Introduction

Airborne dust is one of the important atmospheric constituents which significantly influences the global energy, carbon cycles and solar radiation1,2,3. These dusts affect the radiative flux directly and indirectly through short-wave and long-wave radiation4,5. The long-range transport of airborne dust and its impacts on biogeochemical cycles have been widely concerned6,7,8,9. However, dust aerosols are still inadequately observed over the global source areas10,11,12,13,14. This greatly limits our deep understanding of dust movement on earth surfaces as well as the global dust cycles. Therefore, it requires more measurements and sufficient observations over the data-sparse arid environment.

The Tarim Basin in western China is encircled by the Tianshan Mountains (average altitude above 4000 m a.s.l), the Pamir Plateau (average altitude above 4500 m a.s.l), the Kunlun Mountains (average altitude above 5500 m a.s.l), the Altun Mountains (average altitude above 4500 m a.s.l), and the Tibetan Plateau (average altitude above 5500 m a.s.l). The basin distributes the world’s second-largest desert of moving sand dunes, the Taklimakan Desert (Fig. 1). Aeolian dust events frequently occur in the dryland area with above 60 d per year15,16. This desert region is consistently recognized as one of the primary sources of spring dust storms in Asia and has received widespread attention14,17,18. It is reported that the soil dust emitted from the Taklimakan Desert can be transported up to 8000 km away to the North Pacific Ocean, Arctic region and the North American continent19,20. Thus, the dynamic processes of windblown dust movement in the Tarim Basin are of interest to the wind erosion, atmospheric physics, and dust research communities. The early modern record on desert landscape and aeolian dust in the Tarim Basin was made by Richthofen21. Scientific quantitative studies of windblown dust over the Tarim Basin began in the late twentieth century. Zhang et al.10 made the first quantitative estimation of dust release in the Tarim Basin based on in-situ observations and numerical calculation. Ginoux et al.11 simulated global dust emissions and suggested uncertainties for the Tarim Basin. Huneeus et al.22 further proposed uncertainties in dust simulation based on outputs from 15 global dust aerosol models. To narrow these uncertainties, physical mechanisms of aeolian dust processes have been studied step by step11,13,23,24,25,26,27,28,29,30,31,32,33,34,35. Accurate estimation of dust fluxes requires sufficient multi-source observations, improved parameterization schemes and adequate model validation14. Laurent et al.36 established a map of the aerodynamic roughness length Z0 in East Asia from the composition of protrusion coefficient derived from the POLDER-1 bidirectional reflectance distribution function (BRDF), and retrieved erosion thresholds of 7 m/s in the sandy deserts of the Taklimakan. Yang et al.15,37,38 determined the wind speed threshold range of 4.9–9.2 m/s based on field experiments in the Tarim Basin, and suggested that soil moisture (relative soil moisture above 0.025) may significantly affect the threshold friction velocity. Li et al.39,40 developed a land-use dataset from the latest Land Cover data to improve the simulation capability of the Weather Research and Forecasting model coupled to Chemistry (WRF-Chem) in the Tarim Basin. Meng et al.41 observed and modeled a severe sand and dust storm event occurring in the Tarim Basin and suggested that more than 51% of the emitted soil dusts is persistently suspended over the atmosphere. The dynamical, thermal, and topographic factors together influence the meridional transport of windblown dust in the Tarim Basin15,41,42. Both meteorological conditions and large-scale circulation affect the aeolian dust activity in the Tarim Basin41. In addition, the up-and-down terrains of sand-dune landscapes may considerably impact regional dust emissions and the micrometeorological environment37,38. The large uncertainty in the current estimates of dust emission, transport, and deposition fluxes over source regions, reflects one aspect that, could be attributed to the shortage of surface observational data11,14. Ground-based observations in the Tarim Basin are sporadic and inadequate. Space-borne satellite measurements make up this situation to some extent and provide a timely way to portray the synoptic process of aeolian dust events. Various satellite instruments, such as the Multi-angle Imaging Spectroradiometer (MISR), the Ozone Monitoring Instrument (OMI), the Total Ozone Mapping Spectrometer (TOMS), the Moderate-resolution Imaging Spectroradiometer (MODIS), and the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO), allow the detection and monitoring of the spatiotemporal variation of atmospheric dust aerosols over a large area and with a high observation frequency. The dust aerosol optical depth (AOD) and the Ångström exponent (AE) measured by the above satellite instruments are helpful to reflect the aeolian dust activities, in particular the horizontal and vertical dust distribution and dust transport1,41,43,44. Current monitoring measurements via space satellites have their limitations in estimating aeolian dust, e.g., the relative coarse resolution of space-grid observations, insufficient detection due to abundant cloud cover, and discontinuous sampling. Therefore, there is an urgent need for quantitative analysis and multi-source observations in dust measurements.

Fig. 1
figure 1

Location of the Tarim Basin in western China (a). Three-dimensional terrain and satellite image of the Tarim Basin (b). Desert environment in Tarim Basin (c). The environmental monitoring stations and meteorological stations are marked with red triangles and green dots, respectively. Land use types are identified across the Tarim Basin according to Wang et al.45. The map was created using ArcGIS 10.6 (https://www.esri.com/en-us/arcgis/geospatial-platform/overview).

Full size image

Ground-based observations are still limited in the desolate desert regions of western China14. Continuous and long-term multi-source observations contribute to our further insight into the characterization of aeolian dust activities over the hyper-dry areas. This study presents an investigation based on consecutive in-situ field observations, meteorological records, environmental monitoring data, and satellite products in the Tarim Basin for a full year in 2015. The spatiotemporal variability of airborne dust and its associated influences (e.g., wind regimes, precipitation, atmospheric circulation, etc.) are analyzed and discussed. It is intended to give a detailed context of the windblown environment in this specific Asian dust source region. Meanwhile, we hope to provide insights into aeolian dust activities over complex basin terrain and improve dust simulation and model validation.

Materials and methods

Study area

The Tarim Basin in western China has an area of more than 4 × 105 km2. Mobile dunes, oases, yardangs, gravel Gobi, and saline and alkaline soil crusts are widely distributed over the basin. It has the world’s second-largest mobile dune desert, the Taklimakan desert (Fig. 1), with area of 3.3 × 105 km[2 15. In the sand desert, there are various types of sand dunes, e.g., barchan dune, longitudinal dune, and pyramid dune, etc. This region suffers from perennial climatic drought, aeolian land desertification, and severe sand and dust storms38. The average annual precipitation is less than 100 mm in the desert areas, while its potential evapotranspiration (PET) can exceed 3000 mm/yr14. It’s a hyper-dry continental arid basin with intense wind erosion, dust emission and aeolian dust storms30,46. Windblown sand and dust storms dominate the Tarim Basin throughout the year15.

Ground-based observation data

Surface observation data in the Tarim Basin were mainly collected from meteorological observations, environmental monitoring records, and in-situ field investigations. Meteorological data (e.g., wind speed, air temperature, precipitation, atmospheric visibility, air humidity, and barometric pressure) are regularly observed by the China Meteorological Administration (CMA). The 13 meteorological stations in the Tarim Basin for this study are illustrated in Fig. 1c and summarized in the Supplementary Material Table S1. The dust weather is primarily classified into four types according to the criteria of atmospheric visibility and surface wind speed as listed in the Supplementary Material Table S2. Those four types of dust weather are dust in suspension (visibility ≤ 10 km), blowing dust (1 km ≤ visibility < 10 km), dust storm (0.2 km ≤ visibility < 1 km), and severe dust storm (visibility < 0.2 km)16,17.

Environmental monitoring data for ambient PM10 and PM2.5 concentrations, and atmospheric dust deposition at four environmental monitoring cities (Korla, Aksu, Kashgar and Hotan) in the Tarim Basin were obtained from the Ministry of Ecology and Environment (MEE) of the People’s Republic of China. Meanwhile, we observed the atmospheric dust deposition in Tazhong, the centre of the Taklimakan (Fig. 1). Ambient PM10 and PM2.5 concentration were measured by β-ray absorption method47. The sampling time for PM concentration measurement is 1 h and the detection limit is 10,000 µg/m3·h. For statistical analysis, the PM10 concentration was considered to be 10,000 µg/m3·h if it exceeded the maximum limit of the instrument. Dust deposition is collected by means of glass cylinders using the gravimetric method48. The cylinder is filled with a solution of ethylene glycol (C2H6O2) and water before the sampling experiment. This allows dust particles to be effectively trapped and evaporation to be reduced. The glass cylinder is installed on a 10 m high tower. The deposited dust particles were gravimetrically weighed following collection from glass cylinders, which involved rinsing their contents and oven-drying at 105℃ to remove moisture prior to measurement48.

Satellite data

We used the dust AOD and AE data from MISR satellite instrument49. The satellite instruments were launched on December 18, 1999 and observed at wavelength of 550 nm. MISR can distinguish different types of atmospheric particles, which contribute to study cloud formation and the various coverings on the Earth’s surface50. Vertical distribution data of dust AOD were derived from the CALIPSO satellite product of Vertical Feature Mask (VFM) level 2 version 4.10 at 532 nm (https://eosweb.larc.nasa.gov/project/calipso/cal_lid_l2_vfm-standard-v4-10). The CALIPSO satellite can detect the vertical structure and properties of global aerosols using active lidar, providing a three-dimensional view of cloud and aerosol formation and evolution51.

Results and discussion

Windblown dust in the Tarim basin

Wind erosion and sand dust weather occur frequently in the Tarim Basin15,52. Figure 2 shows the monthly variation of aeolian dust frequency in the Tarim Basin in 2015, including dust in suspension, blowing dust, dust storm and severe dust storm. Dust in suspension accounted for 77.5% of the total dust days, followed by blowing dust (19.1%), dust storm (2.3%) and severe dust storm (1.1%). Aeolian dust events frequently occurred between March and August as depicted in Fig. 2. Spring (March–April–May) is the windblown dust active season of the year, followed by summer (June–July–August), autumn (September–October–November) and winter (December–January–February). Table 1 lists the main windblown dust events in the Tarim Basin during spring dust active season in 2015. These dust episodes were accompanied by strong winds, low visibility, dense dust clouds, and harmful air pollution53. The maximum observed surface wind speed was 26.2 m/s, while the atmospheric visibility was extremely low (less than 10 m). At the same time, the strong winds blew up large amount of dust particles, which seriously deteriorated the local air quality. Robust dust signals were detected during the windblown episodes, where ambient PM10 concentrations were correspondingly higher than 10,000 µg/m3·h as observed in late April and mid-May, respectively. This high concentration of PM10 in the air greatly exceeds the Air Quality Guideline (AQG, 45 µg/m3 averaged over 24 h) suggested by the World Health Organization (WHO). MODIS Terra satellite images captured clearly several strong sand and dust storms over the Tarim Basin in spring 2015, as shown in Fig. 3. Desert dust swept the entire basin, and were uplifted, suspended, and blown toward to the east. It should be noted that in May 2015, severe dust storm blew soil particles into the atmosphere, producing dense yellow dust clouds that looked like the “eye of the Dust Devil” (Fig. 3c). Those satellite images and dust records reveal the Tarim Basin is seriously affected by frequent aeolian dust events.

Fig. 2
figure 2

Monthly mean of the number of days of windblown dust occurrence at 13 meteorological stations over the Tarim Basin in 2015. The pie chart in the right upside represents the proportion of each type of dust event over the total number of dust events.

Full size image
Fig. 3
figure 3

True-colour images of several strong sand and dust storms over the Tarim Basin in 2015 as observed from the MODIS Terra satellite.

Full size image
Table 1 Major strong windblown dust events observed in the Tarim basin in spring 2015.
Full size table

Surface observation

Fig. 4
figure 4

Spatial distribution of total dust events (a), dust in suspension (b), blowing dust (c), dust storm (d), and severe dust storm (e) over the Tarim Basin in 2015. The map was created using ArcGIS 10.6 (https://www.esri.com/en-us/arcgis/geospatial-platform/overview).

Full size image

Figure 4 presents the distribution of spatial occurrence of windblown dust at 13 meteorological stations in the Tarim Basin in 2015, using the spatial analysis tool of ArcGIS. Dust events primarily occurred in Tazhong with 156 d, followed by Pishan, Ruoqiang, Bachu, and Hotan. The yearly mean of dust incidence is 75 d in Tarim Basin. The outbreak occurrence for dust in suspension, blowing dust, dust storm, and severe dust storm are shown in Fig. 4b,e. Significant high incident of dust in suspension was recorded in Pishan (southern Taklimakan Desert) over 119 d. Blowing dust mainly occurs in Tazhong and Tieganlike, while dust storm appears in Tazhong and Alar, and severe dust storm takes place in Ruoqiang and Tazhong. The distribution of spatial occurrence of windblown dust as displayed in Fig. 4 suggests that there is a high incidence of aeolian dust activities within the basin, which in turn may affect the atmospheric environment in the surrounding areas.

The Tarim Basin is in the northern hemisphere’s mid-latitudes and is strongly influenced by the westerly circulations, the Asian jet streams, and Mongolian cyclones40,51,54. Erosive soil dust is easily blown into the atmosphere by wind turbulence38. Figure 5a showed the spatial pattern of surface wind speed and direction in the Tarim Basin in 2015. The surface wind directions prevail in NNE in the northeastern and the middle of the basin, WSW in the northwestern basin, and WNW in the southwestern basin. The average surface wind speeds in the Tarim Basin range from 1.1 to 2.9 m/s, which is agreement with Yang et al.37 regarding the low wind energy environment in this area. It showed that surface wind speed has a correlation coefficient of 0.15 with the frequency of windblown dust events (Supplementary Material Figure S1). Wind patterns in the Tarim Basin are closely related to its regional landform structures and the surrounding alpine terrain1,55. Basin topography may have a significant impact on regional wind directions and dust emissions14,18,38,41, which results in geostrophic current and local vortex. Wind thresholds and wind directions are, therefore, differed at regional scale in specific sites37. Mountain ranges that are transverse to wind current usually generate wave motion in the atmosphere6. On the edges of the Tarim Basin, Fohn and valley winds could be somewhile responsible for generating dust devil or dust storms over the fluvio-lacustrine sedimentary areas. The environment of the Tarim Basin is dominated by desert climate due to the scarcity of precipitation. In 2015, the annual average precipitation in the arid region of Tarim Basin was 132.5 mm (excluding the alpine stations Baluntai and Aheqi), with the lowest observed records being 54.1 mm in Ruoqiang and 63.4 mm in Tazhong (Fig. 5b). Precipitation peaks in summer, followed by spring, autumn, and winter. The lack of rainfall severely restricts the growth of local plants, fungi, and algae, resulting in sparse vegetation throughout the basin. Biological soil crusts are, therefore, fragile and the surface aeolian sandy soils are easily to emitted into the atmosphere56.

Fig. 5
figure 5

Spatial distribution of average wind speed and wind direction (a) and precipitation (b) over the Tarim Basin in 2015. The amount of total precipitation at each meteorological station has been labelled. The map was created using ArcGIS 10.6 (https://www.esri.com/en-us/arcgis/geospatial-platform/overview).

Full size image

Figure 6 shows the time series changes of hourly ambient particulate matter (PM) concentrations at four major environmental monitoring cities over the Tarim Basin in 2015. PM concentrations were highest in spring and lowest in winter, while they varied from Korla to the other cities in summer and autumn. The maximum PM concentrations were detected in Kashgar in early May, exceeding 10,000 µg/m3. Consistently elevated dust concentrations were monitored from March to August in Hotan at the south margin of the basin. The high PM concentrations in the Tarim Basin are associated with extreme arid environment caused by intense wind erosion and prolonged periods of scarce precipitation38.

Fig. 6
figure 6

Time series variation of hourly ambient particulate matter concentration (unit: µg/m3·h) in four environmental monitoring cities in the Tarim Basin in 2015.

Full size image

High concentrations of PM detected in the Tarim Basin lead to large dust depositions and ecological effects1,16. Figure 7 depicted the intra-annual variability of atmospheric deposition in Tarim Basin in 2015. Monthly dust deposition fluxes in the Tarim Basin, northwest China, vary between three orders of magnitude (1 to 1000 g/m2·mon). Korla and Aksu have the lowest records with average deposition of less than 22 g/m2·mon, while Kashgar, Hotan, and Tazhong have a relatively high values of 91, 147 and 317 g/m2·mon, respectively. Dust deposition was highest in the summer (June-July-August), followed by spring (March-April-May), autumn (September-October-Novermber) and winter (December-January-February). It should be noted that anthropogenic pollution in populated oasis cities around the basin (e.g., Korla and Kashgar) may contribute to atmospheric deposition. The monthly maximum dust deposition was recorded in Tazhong of the central Taklimakan Desert with 976.5 g/m2 and 778.2 g/m2 in July and May, respectively. Among these five environmental monitoring sites, Hotan has the second largest deposition rate in the basin, with 435.5 g/m2 and 358.8 g/m2 in June and July, respectively. This high dust depositions are closely associated with ambient PM concentrations as depicted in Fig. 8. Statistical analysis reveals that the higher the concentration of ambient PM, the greater the atmospheric dust deposition. An equation describing the potential relationship between PM concentration and deposition rate is presented in Eq. (1) according to these observational data. Overall, atmospheric dust deposition rates in the basin present a pattern of high levels during May-June-July. This deposition rate may have close linkage with the dry and hot climate in the basin. Specifically, consecutive dust in suspension can contribute to significant dry deposition rates in the basin15. The maximum ambient PM concentration and dust deposition in the Tarim Basin exceeded 10,000 µg/m3 per hour, and 3800 g/m2 per year, respectively. Compared with global observations of atmospheric deposition for the major arid desert areas46,57,58,59,60,61, the above records suggest relatively high dust fluxes in the Tarim Basin. It is noteworthy that the deposition rates in the Tarim Basin are at a similar order of magnitude level to the largest recorded surface observations in Central Asia as reported by Groll et al.59.

Fig. 7
figure 7

Monthly atmospheric dust deposition over the Tarim Basin in 2015.

Full size image
Fig. 8
figure 8

Relationship between monthly dust deposition and ambient PM concentration in Tarim Basin.

Full size image
$$Q_{d} = \left\{ {\begin{array}{*{20}c} {11.79e^{{000046c}} } & {\left( {50 < c \le 400} \right)} \\ {8.14e^{{.9}} c^{{3.83}} ,} & {\left( {400 < c < 600} \right)} \\ \end{array} } \right.$$
(1)

Where Qd is the monthly dust deposition flux (unit: g/m2), c is the monthly average ambient PM10 concentration (unit: µg/m3).

Satellite measurement

Remote monitoring by satellite instruments is fundamental to understand spatio-temporal change of dust aerosols43,51. Figure 9 illustrates the seasonal AOD spatial pattern obtained from the MISR satellite product over the Tarim Basin, west China in 2015. Dense dust aerosols were evident in Tarim Basin in spring and summer seasons. In autumn and winter, the areas with high AOD values are primarily in the western and eastern periphery of the basin, respectively. This intra-annual AOD variation shows that spatial patterns of windblown dust in the basin, which reflected seasonal dynamic distribution and changes. The Ångström exponent derived from the satellite product is associated with aerosol particle size, and is usually utilized as a qualitative index of atmospheric aerosol particle size62. Figure 10 presents the spatial pattern of seasonal Ångström exponent over the Tarim Basin in western China in 2015. It reflected that large particles were measured over most of the basin in the period of spring, autumn and winter, while relative fine particles were measured during summer. This anomaly may be related to the prolonged suspension of fine dust particles in the summer air of the basin, as well as the high temperatures and low winds season (Fig. 2). During the hot summer month, daytime surface temperature in the arid regions of the Tarim Basin can reach over 45 degrees Celsius37. Convective dust emission may produce the suspended dust aerosols due to the regional thermodynamic environment. Particularly, frequent dust whirls (dust vortices) can lead to increased amounts of entrained and uplifted desert dust as reported by Yang et al.14.

Fig. 9
figure 9

Spatial distribution of AOD derived from the MISR satellite product over the Tarim Basin in west China in 2015.

Full size image
Fig. 10
figure 10

Spatial distribution of Ångström exponent derived from MISR satellite product over the Tarim Basin in 2015.

Full size image

Figure 11 illustrated the vertical profiles of windblown dust in the Tarim Basin as observed by CALIPSO satellite instruments. It revealed that dust particles were blown into the atmosphere and suspended over the basin. Windblown dust particulate matter can even rise as high as an altitude of 12 km. Previous studies reported that variable vertical wind may result to strong atmospheric turbulence, and affect the up-welling and transport of stratospheric aerosols63. The vertical distribution and variation of windblown dust observed in the Tarim Basin indicated that widespread mineral dust dispersion and mixed air pollution (Fig. 11). The satellite measurement provides solid evidence of persistent dust suspension occurring in Tarim Basin, which may result in high dry deposition rates as illustrated in Fig. 7. In addition, windblown dust over the Tarim Basin may be further transported eastwards across the high-rolling mountains of the Qinghai-Tibet Plateau, affecting the atmospheric environment of East China, Korea Peninsular, Japan and far oceans64,65.

Fig. 11
figure 11

Vertical profiles of atmospheric features measured by CALIPSO satellite over the Tarim Basin during spring dust active season in 2015. (ND: not determined, CM: clean marine, D: dust, PC: polluted continental, CC: clean continental, PD: polluted dust, S: smoke, O: other).

Full size image

Conclusions

This study presents a continuous full-year observation in Tarim Basin, which help us to gain insight into the dust background in the source region. Both ground-based observations and remote satellite measurements show high dust emissions, ambient PM concentrations and dust depositions in the Tarim Basin. The atmospheric environment of the Tarim Basin in western China is strongly influenced by windblown dust. The most active times of the year for windblown dust are spring and summer, followed by autumn and winter. It shows that frequency of aeolian dust event is higher in Tazhong (center of the basin) and Pishan (southern edge of the basin). Significant deposition of atmospheric dust occurs mainly during the period from April to August. Vertically, dust aerosols have been detected suspended in the atmosphere at altitudes of 3–12 km. During the study period, we observed high dust frequency (> 100 d per year), ambient PM concentrations (> 10,000 µg/m3 per hour) and dust deposition (> 3000 g/m2 per year) in the Tarim Basin, northwest China. This notable flux level of dust deposition in the Tarim Basin is of the same order of magnitude as the largest known global surface observations in the arid regions (Central Asia), and higher than in the sub-Sahara, the Middle East and Australia.

Currently, in-situ observations are still limited to a few sites in the Tarim Basin. Therefore, additional multi-source observations from the ground to high altitude need to be initiated in this hyper dry inland region. More details of the formation processes and thermodynamic mechanisms of aeolian dust weather as well as its regional and global impact are required to further study. The convective dust emission, persistent dust in suspension, remote transport of fine dust particles, mixed physicochemical pollution, and dry/wet dynamic deposition process in the Tarim Basin need to be deeply investigated. Furthermore, the continuous spatiotemporal observations of the data-sparse desert region provided by this work would contribute to the validation of aeolian dust model and parameter adjustment on numerical scheme in a basin-like terrain.