Impacts of wind velocity on sand and dust deposition during dust storm as inferred from a series of observations in the northeastern Qinghai–Tibetan Plateau, China

5 (2007) 82–89www.elsevier.com/locate/powtec

Powder Technology 17

Impacts of wind velocity on sand and dust deposition duringdust storm as inferred from a series of observations in the

northeastern Qinghai–Tibetan Plateau, China

Mingrui Qiang ⁎, Fahu Chen, Aifeng Zhou, Shun Xiao, Jiawu Zhang, Zhenting Wang

Center for Arid Environment and Paleoclimate Research (CAEP), Key Laboratory of Western China's Environmental System (Ministry of Education),Lanzhou University, Lanzhou 730000, PR China

Received 17 July 2006; received in revised form 25 November 2006; accepted 28 December 2006Available online 12 January 2007

Abstract

The monthly sand and dust deposition flux and modern dust storms were monitored in the northern Qaidam Basin of the Qinghai–TibetanPlateau. The monthly sand and dust flux varied between 0.57 and 18.12 mg cm−2 month−1 from June 2003 to April 2005, and was well correlatedwith the monthly extreme wind velocity (Vextr) (r

2=0.60, n=23). Sand and dust was mainly deposited in spring and early summer in the studyarea. The weight of settled sand and dust collected during dust storms exhibited a positive correlation with the mean 10-min wind velocity(r2=0.60, n=16) during the dust storms. For the typical dust storms, the weight and flux of settled sand and dust will linearly increase with theincreasing wind strength and fluctuation amplitude of wind velocities. The coarse fraction (N63 μm) also increases with them, in contrast, the fine-grained fraction (b63 μm) decreases. It is plausible to assume that most of the fine-grained dust particles are lifted and transported far from theregion under dust storm conditions, especially under the stronger and more variable wind conditions. The results demonstrate that the wind regime(strength and variability) is a key control on the sand and dust deposition during dust storm; dust can be emitted from the Qaidam Basin as one ofdust source areas in China.© 2007 Published by Elsevier B.V.

Keywords: Dust storm; 10-min wind velocity; Sand and dust deposition; Qaidam Basin

1. Introduction

Atmospheric dust plays a significant role in biogeochemicalcycles in the oceans and participates in the regulation of levelsof atmospheric CO2 [5,10]. Dust may be more than an indicatorof climate change; dust itself could be an agent of climatechange [27]. Northwestern China has been recognized as one ofthe key dust source areas for global atmospheric dust, which hasbeen recorded in areas such as the Greenland ice sheets [2] andthe pelagic sediments of the North Pacific [20,21]. Bory et al.[3] also found that the provenance of dust during major periods

⁎ Corresponding author. Key Laboratory of Western China's EnvironmentalSystem (Ministry of Education), College of Environment and ResourcesSciences, Lanzhou University, No. 222 Tianshui South Road, Lanzhou 730000,PR China. Tel.: +86 931 8914529; fax: +86 931 8912330.

E-mail address: [email protected] (M. Qiang).

0032-5910/$ - see front matter © 2007 Published by Elsevier B.V.doi:10.1016/j.powtec.2006.12.020

of spring and summer deposition in Greenland was theTaklamakan desert of northwestern China based on isotopic(Sr and Nd) and mineralogical analyses of dust collected fromsnow deposits. However, their isotopic analyses did not includesamples from the Qinghai–Tibetan Plateau, therefore thepossibility of the Plateau as a dust source area for the globaldust loading cannot be excluded. Fang et al. [7] demonstratedthat the Qinghai–Tibetan Plateau is most likely to be animportant source of atmospheric dust, since the altitude of thePlateau is so high (typically more than 3000 m above sea level),and dust particles are prone to being lifted into high-levelwesterly circulation above the Plateau. The Qaidam Basin in thenortheastern part of the Qinghai–Tibetan Plateau is mainlycovered by large areas of playa, Gobi and sand desert, and thisbasin was believed to be a potentially important dust source area[16,19]. In addition, most of the Qaidam Basin is sparselypopulated and vegetation there is scarce. On the other hand, in

Fig. 1. Location map of the observational site.

Fig. 2. The variations in monthly sand and dust deposition flux (bars) and themonthly frequency (days) of strong winds (≥17 m s−1) (triangles, with a rangeof 0–10 days) from June 2003 to April 2005, and the monthly frequency (days)of dust storms and blowing dusts (circles, with a range of 0–4 days) fromNovember 2003 to April 2005.

83M. Qiang et al. / Powder Technology 175 (2007) 82–89

this basin there are lots of hydrologically closed lakes with arelatively simple hydrological cycle. These lakes are mainly fedby meltwater and groundwater and lake sediments can recordsthe long-term evolution of dust storm and dust emission in thisregion [18]. Therefore, it is necessary to continuously monitordust storms to further understand the importance of this regionas a dust source area and the processes of dust storm and dustemission.

Deposition of airborne dust is dynamic and episodic due tovariations in the frequency and magnitude of dust storm occur-rences [17]. A few earlier studies investigated dust depositionduring short intervals (2 to 3 years) or during typical dust stormsin the Hexi Corridor [6], in the Lanzhou region [12], and in theHotian region of Xinjiang province [15]. Chemical and physicalproperties of the particles in a single dust storm have also beeninvestigated [15,29]. However, these researchers did not suf-ficiently consider how variations in wind velocity affected dustemission, transportation, and deposition when dust storms tookplace.

In the paper, we present the results of sand and dust de-position (here, we use the “sand and dust” to refer to the settledmaterials since they contain the abundant sand component)during 16 dust storms (visibility less than 1 km) and 7 blowingdusts (1 kmbvisibilityb10 km) from November 2003 to May2005, together with monthly sand and dust deposition data forthe northern Qaidam Basin. The relationships between theweights, fluxes and different grain-size components of sand anddust that settled during dust storms and the 10-min windvelocities were analyzed. Our study site (38°51′N, 93°54′E)was located at the Lenghu Meteorological Station (LMS) in thenorthern Qaidam Basin, Northeastern Qinghai–Tibetan Plateau(Fig. 1). Based on the records collected by LMS, this site had amean annual precipitation of 15.8 mm and a mean annualpotential evaporation of 2793 mm, and experienced an averageof 55 days per year with strong wind (≥17 m s−1 at a heightof 10 m above ground) from 1957 to 2000, and the number of

dust storm days per year has ranged from 3 to 16 during thisperiod.

2. Methods

Sand and dust samples were collected using a gravimetricmethod. Three cylindrical glass vessels (15 cm in diameter and30 cm deep), each containing two layers of glass balls with adiameter of 1.5 cm at the bottom of the vessel, were placed on aroof 3.5 m above the ground to trap the sand and dust depositedduring dust storms, a season and a whole year at LMS. A fourthvessel of this type was positioned on a 1.5-m-tall trestle table totrap sand and dust deposited nearer to the ground and providedmonthly values. The vessel for trapping sand and dust of duststorm was opened at the beginning of each dust storm andclosed at the end. The vessels and the glass balls were thenwashed with distilled water to collect the trapped sand and dust.The wash water was then dried at 105 °C in an oven until nowater remained, then the samples were weighed. The rate ofsand and dust deposition (the deposition flux) is expressed asthe mass of sand and dust that settled in the vessels per unit areaand unit time.

84 M. Qiang et al. / Powder Technology 175 (2007) 82–89

Wind regime is an important dynamic agent responsible forthe emission, transportation, and deposition of aeolian dust. Inthis study, LMS recorded the mean monthly wind velocity(Vmean), maximum monthly wind velocity (Vmax), extrememonthly wind velocity (Vextr) and monthly frequency of strongwinds (≥17 m s−1) at the height of 10 m from June 2003 to May2005. Vmax is defined as the maximum of the 10-min mean windvelocities recorded within a month, whereas Vextr represents themaximum instantaneous wind velocity recorded in a month.The 10-min wind velocity (mean wind velocity within a 10-mininterval) data was obtained from the recording papers producedby an automatic anemometer during each dust storm.

Grain size of the settled sand and dust samples collectedduring the two seasons and six typical dust storms wasmeasured with a laser grain-size analyzer of Malven Mastersizer2000, which has a measurement range of 0.02–2000 μm and100 size classes. Samples were pretreated according to themethod of Lu and An [13]. The results were presented as avolume percentage.

3. Results and discussion

Fig. 2 shows variations in the monthly sand and dustdeposition flux from June 2003 to April 2005. These fluxesranged from 0.57 to 18.12 mg cm−2 month−1, with a mean of3.58 mg cm−2 month−1. The cumulative sand dust depositionfrom February to June 2004 (28.31 mg cm−2) accounted for52% of the total for the year, suggesting that the sand and dust ismainly settled in spring and early summer, as has been reportedfor other regions, such as China's Hexi Corridor, Loess Plateauand Greenland [3,6,12,23]. The monthly frequency of strongwinds (≥17 m s−1) varied in the range of 0–10 days fromNovember 2003 to April 2005 in this area and was apparentlyhigher than that of dust storm and blowing dust (Fig. 2), whichsuggests that the occurrences of dust storm or blowing dust mayneed stronger wind strength in this area. However, the higherfrequencies of strong winds during spring and early summer(e.g. from February to June 2004) nearly coincided with thefluxes of settled sand and dust (Fig. 2). Most of the dust stormsin the northwestern China are triggered by cold surges fol-lowing the passage of a cold front [25]. The January–Aprildust storms mainly developed when the circumpolar vortex

Fig. 3. The relationships between the monthly sand and dust deposition fluxes (Fd)velocity (Vmax), and (c) monthly mean wind velocity (Vmean).

showed a large 500-hPa trough from Greenland over Siberiainto northern China, which in turn controlled an influx ofmature cyclones from the North Atlantic into China [11]. Thesynoptic processes could account for the higher (sand) dustdeposition in spring and early summer in the different ob-servational sites. The fluxes in the present study varied morewidely than those in China's Hexi Corridor and Loess Plateau[6,23]. The flux was greatest in September 2004. Vegetation isscarce in the northern Qaidam Basin, which differs in thisrespect from the Loess Plateau [24]. More abundant vege-tation in autumn in a region with higher precipitation wouldefficiently prevent dust emission even in the presence of strongwinds.

A significant positive correlation between the fluxes andVextr was found (r2 =0.60, n=23) (Fig. 3a). However, therewere no obvious linear correlations between the fluxes andeither Vmax or Vmean (Fig. 3b, c, respectively), indicating thatVextr is the most important factor that affected sand and dustdeposition in the study area. Aeolian deposits in desert regionsof China contain a great lot of sand component with diametersbetween 100 and 250 μm, and a threshold entrainment velocityat an arid and naked ground surface in the regions determinedbymany field observations was 5m s−1 [26]. Our observationalsite is similar to the occasion as described byWu [26]. Hence, asame threshold entrainment velocity was estimated in thisstudy. The weaker correlation between the fluxes and Vmean

could be well understood because the threshold entrainmentvelocity is an important control on dust emission.

We continuously observed the dust storms and blowing dustsat LMS from November 2003 to April 2005 (Table 1). Wemonitored 16 dust storms and 7 blowing dusts, and collected thesand and dust that settled during each event except event 2(Table 1). The dust storms and blowing dusts occurred between12:20 and 22:00, especially in the interval between 15:00 and16:00. The temporal distribution of these events is related to thedaily changes in atmospheric circulation that occur in responseto changes in the surface's thermal balance. From midday tonightfall, heating of ground surface produces wind and unstableconditions [28]. The weight per unit area of sand and dustcollected during these events ranged between 0.17 and24.06 mg cm−2. Although the monthly sand and dust depositionflux was apparently affected by the frequencies of the dust

and (a) the monthly extreme wind velocity (Vextr), (b) monthly maximum wind

Fig. 4. The relationship between the weight of settled sand and dust (Wd) and themean 10-min wind velocities (V10-min) during 16 dust storms.

Table 1Detailed monitoring of dust storm (DS) and blowing dust (BD) events at the Lenghu Meteorological Station from November 2003 to May 2005

Event no. Date Time Duration (min) Visibility (m) Weight (mg cm−2) Note

1 30 Nov. 2003 14:53–16:40 107 – 15.74 DS, BD after 16:432 14 Feb. 2004 17:24–17:38 14 – – DS, no sample3 10 Mar. 2004 14:46–15:10 24 600 2.26 DS4 27 Mar. 2004 18:38–18:47 9 600 1.70 DS5 24 Apr. 2004 15:15–16:27 – – 1.13 BD6 1 May 2004 19:05–19:55 – – 2.43 BD7 23 May 2004 15:01–15:12 – – 0.57 BD8 28 May 2004 15:40–17:53 133 100 7.93 DS9 3 June 2004 13:03–15:07 4 300 2.53 DS, 13:03–13:07; BD, 13:52–15:0710 4 June 2004 16:51–18:33 12 200 4.25 DS, 16:51–16:58, 17:10–17:15;

BD, 17:20–17:32, 17:40–18:3311 24 June 2004 17:46–21:37 231 – 0.85 BD12 25 June 2004 18:03–18:20 17 – 0.85 DS13 11 July 2004 17:44–18:07 19 600 0.85 DS, 17:44–18:03; BD, 18:05–18:0714 26 July 2004 15:14–15:38 24 400 3.96 DS, 15:14–15:38; BD, 15:23–15:3415 11 Sept. 2004 15:04–15:38 34 100 24.06 DS16 20 Dec. 2004 17:04–18.38 48 500 13.02 DS, 17:36–18:24; BD, 17:04–17:35; 18:24–18:3817 1 Mar. 2005 15:31–15:38 – – 0.57 BD18 6 Apr. 2005 19:20–21:10 80 500 3.96 DS, 19:50–21:10; BD, 19:20–19:5019 7 Apr. 2005 18:20–18:40 20 – 0.85 DS, 18:20–18:40; BD, 8:11–13:27, 18:40–18:5020 18 Apr. 2005 14:41–14:44 3 600 1.42 DS21 4 May 2005 12:17–12:23, 13:21–15:21 126 200 18.97 DS22 11 May 2005 19:42–19:57 – – 0.17 BD23 15 May 2005 14:45–15:21 36 400 14.15 DS, 14:45–15:21; BD, 15:22–15:36

85M. Qiang et al. / Powder Technology 175 (2007) 82–89

storm and blowing dust (Fig. 2), the magnitude of eachindividual storm contributed more to total sand and dustdeposition. For example, event 15 in September 2004 causedthe maximum deposition flux during our observation period,even though only one dust storm took place this month (Fig. 2).

The weight of settled sand and dust collected during duststorm is positively correlated with the mean 10-min windvelocities during the dust storm (Fig. 4) (r2 =0.60, n=16),which supports the existence of a relationship between themonthly deposition flux and the Vextr that we described pre-viously, indicating that the strength of the wind (i.e. its velocity)has an important impact on dust emission, transportation, anddeposition. Generally, the vertical dust flux (dust-emission rate)is proportional to U⁎

n (U⁎, friction velocity) with n varyingbetween 2 and 5 [1,8,9,14,26,]. Shao et al. [22] have reportedthat there was little dust emission even at the maximum flowspeed which the tunnel generated (20 m s−1) if no saltationparticles were introduced, while strong dust emission occurredif sand particles were propelled over the dust surface. Withincrease in wind velocity the sand flux will increase in the upperlayer of airstream and decrease in the bottom layer, limit of thedifferent layers is at 3–4 cm above bed, although the fluxexponentially decreases with heights [26]. Here, we think theseexperiments mainly focused on the near surface conditions,which seems not to be fit for our observations. In this study, wedeal with the relationship between the settled sand and dust atthe heights of 1.5 and 3.5 m and the changes in wind velocity.The results show that the weights of settled sand and dust atthese heights linearly increase with increasing wind velocities.Actually, the low precipitation and higher evaporation in theQaidam Basin determines that there are no significant seasonal

changes of the basin surface, which suggests that the soilmoisture, soil ice, vegetation cover and snow cover, etc. seem tohave negligible impacts on the dust emission. The wind regimeshould be an important control on the dust deposition flux in thestudy area.

We investigated six typical dust storms to better under-stand how wind velocity affects the sand and dust deposition.Variations in the 10-min mean wind velocity during these sixevents are shown in Fig. 5. The durations and total sand anddust deposition during these events ranged from 34 to 133 minand from 7.93 to 24.06 mg cm−2, respectively (Table 2). Thefluxes of sand and dust during these events range from 0.06 to0.71 mg cm−2 min−1 (Table 2). Event 15 lasted 34 min andproduced 24.06 mg cm−2 sand and dust, whereas event 8 lasted133 min and produced only 7.93 mg cm−2 of sand dust. Thevisibilities during events 8 and 15 were both less than 100 m

Fig. 5. The changes in the 10-min mean wind velocity during six typical dust storms.

86 M. Qiang et al. / Powder Technology 175 (2007) 82–89

(Table 2). At the beginning of both events, there was a largeincrease in wind velocity and a steep gradient for this increase.The wind velocity increased by 8.3 m s−1 within 30 min of thestart of event 8, versus 10 m s−1 within 20 min of the start ofevent 15 (Fig. 5b,c). This rapid increase in wind velocity couldresult in intensive dust emission from the areas covered by silty

Table 2Comparison of the characteristics of the six typical dust storms in Fig. 4

Event no.(from Table 1)

Duration(min)

Visibility(m)

Weight(mg cm−2)

Flux (mg cm−2

min−1)Grain-siz

b2 μm

1 107 – 15.74 0.15 3.108 133 100 7.93 0.06 4.1315 34 100 24.06 0.71 3.0516 48 500 13.02 0.17 3.3821 126 200 18.97 0.15 2.5923 36 400 14.15 0.39 2.49

a Cv (variation coefficient) indicates the amplitude of the fluctuation in the 10-mi

Cv ¼ rl

where σ is the standard deviation, and μ is the mean value of the 10-min wind velo

r ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXni¼1

ðXi− X̄ Þ2n−1

vuut ; l ¼Xni¼1

Xi

n

where Xi is the 10-min wind velocity for parameter i, and X̄ is the arithmetical mea

and sandy materials, perhaps leading to a sharp increase in theatmospheric dust concentration in a short interval, accompaniedby a sharp decrease in visibility. The emitted sand and dustwould be transported and deposited in response to changes inthe various parameters for 10-min wind velocity (Table 2). Theweights and fluxes of the settled sand and dust are positively

e components (vol.%) 10-min wind velocity (m s−1)

b10 μm 2–63 μm N63 μm Mean Max. Min. Cva Direction

9.03 33.77 63.13 16.5 20.3 15.0 0.12 WNW9.98 41.42 54.44 15.2 17.3 13.0 0.08 NNW8.23 31.07 65.89 19.2 22.0 15.0 0.28 W9.17 29.54 67.08 15.7 18.0 13.3 0.09 WNW6.83 26.02 71.39 16.1 19.0 12.0 0.11 NW5.35 21.56 75.95 17.3 20.0 14.0 0.11 NW

n sequence, and is calculated as follows:

city.

n value of parameter i.

Fig. 6. The relationships between the weights and fluxes of the settled sand and dust and the mean 10-min wind velocities and the variation coefficients (Cv) of the windvelocity sequences during the six typical dust storms.

87M. Qiang et al. / Powder Technology 175 (2007) 82–89

correlated with the mean 10-min wind velocity (r2 =0.68 and0.88, respectively) during the six events (Fig. 6a,b), indicatingthat for the typical dust storms the sand and dust depositionincrease with the wind velocity. Variation coefficient (Cv) forthe 10-min wind velocity sequences was calculated, whichindicates the amplitude of the fluctuation in the 10-minsequence. For event 8, the wind velocity varied between 13.0and 17.3 m s−1, with a small amplitude (Cv=0.08). In contrast,the wind velocity during event 15 had a higher Cv (0.28) andhigher wind velocity (Table 2). For the six typical dust storms,the sand and dust weights and fluxes also exhibit positivelycorrelations with the variation coefficients of wind velocities(r2 =0.71 and 0.78, respectively) (Fig. 6c,d). Hence, the changesin the weights and fluxes of settled sand and dust during thetypical dust storm are related not only to the variations in windstrength, but also to the amplitude of the fluctuation in windvelocity.

Grain size results of the settled sand and dust during twoseasons fromNovember 2003 to January 2004 and fromFebruaryto April 2004 are shown in Table 3 and Fig. 7. The two samplescontain a great lot of sand components, and the N63 μm fractionsaccount for 81.4% and 74.6% for the winter and spring,respectively. The fine-grained components in these samples arerelatively low, and the b10 μm fractions only account for 6.5%and 9.3% for the two seasons (Table 3). The median grain size ofthe two samples are 137.9 and 124.7 μm, which is three to five

Table 3Grain size results of the settled sand and dust during two seasons (November2003 to January 2004 and February to April 2004)

Season b2 μm(%)

b10 μm(%)

N63 μm(%)

Median grain size(μm)

Nov. 2003–Jan. 2004 1.5 6.5 81.4 137.9Feb. 2004–Apr. 2004 2.1 9.3 74.6 124.7

times the median reported for Lanzhou, in Gansu Province about1400 km from the site of the present study [12]. This differencepresumably resulted from the LMS being closer to the source areathan Lanzhou. Since sand and dust accumulation in the northernQaidamBasin is mainly attributed to dust storm and blowing dust(Fig. 2), the fine-grained particles would be carried fartherdownwind to the southeastern part of the basin or transportedfarther to the southeast under stronger wind conditions or by high-level westerly circulation during these events [7], and thenorthwest or west winds are prevailing in this region (Table 2).The settled sand and dust during the typical dust storms mainlycontains sand fraction as the seasonal deposition. The N63 μmfractions vary between 54.44% and 75.95%, and the b10 μmfactions are less than 10% (Table 2). Moreover, Relationshipsbetween the mean 10-min wind velocities and variationcoefficients of the wind sequences during five events and thedifferent grain size components of the settled materials wereanalyzed. Event 15 was not included in this analysis because theevent is so intensive that it produced the maximum sand and dust

Fig. 7. The grain size distribution of sand and dust samples collected during twoperiods (November 2003 to January 2004 and February to April 2004) at theLenghu Meteorological Station.

Fig. 8. The relationships between the mean 10-min wind velocities during the five typical dust storms (not including event 15 in Table 2) and (a) N63 μm fractions,(b) 2–63 μm fractions, (c) b10 μm fractions and (d) b2 μm fractions of the settled sand and dust during these events.

88 M. Qiang et al. / Powder Technology 175 (2007) 82–89

of the monitoring sequence during 34 min, which perhapsaffected the grain size distribution of the sand and dust, andweaken the relationships. Nonetheless, the obvious relationshipswere found between the wind strength and fluctuation amplitudeand the different components of sand and dust for the other fiveevents (Figs. 8 and 9). The coarse fractions (sand components,N63 μm) increase with increases in the wind strength and Cv

(Figs. 8a and 9a). In contrast, the fine-grained particles aredecrease with increases in the wind strength and Cv (Figs. 8b,c,dand 9b,c,d). It is worthy to note that with the increasing strengthand fluctuation of winds the sand component is effectively settled,and the fine-grained particles may be transported leeward. In fact,

Fig. 9. The relationships between the variation coefficients (Cv) of 10-min wind veTable 2) and (a) N63 μm fractions, (b) 2–63 μm fractions, (c) b10 μm fractions and

loess deposits mainly containing silt fraction are widelydistributed in the Xining region located about 300 km southeastto the Qaidam Basin [4] (Fig. 1). Therefore, the Qaidam Basinactually is a potential dust source area for the global dust loadingas proposed by other researchers [16,19].

4. Conclusions

This study investigated the dynamic impacts of wind velocityon the deposition of settled sand and dust during dust stormsunder such a hyperarid environment of the north Qiadam Basin.Low precipitation there hardly causes seasonal variations of the

locity sequences during the five typical dust storms (not including event 15 in(d) b2 μm fractions of the settled sand and dust during these events.

89M. Qiang et al. / Powder Technology 175 (2007) 82–89

surface conditions, and the wind strength is a key factor affectingthe dust emission, transport and deposition, which is differentfrom other regions. Rapid increases in wind velocity perhapsincrease the atmospheric dust loading and thus decrease visibility.The monthly fluxes of sand and dust deposition in the study areaare linearly related to the variations of monthly extreme windvelocities, and the weight of settled sand and dust during duststorms exhibited a positive correlation with the mean 10-minwind velocity during these events. Moreover, within the durationof a dust storm, the weight and flux of settled sand and dust willlinearly increase with the increasingwind strength and fluctuationamplitude. The coarse fraction (N63 μm) is effectively settledunder stronger and more variable wind conditions, however,under which the fine-grained particles (b63 μm) settle a little atthe study site and are transported leeward. The results obtained inthe present study demonstrate that wind regime is a dynamicfactor that governs dust transportation and deposition in the studyarea; the Qaidam Basin is a potential dust source area. To furtherinvestigate the impacts of wind velocity on dust deposition, moreobservation sites should be monitored along the modern duststorm and blowing dust tracks in this area. Furthermore, a possiblelinkage between this dust source area and remote dust depositionsites might be established on lager scale of atmospheric circu-lations since the wind conditions have an important role on thesand and dust deposition in this region.

Acknowledgements

The authors thank members of the Lenghu MeteorologicalBureau of Qinghai Province for their aid in collecting samples.This research was supported by the National Science Founda-tion of China (NSFC) (grants 40301051 and 40421101).

References

[1] R.A. Bagnold, The Physics of Blown Sand and Desert Dunes, WilliamMorrow and Company, Newyork, 1941, pp. 104–106.

[2] P.E. Biscaye, P.E. Grousset, M. Revel, S. Van der Gaast, G.A. Zielinski, A.Vaars, G. Kukla, Asian provenance of glacial dust (stage 2) in the GISP2ice core, Summit, Greenland, Journal of Geophysical Research 102 (C12)(1997) 26765–26781.

[3] A.J.-M. Bory, P.E. Biscaye, A. Svensson, F.E. Grousset, Seasonalvariability in the origin of recent atmospheric mineral dust at NorthGRIP,Greenland, Earth and Planetary Science Letters 196 (2002) 123–134.

[4] F.H. Chen, M.R. Qiang, Z.D. Feng, H.B. Wang, J. Bloemendal, Stable EastAsian monsoon climate during the Last Interglacial (Eemian) indicated bypaleosol S1 in the western part of the Chinese Loess Plateau, Global andPlanetary Change 36 (2003) 171–179.

[5] H.J.W. De Baar, J.T.M. De Jong, D.C.E. Bakker, B.M. Löscher, C. Veth, U.Bathmann, V. Smetacek, Importance of iron for plankton blooms and carbondioxide drawdown in the Southern Ocean, Nature 373 (1995) 412–415.

[6] E. Derbyshire, X.M. Meng, R.A. Kemp, Provenance, transport andcharacteristics of modern aeolian dust in western Gansu Province, China,and interpretation of the Quaternary loess record, Journal of AridEnvironments 39 (1998) 497–516.

[7] X.M. Fang, Y.X. Han, J.H. Ma, L.C. Song, S.L. Yang, X.Y. Zhang, Duststorms and loess accumulation on the Tibetan Plateau: A case study of dustevent on 4 March 2003 in Lhasa, Chinese Science Bulletin 49 (9) (2004)953–960.

[8] D.A. Gillette, Fine particulate emissions due to wind erosion, Transactionsof the American Society of Civil Engineers 20 (1977) 890–987.

[9] D.A. Gillette, R. Passi, Modeling dust emission caused by wind erosion,Journal of Geophysical Research 93 (D11) (1988) 14233–14242.

[10] N. Kumar, R.F. Anderson, R.A. Mortlock, P.N. Froelich, P. Kubik, S.Dittrich-Hannen, M. Suter, Increased biological productivity and exportproduction in the glacial Southern Ocean, Nature 378 (1995) 675–680.

[11] T. Littmann, Dust storm frequency in Asia: climatic control and variability,International Journal of Climatology 11 (1991) 393–412.

[12] L.Y. Liu, P.J. Shi, S.Y. Gao, X.Y. Zou, H. Erdon, P. Yan, X.Y. Li, W.Q. Ta,J.H. Wang, C.L. Zhang, Dustfall in China's western loess plateau asinfluenced by dust storm and haze events, Atmospheric Environment 38(2004) 1699–1703.

[13] H.Y. Lu, Z.S. An, An experimental study on the impacts on measurementof the grain size of loess sediment from pretreatment, Chinese ScienceBulletin 42 (23) (1997) 96–101 (in Chinese).

[14] W.G. Nickling, J.A. Gillies, Emission of fine grained particulates fromdesert soils, in: M. Leinen, M. Sarnthein (Eds.), Paleoclimatology andPaleometeorology: Modern and Past Patterns of Global AtmosphericTransport, Kluwer Academic Publishers, Dordrecht, 1989, pp. 133–165.

[15] M. Nishikawa, I. Mori, M. Morita, Q. Hao, H. Koyanagi, K. Haraguchi,Characteristics of sand storm dust sampled at an originating desert—caseof the Taklamakan Desert, Journal of Aerosol Science 31 (Suppl. 1) (2000)S755–S756.

[16] F.M. Phillips, M.G. Zreda, T.L. Ku, S.D. Luo, Q. Huang, D. Elmore, P.W.Kubik, P. Sharma, 230Th/234U and 36Cl dating of evaporite deposits fromwestern Qaidam Basin, China: Implications for glacial-period dust exportfrom Central Asia, Geological Society of America Bulletin 105 (1993)1606–1616.

[17] K. Pye, Aeolian Dust and Dust Deposits, Academic Press, London, 1987,p. 334.

[18] M.R. Qaing, F.H. Chen, J.W. Zhang, R.P. Zu, M. Jin, S. Xiao, Grain size insediments from Lake Sugan: a possible linkage to dust storm events at thenorthern margin of the Qinghai–Tibetan Plateau. Environmental, Geology51 (7) (2007) 1229–1238, doi:10.1007/s00254-006-0416-9.

[19] X.F. Qiu, Y. Zeng, Q.L. Miao, Temporal and spatial distribution, sourcesand movement paths of the sand storms in China, Acta Geographica Sinica56 (13) (2001) 316–322 (in Chinese).

[20] D.K. Rea, S.A. Hovan, Grain size distribution and depositional processesof the mineral component of abyssal sediments: lessons from the NorthPacific, Paleoceanography 10 (1995) 251–258.

[21] D.K. Rea, H. Snoeckx, H. Joseph, Late Cenozoic eolian deposition in theNorth Pacific: Asian drying, Tibetan uplift, and cooling of the northernhemisphere, Paleoceanography 13 (3) (1998) 215–224.

[22] Y.P. Shao, M.R. Raupach, P.A. Findlater, The effect of saltationbombardment on the entrainment of dust by wind, Journal of GeophysicalResearch 98 (1993) 12719–12726.

[23] D.H. Sun, F.H. Chen, J. Bloemendal, R.X. Su, Seasonal variability ofmodern dust over the Loess Plateau of China, Journal of GeophysicalResearch 108 (D21) (2003) 4665–4674, doi:10.1029/2003JD003382.

[24] S.Z. Sun, Concerning the vegetation Chinese regionalization map as a partof the physical geographical atlas of the Peoples Republic of China, ActaPhytoecologica Sinica 22 (6) (1998) 523–537 (in Chinese).

[25] I. Watts, Climates of China and Korea, in: H. Arakawa (Ed.), Climates ofNorthern and Eastern Asia, World Surv. Climatol., vol. 8, Elsevier, NewYork, 1969, pp. 1–118.

[26] Z. Wu, Aeolian Geomorphology, Science Press, Beijing, 1987, pp. 18–68,(in Chinese).

[27] L.Y. Yung, T. Lee, C.H. Wang, Y.T. Shieh, Dust: A diagnostic of thehydrologic cycle during the Last Glacial Maximum, Science 271 (1996)962–963.

[28] K.C. Zhang, J.J. Qu, R.P. Zu, H.Y. Fang, Temporal variations ofsandstorms in Minqin oasis during 1954–2000, Environmental Geology49 (2005) 332–338.

[29] R.J. Zhang, M.X. Wang, X.Y. Zhang, G.H. Zhu, Analysis on the chemicaland physical properties of particles in a dust storm in spring in Beijing,Powder Technology 137 (2003) 77–82.