the effects of surface waves were reported by Rossby C1936) and Keulegan 

 C1951), but had not received much attention. Stewart's (1961) sugges- 

 tion received more attention than the earlier suggestions by Rossby and 

 Keulegan, partly because the wave generation theories of Phillips (1957) 

 and Miles (1957) had shown how waves might act as an intermediate step 

 in the transfer of momentum from air to water and partly because a great 

 deal of empirical evidence which could be used to support this hypothesis 

 had been reported in the intervening years. According to one of the 

 latest and most quantitative field studies of wave generation (Hasselmann, 

 et al., 1973), 80 + 20 percent of the momentum transferred across the 

 air-sea interface at short fetches enter. the wave field. About 80 to 90 

 percent of the wave-induced momentum flux from air to water passes into 

 currents through nonlinear transfer processes. However, the inter- 

 pretation of the energy balance is more ambiguous at long fetches. 



Kraus (1972, ch. 5) reviewed the physical processes by which momentum 

 is transferred from the atmosphere to the sea and concluded that Cj 

 is likely to be determined by different processes in different ranges of 

 wind velocities, and that Cj is likely to vary with both the fetch 

 and duration. 



Kitaigorodskii (1970) reviewed much of the recent research dealing 

 with momentum exchange between the atmosphere and the sea. He foimd that 

 when values of the drag coefficient, Cj, are grouped by small ranges of 

 the variable (cc)/Uj^.) , both the mean value and standard deviation decrease 

 systematically with increasing values of (cq/u^j.) . In this expression, Cq 

 is the phase -speed of the waves with maximum energy density. From a phys- 

 ical point of view, this means that the stress coefficient is greatest when 

 waves are growing rapidly, and decreases as the waves approach full growth. 

 Thus, the stress coefficient above an open sea should decrease with increas- 

 ing fetch for reasons and in a way quite different from the decrease observ- 

 ed with rigid boundaries in a wind tunnel. 



4. Summary . 



It was recognized in 1957 that wave generation on a calm water surface 

 is initiated by pressure pulses resulting from turbulence in the airstream 

 above the water. Once the water surface is covered by waves, the roughness 

 of the water surface induces air motions favorable to wave growth. However, 

 the rate of this growth depends on the vertical profile of the horizontal 

 wind near the water surface. 



The turbulence in the airstream responsible for initiation of wave 

 motion also controls the wind profile, and is controlled by the boundary 

 layer phenomena at the base of the atmosphere. 



The later stages of wave growth are the result of interactions between 

 various components of the wave train with little direct influence of the 

 wind. 



The mean momentum which the water derives from the wind is, to a great 

 extent, derived from wind-generated waves, not directly from the wind. 

 Thus, the mean water motions can be related directly to the overlying wind 



25 



