Stratospheric Processes And their Role in Climate
A project of the World Climate Research Programme

Home Initiatives Organisation Publications Meetings Useful Links

Processes governing the chemical
composition of the extratropical UTLS
A report from the joint SPARC-IGAC Workshop

18-20 May 2005
Max Planck Institut für Chemie, Mainz, Germany

K. Law, Service d’Aéronomie du CNRS/IPSL, France (kathy.law@aero.jussieu.fr)
L. Pan, NCAR, USA (liwen@ucar.edu)
H. Wernli, Johannes Gutenberg University Mainz, Germany (wernli@mail.uni-mainz.de)
H. Fischer, Max Plank Institute für Chemie, Germany (hofi@mpch-mainz.mpg.de)
P. H. Haynes, DAMTP, University of Cambridge, UK (P.H.Haynes@damtp.cam.ac.uk)
R. Salawitch, Jet Propulsion Laboratory, USA (Ross.Salawitch@jpl.nasa.gov)
B. Kärcher, DLR Oberpfaffenhofen, Germany (bernd.kaercher@dlr.de)
M. Prather, University of California, Irvine, USA (mprather@uci.edu)
S. Doherty, University of Washington, USA (igac.seattle@noaa.gov)
A. R. Ravishankara, NOAA, USA (A.R.Ravishankara@noaa.gov)

This article also appears in the IGAC newsletter.

Introduction and Background

The links between atmospheric chemistry and climate are receiving increasing attention on several fronts. One region where the two are tightly coupled is the Upper Troposphere/Lower Stratosphere (UTLS), which spans the altitude range from ~8-16 km (depending on latitude). Transport in this region and, in particular, exchange between the troposphere and stratosphere occurs through a combination of processes including, in the tropics, cumulus convection, and in the extratropics, synoptic-scale weather systems, together with the large-scale Brewer-Dobson circulation. It is recognized that net exchange from troposphere to stratosphere in the tropics and from stratosphere to troposphere in the extratropics is under large-scale dynamical control (Holton et al. 1995). However, the net exchange alone does not determine many important aspects of chemical distributions in the UTLS region. Recent observational and modelling studies have further revealed important complexities in UTLS dynamical processes and chemistry, the interplay between the two, and consequences for chemical distributions in the UTLS. In particular these studies have raised questions about the best definition of the boundary between the troposphere and stratosphere, i.e. the tropopause. This applies both to the tropics and to the extratropics. The processes and scientific questions in the two regions are rather different and confining attention to one or the other has some advantages. The subject of this report is the extratropical UTLS, i.e. poleward of the subtropical jets. Many important aspects of the tropical UTLS are discussed in recent papers by Folkins (2005), Folkins and Martin (2005), Gettelman et al. (2004) and Küpper et al. (2004) and references therein.

Based on this new information, a more sophisticated picture is being put together of the factors controlling UTLS chemistry and climate feedbacks. Perturbations to the distributions of trace gases such as O3, H2O, and aerosols in this region can lead to direct forcing of climate. Indirect effects through, for example, changing cirrus following new particle production or contrail formation from aircraft emissions can also impact the radiative balance in this region. In turn, climate change, through changing temperatures and transport patterns, has the potential to effect the chemical composition of the extratropical UTLS and thus the composition of the troposphere and stratosphere. Transport of ozone from the stratosphere to the troposphere may change in response to ozone recovery and greenhouse gas impacts in the stratosphere. Also, as noted in the WMO 2003 Ozone Assessment, transport in the extratropics from the troposphere to the stratosphere of very short-lived halogenated species (VSLS; in particular bromine-containing compounds) and pollutants may be important for understanding current and future stratospheric ozone change.

In an effort to integrate and synthesize new findings and their implications, the IGAC Project (International Global Atmospheric Chemistry; under IGBP and CACGP) and the SPARC Project (Stratospheric Processes and their Role in Climate; under WCRP) held a joint workshop at the Max Plank Institut für Chemie, Mainz in May 2005 to discuss processes governing the chemical composition of the Upper Troposphere and Lower Stratosphere (UTLS) in the extratropics. One aim of the workshop was to update our current state of knowledge following previous workshops discussing the tropopause (i.e. in Bad Tölz, Germany, 2001; Haynes and Shepherd, 2001) and chemistry- climate interactions (Giens, France, 2003; Ravishankara et al. 2004) which both included some discussion about extratropical UTLS composition. It was also felt that it is timely to review these issues given the upcoming WMO assessment in 2006 and given the issues raised in the previous ozone assessment (WMO, 2003). It was also noted that it is nearly 10 years since the publication of the very influential review by Holton et al. (1995), which summarized the state of knowledge at that time related primarily to dynamical drivers of stratosphere-troposphere exchange (STE). Recent observations and modelling studies allow for refinement of these concepts, especially with respect to small(er)-scale dynamics and coupling to chemical composition.

Workshop Design & Discussion Topics

The workshop discussions were designed around four major scientific questions (outlined below) pertinent to improving our understanding about UTLS extratropical chemical composition. Invited overview presentations were given on sub-themes identified within each topic and these were followed by lively discussions in plenary. Discussions were also held in breakout sessions where it was decided to combine the first two topics and discuss the roles of dynamical and chemical processes together. Rapporteurs summarized the discussions on the last day of the meeting. This report summarizes these discussions, focusing on the main highlights from the workshop.

The four framing questions for the workshop were:

1) Which dynamical and meteorological processes govern the chemical composition, especially ozone and water vapour, of the extratropical UTLS on seasonal and interannual timescales? On a large scale, both temporally and spatially, the chemical composition of the extratropical UTLS is influenced by the downward transport of trace gases via the large-scale stratospheric circulation and the upward transport of trace constituents from the troposphere by dynamical processes such as frontal uplift and deep convection. Coupling of air masses between the subtropical UT and the extratropical LS may also be important. Many important details of these transport processes still need to be understood. Analyses of various data sets are now providing insights into the causes of large-scale seasonal and possibly interannual variability in transport processes and chemical composition. The extent to which small-scale processes (e.g. gravity wave breaking near the tropopause, turbulence in the vicinity of jet-streams, radiative processes associated with upper level clouds and condensation) play a role in governing the composition and exchange within the extratropical UTLS is also not well known. In addition, there is increasing evidence that deep convection or convection embedded in frontal systems could be important.

2) What is the relative importance of chemical versus dynamical processes in governing the chemical composition of the extratropical UTLS? Analysis of observational data sets has shown that an extratropical tropopause layer (ExTL) exists in chemical composition between the stratosphere and troposphere which exhibits characteristics of both regions. The extent to which dynamical and/or chemical processes are influencing the composition of this region still remains to be quantified. A better characterization of how strongly the 3-D spatial (latitudinal, longitudinal, altitudinal) and seasonal chemical fields in this region are perturbed by exchange processes between the stratosphere and troposphere is needed in order to identify the relative importance of the chemical and dynamical processes. The impact of different processes such as (pyro-) convection and small-scale mixing on chemical composition are very uncertain and require better quantification.

3) Which chemical/physical processes are important in governing UTLS composition? The physical conditions of the UTLS region (low T, decreasing pressure) give rise to particular conditions such that chemical reactions proceed at different rates than in the lower troposphere or the main bulk of the stratosphere. Large uncertainties still surround our knowledge about many reaction rates and pathways (e.g. VOC degradation, VSLS degradation) which could be important for the chemical composition of this region and which influence the distributions and budgets of HOx, NOx, BrOx and O3, for example. Very little is known about the aerosol budget in this region. In addition, heterogeneous reactions on ice/aerosols are also very uncertain, as are the processes governing aerosol formation/ ageing, ice super-saturation and cirrus properties. Scaling up from the process scale to realistic parameterizations in global models is also an issue.

4) How do we better quantify the net exchange of ozone and other trace constituents between the stratosphere and the troposphere? The flux of ozone from the stratosphere is an important term in the tropospheric ozone budget but global model estimates of net flux still vary by more than a factor of two (EU Chemistry-Climate report, 2004). In addition to the climate impacts, intrusions of ozone-rich stratospheric air into the troposphere can occasionally have significant implications for local regulation of allowable ground-level ozone concentrations and the achievability of established limits. Given that the flux may already have changed or may change in a future climate, it is important to quantify this flux more accurately using new, better metrics. Variations in the methods used to determine the flux together with the paucity of independent estimates based on observations is contributing to these uncertainties. In particular, there is a need to define more meaningful parameters
by which to quantify STE; i.e. ones which can be derived from observations and calculated in models. The concept of a chemical tropopause or exchange boundary between the stratosphere and troposphere is an important issue for defining the exchange, and the choice made for this boundary often influences the conclusions of STE studies. Advances in our knowledge about the processes governing the chemical composition of the UTLS region, refined methods to diagnose fluxes from meteorological data sets and the use of new observational data sets could lead to improved quantification of fluxes. There is also a more basic need to continue the evaluation of global model performance in the UTLS region given that these are the tools being used to integrate our current knowledge and provide predictions of future composition and climate to policy makers.

Discussion Summaries

Summaries of the plenary talks and breakout sessions follow. Reference is made in some cases to talks given by specific speakers,
but we note that this does not preclude the valuable contributions on the topic made by other participants. A full list of workshop participants and talks can be viewed on the workshop web page: http://www.atmosp.physics.utoronto.ca/SPARC/Newsletter26/UTLS%20IGAC/Index.htm As many acronyms (for field projects, satellites, etc.) are used herein, an acronym list with translations is also provided at the end of the paper.

Chemistry and Dynamics: Indicators and Controlling Factors of UTLS Chemistry

The Extratropical Tropopause Layer (ExTL)

While the boundary between the troposphere and stratosphere is generally considered to be defined by the thermal tropopause, this definition is not necessarily appropriate or meaningful when discussing chemical composition. The chemical and thermal tropopause are not generally coincident and further, the chemical transition from UT to LS is not as abrupt or well-defined as the temperature transition. The workshop discussions followed the progress made in the last five years to identify and characterize the ExTL from various in situ and satellite observations of chemical tracers. Trace gas profiles of O3, CO, CO2, N2O and H2O, as well as scatter plots among these species, obtained from recent observations made as part of the airborne MOZAIC, CARIBIC, SPURT, STRAT/POLARIS and AIRS satellite projects, clearly reflect the existence of a transition layer in the upper troposphere/ lowermost stratosphere (UT/LMS) where the chemical composition gradually changes from tropospheric (e.g. high CO, low O3) to stratospheric (low CO, high O3). Figure 1 shows an example of a CO-O3 correlation in the tropopause region where mixing lines (light blue) in the ExTL connect a tropospheric (gray) and a stratospheric (black) trace gas reservoir. Number density distributions relative to the thermal tropopause show that the lower bound of the ExTL extends into the UT. The exact position is hard to determine, since it is neither associated with the thermal tropopause nor with a fixed value of potential vorticity (PV). The upper bound of the ExTL (or the depth of the layer) depends to some extent on the residence time of the tracer under investigation. It is generally higher for species that have a long photochemical lifetime in the LMS (e.g. H2O) than it is for short-lived species like CO, whose tropospheric signature is erased on a time-scale of a few months due to net oxidation by OH in the LMS.

Figure 1. CO-O3 correlation has been used to identify the location and thickness of the ExTL. The top panels display the relationship of stratospheric tracer O3 and tropospheric tracer CO for the two extratropical locations sampled by the in situ measurements on board NASA research aircraft ER-2 during STRAT and POLARIS field campaigns (1995-1997).  The solid lines represent the empirical stratospheric and tropospheric O3-CO relationships, determined imperically from the data.  The dash lines mark the 3s of the respective distribution.  The identified stratospheric, tropospheric, and transitional points are represented by black, gray and cyan.  The centre panels show the altitude distribution of transition points (blue) relative to the thermal tropopause.  In the case of 40° N, the distributions are given as two populations, depending on whether the respective thermal tropopause height is below or above 14 km. The bottom panels show the potential vorticity distribution of the transition points. The 40°N distributions aregiven as two populations, similar to the center panels. (Adapted from Pan et al. 2004)

For better resolution, please contact the SPARC IPO


As outlined in a talk by K. Rosenlof, the chemical composition of the LMS is a function of the relative strength of several processes, such as episodic diabatic upwelling, in particular in NH summer, quasi-isentropic cross-tropopause transport, and diabatic downwelling from the overworld in the Brewer Dobson circulation. The first process is associated with deep overshooting convection and pyro-convection, and its bulk impact is largely unknown. In contrast, the upwelling is relatively easy to quantify via the calculation of EP fluxes1, however there is still
considerable uncertainty about the main forcing that drives the upwelling and about the observed trend of increased tropical upwelling during the last 7 years.

Analysis of seasonal variations of trace gas measurements, presented by P. Hoor for the SPURT project, reveals the importance
of three reservoirs for the understanding of the chemical composition of the ExTL (Figure 2). The seasonal cycle of CO2 in the UT (Figure 2, black) and the ExTL (light blue) is in phase, demonstrating the strong coupling between the ExTL and the UT due to frequent cross-tropopause exchange. Above the ExTL in the LMS (gray) the CO2 maximum is shifted by approximately 3-4 months indicating transport from the
tropical LMS. This transport to extratropical latitudes occurs within 2-4 months and leads to mixing with photochemically aged air diabatically descending from the overworld (Rosenlof et al. 1997).

Figure 2. Seasonal variation of CO2 concentrations as a function of distance in potential temperature relative to the tropopause (2 PVU surface) (a) and the potential temperature (b) (Hoor et al. 2004).

For better resolution, please contact the SPARC IPO

Meteorological processes

Several key meteorological processes in the troposphere contribute to the aforementioned episodic diabatic upwelling into the LMS. These processes include synoptic- scale transport events referred to as conveyor belts as well as smaller-scale deep convective systems. Both conveyor belts and deep convective events are associated with significant latent heat release due to condensation of water vapour and therefore they are distinct from isentropic transport. The role of these non-isentropic transport events for stratosphere-troposphere exchange (STE) has gained increased attention during the last years and hence constituted the main items of the presentations by A. Stohl and M. Lawrence.

The discussion of meteorological processes that are associated with significant transport events from the stratosphere to the troposphere (STT) or vice versa (TST) started on the synoptic scale. In his presentation on this topic, A. Stohl focused on the Lagrangian perspective and first suggested
the terminology introduced during the STACCATO project, whereby STE is regarded as the overall STT plus TST processes, and deep exchange refers to rapid transport on synoptic time scales between the boundary layer and the LS. Deep exchange defined in this way is regarded
as particularly important because it brings together stratospheric and boundary layer air masses, with strongly differing chemical compositions, and does so on a short timescale (≤ 1 day, e.g. Stohl et al. 2003). The concomitant occurrence of deep STT and deep TST can lead to a vertically inverted pattern with air of stratospheric origin close to the ground and polluted boundary layer air at the tropopause (Stohl and Trickl, 1999). Particular attention was given to the role of warm conveyor belts (WCBs) that occur ahead of intense cold fronts and which transport warm and moist air from the subtropical boundary layer to the northern extratropical UT within 1-2 days. According to the recent WCB climatology
of Eckhardt et al. (2004), boundary layer starting points frequently occur near very polluted areas (east coasts of North America and Asia). About 5% of the WCBs eventually enter the LMS. The processes associated with the diagnosed increase in potential vorticity of the WCB air parcels entering the stratosphere is not yet well understood. One hypothesis is that diabatic potential vorticity changes occur due to radiative processes
in the WCB outflow regions, characterized by strong vertical humidity gradients and clouds. V. Wirth showed results from idealized studies on this issue which indicated that significant PV changes can occur due to radiative processes near the interface of humid upper tropospheric and dry stratospheric layers. Other synoptic-scale processes that are relevant for STE in the midlatitudes (e.g. the formation of tropopause folds, Rossby wave breaking) were not discussed in detail.


M. Lawrence presented a concise overview on the role of deep convection for STE. He showed that observations, parameterizations and cloud resolving models (CRMs) have been used to study this process. Almost no direct observations exist for STT due to convection, but idealized
model simulations – for instance with the WRF model – show that convection can trigger STT. For upward transport across the extratropical tropopause (TST) associated with convection, there are several observational and model studies that provide clear evidence for the existence of this process, in particular during the summer months. However, the net quantitative impact of this process is still largely unknown and requires further investigation.

Analysis of observations (e.g. STERAO, EULINOX, TRACE-P) has shown that transport of pollutants by this mechanism is important at extratropical latitudes, especially over Asia and central North America, leading to perturbations in upper tropospheric trace constituent
budgets (e.g. O3, HOx). Interestingly, data collected by the MOZAIC programme over the last three years shows significant enhancements in NOy and O3 in the upper troposphere over North America, and these are often not correlated with CO (Petzoldt et al. 2005; Figure 3; see
colour plate I). This indicates that lightning or possibly aircraft emissions may be a principal source of NOx in the UT over continental regions, something that was also suggested by analyses of previous aircraft campaign data (NOXAR, SONEX; e.g. Jeker et al. 2000). This is in contrast
to regions downwind of continents, where frontal uplift of surface pollutants may be more important. More recent campaigns have shown that trace gases, including short-lived VOCs or OVOCs, can also be transported into the lower stratosphere (e.g. MINOS, H. Fischer). However, further study is needed on the significance of these measurements and processes for lower stratospheric composition (e.g. transport of short-lived bromine containing species). Continued analysis of data collected during previous campaigns has also led to reductions in the range of estimates for the global amount of NOx from lighting emissions to 2-9 Tg N per year. The combination of cloud-resolved modelling of convection/chemistry (DeCaria et al. 2005) and anvil NOx observations suggests that on average an intra-cloud flash produces
nearly as much NO as a cloud-to-ground (CG) flash (K. Pickering, Figure 4). This is very different from previous estimates which assumed that an intra-cloud flash produced only one tenth of that of a CG flash (Price et al. 1997). Also, the newer estimates of the number of moles of NO
per flash in CG lightning are significantly lower than previous estimates.

Figure 3. Correlation of NOy against CO (left) and O3 (right) over North America as a function of season (3 years of data, seasonal cycle removed). High NOy with low CO in summer suggests lightning influence and possibly aircraft emissions, whereas high NOy with high CO indicates the influence of convective transport of boundary layer pollution (anthropogenic emissions) into the UTLS over this region. In contrast, high O3:CO correlations exists in UT over Asia/Siberia in certain years such as 2003. (Petzoldt et al. 2005).

For better resolution, please contact the SPARC IPO

Figure 4. From the presentation by K. Pickering, this figure shows recent estimates of the ratio of NO production via intra-cloud (IC) versus cloud-ground (CG) lightning for different data sets in the northern hemisphere extratropics. These ratios are much higher (mean: 0.86) than previously assumed by, for example, Price and Rind (1997) who used a ratio of 0.1 in their lightning parameterization in global chemical models.  Also shown as a black line is the NO production rate per flash for CG flashes (PCG) for the median peak current for North America.  This value of just over 500 moles NO/flash is much lower than was assumed by Price et al. (1997; >1000 moles/flash).

For better resolution, please contact the SPARC IPO

Recent evidence indicates that convection associated with forest fires, so-called pyroconvection, may also have a significant impact on mid-latitude UTLS composition. New modelling work presented by G. Luderer using the ATHAM model showed that the initiation of a deep pyro-convection event is very dependent on background meteorological conditions (e.g. cold frontal passage) as well as the sensible and latent heat budgets of the storm, the fire and the environment. There is a wealth of new evidence from airborne instruments (e.g. recent ICARTT campaign over North Atlantic in 2004; MOZAIC data over Siberia in 2003; MOPITT CO satellite data – 2003/2004) of significant enhancements to the levels of trace gases such as CO during summertime periods of boreal forest burning. Whilst mainly confined to the free troposphere, certain very large pyro-convective storms can also penetrate above the tropopause, injecting material into the LMS. M. Fromm showed examples of enhanced values of aerosol (as viewed in terms of aerosol index by the POAM II satellite) several kilometers above the local tropopause in the LS. Enhanced CO and acetonitrile concentrations have also been observed in the LS and are attributed to forest fire emissions (e.g. Crystal-FACE, Jost et al. 2004; Livesey et al. 2004; MOZAIC, J.-P. Cammas; Ray et al. 2004). The significance of these events – which may be occurring
several times per year – is the topic of ongoing research, as is their impact on stratospheric composition (aerosols, O3) and the radiative budget. It is possible that even one large event per year may cause significant perturbations to background aerosol levels (M. Fromm) Figure 5.

In addition to the new information emerging from field campaigns and satellites there have also been significant developments in the complexity of processes included in Cloud Resolving Models (CRMs) and mesoscale models, such as the inclusion of detailed chemical schemes including soluble species as well as aerosol and microphysical processes. However, many discrepancies still exist between different models, as shown by recent comparisons (M. Barth). In particular, further validation is required of trace gas transport by convective systems into the LS and for
this purpose new data is needed, particularly on short-lived species above convective systems. In addition, many of the mechanisms
being studied/evaluated using CRMs or mesoscale models are not included in global models. For example, downdrafts and gravity wave breaking at the tropopause, associated with deep convection, may be leading to STT in the extratropics. Embedded convection in frontal systems can be important for transporting trace gases such as CO into the UT and possibly the LS, although the overall role of this mechanism still needs to be quantified and validated in models. There is also a need to compare results from CRMs with those from global models using the single column modelling approach and to continue with improvements to parameterizations of deep convective transport of tracers in global models, and in particular treatments of soluble species and lightning emissions.


Mixing processes

During the workshop discussions it was evident that there is a need to clarify the terminology around small-scale mixing phenomena.“Mixing” is sometimes used to mean“molecular mixing” and sometimes used to mean “stirring”, i.e. deformation of material surfaces (and hence concentration fields of chemical species) by differential advection so that molecular diffusion is potentially enhanced, but without that diffusion necessarily
acting to homogenize chemical concentrations. Stirring is a route to molecular mixing, but does not itself imply molecular mixing. This distinction is important because it is only molecular mixing that leads to chemical reactions (e.g. between species in previously chemically distinct airmasses).
The distinction between stirring and mixing is particularly important in the context of models. Lagrangian models can predict large gradients in chemical concentrations as a result of stirring but cannot (without significant modification from their usual form) describe the final step of molecular mixing. Eulerian models, on the other hand, assume that chemical concentration fields are constant – in other words wellmixed
– within a grid box (typically 100 km x 100 km x 1 km for a global model). On the other hand, in situ atmospheric data shows that chemical concentrations vary significantly in space – essentially down to the resolved scale of the observations (~1 km for horizontal sections and a few tens of meters for vertical sections).

Two different types of stirring may be important to molecular-level mixing. The first is stirring via the large-scale flow, which can be resolved by global climate models and in global meteorological data sets. Here the distinction between stirring and mixing is important because the time-scale for molecular mixing may be significantly larger than the time-scale for stirring, as estimated by stretching rates. The second is stirring by three-dimensional turbulence arising in convective clouds, through breaking of gravity waves, and other such processes. The nature of threedimensional
turbulence, where stretching is dominated by the smallest eddies, is such that the time-scale for molecular mixing (again for chemical species whose molecular diffusivity is similar to the viscous diffusivity) is similar to the time-scale for stirring. In this case, distinguishing between
stirring and mixing is not as critical.

J. Whiteway, G. Vaughan and others noted that inertia-gravity waves are likely to play a significant role in mixing in the tropopause region and above since their breaking gives rise to intermittent layers of three-dimensional turbulence (Figure 6; see colour plate I). These waves may be generated by topography, by convection, or by synopticscale processes. However, the importance of convection for gravity-wave generation
in the extratropics is not clear, and furthermore the generation of inertia-gravity wave breaking by synoptic-scale systems is still poorly understood, though the fact of the generation is not in dispute. The tropopause level and above is a preferential region for breaking because of the change in static stability when going from the troposphere to the stratosphere. Wave breaking is regularly observed, such as over relatively
weak topography in the U.K. mountains, and sometimes results from interactions between short wavelength gravity waves (perhaps directly generated by topography) and longer wavelength inertia-gravity waves (generated by synoptic scale processes). The resulting turbulent layers may
be greater than 1 km in depth and hence imply substantial vertical transport.

Figure 6.(a) Measurements of vertical wind by the Aberystwyth VHF radar. (b) The spectral width of the radar signal averaged between 12:30 and 01:00 UTC. (c) Vertical wind measured on the Egrett.  Each flight leg is placed at its height relative to the vertical scale in (a). The topographic height below the Egrett track is shown in green at the bottom with the same relative vertical scale as in (a).  The coast of Wales is at 4.1° longitude; the position of the Aberystwyth radar is indicated by the vertical dotted line at 4.0° longitude.  Crosses in (a) indicate the time and height when the Egrett passed directly above the radar.  The turbulent layer between 11-12 km is estimated to have an internal turbulent diffusivity of about 2 m2/s (Whiteway, 2003).

For better resolution, please contact the SPARC IPO

One way of assessing the quantitative aspects of mixing is direct observations of the mixing processes themselves (J. Whiteway). Another is to try to infer the characteristics of mixing from the observed structure of chemical concentrations fields, and determining which model representation
of mixing gives the best fit to observations (B. Legras). An interesting conclusion that comes out of this approach is that the strength of mixing processes is highly variable, as might be expected from the intermittency of three-dimensional turbulence and the likely association with particular
geographic features such as topography. A related approach was used in the incorporation of mixing into the CLAMS model (a Lagrangian model with adaptive generation/destruction of parcels) where it is possible to optimize the mixing formulation to give best agreement with chemical
observations (P. Konopka).

We know that global models with horizontal resolution of 100 km or greater and satellite observations with resolutions of tens of kilometers cannot represent observed chemical concentration variations on scales of 1 km or less. However, a more important question is whether the neglect of these variations leads to systematic large-scale errors in chemical predictions. This has been investigated in three different ways: (i) the implications of changing model resolution have been explored (Esler et al. 2004); (ii) the chemical implications of smoothing in situ observations to give spatial resolutions typical of global models have been investigated (Crowther et al. 2002, Esler et al. 2004); (iii) the effect of mixing
between different boxes in Lagrangian calculations has been explored (Esler et al. 2001). Here the strongest effects are seen when the different boxes have very different initial chemical concentrations. In the UTLS context, this occurs when mixing air that originated in the boundary layer with air with the characteristics of the lower stratosphere. The extent to which this actually happens is not clear (G. Vaughan). Approach (i) is the most straightforward to interpret with respect to implications for global-scale models and suggests that at current resolutions models may be making errors of up to 15% in key chemical quantities such as ozone production efficiency.

At present there are, as noted above, clearly several limitations to the representation of mixing in models. With Lagrangian models the difficulty is how to represent mixing effects without losing the essential simplicity of the Lagrangian approach. With Eulerian models the difficulty is how to
reduce mixing to avoid unrealistic smoothing of important chemical contrasts (such as the tropopause itself). It is clear that mixing is, in reality, intermittent, but whether or not the details of that intermittency are important for large-scale chemical distributions or whether they must simply be
taken into account to interpret individual observations remains to be determined.

In situ Chemical and Microphysical Processes

The large- and small-scale dynamical processes discussed in the previous section alter the extratropical UTLS chemical composition by moving and mixing air masses between the troposphere and stratosphere. In situ chemical and microphysical processes in this thermodynamically and chemically unique region further alter its composition. Here we discuss several key species of particular importance to chemistry/climate interactions, controlling processes, and what steps are needed to better constrain them. Discussions below are based on presentations given by K. Carslaw, J. Crowley, M. Dorf, A. Gettelman, D. Murphy, T. Peter, A. Ravishankara, H. Singh, B. Thornton, and R. von Glasow.


Upper tropospheric HOx and NOx: An accurate knowledge of the abundances of HOx and NOx in the upper troposphere is critical, since photochemical production of O3 is controlled by the reactions of NO with HO2 and RO2. Recent observations in the field and laboratory have yielded insights to some important controlling processes. Observations from many tropospheric aircraft flights indicate that models tend to overestimate HOx. These data were generally obtained at lower altitudes and at a higher ambient humidity than earlier observations that exhibited a discrepancy in the opposite direction. Observations also indicate that models tend to underestimate the HO2/OH ratio at high levels of NO by large amounts (Figure 7). Recent laboratory observations show that, at high NO concentrations, the production of a few percent yield of HNO3 by the NO+HO2 reaction may alter the HO2/OH ratio to be more consistent with observations (Butkovskaya et al. 2005). Finally, laboratory data have shown that acetone photolysis may be a less efficient source of HOx than was previously believed (Blitz et al. 2004). Future approaches for constraining controlling processes on UT HOx and NOx include: i) efforts to validate measurements of HOx precursors via simultaneous
observations by different instruments as well as budget studies; ii) determining the level of agreement between modelled and measured OH and HO2 if, in the models, only sources from O(1D)+H2O and CH4 oxidation as a function of NO or NOx are considered (e.g. how important are nonwater and non-methane sources of HOx?), such as is done by Olson et al. 2004; and iii) comparing observations of NO, HNO3, and CO with CCM and CTM output in order to better quantify the efficiency of production of NO by lightning (H. Singh, R. Salawitch, breakout discussions).

Figure 7.  Comparison of measured and modelled HOx, as a function of NOx, for data collected during INTEX, TRACE-P, and PEM Tropics B. (Presented by H. Singh; Courtesy Bill Brune, private communication).

For better resolution, please contact the SPARC IPO

Chlorine Activation: Recent aircraft data show that levels of ClO between 30 and 40 pptv are quite commonly observed at high latitudes in the northern hemisphere for stratospheric air masses within several kilometres of the tropopause (Thornton et al. 2003). Levels of ClO between 20 and 30 pptv are also observed in the extratropical, UTLS region (Figure 8). These observations suggest that Cl activation on sub-visible cirrus, or on cold sulphate aerosols, might be responsible for a significant component of observed depletion of lower stratospheric ozone (Solomon et al. 1997; Bregman et al. 2002), in contrast to earlier studies in dry, particle-poor regions of the extratropical UTLS (Smith et al. 2001). The global significance of these regions of activated ClO is unclear. The observations of high ClO tend to occur in a spatially non-homogeneous manner. This could be due to variations in available chlorine along the flight track, which is difficult to assess without accurate, precise, high-temporal
resolution measurements of HCl, a surrogate for Cly. On the other hand, the patchiness could be related to the sporadic character of Cl activation, such as could be induced by mixing that combines particle or water-rich air with air that has high levels of Cly. It remains unclear whether
formulations for Cl activation by PSCs (polar stratospheric clouds) can be applied to the heterogeneous activation of ClO on extratropical cirrus, given the nature of the water-rich aerosols and particles that form in the UTLS. Efforts needed to resolve these issues include simultaneous measurements of ClO and HCl in the UTLS, analysis of ice frost point temperature and cloud data from satellite data to assess global significance, and the modelling of existing ClO measurements to evaluate the heterogeneous chemistry schemes used in CTMs and CCMs (B. Thornton, plenary and breakout discussions).

Figure 8: ClO (solid) and altitude (dotted) for portion of 11 April 1998 WB-57F flight over central United States. Time is UT seconds; ClO is averaged for 120 sec (Courtesy Brett Thornton).

For better resolution, please contact the SPARC IPO


Bromine and Iodine: Measurements of total column BrO by the GOME instrument reveal abundances that are more than a factor of two higher than found in typical models (Figure 9). The first issue raised by these observations is the need to define the relative contribution of tropospheric BrO and stratospheric BrO to this discrepancy. Results to date are not consistent. Groundbased measurements of the variation with solar zenith angle of differential slant column BrO suggest most of the discrepancy is caused by a global, ubiquitous, 2 to 3 pptv level of background BrO in the free troposphere (e.g. Müller et al. 2002). On the other hand, ground-based measurements of diffuse and direct solar radiation indicate an upper limit for tropospheric BrO of 0.9 pptv at 45°S, with mean values of ~0.2 pptv (Schofield et al. 2004). This suggests the discrepancy between measured and modelled column BrO might be the result of significantly higher levels of bromine in the LS than are commonly found in models. If BrO really is ~2-3pptv throughout the troposphere as suggested by the former study, then the BrO+HO2 cycle could represent an important sink for O3 (von Glasow et al. 2004), the hydrolysis of BrONO2 could be an efficient route for production of HNO3 (Lary, 2004), and BrO could be a significant oxidant for DMS (and perhaps other species) in the marine boundary layer (Boucher et al. 2003). If the“excess” bromine is in fact in the LS, this bromine could be supplied by the decomposition products of very short lived (VSL) bromocarbons and could have important consequences for our understanding of ozone trends (WMO, 2003). The substantial organic content of many aerosol particles
just above the tropopause suggests there is injection of tropospheric particles into the stratosphere, and the presence of Br on these particles provides the possibility of cross-tropopause transport of bromine, in both directions, by aerosols (Murphy and Thomson, 2000). Also, the
presence of iodine on aerosols may explain the lack of stratospheric IO (e.g. via aerosol uptake of Iy species). Resolution of these issues requires accurate and precise measurements of BrO in the UTLS region (i.e. that have sensitivity as low as 0.5 pptv), the simultaneous measurement of a suite of organics and inorganic decomposition products, and laboratory measurements of heterogeneous chemical reactions of inorganic bromine species and the kinetics of the organic decomposition products of VSL bromocarbons (R. Salawitch, M. Dorf, R. von Glasow, D. Murphy, and plenary and breakout discussions).

Figure 9. Comparison of estimated stratospheric column BrO from GOME, October 1997, assuming a uniform 1 ppt distribution of BrOin the troposphere (close to the upper limit of 0.9 reported by Schofield et al. JGR, 2004) compared to stratospheric column BrO from the AER 2D model, for the WMO 2003 baseline Bry scenario (labeled BryTROP =0) and for model simulations assuming non-zero levels of Bry at the tropopause (BryTROP) equal to 4 ppt and 8 ppt, respectively.  Top panel: raw GOME data.  Bottom panel: mean, standard deviation (thick error bars), and extremma (thin error bars) of GOME data in 5º wide latitude bins. After Salawitch et al. GRL, 2005.

For better resolution, please contact the SPARC IPO

Humidity and Microphysics

Water abundance and supersaturation: An accurate knowledge of the abundance of H2O and ambient temperature is crucial for understanding cirrus cloud formation, estimating radiative forcing, and accurately retrieving aerosols and trace chemical species from satellites. Ice super-saturation has been frequently detected in clear air and inside cirrus clouds, predominantly in the UT (Jensen et al. 2001; Haag et al. 2003). Satellite observations point to a high variability of relative humidity in the ExTL in regions of major storm tracks (Figure 10; see colour plate II). These are regions of significant dynamical perturbations, likely coinciding with enhanced mixing of tropospheric H2O across the tropopause. This picture is corroborated by a few case studies of cross-tropopause tracer transport. Mixing ratios of H2O well above stratospheric background levels are observed, reaching far into the LMS, especially in summer (C. Schiller, breakout discussions). Supersaturation and the nucleation of the ice phase appears to be confined to a vertically narrow layer (up to 1 km thick at mid-latitudes and more extended polewards) above the tropopause (Pan et al. 2000). In situ processes affecting H2O amount and cloud formation/frequency near the ExTL do not seem to influence
the observed trends in mid-latitude stratospheric H2O (A. Gettleman, breakout discussions). The quantification of the different microphysical and dynamical sinks and sources of H2O is still very uncertain. It remains to be determined how often cirrus formation takes place in ice-supersaturated regions.

Figure 10. From the presentation by A. Gettelman: Relative humidity over ice (RHI) from the Atmospheric Infrared Sounder (AIRS) at 250hPa averaged for December-February (top) and standard deviation of daily RHI from AIRS at 250hPa for an average of December-February 2002-2005 (bottom). The thick red line marks the thermal tropopause at 225hPa, corresponding to the AIRS layer pictured.  The extratropical stratosphere poleward of the tropopause is very dry.  The tropopause marks the boundary of high humidity regions, and the upper troposphere has high humidities, particularly in convective regions. The daily variance of RHI maximizes around the tropopause at this level, and is highest in the North Atlantic and North East Pacific, mostly equatorward of the thermal tropopause. High variance is also found along the tropopause in the southern hemisphere. Variations do not imply transport, but fluctuations between tropospheric and stratospheric air at this level.

For better resolution, please contact the SPARC IPO

Aerosol transport and composition: Aerosol precursor gases and primary aerosol particles are injected into the UTLS by rapid vertical transport processes such as WCBs and deep convection (including pyro-convection), thereby influencing the aerosol budget and high cloud occurrence around the ExTL (K. Carslaw). Besides organics, many UT particles contain both sulfate and carbon and a large fraction contain insoluble inclusions such as mineral dust and soot (Murphy et al. 1998; Kojima et al. 2004). A small number of such particles may act as efficient heterogeneous ice nuclei, affecting cirrus formation by freezing at lower supersaturations than for liquid particles. The influence of aerosols
originating in the troposphere on the highly variable and non-uniform UTLS particle composition is seen in measurements at up to 5 km above the tropopause (D. Murphy). This challenges the conventional wisdom that those aerosol particles in this region are entirely composed of H2SO4 and H2O. It remains unclear to what degree vertical transport affects the UTLS aerosol, how lofted aerosols are modified by interacting with gases and hydrometeors in convective clouds, and how these aerosols in turn modify the evolution of deep convective clouds and the formation
of cirrus. A global, speciated mass budget of the UTLS aerosols including sources and sinks is lacking, and therefore it is currently not possible to accurately validate recently developed global aerosol models.

Ice formation from aerosols: Ice cloud formation and characteristics may be changing due to two influences: a change in the abundance and properties of ice-nucleating aerosols (i.e. the aerosol indirect effect), and changes in the small-scale dynamical forcing patterns (Kärcher and Ström, 2003). The relative importance of these two is not well known. The dependence of the number of ice crystals on the updraft speed in a rising air parcel is much stronger than in liquid clouds, making cloud formation processes more susceptible to small dynamical changes than in the mid- to lower-troposphere. Frequent observations of high ice supersaturation in conjunction with high ice crystal number densities suggest a global-scale predominance of homogeneous freezing in the UTLS (DeMott et al. 2003; Gayet et al. 2004; Hoyle et al. 2005). Homogeneous freezing is sensitive to changes in the variability of vertical air motion on spatial and temporal scales unresolved by global models (Figure 11; see
colour plate II). The organic aerosol fraction does not seem to contribute significantly to ice formation (Cziczo et al. 2004; T. Peter), but a few heterogeneous ice nuclei could modify cirrus development and high cloud cover if they cause ice formation at lower supersaturations than required for homogeneous freezing (Figure 12; see colour plate III). Changes in dynamical forcing could easily mask changes in cloud properties induced by ice nuclei, and these two influences are difficult to separate in measurements. Discriminating between natural and anthropogenic causes of cirrus changes in a future climate requires that mesoscale temperature fluctuations to be understood and that their sources (typically
gravity waves) be accurately parameterized in global models. It is furthermore important to know to what extent ice nuclei modify radiatively important cirrus cloud properties.

Figure 11.  Calculated frequencies of cirrus cloud occurrence during fall 2000 based on meteorological fields taken from the ECMWF model in T511/L60 resolution (Haag and Kärcher, 2004).  The regions in which the relative humidity over ice (RHI) exceeds 95% are evaluated along synoptic trajectories driven by the ECMWF winds and are used as a measure for cirrus cloud cover (top, ECMWF).  The forecast model uses a thermodynamically-based cloud scheme and forms cirrus at ice saturation.  The other panels show results from explicit calculations of aerosol and cirrus cloud microphysics along the trajectories.  This approach takes into account that cirrus form at significant supersaturations via homogeneous freezing and consider kinetic effects during growth and evaporation of ice crystals.  The microphysical simulations use the synoptic temperatures (middle, HOM-S) and synoptic temperatures with superimposed small-scale temperature oscillations (bottom, HOM) caused by parameterized gravity waves.  The occurrence frequency is lower in case HOM-S than in HOM, because average sizes of ice crystals are larger and their sedimentation speeds are faster in HOM-S, decreasing average cloud lifetimes.

For better resolution, please contact the SPARC IPO

Figure 12.  Changes of the frequency of cirrus cloud occurrence relative in case HOM shown in Figure 11c caused by additions of heterogeneous ice nuclei (IN) forming ice at 130% RHI.   Total IN concentrations are x cm-3 in the cases MIX x (first 3 panels).  The case MIX-IN (bottom) assumes 0.01 cm-3 with extremely efficient IN that nucleate ice at 105% RHI.  Field measurements suggest that background IN concentrations do not exceed 0.01 cm-3, but higher values may occur locally.  The cloud occurrence is a nonlinear function of ice nucleation thresholds and IN concentrations (for details see Haag and Kärcher, 2004).   Changes in gravity wave properties also strongly modify the cloud occurrence (not shown).

For better resolution, please contact the SPARC IPO


Gas uptake in cirrus clouds: Uptake of chemically active trace gases by cirrus ice crystals could possibly lead to vertical redistribution or even irreversible removal of the gas from UT air masses, potentially altering the ozone budget there (J. Crowley). Molecules residing at the surfaces of ice crystals might alter ice particle growth rates by modifying the super-saturation over individual crystal facets (Gao et al. 2004). Cubic ice may alter ice crystal nucleation and growth, possibly over a wider range of temperatures than previously thought (Murray et al. 2005). A number of field measurements indicate there is substantial uptake of HNO3 in low temperature cirrus clouds, in one case even in concert with enhanced in-cloud supersaturation. According to recent laboratory measurements, equilibrium uptake models frequently used in the past to calculate the
uptake of HNO3 on ice are inapplicable at the low HNO3 partial pressures typical for the ExTL (Ullerstam et al. 2005). Perhaps more important, atmospheric ice is not in equilibrium. Both laboratory studies examining HNO3 and HCl uptake and theoretical work suggest that growth and evaporation of ice may strongly affect the amountspecies taken up (Kärcher and Basko, 2004). Growth models for small ice crystals that are valid for UTLS conditions and which are capable of accounting for habit changes and surface pollution are not available. It is unclear to what extent non-equilibrium processes connected to ice growth in cirrus conditions affect trace gas uptake and heterogeneous reaction rates.

Quantifying Net Exchange of Trace Constituents

Quantifying the global stratosphere-troposphere exchange (STE) of atmospheric species is a prerequisite for identifying the roles of different dynamical and photochemical processes in controlling this flux. In particular, the net flux of ozone (from stratosphere to troposphere) and of water, as well as aerosols (from troposphere to stratosphere) are critical elements in the stratosphere-troposphere coupling and thus in the overall chemistry-climate coupling. As has been discussed, the area of transition from the troposphere to stratosphere is a region of partial mixing, small-scale dynamical processes, and unusual chemistry the stratosphere and the troposphere. Key questions now being asked include:

(1) How important is O3 STE to the tropospheric O3 budget and the overall tropospheric oxidative capacity (i.e. OH)?
(2) How will climate change alter the flux of H2O into the stratosphere?
(3) Do chemical processes in the tropopause transition region alter the STE of key species like O3 and aerosols?
(4) How important are the large-scale, planetary disturbances vs. the smallscale dynamical processes in controlling this STE?
(5) What dynamical-chemical measurements would be needed to detect a significant change in STE?

Answers to the above questions form the knowledge base required for estimating the chemical feedback in a changing climate.

Over the last decade, significant progress has been made in quantifying STE flux on both global and regional scales, and over both annual and synoptic times. Studies have ranged from high-resolution process studies to global integrations. In terms of the global pattern and magnitude of STE there is increasing convergence from the knowledge base of a decade ago, but complete agreement has not yet been reached. An example is given in Figure 13 (see colour plate III), where mass flux calculations using two different models (one Eulerian and one Lagrangian) show similarity in the preferred location of the net diabatic flux (Figure 13a; Olsen et al. 2004) and the downward flux (STT, Figure 13b; Sprenger and Wernli 2003). These two quantities are comparable since STT is the dominant component of the net flux in the extratropics. The knowledge base is such that it is possible to generate maps of the O3 STE on regional and monthly scales and to produce the now classic latitude by-month plot of zonal mean O3 STE to match the similar O3 vertical column plots, as shown in Figure 14 (Hsu et al. 2005). With increasing model resolution and the use of analyzed meteorological fields, global CTMs are beginning to be able to reproduce the spatial and temporal variability observed in trace gas distributions, and they have become a useful tool for case studies of STE events. Examples from several field campaigns and intensive modelling studies have shown that in some cases we can model the fine, filamentary structure of ozone folds at the tropopause. Nonetheless, this remains a difficult task, as shown in Figure 15 (see colour plate III) (Wild et al. 2003), due to the fact that current CTMs still lack the full resolution of the observed structures. (M. Prather, M. Olsen, L. Pan, A. Gettelman, A. Stohl, K. Law).

Figure 13.  Examples of stratospheric flux by Eulerian and Lagrangian models.  (a) Five-year mean extratropical diabatic flux of mass (colour shading) and ozone (white contours) from the Goddard model.  The ozone flux contour interval is 0.5 kg/s beginning at 0.4 kg/s (adapted from Olsen et al. 2004). (b) 15 year climatology of STT mass flux for the Northern Hemisphere based on Lagrangian calculation using ECMWF meteorological fields.  (Adapted from Sprenger and Wernli, 2003.)

For better resolution, please contact the SPARC IPO

Figure 14. STE flux in units of g-O3 m-2 y-1 as a function of latitude and month for (top) year 2000/2001 and (bottom) year 1997 as calculated using EMWF pieced forecast met fields and the UCI CTM. (Adapted from Hsu et al. 2005).

For better resolution, please contact the SPARC IPO


Figure 15. Comparison of FRSGC/UCI chemistry-transport model and DC-8 LIDAR ozone profiles for Flight 18 from Hong Kong to Hawaii on 3 Apr 2001 showing stratospheric O3 intrusions.  The colour scale highlights O3 abundances less than 100 ppb, with 100-500 ppb shown as black, and greater than 500 ppb masked (white).  The flight track of the DC-8 is shown in white, and black contours indicate approximate cloud optical extinction (per km) specified from the met fields. (Adapted from Wild et al. 2003).

For better resolution, please contact the SPARC IPO


Ten years ago, an important observational constraint to the calculated stratospheric ozone flux was given by the relationship of ozone with N2O (Murphy and Fahey, 1994). Tracer correlations in the LMS have proven useful in deriving global, annual mean fluxes of many constituents between the stratosphere and troposphere and in understanding the age of stratospheric air (i.e. time since it last was in the troposphere).
Recently, a new observational study has shown the potential of O3-HCl correlations to be a more accurate tracer relationship for constraining the amount of UT ozone that is of stratospheric origin (Marcy et al. 2004). (D. Fahey)

Troposphere-to-stratosphere transport of water vapour and aerosols across the extratropical tropopause is an important yet not fully investigated aspect of STE. While evidence of “fresh” tropospheric air can be readily seen in the tropopause region where tracer correlations identify a mixed
stratosphere-troposphere chemical regime, it is not clear from models or measurements how large the flux of this fresh material is and whether it influences the middle stratosphere. Volcanic eruptions provide a test of our ability to model the reverse flux in this region, such as for simulations of Mt. Pinatubo aerosols mixing across the extratropical tropopause from the LS and thus contributing to the UT aerosol burden. (J. Penner)

Despite recent progress, the community has yet to digest and incorporate this new knowledge into current applications. For example, the STE terms in the tropospheric ozone budgets among major models still differ by a factor of 2 to 3. This raises the important question: How can our
improved knowledge of the extratropical UTLS actually be implemented to improve the models? (M. Prather)

One key issue is what metric to use to calibrate the performance of global models in calculating STE flux. The use of newly-established tracer-tracer correlations across the tropopause is one option, although the theory of tracer relationships within the troposphere is incomplete. Many intensive field studies (e.g. from TRACE-P to MOZAIC) clearly demonstrate that O3-H2O, CO-O3, or HCl-O3 correlations can be used to define purely stratospheric, purely tropospheric, and mixed air masses. What is uncertain is whether a CTM simulation that reproduces these correlations necessarily implies the correct STE. New generations of satellite data provide the opportunity of using tracer-tracer correlations on a global scale and with spatial resolutions comparable with that of global models. AIRS (on the Aqua satellite) O3- H2O correlations and MLS (on the Aura satellite) O3-CO correlations are two examples of such data sets. These data sets, however, often represent spatial averages over small-scale features. It is important to compare the satellite data with in situ data sets like MOZAIC to understand the limitations of the data due to spatial averaging. (K. Law, M. Prather, L. Pan)

A confounding factor in determining STE in models is that observations of chemical discontinuities show that transport barriers appear to exist across the tropopause (Figure 16) and the choice of the precise transport boundary may make a significant difference in the calculated flux. Models, on the other hand, often produce much smoother chemical transitions, in part due to numerical diffusion within the models. A key question is whether the calculated flux depends on the choice of boundary, which would imply that chemical transformations in the transition zone are important. (L. Pan, M. Prather, A. Gettelman)

Figure 16. Chemical discontinuity across the tropopause:  The CO and O3 data are from ER-2 measurements during POLARIS campaign near Fairbanks, Alaska, April-August 1997.  When altitude relative to the thermal tropopause is used as the vertical coordinate, the tracer profiles show abrupt change near the tropopause. (Adapted from Pan et al. 2004).

For better resolution, please contact the SPARC IPO

Further, defining a correct location for the“boundary” between the stratosphere and troposphere can be ambiguous because of the presence of ExTL, which has a mix of stratospheric and tropospheric chemical characteristics. Should we determine a new way of defining STE flux with consideration of this transitional layer? Would accurate simulation of the ExTL change the STE flux? This is only important if there are chemical sources/sinks in this layer, because in the absence of chemical processes, the ozone flux is conserved across the ExTL. (K. Law, M. Prather)

Over the past 25 years, there have been significant long-term declines in midlatitude LS ozone levels, and this is an important factor in changing the STT flux of ozone. Both dynamics and chemistry likely contribute to this long-term ozone depletion. The possible importance of VSLS to enhancing the chemical loss of O3 due to Cly and Bry species was discussed. More observations are needed to quantify the significance of VSLS-related long-term ozone depletion, as well as the relative contribution of chemical and dynamical forcings to the observed longterm changes in ozone. (J. Logan, R.Salawitch)

New satellite data provide an exciting opportunity for validating and constraining models in the UTLS region. In particular, the AIRS instrument
on Aqua and TES, MLS instruments on Aura all provide global ozone field in the UTLS region (see Figure 17, colour plate IV). To date we have only begun to explore the use of these data sets for characterizing and quantifying the integrated effect of STE. (L. Pan, A. Gettelman)

Figure 17. The colour image shows a cross-section of ozone data from satellite instrument AIRS.  The data shown are 1°x1° binned averages. The white contours represent the zonal wind, highlighting the subtropical jet and polar jet locations. The light yellow dash contours are potential temperature.  Orange contours are 2 and 4 PVU potential vorticity. These meteorological fields are from 1°x1° degree and 26 level NCEP GFA data.

For better resolution, please contact the SPARC IPO

Concluding Remarks

While perhaps more questions than answers emerge from the above discussions, the convergence of knowledge at the workshop was very useful in helping better define what is known regarding processes controlling the composition of the extratropical UTLS. Just as important, the workshop
helped to identify the remaining outstanding questions.

It is clear that consistent use of welldefined terminology (c.f. STE=STT+TST) is imperative, so that disparate studies can be integrated for a larger-scale picture. In this regard there is especially a need to better understand the newlyidentified ExTL. Given the complex thermodynamic
and chemical state in this region, what metrics should be used to define the ExTL? What is its special role in the chemical, physical and dynamical
state of the extratropical UTLS?

In some cases, focused measurement campaigns would allow us to clarify which processes are significant to the extratropical UTLS region and therefore warrant more extensive study. For example, targeted measurements of aerosol composition in the northern hemisphere UTLS, in conjunction with satellite data analysis, could help determine how pyro-convection is influencing the chemical and optical properties of particles
in this region. In situ measurements could also be used to investigate, for example, the effects of short-lived bromine containing species transported to the UTLS.

Laboratory studies of reaction rates and heterogeneous ice cloud formation processes under conditions appropriate for the mid-latitude UTLS region are needed for more accurate model representation. Focused studies are also needed to understand how the coupling of dynamical
processes over a range of scales control the chemical mixing and microphysical cloud formations in the extratropical UTLS (i.e. How “mixed” is the air in this region?) and to improve our modelling capability in this region (i.e. What processes are essential to include in order to represent the chemical and microphysical state of this region?). While the importance of deep convection in this region is now recognized, the measurements needed to quantify its effect on a global scale remain to be identified.

Finally, there is a need to incorporate existing knowledge into models in order to assess regional and global-scale impacts on, for example, cirrus cloud formation. In particular, while some key species and processes are starting to be included in CTMs and CRMs there is still the need to determine appropriate parameterizations for GCMs and CCMs. While models’ predictions of STE across the extratropical tropopause have recently improved, large uncertainties in flux calculations still exist. New metrics must be found for validating these models against observations.



We would like to thank IGAC, SPARC and the European ACCENT projects for their financial support of the workshop, and Claudia Keller, Bettina Krueger, Gudrun Schlaf, Christian Gurk and Markus Jonas for their very helpful on-site support in Mainz.

AIRS – the Atmospheric Infrared Sounder on the Aqua satellite (http://www-airs.jpl.nasa.gov/)
Aura – One of NASA’s EOS (Earth Observation System) satellites (http://eosdatainfo.gsfc.nasa. gov/eosdata/aura/mls/mls.html)
CARIBIC – Civil Aircraft for the Regular Investigation of the atmosphere Based on an Instrument Container (http://www.caribic-atmospheric.
CCM – chemistry-climate model
CLAMS – Chemical Lagrangian Model of the Stratosphere
Crystal-FACE – The cirrus Regional Study of Tropical Anvils and Cirrus Layers – Florida Area Cirrus Experiment (http://cloud1.arc.nasa.gov/
EULINOX – The European Lightning Nitrogen Oxides Project (http://www.pa.op.dlr.de/eulinox/)
GOME – instrument on the ERS-2 satellite for global monitoring of ozone (http://earth.esa. int/ers/gome/)
ICARTT – International Consortium for Atmospheric Research on Transport and Transformation (http://www.al.noaa.gov/ICARTT/)
MINOS – Mediterranean Intensive Oxidant Study
MLS – Microwave Limb Sounder on the Aura satellite (http://mls.jpl.nasa.gov/)
MOPITT – Measurements of Pollution in the Troposphere (http://terra.nasa.gov /About/MOPITT/
about_mopitt.html) instrument on the TERRA satellite (http://terra.nasa.gov/About/)
MOZAIC – Measurement of Ozone and Water vapour by Airbus In-service Aircraft (http://www.aero.obs-mip.fr/mozaic/)
NOXAR – Measurements of Nitrogen Oxides and Ozone Along Air Routes (http://www.iac.ethz.ch/en/research/chemie/tpeter/Noxar.html)
POAM II – Polar Ozone and Aerosol Measurements (http://wvms.nrl.navy.mil/POAM/poam.html)
POLARIS – field study; Photochemistry of Ozone Loss in the Arctic Region in Summer
SONEX – SASS Ozone and Nitrogen Oxide Experiment
STRAT – Stratospheric Tracers of Atmospheric Transport (http://cloud1.arc.nasa.gov/strat/strat.html)
SPURT – SPURenstofftransport in der Tropopausenregion (Tracegas transport in the tropopause region; http://www.meteor.uni-frankfurt.
STACCATO - Stratosphere-Troposphere exchange in a Changing Climate on Atmospheric Transport and Oxidation Capacity
STE – stratosphere/troposphere exchange
STERAO – Stratosphere-Troposphere Experiment - Radiation, Aerosols and Ozone (http://chill.colostate.edu/sterao.html)
STT – stratosphere-to-troposphere transport
TES – Tropospheric Emission Spectrometer instrument on the (http://tes.jpl.nasa.gov/), on the Aura satellite
TRACE-P –TRAnsport & Chemical Evolution over the Pacific field campaign (http://code916.gsfc.nasa.gov/Missions/TRACEP/)
TST – troposphere-to-stratosphere transport
(O)VOCs – (oxygenated) volatile organic compounds
VSLS – very short-lived species: e.g. lifetime with respect to photochemical removal <~0.5 year


Blitz, M. A., et al. Pressure and temperaturedependent quantum yields for the photodissociation of acetone between 279 and 327.5 nm,
Geophys. Res. Lett., 31, L06111, doi:10.1029/ 2003GL018793, 2004.

Boucher, O., et al. DMS atmospheric concentrations and sulphate aerosol indirect radiative forcing: a sensitivity study to the DMS source
representation and oxidation, Atmos. Chem. Phys., 3, 49-65, 2003.

Bregman B., et al. Chemical ozone loss in the tropopause region on subvisible ice clouds, calculated with a chemistry-transport model, J.
Geophys. Res., 107, Art. No. 4032, 2002.

Butkovskaya, N. I., et al. Formation of nitric acid in the gas-phase HO2+NO reaction: effects of temperature and water vapor, J. Phys. Chem. A,
109, 6509-6520, 2005.

Crowther R., et al. Characterising the effect of large-scale model resolution upon calculated OH production using MOZAIC data, Geophys.
Res. Lett., 29 (12), doi:10.1029/2002GL014660, 2002.

Cziczo, D.J., et al. Observations of organic species and atmospheric ice formation, Geophys. Res. Lett., 31, L12116, oi:10.1029/2004GL019822,

DeCaria, A. J., et al. Lightning-generated NOx and its impact on tropospheric ozone production: A three-dimensional modeling study of
a Stratosphere-Troposphere Experiment: Radiation, Aerosols and Ozone (STERAO-A) thunderstorm, J. Geophys. Res., 110, D14303,
soi:10.1029/2004JD005556, 2005.

DeMott, P.J., et al. Measurements of the concentration and composition of nuclei for cirrus formation, Proceed. Natl. Acad. Sci., 100 (25),
14655-14660, 2003

Eckhardt S., et al. A 15-year climatology of warm conveyor belts, J. Climate, 17, 218-237, 2004

Esler, J. G., et al. Stratosphere-troposphere exchange: Chemical sensitivity to mixing, J. Geophys. Res., 106(D5), 4717-4732, 10.1029/2000JD900405, 2001.

Esler, J. G., et al. A quantitative analysis of gridrelated systematic errors in oxidising capacity and ozone production rates in chemistry transport
models, Atmos. Chem. Phys. Discuss., 4, 2533-2568, 2004.

European Commission report on Ozone-Climate Interactions, Air pollution research report No 81, EUR 20623, 143pp, 2003.

Folkins, I., Temperatures, Transport, and Chemistry in the TTL, SPARC Newsletter 25, 23-26, 2005.

Folkins, I. and R. V. Martin, The vertical structure of tropical convection and its impact on the budgets of water vapor and ozone, J. Atmos. Sci.,
62, 1560-1573, 2005.

Fromm, M., et al. Observations of boreal forest fire smoke in the stratosphere by POAM III, SAGE II, and lidar in 1998, Gephys. Res. Let., 27,
1407–1410, 2000.

Fromm, M., et al. Pyro-cumulonimbus injection of smoke to the stratosphere: observations and impact of a super blowup in northwestern Canada
on 3-4 August 1998, J. Geophys. Res., 110, D08205, doi:10.1029/2004JD005350, 2005.

Gao, R.S., et al. Evidence that ambient nitric acid increases relative humidity in low-temperature cirrus clouds, Science, 303, 516-520, 2004.

Gayet, J.-F., et al. Cirrus cloud microphysical and optical properties at southern and northern midlatitudes during the INCA experiment,
J. Geophys. Res., 109, D20206, doi:10.1029/ 2004JD004803, 2004.

Gettelman, A., et al. Radiation balance of the tropical tropopause layer, J. Geophys. Res., 109 (D7), D07103, doi:10.1029/2003JD004190,

Haag, W., et al. Freezing thresholds and cirrus cloud formation mechanisms inferred from in situ measurements of relative humidity, Atmos.
Chem. Phys., 3, 1791-1806, 2003.

Haag, W., and B. Kärcher, The impact of aerosols and gravity waves on cirrus clouds at midlatitudes. J. Geophys. Res., 109, D12202, doi:10.1029/ 2004JD004579, 2004.

Haynes P. and T. Shepherd, Report on the SPARC Tropopause Workshop, Bad Tölz, Germany, 17-21 April 2001, SPARC Newsletter, 17,
3-10, 2001.

Holton, J.R., et al. Stratosphere-troposphere exchange, Rev. Geophys., 33, 403-439, 1995.

Hoor, P., et al. Seasonality and extend of extratropical TST derived from in-situ CO measurements during SPURT, Atmos. Chem. and Phys.,
4, 1427-1442, 2004.

Hoor, P., et al. Tropical and extratropical tropospheric air in the lowermost stratosphere over Europe : A CO-based budget, Geophys. Res. Lett.,
32, LO7802, doi:10.1029/2004GL022018, 2005.

Hoyle, C.R., et al. The origin of high ice crystal number densities in cirrus clouds, J. Atmos. Sci., 62, 2568-2579, 2005.

Hsu, J., et al. Diagnosing the stratosphere-totroposphere flux of ozone in a chemistry transport model, J. Geophys. Res., 110, 2005JD006045,

Jensen, E.J., et al. Prevalence of ice-supersaturated regions in the upper troposphere: Implications for optically thin ice cloud formation, J. Geophys. Res., 106 (D15), 17253-17266, 2001.

Jeker, D. P., et al. Measurements of nitrogen oxides at the tropopause: attribution to convection and correlation with lightning, J. Geophys. Res.,
105, 3679-3700, 2000.

Jost, H-J., et al. In-situ observations of midlatitude forest fire plumes deep in the stratosphere, Geophys. Res. Lett., L11101, doi:10.1029/
2003GL019253, 2004.

Kärcher, B. and J. Ström, The roles of dynamical variability and aerosols in cirrus cloud formation, Atmos. Chem. Phys., 3 (3), 823-838, 2003. Kärcher, B. and M.M. Basko, Trapping of trace gases in growing ice crystals, J. Geophys. Res., 109, D22204, doi:10.1029/2004JD005254, 2004.

Kojima, T, et al. Aerosol particles from tropical convective systems: Cloud tops and cirrus anvils, J. Geophys. Res., 109, D12201, doi:10.1029/
2003JD004504, 2004.

Küpper, C., et al. Mass and water transport into the tropical stratosphere: A cloud-resolving simulation, J. Geophys. Res., 109, D10111,
doi:10.1029/2004JD004541, 2004.

Lary, D. J., Halogens and the chemistry of the free troposphere, Atmos. Chem. Phys. Discuss., 4, 5367-5380, 2004. Livesey, N., et al. Enhancements in lower stratospheric CH3CN observed by UARS MLS following boreal forest fires, J. Geophys. Res., 109, D06308, doi:10.1029/2003JD004055, 2004.

Marcy et al. Quantifying stratospheric ozone in the upper troposphere with in situ measurements of HCl, Science, 304, 261-265, 2004.

Müller, R.W. et al. Consistent interpretation of ground based and GOME BrO slant column data, Adv. Space Res., 29, 1655-1660, 2002.

Murphy, D. M., and D. W. Fahey, An estimate of the flux of stratospheric reactive nitrogen and ozone into the troposphere, J. Geophys. Res., 99,
5325-5332, 10.1029/93JD03558, 1994.

Murphy, D.M., et al. In situ mesurements of organics, meteoritic material, mercury, and other elements in aerosols at 5 to 19 kilometers, Science,
282, 1664-1669, 1998.

Murphy, D. M. and D. S. Thomson, Halogen ions and NO+ in the mass spectra of aerosols in the upper troposphere and lower stratosphere,
Geophys. Res. Lett., 27, 3217-3220, 2000.

Murray, B.J., et al. The formation of cubic ice under conditions relevant to Earth’s atmosphere, Nature, 434, 202-205, 2005.

Olsen, M. A., et al. Stratosphere-troposphere exchange of mass and ozone, J. Geophys. Res., 109, D24114, 2004, doi:10.1029/2004JD00186.

Olson, J. R., et al. Testing fast photochemical theory during TRACE-P based on measurements of OH, HO2, and CH2O, J. Geophys. Res., 109, D15S10, doi:10.1029/2003JD004278, 2004.

Pan, L. L., et al. The seasonal cycle of water vapor and saturation vapor mixing ratio in the extratropical lowermost stratosphere, J. Geophys. Res., 105, 26519-26530, 2000.

Pan, L. L., et al. Definitions and sharpness of the extratropical tropopause: A trace gas perspective, J. Geophys. Res., 109, D23103, doi:10.1029/
2004JD004982, 2004.

Petzoldt, K., et al. Four years of NOy measurements in the UTLS by MOZAIC aircraft, EGU Poster 2nd General Assembly, Wien, 24– 29 April 2005; Abstract: EGU05-A-08034. Price, C., J. Penner, and M. Prather, NOx from lightning: 1. Global distribution based on lightning physics, J. Geophys. Res., 102, 5929-5941, 1997.

Ravishankara, A.R., et al. Chemistry Climate Interactions : A report from the joint SPARC/IGAC workshop, IGACtivities Newseltter, 30, 2004.

Ray, E. A., et al.Evidence of the effect of summertime midlatitude convection on the subtropical lower stratosphere from CRYSTAL-FACE tracer measurements. J. Geophys. Res., 109, D18304, doi:10.1029/2004JD004655, 2004.

Rosenlof, K. H., et al. Hemispheric differences in water vapor and inferences about the transport in the lower stratosphere, J. Geophys. Res., 102,
13,213-13,234, 1997

Salawitch, R. J. et al. Sensitivity of ozone to bromine in the lower stratosphere, Geophys. Res. Lett., 32, 10.1029/2004GL021504, 2005.

Schofield, R. et al. Retrieved tropospheric and stratospheric BrO columns over Lauder, New Zealand, J. Geophys. Res., 109, D14304,
doi:10.1029/2003JD004463, 2004.

Smith, J. B., et al. Mechanisms for midlatitude ozone loss: Heterogeneous chemistry in the lowermost stratosphere?, J. Geophys. Res., 106,
1297– 1309, 2001.

Solomon S., et al. Heterogeneous chlorine chemistry in the tropopause region, J. Geophys. Res., 102, 21411-21429, 1997.

Sprenger, M., and H. Wernli, A northern hemispheric climatology of cross-tropopause exchange for the ERA15 time period (1979–
1993), J. Geophys. Res., 108, 8521, doi:10.1029/ 2002JD002636, 2003.

Stohl, A., et al. A New Perspective of Stratosphere– Troposphere Exchange, Bull. Amer. Met. Soc., 84, 1565-1573, 2003.

Stohl A. and T. Trickl, A text-book example of long-range transport: Simultaneous observation of ozone maxima of stratospheric and
North American origin in the free troposphere over Europe. J. Geophys. Res., 104, 30445-30462, 1999.

Thornton B.F. et al. In situ observations of ClO near the winter polar tropopause, J. Geophys. Res., 108, Art. No. 8333, 2003.

Ullerstam, M., et al. Uptake of gas-phase nitric acid to ice at low partial pressures: evidence for unsaturated surface coverage, Faraday Discuss.,
130, 211-226, 2005.

von Glasow R., et al. Impact of reactive bromine chemistry in the troposphere, Atmos. Chem. Phys., 4, 2481–2497, 2004.

Whiteway J. A., et al. Airborne measurements of gravity wave breaking at the tropopause, Geophys. Res. Lett., 30, 2070, doi:10.1029/2003GL018207, 2003.

Wild, O., et al.CTM ozone simulations for spring 2001 over the western Pacific: Comparisons with TRACE-P lidar, ozonesondes, and TOMS columns, J. Geophys. Res., 108(D21), 8826, doi:10.1029/2002JD003283, 2003.

WMO, ‘Scientific Assessment of Ozone Depletion: 2002’, World Meteorological Organization Global Ozone Research and Monitoring Project
– Report No. 47, March, 2003.