Measuring the Urban Particulate Matter Island Effect with Rapid Urban Expansion

International Journal of Environmental Research and Public Health Article Measuring the Urban Particulate Matter Island E ff ect with Rapid Urban Expansion Yu Cao , Xiaoqian Fang, Jiayi Wang, Guoyu Li, Yu Cao and Yan Li * Department of Land Management, School of Public A ff airs, Zhejiang University, Hangzhou 310058, China; caoyu@zju.edu.cn (Y.C.); fangxiaoqian@zju.edu.cn (X.F.); 21822128@zju.edu.cn (J.W.); liguoyu@zju.edu.cn (G.L.); caoyu 98@zju.edu.cn (Y.C.) * Correspondence: liyan 522@zju.edu.cn; Tel.: + 86-571-56662139 Received: 24 June 2020; Accepted: 27 July 2020; Published: 31 July 2020 Abstract: Rapid urbanization has posed numerous negative impacts on the environment, including fine particulate matter (PM 2.5 ) pollution However, quantitative investigations of the PM 2.5 concentration trends over an urban-rural gradient at the local level are still lacking. The urban particulate matter island (UPI) e ff ect, representing the phenomenon that high particle concentrations in urban areas are gradually attenuated to surrounding areas, was adopted and modified in this paper to study the Hangzhou Bay area from 2000 to 2015. We found the following: (1) every urban area in the Hangzhou Bay area experienced rapid expansion, especially during 2000–2005; (2) more than half of the urban areas su ff ered UPI problems, and these urban areas had relatively high and stable UPI intensity (UPII) values, although the UPI footprint (UPIFP) values decreased with urban expansion; and (3) urban areas could be divided into three categories: plain areas, hilly areas and the junction of plains and hills, and the probability of the UPI e ff ect varied significantly for di ff erent categories. This paper can compensate for the lack of research on the UPI e ff ect at the local level and provide scientific evidence for air pollution control during urban agglomeration planning Keywords: urban particulate matter island (UPI) e ff ect; air pollution; urban expansion; Hangzhou Bay area 1. Introduction Urbanization, a major anthropogenic modification of Earth, occurs at an unprecedented speed in recorded human history worldwide [ 1 , 2 ]. According to the Department of Economic and Social A ff airs of the United Nations Secretariat, more than half of the global population currently lives in urban areas (55%, up from 30% in 1950), and by the middle of the century, the proportion is projected to reach 68% Generally, no other anthropogenic activity alters the environment more radically and persistently than urbanization, and air pollution [ 3 ], the urban heat island (UHI) e ff ect [ 4 ], global climate change [ 5 ], etc. have been proven to be related to rapid urbanization. Among these alarming problems, fine particulate matter with an aerodynamic diameter smaller than 2.5 µ m (PM 2.5 ) has elicited considerable public attention because of its harmful e ff ects on public health. There is convincing evidence that PM 2.5 can cause respiratory and cardiovascular problems, lung disease and other health problems [ 6 , 7 ]. Therefore, obtaining a better understanding of PM 2.5 pollution control is of great importance. Many studies have investigated meteorological [ 8 , 9 ] and socioeconomic factors [ 10 , 11 ] that may influence the spatiotemporal dynamics of the PM 2.5 concentration At the same time, urban land has extensively expanded outward across global cities to meet the needs of the soaring number of city dwellers [ 1 ]. A worldwide meta-analysis performed by Seto et al. [ 12 ] indicated that the urban land area had increased by 58,000 km 2 from 1970 to 2000, with the highest rate of urban expansion observed in India, China, and Africa. Spatiotemporal analysis Int. J. Environ. Res. Public Health 2020 , 17 , 5535; doi:10.3390 / ijerph 17155535 www.mdpi.com / journal / ijerph

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 2 of 18 of the urban expansion pattern and dynamics is important not only in urban geography but also in urban planning, environmental, and sustainability studies [ 13 ]. Many methods for measuring urban expansion have been proposed. Among these methods, quantitative metrics generated through the size and distribution of new urban land based on land cover maps have been widely applied to assess the physical growth of cities [ 14 , 15 ]. It has been proven that rapid urban expansion exerts many negative impacts on the environment that can transcend far beyond its physical boundary [ 16 ]. Among these impacts, the Urban Heat Island (UHI) e ff ect, which is the phenomenon whereby urban areas tend to have a higher temperature than surrounding areas, has been relatively widely researched [ 2 ]. In addition to heat energy, cities are also considered a major container of air pollutants. A previous study found that PM 2.5 pollution presented similar characteristics to the temperature; thus, a new concept—the urban particulate matter island (UPI) e ff ect—was introduced. Its definition is the phenomenon whereby a higher particle concentration is observed in urban areas that is gradually attenuated to surrounding areas. In addition, two metrics—UPII (UPI intensity) and UPIFP (UPI footprint)—were designed to quantitatively delineate the UPI phenomenon of 338 prefectures throughout China [ 17 ]. PM 2.5 data can be divided into two categories—ground monitoring site data and remote sensing data—according to the data source. For the data obtained from ground monitoring sites in China, the PM 2.5 concentration was incorporated in the monitoring system by the National Atmospheric Environment Monitoring System in 2012, and national monitoring of the PM 2.5 concentration was absent until 2013. Therefore, it is not feasible to use ground observation data for studies on the PM 2.5 pollution in China before 2013. In addition, due to the uneven coverage of the ground monitoring sites, limited data are available for studying the spatial and temporal distribution features of PM 2.5 . Luckily, satellite-based observations can complement the existing network of ground monitoring stations. PM 2.5 concentrations, however, cannot be measured directly by satellite imagery. Numerous studies have demonstrated that the aerosol optical depth (AOD) obtained by using satellite instruments, such as Moderate-resolution Imaging Spectroradiometer (MODIS) and Multi-angle Imaging Spectroradiometer (MISR), exhibited a notable correlation with surface PM 2.5 concentrations through statistical and mathematical models [ 18 – 21 ]. Owing to the improvements in remote sensing technology, the PM 2.5 concentration data estimated by satellite AOD products have been widely used in research [ 22 – 24 ]. However, the rationale of the UPI e ff ect and the interaction between cities have been ignored in previous research. To improve the model of the UPI e ff ect, three areas (urban, surrounding, and background areas) and three PM 2.5 processes (emission, dispersion, and deposition) are investigated in this paper. Therefore, all analyses of the UPI e ff ect can be pinpointed to specific areas and processes The surrounding and background areas constantly changed with the urban area due to urban expansion Moreover, studies on the UPI e ff ect at the local level are still lacking. To bridge this gap, the Hangzhou Bay area, an urban agglomeration that has experienced rapid urbanization and has su ff ered severe PM 2.5 pollution over the past two decades, was selected as the study area. Considering that the county or district of every prefecture is the most basic economic and social unit in China, the UPI e ff ect of every county or district in the Hangzhou Bay area was studied in this paper. The Hangzhou Bay area is an area with many cultural heritages, such as the West Lake, Liangzhu cultural sites, and numerous museums. PM 2.5 pollution can have a certain impact on cultural heritage. Therefore, this paper analyzed the pollution island e ff ect of the Hangzhou Bay area on a relatively macro scale. In this study, a total of 27 urban areas were extracted from the Hangzhou Bay area in 2000, 2005, 2010, and 2015. Then, the occurrence of the UPI e ff ect was determined—the urban areas whose PM 2.5 concentration gradually attenuated to the surrounding bu ff er zones were regarded as areas experiencing the UPI e ff ect. In addition, the spatiotemporal dynamics of UPII and UPIFP were analyzed. Finally, we evaluated the results in terms of the emission, dispersion, and deposition processes. This study aims to answer the following three questions: (1) does the UPI e ff ect exist in the Hangzhou Bay area? (2) If so, what is its magnitude and extent? (3) Is there any regularity in the UPI e ff ect?

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 3 of 18 By dividing the study area into urban, surrounding, and background areas, this paper regarded the study area as a whole. By dividing the dynamic processes of pollutants into emission, dispersion, and deposition, this paper analyzed the UPI e ff ect at the process level. By choosing an urban agglomeration as the study area, the UPI e ff ect was studied at the local level. Although the Hangzhou Bay area is just a small part of eastern China, it is a typical urban agglomeration su ff ering severe PM 2.5 pollution, with both plains and hills. Hence, this paper is not just a case study in the Hangzhou Bay area, because the methods can be widely applied to air pollution control during urban agglomeration planning. Compared with existing models, the modified model in this paper can help to analyze the spatial heterogeneity of the UPI e ff ect from the mechanism perspective 2. Study Area and Materials 2.1. Study Area Hangzhou Bay is in the coastal area of southeastern China (Figure 1 a), which is both the bay and the entrance to the Qiantang River. In this paper, the Hangzhou Bay area refers to five prefecture-level cities in Zhejiang Province (including Hangzhou, Ningbo, Shaoxing, Jiaxing, and Huzhou, as shown in Figure 1 b), which is part of the Yangtze River Delta region. According to the data provided by the Zhejiang Statistics Bureau, the Hangzhou Bay area covers an area of 44,988 km 2 in 2018, with a total resident population of 30.796 million. As part of the northern subtropical region, the Hangzhou Bay area is characterized by a mild and humid climate with abundant rainfall. The annual average temperature in the Hangzhou Bay area is between 15 and 20 ◦ C. The landforms in the Hangzhou Bay area are diverse, dominated by hills and coastal plains, and the overall terrain is high in the west and low in the east. The land use of the Hangzhou Bay area in 2015 indicates that construction land was mainly concentrated in the coastal plain (Figure 1 c). Int. J. Environ. Res. Public Health 2020 , 17 , x 3 of 17 By dividing the study area into urban, surrounding, and background areas, this paper regarded the study area as a whole. By dividing the dynamic processes of pollutants into emission, dispersion, and deposition, this paper analyzed the UPI effect at the process level. By choosing an urban agglomeration as the study area, the UPI effect was studied at the local level. Although the Hangzhou Bay area is just a small part of eastern China, it is a typical urban agglomeration suffering severe PM 2.5 pollution, with both plains and hills. Hence, this paper is not just a case study in the Hangzhou Bay area, because the methods can be widely applied to air pollution control during urban agglomeration planning. Compared with existing models, the modified model in this paper can help to analyze the spatial heterogeneity of the UPI effect from the mechanism perspective. 2. Study Area and Materials 2.1. Study Area Hangzhou Bay is in the coastal area of southeastern China (Figure 1 a), which is both the bay and the entrance to the Qiantang River. In this paper, the Hangzhou Bay area refers to five prefecturelevel cities in Zhejiang Province (including Hangzhou, Ningbo, Shaoxing, Jiaxing, and Huzhou, as shown in Figure 1 b), which is part of the Yangtze River Delta region. According to the data provided by the Zhejiang Statistics Bureau, the Hangzhou Bay area covers an area of 44,988 km 2 in 2018, with a total resident population of 30.796 million. As part of the northern subtropical region, the Hangzhou Bay area is characterized by a mild and humid climate with abundant rainfall. The annual average temperature in the Hangzhou Bay area is between 15 and 20 °C. The landforms in the Hangzhou Bay area are diverse, dominated by hills and coastal plains, and the overall terrain is high in the west and low in the east. The land use of the Hangzhou Bay area in 2015 indicates that construction land was mainly concentrated in the coastal plain (Figure 1 c). As one of the demonstration areas of the reform and opening of China, its geographical conditions and policies make the Hangzhou Bay area an important growth pole for economic and social development not only in the Yangtze River Delta region but also in China as a whole [25]. However, in addition to rapid economic growth and a high urbanization rate, the Hangzhou Bay area has also experienced severe PM 2.5 pollution in recent years [11]. In this study, the Hangzhou Bay area was selected as the study area to examine the spatiotemporal dynamics of the UPI effect within the context of rapid urban expansion. Figure 1. ( a ) Geographical location of the Hangzhou Bay area; ( b ) five prefecture-level cities in the Hangzhou Bay area; ( c ) land use of the Hangzhou Bay area in 2015. Figure 1. ( a ) Geographical location of the Hangzhou Bay area; ( b ) five prefecture-level cities in the Hangzhou Bay area; ( c ) land use of the Hangzhou Bay area in 2015 As one of the demonstration areas of the reform and opening of China, its geographical conditions and policies make the Hangzhou Bay area an important growth pole for economic and social development not only in the Yangtze River Delta region but also in China as a whole [ 25 ]. However, in addition to rapid economic growth and a high urbanization rate, the Hangzhou Bay area has also experienced severe PM 2.5 pollution in recent years [ 11 ]. In this study, the Hangzhou Bay area was selected as the study area to examine the spatiotemporal dynamics of the UPI e ff ect within the context of rapid urban expansion.

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 4 of 18 2.2. Data Source Several datasets, including PM 2.5 data, land use data, administrative boundaries, historical Google Earth (Google, Mountain View, CA, USA) images, and digital elevation model (DEM, NASA, Washington, DC, USA) data, are listed in Table 1 . The annual mean PM 2.5 data were retrieved from the Atmospheric Composition Analysis Group at Dalhousie University ( http: // fizz.phys.dal.ca / ~{}atmos / martin / ). The dataset was obtained by combining the aerosol optical depth (AOD, NASA, Washington, DC, USA) retrievals from NASA Moderate-resolution Imaging Spectroradiometer (MODIS, NASA, Washington, DC, USA), Multi-angle Imaging Spectroradiometer (MISR, NASA, Washington, DC, USA), and Sea-viewing Wide Field-of-view Sensor (SeaWiFS, NASA, Washington, DC, USA) instruments with the GEOS-Chem (NASA, Washington, DC, USA) chemical transport model and subsequently calibrated against regional ground-based observations of both the total and compositional mass via geographically weighted regression (GWR, ESRI, Redlands, CA, USA). The ground-based PM 2.5 measurements of mainland China were obtained from http: // beijingair.sinaapp.com / , which were captured by individuals from instantaneous data records on the website of the Chinese Environmental Protection Agency [ 26 , 27 ]. Table 1. Description of the data used in this study Data Data Description Source PM 2.5 data Annual mean PM 2.5 concentration in China with a spatial resolution of 0.01 ◦ × 0.01 ◦ in 2000–2015 http: // fizz.phys.dal.ca / ~{}atmos / martin / Land use data Land use data in 2000, 2005, 2010, and 2015 with a spatial resolution of 30 m × 30 m http: // nnu.geodata.cn:8008 / Administrative boundaries Shape files of prefectures and counties in Zhejiang Province http: // www.resdc.cn / Historical Google Earth images Google Earth images of the Hangzhou Bay area in 2000, 2005, 2010, and 2015 Download the software from https: // www. google.com / earth / versions / #earth-pro and install it, then follow these instructions: https: // support.google.com / earth / answer / 148094?hl = en / Digital elevation model (DEM) data ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) GDEM (Global Digital Elevation Model) data with a spatial resolution of 30 m × 30 m https: // lpdaac.usgs.gov / The dataset has been documented with a suitable accuracy on a global scale (the cross-validated R 2 value is 0.81) [ 28 ] and has been widely used [ 17 , 29 ]. Therefore, a subset covering the Hangzhou Bay area in 2000, 2005, 2010, and 2015 was adopted in this study. The data downloaded from the website have the World Geodetic System—the 1984 Coordinate System—with a spatial resolution of 0.01 ◦ . The projection transformation, extraction, and zonal statistics tools in geographical information systems (ESRI, Redlands, CA, USA) were implemented to successively process the annual mean PM 2.5 concentration data Land use data and historical Google Earth images were used to extract the urban area, and historical Google Earth images were supplements to land use data. Unlike the other datasets, the source of historical Google Earth images is the Google Earth Pro software (Google, Mountain View, CA, USA). Moreover, administrative boundaries data and digital elevation model (DEM) data were used in this study 3. Methods 3.1. Rationale The rationale of this paper can be summarized as the emission, dispersion, and deposition of PM 2.5 . Principal component analysis performed on the aerosol data has revealed that vehicular emissions, coal / biomass combustion, industry sources, soil dust, and secondary formation are the main

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 5 of 18 potential sources for the ionic components of PM 2.5 in China [ 30 ]. The economic growth, urbanization, and industrialization of China have turned urban areas into the main sources and victims of air pollution [ 31 ]. By emitting the PM 2.5 pollutant into the atmosphere, the process of dispersion results in transboundary air pollution. Dispersion is always accompanied by deposition, which is a process that removes pollutants from the atmosphere and transfers them to the ground surface. After emission, some of the particulates settle due to gravitation, some are eluted from the atmosphere by precipitation, and others remain in the atmosphere for a long time. It has been assumed that meteorological conditions, particle properties, and underlying surface characteristics can a ff ect the dispersion and deposition of PM 2.5 [ 32 ]. The UPI e ff ect, referring to the phenomenon whereby the high PM 2.5 concentrations in urban areas are gradually attenuated to surrounding areas, is essentially a comprehensive manifestation of the emission, dispersion and deposition of PM 2.5 (Figure 2 b). Int. J. Environ. Res. Public Health 2020 , 17 , x 5 of 17 3. Methods 3.1. Rationale The rationale of this paper can be summarized as the emission, dispersion, and deposition of PM 2.5 . Principal component analysis performed on the aerosol data has revealed that vehicular emissions, coal/biomass combustion, industry sources, soil dust, and secondary formation are the main potential sources for the ionic components of PM 2.5 in China [30]. The economic growth, urbanization, and industrialization of China have turned urban areas into the main sources and victims of air pollution [31]. By emitting the PM 2.5 pollutant into the atmosphere, the process of dispersion results in transboundary air pollution. Dispersion is always accompanied by deposition, which is a process that removes pollutants from the atmosphere and transfers them to the ground surface. After emission, some of the particulates settle due to gravitation, some are eluted from the atmosphere by precipitation, and others remain in the atmosphere for a long time. It has been assumed that meteorological conditions, particle properties, and underlying surface characteristics can affect the dispersion and deposition of PM 2.5 [32]. The UPI effect, referring to the phenomenon whereby the high PM 2.5 concentrations in urban areas are gradually attenuated to surrounding areas, is essentially a comprehensive manifestation of the emission, dispersion and deposition of PM 2.5 (Figure 2 b). Figure 2. Model of the urban particulate matter island (UPI) effect. ( a ) Division of the study area; ( b ) the three processes of particulate emission, dispersion, and deposition; ( c ) evaluation criterion and two metrics of the UPI effect. 3.2. Extraction of the Urban and Surrounding Areas To determine the occurrence of the UPI effect, the study area was divided into urban, surrounding, and background areas (Figure 2 a). Considering that the county/district of every prefecture-level city or municipality is the most basic economic and social unit in China, the urban area in this paper was defined as the urban land of every county or district in these five prefecturelevel cities. Based on the land use data and historical Google Earth images, the urban areas in 2000, 2005, 2010, and 2015 were identified and extracted. Specifically, for each urban area, eight buffer zones surrounding the urban area were generated using geographical information systems software, and each buffer zone covered half the size of the Figure 2. Model of the urban particulate matter island (UPI) e ff ect. ( a ) Division of the study area; ( b ) the three processes of particulate emission, dispersion, and deposition; ( c ) evaluation criterion and two metrics of the UPI e ff ect 3.2. Extraction of the Urban and Surrounding Areas To determine the occurrence of the UPI e ff ect, the study area was divided into urban, surrounding, and background areas (Figure 2 a). Considering that the county / district of every prefecture-level city or municipality is the most basic economic and social unit in China, the urban area in this paper was defined as the urban land of every county or district in these five prefecture-level cities. Based on the land use data and historical Google Earth images, the urban areas in 2000, 2005, 2010, and 2015 were identified and extracted Specifically, for each urban area, eight bu ff er zones surrounding the urban area were generated using geographical information systems software, and each bu ff er zone covered half the size of the urban area. The purpose of the equal-area bu ff er zone instead of the equal-distance bu ff er zone is to ensure that each urban area is comparable. Based on both, the size of each urban area and the distance among the urban areas, eight bu ff er zones eventually proved to be the best choice. However, some urban areas were geographically very close, leading to overlapping bu ff er zones. Instead of removing these urban areas directly from this experiment as in a previous study [ 17 ], the method of merging was

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 6 of 18 applied in this paper, which means that the urban areas with overlapping bu ff er zones were regarded as one urban area To illustrate the above more clearly, the eight districts of Hangzhou (Shangcheng, Xiacheng, Jianggan, Gongshu, Xihu, Binjiang, Xiaoshan, and Yuhang) were selected as an example (Figure 3 a). In 2000, Jianggan District and Yuhang District, located northeast of Hangzhou, were relatively far from the other six districts. However, with rapid urban expansion, overlapping bu ff er zones emerged in 2005. Therefore, all these districts were merged into one urban area. Based on this principle, there are 27 urban areas in the study area, and each urban area has its own corresponding number, such as 1-1 (the first number represents the prefecture-level city) (Figure 3 b). In this paper, these numbers were adopted to represent the 27 urban areas in the following sections Int. J. Environ. Res. Public Health 2020 , 17 , x 6 of 17 urban area. The purpose of the equal-area buffer zone instead of the equal-distance buffer zone is to ensure that each urban area is comparable. Based on both, the size of each urban area and the distance among the urban areas, eight buffer zones eventually proved to be the best choice. However, some urban areas were geographically very close, leading to overlapping buffer zones. Instead of removing these urban areas directly from this experiment as in a previous study [17], the method of merging was applied in this paper, which means that the urban areas with overlapping buffer zones were regarded as one urban area. To illustrate the above more clearly, the eight districts of Hangzhou (Shangcheng, Xiacheng, Jianggan, Gongshu, Xihu, Binjiang, Xiaoshan, and Yuhang) were selected as an example (Figure 3 a). In 2000, Jianggan District and Yuhang District, located northeast of Hangzhou, were relatively far from the other six districts. However, with rapid urban expansion, overlapping buffer zones emerged in 2005. Therefore, all these districts were merged into one urban area. Based on this principle, there are 27 urban areas in the study area, and each urban area has its own corresponding number, such as 1-1 (the first number represents the prefecture-level city) (Figure 3 b). In this paper, these numbers were adopted to represent the 27 urban areas in the following sections. Figure 3. Urban areas in the Hangzhou Bay area. ( a ) Urban areas and buffer zones in Hangzhou (urban 1-1) in 2000, 2005, 2010, and 2015; ( b ) location of the 27 urban areas. 3.3. Model of the UPI effect The average PM 2.5 concentration in each urban and buffer was calculated. For each urban area, Δ C i was defined as the PM 2.5 concentration difference between the i-th buffer zone C i and the outermost buffer zone C 8 (the 8 th buffer zone). It was calculated with the following Equation (1): Δ C i = C i − C 8 , i = {0, 1, 2…, 8} (1) where C 0 (when i = 0) is the PM 2.5 concentration of the urban area, and C i is the PM 2.5 concentration of the i-th buffer zone. Δ C is a series of differential values for the urban area and all its buffer zones. The trends of Δ C for each urban area in 2000, 2005, 2010, and 2015 were analyzed to examine the attenuation of the PM 2.5 concentration. The urban area whose PM 2.5 concentration was gradually attenuated from the urban area to the surrounding buffer zones was regarded as an urban area experiencing the UPI effect. Figure 3. Urban areas in the Hangzhou Bay area. ( a ) Urban areas and bu ff er zones in Hangzhou (urban 1-1) in 2000, 2005, 2010, and 2015; ( b ) location of the 27 urban areas 3.3. Model of the UPI E ff ect The average PM 2.5 concentration in each urban and bu ff er was calculated. For each urban area, ∆ C i was defined as the PM 2.5 concentration di ff erence between the i-th bu ff er zone C i and the outermost bu ff er zone C 8 (the 8 th bu ff er zone). It was calculated with the following Equation (1): ∆ C i = C i − C 8 , i = {0, 1, 2 . . , 8} (1) where C 0 (when i = 0) is the PM 2.5 concentration of the urban area, and C i is the PM 2.5 concentration of the i-th bu ff er zone ∆ C is a series of di ff erential values for the urban area and all its bu ff er zones. The trends of ∆ C for each urban area in 2000, 2005, 2010, and 2015 were analyzed to examine the attenuation of the PM 2.5 concentration. The urban area whose PM 2.5 concentration was gradually attenuated from the urban area to the surrounding bu ff er zones was regarded as an urban area experiencing the UPI e ff ect To quantitatively delineate the UPI e ff ect and its spatiotemporal pattern, two metrics have been defined. The first metric is the UPI intensity (UPII), which reflects the magnitude of the UPI e ff ect and can be formulated in Equation (2) as follows: UPII = ∆ C 0 / C 0 (2)

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 7 of 18 This means that the PM 2.5 concentration ratio decayed from the urban area to the outermost bu ff er zone. In this paper, 0.1 is set as the UPII threshold. Considering PM 2.5 standard from the World Health Organization (WHO), urban areas with C 0 > 10 µ g / m 3 and UPII > 0.1 are regarded as those with a high UPII. It is well known that urban pollution stems from both natural and anthropogenic sources [ 33 ]. Therefore, by removing the influence of the natural background, UPII can be viewed as the urban pollution caused by anthropogenic activities [ 17 ]. The second metric, the UPI footprint (UPIFP), was defined as the number of bu ff er zones across which the urban PM 2.5 concentration decreases by 5% (half of the UPII threshold) from the urban area It can be calculated in Equation (3) as follows: C UPIFP < C 0 × (1 − 5%) < C UPIFP-1 (3) UPIFP quantifies the extent of the continuous e ff ects from the urban area to the surrounding areas In this paper, UPIFP is an integer ranging from 1 to 8 (Figure 2 c). A higher UPIFP value indicates that urban pollution is transported to a further bu ff er zone. In other words, the higher the value of UPIFP is, the lower the degree of PM 2.5 accumulation in the urban area is compared to that in the surrounding area, and the less di ffi cult it is for PM 2.5 pollution to spread outward. In this paper, urban areas with UPIFP values higher than 4 are regarded as areas with a large pollution PM 2.5 e ff ect range The overall concept of this paper is shown in Figure 2 . Based on the rationale mentioned above, the method can be defined as according to several steps: (1) extract the urban areas; (2) generate the equal-area bu ff er zones as the surrounding areas; (3) calculate the average PM 2.5 concentration in each urban and surrounding area; (4) analyze the concentration di ff erence from the urban area to the outermost bu ff er zone; (5) determine whether the UPI phenomenon exists or not; and (6) if so, calculate the UPII and UPIFP of each urban area 4. Results 4.1. Spatial Patterns of Urban Expansion The urban expansion that occurred in the Hangzhou Bay area from 2000 to 2015 is shown in Figure 4 a. The total area of the extracted urban areas was 470.41 km 2 in 2000, 1087.96 km 2 in 2005, 1325.15 km 2 in 2010, and 1584.68 km 2 in 2015. The expansion that occurred from 2000 to 2005 occupied 55.42% of the total urban expansion from 2000 to 2015, making 2000–2005 the period with the most rapid urban growth in the Hangzhou Bay area, followed by 2010–2015 (23.29%), and the period with the least rapid growth was 2005–2010 (21.29%). Violin plots of the study area and multiple expanded areas are shown in Figure 5 a,b, respectively. Combining the kernel density estimation curve and box plot, the violin plots indicated that the median area increased from 2000 to 2015, while the median expanded area decreased over this period. Although violin plots are less common than box plots, they show similar information. In addition to the median and interquartile range, a violin plot shows the distribution of the data with a rotated smoothing curve To visualize the expansion of each urban area more clearly, the specific expansion of each urban area is shown in Figure 4 b (the scale of each urban area varies). Each urban area expanded from 2000 to 2015, but considering the expansion form and size, each urban area di ff ered. In terms of the expansion form, basically, every urban area expanded outward from the original urban area (referring to the urban area in 2000) in all directions over this period without terrain and hydrological restrictions. In addition, the area of each urban area was analyzed as well. According to the calculation of the area, urban 1-1 (which is the urban area whose corresponding number is 1-1, the same as below) possessed the largest area all the time, followed by urban 3-1. Four urban areas expanded more than three times from 2000 to 2005, i.e., urban 4-6 expanded by 348.31%, urban 4-2 expanded by 659.02%, urban 5-2 expanded by 369.92%, and urban 5-6 expanded by 449.12%.

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 8 of 18 Int. J. Environ. Res. Public Health 2020 , 17 , x 8 of 17 expanded more than three times from 2000 to 2005, i.e., urban 4-6 expanded by 348.31%, urban 4-2 expanded by 659.02%, urban 5-2 expanded by 369.92%, and urban 5-6 expanded by 449.12%. Figure 4. Urban expansion from 2000 to 2015 in the Hangzhou Bay area. ( a ) All urban areas; ( b ) each urban area. Figure 4. Urban expansion from 2000 to 2015 in the Hangzhou Bay area. ( a ) All urban areas; ( b ) each urban area.

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 9 of 18 Int. J. Environ. Res. Public Health 2020 , 17 , x 9 of 17 Figure 5. Statistics on the urban expansion of the 27 urban areas in the Hangzhou Bay area. ( a ) Area in 2000, 2005, 2010, and 2015; ( b ) multiple expanded areas during 2000–2005, 2005–2010, and 2010– 2015. 4.2. PM 2.5 Concentration in the Urban Areas and Buffer Zones The annual mean PM 2.5 concentration within each urban area in 2000, 2005, 2010, and 2015 is shown in Figure 6. From the perspective of the spatial distribution, both the northeast and southwest regions of the Hangzhou Bay area presented relatively low PM 2.5 concentrations. In terms of the PM 2.5 concentration of each urban area, the PM 2.5 concentration in urban 1-4, urban 1-5, urban 1-6, urban 3-4, urban 3-5, and urban 5-5 remained low from 2000 to 2015, while that in urban 1-1 remained high throughout in this period. In addition, the cluster formed by urban 4-1, urban 4-2, urban 4-3, urban 4- 4, urban 4-5, and urban 4-6 experienced a notable increase in PM 2.5 concentration from 2000 to 2015. These six urban areas are all located in Jiaxing and are geographically close to one another. By evaluating the PM 2.5 concentrations in these four years, the most significant point is that the PM 2.5 concentrations in 2000 were considerably lower than that in the other three years. The minimum concentration in 2000 was 25.81 μ g/m 3 , while it was 36.12 μ g/m 3 in 2005, 36.37 μ g/m 3 in 2010 and 34.16 μ g/m 3 in 2015. At the same time, the maximum concentration in 2000 was 54.03 μ g/m 3 , while it was 65.75 μ g/m 3 in 2005, 59.31 μ g/m 3 in 2010, and 58.57 μ g/m 3 in 2015. In summary, both the minimum and maximum concentrations were the lowest in 2000. Regarding each urban area, Δ C (the PM 2.5 concentration difference between the urban area/buffer zone and the outermost buffer zone) of the 27 urban areas in 2000, 2005, 2010, and 2015 are shown in Figure 7 a. In 15 of the 27 urban areas, Δ C clearly decayed from the urban area to the buffer zones, including urban 1-1, urban 1-2, urban 1-3, urban 2-1, urban 2-2, urban 2-3, urban 2-4, urban 3-1, urban 3-2, urban 3-3, urban 3-4, urban 3-5, urban 5-1, urban 5-4, and urban 5-6, which were regarded as urban areas with UPI effects in this paper. The distribution of the urban areas with the UPI effect is shown in Figure 7 b. Except for urban 2-2, for each of the remaining 14 urban areas, the shapes of the four curves are roughly similar. When Δ C of all 27 urban areas are combined (labeled “all” in Figure 7 a), the four decay curves for 2000, 2005, 2010, and 2015 can be obtained. Except for 2000, the curves for the other three years are very similar in both shape and value, and their Δ C 0 values are all higher than 4. The slope of the attenuation decreases with the distance in 2005, 2010, and 2015. Conversely, Δ C 0 in 2000 is 3.55, which is slightly lower than that in the other three years, and the slope of the attenuation remains essentially the same. Figure 5. Statistics on the urban expansion of the 27 urban areas in the Hangzhou Bay area. ( a ) Area in 2000, 2005, 2010, and 2015; ( b ) multiple expanded areas during 2000–2005, 2005–2010, and 2010–2015 4.2. PM 2.5 Concentration in the Urban Areas and Bu ff er Zones The annual mean PM 2.5 concentration within each urban area in 2000, 2005, 2010, and 2015 is shown in Figure 6 . From the perspective of the spatial distribution, both the northeast and southwest regions of the Hangzhou Bay area presented relatively low PM 2.5 concentrations. In terms of the PM 2.5 concentration of each urban area, the PM 2.5 concentration in urban 1-4, urban 1-5, urban 1-6, urban 3-4, urban 3-5, and urban 5-5 remained low from 2000 to 2015, while that in urban 1-1 remained high throughout in this period. In addition, the cluster formed by urban 4-1, urban 4-2, urban 4-3, urban 4-4, urban 4-5, and urban 4-6 experienced a notable increase in PM 2.5 concentration from 2000 to 2015. These six urban areas are all located in Jiaxing and are geographically close to one another By evaluating the PM 2.5 concentrations in these four years, the most significant point is that the PM 2.5 concentrations in 2000 were considerably lower than that in the other three years. The minimum concentration in 2000 was 25.81 µ g / m 3 , while it was 36.12 µ g / m 3 in 2005, 36.37 µ g / m 3 in 2010 and 34.16 µ g / m 3 in 2015. At the same time, the maximum concentration in 2000 was 54.03 µ g / m 3 , while it was 65.75 µ g / m 3 in 2005, 59.31 µ g / m 3 in 2010, and 58.57 µ g / m 3 in 2015. In summary, both the minimum and maximum concentrations were the lowest in 2000 Regarding each urban area, ∆ C (the PM 2.5 concentration di ff erence between the urban area / bu ff er zone and the outermost bu ff er zone) of the 27 urban areas in 2000, 2005, 2010, and 2015 are shown in Figure 7 a. In 15 of the 27 urban areas, ∆ C clearly decayed from the urban area to the bu ff er zones, including urban 1-1, urban 1-2, urban 1-3, urban 2-1, urban 2-2, urban 2-3, urban 2-4, urban 3-1, urban 3-2, urban 3-3, urban 3-4, urban 3-5, urban 5-1, urban 5-4, and urban 5-6, which were regarded as urban areas with UPI e ff ects in this paper. The distribution of the urban areas with the UPI e ff ect is shown in Figure 7 b. Except for urban 2-2, for each of the remaining 14 urban areas, the shapes of the four curves are roughly similar When ∆ C of all 27 urban areas are combined (labeled “all” in Figure 7 a), the four decay curves for 2000, 2005, 2010, and 2015 can be obtained. Except for 2000, the curves for the other three years are very similar in both shape and value, and their ∆ C 0 values are all higher than 4. The slope of the attenuation decreases with the distance in 2005, 2010, and 2015. Conversely, ∆ C 0 in 2000 is 3.55, which is slightly lower than that in the other three years, and the slope of the attenuation remains essentially the same.

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 10 of 18 Int. J. Environ. Res. Public Health 2020 , 17 , x 10 of 17 Figure 6. The annual mean fine particulate matter (PM 2.5 ) concentration in each urban area. 4.3. Spatiotemporal Dynamics of UPII and UPIFP The spatiotemporal dynamics of the UPII and UPIFP for the urban areas with UPI effects in the Hangzhou Bay area are shown in Figure 8. For comparison purposes, the UPII in the four years are all classified with 0.1, 0.15, and 0.2. Overall, the UPII in each urban area remained relatively stable over time (Figure 9). From the perspective of the spatial distribution, the urban areas with a higher UPII were distributed in the central and southeast regions of the Hangzhou Bay area. The urban area with the highest UPII was urban 3-4, whose UPII exceeded 0.2 every year. In addition, the UPII of urban 1-1, urban 1-2, urban 1-3, urban 2-1, urban 2-3, urban 2-4, urban 3-1, urban 3-2, urban 3-3, urban 3-5, and urban 5-6 were higher than 0.1 in these four years. In contrast to the UPII, not all urban areas have a UPIFP. Among the 15 urban areas, 14 urban areas have a UPIFP all four years, while urban 2-2 only has a UPIFP in 2005, 2010, and 2015. The UPIFP in 2000 includes 1, 2, 3, and 4. In 2005 and 2010, 1, 2, and 3 still occurred, while in 2015, only 1 and 2 remained. The most notable UPIFP phenomenon is that from 2000 to 2015, the UPIFP generally showed a downward trend, especially from 2000 to 2005, which implied that compared to the surrounding area, the PM 2.5 accumulation in the urban area became more severe in this period. Figure 6. The annual mean fine particulate matter (PM 2.5 ) concentration in each urban area 4.3. Spatiotemporal Dynamics of UPII and UPIFP The spatiotemporal dynamics of the UPII and UPIFP for the urban areas with UPI e ff ects in the Hangzhou Bay area are shown in Figure 8 . For comparison purposes, the UPII in the four years are all classified with 0.1, 0.15, and 0.2. Overall, the UPII in each urban area remained relatively stable over time (Figure 9 ). From the perspective of the spatial distribution, the urban areas with a higher UPII were distributed in the central and southeast regions of the Hangzhou Bay area. The urban area with the highest UPII was urban 3-4, whose UPII exceeded 0.2 every year. In addition, the UPII of urban 1-1, urban 1-2, urban 1-3, urban 2-1, urban 2-3, urban 2-4, urban 3-1, urban 3-2, urban 3-3, urban 3-5, and urban 5-6 were higher than 0.1 in these four years In contrast to the UPII, not all urban areas have a UPIFP. Among the 15 urban areas, 14 urban areas have a UPIFP all four years, while urban 2-2 only has a UPIFP in 2005, 2010, and 2015. The UPIFP in 2000 includes 1, 2, 3, and 4. In 2005 and 2010, 1, 2, and 3 still occurred, while in 2015, only 1 and 2 remained. The most notable UPIFP phenomenon is that from 2000 to 2015, the UPIFP generally showed a downward trend, especially from 2000 to 2005, which implied that compared to the surrounding area, the PM 2.5 accumulation in the urban area became more severe in this period.

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 11 of 18 Int. J. Environ. Res. Public Health 2020 , 17 , x 11 of 17 Figure 7. The occurrence of the UPI effect. ( a ) Trends of concentration difference ( Δ C) from the urban area to the buffer zones for the 27 urban areas in the Hangzhou Bay area in 2000, 2005, 2010, and 2015; ( b ) spatial distribution of the urban areas with the UPI effect. Figure 7. The occurrence of the UPI e ff ect. ( a ) Trends of concentration di ff erence ( ∆ C) from the urban area to the bu ff er zones for the 27 urban areas in the Hangzhou Bay area in 2000, 2005, 2010, and 2015; ( b ) spatial distribution of the urban areas with the UPI e ff ect.

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 12 of 18 Int. J. Environ. Res. Public Health 2020 , 17 , x 12 of 17 Figure 8. Spatial distribution of the UPI intensity (UPII) and UPI footprint (UPIFP) for the urban areas with UPI effects in 2000, 2005, 2010, and 2015. Figure 9. UPI intensity (UPII) and UPI footprint (UPIFP) in 2000, 2005, 2010, and 2015. 0 1 2 3 4 0 0.05 0.1 0.15 0.2 0.25 1-1 1-2 1-3 2-1 2-2 2-3 2-4 3-1 3-2 3-3 3-4 3-5 5-1 5-4 5-6 UPIFP UPII Urban UPIFP 2000 UPIFP 2005 UPIFP 2010 UPIFP 2015 UPII 2000 UPII 2005 UPII 2010 UPII 2015 Figure 8. Spatial distribution of the UPI intensity (UPII) and UPI footprint (UPIFP) for the urban areas with UPI e ff ects in 2000, 2005, 2010, and 2015 Int. J. Environ. Res. Public Health 2020 , 17 , x 12 of 17 Figure 8. Spatial distribution of the UPI intensity (UPII) and UPI footprint (UPIFP) for the urban areas with UPI effects in 2000, 2005, 2010, and 2015. Figure 9. UPI intensity (UPII) and UPI footprint (UPIFP) in 2000, 2005, 2010, and 2015. 0 1 2 3 4 0 0.05 0.1 0.15 0.2 0.25 1-1 1-2 1-3 2-1 2-2 2-3 2-4 3-1 3-2 3-3 3-4 3-5 5-1 5-4 5-6 UPIFP UPII Urban UPIFP 2000 UPIFP 2005 UPIFP 2010 UPIFP 2015 UPII 2000 UPII 2005 UPII 2010 UPII 2015 Figure 9. UPI intensity (UPII) and UPI footprint (UPIFP) in 2000, 2005, 2010, and 2015 5. Discussion 5.1. Emissions In this paper, urban areas are considered the source and container of PM 2.5 pollution, and the PM 2.5 concentration in the urban areas is directly related to local emissions. Because it is the result of

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 13 of 18 economic growth, urban expansion can reflect the development of cities to a certain extent. Pearson correlation analysis between the area and PM 2.5 concentration of the urban areas indicates that there exists a positive correlation between the area and concentration, especially when the area is smaller than 100 km 2 (Figure 10 ). This is consistent with previous observations [ 29 ]. However, the environmental Kuznets curve (EKC) hypothesis claims that there exists an inverted U-shaped connection between environmental degradation and economic growth [ 34 ], which means that the Hangzhou Bay area is still at the initial stage of economic growth. Therefore, although PM 2.5 pollution can be fundamentally reduced by controlling emissions in the short term, it comes at a significant cost and requires e ff orts from both the local government and polluting companies [ 35 ]. In addition, studies have shown that urban form can influence air pollution in China, and moderately scattered and polycentric urban development is encouraged [ 36 ]. Int. J. Environ. Res. Public Health 2020 , 17 , x 13 of 17 5. Discussion 5.1. Emissions In this paper, urban areas are considered the source and container of PM 2.5 pollution, and the PM 2.5 concentration in the urban areas is directly related to local emissions. Because it is the result of economic growth, urban expansion can reflect the development of cities to a certain extent. Pearson correlation analysis between the area and PM 2.5 concentration of the urban areas indicates that there exists a positive correlation between the area and concentration, especially when the area is smaller than 100 km 2 (Figure 10). This is consistent with previous observations [29]. However, the environmental Kuznets curve (EKC) hypothesis claims that there exists an inverted U-shaped connection between environmental degradation and economic growth [34], which means that the Hangzhou Bay area is still at the initial stage of economic growth. Therefore, although PM 2.5 pollution can be fundamentally reduced by controlling emissions in the short term, it comes at a significant cost and requires efforts from both the local government and polluting companies [35]. In addition, studies have shown that urban form can influence air pollution in China, and moderately scattered and polycentric urban development is encouraged [36]. Figure 10. Relationship between the area and PM 2.5 concentration in the urban areas. ( a ) All urban areas; ( b ) urban areas divided at an area of 100 km 2 . Figure 10. Relationship between the area and PM 2.5 concentration in the urban areas. ( a ) All urban areas; ( b ) urban areas divided at an area of 100 km 2 5.2. Dispersion and Deposition With increasing economic development and urbanization, the underlying surface of dispersion and deposition has changed. On the one hand, due to the rapid urban expansion, parts of the former surrounding area have become urban areas, and parts of the background area have been transformed into surrounding areas. On the other hand, the land use of the surrounding and background areas has also changed in this period, such as the transition from farmland to rural settlements. To further

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 14 of 18 examine the mechanism behind the UPI e ff ect, these 27 urban areas were divided into three categories (Figure 11 ). Int. J. Environ. Res. Public Health 2020 , 17 , x 14 of 17 5.2. Dispersion and Deposition With increasing economic development and urbanization, the underlying surface of dispersion and deposition has changed. On the one hand, due to the rapid urban expansion, parts of the former surrounding area have become urban areas, and parts of the background area have been transformed into surrounding areas. On the other hand, the land use of the surrounding and background areas has also changed in this period, such as the transition from farmland to rural settlements. To further examine the mechanism behind the UPI effect, these 27 urban areas were divided into three categories (Figure 11). Plain area: the urban areas were distributed in clusters with a high integrity. Due to the flat terrain, cities were generally densely distributed, and urban expansion resulted in a major reduction in the distance between neighboring urban areas, which might enhance the interaction of pollution. In addition, the surrounding and background areas were usually dominated by cultivated land and rural settlements. Therefore, the PM 2.5 concentration in the plain area was relatively even, and the probability of the UPI effect was relatively low in this area. Hilly area: the urban areas were scattered with notable independence. Because of the woodlands in the surrounding areas, the pollutants emitted from the urban areas could be effectively settled. In addition, the large forests in the background areas diminished the effect of urban expansion on the distance between the urban areas. However, due to the terrain, urban expansion was limited, which might make urban pollution accumulation less distinct. Therefore, the probability of the UPI effect was relatively random in this area, and the decrease in UPIFP might be due to the shorter distance between the urban boundaries and woodlands. Junction of plains and hills: the urban areas were zonally distributed, with both integrity and independence. Compared with the urban areas surrounded by hills, the expansion was relatively less restricted by the terrain but more susceptible to the influence of neighboring urban areas. Compared with the plain area, the surrounding woodlands could effectively settle pollutants. In this paper, the urban areas located in the junction area exhibited a high probability of the UPI effect, and the decrease in UPIFP was also due to the shorter distance between the urban boundaries and woodlands. Figure 11. Division of the urban areas by the terrain. Figure 11. Division of the urban areas by the terrain Plain area: the urban areas were distributed in clusters with a high integrity. Due to the flat terrain, cities were generally densely distributed, and urban expansion resulted in a major reduction in the distance between neighboring urban areas, which might enhance the interaction of pollution. In addition, the surrounding and background areas were usually dominated by cultivated land and rural settlements. Therefore, the PM 2.5 concentration in the plain area was relatively even, and the probability of the UPI e ff ect was relatively low in this area Hilly area: the urban areas were scattered with notable independence. Because of the woodlands in the surrounding areas, the pollutants emitted from the urban areas could be e ff ectively settled. In addition, the large forests in the background areas diminished the e ff ect of urban expansion on the distance between the urban areas. However, due to the terrain, urban expansion was limited, which might make urban pollution accumulation less distinct. Therefore, the probability of the UPI e ff ect was relatively random in this area, and the decrease in UPIFP might be due to the shorter distance between the urban boundaries and woodlands Junction of plains and hills: the urban areas were zonally distributed, with both integrity and independence. Compared with the urban areas surrounded by hills, the expansion was relatively less restricted by the terrain but more susceptible to the influence of neighboring urban areas. Compared with the plain area, the surrounding woodlands could e ff ectively settle pollutants. In this paper, the urban areas located in the junction area exhibited a high probability of the UPI e ff ect, and the decrease in UPIFP was also due to the shorter distance between the urban boundaries and woodlands 5.3. Main Findings In this paper, we proposed a framework to characterize the UPI e ff ect of the Hangzhou Bay area during 2000–2015. The urban areas experienced a rapid expansion during this period. However, the existence of the UPI e ff ect in each urban area di ff ered slightly. More than half of the urban areas

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 15 of 18 exhibited the UPI e ff ect, and the existence of the UPI e ff ect was not related to the PM 2.5 concentration, but instead to the terrain, and urban areas could be classified into three categories: plain areas, hilly areas, and the junction of plains and hills. Terrain not only a ff ects the dispersion and deposition of pollutants, but also a ff ects the expansion of urban areas, which is the source of emissions. For those urban areas with the UPI e ff ect, the UPII value in each urban area remained relatively stable over time The UPIFP value generally showed a downward trend, implying that compared to the surrounding areas, PM 2.5 accumulation in the urban areas became more severe in this period. Compared with existing literature, at the national level, the concentration of monitoring sites around and within the cities were highly correlated [ 37 ]. Remote sensing data have also been used to measure the di ff erence in the PM 2.5 concentration between urban and non-urban areas, but without the concentration trends [ 38 ]. Some studies focused on only a single megacity: there was no significant di ff erence in concentration between the urban and suburban areas of Shanghai [ 39 ]; however, the di ff erence was obvious in Beijing [ 40 ]. In this paper, the study area is an urban agglomeration, which between a nation and a single city. Previous studies have generated equidistant bu ff er zones around cities but did not measure the specific concentration trends of PM 2.5 [ 11 ]. The impact of land use on the spatial distribution of concentration was previously studied, and forests were found to have the greatest impact on PM 2.5 removal [ 11 , 41 , 42 ], which is consistent with the results in this paper. In general, the results of this paper are consistent with existing literature, and the urban-rural concentration trends on the district / county scale in this paper can provide a theoretical basis for pollution control during urban agglomeration 6. Conclusions The UPI e ff ect, which refers to the phenomenon whereby the high particle concentration in the urban areas is gradually attenuated to the surrounding areas, was defined and investigated in this paper. The existing literature has revealed that the rationale for the UPI e ff ect is insu ffi cient, and the mutual influence between cities was often ignored. Therefore, this study tried to fill the gap by dividing the study area into urban, surrounding, and background areas and by dividing the dynamic processes of pollutants into emission, dispersion, and deposition. Moreover, to better study the UPI e ff ect at the local level, the urban area in this paper was changed from the urban land of every prefecture to every county or district. As an area with many cultural heritages, the Hangzhou Bay area was selected as the study area. Based on the observations in the 27 urban areas in the Hangzhou Bay area from 2000 to 2015, we examined the urban expansion and the occurrence of the UPI e ff ect for each urban area. The main findings are as follows: 1 Every urban area in the Hangzhou Bay area experienced rapid expansion from 2000 to 2015. For most of the urban areas, 2000–2005 was the period with the most rapid expansion, and several urban areas even attained growth rates above 300%. Although the form and magnitude of urban expansion varied, single-core expansion occurred in most urban areas 2 By comparison, the PM 2.5 concentrations in the urban areas in 2000 were considerably lower than that in 2005, 2010, and 2015. In addition, from the perspective of the spatial distribution, spatial heterogeneity of the PM 2.5 concentrations were observed in the Hangzhou Bay area. Overall, the PM 2.5 concentration in the urban area had a positive correlation with its area 3 More than half of the urban areas exhibited the UPI e ff ect in the Hangzhou Bay area during this period. Most urban areas with the UPI e ff ect had a relatively high and stable UPII as well. In contrast to the UPII, the UPIFP values showed a downward trend along with urban expansion 4 The urban areas could be divided into three categories according to the terrain, i.e., plain areas, hilly areas, and the junction of plains and hills, and the probability of the UPI e ff ect for the di ff erent categories varied notably Overall, this paper can compensate for the lack of research on the UPI e ff ect at the local level and provide scientific evidence for air pollution control during urban agglomeration planning.

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 16 of 18 Author Contributions: Conceptualization, Y.C.1 and Y.L.; methodology, Y.C.1 and X.F.; investigation and software, X.F., J.W., G.L., and Y.C.2; writing—original draft preparation, X.F., Y.C.1; supervision and writing—review and editing, Y.C.1 and Y.L. All authors have read and agreed to the published version of the manuscript Funding: This research was funded by the Natural Science Foundation of Zhejiang Province, grant number LY 19 D 010012, and supported by the Fundamental Research Funds for the Central Universities of China Conflicts of Interest: The authors declare no conflict of interest References 1 Angel, S.; Parent, J.; Civco, D.L.; Blei, A.; Potere, D. The dimensions of global urban expansion: Estimates and projections for all countries, 2000–2050 Progr. Plan 2011 , 75 , 53–107. [ CrossRef ] 2 Zhou, D.; Zhao, S.; Zhang, L.; Sun, G.; Liu, Y. The footprint of urban heat island e ff ect in China Sci. Rep 2015 , 5 , 1–11. [ CrossRef ] 3 Raj é , F.; Tight, M.; Pope, F.D. Tra ffi c pollution: A search for solutions for a city like Nairobi Cities 2018 , 82 , 100–107. [ CrossRef ] 4 Li, X.; Zhou, Y.; Asrar, G.R.; Imho ff , M.; Li, X. The surface urban heat island response to urban expansion: A panel analysis for the conterminous United States Sci. Total Environ 2017 , 60–606 , 426–435. [ CrossRef ] [ PubMed ] 5 Maheshwari, B.; Pinto, U.; Akbar, S.; Fahey, P. Is urbanisation also the culprit of climate change?–Evidence from Australian cities Urban Clim 2020 , 31 , 100581. [ CrossRef ] 6 Loftus, C.; Yost, M.; Sampson, P.; Arias, G.; Torres, E.; Vasquez, V.B.; Bhatti, P.; Karr, C. Regional PM 2.5 and asthma morbidity in an agricultural community: A panel study Environ. Res 2015 , 136 , 505–512. [ CrossRef ] 7 Gong, T.; Sun, Z.; Zhang, X.; Zhang, Y.; Wang, S.; Han, L.; Zhao, D.; Ding, D.; Zheng, C. Associations of black carbon and PM 2.5 with daily cardiovascular mortality in Beijing, China Atmos. Environ 2019 , 214 , 116876 [ CrossRef ] 8 Tai, A.P.K.; Mickley, L.J.; Jacob, D.J. Correlations between fine particulate matter (PM 2.5) and meteorological variables in the United States: Implications for the sensitivity of PM 2.5 to climate change Atmos. Environ 2010 , 44 , 3976–3984. [ CrossRef ] 9 Liu, W.; Guo, G.; Chen, F.; Chen, Y. Meteorological pattern analysis assisted daily PM 2.5 grades prediction using SVM optimized by PSO algorithm Atmos. Pollut. Res 2019 , 10 , 1482–1491. [ CrossRef ] 10 Wang, S.; Zhou, C.; Wang, Z.; Feng, K.; Hubacek, K. The characteristics and drivers of fine particulate matter (PM 2.5) distribution in China J. Clean Prod 2017 , 142 , 1800–1809. [ CrossRef ] 11 Xu, G.; Ren, X.; Xiong, K.; Li, L.; Bi, X.; Wu, Q. Analysis of the driving factors of PM 2.5 concentration in the air: A case study of the Yangtze River Delta, China Ecol. Indic 2020 , 110 , 105889. [ CrossRef ] 12 Seto, K.C.; Fragkias, M.; Güneralp, B.; Reilly, M.K. A meta-analysis of global urban land expansion PLoS ONE 2011 , 6 , e 23777. [ CrossRef ] [ PubMed ] 13 Sahana, M.; Hong, H.; Sajjad, H. Analyzing urban spatial patterns and trend of urban growth using urban sprawl matrix: A study on Kolkata urban agglomeration, India Sci. Total Environ 2018 , 628–629 , 1557–1566 [ CrossRef ] [ PubMed ] 14 Yue, W.; Liu, Y.; Fan, P. Measuring urban sprawl and its drivers in large Chinese cities: The case of Hangzhou Land Use Pol 2013 , 31 , 358–370. [ CrossRef ] 15 Jiao, L.; Mao, L.; Liu, Y. Multi-order Landscape Expansion Index: Characterizing urban expansion dynamics Landsc. Urban Plan 2015 , 137 , 30–39. [ CrossRef ] 16 Grimm, N.B.; Faeth, S.H.; Golubiewski, N.E.; Redman, C.L.; Wu, J.; Bai, X.; Briggs, J.M. Global Change and the Ecology of Cities Science 2008 , 319 , 756–760. [ CrossRef ] 17 Huang, X.; Cai, Y.; Li, J. Evidence of the mitigated urban particulate matter island (UPI) e ff ect in China during 2000–2015 Sci. Total Environ 2019 , 660 , 1327–1337. [ CrossRef ] 18 Gupta, P.; Christopher, S.A. Seven year particulate matter air quality assessment from surface and satellite measurements Atmos. Chem. Phys 2008 , 8 , 3311–3324. [ CrossRef ] 19 Lee, H.J.; Franklin, S.A.; Turkel, J.E.; Goetz, A.R.; Woods, D.W. Facilitated attentional disengagement from hair-related cues among individuals diagnosed with trichotillomania: An investigation based on the exogenous cueing paradigm J. Obsessive-Compuls. Relat. Disord 2012 , 1 , 8–15. [ CrossRef ]

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 17 of 18 20 Chudnovsky, A.; Lyapustin, A.; Wang, Y.; Tang, C.; Schwartz, J.; Koutrakis, P. High resolution aerosol data from MODIS satellite for urban air quality studies Cent. Eur. J. Geosci 2014 , 6 , 17–26. [ CrossRef ] 21 Chelani, A.B. Estimating PM 2.5 concentration from satellite derived aerosol optical depth and meteorological variables using a combination model Atmos. Pollut. Res 2019 , 10 , 847–857. [ CrossRef ] 22 Yeganeh, B.; Hewson, M.G.; Cli ff ord, S.; Knibbs, L.D.; Morawska, L. A satellite-based model for estimating PM 2.5 concentration in a sparsely populated environment using soft computing techniques Environ. Modell Softw 2017 , 88 , 84–92. [ CrossRef ] 23 He, Q.; Huang, B. Satellite-based high-resolution PM 2.5 estimation over the Beijing-Tianjin-Hebei region of China using an improved geographically and temporally weighted regression model Environ. Pollut 2018 , 236 , 1027–1037. [ CrossRef ] [ PubMed ] 24 Gui, K.; Che, H.; Wang, Y.; Wang, H.; Zhang, L.; Zhao, H.; Zheng, Y.; Sun, T.; Zhang, X. Satellite-derived PM 2.5 concentration trends over Eastern China from 1998 to 2016: Relationships to emissions and meteorological parameters Environ. Pollut 2019 , 247 , 1125–1133. [ CrossRef ] [ PubMed ] 25 Xiao, R.; Lin, M.; Fei, X.; Li, Y.; Zhang, Z.; Meng, Q. Exploring the interactive coercing relationship between urbanization and ecosystem service value in the Shanghai–Hangzhou Bay Metropolitan Region J. Clean Prod 2020 , 253 , 119803. [ CrossRef ] 26 Van Donkelaar, A.; Martin, R.V.; Brauer, M.; Boys, B.L. Use of satellite observations for long-term exposure assessment of global concentrations of fine particulate matter Environ. Health Perspect 2015 , 123 , 135–143 [ CrossRef ] 27 Van Donkelaar, A.; Martin, R.V.; Li, C.; Burnett, R.T. Regional Estimates of Chemical Composition of Fine Particulate Matter Using a Combined Geoscience-Statistical Method with Information from Satellites, Models, and Monitors. Environ Sci. Technol 2019 , 53 , 2595–2611. [ CrossRef ] 28 Van Donkelaar, A.; Martin, R.V.; Brauer, M.; Hsu, N.C.; Kahn, R.A.; Levy, R.C.; Lyapustin, A.; Sayer, A.M.; Winker, D.M. Global Estimates of Fine Particulate Matter using a Combined Geophysical-Statistical Method with Information from Satellites, Models, and Monitors Environ. Sci. Technol 2016 , 50 , 3762–3772. [ CrossRef ] 29 Liu, Y.; Wu, J.; Yu, D.; Ma, Q. The relationship between urban form and air pollution depends on seasonality and city size. Environ Sci. Pollut. Res 2018 , 25 , 15554–15567. [ CrossRef ] 30 Zhou, J.; Xing, Z.; Deng, J.; Du, K. Characterizing and sourcing ambient PM 2.5 over key emission regions in China I: Water-soluble ions and carbonaceous fractions Atmos. Environ 2016 , 135 , 20–30. [ CrossRef ] 31 Li, G.; Fang, C.; Wang, S.; Sun, S. The E ff ect of Economic Growth, Urbanization, and Industrialization on Fine Particulate Matter (PM 2.5) Concentrations in China Environ. Sci. Technol 2016 , 50 , 11452–11459. [ CrossRef ] [ PubMed ] 32 Łowicki, D. Landscape pattern as an indicator of urban air pollution of particulate matter in Poland Ecol Indic 2019 , 97 , 17–24. [ CrossRef ] 33 Zhang, R.; Wang, G.; Guo, S.; Zamora, M.L.; Ying, Q.; Lin, Y.; Wang, W.; Hu, M.; Wang, Y. Formation of Urban Fine Particulate Matter. Chem. Rev 2015 , 115 , 3803–3855. [ CrossRef ] [ PubMed ] 34 Grossman, G.M.; Krueger, A.B. Economic Growth and the Environment Quat. J. Econ 1995 , 110 , 353–377 [ CrossRef ] 35 Shen, Y.; Lisa Ahlers, A. Local environmental governance innovation in China: Staging ‘triangular dialogues’ for industrial air pollution control J. Chin. Gov 2018 , 3 , 351–369. [ CrossRef ] 36 Li, F.; Zhou, T. E ff ects of urban form on air quality in China: An analysis based on the spatial autoregressive model Cities 2019 , 89 , 130–140. [ CrossRef ] 37 Xu, L.; Batterman, S.; Chen, F.; Li, J.; Zhong, X.; Feng, Y.; Rao, Q.; Chen, F. Spatiotemporal characteristics of PM 2.5 and PM 10 at urban and corresponding background sites in 23 cities in China Sci. Total Environ 2017 , 599–600 , 2074–2084. [ CrossRef ] 38 Han, L.; Zhou, W.; Li, W.; Li, L. Impact of urbanization level on urban air quality: A case of fine particles (PM 2.5 ) in Chinese cities Environ. Pollut 2014 , 194 , 163–170. [ CrossRef ] 39 Liu, C.; Henderson, B.H.; Wang, D.; Yang, X.; Peng, Z. A land use regression application into assessing spatial variation of intra-urban fine particulate matter (PM 2.5 ) and nitrogen dioxide (NO 2 ) concentrations in City of Shanghai, China Sci. Total Environ 2016 , 565 , 607–615. [ CrossRef ] 40 Zhao, X.; Zhang, X.; Xu, X.; Xu, J.; Meng, W.; Pu, W. Seasonal and diurnal variations of ambient PM 2.5 concentration in urban and rural environments in Beijing Atmos. Environ 2009 , 43 , 2893–2900. [ CrossRef ]

Int. J. Environ. Res. Public Health 2020 , 17 , 5535 18 of 18 41 Nowak, D.J.; Hirabayashi, S.; Bodine, A.; Greenfield, E. Tree and forest e ff ects on air quality and human health in the United States Environ. Pollut 2014 , 193 , 119–129. [ CrossRef ] [ PubMed ] 42 Lin, Y.; Yuan, X.; Zhai, T.; Wang, J. E ff ects of land-use patterns on PM 2.5 in China’s developed coastal region: Exploration and solutions Sci. Total Environ 2020 , 703 , 135602. [ CrossRef ] [ PubMed ] © 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 / ).

Other Environmental Sciences Concepts:

[back to top]

Discover the significance of concepts within the article: ‘Measuring the Urban Particulate Matter Island Effect with Rapid Urban Expansion’. Further sources in the context of Environmental Sciences might help you critically compare this page with similair documents:

Emission, Geographical location, Deposition, Cultural heritage, Surrounding area, Urban area, Public health, Physical growth, Land management, Heat energy, Air pollution, Urbanization, Economic growth, Urban expansion, Spatial distribution, Land use, Coastal area, Urban growth, Negative impact, Economic and social development, World Health Organization, Scientific evidence, Principal component analysis, Air pollution control, Particulate matter, Urban Air Pollution, Air quality, Geographical Condition, Nitrogen dioxide, Rapid urbanization, Study area, Anthropogenic activities, Harmful Effect, Global population, Buffer zone, Global climate change, Potential source, Meteorological conditions, Rapid expansion, Water-soluble ions, Urban pollution, Administrative boundaries, Quantitative investigation, Dispersion, Natural background, Rapid economic growth, Local level, Satellite imagery, Spatial heterogeneity, Spatial pattern, Coastal Plain, Hilly area, Environmental governance, Specific area, Air quality assessment, Environmental Kuznets curve, Geographical Information System, Fine particulate matter, Spatiotemporal analysis, Remote sensing data, Zhejiang province, Natural Science Foundation, Particle concentration, Urban form, PM 2.5, PM 2.5 concentration, Urban heat island effect, Urban heat island, Spatial autoregressive model, Land use regression, Soft Computing Techniques, Land use data, Kernel density estimation, AOD, Particle properties, Dynamic processes, Concentration difference, Eastern China, Urban agglomeration, Transboundary air pollution, Global change, Remote Sensing Technology, Construction Land, Prefecture-Level Cities, Urban air quality, Particulate matter exposure, Urban Spatial Pattern, Terrain, Landscape Pattern, Land-Use Pattern, Urban sprawl, Ecosystem service value, Beijing-Tianjin-Hebei region, PM 2.5 and PM 10, Yangtze River Delta Region, Demonstration area, Rapid Urban Expansion, Urban expansion dynamics, Local Environmental Governance, Urban agglomeration planning, Urban-rural gradient, Moderate-Resolution Imaging Spectroradiometer, PM 2.5 pollution, Google Earth image, Urban land area, High urbanization rate, Aerosol optical depth, Anthropogenic modification, Satellite measurements, Ecology of cities, Meteorological variable, Large Chinese cities, Landscape Expansion Index, Underlying surface characteristics, Plain area, Local emission, Zhejiang University, Hangzhou Bay, Spatiotemporal Dynamic, Digital elevation model data, Economic and social unit, Evaluation criterion, Urban pollution island, Satellite-based Model.