Consistent Transport of Terrestrial Microplastics to the Oceanthrough AtmosphereKai Liu, Tianning Wu, Xiaohui Wang, Zhangyu Song, Changxing Zong, Nian Wei, and Daoji Li*

State Key Laboratory of Estuarine and Coastal Research, East China Normal University, 500 Dongchuan Road, Shanghai 200062,China

*S Supporting Information

ABSTRACT: Although atmospheric transport and deposition couldbe an important pathway of terrestrial pollutants to the ocean, littleinformation concerning the presence and distribution of thesesuspended atmospheric microplastics in marine air is available. Weinvestigated, for the first time, the occurrence and distribution ofsuspended atmospheric microplastics (SAMPs) in the west PacificOcean. In this study, the spatial distribution, morphologicalappearance, and chemical composition of suspended atmosphericmicroplastics were studied through continuous sampling during acruise. SAMPs abundance ranged from 0 to 1.37 n/m3, the median of0.01 n/m3. Fiber, fragment, and granule SAMPs quantitivelyconstituted 60%, 31%, and 8% of all MPs, respectively. Interestingly,plastic microbeads with numerical proportion of 5% were alsoobserved. A high suspended atmospheric microplastics abundance was found in the coastal area (0.13 ± 0.24 n/m3), whilethere was less amount detected in the pelagic area (0.01 ± 0.01 n/m3). The amount of suspended atmospheric microplasticscollected during the daytime (0.45 ± 0.46 n/m3) was twice the amount collected at night (0.22 ± 0.19 n/m3), on average. Ourobservations provide field-based evidence that suspended atmospheric microplastics are an important source of microplasticspollution in the ocean, especially the pollution caused by textile microfibers.

■ INTRODUCTION

It is well-known that microplastic (MPs) pollution isubiquitous in the marine environment,1−3 and its potentialecological risk is a global concern.4 Due to their small size andaffinity for persistent organic pollutants,5 MPs cause physicaland/or physiological damage of the biota exposed to a highdose.6,7

The pathway and quantitive analysis of MPs transport fromthe terrestrial source to ocean has been a major challenge,considering the complex and multiple factors from humanactivities and environment. Traditionally, riverine input8 andcoastal discharge9 are considered major sources of marineplastic pollution. Recently, several studies suggested thatatmospheric MPs could be the potential source for marineMPs pollution.10,11 However, atmospheric transport aspotential pathway for inland MPs to the ocean has been rarelyinvestigated and is poorly understood. A recent study revealedthat atmospheric MPs was detected in both urban andsuburban areas.10 The transportation of SAMPs may beinfluenced by the roles of atmospheric circulation andatmospheric dynamic.12 However, most of the research wasperformed in central urban cities13 or coastal areas,11 whichonly revealed the MPs migration in a terrestrial environment.The presence, distribution, and composition of MPs within themarine atmosphere are still relatively unknown. Therefore, wesampled sea air during a cruise in the west Pacific Ocean. It is

hoped that this research would provide a better understandingof the transport of terrestrial MPs to the ocean.

■ MATERIALS AND METHODS

Study Area and Sampling. The distribution of suspendedatmospheric microplastics (SAMPs) was investigated in thewest Pacific Ocean from November 24, 2018 to January 3,2019, when the winter monsoon prevailed in the studied area.Sampling was continuously conducted throughout the cruise(Figure 1a and Supporting Information (SI) SI-1). In order toavoid the contamination from the ship, it headed eitherforward or sideways against the wind throughout the cruise.As described by Liu et al.,11 we used the same technique for

SAMPs sampling. In brief, the samples of SAMPs wereconducted using the following procedures. Atmosphericsamples were collected using a KB-120F type intelligentmiddle flow total suspended particulate sampler (Jinshida,Qingdao, China) with an intake flow rate of 100 ± 0.10 L/min.Three instruments were set horizontally on the top of the shipat intervals of 1.70 m (horizontally)(Figure 1b). The weatherconditions were continuously recorded by a portable

Received: June 8, 2019Revised: August 9, 2019Accepted: August 13, 2019Published: August 13, 2019

Article

pubs.acs.org/estCite This: Environ. Sci. Technol. 2019, 53, 10612−10619

© 2019 American Chemical Society 10612 DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

Dow

nloa

ded

via

EA

ST C

HIN

A N

OR

MA

L U

NIV

on

Oct

ober

22,

201

9 at

01:

28:5

9 (U

TC

).Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

meteorological station (Kestrel 5500L) every 20 min. What-man GF/A glass microfiber filters (1.60 μm pore size, 90 mmdiameter) were used during the sampling procedure. All of theapparatus was calibrated prior to use. Sampling was conductedin triplicate every 4−24 h, depending on the weatherconditions (i.e., if there was rain). Once the sampling wasfinished, filter in the instrument was carefully removed andtransferred to precleaned sampling cassettes with stainless-steeltweezers. Unfortunately, malfunction of the three samplingdevices occurred on December 1−2 and December 18−20,2018, because of the rain, resulting in missing samples duringthat period. There was also a malfunction of two of thesampling devices, so there were no replicated samples afterDecember 20, 2018. Overall, there were 89 samples (filters)collected during this cruise (ab = 38; bc = 30; cd = 21)SAMPs Sample Identification. In the laboratory, all of

suspected microplastics on the filters were photographed andmarked under a stereomicroscope (Leica M165 FC, Germany)with a Leica DFC 450C camera on the top.A Micro-Fourier Transform Infrared Spectrometer (Thermo

Nicolet iN10) with an internal mercury cadmium telluride(MCT) detector was used to identify the presence of themarked substance.14 In the present work, all of the markedsubstance was identified. The micro-FTIR spectra wereobtained under the transmission mode. Calibration to theambient carbon dioxide and water vapor levels was conductedto correct the background interference before analysis.Background signal was obtained by scanning the other partof the diamond window in the pressure pool and then used toautomatically calibrate the background interference during theSAMPs identification.The spectra were then analyzed using the OMNIC 9

software package and evaluated using the OMNIC spectralibrary. Samples with matching values >60% were consideredplastic materials.

Source and Transport Modeling of the SAMPs. Thehybrid single particles Lagrangian integrated trajectory(HYSPLIT) model in the backward direction has been widelyapplied to identify the origin of air masses.15 It could be usedfor the source identification of fine particulate matters16,17 andsuccessfully adopted for verifying the potential source ofSAMPs in remote areas.18 During the modeling, the number ofdays that was used to calculate the trajectory frequencies wasset to 48 h. The level adopted for the reversion was 10 m dueto the approximate sampling height (9.42 m). Meteorologicaldata from GDAS (Global Data Assimilation System) was usedto simulate the transport of SAMPs.Parameters used in the modeling were based on the res 16

and 17, but a higher duration period in the current study waschosen because of possibly longer transport distances ascribedto the relative lower density (0.01−1.7 g/cm3) of the commonplastic matrix.19 Source points for the backward trajectory were122.89°E, 30.85° N (ab), 131.88°E, 30.37°N (bc); cd:144.80°E, 16.93°N (cd).

Weight Estimation of the SAMPs. For better inter-pretation of the terrestrial input of SAMPs to the ocean, roughestimation of the mass weight of the SAMPs was conducted.The weight of the SAMPs could be roughly estimatedaccording to the methods by Liu et al.11 The total weight ofthe SAMPs was roughly estimated using the following formula:

M A m h Sn

SAMPs0

∑ρ= δ δ δ δ(1)

where MSAMPs, mδ, and Sδ represent the total weight (g) of theSAMPs, mass of a single piece of the SAMPs, and samplingarea (m2), respectively; Aδ and ρδ represent the meanabundance (n/m3) and density (g/cm3) of the SAMPs withspecific polymer compositions, respectively. h was the constantvalue of 9.42 m, the sampling height above sea level.

Figure 1. Geo-location of the sampling track (a) and sampling device used in the study (b) The red dotted line and black dots indicate the cruisetrack and beginning or ending points of the SAMPs sampling, respectively (more information in SI-1). The division of the sampling areas (ab, bc,and cd) was based on the general distance from the coastline, representing the nearshore (ab), pelagic (bc), and remote (cd) region.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

10613

Fibrous and fragmented MPs are most abundant both interrestrial20 and marine environment.21 Therefore, we roughlycalculated the mass of these two kinds of SAMPs. The fibrousSAMPs were generally seen as cylinders, whereas thefragmented SAMPs were deemed cuboid.11 The mass of asingle piece of fibrous SAMP was computed with the followingformula:

m r L2π=δ δ δ (2)

where rδ and Lδ indicate the transect length and size (along itslongest dimension), respectively, of the SAMP. The mass of asingle piece of fragmented SAMP was calculated with thefollowing formula:

m z L2=δ δ δ (3)

where Zδ and Lδ are the thickness and size (along its longestdimension) of the fragmented SAMP.Quality Assurance and Quality Control. All the filters,

stainless tweezers, and glass vessels were wrapped withaluminum foil and heated at 450 °C overnight before usage.The sample cassettes were totally submerged in the acid media(HCl: H2O, volume ratio 1:10) overnight and then washed bydistilled water (resistivity: 9.5 MΩ·cm, 25 °C) until theyreached neutrality. Finally, these cassettes were thoroughlyrinsed with Milli-Q water (resistivity: 18.2 MΩ·cm, 25 °C) 5−7 times and air-dried in an SW-CJ-1FB type ultracleanworktable (Sujing, Shaoxing) with a vertical wind (0.6 m/s).All of the solutions used in this study were filtrated with a

1.6 μm pore size GF/A glass microfiber membrane (N0 1820−047, Whatman, UK) prior to use. The identification process ofthe SAMPs were performed in an ultraclean stainless-steelroom (researchers used an air shower before entering), andcotton and nitrile gloves were worn to prevent externalcontamination during the sampling and analysis procedure.Statistical Analyses. Data analysis was performed using

the SPSS 23.0 software. Normality of data set was tested withShapiro-Wilk’s test. If SAMPs was not normally distributed,nonparametric tests (Kruskal−Wallis test) was conducted to

determine the difference of spatial and temporal distribution ofSAMPs. Statistical significance and extreme difference wererepresented with * = P < 0.05 and ** = P < 0.01, respectively.The sizes (along its longest dimension) of collected MPs

were measured using the ImageJ software (version 1.51j8).Figure 1 was produced with Ocean Data View (Schlitzer,2018)22 and the other graphs were generated by Origin Pro2017.

■ RESULTS AND DISCUSSION

Distribution Pattern. For the first time, we demonstratedthe presence of MPs in sea air and their spatial distributionfrom the coastal area to the open ocean. In this study, 26 filterswere free of SAMPs contamination, and 88% of these sampleswere from the bc and cd areas. Overall, SAMPs abundanceranged from 0 to 1.37 n/m3, with the median of 0.01 n/m3

(Shapiro-Wilk test, P = 0.00 < 0.01). Similar to previousstudy,11 heterogeneous distribution of atmospheric MPs wasalso found within the replicate collected samples (coefficient ofvariation: 0−173%). A high SAMPs abundance was found inthe coastal area (0.13 ± 0.24 n/m3) (ab), whereas theminimum (no particles were identified) was detected in thepelagic area (bc and cd region) (Figure 2a). Generally, theabundance of the SAMPs tended to decrease and then reachthe plateau as the distance away from the continent increased.Based on the occurrence (depositional and suspended) of

SAMPs, detailed comparison of the atmospheric MPs amongstudies was illustrated in Table.1. Overall, a higher SAMPsabundance was observed within urban cities, and lowabundance with apparently lower values was generally foundin the coastal area or the pelagic environment.Significant difference (Kruskal−Wallis test, χ2 = 36.69, df =

2, P = 0.00 < 0.01) was spatially found among the samplingarea (ab, bc, and cd), implying the nonconservative behavior ofSAMPs. Although higher median abundance (0.27 n/m3) wasobserved during the daytime, compared with data collected atnight (0.22 n/m3) (Figure 2b), no apparent variation wasstatistically found between samples collected from daytime and

Figure 2. Spatial (a) and temporal (b) distribution of SAMPs abundance. In Figure 1, day and night SAMPs were collected on November 24 and25, 2018, before the departure of the ship. The pie charts show the corresponding shape composition of every sample area or period. Extremesignificant difference (P = 0.000) was found between the SAMPs abundance from ab and other regions (bc and cd). No apparent difference ofSAMPs abundance was spatially found between bc and cd area.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

10614

Table

1.Physicochem

ical

Characteristics

ofAtm

osph

eric

Microplastics

amon

gStud

ies

physicalandchem

icalcharacterizatio

nof

MPs

studiedarea

samplestate

MPs

abundance

shape

size

(mm)

colora

polymer

typesb

Paris(urban

city)1

3depositio

nal

0.29

×10

2 −2.80

×10

2n/

(m2 ·d

)fiber

0.10−5

N/A

N/A

Paris(urban

andsuburban

area)1

0depositio

nal

0.02

×10

2 −3.55

×10

2n/

(m2 ·d

)fiber

0.05−0.60

N/A

PET,P

A,P

U

Paris(urban)2

3depositio

nal

1.59

×10

3 −1.11

×10

4n/

(m2 ·d

)fiber

0.05−3.25

N/A

PA,P

E,PP

Dongguan(coastalcity)2

4depositio

nal

1.75

×10

2 −3.13

×10

2n/

(m2 ·d

)fiber,foam

,film,

fragment

N/A

black,blue,p

ink,red,

white,yellow

PE,P

P,PS

Yantai(coastalcity)2

5depositio

nal

1.30

×10

3 −1.10

×10

4n/

(m2 ·d

)fiber,foam

,film,

fragment

0.05−1

black,red,

transparent,white

PET,P

E,PV

C,P

S

Pyrenees

mountains

(rem

ote

mountain)

18depositio

nal

365±

69n/(m

2 ·d)

fiber,fragment,film

0−3

N/A

PET,P

E,PP

,PVC,P

S

Asaluyehcounty,Iran(coastal

city)2

6suspended

0.30−1.10

n/m

3 ;1.00

n/m

3(average)

fiber,fragment

0−5

black,blue,green,gray,orange,p

ink,red,

transparent,white,yellow

N/A

Shanghai(coastalcity)1

1suspended

0−4.18

n/m

3 ;1.42

±1.42

n/m

3(average)

fiber,fragment,

granule

0.02−9.55

black,blue,b

rown,

green,

gray,red,transparent,

yellow

PET,P

E,PE

S,PA

N,P

MA,E

VA,E

P,ALK

WestPacificOcean

(open

ocean)

(present

study)

suspended

0−1.37

n/m

3 ;0.06

±0.16

n/m

3(average)

fiber,fragment,

granule,microbead

0.02−2

black,blue,brown,green,gray,orange,pink,purple,

red,

transparent,white,yellow

PET,P

E,PE

−PP

,PES

,ALK

,EP,

PA,P

AN,

Phe,PM

A,P

P,PS

,PVA,P

VC

aN/A

stands

for“not

reported.”bPE

T:polyethylene

terephthalate;PE

:polyethylene;PE

−PP

:polyethylene-polypropylene;PP

:polypropylene;

PES:

polyester;PA

N:polyacrylonitrile;PM

A:poly(N

-methylacrylamide);EP

:epoxyresin;

ALK

:alkydresin;

Phe:

phenoxyresin;

PVA:poly(vinylacetate);PS

:polystyrene;

PVC:poly(vinylchloride);

PA:polyam

ide;

andPh

e:phenoxyresin.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

10615

night period (Kruskal−Wallis test, χ2 = 0.92, df = 1, P = 0.34 >0.05). Relative lower abundance of SAMPs at night couldpossibly have resulted from the SAMPs settlement caused bythe high relative humidity at night, especially on the sea. Liu etal.11 found a negative relationship between the relativehumidity and SAMPs abundance during their investigation.Higher value of relative humidity at night could potentiallycontribute to the settlement of the SAMPs in present study.Therefore, this period could be crucial for the SAMPs enteringthe seawater from the sea−air interface. Based on ourobservation, it was roughly speculated that 52% of SAMPsduring the daytime could possibly deposit and dispersed in themarine aquatic environment.Morphological Features. Fiber, fragment, and granule

SAMPs were observed, which constituted 60%, 31%, and 8% ofall MPs by quantity, respectively. Surprisingly, the rest of theSAMPs were plastic microbeads (N = 10), which presence inthe sea air is first reported. Plastic microfibers and fragmentswere observed during the whole cruise, while no trace ofmicrobeads was found in the samples from the cd (pelagic)area. Generally, the SAMPs shape composition becomes lessdiverse as the distance away from the coastline increases. Interms of the temporal SAMPs shape composition, the totalnumber of SAMPs decreased at night, of which the numericalproportion of fibrous SAMPs decreased the most (12%)(Figure 2b). It was roughly speculated that plastic microfibersare more easily subjected to settlement, which could possiblyindicate a close relationship with the fibers’ density and size.The size distributions of SAMPs of every shape are shown in

SI Figure 1S. The size of the SAMPs was 16.14−2086.69 μm,

with the average of 318.53 μm. Overall, the size of the SAMPsin our research tended to be smaller than those observed inother studies (Table 1). Therefore, it speculated that relativelysmall sized SAMPs could be easily transported to the pelagicenvironment, and their migration to more remote areas is likelydriven by the wind.18

The fibrous SAMPs had the highest mean and variation size,while the microbeads were the smallest in terms of the averagesize (Figure 3a). The overall size order was as follows: fiber(474.81 ± 416.14 μm) > fragment (142.16 ± 98.92 μm) >granule (93.65 ± 32.96 μm) > microbead (39.09 ± 21.70 μm).The temporal variation of the SAMPs size is shown in Figure3b. The size of the SAMPs varied from 16.14 μm to 1,785.31μm during the day and 35.62 μm to 2086.09 μm at night.Relative higher size was found during the daytime (391.79 ±378.44 μm) than at night (307.72 ± 413.22 μm) on average.The absence of large SAMPs could possibly be attributed tothe atmospheric deposition to the ocean surface during thenight. Small sized SAMPs could be easily spread to moreremote areas by the wind.

Polymer Composition. We observed 201 particles ofSAMPs in this study and identified 14 polymer types (PET:polyethylene terephthalate, EP: epoxy resin, PE−PP: poly-ethylene−polypropylene, PS: polystyrene, PE: polyethylene,PVC: polyvinyl chloride, Phe: phenoxy resin, ALK: alkyd resin,RY: rayon, PMA: poly(N-methyl acrylamide), PA: polyamide,PVA: poly(vinyl acetate), PAN: polyacrylonitrile, and PP:polypropylene) through spectral analysis (SI Table 1S). PET,EP, and PE−PP SAMPs comprised 57%, 10%, and 6% of theverified MP particles, respectively. Meanwhile, 190 pieces of

Figure 3. Spatial (a, c) and temporal (b, d) variation of SAMPs morphological appearance (shape and size) and polymer compositions. In panels aand c, the spatial variation of the SAMPs size and polymers was illustrated, respectively; In panels b and d, the temporal distribution of the SAMPssize and chemical composition was demonstrated, respectively.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

10616

cellulose were verified during the spectrum analysis andconstitute 24% of all observed particles by quantity. A detailedcomparison with the results of other studies is illustrated inTable 1.An exceptionally high percentage of PET was observed in

the sea air (54%). The spatial and temporal variation of theSAMPs chemical composition is shown in Figure 3a and b,respectively. PET SAMPs were the most common particlesfound in the studied area and comprised the major part of totalSAMPs in every sampled area. Statistical analysis revealed thatthere was no significant difference of the polymer compositionsamong these three regions (Kruskal−Wallis test, χ2 = 0.89, df =2, P = 0.64 > 0.05) and no apparent temporal variation ofmajor polymer components of the SAMPs was found(Kruskal−Wallis test, χ2 = 0.08, df = 1, P = 0.77 > 0.05).Drastically small amount of PET SAMPs in the bc area could

possibly be due to complex joint action of significant distanceaway from the emission source and environmental factors.SAMPs collected in bc area could be potentially derived fromtextile sources of remote regions, and transport would beindependent of the precipitation.27 Though the total numberof isolated SAMPs decreased by quantity, the numerical ratioof the PET SAMPs was higher in the cd region, where thereare a few islands with inhabitants. A higher relative content ofPE−PP was observed in the bc and cd areas (pelagic area)compared with the ab area (coastal area). The SAMPs with alow density and small size could possibly have beentransported farther out. Another explanation might be thatthe fine SAMPs with a lower density originated from thedeposition of atmospheric MPs at higher altitudes.18

Interesting temporal variations of the polymer compositionof SAMPs were found (Figure 3b). The proportion of PET andPE−PP decreased at night compared with the results from thedaytime, which was speculated to have a close relationship withits physical−chemical property (density and hydroscopicity)and environmental factors (pollution source and humidity).The EP polymer made up about four times the number ofSAMPs collected at night than collected during the day. RYand PVC were not observed in the sample collected during thenight, which was possibly influenced by the density of thematerials. Overall, the spatial and temporal variation of theSAMPs may be mainly affected by the combination of thepolymer density and relative humidity from the surroundings.

Sources and Weight Estimations of SAMPs. Sources ofSAMPs may be complex and hard to trace, but their physicalappearance (Figure 4 and chemical compositions can be usedto tentatively explore their origin. The primary sources ofcolorful fiber-shaped SAMPs probably originated from textilematerials. The highest quantitative proportion of the PETmicrofiber probably originated from the abrasion or break-down of clothes fabricated with synthetic fibers. PE−PPfragments with a black color and irregular shape could haveresulted from the incomplete combustion of plastic debris inthe terrestrial environment. Once it was released into the airthrough a chimney, the fragments may have been transportedto remote areas through the wind, with some of the fragmentsentering the marine ecosystem.11

In the present study, there were 14 pieces of EP SAMPs withgranule and microbead shapes. Based on their sphericalappearance and black color, these fragments may have beenthe product of a thermal reaction. Considering their small size(85.69 ± 55.83 μm, on average) and low density, wespeculated that the SAMPs with the EP component mainlyoriginated from the atmospheric deposition at a higheraltitude.18 This speculation is consistent with the findings byAllen et al.18 During the investigation of Pyrenees mountain-ous catchment (few human activities), higher numericalproportion of smaller sized atmospheric MPs (<125 μm)was found at 1425 m above sea level.The general emission source and transport of SAMPs were

roughly inferred using the monthly wind field (SI Figure 2S).The SAMPs in area ab could have resulted from Korean andnortheastern Chinese emissions. It was speculated that Japancould be a source of the SAMPs found in the bc area. For thesoutheast part of the sampling area (cd), a potential source ofthese synthetic particles could be the adjacent Mariana Islands.This previous conclusion was explicitly confirmed by furtherbackward trajectory analysis of HYSPLIT model. Generalemission and dispersion pattern of the SAMPs pollution foreach area (ab, bc, and cd) could be traced (SI Figure 3S).Due to the prevalence (91% of the total SAMPs) of the

fibrous and fragmented SAMPs, the total weight of the SAMPswas approximated by the sum of the SAMPs in the studiedarea. Based on the previous eq 1, the total weight of theSAMPs input from the adjacent continent was estimated. Therough estimation of the SAMPs weight was found using theabove formulas (SI Table 2S). Based on our preliminary

Figure 4. Photos of typical SAMPs sampled from the atmosphere in the study. a, b: fibers; c, d: fragments; e: granule; and f: microbead

Environmental Science & Technology Article

DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

10617

modeling, there were 101 kg SAMPs within the top 9.42 m ofsea air from November to December 2018. Assuming all ofthese SAMPs were spread from the nearby continent, there areabout 1.21 t of SAMPs entering the marine ecosystem withinthe studied area annually. Based on the temporal (day andnight) variation of the present study, there are 0.63 tons ofatmospheric MPs deposition through the air−sea interface atnight. In the event of rain, more SAMPs will probably enter theseawater.The pathway and quantitive analysis of MPs transport from

the terrestrial region to marine environment has been a majorchallenge. Although considerable amount of these pollutant inthe ocean was ascribed to the riverine and coastal discharge8,9

through modeling, MPs could potentially and consistentlycontribute to the marine MPs pollution through atmosphere.In present study, for the first time, we provide the field-basedevidence testifying the prevalence and distribution of MPs inthe sea air. Invisible but not negligible amount of SAMPs wasobserved during the cruise of the west Pacific Ocean, rangingfrom 0 n/m3 to 1.37 n/m3. Our observation implied SAMPscould be another vital source for the marine MPs pollution,especially for the smaller sized MPs. Through atmosphericcirculation, these atmospheric MPs could be possibly trans-ported to the polar region. Based on our preliminaryestimation, 1.21 t of SAMPs from the terrigenous sourcewould be annually transported to the marine environment,leading to further unexpected ecological consequences. Ourstudy aimed to provide the baseline for a better understandingthe biogeological−chemical cycles of MPs on earth.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.est.9b03427.

Normalized size distribution of fibrous, fragment,granule,and microbead SAMPs observed in the study(Figure 1S). Polymer types of identified SAMPs in thestudy (Table 1S). Monthly wind field at 10 m height (a:November; b: December) during the cruise in the study(Figure 2S). 48 h backward trajectory of SAMPs in thestudy (Figure 3S). Total weight of SAMPs collected inthe present study (Table 2S) (PDF)Detailed information on the atmospheric samples andphysic-chemical property of observed SAMPs (PDF)

■ AUTHOR INFORMATIONCorresponding Author*Phone: +86 (21)-62231085; e-mail: daojili@sklec.ecnu.edu.cn.ORCIDKai Liu: 0000-0003-3555-4567NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis research was financially supported by the National KeyResearch and Development Program (2016YFC1402205), theNational Natural Science Fund of China (41676190), and theECNU Academic Innovation Promotion Program for ExcellentDoctoral Students (YBNLTS2019-007). We thank Zheng Huiand Wu Jianbao from the Rainbow Fish Company for their

sampling assistance. We also offer our sincere gratitude to theNOAA Air Resources Laboratory for the provision of theHYSPLIT transport and dispersion model (http://www.ready.noaa.gov) used in this study.

■ REFERENCES(1) Law, K. L.; More t-Ferguson, S.; Maximenko, N. A.;Proskurowski, G.; Peacock, E. E.; Hafner, J.; Reddy, C. M. Plasticaccumulation in the North Atlantic subtropical gyre. Science 2010, 329(5996), 1185−1188.(2) Law, K. L.; Thompson, R. C. Microplastics in the seas. Science2014, 345 (6193), 144−145.(3) Cozar, A.; Echevarría, F.; Gonzalez-Gordillo, J. I.; Irigoien, X.;Ubeda, B.; Hernandez-Leon, S.; Palma, A. T.; Navarro, S.; García-de-Lomas, J.; Ruiz, A.; Fernandez-de-Puelles, M. L.; Duarte, C. M. Plasticdebris in the open ocean. Proc. Natl. Acad. Sci. U. S. A. 2014, 111(28), 10239−10244.(4) Andrady, A. L. Microplastics in the marine environment. Mar.Pollut. Bull. 2011, 62 (8), 1596−1605.(5) Jang, M.; Shim, W. J.; Han, G. M.; Rani, M.; Song, Y. K.; Hong,S. H. Widespread detection of a brominated flame retardant,hexabromocyclododecane, in expanded polystyrene marine debrisand microplastics from South Korea and the Asia-Pacific coastalregion. Environ. Pollut. 2017, 231, 785−794.(6) Paul-Pont, I.; Lacroix, C.; Fernandez, C. G.; Hegaret, H.;Lambert, C.; Le Goïc, N.; Frere, L.; Cassone, A.; Sussarellu, R.;Fabioux, C.; Guyomarch, J.; Albentosa, M.; Huvet, A.; Soudant, P.Exposure of marine mussels Mytilus spp. to polystyrene microplastics:toxicity and influence on fluoranthene bioaccumulation. Environ.Pollut. 2016, 216, 724−737.(7) Sussarellu, R.; Suquet, M.; Thomas, Y.; Lambert, C.; Fabioux, C.;Pernet, M. E. J.; Goïc, N. L.; Quillien, V.; Mingant, C.; Epelboin, Y.;Corporeau, C.; Guyomarch, J.; Robbens, J.; Paul-Pont, I.; Soudant, P.;Huvet, A. Oyster reproduction is affected by exposure to polystyrenemicroplastics. Proc. Natl. Acad. Sci. U. S. A. 2016, 113 (9), 2430−2435.(8) Lebreton, L. C.; Van der Zwet, J.; Damsteeg, J. W.; Slat, B.;Andrady, A.; Reisser, J. River plastic emissions to the world’s oceans.Nat. Commun. 2017, 8, 15611.(9) Bai, M.; Zhu, L.; An, L.; Peng, G.; Li, D. Estimation andprediction of plastic waste annual input into the sea from China. ActaOceanol. Sin. 2018, 37 (11), 26−39.(10) Dris, R.; Gasperi, J.; Saad, M.; Mirande, C.; Tassin, B. Syntheticfibers in atmospheric fallout: a source of microplastics in theenvironment? Mar. Pollut. Bull. 2016, 104 (1−2), 290−293.(11) Liu, K.; Wang, X.; Fang, T.; Xu, P.; Zhu, L.; Li, D. Source andpotential risk assessment of suspended atmospheric microplastics inShanghai. Sci. Total Environ. 2019, 675, 462−471.(12) Lusher, A. L.; Tirelli, V.; O’Connor, I.; Officer, R. Microplasticsin Arctic polar waters: the first reported values of particles in surfaceand sub-surface samples. Sci. Rep. 2015, 5, 14947.(13) Dris, R.; Gasperi, J.; Rocher, V.; Saad, M.; Renault, N.; Tassin,B. Microplastic contamination in an urban area: a case study inGreater Paris. Environ. Chem. 2015, 12 (5), 592−599.(14) Peng, G.; Zhu, B.; Yang, D.; Su, L.; Shi, H.; Li, D. Microplasticsin sediments of the Changjiang Estuary, China. Environ. Pollut. 2017,225, 283−290.(15) Stein, A. F.; Draxler, R. R.; Rolph, G. D.; Stunder, B. J.; Cohen,M. D.; Ngan, F. NOAA’s HYSPLIT atmospheric transport anddispersion modeling system. Bull. Am. Meteorol. Soc. 2015, 96 (12),2059−2077.(16) Wang, Y.; Stein, A. F.; Draxler, R. R.; Jesus, D.; Zhang, X.Global sand and dust storms in 2008: Observation and HYSPLITmodel verification. Atmos. Environ. 2011, 45 (35), 6368−6381.(17) Saliba, N. A.; Kouyoumdjian, H.; Roumie, M. Effect of localand long-range transport emissions on the elemental composition ofPM10−2.5 and PM2.5 in Beirut. Atmos. Environ. 2007, 41 (31), 6497−6509.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

10618

(18) Allen, S.; Allen, D.; Phoenix, V. R.; Le Roux, G.; Jimenez, P. D.;Simonneau, A.; Binet, S.; Galop, D. Atmospheric transport anddeposition of microplastics in a remote mountain catchment. Nat.Geosci. 2019, 12, 339−344.(19) Crawford, C.; Quinn, B. Microplastic Pollutants; Elsevier Inc.,2016, 118.(20) Zhang, K.; Shi, H.; Peng, J.; Wang, Y.; Xiong, X.; Wu, C.; Lam,P. K. Microplastic pollution in China’s inland water systems: A reviewof findings, methods, characteristics, effects, and management. Sci.Total Environ. 2018, 630, 1641−1653.(21) Cole, M.; Lindeque, P.; Halsband, C.; Galloway, T. S.Microplastics as contaminants in the marine environment: a review.Mar. Pollut. Bull. 2011, 62 (12), 2588−2597.(22) Schlitzer, R. Ocean Data View, 2018. https://odv.awi.de.(23) Dris, R.; Gasperi, J.; Mirande, C.; Mandin, C.; Guerrouache,M.; Langlois, V.; Tassin, B. A first overview of textile fibers, includingmicroplastics, in indoor and outdoor environments. Environ. Pollut.2017, 221, 453−458.(24) Cai, L.; Wang, J.; Peng, J.; Tan, Z.; Zhan, Z.; Tan, X.; Chen, Q.Characteristic of microplastics in the atmospheric fallout fromDongguan city, China: preliminary research and first evidence.Environ. Sci. Pollut. Res. 2017, 24 (32), 24928−24935.(25) Zhou, Q.; Tian, C.; Luo, Y. Various forms and deposition fluxesof microplastics identified in the coastal urban atmosphere. KexueTongbao 2017, 62 (33), 3902−3910.(26) Abbasi, S.; Keshavarzi, B.; Moore, F.; Turner, A.; Kelly, F. J.;Dominguez, A. O.; Jaafarzadeh, N. Distribution and potential healthimpacts of microplastics and microrubbers in air and street dusts fromAsaluyeh County, Iran. Environ. Pollut. 2019, 244, 153−164.(27) Stanton, T.; Johnson, M.; Nathanail, P.; MacNaughtan, W.;Gomes, R. L. Freshwater and airborne textile fibre populations aredominated by ‘natural’, not microplastic, fibres. Sci. Total Environ.2019, 666, 377−389.

Environmental Science & Technology Article

DOI: 10.1021/acs.est.9b03427Environ. Sci. Technol. 2019, 53, 10612−10619

10619