Houze Robert A., Jr. Cloud Dynamics

Подождите немного. Документ загружается.

6.1 Stratiform Precipitation 199

tion and riming (Sec. 3.2.4) can occur. Most importantly, the particles begin to

aggregate and form large, irregularly shaped snowflakes. The particles may also

grow by riming, since at these warmer levels the vertical air motions of a few tens

of centimeters per second are sometimes strong enough to maintain a small num-

ber of liquid-water drops in the presence of the falling ice particles. As noted in

Sec. 3.2.4, the aggregation becomes more frequent within about 1 km of the

O°C

level. Aggregation, of course, does not add mass to the precipitation but rather

concentrates the condensate into large particles, which, upon melting, become

large, rapidly falling raindrops. As will be explained below, the layer in which the

large snowflakes melt is marked on radar by a

bright

band

of intense echo in a

horizontal layer about 1/2 km thick located

just

below the

O°C

level. The bright

band produced by melting particles and other characteristics of the vertical struc-

ture of the radar echo in stratiform precipitation are particularly informative about

the dominant precipitation mechanisms and will be discussed in more detail in

Sec. 6.1.2.

The convective precipitation process is depicted in Fig. 6.1b.

differs sharply

from stratiform precipitaiton.

is defined as a precipitation process in which

condition (6.1) is not met. Instead it has a vertical air velocity scale of

w -

I-10m

S-I,

which equals or exceeds the typical fall speeds of ice crystals and snow.

is observed that the time available for the growth of precipitation particles in

convective precipitation is limited; often rain reaches the ground within a half an

hour of cloud formation. This time is much shorter than the 1-3 h available for

growth of precipitation particles in stratiform precipitation.

159 Since the time is so

short, the precipitation particles must originate and grow not far above cloud base

at the time the cloud forms (time

to in Fig. 6.1b).

is possible for the growth to

begin at that time since updrafts are strong enough to carry the growing particles

upward until they become heavy enough to overcome the updraft and begin to fall

relative to the ground (see the particle trajectory in Fig. 6.1b). The only micro-

physical growth mechanism rapid enough to allow the particles to develop this

quickly is accretion of liquid water. In contrast, accretion of liquid water is at

most of minor importance in stratiform precipitation, where the dominant micro-

physical mechanisms are vapor deposition and aggregation of ice particles. Since

the strong convective updrafts carrying the particles upward during the growth

phase of the cloud condense large amounts of liquid water, the larger particles

(whether they be liquid or ice) in the rising parcels of air can grow readily by

159 The idea of distinguishing convective from stratiform precipitation on the basis of the magnitude

of the in-cloud vertical air motion and time scale of the microphysical precipitation growth process was

suggested by the pioneering cloud physicist Henry Houghton in one of his last papers

(1968).

official definition of the term stratiform precipitation does not exist; it is absent from the

Glossary

Meteorology (Huschke, 1959). Houghton's distinction, however, captures the physical essence of the

two precipitation types. Further aspects of the difference were articulated by the early radar meteorol-

ogist Louis Battan in his textbooks on meteorological radar (1959, 1973). He denoted the two types of

precipitation as

"convective"

and

"continuous,"

the latter corresponding to our stratiform precipita-

tion. In describing continuous precipitation, he emphasized the radar bright band. the existence of

embedded nonuniform echo structure in the continuous precipitation, and how convective echoes take

on a bright band structure in their late stages. The views depicted in Figs.

6.la

and b are largely a

combination of those of Houghton and Battan.

200 6 Nimbostratus

accretion of cloud liquid water. Aircraft observations of ice particles in convective

clouds confirm that

the .particles in the updrafts grow largely by riming. Since the

strong updrafts in convective clouds are usually narrow (typically

-1

km or less in

width), radar echoes from precipitation associated with active convection form

well-defined vertical cores of maximum reflectivity, which contrast markedly with

the horizontal orientation of the radar bright band seen at the melting level in

stratiform precipitation (compare the reflectivity patterns in Figs. 6.1a and b).

In the dissipating stages of precipitating convective clouds (after time

ts in Fig.

6.lb), strong upward motions cease and no longer carry precipitation particles

upward or suspend them aloft.

The

fallout of the particles left aloft by the dying

updrafts can take on a stratiform character, including a radar bright band.

6.1.2 Radar-Echo Structure

The schematic in Fig. 6.1a depicts qualitatively the radar-echo structure of strati-

form precipitation, with the melting layer denoted by the radar bright band being

the most prominent feature. Closer, quantitative examination of the vertical pro-

file of radar data in stratiform precipitation allows us to divide the precipitation

into distinct layers in which different microphysical processes dominate. These

layers are bounded by points

0-4

in Fig. 6.2.

The zone from 0 to 1 in Fig. 6.2 is associated primarily with ice particles

growing by

vapor deposition, which is the slowest microphysical growth mode.

Since the ice particles settling downward through the nimbostratus are not grow-

ing rapidly, the equivalent radar reflectivity

Ze,

which is proportional to the sixth

moment of the particle size distribution [recall Eqs. (4.3) and (4.8)], does not

increase rapidly with decreasing altitude between 0 and 1. In the layer between 1

VR '" Vif radar pointing vertically

dBZ

dBZ _

Figure 6.2 Schematic of vertical profile of radar data in stratiform precipitation. Solid curve

shows reflectivity. Dashed curve shows Doppler radial velocity

with the antenna at vertical

incidence. Under stratiform conditions V

is related approximately to the mass-weighted terminal fall

speed of the particles,

Layers in which different microphysical processes dominate are bounded by

points

0-4

(see text for further discussion).

6.1 Stratiform Precipitation

201

and 2, the particles continue to grow by deposition and possibly some riming.

However, they also undergo aggregation with increasing frequency as they ap-

proach the melting layer. The aggregation has the effect of producing very large

particles. Since

depends on the sixth power of the particle dimension, the

formation of these large particles sharply increases the radar reflectivity between

levels 1 and 2.

The changes in the radar reflectivity profile between levels 2 (the

O°C

level) and

4 are all associated with

melting. The center

ofthis

layer (level 3) is marked by a

sharp maximum

Ze,

which gives the melting layer its identity as the bright band.

Several processes strongly affect the radar reflectivity profile in the melting layer.

The sharp increase in the

radar

reflectivity downward from level 2 to 3 is thought

to be the result of two effects. First, aggregation continues to occur, as in the layer

from 1 to 2. This conclusion is arrived at inductively, since other effects cannot

fully explain the increase in reflectivity factor from point 2 to 3. The second effect

that appears to be significant is that the magnitude of the complex index of

refraction of the particles

[IKI2

in (4.6)] changes as they melt from 0.197 (for ice

particles) to

0.93 (for liquid water drops). If in the first stage of melting, the

particles take on the character of water but do not collapse to form smaller drops

until the end of the melting process, then Z, which is proportional to the sixth

moment of the particle size distribution [Eq.

(4.3)], remains constant. However,

Z, increases by a factor of

0.93/0.197 = 5 since, according to (4.3) and (4.6), Z, =

(IKI2/0.93)Z. As noted above, this amount of increase in

between points 2 and 3

is insufficient by itself to explain the magnitude of the peak of reflectivity in the

bright band, and it is for this reason that it is thought that the aggregation acts in

concert with the index of refraction change to produce the peak at 3.

The

sharp dropoff

Ze in the lower portion of the melting layer between levels

3 and 4 is produced by two effects. If the melting is completed at point 3 and all the

particles collapse to form small raindrops, two things happen to

Ze.

First, the

particles are now all smaller, and

is decreased in accordance with the sixth

power

all the particle diameters. Second, the fall speed of the particles suddenly

increases from

~1-3

S-I

for snow to

-5-10

S-I

for raindrops.

the down-

ward flux

precipitation mass is the same at levels 3 and 4, as it would be in

steady stratiform rain, then the mean concentration of rainwater (mass of water

per

volume

air) must decrease sharply from level 3 to level 4. Since the concen-

tration

rainwater mass is the third moment of the drop size distribution and Z, is

proportional to the sixth moment of the distribution [recall

(4.3) and (4.10)], the

decrease in mass concentration between levels 3 and 4 corresponds to a decrease

through the same layer.

The layer below point 4 in Fig. 6.2 is characterized by rain. The microphysical

processes that

can

occur

in this lowest layer are quite varied and depend on the

meteorological context. In some cases, the precipitation particles in this lower

region simply fall to the ground at a constant rate. In other cases, the rain falls

through a lower layer of cloud being continually regenerated by upward air mo-

tion, and the raindrops falling below the melting layer continue to grow by vapor

diffusion and collection of cloud droplets. In still other cases, the rain falling out of

202 6 Nimbostratus

the melting layer falls into a layer of dry air into which the drops partially evapo-

rate. Thus, the flux of rain reaching the surface is less than that exiting the melting

layer, and the air below the melting layer is cooled by the evaporation. The

cooling of the air both by this evaporation and by the melting can have important

feedbacks to cloud and storm dynamics, as we will see in later chapters.

6.1.3 Microphysical Observations

Observations of the sizes and shapes of precipitating ice particles above the ooe

level have been made aboard research aircraft flying in nimbostratus associated

with mesoscale convective systems, fronts, and hurricanes. These in situ mea-

surements verify the microphysical layering of the cloud that we have inferred

from the radar echo profile depicted in Fig. 6.2. An example of data obtained in

nimbostratus associated with tropical mesoscale convective systems is shown in

Fig. 6.3. Particles were observed at flight levels ranging in temperature from

-23°

+4°e.

Two-dimensional images in the form of shadows were formed by the ice

particles as they flowed through a laser-illuminated space under the wing of the

aircraft. The crystalline structures of most of the images seen by this device are

impossible to determine. However, occasionally a definable particle shape is

noted. The frequency with which particles of various types could be recognized

was noted and plotted as a function of flight-level temperature.

The particle types indicated in Fig. 6.3a and b (needles, columns, plates, and

dendrites) are known to grow at certain ambient temperatures (Table 3.1). If the

observed particles fell about 1/2-1 km from their altitude of growth before being

encountered by the

aircraft.l'" then their growth temperature

)

would be ex-

pected to be -

3-6°e

lower than the flight-level temperatures indicated in Fig. 6.3.

Taking this difference into account, we see that the maxima and minima of the

curves in Fig. 6.3a and b for needles, columns, plates, and dendrites all

occur

flight-level temperatures that are consistent with their respective

indicated in

Table 3.1. This result indicates that, as we concluded from the radar-echo profile

shown in Fig. 6.2, growth of ice particles by vapor diffusion occurs above the ooe

level. Some riming could also have been occurring, as riming would be consistent

with the unidentifiable shapes

many of the particle images. No attempt was

made, however, to determine whether the irregularities in particle image shape

were caused by riming. In Fig. 6.3c, it can be seen that the frequency of large

aggregates of crystals exhibits a broad maximum strongly overlapping with the

peak of frequency of dendritic crystals seen in Fig. 6.3b. This result is consistent

with aggregation becoming more frequent as the downward-settling crystals ap-

proach the ooe level. In addition, it appears that a primary aggregation mecha-

nism, at least in the tropical nimbostratus represented by these particular data, is

the arms of the dendrites becoming entangled as the ice crystals drift downward.

The nearly round particles indicated in Fig. 6.3c increase in frequency suddenly as

the ice particles melt and form raindrops below the ooe level.

The microphysical picture of nimbostratus indicated by the in situ data summa-

160 This amount of downward displacement would require only 5-15 min at fall speeds of 1-3 m

S-I,

6.1

Stratiform Precipitation

203

(a)

(b)

(c)

-25

---

::J

Columns

i -20

-20

-15

}

,...

-10

Needles

-5

...J

I-

:I:

0 2

3 4

5 6

7 8

0 2

3 4 5

12 16 20

24 28

PARTICLE FREQUENCY (min-

Figure 6.3 lee particle data obtained on aircraft flights through nimbostratus in tropical mesoscale

convective systems over the Bay of Bengal. Plots show relative frequency of observation of ice

particles of a particular type per minute of in-cloud flight time as a function of flight-level temperature.

(From Houze and Churchill, 1987. Reproduced with permission from the American Meteorological

Society.)

rized in Fig. 6.3 is the same as that inferred from the vertical profile of radar

reflectivity in Fig. 6.2. The region above the melting layer is characterized by

particles growing by vapor diffusion and drifting downward. Some riming could

occur along with the vapor diffusion, if the vertical motion becomes strong enough

to condense enough water to maintain the air at saturation with respect to liquid

water. Riming is most likely to occur

just

above the

O°C

level, since the vapor

content of the air increases exponentially with decreasing altitude. Although the

vertical air motions might become large enough to allow some riming to occur

along with the vapor diffusion, they must remain weak enough to allow the parti-

cles to continue to fall [i.e., condition (6.1) must remain satisfied to explain the

observations]. As the particles settle downward, aggregation produces large ice

particles within

-2-2.5

km of the top of the melting layer, with the most aggre-

gates being found within 1 km of the

O°C

level. What this structure implies about

the dynamics of the nimbostratus is that there must be widespread gentle uplift of

the air throughout the region

ofthe

cloud above the

O°C

level. This ascent must be

strong enough to supply vigorous growth by vapor diffusion (and possibly some

riming), but weak enough to allow the sedimentation of the ice particles. Since

these particles fall at speeds

1 m

S-I,

it is readily concluded that nimbostratus

must have typical vertical velocities of a few to a few tens of centimeters per

second. Though rather weak compared to the vertical velocities in convective

clouds, these gentle upward motions, unlike highly localized convective updrafts,

extend over broad areas and thus condense large amounts of water.

The widespread ascent in nimbostratus can be produced by various mecha-

nisms, depending on the meteorological context. In mesoscale convective sys-

tems (Chapter 9), a widespread cloud deck occurs at upper levels in association

with the deep convective clouds. This cloud deck is slightly buoyant and a hydro-

static pressure pattern and associated circulation develops. The upward compo-

nent of this circulation is characterized by vertical velocities

10-50 em s

-I.

hurricanes (Chapter 10), the pattern of vertical overturning in the inner core

ofthe

204 6 Nimbostratus

storm is a cross-vortex circulation required to maintain gradient-wind and hydro-

static balance. After the inflow toward the storm center rises sharply in the con-

vective eyewall, it slopes more gently outward at upper levels. This sloping ascent

produces a nimbostratus cloud layer with stratiform precipitation in a region

surrounding the eyewall. In frontal clouds (Chapter 11), the widespread gentle

ascent in the nimbostratus is part of the vertical overturning produced by the

dynamics of a baroclinic wave and frontogenesis. We will examine the dynamics

of these sources of gentle ascent in nimbostratus as we consider these specific

phenomena in later chapters.

6.1.4 Role of Convection

Whether nimbostratus is observed in connection with mesoscale convective sys-

tems, fronts, or hurricanes, it is always found to be associated closely with con-

vective clouds. Thus, the dynamics of nimbostratus are not solely a matter of

gentle ascent over a wide area. 'Rather, there appears to be a symbiotic relation-

ship in which both strong convective-scale vertical motions and widespread as-

cent

playa

role. How the convection is important to the nimbostratus is seen by

further consideration of the microphysical processes in the stratiform precipi-

tation.

We have seen in Sees. 6.1.2 and 6.1.3 that the precipitation particles that fall to

the ground from the nimbostratus grow by vapor deposition at upper levels (be-

tween levels 0 and 1 in Fig. 6.2). We have not yet considered the origin of these

particles. The radar and aircraft instrumentation, which indicate the orderly and

distinct physical profiles in Figs. 6.2 and 6.3, detect particles only after they have

formed and reached certain threshold sizes. Nucleation of ice particles certainly

occurs readily at the upper levels of the nimbostratus cloud, since the temperature

is low enough for primary nucleation of ice crystals to be quite active (Sec. 3.2.2).

However, there is a difficulty in envisioning nimbostratus in which particles sim-

ply nucleate in the context of the widespread gentle ascent, begin growing by

vapor deposition, and eventually start to settle downward through the cloud. The

stratiform precipitation areas

nimbostratus are often limited in horizontal ex-

tent, and there is usually a relative wind blowing through the cloud. Growth of the

recently nucleated crystals by vapor diffusion, moreover, is a rather slow pro-

cess.l'" Thus, there is a tendency for the ice particles nucleated within the nimbo-

stratus itself to be advected across the cloud before they can grow to a large

enough size to contribute to the precipitation reaching the ground below the

cloud. The efficacy of the precipitation process within the nimbostratus clearly

would be enhanced by any mechanism that sped up the early growth of the

particles.

Convection appears to be such a mechanism. The stronger vertical motions in

even the smallest of convective updrafts can produce precipitable particles rela-

tively quickly, since large concentrations of water substance are condensed and

161 At the relatively warm temperature of

-5°C,

a small ice particle can grow by deposition to

drizzle size in

-1/2

h, but after that its growth rate slows down (Wallace and Hobbs, 1977).

6.2 Nimbostratus with Shallow Embedded Convection Aloft 205

riming and aggregation are quickly established in the convection. If particles

produced in convection are transferred into the upper levels of a nimbostratus

cloud, they subsequently drift downward through the region of widespread gentle

ascent growing and changing phase by the sequence of processes discussed in

connection with Figs. 6.1a, 6.2, and 6.3. Thus, convection can serve as the major

source of ice particles which subsequently fall out of the nimbostratus in the

orderly pattern indicated by these figures.

There are two major ways in which convective and nimbostratus clouds be-

come joined to produce the cooperative mechanism of stratiform precipitation. In

one case, the convection occurs in a shallow layer embedded in the upper portion

of the nimbostratus and drops ice particles into the layers of nimbostratus below.

In the second case, the nimbostratus is located next to an area of deep convective

clouds. Ice particles grown and carried up to upper levels by the strong convective

updrafts of the deep convective clouds fall out in the neighboring stratiform re-

gion. In the next two subsections we examine these two forms of nimbostratus.

6.2 Nimbostratus with Shallow Embedded

Convection Aloft

The occurrence of a layer of shallow, weak convective cells in a layer aloft is

frequently noted in the nimbostratus of certain parts of frontal clouds. Whether

these

generating cells also occur in the nimbostratus cloud decks of mesoscale

convective systems or hurricanes is not clear at this time, although there is some

Horizontal---+

Figure 6.4 Idealized structure of the radar echo associated with stratiform precipitation with

generating cells aloft.

o 10 km

'-----'

HORIZONTAL SCALE

Figure 6.5

Figure 6.6

...J

dHp'

CJ)

s/~f~,!rm

::::>

I-

Figure 6.7

'~'~SE:E

i:i

E:R;~':

zON

:(;:';2'6'~/~i

::c.:::.

::~~k:

~~~~;~~i~~

~;I;;"

~;;'~~J~;

, .: .: . '.:

cirros/ra/us cloud·.·.·.:: .'

.,~r:~~~;;;;:Z-~o\,

':,''','::,'::

:.:'.::

:....;:::.::-:.:

s/ra~~rm

cloud

-30

-20

-10

IX:

:::>

I-

IX:

a..

I-

6.2 Nimbostratus with Shallow Embedded Convection Aloft

207

evidence to suggest that this may be the case. The present discussion is based

primarily on studies of frontal clouds.

The structure of the radar echo associated with stratiform precipitation with

generating cells aloft is shown schematically in Fig. 6.4. Except for the generating

cells themselves, the radar echo is similar to that in Fig. 6.1a, with the basic

stratiform structure of the precipitation indicated by the bright band in the melting

layer. A shallow layer of potentially unstable air

(fJfJe!

fJz

< 0) is located aloft, and

within this layer convective cells form. As indicated in the inset of Fig. 6.4, the

vigorous updrafts in these cells allow rapid growth of precipitable ice particles,

which then fall out in a fallstreak below the cell. The fallstreak is typically curved

as a result of the wind shear in the layer below the generating cells. The structure

of the generating cell and its fallstreak are essentially the same as the cirrus

uncinus described in Sec. 5.3; the main difference is that the generating cells are

embedded in precipitation, while the cirrus uncinus elements exist in clear air.

162

Examples of radar data showing generating cells embedded in stratiform precipita-

tion are shown in Figs. 6.5 and 6.6.

Two analyses of frontal stratiform precipitation regions are shown in Figs. 6.7

and 6.8. These analyses were based on Doppler radar measurements of the reflec-

tivity and air motions, radiosonde soundings showing the thermodynamic struc-

ture, and aircraft measurements of the cloud microphysical characteristics of the

nimbostratus. In both cases, the water budget of the analyzed region has been

examined. In the case shown in Fig. 6.7, the melting layer was located at about the

height of a cold front, below which the vertical air motion was weak. The zone of

upward motion producing the condensation above this level was located entirely

above the

O°C

level. The region of stronger updraft was, moreover, limited in

horizontal extent to the 50-km-wide zone shaded in Fig. 6.7. As we will see in

Chapter

11, it is typical for frontal lifting to be concentrated into such mesoscale

"rainbands"

rather than to be spread evenly across the frontal cloud. In Fig. 6.7,

is indicated that 20% of the mass of precipitation that fell from the rainband was

produced in the form of ice particles that fell out of the generating cells. The layer

162 The generating cells in stratiform precipitation were first noted by Marshall

(1953).

He pointed

out their similarity to cirrus uncinus by referring to the "characteristic 'mares' tail' pattern of falling

snow observed in vertical section by radar.

" and noting that the cells and their tails resembled the

"hair-like texture of virga. "

Figure 6.5 Time-height record of radar-echo intensity showing generating cells (at about 6 km)

embedded in stratiform precipitation. Data were obtained with a vertically pointing radar in Quebec,

Canada. (Courtesy of Roddy Rogers, Stormy Weather Group, McGill University.)

Figure 6.6 Time-height cross section of the radar reflectivity measured with an 8.6-mm-

wavelength radar in stratiform precipitation on the Pacific Coast of Washington State. Generating cells

are indicated by C. Dashed lines indicate fallstreaks from the cells. (From Businger and Hobbs, 1987.

Reprinted with permission from the American Meteorological Society.)

Figure 6.7 Schematic cross section of a wide cold-frontal rainband in a front passing over western

Washington State. (From Hobbs

et al., 1980. Reproduced with permission from the American

Meteorological Society.)

208 6 Nimbostratus

Depositional

Growth

Further Deposition,

Riming

Slight Aggregation

Heavy Aggregation

Collection

Enhancement

Lower Cloud by

Strong Low Level

Convergence

Figure 6.8 Schematic of the dynamical and microphysical processes in a stratiform frontal

rainband

over

western Washington State. (From Houze et al., 1981.)

of the generating cells is dubbed the seeder zone, since it seeds the layer of cloud

below with ice particles. The

zone

below is called the feeder

zone,163

because the

ice particles from the seeder zone increase greatly in mass as they pass through

this layer of active depositional (and possibly riming) growth. The remaining 80%

of the mass of precipitation was condensed in this layer, where the enhanced

general uplift associated with the rainband was making vapor available to the ice

crystals produced in the generating cells. Thus, the convective generating cells

and the rainband dynamics, which produced the region of enhanced general as-

cent

in the nimbostratus, combined to produce the stratiform rainfall.

The nimbostratus situation depicted in Fig. 6.8 is

another

example of a frontal

rainband. Fallstreaks from generating cells aloft are seen above the radar bright

band.

This case differed from Fig. 6.7 in that the mesoscale ascent in the

rainband extended to low levels so that the precipitation particles continued to

grow by collection of cloud

water

in and below the layer of the bright band. As

indicated in the figure,

65% of the mass of the precipitation was added to the mass

163 The basic idea

ofthe

feeder-seeder

mechanism is attributed to T. Bergeron (1950). He proposed

it specifically to explain the behavior of orographic clouds. The concept, however, has found wider

applicability and is useful in explaining the behavior of nimbostratus in general.

164 The fallstreaks probably all originated in generating cells at some common height (e.g.,

5-6

km);

however, in the cases of the first two fallstreaks on the left of Fig. 6.8, the parent cells had apparently

already died or were located somewhere out of the plane of the cross section.