mixing. Ih the li[-;ht oi' these facta, it would seem that, based 

 on our present knowledpje of the present processes Involved, a 

 precise empirical prediction of the thermocline depth is not now 

 possible . 



However, it is possible to make approximations. One of the 

 things that we know about the diurnal and semidiurnal oscillations 

 is that their amplitudes are dependent to a large extent on the 

 depth at which these oscillations occur. For approximation 

 purposes, therefore, the problem may be devided into two parts: 

 (a) to predict the mean depth about which the thermocline oscill- 

 ates daily and (b) to predict the amplitude o£ the oscillations 

 about this depth. 



The first of these is partly seasonal, partly random, and 

 may be partly monthly. The problem, involved is essentially one 

 of extrapolating the curve shown in figure 19. Attempts have 

 been made to devise a successful technique based on such para- 

 meters as persistency, surface temperatures, sharpness of the 

 thermocline, and a departure of the depth of the thermocline 

 from normal. As yet a relationship which can be considered 

 sufficiently accurate has not been devised, but work on a success- 

 ful method of approximation will be continued in the near future. 



The second problem mentioned above has been attacked with 

 some success. First, at several locations in the oceans at 

 which many time-series BT observations were available, the 

 mean amplitudes of the semidiurnal and diurnal oscillations 

 were computed for each of several categories of depth 

 at which the thermocline was found. The results are shoT-m 



50 



