COUPLING BETWEEN SEA AND AIR 407 



be used to increase the mechanical energy of the atmosphere. If 

 more than 1% of this additional efTect could be us3d to drive the 

 ocean currents, we would give back to the ocean more than 10^ 

 kw, and instability may result. In the present case, however, no 

 such amplifications are likely to occur, since certainly less than 

 1% of the total wind energy is used to drive the currents. 



Also, if an amplification had occurred, oscillations would not 

 necessarily be generated. Any original change in the wind stress 

 would most likely give a reverse effect through the atmosphere. 

 (An increase in the wind stress would, for example, increase the 

 advection of warm water toward the poles and cold water toward 

 the equator and so decrease the differential meridional heating of 

 the atmosphere and the winds would slow down.) 



Such a reverse feedback could cause oscillations, but only, as 

 is shown by a closer theoretical analysis, when the characteristic 

 response times of the atmosphere and the sea are comparable. In 

 reality these response times dififer greatly. The atmosphere will 

 adjust to a change in the heating function within a few weeks, 

 whereas the response in the ocean surface water, with a change in 

 the wind stress, will take several years. 



In the above example we see a feedback link that is most likely 

 stable. The question arises whether there may be other dynamic- 

 thermodynamic feedback couplings between the atmosphere and 

 the ocean that may result in oscillatory states. The investigation 

 of such a question is, of course, very difiicult, since we do not have 

 a real knowledge of all the complicated couplings that exist between 

 the atmosphere and the ocean. One can, however, make some 

 estimates based on our general experience with feedback systems. 

 Such systems have, in fact, been studied a great deal in mechanical 

 and electrical engineering. 



A self-sustained oscillation in a feedback system requires, to 

 begin with, that an energy flow take place through the system, 

 to overcome the dissipation that otherwise will damp the oscilla- 

 tions. The atmosphere-ocean system certainly fulfills this require- 

 ment, the energy flow being sustained by the incoming solar 

 radiation. A second requirement is, as said earlier, that the two 

 processes in the atmosphere and the sea have comparable response 



