906 
the computed ascent curve of air from the mixed layer. 
It followed that the entire rain area of a mature storm 
consists of air that has risen from the vicinity of the 
surface. The low-level convergence is restricted to the 
mixed layer, and a level of nondivergence is situated at 
its top. 
0 
Fig. 2.—Vertical cross section through mature tropical 
storm showing hypothetical model of the vertical circulation. 
Heavy solid lines are eye boundaries and tropopauses. The 
model is based in part on Durst and Sutcliffe [13], Wexler [42], 
and Palmén [23]. 
Heat Transfer. We can estimate the size of the area 
from which a hurricane must draw surface air in order 
to build up the thermal structure actually encountered 
in the interior. Let 7) denote the radius of this area and 
r the radius of the rain area as encountered. The eye 
may be neglected for this computation since its area is 
very small (10 per cent or less of the rain area). The 
depth of the mixed layer is about 50 mb and the clouds 
extend at least to 200 mb. It follows that Ay = 164, 
where A> is the initial area and A the rain area. Also, 
Ay = aro and A = ar*. Therefore, ro = 4r. If r = 
100 km, a moderate value, 7» = 400 km, and the diam- 
eter of the initial area is about 8 degrees of latitude. 
This is a large value and shows how far the influence of 
even a small storm extends during the time of formation. 
In the course of its further life history, a very sizable 
fraction of the surface air initially situated over a tropi- 
cal ocean is drawn into the circulation. 
Up to now, no one has calculated the total effect that 
storms can exert on the general circulation as a result 
of the upward transfer of heat. Evidently an enormous 
gain of heat—both sensible and latent—accrues to the 
upper troposphere from the large-scale funnelling. It 
has been suggested that the frequency of tropical storms 
should be great when the normal processes of heat ex- 
change between low and high levels are weaker than 
usual for some reason. But so far it has not been pos- 
sible to establish a clear relation between hurricane 
frequency or duration and other factors of the general 
circulation. 
Wind Distribution near the Surface. Estimates of 
lateral convergence of air into tropical storms and the 
amount of vertical motion present usually treat in- 
stantaneous values only, rather than totals integrated 
over some unit of time. Many observations have been 
published on the rate of inflow near the ground, as de- 
TROPICAL METEOROLOGY 
termined from surface-wind measurements. General _ 
laws, however, have not been formulated. The rate of 
inflow is quite variable. Usually it is not symmetrical, 
but confined to one or two quadrants. During the life 
of one storm, the quadrant of greatest inflow will often 
change. 
Various forecast rules have been offered concerning 
the relation between storm movement and quadrant of 
greatest inflow. One such rule states that a storm will 
move toward the region of heaviest rainfall (greatest 
convergence); another states that the heaviest precipi- 
tation occurs to the rear of a center. Observations can 
be found to demonstrate practically any relation be- 
tween direction of motion and rainfall concentration. 
This merely shows that a variety of synoptic influences 
are operative. A considerable advance could be gained 
through a serious attack on the connection between 
synoptic situation, distribution of convergence, and 
storm motion. 
Our knowledge regarding the wind distribution within 
tropical storms and the dynamical laws that guide the 
air from the outskirts to the center of the cyclones is so 
deficient as to be deplorable. Deppermann is one of the 
few writers who has made a detailed effort to calculate 
radial and tangential velocity components. He also has 
presented a theoretical model [11] which discusses the 
tropical storm as a Rankine vortex with two “rings” 
of convection. Apart from Deppermann, writers have 
contented themselves with application of simple hydro- 
dynamics, generally only a treatment of the vr vortex. 
The best discussion is found in Brunt’s textbook [4, 
pp. 298-306]. Even Brunt, however, treats only a few 
elementary concepts—circulation, vorticity, and the 
wind distribution in a stationary vortex with point 
convergence at the center. 
It is easy to understand why the synoptic meteorol- 
ogist is discouraged from attacking the problem of the 
dynamics of the tropical storm. He rarely has data at 
his disposal for which he can claim general validity. 
But the apathy on the part of theorists is hard to 
comprehend. To the best of the writer’s understanding, 
no laboratory experiment has been carried out to deter- 
mine whether ‘‘simple” vortices can be generated in air 
as in liquids. It would be quite feasible to construct 
theoretical hurricane models, dropping some of the 
assumptions contained in the simple theory. If the air 
did move under conservation of momentum, the abso- 
lute rather than the relative motion of air about a hurri- 
cane center should be considered. The absolute rather 
than the relative vorticity should become zero through- 
out the body of a hurricane. This, however, never hap- 
pens. There are a few famous cases—often quoted—in 
which the relative vorticity was found to be nearly zero. 
Since we know of no physical law that demands con- 
servation of relative angular momentum in a hurricane, 
the rare cases quoted must be regarded as accidental. 
It is certain that the absolute angular momentum is 
not conserved in the air flowing toward a center. Ground 
and eddy friction can be responsible for this, as well as 
the asymmetrical shape of many storms. None of these 
factors has been properly investigated, especially the 
