910 
followed by convergence at the ground and ascent of 
the low-level air in the initial tropical depression. 
4. The vertical cross-stream circulation described is 
“direct,” therefore kinetic-energy producing, as the 
air is less dense in the tropical disturbance where ascent 
takes place than in the surroundings. Maintenance of 
the vertical circulation is accomplished by continuous 
advection of more polar air at the left of the upper cur- 
rent and moist-adiabatic ascent at the right, looking 
downstream. 
INITIAL DIV, — INITIAL CONV. 
MOIST 
ADIABATIGC 
ASCENT 
SUBSIDENCE 
ENFORCED p FALL AND DIV. OF 
AND CONV. 
OLD POLAR AIR 
Fic. 6.—Vertical cross section taken eastward from position 
of tropical depression shown in Fig. 4. Model illustrates hypo- 
thetical development of vertical cross-stream circulation lead- 
ing to deepening of the surface depression. 
5. Several energy sources have been used in creating 
and maintaining the direct vertical circulation cell. 
The in-phase superposition of high- and low-latitude 
wave trains in the upper troposphere directly intensi- 
fies the speed of the high-level current above the surface 
depression. Use has also been made of old polar air, but 
in a manner quite different from that implied in the 
frontal theory of hurricane formation. The polar air 
does not enter the forming storm but moves equator- 
ward at some distance and subsides. This subsidence is 
partly responsible for creating an organized circulation 
cell. 
6. The existence of the direct circulation, initiated 
by the large-scale synoptic picture, imecreases the effi- 
ciency of the convective motion in the surface disturb- 
ance in generating kinetic energy. A priori there is no 
preferred direction for motion produced from conden- 
sation. The small-scale circulation cells initiated im 
different parts of a disturbed area tend to cancel each 
other. Introduction of the large-scale circulation cell, 
however, aligns the smaller convection cells in a definite 
direction. The air that rises in the zone of convection 
is removed a considerable distance before it sinks again. 
At this time it is not possible to tell whether this 
latest model will withstand the test of experience. The 
sequence described is not the only one that can lead to 
cyclogenesis. Both in high and low latitudes, more than 
one synoptic pattern can lead to deepening. It is fair to 
say, however, that it is reasonable to attack the problem 
by placing emphasis on the upper divergence as an 
initiating mechanism. All efforts to explain deepening 
from low-level considerations have failed to date, since 
none of them can explain the surface pressure falls. 
The future evidently must concern itself with further 
TROPICAL METEOROLOGY 
development of models that lead to cyclogenesis. It is 
probable that the interrelation between different lati- 
tude belts will ultimately be recognized as an important 
factor. If that holds true, there is also some hope for 
preparation of extended forecasts of hurricane forma- 
tion. The interrelation between high and low latitudes 
is a function of the state of the general circulation. If it 
should become possible to predict longer term trends of 
the general circulation, success in forecasting hurricanes 
on an extended basis could also follow. 
MOVEMENT OF HURRICANES 
This is the most widely discussed aspect of tropical 
forecasting, certainly one of the most vital. Contrary 
to the topic of formation, recognition of the importance 
of external forces for predicting the motion of storms 
dates from the last century. The close relation between 
storm tracks and the position and intensity of the sub- 
tropical highs (steering) was recognized very early 
[16]. Mitchell [21] was the first to provide an extensive 
set of rules for forecasting recurvature, based largely 
on the effect of travelling surface lows and highs in the 
westerlies on the tropical storms. Dunn [12] used similar 
reasoning, but applied it to 10,000-ft streamline charts. 
His attack was extended by Riehl and Shafer [33], who 
sought to determine under what circumstances hurri- 
canes that meet a polar trough will recurve and when 
they will bypass this trough and continue westward. 
The height of the base of the polar westerlies played the 
principal role in their argument. 
Recently, Simpson [88] introduced the ‘warm 
tongue” steering method, which from the dynamical 
point of view is similar to that of Dunn [12] and Riehl 
and Shafer [33]. Mintz [20], and Moore [22] applied the 
ideas of Bjerknes and Holmboe [2] to make quantita- 
tive forecasts of displacement. This is an interesting 
proposal and has led to some good forecasts. A purely 
dynamical approach is that of Yeh [44], who at first 
computes the path of a hurricane situated m an easterly 
current and then superimposes a general meridional 
component. He finds that storms execute oscillations of 
minor but definite amplitude while in the easterlies. 
There is, however, no total deviation of the storm track 
to the right of the steering current as maintained in 
some previous studies. When a storm comes under the 
influence of a polar trough, a northward displacement 
occurs. But this displacement will vary widely, de- 
pending on certain initial conditions, when the polar 
trough appears. 
Riehl and Burgner [31] made a quantitative test of 
the steering principle. The method used was applicable 
to the zonal component only. An area 90 degrees of 
longitude long and 5 degrees of latitude wide, centered 
on the storm, was defined as the region of influence. 
Within this area, the mean zonal component of motion 
at 700 mb was computed and correlated against the 
24-hr zonal component of storm displacement. The re- 
sulting scatter diagram showed that in the mean the 
zonal components of storm movement and steering 
current are equal. Numerous deviations, however, oc- 
curred and these were relatively large in that portion 
