Analysis on the temporal-spatial distribution character and effect factors of PM10 in the hinterland of Taklimakan Desert and surrounding area

XinChun Liu , YuTing Zhong , Qing He , XingHua Yang ,Ali Mamtimin , Wen Huo

1. Institute of Desert Meteorology, CMA, Xinjiang Laboratory of Tree Ring Ecology, Key Laboratory of Tree-ring Physical and Chemical Research of China Meteorological Administration, Urumqi, Xinjiang 830002, China

2. Desert Atmosphere and Environment Observation Experiment of Taklimakan Station, Tazhong, Xinjiang 831000, China.

Analysis on the temporal-spatial distribution character and effect factors of PM10in the hinterland of Taklimakan Desert and surrounding area

XinChun Liu1,2*, YuTing Zhong1,2, Qing He1,2, XingHua Yang1,2,Ali Mamtimin1,2, Wen Huo1,2

1. Institute of Desert Meteorology, CMA, Xinjiang Laboratory of Tree Ring Ecology, Key Laboratory of Tree-ring Physical and Chemical Research of China Meteorological Administration, Urumqi, Xinjiang 830002, China

2. Desert Atmosphere and Environment Observation Experiment of Taklimakan Station, Tazhong, Xinjiang 831000, China.

In recent years, the physical and chemical properties of dust aerosols from the dust source area in northern China have attracted increased attention. In this paper, Thermo RP 1400a was used for online continuous observation and study of the hinterland of Taklimakan, Tazhong, and surrounding areas of Kumul and Hotan from 2004 to 2006. In combination with weather analysis during a sandstorm in the Tazhong area, basic characteristics and influencing factors of dust aerosol PM10have been summarized as below: (1)The occurrence days of floating dust and blowing dust appeared with an increasing trend in Kumul, Tazhong and Hotan, while the number of dust storm days did not significantly change. The frequency and intensity of dust weather were major factors affecting the concentration of dust aerosol PM10in the desert. (2) The mass concentration of PM10had significant regional distribution characteristics, and the mass concentration at the eastern edge of Taklimakan, Kumul, was the lowest; second was the southern edge of the desert,Hotan; and the highest was in the hinterland of the desert, Tazhong. (3) High values of PM10mass concentration in Kumul was from March to September each year; high values of PM10mass concentration in Tazhong and Hotan were distributed from March to August and the average concentration changed from 500 to 1,000 μg/m3, respectively. (4) The average seasonal concentration changes of PM10in Kumul, Tazhong and Hotan were: spring ＞ summer ＞ autumn ＞ winter; the highest average concentration of PM10in Tazhong,was about 1,000 μg/m3in spring and between 400 and 900 μg/m3in summer, and the average concentration was lower in autumn and winter, basically between 200 and 400 μg/m3. (5) PM10concentration during the sandstorm season was just over two times the concentration of the non-sandstorm season in Kumul, Tazhong and Hotan. The average concentrations of sandstorm season in Tazhong were 6.2 and 3.6 times the average concentrations of non-sandstorm season in 2004 and 2008, respectively. (6) The mass concentration of PM10had the following sequence during the dust weather: clear day ＜ floating dust ＜ floating and blowing dust ＜ sandstorm.The wind speed directly affects the concentration of PM10in the atmosphere, the higher the wind speed, the higher the mass concentration. Temperature, relative humidity and barometric pressure are important factors affecting the strength of storms, which could also indirectly affect the concentration change of PM10in the atmosphere.

dust aerosol; dust weather; mass concentration; effect factors; Taklimakan Desert

1. Introduction

Atmospheric aerosols and gaseous atmospheric pollutants are two basic factors affecting air quality. Atmospheric aerosol pollutants in China are very serious and have attracted increased attention, especially air quality in the sandstorm source area and impacted large cities. Aerosol particles suspended in the atmosphere not only pose hazards to human health, but also produce a significant effect on global and regional climate and environmental systems (Liet al.,2005; Li, 2006).

When sandstorms and blowing sand weather occurs, the wind propels large amounts of dust and sand into the air,causing serious air pollution for the local environment. Fine particles in the dust and sand are hoisted to high-altitudes and spread over long distances with the upper airflow, which impact air quality over a wide range. Floating dust weather caused by sandstorms and blowing sand includes two types:one is formed by suspended, small sized dust and sand particles which have not settled after occurrence of the local sand and dust weather; the other is formed by fine dust and sand particles which is hoisted upwards in the source area by the dust and sand weather, and then spreads with the upper airflow, and falls to the ground in distance areas. In northwestern China, the Taklimakan Desert is the main source area of dust weather which produces serious air pollution for the local area or even the whole northwestern region. At present,there are numerous studies on the concentration, composition and its impact on of atmospheric aerosol PM10for the mainland and coastal cities of China (Kanget al., 2002;Zhanget al., 2003, 2009; Guoet al., 2004; Xiet al., 2004;Huoet al., 2006; Liet al., 2006; Liuet al., 2008; Caoet al.,2009; Junet al., 2009). Studies on dust weather in the Taklimakan Desert and surrounding areas focus mainly on production conditions and variation characteristics (Hanet al., 2005; Li, 2007; Liet al., 2007, 2008; Zhanget al., 2007;Chenet al., 2008; Wanget al., 2008). There is limited data on concentration changes, chemical composition and influencing factors of dust aerosols (Shenet al., 2006; Gaoet al.,2008; Yueet al., 2009). In this paper, Thermo RP 1400a was used for online continuous observation and study of the hinterland of Taklimakan, Tazhong, and surrounding areas of Kumul and Hotan, from 2004 to 2009. Also, basic characteristics and influencing factors of dust aerosol PM10in this area is summarized. This study will contribute to further understanding of distribution characteristics and the influence degree on ambient air quality of dust aerosols in the dust and sand source areas and surrounding cities. Also, this study provides a scientific basis for environmental pollution improvements in this area.

2. Research methods

2.1. Observation instruments and observation areas

In this study, a Thermo RP 1400a, type PM10automonitor (Ample Technology Center, USA) was used to observe dust aerosol particles in the atmosphere. The observation areas were located in the hinterland of Taklimakan Desert,Tazhong, and the surrounding cities of Kumul and Hotan.The observation sites in Kumul, Tazhong and Hotan are selected as Kumul Meteorological Bureau (42°49'N,93°31'E, elevation 737.0 m), Tazhong Meteorological Station (38°58′N, 83°39′E, elevation 1,090.0 m) and Hotan Meteorological Bureau (37°08′N, 79°56′E, elevation 1,374.7 m), respectively. The Tazhong station is built in the hinterland of Taklimakan Desert, about 200 km deep into the desert, which is the first Chinese Desert Meteorological Station built in this region. This station undertook the task of monitoring desert weather and climate, and the resulting data fills a blank for weather data in the Taklimakan Desert.The observation site in Kumul is located at the eastern margin of the Tarim Basin, where the impact of strong sand-dust weather from the east on dust aerosols is obvious.The observation site in Hotan is located at the southwestern margin of the Tarim Basin and the impact of dust weather on this area is obvious. All PM10monitors were placed in a room of the sandstorm observation station, and the intake pipe was switched and extended to the roof,about 1.5 m above the roof and about 4 m above the ground. Observational data from the monitor is representative of the surrounding topography of the station.

2.2. Observation method and data processing

In this study, the specific observation periods are as follows: the observation of PM10in Kumul was from January 13, 2004 to December 31, 2009, during which no continuous absence of measurements occurred, and the longest absence of measurements did not exceed 48 h and the observation data had good continuity. The observation period in Tazhong was from January 8, 2004 to December 31, 2009,with loss of data from June 27, 2004 to January 25, 2005 and April 27, 2005 to August 14, 2005 due to observation instrument failure, but the other observation periods were continuous. The observation period in Hotan was from January 9, 2004 to December 31, 2009, with loss of data from August 26, 2004 to January 18, 2005, July 1 to July 6, 2005,January 18, 2006 to February 1, and April 10, 2006 to April 23, 2006 due to observation instrument failure, but the other observation periods were continuous. Raw data of the PM10mass concentration was examined to exclude obvious incorrect data (mainly negative and missing PM10mass concentration data), in which Kumul had a total of 598,333 observation values, excluding 62,816; Tazhong had a total of 502,372 observation values, excluding 60,202; and Hotan had a total of 462,047 observation values, excluding 12,770.PM10was recorded once every five minutes, with hourly and daily average concentrations firstly obtained, and then monthly averages.

3. Analysis of results

3.1. Characteristics of dust weather in the study area

Statistical analysis of the number of sand-dust weather days from 2004 to 2009 in Kumul, Tazhong and Hotan shows that (Table 1): (1) occurring days of floating and blowing dust in Kumul, Tazhong and Hotan had an increasing trend, in which days of floating dust weather in Tazhong and Hotan were equivalent, and the least in Kumul; (2) occurring days of blowing sand in Tazhong were quite high compared to those in Kumul and Hotan, and Kumul had the least; (3) occurring days of sandstorms did not change significantly, with 2007 containing the least number of days for 2004-2009, and there was no sandstorms in Kumul. The frequency and intensity of dust weather are the main factors affecting dust aerosol particles.

Table 1 Total number of sand-dust weather days during 2004-2009 in Kumul, Tazhong and Hotan

The frequency of dust weather in Kumul from 2004 to 2009 was the least of the three study areas. The number of floating dust days was slightly higher than those of blowing sand, with no sandstorms, and no dust weather during 2004-2007. The number of consecutive dust and sand days gradually increased was two days, especially in 2009 when floating dust occurred 11 times and blowing sand occurred eight times. 2009 had the most historical observation data in Kumul.

The number of floating dust days in Tazhong changed dramatically, with an average of 143.2 days from 2004-2009; 2004 had the least (99 days), then it significantly increased after 2005, and 2007 had the most (157 days).The average days of blowing sand for 2004-2009 was 71.3 days, the annual change of blowing sand days presented a fluctuating increasing trend, and the highest value was achieved in 2009 (95 days) which is 1.83 times that of 2004(52 days). The average sandstorm days was 11.2 days, the annual change in sandstorms presented a slow fluctuating decreasing trend, and the number of sandstorm days in 2006 and 2008 were 14 days, respectively, which is five days more than the lowest in 2005. Occurring days of floating dust and blowing sand during dust weather in Tazhong shows an increasing trend, but the number of sandstorm days shows a decreasing trend. The variation characteristics of dust weather are greatly affected by changes of the surrounding environment of the Tazhong station. Since 2001,the surrounding green belts around Tazhong have continually expanded to various sand hill tops, forming a local basin microclimate. With an increase of oil exploration in Tazhong,large planting tracts have appeared along both sides of the desert highway and the eastern and western sand slopes of Tazhong. Increased human activity has caused damage to the underlying surface, while the occurrence of floating dust and blowing sand has increased, and the occurrence of sandstorms have been inhibited.

Hotan was one of the most serious affected areas suffering from dust weather disasters at the edge of the Taklimakan Desert, and such a proverb was widespread in this area: "People in Hotan are miserable, who eat half a catty of soil during the day and still finish the remainder at night", which indicates that the concentration of dust aerosols in Hotan’s atmosphere is very high. Floating dust weather in 2004-2009 changed very little, with an average floating dust day of 121.3 days,and more than one third of the time as floating dust weather.2006 and 2007 were years with the most, 134 and 133 days,respectively, and 2005 was the lowest with 107 days. The number of blowing sand days in 2004-2009 shows a slow increasing trend, with the lowest in 2006 and 2007 of 15 days,respectively, and the highest in 2008 of 25 days, exceeding the average value of 19.3 days. The number of sandstorm days in 2004-2009 rapidly changed with an average of 8.2 days, the lowest in 2007 of only one day, the second was five days in 2006, and the highest was 12 days in 2004. The frequency and intensity of dust weather are the main factors affecting dust aerosol concentrations in the hinterland of Taklimakan Desert and surrounding areas.

3.2. Mass concentration changes of dust aerosol PM10

3.2.1 Monthly average distribution and changes of PM10 mass concentration

PM10mass concentration had significant regional distribution characteristics, where Kumul at the eastern edge of the Taklimakan Desert had the lowest concentration, second was Hotan at the southern edge of the Taklimakan Desert,and the highest was at Tazhong in the hinterland of the desert (Figure 1). The regional distribution characteristics were mainly affected by the frequency and intensity of the dust weather. The average concentration variation diagram of Kumul in 2004-2009 shows that from March to September every year there was a high area of PM10mass concentration,the highest average concentration from June to August fluctuates around 200 μg/m3, and the concentrations in other months were lower. Annual average concentrations are slightly greater than 100 μg/m3. Concentration changes from March to September were mainly affected by windy and sandy weather, and that urban dust and plant dust may be the main sources. Winter heating has a certain influence on the PM10mass concentration, but the effect was not obvious.

Figure 1 Variation of PM10 monthly average concentrations of Kumul, Tazhong and Hotan from 2004 to 2009

The monthly average mass concentration in 2004 was the highest for 2004-2009, in which the concentration from June to August was particularly prominent; the concentration in July was 349.0 μg/m3, 1.9 times the average value of 184.3 μg/m3. August and June were 304.8 μg/m3and 271.7 μg/m3, respectively. In 2004-2009, the monthly average mass concentration in 2009 was the lowest with 111.9 μg/m3,which was much lower than in 2004 with the highest concentration in April of 234.9 μg/m3and the lowest in August of 61.7 μg/m3. Compared with 2009, there was no dust weather in 2004, although the number of dust days in 2009 was the highest, and total days were low, which indicates that dust aerosols brought in by dust weather was not the main source of PM10in Kumul. Thus, dust and plant dust within the city may be the main source.

There were two peak values in the variation of monthly average mass concentrations of Tazhong during 2004-2009,from March to May, and from July to August. Because dust weather in the desert mainly appears from March to August,this period has a high mass concentration of PM10, in particular, the concentration in every spring from March to May is very high. Dust weather is mainly affected by the system weather, and from June to August, the mass concentration of PM10is mainly affected by local weather. In addition to 2009,the highest value of PM10mass concentration appeared in May each year, and its concentration was essentially above 1,000 μg/m3. From annual changes, the monthly mass concentration in 2006 was basically the highest from 2004 to 2009, and the PM10average mass concentration from March to August during the windy and sandy season was 1,053.3 μg/m3. The average value in other months changed between 70 and 650 μg/m3. The monthly average mass concentrations were lower in 2008 and 2009, and PM10average mass concentrations from March to August were 694.4 μg/m3and 678.1 μg/m3, respectively, and the annual average mass concentrations were 444.5 μg/m3and 481.1 μg/m3,respectively.

High values of PM10mass concentration in Hotan was also distributed from March to August, and the average concentration changed between 500 and 1,000 μg/m3. Because Hotan is located at the southern edge of the Taklimakan Desert, PM10is mainly from dust aerosols formed during the dust weather, its concentration from March to August with frequent wind and sand is very high.During 2004-2009, the average concentrations of PM10were relatively higher in 2004 and 2005, 594.1 μg/m3and 567.0 μg/m3, respectively, with the monthly highest concentration in 2004 of 1,224.3 μg/m3, and the lowest January concentration of 218.2 μg/m3, which greatly exceeded the lowest value of PM10at API index of 500 (grade No.5 heavy pollution) (the daily average concentration of PM10was more than 600.0 μg/m3). PM10monitoring concentration indicated the pollution severity of the dust weather on the atmospheric environment. In contrast, 2009 was the year with lower average concentration, with the annual average concentration of 383.8 μg/m3, the highest of 754.1 μg/m3in July and the lowest of 116.6 μg/m3in December, which still far exceeded the capital city of Urumqi with higher aerosol concentration in China.

3.2.2 Seasonal variations of PM10 mass concentration

The seasonal average concentrations of PM10in Kumul,Tazhong and Hotan have the same change law, spring ＞summer ＞ autumn ＞ winter (Figure 2). The PM10concentration in Kumul is substantially lower than those in Tazhong and Hotan. During 2004-2009 the largest change occurred in summer, the second largest change occurred in winter,and fluctuation ranges were the smallest in spring and autumn. The average concentration in spring changed from the lowest of 145.4 μg/m3in 2009 to the highest of 187.5 μg/m3in 2004. Dramatic changes occurred in summer with the highest concentration of 308.5 μg/m3in 2004 which is 4.7 times the lowest of 66.0 μg/m3in 2009. The average concentration in autumn from 2004 to 2009 was 128.2 μg/m3without great changes every year, and the average concentration in winter was 91.1 μg/m3with relatively stable range of variation. Because particulate matter in the atmosphere in winter is mainly affected by urban heating, variation in PM10concentration could reflect relatively stable winter heating in Kumul.

Spring and summer in Tazhong had increased occurrence of dust and sand weather, with more than 70% of floating dust, more than 80% of blowing sand, and 100% of sandstorms. Thus, the average concentrations of PM10in Tazhong in spring and summer were very high, with especially high in spring. The average concentrations in spring in 2004 and 2006 were respectively equivalent to and more than 1,200 μg/m3, the highest was during 2004-2009. In 2007, the average mass concentration was 876.2 μg/m3,slightly higher than those in 2008 and 2009, and the lowest concentration of 410.5 μg/m3was in 2005. The average concentration of PM10in summer was basically lower than in spring, with the lowest of 406.3 μg/m3in 2005, which was equivalent to that in spring. 2007 was slightly higher than 2005, 2008 was slightly higher than 2009, and the highest was 870 μg/m3in 2006. Comparatively, autumn and winter are relatively better seasons in desert areas, with few dust weather, and the average concentration of PM10is lower,thus air quality is better. During 2004-2009, except for a higher average concentration of PM10in winter of 2005,there were no significant variation in autumn and winter,basically changing between 200 and 400 μg/m3.

The variation in average concentration of PM10during the four seasons in Hotan was very distinct, with significantly higher concentrations in spring and summer than those in autumn and winter; the average concentration in spring for 2004-2009 was 698.2 μg/m3, in summer was 555.6 μg/m3, and in winter was slightly higher than in autumn, 284.3 μg/m3and 253.5 μg/m3, respectively. The concentration was the highest of 891.4 μg/m3in spring of 2004,which is 1.85 times the lowest concentration of 2005; the average concentration in summer changed from the lowest of 439.1 μg/m3in 2009 to the highest of 740.9 μg/m3in 2005;the concentration changed slightly in autumn and winter,which was basically between 200 and 300 μg/m3. This analysis shows that the process of dust weather in spring and summer is the main factor causing PM10mass concentration,and winter heating is an important factor affecting variation of PM10mass concentration.

3.2.3 PM10 mass concentration variation in sandstorm and non-sandstorm seasons

From March to August every year is basically the sandstorm season in the Taklimakan Desert and its surrounding areas, and windy and dusty weather from September to February rarely occurs. PM10concentrations in sandstorm seasons in Kumul, Tazhong and Hotan are significantly higher than those in non-sandstorm seasons, and the concentrations are basically more than two times that of non-sandstorm seasons; the average concentrations in sandstorm seasons in 2004 and 2008 in Tazhong were 6.2 and 3.6 times that of non-sandstorm seasons, respectively (Table 2).

Figure 2 Variation of PM10 seasonal average concentrations of Kumul, Tazhong and Hotan from 2004 to 2009

Table 2 Variation of PM10 average concentrations between dust storm and non-dust storm seasons of Kumul,Tazhong and Hotan from 2004 to 2009 (μg/m3)

During 2004-2009, the average mass concentrations in sandstorm seasons in Kumul were lower than those in non-sandstorm seasons except for 2009, and the concentrations in other years were significantly higher. The performance was outstanding in 2004: PM10average mass concentration in sandstorm seasons was 248.0 μg/m3, 2.1 times that of non-sandstorm seasons, which is absolutely the same as in other years. The data in Table 2 shows that dust weather has only a slight impact on PM10mass concentration distribution and variation in Kumul. PM10average mass concentration in sandstorm seasons in Tazhong is significantly higher than in non-sandstorm seasons, the PM10average mass concentration in sandstorm seasons in 2004 was 1,230.2 μg/m3, higher than that of 1,030.6 μg/m3in non-sandstorm seasons, so as that in 2006, the PM10average mass concentration in sandstorm seasons was 1,053.3 μg/m3,3.6 times that of non-sandstorm seasons, and more than two times that of other years. The difference of PM10average mass concentrations between sandstorm and non-sandstorm seasons in Hotan was not as significant as in Tazhong, except that the PM10concentrations in the two seasons of 2005 were closer, and the differences in other years were very obvious; PM10average mass concentrations in sandstorm seasons changed between 500 and 800 μg/m3, and in non-sandstorm seasons changed between 200 and 300 μg/m3.The sandstorm season has significant effects on the variation of mass concentration of dust aerosol PM10in Hotan.

3.2.4 PM10 mass concentration variation in the process of dust weather

In this paper, a sandstorm that occurred in Tazhong from April 15, 2008 to April 23, 2008 was selected to analyze PM10mass concentration variation during the process of dust weather and impacts on weather conditions. Floating dust weather began on April 15 then gradually increased after April 16, and was accompanied by blowing sand on April 17 and on April 18. A sandstorm occurred on April 19,with decreasing dust and continued blowing sand on April 20. During April 21-22, floating dust was still present but visibility gradually increased, and on April 23 the dust dissipated, presenting a clear day.

The variation of PM10mass concentration reflected the process and intensity of the dust weather (Figure 3), the floating dust weather began to appear after April 17, and PM10concentration in the atmosphere basically remained at about 400 μg/m3, 3 to 4 times that of Urumqi’s concentration. The amount of fine particulate in the atmosphere gradually increased in the evening of April 18 with the incidence of blowing sand, and PM10mass concentration levels gradually increased. On the morning of April 19, a strong sandstorm began to appear, and PM10mass concentration rapidly increased to its first and largest peak value of 8,339.0 μg/m3at 10:00, nearly 20 times that of the hourly average value for April 17-18. At the end of the sandstorm, atmospheric fine particulate matter gradually settled, and PM10mass concentration decreased rapidly, reaching a normal level at about 20:00, April 19 until 5:00, April 20. Intermittent blowing sand appeared again, and PM10concentrations fluctuated during which two peak values appeared, respectively, at 11:00 (1,160.4 μg/m3) and 14:00 (7,606.4 μg/m3). From 15:00 in the afternoon of April 20, blowing sand was absent,floating dust gradually weakened, and PM10concentration decreased to a normal level.

Figure 3 The hourly average mass concentration of PM10 in dust weather from April 17 to 23, 2008

3.3. Impact of weather conditions on dust aerosol (PM10)concentration

To study the impact of weather conditions on dust aerosol concentration variation, we selected Tazhong as a key focus. Dust weather occurred during April 17 to April 23,2008. The wind speed suddenly decreased before formation of a sandstorm, or even in a calm wind condition. The average wind speed was 0 on April 17, then suddenly increased to 1.8 m/s on April 18; during the sandstorm formation stage,the wind speed suddenly increased and soon reached the critical wind velocity to form the sandstorm (Liet al., 2007).The dust critical wind velocity of sandstorms in the Taklimakan Desert is 8.0 m/s (an average wind speed at a height of 10-m above the surface within 10 min) (Liet al.,2007). In this experiment, the wind speed at the observation site was 2-m-height wind speed and the dust critical wind velocity at a height of 2-m was converted to 5.6 m/s according to the empirical formula. The daily average wind speed on April 19 was 3.6 m/s, but the instantaneous wind speed was up to 7.2 m/s. The wind speed during the sandstorms sustainable stage fluctuated slightly and was generally higher than the critical wind velocity. On April 19, during the sandstorms sustainable stage, wind speed stabilized at about 6.0 m/s, and during the sandstorms regression stage, wind speed decreased slowly and gradually fell below the critical wind velocity. On April 20, the wind speed gradually weakened to 2.5 m/s. The wind speed was directly related to production of dust weather, and dust weather has a direct impact on the concentration of particulate matter in the atmosphere, so the greater the wind speed, the higher the concentration of particle matter in the atmosphere.

From April 17 to 23, the dominant wind during the dust weather was northerly, the wind direction during the sandstorms early formation stage was NNW, the wind direction shifted north during the sandstorms formation, then shifted to NNE, E and ENE, and then gradually became consistent with the main wind direction. Because the observation site was located in the hinterland of Taklimakan Desert, where lacks all obstacles, the wind direction did not directly affect measurement results of atmospheric concentrations of fine particulate matter. The change of weather conditions has a significant impact on the atmospheric accumulation of fine particle PM10and the concentration of particulate matter in weather events follows order as: clear day ＜ floating dust ＜floating and blowing dust ＜ sandstorm. Before and after the sandstorms formation, the corresponding changes of temperature, relative humidity and barometric pressure were observed, which are important factors affecting the strength of the sandstorm, and thus the distribution of atmospheric particulate matter concentration was indirectly affected.

4. Conclusions

In this paper, observation data of dust aerosol PM10in the hinterland of Taklimakan Desert, Tazhong, and surrounding areas of Kumul and Hotan, were used in combination with weather data during the process of a sandstorm in Tazhong. Also, spatial and temporal distribution characteristics and the influencing factors of dust aerosol PM10were analyzed to obtain the following conclusions:

(1) The incidence days of floating dust and blowing sand in Kumul, Tazhong and Hotan presented an increasing trend,in which the days of floating dust in Tazhong and Hotan were equivalent, and the incidence days were the least in Kumul; the days of blowing sand in Tazhong were significantly higher than in Kumul and Hotan, and the least in Kumul; sandstorm days did not change significantly, the incidence days in 2007 were the least for 2004-2009, and no sandstorms occurred in Kumul. The frequency and intensity of dust weather were the main factors affecting dust aerosol concentration in the desert and surrounding areas.

(2) PM10mass concentration has significant regional distribution characteristics, the concentration in Kumul at the eastern edge of the Taklimakan Desert was the lowest of the three study areas, the second was Hotan at the southern edge of the desert, and the highest was in the hinterland of the desert, Tazhong; the regional distribution characteristics were mainly affected by the frequency and intensity of dust weather.

(3) From March to September every year experienced high PM10mass concentration values in Kumul, the highest average concentration from June to August fluctuated at 200 μg/m3, the concentrations in other months were lower, and the annual average concentration was slightly greater than 100 μg/m3; the high PM10mass concentration values in Tazhong and Hotan were distributed from March to August,the average concentration changed in range of 500-1,000 μg/m3, and the annual average concentration in Tazhong was higher than in Hotan with concentrations respectively distributed in range of 400-900 μg/m3and 300-600 μg/m3.

(4) PM10seasonal average concentrations in Kumul, Tazhong and Hotan have the same variation law, spring ＞summer ＞ autumn ＞ winter; the average concentration of PM10was the highest in Tazhong and changed at about 1,000 μg/m3in spring, between 400 and 900 μg/m3in summer, and the concentrations in autumn and winter were lower and basically changed between 200 and 400 μg/m3.

(5) PM10concentrations during the sandstorm seasons in Kumul, Tazhong and Hotan were significantly higher than concentrations in non-sandstorm seasons. The concentrations in sandstorm seasons were basically more than two times of those in non-sandstorm seasons; the average concentrations in sandstorm seasons in 2004 and 2008 were,respectively, 6.2 and 3.6 times of those in non-sandstorm seasons.

(6) During the process of dust weather, PM10mass concentration variation has the following rule: clear day ＜floating dust ＜ floating and blowing dust ＜ sandstorm. Wind speed directly affects the concentration of PM10in the atmosphere, the higher the wind speed, the higher the mass concentration. Temperature, relative humidity and barometric pressure are important factors affecting the strength of storms, which also indirectly affect the concentration variation of PM10in the atmosphere.

This project is supported by Natural Science Foundation of China (41175017, 41175140) and Public Service Sectors(Meteorology) Research and Special Funds (GYHY201006012,GYHY201106025). We also thank anonymous reviewers for English improvements.

Cao JJ, Shen ZX, Judith CC, 2009. Seasonal variations and sources of mass and chemical composition for PM10aerosol in Hangzhou, China. China Particuology, 7: 161-168.

Chen X, Wei WS, Liu MZ, 2008. Influence of sand-dust over Tarim Basin on shortwave radiation: a case study in Tazhong area. Journal of Desert Research, 28(5): 920-926.

Gao WD, Yuan YJ, Liu ZH, Wei WS, 2008. Status of dust sources and aerosol formatting condition analysis in Xinjiang. Journal of Desert Research, 28(5): 969-973.

Guo JH, Rahn KA, Zhuang GS, 2004. A mechanism for the increase of pollution elements in dust storms in Beijing. Atmospheric Environment,38: 855-862.

Han ZW, Wang T, Dong ZB, Wu QJ, Yao ZY, 2005. Spatial-temporal distribution of blown sand activities along Taklimakan Desert highway. Scientia Geographica Sinica, 18(4): 490-496.

Huo W, Li X, Ali Mamtimin, Wang J, Zhao XC, 2006. Analysis on the features of sandstorms in the Tarim Basin in spring 2004. Arid Zone Research, 23(2): 210-215.

Jun T, Ho KF, Chen LG, 2009. Effect of chemical composition of PM2.5on visibility in Guangzhou, China, 2007 spring. Particuology, 7: 68-75.

Kang XF, Zhang JH, Liu HH, 2002. Influences of sandy weather on city’s air quality in China: scope and scale. Resources Science, 24(4): 1-4.

Li F, 2007. Research progresses of trace methods and distribution of dust source areas. Arid Zone Research, 24(3): 335-338.

Li JC, Dong ZB, Wang XM, He SS, 2008. Seasonal distribution and causes of dust events in Tarim Basin, China. Journal of Desert Research, 28(1):142-148.

Li SY, Lei JQ, Xu XW, 2007. Genetic conditions of sandstorms in the hinterland of Taklimakan Desert. Journal of Applied Meteorological Science,18(4): 490-496.

Li X, Shen Z, Cao JJ, 2006. Distribution of carbonaceous aerosol during spring 2005 over the Horqin Sandland in northeastern China. China Particuology, 4(6): 316-322.

Li X, 2006. Effects of meteorological factors on optical properties of aerosol in Urumqi. Arid Zone Research, 23(3): 484-488.

Li Y, Zhang XY, Che HZ, Sun JY, 2005. Comparison test of GRIMM1.107 with RP2000 in Beijing City. China Powder Science and Technology, 3:8-10.

Liu ZD, Liu JG, Wang B, 2008. Observation aerosol observation in Fengtai area, Beijing. Particuology, 6: 214-217.

Shen JG, Liu F, Niu SJ, Jiang XG, 2006. Successive observation and analysis of TSP concentration in a sand-dust storm case. Journal of Desert Research, 26(5): 786-791.

Wang MZ, Wei WS, Yang LM, Li YH, Xiao SJ, 2008. Analysis on circulation dynamical structure of a strong sand-dust storm case from east in Tarim Basin. Journal of Desert Research, 28(3): 370-376.

Xi XX, Li J, Quan JN, 2004. Study on the mass concentration of the atmospheric dust and aerosol over Lanzhou city in spring. Arid Zone Research,21(2): 112-116.

Yue P, Niu SJ, Shen JG, Ge ZP, 2009. Observation and analysis of micro meteorology parameters and PM10for an ultra strong dust-storm. Journal of Natural Disasters, 18(1): 118-123.

Zhang RJ, Han ZW, Cheng TT, 2009. Chemical properties and origin of dust aerosols in Beijing during springtime. Particuology, 7: 61-67.

Zhang RJ, Xu YF, Han ZW, 2003. Inorganic chemical composition and source signature of PM2.5in Beijing during ACE-Asia period. Chinese Science Bulletin, 48(10): 1002-1005.

Zhang ZC, Dong ZB, Zhao AG, Han LY, Qian GQ, Luo WY, 2007. Features of sand drift movement in desert—a case study at sand drift observation station of Chinese Academy of Sciences. Arid Zone Research, 24(4):550-555.

10.3724/SP.J.1226.2011.00526

*Correspondence to: XinChun Liu, Associate professor of Institute of Desert Meteorology, CMA. No. 46, Jianguo Road, Urumqi,Xinjiang 830002, China. Tel: +86-991-2651450; Fax: +86-991-2621387; Email: liuxinchun2001@163.com

18 May 2011 Accepted: 14 August 2011