TM No. 377 



Finally, with respect to the relationship of the two spectra associated 

 with the spectral smoothing of averaging function, as observed from actual 

 plots, the spectrum analyzer seems to over-estimate the values of the spec- 

 tral densities by a factor of "about 3"» 



Fourteen wave height auto-spectra made with the CERC Laboratories analog 

 computer were compared with spectra of wave particle motions recorded over 

 the same period (this series was BBELS-12 recorded on 25-26 May 1965). 

 Comparison of the analog wave height data with the digitally derived spectra 

 was difficult because of differences in frequency bandwidth studies. The 

 band of spectral estimates defined by having £T=0'2second and m=50 gives a 

 spectral range of 0-2500 mcps, whereas the bandwidth of the analog spectra 

 was roughly 0-300 mcps. This lower frequency range was chosen because the 

 CERC workers are more interested in resolving low-frequency swell than wind 

 waves The spectral curves were converted from English to metric units and 

 the appropriate factors (arrived at by Mobarek) were applied to produce the 

 plots shown in the appendix E (figures WS-1 to WS-14). 



The absolute values of the spectra may be somewhat inaccurate., since there 

 appears to be some arbitrariness in deciding the factors for conversion from 

 analog to digital spectra* However, the curves do permit a rough comparison 

 of the position of the maximum spectral energy. The shifts in the spectra 

 can be compared with spectra variations depicted in the spectral records of 

 BBELS-13 . 



For a direct comparison of the Tukey analysis with the CERC analog process, 

 two wave staff records on paper strip charts were obtained and processed at 

 the NUWS computer laboratory (using a AT" = 0*2 second). The auto-spectra 

 curves are shown in appendix B in figures WS-1 5 and WS-16. The spectra are 

 also listed in appendix B. The relation of this data to the wave motion is 

 discussed in chapter V. It was planned to obtain the CERC spectra of these 

 two records, but it was learned later that the magnetic tape system was inop- 

 erative while the paper tape records were being made. Hence^ a direct simul- 

 taneous comparison of the two specira was not possible, 



9h 



