Aerosol particles, including airborne microorganisms, are transported through the free troposphere from the Asian continental area to the downwind area in East Asia and can influence climate changes, ecosystem dynamics, and human health. However, the variations present in airborne bacterial communities in the free troposphere over downwind areas are poorly understood, and there are few studies that provide an in-depth examination of the effects of long-range transport of aerosols (natural and anthropogenic particles) on bacterial variations. In this study, the vertical distributions of airborne bacterial communities at high altitudes were investigated and the bacterial variations were compared between dust events and non-dust events.
Aerosols were collected at three altitudes from ground level to the free
troposphere (upper level: 3000 or 2500 m; middle level: 1200 or 500 m; and
low level: 10 m) during Asian dust events and non-dust events over the Noto
Peninsula, Japan, where westerly winds carry aerosols from the Asian
continental areas. During Asian dust events, air masses at high altitudes
were transported from the Asian continental area by westerly winds, and laser
imaging detection and ranging (lidar) data indicated high concentrations of
non-spherical particles, suggesting that dust-sand particles were transported
from the central desert regions of Asia. The air samples collected during the
dust events contained 10–100 times higher concentrations of microscopic
fluorescent particles and optical particle counter (OPC) measured particles
than in non-dust events. The air masses of non-dust events contained lower
amounts of dust-sand particles. Additionally, some air samples showed
relatively high levels of black carbon, which were likely transported from
the Asian continental coasts. Moreover, during the dust events, microbial
particles at altitudes of
High-throughput sequencing technology targeting 16S rRNA genes (16S rDNA)
revealed that the bacterial communities collected at high altitudes (from 500
to 3000 m) during dust events exhibited higher diversities and were
predominantly composed of natural-sand/terrestrial bacteria, such as
Airborne microorganisms (bioaerosols) associated with desert-sand and anthropogenic particles were transported through free troposphere from the Asian continents to downwind regions of East Asia and can influence climate changes, ecosystem dynamics, and human health (Iwasaka et al., 2009). Natural dust events from the Asian desert regions carry airborne microorganisms, supporting atmospheric microbial dispersals (Griffin, 2007; Maki et al., 2010; Pointing and Belnap, 2014). Haze days caused by anthropogenic particles from Asian continents also affect airborne microbial abundance and endotoxin levels (Wei et al., 2016). Some studies demonstrated that Asian dust events, including natural and anthropogenic particles, cause vertical mixture of bioaerosols in downwind areas, such as Japan (Huang et al., 2015b; Sugimoto et al., 2012; Maki et al., 2015).
Bioaerosols, which include bacteria, fungi, and viruses, are transported from
ground environments to the free troposphere and account for a substantial
proportion of organic aerosols (Jaenicke, 2005). Bioaerosols are thought to
influence atmospheric processes by participating in atmospheric chemical
reactions and in the formation of cloud-nucleating particles (Pratt et al.,
2009; Morris et al., 2011; Hara et al., 2016b). Indeed, airborne
microorganisms act as ice nuclei that are related to ice-cloud formation
processes (Möhler et al., 2007; Delort et al., 2010; Creamean et al.,
2013; Joly et al., 2013). In particular, ice-nucleation activating proteins
of some microorganisms, such as
In downwind areas of East Asia, the atmospheric bacterial dynamics at high altitudes should be investigated in order to understand the ecological and meteorological influences of airborne bacteria as well as their long-range dispersion. Meteorological shifts and dust events can dramatically alter airborne bacterial communities at high altitudes in Japan (Maki et al., 2013, 2015) because of air masses that originate from heterogeneous environments, including marine, mountainous, urban, and desert areas. The airborne microorganisms around North American mountains (2700 m above sea level) were also found to increase their species diversities in response to Asian dust events (Smith et al., 2013). High-throughput sequencing technology can generate large numbers of nucleotide sequences and the sequencing database has played an important role for investigation of airborne bacterial compositions (Brodie et al., 2007; Woo et al., 2013). Indeed, the analyses using high-throughput sequencing has demonstrated that airborne bacterial populations at ground levels change in response to pollutants from Beijing (Cao et al., 2014) and African dust events (Mazar et al., 2016). To investigate their long-range transported bacteria while avoiding the ground-surface contaminations, the bioaerosol samples collected at high altitudes by aircrafts were analysed using high-throughput sequencing, showing the airborne microbial diversities at high altitudes, ranging from 1000 to 3000 m (DeLeon-Rodriguez et al., 2013; Maki et al., 2015). There are also a few studies on the vertical bacterial distribution from the ground level to the troposphere (DeLeon-Rodriguez et al., 2013; Maki et al., 2015). Nonetheless, while some variations were observed, the specific changes in tropospheric bioaerosols over East Asia, and, in particular, differences between Asian dust and non-dust events remain poorly understood.
Sampling location and helicopter flight routes during the sampling
periods on
Organic aerosol particles, such as bioaerosols, account for high rates of
tropospheric aerosols, ranging from 30 to 80 % (Jaenicke, 2005), and
fluctuate at high concentrations, ranging from 10
In this study, the bacterial communities from different altitudes around the Japanese islands were compared to identify the potential influences of long-range transported air masses on tropospheric bacteria. We used a helicopter for collecting air samples at altitudes ranging from 1200 to 3000 m over the Noto Peninsula, Japan. Helicopter sampling was used to collect chemical components at high altitudes, which has previously been used to avoid contamination from the downwash created by spinning rotors (Watanabe et al., 2016). This air-sampling method can directly collect aerosols moving from Asian continents or marine areas to Japan. We estimated the air mass conditions using the meteorological data obtained during the sampling periods and determined aerosol amounts by using meteorological monitoring and epifluorescence microscopic observation. Bacterial community structures were analysed by using high-throughput sequencing targeting bacterial 16S rRNA genes (16S rDNA).
Aerosol sampling using a helicopter (R44; Robinson, CA, USA) was performed
over coastal areas from Uchinada (36
Sampling information during the sampling periods.
Air samples were collected through sterilized polycarbonate filters
(0.22
In total, 18 air samples were obtained during the sampling periods (Table 1).
Of the two filters used to collect each sample, one filter was used to
determine the particulate abundances under fluorescence microscopy and the
other was stored at
Information regarding weather conditions (temperature, relative humidity, and pressure) was gathered. During the helicopter flight, outside air was transferred from a window into the meteorological-measurement inlet, into which the adaptor of the measurement device (TR-73U; T&D Corporation, Matsumoto, Japan) was inserted, and the temperature, relative humidity, and pressures were sequentially measured. The temperature and relative humidity at an altitude of 10 m were also measured on the roof of a building in Hakui. The depolarization ratio, which was measured by laser imaging detection and ranging (lidar) measurements at Toyama, has been used for the detection of non-spherical aerosols, such as mineral dust particles and/or sea salts.
To track the transport pathways of air masses, 72 h back trajectories were
calculated using the National Oceanic and Atmospheric Administration (NOAA)
HYbrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model
(
The air particles at each altitude were measured using an optical particle counter (OPC: Rion, Tokyo, Japan). The OPC was connected to the meteorological-measurement inlet. The air particles at an altitude of 10 m were also counted using the OPC placed on the roof of a building.
Fluorescent particles stained with DAPI were also counted via epifluorescence
microscopy. Within 2 h of sampling, 1 mL of 1 % paraformaldehyde was
added to one of the filters to fix the aerosols. After a 1 h incubation, the
filter was stained with DAPI at a final concentration of
0.5
After the aerosol particles on the other two filters were suspended in 3 mL
of sterile 0.6 % NaCl solution, the particles were pelleted by
centrifugation at 20 000
Before the analysis of bacterial community structures, USEARCH v.8.01623 (Edgar, 2013) was used to process the raw Illumina sequencing reads. Anomalous sequences were removed with the following workflow. First, the forward and reverse paired-end reads were merged, and the merged reads with lengths outside of the 200–500 bp range or those exceeding six homopolymers were discarded using Mothur v1.36.1 (Schloss et al., 2009). Next, the sequences were subjected to Q-score filtering to remove reads with more than one expected error. Reads occurring only once in the entire dataset (singleton) were then removed. These sequences were clustered de novo (with a minimum identity of 97 %) into 204 operational taxonomic units (OTUs) among the 18 samples. The taxonomy of the representative OTU sequences was assigned using the RDP classifier (Wang et al., 2007) implemented in QIME v1.9.1 (Caporaso et al., 2010). Non-metric multidimensional scaling (NMDS) plot of the pairwise Bray–Curtis distance matrix were used for the classification of all air samples. Greengenes release 13_8 (McDonald et al., 2012) was used as the reference taxonomic database.
All data obtained from MiSeq sequencing data have been deposited in the DDBJ/EMBL/GenBank database (accession number of the submission is PRJEB17915).
Lidar observation of the depolarization ratio in Toyama city, as well
as vertical changes in temperature, relative humidity, and potential
temperature and vertical distributions of concentrations of OPC-counted
particles and DAPI-stained particles from the four sampling events on
19 March 2013
Three-day back trajectories of aerosols that arrived at 2500 m (blue-type lines) and 1200 m (red-type lines) in Hakui, Japan, every hour for 5 h before the completion of sampling time on four dates: 19 March 2013, 28 April 2013, 28 March 2014, and 20 March 2015.
The vertical distributions of the depolarization ratio determined by lidar measurements were assessed for the four sampling events (19 March 2013, 20 March 2015, 28 April 2013, and 28 March 2014). The depolarization ratio increased at the altitude of 3000 m on 19 March 2013 (Fig. 2a), while it decreased at the middle altitude of 1000 m. The air mass on 20 March 2015 showed high values of depolarization ratio at altitudes of 2500 and 500 m, consistent with the vertical distribution of non-spherical (mineral dust) particles over the Noto Peninsula (Fig. 2d). A 3-day back-trajectory analysis indicated that the air mass at 3000 m on both sampling dates came from the Asian desert region to the Noto Peninsula (Hakui) immediately across the Sea of Japan (Fig. 3). These results indicated the dust-event occurrence on 19 March 2013 was specific to the upper altitude of 3000 m, while the dust event on 20 March 2015 occurred between the altitudes of 2500 and 500 m. Moreover, samples collected on 28 April 2013 and 28 March 2014 exhibited a low depolarization ratio (Fig. 2b–c), and the air masses on these two sampling dates came from areas of North Asia, including eastern Siberia (Fig. 3).
The air-sampling periods from the March 2014 time series (from 23 to 29 March 2014) and the March 2015 time series (from 16 to 21 March 2015) showed different patterns of depolarization ratio and air mass trajectory roots between the two series (Figs. 4 and 5). Depolarization ratio from March 2014 maintained lower values (Fig. 4a) and the trajectory lines changed the roots from eastern Siberia to the Korean Peninsula before surrounding the Japanese islands (Fig. 4c). In contrast, the sampling period during March 2015 had substantially higher depolarization ratio, indicating a strong presence of mineral dust particles (Fig. 5a), and air masses at 3000 m consistently originated from the Asian desert regions (Fig. 5c).
Temperatures from 19 March, 28 April 2013, 28 March 2014, and 20 March 2015 increased from approximately 290 K to approximately 300 K at middle altitudes (500 and 1200 m) (Fig. 2). The temperature profile clearly indicated the presence of a thin boundary under the upper altitudes (2500 and 3000 m), which suggested that there is a difference in air qualities between the middle and upper altitudes (Table 1). During the March 2014 time series, temperatures dynamically changed at altitudes of approximately 1200 m, while those from the March 2015 time series (16, 17, and 21 March 2015) were stable at 1200 m (Figs. S1 and S2 in the Supplement). These results indicate that the boundary layers were located at 1200 m during the March 2014 time series, whereas the tropospheric air transported by westerly winds was suspended at the sampling altitudes (500 and 1200 m) used during the March 2015 time series.
Aerosol particle concentrations from the ground level to the troposphere were
measured using OPC to compare the vertical distributions of aerosols from the
four sampling events. The OPC-measured particles on 19 March 2013 and
20 March 2015 maintained similar concentrations below the troposphere
(Fig. 2a, d), while the concentrations on 28 April 2013 and 28 March 2014
decreased 1 or 2 orders of magnitude between the troposphere and ground
level (Fig. 2b, c). At high altitudes (2000 to 2500 m), the course particles
(greater 1.0
OPC measurements indicated that air samples collected at 1200 m during the
March 2015 time series consistently contained course particles at 1 or 2
orders of magnitude higher in concentration (
Based on the above observations, the sampled air masses that were influenced by Asian dust events and included dust particles were categorized as “dust samples”. The sampled air masses that were not influenced by dust events or contained less dust particles were categorized as “non-dust samples”, in relation to the presence or absence of dust events as the source of the aerosol samples (Table 1).
Using epifluorescence microscopy with DAPI staining, the aerosol particles in
the 18 air samples emitted several types of fluorescence, categorized as
white, blue, yellow, or black (Fig. S3). White fluorescence particles, (white
particles) were indicative of mineral particles originating from the sand or
soil. Microbial (prokaryotic) particles stained with DAPI emitted blue
fluorescence, forming coccoid- or bacilli-like particles with a diameter
The dust samples from upper altitudes (2500 and 3000 m) contained 5 to
100 times higher concentrations of microbial, organic, and white particles
than the concentrations detected in the non-dust samples (Fig. 2). In the
upper-altitude dust samples, the concentration of mineral particles ranged
from 7.77
All types of fluorescence particles were also observed in the sequentially
collected air samples at 1200 m in the March 2015 time series (except for
2500 m on 20 March) and the March 2014 series. The dust samples examined
from the March 2015 series had higher concentrations of total particles than
the non-dust samples of the March 2014 series (Figs. 4 and 5). The mineral
particles detected in the March 2014 series fluctuated at low concentrations
from 3.39
Vertical variations in bacterial compositions at
Changes in bacterial compositions at
Comparison of the bacterial compositions among all air samples
collected over the Noto Peninsula.
For the analysis of the prokaryotic composition in the 18 samples, we
obtained 645 075 merged paired-end sequences with the lengths ranging from
244 to 298 bp after quality filtering, and the sequence library size for
each sample was normalized at 1500 reads. The 16S rDNA sequences were divided
into 204 phylotypes (sequences with
The vertical distributions of bacterial compositions showed different patterns between dust-event days and non-dust days (Fig. 6). In the dust samples collected at upper altitudes, phylotypes belonging to the phylum Bacilli accounted for more than 60.5 % of the total and were mainly composed of members of the families Bacillaceae and Paenibacilliaceae (Fig. 6). Bacterial numbers from the phylum Bacilli decreased at lower altitudes during dust events, and the phylotypes of Cyanobacteria, Actinobacteria, and Proteobacteria increased in relative abundance in the samples collected at middle and low altitudes (13H319-m, 13H319-l, and 15H320-m).
Cyanobacteria, Actinobacteria, and Proteobacteria sequences also dominated in the air samples collected during non-dust events (13H428-m, 14H328-u, 14H328-m, and 14H328-l). Specifically, Actinobacteria phylotypes increased in their relative abundance, ranging from 14.1 to 24.7 % in the non-dust samples collected on 28 March 2014. Proteobacteria phylotypes containing several bacterial families occupied a high relative abundance, ranging from 60.5 to 85.3 % in the non-dust samples 13H428-u, 13H428-m, 14H328-u, 14H328-m, and 14H328-l. In particular, the non-dust samples collected on 28 March 2014 included the Alphaproteobacteria phylotypes, which are composed of members of the families Phyllobacteriaceae and Sphingomonadaceae. Most Betaproteobacteria, phylotypes including the families Oxalobacteraceae and Comamonadaceae, were specific to the non-dust samples collected at 1200 and 2500 m on 28 April 2013.
Cyanobacteria phylotypes, which were randomly detected from both dust samples and non-dust samples, particularly increased in both the non-dust sample collected at 10 m on 28 April 2013 and the dust sample collected at 3000 m on 20 March 2015, with a relative abundance of 15.3 and 74.6 %, respectively. Bacteroidia phylotypes also randomly appeared in several air samples, regardless of the dust-event influences, and were present at maximal levels in the non-dust sample 13H319-m, with a relative abundance of 35.6 %.
Researches targeting bacterial communities associated with Asian dust events.
Sequential variations in the bacterial composition of air samples at altitudes of 1200 or 2500 m were compared between dust-event periods (March 2015 series) and non-dust periods (March 2014 series). During the March 2015 dust event, phylotypes of the family Bacillaceae in the class Bacilli occupied more than 53.0 % of the relative abundance in the four dust samples collected (Fig. 7). Cyanobacteria phylotypes related to the marine cyanobacterium Synechococcus uniquely appeared in the dust samples of the March 2015 series; their abundance fluctuated, with values ranging from 12.5 to 14.8 % between 16 and 20 March 2015 before decreasing to 1.5 % on 20 March.
During the non-dust periods of the March 2014 series at the middle altitude, the relative abundance of Actinobacteria phylotypes belonging to the family Micrococcaceae was occupied 59.9 % on 23 March, decreased to 19.5 % on 24 March, and disappeared from samples collected on 29 March. Corresponding to the decrease in Actinobacteria phylotypes, Alpha- and Gammaproteobacteria phylotypes showed an increasing trend from 30.6 to 96.8 % between 23 and 29 March 2014 (Fig. 7a). Alphaproteobacteria phylotypes belonging to the families Sphingomonadaceae, and Phyllobacteriaceae consistently appeared throughout the sampling periods of the March 2014 series and occupied a maximum relative abundance of 72.9 and 22.3 %, respectively. For Gammaproteobacteria, the Xanthomonadaceae sequences dominated at a relative abundance of 18.3 and 5.4 % in the non-dust samples 14H325-m and 14H329-m, respectively, during the air mass was suspended the Japanese islands for a few days.
Westerly winds blowing over East Asia disperse airborne microorganisms associated with dust mineral particles (Maki et al., 2008) and anthropogenic particles (Cao et al., 2014; Wei et al., 2016), influencing the abundances and taxon compositions of airborne bacteria at high altitudes over downwind areas, such as Noto Peninsula (Maki et al., 2013). In this investigation, the increases in aerosol particles (dust particles) and associated microbial particles were observed over the Noto Peninsula during the dust events of 19 March 2013 and 20 March 2015 (Figs. 2 and 4). At the two sampling dates, the air mass including microbial particles had travelled from the Asian desert region throughout the anthropogenic polluted areas (Fig. 2), and the dust particles entered the Japanese troposphere and were maintained at high altitudes (19 March 2013) or mixed with the ground-surface air (20 March 2015). During non-dust days, the air masses at high altitudes came from several areas, including the eastern region of Siberia, Asian continental coasts (Korean Peninsula), the Sea of Japan, or surrounding Japanese islands, and mixed with ground-surface air over the Noto Peninsula. The air samples collected during dust and non-dust events were valuable for understanding the westerly wind influences on vertical distributions and sequential dynamics of airborne bacteria at high altitudes over the downwind regions.
The microscopic fluorescence particles of all samples could be separated into four categories: mineral (white), microbial (blue), organic (yellow), and black carbon (black) particles (Fig. S3), which were observed in the previous air samples collected during dust events (Maki et al., 2015). The amount of microbial particles increased at high altitudes during dust events, suggesting that the dust events directly carried bacterial particles to the troposphere over downwind areas. At low altitudes, similar concentrations of fluorescent particles were observed in air samples collected between dust events (13H319-l) and non-dust events (13H428-l) (Fig. 2) because the dust particles did not reach the ground surface on the dust-event days. In the absence of the influences of dust events, the aerosols mainly originated from local environments in Japanese areas.
Organic particles also increased during dust events and in the ratios between
all particles related to the dust events. The organic particles originate
from proteins and other biological components (Mostajir et al., 1995). The
tropospheric aerosols would be composed of organic particles at high rates
ranging from 30 to 80 % (Jaenicke, 2005), and organic particle
concentrations fluctuated from 10
The appearance of black carbon most likely originated from anthropogenic activities, such as biomass burning, industrial activities, and vehicle exhaust (Chung and Kim, 2008). In the anthropogenic regions of eastern China, anthropogenic particles originating from human activities are expected to comprise more than 90 % of dust particles (Huang et al., 2015a). When the westerly winds are strongly blowing over the Noto Peninsula, the black carbon particles at upper altitudes (3000 m) are thought to mainly derive from continental anthropogenic regions.
Dust events and air pollutant occurrences changed the airborne bacterial communities over the downwind areas, such as Beijing (Jeon et al., 2011; Cao et al., 2014) (Table 2) and eastern Mediterranean areas (Mazar et al., 2016). The westerly winds blowing over East Asia would transport airborne bacteria to the high-altitude atmosphere over the Noto Peninsula (Maki et al., 2015) and North American mountains (Smith et al., 2013). Our box plots analysis suggested that changes in the bacterial diversity in the dust samples would be more stable than in the non-dust samples (Fig. 8a). Furthermore, using a NMDS plot, the bacterial compositions in the dust samples could be distinguished from non-dust samples (Fig. 8b). Thus, the aerosol particles transported by Asian dust events changed the atmospheric bacterial composition at higher altitudes over downwind areas.
The phylotypes in the dust samples were predominately clustered into the
class Bacilli (Fig. 4a), while the non-dust samples mainly included the
phylotypes of the classes Alpha, Beta, and Gammaproteobacteria and
Actinobacteria. Our previous investigations indicated that the bacterial
communities at an altitude of 3000 m over the Noto Peninsula included more
than 300 phylotypes, which were predominantly composed of Bacilli phylotypes
(Maki et al., 2015). Bacterial groups belonging to Bacilli, Proteobacteria,
and Actinobacteria have been reported as airborne bacteria around European
mountains (Vaïtilingom et al., 2012) and over Asian rural regions
(Woo et al., 2013). Some Bacilli isolates were found to act as ice-nucleating
agents and to be involved in ice cloud (Matulova et al., 2014; Mortazavi et
al., 2015). Isolates of Gammaproteobacteria isolates were obtained from
mineral dust particles (Hara et al., 2016a), glaciated high-altitude clouds
(Sheridan et al., 2003), and plant bodies (Morris et al., 2008), and some
isolate species, such as
In some dust-event samples collected at high altitudes (13H319-u, 15H320-u,
and 15H320-m), Bacilli sequences accounted for more than 52.7 % of the
total number of sequences (Fig. 6). Back trajectories on 19 March 2013 and
20 March 2015 came from the Asian desert region to the Noto Peninsula. Some
Bacilli members can form resistant endospores that support their survival in
the atmosphere (Nicholson et al., 2000). The
The Bacilli sequences showed different vertical variations between the two dust events on 19 March 2013 and 20 March 2015. On 19 March 2013 (13H319-m), the relative abundances of Bacilli sequences decreased dynamically from 3000 to 1200 m, while unstable atmospheric layers on 20 March 2015 most likely mixed the long-range transported bacteria with the regional bacteria over the Noto Peninsula. A previous investigation also demonstrated the vertical mixture of airborne bacteria over Suzu in the Noto Peninsula (Maki et al., 2010).
Actinobacteria sequences decreased in relative abundance between 23 and 29 March 2014, corresponding with changes in the air mass trajectory roots from north Asian regions, such as eastern Siberia and Japan (Fig. 7). Furthermore, Actinobacteria sequences appeared in the samples collected from air masses that were transported throughout the Korean Peninsula on 19 March 2013, 28 April 2013, and 20 March 2015. Actinobacteria members are frequently dominant in terrestrial environments but seldom survive in the atmosphere for a long time, because they cannot form spores (Puspitasari et al., 2015). However, the family Micrococcaceae in Actinobacteria was primarily detected from anthropogenic particles collected in Beijing, China (Cao et al., 2014). Over anthropogenic source regions for Asian continents, anthropogenic particles occupy more than 90 % of dust particles and originate from soils disturbed by human activities in cropland, pastureland, and urbanized regions (Huang et al., 2015a; Guan et al., 2016). Air masses transported from the continental coasts are expected to include a relatively high abundance of Actinobacteria members associated with anthropogenic particles.
Natural dust particles from Asian desert areas (Taklamakan and Gobi) are transported in the free troposphere (Iwasaka et al., 1988) and vertically mixed with anthropogenic particles during the transportation processes (Huang et al., 2015a). In some cases, short-range transport of air masses would carry only anthropogenic particles to Japan, because the anthropogenic particles are often dominant in Asian continental coasts (Huang et al., 2015a). Actinobacteria members may have been transported with anthropogenic particles from continental coasts.
Proteobacteria sequences increased in their relative abundances at high altitudes during non-dust sampling dates (13H428-u, 13H428-m, 14H328-u, 14H328-m, and March 2014 series), when air mass origins at 1200 m changed from the Korean Peninsula to Japan (Fig. 7). Proteobacteria members were the dominate species in the atmosphere over mountains (Bowers et al., 2012; Vaïtilingom et al., 2012; Temkiv et al., 2012), in the air samples collected on a tower (Fahlgren et al., 2010), and from the troposphere (DeLeon-Rodriguez et al., 2013; Kourtev et al., 2011). In the phylum proteobacteria, the families Phyllobacteriaceae, Methylobacteriaceae, and Xanthomonadaceae were predominately detected from the non-dust samples and are associated with plant bodies or surfaces (Mantelin et al., 2006; Fürnkranz et al., 2008; Khan and Doty, 2009; Fierer and Lennon, 2011). The Betaproteobacteria sequences in the non-dust samples mainly contained the Oxalobacteraceae and Comamonadaceae families, which are commonly dominate in freshwater environments (Nold and Zwart, 1998) and on plant leaves (Redford et al., 2010). In addition, the class Alphaproteobacteria in the non-dust samples also included marine bacterial sequences belonging to the family Sphingomonadaceae (Cavicchioli et al., 2003). Bacterial populations originating from marine areas are prevalent in cloud droplets (Amato et al., 2007), in air samples collected from the seashores of Europe (Polymenakou et al., 2008), in storming troposphere (DeLeon-Rodriguez et al., 2013), and at high altitudes in Japanese regions (Maki et al., 2014), suggesting that the marine environments represent a major source of bacteria in clouds. The air masses suspended over the Sea of Japan or Japanese islands during non-dust events (the March 2014 series) could include a high relative abundance of airborne bacteria, which were transported from the surface-level air over the marine environments and the regional phyllosphere.
Sequences originating from Synechococcaceae (in the class Cyanobacteria)
randomly appeared in the MiSeq sequencing databases results obtained from air
samples, regardless of dust-event occurrences.
Bacteroidetes sequences were detected in some air samples collected during Asian dust and non-dust events. Members of the phylum Bacteroidetes, which were composed of the families Cytophagaceae, associate with organic particles in terrestrial and aquatic environments (Turnbaugh et al., 2011; Newton et al., 2011). Furthermore, these bacterial populations dominate the atmosphere and sand of desert areas, where plant bodies and animal feces are sparsely present (Maki et al., 2016). These bacterial groups possibly originated from organic-rich microenvironments in several areas, such as desert and marine areas.
Air samples including airborne bacteria were sequentially collected at high altitudes over the Noto Peninsula during dust events and non-dust events. The sampled air masses could be categorized based on sample types with (dust samples) and without (non-dust samples) dust-event influences. Bacterial communities in the air samples displayed different compositions between dust events and non-dust events. The dust samples were dominated by terrestrial bacteria, such as Bacilli, which are thought to originate from the central desert regions of Asia, and the bacterial compositions were similar between the dust samples (Table 2). In contrast, the air masses of non-dust samples came from several areas, including northern Asia, continental coasts, marine areas, and Japan regional areas, showing different variations in bacterial compositions between the sampling dates. Some scientists have attempted to apply airborne bacterial composition as tracers of air mass sources at ground level (Bowers et al., 2011; Mazar et al., 2016). In this study, the terrestrial bacteria, such as Bacilli and Actinobacteria members (Bottos et al., 2014), were dominant populations in the air samples transported from Asian continental areas. The air samples when the air mass was suspended around Japanese islands, mainly included the members of the classes Alpha- (Phyllobacteriaceae and Methylobacteriaceae), Gamma-, and Betaproteobacteria, which are commonly dominated in phyllosphere (Redford et al., 2010; Fierer and Lennon, 2011) or freshwater environments (Nold and Zwart, 1998). The atmospheric aerosols transported via marine areas include a high relative abundances of marine bacteria belonging to classes Cyanobacteria (Choi and Noh, 2009) and Alphaproteobacteria (Sphingomonadaceae) (Cavicchioli et al., 2003). This study suggested that bacterial compositions in the atmosphere can be used as air mass tracers, which could identify the levels of mixed air masses transported from different sources.
However, one limitation of our investigation is that the number of samples analysed was not sufficient to cover entire changes in airborne bacteria at high altitudes over the Noto Peninsula. Although the airborne bacterial composition during non-dust periods was found to change dynamically, only a few types of variation were followed in this investigation. In the future, greater numbers of samples, which are sequentially collected at high altitudes using this sampling method, will need to be originated to more accurately evaluate bioaerosol tracers. Since helicopter sampling procedures require sophisticated techniques and are expensive, the sample numbers at high altitudes are difficult to increase. Air sampling at high altitudes should be combined with sequential ground-air sampling to advance the understanding of the influence of westerly winds on airborne bacterial dynamics in downwind areas. Metagenomic analyses and microbial culture experiments would also provide valuable information about airborne microbial functions relating to ice-nucleation activities, chemical metabolism, and pathogenic abilities.
The lidar measurement data are accessible from
The authors declare that they have no conflict of interest.
This article is part of the special issue “Anthropogenic dust and its climate impact”. It is not associated with a conference.
We are thankful for the advice given by Richard C. Flagan of California
Institute of Technology and the sampling support from Atsushi Matsukia and
Makiko Kakikawa of Kanazawa University. Trajectories were produced by the
NOAA Air Resources Laboratory (ARL), which provided the HYSPLIT transport and
dispersion model and/or READY website (