atmosphere
Article
Particulate Characteristics during a Haze
Episode Based on Two Ceilometers with
Different Wavelengths
Jing Yuan 1,2 , Lingbing Bu 1, *, Xingyou Huang 1 , Haiyang Gao 1 and Rina Sa 1
1
2
*
Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters,
Key Laboratory for Aerosol-Cloud-Precipitation of China Meteorological Administration,
Key Laboratory of Meteorological Disaster of Ministry of Education, Nanjing University of
Information Science and Technology, Nanjing 210044, China; [email protected] (J.Y.);
[email protected] (X.H.); [email protected] (H.G.); [email protected] (R.S.)
Meteorological Bureau of Wu Jiang County in Jiangsu Province, Suzhou 215200, China
Correspondence: [email protected]; Tel.: +86-25-5869-9863
Academic Editor: Robert W. Talbot
Received: 10 December 2015; Accepted: 19 January 2016; Published: 29 January 2016
Abstract: To investigate the particulate characteristics of a haze episode, data from two ceilometers
with wavelengths of 532 nm and 910 nm, respectively, were studied intensively. By combining the
data from the ceilometers with data from a sounding balloon, an automatic meteorological station,
and a Grimm 180 PM instrument, analyses of the haze process of a short haze event were performed.
The results showed that the relatively calm weather conditions were favorable to the occurrence of
the haze and that higher relative humidity had a great influence on visibility. The extinction profiles
from the ceilometers reflected the existence of an inverted structure of the temperature profiles
and demonstrated the extinction differences at two different wavelengths. Because extinction has
a positive correlation with relative humidity, the effect of hygroscopic growth was analyzed at the
two different wavelengths. As hygroscopic growth of the particles proceeded, the longer wavelength
became more sensitive to the large particles, and vice versa. The hygroscopic growth factor and the
Angstrom exponent showed a negative correlation, and the correlation coefficients at 532 nm and
910 nm were 0.54 and 0.86, respectively. The accumulation mode particles were more stable through
time than the coarse mode particles, and the variation of the coarse mode particles coincided well
with the variation of the Angstrom exponent from the two ceilometers.
Keywords: ceilometer; extinction; haze; hygroscopic growth
1. Introduction
Haze is an important environmental issue worldwide and has received increasing attention due
to its impacts on visibility, air quality, radiative forcing, regional climate, and human health. Aerosols
affect Earth’s radiation balance directly by scattering and absorbing solar shortwave and longwave
radiation. They also act as condensation nuclei for clouds, thereby changing the clouds’ microphysical
properties and life cycles, and thus affect the climate system indirectly [1,2]. Additionally, the World
Health Organization has indicated that aerosols (especially fine particles) can produce adverse effects
on human health [2]. The China Meteorological Administration (CMA) defines haze as a condition
in which atmospheric visibility is less than 10 km and relative humidity is less than 90% [1]. China’s
rapid economic growth, urbanization, population expansion, and increased industrial activity have
led to increasing amounts of particles being discharged into the environment, resulting in frequent air
pollution events [3].
Atmosphere 2016, 7, 20; doi:10.3390/atmos7020020
www.mdpi.com/journal/atmosphere
Atmosphere 2016, 7, 20
2 of 11
Because of the advantages of short wavelength, light can interact better with small particulates
than can microwaves, and optical instruments, such as Lidar and optical particulate counters, are often
used to observe the properties of haze. Huang et al. statistically analyzed the monthly and quarterly
variations in aerosol extinction coefficients over Shanghai using Micropulse Lidar (MPL) data from
July 2008 to January 2009 [4]. The results showed that most of the aerosol was concentrated in the
layer below 3 km and that the optical extinction coefficient in the layer below 2 km contributed 84.25%
of that below 6 km [4]. Haze events in Shanghai were analyzed by Pan et al. Their analysis showed
that severe haze events in Shanghai often occurred under conditions of calm wind and weak radiation.
The humidity had an important impact on visibility during the severe haze events, and the altitude
variation of the Planetary Boundary Layer (PBL) determined the intensity of the haze [5]. Additionally,
Zhang used spaceborne and ground-based Lidar observations over the Jinhua Basin in the province
of Zhejiang to analyze the aerosol vertical distribution during a haze event both temporally and
spatially [6].
The laser ceilometer is a simple elastic scattering lidar system in which a small power laser is
used. Compared with common lidars, the ceilometer often has a low signal-to-noise ratio (SNR)
that is about 1200 in clear air. However, for the increased atmospheric particulates during a haze
episode, the ceilometer can obtain a sufficient signal-to-noise ratio to invert the optical properties
of the particulates. The SNR at an altitude of 500 m during the hazy episode in this article is about
4000. The most important advantage of a ceilometer, as a simple lidar system, is that it can function
in harsh circumstances and poor weather conditions under which other types of lidar are often
invalid. Furthermore, according to its schedule, CMA will soon construct a ceilometer network for
cloud observation. If information about haze could be retrieved from the ceilometer network in
addition to that of cloud, it would benefit the monitoring of haze around China and permit full use
of the ceilometer network data. In previous research, we demonstrated the feasibility of aerosol
detection during fog-haze processes using a ceilometer [7]. In this study, based on data from two
ceilometers with two different wavelengths, more microphysical particulate characteristics of a haze
were studied intensively.
2. Facilities and Method of Analysis
In principle, a ceilometer is a simple lidar system. For the design purpose of detection of cloud
base height, which often has a large backscatter coefficient, the ceilometer often has a more simple
structure than common lidar. As the laser pulse is emitted by the lidar system flying in the atmosphere,
the laser pulse is scattered by aerosol, molecular components, or cloud. The partial scattered signal
is sensed by the high gain detector (usually a PMT) and digitalized by a photon counter. Optical
properties such as the backscatter coefficient and the extinction coefficient can be retrieved from the
intensity of the backscattered signal; thus, cloud distribution information can be obtained.
In this study, two ceilometers were used. One was the MPL-4B, with a laser wavelength of 532 nm.
Other detail specifications of this system can be found in various sources and are not listed here [8]. The
capability of the MPL-4B has been verified through aerosol and cloud projects, such as the Micro-Pulse
Lidar Network (MPLNET) and Atmospheric Radiation Measurement (ARM) projects [9]. The other
ceilometer used was the Vaisala CL31, which uses a 910 nm diode with a pulse frequency of 6.5 kHz
as an optical source. The lens is divided into transmitting and receiving areas. The middle of the
lens focuses the outgoing laser beam, and the outer part of the lens collimates the backscattered light
onto the receiver. Because the beam is transmitted coaxially with the receiver field of view, the system
provides robustness against changes in mechanical alignment. The CL31 has a shield with a blower
and a heater, which enable reliable operation in precipitation and in extreme temperatures. The CL31
provides an aerosol signal profile as well as the height of the cloud base. Although in most cases the
backscatter signal of aerosol is too weak to retrieve extinction, the optical properties of haze can be
obtain by using the Klett method [7]. And the overlap correction before the data inversion had carried
Atmosphere 2016, 7, 20
3 of 11
on. So, above the first bin (blind range: the lowest 30 m), the ceilometers’ data can be used to analyze
the properties of aerosol. The main parameters of the two ceilometers are shown in Table 1.
Table 1. Parameters of the CL31 and MPL-4B.
Laser
Measurement range
Reporting resolution
Measurement interval
Pulse frequency
Input power
Power consumption
Dimensions
Weight
CL31
MPL-4B
910 nm
7.7 km
5/10 m
2 . . . 120 s
6.5 kHz
100/115/230 V
310 W
Total: 1190 ˆ 335 ˆ 324 mm
Measurement unit: 620 ˆ 235 ˆ 200 mm
Total: 32 kg Measurement unit: 13 kg
532 nm
10 km
15/30/75 m
1 s . . . 900 s
2.5 kHz
100/240 V
350 W
380 ˆ 305 ˆ 480 mm
16 kg
The Grimm 180 is a field measurement instrument based on the principle of light scattering.
The pump inputs an ambient air sample including particles into a gas chamber at a constant rate
of flow. The laser irradiation is scattered by the particulates in the air flow and then sensed by the
detectors. The received pulse signal frequency and intensity are then related to the number and
diameter of the particles. The Grimm 180 is an online environmental particulate matter monitor that
can simultaneously measure PM10 , PM2.5 , and PM1 in the environment. The particle size range of
measurement is 0.25 um to 32 um (31 channels), and the measurement range of mass concentration is
0.1 to 6000 µg/m3 [10].
To select a tool for routine observation of cloud base height during meteorological operations,
campaign observation of cloud using several kinds instruments, including ceilometers, was organized
by the CMA at the Nanjiao Surface Meteorological Station from June to October 2014. During the
observation, the CL31 was a candidate tool for cloud observation. It showed notable capability and
stability during the entire observation. The MPL-4B was installed in an air-conditioned room with
a window in the ceiling to emit and collect backscattered light. The capability of the MPL has been
demonstrated in other projects. Here, it functioned as the standard tool to evaluate the candidate tools.
The distance between the MPL-4B and the CL31 was set as 3 m to ensure that the two instruments
would detect the same cloud or aerosol in the same atmospheric column. Routine observation results
of meteorological elements from the Nanjiao surface station were also collected to analyze the haze
process. The Grimm system was mounted at Chaoyang Meteorological Station, another surface station
that is about 10 miles from the Nanjiao surface station. For the scale of the haze phenomenon, the
work assumed that all of the instruments detected the same sample in the haze. In the haze situation,
the Klett method was used to retrieve the extinction coefficient profiles of the pollutants according to
the following formula [11]:
σpzq “
exprpS ´ Sm q{ks
şz
2
´1
tσm
` z m exprpS ´ Sm q{ksdz1 u
k
(1)
in which K is related to the wavelength of the laser radar and the properties of the aerosol, and the
range is 0.67–1. Sm = S(Zm ), σm = σ(Zm ), and Zm is the greatest measurement range. Using the
Klett method, extinction coefficients can be inverted from the MPL-4B and the CL31. Based on the
extinction coefficients at different wavelengths, the microphysical evolution of the haze was analyzed
by combining data with the size spectrum from the Grimm.
Atmosphere 2016, 7, 20
4 of 11
3. Observed Characteristics
In this study, two ceilometers with different wavelength were used to observe a haze episode
that occurred on 2 October 2014. A weather condition analysis showed that high-altitude vertical
wind shear and long-distance transport of particles were the main reason for the formation of the haze.
During the haze process, the upper airflow was mainly from the area of Inner Mongolia, while the
low-level airflow was mainly from the northeast, with variation in direction ranging from northeast
to southeast. Pollutants from Hebei Province were transported to the city of Beijing. With relatively
Atmosphere 2016, 7, 20
11
stable weather
parameters in Beijing, both the transported pollutants and the pollutants4 ofgenerated
by
the mega-city
contributed
to
the
main
part
of
the
haze
process.
the low-level airflow was mainly from the northeast, with variation in direction ranging from
northeast to southeast. Pollutants from Hebei Province were transported to the city of Beijing. With
stable weather
parameters in Beijing, both the transported pollutants and the pollutants
3.1. Surfacerelatively
Meteorological
Elements
generated by the mega-city contributed to the main part of the haze process.
Figure 1 presents the surface meteorological elements during the haze from 16:00 on 2 October to
Surface 2014.
Meteorological
Elements of relative humidity, visibility, and wind velocity are provided
11:00 on 3 3.1.
October
The variations
in the figure, Figure
respectively.
figuremeteorological
shows thatelements
the wind
velocity
remained
as 1 m/s in
1 presentsThe
the surface
during
the haze
from 16:00 as
on 2small
October
to 11:00 on
3 October
The variations
relative
humidity,
visibility,diffusion
and wind velocity
are
the time segment
from
16:002014.
to 01:00.
The lowofwind
velocity
inhibited
of the pollutants,
as
provided in the figure, respectively. The figure shows that the wind velocity remained as small as 1
reported by Pan et al. [5]. As pollutants accumulate, absorption of moisture by nuclei leads to further
m/s in the time segment from 16:00 to 01:00. The low wind velocity inhibited diffusion of the
deterioration
of visibility.
As by
shown
1, in thisaccumulate,
time segment,
the of
relative
humidity
pollutants,
as reported
Pan et in
al. Figure
[5]. As pollutants
absorption
moisture
by nuclei increased
from 55% leads
to 93%
whiledeterioration
the visibility
decreased
from 12
km to1,1.8
km.time
According
to relative
the definition of
to further
of visibility.
As shown
in Figure
in this
segment, the
humidity
increased
from
55%
to
93%
while
the
visibility
decreased
from
12
km
to
1.8
km.
According
haze that was stated previously, haze began at 17:00 when the visibility was less than 10 km and the
to the definition of haze that was stated previously, haze began at 17:00 when the visibility was less
relative humidity was 60%.
than 10 km and the relative humidity was 60%.
Figure 1. Evolution of relative humidity, visibility, and wind velocity during the haze episode.
Figure 1. Evolution of relative humidity, visibility, and wind velocity during the haze episode.
3.2. Analysis of Aerosol Optical Properties
3.2. Analysis ofFigure
Aerosol
OpticaltheProperties
2 presents
variations of the extinction coefficient inverted from the CL31. The slope
method is used to retrieve the boundary layer height from the extinction profiles of aerosol. It is
Figure
2 presents
the variations
of theprogressively
extinction lower
coefficient
inverted
from the
clear
that the boundary
layer becomes
after 18:00
on 2 October.
DueCL31.
to the The slope
method isexistence
used toofretrieve
the
boundary
layer
height
from
the
extinction
profiles
of
aerosol. It is
this boundary layer, the pollutants could not diffuse and accumulated in the lower
atmosphere.
As
shown
in
Figure
2a,
the
color
indicating
the
magnitude
of
the
extinction
coefficient
clear that the boundary layer becomes progressively lower after 18:00 on 2 October. Due to the
especially in the altitude range below 0.5 km. During the night of 2 October,
existence becomes
of this greater,
boundary
layer, the pollutants could not diffuse and accumulated in the lower
radiation inversion restricted pollutant diffusion in the vertical direction. Pollutants concentrated at
atmosphere.
Asofshown
in Figure
indicating
the 22:00,
magnitude
of the
extinction
the top
the boundary
layer2a,
untilthe
thecolor
haze dissipated.
Before
the relative
humidity
increasecoefficient
−1 to 1.84
−1 while radiation
becomes greater,
altitude
range below
0.5increased
km. During
thekm
night
of 2km
October,
from 55%especially
to 90%, andin
thethe
extinction
coefficient
at 532 nm
from 0.65
extinction
coefficientdiffusion
at 910 nm in
increased
from 0.2
km−1 to 0.8Pollutants
km−1, at the concentrated
height of 90 m, at
as the top of
inversion the
restricted
pollutant
the vertical
direction.
shown in Figure 2b.
the boundary layer until the haze dissipated. Before 22:00, the relative humidity increase from 55%
Figure 3 is a radiosonde temperature profile acquired at 08:00 on 3 October,
the vertical
to 90%, and
the extinction
at 532temperature
nm increased
from
0.65height
km´of1 to
km´is1 while the
resolution
is 30 m. Itcoefficient
shows an obvious
inversion
at the
5001.84
m, which
´1
extinctionconsistent
coefficient
910
nm of
increased
from
0.2inkm
km´1 ,08:00
at the
of At
90this
m, as shown
withatthe
height
the boundary
layer
Figureto
2a0.8
at around
on height
3 October.
time,
affected
by
the
inversion
layer,
the
atmosphere
tended
to
be
stable,
and
convection
could
not
in Figure 2b.
occur easily. Because of the low wind speed and the inversion phenomenon caused by temperature,
Figure
3 is a radiosonde temperature profile acquired at 08:00 on 3 October, the vertical resolution
the aerosol particles near the ground could not diffuse, and the accumulation of the aerosol particles
is 30 m. It promoted
shows antheobvious
temperature
emergence
of the haze. inversion at the height of 500 m, which is consistent with the
Atmosphere 2016, 7, 20
5 of 11
height of the boundary layer in Figure 2a at around 08:00 on 3 October. At this time, affected by the
inversion layer, the atmosphere tended to be stable, and convection could not occur easily. Because of
the low wind speed and the inversion phenomenon caused by temperature, the aerosol particles near
the ground could not diffuse, and the accumulation of the aerosol particles promoted the emergence of
the haze.Atmosphere 2016, 7, 20
5 of 11
Atmosphere 2016, 7, 20
5 of 11
Figure 2.Figure
Sequence
diagram
of backscattering
anddiurnal
diurnal
change
of extinction
coefficients
(b).
2. Sequence
diagram
of backscattering (a)
(a) and
change
of extinction
coefficients
(b).
Figure 2. Sequence diagram of backscattering (a) and diurnal change of extinction coefficients (b).
Figure 3. Radiosonde temperature profile at 08:00 on 3 October 2014.
Figure 3. Radiosonde temperature profile at 08:00 on 3 October 2014.
Figure 3. Radiosonde temperature profile at 08:00 on 3 October 2014.
Atmosphere 2016, 7, 20
6 of 11
Atmosphere 2016, 7, 20
6 of 11
3.3. Properties of Hygroscopic Growth
3.3. Properties of Hygroscopic Growth
The relationships between the relative humidity and the extinction are shown in Figure 4. The
Theare
relationships
between
humidity
the results.
extinction
shown
in Figure
4. The
squares
measurement
data,the
andrelative
the lines
are theand
fitting
A are
strong
positive
correlation
squares
measurement
and the lines
are theisfitting
results.
A strong
positive correlation
betweenare
relative
humidity data,
and extinction
coefficient
apparent
at both
laser wavelengths
(532 nm
between
relative
humidity
and
extinction
coefficient
is
apparent
at
both
laser
wavelengths
nm
and 910 nm). The correlation coefficients between relative humidity and extinction coefficient(532
are 0.91
and
910 nm). The
correlation
between
and correlations
extinction coefficient
are
at wavelength
of 532
nm and coefficients
0.9 at wavelength
of relative
910 nm.humidity
The positive
indicate that
0.91
at
wavelength
of
532
nm
and
0.9
at
wavelength
of
910
nm.
The
positive
correlations
indicate
that
the extinction coefficients increased simultaneously with the growth of aerosols caused by moisture
the
extinctionThe
coefficients
increasedhad
simultaneously
with the growth
of aerosolsand
caused
by ability
moisture
absorption.
aerosol particles
dominantly hydrophilic
compositions
strong
to
absorption.
The which
aerosolwas
particles
had
hydrophilic
compositions
strong When
abilitythe
to
absorb moisture,
ideal for
an dominantly
analysis of the
characteristic
of moisture and
absorption.
absorb
was
ideal
an analysis
of the and
characteristic
ofhumidity
moisture showed
absorption.
When
relativemoisture,
humidity which
reached
70%,
thefor
extinction
coefficient
the relative
a negative
the
relative
humidity
reached
70%,
the
extinction
coefficient
and
the
relative
humidity
showed
a
correlation, mainly because of an increase of wind speed at this time. As there was a decrease of
negative
correlation,
mainlyparticles,
because the
of an
increasecoefficient
of wind decreased
speed at this
time. As there
was a
scattered signal
from surface
extinction
correspondingly.
However,
decrease
of
scattered
signal
from
surface
particles,
the
extinction
coefficient
decreased
with the continued increase of relative humidity, the solid-liquid surface chemical reaction caused the
correspondingly.
the continued
increase
of relative
humidity,
thehad
solid-liquid
surface
aerosol extinction However,
coefficient with
to increase
significantly,
which
means that
the haze
entered the
stage
chemical
reaction
caused
the
aerosol
extinction
coefficient
to
increase
significantly,
which
means
of hygroscopic growth [12].
that the
had entered
the stage
hygroscopic
growth
[12]. as the quotient between a particle’s
Thehaze
hygroscopic
growth
factorof(GF)
of a particle
is defined
The hygroscopic
growth
factor
(GF) humidity
of a particle
is defined
as the quotient
between
particle’s
extinction
coefficient at
a given
relative
(RH)
and its extinction
coefficient
at aa (typically
extinction
coefficient
at
a
given
relative
humidity
(RH)
and
its
extinction
coefficient
at
a
(typically
lower) reference humidity (RH0) [13,14]. According to this definition, GF can be written as
lower) reference humidity (RH0) [13,14]. According to this definition, GF can be written as
GF “ σpRHq{σpRH q
(2)
(2)
GFH H (RH ) / (RH0 0 )
in which
which σ(RH)
σ(RH) is
is the
the extinction
coefficient at
is the
the extinction
extinction
in
extinction coefficient
at aa given
given relative
relative humidity
humidity and
and σ(RH
σ(RH00)) is
coefficient at
at aa given
given reference
reference relative
relative humidity.
humidity. For
the case
case of
of this
this haze,
haze, aa relative
relative humidity
humidity of
of 55%
55%
coefficient
For the
was
selected
as
the
reference
relative
humidity
[15].
was selected as the reference relative humidity [15].
Figure 4. Comparison between relative humidity and extinction coefficient for (a) 532 nm and (b) 910 nm.
Figure 4. Comparison between relative humidity and extinction coefficient for (a) 532 nm and
(b) 910 nm.
The hygroscopic growth factors at different values of relative humidity are presented in Figure
5. From the figure, although the hygroscopic growth factors increase with increasing relative
The hygroscopic
factors
values
of relative
are presented
in Figure
5.
humidity,
the detailedgrowth
variations
at at
thedifferent
different
wavelengths
arehumidity
not identical.
Before the
relative
From
the
figure,
although
the
hygroscopic
growth
factors
increase
with
increasing
relative
humidity,
humidity of 70%, the hygroscopic growth factor increases slowly. The growth was relatively slow in
the detailed
variations
at the
aresmall
not identical.
the relative
humidity
this
time segment
because
the different
size of thewavelengths
pollutants was
and both Before
wavelengths
were relatively
of 70%,than
the the
hygroscopic
factor
slowly. The
growth
relatively
slow in
longer
size of thegrowth
pollutants.
In increases
this case, scattering
is much
lesswas
sensitive
to growth
of this
the
time
segment
because
the
size
of
the
pollutants
was
small
and
both
wavelengths
were
relatively
pollutants, especially for the relatively long laser wavelength of 910 nm. The hygroscopic growth
longer than
the
size
of the
this1.89,
case,respectively,
scattering is which
much less
sensitive to
growth
of the
factors
at 532
nm
and
910 pollutants.
nm are 1.88Inand
is consistent
with
the results
pollutants,byespecially
for thethe
relatively
laser of
wavelength
of 910 nm. growth
The hygroscopic
provided
Liu [16]. After
relative long
humidity
72%, the hygroscopic
factor for growth
the 910
nm wavelength increased rapidly and become larger than that of the 532 nm wavelength. The reason
for this change is that, as the pollutants became larger, the longer wavelength became more sensitive
Atmosphere 2016, 7, 20
7 of 11
factors at 532 nm and 910 nm are 1.88 and 1.89, respectively, which is consistent with the results
provided by Liu [16]. After the relative humidity of 72%, the hygroscopic growth factor for the 910 nm
wavelength increased rapidly and become larger than that of the 532 nm wavelength. The reason for
Atmosphere 2016, 7, 20
7 of 11
this change is that, as the pollutants became larger, the longer wavelength became more sensitive to
them
thanthan
the shorter
wavelength
[17]. The
hygroscopic
growth efficiency
of the particles’
to them
the shorter
wavelength
[17].
The hygroscopic
growth efficiency
of theextinction
particles’
coefficient
is
described
by
the
normalized
hygroscopic
growth
factor
GF’.
Generally,
the
definition
of
extinction coefficient is described by the normalized hygroscopic growth factor GF’. Generally, the
GF’
is
the
hygroscopic
growth
factor
GF
divided
by
the
change
of
relative
humidity
[16].
From
the
definition of GF’ is the hygroscopic growth factor GF divided by the change of relative humidity
normalized
growth
factor GF’ that
was calculated
the relative
humidity as
increased
from
[16]. From hygroscopic
the normalized
hygroscopic
growth
factor GF’asthat
was calculated
the relative
73%
to
90%,
the
hygroscopic
growth
efficiency
at
910
nm
and
532
nm
was
16.35
and
9.55,
respectively.
humidity increased from 73% to 90%, the hygroscopic growth efficiency at 910 nm and 532 nm was
The
hygroscopicThe
growth
factor GF’
of the twogrowth
wavelengths
showed
GF’ at the
16.35normalized
and 9.55, respectively.
normalized
hygroscopic
factor GF’
of thethat
twothe
wavelengths
wavelength
910GF’
nmatwas
two times
thanalmost
that attwo
the times
wavelength
532that
nm.atThe
showed thatofthe
the almost
wavelength
of 910larger
nm was
larger of
than
the
difference
between
the
GF’
values
shows
that
the
910
nm
wavelength
becomes
more
sensitive
as
wavelength of 532 nm. The difference between the GF’ values shows that the 910 nm wavelength
particle
size
increases.
becomes more sensitive as particle size increases.
Figure 5. Change of hygroscopic growth factors with relative humidity.
Figure 5. Change of hygroscopic growth factors with relative humidity.
3.4. Relationship between the Hygroscopic Growth Factors and Angstrom Exponent
3.4. Relationship between the Hygroscopic Growth Factors and Angstrom Exponent
With the assumption of the Junge size distribution, the aerosol scattering Angstrom exponent
With the assumption of the Junge size distribution, the aerosol scattering Angstrom exponent
(AE) reveals aerosol size information according to Mie theory. AE has an adverse relationship with
(AE) reveals aerosol size information according to Mie theory. AE has an adverse relationship with
variation of the size distribution of the aerosol, such that the larger the size of the aerosol, the smaller
variation of the size distribution of the aerosol, such that the larger the size of the aerosol, the smaller
the AE. Based on two ceilometers used
the AE. Based on two ceilometers used.
In this study, AE can be derived from the extinction coefficients at 532 nm and 910 nm using the
In this study, AE can be derived from the extinction coefficients at 532 nm and 910 nm using the
following formula:
following formula:
σ532
532 532
lnp  532q{lnp
q
(3)
 δ“ln(
)
(3)
σ910 ) / ln(
910
 910
910
The variation of AE with the change of relative humidity during a portion of the time of the haze
variation in
of Figure
AE with
change 6ofshows,
relative
during
portion
of thethe
time
of the
eventThe
is presented
6. the
As Figure
AEhumidity
decreased
from a2.6
to 1.2 with
increase
haze
eventhumidity.
is presented
Figure of
6. AE
As corresponds
Figure 6 shows,
from
2.6 toof1.2
the
of
relative
The in
variation
withAE
the decreased
hygroscopic
growth
thewith
aerosol,
increase ofthat,
relative
humidity.
Thebecame
variation
of AE
corresponds
withgrowth,
the hygroscopic
growth of the
indicating
as the
aerosol size
larger
due
to hygroscopic
the AE simultaneously
aerosol, indicating
that,
as the
aerosol reached
size became
to to
hygroscopic
growth,
the AE
decreased.
When the
relative
humidity
75%, larger
the AEdue
began
rise. The AE
also showed
decreased.
When
the relative
reached
75%, of
theabout
AE began
rise.changes
The AE
asimultaneously
similar trend with
extinction
coefficient
afterhumidity
the relative
humidity
75%. to
These
also showed
a similar
trendparticulates
with extinction
coefficient
after
the relative
humidity
about
75%.
occurred
because
the larger
that resulted
from
hygroscopic
growth
were of
blown
out
by
Theseand
changes
occurred
becauseparticulates
the larger particulates
resulted
from hygroscopic
were
wind,
the newly
generated
had smaller that
sizes.
In conclusion,
the AE andgrowth
the relative
blown outshowed
by wind,
and the newly
generated
had
smaller sizes.
In within
conclusion,
the AE
humidity
a negative
correlation,
withparticulates
a correlation
coefficient
of 0.83
the relative
and the relative
humidity
showed a negative correlation, with a correlation coefficient of 0.83 within
humidity
range of
55% to 90%.
the relative humidity range of 55% to 90%.
Atmosphere 2016, 7, 20
8 of 11
Atmosphere 2016, 7, 20
8 of 11
Atmosphere 2016, 7, 20
8 of 11
Figure 6. Comparison between relative humidity and Angstrom exponent.
Figure 6. Comparison between relative humidity and Angstrom exponent.
Figure 6. Comparison between relative humidity and Angstrom exponent.
Figure 7 shows the relationship between the hygroscopic growth factor and the AE. During the
Figure
shows
the16:00
relationship
between
the hygroscopic
hygroscopic
growth
factor
andhumidity
the AE.
AE. During
During
the
77 shows
the
relationship
between
the
growth
factor
and
the
the
timeFigure
segment
between
and 22:00
on 2 October
2014, in which
the
relative
increased
time
segment
between
16:00
and 22:00
22:00at
on532
October
2014,
in which
which
the relative
relative
humidity
increased
time
segment
16:00
and
on
22 October
in
humidity
increased
gradually,
thebetween
correlation
coefficients
nm and2014,
910 nm
were the
0.54
and 0.86,
respectively.
The
gradually,
the
correlation
coefficients
at
532
nm
and
910
nm
were
0.54
and
0.86,
respectively.
The
gradually,
theshow
correlation
532 nmhygroscopic
and 910 nmgrowth
were 0.54
relationships
that thecoefficients
correlationat
between
and and
that 0.86,
AE isrespectively.
higher at 910The
nm
relationships
show
that
the
correlation
between
hygroscopic
growth
and
that
AE
is
higher
at
910
nm
relationships
show
that
the
correlation
between
hygroscopic
growth
and
that
AE
is
higher
at
910
nm
than at 532 nm. The reason for the difference is that the 910 nm laser is more sensitive to the larger
than
at 532
532 nm.
nm.
Thethe
reason
for it
the
difference
isthat,
thatas
therelative
910 nm
nmhumidity
laser is
is more
more
sensitive
tosize
the larger
larger
than
at
The
reason
for
the
difference
that
the
910
laser
sensitive
the
particulates.
From
figure,
can
be seenis
increased,
theto
of
the
particulates.
From
the
figure,
it
can
be
seen
that,
as
relative
humidity
increased,
the
size
of
the
particulates.
From the
figure,
can
seen that,
as relative humidity increased, the size of the
particulates become
larger
and itthe
AEbe
become
smaller.
particulates become
become larger
larger and
and the
the AE
AE become
particulates
become smaller.
smaller.
Figure 7. Comparison between hygroscopic growth factor and Angstrom exponent at (a) 532 nm and
Figure
Comparison between hygroscopic growth factor and Angstrom exponent at (a) 532 nm and
(b) 9107.
nm.
Figure 7. Comparison between hygroscopic growth factor and Angstrom exponent at (a) 532 nm and
(b) 910 nm.
(b) 910 nm.
3.5. Distribution Characteristics of Aerosol Particle Spectrum
3.5. Distribution Characteristics of Aerosol Particle Spectrum
Ulevicius etCharacteristics
al. reported the
variations
of the
aerosol size distribution during the haze episode.
3.5. Distribution
of Aerosol
Particle
Spectrum
Ulevicius
et
al.
reported
the
variations
of
the
aerosol
size distribution
during
the haze
episode.
Aerosol effects such as Brownian diffusion, collision, agglomeration,
and
deposition
during
the
Ulevicius
et
al.
reported
the
variations
of
the
aerosol
size
distribution
during
the
haze
episode.
Aerosol
effects
suchhumidity
as Brownian
diffusion,
collision,
agglomeration,
andchanges
deposition
during
the
process of
relative
change
will result
in spectrum
distribution
of atmospheric
Aerosol of
effects
such
as Brownian
diffusion,
collision,
agglomeration,
and deposition
during the
process
relative
humidity
change
will (model:
result
inGrimm
spectrum
atmospheric
aerosol particles
[18].
A particle
monitor
180)distribution
was used tochanges
obtain of
aerosol
particle
aerosol
particles
[18].
A
particle
monitor
(model:
Grimm
180)
was
used
to
obtain
aerosol
particle
concentrations at different aerosol sizes. The characteristics of the aerosol size distribution during
concentrations at different aerosol sizes. The characteristics of the aerosol size distribution during
Atmosphere 2016, 7, 20
9 of 11
process of relative humidity change will result in spectrum distribution changes of atmospheric
aerosol particles [18]. A particle monitor (model: Grimm 180) was used to obtain aerosol particle
Atmosphere 2016, 7, 20
9 of 11
concentrations at different aerosol sizes. The characteristics of the aerosol size distribution during the
haze
that occurred
on 2 October
2014 were
and are and
shown
Figurein
8. Figure
As shown
in Figure
the haze
that occurred
on 2 October
2014analyzed
were analyzed
areinshown
8. As
shown8a,
in
the
peak
for
each
time
at
the
corresponding
particle
radius
is
basically
the
same.
In
the
lower
Figure 8a, the peak for each time at the corresponding particle radius is basically the same. Inpart
the
of
the accumulation
mode (r < 0.3
µm),
aerosol
concentrations
increase monotonously
with time.
lower
part of the accumulation
mode
(r the
< 0.3
µ m), the
aerosol concentrations
increase monotonously
However,
the
distribution
of
the
larger
particulates
is
more
complex
because
this
distribution
with time. However, the distribution of the larger particulates is more complex because has
this
adistribution
relation withhas
hygroscopic
growth
and
other
meteorological
parameters,
as
mentioned
previously.
a relation with hygroscopic growth and other meteorological parameters, To
as
demonstrate
this
phenomenon,
the particulate
distributions atthe
three
points in distributions
time are presented
in
mentioned previously. To demonstrate
this phenomenon,
particulate
at three
Figure
From
Figure
8b, it can
be seen
theFigure
distribution
at 20:00
has that
the highest
concentration
of
points 8b.
in time
are
presented
in Figure
8b.that
From
8b, it can
be seen
the distribution
at 20:00
particles
smaller than
0.3 µm, but
concentration
larger
is not
the highestofat larger
20:00.
has the highest
concentration
of the
particles
smaller of
than
0.3 particulates
µ m, but the
concentration
The
particulate
distributions
show
that
the
particulates
with
small
size
are
stable,
while
the
particulates
particulates is not the highest at 20:00. The particulate distributions show that the particulates with
with
are unstable.
smalllarge
size size
are stable,
while the particulates with large size are unstable.
Figure 8. Aerosol particle spectrum distribution during the fog-haze (a) and at three different points
Figure 8. Aerosol particle spectrum distribution during the fog-haze (a) and at three different points in
in time
time
(b).(b).
In Figure 8b, the relative humidity is 55% at 16:00, 77% at 19:00, and 90% at 22:00. As shown in
In Figure 8b, the relative humidity is 55% at 16:00, 77% at 19:00, and 90% at 22:00. As shown
Figure 8a, when the relative humidity is relatively lower at 16:00, the particle concentration is also
in Figure 8a, when the relative humidity is relatively lower at 16:00, the particle concentration is
relatively small, especially for fine particles within the size range of 0.15 µ m to 0.8 µ m. As the haze
also relatively small, especially for fine particles within the size range of 0.15 µm to 0.8 µm. As the
episode develops, the fine aerosol concentration increases rapidly to the blue curve in Figure 8b. At
haze episode develops, the fine aerosol concentration increases rapidly to the blue curve in Figure 8b.
19:00 and 22:00, the fine mode radius (0.125–0.2 µ m) particle concentrations are similar, but the
At 19:00 and 22:00, the fine mode radius (0.125–0.2 µm) particle concentrations are similar, but the
coarse mode particle concentration is significantly greater at 19:00 than at 22:00. Mie theory was used
coarse mode particle concentration is significantly greater at 19:00 than at 22:00. Mie theory was used
to study the scattering characteristics of particles with different sizes. Firstly, we investigated the
to study the scattering characteristics of particles with different sizes. Firstly, we investigated the
scattering characteristics of particles with a size of 1.5 µ m. The Mie calculation results showed that
scattering characteristics of particles with a size of 1.5 µm. The Mie calculation results showed that the
the phase function at 532 nm has characteristic of large oscillation while scattering of 910 nm is more
phase function at 532 nm has characteristic of large oscillation while scattering of 910 nm is more stable
stable than that at 532 nm. In addition, a comparison between scattering characteristics of particles
than that at 532 nm. In addition, a comparison between scattering characteristics of particles with a
with a size of 1.5µ m and that of particles with a size of 2.5 µ m was conducted. With particle size
size of 1.5 µm and that of particles with a size of 2.5 µm was conducted. With particle size changes
changes from 1.5µ m to 2.5 µ m, the relative changes of scattering intensities at 532 nm and 910 nm
from 1.5 µm to 2.5 µm, the relative changes of scattering intensities at 532 nm and 910 nm were 66%,
were 66%, 89% respectively. Examining the change of AE with time on 2 October (Figure 9), it can be
89% respectively. Examining the change of AE with time on 2 October (Figure 9), it can be seen that,
seen that, because the coarse mode particle concentration begins to decrease at 19:00, the variation of
because the coarse mode particle concentration begins to decrease at 19:00, the variation of AE changes
AE changes to an upward trend. As mentioned previously, an analysis revealed that the decrease of
to an upward trend. As mentioned previously, an analysis revealed that the decrease of the aerosol
the aerosol coarse mode particulate concentration occurred mainly because an increased wind blew
coarse mode particulate concentration occurred mainly because an increased wind blew away the
away the particles that had grown larger. The effect of the wind caused the decrease of the coarse
particles that had grown larger. The effect of the wind caused the decrease of the coarse mode particles
mode particles and the increase of the AE.
and the increase of the AE.
Atmosphere 2016, 7, 20
Atmosphere 2016, 7, 20
10 of 11
10 of 11
Figure 9. Change of Angstrom exponent with time.
Figure 9. Change of Angstrom exponent with time.
4. Conclusion and Discussion
4. Conclusions and Discussion
Combining the data from the MPL (532 nm) and the CL31 (910 nm) ceilometers with the aerosol
Combining the data from the MPL (532 nm) and the CL31 (910 nm) ceilometers with the aerosol
size distribution monitor, the characteristics of the aerosols during the haze episode were analyzed.
size distribution monitor, the characteristics of the aerosols during the haze episode were analyzed.
For the two different wavelengths, the particles showed different optical characteristics, and the
For the two different wavelengths, the particles showed different optical characteristics, and the
longer wavelength was more sensitive to larger particles. As well as it affects visibility, relative
longer wavelength was more sensitive to larger particles. As well as it affects visibility, relative
humidity showed a close relationship with hygroscopic growth and AE. The characteristics of the
humidity showed a close relationship with hygroscopic growth and AE. The characteristics of the size
size concentration spectrum verified the analyzed results for hygroscopic growth and AE. From the
concentration
analyzedthat
results
for hygroscopic
growth and AE.
From
the optical
results
results of thisspectrum
study, itverified
can bethe
concluded
more
particulate characteristics
than
only
of this study, it can be concluded that more particulate characteristics than only optical characteristics
characteristics can be retrieved from ceilometer data. Furthermore, the aforementioned ceilometer
can
be retrieved
ceilometer
data.
the aforementioned
network
be
network
will be from
constructed
in the
nearFurthermore,
future according
to the scheduleceilometer
of the CMA.
Based will
on the
constructed
in the
near the
future
according
the schedule
of the
CMA.
Basedand
on often
the facts
mentioned
facts mentioned
above,
haze
in China,to
which
is a serious
social
problem
brings
adverse
above,
the
haze
in
China,
which
is
a
serious
social
problem
and
often
brings
adverse
effects,
can
effects, can be monitored more effectively by combining data from the ceilometer network with
be
monitored
more
effectively
by
combining
data
from
the
ceilometer
network
with
those
from
the
those from the existing particulate monitoring instruments.
existing particulate monitoring instruments.
Acknowledgments: The work was supported by the Natural Science Foundation of Jiangsu Province
Acknowledgments:
The work
supported
by the Naturaland
Science
Foundation
of Jiangsu
Province (Grant
No.
(Grant No. BK20141480
andwas
Grant
No. BE2015003-4)
the National
Science
Foundation
of China
BK20141480 and Grant No. BE2015003-4) and the National Science Foundation of China (Grant No. 41304124).
(Grant No. 41304124).
Author Contributions: All authors contributed immensely. Jing Yuan collected and analyzed the data and wrote
Author
Contributions:
All authors
contributed
immensely.
Jing Yuan
collected
andHuang
analyzed
the data
the
paper.
Lingbing Bu defined
the major
working scheme
and modified
the paper.
Xingyou
provided
key
data.
Haiyang
Gao
was inLingbing
charge of the
of optical
Rina Sa and
processed
and analyzed
the
and wrote
the
paper.
Bu observation
defined the
major parameters.
working scheme
modified
the paper.
haze data.
Xingyou Huang provided key data. Haiyang Gao was in charge of the observation of optical
Conflicts
of Interest:
authors declare
no conflicts
interest.
parameters.
Rina SaThe
processed
and analyzed
theof
haze
data.
Conflicts of Interest: The authors declare no conflicts of interest.
References
1.
Zhang, Q.; Shen, Z.; Cao, J.; Zhang, R.; Zhang, L.; Huang, R.J.; Zheng, C.; Wang, L.; Liu, S.; Xu, H.; et al.
References
Variations in PM2.5 , TSP, BC, and trace gases (NO2 , SO2 , and O3 ) between haze and non-haze episodes in
1.
Zhang,over
Q.; Shen,
Cao, Atmos.
J.; Zhang,
R.; Zhang,
L.; Huang,
R.J.; Zheng, C.; Wang, L.; Liu, S.; Xu, H.; et al.
winter
Xi’an,Z.;
China.
Environ.
2015, 112,
64–71. [CrossRef]
Variations
in
PM
2.5
,
TSP,
BC,
and
trace
gases
(NO
2
,
SO
2
,
and
3) between
haze2014,
and 189,
non-haze
in
2.
Li, M.; Zhang, L. Haze in China: Current and future challenges.OEnviron.
Pollut.
85–86.episodes
[CrossRef]
winter over Xi’an, China. Atmos. Environ. 2015, 112, 64–71.
[PubMed]
2.
Li, M.; Zhang,
HazeGong,
in China:
Current
and future
challenges.
Environ.
Pollut.
2014, 189,
85–86. based on
Zhang,
M.; Ma,L.Y.Y.;
W.; Zhu,
Z. Aerosol
optical
properties
of a haze
episode
in Wuhan
3.
3.
Zhang,
M.;
Ma,
Y.Y.;
Gong,
W.;
Zhu,
Z.
Aerosol
optical
properties
of
a
haze
episode
in
Wuhan
based on
ground-based and satellite observations. Atmosphere 2014, 5, 699–719. [CrossRef]
ground-based
and
satellite
observations.
Atmosphere
2014,
5,
699–719.
4.
Huang, X.; Yang, X.; Geng, F.; Zhang, H.; He, Q.; Bu, L. Aerosol measurement and property analysis based on
4.
Huang,
X.; Yang,
Geng, F.; Zhang,
He,
Q.; Bu, L.China.
Aerosol
measurement
and property
analysis
based
data
collected
by aX.;
micro-pulse
LIDARH.;
over
Shanghai,
J. Opt.
Soc. Korea 2010,
14, 185–189.
[CrossRef]
on data collected by a micro-pulse LIDAR over Shanghai, China. J. Opt. Soc. Korea 2010, 14, 185–189.
Atmosphere 2016, 7, 20
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
11 of 11
Pan, H.; Geng, F.H.; Chen, Y.H.; He, Q.; Zhang, H.; Kang, Y.; Mao, X.; Wang, H. Analysis of a haze event by
micro-pulse light laser detection and ranging measurements in Shanghai. Acta Sci. Circumstantiae 2010, 30,
2164–2173.
Zhang, W. Spatial and temporal variability of aerosol vertical distribution based on lidar observations:
A haze case study over Jinhua basin. Adv. Meteorol. 2015, 2015, 349592. [CrossRef]
Bu, L.B.; Yuan, J.; Gao, A.Z.; Lei, Y.; Guo, W.; Gao, H.; Huang, X. Analysis of haze-fog events based on laser
ceilometer. Acta Photonica Sin. 2014, 43. [CrossRef]
He, Q.S.; Mao, J.T. Micro-pulse lidar and its applications. Meteorological. Sci. Technol. 2004, 32, 219–224.
Dai, Y.J.; Cai, X.P.; Chen, X.J.; Chen, Z.L. Micro pulse lidar for measuring clouds. Infrared Laser Eng. 2000, 29,
1–5.
Grimm. Mobile Dust Monitor ENVIRON CHECK 180 User’s Guide; Grimm: München, Germany, 2010.
Kleet, J.D. Stable analytical inversion solution for processing lidar returns. Appl. Opt. 1981, 20, 211–220.
[CrossRef] [PubMed]
Yin, K.X.; Fan, C.Y.; Wang, H.T.; Qiao, C.; Zhang, P. Aerosol extinction characteristics in haze and fine weather
in southeast coastal area of China. High Power Laser Part. Beams 2015, 27. [CrossRef]
Meier, J.; Wehner, B.; Massling, A.; Birmili, W.; Nowak, A.; Gnauk, T.; Brüggemann, E.; Herrmann, H.; Min, H.;
Wiedensohler, A. Hygroscopic growth of urban aerosol particles in Beijing (China) during wintertime:
A comparison of three experimental methods. Atmos. Chem. Phys. 2009, 9, 6865–6880. [CrossRef]
Kotchenruther, R.A.; Hobbs, P.V.; Hegg, D.A. Humidification factors for atmospheric aerosols off the
mid-Atlantic coast of the United States. J. Geophys. Res. Atmos. 1999, 104, 2239–2251. [CrossRef]
Bo, G.Y.; Liu, D.; Wu, D.C.; Wang, B.; Zhong, Z.; Xie, C. Two-Wavelength Lidar for observation of aerosol
optical and hygroscopic properties in fog and haze days. Chin. J. Lasers 2014, 41, 0113001.
Liu, X.G.; Zhang, Y.H. Advances in research on aeroslo hygroscopic properties at home and abroad.
Clim. Environ. Res. 2010, 15, 808–816.
Chen, L.Y.; Jing, F.T.; Chen, C.H.; Hsiao, T. Hygroscopic behavior of atmospheric aerosol in Taipei.
Atmos. Environ. 2003, 37, 2069–2075. [CrossRef]
Ulevicius, V.; Trukmas, S.; Girgzdys, A. Aerosol size distribution transformation in fog. Atmos. Environ. 1994,
28, 795–800. [CrossRef]
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons by Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).