gradient south of the region of the maximum stress values extends northeast- 

 wards from the African coast to about 65°E. The general pattern of the 

 contours is similar for August and September, however, with a continued 

 weakening of stress to 0.2 dynes cm"2 or less by October. 



There are apparently fluctuations of a shorter duration than could be 

 shown by monthly averages during the build-up of the southwest monsoon wind 

 field. Schott and Fernandez-Partagas (1980) have found that the wind for May 

 and June 1979 varies over a period of a few days in both speed and direction-. 

 The variations were evident in three-day averaged ship observations as well as 

 records from shore stations and cloud-level winds from satellite data. 



The first evidence of wind reversal and the commencing of the northeast 

 monsoon is also indicated off the Somali coast about 60N to lO^N during Novem- 

 ber. By January at the maximum strength of this monsoon values of -1.0 dynes 

 cm"2 occur, these being along the Somali coast approximately between O^N and 

 lO^N. The wind stress then diminishes by March. April and October, the 

 months falling between monsoons, have relatively low values over most of the 

 northwestern Indian Ocean. 



6. CURL OF WIND STRESS DURING MONSOONS 



The increase of the curl of the wind stress during the commencement of the 

 southwest monsoon starting in May and reaching maximum values in July is shown 

 in the maps of figures 15, 16, and 17. The curl maps were constructed using 

 the gradients of the wind stress field given in figure 14 with a grid spacing 

 of 2° for both latitude and longitude. 



The most negative values first appear in the Somali Basin (B^N to 80N, 

 53OE) by May and increase through July. They tend to lie in a band extending 

 from the central Somali Basin on toward the northeast to about 65°E. 



Values during July of -6 x 10"8 dynes cm-3 in the region, 50N to lO^N, 

 54OE to 590E, occur approximately where the thermocline deepens in the central 

 part of the large Somali anticyclonic eddy observed each southwest monsoon 

 (Bruce, 1968, 1979). There is insufficient data concerning the prevalence and 

 structure of the eddy field to the northeast of the Somali Basin, lO^N to 

 I50N, 6OOE to 65OE, where the curl values up to -10 x lO'^ dynes cm-3 are 

 shown for July, to estimate the probable location of these mid-ocean eddies. 

 However, there is evidence from the near-surface dynamic topography that they 

 occur here (Bruce, 1968). Also from Robinson et al- (1979) (figures 18 and 

 19) the depth to the top of the thermocline reaches a maximum during the 

 southwest monsoon in the region of large negative curl values and the patterns 

 of the maps of curl and thermocline depth are somewhat similar. 



A region of positive curl to the north of this band is 

 Arabian coast (values over 10 x 10"^ dynes cm"^) and extenc 



found off the 

 ids down to the 



southwest just off the Somali coast to about 40N. The region where the Somali 

 current turns offshore has been observed to be between about 4°N to 9°N 

 (Bruce, 1979) within the area of positive curl. Here upwelling and the upward 

 vertical velocity reaches a maximum in the near surface water (calculated by 

 Swallow and Bruce (1966) to be 7 x 10-3 ^m sec-^ while the thickness of the 

 mixed layer decreases. 



