A PROCEDURE OF SHORT-RANGE WEATHER FORECASTING 
or less) is superimposed on a horizontally narrow wedge 
of abnormally humid air at the ground (a dew point of 
+15C or more at the surface of the earth) which is 
possibly both conditionally and potentially unstable 
(hence, for the humid layer, a level of free convection 
at 650 mb, or lower). 
2. Heading obliquely to the axis of the humid wedge, 
there must be a very narrow band of strong westerlies 
aloft (>35 knots at 600 mb). 
3. The accurate analysis of forerunning upper cold 
fronts is also an important prerequisite for tornado 
forecasting. Forced lifting of the lower wedge of humid 
air must take place—usually as a result of the invad- 
ing cold front to which the forerunning upper cold 
front or cold quasi-front is associated. 
TEMPERATURE (°C) 
= 30) = 25) G20) = 25) -20 
793 
only in that part of a deep, subfreezing water cloud 
where supersaturation with respect to ice also occurred. 
As shown by Godske and others [33], supersaturation 
with respect to ice exists whenever 7’ > —8D, where 
the temperature 7’ and the dew-point deficit D are 
expressed in degrees centigrade. On an aerological dia- 
gram, the icing layer is determined by the subfreezing 
cloud area enclosed by the temperature ascent-curve 
on the left and the —8D curve” on the right, as shown 
in Fig. 12. Moreover, the intensity of the icing is indi- 
cated by the size of this enclosed area. The cloud type 
and precipitation, both observed at the surface of the 
earth, will show the degree of colloidal-thermodynamic 
instability and hence the type of icing—rime or glaze. 
In particular, whenever the ascent curves of tempera- 
}===~ Lieut, 
RIME C3 
4 
7 
a 
MILLIBARS 
(a) 
(c) 
(b) 
Fic. 12.—Three sets of ascent curves for temperature (solid) and—8D (dashed), plotted on a skew 1’, log p-diagram. (a) Radio- 
sonde observations for 0300 GMT, April 25, 1949, Yakutat, Alaska. Image points labelled with values of temperature in degrees 
centigrade (upper left) and dew point; the —8D curve labelled with corresponding values of —8D. (b) Aircraft ascent for 1400 
GMT, February 20, 1949, at 41.2°N, 56.2°W, with the observed icing and clouds plotted to the left. (c) Aircraft ascent for 1700 
GMT, February 7, 1949, at 40.1°N, 46.6°W, with the observed icing and clouds plotted to the left. 
An empirical rule for forecasting the movement of 
an individual tornado has been given by both Showalter 
[74] and Willett [82], namely that a tornado moves 
roughly with the direction and speed of the m7’ stratum 
just beneath the inversion. Others state simply that 
the tornado travels roughly parallel to the path of 
the accompanying cyclone. 
The Prediction of Aircraft Icing. For icing on air- 
craft in flight, the subfreezing cloud containing water 
in the liquid phase must be supersaturated with respect 
to ice. Although supercooled cloud droplets occur in 
air subsaturated with respect to (a plane surface of) 
water, sublimation starts only after the moist air at- 
tains saturation equilibrium with respect to ice. In a 
layer subsaturated with respect to ice, the physically 
unstable, supercooled fog droplets, upon striking the 
leading edges of the aircraft, are induced mechanically 
to congeal as light rime ice which almost instantly 
evaporates into the air streaming very rapidly over the 
rime-forming surface of the aircraft. For example, the 
two aircraft soundings in Figs. 12b and 12c report icing 
ture and dew point coincide, the —8D curve must 
necessarily coincide with the OC isotherm on the dia- 
gram. In the case where the ascent curves of tempera- 
ture and dew point coincide in a subfreezing layer, 
the air would be saturated with respect to water, 
supersaturated with respect to ice. Only light rime 
icing would occur in the altostratus-nimbostratus cloud 
system of this layer; but moderate rime icing would 
be encountered in cumulonimbus virga which is lo- 
cated in this layer. Severe clear icing—for the same 
sounding as above—would occur in the stratocumulus 
virga, cumulus virga—particularly when transforming 
to cumulonimbus—and stratus. Whenever, on the other 
hand, the ascent curves of temperature and dewpoint 
do not coincide but the temperature curve lies to the 
left of the —8D curve in a subfreezing layer, this 
12. For most practical purposes, the dew-point ascent curve, 
which is frequently plotted on the aerological diagrams, could 
be replaced by the —8D curve, since the latter, together with 
the OC isotherm, may be used in place of the ascent curves of 
temperature and dew point for showing the vertical variation 
of the dew-point deficit 
