To investigate the spatiotemporal variability of the mixing layer
height (MLH) on the North China Plain (NCP), multi-site and long-term
observations of the MLH with ceilometers at three inland stations (Beijing, BJ; Shijiazhuang, SJZ; Tianjin, TJ) and
one coastal site (Qinhuangdao) were conducted from 16 October 2013 to 15 July 2015. The MLH of the inland
stations in the NCP were highest in summer and lowest in winter, while the
MLH on the coastal area of Bohai was lowest in summer and highest in spring.
As a typical site in southern Hebei, the annual mean of the MLH at SJZ was
464
The convective boundary layer is the region where turbulence is fully developed. The height of the interface where turbulence is discontinuous is usually referred to as the mixing layer height (MLH; Stull, 1988). The mixing layer is regarded as the link between the near-surface and free atmosphere, and the MLH is one of the major factors affecting the atmospheric dissipation capacity, which determines both the volume into which ground-emitted pollutants can disperse, as well as the convective timescales within the mixing layer (Seidel et al., 2010). In addition, continuous MLH observations will be of great importance for the improvement of boundary layer parameterization schemes and for the promotion of meteorological model accuracy.
Conventionally, the MLH is usually estimated from radiosonde profiles (Seidel et al., 2010). Although meteorological radiosonde observations can provide high-quality data, they are not suitable for continuous fine-resolution MLH retrievals due to their high cost and limited observation intervals (Seibert et al., 2000). As the most advanced method of MLH detection, remote sensing techniques based on the profile measurements from ground-based instruments such as sodar, radar, or lidar that have unique vertically resolved observational capabilities are becoming increasingly popular (Beyrich, 1997; Chen et al., 2001; He and Mao, 2005). Because sound waves can be easily attenuated in the atmosphere, the vertical range of sodar is generally limited to within 1000 m. However, the optical remote sensing techniques can provide higher height ranges (at least several kilometers). The single-lens ceilometers developed by Vaisala have been widely used in a variety of MLH studies (Emeis et al., 2004, 2009, 2011; Münkel and Räsänen, 2004; Münkel et al., 2007; Eresmaa et al., 2006; Wagner et al., 2006; Wagner and Schäfer, 2015; Muñoz and Undurraga, 2010; Schween et al., 2014; Sokół et al., 2014; Tang et al., 2015, 2016; Geiß et al., 2017). Compared with other remote sensing instruments, this type of lidar has special features favorable for long-term and multi-station observations (Emeis et al., 2009; Wiegner et al., 2014; Tang et al., 2016), including the low-power system, the eye-safe operation within a near-infrared laser band and the low cost and ease of maintenance during all weather conditions (excluding rainy, very windy or sandstorm weather conditions) with only regular window cleaning required (Emeis et al., 2004; Tang et al., 2016).
The North China Plain (NCP) region is the political, economic and cultural center of China. With the rapid economic development, energy use has increased substantially, resulting in frequent air pollution episodes (Guo et al., 2011; Li et al., 2013; Y. S. Wang et al., 2013; Wang et al., 2014; Zhang et al., 2014; Liu et al., 2016; Xu et al., 2016; Tang et al., 2017b). The haze pollution has had an adverse impact on human health (Tang et al., 2017a) and has aroused a great deal of concern (Tang et al., 2009; Ji et al., 2012; H. Zhang et al., 2015). To achieve the integrated development of the Jing–Jin–Ji region, readjustment of the regional industrial structure and layout is imperative. To this end, the industrial capacity of heavily polluting enterprises in the areas with unfavorable weather conditions should be reduced, and these heavily polluting enterprises should be removed to improve the air quality. For the remaining enterprises, the industrial air pollutant emissions structure should be changed, and strong emission reduction measures must be implemented. Although the government has carried out some strategies for joint prevention and control, with the less well-understood distributions of regional weather condition on the NCP, how and where to adjust the industrial structures on the NCP are questions in pressing need of answers. As one of the key factors influencing the regional heavy haze pollution (Tang et al., 2012, 2016, 2017b; Quan et al., 2013; Hu et al., 2014; W. Zhang et al., 2016; Zhu et al., 2016), the MLH to some extent represents the atmospheric environmental capacity, and the regional distribution and variation of MLH on the NCP can offer a scientific basis for regional industrial distribution readjustment, which will be of great importance for regional haze management.
Nevertheless, due to the scarcity of MLH observations on the NCP, reliable and explicit characteristics of MLH on the NCP remain unknown. Tang et al. (2016) utilized the long-term observation data of MLH from ceilometers to analyze the characteristics of MLH variations in Beijing (BJ) and verified the reliability of ceilometers. The results demonstrated that MLH in BJ was high in spring and summer and low in autumn and winter with two transition months in February and September. A multi-station analysis of MLH on the NCP region was conducted in February 2014, and the characteristics of high MLH at coastal stations and low MLH at southwest piedmont stations were reported (Li et al., 2015). Miao et al. (2015) modeled the seasonal variations of MLH on the NCP and discovered that the MLH was high in spring due to the strong mechanical forcing and low in winter as a result of the strong thermodynamic stability in the near-surface layer. The mountain–plain breeze and the sea breeze circulations played an important role in the mixing layer process when the background synoptic patterns were weak in summer and autumn (Tang et al., 2016; Wei et al., 2017).
To overcome previous studies' deficiencies, our study first conducted a 22-month (from 16 October 2013 to 15 July 2015) observation of MLH with ceilometers on the NCP. The observation stations included three inland stations (BJ; Shijiazhuang, SJZ; Tianjin, TJ) and one coastal site (Qinhuangdao, QHD). First, we will describe the spatial and temporal distribution of MLH on the NCP. Subsequently, reasons for spatial difference of MLH on the NCP will be explained in the discussion section. Finally, the meteorological evidence of serious air pollution in southern Hebei will be studied.
To study the MLH characteristics on the NCP, observations with ceilometers
were conducted at the BJ, SJZ, TJ and QHD stations from 16 October 2013 to
15 July 2015 (Fig. 1 and Supplement Table S1). The SJZ, TJ and QHD sites were set
around Beijing in the southwest, southeast and east directions,
respectively. The BJ station was at the base of the Taihang and Yan
mountains on the northern NCP. The MLH observation site was built in the
courtyard of the Institute of Atmospheric Physics, Chinese Academy of
Sciences (116.32
Locations of the ceilometers observation sites (BJ, SJZ, TJ and QHD)
are marked with red and bold abbreviations; other PM
The instrument used to measure the MLH at the four stations was an enhanced
single-lens ceilometer (Vaisala, Finland), which utilized the strobe laser
lidar (laser detection and range measurement) technique (910 nm) to measure
the attenuated backscattering coefficient profiles. As large differences
existed in the aerosol concentrations between the mixing layer and the free
atmosphere, the MLH can be determined from the vertical attenuated
backscattering coefficient (
To ensure the consistency of the MLH measurements with the two different
ceilometer versions, before we set up the ceilometer observation network in
the NCP, we made a comparison of the MLHs observed by CL31 and CL51 at BJ
from 1 to 8 October 2013 (Fig. S1 in the Supplement). The MLH observed by CL31 was
highly relevant to those observed by the CL51 ceilometers, with correlation
coefficients (
Since the ceilometers can reflect rainy conditions and the precipitation
will influence the MLH retrieval, the precipitation data were excluded. In
addition, a previous study has compared MLH measurements retrieved from
ceilometers and sounding data (Tang et al., 2016). The results revealed that
the ceilometers underestimate the MLH under neutral conditions caused by
strong winds and overestimate the MLH when sandstorms occur. Therefore,
data points for these three special weather conditions were eliminated
manually. The criterion to exclude these data points is as follows: (a) precipitation,
i.e., a cloud base lower than 4000 m and the attenuated
backscattering coefficient of at least 2
Instrument properties of CL31 and CL51.
The hourly data of near-ground relative humidity (RH) and temperature (
The near-ground PM
The aerosol optical depth (AOD) data within the NCP region were retrieved
with the dark target algorithm from the Moderate Resolution Imaging
Spectroradiometer aerosol products on board the National Aeronautics and
Space Administration Earth Observing System Terra satellite from December
2013 to November 2014 (Q. Zhang et al., 2016;
The gradient Richardson number (
Frequency distribution of the daily maximum MLH at the
Using
Since October 2013, continuous operation of the ceilometers observation network in the NCP has provided 22 months of MLH data. For the purpose of analyzing the MLH temporal and spatial variation, the hourly averages of MLH for a whole year (from December 2013 to November 2014) at the BJ, SJZ, TJ and QHD stations were chosen in the following sections. Hourly means of MLH under rainy, sandstorm and windy conditions were removed, resulting in data availability of 81, 89, 83 and 77 % at the BJ, SJZ, TJ and QHD stations, respectively. In this study, March, April and May are defined as spring; June, July and August are defined as summer; September, October and November are defined as autumn; and December, January and February are defined as winter.
To study the regional distribution characteristic of MLH on the NCP, we analyzed the frequency of the daily maximum MLH distribution in Fig. 2. The daily maximum MLH at the BJ, SJZ and TJ stations could reach 2400 m. The large daily maximum values mostly existed in spring and summer, while the low values always appeared in autumn and winter and were as low as 200 m. The daily maximum MLH values at the BJ, SJZ and TJ stations were mainly distributed between 600 and 1800, 400 and 1600, and 800 and 1800 m, accounting for 74.2, 72.0 and 67.0 % of the total samples, respectively. Notably, the daily maximum MLH in SJZ was lower than at the MLHs at the BJ and TJ stations in spring, autumn and winter. Values below 600 m at the SJZ station occurred primarily in autumn and winter. The most frequent daily maximum MLH existed in the range of 1000–1200 m, which was 200–600 m lower than that at the TJ station. This demonstrated a weaker atmospheric diffusion capacity at the SJZ station in spring, autumn and winter than the northern NCP stations.
The frequency distribution of the daily maximum MLH at the coastal site showed different features. The daily maximum MLH in QHD was mainly distributed between 800 and 1800 m with a relatively small seasonal fluctuation (Fig. 2d). Values lower than 600 m were mainly distributed in summer, which was probably influenced by the frequent occurrence of a thermal internal boundary layer (TIBL) in summer. Reasons for this are illustrated in Sect. 4.1.
Monthly variations of MLH at the BJ, SJZ, TJ and QHD stations are shown in Fig. 3. The monthly means of the regional MLH ranged between 300 and 750 m. The maximum and minimum MLH existed in June 2014 at the BJ station and in January 2014 at the SJZ station, with values of 741 and 308 m, respectively. Most of the monthly averages were between 400 and 700 m, which accounted for 81.3 % of the total samples.
The MLH at the BJ, SJZ and TJ stations showed obvious seasonal variations
with high values in spring and summer and low values in autumn and winter.
Seasonal means of MLH at the three stations followed the same order:
summer > spring > autumn > winter, with
maximum values of 722
Nevertheless, the seasonal variation of MLH at the coastal site of Bohai was
different from that at the inland stations. The MLH in QHD exhibited a
decreasing trend from spring to summer and an increasing trend from autumn
to winter, with the maximum seasonal mean of 498
Monthly variations of MLH at the BJ, SJZ, TJ and QHD stations from December 2013 to November 2014.
Diurnal variations of MLH at the BJ, SJZ, TJ and QHD stations in
Seasonal variations of diurnal MLH change patterns were investigated to reveal the 24 h evolution characteristics of the MLH on the NCP. As shown in Fig. 4, diurnal variations of MLH in different seasons all had single peak patterns. With sunrise and increased solar radiation, MLH at the four stations started to develop and peaked in the early afternoon. After sunset, turbulence in the MLH decayed quickly, and the mixing layer underwent a transition to the nocturnal stable layer (less than 400 m). The annual averaged diurnal ranges of MLH at the BJ, SJZ, TJ and QHD stations were 782, 699, 914 and 790 m, respectively. The annual averaged diurnal range of MLH in SJZ was approximately 100–200 m smaller than those at the other stations, which was associated with its shallow daytime MLHs in spring, autumn and winter (Fig. 4a, c and d). This indicated that SJZ has the worst pollutant diffusivity.
Growth rates averaged over the four stations during each season were plotted
with gray columns in Fig. 4. It was obvious that the growth rates of the MLH
varied by season. The MLH developed the earliest in summer (at approximately
07:00 LT) and reached the highest growth rates (164.5 m h
Annual averages of MLH at the BJ, SJZ, TJ and QHD stations were also
calculated, and the values were 594
Through preliminary study of the spatiotemporal variation of MLH on the NCP region, we found something interesting: (a) the MLH at the coastal site was lower than the inland sites in summer; (b) the MLH in southern Hebei was lower than the northern NCP in spring, autumn and winter, but was almost consistent between these two areas in summer. Reasons for these two phenomena will be illustrated in the following Section (4.1 and 4.2). Finally, we will investigate the meteorological evidence for serious haze pollution in southern Hebei in Sect. 4.3.
Monthly diurnal wind vectors at the BJ, SJZ, TJ and QHD stations from December 2013 to November 2014.
From the studies in Sect. 3.1 and 3.2, we found that the maximum MLH at the QHD station was larger and arrived earlier than the BJ, SJZ and TJ stations in summer (Fig. 4b). However, this characteristic was not evident in other seasons (Fig. 4a, c and d). The sea–land breeze was a local circulation that occurs when there is no large-scale synoptic system passes. In our study, we first excluded days with large-scale synoptic systems. Then, according to the coastline orientation, if the southeast wind at the TJ station and south and southwest winds at the QHD station occurred at approximately 11:00 LT, and the northwest wind started to blow at approximately 20:00 LT, and then this type of circulation was supposed to be a sea–land circulation. The prevailing southeast wind at the TJ station and the south and southwest wind at the QHD station were regarded as sea breezes (Fig. 5).
The sea breeze usually brings a cold and stable air mass from the sea to the coastal region. When the top of the local mixing layer was higher than the top of the air mass, a TIBL will develop within the mixing layer under the influence of the abrupt change of aerodynamic roughness and temperature between the land and sea surfaces. Then, the local mixing layer will be replaced by the TIBL. In the presence of warm air on land, the cold sea air advects downwind and is warmed, leading to a weak temperature difference between the air and the ground. As a consequence, the TIBL warms less rapidly due to the decreased heat flux at the ground, and the rise rate is reduced. In addition, since the TIBL deepens with distance downwind and usually cannot extend all the way to the top of the intruding marine air, the remaining cool marine air above the TIBL will hinder vertical development of the TIBL (Stull, 1988; Puygrenier et al., 2005; Sicard et al., 2006; Tomasi et al., 2011). With distance inland, the top of the intruding marine air will enhance and exceed the local MLH; if so, the TIBL will not form, and the TIBL impact will be impaired with distance inland (Stull, 1988). Accompanied by the weak synoptic system and the frequent occurrence of sea breezes in summer, the TIBL formed easily and the MLH peak time and value at the QHD station were earlier and lower than other stations (Figs. 3 and 4). For the TJ station, with a distance of approximately 50 km out to sea, the TIBL will not extend so far. Therefore, although the TJ station can be affected by the sea breeze, the local MLH cannot be influenced by the TIBL.
Turbulent stability was mainly responsible for the MLH development, and the
generation of turbulent energy was highly correlated with the heat flux
(mainly sensible heat fluxes) produced by radiation and the momentum flux
caused by wind shear (Stull, 1988). As presented in Sect. 2.4, the
Vertical profiles of
Using the winter and summer as examples, when we analyzed the seasonal means
of shear term and the buoyancy term between the XT and the BJ stations, some
distinct features were observed. As shown in Fig. 6f and g, the shear term
and the buoyancy term in XT were 2.8 times lower and 1.5 times higher than
that in BJ within 0–1200 m in winter, respectively. The largest
discrepancies of the wind shear term and buoyancy term between southern
Hebei and the northern NCP reached 2.84
As a result, the lower MLH in southern Hebei was due to a more stable atmospheric turbulent structure than the northern NCP in spring, autumn and winter. This probably resulted from the frequent effect of cold air on the northern NCP, and such cold air was usually too weak to reach southern Hebei (Su et al., 2004). Then the cold front resulting from the cold air system will enhance the wind shear over the northern NCP. In addition, a previous study has revealed that the warm advection from the Loess Plateau usually developed from southwest to northeast, and the higher buoyancy term (Fig. 6g) and lower MLH in southern Hebei will be partially related to the enhanced thermal inversion at the altitude of 1500 m (Hu et al., 2014; Zhu et al., 2016). In summer, due to the northward lift and westward intrusion of the subtropical high on the NCP, the diminishing existence of the weak cold air on the northern NCP accompanied with the regional-scale strong solar radiation and strong turbulent activities will lead to a small turbulent stability contrast between southern Hebei and the northern NCP.
Distribution of the annual mean values of AOD from December 2013 to
November 2014 in the NCP. The PM
In addition, other researchers proposed that absorbing aerosols above the
MLH is another factor affecting the MLH (Y. Wang et al.,
2013; Li et al., 2016; Peng et al., 2016). Absorbing aerosols give rise to increasing
temperature aloft but decreasing temperature at the surface, which
enhances the strength of capping inversion and inhibits convection. In contrast, absorbing aerosols within the mixing layer could
reduce the capping inversion intensity despite the reduction in the surface
buoyancy flux and raise the MLH (Yu et al., 2002). Considering the higher
concentrations of surface PM
When we analyzed the near-ground PM
Previous studies revealed that the most significant meteorological factors
for regional heavy haze formation in the NCP were RH and MLH (Tang et al.,
2016; Zhu et al., 2016). In addition, the
Distribution of annual means of
Distributions of annual means of
As MLH and WS can represent the atmospheric dissipation potential in the vertical and horizontal directions, respectively, in addition to the MLH, we analyzed the WS variations on the NCP. Similar to our analysis in Sect. 4.2, as SJZ and QHD had no sounding data and due to the close geographic proximity among SJZ and XT as well as LT and QHD, sounding data from the XT and LT stations were used instead of the data at SJZ and QHD, respectively. The WS profiles were averaged every 100 m at each station and are depicted in Figs. 6 and S3. Except for summer, the WS in southern Hebei was far less than that on the northern NCP in spring, autumn and winter (Figs. 6e, S3a and S3e) but was nearly consistent in summer (Fig. 6a). This finding indicated a weaker horizontal diffusion potential in southern Hebei than that on the northern NCP.
The
When we utilized the wind profiles in Figs. 6 and S3 with equal spacing in
the vertical direction,
Seasonal variations of
Therefore, with lower MLH, lower WS and higher RH occurrence in southern Hebei
compared to the northern NCP, the near-ground PM
The schematic diagram of the meteorological causes for heavy haze in southern Hebei.
To gain new insight into the spatiotemporal variation of the regional MLH, in the present study we conducted simultaneous observation with ceilometers at three inland stations (BJ, SJZ and TJ) and one coastal site (QHD) to obtain MLH data at high spatial and temporal resolution. The experiment period lasted for 22 months from 16 October 2013 to 15 July 2015, and a whole year of data (from December 2013 to November 2014) were utilized for further study. Conclusions were drawn as follows.
The ceilometers retrieve both inland and coastal MLH properly. The MLHs in the inland areas of the NCP were high in spring and summer and low in autumn and winter. While under the impact of TIBL, the coastal MLH had an opposite variation trend than inland sites and the lowest MLH in QHD occurred in summer. The TIBL impaired the local MLH development at the coastal site and caused the mixing layer to peak early in summer; this effect weakened with distance inland.
The MLH in southern Hebei was lower than that on the northern NCP,
especially in spring, autumn and winter. This mainly resulted from the more
stable turbulent structure (weak shear term, higher buoyancy term and larger
frequency of
The lower MLH and WS in southern Hebei restricted the atmospheric environmental potential and the pollutant dissipation potential, respectively. Accompanied by higher RH values (stronger pollutant growth potential), the adverse weather conditions will cause severe haze to occur easily in southern Hebei, and the industrial layout in the NCP is in need of restructuring. Heavily polluting enterprises should be relocated to locations with better weather conditions (e.g., some northern areas and coastal areas), and strong emission reduction measures should be implemented in the remaining industrial enterprises to improve air quality.
Overall, the present study is the first to conduct a long-term observation
of the MLH with high spatial and temporal resolution on a regional scale.
The observation results will be of great importance for model
parameterization scheme promotion and provide basic information on the
distribution of weather conditions in the NCP region. A weakness of this
study is that we did not account for the transport effect on PM
The datasets used in this study can be accessed by contacting the corresponding author (Guiqian Tang: tgq@dq.cern.ac.cn).
The supplement related to this article is available online at:
The authors declare that they have no conflict of interest.
This article is part of the special issue “Regional transport and transformation of air pollution in eastern China”. It is not associated with a conference.
This work was supported by the National Key R&D Program of China (2017YFC0210000), the National Natural Science Foundation of China (41705113), the Beijing Municipal Science and Technology Project (ZL171100000617002) and the National Earth System Science Data Sharing Infrastructure, National Science & Technology Infrastructure of China. Edited by: Renyi Zhang Reviewed by: three anonymous referees