2.1. Details of Ozone Sounding Site and Ozone Sounding Experimental Procedure

The ozone sounding observation in the Pearl River Delta was carried out at Qingyuan National Meteorological Station (23.692°N, 113.016°E, Figure 1). The station has maintained conventional ground meteorological and sounding observations. During the test period from December 2021 to February 2022, a total of 13 atmospheric ozone high-altitude sounding balloons were placed once a week. The application time was selected at about 14 : 00 h Beijing time. The parameters determined were atmospheric ozone partial pressure (concentration), wind, temperature, air pressure, and humidity. The data were recorded every second and transmitted to the receiver on the ground. Throughout the period, only nine valid data were obtained as the ozonesonde observation was affected by receiver outage and signal interferences (the data were taken on December 1, December 8, December 22, December 29, 2021, January 5, January 12, January 26, February 9, and February 16, 2022).

2.2. Ozone High-Altitude Detection System

Consequently, the amount of ozone reacted with iodine in the solution at a unit time is proportional to the generated current. Therefore, the ozone content in the atmosphere can be determined by measuring the electric current flowing through the solution.

4.1. Analysis of ECC Observation Results

Figure 3 shows the vertical distribution profile of ozone partial pressure, temperature, and relative humidity during the test period from December 2021 to February 2022 (the observation results were collected on 1 December, 8 December, 22 December, 29 December 2021, 5 January, 12 January, 26 January, 9 February, and 16 February 2022).

In general, the sounding balloon bursts naturally at about 35 km altitude; therefore, only vertical ozone characteristics from the ground to 35 km were analyzed. The trend of ozone partial pressure change was the same up to 16 km of the top of the troposphere; however, the change was rapidly increased above 16 km of the top of the troposphere. The maximum value of the ozone layer in the stratosphere was about 25 km from the ground (the average maximum value is 14 mPa), and afterward, it decreased. The distribution of ozone partial pressure was mainly unimodal from the troposphere to the stratosphere except for February 6, 2022, which was bimodal.

Figure 4 shows that the ozone content in the troposphere was lower than that in the stratosphere. In the troposphere, the ozone content near the ground was slightly higher, and the low ozone partial pressure was observed between 5 and 15 km. The content of ozone peaked in the lower troposphere (0-2 km), and peak structure was mainly divided into single peak (22%) and multimodal (77%) in the upper troposphere. During the observation period, the surface ozone concentration exceeded 100 μg/m³ for two days (Figure 3), which were 127 μg/m³ and 115 μg/m³ on December 8, 2021, and January 5, 2022, respectively. The ozone profile in the middle and low troposphere on 8 December 2021 showed a multipeak pattern, whereas on 5 January 2022, it showed a single-peak pattern (Figure 3).

In the vertical direction, the correlation coefficient between ozone partial pressure and temperature was −43.2%, and that between ozone partial pressure and relative humidity was −52.5%. The temperature in the troposphere decreased with the increase in height, and the average temperature decline rate was 1.3°C/km. The average temperature at the top of the troposphere was −80°C. Ozone absorbs solar radiation thereby heating the atmosphere; consequently, when ozone content increased from top to bottom of the troposphere, the temperature in the troposphere also gradually increased. As water vapor was mainly concentrated below 10 km, and the relative humidity was stable at the low level of the stratosphere.

4.2. Comparison of Ozonesonde Observation and Ozone LiDAR Data

LiDAR uses a differential absorption method to retrieve ozone concentration. The principle is to use the measured gas’s absorption characteristics to measure the gas’s attention.

During the ozone sounding observation, a comparison test was performed using the ground ozone LiDAR. Light detection and ranging (LiDAR) technique was carried out to test the detection and inversion performance of the ozone and to “fill in the blind” at the near-ground (0-200 m) for LiDAR ozone radar detection. Owing to the off-line laser radar equipment and the effective data of laser radar, the LiDAR observations were carried out only on 29 December 2021, 5 January, 26 January, and 26 February 2022. Figure 5 shows the comparison between the ozone profile retrieved by LiDAR and the sounding experiment. The ozone measurements determined from LiDAR radar were attenuated at altitudes above 1000 m. There was a significant attenuation at a height of above 3000 m on January 5, 2022. As a result of attenuation at a height above 3000 m or even lower (above 1000 m) by the LiDAR technique, it was unable to detect the vertical distribution of ozone in altitudes above 1000 m. According to Figure 5, the vertical profile of ozone detected by LiDAR was significantly different from that detected by the sounding system. The vertical profile of ozone on 29^th December 2021 determined from LiDAR was a single-peak type, and the ozone concentration was 59.9 ppb at 517.5 m (Figure 5(a) and Table 1). On the same day, the vertical profile of ozone determined by the sounding system was a double-peak type, and the ozone concentration was 48.9 ppb at 1108.1 m (Figure 5(a) and Table 1). The difference between the two at the same height ranged between 10 ppb and 50 ppb. The ozone profile structure determined on 5^th January 2022 by both LiDAR and the sounding system techniques showed a fluctuating multipeak structure, and the changing trends of the two systems were relatively consistent. The ozone concentration determined from the LiDAR technique was 105.3 ppb at 532.5 m (Figure 5(b) and Table 1). It was 90.4 ppb for the sounding system technique at the height of 83.1 m (Figure 5(b) and Table 1). The difference of the concentrations between the two techniques at the same height ranged between 20 ppb and 40 ppb. The ozone profile structure on 26^th January 2022 determined from the LiDAR technique was unimodal (Figure 5(c)). The ozone concentration in the near-surface layer (below 500 m) increased with height, and the ozone concentration from 500 m to 1000 m decreased with the height. The ozone concentration below 300 m in the near-surface layer increased with height, and the ozone concentration from 300 m to 500 m decreased slightly with height. The ozone concentration from 500 m to 1000 m increased with height (which was opposite to the LiDAR ozone profile). The ozone concentration determined from the LiDAR technique on 26^th January 2022 was 83.2 ppb at 570 m (Figure 5(c) and Table 1). The ozone concentration determined from the sounding system was 56 ppb for the same day at 725 m (Figure 5(c) and Table 1). The difference ranged between 20 ppb and 40 ppb at the same height. The ozone profile structure on 16^th February 2022 determined from the LiDAR technique showed a unimodal pattern with a single peak while it was a bimodal pattern when determined using the sounding system (Figure 5(d)). The ozone concentration on the 16^th of February 2022 determined from the LiDAR technique was 94.4 ppb at 517 m (Figure 5(d) and Table 1). The sounding system ozone peak concentration was 79.6 ppb at 83.2 m (Figure 5(d) and Table 1). The difference between the two systems at the same height was ranged between 10 ppb and 60 ppb.

Figure 6 shows the mean ozone profiles of LiDAR and sounding systems during the experiment. Because LiDAR has apparent attenuation at the height above 1400 meters, the difference between the two is mainly compared for heights below 1400 meters. As seen from Figure 6, at the height of 100 meters to 1400 meters, the difference between the vertical profile of ozone detected by LiDAR and that detected by the sounding system is noticeable. At the same height, the difference between the two is between 10 ppb and 60 ppb, and the profile of the radar shows a single peak. The peak value of the ozone profile of LiDAR mainly occurred at about 500 m. The ozone concentration from 200 to 500 m increased with height, and the ozone concentration from 500 m to 1000 m decreased with height. The peak value of the ozone profile of the sounding system mainly occurred near the ground or 1000 m. The blind area under 100 meters was detected by the LiDAR; hence, ozone sounding can be supplemented with the near-surface blind spot for LiDAR ozone detection and calibrated the LiDAR inversion results.

4.3. Stratosphere-Troposphere Exchange

Figure 4 shows that stratospheric ozone content was much higher than tropospheric ozone content. Does stratospheric high ozone content transport to the troposphere? In order to explore this problem, potential vorticity was used to characterize the change of tropospheric top [7].

PVU is called a potential vortex unit. World Meteorological Organization (WMO) defined 1.6 PVU as the tropospheric top height. Some studies suggest that the top height of the troposphere is between 1 PVU [9] and 3 PVU [10], and there is no well-defined PVU value to distinguish the stratosphere from the troposphere. In this paper, 1.6 PVU was taken as the tropospheric top height. Figure 7 shows the vertical profile of potential vorticity during the sounding test. During the observation period (nine detection times from December 2021 to February 2022, except 5^th January 2022), the potential vorticity 1.6 PVU was mainly located near 100 hPa (Figures 7(a) to 7(i)). However, the potential vorticity, 1.6 PVU, on January 5, 2022, dropped to 400 hPa (Figure 7(e)), indicating that there was a downward transport of stratospheric air mass on this day. Figure 8 shows the distribution map of the wind field at 14 : 00, 200 hPa on January 5, 2022. The central wind speed of the subtropical jet over Qingyuan was too high, exceeding 70 ms^-1. The left entrance area of the jet was the convergence area, and there was a strong subsidence movement. Figure 9 shows the profile of the wind field streamlines along 113°E. According to Figure 9, the stratosphere (24°N) over Qingyuan to the top of the troposphere (around 400 hPa) was a strong downdraft. The airflow initially gathered around 400 hPa and then spread to the ground through the downdraft movement, resulting in a rapid rise of near-surface ozone concentration in a short time. The ground ozone concentration exceeded 100 μg/m³ on 5 January 2022, which was the day of high ozone concentration observed during the sounding test (Figure 3).

The US Environmental Protection Agency (hereinafter referred to as EPA) defined the Stratospheric Ozone Intrusion Exceptional Events (SOIEE) [11] under “Exceptional Events” in 2007 and revised in 2016: Exceptional Events are triggered by meteorological conditions such as high-altitude frontogenesis and strong convection. The lower stratosphere and the upper troposphere in the midlatitude region are prone to the “folding” of the top of the troposphere [12–14]. This phenomenon makes the ozone rich in the stratospheric atmosphere thereby invading the free troposphere. If there is an obvious sinking airflow, it will further diffuse through the inversion layer to the near-surface, resulting in a rapid rise in ozone concentration near-surface in a short time, which can lead to an excessive ozone concentration in serious cases. “Exceptions” are uncontrollable and preventable. The stratospheric ozone intrusion is mainly affected by atmospheric circulation dynamics and heat, which is different from the near-surface anthropogenic emissions. The “exceptional events” in the United States generally define stratospheric ozone intrusion as a natural event, pointing out that the excess ozone concentration caused by this event is essentially beyond human control and reasonable prevention and does not belong to the category of legal control [15]. It can be treated according to the “exceptional events,” “Clean Air Act,” relevant general rules, and “Guidelines” standards [15]. The stratospheric ozone intrusion into the troposphere on January 5, 2021, can be treated as an “exceptional event.” When formulating ozone pollution prevention and control measures, this factor should be considered to avoid underestimating the effectiveness of ozone pollution prevention and control measures and overestimating the severity of ozone pollution.