usually necessary to apply a high frequency cutoff, above which the pressure 

 contributions are discarded. The proper selection of this high frequency 

 cutoff depends on the signal to noise characteristics of the pressure sensor 

 and the signal conditioning system. In the present program, the high fre- 

 quency cutoff was established at a wave period of 3.0 seconds. Wave gage 

 analyses by Thompson (1980) have shown that a 3.0-second high frequency 

 spectral cutoff value provides reasonable estimates of total wave energy at 

 west coast (U.S.) locations. 



Denoting hereafter the FFT coefficients for the water surface as a(n) 

 and b(n) , it is noted that the coefficients have the following properties: 



N-i 



<n 2 > = I [a 2 (n) + b 2 (n)] 



and 



and thus 



n=i 



a(|+ n)-a(|- n) 



N/2 

 <n 2 > = 2 I [a 2 (n) + b 2 (n)] 

 n=i 



(9) 

 (10) 

 (11) 

 (12) 



Thus, the total (kinetic and potential) energy E(n) associated with a par- 

 ticular wave frequency component, n, is 



E(n) = 2y[a 2 (n) + b 2 (n)] 



(13) 



Now consider two wave or pressure sensors located at (x , y ) and (x , y ) 

 (see Fig. 1). The results will be developed considering discrete frequencies. 



/3 - Beta 



( paral lei to shoreline 



^ 



<^ ( normal to bottom contours ) 



Gage 1 ( x, , y, ) 



Figure 1. Definition sketch for two sensor array. 



10 



