these assumptions, the transport equations may be written: 



dx g \dx ) z=Zb 



fU 3y ~W)^ b + T ° y - (14a ' b) 



There are no existing methods of determining open ocean transports with sufficient pre- 

 cision to account for mass exchange at the sea surface by continuity methods. To a 

 sufficient degree of accuracy the divergence of the transport is expressed by 



dU , dV n 



Equations (14a, b) and (15) are the basic ones for this problem. 



After Fofonoff (1962), the components U and V of the transport can be separated into 

 barotropic, baroclinic, and Ekman modes: 



U=U b + U g + U E 



(16a, b) 



V=V b + V g + V E 



The barotropic mode is interpreted as a velocity, v b , constant throughout the water column, 

 so that 



«=*» 



Ub=\ pu b dz, Vb= pv b dz. (17a, b) 



Jz b Jz b 



since m and v b are not functions of z 



Ub = m> pdz, Vb = Vb pdz. 



By the hydrostatic equation 



Ub^l^dP, Vb = ^\" b dP; 



s Jp e Jp 



n _UbPb y, _VbPb 



(Jb — , Vb — 



g g 



By substituting for m and v b in equations (17a, b) 



g \dy)z=z b g \dx j z=Zb 



(18a, b) 



The baroclinic mode is interpreted as that part of the transport that is a consequence of 

 pressure gradients arising from field variations in the density structure of the water. The 

 velocities associated with this mode of transport are the relative velocities resulting from 

 dynamic height computations. This mode of transport can be related to the pressure 



22 



