338 THE DEEP SEA 



the simplest experimentally realizable types of turbulent transport 

 which can be studied theoretically. In these controlled experiments 

 one can hope to establish the range of validity of quantitative 

 hypotheses concerning the field of motion. In this paper I shall (1) 

 explore the possibility of isolating those turbulent processes in the 

 sea which one might hope to treat theoretically, (2) explore some of 

 the continuity relations which must exist between the separated 

 turbulent processes, and (3) show that the more detailed mechanistic 

 studies of the last decade support the inferred relations. In particu- 

 lar, I wish to paint an idealized picture of the North Atlantic c'rcu- 

 lation suggesting the various turbulent mechanisms which control 

 the flow and to predict certain integrals of mass, heat, and mo- 

 mentum transport which one can compare both with observation 

 and other theories. 



The Turbulent Ocean versus an Idealized Ocean 



Many different length scales characterize the many turbulent 

 transport mechanisms of the sea. The momentum transport from 

 air to sea produces both the smallest and the largest of these turbu- 

 lent motions. This momentum transport, often called wind stress 

 or wind friction, leads to turbulent wave fields on the small scale 

 and turbulent fluctuations in the total southerly mass transports 

 on the large scale. Two important turbulent heat transport processes, 

 with length scales between these wind-dri\"en motions are the verti- 

 cal convection of heat due to winter cooling of surface water and 

 the Gulf Stream fragments or eddies which are an example of 

 baroclinic turbulence due to horizontal temperature differences. 



In his exploration of the density field of the sea, the oceanographer 

 has found the one variable which least reflects the underlying 

 turbulent dynamics. It was plausible that the relati\-ely steady 

 density structure would suggest a rationalization of the dynamics 

 in terms of steady laminar-like flows. Interpreted as average flows, 

 the laminar idealizations are both sound and illuminating. However, 

 the structure and amplitude of these flows can be determined from 

 the winds and surface heating only if we include in our analysis the 

 turbulent transports of momentum and of heat. L^nfortunately the 

 length and time scales of this turbulence can be \'ery large, for 



