Research on Ozone Pollution Characteristics and Source Apportionment During the COVID-19 Lockdown in Jilin City in 2022

Abstract

The increasing Ozone (O₃) concentration in various regions of China has garnered significant attention, highlighting the need to understand the mechanisms of O₃ formation. This study focuses on the source apportionment of O₃ in Jilin City during and after the COVID-19 lockdown countermeasure, and also the influence of anthropogenic emissions on O₃ concentration. The contributions of different O₃ emission sources were quantified using the Weather Research and Forecasting Community Multi-Scale Air Quality (WRF-CMAQ) model in conjunction with the Integrated Source Apportionment Method (ISAM). The results indicate a significant increase in O₃ concentrations during the lockdown in Jilin City, which were particularly characterized by long-distance transportation. Transportation is identified as the primary direct source of O₃ in Jilin City, with Yongji County contributing the most among the six designated regions. This study highlights variations in the causes and sources of O₃ pollution among the different regions of Jilin City. Simply controlling anthropogenic emissions is inadequate for effectively managing O₃ pollution and may even worsen the situation. It is more effective to focus on controlling O₃’s precursors. These findings improve the understanding of O₃ pollution in Jilin City and provide valuable insights for developing O₃ control policies. Similarly, this research is applicable to other countries and regions.

1. Introduction

In recent years, industrialization, urbanization, and population growth have rendered air pollution a significant environmental concern [1]. In December 2019, a novel coronavirus disease (COVID-19) spread rapidly in China and around the world [2]. The COVID-19 pandemic had a profound impact on both social and environmental dimensions globally [3]. To mitigate the spread of COVID-19, China enacted stringent lockdown policies and measures. The widespread adoption of these measures provided a unique opportunity to investigate the effects of reduced anthropogenic emissions on air quality. The influence of COVID-19 on air pollutants has been termed “the largest air quality control experiment ever conducted” [4]. In March 2022, a COVID-19 outbreak emerged in Jilin City. To curb the spread of the virus, the Jilin City government enacted decisive measures that significantly reduced unnecessary human activities, including operations in industry, agriculture, commerce, and transportation [5].

Consequently, numerous scholars have utilized various methods to quantify changes in pollutant concentrations during the COVID-19 pandemic, including carbon dioxide (CO₂), the particulate matter smaller than 2.5 μm (PM_2.5), the particulate matter smaller than 10 μm (PM₁₀), carbon monoxide (CO), sulfur dioxide (SO₂), and nitrogen dioxide (NO₂), revealing differing degrees of reduction in these pollutants. For instance, Cheng et al. [6] analyzed changes in anthropogenic CO₂ emissions in Nanchang City before, during, and after the COVID-19 pandemic, investigating how different emission categories contributed to total CO₂ emissions. Their findings indicated a 37.1–40.2% decrease in emissions, with CO₂ from the power sector and manufacturing declining by 54.5% and 18.9%, respectively. Nie et al. [7] investigated 31 provincial capitals and identified an 8–17% decrease in PM_2.5, PM₁₀, SO₂, NO₂, and CO concentrations compared to 2019 levels. Wang et al. [8] observed that, due to the government’s lockdown policy, the concentrations of five pollutants (PM_2.5, PM₁₀, CO, SO₂, and NO₂) in Jilin City decreased significantly in 2022, ranging from 27% to 41%, excluding O₃.

In contrast, previous studies reported an increase in O₃ concentrations; Bi et al. [9] studied Beijing, Delhi, Guangzhou, London, Paris, Rome, Sao Paulo, Seoul, and Wuhan, and all selected cities showed significant increases in O₃ concentrations during the lockdown. In comparison to the average concentration observed over the past five years, the most substantial increase was recorded in Wuhan at 49.3%, succeeded by London at 47.5%, Paris at 31.8%, Sao Paulo at 28.9%, Guangzhou at 20.1%, Delhi at 20.0%, Beijing at 18.4%, Seoul at 17.9%, and Rome at 4.8%. It is important to note that Wuhan, London, Sao Paulo, and Paris are cities with high levels of pollution, where anthropogenic activities are significant contributors to emissions. Sicard et al. [10] analyzed average deviations during the lockdown (March–April 2020) compared to the same period over the previous three years and found that, in four Southern European cities, O₃ concentrations increased by an average of 17%.

Moreover, in recent years, air quality models with tagging capabilities have been increasingly used to analyze regional O₃ sources. For instance, technologies like Comprehensive Air Quality Model with Extensions (CAMx)’s Ozone Source Apportionment Technology (OSAT) for O₃ source resolution [11,12], and Community Multi-Scale Air Quality (CMAQ)’s integrated source allocation method Integrated Source Apportionment Method (ISAM) [13,14], have been applied. Air quality models have been extensively used to study O₃ pollution. Yang et al. [15] utilized WRF-CMAQ-ISAM simulations to analyze the evolution mechanism of extreme O₃ events in the Sichuan Basin and conducted a source analysis. Their results indicate that industrial and transportation precursor emissions have the largest impact on increasing O₃ concentrations, while power sources contribute minimally to O₃ pollution. Collet et al. [16] employed CMAQ-ISAM to predict O₃ sources in the United States for July 2030, revealing that boundary conditions will be the largest contributors to O₃, followed by point sources, with the eastern United States typically contributing more than the western United States. Liu et al. [17] used CMAQ-ISAM to study O₃ concentration variations and pollutant sources during heavy O₃ pollution in Beijing, finding that boundary conditions contribute significantly (77.45%), while contributions from Beijing and nearby areas are relatively small. During heavy O₃ pollution in Beijing, pollutant transport from Tianjin, Hebei, and Shandong accounted for 9.63%, 19.02%, and 13.01%, respectively.

These studies indicate that, while concentrations of most pollutants decreased during COVID-19 lockdowns, O₃ concentrations increased. Currently, studies focused on Northeast China remain limited. Therefore, this study utilizes the WRF-CMAQ-ISAM air quality model to analyze O₃ sources in Jilin City during the COVID-19 lockdown in 2022, compared to the same period in 2019. It further examines the contribution rates of various industries and the transmission contributions from different regions within the city.

2. Data Sources and Methods

2.1. Overview of the Study Area

Jilin City, commonly referred to as “Beiguo River City”, is a prefecture-level city located in Jilin Province, northeastern China. Covering an area of 17,200 square kilometers, it is one of the primary cities in the province. Recognized by the State Council, Jilin City is a key industrial center, integral to the old industrial base of Northeast China, and an emerging industrial hub within Jilin Province. Geographically, it is situated between latitudes 42°31′ N and 44°40′ N and longitudes 125°40′ E and 127°56′ E, with a registered population of approximately 4.116 million, as of the end of 2020 [18].

Jilin City administers four districts and five county-level cities: Changyi District, Chuanying District, Fengman District, Longtan District, Shulan City, Panshi City, Jiaohe City, Huadian City, Yongji County. It borders Harbin to the north, Yanbian to the east, Changchun and Siping to the west, and Baishan, Tonghua, and Liaoyuan to the south. Geographically, the city is elevated in the southeast and lower in the northwest, bordered by mountains on one side and rivers on three sides. This topography results in a dry and warm climate during spring and autumn, accompanied by cold and prolonged winters. The precipitation in the city is mainly in the form of rainfall and snowfall, and the precipitation in the four seasons shows different characteristics, with the precipitation increasing significantly in winter and summer, while the precipitation in spring and autumn is less obvious. The city’s complex climate leads to highly uneven temperature distribution, with warmer temperatures generally observed in the west and northwest. The overall climate is characterized by low temperatures, with an average high temperature of 13 °C and an average low temperature of −1 °C in 2022, with the highest average temperature in July, an extreme high temperature of 33 °C on 26 July, the lowest average temperature in January, and an extreme low temperature of −28 °C on 13 January.

2.2. Data Sources

This study obtained air pollutant concentrations data from the National City Air Quality Real-Time Publishing Platform of the Ecological Environment Bureau (https://air.cnemc.cn:18007/, accessed on 5 July 2022). Meteorological data were sourced from the China Meteorological Data Service Center (http://data.cma.cn, accessed on 5 July 2022). We collected observational data from seven air quality monitoring stations and one meteorological station in Jilin City (Table S1). The air quality monitoring network consists of one background station in Fengman (FM) and six urban stations: Hada Bay (HDW), East Bureau (DJZ), Electric Power College (DJXY), Jiangbei (JB), Jiangnan Park (JNGY), and Nine Stations (JZ). Ozone concentrations (O₃-8H) were recorded using an eight-hour sliding average, while CO, SO₂, NO₂, PM_2.5, and PM₁₀ concentrations were monitored hourly. Additionally, meteorological variables, including air pressure, wind speed, wind direction, and temperature, were recorded at meteorological stations across the country.

The emission inventory in this study is based on the China Multiresolution Emission Inventory, developed by the research team at Tsinghua University [19,20] (MEIC, http://meicmodel.org/ accessed on 11 March 2022), with a spatial resolution of 0.25° × 0.25°. This inventory is recognized for its versatility and accuracy [21]. MEIC anthropogenic emissions are classified into five sectors: industrial, power, transportation, residential, and agricultural. The pollutants analyzed include SO₂, NO₂, PM_2.5, CO, and volatile organic compounds (VOCs) covering the region from 70° E to 150° E and 10° N to 60° N. We employed the Inventory Spatial Allocation Tool (ISAT) to spatialize the emission inventory and input the resulting data into the Community Multi-Scale Air Quality (CMAQ) model for numerical simulations. Emissions were allocated based on population, GDP, and land use grid data sourced from the Resources and Environmental Sciences Data Center of the Chinese Academy of Sciences [22].

2.3. WRF-CMAQ-ISAM Model

This study employed the WRF model (version 4.3.3) and the CMAQ model (version 5.4) [23] to simulate Jilin City and its surrounding regions, both of which are widely employed for mesoscale numerical simulations [24]. The ISAM module, a label-tracking method used in source apportionment, was employed within the CMAQ model to trace pollutants. WRF was used to simulate meteorological conditions, with initial weather fields sourced from the FNL Global Reanalysis dataset (http://rda.ucar.edu/datasets/ds083.2, accessed on 5 July 2022) provided by the National Center for Environmental Prediction. This dataset has a spatial resolution of 1° × 1° and a temporal resolution of 6 h. Simulations were conducted across three nested domains, with the outermost domain providing initial and boundary conditions for the inner domains. The spatial resolutions of Domain 1, Domain 2, and Domain 3 were 27 km × 27 km, 9 km × 9 km, and 3 km × 3 km, respectively, covering Northeast China, Jilin Province, and Jilin City. The simulation domains are illustrated in Figure 1.

2.4. Research Methods

To visually illustrate the changes in air quality and pollution sources in Jilin City during the 2022 epidemic, the research period was divided into four stages based on the progression of the outbreak and corresponding prevention and control measures issued by the Jilin City People’s Government on its official website. On 3 March 2022, COVID-19 was circulating again in Jilin City, which reported its first local COVID-19 case and three asymptomatic infections. Following initial contact tracing, it became clear that the virus had spread within the city. To contain the outbreak, traffic control was enforced at all entry and exit points starting on 4 March, and public transport in urban and rural areas was temporarily suspended on 6 March. Key urban bus routes remained operational, while others were temporarily suspended. From 7 to 11 March, all government offices, businesses, and institutions suspended operations, major schools and kindergartens closed, and all commercial venues shut down. On 14 March, Jilin City reported 2601 new confirmed cases, including 18 asymptomatic infections that were reclassified as confirmed cases. The Jilin Provincial Government’s Information Office held a press conference on epidemic prevention and control, announcing measures to prevent further spread. Starting 14 March 2022, cross-provincial and cross-city travel was restricted, especially in Changchun and Jilin City. On 8 April, the Jilin City Epidemic Prevention Headquarters announced that the city had achieved its goal of eliminating community transmission. By 28 April, social control measures were gradually lifted, and the normal order of production and daily life was systematically restored. On 12 May, with no new local confirmed cases or asymptomatic infections, the government fully lifted the lockdown. Based on this timeline, the research period was divided into four stages: Stage 1 (3 February to 3 March), Stage 2 (4 March to 27 April), Stage 3 (28 April to 11 May), and Stage 4 (12 May to 31 May). The first stage represents the pre-containment period, the second is the strict containment phase, the third is the orderly recovery phase, and the fourth is the post-containment observation phase.

This research employs the WRF-CMAQ model to simulate air quality during the strict containment period (4 March to 27 April). Firstly, the WRF model is used to simulate the meteorological effect and verify it to ensure that the output results of the model can provide more accurate meteorological data for the subsequent CMAQ simulation, and then the CMAQ model is used to simulate and verify the O₃ concentration in the study period to ensure that the model can accurately present the change trend of O₃ concentration. The ISAM module is employed to analyze industrial and regional O₃ sources. The same simulation was also conducted for the same period in 2019. The contribution of different O₃ emission sources and the impact of shutdown on O₃ concentration were quantified, allowing for a comparison of the pollutant sources. The impact of industrial shutdowns during the epidemic on O₃ levels was then evaluated. In CMAQ version 5.4, the MCIP must begin after the WRF weather field output. To minimize the impact of initial conditions, the WRF simulation was initiated five days earlier.

3. Results and Discussion

3.1. Analysis of Atmospheric O₃ Concentrations

Although directly comparing O₃ concentrations across different periods does not fully capture changes in regional O₃ emissions, it still provides insight into long-term emission trends. To assess the impact of government measures on O₃ concentrations during lockdown, we analyzed daily O₃ concentration changes from 3 February to 31 May, 2019–2022. Figure 2 presents the daily and monthly average O₃ concentrations during Levels 1 through 4 from 2019 to 2022, comparing O₃ concentrations between 2019 (pre-epidemic) and 2022 (during the epidemic). In 2019, 2020, 2021, and 2022, the average concentration of O₃ during Level 1 was 66.07 μg/m³, 70.47 μg/m³, 67.73 μg/m³, and 63.81 μg/m³, respectively. The average concentration of O₃ during Level 2 was 77.25 μg/m³, 78.12 μg/m³, 75.05 μg/m³, and 92.21 μg/m³, respectively. The average concentration of O₃ during Level 3 was 98.01 μg/m³, 90.17 μg/m³, 73.11 μg/m³, and 101.78 μg/m³, respectively. The average concentration of O₃ during Level 4 was 100.09 μg/m³, 78.42 μg/m³, 90.53 μg/m³, and 91.41 μg/m³, respectively. During Level 1 in 2022, O₃ concentrations were reduced by 3.42%, 9.45%, and 5.79% compared to the same period in 2019, 2020, and 2021, respectively. During Level 2 in 2022, O₃ concentrations increased by 19.37%, 18.05%, and 22.87% compared to the same period in 2019, 2020, and 2021, respectively. During Level 3 in 2022, O₃ concentrations increased by 3.84%, 12.88%, and 39.21% compared to the same period in 2019, 2020, and 2021, respectively. During Level 4 in 2022, O₃ concentrations increased by −8.67%, 16.56%, and 0.97% compared to the same period in 2019, 2020, and 2021, respectively. Figure 3 compares O₃ concentrations between 2019 and 2022. The remarkable increase in O₃ concentrations is primarily due to a reduction in the titration reaction, which typically consumes O₃ near the ground [25,26]. During the lockdown, restrictions on mobility and driving led to a significant reduction in vehicle exhaust emissions, which in turn reduced nitrogen oxide (NO_x) concentrations [8]. This weakening of the titration reaction caused a notable increase in O₃ concentrations compared to the same period in previous years.

Ozone-induced photochemical reactions can encompass thousands of species and more than 20,000 reactions. At present, the formation process of photochemical smog is usually summarized using a simplified mechanism composed of the following reactions [27,28].

(1)

Photochemical cycle of O₃ with NO and NO₂:

NO₂ + hv (λ < 420 nm) → NO + O (³P)
O (³P) + O₂ + M → O₃
O₃ + NO → NO₂ + O₂

(2)

Free radicals trigger a reaction:

O₃ + hv (λ < 320 nm) → O (¹D) + O₂
O (¹D) + H₂O → OH + OH
HONO + hv (λ < 400 nm) → OH + NO

(3)

Free radical transport reactions:

CO + OH → H₂O + CO₂
VOCs + OH → RO₂ + H₂O
RCHO + OH → RC(O)O₂ + H₂O
RCHO + hy → RO₂ + HO₂ + CO
HO₂ + NO → NO₂ + OH
RO₂ + NO → NO₂ + RCHO + HO₂
RC(O)O₂ + NO → NO₂ + RO₂ + CO₂

(4)

Free radical termination:

HO₂ + HO₂ + M → H₂O₂ + O₂ + M
HO₂ + RO₂ → ROOH + O₂
OH + NO₂ + M → HNO₃ + M
RC(O)O₂ + NO₂

Figure 4 illustrates the hourly O₃ concentration trends across the four blockade stages. O₃ concentration is lowest during Level 1, highest during Level 3, and peaks between 12:00 and 18:00. This peak occurs because temperatures are highest during this period, enhancing the photochemical reactions necessary for O₃ formation. As temperatures rise, solar radiation intensifies, accelerating photochemical reactions and increasing O₃ concentrations.

3.2. WRF Model Simulation Effect Evaluation

In addition to human activities and emission sources, meteorological conditions significantly influence the production, degradation, and transport of atmospheric pollutants, making their impact on air quality crucial. Adverse meteorological conditions can exacerbate air pollution and affect the accuracy of model simulation results [29]. Air pollution frequently correlates with meteorological changes. As surface temperatures rise, atmospheric molecules’ thermal movement increases, enhancing air mass convection. This process transports pollutants to higher altitudes, reducing surface concentrations. As wind speed and humidity increase, pollutants tend to diffuse, migrate, and transform horizontally, particularly during storms [30,31]. Seasonal variations also influence pollutant distribution. Severe pollution incidents are more likely in cold weather with low temperatures, while higher temperatures tend to correlate with lower pollutant concentrations. Severe weather conditions, such as haze, can exacerbate pollutant migration. Wind transports pollutants downwind, aiding local dispersion while also spreading them to other regions [32,33].

To evaluate the model’s performance, we verified and analyzed the simulation results for April 2019 and April 2022. The accuracy of the WRF model is essential for the reliability of the CMAQ model’s output. Therefore, we validated the monitoring data from Jilin City’s meteorological station by comparing it with WRF-simulated data. We selected the 2 m temperature (T2) and 10 m wind speed (WS10) for comparison. The statistical indicators used include linear correlation coefficient (R), normalized mean error (NME), normalized mean bias (NMB), and root mean square error (RMSE). NME quantifies the deviation between simulated and observed values, while NMB assesses the proximity of the simulation to the observations. A positive NMB indicates that the simulation overestimates the observed values, while a negative NMB indicates an underestimation. R represents the correlation between the simulation and observation trends, with a higher R indicating stronger correlation. RMSE assesses the magnitude of deviation between simulation and observation, with a smaller RMSE indicating less deviation.

The precise computational formulas for the four metrics are detailed as follows:

$$\mathrm{R}={\displaystyle \frac{{\sum }_{i=1}^N\left(P_i-\overline{P}\right)\left(O_i-\overline{O}\right)}{\sqrt{{\sum }_{i=1}^N{\left(P_i-\overline{P}\right)}^2}\sqrt{{\sum }_{i=1}^N{\left(O_i-\overline{O}\right)}^2}}}$$

(1)

$$\mathrm{N}\mathrm{M}\mathrm{B}={\displaystyle \frac{{\sum }_{i=1}^N\left(P_i-O_i\right)}{{\sum }_{i=1}^N{O}_i}}$$

(2)

$$\mathrm{N}\mathrm{M}\mathrm{E}={\displaystyle \frac{{\sum }_{i=1}^N\left|P_i-O_i\right|}{{\sum }_{i=1}^N{O}_i}}$$

(3)

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}=\sqrt{{\displaystyle \frac{{\sum }_{i=1}^N{\left(O_i-P_i\right)}^2}{N}}}$$

(4)

where N represents the total number of samples, P_i is the simulation result at moment i, O_i is the monitoring result at moment i, $\overline{P}$ is the average of the simulation results, and $\overline{O}$ is the average of the monitoring results.

The simulation results are output every 6 h, and the observations are also verified using the data recorded every 6 h. The results are shown in Figure 5 and Figure 6 and

Table 1. Statistical analysis of T2 and WS10 simulation results and observation results of the meteorological station.

Month	Element	NMB	NME	R	RMSE
April 2019	T2 (°C)	0.08	0.30	0.92 **	2.91
	WS10 (m/s)	0.46	0.62	0.62 **	2.27
April 2022	T2 (°C)	0.0002	0.26	0.88 **	3.35
	WS10 (m/s)	0.44	0.62	0.56 **	2.71

** There was a significant association at 0.01 level (bilateral).

; the simulated meteorological values closely align with the observed values, demonstrating the model’s strong performance. So, in the calculation, N is 124 in March and 120 in April. The T2 simulation performs well, with R-values ranging from 0.88 to 0.92, passing the 0.01 significance level test, which indicates a strong correlation between simulated and observed values. The NMB ranges from 0.0002 to 0.08, and the NME ranges from 0.26 to 0.30, indicating that the simulated values slightly overestimate the observed values but remain closely aligned. The RMSE ranges from 2.91 to 3.35 °C. The R-value for WS10 ranges from 0.56 to 0.62, passing the 0.01 significance level test, suggesting a moderate correlation between the simulated and observed values. The NMB values range from 0.44 to 0.46, and the NME is 0.62, indicating that the simulated values overestimate but are reasonably close to the observed values. The RMSE ranges from 2.27 to 2.71 m/s.

This indicates the successful simulation of meteorological elements, providing accurate data for subsequent CMAQ simulations.

3.3. CMAQ Model Simulation Effect Evaluation

Given the varying terrain and meteorological conditions across different regions, along with uncertainties in emission sources, inventory accuracy, and model parameter settings, it is essential to evaluate the CMAQ model’s simulation results and assess its performance. This study focused on O₃, comparing and analyzing simulated values from March and April of 2019 and 2022 against observed values from air quality monitoring stations. Statistical indicators, including the correlation coefficient (R), mean fractional bias (MFB), and mean fractional error (MFE), were employed. MFB and MFE are statistical metrics for assessing the performance of air quality models. Studies indicate that, when MFB is between −60% and 60%, and MFE is below 75%, the model’s simulation accuracy meets the required standards [34]. When MFB is between −30% and 30%, and MFE is under 50%, the model’s performance is considered better [35].

The precise computational formulas for the two metrics are detailed as follows:

$$\mathrm{M}\mathrm{F}\mathrm{B}={\displaystyle \frac{1}{N}}{\sum }_{i=1}^N{\displaystyle \frac{\left(P_i-O_i\right)}{{(O}_i+P_i)/2}}$$

(5)

$$\mathrm{M}\mathrm{F}\mathrm{E}={\displaystyle \frac{1}{N}}{\sum }_{i=1}^N{\displaystyle \frac{\left|P_i-O_i\right|}{{(O}_i+P_i)/2}}$$

(6)

where N represents the total number of samples, P_i is the simulation result at moment i, and O_i is the monitoring result at moment i.

As mentioned in 3.2, N is 124 in March and 120 in April.

Table 2. Statistical analysis of O₃ simulation results and observation results.

	March 2019	April 2019	March 2022	April 2022
Simulation value (μg/m3)	48.15	53.56	62.57	69.55
Observation value (μg/m3)	83.77	92.62	82.49	104.09
R	0.36 **	0.38 **	0.34 **	0.33 **
MFB	−41.11%	−37.45%	−29.32%	−35.99%
MFE	44.42%	42.19%	33.37%	39.63%

** There was a significant association at 0.01 level (bilateral).

shows the monthly averages of the simulated and observed data, as well as the statistical indicators. As shown in Table 2, the correlation coefficient (R) between simulated and observed O₃ values over the four months ranges from 0.33 to 0.38. The mean fractional bias (MFB) for March and April 2019 and April 2022 falls below 60%, with MFE remaining under 75%, satisfying the criteria for model accuracy. For March 2022, the MFB is between −30% and 30%, and MFE is below 50%, indicating strong model performance. Figure 7 illustrates the temporal series of simulated and observed O₃ values, revealing a close alignment between simulation trends and observation results. Overall, the CMAQ model developed in this study meets accuracy standards and effectively captures the trends in O₃ concentrations.

3.4. Analysis of Ozone Source Industry in Different Years

To assess the contribution of O₃ emissions from different industries in Jilin City during the COVID-19 epidemic in 2022 compared to 2019, an analysis of O₃ source characteristics by industry was conducted. This study utilized the ISAM source analysis tool in the CMAQ model to classify emission sources into IH (industrial), PH (power), RH (residential), and TH (traffic), with ICON (initial condition) and BCON (boundary condition) included in the emission inventory. Sources outside the simulation area were classified as OTH (other sources). Six districts in Jilin City were identified and labeled: CYING (Chuanying District), CYI (Changyi District), FMAN (Fengman District), LTAN (Longtan District), PSHI (Panshi City), and YJI (Yongji County). Refer to Figure S1 and Table S3 for labeling details. The statistical results in Table S4 indicate a significantly higher contribution from BCON compared to local industrial sources. In 2019, regional contribution rates ranged from 67.36% to 78.52%, with concentrations contributions between 33.562 and 45.277 µg/m³. Among industrial sources, TH made the highest contribution (2.31% to 3.57%), while RH had the lowest (1.22% to 1.83%). In 2022, the regional contribution rate increased to between 68.39% and 79.11%, with concentrations ranging from 36.67 to 48.82 µg/m³. Similarly, TH made the highest contribution (2.58% to 3.53%), while RH had the lowest (1.64% to 1.97%). Figure 8 illustrates the distribution of these sources in Jilin City, and Figure 9 shows the distribution characteristics of 2019 and 2022, concluding that traffic emissions are the primary direct contributor to local O₃ levels, aside from boundary conditions and unclassified sources, with minimal variation in contributions across regions.

This study used the ISAM source analysis tool in the CMAQ model to classify emission sources into IH (industrial), PH (power), RH (residential), and TH (traffic), with ICON (initial condition) and BCON (boundary condition) noted in the emission inventory. Sources outside the simulation area were categorized as OTH (other sources). Six districts in Jilin City were identified and coded: CYING (Chuanying District), CYI (Changyi District), FMAN (Fengman District), LTAN (Longtan District), PSHI (Panshi City), and YJI (Yongji County). See Figure S1 and Table S3 for labeling details. The statistical results in Table S4 show a much higher contribution from BCON compared to local industrial sources. In 2019, regional contribution rates ranged from 67.36% to 78.52%, with concentrations contributions between 33.562 and 45.277 µg/m³. Among industrial sources, TH had the highest contribution (2.31% to 3.57%), while RH had the lowest (1.22% to 1.83%). In 2022, the regional contribution rate increased from 68.39% to 79.11%, with concentrations between 36.67 and 48.82 µg/m³. Similarly, TH had the highest contribution (2.58% to 3.53%), while RH had the lowest (1.64% to 1.97%). Figure 8 shows the distribution of these sources in Jilin City, concluding that traffic emissions are the main direct contributor to local O₃ levels, aside from boundary conditions and unclassified sources, with little variation in contributions across regions.

Figure S2 illustrates the spatial distribution of emission concentration contributions from various industries in Jilin City for 2019 and 2022. The data indicate that O₃ concentrations in 2022 are higher than in 2019, with traffic heating (TH) being the largest contributor and residential heating (RH) the smallest, consistent with previous research [36,37]. The figure illustrates severe O₃ pollution in YJI and LTAN, with higher concentrations in suburban areas than in the urban core, confirming the conclusions of Tidblad et al. [38]. This can be attributed to limited vegetation in urban centers, where natural precursors from plants promote O₃ formation, making suburban areas more susceptible to O₃ pollution. O₃ pollution in Jilin City exhibits regional characteristics, placing suburban and downwind residents at higher risk from elevated O₃ levels.

3.5. Analysis of Ozone Source Regions in Different Years

This study performed a quantitative simulation of O₃ transmission contributions across six regions of Jilin City, comparing O₃ source characteristics during the COVID-19 epidemic in 2022 with the same period in 2019. The transmission matrix refers to the source of O₃ in each region of the simulation range. The results of the O₃ source region analysis are presented in Table S5.

The results indicate that O₃ concentrations in Jilin City are heavily influenced by long-distance transmission. In 2019, the BCON contribution ranged from 67.36% to 78.52%, while in 2022, it ranged from 68.39% to 79.11%, reflecting distinct regional transmission patterns. Previous studies have identified long-distance O₃ transmission as a key factor in rising O₃ concentrations [16]. BCON primarily transports NO_x, VOCs, and other precursors from outside the inner simulation area to Jilin City over long distances. Additionally, stratospheric O₃ descends and is easily transported by air masses, persisting for extended periods. Given that O₃ formation takes time, BCON, which represents long-distance transmission, significantly contributes to total O₃ levels [39]. The OTH contribution is largely associated with the effect of background methane on O₃ within the inner simulation area, resulting in relatively uniform contributions across receptor regions: 13.52% to 22.72% in 2019 and 12.48% to 20.50% in 2022. ICON reflects the initial conditions, with contributions ranging from 0.41% to 0.52% in 2019 and from 0.39% to 0.49% in 2022, representing a minor proportion of the total.

In 2019, CYING’s contribution to other regions ranged from 1.25% to 2.84%, while CYI contributed between 0.37% and 0.51%, and FMAN’s contribution ranged from 0.30% to 0.51%. LTAN’s contribution ranged from 0.32% to 0.55%, PSHI contributed between 1.10% and 2.13%, and YJI’s contribution ranged from 2.47% to 4.93%. In 2022, CYING’s contribution to other regions ranged from 1.22% to 2.41%, while CYI contributed between 0.26% and 0.59%, and FMAN’s contribution ranged from 0.49% to 0.69%. LTAN’s contribution ranged from 0.46% to 0.65%, PSHI contributed between 1.55% and 2.68%, and YJI’s contribution ranged from 2.35% to 4.46%.

Figure 10 illustrates the transmission matrix of O₃ contributions across various simulation regions in Jilin City, excluding ICON, BCON, and OTH. In the PSHI and YJI regions, local sources are the main contributors to O₃ levels. Conversely, the other four regions are significantly influenced by long-distance transmission and sources from YJI. PSHI emerges as the primary local contributor to O₃ emissions, accounting for 36.7% in 2019 and 39.6% in 2022. YJI is another significant local contributor, with 43.3% in 2019 and 40% in 2022. Furthermore, CYING, CYI, FMAN, and LTAN significantly contribute to O₃ emissions in YJI, with contributions of 40.6%, 38.1%, 41.9%, and 38.7% in 2019, and 38.2%, 33.6%, 39%, and 33.7% in 2022, respectively.

There are significant disparities in the causes and sources of O₃ pollution across various regions of Jilin City, along with notable interactions among these areas. YJI has the greatest influence on surrounding areas; however, this contribution is expected to decline during the lockdown period in 2022. This decrease is attributed to YJI’s role as the primary industrial hub in Jilin City, which promotes the development of a high-quality industrial economy. Industrial activity significantly decreased during the lockdown. To tackle future challenges of O₃ pollution in Jilin City, implementing a regional collaborative prevention and control mechanism is essential.

4. Conclusions

This research analyzes O₃ concentration trends in Jilin City during the COVID-19 lockdown of 2022, comparing them to O₃ levels from the same period in 2019. The WRF-CMAQ model, combined with the ISAM module, was employed to assess industrial and regional contributions to O₃ levels in Jilin City during the 2022 lockdown, in comparison to 2019.

Studies indicate a significant increase in O₃ concentrations following the lockdown compared to before. This increase is attributed to decreased human activities, reduced traffic, and a significant decline in vehicle emissions. Consequently, NO_x concentrations decline, weakening the titration process and decreasing O₃ consumption near the surface. The findings also reveal significant long-range O₃ transport in Jilin City, with BCON’s contribution slightly higher during the lockdown than in the non-lockdown period. Among the four industrial sources, transportation is the primary contributor to local O₃ levels in Jilin City, with minimal variation in contributions across regions. This study identifies six districts in Jilin City, indicating that Panshi City and Yongji County are primarily influenced by local sources, while the other four districts are mainly affected by contributions from Yongji County. Yongji County’s contribution during the lockdown is slightly lower than in the non-lockdown period, as it serves as the main hub for industrial development in Jilin City.

Overall, the sources and causes of O₃ pollution differ across various areas of Jilin City. These findings offer crucial insights for air quality management. Merely reducing anthropogenic NO_x emissions is inadequate and may exacerbate O₃ pollution; thus, controlling precursor emissions represents a more effective strategy. This analysis is also applicable to other countries and regions.

Future research will focus on analyzing the origins of O₃ precursors in Jilin City and exploring variations in O₃ sources across different regional transport pathways. Plans include collecting additional data, refining regional emission inventories, and employing a broader range of methodologies.