The Impact of Meteorology and Emissions on Surface Ozone in Shandong Province, China, during Summer 2014–2019

China has been experiencing severe ozone pollution problems in recent years. While a number of studies have focused on the ozone-pollution-prone regions such as the North China Plain, Yangtze River Delta, and Pearl River Delta regions, few studies have investigated the mechanisms modulating the interannual variability of ozone concentrations in Shandong Province, where a large population is located and is often subject to ozone pollution. By utilizing both the reanalysis dataset and regional numerical model (WRF-CMAQ), we delve into the potential governing mechanisms of ozone pollution in Shandong Province—especially over the major port city of Qingdao—during summer 2014–2019. During this period, ozone pollution in Qingdao exceeded the tier II standard of the Chinese National Ambient Air Quality (GB 3095-2012) for 75 days. From the perspective of meteorology, the high-pressure ridge over Baikal Lake and to its northeast, which leads to a relatively low humidity and sufficient sunlight, is the most critical weather system inducing high-ozone events in Qingdao. In terms of emissions, biogenic emissions contribute to ozone enhancement close to 10 ppb in the west and north of Shandong Province. Numerical experiments show that the local impact of biogenic emissions on ozone production in Shandong Province is relatively small, whereas biogenic emissions on the southern flank of Shandong Province enhance ozone production and further transport northeastward, resulting in an increase in ozone concentrations over Shandong Province. For the port city of Qingdao, ship emissions increase ozone concentrations when sea breezes (easterlies) prevail over Qingdao, with the 95th percentile reaching 8.7 ppb. The findings in this study have important implications for future ozone pollution in Shandong Province, as well as the northern and coastal areas in China.


Introduction
China has been experiencing severe environmental pollution-particularly ozone pollution-in the summer. Although the Clean Air Action Plan of 2013 has led to reductions in the majority of air pollutant emissions [1], and subsequent decreases in PM 2.5 concentrations [2][3][4], the trend of ozone concentrations is increasing [5,6]. Surface ozone may pose threats to human health [7,8] and plant growth [9,10]; thus, it is of vital importance to tackle the ozone pollution issue.
Surface ozone is produced mainly through complex photochemical reactions. Nitrogen oxides (NO X ≡ NO + NO 2 ) and volatile organic carbons (VOCs) are the primary precursors This study uses the Community Multiscale Air Quality (CMAQ, version 5.3.1) model for atmospheric chemistry simulation. Aerosol Module Version 7 (AERO7) and Carbon Bond Version 6 (CB6) [42] represent the aerosol-and gas-phase chemistry. The Model for Ozone and Related Chemical Tracers Version 4 (MOZART-4) [43] was used to downscale chemical initial and boundary conditions (refer to Ma et al. [44] for more detailed information). The anthropogenic emission inventory of 2016 was taken from the 0.25° × 0.25° Multi-resolution Emission Inventory for China (MEIC, http://www.meicmodel.org; accessed on 2 February 2022) [45]. Ship emissions were calculated from the shipping emission inventory model [46,47]. The Model of Emissions of Gases and Aerosols from Nature (MEGAN, version 2.1) [48] was used to estimate hourly biogenic emissions. More information can be found in the work of Zeng et al. [35]. We also performed three sensitivity experiments with different emissions (Table 1) to explore the influence of emissions on ozone concentrations in Shandong Province. The biogenic contribution is defined as the simulation of SEN_ASB minus SEN_AS; the ship contribution is defined as the simulation of SEN_AS minus SEN_A; the contribution of biomass burning is defined as the simulation of CTRL minus SEN_ASB. This study uses the Community Multiscale Air Quality (CMAQ, version 5.3.1) model for atmospheric chemistry simulation. Aerosol Module Version 7 (AERO7) and Carbon Bond Version 6 (CB6) [42] represent the aerosol-and gas-phase chemistry. The Model for Ozone and Related Chemical Tracers Version 4 (MOZART-4) [43] was used to downscale chemical initial and boundary conditions (refer to Ma et al. [44] for more detailed information). The anthropogenic emission inventory of 2016 was taken from the 0.25 • × 0.25 • Multi-resolution Emission Inventory for China (MEIC, http://www.meicmodel.org; accessed on 2 February 2022) [45]. Ship emissions were calculated from the shipping emission inventory model [46,47]. The Model of Emissions of Gases and Aerosols from Nature (MEGAN, version 2.1) [48] was used to estimate hourly biogenic emissions. More information can be found in the work of Zeng et al. [35]. We also performed three sensitivity experiments with different emissions (Table 1) to explore the influence of emissions on ozone concentrations in Shandong Province. The biogenic contribution is defined as the simulation of SEN_ASB minus SEN_AS; the ship contribution is defined as the simulation of SEN_AS minus SEN_A; the contribution of biomass burning is defined as the simulation of CTRL minus SEN_ASB.

Observed Ozone Characteristics
The ozone distribution in Shandong Province (purple box in Figure 2a) was first calculated based on observations over each month during summer 2014-2019, as shown in Figure 2. In general, the ozone concentrations over Shandong Province were the highest in June, followed by July and August. The MDA8 ozone concentration exceeded 80 ppb in inland areas, such as parts of northwest Shandong Province. The seasonal average ozone concentration during summer 2014-2019 over the coastal city of Qingdao (the approximate colored area inside the blue box in Figure 2a) was 61.7 ppb-about 10 ppb lower than those in inland areas, likely associated with the relatively clean air from the ocean. As a result, ozone pollution was more severe in inland cities than on the coast.

Observed Ozone Characteristics
The ozone distribution in Shandong Province (purple box in Figure 2a) was first calculated based on observations over each month during summer 2014-2019, as shown in Figure 2. In general, the ozone concentrations over Shandong Province were the highest in June, followed by July and August. The MDA8 ozone concentration exceeded 80 ppb in inland areas, such as parts of northwest Shandong Province. The seasonal average ozone concentration during summer 2014-2019 over the coastal city of Qingdao (the approximate colored area inside the blue box in Figure 2a) was 61.7 ppb-about 10 ppb lower than those in inland areas, likely associated with the relatively clean air from the ocean. As a result, ozone pollution was more severe in inland cities than on the coast.

Meteorological Conditions Contributing to Ozone Pollution
In this section, the meteorological variables from the ERA5 dataset are averaged over Shandong Province and Qingdao. Table 2

Meteorological Conditions Contributing to Ozone Pollution
In this section, the meteorological variables from the ERA5 dataset are averaged over Shandong Province and Qingdao. Table 2 displays the possible meteorological factors that may be connected with the daily variations of MDA8 ozone concentrations in Shandong Province during summer 2014-2019. The surface shortwave radiation and near-surface relative humidity at 1000 hPa are strongly correlated with MDA8 ozone concentrations in Shandong Province, with correlation coefficients of 0.67 and −0.66 during summer 2014-2019, respectively, stressing the crucial role of intense sunlight in photochemical reactions [51][52][53]. Moreover, regional transport is also an essential factor affecting ozone concentrations. Although the wind speed at 10 m shows no significant relationship with MDA8 ozone concentrations in Shandong Province, zonal wind speed at 10 m is closely associated with MDA8 ozone concentrations in Shandong Province, especially in Qingdao, with a correlation coefficient of 0.42 (displayed in Table 2). The correlation indicates that the stronger westerly can transport gaseous pollutants from the upstream, such as the North China Plain, to Qingdao, aggravating local ozone pollution, akin to regional ozone transport from the polluted Pearl River Delta to another coastal city-Hong Kong [54].
Notably, although the maximum daily temperature is significantly correlated with MDA8 ozone concentration in Qingdao, the correlation is lower than that based on surface shortwave radiation and relative humidity, likely due to the high relative humidity (i.e., 75% during summer 2014-2019) in coastal areas [52]. Therefore, the surface shortwave radiation and relative humidity are the most important factors affecting the ozone concentrations in Shandong Province. The correlation coefficient between maximum daily temperature and ozone concentrations is smaller in coastal cities than in inland cities.
To delve into the large-scale weather systems influencing ozone pollution in summer, a simple method was designed to identify high-ozone events objectively. A high-ozone event is detected when ozone concentrations in an area exceed 82 ppb (tier II standard of the Chinese National Ambient Air Quality GB 3095-2012) for at least four consecutive days. Geopotential height at 500 hPa and ozone anomalies compared with the summer mean of 2014-2019 were composited for five high-ozone events in Qingdao to delineate the atmospheric circulations ( Figure 4). A conspicuous high-pressure ridge dominated over Lake Baikal and its northeast, concurrent with a cyclonic system over northern Japan and Northeast China. Controlled by the high-pressure ridge, the relative humidity and surface shortwave radiation anomalies in Qingdao during high-ozone events were −12% and 39 W/m 2 , respectively, compared with the summer mean during 2014-2019, conducive to the occurrence of high-ozone events.

Contribution of Different Emissions to Ozone Concentrations
The observed data were first used to evaluate the meteorological factors in WRF-CMAQ during summer 2014-2019, e.g., hourly air temperature at 2 m, wind direction at 10 m, and wind speed at 10 m ( Table 3). The performance of the WRF model was reasonable over Shandong Province, generally meeting the benchmark [55], except for the mean gross of wind direction at 10 m. Meanwhile, the relatively large gross error in wind direction at 10 m-i.e., 44.0 • for Shandong Province and 46.8 • for Qingdao-may be caused by values close to 0 or 360 degrees [56].
the trend of increasing ozone in recent years.
Geopotential height at 500 hPa and ozone anomalies compared with the summer mean of 2014-2019 were composited for five high-ozone events in Qingdao to delineate the atmospheric circulations ( Figure 4). A conspicuous high-pressure ridge dominated over Lake Baikal and its northeast, concurrent with a cyclonic system over northern Japan and Northeast China. Controlled by the high-pressure ridge, the relative humidity and surface shortwave radiation anomalies in Qingdao during high-ozone events were −12% and 39 W/m 2 , respectively, compared with the summer mean during 2014-2019, conducive to the occurrence of high-ozone events.

Contribution of Different Emissions to Ozone Concentrations
The observed data were first used to evaluate the meteorological factors in WRF-CMAQ during summer 2014-2019, e.g., hourly air temperature at 2 m, wind direction at 10 m, and wind speed at 10 m ( Table 3). The performance of the WRF model was reasonable over Shandong Province, generally meeting the benchmark [55], except for the mean gross of wind direction at 10 m. Meanwhile, the relatively large gross error in wind direction at 10 m-i.e., 44.0° for Shandong Province and 46.8° for Qingdao-may be caused by values close to 0 or 360 degrees [56].  The model evaluation was also conducted between observed and simulated MDA8 ozone concentrations in Shandong Province each year during summer 2014-2019 ( Figure 5). The mean fractional bias (MFB) and mean fractional error (MFE) during 2014-2019 were less than the benchmarks of 15% and 35% applied from a previous study [57], respectively, indicating that the simulated MDA8 ozone concentrations in the CTRL experiment can reproduce the observed MDA8 ozone concentrations in Shandong Province.  Figure 6 shows the MDA8 ozone concentrations from biogenic emissions (see Section 2) during summer 2014-2019. The ozone enhancement caused by biogenic emissions was highest in July-close to 10 ppb in the west and north of Shandong Province. The ozone concentrations in the east and south of Shandong-particularly coastal cities-were relatively low, ranging from 3 to 5 ppb. In June and August, the ozone concentrations from biogenic emissions were relatively uniform in coastal and inland areas, ranging from 3 to 4 ppb. The concentrations in the inland area were about 1 ppb higher than those in the coastal area. The seasonal average of ozone concentrations in Qingdao from biogenic emissions during summer 2014-2019 was 5.5 ppb, with the higher percentile-i.e., 95th percentile-reaching 15.5 ppb, further aggravating the ozone pollution in Qingdao. Please note that the selection of the 95th percentile is traditionally common in previous studies [44,58].  Figure 6 shows the MDA8 ozone concentrations from biogenic emissions (see Section 2) during summer 2014-2019. The ozone enhancement caused by biogenic emissions was highest in July-close to 10 ppb in the west and north of Shandong Province. The ozone concentrations in the east and south of Shandong-particularly coastal cities-were relatively low, ranging from 3 to 5 ppb. In June and August, the ozone concentrations from biogenic emissions were relatively uniform in coastal and inland areas, ranging from 3 to 4 ppb. The concentrations in the inland area were about 1 ppb higher than those in the coastal area. The seasonal average of ozone concentrations in Qingdao from biogenic emissions during summer 2014-2019 was 5.5 ppb, with the higher percentile-i.e., 95th percentile-reaching 15.5 ppb, further aggravating the ozone pollution in Qingdao. Please note that the selection of the 95th percentile is traditionally common in previous studies [44,58].  To unveil the effects of isoprene emissions-either in or outside Shandong Province-on ozone pollution in Shandong Province, the ozone episode (18-28 July 2017) with the largest contribution from isoprene emissions during 2014-2019-i.e., more than 10 ppb averaged in Shandong Province continuously for 11 days (Figure 8a)-was selected. We designed two additional scenarios during this period, by turning off the isoprene emissions in and outside Shandong Province. These two scenarios were each compared with the case SEN_ASB, so as to elucidate the effects of isoprene emissions in or outside of Shandong Province on ozone concentrations in Shandong Province. Considering that isoprene is generally the main biogenic VOC contributing to ozone production [59,60], we plotted the isoprene emissions during summer 2014-2019 (Figure 7a), finding that the isoprene emissions in the majority of Shandong Province were relatively low, i.e., 0.2 mol/s or less during summer 2014-2019 (Figure 7a). In contrast, regions around Shandong Province-such as the North China Plain and the Yangtze River Delta-had relatively high isoprene emissions. The primary reason for the low isoprene emissions in Shandong Province lies in the dominant vegetation type of crops (Figure 7b), which mainly emit methanol and other reactive VOCs, but have low isoprene emission rates [48,61].  Considering that isoprene is generally the main biogenic VOC contributing to ozone production [59,60]  To unveil the effects of isoprene emissions-either in or outside Shandong Province-on ozone pollution in Shandong Province, the ozone episode (18-28 July 2017) with the largest contribution from isoprene emissions during 2014-2019-i.e., more than 10 ppb averaged in Shandong Province continuously for 11 days (Figure 8a)-was selected. We designed two additional scenarios during this period, by turning off the isoprene emissions in and outside Shandong Province. These two scenarios were each compared with the case SEN_ASB, so as to elucidate the effects of isoprene emissions in or outside of Shandong Province on ozone concentrations in Shandong Province. To unveil the effects of isoprene emissions-either in or outside Shandong Provinceon ozone pollution in Shandong Province, the ozone episode (18-28 July 2017) with the largest contribution from isoprene emissions during 2014-2019-i.e., more than 10 ppb averaged in Shandong Province continuously for 11 days (Figure 8a)-was selected. We designed two additional scenarios during this period, by turning off the isoprene emissions in and outside Shandong Province. These two scenarios were each compared with the case SEN_ASB, so as to elucidate the effects of isoprene emissions in or outside of Shandong Province on ozone concentrations in Shandong Province. The contribution of isoprene emissions in Shandong to MDA8 ozone concentrations therein was generally less than 1 ppb during this period (Figure 8b), due to low isoprene emissions, whereas natural forests in southern China emit large amounts of isoprene [63], accounting for approximately half of ozone enhancement from biogenic VOCs in southern China (Figure 8a vs. Figure 8c). Moreover, the enhanced ozone outside Shandong Province due to isoprene emissions is likely transported to Shandong Province ( Figure  8c), with the detailed transport pathway discussed below.
To reveal the effect of regional transport on the ozone concentrations contributed by biogenic emissions in Shandong Province, the hourly ozone concentrations during 09:00-16:00 local time due to isoprene emissions outside Shandong Province, as well as an 850 hPa wind vector, were composited during 18 July to 28 July 2017, and are presented in Figure 9. During these 11 days, eastern China was controlled by the Northwest Pacific Subtropical High. Dominated by the prevailing southwesterly wind, the ozone produced by biogenic emissions was transported from the southern flank of Shandong Province to Shandong Province. Therefore, biogenic emissions in regions around Shandong Province can affect Shandong Province through regional transport under favorable meteorological conditions. The contribution of isoprene emissions in Shandong to MDA8 ozone concentrations therein was generally less than 1 ppb during this period (Figure 8b), due to low isoprene emissions, whereas natural forests in southern China emit large amounts of isoprene [63], accounting for approximately half of ozone enhancement from biogenic VOCs in southern China (Figure 8a vs. Figure 8c). Moreover, the enhanced ozone outside Shandong Province due to isoprene emissions is likely transported to Shandong Province (Figure 8c), with the detailed transport pathway discussed below.
To reveal the effect of regional transport on the ozone concentrations contributed by biogenic emissions in Shandong Province, the hourly ozone concentrations during 09:00-16:00 local time due to isoprene emissions outside Shandong Province, as well as an 850 hPa wind vector, were composited during 18 July to 28 July 2017, and are presented in Figure 9. During these 11 days, eastern China was controlled by the Northwest Pacific Subtropical High. Dominated by the prevailing southwesterly wind, the ozone produced by biogenic emissions was transported from the southern flank of Shandong Province to Shandong Province. Therefore, biogenic emissions in regions around Shandong Province can affect Shandong Province through regional transport under favorable meteorological conditions. The distribution of MDA8 ozone concentrations contributed by ship emissions (see Section 2) was analyzed in Shandong Province and the nearby seas ( Figure 10). Ozone concentrations affected by ship emissions over the Yellow Sea were the highest in June, followed by July and August. In June, the MDA8 ozone concentrations over the Yellow Sea exceeded 10 ppb, gradually decreasing over the land, such as over Shandong Province. In terms of the land areas affected by ship emissions, the largest period occurred in August. Zooming into the port city of Qingdao, the average MDA8 ozone concentrations due to ship emissions were 2.4 ppb, with the 95th percentile of 8.7 ppb under favorable wind directions. In particular, when easterlies are dominant, ship emissions are prone to enhancing ozone concentrations in Qingdao.
Overall, biogenic and ship emissions play important roles in ozone concentration enhancement under favorable weather conditions, paving the way for ozone control and prediction. Under a changing climate-in particular of the increase in the frequency and intensity of extreme weather events, i.e., heat waves and stagnation [58,64]-it is of great importance to further examine how the interactions between extreme weather events and emissions-such as biogenic emissions-modulate the future ozone concentrations, as well as the associated uncertainties in constraining the signal in future changes. The distribution of MDA8 ozone concentrations contributed by ship emissions (see Section 2) was analyzed in Shandong Province and the nearby seas ( Figure 10). Ozone concentrations affected by ship emissions over the Yellow Sea were the highest in June, followed by July and August. In June, the MDA8 ozone concentrations over the Yellow Sea exceeded 10 ppb, gradually decreasing over the land, such as over Shandong Province. In terms of the land areas affected by ship emissions, the largest period occurred in August. Zooming into the port city of Qingdao, the average MDA8 ozone concentrations due to ship emissions were 2.4 ppb, with the 95th percentile of 8.7 ppb under favorable wind directions. In particular, when easterlies are dominant, ship emissions are prone to enhancing ozone concentrations in Qingdao. Overall, biogenic and ship emissions play important roles in ozone concentration enhancement under favorable weather conditions, paving the way for ozone control and prediction. Under a changing climate-in particular of the increase in the frequency and  The distribution of MDA8 ozone concentrations contributed by ship emissions (see Section 2) was analyzed in Shandong Province and the nearby seas ( Figure 10). Ozone concentrations affected by ship emissions over the Yellow Sea were the highest in June, followed by July and August. In June, the MDA8 ozone concentrations over the Yellow Sea exceeded 10 ppb, gradually decreasing over the land, such as over Shandong Province. In terms of the land areas affected by ship emissions, the largest period occurred in August. Zooming into the port city of Qingdao, the average MDA8 ozone concentrations due to ship emissions were 2.4 ppb, with the 95th percentile of 8.7 ppb under favorable wind directions. In particular, when easterlies are dominant, ship emissions are prone to enhancing ozone concentrations in Qingdao.

Conclusions and Discussions
This study investigated the underlying mechanisms affecting the surface ozone concentrations in Shandong Province, including the major port city of Qingdao, during summer 2014-2019, by utilizing observation, ERA5 reanalysis datasets, and WRF-CMAQ modeling. The summer ozone concentrations in Qingdao were the highest in June, consistent with but lower than those in other inland cities in Shandong Province. The MDA8 ozone concentrations over Qingdao during summer 2014-2019 were highest in 2019, with 30 days exceeding 82 ppb, in contrast to an average of 9 days during summers in 2014-2018.
Surface shortwave radiation and relative humidity were most closely related to MDA8 ozone concentrations in Shandong Province. The high-pressure ridge appearing over Lake Baikal and to its northeast resulted in relatively low humidity and sufficient sunlight in Qingdao, critical for the occurrence of local high-ozone events.
The ozone concentrations contributed by biogenic emissions were highest in July during 2014-2019-close to 10 ppb in the west and north of Shandong Province. Numerical experiments indicate that the local impact of biogenic emissions on ozone in Shandong Province is relatively small. In contrast, the biogenic emissions over the southern flank of Shandong Province trigger ozone enhancement and further transport northeastward, leading to an increase in ozone concentration in Shandong Province. Additionally, in Qingdao, the ozone concentrations due to ship emissions are associated with sea breeze.  Data Availability Statement: The ERA5 reanalysis dataset was downloaded from the ECMWF (https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels-preliminary-backextension?tab=overview; accessed on 2 February 2022). Ozone observations were obtained from the China National Environmental Monitoring Centre (http://www.pm25.in; accessed on 2 February 2022). Other data are available upon request from the corresponding author (yanggao@ouc.edu.cn).

Conflicts of Interest:
The authors declare no conflict of interest.