THE WATER MASSES 181 



of mass within that water. Parr (19386) has already emphasized the importance of these points in 

 interpreting charts of dynamic height anomalies. 



In the waters off South-west Africa, it is quite obvious that the requirements of Bjerknes theorem 

 are not satisfied. Clearly the wind in this region has a considerable effect on the sea-surface, and there 

 are undoubtedly considerable vertical movements of the water masses. It is therefore considered 

 unwise to place too much reliance on the charts as depicting actual currents, but rather to look upon 

 them as plans of the distribution of mass within the water above an arbitrary level surface. 



In the construction of the charts it is desirable to choose a surface of no motion, above which the 

 mass of water can be calculated. In the present observations it is impossible to be certain that such a 

 surface does exist. Normally such a surface is chosen at a considerable depth, where the movements, if 

 any, will be so small relative to those at the surface that they can be ignored. The 'William Scoresby's' 

 observations lie, however, for the greater part in shallow waters, and there is little likelihood that any 

 level will present a surface of no motion in these observations 1 . As the best compromise, however, the 

 600 decibar surface has been chosen, and the topographies of the sea-surface, and the 200 db. surface, 

 are shown relative to this in the charts in Fig. 34. Many of the present observations have been made 

 in the relatively shallow waters of the continental shelf where they cannot extend to the depth of the 

 reference surface chosen in the deeper water offshore. Helland-Hansen (1918) proposed a method to 

 overcome this difficulty. He suggested that the land mass in the vertical section representing the 

 continental shelf and slope, should be replaced by a fictitious body of water which is considered to be 

 at rest, and in which the isosteric surfaces would, therefore, be horizontal. In the adjacent water 

 mass the isosteric surfaces would tend to become horizontal as they approach the continental slope, 

 for, owing to friction, the velocity along the sea-bed must approach to zero. For this reason Helland- 

 Hansen suggested that the isosteres as they approach the sea-bed should be drawn horizontal and 

 produced horizontally through the land mass, thus furnishing arbitrary levels for the calculation of 

 the heights of the isobaric surfaces over the land mass. In face of the difficulties this seems a fairly 

 reasonable assumption to make, but it has no physical basis whatsoever, and it is indeed difficult to 

 imagine the existence of such an arrangement. Groen (1948) has discussed another method which 

 was suggested by Sverdrup, and proposed yet another, that of continuing the isosteric slopes into 

 the land mass, but it is doubtful whether these produce a picture any more realistic than that given by 

 Helland-Hansen's method which has been used in the following. 



Although a precise picture of the currents at different levels cannot be attained, it is intended that 

 the charts should at least serve as a check on the water-movements deduced from the distribution of 

 temperature, salinity and other properties of the water masses. The movements shown will probably 

 be more correct during survey I when there was less vertical movement, than during survey II. 



Surface topography (Fig. 34) 

 On survey I the most pronounced feature of the chart shows the tendency to a movement towards 

 the coast in the latitude of Walvis Bay (23 S.). This flow appears to bend southwards along the coast 

 and is no doubt responsible for the mixed nature of the water inshore at Walvis Bay, and the pro- 

 nounced oceanic influence at the inshore stations on the 25 S. line. In its southerly movement, it 

 displaces away from the coast the north-westerly flow of coastal waters from the south, so that in 



25 ° S. there is a southerly movement of mixed oceanic water inshore, and a northerly movement of 

 coastal water offshore. 



1 Attempts to define a level of no motion using techniques such as that of Sverdrup and Flemming (1941) have given 

 little guidance with the present observations beyond suggesting that the 600 decibar surface may be one of relatively little 

 motion. 



8-2 



