East Asian summer monsoon delivers large abundances of very short-lived organic chlorine substances to the lower stratosphere

Significance

Recent measurements onboard high-altitude research aircraft detected large abundances of very short-lived ozone-depleting substances (VSLSs) over the western Pacific in the outflow of East Asian summer monsoon convection. The measurements in the lower stratosphere above the East Asian monsoon are more than double the concentration of those previously reported from the same layer of the atmosphere in the tropics. These observations suggest that 1) the contribution of VSLSs to the stratospheric chlorine budget may be larger than currently estimated, and 2) when combined with intense emissions of VSLSs in east Asia, East Asian monsoon convection provides a significant yet overlooked pathway for these anthropogenic ozone-depleting substances to enter the stratosphere.

Abstract

Deep convection in the Asian summer monsoon is a significant transport process for lifting pollutants from the planetary boundary layer to the tropopause level. This process enables efficient injection into the stratosphere of reactive species such as chlorinated very short-lived substances (Cl-VSLSs) that deplete ozone. Past studies of convective transport associated with the Asian summer monsoon have focused mostly on the south Asian summer monsoon. Airborne observations reported in this work identify the East Asian summer monsoon convection as an effective transport pathway that carried record-breaking levels of ozone-depleting Cl-VSLSs (mean organic chlorine from these VSLSs ~500 ppt) to the base of the stratosphere. These unique observations show total organic chlorine from VSLSs in the lower stratosphere over the Asian monsoon tropopause to be more than twice that previously reported over the tropical tropopause. Considering the recently observed increase in Cl-VSLS emissions and the ongoing strengthening of the East Asian summer monsoon under global warming, our results highlight that a reevaluation of the contribution of Cl-VSLS injection via the Asian monsoon to the total stratospheric chlorine budget is warranted.

The Asian summer monsoon (ASM), a significant element of the climate system, has been studied as a regional weather pattern for centuries. Only in recent decades has its role in global constituent transport been recognized, largely owing to observations made from satellites (1). The ASM is of particular interest because its associated deep convective systems rapidly transport air masses from the planetary boundary layer (PBL) in one of the most polluted regions on the planet to the upper troposphere and lower stratosphere (UTLS). An annually recurring layer with a distinct chemical signature and aerosol composition forms at the tropopause level within the ASM UTLS anticyclone during boreal summer (2–5). Detailed information on the chemical and microphysical changes in the UTLS induced by this annually recurring transport process is necessary for the accurate representation of the role of the ASM in chemistry-climate models. Targeted measurements using high-altitude research aircraft, in addition to ground-based and balloon-borne observations, provide data essential for understanding physical processes and constraining chemistry-climate models. Two prior experiments provided valuable measurements of trace gases and aerosols in the region of the ASM anticyclone (6–8). The Asian summer monsoon Chemical and Climate Impact Project (ACCLIP), a field campaign conducted in summer 2022 using two high-altitude aircraft, the NSF National Center for Atmospheric Research (NCAR) Gulfstream V (GV) and the NASA WB-57, represents the latest large-scale effort of this kind. The ACCLIP campaign conducted a total of 29 research flights over the northwestern Pacific from July 31 to September 1, 2022, from an airbase in the Republic of Korea (see SI Appendix, Fig. S1 for flight tracks). The flights sampled air masses lofted by convection both inside and near the eastern edge of the ASM anticyclone. The observations include a large suite of trace gases and aerosols, mostly at altitudes between 12 and 19 km above sea level (asl), a layer not accessible by commercial aircraft. The data represent the largest set of in situ atmospheric composition measurements in the ASM region (see SI Appendix for further details).

The ASM system has two distinct components, the South Asian summer monsoon (SASM) and the East Asian summer monsoon (EASM). These components are controlled by different heat sources and respond differently to global warming. The circulation of the EASM is projected to strengthen while that of the SASM is projected to weaken, although both systems are projected to have increased precipitation due to increased moisture content (9). The schematic in Fig. 1A shows the key elements of the two subcomponents, highlighting two convergence zones: the Monsoon Trough and the East Asia Subtropical Front. Up to now, the SASM has been considered the dominant process for injecting PBL air masses into the ASM anticyclone at the tropopause level (6, 10–13). The Monsoon Trough was identified as the central region for convective lofting of PBL air masses (13). A key outcome of ACCLIP is the identification of the EASM convection as a more impactful transport pathway for UTLS chemical composition and stratospheric ozone chemistry than previously considered.

Fig. 1.

ASM subcomponents and the active East Asian monsoon in August 2022. (A) Schematic of Asian summer monsoon subcomponents (14, 15). Flow patterns are shown at three levels represented by near surface (1,000 hPa “Sfc Level”), low level (850 hPa), and upper level (200 hPa). The East Asia Subtropical Front is the major rainfall-producing system; it is located on the northwest flank of the Western Pacific Subtropical High and migrates seasonally from around 25°N in May to 40°N in August. From mid-June to mid-July, the front is generally located from the Yangtze River basin to southern Japan, and the corresponding rainy season is called Meiyu in China or Baiu in Japan (16). (B) August 2022 anomalies of the occurrence frequency of convection defined by infrared brightness temperature T_B < 235 K using GPM MERGIR satellite data (17). The criterion identifies convective cloud tops above ~12 km altitude. Anomalies are calculated with respect to the climatological (2006–2022) August mean occurrence frequency at each location and scaled by the SD (σ). The position of the East Asia Subtropical Front for August 2022 is highlighted (black dashed curve). Also shown are 150 hPa geopotential height from the NCEP Climate Forecast System (CFS) data (18) 14,350-m contours for the 44-y (1979–2022) August climatology (gray) and August 2022 (purple)

In this study, we report higher concentrations of PBL pollutants at the tropopause level than any previously measured in the UTLS and link these high values to convective lofting by deep convection in the EASM. Our measurements include large abundances of very short-lived chlorinated substances (Cl-VSLSs) from anthropogenic sources. Specifically, we report the measurements of five short-lived chlorine compounds: dichloromethane (CH₂Cl₂), chloroform (CHCl₃), 1,2-dichloroethane (C₂H₄Cl₂), tetrachloroethene (C₂Cl₄), and 1,2-dichloropropane (C₃H₆Cl₂). The first four of these species have been observed in the UTLS and are known to contribute to the stratospheric chlorine budget and ozone depletion (19). The fifth gas, 1,2-dichloropropane, is observed for the first time in the UTLS. The atmospheric concentrations of these species are generally of concern because of their increasing emissions in recent years (20–23). In reporting these observations, we bring further attention to the transport capacity of the EASM in a region of high emissions and the system's potential relevance to the recovery of the ozone layer in a changing climate.

Results

Active 2022 East Asian Summer Monsoon.

The 2022 ASM season was highly anomalous, marked by the historic flooding in Pakistan and an unprecedented heatwave over the Yangtze River Valley (24). Drastic changes in the ASM precipitation pattern were linked to the “triple-dip” La Niña-induced atmospheric circulation changes: the westward extension of the Western Pacific Subtropical High in the lower-mid troposphere and the northeastward expansion of the intensified ASM anticyclone in the upper troposphere (24, 25). While the circulation change created the drought conditions dominating the Yangtze River Valley, northern China and the Korean peninsula experienced abundant rainfall (26). Fig. 1B shows the anomalies of EASM convection in August 2022 scaled by the interannual variability using the monthly SD (σ). Here, the distribution of the deep convection anomaly is derived from satellite infrared (IR) brightness temperature (17) and is generally consistent with the observed precipitation anomaly (24, 26). The 2 to 3 σ positive anomalies highlight the active EASM convection in late summer along the East Asia Subtropical Front, which migrated northward from the location of the typical Meiyu front (16). Also shown in the figure is the shape and location of the anticyclone in August 2022 together with the 44-y climatology. Compared to the climatology, the 2022 anticyclone was enlarged and much more elongated, with its center shifted further east, such that it encompassed more of the East Asian convergence zone than in a typical season. The anomalies highlighted in Fig. 1B are the key elements of the circulation pattern change that contributed to the findings of the ACCLIP campaign.

To demonstrate the large-scale impact of the circulation pattern change and the active EASM convection on UT composition, we show, in Fig. 2, the 150 hPa distribution of carbon monoxide (CO) from the Microwave Limb Sounder (MLS) on the NASA Aura satellite (27). Chemically, UT CO is an ozone precursor and a major chemical sink of OH (28). The concentration of CO in this layer has a direct impact on the atmospheric oxidation capacity. In this work, we use CO mostly as a transport tracer to characterize the convection. CO is an effective tracer of convective transport of polluted PBL air because it has a relatively short chemical lifetime (1 to 2 mo) and strong emission sources from both human activities and biomass burning (13, 29). Because of common source regions and photochemistry in the PBL, a wide range of reactive species are expected to be elevated in the high-CO air mass (30, 31). Also shown in Fig. 2 are selected contours of the 150 hPa stream function, which characterizes the UT flow pattern. Two stream function contours are selected to mark the outer edge and the center of the ASM anticyclone. They are selected empirically based on the horizontal gradient of the stream function field, with the edge contour (4 × 10⁶ m²/s) identifying the region of maximum gradient and the center contour (25 × 10⁶ m²/s) surrounding the region of minimum gradient. The average August 2022 CO distribution and the location of the ASM anticyclone contrast strongly with the 17-y (2005–2021) average. The anomalous distribution shown here provides the large-scale chemical signature of the convective transport and the background for the airborne campaign.

Fig. 2.

Chemical signature of the strong EASM convective transport in 2022 from space. August average upper tropospheric carbon monoxide (CO) mixing ratios from satellite retrievals (MLS version 5 data at 147 hPa; see ref. 32 for details of MLS data quality and resolution) for (A) 2022 and (B) 2005–2021. Also shown are two selected contours of the 150 hPa stream function ([4, 25] × 10⁶ m²/s; based on the NCEP CFS data), which mark the outer edge and the center location of the ASM anticyclone.

EASM Convection as an Efficient Transport Pathway for PBL Pollutants to Reach the UTLS.

ACCLIP airborne observations sampled ASM-lofted air masses mostly downwind of the monsoon convection. The regions of convective transport that contributed to air masses sampled by the flights, based on Lagrangian back trajectory model calculations (Materials and Methods), are shown in Fig. 3A. The distribution shows that both EASM and SASM convection contributed to the airborne samples, but the EASM region shows a higher frequency of convective influence, as a result of both active seasonal convection and the proximity of the flights to the East Asia Subtropical Front. In Fig. 3B, in situ measurements of CO are used to characterize the level of pollution in the convection-influenced air masses from different regions. In general, the air masses influenced by EASM convection are associated with higher measured CO mixing ratios. To contrast the regional pollutants from the EASM and SASM, we calculated domain-average CO mixing ratios from the in situ CO measurements grouped into each domain using back trajectories (black boxes in Fig. 3B). The average CO mixing ratio identified with the EASM domain (130 ppbv) is about 35% higher than that identified with the SASM (95 ppbv).

Fig. 3.

Airborne observations of convective transport from SASM and EASM. (A) ACCLIP airborne campaign flight tracks (July 31 to September 1, 2022) and trajectory model derived convective influence (Materials and Methods). The distribution of convective influence for the airborne measurements at high altitudes (<200 hPa), estimated using trajectory calculations to connect the flight tracks with convection. Shaded pixels (1 × 1 degree) show the relative frequency of the colocation between convective clouds and back trajectories from the flight tracks (Materials and Methods). (B) The distribution of average CO mixing ratios for all airborne samples with convective influence within each pixel. The average CO for samples connected to the SASM and EASM domains (gray boxes) are approximately 95 and 130 ppbv, respectively. Note that the SASM average is consistent with the StratoClim measurements (33, 34). The blue and cyan lines show the flight tracks for the two research aircraft: the NASA WB-57 and the NSF/NCAR GV, respectively.

The much higher CO mixing ratios at the top of the EASM convection indicate that the EASM convergence zone is aligned with a far more polluted PBL. This is further demonstrated in Fig. 4, where vertical distributions of CO and CH₂Cl₂ are shown. The maximum 1-min average CO mixing ratio shown in Fig. 4A, ~ 320 ppbv near 15 km asl, is more than double the previous maximum high-altitude CO measured during the StratoClim campaign in the SASM region (~150 ppbv). The mean CO mixing ratio at this level, ~125 ppbv, is about 25% higher than the mean previously measured at this level in the SASM region (~100 ppbv).

Fig. 4.

Record-breaking CO and Cl-VSLS observed in the UTLS. (A) Vertical distribution of CO mixing ratios measured on the WB-57 (blue) and GV (cyan) aircraft (Materials and Methods) from all research flights (July 31 to September 1, 2022). The mean value and the 5th to 95th percentile for the 1-km layer near the level of zero clear-sky radiative heating (LZRH, magenta dashed line), approximated here by 360 K potential temperature (35), are marked by the filled black square and the thin black line, respectively. The average tropopause height is 16.5 km or ~380 K (red dashed line labeled TPH). The layer between the LZRH and the average tropopause is shaded in beige. The mean and range, represented by the 5th and 95th percentiles, for the StratoClim CO data (33) at the same levels are marked by the red vertical and horizontal lines, respectively. (B) Same as A but for dichloromethane (CH₂Cl₂) measured by WAS on WB-57 and TOGA and AWAS from the GV (Materials and Methods). For reference, the red symbols show the mean and range from previous measurements from the tropical tropopause region (see Tables 1-5 in ref. 19) and for the tropopause layer, 16.5 to 17.5 km. (C) Vertical distribution of total chlorine from WB-57 WAS measurements (Materials and Methods) of five primary Cl-VSLSs in tropopause-relative coordinates, adjusted by the mean tropopause height of 16.5 km (ref. 36 their equation 1). The red symbol marks the mean and range from previous measurements (see tables 1-5 in ref. 19). (D) A general positive correlation between CH₂Cl₂ and CO exemplifies the strong correlation between CO and many PBL pollutants measured during ACCLIP.

Record-breaking Observations of Very short-lived Chlorinated Ozone-depleting Substances.

Among the large number of pollutants transported in the high-CO PBL air masses, we focus on a group of species known as very short-lived ozone-depleting substances. This is a subgroup of ozone-depleting substances with local lifetimes shorter than ~0.5 years (19, 37, 38). In particular, we focus on the chlorinated VSLSs (Cl-VSLSs). Unlike long-lived ozone-depleting substances, VSLSs are not controlled under the Montreal Protocol. Nevertheless, these substances are under close watch and of increasing concern in recent years because many Cl-VSLSs originate primarily from anthropogenic sources, and their emissions have been rapidly increasing in recent decades. As a result, their contribution to stratospheric chlorine has increased by 61% between 2000 and 2017 (39). The increase is largely attributed to dichloromethane (CH₂Cl₂), the dominant component of Cl-VSLSs, comprising 70% of the Cl-VSLSs contribution to stratospheric chlorine (21, 37, 39). There are signs that the increasing VSLS levels may have impacts on the decrease of stratospheric chlorine. For example, between 2004 and 2017 the increase in Cl-VSLSs is estimated to have slowed the decline of upper stratospheric HCl (i.e., the major chlorine species in that region) by about 15% (39). Recent airborne observations have identified the impacts of ASM transport on the elevated Cl-VSLS levels in the UTLS (8, 40, 41).

Extremely high values of CH₂Cl₂ were observed near the tropopause in the region of East Asia during ACCLIP (Fig. 4B). Their vertical structure is similar to that of the CO distribution (Fig. 4A). Indeed, there is a strong positive correlation between CH₂Cl₂ and CO in the ACCLIP observations, as shown in Fig. 4D. The CH₂Cl₂ mixing ratios observed in the peak convective outflow layer near 15 km are as high as 600 pptv. This represents the highest value observed at the near-tropopause level; they are more than double those previously observed in the UT (8, 40). These strongly enhanced abundances arise from a combination of active deep EASM convection and the alignment of the convergence zone with intense Cl-VSLS emission sources. The occurrence in August 2022 of large anomalies in deep convection over the highly industrialized regions of Shandong and Hebei in the North China Plain, known to be among the largest CH₂Cl₂ emission sources in the world (22), is shown in Fig. 1B.

Injection of Cl-VSLSs to Layers above the Level of Zero Clear-sky Radiative Heating and the Tropopause.

Theoretically, the level of zero clear-sky radiative heating (LZRH) marks the change from net radiative cooling below to net radiative heating above, which indicates the change of air mass large-scale vertical motion from subsidence to ascent (35). In Fig. 4 A and B, the layer between 360 K and 380 K potential temperature (θ) is shaded to highlight the measurements below the average tropopause, ~16.5 km or ~380 K, but above the estimated LZRH, ~360 K. The “C”-shape of the CO vertical distribution and the observed CO maximum of 320 ppbv near θ ~ 365 K show that the level of the main EASM convective outflow is above the LZRH. This leads to the important recognition that although the EASM convective transport did not reach the tropopause [defined by the temperature lapse rate, see WMO (42)], it lifted PBL pollutants to levels above the LZRH, where air masses undergo rising motion and are expected to enter the stratosphere.

Next, we quantify the concentration of total VSLS-sourced chlorine observed in the 1-km layer above the tropopause using tropopause-relative altitude coordinates (Fig. 4C). In this analysis, only WB-57 WAS data are used since the GV mostly sampled in the troposphere. Each sample’s tropopause-relative coordinate is derived in reference to the nearest tropopause height determined from the WB-57 in situ sensor temperature data (Materials and Methods). The coordinate is then adjusted to the average tropopause height among all samples (36). The average in situ data-based tropopause altitude is 16.4 km and 378.3 K in θ. In Fig. 4C, the total chlorine includes the contributions from the five species listed in the Introduction: dichloromethane (CH₂Cl₂), chloroform (CHCl₃), 1,2-dichloroethane (C₂H₄Cl₂), 1,2-dichloropropane (C₃H₆Cl₂), and tetrachloroethene (C₂Cl₄). Among the 16 Cl-VSLSs observed during ACCLIP (SI Appendix, Table S1), these top five species have measured median concentrations above 1 pptv. The vertical distributions of these five Cl-VSLSs are shown in SI Appendix, Fig. S2.

In general terms, both StratoClim and ACCLIP campaign data show that transport through the ASM is a significant pathway for stratospheric chlorine injection. However, the ACCLIP Cl-VSLS measurements highlight the importance of the EASM in that process. In Fig. 5, we provide relative-frequency distributions of total organic chlorine from Cl-VSLS source gases in the 1-km layer above the tropopause (LS) and the layer between the tropopause and the LZRH (UT), approximated by 360 to 380 K, together with the distributions of additional measurements from recent high-altitude aircraft campaigns, namely POSIDON data from 2016 (40) and StratoClim data from 2017 (8). Also given in the Fig. 5 insets are values of mean, median, range (defined as mean ± 1σ), and maximum for these datasets in both layers. The mean and range of ACCLIP observations in the LS layer are approximately a factor of 2 larger than those of the previous observations near the tropical tropopause reported in the WMO Ozone Assessment (19) (their Tables 1-5). They are ~50% larger than the StratoClim findings. In the UT layer, the ACCLIP mean chlorine value is about a factor of 4 larger than the tropical measurements and about a factor of 2 larger than the StratoClim data.

Fig. 5.

Total VSLS-sourced organic chlorine. The relative frequency distributions of total organic chlorine from five Cl-VSLS source gases (dichloromethane, chloroform, 1,2-dichloroethane, 1,2-dichloropropane, and tetrachloroethene) from ACCLIP measurements (orange, right-hand y axis) in two layers: (A) within 1 km above the tropopause, representing the lower stratosphere (LS), and (B) 360 to 380 K potential temperature, representing the upper troposphere (UT) between the LZRH and the tropopause. The ACCLIP UT data in B were obtained by two instruments: TOGA-TOF (shaded) and AWAS/WAS (striped) (Materials and Methods). The LS data in A are from the WB-57 WAS measurements (Materials and Methods). Also shown for comparison in both (A) and (B) are WAS measurements over the tropical western Pacific during the October 2016 POSIDON campaign (gray; note that 1,2-dichloropropane was not included) and WAS measurements mostly sampling the SASM contribution during the summer 2017 StratoClim campaign (purple, sharing gray axis with POSIDON; note that 1,2-dichloropropane and tetrachloroethene were not included). The estimated contribution of Cl-VSLS source gases to the stratospheric chlorine budget in 2020 used in the 2022 WMO Ozone Assessment (105 ± 20 ppt) is shown in A (black). Mean, median, range (mean ± 1 σ), and maximum values of the total Cl-VSLSs for each dataset are given in the insets. Note that in (A) the POSIDON samples (106 cans) are selected within the 16.5 to 17.5 km layer, whereas the ACCLIP LS samples (69 cans) are selected within 1 km above the tropopause as determined using the nearest WB-57 in situ measurements. StratoClim samples (29 cans) are selected within 1 km above the tropopause using the ERA5 tropopause product (43).

The Asian Monsoon Tropopause Structure and the Entry Point to the Stratosphere.

Past model estimates of stratospheric chlorine entry values have typically used high-altitude aircraft measurements made at the tropical tropopause, i.e., at 16.5 to 17.5 km altitude and 20°S to 20°N latitude (39). The rationale is that tropospheric air primarily enters the stratosphere through the tropical tropopause. However, recent progress in the understanding of ASM UTLS dynamics and transport indicates that the ASM (rather than the tropical) tropopause is the dominant stratosphere entry point during boreal summer (1, 12, 13, 44, 45). Dynamically, the ASM region has the highest tropopause globally during the season (12, 46, 47). While the tropopause within the ASM anticyclone is higher than 380 K on average, the equatorial tropical tropopause is at or below 375 K. The monthly mean large-scale tropopause structure for the ACCLIP campaign season is shown in Fig. 6A.

Fig. 6.

August 2022 tropopause structure in potential temperature and the trajectory model diagnosis of the UT air mass entering the stratosphere. (A) The August 2022 monthly mean tropopause in potential temperature (color shaded contours) based on the ERA5 tropopause product (43). Also shown are the ERA5 horizontal winds at the 380 K level (white overlaid arrows). The rectangle (black) indicates the nominal ACCLIP flight region (Fig. 3). (B) The relative frequency map of the air mass tropopause crossing locations for all WB-57 WAS samples in the UT layer (between 360 K and the tropopause) (Materials and Methods). The WAS sampling locations are distributed along the WB-57 flight tracks (blue lines). To highlight the relationship between tropopause crossing frequency and the tropopause structure associated with the ASM anticyclone shown in panel A, the August 2022 monthly mean tropopause potential temperature at 380 K is shown (black contour). (C) Histogram of the fractions of trajectories crossing the tropopause within 20 d and their transit time distributions: the total fraction of sampled air masses that crossed the tropopause (gray), the fraction that followed the anticyclonic flow and entered the tropical lower stratosphere (red), and the fraction that entered the mid- to high-latitude lowermost stratosphere (orange). The tropical/extratropical separation is approximated by whether a trajectory’s tropopause intercept was south or north (respectively) of the WAS measurement point at which a given trajectory was initialized.

On daily time scales, the anticyclonic flow at the tropopause level acts to produce large-scale stirring, enabling convectively lofted air at the near-tropopause level to mix with the surrounding lower stratospheric air via wave breaking, similar to the situation described by Holton et al. (48) for the extratropical lowermost stratosphere. This mixing process effectively transports chemical constituents in the monsoon upper troposphere into the surrounding lower stratosphere quasi-isentropically. A 40-y climatology constructed using trajectory calculations shows that UT air masses at the eastern edge of the ASM over the western Pacific enter the lower stratosphere within time scales of a week (49). To verify this behavior for August 2022, we have analyzed 20-d forward trajectories, with the vertical velocity based on diabatic heating, from WB-57 WAS samples collected in the UT layer between the LZRH and the local tropopause. The result (Fig. 6 B and C) shows that nearly 90% (~74%) of the sampled air masses enter the stratosphere within 20 (5) d. About 70% follow the anticyclonic flow to the tropical lower stratosphere and ~20% to the lowermost stratosphere in the Northern Hemisphere middle to high latitudes.

This convective transport behavior is well analyzed in a model study using synthetic regional tracers (44). Those authors estimated that nearly 20% of tropical lower stratospheric air is of Asian PBL origin during late boreal summer and fall. Therefore, to obtain an approximation of the impact of enhanced Cl-VSLSs measured during ACCLIP, we assume a 20% ASM contribution to the tropical lower stratosphere airmass and that the mean Cl-VSLS concentration in the layer above the LZRH is representative of a chemically aged Asian PBL. A mean of 500 pptv total VSLS-sourced organic chlorine during ACCLIP represents a 395 pptv greater source gas injection than the 105 pptv VSLS-sourced chlorine near the tropical tropopause reported in the 2022 WMO Ozone Assessment (19). This difference results in nearly 80 pptv more organic chlorine in the tropical lower stratosphere during summer and fall. Furthermore, a Lagrangian model study (45) estimates that the tropical lower stratosphere contains 5% of ASM air masses on an annual basis. Applied to the ACCLIP data, the 5% contribution of seasonal ASM transport would add 19 pptv to the organic chlorine in the tropical stratosphere annually, an 18% increase to the tropical lower stratosphere over that based solely on the tropical Cl-VSLS value (105 pptv) from the 2022 WMO Ozone Assessment (19). The key message is that the VSLS concentrations in the ASM UTLS should be used to constrain and verify model estimates of the amount of Cl-VSLSs injected into the stratosphere during the ASM season.

Discussion

Stratospheric ozone is a key component of the Earth’s atmosphere and has direct links to the climate system. The success of the Montreal Protocol has led to the beginning of stratospheric ozone hole recovery (50). There is concern, however, that a rapid increase of anthropogenically sourced VSLSs, unregulated by the Montreal Protocol, could slow ozone recovery (38) and, together with naturally emitted halogenated VSLSs, induce stratospheric ozone loss that in turn has implications for future climate change (51, 52). The observations we report provide an important piece of information for quantifying the Cl-VSLS contribution to the stratospheric chlorine budget. The ACCLIP campaign results identify the EASM to be a significant transport pathway relevant to stratospheric chemistry and the ozone layer, as it facilitates rapid injection of a large amount of PBL pollution into the stratosphere.

Globally, the largest emission sources of CH₂Cl₂ are known to be in South and East Asia (53). CH₂Cl₂ emissions are reported to have increased by 85% between 2006 and 2017, mostly due to increasing Asian emissions (21). It has also been hypothesized that the emissions from South Asia and East Asia have strong potential to be injected into the stratosphere, considering the depth of Asian monsoon convection (22, 39, 53, 54). While previous Asian monsoon observations have shown clear evidence of SASM contributions (8, 40), ACCLIP observations not only provide further evidence for this hypothesis but also highlight the key role of the EASM in injecting very short-lived ozone-depleting substances to the stratosphere.

The convection anomalies in 2022 reflect both temporal and spatial variations to the precipitation pattern, specifically the increased precipitation late in the ASM season and the anomalous northward migration of the convergence zone. In a typical year, active convection occurs earlier in the season, and the convergence zone (and thus strong deep convection) lies at lower latitudes (55). However, even then the climatological convergence zone still crosses intense Cl-VSLS sources along coastal east Asia, and emissions in those regions show signs of an increasing trend (22, 23). Therefore, although ACCLIP observations were made during an anomalous ASM season, large convective transport contributions from the EASM can be expected to occur even in less meteorologically extreme seasons.

The observations also underscore the need to examine the representation of deep convection and associated constituent transport in the current generation of global chemistry-climate models. Estimating VSLS source gas injection using models (39) requires evaluation and constraints. The ACCLIP campaign provided a unique set of dynamical and chemical observations for assessing the ability of models to accurately represent the injection of reactive chemical species into the stratosphere.

Adequate model representation of the EASM is also important for 21st-century chemistry-climate projections because the EASM is predicted to strengthen in a warming climate (56, 57). Globally, among the eight regional monsoons, the EASM is projected to have the highest sensitivity with respect to global warming (57). Stronger EASM convection will lead to higher Cl-VSLS contributions to stratospheric ozone-depleting substances, and subsequently ozone depletion, which need to be considered as a scenario in climate change predictions. The SASM is also situated above the largest Cl-VSLS source region and projected to have high sensitivity to the global warming trend (57). This work, therefore, provides evidence of monsoon–climate change linkages through enhanced Cl-VSLS injection and stratospheric ozone chemistry.

Materials and Methods

In Situ Measurements of CO and Temperature.

CO mixing ratios were measured on both the GV and the WB-57 aircraft. The WB-57 CO measurements are provided by three different sensors: COLD2 (Carbon Oxide Laser Detector 2; ref. 58), COMA (Carbon monOxide Measurement from Ames), and ACOS (NOAA Airborne Carbonic Oxides and Sulfide Spectrometer). The data from the three sensors are in good agreement and consistent in supporting the presented result. For the purposes of this study, the COLD2 CO data are displayed in Fig. 4.

The instrument COLD2, installed inside the right spearpod of the WB-57, is a midinfrared quantum cascade laser spectrometer based on the direct absorption technique in combination with a multipass cell. The instrument provides in situ CO mixing ratios with a time resolution of 1 s and an accuracy of 4%, without need of in-flight calibration. COLD2 in-flight sensitivity and precision are proved to be 1.5 ppb, increasing to about 900 ppt for 10-s integration time, and the 1-h long-term stability is ±2%. The CO values are inferred, during a postflight analysis, by the transmission spectra acquired during the flight.

The GV CO instrument operated by NCAR is an Aerodyne quantum cascade laser spectrometer (59) that utilizes direct absorbance of infrared radiation near 4.6 µm to quantify CO and nitrous oxide (N₂O). The instrument has an optical path length of 76 m, and the free space optics are continuously purged with dry nitrogen from which CO has been scrubbed to contain less than 1 ppbv. Hourly in-flight calibration checks confirmed the accuracy of flight data. The ACCLIP observations have a time resolution of 1 s, a precision of 0.3 ppbv, and an uncertainty of ±1 ppbv.

In situ static temperature data measured on the WB-57 are used in this work to calculate sampled tropopause height. The measurements are made by the NASA Ames Meteorological Measurement System (MMS), which provides calibrated static pressure, static temperature, and 3D winds at either 1 Hz or 20 Hz sampling rate. The 1-s time resolution data are used for this work. The temperature sensing element is NIST traceable, and inflight adiabatic compression compensation is determined from 3D wind calibration. Since 1984, the MMS has operated on several airborne platforms: the ER-2, Global-Hawk, DC-8, Sierra UAV, and AlphaJet. Based on the analyses of Scott et al. (60) and Gaines et al. (61), MMS static temperature has an accuracy of ±0.3 K at all altitudes, including the critically important UTLS region studied during ACCLIP.

Measurements of Chlorinated VSLSs by the TOGA-TOF, WAS, and AWAS.

The NCAR Trace Organic Gas Analyzer with a TOFWERK high-resolution Time-of-Flight mass spectrometer (TOGA-TOF) was deployed on the NSF/NCAR GV during ACCLIP and provided measurements of Cl-VSLSs including CH₂Cl₂, CHCl₃, 1,2-dichloroethane, 1,2-dichloropropane, and C₂Cl₄. The TOGA-TOF is a fast online gas chromatograph/mass spectrometer (GC/MS) that sampled and processed in situ air for 35 s every 2 min during ACCLIP GV flights. Operational details of the TOGA system have been described previously (62). The TOGA-TOF was calibrated in flight during ACCLIP, as well as in the laboratory prior to and following the deployment, using a series of ppb-level gas mixtures (Apel-Reimer Environmental, Inc.) diluted to typically observed mixing ratios using a clean-air generator.

Whole air samplers (WAS) were deployed on both the NSF/NCAR GV and the NASA WB-57 aircraft. The GV instrument (Advanced Whole Air Sampler, AWAS) collected samples into stainless steel canisters under computer control. Three sample modules contained 12 canisters each, and two sample modules contained 10 samples each. The latter sample modules had a different geometry but similar volume (1.3 L), and they were previously deployed in the NASA Global Hawk Aircraft (63). Samples were pressurized to 40 to 45 psia using two 2-stage metal bellows pumps. Sample collection times varied with altitude and module type, varying from about 4 to 60 s for the AWAS 12-can modules, with an average of about 26 s at the 12 to 15 km altitude range. The sample collection time for the 10-can modules was approximately 40 s and was nearly constant over the 12 to 15 km range. The WB-57 WAS instrument deployed 50 canisters per flight, as described in ref. (40). The sample collection time varied from about 14 s at 10 km to 120 s at 20 km, with an average of about 44 s between 14 km and 18 km (~360 to 400 K potential temperature). All sample canisters were precleaned in the field by repeated flushing/evacuation with wet nitrogen gas. After final evacuation, approximately 15-Torr water vapor was added to each canister to help passivate the canister interior surface. All samples from both aircraft were sent to the home laboratory at the completion of the field mission. Analyses took place from mid-October 2022 to mid-January 2023. Trace gas calibrations were based on in-house and commercial standards as described in refs. 64–66.

Trajectory Modeling and Satellite Data–based Convective Influence.

To relate the ACCLIP observations to convective transport, a trajectory model is used to calculate 3D backward trajectories of idealized air parcels from the measurement locations and times. In the analysis shown in Fig. 3, “Convective influence” is identified when the air parcel’s back trajectory is colocated with convection. Specifically, back trajectories are initialized in 1-min time intervals along the GV and WB-57 flight tracks and calculated with the TRAJ3D model (67) using three-dimensional kinematic wind fields from ERA5 reanalysis (68). Each 1-min trajectory initialization features 75 individual trajectories to account for trajectory path uncertainty. These are initially arranged in a 5 × 5 × 3 grid centered on the aircraft location, with a total dimension of 0.1 degrees latitude × 0.1 degrees longitude × 0.2 hPa. To investigate the ASM air mass contribution to the UTLS layer, trajectories are only included for aircraft measurements taken at pressures less than 200 hPa. Convection is represented by a convective cloud top altitude product, which is produced at 3-h intervals with a resolution of 0.25° × 0.25°. The method of constructing the database is described in ref. 69. Each convective influence case is identified by the first (i.e., most recent) “encounter” of a trajectory with convection within the previous 20 d, as determined by having an altitude lower than the colocated convective cloud top altitude derived from satellite observations. This approach is analogous to that used in refs. 70 and 71 for airborne campaigns over the tropical western Pacific.

To diagnose the tropopause crossing of air masses sampled in the UT (Fig. 6 B and C), forward trajectories are calculated using the TRAJ3D model and the ERA5 radiative temperature tendencies for vertical motion. 3D forward trajectories of idealized air parcels from the measurement locations and times of all WB-57 WAS sampling points within the UT layer (between the 360-K isentrope and the tropopause) are calculated for 20 d. Similar to the convective influence analysis, here the tropopause crossing is identified by colocation of the air parcels’ forward trajectories with the tropopause, which is based on the ERA5 tropopause data product (43).

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

NCAR GV CO and VSLSs data can be ordered from the NCAR EOL ACCLIP Long Term Archive: https://data.eol.ucar.edu/project/ACCLIP. Analyses in this work used TOGA merge, AWAS/WAS merges, and 60-s merge: for TOGA-TOF merge: https://doi.org/10.26023/9TJ3-DT16-VZ00; for AWAS merge: https://doi.org/10.26023/2M8D-9C0V-AV0J; and for 60-s merge: https://data.eol.ucar.edu/dataset/620.053. NASA WB-57 data are available from NASA ESPO archive: https://www-air.larc.nasa.gov/cgi-bin/ArcView/acclip?WB57=1. Climate Forecast System (reanalysis and V2) data are available at the National Centers for Environmental Information archive: https://www.ncei.noaa.gov/products/weather-climate-models/climate-forecast-system. The GPM MERGIR satellite data are available here: https://disc.gsfc.nasa.gov/datasets/GPM_MERGIR_1/summary?keywords=Merged%20IR. DOI: 10.5067/P4HZB9N27EKU. The convective cloud top data can be accessed here: https://bocachica.arc.nasa.gov/nasaarc_cldalt/. ERA5 Tropopause Data (43) are available here: https://datapub.fz-juelich.de/slcs/tropopause/index.html. ERA5 products can be accessed from the ECMWF Climate Data Store: https://cds.climate.copernicus.eu/cdsapp#!/search?text=ERA5&type=dataset.