14 



successive traverses by ships. These devices may also link observations from 

 aircraft to the occurrences in the upper layers of the ocean. 



In order to decide what is to be required of these instruments and how 

 they may be designed, it seems necessary to review present ideas concerning 

 some of the phenomena of oceanic circulation to see what may be observed. 



THE STEADY AND TRANSIENT STATES OF OCEANIC CIRCULATION 



Stommel (1948) and Munk (1950) have offered a very satisfying theoreti- 

 cal description of a primary oceanic circulation system driven entirely by the 

 mean zonal wind stress. Recently it has been possible to simulate qualitatively 

 not only the graphical results of this mathematical model, but a number of sec- 

 ondary details of the ocean prototype in a rotating model (von Arx, 1952a). In 

 these experiments a pattern of motion similar to the surface circulation of the 

 ocean prototype has been produced in homogeneous water. The circulation de- 

 velops almost immediately after a steady wind torque is applied, and coasts to 

 a stop nearly as rapidly when the wind torque is removed. This result appears 

 to support the growing impression that the superficial motions respond to a sea- 

 sonal change in wind torque with little delay (Fuglister, 1951), but that large 

 scale adjustments of the density structure take place much more slowly. The 

 adjustments of the density structure may be effected by two mechanisms, or 

 possibly three, having different times of response and different areas of influ- 

 ence. Since the Gulf Stream has been studied in some detail by many investiga- 

 tors, it serves especially well as a focus for discussion of these effects. 



Seasonal variations in the transport of the Gulf Stream (Iselin, 1940) 

 have been observed by the method of dynamic sections and are indicated to some 

 extent in tide gauge records. Yet, in different authors' opinions, theoretical 

 estimates of the rate of response of the density structure to a change in wind 

 stress range from less than a decade to a millenium. In any case, it is longer 

 than a season. Since the sun may not supply heat rapidly enough to change the 

 volume of specifically light water in the Sargasso Sea by the amounts required to 

 compensate the seasonal rate of change of wind stress, the compensating pres- 

 sure gradients may be derived from a temporary deformation of the Sargasso 

 water mass. Iselin has suggested (1940) that in periods of greater than average 

 wind stress the horizontal extent of the Sargasso Sea may diminish in order that 

 the pressure gradients may be temporarily increased through elevation of the 

 free surface and depression of the main thermocline. These effects are thought 

 to be reversed in the event of less than average wind stress. This response of 

 the density structure would tend to distort the circulation pattern and cause it not 

 to fit exactly within the geometry of the ocean basin. To compensate for this a 

 change in volume of the Sargasso water mass is required which will take place 

 at a secular rate determined by the ability of the sun to heat the added volume 

 or by the ability of dissipative effects to cool enough water to restore the cir- 

 culation to its normal geographic course. Were the rate of change of wind 

 stress slow enough for the secular rate of accumulation or dispersal of specifi- 

 cally light water to keep in step, the volume of the Sargasso water might change 

 while its horizontal extent remained constant. 



While the elastic deformation of the whole Sargasso water mass may be 

 rapid in comparison with the rate of accumulation or dissipation of specifically 

 light water, there may be localized deformations of the pressure field that oc- 

 cur still more quickly; perhaps as rapidly as the response of the superficial 

 circulation to a change in wind stress. A sudden increase in the velocity of the 

 Gulf Stream would be accomplished by a corresponding increase in the Coriolis 

 force. This would cause the motion to be directed to the right of the isobars 



