Potter T.D., Colman B.R. (co-chief editors). The handbook of weather, climate, and water: dynamics, climate physical meteorology, weather systems, and measurements

Подождите немного. Документ загружается.

Tornadoes are arguably the most hazardous weather event associated with

convective storms. Although tornadoes are by no means limited to supercells,

those tor nadoes associated with supercells have a much greater likelihood to be

intense and long-lived than those produced by nonsupercell thunderstor ms. In

turn, this means that such tornadoes have the highest potential for damage and

casualties. This is exempliﬁed by events on May 3, 1999, when an outbreak of 69

tornadoes across Oklahoma and Kansas was produced by only 10 supercell storms.

One tornado from the ﬁrst supercell of the day caused $1 billion in damage and

36 fatalities as it tracked ﬁrst through rural areas southwest of Oklahoma City,

Oklahoma, and then on into the metropolitan area.

As with tornadoes, the hail produced by supercells is much more likely to exceed

2 in. (5 cm) in diam eter than in nonsupercell storms. There is strong scientiﬁc

evidence to believe that updrafts are enhanced by the presence of a mesocyclone,

and giant hail requires an inte nse updraft for its formation. Some supercells are

proliﬁc hail producers, creating long swaths of hail up to 10 cm in diameter.

When such storms interact with populated areas, they can cause damage on the

order of $100 million or more from broken glass, dented motor vehicles, roof

destruction, and vegetation damage. Occasionally, people are seriously injured or

even killed by being caught outdoors during a fall of giant hail.

Windstorms of a nontornadic nature in supercells can also reach extreme propor-

tions. A few times per year in North America, supercells produce swaths of wind up

to 20 km wide and more than 100 km long, within which winds can exceed 25 m=s

for more than 30 min, with peak gusts approaching 50 m=s. In forested areas, such

events produce vast blowdowns of trees. Interactions with populated areas are rare,

but the potential for destruction is enormous. In such events, hail can accompany the

strong winds, adding to their destructive potential.

Finally, a few supercells are responsible for prodigious rainfall production. It

appears that most supercells are not very efﬁcient at producing rainfall because

they often are associated with processes promoting evaporation of precipitation.

Moreover, supercells are mostly isolated storms that move over a given location

relatively quickly. Nevertheless, their intense updrafts can process a lot of water

vapor into precipitation, however inefﬁciently it is accomplished. Thus, supercells

have produced bursts of precipitation exceeding 200 mm=h, even if only for a short

time. Such intense rainfall rates can create ﬂash ﬂoods, particularly in urban areas

where runoff is so high owing to the lack of per meability of urban environments

composed mostly of concrete, buildings, and other hard surfaces.

Variations on the Theme

Supercells are not all the same, but they have certain common features. The presence

of a deep, persistent mesocyclone is the deﬁning feature of a supercell, so it is

possible to look only at storms meeting that criterion. Using this broad deﬁnition,

we should not be surprised to learn that variations on the supercell theme exist.

596 SEVERE THUNDERSTORMS AND TORNADOES

The most widely used classiﬁcation scheme for supercells is based loosely on the

notion that mesocyclones vary in the amount of precipitation they contain. The

prototypes are called the low-precipitation, classic, and high-precipitation supercell.

Low-precipitation (LP) supercells characteristically have little or no precipitation

within their mesocyclones. The absence of heavy precipitation means they typically

do not produce strong downdrafts and outﬂow, nor are they likely to be tornadic.

However, they can produce falls of giant hail and, like other supercells, tend to do so

in relatively long swaths. They typically are not very large storms and tend to be

observed mostly in the transitional environments between arid and moist regions

(e.g., the western Great Plains of North America). Figure 12 shows a schematic

illustration of the appearance of an LP supercell, both visually and on radar.

Classic (CL) supercells most clearly resemble those described by Browning and

his coll aborators. They often exhibit most, if not all, of the traditional radar echo

morphology associated with them in the literature. They produce all for ms of severe

weather, including the most extreme examples, although they are not likely to be

ﬂash ﬂood producers. Figure 13 illustrates the archetypical appearance of a CL

supercell.

Finally, high-precipitation (HP) supercells have mesocyclones deeply embedded

in precipitation, although there may be a narrow corridor that, at low levels, remains

free of precipitation (sometimes referred to as an inﬂow notch). Figure 14 shows the

typical appearance of HP supercells. With heavy precipitation falling into the down-

drafts on a supercell’s rear ﬂank, the potential for strong winds is quite high. Clearly,

HP supercells also have a potential to be heavy rainfall producers, as well as giant

hail. HP supercells stand somewhere between CL and LP supercells in their tornado

potential. It is rare for a violent, long-track tor nado to be associated with storms

displaying a predominantly HP structure.

6 TORNADOES

Tornadoes are rotating columns of air that extend from the surface to the interior of a

thunderstorm cloud (or, more correctly, a cloud associated with deep moist convec-

tion, with or without thunder). The distinction between tornadoes and other vortices

within thunderstorms is not always clear, so tornadoes are sometimes deﬁned as

having wind speeds near the surface that have the potential to cause damag e. Many

tornadoes produce cylindrical or conical clouds of condensed water (Fig. 15a), but a

funnel cloud is not always present. Some tornadoes are made visible only by lofted

debris (Fig. 15b).

The number of reported tornadoes in the United States now typically exceeds

1000 per year. Annual totals of fatalities and property damage caused by tornadoes

vary considerably from year to year. The death toll, as a fraction of the total popula-

tion o f the United States, has been decreasing on average since the mid-1920s.

Improved forecasting, detection, and warning of tornadoes; improved communica-

tion of warnings; increased public awareness of safety precautions; urbanization; and

changes in building standards may be some of the important factors in explaining

6 TORNADOES 597

this trend. Tornadoes typically last less than 10 min but have been known to persist

for over an hour. Wind speeds may exceed 120 m=s (270 mph) in particularly strong

tornadoes but are less than 60 m=s (135 mph) in most cases. In some tornadoes, the

distribution of strong winds around the vortex core is relatively uniform; in others,

Figure 12 Schematic of a low-precipitation supercell. (Courtesy of National Severe Storms

Laboratory.) See ftp site for color image.

598

SEVERE THUNDERSTORMS AND TORNADOES

Figure 13 Schematic of a classic supercell. (Courtesy of National Severe Storms

Laboratory.) See ftp site for color image.

6 TORNADOES 599

Figure 14 Schematic of a high-precipitation supercell. (Courtesy of National Severe Storms

Laboratory.) See ftp site for color image.

600

SEVERE THUNDERSTORMS AND TORNADOES

()b()a ()c

Figure 15 Tornado photographs: (a) Tornado with a cone-shaped condesation funnel near Stockton, Kansas, on May 15,

by David Dowell). See ftp site for color image.

601

the strongest winds may occur within relatively small subvortices (Fig. 15c).

Although tornadoes may extend to over 10 km above ground within the cloud, the

strongest horizontal winds in mature tornadoes occur within 100 m of the surface.

Near the radius of strongest horizontal winds, air rises at speeds comparable to those

of the horizontal winds; the extreme vertical velocities may loft debris to great

heights. Near the center of the tornado, there may be descending motion. On rare

occasions, the large-scale atmospheric pattern may be favorable for the formation of

tornadic thunderstorms over a broad region covering multiple states. The ‘‘super

outbreak’’ of April 3–4, 1974, is an extreme example. During this outbreak, at least

148 tornadoes occurred within 14 states from the Midwest to the Southeast. The

majority of tornadoes, however, are isolated occurrences. Most thundersto rms for m

in environments that are not favorable for tornadoes, or when the environment would

support the development of tornadoes, storms do not always form.

Scientists continue to be challenged to explain the detai ls of how tor nadoes form

(and to explain why tornadoes do n ot always form on days when they are antici-

pated). Organized ﬁeld programs to study tornadoes began in the early 1970s and

continue today. Scientists use a number of mobile devices (Doppler radar, instru-

mented automobiles, weather balloons, etc.) to collect measurements of wind,

temperature, pressure, humidity, and precipitation in and near tornadoes

(e.g., Fig. 16). To assess how a tornado forms in a particular thunderstorm, scientists

Figure 16 (see color insert) High-resolution measurements of the reﬂectivity factor (an

indicator of precipitation size and number concentration) by the Doppler on Wheels mobile

radar in a tornadic storm with a hook echo near Rolla, Kansas, on May 31, 1996. The thin blue

rings indicate range from the radar at inter vals of 5 km. See ftp site for color image.

602

SEVERE THUNDERSTORMS AND TORNADOES

require accurate measurements at high temporal and spatial resolution; the need for

such complete observations continues to challenge observational capabilities.

Signiﬁcant advances in our understanding of tornadoes and their parent storms

have also come from numerical simulations with high-speed computers. In a numer-

ical model, the thundersto rm and its environment are represented on a three-dimen-

sional grid. During each time step of the simulation, the dynamical equations

governing the change s of wind, temperature, pressure, humidity, liquid water

content, and ice content are solved at each grid point. Numerical simulations

(e.g., Fig. 17) have reproduced storm features analogous to those in observed storms.

The prevailing hypotheses for how tornadoes form all involve horizontal conver-

gence of low-level air that has signiﬁcant angular momentum. As air retains its

angular momentum with respect to the center of the region of convergence while

spiraling inward, the rate of rotation increases. The sequence of events leading up to

the formation of a tornado is not identical in all cases. If the air near the surface in

the environment of a thunderstorm already had signiﬁcant rotat ion before the thun-

derstorm formed, then the mechanism of tornado formation could be relatively

simple (Fig. 18). For example, a low-level boundary, along which there is a shift

in the wind direction and speed, may be present in the environment. The wind shift

may be organized into circulations that are initially a few hundred meters to a few

kilometers wide (Fig. 18). If one of these circulations coincides with a growing

Figure 17 Simulated low-level reﬂectivity factor (shading) and horizontal wind (vectors) in

a numerical simulation of a tornadic storm. (Courtesy of Matt Gilmore of Cooperative

Institute of Mesoscale Meteorological Studies.) See ftp site for color image.

6 TORNADOES 603

thunderstorm updraf t, then a tornado may for m as the circulating air at the base of

the updraft converges to smaller and smaller radii. This mechanism involving the

interaction of an updraft with a preexisting low-level circulation may apply to

tornadoes that form within otherwise nonsevere deep moist convection over water

(‘‘waterspouts’’) but may also explain the formation of some tornadoes over land.

The formation of most of the tornadoes in supercell thunderstorms appears to be

more complicated. Supercells develop in environments in which there is strong

vertical shear (i.e., a change in direction and=or speed with height) of the horizontal

wind; such shear is associated with rotation about a horizontal axis. In addition,

rotation about a horizontal axis may develop within a mature thunderstorm when

there are signiﬁcant horizontal variations of air density in the region where rain and

hail are falling. Updr afts and downdrafts within the superc ell thunderstorm may tilt

the orientation of the rotation such that a component of it becomes rotation about a

vertical axis. If the air that is rotating about a vertical axis is drawn into the region of

horizontal convergence at the base of the thunderstorm updraft, then a mesocyclone

(a region of rotation a few kilometers wide) and perhaps tornado (a narrower, more

intense vortex) may form. A major current research problem is to determine why

tornadoes form in some supercell thunderstorms but not in others.

7 RADAR CHARACTERISTICS OF SEVERE STORMS

Use of Radar in Observing Severe Storms

Radar has been used to identify and track severe local storms since its invention in

the 1940s. Weather radar detects hydrometeors within storms with a series of pulses

of electromagnetic energy, directed by an antenna that is mechanically rotated in

azimuth and elevation. It alternates transmitting and receiving those pulses and

measures the distance to the target by the time delay between a pulse transmission

and echo arrival. Figure 19 illustrates the typical radar beam geometry. Because of

the hazards within severe storms to aircraft penetrations, radar has been the primary

tool that has been used by meteorologists to probe the inner structure and circula-

tions of storms to better understand the physics of their behavior.

Figure 18 Schematic of nonsupercell tornado formation (by R. Wakimoto and J. Wilson;

direction. Numbered loops represent circulations along a low-level boundary, which is

indicated by a black line.

604

SEVERE THUNDERSTORMS AND TORNADOES

Early techniques utilized ‘‘incoherent’’ technology of simply displaying returned

power normalized for range on a horizontal display device (termed a plan position

indicator, or PPI). Severe storms, sometimes producing hazardous weather such as

large hail, high winds, and tornadoes, very often exhibit a characteristic structure

that could easily be tracked on a PPI display, and suitable warnings could be made

for regions in the storm’s path by simple extrapolation of storm motion. One of the

key developments in radar technology that made radar valuable for distinguishing

between tornado-producing storms and other less severe events such as hail storms

was the development of Doppler radar techniques. Doppler radars not only detect

and measure the mean power received from a target but also its relative motion. Thus

regions of a storm can be seen to approach or recede from a radar site and inferences

about rotation within the storm (e.g., tornado mesocyclones that are the parent

circulation of tor nadoes) can easily be made. This motion detection capability

greatly assists weather forecasters to better distinguish between storms that produce

tornadoes from ones that do not.

The U.S. National Weather Service (NWS) was quick to recognize the value of

radar in providing timely public warnings of severe weather. In the 1950s a network

of incoherent radars was deployed (called the WSR-57). With the realization by the

1980s that Doppler technology could offer signiﬁcant improvement in torna do warn-

ing lead time, a new generation of operational weather radars (termed at the time

NEXRAD, now called the WSR-88D radar) was deployed to meet the warning

missions of the NWS, U.S. Air Force, and the Federal Aviation Administration.

The network consists of over 130 radars providing nearly complete coverage of

the continental United States (Fig. 20). The WSR-88D has greatly increased the

tornado warning lead time from nearly zero before the deployment of the WSR-88D

network to nearly 10 min today.

Figure 19 Radar beam geometry. The size of the beam is a function of antenna size and

radar wavelength.

7 RADAR CHARACTERISTICS OF SEVERE STORMS 605