SECT. 5J INTERNAL WAVES 757 



6. Internal Waves and Turbulence 



Internal motions in the sea are not necessarily due to free internal waves but 

 may also be due to forced waves or have the character of less well organized 

 motions such as convection of turbulent eddies past the measuring apparatus. 

 Indeed, the distinction between internal waves and turbulence is an arbitrary 

 one, for internal waves, if grown too large, must break and dissipate their 

 energy into eddying motion, while turbulent motions which involve vertical 

 oscillations must be coupled to internal waves, particularly when the turbulent 

 eddies become too weak to turn over. 



It is possible, however, to make a practical distinction on the basis of the 

 transport of energy. Turbulent eddies, whether due to convection of heat or 

 shear instability of ocean currents, transport energy in the form of kinetic 

 energy of rotating water-masses. The speed with which the energy moves is, 

 therefore, limited to the speed of the flow which transports the masses. Internal 

 waves, on the other hand, transport energy at a group velocity which is neces- 

 sarily dijfferent (and usually much greater) than the speed of the ocean currents 

 within which they travel. Unlike turbulent eddies, waves would be expected to 

 preserve coherence over a considerable distance. 



A . Coherence of Isotherm Fluctuations at Separated Localities 



A distinction between fluctuations due to internal waves and other less 

 regular causes can, therefore, be made only if observations are made at more 

 than one position in the sea. UflFord (1947) appears to have been the first to 

 make observations of this sort. He studied internal waves by making repeated 

 lowerings of bathythermographs from bow and stern of a ship and from three 

 separated ships. In all cases, involving a total of a few hours of observing time, 

 he found similar fluctuations at each station with a phase shift appropriate to 

 the phase velocity of internal waves. 



Much longer series of isotherm-depth fluctuations have been measured at three 

 positions on and near the NEL oceanographic tower by use of isotherm followers. 

 Two days of data have been analyzed by statistical methods (Fig. 25). 

 The sea bottom is practically level at 18 m depth within the observing region 

 and featureless off-shore. Therefore it is to be expected that the waves 

 which were incident from seaward would be uniform statistically at the three 

 stations. 



The observed spectra of depth fluctuations were the same (within expected 

 statistical fluctuations) at each station ; only their mean value is shown in 

 Fig. 25. The spectrum shows the same monotonic decrease with frequency as 

 the temperature spectrum from Castle Harbor. (At frequencies above 0.4 cycles 

 per minute [c/min] the spectrum becomes flat. This presumably shows the 

 effect of a random recording error of 0.12 m r.m.s.) 



Vaisala's frequency in the water column had a maximum of 0.65 c/min near 

 the surface, decreasing to 0.15 c/min near the bottom. The analysis extends up 

 to 0.5 c/min but significant coherence is detected only at frequencies well 

 below the mean value of Vaisala's frequency. Under these conditions the phase 



