SECT. 3] DYNAMICS OF OCEAN CXJKRENTS 363 



Cyclonic inertial circulation differs sharply from the anticyclonic flow. The 

 Coriolis force in the boundary current and eastward jet is directed outwards 

 towards the boundary. Hence, no separation of the flow from the coast can 

 occur. If the interior flow is sufficiently rapid, water in the lower layer may be 

 exposed to the surface between the eastward jet and the westward drift. i 

 However, this case cannot be examined with a linear interior transport function 

 as the westward velocity would have to be infinite at the intersection of the 

 interface with the surface. For slower westward velocities, the boundary-layer 

 solutions can be obtained using the same procedures as those outlined for the 

 anticyclonic circulation. Meridional profiles of the upper layer along the 

 western boundary and in the interior are shown in Fig. 4. 



The characteristics of the boundary flow for constant potential vorticity 

 were first obtained and described by Stommel (1954) as an example of a simple 

 model giving a velocity profile similar to that of the Gulf Stream. Morgan (1956) 

 examined the formation of the inertial current along the western boundary 

 assuming an interior solution similar to (78). He chose a parabolic distribution 

 of the zonal component of wind stress so that the westward component of 

 transport was independent of y. Thus, although both the transport component 

 and the transport function depended on x, the flow entering the western- 

 boundary region was essentially identical to inertial drift assumed in (128). 

 Morgan's analysis of the boundary current is equivalent to the present treat- 

 ment with the exception of the eastward jet, which he did not examine. Many 

 of the basic relations for the boundary current are given in his paper. Charney 

 (1955) considered a slightly more complex example of interior flow in his 

 analysis of the inertial intensification in the western boundary region. He 

 assumed an interior transport function that varied parabolically with y. The 

 procedure he used to obtain the boundary-layer solution is similar to that given 

 here. He was able to show that the distribution of depth of the upper layer 

 given by the boundary-layer solution was in reasonably good agreement with 

 that of the 10°C isotherm between Florida Straits and Cape Hatteras. 



It is clear that the dimensions of the inertial baroclinic circulation will be 

 determined in part by the volume of water available in the upper layer. For 

 the elementary flows considered here, the volume of the upper layer can be 

 taken roughly proportional to ^o2/max. Hence, from (130), we see that the 

 meridional extent of the circulation, ?/max, will be proportional to F^/', where V 

 is the volume, and inversely proportional to i/'max*'^'. Thus, to attach an inertial 

 boundary current to an interior flow of the type given by (78), we must assume 

 that enough upper-layer water is available to support the flow as a baroclinic 

 mode on a scale that is determined by the wind field. There does not appear to 

 be any a priori reason for believing that the volume in nature will be just 

 sufficient to support the circulation. If there is insufficient upper-layer water, 

 there would be a complex breakdown of the flow in the boundary region as the 

 maximum allowable baroclinic transport is reached. The breakdown could 



1 Provided a western-boundary current can exist with a "free" stream-line on the 

 seaward edge. 



