Socotra. Note that the mass transport calculations suggest that the Socotra 

 eddy is the larger. Observations also show that during certain years (as 1970, 

 1976, 1979) (Bruce, 1973, 1979; Brown et a^. , 1980; Duing et. al- , 1980) a 

 portion of the Somali current may turn offshore about 40N to 60N forming a 

 separate near-equatorial eddy or loop. There is no indication of this mode of 

 circulation in the mass transport map. 



To the east of the Somali Basin observations from earlier hydrographic 

 surveys (sections along S^N and lOON) (Bruce, 1968) give evidence for the 

 occurrence of other eddies formed during the southwest monsoon. The geostro- 

 phic velocity and transport across these sections are shown in figures 23, 24, 

 25, and 26. Drawing from these data and previous studies of the southwest 

 monsoon circulation (Bruce, 1968; Duing, 1970) a schematic drawing indicating a 

 possible circulation pattern of the eddy field which occurs is given in figure 

 27. 



Off the Arabian coast during July (figure 21) a southward transport would 

 be required. From hydrographic observations (Bruce, 1968) it is not clear that 

 such might be the case. The section along IS^N during early August 1963 (Bruce, 

 1968) indicates a northward geostrophic near-surface current along the coast 

 with a transport amounting to approximately 12 x 10^2 g sec"^. Pilot charts 

 and the Dutch atlas (1952) show a northward current along the coast. 



During January the Sverdrup transport (figure 22) values are relatively 

 weak compared with that of July. A southward transport along the Somali coast 

 south of about 50N amounting to greater than 10 x 1012 g-1 sec would be re- 

 quired. This flow is in good agreement with the pilot charts (H.O. Pub. 566) 

 which show southward coastal currents ranging up to 125 cm sec"^. The meri- 

 dional geostrophic transport across a 5ON section just after the northeast 

 monsoon (figure 28) gives relatively low values compared with the transports 

 during the southwest monsoon (figure 24) as well as indicating a small north- 

 ward transport off the Somali coast. 



8. CONCLUSIONS 



From the evidence on hand to date, it appears that both the northern (or 

 "prime") eddy in the Somali Basin and the eddy east of Socotra are probably 

 generated each southwest monsoon. These eddies have been observed for five 

 consecutive years (1975 through 1979) by XBT temperature sections along the 

 tanker sea lane. They have also been present during all known earlier surveys 

 (Bruce, 1979). During some years (1970, 1976, 1979) the Somali current has 

 been found to turn offshore between about 40N to 6ON forming a southern loop or 

 eddy. When this circulation pattern is established, a region of cold upwelled 

 water occurs both at the location where the northern eddy turns offshore 

 (approximately 8ON to lO^N) and where the southern turnoff is observed (approxi- 

 mately 40N to 6ON) along the coast. During late August and early September 

 during 1970 and 1979 the southern eddy was observed to flow into and coalesce 

 with the northern one. 



During the southwest monsoon the strong signal in the sea surface dynamic 

 topography of the northern eddy develops each year. The slope on the northern 

 edge of the , eddy (S^N to 120N) amounts to about 2 x lO"^ dynes g""! (comparable 



10 



