The impact of Climate Change on the Western Pacific Subtropical High and the related ozone pollution in Shanghai, China

Severe ozone (O3) episodes occur frequently in Shanghai during late-summers. We define geopotential height averaged over the key area region (122.5°E-135°E, 27.5°N -35°N) at 500 hPa as a WPSH_SHO3 index which has high positive correlation with surface O3 concentration in Shanghai. In addition, the index has a significant long-term increasing trend during the recent 60 years. Analysis shows the meteorological conditions under the strong WPSH_SHO3 climate background (compared to the weak background) have several important anomalies: (1) A strong WPSH center occurs over the key area region. (2) The cloud cover is less, resulting in high solar radiation and low humidity, enhancing the photochemical reactions of O3. (3) The near-surface southwesterly winds are more frequent, enhancing the transport of upwind pollutants and O3 precursors from polluted regions to Shanghai and producing higher O3 chemical productions. This study suggests that the global climate change could lead to a stronger WPSH in the key region, enhancing ozone pollution in Shanghai. A global chemical/transport model (MOZART-4) is applied to show that the O3 concentrations can be 30 ppbv higher under a strong WPSH_SHO3 condition than a weak condition, indicating the important effect of the global climate change on local air pollution in Shanghai.

Furthermore, synoptic-scale weather pattern, with spatial scale less than 1000 km, is also an important factor in controlling ozone variability [17][18][19][20][21] . For example, anticyclones (i.e., high pressure systems) produce favorable conditions for ozone production. At the center of anticyclones, it is normally sunny weather, with low wind velocity, causing high O 3 production and accumulation 22,23 . In coastal cities, sea-land breeze can be an important meteorological factor to affect ozone distributions. Tie et al. (2009) finds that the impact of sea breeze on O 3 concentration is noticeable in the city of Shanghai under calm weather condition.
It is well known that the West Pacific Subtropical High (WPSH) is evident as a semi-permanent, sub-tropical anticyclone high pressure over the western North Pacific, affecting the summertime weather and climate in China [24][25][26] . Despite of the effect on weather conditions (such as summertime precipitation and temperature), WPSH has also important effects on vegetation coverage and other research field 27,28 . Currently, the impact of the position and intensity of the WPSH on summertime ozone pollution over eastern China have been paid more attention. Most of the previous studies examine the relationship between the WPSH and ozone on a daily scale. For example, He et al. 29 focus on an short-term ozone pollution event in Shanghai and finds that ozone mixing ratios in summertime at Chongming (a surface site in the northeast of Shanghai) are often higher during the days when the center of the WPSH locates to the southeast of that site, with a weak intensity. Some results 30 show the subsidence air caused by the WPSH plays a crucial role in the formation of high-level O 3 . Zhao et al. 31 studies the impact of WPSH on surface ozone daily variability over eastern China and demonstrates that a stronger WPSH is associated with lower ozone in South China but with higher ozone in North China, suggesting that this south-north difference can be explained by changing moisture transport associated with the WPSH variability.
Generally, the WPSH is closely associated with the timing and spatial distribution of summer ozone concentrations in East Asia and may intensify in a warming climate background 32 . However, there is s lack of study to analyze the effect of global changes on WPSH and the consequence on the O 3 concentrations in the Shanghai region. The focus of this study is to investigate a strong inter-annual variability of ozone pollution and the effect of global climate change on summertime ozone pollution in Shanghai. The paper is organized as follows: in Section 2, we describe the information of the methodology. In Section 3, some results and analysis are analyzed. Section 4 shows a brief conclusion of the results. Gao et al. (2017) and Lin et al. (2017) have demonstrated that the ozone concentrations in Shanghai steadily increase, with a strong seasonal variation. Because the effect of WPSH on O 3 concentrations often occurs in summer time, the monthly variation of ozone during 2013 to 2017 in the Shanghai region is analyzed and is presented in Fig. 1. Figure 1 shows that the 8h-averaged ozone (MDA8 O 3 ) concentrations have highest occurrences in late-summer (July and August) compared with those in other months. The new Chinese national ambient air quality standards (CNAAQS2012, GB 3095-2012) has defined the severe, moderate, and slight ozone pollution 33 . The total pollution days on July and August from 2013 to 2017 are 82 and 56 days, respectively. According to the study by Gao et al. (2017), the ozone concentrations are low in early summer (June), because of the occurrence of frequent precipitation in June which is named Meiyu-Baiu-Changma rain belt in China. Previous studies 34 have demonstrate that the seasonal northward shifts of the WPSH are closely associated with the onset and withdrawal of the EASM, during July and August, the ridge of WPSH locate around Yangtze-Huaihe River region, and the www.nature.com/scientificreports www.nature.com/scientificreports/ weather conditions here are strongly controlled by the main body of WPSH system. As a result, it is important to understand the impact of WPSH on surface ozone in Shanghai during late-summer.

Discussion
As we all know, WPSH is a multi-dimensional climate system, its temporal and spatial circulation characteristics are more complex, so it is often simplified by one dimension monitoring index to analyze its effect on ozone. Five operational indices of WPSH based on the 5880 geopotential height (gpm) at 500 hPa, including the area, intensity, ridge lines, northern extension, and western boundaries are announced monthly by the National Climate Center (NCC) in China to describe the WPSH's evolution 35 . But the correlations between ozone concentrations in Shanghai and indices on a monthly scale is very poor, Consequently, it is necessary to develop a new objective index, which is not only with clear physical meanings to objectively describe the location and intensity of the WPSH, but also has significant associations with surface ozone in Shanghai. Figure 2 shows the distribution of correlation coefficients between monthly mean MAD8 O 3 in Shanghai and the 500 hPa geopotential height over Eastern Asia during late-summer from 2013 to 2017. To obtain correlation analysis data as more as possible, monthly mean data are used rather than late-summer mean data. Correlations between ozone concentrations and 500 hPa geopotential height are positive in most parts of Eastern Asia. The significant positive correlation coefficients with 500 hPa geopotential height (exceeding the 95% confidence level based on student-t test) are located over eastern ocean of China. The significant correlation coefficients related to the geopotential height averaged over the key area region (122.5°E-135°E, 27.5°N-35°N) at 500 hPa shows a maximum positive correlation coefficient (0.78). Thus, we define geopotential height averaged over the key area region at 500 hPa as the new definition index of WPSH's effect on ozone (called WPSH_SHO 3 ). It not only represents the activity center and intensity of the WPSH over the key area region, but also it has high positive correlation with the surface ozone concentration in Shanghai. Figure 3 displays the inter-annual variability of monthly mean O 3 (MAD8) in Shanghai and the corresponding WPSH_SHO 3 index from 2013 to 2017. The characteristics of the two parameters were highly correlated (except during 2015), with a correlation coefficient of 0.78 (exceeding the 99% confidence level). The high correlation between WPSH_SHO 3 index and O 3 concentrations indicated that the intensity of the WPSH_SHO 3 can be used as an indictor to predict a general tendency of late-summer O 3 concentrations in Shanghai.
West Pacific Subtropical High (WPSH) is mainly affected by large-scale circulations, previous studies 36, 37 have shown different results of how the ongoing global warming would change the WPSH. To understand the impact of Climate Change on the WPSH and its impact on ozone pollution in Shanghai, Fig. 4 shows the long-term trend of late-summer mean WPSH_SHO 3 index from 1958 to 2017. As shown in Fig. 4, the index of WPSH_SHO 3 significantly increased during the recent years , with an increasing rate of about 5 gpm decade −1 . For example, the WPSH_SHO 3 was about 5885 gpm in 2017 and 5855 gpm in 1958, respectively. This rapid increase suggested that the global climate change had strong impacts on the WPSH_SHO 3 , which might have important effects on air pollutants (such O 3 ) in Shanghai (as shown in Fig. 3).
To systematically study the impact of WPSH_SHO 3 on O 3 concentrations in Shanghai during late-summers, two different late-summer cases were selected and compared. The first case (Jul. 2017) was a typical strong WPSH case, with the maximum anomalies geopotential height, while the second case (Aug. 2016) was a typical weak WPSH, with the minimum anomalies geopotential height. It is interesting to note that the difference of geopotential height between the two cases is about 30 gpm, which is equivalent to the value between the values of WPSH_SHO 3 in 1957 and 2017 (see Fig. 4). As a result, comparing the two extreme cases can provide useful insights to understand the impact of global climate change on WPSH_SHO 3 and the corresponding effects on O 3 concentrations in Shanghai. Figure 5 shows the monthly mean O 3 (MAD1) concentrations during the strong (Jul. 2017) and the weak (Aug. 2016) WPSH_SHO 3 cases in Shanghai and its surrounding regions. The results show that there was significantly difference for the O 3 concentrations in these 2 cases. During the weak case, the maximum O 3 concentration was located in the inland and the west of Shanghai (SH), with a highest value of >180 μg m −3 . As a result, the O 3 concentrations in SH were low, ranging from 120-140 μg·m −3 . In the contrast, during the strong WPSH case, the maximum O 3 concentration was higher than the weak WPSH case, and the highest O 3 located in the city of Shanghai (SH), with a highest value of >200 μg m −3 . As a result, the O 3 concentrations in SH were high, ranging from 180-200 μg m −3 , resulting in a large anomaly of O 3 concentrations (60 μg m −3 ) between the strong and the weak WPSH cases in Shanghai.
To understand the impact of WPSH_SHO 3 on the anomalies of ozone concentrations between strong and weak WPSH cases, detailed meteorological conditions were analyzed.  www.nature.com/scientificreports www.nature.com/scientificreports/ Figure 6 shows the large-scale atmospheric circulations over 500 hPa between the two cases. It shows the mid-level (500 hPa) circulations of weak case (Fig. 6a) and strong case (Fig. 6b) and its anomalies (the monthly mean values in 2017 minus the monthly mean values in 2016, shown in Fig. 6c). Compared to climatological WPSH_SHO 3 (solid lines), the WPSH_SHO 3 had characteristics with bigger area and stronger intensity under strong WPSH_SHO 3 climate background (Fig. 6b). Moreover, there was significant sub-center of WPSH_SHO 3 over the key area regions (Fig. 3). In contrast, the weak WPSH_SHO 3 had no high center over the West Pacific Ocean (Fig. 6a). As a result, there were significant positive anomalies of geopotential height over 500 hPa in the West Pacific Ocean region, which was approaching the key area region mentioned in Fig. 2. For example, the geopotential height over 500 hPa in the key area region of strong case is about 30-50 gpm higher than weak case. Tie et al. (2009) find that radiation and wind (speed and direction) are the most important meteorological factors for affecting surface ozone pollution in Shanghai, which directly affect the photochemical reaction, diffusion and transport of surface ozone. In this study, the anomalies of the large-scale circulations over 925 hPa, relative humidity, solar radiation, and surface daily-max temperature between the weak and strong cases are investigated.
As shown in Fig. 7a, in the strong case (July 2017), there was is a strong anomalous southwesterly winds in Shanghai, which was enhanced by 4-5 m·s −1 over 925 hPa. The surface temperature was higher 3 °C than the weak case (Fig. 7d). The relative humidity was weaker by 20% than the weak case (Fig. 7b), and the radiation was stronger by 20 W•m −2 than the weak case in Shanghai (Fig. 7d). During the southwesterly winds condition, the upwind region is a large-scale forest. As a result, the enhanced southeasterly winds transported upwind biogenic emissions to Shanghai, causing and the increasing the ozone pollution in Shanghai 38 . However, it's essential to give some supplementary explanation that although the anomaly of specific humidity was small between the strong case and the weak case, the local solar radiation and temperature were enhanced during the strong case (Fig. 7c,d). As a result, the higher solar radiation and temperature reduced relative humidity (RH). Furthermore, the meteorological conditions, with high temperature, low humidity, and high solar radiation were favorable for producing high-level O 3 episodes 39 . And the possible effects of relative humidity (RH) on the ozone formation can be well explained by Yu (2018), as the paper shows, more air humidity could inhibits O 3 formation by lowering air temperature and some complicated chemistry processes, like decreasing the chain length of peroxy radical chemical amplifiers (HO 2 , RO 2 , and RC(O)O 2 ), and decreasing the chain length of NO 2 by enhancing particle water, and destroys the existing O 3 photo-chemically by water vapor through catalytic O 3 destruction cycle 40 . As a result, during the strong WPSH_ SHO 3 case, the photochemical production of O 3 was more active than the weak case.
Same characteristics are found in local weather conditions in the Baoshan station in Shanghai. Table 1 and Fig. 8 show the local weather conditions in the city of Shanghai. The average meteorological conditions were calculated for the both strong and weak cases in the Baoshan observation station. As shown in Table 1, the relative humidity was lower in the strong case (65% compared with the 71% in the weak case). The hourly maximum temperature was higher in strong case (30 °C) than the weak case (32 °C). The measured photolysis rate of NO 2 www.nature.com/scientificreports www.nature.com/scientificreports/ (J[NO 2 ]) was higher (1.9 × 10 -3 s −1 ) than the weak case (1.1 × 10 −3 s −1 ), producing higher ozone photochemical production. Another important factor that increases the O 3 concentrations was due to the wind directions. In the strong WPSH_SHO 3 case, southwest wind occurred frequently in Shanghai. In the contrast, in the weak WPSH_SHO 3 case, east and southeast winds were dominated (Fig. 8), which transported relative clean air from ocean, resulting in lower O 3 concentrations in the weak case. In addition, the areas with subtropical high-pressure control (the strong WPSH_SHO 3 case) are dominated by subsidence air flow, producing weak convection and cumulus clouds. Under less cloud conditions, solar radiation is high, which may greatly enhance the photochemical reactions of O 3 .
In order to better investigate the effects of the WPSH_SHO 3    www.nature.com/scientificreports www.nature.com/scientificreports/ Figure 10 shows the calculated anomalies of the contributions of individual processes to O 3 formation in the Shanghai region under the different WPSH_SHO 3 conditions. In the strong WPSH_SHO 3 case, the vertical diffusion produced significant reduction for the O 3 concentrations at the surface. This was due to the fact that with less cloud condition, the solar radiation was stronger, producing higher thermal turbulence and vertical diffusion. As a result, the surface O 3 concentrations were vertical mixed in the upper planetary boundary layer, causing the decrease of O 3 concentrations at surface layer in Shanghai (with a maximum reduction of 80-160 μg·m −3 ·day −1 , shown in Fig. 10a). As we mentioned before, the wind direction in the strong case was favorable for the O 3 concentrations, which resulted in about 20-40 μg·m −3 ·day −1 increase of surface O 3 concentrations (Fig. 10b). The highest enhancement of O 3 concentrations was due to the process of photochemistry. As shown in Fig. 10c, the maximum increase of O 3 concentrations in Shanghai ranges about 100-180 μg·m −3 ·day −1 , resulting from the higher O 3 chemical productivity. In total, the surface O 3 concentrations were higher in the strong WPSH_SHO 3 case than the weak WPSH_SHO 3 case in Shanghai.  www.nature.com/scientificreports www.nature.com/scientificreports/

Conclusions
It is well known that WPSH is a strongest subtropical anticyclone high pressure system controlling monthly weather conditions in YRD during late-summer and may intensify in a warming climate background. To systemically study the impact of WPSH on inter-annual variability of ozone pollution in Shanghai under the climate change background, intensive surface measurements of ozone and meteorological data and the NCEP/NCAR re-analyzed meteorological data are used in the analysis. The highlights of this study are as follows: (1) To better understand the relationship between WPSH and O 3 in Shanghai, we define geopotential height averaged over a key area (122.5°E-135°E, 27.5°N-35°N) at 500 hPa. In this region, the correlation coefficients between related to ozone concentration in Shanghai shows a maximum positive value (0.78). As a result, the WPSH's in this region (defined as WPSH_SHO 3 ) can better represent the effect of WPSH on the O 3 concentrations in Shanghai, and are used in this study.  The results show that in the strong WPSH_SHO 3 case, the lower relative humidity, higher temperature, less cloud, and higher solar radiation produced higher ozone photochemical production than the weak WPSH_SHO 3 case. In addition the wind directions in the strong case were dominated by southwest wind. In the contrast, in the weak WPSH_SHO 3 case, east and southeast winds were dominated, which transported relative clean air from ocean, resulting in lower O 3 concentrations in the weak case.

Materials and Methods
Measured surface chemical observations. Two measured surface O 3 datasets are used in the study: ( The large-scale weather conditions, such as general circulation are used the data from the National Center for Environmental Prediction (NCEP) and National Center for Atmospheric Research (NCAR) reanalysis data 41 . The data have a horizontal resolution of 2.5° × 2.5°, including the geo-potential height, specific humidity and winds data. In addition, the solar radiation data with a horizontal resolution of 1.875° × 1.904° is also used in this article.

Global chemistry transport model (Mozart-4) description.
A global chemistry transport model (MOZART-4; Model for Ozone and Related chemical Tracers, version-4) is used in this study. The detailed model description is shown by Emmons et al. 42 , and the detailed aerosol modules are shown by Tie et al. 43 . The MOZART-4 model is a global chemical transport model. The model is designed to study the global distributions of tropospheric trace gases and aerosol particles. In this study, the horizontal resolution of the model is 0.7° × 0.7°, with 42 vertical levels. Advection of tracers is performed using the flux-formed semi-Lagrangian advection scheme 44 . The deep convection scheme developed by Zhang and McFarlane 45 is included in the model. The model transport is driven by the European Centre for Medium-Range Weather Forecasts (ECMWF) assimilated wind fields, with 0.5° × 0.5° horizontal resolution 46 . The meteorological data is interpolated to fit the model horizontal resolution by using a bilinear interpolation method. The MOZART model with such configurations has been successfully employed by Chang et al. 47 to investigate the impact of El Niño event on regional air quality of China. The MOZART-4 model is applied to calculate the contributions of individual processes to O 3 formation, including the total diffusion (DIF), advection (ADV) and gas-phase chemistry (CHEM).

Data availability
All data is available on-line and free of charge.