Similar to the ECC type developed by Komhyr (1969), the IAP ozonesonde is also based on an electrochemical method. For a detailed description of the ozonesonde, readers are referred to Zhang et al. (2013). The ozonesonde-based ozone partial pressure can be computed from the following equation:

, (1)

where P is the ozone partial pressure (mPa); I is the cell output current (µA); I_BG is the cell background current (µA); T is the temperature of the ozonesonde box (K); t is the time for the pump to bubble 100 milliliters (mL) of airflow through the cathode solution (s); η is the conversion efficiency; and is the efficiency of the pump flow rate.

In the conventional method, a pressure dependent background correction is applied for the ECC ozonesonde. However, many studies conducted at the laboratory (e.g., Thornton and Niazy, 1983; Smit et al., 1994) for the ECC ozonesonde were unable to confirm the pressure dependence on the background current; therefore, a constant background current correction was suggested to be applied throughout the entire ozone profile. Based on this, a constant background current measured during pre-flight preparations at surface pressure was applied for the IAP ozonesonde background current correction during the post-flight data processing throughout the entire ozone profile in this study. η is assumed to be 1 during normal operation for the ECC ozonesonde, which is also deployed in the IAP ozonesonde. The pump volumetric flow rate is relatively constant from surface pressure to about 300 hPa during the balloon flight; however, one would expect a steady decrease in the pump flow rate with altitude higher than 300 hPa due to an increase in resistance from pumping against the cathode solution fluid head, dead space in the cylinder of the piston pump, and pump leakage ( Komhyr et al., 1995; Steinbrecht et al., 1998; Johnson et al., 2002). A slight ozone deficiency in the IAP ozonesonde measurements, generally less than 10% compared to the ECC ozonesonde retrievals, was observed in the upper atmosphere, which was likely caused by a decrease in the pump flow rate of the IAP ozonesonde ( Zhang et al., 2013). It should be noted that further study of the pump correction factor as a function of altitude is still required for the IAP ozonesonde; however, in the meantime, it was assumed to be 1 in the present study, which may have resulted in a bias of less than 10% in the upper atmosphere on the basis of our previous results obtained in the preliminarily evaluation campaign.

A total of 22 and 23 releases of the IAP ozonesonde were carried out at Beijing Observatory (BJ, (39.81°N, 116.47°E), 31 m above mean sea level (MSL)) and Changchun Observatory (CC, (43.9°N, 125.2°E), 237 m above MSL) in June 2013. The radiosonde used for the data transmission was provided by Changfeng Company, which participated in the Eighth World Meteorological Organization International Radiosonde Comparison held at Yangjiang, China in 2010 ( Nash et al., 2011). An ozonesonde was generally launched at around 14:00 Local Standard Time (LST) per day; in addition, three ozonesondes were released at 07:30, 14:30, and 19:30 LST on 26 June in BJ to investigate the variation in vertical ozone distributions during different periods for a particular day. The maximum detection altitudes exceeded 30 km above ground level (AGL) over the two sites, except for two launches (one at BJ and one at CC).

The two ozonesonde performance parameters, i.e., background current and response time, obtained during the pre-flight preparation procedures at BJ and CC, are shown in Fig. 1. The background current was less than 0.05 μA for 18 out of 22 releases at BJ (Fig. 1a), and 15 out of 23 releases at CC (Fig. 1b). The mean background currents were 0.03 μA at BJ and 0.04 μA at CC, and their standard deviations of the background current were 0.02 μA and 0.03 μA, respectively. Figures 1c and 1d illustrate the response times, which were generally less than 35 seconds. The response time was less than 30 seconds for 16 and 11 releases at BJ and CC, respectively. The average response time and its standard deviation were 27.8 s and 4.0 s at BJ, and 30.4 s and 3.7 s at CC, respectively. The ascent of the IAP ozonesonde was about 5 m s^-1, thus translating into a vertical altitude resolution of about 140 m (27.8 m s^-1 × 5 m s^-1) at BJ and 150 m (30.4 m s^-1 × 5 m s^-1) at CC, as well as resulting in a slight displacement of the observational profile in the vertical direction compared to the actual profile. In general, no corrections are made for the response lag during usual practice, which was also adopted by the IAP ozonesonde measurements at BJ and CC. The background current and response time over the two sites were, respectively, 0.04±0.02 μA and 29.1±4.0 s, which is comparable to results for the ECC ozonesonde (typically 0.03-0.11 μA for the background current and 20-30 s for the response time) ( Smit et al., 2013) (Table 1).

Note that the same pre-flight preparations were conducted for the IAP ozonesondes released at BJ and CC; the background currents at the two stations were close to each other and were also close to that of the ECC ozonesonde. The response time of the IAP ozonesonde at BJ was close to that of the ECC ozonesonde; however, it was slightly larger by ~2.6 s at CC as compared to that at BJ. The different atmospheric environment at the two stations, such as the temperature and/or relative humidity, was the likely cause for the difference in response time, which needs further study.

Figure 1

	Figure Option ViewDownloadNew Window
	Figure 1 The background current (upper panels) and response time (lower panels) for the ozonesondes released over the Beijing (left panels) and Changchun (right panels) site.

Table 1

Table 1

Table 1 The average background current (µA) and response time (s) for the IAP ozonesonde at the Beijing site, Changchuan site, and the two sites combined, as well as their typical values for the ECC ozonesonde. IAP ozonesonde ECC ozonesonde Background current Response time Background current Response time Beijing site 0.03±0.02 27.8±4.0 0.03-0.11 20-30 Changchun site 0.04±0.03 30.4±3.7 Two sites combined 0.04±0.02 29.1±4.0

Table 1 The average background current (µA) and response time (s) for the IAP ozonesonde at the Beijing site, Changchuan site, and the two sites combined, as well as their typical values for the ECC ozonesonde.

3.2 Vertical ozone distributions

The ozonesonde measures data every second with average ascent about 5 m s^-1 that results in vertical resolution of about 5 m (5 m s^-1 × 1 s). The observed profiles are projected to layers spaced at 50-m intervals from the surface to the upper atmosphere. Figure 2 shows the ozone partial pressure profiles from all ozonesonde launches. In general, the ozone structure at both stations was similar, with large ozone partial pressure located in

the middle of the stratosphere ranging from 20 to 28 km. The maximum ozone partial pressure was about 14 mPa at the two sites. However, a rather larger difference can be revealed by a more detailed inspection. The ozone partial pressure in the planetary boundary layer occasionally exceeded 10 mPa at BJ; however, it was generally less than 8 mPa at CC. The ozone partial pressure in the upper layer of the troposphere (10-15 km) at CC occasionally exceeded 10 mPa. This is indicative of a strong stratosphere-troposphere exchange due to the outbreak of a cold vortex over Northeast China in this season Yang and Lü, 2004).

Figure 2

	Figure Option ViewDownloadNew Window
	Figure 2 Vertical distributions of the ozone partial pressure retrieved from ozonesondes launched in (a) Beijing and (b) Changchun; spaced at 50-m intervals. The areas colored in black denote the altitude ranges without detections, and the other colors represent the ozone partial pressure (in units of mPa).

Figure 3 shows the average profile and its standard deviation of ozone partial pressure. The mean ozone partial pressure profile at BJ showed a bimodal distribution, with a central peak in the middle of the stratosphere and a side peak in the planetary boundary layer. The maximum magnitude of ozone partial pressure was 11.4 mPa, which was located at 24.8 km. The pattern of the vertical ozone distribution at CC was similar to that at BJ, especially in the middle of the stratosphere. Larger ozone variation (as shown by the larger standard deviation) was observed in the upper layers of the troposphere, as compared to values in the planetary boundary layer and stratosphere. A large discrepancy was observed within the planetary boundary layer, in which the ozone partial pressure increased sharply from the surface to 1.1 km AGL at BJ (3.4 mPa km^-1); however, it was 0.4 mPa km^-1 at CC. Most of the ozone partial pressures were < 5 mPa between 10 and 20 km at BJ; however, they were generally > 5 mPa at CC.

A comparison of the observed ozone distributions at 07:30, 14:30, and 19:30 LST on 26 June is shown in Fig. 4. In general, the patterns of vertical ozone distributions collected from three profiles were very close. However, large discrepancies existed in the lower atmosphere where the ozone partial pressure was larger in the afternoon (14:30 LST) and evening (19:30 LST) than that in the morning (07:30 LST); the maximum difference was about 5 mPa. This is indicative of strong ozone production in the boundary layer during daytime due to photochemical reactions ( Ma and Zhang, 2000).

Figure 3

	Figure Option ViewDownloadNew Window
	Figure 3 Average profiles of ozone partial pressure (black line) and their standard deviations in (a) Beijing and (b) Changchun. The vertical resolution is 50 m; shaded grey areas represent the standard deviation.

Figure 4

	Figure Option ViewDownloadNew Window
	Figure 4 Comparison of ozone partial pressure profiles collected from ozonesonde launches at 07:30 (blue line), 14:30 (red line), and 19:30 (green line) LST on 26 June in Beijing.