Atlantic Stratocumulus Transition Experiment

What is ASTEX?

The Atlantic Stratocumulus Transition Experiment (ASTEX) was conducted in June 1992 off North Africa (rectangles in Fig. 1), in the area of Azores and Madeira Islands ( Fig. 4). ASTEX was based on two islands and several ships in an area where the total cloud cover (mostly stratocumulus) ranges from 50 - 60% ( Fig. 2). The region is dominated by low-level clouds with moderate optical thicknesses, from about 1 to 10 on average ( Fig. 3). The optically thinner (more highly broken) clouds generally have cloud tops below the 800 mb level). The optically thicker clouds have lower top pressures down to about 700 mb. The region is characterized by broken low cloudiness and strong gradients of low level cloud amount. Satellite studies show cloud conditions ranging from solid stratocumulus decks to broken trade cumulus. The region is not directly influenced by continental effects, and islands provide suitable sites for surface observations and aircraft operations. ASTEX was thus able to address issues related to the stratocumulus to trade- cumulus transition and cloud-mode selection.

ASTEX involved intensive measurements from several platforms and was designed to study how the transition and mode selection are affected by 1) cloud-top entrainment instability, 2) diurnal decoupling and clearing due to solar absorption, 3) patchy drizzle and a transition to horizontally inhomogeneous clouds through decoupling, 4) mesoscale variability in cloud thickness and associated mesoscale circulations, and 5) episodic strong subsidence lowering the inversion below the lifting condensation level. From a broader perspective ASTEX was designed to provide improved dynamical, radiative, and microphysical models and an improved understanding of the impact of aerosols, cloud microphysics, and chemistry on large-scale cloud properties.

From a broader perspective ASTEX was designed to provide improved dynamical, radiative, and microphysical models and an improved understanding of the impact of aerosols, cloud microphysics, and chemistry on large-scale cloud properties.

A telescoping approach was used in ASTEX to investigate connections between scales ranging from microns to thousands of kilometers. Satellites and upper- level aircraft provided a description of large-scale cloud features, and instrumented aircraft flying in the boundary layer and surface- based remote sensing systems provided a description of the mean, turbulence, and mesoscale variability in cloud microphysical properties of boundary layer clouds. A major deficiency of the FIRE observations, however, was an inadequate definition of the large-scale fields of temperature, moisture, and winds. This deficiency was removed for ASTEX by making 4 - 8 soundings per day from the surface sites and ships, and including many of these upper-air observations on the Global Telecommunications System (GTS) for assimilation into the ECMWF and NMC analyses. Furthermore, based on the demonstrated utility of surface-based remote sensing during FIRE (Albrecht et al. 1990), the use of such systems was expanded during ASTEX.

An overview of ASTEX is given by Albrecht et al. (1995). Highlights of the ASTEX field deployment include the following:

The boundary layers observed in ASTEX were extremely deep (1000-2000m) and the most common cloud type observed was cumulus clouds under a stratocumulus layer of variable thickness and extent. On nearly all occasions the thermodynamic structure of the boundary layer was very complex. Only in a few cases was the boundary layer well mixed. More typically, as shown in Fig. 5, the boundary layer was multi-layered, with a subcloud mixed layer (SML) decoupled (both day and night) from a subcloud layer and the cloud layer itself.

Several different boundary-layer cloud types were observed near the Azores, including extensive sheets of stratocumulus, clear skies and fields of small cumulus clouds. In nearly all cases, however, the thermodynamic structure of the boundary layer was similar to that show in Fig. 5, except drier under the subsidence inversion when there was no stratocumulus present. The cumulus clouds always had their bases very close to the top of the SML and were produced when the SML became conditionally unstable as the result of the build up of moisture in the layer. Although none of the individual experiments during ASTEX, including the Lagrangian experiments, observed the transition from persistent stratocumulus to trade wind cumulus, the gross features of the evolution can be pieced together from the hundreds of profiles that were made through the boundary layer at different locations at different times. In general in the northern part of the area where the sea surface temperatures were at their lowest, the boundary layer was shallowest and for at least part of the day it was well mixed. Diurnal variation here keeps the SML decoupled from the layer cloud for most of the day. However as the air moves over warmer sea surface temperatures (SSTs) the boundary layer is observed to deepen. Here the TKE production is not great enough to mix the whole of the boundary layer either during the day or at night and as a result the stratocumulus starts to thin and break up. At the same time though the SML becomes conditionally unstable and the resulting cumulus cloud locally re-couple the SML and the cloud layer. This can either help maintain the stratocumulus by supplying it with moisture from the SML or quicken its dissipation by enhancing the entrainment of free tropospheric air into the cloud layer. Over the warmest SST's the latter dominates leaving only trade wind cumulus in the boundary layer. Obviously there will be other processes which will have a modifying effect on this evolution. These include temporal variation of the subsidence in the free troposphere, cloud top entrainment instability, drizzle, boundary layer wind speeds and aerosol characteristics of the air mass. However, it is clear from the observations to simulate the break up of marine stratocumulus in the sub tropics the permanent decoupling of the boundary layer has to be well represented.

It was very commonly observed during ASTEX that cumulus clouds merged into a stratocumulus cloud base; these observations have mad it possible to study the interactios of cumulus clouds with stratocumulus layers. The dynamical processes associated with cumulus clouds and stratocumulus are quite different and thus they experience different entrainment and mixing processes.

The cumulus clouds feed on the moisture supply in the SML and locally re-couple the SML to the cloud layer. This has been observed to significantly thicken the stratocumulus layer and help maintain the layer for much longer than would be expected if it were permanently decoupled from the sea surface. It also modifies the thermodynamic structure of the boundary layer.

Perhaps the most significant effect of the different dynamical processes associated with cumulus clouds and stratocumulus is on the shape of the droplet size spectra in the different cloud types (Martin et al., 1994). In the relatively narrow and feeble cumulus clouds, mixing from the environment penetrates to the cloud's core and significantly modifies the droplet size spectra. Thus, the typical shapes of the cumulus and stratocumulus droplet size spectra are going to be different. When the cumulus cloud penetrates the base of the stratocumulus the two droplet spectra will start to mix and interact. The effects on the stratocumulus are dependent on the initial thickness of the stratocumulus. In general, the liquid water content is increased as the cumulus clouds are much deeper than the stratocumulus. Also the droplet concentration is increased as initially the vertical velocity at the base of the cumulus cloud is higher than at the base of the stratocumulus; therefore more CCN are activated into cloud drops because of the slightly higher maximum supersaturation. If the stratocumulus is initially very thin then the mean droplet size will only be very small therefore the penetration of the cumulus cloud has the potential to increase the mean droplet size in the stratocumulus. However, if the layer is initially thick with relatively large droplets the interaction with cumulus clouds will most likely result in a decrease in the mean droplet size. From the perspective of the radiative transfer characteristics of the stratocumulus layer the change in the liquid water path dominates the interaction, with the resultant effect that the stratocumulus layer's reflectivity is increased.

During ASTEX the cumulus clouds were observed to form in clusters which on occasions had lifetimes of several hours, and it was frequently noted that drizzle accompanied the penetration of these clusters into the stratocumulus base. The cumulus clouds on their own were not deep enough (the cloud top temperature was always above freezing) to produce precipitation, and generally the stratocumulus was too thin to produce drizzle. Thus the interaction between the cumulus clouds and the stratocumulus was in some way initiating the drizzle formation.

The reflectivity of stratocumulus clouds is very susceptible to changes in cloud condensation nuclei concentration (Platnick and Twomey, 1994; Taylor and McHaffie, 1994) due to their effect on cloud droplet size. Slingo (1989) showed that the sensitivity of the shortwave radiative properties of clouds to changes in droplet size will be greatest for clouds with liquid water paths between 10 and 100 g m-2. Other experiments (Fouquart et al., 1990) have demonstrated that many stratocumulus sheets lie in this range. Stratocumulus clouds therefore have a great potential for changing the global energy budget and thus the climate through the indirect effects of aerosols (Twomey, 1977; Charlson et al., 1987). During ASTEX, a variety of airmass types were encountered, making possible several studies of the effects of aerosol on stratocumulus clouds.

ASTEX had a surprising mixture of very clean airmasses and highly polluted ones which had come from the European and African landmasses. Considering the distance of the Azores from the continents it was quite remarkable how the continental airmasses maintained their characteristics over such a long sea track. On a couple of occasions during June 1992, outbreaks of continental air swept out towards the Azores in strong north easterly winds (Johnson et al., 1993). These were associated with very sharp boundaries between the clean maritime airmass and the continental airmass, both in the boundary layer and in the free troposphere. However back trajectories showed that the air in the boundary layer had originated over the industrial areas of northern Europe while the air above the subsidence inversion had been advected from over the Sahara. Thus the aerosol characteristics had significant variations in the vertical. Both airmasses had well defined stratocumulus sheets but even though they had very similar liquid water profiles, the cloud droplet concentrations and sizes within the layers were completely different. The aerosol concentrations (in the size range 0.1 - 3.0 mm diameter) slightly below cloud base varied between 60 cm-3 in the maritime airmass, and 1500 cm-3 in the continental airmass. This resulted in cloud droplet concentrations of 50 cm-3and a maximum effective radius (near cloud top) of 13 mm in the maritime airmass, and 270 cm-3 and 8 mm in the continental airmass. Because of this the reflectivity of the stratocumulus in the continental airmass was significantly higher than that in the maritime airmass.

The stratus observed during ASTEX was generally associated with decoupled boundary layers. This is clearly illustrated in Fig. 6 where composite temperature and moisture profiles from radiosondes collected in nearly solid status during FIRE and more broken conditions during ASTEX are presented (Jensen, 1993). The moisture structure of the two ASTEX composite soundings from the island of Santa Maria (37N, 25W) and the German ship Vldivi (28N, 24W) clearly show a well defined subcloud layer structure with a decrease in moisture at the base of the cloud layer compared with the FIRE composite sounding from San Nicolas Island (33N, 120W). In addition to the FIRE and ASTEX composite soundings, a sounding obtained at the equator over the central Pacific (0S, 140W) is included in Fig. 6. The soundings for this composite were obtained during the Tropical Instability Wave Experiment (TIWE, Charlock and Fairall, 1993) during December of 1991 where a classic trade-wind boundary layer structure was observed for a three-week period. The tropical composite sounding shows a deeper boundary layer an even more pronounced transition layer separating the clouds and the subcloud layer.

The cloud cover corresponding to each of the composite soundings shown in Fig. 6 was estimated using a laser ceilometer operating at each location. The cloud cover was estimated by classifying each 30-second observation as either clear if no clouds were detected or cloudy if clouds with bases less that 3 km are observed. The cloudiness for each hour was then calculated using these 30-second classifications. The cloudiness during ASTEX is about 67% at Santa Maria and 40% at the Vldivi compared with 82% at San Nicolas and 26% during TIWE. Thus ASTEX is clearly intermediate between the solid stratocumulus clouds and the broken fair-weather cumulus of the undisturbed trades.

The decoupled conditions observed during ASTEX result in a moistening of the subcloud layer relative to the cloud layer. Although the stable layer often observed at cloud base limits the turbulent exchange between the cloud and the subcloud layer, the moistening of the subcloud layer increases convective available potential energy. Thus in areas where an updraft in the subcloud layer reaches the lifting condensation level and penetrates the weak inversion at cloud base, there is the potential for the development of relatively vigorous cumulus clouds. These penetrating cumulus were often observed during ASTEX to help supply liquid water to the overlying stratus through detrainment at the base of the inversion that caps the cloud layer. These detrained cloud masses often have the appearance of the anvils associated with thunderstorms. Substantial drizzle is often associated with these marine boundary layer convective complexes (MBLCCs).

The structure of the MBLCCs was documented by the cloud radars located on Santa Maria and Porto Santo. Aircraft observations were also made in and around these systems. The NOAA Wave Propagation Laboratory tracked several of these systems using their 35 GHz cloud radar and found that they persisted for a number of hours (Kropfli et al., 1992). A cloud radar operated from Santa Maria by Penn State University (Peters et al., 1993) probed several of these systems as they passed over the island. The radar returns shown in Fig. 7 clearly show the anvil-like structure of the detrained stratus, the over- shooting cloud top, and possible entrainment along the edges of the over-shooting cloud top. It is possible that these MBLCCs are related to closed mesoscale cellular convection.

These preliminary results indicate that the transition is not a simple and rapid transition from solid stratus to broken fair weather cumulus. Instead, the transition is from solid stratus associated with well-mixed conditions to stratus that can be generated by long-lived, intermittent strong convective systems feeding on moist air near the surface in decoupled boundary layers.

Extreme variations in aerosol conditions were observed during ASTEX. During the second week of the experimental period very clean air was present in the study region. The chemists on the Electra noted that the air was as clean or cleaner than air they have sampled over the central Pacific. During this period drizzle was observed frequently from the aircraft and the islands -- often in association with the MBLCC's discussed previously. This clean air was replaced by a cloud mass moving westward from the European continent. This air mass provided a very sharp boundary between the clean and dirty air. This boundary was thoroughly sampled with the aircraft. The contrast in cloud structure was striking as illustrated in Fig. 8. The continental air was characterized by substantially higher droplet concentrations and larger droplets than observed in the clean air mass. Drizzle was generally suppressed in the continental air mass. These conditions provided an exceptional data set for characterizing cloud characteristics associated with substantial differences in aerosol concentrations.

Two Lagrangian experiments were performed during the two- week period in the middle of the experiment. During these experiments an attempt was made to follow a tagged air mass for two complete days. The first Lagrangian was conducted in the clean air mass discussed previously and the second was in the dirty air mass. During the first Lagrangian the six constant- level balloons that we used to tag the air mass ended up in the ocean after a few hours due to loading by the drizzle. Measurements were made following a trajectory based on real time winds from the aircraft. This trajectory was in reasonable agreement with a trajectory from calculated using ECMWF analyses. During the second Lagrangian (when there was relatively less drizzle) two balloons were tracked for nearly 48 hours. These experiments allow budgets to be made without evaluating advective effects and provide a unique data set for testing one-dimensional models. Bretherton and Pincus (1995) and Bretherton, Austin, and Siems (1995) described the synoptic setting of the ASTEX Lagrangian Experiments, and the observed cloudiness, surface fluxes, drizzle, and entrainment rate.

The ASTEX data were collected well away from the direct effects of the European and African continents. Island effects can also disrupt the diurnal variations relative to open ocean conditions, however, so a careful comparison of the diurnal variations on the islands and those observed from the ships over the open ocean will be made. The ASTEX Lagrangian experiments provide an excellent opportunity to study diurnal effects, since aircraft measurements were made through the night during these experiments. Furthermore, the effects of advection can be removed from these measurements.

Eight radiosondes per day were launched from Santa Maria, Porto Santo, the R/V Vldivi from Germany during the first three weeks of the experiment and NOAA's R/V Malcolm Baldrige during the last week. Standard and significant level data for most of these soundings were transmitted to Santa Maria where they were transmitted to the Global Telecommunication System (GTS) by technicians from Lisbon's INMG (the Portuguese NWS). These data were then assimilated into ECMWF and other global analyses. Approximately 650 of the 820 soundings were placed on the GTS and about 90% of these were assimilated into the ECMWF analysis. This was clearly a difficult but major accomplishment. Assessments are in progress to determine how well the ECMWF analyses represent the boundary layer structure and other fields during ASTEX. The ECMWF analyses will be used to define large- scale divergence and other parameters needed to test regional and large-scale models.

An extensive deployment of remote sensors was made for ASTEX to study the cloudy marine boundary layer. These sensors included two cloud radars, two wind profilers, a RASS, five microwave radiometers, four ceilometers, and a several upward- looking radiometers (Cox et al., 1993a,b). This instrumentation provided data for characterizing clouds and the environment in which they form.

ASTEX marked the first deployment of cloud radars in a marine environment. They provided estimates of cloud-top height, reflectivity profiles, in-cloud turbulence, and drizzle characteristics. The scanning radar on Porto Santo was used to provide a horizontal mapping of the clouds to track cloud features of interest. The cloud base height from ceilometers are being combined with the radar cloud top to define cloud thickness. Simultaneous measurements of cloud liquid water path were obtained with microwave radiometers. At the same time, microwave radiometers provide integrated liquid water content. These measurements are being used to define the ratio of the observed liquid water path to the adiabatic liquid water path calculated from the cloud thickness following the technique described by (Albrecht et al., 1990). This instrumentation is being used to investigate the structure of the marine boundary layer convective complexes described previously.


Albrecht, B. A., C. W. Fairall, D. W. Thomson, A. B. White, J. B. Snider and W. H. Schubert, 1990: Surface-based remote sensing of the observed and the adiabatic liquid water content of stratocumulus clouds. Geophys. Res. Lett., 17, 89-92.

Albrecht, B. A., C.S.Bretherton, D.W.Johnson, W.H.Schubert and A.Shelby Frisch, 1995: The Atlantic Stratocumulus Transition Experiment -- ASTEX. Bull. Amer. Metor. Soc., 76, 889-904.

Bretherton, C. S., and R. Pincus, 1995: Cloudiness and marine boundary layer dynamics in the ASTEX Lagrangian Experiments. Part I: Synoptic setting and vertical structure. J. Atmos. Sci., in press.

Bretherton, C. S., P. Austin, and S. T. Siems, 1995: Cloudiness and marine boundary layer dynamics in the ASTEX Lagrangian Experiments. Part II: Cloudiness, drizzle, surface fluxes, and entrainment. J. Atmos. Sci., in press.

Charlock, B. and C. W. Fairall, 1993: Surface based measurements and satellite retrievals of broken cloud properties in the equatorial Pacific. J. of Climate. (submitted)

Charlson, R. J., J. E. Lovelock, M. O. Andreae, and S. G. Warren, 1987: Oceanic phytoplankton, atmospheric sulfur, cloud albedo and climate. Nature, 326, 655-661.

Cox, S. K., C. Cornwall, W. Cotton, J. Davis, J. Kleist, T. McKee, Q. Shao, D. Randall, W. Schubert, D. Wood, S. Frisch, M. Hardesty, R. Kropfli, J. Snider and P. Anikin, 1993a: CSU/NOAA/WPL FIRE II - ASTEX field experiment description of field deployment phase. Atmospheric Science Paper No. 523, Colorado State University, Fort Collins, CO 80523, 100 pp.

Cox, S. K., S. Gillies, A. Heidinger, and C. Keith, 1993b: CSU ASTEX surface data sets from Porto Santo (June 1-28, 1992). FIRE Series No. 9, Atmospheric Science Paper No. 530, Colorado State University, Fort Collins, CO 80523, 146 pp.

Fouquart, Y., J. C. Buriez, M. Herman and R. S. Kandel 1990: The influence of clouds on radiation: a climate modelling perspective. Rev. Geophys., 28, 145-166.

Jensen, M. P., 1993: Thermodynamic and cloud characteristics of the undisturbed boundary layers. Penn State University, M.S. Thesis, 81 pp.

Johnson, D. W., G. M. Martin, J. P. Taylor, C. A. Friehe and D. P. Rogers, 1993: The effect of an outbreak of continental air from Europe on marine stratocumulus clouds near the Azores. Proceedings of the 1993 Spring Meeting of the American Geophysical Union.

Kropfli, R. A., S. Y. Matrosov, T. Uttal, A. S. Frisch, B. W. Orr, J. B. Snider, 1992: The use of radar on cloud/radiation study: Initial results from CLARET 1, FIRE 2 and ASTEX.

Martin, G. M., D. W.Johnson and A. Spice, 1994: The measurement and parameterization of effective radius of droplets in warm stratocumulus clouds. J. Atmos. Sci., xxx.

Platnick, S. and S. Twomey, 1994: Determining the susceptibility of cloud albedo to changes in droplet concentration with the Advanced Very High Resolution Radiometer. J. Appl. Meteor., 33, 334-347.

Slingo, A., 1989: A GCM parametrization for the shortwave radiative properties of water clouds. J. Atmos. Sci., 46, 1419-1427.

Taylor, J. P. and A. McHaffie 1994: Measurements of cloud susceptibility. J. Atmos. Sci. In press, April 1994.

Twomey, S., 1977: The influence of pollution on the shortwave albedo of clouds. J. Atmos. Sci., 34, 1149-1152.