Introduction
Aerosols can have important effects on the Earth's climate over regional to
global scales, but there are still uncertainties in the strengths and
significance of these impacts (Hansen et al., 2000; Solomon et al., 2007;
Schwartz and Andreae, 1996). Aerosols affect the radiative balance of the
Earth–atmosphere system by directly scattering and absorbing solar
radiation (Charlson et al., 1992; Ackerman and Toon, 1981), and they can
affect climate indirectly through aerosol–cloud interactions (Twomey et
al., 1984; Albrecht, 1989; Z. Q. Li et al., 2016).
The physical and optical properties of aerosol particles determine their
radiative effects, and information on these properties can be used to
predict and assess global and regional changes in the Earth's climate (Eck
et al., 2005; Myhre, 2009; IPCC, 2013; Panicker et al., 2013).
Long-term, ground-based observations have contributed greatly to our
understanding of the spatial variations in aerosols and their effects on the
Earth's climate (Holben et al., 2001; Kaufman et al., 2002; Sanap and
Pandithurai, 2014; Z. Q. Li et al., 2016). Ground-based monitoring networks have
been established worldwide – for instance, AERONET (Aerosol Robotic Network)
(Holben et al., 1998; Goloub et al., 2007), SKYNET (SKYrad Network) (Takamura
and Nakajima, 2004), EARLINET (European aerosol Lidar network) (Pappalardo et al.,
2014) and the GAW-PFR Network (Global Atmosphere Watch Programmer-Precision
Filter Radiometers) (Wehrli, 2002; Estellés et al., 2012). In China, CARSNET
(the China Aerosol Remote Sensing NETwork) and CSHNET (the Chinese Sun
Hazemeter Network) were established to obtain data on aerosol optical
characteristics (H. Che et al., 2009, 2015; Xin et al., 2007).
High-frequency, ground-based measurements of aerosol optical properties made
at these stations have improved our understanding of the sources, transport
and diurnal variations of air pollutants, and they have provided insights
into the aerosols' effects on climate. Ground-based observations are also
useful for the validation of satellite retrievals (Holben et al., 2017; Xie
et al., 2011).
Most of the ground-based studies of the optical properties of aerosols in
China have been conducted in urban regions that have been undergoing rapid
economic development. Those sites typically have had high aerosol loadings
and in many cases serious environmental problems (Cheng et al., 2015; Pan et
al., 2010; Xia et al., 2013; L. C. Wang et al., 2015; H. Z. Che et al., 2015). Detailed
information on aerosol optical depth (AOD), the types of aerosols and
especially the size and absorption properties of ambient populations over a
wide sampling of regions is needed to understand the effects of aerosols on
the Earth's climate and the environment (Giles et al., 2011; H. Z. Che et al.,
2009; Wang et al., 2010; Zhu et al., 2014). In particular, the aerosol
direct radiative forcing is sensitive to the aerosol radiation absorptivity
(Haywood and Shine, 1995). Therefore, it is important to understand the
connections between the aerosol types and absorption properties because that
information can be used for comparisons and validation of chemical transport
models and satellites (J. Lee et al., 2010).
The Yangtze River Delta (YRD) region in eastern China has recently undergone
rapid economic growth, and the loadings of aerosols in the region can be
very high during heavy pollution episodes (Fu et al., 2008; Zhang et al.,
2009). Studies of aerosol optical properties in eastern China have
contributed to our understanding of local air quality and regional climate
impacts (Duan and Mao, 2007; Pan et al., 2010; Ding et al., 2016). In the
YRD, investigations of aerosol optical properties have been conducted in
Nanjing, Hefei, Shanghai, Shouxian and Taihu (Zhuang et al., 2014; S. Li et al.,
2015; Wang et al., 2015; He et al., 2012; K. H. Lee et al., 2010; Cheng et al.,
2015; Xia et al., 2007). Those studies mostly involved sampling at single
sites ∼ 100 km apart from one other without synchronous
observations, and many have been of relatively short duration, and so there
remains a need for more extensive ground-based measurements.
For the present study, sampling was conducted over a period of several years
to better characterize the climatology of the aerosol microphysical and
optical properties, including aerosol absorptivity, and to improve estimates
of direct aerosol radiative forcing. For these studies, sun photometer
measurements were made at 3 min intervals from 2011 to 2015 at seven CARSNET
sites (one densely populated urban site, five urban center sites in smaller
cities and one rural site) ∼ 10–40 km apart in the YRD. The
dense network of ground-based sun-scanning and sky-scanning spectral radiometers
improves the temporal and spatial coverage of the data, and that has enabled
us to capture small-scale variations in the aerosols. The results not only
contribute to our understanding of regional climate and local air quality
impacts, but they also will be useful for validation of satellite data and
improving the performance of models and remote sensing algorithms in the
future.
This paper is organized as follows: Sect. 2 describes the sites, the
methods used for retrieving the aerosol optical properties and their
uncertainties and the calculation of aerosol direct radiative forcing from
the retrieved aerosol optical parameters. Section 3 presents the aerosol
microphysical properties, optical properties and calculations of direct
radiative forcing. The aerosol type classification is also presented based
on the aerosol optical parameters. Then a brief discussion is made about the
analysis of this study. Section 4 details the conclusions.
Site descriptions, measurements and methods
Locations and elevations of the seven CARSNET sites in the Yangtze
River Delta.
![]()
Figure 1 shows the locations of the seven CARSNET sites of the YRD, and
detailed information on the sites is included in Table 1. Hangzhou is a
densely populated urban site with heavy vehicular traffic, and it is
affected by various types of anthropogenic emissions. LinAn, Fuyang, Jiande,
Xiaoshan and Tonglu are urban center sites in smaller cities, and they are
all affected to varying degrees by anthropogenic activities, especially
pollutants from industries and agriculture. The rural site of ChunAn has a
small population, and there are few industrial sources nearby, so the effects
from local or regional pollution are relatively small.
Sun photometers (CE-318, Cimel Electronique, Paris, France) were installed
at each of the seven sites and operated from 2011 to 2015. These instruments
were standardized and calibrated using CARSNET reference instruments (H. Che et
al., 2009), which in turn were periodically calibrated at Izaña,
Tenerife, Spain, in conjunction with the AERONET program. The cloud-screened
AODs (based on the work of Smirnov et al., 2000) at 340, 380, 440, 500, 670,
870, 1020 and 1640 nm with uncertainties less than 0.01 (Eck et al., 1999)
were obtained using ASTPwin software (Cimel Electronique). The water vapor
expressed as precipitable water in the column was derived from the 940 nm
channel with uncertainties less than 10 % (Eck et al., 1999). Daily
averages and statistical analysis were calculated for days on which
instantaneous AOD measurements were made more than 10 times (H. Che et al.,
2015). The extinction Angström exponent values (EAE) were calculated
from AOD values at 440 and 870 nm.
Aerosol microphysical properties were retrieved from the almucantar sky
irradiance measurements in conjunction with measured spectral AOD, following
the methods of Dubovik and King (2000) and Dubovik et al. (2002, 2006). The
dataset contained information on (1) volume size distributions in 22 size
bins for particle radii 0.05–15 µm; (2) fine-mode and coarse-mode
aerosol effective radii; and (3) aerosol optical properties – including the
wavelength-dependent single-scattering albedo (SSA), the complex refractive
index, the absorption AOD (AAOD), and the absorption Angström exponent
(AAE). For the retrieval process, the surface albedo (SA) was
interpolated/extrapolated to 440, 670, 870 and 1020 nm from the daily
Moderate Resolution Imaging Spectroradiometer (MODIS) surface reflectance
product of MCD43C3 (https://ladsweb.modaps.eosdis.nasa.gov/). Following
the procedures of Dubovik et al. (2002, 2006), all particles smaller than
0.992 µm were considered fine-mode particles, while those larger
than 0.992 µm were considered coarse-mode particles. And the effective radii
for the total (Refft), fine-mode
(Refffine) and coarse-mode
(Reffcoarse) aerosols are calculated as follows:
Reff=∫rminrmaxr3dNrdlnrdlnr∫rminrmaxr2dNrdlnrdlnr,
where rmin is 0.05, 0.05 and 0.992 µm and rmax is 15, 0.992
and 15 µm for the total, fine-mode and coarse-mode aerosols, respectively.
The inversion algorithm used for calculating the aerosol volume distribution
(dV/lnr) assumed a homogeneous distribution of nonspherical aerosol particles
as in the work of Dubovik et al. (2006); this approach has been widely applied in
studies of many different areas of the world. The SSA was retrieved using
only AOD440nm > 0.40 measurements to avoid the large
uncertainties inherent in low AOD retrievals (Dubovik et al., 2002, 2006).
Real and imaginary parts of the refractive index at 440, 675, 870 and 1020 nm
were constrained to the ranges of 1.33–1.60 and 0.0005–0.50, respectively
(Dubovik and King, 2000; H. Che et al., 2015). The complex refractive index is
assumed to be independent of particle size. This assumption is valid for fine-dominated or
coarse-dominated cases; however, it could cause some errors in SSA and particle
size retrievals for mixed aerosol scenarios (Xu and Wang, 2015; Xu et al
2015). The AOD, AAOD and AAE are related to one another as shown in
Eqs. (2) and (3):
AAOD(λ)=[1-SSA(λ)]×AOD(λ)AAE=-dln[AAOD(λ)]/dln(λ).
The inversion algorithms mentioned above have been used for AERONET and
CARSNET, and the accuracies of the volume size distribution were 15–25 %
for 0.1 µm ≤r≤7.0 µm and 25–100 % for r<0.1 µm and r > 7 µm. The accuracies for both AOD and
AAOD are ∼ 0.01. The errors for the total, fine-mode and coarse-mode SSA are about 0.030, 0.037 and 0.085, respectively. The imaginary and
real parts of the complex refractive index for the AOD at 440 nm > 0.40 and a solar zenith angle > 50∘ have
errors of ∼ 0.0025–0.0042 and 0.04, respectively (Dubovik et
al., 2000; J. Li et al., 2015).
The direct aerosol radiative forcing (DARF) values in units of W m-2
were calculated using the radiative transfer module in the AERONET inversion
(García et al., 2008, 2012) under the assumption of cloud-free
conditions. The DARF is defined as the difference in the shortwave radiative
fluxes between the two energy levels including and excluding aerosol effects
at the Earth's surface (bottom of the atmosphere, BOA) and the top of the
atmosphere (TOA) in Eqs. (4) and (5) as follows:
DARFTOA=FTOA↑0-FTOA↑DARFBOA=FBOA↓-FBOA↓0,
where F and F0 represent the broadband fluxes with and without aerosols
at BOA and TOA, respectively. The arrows in these equations indicate the
direction of the fluxes for the downward and upward cases. Defined this way,
a negative value for ΔF indicates aerosol cooling effects,
while positive values imply warming, both at the BOA and the TOA.
In the radiative transfer module used here, the flux calculations accounted
for absorption and multiple scattering effects using the discrete ordinates
(DISORT) approach (Stamnes et al., 1988; Nakajima and Tanaka, 1988). The
solar broadband fluxes from 0.2 to 4.0 µm were simulated by using
information on aerosol properties (size distribution, spectral AOD, SSA and
phase function) obtained from the ground-based measurements. The spectral
refractive indices (both real and imaginary parts) were
interpolated/extrapolated from the values retrieved at four distinct
wavelengths (440, 670, 870, 1020 nm) from the ground-based sun photometers.
Likewise, the spectral dependence of surface reflectance was
interpolated/extrapolated from surface albedo values used in the aerosol
property retrieval process for the same wavelengths.
The integrated effects of atmospheric aerosol scattering and absorption,
gaseous absorption and molecular scattering and underlying surface
reflection effects were evaluated using the Global Atmospheric ModEl (GAME)
code (Dubuisson et al., 1996; Roger et al., 2006). In the GAME code, gaseous
absorption (mainly H2O, CO2 and O3) is calculated from the
correlated k distribution (Lacis and Oinas, 1991). The instantaneous column
water vapor content was retrieved by the absorption differential method from
the 0.94 mm channel (Smirnov et al., 2004). The total ozone content was
taken from monthly climatology values based on the Total Ozone Mapping
Spectrometer (TOMS) measurements. The GAME model accounts for spectral
gaseous absorption; that is, ozone in the ultraviolet–visible spectral range
(0.20–0.35 and 0.5–0.7 µm) and water vapor in the shortwave
infrared spectrum (0.8–3.0 µm).
The flux calculations were performed for a multi-layered atmosphere with the
US standard 1976 atmosphere model for gaseous distributions and single fixed
aerosol vertical distribution (exponential with an aerosol height of 1 km)
(García et al., 2008). As these authors have pointed out, solar fluxes
calculated using the module described above show excellent agreement with
ground-based measurements of solar radiation (slope of 0.98 ± 0.00 and
bias of -5.32 ± 1.00 W m-2) with a correlation of 99 %. There is
a small overestimation of +9 ± 12 W m-2 of the observed solar
radiation at the surface in global terms, and this corresponds to a relative
error of +2.1 ± 3.0 %. The differences range from +14 ± 10 to +6 ± 13 W m-2 for urban/industrial and biomass
burning aerosols, respectively. The errors are expected to be of the same
magnitude at the TOA, since the same methodology and inputs are used at both
levels (gaseous and aerosol distribution, radiative model, etc).
Results and discussion
Aerosol microphysical properties: particle radius and volume size
distributions
Temporal variations in the aerosol volume-size distributions at
(a) Hangzhou, (b) Xiaoshan, (c) Fuyang,
(d) LinAn, (e) Tonglu, (f) Jiande and
(g) ChunAn.
![]()
Figure 2 shows the monthly aerosol size distribution (dV/dlnr) for all seven
sites in the YRD. The annual mean values for the effective radii of the
total particles (Refft) were ∼ 0.30 µm, and the
average annual volume was ∼ 0.18 µm3µm-2 (Table 1). The fine-mode effective radii averaged ∼ 0.16 µm
in the YRD with a fractional volume of
0.10–0.11 µm3µm-2, while the coarse-mode average effective radii were ∼
2.2 µm with a fractional volume
∼ 0.08 µm3µm-2. These results show that there was a larger contribution of
fine-mode particles to the aerosol volume compared with the coarse-mode particles at
all sites. The total and fine-mode aerosol volumes and effective radii
showed small differences from the densely populated urban site (Hangzhou)
to the urban center sites in smaller cities (Xiaoshan, Fuyang, LinAn,
Tonglu, Jiande) or the rural site (ChunAn), and this reflects a generally
homogeneous distribution of the aerosol in the YRD. The coarse-mode aerosol
volumes also showed small differences among the sites, but the range of effective
radii varied 2.16–2.30 µm.
Higher volumes and larger effective radii of the fine-mode particles were observed in
June and September at most sites, and again there were relatively small
differences among sites; the exception was Xiaoshan where the month-to-month
differences in these variables were less pronounced (Fig. 2). The increases
in submicron particles during the summer may have been caused in part by the
hygroscopicity of the aerosols. In this regard, Saha and Morty (2004) reported
that the volume of accumulation-mode particles increased faster than the
coarse-mode particles under high relative humidity conditions. At our sites, the
precipitation in June and September was greater than in the other months
because of the Meiyu flood period and the typhoon and autumn rain
period, respectively. Indeed, the major aerosol components, which include
sulfate, nitrate, ammonium and organic compounds, can cause severe haze-fog
events during high relative humidity conditions. Fine particles containing
sulfate, nitrate and ammonium are hygroscopic, and their sizes are strongly
affected by the relative humidity (Fu et al., 2008; Shen et al., 2015; S. Li et
al., 2015; Huang et al., 2016). Additionally, broad fine-mode distributions
may result from the occurrence of fog or low-altitude cloud dissipation
events (Eck et al., 2012; Li et al., 2010, 2014). Eck et al. (2012) also
pointed out that a large range of fine-mode aerosol sizes may result from
cloud processing, and that also could contribute to a shoulder of larger
size particles in the accumulation mode, especially in regions where sulfate
and other water-soluble aerosols exist. Another interesting observation is
that the fine-mode aerosol volume in July was relatively low at all sites
and is coincident with lower relative humidity in July (∼ 60 %) compared with that in June (∼ 80 %) in the YRD.
Therefore, the hygroscopic effects on fine particles in July evidently are
not as obvious as in June or September.
High volumes for coarse-mode aerosol occurred in March to May at all sites,
and this suggests that more large particles occurred in spring than in the
other seasons. The most likely explanation for this is the presence of
mineral dust in the YRD region at that time of year. High PM10 mass
concentrations of Hangzhou during 2012–2015 showed 78.5 ± 15.4, 84.7 ± 11.3 and 83.6 ± 15.5 µg m-3 in March, April and
May, respectively (see the Supplement), which is consistent with the results
of relative large coarse-mode aerosol volumes in this study. As Cao et al. (2009) pointed out, fugitive dust can account for about one-third of
PM10 mass concentration in Hangzhou, and that mainly came from
resuspended road dust and construction soil. The long-range transportation
of dust in spring from northern/northwestern China could also contribute to
the high coarse-mode aerosol volume. For instance, Fu et al. (2014) and Sun
et al. (2017) found that dust particles could be transported long distances,
and the impacts were apparent in the YRD. Low volumes and large bandwidths
of coarse-mode aerosol were found in July and August at all sites, and that
may have been due to the wet removal of coarse particles by the heavy
precipitation in June. In July and August, strong convection associated with
subtropical anticyclone can disperse the aerosol. These results also are
consistent with previous reports showing minimum PM10 concentrations in
summer of the YRD region (Cao et al., 2009). It was found the PM10 mass
concentrations of Hangzhou during 2012–2015 were relatively low, with values
of 53.2 ± 6.8 and 56.7 ± 4.4 µg m-3 in July and August,
respectively (see the Supplement). Sun et al. (2013) found that aerosol size
distributions can broaden under unstable weather conditions, and in July and
August, the weather in the YRD can become unstable due to the high
temperatures, and this could be another factor that contributed to the large
bandwidth of both fine-mode and coarse-mode particles in our study. However,
these possible connections should be re-visited in the future.
Aerosol optical properties: AOD and EAE
The arithmetic mean annual values for AOD440nm at the six urban sites
(Hangzhou, Xiaoshan, Fuyang, LinAn, Tonglu, Jiande) were 0.71–0.76 and 0.68
at the rural site of ChunAn (Table 1). The difference in AOD between urban
and rural sites was < 10 %, and this indicates that there were
widespread anthropogenic impacts on the aerosol populations in the YRD and
that the high particle concentrations extend beyond the local to the
regional scale. Nonetheless, the AOD440nm generally decreased from the
east coast to inland areas towards the west (0.76 at Hangzhou, 0.73 at
LinAn, 0.71 at Jiande and 0.68 at ChunAn), and this can be explained by
stronger anthropogenic impacts in the more urbanized east. The high
AOD440nm at Hangzhou was likely the result of the greater industrial
activity and higher population density in the eastern part of that
metropolitan region; both of those factors could lead to larger aerosol
emissions compared with the less populated urban and rural sites. The
coarse-mode AOD values were just ∼ 0.06 to 0.08, and the ratio of
the fine-mode AOD440nm to the total AOD440nm varies from
0.89 to 0.91 at the sites, and therefore, fine-mode particles clearly were the
main contributors to light extinction in the region. The lower coarse-mode
fraction of total aerosol extinction (∼ 10 %) indicated that
the contribution of coarse particles to aerosol loading in the YRD region is
not as obvious as in other northern/northeastern China region (Zhang et al.,
2012).
Geographical location and annual arithmetic mean optical parameters
for aerosols from seven sites in the Yangtze River Delta.
Hangzhou
Xiaoshan
Fuyang
LinAn
Tonglu
Jiande
ChunAn
Site type
Urban
Suburban
Suburban
Suburban
Suburban
Suburban
Rural
Longitude (∘ E)
120.19
120.25
119.95
119.72
119.64
119.27
119.05
Latitude (∘ N)
30.26
30.16
30.07
30.23
29.80
29.49
29.61
Altitude (m)
41.9
14.0
17.0
139
46.1
88.9
171.4
Ndaya
485
180
217
562
498
480
439
Ninst.b
2052
752
906
2410
2255
1952
1731
AOD440nmc
0.76 ± 0.42
0.76 ± 0.43
0.76 ± 0.45
0.73 ± 0.44
0.71 ± 0.41
0.73 ± 0.40
0.68 ± 0.38
AODfine(440nm)c
0.68 ± 0.42
0.69 ± 0.41
0.69 ± 0.44
0.66 ± 0.43
0.64 ± 0.41
0.66 ± 0.40
0.61 ± 0.38
AODcoarse(440nm)c
0.08 ± 0.06
0.07 ± 0.06
0.07 ± 0.06
0.07 ± 0.07
0.07 ± 0.06
0.07 ± 0.07
0.06 ± 0.05
EAE440-870nmd
1.29 ± 0.26
1.37 ± 0.24
1.32 ± 0.24
1.29 ± 0.27
1.30 ± 0.26
1.32 ± 0.28
1.22 ± 0.25
SSA440nmc
0.91 ± 0.06
0.93 ± 0.04
0.94 ± 0.04
0.93 ± 0.05
0.92 ± 0.04
0.92 ± 0.05
0.94 ± 0.03
SSA670nme
0.92 ± 0.06
0.91 ± 0.06
0.93 ± 0.06
0.92 ± 0.05
0.93 ± 0.05
0.92 ± 0.07
0.94 ± 0.03
SSA870nmf
0.90 ± 0.07
0.90 ± 0.07
0.91 ± 0.08
0.91 ± 0.06
0.91 ± 0.06
0.90 ± 0.08
0.93 ± 0.04
SSA1020nmg
0.89 ± 0.08
0.89 ± 0.08
0.89 ± 0.09
0.90 ± 0.07
0.90 ± 0.07
0.90 ± 0.09
0.92 ± 0.05
AAOD440nmc
0.06 ± 0.05
0.05 ± 0.04
0.04 ± 0.04
0.05 ± 0.04
0.05 ± 0.04
0.06 ± 0.04
0.04 ± 0.03
AAE440-870nmd
1.13 ± 0.46
0.88 ± 0.42
0.85 ± 0.43
0.98 ± 0.35
1.11 ± 0.49
1.16 ± 0.44
0.93 ± 0.31
Refft (µm)c
0.30 ± 0.10
0.29 ± 0.09
0.30 ± 0.09
0.29 ± 0.10
0.29 ± 0.10
0.29 ± 0.09
0.30 ± 0.10
Refffine (µm)c
0.16 ± 0.04
0.16 ± 0.03
0.17 ± 0.04
0.16 ± 0.04
0.16 ± 0.04
0.17 ± 0.04
0.17 ± 0.04
Reffcoarse (µm)c
2.21 ± 0.40
2.26 ± 0.35
2.30 ± 0.39
2.24 ± 0.44
2.19 ± 0.41
2.16 ± 0.39
2.27 ± 0.42
Volume (µm3µm-2)c
0.19 ± 0.09
0.19 ± 0.09
0.19 ± 0.09
0.18 ± 0.09
0.17 ± 0.09
0.18 ± 0.09
0.17 ± 0.07
Volumefine (µm3µm-2)c
0.10 ± 0.06
0.11 ± 0.06
0.11 ± 0.07
0.10 ± 0.06
0.10 ± 0.06
0.10 ± 0.06
0.10 ± 0.06
Volumecoarse (µm3µm-2)c
0.09 ± 0.06
0.08 ± 0.05
0.08 ± 0.06
0.08 ± 0.05
0.08 ± 0.06
0.08 ± 0.07
0.07 ± 0.05
DARF-BOA (W m-2)c
-93±44
-84±41
-80±40
-81±39
-79±39
-82±40
-74±34
DARF-TOA (W m2)c
-35±20
-36±21
-37±21
-36±21
-35±20
-35±21
-40±19
a Number of available observation days.
b Number of instantaneous observations.
c Optical parameters at a wavelength of 440 nm.
d Angström exponents between 440 and 870 nm.
e Optical parameters at a wavelength of 670 nm.
f Optical parameters at a wavelength of 870 nm.
g Optical parameters at a wavelength of 1020 nm.
Unlike one peak AOD440nm distribution found in June–August over
northern and northeastern China (H. Che et al., 2015), the monthly averaged
AODs at 440 nm at all seven sites showed two peaks, one in June and the
other in September (Fig. 3), with mean values of ∼ 1.26 ± 0.50 (FMF ∼ 0.93) and ∼ 1.03 ± 0.57 (FMF ∼ 0.95),
respectively. The FMF is defined as the fine-mode
particle AOD440nm fraction (AODfine(440nm)/ AOD440nm). Low
AOD440nm values of ∼ 0.40–0.60 were found in July
and August throughout the region, and these low values are consistent with
the discussion above concerning aerosol microphysical characteristics. That
is, hygroscopic effects and/or cloud processing of fine-mode aerosol
particles caused greater extinction in June and September compared with July
and August. Zhang et al. (2015) reported that the hygroscopic particles under
high relative humidity conditions could cause strong aerosol light
scattering in the Yangtze River Delta. Other meteorological factors also may
have played a role in the variations in AOD440nm during summer
because in July and August, subtropical high-pressure systems prevail, and
the planetary boundary layer (PBLH) at Hangzhou is deep, ∼ 1.5–2.0 km (Sun et al., 2017). The large PBLH associated with subtropical
anticyclones favors aerosol dispersion, and this can help explain the
relatively low aerosol extinction observed in July and August. In January
and February, high AOD440nm values (0.71–0.89) were observed at all
sites, and this can be attributed to emissions from residential heating and
the stability of the atmosphere, which can cause the near-surface
accumulation of aerosol particles. The high AOD440nm values in
winter also are consistent with studies by Cao et al. (2009), who reported
that the PM10 mass concentrations at Hangzhou were highest in winter.
According to the PM10 mass concentration measurements during 2012–2015
at Hangzhou, it was found that the PM10 mass concentrations were about
82.2 ± 9.5, 59.1 ± 6.0, 82.8 ± 13.9 and 92.3 ± 14.3 µg m-3
in spring, summer, autumn and winter, respectively (see the
Supplement).
The extinction Angström exponent (EAE =-dln[EAOD(λ)]/ dln(λ)) can be regarded as an indicator of aerosol size; that
is, EAE440-870nm>1.00 typically indicates that
the aerosol particles are small. The mean extinction Angström exponent
at all seven CARSNET sites was higher than 1.20 throughout the year (Table 1), which means that small particles were predominant. This finding is
consistent with the reported dominance of small particles from anthropogenic
emissions and agricultural activity in the region (Tan et al., 2009).
Variations in the total, fine-mode and coarse-mode
AOD440nm at (a) Hangzhou, (b) Xiaoshan,
(c) Fuyang, (d) LinAn, (e) Tonglu,
(f) Jiande and (g) ChunAn. The boxes represent the 25th to
75th percentiles of the distributions while the dots and solid lines within
each box represent the means and medians, respectively.
![]()
Variations in the single-scattering albedo at 440, 670, 870 and
1020 nm at (a) Hangzhou, (b) Xiaoshan,
(c) Fuyang, (d) LinAn, (e) Tonglu,
(f) Jiande and (g) ChunAn. See Fig. 3 for an explanation of
the symbols.
![]()
Aerosol optical properties: single-scattering albedo
The SSAs at 440 nm at our seven sites in the YRD region varied from 0.91 to 0.94
(Table 1), and box plots of the monthly SSAs at wavelengths of 440, 670, 870
and 1020 nm are shown in Fig. 4. Eck et al. (2005) reported that the SSAs
at 440 nm from AERONET retrievals were confined to a relatively narrow range
of values globally from ∼ 0.82 to 0.98. Therefore, the SSA
values in this study may be explained by moderately to strongly absorbing
aerosols from industrial emissions and other anthropogenic sources. The
SSA440nm at Hangzhou site was 0.91 ± 0.06, which is lower than
that at the rural ChunAn site (0.94 ± 0.03). As Dubovik et al. (2000,
2002, 2006) reported, the SSA depends on two factors – particle size and
composition. From Table 1, one can see that the differences in effective
radii for total, fine-mode and coarse-mode particles between Hangzhou and ChunAn
are quite small, and therefore, the differences in SSAs between the two
sites can best be explained by differences in composition. Furthermore, the
differences in SSAs between these sites indicate that there was a higher
percentage of absorbing aerosols at urban sites than the rural one.
The SSAs for seven sites showed significant month-to-month variations. The
increased scattering (light-absorbing) effects seen in June can be
attributed to hygroscopic growth, which can modify aerosol properties
greatly (Xia et al., 2007). The presence of light-absorbing dust aerosols in
spring and of absorbing aerosols from biomass burning in August was probably
responsible for the differences in SSA values observed between those months
because of the distinct differences in the intensive optical properties of
dust and biomass burning products (Yang et al., 2009). At Hangzhou, the
monthly average SSA values at 440 nm were relatively high in February
(∼ 0.94 ± 0.05) and June (∼ 0.92 ± 0.06) and more moderate in March (∼ 0.90 ± 0.06) and
August (∼ 0.89 ± 0.09). In comparison, the differences
in monthly SSA values at the rural ChunAn site were smaller, only varying
from 0.92 to 0.95. We conclude from the temporal patterns of the SSAs that
the types of aerosols at the urban/suburban sites were more variable than at
the rural sites.
The SSA wavelength dependence is a function of the specific
absorption/scattering properties of different aerosol types (Sokolik and
Toon, 1999; Eck et al., 2010). The SSA for mineral dust particles typically
shows a strong wavelength dependence from 440 to 1020 nm, with a low value at
440 nm due to iron oxide absorption (Cheng et al., 2006; Dubovik et al.,
2002). In spring, especially in March, the SSA was obviously lower at
shorter wavelengths than at the longer ones, and this implies absorption by
dust particles. This conclusion is consistent with the discussion above
concerning the impact of dust on aerosol size distributions. In addition,
there was a significant decrease in SSA at shorter wavelengths in July and
August at most sites, and this supports the presence of aerosol particles
with strong absorption, especially at infrared wavelengths. The decreases in
those months can be explained by strongly absorbing aerosols from biomass
burning or possibly industrial emissions. As Ding et al. (2013a, b) and Wang
and Zhang (2008) have reported, plumes from agricultural burning typically
contain light-absorbing carbonaceous aerosols, and these pollutants can
seriously impair air quality. Indeed, aerosols from biomass burning were
more than likely responsible for the low in SSAs found in our study during
July and August.
Aerosol optical properties: AAOD and AAE
(a) Annual average absorption aerosol optical depths at
440 nm (AAOD440nm) at the CARSNET sites and month-to-month
variations in AAOD440nm at (b) Hangzhou,
(c) Xiaoshan, (d) Fuyang, (e) LinAn,
(f) Tonglu, (g) Jiande and (h) ChunAn. See Fig. 3
for an explanation of the symbols.
![]()
The AAODs at 440 nm at the seven sites were similar ∼ 0.04–0.06 (Fig. 5a and Table 1), and therefore, absorbing aerosols were
apparently widely distributed throughout the YRD. There were large
uncertainties in the AAOD440nm, however; in fact, the standard
deviations (0.03 to 0.05) were comparable to the means, and this reflects
the large temporal variability in absorbing aerosol particle loadings. The
average AAOD440nm at Hangzhou was about 0.02 higher than that at
ChunAn, and this shows that the relative proportion of absorbing particles
at the urban area was larger than at the rural site, presumably due to
greater anthropogenic emissions. From Fig. 5, one can see that the monthly
AAOD440nm at the urban sites from March to November usually exceeded
0.05, which implies there were more absorbing species in spring and autumn
compared with winter (December to February) when the AAOD440nm
tended to be lower, < 0.05. This result suggests that the relative
abundances of light-absorbing particles were lower in winter compared with
other seasons, and this is different from many regions in northern China
where the AAODs are highest in winter due to the emission of absorbing
particles from residential heating and other sources (Zhao et al., 2015). At
the ChunAn site, the variations in AAODs were smoother than the more heavily
impacted urban sites, and the monthly mean AAODs at ChunAn were < 0.05 throughout the year.
The AAE can be viewed as an indicator of the type of dominant absorbing
aerosol particles, which include black carbon, organic matter and mineral
dust (Giles et al., 2012). Generally, an AAE < 1 indicates mixing,
coating and coagulation of black carbon with organic and inorganic
materials; an AAE close to 1 indicates absorbing black carbon aerosols from
the fossil fuel burning; and an AAE > 1.10 indicates absorbing
aerosols mainly from biomass burning or mineral dust (Russell et al.,
2010; Bergstrom et al., 2007; Lack and Cappa, 2010). The annual mean AAEs at
Hangzhou, Xiaoshan, Fuyang, LinAn, Tonglu, Jiande and ChunAn were 1.13 ± 0.46, 0.88 ± 0.42, 0.85 ± 0.43, 0.98 ± 0.35,
1.11 ± 0.49, 1.16 ± 0.44 and 0.93 ± 0.31, respectively (Table 1), and for discussion purposes, the seven sites were grouped into three
categories based on their average AAEs. The mean AAE values at Xiaoshan and
Fuyang were < 1.00, which suggests that coated black carbon
particles dominated at these two sites. However, this also could be due to
measurement uncertainties due to smaller numbers of samples from those sites
or to slightly larger values of the imaginary part of the refractive index
at longer wavelengths for certain particles (Bergstrom et al., 2007). There
would need to be more observations to confirm the low AAEs at these two sites.
The AAE values ∼ 1.00 at LinAn and ChunAn indicate that the
absorbing aerosol population was mainly composed of black carbon from fossil
fuel burning. LinAn has a more developed economy compared with Tonglu and
Jiande, and in comparison, LinAn has many more motor vehicles and is more
heavily impacted by industrial emissions and fossil fuel combustion. In
contrast, the ChunAn site is located in the Qiandao Lake National Water
Resources Protection Zone where biomass burning and industrial activities
are banned. Thus, emissions from motor vehicles are probably the main source
of absorbing carbon aerosols at ChunAn. Finally, the AAE values at Hangzhou,
Tonglu and Jiande were > 1.10, indicating a predominance of
light-absorbing aerosols from either biomass burning or mineral dust.
Hangzhou has a population of ∼ 10 million and more than one
million vehicles, and the city can be impacted by fugitive dust and biomass
burning emissions (Cao et al., 2009). Tonglu and Jiande have small economies
compared with Hangzhou, but there is more agricultural production near these
two sites, and they can be impacted by biomass burning. These inter-site
differences in the AAEs reflect a degree of spatial heterogeneity in the
distributions of absorbing aerosols even though the AODs were relatively
similar, and these differences likely result from the many types of emission
sources that can impact the sites.
Direct aerosol radiative forcing at the Earth's surface and TOA
In this study, only clear-sky direct aerosol radiative forcings could
be investigated because the aerosol microphysical and optical parameters were
derived from ground-based retrievals under cloud-free conditions. For the
direct aerosol radiative forcing calculations, the solar fluxes are only
evaluated for solar zenith angles (SZA) between 50 and
80∘, which is where the solar geometry conditions are the
most appropriate for retrieving the aerosol properties (Dubovik et al.,
2000, 2002).
The annual direct DARF-BOA values under clear conditions for Hangzhou,
Xiaoshan, Fuyang, LinAn, Tonglu, Jiande and ChunAn were -93 ± 44,
-84 ± 40, -80 ± 40, -81 ± 39, -79 ± 39,
-82 ± 40 and -74 ± 34 W m-2, respectively (Fig. 6a).
The DARF-BOA at Hangzhou was ∼ -20 W m-2 lower than that
at the rural ChunAn site. The DARFs at the Earth's surface and TOA are
governed by the aerosol microphysical and optical properties, primarily
particle size distributions, AODs and SSAs. The large number of negative DARF-BOA
values at Hangzhou can be attributed to the high aerosol extinction
(AOD440nm∼ 0.76) and small absorption (SSA at 440 nm
∼ 0.91). These calculations indicate that effects of those
particles on radiative fluxes can cause significant surface cooling at that
urban site. In comparison, at the rural ChunAn site, less negative DARF-BOA
values result from lower aerosol extinction (AOD440nm ∼ 0.68) and higher scattering (SSA440nm ∼ 0.94).
(a) Annual variations of monthly mean direct aerosol
radiative forcing (50–80∘ SZA) at the bottom of the atmosphere over
(b) Hangzhou, (c) Xiaoshan, (d) Fuyang,
(e) LinAn, (f) Tonglu, (g) Jiande and
(h) ChunAn. See Fig. 3 for an explanation of the symbols.
![]()
The monthly DARF-BOA at the seven sites was most strongly negative in June,
followed by March and September. The strong cooling effect at the surface in
June was due to the high aerosol extinction (in Sect. 3.2) and in
particular the high volumes of fine-mode particles as shown in Fig. 2. Weakly
absorbing particles with SSA440nm ∼ 0.90–0.95 (Fig. 6)
also would reduce the flux of solar radiation to the surface. A large
surface cooling effect also was found in March and April. Although the AOD
in spring was lower than in winter, there were more strongly absorbing
aerosols with relatively smaller SSAs at shorter wavelengths then, and this
can be attributed to the presence of mineral aerosol particles. These coarse-mode particles with high volumes and large radii are common in spring (see
Sect. 3.1), and they can absorb and scatter solar radiation, and in
doing so cool the surface. Surface cooling in September can be explained by
high aerosol extinction that resulted from the high volumes of weakly
absorbing fine-mode particles (SSA440nm ∼ 0.90–0.95).
These results indicate that the attenuation of the solar radiation by the
aerosols leads to significant surface cooling over the YRD.
The DARF-TOA annual mean values under clear conditions were ∼-40 W m-2 at all sites (Fig. 7a), and these negative DARF-TOA
values indicate that the aerosols caused cooling of the whole earth–atmosphere
system in the YRD. This is different from the case in northern/northeastern
China where the instantaneous DARF-TOA value can be positive in winter due
to the high surface reflectance of short wavelength radiation combined with
atmospheric heating caused by absorbing aerosols (Zhao et al., 2015; Che et
al., 2014).
(a) Annual variations in the monthly mean direct aerosol
radiative forcing (50–80∘ SZA) at the top of the atmosphere (TOA) at
(b) Hangzhou, (c) Xiaoshan, (d) Fuyang,
(e) LinAn, (f) Tonglu, (g) Jiande and
(h) ChunAn. See Fig. 3 for an explanation of the symbols.
![]()
The monthly DARF-TOA means under clear conditions varied smoothly during two
periods: (1) from January to May and (2) from October to December (Fig. 7). The
DARF-TOA values were found to be approximately -40 W m-2 at all sites in the above
two periods. However, the monthly DARF-TOA means in June and September
exceeded -40 W m-2 at most sites, which indicated more cooling effects
on the whole earth–atmosphere system due to large aerosol extinctions in
the YRD. The DARF-TOA means were about -20 W m-2 at the seven sites in
July/August, which suggests weak cooling at that time. In contrast, the
DARF-TOA values under clear conditions at Shenyang (urban area of
northeastern China), Beijing (urban area of northern China) and Xianghe
(rural area of northern China) showed a negative peak during June to
August due to the large aerosol extinctions in the summer season (Zhao et al.,
2015; Xia et al., 2016). As noted in Sect. 3.3, the SSA was low in July
and August, and that was attributed to absorbing particles from biomass
burning. Ding et al. (2016) found that large quantities of black carbon can
be emitted from biomass burning in the YRD during the summer, and because
these particles strongly absorb at infrared wavelengths, they can cause a
heating up of the atmosphere, resulting in lower negative DARF-TOA. Indeed,
the positive DARF-TOA values we found under clear conditions from April to October
were mainly due to the effects of strongly absorbing particles; that is, the
SSAs at 440 nm were < 0.80, and they showed a strong decrease with
wavelength (not shown here) when DARF-TOA values were positive. Moreover,
strongly absorbing aerosol particles can heat the atmosphere column and the
TOA at the same time.
Aerosol type classification based on the aerosol optical
properties
In previous studies, aerosol types have sometimes been simply classified as
dust (high AOD, low AE (Angström exponent)) and anthropogenic aerosols (high AOD, high AE) (H. Z. Che
et al., 2009; H. Che et al., 2009; Wang et al., 2010). However, aerosol direct radiative
forcing is affected by the absorptivity of aerosol and the underlying
surface conditions (Haywood and Shine, 1995), and there are advantages to the
better characterization of the aerosol populations. For example, the
classification of ground-based aerosol types can be used to compare and
validate aerosol types in chemical transport models and satellites
retrievals (J. Lee et al., 2010). Thus, it is advantageous to categorize
aerosols as absorbing or non-absorbing based on ground-based optical
parameters, including SSA, fine-mode fraction of AOD, and EAE etc. In this
study, we used the SSA, FMF and EAE values to classify the fine-mode and
coarse-mode particles from each site into eight groups of particles following the
method of Zheng et al. (2017). The eight types of particles were (I)
highly absorbing fine-mode particles (AE > 1.2, SSA ≤ 0.85);
(II) moderately absorbing fine-mode particles (AE > 1.2, 0.85 ≤ SSA < 0.9); (III) slightly absorbing fine-mode particles (AE > 1.2, 0.9 ≤ SSA < 0.95); (IV) weakly absorbing
fine-mode particles (AE > 1.2, SSA > 0.95); (V)
mixed-absorbing particles (0.6 ≤ AE < 1.2, SSA ≤ 0.95);
(VI) mixed slightly absorbing particles (0.6 ≤ AE < 1.2, SSA > 0.95); (VII) strongly absorbing coarse-mode particles – mainly
dust (AE ≤ 0.6, SSA ≤ 0.95); and (VIII) weakly absorbing coarse-mode
particles (AE ≤ 0.6, SSA > 0.95).
The aerosol type classification using SSA as a function of
AE440-870nm over (a) Hangzhou, (b) Xiaoshan,
(c) Fuyang, (d) LinAn, (e) Tonglu,
(f) Jiande and (g) ChunAn. See text for description of
groups I–VII.
![]()
From Figs. 8 and 9, one can see that the absorbing fine-mode particles
(Type I, II and III) accounted for ∼ 30 to 50 % of the
aerosol in the YRD region with the FMF ∼ 0.86–0.94 %. The
percentage of highly absorbing fine-particles (Type I) was obviously
larger at Hangzhou (∼ 8 %) (FMF ∼ 0.90) than
the smaller city sites (∼ 2–3 %), and at the ChunAn rural
site, the percentage of Type I particles was only ∼ 0.16 %
(FMF ∼ 0.91). This indicates that there were greater emissions
of strong absorbing aerosols from sources such as biomass burning and/or
urban/industrial activities at the urban site compared with the rural one.
The proportion of weakly absorbing fine-mode particles (Type IV) varied from
∼ 20 to 30 % for all sites, and the FMF varies from
0.89 to ∼ 0.95 at the same time, which suggests that the second largest
aerosol type in the area is weakly absorbing fine-mode particles. The
percentage of mixed absorbing particles (Type V) was about ∼ 25–26 % both at Hangzhou and the rural ChunAn site, which is slightly
higher than that at small city sites (∼ 10–20 %) where the
FMF of these particles was ∼ 0.81–0.89. The higher FMF at
Hangzhou was probably due to the complex aerosol emission sources that
impact this megacity, while at ChunAn, which is surrounded by mountains, the
basin topography promotes particle mixing. The mixed slightly absorbing
particles (Type VI) showed the highest percentage of total aerosols at Fuyang (18.05 %) and ChunAn (15.32 %), and the FMF of this
group varied from 0.84 to 0.91 at all sites. The contribution of Type VI mixed
slightly absorbing particles at Hangzhou was ∼ 11.49 %, and
the FMF (0.86) there was not as high as at ChunAn (0.91). The proportion of
strongly absorbing coarse-mode particles – mainly dust (Group VII) – was only
∼ 1.04 % of the total particle count at Hangzhou, while at
the other sites the percent abundances were < 1 %, and the FMF for
these particles was ∼ 0.50–0.63 at all sites. These patterns
show that the YRD region is different from regions in northern China, including
Beijing, where dust particles contribute significantly to the coarse-mode
absorption (Zheng et al., 2017). The percentage of weakly absorbing
coarse-mode particles (Group VIII) at all sites was < 0.54 %,
which shows that this aerosol type was rare. In addition, the FMF of Group
VIII particles was 0.5–0.8, with large uncertainties at all sites. Overall,
this analysis of aerosol types shows that the aerosol absorption is
relatively weak in the YRD region, and the fine mode makes up an especially
large contribution of the absorbing particles at Hangzhou.
Discussion
The distribution of the aerosol type classification over
(a) Hangzhou, (b) Xiaoshan, (c) Fuyang,
(d) LinAn, (e) Tonglu, (f) Jiande and
(g) ChunAn. See text for description of types I–VII.
![]()
Compared with previous studies on the climatology of aerosol microphysical
and optical properties in China, large volumes and effective radii of
fine-mode aerosol, as well as high AODs at 440 nm, were found at most sites
in June and September but not in July or August. This is remarkably
different from studies conducted in northeastern/northern China, where these
properties showed the maximum values in July or August (Eck et al., 2005;
Zhao et al., 2013; H. Che et al., 2015). This reflects the distinctive
climatology of aerosol microphysical and optical properties in the YRD
region of eastern China. This information on the distributions and
variations of aerosol microphysical and optical properties obtained in our
study should be taken into account in the validation of satellite retrievals
and aerosol modeling studies in the future.
The AOD440nm values in our study (0.71–0.76) are similar to other
urban areas of China, such as Shenyang (0.75), Beijing (0.76), Tianjin
(0.74), Shanghai (0.70) and Hefei (0.69) (Pan et al., 2010; He et al.,
2012; Zhao et al., 2013; H. Che et al., 2015; Liu et al., 2017). This
indicates that high aerosol loadings caused by anthropogenic activities
occur over many urban, suburban and even rural areas of eastern China. The
AOD440nm at the rural ChunAn site (∼ 0.68) is
∼ 2–6 times higher than in other rural sites in China, such
as Longfengshan (0.35; northeastern China), Xinglong (0.28, northern China),
Akedala (0.20, northwestern China) and Shangri-La (0.11, southwestern China)
(Wang et al., 2010; Che et al., 2011; Zhu et al., 2014; H. Che et al., 2015).
Therefore, strong aerosol effects on light extinction occur not only in
urban areas but also in much or possibly all of the YRD. The SSA440nm in our study ranged from 0.91 to 0.94, which is similar to other
regions of China, such as Wuhan (0.92), Beijing (0.89) and Xinglong (0.92)
(L. C. Wang et al., 2015; Xin et al., 2014; Zhu et al., 2014); this suggests
aerosol particles in the YRD are slight to relatively strong absorbers of
violet/indigo wavelengths.
Although the AAE has been used as an indicator of the dominant absorbing
aerosol type, the SSA, FMF and EAE were also used to classify the absorption
characteristics of fine-mode and coarse-mode particles. It should also be
mentioned that uncertainties in the AAE calculations due to uncertainties in
the SSAs may have contributed to the observed differences between sites
(Bergstrom et al., 2007; Giles et al., 2011). Ideally, the aerosol physical
and chemical characteristic measurements will be combined at some point to
more definitively classify the aerosol types. In addition, even though
Zhuang et al. (2017) pointed that the DARFs are not very sensitive to
vertical profiles under clear-sky conditions, future research should take
into account the vertical distributions of aerosols to more accurately
assess the direct aerosol radiative forcing effects.
The useful observations at most sites in this study were made on
439–562 days, but smaller numbers of observation were made in
Xiaoshan and Fuyang because of instrument failures, on 180 and 217 days,
respectively. There were fewer than 15 days of useful data for October and
November at Fuyang and for June, July, September and November at Xiaoshan.
While the available data do provide insights into the aerosol
characteristics at the sites, more extended observations should be conducted
at the two sites in future.
Conclusions
A detailed study of aerosol microphysical and optical properties retrieved
from synchronous ground-based sun photometer observations was conducted at
seven sites in the Yangtze River Delta region of eastern China from 2011 to
2015. The aerosols were classified into eight types, and calculations were
made for the direct aerosol radiative forcing (DARF) at the top and bottom
of the atmosphere. The conclusions of the study can be summarized as
follows.
A relatively homogeneous distribution of aerosol microphysical properties
was found for a megacity, five small cities and a rural site in the YRD
region. High particle volumes of coarse-mode aerosol occurred in March to
May, which reflects the existence of large mineral particles from springtime
dust storms. High volumes and large effective radii of fine-mode aerosol in
June and September were found at most sites, and this was attributed to
aerosol hygroscopicity and cloud processing. The low volumes and large
bandwidth of both fine-mode and coarse-mode aerosol found in July and August at
all sites was explained by the wet removal of coarse particles during the
heavy precipitation in June and the influences of subtropical anticyclones
in summer.
The AOD440nm generally decreased from the east coast
(0.76 – Hangzhou) to the inland areas towards the west (0.68 – ChunAn), and
this can be explained by anthropogenic impacts of the more urbanized regions
in the east YRD region. The AOD440nm values at urban and rural sites
of YRD were 0.68–0.76, and the fine-mode fraction was ∼ 0.90, which indicates that fine-mode particles were more important than the
coarse mode for the light extinction. The difference in AODs between the
urban and rural sites was less than 10 %, and this can be explained by
somewhat stronger effects of anthropogenic activities in the urban area. The
monthly averaged AODs at 440 nm showed peaks in June and September that
resulted from increases in fine-mode aerosol particles. However, AODs at 440 nm
in July and August were the lowest over the year, and this was related to
conditions favorable for aerosol dispersion. The mean extinction
Angström exponent was > 1.20 all year, indicating that small
particles were predominant in the region. The SSAs at 440 nm varied from
0.91 to 0.94 at the urban and rural sites, indicating that the aerosol
particles were moderately absorbing, and this is almost surely a result of
impacts from the high industrial emissions and other anthropogenic
activities in the region. There was an obvious wavelength dependence for the
SSA in July and August, and aerosols absorbed most strongly at infrared
wavelengths. The AAODs at 440 nm at the seven sites were ∼ 0.04–0.06, which suggests that absorbing aerosols are distributed
more or less homogeneously throughout the YRD region. The averaged
AAOD440nm at Hangzhou is about 0.02 higher than that at ChunAn,
which indicates that the relative proportion of absorbing particles in the urban
area is larger compared with the rural area. The AAOD440nm in winter
was < 0.05, which suggests that light absorption by the particles
was low compared with the other seasons. The geographical variability in the
distributions of the AAEs suggests that the absorbing aerosols possibly have
different optical characteristics related to the local emission sources in
the YRD.
The DARF-BOA at Hangzhou was ∼-20 W m-2 lower than that
at the rural ChunAn site, which shows stronger aerosol cooling at the
megacity. The monthly DARF-BOA was strongly negative in June due to the high
aerosol extinction and especially the high fine-mode volume. The DARF-TOAs
under clear conditions were <-40 W m-2 but were about -20 W m-2 in July/August, which suggests weaker cooling in midsummer. The
DARF-TOAs were positive from April to October when the SSA440nm<0.80, and the greater effects at shorter wavelengths were likely
due to emissions of carbonaceous particles from the burning of crop
residues.
The SSA, FMF, and EAE values were used to classify the particles as
absorbing or non-absorbing. Relatively large emissions of strongly absorbing
aerosols in the Hangzhou urban area was the result of biomass burning and/or
urban/industrial activities. The aerosol type classification showed overall
that the aerosol absorption is weak to moderate in the YRD, and the
fine mode has a large contribution to the higher percentage of absorbing
particles at the Hangzhou site.