UPWELLING 187 



breakdown of the relative current into eddies may be largely dependent on the localized coastal winds. 

 Clearly, however, a correlation between such observations as wind observed at coastal stations, within 

 some modifying influence of the land, and hydrographical observations at sea must at best be rather 

 rough, and a more satisfactory correlation might be expected if it were possible to compare the winds 

 at sea with the coincident hydrographical changes. The only opportunity which we have of making 



30 -1 



+ 



o 

 z 



o 



UJ 



> 



Q 



z 



5 



30 



f EBBUAftr MARCH 



14 21 IS I 7 



zjsry 



+ 



30- 



+ 



O-- 



30 J 



^ n 



Fig. 35. Predominant coastal wind vectors before and during the two 'William Scoresby' surveys at three points on the 

 coast — Walvis Bay, Luderitz Bay and Orange river mouth. Winds which would produce a longshore northwards, or off- 

 shore drag on the sea-surface are taken as positive and other winds as negative. Wind speeds are in knots. Heavy black lines 

 indicate the beginning of the survey in each position. 



such an estimate of the short-term effect of the wind on the sea-surface was the occasion of a S.S.E. 

 gale off Walvis Bay. While steaming west along the Walvis Bay line on the second survey the ship 

 was forced to heave-to for nearly 48 hr. When the gale subsided the ship returned eastwards, and a 

 comparison of the sea-surface temperatures before and after the gale reveals some interesting features. 

 In Fig. 36 the sea-surface isotherms have been drawn from the station observations and the 

 distant- reading thermograph records; they are in black for the westward passage preceding the 

 gale, and in red for the eastward passage following the gale. 



