Since the wind speed makes the major contribution in describing the 

 frequency spectriim of wind generated sea, considerable work has been 

 done on the problem of wind observations. The variation of wind with 

 height over the sea has been studied and wind profiles obtained based 

 on different drag coefficients. When these profiles are studied, it can 

 readily be seen that the variation of wind with height is logarithmic. 

 An equation for determining wind at various heights is ' 



'^ z, 



where 



K =0.4 (von Karman constant) 



C^ = drag coefficient 



jj^ z height at which speed v 2 is to be determined 



In general the wind speed at a given height is not linear with some other 

 level because the drag coefficients are also functions of wind speed. 

 For example, the Sheppard drag coefficients are given by 



(8) ca = (0.80 + o.iii^V)io2 



I^/hether or not the drag coefficients change with height has not yet been 

 determined. If the air has neutral stability, the coefficient would 

 generally be constant with height and the wind profile purely logarithmic. 

 If the air were stable or unstable, such might not be true. 



The large variability of vnnd observations from ships may be caused 

 by many factors, such as the presence or absence of anemometers, the 

 differences in the heights of the anemometers, the length of time used 

 to average the wind, etc. There is a need for some consistent standard 

 height for wind measurements or correction to a standard height as is 

 done for pressure measurements so that some of the wind variability may 

 be eliminated. 



From calculations using the preceding equations together with the 

 proper relationship between significant wave height and wind speed, it 

 was possible to draw curves of significant heights versus wind speeds 

 for wind speeds at various elevations (Fig. 10). These curves show that 

 the lower the elevation the smaller the wind speed required to produce a 

 wave of a given height. 



In comparing the various theoretical spectra that have been proposed 

 there have been considerable differences among them. It is quite likely 

 that some of these differences may be caused by lack of attention to the 

 variation of wind speed with height. When this factor was taken into 

 account the various spectral models did not differ nearly as much. A 

 very small change in wind speed leads to considerable changes in wave 

 heights computed from the spectrum, in the total energy for fully- 

 developed sea (3; 4)^ and in the location of the spectral peak. These * 

 small changes in wind may easily be present if the observed wind is at 



?3L 



