Introduction
Single-scattering albedo (ω) is one of the most important aerosol
optical properties. It influences the aerosol's radiative effect and is a
significantly uncertain factor. Defined as the ratio of absorption to the
sum of scattering plus absorption, ω represents the combined effect
of the two processes and acts as an indicator of aerosols' net radiative
effect.
Under dry conditions (< 30 % RH), the value of ω is
determined by the particle number–size distribution, the complex refractive
index, and the particle shape (Covert et al., 1972). Due to the complexity
of aerosol processes such as production, transformation, in situ chemical
reactions, and removal, the value of ω is highly variable
(Heintzenberg et al., 1997). Especially, aerosol scattering can be
significantly enhanced by elevated relative humidity (RH). The hygroscopic
growth of aerosol particles determines the water content in the particles
and changes the composition and the size of aerosols. As a result, the value
of ω varies with RH. In polluted areas, compared with dry state,
scattering increases by at least 50 % at RH of around 90 %, mainly from
the increase of water (Cheng et al., 2008; Pan et al.,
2009; Fierz-Schmidhauser et al., 2010b; Langridge et al., 2012; Li et al.,
2013). This enhancement of scattering is stronger for marine aerosols or in
clean regions (Fierz-Schmidhauser et al., 2010b; Carrico et al., 2000; Adam et
al., 2012), and weaker for dust aerosol (Pan et al., 2009; Carrico et al.,
2000). Aerosol absorption is often considered to vary slightly with RH,
while Brem et al. (2012) reported the enhancement of aerosol absorption at
high RH. Thus, the value of ω can be RH dependent and increase by at
least 0.05 at high RHs in a polluted atmosphere (Cheng et al.,
2008; Fierz-Schmidhauser et al., 2010b; Li et al., 2013; Jung et al., 2009).
Because ambient air is most often sampled in a shelter or structure, it is
very important to measure and report the RH at the point of measurement and
to apply coincident measurement of aerosol hygroscopicity (or a model
thereof) to quantify the ambient ω (Nessler et al., 2005).
Due to the high sensitivity of radiative forcing to the variation of
ω, it is essential to obtain atmospherically relevant values of
ω for climate models and photochemical models. It has long been
known that the aerosol radiative forcing is sensitive to ω, and the
transition between positive and negative forcing of direct aerosol effect
takes place when the value of ω is about 0.85 (Heintzenberg et al.,
1997; Cheng et al., 2008; Wang et al., 2007). Combined with the other aerosol
optical properties, the change of ω at high RHs can strengthen the
forcing by a factor of 2 or more (Stock et al., 2011; Fierz-Schmidhauser et
al., 2010a; Cheng et al., 2008; Massoli et al., 2009).
As for the tropospheric photochemical process,
ω as well as the aerosol optical depth (τ) are the relevant
parameters in the determination of ultraviolet (UV) radiation and the photolysis
rate coefficient (Reuder and Schwander, 1999). There are many important
photolysis reactions in the troposphere, such as of NO2, ozone, etc. Among
these reactions, the photolysis of NO2 accounts for the most ozone
production in the troposphere and is the most representative. The NO2
photolysis rate coefficient (JNO2) is widely used in the analysis of
ozone photochemistry (Seinfeld and Pandis, 2006; Dickerson et al.,
1997; Palancar et al., 2013). Ozone photochemistry can be either inhibited or
enhanced by aerosols also depending on ω (Palancar et al., 2013; Li
et al., 2011; Dickerson et al., 1997; Tang et al., 2003; Liu et al., 2013).
Sensitivity studies show that RH is as important as the aerosol loading in
the influence of aerosol on ozone photolysis (He and Carmichael,
1999; Jacobson, 1998).
The North China Plain (NCP), with several megacities and as the location of
many industries, suffers frequent severe aerosol pollution episodes (Xu
et al., 2011; Ran et al., 2011, 2012; Liu et al., 2009). The rapid
industrial development created numerous sources of primary aerosols and the
precursors of secondary aerosol production. The intensive use of coal and
biomass fuels makes the NCP a region of high concentration of black carbon.
Clouds and precipitation in this region might be modified by high aerosol
loading (Zhao et al., 2006; Deng et al., 2009). Strong absorption and
core-shell mixing state of light-absorbing carbonaceous (LAC) were found (Ma
et al., 2011, 2012). The growth factors at RHs of up to 98.5 %
measured by a high humidity tandem differential mobility analyzer (HH-TDMA)
indicated the existence of a dominant more-hygroscopic group of aerosols
(Liu et al., 2011). This result agreed well with the retrieved values from
the microbalance UMT-2 (Mettler Toledo, Switzerland) (Liu et al., 2014) and
in combination contributed to enhancing extinction at high RHs and
to the low visibility on hazy days (Chen et al., 2012). Regional ozone
pollution occurred in the NCP, but the role of the radiation in ozone
photochemistry at high aerosol conditions is still unresolved (Ran et al.,
2011, 2012).
In this study, the RH dependence of aerosol optical properties are
represented and their influences on UV radiation are investigated. A Mie code
considering the coating of aerosols (Y. Cheng et al., 2009) and a radiation
transfer model (Madronich and Flocke, 1997) are used. The descriptions of
data, calculations and models are presented in Sect. 2; overviews of aerosol
optical properties are in Sect. 3; results of modeled UVB irradiance and
JNO2 are represented in Sect. 4; and there is a summary in
Sect. 5.
Impact of hygroscopic growth on aerosol optical properties
Overview of the ambient aerosol optical properties
Ambient aerosol optical characteristics, including σep,
σsp, σap and ω, are calculated using
the method introduced in Sect. 2 for the data set measured during the HaChi
summer campaign. The aerosol optical properties as well as the meteorological
parameters during the observation period are shown in Fig. 1. For the
majority of the observation, the 1 min wind speed is mostly less than
5 m s-1 (Fig. 1a) and the ambient RH is mostly between 60 and 95 %
(Fig. 1b). The overall σsp is within the range of
100–4000 Mm-1, while the σap is mainly lower than
200 Mm-1 (Fig. 1d). As a result, the majority of ω is higher
than 0.85 and is always close to 0.95 (Fig. 1c), which is in good agreement
with the retrieved values from 0.85 to 0.95 by the Aerosol Robotic Network
(AERONET) (Dubovik et al., 2002).The weakening effect of the wind on both
σap and σap can be found, for example, on
25 July and 14 August. But at very high ambient RH, the decrease of
σsp could be neutralized by the enhancement due to the
hygroscopicity. As a consequence, the ω is affected pronouncedly by
the RH and unaffected by the wind speed. From 17 to 18 July, when both the RH
and the wind speed were high, ω approached the high value of 0.96
during the day, with relatively lower σap compared with
σsp. By contrast, for example, on 16 and 22 July and
12 August, with typical wind speed (about 3 m s-1) and relatively low
RH (about 50 %), the decrease of σsp was stronger than
that of σap, and ω reached the low values of 0.75,
0.7 and 0.75, respectively, on those three days. The value of ω was
not sensitive to the direction of the wind. Significant diurnal patterns of
ω can also be found and is confirmed as follows.
In Fig. 2, the diurnal variations of aerosol optical properties are analyzed
and verified by calculating the autocorrelation coefficient. Pronounced
diurnal patterns are found in all variables with a maximum at about 06:00 LT
(local time)
and a minimum at about 16:00 LT. Considering the 25 and 75th percentiles,
values of σsp ranged from about 500 to about
2000 Mm-1 (Fig. 2b), and σap ranged from about
40 to about 100 Mm-1 (Fig. 2c). The ambient ω reached
its maximum (about 0.95) at about 06:00 LT and minimum (about 0.9) at about
18:00 LT (Fig. 2d), mainly attributed to the diurnal variation of ambient
RH. These clear diurnal patterns are consistent with the high values (larger
than 0.1) of the autocorrelation coefficient at the time intervals of 24 h
(Fig. 2f). This obvious diurnal pattern of ambient ω may cause large
variation of aerosol radiative forcing in climate models.
The diurnal patterns of the ambient σep,
σsp and σap are similar to their diurnal
patterns in the dry state found by Ma et al. (2011). However, as Ma et
al. (2011) reported, ω in the dry state reaches its peak at noon and a
minimum in the morning and again in the evening, and the diurnal pattern of
the ω is less notable. By taking aerosol hygroscopicity into account,
the diurnal pattern of σsp is amplified because of the
similar diurnal variation of ambient RH. However, the absorption is less
influenced by the hygroscopic growth of particles. This large difference of
amplification between σsp and σap leads to
the large modification of the diurnal pattern of ambient ω from that
in dry state.
The RH-dependent aerosol optical properties
The aerosol optical properties are shown along with the ambient RHs in Fig.
3. The sensitivity of the σsp to the ambient RH is strong.
The increases of σep and σsp with the
increase of RH are significant, although slight decreases occur at a RH of around
90 % (Fig. 3a, b). So do the standard deviations (SD) of
σep and σsp. However, σap
is not sensitive to RH and fluctuates slightly with the increase of RH
(Fig. 3c). As a result, the enhancement of ω is sustained from about
0.89 at a RH of 55 % to about 0.95 at a RH of 94 % (Fig. 3d). More
detailed statistical characteristics of these aerosol optical properties at
different ranges of ambient RHs are listed in Table 1. It was found that the
increase of the ambient ω from 0.87 to 0.96 brought about nearly 4.5
times the enhancement of aerosol direct radiative forcing (Cheng et al., 2008).
In the NCP, it is reasonable that the negative radiative forcing will be
strengthened by the increase of ω.
Statistical values of aerosol optical properties measured at
different ranges of RH.
RH/ (%)
50 ∼ 60
60 ∼ 70
70 ∼ 80
80 ∼ 90
> 90
σep (Mm-1)
Mean
478
654
953
1384
1479
SD
252
363
535
839
746
Median
415
558
771
1185
1349
σsp (Mm-1)
Mean
429
599
885
1309
1421
Std
230
335
503
802
715
Median
371
515
729
1120
1305
σap(Mm-1)
Mean
50
56
68
75
58
Std
30
34
39
48
40
Median
38
46
60
65
49
ω
Mean
0.887
0.912
0.924
0.943
0.961
SD
0.050
0.028
0.028
0.022
0.015
Median
0.892
0.919
0.931
0.949
0.966
In addition, a specific case (the AVG-PRM case; dotted lines in Fig. 3)
which used average parameters is calculated at different RHs. This case
proves to be representative of the NCP by the analysis as follows, in Fig. 4, and is
compared with the ambient aerosol optical properties here. The variations of
the optical quantities in this case are smooth and monotonic. The congruence
between the ω in the AVG-PRM case and the ambient mean ω is
achieved, especially at high RHs. The AVG-PRM case is representative of the
average status of the aerosol at various RHs in the NCP in summer and will be
used in the following analysis of the JNO2 profile.
The absorption coefficient of the aerosol in the NCP seems to be independent
of RH in Fig. 3c, which is expected according to Pan et al (2009).
Therefore, the value of ω at ambient conditions can be estimated from
the two independent parameters: the ambient RH and ω at dry state
(ω0) (Pan et al., 2009). The relationship between ω, RH and
ω0 is presented in Fig. 4b, along with the frequency distribution
of the measured ω0 (Fig. 4a) and RH (Fig. 4c). As expected,
ω approaches higher values at either higher ω0 or higher
RH. Specifically, ω is more sensitive to ω0 at lower RH
and more sensitive to RH at higher RH.
Frequency distributions of ω0 and RH (respectively, a and
c). Calculated ω at given ω0 and RH
(b).
The distribution of ω0 is mainly in the range of 0.80–0.95 and
has an average of about 0.86 (Fig. 4a), which agrees with the result in Ma et
al. (2011). Considering that over half of the ω0 values are in the
range of 0.85 ∼ 0.9, the value of 0.863 is representative for the NCP.
The ambient RHs are distributed almost evenly between 60 and 95 %, except
for the higher frequency at around 95 % (Fig. 4c). It is essential to take
the enhancement of ω at high RHs into account. To sum up, the majority
of ω in the NCP can be described as the RH dependence at ω0
of 0.863, i.e., ω(ω0 = 0.863, RH), here referred to as the AVG-PRM
case.
The overall influence of input parameters of the Mie model on ω at
different RHs is investigated by a Monte Carlo simulation. As shown in
Table 2, the uncertainties of both the measurements and the constants are
assumed based on previous studies (Wiedensohler et al., 2012; Petzold and
Schonlinner, 2004; Cheng et al., 2006). A detailed description can be found of
Ma et al. (2012). The κi in each mode and the rext-LAC
mentioned in Sect. 2.2 are obtained from the results of relevant studies (Ma
et al., 2012; Liu et al., 2011). Considering the wide range of the ambient
RHs, the Monte Carlo simulations are conducted independently at different
RHs. There are 2000 runs in each simulation and the standard deviations of
ω reveal the uncertainty in the calculation of ω.
The results of the Monte Carlo simulations are listed in Table 3. The
standard deviation of ω is smaller at higher RH, ranging from 0.0308
at dry state to 0.0124 at the RH of 93 %. Moreover, we calculated
dω/dRH and analyzed its influence on the standard deviation of
ω. dω/dRH is multiplied with the standard deviation of
the RHs and then divides the standard deviation of the simulated ω,
i.e., σRHdωdRH/σω.
This variable represents the contribution of the uncertainty of measured RH
on the uncertainty of calculated ω. The results are characterized by
the extremely low contribution (< 5 %) at low RHs and the main
contribution (> 50 %) at high RHs. These main contributions
of the uncertainty of RH to the uncertainty of ω reflect the
importance of the aerosol hygroscopic growth on the calculation of ω
at high RHs.
Uncertainties of the input parameters in the Monte Carlo
simulations.
Item
Relative standard
deviation %
Dp,TDMPS
1.1
Dp,APS
3
NTDMPS,3-20nm
10
NTDMPS,20-200nm
3.3
NTDMPS,200-700nm
8.3
NAPS
3.3
σap
4
MAE = 6.6 m2 g-1
9.1
ρLAC = 1.5 g cm-3
11
nLAC = 1.80
0.5
iLAC = 0.54
0
nnon = 1.55
4
inon = 1e-7
6.6
rext-LAC
40
κ50nm = 0.25
24
κ100nm = 0.27
15
κ200nm = 0.38
13
κ250nm = 0.39
13
RH
3
Impact of the aerosol hygroscopic growth on the JNO2 profile: an application
The relationship between the modeled UVB irradiance and ω
JNO2 is affected pronouncedly by the UVB irradiance, which is
determined by the solar zenith (θ), ω and τ. As shown in
Fig. 5a, both the measured (dots) and the calculated (lines) UVB irradiance
decreases with increasing (secant of) solar zenith angle (sec(θ)) and
τ. When ω is higher, UVB irradiance becomes larger. Clearer
details are shown in Fig. 5b. The measured UVB at higher τ (the dots
with warmer colors) is closer to the modeled UVB at higher ω (solid
lines), while the measured UVB at lower τ (colder dots) is closer to the
modeled UVB at lower ω (dashed lines). The relative deviations of UVB
between the two ω conditions increase from about 10 %
(τ = 0.5) to about 700 % (τ = 4.5). At higher τ (or RHs), the
larger difference between high ω and low ω reveals the
importance of an accurate value of ω for the UVB irradiance simulation.
It is essential to take the aerosol hygroscopic growth into consideration in
UVB simulations.
The standard deviation of ω (σω), the rate of
change of ω with RH dωdRH and
the contribution of RH to the uncertainty of
ωσRHdωdRH/σω
at different RHs.
RH (%)
σω
dωdRH
σRHdωdRH/σω (%)
0
0.0308
0.024
2.3
10
0.0285
0.034
3.5
20
0.0289
0.057
5.9
30
0.0282
0.058
6.1
40
0.0260
0.065
7.51
50
0.0247
0.092
11.2
60
0.0227
0.114
15.1
70
0.0200
0.155
23.2
80
0.0174
0.192
33.0
82
0.0159
0.240
45.3
84
0.0154
0.240
46.8
86
0.0144
0.260
54.2
88
0.0136
0.285
62.9
90
0.0133
0.313
70.7
91
0.0128
0.320
75.0
92
0.0127
0.320
75.6
93
0.0124
0.325
78.6
In order to understand the influence of RH-dependent ω on the UVB
irradiation, the UVB irradiations both near the ground and above the boundary
layer are calculated in three cases, i.e., the high-ω case, the
low-ω case and the AVG-PRM case. In the high-ω case and the
low-ω case, UVB irradiations are calculated at a fixed ω with
different τ, similar to Dickerson et al (1997). In the AVG-PRM case,
both τ and ω vary a lot with RHs, as presented in Table 4. As
shown in Fig. 6, the UVB irradiation in all cases near the ground (the solid
lines) decrease as τ rises. The decrease of the low-ω case is
stronger than that of the high-ω case. For the AVG-PRM case, a transformation
of the UVB irradiation from the low-ω case to high-ω case can
be recognized. Specifically, when τ is about 1.6, ω is higher than
0.96 and the closer value to the high-ω case is expected. Considering
this stronger increase of ω than τ at RHs lower than 90 %,
the transformation mentioned above is rapid. At the top of the boundary layer
(dashed lines), as τ increases, the UVB irradiations in the low-ω
case decrease while the UVB irradiations in the high-ω case decrease
slowly only at high τ. Similarly to the result at the surface, UVB
irradiation in the AVG-PRM case gets closer to the high-ω case quickly as
τ increases and stays undiminished until τ becomes higher than 2.5.
In the polluted NCP in moist summers, both the aerosol loading and the ambient
RHs are always high, resulting in both high τ and high ω. High
UVB irradiance is likely to happen at high τ and will affect
relevant radiative processes, such as the photolysis of ozone and NO2.
(a) The dots represent the measured UVB irradiance. The
lines represents the modeled UVB irradiance at ω of 0.863 (dashed
lines) and 0.985 (solid lines). Colors represent the value of τ, i.e., the
warmer/colder the color is, the higher/lower τ is revealed.
(b) Identical to (a) but with finer ranges of τ as shown in
the figures.
The values of τ and ω at selected RHs from the AVG-PRM case.
RH/%
τ
ω
0
0.51
0.863
15
0.53
0.872
30
0.55
0.878
45
0.59
0.887
55
0.64
0.895
64
0.69
0.904
73
0.79
0.916
86
1.12
0.942
90
1.39
0.954
92
1.61
0.961
94
1.99
0.968
95
2.24
0.972
96
2.64
0.976
97
3.19
0.980
98
4.16
0.985
The influence of aerosol hygroscopic growth on the JNO2 profile
In Fig. 7a and b, we represent the modeled JNO2 at
different conditions to study the influence of RH-dependent ω on
photolysis. The ω values are 0.863 (lowest value in the NCP) and 0.985
(highest value in the NCP). For the JNO2 profile
at τ of 0 (or the original profile), the surface JNO2 is
about 0.011 s-1, and the maximum of JNO2 is at the
higher level, with its value of about 0.014 s-1. As τ increases,
JNO2 decreases the most in the low-ω case, especially in the
range of 2 km above the ground (Fig. 7a). For the high-ω case,
JNO2 decreases near the ground but increases at heights higher
than 2 km. At the height of 2 km, JNO2 at τ of 4.5 is
about 0.019 s-1, enhanced by about 58.3 % compared with
JNO2 at τ of 0 (Fig. 7b). At heights lower than 2 km
and above the surface, JNO2 increases first and decreases later.
At 1 km above the ground, JNO2 increases by 39 % and then
decreases by 6 %. These results are in accordance with the study of
Dickerson et al. (1997).
In Fig. 7c, the vertical profiles of JNO2 at different τ
values are calculated with the AVG-PRM case. Similarly to the high-τ case,
JNO2 at high altitudes become larger at higher τ. At the
height of 1 km, JNO2 increases by 58.3 % at τ of 4.16,
which is larger than in the high-ω case. At heights lower than 2 km
and above the surface, the decreases of JNO2 start at lower
altitudes and are smaller than in the high-ω case. And at the height of
1 km, JNO2 increases by 30.4 % at τ of 4.16. This
feature may lead to a higher JNO2 at the top of the boundary
layer in the AVG-PRM case.
The dependence of UVB irradiance with τ at the surface (solid
lines) and at the top of the boundary layer (dashed lines) in three cases:
ω = 0.985 (circle), ω = 0.863 (square) and RH-dependent
ω (AVG-PRM case; triangle).
dJNO2/dτ at the height of 1 km for different
τ values are shown in Fig. 8. In the low-ω case, the negative
dJNO2/dτ is maintained and reaches its minimum when
τ is between 2 and 2.5. In the high-ω case, the growth rate is
positive when τ is lower than 2.5 and becomes negative when τ is
higher than 3, resulting in a maximum of JNO2 when τ is
between 2.5 and 3. In the AVG-PRM case, dJNO2/dτ is
positive when τ is lower than 4 and is likely to stay positive as τ
increases. Compared with the high-ω case,
dJNO2/dτ in the AVG-PRM case is higher and stays
positive, resulting in higher JNO2 at higher τ conditions.
This higher increase of JNO2 may result in a weaker decrease of
ozone photolysis in polluted conditions and the heavy ozone pollution may
take place along with high τ in the moist and polluted NCP.
Summary
In this paper, the aerosol optical properties at different RHs, including
σep, σsp, σap and
ω, are calculated with a Mie model based on the aerosol measurements
during the HaChi project. The impact of the aerosol hygroscopic growth from
HH-TDMA on ω and the corresponding uncertainty are analyzed. A derived
parameter, the influence of RH-dependent ω on the UVB irradiance and
the JNO2 are also investigated.
The hygroscopic growth influences not only aerosol PNSDs but also the
refractive index of the aerosol. In this study, the shell of the core-shell
mixed aerosol is assumed to be composed of less-absorbing, water-soluble
components with the refractive index of 1.53–10-7i. As the ambient RH
increases, water vapor is taken up by the shell and the less-absorbing
components dissolve to maintain water vapor equilibrium between the ambient
air and the liquid shell. The refractive index of the shell is determined by
the water content and the solute together, and decreases with the RH. At the
RH of 90 %, the refractive index of the shell of the accumulation-mode
aerosol decreases to 1.37–10-7i, which is close to the refractive
index of water. The variations of the refractive indices and PNSDs with RH
will modify the aerosol optical properties by Mie theory.
Altitude profiles of JNO2 in three cases (a,
b, c), as in Fig. 6. Colors represent the value of τ, as
expected. The black lines indicate the height of 2 km.
The increase of JNO2 with τ for three cases:
ω = 0.985 (circle), ω = 0.863 (square) and an RH-dependent
ω (AVG-PRM case; triangle).
Ambient aerosol optical characteristics, such as σep,
σsp, σap and ω, during the HaChi
summer campaign are calculated. A significant sensitivity of ω to the
ambient RH is recognized, which is mainly attributed to the variations of
refractive indices and the aerosol size due to hygroscopic growth. Because of
the sensitivity to the RH, the diurnal patterns of σep and
σsp are evident and similar to the diurnal pattern of
ambient RH. Compared with the diurnal patterns at dry state in Ma et
al. (2011), the variations of the ambient σep and the
ambient σsp are amplified pronouncedly. The diurnal
variation of the ambient σap is gentle and similar to that
at dry state in Ma et al. (2011). Due to the strong enhancement of
σsp and the slight increase of σap with RH,
the diurnal pattern of ω is significant and changes a lot compared
with that at dry state reported by Ma et al. (2011). Considering the
insensitivity of σap to RH, the ambient ω can be
determined by its value at dry state, i.e., ω0, and RH. By
analyzing its frequency, ω0 during the HaChi campaign concentrated
mostly at the value of 0.863. Therefore, the RH dependence of ω in the
NCP can be represented by a dry state ω of 0.863, increasing with the
RH following a characteristic RH dependence curve (the AVG-PRM case). This
representative RH-dependent ω can be used in the calculation of the
radiative transfer process. The uncertainty of the calculation of ω
due to the uncertainty of the input parameters in the Mie model is also
investigated by Monte Carlo simulations. The result shows that the standard
deviation of ω decreases from 0.03 at lower RHs to 0.013 at RHs higher
than 90 %.
The RH-dependent ω is applied in the analysis of the JNO2
(the photolysis rate coefficient of NO2, whose photolysis accounts for
the most tropospheric ozone production) profile to evaluate
the impact of aerosol hygroscopic growth on ozone photochemistry.
JNO2 depends on the UV irradiances and is thus affected by
aerosol optical properties. The influence of ω on the UVB irradiances
is investigated by comparing the modeled UVB irradiances and measured UVB
irradiances. A good agreement between the model result and the observation is
reached. It is demonstrated that the modeled UVB irradiances are sensitive to
ω, especially at high τ, indicating the importance of the
accuracy of ω in the calculation of UVB irradiances. Then the UVB
irradiation at the RH-dependent ω condition (the AVG-PRM case) is
analyzed. The variations of the UVB irradiation with τ at the
RH-dependent ω are close to those at a fixed high ω. This
similarity between the RH-dependent ω case and fixed high-ω
case results from the stronger enhancement of ω than τ at RHs
lower than 90 %, and is important in the study of JNO2.
Previous studies show that at a fixed high ω, JNO2 at the
top of the boundary layer increases with τ. This amplification of
JNO2 can weaken the inhibition of the aerosol on ozone
photolysis and may bring about simultaneous high aerosol loading and high
ozone concentration. In this study, JNO2 at the RH-dependent
ω is found to increase with τ as well. At τ of 4.16,
JNO2 at the height of 1 km increases by 30.4 % compared
with that at τ of 0.51. The weakening of suppression of ozone production
by aerosol is likely to happen in the polluted and moist NCP, and may lead to
more ozone production in polluted conditions. The increase of
JNO2 due to the aerosol hygroscopic growth above the upper
boundary layer may affect the ozone photochemistry and this should be
introduced and evaluated in the atmospheric chemical models.