(colder than -1.8°C (28. 76°F)), with salinities between 34.60%o and 

 34.6 8 %o, and corresponding densities (sigma-t values) slightly less than 

 that of Shelf Water. Previously collected data in this area during DEEP 

 FREEZE 61 show similar temperature and salinity values. The formation 

 of this characteristic water mass may be due to less stable winter con- 

 ditions than exist for formation of Shelf Water or the proximity to 

 Circumpolar Water intrusion. 



F. Sound Velocity Structure 



Hunkins (1962) indicated that a surface SOFAR channel should exist 

 in the Antarctic as well as it does in the Arctic Ocean. Analysis of 

 data from 108 DEEP FREEZE 63 and 53 DEEP FREEZE 64 oceanographic stations, 

 indicates a surface SOFAR channel does exist in certain areas of the Ross 

 Sea during at least part of the austral summer (Figs. 20, 21, 22). This 

 SOFAR channel normally is associated with the Winter Water layer. Figure 

 23 shows a typical sound velocity trace of an undisturbed water column, 

 consisting of Antarctic Surface Water, Winter Water, and Shelf Water. 



Intrusion of Antarctic Circumpolar Water provides the basis for the 

 formation of a secondary sound channel. The changes in the water column 

 caused by Antarctic Circumpolar Water and resulting change in the sound 

 velocity structure are illustrated in Figures 24 through 26. Because 

 Antarctic Circumpolar Water is present at nearly all stations in varying 

 amounts, it may be assumed that most sound velocity traces will be modi- 

 fied and show traces of a secondary sound channel. 



A statistical evaluation of sound velocities compiled from data taken 

 on DEEP FREEZE 63 shows that a surface sound channel existed at 73% of the 

 stations. The surface configuration of sound velocity structure undoubt- 

 ably changes during the season and from year to year. For example, an 

 analysis of DEEP FREEZE 64 data showed only 19% of the stations had a 

 surface channel. 



Of all stations with sound channels, the following characteristics 

 of the channel were noted: mean sound velocity was 1444.3 m/sec (4738.5 

 ft/sec) with a range of 1441.1 to 1448.3 m/sec (4728.0 to 4751.6 ft/sec); 

 mean axial depth was 230 meters (755 feet), with a range of 75 to 400 

 meters (246 to 1,312 feet). 



Correlation of sound channels from a local to regional scale is dif- 

 ficult because of the transient character of the parameters measured. 

 Formation of sea ice results in an increase of salinity at the surface. 

 The resultant density inversion at the surface causes convective mixing 

 to occur, thus modifying the entire water column. Melt water from the 

 ice cover dilutes the surface water during the summer season, causing 

 localized sound channels at the surface. 



Data are available only for the summer and early fall of the austral 

 year. Lack of observations during the other seasons prevents a full 

 understanding of the sound velocity structure of the Ross Sea. 



29 



