TM No. 377 



unity to give credence to the spectrum plot. However, if DF becomes less than 

 10, the confidence limit bands broaden rapidly. 



Before discussing some of the spectra of the individual series of observations, 

 let us examine some typical auto-spectra associated with different wind-sea condi- 

 tions. Figure V-15 shows four spectra of the vertical velocity w measured at one 

 meter below the trough level of the waves. 



A linear ordinate scale is used, in lieu of a log scale, to better show the 

 gross variations of the energy distribution. The frequency scale was cut at 800 

 mcps, since beyond 600 mcps the energy is below an order of magnitude of the peak 

 values. The variances (cm^ sec"2) and wind velocity (cm sec~l) are listed on the 

 graphs for comparison. Values of the estimated height H and a period T of the wind 

 waves are also given. These last two quantities were estimated visually from the 

 BBELS catwalk (some 17 m above the waves), using a "seaman's eye" and a stopwatch „ 

 The estimates must be considered extremely crude and only suggestive of the wave 

 conditions present. 



The spectra in figure V-15 reveal a strong correlation of the auto- spectra 

 and their integrated areas (the variances) with the gross wind speeds and sea 

 conditions. Peaks in the spectra generally occur at frequencies coinciding with 

 those of the visually observed waves. 



By far the most wave energy occurs below U00 mcps (above 2.5 sec periodj . 

 This "cutoff" effect may be caused by the instrument; but, more probably, it is a 

 real effect, since the response time of the impellers was shown (in chapter II ) to 

 be greater than 10U mcps, or less than 0.1 second period. The interpretation of 

 the slope of $(■£) with f near the cutoff range is discussed later in this chapter. 



Another interesting effect is the occurrence of the double peaks in the 06l 

 plot as compared to the single peak in 080, whereas the total energy is roughly 

 equal. In 06l the winds from the WWW were fetch-limited (see figure IV- 3) a^id 

 steadily increasing in speed. Just prior to this WWW wind, a NE storm had passed, 

 generating longer than usual waves from the open sea to the SW. On the other hand, 

 when the 080 observations were made, the SW wind had been blowing many hours, and 

 the seas tended to be fully developed. However, no swell was observed, and the 080 

 spectrum appears to portray fully developed but locally originated wind waves (sea). 



The disproportion between the variances of the u and w velocity components in 

 the waves (discussed in the previous section") shows up distinctly in the plots of 

 the auto-spectra pairs ( (^^ and ^^ ). The curves in appendix B show strong 

 variations in the values of c^^ and (^^ for BBELS- 5 and 7 at ail depths measured. 



In BBELS-5 (serial 001, 002, 005, and 008) the u spectra contain double peaks 

 located at about 250 mcps (k sec) and 400 mcps (2.5 sec), whereas the w spectra 

 have only a single peak at about 200 to 250 mcps. This same effect of a double 

 peaking of ^^ also occurs in most of the BBELS-7 plots. 



107 



