lable at ScienceDirect

Environmental Pollution 230 (2017) 1030e1039

Contents lists avai

Environmental Pollution

journal homepage: www.elsevier .com/locate/envpol

Trans-Pacific transport of dust aerosols from East Asia: Insights gainedfrom multiple observations and modeling

Jianping Guo a, *, Mengyun Lou a, c, Yucong Miao a, **, Yuan Wang b, Zhaoliang Zeng a,Huan Liu a, Jing He a, Hui Xu a, Fu Wang d, Min Min d, Panmao Zhai a

a State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, Chinab Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USAc College of Earth Sciences, University of Chinese Academy of Sciences, Beijing 100049, Chinad Key Laboratory of Radiometric Calibration and Validation for Environmental Satellites (LRCVES/CMA), National Satellite Meteorological Center, Beijing100081, China

a r t i c l e i n f o

Article history:Received 27 February 2017Received in revised form3 July 2017Accepted 19 July 2017

Keywords:DustCALIPSOPacificTransportWRF-ChemPM10Wet deposition

* Corresponding author.** Corresponding author.

E-mail addresses: jpguocams@gmail.com (J. G(Y. Miao).

http://dx.doi.org/10.1016/j.envpol.2017.07.0620269-7491/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t

East Asia is one of the world's largest sources of dust and anthropogenic pollution. Dust particles orig-inating from East Asia have been recognized to travel across the Pacific to North America and beyond,thereby affecting the radiation incident on the surface as well as clouds aloft in the atmosphere. In thisstudy, integrated analyses are performed focusing on one trans-Pacific dust episode during 12e22 March2015, based on space-borne, ground-based observations, reanalysis data combined with Hybrid SingleParticle Lagrangian Integrated Trajectory Model (HYSPLIT), and the Weather Research and ForecastingModel coupled with Chemistry (WRF-Chem). From the perspective of synoptic patterns, the location andstrength of Aleutian low pressure system largely determined the eastward transport of dust plumestowards western North America. Multi-sensor satellite observations reveal that dust aerosols in thisepisode originated from the Taklimakan and Gobi Deserts. Moreover, the satellite observations suggestthat the dust particles can be transformed to polluted particles over the East Asian regions afterencountering high concentration of anthropogenic pollutants. In terms of the vertical distribution ofpolluted dust particles, at the very beginning, they were mainly located in the altitudes ranging from1 km to 7 km over the source region, then ascended to 2 kme9 km over the Pacific Ocean. The simu-lations confirm that these elevated dust particles in the lower free troposphere were largely transportedalong the prevailing westerly jet stream. Overall, observations and modeling demonstrate how a typicalspringtime dust episode develops and how the dust particles travel over the North Pacific Ocean all theway to North America.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

Dust storms are prevalent in many East Asian regions in spring,including China, South Korea, and Japan (Murayama et al., 2001).Recently, increasing attention has been paid to the trans-Pacifictransport of dust originating from East Asia (Uno et al., 2008;Shang et al., 2017) due to its substantial impacts on humanhealth, environment, ecosystems, weather and climate in thedownwind areas or even the entire Pacific Ocean (Tegen and Lacis,

uo), miaoyucong@yeah.net

1996; Prospero, 1999; Kim and Park, 2001; Creamean et al., 2013;Wang et al., 2014; Miao et al., 2015; Guo et al., 2016a).

It has been shown that dust storms travelling thousands of ki-lometers downwind occur approximately two to three times morefrequently each spring compared with other seasons (VanCurenand Cahill, 2002). Major eastern Asian dust source regions,including the Taklimakan Desert, Gobi Desert, and Loess Plateau ofChina, account for ~25% of global dust emissions (Ginoux et al.,2004). After being emitted into the atmosphere, 26% of Asiandust was found to outflow eastward, and roughly 11.5% can bedetected in the atmosphere of North America due to trans-Pacifictransport (Zhao et al., 2006). Moreover, dust plumes arefrequently contaminated by anthropogenic pollutant and biomasssmoke over land (J. Huang et al., 2015a). Such an aging process in

J. Guo et al. / Environmental Pollution 230 (2017) 1030e1039 1031

the atmosphere induces significant changes in aerosol propertiesand alters cloud formation and even global distribution of precip-itation (Wang et al., 2013). Hence the global radiation budget andhydrological cycle are susceptible to the long-range transport andtransformation of dust aerosols (Feingold et al., 2016).

The studies concerning the long-range transport of East Asiandust emerged in the 1980s, when either ground-based observation(e.g., Iwasaka et al., 1983; Murayama et al., 2001) or model simu-lation (e.g., Nakajima et al., 1989) has been successfully applied.Dust particles are generally ejected to the boundary layer, some-times up to the free troposphere due to the convection caused byunstable atmospheric conditions (Huang et al., 2008). Therefore,the vertical distribution of dust particles plays a key role inimproving our understanding of long-range transport (Guo et al.,2010, 2016b; J. Huang et al., 2015a, Guo et al., 2016b). Addition-ally, an accurate estimation of the dust climate effect strongly de-pends on aerosol vertical profiles (e.g., Liao and Seinfeld, 1998).Given the significance of the vertical structure of dust aerosols,Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) on-board the Cloud-Aerosol Lidar and Infrared Pathfinder SatelliteObservations (CALIPSO) (Winker et al., 2010) has been widelyapplied to investigate the long-range transport of dust (McKendryet al., 2008, Huang et al., 2008; Uno et al., 2008; Guo et al., 2013).Based on the integrated analyses involving satellite and reanalysisdata, mid-latitude westerly wind has been recognized as one of themajor factors carrying the dust from northwestern China to NorthAmerica across the Pacific Ocean (e.g., Wilkening et al., 2000; Yuet al., 2012).

CALIPSO measurements, combined with aerosol transportmodels have been extensively used to analyze the trans-Pacifictransport of dust (Eguchi et al., 2009). However, few studiesregarding the trans-Pacific transport explicitly characterize the dustneighboring source regions over eastern China. Recently, Guo et al.(2016c) suggested that relatively high wind speed and enhancedheight of planetary boundary layer (PBL) tend to occur in spring inthe northern China, which are favorable for the dust particles to belofted into the free troposphere. Interestingly, one intensive sand/dust storm (SDS) episode occurred in northwestern China on 12March 2015, and then spread throughout eastern China, the KoreanPeninsula, Japan, and the western part of North America on 21e22March 2015. It provides us a golden case to systematically investi-gate the dust storm formation and its transport route, as well as theunderlying physical mechanisms. Therefore, the objective of thiswork is to elucidate the SDS episode from the perspective of bothobservations and models and to ultimately shed light on how thisepisode varies over time and space with a focus on the sourceregions.

2. Data and methods

Multiple ground-based comprehensive observations, the space-borne CALIPSO and the Ozone Monitoring Instrument (OMI) mea-surements, in combination with Hybrid Single Particle LagrangianIntegrated Trajectory Model (HYSPLIT) (Draxler and Rolph, 2013)and the Weather Research and Forecasting Model couple withChemistry (WRF-Chem) (Grell et al., 2005) were applied to eluci-date the trans-Pacific transport of one dust episode, including (1)the general evolutions of this SDS episode based on the OMI andCALIPSO data, and HYSPLIT backward trajectories (section 3.1); (2)the detailed transport processes over source region of SDS episodein East Asia using the surface observations, soundings, and CALIPSOmeasurements (section 3.2); and (3) the detailed processes over thePacific Ocean using the WRF-Chem simulations (section 3.3).

2.1. Satellite observations

Because of the lack of surface observations over the PacificOcean, the satellite data, including the CALIPSO data and the OMIdata, were used to examine this tran-Pacific dust transport event.The CALIPSO satellite was launched on April 2006, which passesover the equator at around 1:30 p.m. and 1:30 a.m. local time (LT).The CALIOP is a primary instrument onboard CALIPSO, and is adual-wavelength polarization lidar designed to acquire verticalprofiles of attenuated backscatter from a near nadir-viewing ge-ometry (Winker et al., 2007). Unlike other space-borne passivesatellite sensors, CALIOPSO can detect aerosols both in clear skyconditions and beneath thin cloud layer (Winker et al., 2007; Omaret al., 2009).

To identify the dust aloft in the atmosphere, the CALIPSO Level 2Vertical Feature Mask (VFM) product is used, which has a verticallyvarying resolution: 30 m below 8.2 km in altitude versus 60 m forthe altitudes within 8.2e20.2 km (Winker et al., 2007). The featuresidentified by CALIOP are first classified into aerosol and cloud usinga cloud-aerosol discrimination (CAD) algorithm. The level of con-fidence for the accuracy of aerosol and cloud classification isassessed by a CAD score, which generally ranges from�100 to 0 foraerosol and 0 to 100 for cloud. The larger the CAD value is, thehigher confidence we have. When an aerosol layer is identified, thescene classification algorithm further categorizes the aerosol layerto one of the six aerosol types: smoke, polluted continental,polluted dust, dust, clean continental, and cleanmarine (Z. Liu et al.,2009). Although the accuracy of CALIPSO dust products may beaffected by clouds, previous studies (e.g., P. Liu et al., 2009; Amiridiset al., 2013) have proven that these products were good enough tostudy the large-scale transport of dust. CALIPSO version 3 data usedhere have significant improvements over previous versions, whichhave been demonstrated by extensive validation studies (e.g.,Kacenelenbogen et al., 2011). Note that CALIOP has better signal-to-noise ratios during nighttime than during daytime due to the noisecaused by daytime solar illumination (Z. Liu et al., 2009). Therefore,the entire dust and polluted dust aerosols used in this study refersto those obtained from nighttime CALIOP VFM data unless other-wise noted. In addition, only dust and polluted dust aerosols thatmeet the criteria described in Table S1 were considered to obtainrobust analysis. Based on the CALIOP VFM data, the occurrence/fractional frequency of dust was calculated to better characterizethe vertical structure of dust (Huang et al., 2013). Given most dustparticles located in the lower troposphere, only the aerosol featuresdetected below 8.2 km in altitude considered. Therefore, the frac-tional frequency of dust (or polluted dust) in the ith 30 m verticalbin (fi) can be derived from the following equation:

fi ¼NiPni¼0Ni

(1)

where Ni is the number of times when dust (or polluted dust)detected by CALIPSO fell in the ith 30 m vertical bin, and n is thetotal number of 30 m vertical bins.

As an active satellite sensor, CALIPSO only makes nadir mea-surements, resulting in limited spatial coverage (Chen et al., 2002;Vaughan et al., 2009). To compensate for this limitation, the OMIdata were used in this study, which has a 2600 kmwide swath andprovides daily global coverage at a spatial resolution of 13 � 24 km(Torres et al., 2007). The OMI is a new-generation instrumentaboard Aura, mainly designed for measuring ozone. It can alsoprovide daily absorbing aerosol products such as absorbing aerosolindex (AAI) (Torres et al., 2007). The AAI can serve as a qualitativeindicator of the presence of the absorbing aerosols, such as dust andbiomass burning aerosols, even above the clouds, depending on the

J. Guo et al. / Environmental Pollution 230 (2017) 1030e10391032

absorbing aerosol concentration (aerosol type) and altitude (Alfaro-Contreras et al., 2016). Since the dust episode of our interestoccurred in March, high AAI value can be thought as high dustloading in the atmosphere. It is noteworthy that both CALIPSO andAura fly in a sun-synchronous orbit and have an equatorial crossingtime of approximately 1:30 p.m. and 1:30 a.m. LT, providing aperfect opportunity to simultaneously observe the dust episodes.

2.2. Meteorological measurements and observations of dustparticles

As a complement to satellite observations, comprehensiveground-based measurements such as PM2.5 (particulate matter lessthan 2.5 mm in aerodynamic diameter) and PM10 mass concentra-tion, meteorological records of weather phenomena (e.g., duststorm, floating dust and blowing dust), and soundings were furtherused to characterize the detailed evolutions of SDS over land fromlocal to regional scales. All of these data were obtained from theNational Meteorological Information Center (NMIC), China Meteo-rological Administration (CMA).

The hourly PM10 and PM2.5 mass concentrations at four airquality stations (Table S2) were collected over the potential sourceregions, including the Kuerle (KEL), Zhangye (ZYS), Zhangjiakou(ZJK), and Yanjin (YJS). In addition, the vertical profiles of temper-ature and wind collected at four sounding sites (Table S2, and Fig.1)adjacent to these air quality stations, were used to understand thevertical thermodynamic structures/processes during this SDSepisode. The soundings are launched twice a day at 0800 BeijingTime (BJT ¼ UTC þ 8 h) and 2000 BJT. Combined with the verticaldistribution of dust particles derived from CALIOP products, thedetail processes of this SDS event could be unraveled. The spatialdistribution of CALIPSO ground tracks relative to the sounding siteswas shown in Fig. 1. A 100 km-radius circle around each soundingsite was used to determine the segments of CALIOP measurementsused for further analyses.

Besides, the dust weather phenomena (i.e., dust storm, floatingdust, and blowing dust) recorded at 98 surface meteorologicalstations across China are used to examine the eastward propaga-tion of dust particles (Goudie and Middleton, 2006).

2.3. Model description and configurations

The WRF-Chem model was employed to advance our

Fig. 1. Spatial distribution of ground-based air quality stations (blue dots), sounding sites (re(black slant lines) during the period 12e15 March 2015. The green asterisks represent sourceused to determine the segments of along-track CALIPSO measurements. The red rectanglereferences to colour in this figure legend, the reader is referred to the web version of this

understanding the long-range transport processes of dust particles.With respect to the atmospheric chemistry configurations, theRADM2-MADE/SORGAM chemical mechanism (Stockwell et al.,1990) was used. The anthropogenic emissions were set based onthe Emissions Database for Global Atmospheric Research (EDGAR)(http://edgar.jrc.ec.europa.eu/). And the emission, transports,mixing, and deposition of dust particles were calculated onlinefollowing X Huang et al. (2015b). The simulation was conductedfrom 0000 UTC 10 March (no dust state) to 0000 UTC 23 March2015, and the first two days were referred to as the spin-up period,and excluded from further analyses. The initial and boundaryconditions of meteorology were set by the 6-hourly 1� � 1� NCEP-FNL reanalysis data (https://rda.ucar.edu/datasets/ds083.2/).

The target domain (Figs. 2 and 3a) covered the region boundedby 70 �E� 120 �Wand 18 �N e 70 �Nwith a horizontal grid spacingof 1� � 1�. In the vertical dimension, 48 vertical layers were set fromthe surface to the 100-hPa level, with resolution varying withaltitude. In terms of major physical processes, the parameterizationschemes used included the WRF single-moment 5-class micro-physics scheme (Hong et al., 2004), the Rapid Radiative TransferModel for General circulation models (RRTMG) longwave/short-wave scheme (Iacono et al., 2008) with the aerosol direct effect, theYonsei University (YSU) boundary layer scheme (Hong et al., 2006),the Grell-Freitas cumulus scheme (Grell and Freitas, 2014), and theNoah land surface scheme (Chen and Dudhia, 2001).

The air mass backward trajectories were calculated using theNOAA HYSPLIT model, developed by NOAA's Air Resources Labo-ratory (Draxler and Rolph, 2013). The HYSPLIT is widely used forcomputing air/pollutant parcel trajectories, as well as the transport,dispersion, chemical transformation, and deposition processes. Thetrajectories of air/pollutant are calculated based on a hybridmethod of the Lagrangian and Eulerian approaches. The Lagrangianapproach uses a moving frame of reference for the advection anddiffusion calculations as the trajectories or air parcels move fromtheir initial location, whereas Eulerian approach employs a fixedthree-dimensional grid as a frame of reference to computepollutant air concentrations (Stein et al., 2015). The ending pointsand time of HYSPLT (Table S3) were set according to the region ofinterest where the dust particles were detected from CALIPSOmeasurements. As such, the ending points were fallen in two targetregions (i.e., the “A” and “B” in Fig. 2). Using the reanalysis data ofGlobal Data Assimilation System (GDAS), eight 216-hr backwardtrajectories ending at 1300 UTC 21 March 2015 were calculated for

d dots), in combination of the overpassing ground tracks (descending node) of CALIPSOs of dust (Guo et al., 2013). The 100 km-radius red circles around the sounding sites aremarks the region for cross sections shown in Figs. 9 and 10. (For interpretation of thearticle.)

Fig. 2. Spatial distribution of OMI/Aura derived absorbing aerosol index showing the temporal evolution of dust episode originating from northeastern Asia. Two target domainswere selected for the subsequent backward trajectory and model simulation analyses, one is located at Alaska (66 �N, 160 �W, red cross symbol A), and another is located at westerncoastal regions of Canada (57�N, 128�W, red cross symbol B). To highlight the dust episode, only the spatial grids with AAI value greater than 1 were given here. (For interpretationof the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Spatial distribution of geopotential height (GH) fields and wind vectors at 500-hPa level at 1400 BJT on (a) 12 March, (b) 14 March, (c) 16 March, (d) 18 March, (e) 20 March,and (f) 22 March 2015. The GH fields and wind vectors were derived from NCEP-FNL reanalysis data. The locations of troughs on 12 March and 14 March are marked by the blackdash lines in (a) and (b).

J. Guo et al. / Environmental Pollution 230 (2017) 1030e1039 1033

the target region “A”, and thirteen 240-hr backward trajectoriesending at 1100 UTC 22 March 2015 were made for the target region“B”.

3. Results and discussion

3.1. Multi-sensor satellite observations of trans-Pacific SDStransport and associated synoptic patterns

Fig. 2 illustrates the spatial distribution of OMI/Aura derivedabsorbing aerosol index, which captures the temporal evolution ofSDS episode originating from the northeastern Asian regions. TheSDS episode initially developed in the Taklimakan Desert (~80 �E,35�N) of northwestern China on 12March 2015. As shown in Fig. 3a,there was a deep trough developed over the northwest of China on12 March 2015, which induced strong westerly winds and favoredthe outbreak of SDS (Aoki et al., 2005). Part of the dust particlescould be lifted into the free troposphere, and be transported to the

downwind regions (Fig. 2), forced by the westerly jets between 30�N and 50 �N (Fig. 3). On 14 March, the elevated dust particles weretransported to the North China Plain (~120 �E) (Fig. 2). And duringthe following seven consecutive days, the elevated dust particlescould travel across the northern Pacific Ocean driven by the west-erly jets, and reach the North American continent on 21e22 March2015 (Figs. 2 and 3). During this SDS episode, another synopticsystem that should be noted was the Aleutian low-pressure (ALP),which hovered over the northeastern Pacific Ocean between 150�W and 180 �W from 16 March to 20 March (Fig. 3c-e). When theelevated dust particles reached the ALP-dominant region, thecyclonic circulation could carry the elevated dust particles to thetarget regions “A” and “B”.

To better understand the tran-Pacific transport process, thevertical curtains of the CALIOP VFM and the backward trajectoriesof HYSPLIT were shown in Fig. 4. On 12 March, a mass of elevateddust particles could be found over the Taklimakan Desert (~80 �E).And then the elevated dust particles were eastwardly advected to

Fig. 4. Three-dimensional trans-Pacific transport route of the SDS episode. The aerosol vertical curtains were illustrated using CALIPSO nighttime measurements. The red lines arethe backward trajectories from the target region “A”, and the blue lines are the backward trajectories from the target region “B”. (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.)

Fig. 5. Spatial distribution of OMI/Aura derived AAI (shade), overlaid with the weather phenomena (circles) reported by surface stations in East Asia on (a) 12 March, (b) 13 March,(c) 14 March, and (d) 15 March 2015.

J. Guo et al. / Environmental Pollution 230 (2017) 1030e10391034

the Inner Mongolia (~95 �E) on 13 March, and reached the GobiDesert (~110 �E) on 14 March. On 15 March, the dust particlesproceeded to the Northeast China and the Korean Peninsula, andthen traveled over the Pacific Ocean, and finally reached the NorthAmerican continent a few days later. This long-range transportprocess unraveled by the CALIPSO product and HYSPLIT trajectoriesshowed a good agreement with that identified from the OMI data.

Interestingly, during this SDS episode, part of the dust particleswere marked as polluted dust by the CALIOP VFM products whenthe dust particles traveled over East Asia from 12 March to 15March, which may be relevant to the high emissions of anthropo-genic pollutants and the aging processes of dust particles (Penget al., 2016; Miao et al., 2017). Considering that the aging pro-cesses of dust particles are extremely complex, which involvesheterogeneous reactions of dust particles with gases (Sullivan et al.,2007), condensations of gases (Clarke et al., 2004), coagulationswith other types of particles (Chou et al., 2003), and interactionswith water vapor (Kim and Park, 2012), it is too hard to use thesatellite data alone to investigate the aging processes. Thus, furtherexplicit laboratory studies are warranted to provide better under-standing of the aging processes of dust particles (Kim and Park,2012).

3.2. SDS characterization in East Asia

In this section, the ground-based observations, along with thesoundings and CALIOP data, were used to further investigate thedetailed processes in the potential source regions of East Asia. Asillustrated in Fig. 5, the SDS propagationwas well manifested by theweather phenomena recorded in the surface stations. On 12 March,the dust events were mainly observed in the northwest of China(Fig. 5a). Two days later, the dust events were found in most sta-tions of eastern China (Fig. 5c).

Additionally, the passage of dust plume could cause extremelyhigh aerosol concentration observed at the air quality stations,which was typically characterized by extremely high PM10 con-centration (Husar et al., 2001). As illustrated in Fig. 6, the sharpincrease of PM10 concentration and decrease of PM2.5/PM10 ratiowere first observed at TZS at ~0400 BJT on 13 March, followed bythe YZS (~2000 BJT on 13March), ELH (~1000 BJT on 14 March), andHEB (~1200 BJT on 15 March). Note that both PM2.5 and PM10concentrations increased significantly during the SDS episode andthe increments in PM10 were largely higher than those in PM2.5. On13 March, the maximum PM10 concentration was found at TZS,which could reach as high as ~7000 mg m�3 (Fig. 6a). These sharp

Fig. 6. Time series of PM2.5 (red curves), PM10 concentration (black curves) and their ratios (PM2.5/PM10, blue curves) during the SDS episode at (a) TZS, (b) YZS, (c) ELH and (d) HEB,spreading from east to west in China. The corresponding red and black horizontal dashed lines represent the average PM2.5 and PM10 concentrations in March 2015, respectively,and two vertical green lines mark the passage of SDS. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

J. Guo et al. / Environmental Pollution 230 (2017) 1030e1039 1035

changes of PM10 concentration coincided well with the SDS passagederived from OMI measurements.

To further characterize the transport and vertical distributionof the dust particles, the dust fractional frequencies derived fromthe nighttime CALIOP data were shown in Fig. 7, as well as theobserved wind profiles. It was found that the prevailing westerlywinds drove the dust particles eastward to the Pacific Ocean.Along the propagation route, at ZYS station, the dust layer wereobserved at a higher altitude (3.4e5.0 km), compared with othertwo downwind stations, which may be caused by the localorographic condition.

In addition, the vertical profiles of potential temperature (PT)and wind speed at sounding sites of KEL, ZYS, ZJK, YJS before andduring the SDS were shown in Fig. S1 of the supplementary ma-terials to unravel the associated local thermodynamics features ofPBL during this SDS. The PT profiles varied greatly before andduring the SDS for these four sites. At the KEL site, during the SDSthe PT profile was generally cooler than that of pervious day(Fig. S1a). In contrast, the PT profile of SDS was almost the same tothat of previous day at ZYS (Fig. S1b). Additionally, at ZJK and YJS, nosignificant difference could be found for the PT profiles below 1 kmabove ground level (AGL) between the dust day and the previousday (Figs. S1ced); while for the upper level (1e2 km AGL), the PT atZJK was cooler during the dust episode (Fig. S1c), and that of YJSwas warmer (Fig. S1d).

In comparison with differing changes of PT profiles at differentsites, consistent changes of wind speed profiles could be noted e

the winds below ~1 km AGL were significantly stronger during theSDS than in the previous day (Fig. S1). Such a consistent change ofwind speed within the PBL indicated that the dynamic factorscaused by the large-scale synoptic forcings played an importantrole in the propagation of SDS. And the responses of thermalstructure of PBL were more complex, which may be affected bothby the large-scale synoptic forcings and the various local-scalesurface forcings.

3.3. Simulated trans-Pacific transport of dust particles

Due to the lack of observations over the Pacific Ocean, in thissection we used the WRF-Chem outputs to further investigate thedetail processes of the tran-Pacific transport of dust particles.Model simulation results (Fig. 8) indicate that the spatial distribu-tion of the elevated dust particles (at the 500-hPa level) wasgenerally consistent with the satellite observations (Fig. 2). Thesimulated dust particles were most limited to the East Asia before14 March 2015. With respect to the dust intensity, hotspots wereevident over the Taklimakan Desert (~80 �E) on 12 March and theGobi Desert (~110 �E) on 14 March (Fig. 8a and b).

On 16 March, huge fanlike dust plumes were observed over thePacific Ocean (Fig. 8c). Beginning from 20 March till 22 March, thepersistent pronounced cyclone systems linked to ALP, which wasobserved to persist over the northeastern Pacific Ocean, facilitatedthe transport of dust to the two target regions (Figs. 3 and 8). Thisfavorable large-scale synoptic condition resulted in one leadingedge of dust plumes reaching Alaska on 20March and another westcoast of Canada on 22 March.

Altitude-longitude cross sections (Fig. 9) averaged over the beltof 30�N - 55�N revealed that the dust plume can be elevated to theheight of ~3e5 km above sea level (ASL) over the source regionsforced by orographic conditions. By and large, the simulatedheights of the dust plumes agreed well with those as observed bythe CALIOP (Fig. 4). As illustrated in Fig. 4b, on 14 March 2015 therewas a trough developed over the eastern China at 550-hPa level,which would induce upward motions (Fig. S2b) and carry the dustparticles produced in the boundary layer to the free troposphere(Fig. 9b), and part of dust particles could be lifted to an altitude ashigh as 8 km. Combining Figs. 9 and 10, it can be clearly seen thatduring the tran-Pacific transport process, part of dust parties werewell above the underlying stratocumulus or stratus clouds overnorthern Pacific Ocean. As a result, although most of dust particleswere removed through the dry and wet deposition processes, most

Fig. 7. Vertical profiles of fractional frequency of (a) pure dust and (b) polluted dust derived from CALIOP nighttime data from 12 March to 15 March 2015, along with the observedwind profiles. The horizontal black line represents the terrain height of each site. Top 1% (red dotted line) means the highest height where the fractional frequency is no less than99%. Bottom 1% (green dotted line) means the lowest height where the fractional frequency is no more than 1%. The sampling locations of CALIOP data and sounding sites aremarked in Fig. 1. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 8. Spatial distributions of WRF-Chem simulated dust concentration and wind vector field at 500-hPa level at 1400 BJT on (a) 12 March, (b) 14 March, (c) 16 March, (d) 18 March,(e) 20 March, and (f) 22 March 2015. The red dash lines mark the regions (30�N-55�N) of cross sections shown in Figs. 9 and 10. (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.)

J. Guo et al. / Environmental Pollution 230 (2017) 1030e10391036

of which occurred in the downwind regions nearest to dust sourcesin eastern Asia (Fig. S3), the rest of elevated small dust particles cantravel across the Pacific Ocean under the favorable synoptic con-ditions (Figs. 2 and 8).

4. Conclusions

In this paper, multiple observational data, including ground-based PM2.5 and PM10 concentrations, weather phenomena

Fig. 10. Similar as Fig. 9, but for the vertical cross sections of WRF-Chem simulated cloud water mixing ratio.

Fig. 9. Vertical cross sections of WRF-Chem simulated dust concentration and wind vector field averaged along the belt of 30 �N to 55 �N (denoted by the two horizontal paralleldashed lines in Fig. 8) at 1400 BJT on (a) 12 March, (b) 14 March, (c) 16 March, (d) 18 March, (e) 20 March, and (f) 22 March 2015. Note that the vertical velocity is multiplied by afactor of 30 to enhance the visual interpretation when plotting wind vectors.

J. Guo et al. / Environmental Pollution 230 (2017) 1030e1039 1037

record, atmospheric soundings, space-borne dust measurementsfrom CALIOP/CALIPSO and OMI/Aura, in combination of the HYS-PLIT model, and WRF-Chem model were comprehensively exam-ined to characterize a dust episode in spring of 2015.

Over the source region, pervasive dust plumes were found overthe Taklimakan Desert and Gobi Desert at the initial phase (12e14March 2015) by simultaneous measurements of CALIOP/CALIPSOand OMI/Aura. Further, ground-based weather observations, alongwith PM10 concentration and PM2.5/PM10 ratio, revealed the tem-poral evolution of the SDS episode. Satellite data also identified thequick transition of dust to polluted dust in the East Asia. Beyond thecontinental region, almost no ground-based observations are

available over the Pacific Ocean. To fill in the data gap, the uniquecapability of CALIOP in providing the vertical distribution of dustparticles offered us the unique opportunity to be able to seamlesslymonitor the evolution of dust episode.

Combining multiple observations and explicit modeling simu-lation, the physical mechanism underlying the long-range trans-port of dust episode has been put forward: In the dust sourceregion, upon the dust particles were uplifted into free troposphereas evident in CALIOP measurements and the WRF-Chem simula-tions. Over the Pacific Ocean, the prevailing westerly winds (at highaltitudes) in northern hemisphere, on one hand, tend to keep thedust plume well above the underlying stratocumulus or stratus

J. Guo et al. / Environmental Pollution 230 (2017) 1030e10391038

clouds over Pacific Ocean, making it being devoid of dry and wetdeposition. In particular, the Aleutian cyclonic circulation overnorthern Pacific Ocean further carried dust plumes eastward toAlaska and western coastal region of Canada on 20 March 20, and22 March, respectively. All of the abovementioned processes facil-itate the long-range transport of dust plumes.

As a first step in a series of efforts, integrated analyses based onmulti-source observations and process-level model simulationhave been applied in attempt to shed light on physical details in oneindividual SDS episode over time and space. This work has signif-icant implications for long-range transport studies over other re-gions affected by dust or anthropogenic aerosols, although futureobservational and modeling work are warranted in order to assessthe long-term climatological features regarding trans-Pacifictransport of SDS, let alone the complicated interactions betweendust aerosols and clouds during the transport processes.

Acknowledgements

This work was supported by the National Natural ScienceFoundation of China under Grants 41471301 and 91544217, theMinistry of Science and Technology under Grant 2017YFA0603501,Central Leading Local Development of Science and TechnologyProject in China under Grant HN 2016-149, the Climate ChangeProject of China Meteorological Administration (CMA) under GrantCCSF201732, and Chinese Academy of Meteorological Sciencesunder Grants 2017Z005, 2017Y002 and 2017R001. The PM2.5/PM10

and radiosonde data used in this paper were acquired from ChinaMeteorological Administration, whereas the OMI/Aura and CALIOP/CALIPSO data were obtained from NASA. The authors appreciatevery much the NASA team for providing the reliable data used inthis study. Last but not least, we are grateful to the editor and thefour anonymous reviewers for their constructive comments, whichhelp significantly improve the quality of this manuscript.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.envpol.2017.07.062.

References

Alfaro-Contreras, R., Zhang, J., Campbell, J.R., Reid, J.S., 2016. Investigating the fre-quency and interannual variability in global above-cloud aerosol characteristicswith CALIOP and OMI. Atmos. Chem. Phys. 16, 47e69. http://dx.doi.org/10.5194/acp-16-47-2016.

Amiridis, V., Wandinger, U., Marinou, E., Giannakaki, E., Tsekeri, A., Basart, S.,Kazadzis, S., Gkikas, A., Taylor, M., Baldasano, J.M., Ansmann, A., 2013. Opti-mizing CALIPSO Saharan dust retrievals. Atmos. Chem. Phys. 13, 12089e12106.http://dx.doi.org/10.5194/acp-13-12089-2013, 14749-14795.

Aoki, I., Kurosaki, Y., Osada, R., Sato, T., Kimura, F., 2005. Dust storms generated bymesoscale cold fronts in the Tarim Basin, Northwest China. Geophys. Res. Lett.32 http://dx.doi.org/10.1029/2004GL021776.

Chen, F., Dudhia, J., 2001. Coupling an advanced land surfaceehydrology model withthe penn stateeNCAR MM5 modeling system. Part I: model implementationand sensitivity. Mon. Weather Rev. 129, 569e585. http://dx.doi.org/10.1175/1520-0493(2001)129<0587:CAALSH>2.0.CO;2.

Chen, Z., Torres, O., McCormick, M.P., Smith, W., Ahn, C., 2002. Comparative study ofaerosol and cloud detected by CALIPSO and OMI. Atmos. Environ. 51, 187e195.http://dx.doi.org/10.1016/j.atmosenv.2012.01.024.

Chou, C.C.K., Chen, T.K., Huang, S.H., Liu, S.C., 2003. Radiative absorption capabilityof asian dust with black carbon contamination. Geophys. Res. Lett. 30 (12),18e21.

Clarke, A.D., Shinozuka, Y., Kapustin, V.N., Howell, S., Huebert, B., Doherty, S.,Anderson, T., Covert, D., Anderson, J., Hua, X., Moore II, K.G., McNaughton, C.,Carmichael, G., 2004. Size distributions and mixtures of dust and black carbonaerosol in asian outflow: physiochemistry and optical properties. J. Geophys.Res.- Atmos. 109 (15). D15S09 1e20.

Creamean, J.M., Suski, K.J., Rosenfeld, D., Cazorla, A., DeMott, P.J., Sullivan, R.C.,White, A.B., Ralph, F.M., Minnis, P., Comstock, J.M., Tomlinson, J.M., Prather, K.A.,2013. Dust and biological aerosols from the Sahara and Asia influence precip-itation in the western U.S. Science 339, 1572e1578. http://dx.doi.org/10.1126/

science.1227279.Draxler, R., Rolph, G., 2013. HYSPLIT (Hybrid Single-particle Lagrangian Integrated

Trajectory) Model Access via NOAA ARL READY Website. NOAA Air ResourcesLaboratory, College Park, MD available at: http://ready.arl.noaa.gov/HYSPLIT.php(Last access: November 2015).

Eguchi, K., Uno, I., Yumimoto, K., Takemura, T., Shimizu, A., Sugimoto, N., Liu, Z.,2009. Trans-pacific dust transport: integrated analysis of NASA/CALIPSO and aglobal aerosol transport model. Atmos. Chem. Phys. 9, 3137e3145. http://dx.doi.org/10.5194/acp-9-3137-2009.

Feingold, G., McComiskey, A., Yamaguchi, T., Johnson, J.S., Carslaw, K.S.,Schmidt, K.S., 2016. New approaches to quantifying aerosol influence on thecloud radiative effect. Proc. Nat. Acad. Sci. U. S. A. 113 (21), 5812e5819. http://dx.doi.org/10.1073/pnas.1514035112.

Ginoux, P., Prospero, J.M., Torres, O., Chin, M., 2004. Long-term simulation of globaldust distribution with the GOCART model: correlation with North AtlanticOscillation. Enviro. Modell. Softw. 19 (2004), 113e128. http://dx.doi.org/10.1016/S1364-8152(03)00114-2.

Goudie, A.S., Middleton, N.J., 2006. Desert Dust in the Global System. Springer.Grell, G.A., Freitas, S.R., 2014. A scale and aerosol aware stochastic convective

parameterization for weather and air quality modeling. Atmos. Chem. Phys. 14(10), 5233e5250. http://dx.doi.org/10.5194/acp-14-5233-2014.

Grell, G.A., Peckham, S.E., Schmitz, R., McKeen, S.A., Frost, G., Skamarock, W.C.,Eder, B., 2005. Fully coupled “online” chemistry within the WRF model. Atmos.Environ. 39, 6957e6975. http://dx.doi.org/10.1016/j.atmosenv.2005.04.027.

Guo, J.P., Zhang, X.Y., Cao, C.X., Che, H.Z., Liu, H.L., Gupta, P., Zhang, H., Xu, M.,Li, X.W., 2010. Monitoring haze episodes over Yellow Sea by combining multi-sensor measurements. Int. J. Remote Sens. 31 (17e18), 4743e4755.

Guo, J.P., Niu, T., Wang, F., Deng, M.J., Wang, Y.Q., 2013. Integration of multi-sourcemeasurements to monitor sand-dust storms over North China: a case study.Acta Meterol. Sin. 27 (4), 566e576. http://dx.doi.org/10.1007/s13351-013-0409-z.

Guo, J., Deng, M., Lee, S.S., Wang, F., Li, Z., Zhai, P., Liu, H., Lv, W., Yao, W., Li, X., 2016a.Delaying precipitation and lightning by air pollution over the Pearl River Delta.Part I: observational analyses. J. Geophys. Res. - Atmos. 121, 6472e6488. http://dx.doi.org/10.1002/2015JD023257.

Guo, J., Liu, H., Wang, F., Huang, J., Xia, F., Lou, M., Wu, Y., Jiang, J.H., Xie, T., Zhaxi, Y.,Yung, Y.L., 2016b. Three-dimensional structure of aerosol in China: a perspec-tive from multi-satellite observations. Atmos. Res. 178e179, 580e589. http://dx.doi.org/10.1016/j.atmosres.2016.05.010.

Guo, J., Miao, Y., Zhang, Y., Liu, H., Li, Z., Zhang, W., He, J., Lou, M., Yan, Y., Bian, L.,Zhai, P., 2016c. The climatology of planetary boundary layer height in Chinaderived from radiosonde and reanalysis data. Atmos. Chem. Phys. 16,13309e13319. http://dx.doi.org/10.5194/acp-16-13309-2016.

Hong, S.-Y., Dudhia, J., Chen, S.-H., 2004. A revised approach to ice microphysicalprocesses for the bulk parameterization of clouds and precipitation. Mon.Weather Rev. 132, 103e120. http://dx.doi.org/10.1175/1520-0493(2004)132<0103:ARATIM>2.0.CO;2.

Hong, S.-Y., Noh, Y., Dudhia, J., 2006. A new vertical diffusion package with anexplicit treatment of entrainment processes. Mon. Weather Rev. 134,2318e2341. http://dx.doi.org/10.1175/MWR3199.1.

Huang, L., Jiang, J.H., Tackeet, J.L., Su, H., Fu, R., 2013. Seasonal and diurnal variationsof aerosol extinction profile and type distribution from CALIPSO 5-year obser-vations. J. Geophys. Res. - Atmos. 118, 4572e4596. http://dx.doi.org/10.1002/jgrd.50407.

Huang, J., Minnis, P., Chen, B., Huang, Z., Liu, Z., Zhao, Q., Yi, Y., Ayers, J.K., 2008.Long-range transport and vertical structure of Asian dust from CALIPSO andsurface measurements during PACDEX. J. Geophys. Res. - Atmos. 113 (23) http://dx.doi.org/10.1029/2008JD010620.

Huang, J., Guo, J., Wang, F., Liu, Z., Jeong, M.-J., Yu, H., Zhang, Z., 2015a. CALIPSOinferred most probable heights of global dust and smoke layers. J. Geophys. Res.- Atmos. 120 (10), 5085e5100. http://dx.doi.org/10.1002/2014JD022898.

Huang, X., Song, Y., Zhao, C., Cai, X., Zhang, H., Zhu, T., 2015b. Direct radiative effectby multicomponent aerosol over China. J. Clim. 28, 3472e3495. http://dx.doi.org/10.1175/JCLI-D-14-00365.1.

Husar, R.B., Tratt, D., Schichtel, B.A., Falke, S., Li, F., Jaffe, D., Gasso, S., Gill, T.,Laulainen, N.S., Lu, F., 2001. Asian dust events of April 1998. J. Geophys. Res. -Atmos. 106, 18317e18330. http://dx.doi.org/10.1029/2000JD900788.

Iacono, M.J., Delamere, J.S., Mlawer, E.J., Shephard, M.W., Clough, S.A., Collins, W.D.,2008. Radiative forcing by long-lived greenhouse gases: calculations with theAER radiative transfer models. J. Geophys. Res. - Atmos. 113 http://dx.doi.org/10.1029/2008JD009944.

Iwasaka, Y., Minoura, H., Nagaya, K., 1983. The transport and spatial scale of Asiandust-storm clouds: a case study of the dust-storm event of April 1979. Tellus35B, 189e196.

Kacenelenbogen, M., Vaughan, M.A., Redemann, J., Hoff, R.M., Rogers, R.R.,Ferrare, R.A., Russell, P.B., Hostetler, C.A., Hair, J.W., Holben, B.N., 2011. An ac-curacy assessment of the CALIOP/CALIPSO version 2/version 3 daytime aerosolextinction product based on a detailed multi-sensor, multi-platform case study.Atmos. Chem. Phys. 11, 3981e4000. http://dx.doi.org/10.5194/acp-11-3981-2011.

Kim, B.G., Park, S.U., 2001. Transport and evolution of a winter-time Yellow sandobserved in Korea. Atmos. Environ. 35 (18), 3191e3201.

Kim, J.S., Park, K.H., 2012. Atmospheric aging of asian dust particles during longrange transport. Aerosol Sci. Technol. 46 (8), 913e924.

Liao, H., Seinfeld, J.H., 1998. Radiative forcing by mineral dust aerosols: sensitivity to

J. Guo et al. / Environmental Pollution 230 (2017) 1030e1039 1039

key variables. J. Geophys. Res. - Atmos. 103, 31637e31645.Liu, P., Zhao, C., Zhang, Q., Deng, Z., Huang, M., Ma, X., Tie, X., 2009. Aircraft study of

aerosol vertical distributions over Beijing and their optical properties. Tellus B61 (5), 756e767.

Liu, Z., Vaughan, M., Winker, D., Kittaka, C., Getzewich, B., Kuehn, R., Omar, A.,Powell, K., Trepte, C., Hostetler, C., 2009. The CALIPSO lidar cloud and aerosoldiscrimination: version 2 algorithm and initial assessment of performance.J. Atmos. Ocean. Technol. 26, 1198e1213. http://dx.doi.org/10.1175/2009JTECHA1229.1.

McKendry, I.G., Macdonald, A.M., Leaitch, W.R., van Donkelaar, A., Zhang, Q.,Duck, T., Martin, R.V., 2008. Trans-Pacific dust events observed at whistler,british Columbia during INTEX-B. Atmos. Chem. Phys. 8, 6297e6307. http://dx.doi.org/10.5194/acp-8-6297-2008.

Miao, Y., Hu, X.-M., Liu, S., Qian, T., Xue, M., Zheng, Y., Wang, S., 2015. Seasonalvariation of local atmospheric circulations and boundary layer structure in theBeijing-Tianjin-Hebei region and implications for air quality. J. Adv. Model.Earth Syst. 7 (1), 1e25. http://dx.doi.org/10.1002/2015MS000522.

Miao, Y., Guo, J., Liu, S., Liu, H., Li, Z., Zhang, W., Zhai, P., 2017. Classification ofsummertime synoptic patterns in Beijing and their associations with boundarylayer structure affecting aerosol pollution. Atmos. Chem. Phys. 17, 3097e3110.http://dx.doi.org/10.5194/acp-17-3097-2017.

Murayama, T., et al., 2001. Ground-based network observation of Asian dust eventsof April 1998 in east Asia. J. Geophys. Res. - Atmos. 106 (D16), 18345e18359.http://dx.doi.org/10.1029/2000JD900554.

Nakajima, T., Tanaka, M., Yamano, M., Shiobara, M., Arao, K., Nakanishi, Y., 1989.Aerosol optical characteristics in the yellow sand events observed in May, 1982at Nagasaki, part II, Models. J. Meteorol. Soc. Jpn. 67, 269e291.

Omar, A.H., Winker, D.M., Kittaka, C., Vaughan, M.A., Liu, Z., Hu, Y., Trepte, C.R.,Rogers, R.R., Ferrare, R.A., Lee, K.-P., Kuehn, R.E., Hostetler, C.A., 2009. TheCALIPSO automated aerosol classification and lidar ratio selection algorithm.J. Atmos. Ocean. Technol. 26, 1994e2014. http://dx.doi.org/10.1175/2009JTECHA1231.1.

Peng, J., Hu, M., Guo, S., Du, Z., Zheng, J., Shang, D., Zamora, M., Zeng, L., Shao, M.,Wu, Y., Zheng, J., Wang, Y., Glen, C., Collins, D., Molina, M.J., Zhang, R., 2016.Markedly enhanced direct radiative forcing of black carbon particles underpolluted urban environments. Proc. Natl. Acad. Sci. U. S. A. 113 (16), 4266e4271.

Prospero, J.M., 1999. Long-range transport of mineral dust in the global atmosphere:impact of African dust on the environment of the southeastern United States.Proc. Natl. Acad. Sci. U. S. A. 96 (7), 3396e3403. http://dx.doi.org/10.1073/pnas.96.7.3396.

Shang, H., Chen, L., Letu, H., et al., 2017. Development of a daytime cloud and hazedetection algorithm for Himawari-8 satellite measurements over central andeastern China. J. Geophys. Res.- Atmos. 122 (6), 3528e3543.

Stein, A., Draxler, R., Rolph, G., Stunder, B., Cohen, M., Ngan, F., 2015. NOAA's HYS-PLIT atmospheric transport and dispersion modeling system. Bull. Amer.Meteor. Soc. 96, 2059e2077. http://dx.doi.org/10.1175/BAMS-D-14-00110.1.

Stockwell, W.R., Middleton, P., Chang, J.S., Tang, X., 1990. The second generationregional acid deposition model chemical mechanism for regional air qualitymodeling. J. Geophys. Res.-Atmos. 95, 16343. http://dx.doi.org/10.1029/JD095iD10p16343.

Sullivan, R.C., Guazzotti, S.A., Sodeman, D.A., Prather, K.A., 2007. Direct observationsof the atmospheric processing of asian mineral dust. Atmos. Chem. Phys. 7 (5),1213e1236.

Tegen, I., Lacis, A.A., 1996. Modeling of particle size distribution and its influence onthe radiative properties of mineral dust aerosol. J. Geophys. Res. - Atmos. 101,19237e19244. http://dx.doi.org/10.1029/95JD03610.

Torres, O., Tanskanen, A., Veihelmann, B., Ahn, C., Braak, R., Bhartia, P.K., Veefkind, P.,Levelt, P., 2007. Aerosols and surface UV products from Ozone Monitoring In-strument observations: an overview. J. Geophys. Res. - Atmos. 112 http://dx.doi.org/10.1029/2007JD008809.

Uno, I., Yumimoto, K., Shimizu, A., Hara, Y., Sugimoto, N., Wang, Z., Liu, Z.,Winker, D.M., 2008. 3D structure of Asian dust transport revealed by CALIPSOlidar and a 4DVAR dust model. Geophys. Res. Lett. 35 http://dx.doi.org/10.1029/2007GL032329.

VanCuren, R.A., Cahill, T.A., 2002. Asian aerosols in North America: frequency andconcentration of fine dust. J. Geophys. Res. - Atmos. 107 (24), 4804e4844.http://dx.doi.org/10.1029/2002JD002204.

Vaughan, M.A., Powell, K.A., Kuehn, R.E., Young, S.A., Winker, D.M., Hostetler, C.A.,Hunt, W.H., Liu, Z., McGill, M.J., Getzewich, B.J., 2009. Fully automated detectionof cloud and aerosol layers in the CALIPSO lidar measurements. J. Atmos. Ocean.Technol. 26, 2034e2050. http://dx.doi.org/10.1175/2009JTECHA1228.1.

Wang, Y., Khalizov, A., Levy, M., Zhang, R., 2013. New directions: light absorbingaerosols and their atmospheric impacts. Atmos. Environ. 81, 713e715.

Wang, Y., Wang, M., Zhang, R., Ghan, S.J., Lin, Y., Hu, J., Pan, B., Levy, M., Jiang, J.,Molina, M.J., 2014. Assessing the effects of anthropogenic aerosols on Pacificstorm track using a multi-scale global climate model. Proc. Natl. Acad. Sci. U. S.A. 111 (19), 6894e6899.

Wilkening, K.E., Barrie, L.A., Engle, M., 2000. Trans-Pacific air pollution. Science 290(5489), 65e67. http://dx.doi.org/10.1126/science.290.5489.65.

Winker, D.M., Hunt, W.H., McGill, M.J., 2007. Initial performance assessment ofCALIOP. Geophys. Res. Lett. 34 (19) http://dx.doi.org/10.1029/2007GL030135.

Winker, D., Pelon, J., Coakley Jr., J., Ackerman, S., Charlson, R., Colarco, P., Flamant, P.,Fu, Q., Hoff, R., Kittaka, C., 2010. The CALIPSO mission: a global 3D view ofaerosols and clouds. Bull. Amer. Meteor. Soc. 91, 1211. http://dx.doi.org/10.1175/2010BAMS3009.1.

Yu, H., Remer, L.A., Chin, M., Bian, H., Tan, Q., Yuan, T., Zhang, Y., 2012. Aerosols fromoverseas rival domestic emissions over North America. Science 337 (6094),566e569. http://dx.doi.org/10.1126/science.1217576.

Zhao, T., Gong, S., Zhang, X., Blanchet, J.-P., McKendry, I., Zhou, Z., 2006. A simulatedclimatology of Asian dust aerosol and its trans-Pacific transport. Part I: meanclimate and validation. J. Clim. 19, 88e103. http://dx.doi.org/10.1175/JCLI3605.1.