LARGE-SCALE ASPECTS OF ENERGY TRANSFORMATION OVER THE OCEANS 
known temperature gradient. For these reasons the 
equations are not in convenient form for application 
in the large-scale climatological sense. This statement 
is not to be construed as a criticism of the method, 
however, since it has been designed for use by forecasters 
and not for the evaluation of large-scale energy trans- 
formation processes over the oceans. Nevertheless, the 
method might prove invaluable for investigations, on 
a limited areal scale, of the synoptic aspects of heat 
transfer between sea surface and atmosphere. It might 
also be applied in the climatological sense within those 
peripheral ocean areas where, during one or more of 
the seasons, air trajectories remain fairly constant and 
the initial continental air-mass characteristics do not 
show great variation (as, perhaps, over the Sea of 
Japan in winter). 
RATE OF EXCHANGE OF ENERGY IN THE 
LATENT FORM OF WATER VAPOR 
Of the total solar energy absorbed at the sea surface 
during the course of a year, approximately fifty per 
cent is used for evaporating sea water and, therefore, is 
made available to the atmosphere in the latent form 
of water vapor.‘ In view of this, and considering the 
relative magnitudes of land and sea areas, it would 
appear that the latent energy represented by water 
vapor derived from ocean evaporation constitutes the 
most important single component of the atmospheric 
heat budget. It is obvious that any paper which pro- 
poses to deal with atmospheric energy transformation 
processes should give primary attention to the problem 
of ocean evaporation. The fundamental details con- 
cerning this problem are so well covered by another 
article in this Compendium! that it will not be neces- 
sary here to expound the theoretical aspects of the 
processes governing the transfer of water vapor from 
sea surface to atmosphere. For this reason the present 
discussion will be limited to consideration of the large- 
scale and seasonal aspects of ocean evaporation in terms 
of its energy equivalence. 
It has long been known that evaporation is governed 
by atmospheric humidity, surface water temperature, 
and wind speed. Nevertheless, empirical methods which 
have been used in the past to relate evaporation to 
these factors generally have not been successful, either 
because the equations contained unevaluated functions 
or because they were constructed to fit a special limited 
set of observations. In addition, the several evaporation 
equations derived on the basis of theoretical considera- 
tions have given equally divergent results, and no 
single theoretical method for computing evaporation 
has yet been devised which has general application 
when use is made of temperature, humidity, and wind 
data obtained under a variety of observational circum- 
stances. 
4. According to Mosby [18], 41 per cent of the total solar en- 
ergy absorbed at the sea surface between the 70th parallels 
is radiated directly back to space, 53 per cent is used for evapo- 
ration and 6 per cent is conducted to the atmosphere as sen- 
sible heat. However, on the basis of the author’s data, it ap- 
pears that the last figure should be revised upward. 
1061 
This state of affairs led Sverdrup in 1940 to suggest 
to the author that it might be possible to compute the 
average evaporation over the oceans, using available 
marine climatic data, by applying an equation of the 
type: 
E = K (€w — €a) Wa, (4) 
where K is an empirical ‘evaporation factor” arrived 
at by comparing the long-term annual ocean evapora- 
tion computed through the use of the energy equations 
with that computed for the same period through the 
use of the several existing equations which involve 
theoretical considerations of the interchange of water 
vapor within the turbulent boundary layer near the 
sea surface. In equation (4), e» is the vapor pressure at 
the sea surface, @ is the vapor pressure at height a 
above the sea surface, and W, is the wind speed at 
height a. 
In following this suggestion, the mean annual evapo- 
ration was computed for each of four selected areas in 
the North Atlantic and North Pacific by two methods: 
first, by using an energy-budget method similar to that 
employed by Mosby [18] and arriving at values desig- 
nated as H,, and second, by using the evaporation 
equations of Sverdrup [31] and Montgomery [16] and 
arriving at values for K (and #). Since, as a first step 
in computing evaporation by the energy equations, it 
was necessary to disregard the oceanic heat advection 
term, the areas selected were those within which the 
latitudinal transport of surface waters is at a minimum 
and, at the same time, those allowing a rather wide 
sampling of latitudinal differences in available solar 
energy (Table III). The marine climatic data which 
entered into the computations were obtained from U.S. 
Weather Bureau sources [88]. 
Of the two theoretical equations which were used to 
arrive at the final values for K in equation (4), the one 
presented by Montgomery is in somewhat simpler form 
than that presented by Sverdrup and for that reason 
it will be exemplified in the following discussion. Ac- 
cording to Montgomery, 
EH = pkoyala Qu — Gh) Villes (5) 
where p is air density, ky) = 0.4 is the Karman constant, 
va the resistance coefficient, I, the evaporation coefli- 
cient, gw the specific humidity at the sea surface, and 
da the specific humidity at height a. With all quantities 
in cgs units, the evaporation is expressed in grams per 
square centimeter per second. It should be pointed 
out that equation (5) does not deal with instantaneous 
values of humidity and wind speed but probably ap- 
plies to these quantities averaged over a period of not 
less than an hour. 
The coefficients y. and I, in equation (5) depend upon 
the height a at which humidity and wind speed are 
measured, and also upon the character of the sea 
surface. The sea surface can be considered hydrody- 
namically smooth at wind speeds less than about 650 
cm sec~! (measured at a height of 600 em) and hydro- 
dynamically rough at higher wind speeds [19, 24, 25]. 
If averages for a smooth surface and maxima for a 
