3.2. Vertical Velocity 



dw 

 The rate of change of vertical velocity with depth, : — , was determined from the ad- 



oz 



justed values of meridional velocity for May, 1963, and the expression 



dw B 



With w assumed to be zero at the bottom, — — was graphically integrated between 5,000 and 



az 



250 meters. The distributions of vertical velocity with depth are shown in figure 55. The 



maximum values determined at the four latitudes investigated varied from 3 ■ 10~ 5 cm/sec 



to 14-10" 5 cm/sec. Though the magnitudes agree with those inferred by Knauss (1962), 



the directions do not. Knauss inferred upward velocities over the depth interval 2,500 



meters to 5,000 meters, whereas a consistent downward direction was found in this study 



at all levels below 1,000 meters. This disagreement may arise from the fact that Knauss's 



values reflect the net circulation, while values from the present determination are asso- 



dw 

 ciated with a specific period. It should also be noted that the values found for — are of the 



same magnitude as the limit of sensitivity accepted for the determination. Uncertainties 

 in the present determinations of vertical velocities and in the relationship of these veloci- 

 ties to net vertical circulation notwithstanding, it is noteworthy that the upward velocity 

 found near 50°N above 1,000 meters is in accordance with flow inferred from oxygen 

 profiles. In figures 5, 11, 14, and 17, the local thickening of the low-oxygen interval and 

 the decrease in depth of oxygen isopleths toward the north in the depth interval 200 to 

 1.000 meters suggest upward flow in this area. 



3.3 Depth of No Horizontal Motion 



In the present formulation, a component of the geostrophic velocity is zero at a depth 

 where the barotropic mode is equal and opposite to the baroclinic mode. The depths of no 

 zonal motion or of no meridional motion are shown as functions of latitude in figures 56 

 through 62. 



If, near the depth where the baroclinic and barotropic velocities add up to zero, the 

 baroclinic velocity changes slowly with depth, the determination of the depth of no motion 

 will be sensitive both to observational errors in the oceanographic data and to discrepancies 

 in the values for wind-driven transport. The sensitivity in the determination together with 

 the paucity of data leads to uncertainty in the plots of the surfaces of no zonal or meridional 

 motion. However, figures 56 through 62 do indicate that the depths of no zonal or me- 

 ridional motion vary widely with time and place, and that flow at depth is frequently op- 

 posite to that near the surface. 



3.4 Annual Cycle of Horizontal Velocities 



Two features of the flow at 4,000 and 5.000 meters persisted through each determina- 

 tion: The well-developed westerly flow north of 45°N, and the southerly flow in the zone 

 from 30° to 35°N (figs. 28, 29, 31, 32, 37, 38, 53, and 54). The southerly flow is in agreement 

 with that inferred by Wooster and Volkmann (1960). Their data are not definitive north 

 of 45°N. However, an easterly velocity can be inferred from the distributions of potential 

 temperature and dissolved oxygen observed in the present study, because potential tem- 

 perature increases and dissolved oxygen decreases from west to east. So it must be con- 

 cluded that the velocities determined for September 1961, September 1962, and May 1963 

 are not representative of the net circulation. 



12 



