general level of transport of major currents, geographical and vertical 

 variation of the main thermocline, the general distribution of properties, 

 and features of the abyssal circulation have been remarkable. But expla- 

 nation of these gross features of the circulation does not involve a crucial 

 testing of the internal dynamics of the models. The successes may be at- 

 tributable simply to the identification of primary driving mechanisms and 

 the inclusion in the models of dominant constraints of geometry, rotation, 

 and stratification. 



However, despite the efforts of an increasing number of skilled and 

 dedicated scientists, development of understanding in the last few years 

 has been disappointing. Even relatively simple physical refinements of the 

 models led to difl!icult nonlinear fluid dynamical problems; the deduction 

 of general consequences from simple hypotheses has been precluded. Al- 

 though studies of transient processes have been initiated, circulation models 

 including the eddy processes in a general way have not yet been formulated. 

 The problems involved are not overcome by the direct exploitation of nu- 

 merical models in their present state of development. Present limitations 

 of machine capability and computational techniques require the specifica- 

 tion of model dynamics in such a way as to prohibit the occurrence of im- 

 portant nonlinear and time-dependent processes. For example, large posi- 

 tive values for horizontal eddy viscosities are often employed, and grid 

 spacings in general circulation models are too large to allow the mesoscale 

 eddy processes to occur spontaneously. 



With regard to the dynamics of the general circulation of the atmosphere, 

 the importance of mesoscale eddy transport mechanisms has been estab- 29 



lished on a firm theoretical and observational basis. The eddy dynamics 

 replaced the long-prevalent concept of circulation based upon Hadley cell 

 dynamics, which was first abandoned on empirical grounds. The late 1940's 

 saw both the first direct calculation of the eddy transport of angular mo- 

 mentum from adequate data and the discovery of the fundamental process 

 of baroclinic instability. Since that time, great strides forward have been 

 made in the understanding of the general circulation and the ability to 

 predict and to monitor the appropriate variables. 



The evolution of the understanding of atmospheric circulation pro- 

 vides valuable guidance in present studies of ocean circulation. It is un- 

 likely that there exists a strong direct analogy between atmospheric and 

 oceanic circulation, but, as has already been indicated, the fundamental 

 importance of eddy processes is strongly suggested. The space and time 

 scales of the oceanic mesoscale are expected to be respectively shorter and 

 longer than the corresponding atmospheric scales (based upon the radius 

 of deformation, observed speeds, and the advective time scale) . The strength 

 of the physical analogy between atmospheric and oceanic circulation will 

 depend upon the as yet unknown energy source (s) of oceanic eddies. This 

 energy may come from baroclinic instability of the open ocean, from major 

 ocean currents which shed long-lived eddies, from transient surface forc- 

 ing with scale transfer due to interference with bottom topography, from 

 nonlinear transfer of energy from inertial motions, or from internal waves 

 and tides. 



Direct observational evidence for the importance of the mesoscale 

 oceanic eddy motions, although scanty and nondefinitive, provides some in- 



