TORNADOES AND RELATED PHENOMENA 
mph. This depression is also equal to the theoretical 
wind pressure of the corresponding wind. Hence, the 
wind pressure and atmospheric pressure effects on closed 
structures should be of the same order of magnitude. 
Flow Pattern. Although the rotation of air m a tor- 
nado is generally accepted as a fact today, there was 
no such agreement one hundred years ago [4, 17, 18]. 
With a few doubtful exceptions [12] the rotation in 
tornadoes is believed to be always cyclonic. The evi- 
dence for this is found in a wide belt of damage usually 
on the southeast side of a tornado path in the Northern 
Hemisphere. In this belt the tornado wind must be 
increased because of the prevailing wind Gn approxi- 
mately the same direction) which in some cases may be 
responsible for even more damage than the rotation. 
Theoretically, the tangential velocity should be in- 
versely proportional to the distance from the center for 
frictionless inflow, according to the principle of the 
conservation of angular momentum. Actually, the pres- 
ence of a frictional torque acting against the wind 
makes the velocity profile somewhat less steep. Conse- 
quently, the vortex is not irrotational as it would be 
for zero friction, but has a cyclonic vorticity. The direc- 
tion in which a house is twisted on its foundation is 
not a reliable indication of the sense of the tornado’s 
rotation nor of its vorticity, as the twisting is dependent 
on variations (along the windward wall) in the struc- 
tural resistance and in wind speed, which may be in- 
fluenced by trees or by a strictly local whirl. 
The radial motion in a tornado is inward near the 
ground. The winds in different tornadoes or even within 
one tornado probably blow inward at different angles 
to the radius, depending on the ratio of inflow and 
tangential speeds. Along many tornado tracks, evi- 
dence for the inflow is much clearer than evidence for 
the rotation, which suggests that, in these cases, inflow 
- speeds may be greater than tangential speeds except 
near the center. The inflow winds are often strong 
enough to cause damage by suction over lee roofs or 
around lee walls, that is, on the side facing the tornado 
center [11]. This suction, added to the atmospheric 
pressure reduction, builds up an intolerable pressure 
deficit outside. On the other hand, the windward roofs 
and walls are damaged by excess pressure due to the 
wind (counteracted somewhat by the atmospheric pres- 
sure deficit) and by battering from missiles. 
Continuity demands a strong vertical motion upward 
within the vortex. A remarkably strong upflow was 
once witnessed (Abilene, Texas, June 1938) when a 
tornado got ahead of its cumulonimbus cloud and 
immediately created a new cumulonimbus cloud which 
shot up to an elevation of 35,000 ft in one minute. The 
rapid vertical acceleration of the air requires a vertical 
pressure-gradient force upward much stronger than 
gravity. The central depression of the tornado must be 
much more pronounced just above the ground than at 
the ground. Buildings in the path of a tornado receive 
greatest damage in their upper floors. 
Redfield visualized a double spiral, the air from out- 
side the vortex spiralling first downward and inward 
toward the center, then within the vortex upward and 
675 
outward. The latter motion was once observed [17]. 
Whirling air coming down at 1800 feet per minute has 
actually been encountered by a glider pilot in a desert 
dust whirl; and hot puffs of air about 20F warmer 
(because of adiabatic compression) than the prevailing 
air have been felt at the ground near some tornadoes. 
The turbulence in tornadoes is extremely large be- 
cause of strong surface and internal frictional effects. 
Mechanical eddies may contain destructive gusts of 
wind. In the central region of rapid convergence and 
upflow, speeds of secondary whirls increase as their 
radu decrease. The presence of such miniature torna- 
does may be indicated by short pendant clouds overhead 
close to the main pendant cloud (for a photograph of 
this, see [7]). 
Appearance and Visibility of a Tornado. A tornado 
may appear as two columns: (1) a pendant cloud ex- 
tending downward from the general cumulonimbus 
cloud base, and (2) an ill-defined mass of dust and 
debris extending a short distance upward from the 
ground. The air within the pendant cloud has expanded 
adiabatically to pressures and temperatures lower than 
the condensation pressure and temperature of the air 
mass, which is assumed to have nearly uniform specific 
humidity before condensation. The freezing level [6] 
as well as the condensation level is lowered by adiabatic 
cooling. The core of a tornado felt ice-cold to one ob- 
server. At a central pressure of 900 mb the temperature 
could be as much as 10C or 18F cooler (rather than 
warmer [4]) than the surroundings. 
The shape of the pendant cloud is determined by the 
slope of the condensation-pressure isobaric surface. The 
slope is usually steepest near the axis of rotation, but 
sometimes excessive friction greatly reduces the imner- 
most slope, prevents the cloud from reaching the ground, 
and changes it from a funnel shape to a basket or balloon 
shape. If the air is dry enough or the central depression 
is weak enough for the central pressure to be just slightly 
below the condensation pressure, the pendant cloud is 
narrow, resembling an elephant trunk, a rope, or a 
serpent. After the waterdrops condense in the pendant 
cloud they tend to whirl outward from the center owing 
to their centrifugal force. This leaves a very small 
“eye” at the center [13], characterized by cloudless air 
and relatively light wind, as has actually been observed. 
The dust column is usually wider than the pendant 
cloud and surrounds its base, because the wind normally 
reaches a speed great enough to raise dust before the 
pressure drops as low as the condensation pressure. 
Exceptions occur where the soil is too wet for dust or 
where the air is so moist that the pendant cloud is very 
broad, extending outward beyond the edge of the high 
winds. The black dust column may be mistaken for 
smoke rising from a fire. As it rises, most of the debris 
is thrown outward in all directions by centrifugal force, 
sometimes giving the appearance of a dark fountain. 
The condition for the occurrence of damage is not 
whether the pendant cloud strikes, but whether the 
usually dusty column of high winds strikes. 
As many as nine pendant clouds have been seen from 
one place during a tornado situation [1, 12]. In many 
