Article Full Text - Aerosol and Air Quality Research

Transcription

Aerosol and Air Quality Research, 13: 957–976, 2013
Copyright © Taiwan Association for Aerosol Research
ISSN: 1680-8584 print / 2071-1409 online
doi: 10.4209/aaqr.2012.07.0184
Characteristics of PM2.5 Haze Episodes Revealed by Highly Time-Resolved
Measurements at an Air Pollution Monitoring Supersite in Korea
Seung-Shik Park1*, Sun-A Jung2, Bu-Joo Gong2, Seog-Yeon Cho3, Suk-Jo Lee2
1
Department of Environmental Engineering, Chonnam National University, 300 Yongbong-dong, Gwangju 500-757, Korea
Division of Climate & Air Quality Research, National Institute of Environmental Research, Kyungseo-dong, Seo-Gu,
Incheon 404-170, Korea
3
Department of Environmental Engineering, Inha University, 100 Inha-Ro, Nam-gu, Incheon 402-751, Korea
2
ABSTRACT
Hourly measurements of PM2.5, organic and elemental carbon (OC and EC), inorganic ionic species, and elemental
constituents were made between February 1 and March 31, 2011, at a South Area Supersite at Gwangju, Korea. Over the
two-month study period, daily PM2.5 mass concentration exceeded the 24-hr average Korean NAAQS of 50.0 μg/m3 on 20
days, of which two pollution episodes (episodes I and II) are investigated. Episode I (February 01–08) is associated with
regional pollution along with wild fire smoke emissions over southern China, and also characterized by high CO/NOx
ratios and high K+ concentrations. While episode II (March 11–12) is characterized by locally produced pollution with low
CO/NOx ratios, and broad variations in OC, EC, and NO3– concentrations.
For episode I, the 1-hr PM2.5 mass concentration ranged from 27 to 159 μg/m3 with a mean of 88 μg/m3. Although a low
OC/EC ratio (3.4) was observed for February 04–05 when high K+ concentrations occurred, significant linear relationships
between EC – OC (R2 = 0.82), OC – K+ (R2 = 0.77), and K+ – K+/OC (R2 = 0.77) suggest the contribution of carbonaceous
aerosols from long-range transport of biomass burning plumes occurring in the southern regions of China during episode I,
as well as local traffic emissions.
For episode II, the hourly PM2.5 mass concentration ranged from 23 to 116 μg/m3 with a mean of 78 μg/m3. Hourly
maximum contributions of SO42– and NO3– to the PM2.5 mass were 54 and 66%, respectively. An elevated NO3– concentration
was observed along with high OC and EC concentrations, suggesting the influence of local emissions. The pattern of SO42–
variations in relation to wind direction and the strong correlation between SO2 and SO42– suggest local SO2 emissions were
likely an important source of SO42– at the site.
Keywords: Korean PM Supersites; Semi-continuous measurements; Wildfires burning emissions; OC/EC ratio; OC – K+
relationship.
INTRODUCTION
Particulate matter (PM) smaller than 2.5 μm in
aerodynamic diameter (PM2.5) is of great concern due to the
transportability in the human body. PM2.5 in the atmosphere
has been linked to the formation of haze and changes in the
atmosphere’s radiative balance (Charlson et al., 1987), as
well as adverse effects on human health, crops, and materials
(Chameides et al., 1999). Haze pollution has been studied
intensively on its impact on air quality, visibility, climate, and
public health (Chameides et al., 1999; Kim et al., 2001;
Okada et al., 2001; Schichtel et al., 2001; Menon et al., 2002;
*
Corresponding author. Tel.: 82-62-530-1863;
Fax: 82-62-530-1859
E-mail address: [email protected]
Watson, 2002; Watson and Chow, 2002; Chen et al., 2003;
Weber et al., 2003; Yadav et al., 2003; Kang et al., 2004;
Park et al., 2006a; Sun et al., 2006; Kim et al., 2008; Jung
and Kim, 2011; Oanh and Leelasakultum, 2011). Results from
these studies have revealed that haze pollution is strongly
correlated with meteorological conditions, emissions of
pollutants from anthropogenic sources, and gas-to-particle
conversion. Emissions of air pollutants and air stagnation
conditions are favorable to haze formation which alters the
chemical composition of atmospheric aerosol particles. It
has been suggested that haze is not restricted to urban areas
and is often regional in nature. Urban PM2.5 is typically
composed of local and regional sources such as carbon and
nitrate are contributed significantly from local sources,
while sulfate particles are mostly from regional sources.
Therefore, understanding the association between local and
regional emissions and the formation of haze is critical to
effective air quality regulation.
958
Park et al., Aerosol and Air Quality Research, 13: 957–976, 2013
Although long integrated filter-based measurements in the
above-mentioned studies have mainly been made to determine
aerosol chemical constituents during haze events, their
temporal resolution is typically insufficient to observe shortterm variations. The daily averaging times used in long
integrated filter-based measurements tend to smooth out
much of the variability, thus limiting the understanding of
the factors influencing of haze formation dynamics and the
short-term exposures that result in adverse health impacts.
For these reasons, time-resolved measurements of PM2.5 and
its chemical components were conducted to better understand
PM2.5 sources and the dynamics of haze formation because
chemical composition of atmospheric aerosol particles is
highly variable in space and time, and is dependent on
meteorological parameters (Weber et al., 2003; Park et al.,
2005b, 2005c, 2006a; Solomon and Sioutas, 2008). It has
been indicated that during haze events, excess PM2.5 is often
largely due to elevations in the concentration of one or two
of the major chemical species. Such elevations are often
induced by near-by sources and/or regional sources. These
studies indicated that short-term transients in carbonaceous
species, nitrate, CO, and NOx levels were often observed in
the morning hours or sometimes in evening hours. In contrast,
sulfate concentrations peak during summertime afternoons
when ozone levels are elevated, or can be strongly associated
with the regional haze events.
As a part of the development of effective control strategies
to reduce the impacts of atmospheric PM including yellow
sand and long-range transported aerosols, the Korean EPA in
2007 launched an intensive air pollution monitoring research
program known as the Supersite program, which is similar
to the U.S. EPA Supersite program (Solomon and Sioutas,
2008). The primary objectives of the Korean Supersite
program are to monitor the physicochemical characteristics of
the yellow sand and long-range transported aerosols, estimate
the domestic background concentrations of the aerosol
particles at remote background sites, provide unprecedented
physicochemical characterization of ambient PM in urban
areas, and evaluate semi-continuous measurement methods
of PM and its chemical speciation data. This study was carried
out at a South Area Supersite in Gwangju, Korea, which
began in 2009. The South Area air pollution supersite study
was designed to develop an extended set of time-resolved and
compositionally resolved PM2.5 data for use in apportioning
sources and resolving their local and regional contributions.
The objectives of this study was to compare results between
filter-based integrated and semi-continuous measurements of
PM2.5 mass and its major chemical constituents from data
gathered for approximately two months in 2011 at a South
Area Supersite in Gwangju, Korea. In addition to evaluating
the near continuous methods, an attempt was made to
elucidate the key chemical characteristics of two PM2.5 haze
pollution episodes revealed by time-resolved measurement
methods.
EXPERIMENTAL
Description of the South Area Supersite at Gwangju
The South Area Supersite (latitude 35.23°N, longitude
126.85°E) is located in the southern part of Korea, as
depicted in Fig. 1. The site is located in a semi-urban
commercial and residential area which is surrounded by
agricultural lands and traffic roads, and is ~20 km northwest
of downtown Gwangju and northeast of two industrial
complexes. Emissions from these areas to the site could be
resolved during typical western and southern flow regimes.
The closest major traffic road lies approximately 0.3 km
southeast of the site, and the Honam express highway is
located 1.5 km west of the site. Meteorological data were
monitored at the Gwangju regional meteorological station,
located 7 km south of the site. Gwangju has a population of
about of about 1.4 million in a 501.4 km2 area, with 350,000
motor vehicles, of which gasoline vehicles account for ~70%.
In addition, ~85% of total air pollution emissions in the city
are attributed to vehicular sources. Air quality in Gwangju has
been reported to be influenced by local pollution and longrange transported pollution from polluted regions of China
(Kim et al., 2001; Park et al., 2005a, 2006b, 2010, 2012;
Jung et al., 2010; Jung and Kim, 2011; Park and Cho, 2011).
High levels of PM pollution are also observed due to the
burning of agricultural wastes in farming areas nearby the
Supersite (Ryu et al., 2004, 2007; Park et al., 2006b; Jung
and Kim, 2011). Thus, the South Area Supersite located in
Gwangju is an excellent choice for studying the properties
of local, regional, and biomass burning aerosol emissions
and their effect on urban air quality, in order to investigate
our hypothesis regarding time-resolved measurements and
aerosol age.
Highly time-resolved measurements of PM2.5, organic and
elemental carbon (OC and EC), ionic species, and elemental
species have been made at the Gwangju Supersite at 1-hr
intervals since January 2010 using commercial semicontinuous instruments. However, in the present study,
only the measurement data recorded between February 1 and
March 31, 2011 were utilized due to frequent malfunction of
some instruments. Hourly ambient temperature ranged from
–7.4 to 16.7°C (a mean of 3.1°C) during February and from
–2.9 to 19.7°C (a mean of 5.8°C) during March. Average
relative humidity (RH) during February and March was
64.8% (18.2–99.9%) and 55.5% (16.9–99.5%).
24-hr PM2.5 Mass and Speciation Measurements
For comparison of the semi-continuous measurements,
24-hr integrated filter-based PM2.5 samples were collected
between midnight and midnight with three sets of sequential
PM2.5 samplers (PMS103, APM Korea) and then used for
determining the total mass concentration, the amounts of
carbonaceous species (OC and EC), eight ionic species (Cl–,
NO3–, SO42–, Na+, NH4+, K+, Ca2+, and Mg2+), and 29
elemental species (Si, S, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co,
Ni, Cu, Zn, As, Se, Br, Rb, Sr, Mo, Cd, Sn, Sb, Te, Cs, Ba,
Hg, Pb, and Bi). During the filter-based sampling of aerosol
particles, the artifacts that can arise due to particle-to-particle
interactions, gas-particle interactions and the dissociation
loss of semi-volatile species can change the collected aerosol
composition. Positive and negative artifacts may have been
created in this study because the sequential sampler used
did not have a denuder to remove gases prior to sampling
959
Fig. 1. Area map of Gwangju PM Supersite.
particles or a backup filter to collect vapor from particles
collected on the filter. Issues arising from the sampling
artifacts are discussed below.
The Teflon filters (Teflo membrane, 2.0-μm pore size,
Gelman Science, USA) were weighed before and after
sample collections with a microbalance of 1 μg sensitivity
and analyzed for elemental components using X-ray
fluorescence (XRF). The filters were conditioned for about
48hr in a clean chamber maintained at RH of 40% and
temperature of 20°C. Aerosol samples collected on 47-mm
diameter quartz fiber filters were analyzed for OC and EC
using the NIOSH thermal-optical transmittance (TOT)
method. In our protocol, OC and EC were determined as
follows. OC was evolved under a stream of high purity He
(99.999% higher). The sample was heated in five steps to a
final temperature of 800°C in order to volatilize OC and
EC in a helium atmosphere. To evolve EC and pyrolyzed
OC, which were removed in the second part of the analysis,
the sample was first cooled to 550°C, and then heated under
a mixture of 2% O2/98% He, in three temperature steps, to
a final temperature of 800°C. The TOT analyzer utilized
laser transmission to correct for OC charring. EC was
determined as the carbon evolved after the filter transmittance
returned to its initial value. The detection limits of OC and
EC, which are defined as twice the average of the field blanks,
were 0.15 and 0.03 μg C/m3, respectively. The precision of
the OC and EC measurements was 4.0 and 7.5%, respectively.
Zefluor filters (Zefluor, 2 μm pore size, Gelman Science,
USA) from a sequential sampler were analyzed for eight
ionic species (Cl–, NO3–, SO42–, Na+, NH4+, K+, Mg2+, and
Ca2+). In order to extract the ionic species from the sampled
filters, each filter was first put into a 125-mL HDPE vial,
and wetted in 0.2 mL of HPLC-grade C2H5OH and then 30
mL of distilled deionized water. The solution was extracted
using ultra-sonication for 30 min and mechanically extracted
again for 60 min. All the extracts were filtered using 0.4-μm
960
Teflon filters and analyzed by ion chromatography (Metrohm
Modula IC, model 818 IC pump/819 IC detector). Filter
blank values were also applied for background subtraction
for each of the filters. The IC system consisted of an anion
IC (Metrohm Modula with a suppressor, equipped with a
Metrohm Metrosep A Supp-5, 4 × 150 mm, anion column)
and a cation IC (Metrohm Modula without a suppressor,
equipped with a Metrohm Metrosep C4, 4 × 150 mm, cation
column). The eluants were 3.5 mM sodium carbonate
(Na2CO3)/1.0 mM sodium bicarbonate (NaHCO3) for the
anion IC and 4.0 mM nitric acid (HNO3)/0.7 mM dipicolinic
acid for the cation IC. The eluant flow rate for the anion and
cation was maintained at 0.7 and 0.9 mL/min, respectively.
The measured detection limits for Cl–, NO3–, SO42–, Na+,
NH4+, K+, Ca2+, and Mg2+ were 12.3, 12.1, 14.9, 11.3, 4.7,
8.5, 12.4, and 4.7 ng/m3, respectively.
Semi-Continuous PM2.5 Measurements
PM2.5 Mass Concentration
PM2.5 mass concentration was determined continuously
every one hour with a beta-attenuation monitor (BAM1020,
MetOne Instrument Inc., USA). Integrated 24-h averages
of 1-hr PM2.5 data were compared with those from 24-h
integrated PM2.5 speciation sampler. The corresponding
regression relationships: PM2.5BAM (μg/m3) = (1.02 ± 0.03)
PM2.5FRM (μg/m3) + (1.50 ± 1.27), R2 = 0.97, indicated that
the continuous BAM1020 measurements provided excellent
agreement with 24-h integrated filter-based PM2.5 mass.
Organic and Elemental Carbon Measurements
Concentrations of OC and EC in PM2.5 were measured
using a model-4 semi-continuous OC-EC field analyzer
(Sunset Laboratory Inc., USA) with 1-hr time resolution.
This analyzer is based on the NIOSH method 5040 (NIOSH,
1996). In the carbon analyzer, OC and EC were determined
as follows: OC was evolved under a stream of ultrahigh
purity He (99.999% higher) while the sample was heated in
two temperature steps (600°C for 80 s and 840°C for 90 s).
The evolved carbon was oxidized to CO2 in an oxidizing
oven. The CO2 is then swept out of the oxidizing oven by a
helium stream and measured directly by a self-contained
non-dispersive infrared (NDIR) detector system. To evolve
EC and pyrolyzed OC, the sample was cooled to 650°C, and
then heated under a mixture of He/O2 in two temperature
steps (650°C for 35 s and 880°C for 105 s). The analyzer
accounted for this pyrolysis by defining the split between the
OC and EC as the point when the analysis transmittance is
equal to the transmittance before analysis. The EC is
determined as the carbon evolved after the filter transmittance
returned to its initial value.
Ambient air was drawn at 8.0 L/min through a PM2.5 sharpcut cyclone. A carbon impregnated multi-channel parallelplate diffusion denuder was used to remove any semi-volatile
organic vapors absorbed on the quartz filter media (Turpin
et al., 2000). However, the use of a denuder also disturbs
the gas/particle equilibrium at the quartz fiber filter, which
can cause OC to volatilize from the filter (Turpin et al.,
2000). During the 2-month study period, a total of 1,338 1hr measurements, out of a possible 1,415, were recorded,
yielding a data capture efficiency of 95.0%. However, 89.0%
of these data were flagged as good data and the remaining
11% as invalid because of filter replacement, maintenance,
regular check-up, filter blowouts during sampling, power
failure, etc. To determine the degree of equivalence, 24-hr
integrated OC and EC concentrations (52 measurements
available) were regressed against 24-hr averages of the 1-hr
OC and EC measurements from the semi-continuous carbon
analyzer. In calculating the 24-hr averages, regression
analyses were made using only valid OC and EC data. Data
pairs were used in the regressions only if < 30% of the 24
1-hr measurements were missing. Regressions were made
for 43 data pairs during the study period.
Fig. 2 represents the relationship between 24-hr averages
of the 1-hr data and 24-hr integrated data for OC and EC
measurements. Sunset OC concentrations were ~27% lower
than 24-hr integrated OC concentrations [OCsunset (μg C/m3)
= (0.73 ± 0.04)OCfilter (μg C/m3) + (0.05 ± 0.20), R2 = 0.94].
Sunset EC data were also ~27% lower than 24-hr integrated
EC concentrations [ECSunset (μg C/m3) = (0.73 ± 0.03)ECfilter
(μg C/m3) + (0.12 ± 0.05), R2 = 0.90]. The difference in the
OC data could have been due to positive and negative artifacts
between the semi-continuous carbon analyzer (operated
with the denuder) and the PM2.5 speciation sampler without
the denuder (Park et al., 2005b; Polidori et al., 2006), to
the different face velocities of air, or to the different TOT
temperature programs (Schauer et al., 2003). The face
velocity of air passing through the filters was 20.1 and
111.1 cm/s for the speciation sampler and semi-continuous
carbon analyzer, respectively. Therefore, the OC data from
the PM2.5 speciation sampler were expected to be clearly
higher than the semi-continuous measurements. The large
discrepancy for the EC data was attributed to the high
detection limit of EC in the semi-continuous carbon analyzer
at the 1–2 hr time resolution (Lim and Turpin, 2002) and
differences in the temperature programs. A previous study
has indicated that EC measurements are sensitive to
operating conditions despite the use of the same analytical
method, i.e., TOT (Schauer et al., 2003). In this work, the
original OC and EC data were used to test the mass balance
closure of PM2.5 and to estimate the relative contribution of
carbonaceous particles to the PM2.5.
Ionic Species Measurements
Hourly concentrations of eight ionic species in PM2.5 were
measured using an ambient ion monitor (AIM, URG9000D,
URG Corporation). The AIM monitor drew air in through a
PM2.5 sharp-cut cyclone at a volumetric-flow controlled rate
of 3 L/min to remove the large particles from the air stream.
In order to minimize the loss of particles during sampling,
a Teflon-coated aluminum pipe was set up in the vertical
direction. The sample was then drawn through a liquid
diffusion denuder where the interfacing acid and alkaline
gases were removed. In order to achieve high collection
efficiencies, the particle-laden air stream then entered the
aerosol super-saturation steam chamber to enhance the
particle growth. Grown particles, i.e., droplets, were collected
every hour by an inertial particle separator and were then
injected into the ion chromatograph. Both particles and
5.0
18
Sunset carbon monitor vs filter-based OC
2
Linear fit: y = 0.73x + 0.05, R =0.94
16
24-hr average Sunset EC conc (g C/m 3)
24-hr average Sunset OC conc (g C/m3)
20
14
12
10
8
6
4
2
0
4.5
Sunset carbon monitor vs. filter-based EC
Linear fit: y = 0.73x + 0.12, R2=0.90
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0
2
4
6
8
10
12
14
16
18
20
0.0
0.5
24-hr average filter-based OC conc (g C/m3)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
36
40
3.6
4.0
40
27
AIM vs filter-based
24
Linear fit: y = 0.89x + 0.49, R2=0.90
36
SO42-
24-hr average AIM NO3- conc (g/m3)
24-hr average AIM SO42- conc (g/m3)
1.0
24-hr average filter-based EC conc (g C/m3)
30
21
18
15
12
9
6
3
-
AIM vs. filter-based NO3
32
Linear fit: y = 0.76x + 0.28, R2=0.95
28
24
20
16
12
8
4
0
0
3
6
9
12
15
18
21
24
27
0
30
0
4
24-hr average filter-based SO42- conc (g/m3)
8
12
24-hr
16
20
24
average filter-based NO3
28
32
3
conc (g/m )
4.0
20
18
3.6
AIM vs. filter-based NH4
16
+
24-hr average AIM K+ conc (g/m3)
24-hr average AIM NH 4 + conc (g/m 3)
961
2
Linear fit: y = 1.00x + 0.00, R =0.91
14
12
10
8
6
4
2
AIM vs. filter-based K+
Linear fit: y = 090x - 0.02, R2=0.97
3.2
2.8
2.4
2.0
1.6
1.2
0.8
0.4
0
0
2
4
6
8
10
12
14
16
18
20
24-hr average filter-based NH4+ conc (g/m3)
0.0
0.0
0.4
0.8
1.2
1.6
2.0
2.4
+
2.8
3.2
3
24-hr average filter-based K conc (g/m )
Fig. 2. Comparison between filter-based and semi-continuous measurements for PM2.5 major chemical components.
gases were collected and injected using syringe pumps,
PEEK valves, Teflon® tubing and PEEK tubing. A
detailed description of the monitor can be found elsewhere
(http://www.urgcorp.com).
The 24-hr averages of hourly Cl–, SO42–, NO3–, NH4+,
and K+ were regressed against 24-hr integrated ionic
species data measured with a PMS103 speciation sampler.
Comparison results between the two methods for SO42–,
NO3–, NH4+, and K+ concentrations are shown also in Fig. 2.
Very strong correlations were found between the two methods
with R2 of 0.90–0.97. AIM SO42– concentrations were about
10% lower than the filter-based SO42– measurements with
962
an R2 of 0.90. The SO42– from AIM has been reported to
exhibit a positive interference at SO2 concentration exceeding
30 ppbv and negative readings at SO42– levels exceeding 20
µg/m3 (Wu and Wang, 2007). Although not shown here,
however, no SO2 interference was observed in the present
study. The observed AIM SO42– data were corrected based
on the regression slope between the two methods to check
the mass balance closure of PM2.5. AIM NO3– concentrations
were strongly correlated with the filter-based NO3–
measurements with a slope of 0.76 and an R2 of 0.95,
indicating that the AIM NO3– concentrations were about
24% lower than those from the filter measurements. The
higher NO3– concentrations on the filter samples may have
been due to absorption of HNO3 on the filter, causing positive
bias. The sampling artifacts of nitrate and ammonium also
occurred due to evaporation of semi-volatile NH4NO3 from
particles collected on the filter (Koutrakis et al., 1992; Zhang
and McMurry, 1992), leading to a negative bias which can
be substantial (45–75%) at high temperature and at low
aerosol loading (Nie et al., 2010; Saarnio et al., 2010).
However, the negative artifact of NO3– particles in this study
may not have been significant because of low ambient
temperatures and very high aerosol loading. Low temperature
in winter favors the partition of ammonium nitrate in the
particulate phase. Hence, the sampling artifacts probably
were not a major problem in winter. AIM hourly NO3– data
were corrected using a regression relationship (slope and
intercept) between the 24-hr averages of AIM hourly data
and 24-hr integrated speciation data and used to investigate
the characteristics of ionic species observed during two
haze pollution episodes, which are discussed below.
Measurements of Elemental Constituents
Hourly concentrations of elemental constituents (K, Ca,
Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Ag, Cd, Sn, Sb,
Ba, Ag, Tl, and Pb) in PM2.5 were measured with an
ambient metals monitor (model: XactTM 620, Cooper
Environmental Services LLC, USA). The Xact 620
monitoring system uses reel-to-reel filter tape sampling and
nondestructive XRF (Ed- this acronym has already been
defined above) analysis to monitor ambient air. The air was
sampled at a flow rate of 16.7 L/min through a PM2.5 inlet
and drawn through a filter tape. The resulting PM2.5 deposit
was automatically advanced and analyzed by XRF for
selected metals as the next sample was being collected.
Sampling and analysis were performed continuously and
simultaneously, except during advancement of the tape
(~20 sec) and during daily automated quality assurance
checks. The measurement method of the metal species in
ambient air particles was based on EPA Method IO 3.3
(“Determination of metals in ambient PM using XRF). The
minimum detection limits of K, Ca, Ti, V, Cr, Mn, Fe, Co,
Ni, Cu, Zn, As, Se, Cd, Sn, Sb, Ba, Tl, and Pb for time
resolution of 1-hr were 0.81, 0.32, 0.18, 0.14, 0.11, 0.07,
0.08, 0.05, 0.04, 0.06, 0.04, 0.03, 0.03, 1.35, 2.54, 0.67,
0.40, 0.05, and 0.05 ng/m3, respectively. The 24-hr averages
of hourly elemental species measurements (K, Ca, Fe, Mn,
Cu, Zn, As, and Pb) were compared with those derived from
24-hr integrated filter sample (results from other elemental
species are not provided here because of insufficient paired
data). Agreement between elemental species concentrations
measured by the two methods showed correlation coefficients
(R2) greater than 0.80, with the exception of Fe (R2 = 0.70).
Slopes between the measured concentrations from the two
methods ranged from 0.74 (Zn) to 1.80 (Ca), with the majority
in the range of 0.83–1.15, suggesting high accuracy of
continuous XRF measurements for these species.
RESULTS AND DISCUSSION
Reconstructed PM2.5 Mass Balance Closure
Hourly-based chemical mass balance closure was assessed
by comparing BAM1020 PM2.5 mass concentration. The
reconstructed PM2.5 mass was computed by summing the
concentrations of organic matter (OM), EC, SO42–, NO3–,
NH4+, Cl–, crustal material, and trace metals. OM was
obtained by applying a multiplicative factor of 1.6 (i.e.,
urban aerosols) to OC to account for the oxygen and
hydrogen associated with OC (Turpin and Lim, 2001). We
estimated the concentration of crustal material by using the
following equation reported by Malm et al. (1996): Crustal
material (μg/m3) = 2.20[Al] + 2.49[Si] + 1.63[Ca] +
2.42[Fe] + 1.94[Ti]. The sum of the concentrations of all
the other elements is reported here as the trace metals
value. In this study, because Al and Si data were not
available, the contribution of these elements was not
included in the estimated concentrations of the crustal
material, thus resulting in an underestimation of the crustal
material in the reconstructed fine mass. Fig. 3 compares the
BAM and reconstructed PM2.5 concentrations. As shown,
the reconstructed PM2.5 was highly correlated with the
measured BAM PM2.5 concentrations with an R2 of 0.92.
However, the reconstructed PM2.5 was about 19% lower than
the measured PM2.5. As discussed before, the reconstructed
PM2.5 mass concentrations may have been underestimated
due to low contributions of crustal materials, OM, and EC
to PM2.5 mass and particle water content.
Descriptions of Haze Pollution Episodes
As shown in Fig. 3, it is apparent that PM2.5 mass was
highly variable on a time scale of 1 hr, whereas the use of
24-hr average largely reduces this variability. For example,
the hourly averaged PM2.5 mass concentration on February
4 reached 159.0 μg/m3, i.e., approximately 1.3 times
greater than the 24-hr average mass of 122.8 μg/m3. Over
the two-month study period, the average 1-hr PM2.5 mass
concentration was 44.4 ± 30.0 µg/m3 with a maximum of
159.0 μg/m3, and the 24-hr averaged BAM PM2.5 mass
ranged from 14.6–122.8 μg/m3. The 24-hr averaged Korean
PM2.5 NAAQS of 50.0 μg/m3 was exceeded on 20 days
during the study period. These daily PM2.5 mass concentration
ranged from 50.4 μg/m3 on March 20th to 122.8 μg/m3 on
February 4th. These pollution episodes were associated with
regional haze sulfate and nitrate events, biomass burning
emissions, and local traffic emissions under air stagnation and
low-level inversions conditions. Out of 20 days exceeding
the 24-hr averaged Korean PM2.5 NAAQS, two distinct
PM2.5 haze episodes are discussed herein: February 1–8
963
200
Reconstructed PM2.5 conc (g/m3)
180
Reconstructed versus BAM PM2.5
Linear fit: y = 0.81x - 2.3, R2=0.91
160
140
120
100
80
60
40
20
0
0
200
20
40
80
100
120
140
160
180
200
3
MetOne BAM PM2.5 conc (g/m )
180
PM2.5 mass concentration (g/m3)
60
MetOne BAM PM2.5
160
Reconstructed PM2.5
140
120
100
80
60
40
20
0
02/01
02/08
02/15
02/22
03/01
03/08
03/15
03/22
03/29
Measurement period (2011)
Fig. 3. Comparison of hourly measured and reconstructed PM2.5 mass concentrations.
(episode I) and March 11–12 (episode II), 2011. Two haze
episodes were classified based on CO/NOx ratios and K+
concentrations. Haze episode I which developed between
February 01 and 08, was strongly associated with regional
haze pollution along with wildfire smoke emissions over
southern China. This episode was further characterized by
elevated afternoon SO42– and O3 concentrations (O3 = 46–
60 ppb on February 01–05). Episode II was from 14:00 March
11–17:00 March 12 during which hourly PM2.5 increased
from a background level of 23 μg/m3 to a maximum of 116
μg/m3. This episode was characterized by locally produced
pollution with low CO/NOx ratios and broad peaks for OC,
EC, and NO3– concentrations. During episode I, local wind
speeds were generally weak to moderate averaging 1.1 m/s
(range: 0.1 to 3.7 m/s). Ambient temperatures ranged from
–7.4 to 9.9°C (mean: 2.7°C). Relative humidity (RH) had a
range sufficient to cause hygroscopic growth (26.1 to 96.6%;
mean = 67.1 ± 19.6%), and ranged from > 75–80% in the
evening/morning hours. During episode II, temperatures
were in the range of –0.1 to 16.2°C (mean = 7.6°C). Winds
ranged from 0.1 to 6.3 m/s (mean = 2.2 m/s), and were weak
during the night (0.1–1.6 m/s). Relative humidity ranged
from 19.8–92.9% with a mean of 56.3%. Daytime RH was
low (< 40%), but was > 75% during the night, especially on
the evening/morning.
To examine the differences in the chemical composition
of PM2.5 observed during the two episodes, we analyzed
MODIS images of two types, i.e., Aqua & Terra MODIS
image (http://earthdata.nasa.gov/firms) and MODIS true color
images (http://lance-modis.eosdis.nasa.gov/cgi-bin/imagery/
realtime.cgi), and transport pathways of the air mass, as
illustrated in Fig. 4. The MODIS images show that during
episode I, numerous fire counts were observed and a thick
haze layer lingered over northern and southern China, and
the Korean peninsula. While during episode II, a clear sky
was observed over northern China and the Korean peninsula.
The transport pathways of the air mass prior to arriving at
the site were determined using the Hybrid Single-Particle
Lagrangian Integrated Trajectories (HYSPLIT 4.5) model
(Draxler and Ralph, 2012). Isentropic five-day backward
trajectories were computed for every 12-hr (0000 UTC and
1200 UTC time) interval for each day of the two episodes,
at three altitudes (500, 1000, and 1500 m) above ground
level (AGL). Two of the trajectory results are also presented
in Fig. 4, one for episode I (1200 UTC February 04) and
the other for episode II (1200 UTC March 11). For episode
I, air masses originating from the polluted regions of
northern China (at 500 and 1000 m AGL) or regions of
Southern China (at 1500 m AGL), passed over the Yellow
Sea and reached the site. Air mass trajectories for episode I
964
Fig. 4. MODIS images and back trajectories of air mass arriving at the site for the two haze episodes. Symbols in
trajectories; : 500 m AGL, : 1000 m AGL, : 1500 m AGL.
indicated a combination of transport directions of northern
and southern China and the air mass was classified as a
mixture of wild-fire plumes, regionally transported and local
pollution. Similar to the transport pathway for episode I,
the air mass for episode II originated in the regions of
northeastern China and arrived at the site, but without any
observable influence of wild-fire plumes.
In this study, ratio of CO to NOx (= NO + NO2) was also
used to examine the evolution of the aerosol chemical
composition with respect to the age of the air masses as a
proxy for proximity to major pollution sources and
atmospheric processing (Mogan et al., 2010; Freney et al.,
2011). Morgan et al. (2010) classified CO/NOx ratios of 5–
15, 10–50, and > 50 as urban, near-source and aged regional
conditions, respectively. Temporal variation in the CO/NOx
ratio is shown in Fig. 5. The mixing ratios of CO and NOx
for episodes I and II were 896 (475–1319 ppb) and 32 (6–
139 ppb), and 482 (247–688 ppb) and 40 (7–97 ppb),
respectively, with greater impact of CO during episode I.
The CO/NOx ratios ranged from 9 to 159 with an average
value of 41 ± 26 in episode I and from 6 to 44 with an
average of 18 ± 11 in episode II, respectively, confirming
that the air masses arriving at the site during episode I were
associated with more aged regional components than those
during episode II. Fig. 6 shows the mass factions of EC,
OM, SO42– and NO3– to the PM2.5 mass against the CO/NOx
ratio for the two episodes. As shown in Fig. 6, the significant
amounts of EC, OM, SO42– and NO3– observed during
episode I were associated with high CO/NOx ratio (> 50).
Higher EC and OM contributions to the PM2.5 favor low
CO/NOx ratios arising from presumably local emission
sources. However, their absolute concentrations were not
greatly affected by the CO/NOx ratio (not shown here). That
is, it is found that high concentrations of EC, OC, SO42–
and NO3– were also related to either the boundary layer air
or regional aged air masses.
Table 1 summarizes the PM2.5 concentrations and the
chemical components for episodes I and II. As shown in
Table 1, as much as 50% of the PM2.5 mass is attributed to
secondary SO42–, NO3–, and NH4+ particles, with NO3– being
most significant fraction of the PM2.5 mass for the two
episodes. The large difference in chemical compositions
between the two haze episodes is due to the NO3– and K+
concentrations. Fig. 7 shows temporal profiles of PM2.5,
OC, EC, K+, and SO42– concentrations for the two pollution
episodes. Fig. 8 shows correlation relationships between
OC and K+, and between K+ and K+/OC ratio for the period
of February 04–05. Temporal profiles of NO3–, NO2 and O3
concentrations, ambient temperature, and RH are shown in
Fig. 9.
Characteristics of Carbonaceous Particles
OC and EC concentrations were 6.9 (2.6–14.0 μg C/m3)
and 2.2 (0.5–4.5 μg C/m3) for episode I and 6.3 (2.0–10.3 μg
C/m3) and 2.0 (0.7–5.1 μg C/m3) for episode II, respectively.
OM and EC contributed 13.1 ± 2.7% and 3.3 ± 0.7% to the
measured PM2.5 for episode I, and 14.2 ± 3.8% and 3.1 ± 0.8%
for episode II, respectively. The EC and OC concentrations
were strongly correlated with the following regression
relationships; for episode I: OC = 2.76 EC + 0.95, R2 = 0.71,
for episode II: OC = 3.01 EC + 0.46, R2 = 0.83, suggesting
common source emissions (i.e., vehicles) during episodes I
and II. The OC/EC ratio was 3.2 ± 0.6 (1.8–5.4) and 3.3 ±
0.6 (2.1–4.8) for episodes I and II, respectively. Diurnal OC
and EC peak concentrations over the two episodes occurred
mainly during morning and evening rush-hour periods,
indicating the significant impact of traffic. As the boundary
layer rose throughout the morning and early afternoon, the
OC and EC concentrations tended to decrease. However,
close inspection of Fig. 7 revealed high background OC
and EC levels during episode I with values of 3.0–4.0 and
~2.0 μg C/m3, respectively, probably due to long-range
transport of polluted air masses (shown in Fig. 4) and air
stagnation conditions. The OC and EC background levels
during episode II were approximately 1.0 and 0.5 μg C/m3,
respectively.
According to MODIS satellite images and air mass back
trajectory analysis (Fig. 4), huge forest wildfires occurred
in the southern regions of China on February 01–02, and
the air mass passing over these wildfire regions reached the
200
200
180
180
CO/NOx ratio
160
160
February 1~8
120
100
80
120
100
80
60
60
40
40
20
20
0
02/01
02/02
02/03
March 11~12
CO/NOx ratio
140
CO/NOx ratio (-)
CO/NOx ratio (-)
140
965
02/04
02/05
02/06
02/07
02/08
02/09
0
03/11 00
03/11 08
03/11 16
03/12 00
03/12 08
Fig. 5. Variations of CO/NOx ratio for the two haze episodes.
03/12 16
03/13 00
966
10
10
9
9
EC/PM2.5 vs. CO/NOx ratio
8
7
EC/PM2.5 (%)
EC/PM2.5 (%)
7
6
5
4
2
1
1
0
20
40
60
80
100
120
140
160
180
200
0
CO/NOx ratio
27
20
30
40
50
60
70
80
90
100
90
100
CO/NOx ratio
36
OM/PM2.5 vs. CO/NOx ratio
24
February 01~08
21
10
40
Organic mass (OM)/PM2.5 (%)
Organic mass (OM)/PM2.5 (%)
4
3
30
18
15
12
9
6
3
OM/PM2.5 vs. CO/NOx ratio
32
28
March 11~12
24
20
16
12
8
4
0
0
0
20
40
60
80
100
120
140
160
180
200
0
CO/NOx ratio
70
63
10
20
30
40
50
60
70
80
CO/NOx ratio
70
63
2-
SO4 /PM2.5 vs. CO/NOx ratio
56
2-
56
February 01~08
SO4 /PM2.5 vs. CO/NOx ratio
49
SO4 /PM2.5 (%)
49
42
35
2-
2-
5
2
0
SO4 /PM2.5 (%)
March 11~12
6
3
0
28
March 11~12
42
35
28
21
21
14
14
7
7
0
0
0
20
40
60
80
100
120
140
160
180
0
200
CO/NOx ratio
70
10
20
30
40
50
60
70
80
90
100
CO/NOx ratio
70
63
63
-
NO3 /PM2.5 vs. CO/NOx ratio
56
NO3-/PM2.5 vs. CO/NOx ratio
56
February 01~08
49
NO3 /PM2.5 (%)
49
42
March 11~12
42
35
-
35
-
NO3 /PM2.5 (%)
EC/PM2.5 vs. CO/NOx ratio
8
February 01~08
28
28
21
21
14
14
7
7
0
0
0
20
40
60
80
100
120
140
160
180
200
CO/NOx ratio
0
10
20
30
40
50
60
70
80
90
100
CO/NOx ratio
Fig. 6. Fractions of EC, OM, SO42–, and NO3– to the PM2.5 against CO/NOx ratio for the two episodes.
sampling site on February 04. Therefore, relationships among
OC, K+, and the K+/OC ratio from February 04–05, when
the K+ concentration significantly increased, were examined
to evaluate the effect of the forest fire plumes on carbonaceous
particles observed at the site (Fig. 8). It has been known
that biomass burning emissions contribute significantly to
atmospheric OC, EC, and K+ concentrations (Andreae, 1983;
Echalar et al., 1995; Andreae and Merlet, 2001; Reid et al.,
2005; Park et al., 2006b; Ram and Sarin, 2010, 2011).
Vehicles and fossil fuel emissions are the dominant sources
967
Table 1. Concentration summary of chemical species in PM2.5 for the two episodes.
Parameter
Unit
Episode I1)
Episode II2)
Mean
Range
Mean
Range
PM2.5
μg/m3
88
27–159
78
23–116
EC
μg C/m3
2.2
0.5–4.5
2.0
0.7–5.1
OC
μg C/m3
6.9
2.6–14.0
6.3
2.0–10.3
SO42–
μg/m3
18.5
3.1–48.4
16.5
8.4–27.7
–
NO3
μg/m3
22.3
6.2–53.1
32.3
5.7–49.5
Cl–
μg/m3
1.7
0.3–4.8
1.5
0.3–3.0
Na+
μg/m3
0.9
0.6–1.1
0.8
0.7–0.9
NH4+
μg/m3
8.9
4.0–15.1
6.3
2.8–8.2
K+
μg/m3
1.8
0.3–5.6
0.5
0.2–0.7
OC/EC
3.2
1.8–5.4
3.3
2.1–4.8
K+/OC
0.26
0.07–0.63
0.10
0.06–0.20
SOR3)
0.55
0.25–0.76
0.50
0.37–0.71
K
μg/m3
2.0
0.3–5.9
0.9
0.4–1.3
Ca
μg/m3
0.27
0.03–0.74
0.26
0.14–0.41
Fe
μg/m3
0.46
0.09–0.92
0.54
0.24–1.31
Ti
ng/m3
17
2–40
15
8–28
V
ng/m3
8
3–29
8
2–13
Cr
ng/m3
9
2–16
13
4–28
Mn
ng/m3
39
10–87
56
21–130
Ni
ng/m3
6
3–12
8
2–14
Cu
ng/m3
23
6–54
28
11–64
Zn
μg/m3
0.13
0.05–0.32
0.30
0.10–0.60
As
ng/m3
17
3–34
16
6–28
Se
ng/m3
8
2–13
6
2–8
Cd
ng/m3
6
1–46
23
2–57
Sb
ng/m3
73
23–111
86
46–121
Ba
ng/m3
97
11–325
36
17–61
Pb
ng/m3
130
28–308
107
41–151
Note) 1)Episode I covers the period of February 01–08, 2011; 2)Episode II covers the period of March 11–12, 2011; 3)SOR
indicates SO42–/(SO2 + SO42–).
of OC and EC, but have little impact on K+ concentrations.
In this study, OC was strongly correlated with EC (R2 =
0.82) during the 2-days period, but the OC/EC ratio (range:
2.8–5.3, mean: 3.4) did not increase, which was significantly
lower than those reported for Savannah burning, forest fires,
and agricultural waste burning emissions (Andreae, 1983;
Park et al., 2005b, 2006b; Reid et al., 2005; Ram and Sarin,
2011; Ram et al., 2012a, 2012b; Zhang et al., 2012b), or
slightly lower than that (5.7) from emissions of forest fires
in the Indochina (Chuang et al., 2012). Other studies have
shown even when biomass materials are burned, a low
OC/EC ratio is possible because it depends greatly on the
type of materials burned and combustion conditions (Turn
et al., 1997; Saarnio et al., 2010; Bae et al., 2012; Zhang et
al., 2012a). For example, Turn et al. (1997) indicated that
for wood fuels such as Walnut prunings, Almond prunings,
and Ponderosa pine slash, the OC/EC ratio in PM2.5 was in
the range of 1.7–2.1. It was also found that when pine needles,
cherry trees, and acacia trees are combusted, their OC/EC
ratios in PM2.5 were 0.83, 0.20, and 0.21, respectively (Bae
et al., 2012). As shown in Fig. 7, the concentration of K+,
an indicator of biomass burning emissions (Simoneit et al.,
2004; Park et al., 2006b; Zhou et al., 2009; Saarnio et al.,
2010; Chuang et al., 2012), increased from 1.5 µg/m3 at
00:00 to 5.6 µg/m3 at 12:00 on February 04, and remained
high for 2 days. K+ concentration accounted on average
2.9% of the PM2.5 (1.2–4.8%), which was comparable to
that (2.5%) reported by a recent study of Chuang et al.
(2012). In addition to the biomass burning, the abundance
of K in urban air is attributed to soil dust. To assess the
origin of K in this study, the relationship between total K
and K+ was examined. Previous studies indicated quite low
K+/K ratio of 0.1–0.2 in geological material (Gertler et al.,
1995) and 0.01–0.02 in Mongolia soil dust (Kim et al.,
2010). Water-soluble K+ and total K were strongly correlated
(slope = 0.82; R2 = 0.96) for episode I (not shown here),
suggesting these two constituents were clearly derived from a
similar emission source, most likely forest fire emissions.
This was evidenced by the MODIS satellite image, in which
numerous fire counts were observed during episode I, and
transport pathway of air masses. Scatter plots between OC –
K+ and K+ – K+/OC for high K+ episode (February 04–05)
were strongly correlated (R2 of 0.77 and 0.77, respectively)
(Fig. 8), suggesting an emission source from biomass burning
(Zhang et al., 2012b). The K+/OC ratio for the 2-days period
showed a range of 0.15–0.63 with a mean of 0.38, which
was significantly higher than those reported for biomass
burning aerosols. Previous studies have reported K+/OC
968
200
200
180
Februry 01~08
PM2.5 concentration (g/m3)
PM2.5 concentration (g/m3)
140
120
100
80
60
120
100
80
60
40
20
20
02/02
02/03
02/04
02/05
02/06
02/07
02/08
0
03/11 00
02/09
03/11 08
20
9
18
8
16
7
OC concentration (g C/m3)
14
10
EC concentration (g C/m3)
16
OC concentration (g C/m3)
February 01~08
OC
EC
03/12 00
03/12 08
03/12 16
03/13 00
10
March 11~12
OC
EC
9
8
14
7
12
6
10
5
8
4
6
3
12
6
10
5
8
4
6
3
4
2
4
2
2
1
2
1
0
02/01
02/02
02/03
02/04
02/05
02/06
02/07
02/08
0
03/11 00
0
02/09
03/11 08
K (right "y" axis)
03/12 00
03/12 08
03/12 16
0
03/13 00
60
9
54
8
48
2.0
SO42+
1.8
March 11~12
K (right "y" axis)
1.6
7
36
6
30
5
24
4
18
3
2-
42
K+ concentration (g/m3)
3
48
February 01~08
10
SO4 concentration (g/m )
54
2SO4
+
03/11 16
60
SO42- concentration (g/m3)
03/11 16
20
18
March 11~12
140
40
0
02/01
PM2.5 time-series plot
160
EC concentration (g C/m3)
PM2.5 time-series plot
160
42
1.4
36
1.2
30
1.0
24
0.8
18
0.6
0.4
0.2
12
2
12
6
1
6
0
02/01
02/02
02/03
02/04
02/05
02/06
02/07
02/08
0
03/11 00
0
02/09
03/11 08
03/11 16
03/12 00
03/12 08
03/12 16
K+ concentration (g/m3)
180
0.0
03/13 00
+
2–
Fig. 7. Temporal trends of PM2.5, OC, EC, K and SO4 concentrations for the two haze episodes.
ratios from 0.08–0.10 for savanna burning (Echalar et al.,
1995), 0.04–0.13 for agricultural waste burning (Andreae and
Merlet, 2001), 0.19–0.21 for wheat straw burning/fertilizer
(Duan et al., 2004), 0.02-0.14 for biomass burning (Ram
and Sarin, 2010, 2011; Ram et al., 2012a), and 0.04–0.24
for wheat and/or stalk burnings (Zhang et al., 2012b). The
high K+/OC ratio found in this study may be attributed to
traffic and fossil fuels emissions, as well as biomass burning
emissions. Accordingly, even though low OC/EC and high
K+/OC ratios were observed for the February 04–05 episode,
relatively high K+ concentrations and the significant linear
relationships between EC – OC (R2 = 0.82), OC – K+ (R2 =
0.77), and K+ – K+/OC (R2 = 0.77) suggest contribution of
carbonaceous aerosols from long-range transport of biomass
burning plumes occurring in the southern regions of China
during episode I, as well as the traffic emissions.
Sulfate Concentration
The concentrations of SO42– observed during episodes I
and II were 18.5 ± 8.7 μg/m3 (3.1–48.4 μg/m3) and 16.5 ±
1.0
9
0.9
+
8
K vs. OC
2
Linear fit: y = 0.46x- 0.14, R =0.77
7
0.7
February 04~05
6
February 04~05
K+/OC vs. K+
Linear fit: y = 0.09x + 0.07, R2=0.77
0.8
K+/OC ratio (-)
Water-soluble K+ concentration (g/m3)
10
969
5
4
0.6
0.5
0.4
3
0.3
2
0.2
1
0.1
0
0.0
0
2
4
6
8
10
12
14
16
18
20
0.0
0.7
1.4
3
2.1
2.8
3.5
+
OC concentration (g C/m )
4.2
4.9
5.6
6.3
7.0
3
Water-soluble K concentration (g/m )
+
Fig. 8. Regression relationships between K and OC (left) and between K /OC and K+ (right) for February 04–05.
5.7 μg/m3 (8.4–27.7 μg/m3), respectively, contributing 20.7
% (6.4–53.4%) and 25.2 % (9.4–53.5%) to the PM2.5. The
SO42– aerosols in fine particles are formed through gas phase
reactions, in-cloud processes and aerosol droplet processes
(Guo et al., 2010). The sulfur oxidation ratio [SOR = SO42–
/(SO2 + SO42–)] has been used as a good indicator to assess
the atmospheric transformation of SO2 to SO42– (Kaneyasu
et al., 1995; Sahu et al., 2009). SOR in winter is expected
to be low due to the low temperature and low photochemical
flux, lowering the possibility of gas-phase oxidation of SO2
to SO42–. Non-sea salt SO42– was calculated by subtracting the
sea-salt SO42– from the measured SO42–. SO42– in seawater
was estimated from the SO42–/Na+ ratio in seawater.
Previously reported 24-hr measurements made at a Gwangju
urban site of Korea, which is 10 km away from our site,
indicated an SOR of 0.51 in summer and 0.14 in winter
(Park et al., 2010, 2012). The calculated SOR was 0.55 ±
0.11 (0.25–0.76) and 0.50 ± 0.09 (0.37–0.71) for episodes I
and II, respectively, which were substantially high even in
winter. Urban SO42– in PM2.5 is composed of urban and
regional components. The regional SO42– contribution can be
evaluated by its diurnal behavior and background levels. For
episode I, the temporal profile of SO42– concentration (Fig. 7)
exhibited a continuous increase in the SO42– background level
from 3.0 μg/m3 on February 1 to 20.0 μg/m3 on February 5
suggest a greater regional contribution, and are probably
due to long range transport and regional sulfate formation,
or air stagnation. An obvious diurnal pattern was observed
with maximum SO42– in the afternoon and the SO42–
transient increased gradually from 26.6 μg/m3 on February
1 to 48.4 μg/m3 on February 4. The maximum hourly O3
and SO2 concentrations on February 1, 2, 3, 4, and 5 were
46 and 13 ppb, 50 and 16 ppb, 58 and 14 ppb, 52 and 19
ppb, and 48 and 12 ppb, respectively, which were relatively
high even in winter. The SO42– concentration was highly
correlated with SO2 with an R2 of 0.64 (data not shown
here). This implies that local SO2 emissions were also an
important source of SO42–formation, and that slow gasphase oxidation of SO2 probably played a critical role in
+
the SO42– production. During episode I, the temporal variation
of sulfate concentrations and air mass trajectories suggested
that some of the sulfate excess over background could be
attributed to sources in the local region. As a result, the high
SOR and highly elevated SO42– concentration observed during
episode I suggest that in addition to local production, SO42–
was probably caused by long-range secondary atmospheric
processing.
For episode II, SO42– concentrations continued to increase
from 14:00 on March 11 until 19:00 to a maximum of 23.6
μg/m3, accounting for 33.7% of the PM2.5, after which the
SO42– dropped to 9.3 μg/m3 at 07:00 on March 12. At this
time it increased again and peaked at 13:00 hr with a
maximum of 27.7 μg/m3, constituting 40.2% of the PM2.5,
then decreased gradually. Hourly O3 and SO2 concentrations
were in the ranges of 47–50 and 5–11 ppb between 14:00
and 19:00 on March 11, respectively. They subsequently
decreased to 17 and 2 ppb at 07:00 hr of March 12 and
again increased to 66 and 13 ppb at 13:00 hr, respectively.
Hourly temperature and RH were 8.2–13.0°C and 20–43%
between 14:00 and 19:00 on March 11, and 4.6–15.0°C and
28–75% between 07:00 and 13:00 on March 12,
respectively. The SO42– peaks occurring in the afternoon
hours on both March 11 and 12 coincided with maximum
ozone concentrations, thus, both SO42– peaks were driven
by photochemistry. Throughout episode II, the SO42–
concentration was strongly correlated with SO2 (R2 = 0.85)
(not shown here). Local wind directions in the afternoon
hours were between 230 and 250° (not shown here) where
industrial sources are located. This suggests a number of
industrial sources located in southwest direction from the
site may have contributed to the observed SO42– concentration.
Consequently, the pattern of these SO42– excursions in
relation to wind direction and the strong correlation between
SO2 and SO42– suggests local SO2 emissions were likely an
important source of SO42– at the site during episode II.
Nitrate Concentration
Concentrations of NO3– particles during episodes I and II
970
were 22.3 ± 9.7 μg/m3 (6.2–53.1) and 32.3 ± 13.3 μg/m3 (5.7–
49.5), respectively, accounting for 24.2% (9.5–40.4) and
40.9% (19.0–66.3%) of the PM2.5 mass. Atmospheric nitrate
particles are mainly formed through gas-phase oxidation
process of NO2 and OH during daytime and heterogeneous
reactions involving nitrate radical and N2O5 during nighttime
(Smith et al., 1995; Wittig et al., 2004; Park et al., 2005c).
The nitrate ambient concentration is also driven by gas-toparticle partitioning of ammonium nitrate and boundary
layer dynamics (Seinfeld and Pandis, 2006). For episode I,
the NO3– concentration of 7.3 μg/m3 at 10:00 on February
01 was gradually increased to reach a maximum of 53.1
μg/m3 at 12:00 on February 04, accounting for 34.1% of
the PM2.5 mass. As shown in Fig. 9, during this period, the
NO3– pollution event exceeding 20 μg/m3 lasted about 5
days and relatively high NO3– background levels were also
maintained, which could be attributed to the thick haze
layers lingering over southern and northern China and the
Korean peninsula, and to transport pathways of air masses
arriving at the site (Fig. 4). Some studies have indicated
that a strong diurnal pattern in PM2.5 nitrate concentration
was exhibited with a maximum in the morning and minimum
in the afternoon (Wittig et al., 2004; Park et al., 2005c).
Low nitrate concentrations during the daytime are typically
because the nitrate in PM2.5 is mostly in the gas phase as
nitric acid due to the higher temperature and lower RH
compared to the morning and evening. However, as shown
in Fig. 9, morning peaks and afternoon minimums in the
nitrate concentrations were not readily observed at the site.
Instead, the daytime nitrate transients with very high levels
over background were often found around 12:00–13:00.
Temperatures, RH, and wind speeds in the afternoon were
4.0–10.0°C, 26–59%, 1.5–2.5 m/s, respectively. The nitrate
peaks were more enhanced in the daytime than in the
nighttime. The nitrate daytime peak concentrations on
February 2, 3, 4, 5 and 6 were 42.0, 40.8, 53.1, 38.0, and
21.4 μg/m3, accounting for 40.4, 40.0, 34.1, 27.0, and
24.3% of the PM2.5, respectively. Periods showing these
peaks were reasonably consistent with those showing high
O3 concentrations. Concentrations of NO2 and O3 at these
periods were 34 and 41, 15 and 47, 21 and 48, 11 and 48,
and 11 and 60 ppb, respectively. These O3 levels were
comparable to the highest O3 concentrations on these days,
which were observed around 15:00–17:00, with levels of
50, 58, 54, 48, and 60 ppb, respectively. During the
daytime excursion periods, the concentrations of NOx and
carbonaceous species (OC and EC) were reduced to their
background levels on the corresponding day. Typically the
afternoon peak concentrations of fine nitrate particles have
been attributed to the photochemical formation of nitric
acid due to the increased ozone concentration, especially in
summertime (Seinfeld and Pandis, 2006). However, the
present situation, in which ozone levels of 40-60 ppb were
observed in winter, facilitated the formation of nitric acid
through gas-phase oxidation of NO2. These daytime nitrate
transients during episode I were probably associated with
transports of severe haze lingering over the upwind
locations of China (Fig. 4) to the sampling site, as well as
photochemical secondary aerosol production of nitrate and
low temperatures of 4.0–10°C. This hypothesis is further
supported by the relatively high CO/NOx ratio, as illustrated in
Fig. 5. A closer look at the CO/NOx ratio for episode I shows
that the ratios at the times of the elevated NO3– levels were
46–80, 44–78, 42–91, 81–93, and 72–91, suggesting that
the air masses arriving at the site were in regional aged
conditions. Evening NO3– peaks were observed four times,
at 20:00 on each of 02, 04, 05, and 07 February, with levels
of 26.9, 38.3, 22.1, and 24.9 μg/m3, accounting for 26.1,
24.1, 25.1, and 26.3% of the PM2.5, respectively. NO2
concentrations at the times of these nighttime peaks were
26, 25, 35, and 43.3 ppb, respectively. The increases in
NO3– in the evening were probably related to the lower
mass concentrations of sulfate at night, which allowed excess
NH3 to react with HNO3 (Seinfeld and Pandis, 2006).
These evening peaks coincided with the enhanced NO2
concentrations, but the ambient concentrations of the nitrate
were not necessarily proportional to the concentrations of
precursor emissions because the rates at which they form
and their gas-to-particle partitioning are controlled by
meteorological parameters other than the concentration of
the precursor gas. A similar result was found elsewhere (Park
et al., 2005c). After its evening peaks, nitrate decreased for
some time, but was maintained at high levels due to
continuous haze phenomena and poor atmospheric dispersion
conditions at night. Because the RH during the nighttime
(70–93%) was high enough to form aqueous nitric acid, the
nitrate observed at night may have accumulated in wet aerosol
particles. Thus, the significantly elevated concentrations of
NO3– particles observed during the nighttime may have
been related to low ambient temperatures, high RH, and the
regional aged air masses. In summary, the high NO3– values
during episode I probably reflect long-range secondary
atmospheric processing of nitrate particles, in addition to
local emission sources.
Episode II was characterized by a rapid increase of nitrate
starting at 14:00 on March 11 and persisting until 12:00 on
March 12, with a small drop in concentration between
01:00 and 07:00 on March 12. As shown in Figs. 7 and 9,
OC, EC, and NO3– concentrations peaked in the morning of
March 11 prior to the onset of episode II, and were probably
associated with morning rush hour traffic patterns. The
enhancement of morning NO3– concentration, to a maximum
of 18.4 μg/m3 at 09:00, was attributed to high RH and low
temperatures. RH and temperatures between 02:00 and
09:00 on March 11 were 84–93% and 0.0–1.0°C, respectively.
After 09:00, the NO3– concentration decreased to 4.6 μg/m3 at
13:00, started to increase again until 20:00–21:00, with a
maximum of 49.5 μg/m3, accounting for 46.3% of the PM2.5,
and then decreased to about 36.0 μg/m3, which was still
high. From 14:00 to 21:00 on March 11, as shown in Fig. 7,
the temporal behaviors of OC, EC, and SO42– concentrations
were similar to that of NO3–, probably due to local emissions.
Moreover, NO2 increased from 12 to 57 ppb, and O3 was
maintained at 40–50 ppb until 20:00 hr, but then was sharply
reduced to 10–20 ppb (Fig. 9). Hourly ambient temperature
and RH were 7.8–13.0°C and 24–48%, which is below the
deliquescence RH (DRH) of NH4NO3 (61% @25°C),
respectively. Hourly wind speed was 2.2–4.6 m/s. Based on
60
60
NO3
54
-
February 01~08
48
NO3-
3
42
36
42
36
30
-
30
24
2-
24
18
18
12
12
6
6
02/02
02/03
02/04
02/05
02/06
02/07
02/08
0
03/11 00
02/09
03/11 08
100
O3
February 01~08
NO2 and O3 concentrations (ppb)
NO2
80
70
60
50
40
30
20
03/12 16
03/13 00
90
NO2
80
O3
March 11~12
60
50
40
30
20
10
02/02
02/03
02/04
02/05
02/06
02/07
02/08
0
03/11 00
02/09
03/11 08
February 01~08
120
21
110
18
100
15
Relative humidity (%)
Ambient temperature (deg C)
Ambient temperature
Relative humidity
03/11 16
03/12 00
03/12 08
03/12 16
03/13 00
21
15
03/12 08
70
10
18
03/12 00
100
90
0
02/01
03/11 16
120
Ambient temperature
Relative humidity
110
March 11~12
100
12
90
9
80
6
70
3
60
0
50
-3
40
12
90
9
80
6
70
3
60
0
50
-3
40
30
-6
30
-6
-9
02/01
02/02
02/03
02/04
02/05
02/06
02/07
02/08
20
02/09
-9
03/11 00
03/11 08
03/11 16
03/12 00
03/12 08
03/12 16
Relative humidity (%)
0
02/01
NO2 and O3 concentrations (ppb)
March 11~12
48
SO4 and NO3 conc (g/m )
SO42- and NO3- conc (g/m3)
54
Ambient temperature (deg C)
971
20
03/13 00
Fig. 9. Temporal profiles of NO3– concentrations along with NO2, O3, temperature, and relative humidity for the two haze
episodes.
the concentrations of precursor gases and meteorological
conditions, the continuous increase of NO3– concentration
between 14:00 and 20:00 on March 11 was probably due to
gas-phase production of HNO3 via the reaction of NO2 and
OH at low RH (< 50%), with the majority of the HNO3
residing in the particle phase. In addition to the photochemical
production of HNO3, air masses transporting from regions
of northern China to the site could have partially contributed
to the enhanced NO3– concentrations in the afternoon (see
Fig. 4). As shown in Fig. 9, the nitrate concentrations
remained high between 22:00 March 11 and 07:00 March
12, ranging from 36.2 to 46.9 μg/m3. This NO3– excursion
was associated with low NO/high NO2, and high RH,
following high afternoon O3 levels of about 50 ppb,
presumably due to nighttime radical chemistry. During
these periods, the wind speed was 0.1–2.2 m/s with a mean
of 1.0 m/s, and RH and the temperature were 72–79% and
3.4–8.4°C, respectively. The mixing ratios of NO and NO2
at 22:00 were 29 and 73 ppb, respectively. Heterogeneous
reactions involving nitrate radical and N2O5, which are
972
formed during nighttime (Smith et al., 1995), are a potential
source of the nighttime nitrate. Nitric acid formed through
the heterogeneous hydrolysis of N2O5 at night can retain its
molecular form in the gas phase, or be transferred to the
particulate phase. The NO3– observed between 22:00 and
07:00 may have existed in the aqueous phase because of the
RH of 72–79%. The consequently elevated NO3–, which
induced the high PM2.5 levels during episode II, was attributed
to gas-phase photochemical reactions during the daytime and
heterogeneous hydrolysis during the nighttime.
Possible Sources of Crustal and Trace Metals in Ambient
Particles
One of the major anthropogenic sources in urban air is
vehicle emissions. Local sources have a greater effect on
EC concentrations. Variations in local diesel engines or
combustion sources could affect EC at a site. Like EC
concentrations, ambient concentrations of metals were
dominated by local sources. Road dust is generally resuspended by the turbulent passage of motor vehicles over
local roads. The road dust profile contains crustal elements
(Al, Si, Ca, Fe), along with metal constituents such as Mo,
Sb, Cu, Zn, Ba, Mn, Cr, and Pb, reflecting enrichment from
other sources and urban pollutants (Schauer et al., 2006).
These metals are known to be emitted from road traffic
abrasion processes or are important constituents of road dust
(Harrison et al., 2003; Schauer et al., 2006; Bukowiecki et al.,
2009; Amato et al., 2010). Brake wear emissions have been
reported to contain significant amounts of Ba, Zn, Cu, and
Sb, as well as Fe and other crustal elements (Garg et al.,
2000; Sternbeck et al., 2002; Thorpe and Harrison, 2008).
Previous studies indicated also that Sb in urban environments
is associated with traffic emissions, and more specifically with
emissions emerging from brake wear (Weckwerth, 2001;
Furuta et al., 2005; Gomez et al., 2005; Johansson et al.,
2009).
As shown in Table 1, the concentration differences of Sb
and the other brake wear-related elements (Fe, Cr, Mn, Cu,
and Pb) between episodes I and II were not significant. The
concentration of Sb was 73 ± 16 ng/m3 (22.5–111.0) and
86 ± 19 ng/m3 (46–121) for episodes I and II, respectively.
The mean concentration of Fe, one of the most abundant
measured elements, was similar during the two episodes,
ranging from 0.09 to 0.92 µg/m3 with a mean of 0.46 µg/m3
and from 0.24 to 1.31 µg/m3 with a mean of 0.54 µg/m3,
respectively. However, Ba, another marker of brake wear,
was about three times higher in episode I (97 ng/m3) than
in episode II (36 ng/m3), while Zn was two times higher in
episode II (0.30 µg/m3) than in episode I (0.13 µg/m3). This
suggests that in addition to the road dust, there were relevant
local emissions of these elements at the site. Taking EC,
Fe, and Sb as measures of road traffic influence, correlation
analyses among EC, crustal elements and other road trafficrelated metals were performed and their results are presented
in Table 2. During episode I, Sb was correlated with Fe (R2
= 0.71), Ca (R2 = 0.67), Cr (R2 = 0.64), and Mn (R2 = 0.53)
and to a lesser extent with Zn (R2 = 0.42), Pb (R2 = 0.41),
Cu (R2 = 0.34), and Ba (R2 = 0.26), which is suggestive of
other sources such as road dust for these elements. During
episode II, Sb had strong correlations with Cr, Ca, Ba, Zn,
Pb, Mn, Fe, and Cu (R2 = 0.88, 0.86, 0.83, 0.77, 0.76, 0.76,
0.64, and 0.46, respectively). Ca, Fe, Cr, Mn and Pb are
typically present at significant levels in crustal material and
soils, but the relatively high correlations between Sb and
the metals indicated that the concentrations of these metals
must be affected by local sources other than soil. Also EC
had moderate correlations with the metals during episodes I
and II. Because re-suspended road dust is generally too
large to dominate PM2.5, the good relations of Sb and EC
with the other metals indicate that tailpipe and brake wear
emissions were important sources for these elements at the
site, with a greater effect in episode II. Sb and Cd were very
weakly correlated (R2 = 0.04) during episode II, indicating
that other sources, rather than vehicle-related emissions,
dominated the concentration of Cd at the site.
SUMMARY AND CONCLUSION
Hourly measurements of PM2.5 mass, elemental and organic
carbon (EC and OC), ionic species, and elemental species
were made at the South Area Supersite at Gwangju, Korea,
using commercial semi-continuous instruments. During the
2 months between February 01 and March 31, 2011, a total
of 20 days wherein daily PM2.5 mass concentrations
exceeding 24-hr average Korean NAAQS of 50.0 μg/m3 were
identified. Two haze episodes out of them were discussed;
the episode I (February 01–08) was associated with regional
pollution along with severe haze layers and wildfires
emissions, and characterized further by high CO/NOx ratios
and high K+ concentrations. The episode II (March 11–12)
was characterized by locally produced pollution with low
CO/NOx ratios, short-term SO42– excursions, and broad peaks
in OC, EC, and NO3– concentrations.
The EC and OC concentrations were strongly correlated
with R2 of 0.71 and 0.83 for episodes I and II, respectively,
suggesting common source emissions (i.e., vehicles). Their
respective OC/EC ratio was 3.2 and 3.3. However, the
concentration of K+, an indicator of biomass burning
emissions, was increased for February 04–05 during the
episode I and reached an hourly maximum of 5.6 μg/m3,
accounting for ~93% of the total K concentration. This
enhancement was clearly evidenced by the MODIS satellite
image, in which numerous fire counts were observed over
the southern regions of China, and transport pathways of air
masses arriving at the site. Abnormally, however, the OC/EC
ratio (mean: 3.4) did not increase for the 2-days period.
Previous studies suggested that a low OC/EC ratio is possible
because it depends greatly on the type of materials burned and
combustion conditions. Accordingly, although low OC/EC
ratio was observed when relatively high K+ concentrations
occurred, the significant linear relationships between EC –
OC (R2 = 0.82), OC – K+ (R2 = 0.77), and K+ – K+/OC (R2 =
0.77) between February 04 and 05 suggest contribution of
carbonaceous aerosols from long-range transport of biomass
burning plumes occurring in the southern regions of China
during episode I, as well as the traffic emissions.
The good correlations between SO2 and SO42– (R2 = 0.64
and 0.85 for episodes I and II) and local wind directions
973
Table 2. Correlation coefficients among EC, crustal elements, and metals.
(a) Episode I
Component
EC
Ca
Fe
Sb
Ba
EC
0.44
0.64
0.51
0.33
Ca
0.44
0.76
0.82
0.67
Fe
0.64
0.76
0.84
0.54
(b) Episode II
Component
EC
Ca
Fe
EC
0.52
0.63
Ca
0.52
0.74
Fe
0.63
0.74
Sb
0.66
0.93
0.80
Ba
0.76
0.81
0.89
All figures are significant at p < 0.01.
Ti
0.56
0.91
0.81
0.79
0.46
Ti
0.62
0.92
0.88
0.89
0.90
Cr
0.58
0.80
0.87
0.80
0.82
Cr
0.65
0.88
0.91
0.94
0.94
suggest that local SO2 emissions for the two episodes were
an important source of SO42– formation, and that the gas-phase
oxidation of SO2 due to high O3 probably played a critical
role in the SO42– production. The time-series plot of SO42–,
the SO2-to-SO42– relationship, local wind directions, and
the air mass trajectories indicated that regional contributions
during episode I probably dominated increased SO42–
concentrations, while the local urban contribution dominated
the SO42– concentration during episode II. For NO3– particles,
the daytime NO3– concentration was often significantly
elevated over background levels during episode I, which
suggests that these particles can be attributed to the transport
of severe haze lingering over the Chinese locations upwind
of the sampling site, as well as the photochemical production
of nitric acid and low temperatures of 4.0–10°C. Similar to
the formation processes for episode I, the elevated NO3–
leading to high PM2.5 during episode II was attributed to
gas-phase photochemical reactions during the daytime and
heterogeneous hydrolysis during the nighttime.
To investigate the influence of road traffic on the
concentrations of elemental constituents observed in PM2.5
during the two episodes, correlation analyses among crustal
elements (Ca, Fe, and Mn) and road traffic-related metals
(Sb, Cu, Zn, Ba, Mn, Cr, and Pb) were performed. Good
correlations of Sb, a specific marker for brake wear, with
crustal elements and other metals revealed brake wear
emission to be an important source for these elements during
episodes I and II, with greater road traffic influence in
episode II than in episode I. These metal correlation analyses
support the conclusion that the enrichment of crustal and
trace elements at sites close to the roadside can be mainly
attributed to vehicle wear products rather than tailpipe exhaust
emissions.
ACKNOWLEDGEMENTS
This work was supported by project “Investigation on
chemical characteristics of PM2.5 and its formation processes
by areas” funded from the National Institute of Environmental
Research. This work was also partially supported by the
Mn
0.36
0.70
0.84
0.73
0.64
Mn
0.64
0.81
0.97
0.87
0.91
Cu
0.51
0.63
0.61
0.58
0.90
Cu
0.55
0.46
0.50
0.68
0.60
Zn
0.72
0.74
0.85
0.88
0.88
Zn
0.62
0.54
0.82
0.65
0.45
Cd
0.35
0.15
0.64
0.20
0.35
Sb
0.51
0.82
0.84
0.51
Sb
0.66
0.93
0.80
0.91**
Ba
0.33
0.67
0.54
0.51
Ba
0.76
0.81
0.89
0.91
-
Pb
0.74
0.71
0.74
0.64
0.66
Pb
0.69
0.68
0.68
0.87
0.78
General Researcher Program through a NRF grant funded by
the Korea government (MEST) (No. 2011-0007222). The
authors also gratefully acknowledge the NOAA Air Resources
Laboratory (ARL) for the provision of the HYSPLIT transport
and dispersion model and/or READY website (http://www.
arl.noaa.gov/ready.php) used in this publication.
REFERENCES
Amato, F., Nava, S., Lucarelli, F., Querol, X., Alastuey, A.,
Baldasano, J.M. and Pandolfi, M. (2010). Comprehensive
Assessment of PM Emissions from Paved Roads: Realworld Emission Factors and Intense Street Cleaning Trials.
Sci. Total Environ. 408: 4309–4318.
Andreae, M.O. (1983). Soot Carbon and Excess Fine
Potassium: Long Range Transport of Combustion-Derived
Aerosols. Science 220: 1148–1151.
Andreae, M.O. and Merlet, P. (2001). Emission of Trace
Gases and Aerosols from Biomass Burning. Global
Biogeochem. Cycles 15: 955–966.
Bae, M.S., Shin, J.S., Lee, K.Y. and Kim, Y.J. (2012)
Characteristics of Carbonaceous Aerosols Analyzed by the
Analysis of Organic Molecular Markers at Gosan Supersite,
Proce. 54th Meeting of KOSAE, p. 30.
Bukowiecki, N., Lienemann, P., Hill, M., Figi, R., Richard,
A., Furger, M., Rickers, K., Falkenberg, G., Zhao, Y.,
Cliff, S.S., Prevot, A.S.H., Baltensperger, U., Buchmann,
B. and Gehrig, R. (2009). Real-world Emission Factors
for Antimony and Other Brake Wear Related Trace
Elements: Size-Segregated Values for Light and Heavy
Duty Vehicles. Environ. Sci. Technol. 43: 8072–8078.
Chameides, W.L., Yu, H., Liu, S.C., Bergin, M., Zhou, X.,
Mearns, L., Wang, G., Kiang, C.S., Saylor, R.D., Luo, C.,
Haung, Y., Steiner, A. and Giorgi, F. (1999). Case Study
of the Effects of Atmospheric Aerosols and Regional Haze
on Agriculture: An Opportunity to Enhance Crop Yields
in China through Emission Controls? Proc. Nat. Acad.
Sci. U.S.A. 96: 13626–13633.
Charlson, R.J., Lovelock, J.E., Andreae, M.O. and Warren,
S.G. (1987). Oceanic Phytoplankton, Atmospheric Sulfur,
974
Cloud Albedo and Climate. Nature 326: 655–661.
Chen, L.W.A., Chow, J.C., Doddridge, B.G., Dickerson,
R.R., Ryan, W.F. and Mueller, P.K. (2003). Analysis of
a Summertime PM2.5 and Haze Episode in the MidAtlantic Region. J. Air Waste Manage. Assoc. 53: 946–956.
Chuang, M.T., Chou, C.C.K., Sopajareepom, K., Lin, NH.,
Wang, J.L., Sheu, G.R., Chang, Y.J. and Lee, C.T. (2012)
Characterization of Aerosol Chemical Properties from
Near-Source Biomass Burning in the Northern Indochina
during 7-SEAS/Dongsha Experiment. Atmos. Environ., In
Press, http://dx.doi.org/10.1016/j.atmosenv.2012. 06.056.
Draxler, R.R. and Rolph, G.D. (2012). HYSPLIT (HYbrid
Single-Particle Lagrangian Integrated Trajectory) Model
Access via NOAA ARL READY Website (http://ready.
arl.noaa.gov/HYSPLIT.php), NOAA Air Resources
Laboratory, Silver Spring, MD.
Duan, F., Liu, X., Yu, T. and Cachier, H. (2004).
Identification and Estimate of Biomass Burning
Contribution to the Urban Aerosol Organic Carbon
Concentrations in Beijing. Atmos. Environ. 38: 1275–1282.
Echalar, F., Gaudichet, A., Cachier, H. and Artaxo, P.
(1995). Aerosol Emissions by Tropical Forest and Savanna
Biomass Burning: Characteristic Trace Elements and
Fluxes. Geophys. Res. Lett. 22: 3034–3042.
Freney, E.J., Sellegri1, K., Canonaco, F., Boulon1, J. Hervo1,
M., Weigel, R., Pichon, J.M., Colomb, A., Prevot, A.S.H.
and Laj, P. (2011). Seasonal Variations in Aerosol Particle
Composition at the Puy-de-Dome Research Station. Atmos.
Chem. Phys. Discuss. 11: 27139–27170.
Furuta, N., Iijima, A., Kambe, A., Sakai, K. and Sato, K.
(2005). Concentrations, Enrichment and Predominant
Sources of Sb and Other Trace Elements in Size Classified
Airborne Particulate Matter Collected in Tokyo from 1995
to 2004. J. Environ. Monit. 7: 1155–1161.
Garg, B.D., Cadle, S.H., Mulawa, P.A., Groblicki, P.J.,
Laroo, C. and Parr, GA. (2000). Brake Wear Particulate
Matter Emissions. Environ. Sci. Technol. 34: 4463–4469.
Gertler, A.W., Lowenthal, D.A. and Coulombe, W.G.
(1995). PM10 Source Apportionment Study in Bullhead
City, Arizona. J. Air Waste Manage. Assoc. 45: 75–82.
Gomez, D.R., Gine, M.F., Bellato, A.C.S. and Smichowski,
P. (2005). Antimony: A Traffic-Related Element in the
Atmosphere of Buenos Aires, Argentina. J. Environ. Monit.
7: 1162–1168.
Guo, S., Hu, M., Wang, Z.B., Slanina, J. and Zhao, Y.L.
(2010). Size-Resolved Water-Soluble Ionic Compositions
in the Summer of Beijing: Implication of Regional
Secondary Formation. Atmos. Chem. Phys. 10: 947–959.
Harrison, R.M., Tilling, R., Romero, M.S.C., Harrad, S. and
Jarvis, K. (2003). A Study of Trace Metals and Polycyclic
Aromatic Hydrocarbons in the Roadside Environment.
Atmos. Environ. 37: 2391–2402.
Johansson, G., Norman, M. and Burman, L. (2009). Road
Traffic Emission Factors for Heavy Metals. Atmos.
Environ. 43: 4681–4688.
Jung, J. and Kim, Y.J. (2011). Tracking Sources of Severe
Haze Episodes and Their Physicochemical and
Hygroscopic Properties under Asian Continental Outflow:
Long-Range Transport Pollution, Post-harvest Biomass
Burning, and Asian Dust. J. Geophys. Res. 116: D02206.
Jung, J., Kim, Y.J., Lee, K.Y., Cayetano, M.G., Batmunkh,
T., Koo, J.H. and Kim, J. (2010). Spectral Optical
Properties of Long-Range Transport Asian Dust and
Pollution Aerosols over Northeast Asia in 2007 and
2008. Atmos. Chem. Phys. 10: 5391–5408.
Kaneyasu, N., Ohta, S. and Murao, N. (1995). Seasonal
Variation in the Chemical Composition of Atmospheric
Aerosols and Gaseous Species in Sapporo, Japan. Atmos.
Environ. 29: 1559–1568.
Kang, C.M., Lee, H.S., Kang, B.W., Lee, S.K. and Sunwoo,
Y. (2004). Chemical Characteristics of Acidic Gas
Pollutants and PM2.5 Species during Hazy Episodes in
Seoul, South Korea. Atmos. Environ. 38: 4749–4760.
Kim, D.R., Kim, J.S. and Ban, S.J. (2010). A Study on the
Characteristics of Soil in the Asian Dust Source Regions
of Mongolia. J. Korean Soc. Atmos. Environ. 26: 606–
615 (in Korean).
Kim, K.W., Kim, Y.J. and Oh, S.J. (2001). Visibility
Impairment during Yellow Sand Periods in the Urban
Atmosphere of Kwangju, Korea. Atmos. Environ. 35:
5157–5167.
Kim, K.W., Kim, Y.J. and Bang, S.Y. (2008). Summer
Time Haze Characteristics of the Urban Atmosphere of
Gwangju and the Rural Atmosphere of Anmyon, Korea.
Environ. Monit. Assess. 141: 189–199.
Koutrakis, P., Thompson, K.M., Wolfson, J.M., Spengler,
J.D., Keeler, G.J. and Slater, J.L. (1992). Determination
of Aerosol Strong Acidity Losses due to Interactions of
Collected Particles: Results from Laboratory and Field
Studies. Atmos. Environ. 26A: 987–995.
Lee, Y.L. and Sequeira, R. (2001). Visibility Degradation
across Hong Kong: Its Components and Their Relative
Contributions. Atmos. Environ. 34: 5861–5872.
Lim, H.J. and Turpin, B.J. (2002). Origins of Primary and
Secondary Organic Aerosol in Atlanta: Results of TimeResolved Measurements during the Atlanta Supersite
Experiment. Environ. Sci. Technol. 36: 4489–4496.
Malm, W.C., Molenar, J.V., Eldred, R.A. and Sisler, J.F.
(1996). Examining the Relationship among Atmospheric
Aerosols and Light Scattering and Extinction in the
Grand Canyon Area. J. Geophys. Res. 101: 19251–19265.
Menon, S., Hansen, J., Nazarenko, L. and Luo, Y. (2002).
Climate Effects of Black Carbon Aerosols in China and
India. Science 297: 2250–2253.
Morgan, W.T., Allan, J.D., Bower, K.N., Highwood, E.J.,
Liu, D., McMeeking, G.R., Northway, M.J., Williams,
P.I., Krejci, R. and Coe, H. (2010). Airborne Measurements
of the Spatial Distribution of Aerosol Chemical
Composition across Europe and Evolution of the Organic
Fraction. Atmos. Chem. Phys. 10: 4065–4083.
National Institute of Occupational Safety and Health
(NIOSH) (1996). Method 5040 Issue 1: Elemental Carbon
(Diesel Exhaust), NIOSH Manual of Analytical Methods,
4th ed., Cincinnati, OH.
Nie, W., Wang, T., Gao, X., Pathak, R.K., Wang, X., Gao,
R., Zhang, Q., Yang, L. and Wang, W. (2010). Comparison
among Filter-Based, Impactor-Based and Continuous
Techniques for Measuring Atmospheric Fine Sulfate and
Nitrate. Atmos. Environ. 44: 4396–4403.
Oanh, N.T.K. and Leelasakultum, K. (2011). Analysis of
Meteorology and Emission in Haze Episode Prevalence
over Mountain-Bounded Region for Early Warning, Sci.
Total Environ. 409: 2261–2271.
Okada, K., Ikegami, M., Zaizen, Y., Makino, Y., Jensen,
J.B. and Gras, J.L. (2001). The Mixture State of Individual
Aerosol Particles in the 1997 Indonesian Haze Episode.
J. Aerosol Sci. 32: 1269–1279.
Park, S.S., Bae, M.S., Schauer, J.J., Ryu S.Y., Kim, Y.J.,
Cho, S.Y. and Kim, S.J. (2005a). Evaluation of the TMO
and TOT Methods for OC and EC Measurements and
Their Characteristics in PM2.5 at an Urban Site of Korea
during ACE-Asia. Atmos. Environ. 39: 5101–5112
Park, S.S., Harrison, D., Pancras, P. and Ondov, J.M.
(2005b). Time Resolved Elemental and Organic Carbon
Measurements at the Baltimore Supersite in 2002. J.
Geophys. Res. 110: D07S06, doi: 10.1029/2004JD004610.
Park, S.S., Ondov. J.M., Harrison, D. and Nair, N.P. (2005c).
Seasonal and Shorter-term Variations in Particulate
Atmospheric Nitrate in Baltimore. Atmos. Environ. 39:
2011–2020.
Park, S.S., Kleissl, J., Harrison, D., Kumar, V., Nair, N.P.,
Adam, M., Ondov, J. and Parlange, M. (2006a)
Characteristics of PM2.5 Episodes Revealed by SemiContinuous Measurements at the Baltimore Supersite at
Ponca St. Aerosol Sci. Technol. 40:845–860.
Park, S.S., Bae, M.S., Schauer, J.J., Kim, Y.J., Cho, S.Y.
and Kim, S.J. (2006b). Molecular Composition of PM2.5
Organic Aerosol Measured at an Urban Site of Korea
during the ACE-Asia Campaign. Atmos. Environ. 40:
4182–4198.
Park, S.S., Cho, S.Y. and Kim, S.J. (2010). Chemical
Characteristics of Water-Soluble Components in Fine
Particulate Matter at a Gwangju Area. Korean J. Chem.
Eng. 48: 20–26 (in Korean).
Park, S.S. and Cho, S.Y. (2011). Tracking Sources and
Behaviors of Water-Soluble Organic Carbon in Fine
Particulate Matter Measured at an Urban Site in Korea.
Park, S.S., Kim, J.H. and Jeong, J.U. (2012). Abundance
and Sources of Hydrophilic and Hydrophobic WaterSoluble Organic Carbon at an Urban Site in Korea in
Summer. J. Environ. Monit. 14: 224–232.
Polidori, A., Turpin, B.J., Lim, H.J., Cabada, J.C.,
Subramanian, R., Pandis, S.N. and Robinson, A.L. (2006).
Local and Regional Secondary Organic Aerosol: Insights
from a Year of Semi-Continuous Carbon Measurements
at Pittsburgh. Aerosol Sci. Technol. 40: 861–872.
Ram, K. and Sarin, M.M. (2010). Spatio-Temporal
Variability in Atmospheric Abundances of EC, OC and
WSOC over Northern India. J. Aerosol Sci. 41: 88–98.
Ram, K. and Sarin, M.M. (2011). Day-Night Variability of
EC, OC, WSOC and Inorganic Ions in Urban Environment
of Indo-Gangetic Plain: Implications to Secondary Aerosol
Formation. Atmos. Environ. 45: 460–468.
Ram, K., Sarin, M.M. and Tripathi, S.N. (2012a). Temporal
Trends in Atmospheric PM2.5, PM10, Elemental Carbon,
Organic Carbon, Water-Soluble Organic Carbon, and
975
Optical Properties: Impact of Biomass Burning Emissions
in the Indo-Gangetic Plain. Environ. Sci. Technol. 46:
686–695.
Ram, K., Sarin, M.M., Sudheer, A.K. and Rengarajan, R.
(2012b). Carbonaceous and Secondary Inorganic Aerosols
during Wintertime Fog and Haze over Urban Sites in the
Indo-Gangetic Plain. Aerosol Air Qual. Res. 12: 359–370.
Reid, J.S., Koppmann, R., Eck, T.F. and Eleuterio, D.P.
(2005). A Review of Biomass Burning Emissions Part
II: Intensive Physical Properties of Biomass Burning
Particles. Atmos. Chem. Phys. 5: 799–825.
Ryu, S.Y., Kim, J.E., Kim, Y.J. and Gang, G.U. (2004).
Chemical Composition of Post-Harvest Biomass Burning
Aerosols in Gwangju, Korea. J. Air Waste Manage. Assoc.
54: 1124–1137.
Ryu, S.Y., Kwon, B.G., Kim, Y.J., Kim, H.H. and Chun,
K.J. (2007). Characteristics of Biomass Burning Aerosol
and Its Impact on Regional Air Quality in the Summer of
2003 at Gwangju, Korea. Atmos. Res. 84: 362–373.
Saarnio, K., Aurela, M., Timonen, H., Saarikoski, S.,
Teinila, K., Makela, T., Sofiev, M., Koskinen, J., Aalto,
P.P., Kulmala, M., Kukkonen, J. and Hillamo, R. (2010).
Chemical Composition of Fine Particles in Fresh Smoke
Plumes from Boreal Wild-land Fires in Europe. Sci. Total
Environ. 408: 2527–2542.
Sahu, L.K., Kondo, Y., Miyazaki, Y., Kuwata, M., Koike,
M., Takegawa, N., Tanimoto, H., Matsueda, H., Yoon,
S.C. and Kim, Y.J. (2009). Anthropogenic Aerosols
Observed in Asian Continental Outflow at Jeju Island,
Korea, in Spring 2005. J. Geophys. Res. 114: D03301,
doi: 10.1029/2008JD010306.
Schauer, J.J., Mader, B.T., Deminter, J.T., Heidemann, G.,
Bae, M.S., Seinfeld, J.H., Flagan, R.C., Cary, R.A., Smith,
D., Huebert, B.J., Bertram, T., Howell, S., Kline, J.T.,
Quinn, P., Bates, T., Turpin, B., Lim, H.J., Yu, J.Z., Yang,
H. and Keywood, M.D. (2003). ACE-Asia Intercomparison
of a Thermal-Optical Method for the Determination of
Particle-Phase Organic and Elemental Carbon. Environ.
Sci. Technol. 37: 993–1001.
Schauer, J.J., Lough, G.C., Shafer, M.M., Christensen,
W.F., Arndt, M.F., DeMinter, J.T. and Park, J.S. (2006).
Characterization of Metals Emitted from Motor Vehicles,
Health Effect Institute.
Schichtel, B.A., Husar, R.B., Falke, S.R. and Wilson,W.E.
(2001). Haze Trends over the United States, 1980–1995.
Seinfeld, J.H. and Pandis, S.N. (2006). Atmospheric
Chemistry and Physics: From Air Pollution to Climate
Change, 2nd Ed., John Wiley & Sons, Inc.
Simoneit, B.R.T., Kobayashi, M., Mochida, M., Kawamura,
K., Lee, M., Lim, H.J., Turpin, B.J. and Komazaki, Y.
(2004). Composition and Major Sources of Organic
Compounds of Aerosol Particulate Matter Sampled during
the ACE-Asia Campaign. J. Geophys. Res. 109: D19S10.
Smith, N., Plane, J.M.C., Nien, C. and Solomon, P.A.
(1995). Nighttime Radical Chemistry in the San Joanquin
Valley. Atmos. Environ.29: 2887–2897.
Solomon, P.A. and Sioutas, C. (2008). Continuous and
Semicontinuous Monitoring Techniques for Particulate
976
Matter Mass and Chemical Components: A Synthesis of
Findings from EPA’s Particulate Matter Supersites
Program and Related Studies. J. Air Waste Manage. Assoc.
58: 164–195.
Sternbeck, J., Sjodin, A. and Andreasson, K. (2002). Metal
Emissions from Road Traffic and the Influence of
Resuspension – Results from Two Tunnel Studies. Atmos.
Environ. 36: 4735–4744.
Sun, Y.L., Zhuang, G.S., Tang, A.H., Wang, Y. and An,
Z.S. (2006). Chemical Characteristics of PM2.5 and PM10
in Haze–Fog Episodes in Beijing. Environ. Sci. Technol.
40: 3148–3155.
Thorpe, A. and Harrison, R. M. (2008). Sources and
Properties of Non-Exhaust Particulate Matter from Road
Traffic: A Review. Sci. Total Environ. 400: 270–283.
Turn, S.Q., Jenkins, B.M., Chow, J.C. and Pritchett, L.C.
(1997). Elemental Characterization of Particulate Matter
Emitted from Biomass Burning: Wind Tunnel Derived
Source Profiles for Herbaceous and Wood Fuels. J.
Geophys. Res. 102(D3): 3683–3699.
Turpin, B.J., Saxena, P. and Andrews, E. (2000).
Measuring and Simulating Particulate Organics in the
Atmosphere: Problems and Prospects. Atmos. Environ.
34: 2983–3013.
Turpin, B.J. and Lim, H.J. (2001). Species Contributions to
PM2.5 Mass Concentrations: Revisiting Common
Assumptions for Estimating Organic Mass. Aerosol Sci.
Technol. 35: 602–610.
Watson, J.G. (2002). Visibility: Science and Regulation. J.
Air Waste Manage. Assoc. 52: 628–713.
Watson, J.G. and Chow, J.C. (2002). A Wintertime PM2.5
Episode at the Fresno, CA, Supersite. Atmos. Environ.
36: 465–475.
Weber, R., Bergin, M., Kiang, C.S., Chameides, W., Orsini,
D., St, J.J., Chang, M., Bergin, M., Carrico, C., Lee,
Y.N., Dasgupta, P., Slanina, J., Turpin, B., Edgerton, E.,
Hering, S., Allen, G. and Solomon, P. (2003). ShortTerm Temporal Variation in PM2.5 Mass and Chemical
Composition during the Atlanta Supersite Experiment. J.
Air Waste Manage. Assoc. 53: 84–91.
Weckwerth, G. (2001). Verification of Traffic Emitted
Aerosol Components in the Ambient Air of Cologne
(Germany). Atmos. Environ. 35: 5525–5536.
Wittig, A.E., Takahama, S., Khlystov, A.Y., Pandis, S.N.,
Hering, S., Kirby, B. and Davidson, C. (2004). Semicontinuous PM2.5 Inorganic Composition Measurements
during the Pittsburgh Air Quality Study. Atmos. Environ.
38: 3201–3213.
Wu, W.S. and Wang, T. (2007). On the Performance of a
Semi-continuous PM2.5 Sulphate and Nitrate Instrument
under High Loadings of Particulate and Sulphur Dioxide.
Atmos. Environ.41: 5442–5451.
Yadav, A.K., Kumar, K., Kasim, A., Singh, M.P., Parida,
S.K. and Sharan, M. (2003). Visibility and Incidence of
Respiratory Diseases during the 1998 Haze Episode in
Brunei Darussalam. Pure Appl. Geophys. 160: 265–277.
Zhang, X.Q. and McMurry, P.H. (1992). Evaporative Loss
of Fine Particulate Nitrates during Sampling. Atmos.
Environ. 26A: 3305–3312.
Zhang, Y., Shao, M., Lin, Y., Luan, S., Mao, N., Chen, W.
and Wang, M. (2012a). Emission Inventory of
Carbonaceous Pollutants from Biomass Burning in the
Pearl River Delta Region, China. Atmos. Environ., In
Press, doi: 10.1016/j.atmosenv.2012.05.055.
Zhang, R., Tao, J., Ho, K.F., Shen, Z., Wang, G., Cao, J.,
Liu, S., Zhang, L. and Lee, S.C. (2012b). Characterization
of Atmospheric Organic and Elemental Carbon of PM2.5
in a Typical Semi-Arid Area of Northeastern China.
Aerosol Air Qual. Res. 12: 792–802.
Zhou, Y., Wang, T., Gao, X., Xue, L., Wang, X., Wang, Z.,
Gao, J., Zhang, Q. and Wang, W. (2009). Continuous
Observations of Water-Soluble Ions in PM2.5 at Mount
Tai (1534 m a.s.l.) in Central-Eastern China. J. Atmos.
Chem. 64: 107–127.
Received for review, July 12, 2012
Accepted, December 16, 2012

Article Full Text - Aerosol and Air Quality Research

Transcription

Similar documents

Peace River - AEP - Government of Alberta

iGeneration - Landmark Media

13th Conference

Street Games Pitch Presentation

4-km AIRPACT vs 12

Woodheater-Car Comparison

Air Pollution in China: Mapping of Concentrations

Selected Filmography

TECHNIKA TRANSPORTU SZYNOWEGO - tts