1058 
from the sea surface were incidental to computing 
oceanic evaporation on the basis of energy consid- 
erations.! In these cases very simple assumptions 
were made as to the magnitude of the ratio R be- 
tween the amount of surplus heat given off directly 
to the atmosphere as sensible heat Q; and the amount 
used for evaporation Q,. Schmidt [26], in his deter- 
minations of the annual latitudinal evaporation from 
the oceans, accepted values for & which ranged from 
about 1.67 to 0.28, but subsequent analysis by Ang- 
strém [2] indicated that these values were much too 
high to represent average conditions over the oceans. 
Angstrém concluded that only about 10 per cent of 
the heat surplus is given off to the atmosphere by 
conduction and that about 90 per cent is used for 
evaporation. Mosby [18], in recomputing the annual 
latitudinal evaporation over all oceans, accepted Ang- 
strém’s supposition and considered F& to be constant 
at 0.10 at all latitudes. McEwen [13], on the other 
hand, assumed F# to be constant at 0.20 when deter- 
mining the annual latitudinal evaporation over the 
North Pacific. 
Up to 1940 none of the investigators had made any 
attempt to evaluate separately the sea-surface and time 
variations in @, or in the ratio R. Nevertheless, 
Bowen [3] in 1926 had derived a formula which pro- 
posed to establish the relation between evaporation and 
heat exchange at a water surface and this formula had 
been applied by Cummings and Richardson [5] in de- 
termining the evaporation from lakes. In considering 
the rates at which heat and water vapor are transported 
across the lower and upper surfaces of a volume of air 
in contact with a water surface, Bowen had concluded 
that the ratio between the heat loss by conduction and 
that by evaporation can be obtained from the expres- 
sion 
et Pp item = ta 
mie a: 1000 . = =I) 
where ¢, and ¢, are the water and air temperatures re- 
spectively, e~ is the vapor pressure of the water surface, 
€. is the vapor pressure of the air, and p is atmospheric 
pressure (all pressures in millibars). Bowen’s derivation 
of this formula is quite involved but the same equation 
has been derived quite simply by Sverdrup [84] and 
the method will not be repeated here. 
Sverdrup [85] subsequently pointed out that the ap- 
plication of the Bowen ratio might be invalidated if it 
proved necessary to consider the effects of the radiative 
transfer of heat through the laminar layer next to the 
sea surface, and if the evaporation from the sea surface 
is greatly increased at wind velocities high enough to 
carry spray into the air. However, Sverdrup [86] has 
more recently presented data to indicate that near the 
boundary surface the radiative transfer of heat is rela- 
tively unimportant. In addition, through a comparison 
between the observed humidity gradients existing above 
the sea surface and those derived on the basis of theo- 
(1) 
1. Consult ‘Evaporation from the Oceans” by H. U. Sver- 
drup, pp.1071-1081 in this Compendium. 
MARINE METEOROLOGY 
retical considerations of the transfer of water vapor 
within the boundary layer, he has also concluded that 
the effect of the evaporation from spray 1s relatively 
unimportant. He states that “The successful applica- 
tion of the theory of turbulence to the problem of air 
mass transformation indicates that there exists no large 
difference in the processes by which heat and water 
vapor are diffused.” 
Although the Bowen formula has suffered extensive 
criticism from meteorologists, the available observa- 
tions to date indicate that it is capable of giving values 
for the ratio Q;/Q. of the approximate order of mag- 
nitude, at least if reasonably correct temperature and 
humidity observations are obtained near the sea sur- 
face and if the computations involve use of data ob- 
tained over extensive water bodies where the effects of 
lateral mixing can be neglected. 
The author [6], following a suggestion given by Sver- 
drup, applied the Bowen ratio to seasonal evaporation 
values computed for five-degree squares over the North 
Atlantic and North Pacific and demonstrated that this 
ratio is a highly variable quantity, both seasonally and 
with respect to the regional distribution. The ratio 
proves to be low, or even negative, in regions overlain 
by dry air and where the temperature differential be- 
tween sea surface and atmosphere is small (as in the 
trade-wind areas). High values are found in regions 
where warm water is overlain by relatively moist (often 
cool) air (as over the Gulf Stream and Kuroshio during 
winter). The average seasonal five-degree zonal values 
are shown to range from —1.5 to +0.6, although most 
values were within the range from —0.4 to +0.5. 
The seasonal and annual values for the rate of ex- 
change of sensible heat with each five-degree square 
over the North Atlantic were then obtained by apply- 
ing the Bowen ratio as follows: 
Q, = RL,E cal em? day, (2) 
where L; is the latent heat of vaporization at average 
sea-surface temperature ¢, and F# is the evaporation 
rate in centimeters per day. The results of the compu- 
tations on the basis of the annual data only are given 
in Fig. 1; the full set of seasonal charts is to appear in 
a later publication [10]. During all seasons the isolimes 
for Q; show, roughly, the same configuration as those 
for EL (or Q.); these values being at their maximum in 
winter and along the western sides of the oceans in 
mid-latitudes. However, one important difference is 
shown. No tropical areas of maximum Q,, appear within 
the trade-wind regions to correspond to the areas of 
maximum evaporation found in those areas. 
The average seasonal values of Q, for the various 
latitude zones are given in Table I. One interesting 
aspect of these data is the fact that the values are 
higher in the North Pacific than in the North Atlantic 
at all latitudes and during all seasons except winter for 
the areas north of latitude 35°N. From the data so far 
accumulated it appears that the atmosphere is being 
directly heated by the sea surface at significant rates 
only in the middle and high latitudes, along the eastern 
sides of the continents and principally during the winter 
