with a 14.8-kilometer interval between casts. Upon completion of the 

 third SXBT pattern, three sections were made across an area of warm water 

 with SXBT probes spaced approximately 7.4 kilometers apart. Phase II 

 was modified to obtain two profiles using Nansen casts spaced at intervals 

 of 14.8 kilometers, separated by three sections across the warm water 

 with probes spaced at 7.4-kilometer intervals. The initial Nansen section 

 included five casts, the repeat section included three. Two additional 

 SXBT sections were made during transit to other areas; the first consisting 

 of hourly observations between 36''22'N,74°50'W and 35''59'N,73°50'W and 

 the second consisting of half-hourly observations along 36''56'N between 

 72°59'W and 75''O0'W. 



The SXBT probes had a temperature accuracy of +0.4''C and a depth 

 accuracy of +4.6 meters or 2 percent (whichever is greater). SST was 

 continuously recorded by the NSRT system. Previous evaluation of the 

 NSRT system showed a mean difference of 0.4°C from the uppermost bottle 

 of a Nansen cast (Fisher, 1967). 



Meteorological data recorded every 6 hours and SST data recorded 

 daily by Coast Guard personnel at Chesapeake Light Station (36°58.7'N, 

 75*'42.2'W) are used to describe prevailing weather in the survey area. 



DATA ANALYSIS 



SST patterns observed during the flights of 20 and 21 September and 

 6 October are shown in figures 4 through 6. Ship data were not used in 

 the surface analyses because of the differences in time span of the 

 surveys. Corrected ART values were averaged over one-minute periods for 

 plotting, except where strong gradients occurred. Where strong gradients 

 occurred, the actual gradient was plotted. The 20 September analysis is 

 admittedly subjective, because stratus clouds and fog obscured the east- 

 ernmost leg. This analysis is contoured to best maintain continuity with 

 the pattern of 21 September. Two alternative analyses for 20 September 

 are shown in figure 7. The first shows warm water as an eddy not connected 

 with the Gulf Stream; the second shows warm water as a tongue extending 

 northward from the Gulf Stream rather than from the east as in figure 4. 



The major features of the three SST analyses are: (1) a tongue of 

 warm water extending into the survey area from the east, (2) cold water 

 extending into the survey area from the north, (3) a band of warm water 

 along the coast, and (4) location of the Gulf Stream by means of a strong 

 gradient in the southeast corner of the survey area. SST changes of 2°C 

 or greater were observed on 41 overflights of the boundary between the 

 warm water extending into the survey area from the east and the cold water. 

 The mean temperature gradient across the boundary was 0.92°C/kiTi with a 

 minimum of 0.12"'C/km and a maximum of 4.70°C/km. Because the flight tracks 

 were not necessarily normal to the boundary, the actual gradients may 

 have been much greater. 



