lable at ScienceDirect

Atmospheric Environment 44 (2010) 4638e4644

Contents lists avai

Atmospheric Environment

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

Aerosol concentrations observed at Mt. Haruna, Japan, in relationto long-range transport of Asian mineral dust aerosols

Hiroshi Takahashi a,*, Hiroaki Naoe a, Yasuhito Igarashi a, Yayoi Inomata a,1, Nobuo Sugimoto b

aMeteorological Research Institute, 1-1 Nagamine, Tsukuba 305-0052, JapanbNational Institute for Environmental Studies, Tsukuba 305-8506 Japan

a r t i c l e i n f o

Article history:Received 12 January 2010Received in revised form30 July 2010Accepted 2 August 2010

Keywords:AerosolsAsian dustLong-range transportMountain observationFree troposphereBoundary layer

* Corresponding author. Fax: þ81 29 855 7240.E-mail address: htakahas@mri-jma.go.jp (H. Takah

1 Present Affiliation: Asia Center for Air PollutionJapan.

1352-2310/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.atmosenv.2010.08.007

a b s t r a c t

As a part of the effort to understand the structure of long-range transported aerosol plumes and localpollution, aerosol observations monitored the mass concentrations and number-size distributions duringthe period August 2006 to July 2009 near the top of Mt. Haruna (1365 m), an isolated mountain in theKanto Plain in Japan. The mass concentrations observed at Mt. Haruna and plain sites showed a seasonalvariation with a maximum in spring and summer, respectively. The spring peaks in aerosols atMt. Haruna were probably caused by long-range transport of mineral dust and anthropogenic particlesfrom the Asian continent. The summer peaks at the plain sites was attributed to local pollution from theTokyo metropolitan area. Three examples of 2007 Asian dust events were investigated to show thataerosols may be dispersed in a complicated three-dimensional structure and that delayed arrivals of thedust plumes at plain sites compared to Mt. Haruna were not a rare case. Because of the boundary layerbeing stable at night, the dust layer was advected eastward without the vertical mixing before sunrise.This study suggests that after thermal convection activated by sunlight during daytime Asian dusttransported in the free troposphere may be brought down into the atmospheric boundary layer,increasing the dust concentration there.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Many studies have reported long-range transport of natural andanthropogenic aerosols originating from the Asian continent (e.g.,Osada et al., 2009). For instance, Asian “yellow dust” or Kosaparticles are transported to Japan and even to North America (e.g.,Husar et al., 2001). Asian dust, which is composed of mineralparticles uplifted by frontal activities, has drawn attention not onlyas a cause of low visibility (Okada et al., 1987; Tanaka et al., 1989)but also as a possible source of climate change through its radiativeforcing effect (Aoki et al., 2005) and as a major carrier of anthro-pogenic pollutants such as SO2, NO2 and PAHs (e.g., Tamamuraet al., 2007).

Atmospheric aerosols of anthropogenic origin have a greaterimpact on human health than previously suspected (Schwartz,1994). Japanese economic development during the 1960s and1970s led to widespread air pollution, and suspended particulate

ashi).Research, Niigata 950-2144,

All rights reserved.

matter (SPM), which is roughly equivalent to PM10, has beenregulated by the Ministry of Environment since 1972. Althoughtheir definitions differ, SPM (100% cutoff at 10 mm particle diam-eter) and PM10 (50% cutoff at 10 mm) havemeasured concentrationsthat are generally in good agreement. A new regulation for PM2.5,whose 50% cutoff diameter is 2.5 mm, has been under considerationsince September 2009, because PM2.5 also has a serious impact onhuman health (e.g., Samet et al., 2000).

Although there are many monitoring stations in Japan for PM10or SPM concentrations, most of these are located in urban areas oralong roadsides in areas of heavy traffic to measure local pollution.However, trans-boundary pollution can greatly affect atmosphericquality in Japan. A large fraction of aerosols observed in Japanconsists of material transported from the Asian continent, wherethere are huge emission sources such as mineral particles from aridor semiarid regions (Xuan et al., 2000) and where anthropogenicaerosols and precursors arise from the consumption of fossil fuels(Streets et al., 2003). Hayasaki et al. (2008) found that highconcentrations of photochemical ozone over Japan in May 2007were caused by air pollutants from both the Asian continent andlocal urban areas. For assessing the effects of emission reduction indomestic pollutants on air quality in Japan, it is essential to allowfor aerosols from continental Asia. Many monitoring studies have

H. Takahashi et al. / Atmospheric Environment 44 (2010) 4638e4644 4639

been investigated Asian dust, anthropogenic aerosols, and theirprecursor gases from the Asian continent. The monitoring is per-formed mostly in the East China Sea or isolated islands to avoidlocal emission sources. Aerosols observed by multi-axis differentialoptical absorption spectroscopy (MAX-DOAS) at 0e1 km altitudeover Okinawa Island in Japan showed an annual minimum inAugust and September and a maximum from November to May,when air pollution from the Asian continent was dominant(Takashima et al., 2009).

Aerosols may be dispersed in a complicated three-dimensionalstructure. Iwasaka et al. (1983) detected dust layers in themiddle tolower troposphere using a lidar system. Chen et al. (2009) reporteda two-layer structure of pollutants over the Beijing region ata height of 2500e3500 m in the lower free troposphere and in theboundary layer. Tsunematsu et al. (2006) found a dust layerdistributed at 1e5 km altitude above the boundary layer in theKanto area in Japan. Because ground observations are usuallylimited to near sea level, it is difficult to map the three-dimensionalstructure and mechanism of aerosol transport. To better under-stand these multilayered structures of atmospheric aerosols origi-nating frommany sources, aerosol observations at several differentaltitudes would be useful.

Mountain observations are well suited for year-round in situobservations. However, orographic effects commonly cause mois-ture condensation so that orographic clouds and precipitation canscavenge aerosols. Diurnal wind circulation patterns also uplift airmasses from lower altitudes during the day and mix themwith airof the free troposphere. Some studies of mountain aerosols (e.g.,Bigg, 1977; Kido et al., 2001; Naoe et al., 2003) have used onlynocturnal data to avoid interference with the boundary layeraerosols.

Fig. 1 shows the locations of observation sites at Mt. Haruna, Mt.Happo, and Tsukuba. We chose Mt. Haruna for the followingreasons. Because theMt. Haruna observation site is in the transitionregion between the boundary layer and the free troposphere, it iswell situated to detect aerosols from both of these layers. Also, itcan observe both local and Asian aerosols, helping in evaluating therelative contribution of these two aerosol types. Therefore, the aimof this study was to understand the three-dimensional structure oflong-range transported aerosol plumes and local pollution basedon mass concentrations and number-size distributions observed atMt Haruna. In addition, we used PM10 data measured at Mt. Happo

Fig. 1. Map of central Japan, showing the locations of observation sites: Mt. Haruna(36.48�N, 138.88�E, 1365 m altitude), Mt. Happo (36.70�N, 137.80�E, 1850 m), andTsukuba (36.03�N, 140.08�E, 31 m). Topography is shaded for >1000 and >2000 m;borders of prefectures (some named) are indicated by dashed lines.

and SPM data fromKanto Plain sites and outside Kanto Plain sites inthis study.

2. Methods

Our aerosol monitoring started in August 2006 at Haruna-Fujimountain (36.48�N, 138.88�E, 1390 m), a quiescent volcanodormant for more than 1300 years. We used data from August 2006to July 2009 in this study. The mountain is located on the north-western edge of the Kanto Plain, about 100 km northwest of theTokyometropolitan area. The observation site is near the summit atan altitude of 1365 m. Forests cover the steep, conicalmountain, andthe highest roadwayon themountain is about 200 mbelow the site.There are no large urban or industrial aerosol or precursor sourcesnearby. Thus, we assume that the influence of local pollution or dustis small as compared to that in the Kanto Plain. In addition, we usedPM10 mass concentrations measured at Mt. Happo and SPM massconcentrations from the Kanto plain area (Fig. 1).

A tapered element oscillating microbalance (TEOM) (TEOM-1400a, Thermo Scientific, formally R&P Co. Inc.) was used tomeasure PM10 mass concentration. This mass monitor hasa vibrating hollow tube, called the tapered element, which issecured at its base but free to oscillate at the opposite end. Thetapered element is set into oscillation at its resonant frequency. Asmall filter mounted on its free end collects particulate matter fromthe air passing into the chamber. As particulate matter collects onthe filter, the oscillation frequency decreases and the collectedmass is determined (Page et al., 2007).

An optical particle counter (OPC), which is widely used formonitoring aerosol number concentrations and size distributions,has a multichannel pulse height discriminator that receives laserlight scattered by aerosol particles. For this study, we used an OPC(RION Co. Inc., Model KC01-E) to measure the number concentra-tions of aerosol particles with five diameter ranges: >0.3 mm,>0.5 mm, >1 mm, >2 mm, and >5 mm. PM10 mass concentrationswere obtained at a temperature of 293 K and a pressure of 1 atm.The TEOM and OPC were mounted in a small housing withcontrolled temperature (25 �C). The air inlets were positioned overthe roof (5 m above ground) for the TEOM and beside the wall(2.5 m above ground) for the OPC. To avoid the hygroscopic growthof aerosol particles, aerosols were measured by the TEOM at rela-tive humidities of less than 30% by warming the TEOM sensor to30 �C and equipping the sample equilibrium system with Nafionfilm. Although the OPC has no dehumidifying mechanism inside,the controlled temperature in the housing was higher than that ofoutside most of the observing period. That is, the humidity in theOPC would be lower than that of outside. TEOM and OPC data wereacquired every 30 and 20 min, respectively.

Meteorological data such as temperature, relative humidity,pressure, wind direction, wind speed, and precipitation were alsomonitored with an automatic weather station (WXT510, Vaisala Co.Inc.). Despite some data loss due to instrumental problems, ourdata (TEOM: 18095 h, OPC: 9852 h) covered roughly all seasonsfrom August 2006 to July 2009.

Year-round PM10 observation were carried out at Mt. Happo(36.70�N,137.80�E,1850 m), on the eastern slope of the North JapanAlps approximately 100 km WNW of Mt. Haruna, by the AcidDeposition and Oxidant Research Center (ADORC) as a part of AcidRain Monitoring Survey conducted by the Ministry of Environment.The mass concentrations at Mt. Happo were standardized in thesame manner as those at Mt. Haruna (see below). The higher alti-tude location of the Happo site is favorable for measuring aerosolsin the free troposphere.

Number-size distributions of aerosols weremonitored inTsukuba(110 km ESE of Mt. Haruna, 36.03�N, 140.08�E, 31 m altitude) by an

0.0

10

20

30

40

0

0.5

1

1.5

2

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

HarunaHappoKantoKanto_out

Happo SO2

Mas

s co

ncen

tratio

n (µ

g m

-3)

SO2 (p

pb) a

t Hap

po

month

Fig. 2. Monthly mean PM10 mass concentrations (mgm�3) at Mt. Haruna andMt. Happo sites and monthly mean SPM mass concentrations of aerosols over groupsof the observation sites in the Kanto Plain and outside it. Monthly mean SO2 mixingratios (ppb) at Happo are also shown. All data are for August 2006 to July 2009.

H. Takahashi et al. / Atmospheric Environment 44 (2010) 4638e46444640

OPC of the same model as that used at Mt. Haruna. The instrumentwas installed in a room with controlled temperature on the top(sixth) floor of the Meteorological Research Institute (MRI).

As part of local pollution monitoring efforts in Japan, SPM massconcentrations (mostly measured by using ß-ray absorption) weremonitored at more than 1000 observation sites by local govern-ments andby theMinistry of the Environment,most of them locatedin areas of heavy human activities such as major roads. Hourly SPMdata are presented on the “Soramame-kun” website (http://soramame.taiki.go.jp/), where stations are classified as “roadsideair pollution monitoring stations” and “ambient air monitoringstations”. This study used SPM data for the second category. Forcomparison with PM10 data at Mt. Haruna, we classified the SPMdata into two groups: sites in the Kanto Plain (105 stations inTochigi, Gunma, and Saitama Prefectures), and sites outside theKanto Plain (44 stations in Nagano and Niigata Prefectures). TheSPM monitoring stations used in this study ranged from 1 m to900 maltitude,most of them lower than 300 m. Thus, as all the SPMstationswere lower than theMt. Haruna site, they are considered asthe “plain sites” within the atmospheric boundary layer.

The definitions of PM10 and SPM differ in that the particle sizerange of PM10 has a 50% cutoff at 10 mm diameter and that of SPMhas a 100% cutoff at 10 mm diameter. However, because PM10 andSPM are generally in good agreement and the systematic differencebetween PM10 and SPM is small, we compared them without anycorrection in this study.

In order to compare aerosol mass concentrations betweengroups of sites at different altitudes, the volume of the air inlet wascorrected for standard temperature and pressure. Because pressureand temperature are not observed at most of the SPM monitoringsites, we applied an altitude-dependant factor as a parameter

Cssl ¼ Cobs � FðhÞ ¼ Cobs � fT0=ðT0 � GhÞgg=RG;where Cssl and Cobs are the mass concentration of aerosols (mgm�3)corrected to sea level and the observation point altitude, respec-tively, T0 is the temperature of 293 K, G is the constant temperaturelapse rate of 6.5 K km�1, g is the gravitational acceleration of9.8 m s�2, R is the gas constant as 287 J K�1 kg�1, and h is the alti-tude of the site. In this study, Cssl was averaged over each group ofsites in the Kanto Plain and outside it. The differences in the factor F(h) varied with time and place. For example, the difference in F(h)ranged between �0.06 and þ0.03 at Karuizawa (999 m) in NaganoPrefecture and between �0.05 and þ0.03 in Tokyo (6 m). Thus, thecorrected SPM mass concentrations varied within only a fewpercent of the measured concentrations.

A Mie scattering lidar installed by the National Institute forEnvironmental Studies (NIES) in Tsukuba was used to measure theaerosol backscatter coefficient and depolarization ratio, whichwere then converted into vertical distributions of spherical andnonspherical aerosols by the retrieval processes (Shimizu et al.,2004). The time-height distributions of the backscattering coeffi-cient and depolarization ratio were used for distinguishing the dustplume from other aerosols.

3. Results and discussion

3.1. Features of seasonal variations of mass concentrations

Fig. 2 depicts themonthly averages of PM10 concentrations at theMt. Haruna and Mt. Happo sites, along with monthly SPM concen-trations averaged over the groups of sites in Kanto Plain (Kanto) andoutside it (Kanto_out). The annual averaged PM10 mass concentra-tions at Haruna andHappowere 21 and 12 mgm�3, respectively, andthe averaged SPM mass concentrations in the Kanto Plain and

outside it were 27 and 20 mgm�3, respectively. Although the aver-aged mass concentrations at the mountain sites were lower thanthose within the Kanto Plain because of local pollution from theTokyo metropolitan area and because of the different observationaltitudes, the annual averaged mass concentration at Mt. Harunawas slightly higher than that outside the plain.

The PM10 concentration at Mt. Haruna exhibited a maximum of33 mgm�3 in May and a minimum of 11 mgm�3 in January. At theMt. Happo site, similar seasonal variation was found, but themaximum in average PM10 was in April. If both sites experiencedthe same Asian air masses during spring, the monthly patterns ofPM10 should be similar. In general, Asian dust outbreaks often occurfrom late winter to spring. We investigated PM10 mass concentra-tions at Mt. Haruna fromhourly data and found that the one-monthlag of the PM10 maximum at Haruna was due to Asian dustoutbreaks rather than local effects. Monthly averaged PM10 massconcentrations in May of 2008 and 2009 were 30 and 28 mgm�3,respectively, and they were similar to the averaged value for April(26 mgm�3). However, the PM10 mass concentration in May 2007was 41 mgm�3. Two major dust events were recognized in May2007, on 8e10 and on 26e27. In both events, daily averaged massconcentrations exceeded 100 mgm�3.

In contrast to the mountain sites, the maximum of aerosol massconcentration in the Kanto Plain was recorded in summer. Thissummer maximum has been attributed to weak advection of localprecursor pollution such as SO2 and oxidants within the boundarylayer, which leads to efficient gas-to-particle conversion associatedwith active photochemical reactions in summer (Kumagai et al.,2009). The SPM mass concentrations outside the Kanto Plain alsohas a summer peak, but were lower than those in the Kanto Plain atall times of the year.

In winter, the mass concentrations were reduced at all the sites.This is explained by the fact that the oxidation of SO2,which is one ofthe dominant precursors of anthropogenic aerosols (e.g., Chin et al.,1996), is limited in winter because of inactive photochemical reac-tions. The aerosols produced by diesel automobiles (so-called DEP)and combustion facilities, which were dominant contributors toserious pollution in the wintertime urban atmosphere in Japan inpast decades, have decreased due to restrictions on automobileemissions and dioxins (Mizuno, 2006) during the late 1990s andearly 2000s.

It should be noted that PM10 concentrations at Mt. Haruna inMay were higher than those in the Kanto Plain. Although aerosol

0.0

10

20

30

40

50

0-10 10-20 20-30 30-40 40-50 >50

HarunaHappoKantoKanto_out

Freq

uenc

y (%

)

PM10

/SPM (µg m-3)

0.0

10

20

30

40

50

0-10 10-20 20-30 30-40 40-50 >50

HarunaHappoKantoKanto_out

Freq

uenc

y (%

)PM

10/SPM (µg m-3)

Spring

Summer

a

b

Fig. 3. Frequency distributions of average aerosol mass concentrations from August2006 to July 2009 at Mt. Haruna, Mt. Happo, and groups of observation sites in theKanto Plain and outside it. (a) Spring months March, April, and May. (b) Summermonths June, July, and August.

H. Takahashi et al. / Atmospheric Environment 44 (2010) 4638e4644 4641

mass concentrations at the mountain sites tend to increase duringdaytime due to advection of polluted boundary-layer air by valleywinds as well as thermal convection, the concentration atMt. Haruna is unlikely to exceed those at plain sites solely from themixing of air masses. In hourly data, both the Haruna and Happosites showed diurnal variations with a maximum at 1000e1800 JST(Japanese Standard Time) and a minimum at 2100e0600 JST. Themaximum concentrations at Happo occurred only in spring,whereas those at Haruna occurred in both spring and summer. Thisresult suggests that the springtime high concentrations at themountain sites were caused by long-range transportation of Asiancontinental dust. On the other hand, the summertime highconcentrations at Haruna were more likely caused by local trans-port of polluted air by vertical mixing.

Fig. 2 also presents themonthlymean SO2 mixing ratios (ppb) atHappo. There was clear seasonal variation, with maxima in winterand spring and a minimum in summer. SO2 and other precursors tosecondary aerosol formation can be transported from the Asiancontinent during winter and spring (e.g., Igarashi et al., 2004).Therefore, aerosols from the Asian continent can contribute to thehigh SO2 concentrations in spring at the mountain site.

As primary and secondary aerosol sources in Japan are distrib-uted mainly in urban areas, the concentrations of atmosphericaerosols generally decrease with increasing altitude. However, theaerosol mass concentrations at Mt. Haruna, except in winter, werecomparable to those outside the Kanto Plain. For determining thereason, the correlation coefficients of mass concentrations betweenSPM at the surrounding sites and PM10 at Mt. Haruna were evalu-ated. Slightly higher coefficients (r¼ 0.3w0.5) with Mt. Harunawere found at sites in Nagano and Niigata Prefectures as well assites at higher altitudes within the Kanto Plain. Thus, one possiblereason is the laminar structure of aerosol distribution which wouldmakes synchronized variation in aerosol concentration amongMt. Haruna and other sites located in relatively high altitude. Otherreason is that the air quality at Mt. Haruna may also be affected bypollution advected from the Tokyometropolitan area. Many studiesof land-sea breeze circulation in the Kanto Plain (e.g., Fujibe andAsai, 1979) and strong thermal low in the inland mountain regionduring daytime (Kurita et al., 1985) have suggested that theadvection of air masses across the Kanto Plain during daytime onfine days can reach the mountains. Takeuchi et al. (2004) observedsoluble components of atmospheric aerosols in Yokohama and Mt.Oyama in Kanagawa Prefecture and found that ammonium sulfatewas dominant in aerosol components in summer, as it was inpolluted air in the Kanto Plain. These circumstances account for thesmall concentration differences between Mt. Haruna and the areasoutside the Kanto Plain.

3.2. Frequency distributions of aerosol mass concentration

Fig. 3 shows the frequency distributions of PM10 concentrationsat Mt. Haruna and Mt. Happo and those of SPM in the Kanto Plainand outside it. The mass concentration bins are at intervals of10 mgm�3. In spring, the distributions at Mt. Haruna exhibited highfrequencies in both the lower concentration bins and the highestconcentration bin (Fig. 3a). Mt. Happo showed the same pattern.The high frequencies in the lowest concentration bin at themountain sites are probably a reflection of their typical pristineenvironment. The high frequencies in the highest concentration binare probably caused by long-range transport of aerosols. However,the maximum frequencies in the Kanto Plain and outside it (26%and 28% respectively) were in the 10e20 mgm�3 bin. The relativelyhigh frequencies in the intermediate and high concentration rangesin the plain sites were probably due to domestic air pollution in theboundary layer.

Although the frequency distribution of SPM concentrationsoutside the Kanto Plainwas similar to that in the Kanto Plain, it wasshifted slightly toward the lower concentration bins. In comparingthe hourly data from Mt. Haruna with weather charts, we foundthat in spring at Mt. Haruna there were about 7 advection eventsfrom the Asian continent during 2007e2009. Most of them hadmaxima of mass concentrations greater than 100 mgm�3. Thisresult is consistent with the high frequencies in the highest bin(Fig. 3a). The high concentrations (dense plumes) in the freetroposphere due to long-range transport events were diluted bymixing with air of lower concentrations in the boundary layer,which led to the relatively low frequency (9%) in the highestconcentration bin in the Kanto Plain compared to mountain sites.

In the study of the variation in aerosol properties with height, itis important to estimate atmospheric instability which is closelyrelated to vertical mixing of the atmosphere. We calculated it asa Showalter stability index (SSI), which is defined by

SSI ¼ Tat 500 hPa � Tparcel

where Tat 500 hPa is the temperature at 500 hPa and Tparcel is thetemperature of a parcel lifted from 850 hPa to 500 hPa. A low indexindicates an unstable atmosphere.We found that in summer the SSIwas low (1e2 K) whereas inwinter the SSI was high (10e15 K). Thismeans that the atmospheric conditions in summer tended to beunstable and developed a mixed boundary layer, favoring verticaltransport of domestic pollution aerosols to the top of Mt. Haruna,unlike the wintertime conditions.

The most notable feature in summer (Fig. 3b) is a significantdiscrepancy in the highest concentration bin between Mt. Haruna(13%) and Mt. Happo (1%), suggesting that domestic pollution from

H. Takahashi et al. / Atmospheric Environment 44 (2010) 4638e46444642

the Tokyometropolitan areawas transported and then lifted duringthe development of the mixed boundary layer during daytime insummer. The low frequency in the highest bin at Mt. Happo indi-cates that domestic aerosols did not reach the observation altitudeof Mt. Happo even during the day in summer. This was probablydue to its higher altitude and the absence of populated areasaround the mountain. However, it may be that some events insummer instead were advection events from the Asian continent.Although further analyses are needed, the issue is beyond the scopeof this study.

3.3. Arrival of Asian dust plume to the mountain and plain sites

Katsuno (2006) observed soluble ions in aerosols at Mt. Happoduring the period 1993e2000 found that calcium ion was the mostabundant ion in April, which means that the aerosols obtained atMt. Happo in that month were dominated by Asian mineral dust.The aerosol size distributions of Asian dust events have beenobserved to have a coarse mode (e.g., Arao et al., 2006). Fig. 4 showsthe relationship between the PM10 concentration and volume ratiomeasured at Mt. Haruna in spring and summer. Here, we introducea volume ratio Vol2e5/Vol0.5e1.0 as an aerosol size index, whereVol2e5 and Vol0.5e1.0 are the hourly volume averaged concentra-tions in the diameter ranges of 2e5 mm and 0.5e1.0 mm, respec-tively. The aerosol volume concentrations were estimated byassuming that all the particles measured with the OPC are spher-ical. A higher index should signify a greater proportion of coarseaerosols and a lower index should signify fine aerosols. AtMt. Haruna in summer (Fig. 4b), fine particles were dominantduring highmass concentration events (say,>50 mgm�3). In spring,high mass concentration events were strongly associated with anabundance of coarse particles (Fig. 4a). This suggests that in springcoarse Asian dust particles contributed to the high concentrationevents at Mt. Haruna site.

Three examples of 2007 Asian dust events are shown in Fig. 5, bygraphs of the temporal variation in the number concentration ofcoarse aerosols (>5 mm) at Mt. Haruna and Tsukuba (Meteorolog-ical Research Institute site). In the first event, on 25e26 March, thenumber concentration of coarse particles at Mt. Haruna increased,followed 11 h later by an increase at the Tsukuba site (Fig. 5a). In thesecond event, on 31 Marche1 April, the number concentration atMt. Haruna increased 9 h before increasing at Tsukuba (Fig. 5b). Inthe third event, on 25e26 May 2007, the number concentration atTsukuba increased 6 to 9 h later than at Mt. Haruna (Fig. 5c). As wediscuss later, the controlling factor of the dust arrival at the plain

Fig. 4. Relationship between the PM10 mass concentration and Vol2e5/Vol0.5e1.0 (ratio ofHaruna in (a) spring months and (b) summer months.

sites seemed to be vertical mixing initiated in the morning by solarradiation.

From 31March to 1 April 2007 (Fig. 5b), an Asian dust event wasobserved in most of Japan in routine surface meteorologicalobservations of the Japan Meteorological Agency. At Mt. Haruna,the number concentration of coarse particles (>5 mm) increasedrapidly during the night of 31 March to 1 April. The dust transportroute in this case was simulated by the NOAA hybrid single-particleLagrangian integrated trajectory (HYSPLIT) model. The forwardtrajectory analysis (not shown) that started at 0900 JST on 1 April2007 at a height of 1500 m indicated that the dust plume wasadvected toward the east-southeast at a speed of 30 m s�1. Thisspeed was high enough to transport the air mass from Mt. Harunato Tsukuba in 2 h, but the observed time difference was 9 h.

Fig. 6 shows the depolarization ratios observed by lidar at theNIES in Tsukuba during this dust event. Before 0600 JST on 1 April,the dust layer could not be observed owing to interference fromlow clouds. Once the cloud layer disappeared, the dust plume wasapparent in high depolarization ratios in an aerosol layer between 3and 4 km altitude. During the period 0700e0900 JST, the layerexpanded downward, suggesting that the dust plume was beinginjected into the boundary layer. At 0900 JST, a start of rapidincrease in coarse particle concentration, which means the arrivalof dust plume, was observed by the ground-based OPC in Tsukuba.In the middle of the continuing increase in coarse particleconcentration in the daytime, the depolarization ratio at low alti-tude over Tsukuba showed a rapid increase around 16 JST. Thischange was synchronized with the passing of horizontal windshear followed by southeasterly which could bring the dust plumevia a different route near the surface. Here we discuss the increasein themorning. Because the boundary layer is stable at night, beforesunrise the dust layer was advected eastward without the verticalmixing. As vertical mixing began after sunrise and proceeded, thedust layer reached the surface level. This dust injection into theboundary layer was faster thanwould be due to the gravity settling.

In this connection, Tsunematsu et al. (2006) described a typicalevent in which a dust layer lying over the boundary layer absorbedinfrared radiation from the surface at night, which enhanced thestability at the bottom of the dust layer while resulting in a highconcentration of pollutants within the boundary layer. The increasein stability in such cases may prevent dust aerosols from beinginjected into the boundary layer by convection. Once thermalconvection is activated by sunlight during daytime, verticalexchange between the free troposphere and the boundary layercould bring the dust down into the boundary layer, increasing the

volume concentrations in diameter ranges 2e5 mm and 0.5e1.0 mm) measured at Mt.

Fig. 6. Time-height cross-section of aerosol depolarization ratio measured by the NIESlidar over Tsukuba during the period 30 Marche1 April 2007. The Asian dust eventstarted at 0900 JST on 1 April 2007 in Tsukuba.

10-1

100

101

102

Num

ber C

once

ntra

tion

(# L

-1)

01225May 26May

Haruna

Tsukuba

0 012

25-26 May

10-1

100

101

102

Num

ber C

once

ntra

tion

(# L

-1)

01231M ar 1Apr

Haruna

Tsukuba

0 012

31 March-1 April

10-1

100

101

102

Num

ber C

once

ntra

tion

(# L

-1)

01225Mar 26Mar

Haruna

Tsukuba

0 012

25-26 March

a

b

c

Fig. 5. Temporal variation in the number concentration of coarse (>5 mm) aerosols atMt. Haruna and Tsukuba during Asian dust events (a) on 25e26 March, (b) on 31Marche1 April, and (c) on 25e26 May 2007.

H. Takahashi et al. / Atmospheric Environment 44 (2010) 4638e4644 4643

dust concentration there. The delayed arrival of Asian dust in Tsu-kuba is not a rare case, and is consistent with this explanation. Themechanism controlling the three-dimensional distribution ofaerosol plumes in the free troposphere and boundary layer shouldbe investigated in more detail.

4. Conclusions

As a part of our effort to understand the three-dimensionalstructure of long-range transported aerosol plumes and localpollution, aerosol observations were performed for mass concen-tration and number-size distribution at Mt. Haruna, an isolatedmountain in the Kanto Plain, Japan. We also used PM10 datameasured at Mt. Happo and SPM data from stations within andoutside the Kanto Plain. The seasonal variation at Mt. Haruna wasdifferent from those at the plain sites, which are located in thelower boundary layer. In particular, the averaged mass concentra-tion in May at Mt. Haruna exceeded those at the plain sites, causedby long-range dust transport events. It is notable that mountainobservations can detect long-range transport events of mineraldust as well as pollution from the Asian continent with greatersensitivity than observations at the plain sites at lower altitudes.We demonstrated consistently delayed arrivals of Asian dustplumes at the plain sites compared to the mountain sites, which wetentatively attribute to diurnal mixing effects.

Acknowledgements

This work was partly supported by a Grant-in-Aid for ScientificResearch on Innovative Areas, under the A02-P05 research team“Study of the distribution and movement of aerosols using lidar andground-based monitoring networks” in “Impacts of Aerosols in EastAsia onPlants andHumanHealth”.We thank theAcidDeposition andOxidant Research Center for its permission to use the valuable PM10data from Mt. Happo. We also thank the staffs of Haruna Ropeway,Tobu-Kogyo Co., who kindly aided our observations at Mt. Haruna.We received generous support from Dr. Okada and Mr. Chiba.

References

Aoki, T., Tanaka, T.Y., Uchiyama, A., Chiba, M., Mikami, M., Yabuki, S., Key, J.R., 2005.Sensitivity experiments of direct radiative forcing caused by mineral dustsimulated with a chemical transport model. Journal of the MeteorologicalSociety of Japan 83A, 315e331.

Arao, K., Ishizaka, J., Sugimoto, N., Matsui, I., Shimizu, A., Mori, I., Nishikawa, M.,Aoki, K., Uchiyama, A., Yamazaki, A., Togawa, H., Asano, J., 2006. Yellow sanddust event on 13 April 2003 over western Kyushu, Japan. SOLA 2, 100e103.

Bigg, E.K., 1977. Some properties of the aerosol at Mauna Loa observatory. Journal ofApplied Meteorology 16, 262e267.

Chen, Y., Zhao, C., Zhang, Q., Deng, Z., Huang, M., Ma, X., 2009. Aircraft study ofmountain chimney effect of Beijing, China. Journal of Geophysical Research 114,D08306. doi:10.1029/2008JD010610.

Chin, M., Jacob, D.J., Gardner, G.M., Foreman-Fowler, M.S., Spiro, P.A., Savoie, D.L.,1996. A global three-dimensional model of tropospheric sulfate. Journal ofGeophysical Research 101, 18667e18690.

Fujibe, F., Asai, T., 1979. A study of local winds in Kanto District. Part I: Structures ofwind systems with diurnal variation (in Japanese). TENKI 26, 595e604.

Hayasaki, M., Ohara, T., Kurokawa, J., Uno, I., Shimizu, A., 2008. Episodic pollution ofphotochemical ozone during 8e9 May 2007 over Japan: observational dataanalyses (in Japanese). Journal of Japan Society for Atmospheric Environment43, 225e237.

Husar, R.B., Tratt, D.M., Schichtel, B.A., Falke, S.R., Li, F., Jaffe, D., Gassó, S., Gill, T.,Laulainen, N.S., Lu, F., Reheis, M.C., Chun, Y., Westphal, D., Holben, B.N.,Gueymard, C., McKendry, I., Kuring, N., Feldman, G.C., McClain, C., Frouin, R.J.,Merrill, J., DuBois, D., Vignola, F., Murayama, T., Nickovic, S., Wilson, W.E.,Sassen, K., Sugimoto, N., Malm, W.C., 2001. Asian dust events of April 1998.Journal of Geophysical Research 106, 18317e18330.

Igarashi, Y., Sawa, Y., Yoshioka, K., Matsueda, H., Fujii, K., Dokiya, Y., 2004. Moni-toring the SO2 concentration at the summit of Mt. Fuji and a comparison withother trace gases during winter. Journal of Geophysical Research 109, D17304.doi:10.1029/2003JD004428.

H. Takahashi et al. / Atmospheric Environment 44 (2010) 4638e46444644

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

Katsuno, T., 2006. Monitoring of air pollutants and acid deposition at Mt. Happo(in Japanese).14th National Acid Deposition LectureMeeting(Nagano), pp. 25e34.

Kido, M., Osada, K., Matsunaga, K., Iwasaka, Y., 2001. Diurnal variation of ionicaerosol species and water-soluble gas concentrations at a high-elevation site inthe Japanese Alps. Journal of Geophysical Research 106, 17335e17345.

Kumagai, K., Iijima, A., Tago, H., Tomioka, A., Kozawa, K., Sakamoto, K., 2009.Seasonal characteristics of water-soluble organic carbon in atmospheric parti-cles in the inland Kanto plain. Japan. Atmospheric Environment 43, 3345e3351.

Kurita, H., Sasaki, K., Muroga, H., 1985. Long-range transport of air pollution underlight gradient wind conditions. Journal of Climate and Applied Meteorology 24,425e434.

Mizuno, T., 2006. Study on causes of SPMconcentration trends in recent years in urbanarea of the Kanto District based on data at air quality monitoring stations(in Japanese). Journal of Japan Society for Atmospheric Environment 41, 249e258.

Naoe, H., Heintzenberg, J., Okada, K., Zaizen, Y., Hayashi, K., Tateishi, T., Igarashi, Y.,Dokiya, Y., Kinoshita, K., 2003. Composition and size distribution of sub-micrometer aerosol particles observed on Mt. Fuji in the volcanic plumes fromMiyakejima. Atmospheric Environment 37, 3047e3055.

Okada, K., Kobayashi, A., Iwasaka, Y., Naruse, H., Tanaka, T., Nemoto, O., 1987.Features of individual Asian dust-storm particles collected at Nagoya, Japan.Journal of the Meteorological Society of Japan 65, 515e521.

Osada, K., Ohara, T., Uno, I., Kido, M., Iida, H., 2009. Impact of Chinese anthropogenicemissions on submicrometer aerosol concentration at Mt. Tateyama, Japan.Atmospheric Chemistry and Physics 9, 9111e9120.

Page, S.J., Tuchman, D.P., Vinson, R.P., 2007. Thermally induced filter bias in TEOMmass measurement. Journal of Environmental Monitoring 9, 760e767.doi:10.1039/b704424k.

Samet, J.M., Zeger, S.L., Dominici, F., Curriero, F., Coursac, I., Dockery, D.W.,Schwartz, J., Zanobetti, A., 2000. The national morbidity, mortality, and air

pollution study. Part II: Morbidity and mortality from air pollution in the UnitedStates. Health Effect Institute Research Report 94, 5e70.

Schwartz, J., 1994. Air pollution and daily mortality: a review and meta analysis.Environmental Research 64, 36e52.

Shimizu, A., Sugimoto, N., Matsui, I., Arao, K., Uno, I., Murayama, T., Kagawa, N.,Aoki, K., Uchiyama, A., Yamazaki, A., 2004. Continuous observations of Asiandust and other aerosols by polarization lidars in China and Japan during ACE-Asia. Journal of Geophysical Research 109, D19S17. doi:10.1029/2002JD003253.

Streets, D., Bond, T.C., Carmichael, G.R., Fernandes, S.D., Fu, Q., He, D., Klimont, Z.,Nelson, S.M., Tsai, N.Y., Wang, M.Q., Woo, J.-H., Yarber, K.F., 2003. An inventoryof gaseous and primary aerosol emissions in Asia in the year 2000. Journal ofGeophysical Research 108, 8809. doi:10.1029/2002JD003093.

Takashima, H., Irie, H., Kanaya, Y., Shimizu, A., Aoki, K., Akimoto, H., 2009. Atmo-spheric aerosol variations at Okinawa Island in Japan observed by MAX-DOASusing a new cloud-screening method. Journal of Geophysical Research 114,D18213. doi:10.1029/2009JD011939.

Takeuchi, M., Okochi, H., Igawa, M., 2004. Characteristics of water-soluble compo-nents of atmospheric aerosols in Yokohama and Mt. Oyama, Japan from 1990 to2001. Atmospheric Environment 38, 4701e4708.

Tamamura, S., Sato, T., Ota, Y., Wang, X., Tang, N., Hayakawa, K., 2007. Long-rangetransport of polycyclic aromatic hydrocarbons (PAHs) from the eastern Asiancontinent to Kanazawa, Japan with Asian dust. Atmospheric Environment 41,2580e2593.

Tanaka, M., Shiobara, M., Nakajima, T., Yamano, M., Arao, K., 1989. Aerosol opticalcharacteristics in the yellow sand events observed in May, 1982 at Nagasaki-Part I observations. Journal of the Meteorological Society of Japan 67,267e278.

Tsunematsu, N., Sakai, T., Nagai, T., 2006. Strong capping inversion over the Tokyometropolitan area associated with airborne Asian dust. Geophysical ResearchLetter 33, L19806. doi:10.1029/2006GL026645.

Xuan, J., Liu, G., Du, K., 2000. Dust emission inventory in Northern China. Atmo-spheric Environment 34, 4565e4570.