Study on the mixed layer, entrainment zone, and cloud feedback based on lidar exploration of Nanjing city



[1] An experiment devoted to the lidar (LIght Detection And Ranging) study of the urban boundary layer (UBL) over Nanjing city was executed from February 20 to March 5, 2006. Many other techniques including radiosonde, meteorological towers, and turbulence measurements were also performed in order to explore the UBL. In this paper, the results of the lidar experiment with other observation data are presented. The experimental results demonstrate the daily transition features of the mixed layer (ML) and the entrainment zone (EZ). The different features of the ML between urban and suburban areas have been revealed by comparison. Furthermore, cloud feedback on the mixed layer depth (MLD) and the entrainment zone thickness (EZT) has been analyzed.

1. Introduction

[2] Pollution increases quickly in urban areas due to industry and traffic. The atmosphere boundary layer (ABL) is the lowest part of the troposphere that has a direct impact on the concentration and transformation of pollutants. This layer has been of great interest in recent years as it plays an important role in numerical weather prediction, radiation budgets, climate modeling, and flight operations.

[3] Many techniques have been developed to understand the mechanisms presented in the ABL and EZ, such as laboratory experiments [Sun et al., 2005; Deardorff et al., 1980], various parameterization schemes [Gryning and Batchvarova, 1994], simulation investigations [Bretherton et al., 1999; Sullivan et al., 1998; Abdella and McFarlane, 1997], and numerous field studies [Mao et al., 2006; Cohn et al., 1998; Flamant et al., 1997; Steyn et al., 1997].

[4] The ABL can be measured with a high spatial and temporal resolution by using remote sensing systems (lidar; sodar, Sonic Detection and Ranging), and relatively simple parameterizations can be used to calculate other quantities such as the entrainment velocity [Stull, 1988]. In this study, the lidar data over Nanjing city coupled with radiosonde and meteorological tower data are used to investigate the daily behaviors of the ML and EZ. The study reveals the ML and the EZ have obvious daily transition features, which show the different responses of the MLD and the EZT to radiation flux. The results of analysis also show that the ML transitions of the urban areas are fairly different from those of the suburban areas. Cloud feedback on the MLD and EZT is distinct, and the transition of the EZ demonstrates a relatively real-time response to cloud change, while the transition of the ML exhibits a delayed accumulation response.

2. Inspection Method

[5] Nanjing (32°0′N, 118°7′E) is a densely populated city and located in the South-east of China. Measurements of lidar and meteorological tower were executed at sites of Dangxiao and the Nanjing University campus. Dangxiao is located in the downtown area, and the Nanjing University campus is located approximately 20km from the downtown area and therefore experiences far fewer local pollution sources. The Yangtse River is located between the two observation sites, as shown in Figure 1. All other measurements, such as radiosonde and turbulence, were located at Dangxiao.

Figure 1.

The locations of the observation sites.

[6] The Earth's surface is the primary source of aerosols, which are concentrated at the top of the ABL and often form an inversion layer. The concentrated aerosols at this height induce a very clear signature of the lidar backscatter coefficient, allowing a clear definition of the top of the ABL. According to this feature, the border between the ABL and the free troposphere (FA) can be calculated by the logarithmic derivative of the lidar range-corrected signal (RCS) [Senff et al., 1996]. The EZT can then be determined by the visual inspection method [Nelson et al., 1989; Boers and Eloranta, 1986]. In this paper, the normalized entrainment zone depth, EZT/MLD, is employed to express the EZT.

[7] The detecting range of the lidar is about 20km. The laser of the lidar was Big Sky Ultra Nd:YAG laser operating at the second harmonic wavelength of 532nm with beam divergence of 1.5mrad at a repetition rate of 20 Hz, whose output energy was measured to be approximately 30mJ. A Schmidt-cassegrain telescope with 200mm diameter, f/10 focus, the window glass with increased transmission coating, was made of steel. The captured light was focused by a convex on a 8mm, low-noise, high-gain R7400U PMT whose photoelectron efficiency can reach 25%. The signal was finally acquired by a Transient Recorder [Mao et al., 2007].

[8] Figure 2 shows the relationship between the MLDs derived from the lidar and radiosonde data during the observation time as well as their linear relationship, which can be expressed by the function: MLD(radio) = 0.033 + 0.972MLD(lidar). This correlation substantiates the ability of the lidar to determine the MLD and the EZT. Furthermore, Figure 2 reveals that the MLDs of Nanjing during the observation period are usually less than 2 km in height.

Figure 2.

Relationship between the MLDs derived from the lidar and radiosonde data during the observation period.

3. Results

3.1. ML Transition Study

[9] Figure 3 is the revolution diagram of a series of time-altitude cross-sections of the lidar data at the Dangxiao site on March 2, 2006. The color scale represents the intensity of the RCS and warm color indicates stronger scatter. The black circles and white circles represent the MLDs derived by lidar and radiosonde, respectively. The stars represent the nocturnal boundary layer. The lack of a circle demonstrates that the data regarding the inversion temperature layer at these moments are not available.

Figure 3.

Revolution diagram of a series of time-altitude cross-sections of the lidar data at the Dangxiao site on March 2, 2006. The color scale represents the intensity of the RCS and warm color indicates stronger scatter.

[10] Figure 3 shows that the nocturnal boundary layer is slowly dropping before sunrise, which is in accordance with the change of the inversion temperature altitude. Due to the existence of a residual layer in the backscatter profile, the ML increase around sunrise is not so clear, but it is obvious that the ML is at its lowest position during this period. After sunrise, the ML grows rapidly from 600 m at 0800 LST (Locate Standard Time) to 1300 m at 1600 LST and then gradually decreases until sunset because of radiative cooling and a decrease in the turbulent kinetic energy. After sunset, the nocturnal boundary layer begins to build as shown by an inversion temperature. In addition, the EZT also grows quickly after sunrise and reaches its highest level at noon when the radiation flux is at its maximum (not shown).

[11] Along with the new inversion temperature formation shortly before sunset, the main energy is cut off from the ML. As a result, there are no rising thermals to sustain the relatively sharp transition zone between the ML and the FA. The local shear-driven turbulence is the main source for nocturnal mixing, and sometimes a broadening of the transition zone between the ML and the FA can be observed. This process can be considered as a shear-driven detrainment [Deardorff et al., 1980; Nelson et al., 1989]. In Figure 3, the transition zone denoted by the arrow is a possible result of detrainment. The detrainment begins from the upper layer of the ML at 1800 LST and finally reaches the lowest layer three hours later. Furthermore, it is found that the extent of detrainment gradually decreases from the upper layer to lower layer. This phenomenon can be explained in two ways: Either the shear-drive is weaker in the lower layer, or the higher density makes it difficult for the lower layer to ascend.

[12] Figure 4 shows the observation results at the Nanjing University campus on March 2, 2006. The color scale represents the same meaning of the Figure 3's. Although the ML behaves the same daily transition trend as in Figure 3, there are some obvious differences. The MLD of the urban region is generally around 100m higher than that of the suburban region. Furthermore, compared with the suburban ML transition, the urban ML transition is relatively less smooth. In addition, the detrainment is not found from the observation results in the Nanjing University campus site. The MLD increase in the urban region may come from the difference in the thermal-drive or shear-drive between the two sites, but the unsmooth ML transition and the detrainment detected during night time are definitely affected by the shear-drive diversity. Throughout the above analysis, it is assumed that the hetero-surface of the city has an important influence on shear-drive diversity, which is exhibited by the ML transition diversity of urban and suburb.

Figure 4.

Revolution diagram of a series of time-altitude cross-sections of the lidar data at the Nanjing University campus site on March 2, 2006.

[13] Figure 5 shows the transitions of the mean MLD and the mean normalized EZT during the daytime for all clear days. It can be seen in Figure 5 that the transition of the mean normalized EZT is correlated with the transition of the radiation flux. The stage from 1100 to 1300 LST is marked by an enlarged EZ, which is also the stage in which the radiation flux is increased. These increases force the mean normalized EZT to remain nearly stable, which creates a new equilibrium entrainment regime that is limited by the gradually forming inversion. In order to make the ML ascend, it is necessary to obtain an adequate drive from the radiation flux, and, therefore, the MLD exhibits a different behavior from the EZT. The MLD displays a slow increase when the radiation flux reaches its maximum value, and then the MLD increases rapidly after the stage of the enlarged radiation flux. This phenomenon demonstrates that the EZ has a real-time response to the radiation flux, while the ML exhibits a delayed accumulation response.

Figure 5.

Asterisks and black circles represent the mean normalized EZT and MLD for all clear days, respectively.

3.2. Cloud Feedback on the ML Transition Study

[14] Cloud feedback is a complex issue. There are various types of clouds, with varying sizes and at varying heights. Furthermore, the microphysical, optical, and radiative properties of clouds as well as the formation of clouds have complicated physical mechanisms [Khvorostyanov, 1995]. For example, clouds generally demonstrate a negative feedback to solar radiation during the daytime and therefore directly impact the transition of the ML. On the other hand, the positive feedback of clouds during the nighttime indirectly influences the transition of the ABL through its impact on the inversion temperature formation.

[15] Figure 6a is the revolution diagram of a series of time-altitude cross-sections of the lidar data on March 4, 2006. The color scale also represents the same meaning of the Figure 3's. This picture shows that a layer of clouds exists from 5.5–8 km at the beginning of the time scale that then devolves into two distinct layers at 0600 LST. The height of the upper layer displays a slowly increasing trend; on the contrary, the height of the lower layer is gradually decreasing over time. After 1400 LST, the sky is relatively clear, and only sporadic clouds appear.

Figure 6.

(a) Revolution diagram of a series of time-altitude cross-sections of the lidar data; (b) the normalized EZT and the MLD from 0800 LST to 2000 LST on March 4, 2006.

[16] Figure 6b is the normalized EZT and the MLD from 0800 LST to 2000 LST on the same day. Despite minor fluctuations, the normalized EZT displays a decreasing trend from 0800 to 1400 LST. At 1100 LST, the normalized EZT peaks due to a decrease in the density of the lower cloud layer. After 1400 LST, it can be seen that the normalized EZT is increasing with the disappearance of the clouds.

[17] According to the above analysis, the EZT exhibits nearly a real-time feedback to cloud existence. Furthermore, the lower the clouds' altitude is, the more pronounced the effect of cloud feedback will be. The cloud of different altitude means to have the different phase and component; therefore, the phases and components of clouds have an important influence on cloud feedback. As previously discussed, the clouds also influence the transition of the ML through their influence on radiation flux; however, The ML does not decrease with the descent of cloud altitude, and neither does an increase present after the clouds disappear, because the transition of the ML displays a delayed accumulation response to the radiation flux, which is a different behavior from the EZ, and the surface has not received adequate heat radiation flux to uplift the ML after 1400 LST.

4. Discussion and Conclusions

[18] Backscatter lidar is a proven tool for remote sensing the ML due to its high spatial and temporal resolution. From our experimental results, it has been shown that the MLDs derived by lidar and radiosonde demonstrate a high correlation coefficient of 0.947.

[19] In general, the daily transition of the MLD demonstrates a minimum height around sunrise and rapidly increases in height after sunrise. Two hours later from noon, it reaches the maximum height and then begins to drop. The EZT, however, is almost immediately influenced by the radiation flux such that the highest EZT coincides with the moment of maximum solar radiation.

[20] The comparison of the ML transition between urban and suburban regions illuminates that the hetero-surface of the city has a significant impact on the urban ML transition.

[21] Clouds influence the transition of the ML and EZ through their influence on the radiation flux. The transition of the EZ demonstrates a relatively real-time response to cloud change, while the transition of the ML exhibits a delayed accumulation response. The extent of the cloud feedback is impacted by the phases and components of the clouds.

[22] It is important to note two features of the experimental procedure utilized in this work. First, the detection frequency of each lidar profile makes it difficult to obtain other information such as entrainment velocity. Secondly, cloud feedback on the transition of the ML and EZ is an important and valuable subject. In order to quantitatively analyze cloud feedback, it is necessary to take a further step in experimental observation.


[23] This work was supported by the National Science Key Foundation of China 40333027 and the Chinese Academy of Sciences Knowledge Innovation Program CX0201.