904 
to yield results beyond the obvious. But the subject is 
popular, and there will be no shortage of studies of the 
stability consideration in the future. 
Base of the Westerlies. Preparation of charts showing 
contours of the base of the westerlies (axis of the sub- 
tropical ridge) have been proposed by Riehl and Shafer 
[33]. Such charts can be a potent analysis tool. They 
furnish indications of temperature distribution and baro- 
clinity of the atmosphere that greatly help to determine 
the upper-contour analysis where there are no raobs. 
The base of the westerlies is tied to many empirical 
forecast rules on hurricanes. For example, many storms 
develop from just south of the latitude where the base 
reaches the high troposphere (200 mb). If there is no 
base at all and the easterlies increase upward in in- 
tensity through the troposphere, hurricane develop- 
ment is not common. In the same situation, existing 
hurricanes will be steered strictly by the deep easterly 
current. The writer suggests that this simple tool be 
tried out in practice. 
STRUCTURE OF TROPICAL STORMS 
The early literature carries long and sometimes acri- 
monious debates on the structure of tropical storms. 
In recent years the arguments have abated. It has 
become clear that the structure varies with the age of 
a storm. Hurricanes that are just forming differ from 
those that have attained full intensity and those that 
are leaving the tropics. The stage of the fully grown 
storm has attracted the greatest amount of attention. 
Mature storms may travel wide distances over the 
tropical oceans with relatively little change of struc- 
ture, and it is possible to make attempts at application 
of steady-state dynamics. In this section we shall also 
concern ourselves with mature storms. 
Therma! Structure of the Rain Area. Observations 
taken during the last ten years have confirmed the 
classical idea that the air in the interior of tropical 
storms is less dense than its surroundings. A detailed 
discussion of radiosonde observations in hurricanes has 
been given by Schacht [87], Riehl [26], and Palmén 
[23]. The principal result is the following: Inside the 
rain area the vertical temperature distribution is the 
same as that obtained by raising an air particle dry- 
adiabatically from the sea surface to the condensation 
level, and then moist-adiabatically to 200 mb or even 
higher (Fig. 1a). Byers [6] adequately likens the hurri- 
cane to a huge parcel of air. Temperatures observed 
aloft in the rain area are the highest possible that can 
be obtained through ascent. 
The situation is quite different from the ordinary 
zones of organized convection in low latitudes (wave 
troughs, shear lines). In these disturbances some ascent 
of surface air takes place. But convergence extends 
to relatively high levels (700-500 mb). Air initially sit- 
uated anywhere between sea level and 500 mb is en- 
trained and moves upward. Usually, this air is un- 
saturated at first and its moisture content is lower than 
that of the surface air that has risen. As shown by 
Stommel [41] and others, entrainment lowers the virtual 
temperature aloft in cumulus clouds compared to that 
TROPICAL METEOROLOGY 
which would prevail without entrainment. Very often 
air at upper levels in the convective zones is actually 
denser than the surrounding air. As shown by rawins, 
the cyclonic circulation increases upward in intensity. 
TEMPERATURE °G 
Fie. la.—Tephigram showing (A) U. 8. Standard Atmos - 
phere, (B) mean tropical atmosphere of the Caribbean in sum- 
mer [37], (C) sounding in rain area of hurricanes, and (D) eye 
sounding [28]. In the diagram vertical lines are isotherms (°C) 
and slanting lines isobars (mb). Curve C follows a moist-adi- 
abatic path from 900 to 200 mb. 
Comparison of the vertical structure of these disturb- 
ances with tropical storms forces us to the conclusion 
that entrainment does not take place in the core of 
hurricanes and that convergence is restricted to the 
mixed ‘“‘subcloud” layer [5]. 
The question arises, Does the foregoing hold true 
over the entire rain area? If it does, horizontal tem- 
perature gradients should not exist in the rain area, 
41 41 
=40 =20 0 20 
TEMPERATURE °C 
-80 -60 
Fic. 1b.—Tephigram showing (A) ascent curve of surface 
air with 7’ = 26C, spec. hum. = 18 g kg™ (repetition of curve 
C of Fig. 1a), (B) ascent of the same air after isothermal ex- 
pansion to 960 mb and moisture increase of 1.5 g kg4, and 
(C) ascent of surface air with 7 = 24C, spec. hum. = 16 gkg?. 
and there should be a frontlike boundary at its outer 
edge. Deppermann [11] postulated such a boundary 
when, on the basis of empirical evidence from the 
Philippine area, he spoke of an outer and an inner ring 
of convection. The notion of a ‘‘wall” of hurricanes, 
often mentioned, supports Deppermann. 
Palmén [23] was the first to venture a detailed verti- 
cal cross section through a hurricane. He showed, con- 
