COMPARISON OF THE BENGUELA CURRENT WITH OTHER UPWELLING REGIONS 281 



current. Such upwelling may vary with the velocity of the current independently of the winds, and 

 we may distinguish it by referring to it as geostrophic upwelling. 



On the other hand upwelling may be brought about by the direct influence of winds on the surface- 

 waters, and it is into this category that the major upwelling regions of the world fit. This type of 

 upwelling depends on the lie of the coastline and direction of the winds in relation thereto. It may occur 

 at almost any time throughout the year (e.g. Peru, Benguela, etc.), but in certain regions of seasonal 

 winds it may be a more restricted seasonal phenomenon (e.g. off south-east Arabia or the Indian 

 coasts, Somali coast, Red Sea, etc.). The greater importance of this coastal upwelling derives from 

 the greater depths which it affects. In geostrophic upwelling an upward tilt of isopycnals does occur, 

 and the water brought to lie on the surface depends on the degree of tilt. In coastal upwelling, how- 

 ever, the isopycnals may be thrown up very steeply in response to the action of wind on the sea- 

 surface and deep-lying density layers may be transported right up to the surface. In these conditions 

 waters with a higher nutrient content are brought up to the surface, and the subsequent growth of 

 life makes their presence all the more evident. 



There are, as we have said, four extensive areas of coastal upwelling, all on the western coasts of the 

 continents of North and South America and Africa, and it is in these that we may look for analogies 

 and differences in the phenomenon. 



These currents have certain visible features in common — a negative surface-temperature anomaly, 

 characteristic coastal winds, frequent fogs over the cold water, and arid or desert conditions over the 

 adjacent land. These phenomena, however, may be modified by the topography of the land. Thus, in 

 South and West Africa, there is relatively low-lying land for some 80-100 miles from the sea; here the 

 Namib desert occupies quite a broad strip along the coast, and the sea-breeze can develop a consider- 

 able reach over the land. But in Peru the high Cordillera of the Andes is nearer to the coast ; it confines 

 the desert to a much narrower strip, and causes the coastal winds to blow more or less parallel to the 

 coast and axis of the Cordillera. 



There are also striking differences in the submarine topography of the various regions, which might 

 be expected to influence the process of upwelling. Off South-west Africa the continental shelf is 

 relatively broad compared with that off South America, where the steep slopes of the Andes are 

 projected below the sea-surface to abyssal depths, with only a relatively narrow continental shelf. 

 McEwen (19 12) suggested that the upwelling off California showed a striking correlation between the 

 areas of cold water and submerged valleys or other regions where the depth of the sea-bed increased 

 rapidly with distance from the coast. Gunther (1936) could find no evidence of such a correlation 

 off Chile and Peru, and the present work has indicated little or none off South-west Africa. Recently, 

 Yoshida and Mao (1957) have constructed a realistic mathematical model of upwelling off California 

 with little consideration for bottom-effects. There is, therefore, an accumulating amount of evidence 

 which suggests that the actual process of upwelling may not be greatly influenced by the width of the 

 continental shelf in these regions, although there may be effects on the attendant phenomena of the 

 upwelling. 



Essentially, the process of upwelling takes the same form in all of these regions. The discrete, eddy- 

 like structure of the surface isotherms is common to all (Peru, Gunther, 1936; California, Sverdrup 

 and Fleming, 1941; Morocco, Currie, unpublished), and the disposition of isopycnals suggests that 

 the basic mechanism is similar in all cases, conforming to the pattern outlined by Sverdrup (1938). 

 We have shown that the same basic mechanism probably occurs in the Benguela current (Fig. 37) and 

 Gunther's density figures for the San Juan line off Peru suggest that a similar vertical circulation is 

 present there also. It is a mechanism, however, which involves rather complex water movements in 

 three dimensions. Fig. 96 may perhaps help to clarify them. It is based on our inferences as to the 



