December a relaxation of the eddy field is evident. During some years, how- 

 ever, the prime eddy is still discernible in the upper layer as late as January 

 (i .e. , January 1976) . 



Although a large eddy is found each year within the northern Somali Basin, 

 there are some differences from year to year. Probably the most evident is the 

 formation of a southern loop or eddy in 1976 and 1979. The pattern at that 

 time was similar to that found in 1970 (Bruce, 1973). The coastal surface 

 current turned offshore about 5°N to the south of the prime eddy. Normally in 

 the case of a single large eddy as observed during 1976 and 1977 the strong 

 upwelling associated with the divergence of the coastal current from the coast 

 occurs at approximately 9°N to 11°N. When the southern circulation or eddy 

 forms and becomes well developed, strong upwelling also occurs at 4°N to 6°N. 

 Although this circulation is termed an eddy, some of the flow extends south of 

 the equator and may in fact not all return to the coastal current regime. 



Judging from XBT data from 1975 through July 1979, it appeared that once 

 the general pattern of circulation becomes established during the monsoon, it 

 then tends to continue throughout the duration of the monsoon. For example, 

 between late May and mid-October 1976 (see figure 2), all eight tanker XBT 

 sections obtained throughout this period indicate that both the southern eddy 

 and prime eddy were present. Hov/ever, this might not always hold true. 

 During late August 1979 (figure 2), a northward shift in the northern boundary 

 of the southern eddy near 5^M occurred. This shift occurred in the following 

 manner: during the early stages of the development of the 1979 southwest 

 monsoon eddies, a large southern eddy and the northern prime eddy were both 

 clearly established by June in the Somali Basin. This circulation resulted in 

 the current turning offshore and forming a wedge of cold upwelled water in two 

 locations: 40N to B^N, and 90N to IQON. The general location and size of the 

 eddies tended to remain approximately the same through early August. The ESSO 

 HONOLULU XBT section, 14-18 July 1979 (figure 2), is representative of the 

 temperature structure along the tanker sea lane. The strong gradients above 

 150 m depth near 4°N occur at the northern edge of the southern eddy. The cold 

 near-surface water advected offshore by the anticyclonic eddy extends through 

 the section here. It may be seen that the southern eddy is relatively shallow 

 whereas the northern prime eddy (4°N to 10°N) exhibits horizontal temperature 

 gradients at least to 400-500 m depth. The upwelled cold water also extends 

 offshore through the section near lO^N. As found in each previous southwest 

 monsoon during which data in this region have been collected (Bruce, 1979), the 

 Socotra eddy also occurred in 1979 (lO^N to M^N). Because of the relatively 

 fresh (35.135.3 °/oo) near-surface water entrained in the system of eddies from 

 the Somali coastal current, a surface salinity map serves as a remarkably good 

 method for examining the circulation pattern. By mid-August the northern front 

 and cold wedge associated with the southern eddy began a northward translation 

 at a rate of 15-30 cm s" ' . By late August the southern eddy apparently had 

 merged with the northern prime eddy as indicated by the temperature section 

 taken 25-31 August 1979 from the ESSO CARRIBBEAN (figure 2) and a map of sur- 

 face salinity (figure 3) obtained during the survey aboard USNS WILKES. The 

 satellite imagery of sea surface temperature also indicated this translation 

 during August. The coalescence of the southern eddy with the northern prime 

 eddy is somewhat similar to that observed during August and September 1970 by 

 Bruce (1973), although for 1979 the data was obtained more frequently during 

 the occurrence. 



