Zeng 2020

JID: JES
ARTICLE IN PRESS [m5G;May 11, 2020;19:45]

Journal of Environmental Sciences xxx (xxxx) xxx
Contents lists available at ScienceDirect
Journal of Environmental Sciences

journal homepage: www.elsevier.com/locate/jes
Relationships between chemical elements of PM2.5 and O3 in Shanghai

atmosphere based on the 1-year monitoring observation
Junyang Zeng 1, Luying Zhang 1, Chuanhe Yao 1, Tingting Xie 1, Lanfang Rao 1, Hui Lu 2,
Xinchun Liu 2,∗, Qingyue Wang 3, Senlin Lu 1,∗
1
School of Environmental and Chemical Engineering, Shanghai University, Shanghai 200444, China
2
Institute of Desert Meteorology, China Meteorological Administration, Urumqi 83002, China
3
School of Science and Engineering, Saitama University, Saitama 338-8570, Japan
a r t i c l e i n f o a b s t r a c t
Article history: Mass level of fine particles (PM2.5 ) in main cities in China has decreased significantly in recent years
Received 10 September 2019 due to implementation of Chinese Clean Air Action Plan since 2013, however, O3 pollution is getting
Revised 3 March 2020
worse than before, especially in megacities such as in Shanghai. In this work, O3 and PM2.5 were con-
Accepted 17 March 2020
tinuously monitored from May 27, 2018 to March 31, 2019. Our data showed that the annual aver-
Available online xxx
age concentration of PM2.5 and O3 (O3 -8 hr, maximum 8-hour moving average of ozone days) was
Keywords: 39.35 ± 35.74 and 86.49 ± 41.65 μg/m3 , respectively. The concentrations of PM2.5 showed clear seasonal
Shanghai PM2.5 trends, with higher concentrations in winter (83.36 ± 18.66 μg/m3 ) and lower concentrations in summer
O3 (19.85 ± 7.23 μg/m3 ), however, the seasonal trends of O3 were different with 103.75 ± 41.77 μg/m3 in
Chemical elements summer and 58.59 ± 21.40 μg/m3 in winter. Air mass backward trajectory, analyzing results of potential
source contribution function model and concentration weighted trajectory model implied that pollutants
from northwestern China contributed significantly to the mass concentration of Shanghai PM2.5 , while
pollutants from areas of eastern coastal provinces and South China Sea contributed significantly to the
mass level of ozone in Shanghai atmosphere. Mass concentration of twenty-one elements in the PM2.5
were investigated, and their relationships with O3 were analyzed. Mass level of ozone had good correla-
tion with that of Ba (r = 0.64, p < 0.05) and V (r = 0.30, p > 0.05), suggesting vehicle emission pollutants
contribute to the increasing concentration of ozone in Shanghai atmosphere.
© 2020 Published by Elsevier B.V. on behalf of The Research Center for Eco-Environmental Sciences,
Chinese Academy of Sciences.
Introduction and PMs to the atmosphere from industrial and commercial ac-
tivities (Xu et al., 2019). Previous studies showed that mass con-
Particulate matters (PMs) were widely studied due to their en- centration of PM2.5 in Shanghai atmosphere showed seasonal vari-
vironmental effects and various harmful effects on human health ation with much higher values in winter and spring, lower val-
in the last decades (Lu et al., 2008, 2011). Based on their aerody- ues in summer, and the lowest in autumn (Lu et al., 2008). Its
namic diameter (D), particulate matters can be divided into coarse annual average concentration (from 2013 to 2015) of PM2.5 was
particles (PM10 , D ≤ 10 μm), fine particles (PM2.5, D ≤ 2.5 μm) 52.7 μg/m3 (Wang et al., 2019). The PM2.5 was associated with
and ultrafine particles (D ≤ 0.1 μm). Due to the fine particles all-cause mortality and respiratory disease mortality in Shanghai
have smaller particle size and larger surface area per unit mass (Wang et al., 2019). Additionally, PM2.5 chemical components (such
compared with that of coarse particles, various harmful gasses as NH4 NO3 and (NH4 )2 SO4 ) were contributed to the light extinc-
and heavy metals can be easily adsorbed on PM2.5 . Shanghai, the tion efficient in Shanghai (Zhou et al., 2016), and sources of PM2.5
biggest city in China, has a fleet of over 3.6 million vehicles and was mainly contributed by secondary aerosol and vehicle exhaust
a population of over 2400 million permanent residents, which re- (Li et al., 2019a).
sults in high emissions of NOx , volatile organic compounds (VOCs), O3 is a secondary pollutant formed via photochemical re-
actions among precursors including nitrogen oxides (NOx ) and
VOC under solar radiation. Ground-level ozone is prescribed as
one of six criteria air pollutants by the United States Environ-
∗
Corresponding authors.
mental Protection Agency (US EPA) (Zhang et al., 2019a, 2019b;
E-mail addresses: [email protected] (X. Liu), [email protected] (S. Lu). Zhao et al., 2018), and regarded as a threat to human health
https://doi.org/10.1016/j.jes.2020.03.043
1001-0742/© 2020 Published by Elsevier B.V. on behalf of The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences.
Please cite this article as: J. Zeng, L. Zhang and C. Yao et al., Relationships between chemical elements of PM2.5 and O3 in Shanghai
atmosphere based on the 1-year monitoring observation, Journal of Environmental Sciences, https://doi.org/10.1016/j.jes.2020.03.043
JID: JES
2 J. Zeng, L. Zhang and C. Yao et al. / Journal of Environmental Sciences xxx (xxxx) xxx
(https://www.who.int/airpollution/en/), vegetation (Dolker et al., tat to identify spatial sources of PM2.5 and O3 (Zhao et al., 2019a,
2019), and infrastructure. 2019b). The PSCF values for i × j grid cells in the domain were cal-
Mass level of fine particles (PM2.5 ) in main cities in China has culated by counting the trajectory segment endpoints terminating
decreased significantly in recent years due to implementation of within each cell. By defining the number of endpoints that falls in
Chinese Clean Air Action Plan since 2013, however, reductions in the ijth cell as nij and the number of endpoints that corresponds
PM2.5 and O3 were not apparent over the last years (Li et al., to pollutant concentrations larger than an arbitrarily set criterion
2019b). O3 pollution is getting worse than before, especially in when arriving at the receptor site in the same grid cell as mij , the
megacities such as Beijing and Shanghai. The days that O3 as the PSCF value for the ijth cell is defined as (Ashbaugh et al., 1985):
major pollutant already outnumbered PM10 (Xu et al., 2019). Ac- mi j
cording to the Shanghai Environmental Bulletin (2018), the 90th PSCFi j =
ni j
percentile of the average of maximum daily 8-hr ozone concentra-
tion (O3 -8 hr) was 84.47 ppbV and increased 10.4% in 2017 com- The PSCF value represents a conditional probability describing
pared with that in 2016. Li et al. (2019b) reported that O3 pol- the potential contribution of a grid cell to the high pollutant load-
lution in Shanghai, was mainly caused by the joint effect of lo- ings at the receptor site. In order to reduce the uncertainty in cells
cal to regional precursor emissions combined with super-regional by decreasing the effect of small nij value, the value of PSCF was
O3 transport. Shu et al. (2019) demonstrated that super-regional multiplied with an arbitrary weighting function Wij as following
transport was prominent (30.3%–63%) in O3 pollution in Shanghai. (Wang et al., 2009):
However, most of the studies focused on specific ozone episodes,

0.7 20 < ni j < 80
thus, the results were applicable for these specific cases. Until now, Wi j = 0.42 10 < ni j < 20
reports on long term observation of O3 and PM2.5 , especially, rela- 0.17 ni j ≤ 10
tionship between O3 and chemical elements in the PM2.5 has not
been reported. Therefore, our objective was to investigate poten- The PSCF value only gives the spatial distribution of poten-
tial sources and the inter-relationships between the chemical ele- tial emission sources without information on the relative sig-
ments in PM2.5 and O3 basing on long term observation. We tried nificance of the source regions. To compensate the limitation,
to provide fundamental data for the prevention and control mea- a concentration-weighted trajectory (CWT) method developed by
surement of PM2.5 and O3 . Hsu et al. (2003) was used to calculate the trajectory weighted
concentration. The studied domain was in the range of 10°N to
1. Methodology 60°N and 80°E to 150°E with the resolution of 0.5° × 0.5°, which
contained almost all regions overlaid with entire airflow transport
1.1. Sampling site and monitoring methods pathways. The CWT could be defined as:
k
M
Sampling site was located on the roof of a teaching building in Cij = M Ck τijk
the campus of Shanghai University (31°19 0 N, 121°23 25 E) and k=1 τijk k=1
the height was ~15 m above the ground level. PM2.5 samples were where, Cij was the CWT value in the ijth cell, k was the index of
collected by using of PM2.5 sampler (FLD-1, SIBATA, Japan) with a the trajectories that passed the ijth cell, M was the total number
flow rate of 1.7 L/min from 12:00 at noon and lasted for about of the trajectories, Ck was the concentration of trajectory l, and τi jk
72 hr. 103 samples were collected on quartz fiber filters (SIBATA was the number of points falling in the ijth cell by trajectory k. A
Company, with diameter of 47 mm). All filters were baked at 450°C high CWT value in a cell implied that air parcels traveling over the
for 6 hr and kept in aluminum foil envelopes before sampling. The cell would be associated with high concentrations at the receptor
filters were weighted before and after sampling by a microbal- station. The weighting function described above was also used in
ance (balance sensitivity: ±0.001 mg) under the condition of con- the CWT analyses. Weighted PSCF and CWT (WPSCF and WCWT)
stant temperature (25°C) and relative humidity (40%). Then all fil- were used in the results section.
ters were stored in the freezer at −20°C until chemical analysis.
Ozone analyzer (Thermo 49i, ThermoFisher Scientific, USA) was 1.4. Chemical elements in PM2.5
used to continuously and automatically monitor ozone concentra-
tion in ambient air. The O3 -8 hr value was calculated from the on- Based on the air mass trajectory and seasons, 25 samples were
line data monitored by zone analyzer (Thermo 49i, ThermoFisher selected from the total 103 samples for chemical elements analysis.
Scientific, USA). Parameters of the ozone analyzer are listed in Ap- The 25 samples were divided into six types of samples i.e. spring
pendix A Table S1. samples, summer samples, autumn samples, winter samples, north
samples (the air mass from north direction) and northwest sam-
1.2. Backward trajectory model and clustering analysis ples (the air mass from northwest direction). It was worth noting
that the autumn samples were also north samples. The selected
HYSPLIT model (http://www.noaa.gov) was employed to analy- samples’ information is shown in Table 1.
sis the air backward trajectories during the monitoring period. The Chemical analyses of the particles were carried out using an Ep-
model was run at starting times of 04:0 0, 10:0 0, 16:0 0 and 22:00 silon 4 energy dispersive X-ray fluorescence spectrometer (EDXRF,
coordinated universal time (UTC) (12:0 0, 18:0 0, 0 0:0 0 and 06:0 0 Malvern Panalytical, Netherlands). Chemical elements from sodium
for local time, respectively) every day during sampling period. The (Na) to americium (Am) could be analyzed by using of the EDXRF.
duration of the calculation was 72 hr, and the height was 500 m A blank quartz filter was used as a control in the experiment.
above ground level. In this study, the used criterions of the PM2.5
and O3 -8 hr were 75 and 160 μg/m3 , respectively. 1.5. Statistical analysis
1.3. Potential source contribution function (PSCF) and concentration Data of mass level of PM2.5 and O3 were analyzed with Graph-
weighted trajectory (CWT) Pad InStat software (Version 5.0, GraphPad Software, USA). Statis-
tical significance between O3 and chemical elements was carried
Potential source contribution function (PSCF) and concentration out by using one-way analysis of variance with post hoc Tukey’s
weighted trajectory (CWT) were analyzed by using software TrajS- pairwise comparisons.
JID: JES
J. Zeng, L. Zhang and C. Yao et al. / Journal of Environmental Sciences xxx (xxxx) xxx 3
Table 1
Sampling information of the selected PM2.5 samples.
No. Summer Autumn/North Winter Spring Northwest
1 2018/06/26–2018/06/29 2018/11/07–2018/11/10 2019/01/11–2019/01/14 2019/03/17–2019/03/20 2018/12/02–2018/12/05

2 2018/06/29–2018/07/02 2018/11/10–2018/11/13 2019/01/14–2019/01/17 2019/03/20–2019/03/23 2018/12/05–2018/12/08
3 2018/07/02–2018/07/05 2018/11/13–2018/11/16 2019/01/17–2019/01/20 2019/03/23–2019/03/26 2018/12/17–2018/12/20
4 2018/07/05–2018/07/08 2018/11/16–2018/11/19 2019/01/20–2019/01/23 2019/03/29–2019/04/01 2018/12/20–2018/12/23
5 2018/07/08–2018/07/11 2018/11/19–2018/11/22 2019/01/23–2019/01/26 2019/04/01–2019/04/04 2018/12/23–2018/12/26
Fig. 1. Variations of daily PM2.5 concentration and daily O3 -8 hr concentration during sampling period. Histograms were seasonal average value of PM2.5 and O3 -8 hr.
2. Results and discussion ber 2018 and March 2019. The percentage that O3 -8 hr pol-
luted days exceeded the Chinese pollution standard (160 μg/m3 )
2.1. Mass concentrations of PM2.5 and ozone was varied from 0% to 8.84%. It is worth noting that the va-
riety of mass level of PM2.5 and O3 was similar during the
Fig. 1 shows the daily average mass concentration of PM2.5 and sampling data from 25th May to 26th October 2018, coefficient
O3 at the sampling site during the sampling period. The mass con- between mass level of PM2.5 and O3 was 0.47, suggesting the
centration of PM2.5 ranged from 6.04 to 220.74 μg/m3 , with an av- mass concentration of the two kinds pollutant have a similar
erage of 39.35 ± 35.74 μg/m3 . Among the whole sampling days variation. This phenomenon was also reported by Zhao et al.
(293 days), the PM2.5 mass concentration in 44 days exceeded (2015). However, the coefficient between mass level of PM2.5 and
the Chinese pollution standard of 75 μg/m3 (PM2.5 polluted day), O3 was 0.18 from 27th October to 31st March 2018, reflecting
and the polluted days ratio was 15.02%. High PM2.5 concentration the variation of the mass concentration of PM2.5 and O3 was
was found in autumn and winter, the high PM2.5 concentration different.
(55.52 ± 40.82 μg/m3 ) was found from November 2018 to March
2019, with a polluted days percentage of 27.97%. The low PM2.5
concentration with the average of 22.08 ± 15.85 μg/m3 was found 2.2. Cluster analysis
in spring and in summer. During this period, only two days could
be identified as PM2.5 polluted days (on the 5th June, 91.21 μg/m3 ; The back trajectories were clustered into four groups based on
on the 24th October, 78.00 μg/m3 ) (detail information is presented air mass directions (Table 2): Cluster 1 was from east, Cluster 2
in Appendix A Table S2). was from north, Cluster 3 was from southwest and Cluster 4 was
The variation of O3 -8 hr at the sampling site ranged from 6.27 from northwest (Fig. 2). Cluster 1 accounted 37.8% among all the
to 252.91 μg/m3 , with an average of 86.49 ± 41.65 μg/m3 during analyzed trajectories, while Cluster 2, Cluster 3 and Cluster 4 ac-
the sampling period. The variation of O3 -8 hr value was differ- counted 27.2%, 12.5% and 22.5% respectively. The cluster 1 origi-
ent with that of PM2.5 , showing high concentration in spring and nated from eastern sea surface. Cluster 2 originated from Hebei
summer, and low concentration in autumn and winter (Novem- and Liaoning provinces. Cluster 3 was from Guangdong Province
ber 2018 to February 2019). A period of low O3 -8 hr with the and the pathway traveled through Zhejiang and Fujian Province.
average of mass level (58.59 ± 21.40 μg/m3 ), and high O3 -8 Cluster 4 arrived in Shanghai after travelling through Mongolia,
hr level (103.75 ± 41.77 μg/m3 ) was found in May to Octo- Hebei and Shandong Province.
JID: JES
Table 2
Cluster statistics of 4 clusters with different air directions.
Cluster Trajectory Mean concentration Standard deviation Polluted trajectory Polluted trajectory Mean concentration Standard deviation
number number of trajectories of trajectories number ratio of polluted of polluted
trajectories trajectories
1 436 30.67 31.01 50 11.5% 103.68 25.66

2 313 47.40 45.44 62 19.8% 124.48 41.53
3 144 35.28 23.35 13 9.0% 92.98 4.25
4 259 48.44 32.20 51 19.7% 99.73 22.28
Total 1152 39.79 35.91 176 15.3% 109.07 32.65
Fig. 2. Air mass backward trajectories for Shanghai during the sampling time from 27th May 2018 to 31th March 2019 (No. GS (2016) 1585 (http://bzdt.ch.mnr.gov.cn/index.
html)).
The number of trajectories assigned to Cluster 1 was the high- ground surface was mainly formed by some complex chemical re-
est (37.8%) and the air masses associated with this cluster led to actions between nitrogen oxides (NOx ) and volatile organic com-
the lowest PM2.5 loading (30.67 ± 31.01 μg/m3 ) among all four pounds (VOCs) from sources such as motor vehicles and industry
clusters. Most trajectories in the Cluster 1 were relatively clean (Wang et al., 2017). Geng et al. (2008) reported that automobiles
and came from marine. Cluster 2 with high PM2.5 concentration played important roles in the O3 formation in the city of Shang-
(47.40 ± 45.44 μg/m3 ) was a short-range trajectory. Most trajec- hai, therefore, pollutant sources on this cluster might largely con-
tories of Cluster 3 originated from economically developed areas tributed to the increasing mass level of O3 . Additionally, in sum-
such as Zhejiang province, Fujian province, Guangdong province, mer, Cluster 3 traveled through the developed coastal provinces
and the air masses associated with this cluster led to relatively low (Guangdong, Fujian and Zhejiang provinces) with the huge traffic
(35.28 ± 23.35 μg/m3 ) of PM2.5 . The highest average PM2.5 concen- flow, and then the ozone precursors such as NOx and VOCs from
tration (48.44 ± 32.20 μg/m3 ) associated with Cluster 4 (from the automobile exhaust and industrial emissions were higher than that
northwest to Shanghai) was found in winter and spring. Zhao et al. in other regions (Zhao et al., 2015). Therefore, the high concentra-
(2015) reported that the air masses originated from these regions, tions of NOx and VOCs, high temperature and strong light irradi-
such as the Inner Mongol and the Gobi desert might carry dust ation triggered the higher ozone mass concentration in the areas
and mixed with other pollutants (SOx , NOx ) emitted from com- than in other regions.
bustion on their pathways (including Shanxi and Hebei Province,
where large amounts of coal were used for heating) and arrived at 2.3. PSCF and CWT
Shanghai, and led to high mass concentration of PM2.5 .
It is worth noting that the mean concentration of Cluster 3, Spatial distributions of PM2.5 and ozone potential sources ob-
polluted trajectory mean value and polluted ratio of Cluster 3 tained by using the software of TrajStat are shown in Figs. 3
were much higher than other clusters. Because the ozone near the and in 4, respectively. The higher CWT and PSCF values indicated
JID: JES
Fig. 3. Potential source contribution function (PSCF) and concentration weighted trajectory (CWT) maps for PM2.5 arriving at 25-m altitude at Shanghai city showing potential
source regions contributing to PM2.5 concentrations during sampling period (No. GS (2016) 1585 (http://bzdt.ch.mnr.gov.cn/index.html)).
JID: JES
Fig. 4. PSCF and CWT maps for O3 arriving at 25-m altitude at Shanghai city showing potential source regions contributing to O3 concentrations during sampling period
(No. GS (2016) 1585 (http://bzdt.ch.mnr.gov.cn/index.html)).
JID: JES
Table 3
Concentration (ng/m3 ) of different elements in different PM2.5 sample.
Elements Summer Autumn/North Winter Spring Northwest r
Na 4296.77±241.18 2749.73±1152.02 355.93±249.62 580.04±233.43 3832.49±523.45 −0.26∗

Mg 1461.69±19.89 1168.73±498.40 196.02±123.63 203.12±106.08 1189.41±235.51 −0.26∗∗
Al 2662.21±185.24 2539.81±477.78 258.69±201.65 449.46±199.21 2593.37±412.35 −0.32∗
P 19.68±2.96 21.22±10.43 6.69±1.46 10.80±3.24 16.98±8.11 −0.09
S 954.80±247.31 943.69±175.35 1485.15±413.51 805.39±120.66 892.76±223.44 −0.45
Cl 313.33±142.35 275.51±54.36 60.81±25.86 72.54±53.33 404.09±154.14 −0.45
K 625.42±60.06 2956.94±2415.64 868.06±80.96 661.54±67.85 940.37±232.13 −0.29
Ca 5975.16±245.84 5026.64±2620.34 824.22±90.56 385.87±299.66 6475.33±797.78 −0.43∗
Ti 10.80±5.94 23.54±21.54 7.10±6.22 9.26±1.95 15.43±12.99 −0.33
V 9.57±4.30 4.63±1.89 3.47±0.67 3.34±1.80 3.01±2.40 0.30
Cr 20.07±3.24 18.21±6.20 7.10±2.31 3.70±2.31 21.61±5.61 −0.50
Mn 17.60±6.52 30.25±8.87 44.45±14.06 27.47±3.30 43.22±26.56 −0.75
Fe 310.55±93.14 417.05±128.67 456.56±131.70 350.37±31.44 572.94±282.96 −0.85
Ni 9.57±2.05 8.03±2.66 6.48±1.80 5.56±1.23 7.72±3.78 −0.22∗
Cu 8.64±3.98 12.04±5.11 17.90±7.91 8.33±2.09 14.20±8.59 −0.83
Zn 31.18±24.65 73.78±22.27 142.93±45.80 79.03±11.53 104.65±64.57 −0.56
Br 5.87±1.16 8.95±0.62 16.67±5.38 12.04±2.05 12.35±6.02 −0.35
Sr 13.27±3.32 11.73±7.41 8.64±2.09 8.64±3.73 16.98±7.43 −0.51∗
Ba 25.00±2.99 16.59±12.75 27.17±4.43 28.40±5.03 16.98±7.04 0.64∗
Pb 8.64±6.41 18.37±6.72 41.52±10.30 24.85±3.52 38.12±28.07 −0.58
O3 85.06±48.55 62.19±18.98 53.44±20.64 108.77±17.07 42.29±17.04
∗
Correlation (r) is significant at the 0.05 level.
∗∗
Correlation (r) is significant at the 0.01 level.
more remarkable potential source regions, which were repre- level of Ca in summer and in autumn PM2.5 was probably caused
sented by the color of legend. The deeper in red represented the by Shangda road maintenance activities, which could coarse parti-
higher level of pollution while the deeper in blue was on the cles suspended in the air.
opposite. K was regarded as a marker for biomass burning. The highest
Fig. 3 showed that the larger values of PSCF were mainly mass level of K (2956.94 ± 2415.64 ng/m3 ) was found in autumn
concentrated in Shaanxi Province, Ningxia Hui Autonomous Re- samples, this data agreed with previous reports (Wang et al., 2013;
gion, Gansu Province, western Inner Mongolia and northeastern Zhao et al., 2015). It is worth noting that the mean concentration
Xinjiang. The larger values of CWT were mainly concentrated in of K in autumn samples is much higher than that in other sea-
Jiangsu province, Shanxi province, Shaanxi province, Ningxia Hui sons (4.7, 3.4 and 4.5 times higher than that in summer, winter
Autonomous Region, Gansu province, western Inner Mongolia and and spring samples respectively). On the one hand, Zhang et al.
northeastern Xinjiang. Weak CWT values were mainly found in sea (2019a, 2019b) reported that high emissions of crop residue burn-
surface in eastern Shanghai. In general, both PSCF and CWT al- ing were generally located in northern and northeastern China,
gorithm indicated Shanxi province, Gansu province, Ningxia Hui, with outstanding peaks in October. On the other hand, the con-
Autonomous Region, Inner Mongolia, and northeastern Xinjiang as centration of K was affected by burning a large amount of coal
major potential source areas of PM2.5 . and biomass fuels every autumn and winter in northern China. The
The spatial distributions of potential source areas and transport highest of S (1485.15 ± 413.51 ng/m3 ), which was a trace marker
pathways of ozone are illustrated in Fig. 4. The larger PSCF and for coal combustion, was found in winter samples. It was noted
CWT value was mainly found in Zhejiang province, Fujian province, that seasonal variety of the S was not significantly different, sug-
Guangdong province and South China Sea, while moderate CWT gesting the S in Shanghai atmosphere may be contributed by other
value occurred in the northeast region, such as in Jilin, Liaoning pollutant sources, such as industrial emission or vehicle emission.
province. The highest mass level of Cl (404.09 ± 154.14 ng/m3 ) was found
in northwest samples, and lowest (60.81 ± 25.86 ng/m3 ) in win-
ter samples. However, Zhou et al. (2016) reported that Cl ion was
2.4. Chemical compositions of PM2.5 higher in winter in Shanghai atmosphere. Chlorine in air can be
from sea salt and anthropogenic emission. We noted that Na was
Twenty-one elements (Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, near 10 times higher than that of Cl, and especially, no relationship
Mn, Fe, Ni, Cu, Zn, Br, Sr, Ba, Pb) in PM2.5 were investigated in this between Na and Cl could be found. Therefore, mass level of Cl in
study. The mean concentrations of the chemical elements are listed Shanghai air not only contributed by sea salt, but also by anthro-
in the Table 3. pogenic emission, such as industrial emission.
X-ray fluorescence spectrometer (XRF) data showed that to- The seasonal variabilities of analyzed trace elements were listed
tal mass concentrations of analyzed chemical elements in Shang- in Table 3. Average concentrations of Ti, V, Mn, Fe, Cu, Zn, Sr, and
hai PM2.5 collected in spring, summer, autumn and winter were Pb in the four seasons were lower than that in previous studies
3149.71, 12483.05, 16325.44, and 4479.63 ng/m3 , respectively. It is (Wang et al., 2013; Ming et al., 2017).
note that average mass concentration of PM2.5 in winter was the The highest average mass level of trace elements in the PM2.5
highest (83.36 ± 18.56 μg/m3 ) among the four seasonal samples, was Zn, its annual variable trends followed the pattern, winter
however, mass level of the total chemical elements in winter sam- (142.93 ± 55.80 ng/m3 ) > northwest (104.65 ± 64.57 ng/m3 ) >
ples was relative lower than that of summer and autumn samples. spring (79.03 ± 11.53 ng/m3 ) > autumn (73.78 ± 22.27 ng/m3 ) >
The main reason might be there were lots of unidentified fractions Summer (31.18 ± 24.65 ng/m3 ). High mass concentration of Zn
by XRF, such as Si, C, in the PM2.5 . in ambient particles has also been reported in the atmosphere
Among the analyzed elements, calcium was the most abundant Shanghai (Lu et al., 2008, 2011). Mass level of other heavy
element. It’s highest mass concentration (6475.33 ± 797.78 ng/m3 ) metal elements in the fine particles ranked in the following or-
was found in the samples came from Northwest in winter, and the der: Mn (32.60 ± 17.58 ng/m3 ) > Pb (27.04 ± 18.64 ng/m3 )
lowest (385.87 ± 299.66 ng/m3 ) was in spring samples. The high
JID: JES
> Ba (23.09 ± 8.58 ng/m3 ) > Cr (14.14 ± 8.45 ng/m3 ) > joint effect of local to regional precursor emissions and super-
Cu (12.22 ± 7.02 ng/m3 ) > Sr (11.85 ± 6.14 ng/m3 ) > Ni regional O3 transportation.
(7.47 ± 2.82 ng/m3 ) > V (4.90 ± 3.60 ng/m3 ). V and Ni were
probably associated to fossil fuel combustion including ships and 3. Conclusions
power plants (Zhang et al., 2019a, 2019b). Barium was most abun-
dant in gasoline exhaust PM2.5 samples (Cui et al., 2019). Cu and The annual mean PM2.5 mass concentration in Shanghai has de-
Zn were the markers for traffic emission sources. These elements creased under implementation of Chinese Clean Air Action Plan
were found in PMs because of tire wear, lubricating oil combus- since 2013. However, O3 pollution has been continuously worsen-
tion, and vehicle brake abrasion (Liu et al., 2015). Mn, Cu, Zn, Pb ing and has become the primary air pollutant affecting the ambi-
came from industrial emission (Xue et al., 2019; Shridhar et al., ent air quality instead of PM2.5 in Shanghai. Better understanding
2010). The mass concentrations of Mn, Cu, Zn, and Pb were the of the elevated O3 and its potential sources is important for mak-
highest in winter were contributed by the PM2.5 carried by the air ing policy to improve air quality. Based long term field observation
mass from northern China. The air mass covered many industrial of PM2.5 and O3 , our main conclusions are summarized as follows:
parks in northern China. The low wind speeds (average speed was (1) Our data showed that the annual average concentration
1.2 m/sec during the sampling time) and low temperatures in win- of PM2.5 and average O3 -8 hr was 39.35 ± 35.74 and
ter might lead to air poor dispersion, and increased the concen- 86.49 ± 41.65 μg/m3 respectively. The concentration of PM2.5
trations of these species in winter PM2.5 samples (Ancelet et al., showed clear seasonal trends, with higher concentrations in
2014). autumn/winter and lower concentrations in spring/summer,
As the dominant species of the transition metals in the at- however, the seasonal trends of O3 was different.
mosphere (Zhang et al., 2014), mass concentration of iron (Fe) (2) Air mass backward trajectory, PSCF and CWT model anal-
in ambient particles was higher than that of other trace met- yses implied that areas of northwestern China contributed
als. Fe in Shanghai ambient particles could be from natural significantly to the mass concentration of Shanghai PM2.5,
sources (such as fugitive soil, dust storm) and anthropogenic while pollutants from areas of eastern coastal provinces and
sources (such as industrial process, coal combustion emissions). South China Sea contributed significantly to the mass level
Its annual variety character followed the pattern, northwest of ozone.
(572.94 ± 282.96 ng/m3 ) > winter (456.56 ± 131.70 ng/m3 ) > au- (3) Ozone concentrations had good correlation with Ba
tumn (417.05 ± 128.67 ng/m3 ) > spring (350.37 ± 31.44 ng/m3 ) > (r = 0.64, p < 0.05) and V (r = 0.30, p > 0.05) suggesting
summer (310.55 ± 93.14 ng/m3 ). Compared with the previous re- the O3 precursor (VOCs) from traffic emissions contributed
ports (Lu et al., 2008; Wang et al., 2013), mass level of Fe in Shang- to the increasing O3 level in Shanghai atmosphere.
hai PM2.5 has decreased greatly, and seasonal variety of Fe in the
PM2.5 had no significantly different. Declaration of competing interest
2.5. Correlation between ozone level and the concentration of We declare that we do not have any commercial or associative
chemical elements interest that represents a conflict of interest in connection with the
work submitted.
Our O3 monitoring data showed O3 high concentration was
found in spring and summer, and low concentration in autumn Acknowledgments
and winter. The average concentration of ozone in summer, au-
tumn, winter, and spring was 85.06 ± 48.55, 62.19 ± 18.98, The National Natural Science Foundation of China (No.
53.44 ± 20.64, and 108.77 ± 17.07 μg/m3 respectively. It was 21477073) for supporting us to conduct this research.
worth noting that average mass level of O3 from northwest sam-
ples (42.29 ± 17.04 μg/m3 ) was lower than that of seasonal sam- Appendix A. Supplementary data
ples. The relationship between the ozone concentration and the
chemical composition level was calculated and listed in Table 3. Supplementary material associated with this article can be
The relationship data showed that the ozone concentrations found in the online version at doi:10.1016/j.jes.2020.03.043.
had positive correlation with Ba (r = 0.64, p < 0.05) and V
(r = 0.30, p > 0.05), and negative correlation with other mea- References
sured chemical elements. Ba was regarded from traffic road emis-
Ancelet, T., Davy, P.K., Trompetter, W.J., Markwitz, A., 2014. Sources of particulate
sion (Cui et al., 2019; Luo et al., 2018), and V was associated to matter pollution in a small New Zealand city. Atmos. Pollut. Res. 5 (4), 572–580.
fossil fuel combustion (Rodríguez et al., 2004). The ozone near the Ashbaugh, L., Malm, W.C., Sadeh, W.Z., 1985. A residence time probability analysis
ground surface was mainly formed by some complex chemical re- of sulfur concentrations at grand Canyon National Park. Atmos. Environ. 19 (8),
1263–1270.
actions between nitrogen oxides (NOx ) and volatile organic com- Cui, X., Zhou, T., Shen, Y., Rong, Y., Zhang, Z., Liu, Y., et al., 2019. Different biological
pounds (VOCs) (Wang et al., 2017). Considering that the O3 pro- effects of PM2.5 from coal combustion, gasoline exhaust and urban ambient air
duction in Shanghai city was clearly under a VOC-limited regime, relate to the PAH/metal compositions. Environ. Toxicol. Pharmacol. 69, 120–128.
Dolker, T., Agrawal, M., 2019. Negative impacts of elevated ozone on dominant
in which the aromatics and alkenes played the dominant roles species of semi-natural grassland vegetation in Indo-Gangetic plain. Ecotox. En-
(Geng et al., 2008), therefore, our data indirectly demonstrated that viron. Saf. 182, 109404.
O3 in Shanghai atmosphere was largely formed from VOCs. Ad- Geng, F., Tie, X., Xu, J., Zhou, G., Peng, L., Gao, W., et al., 2008. Characterizations of
ozone, NOx, and VOCs measured in Shanghai, China. Atmos. Environ. 42 (29),
ditionally, our potential source model results showed that larger
6873–6883.
PSCF and CWT value was mainly found in Zhejiang province, Fujian Hsu, Y.K., Holsen, T.M., Hopke, P.K., 2003. Comparison of hybrid receptor models to
province, Guangdong province and South China Sea. Zhang et al. locate PCB sources in Chicago. Atmos. Environ. 37 (4), 545–562.
Li, H., Wang, D., Cui, L., Gao, Y., Huo, J., Wang, X., et al., 2019a. Characteristics of at-
(2019a, 2019b) reported that high O3 level could be found when
mospheric PM2.5 composition during the implementation of stringent pollution
the southeastern wind (from Hangzhou Bay, Zhejiang Province and control measures in shanghai for the 2016 G20 summit. Sci. Total Environ. 648,
East China Sea) was dominant, this cluster could be the potential 1121–1129.
exogenous sources of high O3 . Li, L., An, J., Huang, L., Yan, R., Huang, C., Yarwood, G., 2019b. Ozone source appor-
tionment over the Yangtze River Delta region, China: investigation of regional
Above all, combining the characterization of VOCs, Ba and V, transport, sectoral contributions and seasonal differences. Atmos. Environ. 202,
high O3 level in Shanghai atmosphere was mainly caused by the 269–280.
JID: JES
Liu, G., Li, J., Wu, D., Xu, H., 2015. Chemical composition and source apportionment Wang, Y., Shi, Z., Shen, F., Sun, J., Huang, L., Zhang, H., et al., 2019. Associations
of the ambient PM2.5 in Hangzhou. China. Particuology 18, 135–143. of daily mortality with short-term exposure to PM2.5 and its constituents in
Lu, S., Yao, Z., Chen, X., Wu, M., Sheng, G., Fu, J., et al., 2008. The relationship be- Shanghai, China. Chemosphere 233, 879–887.
tween physicochemical characterization and the potential toxicity of fine par- Xu, J., Tie, X., Gao, W., Lin, Y., Fu, Q., 2019. Measurement and model analyses of the
ticulates (PM2.5 ) in Shanghai atmosphere. Atmos. Environ. 42 (31), 7205–7214. ozone variation during 2006 to 2015 and its response to emission change in
Lu, S., Feng, M., Yao, Z., An, J., Zhong, Y., Wu, M., et al., 2011. Physicochemical char- megacity Shanghai. China Atmos. Chem. Phys. 19, 9017–9035.
acterization and cytotoxicity of ambient coarse, fine, and ultrafine particulate Xue, H., Liu, G., Zhang, H., Hu, R., Wang, X., 2019. Similarities and differences in
matters in Shanghai atmosphere. Atmos. Environ. 45 (3), 736–744. PM10 and PM2.5 concentrations, chemical compositions and sources in Hefei
Luo, Y., Zhou, X., Zhang, J., Xiao, Y., Wang, Z., Zhou, Y., et al., 2018. PM2.5 pollution in City. China Chemosphere 220, 760–765.
a petrochemical industry city of northern China: Seasonal variation and source Zhang, G., Bi, X., Lou, S., Li, L., Wang, H., Wang, X., et al., 2014. Source and mix-
apportionment. Atmos. Res. 212, 285–295. ing state of iron-containing particles in Shanghai by individual particle analysis.
Ming, L., Jin, L., Li, J., Fu, P., Yang, W., Liu, D., et al., 2017. PM2.5 in the Yangtze River Chemosphere 95, 9–16.
Delta, China: Chemical compositions, seasonal variations, and regional pollution Zhang, J., Chen, Q., Wang, Q., Ding, Z., Sun, H., Xu, Y., 2019a. The acute health effects
events. Environ. Pollut. 223, 200–212. of ozone and PM2.5 on daily cardiovascular disease mortality: a multi-center
Rodrıǵuez, S., Querol, X., Alastuey, A., Viana, M., Alarcón, M., Mantilla, E., et al., 2004. time series study in China. Ecotox. Environ. Safe. 174, 218–223.
Comparative PM10 –PM2.5 source contribution study at rural, urban and indus- Zhang, X., Lu, Y., Wang, Q., Qian, X., 2019b. A high-resolution inventory of air
trial sites during PM episodes in Eastern Spain. Sci. Total Environ. 328 (1-3), pollutant emissions from crop residue burning in China. Atmos. Environ. 213,
95–113. 207–214.
Shridhar, V., Khillare, P.S., Agarwal, T., Ray, S., 2010. Metallic species in ambient par- Zhao, M., Huang, Z., Qiao, T., Zhang, Y., Xiu, G., Yu, J., 2015. Chemical characteriza-
ticulate matter at rural and urban location of Delhi. J. Hazard. Mater. 175 (1-3), tion, the transport pathways and potential sources of PM2.5 in Shanghai: sea-
600–607. sonal variations. Atmos. Res. 158–159, 66–78.
Shu, L., Wang, T., Xie, M., Li, M., Zhao, M., Zhang, M., et al., 2019. Epsiode study Zhao, S., Yu, Y., Qin, D., Yin, D., Dong, L., He, J., 2019a. Analyses of regional pollution
of fine particle and ozone during the CAPUM-YRD over Yangtze River Delta of and transportation of PM2.5 and ozone in the city clusters of Sichuan Basin.
China: characteristics and source attribution. Atmos. Environ. 203, 87–101. China Atmos. Pollut. Res. 10 (2), 374–385.
Wang, J., Hu, Z., Chen, Y., Chen, Z., Xu, S., 2013. Contamination characteristics and Zhao, W., Tang, G., Yu, H., Yang, Y., Wang, Y., Wang, L., et al., 2019b. Evolution of
possible sources of PM10 and PM2.5 in different functional areas of Shanghai. boundary layer ozone in Shijiazhuang, a suburban site on the North China Plain.
China Atmos. Environ. 68, 221–229. J. Environ. Sci. 83, 152–160.
Wang, T., Xue, L., Brimblecombe, P., Lam, Y.F., Li, L., Zhang, L., 2017. Ozone pollu- Zhao, T., Markevych, L., Romanos, M., Nowak, D., Heinrich, J., 2018. Ambient ozone
tion in China: a review of concentrations, meteorological influences, chemical exposure and mental health: A systematic review of epidemiological studies.
precursors, and effects. Sci. Total Environ. 575, 1582–1596. Environ. Res. 165, 459–472.
Wang, Y., Zhang, X., Draxler, R.R., 2009. TrajStat: GIS-based software that uses var- Zhou, M., Qiao, L., Zhu, S., Li, L., Lou, S., Wang, H., et al., 2016. Chemical charac-
ious trajectory statistical analysis methods to identify potential sources from teristics of fine particles and their impact on visibility impairment in Shanghai
long-term air pollution measurement data. Environ. Modell. Softw. 24 (8), based on a 1-year period observation. J. Environ. Sci. 48, 151–160.
938–939.

Zeng 2020

Uploaded by

Copyright:

Available Formats

Zeng 2020

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Zeng 2020

Uploaded by

Copyright:

Available Formats

JID: JES

ARTICLE IN PRESS [m5G;May 11, 2020;19:45]

Contents lists available at ScienceDirect

Journal of Environmental Sciences

Relationships between chemical elements of PM2.5 and O3 in Shanghai

No. Summer Autumn/North Winter Spring Northwest

1 2018/06/26–2018/06/29 2018/11/07–2018/11/10 2019/01/11–2019/01/14 2019/03/17–2019/03/20 2018/12/02–2018/12/05

1 436 30.67 31.01 50 11.5% 103.68 25.66

Elements Summer Autumn/North Winter Spring Northwest r

Na 4296.77±241.18 2749.73±1152.02 355.93±249.62 580.04±233.43 3832.49±523.45 −0.26∗

You might also like