Elsevier Editorial System(tm) for

Atmospheric Environment or its open access mirror

Manuscript Draft

Manuscript Number:

Title: Observed interaction among haze, fog and atmospheric boundary

layer during a haze-fog episode in the Yangtze River Delta region,
eastern China

Article Type: Research Paper

Keywords: haze-fog episode; interaction; potential temperature jump;

planetary boundary layer height; PM2.5; vertical distribution.

Corresponding Author: Dr. Jun Zou, Ph.D

Corresponding Author's Institution:

First Author: Jun Zou, Ph.D

Order of Authors: Jun Zou, Ph.D; Jianning Sun; Te Li; Xiaomen Han; Zixuan
Xiang; Jie Sha

Abstract: A severe haze-fog episode occurred in the Yangtze River Delta

region in eastern China during 22 - 30 November 2018. In this period, the
PM2.5 mass concentration and meteorological parameters at surface were
collected at the Station for Observing Regional Processes of the Earth
System (SORPES) site in Nanjing. The vertical distribution of PM2.5,
humidity and potential temperature below 500 m were observed
simultaneously by an Unmanned Aerial Vehicle, and the profile of
potential temperature at 1400 LST in each day was also observed by
radiosonde at the same site. These observations allow us to investigate
the interaction among haze, fog and atmospheric boundary layer. In the
former four days, the PM2.5 mass concentration increases, the daytime
maximum planetary boundary layer height (PBLH) decreases, and the air
humidity increases. These are favorable conditions for fog formation. In
the latter five days, fog forms in four days, the PBLHs in the fog days
are much lower than those in the no-fog days, and the PM2.5 mass
concentration increases further. We find that fog induces a potential
temperature jump (i.e., sharp increase of potential temperature) at the
top of fog layer, and this special thermal structure maintains to the end
of fog period, which significantly suppresses the daytime development of
planetary boundary layer after fog dissipation. The fog-induced reduction
of PBLH further increases the PM2.5 mass concentration. We also find that
the wet deposition of fog on PM2.5 is negligibly small. The scavenging
effect of fog on aerosols only acts during fog period. When fog
dissipates, the aerosols come back from the fog droplets to the
atmosphere. Our observations suggest that fog plays an important role in
the enhancement of air pollution in the haze-fog episode.

Research Data Related to this Submission

--------------------------------------------------
Title: Data for: Observed interaction among haze, fog and atmospheric
boundary layer during a haze-fog episode in the Yangtze River Delta
region, eastern China
Repository: Mendeley Data
https://data.mendeley.com/datasets/6bctv57cpk/draft?a=ebd2c04e-3160-4339-
889c-05e7ef72edd3
*Cover Letter

Dear Dr. James J. Schauer,

Please find our manuscript entitled “Observed interaction among haze, fog and atmospheric

boundary layer during a haze-fog episode in the Yangtze River Delta region, eastern China”,

which we would like to submit for publication as a Research Paper in Atmospheric Environment.

This work has not been submitted elsewhere for publication, in whole or in part, and all the

authors listed have approved the manuscript that is enclosed.

This study focuses on the interaction among haze, fog and atmospheric boundary layer during

a haze-fog episode based on observations at the SORPES site in Nanjing, which is located in

northeast part of YRD region. We find that fog induces a potential temperature jump (i.e., sharp

increase of potential temperature) at the top of fog layer, and this special thermal structure

maintains to the end of fog period, which significantly suppresses the daytime development of

planetary boundary layer after fog dissipation. The fog-induced reduction of PBLH further

increases the PM2.5 mass concentration. We also find that the wet deposition of fog on PM2.5 is

negligibly small. The scavenging effect of fog on aerosols only acts during fog period. When fog

dissipates, the aerosols come back from the fog droplets to the atmosphere.

Our results suggest that fog plays an important role in the enhancement of air pollution in the

haze-fog episode. This study is of interest to researchers in the issue of interaction among haze,

fog and atmospheric boundary layer. This study gives an observational result that fog plays an

important role in the formation of heavy air pollution in moist areas. Therefore, we feel this

manuscript is particularly suitable for Atmospheric Environment and of great interest to its readers.

We deeply appreciate your consideration of our manuscript, and we look forward to receiving

comments from the reviewers.

Best regards

Jun Zou, Jianning Sun, Te Li, Xiaomen Han, Zixuan Xiang, and Jie Sha
*Declaration of Interest Statement

Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:
FIG1.TIFF
Click here to download high resolution image
FIG2.TIFF
Click here to download high resolution image
FIG3.TIFF
Click here to download high resolution image
FIG4.TIFF
Click here to download high resolution image
FIG5.TIFF
Click here to download high resolution image
*Manuscript
Click here to download Manuscript: manuscript.docx Click here to view linked References

1 Observed interaction among haze, fog and atmospheric boundary

2 layer during a haze-fog episode in the Yangtze River Delta region,
3 eastern China
4

5 Jun Zoua,b,c,*, Jianning Suna,b,c,**, Te Lia,b, Xiaomen Hana,b, Zixuan Xianga,b, Jie Shaa,b
a
6 Joint International Research Laboratory of Atmospheric and Earth System Sciences, Nanjing
7 University, Nanjing, China
b
8 School of Atmospheric Sciences, Nanjing University, Nanjing, China
c
9 Collaborative Innovation Center of Climate Change, Jiangsu Province, China
10
11 *Corresponding author
12 Tel.: +86 18551752048; fax: +86 02589681162; E-mail address: junzou@nju.edu.cn (J. Zou)
13 Postal address: No. 163 Xianlin Avenue, Qixia District, Nanjing 210023, China.
14 ** Corresponding author
15 Tel.: +86 13851723266; fax: +86 02589681162; E-mail address: jnsun@nju.edu.cn (J. Sun)
16 Postal address: No. 163 Xianlin Avenue, Qixia District, Nanjing 210023, China.
17
18 Highlights:
19  Potential temperature jump at fog-top suppresses PBLH after fog dissipation.
20  Fog-induced reduction of PBLH increases the PM2.5 mass concentration.
21  The scavenging effect of fog on aerosols only acts during fog period.
22

23 Abstract: A severe haze-fog episode occurred in the Yangtze River Delta region in eastern China

24 during 22 – 30 November 2018. In this period, the PM2.5 mass concentration and meteorological

25 parameters at surface were collected at the Station for Observing Regional Processes of the Earth

26 System (SORPES) site in Nanjing. The vertical distribution of PM2.5, humidity and potential

27 temperature below 500 m were observed simultaneously by an Unmanned Aerial Vehicle, and the

28 profile of potential temperature at 1400 LST in each day was also observed by radiosonde at the

29 same site. These observations allow us to investigate the interaction among haze, fog and

30 atmospheric boundary layer. In the former four days, the PM2.5 mass concentration increases, the

31 daytime maximum planetary boundary layer height (PBLH) decreases, and the air humidity

32 increases. These are favorable conditions for fog formation. In the latter five days, fog forms in

33 four days, the PBLHs in the fog days are much lower than those in the no-fog days, and the PM2.5

34 mass concentration increases further. We find that fog induces a potential temperature jump (i.e.,
35 sharp increase of potential temperature) at the top of fog layer, and this special thermal structure

36 maintains to the end of fog period, which significantly suppresses the daytime development of

37 planetary boundary layer after fog dissipation. The fog-induced reduction of PBLH further

38 increases the PM2.5 mass concentration. We also find that the wet deposition of fog on PM2.5 is

39 negligibly small. The scavenging effect of fog on aerosols only acts during fog period. When fog

40 dissipates, the aerosols come back from the fog droplets to the atmosphere. Our observations

41 suggest that fog plays an important role in the enhancement of air pollution in the haze-fog

42 episode.

43 Key words: haze-fog episode, interaction, potential temperature jump, planetary boundary layer

44 height, PM2.5, vertical distribution

45

46 1. Introduction

47 Because of dramatically increasing energy consumption associated with rapid urbanization

48 and industrialization, many regions of China, especially megacities and city clusters in the coastal

49 eastern China, have experienced heavy air pollution (He et al., 2002; Chen et al., 2008; Tie and

50 Cao, 2009; Ding et al., 2016a). In these regions the fine particular matter, PM2.5 (particles of 2.5

51 microns or less in aerodynamic diameter in the ambient air), often show extremely high

52 concentrations during the winter haze period (Ding et al. 2013a; Huang et al., 2014; Guo et al.,

53 2014; R. Zhang, 2014). Heavy air pollution can significantly modify the weather by changing air

54 temperature, cloud and even precipitation (Ding et al., 2013b, Huang et al., 2016). The increase in

55 atmospheric aerosols over China is shown to heat the atmosphere and generates a cyclonic

56 circulation anomaly that weakens the East Asian winter monsoon (F. Niu et al., 2010). The

57 weakened East Asian winter monsoon circulation brings less cold and dry air to the region, and

58 favor the formation of haze and fog over eastern China (F. Niu et al., 2010; R. H. Zhang et al.,

59 2014).

60 The high PM2.5 levels induce a decrease of surface air temperature by shadowing effect of

61 aerosols and an increase of air temperature in the upper planetary boundary layer (PBL) by

62 heating effect of absorbing aerosols, leading to a stabilized PBL. The enhanced PBL stability

63 suppresses the vertical mixing and dispersion of air pollutants, resulting in higher PM2.5

64 concentrations (Wang et al., 2014; Gao et al., 2015). Actually, the scattering and absorbing effects
65 of aerosols can reduce surface heat flux, leading to the decrease of PBL height (PBLH) and

66 consequently the increase of PM2.5 concentrations (Zou et al., 2017). Ding et al. (2016b)

67 demonstrate that black carbon (BC) aerosols induce heating in the PBL, particularly in the upper

68 PBL, and the resulting stabilized stratification and decreased surface heat flux substantially

69 depress the development of PBL and consequently enhance the occurrences of extreme haze

70 pollution episodes. They define this process as “dome effect” of BC and suggest an urgent need

71 for reducing BC emissions as an efficient way to mitigate the extreme haze pollution in megacities

72 of China.

73 Owing to heterogeneous nucleation, aerosols can promote fog formation (Sachweh and

74 Koepke, 1995; Mohan and Payra, 2006, 2009) to such an extent that supersaturation or saturation

75 is not even needed for fog to form in the polluted air (Kumala et al., 1997). Therefore in moist

76 areas heavy air pollution often induces more fog events. Observations show that the frequencies of

77 fog events in wintertime over eastern-central China have been doubled over the past three decades

78 (F. Niu et al., 2010). Yangtze River Delta (YRD) is one of the regions of high fog occurrence in

79 China (S. Niu et al., 2010). In the YRD region, a common type of fog is advection fog, which

80 forms when quite warm, moist air from East China Sea comes across this region (Shi et al., 2010;

81 Lin et al., 2017). Ding et al., (2019) find that BC plays an important role in advection-radiation

82 fog formation in the YRD region by inducing a mesoscale cyclonic wind anomaly, which can

83 transport moister air from sea into this region and help to form heavy fog at night.

84 The fog water can provide a medium for aqueous phase reactions and transform primary

85 pollutants into secondary aerosols (Graedel et al., 1985; Dall'Osto et al., 2009). Wet scavenging

86 effect of fog can also significantly reduce aerosol concentration in the atmosphere (Yuskiewicz et

87 al., 1998; S. Niu et al., 2010; Han et al., 2018). The interrelation between fog and aerosols is

88 crucial in the understanding of the formation and evolution of both haze and fog episodes, which

89 are associated with the PBL processes (Ding et al., 2019). However, the interaction among haze,

90 fog and PBL has been rarely reported in previous studies. During 22 – 30 November 2018, the

91 YRD region suffered from a severe haze-fog episode with high PM2.5 mass concentration and

92 multi-day fog occurrence. In this study, we focus on the interaction among haze, fog and PBL

93 processes based on observations.

94
 95 2. Observation and data processing

96 During 22 – 30 November 2018, a severe haze-fog event occurred in the YRD region. PM2.5

97 mass concentration, surface heat flux, and meteorological data (including wind, temperature,

98 humidity and precipitation) in this period that were collected at the Station for Observing Regional

99 Processes of the Earth System (SORPES) site (118º57′10″E, 32º7′14″N) in Nanjing (Ding et al.,

100 2013a, b; Zou et al., 2017) are used in this study. The PM2.5 mass concentration was measured by

101 a mass analyzer (SHARP-5030, Thermo Fisher Scientific, Massachusetts, USA), which was

102 housed on the top floor of a laboratory building. The surface sensible and latent heat fluxes were

103 measured by an eddy covariance system (CSAT 3A, Campbell Scientific, Utah, USA), which was

104 mounted in the observation yard at the height of 2.2 m above ground level (AGL). The signals for

105 the eddy covariance calculations were sampled at 10 Hz, and the data were subjected to spike

106 detection and rejection algorithms (Zou et al., 2017). The sensible heat flux (Hs) ware computed

107 at half-hour interval, which is the widely-accepted duration for calculating the turbulent fluxes

108 (Nordbo et al., 2013). Corrections were applied for sonic virtual temperature (Schotanus et al.,

109 1983). The data of wind, temperature and humidity ware measured by sensors installed at the

110 height of 9 m AGL on the 75 m-tower (Actually, there are 6 level measurements on this tower.

111 Here we only use the lowest level data). The wind speed (WS) and direction (WD) ware measured

112 by an anemometer, including a 010C wind speed sensor (Campbell, USA) and a 020C wind

113 direction sensor (Campbell, USA). The temperature (T), relative humidity (RH) and specific

114 humidity (q) ware measured by a HMP155A Temperature and Relative Humidity Probe (Campbell,

115 USA). The precipitation was measured by a TE525WS-L Rain Gage (Campbell, USA) in the

116 observation yard.

117 In order to know the maximum PBLH in each day during this haze-fog event, radiosonde

118 measurements were carried out at 1400 local standard time (LST) at the SORPES site. The iMet-4

119 radiosonde (International Met Systems, USA) was used to measure the temperature, pressure,

120 humidity and GPS information, and the sounding data is received by a 3050A receiver

121 (International Met Systems, USA). The GPS data was used to calculate the wind speed and

122 direction, supposing that the radiosonde moves with the wind. To obtain the vertical profiles of

123 PM2.5 concentration and meteorological parameters, measurements were achieved by an

124 Unmanned Aerial Vehicle (UAV). The UAV is a M600 Pro hexacopter (DJI, China). The
125 maximum flight height is 500 m AGL, and the flight time for going down with the payload is

126 16-18 min (the moving speed is about 1 m s–1). An iMet-4 radiosonde was mounted on the UAV to

127 measure the air temperature and humidity, and an AM520 PM2.5 sensor (TSI, USA) was used to

128 monitoring PM2.5 concentration. The profiles measured in the UAV rising process are sued in this

129 study. The UAV measurement times are listed in Table 1.

130 The PBLH can be derived from the radiosonde data, as well as the UAV observation data

131 (when the PBLH is below 500 m). In this study, when there is no fog, we determine the daytime

132 PBLH as the height at which the vertical gradient of potential temperature has the maximum value

133 (Stull, 1988; Batchvarova and Gryning, 1991), and determine the nighttime PBLH as the height of

134 near-surface inversion layer top (Stull, 1988). However, when there exists fog, we define the

135 PBLH as the height at the top of fog layer, above which an elevated inversion or stable layer can

136 be identified, or the air moisture reduces significantly (Seibert et al., 2000).

Time 11/22 11/23 11/24 11/25 11/26 11/27 11/28 11/29 11/30
0800 × × × × × × × × ×
0830 × × × × × × × ×
0900 × × × × × × × × ×
0930 × × × ×
1000 × × × × × ×
1030 × × × ×
1100
1130
1200 ×
1230 ×
1300 × × ×
1330 × × × × × ×
1400
1430 × × ×
1500 × × × × × ×
1530 × × × × ×
1600 × × × × ×
1630
1700
1730 ×
1800 ×
1830 × × ×
 1900 × × × × × ×
1930 × × × × × × ×
2000 × × × × × × ×

137 Table 1. The UAV measurement time in the fog and haze episode during 22 – 30 November, 2018.

138 The symbol × denotes the beginning time of each measurement.

139

140 3. Results and discussion

141 The time series of wind (WS and WD), humidity (q and RH), temperature (T) and surface

142 sensible heat flux (Hs), PBLH, PM2.5 mass concentration, and precipitation in the haze-fog event

143 during 22 – 30 November 2018 are shown in Figure 1, in which the periods of fog occurrence are

144 shaded. In the air pollution area, the relative humidity with the value of 90% is used to distinguish

145 light fog from haze (Schichtel et al., 2001; Doyle and Dorling, 2002). We use this criterion to

146 separate fog from haze and determine the onset time of fog. In the nine days the WS is often less

147 than 2 m s–1 and rarely reaches 3 m s–1, while the WD covers almost all the orientations and is

148 mostly in northeastern or southeastern. Small wind speed is a favorable condition for the

149 formation of haze-fog episodes. In the former four days there is no fog, and the relative humidity

150 RH appears diurnal variation, showing the lower values in the daytime and higher values in the

151 nighttime, due to the diurnal change of temperature. However, the specific humidity q increases

152 steadily from 4 g kg–1 to about 7 g kg–1 in these four days (and an obvious increasing trend of RH

153 can also be seen), suggesting that the air moisture accumulates in this period. This is a favorable

154 condition for fog formation. In the latter five days fog occurs in each day, although the beginning

155 time and the duration are different. On 26 November, the specific humidity q drops back to 4 g

156 kg–1 (a relatively low level of air moisture). The reason is likely that precipitation occurs in the

157 night before and remove some water vapor from the near-surface atmosphere, although the

158 precipitation is light. However, in the following days the air moisture accumulates again, and thus

159 fog can still occur. The near surface air temperature shows diurnal variation between 6 °C (279 K)

160 and 20 °C (293 K). The minimum temperature at night is not very low. The fog can occur due to

161 relatively high level of air moisture. It can be seen that, when the air moisture is not very high, fog

162 occurs later, because a longer radiative cooling time is needed so that a lower temperature can

163 result in a nearly saturated air moisture. For example, fog occurs at about 0800 LST on 28
164 November, the time of sunrise at which the near-surface air temperature reaches the minimum

165 value. The maximum PBLHs at 1400 LST in the former four days are systematically higher than

166 those in the latter five fog days. It can also be seen that in the former four days the PBLH has a

167 decreasing trend while PM2.5 mass concentration increases from 50 μg m–3 to approximately 150

168 μg m–3. Although PM2.5 mass concentration drops back to about 50 μg m–3 on 26 November due to

169 wet deposition by the light precipitation in the night before, it increases rapidly in the following

170 days to more than 200 μg m–3, suggesting that the lower PBLH is a favorable condition for

171 accumulation of air pollutants, as well as air humidity, in the lower atmosphere.

172
173 Figure 1. The time series of (a) wind speed (WS) and wind direction (WD), (b) specific humidity
174 (q) and relative humidity (RH), (c) temperature (T) and surface sensible heat flux (Hs), (d) PBLH,
175 and (e) PM2.5 mass concentration and precipitation in the haze-fog event during 22 – 30 November
176 2018.
177

178 On 29 November there is no fog before sunrise. Therefore the daytime PBLH develops

179 immediately after sunrise, which is similar to the situation in the former four no-fog days. Here we

180 attribute 29 November to a no-fog day, since in the daytime and the whole night before there is no

181 fog. However, on 26-28 and 30 November, fog lasts after sunrise (the fog just begins at sunrise on

182 28 November), and it is inevitable that the development of daytime PBLH is influenced by the fog

183 in these days. We call these four days the fog days. Figure 2 shows the maximum PBLH (at 1400

184 LST) in each day during 22-30 November, which is plotted against the daytime mean PM2.5 mass

185 concentration. The maximum PBLHs in the no-fog days are significantly higher than those in the
186 fog days, suggesting that the development of daytime PBL is suppressed by fog. As shown in

187 Figure 1, the fog lasts to later morning and even afternoon, which must delay the development of

188 daytime PBL and consequently reduce the maximum PBLH. The decrease of PBLH is a favorable

189 condition for air pollutants to accumulate in the PBL, which can lead to heavy air pollution. From

190 27 to 30 November the PM2.5 mass concentration experiences an accumulating process and

191 increases from 50 μg m–3 to more than 200 μg m–3 at the end of this haze-fog episode (Fig. 1e).

192 Figure 2 also shows that the maximum PBLH decreases with PM2.5 mass concentration at the

193 surface in both no-fog and fog days. In the no-fog days, the shadowing effect of aerosols reduces

194 the surface sensible heat flux, and consequently deceases the PBLH (Barbaro et al., 2013; Zou et

195 al., 2017). Our observed results of the daytime maximum PBLH are consistent with the previous

196 studies. However, the situation in the fog days may be more complex. The development of PBL

197 forced by surface heating begins after fog dissipation. On the one hand, the maximum PBLH is

198 influenced by the dissipation time of fog. The later the fog dissipates in the morning, the later the

199 PBL develops, and thus the lower the maximum PBLH can reach (of course, if the fog can last to

200 the afternoon, as on 26 November, the daytime PBL cannot develop, and the PBLH is almost as

201 the same as the depth of the fog layer). On the other hand, it can be expected that the shadowing

202 effect still acts in the delayed developing process of convective PBL, which may further decrease

203 the maximum PBLH. Our observed results in the fog days show a decreasing trend of the

204 maximum PBLH with increasing PM2.5 mass concentration, which seems to support this

205 argument.
206
207 Figure 2. The maximum PBLHs at 1400 LST in fog days (black dots) and no-fog days (circles)
208 during 22-30 November 2018.
209

210 When neglecting large-scale synoptic forcing, the daytime PBL grows mainly driven by

211 surface heating and upper-PBL entrainment (Stull, 1988; Sühring et al., 2014). In a simplified

212 mixed-layer model, the PBLH can be predicted when the surface heat flux and the profile of

213 potential temperature are known (Stull, 1988; van Zanten et al., 1999; Fedorovich et al., 2004; Sun,

214 2009). The growth rate of the PBLH can then be characterized by a simplified parameterization

215 from Sun (2009), in which only the surface heat flux and the background stratification are needed.

216 It is expressed as:

217 (1)

218 where is the PBLH, is the entrainment flux ratio (the ratio of the heat flux at the top of

219 convective PBL to that at the surface), which is approximately a constant of 0.2 (Stull, 1976;

220 Deardorff et al., 1980), is the surface kinematic heat flux, is the lapse rate of potential

221 temperature in the free atmosphere above the convective PBL, and is a parameter, which is

222 approximately a constant of 0.9 for a purely buoyancy-driven convective PBL (Deardorff et al.,

223 1980; Fedorovich et al., 2004; Sun, 2009). Equation (1) shows that the growth rate of PBLH is

224 inversely proportional to the lapse rate of potential temperature in the free atmosphere above the

225 convective PBL. Therefore the profile of potential temperature plays an important role in the

226 daytime PBL development.

227 To understand the effects of haze and fog on the thermal structure of BPL, we compare the

228 profiles of potential temperature at 0800 LST and 1400 LST in each day and at the end of fog

229 period during 22-30 November 2018. These profiles are plotted in Figure 3. The upper panels

230 show that in the no-fog days at 0800 LST the near-surface atmosphere becomes more stable when

231 the PM2.5 mass concentration increases to a high level, while the profiles of potential temperature

232 at 1400 LST present a decreasing trend of the daytime maximum PBLH. This situation is more

233 evident on 29 November, a heavy air pollution day, in which the lapse rate of potential

234 temperature in the near-surface layer at 0800 LST is significantly larger, while the PBLH at 1400

235 is significantly lower than those in the previous no-fog days. A larger lapse rate of potential

236 temperature in the near-surface layer at 0800 LST means that after sunrise the PBL will develop in

237 a more stably stratified background atmosphere and the growth rate of PBLH will be suppressed.

238 The result that the near-surface atmosphere at the end of night becomes more stable in a heavier

239 air pollution day is consistent with the long term observations presented in Zou et al. (2017).

240 Therefore air pollution can impact the PBLH by different processes. One the one hand, high level

241 of aerosol loading in the PBL is apt to form a more stable nocturnal PBL, which will suppress the

242 growth rate of PBLH in the morning. On the other hand, high PM2.5 mass concentration increases

243 the shadowing effect of aerosols on solar radiation and reduces the surface sensible heat flux,

244 which will lead to a lower daytime PBLH. Actually, the heating effect of aerosols also influence

245 the development of daytime PBL. The absorbing aerosols induce solar heating in the atmosphere,

246 and the heating effect is stronger in the upper part of PBL (Barbaro et al., 2013; Ding et al.,

247 2016b). To our knowledge for convective PBL, the stabilized upper part of PBL can suppress the

248 overshooting of thermals and weaken the entrainment process, which means that the entrainment

249 flux ratio is reduced. According to Eq. (1), a smaller leads to a smaller growth rate of

250 PBLH.
251
252 Figure 3. Profiles of potential temperature at 0800 LST (blue curves), 1400 LST (red curves) and
253 the moment of fog dissipation (black curves) in the fog days (lower panels) and the no fog days
254 (upper panels) during 22-30 November. The fog dissipates at 1513 LST on 26 Nov., 1214 LST on
255 27 Nov., 1048 LST on 28 Nov., and 1027 LST on 30 Nov., respectively. The mean mass
256 concentration of PM2.5 in each day is given below each panel.
257

258 The lower panels in Figure 3 show the profiles of potential temperature at different time in

259 the fog days. 26 November is somewhat special, because the fog lasts to afternoon and dissipates

260 at about 1530 LST in this day. When the fog dissipates, the profile of potential temperature (the

261 black line) shows that the PBLH is about 350 m. At 1400 LST, the profile of potential temperature

262 from radiosonde (the red line) indicates that the PBLH is 400 m (here we choose the middle height

263 of the lowest inversion layer as the daytime PBLH, when the profile of potential temperature

264 appears more than one inversion layer). Figure 1 shows the time series of PBLH in this day. The

265 PBLH is actually just the depth of fog layer during the fog period. Since the fog dissipates in the

266 later afternoon, the PBLH decreases slightly in the following afternoon time. Therefore the fog

267 layer depth determines the daytime PBLH in this day. On 27 November, the profile at 0800 LST

268 (the blue line) shows a large potential temperature jump at the height of about 150 m, where is the

269 top of fog layer. At about 1200 LST the fog dissipates, but the potential temperature jump at this
270 height still exists. The fog layer depth is shallow and potential temperature jump becomes smaller

271 at this moment. Additionally, the surface sensible heat flux increases rapidly to about 200 W m–3

272 after fog dissipation (Figure 1). Thus in the following time the surface heating is relatively strong,

273 and the potential temperature jump is easy to be consumed. The potential temperature profile (the

274 red line) shows that at 1400 LST the PBLH reaches a relatively high altitude of about 800 m. The

275 similar situations appear on 28 and 30 November. In the two days, at 0800 LST there is a strong

276 capping inversion layer above the fog layer, and at the time of fog dissipation a step-like potential

277 temperature jump still exists at the top of fog layer. Since the fog layer depth (about 350 m) in

278 these two days is thicker and the potential temperature jump at the moment of fog dissipation is

279 somewhat larger than that on 27 November, surface heating after fog dissipation is to increase air

280 temperature and maintain turbulent motion in the whole mixed layer, but the BPLH is not changed

281 until the potential temperature jump is consumed. This situation is similar to that in many

282 large-eddy simulation studies of convective PBL, in which a large potential temperature jump is

283 set at a chosen height so that the top of mixed layer is approximately fixed at the chosen height

284 during the simulation. To our knowledge, Eq. (1) is not suitable for this situation, because the

285 potential temperature jump cuts off the interaction between the mixed layer and the free

286 atmosphere. Here we distinguish the potential temperature jump from the potential temperature

287 difference across the entrainment zone for a convective PBL that is developing in the equilibrium

288 state. The former corresponds to a step-like shape of potential temperature profile, however, the

289 latter corresponds to a smooth curve of potential temperature profile. Therefore Eq. (1) is only

290 applicable when the convective PBL reaches an equilibrium state, in which the interaction

291 between the mixed layer and the free atmosphere forms a stable interfacial layer, i.e., the

292 entrainment zone, or the so-called capping inversion layer.

293 The potential temperature profile at 1400 LST on 28 November shows that the step-like

294 shape disappears but the PBLH almost does not change, suggesting that the potential temperature

295 jump is consumed just before this moment. On 30 November, the potential temperature profile at

296 1400 LST shows that the PBLH increases only about 150 m from the moment of fog dissipation,

297 implying that most time between 1030 LST and 1400 LST is used to consume the potential

298 temperature jump. Therefore our observations indicate that the potential temperature jump formed

299 at the fog layer top will delay the PBL development after fog dissipation, because a time period is
300 needed for increasing the air temperature in the whole mixed layer by surface heating so that the

301 potential temperature jump is eliminated, and then the PBL begins to develop. Why can a potential

302 temperature jump form at the fog layer top? The most possible reason is that longwave radiative

303 cooling occurs at fog layer top generates well-mixed fog with a well-defined top edge and this

304 sharp top concentrates the radiative divergence closer to that region, which reinforces vertical

305 mixing in the fog layer (Stull, 1988). Therefore a well-mixed fog appears more uniform in the

306 vertical direction and has a well-defined top edge, which forms a sharp potential temperature jump

307 at the fog layer top. Our observations show that the step-like jump corresponds to 3-5 K

308 temperature increase over a thin layer with its depth only about 20 m. Such a large lapse rate

309 cannot be resolved by the weather and climate models, and the growth rate of PBLH after fog

310 dissipation may be overestimated.

311
312 Figure 4. The mean lapse rate of potential temperature at the fog layer top in each temperature
313 profile measured by UAV. The shaded areas represent the daytime duration of fog. The green stars
314 denote the lapse rate in the layer with a depth of 20 m, while the blue stars denote the lapse rate in
315 the layer with a depth of 100 m.
316

317 In order to know the exact strength of potential temperature jump at the fog layer top, we

318 calculate the lapse rate in the thin layer with a depth of 20 m from each temperature profile

319 measured by UAV. The results are shown in Figure 4 and denoted by green starts. On 26

320 November the potential temperature jump is relatively small (as shown in Fig. 3), but the value of

321 lapse rate at the fog layer top is larger than 50 K km–1 and can reach 100 K km–1. In the following
322 three fog days, the potential temperature jump is relatively strong, and the lapse rate at the fog

323 layer top is usually larger than 100 K km–1 and can even reach 250 K km–1. Observations show

324 that a very strong stable layer is formed in a thin layer at the fog top, although the lapse rate in this

325 layer varies with time and the mean strength is different in different case. For the high resolution

326 numerical models, the vertical grid distance is about 100 m. To know the strength of stratification

327 in the layer with a depth of 100 m at the fog layer top, we also calculate the mean lapse rate of

328 potential temperature in this layer and plot the results in Figure 4 as blue starts. On 26 November

329 the lapse rate is about 25 K km–1, while in the following three fog days the lapse rate is about 50 K

330 km–1. Such a large lapse rate maintains to the end of fog period. It means that the convective PBL

331 will develop in a much stably stratified background atmosphere after fog dissipation and the

332 growth rate of PBLH is significantly suppressed. In this situation, the overshooting of rising

333 thermals from mixed layer is inhibited by the strongly stabilized layer above the mixed layer top,

334 which eliminates the entrainment process. Therefore we can use the encroachment model (Boers

335 et al. 1984) to estimate PBLH growth rate. It reads , where is the lapse

336 rate of potential temperature immediately above the mixed layer. This model neglects the

337 entrainment process, implying that the mixed layer top erodes into the strongly stable layer due to

338 the increase of air temperature in the mixed layer. Assuming that K km–1 (considering the

339 situation in the later three fog days), K m s–1 (the corresponding the surface heat

340 flux is 180 W m–2, and this value is estimated according to our measurements at the noontime),

341 and a initial PBLH of 350 m at 1030 LST (according to observed fog layer depth), calculation

342 shows that it needs about 3.7 h for the PBLH to increases to 450 m. This estimation agrees with

343 the observations in the latter two fog days. Our observations on 28 and 30 November indicate that

344 it spends about 3.5 h for the mixed layer top to penetrate this stable layer. Therefore fog can

345 significantly suppress the daytime PBL development by inducing a strongly stabilized layer at the

346 top of fog layer.

347
348 Figure 5. The vertical profile of relative humidity (left) and PM2.5 mass concentration (right) in
349 the fog days. The different colors represents different measurement time. The fog dissipates at
350 1513 LST on 26 Nov., 1214 LST on 27 Nov., 1048 LST on 28 Nov., and 1027 LST on 30 Nov.,
351 respectively.
352

353 Fog droplets interact with aerosol particles and soluble gases in the atmosphere. Thus fogs

354 can affect pollutant formation, transformation, and removal. In this study we observed the vertical
355 distribution of PM2.5 mass concentration in the lower atmosphere both during fog period and after

356 fog dissipation. Figure 5 shows the profiles of PM2.5 mass concentration and relative humidity at

357 different time in the fog days. On 26 November, the profiles of relative humidity at 0830 LST and

358 1000 LST show that the fog layer top reaches about 300 m AGL. Meanwhile, the profiles of PM2.5

359 mass concentration at the two moments show that in the upper part of fog layer the PM2.5 mass

360 concentration in the atmosphere is very small (only about 25 μg m–3), although it increases when

361 approaching to the ground level. However, at 1430 LST, the profiles show that in the fog layer the

362 PM2.5 mass concentration increases to about 100 μg m–3, while the relative humidity in this layer

363 decreases to about 90%, suggesting that at this moment the fog still exists, but evaporation of fog

364 droplets releases a lot of PM2.5 back to the atmosphere. At 1530 LST, the fog dissipates and

365 evaporation ends at this moment, and the PM2.5 mass concentration further increases to about 120

366 μg m–3 that is almost the same level before fog formation, implying that almost all the PM2.5 in the

367 fog droplets comes back to the atmosphere. Figure 5 show that the same process repeats in the

368 following fog days. We note that on 27 November at the moment of fog dissipation the PM2.5 mass

369 concentration reduces to a lower level in comparison with the value before fog formation, which is

370 evidently caused by the light precipitation during fog period (Figure 1). However, in the

371 no-precipitation fog days the PM2.5 mass concentration at the moment of fog dissipation comes

372 back to the level before fog formation. The scavenging effect of fog can significantly reduce

373 aerosol concentration in the atmosphere (Yuskiewicz et al., 1998; S. Niu et al., 2010; Han et al.,

374 2018). Our observations suggest that the wet deposition effect of fog on PM2.5 is negligibly small

375 unless precipitation can occur during fog period. It means that wet scavenging effect of fog on

376 aerosols can only act during fog period. Once the fog dissipates, the aerosols will come back from

377 the fog droplets to the atmosphere. The fog droplets provide a liquid phase reaction environment

378 for the PM2.5 with different chemical compositions and the soluble gases. But this problem is

379 beyond the scope of this study. We will investigate this problem based on observations in our

380 future research.

381

382 4. Concluding remarks

383 A severe haze-fog event occurred in the YRD region, eastern China during 22 – 30

384 November 2018. In this study we focuses on the interaction among haze, fog and atmospheric
385 boundary layer during the haze-fog episode based on observations at the SORPES site in Nanjing,

386 which is located in northeast part of YRD region. Observations show that during this period the

387 PM2.5 mass concentration increases from about 50 μg m–3 to more than 200 μg m–3 while the

388 daytime maximum PBLH decreases from about 1500 m to 500 m and even lower height. During

389 the former half period, i.e., in the clear-sky days, PM2.5 accumulates due to small wind speed, and

390 the air humidity increases to a high level. During the latter half period, fog occurs almost every

391 day, the PM2.5 mass concentration increases, and the daytime maximum PBLHs in the fog days are

392 significantly lower than those in the clear-sky days. These observational results suggest that the

393 interaction among haze, fog and atmospheric boundary layer enhances air pollution. On the one

394 hand, the increase of aerosol loading reduces PBLH by decreasing the daytime surface heating and

395 stabilizing the atmospheric boundary layer, and the aerosol-induced reduction of PBLH limits the

396 aerosols in a smaller space and results in higher PM2.5 mass concentration. This is a well-known

397 interaction process in the clear-sky days. On the other hand, the increase of aerosol loading and air

398 humidity provides favorable conditions for fog formation, and the daytime PBLH is significantly

399 reduced after fog dissipation, which can further increase PM2.5 mass concentration and lead to

400 heavy air pollution.

401 Observations show that the daytime maximum PBLHs in the fog days are much lower than

402 those in the no-fog days, suggesting that the occurrence of fog significantly suppresses the

403 development of daytime PBL. One the one hand, the convective PBL develops after sunrise in

404 clear-sky days, whereas it develops after fog dissipation in the fog days. Therefore fog delays the

405 development of convective PBL. On the other hand, the existence of fog results in a step-like

406 potential temperature jump at the top of fog layer by longwave radiative cooling and remains it to

407 the moment of fog dissipation. Thereafter the convective PBL will firstly spends a few hours on

408 increasing the air temperature in the mixed layer by surface heating to eliminate the step-like

409 potential temperature jump, and then it can develop freely in the equilibrium state. This process

410 further delays the development of convective PBL, which leads to a much lower daytime PBLH

411 and consequently a high level of PM2.5 mass concentration in the fog days.

412 Observations also show that the wet deposition effect of fog on PM2.5 is negligibly small if

413 the fog cannot form precipitation. Previous studies point out that the scavenging effect of fog can

414 significantly reduce aerosol concentration in the atmosphere. Our observations suggest that this
415 effect only acts during the fog period. When the fog dissipates, the aerosols will come back from

416 the fog droplets to the atmosphere. Along with the effect of fog on the thermal structure of PBL,

417 which can significantly reduce PBLH, fog plays an important role in the formation of heavy air

418 pollution in moist areas, such as the YRD region, eastern China.

419

420 Acknowledgements The work is supported by Ministry of Science and Technology of China

421 (2016YFC0200500) and National Natural Science Foundation of China (91544231).

422

423 References

424 Barbaro, E., J. de Arellano, V.-G., Krol, M. C., Holtslag, A. A. M., 2013. Impacts of aerosol

425 shortwave radiation absorption on the dynamics of an idealized convective atmospheric

426 boundary layer. Bound. Layer Meteorol. 148, 31-49. doi: 10.1007/s10546-013-9800-7.

427 Batchavarova, E., Gryning, S. E., 1991. Applied model for the growth of the daytime mixed layer.

428 Bound. Layer Meteorol. 56, 261-274. doi: 10.1007/bf00120423

429 Boers, R., Eloranta, E. W., Coulter, R. L., 1984. Lidar observations of mixed layer dynamics: Tests

430 of parameterized entrainment models of mixed layer growth rate. J. Climate Appl. Meteor. 23,

431 247-266. doi: 10.1175/1520-0450(1984)023<0247:LOOMLD>2.0.CO;2

432 Dall'Osto, M., Harrison, R. M., Coe, H., Williams, P., 2009. Real-time secondary aerosol

433 formation during a fog event in London. Atmos. Chem. Phys. 9, 2459-2469. doi:

434 10.5194/acp-9-2459-2009.

435 Deardorff, J. W., Willis, G. E., Stockton, B. H., 1980. Laboratory studies of the entrainment zone

436 of a convectively mixed layer. J. Fluid. Mech. 100, 41-64. doi:

437 10.1017/S0022112080001000.

438 Ding. A. J., Fu, C. B., Yang, X. Q., Sun, J. N., Zheng, L. F., Xie, Y. N., Herrmann, E., Nie, W.,

439 Petäjä, T., Kerminen, V. M., Kulmala, M., 2013a. Ozone and fine particle in the western

440 Yangtze River Delta: an overview of 1 yr data at the SORPES station. Atmos. Chem. Phys.

441 13, 5813–5830. doi: 10.5194/acp-13-5813-2013.

442 Ding, A. J., Fu, C. B., Yang, X. Q., Sun, J. N., Petäjä, T., Kerminen, V. M., Wang, T., Xie, Y.,

443 Herrmann, E., Zheng, L. F., Nie, W., Liu, Q., Wei, X. L., and Kulmala, M., 2013b. Intense

444 atmospheric pollution modifies weather: a case of mixed biomass burning with fossil fuel
445 combustion pollution in eastern China. Atmos. Chem. Phys. 13, 10545-10554. doi:

446 10.5194/acp-13-10545-2013.

447 Ding, A,, Nie, W., Huang, X., Chi, X., Sun, J., Kerminen, V. M., Xu, Z., Guo, W., Petäjä, T., Yang,

448 X., Kulmala, M., Fu, C., 2016a. Long-term observation of air pollution-weather/climate

449 interactions at the SORPES station: a review and outlook. Front. Environ. Sci. Eng. 10, 15.

450 doi: 10.1007/s11783-016-0877-3.

451 Ding, A. J., Huang, X., Nie, W., Sun, J. N., Kerminen, V. M., Petäjä, T., Su, H., Cheng, Y. F., Yang,

452 X. Q., Wang, M. H., Chi, X. G., Wang, J. P., Virkkula, A., Guo, W. D., Yuan, J., Wang, S. Y.,

453 Zhang, R. J., Wu, Y. F., Song, Y., Zhu, T., Zilitinkevich, S., Kulmala, M., Fu, C. B., 2016b.

454 Enhanced haze pollution by black carbon in megacities in China. Geophys. Res. Lett. 43,

455 2873–2879. doi: 10.1002/2016GL067745.

456 Ding, Q. J., Sun, J. N., Huang, X., Ding, A. J., Zou, J., Yang, X.-Q., and Fu, C. B., 2019. Impacts

457 of black carbon on the formation of advection-radiation fog during a haze pollution episode

458 in eastern China. Atmos. Chem. Phys. 19, 7759–7774. doi: 10.5194/acp-19-7759-2019.

459 Doyle, M., Dorling, S., 2001. Visibility trends in the UK 1950–1997. Atmos. Environ. 36,

460 3161–3172. doi: 10.1016/S1352-2310(02)00248-0.

461 Fedorovich, E., Conzemius, R., Mironov, D., 2004. Convective entrainment into a shear-Free,

462 linearly stratified atmosphere: Bulk models reevaluated through large eddy simulations. J.

463 Atmos. Sci., 61, 281-295. doi: 10.1175/1520-0469(2004)061<0281:CEIASL>2.0.CO;2.

464 Gao, Y., Zhang, M., Liu, Z., Wang, L., Wang, P., Xia, X., Tao, M., Zhu L., 2015. Modeling the

465 feedback between aerosol and meteorological variables in the atmospheric boundary layer

466 during a severe fog–haze event over the North China Plain. Atmos. Chem. Phys. 15,

467 4279–4295. doi: 10.5194/acp-15-4279-2015.

468 Graedel, T. E., Weschler, C. J., Mandich, M. L., 1985. Influence of transition metal complexes on

469 atmospheric droplet acidity, Nature, 317, 240-242. doi: 10.1038/317240a0.

470 Guo, S., Hu, M., Zamora, M. L., Peng, J., Shang, D., Zheng, J., Du, Z., Wu, Z., Shao, M., Zeng, L.,

471 Molina, M. J., Zhang, R., 2014. Elucidating severe urban haze formation in China. Proc. Natl.

472 Acad. Sci. U. S. A. 111, 17373–17378. doi: 10.1073/pnas.1419604111.

473 Han, S. Q., Liu, J. L., Hao, T. Y., Zhang, Y. F., Li, P. Y., Yang, J. B., Wang, Q. L., Cai, Z. Y., Yao,

474 Q., Zhang, M., Wang, X. J., 2018: Boundary layer structure and scavenging effect during a
475 typical winter haze-fog episode in a core city of BTH region, China. Atmos. Environ. 179,

476 187–200. doi: 10.1016/j.atmosenv.2018.02.023.

477 Huang, X., Ding, A., Liu, L., Liu, Q., Ding, K., Niu, X., Nie, W., Xu, Z., Chi, X., Wang, M., Sun,

478 J., Guo, W., Fu, C., 2016. Effects of aerosol–radiation interaction on precipitation during

479 biomass-burning season in East China. Atmos. Chem. Phys. 16, 10063-10082. doi:

480 10.5194/acp-16-10063-2016.

481 Huang, R., Zhang, Y., Bozzetti, C., Ho, K. F., Cao, J., Han, Y., Daellenbach, K. R., Slowik, J. G.,

482 Platt, S. M., Canonaco, F., Zotter, P., Wolf, R., Pieber, S. M., Bruns, E. A., Crippa, M.,

483 Ciarelli, G., Piazzalunga, A., Schwikowski, M., Abbaszade, G., Schnelle-Kreis, J.,

484 Zimmermann, R., An, Z., Szidat, S., Baltensperger, U., El Haddad, I., Prévôt, A. S. H., 2014.

485 High secondary aerosol contribution to particulate pollution during haze events in China.

486 Nature. 514, 218–222. doi: 10.1038/nature13774.

487 Lin, C., Zhang, Z., Zhao, P. U., Wang, F., 2017. Numerical simulations of an advection fog event

488 over Shanghai Pudong international air-port with the WRF model. J. Meteorol. Res-PRC. 31,

489 874-889. doi: 10.1007/s13351-017-6187-2.

490 Mohan, M., Payra, S., 2006. Relation between fog formation and accumulation mode aerosols in

491 urban environment of megacity Delhi. Indian Journal of Environmental Protection, 26,

492 294–300.

493 Mohan, M., and Payra, S., 2009. Influence of aerosol spectrum and air pollutants on fog formation

494 in urban environment of megacity Delhi, India. Environ. Monit. Assess. 151, 265–277. doi:

495 10.1007/s10661-008-0268-8.

496 Niu, F., Li, Z., Li, C., Lee, K.-H., Wang, M. 2010. Increase of wintertime fog in China: Potential

497 impacts of weakening of the Eastern Asian monsoon circulation and increasing aerosol

498 loading. J. Geophys. Res. Atmos. 115, D00K20. doi: 10.1029/2009JD013484.

499 Niu, S., Lu, C., Yu, H., Zhao, L., Lu, J., 2010. Fog research in China: An overview. Adv. Atmos.

500 Sci. 27, 639–662, doi: 10.1007/ s00376-009-8174-8.

501 Nordbo, A., Järvi, L., Haapanala, S., Moilanen, J., Vesala, T., 2013. Intra-City Variation in Urban

502 Morphology and Turbulence Structure in Helsinki, Finland. Bound. Layer Meteorol. 146,

503 469-96. doi: 10.1007/s10546-012-9773-y.

504 Sachweh, M., Koepke, P., 1995. Radiation fog and urban climate. Geophys. Res. Lett. 22,
505 1073–1076. doi: 10.1029/95GL00907.

506 Schichtel, B. A., Husar, R. B., Falke, S. R., Wilson, W.E., 2001. Haze trends over the United

507 States 1980–1995. Atmos. Environ. 35, 5205–5210. doi: 10.1016/S1352-2310(01)00317-X.

508 Schotanus, P., Nieuwstadt, F. T. M., Bruin, H. A. R. D., 1983. Temperature measurement with a

509 sonic anemometer and its application to heat and moisture fluxes. Bound. Layer Meteorol. 26,

510 81-93. doi: 10.1007/BF00164332.

511 Seibert, P., Beyrich, F., Gryning, S.-E., Joffre, S., Rasmussen, A., Tercier, P., Review and

512 intercomparison of operational methods for the determination of the mixing height. Atmos.

513 Environ. 34, 1001-1027. doi: 10.1016/S1352-2310(99)00349-0.

514 Shi, C., Yang, J., Qiu, M., Hao, Z., Su, Z., Li, Z., 2010. Analysis of an extremely dense regional

515 fog event in Eastern China using a mesoscale model. Atmos. Res. 95, 428-440. doi:

516 10.1016/j.atmosres.2009.11.006.

517 Stull, R. B., 1976. The energetics of entrainment across a density interface. J. Atmos. Sci. 33,

518 1260-1267.

519 Stull, R. B., 1988. An introduction to boundary layer meteorology. Dordrecht, Kluwer Academic

520 Publishers, 670 pp.

521 Sun, J. N., 2009. On the parameterization of convective entrainment: Inherent relationships among

522 entrainment parameters in bulk models. Adv. Atmos. Sci. 26, 1005-1014. doi:

523 10.1007/s00376-009-7222-8.

524 Sühring, M., Maronga, B., Herbort, F., Raasch, S., 2014. On the effect of surface heat-flux

525 heterogeneities on the mixed-layer-top entrainment. Bound. Layer Meteorol. 151, 531-556.

526 doi: 10.1007/s10546-014-9913-7.

527 van Zanten, M. C., Duynkerke, P. G., Cuijpers, J. W. M., 1999. Entrainment parameterization in

528 convective boundary layers. J. Atmos. Sci. 56, 813-828. doi:

529 10.1175/1520-0469(1999)056<0813:EPICBL>2.0.CO;2.

530 Wang, J., Wang, S., Jiang, J., Ding, A., Mei, Z., Zhao, B., Wong, D. C., Zhou, W., Zheng, G.,

531 Wang, L., 2014. Impact of aerosol–meteorology interactions on fine particle pollution during

532 China’s severe haze episode in January 2013. Environ. Res. Lett. 9, 094002.

533 doi:10.1088/1748-9326/9/9/094002.

534 Yuskiewicz, B. A., Orsini, D., Stratmann, F., Wendisch, M., Wiedensohler, A., Heintzenberg, J.,
535 Martinsson, B. G., Frank, G., Wobrock, W., Schell, D., 1998. Changes in submicrometer

536 particle distributions and light scattering during haze and fog events in a highly polluted

537 environment. Contri. Atmos. Phys. 71, 33-45.

538 Zhang, R, Li, Q., Zhang, R, 2014. Meteorological conditions for the persistent severe fog and haze

539 event over eastern China in January 2013. Sci. China-Earth Sci. 57, 26-35.

540 doi:10.1007/s11430-013-4774-3.

541 Zou, J., Sun, J., Ding, A., Wang, M., Guo, W., Fu, C., 2017. Observation-based estimation of

542 aerosol-induced reduction of planetary boundary layer height. Adv. Atmos. Sci. 34,

543 1057–1068. doi: 10.1007/s00376- 016-6259-8.