Appendix III Implementation Plan

Convective Excitation of Gravity Waves Experiment - CEGWE

A Proposed Field Campaign to Characterise the Gravity Wave Fields Generated by Tropospheric Moist Convection in the Tropics

1. Introduction
2. Scientific Background
3. Scientific Goals
4. Brief Review of Earlier Observations
5. Proposed Venue and Timing
6. Observational Requirements
7. Anticipated Contributions from Instrumentation on the ER-2 Aircraft
8. Anticipated Contributions from Other Observing Systems
9. Role of Limited-Area Modelling
10. Additional Research Efforts During the CEGWE Period - Possible Connections with Other Programs and Satellite Missions
11. References

Gravity waves exert major influences on the large-scale circulation and structure of the atmosphere and are responsible for much of the spatial and temporal variability in the flow above the tropopause. The divergence of the heat and momentum fluxes associated with dissipating gravity waves is one of the major driving forces of the middle atmosphere. In order to produce credible simulations of the earth's climate and predictions of climate response to anthropogenic forcing, comprehensive numerical models of the atmosphere must include physically justifiable parameterisations of gravity wave momentum, heat, and constituent transport. This issue is now recognised as one of the most important challenges in dynamical meteorology. A NATO Advanced Research Workshop, cosponsored by SPARC and STEP, was held in Santa Fe, New Mexico in April 1996 to address this problem. The workshop included experts in observations, gravity wave theory, and atmospheric modelling. While many important issues were discussed, the participants acknowledged that the accuracy and effectiveness of gravity wave parameterisations will be limited, ultimately, by our knowledge of the wave field propagating up into the stratosphere, including the spectrum, the degree of temporal intermittency and geographic variability, and the relation of the stratospheric waves to tropospheric sources such as convection, topography, jet stream instabilities etc. Our ability to understand and model the large-scale circulation and consequent effects on the chemistry of the upper-troposphere/lower stratosphere (UT/LS) region thus requires an improved knowledge of the generation mechanisms for the gravity waves which propagate into the middle atmosphere.

In situ aircraft observations and numerical models of convectively-forced gravity waves have provided some insight into the wave spectrum in the lower stratosphere. However, the existing database of observations is very sparse and certainly not sufficient for global modelling studies. There was general agreement among the Santa Fe Workshop participants that significant progress in understanding, characterising, and modelling gravity wave forcing in the lower and middle atmospheres would be made only by conducting co-ordinated observational campaigns that were carefully planned and executed with guidance from existing theory and models. Workshop participants noted the likely importance of tropical convection as a source of stratospheric gravity waves and remarked that very little observational work has been done on this wave excitation mechanism. It was concluded that an experiment focusing on the response of the middle atmosphere to convection should be given the highest priority.

At the Adelaide meeting of the SPARC Scientific Steering Group (SSG) in December 1996, it was proposed that SPARC endorse and cosponsor an international project to characterise both the gravity wave field generated by convection and the associated wave effects on middle atmosphere momentum, heat, and constituent transport. The project would include model studies of tropospheric convection which would be used to help plan a major field program (and ultimately guide the interpretation of the results). The SSG agreed to sponsor a workshop involving a small group of experts who would be charged with developing a preliminary plan for the proposed field program. The workshop, organised by Kevin Hamilton, was held in Victoria, Canada June 19-20, 1997. Participants were Joan Alexander (U. of Washington), Chet Gardner (U. of Illinois), Kevin Hamilton (GFDL/NOAA), Jim Holton (U. of Washington), Norm McFarlane (Canadian Climate Centre), Lenny Pfister (NASA Ames Research Center), Gary Swenson (U. of Illinois), Toshitaka Tsuda (Kyoto U.), Bob Vincent (U. of Adelaide) and Jim Whiteway (U. of Toronto).

The result of this meeting was the outline for the ambitious field experiment that is described here. The basic proposal is for an intensive period of observations in a limited region in the deep tropics where a wide variety of convective events might be anticipated. Critical to the plan as now conceived is the participation of a high-flying aircraft such as the NASA ER-2 which can make detailed measurements of temperature and of the horizontal and vertical components of the wind in the lower stratosphere above the convective towers. These measurements would be supplemented by, and co-ordinated with, a large number of observations taken by other techniques, including meteorological radars, balloon-borne radiosondes, microwave profilers, lidars, airglow imagers and middle atmospheric radars. The period of intensive observations would last at least several weeks and would be preceded by a longer period of "baseline" observations with available ground-based instruments. A particularly attractive venue for this experiment in Northern Australia has also been identified. Practical considerations suggest either October 2000 or October 2001 as the target date for the beginning of the intensive observational campaign.

There is now widespread appreciation that the distribution of important trace constituents, including ozone, in the atmosphere is very significantly affected by dynamical transport and that any complete model of the chemistry of the stratosphere and of the UT/LS region needs to have a credible treatment of the dynamics. The global-scale mean meridional circulation in the middle atmosphere is driven by a competition between the in situ radiative heating which tries to create a state with very cold winter polar temperatures, and eddy processes that drive the temperatures away from this state. This view is summarised by Fels (1987), for example, and has been made precise by the derivation of the "downward control" principle by Haynes et al. (1991). This roughly states that (away from the frictionally-dominated region near the ground) the mean meridional circulation at any location is largely controlled by the eddy flux divergences at higher levels in the atmosphere.

A realistic simulation of the stratospheric transport in a numerical model thus requires an accurate treatment of the radiative transfer and an adequate representation of the eddies. It is believed that the radiative transfer codes used in comprehensive general circulation models (GCMs) are now sufficiently accurate, but there are still major deficiencies in the simulated stratospheric transport in current GCMs (see Hamilton, 1996, for a review). In particular GCMs will generally produce a simulation characterised by an unrealistically weak meridional circulation and a temperature structure that is too close to radiative equilibrium (i.e. colder than observed in the winter polar stratosphere and mesosphere, and warmer than observed in the summer polar mesosphere). It has become increasingly clear that these discrepancies result in large measure from the omission of the effects of eddies with scales too small to be resolved in current GCMs (e.g., Garcia and Boville, 1994). In particular, gravity waves with relatively small horizontal wavelengths (say from ~10 km to ~several hundred km) play a crucial role in driving the meridional circulation in the stratosphere. Lindzen (1981) pointed out that the qualitative effects of gravity waves on the mesospheric circulation should be to act as a drag on the zonal jets in the extra-tropical mesosphere and act to increase the strength of the mean meridional circulation. The effects of the wave drag are significantly non-local, and so gravity waves which break in the mesosphere affect the stratospheric circulation. Recently evidence has also been presented that gravity wave drag may be acting directly in the extra-tropical stratosphere (Rosenlof, 1996; Alexander and Rosenlof, 1996). Some of this drag could result from topographically-generated gravity waves (particularly in the Northern Hemisphere extra-tropics in winter), but nonstationary gravity waves excited by mechanisms such as convection or jet stream excitation must also be very important (and likely dominant in summer and in the Southern Hemisphere).

Convectively-generated gravity waves probably have the greatest direct impact in the tropics, due to the strong and widespread convection there. There is now renewed interest in the notion that relatively small scale gravity waves may be important in forcing the quasi-biennial oscillation (QBO) of the tropical stratosphere (e.g., Sato and Dunkerton, 1996; Alexander and Holton, 1997). The tropical QBO itself is now understood to be an important component of interannual variability in middle atmospheric climate, with effects felt even in the extra-tropics and polar regions. In fact, all major issues in stratospheric dynamics and transport are now believed to involve the interaction of relatively small-scale gravity waves with the larger-scale circulation. In order to obtain a realistic simulation of the global-scale transport circulation, GCMs must include a parameterisation of the eddy transports associated with unresolvable gravity waves.

Current attempts to incorporate the effects of unresolved gravity waves into global models (e.g. Lindzen, 1981; Fritts and Lu, 1993; Hines, 1997) suffer from an almost complete lack of knowledge of the actual spectrum of waves in the stratosphere, its variability and its dependence on tropospheric conditions. Typical of approaches adopted in some current climate models is an assumption that the input spectrum is spatially isotropic and geographically uniform over the entire globe (e.g., McFarlane et al., 1997). Slightly more ambitious is the approach of Manzini et al. (1997) who assume that the wave flux has a weak geographical dependence related to the long-term climatological precipitation at each location. At the other extreme are models which assume that the flux of waves upward into the stratosphere at each individual gridpoint and timestep depends entirely on the instantaneous convective stability of the tropospheric layers below (Rind et al., 1988a,b; Kershaw, 1995). At present there is no way of knowing how reasonable any of these approaches is. The enormous uncertainties involved in formulating subgrid-scale gravity wave drag parameterisations must be reduced if the field of global dynamical/chemical modelling of the stratosphere and UT/LS region is to progress on a sound basis.

Recently there has been considerable interest in the idea that the gravity waves in the UT/LS region could directly impact ozone chemistry. In particular the very temperature-sensitive heterogeneous chemistry on aerosols could be affected by the temperature fluctuations associated with gravity waves (e.g., Murphy and Gary, 1995; Tabazadeh, 1996). Detailed chemical models (whether employing self-consistently determined temperatures or "observed" temperatures taken from global meteorological data sets) may need parameterisations of the temperature fluctuations associated with gravity waves. For this, however, one needs a detailed knowledge of the spatial and temporal temperature spectra, including geographical variation and temporal intermittency. Information on the gravity wave spectrum near the tropopause may also be needed for a complete understanding of the cross-tropopause water vapour transport. In the tropical regions, where subvisible cirrus clouds near the tropopause are believed to play an important role in the dehydration of air entering the stratosphere (Jensen et al, 1996), rapid cooling rates due to gravity wave induced temperature fluctuations may limit such dehydration. This is because the size of ice particles formed (and thus the fallout rate) is inversely related to the cooling rate.

The overall goal of the proposed experiment is to characterise the gravity wave field generated by continental and oceanic convection in sufficient detail to aid in the formulation of parameterisations of the associated gravity wave effects in global atmospheric models. The campaign would focus on understanding and modelling the gravity wave source spectrum including energy, vertical eddy momentum flux, wave field anisotropies, and the temporal and spatial intermittency. Secondary goals include gaining a better understanding of wave propagation in the middle atmosphere (including the relationship to the background wind and temperature structure) and improved insights regarding the interaction of gravity waves with the mean flow in the stratosphere and mesosphere.

Some specific issues and questions that can be addressed by the intense field program proposed here include:

- How does the wave momentum flux entering the stratosphere scale with the intensity of convection? The answer to this is crucial for determining how much temporal intermittency and geographical variability should be assumed for the source spectra used in gravity wave drag parameterisations. Some recent work (Alexander and Holton, 1997) has suggested that a very limited number of intense African-wave thunderstorms could account for a very significant part of the mean flow acceleration in the QBO. More generally, how does the spectrum of waves entering the lower stratosphere depend on the intensity of convection?

- Still more generally, one could ask how the properties of the storms (e.g., peak intensity, suddenness of onset, duration, aspect ratio, degree of mesoscale organisation) as well as the background wind in the UT/LS region affect the spectrum of waves entering the lower stratosphere?

While efforts should be made to sample as wide a variety of conditions, any field experiment will provide only a small sample of seasonal and geographical variability in the real atmosphere. Thus it is critically important that the results be related to ongoing monitoring and numerical modelling efforts. In particular the following issues should be addressed.

- There are now limited-area numerical models which explicitly simulate tropospheric moist convection and stratospheric gravity wave generation. An important goal of a field experiment would be to provide detailed results for testing/validation of these models. The current state and immediate prospects for limited-area modelling and the role of the proposed field experiment in model testing are described in more detail in Section 9 below.

- Tropical clouds are monitored continuously with visible and IR satellite imagery and will be even more completely observed with the advent of the Tropical Rainfall Measuring Mission (TRMM, see Simpson et al., 1988). How well can we relate detailed observations of the wave flux in the lower stratosphere to satellite observations of cloud properties? This important question can really only be answered with a field program involving detailed in situ measurements above the cloud tops.

- A great deal of effort recently has been aimed at deriving properties of dominant lower stratospheric gravity waves from routinely-available single station radiosonde observations. Impressive results have been obtained (e.g., Allen and Vincent, 1995; Vincent et al., 1997; Guest et al., 1997; Whiteway and Duck, 1997), although there is some degree of controversy concerning interpretation of single station data (e.g., Eckermann and Hocking, 1989). Particularly interesting has been recent work using single station radiosonde data to make inferences concerning the role of gravity waves in driving the tropical QBO (Dunkerton and Sato, 1996; Sato, 1997). The proposed experiment will provide a possibility of comparing a more detailed picture of the wave field (obtained from aircraft, multiple balloons, ground-based profilers etc.) with the inferences from single balloon data. This is especially important as efforts (supported by SPARC) are now underway to gather as much radiosonde data from as many locations as possible, in order to characterise the seasonal and world-wide geographical variability of the gravity wave field in the lower stratosphere (Hamilton and Vincent, 1995).

There are also upper stratospheric and mesospheric issues that could be addressed by the proposed experiment:

- Observations with rockets, lidars and airglow imagers often show the presence of quasi-monochromatic waves in the middle atmosphere (e.g., Swenson et al., 1995). A key question still outstanding is whether such features can be related in some fairly simple way to intermittent sources. There have been speculations along these lines for at least two decades (e.g., Clark and Morone, 1981; Taylor and Hapgood, 1988), but clear evidence is lacking. By including upper stratospheric and mesospheric measurements (lidars, imagers, radars) in conjunction with the proposed tropospheric and lower stratospheric measurements, it may be possible to trace the effects of a fairly isolated gravity wave source on the wave field throughout the middle atmosphere. This task will be complicated by the restriction of some observational techniques to night-time and the large horizontal spread anticipated for the gravity waves by the time they reach the mesosphere, but a systematic effort of this kind should be undertaken and has the possibility of yielding a major advance in middle atmospheric meteorology.

The proposed experiment will provide an excellent opportunity to address these questions. In addition to the proposed intensive observations of the storm environment and storm-generated waves in the stratosphere, observations of winds in the mesosphere and lower thermosphere from the TIMED (Thermosphere-Ionosphere-Mesosphere-Energetics and Dynamics) satellite, due to be launched in early 2000, should be available. Combining these data sources will provide an unprecedented opportunity to place the mesosphere airglow observations in the detailed context needed for their interpretation, and also allow testing the middle atmosphere wave propagation models that now form the basis for gravity wave parameterisations.

There have been numerous observations of tropical gravity waves in the past, using a variety of techniques, including radar (Vincent and Lesicar, 1991), satellite temperatures (Fetzer and Gille, 1994; Wu and Waters, 1997), radiosondes (Tsuda et al., 1994; Allen and Vincent, 1995; Sato, 1996), and aircraft over-flights (Pfister et al., 1993). Each of these approaches has inherent limitations, however, in that they can observe only a limited region of the vertical wavenumber, horizontal wavenumber, and frequency spectrum. In particular, except for aircraft, none of the techniques can directly observe the horizontal structure of mesoscale gravity waves (say ~20-200 km horizontal wavelengths) directly, though the radar techniques can obtain the momentum fluxes associated with mesoscale gravity waves at mesospheric altitudes. The fact that all these techniques have detected strong gravity wave variance demonstrates an important fact, namely that gravity waves are a very broad-band phenomenon in the tropical atmosphere. Another limitation in previous observations is that, particularly for the aircraft campaigns, the gravity wave observations were often a by-product of experiments focused on other objectives.

This points to the need for a more comprehensive approach, employing a variety of measurement techniques simultaneously, in order to understand the spectrum of vertical momentum flux due to gravity waves. Since a global approach with all techniques is obviously not practicable, it makes sense to focus on the primary source of gravity waves in the tropics, namely individual convective systems.

The fact that gravity wave activity in mid-latitudes is related to convection was shown clearly by Fritts and Nastrom (1992), who detected substantial (factors of 10 or more) enhancements in mesoscale temperature and horizontal wind variance in upper tropospheric aircraft observations as the aircraft passed over convective systems. This is also apparent in the tropics, as Fig. 1 (Pfister et al. in preparation) shows. This figure is from an over-flight of the ITCZ during the recent NASA Stratospheric Tracers of Atmospheric Transport (STRAT) campaign. It shows a clear enhancement of the vertical wind variance over the region of low brightness temperatures, which is the locus of a mature convective system. The data also show the richness of the horizontal wavenumber spectrum, with vertical wind variance peaking in the 5-10 km wavelength range, and temperature variance more apparent in the 50 km wavelength range. Vertical wavelengths for the latter were estimated at about 5.5 km. Fig. 1, and other previous work (Pfister et al., 1993) from the Stratosphere-Troposphere Exchange Project (STEP) show the ability of the aircraft to capture particular mesoscale wave phenomena. Another approach, used by Alexander and Pfister (1995), evaluated the overall momentum flux from a broad spectrum of mesoscale waves on a particular over-flight during the 1987 STEP campaign. These two approaches are a beginning in establishing the spectrum of upward flux of momentum as a function of horizontal phase velocity. This phase velocity spectrum is critical to solving the gravity wave parameterisation problem since the phase velocity governs the level at which a gravity wave can break (and deposit momentum).

Fig. 1 also shows the limitation of current data sets. First, since the focus of the STRAT program was not on gravity waves, there are only two (fortuitous) over-flights, 3 hours apart, of a convective system with a characteristic variation time scale (as seen by half-hourly satellite photos) of perhaps 1-2 hours. One feature of the observations was that there was no evidence of the large vertical wavelength (10 km) gravity waves seen in limited-area models of explicitly-resolved convection. However, this may be entirely due to our inability to observe the system throughout its entire evolution, especially since these large vertical wavelength waves propagate upward quite rapidly. Second, there is no auxiliary data (other than meteorological satellite data) to obtain corroborating vertical wavelength information and information about the interior evolution of the convective system. These sorts of auxiliary data are critical to relating real observed convective events to modelled convection.

The Victoria workshop participants identified the region around Darwin Australia (12°S, 131°E) as a uniquely desirable venue for the proposed experiment. This choice was motivated by a combination of geophysical characteristics of the site and some practical considerations.

The key point about Darwin is that it is in the deep tropics and typically experiences a wide range of convective events, including some of the most intense thunderstorms and deepest convective towers generally seen in the tropics. This is particularly important, given the serious possibility that much of the total gravity wave momentum flux entering the stratosphere could be caused by a small number of extremely intense (and penetrative) convective events. The relatively low latitude of Darwin also implies that the results of a field experiment held there should be applicable to the gravity waves that may be driving the QBO.

Fig. 2 shows the geography of the Darwin area. In the late pre-monsoon period there is a very regular diurnal cycle of intense convection on Bathurst and Melville Islands (Keenan et al., 1990; Keenan and Carbone, 1992). This phenomenon is driven by the sea breeze and is so regular as to have earned the local nickname of "Hector". Hector typically begins shortly after local noon and builds into intense thunderstorms that penetrate the tropopause before moving off-shore to the west of Bathurst Island where it then dissipates by mid-evening. Updrafts as strong as 40 m/s and cloud tops to almost 20 km have been observed during Hector. The usual lifetime of Hector is about 8 hours.

In addition to the regular sea-breeze convection over Bathurst and Melville Islands, there is less intense oceanic convection that develops north of Darwin typically starting in early January. Finally, there is also continental convection south of Darwin that becomes more frequent in the monsoon season.

The interesting character of the convection in the area led to Darwin being chosen as the site for the MCTEX (Maritime Continent Thunderstorm Experiment) campaign conducted in November and December 1995. This was a joint project of the Australian Bureau of Meteorology and researchers at NCAR, NASA Goddard Space Flight Center, Colorado State University, and Monash University, and included contributions from groups at several other institutions in the USA, Australia and Japan. MCTEX has a Web page at http://www.bom.gov.au/bmrc/meso/Project/research/index.htm. The MCTEX experience will provide a very valuable base for understanding the tropospheric convection in the region. The proposed 2000 or 2001 date for the CEGWE gravity wave experiment will allow the MCTEX results to be thoroughly analysed and then used in the detailed planning of the new field campaign.

The MCTEX experiment observing period lasted from November 15 to December 8. On average the beginning of the summer monsoon at Darwin occurs about December 10, but there is considerable variability (with the monsoon starting later in El Nino years and earlier in La Nina years). It seems likely that an expedition based in Darwin and lasting, say, from late October into mid-December would be virtually guaranteed a large number of days with only the localised Hector convection, and would possibly extend into the beginning of the monsoon period, allowing some sampling of days with more widespread continental and/or oceanic convection. An experiment in October-December 2000 would overlap with the nominal three-year TRMM satellite mission. If CEGWE were to be scheduled for 2001 it would presumably entail an extension to the TRMM mission. The Australian Bureau of Meteorology has indicated its willingness to keep the ground based instruments in the Darwin area in place beyond the nominal TRMM mission period if they are needed for CEGWE

Another key advantage of the proposed venue is that Darwin is a TRMM validation site and, as a consequence, is extensively instrumented for meteorological monitoring of convection (Keenan and Manton, 1996; also see the Web page of the Darwin Climate Monitoring Research Station at http://www.bom.gov.au/bmrc/meso/Project/Darwin.html). Fig. 2 shows the instruments that are now deployed for the TRMM monitoring, including a dual polarised Doppler radar, 50 MHz and 920 MHz profilers, rainguage networks, disdrometers, automatic surface weather stations, radiation instruments and lightning detectors. In addition, regular rawindsonde observations are taken near Darwin. The Bureau of Meteorology and other research groups in Australia are enthusiastic about the experiment proposed here, and so it is certain that the Darwin monitoring network would be significantly enhanced during the expected intensive observing period.

Darwin has an excellent civilian infrastructure, including weather forecast facilities and access to real time satellite data. There has been considerable local experience in mounting a large field campaign developed during MCTEX. Darwin was also a base for the tropical campaign of the Stratosphere-Troposphere Exchange Project (STEP), during which the NASA ER-2 flew out of Darwin Airport. There is also a military air base at Tindal, 275 km south of Darwin, that can also be used (and may be preferable at times when bad weather could restrict flying out of Darwin).

Fig. 2. Meteorological monitoring network deployed at the Darwin site. The circles show the regions in which dual-Doppler synthesis of the storm wind field is possible with the two ground-based radars which are located at the interection points of the circles. A much larger area of dual-Doppler synthesis would be possible with the aircraft-mounted Eldora radar.

For the experiment to be most successful, it would be desirable to measure the complete wind and temperature fields as well as important minor constituents (such as water vapour and ozone) over as large a height range as possible. High-resolution observations within and near the convection cells and to altitudes in excess of 30 km are essential. It is envisioned that the field program would include a wide variety of ground-based and airborne remote sensing and in situ observations as well as observations from satellites. Balloon sondes, dropsondes, radars, lidars, imagers, radiometers, spectrometers, and satellites sensors are all expected to make important contributions.

The characteristics of the gravity wave fields generated by convection are related to many features of the source, including geometry and lifetime, large scale motion, and strength of the updrafts. The convection must be characterised in sufficient detail to permit the strength and properties of the gravity wave response to be related to the strength and properties of the convective activity. An extensive network of relevant meteorological observations will be essential for monitoring key parameters of the convective activity, its physical extent, and temporal evolution. Oceanic convection is characterised by weak updrafts (<10 m-s-1), low convective available potential energy (CAPE) and glaciated conditions at upper levels. Continental and island convection has strong updrafts (>20 m-s-1) and can extend well into the stratosphere. The associated gravity wave fields generated by these various types of convection are expected to differ as well.

Desirable monitoring capabilities include satellite imagery to characterise the distribution, physical extent, and temporal evolution of the convection cells, surface measurements of winds, temperature, and barometric pressure to characterise the surface conditions leading to and sustaining the convective activity, radiosonde and dropsonde measurements of wind and temperature profiles to characterise the upper level conditions associated with the convective activity, meteorological radars to monitor the rainfall and updrafts in the interior of the convection cells, and airborne radars and in situ microphysical sampling systems to define the structure and evolution of the convection.

In order to completely characterise the gravity wave fields generated by convective activity it is necessary to determine the distribution of wave energy as a function of vertical and horizontal wavelengths, intrinsic period, and propagation direction. While knowledge of the wave perturbations in winds, temperature, and density are all required to deduce the momentum, heat, and constituent transport associated with dissipating waves, the source spectrum may be characterised by employing a combination of various parameter observations and the gravity wave polarisation and dispersion relations. One of the most important features of the wave field is the distribution of energy as a function of propagation direction. Unfortunately, this feature is also the most difficult one to measure.

The wave spectrum can be measured by using radiosonde balloon observations of the wind and temperature profiles, radar observations of wind profiles, lidar measurements of temperature profiles, and in situ airborne measurements of the winds and temperatures. To be most useful these observations should extend well above the tropopause to altitudes in excess of 30 km if possible. Vertical and horizontal profiles as well as time series are desirable so that the full vertical wave number, horizontal wave number, and temporal frequency spectrum can be characterised. Wave propagation directions in the troposphere and lower stratosphere can be derived from radiosonde balloon profiles of temperature and winds (e.g., Vincent et al., 1997) and from airborne in situ and remote sensing observations (e.g., Gardner, 1991). In order to adequately characterise the anisotropy of the wave field, balloons must be launched from several sites surrounding the convection region and aircraft must make several transects at various azimuths over the convection cell.

As the wave field propagates to higher altitudes in the middle atmosphere, some components of the spectrum will grow in amplitude as the density decreases while other components will be damped by instability processes and critical layer interactions associated with the height varying background wind field. Measurements of the wind and temperature profiles throughout the middle atmosphere as well as observations of the wave spectrum are required to characterise and quantify these effects. Ground-based radars, and ground-based and airborne lidars, airglow imagers, and airglow spectrometers can provide much of the requisite data.

Accurate high resolution wind, temperature, and constituent profiles are required to quantify the momentum, heat, and constituent transport associated with dissipating gravity waves. Ground-based radars and lidars can provide much of the data while airglow imagers can be used to determine the relationships to wave propagation directions.

The NASA ER-2 aircraft is the platform of choice for obtaining the all-important horizontal perspective on the mesoscale gravity waves generated by convective systems. This aircraft has a long track record, and it is probably the only large payload aircraft that can cruise at high enough altitudes to easily clear deep systems such as Hector (which may exceed 17-18 km in altitude). It can stay in the air up to 8 hours, so it can easily fly continuously throughout the life cycle of this diurnal thunderstorm. With a cruising speed of about 200 m-s-1, it can cross a large tropical convective system with room to spare in 20 minutes, allowing 15 or more over-flights of a convective system during its life cycle. It is sensitive to surface weather conditions, particularly crosswinds. Nevertheless, this aircraft operated successfully out of Darwin, Australia during the 1987 STEP campaign. It will be the primary means by which the spectrum of vertical momentum flux as a function of horizontal phase velocity is obtained. Recently, NASA has begun to operate the WB-57 aircraft which has somewhat similar characteristics and can be equipped with much of the instrumentation developed for the ER-2. It appears that the WB-57 could also serve as the high-altitude platform for CEGWE. The following paragraphs describe the proposed payload for a high-altitude aircraft.

The Meteorological Measurement System (MMS) is an on-board system for measuring the 3D wind vector, temperature, and pressure at a 5 Hz data rate. The accuracy claimed is 0.5 K for temperature, 0.5 m-s-1 for horizontal wind, and 0.2 m-s-1 for vertical wind. Temperatures are measured via platinum resistance thermistors having a response time scale of about 0.4 seconds. Winds are derived by evaluating the velocity of the air with respect to the aircraft using flow-angle sensors, and the velocity of the aircraft with respect to the ground using an inertial navigation system (INS), the global positioning system (GPS), and pressure sensors (for vertical speed of the aircraft). Details of the instrument are given in Scott et al. (1990), and an intercomparison with radiosondes is described in Gaines et al. (1992).

Given the importance of the vertical flux of momentum, the most critical measurement is the vertical wind. It is also the most difficult, for two reasons. The vertical speed of the aircraft cannot be derived only using the integral of the INS vertical accelerometer, since such integrations tend to drift. Nevertheless, the MMS has a demonstrated ability to measure vertical winds associated with moderate amplitude mountain waves (Chan et al., 1994). The amplitudes of convectively-generated waves tend to be substantially smaller, however, with peak-to-peak displacements of order 300 meters rather than the 1-2 km typical of mountain waves. Also, and this is perhaps the most difficult aspect of the measurement, the convectively generated mesoscale gravity waves of greatest importance to the stratosphere are not necessarily those that have the largest vertical wind amplitude. For example, the largest amplitude vertical winds in Fig. 1 (1.5 m-s-1 peak-to-peak near 6.5-7.5 degrees latitude) occur at fairly short horizontal wavelengths (8-10 km). Changes in the horizontal wind as these waves propagate upward could easily increase the intrinsic wave frequency above the Brunt-Vaisala frequency, thus inhibiting vertical propagation. Of greater importance are the longer waves (seen in the temperature between 7.8 and 9.8 degrees latitude), which are unlikely to be reflected. However, these have substantially smaller amplitudes in the vertical wind (peak-to-peak values of 0.2-0.3 m-s-1 - see the low-pass filtered vertical winds in Fig. 3). Vertical aircraft motions on these spatial scales are substantially larger than this, so one is faced with the classic "small difference between large terms" problem. Still, as Fig. 3 shows, the northward and upward pointing phase lines evident from the temperature cross-section in the bottom panel, along with the positive correlation of meridional wind and the low pass filtered vertical wind (as well as the quadrature relationship between the temperature and the meridional wind) form a consistent picture. In fact, the vertical wavelengths derived from the relationship of horizontal and vertical wind are consistent with those derived from the relationship of temperature and vertical temperature gradient (about 5.5 km). It is clearly necessary, though, to solidify this measurement, and efforts are underway to improve the evaluation of the vertical aircraft speed, including the use of differential GPS techniques.

The microwave temperature profiler (MTP) is a remote passive instrument that uses the emissions of oxygen lines at 57.3 and 58.8 GHz to provide a vertical temperature profile in a layer up to 10 km deep. Instrument details are given in Gary (1989). The weighting functions increase in depth with distance from the aircraft, so the vertical resolution of the data is best at aircraft altitude. It reports data every 10 seconds, rather than at 5 Hz. This translates into a profile every 2 km, not a serious limitation for mesoscale waves with wavelengths of 20 km or greater. The function of the MTP will be to provide a vertical cross-sectional picture of the wave field, thus aiding in the interpretation of the MMS data. Also, since the aircraft does not fly at exactly constant altitude, the MTP vertical temperature gradients can be used to estimate that part of the MMS-measured temperature variation that is due to vertical excursions of the aircraft (and thus extract the actual temperature variations at constant altitude).

Ozone has a substantial vertical gradient in the lower tropical stratosphere, and so measurements can provide useful information on vertical air parcel displacements. The measurement techniques are well established, and the existing ozone instrument (Proffitt et al., 1983) has excellent time resolution, accuracy, and precision.

Water vapour has only weak vertical gradients in the tropical stratosphere. However, it increases rapidly below the tropopause. Thus, it is a good indicator of the injection of tropospheric air into the stratosphere. It will be useful in distinguishing an irreversible exchange event from a wave feature. The details of the existing water vapour instrument, which provides 1 Hz data, are described in Kelly et al. (1989).

This instrument, described in Liou et al. (1990) will yield high resolution information on the cloud top brightness temperature. It is an important step in relating observed momentum fluxes to an indicator that can be measured globally from satellites. It will also yield important information on the cloud structure.

This instrument (Heymsfield et al., 1996) can provide a vertical cross section under the aircraft of cloud droplet and ice particle structure, as well as the locations of updrafts and downdrafts. It will enable a more thorough comparison of observed and modelled convective structures. In this way, one can evaluate the significance of similarities and differences between the observed and modelled convectively generated gravity waves.

Top panel : Solid (left axis) - temperature; dotted (right axis) - meridional wind.

Middle panel : Solid (left axis) - Low-pass filtered vertical wind; dotted (right axis) - vertical temperature gradient.

Dotted line in bottom panel is the altitude of the aircraft. Vertical temperature gradient and temeprature cross section are from the Microwave Temperature Profiler.

Balloon-borne radiosondes provide vertical profiles of temperature, humidity and horizontal wind. There is considerable experience in launching balloons in the general area around the proposed CEGWE venue, both from standard meteorological operations and in intense campaigns such as MCTEX and Japanese/Indonesian efforts in Indonesia (e.g., Tsuda et al., 1994). The experience of the Indonesia campaigns showed that the maximum height of the balloon measurements exceeded the tropopause (located around 16-18 km) for about 90% of soundings, and that more than 70% of soundings reached 30 km altitude. The effective height resolution of the data with current technology is about one hundred meters for temperature and several hundred meters for wind velocity, By the time of the CEGWE experiment, GPS tracking will likely improve the resolution and accuracy of the wind measurements. For CEGWE a network of stations around the area would be established to supplement the routine launches at the Darwin weather station. The experience gained in MCTEX will be very valuable in planning the locations of the launch sites. A number of Australian and New Zealand colleagues are interested in conducting the balloon operations for CEGWE.

It would be desirable to supplement the ER-2 measurements by other in situ aircraft observations. The Australian government has recently augmented its fleet of research aircraft and created an agency called Airborne Research Australia ("ARA", see http://www.es.flinders.edu.au/ARA). Their fleet now consists of four aircraft including the Beech 200T "King Air" recently purchased from NCAR, a Grob " G520T "Egrett" and a Cesna 404 "Titan" (which was used for in situ observations during in the MCTEX experiment). The Egrett is a particularly interesting plane since it can carry a payload of 750 kg to altitudes in excess of 15 km. The Australian ARA are also very involved in development of high quality probes for in situ airborne observations. An attempt would certainly be made to get one or more of the ARA planes for CEGWE, both for their ability to perform in situ measurements and as platforms for some of the optical instruments such as lidars and airglow imagers (see below). Several colleagues at Australian universities have expressed an interest in applying for grant funds to at least partially defray the operating costs of the any ARA aircraft that could be used in CEGWE.

A strong effort will also be made to obtain the NCAR Research Aviation Facility "Electra" aircraft for CEGWE. This would serve principally as a platform for the ELDORA Doppler radar (see below), but it could also be equipped for some in situ measurements.

The Australian Bureau of Meteorology is heavily involved in development of small, unmanned fixed-wing aircraft for meteorological observations. The "aerosonde" which they have developed was deployed as part of MCTEX. This technique could provide another important contribution to the tropospheric meteorological observations in CEGWE.

Monitoring overall storm production in the vicinity of Darwin is crucial to CEGWE. The BMRC/NCAR C-Band polarmetric/Doppler (C-POL) radar will be located at Gunn Pt approximately 50-100 km south of all convection that develops over Bathurst and Melville Islands and well within range of all convection affecting Darwin. This radar will undertake a primary monitoring role to reveal the structure and kinematic evolution of the storms in the region. With the polarimetric capability inferences on the cloud microphysical and precipitation processes will be possible. This will be particularly important given the differences in convection that occur i.e. continental and oceanic type storms. Little is known about the relation of these two storm types to generation of gravity waves.

Based on MCTEX and other studies there is strong evidence for a microphysical link in the dynamics of continental storms. Keenan et al. (1998) show evidence for the existence of supercooled water before the onset of explosive deep storm growth over the Bathurst and Melville Islands. This is the stage in continental type storms when large updraft development first occurs and when the storms become capable of penetrating into the stratosphere. At the other extreme, oceanic storms have much weaker updrafts and are associated with glaciated conditions aloft. What are the differences in gravity wave production? Before one can understand and model updraft growth that is important in creating stratospheric gravity waves a complete understanding of the storms microphysical characteristics may be necessary. C-POL will provide this information in a variety of storm types.

Later in the convective lifecycle there is organisation and development of a propagating mesoscale system. Yang and Houze (1995) propose that convective cells in these multicellular complexes are initiated by gravity waves propagating from both sides of the gust front. The multicellular structure is associated with trapped gravity waves. Detailed knowledge is therefore required of the mesoscale organisation and the pulsating updrafts that perturb and penetrate the tropopause to initiate vertically propagating stratospheric waves. C-POL will provide the reflectivity and microphysical picture including the storm structure in the upper troposphere.

The relation of gravity wave production to the onset of new convection is another important feedback mechanism. The ground based radars will again be important in terms of defining where and when new convection develops and the associated triggering mechanisms.

Hence detailed four dimensional knowledge of the kinematic, microphysical and dynamical factors influencing storm evolution is necessary to appreciate how and when gravity wave generation occurs. The polarimetric radar will provide overall storm monitoring with an emphasis on the microphysical and rainfall production aspects. Variables measured by C-POL include: radar reflectivity, differential reflectivity, radial Doppler velocity, specific differential phase shift, zero-lag correlation coefficient and linear depolarisation ratio.

The second BMRC C-Band Doppler radar located 30 km south of Gunn Pt. at Berrimah will provide additional storm surveillance information, and, when coupled with C-POL, will enable dual-Doppler synthesis of storms near Darwin in the areas shown in Fig. 2. This capability will enable the full four dimensional evolution of storms to be observed within the limited dual-Doppler areas indicated.

As part of MCTEX the University of Massachusetts deployed a vertically-pointing cloud-profiling radar (operating at 32 and 95 GHz) which usefully supplemented the other radar measurements. An attempt would be made to have a similar deployment during CEGWE.

Not all convection develops in one particular location, and at one time (although Hector does at this location) where an observational framework can be put in place to monitor the full dynamical evolution and the relation to gravity wave generation. As stated above some Dual-Doppler observations will be available over the mainland and surrounding ocean but not over the key areas of Bathurst and Melville Islands and within oceanic convection. The airborne flexibility of ELDORA makes it possible to undertake monitoring of the dynamical evolution of such storms from start to end where ever they occur i.e. the oceanic, island and continental storms. Coupled with the microphysical and routine environmental monitoring of C-POL radar this provides the full four dimensional evolution of the flow field up into the stratosphere by dual-Doppler synthesis. A very strong observational basis is therefore provided for modelling storm characteristics and the subsequent growth of gravity waves originating within 100 km of C-POL.

The focus of attention will be Bathurst and Melville Islands. ELDORA will typically fly some 15-20 km away from the storm near 4-5 km height in straight line legs or curved patterns for airborne dual-Doppler measurement (7-8 km-min-1). It will be able to monitor the complete storm life cycle (~8 hours). It will provide unprecedented complete dual-Doppler synthesis of the initial growth, merger, and convective scale processes such as penetrative convection involved in the formation of gravity waves. Full kinematic details on explosive cell growth and perturbation of the tropopause will be available. Its high sensitivity will enable continuous monitoring of the upper characteristics of the storms. Attenuation effects are important, but with C-POL and employing different look angles these factors can be overcome in most cases, especially in the upper regions of the storms.

The original plan for MCTEX included the deployment of the ELDORA radar on the NCAR Electra aircraft. In the end NCAR could not make the Electra available and this represented a significant disappointment in terms of the results available. A major effort will be undertaken to lobby for the provision of the Electra for CEGWE. The tropospheric results from CEGWE would then represent a significant improvement over those available from MCTEX. The C-POL ground based radar for environmental monitoring and discerning microphysical characteristics of storms, coupled with ELDORA for complete dual-Doppler synthesis of the storms, will produce an unprecedented data base for the study of tropical moist convection.

As part of MCTEX the U.S. Department of Energy ARM program supported deployment of the University of Massachusetts Cloud Profiling Radar System (CPRS) on Melville Island. CPRS operates simultaneously at 33 GHz and 95 GHz through a single antenna aperture and is fully polarimetric, capable of recording LDR and ZDR. Doppler spectrum moments can be obtained via pulse-pair measurements, or full Doppler spectra can be recorded. CPRS is also mounted on a elevation-over-azimuth pedestal that can be programmed to perform various scan patterns, including VAD scans from which windestimates can be obtained. During MCTEX co-ordinated C-pol radar and CPRS scans were made to detail initial stages of convection.

Multi-wavelength analysis of CPRS MCTEX data has permitted estimation of particle size distributions in ice clouds and vertical wind estimation in precipitation. Recent work with full Doppler spectra FFT measurements of precipitation has shown that vertical air motions below the melting layer can be resolved with high precision and high time resolution (less than two seconds). There is very strong interest from the U. Massachusetts group in deploying their radar as part of CEGWE.

Two wind-measuring profilers are located in the Darwin area and are jointly operated by the Bureau of Meteorology and NOAA Aeronomy Laboratory. One profiler, which operates at 50 MHz, is capable of measuring horizontal winds with range resolutions of between 0.45 and 1 km up to heights between 13 and 16 km. The maximum height coverage can extend up to 20 km for vertical velocities measured with a vertically pointing beam. The other profiler operates at 920 MHz and is used for boundary-layer studies for heights ranging from 100 m to between ~2-9 km, depending on the conditions. Both radars are used to obtain other important meteorological information, including the vertical structure of rain and ice dropsize distributions. In the event that CEGWE proceeds, the University of Adelaide plans to place a 50 MHz boundary-layer radar on either Bathurst or Melville Island with the aim of measuring winds and atmospheric structure directly in the region of generation of Hector. This instrument will be capable of observing up to the mid-troposphere.

The University of Adelaide plans to place an MF radar operating at a frequency of 2 MHz in the vicinity of Darwin. The radar will be capable of continuously measuring horizontal winds between 60 and 100 km during the day and between 80 and 100 km at night. It is planned to operate the radar for periods of up to a year either side of the core period of the experiment in order to provide background information on the dynamical state of the mesosphere and lower thermosphere. The good height (2-4 km) and time resolutions (~2-4 min) would be very suitable for determining a climatology of gravity wave activity as well as the mean structure and tides.

In layers where the air does not contain a substantial amount of aerosol, the Rayleigh lidar technique can be used to obtain an absolute measure of atmospheric temperature (Hauchecorne and Chanin, 1980). Fig. 4 shows an example of a temperature profile obtained with a Rayleigh lidar, and clearly demonstrates the ability of this technique to resolve the temperature fluctuations induced by atmospheric gravity waves. Typical Rayleigh lidars employ averaging times of order 10 minutes, with a vertical resolution of 300 meters for measurements up to about 60 km. If averaging times of several hours are used then temperature can be measured up to about 90 km. These measurements are limited to air which does not contain a substantial amount of aerosol, so the lower boundary is at about 25 km, and, of course, useful measurements are restricted to night-time.

Rayleigh lidar observations have made major contributions to the study of gravity waves in the middle atmosphere (e.g., Wilson et al., 1991; Beatty et al., 1992; Whiteway et al., 1997). Measurements have been carried out from ground and aircraft and it is envisioned that both platforms will be employed for CEGWE. The lidar group at the University of Illinois has developed a system which combines Rayleigh and sodium resonance backscatter (see next section). This lidar was successfully operated aboard the NCAR Electra aircraft during the ALOHA-90 campaign (Gardner, 1991) and it would be applied in a similar manner for CEGWE. Airborne lidar measurements make a unique contribution to gravity wave studies since both the vertical and horizontal spectra can be measured (Hostetler and Gardner, 1994). A ground based lidar could employ a larger telescope and would be capable of providing finer resolution to complement the airborne observations.

Rayleigh lidar may also be used to measure wind speed in the height range of 25 km to about 60 km. Incoherent Doppler lidar measurements of wind speed have been demonstrated (Tepley et al., 1991; Chanin et al., 1989) and such a system could make an important contribution to CEGWE. For instance, a credible test of numerical models of the generation and propagation of gravity waves will require the wind velocity to be measured from ground to the mesopause. If only radiosondes and radar are available for wind measurements then there will be a significant gap between 30 and 60 km.

Rayleigh lidars are capable of detecting the excess scattering from aerosol and cloud particles. Measurements of aerosol scattering profiles would be useful in studies of tropopause dynamics by providing a tracer which could indicate exchange of air between the upper troposphere and lower stratosphere.

Airglow emissions from atomic oxygen recombination processes originate from four separate layers in the 80-110 km region. These include the OH Meinel (near IR and IR), Na (589.2 nm), O2 atmospheric band (865 nm), and OI (557.7 nm). The altitudes of peak emission lie between emission layer thicknesses of ~8 km. Waves propagating through the layers redistribute the constituents responsible for the airglow, and subsequently produce brightness distributions related to the dynamics associated with waves. Measurement and imaging of the airglow intensity thus may be useful for observing gravity waves near mesopause heights, at least those waves with sufficiently long vertical wavelengths (say greater than ~15 km). OH emission has been studied most often as it is the lowest emission layer, but airglow wave observing has been successfully demonstrated for all emissions using instruments on aircraft platforms (Swenson and Espy, 1995). Ground-based methods have matured over the past 10 years to include superb signal-to-noise in broad-band (Taylor et al., 1987; Swenson and Mende, 1994) as well as narrow-band for OH and O2 bands (Hecht et al. 1995). A few instruments have become available to image the OH and O2 images to accomplish temperature projections in the respective layers.

Fig. 4. (a) Profile of temperature from a Rayleigh lidar at Eureka (80°N, 86°W).

(b) Temperature perturbations from the smoothed curve shown in (a), expressed as percentages

Airglow imagers would be fielded on an aircraft as well as on the ground at a minimum of two locations for the CEGWE campaign. The imagers produce maps of phase front projections and propagations for the study of sources. Taylor and Hapgood (1988) have used this to identify cumulonimbus sources of gravity waves. Swenson and Espy (1995) tracked a large wavefront in airglow using similar information. Airglow intensity measurements for OH have been well calibrated to the atmospheric response so that 2D unambiguous spectra of image sequences can be obtained (Gardner et al., 1996). Methods have been developed to extract observed phase speeds from moving waves, and assuming mean winds can be measured (e.g., using MF radar or lidar); this results in a calculated intrinsic phase speed from both ground based and aircraft imagery (Taylor et al.., 1995; Swenson et al., 1995).

Even the ambitious field program proposed here cannot hope to observe all of the details of the three-dimensional, time-dependent development of a convective system and the gravity wave field surrounding it. Limited-area, high-resolution models will be needed to place the various observations in context, thereby greatly aiding in their interpretation. Such models to date have been used in rather highly idealised studies, lacking in sufficient observational data to constrain them. Intensive observations and numerical model simulations will be used together in this campaign to characterise the convection and the convectively-generated gravity waves in unprecedented detail, permitting quantitative estimates of the global effects of these waves in the middle atmosphere. The results will also test some previous, less well constrained, model-based estimates of these effects.

High-resolution numerical models have been applied to simulation of extra-tropical mesoscale convective systems with some degree of success since the 1970's (e.g., Klemp and Wilhelmson, 1978). Simulations of deep tropical convection soon followed (e.g., Soong and Tao, 1984; Redelsberger and Lafore, 1988; Lafore and Moncreiff, 1989; Tao and Simpson, 1989). These models typically employ grid spacing of the order of one km in the horizontal and 300 m in the vertical, and explicitly include cloud water, rain water and ice phase precipitation as variables. The cloud physics in such models is handled by bulk parameterisations. At present there are at least ten groups world-wide that are engaged in detailed numerical modelling of atmospheric moist convection.

The investigators who have conducted such numerical experiments have generally not studied the upward flux of gravity waves produced by the convection in their models, and most existing simulations have not produced useful information on this issue. In some cases the models use an approximate "anelastic" form of the governing equations that limits the validity of the solutions to some fairly shallow depth. And, of course, modellers have been reluctant to devote a significant fraction of their model domain to the region above the tropospheric convection that they aimed to simulate. The first papers to examine gravity waves in cloud model simulations appear to be those of Clark et al. (1986) and Hauf and Clark (1989) who studied gravity waves near the tropopause resulting from fairly shallow convection as simulated by a 2D model and a 3D model, respectively. Fovell et al. (1992), Holton and Durran (1993), Alexander et al. (1995), and Alexander and Holton (1997) have performed more extensive investigations of the stratospheric gravity wave field forced by deep convection in high-resolution 2D models. Fig. 5 shows instantaneous results from two of the Alexander and Holton (1997) simulations of intense tropical squall lines. In each panel the heavy solid curve outlines the simulated cloud, the thin lines show isentropic surfaces and the background shading represents the vertical velocity field. The two cases shown are for nearly identical tropospheric squall lines, but with different stratospheric wind profiles: one appropriate for the extreme west phase of the QBO, and one for the east phase. In each case the convection has disturbed the material surfaces in the lower stratosphere and produced a complicated pattern of upward-propagating gravity waves. As Alexander and Holton (1997) note, the differing patterns of gravity waves in west phase and east phase QBO winds is associated with a strong QBO modulation of the eddy momentum transports associated with the waves. This in turn provides a mechanism by which the convectively-excited gravity waves can contribute to the dynamical forcing of the QBO.

The recent interest in numerical simulation of convectively-forced gravity waves reflects increased appreciation of the importance of gravity waves in climate, increased experience with the detailed simulation of moist convection in the troposphere, and an improvement in available computer resources. There seems to be agreement that the models need to have horizontal resolution no coarser than ~1-1.5 km in order to produce adequate simulations of the tropospheric convective elements. Such high-resolution models are very expensive to run in 3D, particularly if a model domain adequate to study the gravity waves in the lower stratosphere is employed. It is only recently that such integrations could be regarded as practical, and it is expected that over the next few years numerical simulation of convectively-excited gravity waves will be addressed by more research groups. At present, in addition to the project of the University of Washington group (J. Alexander, J. Holton, D. Durran), there are efforts underway at GFDL/NOAA (V. Balaji, K. Hamilton), the University of Wisconsin (M. Hitchman) and the UK Meteorological Office (R. Kershaw).

The kinds of numerical experiments of interest to CEGWE fall into two broad categories. One would be somewhat idealised experiments designed to examine the gravity wave spectrum above convection in a variety of conditions (mean winds, initial mean sounding, initial perturbations). Kershaw (1995) has attempted something along these lines for rather shallow extra-tropical convection, and has even used the results as a basis for a simple gravity wave parameterisation for GCMs. This kind of work should be extended to a range of tropical convection situations, and the results should be examined in light of the observations that ultimately would be available from the CEGWE field campaign. The second category of experiments would be those aimed at gaining the ability to simulate a particular observed convective event as closely as possible. This is a major challenge at present, and would require considerable development of the models themselves (treatment of topography and exchanges with the surface, implementation of lateral boundary conditions) and of methods of initialisation. There have been some fairly encouraging recent efforts in terms of simulating specific observed tropical squall lines (e.g., Trier et al, 1996), but this remains a major challenge. The problem of simulating the sea-breeze-driven diurnal convection (of the sort seen north of Darwin) for a particular day is just beginning to be attacked now. It is hoped that financial support for numerical model development and experimentation could be obtained as part of the CEGWE project during the years leading up to the actual field experiment.

10. Additional Research Efforts During the CEGWE Period - Possible Connections with Other Programs and Satellite Missions

The detailed characterisation of the convection that is expected during the CEGWE intensive observational period and the availability of airborne platforms makes it very likely that research groups with other interests in tropical convection would join. The obvious candidates are those interested in cloud microphysics/chemistry, cloud radiative properties or stratosphere-troposphere exchange. Many of the same features that make the proposed Darwin venue and October-December timing attractive for CEGWE would also be valuable for other projects related to tropical convection. One obvious potential partner for the CEGWE experiment is FIRE ("First ISSCP Regional Experiment"). In their Phase III Implementation Plan (written in May 1996) the FIRE Science Team outlined the general need for an ambitious field experiment to study the microphysics and radiative properties of tropical cirrus. The isolated intense convection over Bathurst and Mellville Islands provides a particularly simple system in which to study the formation of tropical cirrus from cumulus anvils.

The CEGWE proposal has obvious relevance to two important satellite missions. The TRMM mission is scheduled to be launched in November 1997 and the planned three year mission lifetime would overlap the proposed 2000 CEGWE period. If CEGWE were to be held in 2001 it would be very desirable to have the TRMM mission continued. While there are other TRMM validation missions scheduled for the ER-2, it seems likely that the extensive array of observations proposed for CEGWE would aid in the validation/interpretation of the TRMM measurements. Certainly from the CEGWE viewpoint it would be very desirable to have the TRMM radar, microwave and IR instrument data available.

The proposed timing for CEGWE in either 2000 or 2001 would put it within the planned mission lifetime of the Thermosphere Ionosphere Mesosphere Dynamics and Energetics (TIMED) satellite (http://sd-www.jhuapl.edu/TIMED). The expected concentration at the Darwin site of radar and optical instruments capable of probing the mesosphere and lower thermosphere would make CEGWE a valuable opportunity for comparison with the observations from the TIMED satellite.

Alexander, M.J., J. R. Holton and D.R. Durran, 1995: The gravity wave response above deep convection in a squall line simulation. Journal of the Atmospheric Sciences, 52, 2212-2226.

Alexander, M.J., and J.R. Holton, 1997. A model study of zonal forcing in the equatorial stratosphere by convectively induced gravity waves. Journal of the Atmospheric Sciences, 54, 408-419.

Alexander, M.J. and L. Pfister, 1995: Gravity wave momentum flux in the lower stratosphere over convection, Geophysical Research Letters, 22, 2029-2032.

Alexander, M.J. and K. Rosenlof, 1996: Nonstationary gravity wave forcing of the stratospheric zonal mean wind. Journal of Geophysical Research, 101, 23465-23474.

Allen, S. and R.A. Vincent, 1995: Gravity wave activity in the lower atmosphere: seasonal and latitudinal variations. Journal of Geophysical Research, 100, 1327-1350.

Beatty, T. J., C. A. Hostetler, and C. S. Gardner, 1992: Lidar observations of gravity waves and their spectra near the mesopause and stratopause at Arecibo. Journal of the Atmospheric Sciences, 49, 477-496.

Chan, K. R., L. Pfister, T. P. Bui, S. W. Bowen, J. Dean-Day, B. L. Gary, D. W. Fahey, K. Kelly, C. R. Webster, and R. D. May, 1993: A case study of the mountain lee wave event of January 6, 1992, Geophysical Research Letters, 20, 2551-2554.

Chanin, M. L., A. Granier, A. Hauchecorne, and J. Porteneuve, 1989: A Doppler lidar for measuring winds in the middle atmosphere. Geophysical Research Letters, 16, 1273-1276.

Clark, J.H.E. and L.T. Morone, 1981: Mesospheric heating due to convectively excited gravity waves, a case study, Monthly Weather Review, 109, 990-1001.

Clark, T.E., T. Hauf, and J.P. Kuettner, 1986: Convectively forced internal gravity waves: results from two-dimensional numerical experiments. Quarterly Journal of the Royal Meteorological Society, 112, 899-925.

Eckermann, S.D. and W.K. Hocking, 1989: Effect of superposition on measurements of atmospheric gravity waves: A cautionary note and some reinterpretations. Journal of Geophysical Research, 94, 6333-6339.

Fels, S.B., 1987: Response of the middle atmosphere to changing O3 and CO2 - A speculative tutorial. Transport Processes in the Middle Atmosphere (G. Visconti ed.), D. Riedel, pp. 371-386.

Fetzer, E.J. and J.C. Gille, 1994: Gravity wave variances in LIMS temperatures. part I: variability and comparison with background winds. Journal of the Atmospheric Sciences, 51, 2461-2483.

Fovell, R., D.R. Durran and J.R. Holton, 1992: Numerical simulations of convectively generated stratospheric gravity waves. Journal of the Atmospheric Sciences, 49, 1427-1442.

Fritts, D.C., and W. Lu, 1993: Spectral estimates of gravity wave energy and momentum fluxes. Part II: Parameterization of wave forcing and variability. Journal of the Atmospheric Sciences, 50, 3695-3713.

Fritts, D.C. and G.D. Nastrom, 1992: Sources of mesoscale variability of gravity waves, Part II: Frontal, convective and jet stream excitation. Journal of the Atmospheric Sciences, 49, 111-127.

Gaines, S. E., S. W. Bowen. R. S. Hipskind, T. P. Bui, and K.R. Chan, 1992: Comparison of the NASA ER-2 Meteorological Measurements System with radar tracking and radiosonde data, Journal of Atmospheric and Oceanic Technology. 9, 210-225.

Garcia, R.R., and B. Boville, 1994: ``Downward control" of the mean meridional circulation and temperature distribution of the polar winter stratosphere. Journal of the Atmospheric Sciences, 51, 2238-2245.

Gardner, 1991: Introduction to ALOHA-90: The airborne lidar and observation of the Hawaiian airglow campaign. Geophysical Research Letters., 18, 1313-1316.

Gardner, C. S., M. Coble, G. C. Papen, and G. R. Swenson, 1996: Observations of the unambiguous two-dimensional horizontal wave number spectrum of OH intensity perturbations, Geophysical Research Letters., 23, 3739-3742.

Gary, B., 1989: Observational results using the Microwave Temperature Profiler during the Airborne Antarctic Ozone Experiment, Journal of Geophysical Research, 94, 11223-11232.

Guest, F.M., M.J. Reeder, C.J. Marks and D.J. Karoly, 1997: Analyses of stratospheric gravity waves over Macquarie Island. Proceedings of the First SPARC General Assembly, in press.

Hamilton, K., 1996: Comprehensive meteorological modelling of the middle atmosphere: a tutorial review. Journal of Atmospheric and Terrestrial Physics, 58, 1591-1627.

Hamilton, K. and R.A. Vincent, 1995: High-resolution radiosondes offer new prospects for research. Eos, 74, 497-507.

Hauchecorne A., and M. L. Chanin, 1980: Density and temperature profiles obtained by lidar between 35 and 70 km. Geophysical Research Letters, 7, 565-568.

Haynes, P. H, C.J. Marks, M.E. McIntyre, T.G. Shepherd and K.P. Shine, 1991: On the "downward control" of extratropical diabatic circulations by eddy-induced mean zonal forces. Journal of the Atmospheric Sciences, 48, 651-678.

Hecht, J. H., S. K. R. Hawat, R. L. Walterscheid, and J. R. Isler, 1995: Observations of variations in airglow emissions during ALOHA-93, Geophysical Research Letters, 22, 2817-2820.

Heymsfield, G. M., I. J. Caylor, J. M. Shepher, W. S. Olson, S. W. Bidwell, W. C. Boncyk, and S. Ameen, 1996: Structure of Florida thunderstorms using high-altitude aircraft radiometer and radar observations, Journal of Applied Meteorology, 35, 1736-1762.

Holton, J.R., and D.R. Durran, 1993: Convectively generated stratospheric gravity waves: the role of mean wind shear. Coupling Processes in the Lower and Middle Atmosphere (Thrane, Blix and Fritts, eds.), Kluwer Academic, 175-189.

Hostetler, C. A., and C. S. Gardner, 1994: Observations of horizontal and vertical wave number spectra of gravity wave motions in the stratosphere and mesosphere over the mid-Pacific. Journal of Geophysical Research, 99, 1283-1302.

Jensen, E.J., O.B. Toon, H.B. Selkirk, J.D. Spinhirne and M.R. Schoeberl, 1996: On the formation and persistence of subvisible cirrus clouds near the tropical tropopause. Journal of Geophysical Research, 101, 21361-21375.

Keenan, T., and R. Carbone, 1992: A preliminary morphology of precipitation systems in tropical Northern Australia. Quarterly Journal of the Royal Meteorological Society, 118, 283-336.

Keenan, T. and M. Manton, 1996: Darwin Climate Monitoring and Research Station: Observing precipitating systems in a Monsoon Environment, BMRC Research Report, 53, 31pp.

Keenan, T., K. Glasson, F. Cummings, T. Bird, J. Keeler and J. Lutz, , 1998: The BMRC/NCAR C-Band Polarimetric (C-POL) Radar System, Journal of Atmospheric and Oceanic Technology, in press..

Keenan, T.D., B.R. Morton, Y.S. Zhang, and K. Nyguen, 1990: Some characteristics of thunderstorms over Bathurst and Melville Islands near Darwin, Australia, Quarterly Journal of the Royal Meteorological Society, 116, 1153-1172.

Kelly, K. K., et al, 1989: Dehydration in the lower antarctic stratosphere during late winter and early spring, 1987, Journal of Geophysical Research, 94, 317-357.

Kershaw, R., 1995: Parameterization of momentum transport by convectively generated gravity waves. Quarterly Journal of the Royal Meteorological Society, 121, 1023-1040.

Klemp, J.B., and R.B. Wilhelmson, 1978: The simulation of three-dimensional convective storm dynamics. Journal of the Atmospheric Sciences, 35, 1070-1096.

Knudsen, B.M., 1996: Accuracy of Arctic stratospheric temperature analyses and the implications for the prediction of polar stratospheric clouds, Geophysical Research Letters, 23, 3747-3750.

Lafore, J.-P. and M.W. Moncreiff, 1989: A numerical investigation of the organization and interaction of the convective and stratiform regions of tropical squall lines. Journal of the Atmospheric Sciences, 46, 521-544.

Lindzen, R. S., 1981: Turbulence and stress owing to gravity wave and tidal breakdown. Journal of Geophysical Research, 86, 9707-9714.

Liou, K.N., S. C. Ou, Y, Takano, F. P. J. Valero, and T. P. Ackerman, 1990: Remote sounding of the tropical cirrus cloud temperature and optical depth using 6.5 and 10.5 micron radiometers during STEP, Journal of Applied Meteorology, 29, 716-726.

McFarlane, N., C. McLandress and S. Beagley, 1997: Seasonal simulations with the Canadian middle atmosphere model: sensitivity to a combination of orographic and Doppler spread parameterizations of gravity wave drag. Gravity Wave Processes - Their Parameterization in Global Climate Models (K. Hamilton, ed.) Springer-Verlag, pp. 351-366.

Manzini, E., N.A. McFarlane and C. McLandress, 1997: Middle atmosphere simulations with the ECHAM4 model: sensitivity to Doppler-spread gravity wave parameterization. Gravity Wave Processes - Their Parameterization in Global Climate Models (K. Hamilton, ed.) Springer-Verlag, pp. 367-381.

Moncrieff, M.W. 1992: Organized convective systems: Archetypal dynamical models, mass and momentum flux theory, and parameterization, Quarterly Journal of the Royal Meteorological Society, 118, 819-850.

Murphy, D.M and B.L. Gary, 1995: Mesoscale temperature fluctuations and polar stratospheric clouds, Journal of the Atmospheric Sciences, 52, 1753-1760.

Pfister, L., K.R. Chan, T.P. Bui, S. Bowen, M. Legg, B. Gary, K. Kelly, M. Proffitt and W. Starr 1993: Gravity waves generated by a tropical cyclone during the STEP tropical field program. Journal of Geophysical Research, 98, 8611-8638.

Proffitt, M. H. and R. J. McLaughlin, 1983: Fast-response dual-beam UV-absorption ozone photometer suitable for use on stratospheric balloons, Reviews of. Scientific Instruments, 43, 1719-1728..

Redelsberger, J.-L. and J.-P. Lafore, 1988: A three-dimensional simulation of a tropical squall line: Convective organization and thermodynamic vertical transport. Journal of the Atmospheric Sciences, 45, 1334-1356.

Rind, D., R. Suozzo, N.K. Balachandran, A. Lacis, and G. Russell, 1988a: The GISS global climate-middle atmosphere model. Part I: Model structure and climatology. Journal of the Atmospheric Sciences, 45, 329-370.

Rind, D., R. Suozzo, and N.K. Balachandran, 1988b: The GISS global climate-middle atmosphere model. Part II: Model variability due to interactions between planetary waves, the mean circulation and gravity wave drag. Journal of the Atmospheric Sciences, 45, 371-386.

Rosenlof, K., 1996, Summer hemisphere differences in temperature and transport in the lower stratosphere, Journal of Geophysical Research, 101, 19129-19136.

Sato, K., 1997: Observational studies of gravity waves associated with convection. Gravity Wave Processes - Their Parameterization in Global Climate Models (K. Hamilton, ed.) Springer-Verlag, pp. 63-68.

Sato, K. and T.J. Dunkerton, 1996: Estimates of the momentum flux associated with equatorial Kelvin and gravity waves. Journal of Geophysical Research, in press.

Sato, K., D. O'Sullivan and T.J. Dunkerton, 1997: Low-frequency inertia-gravity waves in the stratosphere revealed by three-week continuous observation with the MU radar, Geophysical Research Letters, in press.

Scott, S.G., T. P. Bui, K. R. Chan, and S. W. Vowen, 1990: The Meteorological Measurement System on the NASA ER-2 aircraft, Journal of Atmospheric and Oceanic Technology. 7, 520-525.

Simpson, J., Adler R.F and North G.R., 1988: A proposed Tropical Rainfall Measuring Mission (TRMM) Satellite, Bulletin of the American Meteorological Society, 69, 278-295.

Soong, S.-T. and Tao, W.-K., 1984: A numerical study of the vertical transport of momentum in a tropical rainband. Journal of the Atmospheric Sciences, 41, 1049-1061.

Swenson, G. R., and S. B. Mende, 1994: OH emission and gravity waves (including abreaking wave) in all sky imagery from Bear Lake, UT., Geophysical Research Letters, 21, 2239-2242.

Swenson, G. R, M.J. Taylor, P.J. Epsy, C. Gardner and X. Tao, 1995: ALOHA-93 measurements of intrinsic AGW characteristics using airborne airglow imager and groundbased Na wind/temperature lidar. Geophysical Research Letters, 22, 2841-2844.

Swenson, G. R., M. J. Taylor, P. J. Espy, C. S. Gardner, and X. Tao, 1995: ALOHA-93 measurements of intrinsic AGW characteristics using airborne airglow imager and groundbased Na wind/temperature lidar, Geophysical Research Letters, 22, 2841-2845.

Tabazadeh, A., O.B. Toon, B.L. Gary, J.T. Bacmeister and M.R. Schoeberl, 1996: Observational constraints on the formation of Type Ia polar stratospheric clouds, Geophysical Research Letters, 23, 2109-2112.

Tao, W.-K., and J. Simpson, 1989: Modeling study of a tropical squall-type convective line. Journal of the Atmospheric Sciences, 46, 177-202.

Taylor, M. J., M. A. Hapgood, and P. Rothwell, 1987: Observations of gravity wave propagation in the OI(557.7 nm), Na(589.2 nm) and the near infrared OH nightglow emissions, Planetary and Space Science, 35, 413-423.

Taylor, M. J., and M. A. Hapgood, 1988: Identification of a thunderstorm as a source of short period gravity waves in the upper atmospheric nightglow emission, Planetary and Space Science, 36, 975-985.

Taylor, J. J., Y. Y. Gu, X. Tao, C. S. Gardner, and M. B. Bishop, 1995: An investigation of intrinsic gravity wave signatures using coordinated lidar and nightglow image measurements, Geophysical Research Letters, 22, 2853-2856.

Tepley, C. A., S. I. Sargoytchev, and C. O Hines, 1991: Initial Doppler Rayleigh lidar results from Arecibo. Geophysical Research Letters, 18, 167-170.

Trier, S.B., W.C. Skamarock, M.A. LeMone, D.B. Parsons and D.P. Jorgensen,1996: Structure and evolution of the 22 February 1993 TOGA COARE squall line: numerical simulations. Journal of the Atmospheric Sciences, 53, 2861-2886.

Tsuda, T., Y. Maruyama, H. Wiryosumarto, S.W.B., Harijono and S. Kato, 1994: Radiosonde observations of equatorial atmospheric dynamics over Indonesia, I, equatorial waves and tides. Journal of Geophysical Research, 99, 10491-10505.

Vincent, R.A. and D. Lesicar, 1991: Dynamics of the equatorial middle atmosphere: first results with a new generation partial reflection radar. Geophysical Research Letters, 18, 825-828.

Vincent, R.A., S.J. Allen and S.D. Eckermann, 1997: Gravity wave parameters in the lower stratosphere. Gravity Wave Processes - Their Parameterization in Global Climate Models (K. Hamilton, ed.) Springer-Verlag, pp. 7-25.

Whiteway, J. A. and T. J. Duck, 1997: Evidence for critical level filtering of atmospehric gravity waves. Geophysical Research Letters, 24, 145-148.

Whiteway, J. A., T. J. Duck, D. P. Donovan, J. C. Bird, S. R. Pal, A. I. Carswell, 1997: Measurements of gravity wave activity within and around the Arctic stratospheric vortex. Geophysical Research Letters, 24, 1387-1390.

Wilson, R., M. L. Chanin, and A. Hauchecorne, 1991: Gravity waves in the middle atmosphere observed by Rayleigh lidar, 2, Climatology. Journal of Geophysical Research, 96, 5169-5183.

Wu, D.L. and J.W. Waters, 1997: Observations of gravity waves with the UARS microwave limb sounder. Gravity Wave Processes - Their Parameterization in Global Climate Models (K. Hamilton, ed.) Springer-Verlag, pp. 103-120.

Yang, M.-J., and R.A. Houze, 1995: Multicell squall-line structure as a manifestation of vertically trapped gravity waves. Monthly Weather Review, 123, 641-661.

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

`Implementation Plan`

Appendix III

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

Implementation Plan

Appendix III

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

`Implementation Plan`