Spatio-temporal Variations in NO 2 and PM 2 . 5 over the Central Plains Economic Region of China during 2005-2015 Based on Satellite Observations

The Central Plains Economic Region (CPER) is located in central eastern China and has experienced tremendous economic growth in the last decade. Like many areas in China, the rapid economic development of the CPER has led to serious air pollution problems, including extremely high concentrations of nitrogen dioxide (NO2) and particulate matter with an aerodynamic diameter less than 2.5 μm (PM2.5). However, the current air pollution monitoring system lacks good spatial and temporal coverage. Therefore, we used high-resolution satellite remote sensing techniques to analyze the variation in NO2 and PM2.5 in the CPER from 2005 through 2015. The Ozone Monitoring Instrument (OMI) and the Moderate Resolution Imaging Spectroradiometer (MODIS) were used to retrieve the tropospheric NO2 columns and ground-level PM2.5 concentrations, respectively. High NO2 and PM2.5 concentrations were mainly located in the central and northern areas of the CPER. The highest 11-year average concentrations were found in the city of Jiaozuo for NO2 (19.54 × 10 molecules cm) and in the city of Hebi for PM2.5 (107.8 μg m). The western and southern mountainous areas had lower NO2 and PM2.5 concentrations. The average seasonal NO2 and PM2.5 concentrations were both highest in winter and lowest in summer. The average monthly concentrations of NO2 and PM2.5 were as high as 24.60 × 10 molecules cm and 152.2 μg m, respectively, in January 2013 and as low as 43.86 × 10 molecules cm and 40.2 μg m, respectively, in July 2006. During the 11-year study period (2005–2015), the CPER concentrations of both NO2 and PM2.5 decreased from 2011 to 2015 by 31.5% and 36.8%, respectively. This study successfully applies satellite remote sensing data to quantitatively analyze the spatial-temporal distributions of tropospheric NO2 and ground-level PM2.5. This approach can support air pollution monitoring in the CPER, and the estimated concentrations can provide references for environmental policymaking.


INTRODUCTION
In recent decades, China has experienced both rapid economic development and dramatic growth of air pollution problems.Living in one of the most polluted countries, nearly all of China's population of 1.3 billion resides in areas that do not meet the World Health Organization (WHO)'s Air Quality Guidelines (AQG) of 10 µg m -3 (Apte et al., 2015;Ma et al., 2015;Van Donkelaar et al., 2015;West et al., 2016).Nitrogen dioxide (NO 2 ) and particulate matter with an aerodynamic diameter less than 2.5 µm (PM 2.5 ) are the two major air pollutants in China (Streets et al., 2013).
The environmental effects of ambient NO 2 have received considerable attention.To reduce the environmental damage associated with acidification, eutrophication and ozone depletion, the United Nations European Economic Commission (UNECE) issued the Sofia Agreement in 1988 and the Goteborg Protocol in 1999 to reduce the emissions of NO 2 by human activities.In the United States, both nitrogen oxide (NO x ) emissions and ambient NO x concentrations were reduced by approximately 30% from 1990 to 2006 (United States Environmental Protection Agency (USEPA), 2010).Compared to developed countries, China was late in initiating control of NO 2 emissions.In 2010, only 18% of China's power plants were using denitrification technology to reduce NO 2 emissions (Wang and Hao, 2012).Previous studies have analyzed the longterm NO x trends in China.For example, Zhao et al. (2013) showed that the NO x emissions in China increased rapidly from 11.0 Mt in 1995 to 26.1 Mt in 2010. De Foy et al. (2016) suggested that there was an increase in Ozone Monitoring Instrument (OMI)-retrieved NO 2 columns in most areas from 2005 to approximately 2011, which was followed by a strong decrease continuing through 2015.Gu et al. (2013) found large regional, seasonal, and urban/rural variations in NO x emission trends based on an analysis of OMI observations of NO 2 columns over China from 2005 to 2010.Meanwhile, studies have shown that the concentrations of NO 2 in China have rapidly increased because of the development of large industries and a greater number of motor vehicles in the recent decade (van der A et al., 2006;Zhang et al., 2007).
PM 2.5 particles can directly enter human alveoli and can cause adverse health effects (Duki et al., 2003;Pope and Dockery, 2006).In addition, China's high PM 2.5 concentrations have received global attention (Wang et al., 2014b).Van Donkelaar et al. (2006) showed that the North China Plain was the area with the highest concentrations of PM 2.5 in the world.According to the "China Environmental Status Bulletin 2014," the average concentration of PM 2.5 in China was 62 µg m -3 , and few Chinese cities met the WHO's air quality standard (10 µg m -3 as an annual mean) in the statistical year of 2012.
In 2012, NO 2 and PM 2.5 were added to the Chinese Ambient Air Quality Standard as criteria air pollutants.The Ministry of Environmental Protection (MEP) of China started to publish the monitored mass concentrations of NO 2 and PM 2.5 at each China Environmental Monitoring Center (CEMC) located in major Chinese cities in January 2013 (Wang et al., 2014a;Zhang and Cao, 2015;Zhao et al., 2016).However, the monitored data might not be representative because of the limited number and spatial distribution of monitors.For example, most monitors were located in urban areas; hence, those monitors provided inadequate information about the air pollution in rural areas, where the air quality was still likely to be affected by industrial complexes.In addition, the earliest date of air pollution data we could access from the China's monitoring system was in 2013.Thus, historical air pollution data are not available for continental China.
The understanding of the spatial distribution of air pollution in China has been recently improved by the application of advanced assessment tools, such as satellite remote sensing.Satellite-retrieved products have many advantages, including global coverage, high spatial-temporal resolution, and historical datasets (Zhang et al., 2004).The application of satellite remote sensing in retrieving tropospheric NO 2 columns and predicting ground-level PM 2.5 has become a popular research area.Zhou et al. (2015Zhou et al. ( , 2016a, b) , b) analyzed the temporal and spatial variations in NO 2 concentrations and the relevant influential factors in the Beijing-Tianjin-Hebei district, Yangtze River Delta, and Shandong regions using the OMI product.Zhang et al. (2017b) used OMI data to analyze the temporal and spatial changes in the column concentrations of NO 2 and SO 2 in Henan Province over the past decade.Yao and Lu (2014) used the Moderate Resolution Imaging Spectroradiometer (MODIS) aerosol optical depth (AOD) to estimate the distribution of PM 2.5 concentrations in China from 2006 through 2010.Sun et al. (2017) retrieved 1-km MODIS AOD in the Beijing-Tianjin-Hebei region to analyze the spatial and temporal distribution of aerosols.Ma et al. (2016) combined the MODIS AOD and the two-step generalized additive model approach to estimate the ground-level PM 2.5 concentration in China for the study year 2004-2013.Therefore, the application of satellite remote sensing has extended the assessment of air pollution to areas without monitors.
The Central Plains Economic Region (CPER) is located in central eastern China.With a population of 160 million, the CPER is one of the most populous areas in the world.The intense consumption of fossil fuels driven by the large population and rapid economic growth has led to a serious air pollution problem in the CPER.The economic growth of the CPER ranks fourth behind the Yangtze River Delta, the Pearl River Delta, and the Beijing-Tianjin-Hebei region in recent years.In addition, the CPER is an area that has one of the highest coal consumption rates in China (Li et al., 2012).As noted by the Civil Aviation Administration of China (CAAC)'s air quality report in 2015, the majority of the CPER was classified as a heavily polluted area in terms of PM 2.5 (CAAC, 2015).The air pollution of the CPER was ranked as the second worst in China in terms of the concentration of PM 2.5 pollution, behind the Beijing-Tianjin-Hebei district (Natural Resources Defense Council (NRDC), 2015).The cities of Xingtai, Zhengzhou, Anyang, and Handan within the CPER have frequently been ranked among the top 10 most polluted cities in China.In 2017, China's MEP implemented an emission control program to reduce the air pollution in Beijing, Tianjin and 26 other heavily polluted cities in China, and the program was referred to as the "2+26" program (Ministry of Environmental Protection, 2017).Nine cities in the CPER were included in the "2+26" program, and monitoring and emission reductions were strengthened in order to reduce the concentrations of NO 2 and PM 2.5 in the air.
Determining the air pollution in the CPER is of great importance because the air quality of the CPER is likely affected by many factors.First, the CPER's air pollution represents a mixture of natural sources and anthropogenic sources.Second, the CPER is near two major economic zones in China, the Beijing-Tianjin-Hebei region and the Yangtze River Delta.The transport of air pollutants between economic zones may have a large impact on the air pollution level in the CPER.Nevertheless, the current air pollution monitoring network has a poor spatial resolution, and no historical data prior to 2013 are available.Therefore, we used satellite data to retrieve tropospheric NO 2 columns and ground-level PM 2.5 concentrations at a high spatial resolution in the CPER domain.By incorporating factors that influence air pollution and statistical models, we are able to estimate the source impact of NO 2 and PM 2.5 from 2005 through 2015.The approaches we used in this study can provide exposure assessment results for air quality management and policymaking in the CPER.

Study Area
The CPER is an economic development zone covering the entire administrative area of Henan Province and surrounding areas.As shown in Fig. 1, the CPER, located in central China (110°E-118°E, 32°N-38°N), has a total area of approximately 2.89 × 10 5 km 2 and contains 28 cities in five provinces (i.e., Henan, Shandong, Anhui, Hebei, and Shanxi).In addition, the Central Plains Urban Agglomeration (CPUA) consists of 12 cities within the CPER (i.e., Zhengzhou, Luoyang, Kaifeng, Xinxiang, Jiaozuo, Xuchang, Pingdingshan, Luohe, Jiyuan, Hebi, Shangqiu, and Zhoukou).In particular, Zhengzhou, as a key connection center for China's high-speed rail network, is an important ground transportation hub and has experienced greater economic growth than other cities in the CPER.

OMI-retrieved NO 2
The OMI is a sensor installed on the Aura satellite, which was launched by the U.S. National Aeronautics and Space Administration (NASA) in July 2004.The OMI was co-produced by the Netherlands, Finland and NASA (Celarier et al., 2008).It was designed to detect trace gases in the atmosphere and to study the roles that trace gases play in climate change.The orbit of the OMI is sun-synchronous and has a swath as wide as approximately 2600 km.The products of the OMI have a spatial resolution of 13 km × 24 km and daily retrieval (Boersma et al., 2007;Wenig et al., 2008).Thus, these products can be used to measure the daily concentrations of trace gases at high spatial resolution.In this study, the satellite product used to measure the ground-level NO 2 concentrations was the monthly average tropospheric NO 2 column dataset (DOMINO version 2.0) obtained from the Tropospheric Emission Monitoring Internet Service (TEMIS) (http://www.temis.nl/airpollution/no2.html).This product is a level-3 OMI product and was globally gridded at a spatial resolution of 0.125°.The OMI NO 2 data provided by this product were the tropospheric vertical column density data, which is labeled VCD_NO 2 (or simply NO 2 ) and represents the total number of NO 2 molecules, and these data were further processed to seasonal and annual averages for investigation.Similar to the popular NASA NO 2 product, the DOMINO product is also based on the differential optical absorption spectroscopy (DOAS) algorithm.The differences between the DOMINO and NASA products include different methods for adjusting the stratospheric contribution and the use of different terrain and profile datasets for conversion of the tropospheric slant column into a vertical column (Russell et al., 2012).DOMINO subtracts stratospheric contributions as determined by a data assimilation system, while the NASA product estimates stratospheric NO 2 from OMI data without using stratospheric chemical transport models directly.Additionally, DOMINO calculates the air mass factor (AMF) with a priori NO 2 monthly mean vertical profile shapes from the Global Modeling Initiative (GMI) model (Bucsela et al., 2013).Despite the differences, both algorithms produce statistically similar regional trends (Krotkov et al., 2016).Ialongo et al. (2016) found that the retrieval uncertainties render these two products indistinguishable.
Several studies have validated the OMI-retrieved NO 2 by using ground-based measurements in China.For example, Lin et al. (2014)  sites in Beijing.Their findings revealed consistent positive correlations across all sites, and the overall R 2 was 0.92.Notably, lightning in the upper troposphere can produce 10-20% of the NO x generated on a global scale (Miyazaki et al., 2014).However, in China, previous studies (Guo et al., 2016;2017) found that the annual mean contribution of lightning to the total tropospheric NO x is only 7.5%.Furthermore, the lightning contribution is even lower in eastern China than in western China.Therefore, the effect of lightning on NO x emissions over the CPER is assumed to be negligible in this study.Considering that the DOMINO product has been widely used in many studies on NO 2 emissions trends (Irie et al., 2016;Liu et al., 2016), we believe that choosing the DOMINO NO 2 product will not cause a substantial change in our results.

MODIS-estimated PM 2.5
We followed the two-stage approach developed by Ma et al. (2016) to estimate the daily-level PM 2.5 concentrations using satellite-retrieved AOD in the CPER for 2005-2015.The first stage of the statistical model adopted a linear mixed-effect model, which was adjusted with the monitored PM 2.5 concentrations and the same grid cell's AOD, as well as with meteorological factors collected from weather stations near the grid cell.The second stage adopted a generalized additive model, which used a smoothing function to optimize the land-use parameters and geographical coordinates to improve the model's performance in predicting PM 2.5 .
In this model, the AOD data were generated using the dark target algorithm from MODIS.The process of the MODIS Level 2.0 AOD product has been described at length by its development group (Levy et al., 2013).The level-2 products with 10-km-resolution AOD data were downloaded from the Atmosphere Archive and Distribution System database (LAADS web: https://ladsweb.modaps.eosdis.nasa.gov/)for the study period 2005-2015.The ground-based PM 2.5 data used for model fitting were obtained from the website of the CEMC for all available PM 2.5 sites located inside the CPER.Overall, data were collected from more than 150 sites.We matched the AOD values to the ground-level measurements of PM 2.5 concentrations in the same 10-km grid cell and on the same day that the PM 2.5 data were collected.This model also used meteorological parameters, such as planetary boundary layer (PBL) height, relative humidity (RH), wind speed, and surface pressure, from the Goddard Earth Observing System Data Assimilation System (GEOS-5), as well as 300-m-resolution land use data from the European Space Agency Land Cover data portal (http://due.esrin.esa.int/pa ge_globcover.php).Ma et al. (2016) has validated the two-stage model PM 2.5 data using the corresponding ground-based observations throughout China.For example, the correlation between satellite and in situ measurements in 2014 in China was 0.85; seasonally, this correlation was 0.89.Our approach here also involved a focused validation study to illustrate the applicability of the estimated PM 2.5 data from the twostage statistical model over the study area.Fig. 2 is a scatterplot of the monthly averaged PM 2.5 concentrations from ground-based observations and MODIS-estimated PM 2.5 concentrations during 2013-2015.Overall, the satellite estimates have a strong correlation with the ground-based data (y = 0.71x + 27.96, r = 0.77).Therefore, the two-stage statistical model can reliably estimate historical PM 2.5 data and effectively reflect the concentration levels of the nearground PM 2.5 in the research area.

Spatial distributions of NO 2 and PM 2.5 in the CPER
The spatial distributions of NO 2 and PM 2.5 concentrations in the CPER are shown in Fig. 3.As shown in Fig. 3(b), the high NO 2 concentrations were mainly found in relatively developed and populous areas, such as Zhengzhou, northern Luoyang, Jiaozuo, Jiyuan, western Xinxiang, northern Anyang, and eastern Hebei Province.In addition, high concentrations of NO 2 were also found in areas with high industrial emissions.For example, Jiaozuo has the highest average concentration of NO 2 , and it is also the location of the largest coal mine in central China.The NO 2 concentrations in the mountainous areas of western and southern Henan were estimated to be lower than those in the cities, which may be due to the high vegetation coverage and the lack of anthropogenic emissions.
As shown in Fig. 3(b), the distribution of PM 2.5 concentrations exhibited great spatial variation.The annual average concentrations of PM 2.5 in our study were consistent with the measured PM 2.5 concentrations in the cities with air monitors.The cities ranked by air quality from worst to best are Zhengzhou, Kaifeng, Xuchang, Luohe, southwestern Luoyang, Jiaozuo, Xinxiang, Hebi, Anyang, and Puyang.Across the CPER domain, high PM 2.5 concentrations mainly occurred along the southern side of the Taihang Mountains and in areas with perennially dominant northeast and southwest winds.These topographical and meteorological conditions prevent air pollutants from dispersing, resulting in an accumulation of air pollutants on the southern side of the Taihang Mountains (Tian et al., 2011).The areas with relatively low concentrations of PM 2.5 included the city of Sanmenxia, which is located in western Henan Province, and southwestern Luoyang and Xinyang, which are located in southern Henan Province and Shanxi Province, respectively.
The twelve core cities in the CPUA serve as the industrial and economic center of the CPER and were also the most polluted areas in the CPER.Therefore, a statistical analysis of the concentrations of NO 2 and PM 2.5 in these twelve cities in the CPUA was performed (Tables 1 and 2).In the results, Luoyang had the lowest average concentrations of both NO 2 (7.74 × 10 15 molecules cm -2 ) and PM 2.5 (75.5 µg m -3 ) during the 11 study years.Jiaozuo had the highest NO 2 concentration (19.54 × 10 15 molecules cm -2 ), and Hebi had the highest PM 2.5 concentration (107.8 µg m -3 ) among the 12 CPUA cities.
The two major causes of the air pollution in the CPUA cities were the high density of industrial facilities and difficulty dispersing the air pollution due to their geographic locations.For example, the economy of the city with the highest NO 2 concentrations for the 11 study years, Jiaozuo, relies heavily on coal mining and metallurgical industry.According to Henan Province's Governmental Statistics, although Jiaozuo contained only 3.7% of Henan Province's population, it contributed 5.2% of the province's total gross domestic product (GDP) and 7.1% of the province's industrial production in the year 2015.During the same year, although the GDP contributed by Hebi was not as high as that of Jiaozuo, 62% of the city's GDP was contributed by heavy industry, such as automotive manufacturing.Hebi built two large power plants (a 2,200-MW thermal power plant and a 4,100-MW coal-fired power plant) to provide electricity for the heavy industry facilities, and both of the power plants consumed large amounts of fossil fuels, consequently releasing large amounts of air pollution into the city.In addition to containing areas with the highest air pollution levels in China, Hebei Province also acted as a source of air pollutants to nearby cities, lowering their air quality (Wang et al., 2007).

Monthly and Seasonal Patterns of NO 2 and PM 2.5 in the CPER
The spatial distributions of the monthly average concentrations of NO 2 and PM 2.5 in the CPER over the 11 study years are shown in Figs. 4 and 5, respectively.As shown in Fig. 4, January and December had higher NO 2 concentrations than the other months.Throughout almost all of the CPER, the NO 2 concentrations were highest (NO 2 > 15 × 10 15 molecules cm -2 ) in January and December, and  The variation in the monthly average concentrations of PM 2.5 was similar to that of NO 2 across the different months (Fig. 5).The PM 2.5 concentrations were higher in the months of December, January and February than in the other months.Fig. 5 showed that almost the entire domain of the CPER was in the highest category for PM 2.5 (PM 2.5 > 100 µg m -3 ).The PM 2.5 concentrations decreased in March, as shown by the decrease in the area of high PM 2.5 concentrations in Fig. 5.Only Zhengzhou and northern Kaifeng were above the lowest level of PM 2.5 .The PM 2.5 concentrations continued to decrease in April and May, and the decreases were greater in Sanmenxia and in the western, mountainous area of Luoyang.The variations in PM 2.5 concentrations were not consistent across different CPER areas and were not straightforward in different summer months (June, July, and August).In June and July, the predicted concentrations of PM 2.5 increased greatly in the northern and middle areas of the CPER.In addition to the widespread occurrence of crop residue burning, the increase in aerosols might be due to the relatively strong solar radiation in summer, which can promote more rapid photochemical reactions by accelerating the transformation of NO 2 to NO 3 -and SO 2 to SO 4 2-.On one hand, these transformations reduce the lifetime of gas-phase air pollutants in the atmosphere, but on the other hand, they also cause high concentrations of nitrate and sulfate particles (Wang et al., 2011;Zhang et al., 2012;Cheng et al., 2013).In August, the predicted concentrations of PM 2.5 decreased in most areas in Henan Province, and the lowest values were found in the western and southern regions of Henan Province and in northern Shanxi Province (red polygon in Fig. 5).From September to November, the areas with high predicted concentrations of PM 2.5 expanded from individual cities to the majority of the eastern CPER.The increase in PM 2.5 concentrations in the late fall might be associated with straw burning in Henan Province.The widespread grain farmlands in Henan Province leave thousands of tons of straw in the field after the harvest.Because no appropriate alternative crop waste management approach is available, this farmland straw is incinerated.The burning straw is a major source of PM 2.5 and other air pollutants (e.g., SO 2 , NO x , and VOC) in the late fall and early winter period in central China.A previous study has shown that the concentrations of inhalable particles near farmlands were three times higher than usual during the straw burning period (Li et al., 2008).
The seasonal variations in the NO 2 concentrations and the PM 2.5 concentrations are shown in Fig. 6.The seasonal patterns of NO 2 and PM 2.5 found in our study were consistent with the seasonal patterns of NO 2 and PM 2.5 reported in other studies.The concentrations of NO 2 and PM 2.5 were generally higher in the fall and winter seasons.In addition to the high precursor emissions due to heating and coal burning, the wintertime adverse weather conditions also contribute significantly to the high surface PM 2.5 concentrations in the CPER.For example, the PBL height over the North China Plain is usually less than 500 m in winter (Li et al., 2016), resulting in less efficient vertical transport and mixing of particles to higher altitudes.In addition, the limited precipitation in the cold season hinders the wet deposition of PM 2.5 , so high concentrations of air pollutants can remain in the atmosphere for a longer time.Compared to the PM 2.5 concentrations, the NO 2 concentrations exhibited a greater decrease from winter to summer.During the summer season, the atmospheric temperatures and humidity were relatively high, which accelerated the oxidization process of NO 2 in the air (Khoder, 2002).In addition, the CPER features a warm temperate subtropical climate.The frequent rainfall in the summer season greatly reduced the NO 2 concentrations in the atmosphere (Yan et al., 2015).
The seasonal variations in NO 2 concentrations and PM 2.5 concentrations varied among the different areas in the CPER (Fig. 6).Zhengzhou, Jiaozuo, Xinxiang, Hebi, and Anyang are all located in northern Henan Province, and Handan and Xingtai are located in southern Hebei Province.During the spring (Fig. 6(a)), the NO 2 concentrations in these regions were high, and the PM 2.5 concentrations were generally low.During the summer (Fig. 6b), most regions had low NO 2 concentrations, but Zhengzhou, Jiaozuo, Xinxiang, Handan, and Xingtai, which formed the center of the developed area, had higher concentrations.The PM 2.5 concentrations showed two prominent low-value areas in the western and southern mountain regions.During the autumn (Fig. 6(c)), the high-concentration area was widespread, with high NO 2 concentrations in the northern region and high PM 2.5 concentrations extending over a large area of the central eastern section of the study region.During the winter (Fig. 6(d)), most regions still showed high NO 2 concentrations.Additionally, the concentrations decreased from north to south and from east to the west, and the concentrations of PM 2.5 across the entire province were high, except over Pingdingshan, the mountainous area of Luoyang in Henan Province, and northern Shanxi Province.NO 2 and PM 2.5 concentrations in the CPER.Overall, the variations in NO 2 were greater than those in PM 2.5 , but the trends were similar between the two pollutants.As shown in Fig. 7, the concentrations of NO 2 and PM 2.5 increased from 2005 to 2011 and then decreased after 2011.Our results before 2011 are consistent with the findings of Itahashi et al. (2014), who concluded that there was an increase in OMI NO 2 columns during 2000-2010 in central eastern China.The rapid growth in NO 2 column density before 2011 was driven by the increase in anthropogenic NO x emissions in China (Zhao et al., 2013).For example, the total capacities of coal-fired power generation increased by 48.8% in 200548.8% in -200748.8% in (Wang et al., 2012)).The turning point of NO 2 and PM 2.5 concentrations in the year 2011 may be related to the Chinese government's "12 th Five-Year Plan", which included policies related to emission reductions to control serious air pollution.As a result, the national NO 2 level has decreased by 6% year -1 since 2011 (Irie et al., 2016).In Henan Province, industrial NO 2 emissions amounted to 719,500 tons in 2015, which was 28% lower than that in 2010.Correspondingly, the NO 2 concentration in Henan Province decreased by 27.4% during those five years.Although there was no reduction plan directly for PM 2.5 emissions, our results indicated that the PM 2.5 concentration decreased by 28.2% from 2010 to 2015.The decrease in air pollution accelerated after 2014, when straw burning was forbidden in certain areas in Henan Province.For example, the number of straw-burning events calculated by the Office of the Ministry of Environmental Protection was 37.18% lower in 2015 than in 2014 (Office of the Ministry of Environmental Protection, 2015).

Eleven-year Variations in NO 2 and PM 2.5 in the CPER
In this study, we further processed the Multi-resolution Emission Inventory for China (MEIC, http://meicmodel.org)data with a spatial resolution of 0.25° × 0.25° to estimate anthropogenic emissions in the CPER.Fig. 8 shows the spatial distribution of the MEIC anthropogenic NO x and PM 2.5 inventory in 2012.Figs. 8(a) and 8(b) show that the areas with high emissions are mainly concentrated in megacities, such as Zhengzhou, Jiaozuo, and Xinxiang, as well as in areas with high densities of industry and power plants.The PM 2.5 from residential sources, as depicted in Fig. 8(c) (2), had a very similar distribution as the transportation NO x sources (Fig. 8(d), 1), and both were mostly concentrated in urban areas with dense populations.As shown in Table 3, we calculated the total anthropogenic emissions of NO x and PM 2.5 in Henan in three years (2008, 2010, and 2012).For NO x , industry, power plants, and transportation were the major sources of NO x emissions, with each accounting for approximately one-third of the total NO x emissions.Both industrial and power plant emissions increased from 2008 to 2010 and decreased from 2010 to 2012, suggesting a close connection to the NO x emission-reduction measures established in the 12 th Five-Year Plan (Foy et al., 2016;Liu et al., 2016).In contrast, the NO x emissions from transportation increased by 4 × 10 4 t along with the increase in the number of vehicles during the four years.Unlike NO x , anthropogenic PM 2.5 emissions are mainly contributed by industrial and residential sources, which together accounted for more than 85% of the emissions in 2012.Except for the increase in residential emissions from 2010 to 2012, the industrial, power and transportation emissions all showed significant declines from 2008 to 2012.
The monthly average concentrations of NO 2 and PM 2.5 over the CPER are shown in Fig. 9.The monthly average concentrations of NO 2 and PM 2.5 showed cyclical variation throughout the year.Low concentrations of NO 2 and PM 2.5 generally occurred during the summer, i.e., July and August.The predicted monthly average NO 2 concentration reached a minimum value of 4.39 × 10 15 molecules cm -2 in July 2006, and the predicted monthly average PM 2.5 concentration reached a minimum level of 40.2 µg m -3 in August 2005.The peak predicted concentrations generally appeared during winter, i.e., December and January.In January 2013, both NO 2 and PM 2.5 concentrations reached their highest values of 31.68 × 10 15 molecules cm -2 and 181.6 µg m -3 , respectively.The air quality in the CPER was also affected by the short-term emission reduction program implemented by the Chinese government during international events.For example, in Henan Province, the  The correlation analysis of NO 2 and PM 2.5 indicated that the concentrations of the two air pollutants were highly correlated (0.84) (Fig. 10).This significantly positive relationship might provide a reference for future emission control plans.Based on the current results, the control of NO 2 emissions could lead to a reduction in ambient PM 2.5 concentrations as well.

CONCLUSION
This study estimated the concentrations of NO 2 and ground-level PM 2.5 in the CPER using satellite remote sensing over 11 study years.The spatial distributions of NO 2 and PM 2.5 concentrations were examined, including the monthly and seasonal concentration variations in the CPER and in 12 core cities.The areas with high NO 2 concentrations were mainly located in eastern Hebei Province and northern Henan Province.Among the investigated cities, Jiaozuo had the highest 11-year average concentration of NO 2 (19.54 × 10 15 molecules cm -2 ).The spatial distribution of PM 2.5 is similar to that of NO 2 .The city of Hebi and the surrounding area had the highest concentrations of PM 2.5 (107.8 µg m -3 ).In the CPER, Sanmenxia, southwestern Luoyang, and Xinyang had lower levels of air pollution.These cities are all located in the mountainous region of Henan Province.
The predicted concentrations of NO 2 and PM 2.5 showed cyclical variations by month and season.Within a calendar year, the winter was usually associated with higher concentrations of NO 2 and PM 2.5 , while summer was usually associated with lower concentrations of NO 2 and PM 2.5 .Overall, the pollution levels of the four seasons were ranked as follows (in descending order): winter, autumn, spring, and summer.Fossil fuel consumption and weather were the two most important factors in the seasonal variations in air pollution in the CPER.The high demand for heating in winter increased the consumption of coal, and the static atmospheric conditions in winter prevented pollutants from dispersing.Therefore, winter had the highest air pollution levels among the four seasons.
Both the NO 2 and PM 2.5 concentrations fluctuated during the 11 years we studied, but the variation in NO 2 was greater than in PM 2.5 in the CPER.The average annual concentration of NO 2 was lowest in 2006 and then increased greatly in 2007.The average annual concentrations of NO 2 increased gradually from 2007 till 2010 and peaked in 2011.However, due to emission control efforts in recent years, the average annual concentration of NO 2 decreased in 2014 and 2015, approaching the low concentration observed in 2006.The variation in the PM 2.5 concentrations was consistent with that in the NO 2 concentrations.The average annual concentration of PM 2.5 peaked in 2012 and continued to decline over subsequent years, reaching its lowest level in 2015.
Fig. 1.The location of the study area (CPER) in China.Henan Province is outlined in red.The twelve cities of the Central Plains Urban Agglomeration (CPUA) are located in northern Henan Province.

Fig. 7 .
Fig. 7.The annual variations in NO 2 columnar concentration and PM 2.5 mass concentration in the CPER during 2005-2015.

Fig. 10 .
Fig. 10.Scatterplot of the monthly NO 2 columnar concentration and PM 2.5 mass concentration in the CPER during 2005-2015.

Table 1 .
The mean (plus standard deviation) NO

Table 2 .
The mean (plus standard deviations) PM

Table 3 .
Four sources of anthropogenic NO x and PM 2.5 emissions (i.e., industrial, power, residential and transportation) in Henan in 2008, 2010 and 2012.(Unit: 10 4 t yr -1 .) of NO 2 and PM 2.5 reached their lowest levels for the December period during the 11-year study period in December 2015 following the implementation of a short-term emission reduction program related to the Shanghai Cooperation Organization Summit, which occurred from December 14 to 16, 2015.Another example was the 2008 Beijing Summer Olympic Games.The monthly average concentrations of NO 2 in 2008 were significantly lower than those in the same period in preceding years.
Fig. 9.The monthly variations in NO 2 columnar concentration and PM 2.5 mass concentration in the CPER during 2005-2015.concentrations