longer waves than they actually are. Suppose that m lags are to be used in 

 the X and y directions of the rectangular grid. Then the E value contributed 

 by the waves with lengths from infinity to 4mAx will all be concentrated at the 

 zero wave number of the spectral coordinate system., The next wave number 

 will correspond to a wavelength of 2m Ax and will actually cover a range from 

 4mAx to 4mAx/3. On a line at 45° to the grid system of the spectrum, it is 

 necessary to shorten the above wavelengths by 1/2/2. Finally, if N^ and N 

 are the nunnber of points on the grid system in the x and y directionsj, then 



Wr) 



the number of degrees of freedom is given by 



f = 1.581 -^ -^ 



where for purposes of symmetry it was decided to let m^ = my. 



The significant wave heights reported by the wave pole observations cor- 

 responded to a wind of about 17 knots, and an attempt was made to choose a 

 method of analysis such that a theoretical Neumanjn spectrum. (Neumiann [1954]) 

 for 17 knots would be adequately resolved. Also since winds of 20 knots had 



. i 



occurred previously, it was decided to guard against wavelengths due to a wind 

 of 20 knots in addition to those due to 17 knots. 



It was estinnated that periods from 2.25 to 10 seconds would be present 

 and that approximately 10 percent of the energy would be at frequencies above 

 0.29 cycles per second (or a period of 3.45 seconds). A wavelength of 60 ft 

 corresponds to this period, and hence a spacing of 30 feet between points 

 would be needed to insure no more than 10 percent aliasing. 



54 



