EN中文

Journal of Geographical Sciences >

2021 , Vol. 31 >Issue 6: 878 - 898

DOI: https://doi.org/10.1007/s11442-021-1876-2

Research Articles

Spatiotemporal evolution of PM2.5 concentrations in urban agglomerations of China

  • WANG Zhenbo , 1, 2 ,
  • LIANG Longwu 1, 2 ,
  • WANG Xujing 3
Expand
  • 1. Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China
  • 2. College of Resources and Environment, University of Chinese Academy of Sciences, Beijing 100049, China
  • 3. College of Geographical Science, Shanxi Normal University, Linfen 041004, Shanxi, China

Wang Zhenbo, PhD and Associate Professor, specialized in urbanization. E-mail:

Received date: 2021-01-02

  Accepted date: 2021-03-20

  Online published: 2021-08-25

Supported by

National Natural Science Foundation of China(41771181)

National Key Research and Development Plan(2017YFC0505702)

Open Fund Project of New Urbanization Research Institute of Tsinghua University(TUCSU-K-17015-01)

Fold

Abstract

As the main form of new urbanization in China, urban agglomerations are an important platform to support national economic growth, promote coordinated regional development, and participate in international competition and cooperation. However, they have become core areas for air pollution. This study used PM2.5 data from NASA atmospheric remote sensing image inversion from 2000 to 2015 and spatial analysis including a spatial Durbin model to reveal the spatio-temporal evolution characteristics and main factors controlling PM2.5 in China’s urban agglomerations. The main conclusions are as follows: (1) From 2000 to 2015, the PM2.5 concentrations of China’s urban agglomerations showed a growing trend with some volatility. In 2007, there was an inflection point. The number of low-concentration cities decreased, while the number of high-concentration cities increased. (2) The concentrations of PM2.5 in urban agglomerations were high in the west and low in the east, with the “Hu Line” as the boundary. The spatial differences were significant and increasing. The concentration of PM 2.5 grew faster in urban agglomerations in the eastern and northeastern regions. (3) The urban agglomeration of PM2.5 had significant spatial concentrations. The hot spots were concentrated to the east of the Hu Line, and the number of hot-spot cities continued to rise. The cold spots were concentrated to the west of the Hu Line, and the number of cold-spot cities continued to decline. (4) There was a significant spatial spillover effect of PM2.5 pollution among cities within urban agglomerations. The main factors controlling PM2.5 pollution in different urban agglomerations had significant differences. Industrialization and energy consumption had a significant positive impact on PM2.5 pollution. Foreign direct investment had a significant negative impact on PM2.5 pollution in the southeast coastal and border urban agglomerations. Population density had a significant positive impact on PM2.5 pollution in a particular region, but this had the opposite effect in neighboring areas. Urbanization rate had a negative impact on PM2.5 pollution in national-level urban agglomerations, but this had the opposite effect in regional and local urban agglomerations. A high degree of industrial structure had a significant negative impact on PM2.5 pollution in a region, but this had an opposite effect in neighboring regions. Technical support level had a significant impact on PM2.5 pollution, but there were lag effects and rebound effects.

Key words： urban agglomeration; PM2.5; spatiotemporal evolution; influencing factor; spatial Durbin model

Cite this article

WANG Zhenbo , LIANG Longwu , WANG Xujing . Spatiotemporal evolution of PM2.5 concentrations in urban agglomerations of China[J]. Journal of Geographical Sciences, 2021 , 31(6) : 878 -898 . DOI: 10.1007/s11442-021-1876-2

1 Introduction

Since its reform and opening up, China has been making great progress in urbanization and industrialization; in the last 40 years, it has achieved progress in urbanization and industrialization that took developed countries more than 200 years (Fang, 2018). The urbanization rate of China increased from 17.92% in 1978 to 58.52% in 2017. However, long-term extensive development has degraded the eco-environment and brought about many pollution issues, especially air pollution (Wang et al., 2018, 2019). In recent years, air-quality indexes in many cities in central and eastern China have often exceeded the maximum values that could be monitored by the equipment (Tian et al., 2014). This has attracted increasing concern about air pollution in China, and PM2.5 in particular.
Studies have indicated that PM2.5 can have a great impact on human health, both physically and mentally. Both short- and long-term exposure to PM2.5 increase the risk of morbidity from respiratory diseases such as asthma (Tecer et al., 2018), cerebrovascular diseases (Santibañezet al., 2013), and mental illnesses (Massimiliano et al., 2018). Haze events have also been observed in many other countries across the world, such as the United States (Franklin et al., 2007), Germany (Kourtchev et al., 2008), Canada (Lee et al., 2003), and India (Chowdhury et al., 2007). As China is facing more severe haze pollution, how to tackle such a serious situation has aroused international concern. In-depth studies are therefore urgently needed to find out the main sources of PM2.5 in China, its spatiotemporal evolution patterns, its influencing factors, and potential methods for its prevention and control.
In the past decade, a great number of studies have focused on PM2.5, especially on its chemical characteristics (Sun et al., 2006), spatial clustering (Xiong et al., 2017), spatial variability (Pinto et al., 2004), and inhalable microorganisms during haze pollution (Cao et al., 2014). The main human and social factors that have been found to affect PM2.5 concentrations include per capita GDP and urbanization rate (Wang and Fang, 2016), population density and intensity of public transportation (Yang and Wang, 2017), foreign direct investment (Yan and Qi, 2017), and energy consumption (Wu et al., 2017), while physical geographical factors mainly cover barometric pressure, temperature, relative humidity, wind speed, precipitation, sunshine duration and gases such as SO2, NO2, CO, and O3 (He and Lin, 2017). Previous studies have relied on models including grey correlation (He et al., 2016), geographic detectors (Zhou et al., 2017), land-use regression (Eeftens et al., 2012), principal component analysis (Song et al., 2006), hybrid regression (Kloog et al., 2012), and spatial econometrics (Liu et al., 2017). Scholars have made many suggestions, such as pollution control across administrative regions at different levels (Wang et al., 2017), coordinated management by multiple subjects (Liu and Lei, 2018), compensation mechanisms by multiple parties for haze control (Zhou et al., 2017), and the promotion of meteorological science and technology (Mu and Zhang, 2014). Overall, most of the existing research has analyzed PM2.5 pollution in provinces or cities, but there have been notably few studies examining this pollution on an urban agglomeration basis.
An urban agglomeration is a specific area comprising a megalopolis as a central place and at least three metropolitan areas or large and medium-sized cities as the basic units. These city groups are characterized by urban integration and are relatively compact spaces with close economic ties. These are highly reliant on well-developed traffic, communication, and other infrastructure networks. Urban agglomerations are a highly developed form of spatially integrated cities with elements and resources highly concentrated in metropolises (Fang, 2014). They have become a key component of global competition strategy and international division (Wang and Fang, 2011).
The Central Urbanization Work Conference in 2013 was the first to regard urban agglomerations as the main body in a new type of urbanization. They have been continuously considered as the main pillar of growth in economic and social development for the last 15 years in reports at the 17th (2007), 18th (2012), and 19th (2017) national congresses of the Communist Party of China. Compared with urban districts, conflicts between urbanization and environmental protection seem to be more obvious in urban agglomerations; that is, environmental pollution phenomena are more complicated (Fang et al., 2016a) and high PM2.5 concentrations are usually observed in urban agglomerations (Fang et al., 2016b). This is the reason that quantitative research into the spatiotemporal evolution patterns of PM2.5 concentrations and their influencing mechanisms in China’s urban agglomerations is urgently needed. Data from such studies will lead to proposals for the prevention and control of this pollution and enable the application of more targeted measures.
This study revealed the spatiotemporal evolution patterns of PM2.5 in 19 urban agglomerations in China’s 13th Five-Year Plan (2016-2020) based on panel data in the period 2000-2015. Spatial autocorrelation was adopted to examine the spatial clustering of PM2.5 in the study areas, and a Tobit model was used to determine the main factors controlling PM2.5 concentrations in different urban agglomerations and the mechanisms influencing these concentrations. The results will provide an important reference for the prevention and control of air pollution in China.

2 Study areas, influencing factors and data sources

2.1 Study areas

China’s National New-type Urbanization Plan (2014-2020) proposed that there should be 19 urban agglomerations under construction in the future. This paper takes these agglomerations as cases for discussing the spatiotemporal characteristics of PM2.5 in cities and analyzing its main human influencing factors. Of the 19 agglomerations, five are national level, namely, Beijing-Tianjin-Hebei (BTH), Yangtze River Delta (YRD), Pearl River Delta (PRD), Middle Reaches of Yangtze River (MYR), and Chengdu-Chongqing (CC); eight are regional level with medium sizes, namely, Central and Southern Liaoning (CSL), Shandong Peninsula (SP), West Coast of the Taiwan Straits (WTS), Harbin-Changchun (HC), Central (China) Plains (CCP), Guanzhong (GZ), Beibu Gulf (BG), and Northern Slope of Tianshan Mountains (NTM); and six are local level with relatively small sizes, which are Central Shanxi (CS), Hohhot-Baotou- Erdos-Yulin (HBEY), Central Yunnan (CY), Central Guizhou (CG), Lanzhou-Xining (LX), and Cities in Ningxia along the Yellow River (NX). These are shown in Figure 1 (Fang et al., 2015).
Figure 1 Map of urban agglomeration planning areas in China

2.2 Influencing factors and data sources

The spatiotemporal disparities in PM2.5 concentrations are highly dependent on multiple factors including geographical differences and human and economic elements. According to the environmental Kuznets curve theory and relevant research outcomes (Guang and Xu, 2014; Li et al., 2018; Liu et al., 2018), eight human elements were selected for this study, namely, per capita GDP (PGDP), population density (PD), urbanization rate (UR), industrialization rate (IR), high degree of industrial structure (HDIS), foreign direct investment (FDI), technical support (TS), and energy consumption (EC) (Figure 2). Additionally, five geographical elements, namely, wind speed, sunshine duration, air humidity, temperature, and precipitation were selected, along with two political elements, namely, government research and development (R&D) input, and environmental governance investment as control variables. Due to their limited characteristics, the study did not analyze the impacts of the two control variables on PM2.5 pollution in detail.
Figure 2 Interactive coupling of human elements comprehensively affecting PM2.5 pollution in China’s urban agglomerations
The following metrics are used for particular factors: IR is the proportion of industrial output value to GDP; HDIS is the angle between the specific gravity vector of the three industrial output values and the corresponding unit vector; TS is the proportion of science and technology in the GDP; and EC is per capita natural gas supply rather than coal consumption, because it is difficult to obtain accurate per capita coal-consumption data (Huang et al., 2019). It can be concluded that the internal human elements in a city jointly affect the PM2.5 concentration; multiple cities in an urban agglomeration develop in an interactive and coupled way, and this comprehensively influences the PM2.5 within an agglomeration. Similarly, the interactive coupling of multiple urban agglomerations also affects the PM2.5 concentrations across the country as a whole (Figure 2).
The data in this paper include vector data indicating the administrative boundaries of prefecture-level cities in urban agglomerations from fundamental geographic information of China at a scale of 1:4,000,000 provided by the National Geomatics Center of China, inversion of raster data of the annual average concentration of PM2.5 from the NASA Socioeconomic Data and Applications Center for 2000-2015 (Lee et al., 2011), and data of its influencing factors, which are mainly from the China City Statistical Yearbook from 2001 to 2016 (some of the missing data are taken from provincial or prefecture-level statistical yearbooks or annual reports, Hong Kong, Macao and Taiwan are excluded).

2.3 Examination of the statistical significance of the indexes

To study the collinearity of the indexes, the SPSS software package was used to analyze the correlation of variables before the quantitative regression analysis of the spatial panel data. The test results are shown in Table 1. The variance inflation factor and condition indexes demonstrate that there is almost no collinearity among the variables described above.
Table 1 The collinearity test results of indexes
Index Index
Per capita GDP Population density Urbanization rate Industrialization rate High degree of industrial structure Dependence on foreign trade Technical support Energy consumption
Per capita GDP 1
Population density 0.017 1
Urbanization rate 0.002 0.061** 1
Industrialization rate 0.009 0.045** 0.000 1
High degree of industrial structure 0.036* 0.166** 0.054** 0.049** 1
Dependence on foreign trade 0.019 0.226** 0.049** 0.036* 0.183** 1
Technical support 0.161** 0.096** -0.018 0.006 0.070** 0.109** 1
Energy consumption 0.010 0.014 -0.007 0.002 0.031 0.003 0.019 1

Note: ** represents significant correlation at the 0.01 level on both sides and * represents the 0.05 level. The mean value of the condition index is 43.47; the variance inflation factors of each variable are all less than 3, with a mean value of 2.51.

3 Methods

3.1 Spatial autocorrelation

The PM2.5 concentrations show that there is significant spatial correlation under the effects of atmospheric circulation, which is of great significance for examining the development of this pollution (Fang et al., 2016a). Hence, spatial correlation, especially global and local spatial correlation, is frequently adopted for studies of the spatial agglomeration and variation of air pollution (Hu et al., 2013).

3.1.1 Global spatial correlation

In spatial correlation, the global Moran’sI index is a common way to estimate the average similarity of PM2.5 concentrations in adjacent areas. The calculation equation is as follows:
$I=\frac{n}{S_{0}} * \frac{\sum_{i=1}^{n} \sum_{j=1}^{n} w_{i j} z_{i} z_{j}}{\sum_{i=1}^{n} z_{i}^{2}} \quad S_{0}=\sum_{i=1}^{n} \sum_{j=1}^{n} w_{i j} \quad z_{i}=Y_{i}-\bar{Y}, \quad z_{j}=Y_{j}-\bar{Y}$
where I represents the global spatial correlation index; Yi and Yj represent the observed values of air quality in city i and city j, and $\bar{Y}$. is the mean air-quality value; wij is spatial weight matrix for which the values of neighbors are 1, while those of others are 0. When I∈[-1, 1],I∈[-1, 0) indicates that there is a negative correlation between regional units;n is the number of units studied; while I∈(0, 1] indicates a positive correlation;I = 0 indicates there is no correlation. The closer the Moran’sI index to 1, the closer the relationship between the attribute values in regional units and vice versa.

3.1.2 Local spatial correlation

Local spatial correlation is used to estimate the effects of spatial correlation on local spatial units compared to those on global spatial units. This is equal to the degree of correlation between the air quality of a regional unit and its surrounding units. The equation is expressed as:
$\begin{matrix} Local & \begin{matrix} Mora{n}'s & I=\frac{n({{x}_{i}}-\bar{x})\sum\limits_{j=1}^{m}{{{W}_{ij}}({{x}_{j}}-\bar{x})}}{\sum\limits_{i=1}^{n}{{{({{x}_{i}}-\bar{x})}^{2}}}},(i\ne j) \\\end{matrix} \\ \end{matrix}$
where xi and xj represent the observed values of air quality in city i and city j; n is the number of cities; Wij represents the space weight; and $i=1,2,...,n,j=1,2,...,m$, in which m is the number of cities adjacent to city i. Within the academic community, the standardized statistic Z is usually used to explore whether there exists spatial correlation in Moran’sI index, which is expressed as
$Z_{i}=\frac{I-E[I]}{\sqrt{V[I]}}, \quad E[I]=-1 /(N-1), \quad V[I]=E\left[I^{2}\right]-E[I]^{2}$
In this study, a significance level of 0.01 was adopted to enhance the accuracy of the results. When $\left| Z(I) \right|<\text{2}\text{.58}$, this means that the spatial correlation of PM2.5 concentration is not significant, and it has an independent, random distribution. When $Z(I)<\text{2}\text{.58}$, this shows a negative correlation in the spatial distribution of PM2.5 concentration, and the attributes of the concentration are inversely distributed, and this includes “high–low” correlations and “low–high” correlations. When $Z(I)>\text{2}\text{.58}$, this indicates a positive correlation with a manifestation of spatial agglomeration among similar high/low values, which are also known as hot- or cold-spot areas.

3.2 Spatial econometric model

Based on the core of spatial variation in geography, the spatial econometric model uses a spatial weight matrix that takes the spatial correlation of elements into consideration. This is more objective than classical econometric models. Pollution by PM2.5 in cities is not isolated but is closely related to pollution in the surrounding areas; it has relatively strong spatial spillover. Therefore, spatial effects should not be ignored when analyzing the driving force of human factors. The spatial econometric model is appropriate here as it analyzes cross-section and panel models. The cross-section model only uses the data from a specific year and ignores any lag effects in the factors. The panel model increases the number of indicators, and this satisfies the demand for the asymptotic properties for large samples and makes full use of the data. Clearly, the latter has a higher degree of accuracy (Xi and Li, 2015). The spatial lag model (SLM), spatial error model (SEM), and spatial Durbin model (SDM) are included in the spatial panel econometric model (Cheng et al., 2014). The equation for the SDM is
$Panel-SDM:\ln P{{M}_{it}}=\alpha W\ln P{{M}_{it}}+\varphi \ln P{{M}_{it-1}}+{{\beta }_{0}}+{{\beta }_{i}}{{X}_{it}}+\theta W{{Z}_{it}}+{{a}_{i}}+{{\gamma }_{t}}+{{\mu }_{it}}$
where lnPMit, lnPMit-1, and WlnPMit represent the logarithmic values of PM2.5 concentrations in urban areas and their time- and space-lags, respectively; Xit is the panel data representing the explanatory variable; WZit is the space-lag of the explanatory variable; ai, γt, and μit represent individual effects, time effects, and an error term, respectively; φ and α are coefficients of the time- and space-lags of the explanatory variable; β0 and βi are the parameter vectors of order K×1 to be estimated; andθ represents the coefficient of the space-lag of the explanatory variable.
When error terms in the model have spatial correlation, the SEM is better; when the spatial dependence of an explanatory variable plays an important role in the model and shows significant spatial correlation, the SLM is better. The SDM is the general form of the SEM and SLM. The SLM focuses on the endogenous interaction effect of an explanatory variable, while the SEM focuses on the interaction effect of error terms; the SDM focuses on both endogenous interaction effects and exogenous interaction effects (Elhorst, 2010). Considering the relatively strong spatial correlation of PM2.5 pollution and its influencing factors, this study used SDM. The adjacent space weight matrix was selected for this study, in which adjacent space units have a significant effect on each other while non-adjacent units basically have no effect. The code used for the calculation of the SDM was Elhorst’s spatial econometrics Matlab code.

4 Results

4.1 Spatiotemporal evolution pattern of PM2.5 in China’s urban agglomerations

4.1.1 Analysis of temporal distribution

In 2000-2015, PM2.5 concentrations in China’s urban agglomerations exhibited an increasing trend with fluctuations. The average concentration changed from 21.50 µg/m3 to 33.23 µg/m 3, an increase of 54.56%, which indicates that the air quality deteriorated during this period (Figure 3). In terms of PM2.5 concentrations in different levels of urban agglomerations, concentrations in national urban agglomerations (excluding the Pearl River Delta) were greater than those in regional urban agglomerations (excluding the Shandong Peninsula and the Central Plains), and these were in turn greater than those in local urban agglomerations. From 2000 to 2007, the average PM2.5 concentrations increased, while those between 2007 and 2015 fluctuated. The year 2007 was the inflection point of PM2.5 concentrations in nine urban agglomerations (such as Beijing-Tianjin-Hebei and Middle Reaches of Yangtze River), which is in accordance with the results of Xu et al. (2018).
Figure 3 PM2.5 concentrations time series in China’s urban agglomerations for the period 2000-2015
In 2000-2015, the increase rates of PM2.5 concentrations in Harbin-Changchun and Central and Southern Liaoning were relatively high, and the highest PM2.5 concentrations were observed in the Beijing-Tianjin-Hebei, Shandong Peninsula, and Central Plains urban agglomerations. During this period, Lanzhou-Xining and Cities in Ningxia along the Yellow River were the only two urban agglomerations with improved air quality. The increases in the PM2.5 concentrations in Harbin-Changchun and Central and Southern Liaoning were in excess of 150%, and those in the Northern Slope of Tianshan Mountains, Central Shanxi, Beibu Gulf, Shandong Peninsula, Pearl River Delta, Middle Reaches of Yangtze River, and Beijing-Tianjin-Hebei ranged from 50% to 100%. Those of the other urban agglomerations were less than 50%.
The annual average limit for PM2.5 concentration indicated by China’s Ambient Air Quality Standard (AAQS, GB3095-2012) is 35.00 µg/m 3 (MEP, 2012). The PM2.5 concentrations in Beijing-Tianjin-Hebei, Shandong Peninsula, and Central Plains in 2000-2015 exceeded this limit. The PM2.5 concentrations in the Yangtze River Delta, Pearl River Delta, Middle Reaches of Yangtze River, Chengdu-Chongqing, Central and Southern Liaoning, Harbin- Changchun, Guanzhong, and Central Shanxi exceeded the limit in some of the years, while concentrations were below the limit every year in the remaining urban agglomerations.
The PM2.5 concentration standards set by the World Health Organization (WHO) and China’s AAQS (GB3095-2012) are shown in Table 2. We divided China’s urban agglomerations into seven annual PM2.5 level ranges to analyze the proportions of cities falling within each range every year in the period 2000-2015 (Figure 4). The results firstly indicate that the proportion of cities with PM2.5 concentrations lower than 10 µg/m 3 (the WHO guideline level) decreased from 6.19% in 2000 to 1.77% in 2015, and the proportion with less than 15 µg/m 3 (the annual average limit of interim target 3) reduced from 20.35% to 7.52%. The proportion of cities with PM2.5 concentrations higher than 35 µg/m 3 (the annual average limit in China) almost tripled, increasing from 16.81% to 47.79%; the proportion with PM2.5 concentrations higher than 50 µg/m 3 (the daily average limit of interim target 2) increased by more than six times, from 4.42% to 32.74%. Finally, more than 50% of the cities with PM2.5 annual average concentrations higher than the annual average limit set by GB3095-2012 mainly occurred in the period 2005-2010, followed by 2013-2014. This means that the numbers of low-pollution cities (below 15 µg/m 3) and very low-pollution cities (below 10 µg/m 3) decreased overall, while the number with high pollution (higher than 35 µg/m 3) and extremely high pollution (higher than 50 µg/m 3) showed a rapidly increasing trend.
Table 2 Standard values of PM2.5 concentrations set by the WHO and China
WHO Guidelines for Air Quality (2005) China’s Ambient Air Quality Standard (GB3095-2012) (2016)
Type Annual average (μg/m3) Daily average (μg/m3) Type Annual average (μg/m3) Daily average (μg/m3)
Guideline level 10 25 Standard level 35 75
Interim target 1 35 75 — — —
Interim target 2 25 50 — — —
Interim target 3 15 37.5 — — —

Note: — indicates no corresponding value exists.

Figure 4 Changing trend in proportions of annual average PM2.5 concentration range in China’s urban agglomerations for 2000-2015

4.1.2 Analysis of spatial patterns

Taking the annual average limit of 35 µg/m 3 in GB3095-2012 as a cut-off point, we divided the urban agglomerations into two types, namely high-value areas and low-value areas. From 2000 to 2015, the PM2.5 concentrations in different urban agglomerations varied. On the whole, the PM2.5 concentration decreased from the southeast coast to the northwest inland, with the Hu Line as a divide. Regional differences widened in this period (Figure 5). Specifically, compared with national and regional urban agglomerations, local ones had lower PM2.5 concentrations. Low-value areas were mainly concentrated in the Northern Slope of the Tianshan Mountains, Hohhot-Baotou-Erdos-Yulin, Cities in Ningxia along the Yellow River, Lanzhou-Xining, Central Yunnan, and West Coast of the Taiwan Straits; high-value areas were Harbin-Changchun, Central and Southern Liaoning, Shandong Peninsula, Central Plains, northern Yangtze River Delta, southern Beijing-Tianjin-Hebei, and northern Middle Reaches of Yangtze River. To analyze the spatial differences in PM2.5 concentration, this study selected years with the same time intervals. The main conclusions are as follows.
Figure 5 Evolution of the spatial patterns in PM2.5 concentration (µg/m3) in China’s urban agglomerations for 2000-2015
First, the PM2.5 concentrations had obvious polarization. The areas southeast of the Hu Line are mainly high-value areas, while those northwest are mainly low-value areas. The spatial pattern basically follows the population density, which reflects the fact that human activities have distinct effects on PM2.5 concentrations. Second, PM2.5 concentrations decreased from southeastern coastal areas to northwestern inland areas. High-value areas were highly concentrated, especially in urban agglomerations along the Yangtze River and Yellow River. Beijing-Tianjin-Hebei, Shandong Peninsula, Central Plains, Middle Reaches of Yangtze River, as well as a small number of cities in Yangtze River Delta were highly polluted areas, accounting only for 16.8%. Both the total high-value area and pollution levels increased. In 2015, cities in high-value areas occupied 47.3% of the total, and these were mainly concentrated in Shandong Peninsula, Central Plains, Yangtze River Delta, Middle Reaches of Yangtze River, and southeastern Beijing-Tianjin-Hebei. Third, PM2.5 rapidly increased in the urban agglomerations of eastern and northeastern China. During 2000-2015, PM2.5 concentrations only decreased in Lanzhou-Xining and Cities in Ningxia along the Yellow River, while those in other urban agglomerations all increased, especially in Harbin-Changchun and Central and Southern Liaoning. In addition, the PM2.5 concentration in the Northern Slope of the Tianshan Mountains was below 35 µg/m 3; however, this value doubled. More attention should be paid to this urban agglomeration due to its fragile eco-environment. Fourth, during 2000-2015, PM2.5 concentrations in Beijing-Tianjin-Hebei, Shandong Peninsula, and Central Plains continuously ranked as the top-three most polluted, with characteristics of high concentration, high aggregation, and rapid increase. After 2002, the PM2.5 concentration in Shandong Peninsula was always on the top of the list.

4.2 Spatial clustering characteristic of PM2.5 in China’s urban agglomerations

4.2.1 Analysis of spatial autocorrelation index

This study used ArcGIS to examine the spatial autocorrelation of annual average PM2.5 concentrations in 18 of China’s urban agglomerations (Northern Slope of the Tianshan Mountains was excluded due to data deficiency) in the period 2000-2015 (Table 3). The results show that the Moran’sI indexes in all the urban agglomerations were positive (greater than 0.700), which passed the test at 1% significance level. This means that there was similar spatial clustering of PM2.5 concentrations in China’s urban agglomerations in 2000-2015.
Table 3 Spatial autocorrelation indexes of annual average PM2.5 concentrations in China’s urban agglomerations from 2000 to 2015
Year 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
All 0.82*** 0.79*** 0.77*** 0.84*** 0.74*** 0.75*** 0.80*** 0.82*** 0.76*** 0.76*** 0.78*** 0.79*** 0.79*** 0.83*** 0.81*** 0.80***
BTH 0.33** 0.41** 0.41** 0.39** 0.36** 0.37** 0.36** 0.36** 0.33** 0.33** 0.33** 0.33** 0.34** 0.35** 0.35** 0.34**
YRD 0.68*** 0.66*** 0.68*** 0.67*** 0.68*** 0.65*** 0.66*** 0.64*** 0.69*** 0.66*** 0.66*** 0.68*** 0.67*** 0.67*** 0.67*** 0.68***
PRD 0.13* 0.16* -0.10 0.20* 0.18* -0.16* 0.11 -0.14* 0.13* 0.15* 0.17* 0.12* 0.14* 0.17* 0.25* 0.23*
MYR 0.84*** 0.77*** 0.55*** 0.73*** 0.19*** 0.49*** 0.57*** 0.40*** 0.38*** 0.43*** 0.59*** 0.68*** 0.41*** 0.66*** 0.64*** 0.61***
CC 0.38** 0.54*** 0.32** 0.30* 0.31** 0.27* 0.19 0.23 0.15 0.12 0.19 0.20 0.20 0.19 0.12 0.19
CSL 0.11* 0.11* 0.12* 0.28*** 0.15* 0.15* 0.14* 0.21** 0.15* 0.21** 0.10 0.15* 0.20** 0.18** 0.12* 0.13*
SP 0.44*** 0.49*** 0.46*** 0.51*** 0.48*** 0.54*** 0.53*** 0.50*** 0.49*** 0.51*** 0.58*** 0.53*** 0.53*** 0.52*** 0.53*** 0.47***
WTS 0.25* -0.16* -0.29 0.10* -0.14 0.11* 0.24 0.12* 0.10 0.24* 0.11* 0.10* 0.17 0.11* 0.13* 0.21
HC 0.55** 0.67*** 0.57*** 0.62*** 0.74*** 0.74*** 0.71*** 0.75*** 0.81*** 0.72*** 0.68*** 0.77*** 0.75*** 0.61*** 0.61*** 0.52***
CCP 0.55*** 0.52*** 0.28* 0.41** 0.42** 0.38** 0.41** 0.47** 0.47** 0.49*** 0.48*** 0.42** 0.41** 0.45*** 0.53*** 0.49***
GZ -0.41* -0.31* -0.09* -0.13 -0.23* 0.09* 0.11* 0.19** 0.12* 0.10 0.09* 0.08 0.10* 0.11 0.10* 0.11*
BG 0.79*** 0.76*** 0.85*** 0.92*** 0.93*** 0.93*** 0.93*** 0.87*** 0.97*** 0.95*** 0.96*** 0.96*** 0.97*** 0.92*** 0.93*** 0.90***
CS 0.23* -0.10* -0.44 -0.59* -0.52* -0.10 -0.34* -0.25* 0.20 0.12* 0.31* 0.11* 0.13* 0.10* 0.12 0.15*
HBEY -0.18* 0.15 -0.43 -0.77* -0.78* -0.56* -0.47* -0.36 0.58* 0.79* 0.67* 0.50* 0.91* 0.87* 0.88* 0.93*
CY 0.42* 0.46* 0.57* 0.85* 0.87* 0.65* 0.13* 0.80* 0.61* 0.73* 0.81* 0.45* 0.82* 0.78* 0.67* 0.73*
CG -0.45* 0.10 -0.42* -0.17* -0.16* -0.16* -0.17* -0.10 0.11 0.25* 0.26* 0.24 0.23* 0.38* 0.19* 0.29*
LX 0.81** 0.81** 0.83** 0.79** 0.76** 0.90** 0.87** 0.84** 0.77** 0.67* 0.59* 0.74** 0.66* 0.68* 0.75** 0.76**
NX 0.33* 0.19* 0.13 0.18* 0.23* 0.23* 0.15* 0.25* 0.30* 0.22 0.28* 0.21* 0.25* 0.12 0.21* 0.32*

Note: * represents 10% significance level; ** represents 5% significance level; *** represents 1% significance level.

Statistical analysis can be used to study the hot and cold spots distribution. When applying the autocorrelation to a specific urban agglomeration, Moran’sI indexes can be either positive or negative. The indexes were low in some urban agglomerations. The PM2.5 concentrations in a small number of urban agglomerations were different, showing a negative value over a long period. However, after 2008, the values became positive with a clustering trend, and the indexes in the majority of the urban agglomerations passed the 10% significance level. Hence, it is suggested that the spatial econometric model should be adopted to explore the main factors controlling PM2.5 pollution in these urban agglomerations.

4.2.2 Analysis of clustering characteristics

Overall, the Hu Line is regarded as the boundary between the hot and cold spots of PM2.5 concentration; the hot spots are concentrated to the east of the line, while the cold spots are mainly to its west (Figure 6). From 2000, the number of cities in hot-spot areas increased and the number of cities in cold-spot areas decreased. The hot-spot areas are mainly in the urban agglomerations of eastern and central China such as Beijing-Tianjin-Hebei, Shandong Peninsula, Central Plains, Middle Reaches of Yangtze River, and Yangtze River Delta. In northern China, rapid industrialization and the use of coal in winter aggravate the air pollution (Fang et al., 2016b). The cold-spot areas are the urban agglomerations of western, southwestern, and southeastern China where the air conditions are relatively good, such as the Northern Slope of the Tianshan Mountains, Lanzhou-Xining, Hohhot-Baotou-Erdos- Yulin, Cities in Ningxia along the Yellow River, Central Yunnan, and West Coast of the Taiwan Straits. It is worth noting that Central and Southern Liaoning and the southwestern part of Harbin-Changchun became hot-spot areas in 2015, while before 2010 they were neither obviously hot nor cold spots. This means that the air pollution from burning coal for heat was getting worse in northeastern China. In 2000, 2005, and 2010 neither hot nor cold spots were found in Pearl River Delta and Central Guizhou. However, in 2015, these agglomerations became cold-spot areas, with notably improved air quality as a result of effective prevention and control. Cities in Ningxia along the Yellow River and eastern Lanzhou-Xining were undistinguished in 2000, but they were cold-spot areas in 2005, 2010, and 2015; their air-quality levels were better than those of other urban agglomerations.
Figure 6 Spatial clustering characteristics of PM2.5 concentrations in China’s urban agglomerations for 2000-2015

4.3 Analysis of the factors influencing PM2.5 in China’s urban agglomerations

4.3.1 Comparative analysis of factor estimates

In Equation (4), WlnPMit is the space-lag term, and this is not significant in Lanzhou-Xining but is significant at the 10% level in Central Plains and Beibu Gulf. In the rest of the urban agglomerations, it is significant at the 1% level. This indicates that PM2.5 pollution in urban agglomerations has a strong spatial endogenetic interaction effect; that is, the air pollution in different cities clearly interacts among urban agglomerations. The estimates of this factor exceeded 0.600 in 14 urban agglomerations, including Beijing-Tianjin-Hebei, Pearl River Delta, Middle Reaches of Yangtze River, Chengdu-Chongqing, Central and Southern Liaoning, Shandong Peninsula, West Coast of the Taiwan Straits, Harbin-Changchun, Guanzhong, Central Shanxi, Hohhot-Baotou-Erdos-Yulin, Central Yunnan, Central Guizhou, and Lanzhou-Xining (not significant). The PM2.5 concentration of an area increases by more than 0.6% if the concentration in its surrounding areas increases by 1%. The top-three estimates were found in Shandong Peninsula, Chengdu-Chongqing, and Central and Southern Liaoning.
Per capita GDP is negatively correlated with PM2.5 pollution in Beibu Gulf. It also has a spatial spillover effect in this region. In Yangtze River Delta, Middle Reaches of Yellow River, Chengdu-Chongqing, Central and Southern Liaoning, Shandong Peninsula, West Coast of the Taiwan Straits, and Harbin-Changchun, the effects of per capita GDP are positively correlated with PM2.5 concentrations, and a spatial spillover effect also exists. The effects of per capita GDP on PM2.5 pollution and its spatial spillover effect are not obvious in Beijing-Tianjin-Hebei, Pearl River Delta, Central Plains, Central Shaanxi or six regional urban agglomerations. The main reasons for these differences are the different stages of socioeconomic development and industrialization in these areas, as well as the types of industry and their production efficiencies.
Population density is positively correlated with PM2.5 concentrations in Beijing-Tianjin- Hebei, Pearl River Delta, Middle Reaches of Yangtze River, Chengdu-Chongqing, Central Shanxi, Hohhot-Baotou-Erdos-Yulin, and eight regional urban agglomerations, but this has a negative spatial spillover effect on PM2.5 concentrations in Beijing-Tianjin-Hebei, Chengdu-Chongqing, Central and Southern Liaoning, Shandong Peninsula, West Coast of the Taiwan Straits, Harbin-Changchun, Central Plains, Guanzhong, Beibu Gulf, Central Shanxi, and Lanzhou-Xining. High population density puts much pressure on a local area, but for the areas surrounding it, the situation is the opposite. This indicates that in the core cities of urban agglomerations, the intensity of human activities is strong, which brings about more emissions of pollutants.
The urbanization rate aggravates PM2.5 pollution in Beijing-Tianjin-Hebei, Yangtze River Delta, and Pearl River Delta. However, the situation is reversed in Central and Southern Liaoning, Shandong Peninsula, Harbin-Changchun, Guanzhong, Hohhot-Baotou-Erdos- Yulin, Central Yunnan, and cities in Ningxia along the Yellow River. Urbanization rate has a negative spatial spillover effect on Yangtze River Delta and Pearl River Delta, while it has a positive spatial spillover effect on Central and Southern Liaoning, Shandong Peninsula, Harbin-Changchun, Guanzhong, Hohhot-Baotou-Erdos-Yulin, Central Yunnan, and Central Guizhou. It is suggested that, in national urban agglomerations with higher urbanization rates, the effects are negative, while in regional and local ones, the effects are opposite.
Industrialization has positive effects on Yangtze River Delta, Pearl River Delta, Shandong Peninsula, Central Plains, Hohhot-Baotou-Erdos-Yulin, and Central Yunnan; it has a positive spatial spillover effect on urban agglomerations such as Beijing-Tianjin-Hebei, Pearl River Delta, Yangtze River Delta, Central and Southern Liaoning, Central Plains, Guanzhong, and Hohhot-Baotou-Erdos-Yulin. Industrialization has a significant negative correlation on PM2.5 pollution, which suggests that industrial wastes and contaminants, as well as smoke dust, aggravate the pollution.
A high degree of industrial structure alleviates the PM2.5 pollution in Beijing-Tianjin- Hebei, Pearl River Delta, Central and Southern Liaoning, Shandong Peninsula, West Coast of the Taiwan Straits, Harbin-Changchun, Hohhot-Baotou-Erdos-Yulin, Central Yunnan, and Central Guizhou, while it has positive spatial spillover effect on PM2.5 pollution in Pearl River Delta, West Coast of the Taiwan Straits, Harbin-Changchun, Guanzhong, Central Yunnan, and Central Guizhou. The optimization of the industrial structure reduces the PM2.5 pollution. However, some polluting enterprises move to the surrounding areas, which aggravates the pollution in these areas.
Foreign direct investment reduces the PM2.5 pollution in coastal urban agglomerations such as Beijing-Tianjin-Hebei, Yangtze River Delta, Pearl River Delta, Middle Reaches of Yangtze River, West Coast of the Taiwan Straits, and Beibu Gulf, as well as some urban agglomerations near the borders such as Harbin-Changchun and Central Yunnan. It also has a significant negative spatial spillover effect on PM2.5 pollution in Yangtze River Delta, Pearl River Delta, Harbin-Changchun, Central Shanxi, Hohhot-Baotou-Erdos-Yulin, and Central Guizhou. The influences of FDI include the pollution haven hypothesis and the pollution halo hypothesis (Walter and Ugelow, 1979). As FDI appears to improve the environment, the pollution haven hypothesis does not accord with the situation in China’s urban agglomerations. This is consistent with the results of Xu and Deng (2012) and Jiang et al. (2018).
Technical support improves the PM2.5 pollution in Pearl River Delta, Yangtze River Delta, Central Shanxi and Central Yunnan; however, it aggravates the pollution in the West Coast of the Taiwan Straits and Central Guizhou; it has a positive spatial spillover effect on Guanzhong and Hohhot-Baotou-Erdos-Yulin; it has a negative spatial spillover effect on Beijing-Tianjin-Hebei, Shandong Peninsula, Central Plains, and Central Shanxi. Technology innovation is conducive to PM2.5 governance, prevention, and control, as well as to air-quality improvement. The marketization of scientific and technological achievements can greatly promote the development of new technologies. However, undeveloped technology may lead to the extensive and rapid development of industrial enterprises, which also triggers a rebound effect in energy consumption. This is in accordance with the conclusions of Cheng et al. (2019). The hysteresis effect of technical innovation on science and technology achievements is not obvious over short periods, and thus it has no significant effect on PM2.5 pollution.
Energy consumption aggravates the PM2.5 pollution in Yangtze River Delta, Pearl River Delta, Central and Southern Liaoning, Shandong Peninsula, Harbin-Changchun, Central Plains, Guanzhong, Central Shanxi, Central Yunnan and Central Guizhou urban agglomerations, while it has a positive spatial spillover effect on Beijing-Tianjin-Hebei, Yangtze River Delta, Pearl River Delta, Central and Southern Liaoning, Shandong Peninsula, Harbin- Changchun, Central Plains, Guanzhong, Central Shanxi, Hohhot-Baotou-Erdos-Yulin, Central Guizhou, and Lanzhou-Xining. At the same time, due to the atmospheric motion among different cities, energy consumption in these areas also aggravates the PM2.5 pollution in their surroundings.

4.3.2 Analysis of main influencing factors in urban agglomerations

Overall, for national urban agglomerations, technical support is the main factor that reduces PM2.5 pollution, while industrialization aggravates this pollution. For regional urban agglomerations, technical support, high degree of industrial structure, and population density are the dominant factors leading to a reduction in pollution, while PM2.5 pollution in the surrounding areas may have an opposite effect. For local urban agglomerations, technical support reduces pollution, and PM2.5 pollution in the surrounding areas increases pollution (Table 4). Hence, it is of great importance to increase research input, deepen industrial restructuring, implement green industrialization, optimize the population layout, and strengthen the cooperation between different regions.
Table 4 Results of different factors influencing PM2.5 pollution in China's urban agglomerations from 2000 to 2015

Note: * represents 10% significance level; ** represents 5% significance level; *** represents 1% significance level. Due to significant data missing in Northern Slope of Tian-shan Mountains, this area could not be analyzed.

Technical support plays an important role in a notable reduction of PM2.5 concentrations in Beijing-Tianjin-Hebei, Pearl River Delta, Middle Reaches of Yangtze River, Shandong Peninsula, Central Plains, Central Shanxi, and Central Yunnan. This means that the policies made by governments at all levels are beneficial to the improvement of air quality. A high degree of industrial structure has obvious effects on the reduction of PM2.5 concentrations in Central and Southern Liaoning and Harbin-Changchun, indicating that Northeast China should promote the transformation of industrial structure and optimize polluting industries. Foreign investment remarkably reduces the PM2.5 concentrations in Yangtze River Delta, West Coast of the Taiwan Straits, Hohhot-Baotou-Erdos-Yulin, and Central Guizhou. This indicates the need for further improvement of the investment environment and increasing of measures to incentivize foreign investment.
The PM2.5 concentrations of surrounding areas is the main reason for the increases in PM2.5 concentrations in Chengdu-Chongqing, Central and Southern Liaoning, Shandong Peninsula, Central Shanxi, and Central Yunnan. Hence, cross-regional prevention and control is urgently required. The increasing PM2.5 concentrations in Beijing-Tianjin-Hebei, Yangtze River Delta, and Middle Reaches of Yangtze River are mainly due to the relatively high industrialization in these regions. For them, it is important to focus on industrial transformation and upgrading, as well as pollution control. Technical support is the main controlling factor in the West Coast of the Taiwan Straits and Guanzhong, which means that the two urban agglomerations are facing the problems of immaturity of technology and rapid marketization. Harbin-Changchun and Cities in Ningxia along the Yellow River are undergoing rapid urbanization, and it is important to control the speed of this but also to increase efficiency. Population density, high degree of industrial structure, and energy consumption are the main controlling factors in Central Plains, Pearl River Delta, and Lanzhou-Xining urban agglomerations, respectively, and these factors require further measures to control pollution.

5 Conclusions and discussion

5.1 Conclusions

The PM2.5 concentrations in China’s urban agglomerations from 2000 to 2015 showed an increasing trend with fluctuations, and the overall increase was 54.53%, which is concerning. From 2000 to 2007, there was a significant increase, while from 2007 to 2015, the concentrations fluctuated. The year 2007 was also the inflection point of the PM2.5 concentrations in nine urban agglomerations (including Beijing-Tianjin-Hebei and the Middle Reaches of Yangtze River). Specifically, from 2000 to 2015, the air quality in Lanzhou-Xining and Cities in Ningxia along the Yellow River improved; an increase of 150% was observed in the PM2.5 concentrations of Central and Southern Liaoning and Harbin-Changchun; the most seriously polluted urban agglomerations were Beijing-Tianjin-Hebei, Shandong Peninsula, and Central Plains. The proportion of cities with PM2.5 concentrations below 15 µg/m 3 decreased from 20.35% in 2000 to 7.52% in 2015. Simultaneously, the proportion of cities with PM2.5 concentrations above 35 µg/m 3 increased from 16.81% in 2000 to 47.79% in 2015.
The PM2.5 concentrations of China’s urban agglomerations varied from region to region during the period studied. On the whole, the Hu Line is regarded as a boundary, and PM2.5 concentrations decreased from the southeastern coast to northwestern inland, with increasing regional disparities. A relatively rapid increase in pollution in urban agglomerations of eastern and northeastern China was observed. In detail, the Northern Slope of Tianshan Mountains, Hohhot-Baotou-Erdos-Yulin, Cities in Ningxia along the Yellow River, Lanzhou-Xining, Central Yunnan, and West Coast of the Taiwan Straits had low PM2.5 concentrations, while Harbin-Changchun, Central and Southern Liaoning, Shandong Peninsula, Central Plains, north of Yangtze River Delta, southeast of Beijing-Tianjin-Hebei, and north of Middle Reaches of Yangtze River had high PM2.5 concentrations. Among the urban agglomerations with high PM2.5 concentrations, Beijing-Tianjin-Hebei, Shandong Peninsula, and Central Plains were on the top three. These areas should focus more on eco-environmental protection during vigorous socioeconomic development.
The Moran’sI indexes in all the 19 urban agglomerations of China were greater than 0.700 during the period 2000-2015, indicating significant spatial agglomeration. The hot-spot areas were mainly concentrated to the east of the Hu Line, while the cold-spot areas were largely found to the west of this line. Overall, the proportion of cities in hot-spot areas continuously increased, while that in cold-spot areas decreased. The hot-spot areas were mainly in Beijing-Tianjin-Hebei, Shandong Peninsula, Central Plains, Middle Reaches of Yangtze River, and Yangtze River Delta. The pollution was more severe in the urban agglomerations of eastern China, especially northern China. Cold-spot areas were distributed in the Northern Slope of Tianshan Mountains, Lanzhou-Xining, Hohhot-Baotou-Erdos- Yulin, Cities in Ningxia along the Yellow River, Central Yunnan, and West Coast of the Taiwan Straits; in other words, the urban agglomerations of western, southwestern, and southeastern China had better air quality.
The PM2.5 concentrations in China’s urban agglomerations are characterized by relatively strong endogenetic interactions in space. Establishing cross-regional joint prevention and control mechanisms to better control this pollution is a matter of urgency. Industrialization and energy consumption have positive correlations with PM2.5 concentrations; foreign investment in coastal urban agglomerations and those near the borders have negative correlations, which is consistent with the pollution halo hypothesis. Other factors affecting PM2.5 concentrations in urban agglomerations show spatial disparities. For example, population density has a positive correlation with PM2.5 concentration within a region, but it has a negative correlation in the surrounding areas; urbanization rate has negative correlations with PM2.5 concentrations in national urban agglomerations, but has positive ones on regional and local urban agglomerations; high degree of industrial structure has negative correlations on PM2.5 concentrations within a region but has positive ones on its surrounding areas; technical support’s effects on PM2.5 concentrations are very large, but it is uncertain whether the effects are positive or negative. We need to increase the technology innovation input and strictly regulate the rapid marketization of new technologies.

5.2 Discussion

Since 1978, the rapid urbanization and industrialization of China’s urban agglomerations have resulted in many environmental issues, especially haze pollution. Urban agglomerations have become areas with high PM2.5 concentrations. This study examined PM2.5 pollution in China’s urban agglomerations, its patterns of spatial clustering, and its evolution. This work has clarified the main factors influencing PM2.5 pollution and analyzed its mechanism from the viewpoint of human factors. This may help identifying the reasons for high PM2.5 concentrations, and it is also of great significance for taking targeted control and prevention measures, as well as for scenario stimulations, risk warnings, planning and layout of industries, and healthy development of cities. The improvement of air quality and living conditions requires the participation of different parties of governments at all levels, as well as enterprises, scholars, and the public. It will also be a difficult and protracted battle. Although this study has focused more on human factors, we should not ignore the vital effects of natural factors, and these need further study. Models for the transmission paths of PM2.5 should be established and circulation corridors should be constructed.
The prevention of PM2.5 pollution is a complicated project that requires the cooperation of different departments, organizations, and scholars in various disciplines and fields. This paper discloses the spatiotemporal evolution patterns of PM2.5 pollution and its influencing factors in different urban agglomerations from a macroscopic perspective. In future research, more attention should be paid to the mechanisms from microcosmic perspectives, such as at the levels of urban agglomerations, provinces, prefectures, and counties. It is important to establish the characteristics of PM2.5 pollution and enact policies according to the different situations in different cities and urban agglomerations.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Cao C, Jiang W, Wang B et al., 2014. Inhalable microorganisms in Beijing’s PM2.5 and PM 10 pollutants during a severe smog event . Environmental Science & Technology, 48(3):1499-1507.

DOI

[2]
Cheng Y, Wang Z, Ye X et al., 2014. Spatiotemporal dynamics of carbon intensity from energy consumption in China. Journal of Geographical Sciences, 24(4):631-650.

DOI

[3]
Cheng Zhonghua, Liu Jun, Li Lianshui, 2019. Research on the effects of industrial structure adjustment and technical progress on haze reduction. China Soft Science, (1):146-154. (in Chinese)

[4]
Chowdhury Z, Zheng M, Schauer J J et al., 2007. Speciation of ambient fine organic carbon particles and source apportionment of PM 2.5 in Indian cities . Journal of Geophysical Research: Atmospheres, 112(D15)

[5]
Eeftens M, Beelen R, de Hoogh K et al., 2012. Development of land use regression models for PM 2.5, PM 2.5absorbancePM 10 and PM coarse in 20 European study areas: Results of the ESCAPE project. Environmental Science & Technology, 46(20):11195-11205.

DOI

[6]
Elhorst, Paul J , 2010. Applied spatial econometrics: Raising the bar. Spatial Economic Analysis, 5(1):9-28.

DOI

[7]
Fang Chuanglin, 2015. Important progress and future direction of studies on China’s urban agglomerations. Journal of Geographical Sciences, 25(8):1003-1024.

DOI

[8]
Fang Chuanglin, 2018. Important progress and prospects of China’s urbanization and urban agglomeration in the past 40 years of reform and opening-up. Economic Geography, 38(9):1-9. (in Chinese)

DOI

[9]
Fang Chuanglin, Liu Haimeng, Li Guangdong, 2016a. International progress and evaluation on interactive coupling effects between urbanization and the eco-environment. Journal of Geographical Sciences, 26(8):1081-1116.

DOI

[10]
Fang Chuanglin, Mao Qizhi, Ni Pengfei, 2015. The controversy and exploration of scientific selection and grading development of Chinese urban agglomeration. Acta Geographica Sinica, 70(4):515-527. (in Chinese)

[11]
Fang Chuanglin, Wang Zhenbo, Xu Guang et al., 2016b. Spatial-temporal characteristics of PM 2.5 in China: A city-level perspective analysis. Journal of Geographical Sciences, 26(11):1519-1532.

DOI

[12]
Franklin M, Zeka A, Schwartz J, 2007. Association between PM 2.5 and all-cause and specific-cause mortality in 27 US communities. Journal of Exposure Science and Environmental Epidemiology, 17(3):279.

DOI

[13]
Guan Wei, Xu Shuting , 2015. Spatial energy efficiency patterns and the coupling relationship with industrial structure: A study on Liaoning Province, China. Journal of Geographical Sciences, 25(3):355-368.

DOI

[14]
He Xiang, Lin Zhenshan, 2017. Interactive effects of the influencing factors on the changes of PM2.5 concentration based on GAM model . Environmental Science, (1):22-32. (in Chinese)

[15]
He Xiang, Lin Zhenshan, Liu Huiyu et al., 2016. Analysis of the driving factors of PM 2.5 in Jiangsu province based on grey correlation model. Acta Geographica Sinica, 71(7):1119-1129. (in Chinese)

[16]
Hu M, Lin J, Wang J et al., 2013. Spatial and temporal characteristics of particulate matter in Beijing, China using the empirical mode decomposition method. Science of the Total Environment, 458(3):70-80.

[17]
Huang Xiaogang, Zhao Jingbo, Cao Junji et al., 2018. Spatial-temporal variation of ozone concentration and its driving factors in China. Environmental Science, 40(3):1120-1131. (in Chinese)

DOI

[18]
Jiang Lei, Zhou Haifeng, Bo Ling , 2018. Spatial heterogeneity analysis of impacts of foreign direct investment on air pollution: Empirical evidence from 150 cities in China based on AQI. Scientia Geographica Sinica, 38(3):351-360. (in Chinese)

DOI

[19]
Kloog I, Nordio F, Coull B A et al., 2012. Incorporating local land use regression and satellite aerosol optical depth in a hybrid model of spatio-temporal PM 2.5 exposures in the Mid-Atlantic states. Environmental Science & Technology, 46(21):11913-11921.

DOI

[20]
Kourtchev I, Warnke J, Maenhaut W et al., 2008. Polar organic marker compounds in PM 2.5 aerosol from a mixed forest site in western Germany. Chemosphere, 73(8):1308-1314.

DOI PMID

[21]
Lee H J, Liu Y, Coull B A et al., 2011. A novel calibration approach of MODIS AOD data to predict PM 2.5 concentrations. Atmospheric Chemistry and Physics, 11(15):7991.

DOI

[22]
Lee P K H, Brook J R, Dabek-Zlotorzynska E et al., 2003. Identification of the major sources contributing to PM 2.5 observed in Toronto. Environmental Science & Technology, 37(21):4831-4840.

DOI

[23]
Li R, Hardy R, Zhang W et al., 2018. Chemical characterization and source apportionment of PM 2.5 in a Nonattainment Rocky Mountain Valley. Journal of Environmental Quality, 47(2):238.

DOI

[24]
Liu Haimeng, Fang Chuanglin, Huang Jiejun et al., 2018. The spatial-temporal characteristics and influencing factors of air pollution in Beijing-Tianjin-Hebei urban agglomeration. Acta Geographica Sinica, 73(1):177-191. (in Chinese)

[25]
Liu H M, Fang C L, Zhang X et al., 2017. The effect of natural and anthropogenic factors on haze pollution in Chinese cities: A spatial econometrics approach. Journal of Cleaner Production, 165:323-333.

DOI

[26]
Liu Huajun, Lei Mingyu, 2018. The dilemma of the regional collaborative governance of haze pollution and its solution ideas in China. China Population, Resources and Environment, 28(10):88-95. (in Chinese)

[27]
Massimiliano B, Silvia G, Alice C et al., 2018. Is there a link between air pollution and mental disorders. Environment International, 118:154-168.

DOI

[28]
Ministry of Environmental Protection MEP of the People’s Republic of China. Ambient Air Quality Standards (on trial), 2012. National Environmental Protection Standards of the People’s Republic of China (GB3095-2012), 2012-02-29. (in Chinese)

[29]
Mu Mu, Zhang Renhe , 2014. Addressing the issue of fog and haze: A promising perspective from meteorological science and technology. Science China Earth Sciences, 57(1):1-2.

DOI

[30]
Pinto J P, Lefohn A S, Shadwick D S, 2004. Spatial variability of PM 2.5 in urban areas in the United States. Journal of the Air & Waste Management Association, 54(4):440-449.

[31]
Santibañez D A, Ibarra S, Matus P, 2013. A five-year study of particulate matter (PM 2.5) and cerebrovascular diseases. Environmental Pollution, 181:1-6.

DOI

[32]
Song Y, Xie S, Zhang Y et al., 2006. Source apportionment of PM 2.5 in Beijing using principal component analysis absolute principal component scores and UNMIX. Science of the Total Environment, 372(1):278-286.

PMID

[33]
Sun Y, Zhuang G, Tang A et al., 2006. Chemical characteristics of PM 2.5 and PM 10 in haze fog episodes in Beijing. Environmental Science & Technology, 40(10):3148-3155.

DOI

[34]
Tecer L H, Alagha O, Karaca F, 2008. Particulate matter (PM 2.5, PM 10-2.5, and PM 10) and children’s hospital admissions for asthma and respiratory diseases: A bidirectional case-crossover study. Journal of Toxicology and Environmental Health (Part A), 71(8):512-520.

[35]
Tian S, Pan Y, Liu Z et al., 2014. Size-resolved aerosol chemical analysis of extreme haze pollution events during early 2013 in urban Beijing, China. Journal of Hazardous Materials, 279:452-460.

DOI

[36]
Walter I W, Ugelow J, 1979. Environmental policies in developing countries. Ambio, 8(2/3):102-109.

[37]
Wang Jing, Fang Chuanglin, 2011. A new driving force for the development of urban agglomerations in China. Geographical Research, 30(2):335-347. (in Chinese)

[38]
Wang Zhenbo, Fang Chuanglin, 2016. Spatial-temporal characteristics and determinants of PM2.5 in the Bohai Rim Urban Agglomeration. Chemosphere, 148:148-162.

DOI

[39]
Wang Zhenbo, Liang Longwu, Fang Chuanglin et al., 2018. Study of the evolution and factors influencing ecological security of the Beijing-Tianjin-Hebei Urban Agglomeration. Acta Ecologica Sinica, 38(12):4132-4144. (in Chinese)

[40]
Wang Zhenbo, Liang Longwu, Lin Xiongbin et al., 2017. Control models and effect evaluation of air pollution in Jing-Jin-Ji urban agglomeration. Environmental Science, 38(10):4005-4014. (in Chinese)

[41]
Wang Zhenbo, Liang Longwu, Sun Zhan et al., 2019. Spatiotemporal differentiation and the factors influencing urbanization and ecological environment synergistic effects within the Beijing-Tianjin-Hebei urban agglomeration. Journal of Environmental Management, 243:227-239.

DOI PMID

[42]
Wu Jiansheng, Wang Qian, Li Jiangcheng et al., 2017. Comparison of models on spatial variation of PM 2.5 concentration: A case of Beijing-Tianjin-Hebei region. Environmental Science, 38(6):2191-2201. (in Chinese)

[43]
Xi Qiangmin, Li Guoping, 2015. Characteristics and spillover effects of space division of producer service in the Beijing-Tianjin-Hebei metropolitan region: Based on spatial panel model. Acta Geographica Sinica, 70(12):1926-1938. (in Chinese)

[44]
Xiong Huanhuan, Liang Longwu, Zeng Zeng et al., 2017. Dynamic analysis of PM 2.5 spatial-temporal characteristics in China. Resources Science, 39(1):136-146. (in Chinese)

[45]
Xu Chao, Wang Yunpeng, Li Lili, 2018. Study on spatiotemporal distribution of PM 2.5 in China and its relationship to energy consumption based the remote sensing data from 1998 to 2012. Ecological Science, 37(1):108-120. (in Chinese)

[46]
Xu Helian, Deng Yuping, 2012. Does foreign direct investment lead to environmental pollution in China? Spatial measurement research based on Chinese provincial panel data. Management World, (2):30-43. (in Chinese)

[47]
Yan Yaxue, Qi Shaozhou, 2017. Time-space effect test on foreign direct investment and PM 2.5 pollution at city level . China Population, Resources and Environment, 27(4):68-77. (in Chinese)

[48]
Yang Mian, Wang Yin, 2017. Spatial-temporal characteristics of PM 2.5 and its influencing factors in the Yangtze River Economic Belt. China Population, Resources and Environment, 27(1):91-100. (in Chinese)

[49]
Zhou Liang, Zhou Chenghu, Yang Fan et al., 2019. Spatio-temporal evolution and the influencing factors of PM 2.5 in China between 2000 and 2015. Journal of Geographical Sciences, 29(2):253-270.

DOI

[50]
Zhou Zhen, Xing Yaoyao, Sun Hongxia et al., 2017. Interval game analysis of government subsidies on haze government in Beijing-Tianjin-Hebei. Systems Engineering-Theory & Practice, 37(10):2640-2648. (in Chinese)

Options
Outlines

/

〈 〉