JES-00782; No of Pages 8 J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX X–XXX
Available online at www.sciencedirect.com
ScienceDirect www.elsevier.com/locate/jes
Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing Qian Yin1 , Jinfeng Wang1,2,⁎, Maogui Hu1 , Hoting Wong1 1. LREIS, Institute of Geographic Sciences and Nature Resources Research, Chinese Academy of Sciences, Beijing 100101, China. E-mail:
[email protected] 2. Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, China
AR TIC LE I N FO
ABS TR ACT
Article history:
When investigating the impact of air pollution on health, particulate matter less than 2.5 μm
Received 9 October 2015
in aerodynamic diameter (PM2.5) is considered more harmful than particulates of other
Revised 5 March 2016
sizes. Therefore, studies of PM2.5 have attracted more attention. Beijing, the capital of China,
Accepted 7 March 2016
is notorious for its serious air pollution problem, an issue which has been of great concern to
Available online xxxx
the residents, government, and related institutes for decades. However, in China, significantly less time has been devoted to observing PM2.5 than for PM10. Especially
Keywords:
before 2013, the density of the PM2.5 ground observation network was relatively low, and
PM10 concentration
the distribution of observation stations was uneven. One solution is to estimate PM2.5
PM2.5 concentration estimation
concentrations from the existing data on PM10. In the present study, by analyzing the
Wind speed
relationship between the concentrations of PM2.5 and PM10, and the meteorological
Wind direction
conditions for each season in Beijing from 2008 to 2014, a U-shaped relationship was found between the daily maximum wind speed and the daily PM concentration, including both PM2.5 and PM10. That is, the relationship between wind speed and PM concentration is not a simple positive or negative correlation in these wind directions; their relationship has a complex effect, with higher PM at low and high wind than for moderate winds. Additionally, in contrast to previous studies, we found that the PM2.5/PM10 ratio is proportional to the mean relative humidity (MRH). According to this relationship, for each season we established a multiple nonlinear regression (MNLR) model to estimate the PM2.5 concentrations of the missing periods. © 2016 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.
Introduction With people becoming increasingly keen on healthy living, the levels of atmospheric particulate matter (PM) have received considerable attention. Although the detailed effects of each type of PM on health have not yet been established, epidemiological studies (Ostro et al., 2010; Zhang et al., 2002; Cairncross et al., 2007) have identified a strong exposure–response relationship between PM and both short-term (coughing, sneezing,
runny nose, shortness of breath and lung irritation) and longterm health effects (lung cancer, cardiovascular and cardiopulmonary diseases). PM10 is particulate matter of less than 10 μm in aerodynamic diameter. It can penetrate into the air passages of the lungs. Different from PM10, PM2.5 is less than 2.5 μm in-aerodynamic diameter. It can reach the deeper alveolar spaces and is more harmful than particulate matter of other sizes (Hossain and Easa, 2012; Fan et al., 2013; Chun et al., 2014; Park and Kim, 2014).
⁎ Corresponding author. E-mail:
[email protected] (Jinfeng Wang).
http://dx.doi.org/10.1016/j.jes.2016.03.024 1001-0742/© 2016 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.
Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024
2
J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX –XXX
Moreover, it has been reported that PM2.5 can easily penetrate into other organs, and its removal takes longer. Thus, within the research on the health-related effects of air pollution, PM2.5 has attracted widespread attention. Air pollution has long been a serious problem in China, especially in eastern areas such as the Yangtze River Delta, southern areas such as the Pearl River Delta (Wang and Christopher, 2003) as well as in the Jing-Jin-Ji (Beijing, Tianjin and Hebei provinces) region (Xu et al., 2005; Ding et al., 2008; Zhang et al., 2013). Beijing, the capital of China, has a large population and high energy consumption, both of which have led to serious air pollution. The weather in Beijing is typical of a temperate monsoon climate: summers are hot and rainy, while winters are cold and dry. Particulates in Beijing are very complicated in terms of sources, pollution characteristics, and their effects on the environment and health. Among these particulates, PM2.5 is one of the main air pollutants, and its concentration is extremely high. For example, January 2013 witnessed the most frequent and serious episodes of heavy haze to date, which caused consternation among the public (Che et al., 2014; Wang et al., 2014; Zhang et al., 2014). In 2014, Beijing residents suffered seven periods of heavy pollution that lasted more than 3 days, including a week-long period in February. However, in China the time spent on PM2.5 ground observation has been quite short, only about a few years, and the density of the PM2.5 ground observation network is relatively low. Moreover, the distribution of observation stations is uneven. Particularly before 2013, information about the concentration of PM2.5 in Beijing was not released. The limited number of observation stations meant that the PM2.5 characteristics throughout Beijing could not be well established. One solution to this problem is to estimate the PM2.5 concentrations using the existing PM10 concentration data. In this respect, the existing research has focused on estimating the concentration of PM2.5 based on the aerosol optical depth (AOD), NO2, SO2, CO or O3 concentration (Che et al., 2015; Huo et al., 2011; Zhang et al., 2015; Wang et al., 2013). Furthermore, since PM2.5 is a part of PM10 (Wang et al., 2006), many researchers of previous studies believed that the PM2.5 concentration has a linear relationship with PM10 concentration, therefore they conducted a linear regression analysis to estimate PM2.5 concentration according to the PM10 concentration. In the present study, we first analyzed the relationships between PM2.5 and PM10, as well as the meteorological conditions, including mean relative humidity (MRH), sunshine duration, wind speed, and wind direction. Accordingly, we found that the concentrations of PM2.5 and PM10 are nonlinear with respect to MRH. Secondly, since different seasons have different weather conditions, their influences on PM concentrations are also different. Therefore, according to the above nonlinear relationship, for each season we conducted a multiple nonlinear regression (MNLR) model. Since our PM10 concentration data covered an entire sevenyear period from 2008/01/01 to 2014/12/31, but the PM2.5 data only covered about 4 years and 5 months (from 2010/01/01 to 2014/12/31, with the exception of 2012/05/16 to 2012/12/31), we used the model to estimate the PM2.5 concentration data of the missing period. In addition, we analyzed the relationship between the wind speed from different directions and PM concentrations in detail.
1. Materials and methods 1.1. Study area The study was conducted in Beijing. As shown in Fig. 1, the red area represents the urban region of Beijing and the black dot indicates the location of the observation point.
1.2. Materials Two datasets were used in the research. The first dataset consisted of hourly observations of PM2.5 concentrations at the United States (US) Embassy station from 2010/01/01 to 2014/12/ 31 (with the exception of 2012/05/16 to 2012/12/31), which were obtained from the website of the Embassy of the United States in China (http://chinese.usembassy-china.org.cn/070109air.html), and the daily mean concentrations of PM10 at the same station from 2008/01/01 to 2014/12/31, which were obtained from the website of the Beijing Municipal Environmental Monitoring Center (http://www.bjmemc.com.cn/). The second dataset consisted of daily meteorological data from 2008/01/01 to 2014/12/31 in Beijing, which was obtained from the website of the China Meteorological Data Sharing Service System (http://cdc.nmic.cn/home.do). The meteorological parameters are listed in Table 1.
1.3. Methods In this research, some common statistical methods were employed, for example Spearman correlation analyses and an MNLR model which was utilized to fit the PM2.5 concentration
Fig. 1 – The study region and the observation point.
Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024
3
J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX X–XXX
Table 1 – Meteorological parameters considered in the statistical analysis. Independent variables
Meteorological parameters
x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x11 x12 x13 x14 x15
Daily mean air temperature (°C) Daily maximum air temperature (°C) Daily minimum air temperature (°C) Mean relative humidity (%) Minimum relative humidity (%) Extreme max. wind speed (m/sec) Direction of extreme max. wind speed (16 directions) Maximum wind speed (m/sec) Direction of maximum wind speed (16 directions) Mean wind speed (m/sec) Mean air pressure (hPa) Maximum air pressure (hPa) Minimum air pressure (hPa) Daily total precipitation (from 20:00 to 19:59) (mm) Sunshine duration (hr)
according to the PM10 concentration and meteorological variables shown in Table 1. The assumption of MNLR models is that the relationship between the dependent variable y and the regressor xi is nonlinear. It was defined by the following (Eq. (1)): 15 2 y ¼ α þ ∑15 j¼0 β j x j þ ∑ j>i β ji x j xi þ ∑ j¼0 β jj x j
ð1Þ
where y is the daily mean concentration of PM2.5 and α is the intercept. x0 is the daily mean concentration of PM10. x1,… xj (j = 1,…,15) are the meteorological variables in Table 1, while β0,…βj (j = 0,1,…,15) are the regression coefficients. For each meteorological variable, the regression was done
stepwise to add and delete terms based on Akaike Information Criterion (AIC) statistics to obtain the best model fit. In addition, the mean absolute percentage error (MAPE) was used to measure the accuracy of the model. This is a statistical forecasting method to measure prediction accuracy. It is defined by the formula (Eq. (2)): MAPE ¼ 100% ∑ni¼1 jðAt −Ft Þ=At j =n
ð2Þ
where At is the actual value and Ft is the estimated value. The absolute value in this calculation is summed for every forecasted point in time and divided by the number of fitted points n. Multiplying by 100 gives the percentage error. R provides comprehensive support for MNLR. All the statistical analyses were conducted using R statistical software (version 3.1.2) and the Openair package (version 1.4; Carslaw, 2015).
2. Results and discussion 2.1. Descriptive statistics We collected and analyzed the concentration data of PM2.5 and PM10 at the observation point from January 1, 2008 to December 31, 2014 (with the exception of May 16, 2012 to December 31, 2012). The results showed that the daily mean concentration of PM10 ranged from 8 to 600 μg/m3, while the concentration of PM2.5 ranged from 3 to 480 μg/m3. Fig. 2 shows the plots of the daily PM2.5 and PM10 concentrations at the US Embassy station from January 2010 to December 2010. Fig. 3 shows the daily MRH of Beijing from
Fig. 2 – Plots of daily mean PM2.5 and PM10 concentrations at the US Embassy station in 2010. PM2.5: particulate matter less than 2.5 μm; PM10: particulate matter less than 10 μm.
Fig. 3 – Plot of daily mean relative humidity (MRH) of Beijing from 2010 to 2012. Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024
4
J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX –XXX
Fig. 4 – Annual mean of PM2.5, PM10, PM10–PM2.5 concentrations, PM2.5/PM10 ratio, and the MRH from 2010 to 2014 in Beijing. MRH: mean relative humidity.
2010 to 2012. Fig. 4 shows the annual mean of the PM2.5 concentration, PM10 concentration, PM2.5/PM10 ratio, PM10– PM2.5 concentration and the MRH from 2010 to 2014 (with the exception the period of 2012/05/16 to 2012/12/31).
2.2. Relationship between PM2.5 and PM10 and meteorological factors To better understand the factors influencing PM2.5, we investigated the relationships between the daily mean concentration of PM2.5 and the meteorological factors for each season. Spearman correlations between them were calculated (Table 2). The seasons were defined as: spring (March–May), summer (June–August.), autumn (September–November) and winter (December–February). From Table 2, it can be seen that PM2.5, as a part of PM10, has a strong correlation with PM10 in respect to their concentrations. In addition to PM10, most of the meteorological factors also had correlations with PM2.5 concentrations. Among these meteorological factors, relative humidity (RH, including MRH and minimum relative humidity) and sunshine duration had stronger
correlations with PM2.5 concentrations than the other meteorological factors. The correlation coefficients between MRH and PM2.5 concentrations were 0.59, 0.32, 0.55, and 0.69 for spring, summer, autumn and winter respectively. Similarly, the correlation coefficients between sunshine duration and PM2.5 concentration were −0.51, −0.39, −0.51, and −0.52 for the four seasons respectively. The correlations all had obvious statistical significance. Since MRH and sunshine duration were highly correlated in this data set (their correlation coefficients were 0.64, 0.63, 0.58, and 0.63 for the four seasons respectively), and their p-values were all less than 0.001, we only needed to consider one of them. Fig. 5 shows the scatter plots of the daily mean concentration of PM2.5, the daily mean concentration of PM10 and the daily MRH in four seasons. The above significant positive and negative correlations both had certain practical implications. For MRH, the significant positive correlation showed that higher MRH was often associated with windless, cloudy, and insufficiently sunny days, which aggravated the accumulation and chemical reaction of pollutants (Kang et al., 2013; Csavina et al., 2014). In contrast, the significant negative correlation with sunshine duration showed that with the increase of sunshine hours, the concentration of PM2.5 decreased. A day with more sunshine hours often indicated better weather conditions with little cloud cover or strong winds, which are favorable for the diffusion and elimination of air pollutants (Yang et al., 2009; Sanchez-Romero et al., 2014). Beyond that, a poor correlation was found between PM2.5 concentration and wind speed. This means that the relationship between wind speed and PM concentration may not be a simple positive or negative correlation in different seasons. The impacts of wind speed and wind direction on PM concentration will be discussed in more detail in the next section. Similarly, a poor correlation was found between PM2.5 concentration and air pressure in spring, autumn, and winter
Table 2 – Spearman correlations between PM2.5 and meteorological factors for the four seasons.
x0 (PM10) x1 (Mean temperature) x2 (Maximum temperature) x3 (Minimum temperature) x4 (Mean relative humidity) x5 (Minimum relative humidity) x6 (Extreme maximum wind speed) x7(Direction of extreme maximum wind speed) x8 (Maximum wind speed) x9 (Direction of maximum wind speed) x10 (Mean wind speed) x11 (Mean air pressure) x12 (Maximum air pressure) x13 (Minimum air pressure) x14 (Daily total precipitation) x15 (Sunshine duration)
PM2.5 (spring)
PM2.5 (summer)
PM2.5 (autumn)
PM2.5 (winter)
0.83 ⁎⁎⁎ 0.18 ⁎⁎ 0.16 ⁎ 0.25 ⁎⁎ 0.59 ⁎⁎⁎ 0.43 ⁎⁎⁎ −0.24 ⁎⁎
0.87 ⁎⁎⁎ 0.14 0.14 0.33 ⁎⁎⁎ 0.32 ⁎⁎⁎ 0.43 ⁎⁎⁎
0.96 ⁎⁎⁎ −0.09 −0.09 −0.04 0.55 ⁎⁎⁎ 0.51 ⁎⁎⁎ −0.40 ⁎⁎⁎
0.94 ⁎⁎⁎ 0.40 ⁎⁎⁎ 0.35 ⁎⁎⁎ 0.37 ⁎⁎⁎ 0.69 ⁎⁎⁎ 0.48 ⁎⁎⁎ −0.39 ⁎⁎⁎
N/A −0.38 ⁎⁎⁎ N/A −0.34 ⁎⁎⁎ −0.25 ⁎⁎⁎ −0.23 ⁎⁎ −0.21 ⁎
N/A −0.40 ⁎⁎⁎ N/A −0.43 ⁎⁎⁎ −0.47 ⁎⁎⁎ −0.46 ⁎⁎⁎ −0.43 ⁎⁎⁎
−0.15 −0.51 ⁎⁎⁎
0.11 −0.52 ⁎⁎⁎
N/A −0.23 ⁎⁎ N/A −0.26 ⁎⁎ −0.35 ⁎⁎⁎ −0.33 ⁎⁎⁎ −0.35 ⁎⁎⁎ 0.22 ⁎⁎ −0.51 ⁎⁎⁎
−0.09 N/A −0.07 N/A 0.12 −0.10 −0.12 −0.06 0.01 −0.39 ⁎⁎⁎
PM2.5: particulate matter less than 2.5 μm; PM10: particulate matter less than 10 μm. N/A: Not available. ⁎⁎⁎ p-Value < 0.001. ⁎⁎ p-Value < 0.01. ⁎ p-Value < 0.05.
Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024
J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX X–XXX
5
Fig. 5 – Scatter plots of PM2.5 and PM10 concentration and mean relative humidity (MRH).
(− 0.47 < r < − 0.25, p-value < 0.05); meanwhile, no obvious correlation was found between PM2.5 concentration and air pressure in summer (p-value > 0.1). In contrast, for PM2.5, the correlations with temperature and precipitation were not obvious, which indicated that the links between PM2.5 pollutants and temperature or precipitation were weak in each season.
negative correlation in these wind directions, so that their relationship is similar to a “bathtub curve”. From these U-shaped figures, we found that when the wind speed is low (< 4 m/sec), the PM concentration is obviously high. In
2.3. Relationship between PM2.5 and wind direction To better understand the influence of wind speed and wind direction on PM concentration, the prevailing wind directions were divided into 16 groups: N/NNE/NE/ENE/E/ESE/SE/SSE/S/ SSW/SW/WSW/W/WNW/NW/NNW. Fig. 6 shows the polar plot (Carslaw, 2015) of mean PM2.5 concentrations for the 16 directions. In the plot, (0, 0) is the middle of the plot and the radial distance represents the mean wind speed of each direction, the unit is m/sec. From Fig. 6, we generally found that the relationship between wind speed and PM2.5 concentration is not a simple negative correlation in each direction. That is, at greater wind speed, the PM2.5 concentration is not necessarily smaller. In order to better understand the relationships between PM concentration and maximum wind speed, in Fig. 7, the horizontal axes represent the daily maximum wind speeds (unit: m/sec) and the vertical axes represent the mean concentrations of PM2.5 (red lines) and PM10 (blue lines). From Fig. 7, for most of the wind directions, a U-shaped relationship was found between maximum wind speed and PM concentration. This means that the relationship between wind speed and PM concentration is not a simple positive or
Fig. 6 – Polar plot of mean PM2.5 concentrations at 16 directions.
Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024
6
J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX –XXX
Fig. 7 – In different wind directions, the mean PM2.5 and PM10 concentrations under different maximum wind speed levels.
addition, when wind speed increases (4–8 m/sec), the PM concentration decreases. This indicates that a low wind speed could limit the spread of particulates, but increases in wind speed could restrain the accumulation of particulates and, thus, reduce the PM2.5 and PM10 pollutant concentrations. However, as the wind speed continues to increase, the PM concentration unexpectedly becomes higher again. The reason for this phenomenon may be that a strong wind can bring in pollution from surrounding areas.
2.4. Multiple nonlinear regression Since PM2.5 is a part of PM10, many researchers in previous studies believed that the PM2.5 concentration has a linear relationship with the PM10 concentration (Wang et al., 2006), Therefore, they conducted a linear regression analysis (Model I) to estimate PM2.5 concentrations according to PM10 concentrations. However, from Fig. 4, we found that the PM2.5/PM10 ratio and the MRH have an obvious relationship; that is, the
PM2.5/PM10 ratio is proportional to the MRH. Since different seasons have different MRH conditions, their influences on PM concentration also differ. According to the above relationship between them, for each season, we conducted an MNLR (Model II). Table 3 shows a comparison of the results of Models I and II: In Table 3, y represents the PM2.5 concentration, x1 represents the concentration of PM10, and x4 represents the MRH. By comparing the results of the two models, we can clearly see that in spring, the adjusted R2 (0.80) of Model II was clearly higher than the adjusted R2 (0.73) of Model I, and the error of Model II (13%) was remarkably smaller than the error of Model I (21.5%). In the other seasons, the results of Model II were also better than Model I. In addition, we found that the errors were obviously larger in spring than in the other seasons. One reason for this was that in spring the impacts of climate conditions, such as dust storms, on PM2.5 concentrations were greater than during the other seasons. If there was no dust storm, the PM2.5/PM10 ratio was larger and closer to the ratio of the other seasons;
Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024
7
J O U RN A L OF E N V I RO N ME N TA L S CI EN CE S X X (2 0 1 6 ) XX X–XXX
Table 3 – A comparison of the results of multiple linear regression (Model I) and multiple nonlinear regression (Model II). Adjusted R2
The regression results Model I
Season Spring Summer Autumn Winter
Y Y Y Y
= = = =
−32.41 −34.82 −24.13 −25.40
+ + + +
Model II
0.50x1 + 1.17x4 0.82 x1 + 0.42 * x4 0.82 * x1 + 0.28 * x4 0.85 * x1 + 0.42 * x4
Y Y Y Y
= = = =
Model I Model II
8.76 + 0.12 * x1 + 0.01 * x1 * x4 −8.67 + 0.58 * x1 + 0.0037 * x1 * x4 7.45 + 0.62 * x1 + 0.003 * x1 * x4 −5.75 + 0.65 * x1 + 0.0036 * x1 * x4
however, if a dust storm occurred, then the PM2.5/PM10 ratio became significantly smaller. Therefore, if we only considered PM10 concentrations and MRH as the explanatory variables when estimating the concentrations of PM2.5, the errors would be larger.
2.5. Estimation of PM2.5 concentrations The results of the MNLR can be utilized to estimate the concentrations of PM2.5. Fig. 8 shows the estimated PM2.5 values for the missing period (May 17 to December 31) in 2012. In the figure, the dotted line represents the ground monitoring concentration of PM10. PM2.5_R represents the real ground monitoring concentration of PM2.5, while PM2.5_E represents the estimated PM2.5 concentration value.
0.73 0.87 0.92 0.95
0.80 0.88 0.94 0.97
MAPE (error) Model I 21.5% 11.1% 15.3% 17.1%
Model II
(1.2%, 63.9%) 13.0% (1.2%, 28.8%) (0.5%, 31.5%) 10.0% (0.4%, 25.6%) (0.7%, 41.5%) 9.8% (0.6%, 25.3%) (0.8%, 50.6%) 9.9% (0.7%, 25.7%)
relationship was found between maximum wind speed and PM concentration in most wind directions. That is, the relationship between wind speed and PM concentration is not a simple positive or negative correlation in these wind directions, with higher PM observed at low or high wind and lower PM for moderate wind speeds. In addition, in contrast to previous studies, we found that the PM2.5/PM10 ratio is proportionate to the MRH. According to this relationship between them, for each season, we established an MNLR model to estimate the PM2.5 concentration of the missing period. The cross-seasonal validation suggests that our strategy for reconstructing the PM2.5 time series using PM10 data and meteorological factors in Beijing is satisfactory.
Acknowledgments 3. Conclusions By analyzing the relationships between PM2.5 concentration, PM10 concentration, and the meteorological conditions for each season in Beijing from 2010 to 2014, a U-shaped
This work was supported by the National Science & Technology Major Program of China (No. 2012CB955503) and the National Natural Science Foundation of China (Nos. 41421001; 41301425; 41271404).
Fig. 8 – The estimated PM2.5 values for the missing period in 2012. Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024
8
J O U RN A L OF E N V I RO N ME N TA L S CIE N CE S X X (2 0 1 6 ) XXX –XXX
REFERENCES Cairncross, E.K., John, J., Zunckel, M., 2007. A novel air pollution index based on the relative risk of daily mortality associated with short-term exposure to common air pollutants. Atmos. Environ. 41 (38), 8442–8454. Carslaw, D.C., 2015. The openair manual — open-source tools for analyzing air pollution data. Available at: http://www. openair-project.org/Default.aspx (Date accessed: 18 January 2016). Che, H., Xia, X., Zhu, J., Li, Z., Dubovic, O., Holben, B., 2014. Column aerosol optical properties and aerosol radiative forcing during a serious haze-fog month over North China Plain in 2013 based on ground-based sun photometer measurements. Atmos. Chem. Phys. 14 (4), 2125–2138. Che, H.Z., Xia, X.G., Zhu, J., Hong, W., Wang, Y.Q., Sun, J.Y., et al., 2015. Aerosol optical properties under the condition of heavy haze over an urban site of Beijing, China. Environ. Sci. Pollut. Res. Int. 22 (2), 1043–1053. Chun, Y.S., Kim, H.S., Chun, Y.S., 2014. On large-scale transport of dust storms and anthropogenic dust-falls over East Asia observed in Central Korea in 2009. Asia-Pac. J. Atmos. Sci. 50 (3), 345–354. Csavina, J., Field, J., Felix, O., Alba, Y., Avitia, C., Saez, A.E., et al., 2014. Effect of wind speed and relative humidity on atmospheric dust concentrations in semi-arid climates. Sci. Total Environ. 487, 82–90. Ding, A.J., Wang, T., Thouret, V., Cammas, J.P., Nedelec, P., 2008. Tropospheric ozone climatology over Beijing: analysis of aircraft data from the MOZAIC program. Atmos. Chem. Phys. 8 (1), 1–13. Fan, Q., Shen, C., Wang, X.M., Yuan, L., Huang, W., Liang, G.X., et al., 2013. Impact of a dust storm on characteristics of particle matter (PM) in Guangzhou, China. Asia-Pac. J. Atmos. Sci. 49 (1), 121–131. Hossain, K.M.A., Easa, S.M., 2012. Pollutant dispersion characteristics in Dhaka city, Bangladesh. Asia-Pac. J. Atmos. Sci. 48 (1), 35–41. Huo, M.Q., Sun, Q., Bai, Y.H., Xie, P., Lin, Z.R., Wang, X.S., et al., 2011. Acidic and basic properties and buffer capacity of airborne particulate matter in an urban area of Beijing. Environ. Monit. Assess. 176 (1–4), 355–364. Kang, H.Q., Zhu, B., Su, J.F., Wang, H.L., Zhang, Q.C., Wang, F., 2013. Analysis of a long-lasting haze episode in Nanjing, China. Atmos. Res. 120, 78–87. Ostro, B., Lipsett, M., Reynolds, P., Goldberg, D., Hertz, A., Garcia, C., et al., 2010. Long-term exposure to constituents of fine
particulate air pollution and mortality: results from the California teachers study. Environ. Health Perspect. 118 (3), 363–369. Park, R.J., Kim, S.W., 2014. Air quality modeling in East Asia: present issues and future directions. Asia-Pac. J. Atmos. Sci. 50 (1), 105–120. Sanchez-Romero, A., Sanchez-Lorenzo, A., Calbo, J., Gonzalez, J.A., Azorin-Molina, C., 2014. The signal of aerosol induced changes in sunshine duration records: a review of the evidence. J. Geophys. Res. 119 (8), 4657–4673. Wang, J., Christopher, S., 2003. Intercomparison between satellite-derived aerosol optical thickness and PM2.5 mass: implications for air quality studies. Geophys. Res. Lett. 30 (21), 2095. Wang, X.H., Bi, X.H., Sheng, G.Y., Fu, J.M., 2006. Chemical composition and sources of PM10 and PM2.5 aerosols in Guangzhou, China. Environ. Monit. Assess. 119 (1–3), 425–439. Wang, J.F., Hu, M.G., Xu, C.D., Christakos, G., Zhao, Y., 2013. Estimation of citywide air pollution in Beijing. PLoS ONE 8 (1), e53400. Wang, Y., Yao, L., Wang, L., Liu, Z., Ji, D., Tang, G., et al., 2014. Mechanism for the formation of the January 2013 heavy haze pollution episode over central and eastern China. Sci. China Earth Sci. 57 (1), 14–25. Xu, X., Xie, L., Ding, G., Bian, L., 2005. Beijing air pollution project to benefit 2008 Olympic games. Bull. Am. Meteorol. Soc. 86 (11), 1543–1544. Yang, Y.H., Zhao, N., Hao, X.H., Li, C.Q., 2009. Decreasing trend of sunshine hours and related driving forces in North China. Theor. Appl. Climatol. 97 (1), 91–98. Zhang, J.J., Hu, W., Wei, F., Wu, G., Korn, L.R., Chapman, R.S., 2002. Children's respiratory morbidity prevalence in relation to air pollution in four Chinese cities. Environ. Health Perspect. 110 (9), 961–967. Zhang, R., Jing, J., Tao, J., Hsu, S.C., Wang, G., Cao, J., et al., 2013. Chemical characterization and source apportionment of PM2.5 in Beijing: seasonal perspective. Atmos. Chem. Phys. 13 (14), 7053–7074. Zhang, J., Sun, Y., Liu, Z., Ji, D., Hu, B., Liu, Q., et al., 2014. Characterization of submicron aerosols during a serious pollution month in Beijing. Atmos. Chem. Phys. 14 (6), 2887–2903. Zhang, Z.Y., Zhang, X.L., Gong, D.Y., Quan, W.J., Zhao, X.J., Ma, Z.Q., et al., 2015. Evolution of surface O3 and PM2.5 concentrations and their relationships with meteorological conditions over the last decade in Beijing. Atmos. Environ. 108, 67–75.
Please cite this article as: Yin, Q., et al., Estimation of daily PM2.5 concentration and its relationship with meteorological conditions in Beijing, J. Environ. Sci. (2016), http://dx.doi.org/10.1016/j.jes.2016.03.024