Interannual and seasonal variations of sand-dust events in Tarim Basin, China

Dust events frequently occur in Tarim Basin (TB), China. However, research is scarce on the variation characteristics of sand–dust, oating dust, sand blowing, and sandstorms in the mountains, oases, and deserts in TB. Therefore, we conducted in–depth studies on the distributions and variations of dust events in TB. The results indicate that dust events tend to increase from south to north. The yearly sand–dust event occurrence rate trend is deserts > oases > mountainous regions. In spring, sand–dust, oating dust, sand blowing, and sandstorm days account for 35.9 %, 36.0 %, 38.5 %, and 47.1 % of the whole year, while in summer they account for 33.6 %, 33.5 %, 43.8 %, and 60.3 % of the whole year. From 1961 to 2015, most stations (annual average > 30 stations) showed a downward linear trend in sand dust, oating dust, sand blowing, and sandstorms in TB. All types of sand–dust events in mountains and oases decreased substantially after 1975 to 1978. In spring and summer, the downward dust trends are the most obvious, which the largest annual average downward linear trend rates of oating dust, sand blowing, and sandstorms are -0.33 d/yr (QM station), -0.21 d/yr (TGLK station), and -11 d/yr (KP station), respectively, in spring, and -0.28 d/yr (QM station), -0.17 d/yr (TGLK station), and -0.14 d/yr (KP station), respectively, in summer. It is found that the decreasing and increasing trends of oating dust are the most obvious during sand–dust events. There is a signicant positive correlation between wind speed and relative humidity in mountains, and a signicant positive correlation between wind speed in oases. d/yr (AHQ station), 42.7 d/yr (WQ station), 14 d/yr (AHQ station), and 8 d/yr (AHQ station), respectively; in oasis regions, they are 195.8 d/yr (HT station), 193.6 d/yr (HT station), 66 d/yr (PS station), and 38 d/yr (XT station), respectively; and in desert regions, they are 48.9 d/yr, 40.6 d/yr, 23 d/yr, and 6 d/yr, respectively. These results indicate that oating dust accounts for the highest ratio of dust events in TB, comprising 87.7 %, 98.9 %, and 83.3 % in mountain, oasis, and desert regions, while sandstorms account for the lowest occurrence of dust events in TB, accounting for 16.4 %, 19.4 %, and 12.3 %, respectively, in the three regions. south of hinterland TB), TB),

smaller than 10 microns) on dust days increasing by 1 to 173 % compared with non-dust days, while the mass concentration of PM 2.5 (i.e. particulate matter with diameters smaller than 2.5 microns) increasing by 21 to 172 % (Li et al., 2018). Therefore, it is necessary to understand the basic climatic characteristics of dust weather in the TD and surrounding areas. For example, Yang et al. (2015) analyzed observations of blowing dust events in the TD using 10 desertedge meteorological stations and two inner-desert meteorological stations. Pi et al. (2017) showed atmospheric dust events in central Asia, including a discussion on what is driving their ndings. Although scholars have studied the dust weather in the TD, thorough analysis on the seasonal variation of different types of sand-dust events of the underlying surface has not yet been performed in Tarim Basin (TB). Therefore, the spatial and temporal variations of sand-dust, oating dust, sand blowing, and sandstorm are analyzed here using data obtained at 42 meteorological stations (4, 1, and 37, respectively, from mountain, desert and oasis stations) in the TB. The analysis of the TB is helpful for further our understanding of the distribution characteristics of dust events. The remainder of this paper is organized as follows. In Sect. 2, we describe the study area and data. In Sect. 3, we present the spatial and temporal variations of dust events of the TB, including sand-dust, oating dust, sand blowing, and sandstorm.
Finally, we give our conclusions based on this analysis in Sect. 4.

Study Area And Data
TB, the largest inland basin in China, lies between the Tianshan and Kunlun Mountains. Its widest extent from north to south is 520 km, that from east to west is 1,400 km, and it covers an area of more than 400,000 km 2 (Fig. 1). TB has a warm temperate climate with an annual solar radiation budget of 575-627 kJ/cm 2 , with about 3,000 annual sunshine hours in the north, and fewer than 3,000 hours in the south. All regions contain windy and dusty weather. The average annual temperature is 9-11 ℃, and the south is slightly warmer than the north. In winter, the east is colder than the west. In winter, the average temperature in TB is lower than − 20 ℃ for only 1-2 days. The average temperature in July in TB is 25-27 ℃. The accumulated temperature varies greatly from year to year. The frost-free period is 200-210 days in the north and 220 days in the south.
A dust event refers to a weather phenomenon in which sand particles and dust are suspended in the air due to wind, and move with the wind to make the air cloudy and reduce visibility. According to the horizontal visibility when dust weather occurs, dust weather is divided into ve levels, namely oating dust, sand blowing, sand storm, strong sand storm, and extra-strong sand storm (GB/T 20480 − 2017). Floating dust means that there is no wind or the wind force is no more than level 3 (3.4-5.4 m/s), sand particles and dust oating in the air make the air cloudy, and the horizontal visibility is less than 10 km. Sand blowing means that the wind blows ground sand and dust to make the air quite cloudy, and the horizontal visibility is 1-10 km. A sandstorm refers to wind blowing ground sand and dust to make the air very cloudy, with horizontal visibility < 1 km; a strong sandstorm has a horizontal visibility < 500 m; and an especially strong sandstorm has a horizontal visibility < 50 m.
In this work, we de ne the occurrence of oating dust, sand blowing, or a sandstorm during a 24-hr period as 'dust weather'. Additionally, the occurrence of any two or three kinds of oating dust, sand blowing, and/or sandstorm is calculated according to a day of dust weather.
This study investigates trends in annual and seasons sand-dust, oating dust, sand dust and sandstorms days using the Mann-Kendall (M-K) (Kendall, 1934(Kendall, , 1938Mann, 1945). Figure 2 shows the distribution characteristics of the annual average sand dust, oating dust, sand blowing, and sandstorm days in mountain, oasis and desert regions in TB. The dust event distribution in TB features more in the south and hinterlands, and less in the north (Fig. 2). In mountainous regions, the maximum occurrence of annual average sand-dust, oating dust, sand blowing, and sandstorm days are 48.    d/y (TSKEG station), 1.9 d/y (YQ station, oasis, north of TB), and 0.4 d/y (TEGT station), respectively, in spring, summer, autumn and winter, accounting for 47.0 %, 22.2 %, 8.8 %, and 2.4 %, respectively, of the total annual average sand-dust days. Spring and summer are the seasons with the highest rates of sanddust in TB, accounting for 69.5 to 77.0 % of the annual sand-dust days; meanwhile, the sand-dust days in autumn and winter account for 23.0 to 30.5 % of the total amount of annual sand-dust days.  show the spatial distributions of oating-dust days in spring, summer, autumn, and winter. The average maximum occurrences of oatingdust days are 69.6 d/yr, 64.9 d/yr, 35.9 d/yr, and 23.2 d/yr, respectively, all at HT, in spring, summer, autumn, and winter, accounting for 36.0 %, 33.5 %, 18.5 %, and 12.0 %, respectively, of the total annual average number of oating-dust days. The average minimum occurrences of oating-dust days are 2.3 d/yr (TSKEG station), 0.5 d/yr (YQ station), 0.7 d/yr (YQ station), and 0.1 d/yr (TSKEG station), respectively, in spring, summer, autumn, and winter, accounting for 37.8 %, 5.9 %, 8.2 %, and 0.9 %, respectively, of the total annual average number of oating-dust days. Spring and summer are the seasons with highest amounts of oating dust in TB, accounting for 69.5 to 77.8 % of the total annual oating-dust days; meanwhile, the number of oating-dust days in autumn and winter account for 22.2 to 30.5 % of the total annual oating dust days.   Figure 9 shows that 34 stations, 33 stations, 29 stations and 32 stations present a downward trend, and only 3 stations, 2 stations, 1 station and 1 station present an upward trend, of spring, summer, autumn and winter in oating dust days. Among the stations that negative trend have 25, 29, 23 and 22 stations passed 99% con dence level, of spring, summer, autumn and winter in oating dust days. Positive trend have 1, 0, 3 and 0 stations passed 99% con dence level, of spring, summer, autumn and winter in oating dust days.

Seasonal trend variations of oating dust
The largest annual average downward linear trend rates of oating dust are − 0.33 d/yr (QM station), -0.28 d/yr (QM station), -0.19 d/yr (QM station), and − 0.18 d/yr (HT station), respectively, in spring, summer, autumn, and winter in TB. However, the largest annual average upward trends of oating dust are 0.67 d/yr, 0.36 d/yr, 0.33 d/yr, and 0.17 d/yr, respectively, all at TZ station, in spring, summer, autumn, and winter in TB.   (Table 1).

Seasonal trend variations of sandstorms
On average, the maximum values of sand-dust, oating dust, sand blowing and sandstorm days are 62.0 d, 51.3 d, 11.3 d, and 9. 3 d, respectively, in 1978, 1978, 1975, and 1976 in mountains; 147.7 d, 133.6 d, 39.9 d, and 19.9 d, respectively, in 1979, 1979, 1980, and 1966, in oases; and 188 d, 172 d, 95 d, and 36 d, respectively, in 2010, 2010, 2009, and 1998, in deserts. The annual variations of sand dust, oating dust, sand blowing, and sandstorm in mountains are similar to those in oasis; that is, they were higher before 1985, and smaller thereafter. In other words, before the mid-1980s, it was the peak of sand-dust, oating dust, sand blowing, sandstorms, and after that, it was the low period. Table 1 shows that the sand-dust is positively correlated with wind speed, air temperature, and relative humidity over different surfaces. A further analysis found that the annual and spring sand-dust shows a positive correlation between wind speed, air temperature and relative humidity in mountains. However, summer, autumn and winter sand-dust shows a positive correlation between relative humidity. Annual, summer and autumn sand-dust shows a positive correlation between wind speed, air temperature and relative humidity, relatively in desert. However, spring and winter sand-dust shows no correlation with wind speed, air temperature and relative humidity in desert. Interestingly, annual and seasons sand-dust shows a positive correlation between wind speed.
Moreover, sand-dust shows a positive correlation between air temperature in oases in autumn. Table 1 Trend of sand-dust, oating dust, sand blowing, and sandstorms, and the correlation coe cients between the various meteorological elements in mountains, deserts, and oases in terms of their annual values, and those during spring, summer, autumn, and winter.

Annual
Spring  Figure 13 shows that the numbers of sand-dust, oating dust, sand blowing, and sandstorm days above the different underlying surfaces in TB from January to December follow the ascending trend of deserts > oases > mountainous regions. Figure 13 also shows that sand dust, oating dust, sand blowing, and sandstorms mainly occur from March to October in TB, accounting for 90.9 %, 88.

Conclusions
Using observational data obtained at 41 stations in TB, we analyzed the spatial and temporal distribution characteristics of sand-dust, oating dust, sand blowing, and sandstorms in mountains, oases, and deserts in this region. Our main conclusions are summarized as follows.
(1) Dust weather occurs frequently in southern regions of TD. The number of dust days (sand-dust, oating dust, sand blowing, and sandstorms) follows the order of deserts > oases > mountains. In the sand-dust event, the main is oating dust.  1.62 d/yr, respectively, in deserts. March to September is the peak season for dust events, accounting for 87.0 to 90.9 %, 85.2 to 97.8 %, and 84.5 to 94.0 %, respectively, of the whole year in mountains, deserts, and oases. The occurrence of sand-dust events is closely related to wind speed and relative humidity in mountains, and is closely related to wind speed in oases, while, is complex in desert. Figure 1 Spatial distribution of the meteorological stations in TB.

Figure 2
Annual average sand-dust days, oating dust days, sand blowing days, and sandstorm days from 1961 to 2015 in TB.

Figure 3
Seasonal average sand-dust days from 1961 to 2015 in TB.

Figure 4
Page 11/17 Seasonal average oating-dust days from 1961 to 2015 in TB.

Figure 5
Seasonal average sand-blowing days from 1961 to 2015 in TB.

Figure 6
Seasonal average sandstorm days from 1961 to 2015 in TB.

Figure 7
Annual average linear trend rates of sand-dust days, oating dust days, sand blowing days and sandstorm days from 1961 to 2015 in TB.

Figure 8
Seasonal average linear trend rates of sand-dust days from 1961 to 2015 in TB.

Figure 9
Seasonal average linear trend rates of oating-dust days from 1961 to 2015 in TB.

Figure 10
Seasonal average linear trend rates of sand-blowing days from 1961 to 2015 in TB.

Figure 11
Seasonal average linear trend rates of sandstorm days from 1961 to 2015 in TB.

Figure 12
Interannual variations of sand-dust days, oating dust days, sand blowing days, and sandstorm days from 1961 to 2015 in mountainous regions, desert regions, and oasis regions in TB.

Figure 13
Annual variations of sand-dust days, oating dust days, sand blowing days, and sandstorm days from 1961 to 2015 in mountainous regions, desert regions, and oasis regions in TB.