Role of Stratospheric Processes in Climate Change: Advances and Challenges

In this review, instead of summarizing all the advances and progress achieved in stratospheric research, the main advances and new developments in stratosphere-troposphere coupling and stratospheric chemistry-climate interactions are summarized, and some outstanding issues and grand challenges are discussed. A consensus has been reached that the stratospheric state is an important source of improving the predictability of the troposphere on sub-seasonal to seasonal (S2S) time scales and beyond. However, applying stratospheric signals in operational S2S forecast models remains a challenge because of model deficiencies and the complexities of the underlying mechanisms of stratosphere-troposphere coupling. Stratospheric chemistry, which controls the magnitude and distribution of many important climate-forcing agents, plays a critical role in global climate change. Convincing evidence has been found that stratospheric ozone depletion and recovery have caused significant tropospheric climate changes, and more recent studies have revealed that stratospheric ozone variations can even exert an impact on SSTs and sea ice. The climatic impacts of stratospheric aerosols and water vapor are also important. Although their quantitative contributions to radiative forcing have been reasonably well quantified, there still exist large uncertainties in their long-term impacts on climate. The advances and new levels of understanding presented in this review suggest that whole-atmosphere interactions need to be considered in future for a better and more thorough understanding of stratosphere-troposphere coupling and its role in climate change.


Introduction
Benefiting from continuously increasing amounts of data and more complex climate models, significant progress has been made during the past several decades in understanding Earth's climate, both past and future. A series of climate change assessment reports produced by the Intergovernmental Panel on Climate Change (IPCC) confirm that increases in greenhouse gases (GHGs) and aerosols have been dominant external forcings of climate change during the last two centuries. However, the magnitude and significance of observed climate changes are still under wide debate, and our physical understanding of many components of the climate system and their role in climate change remains incomplete (IPCC, 2001(IPCC, , 2014(IPCC, , 2021. Among the roles of the various components of the climate system in climate change, that of stratospheric processes has been one of the hottest topics in the past three decades. In ※ This paper is a contribution to the special issue on National Report to the 28th IUGG General Assembly by CNC-IAMAS (2019−2022). 1992, a project named Stratosphere-Troposphere Processes and their Role in Climate (SPARC) was founded as a core project of the World Climate Research Programme. This project aims to understand how atmospheric chemical and physical processes in the stratosphere interact with Earth's climate system. After 30 years of implementation of this project, the importance of the stratosphere in modulating and regulating climatic variability and trends is now widely recognized, and some great advances have been achieved despite grand challenges still remaining (SPARC, 2022). In a recent comprehensive research overview by Baldwin et al. (2019), the progress made in understanding the stratosphere and mesosphere in the past 100 years was presented with a large coverage of various directions and aspects in this area. Meanwhile, some other recent review papers have presented more detailed descriptions of the advances made on some specific topics or issues (e.g., Kidston et al., 2015;Butler et al., 2019;Baldwin et al., 2021, Haynes at al., 2021Butchart, 2022). After 100 years of research, our understanding of the dynamical and chemical processes in the stratosphere has advanced greatly. In addition, it is now widely accepted that stratospheric processes have important impacts on tropospheric weather and climate on various time scales, and climate models with a well-resolved stratosphere can give more accurate climate predictions.
Despite the significant progress in understanding stratospheric processes, their role in climate changes remains an issue of debate. The pathways, time scales and underlying mechanisms through which stratospheric processes impact tropospheric weather and climate still need to be further elucidated. In this review, we do not intend to detail all advances and progress made in stratospheric research; rather, we will focus on the coupling between the stratosphere and troposphere and discuss how and to what extent stratospheric processes modulate tropospheric weather or shape climatic variability and trends. In doing so, we also intend to highlight issues of current scientific interest in stratospheric research. In section 2, the impacts of thermodynamic and dynamic processes on weather and climate change are discussed. In section 3, the impacts of stratospheric chemical processes on the troposphere are presented. A summary and concluding remarks are given in section 4.

Role of thermodynamic and dynamic processes
It is well-known that the variations in the stratospheric circulation are strongly governed by the tropospheric processes below. However, the stratosphere is far from being a passive bystander to tropospheric influences. In fact, a twoway coupling exists between the stratosphere and troposphere. The diverse variabilities of the stratospheric circulation and thermal state can exert an influence on the troposphere below at different time scales. Therefore, knowledge of the stratospheric states has been identified as important for improving the predictability of the troposphere on sub-sea-sonal to seasonal (S2S) timescales and beyond (e.g., Sigmond et al., 2013;Rao et al., , 2020aDomeisen et al., 2020a;Huang et al., 2021). In addition to its role in S2S forecasting, a consensus has been reached by the scientific community that stratospheric processes should be well represented in climate models in order to accurately evaluate future climate change (e.g., Baldwin et al., 2019), thereby highlighting the importance of such processes in climatic variability.

Synoptic and S2S timescale impacts
The largest dynamical variability at S2S timescales occurs in the polar stratosphere where the stratospheric polar vortex (SPV) exists. The SPV is formed mainly through radiative cooling. More specifically, the strong contrast in stratospheric air temperatures between the poles and low latitudes leads to the formation of a circumpolar vortex in both hemispheres. The SPV in the Northern Hemisphere exhibits a larger variability compared to its counterpart in the Southern Hemisphere, due to stronger wave forcing that is closely related to larger mountain ranges and greater land-sea contrast in the Northern Hemisphere (e.g., Held et al., 2002;Rao and Ren, 2020). The strength and position of the SPV is largely influenced by planetary-scale Rossby waves, which propagate vertically into the polar stratosphere and break (Charney and Drazin, 1961;Matsuno, 1971). In extreme cases, the effects of planetary wave breaking lead to a dramatic weakening of the SPV with a reversal of the climatological westerlies, which is well-known as sudden stratospheric warming (SSW; Schoeberl, 1978;Quiroz, 1986;Baldwin et al., 2021). SSW occurs more frequently in the Northern Hemisphere than in the Southern Hemisphere; for instance, it occurs roughly six times per decade in the Northern Hemisphere (e.g., Charlton and Polvani, 2007;Hu et al., 2014;Butler et al., 2015), with more events in January and February, whereas only one major event (in September 2002) has been observed since the modern satellite era in the Southern Hemisphere (Krüger et al., 2005). Note that in early September 2019 there was a similar event in the Southern Hemisphere, but it did not meet the established criteria for a major SSW (Rao et al., 2020b).
As the most dramatic form of SPV variability, SSW events have important impacts on tropospheric weather and climate, particularly in the Northern Hemisphere. The stratospheric anomalies associated with SSW extend downwards into the lowermost stratosphere, where they last for one to two months on average, and thence to the surface, affecting tropospheric weather (Baldwin and Dunkerton, 2001;Hitchcock and Simpson, 2014;Karpechko et al., 2017;Zhang et al., 2019b). Composites of the Northern Annular Mode (NAM), as presented in previous studies (e.g., Dunkerton 2001, Hitchcock andSimpson, 2014;Liang et al., 2022b;Lu et al., 2022), remain the common approach to diagnosing the coupling between the stratosphere and troposphere following SSW. Here, we update these composite results with 46 SSW events based on ERA5 reanalysis data, as shown in Fig. 1. The canonical tropospheric response to SSW is a negative phase of the NAM, corresponding to an equatorward shift of storm tracks, severe cold-air outbreaks, and heavy snowfall over most of the midlatitude Northern Hemisphere (Baldwin and Dunkerton, 2001;Charlton and Polvani, 2007;Kolstad et al., 2010;Wang and Chen, 2010;Yu et al., 2015Yu et al., , 2018Kretschmer et al., 2018;Huang and Tian, 2019;Domeisen and Butler, 2020;Huang et al., 2021).
Despite corroborative evidence that the SPV can exert an impact on tropospheric weather, the dynamical mechanisms by which circulation anomalies associated with SSW induce changes to surface weather are still under investigation. Several mechanisms have been proposed in the literature, but no consensus has been reached. Some studies have argued that SSW affects tropospheric weather mainly through modulating the Arctic Oscillation or North Atlantic Oscillation (NAO) in the troposphere (Baldwin and Dunkerton, 2001;Charlton and Polvani, 2007;Scaife et al., 2016;Baldwin et al., 2021), whereas others have reported that SSW can induce changes in tropospheric weather regimes, such as tropospheric blocking and eddy-driven jets (e.g., Maycock et al., 2020;Huang et al., 2017), as well as the hemispheric transport of cold airmasses at mid-high latitudes (Huang et al., 2021), and hence changes in weather patterns. In addition, previous studies have linked the mass or pressure changes at the surface to SPV variations (e.g., Yu et al., 2014;Yu and Ren, 2019;Zhang and Tian, 2019). Other mechanisms have been proven to be of importance to the dynamical coupling between the stratosphere and troposphere, such as tropospheric modulations of stratospheric potential vorticity anomalies (e.g., Hoskins et al., 1985;Hartley et al., 1998), the "downward control" principle (e.g., Haynes et al., 1991), stratospheric influence on tropospheric baroclinic instability (e.g., Wittman et al., 2007), and stratospheric reflection of planetary waves vertically propagating from the troposphere (e.g., Perlwitz and Harnik, 2003). However, none of the abovementioned mechanisms is always active in the coupling between the polar stratosphere and troposphere.
The impacts of the SPV on tropospheric weather and climate are by now well recognized, and so the forecasting community has suggested that SSW provides useful information for improving the forecasting skill for surface weather in the Northern Hemisphere on S2S timescales (e.g., Sigmond et al., 2013;Rao et al., , 2020aDomeisen et al., 2020a;Huang et al., 2021), and even on synoptic timescales (Huang et al., 2022). For a comprehensive discussion on how our improved understanding of the state of the stratosphere and its dynamical links to the troposphere has helped to enhance the predictability of the troposphere on S2S timescales, readers are referred to the review by Butler et al. (2019). However, applying this skill in forecast models, including S2S and operational models, remains a challenge, partly because of the complexity of the physical mechanisms responsible for the impacts of SSW on surface weather in the northern extratropics (see above discussion). Another reason is that robust forecasting of the stratosphere on subseasonal timescales has not yet been achieved (e.g., Domeisen et al., 2020b). More recently, the SNAPSI (Stratospheric Nudging And Predictable Surface Impacts) project has been initiated by the SPARC community (Hitchcock et al., 2022), which aims to further our understanding of the role of SPV disturbances in the surface predictability of S2S models. The dynamics of the SPV and its corresponding influences on the troposphere have been intensively investigated over several decades, and whilst we do not review all developments here, we do highlight the relevant challenges that are of importance to address in the future. For a detailed review of the dynamics of the SPV and its tropospheric influences, readers are referred to the recent reviews by, for example, Kidston et al. (2015), Baldwin et al. (2019Baldwin et al. ( , 2021, Butler et al. (2019), and Butchart (2022).
In the tropical lower stratosphere, a dominant feature exists in the zonal wind field called the Quasi-Biennial Oscillation (QBO), which was discovered during the 1960s and Lag (days) Pressure (hPa) Fig. 1. Composite normalized anomaly of polar-cap-averaged (60°-90°N) geopotential height following 46 SSW events according to ERA5 reanalysis data for the period 1950-2020. SSW is defined according to the method of Charlton and Polvani (2007). The polar-cap-averaged geopotential height anomaly is commonly used as the NAM index (Baldwin and Thompson, 2009 1970s (Reed et al., 1961;Angell and Korshover, 1964;Belmont and Dartt, 1968). The QBO is characterized by a regular downward mitigation of easterly and westerly flow over a period of roughly 28 months (e.g., Ebdon, 1960;Ebdon and Veryard, 1961;, and two QBO disruptions have been respectively observed in 2015/16 and 2019/20 (e.g., Osprey et al., 2016;Anstey et al., 2022). So far, the QBO has been theoretically well understood; the transitioning of equatorial winds is a result of the upward propagation of tropospheric atmospheric waves (e.g., gravity waves, mixed Rossby-gravity waves, and Kelvin waves) in the tropics into the stratosphere, which deposit momentum into the stratospheric mean flow and in turn drive the easterly-westerly transition (Lindzen and Holton, 1968;Wallace and Kousky, 1968;Holton and Lindzen, 1972). Although the QBO is a tropical stratospheric phenomenon, its influences are not confined to the tropical stratosphere, but reach instead to the high latitudes.  summarized three possible pathways for the QBO impacting the troposphere below, as shown in Fig. 2. One of QBO's impacts in the tropics is its modulation of the Madden-Julian Oscillation (MJO; Madden and Julian, 1972;Zhang, 2005), which is the dominant mode of intraseasonal variability in the tropical troposphere and has an important influence on global weather and climate. Studies have found a robust QBO-MJO connection, e.g., the MJO is much stronger and persists about 10 days longer in the QBO easterly phase than in the QBO westerly (Yoo and Son, 2016;Son et al., 2017). In addition, the S2S forecast skill for the MJO is improved by one week during the QBO easterly phase relative to the westerly phase (Marshall et al., 2017;Lim et al., 2019). Despite the robustness of the QBO-MJO connection, there is no consensus as to the mechanisms responsible for the QBO's impacts on the MJO (e.g., Martin et al., 2021). Several mechanisms have been proposed to explain the links, such as QBO-related temperature stratification effects, radiative effects, and its impacts on wave propagation (e.g., Son et al., 2017;Abhik et al., 2019;Densmore et al., 2019;Sakaeda et al., 2020), but their relative importance and robustness are still under scrutiny. In addition to its impacts on the MJO, the QBO has been found to impact tropical deep convection and precipitation via its influence on tropical tropopause temperatures or static stability, as evidenced by both observational and modeling studies (Giorgetta et al., 1999;Collimore et al., 2003;Garfinkel and Hartmann, 2011;Nie and Sobel, 2015;Gray et al., 2018;Rao et al., 2020a, b, c;Huangfu et al., 2021;Wang et al., 2021;Anstey et al., 2022).
The QBO can also induce a meridional circulation at subtropical latitudes, which can influence the horizontal temperature gradient and vertical wind shear near the tropospheric subtropical jet. The modified horizontal temperature gradient/vertical wind shear can further affect the life cycle of synoptic-scale and planetary-scale waves over midlatitudes in the troposphere, which thus helps extend the influence of the QBO down to the surface (Ruti et al., 2006;Garfinkel and Hartmann, 2011;Gray et al., 2018).
The QBO can also modulate the strength of the SPV, particularly in the Northern Hemisphere (e.g., Chen and Wei, 2009;Garfinkel et al., 2012;Zhang et al., 2019aZhang et al., , 2020b;  Rao et al., 2020a), and subsequently influences tropospheric circulation and surface weather patterns. The SPV tends to be colder and stronger during westerly QBO, but warmer and more disturbed during easterly QBO; this is wellknown as the Holton-Tan effect Tan, 1980, 1982), i.e., during easterly QBO, vertically propagating planetary waves from the extratropical troposphere in winter are refracted away from the subtropical region and redirected towards the polar stratosphere where they break and in turn weaken the polar vortex, and vice versa for westerly QBO. Some studies have indicated that the Holton-Tan effect was weakened during 1977-97 (e.g., Lu et a., 2008Lu et a., , 2014, whereas it is projected to strengthen in the future according to the results of CMIP5/6 models (Rao et al., 2020c), suggesting that the Holton-Tan effect has not been steady over time. In addition, the QBO has been reported to have impacts on stationary planetary wave activity in winter in the Northern Hemisphere, and subsequently on East Asian climate (e.g., Chen and Li, 2007;Ma et al., 2021). For a detailed summary of the QBO's impacts on the troposphere, readers are referred to the reviews by Haynes et al. (2021) and Butchart (2022).
A prominent meridional circulation in the stratosphere that bridges the tropical stratosphere and the polar stratosphere is the Brewer-Dobson (BD) circulation (see Fig. 3), which is also referred to as the global mass circulation and features tropospheric air entering into the stratosphere in the tropics, moving upwards and polewards, and then descending in the middle and high latitudes (e.g., Dobson and Harrison, 1926;Brewer, 1949;Dobson, 1956;Butchart, 2014). The BD circulation is driven by the breaking or dissipation of the waves propagating from the troposphere in the midlatitudes, and thus is also expressed by the wave-driven circula-tion McIntyre, 1976, 1978a, b;Boyd, 1976).
More model-based evidence is now emerging that a changing BD circulation may play a significant role in the dynamical coupling between the stratosphere and troposphere, with implications for surface climate and weather (e.g., Baldwin et al., 2007;Karpechko and Manzini, 2012;Scaife et al., 2012). Changes in the BD circulation have been found to be closely connected to changes in the intensity and location of the tropospheric jet stream (e.g., Thompson and Solomon, 2002), surface pressure changes over middle and high latitudes (e.g., Zhang and Tian, 2019), tropospheric storm tracks (e.g., Scaife et al., 2012;Kidston et al., 2015), and precipitation (e.g., Karpechko and Manzini, 2012). As shown in Fig. 3, the QBO, the BD circulation, and the SPV are also dynamically connected. For instance, during easterly QBO, planetary waves are refracted by the critical line and redirected towards the polar regions, where they break, and an anomalous BD circulation that pumps polewards and then descends over the pole is thus induced. The descending branch of this anomalous BD circulation can cause anomalous adiabatic warming over the pole and a subsequent weakening of the SPV. It has been found that easterly QBO together with a weakened SPV are conducive to an increased likelihood of extreme cold-air outbreaks occurring, and vice versa (e.g., . Apart from large-scale stratospheric processes, some small-and mesoscale processes in the lowermost midlatitude stratosphere (e.g., tropopause folding, stratospheric potential vorticity intrusion, isentropic mixing) also play a role in tropospheric weather and climate changes (Fig. 3). Luo et al. (2013) found that frequent tropopause folds over the mei-yu area in China lead to a strong downward transport of airmasses from the lower stratosphere to the upper troposphere before the onset of mei-yu precipitation, suggesting that strong potential vorticity intrusion from the lower stratosphere could provide useful information for estimating the onset of the mei-yu season in China. In addition, tropopause folds are closely associated with surface wind gust extremes (e.g., Škerlak et al., 2015), rapid cyclogenesis (e.g., Wernli et al., 2002), and even the initiation of severe mesoscale weather systems like strong convection (e.g., Russell et al., 2012;Antonescu et al., 2013). A recent model-based study reported a statistically significant change in the tropopause fold frequency in both hemispheres under future climate change (Akritidis et al., 2019). As for small-scale isentropic mixing processes, their impacts on tropospheric weather and climate take place mainly through changing the tracer distributions in the upper troposphere and lower stratosphere, and hence exert an indirect impact via chemical-radiative coupling. However, this issue has not been well addressed to date, due to the complexity of the coupled chemical-radiative-dynamic processes involved and a lack of high-resolution tracer observations.

Role of stratospheric thermodynamics in climatic variability and trends
Owing to increases in anthropogenic emissions of GHGs and ozone depleting substances (ODSs), the stratosphere has experienced a significant cooling trend during the past several decades (e.g., Ramaswamy et al., 2001;Randel et al., 2009;Ren et al., 2015). However, due to limitations in data, there are large uncertainties in the magnitude of linear trends estimated from different data sources. For instance, the global-mean temperature trend in the middle stratosphere (25-45 km) estimated from NOAA data is almost twice that from UK Met Office data (Thompson et al., 2012). This uncertainty in the trend of stratospheric temperature is also apparent when comparing observations and hindcast model simulations (Thompson et al., 2012). A recent study revisited this issue and indicated that the latest satellite-based data show better consistency with each other and better agreement with observations (Maycock et al., 2018). Nevertheless, the well-known stratospheric cooling and tropospheric warming due to increasing concentrations of GHGs will change meridional and horizontal temperature gradients of the atmosphere, as well as vertical temperature gradients, and in turn affect stratospheric processes (e.g., ozone depletion, BD circulation, the polar vortex) owing to the thermodynamic feedback, which further impacts the weather and climate in the troposphere.
Although there is still debate on what causes the change in the BD circulation, the scientific community has reached a broad consensus that there is an acceleration rate of roughly 2.0%-3.2% (10 yr) −1 (i.e., varying with the GHG scenario considered) in the BD circulation in response to stratospheric cooling and tropospheric warming (e.g., Rind et al., 1990;Austin, 2002;Butchart et al., 2010;Lin and Fu, 2013;Fu et al., 2019). In addition, the mechanisms behind the strengthened BD circulation are still under investigation. Apart from the mechanism that increasing wave activities in the troposphere due to tropospheric warming could lead to a strengthening of the BD circulation, Shepherd and McLandress (2011) argued that GHG-induced tropospheric warming pushes the critical layers within the subtropical lower stratosphere upwards, which allows more wave activities to penetrate into the subtropical lower stratosphere and then strengthen the BD circulation. Changing BD circulation has contributed to an increase in column ozone at midlatitudes (e.g., Mahfouf et al., 1994;Shepherd, 2008;Li et al., 2009), faster removal of ODSs (e.g., Butchart and Scaife, 2001;Douglass et al., 2008;Butchart, 2014), and an increasing ozone flux from the stratosphere into the troposphere (e.g., Hegglin and Shepherd, 2009;Neu et al., 2014). Finally, the dynamic coupling between the stratosphere and troposphere is also influenced by a changing BD circulation (e.g., Zhang and Tian, 2019), with important implications for surface weather and climate.
Observed variability and long-term changes in the properties of the polar vortex have also been widely discussed (e.g., Kim et al., 2014;Zhang et al., 2016;Seviour, 2017;Hu et al., 2018). Previous studies have reported an observed strengthening of the Antarctic polar vortex during austral spring (e.g., Zuev and Savelieva, 2019), whereas there are no recent studies about projected changes in the Antarctic polar vortex. In general, ozone recovery will likely lead to a weakening of the Antarctic polar vortex, while increases in GHGs will tend to strengthen it. However, as mentioned earlier, the SPV has large dynamical variabilities. Consequently, determining the SPV trend is rather challenging. Kim et al. (2014) showed that the Arctic polar vortex experienced a weakening trend in midwinter over the past three decades. However, Hu et al. (2018) reported that the SPV has been strengthened since the last decade. These discrepancies are due either to the different lengths of data recorders used in these studies, or the analysis of vortex strength at different altitudes.
As for the changes in SPV position, Zhang et al. (2016) found that the Arctic SPV in February shifted towards Eurasia during 1980-2016. Moreover, notably, the extent of the polar vortex shift in the last decade (2010s) compared to the 1980s was smaller than before (Zhang et al., 2020a), suggesting that there is a decadal variability in the SPV position, which is related to the internal variability in the climate system (Seviour, 2017;Zhao et al., 2022). In addition, various climate variabilities could influence the SPV position Zhang et al., 2019a). The above results suggest that future changes in the strength and position of the Arctic SPV deserve further investigation.
Recent studies have presented corroborative evidence that the weakening and shift of the Arctic polar vortex has contributed to observed cooling in midlatitude Eurasia in boreal winter (e.g., Zhang et al., 2016;Garfinkel et al., 2017;Huang et al., 2018;Kretschmer et al., 2018;Xu et al., 2021;Liang et al., 2022a). Overall, the future changes in the strength of the Arctic polar vortex diverge widely across climate models, with no agreement as to the sign of the change, and a demonstration of large uncertainties (e.g, Karpechko et al., 2022). In addition, the mechanisms responsible for the projected changes in the Arctic polar vortex remain unclear. To add confidence to the projections of the wintertime Arctic polar vortex strength and position, more effort is thus needed to understand the mechanisms and to narrow the large intermodel spread in projections.
Considering the prominent role of SSW in improving the S2S forecast skill of surface weather, several previous studies have focused on their potential future change. However, there does not exist a consistent trend in SSW events (e.g., Mitchell et al., 2013;Kim et al., 2017;Ayarzagüena et al., 2018Ayarzagüena et al., , 2020Charlton-Perez et al., 2018;Rao and Garfinkel, 2021a). Several possible reasons have been proposed to explain the discrepancy: (1) the opposing effects of the projected ozone recovery and increasing GHG concentrations on the Arctic stratosphere (e.g., Hu and Tung, 2003;Ayarzagüena et al., 2013;Hu et al., 2015); (2) the various criteria used to define SSW, since the frequency of SSW is sensitive to the different definitions of SSW (e.g., Butler et al., 2015;Palmeiro et al., 2015;Butler and Gerber, 2018). A recent study analyzed SSW changes in the future using new simulations performed under CMIP6. It was found that most models show a statistically significant trend in the frequency of SSW, but large disagreement exists on the sign of this trend (Ayarzagüena et al., 2020).
Beside the processes mentioned above, there may be other changes in response to tropospheric warming and stratospheric cooling, which awaits the results of further studies. For example, some studies have investigated the projected changes in QBO and suggest that the QBO amplitude in the lower stratosphere has decreased (Kawatani and Hamilton, 2013;Schirber et al., 2015;Naoe et al., 2017;Rao et al., 2020c). Considering our incomplete understanding of changes in wave forcing contributing to the QBO, it is still difficult to draw a robust conclusion about its future changes or whether there will be more frequent QBO interruptions under the future climate.
3. The stratospheric chemistry and climate interaction 3.1. Stratospheric chemical composition changes Atmospheric chemistry plays a critical role in modulating global climate. Climatic impacts of stratospheric ozone, which is both radiatively and chemically active, have been well documented in the literature (e.g., de F. Forster and Shine, 1997;Son et al., 2009Son et al., , 2010Hu et al., 2015;Nowack et al., 2015Nowack et al., , 2018Xie et al., 2016;Maleska et al., 2020;Zhang et al., 2020a;Friedel et al., 2022;Oh et al., 2022). Apart from ozone, ODSs, aerosols, water vapor, and GHGs in the stratosphere all have an impact on global climate through radiative or chemical processes and chemical-radiative-dynamic feedbacks. A conceptual model of the interactions among the various major chemical components in the stratosphere is shown in Fig. 4.
Due to the emissions of ODSs such as chlorofluorocarbons (CFCs), the quantity of global stratospheric ozone decreased between the late 1970s and early 1990s. On a near-global scale, upper-stratospheric ozone is now recovering owing to the effects of the Montreal Protocol and its amendments on ODS levels (Chipperfield et al., 2017;WMO, 2022). However, ozone in the lower stratosphere, between 60°S and 60°N, has continued to decline since 1998, which has been attributed to various dynamical variabilities in the climate system (Ball et al., 2018; Konopka et al., 2022). Long-term changes and the interannual variability of stratospheric water vapor are not only controlled by the oxidation of methane (CH 4 ), but also related to tropical cold point temperatures and overshooting convections that are both influenced by climate change (Sherwood and Dessler, 2000;Holton and Gettelman, 2001;Rosenlof, 2003;Solomon et al., 2010). Stratospheric aerosols are formed when particles and their precursor gases enter the stratosphere.
The stratospheric aerosol loading and aerosol optical depth (AOD) can increase by one order of magnitude or more and reside for one to three years when a large volcanic eruption injects massive quantities of sulfur dioxide (SO 2 ) into the stratosphere (Robock, 2000;Kremser et al., 2016;Aubry et al., 2021). Both historic and recent trends in stratospheric aerosols can be largely attributed to volcanic eruptions (Deshler et al., 2003;Vernier et al., 2011). Solar radiation warms the stratosphere, mainly through absorption of solar radiation by ozone, stratospheric aerosols, and molecular oxygen. Meanwhile, stratospheric aerosols and GHGs, including water vapor (H 2 O), ozone, carbon dioxide (CO 2 ), CH 4 , nitrous oxide (N 2 O), CFCs, hydrochlorofluorocarbons, and brominated chlorofluorocarbons (halons), can absorb and re-emit outgoing long wave radiation from the troposphere. As mentioned earlier, a significant cooling trend has been observed in the stratosphere owing to increasing concentrations of GHGs. On the other hand, there are various photochemical reactions, gas-phase chemical reactions, and heterogeneous chemical processes that take place among the above chemical species. These chemical processes cause changes in the magnitudes and spatial distributions of chemically and radiatively active tracer gases, which in turn induce temperature and circulation changes in the stratosphere via coupled radiative-dynamical processes. Therefore, complex chemical-radiative-dynamical feedbacks exist in the stratosphere. In the following section, the feedback effects of stratospheric ozone, aerosols and water vapor on the climate system are discussed.

Climate forcing of stratospheric ozone
Stratospheric ozone plays an important role in the global radiative balance of Earth's atmosphere because it can heat the stratosphere by absorbing solar ultraviolet radiation and has a strong infrared radiation absorption band at 9.6 μm, which can change the stratospheric temperature by absorbing and emitting longwave radiation. The anthropogenic emissions of ODSs in the 20th century led to a significant depletion of stratospheric ozone. There has since been an expected global recovery of mid-and upper-stratospheric ozone under the control of the Montreal Agreement, but a trend of decreasing lower-stratospheric ozone still exists (WMO, 2022). Stratospheric ozone changes exert a direct radiative forcing on the surface climate. de F. Forster and Shine (1997) calculated an annually and globally averaged radiative forcing of stratospheric ozone of −0.22 ± 0.03 W m −2 for the period 1979-96, when stratospheric ozone depletion was at its strongest. The IPCC's Fifth Assessment Report estimated a global mean effective radiative forcing of −0.05 ± 0.10 W m −2 for the period 1750-2011, which is relatively smaller than that of CO 2 (2.16 ± 0.25 W m −2 ).
Although the above studies suggest that the radiative forcing of stratospheric ozone is marginal, the representation of stratospheric ozone in chemistry-climate models may have a first-order impact on estimates of effective climate sensitivity (e.g., Tian and Chipperfield, 2005;Nowack et al., 2015). Romanowsky et al. (2019) found that including interactive stratospheric ozone chemistry in climate models leads to a significant improvement in their representation of the stratosphere-troposphere coupling induced by Arctic seaice loss, as compared to simulations without interactive chemistry, and consequently the reproduction of more realistic midlatitude atmospheric responses. It has been suggested that climate-ozone feedback may amplify the climatic impacts of stratospheric ozone, since changes in ozone could induce changes in atmospheric circulation and planetary wave propagation through radiative-dynamical feedbacks (Nathan and Cordero, 2007;Albers and Nathan, 2013;Hu et al., 2015). Ozone-induced changes in planetary wave propagation and breaking may act directly on the tropospheric climate through the "downward control" theory (Haynes et al., 1991) or cause convergence and divergence anomalies in the upper tropospheric eddy heat flux and eddy momentum flux (Zhang et al., 2020a), indirectly changing the meridional overturning circulations in the troposphere (Limpasuvan and Hartmann, 2000;Garfinkel, 2020, 2021b).
The Antarctic ozone hole can exert significant impacts on the atmospheric circulations and the surface climate in the Southern Hemisphere. Because of its radiative cooling effects, stratospheric ozone depletion results in a strengthening of the SPV over the Antarctic. Such a strengthened SPV causes the tropospheric Southern Annular Mode (SAM) to shift toward its positive phase (Thompson et al., 2011), which establishes a linkage between the SPV changes and the surface climate (Son et al., 2010;Thompson et al., 2011). The positive phase of SAM is accompanied with a strengthening of the westerly winds at high latitudes, a southward shift of the subtropical jet stream, and a southward expansion of the Hadley cell in austral summer (Son et al., 2010;Min and Son, 2013;Solomon and Polvani, 2016). Corresponding to such changes in large-scale circulations, a poleward shift of storm track in the Southern Hemisphere was also reported (Grise et al., 2014), with an increase in occurrence probability of cyclones at southern high latitudes (Lynch et al., 2006). The stratospheric ozone depletion may also favor more Rossby wave breaking events in austral summer on the equatorward side of the southern mid-latitude jet (Ndarana et al., 2012). In addition, the Antarctic ozone hole could also result in an increase in subtropical precipitation and a shift of the subtropical dry zones toward the South Pole in austral summer (Son et al., 2009;Kang et al., 2011;Polvani et al., 2011).
The Antarctic ozone hole may also influence the Antarctic sea ice. Some studies have suggested that the increase in Antarctic sea-ice extent from 1979 to 2015 could be explained by stratospheric ozone depletion and its associated atmospheric circulation changes (Hall and Visbeck, 2002;Turner et al., 2009). However, other model-based studies have argued that the Antarctic ozone hole may have led to higher surface temperatures in the Southern Ocean and thus a decrease in the Antarctic sea ice (Sigmond and Fyfe, 2010;Bitz and Polvani, 2012;Xia et al., 2020). Such contradicting responses of sea ice to stratospheric ozone changes may be due to processes operating on different timescales (Ferreira et al., 2015). On shorter time-scales (months to years), the Antarctic ozone hole leads to anomalous equatorward Ekman transport, reducing the SST around Antarctica in summer. Consequently, sea ice freeze-up the following winter is earlier, and the sea-ice edge, year-round, is situated more to the north than normal. On longer time-scales (years to decades), the surface cooling is replaced by warming associated with enhanced Ekman pumping of relatively warm water beneath the mixed layer, which reduces the Antarctic sea ice (Ferreira et al., 2015). The impacts of stratospheric ozone depletion on sea ice take place not only through ozone-induced changes in surface winds and ocean drifts, but also in relation to increases in cloud in the upper troposphere (Xia et al., 2018(Xia et al., , 2020. It is worth noting that stratospheric ozone depletion is unlikely to be the primary driver of the observed Antarctic sea-ice trend Landrum et al., 2017;Seviour et al., 2019). More details about the climatic impacts of the Antarctic ozone hole are discussed in the review by Thompson et al. (2011) and the assessment by WMO (2022).
It is well recognized that the Antarctic ozone hole is recovering as a result of the Montreal Protocol (WMO, 2022). Banerjee et al. (2020) found a "pause" or even a weak "reversal" of the Southern Hemispheric climate trend since the early 2000s due to stratospheric ozone recovery. Their findings provide clear evidence that humans can positively influence Earth's climate through international cooperation, i.e., the Montreal Protocol has slowed the rate of climate change related to stratospheric ozone depletion. Following that study, other researchers have also reported the climate change trend related to ozone recovery. Zambri et al. (2021) found that the post-2001 increase in ozone has led to significant changes in the trends of stratospheric temperature and circulation over the Southern Hemisphere. Ivanciu et al. (2022) further indicated that the weakening of residual circulation, the equatorward shift and weakening of surface westerly winds, and a decrease in the transport of the Antarctic Circum-polar Current in the near future are also related to stratospheric ozone recovery. However, not all the climate changes over the Southern Hemisphere in recent decades are associated with Antarctic stratospheric ozone changes. Hu et al. (2022) reported that a weakening of stratospheric planetary wave activities during September since the early 2000s is a response to increases in SST rather than the ozone recovery.
To date, the impacts of Arctic stratospheric ozone changes on the climate system remain insufficiently understood. Using the UK Met Office's model, Cheung et al. (2014) found that considering stratospheric ozone variability does not significantly improve the weather forecasting skill for the troposphere during spring. In addition, Rao and Garfinkel (2020) indicated that, in models, a good ozone forecast does not ensure a good forecast of surface weather. However, other studies, using chemistry-climate models with coupled chemical-radiative-dynamical processes, have found that Arctic stratospheric ozone depletion can cause positivephase anomalies in the NAO, a poleward shift of the tropospheric jet stream in the North Atlantic, and warming anomalies over Eurasia (Smith and Polvani, 2014;Calvo et al., 2015;Ivy et al., 2017;Friedel et al., 2022). It has also been reported that springtime Arctic stratospheric ozone depletion can cause an increase in precipitation over northern Europe and the northwestern United States, and a decrease over southern Europe, Eurasia and central China (Xie et al., 2018;Ma et al., 2019;Friedel et al., 2022). These climate responses are closely related to a strengthened SPV and a delayed break-up of the SPV, which result from a cooling of the SPV due to stratospheric ozone depletion. In addition, stratospheric ozone depletion may also affect the position of the Arctic SPV and favor a shift of the SPV towards Siberia (Zhang et al., 2020a).
The changes in the polar vortex associated with Arctic stratospheric ozone depletion may further affect the cryosphere and ocean. A recent study found that strengthening of the Arctic SPV in association with stratospheric ozone depletion may induce reductions in the Arctic sea-ice concentration and sea-ice thickness over the Kara Sea, Laptev Sea and East Siberian Sea from spring to summer (Zhang et al., 2022). Arctic stratospheric ozone depletion may further affect the climate over the North Pacific Ocean. Xie et al. (2017) found that a decrease in Arctic stratospheric ozone in early spring tends to induce a negative North Pacific Oscillation (NPO) anomaly in April via southward propagation of Rossby waves from high latitudes. Wang et al. (2022) proposed that Arctic stratospheric ozone depletion favors negative NPO anomalies through the interactions between synoptic-scale eddies and the mean flow over the North Pacific. The relative contributions of the two abovementioned mechanisms need to be further evaluated in future studies. Furthermore, ozone-related negative NPO anomalies force positive Victoria mode-like SST anomalies in the North Pacific through modulating the surface heat flux (Xie et al., 2017) and surface ocean current (Wang et al., 2022), which then induces El Nino-like SST anomalies 20 months later . The main climatic impacts of polar stratospheric ozone depletion on tropospheric and surface climate are summarized in Fig. 5.

Climatic impacts of stratospheric water vapor
Although the concentrations of stratospheric water vapor are rather small, it too can modulate the radiative energy budget of the climate system (de F. Forster and Shine, 1999;Solomon et al., 2010;Dessler et al., 2013;Hegglin et al., 2014). Considering the uncertainty due to calculation methods and various forcings, the total stratospheric water vapor forcing reported by the IPCC's Sixth Assessment Report is 0.05 ± 0.05 W m −2 . In addition, tropospheric temperature increases may cause more water vapor to enter the stratosphere, implying the existence of feedback between stratospheric water vapor and tropospheric climate (Garfinkel et al., 2021). Dessler et al. (2013) estimated the strength of this feedback in a chemistry-climate model to be 0.3 W (m 2 K) −1 , which would be a significant contribution to the overall climate sensitivity. However, some studies have argued that its associated surface warming effect is not as significant as previously thought, which is due to compensating effects of cloud feedbacks induced by water vapor Li and Newman, 2020). On the other hand, increases in stratospheric water vapor could reduce the temperature in the stratosphere since it is also a GHG. However, studies have reported that estimated radiative cooling rates of water vapor in the stratosphere vary considerably among different numerical models Banerjee et al., 2019;Li and Newman, 2020;Wang and Huang, 2020). For example, de F. Forster and Shine (2002) found that an average increase of 0.7 ppmv in stratospheric water vapor could result in a maximum stratospheric cooling of approximately 0.8 K. Following that study, Tian et al. (2009) took the chemical-radiative-dynamical feedbacks of stratospheric water vapor into account and estimated that a 2.0 ppmv increase in stratospheric water vapor can cause a maximum cooling of 4.0 K. Maycock et al. (2014) estimated that the net increase in stratospheric water vapor from 1980 to 2010 cooled the lower stratosphere by up to approximately 0.2 K (10 yr) −1 in global and annual mean terms, which was approximately 40% of the observed cooling trend over this period.
Furthermore, stratospheric water vapor changes could significantly affect ozone concentrations in the stratosphere (Kirk-Davidoff et al., 1999;Vogel et al., 2011;Rosenlof, 2018). Increased stratospheric water vapor improves the HO x level and thereby decreases ozone in the upper and lower stratosphere (Dvortsov and Solomon, 2001;Ito and Matsuzaki, 2015). Stratospheric water vapor also affects ozone indirectly by cooling the stratosphere, which in turn modulates the rates of gas-phase reactions (Dessler et al., 2013;Gilford et al., 2016). In addition, increased water vapor may enhance the areas of polar stratospheric cloud particles and increase the heterogeneous reaction rates, leading to more chlorine activation and stronger polar ozone loss (Tian et al., 2009;Vogel et al., 2011;Drdla and Müller, 2012). The above processes may change ozone-climate feedbacks and give rise to further climate changes.

Climatic impacts of GHGs in the stratosphere
Changes in GHGs, including CO 2 , CH 4 , N 2 O and halocar- bons, could also alter the chemical composition and chemical -radiative-dynamical feedback processes in the stratosphere. First, halocarbons are the major gases responsible for stratospheric ozone depletion, while N 2 O increases NO x concentrations after stratospheric photolysis and also decreases stratospheric ozone. WMO (2022) documented that the ozone depletion potential of anthropogenic N 2 O emissions in 2020 were more than two times as large as that of all CFCs in that year. In the future, as halocarbons continue to decrease, the contribution of N 2 O to stratospheric ozone depletion may be more important (Wang et al., 2014). CH 4 can be photolyzed in the upper stratosphere, increasing the concentration of HO x and thus ozone depletion. An increase in GHGs would reduce the stratospheric temperature and weaken the rate of gas-phase chemical reactions, leading to a "super-recovery" of stratospheric ozone (Shepherd, 2008;Li et al., 2009;Chipperfield et al., 2017). Furthermore, greater quantities of GHGs may increase the frequency of planetary waves entering the stratosphere and accelerate the BD circulation in the stratosphere (Rind et al., 1990;Butchart, 2014;Hardiman et al., 2014), thereby affecting the distributions of stratospheric chemical components. As a result, stratospheric ozone changes and chemical-climate interactions in the future are likely to be highly complex. Multiple chemistry-climate models project that the stratospheric ozone recovery will depend on the future climate change scenario (Dhomse et al., 2018;Keeble et al., 2021). However, it is worth mentioning that there is large uncertainty in future projections of Arctic stratospheric ozone changes. Most chemistry-climate model simulations indicate that increases in GHGs will mean Arctic ozone recovers to its 1980 level earlier than Antarctic ozone does, by slowing down the gas-phase chemical depletion rate of ozone and enhancing the BD circulation (WMO, 2022). However, the cooling effects in the polar lower stratosphere induced by increasing GHGs, combined with increasing stratospheric water vapor, may increase the probability of polar stratospheric cloud occurrence, enhance heterogeneous chemistry, and deplete polar ozone (Tian et al., 2009;Vogel et al., 2011;von der Gathen et al., 2021). von der Gathen et al. (2021) showed that local maxima of polar stratospheric cloud formation potential in the Arctic tends to increase in the future under high GHG emission scenarios.

Climatic impacts of stratospheric aerosols
Stratospheric sulfate aerosols are highly reflective in the visible and ultraviolet bands, which reduces the amount of solar radiation reaching Earth's surface at timescales of months to years. Previous studies have reported a significant negative effective radiative forcing of about −20 W m −2 per unit of stratospheric AOD during volcanically active periods (Gregory et al., 2016;Larson and Portmann, 2016;Schmidt et al., 2018;Marshall et al., 2020), although this may not be a linear relationship for large eruptions (Marshall et al., 2020). Consequently, a general surface cooling can be observed after a volcanic eruption, and this surface cooling can last two to three years after a large eruption (Robock and Mao, 1995). For example, global mean air temperatures were reduced by up to 0.5°C at the surface and 0.6°C in the troposphere following the eruption of Mount Pinatubo (Parker et al., 1996). Yu et al. (2016) estimated that stratospheric aerosols have led to a radiative forcing of −0.072 W m −2 since 1850. However, a warmer winter after eruptions in Northern Hemispheric continents has also been reported, which is a result of the enhanced pole-to-equator temperature gradient and thereby a stronger polar vortex (Robock and Mao, 1992;Robock, 2000). However, the significance of this warming is doubtful and uncertain, and some recent studies have argued, based on models with a relatively higher resolution, that the warming of Northern Hemispheric continents is very small for eruptions such as Pinatubo or Krakatau Azoulay et al., 2021;DallaSanta and Polvani, 2022).
Stratospheric aerosols may also affect regional and global precipitation. For instance, it was reported that monsoon precipitation tends to decrease following large eruptions in the same hemisphere (Trenberth and Dai, 2007;Joseph and Zeng, 2011;Iles et al., 2013;Man et al., 2014;Yang et al., 2022), while significantly intensified monsoon precipitation is found when large volcanic forcing occurs in the other hemisphere or the tropics (Liu et al., 2016(Liu et al., , 2022. But what causes this hemispheric asymmetric response of precipitation to volcanic forcing? Stratospheric aerosols emitted by large eruptions can reduce the incoming shortwave radiation and weaken the land-sea thermal contrast in the same hemisphere, leading to decreases in monsoon rainfall (Iles et al., 2013;Man et al., 2014;Yang et al., 2022). On the other hand, an eruption occurring in one hemisphere enhances the hemispheric thermal contrast and drives cross-equatorial flows that converge into the monsoon trough, favoring more precipitation in the counter hemisphere (Liu et al., 2016). Overall, the precipitation response to volcanic aerosols in the stratosphere is the result of a complex combination of multiple processes including radiative and dynamical effects, which remain poorly understood.
In addition, stratospheric aerosols may modulate stratospheric ozone-climate interactions through chemical-radiative-dynamical coupling processes. Volcanic aerosols might provide an additional surface area density for heterogeneous chemistry, thus accelerating the hydrolysis of dinitrogen pentoxide and the release of active halogens, which would decrease the stratospheric ozone (e.g., Aquila et al., 2013;Tilmes et al., 2018;Kilian et al., 2020). On the other hand, volcanic sulfate aerosols heat the lower stratosphere and accelerate the BD circulation that transports more ozone to higher latitudes (Aquila et al., 2012(Aquila et al., , 2013Muthers et al., 2015;Kilian et al., 2020).

Concluding remarks
This review summarizes the main advances in stratosphere-troposphere coupling and stratospheric chemistry-climate interactions, and discusses some relevant outstanding issues and grand challenges. These new advances and devel- AUGUST 2023 TIAN ET AL. opments in stratospheric processes and their role in climate changes imply that climate prediction skill could be improved on sub-seasonal, seasonal, decadal, and even multi-decadal time scales when stratospheric processes and signals are taken into account. The SPV/SSW, QBO, BDC, and stratospheric air intrusion, in particular, have important impacts on tropospheric weather and climate at various time scales ranging from synoptic to interannual. At even longer time scales, the changes in stratospheric composition, particularly that of ozone, aerosols and water vapor, can modulate global climatic variability and trends via complex chemical-radiative-dynamic feedbacks, and are hence potentially useful for improving the long-term predictability of Earth's climate. However, applying stratospheric information in operational S2S forecast models remains a challenge, partly because of the complexity of the physical mechanisms responsible for the stratosphere-troposphere coupling. Another reason is that better forecasting of the stratosphere on S2S time scales has not yet been achieved. The latest generation chemistry-climate models are still unable to reproduce well important aspects of the chemical and physical climate of the stratosphere. For instance, because of the relatively coarse spatial and temporal grid resolutions of some chemistry-climate models, stratospheric gravity waves that occur on scales smaller than the model resolution cannot be resolved, which to some degree introduces uncertainty and potential bias (e.g., Wang et al., 2023). In addition, in most models, the QBO's period and amplitude are not well simulated. For aspects of stratospheric chemistry, previous studies have identified that the latest generation chemistry-climate models have difficulty in simulating polar stratospheric clouds and heterogeneous reactions that take place on modeled polar stratospheric cloud surfaces. Some of the mechanisms through which stratospheric processes modulate tropospheric weather and climate have been revealed in recent years. Stratosphere-troposphere coupling works at various time scales, and different coupled processes have different mechanisms and pathways depending on the time scale. At synoptic to sub-seasonal time scales, robust evidence exists that SPV anomalies, SSW events in particular, correlate closely with tropospheric weather and climate changes at lead times of more than at least 15 days. At interannual time scales, the QBO plays a dominant role in stratosphere-troposphere interaction, especially in the tropics. BD circulation, which is the primary control of stratosphere-troposphere mass exchange (STE), can also act as a dynamical bridge between the tropical and polar stratosphere. Despite the above knowledge, the thermodynamic mechanisms by which temperature and circulation anomalies in the stratosphere impact surface weather and climate are still not fully understood, and more studies are needed in the future.
There also exists convincing evidence that the stratosphere itself is cooling with increasing concentrations of GHGs, while stratospheric ozone has stopped its decline since the 2000s. In response to the increased concentrations of GHGs and the ozone decline since the 1980s, some climatic factors and processes have exhibited evident trends both in the stratosphere and troposphere. The depletion of ozone in the Antarctic, in particular, has induced significant climate changes in the Southern Hemisphere. In recent years, it has been revealed that variations in Arctic stratospheric ozone have also had important impacts on the weather and climate in the Northern Hemisphere. Arctic stratospheric ozone variations could even cause changes in SSTs and ocean currents over the North Pacific. On the other hand, a slow recovery signal in stratospheric ozone has been noted in the past two decades, but its recovery process and significance still possess many uncertainties and close attention should be paid to the climatic impacts of the ozone recovery in the future. Meanwhile, the several extreme low events of the Arctic stratospheric ozone have been observed in recent years. Whether or not extremes of Arctic stratospheric ozone will appear more frequently and whether those extremes will exert an impact on the weather and climate in the Northern Hemisphere should be carefully monitored.
Apart from ozone, the climatic impact of stratospheric water vapor and aerosols has emerged as another hot topic in recent years. Stratospheric water vapor is increasing with climate warming, while stratospheric aerosols have large variability owing to highly varying natural and anthropogenic emissions. So far, the quantitative contributions of stratospheric aerosols and water vapor to radiative forcing have been reasonably well quantified, but large uncertainties in their short-term and long-term impacts on climate remain. The climatic impacts of stratospheric aerosols and water vapor, including the potential for catastrophic climate impacts following a major volcanic event, and stratospheric smoke from extreme fires, have become a challenging issue.
Finally, from these new advances and knowledge with respect to the role of stratospheric processes in climate, we can see that stratosphere-troposphere coupling is also modulated by land surface and oceanic processes (e.g., snow cover and sea ice), the atmospheric processes above the middle atmosphere, the solar cycle, solar wind, and the meteoric rise in smoke particles (see Fig. 6). The interactions of the whole atmosphere need to be considered in future for a better and more thorough understanding of stratosphere-troposphere coupling. The most important issues worth our attention at present include (1) developing high-top models with good presentation of the stratosphere and even the upper atmosphere for the purpose of applying stratospheric information in operational S2S forecasting; (2) close monitoring of short-term and long-term changes in stratospheric constituents for the purpose of improving climate predictions and projections; and (3) in-depth investigations of atmospheric processes in the upper atmosphere, including the stratosphere and beyond, for the purpose of developing/ improving whole-atmosphere climate models.

Acknowledgements.
This work was supported by the National Natural Science Foundation of China (Grant Nos. 42175089, 42121004 and 42142038).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.