398 
in the Gulf of Mexico and have a strong northerly 
component. 
Tropical storms in July are only slightly more fre- 
quent than in June and are slightly larger, but none 
of the outstanding hurricanes have occurred in this 
month. Most develop just to the north or to the south 
of the Antilles and have a more westerly component. 
A marked increase in the frequency and intensity 
of tropical cyclones takes place in August and some of 
the most severe of record have occurred in this month. 
Many form in the Cape Verde region and travel across 
the entire Atlantic. August storms have strong westerly 
components and the most regular paths of any month. 
—F = : 1 
NY ¢ 
\ SEPTEMBER 
-® 
SKS 
cayema ss 
leas 
Sew Se een 
SS 2s pom 
=— fo 
= LAST HALF OF 0. 
SEPTEMBER 
BER 
‘ai 
Ll 
OCTO 
Fic. 8—Continued 
The height of the hurricane season is reached in the 
first half of September. During this period many still 
develop in the Cape Verde region. During the last 
half of the month the area of most frequent origin 
shifts back to the southwestern Atlantic and storms 
again develop strong northerly components as the polar 
westerlies move southward. 
October continues to have a large number of tropical 
storms but the frequency and average intensity decline 
rapidly after mid-October. All have strong northerly 
components and recurve quickly. Many develop in 
the western Caribbean north of Panama and move 
northward across Cuba. Rarely do they reach as far 
west as Texas. 
November is an unimportant hurricane month. The 
few storms mostly develop in the Carribbean and move 
northward immediately. 
TROPICAL METEOROLOGY 
The stronger northerly components in the early and 
late portions of the tropical cyclone season and the 
more westerly component during the middle of the 
season are characteristic of all areas with the possible 
exception of the Bay of Bengal and the Arabian Sea. 
Of course, many exceptions to the mean tracks occur, 
depending upon the mean position of polar troughs 
which influence recurvature. 
Steering Currents. The direction and rate of move- 
ment of a tropical cyclone are dependent upon the 
direction and speed of the air in which the storm is 
imbedded. There is no physical basis for believing the 
current at any single level is responsible for steering a 
storm, but Norton, Riehl, and others have found that 
the winds at or just above the top of the warm lower 
cell of the storm, where the vortical circulation vir- 
tually disappears, best approximate the direction of 
movement of the storm. This level will vary in indi- 
vidual storms from 20,000 ft to 30,000 ft or higher. 
It has been noted that most storms appear to have 
some deflection to the right of the steerig current. 
The angle varies with the speed of movement of the 
storm, being as much as 20 degrees with speeds under 
20 mph. 
At times the pressure situation may be changing 
rapidly at the steering level and the steering current 
may change materially with time. Wide-scale changes 
of quarter-hemisphere dimensions may be taking place 
with resulting strong anticyclogenesis upstream, as in 
the Atlantic hurricane of September 17, 1947. Such a 
development may, in turn, alter the steering current 
over a large area within a 24-hr period. 
Hurricane forecasters in the United States believe 
tropical cyclones usually move at a rate of 60-80 per 
cent of the velocity of the winds at the steering level, 
and that winds ahead of the storm are more repre- 
sentative than those to the rear in computing its move- 
ment. Some forecasters have indicated that the 
relationship between rate of movement and the 20,000- 
ft winds ahead of the storm is always better than the 
70 per cent value. 
Simpson [16] has developed a technique which he 
called ‘““warm-tongue steering” for forecasting the di- 
rection of movement of tropical cyclones. This method 
gave good results when tested on 25 storms during 
the period 1941-45. An analysis of the mean virtual- 
temperature field between the 500- and 700-mb pres- 
sure surfaces indicated that a tongue of warmer lighter 
air is associated with and extends some 800 to 1200 
miles in advance of the tropical cyclone. The tests 
seemed to indicate that a good lag-correlation exists 
(for 24 hr and often 48 hr) between the orientation of 
the warm tongue and the future movement of the 
storm. This correlation was least effective in small 
immature storms. A theoretical explanation of this 
relationship is not apparent. It is understood that 
unpublished tests by other groups fail to show sub- 
stantial agreement with Simpson’s results. 
The problem of recurvature is the most difficult one 
facing the forecaster, since it takes place frequently 
within 24-hr striking distance of the coast line and over 
