890 
only about thirty-five miles, therefore neither minimum 
barometer nor intensity of the storm is necessarily an 
indication of the area of hurricane winds. 
Some writers have suggested various means of com- 
puting the distribution of surface wind velocities in a 
cyclone vortex by developing reasonable theoretical 
velocity profiles. Deppermann [4] has developed an 
idealized profile of the type shown in Fig. 1 as most 
typical of tropical cyclones. Here we have a Rankine 
vortex in which V/R = constant for the innermost cone 
including the eye and its adjacent areas, and VR = 
' constant in the outer vortex. Deppermann describes 
the area between these two as a ring of violent con- 
vection characterized by a still different velocity vari- 
ation. While this may serve as an average distribution 
for a large number of storms of various intensities, 
the range of deviations from this average will probably 
be quite considerable, both from quadrant to quadrant 
and from one stage of storm development to another. 
Tn the small immature storms maximum wind velocities 
will be reached closer to the eye. Since the required 
dense network of observing stations has never been 
available, no rigorous study of surface wind velocities 
and directions and associated pressure gradients has 
ever been made. 
Clouds. Cloud arrangements, in common with other 
tropical-cyclone characteristics, will vary considerably 
with individual storms. As a rule the cloud sequence is 
much the same as in advance of a warm front in middle 
latitudes. First to appear are the cirrus clouds which 
frequently seem to radiate from a point on the horizon 
in the general direction of the storm. The cirrus will 
thicken to cirrostratus and then to altostratus and 
altocumulus. Soon occasional cumulonimbus, extending 
through the upper deck with passing squalls, will ap- 
pear. Finally a dark wall of cloud, sometimes called the 
“bar” of the storm, appears, and squalls become almost 
continuous with stratus based at from 500 to 2500 ft 
or with a precipitation ceiling. 
There is occasional reference in the literature to lurid 
sunsets preceding the advent of a hurricane. This condi- 
tion may precede a hurricane but handsome sunsets 
are frequent in the tropics and subtropics where cirrus 
(nothus) formed from dissipated cumulonimbus are 
present. 
In the hurricane vortex high clouds are rarely visible. 
Low clouds will have about the same direction as the 
surface wind although usually more tangential to the 
isobars. At sea, clouds are occasionally reported down 
to the water, but these reports are considered un- 
reliable since rain and flying spray will frequently lower 
visibility to zero. On land, even at the height of the 
storm, the ceiling is usually reported as above 1000 ft, 
although in later stages of recurvature 500-ft ceilings 
are not infrequent. 
Vines [19] believed that cirrus radiate from and move 
outward from the center of the tropical cyclone. Most 
of the observations during the past twenty years in- 
dicate that, in general, cirrus move forward in the 
general air stream in which the cyclone is imbedded. 
It does not necessarily follow that the hurricane circula- 
TROPICAL METEOROLOGY 
tion does not at times reach the cirrus level since cirrus 
cannot ordinarily be observed over the vortex. How- 
ever, outside the periphery of low and middle cloud 
decks associated with the vortex, cirrus appear to be 
unaffected by the storm’s circulation. 
Precipitation. The distribution of rainfall around a 
tropical cyclone varies markedly from one storm to 
another. In general the more immature the storm and 
the lower the latitude, the more symmetrical the rain 
pattern from the standpoint of intensity and area. As 
the storm begins to recurve and reaches more northern 
latitudes, rainfall, from the standpoint of both intensity 
and area, becomes concentrated in the front quadrants. 
Notable exceptions, however, have occurred. In the 
“Yankee” hurricane of 1935, 0.04 inches of rain fell 
before the arrival of the center and 3.40 inches after- 
ward. In October 1941 a hurricane passed over Miami, 
Dinner Key reporting a maximum wind of 123 mph, 
with only a trace of precipitation. 
A check on three fairly recent Puerto Rican hurri- 
canes reveals that at San Juan, 3.03 inches fell in the 
front half and 7.85 inches in the rear half of the storm 
of September 1926. This was a mature Cape Verde 
storm and the center passed to the south of the station. 
On September 27, 1932, a relatively immature hurricane 
passed over Rio Piedras seven miles southeast of San 
Juan, and 1.17 inches fell before and 1.59 inches after 
the center passed. On September 19, 1931, 0.71 inches 
fell before the calm center and 1.15 inches after it 
passed. This was a very young storm which had reached 
hurricane intensity less than twelve hours before it 
arrived over Rio Piedras. It is believed the distribution 
of rainfall around these three storms is normal for mod- 
erately low latitudes, although the total amount in the 
two later storms was subnormal. 
Vines indicated that in Cuba rainfall extends farther 
in advance of the storm than to the rear, but there are 
also cases when the reverse is true. Most storms affect- 
ing Cuba have a strong northerly component and occur 
late in the season. 
Deppermann has not reached very definite conclu- 
sions on the distribution of precipitation around tropi- 
cal cyclones but states that most of the rain in the so- 
called ‘‘triple-point” typhoons in the Philippine area 
occurs in the rear and most of the rain in the “trade- 
norther” type in the right front quadrant. 
Tropical cyclones in the Gulf of Mexico, according to 
Cline [1], have the greatest precipitation intensity 60-80 
miles in front of the cyclone center and mostly to the 
right of the line along which the cyclone is advancing, 
and with relatively little precipitation in the rear half. 
Cline suggests that this is due to the winds of high 
velocities moving through the right rear quadrant and 
converging with winds in the right front quadrant of ~ 
lesser velocity and greater cross-isobar components. He 
cites a rather extreme example, in the cyclone of October 
10-12, 1909, when Key West received 12.04 inches be- 
fore passage of the center and only 0.24 inches after- 
ward. This storm had already recurved and was moving 
northeastward and, indeed, most of the storms cited by 
Cline were recurving and in stages 3 and 4. 
