The inferred easterly net flow below 4,000 meters does not preclude the westerly 

 flows determined for specific periods in this depth interval. Each of the determinations 

 in the present study was based on wind-driven geostrophic transports associated with 

 the weak atmospheric pressure gradients of the North Pacific summer (that is, May 

 or September). Thus, total geostrophic transport north of 45° N as found, for exam- 

 ple, for the portion of Section 5 between stations 120 and 123 and 125 was small. The 

 mass structure in this horizontal interval is such that a large easterly baroclinic trans- 

 port existed. The velocities associated with this easterly baroclinic transport are 

 strongest above 2,000 meters, decreasing to weak easterly values below 4,000 meters. 

 The combination of a small geostrophic transport and large easterly baroclinic transport 

 results in a large westerly barotropic transport. Above 2,000 meters, westerly barotropic 

 flow is masked by the stronger easterly baroclinic flow; but below 4,000 meters barotropic 

 flow is stronger than the baroclinic mode, and total flow is westerly. This result is reason- 

 able for the periods associated with the present determinations. 



Because the mass structure and the associated baroclinic transport can be expected 

 to be relatively constant, the annual cycle exhibited by the total wind-driven transport 

 must be reflected in the barotropic mode (Fofonoff, 1961c). The baroclinic velocity has 

 been shown to make but a small contribution to the flow below 4,000 meters, therefore 

 a description of the annual cycle in the barotropic flow does much to describe the net 

 flow below 4,000 meters. To gain a qualitative estimate of the annual cycle of the baro- 

 tropic flow the transport associated with the dynamic heights through stations 120-123 

 and stations 123-125 (1963) was successively adjusted to monthly values of the wind- 

 driven geostrophic transport. As this facet of the investigation is qualitative at best, it 

 seemed justifiable to use the most readily available tabulation of wind-driven geostrophic 

 transports for the comparison. The 1961 tabulations based on a drag coefficient of 0.0026 

 and monthly means of the atmospheric pressure distribution were used. Figure 63 

 is a plot of this determination of the annual cycle of the zonal components of velocities 

 at 4,000 meters. Variations of about 1.5 cm/sec in the zonal velocities are indicated but 



the net flow is not easterly. 



The approach used for estimating the annual cycle of flow shown in figure 63 is de- 

 ficient in the following respects: 



a. The monthly values used for wind-driven geostrophic transport are only approxi- 

 mations of the proper values. Transport based on a drag coefficient of 0.0012 and mean 

 stresses computed from twice-daily pressure reports might be significantly greater than 

 those used in constructing figure 63, due to the large variance in atmospheric pressure 

 which accompanies the severe, often short-period storms of the North Pacific winter. 

 These larger wind-driven transports should be easterly and larger than the easterly baro- 

 clinic transports, thus resulting in easterly geostrophic flow at the 4,000-meter level. 



b. The present formulation is based on the geostrophic approximation that a balance 

 exists between pressure forces and Coriolis forces. It must be recognized that this 

 balance is not perfectly achieved. Those time-dependent motions not satisfying the 

 geostrophic approximation are not considered in the present determinations. These 

 could make a significant contribution to the net flow which has been inferred from the 

 distribution of potential temperature and dissolved oxygen. 



c. The present investigation is based on the assumption that the total wind-driven 

 geostrophic transport and the barotropic transport respond significantly only to disturb- 

 ances of the one-month period considered. Veronis and Stommel (1956) indicated that 

 the barotropic response might be significant for disturbances with periods of from one 

 week to two months. 



d. In the comparisons supporting figure 63, it is assumed that baroclinic transport is 

 constant and can be accurately described by dynamic height computations based on 

 oceanographic data from May 1963. Obviously, there may be a significant variation in 

 the mass structure and thus in the baroclinic transport through the section. 



13 



