key: cord-0071314-98thxgd2
authors: Fang, Chunsheng; Wang, Liyuan; Li, Zhuoqiong; Wang, Ju
title: Spatial Characteristics and Regional Transmission Analysis of PM(2.5) Pollution in Northeast China, 2016–2020
date: 2021-11-26
journal: Int J Environ Res Public Health
DOI: 10.3390/ijerph182312483
sha: f001f187db23d0f046914d8be1fcaabd08845c9f
doc_id: 71314
cord_uid: 98thxgd2

Northeast China is an essential industrial development base in China and the regional air quality is severely affected by PM(2.5) pollution. In this paper, spatial autocorrelation, trajectory clustering, hotspot analysis, PSCF and CWT analysis are used to explore the spatial pollution characteristics of PM(2.5) and determine the atmospheric regional transmission pattern for 40 cities in Northeast China from 2016 to 2020. Analysis of PM(2.5) concentration characteristics in the northeast indicates that the annual average value and total exceedance days of PM(2.5) concentration in Northeast China showed a U-shaped change, with the lowest annual average PM(2.5) concentration (31 μg/m(3)) in 2018, decreasing by 12.1% year-on-year, and the hourly PM(2.5) concentration exploding during the epidemic lockdown period in 2020. A stable PM(2.5) pollution band emerges spatially from the southwest to Northeast China. Spatially, the PM(2.5) in Northeast China has a high degree of autocorrelation and a south-hot–north-cool characteristic, with all hotspots concentrated in the most polluted Liaoning province, which exhibits the H–H cluster pattern and hotspot per year. Analysis of the air mass trajectories, potential source contributions and concentration weight trajectories in Northeast China indicates that more than 74% of the air mass trajectories were transmitted to each other between the three heavily polluted cities, with the highest mean value of PM(2.5) pollution trajectories reaching 222.4 μg/m(3), and the contribution of daily average PM(2.5) concentrations exceeding 60 μg/m(3) within Northeast China. Pollution of PM(2.5) throughout the Northeast is mainly influenced by short-range intra-regional transport(,) with long-range transport between regions also being an essential factor; organized integration is the only fundamental solution to air pollution.

Atmospheric PM 2.5 pollution is of great importance to the international community because of its harmfulness, extensive coverage and being difficult to manage. The World Health Organization has confirmed that PM 2.5 is the most harmful to human health and that it can cause cancer [1] , as well as respiratory and cardiopulmonary diseases [2] , increasing the morbidity and mortality of many diseases [3] . As the world's largest developing country and a major manufacturing country, PM 2.5 pollution in China has always been severe [4] [5] [6] . In 2013, China's State Council issued the "Action Plan for Prevention and Control of Air Pollution", with the expectation to reduce PM 2.5 pollution in five years. Compared with 2015, the number of cities with sub-standard PM 2.5 concentrations and the ratio of heavily polluted days dropped by more than 18% and 25% separately in 2020, while the ratio of good air quality days in prefecture-level and above cities has reached 80%.

Heavy pollution events are caused by a combination of high-intensity emissions and unfavorable meteorological conditions; the mobility of the atmospheric environment determines the regional transport characteristics of atmospheric pollutants, which can be transported from more polluted areas to downstream cities under the effect of meteorological factors, forming regional pollution, and becoming one of the important factors affecting transported from more polluted areas to downstream cities under the effect of met logical factors, forming regional pollution, and becoming one of the important facto fecting regional air quality [7, 8] . Spatial models can be used to study the dispersi pollutants in landfills [9, 10] , stacks [11, 12] , wildfires [13, 14] and urban areas [15] , an important to apply spatial models to study PM2.5. However, previous studies on reg pollution have mainly focused on developed regions in China, such as Beijing-Tia Hebei [16, 17] , Yangtze River Delta [18, 19] and Pearl River Delta regions [20, 21] . Reg composite pollution has currently become a characteristic of air pollution in China.

Research has shown that the degree of pollution in Northeast China could n underestimated, especially in winter and spring [22] , which is mainly caused by th crease in emissions from coal-fired power plants due to the burning of agricultural and the central heating from October to April of the following year [23] . The periph Northeast China is the high mountains and hills of the Greater Khingan, the Lesser gan and Changbai Mountain, the central part is the vast Northeast China Plain (Son Plain, Liaohe Plain, Sanjiang Plain). The special geographical environment causes s atmospheric linkage in Northeast China; indeed, it is very feasible to conduct resear the whole for regional pollution identification and pollution transmission. In recent y many large cities in Northeast China have seen high value areas of PM2.5 pollution the special geographical environment of Northeast China has caused strong atmosp linkage effects and exhibited strong regional characteristics. In this context, it is impo to identify potential sources of fine particulate matter pollution, to fundamentally the regional pollution problem of fine particulate matter in Northeast China, and t termine the regional transport pathways for the scientific prevention and control o particulate matter pollution.

Administratively comprising the provinces of Liaoning, Jilin and Heilongjian the four eastern leagues of Inner Mongolia, Northeast China is rich in natural reso and well-developed agriculturally, making it an important grain-producing regi China. The industry is the backbone of the national economy and main source of revenue in Northeast China [24] . The geographical location of the study area is sho Figure 1 . This study used PM 2.5 hourly concentration data from a total of 177 air quality monitoring stations in 40 cities in Northeast China from 2016 to 2020 [25] . The full data analysis of PM 2.5 is based on the National Ambient Air Quality Standard (NAAQS, GB 3095-2012). If the hourly value and annual mean value of PM 2.5 exceed 75 and 35 µg/m 3 , it shall 3 of 15 be regarded as exceeding the standard. Data were excluded including missing hourly PM 2.5 concentration values from the original monitoring station, daily average PM 2.5 concentration values with less than 20 h of valid data per day and monthly average PM 2.5 concentration values with less than 27 days of valid data per month (25 days in February). The effective rate of average daily PM 2.5 in the study area was 96.3% and 90.3% for monthly average values. The fire point data used in this study were obtained from the near real-time surface high temperature anomaly dataset SatSee-Fire published by the Institute of Remote Sensing and Digital Earth of the Chinese Academy of Sciences (CAS). The spatial vector data and arable land patches of county administrative regions in China were derived from the Resource and Environment Science Data Center (RESDC) of the CAS (https://www.resdc.cn/) (accessed on 16 July 2021), where the arable land vector patches were extracted from the 2015 remote sensing monitoring data of the current land use in China.

Spatial autocorrelation refers to the dependence of geographical things or attributes on spatial location, and the closer the spatial location the stronger the correlation [26] . This study uses global and local Moran s I to characterize the spatial autocorrelation of PM 2.5 concentrations in Northeast China. The positive and negative I values indicate a positively or negatively correlated spatial distribution of pollutant concentrations. The magnitude indicates that similar concentration values (high or low values) tend to be spatially clustered or dispersed. High-value adjacent elements clustered near high-value elements are called high-high (H-H); otherwise, low-low (L-L); the adjacent elements clustered near the highvalue elements with low values are called high-low (H-L); otherwise, low-high (L-H), and I = 0 means the spatial distribution of concentration is random. The specific calculation method of I value is shown in Equation (1):

In Equation (1), x i and x j are the fine particulate matter concentration values of cities i and j in the study area, respectively, ω ij is the spatial weight matrix, n is the number of cities in the study area, and X is the mean value of fine particulate matter concentration in the study area. The ω ij was calculated based on whether two cities in the study area are adjacent to each other. If two cities are adjacent, the number "1" is defined, and if they are not adjacent, the number "0" is defined and so on, forming a matrix of spatial weights for all cities in the region.

After using the global Moran s I to reflect the clustering or dispersion dynamics of the whole study area, we analyzed the clustering or dispersion effects of fine particle concentrations in local areas by calculating the G * i values of the Getis-Ord index, and detects the specific locations of hot and cold spot areas to reflect the specific situation of the local area [27, 28] . The calculation method is shown in Equation (2):

In Equation (2), x j is the concentration of fine particles in city j; ω ij is the spatial weight matrix; S is the standard deviation; X is the mean value of fine particles concentration in the study area, and n is the number of cities. The positive or negative G * i value indicates that the area is a concentration zone of high (hot spot) or low (cold spot) values.

The hybrid single particle Lagrangian integrated irajectory (HYSPLIT) model [29] developed by the National Oceanic and Atmospheric Administration (NOAA) and the Australian Bureau of Meteorology (BOM) was used to simulate 48 h backward trajectories at 500 m altitude for three polluted provincial capitals (Harbin, Changchun, and Shenyang) in Northeast China to analyze atmospheric pollutant transport [30] . In this paper, the Euclidean distance clustering algorithm in TrajStat software, which combines the HYSPLIT model and GIS technology, was used to cluster the airflow trajectories arriving at three provincial capitals in Northeast China to obtain different clustering results for the three cities [31, 32] .

This study used potential source contribution function (PSCF) analysis to locate pollution sources using backward trajectories. In order to reduce the error and make the results more accurate and practical, a weighting factor W ij was introduced, the value of which depends on the relationship between the sum of the transmission times of all trajectories in a given grid and the average residence time of each grid, W ij is expressed in Equation (3).

The weighting factor W ij used in this study reduces the uncertainty of PSCF results and is called WPSCF [33] . The grid resolution of PSCF was set to 0.5 • × 0.5 • [34] . The PM 2.5 limit value used in the study is 75 µg/m 3 , which is the limit value of the secondary standard of the National Ambient Air Quality Standard (NAAQS) established by the Ministry of Ecology and Environment of China.

The concentration-weighted trajectory (CWT) method can quantify the concentration contribution level of external transport by obtaining the average value of the concentration of samples corresponding to all trajectories passing through a single grid during the study period [35] . The CWT analysis method can obtain the difference in the pollution level of contaminated trajectories by calculating the weighted degree. In the CWT analysis method, each grid point is assigned a degree of weight [36] . By introducing the same numerical correction from W ij in the PSCF method, the weighted average concentration value (WCWT value) can be used to distinguish the source intensity of potential sources. A higher WCWT value in the grid indicates that the air mass passing through the grid results in a high receiving point concentration, and the area corresponding to this grid can be considered as a potential area of high concentration contribution to the external transport of pollutants from the receiving area. 

Benefiting from the country's determination to reduce pollutant emissions, China's key regions have seen significant improvements in pollution and air quality, with numerous studies reporting that 2018 was the most effective year for combating PM 2.5 pollution in China [37] ; the annual average PM 2.5 concentration reached the lowest value in recent years (31 µg/m 3 ), decreasing by 12.1% year-on-year, and the maximum number of exceedance days was 61 days in Jinzhou, totaling 797 days in all 40 cities (Table S1 in Supplementary Materials), of which Liaoning Province accounted for 55%. However, Figure 2 shows that the average annual PM 2.5 concentration and the total number of polluted days in Northeast China showed a U-shaped variation in the previous five years, with a slight rebound after 2018, and the total number of exceedance days increasing 30% in 2020, with an annual average PM 2.5 concentration of 32 µg/m 3 . Excellent weather conditions are due to China's initiatives to transform, remediate, eliminate and clean up the industry in accordance with its industrial structure [38] . Initiatives in Northeast China have focused on managing open biomass burning and coal combustion emissions during the heating period. Figure 3 shows the fire point data set for March, April, October and November in Northeast China yearly, which has a suitable identification of abnormally high-temperature points, including straw and burning fire points [39] . The result shows that the number of fire points increased significantly to a peak in 2017, with these four months accounting for approximately 83% of the year, of which 91% are concentrated on the subsurface of farmland. key regions have seen significant improvements in pollution and air quality, with numerous studies reporting that 2018 was the most effective year for combating PM2.5 pollution in China [37] ; the annual average PM2.5 concentration reached the lowest value in recent years (31 μg/m 3 ), decreasing by 12.1% year-on-year, and the maximum number of exceedance days was 61 days in Jinzhou, totaling 797 days in all 40 cities (Table S1 in Supplementary Materials), of which Liaoning Province accounted for 55%. However, Figure 2 shows that the average annual PM2.5 concentration and the total number of polluted days in Northeast China showed a U-shaped variation in the previous five years, with a slight rebound after 2018, and the total number of exceedance days increasing 30% in 2020, with an annual average PM2.5 concentration of 32 μg/m 3 . Excellent weather conditions are due to China's initiatives to transform, remediate, eliminate and clean up the industry in accordance with its industrial structure [38] . Initiatives in Northeast China have focused on managing open biomass burning and coal combustion emissions during the heating period. Figure 3 shows the fire point data set for March, April, October and November in Northeast China yearly, which has a suitable identification of abnormally high-temperature points, including straw and burning fire points [39] . The result shows that the number of fire points increased significantly to a peak in 2017, with these four months accounting for approximately 83% of the year, of which 91% are concentrated on the subsurface of farmland. The stringent treatment of local governments in Northeast China has sharply reduced straw burning in the region. By 2020, the number of fires across the region had been reduced by 68% from their peak, with a 97% reduction in the two winter months. It is fair to say that reasonable control and organized straw burning under favorable diffusion conditions have effectively controlled pollution. The statistics in Figure 2 show that the pollution in winter in the Northeast China has been controlled, with an average PM 2.5 concentration of 44.6 µg/m 3 during the heating period. Meanwhile, the PM 2.5 pollution has started to rebound in several regions in Northeast China after 2018, with the 95% quantile being the most pronounced, increasing at a rate of 1-2% per year.

Although emission reduction has been the dominant factor in China's air quality improvement in recent years, and studies in previous years have tended to attenuate the influence of meteorological factors, it is now generally accepted that the implementation of regional joint prevention and control of air pollution has been an effective means of addressing air pollution prevention [40, 41] . As shown in Figure 2 , the PM 2.5 pollution assessment of 177 state-controlled monitoring stations showed, with the exception of individual stations (Jixi Water Company and Baijiu Factory Station, Jixi, China) where PM 2.5 concentrations exceeded the national ambient air quality standards by approximately 15% in 2020, all other stations meet the standards, of which 42% of the stations met the primary standard (35 µg/m 3 ). Higher PM 2.5 concentrations were concentrated in three provincial capitals and surrounding cities, of which Jilin Province is sandwiched in the middle and has slightly better indicators than the other two provinces caused by the transmission from the highly polluted regions [42] . Previous studies concerning the perennial air mass trajectory movement pattern prove have shown that more than 60% of the polluted air masses come from Inner Mongolia in the north direction of Northeast China [43] . In contrast, Liaoning Province borders Bohai Bay and variable weather systems following the reduction of local emissions have resulted in the region being more affected by meteorological transport. Strong northerly winds in winter and spring rapidly import polluted air masses from the Beijing-Tianjin-Hebei region upstream into Liaoning Province. Rapids transport and continuous water vapor accumulation due to the difference in pressure between land and incoming sea in summer favors the transport and stagnation of pollutants over Liaoning Province [44] . The stringent treatment of local governments in Northeast China has sharply reduced straw burning in the region. By 2020, the number of fires across the region had been reduced by 68% from their peak, with a 97% reduction in the two winter months. It is fair to say that reasonable control and organized straw burning under favorable diffusion conditions have effectively controlled pollution. The statistics in Figure 2 show that the pollution in winter in the Northeast China has been controlled, with an average PM2.5 concentration of 44.6 μg/m 3 during the heating period. Meanwhile, the PM2.5 pollution has started to rebound in several regions in Northeast China after 2018, with the 95% quantile being the most pronounced, increasing at a rate of 1-2% per year.

Although emission reduction has been the dominant factor in China's air quality improvement in recent years, and studies in previous years have tended to attenuate the influence of meteorological factors, it is now generally accepted that the implementation of regional joint prevention and control of air pollution has been an effective means of addressing air pollution prevention [40, 41] . As shown in Figure 2 

The yearly hour PM 2.5 pollutant concentration distribution throughout Northeast China ( Figure 4 ) and each city (Figures S1-S5 in Supplementary Materials) shows that the beginning and end of each year are the most polluted periods for PM 2.5 pollution. In a side-by-side comparison, 2018 remains the least polluted year, with maximum hourly PM 2.5 concentrations around 50% lower than 2016. 2020 contributes the highest average January concentrations in the Northeast China in those years at 88µg/m 3 . As a result of the epidemic lockdown, the country was completely closed in February and March in 2020. When operations resumed in various industries in early April, severe PM 2.5 pollution broke out for 7 days in Northeast China with the most polluted day reaching a daily average of 155.9 µg/m 3 , which peaked at 807 µg/m 3 on 13 April in Qiqihar, and hourly PM 2.5 concentrations in individual cities even exceeding 1000 µg/m 3 on 18 April. This outbreak was still associated with straw burning, which is commonly carried out on a large scale in the region before spring plowing. At the same time, unfavorable meteorological conditions of rising relative humidity and ground-level winds of less than 2 m/s in Northeast China occurred during this period, resulting in air quality maintaining heavy pollution levels for a long time [45] . In previous years, PM 2.5 pollution levels in Northeast China always spiked to a peak at the beginning of the heating period and decreased until the end of the following spring, which was higher in the end of each year from October to December than that of the current and following year from January to April. Nevertheless, it becomes the case that PM 2.5 pollution is higher in the first half of the year than in the second half since 2017, a situation that is more clearly reflected in the results of the fan chart in Figure 4 , which shows the average values each month for the 5 years. In general, the highest PM 2.5 pollution period is from October to April, As the temperature rises, the boundary layer rises, turbulent vertical exchange is enhanced and PM 2.5 diffusion is enhanced [46] , a rare low-pollution moment of the day for PM 2.5 is between 15:00 and 17:00. Later in the evening, as the temperature decreases, the boundary layer height decreases, the turbulent vertical exchange weakens and the PM 2.5 concentration increases. the epidemic lockdown, the country was completely closed in February and March in 2020. When operations resumed in various industries in early April, severe PM2.5 pollution broke out for 7 days in Northeast China with the most polluted day reaching a daily average of 155.9 μg/m 3 , which peaked at 807 μg/m 3 on 13 April in Qiqihar, and hourly PM2.5 concentrations in individual cities even exceeding 1000 μg/m 3 on 18 April. This outbreak was still associated with straw burning, which is commonly carried out on a large scale in the region before spring plowing. At the same time, unfavorable meteorological conditions of rising relative humidity and ground-level winds of less than 2 m/s in Northeast China occurred during this period, resulting in air quality maintaining heavy pollution levels for a long time [45] . In previous years, PM2.5 pollution levels in Northeast China always spiked to a peak at the beginning of the heating period and decreased until the end of the following spring, which was higher in the end of each year from October to December than that of the current and following year from January to April. Nevertheless, it becomes the case that PM2.5 pollution is higher in the first half of the year than in the second half since 2017, a situation that is more clearly reflected in the results of the fan chart in Figure 4 , which shows the average values each month for the 5 years. In general, the highest PM2.5 pollution period is from October to April, As the temperature rises, the boundary layer rises, turbulent vertical exchange is enhanced and PM2.5 diffusion is enhanced [46] , a rare low-pollution moment of the day for PM2.5 is between 15:00 and 17:00. Later in the evening, as the temperature decreases, the boundary layer height decreases, the turbulent vertical exchange weakens and the PM2.5 concentration increases. The spatial distribution of annual average PM 2.5 concentrations in Northeast China is shown in Figure 5 . It is evident that the ambient atmosphere in the Northeast China has experienced a contiguous massive patch of PM 2.5 pollution yearly and monthly (Figures 1-5 attached) for the past five years. The heavily polluted areas are concentrated in and around the three provincial capitals, with the polluted areas extending in a band from the Southwest to Northeast China. In 2020, there is a center of high concentration appears in Jixi City, Heilongjiang Province, where the concentration is more than 70% higher than the surrounding area. The analysis of PM 2.5 spatial distribution proves that a linked atmospheric pollution belt has formed in Northeast China and regional transport has become a significant factor in local pollution sources. The influence of regional transmission accounts for more than 70% in the heavy pollution. Northeast China must tighten its grip on joint prevention and control while addressing local pollution in order to achieve PM 2.5 concentrations at new lows, through improvement of air quality and control Jixi City, Heilongjiang Province, where the concentration is more than 70% higher than the surrounding area. The analysis of PM2.5 spatial distribution proves that a linked atmospheric pollution belt has formed in Northeast China and regional transport has become a significant factor in local pollution sources. The influence of regional transmission accounts for more than 70% in the heavy pollution. Northeast China must tighten its grip on joint prevention and control while addressing local pollution in order to achieve PM2.5 concentrations at new lows, through improvement of air quality and control 

The results of the global spatial autocorrelation analysis of PM2.5 concentrations indicated that the Moran′s I > 0 for PM2.5 concentrations in Northeast China from 2016 to 2020, and all passed the significance test of 99%, with the Z scores were well above the critical value of 2.58 [47] . Therefore, the distribution of PM2.5 concentrations in Northeast China has a significant positive spatial autocorrelation, and cities with similar concentrations tend to be clustered, which has become more assertive in recent years. The local spatial autocorrelation analysis in Figure 6 reveals that the northwestern part of Northeast China is always in the low-low (L-L) clustering category, and the high-high (H-H) clustering phenomenon is always found in the pollution belt of Liaoning Province in the northeastward extension of Liaodong Bay. Around these two kinds of clustering areas, low-high (L-H) and high-low (H-L) clustering emerged in 2017 and 2019 respectively, indicating that cities are increasingly influencing each other. With the pollution situation of a single city being influenced by the surroundings evolving into a convergence pattern, the PM2.5 clustering conditions of individual cities in Northeast China are becoming more consistent from 2018 to 2020.

Using Getis-Ord G * to identify clustering areas with positive spatial autocorrelation, the results show that PM2.5 concentrations in Northeast China in the past five years have 

The results of the global spatial autocorrelation analysis of PM 2.5 concentrations indicated that the Moran s I > 0 for PM 2.5 concentrations in Northeast China from 2016 to 2020, and all passed the significance test of 99%, with the Z scores were well above the critical value of 2.58 [47] . Therefore, the distribution of PM 2.5 concentrations in Northeast China has a significant positive spatial autocorrelation, and cities with similar concentrations tend to be clustered, which has become more assertive in recent years. The local spatial autocorrelation analysis in Figure 6 reveals that the northwestern part of Northeast China is always in the low-low (L-L) clustering category, and the high-high (H-H) clustering phenomenon is always found in the pollution belt of Liaoning Province in the northeastward extension of Liaodong Bay. Around these two kinds of clustering areas, low-high (L-H) and high-low (H-L) clustering emerged in 2017 and 2019 respectively, indicating that cities are increasingly influencing each other. With the pollution situation of a single city being influenced by the surroundings evolving into a convergence pattern, the PM 2.5 clustering conditions of individual cities in Northeast China are becoming more consistent from 2018 to 2020. 

The above analysis indicates that the regional transport characteristics of Northeast China are already formed. It is necessary to reveal its pollution transport patterns. Consequently, three heavily polluted cities in Northeast China (Harbin, Changchun and Shen- Using Getis-Ord G * i to identify clustering areas with positive spatial autocorrelation, the results show that PM 2.5 concentrations in Northeast China in the past five years have a clustering characteristic of being hot in the south and cold in the north. The hotspot cities that passed the significance test for annual average PM 2.5 concentrations were mainly distributed in Liaoning Province, indicating that these regions are indeed the high PM 2.5 pollution areas in Northeast China. The cold spot cities are concentrated in the western part of Heilongjiang Province and eastern part of Inner Mongolia, with excellent ambient air quality in Northeast China. The hotspot areas in Liaoning Province have gradually shrunk in the last three years, although the distribution of cold spot cities is stable, besides spread out tendency for PM 2.5 concentration due to pollution from surrounding cities.

The above analysis indicates that the regional transport characteristics of Northeast China are already formed. It is necessary to reveal its pollution transport patterns. Consequently, three heavily polluted cities in Northeast China (Harbin, Changchun and Shenyang), which are also provincial capitals, were selected for more focused discussion in 2020. We apply the HYSPLIT model to determine the transport paths of atmospheric pollutants and the results are shown in Figure 7 [48, 49] . All trajectories were clustered into four entries named cluster C1 to C4. Cluster C2 carries the most significant number of trajectories in Harbin City and accounting for 39.3% of the total 8784 trajectories in 366 days multiplied by 24 h, which also wrapped the most polluting trajectories totaling 787, whose average PM 2.5 pollution concentration reached 186.6 µg/m 3 , nearly two times higher than the total average concentration. Cluster C2 originates from the southwest and the western part of Jilin Province, and then blows straightly through Changchun, with a shorter path and moves more slowly, carrying pollutants from Jilin Province to Heilongjiang Province. The trajectory clustering results for Changchun show that about 58% of trajectories come from the northwest, with the most significant proportion of the Cluster C2 (33%) coming from Northeast Inner Mongolia. However, the Cluster C4 comes from southern Heilongjiang Province with a minor proportion (15.8%) and contributes more polluting trajectories with an average PM 2.5 concentration of 222.4 µg/m 3 . The Cluster C3 in Shenyang accounted for the largest proportion and wrapped the least polluted trajectory (9.9%), but the average PM 2.5 of the polluted trajectory was as high as 113.8µg/m 3 , which is strongly related to the polluted air mass from the intersection of the Yellow Sea and Bohai Sea.

Generally, the majority of the air mass paths from these three cities transmit within the study region, transporting and influencing local air quality with each other. However, 26% of the trajectories originate from Mongolia, which is outside the study area. Numerous studies have demonstrated that long-range transport from Mongolia has been the largest source of air pollution affecting Jilin Province for many years [50] . In this paper, the analysis within Northeast China proves that short-range regional transport instead transmits more air pollutants than long-range cross-regional transport, which indicates that delineating local areas for joint pollution control is an essential option. Generally, the majority of the air mass paths from these three cities transmit within the study region, transporting and influencing local air quality with each other. However, 26% of the trajectories originate from Mongolia, which is outside the study area. Numerous studies have demonstrated that long-range transport from Mongolia has been the largest source of air pollution affecting Jilin Province for many years [50] . In this paper, the analysis within Northeast China proves that short-range regional transport instead transmits more air pollutants than long-range cross-regional transport, which indicates that delineating local areas for joint pollution control is an essential option.

The cluster analysis of the trajectories has clearly determined the transmit direction of trajectories over Northeast China, the proportion and the concentration of the polluted trajectories. In order to specifically determine the relative contribution of potential source areas of the study area, we conducted PSCF analysis of PM2.5 in Northeast China throughout 2020, while the CWT method was used to analyze the concentration weight trajectories, which better reflects the specific contribution concentration of the pollution grid [51, 52] . For the analysis of the results shown in Figure 8 , it was found that the main potential source areas (PSCF > 0.4) in Harbin were distributed at the junction of the Yellow 

The cluster analysis of the trajectories has clearly determined the transmit direction of trajectories over Northeast China, the proportion and the concentration of the polluted trajectories. In order to specifically determine the relative contribution of potential source areas of the study area, we conducted PSCF analysis of PM 2.5 in Northeast China throughout 2020, while the CWT method was used to analyze the concentration weight trajectories, which better reflects the specific contribution concentration of the pollution grid [51, 52] . For the analysis of the results shown in Figure 8 , it was found that the main potential source areas (PSCF > 0.4) in Harbin were distributed at the junction of the Yellow Sea and the Bohai Sea with the eastern part of Korea. The distribution of moderate potential source areas (0.3 < PSCF < 0.4) in Harbin is concentrated in the northwestern part of Jilin Province and the border between Inner Mongolia and Liaoning Province, the high-value areas of the PM 2.5 concentration weighting trajectory are also distributed in this range with a daily average PM 2.5 concentration contribution of 70-80 µg/m 3 . CWT analysis reveals the important influence of short-range regional transport on air pollution, Jixi City, located directly east of Heilongjiang Province, contributes more than 100 µg/m 3 to the daily average PM 2.5 concentrations in Harbin. tial source areas (0.3 < PSCF < 0.4) in Harbin is concentrated in the northwestern part of Jilin Province and the border between Inner Mongolia and Liaoning Province, the highvalue areas of the PM2.5 concentration weighting trajectory are also distributed in this range with a daily average PM2.5 concentration contribution of 70-80 μg/m 3 . CWT analysis reveals the important influence of short-range regional transport on air pollution, Jixi City, located directly east of Heilongjiang Province, contributes more than 100 μg/m 3 to the daily average PM2.5 concentrations in Harbin. Mongolia region. This shows that the influence of regional transmission caused by atmospheric flow is powerful between regions and regions themselves. Therefore, dividing the whole region and organized linkage and integration is the fundamental solution to air pollution.

The results of identifying the characteristics of PM2.5 concentration variations in Northeast China from 2016 to 2020 show that the PM2.5 emission reduction effect of straw burning and coal-fired heating is significant, with an average PM2.5 concentration of 44.6 μg/m 3 in spring and winter. The average annual PM2.5 concentration and the total number of polluted days in Northeast China from 2016 to 2020 show a U-shaped change, with about 42% of monitoring stations meeting a primary air quality standard. Northeast China had the best air quality in 2018. The annual average PM2.5 concentration reached a minimum of 31μg/m 3 and decreased 12.1% year-on-year, after which the indicators rebounded slightly, with the annual average PM2.5 concentration increasing at an annual rate of 1-2%. Inner Mongolia region. This shows that the influence of regional transmission caused by atmospheric flow is powerful between regions and regions themselves. Therefore, dividing the whole region and organized linkage and integration is the fundamental solution to air pollution.

The results of identifying the characteristics of PM 2.5 concentration variations in Northeast China from 2016 to 2020 show that the PM 2.5 emission reduction effect of straw burning and coal-fired heating is significant, with an average PM 2.5 concentration of 44.6 µg/m 3 in spring and winter. The average annual PM 2.5 concentration and the total number of polluted days in Northeast China from 2016 to 2020 show a U-shaped change, with about 42% of monitoring stations meeting a primary air quality standard. Northeast China had the best air quality in 2018. The annual average PM 2.5 concentration reached a minimum of 31µg/m 3 and decreased 12.1% year-on-year, after which the indicators rebounded slightly, with the annual average PM 2.5 concentration increasing at an annual rate of 1-2%. An outbreak of PM 2.5 pollution in January and April in 2020 due to the epidemic lockdown, with the highest daily average value across the Northeast China reached 155.9 µg/m 3 . In contrast to this phenomenon in northeast China, studies in the Beijing-Tianjin-Hebei (BTH) region and the North China Plain (NCP) in China also found an increase in PM 2.5 concentrations during the lockdown, which was related to anthropogenic sources such as heating and fireworks emissions and straw burning during the pandemic [53] [54] [55] . In other countries such as the United States and India, decreases in PM 2.5 concentrations were observed to varying degrees [56, 57] . Spatially, the pollution in Northeast China has become contiguous, with spatial interpolation results indicating the emergence of a pollution belt in Northeast China extending from the Southwest to Northeast China, landing in Liaoning Province from Liaodong Bay, connecting the three provincial capitals and running southwest towards Heilongjiang Province. The whole Northeast China shows a high degree of spatial autocorrelation, and exhibits the clustering characteristics of south-hot-north-cool. Coving all hotspot areas, Liaoning Province is the most polluted region in terms of PM 2.5 and shows H-H clustering.

For the three heavily polluted cities in 2020, the analysis of the results found that more than 74% of the backward trajectories were transported between these three provincial capitals. Forty percent of the air masses in Harbin originate in the western part of Jilin Province and carry the most pollution backward trajectories, with average PM 2.5 concentrations reaching 186.6 µg/m 3 . Most of the backward trajectories transport through Changchun are from Inner Mongolia, but the most pollution trajectories come from southern Heilongjiang province, with average PM 2.5 concentrations in these polluting air masses reaching 222.4 µg/m 3 . The pollution in Shenyang is influenced by multiple factors, including air masses from Inner Mongolia and the pollution belts in the study area, while air masses from the interface between the Yellow Sea and the Bohai Sea also contribute a large proportion of PM 2.5 pollution. Similar results were obtained in the studies for the central cities of Liaoning Province as well as Shenyang City, suggesting some commonality in the results of backward trajectory analysis in Northeast China [58, 59] .

The results of analyzing the potential source contribution and concentration weight trajectories of the three heavily polluted cities in 2020 show that the distribution of potential source areas in Harbin City is relatively concentrated, with the main potential source areas distributed at the junction of the Yellow Sea and the Bohai Sea with the eastern part of North Korea. The moderate potential source areas are distributed in the northwestern part of Jilin Province and the junction of Inner Mongolia and Liaoning Province, consistent with the distribution range of the concentration weight trajectories, which contribute daily average PM 2.5 concentrations over 70 µg/m 3 . There are no major potential source areas of PM 2.5 pollution in Changchun, and the light pollution grids are mainly distributed in the high pollution belt in Northeast China, with a contribution of 60-70 µg/m 3 of daily average PM 2.5 concentration, and a part of the grids with a contribution of more than 90 µg/m 3 scattered in the east side of Korea. The PM 2.5 pollution grid in Shenyang is loosely distributed across Northeast China and Eastern Inner Mongolia. Most of the high PM 2.5 concentration weighting trajectory areas are located in the northeast China pollution belt, with a small proportion coming from the heavily polluted Inner Mongolia region. In the study for Beijing, some similarity was found between the potential source areas and high concentration contribution areas in Beijing and the results for Northeast China; higher results values were also found in Shandong, Henan and Hebei [60] . In the study for the southern cities, the main potential source areas and high contribution areas were concentrated in the Yangtze River Delta region [61] .

Throughout China, there is an apparent regional pollution pattern that has emerged. Meanwhile, numerous large cities in Northeast China are considered high PM 2.5 pollution areas and have shown solid regional characteristics for pollution. In this context, a scientific and detailed anatomizing of the sources of PM 2.5 is essential to resolve the regional pollution in the Northeast China fundamentally and to identify regional transmission pathways for the scientific prevention and control of PM 2.5 pollution. It is of tremendous research significance to explore the spatial pollution characteristics of PM 2.5 in Northeast China and determine its atmospheric regional transmission pattern, providing a basis for the study of linked atmospheric pollution and prevention and control measures. Straw burning has been a major contributor to the occurrence of heavy pollution in the northeast from the beginning to end. Local governments in the northeast have invested heavily in various monitoring, guidance, and accountability approaches and have largely contributed to tackling the straw-burning problem. Importantly, this study clarifies that PM 2.5 pollution in the northeast has spatial clustering characteristics and that Liaoning Province is the most heavily polluted hot spot for PM 2.5 pollution.

The internal pollution trajectories in Northeast China transmit an exceptionally high proportion of pollution; the main potential source areas of pollution are spread across the province and parts of the surrounding countries. This paper also identifies a high pollution belt in the Northeast, which has the most severe PM 2.5 pollution and most robust spatial clustering, seriously affecting the air quality of the surrounding region. The findings of this paper are important for research on the regional linkage of air pollution transmission corridors in the northeast, which is consistent with the strategic trend of China's future air quality objectives, which is the sub-regional joint prevention and control of air pollution. 

Fine and coarse particulate air pollution in relation to respiratory health in Sweden

Associations of Mortality with Long-Term Exposures to Fine and Ultrafine Particles, Species and Sources: Results from the California Teachers Study Cohort

The association between long-term exposure to low-level PM 2.5 and mortality in the state of Queensland, Australia: A modelling study with the difference-in-differences approach

A framework for delineating the regional boundaries of PM 2.5 pollution: A case study of China

Impacts of Industrial Restructuring and Technological Progress on PM 2.5 Pollution: Evidence from Prefecture-Level Cities in China

PM 2.5 pollution-related health effects and willingness to pay for improved air quality: Evidence from China's prefecture-level cities

Influence of meteorological conditions on PM 2.5 concentrations across China: A review of methodology and mechanism

Understanding meteorological influences on PM 2.5 concentrations across China: A temporal and spatial perspective

Spatial modelling of gaseous emissions from two municipal solid waste dump sites

Dynamics of PM 2.5 pollution in the vicinity of the old municipal solid waste dumpsite

Dispersion Effects of Particulate Lead (Pb) from the Stack of a Lead Battery Recycling Plant

Atmospheric Dispersion of PM 2.5 Precursor Gases from Two Major Thermal Power Plants in Andhra Pradesh

Improving spatial resolution of PM 2.5 measurements during wildfires

Application of learning vector quantization and different machine learning techniques to assessing forest fire influence factors and spatial modelling

The role of atmospheric dispersion in the seasonal variation of PM1 and PM 2.5 concentration and composition in the urban area of Milan (Italy)

Regional transport in Beijing-Tianjin-Hebei region and its changes during 2014-2017: The impacts of meteorology and emission reduction

Numerical modeling of regional transport of PM 2.5 during a severe pollution event in the Beijing-Tianjin-Hebei region in

Air pollution characteristics and their relationship with emissions and meteorology in the Yangtze River Delta region during 2014-2016

Real-time numerical source apportionment of PM 2.5 concentrations over the Yangtze River Delta region

Source and exposure apportionments of ambient PM 2.5 under different synoptic patterns in the Pearl River Delta region

Differences in concentration and source apportionment of PM 2.5 between 2006 and 2015 over the PRD region in southern China

Temporal and spatial analyses of particulate matter (PM10 and PM 2.5 ) and its relationship with meteorological parameters over an urban city in northeast China

Distribution, bioaccessibility, and health risk assessment of heavy metals in PM 2.5 and PM10 during winter heating periods in five types of cities in Northeast China

Evolution of the Process of Urban Spatial and Temporal Patterns and its Influencing Factors in Northeast China

China Air Quality Online Monitoring and Analysis Database

Spatiotemporal characteristics of urban air quality in China and geographic detection of their determinants

Modeling spatial patterns of link-based PM 2.5 emissions and subsequent human exposure in a large canadian metropolitan area

Spatio-temporal evolution of ozone pollution and its influencing factors in the Beijing-Tianjin-Hebei Urban Agglomeration

Air Resources Laboratory-HYSPLIT-Hybrid Single Particle Lagrangian Integrated Trajectory Model

Monitoring the long-range transport effects on urban PM10 levels using 3D clusters of backward trajectories

Use of backward trajectories to interpret the 5-year recordof PAN and 03ambient air concentrations at Kejimkujik National Park, Nova Scotia

GIS-based software that uses various trajectory statistical analysis methods to identify potential sources from long-term air pollution measurement data

Indicators reflecting local and transboundary sources of PM 2.5 and PM COARSE in Rome-Impacts in air quality

The aerosol at Barrow, Alaska: Long-term trends and source locations

Air pollution episodes during the COVID-19 outbreak in the Beijing-Tianjin-Hebei region of China: An insight into the transport pathways and source distribution

Potential Source Regions and Transportation Pathways of Reactive Gases at a Regional Background Site in Northwestern China

Spatio-Temporal Variations of the PM 2.5 /PM10 Ratios and Its Application to Air Pollution Type Classification in China

The relationships between PM 2.5 and aerosol optical depth (AOD) in mainland China: About and behind the spatio-temporal variations

Spatial and temporal variability of open biomass burning in Northeast China from

Can regional integration control transboundary water pollution? A test from the Yangtze River economic belt

Energy use, industrial soot and vehicle exhaust pollution-China's regional air pollution recognition, performance decomposition and governance

Evaluating the impacts of burning biomass on PM 2.5 regional transport under various emission conditions

Analysis of Air Quality Changes and Influencing Factors in Changchun during the COVID-19 Pandemic in 2020

Assessing the Impact of Lockdown on Atmospheric Ozone Pollution Amid the First Half of 2020 in Shenyang, China

How aerosol transport from the North China plain contributes to air quality in northeast China

Pollution Remediation by Urban Forests: PM 2.5 Reduction in Beijing

Moran's I statistic-based nonparametric test with spatio-temporal observations

Incorporating Features from the Stochastic Time-Inverted Lagrangian Transport (STILT) Model into the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) Model: A Unified Dispersion Model for Time-Forward and Time-Reversed Applications

NOAA's HYSPLIT Atmospheric Transport and Dispersion Modeling System

Long-term characteristics of criteria air pollutants in megacities of Harbin-Changchun megalopolis, Northeast China: Spatiotemporal variations, source analysis, and meteorological effects

Aerosol vertical distribution and sources estimation at a site of the Yangtze River Delta region of China

Evolution of the vertical structure of air pollutants during winter heavy pollution episodes: The role of regional transport and potential sources

Significant concurrent decrease in PM 2.5 and NO 2 concentrations in China during COVID-19 epidemic

Impacts of COVID-19 on air quality in mid-eastern China: An insight into meteorology and emissions

Substantial changes of chemical composition and sources of fine particles during the period of COVID-19 pandemic in Taiyuan, Northern China

Examining Effects of the COVID-19 National Lockdown on Ambient Air Quality across Urban India

Changes in U.S. air pollution during the COVID-19 pandemic

Air pollutant emission characteristics and HYSPLIT model analysis during heating period in Shenyang

Characteristics of Boundary Layer Structure during a Persistent Haze Event in the Central Liaoning City Cluster, Northeast China

Continuous monitoring, compositions analysis and the implication of regional transport for submicron and fine aerosols in Beijing, China

Meteorological and chemical impacts on PM 2.5 during a haze episode in a heavily polluted basin city of eastern China

The authors would like to thank the Ecological Environment Monitoring Center of Changchun, Jilin Province, for providing the data on pollutants. Additionally, the authors would like to thank the group members of Laboratory 537 and 142 of Jilin University.

The authors declare no conflict of interest.