High tropospheric ozone in Lhasa within the Asian summer monsoon anticyclone 2013: influence of convective transport and stratospheric intrusions

Abstract. Balloon-borne measurements of ozone in Lhasa (29.66° N, 91.14° E, 3650 m above sea level) in August 2013 are investigated using backward trajectory calculations performed with the Chemical Lagrangian Model of the Stratosphere (CLaMS). Measurements show three time periods characterized by high ozone mixing ratios in the troposphere on 8, 11, and 18–20 August 2013 during the Asian summer monsoon season. Here, we could verify two different sources for the enhanced ozone values in the troposphere. First, transport of polluted air from the boundary layer, and second transport from the stratosphere by stratospheric intrusions. Air pollution from South Asia through convective and long-range transport plays a key role in enhancing middle/upper tropospheric ozone mixing ratios up to 90 % on 8 August and up to 125 % on 11 August 2013. Stratospheric air intruded from the northern high-latitude to the southeastern flank of the Asian summer monsoon (ASM) anticyclone to the troposphere and is identified as source of enhanced ozone according to backward trajectory calculation and satellite measurements by the Ozone Monitoring Instrument (OMI) and the Atmospheric Infrared Sounder (AIRS). Air parcels with high ozone moved from the high latitude lower stratosphere to the middle/upper troposphere and are then transported to Lhasa over a long-distance and enhanced upper middle/tropospheric ozone in Lhasa during 18–20 August 2013. Our findings demonstrate that the strong variability of ozone within the ASM anticyclone in the free troposphere is caused by transport from different regions of the atmosphere.


the stratosphere, or upward transport from the boundary layer; (2) photochemical production induced by solar radiation and other chemical reactions involving lighting products or ozone-precursors from biomass burning and anthropogenic pollution.
The Asian summer monsoon (ASM) anticyclone is the most intense circulation pattern in the Northern Hemisphere in the upper troposphere and lower stratosphere (UTLS) during boreal summer (Mason and Anderson, 1963) which is forced by deep convection over South Asia (Hoskins and Rodwell, 1995). Intense monsoon convection can transport tropospheric tracers 10 (such as hydrogen cyanide (HCN) produced by biomass burning, carbon monoxide (CO), nitrogen oxides (NOx) or aerosols) from the lower troposphere into the UTLS of the ASM anticyclone or its edge Vogel et al., 2015;Tissier and Legras, 2016;Li et al., 2017;Fan et al., 2017b). Because of the strong dynamical confinement of the ASM anticyclone in the UTLS region in summer (Ploeger et al., 2015;Pan et al., 2016;Vogel et al., 2016), tropospheric tracer gases show a local maximum near the tropopause layer within the ASM anticyclone according to satellite measurements (e.g. Park et al., 2009;15 Randel et al., 2010;Vernier et al., 2015;Yan and Bian, 2015;Yu et al., 2017). At the edge of the ASM anticyclone, tropospheric tracers within the ASM anticyclone are transported outside, will affect trace gas concentrations in the UTLS result in significant changes in radiative forcings.
The ASM anticyclone is also an active region for stratosphere−troposphere exchange (e.g. Škerlak et al., 2014;Garny and Randel, 2016;Fan et al., 2017a). In-situ balloon measurements in August 2015 in Kunming, China combined with satellite 20 data and model simulations show that anthropogenic emissions from Asia play a significant role in the tropopause aerosol layer formation within the ASM anticyclone (Yu et al., 2017). Vogel et al. (2016) show that the northeastern flank of the ASM anticyclone is a region where air masses from the ASM anticyclone are separated from the anticyclone and are subsequently transported into the extra-tropical lower stratosphere. When air parcels enter into the stratosphere, they have the potential to impact the regional climate in Asia (Vernier et al., 2015;Gu et al., 2016). Furthermore, stratospheric intrusions occur at the 25 northeastern flank of the anticyclone and transport dry ozone-rich air into the troposphere over northern India. Stratospheric intrusions or tropopause folds have the potential to influence surface weather (e.g. monsoon deficit rainfall, Fadnavis and Chattopadhyay, 2017).
Balloon-borne measurements provided highly accurate water vapour and ozone profiles. Measurements with such balloon payloads have been carried out since 2009 in Lhasa and Kunming (Bian et al., 2012;Li et al., 2017), in Nainital (Brunamonti 30 et al., 2018), and in southern India and South America (Vernier et al., 2018). Low ozone values measured in the upper troposphere in Lhasa were present by Li et al. (2017), rapid vertical transport associated with typhoon convection lead to this phenomenon. The ozone profile in Li et al. (2017) also shows anomalies of high ozone values in the middle/upper troposphere over the Tibetan Plateau, however, they gave limited explanation about this phenomenon. It is important to investigate the ozone variation over the Tibet Plateau, in order to quantify the uncertainty of the radiative forcing from tropospheric ozone in 35 2 Atmos. Chem. Phys. Discuss., https://doi.org/10.5194/acp-2018-652 Manuscript under review for journal Atmos. Chem. Phys. Discussion started: 3 August 2018 c Author(s) 2018. CC BY 4.0 License. climate model. In this study, we combined these in-situ measurements with satellite data and trajectory calculations using the Chemical Lagrangian Model of the Stratosphere (CLaMS) model (McKenna et al., 2002;Pommrich et al., 2014) to analyse the high ozone structure found in the middle/upper troposphere in Lhasa over the Tibetan Plateau in August 2013. This paper is organized as follows: Sect. 2 describes the balloon sonde data, satellite data, and the CLaMS model. In Sect. 3, we present three case studies with enhanced tropospheric ozone in August 2013. A summary is given in the final section. The payload consists of an electrochemical concentration cell (ECC) ozonesonde (Komhyr et al., 1995) to measure ozone, 10 a cryogenic frost point hygrometer (CFH) (Vömel et al., 2007(Vömel et al., , 2016 to measure the frost (dew) point temperature for the temperature below (above) −15 • C, and a compact optical backscatter aerosol detector (COBALD, developed at the Swiss Federal Institute of Technology, Zürich) backscatter sonde to detect aerosol or ice cloud backscatter (Brabec et al., 2012).
An iMet radiosonde was used to transmit the CFH and Cobald data as well as to measure the ambient temperature, pressure, relative humidity (RH), and wind speed/direction. Further details about the different balloon flights during the SWOP campaign The relative humidity over ice (RH i ) from CFH is defined as where e is the water vapour pressure calculated from the frost point or dew point temperature and e sat is the saturated vapour pressure with respect to liquid water or ice, which is calculated from the ambient temperature using the Hyland−Wexler 20 equation (Hyland and Wexler, 1983) for liquid water and Goff−Gratch equation (Goff and Gratch, 1946) for ice water. The RH i uncertainty is 5% in the tropopause layer (Vömel et al., 2016).

Satellite data
The ozone monitoring instrument (OMI) is a nadir viewing near ultraviolet/visible charge−coupled spectrometer aboard the National Aeronautics and Space Administration's (NASA's) Earth observing system's Aura satellite (Levelt et al., 2006). In this study we use the TOMS−Like total column ozone level-3 product (OMTO3e) with horizontal resolution 0.25×0.25. The TOMS version 8 algorithm is used to extract the vertical column ozone data using only two wavelengths (317.5 and 331.2 nm).
The strong ozone absorption at 317.5 nm is used to derive total ozone, and the weaker absorb at 331.2 nm is used to estimate the effective surface reflectivity. The relative uncerntainty on OMI-TOMS product is less than 5%.
The atmospheric infrared sounder (AIRS) on NASA's Aqua satellite is on a sun synchronous polar orbit. The instrument 5 employs a cross-track scanning hyper-spectral infrared spectrometer with 2378 spectral channels (Aumann et al., 2003). It is designed to provide twice daily global data sets for different constituents and temperature. Here we use the AIRS Level-2 ozone and water retrieval product version 6.0 (Olsen et al., 2013).
CloudSat is designed to probe the vertical structure of clouds and precipitation using a cloud profiling radar (CPR), as a component of the A-Train (Marchand et al., 2008). The Cloudsat operational 2B Geometric Profile (2B-GEOPROF) data 10 product (Version R04) is used with 480 m vertical resolution. The vertical distribution of radar reflectivity is used to mark cloud layer. The echo mask values are greater than 20 dBZe indicating a false detection value below 16%.  Konopka et al. (2007). The trajectory has been used to focus on the transport process within or around the ASM anticyclone (Vogel et al., , 2018Ploeger et al., 2015). Dynamic fields from the European Centre for Medium-range Weather Forecasts (ECMWF) interim reanalysis (Era-Interim) (Dee et al., 2011) are used to drive the CLaMS 20 model. The input dynamic fields are recorded every 6 hours on a 1 • × 1 • horizontal grid with 60 hybrid vertical levels from the surface to 0.1 hPa. The vertical velocity on hybrid level is calculated using the diabatic heating budget including the cloud radiation, latent heat release, and mixing and diffusion (Ploeger et al., 2010). The trajectory model setup is the same as in Li et al. (2017).

25
The ozone profiles over Lhasa show a pronounced daily variation of ozone mixing ratios (OMR) between 340 K and 430 K from 4 to 27 August 2013 (Fig. 1). The lapse-rate tropopause is calculated from measured temperature profiles using the lapse-rate the OMR relative changes above 30%; in the following, we solely focus on these air masses. Unfortunately, RH i from CFH are not available above 335 K. Instead, the relative humidity (RH) from iMet was used, with useful data just below 350 K (above −40 • C). Below −40 • C, RH will not be used, because of the detection limit of the radiosonde humidity sensor.
The potential vorticity (PV) along the 30-day backward trajectories of air parcels within high ozone layer between 355 K and 363 K was shown in Fig. 2b as a function of time and potential temperature. According to their different pathways, backward 15 trajectories of air parcels could be divided into two clusters. Both of them experienced strong uplift processes with potential temperature increasing from 330 K to 360 K on 20 and 29 July. The tracks of 20-day backward trajectories of air parcels around 355−363 K isentropes initialized on 8 August 2013 are shown colour-coded by date in Fig. 2c. Air parcels started their ascent near the Himalayas and were uplifted to 360 K isentrope between 19 and 21 July and then moved horizontally to East Asia following the ASM anticyclone. Finally, the air parcels moved westerly around the anticyclone circulation before they arrived 20 in Lhasa on 8 August 2013. The transport time for the air parcels from the lower troposphere of the Himalayas to the upper troposphere over Lhasa is less than 20 days for the whole pathway. Figure 2d gives the boundary layer geolocation of air parcels (the same as in Fig. 2b and c), where they experience strong uplift through convection. The uplift rate of air parcel is defined as (θ t+δt − θ δt )/δt. When the uplift rate is greater than 9 K day −1 (an empirical value), the strong uplift process of air parcels will be recognized. The point that air parcels start 25 ascent is marked as the geolocation. We find that most of air parcels were from the Himalayas, where strong uplift occurred frequently. South Asia, the area adjacent to the Himalayas is usually a strong source region of air pollution, which is caused by natural (e.g. biomass burning) and anthropogenic processes (Cong et al., 2015). In addition, ozone is photochemically enhanced by reactions involving ozone precursors from biomass burning. After reaching the upper troposphere, the polluted air masses that were transported over long distances, made the best possible contribution to high tropospheric ozone over Lhasa The vertical variability of OMR, monthly mean ozone, temperature, the OMR relative change, RH i , and colour index on 11 August 2013 is shown in Fig. 3a. Positive ozone anomalies (Fig. 3a left) and the OMR positive variance (Fig. 3a right)    Air pollutants from South Asia or Lhasa may contribute to the positive ozone anomalies over Lhasa through photochemical production, however, due to the detection limit for chemical constituents that produces ozone, we can not answer this question conclusively. During the next days, this broken filament structure moves westward about 3,000 km. It arrives in Lhasa and is captured by  Fig. 4b−d), contributing to the positive ozone value measured in the UTLS region over Lhasa (the black rectangles in Fig. 1). The stratospheric intrusion with high ozone mixing ratios is also reflected in low water vapour values observed by AIRS at 346 hPa during its ascending track on 16 August 2013 ( Fig. 4e and   4f). Figure 5a shows the vertical variability of ozone, monthly mean ozone, temperature, RH i , colour index, and the OMR relative 5 change for 18 August 2013. The positive OMR relative change appears in the troposphere around 350 K and 363−373 K, and the lower stratosphere (Fig. 5a right). Two characteristic minima of OMR relative change occur between 352−357 K and 375−390 K on 18 August. RH i is anti-correlated with OMR anomalies on these isentropic surfaces. RH i near the tropopause (384 K) is greater than 100%. Indeed, an ice cloud layer was observed near the tropopause layer according to the colour index (CI>7) from COBALD backscatter measurements. Super-saturation is observed within the ice cloud.

10
In order to investigate in detail the variance of ozone profiles measured on 18 August 2013, the PV along the 50-day backward trajectories from the CLaMS model is displayed in Figure 5b. Parcels in the upper troposphere (363−373 K) originate from the dry intrusion layer. There is evidence for mixing processes that occurred between air parcels with high PV from stratosphere and low PV value from the troposphere, while air parcels in the middle troposphere (around 350 K) with high ozone and low water vapour originate from the thin intrusion layer. Air parcels around 350 K experienced a weak uplift during 13−14 15 August 2013. Figure  Positive ozone mixing ratio anomalies were also captured on 19 and 20 August 2013. The variability of ozone vertical structure is significant in the middle troposphere on 19 August. Ozone and water vapour show strong anti-correlation below 25 355 K. The OMR relative change shows large increase in the tropopause region (368−408 K), up to 90% (Fig. 6a). The 50-day backward trajectories of air parcels in Fig. 6b indicate diabatic decent transport process, especially the air cluster in the middle troposphere (between 347 K and 355 K). The PV along the trajectories of air parcels between 368 K and 380 K display high PV with values greater than 6 PVU. Figure 6c shows the tracks of backward trajectories of air parcels within high ozone in the middle and upper troposphere on 19 August. The stratospheric intrusion in the middle troposphere has the same transport 30 pathway as on 18 August and also experienced an uplift process around 14 August. Air parcels near the tropopause layer ( Fig.   6a) originated from the northern Hemisphere with high ozone and high PV. The equatorial regions contribute little to air parcels on 19 August compared to the case on 18 August. That is why the OMR relative change near the tropopause layer on 19 August is higher than OMR relative change on 18 August. The ozone vertical structure and RH i also show an anti-correlation in the middle troposphere on 20 August. In the troposphere, the OMR relative change near 355 K is higher than 30%. The OMR relative change on 20 August shows minor increases in the tropopause region (Fig. 7a) and is also weaker than on 19 August. The lapse rate tropopause height on 20 August is lower than on 18 and 19 August. Fig. 7b shows the PV along the backward trajectories of air parcels in the middle troposphere (355 K) on 20 August. The intrusion in the middle troposphere has the same transport pathway as the one on 18   5 August and also experienced an uplift process around 14 August (Fig. 7c).  (Fig. 8a). These parcels moved equatorward and arrived at the poleward edge of westerly wind jet two days later, where a tropopause folding occured (Fig. 8b). Parcels continue to move along the isentropic surfaces and cross the tropopause region from the polar lowermost stratosphere to the upper troposphere in the mid-latitude on 13 August 2013 at 12:00 UTC (Fig. 8c). Overall, it takes 2−3 days for air parcels to cross the tropopause. Both the pathway and timescale of transport are consistent with the analysis 15 of other deep stratospheric intrusions that occurred over North America (Langford et al., 1996;Vogel et al., 2011;Kuang et al., 2012;Lin et al., 2015) or Europe (Stohl and Trickl, 1999;Trickl et al., 2010Trickl et al., , 2014 associated with the polar jet stream. It is interesting that in our case, air parcels are affected by strong convection in the troposphere, after they are transported from the stratosphere downward into the upper troposphere. The strong convection lifted air parcels to high altitude, which can be seen from the CloudSat radar reflectivity (dBZe) (Fig. 8d). This uplift process can be seen clearly on 14 August 2013 in Fig. 5b. The 20 extra-tropical tropopause is located between the upper troposphere and lower stratosphere and intersects the isentropic surface, which acts as a dynamic barrier for tracer transport (Gettelman et al., 2011). Ozone in the extratropics exhibit large gradients in the UTLS. There is a net downward transport of ozone from the stratosphere to the troposphere along the isentrope at the poleward edge of the jets (Yang et al., 2016). The tropopause fold transports air parcels quasi-isentropically from the lowermost stratosphere with high ozone mixing ratio to the troposphere within the ASM anticyclone, contributing to high tropospheric However, the PV along the backward trajectories for the convective transport and the stratospheric intruison are different.
The PV values along the trajectories are less than 2 PVU for the convective transport case, and greater than 6 PVU for the stratospheric intrusion when air parcels are located in the extra-tropical lower stratosphere. The PV decrease when air parcel cross the lapse rate tropopause from the lower stratosphere to the troposphere. Tropical cyclones, which could transport marine