T of 14 sec on 13 November 1981. Table A2 of calibration results for 18 Feb- 

 P 

 ruary 1981 (buoy 66967) gives a difference error d for 14 sec which is 



-0.0779. From Part V the water temperature is estimated to be 12° C. Entering 



Table 10 with the difference error -0.08 and water temperature 12° C, the 



correction is 0.029. This is added to the uncorrected difference error d to 



obtain the corrected difference error D : -0.0489 = -0.0779 + 0.029. The 



corrected difference error (D = -0.0489) is added to the Datawell predicted 



difference error (DW = -0.0220; see Table A2 for T =14 sec), e.g., -0.0709 



P 

 = -0.0489 + (-0.0220), and the sensitivity is computed by adding 1, or 0.9291 



= 1 + (-0.0709). 



12. This sensitivity is used to correct amplitudes and variance spectra 

 coefficients for a l4-sec period. 



Corrected amplitude = Uncorrected amplitude times F(t), or 



0.9291 ^ '^•'^ ^^^» increase) 



and the corrected variance coefficient = 



Uncorrected Variance Coefficient 

 (0.9291)^ 



13. In general, the wave statistics errors are near 5 percent for wave 

 periods less than 12 sec (12 sec is equal to the annual mean plus 1 standard 

 deviation wave period) . Errors of this magnitude are generally tolerable for 

 most engineering applications, although it is worthwhile to know the error 

 bounds for some design considerations. When investigating coastal phenomena 

 involving a very long period swell of 15 sec or greater, such as surf beats 

 and sediment accretion due to swell waves, these corrections will produce 

 significant increases in the magnitudes of the wave parameters. Therefore, it 

 is recommended that the corrections be used. 



A26 



