Impacts of NO₂ on Urban Air Quality and Causes of Its High Ambient Levels: Insights from a Relatively Long-Term Data Analysis in a Typical Petrochemical City in the Bohai Bay Region, China

Abstract

The ambient levels of NO₂ in urban areas in China in recent years have generally shown a downward trend, but high NO₂ concentrations still exist under certain conditions, and the causes for such phenomenon and its impact on air quality remain unclear. Taking Dongying, a typical petrochemical city in the Bohai Bay of China, as an example, this paper analyzed the influence of NO₂ on urban air quality and investigated the causes for the formation of NO₂ with high concentrations. The results indicated that higher daily NO₂ concentrations (>40 μg/m³) mainly occurred during January-April and September-December each year, and higher hourly NO₂ concentrations mainly occurred during the nighttime and morning rush hour in Dongying from 2017 to 2023. With the increase in daily NO₂ concentrations, the daily air pollution levels showed a general increasing trend from 2017 to 2023. The occurrence of high NO₂ values in Dongying was affected by the combination of unfavorable meteorological conditions, local emissions and regional transports, and localized atmospheric chemical generation. High-pressure and uniform-pressure weather patterns in 2017–2022, along with land–sea breeze circulation in 2022, contribute to high NO₂ concentrations in Dongying. Boundary layer heights (BLH) in spring (−0.43) and winter (−0.36), wind direction in summer (0.21), and temperature in autumn (−0.46) are the primary meteorological factors driving NO₂-HH (High hourly NO₂ values), while BLH (−0.47) is the main cause for NO₂-HD (High daily NO₂ values). The titration reaction between NO with O₃ is the main cause for NO₂-HH in spring, summer and autumn, and photochemical reactions of aromatics have a significant influence on NO₂-HD. NOx emissions from the thermal power and petrochemical industry in Dongying and air pollution transports from western and southwestern Shandong Province (throughout the year) and from the Bohai Sea (during spring and summer) had serious adverse impact on high NO₂ values in 2022. The results of the study could help to provide a scientific basis for the control of NO₂ and the continuous improvement of air quality in Dongying and similar petrochemical cities.

1. Introduction

Atmospheric nitrogen oxides (NOx) include compounds such as nitrous oxide (N₂O), nitric oxide (NO), nitrogen dioxide (NO₂), dinitrogen trioxide (N₂O₃), and dinitrogen tetroxide (N₂O₄). Among these, NO₂ is one of the more stable nitrogen oxides, with higher concentration and longer persistence, playing a significant role in atmospheric chemical reactions) [1]. The formation of NO₂ in the atmosphere is primarily related to the conversion of NO. During the day, the photochemical chain reactions involving volatile organic compounds (VOCs) are key to NO₂ generation, while at night, the reaction of ozone (O₃) with NO is the main pathway for NO₂ generation (Figure S1). The photochemical reactions between NO₂ and VOCs during the day are the only chemical source of O₃ [2]. The reaction between NO₂ and hydroxyl radicals (OH) is the primary source of gaseous nitric acid (HNO₃), and the reactions between NO₂ and alkoxy radicals (RO) are the main sources of organic nitrate compounds in the atmosphere. NO₂ reacts with O₃ to form nitrate radical (NO₃) at night when NO concentration is low. NO₃ can react with VOCs or with NO₂ to form dinitrogen pentoxide (N₂O₅), which then reacts with water or aerosols to produce nitric acid (HNO₃) or nitrate salts (NO₃⁻), ultimately leading to the net removal of NO₂ [3]. NO₂ and its related products, such as O₃ and organic nitrate compounds, are key components of photochemical smog, while HNO₃ or NO₃⁻ are major contributors to acid rain. These pollutants pose potential risks to human health and the eco-environment, they could damage human cardiovascular and pulmonary functions, and other organisms in ecosystems; In addition, photochemical smog could irritate the eye membrane of humans, reduce atmospheric visibility, and contribute to global warming, and acid rain could lead to soil acidification and building corrosion [4,5].

Therefore, studying the pollution characteristics of NO₂, its impact on air quality, and the causes of high NO₂ concentrations is crucial for further understanding atmospheric chemical mechanisms, continuously improving air quality, and reducing its adverse effects on human health and the environment.

The distribution pattern of NO₂ in the atmosphere results from the combined effects of pollution sources, meteorological conditions, and chemical generation. NO₂ sources are mainly divided into natural and anthropogenic sources [6]. Natural sources include lightning, soil emissions, and ammonia oxidation, while anthropogenic sources include industrial facilities and mobile emissions. Ambient NO₂ concentrations are greatly influenced by anthropogenic sources [7,8], which is the primary causes for the “higher in the east, lower in the west” distribution pattern observed in China over the years [9]. The average NO₂ concentration in autumn and winter are relatively higher than in spring and summer [10]. The seasonal differences in anthropogenic NOx emissions and meteorological conditions are important factors contributing to the seasonal variation in NO₂ concentration [11]. The diurnal variation of NO₂ concentration usually shows a “two peaks and one valley” pattern. The morning peak can be attributed to motor vehicle emissions, while the evening peak may be related to unfavorable weather conditions and the conversion of NO with high concentrations during the evening rush period [12]. Additionally, NO₂ exhibits a certain transport capacity [13]. Cheng et al. [14] found that implementing an odd-even license plate policy in the surrounding areas of Beijing significantly reduced NO₂ concentration in the city. Wang et al. [15] found that controlling NOx emissions in upwind cities is crucial for mitigating O₃ pollution in specific cities. Relevant studies have found that meteorological conditions, especially wind speeds (WS) and wind direction (WD), often affect the regional transports of NO₂, leading to high urban NO₂ concentrations [16,17]. High NO₂ concentrations at low WS are influenced by nearby pollution sources, while at high WS are influenced by long-distance transport. Xiao et al. [18] found that under weak large-scale meteorological conditions, local circulations related to terrain can also affect pollutant concentrations. The land–sea breeze characteristics in the Bohai Bay region are distinct [19]. During pollution events, offshore land winds can transport pollutants from Shandong Province to the Bohai Sea, while onshore sea winds can create a convergence zone, inducing upward air currents that cause pollutants to cycle over the bay, further deteriorating air quality in coastal areas [20]. Therefore, to fully understand the pollution characteristics of NO₂ and determine the causes for the formation of NO₂ with high concentrations, it is necessary to investigate from multiple perspectives.

Over the past 20 years, NOx emissions in China have initially increased and then decreased annually, peaking in 2012. Although NOx emissions in China have decreased, the emission structure in many regions has shifted to being dominated by traffic sources [21,22,23]. However, the overall NOx emissions remain above ten million tons, leading to persistently high ambient NO₂ concentrations in China [24]. The China Ecological Environment Status Bulletin shows that after a sharp decline in 2014 (34.8 µg/m³), NO₂ concentrations in China’s urban areas fluctuated from 2015 to 2019 (27.5–28.4 µg/m³), continued to decline from 2019 (27 µg/m³) to 2022 (21 µg/m³) but rebounded in 2023 (22 µg/m³) with a year-on-year increase of 4.8% [25].Currently, high NO₂ values in China are concentrated in economically developed and densely populated regions of North China, East China, and Central China [26], primarily in autumn and winter. For example, the monthly average NO₂ concentration in Jinan, Shandong Province, exceeded 60 μg/m³ in January from 2015 to 2021 [27]. Some cities also experienced peak concentration at night during the summer. In 2018, four prefecture-level cities in Shanxi Province experienced pollution peaks during summer, with the average hourly NO₂ concentration in Lvliang exceeding 80 μg/m³ at night [28]. Besides O₃, PM₂.₅, and PM₁₀, NO₂ remains one of the primary pollutants on days with air quality exceedance in urban areas of China [25,29]. For example, the number of days with NO₂ as the primary pollutant and the number of NO₂ exceedance days increased annually from 2018 to 2020 in Chuzhou, Anhui Province, in the Yangtze River Delta region. In 2020, the proportion of days with NO₂ as the primary pollutant reached 5.2–9.0% in coastal cities like Qingdao, Binzhou, Rizhao, and Dongying in Shandong Province [30,31].

Dongying City, located in the northwest of Shandong Province and bordered by Bohai Bay to the north and Laizhou Bay to the east, is one of the representative cities in the Bohai Bay of China. Dongying has a developed petrochemical industry. Transportation primarily relies on road traffic, with extensive motor vehicle use and a high number of trucks in urban areas, leading to substantial NO_X and VOC emissions in Dongying [32], resulting in various environmental issues, such as increased photochemical pollution and frequent smoggy weather [33]. Previous studies on air pollution in Dongying have mainly focused on aerosols, particulate matter, O₃, and their precursors [31,32]. Research on NO₂ has primarily concentrated on pollution characteristics [8,15], with few comprehensive analyses of the impacts and causes of high NO₂ values. This study analyzed the temporal and spatial distribution characteristics of NO₂ concentrations in Dongying based on atmospheric pollutant and meteorological data from 2017 to 2023. The impact of NO₂ on air quality and of high NO₂ values in 2022 on O₃, PM₂.₅, PM₂.₅-related secondary components, and atmospheric oxidation capacity were explored, and the causes of high NO₂ values were investigated. The study results could provide a theoretical basis for NO₂ pollution control and air quality improvement in Dongying and similar petrochemical cities.

2. Materials and Methods

2.1. Study Area and Data Sources

This study focuses on Dongying, a representative city in the Bohai Bay region, China. Dongying is located in the northwest of Shandong Province, bordered by Bohai to the east and north, adjacent to Binzhou City to the west, and neighboring Zibo and Weifang cities to the south (Figure 1a). As a typical resource-based city, Dongying is dominated by heavy industries (petroleum and chemicals), with industrial parks concentrated in the southern region. Consequently, the air quality in Dongying is easily influenced by both maritime and inland pollutant transports. Dongying has jurisdiction over five counties/districts: Dongying District, Hekou District, Kenli District, Lijin County, and Guangrao County. Currently, the city has one Super Atmospheric Observation Station (referred to as “Atmospheric Observatory”), eight state-controlled stations, and four provincial-controlled stations. The Hekou Urban Area and Hekou Development Zone stations are in the northern region, Minfeng Lake and Minfeng Road stations in Kenli District are in the central region, and the remaining stations are in the southern region (Figure 1b).

The routine observation items include NO₂ and meteorological parameters—temperature (T), relative humidity (RH), WS, WD, atmospheric pressure (AP), and boundary layer heights (BLH). The enhanced observation items include routine monitoring parameters (NO₂, NO, SO₂, PM₂.₅, O₃), 115 VOCs, OC (organic carbon), EC (elemental carbon), ionic components (NO₃⁻ and SO₄²⁻), ultraviolet radiation (UR), and emissions data. All observation items were continuously monitored using automatic monitoring devices, with monitoring station, frequency and duration of data collection, quality assurance, and quality control adhering to various technical specifications. Details regarding data usage, sources, and quality control in this study can be found in Table S1, Text S1, and Text S2, respectively.

2.2. Related Definitions

NO₂ evaluation standards: According to the Technical Regulation for Ambient Air Quality Assessment (on Trial) (HJ 663-2013), the daily NO₂ assessment value is the average NO₂ concentration over a 24 h period, the monthly NO₂ assessment value is the average NO₂ concentration for the month, and the annual NO₂ assessment value is the average NO₂ concentration for the calendar year. According to the Technical Regulation on Ambient Air Quality Index (on Trial) (HJ 633-2012), the daily NO₂ concentration corresponding to the “Excellent” and “Good” air quality levels are 0–40 μg/m³ and 40–80 μg/m³, respectively. The daily assessment limit for NO₂ under the National Ambient Air Quality Secondary Standard is 80 μg/m³. The World Health Organization (WHO) recommends annual NO₂ Interim Target 2: IT-2 (30 μg/m³) and IT-3 (20 μg/m³). For 24 h NO₂, IT-2 is 50 μg/m³, and the 2021 Air Quality Guideline (AQG) level is 25 μg/m³ [34].

The significance of NO₂ at different percentiles: Analyzing the arithmetic mean of NO₂ is complemented by evaluating NO₂ at different percentiles to provide a more comprehensive and detailed description of NO₂ pollution characteristics [35,36]. Generally, low percentiles of NO₂ (5th percentile) represent background concentration, while high percentiles (99th and 95th percentiles) indicate concentration during severe NO₂ pollution events [37]. Middle percentiles (75th, 50th, and 25th percentiles) reflect the NO₂ concentration observed during most periods in Dongying.

High daily NO₂ values (NO₂-HD): NO₂-HD is defined as daily average concentration exceeding the 75th percentile for each season in 2022, with seasonal thresholds of >23 μg/m³ (spring), >18 μg/m³ (summer), >45 μg/m³ (autumn), and >44 μg/m³ (winter) (Figure S2a). NO₂-NHD is defined as the daily average concentration not exceeding the 75th percentile.

High hourly NO₂ values (NO₂-HH): NO₂-HH is defined as hourly concentration exceeding the 75th percentile for each season in 2022, with seasonal thresholds of >25 μg/m³ (spring), >19 μg/m³ (summer), >47 μg/m³ (autumn), and >49 μg/m³ (winter) (Figure S2b).

NO₂-HH events and NO₂-NHH periods: Based on diurnal variation characteristics of NO₂, NO₂-HH typically occurs during nighttime to morning rush hours. NO₂-HH events are defined as at least five consecutive NO₂-HH occurrences during this period (with spring and summer from 00:00 to 09:00 and 20:00 to 23:00, and autumn and winter from 00:00 to 10:00 and 19:00 to 23:00) (Figure S8). Periods without five consecutive NO₂-HH occurrences are defined as NO₂-NHH periods. In addition, the remaining hours of each season are referred to as the daytime (with spring and summer from 10:00 to 19:00, and autumn and winter from 11:00 to 18:00).

2.3. Analytical Methods

2.3.1. Methods for Analyzing Impact of NO₂ on Air Quality

In this study, when analyzing the impact of NO₂ on air pollution levels in Dongying, the number of mildly, moderately, heavily, and severely polluted days refers to the total number of polluted days where various basic pollutants were the primary pollutants, in accordance with the Technical Regulation on Ambient Air Quality Index (on Trial) (HJ 633-2012). The proportion of polluted days was then calculated as the ratio of the number of polluted days corresponding to a specific NO₂ concentration range to the total number of all days in that range. To explore the impact of high NO₂ values on different pollutants in detail, analyses were conducted from the perspectives of NO₂-HH and NO₂-HD in different seasons of Dongying. The analytical framework is detailed in Figure S3, and the main analytical methods are as follows:

Calculation of Secondary Organic Carbon (SOC): The concentration of SOC can be calculated using the OC/EC ratio method. The calculation formula is as follows:

SOC = OC − EC × (OC/EC)_min

(1)

In the formula, SOC represents secondary organic carbon (µg/m³); OC and EC denote organic carbon and elemental carbon, respectively, and (OC/EC)_min represents the minimum value of the OC/EC ratio observed during the study period.

Evaluation indicators for atmospheric oxidation capacity: The atmospheric oxidation capacity is typically represented by O_X concentration, which is generally estimated using the concentration of NO₂ and O₃ (NO₂ + O₃). NO₃ is an important nocturnal atmospheric oxidant, and the reaction between NO₃ and NO₂ is the only pathway for the formation of N₂O₅ [38]. The changes in N₂O₅ concentration levels can be indicated by the nighttime O₃ concentration multiplied by the square of the NO₂ concentration ([NO₂]² × O₃), which indirectly represents the nighttime atmospheric oxidation capacity.

Transformation Rate of Nitrates and Sulfates: The nitrogen oxidation rate (NOR) and sulfur oxidation rate (SOR) can be used to assess the conversion status of gaseous precursors such as NO₂ and SO₂ into secondary inorganic ions. Higher values of NOR and SOR indicate a greater extent of secondary conversion of NO₂ and SO₂ in the atmosphere. The formulas are as follows:

NOR = N₁/(N₁ + N₂)

(2)

SOR = S₁/(S₁ + S₂)

(3)

In the formula, N₁ and N₂ represent the concentration of NO₃⁻ and NO₂, respectively (mol/m³); S₁ and S₂ represent the concentration of SO₄²⁻ and SO₂, respectively (mol/m³).

2.3.2. Methods for Analyzing the Causes of High NO₂ Values

The NO₂ concentration distribution in 2022 is not significantly different from that in recent years. Additionally, the Atmospheric Observatory and meteorological station data for 2022 are relatively comprehensive. Therefore, this study selects 2022 for analyzing the causes of high NO₂ levels. The analytical framework is detailed in Figure S4, and the main analytical methods are as follows:

Synoptic classification method: This study utilized the synoptic classification software cost733class developed by the European Union COST733 Project (http://www.cost733.org, accessed on 18 March 2024) and employed T-mode principal component Analysis (PCT) method to perform objective synoptic classification and reveal spatial distribution of the 925 hPa geopotential height field and the horizontal full WS (U and V) from 2017 to 2022. Details regarding the synoptic classification method in this study can be found in Text S3.

Criteria for sea–land breeze identification: Based on domestic and international studies on sea–land breeze classification standards [39,40], along with the natural geographic features of Dongying’s coastline, the criteria for identifying sea–land breeze events in Dongying are established as follows. Based on the orientation of the coastline, airflow within the SSE to NW range (157.5–315°) is defined as a land breeze, while airflow within the NNW to SE range (0–135° and 337.5–360°) is defined as a sea breeze (Figure 1a). A day is classified as a sea–land breeze day if the following criteria are met: (1) the 24 h average surface wind speed is below 10 m·s⁻¹; (2) during the land breeze period (01:00–08:00), the land breeze duration is ≥4 h, while the sea breeze duration is ≤2 h; and (3) during the sea breeze period (13:00–20:00), the sea breeze duration is ≥4 h, while the land breeze duration is ≤2 h.

Random forest model: The random forest model, introduced by Breiman in 2001, is a machine learning algorithm based on classification trees. This model is widely used for handling nonlinear relationships, classification, regression, high-order correlations, and variable importance assessment [41,42]. A linear regression model and a random forest model were constructed using Python 3.13. The models randomly selected 90% of the data as the training set and the remaining 10% as the test set. Model performance was evaluated by calculating the coefficient of determination (R²) and root mean squared error (RMSE) [43]. Furthermore, the mean decrease accuracy (MDA) method was employed to evaluate the importance of each influencing factor. The principle behind MDA is that each variable is randomly assigned new values; if a variable is more important, replacing it randomly will result in a larger prediction error [44].

HYSPLIT mode: The Hybrid Single-Particle Lagrangian Integrated Trajectory Model (HYSPLIT) is primarily used to analyze air mass trajectories and trace the sources of air pollutants. In this study, the Dongying Atmospheric Observatory (37.45° N, 118.59° E) was selected as the starting point for the backward trajectories, with the analysis period covering the four seasons of 2022. The air mass analysis was conducted at an altitude of 300 m, with a backward tracing duration of 24 h. A total of 24 trajectories were calculated per day, resulting in 2208, 2206, 2184, and 2160 effective trajectories for each season. More details can be found in Text S4.

Backward trajectory clustering analysis method: Backward trajectory clustering analysis uses mathematical methods to classify and cluster all air parcel trajectories within a set time based on transport speed and direction. This helps analyze the source and proportion of dominant air masses at the target location, thereby determining the primary pollution source direction [45]. After cluster analysis, this study statistically analyzed the trajectory types and corresponding NO₂ hourly concentration data for each season, and subsequently determined the mean NO₂ concentration and contribution percentage for each trajectory type. More details can be found in Text S5. The formula for calculating the contribution percentage of NO₂ concentration is as follows.

$${\mathrm{P}}_{{\mathrm{N}\mathrm{O}}_2,\mathrm{t}}={\displaystyle \frac{\sum {\mathrm{C}}_{\mathrm{t}}}{\sum \mathrm{C}}}\times 100\mathrm{\%}$$

(4)

$\mathrm{t}$ is the category corresponding to the trajectory, ${\mathrm{C}}_{\mathrm{t}}$ is the hourly mass concentration of NO₂ corresponding to the trajectory with category t, $\sum {\mathrm{C}}_{\mathrm{t}}$ is the sum of hourly mass concentrations of NO₂ corresponding to the trajectory with category $\mathrm{t}$, $\sum \mathrm{C}$ is the sum of hourly mass concentrations of NO₂ of all the trajectories, and ${\mathrm{P}}_{{\mathrm{N}\mathrm{O}}_2,\mathrm{t}}$ is the contribution of NO₂ concentration corresponding to the category of trajectory, $\mathrm{t}$.

Potential source contribution function (PSCF) analysis method: PSCF is a method based on conditional probability functions to identify potential pollution sources [46]. This method relies on backward trajectory results and utilizes a series of functions in the TrajStat plugin to calculate WPSCF values. In this study, NO₂-HH in each season were set as the threshold for the target pollutant. The higher the WPSCF value in an area, the darker the color in the generated graphics, indicating a higher probability that the area contributes to NO₂-HH at the target site. More details can be found in Text S6.

Concentration weight trajectory (CWT) analysis method: The CWT method calculates the weighted concentration of trajectories to quantitatively reflect the pollution severity of different trajectories [47], thus determining the impact of regional transport on NO₂ concentration in Dongying. The higher the WCWT value, the more pollutants transported by the trajectory passing through the grid to the target site, and the greater the contribution to NO₂ concentration at the target site [48]. More details can be found in Text S7.

3. Results and Discussion

3.1. NO₂ Concentration Levels and Variation Characteristics

3.1.1. Concentration Levels

From 2017 to 2023, the annual mean values of NO₂ in Dongying showed a trend of “increase-decrease-stabilization” (Figure 2). The 5th, 25th, 50th, and 75th percentiles of NO₂ in Dongying exhibited a clear downward trend overall, except for a rise in 2019. The 95th and 99th percentiles of NO₂ displayed a fluctuating trend of “increase–decrease–increase,” indicating a significant difference in the trends between low and high percentiles of NO₂. Additionally, in 2023, while the low and middle percentiles continued to decline and the high percentiles rebounded, the annual average remained relatively stable. This indicates that the high NO₂ concentration issue in Dongying has not been resolved. The NO₂ concentration at the 95th percentile in Dongying each year was consistently higher than that at the 75th percentile, further indicating a high NO₂ concentration issue. Notably, in 2019, the annual mean values of NO₂ were the highest, and both the 75th and 95th percentiles were significantly elevated compared to other years. This suggests that the high NO₂ concentration in 2019 may have been driven by increases in the middle and high percentiles. In recent years, the annual NO₂ concentration in Dongying has met IT-2 (30 μg/m³), but there is still a significant gap in reaching IT-3 (20 μg/m³). In summary, Dongying currently faces a significant issue with high NO₂ concentration, and reducing the high values could notably lower the overall NO₂ concentration levels in the city.

3.1.2. Temporal Variation Characteristics

The monthly variation of NO₂ in Dongying exhibits a “U” distribution (Figure 3), which is consistent with the findings of most researchers [49,50]. Overall, NO₂ concentration was highest in December, followed by November, October, and January. A significant decline in NO₂ concentration was observed each year from January to February. The period from March to August marked a rapid decrease in NO₂ levels, reaching the lowest values in July and August, while the period from August to October shows a rapid increase. The calendar of NO₂ in Dongying from 2017 to 2023 (Figure 4) shows that NO₂ with high concentrations (>40 μg/m³) was concentrated from January to mid-March and from September to late December in 2017, with exceedances (>80 μg/m³) occurring in late October. Compared to 2017, NO₂ with high concentrations in 2018 was concentrated from October to December and March. The days in January and February were more dispersed, but days where exceedances occurred in mid-January and were concentrated. In 2019, compared to 2018, the days of NO₂ with high concentrations increased and became more concentrated, with exceedance days mainly occurring in January and being more dispersed. From 2020 to 2023, the days of NO₂ with high concentrations significantly decreased and became more dispersed. In 2023, Dongying still experienced 46 days with NO₂ concentrations exceeding IT-2 (50 μg/m³) and 159 days not meeting the 2021 AQG level (25 μg/m³), indicating a relatively high daily average NO₂ level in the city. Overall, the days of NO₂ with high concentrations in Dongying exhibited a distribution pattern of “few and dispersed days—increasing and concentrated days—gradually decreasing and more dispersed days” from 2017 to 2023. High NO₂ concentrations mainly occurred from September to December and January to April, with exceedances usually happening in late October, November, December, and January.

From 2017 to 2023, the concentrations of NO₂ during high NO₂ concentrations periods (from September to December and January to April) and non-high NO₂ concentrations periods, and the annual mean value of NO₂ concentration in Dongying all exhibited a “single peak and single valley” pattern (Figure 5 and Figure S5). The peak occurred around 8:00 a.m., and the valley appeared around 2:00 PM. During the morning traffic rush (6–9 a.m.), NO₂ concentration slightly increased, then slowly decreased after 9 a.m. After 4:00 p.m., NO₂ concentration rose significantly, remained at high levels, and exhibited a slow upward trend during nighttime to morning rush hours. Most research findings indicate that daily NO₂ variations exhibit a “double peak and single valley” pattern, with distinct morning and evening peaks [23,51]. In comparison, the hourly NO₂ concentration in Dongying did not show a significant peak at night, and the morning peak was smaller, indicating that in addition to vehicle emissions, other pollution sources also influence NO₂ concentration in Dongying. Overall, from 2017 to 2023, the concentrations of NO₂ during high NO₂ concentrations periods and non-high NO₂ concentrations periods, and the annual mean value of NO₂ concentration all showed a downward trend. However, changes in NO₂ concentration mainly occurred during nighttime to morning rush hours, with minimal changes during daytime hours.

3.1.3. Spatial Variation Characteristics

The interannual variation of NO₂ concentration in different regions of Dongying from 2017 to 2023 is shown in Figure 6. From different monitoring sites, during non-high NO₂ concentrations periods, the highest average NO₂ concentration was at the Gengjiangcun site, and the lowest was at the Minfeng Lake site in Kenli District. During periods of high NO₂ concentrations, the highest average NO₂ concentration was also at the Gengjiangcun site, while the lowest was at the Hekou urban site. From the annual variation at different sites, the Gengjiangcun site consistently showed higher average NO₂ concentration, while the Hekou urban site had lower averages. From different regions, during non-high NO₂ concentrations periods, the southern region had the highest average NO₂ concentration (19 μg/m³), and the central region had the lowest (15 μg/m³). During high NO₂ concentrations periods, the southern region had the highest average NO₂ concentration (37 μg/m³), while the northern region had the lowest (34 μg/m³). From the annual variation at different regions, the southern region had the highest average NO₂ concentration (19 μg/m³), while the central region had the lowest (15 μg/m³). Overall, from the perspectives mentioned above, the NO₂ concentration levels in the southern region of Dongying are relatively higher compared to other regions.

3.2. Impact of NO₂ on Air Quality

3.2.1. Impact of NO₂ on Air Quality Levels

From the perspective of the primary pollutant (Figure S6), the number of days with NO₂ as the primary pollutant in Dongying was the lowest in 2017 (1 d), higher from 2018 to 2020 (14–20 days), and relatively lower from 2021 to 2023 (10–13 days) with a fluctuating trend. The proportion of days with NO₂ daily assessment value of “excellent” showed an overall increasing trend, reaching its lowest point in 2019 (64%) and experiencing a slight rebound in 2023. Exceedance days for NO₂ occurred between 2017 and 2020 (3–5 days) and in 2023 (1 day). Overall, as a pollutant in air quality assessment, NO₂ itself currently has a reduced impact on the air quality in Dongying.

According to Figure 7, from 2017 to 2019, the number of mildly polluted days generally increased and then gradually decreased with rising NO₂ concentration. From 2020 to 2023, the number of mildly polluted days steadily decreased with rising NO₂ concentration. The number of moderately polluted days showed an increase, followed by a decrease and then another increase with rising NO₂ concentration. Severe pollution days mainly occurred in 2017 (1 day) and 2021 (4 days), with corresponding NO₂ concentration ranges of 70–80 and 30–40 µg/m³, respectively. Unlike other polluted days, the number of severe pollution days is relatively small, and there is no obvious pattern of variation. From 2017 to 2023, the number of mildly, moderately, and heavily polluted days was higher when the NO₂ concentration were 10–30 μg/m³, 20–30 μg/m³, and 50–80 μg/m³, respectively. When the NO₂ concentration is below 20 μg/m³, the proportion of polluted days is relatively low (<25%). Within the 20–50 μg/m³ range, the proportion remained stable (32–35%). However, when the NO₂ concentration exceeded 60 μg/m³, the proportion of polluted days increased significantly (>60%). In summary, the overall air quality in Dongying showed that both the air pollution level and the proportion of polluted days increased with rising daily NO₂ concentrations.

3.2.2. Impact of High NO₂ Values on O₃, PM₂.₅, and Atmospheric Oxidation Capacity in 2022

According to Figure S7, the number of NO₂-HD days in Dongying in 2022 showed no significant seasonal variation, with similar counts in spring (21 days), summer (21 days), autumn (19 days), and winter (22 days). On a monthly scale, December and March had the highest number of NO₂-HD days (11 days), followed by June and November (9 days), while February, April, and September had the fewest (4 days), indicating some inter-monthly variability. As shown in Figure S8, the proportion of NO₂-HH in spring and summer peaked around 00:00–01:00 (8% and 7%) and again around 07:00 (8% and 10%). In contrast, NO₂-HH in autumn and winter remained relatively stable between 19:00 and 10:00, the following day, ranging from 5% to 7% and 4% to 6%, respectively. Overall, NO₂-HH in Dongying primarily occurred from nighttime to the morning peak hours (with spring and summer being 00:00–09:00 and 20:00–23:00, and autumn and winter 00:00–10:00 and 19:00–23:00).

As shown in Figure 8, compared to NO₂-NHH periods, NO₂-HH events in all seasons lead to significant increases in PM₂.₅ and PM_2.5-bounded NO₃⁻, SO₄²⁻, and SOC concentrations, and a notable decrease in O₃ concentration. This indicates that high NO₂ values can enhance the concentrations of O₃, PM_2.5 and PM_2.5-bounded secondary components during nighttime to morning rush hours, with the most pronounced effects in autumn and winter. The higher the NO₂ concentration in a season, the lower the nighttime O₃ concentration, which is related to the NOx titration reaction [2]. As shown in Figure 9, during NO₂-HH events across all seasons, the concentrations of O_X and N₂O₅ are significantly higher compared to NO₂-NHH periods, indicating that high NO₂ values can enhance the atmospheric oxidation capacity at night, consistent with the findings of Alberto Notario et al. [52]. In spring, summer, and winter, the NOR values decrease during NO₂-HH events, whereas the opposite is observed in autumn. Similarly, SOR values decrease during NO₂-HH events in spring, autumn, and winter, but increase in summer. This may be because the formation of NO₃⁻ and SO₄²⁻ at night is significantly influenced by nighttime O₃ concentration [53]. During NO₂-HH events, high NO concentration results in the titration of O₃ by NO and the concentrations of NO₂ and SO₂ are relatively high, leading to lower NOR and SOR values. In summary, NO₂-HH events can increase O₃, PM_2.5 and PM_2.5-bounded secondary components concentrations, as well as O_X and N₂O₅ concentrations during nighttime to morning rush hours. They also lead to lower O₃ concentration, NOR, and SOR values. Moreover, the analysis of impact of high NO₂ values on daytime O₃, PM_2.5, and atmospheric oxidation capacity in Dongying in 2022 is detailed in Figures S9 and S10 and Text S8.

As shown in Figure 10, compared to NO₂-NHD days, PM_2.5 and PM_2.5-bounded NO₃⁻, SO₄²⁻, and SOC concentrations are significantly higher on NO₂-HD days across all seasons, indicating that NO₂-HD can promote increases in concentrations of PM_2.5 and PM_2.5-bounded secondary components on the same day, with the most pronounced effect in autumn and winter. In spring, autumn, and winter, O₃ concentration on NO₂-HD days is lower than on NO₂-NHD days, whereas the opposite is true in summer. In spring, summer, and autumn, MDA8-O₃ concentration on NO₂-HD days is higher than on NO₂-NHD days, with the opposite observed in winter, likely due to higher daytime NO emissions in winter. As shown in Figure 11, O_X concentration is significantly higher on NO₂-HD days across all seasons, indicating that NO₂-HD can enhance the atmospheric oxidation capacity. In spring and summer, NOR values on NO₂-HD days are lower than on NO₂-NHD days, whereas the opposite is observed in autumn and winter. Conversely, SOR values are higher on NO₂-HD days in summer, autumn, and winter, but lower in spring, indicating a noticeable seasonal difference in the impact of NO₂-HD on NOR and SOR. In summary, NO₂-HD can increase MDA8-O₃, PM_2.5 and PM_2.5-bounded secondary component concentrations, as well as O_X concentration, while reducing the average O₃ concentration on the same day.

3.3. Causes of High NO₂ Values

3.3.1. Meteorological Conditions

Using the PCT method, the circulations of the 925-hPa geopotential height field from 2017 to 2022 were classified into nine patterns (Figure S11)—namely, “offshore high-pressure rear” (Type 1), “high-pressure front” (Type 2), “high-pressure inside” (Type 3), “high-pressure top front” (Type 4), “uniform pressure” (Type 5), “west-high and east-low” (Type 6), “subtropical high” (Type 7), “inverted trough” (Type 8), and “low-pressure field” (Type 9).

From 2017 to 2022, Type 1 occurs most frequently (20.5%), with large variations in the daily average maximum temperature (T_max-mean) and daily average relative humidity (RH-mean). Type 4 (16.6%), Type 2 (14.1%), and Type 3 (12.1%) followed, with these patterns generally characterized by low T_max-mean and high RH-mean. Type 6 (8.4%) and Type 5 (6.7%) exhibited large ranges in T_max-mean, with the former having sparse isobars and lower RH-mean, and the latter having dense isobars and higher RH-mean. Type 7 (9.5%), Type 8 (7.6%), and Type 9 (4.4%) mainly occur in summer, characterized by higher T_max-mean, occasional precipitation, and higher RH-mean (

Table 1. Frequency and major characteristics of nine synoptic patterns of the 925-hPa geopotential height field in Dongying from 2017 to 2022.

Synoptic Pattern	Frequency	Major Characteristics
Type1	20.5%	Dongying near the offshore high-pressure rear, with dominant southeasterly wind, Tmax-mean of 21.8 °C, RH-mean of 56.6%, and Pre-mean of 1.4 mm.
Type2	14.1%	High-pressure center over the Mongolian Plateau. Dongying under the front of the high-pressure, with dominant southeasterly wind, Tmax-mean of 10.4 °C, RH-mean of 53.9%, and Pre-mean of 0.8 mm.
Type3	12.1%	High-pressure center over the North China Plain. Dongying inside the high-pressure, with dominant easterly wind, Tmax-mean of 16.9 °C, RH-mean of 64.0%, and Pre-mean of 1.4 mm.
Type4	16.6%	High-pressure center over the North China Plain. Dongying under the top-front of the high-pressure, with dominant southwesterly wind, Tmax-mean of 14.5 °C, RH-mean of 57.0%, and Pre-mean of 0.9 mm.
Type5	6.7%	Dongying under the control of a weak pressure field with sparse isobars and stable synoptic condition, with dominant southerly wind, Tmax-mean of 17.3 °C, RH-mean of 64.0%, and Pre-mean of 1.9 mm.
Type6	8.4%	Dongying between a high- and low-pressure systems with dense isobars, with dominant southeasterly and southwesterly winds, Tmax-mean of 19.2 °C, RH-mean of 56.8%, and Pre-mean of 0.6 mm.
Type7	9.5%	Dongying in the northwest of the western Pacific subtropical high, with dominant southeasterly wind, Tmax-mean of 29.2 °C, RH-mean of 72.4%, and Pre-mean of 8.0 mm.
Type8	7.6%	Dongying near the inverted trough, with dominant southeasterly wind, Tmax-mean of 25.8 °C, RH-mean of 73.2%, and Pre-mean of 6.6 mm.
Type9	4.4%	Dongying inside the low-pressure field, with dominant southeasterly wind, Tmax-mean of 29.3 °C, RH-mean of 69.1%, and Pre-mean of 2.1 mm.

Notes: T_max-mean refers to the daily average maximum temperature; RH-mean refers to the daily average relative humidity; and Pre-mean refers to the daily average precipitation.

).

As shown in Figure 12, when the proportion of “subtropical high” (Type 7), “inverted trough” (Type 8), and “low-pressure field” (Type 9) was higher, the NO₂ concentration in Dongying was relatively low. Conversely, when the proportion of these types decreased and the proportion of high-pressure synoptic patterns (Type 1, Type 2, Type 3, and Type 4) increased, the NO₂ concentration rose. This is because Type 7, Type 8, and Type 9 usually occur in summer, with higher T_max-mean, occasional precipitation, and higher RH-mean, which facilitate the removal of pollutants. When Dongying was under “offshore high-pressure rear” (Type 1), the impact of the cold high-pressure was waning, resulting in stable synoptic conditions and low RH, which can lead to a higher concentration of accumulated pollutants and higher NO₂ levels. When Dongying was under “high-pressure front” (Type 2) and “high-pressure top front” (Type 4), the passage of cold air masses resulted in a significant drop in T and RH. Additionally, the dominant wind was typically a low-speed southwest wind, which was conducive to the increase in NO₂ concentration. When Dongying was under “high-pressure inside” (Type 3), the activity of the cold high-pressure was weak, leading to conditions such as temperature inversion, which can result in NO₂ exceedances. Moreover, when Dongying was under “uniform pressure” (Type 5), horizontal atmospheric movement is weak, often accompanied by strong radiative inversions or low-level inversions, making high NO₂ concentration more likely [54]. The above analysis indicates that Type 7, Type 8, and Type 9 help to reduce NO₂ concentration, whereas high-pressure synoptic patterns (Type 1, Type 2, Type 3, and Type 4) and “uniform pressure” (Type 5) tend to trigger NO₂ pollution.

The monthly variation in NO₂ concentrations on land–sea breeze and non-land–sea breeze days in Dongying City in 2022 is shown in Figure 13a. Overall, the average NO₂ concentration on land–sea breeze days in Dongying (29.1 µg/m³) was higher than that on non-land–sea breeze days (26.5 µg/m³). In January, March, and September to October, the NO₂ concentration on land–sea breeze days in Dongying was higher than that on non-land–sea breeze days. From June to August and November to December, the NO₂ concentrations on land–sea breeze days were similar to those on non-land–sea breeze days. In February and from April to May, the NO₂ concentration on land–sea breeze days was lower than that on non-land–sea breeze days. The above analysis indicates that during high NO₂ concentrations periods, the land–sea breeze has a significant contribution to the elevated daily average NO₂ concentrations in Dongying City.

The daily variation in NO₂ concentrations on land–sea breeze and non-land–sea breeze days in Dongying City in 2022 is shown in Figure 13b. Overall, the hourly NO₂ concentrations on land–sea breeze days were significantly higher than those on non-land–sea breeze days. From 00:00 to 20:00, the NO₂ concentrations on land–sea breeze days were higher than those on non-land–sea breeze days. From 00:00 to 04:00, the difference between the two was small, ranging from 1.9 to 3.6 µg/m³. From 05:00 to 15:00, the difference gradually increased and then decreased, with the greatest difference observed at 11:00 (7.3 µg/m³). From 16:00 to 18:00, the concentration difference between the two remained between 3.3 and 3.5 µg/m³. From 19:00 to 20:00, the difference gradually decreased, and by 21:00, the NO₂ concentration on land–sea breeze days was lower than on non-land–sea breeze days. The possible explanation for this phenomenon is as follows: during the early phase of land breeze dominance, pollutants from the continent are transported to the air over the Bohai Sea. Meanwhile, ship emissions in the Bohai region also contribute to air pollution [55], and currently, NOx and VOC emissions from ships in China remain relatively high [56]. Consequently, NO₂ gradually accumulates over the Bohai Sea. During the transition from land breeze to sea breeze, NO₂ pollution plume over the Bohai Sea are gradually transported back to Dongying City. Under the combined influence of transport and local emissions, the hourly NO₂ concentration is higher than that on non-land–sea breeze days [20]. During the period of sea breeze dominance, NO₂ is continuously transported from the Bohai Sea to Dongying, resulting in a large difference between the two. During the transition from sea breeze to land breeze, the NO₂ concentration in the sea area decreases and is gradually carried back to the air over the Bohai Sea by the dominant land breeze, resulting in a smaller difference in NO₂ concentrations between land–sea breeze and non-land–sea breeze days, with the concentration on land–sea breeze days at certain hours being lower.

The variation of NO₂ concentrations with T, RH, WS, and WD in Dongying from 2017 to 2022 is shown in Figure 14. High NO₂ concentration (>40 μg/m³) are notably concentrated when T from −5 to 15 °C and RH from 30 to 80%. Daily average NO₂ concentration exceeding standards is more likely when T are below 10 °C and RH ranges from 50 to 70%. It is noteworthy that NO₂ concentration does not show a strong linear relationship with T and RH. This can be explained by the findings of Huang et al. [57], who, through extensive data analysis, discovered that NO₂ concentrations are primarily influenced by BLH and AP, especially by BLH. When WS in Dongying are below 1 m/s, both the level of daily average NO₂ concentration and the frequency of high NO₂ concentration are relatively high. When WD is N to NW and WS range from 2 to 3 m/s, high NO₂ concentration are more frequent. When WD is W to S, the level of daily average NO₂ concentration and the frequency of high NO₂ concentration decrease with increasing WS. Overall, high NO₂ concentration are more likely to occur when WS are below 1 m/s. While when WS exceeds 1 m/s, southwestern winds facilitate a higher NO₂ concentration. This may be due to local pollutant accumulation at low WS and regional pollutant transport at high WS [58,59].

In addition, this study analyzed the correlation between the daily average NO₂ concentrations and meteorological factors from 2017 to 2022 (Table S2). From a seasonal perspective, WS (−0.18 to −0.57) and RH (−0.30 to −0.60, except in winter) exhibited relatively consistent correlations across all seasons, whereas T and AP showed seasonal variations. Previous studies have indicated that the strength and correlation of meteorological factors affecting NO₂ concentrations vary depending on the season [60] or the study period [61]. From an annual perspective, NO₂ concentrations showed a significant negative correlation with WS (−0.43), T (−0.45), and RH (−0.38), while exhibiting a positive correlation with AP (0.46). These results are consistent with findings from Handan [62], as well as studies in Shijiazhuang, Foshan and Ya’an regarding WS, T and AP, although no significant correlation with RH was observed in these regions [60,63,64]. The possible reasons for these findings include higher NO₂ emissions in winter, while elevated T in summer enhances photochemical consumption; low WS hinders pollutant dispersion; high RH may promote the heterogeneous reaction of NO₂ [65]; and high AP leads to descending airflows and suppresses pollutant dispersion.

In summary, high NO₂ concentrations in Dongying tend to occur under meteorological conditions characterized by relatively low T (−5 to 15 °C), moderate RH (30–80%), southwesterly WD, and WS below 1 m/s.

The factor importance obtained after building a random forest model based on meteorological data is shown in Figure 15. Overall, the main factors influencing the hourly NO₂ concentrations in Dongying City in 2022 were BLH, while the dominant factors for the daily average NO₂ concentrations were BLH, WS, and UR. Studies have shown that a decrease in BLH is a key reason for the significant increase in NO₂ concentrations at night and across seasons [66,67]. From the difference between NO₂-HH and NO₂-NHH, the importance of BLH significantly increased during the spring and winter for NO₂-HH (with correlations of −0.43 and −0.36, and importance contribution differences of 10% and 12%, respectively). This is because BLH is generally lower in spring and winter, which facilitates the occurrence of high NO₂ values, thus contributing to the main cause of NO₂-HH formation. In summer, the importance of WD increased significantly (correlation of 0.21, importance contribution difference of 10%), likely due to the transport of NO₂ air masses in Dongying City, which is a major factor in the formation of NO₂-HH. In autumn, the importance of T increased significantly (correlation of −0.46, importance contribution difference of 4%), which is related to Dongying City’s low-temperature environment, making it conducive to the formation of NO₂-HH. From the difference between NO₂-HD and NO₂-NHD, the importance of BLH significantly increased, indicating that BLH is the primary factor for the occurrence of NO₂-HD.

3.3.2. Chemical Generation

The chemical reactions between NO, NO₂, VOCs, and O₃ affect the concentration of NO₂. As shown in Figure 16, compared to NO₂-NHH periods, all types of VOCs increase during NO₂-HH events, with alkanes and halogenated hydrocarbons showing a more pronounced increase, while OVOCs increase to a lesser extent. This suggests that the significant increase in VOCs during NO₂-HH events, particularly alkynes, alkanes, and alkenes, may be associated with the formation of NO₂-HH. In terms of changes in pollutant concentrations, except for a marked decrease in O₃ concentration during NO₂-HH events, concentrations of other pollutants are higher, with NO showing the greatest increase. This is related to NO₂-related chemical reactions, particularly the titration reaction between NO and O₃ and the series of nighttime chemical reactions involving NO₂ [3]. The balance of these reactions significantly influences hourly NO₂ concentration levels. Overall, NO₂-HH events are more likely to occur when VOCs and NO concentrations are high, and O₃ concentration is low.

As shown in Figure 17, on NO₂-HH event days, VOCs and NO concentrations increase significantly during nighttime and morning rush hours, exhibiting a unimodal pattern, with more pronounced photochemical consumption at noon. In contrast, on NO₂-NHH days, VOCs and NO concentrations remain relatively stable at night, with a smaller increase during the early morning rush and less photochemical consumption at noon. The high concentrations of VOCs and NO in Dongying contribute to intense photochemical reactions during the day, leading to higher NO₂ levels. On NO₂-HH event days, the peak NO concentration in the evening is higher, which rapidly reduces nighttime O₃ levels through titration reactions, while the NO₂ concentration remains stable after a rapid increase. Xia et al. [2] found that an increase in the diurnal variation of NO₂ is accompanied by a larger diurnal variation in O₃, which is consistent with the variations in NO₂ and O₃ on NO₂-HH event days in Dongying. Additionally, studies have shown that when NO concentration is high, NO₃ is primarily removed through reaction with NO to form NO₂, which suppresses the nighttime NO₂ net consumption rate about NO₃. Therefore, the NO₃^- formation rate is lower during the evening peak of NO₂-HH event days. The stability of NO₂ in the early morning indicates that on NO₂-HH event days, there is ample NO, and NO₂ is in a dynamic equilibrium between generation and consumption, with O₃ continuously depleted and NO₃⁻ increasing. Conversely, on NO₂-NHH days, after the NO₂ concentration rises during the evening peak, it shows a decreasing trend along with VOCs in the early morning. This is due to insufficient NO at night and higher O₃ concentration, which dominate the NO₂ net consumption reactions. Overall, daytime NO₂ concentration are primarily influenced by intense photochemical reactions, with higher VOCs and NO concentrations leading to higher daytime NO₂ levels. During nighttime and morning rush hours, NO and O₃ significantly affect NO₂ concentration.

The factor importance obtained after building a random forest model based on pollutant data is shown in Figure 18. Overall, the main pollutants influencing the hourly NO₂ concentrations in Dongying City were NO, O₃, and Aromatics, while the dominant factors for the daily NO₂ concentrations were NO, Aromatics, Alkenes, and O₃. From the difference between NO₂-HH and NO₂-NHH, the importance of O₃ and NO increased for NO₂-HH, except in winter, while the importance of Aromatics decreased across all seasons. Other pollutants showed no consistent pattern. This suggests that NO₂-HH in spring, summer, and autumn is primarily driven by the titration reaction between NO and O₃, while the day–night variation in Aromatics is related to the concentrations of oxidants OH and NO₃ [68]. The weakening of reactions with NO₃ reduces the net consumption of NO₂. From the difference between NO₂-HD and NO₂-NHD, the importance of Alkenes decreased, while that of Aromatics increased. This is because Alkenes are key participants in the nighttime chemical reaction chain, consuming NO₃ and leading to net consumption of NO₂ [69]. Both Alkenes and Aromatics exhibit strong photochemical activity during the day, participating in photochemical chain reactions to produce NO₂ [70].

Overall, daytime NO₂ concentration are primarily influenced by intense photochemical reactions, with higher VOCs and NO concentrations leading to higher daytime NO₂ levels. During nighttime and morning rush hours, NO and O₃ significantly affect NO₂ concentration. The titration reaction between NO with O₃ is the main cause for NO₂-HH in spring, summer and autumn, and photochemical reactions of Aromatics have a significant influence on NO₂-HD.

3.3.3. Sources Analysis

As shown in Figure 19, NOx emissions in Dongying are generally higher on NO₂-HD days compared to NO₂-NHD days across all seasons, with the greatest difference in winter and the smallest in summer. This indicates that local NOx emissions have a significant impact on NO₂-HD in Dongying. Seasonal variations in NOx emissions are highest in winter and lowest in spring, which differs from the seasonal variations in NO₂ concentration, suggesting that NO₂ distribution may be influenced by additional factors. Industry-specific NOx emissions data reveal that thermal power (51–57%) and petrochemical industry (26–32%) are the largest contributors to NOx emissions across all seasons. Compared to NO₂-NHD days, on NO₂-HD days, the proportion of emissions from thermal power increases in spring, summer, and autumn, while the proportion from the petrochemical industry increases in winter. Therefore, thermal power and petrochemical industry are likely major sources affecting NO₂-HD. In summary, local emissions (thermal power and petrochemical industry) have a certain impact on NO₂-HD in Dongying.

Cluster analysis (Figure 20 and Figure S12) reveals significant seasonal differences in the direction, speed, and altitude of air masses at Dongying sites. In spring, northwest trajectory 3 accounts for 43.1%, with the shortest trajectory, indicating slower air mass movement and a higher potential for local pollutant accumulation. This trajectory also operates at a lower altitude, facilitating local horizontal pollutant transport. The corresponding average NO₂ concentration for this trajectory was 21.1 μg/m³, contributing 48.1%. Southwest trajectory 2, which accounts for 34.9%, has a relatively short trajectory and the lowest altitude, leading to pollutant transport from southern cities in Shandong Province. Additionally, trajectory 1, with a lower proportion, has a higher altitude and exerts minimal impact on Dongying in spring. In summer, the southeastward trajectory had the highest share (44.9%) and contribution (45.7%). The average concentration for this trajectory matched the summer average NO₂ concentration, indicating that this trajectory is the primary source of air masses maintaining summer NO₂ levels. Trajectory 3 had the lowest share (23.1%) but the highest average concentration (18.4 μg/m³), which facilitates the formation of high NO₂ values. Additionally, Dongying City in summer may be affected by the transport of air masses from the Bohai Bay (32%). Similarly, in autumn and winter, the local influence on Dongying City is significantly stronger, primarily from air masses coming from the south and southwest. This is related to air mass transport and pollutant accumulation [71]. The lifespan of NO₂ is relatively short (1–2 days), primarily influenced by local transport rather than long-distance transport [72].

The distribution of WPSCF and WCWT in Dongying in 2022 is shown in Figure 21. From the WPSCF distribution, in spring, the WPSCF values for the southwestern regions of Dongying and the small area around the Bohai are relatively high, with a focus on cities in western Shandong Province, such as Jinan, Binzhou, and Taian, with western Taian having the highest potential impact on NO₂-HH. In summer, the WPSCF values for nearby cities of Dongying, particularly Binzhou and Zibo, are higher, indicating a significant influence from nearby cities. In autumn, potential source regions affecting NO₂-HH in Dongying are more widely distributed, primarily in central Dezhou, Jinan, Taian, and parts of Linyi. In winter, the main potential source regions affecting NO₂-HH in Dongying are located in northeastern Handan, southwestern Dezhou, and central areas of Jinan and Taian. From the WCWT distribution, the distribution of WCWT values is quite similar to that of WPSCF values, with the areas of high concentration contribution aligning well with the potential source regions that significantly impact NO₂-HH at the observation site. In spring, autumn, and winter, areas with WPSCF values greater than 0.5 generally correspond to areas with high WCWT values, while in summer, areas with WPSCF values greater than 0.25 align with areas of high WCWT values. This correspondence is related to the NO₂ threshold settings in the seasonal PSCF analysis. In addition to the regions corresponding to the high WPSCF values, there are some areas with high WCWT values in the Bohai Sea of China during spring and summer and in the northwest of Dongying during winter.

In summary, the distribution of high WCWT and WPSCF values is quite similar and aligns closely with the main air mass transport pathways. NO₂ regional transport in Dongying primarily occurs within Shandong Province. The potential source regions that contribute most significantly to the occurrence probability and concentration of NO₂-HH at the Dongying site are mainly located in the western and southwestern parts of Shandong Province (with wider distribution in autumn and winter), including Jinan, Binzhou, Taian, and Zibo. Additionally, during spring and summer, there is a certain influence from the Bohai Sea of China.

Dongying City is located in the Bohai Bay region, which has a unique geographical location and meteorological conditions [19,55]. In the future, building on outstanding international research, it is highly necessary to conduct long-term and large-scale studies on NO₂ to provide scientific support for regional NO₂ control and air quality improvement.

4. Conclusions

In this study, we explored the effects of NO₂ on the ambient air quality in Dongying, a typical petrochemical city in Bohai Bay of China, from 2017 to 2023 and the impacts of high NO₂ values on the concentrations of O₃, PM_2.5 and PM_2.5-bounded secondary components, and atmospheric oxidation capacity in 2022, and investigated the reasons for the formation of high NO₂ values in Dongying.

(1)

From 2017 to 2023, the annual assessment values of NO₂ concentrations in Dongying exhibited a trend of initially decreasing and then stabilizing, with monthly variation characterized by a “U” distribution, diurnal variation characterized by a “single-peak, single-valley” distribution, and spatial variation characterized by a “high in the south-central part and low in the north” distribution. Higher daily NO₂ concentrations mainly occurred during January-April and September-December each year, and higher hourly NO₂ concentrations mainly occurred during the nighttime and morning rush hours.

(2)

NO₂ is a contributing factor to the increase in various pollutant concentrations, thereby affecting air quality levels. High hourly NO₂ values (NO₂-HH) events promoted the increase in the concentrations of PM_2.5 and PM_2.5-bounded NO₃^-, SO₄^2- and SOC, and O_X during the nighttime and morning rush hours, and they also facilitated the increase in the concentrations of O₃, PM_2.5 and PM_2.5-bounded NO₃^-, SO₄^2- and SOC, and O_X, as well as NOR and SOR values during the daytime. High daily NO₂ values (NO₂-HD) could promote the increase in the concentrations of O₃ (MDA8-O₃), PM_2.5 and PM_2.5-bounded NO₃^-, SO₄^2- and SOC, and O_X on the same day.

(3)

The occurrence of high NO₂ values was affected by the combination of unfavorable meteorological conditions, local emissions and regional air pollution transports, and localized atmospheric chemical generation. High-pressure and uniform-pressure weather patterns, along with low temperatures (−5~15 °C), moderate humidity (30~80%), wind speeds below 1 m/s, and southwesterly winds, are conducive to high daily NO₂ values. In 2022, land–sea breeze circulation significantly contributed to high NO₂ concentrations. Random forest analysis indicates that in 2022, BLH in spring (−0.43) and winter (−0.36), WD in summer (0.21), and T (−0.46) in autumn were the primary factors contributing to NO₂-HH, while BLH (−0.47) is the main cause for NO₂-HD. The titration reaction between NO with O₃ is the main cause for NO₂-HH in spring, summer and fall, and photochemical reactions of Aromatics have a significant influence on NO₂-HD. NOx emissions from the thermal power and petrochemical industry in Dongying and air pollution transports from western and southwestern Shandong Province (throughout the year) and from the Bohai Sea (during spring and summer) had serious adverse impact on high NO₂ values.

It is recommended that Dongying strengthens its control over NO_X emissions from local thermal power and petrochemical industries, and actively promotes the establishment of an air pollution prevention and control mechanism in collaboration with cities in western and southwestern Shandong Province and the Bohai Bay region of China to effectively reduce NO₂ concentrations and prevent high NO₂ values, thereby continuously improving air quality.