hPa (16 km) with the
to generate. With the gale and blowing-dust occurring,
largest values at 88 E-89 E (west of Lhasa) (where the
Fig. 2 shows that there are two negative allobaric centers
OJ is as big as to -0.5 hPaJS even at 250 hPa and still at
moving eastwards and southwards, respectively. From
11:00 on March 3, one negative allobaric center moved -0.2 hPaJS even at 100 hPa (Fig. 3(b )).
from the western Plateau to its eastern part and finally out The synoptic dynamics analysis of the dust storm
above shows that the Tibetan Plateau possesses all condi-
of the Plateau, passing through Shiquanhe (with value of -6
tions for dust storm generation when the westerly jet ad-
hPa), Gaize (-6 hPa), Dangxun (-6 hPa, Naqv(-8 hPa)
justs. When dust storm occurs, the vertical ascending
and Yushu (-4 hPa) (called south path); the other moved
movement prevails on the Tibetan Plateau. Our case indi-
from the northern Plateau to its southeast, passing through
cates that the OJ at 500 hPa on 4 March, 2003 in Lhasa is
Geermu (-8 hPa)and Zaduo (-6 hPa) (called north path)
much higher than that (-0.53 hPaJS)l) in the Tarim Basin
and joining the west path at Yushu at 02:00 on March 5
with its value decreasing to -4 hPa. In the rear of the
I) Lii, L.Q., Fang, X. M., Lu, H. Y. et aI., Grain size record of millennial winter monsoon variation on the Tibetan Plateau since last glaciation,
955
Chinese Science Bulletin Vol. 49 NO.9 May 2004
A RTICLES
Plateau. Among tliese dust storms, tliose occurring in the
far-reached Qiangtang Plateau are the most strongest
where lots of flowing sand dunes provide plenty dusts for
tlie dust storm generation. Even in the MODIS images on
March 5, patched blowing-dusts and floating-dusts still
can be seen in tlie central Tibetan Plateau and tlie float-
ing-dusts even expanded to tlie Hengduan Shan of the
soutlieastern Tibetan and the Sichuan Basin (Fig. 4(c)).
This observation is quite similar with tliat analyzed by
meteorological dynamics diagnosis, but the scope and
35"
intensity of the dust storm are much beyond our imagina-
tion. The processed MODIS images of floating-dusts on
April 10 in the source areas of the Yellow and Yangtze
Rivers also clearly show tliat tlie dusts were transported
toward tlie downstream regions. The floating-dusts ves-
tured the valley of tlie Yellow River and its main tributar-
ies from tlie Gyaring and Ngoring Lakes to Lanzhou, and
a mass of yellow dusts were deposited in tlie peak area
(highest peak at 6282 m) of tlie Buerhanda Shan and An-
yemaqen Shan. Thus, we estimated that these dust grains
are lifted at least over 5500 minto tlie zone of tlie west-
erly jet.
4 Spatial-temporal distribution of dust storms on the
400 - Tibetan Plateau and the westerly transport of dusts
The spatial-temporal distribution of dust storms be-
500 _ _ _ _
tween 1961 and 2000 on the Tibetan Plateau and in Xin-
72" 78" 84 90
jiang to the north of the Plateau demonstrates that the oc-
Fig. 3. Distribution of ascending speed OJ on the Tibetan Plateau at currence of dust storms on the Tibetan Plateau is princi-
20:00 pm on March 4,2003. (a) OJ at 500 hPa; (b) OJ from 500-100 hPa
pally between December and April and gradually moves
along 70 0 E-1 OooE at 30 0 N.
northwards from the valley of Yarlung Zangbo River to
3 Remote sensing monitoring of the blowing-dust tlie soutliern Qiangtang Plateau, then to tlie central Qiang-
event tang Plateau and tlie soutliern Tarim Basin with the season
passing on. This seasonal movement is closely related
The complex topography and sparse weather stations
with tlie movement of tlie westerly jet over the Tibetan
on the Tibetan Plateau make it impossible routine moni-
Plateau. When tlie westerly jet moves from soutli to north
toring of large-scale regional dust storm. However, this
over the Tibetan Plateau along with the season changing
can be supplemented to a certain extent by remote sensing.
from winter to spring, the southern branch of the westerly
We used ERDAS Imaging processing software firstly to
jet gradually becomes weaker, whereas tlie northern
calibrate the geometry and radiation and to enhance image
of MODIS data, tlien to process the MODIS data on branch becomes stronger (the westerly jet over tlie Tibetan
March 1 (sunshine), March 4 (blowing-dust) and March 5 Plateau is diverged into two branches in tlie winter and
(floating-dust) by monitoring and non-monitoring class spring seasons due to tlie high topography of the Plateau).
methods[24], and finally to map out tlie distributive area Superimposed witli the impact of tlie complicated topog-
and intensity of tlie blowing-dust on the Tibetan Plateau, raphy on the Tibetan Plateau, the convergence center of air
integrated witli surface data observed by weatlier stations currents also moves accordantly from soutli to north in the
and meteorological conditions (Fig. 4). Combined witli Plateau, giving rise to a corresponding northward move-
dust storm images of NOAA-12 at 14:00 on March 2 in ment of tlie center of dust storms (Fig. 5).
the Tarim Basin and at 17 :00 on March 3 in the source The content of >63 flm of fine sands in surface sedi-
areas of the Yellow and Yangtze Rivers (Data were issued ments of the Tibetan Plateau is over 60%-70%[19-21].
by tlie National Meteorological Satellite Center of China Their short distance transport is mostly proceeded witli
Meteorological Administration, and images are not pre-
creeping and bouncing. When gale and dust storm occur,
sented here due to limited space), the processed MODIS
tliese coarse dust grains can be deposited in downstream
images on March 4 shows clearly three paths of tlie blow-
slope and adjacent area and finally tum into loess by pe-
ing-dust, i.e., (1) from tlie NW Plateau to its eastern part,
dogenesis. For example, tlie loess in the valley of Yarlung
(2) from the Yarlung Zangbo River valley to tlie eastern
Plateau and (3) from tlie Qaidam Basin to tlie soutlieastern
956 Chinese Science Bulletin Vol. 49 No.9 May 2004
(al (h)
4 N
4 ' N
'"
35 N
3 N
Cloond & snow
L
c:l Flootin~ duSi
25" N c Blowing dIN.
-Water
95 E 100 E
9!r E
X5'E 9 E
XO E
35 N
3 ' N
Fig, 4, Remote sensing processed MODIS images of dust storms on the Tibetan Plateau, (a) Sunshine image at 12:00 on March 1,2003 in Lhasa; (b)
strong blowing-dust image at 12:00 on March 4, 2003 in Lhasa; (c) images of blowing-dust and floating-dust at 12:00 on March 4,2003 in the western
Sichuan Basin; (d) floating-dust image at 12:00 on April 10, 2003 in the source areas of the Yellow and Yangtze Rivers,
Zangbo River in the Lhasa-Rigaze region is very coarse
with >63 flm fine sands over 60%-70%, and becomes
coarser when approaching sand dunes[20,25 J, In addition,
loess is widely distributed on the eastern Tibetan Plateau
surrounding the east part of source areas of the Yellow and
Yangtze Rivers, The grain size of this loess becomes finer
away from the river source areas, but much coarser than
the loess in the nearby Loess Plateau!), Consequently, the
coarse fractions of the Tibetan loess cannot be transported
from sources out of the Plateau but the Plateau itself
However, the fine fractions of dust grains can be
raised by convergent ascending flow onto the westerly jet
at 500 hPa (5,5 km) which is the main power for long dis-
tance transport of Asian dusts[l8,26-28 J, Because the altitude
of the Tibetan Plateau is higher than that of other source
Fig, 5 Monthly spatial-temporal distribution of dust storms and the
westerly jet on the Tibetan Plateau, Black solid arrow and number:
areas in Asia, such as, 840-1200 m, 1300-1800 m and
Westerly jet and month; open circle: Center of dust storm; number and
1400-1600 m higher than those of the Taklimakan,
(number) in open circle: Abbreviation for month and number of monthly
Badain Jaran and Tengger Deserts, respectively[l7 J, only
occurrence of dust storm; black solid circle in the auxiliary figure: Sta-
tion observing dusts in Japan and Korea; a, Seoul; b, Yashiro; c, Na-
very strong dust storm is able to raise fine dusts in these
gasaki; d, Ishigkiis; e, Okayama,
I) See footnote I) on page 955,
Chinese Science Bulletin Vol. 49 No, 9 May 2004 957
desert source areas onto the westerly jet at 5500 m level. annual days of dust storm generation on the Tibetan Pla-
This indicates that the occurrence of dust storm on the teau are close to the Hobq Desert, only lower than Takli-
Tibetan Plateau over altitudes 3500 to 4500 m approaches makan, Badain Jaran, Tengger and Ulan Buh Deserts, but
or exceeds the main height of biggest dust storms in the higher than the Gurbantunggut and Mu Us Deserts and
Tarim Basin lifting fine dusts, thus fine dusts on the Ti- Otindag, Hulun Buir and Horqin sandy lands (Table 1).
betan Plateau are much easier to be lifted onto the zone of The Tibetan Plateau is between the high dust regions of
the westerly jet for long distance transport, as shown by northwestern deserts and low dust regions of northern
the blowing-dust even in Lhasa above. deserts as defined by ref. [30]. Therefore, judging from
The dust records in Japan and Korea provide further annual days of dust storm generation, the Tibetan Plateau
evidence for the long distance transport of the Tibetan fine can be defined as a moderately dust storm occurring re-
dusts. The monthly distribution of dusts at four stations in gion.
Japan and Korea (Ishigkiis, Nagasaki, Yashiro and Seoul) Table 1 Annual days of dust stonns occurred in all source areas in
shows that high dust records in all four stations appear in China (Data except the Tibetan Plateau come from ref. [17])
December and January to April, and that with latitude Dust stonn
Dust source area Altitude/m
(blowing-dust)/d
higher, the highest frequency firstly appears in the south
4279 16.5(58)
Western Tibet
(24 ON) in March and gradually moves to the north in April
3650 4-5(16)
Southern Tibet
and May[6]. However, in China, this pattern of higher fre-
4674 16(36)
Southern Qiangtang Plateau
quency occurrence of dust storms only appears on the Ti-
4800 12(30)
Source area of Yangtze River
betan Plateau. Therefore, we estimate that a majority of
3324 9(16)
Southern Qinghai Lake
Japanese and Korean dusts in winter may come from the
840-1200 30
Taklimakan Desert
Tibetan Plateau. Further evidence comes from the fact that
200-1000
Gurbantunggut Desert 5
dusts in Japan often have a two-layer structure in early
Badain Jaran Desert
spring, with the lower layer below 4000 m a.s.l. and the
Tengger Desert
upper one above 4000 m a.s.l. This made Iwasaka and
Ulan Buh Desert
others think that the two-layer dusts may derive from dif-
Hobq Desert
ferent areas[26,28]. Our study shows that the upper layer of
Mu Us Desert
dusts may reflect the long-distance transport of fine dusts
Otindag Sandy Land
mainly from the Tibetan Plateau. For example, there oc-
500-1000
Hulun Buir Sandy Land 1
curred a strong dust storm in the source areas of the Yel-
0-500 5-10
Horqin Sandy Land
low and Yangtze Rivers at 18:00 on April 9, 2003. At
12:00 on April 11, a dust layer appeared at altitudes of
The westerly jet at 500 hPa is the main power for the
4000 to 7000 m at Okayama (see Fig. 5 for location) in
long-distance transport of Asian dusts[l8,26-28]. The abso-
Japan and no dusts appeared below 4000 m, providing
lute altitude of fine dust grains raised into the westerly jet
direct evidence for the long-distance transport of fine
at 500 hPa can be calculated from all known altitudes of
dusts from the Tibetan Plateau to the northern Pacific re-
dust source regions in China. The altitudes of twelve dust
gion[29].
source regions are at about 1000 to 2000 ill. Thus, fine
Discussion
5 dust grains need to lift about 3500 to 4500 m onto the
Apart from the geologic evidence, from the westerly jet zone at 500 hPa. This means that only strong
meteorological point of view, whether the Tibetan Plateau dust storms have the possibility. For example, the loess on
is a significant dust source area mostly depends on dust the northern slope of the Kunlun Mountains is the deposi-
storm frequency and the ability lifting dusts onto the zone tion of dust storms from the Taklimakan Desert. The grain
of the westerly jet. size and thickness decrease gradually with the increase of
Figure 1 shows that the Tibetan Plateau has higher altitude. For instance, the thickest loess is found at an al-
dust storm frequency and the high frequency center is in titude of 2900 m in Ruoqiang, then the loess thins upper-
wards until about 4500 m[ll]. This indicates that the power
the Qiangtang Plateau with a trend gradually decreasing
southeastwards. We choose five weather stations at Shi- raising dust grains decreases with rising altitude, thus,
quanhe, Lhasa, Shengzai, Wudaoliang and Xinghai to only strong dust storms can arrive at mountain peaks. By
represent the western Plateau, the valley of Yarlung literature, there totally occurred 32 times of strong dust
Zangbo River, the southern Qiangtang Plateau, the source storms in Hetian in recent 40 years, only 3% of the totality
of dust storms at 1036[31]. Even in the case where 2 and
areas of the Yellow and Yangtze Rivers and the Qinghai
others included more dust storms into strong dust storms[32]
Lake, respectively. These dust stations can reflect princi-
due to different definition of a strong dust storm[33], they
pally dust characteristics on the main part of the Tibetan
Plateau. Compared with the annual days of dust storm reported the strong and very strong dust storms in the
generation in twelve sand desert regions in China[17], the Hexi Corridor also only account for 17% and 3.4% of the
958 Chinese Science Bulletin Vol. 49 NO.9 May 2004
A RTICLES
total dust storms in recent 45 years, respectively[33 J
Therefore the maximum of strong dust storms will be not 8 Blowing-dust
more than 50% of total dust storms. Taking into account D Dust storm
o Observation priod
the height lifting fine dusts onto the westerly jet zone 6
above, the number of dust storms rising onto the westerly
jet in twelve dust source regions in China will reduce at
least 50%. However, dust grains on the Tibetan Plateau
only need to raise about 1000 to 2000 m to get into the
2
zone of the westerly jet. The case of Lhasa above indicates
00
that only a weak dust storm, the blowing dust process, can
lift fine dusts onto the zone of westerly jet. The sum of
Jan Mar Maylul SeptNovJan MarMayJul SeptNov
annual dust storms and blowing dust process on the Ti-
1993 1994
betan Plateau is over thirty (Table 1). Even excluding the
blowing dust process, the number of dust storms able to Fig. 6. Monthly days of dust storms at Wudaoliang between 1993 and
rise onto the zone of the westerly jet is still much higher 1994. Dots mark the period of sampling in ref. [15].
than that in other dust source areas. All these go to a con-
clusion that the Tibetan Plateau is one of the major source
areas for the long-distance transport of dusts. " Shenzai
40
I :".
Zhang and others used the GMW-231O sampler to - - - Wudaoliang ( \
1\ ~,: "
I
- Lhasa
collect aerosol samples at Wudaoliang at 4800 m on the
I~
30
en
I
II I II II
;::l,
central Tibetan Plateau[l4-16J On the postulation that de- /\,I II II
'0,
i I_' \\
4-
" .,v h I"
I I / I
o 20 "
positional dust flux is equal to released dust flux in an
~ 1\ "
area, their data give out a depositional dust flux of about
N,/ ill \1 I
Ci r \ II I \ I
100 g/(m2 a) for the Tibetan Plateau[15 J Comparing this I liV
10 v I,,
datum with those from Chinese deserts, Loess Plateau, "
1955
historic dust-fall regions in northeast and southeast China
at 410, 250, 95 and 17 g/(m2 a), respectively, they con- Fig. 7. Annual dust storm days of representative stations on the Tibetan
cluded that the Tibetan Plateau is not a dust source re- Plateau.
gion[15 J
The very large variation of dust concentration in dif- real mean of the Tibetan dust flux. If viewing from geo-
ferent meteorological processes leads to a great difference logic time scale, the modem grassland and meadow areas
in depositional dust fluxes. For example, according to the on the Tibetan Plateau were mostly source areas of the
observed data from four stations like Jilantan, dust con- past dust storms, because fixed sand dunes are immedi-
centrations in processes of normal weather, floating dust, ately under them, e.g., those in the source areas of the
blowing dust and weak dust storm are 0.083, 0.356, 1.206 Yellow and Yangtze Rivers. Chronology and deposition of
and 3.955 mg/m3, respectively[l7 J And the dust concentra- those sand dunes demonstrate that they were mainly
tion of a very strong dust storm, e.g., that in Jinchang on formed in the last two glaciations when dunes were very
May 5, 1993, even reaches 10 17 mg/m3[3l]. Therefore, active and provided a tremendous amount of dusts for the
Tibetan coarse loess[I,2,19 J (Fig. 1). Taking into account
whether there occur or not dust storms in the sampling
period has great impact on resulting dust flux and thus spatial distribution, the occurrence of dust storms on the
conclusions have been drawn. Dust storms on the Tibetan Plateau is a dynamic process changing from south to north
Plateau occur mostly in winter and early spring. However, (Fig. 5), and is closely related with every synoptic process
the sampling interval in ref. [15] was just in months of and distribution of dust source and topography. As a con-
low dust storm occurrence (Fig. 6), so the calculated de- clusion, because the sampling in ref. [15] was too short
positional dust fluxes cannot present the annual dust flux and only at one site at Wudaoliang, the data in ref. [15]
of the Tibetan Plateau. If viewing for a long period, we can neither represent the annual dust flux nor the
find that the sampling intervals in ref. [15] lie also in the long-term average dust flux for the whole Tibetan Plateau.
nearly lowermost period of dust storm generation over the The potential as a dust source region for the Tibetan Pla-
last 40 years on the Tibetan Plateau as indicated by days teau was greatly underestimated. However, considering
of dust from Wudaoliang (35.2 oN, 93.1 E), Lhasa (29.4 ON, the difficulty to collect dust aerosol samples on the Ti-
91.08 E) and Shengzai (30.57 N, 88.38 E) (Fig. 7), 2 to 6 betan Plateau, the calculated depositional dust flux by ref.
times lower than the mean over the last 30 years. There- [15] can also reflect the state of fluxes in non- or few dust
fore, even for one station at Wudaoliang, observation in storm period.
ref. [15] represents annual dust flux neither in short term Therefore, both geologic evidence and the character-
nor in long term, thus their data are far away from the istics of modem circulation and dust storms on the Tibetan
Chinese Science Bulletin Vol. 49 NO.9 May 2004 959
A RTICLES
Plateau demonstrate that the Tibetan Plateau is a signifi- transportation of atmospheric aerosol over northern Qinghai-
Xizang (Tibetan) Plateau (eds. Tang Maocang, Cheng Guodong,
cant dust source area.
Lin Zhenyao), Contemporary Climatic Variations Over Qinghai-
Acknowledgements Many thanks are due to Niu Zhimin and Ye Yan- Xizang (Tibetan) Plateau and Their Influences on Environments (in
hua for their help. This work was co-supported by Hundred Talents Chinese), Guangzhou: Guangdong Science & Technology Press,
Project of the Chinese Academy of Sciences (Renjiaozi[2000]0282), the
1998, 83-98.
National Natural Science Foundation of China (Grant Nos. 49925101
17. Qin Dahe, Yang Debao, Shang Kezheng et aI., Dust Storm (in Chi-
and 40121303) and the National Key Project for Basic Research on Ti-
nese), Beijing: Meteorological Press, 2003, 50.
betan Plateau (Grant No. GI998040802). Lanzhou Arid Meteorological
18. Duce, R. A., Unni, C. K., Ray, B. 1. et aI., Long-range atmospheric
Institute, China Meteorological Administration provided the data of
transport of soil dust from Asia to the tropical North Pacific: Tem-
MODIS.
poral variability, Science, 1980,209: 1522-1524.
References 19. Dong Yuxiang, Li Sen, Dong Guangrong, Tentative study on the
status and the causes of desertification in Yarlung Zangbo River
I. Fang Xiaomin, The origin and prov
Contact this candidate