SECT. 2] LARGE-SCALE INTERACTIONS 205 



it was being transported vertically at rates of up to +1 m/sec. After another 

 152 sec, the smoke loop at B in Fig. 49b had descended into contact with the 

 sea surface. 



Using these photographic suggestions, we may hypothesize the exchange 

 mechanism. Let us look at a single eddy, roughly the size of the smoke loop. 

 As illustrated schematically in Fig. 50, it first descends to the ocean surface, 

 stays there for a little while and then rises again. Let us suppose that the air 

 on the downward leg is about 0.5-1.0°C colder than the water, as is frequently 

 true. Then the air will be warmed by contact with the ocean and ascend with a 

 temperature a fraction of a degree higher than it had during the descent. It has 

 thus received an increment to its sensible heat. 



One generally assumes that the vapor pressure (or specific humidity) at the 

 ocean surface corresponds to the saturation vapor pressure (or specific humidity) 

 at the temperature of the surface. The descending air is cooler than the water 

 and its relative humidity is generally less than 100%, perhaps 70-80%. Thus 

 the vapor pressure of the eddy will be lower than that of the water and a little 

 evaporation takes place from sea to air. The rising air will carry away with it a 

 slightly larger amount of water vapor than it brought down ; its latent -heat 

 content has been increased. 



Still another type of energy exchange takes place. The air undergoes frictional 

 braking at the surface, and also, therefore, a decrease of its kinetic energy. This 

 decrease is so small compared to the sensible and latent heat gains that, from 

 the energy viewpoint, it is negligible. For instance, a reduction of wind speed 

 from 7 to 6 m/sec produces an energy decrease equivalent to about 2 cal in heat 

 units per kg of air; a temperature rise of 0.1°C for the same mass corresponds 

 to an energy increase of 20 cal and a moisture increase of 0.1 g to 60 cal! From 

 the momentum viewpoint, however, the slowing down is very important. As 

 we saw, a reduction of wind speed in the trades represents a loss of easterly 

 momentum or, as frequently put, the air receives westerly momentum from the 

 sea. For the ocean, in turn, this momentum input humps up the boundary into 

 waves and sets the surface layers in motion. 



Thus, when the ocean temperature exceeds that of the overlying atmosphere, 

 we should observe that air leaving the surface has a slightly higher temperature 

 and moisture content, and a little lower wind speed than when it approached 

 the surface. These predictions need testing by measurements, and if such sized 

 motions are effecting the transfers, we need to know how they are maintained. 

 Of course, it is difficult to follow an individual eddy with measuring equipment. 

 We can, however, determine whether instruments recording temperature, 

 humidity and wind on board a ship reveal fluctuations of these magnitudes as 

 the air moves past ; further, we can find how these fluctuations are correlated. 



In 1946, a second expedition from the Woods Hole Oceanographic Institution 

 went to the Caribbean to study trade cumuli and the layer below the clouds, 

 this time using both an instrumented ship and aircraft (Wyman et al., 1946; 

 Bunker et al., 1949). The normal excess sea temperature over that of the air at 

 ship's deck level (~6 m) was found to be 0.2-0.5°C, with an extreme range 



