Microscopte Structures of Wind Waves 
It is noted that the longitudinal (upwind-downwind) component 
is about one half of the total mean-square sea-surface slope, as short 
ocean waves propagate nearly isotropically. In the laboratory tank, 
the longitudinal component is about the same as the total mean-square 
surface slope, as waves propagate nearly unidirectionally. Needless 
to say, these ratios can only be approximations, because the propa- 
gation of wind waves can neither be exactly isotropical nor be exactly 
unidirectional, 
VI. 2 Sea-surface slope and spectral coefficients. 
Cox and Munk (1956) deduced slopes of the sea-surface from 
the brightness distribution of photographs of sun glitter of the sea- 
surface. The wind velocity at the time of taking sun glitter pictures 
was recorded at two heights, 9 and 41 ft. from the mean sea level. 
The lower height may be too close to the water surface to be free 
from the wave-induced air motion, especially at higher wind veloci- 
ties. Therefore, the wind velocity measured at the upper height was 
used along with the wind-stress coefficient iGr,! formula (Wu 1969), 
=F 0.5" 'U 
Cio dos (a 10 (12) 
to determine from the logarithmic wind profile the corresponding 
wind velocity U at the standard anemometer height. It may be 
worthwhile to note, however, that this correction is very small and 
that the wind velocities U, obtained from both methods are about 
the-same. The results’ of the mean-square surface slope of a clean 
surface s¢ versus U_ are plotted in a semilogarithmic form in figu- 
re 16, Various poumde es layer regimes of the wind (Wu 1968) are 
shown in the same figure. The results of mean-square slopes obtai- 
ned from the interior of an artificial slick are not included in this 
figure. 
It is very obvious from figure 16 that the data are divided 
into two groups: one in the hydrodynamically smooth flow regime 
and the transition region (U,_.< 7 m/sec), and the other in the hydro- 
dynamically rough flow regime (U > 7m/sec). A straightline can 
fit the results in each group rather well. The data are scattered at 
low wind velocities where the wind condition is less stable in the 
transition region, scattering of the data seems to be inevitable. An 
excellent correlation of the mean-square slope with the wind veloci- 
ty is seen in figure 16. 
Most of the results of Cox and Munk were obtained from a 
1451 
