Gulf Stream, Vol. 1, No. 7) and its wavelength is smaller. Figure 7 

 re-emphasizes the wave movement south from the 10th to the 11th, then 

 2 days of northern progression from the 11th to the 13th. The move- 

 ment of the northern boundary is counter to the wind drift on the 12th 

 and 13th. 



Calculations were made to determine vertical thermal structure 

 changes due to atmospheric change (north of the Gulf Stream). Since 

 the technique (Gemmill and Nix, 1965) requires knowledge of the initial 

 ocean thermal structure with depth, a guess of the structure based on 

 historical BT's was made. Several tests of mixed layer depth indicated 

 that the amount of observed cooling would deepen a 15-meter layer to 

 30 meters during the survey period. 



Analyses of ART data are good representations of the sea surface 

 temperature pattern. However, vertical thermal structure can only be 

 implied from ART analyses, and, in order to fully understand the three- 

 dimensional changing ocean, temperature observations with depth must be 

 provided. 



THE USNS GILLISS SURVEY 



The USNS GILLISS survey was conducted in two identical phases (I 

 and II) with the objective of observing changes that occurred over a 2- 

 week period. Of approximately 1,000 BT's collected, half were taken by 

 mechanical BT's to a depth of 300 meters; the remainder were taken by 

 shipboard expendable BT's (SXBT) to 450 meters. Ten zonal legs were 

 surveyed (figure 1) from 36°N to 40°N between 65°W and 69°W, with obser- 

 vations about 8 km apart. Each leg required nearly one day to transit. 

 A special crossing of the Gulf Stream was made from 39°20'N to 38°34'N 

 and back to 39°N along 65°W. Observations were taken at 1-km intervals 

 during this crossing. 



Horizontal temperature analyses were drawn for the sea surface, 

 200 meters, 450 meters, and the mixed layer depth. Cross sections were 

 drawn for each zonal leg for both phases. Changes between phases appear 

 in the horizontal analyses and cross sections, but it should be remem- 

 bered that in composite analyses there are time distortions. This factor 

 is insignificant in the cross sections, because each section is based on 

 data taken within a 24-hour period. 



A. WEATHER INFLUENCES 



During the early part of phase I, two large cool anticyclones moved 

 eastward from the U.S. to dominate the area. The storm track was located 

 close to the east coast of the U.S. and did not influence the area. On 

 30 October an intense cyclone developed near the northeast corner of the 



