Data collected from WILKES during August and September included 415 XBTs 

 and 27 STD stations. In addition to observing the shift of the southern eddy, 

 the prime or northern eddy was surveyed. Also a more detailed study of the 

 Socotra eddy was obtained than has been made to date. 



In 1975, 1977, and 1978, whereas a wel 1 -developed southern eddy was not 

 evident, still there appear to be variations in the near-surface (0-100 m) 

 structure that suggest that some offshore flow might occur between 3°N and 

 5°N. For example, along the 19-23 October 1975 section (figure lb) at 4ON to 

 j°N, there are small scale temperature gradients which suggest a weak eastward 

 flow in the upper 100 m. It seems possible that the returning onshore flow 

 (20N to 5°N) of the prime eddy or northern eddy might well affect changes in 

 the alongshore current flow. Similar observations are described in laboratory 

 scale models for fluidics research (Carbonare ejt aj . , 1970). A small southern 

 eddy was found during the early stage of the southwest monsoon 1978 (Bruce e^ 

 a^l . , 1980) although it did not appear to attain the size or strong horizontal 

 gradients occurring during 1976 or 1979. 



The time series of sections following the development of the eddy struc- 

 ture through the early stages of the southwest monsoon (March through June) 

 indicate that the prime eddy first forms between 5°N to lO^N in the Somali 

 Basin, with the center at approximately 8°M, 530E. The data do not suggest 

 that the northern eddy is formed at the equator and then translates northeast- 

 ward along the coast as postulated by the numerical models of Cox (1976), and 

 Hurlburt and Thompson (1976). 



The strong signal during the southwest monsoon in the surface dynamic 

 topography (70N to 10°N) of the sea surface within the prime eddy is evident 

 in figure 4 and the complete series in figure 5. The same temperature-salinity 

 relationship for the Somali Basin during the southwest monsoon (a mean tempera- 

 ture-salinity curve determined by values obtained from previous surveys 

 during the southwest monsoon period) was used for all the determinations 

 shown. The density gradients that occur in the Somali eddies, as in the Gulf 

 Stream, are largely a function of temperature. The pronounced downward slope 

 of the surface dynamic topography to the north in figure 4 occurs between SON 

 to 12°N with values on the order of 2 x lO'^ dynes g"^ (about the same as 

 found across the Gulf Stream at 36°N). The volume transport of the prime eddy 

 to the east offshore amounts to 38 to 42 x 10^ m^ sec'^ (0-400 dbar, rel . 400 

 dbar) with a comparable return flow inshore to the south between 4ON to 8°N. 

 To the north of the prime eddy, the Socotra eddy occurs each of the five 

 observation years with transports on the order of 9 to 15 x 10^ m-^ sec"^. 

 The temperature sections and surface dynamic topography show that this eddy 

 during 1979 (center of eddy along sections is about 120N) was well developed 

 from July through October (figures 2 and 5). 



The surface temperature (figure 6) and salinity (figure 7) characteristics 

 of the western Indian Ocean, particularly in the region of the Somali Basin, 

 are changed considerably during the southwest monsoon as a result of several 

 factors: 1) the advection into the basin by the Somali Current of relatively 

 cool and fresh South Equatorial Current water, 2) high evaporation, 3) advec- 

 tion of upwelled water (also relatively fresh and cool) off the Somali Coast, 

 4) vertical mixing resulting from the very large wind stress at the sea 

 surface during the southwest monsoon, and 5) horizontal mixing within the 



