Spatial–Temporal Variations in NO2 and PM2.5 over the ... - MDPI

sensors Article

Spatial–Temporal Variations in NO2 and PM2.5 over the Chengdu–Chongqing Economic Zone in China during 2005–2015 Based on Satellite Remote Sensing Kun Cai 1,2 , Qiushuang Zhang 2 , Shenshen Li 3, *, Yujing Li 2 and Wei Ge 3 1 2 3

*

College of Environment and Planning, Henan University, Kaifeng 475004, China; [email protected] College of Computer and Information Engineering, Henan University, Kaifeng 475004, China; [email protected] (Q.Z.); [email protected] (Y.L.) State Key Laboratory of Remote Sensing Science, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100101, China; [email protected] Correspondence: [email protected]

Received: 4 September 2018; Accepted: 12 November 2018; Published: 15 November 2018







Abstract: The Chengdu–Chongqing Economic Zone (CCEZ), which is located in southwestern China, is the fourth largest economic zone in China. The rapid economic development of this area has resulted in many environmental problems, including extremely high concentrations of nitrogen dioxide (NO2 ) and fine particulate matter (PM2.5 ). However, current ground observations lack spatial and temporal coverage. In this study, satellite remote sensing techniques were used to analyze the variation in NO2 and PM2.5 from 2005 to 2015 in the CCEZ. The Ozone Monitoring Instrument (OMI) and the Moderate Resolution Imaging Spectroradiometer (MODIS) aerosol optical depth (AOD) product were used to retrieve tropospheric NO2 vertical columns and estimate ground-level PM2.5 concentrations, respectively. Geographically, high NO2 concentrations were mainly located in the northwest of Chengdu and southeast of Chongqing. However, high PM2.5 concentrations were mainly located in the center areas of the basin. The seasonal average NO2 and PM2.5 concentrations were both highest in winter and lowest in summer. The seasonal average NO2 and PM2.5 were as high as 749.33 × 1013 molecules·cm−2 and 132.39 µg·m−3 in winter 2010, respectively. Over 11 years, the annual average NO2 and PM2.5 values in the CCEZ increased initially and then decreased, with 2011 as the inflection point. In 2007, the concentration of NO2 reached its lowest value since 2005, which was 230.15 × 1013 molecules·cm−2 , and in 2015, the concentration of PM2.5 reached its lowest value since 2005, which was 26.43 µg·m−3 . Our study demonstrates the potential use of satellite remote sensing to compensate for the lack of ground-observed data when quantitatively analyzing the spatial–temporal variations in regional air quality. Keywords: NO2 ; PM2.5 ; Chengdu–Chongqing Economic Zone; OMI; MODIS

1. Introduction The Chengdu–Chongqing Economic Zone (CCEZ) is the fourth largest economic region in China after the Pearl River Delta, the Yangtze River Delta, and the Beijing–Tianjin–Hebei region. The rapid economic development in the CCEZ has resulted in many environmental problems [1], including air pollution. Two important constituents of air pollution are nitrogen oxides (NOx = NO + NO2 ) and particulate matter with an aerodynamic diameter of less than 2.5 µm (PM2.5 ), both of which can inflict substantial harm to human health [2,3]. Studies have shown that high levels of NO2 in the air increase the risk of respiratory infections and lung damage [4]. In addition, NO2 constitutes one of the main factors that leads to the formation of acid rain photochemical smog and causes

Sensors 2018, 18, 3950; doi:10.3390/s18113950

www.mdpi.com/journal/sensors

Sensors 2018, 18, 3950

2 of 16

climate change; accordingly, NO2 considerably damages the atmospheric ecological environment [5]. PM2.5 particles are so small that they can reach the alveoli directly and penetrate into the bloodstream. Hence, long-term exposure to PM2.5 can lead to bronchitis and cardiovascular diseases [6]. In addition to affecting human health, PM2.5 is the primary cause of urban haze conditions [7]. Two main approaches—ground-based observation and satellite remote sensing observation—are employed to acquire long-term continuous observations of atmospheric NO2 and PM2.5 concentrations. However, current ground observation networks suffer from deficiencies in spatial and temporal coverage. By contrast, satellite remote sensing can be used to obtain data for most regions and therefore overcomes the insufficient spatial coverage associated with ground-based observations [8]. In recent years, many researchers have conducted numerous studies on tropospheric NO2 concentrations and PM2.5 mass concentrations using satellite data. For example, Liu et al. analyzed the spatial and temporal distributions of NO2 in urban agglomerations and their adjacent areas in the Pearl River Delta by using Ozone Monitoring Instrument (OMI) data; the results showed that the areas with high NO2 concentrations in the Pearl River Delta region were connected, that the urban group effect was significant, and that much higher concentrations were concentrated in the Pearl River Delta than in the surrounding areas [9]. Long et al. analyzed the effects of changes in tropospheric NO2 concentrations on human activities from 2005 to 2010 by using OMI data and found that the winter NO2 concentration is high and the summer NO2 concentration is low, while the western regions exhibit the opposite situation [10]. Krotkov et al. analyzed the changes in NO2 and SO2 concentrations in the world’s most polluted industrialized regions over the past few years (2005–2015) by using OMI data; their analysis revealed that NO2 and SO2 concentrations have declined in the past 10 years over the eastern U.S. and eastern Europe under the effects of pollution prevention and control policies, while those in China rose and then declined [11]. Guo et al. estimated the PM2.5 mass concentrations in eastern China using Moderate Resolution Imaging Spectroradiometer (MODIS) aerosol optical depth (AOD) product data and compared the values with monitoring data obtained from various sites (i.e., Benxi, Zhengzhou, Lushan Mountain, Guilin, and Nanning); the results showed that the locations with the most and least pollution are Zhengzhou and Nanning, respectively, which are basically consistent with ground station observations, among which the strongest correlation was found in Lushan [12]. The CCEZ was developed from the preexisting industrial base of Chengdu–Chongqing [13]. In 2007, the Chongqing Municipal Government and the Sichuan Provincial Government signed an agreement to promote Sichuan–Chongqing Cooperation and build the CCEZ together. The CCEZ was officially approved by the state council in March 2011, and the following two years witnessed substantial industrial developments. As a region with the highest urbanization level in western China, in recent years, the CCEZ has been playing an important strategic role in China’s economic and social development due to its rapid advancement of industrialization and urbanization. According to the development strategy of the CCEZ, energy is one of its leading industries. During the period between the 11th Five-Year Plan and the 12th Five-Year Plan, coal consumption was the main driving force for the development of the national economy, indicating that great challenges need to be overcome to reduce the associated nitrogen oxide and sulfide emissions [14]. Although the majority of coal-fired power plant boilers in China are currently equipped with low nitrogen combustion technology, the denitration efficiency is low [15], and the desulfurization technology is subject to resource shortages and secondary pollution in addition to the regional environment and other factors [16]. Industrial pollution and transportation pollution are the major sources of nitrogen oxides and sulfides in the CCEZ [17–19]. However, to the best of our knowledge, air pollution in the CCEZ region is not as widely studied compared with that in the Beijing–Tianjin–Hebei region, Yangtze River delta, and Pearl River delta in China [20–22]. In recent years, China has begun to pay attention to environmental monitoring. The characteristics of NO2 and PM2.5 can be measured using ground-based equipment, although the representativeness of individual sites is usually limited in a large domain. More importantly, current PM2.5 and NO2 monitoring stations in China have been established since 2013. Therefore, research on the spatial and temporal distribution and changes over long time series of pollution in the

Sensors 2018, 18, x FOR PEER REVIEW Sensors 2018, 18, 3950

3 of 16 3 of 16

long time series of pollution in the entire CCEZ is still lacking. In this paper, we used NO2 data from OMI and PM estimated MODIS 2005NO to 2015 to statistically analyze the temporal and entire CCEZ is2.5 still lacking. by In this paper,from we used 2 data from OMI and PM2.5 estimated by MODIS spatial distribution characteristics and the changes over long time series of NO2 and PM2.5 in the the from 2005 to 2015 to statistically analyze the temporal and spatial distribution characteristics and CCEZ asover welllong as the main factors influencing these distributions to provide theoretical support for changes time series of NO2 and PM2.5 in the CCEZ as well as the main factors influencing NO 2 and PM2.5 control strategies in this area. these distributions to provide theoretical support for NO and PM control strategies in this area. 2

2.5

2. Materials Materials and and Methods Methods 2. 2.1. 2.1. Study Area Area The The CCEZ, CCEZ, which which is is located located in in the the upper upper reaches reaches of the the Yangtze Yangtze River, River, comprises comprises parts parts of of Sichuan Sichuan and and Chongqing, Chongqing, including 15 prefectures in Sichuan Sichuan and and 31 31 districts districts and and counties counties in in Chongqing. The entire area is located within the Sichuan Basin, which is surrounded by both Chongqing. The entire area is surrounded by both highAs shown shown in in Figure Figure 1, Ya’an, southern Leshan, southern Yibin, high- and and low-elevation low-elevation areas. areas. As southern southern Luzhou, Luzhou, northern northern Mianyang, Mianyang, northern northern Dazhou, Dazhou, and and eastern eastern Chongqing Chongqing are are high-elevation high-elevation areas, whereas the terrain in other areas is relatively low. The area of the CCEZ is approximately areas, whereas the terrain in other areas is relatively low. area of the CCEZ is approximately 206,000 206,000 square square kilometers, kilometers, and and the the population population density density is is high, high, accounting accounting for for nearly nearly 8% 8% of of China’s China’s population, accounts forfor approximately one-third of the region and population,and andits itseconomic economicaggregate aggregate accounts approximately one-third of western the western region ranks first infirst the in west. and Chongqing are the two withcities the largest economic and ranks the Chengdu west. Chengdu and Chongqing are central the twocities central with the largest aggregate westernin China. economic in aggregate western China.

Figure 1.1.The Thegeographic geographiclocation location and terrain of study the study area (Chengdu–Chongqing Economic Figure and terrain of the area (Chengdu–Chongqing Economic Zone) Zone) in China. in China.

2.2. 2.2. OMI-Retrieved OMI-Retrieved NO NO22 The The OMI OMI is is one one of of four four sensors sensors installed installed on on the the Aura Aura Earth Earth Observation Observation System System satellite, satellite, which which was launched by the U.S. National Aeronautics and Space Administration (NASA) on 15 July July 2004. 2004. was launched by the U.S. National Aeronautics and Space Administration (NASA) on 15 The OMI was coproduced by the Netherlands and Finland with NASA, and it is an inheritance The OMI was coproduced by the Netherlands and Finland with NASA, and it is an inheritance instrument Global Ozone Monitoring Experiment (GOME)(GOME) and Scanning Absorption instrumentfor forboth both Global Ozone Monitoring Experiment and Imaging Scanning Imaging Spectrometer for Atmospheric Chartography (SCIAMACHY) measurements. The orbital scanning Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY) measurements. The orbital amplitude of the OMI is 2600 km, the spatial resolution at nadir is 13 km × 24 km, and the whole scanning amplitude of the OMI is 2600 km, the spatial resolution at nadir is 13 km × 24 km, and the world covered a day [23].a The NO2 data used this paper were wholeisworld is once covered once dayOMI-retrieved [23]. The OMI-retrieved NO2indata used in thisderived paper from were the official Royal Dutch product, that is, the OMI tropospheric NO vertical column concentrations 2 derived from the official Royal Dutch product, that is, the OMI tropospheric NO2 vertical column were obtained from theobtained Derivation of OMI data product version 2.0 data [24] 2 (DOMINO) concentrations were from the Tropospheric Derivation ofNO OMI Tropospheric NO2 (DOMINO) from 2005 to 2015. The product was retrieved by the Royal Institute of Meteorology in the Netherlands product version 2.0 [24] from 2005 to 2015. The product was retrieved by the Royal Institute of with a spatial resolution of 0.125◦ . Monthly averageresolution data were downloaded from http://www.temis.nl Meteorology in the Netherlands with a spatial of 0.125°. Monthly average data were in an Environmental Systems Research Institute (ESRI) grid format andResearch then converted into a .csvgrid file downloaded from http://www.temis.nl in an Environmental Systems Institute (ESRI) 13 molecules·cm−2 . format by using the Interactive Data Language (IDL); the units in this format are 10 format and then converted into a .csv file format by using the Interactive Data Language (IDL); the Subsequently, Microsoft SQL Server was used to process the Microsoft quarterly average datawas andused yearly 13 molecules·cm −2. Subsequently, units in this format are 10 SQL Server toaverage process data. Kayana average et al. used multiaxis optical absorption spectroscopy (MAX-DOAS) the quarterly dataa and yearlydifferential average data. Kayana et al. used a multiaxis differential

Sensors 2018, 18, 3950

4 of 16

monitoring network in Russia and Asia to validate OMI-derived tropospheric NO2 vertical column concentrations. The results showed that the monthly average correlation reached an R2 value of 0.84 in urban areas and 0.74 in rural areas, indicating that OMI can better capture NO2 tropospheric levels, especially in NO2 source regions [25]. Mou et al. derived tropospheric NO2 vertical column concentrations for Hefei from ground-based MAX-DOAS observations from August 2013 to July 2014 and compared the results with OMI data; these authors reported good consistency between the data sets, with a correlation coefficient of 0.88 obtained under cloudless weather conditions [26]. Jin et al. reported that the correlation coefficient between star and ground observations in the North China Plain (Gucheng Station, Hebei, China) reached 0.945 in the absence of clouds [27]. These results show that the inversion of OMI data can reflect the spatial and temporal distributions of the tropospheric NO2 column. 2.3. MODIS-Estimated PM2.5 We estimated the ground-level PM2.5 concentrations using satellite-retrieved AOD data with a spatial resolution of 10 km in the CCEZ for 2005–2015. The AOD data were generated using the dark target algorithm based on images retrieved from the MODIS instrument onboard the Terra satellite. The processing of the MODIS AOD data product has been described at length by its development group [28]. We followed the two-stage approach developed by Ma et al. [29] to estimate the daily PM2.5 concentrations using satellite-retrieved AOD. The PM2.5 data used for model fitting were obtained from the website of the China Environmental Monitoring Center (CEMC) for all available PM2.5 sites located inside the CCEZ. We matched the AOD values to ground-level measurements of the PM2.5 concentrations (µg·m−3 ) within the same 10 km grid cell acquired on the same day that the PM2.5 data were collected. The first stage of the statistical model adopted a linear mixed-effect model, which was adjusted with the monitored PM2.5 concentrations and the same grid cell’s AOD as well as 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 the PM2.5 concentration. This model also used meteorological parameters, such as the 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 [30]. To further illustrate the applicability of the near-surface PM2.5 mass concentration in the CCEZ estimated by the two-stage statistical model method, we calculated the monthly average between the PM2.5 mass concentrations released by the China Environmental Monitoring Center (CEMC) during 2013–2015 and estimated the PM2.5 mass concentrations. Until 2013, the Ministry of Environmental Protection (MEP) of China greatly expanded its air pollution monitoring network to measure hourly ground-level mass concentrations of PM2.5 . Thus, we statistically analyzed the correlation between these concentrations, and Figure 2 shows that the correlation coefficient reached 0.74. Thus, the PM2.5 concentrations estimated by this method are consistent with near-surface observations. The historical data estimated by this method can be used to study the concentration level of near-surface PM2.5 in the study area.

Sensors 2018, 18, 3950 Sensors 2018, 18, x FOR PEER REVIEW

5 of 16 5 of 16

Figure ·m−−33)) mass 2.52.5(µg Figure2.2.Scatter Scatterplots plotsofofthe themonthly monthlyPM PM (µg·m mass concentrations concentrations over over CEMC CEMCobservations observations and AOD-estimated PM from 2013 to 2015. 2.5 from 2013 to 2015. and AOD-estimated PM2.5

3. Results 3. Results 3.1. Spatial Distributions of NO2 and PM2.5 in the CCEZ 3.1. Spatial Distributions of NO2 and PM2.5 in the CCEZ Figure 3 shows the 11-year average distributions of the tropospheric NO2 vertical column Figure 3 shows the 11-year average distributions of the tropospheric NO2 vertical column concentration and PM2.5 mass concentration in the CCEZ from 2005 to 2015. The mass concentration concentration and PM2.5 mass concentration in the CCEZ from 2005 to 2015. The mass concentration distribution of PM2.5 was similar but not identical to that of NO2 . In addition, as shown in Figure 3b, distribution of PM2.5 was similar but not identical to that of NO2. In addition, as shown in Figure 3b, the middle- and high-value areas of PM2.5 are more widely distributed than the low-value areas. the middle- and high-value areas of PM2.5 are more widely distributed than the low-value areas. The NO2 high-value areas (NO2 > 750 × 1013 molecules·cm−2 ) were mainly distributed in Chengdu, The NO2 high-value areas (NO2 > 750 × 1013 molecules·cm−2) were mainly distributed in Chengdu, Deyang, and the main city of Chongqing, and the median-value (500 × 1013 molecules·cm−2 < NO2 < Deyang, and the main city of Chongqing, and the median-value 750 × 1013 molecules·cm−2 ) areas were mainly distributed in Mianyang, Neijiang, and northern Yibin. (500 × 1013 molecules·cm−2 < NO2 < 750 × 1013 molecules·cm−2) areas were mainly distributed in According to traffic department statistics, at the end of 2014, motor vehicle ownership in Chengdu Mianyang, Neijiang, and northern Yibin. According to traffic department statistics, at the end of reached 3,857,000, representing a 13.9% increase from the end of the previous year; in addition, Sensorsmotor 2018, 18,vehicle x FOR PEER REVIEW in Chengdu reached 3,857,000, representing a 13.9% increase from 6 of 16 2014, ownership the resident population reached 13,653,000, ranking first in the province [31,32]. the end of the previous year; in addition, the resident population reached 13,653,000, ranking first in the province [31,32]. In Chengdu, which is located in the Sichuan Basin, the RH is very high; consequently, PM suspended in the air undergoes hygroscopic growth, thereby accelerating the accumulation of pollutants. Moreover, the weather conditions in Chengdu are fairly calm and stable, under which pollutants are not easily diluted or dispersed, leading to an increase in air pollution; additionally, inversion layers are common in Chengdu and can be relatively thick [33]. These conditions constitute important reasons why the air pollution in the CCEZ does not readily dissipate. The permanent residents in Chongqing’s main urban area account for nearly 28% of the total population in Chongqing, and they occupy an area of less than 6.7% of Chongqing. Furthermore, Chengdu and the main cities of Chongqing have a relatively well-developed traffic industry, and Deyang is a heavy industry base in China. Each of the abovementioned factors places enormous pressure on the (a) (b) environment, thereby resulting in serious air pollution throughout the CCEZ. Figure 3. Spatial distributions of the 11-year average NO2 and PM2.5 concentrations in the CCEZ Figure2005–2015: 3. Spatial (a) distributions of the 11-year average NOthe 2 and 2.5 concentrations the CCEZ during NO2 columnar concentration based on OMIPM DOMINO product atin a resolution ◦ ◦ ◦ ◦. a 2 columnar concentration on the OMI DOMINO ofduring 0.125 2005–2015: × 0.125 and(a)(b)NO MODIS-estimated PM2.5 massbased concentration at a resolution of product 0.1 × 0.1at resolution of 0.125° × 0.125° and (b) MODIS-estimated PM2.5 mass concentration at a resolution of 0.1° × 0.1°.

According to the 2015 Yearbook of Sichuan Province and Chongqing Municipality, the number of practitioners in the secondary industry is linearly related to air pollution. The Sichuan Province Environmental Protection Agency released a bulletin on the ambient air quality in June 2017; the ambient air quality was measured in the first half of the year, and the results regarding the magnitude of the variation in the air quality were ranked. This information showed that within the CCEZ, the cities of Sichuan Province (i.e., Chengdu, Deyang, and Neijiang) are the most polluted [34]. This result is consistent with the multiyear average mean from 2005 to 2011. The CCEZ is centered around Chengdu and Chongqing, which drive the development of the surrounding cities. According to the development of the central cities in the regional report of the CCEZ, Deyang

Sensors 2018, 18, 3950

6 of 16

In Chengdu, which is located in the Sichuan Basin, the RH is very high; consequently, PM suspended in the air undergoes hygroscopic growth, thereby accelerating the accumulation of pollutants. Moreover, the weather conditions in Chengdu are fairly calm and stable, under which pollutants are not easily diluted or dispersed, leading to an increase in air pollution; additionally, inversion layers are common in Chengdu and can be relatively thick [33]. These conditions constitute important reasons why the air pollution in the CCEZ does not readily dissipate. The permanent residents in Chongqing’s main urban area account for nearly 28% of the total population in Chongqing, and they occupy an area of less than 6.7% of Chongqing. Furthermore, Chengdu and the main cities of Chongqing have a relatively well-developed traffic industry, and Deyang is a heavy industry base in China. Each of the abovementioned factors places enormous pressure on the environment, thereby resulting in serious air pollution throughout the CCEZ. According to the 2015 Yearbook of Sichuan Province and Chongqing Municipality, the number of practitioners in the secondary industry is linearly related to air pollution. The Sichuan Province Environmental Protection Agency released a bulletin on the ambient air quality in June 2017; the ambient air quality was measured in the first half of the year, and the results regarding the magnitude of the variation in the air quality were ranked. This information showed that within the CCEZ, the cities of Sichuan Province (i.e., Chengdu, Deyang, and Neijiang) are the most polluted [34]. This result is consistent with the multiyear average mean from 2005 to 2011. The CCEZ is centered around Chengdu and Chongqing, which drive the development of the surrounding cities. According to the development of the central cities in the regional report of the CCEZ, Deyang is an important manufacturing base for heavy equipment, Zigong is a modern industrial city where machinery is manufactured, and Meishan comprises mainly locomotive manufacturing and constitutes an important transport node. The main industries of Neijiang, which is an important transport node and port city, are metallurgical building materials and automotive parts production. All of these cities exhibit serious pollution. Leshan and Ya’an form part of the clean energy industrial base; appropriately, these two cities are the least polluted areas in the CCEZ [35]. As shown in Tables S1 and S2, Ya’an has the lowest 11-year annual average NO2 and PM2.5 concentrations (i.e., 232.66 × 1013 molecules·cm−2 and 53.26 µg·m−3 , respectively) whereas Chengdu displays the highest NO2 concentration (i.e., 849.46 × 1013 molecules·cm−2 ) and Zigong presents the highest PM2.5 concentration (i.e., 91.34 µg·m−3 ). The pollution in these cities is inseparable from their industrial development. 3.2. Monthly and Seasonal Patterns of NO2 and PM2.5 in the CCEZ This paper presents the monthly average spatial distributions of the NO2 column concentration and PM2.5 mass concentration for the 11 years between 2005 and 2015. The spatial and temporal distributions of NO2 are shown in Figure 4. Highly polluted areas were distributed in the center of Chengdu and the main urban area of Chongqing. Along the border of the CCEZ (Ya’an, northern Mianyang, southern Leshan, northern Dazhou, and northeastern Chongqing), the pollution concentration was very low (NO2 < 250 × 1013 molecules·cm−2 ) throughout the year, and slight pollution (250 × 1013 molecules·cm−2 < NO2 < 500 × 1013 molecules·cm−2 ) was observed in December and January in southern Leshan and northeastern Chongqing. The highest levels of pollution were recorded in November, December, and January, and high-value areas were concentrated in the western and southern CCEZ and in most of the areas in and around the main urban area of Chongqing. The extent of highly polluted areas decreased considerably from February to March, and those of both highly and moderately polluted areas decreased on a monthly basis from April to August. However, the extents of areas with high and moderate pollution increased in September and October. The monthly mean distribution of PM2.5 is shown in Figure 5, which indicates that the overall pollution along the border area of the CCEZ was relatively low (PM2.5 < 60 µg·m−3 ) throughout the year. The highest levels of pollution (PM2.5 > 90 µg·m−3 ) were recorded in January, and even the border area pollution level, which was relatively low compared with those in other areas, was approximated high. The high-value areas in February were similar to those in January, and the

August. However, the extents of areas with high and moderate pollution increased in September and October. The monthly mean distribution of PM2.5 is shown in Figure 5, which indicates that the overall pollution along the border area of the CCEZ was relatively low (PM2.5 < 60 µg·m−3) throughout the year. The highest levels of pollution (PM2.5 > 90 µg·m−3) were recorded in January, Sensors 2018, 18, 3950 7 of 16 and even the border area pollution level, which was relatively low compared with those in other areas, was approximated high. The high-value areas in February were similar to those in January, pollution along the Chengdu–Chongqing boundary decreased to moderate levels.levels. The areas with and the pollution along the Chengdu–Chongqing boundary decreased to moderate The areas high levels of pollution decreased in March and December, and the highand median-value areas with high levels of pollution decreased in March and December, and the high- and median-value 3 < PM2.5 3 ) overlapped. −3 < PM2.5 −3) overlapped. (60 µg·(60 m−µg·m < 90 µg m−µg·m Higher levels levels of pollution were recorded in the border areas < ·90 Higher of pollution were recorded in the area in December than in February. April and May, the extents of thethe highand median-value border area in December than inInFebruary. In April and May, extents of the high-areas and decreased sharply, and areas with medium and high levels of pollution around Zigong and Chengdu median-value areas decreased sharply, and areas with medium and high levels of pollution around intersected. June intersected. to August, the entire area slightly from to November, Zigong and From Chengdu From June to was August, the polluted; entire area wasSeptember slightly polluted; from the extents of the highand median-value areas began to gradually expand to include most of the September to November, the extents of the high- and median-value areas began to gradually expand CCEZ, andmost the extents of the high-pollution in high-pollution December were similar to but slightly to include of the CCEZ, and the extentsareas of the areas in December wereheavier similar than in March. to butthose slightly heavier than those in March.

1

2

3

4

5

6

7

8

9

10

11

12

Figure 4. Monthly average NO2 distributions in the CCEZ from 2005 to 2015. Figure 4. Monthly average NO2 distributions in the CCEZ from 2005 to 2015.

Sensors 2018, 18, 3950

8 of 16

Sensors 2018, 18, x FOR PEER REVIEW

8 of 16

1

2

3

4

5

6

7

8

9

10

11

12

Figure 5. Monthly average PM2.5 distributions in in the the CCEZ CCEZ from from 2005 2.5 distributions 2005 to to 2015. 2015.

The in in thethe NONO concentrations and PM mass varied 2 columnar 2.5 PM The seasonal seasonalvariations variations 2 columnar concentrations and 2.5 concentrations mass concentrations among different areas as shown in Figure months were dividedwere into divided seasons according to varied the among the different areas as shown 6.inThe Figure 6. The months into seasons the climate and phenology as follows: spring comprised March, April, and May; summer comprised according to the climate and phenology as follows: spring comprised March, April, and May; June, July, and August; comprised September, October,September, and November; andand winter comprised summer comprised June,fall July, and August; fall comprised October, November; and December, January, and February. The NOand concentrations Ya’an, southern Leshan,in northern 2 column winter comprised December, January, February. The NO2incolumn concentrations Ya’an, Mianyang, northernnorthern Dazhou,Mianyang, and northeastern Chongqing low; this result is closely related southern Leshan, northern Dazhou,were and very northeastern Chongqing were very to the low elevations of these geographic locations. The NO concentrations in the center of Chengdu low; this result is closely related to the low elevations 2of these geographic locations. The NO2 and in Chongqing’s area wereand veryinhigh, and heavydowntown pollution was during concentrations in thedowntown center of Chengdu Chongqing’s area most wereobvious very high, and the winter. The extents of the highand median-value areas were large, and the overall spatial and heavy pollution was most obvious during the winter. The extents of the high- and median-value temporal inoverall the spring andand falltemporal were fairly similar. However, the and extent the fairly highareas weredistributions large, and the spatial distributions in the spring fallofwere and median-value areas in spring was obviously larger than that in fall; moreover, the lowest values similar. However, the extent of the high- and median-value areas in spring was obviously larger were in moreover, summer, during which most were areas observed were onlyinslightly polluted. than observed that in fall; the lowest values summer, duringSeveral which prominent most areas low-value areas, namely, Ya’an, northern Mianyang, southern Leshan, Ya’an, and the northeastern part were only slightly polluted. Several prominent low-value areas, namely, northern Mianyang, of Dazhou, Chongqing, were observed in terms of the PM concentration. The highest pollution 2.5 southern Leshan, and the northeastern part of Dazhou, Chongqing, were observed in terms of the occurred in winter; some of the pollution borders of Chengdu and Chongqing were moderately polluted, PM2.5 concentration. The highest occurred in winter; some of the borders of Chengdu and while otherswere weremoderately heavily polluted. Thewhile lowest pollution observed in the summer, during which Chongqing polluted, others werewas heavily polluted. The lowest pollution was

observed in the summer, during which the entire CCEZ was only slightly polluted. However,

Sensors 2018, 18, 3950

9 of 16

Sensors 2018, 18, x FOR PEER REVIEW

9 of 16

the entire CCEZ was only slightly polluted. However, differences were observed between spring and differences were observed between spring and fall: the border area was more polluted in spring than fall: the border area was more polluted in spring than in fall but the central area had relatively higher in fall but the central area had relatively higher levels of pollution in fall than in spring. levels of pollution in fall than in spring.

(a) Spring

(b) Summer

(c) Fall

(d) Winter

Figure6. 6. Seasonally averaged NO2 (upper panel) PMpanel) 2.5 (lower panel) concentration Figure Seasonally averaged NO2 (upper panel) and PM2.5and (lower concentration distributions from 2005 to 2015April, (Spring: April, June, and May; Summer: in the CCEZ indistributions the CCEZ from 2005 to 2015 (Spring: March, and March, May; Summer: July, and August;June, Fall: July, andOctober, August; September, October, and November; and Winter: December, September, andFall: November; and Winter: December, January, and February).

January, and February). To analyze the seasonal variation characteristics of the tropospheric NO2 column concentration and the concentration in the CCEZ, we statistically calculated quarterly average 2.5 mass To PM analyze the seasonal variation characteristics of the tropospheric NO2the column concentration NO column and the PMCCEZ, concentrations from 2005the to quarterly 2015. Figure 7 shows 2.5 masswe and2 the PM2.5concentrations mass concentration in the statistically calculated average NO2 the quarterly average forand each the 11 yearsfrom from2005 2005to to 2015. 2015, Figure and the7 error bars mass concentrations shows the column concentrations theseason PM2.5 spanning represent the standard deviation of the mean for a particular season over the 11-year period. The NO quarterly average for each season spanning the 11 years from 2005 to 2015, and the error bars 2 concentrations in the various seasons were ranked high toseason low as over follows: winter, fall, spring, represent the standard deviation of the mean for afrom particular the 11-year period. The and An analysis of various the sizesseasons of the error in thefrom graph indicates difference in the NOsummer. 2 concentrations in the werebars ranked high to lowthat as the follows: winter, fall, average value for a particular season over the 11-year period from 2005 to 2015 was consistent with spring, and summer. An analysis of the sizes of the error bars in the graph indicates that the the NO2 concentration. The PM2.5 was highest in 11-year winter and lowest in2005 summer, and it difference in the average value forconcentration a particular season over the period from to 2015 was was slightly higher in spring than in fall. Within a season, the mean difference over the 11-year period consistent with the NO2 concentration. The PM2.5 concentration was highest in winter and lowest in was rankedand as follows: winter higher > fall >in spring > summer. In the wintertime, cold is dominated summer, it was slightly spring than in fall. Within a season, theweather mean difference over by air flow; thus, pollutants do notwinter easily >disperse, which> issummer. one of the reasons for cold the thedownward 11-year period was ranked as follows: fall > spring In main the wintertime, higher NO and PM concentrations during the winter season [36]. Moreover, the CCEZ is located weather is2dominated 2.5 by downward air flow; thus, pollutants do not easily disperse, which is one of in themain Sichuan Basin,for which receivesNO a considerable of rain and is consequently referred[36]. to the reasons the higher 2 and PM2.5 amount concentrations during the winter season colloquially as the “Huaxi rain screen”. In the CCEZ, weather occurs in winter, droughts occur Moreover, the CCEZ is located in the Sichuan Basin, dry which receives a considerable amount of rain in spring, floods occur duringtosummer, and substantial rainfall annual and is consequently referred colloquially as the “Huaxi rain occurs screen”.inInfall. the The CCEZ, dry rainfall weather distribution is uneven, that is, 70~75% falls from June to October, which plays role in reducing occurs in winter, droughts occur in spring, floods occur during summer, and asubstantial rainfall the amount of pollution. According to research results, precipitation has little ability to remove occurs in fall. The annual rainfall distribution is uneven, that is, 70~75% falls from June October, atmospheric when the precipitation is less than 5 mm. However, when the daily which plays pollutants a role in reducing thedaily amount of pollution. According to research results, precipitation precipitation exceeds mm, the efficiency withpollutants which pollutants increases withisincreasing has little ability to5 remove atmospheric when are theremoved daily precipitation less than precipitation. The greater the amount of daily precipitation is, the greater the clearance rate and the 5 mm. However, when the daily precipitation exceeds 5 mm, the efficiency with which pollutants are better the air quality.with In addition, the precipitation. maximum clearance rate can 48.55% [37–39]. Rainfall affects removed increases increasing The greater thereach amount of daily precipitation is, the greater the clearance rate and the better the air quality. In addition, the maximum clearance rate can reach 48.55% [37–39]. Rainfall affects the concentration of pollutants in spring, summer, and fall, although air pollution is also affected by industrial and human factors [38].

Sensors 2018, 18, 3950

10 of 16

the concentration of pollutants in spring, summer, and fall, although air pollution is also affected by Sensors 2018,and 18, x human FOR PEER REVIEW 10 of 16 industrial factors [38].

Figure concentrations in the CCEZ from 2005 to 2015; Figure 7. 7. Seasonal Seasonal variations variations in in the the NO NO22 and and PM PM2.5 2.5 concentrations in the CCEZ from 2005 to 2015; error bars represent standard deviations. error bars represent standard deviations.

In this study, we further processed the Multiresolution Emission Inventory for China (MEIC, Table 1. Four sources of anthropogenic NOx emissions (i.e., industrial, power, residential, and http://meicmodel.org) data with a spatial resolution of 0.25◦ × 0.25◦ to estimate the anthropogenic transportation) during all seasons of 2012 in the CCEZ (Unit: 104 t·yr−1). emissions in different seasons over the CCEZ. Tables 1 and 2 show four sources of anthropogenic NOx Type Spring power, residential, Summer and transportation) Fall and PM2.5 emissions (i.e., industrial, in Winter different seasons Industry 15.45 (49.9%) 16.05were (52.2%) 15.41sources (43.9%)of NOx of 2012. For NOx, industry, power plants,15.59 and (51.9%) transportation the major Power 5.17 (16.7%) 4.07 (13.5%) 4.29 (13.9%) 6.85 (19.5%) emissions, with industry accounting for approximately a half of the total NOx emissions. In addition, Residential (4.8%) throughout 1.49 (5.0%) 1.47 (4.7%) Table 1 clearly shows that the1.49 emissions the four seasons are similar 3.9 for (11.1%) industry, power, Transportation 8.92 (28.7%) (29.6%) 8.92 (29.2%) 8.94in(25.5%) and transportation, while residents’ emissions 8.91 are significantly higher in winter than the other three seasons. Compared with NOx, anthropogenic PM2.5 emissions are mainly contributed by industrial and Table sources, 2. Four sources of anthropogenic PMfor 2.5 emissions industrial, power, residential, and to residential which together accounted more than(i.e., 90% of the emissions in 2012. Similar 4 t·yr−1). transportation) during all seasons in 2012 in the CCEZ (Unit: 10 NOx, industrial emissions accounted for more than half of the total PM2.5 emissions. As shown in Table 2, residential 64% of the PMFall winter, and these 2.5 emissions in Winter Type sources contributed Spring more than Summer emissions are approximately three times higher than those of the other three seasons. Industry 9.21 (53.4%) 8.96 (53.3%) 9.2 (54.1%) 9.08 (31.7%)

Power

0.53 (3.1%)

0.41 (2.4%)

0.44 (2.6%)

0.70 (2.5%)

Transportation

0.47 (2.7%)

0.46 (2.7%)

0.47 (2.8%)

0.47 (1.6%)

Table 1. Four sources of anthropogenic NOx emissions (i.e., industrial, power, residential, Residential 7.04 (40.8%) 6.99 (41.6%) 6.91 (40.7%) 18.36 (64.2%) and transportation) during all seasons of 2012 in the CCEZ (Unit: 104 t·yr−1 ). Type

Spring

Summer

Fall

Winter

In this study, we further15.45 processed Multiresolution Emission for China (MEIC, Industry (49.9%) the15.59 (51.9%) 16.05 (52.2%) Inventory 15.41 (43.9%) Power data with 5.17 a(16.7%) 4.07 (13.5%) 4.29 6.85 (19.5%) http://meicmodel.org) spatial resolution of 0.25° × (13.9%) 0.25° to estimate the anthropogenic Residentialseasons1.49 (4.8%) 1.49Tables (5.0%) 1 and1.47 (4.7%)four sources 3.9 (11.1%) emissions in different over the CCEZ. 2 show of anthropogenic Transportation 8.92 (28.7%) 8.91 (29.6%) 8.92 (29.2%) 8.94 (25.5%) NOx and PM2.5 emissions (i.e., industrial, power, residential, and transportation) in different seasons of 2012. For NOx, industry, power plants, and transportation were the major sources of Table 2. Four sources of anthropogenic PM2.5 emissions (i.e., industrial, power, residential, NOx emissions, with industry accounting for approximately a half of the total NOx emissions. In and transportation) during all seasons in 2012 in the CCEZ (Unit: 104 t·yr−1 ). addition, Table 1 clearly shows that the emissions throughout the four seasons are similar for industry, power,Type and transportation, Spring while residents’ Summer emissions Fallare significantly Winterhigher in winter than in the other three seasons. Compared 8.96 with NOx, anthropogenic 2.5 emissions are mainly Industry 9.21 (53.4%) (53.3%) 9.2 (54.1%) PM9.08 (31.7%) contributed by industrial and0.53 residential which together accounted for(2.5%) more than 90% of Power (3.1%) sources, 0.41 (2.4%) 0.44 (2.6%) 0.70 Residential 7.04 (40.8%) 6.99 (41.6%) 6.91 (40.7%) 18.36 (64.2%) the emissions in 2012. Similar to NOx, industrial emissions accounted for more than half of the total Transportation (2.7%) 0.46 (2.7%) 0.47 (2.8%) more 0.47 (1.6%) PM2.5 emissions. As shown in0.47 Table 2, residential sources contributed than 64% of the PM2.5 emissions in winter, and these emissions are approximately three times higher than those of the 3.3. Eleven-Year Variations in the NO2 and PM2.5 Concentrations in the CCEZ other three seasons.

Figure 8 shows the quarterly average for each year of the 2005–2015 period. The error 3.3. Eleven-Year Variations in the NO2 and PM2.5 Concentrations in the CCEZ bars represent the standard deviation of the mean of all months within a season. From 2005 Figure 8 shows the quarterly average for each year of the 2005–2015 period. The error bars represent the standard deviation of the mean of all months within a season. From 2005 to 2015, the NO2 concentration in winter exhibited the largest fluctuation with an increase of 348.86 × 1013 molecules·cm−2 and a decrease of 184.95 × 1013 molecules·cm−2. The NO2 concentration increased linearly from 2006 to 2010 and reached a peak in 2010, followed by a decrease and then an

Sensors 2018, 18, 3950

11 of 16

Sensors 2018, 18,NO x FOR concentration PEER REVIEW in winter exhibited the largest fluctuation with an increase 11 ofof 16 to 2015, the 2 13 − 2 13 − 2 348.86 × 10 molecules·cm and a decrease of 184.95 × 10 molecules·cm . The NO2 concentration increase from 2011 to 2006 2013toand linearindecrease from 2013 to 2015. and The then NO2 increased linearly from 2010subsequently and reached aa peak 2010, followed by a decrease concentration in spring decreased slightly from 2005 to 2008 but continued to increase from 2008 an increase from 2011 to 2013 and subsequently a linear decrease from 2013 to 2015. The NOto 2 13 −2 Over the 11-year period, the 2013 and reached a maximum increase of 129.13 concentration in spring decreased slightly from× 10 2005molecules·cm to 2008 but .continued to increase from NO2 to concentration fluctuated slightly during the of summer remained fairly The NO 13 molecules −2 . Over 2008 2013 and reached a maximum increase 129.13 but × 10 ·cmstable. the2 concentration in fall increased from 2006 to 2011 and exhibited an increase of 11-year period, the NO2 concentration fluctuated slightly during the summer but remained fairly 13 molecules·cm−2. After 2011, the NO2 concentration began to decline annually and 164.32 × 10 stable. The NO2 concentration in fall increased from 2006 to 2011 and exhibited an increase of −2 Based on the error bars in the figure, we can exhibited of ·111.39 1013 molecules·cm 164.32 × 10a13decrease molecules cm−2 . ×After 2011, the NO2. concentration began to decline annually and conclude athat the mean monthly was in winter then fall,we spring, 13 molecules −2 . Basedhighest exhibited decrease of 111.39 × 10fluctuation ·cmtypically on the error bars and in the figure, can and summer. The fluctuations in the PM 2.5 mass concentration during the fall and winter were conclude that the mean monthly fluctuation was typically highest in winter and then fall, spring, relatively largeThe with no regularity. The difference between the highest and lowest concentrations and summer. fluctuations in the PM 2.5 mass concentration during the fall and winter were −3 in winter and 41.11 µg·m−3 in fall. The PM2.5 concentration was fairly stable in was 51.93 µg·m relatively large with no regularity. The difference between the highest and lowest concentrations was spring from exhibited a steady decline from 2013 to 2015; this result is 51.93 µgand ·m−3summer in winter and2005 41.11to µg2013 ·m−3but in fall. The PM 2.5 concentration was fairly stable in spring and related to the “Suggestions on Haze Pollution in the Sichuan Basin and Suggestions for summer from 2005 to 2013 but exhibited a steady decline from 2013 to 2015; this result is related to the Countermeasures” formulated by the Sichuan Provincial Department of Environmental Protection “Suggestions on Haze Pollution in the Sichuan Basin and Suggestions for Countermeasures” formulated in the 2013 and the “Chongqing City Blue action implementation plan and (2013–2017) proposed by by Sichuan Provincial Department of Sky” Environmental Protection in 2013 the “Chongqing City Chongqing Municipality. The average monthly fluctuation in the PM 2.5 mass concentration was Blue Sky” action implementation plan (2013–2017) proposed by Chongqing Municipality. The average highest in winter and summer, while the fluctuation in spring was unstable and greater monthly fluctuation in lowest the PMin 2.5 mass concentration was highest in winter and lowest in summer, than that in fall. while the fluctuation in spring was unstable and greater than that in fall.

13 molecules·cm−2 ) (upper panel) and PM Figure Seasonally averaged averagedNO NO ·m−3 ) 13 molecules·cm −2) (upper panel) and PM2.5 (µg·m −3) (lower 2.5 (µg Figure 8.8. Seasonally 22 (10(10 (lower concentration changes in the CCEZ from2005 2005toto 2015; 2015; vertical vertical bars panel) panel) concentration changes in the CCEZ from bars represent represent standard deviations. standard deviations.

We used the monthly average data of 2005–2015 and the regression model developed by Lamsal et al. to compare the interannual changes over the 11-year period in the CCEZ [40–42]. The

Sensors 2018, 18, 3950

12 of 16

Sensors 2018, 18, x FOR PEER REVIEW

12 of 16

used of 2005–2015 and the the seasonal regression model developed by timeWe series of the the monthly monthly average average data consists of three parts: trend, linear trend, and Lamsal et al. to compare the interannual changes over the 11-year period in the CCEZ [40–42]. The irregular fluctuation: time series of the monthly average consists of three parts: the seasonal trend, linear trend, and irregular (1) Ω t S t T t I t fluctuation: t) + T(t)the + I(seasonal t) (1) (t) =InS(general, where t represents time (in months)Ω[43]. trend is composed of sine and cosine functions [44]: where t represents time (in months) [43]. In general, the seasonal trend is composed of sine and cosine functions [44]: 3
sin(2πjt/12 2πjt/12) + c c2j cos cos(2πjt/12 2πjt/12) (2) SS(t)t = cc0 + ∑ cc1j sin (2) j=1

where c0 , c1j , and c2j are constant coefficients. The monthly average value changes with the month; where c0, c1j, and c2j are constant coefficients. The monthly average value changes with the month; therefore, we used the 12-month moving average method to calculate the monthly index. We identified therefore, we used the 12-month moving average method to calculate the monthly index. We and extracted the seasonal trend components by exploiting changes in the measured seasonal pattern identified and extracted the seasonal trend components by exploiting changes in the measured (amplitude and phase) for individual years. Eleven regression lines (2005–2015) are based on annual seasonal pattern (amplitude and phase) for individual years. Eleven regression lines (2005–2015) are observations plus six-month observations in adjacent years. Comparing the 11 regression lines based on annual observations plus six-month observations in adjacent years. Comparing the 11 with high and low amplitudes, two seasonal terms (S(t) = s1 + s2 Equation (1)) were identified regression lines with high and low amplitudes, two seasonal terms (S(t) = s1 + s2 Equation (1)) were and isolated. In this paper, we calculated the changes in the deseasonalized trends of the NO2 identified and isolated. In this paper, we calculated the changes in the deseasonalized trends of the columns (Figure 9a) and PM2.5 (Figure 9b) from 2005 to 2015. The error bars show that the monthly NO2 columns (Figure 9a) and PM2.5 (Figure 9b) from 2005 to 2015. The error bars show that the fluctuation in NO2 was greater than that in PM2.5 . The black-filled circles in the figure show that monthly fluctuation in NO2 was greater than that in PM2.5. The black-filled circles in the figure show the NO2 and PM2.5 concentrations generally exhibited upward trends from 2005 to 2011, followed that the NO2 and PM2.5 concentrations generally exhibited upward trends from 2005 to 2011, by downward trends after 2013. The annual mean NO2 concentration was fairly similar from 2011 followed by downward trends after 2013. The annual mean NO2 concentration was fairly similar to 2013 but exhibited a subsequent downward trend, which was attributed to the environmental from 2011 to 2013 but exhibited a subsequent downward trend, which was attributed to the strategy formulated for the regional planning requirements of the CCEZ. The final year for the environmental strategy formulated for the regional planning requirements of the CCEZ. The final total emission reduction of major pollutants in the 12th Five-Year Plan was 2015, during which year for the total emission reduction of major pollutants in the 12th Five-Year Plan was 2015, during all cities in Sichuan and Chongqing Municipalities accelerated the key emission reduction projects which all cities in Sichuan and Chongqing Municipalities accelerated the key emission reduction pertaining to the total amount of major pollutants, and they achieved remarkable results. By the end projects pertaining to the total amount of major pollutants, and they achieved remarkable results. By of 2015, the NO2 concentration decreased by 15.4% compared with 2011, and the PM2.5 concentration the end of 2015, the NO2 concentration decreased by 15.4% compared with 2011, and the PM2.5 decreased by 38%. In 2011, the concentrations of NO2 and PM2.5 reached their highest values of concentration decreased by−38%. In 2011, the concentrations of NO2 and PM2.5 reached their highest 450.71 × 1013 molecules ·cm 2 and 78.96 µg·m−1 , respectively, and in 2015, they reached their lowest 13 molecules·cm−2 and 78.96 µg·m−1, respectively, and in 2015, they reached their values of 450.71 × 1013 values of 381.19 × 10 molecules ·cm−2 and 48.98 µg·m−3 , respectively. lowest values of 381.19 × 1013 molecules·cm−2 and 48.98 µg·m−3, respectively.

(a) Figure 9. Cont.

Sensors Sensors2018, 2018,18, 18,3950 x FOR PEER REVIEW Sensors 2018, 18, x FOR PEER REVIEW

13 13 of of 16 16 13 of 16

(b) (b) Figure 9. Changes Changes in the the annualaverage average NO22 (a) (a) andPM PM2.5 (b) (b) concentrations concentrations in in the the CCEZ CCEZ during during Figure Figure 9. 9. Changes in in theannual annual averageNO NO2 (a) and and PM2.5 2.5 (b) concentrations in the CCEZ during 2005–2015; gray graycircles circlesrepresent representdeseasonalized deseasonalized monthly columns, while black-filled circles show 2005–2015; monthly columns, while black-filled circles show the 2005–2015; gray circles represent deseasonalized monthly columns, while black-filled circles show the annual means, vertical show the standard deviations. annual means, and and vertical barsbars show the standard deviations. the annual means, and vertical bars show the standard deviations.

The average NO NO and andPM PM concentrations over CCEZ shown in Figure 10. The monthly average 2.52.5 concentrations over thethe CCEZ are are shown in Figure Figure 10. The The The monthly average NO222and PM2.5 concentrations over the CCEZ are shown in 10. The monthly average concentrations NO PM showed cyclicalvariations variationsthroughout throughoutthe the year. year. monthly average concentrations of of NO 2 and PM 2.5 2.5 showed cyclical 2 and monthly average concentrations of NO 2 and PM2.5 showed cyclical variations throughout the year. The troughs of NO and PM in each cycle generally occurred in July and August (i.e., summer); The troughs troughs of of NO NO222 and and PM PM2.5 2.5 The 2.5 in each cycle generally occurred in July and August (i.e., summer); 1313 molecules·cm−−22 in NO aaminimum value ofof230.15 2007, and NO reached minimum value 230.15×××10 1013 molecules·cm−2 in in August August 2007, and PM PM2.5 2.5 reached a a 22 reached NO2 reached a minimum value of 230.15 10 molecules·cm and PM 2.5 reached a − 3 −3 August minimum value ofof 26.43 µg·µg·m m in 2015. The peaks usually occurred in December and January minimum value 26.43 in August 2015. The peaks usually occurred in December and minimum value of 26.43 µg·m−3 in August 2015. The peaks usually in December and −2 inoccurred −2 in 2012, (i.e., winter); maximum value of 975.18 December and PM January (i.e.,NO winter); NO22areached reached maximum valuemolecules.cm of 975.18 975.18 molecules.cm molecules.cm December 2012, 2 reached 2.5 −2 in January (i.e., winter); NO aa−3maximum value of December 2012, −3 reached a2.5maximum of maximum 157.64 µg·mof 157.64 in February 2010. From 2005 to 2008, the tropospheric NO and PM reached a µg·m in February 2010. From 2005 to 2008, the 2 and PM2.5 reached a maximum of 157.64 µg·m−3 in February 2010. From 2005 to 2008, the columnar concentration and PMconcentration mass concentration were fairly stable and exhibited few fluctuations. tropospheric NO 2 columnar and PM 2.5 mass concentration were fairly stable and 2.5 tropospheric NO2 columnar concentration and PM2.5 mass concentration were fairly stable and From 2008few to 2010, the NO2From concentration continued increase, and continued the PM2.5 mass concentration exhibited fluctuations. 2008 to to 2010, 2010, the NO NOto 2 concentration to increase, and the exhibited few fluctuations. From 2008 the 2 concentration continued to increase, and the fluctuated The concentrations of NOThe PM2.5 dropped sharply 2011 compared with 2.5 massslightly. concentration fluctuated slightly. concentrations of NO 2 andin PM 2.5 dropped sharply PM 2 and PM2.5 mass concentration fluctuated slightly. The concentrations of NO2 and PM2.5 dropped sharply those in previous years, and the concentrations in 2012 were unchanged from the 2010 levels. Between in 2011 2011 compared compared with with those those in in previous previous years, years, and and the the concentrations concentrations in in 2012 2012 were were unchanged unchanged from from in 2013 and 2015, the NO2 and PMand decrease periodically, andtoindecrease August the 2010 levels. Between 2013 2015, the NO 2continued and PM 2.5toconcentrations continued 2.5 concentrations the 2010 levels. Between 2013 and 2015, the NO2 and PM2.5 concentrations continued to decrease 2015, they reached lowest2015, values in reached the 11-year periodically, and in inthe August they the period. lowest values values in in the the 11-year 11-year period. period. periodically, and August 2015, they reached the lowest

13 molecules·cm−2−2 3) Figure ·m−−3 Figure 10. 10. Changes Changes in in the themonthly monthlyaverage averageNO NO 2 (1013 molecules·cm ) ) and and PM PM (µg·m 2 (10 2.52.5 (µg Figure 10. Changes in the monthly average NO 2 (1013 molecules·cm−2) and PM2.5 (µg·m−3) concentrations concentrations in in the the CCEZ CCEZ from from2005 2005to to2015. 2015. concentrations in the CCEZ from 2005 to 2015.

Sensors 2018, 18, 3950

14 of 16

4. Conclusions This study analyzed the spatial–temporal distributions in NO2 and PM2.5 using satellite remote sensing data in the CCEZ over 11 study years. Geographically, the area with the highest level of column NO2 is Chengdu, which presented an average value of 848.55 × 1013 molecules·cm−2 in the 11-year study period. The area with the highest near-surface PM2.5 concentration is Zigong, which presented an average value of 91.34 µg·m−3 over the 11-year study period. The area with the lowest concentration of NO2 and PM2.5 is Ya’an, which presented average values of 91.34 µg·m−3 and 53.26 µg·m−3 over the 11-year study period, respectively. Seasonally, the tropospheric NO2 column concentration and near-surface PM2.5 mass concentration in the CCEZ all showed obvious monthly and seasonal variation characteristics. The concentrations of NO2 and PM2.5 reach their highest annual values usually in December and January in winter and their lowest annual values in July and August in summer. For anthropogenic emissions, industrial emissions contributed approximately 50% of NOx and PM2.5 emissions. For PM2.5 , residential emissions contributed more than 64% in winter, and these emissions were approximately three times higher than those in the other three seasons. During our study period, NO2 presented relatively greater fluctuations. From 2005 to 2011, the concentration of column NO2 showed annual increases and reached a high value of 953.33 × 1013 molecules·cm−2 in January 2011, after which the column NO2 concentration showed a downward trend, although it increased in 2013 and then continued to decline. From 2005 to 2011, the concentration of PM2.5 was basically unchanged and presented an annual average of 76.52 µg·m−3 . After 2011, the concentration of PM2.5 began to decline sharply, and the lowest value of 26.43 µg·m−3 was found in August of 2015. Supplementary Materials: The following are available online at http://www.mdpi.com/1424-8220/18/11/3950/ s1: Table S1: The mean (plus standard deviation) NO2 concentrations (1013 molecules·cm−2 ) for the sixteen cities in the CCEZ during 2005–2015; Table S2: The mean (plus standard deviation) PM2.5 concentrations (µg·m−3 ) for the sixteen cities in the CCEZ during 2005–2015. Author Contributions: K.C. and S.L. conceived the study; K.C. and Q.Z. performed the data analyses and wrote the manuscript; Q.Z. and Y.L. conducted the validation; K.C., S.L. and W.G. contributed to the revisions. Funding: This research was financially supported by the National Key R&D Program of China (2016YFC0201507) and the National Natural Science Foundation of China (grant Nos. 41801281 and U1704122). Acknowledgments: We would like to acknowledge the use of the tropospheric NO2 column data from http://www.temis.nl, and we are grateful for the use of the MODIS aerosol data that are freely available at https://ladsweb.modaps.eosdis.nasa.gov/search/. Conflicts of Interest: The authors declare no conflicts of interest.

References 1. 2.

3.

4. 5.

Li, Y.P.; Chen, G.B.; Tong, X.S. Fuzzy Comprehensive Assessment of the Urban Air Quality in the Chengdu-Chongqing Economic Zone. Sichuan Environ. 2012, 31, 107–110. Zyrichidou, I.; Koukouli, M.E.; Balis, D.S.; Kioutsioukis, I.; Poupkou, A.; Katragkou, E.; Melas, D.; Boersma, K.F.; Roozendael, M.V. Evaluation of high resolution simulated and OMI retrieved tropospheric NO2 column densities over Southeastern Europe. Atmos. Res. 2013, 122, 55–66. [CrossRef] Rui, W.; Guan, L.; Zhang, F.; Zhang, W.; Ding, W. PM2.5 -induced oxidative stress increases adhesion molecules expression in human endothelial cells through the ERK/AKT/NF-WeiB-dependent pathway. J. Appl. Toxicol. 2015, 36, 48–59. [CrossRef] [PubMed] Kornartit, C.; Sokhi, R.S.; Burton, M.A.; Ravindra, K. Activity pattern and personal exposure to nitrogen dioxide in indoor and outdoor microenvironments. Environ. Int. 2010, 36, 36–45. [CrossRef] [PubMed] Wang, Z.Y. Study on Nitrogen Dioxide Emission Characteristics of High Pressure Common Rail Diesel Engine. Master’s Thesis, Southwest Forestry University, Kunming, China, 2013.

Sensors 2018, 18, 3950

6.

7.

8.

9.

10. 11.

12. 13. 14.

15. 16. 17. 18. 19. 20.

21. 22. 23. 24.

25.

26.

15 of 16

Öztaner, Y.B.; Güney, B.; Özçomak, D.; Kahya, C.; Balçık, F.B.; Çakır, S.; Kalkan, K. An Investigation of the Relationship between Modis AOD and PM2.5 over Marmara Region, Turkey. In Proceedings of the VII Atmospheric Science Symposium, Istanbul, Turkey, 28–30 April 2015. Xu, Y.F. Analysis of the Influence of Haze and PM2.5 Pollution Levels and Thermal Emission Sources on Atmospheric Environment in Zhengzhou City. Master’s Thesis, Zhengzhou University, Zhengzhou, China, 2012. Zhang, X.; Ai, J.; Jiang, W.; Tan, Y.; Liao, T.; Sun, Y. The Trend and Spatiotemporal Distribution Characteristics of Troposphere NO2 Columns Concentration in Chengdu Region During 2005–2015. J. Chengdu Univ. Inf. Technol. 2016, 7, 656–664. Liu, X.T.; Zheng, T.F.; Wan, Q.L.; Tan, H.B.; Deng, X.J.; Li, F.; Deng, T. Spatial-temporal Characteristics of NO2 in Concentrated PRD Urban Districts and Analysis of Anthropogenic Influences Based on OMI Remote Sensing Data. J. Trop. Meteorol. 2015, 31, 193–201. Long, L.I.; Shi, R.; Chen, Y.; Yongming, X.U.; Bai, K.; Zhang, J. Spatio-temporal Characteristics of NO2 in China and the Anthropogenic Influences Analysis Based on OMI Data. J. Geo-Inf. Sci. 2013, 15, 688–694. Krotkov, N.A.; Mclinden, C.A.; Li, C.; Lamsal, L.N.; Celarier, E.A.; Marchenko, S.V.; Swartz, W.H.; Bucsela, E.J.; Joiner, J.; Duncan, B.N. Aura OMI observations of regional SO2 and NO2 pollution changes from 2005 to 2014. Atmos. Chem. Phys. Discuss. 2015, 15, 26555–26607. [CrossRef] Guo, J.P.; Wu, Y.R.; Zhang, X.Y.; Li, X.W. Estimation of PM2.5 over eastern China from MODIS aerosol optical depth using the back propagation neural network. Environ. Sci. 2013, 34, 817–825. Qin, Z.H. Research on Sustainable Development of Chengdu—Chongqing Old Industrial Base; Minzu University of China: Beijing, China, 2011. Yi, P.; Duan, N.; Xu, Y.X.; Yu, H.T. Effects of Coal-Fired Thermal Power Industry Development on the Atmospheric Environment in the Chengdu-Chongqing Economic Zone. Res. Environ. Sci. 2012, 25, 1107–1114. Yang, Y.L. Emissions Reduction of NOX in Coal-fired Power Plant and Simple Analysis of SCR Flue Gas Denitration Technology. Energy Environ. Prot. 2017, 31, 31–35. Zhu, J.; Zheng, F.; Wang, H.; Wang, F.; Shu, W. Analysis on Development Trend of Desulfurization and Denitration Technologies for Coal-fired Flue Gas. J. Environ. Eng. Technol. 2015, 5, 200–204. Li, L. Analysis and Countermeasure Research on the Atmospheric Environment Changes in the Main City of Chongqing (2009–2014). Wirel. Internet Technol. 2015, 11, 85–86. Yi, D. Total Amount of Nitrogen Oxide Emissions and Countermeasures in Sichuan Province. Master’s Thesis, Southwest Jiaotong University, Chengdu, China, 2013. Ren, L.H.; Zhou, Z.E.; Zhao, X.Y.; Yang, W.; Yin, B.H. Source Apportionment of PM10 and PM2.5 in Urban Areas of Chongqing. Res. Environ. Sci. 2014, 27, 1387–1394. Zhou, C.; Qing, L.I.; Wang, Z.; Gao, Y.; Zhang, L.; Hui, C.; Pengfei, M.A.; Chang, T.; Center, S.E. Spatio-temporal trend and changing factors of tropospheric NO2 column density in Beijing-Tianjin-Hebei region from 2005 to 2014. J. Remote Sens. 2016, 20, 468–480. Sun, L.; Li, R.; Tian, X.; Wei, J. Analysis of the temporal and spatial variation of aerosols in the Beijing-Tianjin-Hebei region with a 1 km AOD product. Aerosol Air Qual. 2017, 17, 923–935. [CrossRef] Lin, C.Q.; Li, Y.; Lau, A.K.H.; Li, C.C.; Fung, J.C.H. 15-Year PM2.5 Trends in the Pearl River Delta Region and Hong Kong from Satellite Observation. Aerosol Air Qual. 2018, 18, 2355–2362. [CrossRef] Zhao, C.; Wang, Y.H. Assimilated inversion of NOx emissions over east Asia using OMI NO2 column measurements. Geophys. Res. Lett. 2009, 36, 150–164. [CrossRef] Boersma, K.F.; Eskes, H.J.; Dirksen, R.J.; van der, A.; Veefkind, J.P.; Stammes, P.; Huijnen, V.; Kleipool, Q.L.; Sneep, M.; Claas, J.; et al. An improved tropospheric NO2 column retrieval algorithm for the Ozone Monitoring Instrument. Atmos. Meas. Tech. 2011, 4, 2329–2388. [CrossRef] Yoshimoto, K. Long-term MAX-DOAS network observations of NO2 in Russia and Asia (MADRAS) during 2007–2012: Instrumentation, elucidation of climatology, and comparisons with OMI satellite observations and global model simulations. Atmos. Chem. Phys. 2014, 14, 2883–2934. Mou, F.S.; Li, A.; Xie, P.H.; Wang, Y.; Zhang, J. August 2013—July 2014 NO2 Tropospheric Column Concentration Analysis and Satellite Comparison in Hefei. Spectrosc. Spectr. Anal. 2017, 37, 1042–1047.

Sensors 2018, 18, 3950

27.

28. 29.

30. 31. 32. 33. 34. 35. 36.

37. 38. 39. 40.

41.

42. 43.

44.

16 of 16

Jin, J.; Ma, J.; Lin, W.; Zhao, H.; Shaiganfar, R.; Beirle, S.; Wagner, T. MAX-DOAS measurements and satellite validation of tropospheric NO2 and SO2 vertical column densities at a rural site of North China. Atmos. Environ. 2016, 133, 12–25. [CrossRef] Levy, R.C.; Mattoo, S.; Munchak, L.A.; Remer, L.A. The Collection 6 MODIS aerosol products over land and ocean. Atmos. Meas. Tech. 2013, 6, 2989–3034. [CrossRef] Ma, Z.; Hu, X.; Sayer, A.; Levy, R.; Zhang, Q.; Xue, Y.; Tong, S.; Bi, J.; Huang, L.; Liu, Y. Satellite-Based Spatiotemporal Trends in PM2.5 Concentrations: China, 2004–2013. Environ. Health Perspect. 2016, 124, 184–192. [CrossRef] [PubMed] European Space Agency Data User Element. Available online: http://due.esrin.esa.int/page_globcover.php (accessed on 22 August 2016). Chongqing Provincial Bureauof Statistics. Statistical Yearbook of Chongqing. Available online: http://www.cqtj.gov.cn (accessed on 1 October 2017). Sichuan Provincial Bureauof Statistics. Statistical Yearbook of Sichuan Province. Available online: http://www.sc.stats.gov.cn (accessed on 1 October 2017). Zhou, S.H.; Chang-Jian, N.I.; Liu, P.C. Characteristics of temperature inversion in atmospheric boundary layer in Chengdu area. J. Meteorol. Environ. 2015, 2, 016. Yi, P.; Duan, N.; Chai, F.; Xu, Y.; He, Y. SO2 emission cap Planning for Chengdu-Chongqing economic zone. J. Environ. Sci. 2012, 24, 142–146. [CrossRef] National Developmentand Reform Commission. The Regional Planning of Chengdu-Chongqing Economic Zone. Available online: http://www.ndrc.gov.cn (accessed on 30 May 2011). Guo, X.M. Observation and Simulation of the Climate Characteristics of Air Quality and the Effects of Large Topography in Sichuan Basin. Master’s Thesis, Nanjing University of Information Engineering, Nanjing, China, 2016. Wang, N.; He, T.R.; Liu, J.P. Analysis of the Effect of Summer Precipitation on Air Pollutants in Chongqing Urban Area. Environ. Eng. 2017, 35, 69–73. Chen, Y.; Sun, J.Z.; Zhang, W.; Meng, X.W.; Wang, J.; Sun, J.Z. Research on the correlation between MODIS aerosol optical depth and air pollution index over Lanzhou area. J. Lanzhou Univ. 2013, 49, 765–772. Xiao, Z. Study on the Characteristics of Atmospheric NO2 in Sichuan Basin. Chin. Environ. Sci. 2011, 31, 1782–1788. Lamsal, L.N.; Duncan, B.N.; Yoshida, Y.; Krotkov, N.A.; Pickering, K.E.; Streets, D.G.; Lu, Z.U.S. NO2 trends (2005–2013): EPA Air Quality System (AQS) data versus improved observations from the Ozone Monitoring Instrument (OMI). Atmos. Environ. 2015, 110, 130–143. [CrossRef] Krotkov, N.A.; Mclinden, C.A.; Li, C.; Lamsal, L.N.; Celarier, E.A.; Marchenko, S.V.; Swartz, W.H.; Bucsela, E.J.; Joiner, J.; Duncan, B.N. Aura OMI observations of regional SO2 and NO2 pollution changes from 2005 to 2015. Atmos. Chem. Phys. 2016, 16, 4605–4629. [CrossRef] Randel, W.J.; Cobb, J.B. Coherent variations of monthly mean total ozone and lower stratospheric temperature. J. Geophys. Res. Atmos. 1994, 99, 5433–5447. [CrossRef] Lafare, A.E.A.; Peach, D.W.; Hughes, A.G. Use of seasonal trend decomposition to understand groundwater behaviour in the Permo-Triassic Sandstone aquifer, Eden Valley, UK. Hydrogeol. J. 2016, 24, 141–158. [CrossRef] Schneider, P.; van der, A.R.J. A global single-sensor analysis of 2002–2011 tropospheric nitrogen dioxide trends observed from space. J. Geophys. Res. Atmos. 2012, 117. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).