Science of the Total Environment 502 (2015) 641649

Contents lists available at ScienceDirect

Science of the Total Environment

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

The impacts of summer monsoons on the ozone budget of the

atmospheric boundary layer of the Asia-Pacic region
Xuewei Hou a,b, Bin Zhu a,b,, Dongdong Fei a, Dongdong Wang a,b
a
b

Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters, Nanjing University of Information Science & Technology, Nanjing 210044, China
Key Laboratory for Aerosol-Cloud-Precipitation of China Meteorological Administration, Nanjing University of Information Science and Technology, Nanjing 210044, China

H I G H L I G H T S
The Asia-Pacic monsoon greatly affects O3 seasonal and inter-annual variations.
The differences of emissions and zonal winds lead to pollutants transition zone.
Advection plays a key role in the monsoon impact on O3 inter-annual variation.

a r t i c l e

i n f o

Article history:
Received 31 May 2014
Received in revised form 10 September 2014
Accepted 22 September 2014
Available online 8 October 2014
Editor: Xuexi Tie
Keywords:
Tropospheric ozone
Transition zone
Asia-Pacic monsoon

a b s t r a c t
The seasonal and inter-annual variations of ozone (O3) in the atmospheric boundary layer of the Asia-Pacic
Ocean were investigated using model simulations (20012007) from the Model of Ozone and Related chemical
Tracers, version 4 (MOZART-4). The simulated O3 and diagnostic precipitation are in good agreement with the
observations. Model results suggest that the Asia-Pacic monsoon signicantly inuences the seasonal and
inter-annual variations of ozone. The differences of anthropogenic emissions and zonal winds in meridional directions cause a pollutants' transition zone at approximately 2030N. The onset of summer monsoons with a
northward migration of the rain belt leads the transition zone to drift north, eventually causing a summer minimum of ozone to the north of 30N. In years with an early onset of summer monsoons, strong inows of clean
oceanic air lead to low ozone at polluted oceanic sites near the continent, while strong outows from the continent exist, resulting in high levels of O3 over remote portions of the Asia-Pacic Ocean. The reverse is true in years
when the summer monsoon onset is late.
2014 Elsevier B.V. All rights reserved.

1. Introduction
Tropospheric ozone is an important greenhouse gas, pollutant, and
source of OH radicals. Its contribution to global warming from the preindustrial era to the present is regarded as the third most important, following those of carbon dioxide and methane (Solomon et al., 2007). The
level of tropospheric ozone also affects human health and natural ecosystems. Elucidation of the processes determining spatial and temporal
variations in tropospheric ozone is important for evaluating the effect of
ozone on regional air quality and climate change.
Monsoons are a seasonal variation of wind, particular of wind directions, which result from the variations of meridional differences in solar
radiation and the thermal difference between the land and sea. It is an
important element of the global climate system. Asian monsoons are
Corresponding author at: Key Laboratory for Aerosol-Cloud-Precipitation of China
Meteorological Administration, Nanjing University of Information Science and
Technology, Nanjing 210044, China. Tel.: +86 2558699785.
E-mail address: binzhu@nuist.edu.cn (B. Zhu).

http://dx.doi.org/10.1016/j.scitotenv.2014.09.075
0048-9697/ 2014 Elsevier B.V. All rights reserved.

composed of three sub-systems, which are the tropical Indian monsoon,

the tropical Western North Pacic (WNP) monsoon, and the subtropical
East Asian (EA) monsoon (Zhu et al., 1986; Wang and Lin, 2002). Over
the Asia-Pacic region, low-level winds seasonally reverse from winter
easterlies to summer westerlies for the WNP monsoon and from winter
northerlies to southerlies for the EA monsoon. The EA and WNP monsoons are considered together as the Asia-Pacic monsoon hereafter
in this study. Whether considering the EA monsoon or the WNP monsoon, the transition of the winter monsoon to the summer monsoon results in a series of changes in weather, such as wind shifts, convection,
precipitation, and air temperature. Together these changes signicantly
affect the transport paths and photochemical production of pollutants.
The observed summer minimum of surface O3 over the Asia-Pacic region was attributed to the incursion of the monsoon, which transports
oceanic air with less background O3 to the region, causing lower O3 concentrations (Chan et al., 1998; Pochanart et al., 2002; Wang et al., 2006;
Yamaji et al., 2006; Zbinden et al., 2006; He et al., 2008). Tanimoto et al.
(2005) and He et al. (2008) identied a relationship between monsoons
and the O3 spring maximum at the surface. He et al. (2008) found that

642

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

the basic common features of O3 seasonal behaviors over the monsoon

region are the pre- and post-monsoon peaks with a summer trough;
these bimodal seasonal patterns become weaker or even disappear outside of the monsoon region. Kurokawa et al. (2009) investigated the inuence of meteorological variability on the interannual variation of the
springtime boundary layer ozone over Japan, and found some correlation between spring ozone over Japan and the El Nio-Southern Oscillation. Zhu (2012) determined that ozone seasonal cycle features were
mainly controlled by the seasonal transitions of Asia-Pacic monsoon
circulations by using monthly surface ozone data and related wind
eld and precipitation data.
This paper addresses the inuence of the Asia-Pacic monsoon on
seasonal and inter-annual variations of boundary layer ozone over the
Asia-Pacic region, based on an analysis of the ozone budget using the
chemical transport model (CTM), Model of Ozone and Related chemical
Tracers, version 4 (MOZART-4). A xed emission experiment was performed to identify the inuence of early/late summer monsoon onset
on the O3 budget. For the Asia-Pacic monsoon region, various indices
based on considerations of thermodynamics and dynamics from different aspects have already been dened. Each monsoon index pays attention to some specic physical processes and represents its own
meanings, independent of others. In this study, meridional wind (Lau
and Li, 1984; Chen et al., 1991) was used to reveal the seasonal march
of the summer monsoon, and a dynamical normalized seasonality monsoon index developed by Li and Zeng (2003) was used to distinguish the
early and late transition periods of the Asia-Pacic summer monsoon
onset. The importance of the impact on the ozone levels by advection,
convection, diffusion, and photochemistry over the Asia-Pacic monsoon region was evaluated.
2. Data and methods
2.1. Observed data
Observations of daily O3 at seven regional stations were taken from
the Acid Deposition Monitoring Network in East Asia (EANET): Rishiri,
Tappi, Sado-Seki, Oki, Yusuhara, Hedo, and Ogasawara. Descriptions of
these sites can be found at http://www.eanet.asia/site/index.html. Observations from two additional sites, Yonagunijima and Minamitorishima,
were obtained from the WMO-World Data Centre for Greenhouse Gases
(WDCGG) (http://ds.data.jma.go.jp/gmd/wdcgg/). The site location information including latitude, longitude and elevation height above sea level
for all sites is provided in Fig. 1. Daily precipitation data collected from
2001 to 2007 are taken from the Global Precipitation Climatology Project
(GPCP) with a resolution of 1 1 (http://precip.gsfc.nasa.gov/). The
GPCP dataset reects the spatial and temporal distributions of precipitation very well, as a combination of GPCC (Global Precipitation Climatology
Center) ground observations from precipitation gauges and a precipitation inversion from satellite remote sensing observations.
2.2. Model setup
A detailed description and evaluation of the standard version of
MOZART-4, and the upgrade over MOZART-2 (Horowitz et al., 2003),
is given by Emmons et al. (2010). MOZART-4 includes an updated chemical scheme of hydrocarbon and bulk aerosols (Tie et al., 2001, 2005),
and improved emissions compared to MOZART-2. It does not include
explicit stratospheric chemistry, but constrains the climatological
mixing ratios of ozone and other species in the stratosphere. The
Synoz (synthetic ozone) scheme (McLinden et al., 2000) is used as a
ux upper boundary condition for ozone in the stratosphere and yields
a cross-tropopause ozone ux of 500 Tg/yr. More details about the
physical and chemical mechanisms were discussed in Emmons et al.
(2010). In this study, MOZART-4 is run with the standard chemical
mechanism [see Emmons et al., 2010 for details], with online calculation
of dry deposition. It is driven by meteorological parameters from the

NCAR reanalysis of the National Centers for Environmental Prediction

(NCEP) forecasts. The output from the model run is available at a temporal resolution of 6 h, a horizontal spatial resolution of approximately 2.8
2.8, and 28 hybrid levels in the vertical. The top model level is located
at approximately 2 hPa. The initial condition and emissions are based on
the NCAR Community Data Portal (http://cdp.ucar.edu/). The model
was run in time steps of 20 min from June 2000 to December 2007,
with the rst seven months used to spin up. The experiments used
xed emissions from 2001 and the meteorological parameters were varied for each year over the simulations, with the rst seven months as
spin-up time. Note that the modeled results are based on the mean
values in the atmospheric boundary layer (the six lowermost layers in
the model, surface to ~ 2 km) except for the model validations in
Section 3.1 and the special denitions in Section 3.2.
2.3. Monsoon index and ozone budget
Previous studies showed that the East Asian monsoon is characterized as a seasonal reversal of the lower-troposphere meridional wind
direction (Lau and Li, 1984; Chen et al., 1991), and the 850 hPa wind
is used to reect the low-level atmospheric circulation (Lu and Chan,
1999). Therefore, the daily meridional wind at 850 hPa is used to
show the seasonal march of the Asia-Pacic monsoon in this study. A
pre-dened monsoon index (MI) can provide a useful insight into quantitatively examining the strength and variation of monsoon circulation
in monsoon regions. For the Asia-Pacic monsoon region, many studies
have already dened various indices based on considerations of dynamics and thermodynamics from different aspects, such as wind eld, precipitation, difference of oceanland temperature, and so on. Each MI is
designed to capture specic physical processes. In this study, we used
a dynamically normalized seasonality MI developed by Li and Zeng
(2003) to examine the wind eld to investigate the inuence of the
Asia-Pacic monsoon on the inter-annual variability of ozone.
The factors controlling ozone levels in the atmospheric boundary
layer (transport, net photochemical production, and deposition) are
discussed in this study. Each term affecting the ozone budget is evaluated quantitatively at 850 hPa in Section 3.2 and in a column from the surface to approximately 2 km (the atmospheric boundary layer) in
Section 3.3. The rate of change of ozone can be expressed as:
dO3
Chem Adv Con Dif Dep
dt
where Chem represents the net chemical production; Adv, Con, and Dif
are the transport uxes associated with advection, convection and diffusion, respectively; Dep is the dry deposition rate. Fig. 4 shows each of
these components of the ozone budget for the study region: net chemistry, net transport ux (advection, convection, and diffusion), and O3S.
O3S is an O3 tracer that tags O3 transported from the stratosphere (Sudo
and Akimoto, 2007).
3. Results and discussion
3.1. Validation of simulated O3 and precipitation
The modeling system described in Section 2.2 has previously been
used for analyzing tropospheric O3 over the northern hemisphere
(Horowitz et al., 2003; Liu et al., 2005; Pster et al., 2008a,b; Emmons
et al., 2010), and, in these studies, the simulated results showed good
agreement with observations. Stevenson et al. (2006) suggested that
the tropospheric O3 budget in MOZART-4 is in good agreement with
the mean of 26 models, whereas the stratospheric input value has
been determined to be realistic by Wild (2007). In this section, we further evaluate the general performance of our modeling system for O3
over the Asia-Pacic region.

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

643

Fig. 1. Comparison of modeled O3 (ppbv, black line) to observation (red line) at selected sites over the Asia-Pacic rim region between 2001 and 2007. All of the data were calculated based
on the mean of seven years' data (similarly hereinafter). R is the correlation coefcient. The degrees of freedom are 363, and p values in the test are 0.05.

Fig. 1 shows a comparison of modeled daily O3 with observations

over the Asia-Pacic region. Seasonal variations with a spring maximum
and a summer minimum are fairly well reproduced by the model. Simulated results of the spring maximum at middle latitude sites (Rishiri,
Tappi, and Sado-Seki) are underestimated relative to observations,
while summer minima at all sites are overestimated. These discrepancies
may result from the coarse horizontal model resolution in meteorology
and pollutant sources due to sub-grid-scale processes that cannot be accurately simulated. Signicant effects of cloud and wet processes may
also be partly responsible for the discrepancies in summer ozone, as
discussed below. The correlation-coefcient (R) between simulated and
observed daily results ranges from approximately 0.47 to 0.87. They all
passed the t test under the 95% condence level. The relatively strong
correlations (an R of approximately 0.8) at four oceanic sites (Ogasawara,
Hedo, Yonagunijima, and Minamitorishima) are primarily attributed to
their remote locations with less impact from continental emissions.
Other sites at mid-latitude can be affected both by local to regional anthropogenic emissions and complex topography, which might account
for the less than satisfactory model performance at these locations.
The increase of precipitation intensity and the northward migration
of the rain belt are key features of the summer monsoon onset. Although
direct precipitation scavenging of ozone is relatively unimportant due to
its low solubility, precipitation can affect ozone chemical production by
altering solar radiation and through wet removal of soluble ozone precursors. Consequently, it is important to evaluate the simulated precipitation. Fig. 2 indicates that the diagnosis precipitation closely
represents the GPCP data in January. In April, July, and October, strong
precipitation over Japan and its northern regions in the GPCP data is
underestimated by the model. Because MOZART-4 is an ofine model,
the accuracy of simulated precipitation depends largely on the accuracy
of the meteorological elds, which may have some errors in mid- and
high-latitudes that lead to an underestimation of precipitation. On the
other hand, the coarse horizontal resolution in the global model may inaccurately diagnose sub-grid-scale processes, leading to these discrepancies. At relatively low latitudes, the diagnosis result represents the
GPCP precipitation very well.
3.2. The impact of the Asia-Pacic monsoon march on O3 seasonal variation
Wang and Lin (2002) demonstrated the onset time of the Asia-Pacic
summer monsoon varies across different regions based on the

precipitation analysis, with the 30th pentad (150th Julian day) reaching
2030N and the 40th pentad (200th Julian day) being north of 30N.
Accordingly, it is necessary to divide the Asia-Pacic monsoon region
to show the onset period. In this study, the monsoon regions covering
120140E longitude and 2030N and 3050N latitude were chosen, considering the spatial extent of the region affected by the monsoon
and the distribution of sites available for estimation. The meridional
wind at 850 hPa can be used to show the seasonal march of the Asia-Pacic monsoon, where positive meridional winds represent the prevailing
summer monsoon (Lau and Li, 1984; Chen et al., 1991; Lu and Chan,
1999). Fig. 3 shows that the sub-tropical summer monsoon onset is in
March at 2030N and is in early April to late May in 3050N. The retreat of the summer monsoon occurs earlier at 3050N than at 20
30N. The onset and retreat of the Asia-Pacic summer monsoon
match the results of Wang and Lin (2002) very well.
As shown in Fig. 4a and b, there is a transition zone of pollutants near
20N in the winter. South of 20N, pollutant concentrations of O3, CO,
NO are relatively low with a lesser impact of continental emissions.
North of 30N, pollutant concentrations are relatively high with a strong
inuence from anthropogenic emissions or transport. The difference in
anthropogenic emissions plays a key role in the formation of a transition
zone. In addition, the difference in prevailing winds is also largely responsible for the formation of a transition zone. South of 20N, where
the tropical monsoon resides, easterly winds prevail in the winter.
North of 20N, where the subtropical monsoon is located, northwesterly
winds prevail in the winter. Because of the differences in zonal wind and
the weak meridional wind, a shear line is located at approximately
20N, which limits the air masses' transport and exchange between
north and south. With the onset of summer monsoons in late March
at 20N, southerly winds prevail, causing the transition zone to move
northward. In summer, the transition zone continuously moves northward with the advance of the summer monsoon. In autumn, the summer monsoon retreats. Southerly winds become weak, and the zonal
wind component increases. The Northsouth differences in the zonal
wind component decrease, causing the transition zone to retreat back
to 20N. Eventually, a transition zone of pollutants moves back and
forth in a south to north direction with the advance of the summer
monsoon.
Tropospheric ozone, as a secondary pollutant, is affected by solar radiation and the emission of precursor species. The transition from the
winter monsoon to the summer monsoon results in a series of changes

644

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

Fig. 2. Comparison of monthly mean precipitation (shaded) to GPCP data (contour) over the Asia-Pacic rim region in January, April, July and October during 20012007, unit: mm/d.

in weather, such as atmospheric circulations and precipitation, which

signicantly affect the pollutants' transport paths and ozone photochemical production. South of 20N, the precursors of ozone are very
low, especially NO, which is as low as 4 pptv, resulting in a net photochemical loss of ozone (see Fig. 4b and c). Precipitation increases
south of 20N in early May (Fig. 4c), slowing down the photochemical
reactions there. In summer, southerly winds prevail and transport low
latitude clean air deeply towards higher latitudes at approximately
45N. Therefore, the O3 seasonal pattern with maximum pollutant levels
in the winter or early spring and a broad minimum in the summer is
present south of 20N.

Fig. 3. The mean meridional wind at 850 hPa (blue points) between 120 and 140E and
different latitude regions (a, 3050N; b, 2030N). The solid black line represents the
wind speed at 0 m/s. The solid red line is a Gaussian tting curve. The green dashed line
represents the 95% prediction bands.

In comparison with that of south 20N, the concentration of NO is

high in the winter over north of 30N, but the net chemical production
of ozone is low under the winter weak solar radiation. In spring, net
ozone chemical production increases with the enhancement of solar radiation, leading to a maximum ozone concentration at 40N in May. Due
to the later onset of the summer monsoons in the regions between 30
and 50N, westerly winds prevail until late June. High precursor concentrations, transported from upwind regions or locally emitted, lead to the
accumulation of ozone through strong chemical production under
strong solar radiation. In summer, cloudiness and precipitation are enhanced at 3050N. The concentrations of ozone precursors decrease,
resulting from the inow of clean marine air masses by prevailing
southerly winds and the wet removal by the rain belt. Net ozone production is reduced by decreased solar radiation and ozone precursors.
The net ozone transport ux during this time is outow, but the outow
is weaker than in the spring. The overall result of the factors discussed
above is a decrease in net chemical production, eventually leading to
an ozone minimum in summer. The direct transport of ozone by wind
has a relatively weak impact on the summer minimum due to the
weak outow or inow. The shift of south winds to westerly winds in
early September reects the relatively early retreat of summer monsoons at 3050N. High concentrations of precursors enhance the net
ozone chemical production. Similarly as in spring, the ozone concentration is higher in autumn than in summer.
The contribution of stratospheric ozone (Fig. 4d) includes a winter
maximum and a summer minimum. Its seasonal variation has an inuence on ozone concentrations, especially south of 20N where the contribution of stratospheric ozone is approximately 8 ppbv in winter, or
approximately 30%.

3.3. The impact of the Asia-Pacic summer monsoon on inter-annual

variation of O3
This section focuses on the inuence of the early/late onset of the
Asia-Pacic sub-tropical summer monsoon on the budget of O3 in the

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

645

(a)

(b)

(c)

(d)

Fig. 4. The zonal mean ozone mixing ratio (ppbv, shaded), wind (m/s, vector), zonal wind speed (m/s, contour) (a); CO (ppbv, shaded), NO (pptv, contour) and wind (m/s, vector) (b);
precipitation (mm/d, contour) and the O3 net chemical production (ppbv/d, shaded) (c); the contribution of stratospheric O3 (ppbv, contour) and the net O3 transport ux (ppbv/d, shaded) (d) at 850 hPa between 1 and 65N. The black solid line is the 40 ppbv of O3 in (a). The dashed lines are the latitudes of 20N, 30N, and 50N.

atmospheric boundary layer over the low-latitude oceanic stations,

where model results were well-correlated with observations (see
Section 3.1). The correlation coefcients of simulated and observed O3
anomalies are 0.56, 0.36, 0.5, and 0.52 at Hedo, Ogasawara,
Yonagunijima, and Minamitorishima, respectively. As shown in Fig. 5,
most of the simulated and observed anomalies of O3 concentration are
signicant at the 95% condence level, and they all pass the test at the
95% condence level. The model results can be used to analyze the

inter-annual variation of ozone in the atmospheric boundary layer.

Ogasawara is located in an area with a strong gradient of wind speeds
and direction, a region that is most sensitive to the monsoon transition
(seen Fig. 8). It is difcult to capture the ozone inter-annual variations
accurately in the global model, leading to the relatively low correlation
coefcient of O3 at Ogasawara.
We calculated the dynamically normalized seasonality monsoon
index dened by Li and Zeng (2003) over the region 2050N and

646

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

Fig. 5. Scatter plots showing the relationship between observed and simulated daily surface ozone anomalies at Hedo, Ogasawara, Yonagunijima and Minamitorishima from 2001 to 2007
(OB and MZ4 represent observations and simulations, respectively). The simulated ozone anomaly is the simulation for the xed emissions. The solid red line is the linear regression equation. The green dashed line is the 95% condence level. N is the number of paired samples. R is the correlation coefcient.

120140E using NCEP reanalysis data, to obtain an Asia-Pacic monsoon index (AMI) and its anomaly (shown in Fig. 6). From 2001 to
2007, AMI is negative in the winter and positive in the summer, reasonably representing the winter and summer monsoons. In March and
April, a negative AMI value implies that the monsoon is still in its winter
phase. In May, it has transitioned to its summer phase, on average. The
AMI and its anomaly are positive in May 2003, 2004 and 2006, indicating the early onset of Asia-Pacic summer monsoon in these years. In
April 2003, the AMI is near zero and has a high positive anomaly implying a much earlier transition of the winter monsoon into the summer
monsoon. To unify the transition period, we chose May in 2004 and
2006 as the early transition periods of the winter monsoon to the summer monsoon (strong monsoon period) and May in 2002 and 2007 as
the late transition periods for the negative anomalies (weak monsoon
period) to discuss the inuence of the summer monsoon on O3 over
the Asia-Pacic Ocean.
Budget analysis from Fig. 7 shows the net chemistry is
0.03 ppbv/d in Hedo and 0.32 ppbv/d in Yonagunijima, while it
is 2.1 ppbv/d in Ogasawara and 2.8 ppbv/d in Minamitorishima.
The absolute value of net chemistry in polluted oceanic sites (Hedo

Fig. 6. The Asia monsoon index (AMI, blue line) and the anomaly of AMI (red bar). The bar
marked with the green line is the anomaly of AMI in March, April, and May, every year.

and Yonagunijima) is lower than in remote oceanic sites (Ogasawara

and Minamitorishima) due to stronger chemical production in polluted oceanic regions. In addition, according to the division of AsiaPacic monsoon regions by Wang and Lin (2002), Hedo and
Yonagunijima are located in the subtropical East Asia (EA) monsoon
region, Ogasawara is at the boundary of the EA and western North
Pacic (WNP) monsoon regions, and Minamitorishima is in the tropical WNP monsoon region. Fig. 7 shows O3 during early transition
periods is lower than average in Yonagunijima and Hedo, and higher
than the mean at Ogasawara and Minamitorishima. In late transition
periods the effect is opposite, except in Ogasawara due to its location
on the boundary of the EA and WNP monsoon regions (Wang and
Lin, 2002) with sharp gradients of wind speed and direction.
As the star reveals in Fig. 7, the contribution of stratospheric O3 is
3.5 ppbv at Hedo in early transition periods, lower by 0.1 ppbv than
the mean value, which cannot account for the decreased 3 ppbv of O3
from the average. In late transition periods, the contribution is higher
by 1.1 ppbv than the mean at Hedo, which does not lead to an O3 increase of 4 ppbv. We can also nd similar facts at Ogasawara,
Yonagunijima, and Minamitorishima. Accordingly, ozone intruded
from the stratosphere has an inuence on the inter-annual variation
of the boundary layer O3 in the spring, but it is not the only responsible
factor.
The controlling sources of ozone in polluted oceanic sites are different from that in remote oceanic sites. As the mean values show in Fig. 7,
controlling sources are convection and diffusion in Hedo, chemistry and
diffusion in Yonagunijima, and convection and advection in Ogasawara
and Minamitorishima. Chemistry and diffusion have relatively low correlation with the variation of the monsoon transition period, as well as
dry deposition. The anomalies of convection are relatively large in polluted oceanic sites. They are correlated with O3 anomalies in
Yonagunijima while they are anti-correlated in Hedo. Therefore, convection may have some contribution to the O3 interannual variability.
The anomalies of advection are relatively large, and positively correlated
with O3 anomalies in all oceanic sites. In polluted oceanic sites,

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

647

Fig. 7. Boundary layer O3 mixing ratios (unit: ppbv), O3S (unit: ppbv) and the daily rates of change in O3 mixing ratios (unit: ppbv/d) at four oceanic sites in May in the multi-year average
(mean), early transition period (early), and late transition period (late). Black dots are O3. Stars are O3S. The values marked on the color bars are the daily rates of change in O3 from individual processes (values lower than 0.1 are not marked). The value on the black box indicates the inuence of the stratospheric tracer (O3S).

advection is O3 outow from the regions, 0.4 ppbv/d. The outow is

strong in the early transition period, about 1 ppbv/d, and weak in
the late transition period, 0.2 ppbv/d. Advection is O3 inow in remote oceanic sites, approximately 1.2 ppbv/d. It is also strong in the
early transition period, approximately 2 ppbv/d, and weak in the late
transition period, approximately 1 ppbv/d. We nd that advection
clearly plays a key role in the impact of the summer monsoon on the
inter-annual variation of the O3 boundary layer in May (spring). Net
chemistry, convection, diffusion, and dry deposition may disturb the
impact.
Winds and their anomalies were analyzed to investigate O3 advection
(export and import) in Fig. 8. As shown in Fig. 8a, the average prevailing
wind is southwesterly in Yonagunijima and Hedo and is southeasterly
in Minamitorishima. This suggests that Yonagunijima and Hedo are
inuenced by continental outow of O3 and O3 precursors, and
Minamitorishima is inuenced by the transport of pollutants around the
Pacic Ocean under the effect of a subtropical high. In the early transition
period, the mean wind eld (Fig. 8a) and wind anomalies (Fig. 8b) in
Yonagunijima and Hedo show that the westerly components and southerly components are both stronger than the corresponding components
of the mean wind eld, resulting in strong net inows of clean oceanic
air in Yonagunijima and Hedo. In Minamitorishima, the northeasterly
wind anomalies indicate the weak southerly components and strong easterly components, making the transport around the Pacic Ocean strong.
Thus the inow of O3 in remote oceanic sites is strong. There are negative
O3 anomalies in polluted oceanic sites, and positive O3 anomalies in remote oceanic sites. In late transition periods, as shown in Fig. 8c, there is
a weak westerly component of winds on the coast of East China and
Japan, due to the northeastern wind anomalies, causing a weak inow
of clean oceanic air through advection, leading to the high O3 value in
Yonagunijima and Hedo. However, polluted air masses from the continent tend to be transported to higher latitudes by the strong southerly
component of the winds east of 160E and north of 30N (Fig. 8c). As a

result, O3 inow is weak in Minamitorishima, leading to low O3 values

in the late transition period. Ogasawara is located at the boundary of the
wind anomaly transition, making the effects of the monsoon transition
on local ozone concentrations apparent.
4. Conclusions
This study investigated the seasonal and inter-annual variations of O3
in the atmospheric boundary layer over the Asia-Pacic region between
2001 and 2007 using multi-year simulations by MOZART-4. The inuence
of the Asia-Pacic monsoon march on ozone seasonal variations was introduced by an analysis of the ozone budget. The relationship between
the Asia-Pacic summer monsoon onset and the inter-annual variation
of O3 was also examined by a seven-year numerical experiment with
xed emissions from 2001 and varied meteorology (from NCEP data)
for 20012007 over the Asia-Pacic Ocean. The daily meridional wind
at 850 hPa was used to show the seasonal march of the Asia-Pacic summer monsoon. A dynamically normalized seasonality MI developed by Li
and Zeng (2003) on the basis of wind eld calculations was used to distinguish the early and late transition periods of the summer monsoon onset.
The main ndings of the study are summarized as follows:
1. The model simulation reasonably reproduced both seasonal and interannual variations of O3, particularly at relatively low latitude oceanic sites where correlation coefcients between modeled and
observed values reached 0.87. The modeled results also represented
GPCP precipitation patterns for the region.
2. The Asia-Pacic monsoon has a great inuence on the seasonal and
inter-annual variations of O3 over the Asia-Pacic Ocean. Differences
of zonal wind in the meridional direction create a transition zone
near 20N between high and low pollutant regions in the winter.
The onset of the summer monsoon with the northward migration
of the rain belt helps move the transition zone northward.

648

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

(a) May

(b) early

(c) late

Fig. 8. The multi-year average of boundary layer (BL) O3 mixing ratios (ppbv) and wind led (m/s) in May from 2001 to 2007 (a), and the anomalies of O3 (ppbv) and wind (m/s) in the
early transition period (b) and in the late transition period (c). The large arrows overlaid on the gure are the wind direction, showing the large scale circulation over the north Pacic.

3. North of 30N, the transport of clean marine air masses by prevailing

southerly winds and the northward migration of the rain belt upon
the onset of the summer monsoon result in low concentrations of
ozone precursors in summer. This makes net O3 chemical production
weak and results in a summer ozone minimum. The direct transport
of ozone by wind has a relatively weak impact on the summer
minimum.
4. In the early transition of the Asia-Pacic winter monsoon to the
summer monsoon phase, strong inows of clean oceanic air tend
to lead to lower levels of O3 than average at polluted oceanic
sites near the continent. At the same time, stronger outows of
continental air masses result in higher levels of O3 at remote oceanic sites. The reverse is true in years when the summer monsoon
onset is late. Advection appears to play a key role in the impact of
the summer monsoon onset on the inter-annual variation of O3 in
spring.

Acknowledgments
This work was supported by the National Natural Science Foundation of China (40875078), the European Union Seventh Framework Program ([FP7/20072013]) under grant agreement no. 606719 (PANDA
project), the Natural Science Foundation of the Higher Education Institutions of Jiangsu Province, China (12KJA170003), the project of Jiangsu
provincial 333 and Six Talent Peaks, and the Program for Postgraduates Research Innovation of Jiangsu Higher Education Institutions
(grant no. CXZZ13_0509). We would like to acknowledge the entire
staff of the EANET and the WDCGG. Special thanks go to Stacy Walters
and Louisa Emmons of NCAR who provided MOZART-4 source codes
and helped us install and run it. We thank Jeff Collett at Colorado State
University in America and Tianliang Zhao at Nanjing University of
Information science & Technology in China for their valuable suggestions on improving the manuscript.

X. Hou et al. / Science of the Total Environment 502 (2015) 641649

References
Chan LY, Liu HY, Lam KS, Wang T. Analysis of the seasonal behavior of tropospheric ozone
at Hong Kong. Atmos Environ 1998;32(2):15968.
Chen LX, Zhu QG, Luo HB. East Asian Monsoons (in Chinese). Beijing: Chinese Meteorological Press; 1991. p. 1362.
Emmons LK, Walters S, Hess PG, Lamarque J-F, Pster GG, Fillmore D, et al. Description
and evaluation of the Model for Ozone and Related chemical Tracers, version 4 (MOZART-4). Geosci Model Dev 2010;3(1):4367.
He YJ, Uno I, Wang ZF, Pochanart P, Li J, Akimoto H. Signicant impact of the East Asia
monsoon on ozone seasonal behavior in the boundary layer of Eastern China and
the west Pacic region. Atmos Chem Phys 2008;8:754355.
Horowitz LW, Walters S, Mauzerall DL, Emmons LK, Rasch PJ, Granier C, et al. A global
simulation of tropospheric ozone and related tracers: description and evaluation of
MOZART, version 2. J Geophys Res 2003;108:4784. http://dx.doi.org/10.1029/
2002JD002853. [D2].
Kurokawa J, Ohara T, Uno I. Inuence of meteorological variability on interannual variations of springtime boundary layer ozone over Japan during 19812005. Atmos
Chem Phys 2009;9:6287304.
Lau KM, Li MT. The monsoon of East Asia and its global associations a survey. Bull Am
Meteorol Soc 1984;65:11425.
Li Jianping, Zeng Qingcun. A new monsoon index and the geographical distribution of the
global monsoons. Adv Atmos Sci 2003;20(2):299302.
Liu Qianxia, Zhang Meigen, Wang Bin. Simulation of tropospheric ozone with MOZART-2:
an evaluation study over East Asia. Adv Atmos Sci 2005;22(4):58594.
Lu E, Chan JC. A unied monsoon index for South China. J Clim 1999;12:237585.
McLinden CA, Olsen SC, Hannegan B, Wild O, Prather MJ. Stratospheric ozone in 3-D
models: a simple chemistry and the cross-tropopause ux. J Geophys Res 2000;
105(14):65314,665.
Pster GG, Emmons LK, Hess PG, Lamarque J-F, Thompson AM, Yorks JE. Analysis of the
Summer 2004 ozone budget over the United States using Intercontinental Transport
Experiment Ozonesonde Network Study (IONS) observations and Model of Ozone
and Related Tracers (MOZART-4) simulations. J Geophys Res 2008a;113:D23306.
http://dx.doi.org/10.1029/2008JD010190.
Pster GG, Emmons LK, Hess PG, Lamarque J-F, Orlando JJ, Walters S, et al. Contribution of
isoprene to chemical budgets: a model tracer study with the NCAR CTM MOZART-4. J
Geophys Res 2008b;113:D05308. http://dx.doi.org/10.1029/2007JD008948.
Pochanart P, Akimoto H, Kinjo Y, Tanimoto H. Surface ozone at four remote inland sites
and the preliminary assessment of the exceedances of its critical level in Japan.
Atmos Environ 2002;36:423550.

649

Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, et al. Intergovernmental Panel on Climate Change: Climate Change 2007: The Physical Science
Basis. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change. New York: Cambridge Univ.
Press; 2007.
Stevenson DS, Dentener FJ, Schultz MG, Ellingsen K, van Noije TPC, Wild O, et al.
Multimodel ensemble simulations of present-day and near-future tropospheric ozone. J Geophys Res 2006;111:D08301. http://dx.doi.org/10.1029/
2005JD006338.
Sudo K, Akimoto H. Global source attribution of tropospheric ozone: long-range transport
from various source regions. J Geophys Res 2007;112(D12302). http://dx.doi.org/10.
1029/2006JD007992.
Tanimoto H, Sawa Y, Matsueda H, Matsueda H, Uno I, Ohara T, et al. Signicant latitudinal
gradient in the surface ozone spring maximum over East Asia. Geophys Res Lett
2005;32(L21805). http://dx.doi.org/10.1029/2005GL023514.
Tie X, Brasseur G, Emmons L, Horowitz L, Kinnision D. Effects of aerosols on tropospheric
oxidants: a global model study. J Geophys Res 2001;106:293164.
Tie X, Madronich S, Walters S, Edwards D, Ginoux P, Mahowald N, et al. Assessment of the
global impact of aerosols on tropospheric oxidants. J Geophys Res 2005;
110(D03204). http://dx.doi.org/10.1029/2004JD005359.
Wang Bin, Lin Ho. Rainy season of the Asian-Pacic summer monsoon. J Clim 2002;15:
38698.
Wang HX, Zhou LJ, Tang XY. Ozone concentrations in rural regions of the Yangtze Delta in
China. J Atmos Chem 2006;54(3):25565.
Wild O. Modeling the global tropospheric ozone budget: exploring the variability in current models. Atmos Chem Phys 2007;7:264360.
Yamaji K, Ohara T, Uno I, Tanimoto H, Kurokawa J, Akimoto H. Analysis of the seasonal
variation of ozone in the boundary layer in East Asia using the Community Multiscale Air Quality model: what controls surface ozone levels over Japan? Atmos Environ 2006;40(10):185668.
Zbinden RM, Cammas J-P, Thouret V, Ndlec P, Karcher F, Simon P. Mid-latitude tropospheric ozone columns from the MOZAIC program: climatology and interannual variability. Atmos Chem Phys 2006;6:105373.
Zhu Bin. Seasonal and interannual variation features of surface ozone over East Asia-Pacic
region and its relation with East Asian monsoon. Trans Atmos Sci 2012;35(5):51323.
(in Chinese).
Zhu Qiangen, He Jinhai, Wang Panxing. A study of circulation differences between East
Asian and Indian summer monsoons with their interactions. Adv Atmos Sci 1986;
3(4):46677.