Introduction
Fine particulate matter with aerodynamic diameter less than 2.5 µm
(PM2.5) is a robust indicator of mortality and other negative health
effects associated with ambient air pollution (Goldberg, 2008; Laden
et al., 2006). It is estimated that more than 3 million people lost
their lives prematurely due to PM2.5 in 2010 (Lim et al.,
2012), of which 1 million occurred in East Asia (Silva et al., 2013). In
China, there have already been several episodes with PM2.5 described as
“beyond index” levels. Thus, it is of paramount importance to monitor
PM2.5 concentration across China. Satellite remote sensing has a high
potential to monitor PM2.5.
Satellite retrievals of aerosol optical depth (AOD), which provide a measure
of the amount of light extinction through the atmospheric column due to the
presence of aerosols, have long been recognized to relate to ground level
PM2.5 (Wang and Christopher, 2003). Many studies have developed
advanced statistical relationships to estimate with high accuracy surface
PM2.5 from satellite AOD (Liu et al., 2009; Kloog et al., 2012; Hu et
al., 2013). For example, Ma et al. (2014) estimated PM2.5 concentrations in China from satellite AOD by developing a national-scale
geographically weighted regression model, and found strong agreement
(r2=0.64) with ground measurements.
In addition to empirical statistical methods, satellite AOD can also be
geophysically related to surface PM2.5 by the use of a chemical
transport model to simulate the PM2.5 to AOD relationship (Liu et al.,
2004; van Donkelaar et al., 2010). This approach was first demonstrated
using data from the Multiangle Imaging Spectroradiometer (MISR) aboard
NASA's Terra satellite over the United States for 2001 (Liu et al., 2004).
Van Donkelaar et al. (2006, 2010) extended this approach to estimate
PM2.5 from AOD retrieved from both the MODIS (Moderate Resolution
Imaging Spectroradiometer) and the MISR satellite instruments, and developed
a long-term global estimate of PM2.5 at a spatial resolution of
approximately 10 km × 10 km. Boys et al. (2014) used AOD retrieved from MISR
and the SeaWiFS (Sea-Viewing Wide Field-of-View Sensor) to produce a 15-year
(1998–2012) global trend of ground-level PM2.5. These previous
studies have proven to be globally effective, but more detailed regional
investigation is needed in densely polluted and populated regions like
China.
The Geostationary Ocean Color Imager (GOCI) is the first geostationary
satellite instrument that offers multi-spectral aerosol optical properties
in Northeast Asia (Park et al., 2014). GOCI has a high observation density
of eight retrievals per day (hourly retrievals from 09:00 to 16:00 Korean Standard
Time) over a location, which exceeds the retrieval density of traditional
low-Earth polar-orbiting satellite instruments. Thus, GOCI is promising for
more detailed investigations on aerosol properties in highly polluted and
populated regions including eastern China.
In this study, we estimate ground-level PM2.5 in eastern China for 2013
at a horizontal resolution of 6 km by 6 km, by using AOD retrieved from
GOCI, coupled with the relationship of PM2.5 to AOD simulated by a
chemical transport model (GEOS-Chem). Section 2 describes the approach and
data. Section 3 evaluates the GOCI AOD, the simulated PM2.5 to AOD
relationship, and the GOCI-derived PM2.5 using recently available
ground-level measurements from the China Environmental Monitoring Center
(http://113.108.142:20035/emcpublish/). We also interpret the GOCI-derived
PM2.5 by using the GEOS-Chem model to estimate its chemical
composition. Section 4 summarizes the major findings and potential future
improvements of the current analysis.
Methods
Aerosol optical depth from the GOCI satellite instrument
GOCI operates onboard the Communication, Ocean, and Meteorology Satellite
(COMS) that was launched in 2010 in Korea (Lee et al., 2010). The spatial
coverage of GOCI is 2500 km × 2500 km in Northeast Asia, including eastern
China, the Korean peninsula and Japan (Kang et al., 2006). GOCI has eight
spectral channels for aerosol retrievals, including six visible bands at
412, 443, 490, 555, 660, 680 nm and two near infrared bands at 745 and 865 nm (Park et al., 2014). The Level 2 AOD products are retrieved at a spatial
resolution of 6 km by 6 km, using a clear-sky composite method for surface
reflectance and a lookup table approach based on AERONET observations (Lee
et al., 2010, 2012).
A challenge using GOCI to detect aerosols in the atmosphere is the absence
of mid-infrared (IR) channels to detect clouds, which means that significant
errors could be induced in the estimates of AOD. The operational GOCI
products screen clouds based on spatial variability and threshold tests at
each 6 km × 6 km pixel in combination with a meteorological imager that has
four IR channels (at 3.7, 6.7, 10.8, 12 µm
wavelengths) at 4 km by 4 km resolution onboard the same satellite (Cho and Youn, 2006). However, as will be shown here, cloud contamination still occurs.
Therefore, we apply a set of spatial filters following Hyer et al. (2011)
and temporal filters to further eliminate cloud contamination in GOCI AOD.
The filters include (1) a buddy check that sets a minimum number of 15 retrievals per 30 km × 30 km grid cell, (2) a local variance check to
eliminate grid cells where the coefficient of variation of AOD is larger
than 0.5 within the surrounding 5 × 5 grid cells and (3) a diurnal variation
check that excludes grid cells with diurnal variation (maximum – minimum)
of AOD larger than 0.74 which is the 90th percentile of diurnal
variation of AERONET AOD in Beijing and northern Taiwan for 2013. In this
study, we use GOCI AOD for January–December 2013 to derive ground-level PM2.5 in
eastern China.
Aerosol optical depth from AERONET ground-based measurements
The Aerosol Robotic Network (AERONET) is a globally distributed network of
CIMEL Sun photometers (Holben et al., 1998) that provide multi-wavelength
AOD measurements with a low uncertainty of < 0.02 (Holben et al.,
2001). Here we use AERONET Level 1.5 cloud screened data (Smirnov et al.,
2000) for January–December 2013 from four stations within the GOCI domain: Beijing,
Beijing-CAMS, Taipei_CWB and EPA-NCU. AERONET Level 2 data
for 2013 are not available for some stations discussed in this paper. Thus,
we use Level 1.5 for consistency. We compared Level 2 and Level 1.5 data for
2013 for stations that do have Level 2 available, and found Level 1.5 AOD is
highly consistent with Level 2 AOD with RMSE of 0.01–0.02 (relative RMSE of
2–7 %). Criteria for selecting an AERONET station are (1) a PM2.5
ground monitor has to be located within 10 km and (2) a complete time series
of AOD data records for the period of study has to be available. Beijing and
Beijing-CAMS stations are located in downtown Beijing, with the closest
available PM2.5 monitors 9.5 and 7.5 km away, respectively. However,
due to interrupted time series of PM2.5 records at both these stations,
we combine the AERONET AOD from the Beijing and Beijing-CAMS stations and
PM2.5 from the corresponding two in situ ground-based sites as a “combined
Beijing” site. Taipei_CWB and EPA-NCU stations are located
in populated northern Taiwan, with nearly collocated PM2.5 monitors
(< 3 km). We similarly combine the Taipei_CWB and
EPA-NCU as “northern Taiwan” site. We use these sites to evaluate GOCI AOD
and the relationship between AOD and PM2.5 simulated by a global
chemical transport model.
Simulation of the relationship between AOD and PM2.5 by
GEOS-Chem
We use the GEOS-Chem chemical transport model (version 9-01-03;
http://geos-chem.org) to calculate the spatiotemporally resolved
relationship between ground-level PM2.5 and satellite-retrieved column
AOD.
Our nested GEOS-Chem simulation at 1/2∘ × 2/3∘ spatial
resolution with 47 vertical levels (14 levels in the lowest 2 km) is driven
by assimilated meteorology from the Goddard Earth Observing System (GEOS-5).
A global simulation at 2∘ × 2.5∘ spatial resolution is
used to provide boundary conditions for the nested domain (Wang et al.,
2004). We spin up the model for 1 month before each simulation to remove
the effects of initial conditions on the aerosol simulation.
GEOS-Chem includes a fully coupled treatment of tropospheric oxidant–aerosol
chemistry (Bey et al., 2001; Park et al., 2004). The GEOS-Chem aerosol
simulation includes the sulfate-nitrate-ammonium system (Park et al., 2004;
Pye et al., 2009), primary (Park et al., 2003) and secondary (Henze and Seinfeld, 2006; Henze et al., 2008; Liao et al., 2007; Fu et al., 2008) organics,
mineral dust (Fairlie et al., 2007), and sea salt (Jaeglé et al., 2011). We
estimate the concentration of organic matter (OM, which includes elements
such as hydrogen, oxygen and nitrogen) from the simulated primary organic
carbon (OC) using spatially and seasonally resolved values from OMI (Ozone
Monitoring Instrument) NO2 and AMS (aerosol mass spectrometer)
measurements following Philip et al. (2014). Gas-aerosol phase partitioning
is simulated using the ISORROPIA II thermodynamic scheme (Fountoukis and
Nenes, 2007). GEOS-Chem calculates AOD using relative humidity-dependent
aerosol optical properties following Martin et al. (2003). Dust optics are
from Ridley et al. (2012).
Anthropogenic emissions are based on the Multi-resolution Emission Inventory
for China (MEIC; http://www.meicmodel.org) for 2010, and the Zhang et al. (2009) inventory for surrounding East Asia regions for 2006. Both
inventories are scaled to the simulation year (2012–2013), following Ohara
et al. (2007). Non-anthropogenic emissions include biomass burning emissions
(GFED-3) (Mu et al., 2011), biogenic emissions (MEGAN) (Guenther et al.,
2006), soil NOx (Yienger and Levy, 1995; Wang et al., 1998), lightning
NOx (Murray et al., 2012), aircraft NOx (Wang et al., 1998;
Stettler et al., 2011), ship SO2 from EDGAR (Olivier and Berdowski, 2001) and
volcanic SO2 emissions (Fischer et al., 2011). HNO3 concentrations
are artificially decreased to 75 % of their values at each timestep
following Heald et al. (2012) to account for regional bias (Wang et al.,
2013). Emissions are distributed into the lower mixed layer, with a
correction to the GEOS-5 predicted nighttime mixing depths following Heald
et al. (2012) and Walker et al. (2012).
We apply GEOS-Chem to simulate daily relationships between ground level
PM2.5 and column AOD, specifically PM2.5 / AOD. PM2.5
concentrations are calculated at 35 % relative humidity for consistency
with in situ measurements. For consistency with GOCI AOD and PM2.5 ground-based measurements, we sample the simulated AOD only from hours
that GOCI has retrievals (00:00–07:00 UTC), and calculate the simulated
daily PM2.5 from 24 h averages as reported for the ground-based
PM2.5 measurements. The simulation period is May 2012–April 2013 as
the GEOS-5 meteorological fields are not available afterward. The mismatch
with observations for May–December 2013 has the potential to degrade performance,
but as will be shown here no clear loss of quality is apparent.
In situ PM2.5 measurements
We collect PM2.5 measurements from 494 monitors to evaluate the
GOCI-derived values. In situ PM2.5 daily measurements in Mainland
China for 2013 are primarily from the official website of the China
Environmental Monitoring Center (CEMC;
http://113.108.142:20035/emcpublish/). Data are also collected from
some provinces (e.g., Shandong, Zhejiang) and municipalities (e.g., Beijing
and Tianjin) with additional sites that are not included in the CEMC website.
Daily in situ PM2.5 data in northern Taiwan for 2013 are from the
Taiwan Environmental Protection Administration (TEPA;
http://taqm.epa.gov.tw). The in situ PM2.5 data in both Mainland
China and northern Taiwan are measured by a collection of the tapered element
oscillating microbalance methods (TEOMs) and beta-attenuation methods (BAMs)
with some TEOMs being heated to 30 ∘C and others to 50 ∘C
(CNAAQS, GB3095-2012, 2012;
http://taqm.epa.gov.tw). The specific instrument (BAMs or TEOMs) used
by each monitoring site is unknown. The effective relative humidity of the
resultant PM2.5 measurement likely varies diurnally and seasonally as a
function of the ambient temperature. Semivolatile losses are expected from
the TEOMs. The network design appears to include compliance objectives that
may affect monitor placement. Despite these issues, we use the monitoring
data to evaluate our satellite-derived PM2.5 since the monitoring data
offer valuable information about ground-level PM2.5 concentrations. We
also collect PM2.5 measurements from a monitor in Beijing as part of the
surface particulate matter Network (SPARTAN; www.spartan-network.org)
using a three-wavelength nephelometer and an impaction filter sampler (Snider
et al., 2015). The SPARTAN, CEMC and TEPA PM2.5 monitoring data combined
with AERONET AOD are used to estimate the empirical relationship between
PM2.5 and AOD, and to further evaluate the relationship simulated by the
model.
Statistical terms
Root mean square error (RMSE), relative root mean square error (rRMSE), mean
fractional bias (MFB) and mean fractional error (MFE) are defined as
RMSE=1N∑i=1N(Si-Oi)2rRMSE=RMSE1N∑i=1NOiMFB=1N∑i=1N(Si-Oi)Si+Oi2×100%MFE=1N∑i=1N|Si-Oi|Si+Oi2×100%,
where Si is the satellite-derived value of the parameter in question,
Oi is the corresponding observed value, and N is the number of
observations.
Coefficient of variation (CV) is defined as
CV=StandarddeviationMean
Results and discussion
Evaluation of satellite AOD and the simulated relationship between
PM2.5 and AOD
Figure 1 shows the effects of our cloud-screening filters on GOCI AOD. The
left panel shows GOCI true color images from 5 July 2013 at 10:30 (top) and
11:30 (bottom) Korean Standard Time. The boxes identify challenging regions
with thick white cloud, dark cloud-free oceans and gray shading that appears
to be thin cloud. The operational GOCI AOD retrievals, shown in the middle
panel, correctly exclude thick clouds, but report high AOD for the
potentially thin clouds. Although these gray regions could contain aerosol,
we err on the side of caution. Application of our additional temporal and
spatial cloud filters removes the suspicious pixels from the original GOCI
data, as shown in the right panel. Our filters reject 10.3 % of all the
operational GOCI AOD data investigated in this study. We evaluate the cloud
filters further below.
Figure 2 (top) shows monthly averages of coincident filtered hourly GOCI and
AERONET AOD for January–December 2013 at combined Beijing and northern Taiwan
stations. GOCI AOD is highly consistent with AERONET observations with MFB
of 6.7 % in Beijing and -1.2 % in Taiwan. GOCI AOD and AERONET AOD are
positively skewed at both stations, and the skewness is reduced in GOCI AOD
at both stations due to more records for extremely small AOD (< 0.04) in GOCI products. The relatively larger rRMSE between GOCI and AERONET
AOD in northern Taiwan may reflect the fewer observations there.
The GOCI granules from 5 July 2013, 10:30 (top) and 11:30
(bottom) Korean Standard Time. From left to right on each panel are the GOCI
true color images, the operational AOD retrievals and the AOD retrievals
after applying temporal and textual filters to reduce cloud contamination.
The boxes highlight examples of challenging cloud fields, and are enlarged
within the lower right subplot of each panel.
We investigate the filtered diurnal variation of GOCI AOD at the above
AERONET stations and find the level of AOD is uniform within a day (e.g., the
coefficient of variation in Beijing is 0.1), similar to AERONET
observations.
The effect of excluding our cloud-screening filters is negligible for
coincident comparisons with AERONET since AERONET is already cloud-screened.
The exclusion of our cloud filters for a non-coincident comparison that
includes all GOCI data would introduce significant error versus AERONET
observations, increasing rRMSE by a factor of 1.7–3.3 in Beijing and
northern Taiwan. Changing the buddy check threshold in our cloud filters
from 15 to 10 would significantly underestimate AOD especially in northern
Taiwan where the MFB would increase from -1.2 to -15.0 %. Decreasing
the threshold of local variance check to 0.4 has little influence (< 0.1 %) for rRMSE, MFB and MFE, but would have larger influence on
GOCI-derived PM2.5 as will be shown later. Limiting the diurnal
variation of GOCI AOD to the 80th percentile of diurnal variations in
observations would introduce bias (rRMSE would increase by 4 % in Beijing)
to GOCI AOD. As will be shown here, GOCI-derived PM2.5 offers an
additional test of cloud screening filters. Figure 2 (bottom) shows the
relationship between the ground level PM2.5 and the columnar AOD as
simulated by GEOS-Chem and from ground-based measurements. The measured
ratio in Beijing has pronounced seasonal variation with values high in
winter and low in spring. The measured ratio in northern Taiwan exhibits
little seasonal variation. The annual mean GEOS-Chem PM2.5 / AOD ratio
well reproduces the ground-based measurements despite the temporal
inconsistency of the two metrics for May–December. The simulation captures the
pronounced seasonal variation in Beijing and the comparably aseasonal
behavior in northern Taiwan. The simulated seasonal variation of
PM2.5 / AOD in Beijing arises from the seasonal variation of mixed-layer
depth (factor of 2 higher in summer than winter) combined with the
near-constant columnar AOD throughout the year as shown in Fig. 2 (top).
Top: monthly time series of AOD from AERONET and GOCI for
January–December 2013. Numbers above the x axis denote the number of coincident
hourly observations in each month. Bottom: monthly averages of
PM2.5 / AOD from ground measurements and the GEOS-Chem simulation at
AERONET sites. The ground-based ratio is sampled from daily ground
PM2.5 coincident with AERONET AOD for January–December 2013. The GEOS-Chem
simulation is for May 2012–April 2013, noncoincident with the ground-based
ratio for May–December 2013. Numbers above the x axis denote the number of
daily ground-based observations in each month. Error bars represent standard
errors. Statistics are root mean square error (RMSE), relative root mean
square error (rRMSE), mean fractional bias (MFB) and mean fractional error
(MFE).
Snider et al. (2015) interpreted coincident measurements of AOD, PM2.5,
and nephelometer measurements of aerosol scattering and found that the
temporal variation of the PM2.5 / AOD ratio in Beijing was primarily
driven by the vertical profile in aerosol scattering. We examine the
seasonal variation in the simulated PM2.5 / AOD and similarly find that
the ratio of ground-level aerosol scatter to columnar AOD contributes most
(89 %) of the monthly variability in the PM2.5 / AOD ratio in Beijing.
Evaluation of ground-level PM2.5 derived from GOCI AOD
Figure 3 shows the seasonal and annual distribution of PM2.5 over East
Asia at a spatial resolution of 6 km by 6 km for 2013. In both GOCI-derived
and measured PM2.5, winter concentrations in eastern China exceed 100 µg m-3 over vast regions, with lower values in summer. Both
GOCI-derived and in situ measurements reveal that PM2.5 in northern China is
higher than in southern China, especially for the Beijing, Hebei and
Shandong provinces where the annual PM2.5 is almost 100 µg m-3 or more. Prior work has attributed this regional enhancement to
high emission rates (Zhao et al., 2013; Zhang et al., 2013) that in part
arises from emissions when producing goods for exports (Jiang et al., 2015).
Figure 4 compares annual and seasonal averages of daily ground-measured
PM2.5 from 494 sites with coincident daily GOCI-derived PM2.5
from pixels that contain the ground-based sites. A significant correlation
(r2= 0.66, N=494) with a slope near unity (1.01) is found in the
annual scatter plot. The slope remains near unity (0.95–1.01) in seasonal
scatter plots. The weaker correlation for all four seasons implies random
representativeness differences between point in situ measurements and area-averaged
satellite values when data density diminishes. Semivolatile losses from some
in situ instruments (TEOMs) might contribute to scatter in winter when nitrate
constitutes a larger fraction of PM2.5. We focus on more meaningful
aggregated measurements. Using the same technique, we also estimated
PM2.5 from MODIS Collection 6 AOD for 2013, and found GOCI-derived
PM2.5 achieves greater consistency than MODIS-derived PM2.5 when
compared with ground-based measurements (slope = 1.1, r2= 0.61).
GOCI-derived PM2.5 also corrects the significant underestimation of
PM2.5 from GEOS-Chem (slope = 0.68, r2= 0.85) when compared with
ground measurements.
Seasonal and annual distribution of PM2.5
concentrations at 6 km by 6 km resolution over East Asia for 2013. The
background color indicates averages of GOCI-derived daily surface PM2.5
concentrations. Filled circles represent averages of daily ground-based
measurements of PM2.5. Gray denotes missing values. Boxes in the annual
map denote regions used for monthly comparisons in Fig. 5 from top to
bottom: Beijing and surrounding areas, Shandong and surrounding regions,
Shanghai and surrounding areas and northern Taiwan.
Figure 5 shows monthly averages of GOCI-derived PM2.5 and in situ measurements
at four regions outlined in Fig. 3. Regions are selected based on the level
of PM2.5 concentration and the population of residents. A high degree
of consistency is found in all regions. Both data sets show more seasonal
variation in northern regions like Beijing and Shandong than southern
regions like Shanghai and northern Taiwan. Both indicate that PM2.5
concentrations in northern regions are generally higher than in southern
regions. The exclusion of our cloud screening filters from the GOCI AOD
would introduce significant bias in GOCI-derived PM2.5 versus
ground-based measurements especially in summer, increasing rRMSE by a factor
of 1.7–5.3 in all four regions. Changing the threshold of local variance
check in our cloud filters to 0.4 would introduce bias by restricting the
variation of PM2.5 concentrations. For example, GOCI-derived PM2.5
would be generally underestimated in Beijing areas (rRMSE = 16.6 % and
MFB =-6.8 %) and Shandong areas (rRMSE = 16.6 % and MFB =-3.42 %).
Scatterplots of the annual mean (left) and seasonal mean
(right) GOCI-derived PM2.5 for 2013 against PM2.5 from 494 ground
monitors over the GOCI domain in eastern China.
Monthly averages of daily PM2.5 from in situ measurements
and daily PM2.5 estimated from GOCI AOD for 2013. Regions are defined
in Fig. 3. Error bars represent standard errors. Statistics are root mean
square error (RMSE), relative root mean square error (rRMSE), mean
fractional bias (MFB) and mean fractional error (MFE).
Monthly variation of GOCI-derived PM2.5 and
in situ PM2.5 for 2013 over eastern China, with chemical composition for
GOCI-derived PM2.5. The in situ PM2.5 is determined from the averages of
all ground stations in eastern China for 2013 and GOCI-derived PM2.5 is
calculated from the average of all grid boxes that contain PM2.5 ground
monitors. The chemical composition is calculated by applying the GEOS-Chem
simulated mass fraction of PM2.5 chemical components to GOCI-derived
PM2.5 mass concentration. Aerosol water is associated with
each PM2.5 component according to its hygroscopicity. Error bars
represent standard errors.
Seasonal variation of PM2.5
Figure 6 shows the monthly averages of coincident daily GOCI-derived and
in situ PM2.5 concentrations for the domain of eastern China. Both the
GOCI-derived PM2.5 and ground-based observations exhibit similar
seasonal variation with values high in winter and low in summer. Exclusion
of our temporal and spatial cloud-screening filters from GOCI-derived
PM2.5 would increase rRMSE by a factor of 3.4. Figure 6 also shows the
chemical composition of GOCI-derived PM2.5, as calculated by applying
the GEOS-Chem simulated mass fraction of PM2.5 chemical components to
GOCI-derived PM2.5 mass concentration. Aerosol water is attached to
each component according to its hygroscopicity. Secondary inorganic aerosols
(SIA; SO42-, NO3-, NH4+) are the most abundant
components throughout the year, accounting for 65 % of PM2.5
concentrations, followed by OM (18 %). The NO3- and OM
concentrations increase by a factor of 2 in winter, together comprising most
of PM2.5 (31 % for NO3- and 26 % for OM). Summer is
predominately controlled by SIA (74 %). Dust plays an important role in
spring (15 %) and fall (15 %). Our seasonal variation of chemical
composition is generally consistent with ground-based measurements in
previous works across eastern China. A number of studies in Beijing, the
Yangtze River delta and Pearl River delta regions all reported that OM and
SIA are the most important components of PM2.5 through the year (He et
al., 2001; Ye et al., 2003; Tao et al., 2012; Zhang et al., 2013). Zhang et
al. (2008) showed consistent seasonal patterns in OM at 18 stations in
China, with a winter maximum, and a summer minimum, similar to the
seasonality of OM in this work. Zhang et al. (2013) studied the chemical
composition of PM2.5 in Beijing and found the percentage of SIA in
PM2.5 is largest in summer, consistent with our result.
Annual PM2.5 concentrations, area-weighted concentrations of
chemical composition and affected population of PM2.5 in regions
outlined in Fig. 3 and in overall eastern China (excluding northern Taiwan)
for 2013. Aerosol water is not associated with each PM2.5 component for
consistency with measurement protocols. PM2.5 concentration is at
35 % relative humidity. IT1 refers to the WHO air quality interim
target-1 of 35 µg m-3.
Region
Beijing
Shandong
Shanghai
Eastern
Northern
China
Taiwan
Population-weighted GOCI-derived
90.8
89.1
56.9
53.8
18.9
PM2.5 (µg m-3)
Area-weighted GOCI-derived
86.5
89.1
51.0
44.3
23.6
PM2.5 (µg m-3)
SO42- (µg m-3)
12.8
14.0
9.2
13.1
5.1
NO3- (µg m-3)
14.5
16.1
8.5
4.2
2.1
NH4+ (µg m-3)
8.9
9.8
5.7
3.3
2.2
OC (µg m-3)
10.3
9.6
4.3
2.9
1.6
BC (µg m-3)
6.3
5.2
2.6
1.6
0.8
Dust (µg m-3)
9.1
8.3
4.9
4.4
2.9
Sea Salt (µg m-3)
0.2
0.4
0.9
1.9
2.2
OM (µg m-3)
17.1
15.7
7.4
5.4
3.0
Population (million people) exposed
37.8
88.8
98.3
432.8
1.5
to PM2.5 exceeding IT-1 level
The seasonal variation of PM2.5 in Fig. 6 is driven by a combination of
meteorological conditions, emissions, and nitrate formation. All three
processes have greater seasonal variation in the north than south. The
mixing height over northeastern China has strong seasonal variation with
summer having an average mixing height from GEOS-5 that is 1.9 times higher
than in winter. The GEOS-Chem simulation reveals that the increase of OM in
winter is primarily driven by biofuel emissions from burning wood, animal
waste and agricultural waste (Bond et al., 2004) for heating in eastern
China. The spatial distribution of biofuel emission is primarily north of
the Yangze River, especially from the North China Plain. The significant
contribution from biofuel emissions to the OM concentration in our work is
consistent with Bond et al. (2004) who found residential biofuel emissions
were responsible for ∼ 70 % of OC emissions in China. The
increase of NO3- in winter in Fig. 6 is consistent with prior
attribution of the increase of NO3- in winter to the favorable
formation of NH4NO3 at low temperatures (Wang et al., 2013).
Figure S1 in the Supplement shows the spatial distribution of PM2.5 chemical
components.
Table 1 shows the annual chemical composition of GOCI-derived PM2.5 in
regions outlined in Fig. 3 and in overall eastern China. SIA and OM are the
most abundant species. Among the SIA components, SO42- and
NO3- concentrations are similar in the Beijing, Shandong and
Shanghai regions, whereas in eastern China and northern Taiwan
SO42- is the dominant component. OM concentrations in the Beijing
and Shandong regions are considerably higher than in the other regions,
similar to or even exceeding the concentrations of SO42- and
NO3-. Our estimation of PM2.5 composition is generally
consistent with in situ measurements in prior studies. In Beijing, the
concentrations of SIA in this work are similar to Zhang et al. (2013)
who measured concentrations for 2009–2010 of 13.6 ± 12.4 µg m-3 for SO42-, 11.3 ± 10.8 µg m-3 for
NO3- and 6.9 ± 7.1 µg m-3 for
NH4+. Our SIA concentrations in Beijing are also comparable with
Yang et al. (2011) who measured concentrations for 2005–2006 of 15.8 ± 10.3 µg m-3 for SO42-,
10.1 ± 6.09 µg m-3 for NO3- and 7.3 ± 4.2 µg m-3 for NH4+. The OC concentration in Beijing in this work is
smaller than Zhang et al. (2013) of 16.9 ± 10.0 µg m-3 and
Yang et al. (2011) of 24.5 ± 12.0 µg m-3. In Shandong and
surrounding regions, our concentrations are smaller than in Cheng et al. (2011) by a factor of about 2, perhaps related to unresolved sources. Our
results in Shanghai cluster are comparable with Yang et al. (2011) for
1999–2000, except the OC concentration (4.3 µg m-3) in this work is considerably lower than that of 16.8 µg m-3 in Yang et al. (2011). In northern Taiwan, our
NO3- is similar to Fang et al. (2002) for 2001–2003, yet our
estimations of SO42- and NH4+ are higher than Fang et
al. (2002) by a factor of two, which could be driven by changes in emissions
over the last decade. In summary, the chemical composition broadly
represents in situ measurements with some location-dependent discrepancies.
Population exposure to ambient PM2.5 in eastern China
We estimate the population exposure to ambient PM2.5 in eastern China
for 2013 at a spatial resolution of 6 km by 6 km using our GOCI-derived
PM2.5 and the Gridded Population of the World (GPW; Tobler et al.,
1997) data for 2010 from the Socioeconomic Data and Applications Center (GPW
version 3; http://sedac.ciesin.columbia.edu/). Table 1 also provides the
population-weighted GOCI-derived PM2.5 for regions outlined in Fig. 3
and for overall eastern China. The population-weighted PM2.5 exceeds
the area-weighted for all regions except northern Taiwan and Shandong and
surrounding regions. The overall population-weighted PM2.5
concentration for eastern China for 2013 is 53.8 µg m-3. The
level of PM2.5 for Beijing and Shandong regions in this study is
similar to Ma et al. (2014) who suggested that the PM2.5 concentration
over the North China Plain for 2013 is 85–95 µg m-3. The
PM2.5 concentration in eastern China in this study is also comparable
with previous works. Van Donkelaar et al. (2015) estimated that the PM2.5 concentration over eastern Asia for 2001–2010
is 50.3 ± 24.3 µg m-3. Geng et al. (2015) estimated that the PM2.5 concentration in
China for 2006–2012 is 71 µg m-3, higher than our work.
According to the World Health Organization (WHO) Air Quality Interim
Target-1, an annual mean PM2.5 concentration of 35 µg m-3
or higher is associated with about 15 % increased risk of premature
mortality. As shown in Table 1, population-weighted PM2.5 for eastern
China considerably exceeds the Interim Target-1 level of PM2.5
concentration, especially in Beijing and Shandong regions where the
PM2.5 concentration is almost triple the Interim Target-1 level. These
elevated concentrations threaten the health of 433 million inhabitants
(Table 1) in eastern China who live in regions that exceed this target.
Conclusions
We estimated the ground-level concentration of PM2.5 in eastern China
for 2013 using AOD retrieved from the GOCI satellite instrument, coupled
with the relationship of AOD to PM2.5 simulated by a global chemical
transport model (GEOS-Chem). GOCI-derived PM2.5 was compared with
in situ measurements throughout eastern China.
We applied a set of filters to GOCI AOD to remove cloud contamination. The
filtered GOCI AOD showed significant agreement with AERONET AOD at Beijing
and northern Taiwan (MFB of 6.7 to -1.2 %). We also evaluated the
simulated relationship of PM2.5 and AOD from GEOS-Chem by using an
empirical relationship calculated from nearly collocated ground-based
PM2.5 monitors and AERONET AOD stations. A high degree of consistency
was observed between the GEOS-Chem simulation and ground-based measurements
with MFB of -0.52 to 8.0 %.
The GOCI-derived PM2.5 were highly consistent with in situ measurements,
capturing the similar seasonal and spatial distribution throughout eastern
China. The highest PM2.5 concentrations were found in winter over
northern regions. The annual averages of GOCI-derived PM2.5 were
strongly correlated (r2= 0.66) with surface measurements with a slope
near unity (1.01). Monthly comparison of GOCI-derived PM2.5 with
ground-based measurements across the entire region of eastern China was also
in good agreement with rRMSE = 18.9 %. The exclusion of our
cloud-screening filters in GOCI retrievals would introduce significant bias
in GOCI-derived PM2.5, especially in summer and would increase the
rRMSE by a factor of 1.7–5.3.
The chemical composition of GOCI-derived PM2.5 revealed that secondary
inorganic aerosols (SIA; SO42-, NO3-, NH4+)
and organic matter (OM) dominated throughout the year. NO3- had a
winter maximum due to aerosol thermodynamics. OM increased by a factor of 2
in winter, which was primarily driven by biofuel emission for heating in
northern China. Dust played an important role in spring and fall.
The population-weighted GOCI-derived PM2.5 for 2013 at 6 km by 6 km
resolution in eastern China was 53.8 µg m-3, suggesting
∼ 400 million people in China live in regions with PM2.5
concentrations exceeding the suggested 35 µg m-3 by the World
Health Organization (WHO) Air Quality Interim Target-1, of which
∼ 130 million people in Beijing and Shandong regions are
seriously threatened by even higher PM2.5 concentrations.
Population-weighted PM2.5 of pixels containing ground-based monitors is
much higher at 82.4 µg m-3, suggesting the value of the newly
established PM2.5 network to monitor these seriously polluted regions.
The satellite measurements of AOD from the GOCI instrument coupled with the
relationship between AOD and PM2.5 simulated by a chemical transport
model have the potential to provide a unique synopsis of ground-level
PM2.5 concentrations at fine spatial resolution in the most polluted
and populated part of China. Further development of this capability will
depend on both the quality of GOCI aerosol products and the aerosol
simulation. Assimilating satellite observations of trace gases from the
forthcoming GEMS (Geostationary Environment Spectrometer) geostationary
platform would provide additional constraints on PM2.5 composition.