Simulation and Analyses of the Potential Impacts of Di ﬀ erent Particle-Size Dust Aerosols Caused by the Qinghai-Tibet Plateau Desertiﬁcation on East Asia

: In this paper on the analysis of the vertical distribution of di ﬀ erent-diameter dust aerosols and the potential impacts on East Asia, the sensitivity simulation tests of dust aerosols during 2002–03 were conducted by changing the underlying surface on the Qinghai-Tibet Plateau in the global atmospheric circulation model Community Atmosphere Model (CAM) 3.1. The results show that dust aerosol particles in East Asia are mainly distributed in the diameters of 0.64–5.12 µ m. The high concentrations of dust aerosols are centered on the surface in the source areas and gradually raised during the eastward transport across East Asia, reaching a height of 4 km at 120 ◦ E. The small dust particles with diameters less than 1.28 µ m are transported higher and farther driven by the midlatitude westerlies. The Qinghai-Tibet Plateau desertiﬁcation leads to increasing concentrations of dust aerosols in all size bins and raisesthe transport height of dust aerosols in East Asia. The long-range transport in the East Asian troposphere is dominated by dust aerosols particles of diameters 0.64–2.56 µ m, as well as a large contribution of dust aerosols with diameters larger than 1.28 µ m.


Introduction
Dust aerosol is a major type of natural atmospheric aerosols. The amount of dust aerosol emitted into the atmosphere is estimated up to 1000-3000 Tg every year, accounting for about half of the total tropospheric aerosol amount [1][2][3]. Dust aerosols significantly influence local, even global, ecological environment, as well as influence climate change [4][5][6]. Dust aerosols can directly change the global radiation balance through reflecting, absorbing and scattering the solar short-wave radiation and the ground long-wave radiation. The direct effect of aerosol depends on the scattering and absorption capacities of particles, which are mainly decided by the composition, size and shape of particles. As cloud-condensation nuclei or ice nuclei, dust aerosols can also indirectly change the cloud microphysical structure and the rainfall development. The indirect effect depends on the ability of particles to act as cloud-condensation nuclei or ice nuclei [7], which is mainly decided by the activation height of cloud droplets, the hygroscopicity and the size distribution of particles. Hence, constraining the size distribution of dust aerosols is necessary (albeit insufficient) to limit both the direct and indirect aerosol forcing. In particular, the vertical distribution of dust aerosols can affect the magnitude of both effects [8,9].

Model Description and Simulation Settings
CAM 3.1 is an advanced atmospheric circulation model jointly developed by the National Center for Atmospheric Research/University Corporation for Atmospheric Research (NCAR/UCAR) as the atmosphere component of the community climate system model 3.0 (CCSM3.0). There are three optional dynamic cores in CAM 3.1: the Eulerian spectral transform method (EUL); the semi-Lagrangian scheme (SLD); and the finite volume approach (FV). The spectrum model CAM3.1 can run simulations in the horizontal spectrum resolutions of triangle truncated wave numbers 85, 63, 42 and 31, naming respectively T85, T63, T42and T31 resolutions. T42 was selected in this paper. In the model, the hybrid σ-P (sigmapressure) vertical coordinate was used. The hybrid σ-P vertical coordinate focused on properties of the vertical derivative of the terrain-following coefficient, which affected the smoothness and shape of layer-thickness profiles and determined the coordinate's monotonicity over variable terrain [18]. This coordinate can take the functional form η = h p, p s , where p s is the surface pressure and h p, p s is a monotonic function of p, such that there is a one-to-one mapping between p and η for a given p s . The functional dependence on p s permitsterrain-following coordinates. The standard convention is to define η to be normalized and dimensionless, satisfying the lower-and upper-boundary conditions: h p s , p s = 1, h p top , p s = 0, respectively, where p top is the pressure at the top model half-layer. The σ coordinate was used at the surface layer; the σ-P transition coordinate was used in the middle layer; and the pure P coordinate was used in the upper layer. The atmosphere was divided into 26 layers vertically, with the top of about 3.5 hPa.
Five main atmospheric aerosols, including black carbon, organic carbon, sulfate, sea salt and sand dust, were taken into account in the CAM [19][20][21].According to the physical setting of CAM3.1, 12 diameter classes from 0.01 to 40.96mm can be only given to represent the size distribution of soil dust (Table 1).All atmospheric aerosol processes, including dust emission fluxes, concentrations and deposition, were calculated in 12 size bins. A size-distributed soil-dust scheme [22][23][24] in the model was modified and driven with a Chinese soil-texture scheme that inferred the size distribution with 12 categories and up-to-date desert distribution in China [19].Combined data sets for the desert distribution/texture, satellite land-use/roughness length and observed soil moisture provided a coherent input parameter set for the soil-dust emission scheme for deserts in East Asia. The transport model used the physics package from the Canadian global climate model (GCM) and a semi-Lagrangian and semi-implicit transport scheme for dynamics and passive tracers [25].The CAM3.1involved the main aerosol processes, such as aerosol emission, transport, chemical transformation, dry deposition and wet deposition, and dealt with the interaction between aerosol and cloud precipitation through working with the cloud module of microphysics. The mass conservation equation of aerosol is expressed as follows [26]: In the equation, x ip is the aerosol mass, in which p is the variation rate of five different dry-particle mass mixing ratios and i represents the 12 bins of aerosol particle size. Local changes of aerosol mass were determined by the aerosol emission, transport, chemical transformation, wet growth, collision growth, dry deposition, incloud scavenging and wash-out. The advective and convective transport and subgrid turbulent diffusion were considered in the aerosol transport. The emission sources contain both natural and artificial aerosol emissions from the surface to the atmosphere. The chemical transformation produces secondary aerosol particles, and the wet growth and collision growth affect the internal and external mixing of aerosols. During the dry removal, aerosol particles are absorbed by vegetation or subsided to the ground by gravity. The aerosols are also removed by incloud scavenging and wash-out under the cloud caused by cloud precipitation, named as wet removal for aerosols, both resulting in the wet deposition of aerosols.
The Asian monsoon is considered as one of the major components of the atmospheric circulation over the northern hemisphere and the globe [27]. The plateau is important for the Asian monsoon, and the variation of the Asian monsoon can also affect the weather and climate of the plateau [28,29]. Moreover, the Asian monsoon is closely related to dust storms and dust aerosols [30,31]. Therefore, to comprehensively illustrate the general situation, the weak (2002) and strong (2003) South-Asian-monsoon years were chosen to carry out the simulation and analyses in this study. A series of contrast sensitivity experiments were carried out by using the model. Experiment 1 set the global desert distribution as the underlying surface of the model (Figure 1), and the simulation experiment of dust aerosols in 2002-03 was carried out as A1. As there is no available data reflecting the desert distribution of the plateau's underlying surface, the sensitivity simulation, which changed the land type of the plateau (polygonal frame area in Figure 1) into deserts, was conducted for two years of 2002-03 in the experiment A2, with the properties of the ground soil modified correspondingly (that is, the desert proportion of the region was changed to 1.0). After expanding the plateau desert area, the potential impact in East Asia of dust aerosols with different particle sizes could be studied.

Dust Aerosol Emission Verification
The annual total emissions simulated by CAM3.1 in the global main region of dust aerosol emission, as well as its proportion to global emissions and its comparison with other simulation results, are shown in Table 2. The results of global and regional dust aerosols during 2002-03 simulated by CAM3.1 are reasonable and can be used for further analyses. However, it should be noted that the simulation results of dust-emission amount from the desert region in North Africa (71%) was slightly more than that in other studies, while that from the desert areas in Central Asia was slightly less (4.5%), which maybe attributed to the two-year simulation with the different resolution of desert areas in this study. Note: the global proportion is in brackets.

Concentration Distribution of Dust Aerosols with Different Particle Sizes
There were no dust aerosol particles with the diameter of 0.01-0.02 µm or 0.02-0.04 µm. Therefore, the concentration characteristics of the dust aerosols with the diameter of 0.04-40.96 µm are discussed in this paper after being divided into 10 bins. Figure 2 shows the concentration of dust aerosols in each size bin and the distribution of their proportions to the total dust concentration in East Asia. The size bins of 0.04-0.08 µm (Figure 2a Figure 3. The concentrations of dust aerosols at sites A, B, C, D and E decreased in turn. Meanwhile, the dust aerosol high-concentration center in the source area was on the surface and was gradually raised when being transported eastward, maintaining a height of 4 km when it arrived at 120 • E. The dust aerosols were transported across the Eurasian continent. First, the dust was lifted into the air from the surface over the dust-source region and then transported in the troposphere, before finally sinking in the surface of the sink region. As a result, the concentrations of dust aerosols and air were changed [37,38]. The concentration of dust aerosol particles with the diameter of 1.28-2.56 µm was higher than that of 0.64-1.28 µm under the height of 1 km at site A. However, the concentrations of dust aerosols with the diameter of 0.64-1.28 µm at other sites and site A above 1 km were the largest. At sites A and B, the dust-aerosol concentration of 1.28-2.56 µm was higher than that of 0.32-0.64 µm under 9 km, while it is opposite above 9 km. The dust-aerosol concentration difference between the two size bins of 0.32-0.64 µm and 1.28-2.56 µm gradually decreased at sites C and D, and the dust-aerosol concentration of 0.32-0.64 µm was higher than that of 1.28-2.56 µm at site E. The second highest concentration level corresponded to the size bins of 2.56-5.12 µm and 0.16-0.32 µm, and the concentrations corresponding to other bins were lower. The dust aerosol particles in the atmosphere were mainly distributed in the size bins 6, 7 and 8 with the diameters ranging from 0.32 to 2.56 µm ( Table 1). The dust particles with the diameter less than 1.28 µm could be transported higher and farther by the atmospheric movement, while those with the diameter greater than 1.28 µm were more likely to be raised and settled. The concentration was higher in the source area and near the surface, and decreased faster in the transport.

Variation of Dust Aerosol Concentration in East Asia
The annual average difference of dust aerosol concentration between experiments A1 and A2 in East Asia is shown in Figure 4 for the heights of 0, 1500 and 6000 m, indicating the impacts of the plateau desertification on the regional dust aerosols. The concentration difference between A1 and A2waspositive in East Asia. The high-value area of ground concentration difference was located in the dust-source areas of the western plateau, the Qaidam Basin, the southern Qinghai Plateau and southern Tibet. The high-value center of concentration difference in the lower troposphere (1500m) was located in the plateau and the surrounding areas, and the concentration difference gradually decreased eastward and southward. The high value of concentration difference in the middle troposphere (6000m) is shaped like a band, one extending from Xinjiang to central and western Inner Mongolia, to Qinghai, to the Sichuan Basin, to the north of the middle and lower reaches of the Yangtze River, and then directly to the Yellow Sea, South Korea and the southern Japan peninsula. The mean altitude of the Tibetan Plateau is above 4000 m. The emitted dust aerosols from the plateau are easily transported into the high levels of the middle troposphere over the downstream areas. The dust source on the plateau mainly affects the low level and the surface in the near-source area, i.e., the plateau and its surrounding areas, while it mainly affects the upper level in areas far away from the source, such as the sea area to the south of Japan Island and the central Pacific.  Figure 5 shows the concentration difference of dust aerosols in each size bin at sites A, B, C, D and E in East Asia simulated in experiments A1 and A2. In general, the difference of dust aerosol concentrations between A1 and A2in each size bin at each site was larger than 0, indicating that the plateau desertification increased the concentration of dust aerosols in each size bin. At site A, the larger difference of dust aerosol concentrations corresponded to the size bins of 0.16-0.32 µm, 0.32-0.64 µm, 0.64-1.28 µm, 1.28-2.56 µm and 2.56-5.12 µm from the surface to the height of 10 km, and the high-value difference concentrated in the size bins of 0.64-1.28 µm and 1.28-2.56 µm from the surface to the height of 3 km. At site B, the larger difference corresponded to the size bins of 0.32-0.64 µm, 0.64-1.28 µm, 1.28-2.56 µm and 2.56-5.12 µm from the surface to the height of 11 km. There were high-value centers: the size bin of 1.28-2.56 µm at the height of 0.4-2 km and the size bins of 0.64-1.28 µm and 1.28-2.56 µm at the height of 4-8 km. At site C, the larger difference corresponded to the size bins of 0.32-0.64 µm, 0.64-1.28 µm, 1.28-2.56 µm and 2.56-5.12 µm at the height of 3-11 km, while the high values concentrated in the size bins of 0.64-1.28 µm and 1.28-2.56 µm at the height of 5-8 km. The difference of dust concentrations at site D and E was similar to that at site C, and the difference became more and more concentrated. The plateau desertification raised the transport height of dust aerosols in East Asia. The main contributing size bins in the source area of plateau desertification were 0.64-1.28 µm, 1.28-2.56µm, 0.32-0.64 µm, 2.56-5.12 µm and 0.16-0.32 µm in that order, and the long-distance transport was mainly contributed by the dust aerosols in size bins of 0.64-1.28 µm and 1.28-2.56 µm. At site A in the source area, the high-value center of the concentration difference between dust concentrations of 0.64-1.28 µm and 1.28-2.56 µm was near the surface, and it was lifted to 5-8 km while being transported eastward. It was further confirmed that the particles in different size bins mainly affect the low level and surface in the near-source area, i.e. The plateau and its surrounding areas, while they mainly affect the high level in the areas away from the source. The large difference concentrated in the surface layer and at the low level and the height of 9-11 km. At sites D and E, the large difference concentrated in the height of 5-10 km. The proportion difference between dust particles with a diameter smaller and greater than 1.28 µm was negative and positive, respectively. It indicates that the plateau desertification contributes more to dust aerosols with a diameter more than1.28 µm than those with a diameter smaller than1.28 µm in East Asia. The size distribution of dust particles along transport may result from a combination of the gravitational removal of larger particles and the re-entrainment of smaller particles, because the removal rate of dust aerosols during transport is size dependent [39].The main contributing size bins of dust aerosols in the source area of plateau desertification were0.64-1.28 µm, 1.28-2.56 µm, 0.32-0.64 µm, 2.56-5.12 µm and 0.16-0.32 µm in that order, and the long-distance transport was mainly contributed by the dust aerosols in size bins of 0.64-1.28 µm and 1.28-2.56 µm. The high-value center of the concentration difference between dust concentrations of 0.64-1.28 µm and 1.28-2.56 µm was near the surface in source region, and it was lifted to 5-8 km while being transported eastward. It was further confirmed that the particles in different size bins mainly affected the low level and surface in the near-source area, i.e. The plateau and its surrounding areas, while they mainly affected the high level in the areas away from the source. The dust aerosol on the plateau can be easily ascended to westerlies, which could become the world's highest efficiency dust-distance transport source. The plateau desertification not only increases the concentration of dust aerosols in each size bin, but also raises the transport height of dust aerosols in East Asia. The regional transport of dust aerosols from the Qinghai-Tibet plateau modifies the size distribution of aerosols in East Asia, which could affect the scattering and absorption capacities of aerosols, for the radiation balance during regional transport. Moreover, dust aerosols as cloud-condensation nuclei and ice nuclei alter cloud microphysical structure and the rainfall, depending on the size distribution of dust particles.

Conclusions
Based on the simulation of dust aerosols in East Asia during 2002-03 by using the global atmospheric circulation model CAM3.1, the vertical distribution characteristics of dust aerosols in different size bins in East Asia were analyzed. Moreover, the potential impact of dust aerosols with different particle sizes on East Asia was analyzed by the sensitivity simulation test of changing the underlying surface on the plateau. The main conclusions are as follows.
First, the concentrations of surface dust aerosol particles in the size bins of 0.04-0.08 µm, 0.08-0.16 µm, 10.24-20.48 µm and 20.48-40.96 µm were the lowest (less than 0.5 µg/kg) in East Asia, and the corresponding proportions were below 0.01. They were followed by the size bins of 0.16-0.32 µm and 5.12-10.24 µm. The concentrations of surface dust aerosol particles with a diameter of 0.32-0.64 µm and 2.56-5.12 µm were the second highest, while those of 0.64-1.28 µm and 1.28-2.56 µm were the highest.
Second, in East Asia, the high-concentration center of the dust aerosols in the source area was on the surface, and the center was gradually uplifted while dust aerosols were transported eastward, maintaining at the height of 4 km when reaching 120 • E. The dust aerosol particles in the atmosphere were mainly distributed in diameters ranging from 0.32 to 2.56 µm. The dust particles with diameters smaller than 1.28 µm could be transported higher and farther by the atmospheric movement, while the dust particles with diameters greater than 1.28 µm were more likely to be raised and settle. Meanwhile, the concentration was higher in the source area and near the surface, and decreased faster in the transport.
Last, the plateau desertification not only increased the concentration of dust aerosols in each size bin, but also raised the transport height of dust aerosols in East Asia. Dust particles in different size bins mainly affected the low level and the surface in the near-source area, i.e., the plateau and its surrounding areas, while they mainly affected the high level in areas far away from the source. In the source area, the main contributing size bins were 0.64-1.28 µm, 1.28-2.56 µm, 0.32-0.64 µm, 2.56-5.12µm and 0.16-0.32 µm in that order. The long-distance transport was mainly contributed by the size bins of 0.64-1.28 µm and 1.28-2.56 µm. The plateau desertification contributed more to dust aerosols with diameters greater than1.28 µm than those smaller than1.28 µm in East Asia.