SECT. 2] 



LARGE-SCALE INTERACTIONS 



155 



several assiini])tions and approximations were required in applying (30), 

 namely, 



(i) The basic ciuTent was taken constant throughout the year. For mass flux, 

 the transport across the Florida Straits of 26 x 10^ m^ sec"i (Sverdrup, 1942) 

 was used, assuming that one-third occurs in the upper 90 m, thus contributing 

 to the heat flow. 



(ii) The mean monthly temperatures across the entrance and exit ends were 

 computed using the temperature depth dependence of Fig. 27 and the surface 

 temperature profiles of Fig. 28. 



(iii) Vertical heat flux was ignored. Stommel's (1958) suggested upwelling 

 (of the order of meters per year) may be shown to contribute a negligible 

 correction to the computed Qvo. 



(iv) Space correlations of mass flux and temperature across the channel 

 were neglected, as were time correlations on a scale less than a month, due to 

 lack of data. Thus the eff'ect of all oceanic scales of motion smaller than the 

 basic advective current and its monthly temperature variations had to be 

 left out. 



With these restrictions, the resulting Q^o was calculated and appears in line 

 10 of Table XI. Positive heat-flux divergence in summer gives way to negative 

 (convergence) in winter because, as shown in Fig. 28, the Gulf of Mexico 



84 



■t 82 



S 81 



I 



*I 80 



o 



5 79 



78 h 



77 



J FMAMJ JASOND 



Fig. 28. Seasonal march of sea-surface teinperature at entrance (East Caribbean — dashed) 

 and exit (Yucatan Channel — solid) ends of Caribbean ellipse. Determined from 

 Fuglister's (1947) charts. (After Colon, 1960, Fig. 11.) 



(current exit) is warmer than the Atlantic (current entrance) in summer and 

 cooler in winter. Actually Qvo proves to be the smallest component in the 

 Caribbean heat balance, so that fairly large errors in its assessment are not 

 important. In June and September, the maximum oceanic heat export is 

 4-9% of R, while in November and December a maximum import of 5-6% R 

 was indicated which, as we saw earlier, are within the uncertainty of radiation- 

 balance computations. 



