 |
Stratospheric Processes And their Role in Climate
|
||||||||
| Home | Initiatives | Organisation | Publications | Meetings | Acronyms and Abbreviations | Useful Links |
|
Stratospheric Processes And their Role in Climate
|
||||||||
| Home | Initiatives | Organisation | Publications | Meetings | Acronyms and Abbreviations | Useful Links |
The subject of Stratosphere-Troposphere Exchange (STE) is crucial to many areas of atmospheric science, and lies at the heart of present concerns about the impact of aircraft emissions on the ozone layer. This last issue is a prime motivation behind the considerable research funds that are currently being invested around the world (particularly in Europe and the USA) in measurement campaigns and in the development of research aircraft. In light of this activity, the SPARC SSG decided at its meeting of September 1994 that it would be useful to bring together a cross-section of experts from the various scientific communities concerned with STE, in order to provide improved coordination between the different national research activities and to work towards a comprehensive measurement and modelling strategy. The resulting workshop was sponsored jointly by the Atmospheric Environment Service of Canada and the WCRP. Participants were invited from the following scientific communities : theory, global modelling, mesoscale modelling, airborne measurements, ground-based measurements, satellite measurements, stratospheric chemistry, diagnostics, and assessment studies. There was also representation from funding agencies. A total of 37 scientists participated, from eight different countries.
The workshop took as its starting point the progress reported at the highly successful NATO ARW on STE held in Cambridge in September 1993 (see report by P. Haynes in SPARC Newsletter No. 2), which led to the excellent review paper on STE by J.R. Holton et al. in Reviews of Geophysics, 33, 4, 403-439, 1995. The Cambridge workshop marked the beginning of a turning point in the subject. Previous research on STE had focused almost exclusively on the behaviour of the tropopause itself, and on the mesoscale phenomenology of strong mixing events such as midlatitude tropopause folds and deep tropical convection. At Cambridge, it was explicitly recognised that STE is but one aspect of a global picture of transport and mixing of chemical species, and one which for many purposes involves the dynamics of the entire stratosphere. In this respect it is clear that rather than focusing on STE per se, it makes more sense to consider the broader question of transport and mixing in what is now referred to as the "lowermost stratosphere" namely that part of the stratosphere whose isentropic surfaces are connected to the troposphere (it is also sometimes called the stratospheric part of the atmospheric "middleworld"). This development was consolidated, and pushed still further, by the recent SPARC workshop described herein.
The workshop began with overview presentations on the theoretical and modelling background. M. McIntyre (Univ. of Cambridge, UK) emphasised the connection between global and local aspects of exchange, as demonstrated so clearly by the "tropical tape recorder" which he had first proposed at the 1993 Cambridge workshop to explain certain puzzling features of observed tropical water vapour profiles. He noted the crucial distinction between the lowermost stratosphere and the "overworld'' (that part of the stratosphere not connected to the troposphere along isentropic surfaces). The overworld is partitioned by a subtropical transport barrier which maintains a contrast between ascending and descending air. Exchange between the overworld and the lowermost stratosphere is dominated by the zonal mean vertical (diabatic) mass flux, which is significantly controlled by extratropical wave driving (termed the "extratropical pump"). Exchange across the tropopause itself is, in contrast, strongly affected by eddy fluxes (TUTTs, folds, monsoon circulation, Cb anvils).
P. Haynes (Univ. of Cambridge, UK) noted that although there is a general connection between tropical ascent and extratropical wave driving (as expressed by the "downward control principle"), a quantitative understanding of the details of tropical upwelling remains elusive. Another outstanding theoretical puzzle is the relation between the Transformed Eulerian Mean (TEM) circulation -- which is the one connected to wave driving -- and the transport (tracer-bearing) circulation ; previous work by Matsuno, and recent work by Mo, suggest there may be important differences between the two, especially within the polar vortex. On a different note, Haynes addressed the issue of stirring and mixing in the lowermost stratosphere, and suggested that this region should contain chemical filaments of smaller scale than potential vorticity (PV) filaments ; one might call such structures "fossils of exchange''. Estimated horizontal scales were 2-20 km, with vertical scales of 20-200 m. The question of the sensitivity of chemical reaction rates to the spatial scale of mixing was raised.
A. Plumb (MIT, USA) also emphasised the importance of the subtropical transport barrier to STE. It is analogous to the polar-vortex transport barrier, though rather more porous ; both are, to a certain extent, a consequence of PV mixing in midlatitudes, creating edges on each side. Plumb proposed two simple models of exchange, which could be fit to chemical tracer observations. The first was a "pipe model", consisting of a tropical pipe (uniform ascent) together with a midlatitude stratospheric surf zone (horizontally sheared descent plus lateral diffusion). The second was a model of the lowermost stratosphere with a mean downward mass flux together with a "hyperventilation" coefficient representing mass-conserving exchange across the tropopause.
J. Holton (Univ. of Washington, USA) began by presenting results from a diagnostic study with Appenzeller and Rosenlof on the seasonal mass budget of the lowermost stratosphere, to assess the relative importance of variations in the amount of total mass. He then proceeded to focus on the tropical "tropopause layer" (15-18 km altitude), representing the transition region between clearly tropospheric and clearly stratospheric dynamics. (Some argued the top of this layer should be more like 20 km.) Holton highlighted a number of critical questions concerning this region, including : What are the mechanisms for exchange between the tropics and midlatitudes ? (And, hence, how fuzzy is the tape recorder imprint ?) How can we account for the observed water vapour distribution ? What is the role of thin cirrus in this ? (It was noted that the existence of subvisible cirrus, supporting the notion of broad tropical ascent, was only recently confirmed by SAGE II and LITE observations.) What is the role of the Indian monsoon in dehydration ? How is the tropopause sharpened ? What is the longitudinal dependence of the tropopause height and temperature ?
K. Hamilton (GFDL, USA) and B. Boville (NCAR, USA) discussed the extent to which general circulation models (GCMs) are able to represent STE. Hamilton noted that the GFDL SKYHI model does get a reasonable diabatic circulation, with upward propagation of water vapour anomalies (consistent with the tropical tape recorder), and a vertical decay of N2O in the tropics presumably due to lateral inmixing. Boville emphasised the extreme difficulty of getting a correct representation of the tropical troposphere-to-stratosphere water vapour transport in a GCM: water vapour mixing ratios vary by nearly four orders of magnitude over a few vertical grid points, and models don't like such large gradients. Also, the details of upper troposphere dehydration in the model depend strongly on the parameterization of small-scale processes, which are poorly understood.
J. Rodriguez (AER, USA) reviewed the present status of 2-D and 3-D assessment models. Historically, 2-D models were developed for the CFC problem, where they could be argued to be somewhat reasonable because of the long tropospheric lifetimes (and hence well-mixed nature) of CFCs and HCFCs. But current assessment questions concern the effect on chlorine and bromine chemistry from the injection of NOx, H2O and aerosols from subsonic (at the tropopause) and supersonic (at 50-60 mb) commercial aircraft. The shorter chemical lifetimes and the importance of lower stratospheric processes, including heterogeneous chemistry, make 2-D models more questionable in this case. For the time being, however, 2-D models remain a "necessary evil" for assessment purposes. The importance of including the effect of transport barriers in 2-D models, and the difficulty of doing so in a reasonable fashion, was emphasised, as was the need for measurement strategies specifically designed to test existing models.
The workshop then moved on to detailed presentations from the measurement community. J. Anderson (Harvard, USA) began with the question of how measurements of meteorological variables and trace species could be used to test theories of STE. The NOy/O3 ratio from ASHOE/MAESA, as a function of latitude, was one example which clearly showed isolated ascent in the tropics. Recently attention has focused on using a greater variety of tracers, with shorter lifetimes, in order to pinpoint seasonal and latitudinal dependencies (see also next paragraph). Factors critical to the effectiveness of such studies include: accuracy, precision, and spatial resolution of the measurements ; good spatial and temporal coverage ; and well-designed aircraft flight paths. With appropriate precision and sampling, one can infer chemical reaction rates. Anderson went on to describe the remotely piloted PERSEUS aircraft presently under development and the sort of measurements it would be able to make.
K. Boering (Harvard, USA) described measurements of tracer correlations obtained in the recent SPADE and ASHOE/MAESA campaigns, as well as in STRAT test flights. The strong seasonal cycle of tropospheric CO2 means that CO2 : N2O correlations in the lowermost stratosphere can be used to infer constraints on tropical-to-midlatitude quasi-isentropic transport, suggesting transport timescales on the order of 1-2 months (see Figure 1). Corroborative evidence using a combination of CO2, H2O and N2O measurements suggests that the principal route by which air reaches the lowermost stratosphere from the troposphere is precisely this quasi-isentropic transport across the subtropical edge of the tropical tropopause layer, rather than across the midlatitude tropopause itself or "up-and-over" in the overworld via the Brewer-Dobson circulation (both of which require diabatic transport). This possible short-circuiting of the Brewer-Dobson circulation is an important effect to quantify.
S. Wofsy (Harvard, USA) described work with Plumb aimed at constructing a simple model of transport and mixing that was crudely consistent with stratospheric dynamics and could be completely constrained by observed tracer distributions. An example constraint might be lateral inmixing rates inferred from the observed decrease of N2O with height in the tropics (also noted by Boering). A "3-pipe" model was described, with prescribed mean vertical transport (based on a radiative calculation), and with vertical diffusion and subtropical lateral exchange (leakage) parameters to be determined. Using CO2, N2O, and CFC11 measurements from ATMOS, ASHOE and SPADE, the model suggests strong lateral exchange below about 480 K (in the 15-17 km layer), and only modest exchange above that. The vertical diffusivity is only weakly constrained by the data, but appears to be much stronger in midlatitudes than in the tropics. There is a pressing need for high-resolution vertical profiles of tracers in the tropics, with seasonal coverage, at altitudes above those the ER-2 aircraft can reach.
P. Newman (NASA Goddard, USA) described the STRAT (Stratospheric TRacers of Atmospheric Transport) experiment which he and Wofsy are leading, including preliminary results from test flights in May 1995. Initial flights with the ER-2 were from NASA Ames, allowing access to latitudes 15N-60N. (Future flights from Hawaii will reach the equator). Contour advection was used to predict PV filaments, then the aircraft was directed to fly through them. Identification of the nature of the air (e.g. tropical or polar vortex, lowermost stratosphere or overworld) was made via tracers, corroborated by backward domain-filling trajectories. It appears that filamentary chemical features can be forecast from contour advection based on assimilated PV fields started between 5 and 15 days in advance. One problem is that in midlatitudes, the ER-2 cannot fly low enough to sample STE events ; there is a need for coordination with low-flying aircraft.
Newman went on to present (on behalf of M. Schoeberl) the upcoming NASA TOTE/VOTE (Tropical Ozone Transport Experiment/Vortex Ozone Transport Experiment) campaign. The aim is to investigate exchange processes across the subtropical and polar transport barriers, focusing in particular on filamentation processes (which are argued to be insufficiently resolved in satellite measurements). The DC-8 aircraft will be flown from Hawaii, and Alaska/Norway, in winter 1995-96.
B. Vincent (Univ. of Adelaide, Australia) presented an overview of ground-based radar techniques. He noted that radar reflectivity is very sensitive to static stability and to water vapour ; thus it can be used to identify airmass boundaries such as the tropopause, and clouds. Measurements of turbulence intensity have previously been used to infer vertical diffusivity. M. McIntyre argued that these inferences are based on dubious assumptions about a ``turbulent Prandtl number''. However, F. Bertin argued that the estimated vertical diffusivities are consistent with the statistical models of Dewan (1979) and Woodman (1984) applied to high-resolution UHF radar observations of turbulent patches. This point remains controversial and requires further elucidation. Radar measurements of mean vertical velocities were also argued to be problematical, because of the requirement for a very narrow beam antenna which is only possible with UHF radars.
L. Gray (RAL, UK) described results from the two TOASTE (Transport of Ozone And STE) experiments. The experiments used a European network of radars, ozone lidars, temperature/water vapour lidars, aircraft, ozonesondes, and radiosondes, in combination with ECMWF analyses and modelling, to study tropopause folds and cut-off lows. Examples were found of dry, ozone-rich air, evidently of stratospheric origin, with no corresponding PV signature ; are these "fossils of exchange" (cf. Haynes presentation described above) ? The degree to which the troposphere (and, indeed, the lowermost stratosphere) is filled with such fossils is clearly worth further study. It was suggested that having tracers other than just O3 and H2O would be useful for discriminating more precisely between tropospheric and stratospheric air.
G. Ancellet (CNRS, France) described the ESTIME (Echanges Stratosphere-Troposphere, Investigations à Moyenne Echelle) experiment, which, much like TOASTE, used a network of ground-based radars, lidars, and ozonesondes (all in France) together with chemical transport modelling (CTM) from KNMI (Netherlands), to study STE in cut-off lows. Despite the obvious limitations (as for TOASTE) of a fixed spatial location of the network, the temporal continuity of the measurements allowed a characterisation of a variety of small-scale erosion processes involved in STE : turbulent mixing by shear instability, wave breaking, and diabatic effects related to convective activity. ECMWF analyses of tropopause height agreed well with the radars ; also, ozone was found to correlate well with analysed PV. It was hoped that the ozone measurements, combined with the CTM, could provide estimates of tracer transport. However, the CTM was found to be too coarse -- a mesoscale model appears to be required ---and the ozone profiles too few.
H. Kelder (KNMI, Netherlands) first presented results on the cross-tropopause mass flux calculated using the formulation of Wei (1987) applied to ECMWF analyses. The results were found to depend sensitively on the horizontal resolution employed. (Some noted that Wei's formulation is computationally ill-conditioned). Kelder went on to describe the recent STREAM (Stratosphere-TRoposphere Experiment by Aircraft Measurements) and POLINAT (POLlution from aircraft emissions In the North ATlantic corridor) experiments, both of them European efforts. STREAM used a twin-engine Cessna jet aircraft to measure a number of chemical constituents, with the objective of assessing the role of NOy in the lower stratosphere, and the impact of aircraft exhausts. POLINAT had rather similar goals, but used the DLR Falcon aircraft, and complemented the study with extensive modelling. Results focused on identifying chemical structures and relating them either to aircraft plumes or to transport from other regions of the atmosphere.
L. Pfister (NASA Ames, USA) described preliminary results using a satellite water vapour channel, in conjunction with STRAT aircraft measurements, to study processes on various scales in the vicinity of the midlatitude tropopause. He focused on identifying the mechanisms by which tropospheric air gets into the lowermost stratosphere. Pfister argued that for this purpose, H2O provides a much better tracer than N2O. Measurements show that midlatitude convection occurs close to the tropopause, but only just bumps into the stratosphere ; convection does not seem able to account for the relatively high H2O values found at the top of the middleworld. Much deeper intrusions into the stratosphere are observed in conjunction with cut-off lows and anticyclones (i.e. lateral mixing), with parcel trajectories extending over tens of degrees of latitude. Breaking inertia-gravity waves were observed in many cases, and are presumably important in the final mix-down.
J.P. Cammas (Laboratoire d'Aérologie, Toulouse, France) presented a case study of a tropopause fold in the Caribbean observed during the TROPOZ II (TROPospheric OZone) airborne experiment. Subtropical folds appear to be commonplace, but are underrepresented in standard climatologies of tropopause folds which are biased towards midlatitudes by their use of PV thresholds. The PV signatures of subtropical folds are comparatively weak, and one requires chemical signatures for their identification. The challenge remains of how to fit isolated observations such as these into a global framework of transport and mixing.
E. Browell (NASA Langley, USA) presented results from his airborne lidar systems, which measure ozone or water vapour, and aerosols and clouds. His ozone DIAL system, mounted on a DC-8, has been used in many campaigns since 1980. The lidar takes a vertical profile, and when employed in an aircraft flight track provides a horizontal-vertical cross-section. An example is provided by Figure 2, which suggests subtropical lateral exchange in the lower stratosphere and well-isolated tropical ascent above about 22 km (cf. presentations by Plumb, Boering, and Wofsy). The existence and detailed structure of the tropopause can be clearly identified in the ozone cross-sections and sometimes in the aerosol cross-sections. By superimposing a number of legs, a large-scale cross-section can be obtained, for example covering the sloping tropopause from equator to midlatitudes. The small-scale cross-sections show plenty of fold structure, as well as more generic filamentation: the tropopause appears to be an extremely choppy (though distinct) surface. Correlations between ozone and PV have been examined in the troposphere to determine the extent of STE. Browell also described his H2O lidar (called LASE) for use in high-altitude studies with the ER-2 ; this instrument can observe subvisible cirrus and the water vapour associated with it, which have been found to be highly spatially inhomogeneous (cf. Holton presentation described above).
K. Hoinka (DLR, Germany) described the two DLR research aircraft, the Falcon and the STRATO-2C. The Falcon is presently being upgraded, which will increase its ceiling (from 38,000 to 41,000 ft), range (from 1,800 to 3,500 km), and flight time (from 3 to 5:45 hr). In addition to measuring meteorological parameters, it has optional chemical sensors as well as a DIAL (lidar) system ; the upgrade will include a dropsonde system, which is presently being developed in collaboration with NCAR and NOAA. Examples were shown of cross-sections through a B-747 wake and through an orographically induced tropopause deformation. The STRATO-2C is still under development, and the first proof-of-concept flight testing took place between the end of March and early August 1995. The aircraft was designed to have a high cruising altitude and long range ; the modified mission aircraft will be able to cruise at about 22 km for 5 hr. This falls well short of the original design specifications, however, and a decision will have to be made soon on whether to proceed with further modifications. The aircraft is now not expected to be operational before 1998. Anticipated cost of flight time is DM 10-20,000 per hour.
B. Gandrud (NCAR, USA) described NCAR's new high-altitude research aircraft, the WB-57F. This is a converted B-57, with a ceiling of 20-21 km and a range of 4,500 km. It is designed to provide measurements right across the upper troposphere/lower stratosphere region, which is inaccessible to both the DC-8 (which cannot get high enough) and the ER-2 (which cannot cruise low enough). This will be particularly important for in-situ water vapour measurements. The aircraft will have a dropsonde capability, and should have its first flight test in January 1996. Flight time will be available through NSF.
M. Shapiro (NOAA Boulder, USA) began by reviewing observational knowledge of tropopause folds. He emphasised the importance of turbulent exchange processes ; for example, the air in a tropopause fold is typically a 50/50 mixture of tropospheric and stratospheric air. Shapiro then proceeded to outline the proposed FASTEX (Fronts and Atlantic STorms EXperiment) campaign. This is an ambitious international project, designed to improve the forecasting of North Atlantic storms. It is scheduled to be carried out in January-February 1997, and will involve four ships and five aircraft, with extensive dropsonde activity. Rather unfortunately, from the perspective of STE, there are apparently no plans to include measurements of chemical tracers.
J.P. Cammas (Laboratoire d'Aérologie, Toulouse, France) described the MOZAIC (Measurement of OZone on Airbus In-service airCraft) programme. It provides measurements of H2O, O3, wind velocity and temperature taken during operational Airbus flights with five aircraft, including vertical profiles during ascent and descent, and along-flight legs. A number of problems with the H2O measurements were noted. At this point about 1000 flights are in the database, which is available for distribution by CD-ROM. A second phase of MOZAIC (proposed for 1996-97) will include measurements of NOy and CO. A possible third phase (1998-2000) is uncertain.
F. Bertin (CETP, France) described the proposed Airbus 340 project. The goal is to have a dedicated, or partially dedicated, A340 aircraft available for research use. The merits of this aircraft include its large payload (20-50 T), long range (15,000 km), and rapid vertical profiling (0-13.5 km in 1/2 hr). Airbus Industrie were initially supportive, but have recently cancelled funding for the required modifications. The project is now on hold, pending funding. Possible use could be in the context of ARGOS (Atmospheric Research Global Observation System), for which 15 European teams have proposed to make 30 species measurements using the A340. Another proposed use is for NEPONA (Nitrogen oxide Emission and Photochemistry Over the North Atlantic), a POLINAT-like experiment. Anticipated cost of flight time is US $ 20,000 per hour, minimum.
R. Stolarski (NASA Goddard, USA) raised the challenge of aircraft assessment studies. These focus on the expected impact of emissions from three categories of aircraft: high-speed (supersonic) commercial transport, or HSCT, cruising at 65,000 ft ; current subsonic aircraft, cruising at 35-39,000 ft ; and possible future higher-altitude subsonic aircraft, cruising at around 45,000 ft. What we think we need is information on the lifetime and accumulation of pollutants, and on the extent of transport to high altitudes, together with some limits or uncertainties. This has led to a three-track approach: measurements and analysis of individual events, to quantify the degree of irreversible mixing ; global analyses and 3-D models, to quantify its frequency and intensity (and provide constraints on parameterisation) ; and 2-D or low-resolution models, to quantify the overall uncertainty. But what do we really need ? In particular, do we need to understand (and parameterise) irreversible mixing in detail? For chemical reactions that are highly nonlinear, the answer could well be yes (cf. Haynes and Rodriguez presentations above).
M. Geller (SUNY Stony Brook, USA) focused on the measurement issues. Given the fact that there are some very expensive aircraft experiments, and aircraft development projects, it is imperative to ask whether there is sufficient communication between them, and whether their science plans are sufficiently coordinated. Is the community yet in a position to agree on what the most important measurement activities might be? And how could such measurements be most effectively used to improve assessment models ? T. Shepherd (Univ. of Toronto, Canada) highlighted some of the key outstanding issues that had arisen during the workshop. These include :
(Note that the last two questions are presently hampered by the
lack of reliable wind measurements in the tropics ; for the time
being, the only way to estimate mixing rates appears to be by
inferring them from chemical tracers.)
More generally, it can be said that the problem of STE is fundamentally a fluid dynamical problem of transport and mixing. A comprehensive understanding of STE therefore requires a comprehensive theoretical framework to understand transport and mixing. Traditionally this framework has consisted of mean (Lagrangian) advection plus diffusion, but this is completely inappropriate, sometimes spectacularly so, in the vicinity of sharp edges in the tracer fields -- which is the essence of the STE problem. (Stirring by large eddies generates strong gradients, quite the opposite of diffusion.) New conceptual models (transport barriers, leaky pipes and so on) are beginning to emerge, but they would have to be considered quite heuristic and certainly preliminary.
All will agree that we need to develop well-conceived global measurement
strategies, using appropriately chosen chemical tracers, in order
to constrain conceptual and numerical models ; the question is,
how? In this regard, the recent use of satellite data in combination
with carefully designed chemical tracer sections from airborne
instruments offers an extremely promising development. Results
presented at the workshop clearly demonstrated the potential for
chemical measurements to provide quantitative information on critical
questions such as the extent of mixing across the subtropical
potential-vorticity edge, as a function of altitude, and the most
likely route for air to get from the troposphere into the lowermost
stratosphere. To many participants, these new results offered
the kind of insight long familiar to oceanographers in their ability
to identify water masses by their T-S relations. It seems that
we may now be in a position to achieve a similar kind of Lagrangian
view of the chemical circulation of the stratosphere -- which
is an exciting prospect.
This raises the question of whether a global measurement strategy (akin to WOCE) might now be appropriate, if not required. That time has not yet come, however ; as argued above, we still lack a satisfactory theoretical framework for defining a critical measurement strategy, and for integrating those measurements in global models. In the meantime, there are a number of more sharply focused questions, including those described above, that are amenable to quantitative study. In any such study there is a pressing need for better design of complementary measurement programmes, in order to avoid the crippling limitations that are often associated with particular measurement platforms (e.g. intrinsic biases and under sampling inherent in fixed-location measurements ; inability of the ER-2 aircraft to cruise low enough to intersect the midlatitude tropopause ; lack of chemical-tracer measurements in "traditional'' meteorological studies of STE ; lack of connection to the global circulation in mesoscale studies ; and so on).
Ted G. Shepherd