were dominant; but because the fetch in the wind direction was much greater 

 than elsewhere, it is not expected that the [results] will apply to open water." 



Aliasing in the direction al spectrum 

 As shown in figure 1J.26 the computed wave pole spectrum at high fre- 

 quencies is a little high compared to the theoretical Neumann spectrum. The 

 computed energy at frequencies greater than a value corresponding to a k of 

 27.5 is 0.57 (ft) , Some of this energy is aliased in the directional spectrum 

 back into longer wavelengths. The amount aliased is certainly less than 0.57 1 

 because part of the above value is probably white noise and part is correctly 

 located in the corners of the rectangular area of the directional spectrum analy- 

 sis. Only about 0.37 (ft)^ lies above k equal to 31 and hence some part of 

 0.20 (ft) is correctly located. Thus as a very crude estimate something of the ' 

 order of 0.25 (ft) is actually aliased over the directional spectrum. When 

 spread out over a wide frequency and angular range^ this aliased energy is un- 

 detectable because of the sam.pling variation in the higher unaliased values. 



Correction to the cova r iance surface 

 The covariance surface given in figure 11,15 still has errors due to 

 white noise and column noise in it,. The correction for white noise is to sub- 



tract 0.56 (mm^ x 100) (that is. 0.54 x 1.032) from iji(0,0) and 0.191 from the 



I 

 I 



central column (0.186 x 1.032). The result is the estimated covariance surface 

 of the sea plus the swell as shown in figure 11,33, The major effects are to re- 

 duce the peak at the center,, and hence increase the correlation of the edges with 



244 



I 



