reasonable, because there is likely to be a long-term trend in the 

 motion related to the annual heating cycle. In addition, the spectra 

 show that the dominant periods of thermocline oscillation are 4 1/2 

 days, 24 hours, and 12 l/2 hours. It can be seen that no peaks exist, 

 in either the wind or tide spectra, that do not also appear in the 

 measured depth spectrum. The 4 1/2 day oscillation appears to be 

 accounted for entirely by the wind and is no doubt a direct result of 

 wind transport. The 24-hour oscillation, however, is over-predicted 

 on the basis of wind alone. That is, the diurnal depth cycle as pre- 

 dicted by wind transport is of higher amplitude than that actually 

 observed. This over-prediction appears as the 24-hour peak in the At 

 spectrum in Figure 10. This is compensated for by the 24-hour peak in 

 the tidal spectrum. This result is in complete agreement with the con- 

 clusion reached earlier, that the diurnal thermocline oscillation is 

 diminished by the tide. The 12 l/2 hour oscillation of the thermocline 

 is seen to be entirely an effect of the semidiurnal component of the 

 tide. It was concluded earlier that the internal tide is four to five 

 times as great as the surface tide. This result is also inferred from 

 the power spectra study. If it is assumed that the oscillations under 

 study are sinusoidal in nature, then the spectral ordinates of the 

 various peaks may be set equal to one-half of the squares of the 

 amplitudes. The tidal spectrum in Figure 10 has been multiplied by 

 42, which is the factor necessary to normalize it to the At spectrum. 

 Thus, the spectra also indicate that the internal tide is four or five 

 times as great in amplitude as the surface tide. 



To confirm the 3-hour phase lag between the surface and internal 

 tides, a cross-correlation of the tide data and the At difference data 

 was computed. This correlation supported the original estimate of a 

 3-hour phase shift between the tides. 



APPLICATION 



The information thus far gained as to the nature and magnitude of 

 the thermocline vertical oscillations finds immediate application in 

 predicting the thermocline depth. It has been shown (Fig 4) that the 

 daily average depth is adequately described by a linear relationship 

 involving only the wind. The standard deviation between the two curves 

 of Figure 4 is ± 5 feet. For more detailed predictions, equation (l) 



130 



