the upper thermocline (about 150 m depth) and 700 m. The similarity of the 

 nature of these failures suggests that they resulted from a common cause 

 rather than from random factors such as sea state, wire insulation penetration, 

 manufacturing defects, and electrical interference. Since the region of the 

 Somali Current during the southwest monsoon is characterized by strong current 

 shear, it seems reasonable to ascribe the persistent wire failure to strains 

 induced by the strongly shearing currents. The sea surface analog trace was 

 examined, and it was found that the area of high XBT failure rate coincided 

 with the strong sea surface temperature gradient associated with trace A 

 (figure 48a). Similarly, this area of frequent XBT failures coincided with the 

 strong sea surface salinity gradient encountered on section 1 (figure 19). _ 

 Almost every other encounter with strong sea surface temperature and salinity 

 gradients produced high XBT failure rates. The locations of XBT failures are 

 plotted in figure 49. 



IX. INTENSE DIURNAL HEATING 



Trace S (figure 48c), made during the afternoon of 3 September 1979, 

 resulted not from a frontal crossing but from intense solar, insolation of the 

 upper few meters of the water column, often termed the "afternoon effect". 

 The wind was calm, and the sea surface was characterized by alternating bands 

 of slick and capillary waves accompanied by a 2- to 3-foot swell. The spikes 

 seen on trace S indicate a highly variable sea surface temperature field. It 

 was found that the spikes on the sea surface temperature trace corresponded to 

 the bands of slick or glassy sea surface, while the minima were encountered in 

 the bands of capillary waves. Alternating bands of smooth glassy water and 

 capillary waves are quite common in tropical and subtropical waters under calm 

 or nearly calm conditions. Although the mechanisms for the formation of these 

 alternating bands are not completely understood. Roll (1965) has proposed 

 internal waves and subsurface convective cells as possible causes of their 

 formation. 



While transiting trace S between 300 and 280 nmi from the end of section 

 10, sea surface bucket temperatures of 30.9°C, 31.0°C, 30.4°C, and 30.1°C were 

 obse^-ved. In spite of these extremely high surface temperatures, the XBT 

 analog traces registered sea surface temperatures of only 28°C. These low 

 readings were probably caused by the thermal lag of the thermistor in the nose 

 of the XBT probe as it fell through the extremely shallow layer of warm sur- 

 face water. To test this hypothesis, a styrofoam jacket was made for an XBT 

 probe to give it a very low sink rate. The resulting XBT record, after being 

 corrected for the low sink rate by comparison with a normal XBT drop (Bruce 

 and Firing, 1974), showed a very thin layer -- less than 2 m -- of warm water, 

 apparently caused by intense solar insolation accompanied by a lack of wind 

 mixing. 



X. DISCUSSION AND CONCLUSIONS 



Although there is a definite consensus among meteorologists and oceanogra- 

 phers as to the existence of a causal link between the southwest monsoon winds 

 and the Somali Current, it is not sufficient to attribute this current solely 

 to local wind forcing. The wind induced coastal upwelling evident in satellite 



19 



