1 Introduction

Ozone (O₃) pollution has become an emerging concern in China, influencing air quality, public health and vegetation productivity (K. Li et al., 2019; Wang et al., 2023, 2025). Chemically produced through reactions of volatile organic compounds and nitrogen oxides, O₃ is also modulated by meteorological conditions and vertical transport processes (Chen et al., 2024; Elbern et al., 1997; Holton et al., 1995; Langford et al., 2022; D. Li & Bian, 2015; D. Li et al., 2015; Stohl et al., 2003). From a source-receptor perspective, at a given location, O₃ variations depend on the local photochemical production, and transport of tropospheric O₃ from regional and across-continental sources, and downward injection of stratospheric O₃ (Chen et al., 2023). Vertical transport mixes airmasses from different sources, and hence shapes the vertical structure of tropospheric O₃ (Emmons et al., 2012; Liu et al., 2009; Shen et al., 2022).

Long-term ozonesonde measurements in China are relatively limited, available only in Beijing and Hong Kong, while vertical O₃ profiles have been measured through the IAGOS program at several airports (Chan et al., 2003; Chen et al., 2023; Ding et al., 2008; Roux et al., 2020; Wang et al., 2021). From these measurements, distinct seasonality of tropospheric O₃ in China has been illustrated (Chen et al., 2023; Ding et al., 2008; Oltmans et al., 2004; Wang et al., 2012; Zhang et al., 2020), which is essentially important for low-level O₃ build-up and occurrences of high-O₃ episodes. For example, in southern China (Chen et al., 2023), the multi-year O₃ profiles reveal that springtime O₃ is typical of high values throughout the troposphere. During summer, high O₃ values persist in the middle-to-upper troposphere, but are sharply reduced in low levels, resulting in the largest vertical variations among seasons. Tropospheric O₃ remains lower in autumn and winter O₃ than its annual mean throughout the tropospheric column (Chen et al., 2023), which is also found in some Chinese megacities in another study (Liao et al., 2021).

The structures of tropospheric O₃ in Beijing, representing the northern China, have characteristics similar to those in southern China during spring, autumn and winter, but show the most prominent O₃ abundance in summer throughout the troposphere among the entire year (Zeng et al., 2023; Zhang et al., 2020, 2021). Impressively, O₃ in the middle-to-lower troposphere increases largely from spring to summer, which is opposite to a decreasing evolution in southern China, such as Hong Kong and Taipei (Chen et al., 2023). Maritime airmasses with low O₃, brought by the Asian summer monsoon (ASM), are responsible for the summertime O₃ valley in southern China (Chen et al., 2023; Liao et al., 2021; Oltmans et al., 2004). But it seems not to be the case in Beijing, though it is also influenced by the ASM. Instead, several researchers found that intensive tropospheric photochemistry dominates the summertime O₃ abundance (Liao et al., 2024; Wang et al., 2012; Zhang et al., 2020). Additionally, biomass burning (BB) in Central Asia and Russia promotes the transport of O₃ and its precursors, and is suggested an important reason for the high summertime tropospheric O₃ (Ding et al., 2008; Zhang et al., 2021). A recent work using quasi-two-decade ozonesonde observations in Beijing calculated that BB can lead to an increase of 60.6 ppbv in tropospheric O₃ (Zeng et al., 2023). Stratospheric intrusion (SI) of O₃-rich airmasses is also an important source of summertime tropospheric O₃ (Oltmans et al., 2004; Zhang et al., 2022), though its contribution remains a question to be resolved. In a deep SI event reaching the ground level in the North China Plain (NCP) triggered by convection, Chen et al. (2022) observed that such stratospheric impact can directly induce surface O₃ pollution over broad areas of northern China. Ding et al. (2008) attributed the O₃ maximum in summer over Beijing to the mixture of multiple sources including SI and BB.

The day-to-day variations of tropospheric O₃, primarily driven by the combined influence of different weather systems, can impact lower-level O₃ and determine the occurrence of O₃ pollution (Dufour et al., 2018, 2021; Huang et al., 2015). To resolve the vertical O₃ structure and its controlling factors in northern China, a highly polluted region, we conducted a field campaign in Binzhou city, Shandong Province located in the NCP, the SHAndong Triggering Lightning Experiment (SHATLE; Qie et al., 2007, 2011; Yang et al., 2010), during June–August 2024. With a focus on the natural processes impacting tropospheric O₃ such as convective storms, lightning and SI, we deployed balloon-borne measurements with ozonesondes to collect in situ O₃ vertical observations.

During the campaign, the cut-off low (COL) weather system is of particular interest, since it is not only an important trigger of stratosphere-troposphere exchange (STE) processes (D. Li et al., 2015; Song et al., 2016), but also frequently induces intense convective episodes over northern China. Correspondingly, several ozonesondes were successively released to track the fine-scale evolution of O₃ during the passage of a COL. The profile observations reveal an anomalous structure of tropospheric O₃, characterized by positive O₃ anomalies from the background in the middle-to-lower troposphere and prominent negative O₃ anomalies in the upper troposphere. Such a structure shows pronounced O₃ deviations from the expected summertime baseline profile over northern China, highlighting the need to better understand how the synoptic situations contribute to short-term tropospheric O₃ variations.

In the following sections, we will illustrate the structural differences of summertime O₃ profile and synoptic processes responsible for such a “bottom-heavy” O₃ distribution. Inspired by the discovery of the vertical O₃ structure in this novel case, we further statistically assess the impact of downward transport of stratospheric O₃-rich airmasses on tropospheric O₃. Therefore, the goals for the present study can be concluded as follows: (a) to characterize the tropospheric O₃ behavior under influences of the COL episodes over northern China during summer; (b) to reveal the fine-scale evolution of meteorological conditions leading to the anomalous structure of tropospheric O₃; (c) to assess the contribution of SI to the middle troposphere based on multi-year satellite observations. The outcomes from this study can complement our understanding of summertime tropospheric O₃ abundance over northern China and their linkages with synoptic weather conditions.

2 Data and Model

The balloon-borne electrochemical concentration cell (ECC) ozonesonde and iMet radiosonde are deployed to measure vertical profiles of O₃ and meteorological variables including air pressure, temperature, relative humidity (RH), and horizontal winds during the SHATLE campaign. The precision of ECC O₃ sensor ranges between 3% and 5% from the surface level to 30 km (Smit et al., 2007). Four ozonesondes were released on 15–17 July 2024 in Binzhou (38°N, 117.9°E) to track tropospheric O₃ variations associated with the evolution of COL (Figure 1). Long-term ozonesonde observations, with a total of 911 profiles, are available in Beijing (39.8°N, 116.5°E) over 2001–2019, which is 230 km away from the campaign. Therefore, we calculate the representative summertime tropospheric O₃ from the Beijing 239 profiles during June, July and August and use them as the summertime O₃ baseline over the northern China.

The satellite-based daily vertical profiles of O₃, CO, and H₂O from the Atmospheric Infra-Red Sounder (AIRS; Fu et al., 2018; Rawat et al., 2023) with a horizontal resolution of 1° × 1° are used to investigate the horizontal and vertical variations in trace gases during the COL episode. We further use multi-year AIRS data to conduct a statistical analysis of stratospheric influence on tropospheric O₃ over northern China (35–55°N, 105–125°E, Figure 1a).

The meteorological conditions are extracted from the three-dimensional MERRA-2 (The Modern-Era Retrospective Analysis for Research and Applications, Version 2) reanalysis data (Gelaro et al., 2017), which are publicly available with a horizontal resolution of 0.5° latitude × 0.625° longitude and 72 model vertical levels from the surface to 0.01 hPa, and a 3-hr temporal resolution.

The transport processes and origin of different airmasses probed during the COL episodes are tracked with the National Oceanic Atmospheric Administration's (NOAA) hybrid single-particle Lagrangian integrated trajectories (HYSPLIT) model (Stein et al., 2015). We run 10-day backward trajectory simulations using the Global Forecast System 0.25° × 0.25° meteorological data set. Variations in PV, O₃, RH and vertical pressure velocity along the trajectories are extracted from the MERRA-2 data.

3 Results

3.1 Summertime Tropospheric O₃ Behaviors Associated With COL Over Northern China

COL is an isolated circulation system forming from a pre-existing cold trough in the middle-upper troposphere. COL is characterized by a low tropopause, and a high-potential-vorticity (PV) tongue extending from the upper troposphere and lower stratosphere (UTLS) region. Episodes of COL can transfer stratospheric air into the troposphere, leading to O₃ enhancements even at lower levels (Price & Vaughan, 1993). High frequencies of COL occur during the warm months over northern China, particularly in May, June, and July (Nie et al., 2023). Figure 1 shows an example of COL episodes during the Binzhou campaign in July 2024. A former COL at mature stage was observed on 10 July, which delivered large amounts of O₃ into the middle-to-lower troposphere via the tropopause folding (Figures 1a and 1b). While this COL propagated eastward and weakened, a new COL formed as the cold trough deepened in the northwest of Binzhou on 15 July (Figures 1c and 1d). The isolated cyclonic vortex was filled with high-PV airmasses, moved slowly into northern China, and was accompanied by a descending dynamical tropopause reaching 400 hPa. The evolution of COL triggered substantial tropospheric O₃ enhancements revealed by AIRS observations (Figures 1e and 1f). The O₃ enhancements were connected with the stratospheric O₃ reservoir, penetrated southward and reached Binzhou on 17 July (Figures 1g and 1h). Along with the development of the new COL, the western Pacific subtropical high (WPSH) was rather intensive over northern China, represented by the extent of the 5880-m geopotential height at 500 hPa. Under the influence of WPSH, the lower troposphere remained quite stable, as seen in the widespread regions with high static stability and the quasi-uniform planetary boundary layer (PBL) heights confined to 850 hPa.

In situ vertical O₃ concentrations were measured by four successive ozonesondes in Binzhou on 15–17 July associated with the evolution of COL as shown in Figure S1 in Supporting Information S1. A climatic summertime O₃ baseline profile over northern China is obtained by averaging the long-term ozonesondes over Beijing collected in June, July and August, and is applied to evaluate the tropospheric O₃ behaviors observed over Binzhou. During the COL episode, a bottom-heavy structure of tropospheric O₃ is noticeable, where distinctive O₃ enhancements overwhelming the baseline appear in the middle-to-lower troposphere and a sharp O₃ reduction in the UTLS region.

Zoomed-in profiles between 0 and 12 km are presented in Figure 2. On 15 July, photochemical production of O₃ was high as seen in the middle-to-lower troposphere under the favorable conditions generated by WPSH. Twenty-four hours later, when the COL approached northern China, O₃ between 4 and 9 km was further promoted beyond the seasonal baseline. Layers with O₃ in excess of the 90th percentile extended in 5–7 km together with the sharp reduction of RH values there. It is worthy note that the RH measured in the O₃ enhancement layer ranged between 1% and 5%, which is indicative of the intrusion of dry and O₃-rich stratospheric airmasses under the influence of COL (Stohl et al., 2003; Trickl et al., 2023). Such concurrence of high O₃ and low RH persisted into the following day, and descended to lower altitudes, though the stratospheric characteristics weakened gradually due to the mixing with tropospheric air. Contrary to the middle-to-lower troposphere, O₃ in the UTLS region was obviously reduced from the baseline, showing negative O₃ anomalies. Quantitatively, the mean O₃ between10–12 km was only 66–85 ppbv on 15–17 July, 33%–48% lower than the normal values. On the other hand, the RH around 10 km experienced substantial increases before it returned to the dryness in high-altitude regions above.

3.2 Fine-Scale Evolution of Synoptic Conditions Responsible for the Bottom-Heavy Structure of Tropospheric O₃

Vertical variations in synoptic situations over Binzhou are examined in Figure 3a to gain deeper insights into the causes of such a bottom-heavy structure of tropospheric O₃. In the UTLS region, MERRA-2 data also show enhanced RH values (>60%) between 300 and 200 hPa on 15–17 July, which agree well with the profile observations (Figure 2) but span over a longer period (over 10 days). These high values of RH were distinctly different from the stratospheric dryness above and the stratospherically intruded airmasses (SIA) below. Together with the abnormally low O₃ values for these altitudes, these moist airmasses in the UTLS region were likely due to their tropospheric origins. The alternating wet and dry features in the vertical direction were also observed in Beijing and Ji'nan (Figure S2 in Supporting Information S1). We further track the transport processes of these low-O₃ and high-RH airmasses in the UTLS region using the HYSPLIT trajectory model (Stein et al., 2015). As shown in Figure S3a in Supporting Information S1, these airmasses originated from the Ngari located in western Tibetan Plateau 3–4 days before they were detected by the ozonesondes. The Tibetan Plateau has been recognized as an important vertical transport window of surface airmasses to enter the UTLS region (Bian et al., 2020; Xu et al., 2022). Ground-based observations in Ngari, with an average altitude of over 4.5 km, show that the surface pressure ranged between 550 and 560 hPa, and the maximum daily 8 hr average (MDA8) O₃ concentrations were 70–90 ppbv on 12–13 July. Convective clouds with intense upward air motions were widely distributed in Ngari (Figure S4 in Supporting Information S1). With reference to the trajectory height and meteorological conditions (Figures S3b and S3c in Supporting Information S1), it can be inferred that the moist and relatively low-O₃ air within the Ngari PBL was transported upward and traveled eastward into northern China, which led to the anomalous declining O₃ in the UTLS there.

When the COL was approaching northern China, the dynamical tropopause began to fall since 15 July (Figure 3a). Downward air motions dominated the middle-to-upper troposphere, favoring the descent of stratospheric air transferred by the COL. A small portion of airmasses with high PV and reduced RH was noticeable between 500 and 600 hPa, corresponding to the O₃ enhancements of stratospheric origins in the ozonesonde on 16 July. These dry airmasses continued to penetrate into lower levels slowly and reached around 700 hPa (Figure 3a). Because of the intense WPSH, the PBL in Binzhou was rather stable as seen in the high static stability N² between 850 hPa and 800 hPa. With reference to the profile observations (Figure 2), the N² was as large as 4 × 10⁻⁴ s⁻², acting as a barrier for the SIA to enter the PBL. As a result, substantial O₃ enhancements only appeared in the middle-to-lower troposphere due to the combined effect of SI and the stable PBL.

Backward trajectory simulations of the airmasses at 6 km also provide evidence for the stratospheric impact, which show that the air previously originated in regions with high values of PV and O₃ but low RH near the tropopause region, and descended into the middle troposphere (Figures S3d and S3e in Supporting Information S1). We further map the distribution of O₃ and H₂O in the tracer-tracer scatterplots using the profile observations (Figures 3b–3e). The tracer-tracer correlation analysis, which applies tracer pairs with distinct differences in their abundance within the troposphere (such as H₂O) and stratosphere (O₃), is a powerful tool to distinguish airmasses of stratospheric origins (Chen et al., 2025; Homeyer et al., 2011; Pan et al., 2004; Schäfler et al., 2023; Schroeder et al., 2014; Tilmes et al., 2010). As shown in Figures 3b–3e, the O₃-H₂O distributions of the SIA observed in the middle troposphere were characteristically different from the normal tropospheric air. Compared with the profile observation on 15 July (Figure 3b), O₃ between 5 and 7 km significantly overwhelmed the threshold of tropospheric O₃ (Figure 3c), while H₂O there were sharply reduced from 5,000 to 6,000 ppmv to 200 ppmv, and were even lower than the convectively transported airmasses at 10 km (700–800 ppmv). On 17 July, the SIA were still noticeable as a thin layer located approximately at 4.5–5 km (Figure 3d), and their stratospheric characteristics continued to fade due to the mixing with surrounding tropospheric air (Figure 3e).

Given the temporally dense ozonesondes (3 launches in a day), it is feasible to track the descending rate and mixing processes of the SIA, providing some vital information to assess the stratospheric influence on tropospheric O₃. During a period of 16 hr (Figures 3c and 3d), the height of extreme dryness associated with the SIA fell from 5.23 to 4.83 km, equal to a descending rate of 25 m per hour. In terms of the variations in O₃ during the same period, the SIA lost O₃ at a rate of 0.8 ppbv per hour. Suppose that the SIA continued to lost their stratospheric characteristics at the same rate, then the height of extreme dryness would be 4.61 km by 2100 LST July 17, and the associated O₃ would be 96.8 ppbv. The profile observation at that time shows that the height of minimum H₂O value below 5 km located at 4.62 km, while the O₃ value was 92.2 ppbv (Figure 3e). Thus, it can be inferred that the SIA descended a constant but slow rate, probably due to the stable conditions brought by the intensive WPSH, as seen in the high N² at between 4 and 5 km (Figure 2). As a comparison, during the passage of tropical cyclones, Das et al. (2016) calculated that the descending rate of enhanced O₃ layers of stratospheric origins was 1 km day⁻¹ using several ozonesondes, which is approximately twice the value of this study (0.6 km day⁻¹). In a SI event observed over Kagoshima, Japan, Oltmans et al. (2004) reported a more dramatic descent rate of 4–5 km in 2 days. On the other hand, the loss of O₃ seemed to be accelerated because of a longer mixing history of tropospheric air. Under such circumstances, the SIA can prolong their stratospheric influence even in the middle troposphere, but fail to penetrate into lower levels owing to more contact times with tropospheric air.

As the ozonesonde observations were performed over limited region, the horizontal distribution and vertical evolution of the SIA are investigated over northern China, in a domain of 35–55°N and 105–125°E (Figure 1a), using AIRS observations to gain a more complete picture of the stratospheric impact on tropospheric O₃. With reference to the summertime statistics of the layered AIRS observations, including O₃, CO, and H₂O (Figure S5 in Supporting Information S1), the airmasses are regarded as SIA when their O₃ exceeds the 90th percentile background value and CO and H₂O below their 10th percentile background values. This proxy for stratospheric characteristics is screened in each layer between 850 and 400 hPa. As shown in Figure S6 in Supporting Information S1, the SIA associated with the development of COL on 15 July penetrated southeastward into northern China, and descended vertically reaching 600 hPa and 700 hPa near Binzhou on 17 and 18 July, respectively. However, none of SIA were detected below 700 hPa during the following days, as a result of mixing with tropospheric air and high stability brought by the intensive WPSH.

3.3 Assessment of Stratospheric Influence on the Middle Troposphere Over Northern China During Summer

The higher frequency of COL during the warm months can, potentially, trigger more stratosphere-to-troposphere (STT) transport events leading to enhanced O₃ concentrations in the troposphere, which are crucial for tropospheric O₃ budget and atmospheric chemical nature (Price & Vaughan, 1993). Given the considerable O₃ inputs of stratospheric origins observed in the STT case analyzed above, we present a quantitative investigation of stratospheric influence on the middle troposphere over northern China, using multi-year AIRS satellite observations. The SIA during summer at 500 hPa are searched following the criteria aforementioned. Days associated SIA existence are screened out, together with their areal extent and O₃ enhancements (Figure 4). These three metrics, that is, the frequency, coverage and exceedance of O₃, are important for a comprehensive understanding of the stratospheric impact. Generally, the occurrences of SIA are more concentrated in June, but drop quickly in July and August. On average, approximately 13% of summer days are associated with the STT transport events. Moreover, these events at 500 hPa seem to be confined over localized regions during summer, with an averaged coverage 3.5% of northern China. Impressively, these SIA can promote substantial short-term O₃ enhancements in excess of 35% above the normal O₃ in the middle troposphere, yielding non-negligible concerns in tropospheric O₃ budget (Chen et al., 2022, 2024; Ding et al., 2008; Oltmans et al., 2004; Trickl et al., 2023; Wang et al., 2012).

4 Conclusions

Seasonally, the summertime tropospheric O₃ over northern China shows prominently high O₃ abundance throughout the troposphere. Factors such as photochemistry, BB and SI are suggested as important reasons for the high summertime O₃. Essentially, vertical transport, carrying airmasses from different O₃ sources, still remains a crucial issue to be documented and better understood due to limited in situ vertical O₃ profile observations.

Focusing on the short-term tropospheric O₃ behaviors in summer, we deployed temporally dense ozonesondes during a field campaign to measure in situ O₃ profiles, which have been rarely conducted over northern China. Particularly, this paper reports an anomalous vertical structure of tropospheric O₃ associated with an approaching COL, an important trigger for the STT transport. The vertical profiles were characterized by a “bottom-heavy” O₃ distribution, where O₃ significantly exceeded their summertime baseline values in the middle-to-lower troposphere but was sharply reduced in the UTLS region in the meantime. Quantitatively, O₃ concentrations between 5 and 7 km were even larger than their 90th percentile at the same altitudes, however, they were 33%–48% lower than the normal values between 10 and 12 km, leading to large shifts from the summertime baseline profile over northern China.

Combining the fine-scale meteorological conditions and trajectory simulations, we find that the anomalous O₃ decrease in the UTLS region was resulted from the convectively lofted PBL pollution originated in the Tibetan Plateau, which is an important vertical transport window of low-level airmasses. The enhanced O₃ layer in the middle-to-lower troposphere, however, was associated with the intrusion of stratospheric O₃-rich airmasses brought by the approaching COL. On the other hand, the WPSH was rather intensive and generate a stable PBL over northern China, inhibiting the downward transport of SIA toward lower troposphere. We calculate the descent rate of SIA from the temporally dense ozonesondes, which was 0.6 km per day, approximately half of the value reported during the passage of tropical cyclones (Das et al., 2016). Such a slow descent rate of SIA can prolong the stratospheric influence in the middle troposphere, but reduce their chances to reach the PBL due to a longer mixing history with tropospheric air.

Inspired by the discovery of the vertical structure of trace gases revealed in the forementioned case, we further assess the frequency, coverage and O₃ enhancements associated with the SIA over northern China during summer, based on multi-year AIRS satellite data. Statistically, days with SIA consist of 13% of the summertime, and are most frequent in June. Though occurring over localized regions, the SIA can add extra O₃ of stratospheric origins in the middle troposphere during the short-term STT periods, and hence play a non-negligible role in tropospheric O₃ budget.

From the ozonesonde observations, the SIA brought by the COL were found above 700 hPa and did not enter the PBL due to the intensive WPSH, suggesting that the stratospheric influence at the ground level depends on local synoptic conditions. Even though, given that the frequency of COL and convective clouds are both higher in warm months, there are potentially injection of SIA reaching the ground and induce high-O₃ episodes, raising concerns about surface air pollution (Chen et al., 2022; Cooper et al., 2006; Dreessen, 2019; Grant et al., 2008). Moreover, such stratospheric influences are pronounced enough to change local environments impacting the chemistry, radiation and cloud formation and hence affect weather and climate (Bian et al., 2020; Xie et al., 2016, 2025; Zhang et al., 2024).

Since it is challenging to track SIA descending into the PBL using satellite observations of O₃, routine ozonesonde measurements conducted over a longer period are of particular value to resolve the fine-scale evolution of SIA, and hence, combining with surface O₃ observations (Chen et al., 2024), can aid to assess the stratospheric contribution to the PBL O₃ variations. Such knowledge is warranted to improve the understanding of surface O₃ pollution related to natural processes such as SI. In future work, provided with the quasi-two-decade ozonesonde observations in Beijing, we will further address the fate and contribution of SIA in the PBL. These results can provide new insights into the linkages between O₃ behaviors and synoptic processes, and hence aid to design effective O₃ pollution mitigation strategies over polluted regions.

Conflict of Interest

The authors declare no conflicts of interest relevant to this study.