THE PHYSICAL BASIS FOR THE GENERAL CIRCULATION 
atmospheric kinetic energy as well as preferred regions 
for its disappearance by conversion into other forms of 
energy through friction or otherwise. 
With the aid of the examination and interpretation 
of the fundamental dynamical principles relating to 
kinetic energy given elsewhere in this volume® we shall 
now attempt to discuss this phase of the energy prob- 
lem. In the reference mentioned there is presented a 
discussion of the process whereby kinetic energy is 
produced in the atmosphere. According to the views 
expressed there, kinetic energy of large-scale motions 
can be generated only through the work of expansion 
by pressure forces. Furthermore, kinetic energy asso- 
ciated with horizontal motions can be generated only 
through work done by horizontal pressure forces. It is 
thus pointed out that the rate at which horizontal 
kinetic energy is generated at a given point in the 
atmosphere is equal to the product of the pressure into 
the horizontal divergence of velocity. This carries the 
implication that regions of horizontal convergence are 
hydrodynamic sinks for kinetic energy which act in- 
dependently of friction. 
Speaking next of the transport of horizontal kinetic 
energy, it has been shown that this process is ac- 
complished either through the work done by the pres- 
sure forces in virtue of the horizontal velocities across 
the boundary or by the advection of sensible kinetic 
energy. Since the latter (meridional) transport is small, 
it follows that the significant (meridional) transport of 
kinetic energy is accomplished through the work done 
by the pressure forces, which im turn is proportional to 
the advection of internal heat energy and occurs in ad- 
dition to it. Applying these ideas to the transfer of 
kinetic energy across the vertical boundaries of an 
equatorial strip between two fixed middle latitudes 
+ and —4, our general information would lead us to 
suppose that there exists a net transport of internal 
heat energy and therefore of kinetic energy poleward 
across these boundaries. Estimates from data tend to 
confirm this supposition, and indeed it also appears 
reasonable from common synoptic experience. We are 
therefore confronted by the very important conclusion 
that regardless of the nature of other details of the 
atmospheric circulations the more tropical regions ap- 
pear to be preferred regions for kinetic-energy genera- 
tion in that they not only produce sufficient kinetic 
energy to overcome frictional losses within these 
regions, but also provide an excess which is transferred 
to the polar caps. 
Apparently we may also conclude that in the long 
run the more polar regions must serve as preferred 
regions for the disappearance of kinetic energy, since 
here the losses must be sufficiently great not only to 
offset whatever amounts of kinetic energy are generated 
locally but also to absorb the kinetic energy which is 
supplied from the more tropical regions. It is therefore 
reasonable to suppose that in the normal state of af- 
5. Consult ‘‘Applications of Energy Principles to the Gen- 
eral Circulation” by Victor P. Starr, pp. 568-574. 
545 
fairs in the atmosphere there are vast amounts of 
kinetic energy generated in the more tropical regions 
and vast amounts consumed in the polar caps. The 
existing kinetic energy is merely a small difference due 
to the fact that the positive generation process in the 
more tropical regions is slightly larger than the losses 
in regions of negative generation in the polar caps. 
It is a matter of interest to examine the import of 
the views expressed above for the nature of the secon- 
dary circulations of the middle latitudes. It would thus 
be suggested that the kinetic energy supply for the main 
cyclone belt of each hemisphere is perhaps provided by 
the poleward flow of such energy from more tropical 
regions. Furthermore this flow is measured by (but 
exists in addition to) the poleward flow of internal 
heat energy. Although this flux of energy has some 
average mean value, there can be no doubt that in its 
details the process is a sporadic one with irregularities 
both im time and in longitude. Hach cyclone may thus 
be considered as producing a spurt or poleward surge 
in this transport locally during its period of develop- 
ment. 
According to the classical view, the kinetic energy for 
a developing cyclone is obtained from the sinking of 
dense masses of cold air in the immediate vicinity, so 
that by and large the decrease in potential and internal 
energy represents a corresponding increase in kinetic 
energy. From the present viewpoint the kinetic energy 
increase could equally well be derived from a pro- 
nounced local poleward flow from the more tropical 
regions, since a developing cyclone is accompanied by a 
marked increase in the net local poleward transfer of 
internal heat energy. 
As a matter of fact it is not too difficult to visualize 
the general outlines of a certain type of instability 
associated with a developing cyclone when viewed in 
this way. For if it is granted that there exists a supply 
of warm and cold air in the vicinity, it may be that the 
pronounced local surge of kinetic energy once started 
serves to increase the intensity of the cyclone which in 
turn further increases the flow of kinetic energy and 
the intensity of the cyclone. The intensification finally 
ceases when complete occlusion takes place and the 
net heat transport disappears. 
In a recent discussion the writer [11] has made an 
endeavor to elaborate further this picture of the kinetic 
energy balance of the atmosphere, making use of the 
assumption that mean meridional circulations do not 
play a significant role in these processes. The plausi- 
bility of this supposition is supported by the study of 
the angular momentum balance outlined earlier. Under 
such circumstances, since data show that near the sur- 
face poleward-moving individual air masses possess a 
higher specific volume than the equatorward-moving 
ones, it follows that there should exist a mean horizontal 
velocity divergence in the more tropical regions and a 
mean horizontal velocity convergence in the polar caps. 
This situation should thus contribute to a net pro- 
duction of kinetic energy, in view of the normal mean 
meridional distribution of pressure along horizontal 
