atmosphere

Article The Study of Emission Inventory on Anthropogenic Air and Source Apportionment of PM2.5 in the Changzhutan Urban Agglomeration, China

Bin Xu 1,2, Xiangyu You 1,*, Yaoyu Zhou 2, Chunhao Dai 2,*, Zhan Liu 1, Shaojian Huang 2, Datong Luo 1 and Hui Peng 2 1 Hunan Provincial Key Laboratory of Water Pollution Control Technology, Hunan Research Academy of Environmental Sciences, 410004, China; [email protected] (B.X.); [email protected] (Z.L.); [email protected] (D.L.) 2 College of Resources and Environment, Hunan Agricultural University, Changsha 410128, China; [email protected] (Y.Z.); [email protected] (S.H.); [email protected] (H.P.) * Correspondence: [email protected] (X.Y.); [email protected] (C.D.)

 Received: 27 May 2020; Accepted: 6 July 2020; Published: 12 July 2020 

Abstract: As one of China’s emerging urban agglomerations, the Changzhutan urban area is suffering from regional composite . Previous studies mainly focus on single cities or world-class urban agglomerations, which cannot provide a scientific basis for air pollution in emerging urban agglomerations. This paper proposes the latest high-resolution emission inventory through the emission factor method and compares the results with the rest of the urban agglomeration. The emission inventory shows that the estimates for (SO2), nitrogen oxides (NOX), particulate matter 10 (PM10), particulate matter 2.5 (PM2.5), volatile organic compounds (VOCs), and (NH3) emission are 132.5, 148.9, 111.6, 56.5, 119.0, and 72.0 kt, respectively. From the 3 3 km emission grid, the spatial difference of air emissions in the Changzhutan urban × agglomeration was more obvious, but the overall trend of monthly pollutant discharge was relatively 2 stable. Depending on the source apportionment, SO4 −, OC, and NO3− are the main chemical 3 constituents of PM2.5, accounting for 13.06, 8.24, and 4.84 µg/m , respectively. Simultaneously, industrial emissions, vehicle exhaust, and dust are still three main sources that cannot be ignored. With the support of these data, the results of this study may provide a reference for other emerging urban agglomerations in air quality.

Keywords: anthropogenic source; Changzhutan urban agglomeration; emission inventory; source apportionment of particulate matter

1. Introduction With the effective management of traditional pollution, the continuous expansion of urban areas, and the long-distance transmission of atmospheric pollution, the characteristics of regional composite air pollution is gradually emerging. Analysis of pollution sources is the precondition for the treatment of regional composite air pollution. The emission inventory method can demonstrate the spatial–temporal distribution of air pollutants emissions in various regions and accurately simulate the atmospheric environmental quality of the region [1]. Therefore, establishing an air pollution source information database and constructing a key pollution source spectrum are of major practical significance for comprehensively grasping regional environmental pollution characteristics and improving regional environmental quality. As one of the economic engines in Central China, the Changzhutan (CZT) urban agglomeration is famous for the metallurgical and chemical industry in China, which may inevitably lead to the deterioration of urban and regional atmospheric quality.

Atmosphere 2020, 11, 739; doi:10.3390/atmos11070739 www.mdpi.com/journal/atmosphere Atmosphere 2020, 11, 739 2 of 19

Air pollutant emission inventory is a crucial basis for pollutant emission control and management. It contributes to establishing an air quality model and provide guidance for decision-makers by supplying information on pollutant emission sources and their characteristics. Over the past ten years, considerable emission inventories have been established covering Asia [2], Europe [3], and America [4]. These studies focus more on the primary air pollutants including NOX, SO2, NH3, etc. In China, the air pollutant emission inventory has also been developed widely. Hao has established a NOx emission inventory utilizing energy consumption over the past decade [5]. Cao and Chen estimate the emission inventory of aerosols throughout China [6,7]. Qiang (2007) has classified the emission of particulate into three size ranges by a technology-based emission inventory approach [8]. After that, more studies have begun to consider volatile organic compounds (VOCs) emission in China due to its potential formation of ozone [9,10]. At present, the research areas of emission inventory are mainly concentrated on single cities or world-class city clusters. These studies cannot provide regional composite air pollution data support or technical assistance for the newly developed urban agglomeration. The CZT urban agglomeration is the first urban agglomeration in China to conduct regional economic integration experiments spontaneously in the early 21st century. It is a typical representative of the newly emerging urban agglomeration in the Yangtze River Economic Belt. Studying the emission inventory and source apportionment of particulate matter 2.5 (PM2.5) of CZT urban agglomeration can provide new ideas for other emerging urban agglomerations. This paper focuses on creating an air pollutant emission inventory of the CZT urban agglomeration (the primary agglomeration for economic development) based on the year 2015. Air pollutants studied in this research include sulfur dioxide (SO2), nitrogen oxides (NOX), particulate matter 10 (PM10), particulate matter 2.5 (PM2.5), volatile organic compounds (VOCs), and ammonia (NH3). Utilizing the main pollution source emission inventory treatment model, we established an air pollution source emission inventory platform serving the total atmospheric pollutant control and air quality simulation of typical regional industries. At the same time, our findings in CZT urban agglomeration can provide data support and theoretical basis for other emerging urban agglomerations to explore development governance and improve urban air quality.

2. Materials and Methods

2.1. Study Domain The CZT urban agglomeration lies in the central-eastern part of Hunan Province, consisting of Changsha, , and . It is located at 111◦530 E–114◦150 E and 26◦030 N–28◦410 N (see Figure1). Owing to its central position in South China, this region is not only worked as a transportation junction but also regarded as an engine of Central China. The terrain is low in the northwest and high in the southeast. Additionally, this region is surrounded by mountains on three sides. Hence, this region is prone to suffering from heavy air pollution during a pollution episode. Atmosphere 2020, 11, 739 3 of 19 Atmosphere 2020, 11, x FOR PEER REVIEW 3 of 19

Figure 1. RegionalRegional diagram diagram of of the the Changzhutan Changzhutan urban urban agglomeration. agglomeration.

2.2. Methodology Methodology 2.2.1. Emission Inventory Methods 2.2.1. Emission Inventory Methods This study considers six conventional air pollution sources(SO2, NOX, PM10, PM2.5, VOCs, and This study considers six conventional air pollution sources(SO2, NOX, PM10, PM2.5, VOCs, and NH3). The research methods in this paper mainly refer to Kurokawa [1], including the emission factor NH3). The research methods in this paper mainly refer to Kurokawa [1], including the emission method used to estimate the emission value and the Monte Carlo method to analyze the uncertainty. factor method used to estimate the emission value and the Monte Carlo method to analyze the According to the actual situation of the air pollution source and emission factor method, we divide our uncertainty. According to the actual situation of the air pollution source and emission factor emission sources into 4 levels, including 10 major industries (i.e., stationary fuel combustion source, method, we divide our emission sources into 4 levels, including 10 major industries (i.e., stationary process source, mobile source, solvent source, agricultural source, dust source, biomass combustion fuel combustion source, process source, mobile source, solvent source, agricultural source, dust source, storage and transportation sources, waste treatment source, and other sources). The detailed source, biomass combustion source, storage and transportation sources, waste treatment source, pollution source categories are summarized in Table S1. and other sources). The detailed pollution source categories are summarized in Table S1. Emission estimation using the emission factor method is achieved by Formula (1): Emission estimation using the emission factor method is achieved by formula (1):   Ei,j =A= Aj,k×EFEFi,j,k×1−η1 η (1)(1) , , × ,, × − i,j, where A is activity data; EF is the average emission factor; ηη is the removal efficiency efficiency of pollution control measures for for pollutants; pollutants; i, i, j, j, k k represent the pollutant type, type, source source category, category, and technology type,type, respectively. respectively. InIn the process of quantitative estimation, the the list list of of emission sources sources is is usually calculated by by selecting representative emission factors and activityactivity level data.data. The uncertainty of emission factors isis inevitable inevitable due due to to the the lack lack of of key key data data and and the the differences differences between between various various pollution pollution sources. sources. Therefore, itit is is important important to quantitativelyto quantitatively estimate estimate the potential the potential uncertainty uncertainty of the emission of the inventory.emission inventory.The Monte The Carlo Monte method Carlo was method used to quantifywas used the to emission quantify uncertainty the emission of eachuncertainty pollution of source. each pollutionThis method source. relies This on method repeated relies and randomon repeated sampling and random to obtain sampling results. to Within obtain the results. 95% confidenceWithin the 95%interval, confidence 10,000 testsinterval, were 10,000 conducted tests towere calculate conduc theted emission to calculate range. the The emission input parameters range. The of input each parameterspollution source of each refer pollution to the studies source ofrefer other to the scholars. studies of other scholars.

2.2.1.2.2.2. Source Source Apportionment Apportionment Methods Methods The Chemical Mass Balance model (CMB) and the Positive Matrix Factorization model model (PMF) (PMF) are the most widely used source resolution receptorreceptor models [11]. [11]. These These two two models can can start from thethe receptor andand estimateestimate thethe contributioncontribution value value of of various various pollution pollution sources sources to to the the receptor receptor based based on on the physical and chemical characteristics of VOCs in the atmospheric environment. In addition, for regional transmission characteristics, the mixed single-particle Lagrangian ejector model

Atmosphere 2020, 11, 739 4 of 19 the physical and chemical characteristics of VOCs in the atmospheric environment. In addition, for regional transmission characteristics, the mixed single-particle Lagrangian ejector model (HYSPLIT) was applied to calculate the atmospheric trajectory, revealing the air mass motion of PM2.5 in the CZT urban agglomerations. The whole research area was divided into a 0.5 0.5 grid, and the potential ◦ × ◦ source contribution factor (PSCF) and concentration weight trajectory (CWT) analysis by the MeteoInfo method were used to semi-quantitatively analyze the PM2.5 emission source intensity. Cij is a numerical value calculated by the model, representing the concentration of the value of ingredient in the grid i j, which can be used to explain the influence of the air volume of grid i j on the concentration of × × pollutants. Models used in this study are described in the supplementary information Table S2.

2.3. Activity Data We used eight major data sources to ensure the refinement of emission inventory data, including longitude/latitude coordinates of pollution sources, product types, fuel categories, technical processes, and pollution control measures. These data cover the vast majority of the required activity level, and the sources of these data used for the CZT inventory are summarized in Table S3. Moreover, the PM2.5 concentration data comes from the automatic air monitoring station of the Hunan Environmental Protection Agency. The source spectrum data required by the CMB model comes from the study by You [12]. The uncertainty required by the PMF model was obtained by fitting the system formula in the model.

3. Results and Discussion

3.1. Emission Factor This study used the emission factor method, which requires the collection of localized emission factors. The Chinese emission inventory was mostly based on foreign emission factors in the past. The difference in actual development has made these databases unable to adapt to the CZT urban agglomeration. It is necessary to select appropriate emission factors from the literature based on the current situation of pollution in CZT urban agglomeration to ensure the accuracy of data estimation. This study summarized the emission factor database suitable for CZT urban agglomeration. The emission factors of SO2, NOX, PM (PM10, PM2.5), VOCs, and NH3 come from the latest literature. Domestic measurement or related research should be given priority in this study. Unspecified emission factors refer to the AP-42 database [13] and the CORINAIR (Core Inventory of Air Emissions) database. The general description of our selection or estimation of specific emission factors is described below.

3.1.1. SO2

SO2 is mainly derived from the combustion of sulfur-containing fuels in power plants and heating plants. This study is based on this feature to update the emission factor database. Among them, Lei [14] and Zhao [15] explore the emission factors of process sources, including the cement and chemical industries that cannot be ignored. The biomass combustion source is also an important source of SO2. Li [16], Wang [17], Jetter [18], and Shen [19] have tested the emission characteristics of domestic household biomass stoves. The pollution emission factor of straw open burning is studied by Li [16]. As for the forest fires and grassland fires section, this study refers to the authoritative data of Andreae and Merlet [20].

3.1.2. NOX

There are many types of NOX, and the main causes of atmospheric pollution are NO and NO2. Combined with the measured data, the AP-42 database and the CORINAIR database can cover most of the NOX emission sources. But for the mobile source, the emission factor of the research in this study can be further localized. The emission factors of mobile sources are mostly from the researches by Liu [21], Zhang [22], Zhou [23], and Hu [24], who have measured the typical urban vehicles in China. Atmosphere 2020, 11, 739 5 of 19

Fu [25] has used the latest portable emission measurement system (PENS) to a large number of localized emission factors.

3.1.3. PM (PM10 and PM2.5) The anthropogenic sources of PM are mainly fuel combustion processes, various industrial process emissions, and vehicle emissions. For most sources, their emission factors are consistent with the sources above. Through research on pollutant emissions in Hunan, we find that the dust source also provides a certain contribution rate. This study directly refers to the Urban Air Pollutant Emission Inventory Weaving Technical Manual.

3.1.4. VOCs The VOCs emission factors for stationary fuel combustion sources are from Bo [26]. The solvent content of some products is limited by Chinese national standards, such as GB18581-2009 (wood coating), GB18582-2008 (interior wall coating), and HBC12-2003 (decorative adhesives), etc., and specific emission factors can be collected in these national standards. The studies by Wei [27] and Fu [28] can cover products outside national standards. Other sources of emissions refer to the categories of emissions that are not included in the above categories, especially the catering fume. After consulting the manual, the VOCs emission factor corresponding to the catering fume is 0.0095 g/(person·d).

3.1.5. NH3

Agricultural sources are the main sources of NH3 emissions. Most of the data come from the study by You [12] and also refers to some Chinese local research [22,29]. Waste treatment source is also a source of NH3 that cannot be ignored. The NH3 produced by the sewage treatment project uses the emission factor recommended by GBT32150. Landfill and incineration emission factors are based on the study by You [12]. The emission factor for flue gas denitration comes from the technical guidelines of the Ministry of Environmental Protection.

3.2. Air Pollutant Emissions Status Table1 summarizes the anthropogenic emission inventory for the 2015 CZT urban agglomeration. The total emissions of SO2, NOX, PM10, PM2.5, VOCs, and NH3 are 132.5, 148.9, 111.6, 56.5, 119.0, and 72.0 kt, respectively.

Table 1. The anthropogenic emission inventories of the Changzhutan urban agglomeration in 2015 (kt).

City SO2 NOX PM10 PM2.5 VOCs NH3 Changsha 28.7 57.5 54.3 23.2 49.9 32.4 Zhuzhou 55.0 39.9 24.4 15.3 44.9 21.0 Xiangtan 48.8 51.5 32.9 18.0 24.2 18.6 Total 132.5 148.9 111.6 56.5 119.0 72.0

Figure2 shows the emission contributions of 10 source categories in the CZT urban agglomeration. Stationary fuel combustion source and process source are the main sources of SO2, contributing 73.75% and 25.61%, respectively, to the total emissions. The largest contribution of NOX emissions is from the mobile source (39.45%) and stationary fuel combustion source (34.91%). It is worth mentioning that the car is the main source of nitrogen oxides and particulate matter on the road. Among them, the truck discharge contribution rate is twice that of the passenger car. Therefore, reducing the number of cars effectively plays a significant role in controlling NOX emission. In addition, thermoelectric production also contributes to the formation of NOX. The process source is the primary source of PM10 and PM2.5, contributing 35.12% to the emission, followed by the dust source (about 26.63%). When it comes to the industrial sectors, the emissions from building materials production and ferrous Atmosphere 2020, 11, 739 6 of 19 metal smelting take up nearly 76.86% of the total industrial emission. The process source and solvent source account for 30.17% and 24.10% to VOCs emissions. So far, an increasing number of researchers have begun to study VOCs, and the results show that VOCs are not only harmful to human health, but also have a significant impact on the Earth’s climate [30]. NH3 emission differs from other sources of pollution. The contribution of agricultural sources to NH3 emission is so large that other sources are nearly negligible. Among agricultural sources, NH3 emission from livestock and poultry accounts for 62.55%,Atmosphere and 2020 the, 11 rest, x FOR is basicallyPEER REVIEW from fertilizer usage (31.73%). 6 of 19

100%

80%

60%

40% percentage(%)

20%

0% SO₂ NOx PM₁₀ PM₂.₅ VOCs NH₃ Pollutant fossil fuel fixed combustion source process source mobile source solvent source agricultural source dust source biomass combustion source storage and transportation source waste disposal source other sources

FigureFigure 2. 2.The The emission emission contributions contributions of ten sourceof ten categories source incategories the Changzhutan in the urbanChangzhutan agglomeration. urban agglomeration. 3.3. Spatial-Temporal Distribution Of Air Pollutants 3.3. Spatial-TemporalWe divided the Distribution emission inventoryOf Air Pollutants grid into a 3 3 km with the application of the air × quality model. Furthermore, we elevated the resolution within the core areas of the CZT urban We divided the emission inventory grid into a 3 × 3 km with the application of the air quality agglomeration to a 1 1 km grid (see Figure3), where has a concentrated population and a high model. Furthermore,× we elevated the resolution within the core areas of the CZT urban level of industrial production. There are great spatial differences in the distribution of SO in CZT agglomeration to a 1 × 1 km grid (see Figure 3), where has a concentrated population and2 a high urban agglomeration. On the whole, the SO2 emission in Zhuzhou and Xiangtan is higher than that level of industrial production. There are great spatial differences in the distribution of SO2 in CZT in Changsha, which may be ascribed to the larger scale of industrial production in Zhuzhou and urban agglomeration. On the whole, the SO2 emission in Zhuzhou and Xiangtan is higher than that Xiangtan. Furthermore, NO emission is not only from industrial activity, but also from mobile sources. in Changsha, which may Xbe ascribed to the larger scale of industrial production in Zhuzhou and The emission of nitrogen oxides from urban central areas and major highways is significantly higher. Xiangtan. Furthermore, NOX emission is not only from industrial activity, but also from mobile In highly developed areas with booming industry and transportation, NO emission shows significant sources. The emission of nitrogen oxides from urban central areasX and major highways is linearity and polygonal distribution. Overall, the spatial distribution of PM2.5 and PM10 is consistent. significantly higher. In highly developed areas with booming industry and transportation, NOX Counties like Ningxiang, Changsha, and Liuyang owning a large number of manufactories will emit emission shows significant linearity and polygonal distribution. Overall, the spatial distribution of more particulate matter, resulting in large regional emission. In addition, the of PM2.5 and PM10 is consistent. Counties like Ningxiang, Changsha, and Liuyang owning a large particulate matter in the central area of CZT urban agglomeration is high due to a large number of number of manufactories will emit more particulate matter, resulting in large regional emission. In construction sites for infrastructure in urban areas. VOCs emission also shows an obvious spatial addition, the emission intensity of particulate matter in the central area of CZT urban distribution. For most regions, the VOCs emission in urban areas are higher than that in suburban agglomeration is high due to a large number of construction sites for infrastructure in urban areas. and rural areas. This phenomenon is likely related to the larger traffic flow and population in urban VOCs emission also shows an obvious spatial distribution. For most regions, the VOCs emission in areas. The usage of architectural coatings and decoration coatings are more frequent in urban areas, urban areas are higher than that in suburban and rural areas. This phenomenon is likely related to which can also contribute to the fugacity of VOCs. Other regions such as Liling, Xiangtan County the larger traffic flow and population in urban areas. The usage of architectural coatings and distribute ceramics, chemicals, and other VOCs emissions enterprises. In addition, thermoelectric decoration coatings are more frequent in urban areas, which can also contribute to the fugacity of production and other large industrial coal-fired sources also have a larger emission grid. Unlike other VOCs. Other regions such as Liling, Xiangtan County distribute ceramics, chemicals, and other sources, the distribution of NH emission areas is extremely uneven, and the large emission is mainly VOCs emissions enterprises. 3In addition, thermoelectric production and other large industrial concentrated in Ningxiang County, Zhuzhou County, and other regions. In these areas, people rely coal-fired sources also have a larger emission grid. Unlike other sources, the distribution of NH3 heavily on farming and livestock breeding for living, so the amount of livestock and poultry feeding, emission areas is extremely uneven, and the large emission is mainly concentrated in Ningxiang nitrogen fertilizer application is relatively large, causing a high emission of NH . County, Zhuzhou County, and other regions. In these areas, people rely heavily3 on farming and livestock breeding for living, so the amount of livestock and poultry feeding, nitrogen fertilizer application is relatively large, causing a high emission of NH3.

Atmosphere 2020, 11, 739 7 of 19 Atmosphere 2020, 11, x FOR PEER REVIEW 7 of 19

(a) (b) (c) (d)

(e) (f) (g) (h)

(i) (j) (k) (l)

Figure 3. Spatial allocation of air pollutant emissions and annual concentrations in the Changzhutan (CZT) urban agglomeration in 2015. (a) sulfur dioxide (SO ) emission (3 3 km). (b) SO emission (CZT) urban agglomeration in 2015. (a) sulfur dioxide (SO2)2 emission (3 × ×3 km). (b) SO2 emission2 (1 (1 1 km). (c) nitrogen oxides (NO ) emission (3 3 km). (d) NO emission (1 1km). (e) particulate × 1× km). (c) nitrogen oxides (NOX)X emission (3 × ×3 km). (d) NOXX emission (1 ×× 1km). (e) particulate matter 2.5 (PM2.5) emission (3 3 km). (f) PM2.5 emission (1 1 km). (g) particulate matter 10 (PM10) matter 2.5 (PM2.5) emission (3 × 3 km). (f) PM2.5 emission (1 × 1 km). (g) particulate matter 10 (PM10) emission (3 3 km). (h) PM10 emission (1 1 km). (i) volatile organic compounds (VOCs) emission emission (3 × 3 km). (h) PM10 emission (1 × ×1 km). (i) volatile organic compounds (VOCs) emission (3 (3 3 km). (j) VOCs emission (1 1 km). (k) ammonia (NH3) emission (3 3 km). (l) NH3 emission × 3× km). (j) VOCs emission (1 × 1× km). (k) ammonia (NH3) emission (3 × 3 ×km). (l) NH3 emission (1 × (1 1 km). 1 km).× Figure4 illustrates the monthly changes in the emission of major air pollutants in the CZT urban Figure 4 illustrates the monthly changes in the emission of major air pollutants in the CZT agglomeration. It can be seen from the figure that the monthly trend of pollutant discharge is relatively urban agglomeration. It can be seen from the figure that the monthly trend of pollutant discharge is stable overall. Nevertheless, the monthly discharge of some pollutant’s peaks in autumn and winter. relatively stable overall. Nevertheless, the monthly discharge of some pollutant’s peaks in autumn Among them, the concentration of SO2 reaches the highest in December, and NO2 emission reaches and winter. Among them, the concentration of SO2 reaches the highest in December, and NO2 two peaks during July–August and November–December throughout the year. VOCs are found to emission reaches two peaks during July–August and November–December throughout the year. peak in October, and the concentration of PM (PM10, PM2.5) rose rapidly in November. Generally, VOCs are found to peak in October, and the concentration of PM (PM10, PM2.5) rose rapidly in it is obvious that multiple pollutants peak in the autumn or winter. This may be caused by the November. Generally, it is obvious that multiple pollutants peak in the autumn or winter. This may following three points: firstly, heating in winter leads to the increase of coal consumption and the be caused by the following three points: firstly, heating in winter leads to the increase of coal emission of a variety of air pollutants; secondly, the weather conditions in this period are not conducive consumption and the emission of a variety of air pollutants; secondly, the weather conditions in this to the spread of pollutants, resulting in constant deposition of pollutants; thirdly, the CZT urban period are not conducive to the spread of pollutants, resulting in constant deposition of pollutants; thirdly, the CZT urban agglomeration is surrounded by mountains on three sides. Atmospheric

Atmosphere 2020, 11, 739 8 of 19 Atmosphere 2020, 11, x FOR PEER REVIEW 8 of 19 agglomerationpollutants in the is surrounded north migrate by mountains to the CZT on urban three sides.agglomeration Atmospheric under pollutants the influence in the north of the migrate winter tomonsoon. the CZT urbanThe mountainous agglomeration areas under block the influence the air ofma thess winterand limit monsoon. the diff Theusion mountainous of atmospheric areas blockpollutants. the air The mass accumulation and limit the diofff atmosphericusion of atmospheric pollutants pollutants. in the CZT The urban accumulation agglomeration of atmospheric causes a pollutantspeak value. in On the the CZT contrary, urban agglomeration the peak of NH causes3 emission a peak appears value. in On the the June–August, contrary, the which peak probably of NH3 emissionrelates to appears a sharpin rise the in June–August, temperatures which in June. probably During relatesthis time, to a high sharp temperatures rise in temperatures will promote in June. the Duringbreeding this of time, livestock high temperatures and increase will the promoteevaporation the breeding of NH3 ofin livestocklivestockand and increase poultry, the as evaporation observed in ofthe NH study3 in livestock of Wang and [31]. poultry, Besides, as ni observedtrogen fertilizer in the study will of be Wang widely [31 ].applied Besides, for nitrogen crop planting, fertilizer which will becan widely also contribute applied for to crop the emission planting, of which NH3. can also contribute to the emission of NH3.

24,000 22,000 20,000 18,000 16,000 14,000 12,000 10,000

concentration(kg) 8,000 6,000 4,000 2,000 0 123456789101112 month SO₂ NOx PM₁₀ PM₂.₅ VOCs NH₃

Figure 4. The emission contributions of ten source categories in the Changzhutan urban agglomeration. TheFigure monthly 4. The changes emission in emissions contributions of major airof pollutantsten source in thecategories CZT urban in agglomeration.the Changzhutan urban agglomeration. The monthly changes in emissions of major air pollutants in the CZT urban 3.4. Sourceagglomeration. Apportionment of PM2.5

3.4.1.3.4. Source Chemical Apportionment Speciation of and PM Source2.5 Apportionment

The composition of PM2.5 in three regions are determined based on the local measurement. There3.4.1. areChemical 23 kinds Speciation of components and Source found Apportionment in the particles as shown in Table S4. On the whole, there is little diThefference composition in the content of PM of2.5 the in mainthreecomponents regions are indetermined three cities. based The major on the species local ofmeasurement. the PM2.5 in 2 CZTThere urban are 23 agglomeration kinds of components in 2015 are found SO4 in−, OC,the part andicles NO 3as−, shown with the in annual Table S4. average On the concentration whole, there reachingis little difference 13.06, 8.24, in and the 4.84 contentµg/m 3of, respectively.the main co Zhuzhoumponents has in three the largest cities. industrial The major area species in the of CZT the urbanPM2.5 agglomeration,in CZT urban soagglomeration it has more SO in2 and2015 NOareX SOemission42₋, OC, from andindustrial NO3-, with coal-burning the annual than average the otherconcentration two cities. reaching The reason 13.06, may 8.24, be dueandto 4.84 the μ conversiong/m3, respectively. of SO2 and Zhuzhou photochemical has the reactionlargest industrial to NOX, andarea then in the formationCZT urban of agglomeration, sulfate and nitrate so [it32 has], which more has SO a2 greatand influenceNOX emission on the from concentration industrial ofcoal-burning PM2.5 in the than three the cities. other Organic two cities. carbon(OC) The reason and elemental may be carbon(EC)due to the areconversion the major of emissions SO2 and inphotochemical the transportation reaction sector, to NO andX,the and number then the of formation motor vehicles of sulfate in Changsha and nitrate is much[32], which larger has than a great that ofinfluence Zhuzhou on and the Xiangtan concentration [12], accounting of PM2.5 in for the about three 50.75% cities. to Organi the totalc carbon(OC) vehicles of theand three elemental cities. Therefore,carbon(EC) the are percentages the major emissions of OC and in EC the are transpor highertation in Changsha sector, and rather the thannumber the of other motor two vehicles cities. Whatin Changsha cannot beis much ignored larger is that than the that crustal of Zhuzhou elements and (Si, Xiangtan Al, Ca, [12], Fe, accounting Mg, Mn, Ti, for V, about etc.) 50.75% occupy to + 16.13%,the total which vehicles is also of an the important three compositioncities. Therefore, of PM the2.5.K percentagesand Cl− are of mainly OC and derived EC fromare higher biomass in burningChangsha [33 ],rather and they than combine the other with two dust cities. generated What bycann constructionot be ignored and is transportation that the crustal activities elements [34] (Si, in + theAl, air. Ca, There Fe, Mg, is little Mn, di Ti,fference V, etc.) in occupy the concentrations 16.13%, which of K is andalso Clan− importantamong three composition cities, suggesting of PM2.5 that. K+ + theand CZT Cl- urbanare mainly agglomeration derived from may havebiomass a similar burning source [33], of and K andthey Cl combine−. However, with the dust content generated of crust by elementsconstruction in Xiangtan and transportation is always higher activities than that [34] in the in other the twoair. cities.There This is maylittle be difference explained byin thethe presenceconcentrations of more of lead K+ and zincCl- among smelting three enterprises cities, suggesting near the Xiangtan that the samplingCZT urban point. agglomeration Figure5 shows may thehave components a similar source of PM2.5 ofin K di+ andfferent Cl-. cities. However, the content of crust elements in Xiangtan is always higher than that in the other two cities. This may be explained by the presence of more lead and zinc smelting enterprises near the Xiangtan sampling point. Figure 5 shows the components of PM2.5 in different cities.

Atmosphere 2020, 11, 739 9 of 19 Atmosphere 2020, 11, x FOR PEER REVIEW 9 of 19

(a) Xiangtan

Zhuzhou City

Changsha

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% percentage(%)

OC EC Cl⁻ NO₃⁻ SO₄²⁻ NH₄⁺ Na Mg Al Si K Ca Fe Ti V Cr Mn Ni Cu Zn Pb As Cd

16 (b) 14

12 ) 3

g/m 10 μ

8

6

Contribution( 4

2

0 OC EC Cl⁻ NO₃⁻ SO₄²⁻ NH₄⁺ Na Mg Al Si K components

Changsha Zhuzhou Xiangtan

FigureFigure 5. PM5. PM2.5 2.5speciation speciation by by cities. cities. (a )(a The) The percentage percentage of of 23 23 components components of PMof PM2.5 2.5in in di ffdifferenterent regions. regions. 3 µ 3 (b)( PMb) PM2.5 2.5main main component component concentration concentration in in di ffdifferenterent regions regions ( g(μ/mg/m). ).

Figure6 shows the average content of major chemical components in PM in different seasons. Figure 6 shows the average content of major chemical components in PM2.5 2.5 in different seasons. SO 2 has the highest proportion in four seasons, with the value reaching 18.43%. Five components SO4 −42- has the highest proportion in four seasons, with the value reaching 18.43%. Five components (including SO 2 , NO , NH +, OC, and EC) are attributed to the increase of PM concentration in (including SO4 −42-, NO3−3-, NH44+, OC, and EC) are attributed to the increase of PM2.5 2.5 concentration in winter, with the most dominant increase in NO and NH +. Coal-fired emissions from industrial winter, with the most dominant increase in 3NO− 3- and NH4 4+. Coal-fired emissions from industrial powerpower generation generation and and industrial industrial production production have have important important contributions contributions to OC to and OCEC and [35 EC]. At[35]. the At samethe time,same CZTtime, urban CZT urban agglomeration agglomeration has more has coldmore air cold activity air activity in winter, in winter, which which is beneficial is beneficial to the to inputthe ofinput exogenous of exogenous pollutants pollutants [36]. The [36]. concentration The concen oftration OC and of OC EC and are theEClowest are the in lowest summer, in summer, which is mainlywhich is due mainly to the due abundant to the abundant precipitation precipitation in summer. in The summer. wash-out The ewash-outffect of rain effect plays of a rain vital plays role a + invital eliminating role in theeliminating atmospheric the particulateatmospheric matter. particulate The concentrationmatter. The concentration of K and Cl −ofin K autumn+ and Cl is- in significantlyautumn is higher,significantly and thehigher, concentrations and the concentrations are 1.9 times are and 1.9 2.2 times times and than 2.2 the times average than ofthe spring, average summer,of spring, and summer, winter, respectively. and winter, Combined respectively. with Combined the analysis with of chemicalthe analysis composition of chemical characteristics composition ofcharacteristics pollution sources of [pollution33], the content sources of [33], these the two content components of these in biomass two components combustion sourcesin biomass is significantlycombustion higher sources than is that significantly of other sources, higher which than indicates that of thatother the sources, influence which of biomass indicates combustion that the on PM in autumn CZT region is strengthened. influence2.5 of biomass combustion on PM2.5 in autumn CZT region is strengthened.

Atmosphere 2020, 11, 739 10 of 19 Atmosphere 2020, 11, x FOR PEER REVIEW 10 of 19

(a) Spring

Summer season Autumn

Winter

0% 20% 40% 60% 80% 100% percentage(%)

OC EC Cl⁻ NO₃⁻ SO₄²⁻ NH₄⁺ Na Mg Al Si K Ca Fe Ti V Cr Mn Ni Cu Zn Pb As Cd

(b) 18 16

14 ) 3 12 g/m μ 10

8

6 Contribution( 4

2

0 OC EC Cl⁻ NO₃⁻ SO₄²⁻ NH₄⁺ Na Mg Al Si K component

Winter Autumn Summer Winter

Figure 6. PM compositions in different seasons. (a) The percentage of 23 components of PM in Figure 6. PM2.52.5 compositions in different seasons. (a) The percentage of 23 components of PM2.52.5 in 3 different seasons. (b) PM2.5 main component concentration in different seasons (µg/m ).3 different seasons. (b) PM2.5 main component concentration in different seasons (μg/m ). Based on the chemical composition data above, the source apportionment of the regional fine Based on the chemical composition data above, the source apportionment of the regional fine particulate matter is carried out by a comprehensive analysis method (combining the CMB model, PMF particulate matter is carried out by a comprehensive analysis method (combining the CMB model, model, and this emission inventory). The difference between comprehensive analysis results of PM PMF model, and this emission inventory). The difference between comprehensive analysis results2.5 in the three cities is obvious, as seen in Figure7. The comprehensive contribution rate of vehicle exhaust of PM2.5 in the three cities is obvious, as seen in Figure 7. The comprehensive contribution rate of is 25% in Changsha, while that of Zhuzhou and Xiangtan is less than 15%. It proves that automobile vehicle exhaust is 25% in Changsha, while that of Zhuzhou and Xiangtan is less than 15%. It proves exhaust is still a source of pollution worthy of attention. In addition, the effect of industrial emissions that automobile exhaust is still a source of pollution worthy of attention. In addition, the effect of from Zhuzhou and Xiangtan is prominent. The overall contribution rate of industrial emissions to industrial emissions from Zhuzhou and Xiangtan is prominent. The overall contribution rate of PM2.5 from Xiangtan and Zhuzhou is about 40%, while that of Changsha is 30%. Among industrial industrial emissions to PM2.5 from Xiangtan and Zhuzhou is about 40%, while that of Changsha is emission sources in Changsha, the highest contribution is from building materials production (8%), 30%. Among industrial emission sources in Changsha, the highest contribution is from building followed by thermoelectric production (7%). Zhuzhou also has the largest contribution of industrial materials production (8%), followed by thermoelectric production (7%). Zhuzhou also has the emission from building materials production (14%), followed by the metallurgical industry and largest contribution of industrial emission from building materials production (14%), followed by chemical industry, both contributing about 10%. The proportion of SO , NO , and VOCs emitted the metallurgical industry and chemical industry, both contributing about2 10%.X The proportion of by Xiangtan’s metallurgical industry accounts for 30–40% of total emissions, which indicates a large SO2, NOX, and VOCs emitted by Xiangtan`s metallurgical industry accounts for 30–40% of total contribution to PM2.5 pollution in Xiangtan. emissions, which indicates a large contribution to PM2.5 pollution in Xiangtan.

AtmosphereAtmosphere2020 2020, 11, 11, 739, x FOR PEER REVIEW 1111 of of 19 19

(a) 6%5% 8% 8% 4% 6% 30% 9% 25% 7% 4% 18%

raise dust vehicle exhaust catering fume biomass burning civil fuel others chemical Industry thermoelectric production other industry metallurgy building materials production

(b)

7% 8% 8% 10% 5% 2% 39% 14% 14% 4% 9% 19% raise dust vehicle exhaust catering fume biomass burning civil fuel others chemical Industry thermoelectric production other industry metallurgy building materials production (c)

6% 6% 8% 1% 4% 18% 13% 40% 10% 5% 6% 23%

raise dust vehicle exhaust catering fume biomass burning civil fuel others chemical Industry thermoelectric production other industry metallurgy building materials production

FigureFigure 7. 7.Analytical Analytical results results of of PM2.5 PM2.5 comprehensive comprehensive source source in in Changsha, Changsha, Zhuzhou, Zhuzhou, and and Xiangtan. Xiangtan. (a()a Changsha.) Changsha. (b ()b Zhuzhou.) Zhuzhou. (c ()c Xiangtan.) Xiangtan.

SeasonalSeasonal variation variation has has been been observed observed in in the the contribution contribution rate rate of of di differentfferent PM PM2.52.5sources sources in in CZT CZT urbanurban agglomeration agglomeration as showedas showed in Figure in Figure8. The 8. contribution The contribution of secondary of secondary sulfate and sulfate vehicle and exhaust vehicle rankexhaust first andrank second first and in eachsecond season. in each The season. secondary The sulfatesecondary contribution sulfate contribution rate is the highest rate is inthe autumn highest (24%),in autumn and the (24%), vehicle and exhaust the vehicle is the exhaust highest is in the summer highest (17%). in summer Since (17%). industrial Since source industrial emissions source contributeemissions the contribute most to the the secondary most to sulfate the secondary concentration sulfate in PM concentration2.5, the impact in of PM industrial2.5, the emissionsimpact of continuesindustrial to emissions be of concern. continues The contribution to be of concern. rate of secondaryThe contribution organic rate carbon of secondary is the highest organic in summer, carbon accountingis the highest for about in summer, 13%. Raise accounting dust and for building about 13 cement%. Raise dust dust have and a relatively building high cement contribution dust have in a springrelatively and summer,high contribution which is in 2–4% spring higher and than summer, autumn which and is winter. 2–4% higher In the casethan ofautumn significantly and winter. less precipitationIn the case of during significantly the autumn less precipitation and winter seasons, during the the autumn reason for and this winter phenomenon seasons, the may reason be that for this phenomenon may be that the construction time in spring and summer is longer than that in autumn and winter. In addition, the relative contribution rate of coal-fired dust in different seasons

Atmosphere 2020, 11, 739 12 of 19

Atmospherethe construction 2020, 11, x timeFOR PEER in spring REVIEW and summer is longer than that in autumn and winter. In addition,12 of 19 the relative contribution rate of coal-fired dust in different seasons is between 8% and 13%, which is isrelatively between stable. 8% Theand contribution13%, which of is catering relatively fume stable. and metallurgical The contribution dust in fourof seasonscatering is fume very small,and metallurgicalbut it cannot bedust ignored. in four seasons is very small, but it cannot be ignored.

Spring Summer 5% 8% 13% 14% 6% 13%

8% 8% 14%

11% 17% 9%

7% 4% 5% 3% 4% 6% 3% 4% 21% 19% raise dust building cement dust raise dust building cement dust coal-fired dust catering fume coal-fired dust catering fume biomass burning metallurgical dust biomass burning metallurgical dust secondary sulfate secondary nitrate secondary sulfate secondary nitrate vehicle exhaust secondary organic carbon vehicle exhaust secondary organic carbon others others

5% Winter Autumn 10% 12% 11% 11% 6% 5% 6% 9% 13% 13% 14% 5%

7% 7% 6% 2% 3% 10% 3%

24% 19% raise dust building cement dust raise dust building cement dust coal-fired dust catering fume coal-fired dust catering fume biomass burning metallurgical dust biomass burning metallurgical dust secondary sulfate secondary nitrate secondary sulfate secondary nitrate vehicle exhaust secondary organic carbon vehicle exhaust secondary organic carbon others others

FigureFigure 8. 8. PMPM2.52.5 sourcesource contribution contribution rate rate in in different different seasons. seasons.

3.4.2.3.4.2. Regional Regional Transmission Transmission Characteristics Characteristics and and Potential Potential Source Source Domain Domain Analysis TheThe pollution levellevel ofof local local atmospheric atmospheric pollutants pollutants is a ffisected affected by many by many factors. factors. Under Under the control the controlof different of different air masses, air masses, the concentration the concentration of pollutants of pollutants varies greatlyvaries greatly [37]. The [37]. clustering The clustering of airflow of airflowsources sources after 72 after h in four72 h in seasons four seasons is illustrated is illustrated in Figure in9 Figure. The CZT 9. The urban CZT agglomeration urban agglomeration is mainly is mainlyaffected affected by the medium-distanceby the medium-distance transmission transmissi air masses.on air masses. The source The source and transmission and transmission path ofpath air ofmasses air masses are greatly are greatly affected affected by the by season. the season. In spring, In spring, the eastern the eastern air mass air originated mass originated from the from central the centralpart of Jiangxipart of accountsJiangxi foraccounts 63% of for the 63% airflow of transfer;the airflow 20% transfer; are attributed 20% toare air attributed mass moving to air north mass by movingpassing north through by Henanpassing and through Hubei, Henan and 17% and of Hubei, the airflow and transfer17% of the is determined airflow transfer by the is southwestern determined byair the mass southwestern passing Guangdong. air mass Inpassing summer, Guangdon southwesterng. In airsummer, masses southwestern from Guangxi air and masses Guangdong from Guangxicontrols 87%and ofGuangdong the airflow controls transfer 87% while of 13% the leftairflow is aff ectedtransfer by while the northeastern 13% left is airaffected mass passingby the northeasternthrough Zhejiang, air mass southern passing Anhui, through and northernZhejiang Jiangxi., southern The Anhui, northeastern and northern air mass, whichJiangxi. passes The northeasternthrough Jiangsu, air mass, Anhui, which and Hubei, passes determines through Jian 92%gsu, of the Anhui, airflow and transfer Hubei, in determines autumn, and 92% the 8%of the left airflowis impacted transfer by the in southwestern autumn, and air the mass 8% passing left is throughimpacted Guangdong. by the southwestern Variation of airair massmass in passing transfer throughdistance Guangdong. and direction Variation is observed of air in mass winter. in transfer The southwestern distance and air direction mass passing is observed through in winter. Anhui, Thesouthern southwestern Hubei, and air northernmass passing Jiangxi through controls Anhui, 72% ofsouthern the airflow Hubei, transfer. and northern Air mass Jiangxi from southerncontrols 72%Guangdong of the airflow affects transfer. 20% of Air the mass airflow from transfer southern while Guangdong the 8% left affects is impacted 20% of bythe air airflow mass transfer passing while the 8% left is impacted by air mass passing through Inner Mongolia, Shanxi, Henan, and Hubei. The relatively short airflow trajectory indicates that the atmospheric environment is relatively stable and prone to cause the accumulation of pollutants. Judging from the transmission

Atmosphere 2020, 11, 739 13 of 19 through Inner Mongolia, Shanxi, Henan, and Hubei. The relatively short airflow trajectory indicates thatAtmosphere the atmospheric 2020, 11, x FOR environment PEER REVIEW isrelatively stable and prone to cause the accumulation of pollutants.13 of 19 Judging from the transmission distance to the air mass after 72 h in each season, the main airflow distanceAtmosphere to 2020 the, 11 air, x FOR mass PEER after REVIEW 72 h in each season, the main airflow trajectory is longer in13 summer of 19 trajectoryand autumn, is longer which in summer is not conducive and autumn, to the which accumulation is not conducive of pollutants. to the accumulation of pollutants. distance to the air mass after 72 h in each season, the main airflow trajectory is longer in summer and autumn, which is not conducive to the accumulation of pollutants.

Spring Summer Spring Summer

Autumn Winter Autumn Winter

FigureFigure 9. 9.Airflow AirflowAirflow trajectorytrajectory trajectory clusteri clustering clusteringng in inin different didifferentfferent seasons. seasons.

PSCFPSCFPSCF and andand CWT CWTCWT are are two two gridding gridding statistical statisticalstatistical analysis analysis analysis methods methods methods basedbased based onon on thethe the HYSPLITHYSPLIT HYSPLIT model model model and wereandand applied were were applied inapplied semi-quantifying inin semi-quantifyingsemi-quantifying the distribution thethe distribution distribution of pollutant of of pollutant pollutant sources sources sources in in this thisin study.this study. study. TheThe The study study area areaarea was was divided divided intointo a grid of 0.5°0.5° ×× 0.5°.0.5°. was divided into a grid of 0.5◦ 0.5◦. PMPM2.52.5 potential potential sourcesource contribution× distribution distribution is is shown shown in inFigure Figure 10. 10. The The threshold threshold of PM of 2.5PM is2.5 is PM potential source contribution distribution is shown in Figure 10. The threshold of PM is 2.5 3 2.5 3535 μ μg/mg/m3, , andand thethe resultresult of PSCFPSCF showsshows thatthat PM PM2.52.5 pollution pollution in inCZT CZT urban urban agglomeration agglomeration has hasa a 35 µg/m3, and the result of PSCF shows that PM pollution in CZT urban agglomeration has a similar similarsimilar pollutant pollutant source,source, mainly originatedoriginated from from2.5 Jiangsu, Jiangsu, Anhui, Anhui, southern southern Hubei, Hubei, and and northwestern northwestern pollutant source, mainly originated from Jiangsu, Anhui, southern Hubei, and northwestern Jiangxi. Jiangxi.Jiangxi.

2.5 FigureFigure 10. 10.PM PM2.5 potentialpotential source source contribution contribution distribution. distribution. Figure 10. PM2.5 potential source contribution distribution.

Atmosphere 2020, 11, x FOR PEER REVIEW 14 of 19

The PM2.5 trajectory weight concentration distribution is shown in Figure 11. The larger Cij value indicates that the air mass passing through the grid i × j will cause a higher concentration of pollutants in the target area. The area corresponding to the grid is the main outer source area that contributes to the high-concentration pollutants in the target area. The trajectory through the grid is the main transmission path that contributes to the pollutants. As the main Atmospherepotential2020 contributing, 11, 739 area in CZT urban agglomeration, northern areas such as Jiangsu, southern14 of 19 Hubei, and northwestern Jiangxi appear higher value. Besides, the southern regions such as southern Guangdong and southern Guangxi present high contribution rates, which cannot be ignored.The PM2.5 trajectory weight concentration distribution is shown in Figure 11.

FigureFigure 11.11. PMPM2.52.5 concentrationconcentration weight trajectory distribution.

3.5. ComparisonThe larger CWithvalue Other indicates Inventories that the air mass passing through the grid i j will cause a higher ij × concentration of pollutants in the target area. The area corresponding to the grid is the main outer source To better understand the emission levels of CZT urban agglomeration, we have compared our area that contributes to the high-concentration pollutants in the target area. The trajectory through the inventory with the inventory of the Yangtze River Delta (YRD) region and the Beijing–Tianjin– grid is the main transmission path that contributes to the pollutants. As the main potential contributing Hebei (BTH) region. The results are shown in Figure 12. The emission data came from the 2010 YRD area in CZT urban agglomeration, northern areas such as Jiangsu, southern Hubei, and northwestern region Fu, Wang, Zhao, Xing, Cheng, Liu, and Hao [28] and the 2013 BTH region [38] emission Jiangxi appear higher value. Besides, the southern regions such as southern Guangdong and southern inventory. These two regions are not only geographically larger but also have larger emissions than Guangxi present high contribution rates, which cannot be ignored. the CZT urban agglomeration, so it will be less precise to directly compare the emission status 3.5.between Comparison CZT withurban Other agglomeration Inventories and other two regions. Therefore, we allocated the air pollutants in the area unit, population, and GDP to better evaluate the emission status in different To better understand the emission levels of CZT urban agglomeration, we have compared our regions. The SO2 emission per unit area of the YRD region and BTH region is about twice that of the inventoryCZT urban with agglomeration, the inventory and of thethe Yangtze per capita River em Deltaission (YRD) and per region unit and GDP the are Beijing–Tianjin–Hebei roughly twice and (BTH)four times region. that The of resultsthe CZT are urban shown agglomeration. in Figure 12. The The emission first reason data came can be from that the central 2010 YRD heating region is Fu, Wang, Zhao, Xing, Cheng, Liu, and Hao [28] and the 2013 BTH region [38] emission inventory. required in the northern region, and the fuel consumption process generates a large amount of SO2; Thesethe second tworegions is that arethe notYRD only region geographically and BTH regi largeron butare alsoaffiliated have largerto industrial emissions provinces than the which CZT urban agglomeration, so it will be less precise to directly compare the emission status between CZT consumes a lot of energy and results in more SO2. For NOX emission, the values per unit area urbanemissions agglomeration of the YRD and region other and two BTH regions. region Therefore, are 2.13we times allocated and 2.65 the times air pollutants higher than in the that area of unit, the population,CZT urban andagglomeration GDP to better respectively. evaluate the This emission phenomenon status in dimayfferent be regions. attributed The to SO 2theemission dense pertransportation unit area of hub the YRDand regiona largeand number BTH regionof moto isr aboutvehicle twice holdings that of in the the CZT YRD urban region. agglomeration, Combined andwith the per per capita capita emission emission and and per per unit GDPGDP, are it roughlyis not difficult twice and to speculate four times that that ofpollution the CZT control urban agglomeration.measures for motor The firstvehicles reason in canthe beYRD that region central have heating significant is required effects, in the which northern are worth region, learning and the fuel consumption process generates a large amount of SO ; the second is that the YRD region and BTH from the CZT urban agglomeration. The total amount2 of PM (PM10, PM2.5) in the CZT urban regionagglomeration are affiliated and the to industrial BTH region provinces is quite whichdifferent, consumes but the a emission lot of energy per unit and area results is consistent. in more SO It2. Formay NO beX attributedemission, to the the values fact perthat unit both area regions emissions belong of to the inland YRD region areas and BTHparticulate region matter are 2.13 can times be and 2.65 times higher than that of the CZT urban agglomeration respectively. This phenomenon may be attributed to the dense transportation hub and a large number of motor vehicle holdings in the YRD region. Combined with per capita emission and per unit GDP, it is not difficult to speculate that pollution control measures for motor vehicles in the YRD region have significant effects, which are worth learning from the CZT urban agglomeration. The total amount of PM (PM10, PM2.5) in the CZT urban agglomeration and the BTH region is quite different, but the emission per unit area is consistent. Atmosphere 2020, 11, x FOR PEER REVIEW 15 of 19 Atmosphere 2020, 11, 739 15 of 19 capable of accumulating better. Except for the East China Sea, the YRD region is highly affected by sea Itwind, may befacilitating attributed the to the transport fact that bothof particulate regions belong matter. to inland There areas are andsignificant particulate differences matter can in be VOCs emissioncapable among of accumulating the three better. regions. Except There for the are East fe Chinawer Sea,VOCs the YRDemission region enterprises is highly affected in CZT by sea urban agglomeration.wind, facilitating The the other transport two ofregions, particulate acting matter. as Therethe Chinese are significant economic differences backbone in VOCs have emission continual industrialamong activities the three regions. and cause There large are feweremissions. VOCs Compared emission enterprises with the indata CZT of urban INTEX-B agglomeration. [39] in Hunan Province,The other the two VOCs regions, emission acting per as theunit Chinese area of economic this study backbone are 13.99 have times continual higher industrial than that activities of INTEX-B, and cause large emissions. Compared with the data of INTEX-B [39] in Hunan Province, the VOCs which shows that the control measures of VOCs in Hunan Province have yet to be strengthened. It emission per unit area of this study are 13.99 times higher than that of INTEX-B, which shows that the is difficultcontrol measuresto acquire of VOCsthe activity in Hunan data Province from havesolvent yet touse be sources strengthened. collected It is didirectlyfficult to at acquire the city the level, andactivity a lack dataof accurate from solvent data usesources sources may collected also be directly responsible at the city for level, differentiation. and a lack of It accurate is worth data noting 3 thatsources the NH may emission also be responsible in the BTH for region differentiation. is much It larger is worth than noting that that in thethe NH CZT3 emission urban agglomeration, in the BTH but regionthey are is much equal larger in NH than3 emissions that in the per CZT unit urban area agglomeration, and per capita but theyemissions. are equal The in NHreason3 emissions is that CZT urbanper agglomeration unit area and per is capita located emissions. near Dongting The reason Lake, is that known CZT urban as the agglomeration "land of fish is located and rice". near NH3 emissionDongting caused Lake, by known nitrogen as the “landfertilizer of fish and and livestock rice”. NH and3 emission poultry caused in this by nitrogen area is fertilizerhigher andthan the averagelivestock value. and poultry in this area is higher than the average value.

(a) 20 18 16 ) 3 14 g/m

μ 12 10 8 6 Contribution( 4 2 0 SO₂ NOx PM₁₀ PM₂.₅ VOCs NH₃ Pollutant the Yangtze River Delta region the Beijing-Tianjin-Hebei region the Changzhutan urban agglomerations

(b) 30

25 ) 3 20 g/m μ

15

10 Contribution( 5

0 SO₂ NOx PM₁₀ PM₂.₅ VOCs NH₃ Pollutant the Yangtze River Delta region the Beijing-Tianjin-Hebei region the Changzhutan urban agglomerations

Figure 12. Cont.

Atmosphere 2020, 11, 739 16 of 19 Atmosphere 2020, 11, x FOR PEER REVIEW 16 of 19

(c) 500 450 400 ) 3 350 g/m

μ 300 250 200 150 Contribution( 100 50 0 SO₂ NOx PM₁₀ PM₂.₅ VOCs NH₃ Pollutant the Yangtze River Delta region the Beijing-Tianjin-Hebei region the Changzhutan urban agglomerations

Figure 12. Comparison of air pollutant emission inventory established by different studies [27,38]. Figure 12. Comparison of air pollutant emission inventory established by different studies [27,38]. (a) Emissions of atmospheric pollutants per unit area (kg/m2). (b) Emissions of atmospheric pollutants (a) Emissions of atmospheric pollutants per unit area (kg/m2). (b) Emissions of atmospheric per capita emissions (kg/person). (c) Emissions of atmospheric pollutants per unit GDP (kg/billion). pollutants per capita emissions (kg/person). (c) Emissions of atmospheric pollutants per unit GDP 3.6.(kg/billion). Uncertainty Analysis This emission inventory is estimated using the Monte Carlo method, which has been widely 3.6.applied Uncertainty for calculating Analysis the uncertainty of emission inventory. As illustrated in Table S5, the total uncertaintyThis emission for the inventory anthropogenic is estimated emission using of NO thX,e SO Monte2, NH 3Carlo, VOCs, method, PM10, and which PM2.5 hasare been 20% towidely 42%, 34% to 58%, 42% to 89%, 46% to 73%, 42% to 67%, 38% to 69%, respectively. Among them, applied −for calculating− the uncertainty− of emission− inventory.− As illustrated in Table S5, the total NH3 and VOCs have higher uncertainty. The low uncertainty of NH3 from livestock breeding and uncertainty for the anthropogenic emission of NOX, SO2, NH3, VOCs, PM10, and PM2.5 are 20% to agricultural system ( 51% to 42% and 43% to 47% respectively) contributes to relatively accurate − − 42%,emission −34% data,to 58%, while −42% the uncertaintyto 89%, −46% value to is large73%, for −42% biomass to 67%, burning −38% ( 82% to to69%, 125%) respectively. ascribed to theAmong − them,lack NH of local3 and data. VOCs For have VOCs, higher high uncertaintiesuncertainty. areThe contributed low uncertainty by the unavailability of NH3 from of livestock specific data breeding andfrom agricultural non-road mobilesystem sources (−51% as wellto 42% as residential and −43% fuel to sources. 47% Moreover,respectively) PM2.5 contributesand PM10 also to sharerelatively accuratea high emission uncertainty, data, which while mainly the results uncertainty from the non-roadvalue is mobilelarge for source biomass and dust burning ( 40% to(− 117%82% andto 125%) − ascribed60% to 121%the lack respectively). of local data. Lower For uncertainty VOCs, high is found uncertainties in SO and are NO contributeddue to the by less the complicated unavailability − 2 X of specificcomposition data and from more non-road effective mobile control measuressources as than well other as pollutants.residential fuel sources. Moreover, PM2.5 and PM10 also share a high uncertainty, which mainly results from the non-road mobile source and 4. Discussion dust (-40% to 117% and -60% to 121% respectively). Lower uncertainty is found in SO2 and NOX due to the lessChangsha complicated should composition emphasize on and controlling more effective the number control of measures the vehicle than in reducing other pollutants. pollutant emissions. Various measures can effectively reduce air pollution, such as accelerating the phase-out of 4. Discussionused cars, increasing fuel quality, and upgrading vehicle emission standards. For heavily polluting diesel vehicles, exhaust gas recirculation and lean-burn NOX traps are recommended to reduce pollutantChangsha emissions. should Theemphasize NOX emission on controlling in Xiangtan the is number significantly of the higher vehicle than in that reducing in Changsha pollutant emissions.and Zhuzhou. Various Therefore, measures restrictingcan effectively industries reduce dominated air pollution, by metallurgy such as a isccelerating an urgent the issue phase-out in of usedXiangtan. cars, Xiangtan increasing should fuel consider quality, rebuilding and upgrading the industrial vehicle structure em byission reducing standards. the production For heavily pollutingcapacity diesel of steel vehicles, and strengthening exhaust gas the recirculation utilization eandfficiency lean-burn of waste NO heat.X traps Moreover, are recommended it is also to necessary to accelerate the construction of denitrification equipment in the metallurgical industry, reduce pollutant emissions. The NOX emission in Xiangtan is significantly higher than that in Changshasuch as theand selective Zhuzhou. non-catalytic Therefore, reduction restricting method industries and the dominated selective catalytic by metallurgy reduction method.is an urgent In addition, SO emission in Zhuzhou accounts for 41.67% of total emissions. Industrial boilers should issue in Xiangtan.2 Xiangtan should consider rebuilding the industrial structure by reducing the be a key factor for managers to consider. Not only do industrial boiler emission standards need to be production capacity of steel and strengthening the utilization efficiency of waste heat. Moreover, it improved to control SO2, but the existing desulfurization facilities should be upgraded to the limestone is also(lime) necessary/gypsum process. to accelerate Film-coated the bagconstruction filters are widelyof denitrificatio used becausen ofequipment their efficient in dustthe metallurgical removal industry,efficiency. such Enterprises as the selective should non-catalytic be strictly required reduction to equip method dust removaland the equipmentselective catalytic and prohibit reduction method.unorganized In addition, emissions. SO2 Besides, emission regarding in Zhuzhou the governance accounts of for VOCs 41.67% and NH of3 ,total the three emissions. cities should Industrial boilers should be a key factor for managers to consider. Not only do industrial boiler emission standards need to be improved to control SO2, but the existing desulfurization facilities should be upgraded to the limestone (lime)/gypsum process. Film-coated bag filters are widely used because of their efficient dust removal efficiency. Enterprises should be strictly required to equip dust removal equipment and prohibit unorganized emissions. Besides, regarding the governance of VOCs and NH3, the three cities should simultaneously develop more stringent standard standards

Atmosphere 2020, 11, 739 17 of 19 simultaneously develop more stringent standard standards and recommend enterprises to use VOCs removal technologies, including regenerative thermal oxidizers and regenerative catalytic oxidation. Last but not least, Changsha should provide more information technology and financial services to help Zhuzhou and Xiangtan save energy and reduce emissions.

5. Conclusions In this study, we developed a high-resolution CZT urban agglomeration air pollutant emission inventory for the year 2015. Conclusions are as follows:

1. The total emissions of SO2, NOX, PM10, PM2.5, VOCs, and NH3 are 132.5, 148.9, 111.6, 56.5, 119.0, and 72.0 kt, respectively. The discharge of atmospheric pollutants in the CZT urban agglomeration shows obvious spatial differences. The monthly variation trend of major air pollutants is relatively stable, and the monthly emission of some pollutants peak in autumn and winter.

2. The chemical composition data indicate that the main species in the PM2.5 of the CZT urban 2 agglomeration in 2015 are SO4 −, OC, and NO3−, and the annual average concentrations are 3 13.06, 8.24, and 4.84 µg/m , respectively. The regional PM2.5 pollution shows obvious seasonal differences, and the PM2.5 concentration in winter varies greatly. The results show that the influence of the source types of Changsha, Zhuzhou, and Xiangtan on PM2.5 is not significant and consistent, but pollution causes of PM2.5 are similar. 3. The source and transmission path of the air mass in the CZT urban agglomeration vary in different seasons. In summer, it is mainly controlled by the air mass transmitted by the middle distance of Guangdong and Guangxi along the coast, accounting for 87%. The autumn and winter are mainly controlled by the transmission air masses in the southeast direction of Jiangsu–Anhui–South Hubei, accounting for 92% and 72%, respectively. Potential source analysis shows that Jiangsu, Anhui, southern Hubei, and northwestern Jiangxi are potential contributing areas for major particulate matter in CZT urban agglomeration. 4. The comparison of this inventory of the YRD region and the BTH region presage that the emissions of six major pollutants in CZT urban agglomeration need to be taken seriously in the next decade. The concentration of PM and NH3 in CZT urban agglomeration is not much different from that in the YRD region and the BTH region.

Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4433/11/7/739/s1, Table S1: Classification of air pollutant emission sources, Table S2. Detailed information of various models, Table S3: Data sources used to derive the parameters needed for the CZT inventory, Table S4: Content of main chemical components in PM2.5 in different cities of Changsha, Zhuzhou and Xiangtan, Table S5: Uncertainty assessment of major emission sources in the Changzhutan urban agglomeration. Author Contributions: Conceptualization, B.X. and C.D.; Data curation, B.X. and X.Y.; Formal analysis, Y.Z.; Funding acquisition, C.D. and Z.L.; Methodology, X.Y.; Software, D.L.; Supervision, Y.Z., Z.L., S.H. and H.P.; Visualization, B.X.; Writing—original draft, B.X.; Writing—review & editing, C.D. and S.H. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Key Research and Development Program of Hunan Province (2019SK2071), Natural Science Foundation of Hunan Province (2018JJ3240), Project of Science and Technology of Hunan Education Department (16C0761). Conflicts of Interest: No conflict of interest exits in the submission of this manuscript, and the manuscript has approved by all authors for publication. The authors would like to declare that the manuscript is original and has not been published previously, and is not under consideration for publication elsewhere, in whole or in part.

References

1. Kurokawa, J.; Ohara, T.; Morikawa, T.; Hanayama, S.; Janssens-Maenhout, G.; Fukui, T.; Kawashima, K.; Akimoto, H. Emissions of air pollutants and greenhouse gases over Asian regions during 2000–2008: Regional Emission inventory in Asia (REAS) version 2. Atmos. Chem. Phys. 2013, 13, 11019–11058. [CrossRef]

2. Kato, N.; Akimoto, H. Anthropogenic emissions of SO2 and NOx in Asia: Emission inventories. Atmos. Environ. Part A Gen. Top. 1992, 26, 2997–3017. [CrossRef] Atmosphere 2020, 11, 739 18 of 19

3. Benkovitz, C.M.; Scholtz, M.T.; Pacyna, J.; Tarrasón, L.; Dignon, J.; Voldner, E.C.; Spiro, P.A.; Logan, J.A.; Graedel, T.E. Global gridded inventories of anthropogenic emissions of sulfur and nitrogen. J. Geophys. Res. Atmos. 1996, 101, 29239–29253. [CrossRef] 4. Goebes, M.D.; Strader, R.; Davidson, C. An ammonia emission inventory for fertilizer application in the United States. Atmos. Environ. 2003, 37, 2539–2550. [CrossRef] 5. Hao, J.; Tian, H.; Lu, Y. Emission inventories of NOx from commercial energy consumption in China, 1995 1998. Environ. Sci. Tech. 2002, 36, 552–560. [CrossRef] − 6. Chen, Y.; Zhi, G.; Feng, Y.; Fu, J.; Feng, J.; Sheng, G.; Simoneit, B.R.T. Measurements of emission factors for primary carbonaceous particles from residential raw-coal combustion in China. Geophys. Res. Lett. 2006, 33, 382–385. [CrossRef] 7. Cao, G.; Zhang, X.; Zheng, F. Inventory of black carbon and organic carbon emissions from China. Atmos. Environ. 2006, 40, 6516–6527. [CrossRef] 8. Qiang, Z.; David, G.S.; Kebin, H.; Zbigniew, K. Major components of China’s anthropogenic primary particulate emissions. Environ. Res. Lett. 2007, 2, 045027. 9. Zheng, J.; Shao, M.; Che, W.; Zhang, L.; Zhong, L.; Zhang, Y.; Streets, D. Speciated VOC Emission Inventory and Spatial Patterns of Ozone Formation Potential in the Pearl River Delta, China. Environ. Sci. Tech. 2009, 43, 8580–8586. [CrossRef] 10. Ou, J.; Zheng, J.; Li, R.; Huang, X.; Zhong, Z.; Zhong, L.; Lin, H. Speciated OVOC and VOC emission inventories and their implications for reactivity-based ozone control strategy in the Pearl River Delta region, China. Sci. Total Environ. 2015, 530–531, 393–402. [CrossRef]

11. Lee, S.; Liu, W.; Wang, Y.; Russell, A.G.; Edgerton, E.S. Source Apportionment of PM2.5: Comparing PMF And CMB Results For Four Ambient Monitoring Sites In The Southeastern United States. Atmos. Environ. 2008, 42, 4126–4137. [CrossRef] 12. You, X. Anthropogenic Ammonia Emission Inventory and Its Spatial Distribution in Chang-Zhu-Tan Region. Environ. Sci. 2016, 37, 94–101. 13. United States Environmental Protection Agency. Compilation of Air Pollutant Emission Factors, AP-42;Office of Air Quality Standards, Emission Inventory Branch, US Environmental: Washington, DC, USA, 1995.

14. Lei, Y.; Qiang, Z.; Nielsen, C.; He, K. An inventory of primary air pollutants and CO2 emissions from cement production in China, 1990–2020. Atmos. Environ. 2011, 45, 147–154. [CrossRef] 15. Zhao, Y.; Wang, S.; Nielsen, C.P.; Li, X.; Hao, J. Establishment of a database of emission factors for atmospheric pollutants from Chinese coal-fired power plants. Atmos. Environ. 2010, 44, 1515–1523. [CrossRef] 16. Li, X.G.; Wang, S.X.; Duan, L.; Hao, J.; Li, C.; Chen, Y.S.; Yang, L. Particulate and trace gas emissions from open burning of wheat straw and corn stover in China. Environ. Sci. Tech. 2007, 41, 6052–6058. [CrossRef] [PubMed] 17. Wang, S.; Wei, W.; Du, L.; Li, G.; Hao, J. Characteristics of gaseous pollutants from biofuel-stoves in rural China. Atmos. Environ. 2009, 43, 4148–4154. [CrossRef] 18. Jetter, J.; Zhao, Y.; Smith, K.R.; Khan, B.; Yelverton, T.; DeCarlo, P.; Hays, M.D. Pollutant Emissions and Energy Efficiency under Controlled Conditions for Household Biomass Cookstoves and Implications for Metrics Useful in Setting International Test Standards. Environ. Sci. Tech. 2012, 46, 10827–10834. [CrossRef][PubMed] 19. Shen, G.; Tao, S.; Wei, S.; Zhang, Y.; Wang, R.; Wang, B.; Li, W.; Shen, H.; Huang, Y.; Chen, Y.; et al. Reductions in Emissions of Carbonaceous Particulate Matter and Polycyclic Aromatic Hydrocarbons from Combustion of Biomass Pellets in Comparison with Raw Fuel Burning. Environ. Sci. Tech. 2012, 46, 6409–6416. [CrossRef] [PubMed] 20. Andreae, M.O.; Merlet, P. Emission of trace gases and aerosols from biomass burning. Glob. Biogeochem. Cycl. 2001, 15, 955–966. [CrossRef] 21. Liu, H. Research on Coupled Mapping Model of Urban Vehicle Activity and Emissions; Tsinghua University, China Academic Journal Electronic Publishing House: Beijing, China, 2008. 22. Zhang, C.; Chai, C.; Zhou, Z.; Lu, X. Research on Ammonia Emission from Agricultural Sources in Chongqing Urban City. Environ. Monit. China 2014, 30, 90–96. [CrossRef] 23. Zhou, Y.; Wu, Y.; Zhang, S.; Fu, L.; Hao, J. Evaluating the emission status of light-duty gasoline vehicles and motorcycles in Macao with real-world remote sensing measurement. J. Environ. Sci. 2014, 26, 2240–2248. [CrossRef][PubMed] Atmosphere 2020, 11, 739 19 of 19

24. Hu, J.; Wu, Y.; Wang, Z.; Li, Z.; Zhou, Y.; Wang, H.; Bao, X.; Hao, J. Real-world fuel efficiency and exhaust emissions of light-duty diesel vehicles and their correlation with road conditions. J. Environ. Sci. 2012, 24, 865–874. [CrossRef] 25. Fu, M.; Ge, Y.; Tan, J.; Zeng, T.; Liang, B. Characteristics of typical non-road machinery emissions in China by using portable emission measurement system. Scie. Total Environ. 2012, 437, 255–261. [CrossRef][PubMed] 26. Bo, Y.; Cai, H.; Xie, S.D. Spatial and temporal variation of historical anthropogenic NMVOCs emission inventories in China. Atmos. Chem. Phys. 2008, 8, 11519–11566. [CrossRef] 27. Wei, W.; Wang, S.; Hao, J. Uncertainty analysis of emission inventory for volatile organic compounds from anthropogenic sources in China. J. Environ. Sci. 2011, 32, 305–312. 28. Fu, X.; Wang, S.; Zhao, B.; Xing, J.; Cheng, Z.; Liu, H.; Hao, J. Emission inventory of primary pollutants and chemical speciation in 2010 for the Yangtze River Delta region, China. Atmos. Environ. 2013, 70, 39–50. [CrossRef] 29. Ying, Z.; Cheng, S.; Lang, J.; Chen, D.; Zhao, B.; Chao, L.; Ran, X.; Li, T. A comprehensive ammonia emission inventory with high-resolution and its evaluation in the Beijing-Tianjin-Hebei (BTH) region, China. Atmos. Environ. 2015, 106, 305–317. 30. Wu, W.; Zhao, B.; Wang, S.; Hao, J. Ozone and secondary organic aerosol formation potential from anthropogenic volatile organic compounds emissions in China. J. Environ. Sci. 2017, 53, 224–231. [CrossRef]

31. Wang, Y.; Dong, H.; Zhu, Z.; Li, L.; Zhou, T.; Jiang, B.; Xin, H. CH4, NH3,N2O and NO emissions from stored biogas digester effluent of pig manure at different temperatures. Agric. Ecosys. Environ. 2016, 217, 1–12. [CrossRef] 32. Jiang, X.; Wiedinmyer, C.; Carlton, A.G. Aerosols from fires: An examination of the effects on ozone photochemistry in the Western United States. Environ. Sci Tech. 2012, 46, 11878–11886. [CrossRef] 33. Chandrasekaran, S.R.; Hopke, P.K.; Newtown, M.; Hurlbut, A. Residential-Scale Biomass Boiler Emissions and Efficiency Characterization for Several Fuels. Energy Fuels 2013, 27, 4840–4849. [CrossRef] 34. Liu, H.; Tian, H.; Zhang, K.; Liu, S.; Cheng, K.; Yin, S.; Liu, Y.; Liu, X.; Wu, Y.; Liu, W.; et al. Seasonal variation,

formation mechanisms and potential sources of PM2.5 in two typical cities in the Central Plains Urban Agglomeration, China. Sci. Total Envion. 2019, 657, 657–670. [CrossRef] 35. Yang, J.; Kang, S.; Ji, Z.; Yang, S.; Li, C.; Tripathee, L. Vital contribution of residential emissions to atmospheric

fine particles (PM2.5) during the severe wintertime pollution episodes in Western China. Environ. Pollut. 2019, 245, 519–530. [CrossRef] 36. Dai, C.; Huang, S.; Peng, H.; Yi, K.; Zhou, Y.; Qin, P. Particulate pollution status and its characteristics during 2015–2016 in Hunan, China. Atmos. Pollut. Res. 2018, 739–748. [CrossRef] 37. Atwood, S.A.; Reid, J.S.; Kreidenweis, S.M.; Cliff, S.S.; Zhao, Y.; Lin, N.-H.; Tsay, S.-C.; Chu, Y.-C.; Westphal, D.L. Size resolved measurements of springtime aerosol particles over the northern South China Sea. Atmos. Environ. 2013, 78, 134–143. [CrossRef] 38. Qi, J.; Zheng, B.; Li, M.; Yu, F.; Chen, C.; Liu, F.; Zhou, X.; Yuan, J.; Zhang, Q.; He, K. A high-resolution air pollutants emission inventory in 2013 for the Beijing-Tianjin-Hebei region, China. Atmos. Environ. 2017, 170, 156–168. [CrossRef] 39. Zhang, Q.; Streets, D.G.; Carmichael, G.R.; He, K.; Huo, H.; Kannari, A.; Klimont, Z.; Park, I.; Reddy, S.; Fu, J. Asian emissions in 2006 for the NASA INTEX-B mission. Atmos. Chem. Phys. 2009, 9, 5131–5153. [CrossRef]

© 2020 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/).