Click for next page ( 5


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 4
2 Scientific Objectives The problem of the interaction between synoptic- and convective-scale mo- tions may be viewed in two different ways. One might seek a complete physical understanding of the mechanisms by which the scale interaction takes place. This is the approach of the physical scientist or the numerical modeler who is interested in convection for its own sake. On the other hand, one might be satisfied with parameterizing* the collective effects of convective-scale motions upon the synoptic-scale disturbances in terms of the synoptic-scale fields, without understanding the detailed structure of the convection. This attitude is typical among numerical modelers who are primarily concerned with predicting the evolution of the synoptic-scale fields and is the viewpoint prevalent in the GARP planning documents. The first approach requires de- tailed knowledge of the internal structure of convection and how it responds to changes in the synoptic-scale environment, while the second approach is mainly concerned with relating the bulk properties of convective ensembles to the synoptic-scale fields. Both approaches are consistent with the overall objectives of GARP, yet each suffers from inherent shortcomings. The mere task of documenting the various types of mesoscale organization that convection is capable of assuming could easily absorb all the efforts of many GARP experiments. The prospect of fully understanding the dynamical mechanisms responsible for each type of organization and how they depend upon the synoptic-scale environment is *"Parameterization" is a simplification introduced into a dynamical model by pre- assigning the magnitude of a physical effect rather than allowing the effect to be realistically determined internally as a consequence of the dynamics of the system.

OCR for page 4
indeed remote. Only through recourse to the bulk properties of convective ensembles can one avoid becoming hopelessly mired in the complexities of convective-scale phenomena. On the other hand, complete reliance on param- eterization techniques is bound to lead to unfounded empiricism and ulti- mately to stagnation. It would appear that the best way of avoiding the pitfalls of the two approaches is not to rely on either one of them entirely. Here, as in the field of atmospheric turbulence, a coordinated effort to un- derstand both the internal structure and the bulk properties represents the path to progress. Thus, an experiment is envisioned that will provide simultaneous measure- ments of: (a) the synoptic-scale fields, (b) the bulk properties of convective ensembles, and (c) the internal structure of convective ensembles. The scientific objectives of the experiment can be expressed in terms of the interrelations between these three sets of measurements: (a-b) to relate the internal structure of convective ensembles to the synoptic-scale environment (see also Figure 1), (b-c) to explain the bulk properties of convective ensembles in terms of their internal structure, and (c-a) to use the bulk properties of convective ensembles to parameterize the effects of convection on the synoptic-scale fields in terms of the synoptic- scale variables. These objectives will not be realized until the results of the experiment are fully analyzed. The remainder of this Chapter is divided into two parts. First, Section 2.1 briefly reviews the present state of knowledge of (a), (b), and (c) in Figure 1 as it relates to the tropical Atlantic. Section 2.2 summarizes what is known concerning the interrelationships between (a-b), (b-c), and (c-a). Wherever possible, questions are posed that should be answered by the experiments. 2.1 Properties of the Observed F ields (a) SYNOPTIC-SCALE SYSTEMS IN THE ATLANTIC It is generally agreed that Atlantic systems are somewhat weaker than their counterparts in the tropical northwest Pacific; wind fluctuations are generally smaller, and rainfall amounts are typically about half of those recorded along the Pacific

OCR for page 4
SYNOPTIC SCALE DISTURBANCES understanding parameterization \ SCALE SYSTEMS r / , CONVECTIVE _\ j i b h—1 i Internal i 1 . Bulk I I 1 Structure | 1 Properties . 1 J 1 1 FIGURE 1 Schematic relationship between the synoptic-scale field and the bulk and internal characteristics of the convective-scale field. ITCZ. Nevertheless, there is evidence of definite synoptic-scale organization in the Atlantic systems and considerable day-to-day continuity. In fact, it appears that many of the Pacific systems can be traced backward across the Atlantic to an origin over Africa. Synoptic systems of the tropical North Atlantic summer season have been classed in four categories: 1. The Intertropical Convergence Zone (ITCZ), which marks the boundary between the northeast trades and low-level, cross-equatorial flow. It is well marked in the eastern Atlantic, where it lies close to ION. It is usually con- centrated in a narrow band 100 to 300 km in width, comprising one or more lines of active mesoscale convection. It undergoes frequent distortion due to the passage of wave disturbances. 2. ITCZ disturbances, which enter the Atlantic at about 10-12N and track

OCR for page 4
west-southwestward to about 8N, 45W, at a rate of about six degrees of longi- tude per day. Near this point about two thirds of these systems lose their identity, while the remaining third reintensify and move northwestward through the Antilles, where some of them reach tropical storm intensity. The cloud patterns associated with the ITCZ disturbances are often vortical in shape and show evidence of deep convection. 3. Tradewind disturbances, which cross the Atlantic at latitudes near 15N, at a rate of about six degrees of longitude per day. These are marked by the characteristic "inverted V" cloud patterns which are believed to consist largely of low clouds. These systems exhibit maximum amplitude in the middle troposphere (about 600 mbar), and have relatively little surface weather asso- ciated with them. 4. Upper-tropospheric vortices, which form in association with the tropi- cal upper-tropospheric trough. (The latter is a recurrent feature of the summer circulation, extending from the Azores southwestward to the Caribbean.) The more intense vortices extend through much of the depth of the troposphere and exhibit characteristic "comma-shaped" cloud patterns. When the westward moving ITCZ and tradewind disturbances come under the influence of the upper-tropospheric trough, in the region of the Antilles, a sudden flareup of the convective activity is often observed. At the present time, considerable uncertainty exists regarding the structure and energetics of all three types of disturbance mentioned above and the nature of the interactions that take place between them because of the lack of radiosonde data. There is also serious question as to whether satellite data alone will be sufficient for defining these disturbances over vast regions of the Atlantic where radiosonde coverage is not available. (b) INTERNAL STRUCTURE OF CONVECTIVE ENSEMBLES The limited data presently available suggest the following view of the internal structure of convective ensembles. Within regions of disturbed weather, active convection is confined to around 10 percent of the surface area. Much of the remaining area is cloudy, but this cloudiness is associated with inactive cloud debris which produces little measurable rain. The active convection sometimes takes the form of isolated cells; at other times it is organized into lines, rings, or horseshoe-shaped patterns which have mesoscale dimensions. Saturated updrafts occupy only about 10 percent of the area of active convection (about 1 percent of the regions of disturbed weather). The re- maining 90 percent is divided between a small area of strong saturated down- drafts and a larger area of weak, unsaturated downdrafts which occupy most of the rain area. Updraft and saturated downdraft speeds are believed to be on the order of 1-10 m sec"1; unsaturated downdraft speeds, perhaps a few

OCR for page 4
tenths of a m sec"1. Updrafts are strong enough to carry liquid water well above its condensation level, into the upper, anvil-shaped portion of the clouds, where vertical wind shear disperses it over a wide area. This explains why the rain areas are large compared to the updraft areas. A portion of the rain evaporates as it passes through relatively dry, midtropospheric air, cooling the latter and causing it to become more dense than its environment; hence the downdrafts. Updrafts consist of air originating in the subcloud layer. Despite their rela- tively small dimensions (updraft radii in excess of 2 km have never been ob- served), there is strong evidence that the inner cores of the updrafts do not mix with their environment. For this reason, they have often been referred to as "undilute hot towers." The actual temperature excess of the updraft over its environment has been estimated in various studies to lie between 0.1 °C and 10°C. There are strong thermodynamic arguments for placing it in the range from 2°C to 5°C at the 500-mbar level. Temperature excesses of this magnitude could produce updrafts on the order of 100 m sec"1 or more were it not for the weight of liquid water and the retarding influence of form drag. Downdrafts have been measured to be on the order of 1°C cooler than their environment with the largest anomalies occurring near the ground. In well-developed mesoscale convective systems there is a strong concen- tration of cyclonic vorticity in the low-level wind field at the base of the up- draft. Values on the order of 1 X 10"3 to 5 X 10"3 sec"1 have been measured by aircraft on a number of occasions. These large vorticities are accompanied by horizontal convergences of comparable magnitude. At the levels of cirrus outflow (about 250 mbar) vorticity and divergence, as inferred from ATS cloud motions, are opposite in sign from those at low levels and are diffused over a larger area. Gusty surface winds and cool temperatures associated with downdrafts cause a local enhancement of sensible heat flux from the sea surface, which results in a rather rapid modification of the downdraft air. Organized cloud groups apparently originate when the low-level mixed layer attains a thickness such that it extends above the condensation level of the air within it, over a distance roughly 10 km on a side. Small cloudlets, 100-300 m in radius, break out where the moisture subcloud eddies reach their condensation level. In undisturbed weather, time-lapse pictures have shown that cloud groups often move much more slowly than the wind, some- times remaining nearly stationary over or just downwind of a sea-surface warm spot. Individual cloudlets form on the upwind side of the group and perish as they become the last cloud on the downwind side. Relatively little is known concerning the variations in the internal structure of mesoconvective elements over their life cycles, particularly

OCR for page 4
during the development stages. Of particular interest is the question of whether large low-level vorticities are present initially or whether these de- velop as a result of conservation of circulation during a prolonged period of boundary-layer convergence. It will also be important to establish the "re- laxation time" over which downdraft air is modified until it assumes typical boundary-layer characteristics. (c) BULK PROPERTIES OF CONVECTIVE ENSEMBLES These have been studied independently by a number of investigators using compositing and spectrum-analysis techniques on radiosonde data from the western Pacific. Many of the following statistics represent averages of many individual systems obtained from these studies. Disturbed weather tends to occur near or slightly east of the low-level troughs in "cloud clusters" with dimensions on the order of 2-10° of latitude on a side. Cloud clusters represent local maxima in cyclonic vorticity (about 10"5 sec"1) and convergence (about 5 X 10"6 sec"1) in the lower-tropospheric flow and anticyclonic vorticity (about 10~s sec"1) and divergence (1 X 10~s to 2 X 10~5 sec"1) at the 200-mbar level. The low-level convergence of mass and moisture is spread through a rather deep layer of the lower troposphere, with only about 30 percent of the mass and about 50 percent of the moisture being supplied by the subcloud layer. In the area average disturbed regions are cool in the lower troposphere, warm near the 300-mbar level, and cool at the tropopause, relative to their surroundings. Temperature anomalies are on the order of 1-2°C at the surface and 0.5-1.0°C at the higher levels. Above the subcloud layer, relative humidi- ties are 10-30 percent higher than in undisturbed regions but are still well below saturation. (Within the subcloud layer there is very little humidity dif- ference between disturbed and undisturbed conditions.) Maximum rising motion occurs near the 300-mbar level, where it reaches about 3 cm sec"1. This is consistent with observed rainfall rates on the order of a few centi- meters per day. Most of the above statistics represent broad averages. It would be highly desirable to know how much variability there is from one cloud cluster to another and over the lifetime of an individual cluster. It will also be of prime importance to determine how much information the satellite can provide regarding the bulk properties of convective ensembles. 2.2 Interrelations between the Observed Fields (a-b) The synoptic-scale environment could influence the intensity and inter- nal structure of convection in at least four different ways:

OCR for page 4
1. It has been proposed that variations in synoptic-scale static stability exert a controlling influence on organized convection in the tropics. Although in a general sense, convection can be viewed as a consequence of the destab- lizing influences on the tropical atmosphere, it does not necessarily follow that convection must take place in those regions where the static stability is lowest. In fact, the reverse appears to be true: regions of disturbed weather are observed to be more stable than their surroundings, as a result of the stabilizing influence of convection that has already taken place. Apparently it is organized convection that controls the synoptic-scale distribution of static stability,* and not the reverse. 2. It has long been recognized that low-level convergence is a necessary condition for the maintenance of deep, penetrative convection which is characteristic of disturbed weather in the tropics. When this ingredient is absent, the size spectrum of clouds sized exhibits a logarithmic distribution, with very few, if any, clouds reaching radii of 1.5-2.0 km (which appears to be the minimum size required to penetrate, undiluted, through the middle troposphere). In the absence of low-level convergence, clouds reach heights of only 3-6 km, and very little precipitation reaches the ground. However, when low-level convergence is present, the size distribution becomes highly bimodal, with just the very large and very small clouds present. Middle-size clouds are thought to be suppressed by the stabilizing influence of the large clouds, whose circulations replace the ambient subcloud layer air with cooler, drier downdraft air. But here again we have a "chicken and egg" problem, for it might equally well be said that the presence of deep cumulus convection is a necessary condition for strong synoptic-scale divergent motions at any level in the tropics. This has been shown to be a consequence of scaling considerations. It has been hypothesized that the mutual dependence of the synoptic-scale divergence field and the distribution of deep cumulus convection in the tropics may, under certain conditions, provide a growth mechanism for synoptic-scale perturbations. This phenomenon is called conditional insta- bility of the second kind (CISK) to distinguish it from ordinary conditional instability which gives rise to cumulus-scale motions. When stated in this way, the CISK hypothesis is irrefutable. The problem is that in order to make use of this concept in numerical prediction models, it is necessary to make additional assumptions in order to relate the subcloud layer convergence to the synoptic-scale motion field and to parameterize the collective effects of the convection upon the synoptic motions. At present, we are lacking a sound physical basis for these assumptions. We will return to this "Variations in parcel static stability are actually very small. 10

OCR for page 4
problem when we discuss the interrelation between the bulk properties of convective ensembles and the synoptic-scale fields. 3. It also seems likely that the synoptic-scale vertical wind shear should have an influence, if not on the amount of convection, at least upon its in- ternal structure and upon the type of mesoscale organization that it assumes. For example, it is observed that strong vertical wind shear is a necessary condition for the development of severe convective storms in middle latitudes. 4. Water temperature through wind-driven mixing. (b-c) The bulk properties of convective ensembles can be deduced from their internal structure, and they can also be measured directly. In this respect, the observational plan for the experiment contains a certain amount of redun- dancy. However, since considerable uncertainty is inherent in both sets of measurements, this redundancy plays a crucial role in the design of the experiment. It has been mentioned that much of the convergence of mass and moisture into regions of disturbed weather takes place above the subcloud layer. The vertical profile of synoptic-scale divergence may represent the superposed effects of three types of internal circulation; buoyant updrafts, which carry subcloud air upward to near the 200-mbar level; entrainment into updrafts; and evaporatively cooled downdrafts, which replace some of the loss from the subcloud layer with air that has entered the region from above. It appears that the strong concentrations of cyclonic vorticity at the bases of updrafts can easily account for the low-level (synoptic-scale) vorticity ob- served in the disturbed regions, even if the remaining some 99 percent of the area contributes nothing. Because of this tendency for vorticity to be strongly concentrated, there may very well be quite large frictional losses of cyclonic vorticity to the ground, even though the synoptic-scale low-level vorticity is not large. This may explain why, in a bulk sense, there is a flux of cyclonic vorticity into disturbed regions at low levels and an export of anticyclonic vorticity at upper-tropospheric levels. (a-c) The bulk properties of convective ensembles provide a convenient means of parameterizing the effects of convection upon the synoptic-scale motions. Using the data on bulk properties, it is possible to compute the rate of gen- eration of synoptic-scale available potential energy by condensation heating and the rate at which this is converted into kinetic energy. This has been done successfully, using compositing and spectrum-analysis techniques, and there is every reason to believe that it can be done for individual cases if the experi- ment is designed properly. It should also be possible to parameterize the vertical flux of momentum by convective-scale motions. 11

OCR for page 4
The most difficult task in parameterizing the convective-scale motions is to predict the future distribution of convective activity on the basis of synoptic- scale information. In the case of the symmetric hurricane, it is possible to ob- tain an analytic boundary-layer solution that prescribes the distribution of low-level convergence, which, in turn, is assumed to be proportional to the intensity of convective activity. However, in the general case, a boundary- layer solution is much more difficult to obtain. Moreover, it is not at all clear that the distribution of convective activity is determined uniquely by fric- tional convergence in the boundary layer. Some other source of vertical motions may yet prove to be of equal or greater importance. 2.3 Concluding Remarks The prospect of the experiment has stimulated a great deal of research in this field. If progress continues at its present rate, it is not unreasonable to expect that in the three years that remain before the experiment, some of the current thinking on the scale-interaction problem will have to be revised. This is why in the previous sections, the observational basis for hypotheses has been stressed rather than the hypotheses themselves. And this is why the scientific objectives of the experiment have been expressed in the terms of consolidating and extending our observational knowledge rather than as a test of some spe- cific hypothesis. In this way, we can assure that the experiment will be highly relevant to the problems existing at the time of the experiment. At the same time, we recognize the importance of hypothesis-making in general. Without hypotheses to be tested, there is no basis for distinguishing among a variety of events of equal a priori relevance. While no one hypothesis is presently capable of explaining the complicated convective phenomena of the tropics, existing hypotheses, such as CISK, at least focus attention on events in the subcloud and frictional boundary layers as deserving of concerted observation. 12