A Case Study on the Impact of East Asian Summer Monsoon on Surface O₃ in China

Abstract

The East Asian summer monsoon (EASM) was extremely strong in 2018, which substantially affected surface ozone (O₃) in China. Taking 2018 and the average synthesis of 2003 and 2010 to represent the strong and weak EASM cases, respectively, GEOS-Chem with constant anthropogenic emission was employed to investigate the impact of the EASM on surface O₃ in the east of China. Simulations show that surface O₃ decreased in the northeast and the eastern coast of China and increased in most of the remaining regions during strong EASM. The difference in surface O₃ between strong and weak EASM was around −15~7 ppbv. After analyzing relevant meteorological fields, it is found that the decrease in northeast China was mainly attributed to the large increase in vertical upward transport. The considerable decrease in the Huang-Huai-Hai region depended on the dilution and diffusion of eastward anomalous horizontal circulation. The increase in Hunan-Hubei-Guangdong Province was largely due to input from the north. In addition, the vast areas between the Yangtze River and the Yellow River were supported by higher temperatures and stronger shortwave solar radiation that promoted photochemical reactions. The reasons for changes in Shanxi-Sichuan-Yunnan Province were relatively more complex and thus require more in-depth exploration.

1. Introduction

China has a greater level of ozone (O₃) exposure to humans and vegetation than other developed nations [1]. Since 2020, O₃ has surpassed PM_2.5 to take the lead as the most significant air pollutant in the highly urbanized and industrialized regions of China [2]. Surface O₃ is a potent oxidant and phytotoxin [3] that poses direct risks to human health [4,5,6] and reduces crop yields [7,8]. O₃ may even lessen terrestrial carbon sinks by inhibiting vegetation photosynthesis, causing more severe global warming [9].

Surface O₃ is mainly produced by photochemical reactions of precursors, such as volatile organic compounds (VOCs) and nitrogen oxides (NOx), in the presence of solar radiation [10,11,12,13]. Thus, O₃ pollution, significantly influenced by precursor emissions, typically occurs in densely populated and industrially developed areas [14,15]. However, high concentrations of O₃ and its precursors could be transported by atmospheric circulation over long distances. Therefore, in addition to factors related to photochemical reactions, regional O₃ pollution is also regulated by atmospheric circulation transport and meteorological conditions associated with it [16,17].

China, bordered by the Pacific Ocean, has one of the most significant monsoon climates in the world [18]. The East Asian summer monsoon (EASM) is one of the most important components of the Asian monsoon system, which controls summer weather and climate in the area to the east of 100° E in China [19]. In addition, the EASM significantly affects the level and frequency of summertime O₃ pollution in the east of China by changing its external-natural factors [20,21,22]. Firstly, the EASM transports relatively clean marine air masses which are, in general, free from O₃ and its precursors from the northwest Pacific, favoring the dilution of O₃ in southeastern China [23,24]. Secondly, increased cloud cover and precipitation under the influence of humid air masses impedes photochemical reactions, causing O₃ concentrations to further decline in coastal areas [25,26]. Finally, the EASM circulation directly changes the horizontal distribution of O₃ by transporting the polluted air mass to downwind regions [23]. Pollutants in highly industrialized southeastern China are transported northward by the EASM, contributing to north-south differences in the background level of summertime O₃ in the east of China [27,28]. Numerous studies have focused on the effect of the EASM on O₃ distribution [22,24,29,30,31]; however, the relevant influencing pathways and dominant processes in different regions have not been thoroughly investigated yet, and are thus worth exploring further.

In this study, we utilize a global three-dimensional model of atmospheric chemistry driven by meteorological input from the Goddard Earth Observing System (GEOS-Chem) to simulate the impact of the EASM on surface O₃ concentrations in China during strong and weak EASM years selected by the East Asian summer monsoon index (EASMI). The interference of anthropogenic emission changes was eliminated since anthropogenic emissions were fixed during the simulations.

2. Materials and Methods

2.1. GEOS-Chem Model

GEOS-Chem, developed by Harvard University [32], has been widely used in the study of atmospheric composition and its transports [33,34,35,36]. In this study, GEOS-Chem (v11-01, http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v11-01#v11-01_public_release, accessed on 1 May 2020) was utilized. The nested-grid GEOS-Chem over Asia (60–150° E, 10° S–55° N) with a horizontal resolution of 0.5° × 0.625° (latitude × longitude) was driven by the MERRA-2 reanalysis meteorological dataset, which also served as boundary conditions with a resolution of 2° × 2.5°. The chemical mechanism used in this study is GEOS-Chem’s tropospheric chemistry (Tropchem) mechanism, which was formerly referred to as “NOx-Ox-HC-aerosol-Br”. The Tropchem mechanism includes NOx-Ox-hydrocarbon-aerosol chemistry, Methyl proxy nitrate chemistry, Halogen chemistry, simplified secondary organic aerosol species, and updated isoprene and monoterpene chemistry [37,38]. The linearized O₃ (LINOZ) chemical mechanism provided the stratospheric O₃ [39]. The non-local scheme was employed for vertical mixing within the boundary layer [40].

The Community Emissions Data System (CEDS) inventory, the default global anthropogenic inventory in GEOS-Chem, was used for global anthropogenic emissions. The MIX inventory for Asia includes the Asian Regional Emissions Inventory (REAS2) for the whole of Asia [41], the Multi-resolution Emission Inventory for China (MEIC) by Tsinghua University, a high-resolution NH₃ emission inventory by Peking University [42], an Indian emission inventory developed by Argonne National Laboratory (ANL-India) [43,44], and the official Korean emission inventory from the Clean Air Policy Support System (CAPSS) [45]. The biomass-burning emission inventory was provided by the Global Emission Database v4 (GFED4). Natural emissions were calculated online using the Model of Emissions of Gases and Aerosol from Nature (MEGAN v2.1). Lightning NOx emissions were parameterized by Price and Rind [46] and then driven by Lightning Imaging Sensor/Optical Transient Detector (LIS/OTD) climatology data sets. Both the lightning and soil NOx emissions were calculated online. Here, the anthropogenic emissions were fixed during the different years.

2.2. East Asian Summer Monsoon Index

The EASM has been measured by various EASMI [47,48,49]. Spatial mean Dynamical Normalized Seasonality (DNS) can indicate the seasonal variation in the horizontal wind field, causing it to be used in defining the EASMI [49,50]. The DNS index (DNSI) is given by

$$\begin{array}{c}DNSI=\frac{\parallel \overline{V_1}-V_i\parallel }{\parallel \overline{V}\parallel }-2\end{array}$$

(1)

where $V_1$ and $V_i$ denote horizontal wind vectors in January and month $i$, respectively. $\overline{V}$ is the average horizontal wind vector of January and July. The expression of the norm $\parallel A\parallel$ is

$$\begin{array}{c}\parallel A\parallel ={\left({{\displaystyle \iint }}^{ }{\left|A\right|}^{2}dS\right)}^{1/2}\end{array}$$

(2)

where $S$ represents the domain of integration. For $\parallel A\parallel$ at a point ($i,j$), it is defined as

$$\begin{array}{c}\parallel A_{i,j}\parallel \approx \Delta s{\left[\left(\left|A_{i-1, j}^2\right|+4\left|A_{i, j}^2\right|+\left|A_{i+1, j}^2\right|\right)cos{\phi }_j+\left|A_{i, j-1}^2\right|cos{\phi }_{j-1}+\left|A_{i, j+1}^2\right|cos{\phi }_{j+1}\right]}^{\frac{1}{2}} \end{array}$$

(3)

where ${\phi }_j$ is the latitude of the point ($i,j$), $\Delta s$ is

$$\begin{array}{c}\Delta s=\frac{a\Delta \phi \Delta \mathsf{\lambda }}{4}\end{array}$$

(4)

where a represents the mean Earth radius, $\Delta \phi$ and $\Delta \mathsf{\lambda }$ in radian are resolutions at meridional and zonal directions, respectively. Here, the EASMI was defined as summer (June-July-August, JJA) mean DNSI at 850 hPa for East Asia (10–40° N, 110–140° N).

As shown in Figure 1, 2018 was the strongest year for the EASM in at least the last 20 years. Chen, et al. [51] concluded that an extremely strong EASM appeared in 2018 based on multiple definitions of the EASMI [47,52,53]. To capture the characteristics of this extremely strong EASM in 2018, we selected it as a separate case of strong events. To increase the representativeness of weak EASM, we synthesized weak events of different intensities in 2003 and 2010 by using their average to represent a weak EASM.

2.3. Other Surface O₃ Data

Observed maximum daily 8-h average (MDA8) O₃ from the China National Environmental Monitoring Centre (CNEMC) was used to evaluate the performance of GEOS-Chem. Stations in Beijing, Shanghai, and Guangzhou, where extreme pollution events occur more frequently [54,55,56], were chosen for comparison. There are 12 stations in Beijing, 9 stations in Shanghai, and 11 stations in Guangzhou. A city’s mean MDA8 O₃ level was represented by averaging observations from all stations within that city. The simulation data was selected from grids of neighboring stations to match the observations.

Tracking Air Pollution in China (TAP, http://tapdata.org.cn/, accessed on 31 March 2023) provides near real-time MDA8 O₃ data products with high a resolution of 10 km from 2013 to the present [57]. TAP integrates multi-source data including ground observations, satellite remote sensing information, high-resolution emission inventory, air quality model simulation, and other ancillary data such as meteorological fields, land use data, population, and elevation.

2.4. Meteorological Data

The monthly mean meteorological data in JJA of 2003, 2010, and 2013–2022 were obtained from the European Centre for Medium-Range Forecasting (ECMWF) Reanalysis v5.0 (ERA5). Data on pressure levels between 1000–500 hPa (16 layers, https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-pressure-levels-monthly-means?tab=form, accessed on 31 March 2023), including u-component (U), v-component (V), and the vertical velocity ($\omega$) of wind, as well as geopotential height (GH), were used. Meteorological variables, including 2 m temperature (T-2m), surface solar radiation downwards (SSRD), total precipitation (TP), total cloud cover (TCC), low cloud cover (LCC), surface pressure (SP), 1000 hPa relative humidity (RH), U, V, the speed of wind at 10m (U-10m, V-10m, and WS-10m, respectively), and boundary layer height (BLH), were selected from ERA5 monthly data on single levels (https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels-monthly-means?tab=form, accessed on 31 March 2023). All meteorological variables have a spatial resolution of 0.5° × 0.5°.

2.5. Regression Model

An exhaustive search with linear regression (ES-LR) was utilized to measure the influence of meteorological processes on MDA8 O₃. Meteorological processes are divided into three categories: processes related to chemical reactions and horizontal and vertical transport processes. Meteorological parameters associated with each process are listed in

Table 1. Categories of meteorological processes affecting MDA8 O₃ and their associated meteorological variables.

Process Categories	Related Meteorological Variables
chemical reaction	T-2M, RH, TP, SSRD, TCC, and LCC
horizontal transport	SP, U-10m, V-10m, WS-10m, BLH, GH-500hPa 1, U-850hPa 1, and V-850hPa 1
vertical transport	W-950hPa 1,2, W-975hPa 1,2, and W-1000hPa 1,2

¹ The suffix specifies the pressure layer where the variable is located. ² To facilitate vector synthesis, $\omega$ (Pa/s) is converted to W (m/s).

. The linear regression equation is

$$\begin{array}{c}\mathrm{MDA}8{ \mathrm{O}}_3\mathrm{t}={\alpha }_0+{\displaystyle {\displaystyle \sum }_1^i}{\alpha }_i{x}_i\left(t\right)+\epsilon \left(t\right)\end{array}$$

(5)

where $\mathrm{MDA}8{ \mathrm{O}}_3\mathrm{t}$ is the detrended time series of JJA monthly mean MDA8 O₃ from TAP (2013–2022), $x_i\left(t\right)$ is detrended time series of meteorological parameter $x_i$, $\epsilon \left(t\right)$ is time series of the residual, ${\alpha }_0$ is the intercept, and ${\alpha }_i$ is the fitting coefficient of $x_i$. Explanatory variables were screened from Table 1 by maximizing the adjusted R-square ($R_{adjusted}^2$) in ES-LR. $R_{adjusted}^2$ represents the fraction of variance in MDA8 O₃ explained by the regression model, and its expression is

$$\begin{array}{c}{R}_{adjusted}^2=1-\frac{\left(1-R^2\right)\left(n-1\right)}{n-m-1}\end{array}$$

(6)

where $n$ is the number of samples, $m$ is the number of parameters, and $R^2$ is the coefficient of determination.

3. Results

3.1. GEOS-Chem Model Evaluation

Figure 2 shows the comparison of observed and simulated MDA8 O₃ in July 2015 in Beijing, Shanghai, and Guangzhou. GEOS-Chem performed the worst in Beijing, failing to present the fluctuation in the observation. The simulation and observation had the greatest agreement in Shanghai, particularly during the first three weeks. While not performing as remarkably as in Shanghai, the simulation in Guangzhou captured the approximate trend and extremums of observed MDA8 O₃. The correlation coefficients between the simulated MDA8 O₃ and the observations were 0.61 (p < 0.01), 0.89 (p < 0.01), and 0.76 (p < 0.01) in Beijing, Shanghai, and Guangzhou, respectively. Overall, GEOS-Chem reproduced the variations of surface O₃ in China to some extent, with higher accuracy in eastern and southern China, which is consistent with the results of the previous evaluation [58,59,60]. As a result, the GEOS-Chem simulation results were applied to the following discussion.

3.2. Comparison of Simulated MDA8 O₃ during Strong and Weak EASM

Figure 3 illustrates the spatial distribution of simulated MDA8 O₃ during strong and weak EASM along with their relative changes. Central and eastern China, as well as the Sichuan Basin, were the most heavily polluted regions. MDA8 O₃ in these regions reached approximately 80 ppbv during both weak and strong EASM (Figure 3a,b). Figure 3c reveals obvious changes in the spatial distribution of summertime surface O₃ induced by different intensities of EASM. Strong EASM greatly reduced surface O₃ mainly in three regions, which we refer to as Regions A−, B−, and C−, respectively (Figure 3c). Region A−, at the junction of southern Gansu Province, eastern Sichuan Province, southwestern Shanxi Province, western Chongqing Municipality, and northwestern Guizhou Province, had the smallest drop (around 3 ppbv) of the three. Covering the southeast of Hebei Province, Shandong-Anhui-Jiangsu Province, and eastern Zhejiang Province, Region B− experienced the largest reduction, reaching −15 ppbv in its coastal areas. Heilongjiang-Jilin-Liaoning Province and the northeast of Inner Mongolia Autonomous Region were divided into Region C−. The decline in MDA8 O₃ exceeded 5ppbv in the northeast of Region C−. Surface O₃ was increased by strong EASM in most of the remaining areas, particularly in the junction of Sichuan-Yunnan Province (Region A+), the junction of Hubei-Hunan Province (Region B+), and the western part of Guangdong Province (Region C+). The rise in MDA8 O₃ reached 5~7 ppbv in these regions. The relative change in MDA8 O₃ between strong and weak EASM relative to its level during weak EASM ((strong-weak)/weak) was −20~16% in the east of China. This variation pattern of surface O₃ caused by EASM has high consistency with the simulation results of RegCM4-Chem [20] and other studies based on GEOS-Chem [21,31].

3.3. Comparison of Meteorological Conditions during Strong and Weak EASM

The difference in surface O₃ between strong and weak EASM cases was mainly attributed to meteorological conditions since the anthropogenic emissions were fixed. The impact of meteorological conditions will be discussed from three perspectives: chemical reactions, horizontal diffusions and transports, and vertical processes. Before quantitatively evaluating changes in meteorological parameters caused by the EASM and qualitatively discussing their impacts on surface O₃ in selected regions, we first assessed fractions of variance in regional MDA8 O₃ explained by each meteorological process. According to $R_{adjusted}^2$ based on ES-LR (

Table 2. Explanatory variables screened by ES-LR used in linear regression models for different meteorological processes and corresponding $R_{adjusted}^2$.

Region	Process Categories	Explanatory Variables	$${\mathit{R}}_{\mathit{a}\mathit{d}\mathit{j}\mathit{u}\mathit{s}\mathit{t}\mathit{e}\mathit{d}}^2$$
A+	chemical	RH + TP + TCC	0.51
	horizontal	V-10m + BLH + GH-500hPa + V-850hPa	0.37
	vertical	W-975hPa + W-950hPa	0.00
	total	TCC + LCC + SP + U-10m + BLH + GH-500hPa + U-850hPa + W-950hPa	0.76
B+	chemical	T-2m + TP + SSRD + TCC	0.42
	horizontal	V-10m + GH-500hPa + U-850hPa	0.39
	vertical	W-1000hPa + W-975hPa + W-950hPa	0.03
	total	RH + LCC + U-10m + WS-10m + BLH + GH-500hPa + V-850hPa	0.72
C+	chemical	RH + TP	0.26
	horizontal	U-10m + GH-500hPa + U-850hPa + V-850hPa	0.56
	vertical	W-975hPa	0.46
	total	T-2m + RH + TCC + SP + U-10m + WS-10m + BLH + GH-500hPa + U-850hPa + W-1000hPa	0.71
A−	chemical	T-2m + TP + SSRD + TCC + LCC	0.46
	horizontal	V-10m + WS-10m + BLH + V-850hPa	0.05
	vertical	W-1000hPa + W-975hPa	0.00
	total	T-2m + RH + SSRD + TCC + LCC + V-10m + WS-10m + V-850hPa + W-975hPa	0.68
B−	chemical	T-2m + RH + SSRD	0.69
	horizontal	U-10m + V-10m + BLH + U-850hPa + V-850hPa	0.67
	vertical	W-1000hPa + W-950hPa	0.28
	total	RH + TP + SP + U-10m + V-10m + BLH + U-850hPa + W-1000hPa + W-975hPa + W-950hPa	0.79
C−	chemical	T-2m + RH + SSRD + LCC	0.68
	horizontal	U-10m + WS-10m + BLH + GH-500hPa + U-850hPa	0.71
	vertical	W-975hPa	0.18
	total	RH + TP + U-850hPa + W-950hPa	0.75

), meteorological parameters totally explained 76%, 72%, 71%, 68%, 79%, and 75% of the summertime MDA8 O₃ variation in Regions A+, B+, C+, A−, B−, and C−, respectively. In Regions A+, B+, A−, and B−, chemical reactions were more influential than transports [61], particularly in Region A− [62]. Regions C+ and C− were more affected by horizontal transports [63,64]. Vertical processes have little effect on Regions A+, B+, and A− and considerable effect on Region C+. The following analysis will focus on relative changes in the explanatory variables screened by ES-LR during strong and weak EASM.

3.3.1. Comparison of Chemical Reactions

Figure 4 shows the spatial distribution of the relative difference in meteorological parameters related to chemical reactions. According to Figure 4, strong EASM caused a positive-negative-positive tripolar response of RH, TP, TCC, and LCC, roughly bounded by the Yangtze and Yellow Rivers. With a consistent boundary, SSRD has an opposite response of negative-positive-negative. In general, the meteorological conditions in the regions between the Yangtze River and the Yellow River were favorable for photochemical reactions to produce O₃ during strong EASM.

Table 3. The spatial mean of relative changes ((strong − weak)/weak, %) in MDA8 O₃ and meteorological elements related to chemical reactions in selected regions with significant changes in surface O₃.

	A+	B+	C+	A−	B−	C−
MDA8 O3	7.72	6.92	6.88	−1.02	−7.19	−5.33
T-2m	0.06	0.22	−0.08	0.30	0.45	0.19
RH	0.31	−0.91	2.34	−0.05	−1.18	3.65
TP	7.19	−34.29	42.24	−6.66	−18.68	14.94
SSRD	3.20	5.61	−8.59	8.35	4.70	−4.55
TCC	−3.27	−15.19	10.27	−9.35	−16.79	0.62
LCC	−2.04	−21.63	5.25	−13.27	−15.62	14.65

lists the spatial mean of relative change in MDA8 O₃ and meteorological elements related to photochemical reactions in selected regions. The meteorological conditions in Region B+ were more favorable for photochemical reactions during strong EASM than during weak EASM [61]. Higher T-2m (0.22%) and lower RH (−0.91%) favored the NO₂ accumulation [65] and natural emissions of biogenic volatile organic compounds (BVOCs) [66,67]. TCC was lowered by −15.19% and SSRD was increased by 5.61%. Increased temperature and solar radiation further promoted O₃ production by accelerating photochemical reactions [10,63]. In addition, considerably decreased TP (−34.29%) weakened the wet scavenging effects of rainfall on O₃ precursors and enhanced solar radiation indirectly [68,69,70]. All these changes in meteorological conditions contributed to the increase in local surface O₃. In comparison, Region A+ gained fewer benefits from meteorologic-related chemical reactions than Region B+ due to the inhibitory effect of increased RH (0.31%) and TP (7.19%) on chemical reactions [71,72]. However, Region A+ got the largest increase in surface O₃ (7.72%) under the premise of a low level of emission [73,74,75], which indicates that Region A+ may be supported by dynamic input more greatly than Region B+ during strong EASM. The contributions of dynamic transport were even more dominant in Regions A−, B−, and C+. Though located in meteorological fields favorable to photochemical reactions with a high level of formaldehyde [76], Regions A− and B− experienced a decline in surface O₃, particularly in Region B− (−7.19%). Similarly, given that photochemical reactions were greatly suppressed by changes in meteorological conditions in Region C+, the local surface O₃ increase (6.88%) was attributed to dynamic processes. In Region C−, since the impact of higher T-2m (0.19%) was inconsistent with the weaker SSRD (−4.55%) during strong EASM, it is difficult to explain local changes in surface O₃ (−5.33%) from the perspective of photochemical reactions regulated by meteorological conditions.

3.3.2. Comparison of Horizontal Diffusions and Transports

Figure 5 shows the spatial distribution of the difference in meteorological elements closely related to diffusions in the boundary layer between strong and weak EASM. The WS-10m (Figure 5b) and BLH (Figure 5c) in Region B− increased by 12.86% and 17.54% (

Table 4. The spatial mean of relative changes ((strong − weak)/weak, %) in MDA8 O₃ and meteorological elements related to horizontal diffusions and transports in selected regions with significant changes in surface O₃.

	A+	B+	C+	A−	B−	C−
MDA8 O3	7.72	6.92	6.88	−1.02	−7.19	−5.33
SP	−0.07	−0.12	−0.31	−0.08	−0.12	0.01
U-10m	−19.62	−10.08	906	73.13	175.59	−2.81
V-10m	15.78	−114.89	−60.65	0.76	−7.25	48.5
WS-10m	−8.42	−7.57	−7.25	−4.62	12.86	1.21
BLH	−2.9	−1.01	−11.82	0.57	17.54	−2.34
GH-500hPa	−0.07	−0.03	−0.32	0.06	0.27	0.51
U-850hPa	33.24	−196.81	−202.17	269.72	229.43	−9.41
V-850hPa	−30.66	−71.55	−36.79	93.88	13.36	−42.44

), respectively, during strong EASM. While the increase in BLH was favorable for the mixing of precursors to produce O₃ [77,78], its promotion of outward diffusion contributed more to the decline of surface O₃ in Region B− [79,80]. In terms of wind vector at 10m (Figure 5a), the eastward wind anomaly located in Region B−, with a relative increase up to 175.59% in U-10m, brought the relatively clean air mass from the ocean and diluted local surface O₃. Similarly, the southern part of Region C− (i.e., Liaoning Province), where V-10m increased by 48.5%, was also affected by the dilution of the northward ocean air. The enhanced southwestward diffusion extended from Region B− toward the south China coast (Figure 5a), further causing surface O₃ to decrease in Region B− and increase in Regions B+ and C+ [21,81]. The influence of the EASM on horizontal diffusion conditions within the boundary layer was relatively insignificant in Regions A+ and A−, particularly in the former (Table 4). Complex geography may have played an important role. Previous research has attributed O₃ pollution in Sichuan Province to downward transport from the Qinghai-Tibet Plateau [82,83].

Figure 6 further illustrates the large-scale cyclonic anomalous circulation over the Northwest Pacific during strong EASM. The east wind anomalies had incredible impacts on southeastern China, including Regions B+, C+, B−, and A+, where the U-850hPa experienced relative changes by around 200% (Table 4). Figure 6 shows that strong EASM induced a north-south dipole response of GH-500hPa with a transition latitude near 31° N, making the Northwest Pacific Subtropical High (WPSH) weakened. Interestingly, surface O₃ in northern and southern China in turn responded to weakened WPSH with a similar dipole-like pattern [67]. Bounded by 32° N, MDA8 O₃ rises significantly in southern China, particularly in the southeast part, and drops in northern China during weaker WPSH [67]. Thus, increased O₃ in Regions A+, B+, and C+ was also additionally supported by weaker WPSH, which led to a significant rise in natural emissions in southern China [67]. According to Table 4, the relative change in the U-850hPa and V-850hPa in Region C− was notably smaller (−9.41% and −42.44%, respectively) than in other regions, indicating that local O₃ change was relatively less dependent on horizontal transport in Region C−.

3.3.3. Comparison of Vertical Processes

The drop of surface O₃ in Region C−, particularly in its northern part, which failed to be explained by changes in chemical reactions and horizontal circulation transport (Figure 4, Figure 5 and Figure 6), may be attributed to vertical processes. Figure 7 shows the vertical cross-section of the differences in regional zonal mean V-W between strong and weak EASM in selected regions. The upward W anomaly in Region C− extended from the ground surface to the top of the boundary layer (Figure 7f), which may greatly account for the decline in local MDA8 O₃. The relative change in W-1000hPa in Region C− was the largest of the selected 6 regions, up to −244.81%, which was nearly 5.9~21.5 times that of the other regions (

Table 5. The spatial mean of relative changes ((strong − weak)/weak, %) in MDA8 O₃ and W within the boundary layer in selected regions with significant changes in surface O₃.

	A+	B+	C+	A−	B−	C−
MDA8 O3	7.72	6.92	6.88	−1.02	−7.19	−5.33
W-950hPa	−25.68	53.40	−125.69	−3.52	−413.15	−47.08
W-975hPa	−14.14	9.44	−81.76	−24.87	549.37	−9.38
W-1000hPa	13.24	−11.39	−50.35	−13.72	11.83	−244.81

). The upward transport within the boundary layer was also found in Region A−, particularly in areas outside the Basin (Figure 7b). The upward transport in the northern part of Region A was even stronger than in Region C−. However, the downward input between 30–32° N partially offset the loss of surface O₃ in Region A− and reduced its decline (Table 5). Strong downward transport in the south part of Region A+ seemed to be an important source of the regional surface O₃ (Figure 7a), which supported the hypothesis that the Tibetan Plateau greatly contributed to high O₃ in the Basin [83]. In contrast, Regions B+, C+, and B− were dominated by strong horizontal transport during strong EASM (Figure 7c–e), which counteracted the effects of vertical transport on regional O₃ (Table 5).

4. Discussion

Based on the similarity in the changes between the surface O₃ integrated process rate caused by chemical reactions and its concentration simulated by RegCM4-Chem, previous research pointed out that the chemical process is the uppermost factor in reducing surface O₃ levels in the eastern part of China (i.e., Region B− in our study) during summer monsoon seasons [20]. However, our study illustrates the important effects of abnormal horizontal and vertical circulation transport, caused by changes in EASM intensity, on the distribution pattern of surface O₃ in central and southeastern China. A variety of reasons can lead to inconsistent outcomes, such as the fact that the objects and methods of these two studies are not identical, the variations in physical and chemical processes between the two models, etc. More in-depth quantitative analysis is required even though our findings are in line with the model analysis [21,31] and viewpoints from space monitoring [30,84]. In addition, the complex response of natural emissions of BVOCs and NOx to the climate system [85] and their impact on surface O₃ [86] need to be further explored in future studies.

5. Conclusions

In this study, we first evaluated the performance of GEOS-Chem in simulating summertime surface O₃ concentrations against the monitoring data from the CNEMN. The results demonstrate that GEOS-Chem performs better in eastern and southern China. Correlation coefficients between simulations and observations are 0.61 (p < 0.01), 0.89 (p < 0.01), and 0.76 (p < 0.01) in Beijing, Shanghai, and Guangzhou, respectively.

To compare the impact of the EASM on surface O₃ in the east of China, GEOS-Chem with fixed anthropogenic emissions was used to simulate summer surface O₃ during strong and weak EASM. We found that strong EASM mainly greatly reduced surface O₃ in the Heilongjiang-Jilin-Liaoning Province in the northeast, the Huang-Huai-Hai coast in the east, and the junction of Shanxi-Sichuan-Chongqing in the west of China. Surface O₃ increased during strong EASM in most of the remaining areas, particularly in the junction of Sichuan-Yunnan Province, the border of Hubei-Hunan Province, and the western part of Guangdong Province. The spatial mean of the relative difference of MDA8 O₃ between the strong and weak EASM in these regions reached approximately −7~7% of their levels during weak EASM. Utilizing MDA8 O₃ data with high accuracy from TAP, it was confirmed by linear regression that the combination of meteorological parameters could explain 68~79% of summertime surface O₃ variability in these regions.

The meteorological process analysis revealed that the decrease in surface O₃ of northeast China during the strong EASM was mainly attributed to the large increase in vertical upward transport. The considerable decrease in the Huang-Huai-Hai region depended on the dilution and diffusion of the eastward anomalous horizontal circulation. The increase in Hunan-Hubei-Guangdong Province was largely due to input from the north. In addition, the vast area between the Yangtze River and the Yellow River has also been supported by higher temperatures and stronger shortwave solar radiation that promoted photochemical reactions [61,65]. The reasons for the changes in Shanxi-Sichuan-Yunnan Province were relatively more complex and required more in-depth exploration.