62 AIR MASS ANALYSIS 
current of warm dry air (Ts); if the 
Te air is shallow, the construction 
of a tephigram with the assumption 
that the surface particle will pene- 
trate the Ts air will often be mis- 
leading, for it suggests there are 
large amounts of available energy 
(positive area) which actually, how- 
ever, cannot be realized because of 
the mixing of this moist stratum with 
‘the dry air above. But generally, 
especially when the TG air is about 
1 km deep, the tephigram gives a 
fairly true picture of the available 
energy, indicating (after extrapola- 
tion of the surface point) positive 
areas at the surface and aloft separ- 
ated by a negative area. This stable 
layer (the negative area) tends to 
resist convection and must in some 
manner be penetrated by the rising 
particle before the Cu clouds can 
develop into thunderstorm propor- 
tions. It is conceivable that in some 
ceases the extra energy needed to 
overcome stability may be supplied 
by the kinetic energy of the rising air 
—this energy causes the cloud to rise 
beyond the level at which it has the 
same density as its surroundings. 
The energy may at times be supplied 
by the lifting action of an incoming 
front or by orographic upthrusting. 
In a special study of aerological 
material Willett; found that the Tc 
layer, whether or not a front enters 
the synoptic field, must be at least 
some 3% km thick for thunderstorms 
to develop. As he points out, however, 
his data contained no ascents made 
during June through September—just 
the time of the year when convective 
activity is the greatest. It is prob- 
able, then, that this conclusion does 
not hold for summer situations.* It 
is also likely that soundings made in 
or near a Cb cloud are not repre- 
sentative of the properties of the 
air mass in which the convective 
overturning is taking place, because 
these turbulent ascending bodies may 
be virtual “fountains” of moist air. 
Further research is required here. 
In making use of the tephigram 
in forecasting local thundershowers it 
is important to consider the chrono- 
logical changes in the positive and 
negative areas. Rapidly increasing 
positive areas and decreasing nega- 
tive areas are highly indicative of 
future thunderstorm activity. If a 
succession of six- to twelve-hourly 
ascents is not available, it is very 
helpful to choose for comparison 
tephigrams of ascents made the pre- 
ceding few days at other stations in 
the same air mass, preferably ones 
lying in the general trajectory of the 
air mass—that is, in the probable 
path of the air current before it 
reached the station for which the 
meteorologist is forecasting. 
The local indicatiors of convective 
thunderstorms are so commonly 
known’ that we need not dwell upon 
them here in detail; some may be 
mentioned: the presence and growth 
of towering Cu during the day, the 
warm, muggy and often stagnant 
atmosphere, and the steep pressure 
fall by afternoon. The distribution 
of the Cu is frequently indicative, for 
it has been shown by Bjerknes® that 
tall Cu separated by fairly large clear 
spaces (Cu. castellatus) are most 
7Discussion and Illustration of Problems Sug- 
gested by Analysis of Atmospheric Cross- 
Sections, M.I.T. and W. H. Ocean. Inst., 
Papers in Phys., Ocean. and Met., Vol. 4, 
No. 2, 1935. 
*For further discussion of this see my article 
in the Jan. 1938, Bulletin Am. Met. Soce., 
Dols 
2Cf. C. F. Brooks: The local, or heat, thun- 
derstorm, Mo. Weather Rev., June, 1922, Vol. 
50, pp. 281-284. (Includes “Questionnaire [of 
16 questions] for predicting local thunder- 
storms.” ) 
’Physikalische Hydrodynamik. Berlin, 1933; 
Hydrodynamique Physique. Paris, 1934 (3 
Vols.). In Q. Jn. Roy. Met. Soc., April 1938, 
pp. 325 ff, Bjerknes shows that Cu convection 
does not release sufficient kinetic energy 
(motion) to cause Cu to develope into Cb un- 
less the spacing—thickness ratio of the Cu 
towers is below a certain critical value. 
