transport stream function, i// 9 . (As noted in appendix I, this procedure eliminates from 

 consideration the Ekman divergence.) Plots of the geostrophic transport function similar 

 to those for the total transport function are made (fig. 68) and the geostrophic transport is 

 extracted by difference. A comparison of alternate values of geostrophic transport for 

 the four sections considered is given below. 



Section Transport considering Transport from geostrophic 



Ekman divergence transport function 



63-59 — 253 • 10 4 metric tons/sec —20- 10 4 metric tons/sec 



29-32 — 795- 10 4 metric tons/sec —310- 10 4 metric tons/sec 



63-29 — 74 ■ 10 4 metric tons/sec — 250 • 10 4 metric tons/sec 



59-32 —550- 10 4 metric tons/sec —540- 10 4 metric tons/sec 



The dynamic topographies arising from a comparison of transport from dynamic heights 

 adjusted to each of these sets of wind-driven transport are shown on figures 23 through 29. 

 For the balance of the numerical example the transport including the effects of diver- 

 gence is used. 



The final terms in equations (35a, b) express mass transport through the sections due 

 to internal density structure. To obtain this baroclinic transport, data from each station 

 is submitted to a machine interpolation program which provides the following values at 

 standard depths: 



temperature, 



salinity, 



sigma-t, 



specific volume anomaly, 5, 



geopotential anomaly, AD, referred to the surface, 



anomaly of potential energy, x, 



and values for other variables represented in the field data. 

 For convenience in the computations advantage is taken of the near numerical equality 

 between mass transport and volume transports, T and T E . The mass transports deter- 

 mined from the curl of the wind stress are entered in subsequent comparisons as though 

 they were volume transports. The final terms in (35a, b) are determined from differences 

 in geopotential anomalies. Equations (35a, b), reflecting these near-equalities, become: 



u b = 



hRAcj) 

 AD„-AD t 



Vb = tb TTT '" (36a, b) 



tin cos <pAX 



Next, values are determined for the length-of-section terms, R cos 4>Ak and RA<p, 

 and for -, the reciprocal of the Coriolis parameter: 



28 



