about 1.0 percent, with a 0.1 percent full-scale resolution; full scale is 

 14 m for Gage 625 and 8.5 m for Gage 645. These gages are susceptible to 

 lightning damage, but protective measures have been taken to minimize such 

 occurrences. A more complete description of the gages' operational character- 

 istics is given by Grogg (1986). 

 Buoy gages 



31. Two Datawell Laboratory for Instrumentation (Haarlem, 



The Netherlands) Waverider buoy gages (Gage 630 and 640) measure the vertical 

 acceleration produced by the passage of a wave. The acceleration signal is 

 double-integrated to produce a displacement signal which is transmitted by 

 radio to an onshore receiver. The manufacturer stated that wave amplitudes 

 are correct to within 3 percent of their actual value for wave frequencies be- 

 tween 0.065 and 0.500 Hz (corresponding 15- to 2-sec wave periods). The man- 

 ufacturer also specified that the error gradually increased to 10 percent for 

 wave periods in excess of 20 sec. The results in this report were not 

 corrected for the manufacturer's specified amplitude errors. However, to en- 

 sure that the buoys were within the manufacturer's specifications, they were 

 calibrated semiannually. 



Digital Data Analysis and Summarization 



32. Thompson (1977) and Harris (1974) describe the procedure used for 

 analyzing and summarizing the digital wave data contained in this report. The 

 procedure is based on a Fast Fourier Transform (FFT) spectral analysis of 

 4,096 data values (1,024 sec sampled at 4 Hz) for each file processed. 



33. The analysis program computes the first five moments of the distri- 

 bution of sea surface elevations then edits the digital data by checking for 

 "jumps" and "spikes" and for data points out of the 0- and 5-V range. A jump 

 is defined as a data value greater than 2.5 standard deviations from the pre- 

 vious data value, while a spike is a data value 5 standard deviations or more 

 from the mean. If less than 5 jumps or spikes in a row are found, the program 

 linearly interpolates between acceptable data and replaces the erroneous data 

 values. If more than 5 jumps or spikes in a row or a total of 100 bad data 

 points for the file are found, the program stops interpolating and editing. 



At this point the program analyzes the data and prints a flag indicating that 



2 

 there is a problem with the file. If the variance is less than 0.001 m , the 



24 



