2.0 THE SCIENTIFIC EXPERIMENT 



A large theoretical literature exists on the possible thermal 

 oscillations of the ocean. In principle, thermal oscillations exist 

 with periods ranging from seconds to eons. A detailed knowledge 

 of the precise behavior of the deep ocean temperature structure is 

 fundamental to questions as diverse as baroclinic instability, air- 

 sea interaction, and the energy dissipation of the moon in the ocean, 



A handful of observations have been made of temperature fluctua- 

 tions in other than coastal waters. A useful summary is contained 

 in Ichiye (Ref. 2). Most of these observations were made from 

 ships, and of necessity were of short duration. It has been assumed 

 that the observed oscillations were due to internal gravity waves. 



On the continental slopes, observations of relatively long durations 

 have been obtained (e.g., from the NEL tower). It seems unlikely 

 that these are representative of deep ocean conditions. 



Array configurations are required for long-term, large-scale thermal 

 observations, since internal waves constitute a 3-dimensional field 

 phenomenon, propagating with specific wave numbers and frequencies 

 through a time-dependent propagation medium. Measurements then must 

 be made from 3-dimensional arrays sufficiently dense to maintain 

 coherence across the array and occupying a sufficient volume of 

 space to have directional resolution. The quasi-random nature of 

 the temperature field demands long-time series of measurements to 

 obtain meaningful statistics. 



The design criteria for sensor stability rule out buoy techniques. We 

 have used instead a miodification of a method due to Stommel, who 

 maintained a cable on the Bermuda slope for 9 months in 1954. Two 

 thermistors were placed on the cable at a depth of 50 and 500 m., on 

 the bottom, about 1 mile apart. The results, reported by Haurwitz, 

 Stommel and Munk (Ref. 8), are the longest deep-ocean measurements 

 that have been made. 



From an array designer's point of view, the results from the cable 

 are discouraging, showing relatively featureless spectra and nearly- 

 zero coherence between the two sensors. In an effort to understand 

 the significance of these results, and to determine the distance over 

 which 2 sensors would be coherent, another experiment at Bermuda 

 was performed in September, 1966, by Wunsch of MIT and Parker 

 of Woods Hole. 



195 



