On the Global Variation of Precipitating Layer Clouds C €

On the Global Variation of Precipitating Layer Clouds c €

B. F. Ryan CSIRO Division of Atmospheric Research, Aspendaie, Victoria, Australia

ABSTRACT

The aim of the Global Energy and Water Cycle Experiment Cloud System Study (GCSS) is to promote the description and understanding of key cloud system processes, with the aim of developing and improving the representation of cloud processes in general circulation models. The GCSS Science Panel identified a need to document important observational gaps in the structure of cloud systems inhibiting the development of cloud-resolving models as a tool for parameterizing cloud systems in general circulation models. The nature of precipitating layer clouds around the world is not well documented. To better quantify this, a synthesis of observations of these types of clouds made during field experiments conducted around the world has been developed. The synthesis draws on observations made in Australia, Canada, China, Israel, Japan, Russia, the Ukraine, the United States, and several European countries. The survey examines the global variation of the horizontal scales of cloud and precipitation, embedded phenomena such as rainbands, conveyor belt characteristics, ice crystal and water droplet concentrations, and raindrop and ice crystal size distributions.

1. Introduction models is data limited and more effort needs to be devoted to observational aspects driven by model The Global Energy and Water Cycle Experiment needs. Consequently, the GCSS Science Panel has (GEWEX) Scientific Steering Group identified one of undertaken the task of identifying areas where obser- the biggest challenges in the World Climate Research vations to support the GCSS modeling objectives are Programme as an improvement in the understanding lacking. of how the wide range of processes within clouds af- The GCSS Science Panel identified layer cloud fects the atmosphere on the large scale. The GEWEX systems as one type of cloud system where there is a Cloud System Study (GCSS) has therefore set out to need to document the important observational gaps in- promote the description and understanding of key hibiting the continued exploitation of cloud-resolving cloud system processes, with the aim of developing models as a tool for developing and testing the param- and improving the representation of cloud processes eterization of cloud systems for application in GCMs. in general circulation models (GCMs) (Browning Precipitating layer clouds are a significant subgroup 1994). of the layer cloud classification. In this survey of the GCSS was conceived initially as a modeling pro- characteristics of precipitating layer cloud systems, I gram that would exploit existing datasets as far as pos- will attempt to identify where observations are most sible. However, it has become increasingly clear that needed and I will identify some of the model cloud the development and validation of cloud-resolving features in GCMs that should be validated against both observations and cloud-resolving models. The paper Corresponding author address: Dr. Brian Ryan, CSIRO, Divi- is limited to a survey of midlatitude storm cloud sys- sion of Atmospheric Research, PMB1, Aspendaie, Victoria 3195, tems and consequently the characteristics of precipi- Australia. tating layer clouds associated with tropical cloud sys- E-mail: [email protected] In final form 6 June 1995. tems; shallow precipitating stratus and stratocumulus ©1996 American Meteorological Society systems are not covered in the paper.

Bulletin of the American Meteorological Society 53

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC In the midlatitudes, layer cloud systems are asso- Major meteorological observing programs are coor- ciated with frontal systems, depressions in the west- dinated studies of midlatitude precipitating cloud sys- erlies, lows cutoff from the westerlies, and tropical tems. Examples of these types of studies are given in extratropical cloud bands. They are strongly modu- appendix A. Major cloud seeding studies, undertaken lated by orography and may be associated with orga- as part of a cloud seeding program, usually focus on nized convection in the form of rainbands. the microphysical properties of the clouds. Examples The precipitating layer clouds form an integral part of major cloud seeding research studies are given in of the global water and energy cycle that drives the appendix B. The locations of the major meteorologi- earth's climate system. In this paper precipitating layer cal observing programs and the major cloud seeding clouds include both nimbostratus and altostratus programs that are relevant to the study of precipitat- clouds. Altostratus clouds are included because of the ing layer clouds are shown in Fig. 1. Figure 1 also importance of the evaporation of virga to the water shows the locations of case studies used in the paper. cycle. Warren et al. (1986) show that over the land The case studies examined are confined to regions the zonal mean of the average amount of altostratus where there are no reported major meteorological cloud between 30° and 60° latitude varies from 15% observing programs or major cloud seeding programs. to 25% in the Northern Hemisphere and from 16% to It is clear from Fig. 1 that there have been exten- nearly 60% in the Southern Hemisphere. Over the sive observational and cloud seeding programs that oceans between 30° and 60° latitude the zonal mean extend across North America and western Europe. altostratus cloud cover varies from 20% to 35% in Over the land and in coastal regions there have been both hemispheres and varies as a function of the sea- many case studies made in the United States, Canada, son (Warren et al. 1988). The precipitating layer the United Kingdom, France, and Germany. The clouds are far from uniform in structure. For example, Beaufort Arctic Storms Experiment is a recent study Stephens and Greenwald (1991) showed that the al- in northern Canada that should contribute information bedo and liquid water path are significantly different on the structure of precipitating layer clouds over the for midlatitude oceanic clouds compared to clouds Arctic region. over the tropical oceans. Heymsfield (1993) in a re- In southern Europe and the Middle East, cloud view of the structure of clouds suggested that the water seeding studies have been undertaken in Spain, Mo- phases in a cloud depends upon the temperature, the rocco, and Israel. In eastern Europe major field stud- age of the cloud, the ambient aerosol population, and ies have been reported from Russia and the Ukraine, the composition and size, as well as a host of other and in China a major cloud seeding study has been possible factors. made in the Xinjiang Province. In Japan, several case An inherent feature of the physical processes in studies provide observations for comparison with the cloud systems is the interaction between scales, which European and North America studies. Except for extend from the synoptic scale down to the mesos- FRONT-87, there are no significant observations over cale. Systematic field observations are needed across either the Atlantic or Pacific Oceans, although the data these scales to develop and validate cloud-scale void over the Atlantic will be addressed during the parameterizations. In the case of the parameterization Fronts and Atlantic Storm Tracks Experiment of precipitating layer clouds, validation is required (FASTEX) (Browning 1994). over both the land and the sea. Where layer clouds In the Southern Hemisphere, the structure of pre- form over the land, parameterizations need to be vali- cipitating layer clouds has been investigated as part dated in the absence and the presence of significant of the Australian Cold Fronts Research Program and orography. through studies made during cloud physics and seeding projects. These studies focused on the southeastern corner of Australia, although there is one unpub- 2.International studies of precipitating lished cloud seeding study north of Perth in Western clouds Australia. One case study of precipitating layer clouds has been reported from Cape Town in South Africa. Field studies of frontal clouds over the past 20 years The Southern Alps Experiment (SALPEX) in New can be classified into three categories: namely, 1) ma- Zealand is a new initiative to investigate the clouds jor meteorological observing programs, 2) major and precipitation upstream of and over the southern cloud seeding programs, and 3) specific case studies. Alps. Apart from satellite observations there are no

54 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC FIG. 1. A map showing the location of the major field experiments (see appendix A), the major cloud seeding studies (see appendix B), and the case studies discussed in the paper. systematic studies over the vast ocean areas, and I the northeast ahead of the upper-level trough. A cold have been unable to find papers on the structure of conveyor belt approaches the cyclone at low levels precipitating layer clouds over South America. from the northeast, rises near the center of the low and merges with the warm conveyor belt as it leaves the system toward the northeast. A third stream of dry 3. Synoptic-scale structure of stream originates in the upper troposphere upstream precipitating clouds of the cyclone and descends into the systems from the northwest. Figure 2 shows the conceptual model ap- The precipitating layer clouds associated with plied to a split (kata) cold front. Browning (1990) fronts and midlatitude cyclones are organized on the shows that both the instant occlusion and classical oc- synoptic scale. If the flow is visualized as occurring clusion can be schematically portrayed by the con- relative to moist or dry isentropic surfaces, cloud for- veyor belt model. The limitation of the conveyor belt mation (dissipation) is associated with regions of as- approach is that it provides only a coarse description cent (descent) of air within the sloping portions of the of the nature of the flow through a midlatitude cy- isentropic surfaces (see, e.g., Browning 1990 or clone. Mass and Shultz (1993) analyzed trajectories Carlson 1991). Harrold (1973) introduced the concept calculated from a high-resolution simulation of a mid- of the "conveyor belt" and identified the warm con- latitude cyclone and showed that the detailed flow veyor belt as the warm moist airstream that ascends is far more complex than that suggested by the con- relative to the warm front. Carlson (1980) extended veyor belt model. Nevertheless on the scale of the these ideas to a conceptual model of a developing cloud systems, the conveyor belt model is a valuable extratropical cyclone. The Carlson conceptual model diagnostic tool. consists of three streams of air. In the Northern Hemi- The warm conveyor belt is the primary midlevel sphere, a warm conveyor belt enters the cyclone from cloud-producing flow. The conceptual model of the the south or southeast, rises in the warm sector and warm conveyor belt has been applied to account for over the surface warm front, and exits the systems to aspects of cloud and precipitation distribution in north-

Bulletin of the American Meteorological Society 55

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC upper-tropospheric air influences the cloud pattern by generating moist instabilities in the warm conveyor belt. If the dry air undercuts the warm frontal zone or warm sector, it leads to evaporation of precipitation and instability in the subcloud layer (Ryan et al. 1990), and if it overruns the warm conveyor belt, it leads to the generation of a shallow moist zone with potential instability at cloud top (Browning and Monk 1982). The cloud structure below the warm conveyor belt is strongly dependent on the thermodynamic properties of the air and hence the geographic location. If the low-level air is subsiding (see Browning 1990) or has risen from a dry continental region, such as oc- curs in southeastern Australia (Ryan et al. 1989), precipitation from the overlying warm conveyor belt evaporates into the subsaturated cold conveyor belt. Clough and Franks (1991) point out that the ice and snow precipitation typically has a fall speed of the order of 1 m s_1, so it may take about one hour to reach the ground. When the freezing level is well below cloud base, substantial sublimation takes place before the snow begins to melt. The sublimating snow is sufficient to maintain an air parcel descending at a rate of 10-30 cm s_1 at near saturation, whereas evaporat- ing rain maintains only the parcel at about 40%-60% FIG. 2. A schematic portrayal of the airflow in an extratropical saturation. Where the air beneath the warm conveyor cyclone in which the warm conveyor belt (solid arrow) is undergoing forward sloping ascent ahead of a kata-cold front, belt rises and reaches saturation as in parts of the cold before rising above a flow of cold air ahead of the warm front. conveyor belt close to the low center and surface front, The dotted arrow is the cold conveyor belt, and midtropospheric deep layer cloud systems form. Under these conditions air is shown by a dashed arrow over running the cold front cloud base may be below 1 km and there is little or generating potential instability in the upper portions of the warm no subcloud evaporation. conveyor belt. Panel (a) is a plan view and panel (b) is vertical section along line AB in (a). Flows are shown relative to the The warm conveyor belt is typically several thou- moving frontal system (from Browning 1990). sand kilometers long, a few hundred kilometers wide, and 1-2 km in depth (see, e.g., Harrold 1973). The scales of the conveyor belt flows are such that they western Europe (Harrold 1973; Browning and Monk are able to be resolved by limited area numerical 1982), the United States (Carlson 1980), and Canada models (LAMs) and probably by general circulation (Stewart et al. 1988). In Australia, studies of the sum- models run at high resolution. Models with explicit mertime cool change (Ryan et al. 1985a) and winter- liquid water schemes should be able to resolve the time cold fronts (May et al. 1991) of all three airflows nonconvective component of the conveyor belt cloud. were identified. In eastern Europe, Matkovskii and However, as will be shown in the next section, much Shakina (1982) have identified the warm conveyor of the cloud in the conveyor belt is coupled to mesos- belt and the dry airstream. Finally, Wang et al. (1992) cale convective activity that is not explicitly resolved identified the warm conveyor belt in a case study of in either GCMs or LAMs. a cold front passing over north China. A fundamental aspect of these layer cloud systems In all of these studies the western and poleward is their water budget. There are very few studies ad- edges of the cloud systems are well defined and the dressing water budgets associated with individual pre- cloud bands follow the storm-relative isentropic cipitating layer cloud systems. The moisture budget streamlines. The cloud cover in the warm conveyor depends both on the moisture flux into the cloud sys- belt consists of mid- and high-level layers of altocu- tem as well as the redistribution of the moisture in the mulus and altostratus or cirrus. The intrusion of dry vertical. With the latter process, the vertical ascent

56 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC associated with convection may be as important as the An important dynamic characteristic of the warm large-scale ascent. Evaporative processes play a very frontal and warm sector clouds is the so-called snow- important role in the vertical redistribution of mois- generating cells. In an early radar-based survey of the ture. Below cloud base, evaporation into the subcloud nature of cloud and precipitation Plank et al. (1955) layer is important, while above the top of the layer found that the ice phase was critical to the formation cloud the evaporation from penetrating embedded of stratiform precipitation and that ice crystals gen- convective towers moistens the upper layers of the erally occurred in the form of streamerlike trails. The atmosphere. Evaporative cooling is also a powerful streamers seeded the lower cloud decks or alterna- feedback mechanism on the dynamics. tively constituted a major portion of the precipitation. Subsequently, "The McGill Stormy Weather Group" made a study of these "generating cells" using range 4.Mesoscale structure of precipitating height indicator radar patterns (see, e.g., Douglas et clouds al. 1957), and later Wexler and Atlas (1959) used combined radar and aircraft observations to show that The vast majority of the recent observational stud- the echo structure associated with the generating cells ies of the mesoscale (20-200 km) structure of precipi- is embedded in the stratiform cloud deck. More re- tating clouds has been associated with large field ex- cently, the University of Washington group showed periments requiring the synthesis of data from Dop- that the relatively intense precipitation associated with pler radar, aircraft, radiosondes, dropsondes, wind wide rainbands appears to be associated with gener- profilers, and satellite imagery. The majority of these ating cells aloft, which in turn seed the feeder layer projects have taken place in North America and west- clouds below (Hobbs 1978). ern and eastern Europe. Most of the experiments iden- In contrast to the wide rainbands, narrow cold fron- tify the presence of rainbands or embedded convec- tal rainbands are associated with a narrow updraft re- tion. In this paper the term "embedded convection" gion. The updraft is coupled to a weaker downdraft is used when there is insufficient data to classify the that is associated with heavy precipitation. The struc- characteristics of the rainband. ture of the narrow cold frontal rainband is characterized by young, vigorous convective cloud elements. a. Rainband structure The strong updraft, with high liquid water content and The current classification of rainbands is based growth of precipitation particles by the accretion of largely on observational studies made in the United liquid water contrasts sharply with the stratiform cloud Kingdom and on the west coast of the United States. and precipitation in which it is embedded (see, e.g., These studies showed that frontal precipitation has an Houze 1981). embedded banded structure (see, e.g., Houze and There are many studies confirming the generality Hobbs 1982; Browning 1985). The bands have been of this classification of rainbands across the continen- classified as narrow or wide rainbands according to tal United States and in other regions of the world. their scale. The CYCLES observations in the U.S. and Table 1 gives the cloud base height, cloud top height, the U.K. observations are representative of rainbands rainband width, melting level, and season for wide associated with fronts that have formed between 40° rainbands observed in the United Kingdom (Brown- and 50°N in air of recent maritime origin. ing and Monk 1982; S. Moss and D. Johnson 1994, Wide rainbands are typically 50-75 km across and personal communication), the United States (Hobbs found embedded within and parallel to the frontal et al. 1980), Russia (Belyakov et al. 1984; Bezrukova zone. Matejka et al. (1980) classified the wide bands and Shmeter 1989; Shmeter 1992), Finland (Saarikivi into warm frontal bands, warm sector bands, wide 1989), China (Gao and Zhang; 1988; You and Liu cold frontal bands, and prefrontal cold surge bands. 1989), Israel (Gagin 1981), and Australia (Ryan and Narrow cold frontal rainbands are 5-25 km wide and Wilson 1985). tend to be generated by forced convection, sometimes This limited survey suggests that the classification referred to as line convection. They tend to occur in of the various types of rainbands appears to be valid the cold season when air ahead of the surface cold over a wide range of environmental conditions. It is front has near-neutral stability (Browning 1990). clear that with almost all the cloud systems, rainbands Postfrontal rainbands form from ascent above the cold and embedded convection are a characteristic of the air behind the front. precipitating midlevel-layer clouds. The wide mesos-

Bulletin of the American Meteorological Society 57

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC TABLE 1. The characteristics of wide rainbands.

Location/reference Cloud top Cloud base (km) Width (km) 0°C References

United Kingdom 7-8 km (all) 1-3 60-150 — Browning and Monk 1982

5.6 km — — 1.8 km S. Moss and D. Johnson 1994

(personal communication)

United States ~7 km ~1 50-75 3 km Hobbs et al. 1980

9-6 km < 1 35-70 — Elliot and Hovind 1964 Russia 4-6 km ~1 50-100 — Belyakov et al. 1984 Berukova and Shmeter 1989 Shmeter 1992

Finland 6.5 km (summer) — 35 — Saavikiv 1983

3.0 km (winter) .— 18

China 1-10 km (winter) 1-3 20-40 < CB Gao and Zang 1988

You and Liu 1989

Israel 3-6 km (winter) — 20-120 -- Gagin 1981

Australia 4-6 km (summer) 1-3 -40 3 km Ryan and Wilson 1985 cale rainbands (warm frontal, warm sector, and wide ration in the subcloud layer. Furthermore, circulations cold frontal rainbands) are typically 50-100 km in associated with the rainband structure modify the width, while wide mesoscale snowbands may be moisture content of the air in the gaps between somewhat narrower at 20-30 km (Table 1). Narrow rainbands. However, from the modeling point of view cold frontal rainbands have been well documented in it is clear that the rainband structure is resolvable only the United States and Europe. Studies of the narrow in high-resolution cloud and mesoscale models and rainbands are needed from other regions of the globe that general circulation models are unable to resolve to show if there are any regional biases in the forma- the rainbands. This implies that rainbands need to be tion of narrow rainbands. parameterized. This is no simple task, because while Rainband structures have also been identified in the observations show that they form as either wide Spain (Goyer and Cunningham 1981); Morocco or narrow rainbands, there are several theories pro- (Baddour et al. 1989); Japan, in both warm frontal posed to explain the mechanisms generating the rainbands (Murakami et al. 1988,1992) and warm sec- rainbands (see Stewart 1994). tor rainbands (Nozumi and Arakawa 1968); and South Africa (Bruintjes 1988). In each of these examples the b. The role of subcloud layer, the melting level, and rainband width was not specified. topography How important are rainbands for numerical Regional differences in surface weather and in the weather prediction and for climate studies, and do mesoscale structure of frontal layer cloud systems are they need to be represented in numerical weather pre- forced by local changes in the temperature and humid- diction models and climate models? The accurate ity of the subcloud structure as well as the height of forecasting of precipitation depends both on forecast- the melting layer above the ground. In addition, the ing the timing and the intensity of rainbands. In cli- mesoscale structure of frontal layer cloud responds to mate studies, the efficiency of precipitation events de- orographic forcing in the form of coastal convergence pends on the intensity of rainbands and on the evapo- and topographic uplift by hills and mountains. Finally,

58 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC local mesoscale circulations, such as land and sea ditions, it is not uncommon for a cloud system to com- breezes, can interact with an approaching front. prise several layers each about 1 km thick and sepa- rated by a few hundred meters (J. Jensen 1994, per- 1) SUBCLOUD-LAYER EVAPORATION sonal communication). On the west coast of the United States both warm sector and wide cold frontal rainbands have cloud base 2) THE DYNAMICAL FORCING BY THE MELTING LAYER at about 1 km when the low-level air beneath the warm In many areas of the world, the precipitation at the conveyor belt is cool and moist (see, e.g., Hobbs et surface characteristically associated with winter al. 1980). These stratiform clouds are deep with the storms varies in form from snow to freezing rain and melting level at typically 1-2 km above cloud base. ice pellets and then finally to rain. However, this tran- When cloud forms beneath the warm conveyor belt, sition zone is not necessarily tied to any surface fea- the layer clouds have very high precipitation efficien- ture. Stewart (1992) points out that a favorable loca- cies and subcloud evaporation is minimized. For ex- tion for it is nevertheless in the vicinity of coastlines ample, Houze et al. (1981) estimates the efficiency of due to the imposed temperature gradient spanning 0° a warm frontal rainband over Washington State to be in that region. In both Canada and Scandinavia heavy between 60% and 90%. However, earlier Hobbs et al. precipitation has been observed and characterized by (1980) found that two wide cold frontal bands char- snow over the land, where temperatures were below acterized by the "seeder-feeder" process had efficien- freezing, a mixture of rain and snow along the coast cies of 80% and 20%. In the latter case, the efficiency and rain over the ocean. Stewart suggests the scale was strongly affected by low-level evaporation. of the transition to be about 25 km extending to A characteristic of the katafronts (see, e.g., Brown- 100 km. He also suggests that the transition region has ing 1990) is that the warm conveyor belt ascends over a dynamical structure that leads to mesoscale circu- a cold conveyor belt, which may be both cold and dry. lations in close proximity to the transition region, The warm frontal precipitation falling from the warm thereby modifying the location and intensity of the conveyor belt evaporates as it falls into the dry sub- precipitation. cloud air (Fig. 3). During summer in southeastern Australia the subcloud layer beneath the warm con- 3) OROGRAPHIC FORCING OF LAYER CLOUDS veyor belt is cloud free. As shown by Ryan et al. There are many examples of the impact of orogra- (1989), this hot northerly airstream in the subcloud phy on cold fronts. It is beyond the scope of this pa- layer can extend several hundred kilometers offshore per to give a detailed review of fronts and orography. before being modified. Relative humidities of less An extensive review of both observations and theory than 30% in the prefrontal subcloud layer air are com- is given by Egger and Hoinka (1992). They list sev- mon (see Ryan et al. 1985a). Over the land, cloud base eral observational programs that have studied the is typically 3-4 km for the warm sector rainband with the freezing level near 3 km. The subcloud layer is cooled and moistened by the virga falling from the midlevel layer cloud. Using a nonhy- drostatic cloud model, Ryan et al. (1990) showed that the evaporative cooling generated by virga falling from the middle-level cloud was sufficient to generate the observed squall line in the cloud layer. By contrast, in southern Australia both during winter and spring the pre- FIG. 3. A schematic representation of the regions of cloud and rain along sections frontal airstream associated with A-B in Fig. 2. The numbers represent precipitation types as follows: (1) warm frontal precipitation; (2) convective precipitation and generating cells associated with an upper cutoff lows is both cooler and cold front; (3) precipitation from the upper cold front falling through the "freezer zone"; moister and cloud base is more (4) shallow moist zone between the upper and surface cold fronts; and (5) shallow typically 1-2 km. Under these con- precipitation at the surface cold front (from Browning 1990).

Bulletin of the American Meteorological Society 59

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC impact of orography on fronts. The majority of these cal cycle in a fully coupled atmosphere-ocean model. observational programs have focused on the dynami- These surface water budgets critically depend on the cal structure of the fronts while relatively few studies parameterization of the rainfall over the large moun- have concentrated on the cloud structure. tain ranges. The review by Egger and Hoinka (1992) It is clear that orographic forcing considerably points to several datasets in both Europe and the modulates the precipitation associated with layer United States that should be suitable to validate the clouds. The direct effects are generally to increase generation of middle-level cloud in the vicinity of the upward flow of air on the windward side, which mountain ranges by GCM and cloud-resolving mod- in turn enhances the formation of clouds and pre- els. In particular, there seem to be very complete cipitation. On the leeward side there is descent that datasets associated with CYCLES, GALE WISP, and suppresses the formation of clouds and precipitation. the German Front Experiment (Hoinka and Volkert The forced ascent over the hill or mountain causes the 1987). However, there is still a need for high quality formation of "feeder" clouds. Above the feeder clouds datasets from other regions of the world. are "seeder" clouds forced by the large-scale ascent. The precipitation generated in the seeder clouds is 4) COASTAL AND DIURNAL FORCING enhanced by the water-rich environment of the Coastal effects modulate frontal cloud cover, and feeder clouds leading to increased rainfall (see, e.g., this has been observed in different geographic loca- Hobbs 1978). tions. For example, Dodge and Burpee (1993) showed Large mountain ranges are an important constraint that on the Mid-Atlantic east coast of the United States on the development of the midlevel cloud systems on rainband activity was modulated by the diurnal cycle the synoptic scale. For example, in the United States, and showed land-ocean differences in the structure and upstream of the Rocky Mountains, the orographic of the rainbands. In particular, there was a greater oc- barrier partially cuts off the moisture to the warm currence at night of both strong echoes (convective conveyor belt coming from the south into the frontal cloud) and weak echoes (stratiform cloud) 100- system. On the leeside of Colorado's Front Range, 200 km offshore and parallel to the coast. In this in- snowfall is linked to cold air damming and frontal stance the cloud field was at least partially attributed overrunning of the trapped cold air by a moist south- to the consequences of the Gulf Stream. A second easterly flow that leads to heavy precipitation (see, example is in Australia where Garratt (1987) showed e.g., Cotton and Anthes 1989). The Winter Icing that coastal effects, such as the development of a sea Storms Project (WISP) over the Colorado Front breeze, lead to a tendency for a high percentage of Range region is concerned with improving the under- fronts to cross the coastline in the afternoon and early standing of the dynamical and microphysical pro- evening hours. cesses leading to the production and depletion of liquid water in such orographically affected systems 5) SUMMARY (Rasmussen et al. 1992). Hobbs et al. (1990) have Observations on the synoptic-scale support the con- shown that cold fronts aloft play an important role in ceptual model of precipitating layer clouds, and there producing significant weather east of the Rocky is in general an accepted classification of rainbands. Mountains. They have developed a conceptual model Furthermore it is clear that layer clouds may be modu- that differs from the classical model of a midlatitude lated by processes occurring in the subcloud layer. cyclone. The Hobbs et al. (1990) conceptual model From a climate-modeling perspective the rainband cir- has a belt of warm moist air aloft (warm conveyor culations and the associated evaporative cooling, belt) underlying a belt of colder and drier air, thereby melting layer effects, orographic lifting, coastal con- generating severe weather along an upper-level cold trasts, and diurnal effects may not be resolved by a front. There is no cold conveyor belt, and the surface GCM and hence they require parameterization. These trough is behind the cold front aloft. effects are important climatologically and should not From the GCM perspective the effect of orography be ignored as they markedly affect precipitation effi- on cloud systems, cloud evolution, and precipitation ciencies of the layer cloud systems and the relative is obviously important and there is a continuing need humidity of the neighboring areas between rainbands. for the proper accounting of such features. The flow The most effective tools for developing the para- of freshwater from large rivers that rise in mountains meterizations to describe these phenomena are mesos- and feeds into the ocean forms part of the hydrologi- cale and cloud-resolving models.

60 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC 5. Microphysical characteristics of thicker and longer lived and they refer to them as "ice- precipitating layer clouds producing Ac" (IP AC) clouds. A feature of the IP AC clouds is that there tends to be very little ice enhance- In climate simulations, the microphysical structure ment. Ice enhancement or ice multiplication is said to of the clouds is critical due to their impact on the ra- have taken place if the observed numbers of ice crys- diation. To calculate the radiative response of a cloud tals substantially exceed the number of ice crystals that it is necessary to know the phase of the water in the would have formed on the observed background ice cloud and to characterize the particle size distribution nucleus population. In contrast to the NIPAC, all but of the clouds, the ice crystal habit, and the thickness one of the examples given by Hobbs and Rangno of of the clouds. In addition to determining the radia- the altocumulus clouds showing ice multiplication had tive feedback of the clouds, the microphysics controls liquid water contents near cloud top exceeding the precipitation efficiency of the cloud system. In 0.3 g nr3. midlatitudes, precipitating layer clouds usually extend Many observations have shown that stratiform above the freezing level. It is convenient therefore to clouds less than 1 km deep have either very few ice classify the layer clouds into non-ice-producing strati- crystals or no ice crystals (Table 2). In particular, form clouds, non-ice-enhancing stratiform clouds, Heymsfield (1977) found that jet stream cloud tended and ice-enhancing stratiform clouds as defined by to have little horizontal variation and that some clouds Hobbs and Rangno (1985) and described in the next possessed liquid water at temperatures as low as section. -35°C. More recently, Heymsfield et al. (1991) examined two thin altocumulus clouds at about -30°C. a. Ice crystal formation in precipitating layer They pointed out that the absence of ice crystals in- clouds fers a dearth of ice nuclei, and this is contrary to the 1) NON-ICE-PRODUCING AND NON-ICE-ENHANCING assumption that ice crystal concentrations increase STRATIFORM CLOUDS with decreasing temperature. They hypothesized that Hobbs and Rangno (1985) described the cloud since ice nuclei concentrations in cirrus cloud are also microphysical structure of altocumulus (Ac), altostra- low while ice crystals are abundant, there is support tus (As), and nimbostratus (Ns). "Non-ice-producing for the idea that ice crystals produced at temperatures Ac" (NIPAC) contained no ice crystals or relatively below -40°C originate through the homogeneous low concentrations of ice crystals (<10 L1)- These freezing of water droplets. clouds were generally thin and often newly formed. The only deep clouds reported to have ice crystal Those clouds containing ice crystals were often numbers close to that predicted by the observed ice

TABLE 2. The characteristics of non-ice-enhancing stratiform clouds.

Location/reference Cloud type Cloud depth (km) n n > 20 /zm d (/zm) Ice Enhancement

United States (Hobbs Ac NIPAC <0.5 50-1000 0 7 no no and Rangno 1985)

AcIPAC 0.5-1.0 25-300 yes 10 yes no(lwc-O.l)

United States (Cooper Cap cloud -0.5 200-600 8 yes no(lwc<1.0) and Vali 1981)

Russia (Borovikov Ac 0.2-0.7 — 5-7 no no et al. 1963)

Israel (Gagin and Cu 3-4 500-1200 no 9 yes no (0.5-2.0) Neumann 1974)

Australia (King Ac and As -0.5 500 — 10 <0.11 no et al. 1979)

Bulletin of the American Meteorological Society 61

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC nucleus spectrum are in Israel (Gagin and Neumann ice crystals is the breadth of the water droplet spec- 1974; Gagin 1975). However, Rangno and Hobbs tra. In particular, they noted that the IP AC clouds had (1988) have questioned these findings as they appear droplets with diameters > 20 /urn in concentrations of to be inconsistent with similar cloud observations found more than 10 cm-3. Table 3 shows observations from in the United States. This criticism is supported by re- the United States, Russia, China, Spain, and Austra- cent measurements made by Levin (1994) in Israel. lia that support their observations. In all cases, the Levin found ice crystals in these clouds at concen- clouds glaciated and, with the exception of the seeder trations of up to three orders of magnitude above that clouds in China (Gao and Zhang 1988) and western expected on the basis of ice nuclei measurements. Victoria in Australia (King 1982), the cloud droplet spectra had observable concentrations of droplets 2) ICE-ENHANCING STRATIFORM CLOUDS greater than 20 /im in diameter. Hobbs and Rangno (1985) found that a difference Generally, the deeper systems were largely glaci- between the non-ice-producing stratiform clouds and ated cloud with segments of embedded convection. deeper stratiform clouds with high concentrations of This is true of most of the observations shown in

TABLE 3. The characteristics of ice-enhancing stratiform clouds.

Location/reference Cloud type Cloud depth n > 20 jj.m d (/an) Ice Enhancement

United States (Hobbs Ac IPAC 0.2-0.6 km 80-500 yes yes yes (lwc > 0.3) and Rangno 1985)

As/Ns 0.4-1.1 km 60-600 yes yes (lwc > 0.3)

United States (Cooper As/Ns Layered 200-400 yes 20 yes yes (0.2-0.5) and Saunders 1980) (orographic uplift)

United Kingdom (S. Moss Ns yes yes and D. Johnson 1994, personal communication)

Canada (Isaac 1991) As/Ns 50-100 yes 17-20 yes yes (0.0-0.4)

Russia (Borovikov Ns several km yes 7-8 yes et al. 1963)

China (Gao As (seeder) 1 km — no 8-18 yes no (0.1-0.5) and Zhang 1988)

St (feeder) 1 km — yes 4-8 yes (0.0-1.2)

Spain (Vali 1980) Ac, As, Ns 1.2-6.0 km 200 yes yes (0.0-0.3)

Morroco (Baddour Ac, As, Ns 3 km — yes no (0.0-0.3) and Rasmunsen 1989)

Australia (King 1982) As/Ns layered 100 6 yes yes (lwc <0.1) (no orographic uplift)

Australia (J. Jensen 1994, As/Ns layered 150-300 yes 10 yes yes (0.1-0.6) personal communication) (orographic uplift)

South Africa (Bruinjes 1988) As/Ns layered 100 20 yes yes

62 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC FIG. 4b. Ice water profiles through the cloud shown in Fig. 4a (from Paltridge et al. 1986). FIG. 4a. Liquid water content of a mixed phase altostratus cloud showing the high liquid water contents near cloud top for two different profiles of the same cloud shown as solid and open found in the regions of embedded convection. Away circles (from Paltridge et al. 1986). from regions of embedded convection the cloud water contents are small and the clouds are glaciated. Recent Table 3. A feature of several recent studies is that observations of 11 frontal cloud systems by S. Moss and while these clouds are largely glaciated, water drop- D. Johnson (1994, personal communication) suggest lets are often observed near the top of the clouds (see, that continental frontal clouds may have a smaller pro- e.g., Hobbs and Rangno 1985). Figure 4 shows an portion of ice in them for a given temperature than example from Paltridge et al. (1986). Calculations by maritime clouds (Fig. 5). The observations shown in Rauber and Tokay (1991) suggest that the liquid lay- Fig. 5 graphically illustrate that frontal layer clouds ers occur when there is an imbalance between the colder than -15°C are significantly glaciated. condensate supply rate and the bulk ice crystal mass Generally, more observations of the ice water con- growth rates. They demonstrated that in the presence tent of layer clouds are required to both test theories of weak updrafts, the small ice crystals at cloud top of ice crystal generation and to develop improved allow the existence of a liquid water layer in weak parameterizations for GCM and limited-area models. updrafts over a wide range of temperatures and ice There are two important gaps where there is a severe crystal concentrations. lack of data. First, there is a dearth of information on The sensitivity of radiation calculations in GCM layer clouds over midoceanic areas, and second, ob- and limited-area models to the presence of these liq- servations over continental areas in nonmountainous uid water layers are largely untested. Analyses of ad- regions are comparatively rare. ditional datasets are needed to verify the generality of these observations, as well as modeling studies to b. Ice crystal numbers in clouds calculate the lifetime and radiative response of the liq- Hobbs and Rangno (1985) made an extensive study uid water layers in the clouds. However, calculations of ice crystal numbers in stratiform cloud that formed by M. Piatt (1994, personal communication) suggest over Washington State, southern Utah, and the Pacific that for the shortwave albedo, a 100-m layer of small Ocean. Their extensive study included altocumulus, water drops is equivalent to 1 km of ice crystals. Con- altostratus, nimbostratus, stratocumulus, and stratus. sequently, if these liquid layers are ubiquitous they In this paper only the observations relating to the al- may be radiatively significant. tostratus and nimbostratus cloud systems are discussed In deep-layer clouds, observations suggest that (see Table 3 of their paper). Hobbs and Rangno con- when cloud tops are warmer than -15°C liquid water clude that there was no significant correlation between contents exceeding 0.2 g m-3 are found frequently in the maximum number of ice crystals and cloud top cloud (see, e.g., Borovikov et al. 1963). This is parti- temperature. However, their results suggest that there cularly true when the layer clouds are modified by oro- is a relation between the number of ice crystals and graphy. Relatively high liquid water contents are the width of the droplet spectrum; namely,

Bulletin of the American Meteorological Society 63

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC crystal number dependence on cloud top temperature for AWSE is more consistent with that of Hobbs and Rangno (1985) than those found by King (1982) in western Victoria and Mossop (1968) in New South Wales. The difference in the droplet spectra found by King (1982) over western Victoria and by J. Jensen (1994, personal communication) near the Aus- tralian Great Dividing Range are shown in Table 3. The western Vic- torian observations show very few droplets greater than 20 jum, a mean diameter of about 6 jum, and low liquid water contents in the cloud of less than 0.1 g nr3. By contrast Jensen's results show droplets greater than FIG. 5. Plot of the phase ratio (the proportion of water to ice) in cloud against 20 jum, a mean droplet diameter of temperature from 11 frontal flights. Each point represents the average over a 2-min 10 jum, and liquid water contents ex- horizontal leg in cloud. Crosses indicate clouds in continental airmasses and squares 3 indicate clouds in maritime airmasses. The dotted line is the best fit to the data for ceeding 0.1 g nr . The ice crystals continental clouds and the dashed line for maritime clouds. The solid line is the current form in synoptically similar systems parameterization in the U.K. Meteorological Organization Atmospheric GCM with the main difference being that (courtesy S. Moss and D. Johnson). the AWSE clouds formed in the airflow perturbed by the dividing range.

6.6 In the observing region for the AWSE, the orogra- Dr phy has a maximum peak of 1.2 km. This modest to- 1 = (i) 19A pographic barrier appears to generate substantial differences in the microphysical structure of the clouds. _1 where / is in L and Dr is the diameter in jum that Since most observations of the ice crystal numbers exceeds a cumulative concentration of 3 cm-3. The in precipitating layer clouds have been made near re- onset of ice particles in the stratiform cloud with tops gions of steep topography, there is a need to confirm colder than -6°C commences with the development their representativeness away from significant topog- of larger droplets and is independent of the cloud top raphy by comparing them with new observations that temperature. The cloud drop spectra of NIPAC are are either over the sea or over flat land. narrow, and the concentration of droplets greater than The observations by Bruintjes (1988) in South Af- 20 jum is less than 1 cm-3, whereas the high ice pro- rica are interesting because above the melting level ducers have concentrations of droplets with diameters in the prefrontal and warm frontal rainbands super- greater than 20 jum exceeding 10 cm-3. In other re- cooled liquid water is encountered on only 5% of oc- gions of the world, there is some evidence that in pre- casions. This is comparable with the observations in cipitating stratiform cloud systems there may be a re- Australia over western Victoria (King 1982). Near lationship between cloud top temperature and the cloud base the droplet concentration is typically number of ice crystals near cloud top. These data 100 cm-3 and the mean diameter of the droplets is come from observations in southeastern Australia 20 /im. On the other hand, the prefrontal cold surge (King 1982; Mossop 1968), Colorado (Cooper and rainbands were substantially more convective than the Saunders 1980), northern Spain (Vali et al. 1988), and prefrontal and warm frontal rainbands. In the convec- Israel (Gagin 1981), and are summarized in Fig. 6. tive elements, liquid water was present in new tow- Also shown in Fig. 6 are the results from Table 3 of ers and ice crystal concentrations between 10 and Hobbs and Rangno (1985) and observations from the 100 L_1 were found. However, in regions where the Australian Winter Storms Experiment (AWSE) convective activity was weak the ice particle concen- (J. Jensen 1994, personal communication). The ice trations were in the range 10-20 Lr1. The cloud-top

64 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC temperature is not given in the paper; however, the cloud top was estimated to be about 6 km. S. Moss and D. Johnson (1994, personal communication) find that during a continuous descent through thick layer clouds, three areas were encountered where ice particle concentrations were significantly higher than in adjacent regions. These occurred at temperatures of -15°C, -6°C, and near the freezing level. The dominant ice crystal habits at-6°C are small ice columns and aggregates, which Moss and Johnson hypothesize to be created by splintering. At -15°C the authors are less sure of the ice-multiplication mechanism. In summary, there is a need for observations that will test the generality of the empirical relation found by Hobbs and Rangno (1985) as well as a need to de- termine under what conditions there is a relationship between cloud top temperatures and ice crystal concentrations. However, the most pressing need is for a physically based ice-multiplication mechanism for frontal clouds. FIG. 6. Observations of ice crystal concentrations (no. L_1) as c. Ice crystal spectra a function of cloud top temperature. The ice nucleus spectrum is The second parameter required to calculate the pre- from Fletcher (1962), the Australian data from Mossop (1968) cipitation and radiative response of precipitating layer and King (1982), the Spanish data from Vali et al. (1988), and the Colorado data from Cooper and Saunders (1980) are shown cloud is the shape of the particle spectrum. M. Piatt as linear fits to the data. Observations from Hobbs and Rangno (1995, personal communication) examined the part- (1985) over the Pacific Northwest of the United States are shown icle size spectra from precipitating clouds provided by as open dots, and observations from the Australian Winter Storms A. Heymsfield. Some of the data from nonprecipi- Experiment (J. Jensen 1994, personal communication) are shown tating cirrus had been used previously by Heymsfield as closed dots. and Piatt (1984) in a study of the particle size spectrum of cirrus clouds. These authors argue that the particle size spectra Figure 7 is an example from M. Piatt (1995, personal in the range from 10 to about 100 jum obey a power communication) showing both the power-law segment law of the form and the exponential segment for particles collected between-25° and-30°C in the same cloud. Similarly

h shaped spectra have been reported by Paltridge et al. N(D) — aD , (2) (1986) in Australia, Houze et al. (1979) in the United States, and You et al. (1990) in China. However, the where N(D) is the number of ice crystals (nr3) and power-law component of the parameterization was not D is the largest dimension of the ice crystal in mi- derived by You et al. (1990). crometers. Heymsfield and Piatt found that "b" var- Many NWP and climate models parameterize the ied from -2.2 to -3.8 over the temperature range - ice phase in the microphysics in terms of precipita- 25° to 60°C. tion and cloud-sized ice crystals. Based on Piatt's ob- M. Piatt (1995, personal communication) found servations, there is a good case for parameterizing pre- similar power laws for precipitating clouds but, for cipitation-sized particles in terms of the Marshall- particle sizes in excess of 100-200 /um, the size dis- Palmer distribution, while parameterizing the smaller tribution follows that of a Marshall-Palmer distribu- cloud-sized particles using the power law. The prob- tion more closely; namely, lem with this formulation is that there is no physically based "autoconversion" parameterization that allows

N(D) = N0 exp(-AD). (3) for generation of precipitation-sized ice particles from

Bulletin of the American Meteorological Society 65

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC FIG. 8. The observed exponent / (m ') for the Marshall-Palmer distribution in clouds as function of temperature. HHHP is from Washington State (Houze et al. 1979); SMPC from California (Stewart et al. 1984), as is GM (Gordon and Marwitz 1986); Piatt from Heymsfield and Piatt (1984); AWSE from Australia (J. Jensen 1994, personal communication); YL from China (You FIG. 7. Observations of the observed particle concentrations and Liu 1989); and from Europe. Bennetts and Ryder (1984) is from a layer cloud in the Heymsfield and Piatt (1984) dataset shown as (B), and Marecal (1993) is shown as (M). (courtesy C.M.R. Piatt). the ice cloud. This is clearly a problem that needs to of the datasets and therefore can be parameterized be addressed. independent of the geographic region. The functional

The Marshall-Palmer component of the distribu- relation between N0 and temperature is less system- tion for precipitating clouds has been studied exten- atic than that for X with local variations being evident. sively in a number of regions of the world, and a consistent picture has evolved. Figure 8 shows examples of the exponent A as a function of temperature for 6. Summary Washington State (Houze et al. 1979), California (Stewart et al. 1984; Gordon and Marwitz 1986), The results of the survey illustrate quantitatively Australia (J. Jensen 1994, personal communication), that the synoptic and mesoscale characteristics of pre- China (You and Liu 1989), and Europe (Bennetts and cipitating layer clouds around the world are not Ryder 1984; Marecal et al. 1993), and Fig. 9 shows strongly dependent on the geographic location. How- examples of the value of N0 as a function of tempera- ever, despite the large number of field experiments ture for Washington State, California, Australia, and undertaken in recent years, it is clear that these con- Europe. In their analysis, Gordon and Marwitz (1986) clusions are biased to observations taken near moun- identified a change in the slope of A with particular tainous terrain or near the coast. There are almost no ice crystal growth regimes. Apart from these subtle observations over the open oceans and too few over changes, A is a simple function of temperature in all relatively flat terrain.

66 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC The simple conceptual conveyor belt model of frontal systems gives a clear description of the flows generating the layer cloud systems on the synoptic scale. Some global models and all limited-area models have the resolution to diagnose the flows in the conceptual model. The efficiency and intensity of rainbands generated by midlatitude cloud systems may be modulated by evaporation and melting in the subcloud layer and by orographic and coastal forcing. It is doubtful if the current generation of general circulation models has either the physics or resolution to be able to model these processes adequately. The microphysical structure of the precipitating clouds is normally dominated by the ice phase in the clouds. The lifetime of liquid water above the melting layer is small away from regions of enhanced convection and orographic uplift. Observations suggest that for ice particles with sizes between 10 and 100 /im the distribution is best described by a power law and for sizes above 200 /im the ice crystal size distribution is best described by a Marshall-Palmer distribution. This survey would suggest that the fundamental knowledge needed to parameterize the number of ice -3 crystals in clouds is still incomplete. FIG. 9. The observed constant(m ) for the Marshall-Palmer distribution in clouds as a function of temperature. The symbols A systematic database for precipitating layer are as in Fig. 8. clouds is required and this is a priority for the GEWEX Cloud System Study Working Group III (Browning 1994). A panel report from a recent World CAO Central Aerological Laboratory Meteorological Organization (WMO) Workshop on Russia, 1977-1979 (Postnov 1983) Cloud Microphysics and Applications to Global CFRP Cold Fronts Research Program, Change (WMO 1992) recommended that WMO Southeastern Australia, 1980-1984 should keep a central inventory of projects and (Ryan et al. 1985b) datasets that are relevant to the effects of cloud mi- CASP I & II Canadian Atlantic Storms Program, crophysics on climate change. The report emphasizes 1986 and 1992, Western Atlantic a number of problems that affect the usefulness and (Stewart et al. 1987; Stewart 1991) compatibility of datasets. ERICA Experiment on Rapidly Intensifying This paper has identified two requirements for a Cyclones over the Atlantic, data blueprint on precipitating layer clouds, namely Western Atlantic, 1986-1991 to provide 1) specifications for datasets used to vali- (Hadlock and Kreizberg 1988) date cloud climatologies generated by climate mod- GALE Genesis of Atlantic Lows Experiment els and 2) the specifications for datasets used in pro- East Coast United States, 1986 cess studies. (Dirks et al. 1988) WISP Winter Icing and Storms Project, Colorado Front Range, United States, Appendix A: Major meteorological 1990-1991 (Rasmussen et al. 1992) observing programs BASE Beaufort and Arctic Storms Experiment (R. Stewart 1994, personal CYCLES CYCLonic Extratropical Storms communication) Project, Pacific Northwest, United FRONTS-87 Eastern Atlantic, 1987 States (Hobbs et al. 1980) (Clough and Testud 1988)

Bulletin of the American Meteorological Society 67

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC Appendix B: Major cloud seeding studies Belyakov, I. E., N. A. Bezrukova, S. N. Burkovskaya, A. A. Postnov, V. I. Silayeva, T. V. Trutko, and V. M. Vostrenkov, ISRAEL I & II Israel, 1961-1967 and 1967-1975 1984: Mesoscale structure of atmospheric fronts and associated cloud precipitation systems over European USSR. Proc. (Gagin 1981) Ninth Int. Conf. on Cloud Phys., Tallinn, USSR, ICCP/ CASCADE Pacific Northwest, United States, IAMAP, 339-342. 1972-1973 (Hobbs 1975) Bennetts, D. A., and P. Ryder, 1984: A study of mesoscale CRBPP Colorado River Basin Pilot Project, convective bands behind a cold front. Part II: Cloud and Colorado, 1974-1975 microphysical structure. Quart. J. Roy. Meteor. Soc., 110, 467-487. (Marwitz 1980) Bezrukova, N. A., and S. M. Shmeter, 1989: Cloud and precipi- SCPP Sierra Cooperative Pilot Project, tation structures in frontal zones. Proc. Fifth WMO Scientific Nevada, 1979-1980 (Heggli et al. Conf on Weather Modification and Applied Cloud Physics, 1983) Beijing, China, World Meteorological Organization, 471^74. WVCSE Western Victorian Cloud Seeding Borovikov, A. M., L. I. Gaivoronskii, E. G. Zak, V. V. Kostarev, I. P. Mazin, V. E. Minervin, A. Kh. Khgain, and S. M. Shmeter, Experiment, Southeastern Australia, 1963: Cloud Physics. Israel Program for Scientific Transla- 1978-1980 (King 1982) tions, 392 pp. COSE Colorado Orographic Cloud Seeding Browning, K. A., 1985: Conceptual models of precipitating sys- Experiment, Colorado, 1979-1981 tems. Meteor. Mag., 114, 293-316. (Rauber and Grant 1986) , 1990: Organization of clouds and precipitation in extrat- TAS II Tasmania II, Tasmania, Australia, ropical cyclones. Extratropical Cyclones: The Erik Palmen Memorial Volume. C. Newton and E. O. Holopainen, Eds., 1979-1982 (Shaw et al. 1984) Amer. Meteor. Soc., 129-153. PEP Precipitation Enhancement Project, , 1994: GEWEX Cloud System Study (GCSS): Science Plan. 1979-1984, Villanubla, Spain IGPO Publication Series, Vol. 11, 62 pp. (Vali et al. 1988) , and G. A. Monk, 1982: A simple model for the synoptic analy- X'JIANG Xinjiang Experiment, 1982-1984, sis of cold fronts. Quart. J. Roy. Meteor. Soc., 108, 435^52. Urumiq Region, China Bruintjes, R. T., 1988: The occurrence of ice in cloud associated with extratropical cyclones. Ann. Meteor., 25, 547-550. (You et al. 1990) Carlson, T. N., 1980: Airflow through mid-latitude cyclones and AME Atlas Mountains Experiment, 1984, the comma cloud pattern. Mon. Wea. Rev., 108, 1498-1509. Morocco (Baddour and Rasmussen , 1991: Midlatitude Weather Systems. Harper-Collins/Aca- 1989) demic Press, 507 pp. AWSE Australian Winter Storms Experiment, Clough, S. A., and J. Testud, 1988: The Fronts-87 Experiment and Mesoscale Frontal Dynamics Project. WMO Bull., 37,276-281. 1988-1992, Southeastern Australia , and R. A. A. Franks, 1991: The evaporation of frontal and (Long and Huggins 1992) other stratiform precipitation. Quart. J. Roy. Meteor. Soc., 117, 1057-1080. Cooper, W. A., and C. P. R. Saunders, 1980: Winter storms over Acknowledgments. The paper is a contribution from the GCSS the San Juan Mountains. Part II: Microphysical processes. Working Group on Layer Clouds to the CCSS Science Panel and J. Appl. Meteor., 19, 127-941. was presented at a meeting in Victoria, Canada, in December , and G. Vali, 1981: The origin of ice in mountain cap clouds. 1994. I would like to express my gratitude to Jorgen Jensen, J. Atmos. Sci., 38, 1244-1259. Martin Piatt, Sarah Moss, and Doug Johnson for giving me ac- Cotton, W. R., and R. A. Anthes, 1989: Storm and Cloud Dynam- cess to unpublished data. I would all so like to thank Ron Stewart, ics. International Geophysics Series, Vol. 44, Academic Press, Sid Clough, Jack Katzfey, and John Garratt for their comments 883 pp. on this paper. Dirks, R. A., J. P. Kuettener, and A. J. Moore, 1988: Genesis of Atlantic Lows Experiment (GALE): An overview. Bull. Amer. Meteor. Soc., 69, 148-172. References Dodge, P. D., and R. W. Burpee, 1993: Characteristics of rainbands, radar echoes, and lightning near the North Caro- Baddour, O., and R. M. Rasmussen, 1989: Microphysical obser- lina coast during GALE. Mon. Wea. Rev., 121, 1936-1955. vations in winter storms over the Atlas Mountains in Morocco. Douglas, R. H., K. L. S. Gunn, and J. S. Marshall, 1957: Pattern Atmos. Res., 24, 103-122. in the vertical of snow generation. J. Meteor., 14, 95-114. , M. Nbou, M. El Mokhtari, R. M. Rasmussen, D. B. Egger, J., and K. P. Hoinka, 1992: Fronts and orography. Meteor. Johnsson, and D. A. Mathews, 1989: Seedability of Moroc- Atmos. Phys., 48, 3-36. can clouds. Proc. Fifth Scientific Conf on Weather Modifi- Elliot, R. D., and E. L. Hovind, 1964: On convection bands within cation and Applied Cloud Physics, Beijing, China, World Pacific coast storms and their relation to storm structure. Meteorological Organization, 77-80. J. Appl. Meteor., 3, 143-154.

68 Vol. 77, No. 1, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC Fletcher, N. H., 1962: The Physics of Rainclouds. Cambridge , J. D. Locatelli, and J. E. Martin, 1990: Cold fronts aloft University Press, 386 pp. and the forecasting of precipitation and severe weather east Gagin, A., 1975: The ice phase in winter continental cumulus of the Rocky Mountains. Wea. Forecasting, 5, 613-626. clouds. J. Atmos. Sci., 32, 1604-1614. Hoinka, K. P., and H. Volkert, 1987: The German Front Experi- , 1981: The Israeli rainfall enhancement experiment: A ment. Bull. Amer. Meteor. Soc., 68, 1424-1427. physical overview. J. Wea. Mod., 13, 103-120. Houze, R. A., 1981: Structures of atmospheric precipitation sys- , and J. Neumann, 1974: Rain stimulation and cloud phys- tems: A global survey. Radio Sci., 16, 671-689. ics in Israel. Weather and Climate Modification. W. N. Hess, , and P. V. Hobbs, 1982: Organization and structure of pre- Ed., Wiley, 842 pp. cipitating cloud systems. Advances in Geophysics, Vol. 24, Gao, Z., and J. Zhang, 1988: The basic characteristics of the winter Academic Press, 225-315. northern Xinjiang stratiform cloud system. Proc. of 10th Int. , , and P. H. Herzegh, 1979: Size distribution of precipi- Cloud Phys. Conf, Bad Homburg, Germany, ICCP/IAMAP, tation particles in frontal clouds. J. Atmos. Sci., 36, 156-162. 389-392. , S. A. Rutledge, T. J. Matejka, and P. V. Hobbs, 1981: The Garratt, J. R., 1988: Summertime cold fronts in southeast Aus- mesoscale and microscale structure and organization of clouds tralia—behaviour and low-level structure of main frontal and precipitation in midlatitude cyclones. Ill: Air motions and types. Mon. Wea. Rev., 116, 636-649. precipitation growth in a warm-frontal rainband. J. Atmos. Sci., Gordon, G. L., and J. D. Marwitz, 1986: Hydrometeor evolution 38, 639-649. in rainbands over the California valley. J. Atmos. Sci., 43, Isaac, G. A., 1991: Microphysical characteristics of Canadian 1087-1100. Atlantic storms. Atmos. Res., 26, 339-360. Goyer, G. G., and R. M. Cunningham, 1981: PEP site-selection King, W. D., 1982: Location and extent of supercooled regions Phase 3 1981 field season. Data catalogue, weather conditions in deep stratiform cloud in western Victoria. Aust. Meteor. and cloud structure. PEP Report 26, WMO, 76 pp. Mag., 30, 81-88. Hadlock, R., and C. W. Kreitzberg, 1988: The Experiment on , M. J. Manton, D. E. Shaw, E. J. Smith, and J. Warner, 1979: Rapidly Intensifying Cyclones over the Atlantic (ERICA) field Prospectus for a cloud-seeding experiment in western Victoria. study: Objectives and plans. Bull. Amer. Meteor. Soc., 69, CSIRO Division of Cloud Physics, CP222, 48 pp. [Available 1309-1320. from Librarian, CSIRO Division of Atmospheric Research, Harrold, T. W., 1973: Mechanisms influencing the distribution PMB1, Aspendale, Victoria 3195, Australia.] of precipitation within baroclinic disturbances. Quart. J. Roy. Levin, Z., 1994: Aerosol composition and its effect on cloud Meteor. Soc., 92, 232-251. growth and cloud seeding. Proc. of the Sixth WMO Scientific Heggli, M. F., L. Vardiman, R. E. Stewart, and A. Huggins, 1983: Conf. on Weather Modification, Siena, Italy, World Meteo- Supercooled liquid water and ice crystal distributions within rological Organization, 367-369. Sierra Nevada winter storms. J. Climate Appl. Meteor., 21, Long, A. B., and A. W. Huggins, 1992: Australian Winter Storms 1875-1886. Experiment (AWSE) I: Supercooled liquid water and precipi- Heymsfield, A. J., 1977: Precipitation development in ice strati- tation enhancement opportunities./. Appl. Meteor., 31,1041— form clouds: A microphysical and dynamical study. J. Atmos. 1055. Sci., 34, 367-381. Marecal, V., D. Hauser, and C. Duroure, 1993: Airborne micro- , 1993: Structures of clouds. Aerosol-Cloud-Climate Inter- physical measurements and radar reflectivity observations actions, International Geophysical Series, P. V. Hobbs, Eds., near a cold frontal rainband observed during the FRONTS 87 Academic Press, 97-121. experiment. Atmos. Res., 29, 170-207. , and C. M. R. Piatt, 1984: A parameterization of the Marwitz, J. D., 1980: Winter storms over the San Juan Mountains. particle size spectrum of ice clouds in terms of ambient Part I: Dynamical process. J. Appl. Meteor., 19, 913-926. temperature and ice water content. J. Atmos. Sci., 41, Mass, C. F., and D. M. Schultz, 1993: The structure and evolu- 846-855. tion of a simulated midlatitude cyclone over land. Mon. Wea. , L. M. Miloshevich, A. Slingo, K. Sassen, and D. O'C. Starr, Rev., 121, 889-917. 1991: An observational and theoretical study of highly super- Matekjka, T. J., R. A. Houze, and P. V. Hobbs, 1980: Microphys- cooled altstratus. J. Atmos. Sci., 48, 923-945. ics and dynamics of clouds associated with mesoscale Hobbs, P. V., 1975: The nature of winter clouds and precip- rainbands in extratropical cyclones. Quart. J. Roy. Meteor. itation in the Cascade Mountains and their modification by ar- Soc., 106, 29-56. tificial seeding. Part I: Natural conditions. J. Appl. Meteor., Matkovskii, B. M., and N. P. Shakina, 1982: Mesoscale struc- 14, 783-804. ture of an occluded front over the center of the European USSR , 1978: Organization and structure of clouds and precipita- from special measurements. Meteor. Gidrol., 1, 24-33. tion on the mesoscale and microscale in cyclonic storms. Rev. May, P. T., K. J. Wilson, and B. F. Ryan, 1991: VHF radar stud- Geophys. Space Phys., 16, 741-755. ies of cold fronts passing across southern Australia. Beitr. , and A. L. Rangno, 1985: Ice particle concentrations in Phys. Atmos., 63, 257-269. clouds. J. Atmos. Sci., 42, 2523-2549. Mossop, S. C., 1968: Comparison between concentration of ice , T. J. Matejka, P. H. Herzegh, J. D. Locatelli, and R. A. crystals in cloud and concentrations of ice nuclei. J. Rech. Houze, 1980: The mesoscale and microscale structure and or- Atmos., 3, 119-124. ganization of clouds and precipitation and midlatitude Murakami, M., T. Matsou, H. Mizuno, and Y. Yamada, 1988: Mi- cyclones. I: A case study of a cold front. J. Atmos. Sci., 37, crophysical structures of Japanese warm-frontal clouds observed 568-596. with a new type of sonde (HYVIS). Ann. Meteor., 25,521-524.

Bulletin of the American Meteorological Society 69

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC , Y. Yamada, T. Matsuo, H. Mizuno, and K. Morikawa, Stephens, G. L., and T. J. Greenwald, 1991: The Earth's radia- 1992: Microphysical structures of warm-frontal clouds—the tion budget and its relation to atmospheric hydrology 2. Ob- 20 June 1987 case study. J. Meteor. Soc. Japan, 70, 877-894. served cloud effects. J. Geophys. Res., 96, 15 325-15 340. Nozumi, Y., and H. Arakawa, 1968: Prefrontal rainbands located Stewart, R. E., 1991: Canadian Atlantic Storms Program: Progress in the warm sector of subtropical cyclones over the ocean. and plans of the meteorological component. Bull. Amer. Me- J. Geophys. Res., 73, 487-492. teor. Soc., 68, 346-371. Paltridge, G. W., W. D. King, and C. M. R. Piatt, 1986: A case , 1992: Precipitation types in the transition region of winter study of ice particle growth in a mixed-phase altostratus cloud. storms. Bull. Amer. Meteor. Soc., 73, 287-296. Aust. Meteor. Mag., 34, 149-154. , 1994: Critical processes within and attributes of frontal Plank, V. G., D. Atlas, and W. H. Paulsen, 1955: The nature and cloud. ECMWF/GCSS Workshop on Modelling and Assimi- detectability of clouds and precipitation as determined by 1.25 lation of Clouds, Reading, United Kingdom, ECMWF, in centimeter radar. J. Meteor., 12, 358-378. press. Postnov, A. A., 1983: The mesoscale wind structure in warm , J. D. Marwitz, J. C. Pace, and R. E. Carbone, 1984: Char- fronts over the European part of the USSR. Meteor. Gidrol., acteristics through the melting layer of stratiform clouds. 2, 25-31. J. Atmos. Sci., 22, 3227-3237. Rangno, A. L., and P. V. Hobbs, 1988: Criteria for the onset of ——, R. W. Shaw, and G. A. Isaac, 1987: Canadian Atlantic significant concentrations of ice particles in cumulus clouds. Storms Program: The meteorological field project. Bull. Amer. Atmos. Res., 22, 1-13. Meteor. Soc., 68, 338-345. Rasmussen, R., and Coauthors, 1992: Winter Icing and Storms , S. R. Macpherson, N. R. Donaldson, and C. A. Lin, 1988: Project (WISP). Bull Amer. Meteor. Soc., 73, 951-974. Melting-induced circulations, frontogenesis and cyclogenesis. Rauber, R. M., and L. O. Grant, 1986: The characteristics and Palmen Memorial Symposium on Extratropical Cyclones. distribution of cloud water over the mountains of north Colo- Helsinki, Finland, Amer. Meteor. Soc., 250-253. rado during wintertime storms. Part II: Spatial distribution and Vali, G., 1980: Cloud microphysics studies at the Spanish site microphysical characteristics. J. Climate Appl. Meteor., 25, for the Precipitation Enhancement Project. The Third WMO 489-509. Scientific Conf. on Weather Modification, Geneva, Switzer- , and A. Tokay, 1991: An explanation for the existence of land, World Meteorological Organization, 259-264. supercooled water at the top of cold clouds. J. Atmos. Sci., 48, , R. Koenig, and T. Yoksas, 1988: Estimation of precipita- 1005-1023. tion enhancement potential for the Duero Basin of Spain. Ryan, B. F., K. J. Wilson, J. R. Garratt, and R. K. Smith, 1985a: J. Appl. Meteor., 27, 829-850. The Australian summertime cool change: III. Subsynoptic and Wang, C., Z. Hu, and L. You, 1992: A case study of frontal cloud mesoscale model. Mon. Wea. Rev., 113, 224-240. precipitation system over North China. Proc. WMO Work- , , , and , 1985b: Cold Fronts Research shop on Cloud Microphysics and Applications to Global Programme: Progress, future plans and research directions. Change. Toronto, Canada, World Meteorological Organiza- Bull. Amer. Meteor. Soc., 66, 1116-1122. tion, 221-228. , , and E. J. Zipser, 1989: Modification of the thermo- Warren, S. G., C. J. Hahn, J. London, R. M. Chervin, and R. Jenne, dynamic structure of the lower troposphere by the evapora- 1986: Global distribution of total cloud cover and cloud type tion of precipitation ahead of a cold front. Mon. Wea. Rev., amount over land. NCAR Tech. Note NCAR/TN-273+STR, 117, 138-153. 29 pp. plus 200 maps. , G. J. Tripoli, and W. R. Cotton, 1990: Convection in high , , , , and , 1988: Global distribution of stratiform cloud bands: Some numerical experiments. Quart. total cloud cover and cloud type amounts over ocean. NCAR J. Roy. Meteor. Soc., 116, 943-964. Tech. Note NCAR/TN-0317+STR, 42 pp. plus 170 maps. Saarikivi, P., 1983: Radar observations of snowbands in Fin- Wexler, R., and D. Atlas, 1959: Precipitation generating cells. land—a preliminary report. Mesoscale Meteorology—Theo- J. Meteor., 16, 327-332. ries, Observations and Model, D. K. Lilly and T. Gal-Chen, WMO, 1992: Procedings, WMO workshop on cloud microphys- Eds., D. Reidel, 282-292. ics and applications to global change. WMP Report 19. , 1989: Characteristics of mesoscale precipitation bands in You, L., and Y. Liu, 1989: The synoptic structures and micro- southern Finland. Mon. Wea. Rev., 117, 2584-2593. physical processes of precipitation systems in winter in Shaw, D. E., W. D. King, and E. Turton, 1984: Analysis of Hy- Wulumuqi area, Xingjang. The Fifth WMO Science Conf on dro-Electric Commission cloud seeding in Tasmania, 1979— Weather Modification and Applied Cloud Physics, Beijing, 83. Report to Hydro-Electric Commission of Tasmania, Cloud China, World Meteorological Organization, 65-68. Physics Laboratory, 33 pp. , S. Wang, D. Wang, and Y. Chi, 1990: The microphysical Shmeter, S. M., 1992: Some peculiarities of the mesostructurel structure of snow clouds and the growth process of snow par- of frontal clouds field. Proc. 11th Int. Conf. on Clouds and ticles in winter Xinjiang. Acta Meteor. Sin., 4, 222-230. Precipitation, Montreal, Canada, ICCP/IAMAP, 601-604.

^ 4

70 Vol. 77, No. I, January 1996

Unauthenticated | Downloaded 10/05/21 05:50 PM UTC