alongshore traverse shows few peaks. The two lower frequency 

 peaks (fig. 20) are for 10. and 19. 2-minute periods (or 1. and 

 1. 9 miles), which are relatively long periods when compared to 

 the 9°C isotherms. 



The power spectrum computed for the 16 °C isotherm for the 

 onshore leg shows low peaks which occur between 3. 2 and 22. 2 

 minutes (0. 3 to 2. 2 miles) (fig. 21). 



2# 1 

 The degree of freedom v =— — - — . When using 1500 depth 



Ti 2 



sample values and 150 lags, u = 19.5. From reference 12, the 



ratio of measured to average value falls between 0. 54 and 1. 60 for 



a 90 per cent confidence limit. 



Thus the 7. 2 minute period peak of the offshore 9°C isotherm 

 depth (fig. 17) has a ratio to background of 58 to 31 or 1.87 which 

 is significant, whereas the onshore 14. 7 minute period peak shown 

 in figure 19 has a ratio of 1. 65 which is barely outside the 90 per 

 cent confidence limit. 



The difference in power spectrum, if caused by a dominant 

 internal wave moving in one direction, sheds some light on which 

 way the waves are moving. The greatest power is concentrated in 

 the long-period fluctuations at the beginning of the spectrum. 



Comparing the power spectra with the corresponding auto- 

 correlation curves indicates that when the autocorrelation curve is 

 irregular the power spectrum shows a greater number of peaks. 

 However, any possible wavelength shown on a correlation curve is 

 at the low frequency end of the power spectrum diagram. 



DIRECTION OF PROPAGATION 

 OF INTERNAL WAVES 



Little is known of the propagation direction of internal waves 

 in deep water. It is reasonable, however, to assume that internal 

 waves do propagate in one or, more likely, many directions. 

 Since the ship is moving at 6 knots with reference to the sea 



35 



