(i.e., total energy or variance spectrum coefficients), the following should 

 be used: 



7. To apply the correction, first the difference error between the 

 Datawell-predicted error and the error measured during calibration is deter- 

 mined. The Datawell-predicted error and the difference error are summed, and 

 the decrease in sensitivity (based on the wave period) is computed by adding 1 

 to the sum. 



8. To demonstrate the use of the calibration results, the Waverider 



buoy recorded an H of 3.7 m and a T of lA sec on 13 October 1984. From 

 m p 



o 



Figure A2 and Table A3 with calibration results for 1 December 1982, buoy 



66967, the d for 14 sec is -0.0635 (interpolating from Table A3). This d 



is added to the DW of -0.0207 (Equation Al minus 1.0), e.g., -0.0635 



+ (-0.0207) = -0.0842 , and the sensitivity is computed by adding 1 or 1 



+ (-0.0842) = 0.9158 . This sensitivity is used to correct the H = 3.7 m , 



m 



T = 14 sec as follows: 



Uncorrected H divided by the sensitivity = corrected H 



m 



o 



3.7 m , „, ,„ , 

 „ Q. CO '^ 4.04 m (o.4 percent increase) 



and the correction for a variance coefficient at this period is applied as; 



Uncorrected variance coefficient 

 (0.9158)^ 



9. In general, the wave statistics errors for 1984 are less than 

 7 percent for wave periods less than 12 sec and less than 10 percent for 

 longer wave periods. Errors of this magnitude are generally tolerable for 

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



A5 



