TROPICAL CYCLONES 
oceanic areas where upper-air data are lacking. Satis- 
factory forecasting of recurvature is impossible without 
accurate constant-pressure charts. Riehl and Shafer 
[14] studied some 57 tropical cyclones occurring be- 
tween 1935 and 1943 and classified them in groups as 
follows: (1) seventeen which curved north or north- 
eastward, (2) twenty-three which continued between 
west and north without recurvature, and (3) seventeen 
which had irregular tracks, particularly motion with a 
southward component for a portion of the track. It 
should be noted that the “abnormal” storm tracks 
of the last group were as frequent as the ‘normal”’ 
tracks of the first group. 
Riehl and Shafer found that a tropical cyclone will 
begin to react when the forward edge of the westerlies— 
at least as low as 8000-14,000 ft—is found 500-700 
miles to the west. As the polar trough advances east- 
ward, the storm turns from a westward to a northward 
path. The most difficult, though frequent, cases oc- 
curred when storms encountered strong troughs and an 
intrusion of the polar westerlies into the tropics similar 
to group (1). In that group, however, the base of the 
westerlies not only lowered but remained low in the 
subtropics and part of the tropics. In group (2), how- 
ever, the westerlies died away quickly to the rear of 
the polar troughs and the easterlies were re-established 
quickly and storms continued in a westerly direction. 
This condition developed when a warm high follows 
or develops to the rear of a polar trough and also if 
the polar trough advances from the west against a 
blocking high resulting in fracture. The tropical cyclone 
usually responds within 24 hr after the easterlies build 
back to 14,000 ft and above. Group (8), the “abnormal” 
tracks, result from a number of causes. A storm which 
has already completely or partially recurved may turn 
back northwestward or westward when blocked by a 
warm high (October 13-16, 1938). A storm recurving 
up a polar trough may be overtaken by the trough, 
become imbedded in a northwesterly current, and turn 
southeastward (October 8-9, 1941). Storms may move 
southwestward when they are overtaken by a narrow 
trough and encounter a deep northeasterly current 
(November 2-5, 1935). Even at the present time suffi- 
cient upper-air information is not available to forecast 
or explain 5-10 per cent of the ‘‘abnormal” movement. 
Forecasting direction of movement, particularly re- 
curvature, is the most acute problem facing the fore- 
caster. Previous checks on the correlation between the 
steering current and actual rate and direction of move- 
ments of tropical storms have not been completely 
satisfactory and, among hurricane forecasters, any 
further research on this problem should have the highest 
priority. 
Forecasting Significance of Storm Swells. The ap- 
pearance of a swell of a particular type may give 
quite reliable indications of a tropical storm as much 
as 500 to 1000 miles or more distant.? The height of 
the waves from which swells develop is determined by 
3. Consult ‘Ocean Waves as a Meteorological Tool” by 
W. H. Munk, pp. 1090-1100 in this volume. 
899 
the force of the wind and the “fetch” or water distance 
over which the wind has blown without significant 
deviation in direction. According to Thomas Stevenson 
[17], the maximum height of a wave as a function of 
“fetch” is H = 1.5\/F, where H is in feet, and F, the 
fetch, is in nautical miles. 
The magnitude of waves is dependent not only upon 
the fetch, but also upon the wind velocity. Over oceanic 
areas with 600-1000 miles or more of sea room, waves 
35-40 ft high are developed in ordinary storms and in 
more intense storms may exceed 45 ft. According to 
Cline [1], the quotient obtained by dividing the wind 
FRONT 
REAR 
Fre. 9.—Schematic development of swells in a tropical 
cyclone: 
A—swells of greatest length and magnitude, traveling in the 
line of advance of the tropical cyclone. 
B—swells and waves of moderate length and magnitude in 
the front segment moving outward to the right and left 
of the line of advance. 
C—swells and waves of smaller length and lesser magnitude 
in the rear segment moving outward to the right and left 
of the line of advance. 
D—swells and waves of least magnitude moving outward 
from the rear of the hurricane. (After Tannehill.) 
velocity (probably average for one hour) in miles per 
hour by 2.05 represents the average height in feet of 
waves developed by the wind. This should be used 
with caution and only as an approximation, since there 
are always other factors to be taken into consideration, 
and a wind, constant in speed and direction (in a 
hurricane at least), does not act on a wave for any 
great length of time. The breaking wave or swell is 
one of the most destructive elements of a tropical 
cyclone, since a cubic yard of water weighs about 
1500 pounds and waves moving forward many feet 
per second may be very destructive to beaches and 
harbor facilities, especially when they contain debris 
such as tree trunks and heavy beams. 
The generation of swells in a tropical cyclone is a 
rather complex phenomenon because of the curving 
wind flow. Therefore, since the wind-generated waves 
move in the same direction as the wind, swells tend to 
emanate in all directions from the center of the hur- 
